Chemical Structure Determination of DNA Bases Modified by Active

Dec 10, 2009 - However, the mechanisms underlying MR carcinogenesis still remain uncertain partly because the contributing constituents have not been ...
0 downloads 0 Views 502KB Size
134

Chem. Res. Toxicol. 2010, 23, 134–141

Chemical Structure Determination of DNA Bases Modified by Active Metabolites of Lucidin-3-O-primeveroside Yuji Ishii,*,† Toshiya Okamura,† Tomoki Inoue,† Kiyoshi Fukuhara,‡ Takashi Umemura,† and Akiyoshi Nishikawa† DiVision of Pathology, National Institute of Health Sciences, 1-18-1 Kamiyoga, Setagaya-ku, Tokyo 158-8501, Japan, and DiVision of Organic Chemistry, National Institute of Health Sciences, 1-18-1 Kamiyoga Setagaya-ku, Tokyo 158-8501, Japan ReceiVed September 2, 2009

Lucidin-3-O- primeveroside (LuP) is one of the components of madder root (Rubia tinctorum L.; MR) which is reported to be carcinogenic in the kidney and liver of rats. Since metabolism of LuP generates genotoxic compounds such as lucidin (Luc) and rubiadin (Rub), it is likely that LuP plays a key role in MR carcinogenesis. In the present study, the chemical structures of Luc-specific 2′-deoxyguanosine (dG) and 2′-deoxyadenosine (dA) adducts following the reactions of dG and dA with a Luc carbocation or quinone methide intermediate derived from Acetoxy-Luc were determined by liquid chromatography with photodiode array and electron spray ionizaion-mass spectrometry (LC-PDA-ESI/MS). The identification of the two measurable adducts as Luc-N2-dG and Luc-N6-dA was confirmed by NMR analysis. Subsequently, using a newly developed quantitative analytical method using LC-ESI/MS, the formation of Luc-N2-dG and Luc-N6-dA from the reaction of calf thymus DNA with Luc in the presence of S9 mixture was observed. The fact that this reaction with Rub also gave rise to the same dG and dA adducts strongly suggests that Rub genotoxicity involves a metabolic conversion to Luc. The precise determination of the modified DNA bases generated by LuP and the method for their analysis may contribute to further comprehension of the mode of action underlying carcinogenesis by MR and related anthraquinones. Introduction 1

Madder root (Rubia tinctorum; MR ) contains red coloring material used as a dye and as a food additive for a variety of foods/drinks in Japan. Our previous study revealed that MR, now prohibited as a food additive, has a potent carcinogenicity targeting the kidneys and livers of F344 rats (1). However, the mechanisms underlying MR carcinogenesis still remain uncertain partly because the contributing constituents have not been identified. We have speculated that lucidin-3-O-primeveroside (LuP) and its metabolites, lucidin (Luc) and rubiadin (Rub) take part in MR carcinogenesis by way of genotoxicity (Figure 1) (2).32P-postlabeling analysis has demonstrated the formation of LuP- and Luc-related DNA adducts in the livers, kidneys, and duodenums of mice, but the chemical structures of these adducts have remained undetermined (3). It has been generally accepted that DNA adduct formation could be a trigger for genotoxic chemical carcinogenesis. However, various base modifications of DNA do not always result in gene mutations because of the * Corresponding author. Tel: +81-3-3700-9819. Fax: +81-3-3700-1425. E-mail: [email protected]. † Division of Pathology. ‡ Division of Organic Chemistry. 1 Abbreviations: MR, madder root; LuP, lucidin-3-O-primeveroside; Luc, lucidin; Rub, rubiadin; DNA, DNA; SULT, sulfotransferase; PAPS, 3′phosphoadenosine-5′-phosphosulfate; dG, 2′-deoxyguanosine; dA, 2′-deoxyadenosine; LC; liquid chromatography; PDA, photodiode array; ESI, electron spray ionization; MS, mass spectrometry; NMR, nuclear magnetic resonance; RNase, ribonuclease; DMF, N,N-dimethylformamide; HPLC, high performance liquid chromatography, Acetoxy-Luc, 1,3-diacetoxy-2-acetoxymethylanthraquinone; TMS, tetramethylsilane; NADPH, reduced nicotinamide adnine dinucleotide phosphate; SPE, solid phase extraction; SIR, selected ion recorder; LOQ, limit of quantification; BP, benzo[a]pyrene; 3-NBA, 3-nitrobenzanthrone.

