Identification and Characterization of a Series of Nucleoside

1606

Chem. Res. Toxicol. 2003, 16, 1606-1615

Identification and Characterization of a Series of Nucleoside Adducts Formed by the Reaction of 2′-Deoxyguanosine and 1,2,3,4-Diepoxybutane under Physiological Conditions Xin-Yu Zhang and Adnan A. Elfarra* Department of Comparative Biosciences and the Molecular and Environmental Toxicology Center, University of Wisconsin-Madison, Madison, Wisconsin 53706 Received June 30, 2003

The carcinogenicity of 1,3-butadiene (BD) has been attributed to its in vivo metabolites, 3,4epoxy-1-butene (EB) and 1,2,3,4-diepoxybutane (DEB). In this study, DEB was demonstrated to react with 2′-deoxyguanosine (dG) under in vitro physiological conditions (pH 7.4, 37 °C) to yield several pairs of diastereomeric adducts, including N-(2-hydroxy-1-oxiranylethyl)-2′deoxyguanosine (P4-1 and P4-2), 7,8-dihydroxy-3-(2-deoxy-β-D-erythro-pentofuranosyl)3,5,6,7,8,9-hexahydro-1,3-diazepino[1,2-a]purin-11(11H)one (P6), 1-(2-hydroxy-2-oxiranylethyl)2′-deoxyguanosine (P8 and P9), 1-[3-chloro-2-hydroxy-1-(hydroxymethyl)propyl]-2′-deoxyguanosine (1AP9 and 2AP9), and 4,8-dihydroxy-1-(2-deoxy-β-D-erythro-pentofuranosyl)-9-hydroxymethyl6,7,8,9-tetrahydro-1H-pyrimido[2,1-b]purinium ion (1BP4 and 2BP4). The 7-alkylation dG adducts (P5 and P5′) were not characterized directly by NMR spectrometry because of their instability. However, their formula weights were determined to be 354, and their acid hydrolysis products were characterized as 2-amino-7-(3-chloro-2,4-dihydroxybutyl)-1,7-dihydro-6H-purin6-one (H3), consistent with the structures of P5 and P5′ being diastereomers of 6-oxo-2-amino9-(2-deoxy-β-D-erythro-pentofuranosyl)-7-(2-hydroxy-2-oxiranylethyl)-6,9-dihydro-1H-purinium ion. Time-course experiments indicated that alkaline pH and/or high DEB:dG molar ratios made the reactions faster without changing the adduct profile. The adducts were detected in the following chronological order: 7- (P5 and P5′), 1- (P8 and P9), N2- (P4-1 and P4-2), and P6. Whereas P4-1, P4-2, and P6 appeared stable during the courses of the reactions, P5, P5′, P8, and P9 were labile and completely decomposed by the time dG was fully consumed. These results may contribute to a better understanding of the chemical reactivity and strong mutagenicity and carcinogenicity of DEB.

Introduction BD1

is a colorless gas that is used primarily in the industrial production of synthetic rubber and plastic. Low levels of BD (0.5-10 ppb) have been detected in ambient air in urban locations (1), and environmental sources include cigarette smoke, gasoline formulations, automobile exhaust, and incinerating plant exhaust (2, 3). In the United States, approximately 52 000 workers are potentially exposed to BD every year (4). Several studies have established a link between longterm inhalation exposure to BD and the development of tumors at multiple sites in mice and rats (5-7). Epidemiological studies demonstrate increased incidences of lymphatic and hematopoietic cancers in industrial workers exposed to BD (8-11). On the basis of both epidemiological and mechanistic data indicating a causal relationship between occupational exposure to BD and increased mortality from lymphatic and/or hematopoietic 1 Abbreviations: BD, 1,3-butadiene; COSY, correlation spectroscopy; DEB, 1,2,3,4-diepoxybutane; DEPT, distortionless enhancement by polarization transfer; dG, 2′-deoxyguanosine; DMSO, dimethyl sulfoxide; EB, 3,4-epoxy-1-butene; ESI, electrospray ionization; HMBC, heteronuclear shift correlations via multiple bond connectivities; HMQC, heteronuclear multiple quantum coherence; MS-MS, tandem mass spectrometry; TFA, trifluoroacetic acid; TMS, tetramethylsilane; TSP, sodium 3-(trimethylsilyl)propionate-2,2,3,3-d4.

cancers, BD is listed by the EPA as a hazardous component of air pollution and has recently been upgraded to “known to be a human carcinogen” by the U.S. Department of Health and Human Services National Toxicology Program (12). BD is bioactivated by cytochrome P450- or myeloperoxidase-mediated oxidation to produce EB,2 also called butadiene monoxide (13-15). Further P450-mediated oxidation of EB leads to the formation of DEB2 (16). Whereas both of these epoxides are direct mutagens, the in vitro mutagenicity of DEB is approximately 100 times higher than that of EB (17-20). The high DEB mutagenicity is attributed to both its high DNA reactivity and its cross-linking ability; both interstrand and intrastrand DNA cross-links have been implicated in DEB mutagenesis (21, 22). Nonetheless, the chemical and molecular basis of the various biological effects of EB and DEB are not fully understood. Our laboratory has studied the reactions between EB and the four nucleoside bases of DNA under in vitro physiological conditions and structurally characterized 2 According to the Chemical Abstracts (CA) nomenclature rules, EB and DEB should be named ethenyloxirane and 2,2′-bioxirane, respectively. However, to maintain consistency with previous literature, the names of EB and DEB are still used in this article.

10.1021/tx0341355 CCC: $25.00 © 2003 American Chemical Society Published on Web 11/15/2003

Reaction of 2′-Deoxyguanosine and 1,2,3,4-Diepoxybutane

Chem. Res. Toxicol., Vol. 16, No. 12, 2003 1607

Scheme 1. Structures of All of the dG Adducts and the Hydrolysis Product (H3) Characterized in This Paper

Materials. Racemic DEB, TFA, deuterium oxide, and DMSOd6 were obtained from Sigma-Aldrich Chemical Co. (Milwaukee, WI). dG (anhydrous or monohydrate) was obtained from SigmaAldrich Chemical Co., ICN Biomedicals, Inc. (Aurora, OH), or Research Organics (Cleveland, OH) depending upon its availability during the course of our studies. HPLC grade acetonitrile was purchased from EM Science (Gibbstown, NJ). NMR supplies were purchased from Wilmad Glass Co. (Buena, NJ). Sephadex LH-20 resin was purchased from Pharmacia LKB Biotechnology (Sweden). Mobile phases were prepared using acetonitrile and water, and the pH was adjusted to 2.5 with 20% (v/v) TFA if necessary. Phosphate buffers (100 mM) containing 100 mM KCl were prepared with KH2PO4 for pH 7.4 buffer or K2HPO4 for pH 10 buffer using KOH to adjust the pH.

29 nucleoside adducts (23-26). In this paper, we describe the reactions between dG and DEB3 under in vitro physiological conditions and the characterization of several nucleoside adducts resulting from alkylation at the 1-, N2-, and 3-positions, including adducts with fused ring systems (P6 and 1BP4/2BP4; Scheme 14). In addition, the previously characterized guanine adduct 2-amino-7(3-chloro-2,4-dihydroxybutyl)-1,7-dihydro-6H-purin-6one (27; H3) was also detected. Although several groups have previously reported on the adducts of guanosine or dG with DEB (27-31), only 7-alkylation guanine products [N-7-(2′,3′,4′-trihydroxybut-1′-yl)guanine, N-7-(2′,4′dihydroxy-3′-chlorobut-1′-yl)guanine (H3), and N-7-(2′,3′dihydroxy-4′-acetoxybut-1′-yl)guanine (27)] were isolated and characterized, possibly because either reactions were carried out in glacial acetic acid or the reaction products were subjected to acid or thermal hydrolysis before product separation and identification.

Experimental Procedures Caution: DEB is a known mutagen and carcinogen and must be handled using proper safety measures. 3 DEB has three isomers, R/S or S/R (meso), R/R, and S/S. In this study, racemic (DL) DEB was used. 4 All compounds shown in Scheme 1 were named according to the CA nomenclature rules. The purine ring, the deoxyribose ring, and the side chain were numbered as shown on the structure of P4-1/ P4-2. Because of the new fused ring systems in P6 and 1BP4/2BP4, these compounds were numbered differently as shown outside the rings of P6 and 1BP4/2BP4. However, to facilitate discussing and comparing the structures of P6 and 1BP4/2BP4 with those of the other adducts, the DEB moieties were also numbered (from 1′′ to 4′′) as shown inside the rings of P6 and 1BP4/2BP4.

