Characterization of the Major Purine and Pyrimidine Adducts Formed

Article pubs.acs.org/crt

Characterization of the Major Purine and Pyrimidine Adducts Formed after Incubations of 1‑Chloro-3-buten-2-one with Single-/ Double-Stranded DNA and Human Cells Ling-Yan Liu,† Jin Zheng,† Cong Kong,‡ Jing An,† Ying-Xin Yu,† Xin-Yu Zhang,*,† and Adnan A. Elfarra*,§ †

Institute of Environmental Pollution and Health, School of Environmental and Chemical Engineering, Shanghai University, Shanghai 200444, China ‡ East China Sea Fisheries Research Institute, Chinese Academy of Fishery Sciences, Shanghai 200090, China § Department of Comparative Biosciences and the Molecular and Environmental Toxicology Center, University of Wisconsin-Madison, Madison, Wisconsin 53706, United States S Supporting Information *

ABSTRACT: We have previously shown that 1-chloro-3buten-2-one (CBO), a potential reactive metabolite of 1,3butadiene (BD), exhibits potent cytotoxicity and genotoxicity that have been attributed in part to its reactivity toward DNA. In an effort to identify the DNA adducts of CBO, we characterized the CBO reactions with 2′-deoxyguanosine (dG), 2′-deoxycytidine (dC), and 2′-deoxyadenosine (dA) under in vitro physiological conditions (pH 7.4, 37 °C). In the present study, we investigated the CBO reaction with 2′deoxythymidine (dT) and compared the rate constants of the reactions of CBO with dA, dC, dG, and dT at both individualand mixed-nucleosides levels. We also investigated the reactions of CBO with single- and double-stranded DNA using HPLC with UV detection after adducts were released by either acid or enzymatic hydrolysis of DNA. Consistent with the results from the nucleoside reactions and the rate constant experiments, 1,N6-(1-hydroxy-1-chloromethylpropan-1,3-diyl)adenine (A-2D) was identified as the major DNA adduct detected after acid hydrolysis, followed by N7-(4-chloro-3-oxobutyl)guanine (CG-2H) and a small amount of 1,N6-(1-hydroxy-1-hydroxymethylpropan-1,3-diyl)adenine (A-1D). After enzymatic hydrolysis, 1,N6-(1-hydroxy-1-hydroxymethylpropan-1,3-diyl)-2′-dexoyadenosine (dA-1), 3,N4-(1-hydroxy-1-hydroxymethylpropan-1,3diyl)-2′-deoxycytidine (dC-1/2), and 1,N2-(3-hydroxy-3-hydroxymethylpropan-1,3-diyl)-2′-dexoyguanosine (CG-1) were detected, with dA-1 being the major product, followed by dC-1/2. When a nontoxic concentration of CBO (1 μM) was incubated with HepG2 cells, no adducts could be detected by LC-MS. However, pretreatment of cells with L-buthionine sulfoximine to deplete GSH levels allowed A-2D to be consistently detected in cellular DNA. These results may contribute to a better understanding of the role of the DNA adducts in CBO genotoxicity and mutagenicity. It also suggests that A-2D could be developed as a biomarker of CBO formation after BD exposure in vivo.

■

INTRODUCTION Carcinogenicity of 1,3-butadiene (BD), an air pollutant, has been attributed to its epoxide metabolites formed via the cytochrome P450-mediated pathway, including 3,4-epoxy-1butene, 3,4-epoxy-1,2-butanediol, and 1,2,3,4-diepoxybutane.1,2 These epoxides are cytotoxic, genotoxic, and mutagenic and can react with DNA to generate multiple adducts.3−12 BD can also be metabolized via an alternative pathway, i.e., the myeloperoxidase (MPO)-mediated pathway, to generate 1chloro-2-hydroxy-3-butene (CHB; Scheme 1),13 which has been demonstrated to be cytotoxic, genotoxic, and mutagenic.14 CHB can be bioactivated by alcohol dehydrogenases to yield 1chloro-3-buten-2-one (CBO; Scheme 1), which can readily react with GSH to form mono- and di-GSH conjugates,15 and © XXXX American Chemical Society

Scheme 1. Formation of CHB and CBO from BD

has higher cytotoxicity and genotoxicity than CHB.14 The alternative metabolic pathway of BD could be toxicologically important because it is expected to be active primarily in the Received: August 21, 2016 Published: December 6, 2016 A

DOI: 10.1021/acs.chemrestox.6b00282 Chem. Res. Toxicol. XXXX, XXX, XXX−XXX

Article

Chemical Research in Toxicology bone marrow and neutrophils, which contain abundant MPO.16,17 Because the target organ for BD carcinogenicity in humans is the lymphohematopoietic system1,18 located in the bone marrow, the results obtained with CHB and CBO suggest a possible association between the MPO metabolic pathway and the carcinogenicity of BD in humans. CBO is a bifunctional alkylating agent that has been shown to readily undergo reactions with nucleobases. We have reported the reactions of CBO with 2′-deoxycytidine (dC),19 2′-deoxyadenosine (dA),20 and 2′-deoxyguanosine (dG)21 under in vitro physiological conditions (pH 7.4, 37 °C). The reaction of CBO with dC yields 3,N4-(1-hydroxy-1-chloromethylpropan-1,3-diyl)-2′-deoxycytidine (dC-4) as the major product, which decomposes to a pair of diastereomers of 3,N4-(1-hydroxy-1-hydroxymethylpropan-1,3-diyl)-2′-deoxycytidine (dC-1/2) under in vitro physiological conditions (Scheme 2); the decomposition occurs through replacement of the chlorine atom by a hydroxyl group with the half-life being 1.7 h.19

Scheme 3. Structures of the CBO-dA Adduct (dA-2) under in Vitro Physiological Conditions, the Decomposition Product (dA-1) of dA-2, and Their Corresponding AcidHydrolysis Products (A-2D and A-1D)

Scheme 2. Structures of the Major CBO-dC Adduct (dC-4) under in Vitro Physiological Conditions and Its Decomposition Product (dC-1/2)

product and is slowly converted to two products under in vitro physiological conditions: 1,N2-(1-hydroxy-1-hydroxymethylpropan-1,3-diyl)-2′-deoxyguanosine (CG-4D1) and N2-(4-hydroxy-3-oxobutyl)-2′-deoxyguanosine (CG-4D2) (Scheme 4).21 To investigate the reactions of CBO with nucleoside residues in DNA and to identify the adducts that could be formed at low CBO concentrations, we first investigated the reaction of CBO with 2′-deoxythymidine (dT) and compared the kinetics of the reactions with individual- and mixed-nucleosides. We then investigated the reactions of CBO with double-stranded DNA (dsDNA) and single-stranded DNA (ssDNA) under in vitro physiological conditions using HPLC with UV detection after either acid or enzymatic hydrolysis of DNA. Lastly, we used more sensitive LC-MS methods to investigate the DNA adducts formed in control HepG2 cells and in cells preincubated with Lbuthionine sulfoximine (BSO) at low CBO concentrations that were not associated with cytotoxicity.