existence of specific DNA repair systems (4). Since the specific activity of repair enzymes is primarily dependent on the stereochemical structures, bulks, and sites of their substrates, confirmation of DNA base modifications is necessary for understanding their biological significance (5, 6). Sulfotransferases (SULTs) are one of the phase II detoxification enzymes which catalyze the transfer of a sulfonyl group from 3′-phosphoadenosine-5′-phosphosulfate (PAPS) to the hydroxyl, sulfhydryl, amino, or N-oxide groups of various substrates (7). As known in some compounds, the dissociation of the sulfate ion following sulfate conjugation of benzylic alcohol results in a reactive carbocation intermediate, and it has been thought that SULT-mediated reactions could generate electrophilic reactants in a similar manner (8, 9). Since Luc possesses benzylic alcohol in its structure, it has been postulated that after sulfate conjugation a reactive carbocation or quinone methide formed in Luc at the benzylic position could react with DNA bases. In this study, as a mimic of the SULT metabolic pathway, we took advantage of an alternative formation of a reactive carbocation generated by acetylation and then acetate dissociation of a benzylic alcohol (10, 11). First, to determine the structures of Luc-specific 2′-deoxyguanosine (dG) and 2′deoxyadenosine (dA) adducts, a reaction mixture of dG and dA with acetylated-Luc was analyzed using liquid chromatography (LC) with photodiode array (PDA) and electron spray ionization (ESI)/mass spectrometry (MS); subsequently, the precise chemical structures of the detected adducts were identified using NMR. Second, a quantitative analytical method for these adducts was developed using LC-ESI/MS, which was applied to confirm adduct formation in the reaction of calf thymus DNA with Luc

10.1021/tx900314c  2010 American Chemical Society Published on Web 12/10/2009

Characterization of Luc-Specific DNA Adducts

Chem. Res. Toxicol., Vol. 23, No. 1, 2010 135

Figure 1. Chemical structure of LuP, Luc, and Rub.

in the presence of S9 mixture. Since Rub, another metabolite of LuP, requires S9 mixture to show genotoxicity in the Ames test (2), we also examined whether adduct formation occurred in calf thymus DNA treated with Rub in the presence of S9 mixture.

Materials and Methods Caution: Lucidin is a mutagen and should be handled accordingly. Chemicals and Reagents. Luc and Rub were synthesized at ALPS Pharmaceutical Ind. Co., Ltd. (Gifu, Japan). The purities of Luc and Rub were 98.5% and 99.6%, respectively. Calf thymus DNA, dG, dA, alkaline phosphatase, PAPS, and ribonuclease (RNase) were purchased from Sigma Chemical Co. (St. Louis, MO), and nuclease P1 was purchased from Yamasa Shoyu Co. (Chiba, Japan). Pooled male rat liver S9 mixture (Sprague-Dawley) was purchased from Oriental Yeast Co., Ltd. (Tokyo, Japan). Stable isotope labeled 15N5-dG and 15N5-dA were obtained from Cambridge Isotope Laboratories (Cambridge, MA). Acetic anhydride, ammonium carbonate, dichloromethane, N,N-dimethylformamide (DMF), 2-propanol, pyridine, and DNA extraction kit were purchased from Wako pure chemicals (Tokyo, Japan). All other chemicals used were specified as analytical or of high performance liquid chromatography (HPLC) grade. Synthesis of 1,3-Diacetoxy-2-acetoxymethylanthraquinone (Acetoxy-Luc). Acetoxy-Luc was synthesized from Luc as described for 1′-acetoxysafrole by Borchert et al. and Drinkwater et al. (11, 12). In brief, Luc (80 mg) was dissolved in 400 µL of pyridine. Acetic anhydride (300 µL) was added dropwise to this solution, and the reaction mixture was stirred for 5 h at room temperature, after which 400 µL of dichloromethane was added. The reaction mixture was extracted several times with 1.0 mL aliquots of 1.0 N HCl. When the aqueous phase reached pH 2-3, the organic layer was immediately extracted with 2.0 mL of 1.0 M sodium carbonate solution (pH 7.6). Acetoxy-Luc: UV λmax 257 and 332 nm. 1H NMR (DMSO-d6): δ 8.13-8.18 (m, 2H, Ar), 8.03 (s, 1H, Ar), 7.90-7.94 (m, 2H, Ar), 5.16 (s, 2H, -CH2OCO-), 2.441 (s, 3H, -OCOCH3), 2.381 (s, 3H, -OCOCH3), 1.963 (s, 3H, -OCOCH3). Determination of a Luc-Specific Nucleoside Adduct. Lucspecific DNA adducts were determined from reactions between Acetoxy-Luc and dG or dA based on the protocol of Phillips et al. (13). The reaction product containing Acetoxy-Luc was diluted 50fold in DMF from which 200 µL was added to 1.8 mL of 10 mM dG or dA solution in 50 mM sodium phosphate buffer (pH 7.4). The reaction mixture was stirred for 24 h at 37 °C. The reaction mixture was then passed through an HLC-DISK syringe filter (Kanto Chemical Co., Inc., Tokyo) and was separated on a LCPDA-ESI/MS system. LC-PDA-ESI/MS analyses were performed using an Alliance HT model 2695 liquid chromatographic system coupled to a 996 photodiode array detector and a Micromass ZQ, a single quadrupole, mass spectrometry system (Waters Corp., Miliford, MA) equipped with an ESI source through a splitter. Twenty microliters of the reaction mixture was injected directly onto a reverse-phase C18 column (Mightysil RP-18 GP, 4.6 × 150 mm, 5 µm, Kanto Chemical Co., Inc.) maintained at 40 °C. Solvent A was water, solvent B was methanol, and solvent C was 0.1% formic acid. The column was equilibrated with a mixture of solvent A/solvent B/solvent C (60/30/10, v/v). A linear gradient was applied from 30% to 90% methanol over 30 min, kept at 90% for 10 min,