Instruments and Methods. Mass spectra were obtained on an MDS Sciex API 365 LC/MS/MS triple quadrupole ESI mass spectrometer. NMR spectra were recorded on a Bruker Instruments DMX-400 Avance console, 9.4 T wide bore magnet NMR spectrometer (400 MHz), or a Bruker Instruments DMX-500 Avance console, 11.74 T standard bore magnet NMR spectrometer (500 MHz). All 1H NMR spectra, including regular onedimensional (1D) spectra and two-dimensional (2D) COSY spectra, were obtained in DMSO-d6 with TMS as the internal standard. For some compounds, 1H NMR spectra were also recorded in D2O, and in these cases, TSP was used as the internal standard. 13C NMR, 13C DEPT 135, 1H-13C HMQC, and 1H-13C HMBC spectra were all recorded in DMSO-d6, and the 13C signal of DMSO at 39.51 ppm was used as the internal standard. The chemical shifts (δ) and the coupling constants (J) are reported in ppm and Hz, respectively. To determine the structures of the products, 1H NMR and COSY spectra in DMSO-d6 were routinely recorded. For some products, 1H NMR and COSY spectra in D2O, 13C NMR, 13C DEPT 135, HMQC, and HMBC spectra were also obtained for further structural characterization. Initial separation of the dG-DEB products was carried out on a Sephadex LH-20 column (∼900 mm × 25 mm) in a cooled chamber (4 °C). The samples were eluted using pure water, the eluate was monitored using a UV detector at 254 nm, and the fractions were collected manually. HPLC separation was performed on a Beckman Ultrasphere 5 µm ODS reverse phase analytical column (250 mm × 4.6 mm) with acidic mobile phase, using a Beckman gradient-controlled HPLC system (Irvine, CA) equipped with a Beckman diode array detector (model 168). The chromatograms were recorded simultaneously at 220 and 260 nm. A Beckman autosampler (model 507e) was used to make injections. A linear gradient program was used, starting at 2 min from 0% pump B to 10% pump B over 1 min [pump A, 1% (v/v) acetonitrile at pH 2.5; pump B, 10% (v/v) acetonitrile at pH 2.5] at a flow rate of 1 mL/min and then at 10 min from 10 to 20% pump B over 1 min, at 11 min from 20 to 30% pump B over 9 min, at 20 min from 30 to 65% pump B over 1 min, at 25 min from 65 to 100% pump B over 12 min, and finally at 37 min from 100 to 0% pump B over 4 min and stopped at 43 min. Preparative separation and purification were carried out on a Beckman Ultrasphere 10 µm ODS reverse phase preparative column (150 mm × 21.2 mm) with acidic or pH-unadjusted mobile phase. A linear gradient program was used, which started at 10 min from 20% pump B to 50% pump B over 20 min [pump A, 1% (v/v) acetonitrile at pH 2.5; pump B, 25% (v/v) acetonitrile at pH 2.5] at a flow rate of 5 mL/min and then at 35 min from 50 to 62% pump B over 3 min and finally at 39 min from 62 to 20% pump B over 1 min and stopped at 40 min. In some cases, a shortened gradient program or other similar programs were also used. When pH-unadjusted mobile phase was needed for preparative separation, the same concentration of acetonitrile without adjusting pH was used on pumps A and B and all gradient programs were unchanged. Samples were injected manually. Fractions were collected using a Gilson fraction collector (model 202) with all collecting beakers kept on ice. The collected fractions were lyophilized on a Labconco lyophilizer (Kansas City, MO).

1608 Chem. Res. Toxicol., Vol. 16, No. 12, 2003

Zhang and Elfarra

Formation, Separation, and Purification of DEB-dG Adducts. dG (260 mg, 0.97 mmol) was dissolved in 100 mL of pH 7.4 phosphate buffer with sonication and heating. DEB (2.26 mL, 29.2 mmol, DEB:dG ) 30:1) was added to the solution. The mixture was incubated at 37 °C with gentle shaking for 4 h, cooled to room temperature, brought to pH 7 using 1 M HCl, and then extracted six times using diethyl ether to remove excess DEB. Nitrogen was bubbled through the solution to purge the remaining ether. The solution was then lyophilized, and the resulting solid was added to ∼15 mL of water. The turbid solution was centrifuged for 5 min (20 800g). Aliquots of the supernatant (2-3 mL) were loaded on the Sephadex LH-20 column for initial separation. Eight or nine peaks were observed on the chromatogram, and the corresponding fractions were collected, analyzed by HPLC, and lyophilized. The seventh or eighth fraction was injected onto a preparative HPLC column to purify P8, P9, 1AP9, and 2AP9 using a pH-unadjusted mobile phase. P5 and P5′ were prepared at pH 7.4 for 1 h using a DEB:dG molar ratio of 80:1 as described above. However, only three peaks were observed from the chromatogram of the reaction mixture on the Sephadex LH-20 column and the corresponding fractions were collected and lyophilyzed. The first fraction, which mostly consisted of P5 and P5′, was purified by HPLC using an acidic (pH 2.5) mobile phase. The reaction between DEB and dG was also carried out at pH 10 and 37 °C for 3 h using a DEB:dG molar ratio of 40:1. The products P4-1, P4-2, P6, 1BP4, and 2BP4 were purified by HPLC from the reaction mixture as described above. For P6, an acidic mobile phase (pH 2.5) was used whereas a pHunadjusted mobile phase was used to elute P4-1, P4-2, because these adducts were not resolved under acidic conditions. Purification of Hydrolyzed Product H3. To determine if 7-alkylation dG adducts were formed during the reaction of dG with DEB, the acid hydrolysis of the reaction mixture was studied. In these experiments, dG (130 mg, 0.456 mmol) was dissolved in 50 mL of pH 7.4 buffer by means of sonication and heating and DEB (1.06 mL, 13.7 mmol, DEB:dG ) 30:1) was added. The reaction mixture was incubated at 37 °C with gentle shaking for 4 h, cooled to room temperature, neutralized by adding concentrated HCl (∼220 µL), and extracted six times with diethyl ether to remove the excess of DEB. To acid hydrolyze the reaction products, concentrated HCl (6.8 mL) was added. Then, the solution was heated in a boiling water bath for 1 h, cooled to room temperature, neutralized with concentrated ammonium hydroxide, and lyophilized. The resulting solid was suspended in 10 mL of water, and the turbid solution was filtered. Half of the solution was loaded on the Sephadex LH-20 column, and 7-8 fractions were collected, analyzed by HPLC, and lyophilized. The sixth or seventh fraction was further separated on a preparative HPLC column with an acidic mobile phase to obtain H3. Time-Course Experiments and Product Characterization before and after Acid Hydrolysis. A typical time-course experiment was performed as follows: dG (7.6 mg, 0.028 mmol) was dissolved in 3.0 mL of buffer (pH 7.4 or 10) using sonication and heating. DEB (88 µL, 1.14 mmol, DEB:dG ) 40:1) was added, and the solution was incubated in a 37 °C water bath shaker. Aliquots (250 µL), withdrawn every 30 min, were neutralized with 1 M HCl, supplemented with water to maintain identical volume, extracted six times with diethyl ether, filtered through a 0.2 µm membrane filter, and then injected (20 µL) onto an analytical HPLC column for analysis. For the time-course experiments with the DEB:dG molar ratios of 80:1 and 40:1 at pH 7.4 and for the experiments at pH 10, samples were withdrawn every 30 min. For experiments where the DEB:dG molar ratios were lower than 40:1 at pH 7.4, the sampling time interval was 1 h because the reactions were slow. For time-course experiments and product characterization after acid hydrolysis, the procedures were very similar to those

Figure 1. Typical HPLC chromatogram of the reaction mixture prepared at pH 7.4 for 2 h (37 °C) (DEB:dG ) 80:1, analytical column; mobile phase: 1% acetonitrile at pH 2.5 on pump A, 10% acetonitrile at pH 2.5 on pump B; detection wavelength, 220 nm). described above followed by the acid hydrolysis step. This was carried out by adding concentrated HCl (31 µL) to each reaction mixture aliquot (250 µL), heating in a boiling water bath for 1 h, cooling to room temperature, and then neutralizing with concentrated ammonium hydroxide. Analyses of the neutralized aliquots were then performed on a preparative HPLC column (injection volume, 250 µL).