The reaction of CBO with dA under in vitro physiological conditions produces 1,N6-(1-hydroxy-1-chloromethylpropan1,3-diyl)-2′-dexoyadenosine (dA-2), which undergoes replacement of the chlorine atom by a hydroxyl group to generate 1,N6-(1-hydroxy-1-hydroxymethylpropan-1,3-diyl)-2′-deoxyadenosine (dA-1) with a half-life of 0.9 h (Scheme 3).20 Acid hydrolysis of dA-2 and dA-1 gives the corresponding deribosylated products A-2D and A-1D, respectively (Scheme 3). Interestingly, in a preliminary experiment, A-2D was detected by LC-MS in DNA from cells incubated with CBO.20 The reaction of CBO with dG is complex, apparently because dG has three reactive sites, i.e., N7-, N1-, and N2-positions. The reaction at the N7-position leads to the formation of a product N7-(4-chloro-3-oxobutyl)-2′-deoxyguanosine (CG-2); CG-2 is so labile that pure compound cannot be obtained, but its acidhydrolysis product N7-(4-chloro-3-oxobutyl)guanine (CG-2H) is stable (Scheme 4).21 Most of CG-2 undergoes a retroMichael addition to produce dG and CBO, and a small part of CG-2 decomposes to N7,8-(3-hydroxy-3-chloromethylpropan1,3-diyl)guanine (CG-3). The reaction at the N1-position generates a pair of diastereomers with the structures of 1,N2-(3hydroxy-3-chloromethylpropan-1,3-diyl)-2′-deoxyguanosine (CG-5/6), which are formed apparently through an initial Michael reaction at the N1-position and subsequent cyclization with the exocyclic amino group. Incubation of CG-5/6 under in vitro physiological conditions yields 1,N2-(3-hydroxy-3-hydroxymethylpropan-1,3-diyl)-2′-deoxyguanosine (CG-1) with the half-life at 2.7 h.21 In comparison with those at the N7- and N1positions, the CBO-dG adduct at the N2-position, that is, N2(4-chloro-3-oxobutyl)-2′-deoxyguanosine (CG-4), is a minor

■

EXPERIMENTAL PROCEDURES

Materials. BD (99%) was purchased from Dalian Da-Te Gas Ltd. (Dalian, China). dA monohydrate was obtained from Alfa Aesar (Ward Hill, MA, U.S.). dG monohydrate, dC hydrochloride, dT, calf thymus DNA (CT DNA, Cat. No. D1501), BSO, 3-(4,5-dimethyl-2thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide (MTT), dimethyl sulfoxide-d6 (DMSO-d6), D2O, phosphodiesterase II from bovine spleen (Cat. No. P9041), micrococcal nuclease from Staphylococcus aureus (Cat. No. N3755), and alkaline phosphatase from calf intestine (Cat. No. P4978) were obtained from Sigma-Aldrich (St. Louis, MO, U.S.). Methanol and acetonitrile (ACN) (HPLC solvent grade), trifluoroacetic acid (TFA), 2-ethoxyethanol, and other reagents (analytical reagent grade) were purchased from Sinopharm Chemical Reagent Company (Shanghai, China). CBO was synthesized as reported previously and was usually used without further purification (purity ∼80%).15 All phosphate buffers (100 mM) contained 100 mM KCl. HepG2 cell line was purchased from the Center of Cell Resources of the Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences (Shanghai, China). Dulbecco’s modified Eagle’s medium (DMEM) and fetal bovine serum (FBS) were purchased from B


Article

Chemical Research in Toxicology

Scheme 4. Structures of the CBO-dG Adducts and Their Decomposition Products, and the Acid-Hydrolysis Product of CG-2 (CG-2H)

Life Technologies (Grand Island, NY, U.S.). Streptomycin and penicillin were purchased from Beijing Dingguo Changsheng Biotech Company (Beijing, China). Biomiga EZgene Tissue gDNA Miniprep Kit was obtained from Biomiga, Inc. (San Diego, CA, U.S.). Instruments and Methods. Mass spectra were obtained on a Bruker microTOF II mass spectrometer with electrospray ionization (ESI). NMR spectra were recorded on a Bruker Instruments Avance AV500 spectrometer (500 MHz). Analyses of samples and isolation of products were performed on a Waters Alliance 2695 HPLC system (Milford, MA, U.S.) equipped with a Waters 2996 Photodiode Array detector and a Waters 2475 Multi λ fluorescence detector. LC-MS analyses were performed on an Agilent 1290 Infinity HPLC system connected to an Agilent 6224 TOF mass spectrometer with ESI. In the present study, multiple HPLC columns and gradient programs were used. The columns used included an Agilent ZORBAX 80 Å StableBond SB-C18 analytical column (4.6 × 250 mm, 5 μm), an Agilent Zorbax Eclipse Plus C18 analytical column (4.6 × 250 mm, 5 μm), an Agilent Poroshell 120 PFP analytical column (3.0 × 100 mm, 2.7 μm), and a Waters SunFire C18 preparative column (10 × 150 mm, 5 μm), which were abbreviated as SB, EP, PS, and SF columns, respectively. To simplify the description and avoid confusion, the combinations of the HPLC columns used and the accompanied gradient programs were designated as “HPLC Methods”, which are detailed in Table 1. Reaction of CBO with dT under in Vitro Physiological Conditions, and at pH 10 and 37 °C. dT (1.3 mg, 5.4 μmol) was dissolved in 2 mL of pH 7.4 phosphate buffer, and CBO (3 μL, purity 83%, 26 μmol, CBO/dT = 5:1) was added. The mixture was incubated at 37 °C. Aliquots (100 μL) were withdrawn every hour and were subjected to HPLC analyses using Method 1. The reaction was also performed at the CBO/dT molar ratio of 3:1 and at pH 10. For preparation of the product, CBO (23.5 μL, 205 μmol) was added to a solution of dT (10 mg, 41 μmol) in 4 mL of pH 10 phosphate buffer. After incubation at 45 °C for 3 h and the pH being adjusted to 3−4, the solution was extracted with ethyl ether, lyophilized, and subjected to preparative HPLC separation using Method 2 to isolate the product dT-1 (the retention time was 10.2 min). Purified dT-1 was dissolved in pH 7.4 phosphate buffer to determine its half-life as described before.20 Acid Hydrolysis of CBO-dG Adducts. The CBO-dG adducts were prepared as described previously.21 For hydrolysis, the products (1−2 mg) were dissolved in 200 μL of water, and then 20 μL of 1.1 M HCl was added. The acidified solutions were incubated at 75 °C for 20 min. The solutions were neutralized to pH ∼6 and subjected to HPLC

analyses using Method 3. The retention times for these acid-hydrolysis products were as follows: CG-1H, 9.2 min; CG-2H, 16.3 min; CG-3, 18.6 min; CG-4H, 18.3 min; CG-5/6H, 18.6 min; CG-4D1H, 10.4 min; and CG-4D2H, 12.2 min. The products were isolated with preparative HPLC using Method 4. Determination of the Rate Constants of the Pseudo-First Order Reactions of CBO with Nucleosides under in Vitro Physiological Conditions. The stock solutions (3 mM) of the four nucleosides were prepared by dissolving dC (2.37 mg, 9 μmol), dA (2.42 mg, 9 μmol), dG (2.57 mg, 9 μmol), and dT (2.18 mg, 9 μmol) in 3 mL of pH 7.4 phosphate buffer, respectively. To determine the rate constants of the reactions of CBO with individual nucleoside, 100 μL of the stock solution of each nucleoside was diluted with 900 μL of pH 7.4 phosphate buffer. After adding 3.5 μL of CBO (purity 83%, 31 μmol), the solutions were immediately incubated at 37 °C. Aliquots (100 μL) were withdrawn every 20 min (the pH was adjusted to ∼6 if needed) and subjected to HPLC analyses using Method 1. The monitoring wavelength was 260 nm. To measure the rate constants of the reaction of CBO with mixed nucleosides, the stock solutions of the four nucleosides (100 μL each) were mixed and then diluted with 600 μL of pH 7.4 phosphate buffer. After CBO (3.5 μL, 31 μmol) was added, the solution was incubated at 37 °C and analyzed as described above. The standard curves of the four nucleosides were obtained through injecting different amounts (0.40, 1.2, 2.0, 4.0, 6.0, 8.0, 10.0, 12.0, and 16.0 nmol) of nucleosides onto HPLC. For each amount, three injections were made. The averages of the three peak areas obtained at 260 nm were plotted against the nucleoside amounts, and the data were subjected to linear regression using the equation y = kx. Excellent linearity was obtained for the standard curves; the slopes (k) and correlation coefficients (R2) were listed as follows: dA, 830236, 0.9986; dC, 383248, 1; dG, 674077, 0.9994; and dT, 479353, 0.9999. The retention time of dC, dA, dG, and dT was 6.3, 7.7, 7.9, and 8.8 min, respectively. The peak areas of the four nucleosides were determined and were converted to the concentrations of nucleosides by using the corresponding standard curves. The concentrations of nucleosides were plotted against the reaction time, and the data were fitted with exponential decay (y = a × e−bx) to obtain the rate constants b. Usually, the first 6 points (i.e., the data collected within 2 h) were used for fitting with the exception of dT in mixed nucleosides, in which 14 points were employed because the data exhibited severe scattering due to a very low reaction rate. The data exhibited excellent fitting (R2 > 0.98) except those for dT in mixed nucleosides, possibly C


