Identification of DNA Adducts of Acetaldehyde - American Chemical

Peter W. Villalta, and Stephen S. Hecht*. University of Minnesota Cancer Center, Minneapolis, Minnesota 55455. Received May 25, 2000. Acetaldehyde is ...
0 downloads 0 Views 106KB Size
Chem. Res. Toxicol. 2000, 13, 1149-1157

1149

Identification of DNA Adducts of Acetaldehyde Mingyao Wang, Edward J. McIntee, Guang Cheng, Yongli Shi, Peter W. Villalta, and Stephen S. Hecht* University of Minnesota Cancer Center, Minneapolis, Minnesota 55455 Received May 25, 2000

Acetaldehyde is a mutagen and carcinogen which occurs widely in the human environment, sometimes in considerable amounts, but little is known about its reactions with DNA. In this study, we identified three new types of stable acetaldehyde DNA adducts, including an interstrand cross-link. These were formed in addition to the previously characterized N2-ethylidenedeoxyguanosine. Acetaldehyde was allowed to react with calf thymus DNA or deoxyguanosine. The DNA was isolated and hydrolyzed enzymatically; in some cases, the DNA was first treated with NaBH3CN. Reaction mixtures were analyzed by HPLC, and adducts were isolated and characterized by UV, 1H NMR, and MS. The major adduct was N2ethylidenedeoxyguanosine (1), which was identified as N2-ethyldeoxyguanosine (7) after treatment of the DNA with NaBH3CN. The new acetaldehyde adducts were 3-(2-deoxyribos1-yl)-5,6,7,8-tetrahydro-8-hydroxy-6-methylpyrimido[1,2-a]purine-10(3H)one (9), 3-(2-deoxyribos-1-yl)-5,6,7,8-tetrahydro-8-(N2-deoxyguanosyl)- 6-methylpyrimido[1,2-a]purine-10(3H)one (12), and N2-(2,6-dimethyl-1,3-dioxan-4-yl)deoxyguanosine (11). Adduct 9 has been previously identified in reactions of crotonaldehyde with DNA. However, the distribution of diastereomers was different in the acetaldehyde and crotonaldehyde reactions, indicating that the formation of 9 from acetaldehyde does not proceed through crotonaldehyde. Adduct 12 is an interstrand cross-link. Although previous evidence indicates the formation of cross-links in DNA reacted with acetaldehyde, this is the first reported structural characterization of such an adduct. This adduct is also found in crotonaldehyde-deoxyguanosine reactions, but in a diastereomeric ratio different than that observed here. A common intermediate, N2-(4-oxobut-2-yl)deoxyguanosine (6), is proposed to be involved in formation of adducts 9 and 12. Adduct 11 is produced ultimately from 3-hydroxybutanal, the major aldol condensation product of acetaldehyde. Levels of adducts 9, 11, and 12 were less than 10% of those of N2-ethylidenedeoxyguanosine (1) in reactions of acetaldehyde with DNA. As nucleosides, adducts 9, 11, and 12 were stable, whereas N2-ethylidenedeoxyguanosine (1) had a half-life of 5 min. These new stable adducts of acetaldehyde may be involved in determination of its mutagenic and carcinogenic properties.

Introduction Acetaldehyde exists commonly in the human environment (1, 2). It is the primary metabolite of ethanol and is also an intermediate in the metabolism of sugars. It is present in alcoholic beverages, in plant juices and essential oils, and in roasted coffee. Its concentration in tobacco smoke (18-1400 µg/cigarette) is more than 1000 times greater than those of polycyclic aromatic hydrocarbons and tobacco-specific nitrosamines (3). It occurs commonly in mobile source emissions, is a product of photooxidation of hydrocarbons, and is found in the atmosphere (1, 2). It is an intermediate in the manufacturing of numerous chemicals; the production capacity in the United States in 1989 was estimated to be 443 000 tons (2). More than 200 000 workers in the United States may be exposed to acetaldehyde (2). Acetaldehyde causes mutations, sister chromatid exchanges, micronuclei and aneuploidy in cultured mammalian cells, and gene mutations in bacteria (1, 2). In rats, inhalation of acetaldehyde results in tumors of the * To whom correspondence should be addressed: University of Minnesota Cancer Center, Box 806 Mayo, 420 Delaware St. SE, Minneapolis, MN 55455. Telephone: (612) 624-7604. Fax: (612) 6265135. E-mail: [email protected].

respiratory tract, particularly adenocarcinomas and squamous cell carcinomas of the nasal mucosa (1, 2). In hamsters, acetaldehyde inhalation produces laryngeal carcinomas and enhances the production of respiratory tract tumors induced by intratracheal instillation of benzo[a]pyrene (1, 2). It has been associated with alcoholrelated cancers in humans who have certain genetic polymorphisms in aldehyde dehydrogenase, which would cause accumulation of acetaldehyde following heavy alcohol intake (2). Acetaldehyde is considered by the International Agency for Research on Cancer to be possibly carcinogenic in humans (2). Acetaldehyde reacts with the exocyclic amino group of dG and DNA to form an unstable Schiff base, which can be stabilized by reduction, producing N2-ethyl-dG (4-6). N2-Ethyl-dG has been detected in the DNA of peripheral white blood cells of alcohol abusers, in human buccal cell DNA following exposure to acetaldehyde, and in human urine (7-9). N2-Ethyl-dG is efficiently incorporated into DNA by mammalian DNA polymerases (9). In reactions of acetaldehyde with deoxyribonucleosides, dG was the most reactive; products were also observed in reactions with dAdo and dCyd (6). The products in the dG reactions were identified after reduction with NaBH4. In addition to N2-ethyl-dG, Vaca et al. detected N2-(3-hydroxybut-1-

10.1021/tx000118t CCC: $19.00 © 2000 American Chemical Society Published on Web 10/10/2000

1150

Chem. Res. Toxicol., Vol. 13, No. 11, 2000

yl)-dG and N2-(4-hydroxybut-2-yl)-dG (6). Cooperative reactions of acetaldehyde and ethanol with deoxyribonucleosides and acetaldehyde and malondialdehyde with dAdo have also been reported (6, 10-12). In ongoing studies of crotonaldehyde-DNA adducts, we have characterized products resulting from the reaction with DNA of 2-(2-hydroxypropyl)-4-hydroxy-6-methyl-1,3-dioxane (paraldol), the dimer of 3-hydroxybutanal, which is present in aqueous solutions of crotonaldehyde (13). Since 3-hydroxybutanal is also readily formed by aldol condensation of acetaldehyde, we hypothesized that paraldol-derived adducts may be produced in the reactions of acetaldehyde with DNA. In this study, we identified three new types of acetaldehyde-DNA adducts. These were distinct from the paraldol-derived adducts which we observed in the crotonaldehyde-DNA reactions.

