pyrene (Dibenzo[def,p]chrysene) to Dibenzo[a,l ... - ACS Publications

Metabolic activation studies of dibenzo[a,l]pyrene (DB[a,l]P) (dibenzo[def,p]chrysene), an extremely potent environmental carcinogen, have been focuse...
0 downloads 0 Views 147KB Size
1596

Chem. Res. Toxicol. 1998, 11, 1596-1607

Metabolic Activation of Racemic and Enantiomeric trans-8,9-Dihydroxy-8,9-dihydrodibenzo[a,l]pyrene (Dibenzo[def,p]chrysene) to Dibenzo[a,l]pyrene-bis-dihydrodiols by Induced Rat Liver Microsomes and a Recombinant Human P450 1A1 System: The Role of the K-Region-Derived Metabolic Intermediates in the Formation of Dibenzo[a,l]pyrene-DNA Adducts† Stephen Nesnow,*,‡ Christine Davis,‡ William Padgett,‡ Michael George,‡ Guy Lambert,‡ Fabienne Meyers,§,| Joycelyn Allison,‡ Linda Adams,‡ and Leon C. King‡ Biochemistry and Pathobiology Branch, National Health and Environmental Effects Research Laboratory, U.S. Environmental Protection Agency, MD-68, Research Triangle Park, North Carolina 27711, and Curriculum in Toxicology, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27514 Received June 30, 1998

Metabolic activation studies of dibenzo[a,l]pyrene (DB[a,l]P) (dibenzo[def,p]chrysene), an extremely potent environmental carcinogen, have been focused on metabolism at the fjord region, a region associated with high mutagenic and carcinogenic activities of the corresponding fjord-region DB[a,l]P-11,12-diol-13,14-epoxides. DB[a,l]P is metabolized by β-naphthoflavone (BNF)- and 3-methylcholanthrene-induced rat liver microsomes and a recombinant human P450 1A1 system to two major dihydrodiols, the K-region dihydrodiol, DB[a,l]P-8,9-dihydrodiol (DB[a,l]P-8,9-diol), and the fjord-region dihydrodiol, DB[a,l]P-11,12-dihydrodiol. We have investigated the further metabolic activation of DB[a,l]P-8,9-diol by BNF-induced rat liver microsomes and a recombinant human P450 1A1 system with epoxide hydrolase to DB[a,l]Pbis-diols and to DNA adducts. (()-trans-DB[a,l]P-8,9-diol was synthesized and resolved into its enantiomers. Racemic trans-DB[a,l]P-8,9-diol was metabolized by BNF-induced rat liver microsomes to six metabolites: two diastereomers of trans,trans-DB[a,l]P-8,9:11,12-bis-diol, two diastereomers of trans,cis-DB[a,l]P-8,9:11,12-bis-diol, and two diastereomers of trans-DB[a,l]P-8,9:13,14-bis-diol as characterized by NMR, MS, and UV spectroscopy. Metabolic studies using both enantiomeric (-)- and (+)-trans-DB[a,l]P-8,9-diol further demonstrated that each diastereomer of trans,trans-DB[a,l]P-8,9:11,12-bis-diol and trans-DB[a,l]P-8,9:13,14-bis-diol was comprised of two enantiomers. Similarly, incubations of enantiomeric or racemic trans-DB[a,l]P-8,9-diol with a recombinant human P450 1A1 system and epoxide hydrolase also gave the same two enantiomeric mixtures of diastereomers of trans,trans-DB[a,l]P-8,9:11,12-bisdiol and the same two enantiomeric mixtures of diastereomers of trans-DB[a,l]P-8,9:13,14bis-diol. This suggested that the microsomal oxidations of (-)- and (+)-trans-DB[a,l]P-8,9-diol were stereospecific. The stereospecific formation of enantiomers of trans-DB[a,l]P-8,9-diol from DB[a,l]P was examined using both BNF-induced rat liver microsomes and a recombinant human P450 1A1 system with epoxide hydrolase. Stereospecificity was observed as both metabolic systems favored the formation of (-)-trans-DB[a,l]P-8,9-diol by 8-9-fold. DNA adduct studies were undertaken using TLC/HPLC 32P-postlabeling techniques. In the presence of a recombinant human P450 1A1 system with epoxide hydrolase, DB[a,l]P gave two groups of calf thymus DNA adducts. The group of later-eluting adducts were identified as arising from synand anti-DB[a,l]P-11,12-diol-13,14-epoxides, while the more polar early-eluting adducts were derived, in part, from the further activation of trans-DB[a,l]P-8,9-diol. Our data indicate that, in P450 1A1-mediated microsomal incubations, DB[a,l]P is metabolized to trans-DB[a,l]P-8,9diol which is further metabolized to DB[a,l]P-bis-diols. trans-DB[a,l]P-8,9-diol is metabolically activated to intermediates that can bind to DNA and give DNA adducts similar to those observed with DB[a,l]P. These results indicate that DB[a,l]P can be metabolically activated by both fjord-region and K-region pathways.

