341
Chem. Res. Toxicol. 1989,2,341-348
Characterization of DNA Adducts Derived from (f)-irans-3,4-Dihydroxy-anti-l,2-epoxy1,2,3,4-tetrahydrodibenz[ a ,\]anthracene and ( &)-'l-Methyl-trans -3,4-dihydroxy-antl-1,2-epoxy1,2,3,4-tetrahydrodibenz[ a ,/]anthracene Raghunathan V. Nair,? Rosalynn D. Gill,+Cecilia Cortez,l Ronald G. Harvey,t and John DiGiovanni**+ Department of Carcinogenesis, University of Texas M . D. Anderson Cancer Center, Science Park-Research Division, P.O. Box 389, Smithville, Texas 78957, and The Ben May Institute, University of Chicago, Chicago, Illinois 60637 Received June 30, 1989 Structural characterizations of the r 1' A adducts derived from reaction of the racemic bay region anti-diol epoxides of dibenz[aj]t .thracene and 7-methyldibenz[aj]anthracenewith calf thymus DNA are presented. Quantities of adducts necessary for spectroscopic characterization were obtained from reactions of the respective diol epoxides with individual deoxyribonucleotides. Both hydrocarbon diol epoxides showed similar adduct profiles upon reaction with calf thymus DNA in vitro which were composed mainly of three deoxyguanosine and four deoxyadenosine adducts. No significant modification of pyrimidine bases in DNA was detected with either of the diol epoxides. Approximately 3 times more deoxyguanosine than deoxyadenosine residues in the DNA were found to be modified by both diol epoxides. The DNA reactions showed very similar stereo- and enantioselectivitieswith both diol epoxides. The stereochemistries of addition of the purine bases to the diol epoxides were determined from analysis of the NMR spectra of individual adducts. The predominant adducts formed were products of trans addition of the exocyclic amino group of purines to the diol epoxides. The enantiomeric nature of the various adducts was determined from reaction of the individual deoxyribonucleotides with the pure (+)-anti-diol epoxide of dibenz[aj]anthracene. The major deoxyguanosine and deoxyadenosine adducts from reactions with DNA were found to arise from the (+)-enantiomer of both hydrocarbon diol epoxides. The high reactivities of both diol epoxides (24-38%) with DNA in 7 lution are consistent with the high tumor-initiating activity exhibited by the diol epoxide of c,.oenz[aj]anthracene relative to the parent hydrocarbon.
Introduction The metabolic activation of carcinogenic alternant polycyclic aromatic hydrocarbons (PAH)' to bay region diol epoxides and subsequent reaction of these diol epoxides with cellular DNA leading to the process of tumor initiation constitute one of the widely accepted present-day theories in chemical carcinogenesis (1,2). Although there have been several reports (3-5) correlating the carcinogenicities of various PAH with their covalent binding to DNA, the complete structural characterization of the major DNA adducts formed has been accomplished only with a select number of PAH (6-10). Thus, a correlation between levels of specific DNA adducts formed in vivo with the varying tumor-initiating potencies of different PAH still remains to be delineated. Early studies with benzo[a]pyrene ( B [ a ] P )in mouse skin in vivo revealed the formation of a significant proportion of deoxyguanosine (dGuo) adducts derived from trans addition of the exocyclic amino group of dGuo with (+)-7&8a-dihydroxy9ci,1O~-epoxy-7,8,9,1O-tetrahydrobenzo[ alpyrene (1). More recent reports ( 11-14) have indicated that quantitative differences in modification of deoxyadenosine (dAdo) residues in cellular DNA may play a critical role in the tumor-initiating capacities of some PAH. The modification of dGuo residues at N-7 with subsequent depurination University of Texas M. D. Anderson Cancer Center.
* University of Chicago.
Chart I
\
\
\
0
/
w H
OH 0
w:
/
R
R
1, R=H 2, R=CH3
3, R=H, (+) 4, R=CHs, (+)
0-H
W
0H .