Experimental Section HPLC Analysis. HPLC was carried out with Waters Associates (Milford, MA) systems equipped with a model 991 or 996 photodiode array detector and an RF-10 AXL spectrofluorometric detector (Shimadzu Scientific Instruments, Columbia, MD). Columns and solvent elution systems were as follows. For system 1, we used two 4.6 mm × 25 cm Supelcosil LC 18-BD columns (Supelco, Bellefonte, PA) with isocratic elution by 5% CH3CN in 10 mM sodium phosphate buffer (pH 7) for 10 min and then a gradient from 5 to 25% CH3CN over the course of 60 min at a rate of 1 mL/min and UV detection (254 nm). This system was used for analysis of all acetaldehyde adducts. For system 2, we used the same columns as system 1, with elution by a gradient from 0 to 30% CH3CN in 10 mM sodium phosphate buffer (pH 7) over the course of 40 min at a rate of 1 mL/min and UV detection (254 nm). This system was used for quantitation of dG. For system 3, we used one of the columns employed in system 1, with elution by a gradient from 10 to 40% CH3CN in 20 mM ammonium acetate buffer (pH 3) over the course of 60 min at a flow rate of 0.5 mL/min with UV detection (254 nm). This system was used for purification of adduct 10 as the guanine base. For system 4, we employed a 4.6 mm × 25 cm, 5 µm OD5 octadecyl column (Burdick and Jackson, Baxter, McGaw Park, IL) with elution by a gradient from 20 to 80% CH3CN in H2O over the course of 40 min at a rate of 1 mL/min with UV detection at 254 nm. This system was used for desalting of adducts collected in system 1. For system 5, we used the same column and flow rate as in system 4, with a gradient from 40 to 60% CH3CN in H2O over the course of 40 min with UV detection (365 nm). This system was employed for analysis of 2,4-dinitrophenylhydrazones. For system 6, we used two 4.6 mm × 25 cm Partisil-10 SCX strong cation exchange columns (Whatman, Clifton, NJ) and an elution medium of 100 mM ammonium phosphate buffer (pH 2) at a rate of 1 mL/min with fluorescence detection (excitation at 290 nm and emission at 380 nm). This system was used for quantitation of adduct 9 (as the guanine base) and guanine. GC Analysis. GC with flame ionization detection was performed with a HP 6890 series gas chromatograph (HewlettPackard, Palo Alto, CA) with a 30 m × 0.32 mm i.d., 3.0 µm film thickness, DB-1 column (J&W Scientific, Folsom, CA). One microliter of sample was injected in the split mode (1:100). The injector temperature was 200 °C, and He was used as a carrier gas (32 cm/s at 40 °C). The initial temperature of the oven was 40 °C, which was maintained for 4 min, and followed by a programmed rate of 10 °C/min to 210 °C. The flame ionization detector was set at 25 °C with flow rates of H2 (40 mL/min), air (400 mL/min), and He (25 mL/min). The retention times of acetaldehyde, crotonaldehyde, and paraldol were 2.52, 9.10, and 11.97 min, respectively. MS Analysis. LC/MS analysis was performed on a FinniganMAT LCQ Deca instrument (Thermoquest LC/MS Division, San