Introduction 1

Dibenzo[a,l]pyrene (DB[a,l]P) (dibenzo[def,p]chrysene) is a product of the incomplete combustion of fossil fuels

(1) and is found in the environment in cigarette smoke condensate (2), in extracts of emissions of indoor air particles from smoky coal combustion (3), and in extracts

10.1021/tx9801561 CCC: $15.00 © 1998 American Chemical Society Published on Web 11/20/1998

DB[a,l]P-bis-diols and DNA Adducts

of some environmental samples (4). DB[a,l]P induces mouse skin tumors when dermally applied (5, 6). Mouse skin tumor initiation promotion assays have shown DB[a,l]P to be a potent tumor initiator inducing papillomas at doses where B[a]P is inactive (6). DB[a,l]P induced mammary adenocarcinomas in rats and was considerably more active than 7,12-dimethylbenz[a]anthracene (5). Structurally, DB[a,l]P contains a fjord region, a bay region, and a K region. All of these regions have been associated with the metabolic activation of PAHs (7). DB[a,l]P is metabolized to trans-8,9-dihydroxy-8,9-dihydroDB[a,l]P (trans-DB[a,l]P-8,9-diol), trans-11,12-dihydroxy11,12-dihydro-DB[a,l]P (trans-DB[a,l]P-11,12-diol),DB[a,l]P-7-phenol, and a DB[a,l]P-quinone by β-naphthoflavone (BNF)- and 3-methylcholanthrene-induced rat liver microsomes with the two dihydrodiols and the phenol as principal metabolites (8, 9). A study of specific recombinant human P450 isoforms indicated that human P450 1A1 was the most active in the metabolism of DB[a,l]P and also possessed the highest enzymatic activity (10). Recombinant human P450 1A1 and human lung and human liver microsomes metabolized DB[a,l]P to the same metabolites as reported for rat liver microsomes with the addition of the distal bay-region dihydrodiol, 13,14-dihydroxy-13,14-dihydro-DB[a,l]P (10). Studies with Salmonella typhimurium strain TA98 using Aroclor-1254-induced rat liver S9 indicated that the syn- and anti-DB[a,l]P-11,12-diol-13,14-epoxide (DB[a,l]PDE) derived from DB[a,l]P-11,12-diol were the most active metabolites examined in these genetic toxicity systems (11). DNA adduct studies in human mammary MCF-7 cells (12), mouse skin (13), and calf thymus DNA (14, 15) suggested that the fjord region was an important site of activation, yielding DNA adducts from the antiand syn-diol-epoxides of DB[a,l]P with a predominance of 2′-deoxyadenosine adducts. In support of these results, high tumorigenic activity of anti-DB[a,l]PDE in newborn mice was observed (16). The K-region dihydrodiol of DB[a,l]P, trans-DB[a,l]P8,9-diol (Scheme 1), is a predominant rat liver metabolite of this PAH, and has been found to be a weak mouse skin tumorigen (5). In a preliminary account, we reported that a recombinant human P450 1A1 system [originally cloned from MCF-7 breast carcinoma cells induced with 2,3,7,8tetrachlorodibenzo-p-dioxin (17)] bioactivated DB[a,l]P to metabolites capable of forming highly polar calf thymus DNA adducts. These DNA adducts did not comigrate with † The research described in this article has been reviewed by the National Health and Environmental Effects Research Laboratory, U.S. Environmental Protection Agency, and approved for publication. Approval does not signify that the contents necessarily reflect the views of the agency, nor does the mention of trade names or commercial products constitute endorsement or recommendation for use. * To whom requests for reprints should be addressed. ‡ U.S. Environmental Protection Agency. § University of North Carolina at Chapel Hill. | F.M. was supported by a postdoctoral fellowship in the UNC Curriculum in Toxicology funded by the UNC/EPA Toxicology Research Program (Training Agreement CT902908). 1 Abbreviations: BNF, β-naphthoflavone; DB[a,l]P, dibenzo[a,l]pyrene; trans-DB[a,l]P-8,9-diol, trans-8,9-dihydroxy-8,9-dihydro-DB[a,l]P; cis-DB[a,l]P-8,9-diol, cis-8,9-dihydroxy-8,9-dihydro-DB[a,l]P; trans-DB[a,l]P-11,12-diol, trans-11,12-dihydroxy-11,12-dihydro-DB[a,l]P; trans,trans-DB[a,l]P-8,9:11,12-bis-diol, trans,trans-8,9:11,12tetrahydroxy-8,9,11,12-tetrahydro-DB[a,l]P; trans-DB[a,l]P-8,9:13,14bis-diol, trans-8,9:13,14-tetrahydroxy-8,9,13,14-tetrahydro-DB[a,l]P; anti-DB[a,l]PDE, anti-DB[a,l]P-11,12-diol-13,14-epoxide; syn-DB[a,l]PDE, syn-DB[a,l]P-11,12-diol-13,14-epoxide; DDQ, 2,5-dichloro-3,6dicyano-1,4-benzoquinone; PAH, polycyclic aromatic hydrocarbon; RRT, relative retention time.