lowered to 30% over 2 min, and re-equilibrated at the initial conditions for 15 min. Mass analysis was performed in scan mode. The cone voltages were 20, 40, and 60 V in the positive ion mode. Synthesis of Luc-Derived dG and dA Adduct (Peak 1 and Peak 2). Luc-dG and dA adducts were synthesized from a reaction between Acetoxy-Luc and dG or dA. Acetoxy-Luc was diluted 50fold in DMF from which 200 mL was added to 1800 mL of 10 mM dG or dA solution in 50 mM sodium phosphate buffer (pH 7.4). The reaction was stirred for 24 h at 50 °C. The yields of peak 1 and peak 2 were 15.4 mg (2%) and 4.0 mg (0.54%), respectively. The reaction mixture was repeated several times as needed to acquire enough product for NMR analysis. Purification was performed with the combined reaction mixtures that were evaporated and reconstituted in 50% methanol/water. The concentrated reaction mixture was separated by a liquid chromatography system equipped with a UV detector (LC-UV) (PU-2080 Plus Intelligent HPLC Pump, AS-2057 plus Intelligent Sampler, CO-966 Intelligent Column Thermostat and UV-970 Intelligent UV/ vis Detector; JASCO Co., Tokyo). Two milliliters of sample was injected directly on to a reverse phase C 18 column (Mightysil RP18 GP, 20 × 250 mm, 5 µm, Kanto Chemical Co., Inc.) maintained at 40 °C. Solvent A was 0.01% formic acid, and solvent B was methanol containing 0.01% formic acid. The column was equilibrated with a mixture of solvent A/solvent B (70/30, v/v). A linear gradient was applied from 30% to 90% methanol over 30 min, kept at 90% for 10 min, lowered to 30% over 2 min, and equilibrated at the initial conditions for 15 min. The products eluting at 22.5 and 31.2 min (UV absorbance at 280 nm; flow rate 10 mL/min) were collected. The fractions were dried in vacuo and weighed. The identity of the product was confirmed by LC-ESI/MS and 1H NMR. 1 H NMR spectra were recorded with a Varian 600 MHz NMR system (Varian Inc. Corp., Palo Alto, CA). Chemical shifts are expressed in ppm downfield-shift from TMS (δ scale). Synthesis of stable isotopically labeled surrogate standards was also performed on a small scale by the same method. The reaction mixture was purified using the LC-UV system. Reaction of Calf Thymus DNA with Luc or Rub. Calf thymus DNA (final conc 2.5 mg/mL) was dissolved in 50 mM sodium phosphate buffer, pH 7.4, containing 160 mM KCl, 5 mM MgCl2, and 2.7 mM reduced nicotinamide adenine dinucleotide phosphate (NADPH). After the addition of 500 µL of 1 mM Luc or Rub (DMF solution) and 2 mL of rat liver S9 mixture (5 mg of microsomal proteins) to calf thymus DNA solution, the reaction was started by the addition of 1 mM PAPS. The reaction mixtures were divided into 10 tubes and were incubated for 24 h at 37 °C. The reactions were stopped by placing the reaction tubes on ice. DNA isolation and digestion were performed as described in the following section. DNA Isolation and Enzymatic Digestion. The reaction mixtures were centrifuged at 2,430g for 5 min. The supernatant was removed, and 1.0 mL of 2-propanol was added. Following mixing by inversion, the mixture was centrifuged at 9,730g for 20 min. The supernatant was discarded, and the pellet containing DNA and proteins was washed with 1.0 mL each of 2-propanol and ethanol with centrifugation at 9,730g for 10 min at each washing step. The pellet was air-dried and then dissolved in 200 µL of enzyme reaction buffer included in the DNA extractor WB kit, and was incubated at 50 °C for 15 min after the addition of 20 µL of RNase. Next, 20 µL of protease K was added, and the mixture was incubated at 50 °C for 60 min. The DNA pellet was obtained by washing with 2-propanol and ethanol with intermediate centrifugation.

136

Chem. Res. Toxicol., Vol. 23, No. 1, 2010

Ishii et al.

Figure 2. Typical LC-PDA chromatograms (280 nm) in the reaction of (A) Acetoxy-Luc alone, (B) Acetoxy-Luc with dG, and (C) Acetoxy-Luc with dA. UV spectra of peaks 1 (D) and 2 (E) obtained from LC-PDA analysis for the reactions of Acetoxy-Luc with dG and dA, respectively. LC-PDA conditions are described in Materials and Methods.