Results Reaction of dG and DEB. The reaction of dG and racemic DEB under in vitro physiological conditions yielded many products detected by HPLC. Figure 1 shows a typical analytical HPLC chromatogram. Ten major peaks were consistently detected, and the compounds that they represent were designated P1, P2, P3, P4, P5, P6, P7, P8, P9, and P10 in the order of their retention times. Whereas P1 and P7 exhibited similar UV absorption spectra with two peaks at 217 and 273 nm for P1 and 219 and 274 nm for P7, UV absorption spectra of the other products (except P10) were similar to that of dG (P2) but with a small or moderate bathochromic shift of λmax (P2, 254 nm; P3, 259 nm; P4, 260 nm; P5, 258 nm; P6, 262 nm; P8 and P9, 258 nm). On the other hand, the UV absorption spectrum of P10 exhibited three peaks at 212, 247, and 284 nm. Time-Course Experiments. Time-course experiments showed that all adducts underwent formation followed by decomposition (Figure 2). P5 was the major product first detected, followed by P8 and P9 and then P4 and P6. These products also decomposed in the same order. The P5 peak completely disappeared within a few hours. P8 and P9 remained detectable slightly longer, whereas P4 and P6 decomposed to a much lesser extent than P5, P8, and P9 during the duration of the reaction. The time-course experiments were performed at either pH 7.4 or pH 10 and at different DEB:dG molar ratios (10:1, 20:1, 30:1, 40:1, and 80:1 for pH 7.4 and 40:1 and 80:1 for pH 10). The qualitative comparison of the results, as shown in Figure 2, revealed that alkaline pH or high DEB:dG molar ratios made the reactions faster, whereas neutral pH or low DEB:dG molar ratios slowed the reactions. At pH 7.4, the reactions with a DEB:dG molar ratio of 10:1 appeared slow as 20-30% of the original dG remained after the incubation with DEB for 18 h,


Figure 2. Time-course of the reactions between dG and DEB at different pH values and at different DEB:dG molar ratios (all analyses were carried out on the analytical column except the time-course for formation of H3, which was carried out on the preparative HPLC column).

whereas dG was completely consumed within 4 h when the DEB:dG molar ratio was 80:1. However, alteration of the pH or the DEB:dG molar ratio did not seem to affect the product profile; the reaction at pH 7.4 with a DEB:dG molar ratio of 10:1 produced all products detected at the DEB:dG molar ratio of 80:1, and similar product molar ratios were observed at the two reaction conditions. Characterization of P4. Product P4 was initially detected at the middle stage of the reaction, and its amount then increased over time (Figure 2). It was one of two final products observed when the reaction was carried out until dG was all consumed. This product appeared as a sharp peak on the HPLC chromatogram (Figure 1). However, when a pH-unadjusted mobile phase was used instead of an acidic mobile phase (pH 2.5), the peak representing P4 was resolved into two slightly overlapping peaks, designated as P4-1 and P4-2 (data not shown). The two compounds were collected, further purified, and characterized as a pair of diastereomers because of their identical molecular weights and nearly identical 1H NMR spectra. The ESI/mass spectra of the purified P4-1 and P4-2 showed protonated molecular ion peaks at m/z 354, suggesting that a DEB molecule with an intact epoxy ring had added to dG. The signals of the protons on deoxyribose ring and at the 8-position in the 1H NMR spectra (see Supporting Information) could be assigned based on the coupling information obtained from the COSY spectra. Alkylation of the purine moiety of dG had little effect on the chemical shifts of the signals of the protons on the deoxyribose ring (in fact, this is true for all of the adducts characterized in this paper; see Table 1). In addition, because the signals of the protons at the


8-position appeared at 7.90 and 7.91 ppm (Table 1) with little deviation from the absorption position of the same proton of dG, the possibility that P4-1 and P4-2 were 7-alkylation adducts was excluded. The signal of the two protons on the exocyclic amino group in dG appeared at 6.5 ppm (32). Because no single peak representing two protons in their 1H NMR spectra could be found, P4-1 and P4-2 were identified as Nalkylation products. The peak representing one proton at 7.64 or 7.66 ppm did not couple to other peaks and thus was assigned as the signal of the proton at the 1-position. It was interesting that the chemical shift of the signal of the proton at the 1-position deviated far from that of the same proton in dG (10.9 ppm). The side chain was determined to be 2-substituted instead of 1-substituted based on the 13C NMR and 13C DEPT 135 data, because the 13C DEPT 135 experiment of P4-2 indicated that among the four carbons of the side chain only the signal of a methylene carbon appeared at high field (41.76 ppm, Table 2) whereas the peak of the other methylene carbon appeared at low field (62.59 ppm). Clearly, the methylene carbon at low field had a strong electronwithdrawing group attached, which in this case was a hydroxyl group. The methylene carbon at high field was thus the one in the intact epoxy ring. The 13C NMR data of P4-2 also supported the structural assignments that P4-1 and P4-2 are N-alkylation products (Table 2). On the basis of the conclusions drawn from the 13C NMR studies of other alkylation nucleosides (33, 34), the possibility of alkylation at the O6- or 7-position was excluded, because O-alkylation would cause significant perturbation of all purine carbon signals relative to dG (>2.0 ppm), whereas modification at the 7-position would result in a downfield shift of the signals of the carbons at the 2- and 6-positions (>3.0 ppm) in addition to a large upfield shift in the signal of the carbon at the 5-position (∼9 ppm) relative to the corresponding carbon signals of dG. Modification at the 1-position would result in an upfield shift in the signal of the carbon at the 4-position (∼2-3 ppm) and a small downfield shift in the signal of the carbon at the 2-position (∼0.5-0.6 ppm) (33, 34). P4-1 and P4-2 could not be 1-alkylation adducts because P4-2 showed a significant upfield shift in the carbon absorption at the 2-position (2.34 ppm) and moderate and small upfield shifts in the signals of the carbons at the 4- and 5-positions (1.38 and 0.85 ppm, respectively). Therefore, the only possibility remaining was N-alkylation, although the pattern of the shifts in P4-2 was not consistent with previous reports that suggested that modifications of the exocyclic amino group of guanosine or dG produce little chemical shift variation on the signals of all purine carbons (33-35). Nonetheless, all of the spectroscopic data of P4-1 and P4-2 are consistent with the structure N-(2-hydroxy-1-oxiranylethyl)-2′-deoxyguanosine (Scheme 1). Characterization of P6. Similar to P4, P6 was also detected at the middle stage of the reaction and was one of the most stable products of the reaction (Figure 2). However, in contrast to P4, the HPLC peak of P6 remained unresolved in either acidic or pH-unadjusted mobile phases. The protonated molecular ion peak of P6 appeared at m/z 354, identical to both P4-1 and P4-2. On the basis of the same reasoning used with P4-1 and P4-2, P6 was also first identified as an N-alkylation product. The 1H NMR spectrum showed a signal representing one proton


Zhang and Elfarra

Table 1. 1H NMR Data of dG-DEB Adducts (ppm) in DMSO-d6 P4-1

P4-2

P6

P8

P9

H-C3′ H-C4′ H-C5′ H-C1′′

7.90 (s, 1H) 6.10 (dd, 1H) 2.19 (m, 1H) 2.52 (m, 1H)a 4.34 (s, 1H) 3.81 (s, 1H) 3.55 (m, 2H)b 3.32 (m, 2H)c

7.91 (s, 1H) 6.11 (dd, 1H) 2.18 (m, 1H) 2.52 (m, 1H)a 4.34 (m, 1H) 3.81 (m, 1H) 3.55 (m, 2H)b 3.34 (m, 2H)c

H-C2′′ H-C3′′ H-C4′′

4.12 (s, 1H) 4.26 (m, 1H) 3.55 (m, 2H)b

4.13 (s, 1H) 4.28 (m, 1H) 3.47 (m, 2H)b

8.02 (d, 1H) 6.13 (t, 1H) 2.21 (m, 1H) 2.52 (m, 1H)a 4.34 (s, 1H)b 3.81 (m, 1H) 3.52 (m, 2H)b 3.04 (m, 1H) 3.32 (m, 1H) 3.54 (m, 1H) 3.54 (m, 1H) 3.76 (m, 1H)b 4.37 (m, 1H)b

