Marked Differences in Base Selectivity between DNA and the Free

Chem. Res. Toxicol. 2000, 13, 883-890

883

Marked Differences in Base Selectivity between DNA and the Free Nucleotides upon Adduct Formation from Bay- and Fjord-Region Diol Epoxides SreenivasaRao Vepachedu, Naiqi Ya, Haruhiko Yagi, Jane M. Sayer,* and Donald M. Jerina Laboratory of Bioorganic Chemistry, National Institute of Diabetes and Digestive and Kidney Diseases, The National Institutes of Health, Bethesda, Maryland 20892-0820 Received March 31, 2000

Distributions of adducts formed from each of the four optically active isomers of 3,4-dihydroxy1,2-epoxy-1,2,3,4-tetrahydrobenzo[c]phenanthrene and of 7,8-dihydroxy-9,10-epoxy-7,8,9,10tetrahydrobenzo[a]pyrene (BcPh and BaP diol epoxides) on reaction with an equimolar mixture of deoxyadenosine and deoxyguanosine 5′-monophosphates were compared with the known adduct distributions from these diol epoxides (DEs) upon reaction with calf thymus DNA in vitro. In the presence of an equimolar (100 mM total) mixture of dAMP and dGMP, the efficiency of formation of all types of adducts relative to tetraols is comparable for both the BaP (∼4060%) and BcPh (∼30-40%) diol epoxides. This is in contrast to the partitioning between tetraols and adducts observed with DNA, where the BcPh DEs form adducts much more efficiently than the BaP DEs. Preference for trans versus cis ring opening by the exocyclic amino groups of the free nucleotides in the dAMP/dGMP mixture is greater for the DE diastereomer in which the benzylic hydroxyl group and the epoxide oxygen are trans (DE-2). This is qualitatively similar to the preferences for trans versus cis adduct formation on reaction of these isomers with DNA, as well as trans versus cis tetraol formation on their acid hydrolysis. For the BcPh DE isomers, competitive reaction between dGMP and dAMP gives 40-62% of the total exocyclic amino group adducts as dA adducts. A similar distribution of dG versus dA adducts had previously been observed on reaction of the BcPh DEs with DNA, except in the case of (+)3(R),4(S)-dihydroxy-1(R),2(S)-epoxy-1,2,3,4-tetrahydrobenzo[c]phenanthrene, which gives ∼85% dA adducts on reaction with DNA. With the BaP DEs, 60-77% of the exocyclic amino group adducts formed upon competitive reaction with the free nucleotides are derived from dGMP. The observed dG selectivity of these BaP DEs is much smaller with the nucleotide mixture than it is with DNA, leading to the conclusion that DNA structure has a much larger modifying effect on the base selectivity of the BaP relative to the BcPh DEs.

Introduction Polycyclic aromatic hydrocarbons are common environmental pollutants, many of which are carcinogenic or mutagenic. Their oxidative metabolism by the cytochrome P450 system and microsomal epoxide hydrolase results in the formation of bay- and/or fjord-region diol epoxides (DEs)1 on an angular benzo ring of the hydrocarbon (1, 2); for those hydrocarbons that are tumorigenic, such DEs have been shown to be responsible for the tumorigenic response (3). Four optical isomers of a given DE are metabolically formed in mammals (Figure 1): two enantiomers in which the benzylic hydroxyl group and the epoxide oxygen are cis to each other (DE-1 or syn) and two in which these groups are trans (DE-2 or anti). An early event in cell transformation induced by DEs is * To whom correspondence should be addressed. Phone: (301) 4969893. Fax: (301) 402-0008. E-mail: [email protected]. 1 Abbreviations: BaP, benzo[a]pyrene; BcPh, benzo[c]phenanthrene; dGMP, deoxyguanosine 5′-monophosphate; dAMP, deoxyadenosine 5′monophosphate; dG, deoxyguanosine; dA, deoxyadenosine; DE-1, diol epoxide-1, in which the benzylic hydroxyl group and epoxide oxygen are cis; DE-2, diol epoxide-2, in which the benzylic hydroxyl group and epoxide oxygen are trans; DMSO, dimethyl sulfoxide; Tris, tris(hydroxymethyl)aminomethane.

