Mechanistic Insights into the Michael Addition of Deoxyguanosine to

Jun 12, 2008 - Dedicated to the occasion of Professor James K. Woodʼs retirement and his 39 years of teaching. , * To whom correspondence should be ...
2 downloads 0 Views 995KB Size
Download PDF
Chem. Res. Toxicol. 2008, 21, 1415–1425

1415

Mechanistic Insights into the Michael Addition of Deoxyguanosine to Catechol Estrogen-3,4-quinones† Douglas E. Stack,* Guangping Li, Anastacia Hill, and Nicholas Hoffman Department of Chemistry, UniVersity of Nebraska at Omaha, 6001 Dodge Street, Omaha, Nebraska 68182-0109 ReceiVed February 23, 2008

The reaction of catechol estrogen quinones with DNA to produce the depurinating adducts, 4-OHE2(E1)1-N7Gua and 4-OHE2(E1)-1-N3Ade, has been linked to the initiation of breast and other human cancers. A better understanding into the mechanism of how these adducts are formed would be useful to studies aimed at correlating adduct formation to DNA damage. Possible reaction intermediates, produced as a result of Michael addition of deoxyguanosine (dG) to catechol estrogen-3,4-quninones, have been modeled using density functional theory to determine likely intermediates on the potential energy surface (PES) of this reaction. Specifically, the sequence of elimination events, glycosidic bond cleavage and rearomatization of the estrogen A ring, was explored. Consistent with known experimental procedures, B3LYP calculations indicate that a proton source is needed to effect the Michael addition. Calculations also indicate that a catalytic mechanism, where one catechol estrogen quinone could adduct multiple purine bases, is unlikely. Experimental investigation toward an observed cationic reaction intermediate was also consistent with a stoichiometric reaction between estrone-3,4-quinone (E1-3,4-Q) and dG. HPLCMS analysis indicates that the cationic reaction intermediate contains the 2′-deoxyribose moiety. Assay of 4-OHE1-1-N7Gua adduct formation and 2′-deoxyribose formation at different times during the reaction of E1-3,4-Q with dG indicates that equimolar amounts of each are produced, further supporting a stoichiometric process with respect the catechol estrogen quinone. Differences in the UV spectroscopy of cationic reaction intermediate and the 4-OHE1-1-N7Gua adduct allowed for kinetic analysis of the glycosidic bond cleavage process. Kinetic scanning analysis indicates that the decomposition of the cationic reaction intermediate is a first-order process with a t1/2 of 40 min at 30 °C. Measurement of the unimolecular rate constant k at different temperatures afforded an Arrhenius plot, which provided values for ∆H+, ∆S+, and ∆G+ of 24.7 kcal/mol, 7.2 eu, and 26.8 kcal/mol, respectively. The computational data in conjunction with experimental results are consistent with a mechanism that involves a proton-assisted Michael addition to form an R-ketoenol ring system, followed by slow loss of the proton at C1 to restore the aromatic A ring, then fast cleavage of the glycosidic bond to form the 4-OHE1-1-N7Gua adduct. Introduction 1

Oxidative metabolites of β-estradiol (E2) and estrone (E1) have been implicated as possible genotoxic agents capable of initiating breast and other human cancers (1, 4). The predominate metabolites of both E2 and E1 are 2-hydroxyestradiol(estrone) [2-OHE2(E1)] and 4-hydroxyestrdadiol(estrone) [4-OHE2(E1)] formed by hydroxylation of the estrogen A ring (Figure 1) (5, 7). 2-OHE2(E1) is the predominate isomer; however, the 4-OHE2(E1) isomer is expressed at abnormally high levels in or near breast tumors in humans (8, 11). In addition, the 4-OHE2(E1) catechols have been shown to be carcinogenic in animal models, whereas the 2-OHE2(E1) metabolites are not (2, 12, 13). The 4-OHE2(E1) catechol is thought to be a procarcinogen that is further oxidized to electrophilic o-quinones capable of binding to DNA. The chemistry of the two isomeric † Dedicated to the occasion of Professor James K. Wood’s retirement and his 39 years of teaching. * To whom correspondence should be addressed. E-mail: dstack@ unomaha.edu. 1 Abbreviations: Ade, adenine; dA, deoxyadenosine; dG, deoxyguanosine; E2, β-estradiol; E2-2,3-Q, estradiol-2,3-quinone; E2-3,4-Q, estradiol-3,4quinone; E1, estrone; E1-2,3-Q, estrone-2,3-quinone; E1-3,4-Q, estrone-3,4quinone; Gua, guanine; 2-OHE2, 2-hydroxyestradiol; 2-OHE1, 2-hydroxyestrone; 4-OHE2, 4-hydroxyestradiol; 4-OHE1, 4-hydroxyestrone; DMF, N,N-dimethylformamide; IBX, o-iodoxybenzoic acid; PES, potential energy surface; SPE, single point energy.