7.94 (s, 1H) 6.13 (dd, 1H) 2.19 (m, 1H) 2.52 (m, 1H)a 4.34 (s, 1H) 3.81 (m, 1H) 3.53 (m, 2H)b 2.69 (t, 1H) 3.02 (dd, 1H) 3.55 (m, 1H) 3.55 (m, 1H) 4.00 (dd, 1H) 4.15 (dd, 1H)

7.95 (s, 1H) 6.13 (dd, 1H) 2.19 (m, 1H) 2.52 (m, 1H)a 4.34 (s, 1H) 3.80 (m, 1H) 3.52 (m, 2H)b 2.68 (t, 1H) 3.02 (m, 1H) 3.55 (m, 1H) 3.54 (m, 1H) 3.99 (dd, 1H) 4.16 (dd, 1H)

H-N1 H-N2/H2-N2 OH-C3′ OH-C5′ OH-C1′′/C2′′ OH-C3′′

7.64 (s, 1H) 5.33 (s, 1H) 5.29 (s, 1H) 4.97 (br s, 1H) 4.97 (br s, 1H)

7.66 (d, 1H) 5.32 (d, 1H) 5.28 (d, 1H) 4.96 (br s, 1H) 4.96 (br s, 1H)

6.85 (d, 1H) 5.31 (d, 1H) 4.96 (t, 1H) 5.34 (t, 1H)d 5.19 (d, 1H)d

6.90 (s, 2H) 5.27 (s, 1H) 4.94 (t, 1H) 5.68 (s, 1H)

6.91 (s, 2H) 5.28 (s, 1H) 4.94 (s, 1H) 5.69 (br s, 1H)

H-C8 H-C1′ H-C2′

H-C8 H-C1′ H-C2′ H-C3′ H-C4′ H-C5′ H-C1′′ H-C2′′ H-C3′′ H-C4′′ H-N1 H-N2/H2-N2 OH-C3′ OH-C5′ OH-C1′′ OH-C2′′ OH-C3′′ OH-C4′′

1BP4

2BP4

1AP9

2AP9

7.88 (s, 1H) 6.09 (t, 1H) 2.18 (m, 1H) 2.52 (m, 1H)a 4.33 (d, 1H) 3.80 (m, 1H) 3.52 (m, 2H)b 3.21 (d, 1H) 3.49 (m, 1H)b 4.25 (s, 1H) 4.71 (m, 1H) 3.37 (m, 1H)c 3.51 (m, 1H)b 3.51 (1H)b 7.63 (d, 1H) 5.27 (d, 1H) 4.94 (t, 1H)

7.88 (s, 1H) 6.10 (t, 1H) 2.17 (m, 1H) 2.52 (m, 1H)a 4.33 (s, 1H) 3.80 (m, 1H) 3.52 (m, 2H)b 3.21 (d, 1H) 3.47 (m, 1H)b 4.25 (s, 1H) 4.71 (m, 1H) 3.35 (m, 1H)c 3.50 (m, 1H)b 3.51 (1H)b 7.63 (d, 1H) 5.27 (d, 1H) 4.95 (t, 1H)

7.95 (s, 1H) 6.13 (dd, 1H) 2.19 (m, 1H) 2.52 (m, 1H)a 4.34 (s, 1H) 3.81 (m, 1H) 3.53 (m, 2H) 3.62 (dd, 1H) 3.69 (dd, 1H) 3.54 (m, 1H)b 3.91 (s, 1H) 3.98 (t, 1H) 4.05 (d, 1H)

7.96 (s, 1H) 6.13 (t, 1H) 2.18 (m, 1H) 2.52 (m, 1H)a 4.34 (m, 1H) 3.81 (m, 1H) 3.52 (m, 2H) 3.63 (dd, 1H) 3.69 (dd, 1H) 3.53 (m, 1H)b 3.91 (s, 1H) 3.98 (t, 1H) 4.06 (d, 1H)

6.82 (s, 2H) 5.28 (d, 1H) 4.95 (t, 1H) 5.46 (d, 1H)

6.81 (s, 2H) 5.27 (d, 1H) 4.94 (t, 1H) 5.44 (d, 1H)

5.24 (d, 1H)

5.24 (d, 1H) 5.34 (d, 1H)

5.36 (d, 1H)

5.08 (t, 1H)

5.08 (t, 1H)

a The values of the chemical shifts might not be accurate because the solvent peak (DMSO) appeared at 2.50 ppm. b The values of the chemical shifts might not be accurate because the peaks overlapped with other peaks. c The peaks were partly covered by the strong signal of HOD (appeared around 3.3); therefore, the values of the chemical shifts might not be accurate. d These assignments are tentative.

Table 2. 13C Chemical Shift Data of dG-DEB Adducts (ppm) and Unmodified DG position 2 4 5 6 8 1′ 2′ 3′ 4′ 5′ 1′′ 2′′ 3′′ 4′′

P4-2

P6

P8

dG (32)

151.13 (-2.34)a 149.31 (-1.38) 115.62 (-0.85) 156.43 (-0.15) 135.33 (0.23) 82.43 (0.02) b 70.77 (0.18) 87.54 (0.13) 61.74 (0.18) 62.59d 57.22 or 59.84 57.22 or 59.84 41.76

157.08 (3.61) 148.28 (-2.41) 117.35 (0.88) 157.61 (1.03) 136.15 (1.05) 82.34 (-0.07) 39.49c (0.05) 70.71 (0.12) 87.60 (0.19) 61.67 (0.11) 46.28 72.05e 69.88e 44.37

154.61 (1.14) 148.94 (-1.75) 116.00 (-0.47) 156.72 (0.14) 135.65 (0.55) 82.31(-0.10) 39.50c (0.06) 70.74 (0.15) 87.55 (0.14) 61.71 (0.15) 43.52e 69.88 53.41 44.52e

153.47 150.69 116.47 156.58 135.10 82.41 39.44 70.59 87.41 61.56

a The data in the parentheses are the difference between the chemical shifts of the adducts and the unmodified dG. b The signal was covered by that of the solvent (DMSO). c The data were obtained by inverted 13C DEPT 135 spectra. d The methylene and methine carbons were distinguished by 13C DEPT 135 experiment. e These assignments are tentative.

at 6.85 ppm (Figure 3A), which disappeared in D2O (data not shown) but strongly coupled to the signals of two other protons absorbing at 3.04 and 3.32 ppm (see the

COSY spectrum, Figure 3B), thus indicating that this was the signal of the proton on the exocyclic amino group. The HMQC spectrum confirmed that the two protons absorbing at 3.04 and 3.32 ppm were the hydrogen atoms on the same carbon atom (Figure 3C), whose chemical shift appeared at 46.28 ppm (Table 2). These results clearly indicate an -NHCH2- structure. The three peaks above 6 ppm in the 1H NMR spectrum were identified as the signals of the protons at the 8-position (8.02 ppm), on the exocyclic amino group (6.85 ppm), and at the 1′-position (6.13 ppm), respectively. In addition, there were four peaks between 4.9 and 5.4 ppm, which all disappeared in D2O (data not shown), indicating that the four peaks must be the signals of the protons on oxygen or nitrogen atoms. The coupling information obtained from the COSY spectrum showed that two of the four peaks were the signals of two hydroxyl group protons on the deoxyribose ring: the signal at 5.31 ppm was identified as that of the 3′-hydroxyl proton and the signal at 4.96 ppm was that of the 5′-hydroxyl proton. The other two peaks (at 5.19 and 5.34 ppm) showed strong couplings with the signals appearing around 3.53 ppm, which were actually a group of peaks representing four protons. The HMQC spectrum, combined with the


Figure 3. (A) 1H NMR, (B) 1H COSY, (C) HMQC, and (D) HMBC spectra of P6 in DMSO-d6. 13