10.1021/tx000073w

presumed to be the covalent modification of DNA by reaction with the benzylic carbon of the epoxide group. DEs react with DNA in vitro by a mechanism involving initial formation of a noncovalent DE-DNA complex (46), which undergoes epoxide ring opening at the benzylic position and partitioning to give tetraols (by DNAcatalyzed addition of water) and covalent DNA adducts (by addition of nucleophilic groups on the DNA). For DEs derived from a number of different hydrocarbons (7-16), the major DNA adducts in vitro have been identified as resulting from epoxide ring opening by the exocyclic amino groups of guanine and adenine (dG and dA adducts, Figure 1). Attack of the amino group occurs either from the same face of the ring as the epoxide oxygen (to give cis adducts) or from the opposite face (to give trans adducts). Adduct distribution (cis versus trans, dA versus dG) on reaction with DNA depends on both the hydrocarbon and the specific optical isomer of the DE. Notably, all but the (-)-(7S,8R,9R,10S) isomer of the four benzo[a]pyrene (BaP) DEs yield a preponderance of dG adducts (11, 17) [as much as 95% for the tumorigenic (+)(7R,8S,9S,10R) isomer], whereas the benzo[c]phenanthrene (BcPh) DEs exhibit a much greater preference for reaction at dA (18).

This article not subject to U.S. Copyright. Published 2000 by the American Chemical Society Published on Web 08/16/2000

884

Chem. Res. Toxicol., Vol. 13, No. 9, 2000

Figure 1. Structures of the hydrocarbons, their optically active bay- and fjord-region DEs, and adducts of the DEs at the exocyclic amino groups of purines in DNA. Heavy lines indicate the angular benzo rings where metabolism to the DEs occurs. Absolute configurations are designated starting from the benzylic hydroxyl-bearing carbon and progressing to the benzylic epoxide carbon. Note that the signs of rotation are reversed for the BaP and BcPh DE-2 enantiomers. Eight dA and eight dG adducts are possible from cis and trans opening of the epoxide ring of the four optically active DEs. Representative adduct structures are shown; absolute configuration is not implied.

The observed differences in adduct distribution upon reaction with DNA could result either from intrinsic differences in the selectivity of DEs from different hydrocarbons for reaction at dA versus dG or from specific effects of the DNA structure, for example, a preference of BcPh DEs for noncovalent complexation with DNA at dA versus dG sites in an orientation that is productive for adduct formation. Notably, the BcPh DEs are much more efficient in their covalent binding to DNA relative to DNA-catalyzed tetraol formation than are the BaP DEs (15, 19). Noncovalent interactions with bases in native DNA could occur either by intercalation into a native DNA structure or by looping out of a base to interact with the hydrocarbon. On the basis of an analysis of nearest neighbor effects on adduct formation (20), it has been suggested that intercalation is significant in the transition state(s) for adduct formation from double-stranded DNA. Some indication that DNA structure has a significant influence is provided by the observation of changes in adduct distribution upon denaturation of DNA (11, 21). Since even single-stranded DNA may retain significant structure, we determined the reactivities of DEs of BaP and BcPh with the free nucleotides, where no structure other than that provided by interactions between the reactant base and the hydrocarbon is possible.

Materials and Methods General Methods. Caution: BaP and BcPh DEs are carcinogenic and mutagenic and must be handled carefully in accordance with NIH guidelines (22). Reactions of DEs with deoxynucleoside 5′-monophosphates were carried out at 37 °C in 10 mM Tris-HCl buffer (pH 7.4), containing either dAMP (50