The dried DNA pellet was dissolved in surrogate standard containing 20 mM sodium acetate buffer, pH 4.8, and was incubated with 8 µL of nuclease P1 (2000 U/mL) at 70 °C for 30 min. Next, 20 µL of 1.0 M Tris-HCl buffer, pH 8.2, was added, and the mixture was incubated with 8 µL of alkaline phosphatase (2500 U/mL) at 37 °C for 90 min. After the addition of 3.0 M sodium acetate buffer, pH 5.1, the digested DNA samples from the 10 tubes were combined into one sample. For dG and dA analysis, a 500 µL portion of the digested DNA sample was passed through a 100,000 NMWL filter (Millipore, Bedford, MA) and injected into the LCUV. Two milliliters of the digested sample was purified by solid phase extraction (SPE) and then injected into the LC-ESI/MS for adduct analysis. Sample Preparation with SPE. For adduct analysis, the digested DNA sample was subjected to SPE using an Oasis HLB column (3.0 cm3, 30 mg, Waters Corp.) connected to a vacuum manifold maintained at a vacuum of 10 mmHg. The column was initially conditioned with 3.0 mL of methanol followed by 3.0 mL of HPLC grade water. The digested DNA sample (2.0 mL) was then loaded onto the column and washed with 3.0 mL of methanol/HPLC grade

water (5:95, v/v). The DNA adducts were eluted from the column with 9 mL of methanol, concentrated using a centrifugal vacuum evaporator, and redissolved in 100 µL of methanol/HPLC grade water (50:50, v/v). LC-ESI/MS Analysis for Luc Adducts. LC-ESI/MS analyses were performed using a Waters Micromass ZQ, a single quadrupole, mass spectrometer, coupled to a Waters 2695 separations module with an autosampler. The mass spectrometer was operated in positive ionization mode with selected ion recorder (SIR) acquisition. The nebulizer gas was obtained from an in house high purity nitrogen source. The temperature of the source was set at 110 °C, and the voltage of the capillary was 3 kV. The cone voltages were 40 and 50 V for the dG adduct and dA adduct, respectively. The gas flow of desolvation and the cone were set at 400 L/h and 150 L/h, respectively. Chromatographic separation was achieved using a reverse phase C 18 column (Mightysil RP-18 GP, 2.0 × 150 mm, 5 µm; Kanto Chemical Co., Inc.) as the stationary phase. The mobile phase consisted of a mixture of methanol/water/0.1% formic acid with an initial ratio of 20/70/10, followed by a linear gradient to a final ratio of 90:10 v/v over 30 min, and was delivered at a

Characterization of Luc-Specific DNA Adducts

Chem. Res. Toxicol., Vol. 23, No. 1, 2010 137

Figure 3. Mass spectra of peaks 1 and 2 obtained from LC-ESI/MS analysis for the reaction of Acetoxy-Luc with dG (A-C) and dA (D-F). Mass analysis was performed in scan mode. The cone voltages were 20 V (A and D), 40 V (B and E), and 60 V (C and F) in the positive ion mode. LCESI/MS conditions are described in Materials and Methods.

Figure 4. Chemical structures of Luc-N2-dG and Luc-N6-dA adducts and their stable, isotopically labeled compounds. The asterisk (*) indicates nitrogen 15-labeling.

constant flow rate of 0.2 mL/min. The molecular ions at m/z 404, 409, 388, and 393 corresponding to [M + H-glycoside]+ were selected for the quantification of Luc-N2-dG, 15N5-Luc-N2-dG, LucN6-dA, and 15N5-Luc-N6-dA, respectively. Under these conditions, the standard retention times were 15.0 and 18.1 min for Luc-N2dG and Luc-N6-dA, respectively. LC-UV Analysis for dG and dA. dG and dA were determined by a LC-UV system (PU-980 Intelligent HPLC Pump, AS-950-10 Intelligent Sampler, CO-1560 Intelligent Column Thermostat, and MD-1515 Multiwavelength Detector; JASCO, Co.). Two milliliters of sample was injected directly on to a reverse phase C 18 column (ULTRASPHERE ODS, 4.6 × 250 mm, 5 µm; Beckman Coulter, Inc., Fullerton, CA) maintained at 40 °C. Solvent A was 0.01% formic acid, and solvent B was methanol containing 0.01% formic acid. The column was equilibrated with a mixture of solvent A/solvent B (98/2, v/v). The compounds were eluted at a flow rate of 1.0 mL/min using a method consisting of a linear gradient from 2% to 10% methanol over 20 min, kept at 10% for 5 min, and lowered to 2% over 2 min, and equilibration at the initial conditions

for 15 min. The wavelength of the UV detector was set at 280 nm for the detection of dG and dA.

Results Characterization of Luc-Specific DNA Adduct. Poginsky et al. have proposed that the benzylic position of Luc could be activated as an electrophilic intermediate in the process of sulfate conjugation. This hypothesis is based on the fact that benzylic sulfate esters are subject to dissociation leading to the generation of highly reactive benzylic carbocations or quinone methide. However, since it was uncertain whether Luc reacts with DNA bases via such an electrophilic intermediate, this mechanism needed to be examined using a chemically synthesized Luccarbocation. Punt et al. have demonstrated that acetylated benzylic alcohols give rise to electrophilic carbocations through acetate dissociation in aqueous solution (10). We took advantage of this chemical reaction as a mimic of sulfate conjugation to

138

Chem. Res. Toxicol., Vol. 23, No. 1, 2010

Ishii et al.