Chem. Res. Toxicol., Vol. 11, No. 12, 1998 1597 Scheme 1. Metabolism of Dibenzo[a,l]pyrene to Bis-Dihydrodiols through the K-Region Dihydrodiol, DB[a,l]P-8,9-dihydrodiol, by Cytochrome P450 (P450) 1A1 and Epoxide Hydrolasea

a

The absolute stereochemistry is not implied.

standards prepared from reaction products of (()-antiand syn-DB[a,l]PDE with 3′-dAMP or 3′-dGMP (9). Furthermore, highly polar unidentified DB[a,l]P-DNA adducts, not related to (()-anti- and syn-DB[a,l]PDEs, have recently been reported: (i) from incubations of Aroclor-1254-induced rat liver microsomes and calf thymus DNA with DB[a,l]P, (ii) in female rat liver DNA after intermammillary injection of DB[a,l]P, and (iii) in DNA from rainbow trout liver slices exposed to DB[a,l]P (15, 18, 19). We also reported that trans-DB[a,l]P-8,9-diol was metabolized by BNF-induced rat liver microsomes (containing P450 1A1) to three bis-diols: two stereoisomers of trans,trans-8,9:11,12-tetrahydroxy-8,9,11,12-tetrahydro-DB[a,l]P (trans,trans-DB[a,l]P-8,9:11,12-bis-diol) and one stereoisomer of trans-8,9:13,14-tetrahydroxy-8,9,13,14-tetrahydro-DB[a,l]P (trans-DB[a,l]P-8,9:13,14-bis-diol) (9). All of these observations implied that a second metabolic activation pathway for DB[a,l]P (not involving strict fjord-region activation) was operable as suggested by Chakavarti et al. (20). Pursuing this avenue of investigation, we now report on the further identification of six BNF-induced rat liver bis-diol metabolites derived from racemic trans-DB[a,l]P8,9-diol: two diastereomers of trans,trans-DB[a,l]P-8,9: 11,12-bis-diol, two diastereomers of trans,cis-DB[a,l]P8,9:11,12-bis-diol, and two diastereomers of trans-DB[a,l]P8,9:13,14-bis-diol. Metabolic studies using both enantiomeric (-)- and (+)-trans-DB[a,l]P-8,9-diol further demonstrated that each diastereomer of trans,trans-DB[a,l]P8,9:11,12-bis-diol and trans-DB[a,l]P-8,9:13,14-bis-diol was comprised of two enantiomers. A recombinant human P450 1A1 system also gave many of the same metabolites. Moreover, we report on the enantiomeric purities of the trans-DB[a,l]P-8,9-diols metabolically formed from DB[a,l]P by BNF-induced rat liver mi-

1598 Chem. Res. Toxicol., Vol. 11, No. 12, 1998

crosomes and a recombinant human P450 1A1 system. Finally, we report on the metabolic activation by a recombinant human P450 1A1 system of (()-trans-DB[a,l]P-8,9-diol to metabolites that react with calf thymus DNA to form adducts that coeleute with DB[a,l]P-formed DNA adducts.

Experimental Procedures Caution: DB[a,l]P and DB[a,l]P-8,9-diol are tumorigenic and must be handled with care using the guidelines for carcinogenic chemicals developed by the National Cancer Institute. Osmium tetroxide is extremely toxic and should be handled with care. Chemicals. DB[a,l]P (99%) was obtained from Chemsyn (Lenexa, KA). NADP, NADPH (grade III), glucose 6-phosphate, glucose-6-phosphate dehydrogenase, micrococcal nuclease, diethyl ether, 3′-dGMP, 3′-dAMP, and sodium citrate were obtained from Sigma Chemical Co. (St. Louis, MO). Recombinant human P450 1A1 (CYP 1A1) microsomes (catalog no. M111r) were commercially obtained from Gentest (Woburn, MA). The source of these microsomes was human AHH TK-1+/lymphoblast cells expressing transfected human P450 1A1 cDNA. These cells also coexpress human P450 reductase. The P450 1A1 gene had been previously cloned from MCF-7 human breast carcinoma cells induced by 2,3,7,8-tetrachlorodibenzo-pdioxin (17). For large scale metabolism studies, recombinant human P450 1A1 Supersomes (catalog no. P211) containing the same P450 1A1 were used. Microsomal epoxide hydrolase (catalog no. M108a) was purchased from Gentest. HPLC/ spectrophotometric grade acetonitrile, acetone, ethyl acetate, hexane, tetrahydrofuran, 2-propanol, water, and methanol were obtained from Fisher Scientific (Pittsburgh, PA). Polyethyleneimine cellulose thin-layer chromatography plates were purchased from Alltech (Deerfield, IL). T4 polynucleotide kinase (3′phosphatase free) was purchased from Boehringer Mannheim Biochemicals (Indianapolis, IN). Calf thymus DNA, calf spleen phosphodiesterase, and nuclease P1 were purchased from Calbiochem (LaJolla, CA), and [γ-32P]ATP was purchased from Amersham (Arlington Heights, IL). Flo-Scint II was obtained from Radiomatic Instruments & Chemical Co. (Meriden, CT). All other chemicals were reagent grades and obtained from commercial sources. Instrumentation. Melting points were taken on a ThomasHoover capillary apparatus and are uncorrected. UV spectra were recorded on a Beckman model DU-70 spectrometer. 1H NMR spectra were recorded on a Bruker AVANCE-300 instrument at 300.13 MHz. Chemical shifts are reported as parts per million referenced to tetramethylsilane. 2D-COSY, delayed 2DCOSY, homonuclear decoupling, and 1D-NOE difference spectra were recorded according to established manufacturers’ protocols. Low-resolution mass spectra were obtained on an Extrel ELQ400 triple-quadrapole mass spectrometer with LC and GC interfaces. Perfluorotributylamine was used as the mass calibration standard. Exact mass determinations were recorded on a VG70-250SEQ hybrid mass spectrometer at 10 000 resolution, using a solids probe. Mass calibration was performed with perfluorokerosene. CD spectra were recorded on a JASCO model 600 spectrometer using a 1 mm path length quartz cell. Analytical and preparative TLC were carried out on fluorescent silica gel plates or C18 reverse phase plates. Bands were visualized with short- and long-wavelength UV lamps. HPLC was conducted on a Shimazdu model 10A liquid chromatograph at 254 nm or on a Hewlett-Packard model 1050 system connected to a Hewlett-Packard model 1050 diode array detector. Effluents were monitored over the range of 220-550 nm. Synthesis of (()-trans-DB[a,l]P-8,9-diol. The synthesis of (()-trans-DB[a,l]P-8,9-diol was performed according to the general methods for the synthesis of K-region dihydrodiols described by Harvey (21). DB[a,l]P (160 mg, 0.53 mmol) and OsO4 (140 mg, 0.55 mmol) were dissolved in pyridine, and the mixture was stirred at 25 °C for 7 days under a nitrogen atmosphere. TLC on silica (methylene chloride) revealed that