Vepachedu et al. mM), dGMP (50 mM) (as the sodium salts), or a solution of 50 mM dAMP and 50 mM dGMP (dAMP/dGMP mixture). Solutions of the optically active DEs, synthesized as described previously (23, 24), were prepared in a 1:9 DMSO/acetonitrile mixture. All other reagents were from commercial sources and were used without further purification. Reactions were initiated by addition of either DE solution (see below for details of individual experiments) to the aqueous nucleotide solutions such that the final ratio of organic solvent to water was 1:5 (1:10 for the BaP DEs); see below. Adduct Distribution upon Reaction of BcPh Diol Epoxides with a dAMP/dGMP Mixture. To 0.5 mL portions of the mixed nucleotide solution were added solutions (2 mg/ mL) of the individual optically active DEs (0.1 mL), and reaction was allowed to proceed at 37 °C for 20 h, at the end of which no residual DE was present as indicated by the absence of a thioether product upon reaction (25) with 2-mercaptoethanol. To remove tetraol hydrolysis products, the reaction mixtures were then extracted twice with ethyl acetate (1 mL) and once with ether (1 mL). Traces of solvent were removed by purging with nitrogen. Each aqueous reaction mixture was adsorbed on a Sep-Pak cartridge (Waters Associates), which was washed with 5 mL of water to remove unadducted nucleotides and eluted with 5 mL of a methanol/water mixture (60:40). The methanol/ water eluates were vacuum-dried, and the residue was dissolved in 300 µL of Tris-HCl buffer (pH 9) containing 10 mM MgCl2. The mixture was incubated for 3 h at 37 °C with 2.7 units of Escherichia coli alkaline phosphatase (Sigma Chemical Co.). Following enzymatic hydrolysis, 25 µL of each sample was analyzed by HPLC on a Beckman Ultrasphere ODS column (4.6 mm × 250 mm, eluted at 1.5 mL/min with a gradient from 85:5:10 to 70:10:20 water/methanol/acetonitrile over the course of 10 min, followed by isocratic elution at the final solvent composition; detection at 250 nm). Adducts were identified by comparison of their chromatographic retention times with those obtained upon reaction of each BcPh DE with the individual nucleotides rather than with the dAMP/dGMP mixture. Adduct Distribution upon Reaction of BaP Diol Epoxides with the dAMP/dGMP Mixture. Reaction of the DEs with the dAMP/dGMP solution was as described for the BcPh DEs, except that 50 µL of each DE solution (1 mg/mL) was added to 0.5 mL of the aqueous dAMP/dGMP mixture, and reaction was allowed to proceed for 1 h. Sep-Pak separation of the unadducted nucleotides was carried out immediately after completion of the reaction; the methanol/water eluates were vacuum-dried, and the residues were dissolved in 0.5 mL of TrisHCl buffer and extracted with water-saturated ether for tetraol removal prior to the enzymatic hydrolysis step. Products were analyzed (11) by HPLC on the Beckman Ultrasphere column eluted at 1.0 mL/min with 50% methanol in water for 10 min, followed by a linear gradient that increased the methanol composition by 1% per minute for 35 min; detection was at 344 nm. Partitioning between Adduct and Tetraol Formation. The extent of dA/dG adduct formation from each DE, relative to the extent of nucleotide-catalyzed hydrolysis to tetraols, was determined from the difference between the tetraols formed upon reaction with the dAMP/dGMP mixture (complete after 1 and 20 h at 37 °C for the BaP and BcPh DEs, respectively) and upon hydrolysis of the DEs in 0.1 M perchloric acid. To ensure reliable HPLC quantitation of the tetraols, 4-(p-nitrophenyl)1-butanol was used as an internal standard. For the BcPh reaction mixtures, retention times (minutes) for the tetraols in the water/methanol/acetonitrile solvent system were as follows: from BcPh DE-1, 10.9 and 13.3; and from DE-2, 12.4. Tetraols were quantitated by absorbance at 255 nm relative to the internal standard (tR ) 21 min). A similar procedure was used to quantitate the tetraols from the BaP DEs. For HPLC analysis, a VYDAC C18 column (4.6 mm × 250 mm) was eluted with the same solvent system. Retention times (minutes) for the tetraols on this column were as follows: from BaP DE-1, 11.8 and 12.7; and from DE-2, 10.7 and 13.0. Tetraols were quantitated by absorbance at 325 nm relative to the internal