\
/
/
R 5, R=H 6, R=CH3
0
/
H
OH
R 7, R=H 8, R=CH3
may also be a contributing factor in the carcinogenicity of certain PAH (15). In attempts to further explore the Abbreviations: PAH, polycyclic aromatic hydrocarbons; B[a]P, benzo[a]pyrene;dGuo, deoxyguanosine;dAdo, deoxyadenosine;DB[aj]A, dibenz[aj]anthracene;7-Me-DB[aJ]A,7-methyldibenz[aj]anthracene; DB[aj]ADE, ( ~ ) - t r a n s - 3 , 4 - d ~ y ~ o x y - a n ~ i - l , 2 - e ~ x y - l , 2 , 3 , 4 ~ t r ~ y ~ ~ dibenz[aj]anthracene; B[c]Ph, benzo[c]phenanthrene; 7-Me-DB[oj]ADE, (~)-7-methyl-trans-3,4-dihydroxy-anti-1,2-epoxy-l,2,3,4-tetrahydrodibenz[aj]anthracene;CD, circular dichroism; DMBA, 7,12-dimethylbenz[a]anthracene;5-MeC, 5-methylchrysene.
0893-228~/89/2702-0341$01.50/0 0 1989 American Chemical Society
342 Chem. Res. Toxicol., Vol. 2, No. 5, 1989
relationship of various dGuo versus dAdo adducts formed in mouse skin with the tumor-initiating ability of a given PAH, we have undertaken the comparative study of model compounds dibenz[aj]anthracene (DB[aj]A) (1) and its 7-methyl homologue (7-Me-DB[aj]A)(2) (Chart I). Previous reports (16,17) from this laboratory have described the tumor-initiating activities of compounds l and 2 and the diol epoxide 3 (DB[aj]ADE) in mouse skin. 7-Me-DB[a,j]A as well as diol epoxide 3 were approximately 3-fold more potent as tumor initiators than the parent hydrocarbon 1. The relative higher tumor-initiating activity of diol epoxide 3 over the parent hydrocarbon 1 found in this earlier study (17) is parallel with observations made by using benzo[c]phenanthrene (B[c]Ph) (18). The high tumor-initiating activities of the diol epoxides of these structurally related hydrocarbons possibly reflects their chemical stability relative to cellular transport and bioavailability (19-21). The comparative quantitative evaluation of the levels of various DNA adducts formed in vivo from the closely related compounds (1 and 2) with established differential biological activity may provide some insight into the significance of modification of dGuo versus dAdo residues in DNA on their ultimate biological response. As part of this objective, we now report the structural characterization of various dGuo and dAdo adducts formed upon reaction of calf thyumus DNA and deoxyribonucleotides with the racemic bay region anti-diol epoxides 3 and 4 (7-Me-DB[a,j]ADE).
Materials and Methods Chemicals. Sodium salts of 2'-deoxyguanosine 5'-mOnOphosphate and 2'-deoxyadenosine 5'-monophaphate, calf thymus DNA, DNase I (bovine pancreas, EC 3.1.21.1), snake venom phosphodiesterase (Crotalus atrox, EC 3.1.4.1), and Escherichia coli alkaline phosphatase (type 111, EC 3.1.3.1) suspension in 3.5 M ammonium sulfate were purchased from Sigma Chemical Co., St. Louis, MO. [3H]dGuoand [3H]dAdohaving specific activities of 10-22 Ci/mmol were obtained from ICN Laboratories, Irvine, CA. Racemic anti-diol epoxides 3 and 4 were prepared according to published procedures (21). The optically pure enantiomers of trans-3,4-dihydroxy-3,4dihydrodibenz[a,j]anthracene were prepared by reaction of the racemic dihydrodiol (250 mg) with (-)-menthoxyacetyl chloride by the usual procedure (22,23). The crude product was dissolved in CH2C12and passed through a short column of Florisil eluted with CHzC12to yield the mixed bis(menth0xyacetate)isomers (500 mg). This mixture was dissolved in a small volume of THF and separated by HPLC on a Du Pont Zorbax silica column (25 cm X 21.2 mm) eluted with 2% THF in hexane, flow rate 9 mL/min, using 30-40 mg/injection. There were obtained 150 mg of the early-eluting (-)-diastereomer ( [ a ] 2=3 -304) and 120 mg of the late-eluting (+)-diastereomer ([a]23 = +261). The purity of the fractions was rechecked by reinjection on a similar analytical HPLC column and found to be greater than 98.5%. The chemical shifts and couplings of the methylene protons of the