Nesnow et al. no substrate remained. The reaction mixture was treated for 3 h with 450 mg of sodium bisulfite previously dissolved in 0.5 mL of methanol and 1 mL of water. The mixture was poured over ice/water and extracted three times with 50 mL of acetone and 100 mL of ethyl acetate. The organic extracts were dried over magnesium sulfate, filtered, and evaporated in vacuo to give cis-8,9-dihydroxy-8,9-dihydro-DB[a,l]P (cis-DB[a,l]P-8,9diol). The crude cis-DB[a,l]P-diol was treated with 2,5-dichloro3,6-dicyano-1,4-benzoquinone (DDQ, 240 mg, 1.06 mmol) in refluxing dioxane for 15 h. The solvent was reduced to 1 mL and cooled; 20 mL of dichloromethane was added, and the DDQhydroquinone was removed by filtration. The filtrate was washed with saturated sodium bisulfite (100 mL) and water (three times) and evaporated to dryness. The residue was triturated with hexane (20 mL, three times) to give DB[a,l]P8,9-dione as a red solid: mp >250 °C; low-resolution MS m/z (relative intensity) 332 [M]+ (15.6), 304 [M - CO]+ (100). The DB[a,l]P-8,9-dione was suspended in 100 mL of 2-propanol and the mixture stirred with 100 mg of sodium borohydride with an air bubbler for 24 h. The solvent was removed in vacuo, and after addition of 200 mL of ice/water, the aqueous solution was extracted with acetone (50 mL) and ethyl acetate (100 mL). The organic extracts were evaporated, and the residue was sonicated with 0.5 mL of acetone. Hexane (5 mL) was added to the acetone solution to precipitate the diol and the solvent decanted to leave a dry solid. This procedure was repeated three times to give 64 mg (36% yield) of (()-trans-DB[a,l]P-8,9-diol which was recrystallized from acetone/hexane as a colorless solid: mp 209-211 °C; high-resolution MS calcd for C24H16O2 336.1150, found 336.1171; low-resolution MS m/z (relative intensity) 336 [M]+ (60), 318 [M - H2O]+ (100), 289 [M - 47]+ (54). The UV, mass, and NMR spectra were consistent with the structures of cisDB[a,l]P-8,9-diol, DB[a,l]P-8,9-dione, and trans-DB[a,l]P-8,9diol. The potential presence of contaminating cis-DB[a,l]P-8,9diol in the final product was excluded by both NMR and HPLC analyses. In the NMR spectra, protons H8 and H9 exhibit different chemical shifts and coupling constants in each isomer: cis-DB[a,l]P-8,9-diol, H8, 5.36 (d), H9, 5.39 (d), J8,9 ) 2.0 Hz; trans-DB[a,l]P-8,9-diol, H8, 4.86 (d), H9, 4.98 (d), J8,9 ) 10.5 Hz. HPLC of trans- and cis-8,9-diacetoxy-8,9-dihydro-DB[a,l]P (obtained from reaction of the dihydrodiols with acetic anhydride in pyridine) using a 4.6 mm × 250 mm Synchrom Synchropak RP-P column (300 Å pore size) (Alltech) and an isocratic solvent system of 100% acetonitrile at a flow rate of 1 mL/min gave baseline separation (retention times): trans-DB[a,l]P-8,9-diol, 11.5 min; and cis-DB[a,l]P-8,9-diol, 13.8 min. No cis-diol was detected in the final product by either method. All procedures were performed under subdued light. Chiral Separation. Racemic (()-trans-DB[a,l]P-8,9-diol was dissolved in tetrahydrofuran and was resolved by HPLC chiral stationary phase chromatography on a covalently bonded (R,R)Whelk-O1 5 µm chiral stationary phase column (4.6 mm × 250 mm) (Regis Technologies, Morton Grove, IL) using 90% hexane and 10% ethanol/methanol (1:1) at 2 mL/min. The early- and later-eluting enantiomers had retention times of 14.9 and 16.4 min, respectively. Preparative amounts were collected by repetitive chromatographic procedures. Metabolism Studies. Male CD-1 rats (60-80 g) were treated ip (0.5 mL) with BNF, 20 mg/kg/day in corn oil, for 4 days and were sacrificed for 24 h after the last injection. The animals were fasted 24 h prior to sacrifice. Liver microsomes were prepared as previously described (22). The protein content was determined by the method of Lowry using bovine serum albumin as the protein standard. Incubation mixtures (5 mL) contained BNF-induced rat liver microsomes (1 mg/mL),1 mM NADP+, 4.5 mM glucose 6-phosphate, 1.8 units of glucose-6-phosphate dehydrogenase, 3 mM magnesium chloride, 50 mM potassium phosphate buffer (pH 7.5), and trans-DB[a,l]P-8,9-diol (40 µM) dissolved in acetone (250 µL). After incubation for 15 min at 37 °C, a bolus of the fresh NADPH-generating system was added (500 µL). After incubation for an additional 15 min, the mixture was extracted with 15 mL of ethyl acetate/acetone (2:1). The