Article

Chemical Research in Toxicology Table 1. Information of the HPLC Methods Used in the Present Study no.

columna

temperature (°C)

flow rate (mL/min)

mobile phase at pump Ab

mobile phase at pump Bb

time (min)

percentage of pump B

1

SB

30

1.0

1% MeOH

30% MeOH

2

SF

30

3.0

1% MeOH

30% MeOH

3

SB

30

1.0

water

ACN

4

SF

30

3.0

1% MeOH

30% MeOH

5

EP

30

1.0

water

MeOH

6

SB

30

1.0

water

ACN

7

EP

60

1.0

water

MeOH

8

PS

30

0.3

water

ACN

9

EP

30

1.0

water

ACN

1 7 8 19 20 0 22 1 20 29 30 0 4 5 19 20 0 15 16 1 23 24 30 31 1 23.4 24.4 1 11 12 1 12 13

0 93 100 100 0 100 100 0.1 10 15.3 0.1 50 50 90 90 50 5 28.8 5 0.1 11.2 40 40 0.1 0 35.8 0 0.1 53.2 0.1 10 90 10

a The detailed information about the columns is in the text. bThe mobile phases except those in Method 8 contained TFA to control the pH. If the mobile phase was water or contained water, the pH values were adjusted to 2.5 with TFA. For pure methanol or ACN, where pH values could not be measured, TFA (0.1% of the solvent volumes) was directly added to the solvents. For mobile phases in Method 8, formic acid at 0.1% of the solvent volumes was added without measuring the pH. The compositions of the mobile phases are indicated with percentages in volume.

because of its low reactivity toward CBO in comparison with the other nucleosides. The experiments were performed in triplicate. Reactions of CBO with CT DNA under in Vitro Physiological Conditions. CT DNA (2 mg/mL) was hydrated in 50 mM, pH 7.2 Tris buffer containing 1 mM MgCl2 overnight at 4 °C. ssDNA was prepared by heating the hydrated DNA at 100 °C for 5 min and rapidly cooling the solution on ice.4 Different volumes of CBO were added, and the solutions were immediately incubated at 37 °C for 7 h. In the experiments, to optimize the incubation time, 3 μL of CBO (purity 83%, 26 μmol) was added to 0.5 mL of DNA solutions, which were then incubated at 37 °C for 3, 7, and 12 h, respectively. After extraction with ethyl ether, the solutions were subjected to either acid or enzymatic hydrolysis. For acid hydrolysis, 1.1 M HCl at one tenth of the solution volumes was added. The acidified solutions were heated at 75 °C for 20 min. After cooling to ambient temperature and pH being adjusted to ∼4 with NaOH, the solutions were analyzed by HPLC using Method 3. However, with Method 3, CG-3 and CG-5/ 6H exhibited virtually identical retention times (both at 18.6 min) and thus were unresolvable; CG-3 and CG-4H, or CG-5/6H and CG-4H could barely be resolved due to the small difference between the retention time of CG-3 or CG-5/6H and CG-4H (which was 18.3 min). To separate the three compounds, a different HPLC method was developed, which was designated as Method 5. Using Method 5, baseline resolution was obtained for CG-4H and CG-5/6H, whose retention time was 12.1 and 12.8 min, respectively. The peak of CG-3,

which had the retention time at 13.2 min, slightly overlapped with that of CG-5/6H. For enzymatic hydrolysis, DNA was first precipitated with 2-ethoxyethanol as described before.22 The precipitated DNA was resolubilized in buffer and then hydrolyzed with micrococcal nuclease, phosphodiesterase, and alkaline phosphatase using a protocol reported in the literature,23 with modifications in the amounts of enzymes, which were 10, 1, and 1 mU/μg DNA, respectively. After enzymatic hydrolysis, the pH values of the solutions were adjusted to ∼6, and four volumes of ice-cold 75% ethanol were added. The mixtures were left at −20 °C for 1 h and then centrifuged (22,000g, 4 °C, 10 min) to remove enzymes. The supernatants were lyophilized to dryness, and the resulting solid was dissolved in water and subjected to HPLC analyses using Method 6. The retention time of the adducts was as follows (in minutes): dA-1, 9.9; dC-1, 10.8; dC-2, 11.7; CG-1, 14.4; CG-3, 18.6; CG-4, 25.0; CG-4D1, 15.8; CG-4D2, 17.0; and CG-1H, 9.2. However, using Method 6, we were unable to resolve CG-1 from dT. Therefore, a different HPLC method, which was designated as Method 7, was developed. Using Method 7, CG-1 and dT were successfully separated with the retention time being 8.7 and 9.0 min, respectively. Identification of Adducts in CBO-Treated Cells. HepG2 cells at ∼80% confluence were incubated with different concentrations of CBO in D-Hank’s balanced salt solutions at 37 °C for 1 h. Alternatively, before incubation with CBO, cells at ∼60% confluence were preincubated with 100 μM BSO at 37 °C for 24 h. Cellular DNA D


Article

Chemical Research in Toxicology was extracted with Biomiga EZgene Tissue gDNA Miniprep Kit. The extracted DNA was either acid-hydrolyzed or enzymatically hydrolyzed as described above and then subjected to LC-MS analyses using Method 8. The use of LC-MS was necessary to increase the sensitivity of adduct detection so that low CBO concentrations not associated with cytotoxicity could be used. The retention time of the adducts was as follows (in minutes): dA-1, 1.9; dA-2, 3.4; dC-1, 1.9; dC-2, 2.0; dC3, 4.5; dC-4, 4.1; CG-1, 5.1; CG-2, 4.1; CG-3, 5.4; CG-4, 6.4; CG4D1, 5.1; CG-4D2, 5.3; CG-5, 6.3; CG-6, 6.4; dT-1, 7.7; A-1D, 1.5; A2D, 2.2; CG-1H, 2.8; CG-2H, 4.9; CG-4H, 6.7; CG-4D1H, 3.7; CG4D2H, 4.2; and CG-5/6H, 5.2. Examination of CBO Cytotoxicity and Genotoxicity in HepG2 Cells. Cytotoxicity and genotoxicity of CBO in HepG2 cells were examined with the MTT assay and comet assay as described before, respectively.11,14 The cells used were either control cells or cells preincubated with 100 μM BSO as described above. Determination of Intracellular GSH Concentrations. The intracellular GSH concentrations were measured with a protocol reported in the literature.24 The samples were analyzed with Method 9, and the retention time of the derivatized product of GSH was 5.2 min. The protein concentrations were measured with the Bradford assay. Statistical Analysis. The difference between two values was examined with Student’s t-test. The difference was considered statistically significant when the p value was smaller than 0.05.

Scheme 5. Structures of dT-1, CG-4H, CG-4D1H, and CG4D2H

RESULTS Reaction of CBO with dT under in Vitro Physiological Conditions and Structural Characterization of the Product. The reaction of CBO with dT under in vitro physiological conditions yielded only a single product with the retention time at 16.0 min and the absorption maximum at 269 nm, which was designated as dT-1. The formation rate of dT-1 was dependent on the CBO/dT molar ratio and the pH of the incubation (Figure 1). At the CBO/dT molar ratio of 5:1 and