Table 1. 1H NMR Chemical Shifts of Peak 1 and Peak 2 Peak 1 H-1′ H-2′ H-2′′ H-3′ H-4′ H-5′ H-5′′ OH-3′ OH-5′ N1H N2-H H-2 H-8 N6-H AQb H-6,7 AQ H-5 AQ H-8 AQ H-4 AQ CH2 c

Peak 2

6.25 (1H, m)a 2.70 (1H, m) 2.28 (1H, m) 4.43 (1H, m) 3.87 (1H, m) 3.63 (1H, m) 3.55 (1H, m) 5.34 (1H, bs) 4.98 (1H, bs) 10.58 (1H, bs) 6.70 (1H, bs) 7.95 (1H, s) 7.92 8.16 8.23 7.27 4.57

(2H, (1H, (1H, (1H, (2H,

m)c m) m) s) m)

a m, multiplet; s, singlet; bs, broad singlet. Overlapped signals.

b

6.34 2.70 2.27 4.38 3.87 3.60 3.51 5.30 5.14

(1H, (1H, (1H, (1H, (1H, (1H, (1H, (1H, (1H,

m) m) m) m) m) m) m) s) bs)

8.31 8.37 7.99 7.87 8.10 8.16 7.08 4.64

(1H, (1H, (1H, (2H, (1H, (1H, (1H, (2H,

bs) s) bs) m) m) m) s) bs)

AQ, anthraquinone.

form a Luc carbocation from acetylated Luc. The reaction mixtures of Acetoxy-Luc with/without dG or dA were separated and analyzed by LC-PDA-ESI/MS. Typical LC-PDA chromatograms are shown in Figure 2. Several unknown peaks were observed in the LC chromatogram (280 nm) after the reaction of Acetoxy-Luc with dG and dA (Figure 2B and C). Simultaneous acquisition of mass spectra and UV-vis absorption spectra were obtained by ESI/MS (cone voltage 20, 40, and 60 V) and PDA. MS spectra of all unknown peaks were analyzed on the basis of Luc and base structures, and the two major peaks were selected from the chromatograms. One of the two unknown peaks (Peak 1) was detected at 18.0 min in the chromatogram of the dG reaction (Figure 2B). The UV-vis absorption spectrum of Peak 1 is shown in Figure 2D and had λmax at 280 and 410 nm and a shoulder at about 240 nm. In the mass spectrum of Peak 1, when the cone voltage was set at 20 V (Figure 3A), the precursor ion ([M + H]+) with a mass of m/z 520 and the product ion with m/z 404 corresponding to a Luc-guanine adduct followed by cleavage of the glycosidic bond were clearly observed. Product ions at m/z 253 and 152

Figure 5. SIM chromatograms of Luc-N2-dG, Luc-N6-dA, and surrogate standards in calf thymus DNA incubated with Luc in the presence of S9 mixture with PAPS. Two hundred femtomoles on column for Luc-N2-dG and Luc-N6-dA standard (A and D); Luc-N2-dG and Luc-N6-dA formed in the reaction (B and E); surrogate standards for Luc-N2-dG and Luc- N6-dA (C and F). LC-ESI/MS conditions are described in Materials and Methods.

Characterization of Luc-Specific DNA Adducts

Chem. Res. Toxicol., Vol. 23, No. 1, 2010 139

Figure 6. SIM chromatograms of Luc-N2-dG, Luc-N6-dA, and surrogate standards in calf thymus DNA incubated with Rub in the presence of S9 mixture with PAPS. Luc-N2-dG and Luc-N6-dA formed in the reaction (A and C); surrogate standards for Luc-N2-dG and Luc-N6-dA (B and D). LC-ESI/MS conditions are described in Materials and Methods.

corresponding to the structures of 1,3-dihydroxy-2-methylanthraquinone and guanine were observed in the mass spectra at 40 and 60 V (Figure 3B and C). The second peak (Peak 2) was detected at 23.4 min in the chromatogram for the reaction of Acetoxy-Luc and dA (Figure 2C). The UV-vis absorption spectrum of Peak 2 is shown in Figure 2E and was virtually identical to Peak 1 with λmax at 280 nm and 410 nm and a shoulder at about 240 nm. In the mass spectra of Peak 2 when the cone voltage was set at 20 and 40 V (Figure 3D and E), the precursor ion ([M + H]+) with a mass of m/z 504 and the product ion with m/z 388 corresponding to a Luc-adenine adduct followed by cleavage of the glycosidic bond were clearly observed. Product ions at m/z 253 and 136 corresponding to the structures of 1, 3-dihydroxy-2-methylanthraquinone and adenine were also observed in the mass spectra at 40 and 60 V (Figure 3E and F). Acetoxy-Luc was used as an electrophilic synthon in a large scale synthesis to identify the chemical structures of peaks 1 and 2 from dG and dA, respectively. Adducts were synthesized in several reactions with each base until a sufficient amount of materials for 1H NMR analysis was acquired. Adducts were purified for 1H NMR analysis by HPLC-UV. 1H NMR chemical shifts in DMSO-d6 for the compounds corresponding to peaks 1 and 2 are shown in Table 1. The presence of the C-8 proton at 7.95 and 8.37 ppm in Peak 1 and Peak 2 and the fact that these protons did not exchange with D2O appeared to exclude C-8 substitution in deoxyguanosine and deoxyadenosine, respectively. Additionally, the signal at 6.70 ppm and 7.99 ppm in Peak 1 and Peak 2 which integrated for 1 proton and was exchangeable with D2O and was assigned to the proton of the substituted-NH group attached to C-2 of guanine and C-6 of adenine, respectively. Then, we determined that the modified