Diol Epoxide Adducts from DNA and Nucleotides standard (tR ) 12.6 min). Correction was made for small amounts of tetraols present at time zero in the optically active BaP DE samples, as determined by treating (25) a 20 µL portion of each BaP DE solution with 50 µL of 0.2 M 2-mercaptoethanol (50% as the thiolate anion) followed by HPLC; thr tR of the thiolate adduct from BaP DE-1 was 14.9 min and that from BaP DE-2 13.8 min. Under our experimental conditions, the yield of all types of adducts as determined by difference relative to tetraols ranged from ∼30 to ∼60% of each DE, and was slightly higher for the BaP relative than for the BcPh DEs, possibly because BaP DEs react with the 5′-phosphate groups as well as with the exocyclic amino groups of the purine nucleotides (see below). Effect of Nucleotide Concentration on Adduct Yields and dA:dG Ratios. Racemic BaP DE-2 was allowed to react as described for 20 h with three dAMP/dGMP mixtures in which each nucleotide was at a concentration of 25, 50, and 100 mM (combined nucleotide concentrations of 50, 100, and 200 mM, respectively). Following workup and enzymatic digestion, identical 50 µL samples of each mixture of digestion products were subjected to HPLC analysis on a ZORBAX Eclipse C18 column (4.6 mm × 250 mm), eluted with 45% methanol in water for 10 min, followed by a linear gradient that increased the methanol composition by 1% per minute. Relative adduct yields and dA: dG adduct ratios were determined by integration of the UV signal at 343 nm. The same experiment was also carried out with (-)-BcPh DE-2. Note that the lowest adduct concentration was half that used under the standard conditions (see above), and consequently, the reaction time was increased to 40 h. The enzymatic hydrolysis products (BcPh DE-2 dA and dG adducts) were analyzed by HPLC on the Eclipse column using 40% methanol for the isocratic segment of the elution, followed by the same rate of increase in the methanol composition as described above. Quantitation of BcPh adducts was carried out at 250 nm. Adduct Formation Relative to Hydrolysis upon Reaction of BcPh DE-2 with a Dilute Nucleotide Mixture. A solution containing the four nucleotides (dGMP, dAMP, dCMP, and TMP), in which each nucleotide was at a concentration of 0.7 mM, was prepared in 10 mM Tris-HCl buffer at pH 8.0. To 1.0 mL of this mixture was added 0.1 mL of a solution of (()BcPh DE-2 in acetone (1 mg/mL), containing ∼0.5 mg/mL p-nitrophenylbutanol as an internal standard. Reaction progress was monitored by trapping of unreacted DE as the 2-mercaptoethanol adduct. Typically, 20 µL of the reaction mixture was treated with 100 µL of the 2-mercaptoethanol reagent described above, incubated at 37 °C for 10 min, and neutralized with 100 µL of 0.1 M HClO4. Samples were analyzed by HPLC on the VYDAC C18 column as described previously (tR of the 2-mercaptoethanol adduct, 8.7 min). The purity of the (()-DE sample was verified by the absence of tetraol upon chromatography after treatment with thiol at time zero. The reaction was allowed to proceed to completion (10 days) at 37 °C. The trans tetraol product was quantitated by integration at 280 nm relative to the internal standard, and compared with the amount of tetraol formed from an identical sample of the DE on hydrolysis in 0.1 M HClO4.

Results and Discussion Substantial information about the reactivity and product compositions obtained upon reaction of DEs with calf thymus DNA under a standard set of conditions is available (7, 16, 19), normally, ∼0.8-1.0 mg/mL DNA in 1:10 organic/aqueous solutions at neutral pH and 37 °C. For example, at a DNA concentration of 0.8 mg/mL, reactions of the isomeric BcPh DEs occur exclusively via DNA catalysis, with a half-life on the order of 5-7 min at neutral pH, and yield covalent DNA adducts corresponding to 55-77% of the reactant DEs (19). In the study presented here, we confirmed (data not shown) that

Chem. Res. Toxicol., Vol. 13, No. 9, 2000 885

the reaction of (()-BcPh DE-2 with ∼1 mg/mL DNA is complete within 1 h. To examine the products from a BcPh DE with a concentration of free nucleotides comparable to that in 0.8 mg/mL DNA (approximately 2.5 mM total bases), an equimolar mixture containing each of the four deoxynucleotides (at 0.7 mM) was prepared. A pH of 8 was chosen to maintain the nucleotides predominantly in the unprotonated, dianion form, since DNA is not appreciably protonated at neutrality. Under these conditions, the disappearance of (()-BcPh DE-2 required 10 days for completion, and within the limits of detectability, all of the DE underwent hydrolysis to tetraol. Thus, relative to similar concentrations of free nucleotides, the structure of DNA and its ability to form a noncovalent complex with the hydrocarbon greatly enhance both the overall reactivity of the DEs and their preference for covalent adduct formation. This result also meant that it was not possible to examine adduct distributions at nucleotide concentrations comparable to those present in DNA under our standard conditions for DNA binding studies, and that much higher concentrations of the free nucleotides were required. The standard conditions for DNA binding utilized 10 mM Tris-HCl buffer, and consequently, the same buffer was used in this study. Although it is known that high chloride ion concentrations can affect the distribution of trans versus cis BaP DE adducts from both DNA (26) and nucleosides (27) as well as reaction rates and the distribution of cis and trans tetraol products in solvolysis reactions (28), due to formation of a chlorohydrin intermediate, this chlorohydrin pathway was expected to constitute only a minor fraction (e10-12%) of the total reaction under our conditions due to the small amount of chloride ion present. Formation of Adducts and Tetraols on Reactions with DNA and Nucleotides. Nucleotides, like DNA, catalyze the ring opening of BaP DEs (29, 30). Rate accelerations observed for general acid-catalyzed ring opening of BaP DE-2 by 5′-AMP and 5′-GMP (∼30-fold relative to that of ribose 5′-phosphate) indicate that charge transfer or hydrophobic interactions between the hydrocarbon and the base must be involved in the reaction of purine nucleotides with polycyclic aromatic hydrocarbon DEs in aqueous solution. The second-order rate constant for 5′-GMP is ∼1.5 times greater than for 5′-AMP, which suggests that interactions of the pyrene moiety with guanine are slightly favored relative to those with adenine in the transition state for ring opening. At pH 7 and nucleotide concentrations of >3 mM, the observed rate of reaction of BaP DE-2 is estimated to be almost exclusively (>90%) due to a nucleotide-catalyzed mechanism, on the basis of known rate constants for the reaction of this DE with the ribonucleotides AMP and GMP (30) as compared to that for hydrolysis in water. Although there have been no detailed kinetic studies of nucleotide-catalyzed reactions of the BcPh DEs, the starting DEs are completely consumed within 24 h at pH 7.4 and 37 °C in the presence of dAMP and dGMP at a total nucleotide concentration of 100 mM (this study). At the same pH in the absence of nucleotides, the half-life of the DEs (25) is on the order of 19-22 h (25 °C, 1:9 dioxane/water mixture). This indicates that nucleotide catalysis is also a major reaction route for the BcPh DEs under our experimental conditions described here. At low nucleotide concentrations, tetraols are the predominant products formed from BaP DE-2, and the