DB[a,l]P-bis-diols and DNA Adducts organic extracts were dried over anhydrous magnesium sulfate and evaporated under nitrogen. To characterize the metabolites formed, the organic extracts from repetitive incubation mixtures described above were pooled for HPLC analyses. For metabolism studies with recombinant human P450 1A1 microsomes (Supersomes), the same method was used except that the P450 1A1 protein concentration was 0.4 mg/mL and the epoxide hydrolase protein concentration was 0.6 mg/mL. Metabolite Isolation and HPLC. The organic extracts from the metabolic incubations were reconstituted in acetone and filtered, and the filtrate was solvent exchanged into methanol. The methanolic solution was extracted with hexane (saturated with methanol), reduced in volume, and analyzed by HPLC using a 4.6 (or 10) mm × 250 mm Synchrom Synchropak RP-P column (300 Å pore size) (Alltech) with a guard column using an isocratic solvent system of 40% methanol/acetonitrile (2:1)/ 60% water at a flow rate of 2 mL/min. In Vitro Modification of Calf Thymus DNA. DB[a,l]P or racemic (()-trans-DB[a,l]P-8,9-diol (50 µM in acetone) and calf thymus DNA [2 mg in 0.1 M potassium phosphate buffer (pH 7.4)] were incubated for 3 h at 37 °C with recombinant human P450 1A1 microsomes (3 mg) and epoxide hydrolase (3 mg). Reactions were carried out in 3 mL of 0.1 M potassium phosphate buffer (pH 7.4) and 4 µmol of NADPH. The incubations were stopped by the addition of 10 mL of ethyl acetate. The incubation mixtures were extracted three times with 10 mL of ethyl acetate. The aqueous portion was cooled to 0 °C, and 30 mL of cold ethanol was added slowly. The DNA was precipitated, spooled out with a glass rod, washed with 10 mL of ethanol, 10 mL of acetone, and 10 mL of ether, and dried under a stream of nitrogen at 37 °C. The DNA was purified by extraction with phenol, chloroform, and isoamyl alcohol. The DNA samples were resuspended in 1.2 mL of HPLC grade water and stored in a -80 °C freezer until the time of DNA adduct analysis. In Vitro Modification of Nucleotides with anti- and synDB[a,l]PDE. 3′-dGMP and 3′-dAMP were dissolved in 0.05 M Tris buffer (pH 7.4), and anti-DB[a,l]PDE and syn-DB[a,l]PDE were each dissolved in 1 µg/µL tetrahydrofuran. Solutions of each nucleotide (10 mg/mL) in a total volume of 2 mL were mixed with 50 µL of anti-DB[a,l]PDE or syn-DB[a,l]PDE and incubated in the dark for 18 h at 37 °C. The reaction was terminated with an equal volume of ethyl acetate and ether (1: 1). The reaction mixtures were extracted three times with 1 volume of ethyl acetate/diethyl ether (1:1) with centrifugation (1500 rpm for 15 min). Removal of the unreacted nucleotides was achieved by loading the reaction mixtures on primed reverse phase Sep-Pak Classic cartridges (Waters, Millipore Co., Milford, MA) washing with 10 mL of HPLC grade water and then 10 mL of methanol to elute the modified nucleotides. The methanol was removed in vacuo, and the samples were resuspended in 1.2 mL of HPLC grade water and stored at -80 °C until the time of adduct analysis. 32P-Postlabeling Analysis. DNA adducts were analyzed by 32P-postlabeling as previously described using the TLC method (23), the nuclease P1 enhancement technique (24, 25), and HPLC modification (26). The DNA (50 µg) from the in vitro recombinant human P450 1A1 incubations was digested to mononucleotides at 37 °C for 3.5 h with micrococcal nuclease and spleen phosphodiesterase. DNA adducts were enriched by nuclease P1 treatment and evaporated to dryness in vacuo. The samples were incubated with 50 µCi of ATP (3000 Ci/mmol) and 3.5 units of T4 polynucleotide kinase for 30 min at 37 °C, and the complete incubation mixture was spotted onto PEI-cellulose TLC plates. TLC/HPLC Analysis of DNA Adducts by 32P-Postlabeling. TLC was performed on 20 cm × 20 cm plates using the following solvent system: D1, 1 M sodium phosphate (pH 6) with overnight development onto a 10 cm Whatman grade 3MM wick, followed by a wash with water. The remaining spot at the origin was excised and placed in scintillation vials containing 5 mL of ethanol, and the amount of total radioactivity was