Wang et al. Jose, CA) interfaced with a Waters 600 HPLC multisolvent delivery system. The API source of the MS was set as follows: voltage, 4.5 kV; current, 7.0 µA; and capillary interface temperature, 350 °C. The analyses were performed in the positive ion electrospray ionization (ESI) mode. HPLC system 4 was used, except that the eluting solvent was a gradient from 25 to 65% methanol in H2O containing 1% acetic acid. This system was used for analysis of all adducts except the guanine analogue of 12, which was analyzed using HPLC system 3. 1H NMR. Spectra were acquired on an 800 MHz instrument (Varian, Inc., Palo Alto, CA) using standard 5 mm tubes or 3 mm Shigemi tubes (Shigemi, Inc., Allison Park, PA). Chemicals and Enzymes. Crotonaldehyde, acetaldehyde, and NaBH3CN were purchased from Aldrich Chemical Co. (Milwaukee, WI). Calf thymus DNA was obtained from Sigma Chemical Co. (St. Louis, MO). Alkaline phosphatase was procured from Boehringer Mannheim Co. (Indianapolis, IN). Paraldol, diastereomers of adduct 9, and N2-ethyl-dG were synthesized as described previously (14-16). All other chemicals and enzymes were obtained from Sigma. N2-(2,6-Dimethyl-1,3-dioxan-4-yl)-dG (11, Scheme 1, peaks 4-6, Figures 1 and 3). This was prepared in reactions with DNA or dG. For the DNA reactions, acetaldehyde (12 mmol) was allowed to react with calf thymus DNA (100 mg) in 6 mL of 0.1 M phosphate buffer (pH 7.0) for 96 h at 37 °C. The DNA was precipitated by addition of ethanol and then hydrolyzed enzymatically as described below. For the dG reactions, acetaldehyde (22.5 mmol) was allowed to react with dG (0.17 mmol) in 10 mL of 0.1 M phosphate buffer (pH 7.0) for 55 h at 37 °C. Peaks 4-6 were collected from HPLC system 1 followed by desalting using system 4 and were obtained in 2% yield based on dG. Peak 4: 1H NMR (DMSO-d6) δ 10.7 (bs, 1H, dG-N1-H), 7.93 (s, 1H, dG-C8-H), 7.24 (bs, 1H, dG-N2-H), 6.14 (dd, J ) 7.2, 7.2 Hz, 1H, 1′-H), 5.40 (dd, J ) 9.6, 10.4 Hz, 1H, dioxane 4-H), 5.25 (s, 1H, 3′-OH), 4.87 (bs, 1H, 5′-OH), 4.83 (m, 1H, dioxane 2-H), 4.36 (m, 1H, 3′-H), 3.80 (m, 2H, 4′-H and dioxane 6-H), 3.57 (m, 1H, 5′-Ha), 3.50 (m, 1H, 5′-Hb), 2.61 (m, 1H, 2′Ha), 2.18 (m, 1H, 2′-Hb), 1.75 (m, 1H, dioxane 5eq-H), 1.30 (m, 1H, dioxane 5ax-H), 1.18 (d, J ) 4.8 Hz, dioxane 2-CH3), 1.15 (d, J ) 4.8 Hz, dioxane 6-CH3); MS (positive ion LC-ESI) m/z (relative intensity) 382 (MH+, 100), 266 (BH+, 33), 248 (BH+ H2O, 8), 222 (BH+ - CH3CHO, 54), 204 (BH+ - CH3CHO H2O, 20), 178 (BH+ - 2CH3CHO, 49), 152 (15); MS/MS of m/z 382; 266 (65), 222 (4); UV (H2O) λmax 254, 275 (sh) nm. Peak 5: 1H NMR (DMSO-d ) δ 10.8 (bs, 1H, dG-N1-H), 7.90 (s, 1H, dG6 C8-H), 7.25 (bs, 1H, dG-N2-H), 6.12 (dd, J ) 6.4, 6.4 Hz, 1H, 1′-H), 5.39 (dd, J ) 9.4, 9.4 Hz, 1H, dioxane 4-H), 5.26 (s, 1H, 3′-OH), 4.84 (bs, 1H, 5′-OH), 4.83 (bs, 1H, dioxane 2-H), 4.37 (m, 1H, 3′-H), 3.82 (m, 1H, dioxane 6-H), 3.77 (dt, J ) 4.2, 4.2 Hz, 1H, 4′-H), 3.53 (m, 1H, 5′-Ha), 3.44 (m, 1H, 5′-Hb), 2.70 (m, 1H, 2′-Ha), 2.18 (m, 1H, 2′-Hb), 1.75 (m, 1H, dioxane 5eq-H), 1.31 (m, 1H, dioxane 5ax-H), 1.18 (d, J ) 4 Hz, 3H, dioxane 2-CH3), 1.14 (d, J ) 6.4 Hz, 3H, dioxane 6-CH3); MS (positive ion LCESI) m/z (relative intensity) 382 (MH+, 100), 266 (BH+, 38), 248 (BH+ - H2O, 8), 222 (BH+ - CH3CHO, 54), 204 (BH+ - CH3CHO - H2O, 20), 178 (BH+ - 2CH3CHO, 48), 152 (15), 117 (9); UV (H2O) λmax 254, 275 (sh) nm. 3-(2-Deoxyribos-1-yl)-5,6,7,8-tetrahydro-8-(N2-deoxyguanosyl)-6-methylpyrimido[1,2-a]purine-10(3H)one (12, Scheme 1, peak 3, Figures 1 and 3). Acetaldehyde was allowed to react with DNA as described above for the preparation of adduct 11. The DNA was enzymatically hydrolyzed, and adduct 12 (peak 3) was collected as described above: 1H NMR (D2O) δ 8.01 (s, 1H, C2-H or C8′-H), 8.00 (s, 1H, C2-H or C8′H), 6.87 (m, 1H, C8-H), 6.44 (dd, J ) 7.3, 6.7 Hz, 1H, 1′-H), 6.35 (dd, J ) 6.7, 6.7 Hz, 1H, 1′-H), 4.68 (m, 1H, 3′-H), 4.17 (m, 2H, 4′-H), 3.86 (m, 5H, 5′-Ha,b and C6-H), 3.12 (m, 1H, 2′-H), 2.86 (m, 1H, 2′-H), 2.56 (m, 3H, 2′-Ha,b and C7-H), 1.87 (td, J ) 13.1, 3.1 Hz, 1H, C7-H), 1.41 (d, J ) 6.7 Hz, 3H, C6-CH3); MS (positive ion LC-ESI) m/z (relative intensity) 587 (MH+, 100), 471 (MH+ - 116, 23), 355 (MH+ - 232, 18), (294, 48); UV (H2O) λmax 258, 280 (sh) nm. Hydrolysis of 12 to the guanine base was

Acetaldehyde DNA Adducts

Chem. Res. Toxicol., Vol. 13, No. 11, 2000 1151

Scheme 1. Product Formation in the Reaction of Acetaldehyde with DNAa

a With the exception of 7, adducts are shown as the structures in DNA to conserve space. In the text, they are identified at the nucleoside level. EH represents enzyme hydrolysis.

Figure 2. Chromatogram obtained upon HPLC analysis (system 1) of an enzymatic hydrolysate of DNA that had been allowed to react with acetaldehyde and then treated with NaBH3CN prior to enzyme hydrolysis.

Figure 1. Chromatogram obtained upon HPLC analysis (system 1) of the products of the reaction of acetaldehyde with dG. The identities of peaks 1-6 are as follows: 1 and 2, two diastereomers of 9 (Scheme 1); 3, 12 (Scheme 1); and 4-6, diastereomers of 11 (Scheme 1). See the text for additional information. carried out by dissolving 12 in 300 µL of 0.1 N HCl and heating the mixture at 37 °C overnight. The resulting mixture was neutralized and then collected using HPLC system 3 (retention time of 34.5 min): MS (positive ion LC-ESI) m/z (relative intensity) 393 (M + K, 100), 355 (MH+, 62), 204 (MH+ guanine); UV (H2O) λmax 246, 275 (sh) nm.