886

Chem. Res. Toxicol., Vol. 13, No. 9, 2000

rate is almost exclusively due to general acid catalysis of hydrolysis by the nucleotide monoanion. For example (30), at neutral pH in the presence of 5 mM GMP monoanion, the rate due to nucleotide catalysis is several hundred times greater than the rate for uncatalyzed and hydronium ion-catalyzed hydrolysis, although >94% of the observed products are tetraols. If the nucleotide catalyzes tetraol formation in a general acid-catalyzed reaction without formation of a carbocation intermediate, increasing its concentration from 5 to 50 mM would be expected to increase the rates of both tetraol and adduct formation proportionally (since both rates depend on nucleotide concentration), and therefore, nucleotide concentration should not affect the product ratio. Thus, tetraol might simplistically have been expected to predominate at high nucleotide concentrations also. Instead, at the nucleotide concentrations used in this study, we observe that a much larger proportion (∼40-60%) of the DE is converted to covalent nucleotide adducts. A possible explanation for the observed effect of nucleotide concentration on product distribution involves a stepwise reaction in which a carbocation intermediate is trapped by either water or the nucleotide after the ratedetermining step. In this hypothetical mechanism, nucleotide-catalyzed ring opening to the carbocation is ratedetermining, leading to the observed steep dependence of the rate on nucleotide concentration even at very low nucleotide concentrations. At low nucleotide concentrations, the subsequent, product-determining step consists predominantly of uncatalyzed trapping of the cation by water, leading to the observed (30) preponderant tetraol formation. At the nucleotide concentrations (g50 mM) used in the study presented here, adduct formation and nucleotide-catalyzed attack of water become the major (and approximately equally significant) paths for trapping of the cation. Consequently, under these conditions, both tetraol and adducts are observed, and since both have identical dependencies on nucleotide concentration, product ratios are relatively insensitive to further increases in nucleotide concentration. An alternative mechanistic interpretation is that at low nucleotide concentrations nucleotide-catalyzed hydrolysis of the free DE to tetraols by a general acid mechanism predominates. As the nucleotide concentration is increased, a different mechanism involving reaction via a noncovalent complex between the DE and the nucleotide may become significant. Equilibrium constants of ∼73 and ∼55 M-1 for ground-state association in water between the purine nucleosides dG and dA, respectively, and a model compound, the trans-opened tetraol derived from BaP DE-2, have been measured by UV spectrophotometric titration (31). On the basis of these equilibrium constants, up to 70-80% of the DE could be in the complexed form at 50 mM nucleotide. We suggest that, as in the case of DNA (7), both hydrolysis to tetraols and covalent adduct formation could occur within a DEnucleotide complex that predominates at equilibrium under the conditions described here. This interpretation is consistent with the observation that the adduct versus tetraol distribution from (()-BaP DE-2 remained relatively constant (