bases were Luc-N2-dG and Luc-N6-dA, respectively. The chemical structures of Luc-N2-dG (A), Luc-N6-dA (B), and stable, isotopically labeled compounds as surrogate standards for LC-ESI/MS analysis are shown in Figure 4. Performance of the LC-ESI/MS Method for Adduct Analysis. On the basis of the fragmentation patterns of LucN2-dG, Luc-N6-dA, and each surrogate standard, an isotope dilution LC-ESI/MS method was developed using SIR to ensure a high sensitivity of detection. The standard retention times were 15 min for Luc-N2-dG and 18.1 min for Luc-N6-dA (Figure 5A and D). The calculated limits of detection of Luc-N2-dG and Luc-N6-dA of the standard solutions were 10 fmol and 6.0 fmol on column, respectively, for LC-ESI/MS detection at a compound signal-to-noise ratio (S/N) of 3. In addition, the limits of quantification (LOQ) calculated with a S/N ) 10 were 40 fmol and 20 fmol on column for Luc-N2-dG and Luc-N6-dA, respectively. For quantitative analyses of Luc-N2-dG and LucN6-dA, calibration curves were constructed in the presence of a 400 fmol internal standard. The peak ratio with respect to each surrogate standard was plotted using five different concentrations of Luc-N2-dG and Luc-N6-dA, and the response was found to be linear over the calibration range with a correlation coefficient (r) of 0.999 (2-500 nM). The average retention times for LucN2-dG and Luc-N6-dA were 15.0 min (RSD ) 0.06%, n ) 5) and 18.1 min (RSD ) 0.02%, n ) 5), respectively, and the accuracies of the RSDs of the peak area were 1.44% and 0.65%, respectively. Determination of Luc-Specific DNA Adducts in Calf Thymus DNA. Luc-specific DNA adducts in calf thymus DNA were determined by our newly developed method using LCESI/MS. Typical SIR chromatograms of calf thymus DNA are shown in Figure 5. Luc-N2-dG and Luc-N6-dA adducts and the

140

Chem. Res. Toxicol., Vol. 23, No. 1, 2010

Ishii et al.

Figure 7. Proposed pathways for the metabolic activation of Luc and Rub.

surrogate standards were detected at 15.0 and 18.1 min in the chromatograms at m/z 404, 388, 409, and 393, respectively, from the calf thymus DNA incubated with Luc in the presence of S9 mixture with PAPS at 37 °C after mixing for 12 h. The LucN2-dG/106dG and Luc-N6-dA/106dA ratios were 235.1 ( 27.5 and 160.1 ( 14.5, respectively. Furthermore, the reaction of DNA and Rub in the presence of S9 mixture with PAPS also formed Luc-specific DNA adducts (Figure 6). The Luc-N2-dG/ 106dG and Luc-N6-dA/106dA ratios were 7.2 ( 2.7 and 5.2 ( 3.0, respectively.

Discussion Poginsky et al. have already demonstrated the formation of LuP and Luc-related DNA adducts in several organs of mice using 32P-postlabeling analysis (3). In addition, they have proposed that the benzylic position of Luc could be activated as an electrophilic intermediate in the process of sulfate conjugation. In the present study, we took advantage of an alternative formation of a reactive carbocation generated by acetylation. In the reaction of dG or dA with Acetoxy-Luc, two peaks including characteristic ions for Luc-specific DNA adducts were found by LC-ESI/MS analysis. Subsequently, 1H NMR analysis of chemically synthesized and HPLC purified samples led to the conclusion that these two adducts were Luc-N2-dG and Luc-N6-dA, respectively.