C NMR and 13C DEPT 135 data, indicated that the four protons consisted of the two protons at the 5′-position and two methine protons of the DEB moiety. Clearly, the two proton signals at 5.19 and 5.34 ppm coupled to the two methine proton signals, indicating that the former two protons were those of the two hydroxyl groups of the DEB moiety. This assignment of the proton signals at 5.19 and 5.34 ppm indicated that the DEB moiety must have formed a fused seven-membered ring with the pyrimidine ring. The 13C NMR data, together with the 13C DEPT 135 experiment, unambiguously indicated that among the four carbons of the DEB moiety, two methylene carbons absorbed at high field (44.37 and 46.28 ppm) and two methine carbons at low field (69.88 and 72.05 ppm). These 13C chemical shift data are consistent with the seven-membered ring structure. Because a seven-membered ring is not as readily formed as a five- or sixmembered ring, additional supporting evidence, i.e., the long-range 1H-13C connectivities observed by the HMBC experiment (Figure 3D), was obtained. In the HMBC spectrum, the expected strong coupling between the signals of the protons on the methylene carbon of the DEB moiety adjacent to the exocyclic amino group (3.04 and 3.32 ppm, i.e., the two protons at the 1′′-position; see Scheme 1) and the signal of the carbon at the 2-position (157.1 ppm) was observed. Similarly, strong coupling between the signal of one of the protons on the other methylene carbon of the DEB moiety (∼4.4 ppm, i.e., one of the two protons at the 4′′-position) and the


signal of the carbon at the 2- or 6-position (157.1 or 157.6 ppm) was also observed; because the resolution of the carbon resonance in this experiment was only 0.5 ppm, it was impossible to distinguish the carbon absorption at the 2- and 6-positions. Consequently, the structure of the seven-membered ring is firmly established and P6 is structurally characterized as 7,8-dihydroxy-3-(2-deoxyβ-D-erythro-pentofuranosyl)-3,5,6,7,8,9-hexahydro-1,3-diazepino[1,2-a]purin-11(11H)one (Scheme 1). Characterization of P8 and P9. Products P8 and P9, represented by two moderately overlapping peaks in the chromatogram of the reaction mixture (Figure 1), exhibited nearly identical 1H NMR spectra and were determined to have molecular weights of 353. In the 1H NMR spectra, a single peak representing two protons appeared at 6.9 ppm (this peak disappeared in D2O, data not shown) and did not couple to any other proton signals, indicating that this peak was the signal of the two protons on the exocyclic amino group. The signal of the proton at the 8-position at 7.94 or 7.95 ppm excluded the possibility of any 7-alkylation having occurred. Furthermore, the signal of the proton at the 1-position was missing; therefore, the two compounds were identified as 1-alkylation products. The identification of 1-alkylation was also supported by the 13C NMR data of P8 (Table 2), in which the signal of the carbon at the 4-position showed a moderate upfield shift (1.75 ppm), consistent with previous observations in the literature (see above) (33, 34). The 13C NMR and the 13C DEPT 135 data of P8 showed that the two methylene carbons absorbed at high field (43.52 and 44.52 ppm), indicating that the side chain was 1-substituted instead of 2-substituted. Therefore, P8 and P9 are characterized as the diastereomeric pair of 1-(2-hydroxy2-oxiranylethyl)-2′-deoxyguanosine (Scheme 1). Characterization of 1BP4 and 2BP4. While purifying P4-1 and P4-2 on HPLC from the crude fraction obtained from the Sephadex LH-20 column, a number of small peaks with shorter retention times were observed. Two of these small peaks were designated as 1BP4 and 2BP4. These peaks were usually not observed in the chromatogram of the reaction mixture because their retention times (14.0 and 14.4 min on the preparative HPLC column, respectively) lie within the large range of the broad bump around P1 (11.3-16.0 min). The ESI/mass spectra of 1BP4 and 2BP4 showed molecular ion peaks at m/z 354, identical with the protonated molecular ion peaks of P4-1, P4-2, P6, P8, and P9. The two compounds are identified as a pair of diastereomers because of their nearly identical 1H NMR spectra. Twenty protons could be counted in their 1H NMR spectra in DMSO-d6, instead of 19 protons seen in the spectra of P4-1, P4-2, P6, P8, and P9. This indicated that 1BP4 and 2BP4 are 3- or 7-alkylation products. However, the signals of the protons at the 8-position appeared at 7.88 ppm (Table 1), a little different from that of dG (7.93 ppm), ruling out the possibility of 7-alkylation. Thus, they must be 3-alkylation adducts. On the other hand, no single peak representing two protons could be found. A peak at 7.63 ppm representing one proton strongly coupled to a doublet peak at 3.2 ppm. The peak also coupled weakly to a peak at ∼3.5 ppm, suggesting that the peak at 7.63 ppm was the signal of the proton on the exocyclic amino group with an -NHCH2- structure. Thus, 1BP4 and 2BP4 are both N- and 3-alkylation products, i.e., the four carbon chain


Figure 4. d6.

1H

NMR and 1H COSY spectra of 2AP9 in DMSO-

of the DEB moiety formed a fused ring involving both the 3-position and the exocyclic amino group. Further analysis of the 1H NMR and COSY spectra of 1BP4 and 2BP4 provided clear evidence that they have the structures represented by 4,8-dihydroxy-1-(2-deoxy-β-D-erythropentofuranosyl)-9-hydroxymethyl-6,7,8,9-tetrahydro-1Hpyrimido[2,1-b]purinium ion (Scheme 1). Characterization of 1AP9 and 2AP9. While purifying P8 and P9 on HPLC from the crude fraction obtained from the Sephadex LH-20 column, 1AP9 and 2AP9 were detected as two small peaks with longer retention times. Their ESI/mass spectra showed protonated molecular ion peaks at m/z 390 and isotope peaks at m/z 392 with intensities consistent with chlorine incorporation. They are determined to be a pair of diastereomers because of their nearly identical 1H NMR spectra. The signals of the protons at the 8-positions at 7.95 and 7.96 ppm were inconsistent with the two compounds being 7-alkylation products (Figure 4). Like P8 and P9, a single peak representing two protons at 6.81 or 6.82 ppm, which did not couple to any other peak, was assigned as the signal of the two protons on the exocyclic amino group. In addition, absence of the proton signal at the 1-position indicates that the two products are 1-alkylation adducts. Only the 1H NMR and COSY spectra were obtained for 1AP9 and 2AP9 due to their limited quantities. Because the two compounds were 1-alkylation products, their spectra were compared with those of P8 and P9 to help determine their structures. In the 1H NMR spectra of P8 and P9, the signals of the two protons at the 1-positions of the side chains were well-separated (2.69 and 3.02 ppm for P8 2.68 and 3.02 ppm for P9), which was probably caused by the hindered rotation of the side chain C-N bond by the exocyclic amino group. A similar situation also occurred on the two protons at the 2-position of the deoxyribose ring (2.19 and 2.52 ppm). This interpretation was further supported by a similar observation when comparing the 1H NMR spectra of P4-1, P42, and P6. The two protons at the 1′′-position of P6