The 1H NMR spectrum of dT-1 in D2O showed 17 protons (Figure S1), consistent with the expected molecular formula. The signals of the protons at the pyrimidine and deoxyribose rings (Table 2) were easily assigned on the basis of the correlation spectroscopy (COSY) and heteronuclear multiplequantum correlation (HMQC) spectra (Figures S2 and S3). The signal at 4.48 ppm was assigned as that of protons at C1″ (Scheme 5) because it was a single peak (Figure S1). The peaks at 2.91 and 4.20 ppm were determined to be the signals of protons at C3″ and C4″, respectively, based on the COSY and heteronuclear multiple-bond correlation (HMBC) spectra (Figure S4) because the protons at 4.20 ppm exhibited coupling with two carbons at 151.4 (C2) and 165.2 (C4) but those at 2.91 ppm did not. Therefore, the structure of dT-1 was characterized as N3-(4-chloro-3-oxobutyl)thymidine (Scheme 5). dT-1 was relatively stable under in vitro physiological conditions with the half-life being 62 ± 2 h. Characterization of the Acid-Hydrolysis Products of CBO-dG Adducts. The structures of CBO-dG adducts have been determined before; however, the acid-hydrolysis products of these adducts have not been characterized previously except for CG-2H.21 Thus, in the present study, these adducts were subjected to acid-hydrolysis (CG-3 was stable under the acidhydrolysis conditions), and the products were characterized to provide the standards for investigation of the reactions between CBO and DNA, as DNA samples are usually acid-hydrolyzed (0.1 M HCl, 75 °C, 20 min) to release the purine adducts.4,25 CG-5 and CG-6, the predominant CBO-dG adducts, underwent acid-hydrolysis to yield a major product with the retention time at 18.6 min, which was designated as CG-5/6H (Scheme 6). It should be noted that CG-5 and CG-6 are a pair of diastereomers, which have slightly different 1H NMR spectra and can be isolated individually.21 Thus, their corresponding acid-hydrolysis products were designated as CG-5H and CG6H, respectively. However, experimentally acid hydrolysis of either CG-5 or CG-6 did not lead to the formation of products with different retention times because CG-5H and CG-6H are a pair of enantiomers; loss of the deoxyribose moieties, which are chiral, results in the conversion from diastereomers to enantiomers. As a pair of enantiomers, CG-5H and CG-6H

■

Figure 1. Time-dependent formation of dT-1 at different CBO:dT molar ratios and different incubation pH values (the points were connected to clearly show the trends of the formation of dT-1).

pH 7.4, the yields of dT-1 were 3-fold higher than those at 3:1; when the pH was increased from 7.4 to 10, not only did the yields of dT-1 increase significantly but also the reaction became faster (Figure 1). At pH 10 and the CBO/dT molar ratio of 5:1, approximately 30% of dT was consumed when the yield of dT-1 reached the maximum. The ESI-mass spectrum of dT-1 showed that the protonated molecular ion (MH+) of dT-1 appeared at m/z 347.1028 with an [MH+2]+ isotope peak (∼1/3 intensity) (data not shown), consistent with the expected structure of one molecule of CBO being added to dT (the calculated value was 347.1010; Scheme 5). E


Article

Chemical Research in Toxicology Table 2. NMR Data of the Products Characterized in the Present Studya dT-1 positionb 2 4 5 6 8 1′ 2′ 3′ 4′ 5′

H

7.64 (s, 1H) (t, 1H, J = 6.5 Hz) (m, 2H) (m, 1H) (dd, 1H, J = 4, 8.5 Hz) (dd, 1H, J = 5, 12.5 Hz) (dd, 1H, J = 3.5, 12.5 Hz) (s, 2H)

CG-5/6H (open chain) 13

1

C

H

151.4 165.2 110.6 135.5

CG-5/6H (fused-ring) 13

C

9.02 (s, 1H)c

174.2 149.2c 108.0c 174.2c 137.4c

48.8

2.09 (s, 2H)

30.0

2.65 (t, 2H, J = 6.5 Hz) 2.37 (t, 2H, J = 6.5 Hz)

207.4 38.0 28.0

1″

6.27 2.36 4.44 4.01 3.75 3.82 4.48

2″ 3″ 4″

2.91 (t, 2H, J = 7 Hz) 4.20 (m, 2H)

204.7 37.1 36.8

1.89 (s, 3H)

12.1

CH3 NH a

1

1

H

8.78 (s, 1H)c

13

C

151.9 149.5c 111.2c 152.5 138.1c

85.9 38.5 70.2 86.4 60.9

d

3.77 (d, 1H, J = 11 Hz) 3.84 (d, 1H, J = 11 Hz) 2.03 (m, 2H) 3.58 (m, 1H) 4.49 (m, 1H)

49.5 79.3 28.6 35.8

8.93 (s, 1H)

13

The C chemical shifts of the carbons carrying hydrogen were obtained from HMQC spectra, and those carrying no hydrogen were determined from HMBC spectra. bAccording to the rule, the carbons at the side chain of CG-5/6H should be numbered 1′, 2′, 3′, and 4′ because the deoxyribose moiety has been lost. However, in the present study the carbons were still numbered 1″, 2″, 3″, and 4″ as CG-5/6 to facilitate the comparison of data among these adducts. The numbering of the atoms is shown in Schemes 5 and 6. cThese assignments are tentative. dThe signal was not observed.

corresponding 13C chemical shifts at 138.1 and 137.4 ppm, respectively (Table 2). These protons could only be assigned as those at C8 (Scheme 6) on the basis of their high chemical shifts (and also the high chemical shifts of the corresponding carbons) and a comparison with the corresponding data of CG5 and CG-6,21 the precursors of CG-5/6H. Because the guanine ring has only one proton on carbon (i.e., the proton on C8), the fact that CG-5/6H exhibited such two protons indicated that it was a mixture of two isomers. The COSY spectrum (Figure S7) showed that the proton at 8.93 ppm (a proton at nitrogen) coupled to the two protons at 2.03 ppm, which, in turn, coupled to two protons at 3.58 and 4.49 ppm. The two protons at 3.58 and 4.49 ppm were on the same carbon with the chemical shift at 35.8 ppm as indicated by the HMQC spectrum (Figure S6). This indicated a −CH2CH2− structure. The HMBC spectrum (Figure S8) showed that the proton at 4.49 ppm coupled to a carbon at 152.5 ppm (however, the coupling signal was weak and thus was invisible in Figure S8 due to the threshold), indicating that the carbon at 35.8 ppm was connected to the purine ring. Thus, the carbon was C4″ (Scheme 6). In addition, the protons at 2.03 ppm (i.e., H-3″, see Scheme 6) exhibited very weak coupling with a carbon at 79.3 ppm based on the HMBC spectrum, and two protons at 3.77 and 3.84 ppm showed strong coupling to this carbon (Figure S8), clearly indicating that C2″ carries a hydroxyl group. Thus, like their precursors CG-5 and CG-6, one of the two CG-5/6H isomers had a fusedring structure. In fact, the NMR data of this CG-5/6H isomer were very similar to those of CG-5/6. Therefore, one isomer of CG-5/6H was characterized as 1,N2-(3-hydroxy-3-chloromethylpropan-1,3-diyl)guanine (Scheme 6). The structure of the other CG-5/6H isomer was determined starting from a singlet peak representing two protons at 2.09 ppm, which did not couple to other protons based on the

Scheme 6. Structures of CG-5/6H and CG-1H

have exactly the same NMR spectra and are considered to be the same compound. Thus, they were designated as CG-5/6H in the present study. The ESI-mass spectrum of CG-5/6H showed that the protonated molecular ion appeared at m/z 256.0606 with an isotope peak having approximately 1/3 intensity at m/z 258.0519 (data not shown), consistent with the expected formula C9H10N5O2Cl (the calculated mass for MH+ was 256.0601). Apparently, CG-5/6H was the product of CG-5/6 after the deoxyribose moiety was lost. The 1H NMR spectrum of CG-5/6H in DMSO-d6 exhibited more protons than expected probably due to impurities (Figure S5). An unusual feature of the spectrum was that four single peaks appeared at low fields (8.78, 8.93, 9.02, and 9.85 ppm), each of which represented one proton (Figure S5). The HMQC spectrum (Figure S6) indicated that only the two protons at 8.78 and 9.02 ppm were on carbons with the F