Reactions. (1) Acetaldehyde and DNA. For characterization of adducts, these reactions were carried out as described above for adduct 11. For quantitation of adducts 1, 9, and 11, acetaldehyde (0.01-40 mM) was allowed to react with DNA (25 mg) in 1.5 mL of 0.1 M phosphate buffer (pH 7.0) for various time periods at 37 °C. DNA was precipitated by addition of ethanol and then washed sequentially with 70% ethanol and 95% ethanol until acetaldehyde could not be detected in the washings by analysis of its 2,4-dinitrophenylhydrazone. (i) Neutral Thermal and Acid Hydrolysis of DNA. Modified DNA (5 mg) was dissolved in 1 mL of 10 mM sodium cacodylate buffer (pH 7.0). The solution was heated at 100 °C for 1 h. HCl (1 N) was added to precipitate the DNA. The supernatant was analyzed for released acetaldehyde (from adduct 1) by derivatization with 2,4-dinitrophenylhydrazine

1152

Chem. Res. Toxicol., Vol. 13, No. 11, 2000

Figure 3. Chromatogram obtained upon HPLC analysis (system 1) of an enzymatic hydrolysate of DNA that had been allowed to react with acetaldehyde. The identities of peaks 1-6 are as follows: 1 and 2, two diastereomers of 9 (Scheme 1); 3, 12 (Scheme 1); and 4-6, diastereomers of 11 (Scheme 1). See the text for additional information. reagent. The 2,4-dinitrophenylhydrazone was quantified using HPLC system 5. The precipitated DNA was dissolved in 0.5 mL of 0.1 N HCl and heated at 85 °C for 1 h. Adduct 9 and its guanine base were analyzed using HPLC system 6. (ii) Enzyme Hydrolysis of DNA. Modified DNA (2.5 mg) was dissolved in 1 mL of 10 mM Tris-HCl/5 mM MgCl2 buffer (pH 7.0). The mixture was incubated at 37 °C for 70 min with DNase I, phosphodiesterase I, and alkaline phosphatase. The hydrolysate was analyzed for adducts 9, 11, and 12. For detection of N2-ethyl-dG, modified DNA (2.5 mg) was dissolved in 1 mL of 10 mM Tris-HCl/5 mM MgCl2 buffer (pH 7.0). NaBH3CN (10 mg) was added to the mixture, and it was allowed to stand for 30 min at room temperature. A second 10 mg aliquot of NaBH3CN was then added, and the mixture was incubated at 37 °C for 30 min. This procedure was repeated one more time; then the pH was adjusted to 7 with 8-10 µL of 1 N HCl, and the DNA was hydrolyzed enzymatically as described above. All adducts were analyzed using HPLC system 1. (2) Acetaldehyde and dG. These reactions were carried out as described above for the preparation of adduct 11.

Results The purity of the acetaldehyde used in this study was established by GC. We did not detect 3-hydroxybutanal, paraldol, or crotonaldehyde. Initially, we investigated the reaction of acetaldehyde with dG. HPLC analysis of this reaction mixture produced the chromatogram illustrated in Figure 1. The major peak eluting at 31 min has a half-life of less than 5 min at 37 °C. Its UV spectrum has a maximum at 254 nm and a shoulder at 275 nm. LC/ESI-MS1 analysis shows a base peak of m/z 312 (M + 1 + H2O), as well as peaks at m/z 294 (M + 1, relative intensity of 2), 587 (2M + 1, relative intensity of 23), and 605 (2M + 1 + H2O, relative intensity of 14). Collection of this peak followed by treatment with 2,4-dinitrophenylhydrazine reagent 1 Abbreviations: LC/ESI-MS, liquid chromatography/electrospray ionization mass spectrometry; paraldol, 2-(2-hydroxypropyl)-4-hydroxy6-methyl-1,3-dioxane.

Wang et al.

gives the 2,4-dinitrophenylhydrazone of acetaldehyde. Treatment of the acetaldehyde-dG reaction mixture with NaBH3CN produces a major product which was identified as N2-ethyl-dG (7, Scheme 1) by comparison of its UV spectrum, MS data, and HPLC retention time to those of a standard. These data confirmed the identity of the major adduct in Figure 1 as N2-ethylidene-dG (1, Scheme 1). Acetaldehyde was then allowed to react with calf thymus DNA. When the DNA was treated with NaBH3CN prior to enzyme hydrolysis, the chromatogram illustrated in Figure 2 was obtained. The major adduct peak eluting at 38.9 min was identified as N2-ethyl-dG by comparison of its UV spectrum and retention time to those of a synthetic standard. Considerable effort was expended to optimize the NaBH3CN reduction conditions, which are described in the Experimental Section. These conditions resulted in virtually complete conversion of N2-ethylidene-dG to N2-ethyl-dG in DNA. When the NaBH3CN step was omitted prior to enzymatic hydrolysis, the chromatogram illustrated in Figure 3 was obtained. Because of its instability, N2-ethylidenedG was not detected in these hydrolysates. However, we did observe peaks 1-6, as in the reactions with dG illustrated in Figure 1. These peaks were identified here for the first time as products of the reaction of acetaldehyde with DNA. The UV spectrum of peak 2 is illustrated in Figure 4. Analysis of peak 2 by LC/ESI-MS gives a base peak at m/z 338. 1H NMR data for peak 2 are summarized in Table 1. These spectral data are identical to those of 3-(2deoxyribos-1-yl)-5,6,7,8-tetrahydro-8-hydroxy-6-methylpyrimido[1,2-a]purine-10(3H)one (9), the structure of which is illustrated in Table 1 and Scheme 1. Two diastereomers of this 1,N2-propano-dG adduct 9, with the 6-methyl and 8-hydroxy groups trans to each other, have been previously identified in the reaction of crotonaldehyde with dG and DNA (15, 17). These standards coelute with peaks 1 and 2 of Figure 3 and have UV spectra identical to those of peaks 1 and 2. The ratio of peak 1 to peak 2 was 1:5, in contrast to the 1.5:1 ratio of the two peaks obtained upon reaction of crotonaldehyde with DNA (13). These 1,N2-propano-dG adducts were also produced in the reaction of acetaldehyde with dG, and the ratio of peak 1 to peak 2 was 1:1.7 (Figure 1). The UV spectrum of peak 3 of Figure 3 is shown in Figure 4. This spectrum is similar to that of peak 2 and other N2-substituted dG derivatives. LC/ESI-MS analysis of peak 3 produces a base peak of m/z 587, which is M + 1, as well as peaks at m/z 471 and 355 corresponding to successive losses of 116 mass units (deoxyribose - 1) from M + 1. These data suggest that there are two deoxyribonucleoside moieties in the molecule. This was confirmed by the 1H NMR spectrum which is summarized in Table 2. This spectrum clearly shows that there are two dG residues in this adduct, indicating that it is a cross-link. The structure (12, Scheme 1 and Table 2) was deduced from the 1H NMR spectrum. Since there are two dG moieties and the molecular weight is 586, the linking portion of the molecule must have a mass of 54, corresponding to one degree of unsaturation. The 1H NMR data are not consistent with the presence of a double bond; therefore, the linking portion must be a ring. The ring protons are analogous in chemical shift and multiplicity to those of the 1,N2-propano-dG adducts (9, Scheme 1 and Table 1). All assignments were confirmed