Zhang and Elfarra

absorbed at 3.04 and 3.32 ppm whereas the two protons at the 1′′-position of P4-1 or P4-2 absorbed at 3.31 or 3.34 ppm as a group of peaks. On the basis of the above observations, the side chains of 1AP9 and 2AP9 were determined to be 2-substituted rather than 1-substituted, because in their 1H NMR spectra no two well-separated peaks were observed in the range of 2.6-3.3 ppm. The 1H NMR spectrum of 2AP9 showed the proton signals of four hydroxyl groups: the peaks at 4.94 and 5.27 ppm were identified as the signals of the protons of the two hydroxyl groups on the deoxyribose ring based on the coupling information obtained from the COSY spectrum (Figure 4), and the peaks at 5.36 and 5.44 ppm were those of the two hydroxyl groups on the side chain. The COSY spectrum showed that one of the proton signals of the two hydroxyl groups on the side chain coupled to a group of peaks around 3.65 ppm representing two protons, while the other proton signal coupled to a group of peaks at 3.91 ppm representing one proton. These results suggested that the group of peaks around 3.65 ppm are the signals of the two protons at the 1′′position. The group of peaks around 3.65 ppm coupled strongly to another group of peaks around 3.53 ppm, which was assigned as the signal of the proton at the 2′′-position. However, the coupling between the signals of the protons at the 2′′- and 3′′-positions was not detected on the COSY spectrum, making it difficult to determine which group of peaks was the signal of the proton at the 3′′-position. Nevertheless, only three groups of peaks at 3.91, 3.98, and 4.06 ppm (one proton per group of peaks) were unassigned; thus, they must be the signals of the three protons at the 3′′- and 4′′-positions. Because the three protons form an ABX system, multiplicities were used to distinguish between the protons. The two groups of peaks at 3.98 and 4.06 ppm were thus assigned as the signals of the two protons at the 4′′-position, and the group of peaks at 3.91 ppm was assigned as the signal of the proton at the 3′′-position. Because the proton signal of the hydroxyl group at 5.36 ppm coupled to the signal of the proton at 3.91 ppm, namely, the proton at the 3′′position, the hydroxyl group was finally determined to be at the 3′′-position and the chlorine atom was at the 4′′-position. Thus, 1AP9 and 2AP9 are structurally characterized as 1-[3-chloro-2-hydroxy-1-(hydroxymethyl)propyl]-2′-deoxyguanosine (Scheme 1). Characterization of P5 and P5′. To obtain a high concentration of P5, the reaction was carried out at pH 7.4 and 37 °C for 1 h with a DEB:dG molar ratio of 80:1 based upon the results obtained from the time-course experiments (see Figure 2). When this mixture was analyzed on the preparative HPLC column, there was a shoulder peak in the tail of P2 besides P5, P8, and P9. To isolate P5, the reaction mixture was loaded on the Sephadex LH-20 column and three major peaks were eluted. The corresponding fractions were collected and lyophilyzed. The chromatogram of the first fraction by preparative HPLC showed two major peaks, which were slightly overlapped, nearly equal in height, and exhibited identical UV absorption spectra. The peak with the longer retention time was identified as P5 by comparing the retention time and UV absorption spectrum with those of P5 in the reaction mixture. The other peak was similarly determined to be the shoulder peak of P2 in the chromatogram of the reaction mixture. The earlier eluting product was designated as P5′ and was assumed to be a diastereomer of P5.



Products P5 and P5′ were separated and purified by HPLC in very small quantities to obtain their mass spectra. However, even in very small quantities (thus short purification duration), it was difficult to obtain pure compounds. ESI/mass spectra of P5 and P5′ after being purified only once by HPLC (their purity was 72 and 77% calculated by the peak areas, respectively), which were almost identical, showed two major peaks at m/z 354 and 458 with the peak at m/z 458 having lower intensity. However, the ion at m/z 354 was not the daughter ion of the peak at m/z 458, because no peak at m/z 354 was detected in the fragmentation pattern of the ion at m/z 458 as determined by the MS-MS spectrum. Instead, a peak at m/z 342 was the major fragment of the ion at m/z 458, indicating both P5 and P5′ samples were mixtures of two compounds. The mass spectra of the P5 and P5′ samples after being purified three times by HPLC (purity, 84 and 86%, respectively) still showed the two peaks at m/z 354 and 458, but the intensities of the peaks at m/z 458 were lower than those of the same peaks in the mass spectra of the samples purified only once. These observations indicated that the peak at m/z 354, rather than m/z 458, was the signal of the molecular ion of P5 or P5′. The formula weight indicated that P5 and P5′ were the products of dG and one molecule of DEB. Because P5 and P5′ were far less stable than the 1- (P8 and P9) and N-alkylation adducts (P4-1 and P4-2), it was speculated that P5 and P5′ were 7-alkylation dG adducts. However, P5 and P5′ were labile and decomposed during purification. For this reason, the stable acid hydrolysis products of P5 and P5′ were characterized. The reaction mixture, obtained at pH 7.4 and a DEB: dG molar ratio of 30:1 at 37 °C for 4 h, was hydrolyzed in the presence of 1 M HCl in a boiling water bath for 1 h. The hydrolyzed reaction mixture showed six major peaks in the HPLC chromatogram (data not shown). From the hydrolyzed reaction mixture, a product, which was represented by the third HPLC peak and designated as H3, was purified. H3 was identified as 2-amino-7-(3chloro-2,4-dihydroxybutyl)-1,7-dihydro-6H-purin-6-one (Scheme 1) because the molecular weight of this product was 273 and contained a chlorine atom and the 1H NMR spectrum is consistent with that reported in the literature for this compound (27). H3 was believed to be the hydrolyzed product of P5 because the time dependence curves of P5 correlated well with those of H3 under all three conditions tested (pH 7.4, 40:1; pH 7.4, 80:1; and pH 10, 40:1; see Figure 2). To provide further evidence for this hypothesis, very small quantities of the solutions of P5 and P5′ were individually acid hydrolyzed immediately following HPLC separation. After complete hydrolysis, only one product was found for both P5 and P5′ as determined by HPLC. This product exhibited an identical UV absorption spectrum and coeluted with purified H3. Thus, P5 and P5′ are determined to be a pair of diastereomers of 6-oxo-2amino-9-(2-deoxy-β-D-erythro-pentofuranosyl)-7-(2-hydroxy2-oxiranylethyl)-6,9-dihydro-1H-purinium ion (Scheme 1). Other Products. Attempts to purify P1 and P7 using Sephadex LH-20 chromatography and HPLC were not successful, because the P1 and P7 fractions contained complex mixtures of many products, possibly due to the broad bumps around P1 and P7 (Figure 1). It was easier to separate the two products at very low DEB:dG molar

ratios (e.g., 10:1), because the bumps almost disappeared under these conditions. However, the yields of P1 and P7 also decreased greatly at low DEB:dG molar ratios. The long reaction time and low yields made preparation of the products impractical. As a result, further efforts to separate and purify P1 and P7 were not attempted. P3 and P10 were minor products, and P3 was labile and decomposed noticeably during the purification steps. As a result, P3 and P10 were not characterized.

Discussion The products characterized in this study, with the exception of P6, were determined to be five pairs of diastereomers, formed because DL-DEB was used in the reactions. It is possible that P6 could be a mixture of diastereomers unresolved under the chromatography conditions used in the experiments. Support for this hypothesis comes from the 13C NMR data of P6. The 13C NMR data show unusual perturbation of all of the purine carbon signals relative to dG (Table 2), which is not explained by the general rules of how alkylation at different positions affects the chemical shifts of the purine carbons (33, 34). Further support for this hypothesis is provided by the finding that a similar perturbation is also observed on the 13C NMR data of P4 (i.e., the 1:1 mixture of P4-1 and P4-2) whereas purified P4-1 and P4-2 did not exhibit this perturbation. The chemical shift data for P4 are as follows: C2, 151.56 (-1.91); C4, 148.82 (-1.87); C5, 113.08 (-3.39); C6, 155.59 (-0.99); C8, 135.49 (0.39) (the values in the parentheses are the differences relative to dG). P6 is an unusual product because it possesses a sevenmembered ring, and it is one of the major products. Normally, five- and six-membered rings are kinetically formed more preferably than seven-membered rings (36). In the case of P6, a steric hindrance effect may have played an important role in facilitating the formation of P6 rather than products with five- or six-membered rings. P6 can be formed through an N-alkylation product [N-(2-hydroxy-2-oxiranylethyl)-2′-deoxyguanosine], the undetected regioisomer of P4-1/P4-2. Alternatively, P6 can be formed through a 1-alkylation precursor (i.e., P8/ P9). Among the two possibilities, the latter is probably more reasonable based on the following observations: (i) P8/P9 were detected earlier than P6; (ii) P6 accumulation was slower than P8/P9; (iii) N-(2-hydroxy-2-oxiranylethyl)-2′-deoxyguanosine was not detected as a major product in the reaction mixture; and (iv) P4-1/P4-2 did not seem to undergo any significant five- or six-membered ring formation reactions, although 1BP4/2BP4, which may be formed from the regioisomers of P4-1/P4-2, were detected as minor products. The formation order of the products reveals the reactivity of dG at different positions. The 7-alkylation products are first formed probably because the nitrogen atom at the 7-position has a high electron density with a small steric hindrance, whereas the nitrogen atoms at the 1-position and the exocyclic amino group are expected to carry partial positive charges at neutral conditions (37). However, when the pH of the reaction mixture increases over time due to the formation of the 7-alkylation dG adducts, the imine group at the 1-position gradually deprotonates (pKa ) 9.52; see 37) and thus becomes electron-rich, leading to the formation of 1-alkylation adducts. A further increase of pH will also cause