To confirm whether Luc-N2-dG and Luc-N6-dA were actually formed in DNA under physiological conditions, we attempted to detect these adducts in the reaction of Luc with calf thymus DNA in the presence of both PAPS and S9 mixture. Since appropriate sample preparations and highly sensitive analytical methods are necessary to detect modified DNA bases in genomic DNA, we developed a new analytical procedure which combines a sample purification method using SPE and an isotope dilution LC-ESI/MS method using SIR. As a result, both Luc-N2-dG and Luc-N6-dA were detected to an appreciable extent, implying that Luc undergoes sulfate conjugation followed by covalent modification of dG or dA via a carbocation and quinone methide intermediate under physiological conditions. In addition, the LOQs of Luc-N2-dG and Luc-N6-dA were determined to be 2-3 adducts/107 unmodified dG or dA bases from a 2.5 mg DNA sample. Judging from these figures, our new method might have applicability to genomic DNA extracted from in vivo samples and could be useful for further research on Luc-induced DNA modifications using various experimental animal models. The N-2 and N-6 positions of dG or dA are known to be sites susceptible to reaction with electrophiles as has been found with other potent mutagens such as benzo[a]pyrene (BP) and 3-nitrobenzanthrone (3-NBA) (14). It is believed that N2-dG and N6-dA adducts formed by BP or 3-NBA are responsible for their

Characterization of Luc-Specific DNA Adducts

carcinogenicity (15-20). The structural features of DNA adducts are believed to determine their mutational properties since the functions of specific DNA polymerases and repair enzymes are dependent on the specific chemical structure (21). In fact, oligonucleotide synthesis with the modified DNA bases is a prerequisite for investigating their biological significance (22). In this respect, information on the precise chemical structures of Luc-specific DNA adducts clarified in this study could be very significant for further research in this area. As Rub is another genotoxic metabolite of LuP, we have proposed that it may also take part in MR carcinogenicity together with Luc. In mutagenicity tests using Salmonella typhimurium, in contrast to Luc being positive independent of the presence of S9 mixture, Rub required S9 mixture to reveal mutagenicity in both strains of TA98 and TA100 (23). As mentioned above, while Luc undergoes metabolism by SULT to generate its active form, the pathway in which the sulfate ion is eliminated from Luc resulting in the production of an electrophilic intermediate has been proposed (3). Considering that the Rub structure differs from Luc by having a methyl group (-CH3) at the C-2 position in place of a hydroxymethyl group (-CH2OH), it is highly probable that Rub is metabolized to Luc through hydroxylation of its methyl group, which consequently would lead to Rub genotoxicity (Figure 7). To confirm this hypothesis, Luc-specific DNA adduct formation in calf thymus DNA treated with Rub in the presence of PAPS and S9 mixture was examined. Luc-N2-dG and Luc-N6-dA adducts were detected from this reaction with Rub by the LC-ESI/MS method in spite of the low level of adducts as compared to the same reaction conditions with Luc. We were also able to confirm the existence of Luc in the Rub solution incubated with S9 mixture by LC-UV (data not shown). In conclusion, the major DNA bases chemically modified by Luc were determined to be Luc-N2-dG and Luc-N6-dA adducts. These adducts were also formed in reactions with calf thymus DNA possibly after SULT conjugation. In addition, the present data suggest that LuP and Rub exert their genotoxicity by metabolic conversion to the proximate metabolite Luc and ultimate carcinogenic sulfooxy-Luc. The new analytical method developed to achieve these data may be a powerful tool in exploring the mechanisms underlying the carcinogenicity of MR and related anthraquinones. Supporting Information Available: 1H NMR spectra of LucN -dG and Luc-N6-dA. This material is available free of charge via the Internet at http://pubs.acs.org. 2

References (1) Inoue, K., Yoshida, M., Takahashi, M., Shibutani, M., Takagi, H., Hirose, M., and Nishikawa, A. (2009) Induction of kidney and liver cancers by the natural food additive madder color in a two-year rat carcinogenicity study. Food Chem. Toxicol. 47, 184–191. (2) Kawasaki, Y., Goda, Y., and Yoshihira, K. (1992) The mutagenic constituents of Rubia tinctorum. Chem. Pharm. Bull. 40, 1504–1509. (3) Poginsky, B., Westendorf, J., Blomeke, B., Marquardt, H., Hewer, A., Grover, P. L., and Phillips, D. H. (1991) Evaluation of DNAbinding activity of hydroxyanthraquinones occurring in Rubia tinctorum L. Carcinogenesis 12, 1265–1271. (4) Moustacchi, E. (2000) DNA damage and repair: consequences on doseresponses. Mutat. Res. 464, 35–40. (5) Choi, J. Y., and Guengerich, F. P. (2005) Adduct size limits efficient and error-free bypass across bulky N2-guanine DNA lesions by human DNA polymerase eta. J. Mol. Biol. 352, 72–90.