Article

Chemical Research in Toxicology COSY spectrum (Figure S7). The corresponding 13C chemical shift was 30.0 ppm, indicating an X-CH2−Y structure. The HMBC spectrum (Figure S8) showed that the two protons coupled to two carbons at 38.0 and 207.4 ppm. Clearly, the methylene group was adjacent to a carbonyl moiety. Therefore, an open side chain was present, i.e., unlike its precursor CG-5 and CG-6, the side chain did not form a fused ring with the purine ring. Consequently, the other isomer of CG-5/6H was N1-(4-chloro-3-oxobutyl)guanine (Scheme 6). Therefore, based on the NMR data, CG-5/6H was a 1:1 mixture of one isomer with an open side chain and another isomer with a fused-ring (Scheme 6). Under in vitro physiological conditions, CG-5/6H was labile with a half-life at 4.5 ± 0.1 h and was converted to another compound (i.e., CG-1H, see below). Acid hydrolysis of CG-1 in 0.1 M HCl at 75 °C for 20 min generated only a product with the retention time at 9.2 min and the absorption maximum at 254 nm, which was designated as CG-1H. The product exhibited the protonated molecular ion at m/z 238.0943 (data not shown), which was consistent with a structure with the deoxyribose moiety being lost. Incubation of CG-5/6H under in vitro physiological conditions produced a compound that was verified to be CG-1H through the coelution experiment. Because CG-1 was the incubation product of CG-5/6 with the chloromethyl group being converted to a hydroxymethyl moiety, obviously, CG-1H was the product of CG-5/6H with the same conversion. Because CG-5/6H was a mixture of the two isomers with ring-closed and ring-opened structures, respectively, CG-1H seemed also to have the two possible structures, i.e., CG-1H was characterized as 1,N2-(3-hydroxy-3-hydroxymethylpropan-1,3-diyl)guanine or N1-(4-hydroxy-3-oxobutyl)guanine, or both (Scheme 6). Acid hydrolysis of CG-4 yielded only a compound eluted at 18.3 min with the absorption maximum at 251 nm (there was a shoulder peak at 281 nm), which was designated as CG-4H. In its ESI-mass spectrum, the protonated molecular ion appeared at m/z 256.0600 with an isotope peak with ∼1/3 intensity at m/z 258.0575 (data not shown), suggesting that CG-4H was the product of CG-4 with the deoxyribose moiety being lost. Similarly, the two incubation products of CG-4, i.e., CG-4D1 and CG-4D2, were also hydrolyzed to corresponding products with the retention time at 10.4 and 12.2 min, which were designated as CG-4D1H and CG-4D2H, respectively. Both compounds exhibited the UV absorption spectra very similar to those of their corresponding precursors, although the absorption maxima showed slightly hypsochromic shifts (261 nm for CG-4D1H vs 264 nm for CG-4D1 and 251 nm for CG4D2H vs 257 nm for CG-4D2; a similar hypsochromic shift was also observed for CG-4H because the absorption maximum for CG-4 was 257 nm). The two compounds exhibited protonated molecular ions with the same masses (m/z 238.0935 and 238.0932, respectively). The evidence showed that CG-4H, CG-4D1H, and CG-4D2H were the corresponding guanine products of CG-4, CG-4D1, and CG-4D2, respectively, after their deoxyribose moieties were simply removed. Thus, their structures were N2-(4-chloro-3-oxobutyl)guanine, 1,N2-(1hydroxy-1-hydroxymethylpropan-1,3-diyl)guanine, and N2-(4hydroxy-3-oxobutyl)guanine, respectively (Scheme 5). Kinetics of the Reactions of CBO with the Four Nucleosides under in Vitro Physiological Conditions. In previous studies, the reactions of CBO with dA, dG, and dC were investigated individually. Although each nucleoside could readily react with CBO, it was unclear how their reactivity compared to each other. Thus, we investigated the kinetics of

the reactions of CBO with the four nucleosides either individually or mixed together. In these experiments, the reactions were performed at the CBO/dN molar ratios of 100:1 to mimic the conditions for the pseudo-first order reaction. Although pseudo-first order reaction kinetics generally requires a greater molar ratio (1000:1), we were restricted to lower molar ratios due to the limited solubility of CBO. Specifically, the concentration of individual nucleoside was 0.3 mM and that of CBO was 31 mM. The rate constants of the reactions measured are listed in Table 3. The results showed that in the reactions of CBO with Table 3. Rate Constants of the Pseudo-First Order Reactions of CBO with Single Nucleoside or a Mixture of Equal Molars of the Four Nucleosides under in Vitro Physiological Conditions (×10−5 s−1)a single nucleosides mixed nucleosides

dA

dC

dG

dT

19 ± 2 23 ± 2

12 ± 1 17 ± 3

4.1 ± 0.4 4.9 ± 0.3

0.9 ± 0.1 0.08 ± 0.01

Values are presented as the means ± SD from three independent experiments.

a

either single or mixed nucleosides, the rates of dA were the highest, followed by dC, and that of dT was the lowest (Figure S9). That is, the rates of the reactions of CBO with the four nucleosides exhibited a descending order of dA > dC > dG > dT. The rates of dA were moderately higher than those of dC, whereas the rates of dA or dC were much greater than those of dG, which, in turn, were considerably higher than those of dT. In the reaction of CBO with mixed nucleosides, the rates of dA, dC, and dG showed slight increases in comparison with the corresponding values for the single nucleosides; however, the increases were not statistically significant (p > 0.05). In contrast, the rate of dT in the mixed nucleosides decreased by 12-fold, and the difference was statistically significant (Table 3). Reactions of CBO with DNA under in Vitro Physiological Conditions: Acid Hydrolysis. The reaction of CBO with DNA was investigated starting from a CBO concentration of 52 mM and at three different reaction time points (3, 7, and 12 h; Figure 2). With acid hydrolysis (0.1 M HCl, 75 °C, 20 min), the reaction of CBO with DNA was found to produce A1D, A-2D, and CG-2H, with A-2D and CG-2H being the major products, and A-1D being a minor product (Figure 2). For the reaction mixtures incubated for 7 and 12 h, a very small peak around 18.1 min was observed (which was indicated with an arrow; see Figures 2B and 2C), which exhibited the UV absorption spectra of both CG-3 (λmax = 216 and 253 nm) and CG-4H (λmax = 251 and 281 nm). However, because the retention time of CG-5/6H was actually the same as that of CG-3 and the UV absorption maximum of CG-5/6H was 254 nm, the small peak could also contain that of CG-5/6H. Using a different HPLC method (Method 5; see the Experimental Procedures section for details), the presence of CG-3, CG-4H, and CG-5/6H in the reaction mixtures was verified. Among the three compounds, CG-5/6H and CG-4H were the products with the largest and smallest amounts, respectively. Therefore, in total six purine adducts could be identified in the reaction mixtures of CBO with DNA; their amounts showed a descending order of A-2D > CG-2H > A-1D ≫ CG-5/6H > CG-3 > CG-4H. G


Article


the CBO concentrations, 7 h was selected as the reaction time for the next experiments. The reaction of CBO with DNA was then investigated with the CBO concentrations varying from 0.087 to 120 mM. This CBO concentration range was used because of the limited sensitivity of the HPLC detection method. The chromatograms were qualitatively similar to those shown in Figure 2. The results showed that the amounts of all adducts increased with the increase in the CBO concentrations but reached a plateau around 70 mM of CBO except for A-1D in dsDNA (Figure 3). Clearly, A-2D was the predominant product at all CBO concentrations in both dsDNA and ssDNA, especially at high CBO concentrations. CG-2H was also a major product with the second largest peak areas in most cases, although at low CBO concentrations (≤9 mM for dsDNA and ≤35 mM for ssDNA) its peak areas were smaller than those of A-1D. CG-3, CG-4H, and CG-5/6H were all trace products; they could be detected only at high CBO concentrations (≥17 mM for both dsDNA and ssDNA). By contrast, A-2D could be detected at the CBO concentration as low as 1.7 mM for dsDNA and 0.87 mM for ssDNA. In addition, the reactivity of ssDNA was significantly higher than that of dsDNA (Figure 3C and D), which was consistent with the observations reported previously.4 Reaction of CBO with DNA under in Vitro Physiological Conditions: Enzymatic Hydrolysis. In the above experiments, acid hydrolysis of DNA yielded only purines. To liberate pyrimidines, concentrated acid and relatively high temperature are required, which can destroy purines.26 Alternatively, DNA can be enzymatically hydrolyzed to free all four nucleosides. Thus, DNA samples were subjected to enzymatic hydrolysis to investigate the formation of pyrimidine adducts. In the experiments, dsDNA and ssDNA were incubated with CBO at 17, 70, and 120 mM at pH 7.4 and 37 °C for 7 h. The results showed that the reactions yielded dA-1, dC-1/2, CG-1,

Figure 2. HPLC chromatograms of the reaction mixtures of CBO (52 mM) with dsDNA (2 mg/mL) at pH 7.4 and 37 °C for 3, 7, and 12 h. The monitoring wavelength was 260 nm.