the exocyclic amino group to deprotonate (pKa > 9.5), which could explain why the N-alkylation products were detected at the middle stage of the reaction. The 7-alkylation adducts are unstable and decompose quickly because they carry a positive charge at the 7-position, making these adducts likely to lose their deoxyribose moiety and to be attacked at the 8-position by nucleophilic agents (37). In the time-course experiments, several DEB:dG molar ratios were examined. The product comparison between the results obtained with the DEB:dG molar ratio of 10:1 and that of the molar ratio of 80:1 (pH 7.4) reveals that product formation at the lower molar ratio is much slower than at the higher molar ratio. Nonetheless, the product profile does not change significantly. That is, the major and minor products, designated for the reaction with a molar ratio of 80:1, are all observed in the reaction with a molar ratio of 10:1, and the approximate molar ratio of the product quantities and the appearing and disappearing order of the products also do not change. Thus, it is reasonable to speculate that the major adducts characterized in this study may be formed in vivo under long-term moderate or low BD exposure conditions. To date, only 7-trihydroxybutyl guanine adducts have been characterized after mice and rats were exposed to BD (30, 31, 38). The results presented in this paper demonstrate that DEB readily reacts with dG under in vitro physiological conditions to yield several major and minor adducts. The adducts also exhibited different stabilities at physiological conditions (pH 7.4 and 37 °C). These results are consistent with our previous findings with the adducts of EB and guanosine (23), although the 7-EB-guanosine adducts seem to be slightly more stable than the 7-DEBdG adducts. As the link between modification of dG in DNA and mutagenesis has been established (39-41), our results may lead to a better understanding of the strong mutagenicity and carcinogenicity of DEB. For example, 1- and N2-dG adducts are expected to disturb the hydrogen bondings between the two helices in DNA, causing base mispairing and mutation. While the N2-dG adducts of EB and 3,4-epoxybutane-1,2-diol were only modestly mutagenic in bacterial systems (42, 43), the intact oxirane ring of P4-1 and P4-2 may make these adducts more mutagenic than the N2-dG adducts of EB and 3,4-epoxybutane-1,2-diol. Furthermore, the mutagenic potential of the EB and 3,4-epoxybutane-1,2-diol adducts in mammalian cells is presently unknown. Studies have demonstrated that for many chemicals, there are significant differences in the rates of removal of dG adducts from DNA depending upon the chemical structures of these adducts and the cell type (44-47). Thus, while the DEB-dG adducts characterized in our study may be enzymatically repaired in various mammalian cells by distinct repair mechanisms, it is likely that the dG adducts with fused ring structures may be more difficult to repair in comparison with the other characterized adducts. Therefore, more studies are warranted to examine the mutagenic potential and repair of the various dG-DEB adducts characterized in this study.

Acknowledgment. This research was supported by NIH Grant ES06841 from the National Institute of Environmental Health Sciences. We thank Dr. Mark Anderson for his great assistance in measuring NMR spectra. NMR studies were carried out at the National

Zhang and Elfarra

Magnetic Resonance Facility at Madison with support from the NIH Biomedical Technology Program (RR02301) and additional equipment funding from the University of Wisconsin, NSF Academic Infrastructure Program (BIR-9214394), NIH Shared Instrumentation Program (RR02781, RR08438), NIH Research Collaborations to Provide 900 MHz NMR Spectroscopy (GM66326), NSF Biological Instrumentation Program (DMB-8415048), and U.S. Department of Agriculture. Supporting Information Available: 1H NMR spectrum of P4-2 in DMSO-d6. This material is available free of charge via the Internet at http://pubs.acs.org.

References (1) Himmelstein, M. W., Acquavella, J. F., Recio, L., Medinsky, M. A., and Bond, J. A. (1997) Toxicology and epidemiology of 1,3butadiene. Crit. Rev. Toxicol. 27, 1-108. (2) Pelz, N., Dempster, N. M., and Shore, P. R. (1990) Analysis of low molecular weight hydrocarbons including 1,3-butadiene in engine exhaust gases using an aluminum oxide porous-layer opentubular fused-silica column. J. Chromatogr. Sci. 28, 230-235. (3) Brunnemann, K. D., Kagan, M. R., Cox, J. E., and Hoffmann, D. (1990) Analysis of 1,3-butadiene and other selected gas-phase components in cigarette mainstream and sidestream smoke by gas chromatography-mass selective detection. Carcinogenesis 11, 1863-1868. (4) NIOSH (1990) National Occupational Exposure Survey (19811983), National Institute for Occupational Safety and Health, Cincinnati, OH. (5) Huff, J. E., Melnick, R. L., Solleveld, H. A., Haseman, J. K., Powers, M., and Miller, R. A. (1985) Multiple organ carcinogenicity of 1,3-butadiene in B6C3F1 mice after 60 weeks of inhalation exposure. Science 227, 548-549. (6) Owen, P. E., Glaister, J. R., Gaunt, I. F., and Pullinger, D. H. (1987) Inhalation toxicity studies with 1,3-butadiene: 3. Two year toxicity/carcinogenicity study in rats. Am. Ind. Hyg. Assoc. J. 48, 407-413. (7) Melnick, R. L., and Huff, J. (1992) 1,3-Butadiene: toxicity and carcinogenicity in laboratory animals and in humans. Rev. Environ. Contam. Toxicol. 124, 111-144. (8) Santos-Burgoa, C., Matanoski, G. M., Zeger, S., and Schwartz, L. (1992) Lymphohematopoietic cancer in styrene-butadiene polymerization workers. Am. J. Epidemiol. 136, 843-854. (9) Divine, B. J., Wendt, J. K., and Hartman, C. M. (1993) Cancer mortality among workers at a butadiene production facility. In Butadiene and Styrene: Assessment of Health Hazards (Sorsa, M., Peltonen, K., Vainio, H., and Hemminki, K., Eds.) IARC Scientific Publications No. 127, pp 345-362, IARC, Lyon. (10) Delzell, E., Sathiakumar, N., Hovinga, M., Macaluso, M., Julian, J., Larson, R., Cole, P., and Muir, D. C. F. (1996) A follow-up study of synthetic rubber workers. Toxicology 113, 182-189. (11) Macaluso, M., Larson, R., Delzell, E., Sathiakumar, N., Hovinga, M., Julian, J., Muir, D., and Cole, P. (1996) Leukemia and cumulative exposure to butadiene, styrene and benzene among workers in the synthetic rubber industry. Toxicology 113, 190202. (12) Department of Health and Human Services (2000) The Ninth Report on Carcinogens, U.S. Public Health Services, National Toxicology Program, Research Triangle Park, NC. (13) Csana´dy, G. A., Guengerich, F. P., and Bond, J. A. (1992) Comparison of the biotransformation of 1,3-butadiene and its metabolite, butadiene monoepoxide, by hepatic and pulmonary tissues from humans, rats and mice. Carcinogenesis 13, 11431153. (14) Duescher, R. J., and Elfarra, A. A. (1992) 1,3-Butadiene oxidation by human myeloperoxidase: role of chloride ion in catalysis of divergent pathways. J. Biol. Chem. 267, 19859-19865. (15) Duescher, R. J., and Elfarra, A. A. (1994) Human liver-microsomes are efficient catalysts of 1,3-butadiene oxidation: evidence for major roles by cytochromes P450 2A6 and 2E1. Arch. Biochem. Biophys. 311, 342-349. (16) Krause, R. J., and Elfarra, A. A. (1997) Oxidation of butadiene monoxide to meso- and (()-diepoxybutane by cDNA-expressed human cytochrome P450s and by mouse, rat, and human liver microsomes: evidence for preferential hydration of meso-diepoxybutane in rat and human microsomes. Arch. Biochem. Biophys. 337, 176-184.