Acetaldehyde DNA Adducts

Chem. Res. Toxicol., Vol. 13, No. 11, 2000 1153

Figure 4. UV spectra of peaks 2-6 of Figure 3. These peaks were identified as follows: 2, adduct 9 of Scheme 1; 3, adduct 12; and 4-6, adduct 11 (see the text). Table 1. 800 MHz 1H NMR Data for Peak 2a

C2-H

N5-Hb

C6-H

C6-CH3

C7-Ha

C7-Hb

C8-H

C8-OHb

7.87 (s)

7.86 (m)

3.69 (m)

1.20 (d)

1.39 (dd, J ) 12, 12 Hz)

2.01 (d, J ) 14 Hz)

6.17 (bs)

6.57 (m)

1′-H

2′-Hc

3′-H

3′-OH

4′-H

5′-Ha,b

5′-OH

6.08 (dd, J ) 6.4, 6.4 Hz)

2.16 (m)

4.31 (bs)

5.20 (m)

3.78 (m)

3.53 (m), 3.49 (m)

4.87 (m)

a

Values are chemical shifts (δ). b These assignments may be interchanged. c The other 2′-H was obscured by DMSO.

by a COSY spectrum. Additional evidence was obtained by HPLC analysis of the sample used to obtain the 1H NMR. This material was virtually 100% pure when the 1H NMR was recorded. It was reanalyzed by HPLC after 30 days standing at 4 °C, and gave the chromatogram shown in Figure 5. The peaks eluting at 15.8 and 42.9 min are dG and the second eluting diastereomer of the 1,N2-propano-dG adduct 9 (e.g., identical to peak 2 of Figure 3), identified by comparison to standards. This demonstrates partial hydrolysis of adduct 12 to adduct 9 upon prolonged storage. Further confirmation of the structure of adduct 12 was obtained by hydrolysis with acid, which produces a new peak with an HPLC retention time of 34.5 min. ESI-MS of this material shows an M + 1 of m/z 355 and a base peak of m/z 204 (M - guanine + 1). The UV spectrum of this product has maxima at 246 and 274 nm, values which are very similar to those of N2-ethylguanine. These results demonstrate that the two deoxyribose moieties of 12 are removed by hydrolysis, giving the guanine analogue of this adduct. Collectively, these data provide unambiguous proof of the structure of peak 3 as 12 (Scheme 1). This

cross-linked adduct was observed in reactions of acetaldehyde with double-stranded DNA and dG, but not with single-stranded DNA or poly(G). Therefore, it is an interstrand cross-link. Adduct 12 was also confirmed as a product of the reaction of acetaldehyde with dG (see peak 3 of Figure 1). The amount of adduct 12 (peak 3) relative to the other stable adducts (peaks 1, 2, and 4-6) was greater in the DNA reaction (Figure 3) than in the dG reaction (Figure 1), consistent with its structure. Reaction of crotonaldehyde with dG gave two diastereomers of adduct 12, identified by their UV and MS spectra (Figure 6). These two diastereomers were formed in equal amounts, as were the 1,N2-propano-dG adducts 9 (Figure 6). This is in contrast to the production of mainly the second eluting diastereomers of 9 and 12 from the reaction of acetaldehyde with DNA (e.g., peaks 2 and 3 of Figure 3). These results indicate that adduct 12 is formed from the same intermediate as adduct 9 (e.g., 6, Scheme 1). The UV spectra of peaks 4-6 of Figure 3 are illustrated in Figure 4. These spectra are similar to those of peaks 2 and 3. Peaks 4-6 were each collected and treated with

1154

Chem. Res. Toxicol., Vol. 13, No. 11, 2000

Wang et al.

Table 2. 800 MHz 1H NMR Data for Peak 3a

purine protons

deoxyribose protons

C2-Hb

C8′-Hb

1′-H

2′a,b

3′-Hc

4′-H

5′-Ha,b

8.01

8.00

6.44 (dd, J ) 6.7, 7.3 Hz) 6.35 (dd, J ) 6.7, 6.7 Hz)

2.56 (m) 2.86 (m) 3.12 (m)

4.68 (m)

4.17 (m)

3.86 (m)

ring protons C6-H

C6-CH3

C7-Ha

C7-Hb

C8-H

3.86 (m)

1.41 (d, J ) 6.7 Hz)

1.87 (td, J ) 3.1, 13.1 Hz)

2.56 (m)

6.87 (m)

a Recorded in D O; values are chemical shifts (δ). b These assignments could be interchanged. c A second 3′-H was obscured by D O 2 2 (4.8 ppm).

Figure 5. Chromatogram obtained upon HPLC analysis (system 1) of the 1H NMR sample of adduct 12, after having been stored in D2O for 30 days at 4 °C. The indicated peaks are dG, adduct 9, and adduct 12.