The results also showed that CBO had high reactivity toward DNA; the amounts of unreacted guanine and adenine decreased significantly with the increase in the reaction time. When the reaction mixture was incubated for 12 h, 80% of guanine and 99% of adenine were consumed (Figure 2C). Thus, to investigate the dependence of the product amounts on

Figure 3. Dependence of the amounts of the major CBO−DNA adducts in dsDNA and ssDNA on the CBO concentrations. DNA was hydrolyzed with HCl, and adducts were resolved by HPLC with UV detection as described in the Experimental Procedures section. The reaction time was 7 h. The amounts of CG-3, CG-4H, and CG-5/6H were extremely small and thus were not shown. H


Article


CG-4D1 was also a trace product; however, its peak areas could not be determined due to severe overlapping with other peaks (Figure 4). Similarly to acid hydrolysis, the amounts of all products with ssDNA were remarkably greater than those with dsDNA. Detection of the Adducts in CBO-Incubated Cells. The above experiments demonstrated that dA, dG, and dC residues in DNA could react with CBO to generate multiple CBO adducts. Because it was of interest to determine which products could be formed in cellular DNA in living cells, HepG2 cells were incubated with different concentrations of CBO as described in the Experimental Procedures section and the extracted DNA was subjected to either acid or enzymatic hydrolysis, followed by LC-MS analyses to increase the sensitivity for adduct detection. At three nontoxic CBO concentrations (1, 0.1, and 0.01 μM), which were selected on the basis of the cytotoxicity and genotoxicity results shown below, no adducts were detected in control cells, whereas cells pretreated with 100 μM BSO consistently exhibited the presence of A-2D in acid-hydrolyzed DNA after treatment with 1 μM CBO, consistent with our preliminary result;20 no adducts were detected in enzymatically hydrolyzed DNA samples, and A-2D was not consistently detected at the lower CBO concentration (0.1 μM). GSH Depletion by BSO, and Cytotoxicity and Genotoxicity of CBO in Control and GSH-Depleted Cells. To help provide a basis for the above data, the intracellular GSH concentrations were measured in control and BSO-treated cells. The GSH concentrations were determined to be 4.0 ± 0.2 and 54 ± 3 nmol mg protein−1 (means ± SD) for cells incubated with 100 μM BSO for 24 h and the control cells (Figure S10), respectively. Thus, GSH was decreased by about 90% after incubation with BSO. Cytotoxicity of CBO in HepG2 cells with the MTT assay was examined at 1 h after the addition of CBO. The results showed that the BSO pretreatment potentiated the cytotoxicity of CBO. CBO at 5 μM exhibited cytotoxicity on cells preincubated with BSO, whereas higher CBO concentrations (≥10 μM) were needed to cause cytotoxicity in control cells (Figure 6A). To determine the CBO concentrations that caused DNA damage in HepG2 cells, the genotoxicity of CBO was examined with the comet assay. Similar to the above-described data obtained with the MTT assay, it was observed that BSO pretreatment significantly enhanced the genotoxicity of CBO. CBO induced statistically significant DNA damage only at 2 μM or higher concentrations as measured by the comet assay (Figure 6B). Whereas at 2 μM CBO caused similar levels of DNA damage in control and BSO-preincubated cells, higher CBO concentrations (5 and 10 μM) induced considerably more severe DNA damage in BSO-preincubated cells in comparison with the control cells (Figure 6B).

and CG-4D1. (Note: CG-1 was masked by the large peak of dT and thus was invisible; as a result, a different HPLC method was developed to separate CG-1 from dT. Therefore, CG-1 was examined with this method; see Experimental Procedures section for details) (Figure 4).) Among them, dA-1 and dC-1/2 were the major products, and CG-1 and CG-4D1 were trace products.

Figure 4. Typical HPLC chromatogram of the reaction mixture of CBO with DNA after enzymatic hydrolysis. CBO (70 mM) was reacted with dsDNA (2 mg/mL) at pH 7.4 and 37 °C for 7 h. The monitoring wavelength was 260 nm.

Among all products, dA-1 was clearly the predominant adduct with either dsDNA or ssDNA, which was detected at all CBO concentrations tested (Figure 5). dC-1 and dC-2 were detectable with moderate peak areas at almost all CBO concentrations except at 17 mM with dsDNA. The amounts of CG-1 were very small, and it was detectable only at the highest CBO concentration (120 mM) with dsDNA, although it could be detected at all three concentrations with ssDNA (Figure 5).

■

DISCUSSION CBO can readily react with free nucleosides to generate multiple adducts (Schemes 2, 3, and 4).19−21 However, these adducts are not necessarily formed in the reaction of CBO with DNA because the nucleophilic ability of the various positions in free nucleosides may greatly change due to Watson−Crick hydrogen bonding in duplex DNA, and the accessibility of these positions may also be limited due to nucleobase stacking in DNA. Thus, it was of interest to investigate which adducts

Figure 5. Dependence of dsDNA and ssDNA adduct detection by HPLC after enzymatic hydrolysis of DNA on the CBO concentrations. The reaction time was 7 h. I


Article


at 0.9 h,20 and the DNA enzymatic hydrolysis process needs ∼24 h to be completed.23) On the other hand, the dC and dG residues reacted with CBO only moderately. The selectivity in the nucleophilic positions was also obvious. The N3-position in the dC residue in DNA was the predominant reaction site, similar to free dC.19 However, the situation was different with dG; for free dG, the N1-position was the predominant reaction site because the corresponding adducts (CG-5/6 and CG-1) had the largest amounts (Scheme 4).21 In sharp contrast, the corresponding adducts obtained in DNA (CG-5/6H or CG-1) were trace products, suggesting that this position was only a minor reaction site in DNA. Actually, among the nucleophilic sites of dG (N1-, N2-, and N7positions), the N7-position was the major reaction site in the reaction of CBO with DNA. That the N1-position of the dG residue in duplex DNA was less accessible in comparison with the N7-position may partly provide an explanation (see below for more discussion of this point). Surprisingly, it was the N1-position of the dA residue, rather than the N7-position of the dG residue, that was the overwhelmingly predominant reaction site in DNA for CBO. The result was consistent with the observation that dA exhibited the largest rate constants in the reactions of CBO with free nucleosides, which were nearly 5-fold greater than those of dG (Table 3). However, the accessibility of the N1position of the dA residue in dsDNA was expected to be much lower than that of the N7-position of the dG residue as the dG N7-position does not participate in Watson−Crick hydrogen bonding and is located in the major groove. However, factors such as reversibility of the reaction by a retro-Michael reaction, the stability of the transition state, and the internal cyclization potential of the N-1 position of dA may enhance the reactivity at this site. It was interesting that GSH depletion potentiated the cytotoxicity of CBO and also enhanced the detectability of A-2D at CBO concentrations that were not associated with cytotoxicity. These results could be explained by scavenging of CBO by GSH. Indeed, we have demonstrated that CBO rapidly reacted with GSH to form mono- and di-GSH conjugates.15 Thus, in GSH-depleted cells more CBO reacted with DNA, and this increased the detectability of A-2D. Interestingly, the reactivity feature of CBO toward nucleobases in DNA was different from structurally related α,β-unsaturated aldehydes, such as acrolein, 4-hydroxynonenal, and crotonaldehyde, which predominantly react with the exocyclic amino group (i.e., the N2-position) of the dG residue in DNA.27,28 Hydroxymethylvinyl ketone and methyl vinyl ketone, two structural analogues of CBO, predominantly react with the dG residue in DNA at the N1-position.29,30 Thus, the selectivity of the reaction of CBO toward nucleobases in DNA appeared distinct from that of structurally related Michael acceptors. Because BSO pretreatment also enhanced CBO genotoxicity as measured by the comet assay, the results suggested that DNA adduct formation may play a role in genotoxicity. The finding that CBO-dA adducts are the predominant products of the reactions of CBO with DNA may be toxicologically significant. The formation of CBO-dA adducts will disrupt Watson−Crick base paring, thus being likely mutagenic. Indeed, multiple studies have showed that the N6dA adducts in DNA can disrupt Watson−Crick base pairing,31−33 potentially inducing mutations.34 Recently, it has been demonstrated that N6-singly substituted dA lesions are not miscoding but that exocyclic N6,N6-dA adducts are strongly mispairing, probably due to their inability to form stable