(17) Gervasi, P. G., Citti, L., Del Monte, M., Longo, V., and Benetti, D. (1985) Mutagenicity and chemical reactivity of epoxidic intermediates of the isoprene metabolism and other structurally related compounds. Mutat. Res. 156, 77-82. (18) Tice, R. R., Boucher, R., Luke, C. A., and Shelby, M. D. (1987) Comparative cytogenetic analysis of bone marrow damage induced in male B6C3F1 mice by multiple exposures to gaseous 1,3butadiene. Environ. Mutagen. 9, 235-250. (19) Cochrane, J. E., and Skopek, T. R. (1994) Mutagenicity of butadiene and its epoxide metabolites: I. Mutagenic potential of 1,2-epoxybutene, 1,2,3,4-diepoxybutane and 3,4-epoxy-1,2-butanediol in cultured human lymphoblasts. Carcinogenesis 15, 713717. (20) Cochrane, J. E., and Skopek, T. R. (1994) Mutagenicity of butadiene and its epoxide metabolites: II. Mutational spectra of butadiene, 1,2-epoxybutene and diepoxybutane at the hprt locus in splenic T cells from exposed B6C3F1 mice. Carcinogenesis 15, 719-723. (21) Lawley, P. D., and Brookes, P. (1967) Interstrand cross-linking of DNA by difunctional alkylating agents. J. Mol. Biol. 25, 143160. (22) Carmical, J. R., Kowalczyk, A., Zou, Y., Van Houten, B., Nechev, L. V., Harris, C. M., Harris, T. M., and Lloyd, R. S. (2000) Butadiene-induced intrastrand DNA cross-links: a possible role in deletion mutagenesis. J. Biol. Chem. 275, 19482-19489. (23) Selzer, R. R., and Elfarra, A. A. (1996) Synthesis and biochemical characterization of N1-, N2-, and N7-guanosine adducts of butadiene monoxide. Chem. Res. Toxicol. 9, 126-132. (24) Selzer, R. R., and Elfarra, A. A. (1996) Characterization of N1and N6-adenosine adducts and N1-inosine adducts formed by the reaction of butadiene monoxide with adenosine: evidence for the N1-adenosine adducts as major initial products. Chem. Res. Toxicol. 9, 875-881. (25) Selzer, R. R., and Elfarra, A. A. (1997) Chemical modification of deoxycytidine at different sites yields adducts of different stabilities: characterization of N3- and O2-deoxycytidine and N3deoxyuridine adducts of butadiene monoxide. Arch. Biochem. Biophys. 343, 63-72. (26) Selzer, R. R., and Elfarra, A. A. (1997) Characterization of four N-3-thymidine adducts formed in vitro by the reaction of thymidine and butadiene monoxide. Carcinogenesis 18, 1993-1998. (27) Tretyakova, N. Y., Sangaiah, R., Yen, T.-Y., and Swenberg, J. A. (1997) Synthesis, characterization, and in vitro quantitation of N-7-guanine adducts of diepoxybutane. Chem. Res. Toxicol. 10, 779-785. (28) Brookes, P., and Lawley, P. D. (1961) The alkylation of guanosine and guanylic acid. J. Chem. Soc. 3923-3928. (29) Tretyakova, N. Y., Lin, Y.-P., Upton, P. B., Sangaiah, R., and Swenberg, J. A. (1996) Macromolecular adducts of butadiene. Toxicology 113, 70-76. (30) Oe, T., Kambouris, S. J., Walker, V. E., Meng, Q. X., Recio, L., Wherli, S., Chaudhary, A. K., and Blair, I. A. (1999) Persistence of N7-(2,3,4-trihydroxybutyl)guanine adducts in the livers of mice and rats exposed to 1,3-butadiene. Chem. Res. Toxicol. 12, 247257. (31) Koivisto, P., Kilpela¨inen, I., Rasanen, I., Adler, I.-D., Pacchierotti, F., and Peltonen, K. (1999) Butadiene diolepoxide- and diepoxybutane-derived DNA adducts at N7-guanine: a high occurrence of diolepoxide-derived adducts in mouse lung after 1,3butadiene exposure. Carcinogenesis 20, 1253-1259. (32) Aldrich Chemical Co., Inc. (1993) In The Aldrich Library of 13C and 1H FT NMR Spectra, 1st ed. (Pouchert, C. J., and Behnke, J., Eds.) Vol. 3, pp 223A, Aldrich Chemocal Co., Inc., Milwaukee, WI.

(33) Chang, C.-J., Ashworth, D. J., Chern, L.-J., DaSilva Gomes, J., Lee, C.-G., Mou, P. W., and Narayan, R. (1984) 13C NMR studies of methylnucleosides. Org. Magn. Reson. 22, 671-675. (34) Box, H. C., Lilga, K. T., Alderfer, J. L., French, J. B., and Potienko, J. (1979) 13C NMR characterization of alkyl derivatives of guanosine. J. Carbohydr. Nucleosides Nucleotides 6, 255-262. (35) Veldhuyzen, W. F., Lam, Y.-F., and Rokita, S. E. (2001) 2′Deoxyguanosine reacts with a model quinone methide at multiple sites. Chem. Res. Toxicol. 14, 1345-1351. (36) March, J. (1992) Kinetic requirements for reaction. Advanced Organic Chemistry: Reactions, Mechanisms, and Structure, 4th ed., pp 209-212, John Wiley & Sons, New York. (37) Kochetkov, N. K., Budovskii, E. I., Sverdlov, E. D., Simukova, N. A., Turchinskii, M. F., and Shibaev, V. N. (1971) Electronic structure and reactivity of the monomer components of nucleic acids. In Organic Chemistry of Nucleic Acids, Part A (Kochetkov, N. K., and Budovskii, E. I., Eds.) pp 121-129, 162, and 165172, Plenum Press, London and New York. (38) Tretyakova, N. Y., Chiang, S.-Y., Walker, V. E., and Swenberg, J. A. (1998) Quantitative analysis of 1,3-butadiene-induced DNA adducts in vivo and in vitro using liquid chromatography electrospray ionization tandem mass spectrometry. J. Mass Spectrom. 33, 363-376. (39) Pischetsrieder, M., Seidel, W., Munch, G., and Schinzel, R. (1999) N(2)-(1-Carboxyethyl)deoxyguanosine, a nonenzymatic glycation adduct of DNA, induces single-strand breaks and increases mutation frequencies. Biochem. Biophys. Res. Commun. 264, 544-549. (40) Fernandes, P. H., Wang, H., Rizzo, C. J., and Lloyd, R. S. (2003) Site-specific mutagenicity of stereochemically defined 1,N2-deoxyguanosine adducts of trans-4-hydroxynonenal in mammalian cells. Environ. Mol. Mutagen. 42, 68-74. (41) Turesky, R. J., Markovic, J., and Aeschlimann, J.-M. (1996) Formation and differential removal of C-8 and N2-guanine adducts of the food carcinogen 2-amino-3-methylimidazo[4,5-f]quinoline in the liver, kidney, and colorectum of the rat. Chem. Res. Toxicol. 9, 397-402. (42) Carmical, J. R., Zhang, M., Nechev, L., Harris, C. M., Harris, T. M., and Lloyd, R. S. (2000) Mutagenic potential of guanine N2 adducts of butadiene mono- and diolepoxide. Chem. Res. Toxicol. 13, 18-25. (43) Minko, I. G., Washington, M. T., Prakash, L., Prakash, S., and Lioyd, R. S. (2001) Translesion DNA synthesis by yeast DNA polymerase η ontemplates containing N2-guanine adducts of 1,3butadiene metabolites. J. Biol. Chem. 276, 2517-2522. (44) Mah, M. C.-M., Boldt, J., Culp, S. J., Maher, V. M., and McCormick, J. J. (1991) Replication of acetylaminofluoreneadducted plasmids in human cells: spectrum of base substitutions and evidence of excision repair. Proc. Natl. Acad. Sci. U.S.A. 88, 10193-10197. (45) Stierum, R. H., van Herwijnen, M. H. M., Maas, L. M., Hageman, G. J., and Kleinjans, J. C. S. (1994) Measurement by 32Ppostlabeling of (()anti-benzo[a]pyrene-diolepoxide-N2-deoxyguanosine adduct persistence in unstimulated human peripheral blood lymphocytes. Mutat. Res. 325, 31-37. (46) Johnson, K. A., Fink, S. P., and Marnett, L. J. (1997) Repair of propanodeoxyguanosine by nucleotide excision repair in vivo and in vitro. J. Biol. Chem. 272, 11434-11438. (47) Bodell, W. J. (2003) Repair of DNA alkylation products formed in 9L cell lines treated with 1-(2-chloroethyl)-1-nitrosourea. Mutat. Res. 522, 85-92.

TX0341355

Identification and Characterization of a Series of Nucleoside

Recommend Documents