These were separated by HPLC, with the following retention times: acetaldehyde, 21.5 min; crotonaldehyde, 33.9 min; and 3-hydroxybutanal, 12.6 min. They were identified by comparison of their UV spectra and retention times to those of standards. Analysis of peaks 4-6 by ESI-MS demonstrates that each has a base peak of m/z 382, which is M + 1. Peaks 4-6 were also observed in the reaction of acetaldehyde with dG (Figure 1). They had the same HPLC retention times and UV spectra as peaks 4-6 of Figure 3. Peaks 4 and 5 from the reaction with dG were also analyzed by ESI-MS and by reaction with the 2,4-dinitrophenylhydrazine reagent, with results identical to those obtained for the corresponding DNA products. 1H NMR spectra were obtained on peaks 4 and 5 from the reaction of acetaldehyde with dG. These data are summarized in Table 3. The 1H NMR spectra are remarkably similar to that of N2-paraldol-dG [see the structure in Table 3, where R is 2-(2-hydroxypropyl)] (13). All assignments were confirmed by COSY spectra. These results demonstrate that peaks 4 and 5, and most likely peak 6, are isomers of N2-(2,6-dimethyl-1,3-dioxan-4-yl)dG (see the structure in Table 3 and 11 in Scheme 1). There are eight possible stereoisomers of 11, at least two of which are observed here. Levels of N2-ethylidene-dG (1) in DNA that had been allowed to react with acetaldehyde are summarized in Table 4. This adduct was quantified by subjecting the DNA to neutral thermal hydrolysis and measuring the amount of released acetaldehyde as its 2,4-dinitrophenylhydrazone, or by treating the DNA with NaBH3CN, followed by enzymatic hydrolysis and quantitation of N2ethyl-dG. Levels of adducts 9 and 11 could be quantified only in the 96 h reaction with 40 mM acetaldehyde, and were approximately 42 µmol/mol of G (adduct 9) and 280 µmol/mol of dG (adduct 11). These amounts were 1 and 7%, respectively, of that of N2-ethylidene-dG.

Figure 6. Chromatogram obtained upon HPLC analysis (system 1) of the products of the reaction of crotonaldehyde with dG. The indicated peaks are dG, adduct 9, and adduct 12.

Discussion

2,4-dinitrophenylhydrazine reagent. In each case, a mixture of the 2,4-dinitrophenylhydrazones of acetaldehyde, crotonaldehyde, and 3-hydroxybutanal was observed.

In this study, we have characterized three new types of stable DNA adducts of acetaldehyde: the 1,N2-propano-dG adduct 9, the N2-dimethyldioxane-dG adduct 11, and the cross-link adduct 12. While these adducts are formed in substantially lower yield than the major adduct

Acetaldehyde DNA Adducts

Chem. Res. Toxicol., Vol. 13, No. 11, 2000 1155

Table 3. 800 MHz 1H NMR Data for Peaks 4 and 5 vs N2-Paraldol-dGa

deoxyguanosyl protons C8-H

N1-H

N2-H

1′-H

2′-Ha,b

3′-H

4′-H

5′-Ha,b

3′-OH

5′-OH

peak 4 (R ) CH3)

7.93 (s) 10.7 (bs) 7.24 (bs) 6.14 (dd, 2.61 (m) 4.36 (m) 3.80 (m) J ) 7.2, 2.18 (m) 7.2 Hz)

3.57 (m) 5.25 (s) 3.50 (m)

4.87 (bs)

peak 5 (R ) CH3)

7.90 (s) 10.8 (bs) 7.25 (bs) 6.12 (dd, 2.70 (m) 4.37 (m) 3.77 (dt, 3.53 (m) 5.26 (s) J ) 6.4, 2.18 (m) J ) 4.2, 3.44 (m) 6.4 Hz) 4.2 Hz)

4.84 (bs)

N2-paraldol-dG 7.93 (s) 10.6 (bs) 6.49 (s) [R ) 2-(2-hydroxypropyl)]

6.13 (dd)

2.59 (m) 4.35 (m) 3.79 (m) 2.17 (m)

3.56 (m) 5.24 (d) 4.85 (m) 3.50 (m)

dioxane protons 2-H

2-CH3

4-H

5-Heq

5-Hax

6-H

6-CH3

peak 4 (R ) CH3)

4.83 (m)

1.18 (d, J ) 4.8 Hz)

5.40 (dd, J ) 9.6, 10.4 Hz)

1.75 (m)

1.30 (m)

3.80 (m)

1.15 (d, J ) 4.8 Hz)

peak 5 (R ) CH3)

4.83 (bs)

1.18 (d, J ) 4 Hz)

5.39 (dd, J ) 9.4, 9.4 Hz)

1.75 (m)

1.31 (m)

3.82 (m)

1.14 (d, J ) 6.4 Hz)

N2-paraldol-dG [R ) 2-(2-hydroxypropyl)]

4.79 (m)

5.37 (m)

1.76 (d)

1.30 (m)

3.79 (m)

1.13 (d)

a

-

Values are chemical shifts (δ). Table 4. N2-Ethylidene-dG (1) in DNA Reacted with Acetaldehydea N2-ethylidene-dG (1) quantitation (µmol/mol of dG)

[acetaldehyde] (mM)

reaction time (h)

by released acetaldehydeb

by reduction to N2-ethyl-dG (7)c

0.01 0.1 1.0 10 20 40

20 20 20 20 20 20 96

155 204 222 467 1200 2430 4150

NQd NQ NQ 544 1430 2800 NQ

a Acetaldehyde was allowed to react with DNA (25 mg) in 1.5 mL of 0.1 M phosphate buffer (pH 7.0) at 37 °C. b DNA was precipitated and then washed until acetaldehyde could no longer be detected. It was subjected to neutral thermal hydrolysis and the released acetaldehyde quantified as its 2,4-dinitrophenylhydrazone as described in the Experimental Section. c The DNA was treated with NaBH3CN as described in the Experimental Section and then enzymatically hydrolyzed and analyzed for adduct 7. d NQ, not quantified.