Chem. Res. Toxicol., Vol. 11, No. 12, 1998 1599 determined. The ethanol was decanted, and the spot was extracted overnight (18 h) in 1 mL of 4 M pyridinium formate (pH 4). The extracts were transferred to 1.5 mL microcentrifuge vials and the particulates sedimented by centrifugation (10 000 rpm for 1 min). The samples were reduced to dryness in vacuo, resuspended with 100 µL of methanol/sodium phosphate buffer (pH 2) (9:1), vortexed, and centrifuged again (10 000 rpm for 1 min). The supernatant (75 µL) was removed and spiked with 4 nmol (8 µL in methanol) of cis-9,10-dihydroxy-9,10-dihydrophenanthrene. The volume was adjusted to 100 µL with a mixture of methanol/0.3 M NaH2PO4 buffer (pH 2) (9:1) and the mixture placed in 100 µL autosampler vials for HPLC analyses. Separation of the 32P-postlabeled 3′,5′-bisphosphate adducts was carried out on a Hewlett-Packard series 1100 HPLC system (Hewlett-Packard Co., Wilmington, DE) equipped with a model HP G1311A quaternary solvent delivery system, a model HP G1322 degasser, a model HP 1315 diode array detector, and a model HP G1313A autosampler. The extracts of the excised adducts (95 µL) were applied to a 5 µm, 4.6 mm × 250 mm Zorbax phenyl-modified column (MAC-MOD Analytical, Inc., Caddies Ford, PA) and eluted with a linear gradient using the following solvent system. Solvent A was 0.3 M NaH2PO4 buffer (pH 2); solvent B was 90% methanol and 10% solvent A. The gradient was 95% A and 5% B from 0 to 3 min, 5 to 48% B from 3 to 40 min, 48 to 60% B from 40 to 80 min, 60 to 95% B from 80 to 100 min, 95% B from 100 to 105 min, and 95 to 5% B from 105 to 120 min. The flow rate was 1 mL/min, and the column was allowed to equilibrate at the initial solvent ratio (95% A: 5% B) for 15 min before subsequent analysis. Analyses for the presence of 32P were carried out using an in-line flowthrough scintillation counter (model A-500 Flow-One, β Radiomatic Instruments & Chemical Co., Inc., Tampa, FL). The retention times of 32P-postlabeled DNA adducts were expressed as relative retention time (RRT) which was calculated by dividing the retention time of the 32P-postlabeled DNA adduct by the retention time of the UV absorbing internal standard (cis-9,10-dihydroxy-9,10-dihydrophenanthrene). The reproducibility of the RRT was (0.03. Computational Studies with Molecular Modeling and Theoretical UV Spectra. Initially, the ground state geometry was fully optimized by means of the semiempirical HartreeFock Austin Model 1 (AM1) method (27). No geometrical constraints were imposed, and all the bond lengths, angles, and dihedral angles were optimized using the SPARTAN molecular modeling program (28). In a second step, on the basis of the optimized geometry, the lowest optical transitions and their corresponding intensities were computed with the semiempirical Hartree-Fock intermediate neglect of differential overlap (INDO) Hamiltonian coupled to a single-configuration interaction (SCI) formalism. The CI expansion was generated by the promotion of an electron from one of the 25 highest occupied levels to one of the 25 lowest unoccupied levels. The ZINDO/S method (29, 30) was parametrized to reproduce spectroscopic transition and, as implemented in HyperChem, was used to calculate the UVvisible spectra (31).

Results Synthesis and Chiral Resolution of (()-trans-DB[a,l]P-8,9-diol. The synthesis of (()-trans-DB[a,l]P-8,9diol was accomplished by the osmium tetroxide, DDQ, sodium borohydride method with the modification of eliminating the two acetylation steps. The use of 2-propanol as the solvent for the sodium borohydride reduction of the quinone facilitated the trans reduction, with no evidence of the cis isomer. The low- and high-resolution mass spectra, NMR spectra (Table 1), and UV spectra (Figure 1) were consistent with its structure, and the UV and NMR data were similar to those of the metabolically formed product (8). The coupling constant of the 8,9dihydrodiol protons of 10.5 Hz suggested a trans quasidiequatorial conformation. Chiral separation on a co-

1600 Chem. Res. Toxicol., Vol. 11, No. 12, 1998

Nesnow et al.

Table 1. 300 MHz 1H NMR Spectra of trans-DB[a,l]P-8,9-diol and Metabolic DB[a,l]P-bis-diolsa DB[a,l]P metabolite (()-trans-DB[a,l]P-8,9-diol

H1

H 2, H 3, and H6

H8

H9

8.92-8.97 7.55-7.73 8.81-8.84 8.67

7.98-8.06 4.86

4.98

m 1H

d 1H J5,6 ) 8.22 8.61

m 1H

d s m 1H 1H 1H J9,8 ) 10.5 5.02 8.21 4.73

m 1H

m 1H

m 1H

7.88

d 1H J8,9 ) 10.5 4.97

4.64

6.26

7.15

d 1H J5,6 ) 8.10

d 1H J7,6 ) 7.26

d 1H J8,9 ) 9.0

d s 1H 1H J9,8 ) 9.0

t of d 1H J12,11 ) 10.9 J12,13 ) J12,14 ) 2.2 4.65

d of d 1H J13,12 ) 2.2, J13,14 ) 10.2

d of d 1H J14,12 ) 2.2, J14,13 ) 10.2

6.25

7.12

m 3H

H4

m 1H

H5

H7

H10

H11

H12

H13

H14

8.27 7.98-8.06 7.55-7.73 7.55-7.73 8.92-8.97

(()-trans,trans-DB[a,l]- 8.43 P-8,9:11,12-bis-diol (4) d 1H J1,2 ) 8.11

7.59-7.77 8.74

(()-trans,trans-DB[a,l]- 8.45 P-8,9:11,12-bis-diol (6) d 1H J1,2 ) 8.08 trans-DB[a,l]P-8,9: 8.59 13,14-bis-diol (2) m 1H