OCOCH20 group in the NMR spectrum of the (-)-isomer were 8m 4.l3,8Agj 4.07, and Jm = 16.76 Hz, JA,Bt = 16.84 Hz, respectively. The corresponding values for the (+)-isomer were 8m 4.14,8Agj 4.08, and Jm = 28.03 Hz, J ~ g t= 39.80 Hz, respectively. On the basis of the empirical rule that the less polar diastereomer with larger negative optical rotation and minor magnetic nonequivalence of the methylene protons of the OCOCH20group has the R,R absolute configuration (24,25),the (-)-diastereomer may be tentatively assigned the 3R,4R configuration. The (-)-bis(menth0xyacetate) (150 mg) was converted to the (-)-3,4-dihydrodiol by treatment with NaOCH3 (25 mg) in CH30H (30 mL) and THF (20 mL) at reflux temperature for 10 min. The usual workup gave the free dihydrodiol, which was further p d i e d by trituration with ether to yield the (3R,4R)-dihydrodio1(60mg): [.Iz3 = -303. A similar procedure was employed to prepare the (3S,4S)-dihydrodiol: [a]23= +287. Epoxidation of the (-)-(3R,4R)-dihydrodiol with m-chloroperbenzoic acid by the procedure employed for preparation of
Nair et al. the racemic diol epoxide (21)furnished the corresponding (+)-diol epoxide which may be assigned 1R,2S,3S,4R absolute stereochemistry on the basis of its mode of preparation. Similarly, epoxidation of the (+)-(3S,4S)-dihydrodiol gave the (-)(lS,2R,3R,4S)-diol epoxide. The diol epoxides should be considered as carcinogens and handled with care. Instrumentation. Ultraviolet absorption spectra were obtained in CH30H on a Shimadzu UV-160 UV-visible recording spectrophotometer. Circular dichroism (CD) spectra were recorded in CH30H on a Jasco J500A spectropolarimeter at the Baylor College of Medicine, Houston, TX. The CD spectra were normalized to the absorbance values at A- of 1.0 and 0.6 for dGuo and dAdo adducts, respectively. Proton NMR spectra were obtained in CD30D on a General Electric GN-500 spectrometer at the University of Texas in Austin. HPLC separations were carried out by using a Shimadzu LC-GA equipped with Shimadazu SCL-6A system controller, SPD-6A W detector, and CR-501 data processor. Radioactivity of HPLC fractions was measured by using a Beckman LS 1800 liquid scintillation counter. DNA Binding. Calf thymus DNA labeled with [3H]dGuoor [3H]dAdo (1.3-3.5 pCi/pg of DNA) was prepared as previously described (26) by using a modification of the method supplied with the nick translation kit from Bethesda Research Laboratories, Gaithersburg, MD. Labeled DNA (13 pg) was combined with the stock of unlabeled DNA (400 pg) and reacted with diol epoxides 3 or 4 (100 pL of a 1mg/mL solution in acetone) in 0.01 M Tris buffer (1.5 mL, pH 7.0) at 37 "C for 16 h. DNA was precipitated by addition of EtOH (2 volumes) and 7.5 M NH40Ac (1/2 volume), washed sequentially with EtOH (2X), HPLC-grade acetone, and anhydrous Et20, and dried (Nz). The samples of DNA were dissolved in 0.01 M Tris-MgC12 buffer (1.5 mL, pH 7.0) and hydrolyzed to deoxyribonucleosides by using DNase I, snake venom phosphodiesterase, and alkaline phosphatase as described previously (27). The hydrolysates were purified on Sephadex LH20 according to published procedures (28) to afford the modified deoxyribonucleosides. Synthesis of DB[a j]ADE Adducts. To a solution of 2'deoxyguanosine 5'-monophosphate sodium salt (1g) in 50 mM Tris-HC1buffer (50 mL, pH 7.0) was added a solution of racemic 3 (2 mg) in THF (1mL). The mixture was slowly shaken for 20 h in a water bath maintained at 37 "C. The reaction mixture was cooled to room temperature and extracted with EtOAc (5 X 20 mL), followed by E t 2 0 (3 X 25 mL). The organic extracts were combined and evaporated under reduced pressure to afford a mixture of isomeric tetrols 5 and 7. (See below for separation and isolation of tetrols.) Following removal of trace organic solvents by purging N2 through the solution, the aqueous phase was