Figure 6. Cytotoxicity (A) and genotoxicity (B) of different concentrations of CBO in control and GSH-depleted HepG2 cells as measured by the MTT assay and the comet assay, respectively. For GSH-depletion, cells were incubated with 100 μM BSO for 24 h (*p < 0.05, **p < 0.01, and ***p < 0.001). The data are presented as the means ± SD.

could be formed in DNA and to compare the results with those obtained with free nucleosides. In this regard, the reaction of CBO with DNA exhibited selectivity in nucleobases and also in the nucleophilic positions. Although adducts from the dA, dC, and dG residues were detected, their relative amounts, as estimated by the peak areas, varied greatly. Among the four nucleobases in DNA, it was clear that adenine was the overwhelmingly predominant reaction site, as supported by the findings that A-2D and dA-1 were the predominant adducts with acid and enzymatic DNA hydrolysis, respectively (Scheme 7). (It should be noted that dA-2 was not expected to be detected with enzymatic hydrolysis; it would be likely converted to dA-1 because dA-2 exhibited a short half-life Scheme 7. Major CBO−DNA Adducts Detected after Acidor Enzymatic Hydrolysis of DNA

J


Article

Chemical Research in Toxicology Watson−Crick pairs with dT.35 In another study, Kotapati et al. found that 1,N6-(2-hydroxy-3-hydroxymethylpropan-1,3-diyl)2′-deoxyadenosine, an adduct structurally related to dA-1, was potentially mutagenic.36 In addition, the dA residue in DNA was the most reactive site, and as a result, its acid-hydrolysis product A-2D was the predominant DNA adduct. Cellular experiments showed that A-2D was a promising biomarker for CBO. Further experiments are needed to examine the utility of A-2D as a biomarker of CBO formation after BD exposure in vivo.

■

calf thymus DNA; dA, 2′-deoxyadenosine; dC, 2′-deoxycytidine; dG, 2′-deoxyguanosine; DMEM, Dulbecco’s modified Eagle’s medium; DMSO, dimethyl sulfoxide; dsDNA, doublestranded DNA; dT, 2′-deoxythymidine; ESI, electrospray ionization; FBS, fetal bovine serum; HMBC, heteronuclear multiple-bond correlation; HMQC, heteronuclear multiplequantum correlation; MPO, myeloperoxidase; MTT, 3-(4,5dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide; ssDNA, single-stranded DNA; TFA, trifluoroacetic acid

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.chemrestox.6b00282. NMR spectra of dT-1 and CG-5/6H, time-dependent decreases in nucleoside concentrations in the reaction of CBO with mixed nucleosides with the molar ratio of CBO/individual nucleoside at 100:1 under in vitro physiological conditions, and the GSH concentrations in control and BSO-incubated cells (PDF)

■

REFERENCES

(1) Koppikar, A. M., Wu, C., Leavens, T., Valcovic, L., Kimmel, C., Faust, R., and Jinot, J. (2002) Health Assessment of 1,3-Butadiene. U.S. Environmental Protection Agency, Office of Research and Development, National Center for Environmental Assessment, Washington Office, Washington, DC, EPA/600/P-98/001F. http://cfpub.epa, gov/ ncea/cfm/recordisplay.cfm?deid=54499 (accessed Aug 21, 2016). (2) Kirman, C. R., Albertini, R. J., Sweeney, L. M., and Gargas, M. L. (2010) 1,3-Butadiene: I. Review of metabolism and the implications to human health risk assessment. Crit. Rev. Toxicol. 40 (Suppl. 1), 1−11. (3) Albertini, R. J., Carson, M. L., Kirman, C. R., and Gargas, M. L. (2010) 1,3-Butadiene: II. Genotoxicity profile. Crit. Rev. Toxicol. 40 (Suppl. 1), 12−73. (4) Selzer, R. R., and Elfarra, A. A. (1999) In vitro reactions of butadiene monoxide with single- and double-stranded DNA: characterization and quantitation of several purine and pyrimidine adducts. Carcinogenesis 20, 285−292. (5) Zhang, X.-Y., and Elfarra, A. A. (2003) 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. Chem. Res. Toxicol. 16, 1606−1615. (6) Zhang, X.-Y., and Elfarra, A. A. (2005) Reaction of 1,2,3,4diepoxybutane with 2′-deoxyguanosine: initial products and their stabilities and decomposition patterns under physiological conditions. Chem. Res. Toxicol. 18, 1316−1323. (7) 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 N3-deoxyuridine adducts of butadiene monoxide. Arch. Biochem. Biophys. 343, 63−72. (8) Tretyakova, N., Lin, Y.-P., Sangaiah, R., Upton, P. B., and Swenberg, J. A. (1997) Identification and quantitation of DNA adducts from calf thymus DNA exposed to 3,4-epoxy-1-butene. Carcinogenesis 18, 137−147. (9) Tretyakova, N. Y., Sangaiah, R., Yen, T.-Y., and Swenberg, J. A. (1997) Synthesis, characterization, and in vitro quantitation of N-7guanine adducts of diepoxybutane. Chem. Res. Toxicol. 10, 779−785. (10) Tretyakova, N., Sangaiah, R., Yen, T.-Y., Gold, A., and Swenberg, J. A. (1997) Adenine adducts with diepoxybutane: isolation and analysis in exposed calf thymus DNA. Chem. Res. Toxicol. 10, 1171−1179. (11) Wen, Y., Zhang, P.-P., An, J., Yu, Y.-X., Wu, M.-H., Sheng, G.-Y., Fu, J.-M., and Zhang, X.-Y. (2011) Diepoxybutane induces the formation of DNA-DNA rather than DNA-protein cross-links, and single-strand breaks and alkali-labile sites in human hepatocyte L02 cells. Mutat. Res., Fundam. Mol. Mech. Mutagen. 716, 84−91. (12) Zhang, P.-P., Wen, Y., An, J., Yu, Y.-X., Wu, M.-H., and Zhang, X.-Y. (2012) DNA damage induced by three major metabolites of 1,3butadiene in human hepatocyte L02 cells. Mutat. Res., Genet. Toxicol. Environ. Mutagen. 747, 240−245. (13) Duescher, R. J., and Elfarra, A. A. (1992) 1,3-Butadiene oxidation by human myeloperoxidase. J. Biol. Chem. 267, 19859− 19865. (14) Liu, X.-J., Zeng, F.-M., An, J., Yu, Y.-X., Zhang, X.-Y., and Elfarra, A. A. (2013) Cytotoxicity, genotoxicity, and mutagenicity of 1-chloro2-hydroxy-3-butene and 1-chloro-3-buten-2-one, two alternative metabolites of 1,3-butadiene. Toxicol. Appl. Pharmacol. 271, 13−19. (15) Elfarra, A. A., and Zhang, X.-Y. (2012) Alcohol dehydrogenaseand rat liver cytosol-dependent bioactivation of 1-chloro-2-hydroxy-3-

AUTHOR INFORMATION

Corresponding Authors

*(X.-Y.Z.) Institute of Environmental Pollution and Health, School of Environmental and Chemical Engineering, Shanghai University, 333 Nanchen Road, Shanghai, 200444, People’s Republic of China. Tel: + 86-21-6613-7736. Fax: + 86-21-66136928. E-mail: [email protected]. *(A.A.E.) University of Wisconsin-Madison, Department of Comparative Biosciences, 2015 Linden Dr., Madison, WI 53706, United States. Tel: + 1-608-262-6518. Fax: + 1-608-2623926. E-mail: [email protected]. ORCID

Xin-Yu Zhang: 0000-0003-3472-8393 Funding

This work was supported by the National Nature Science Foundation of China (Nos. 21077070 and 21377080), the Shanghai Municipal Education Commission (No. 11ZZ90), the Shanghai Leading Academic Discipline Project (No. S30109), and the National Institutes of Health of the U.S.A. (R01 DK044295 and R01 ES06841). Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The NMR spectra of all samples were collected on a Bruker Avance AV500 spectrometer in the Instrumental Analysis & Research Center of Shanghai University. We are grateful to Dr. Hong-Mei Deng of the Nuclear Magnetic Resonance Laboratory in this center for her assistance in measuring NMR spectra. We also thank Dr. Ying Wen, Professor Tao Yi, and Mr. Gang-Feng Tang of the Chemistry Department in Fudan University for their assistance in collecting mass spectral data.