7.61-7.72 8.75

8.68

7.94

4.92

5.04

m 3H

d 1H J3,4 ) 8.01 7.66-7.77 9.02-9.08

d 1H J5,6 ) 8.14 9.02-9.08

d 1H J7,6 ) 7.27 8.13-8.16

d 1H J8,9 ) 10.8 4.42

d s 1H 1H J9,8 ) 10.8 4.82 8.31

d 1H J11,12 ) 11.4 7.66-7.77

d 1H J12,11 ) 11.4 6.71

d 1H J13,14 ) 9.86 4.52

d 1H J14,13 ) 9.86 5.48

m 3H

m 1H

m 1H

m 1H J8,9 ) 10.4

m m 1H 1H J9,8 ) 10.4

m 1H

m 1H J12,13 ) 5.97

m 1H J13,12 ) 5.97

br s 1H

m 3H

d 1H J3,4 ) 8.01

m 1H

d 1H J11,12 ) 10.9

8.28 4.73

a NMR spectra were recorded in acetone-d . The following data are listed: chemical shifts in parts per million, multiplicity, number of 6 protons, and coupling constants in hertz.

valently bonded (R,R)-Whelk-O1 5 µm chiral stationary phase column gave a near-baseline separation of the enantiomeric trans-DB[a,l]P-8,9-diols. The CD spectra of each enantiomeric dihydrodiol indicated that the first Cotton effect of the early-eluting dihydrodiol (diol 1) was positive and the first Cotton effect of the later-eluting dihydrodiol (diol 2) was negative (Figure 2). Identification of DB[a,l]P-bis-diols. Metabolism of racemic (()-trans-DB[a,l]P-8,9-diol by BNF-induced rat liver microsomes gave six metabolites on the basis of reverse phase HPLC separation (Figure 3). Three major metabolites, metabolites 2, 4, and 6, were identified by their UV, NMR, and mass spectral data. Metabolites 4 and 6 had almost identical NMR (Table 1), UV (Figure 1), and mass spectra. Low-resolution mass spectra of both 4 and 6 gave a parent ion (m/z 370) with successive losses of H2O (m/z 352) and H2O/CHO (m/z 305). High-resolution mass spectra of 4 gave the exact mass of 370.1218 (C24H18O4 requires 370.1205), consistent with a DB[a,l]Pbis-diol structure. Metabolites 4 and 6 were identified as trans,trans-DB[a,l]P-8,9:11,12-bis-diols (Scheme 1). The NMR spectra of these metabolites had many of the features of the spectra of both trans-DB[a,l]P-8,9-diol (8) and trans-DB[a,l]P-11,12-diol (8, 32). The NMR spectra of both 4 and 6 revealed three downfield protons (H4, H5, and H1) as doublets, suggesting that oxidation had not occurred at carbons 1-4, as only two downfield protons would be expected had this occurred (Figure 4 and Table 1). The appearance of two coupled olefinic protons (H14 and H13) and two sets of coupled carbinol protons (H8 and H9, and H11 and H12) further supported a bis-diol structure of metabolites 4 and 6. COSY spectra identified significant couplings between protons H13 and H14, H8 and H9, and H11 and H12, and a small, but significant, coupling between protons H5 and H7 (data not shown). In addition, H12 was weakly coupled to H13 and H14 to about the same extent, giving the observed triplet of

doublets for H12 (Table 1). The definitive assignment of metabolites 4 and 6 as (()-trans,trans-DB[a,l]P-8,9:11,12-bis-diols was deduced from 1D-NOE difference spectra recorded for metabolite 4. Irradiation of olefinic proton H14 gave a strong NOE effect (13.3% increase) between H1 and H14 and a weaker NOE effect (8.1% increase) with H13 (Figure 4). Concomitantly, irradiation of H1 gave a similar increase in H14 (data not shown). These results are supported by molecular modeling of this bis-diol which gave H1-H14 and H13-H14 distances of 2.14 and 2.51 Å, respectively (Table 2). The carbinol coupling constants were slightly different in these metabolites with J8,9 equaling 9 and 10.8 Hz in metabolites 4 and 6, respectively. Similarly, the carbinol coupling constants for J11,12 were 10.9 and 11.4 Hz in metabolites 4 and 6, respectively, suggesting that each dihydrodiol moiety in each bis-diol assumed a trans quasi-diequatorial configuration. Metabolite 2 was identified as trans-DB[a,l]P-8,9:13,14-bis-diol (Scheme 1). It exhibited low-resolution mass spectra [parent ion (m/z 370) with successive losses of H2O (m/z 352) and H2O/CHO (m/z 305)] and highresolution mass spectra [exact mass of 370.1205 (C24H18O4)] consistent with a DB[a,l]P-bis-diol structure. Metabolite 2 exhibited a complex NMR spectra of at least two conformers as multiple absorbances were observed for protons H8-H10, H12, and H13 that were not explained by direct coupling. For example, at a reduced temperature (-40 °C), proton H10 was two peaks approximately 1.5 Hz apart with a peak height ratio of 0.8:1. Nominally, there were three high-field protons (H4, H5, and H1), two olefinic protons (H11 and H12), and four carbinol protons (H8, H9, H13, and H14) (Figure 5 and Table 1). In addition to coupling between the aromatic protons, COSY spectra identified an H8-H9 coupling of 10.4 Hz consistent with a trans quasi-diequatorial configuration. Also identified was a coupling between H12 and H13, where J12,13 ) 5.97

DB[a,l]P-bis-diols and DNA Adducts

Chem. Res. Toxicol., Vol. 11, No. 12, 1998 1601

Figure 3. HPLC analysis of (()-trans-DB[a,l]P-8,9-diol metabolites from BNF-induced rat liver incubations. Samples were taken from the combined incubations. Unmetabolized (()-transDB[a,l]P-8,9-diol elutes at 57.9 min.