diluted by addition of 200 mM Tris-HC1 buffer (50 mL, pH 9.0). To the solution was added E. coli alkaline phosphatase type I11 (60 units), and the solution was slowly shaken overnight in a water bath maintained at 37 "C. The reaction mixture was cooled to room temperature, divided into three equal portions, and passed through three Sephadex LH-20 columns (1.5 X 6 cm) previously equilibrated with H20. The columns were eluted with H 2 0 (25 mL each), followed by CH30H (25 mL each). The H20 and CH30H eluates were collected separately. Evaporation of the CH30H solution afforded the isomeric dGuo adducts, which were separated and isolated by HPLC. Analytical HPLC of the adducts on an Ultrasphere ODS column (4.6 mm X 25 cm) eluted with 45% CH30H at 1mL/min showed five component peaks. Individual peaks were isolated by semipreparative HPLC (Ultrasphere ODS 10 mm X 25 cm) using 48% CH30H at a flow rate of 3 mL/min. In the order of their elution, the adducts were labeled G1, G2, G3, G4, and G5. The dAdo adducts were similarly prepared by reacting the racemic diol epoxide 3 with 2'-deoxyadenosine 5'-monophosphate sodium salt. Following separation of the adducts on Sephadex LH-20 columns and concentration of the methanol eluate, analytical HPLC using 50% CH30H at 1mL/min eluted four component peaks. Individual peaks were isolated by using a semipreparative column by elution with 53% CH30H at a flow rate of 3 mL/min. In the order of their elution, the adducts were labeled A l , A2, A3, and A4. The pure (+IDB[a,j]ADE (1 mg) was reacted in a similar manner with the individual deoxyribonucleotides(0.5 g) to afford adducts G2, G3, A l , and A4.
Dibenz[a,j]anthracenediol Epoxide-DNA Adducts
Chem. Res. Toxicol., Vol. 2, No. 5, 1989 343
Synthesis of 7-Me-DB[a j]ADE Adducts. Racemic diol epoxide 4 (2.1 mg in 3 mL of THF) was reacted with 2’-deoxyguanosine 5’-monophosphate sodium salt (1g) under conditions identical with those described for the synthesis of corresponding DB[aJ]ADE adducts. Following alkaline phosphatase treatment, the reaction mixture was subjected to Sephadex LH-20 chromatography as described. Each of the three columns were eluted with 35-mL portions of water followed by 45-mL portions of CH30H. HPLC analysis of the adducts obtained upon evaporation of the combined CH30H eluates showed five component peaks which were isolated (Ultrasphere ODs, 10 mm X 25 cm) by elution with 48% CH30H for 120 min followed by a linear gradient of 48-100% CH30H for 30 min at a flow rate of 3 mL/min. In the order of their elution, the adducts were labeled 7G1, 7G2, 7G3, 7G4, and 7G5. Likewise, racemic diol epoxide 4 (2.3 mg in 3 mL of THF) was reacted with 2’-deoxyadenosine 5’-monophosphate sodium salt (1g) to afford the corresponding dAdo adducts. Individual adducts were isolated by semipreparative HPLC using 53% CH30H for 120 min followed by a linear gradient of 53-100% CH30H for 30 min at a flow rate of 3 mL/min. In the order of their elution, adducts were labeled 7A1, 7A2,7A3, and 7A4. Isolation of Tetrols 5-8. Tetrols 5 and 7 were separated and isolated by semipreparative HPLC (Ultrasphere ODs, 10 mm x 25 cm, 50% CH30H at 3 mL/min) of the residue obtained upon evaporation of the organic extracts following reaction of diol epoxide 3 with the deoxyribonucleotides. Likewise, HPLC separation of organic extractable residue obtained following reaction of diol epoxide 4 with the deoxyribonucleotidesafforded tetrols 6 and 8. Tetrols trans-5 and trans-6 were eluted earlier than tetrols cis-7 and cis-8. Additional quantities of the isomeric tetrols 5 and 7 were obtained by hydrolysis of anti-diol epoxide 3 (2 mg dissolved in 0.6 mL of THF in 200 mL of HzO) at 37 “C for 22 h. EtOAc extraction of the aqueous mixture followed by evaporation of the combined organic extract and subsequent HPLC purification as described afforded tetrols 5 and 7.