of acetaldehyde, N2-ethylidene-dG (1), they are stable at the nucleoside level, in contrast to 1, and may be more stable in DNA. The 1,N2-propano-dG adduct 9 is also formed in reactions of crotonaldehyde with dG and DNA (14, 16). Crotonaldehyde can be produced in aqueous solutions of acetaldehyde by aldol condensation to 3-hydroxybutanal (4) followed by dehydration (18). The crotonaldehyde generated in this way could then react with dG or DNA

to give adduct 9. However, our data indicate that this is not the mechanism by which adduct 9 is formed from acetaldehyde. In the reaction of acetaldehyde with DNA, the ratio of peak 1 to peak 2 (Figure 3), which are the two diastereomers of adduct 9, is 1:5, whereas the corresponding ratio in the reaction of crotonaldehyde with DNA is 1.5:1 (13). Therefore, we propose the mechanism illustrated in Scheme 1 for the formation of adduct 9. The major DNA adduct of acetaldehyde, N2-ethylidenedG (1), reacts with a second molecule of acetaldehyde to give intermediate 6. The transition state for this reaction is diastereomeric, and the approach of the second molecule of acetaldehyde from one direction is clearly favored, producing one diastereomer of 6 in excess over the other. This same preference is observed in the reaction with dG, but here the ratio is only 1:1.7. Vaca et al. isolated two diastereomers of N2-(4-hydroxybut-2yl)-dG from reactions of dG with acetaldehyde, after reduction with NaBH4, and proposed a similar mechanism of formation (6). Although their conditions were different from those employed here, their products may have been produced from 9, as NaBH4 reduction of 9 is known to give N2-(4-hydroxybut-2-yl)-dG (15). Intermediate 6 of Scheme 1 is also the most likely precursor to the cross-linked adduct 12. Reaction of 6 with dG in the opposite strand would produce Schiff base 10, which then reacts with N-1 to give 12. There is precedent for interstrand G-G cross-linking by molecules which form intermediates with linkers similar in size to that in structure 10. Cross-link formation is favored with a G two bases away from the C paired to the G where

1156

Chem. Res. Toxicol., Vol. 13, No. 11, 2000

the initial reaction occurs (19). Our results demonstrate that predominantly one diastereomer of 12 is formed in the acetaldehyde-DNA reaction. On standing for 30 days at 4 °C, this diastereomer is gradually converted back to the major diastereomer of 9 which we have observed in the reactions of acetaldehyde with DNA and dG. Therefore, it is highly probable that 12 and 9 are formed via the same intermediate (6) as illustrated in Scheme 1. In contrast to these results, we detected equal amounts of two diastereomers of 12 in the reaction of crotonaldehyde with dG, consistent with the formation of equal amounts of the two diastereomers of 9 in this reaction (Figure 6). Although we have not yet detected the cross-link adduct 12 in reactions of crotonaldehyde with DNA, it is likely that the two diastereomers of this adduct are present. Acetaldehyde is known to produce interstrand crosslinks in DNA (1, 2, 20, 21). It also causes cross-links in human lymphocytes in vitro without metabolic activation (2, 22). DNA-protein cross-links have been observed as well (2). Adduct 12 represents the first structural characterization of a DNA cross-link resulting from reaction with acetaldehyde. The availability of this structure should facilitate future studies on mechanisms of mutagenesis by acetaldehyde involving cross-linking. Matsuda and co-workers have identified specific tandem GG to TT mutations induced by acetaldehyde and have attributed these to intrastrand cross-linking between adjacent guanine bases, although the structure of the cross-link was not determined (21). We did not observe adduct 12 in single-stranded DNA or poly(G); however, we did not extensively investigate the possibility of intrastrand formation of 12, and it is possible that this would occur under other conditions. The aldol condensation product 3-hydroxybutanal (4) is clearly the initial source of the N2-dimethyldioxanedG adducts 11. This can react with the exocyclic amino group of dG producing Schiff base 5. Reaction of 5 with another molecule of acetaldehyde would produce intermediate 8, which then cyclizes giving 11. We did not investigate the presence of intermediate 5 in this study. Vaca et al. did obtain evidence for 5 in the reaction of acetaldehyde with dG by carrying out NaBH4 reduction, yielding N2-(3-hydroxybut-1-yl)-dG (6). However, it should be noted that NaBH4 reduction of 11 would also be expected to produce this product. An alternative pathway to 11 is illustrated in Scheme 1. 3-Hydroxybutanal can react with acetaldehyde, producing 2,6-dimethyl-4-hydroxy-1,3-dioxane (aldoxane) (2) which can yield 11 via intermediates 3 and 8. This is analogous to the reaction of paraldol with DNA (13). In summary, the results of this study provide new information about DNA adduct formation by acetaldehyde. This carcinogen is ubiquitous in the human environment, and exposures can be exceptionally high, for example, through cigarette smoking or alcohol consumption. The new adducts described here are formed in quantities of less than 10% of that of N2-ethylidene-dG, and require more than one molecule of acetaldehyde per molecule of dG. They were detectable in this study only at relatively high acetaldehyde concentrations, which, however, may partially reflect the relative insensitivity of the analytical methods. It is plausible that adducts resulting from reaction of more than one molecule of acetaldehyde could occur in vivo, and might be able to be detected with highly sensitive methods designed for application to humans. As an example, Tuma and co-

Wang et al.

workers have detected protein adducts resulting from two molecules of acetaldehyde and one molecule of malondialdehyde in the livers of rats treated with ethanol, and in patients with alcohol-induced liver disease (23, 24). As the adducts described here are stable in DNA, they potentially could play a significant role in the mutagenic and carcinogenic effects of acetaldehyde.

Acknowledgment. This study was supported by Grant CA-85702 from the National Cancer Institute. S.S.H. is an American Cancer Society Research Professor, supported by ACS Grant RP-00-138.