■

ABBREVIATIONS ACN, acetonitrile; BD, 1,3-butadiene; BSO, L-buthionine sulfoximine; CBO, 1-chloro-3-buten-2-one; CHB, 1-chloro-2hydroxy-3-butene; COSY, correlation spectroscopy; CT DNA, K


Article

Chemical Research in Toxicology butene to 1-chloro-3-buten-2-one, a bifunctional alkylating agent. Chem. Res. Toxicol. 25, 2600−2607. (16) van der Veen, B. S., de Winther, M. P., and Heeringa, P. (2009) Myeloperoxidase: molecular mechanisms of action and their relevance to human health and disease. Antioxid. Redox Signaling 11, 2899−2937. (17) Nussbaum, C., Klinke, A., Adam, M., Baldus, S., and Sperandio, M. (2013) Myeloperoxidase: a leukocyte-derived protagonist of inflammation and cardiovascular disease. Antioxid. Redox Signaling 18, 692−713. (18) U.S. EPA (2002) 1,3-Butadiene, CASRN 106-99-0. Integrated Risk Information System, U.S. Environmental Protection Agency, National Center for Environmental Assessment, Washington, DC, available at http://cfpub.epa.gov/ncea/iris/iris_documents/ documents/subst/0139_summary.pdf (accessed Aug 21, 2016). (19) Sun, L., Pelah, A., Zhang, D.-P., Zhong, Y.-F., An, J., Yu, Y.-X., Zhang, X.-Y., and Elfarra, A. A. (2013) Formation of fused-ring 2′deoxycytidine adducts from 1-chloro-3-buten-2-one, an in vitro 1,3butadiene metabolite, under in vitro physiological conditions. Chem. Res. Toxicol. 26, 1545−1553. (20) Zeng, F.-M., Liu, L.-Y., Zheng, J., Kong, C., An, J., Yu, Y.-X., Zhang, X.-Y., and Elfarra, A. A. (2016) Identification of a fused-ring 2′deoxyadenosine adduct formed in human cells incubated with 1chloro-3-buten-2-one, a potential reactive metabolite of 1,3-butadiene. Chem. Res. Toxicol. 29, 1041−1050. (21) Zheng, J., Li, Y., Yu, Y.-X., An, J., Zhang, X.-Y., and Elfarra, A. A. (2015) Novel adducts from the reaction of 1-chloro-3-buten-2-one with 2′-deoxyguanosine. Structural characterization and potential as tools to investigate 1,3-butadiene carcinogenicity. Chem.-Biol. Interact. 226, 40−48. (22) Herrmann, K., Engst, W., Appel, K. E., Monien, B. H., and Glatt, H. (2012) Identification of human and murine sulfotransferases able to activate hydroxylated metabolites of methyleugenol to mutagens in Salmonella typhimurium and detection of associated DNA adducts using UPLC-MS/MS methods. Mutagenesis 27, 453−462. (23) Schumacher, F., Herrmann, K., Florian, S., Engst, W., and Glatt, H. (2013) Optimized enzymatic hydrolysis of DNA for LC-MS/MS analyses of adducts of 1-methoxy-3-indolylmethyl glucosinolate and methyleugenol. Anal. Biochem. 434, 4−11. (24) Marutani, E., Kosugi, S., Tokuda, K., Khatri, A., Nguyen, R., Atochin, D. N., Kida, K., Van Leyen, K., Arai, K., and Ichinose, F. (2012) A novel hydrogen sulfide-releasing N-methyl-D-aspartate receptor antagonist prevents ischemic neuronal death. J. Biol. Chem. 287, 32124−32135. (25) Beranek, D. T., Weis, C. C., and Swenson, D. H. (1980) A comprehensive quantitative analysis of methylated and ethylated DNA using high pressure liquid chromatography. Carcinogenesis 1, 595−606. (26) Daly, M. M., Allfrey, V. G., and Mirsky, A. E. (1950) Purine and pyrimidine contents of some desoxypentose nucleic acids. J. Gen. Physiol. 33, 497−510. (27) Voulgaridou, G.-P., Anestopoulos, I., Franco, R., Panayiotidis, M. I., and Pappa, A. (2011) DNA damage induced by endogenous aldehydes: current state of knowledge. Mutat. Res., Fundam. Mol. Mech. Mutagen. 711, 13−27. (28) Tang, M.-S., Wang, H.-T., Hu, Y., Chen, W.-S., Akao, M., Feng, Z., and Hu, W. (2011) Acrolein induced DNA damage, mutagenicity and effect on DNA repair. Mol. Nutr. Food Res. 55, 1291−1300. (29) Powley, M. W., Jayaraj, K., Gold, A., Ball, L. M., and Swenberg, J. A. (2003) 1,N2-Propanodeoxyguanosine adducts of the 1,3-butadiene metabolic, hydroxymethylvinyl ketone. Chem. Res. Toxicol. 16, 1448− 1454. (30) Eder, E., Hoffman, C., and Deininger, C. (1991) Identification and characterization of deoxyguanosine adducts of methyl vinyl ketone and ethyl vinyl ketone. Genotoxicity of the ketones in the SOS Chromotest. Chem. Res. Toxicol. 4, 50−57. (31) Kowal, E. A., Wickramaratne, S., Kotapati, S., Turo, M., Tretyakova, N., and Stone, M. P. (2014) Major groove orientation of the (2S)-N6-(2-hydroxy-3-buten-1-yl)-2′-deoxyadenosine DNA adduct induced by 1,2-epoxy-3-butene. Chem. Res. Toxicol. 27, 1675−1686.

(32) Kowal, E. A., Seneviratne, U., Wickramaratne, S., Doherty, K. E., Cao, X., Tretyakova, N., and Stone, M. P. (2014) Structures of exocyclic R,R- and S,S-N6,N6-(2,3-dihydroxybutan-1,4-diyl)-2′-deoxyadenosine adducts induced by 1,2,3,4-diepoxybutane. Chem. Res. Toxicol. 27, 805−817. (33) Seneviratne, U., Antsypovich, S., Dorr, D. Q., Dissanayake, T., Kotapati, S., and Tretyakova, N. (2010) DNA oligomers containing site-specific and stereospecific exocyclic deoxyadenosine adducts of 1,2,3,4-diepoxybutane: synthesis, characterization, and effects on DNA structure. Chem. Res. Toxicol. 23, 1556−1567. (34) Carmical, J. R., Nechev, L. V., Harris, C. M., Harris, T. M., and Lloyd, R. S. (2000) Mutagenic potential of adenine N6 adducts of monoepoxide and diolepoxide derivatives of butadiene. Environ. Mol. Mutagen. 35, 48−56. (35) Kotapati, S., Wickramaratne, S., Esades, A., Boldry, E. J., Quirk Dorr, D., Pence, M. G., Guengerich, F. P., and Tretyakova, N. Y. (2015) Polymerase bypass of N6-deoxyadenosine adducts derived from epoxide metabolites of 1,3-butadiene. Chem. Res. Toxicol. 28, 1496− 1507. (36) Kotapati, S., Maddukuri, L., Wickramaratne, S., Seneviratne, U., Goggin, M., Pence, M. G., Villalta, P., Guengerich, F. P., Marnett, L., and Tretyakova, N. (2012) Translesion synthesis across 1,N6-(2hydroxy-3-hydroxymethylpropan-1,3-diyl)-2′-deoxyadenosine (1,N6-γHMHP-dA) adducts by human and archebacterial DNA polymerases. J. Biol. Chem. 287, 38800−38811.

L


Characterization of the Major Purine and Pyrimidine Adducts Formed

Recommend Documents