o-quinones, β-estradiol(estrone)-2,3-quinones [E2(E1)-2,3-Q] and β-estradiol(estrone)-3,4-quinones [E2(E1)-3,4-Q], is quite different with respect to their reaction with nucleosides and DNA (1, 14). The catechol estrogen-2,3-quinones produce stable DNA adducts that remain bound to the DNA polymer via linkage to the exocyclic amino groups of guanine (Gua) and adenine (Ade). In contrast, the catechol estrogen-3,4-quinones generate depurinating adducts, adducts that detach from the DNA due to cleavage of the glycosidic bond via reaction at the N7-position of Gua (14) or the N3-position of Ade (4) (Figure 2). The formation of a depurinating adduct results in an apurinic site in the DNA. Copious formation of apurinic sites can lead to misrepair of the DNA and to mutations that initiate the cancer process. Depurinating hydrocarbon-DNA adducts have been linked to mutations in the Harvey-ras oncogene isolated from mouse skin papillomas induced by polyaromatic hydrocarbons (15, 16). There is increased interest in correlating the production of depurinating adducts, formed by reaction of catechol estrogen3,4-quinones with DNA, and the occurrence of tumors in both animal and human studies (4, 10, 17, 18). To correlate the production of estrogen-DNA adducts with the formation of apurinic sites in the DNA, a basic understanding of the mechanism involving the reaction of catechol estrogen-3,4quinones with deoxynucleosides is needed. E2(E1)-3,4-Q reacts regioselectively with dG at the N7-position of Gua and the C1

10.1021/tx800071u CCC: $40.75  2008 American Chemical Society Published on Web 06/12/2008

1416

Chem. Res. Toxicol., Vol. 21, No. 7, 2008

Stack et al.

Figure 1. Metabolism of E2 and/or E1 to catechol estrogens and catechol estrogen quinones.

Figure 2. Reaction of catechol estrogen quinones with dG, dA, and Ade.

carbon of the o-quinone to produce the 4-OHE2(E1)-1-N7Gua adduct (Figure 2) (14). We chose this model system to investigate since reaction of deoxyadenosine (dA) with catechol estrogen-3,4-quinones does not generate adducts under similar conditions. The formation of the standard 4-OHE2(E1)-1-N3Ade adduct requires reaction of Ade with catechol estrogen-3,4quinones (4). Thus, the deoxyguanosine (dG) system could provide information with regard to both the nature and the timing of glycosidic bond cleavage in reactions that form estrogen-DNA depurinating adducts, whereas dA cannot. Presumably, the presence of the sugar moiety sterically restricts reaction at the N3 site under conditions shown in Figure 2. Reaction of catechol estrogen 3,4-quinones with DNA in vitro (4) or in vivo (4, 17, 19) do produce the 4-OHE2(E1)-1-N3Ade adducts, indicating that steric accessibility to the N3 site may be less hindered in the DNA polymer as compared to solutions of dA. In B-DNA, the Ade base adopts an anti conformation with respect to the ribose polymer exposing the N3-position in the minor grove (20). Several other electrophiles, differing in electronic properties and the manner in which they are generated, also produce N3 adducts with DNA but do not react with dA (21, 24). The formation of 4-OHE2(E1)-1-N7Gua from catechol estrogen3,4-quinones requires an acidic medium; a solvent mixture of 50:25:25 acetonitrile:H2O:acetic acid, respectively, is typical (14). Other organic solvents, such as N,N-dimethylformamide (DMF), can be used in lieu of acetonitrile, but all test tube reactions producing the Gua or Ade depurinating adducts require a proton source. Scheme 1 illustrates possible mechanistic steps

in producing the 4-OHE2(E1)-1-N7Gua adduct from catechol estrogen-3,4-quinone’s reaction with deoxyguanoisne. Michael addition at the C1 carbon, followed by proton transfer, would yield intermediate 2. The fate of intermediate 2 is key in determining whether the reaction is stoichiometric with respect to the catechol estrogen-3,4-quinone or whether a catalytic cycle could produce several apurinic sites from the production of one o-quinone. Proton loss, Ha, from C1 would lead to rearomatization of the estrogen A ring generating 3 (path A: 2 f 3 f 5 f 7). Conversely, cleavage of the glycosidic bond on 2 would produce 4, which could rearomatize (path B: 2 f 4 f 6 f 7), or, following the loss of proton Hb, undergo a reverse Michael reaction to regenerate the o-quinone (path C: 2 f 4 f 8 f o-quinone). Path C would represent a catalytic cycle whereby one catechol estrogen-3,4-quinone could generate multiple apurinic sites. Conversely, paths A and B would be stoichiometric with respect to the catechol estrogen-3,4-quinone, and one apurinic site would result from the formation of one depurinating adduct. To gain insight into the thermodynamic differences among paths A, B, and C, we have sampled the potential energy surface (PES) leading to the intermediates shown in Scheme 1 using density functional theory (DFT) calculations of a representative set of model compounds. In addition, experimental studies identifying a key intermediate, referred to as “intermediate A”, were preformed to supplement the theoretical calculations. The combination of these two studies will be discussed in terms of which mechanism(s) is(are) likely to prevail and the implications that these mechanistic insights

Michael Addition of dG to Catechol Estrogen-3,4-quinones

Chem. Res. Toxicol., Vol. 21, No. 7, 2008 1417

Scheme 1. Possible Mechanisms for the Michael Addition of dG to Estradiol-3,4-quinone (E2-3,4-Q)

provide for assays based on using depurinating adducts as biomarkers for estrogen-induced cancers.