Chem. Res. Toxicol., Vol. 23, No. 1, 2010 141 (6) Wei, D., Maher, V. M., McCormick, J. J. (1996) Site-specific excision repair of 1-nitrosopyrene-induced DNA adducts at the nucleotide level in the HPRT gene of human fibroblasts: effect of adduct conformation on the pattern of site-specific repair. 16, 3714-3719. (7) Falany, C. N. (1991) Molecular enzymology of human liver cytosolic sulfotransferases. Trends Pharmacol. Sci. 12, 255–259. (8) Lehner, A. F., Horn, J., and Flesher, J. W. (2004) Mass spectrometric analysis of 7-sulfoxymethyl-12-methylbenz[a]anthracene and related electrophilic polycyclic aromatic hydrocarbon metabolites. J. Mass. Spectrom. 39, 1366–1378. (9) Shibutani, S., Dasaradhi, L., Terashima, I., Banoglu, E., and Duffel, M. W. (1998) Alpha-hydroxytamoxifen is a substrate of hydroxysteroid (alcohol) sulfotransferase, resulting in tamoxifen DNA adducts. Cancer Res. 58, 647–653. (10) Punt, A., Delatour, T., Scholz, G., Schilter, B., van Bladeren, P. J., and Rietjens, I. M. (2007) Tandem mass spectrometry analysis of N2(trans-isoestragol-3′-yl)-2′-deoxyguanosine as a strategy to study species differences in sulfotransferase conversion of the proximate carcinogen 1′-hydroxyestragole. Chem. Res. Toxicol. 20, 991–998. (11) Borchert, P., Wislocki, P. G., Miller, J. A., and Miller, E. C. (1973) The metabolism of the naturally occurring hepatocarcinogen safrole to 1′-hydroxysafrole and the electrophilic reactivity of 1′-acetoxysafrole. Cancer Res. 33, 575–589. (12) Drinkwater, N. R., Miller, E. C., Miller, J. A., and Pitot, H. C. (1976) Hepatocarcinogenicity of estragole (1-allyl-4-methoxybenzene) and 1′hydroxyestragole in the mouse and mutagenicity of 1′-acetoxyestragole in bacteria. J. Natl. Cancer Inst. 57, 1323–1331. (13) Phillips, D. H., Reddy, M. V., and Randerath, K. (1984) 32Ppostlabelling analysis of DNA adducts formed by electrophilic esters of the hepatocarcinogens 1′-hydroxysafrole and 1′-hydroxyestragole in vitro and in mouse liver in vivo, including new adducts at C-8 and N-7 of guanine residues. Cancer Res. 45, 3096–3105. (14) Kim, H. Y., Cooper, M., Nechev, L. V., Harris, C. M., and Harris, T. M. (2001) Synthesis and characterization of nucleosides and oligonecleotides with benzo[a]pyren-6ylmethyl adduct at adenine N6 or guanine N2. Chem. Res. Toxicol. 14, 1306–1314. (15) Wang, J. J., Marchall, W. D., Frazer, D. G., Law, B., and Lewis, D. M. (2003) Characterization of DNA adducts from lung tissue of asphalt fume-exposed mice by nanoflow liquid chromatography quadrupole time-of -flight mass spectrometry. Anal. Biochem. 322, 79–88. (16) Beland, F. A., Churchwell, M. I., Von Tungeln, L. S., Chen, S., Fu, P. P., Culp, S. J., Schoket, B., Gyorffy, E., Minarovits, J., Poirier, M. C., Bowman, E. D., Weston, A., and Doerge, D. R. (2005) Highperformance liquid chromatography electrospray ionization mass spectrometry for the detection and quantitation of benzo[a]pyreneDNA adducts. Chem. Res. Toxicol. 18, 1306–1315. (17) Rodriguez, H., and Loechler, E. L. (1993) Mutational spectra of the (+)-anti-diol epoxide of benzo[a]pyrene in supF gene of an Escherichia coli plasmid: DNA sequence context influences hotspots, mutational specificity and the extent of SOS enhancement of mutagenesis. Carcinogenesis 14, 373–383. (18) Rodriguez, H., and Loechler, E. L. (1993) Mutagenesis by the (+)anti-diol epoxide of benzo[a]pyrene: what controls mutagenic specificity? Biochemistry 32, 373–383. (19) Seo, K. Y., Nagalingam, A., Tiffany, M., and Loechler, E. L. (2005) Mutagenesis strudies with four stereoisomeric N2-dG benzo[a]pyrene adducts in the identical 5′-CGC sequence used in NMR studies: G f T mutations dominate in each case. Mutagenesis 20, 441–448. (20) Zhao, B., Wang, J., Geacintow, N. E., and Wang, Z. (2006) Poleta, Polzeta and Rev1 together are required for G to T transversion mutations induced by the (+)-and (-)-trans-anti-BPDE-N2-dG DNA adducts in yeast cells. Chem. Res. Toxicol. 34, 417–425. (21) Broyde, S., Wang, L., Zhang, L., Rechkoblit, O., Geacintov, N. E., and Patel, D. J. (2008) DNA adduct structure-function relationships: comparing solution with polymerase structures. Chem. Res. Toxicol. 21, 45–52. (22) Shibutani, S., Fernandes, A., Suzuki, N., Zhou, L., Johnson, F., and Grollman, A. P. (1999) Mutagenesis of the N-(deoxyguanosine-8yl)2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine DNA adduct in mammalian cells. Sequence context effects. J. Biol. Chem. 274, 27433– 27438. (23) Blomeki, B., Poginsky, B., Schmutte, C., Marquardt, H., and Westendorf, J. (1992) Formation of genotoxic metabolites from anthraquinone glycosides, present in Rubia tinctorum L. Mutat. Res. 265, 263–272.

TX900314C