References (1) International Agency for Research on Cancer (1985) Allyl compounds, aldehydes, epoxides and peroxides. In Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans, Vol. 36, pp 101-132, International Agency for Research on Cancer, Lyon, France. (2) International Agency for Research on Cancer (1999) Re-evaluation of some organic chemicals, hydrazine and hydrogen peroxide (part two). In Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans, Vol. 71, pp 319-335, International Agency for Research on Cancer, Lyon, France. (3) Hoffmann, D., and Hoffmann, I. (1997) The changing cigarette, 1950-1995. J. Toxicol. Environ. Health 50, 307-364. (4) Hemminki, K., and Suni, R. (1984) Sites of reaction of glutaraldehyde and acetaldehyde with nucleosides. Arch. Toxicol. 55, 186-190. (5) Fang, J.-L., and Vaca, C. E. (1995) Development of a 32Ppostlabelling method for the analysis of adducts arising through the reaction of acetaldehyde with 2′-deoxyguanosine-3′-monophosphate and DNA. Carcinogenesis 16, 2177-2185. (6) Vaca, C. E., Fang, J.-L., and Schweda, E. K. H. (1995) Studies of the reaction of acetaldehyde with deoxynucleosides. Chem.-Biol. Interact. 98, 51-67. (7) Fang, J.-L., and Vaca, C. E. (1997) Detection of DNA adducts of acetaldehyde in peripheral white blood cells of alcohol abusers. Carcinogenesis 18, 627-632. (8) Vaca, C. E., Nilsson, J. A., Fang, J. L., and Grafstrom, R. C. (1998) Formation of DNA adducts in human buccal epithetial cells exposed to acetaldehyde and methylglyoxal in vitro. Chem.-Biol. Interact. 108, 197-208. (9) Matsuda, T., Terashima, I., Matsumoto, Y., Yabushita, H., Matsui, S., and Shibutani, S. (1999) Effective utilization of N2-ethyl-2′deoxyguanosine triphosphate during DNA synthesis catalyzed by mammalian replicative DNA polymerases. Biochemistry 38, 929935. (10) Fraenkel-Conrat, H., and Singer, B. (1988) Nucleoside adducts are formed by cooperative reaction of acetaldehyde and alcohols: possible mechanism for the role of ethanol in carcinogenesis. Proc. Natl. Acad. Sci. U.S.A. 85, 3758-3761. (11) Austin, J., Dosanjh, M. K., and Fraenkel-Conrat, H. (1993) Further studies of the mixed acetals of nucleosides. Biochimie 75, 511-515. (12) Le Curieux, F., Pluskota, D., Munter, T., Sjo¨holm, R., and Kronberg, L. (1998) Formation of a fluorescent adduct in the reaction of 2′-deoxyadenosine with a malonaldehyde-acetaldehyde condensation product. Chem. Res. Toxicol. 11, 989-994. (13) Wang, M., McIntee, E. J., Cheng, G., Shi, Y., Villalta, P. W., and Hecht, S. S. (2000) Identification of paraldol-deoxyguanosine adducts in DNA reacted with crotonaldehyde. Chem. Res. Toxicol. (in press). (14) Wang, M., Upadhyaya, P., Dinh, T. T., Bonilla, L. E., and Hecht, S. S. (1998) Lactols in hydrolysates of DNA reacted with R-acetoxy-N-nitrosopyrrolidine and crotonaldehyde. Chem. Res. Toxicol. 11, 1567-1573. (15) Chung, F.-L., and Hecht, S. S. (1983) Formation of cyclic 1,N2adducts by reaction of deoxyguanosine with R-acetoxy-N-nitrosopyrrolidine, 4-(carbethoxynitrosamino)butanal, or crotonaldehyde. Cancer Res. 43, 1230-1235. (16) Sako, M., Kawada, H., and Hirota, K. (1999) A convenient method for the preparation of N2-ethylguanine nucleosides and nucleotides. J. Org. Chem. 64, 5719-5721. (17) Chung, F.-L., Young, R., and Hecht, S. S. (1984) Formation of cyclic 1,N2-propanodeoxyguanosine adducts in DNA upon reaction with acrolein or crotonaldehyde. Cancer Res. 44, 990-995.

Acetaldehyde DNA Adducts (18) Baigrie, L. M., Cox, R. A., Slebocka-Tilk, H., Tencer, M., and Tidwell, T. T. (1985) Acid-catalyzed enolization and aldol condensation of acetaldehyde. J. Chem. Soc. 107, 3640-3645. (19) Gold, B., and Fan, Y.-H. (2000) Sequence specificity for DNA interstrand cross-linking by R,ω-alkanediol dimethylsulfonate esters: evidence for DNA distortion by the initial monofunctional lesion. 219th ACS National Meeting, San Francisco, CA, Division of Chemical Toxicology, March 2000, Abstract 100. (20) Ristow, H., and Obe, G. (1978) Acetaldehyde induces cross-links in DNA and causes sister-chromatid exchanges in human cells. Mutat. Res. 58, 115-119. (21) Matsuda, T., Kawanishi, M., Yagi, T., Matsui, S., and Takebe, H. (1998) Specific tandem GG to TT base substitutions induced by acetaldehyde are due to intra-strand crosslinks between adjacent guanine bases. Nucleic Acids Res. 26, 1769-1774.

Chem. Res. Toxicol., Vol. 13, No. 11, 2000 1157 (22) Lambert, B., Chen, Y., He, S.-M., and Sten, M. (1985) DNA crosslinks in human leucocyctes treated with vinyl acetate and acetaldehyde in vitro. Mutat. Res. 146, 301-303. (23) Tuma, D. J., Thiele, G. M., Xu, D., Klassen, L. W., and Sorrell, M. F. (1996) Acetaldehyde and malondialdehyde react together to generate distinct protein adducts in the liver during long-term ethanol administration. Hepatology 23, 872-880. (24) Rolla, R., Vay, D., Mottaran, E., Parodi, M., Traverso, N., Arico, S., Sartori, M., Bellomo, G., Klassen, L. W., Thiele, G. M., Tuma, D. J., and Albano, E. (2000) Detection of circulating antibodies against malondialdehyde-acetaldehyde adducts in patients with alcohol-induced liver disease. Hepatology 31, 878-884.

TX000118T