Results T h e reaction of t h e racemic diol epoxides 3 a n d 4 with calf t h y m u s DNA in vitro followed by enzymatic degradation a n d isolation of modified deoxyribonucleosides provided t h e HPLC profiles shown in Figures 1 a n d 2, respectively. Incorporation of tritiated dGuo a n d dAdo residues into DNA in separate experiments by utilizing a nick translation procedure prior t o reaction with diol epoxides unambiguously identified peaks 1-111 as arising from dGuo residues whereas peaks IV-VI1 were unambiguously identified as arising from dAdo residues (Figures 1a n d 2). Estimation (UV) of the extent of reaction of diol epoxides 3 a n d 4 with DNA from t h e recovered tetrols indicated approximately 24% conversion of diol epoxide 3 and 38% conversion of diol epoxide 4 t o their respective DNA adducts. Diol epoxides 3 and 4 afforded qualitatively similar adduct profiles upon reaction with calf thymus DNA in vitro. In this regard, both diol epoxides reacted extensively with dGuo and dAdo residues (dGuo/dAdo ratios of -3.0 for both diol epoxides). T h u s , t h e presence of a methyl group a t position 7 of D B [ a j ] A did not appear t o significantly influence the nature of the reaction with calf thymus DNA in vitro. Quantities of adducts sufficient for structural characterization were prepared from reaction of diol epoxides 3 a n d 4 with t h e respective deoxyribonucleotides as described under Materials and Methods. Reaction of racemic diol epoxide 3 with 2’-deoxyguanosine 5‘-monophosphate sodium salt afforded five products which were separated o n HPLC. Adducts G1, G2, a n d G 3 corresponded with adduct peaks 1-111 derived from the DNA reactions (Figure 1). T h e ultraviolet absorption spectra of these adducts in C H 3 0 H exhibited the following A, values: G1,281.2 nm; G2, 280.8 nm; G3, 281.2 nm; G4, 291.6 nm; G5, 292.4 nm. Likewise, reaction of t h e 7-methyl derivative 4 with 2‘-
26
52
70
104
TIME (mln)
Figure 1. HPLC elution profiles of deoxyribonucleoside adducts obtained from reactions of diol epoxide 3 with (A) calf thymus DNA with elution monitored by W detection; (B) tritium-labeled calf thymus DNA with elution monitored by radioactive scintillation counting; and (C) deoxyribonucleotides with elution monitored by UV detection. Benzo[a]pyrene 9,lO-dihydrodiol (B[a]P-9,lO-diol)was added to the adduct samples as a UV marker. The adducts were eluted on an Ultrasphere ODS column (4.6 mm X 25 cm) with 47% CH30H for 50 min followed by sequential linear gradients of 47-60% CH30H for 50 min and 60-100% CH30H for 15 min. The column flow rate was 1 mL/min. Note that panel B is a composite chromatogram from separate reactions using [3H]dGuo- or [3H]dAdo-labeled DNA and panel C is a similar composite from separate reactions with the individual deoxyribonucleotides. Panels B and C do not represent the actual proportions of dGuo vs dAdo adducts since approximately equal amounts of radioactivity or UV-absorbing material were combined from separate reactions to make up these chromatograms. deoxyguanosine 5’-monophosphate sodium salt resulted in five products having A, values in CH30H as follows: 7G1,285.8 nm; 7G2, 285.2 nm; 7G3, 286.0 nm; 7G4, 296.4 nm; 7G5, 297.6 nm. Adducts 7G1, 7G2, a n d 7G3 corresponded t o peaks 1-111 in Figure 2. T h e reaction of both diol epoxides (3 a n d 4) with 2’-deoxyadenosine 5‘-monophosphate sodium salt afforded four products, corresponding t o adduct peaks IV-VI1 in Figures 1 a n d 2, re-
344
Chem. Res. Toxicol., Vol. 2, No. 5, 1989
Nair et al. -H-1 H-2
E
3
0 Y
Y< m
i