Figure 1. Observed and calculated UV spectra of DB[a,l]P metabolites. The observed UV spectra were obtained in methanol and are represented as curves. The calculated UV spectra were obtained as described in Experimental Procedures, and the transitions are represented as vertical lines: (top panel) (()trans-DB[a,l]P-8,9-diol, (middle panel) trans,trans-DB[a,l]P-8,9: 11,12-bis-diol (4) (metabolite 6 gave identical spectra), and (bottom panel) trans,trans-DB[a,l]P-8,9:13,14-bis-diol (2).

Figure 4. 1H NMR spectra (300 MHz) of trans,trans-DB[a,l]P8,9:11,12-bis-diol (4) in acetone-d6: (top panel) one-dimensional 1H NMR and (bottom panel) one-dimensional NOE NMR difference spectra. The negative band represents the point of NOE irradiation.

Figure 2. Circular dichroism spectra of the early-eluting (diol 1) and late-eluting (diol 2) trans-DB[a,l]P-8,9-diol enantiomers in methanol previously separated by HPLC on an (R,R)-WhelkO1 5 µm covalently bonded chiral stationary phase column.

Hz, and no coupling between H13 and H14 (broadened peaks were observed), suggestive of either a transdihydrodiol in a quasi-diaxial conformation or possibly a cis-dihydrodiol. All of these couplings were confirmed by

homonuclear decoupling studies. A delayed COSY spectra identified a significant coupling between H5 and H7. Spatial orientations of protons H1, H14, H13, and H12 were unambiguously identified by NOE and/or COSY spectra. In detailed NOE studies, NOE irradiation of carbinol proton H14 (5.48 ppm) gave NOE effects with protons H1 (8.59 ppm) and H13 (4.52 ppm) (Figure 5). NOE irradiation of H1 (8.59 ppm) gave NOE effects with H14 (5.48 ppm) and the multiplet at 7.66-7.77 ppm, and NOE irradiation of proton H12 (6.71 ppm) gave NOE effects at proton H13 (4.52 ppm) (data not shown). Proton H12 exhibited a multiplet of two groups of four peaks each, with the major coupling to H13 (J12,13 ) 5.97 Hz). The

1602 Chem. Res. Toxicol., Vol. 11, No. 12, 1998

Nesnow et al.

Table 2. Physical Parameters from Computational Analysis of DB[a,l]P and Its Metabolitesa

PAH

H1-H14 distance (Å)

DB[a,l]P trans-DB[a,l]P-8,9-diol trans,trans-DB[a,l]P-8,9:11,12-bis-diol trans,trans-DB[a,l]P-8,9:13,14-bis-diol

2.06 2.06 2.14 1.93

a

H13-H14 distance (Å)

H8-H9 dihedral angle (deg)

2.51 2.56

11.1 12.6 12.9

H11-H12 dihedral angle (deg)

H13-H14 dihedral angle (deg)

C1-C14 distance (Å)

C1-C1a-C14a-C14 dihedral angle (deg)

67.3

2.94 2.96 2.91 3.07

35.9 37.3 33.8 43.4

15.1

Parameters obtained using the SPARTAN molecular modeling program on minimized conformations.

Figure 5. 1H NMR spectra (300 MHz) of trans-DB[a,l]P-8,9: 13,14-bis-diol (2) in acetone-d6: (top panel) one-dimensional 1H NMR and (bottom panel) one-dimensional NOE NMR difference spectra. The negative band represents the point of NOE irradiation.

multiplet at 7.66-7.77 ppm was integrated for four protons (H2, H3, H6, and H11). The downfield shifts of H11 and H12 are explained by the extensive conjugation of the olefinic C11-C12 to the benzo[g]chrysene nucleus as observed in the red-shifted UV spectra, and the severe distortion of this molecule. With a dihedral angle of 43.4°, the benzo ring (C1-C4) is twisted and above the phenanthrenoid plane and the partially saturated ring (C11-C14) is twisted and below the phenanthrenoid plane. This great distortion with partial ring overlap has the potential to affect specific protons through ring current effects. The distortion also explains the lower than expected chemical shift of the H1 proton (8.59 ppm) as there is less overlap of H14 and H1 than found in DB[a,l]P-11,12,13,14-tetrol (33). The lack of observable coupling between H11 and H12 and H13 and H14 is explained by the rapid interconversion of the conformers. Computational Analysis. Since suitable standards for UV analysis were not available, we applied computational approaches to estimate the UV spectra of these agents. The optimized molecular structure of trans-DB[a,l]P-8,9-diol and the two bis-diols, trans,trans-DB[a,l]P-

8,9:11,12-bis-diol and trans,trans-DB[a,l]P-8,9:13,14-bisdiol, revealed that all exhibited conformational minima with all trans-8,9-dihydrodiol functionalities in the quasidiequatorial conformation (H8-H9 dihedral angle of