Experimental Procedures Chemicals. E1 and E2 were purchased from Steraloids, Inc. (Newport, RI), and dG was purchased from TCI America (Portland, OR). All other chemicals and solvents were purchased from Fisher Scientific Co. (Fair Lawn, NJ) or Aldrich Chemical Co. (Milwaukee, WI). Instrumentation. HPLC was conducted on a Waters 2690 Separations Module equipped with a Waters 2487 Dual λ Absorbance Detector. GC/MS was performed on a Varian Saturn 2000 GC/MS/MS ion trap attached to a Varian 3800 Gas Chromatograph interfaced to a PC running Saturn software. UV spectra were obtained on an Varian, Cary 100 Bio UV-visible spectrophotometer using a dual cell peltier temperature controller. Synthesis of 4-Hydroxyestrone (4-OHE1) and 4-OHE2. 4-OHE1(E2) was synthesized by o-iodoxybenzoic acid (IBX) oxidation of E1 or E2 as described previously (25) except that

reactions were run on a 1 g scale of E1 or E2. The compounds were purified from the 2-OH isomers on a Biotage Flash40i separations module (Biotage, Charlottesville, VA) using a 40 M (90 g) silica gel cartridge employing a solvent pair of hexane and ethyl acetate, 75:25, respectively, containing 1% acetic acid. Reaction of Estrone-3,4-quinone (E1-3,4-Q) with dG and Isolation of Intermediate A. 4-OHE1, 50 mg, was dissolved in 4 mL of acetonitrile and cooled to 0 °C. Activated MnO2, 100 mg, was added, and the mixture was stirred for 15 min to convert the catechol to the o-quinone. The suspension was filtered directly into a solution containing 250 mg of dG in 4 mL of water acetic acid (1:1) held at 0 °C. It was found that conducting the reaction at 0 °C allowed maximum build up of the thermally liable, unknown reaction intermediate A. The reaction was stirred at 0 °C for 4 h, and a 200 uL aliquot was removed, filtered, and separated by HPLC using a YMC, ODS-AQ C18 semipreparative column (5 µm, 120 Å, 250 mm × 10 mm, Waters Corp.). A gradient solvent system starting at 43% methanol and 57% water containing 0.5% acetic acid, at a flow rate of 3 mL/min, was held for 5 min and then linearly increased to 100% methanol at 20 min. The intermediate

1418

Chem. Res. Toxicol., Vol. 21, No. 7, 2008

Stack et al.

Figure 3. Conversion of starting materials to 2m.

A was collected as it eluted from the HPLC in a 4 mL amber vial and immediately immersed in liquid nitrogen. The compound was then stored at -78 °C until needed for UV analysis. A sample of this intermediate was also analyzed at the Nebraska Center for Mass Spectrometry (University of NebraskasLincoln) via HPLC/MS employing a time-of-flight detector (Waters Q-TOF I). Assay of 2′-Deoxyribose and the 4-OHE1-1-N7Gua Adduct. The reaction of E1-3,4-Q with dG was conducted as described above except that reaction with dG was done at room temperature. Mixtures of 4-OHE1-1-N7Gua (obtain by the same HPLC, semipreparative, gradient separation described above) and 4-nitroaniline were analyzed to obtain response factors between the two compounds at 290 nm. Reaction aliquots were quenched by mixing with 200 mg of N-(2-mercaptoethyl)aminomethyl polystyrene resin (NovaBiochem, Gibbstown, NJ), in 1 mL of DMF, to sequester unreacted o-quninone, which reacts rapidly with thiol nucleophiles. The addition of the internal standard was followed by HPLC separation using an analytical column, Kromasil C18 (5 µm, 120 Å, 250 mm × 10 mm, Supelco, Inc.). A gradient of 30% methanol in 5% acetic acid, 1 mL/min for 5 min, and then a linear gradient to 100% methanol at 20 min provided separation of the internal standard and the 4-OHE1-1-N7Gua adduct. At the same time, HPLC aliquots were removed from the reaction, and a separate aliquot for 2′-deoxyribose analysis was obtained and analyzed via GC/ MS using the derivation procedure of Guerrant (26) using Darabinonse as an internal standard. Computational Methods. All density functional calculations were done using Gaussian 03 software (27). Conformational analysis of modeled compounds was done with Spartan 5.0 software using the Osawa algorithm (28). Low-energy candidates were then optimized in the gas phase using DFT with the Becke threeparameter exchange and Lee-Yang-Parr correlation functional (B3LYP) (29, 30) with a 6-31G(d) basis set. The molecular entropies and thermal corrections to the free energies were calculated from the harmonic vibrational constants, scaled to 0.9804, determined by analytic energy second derivatives calculated at the gas phase-optimized structures. Further refinement of electronic energies was obtained by submitting the B3LYP/6-31G(d) geometries to a single point energy (SPE) calculation with a 6-31+G(2d,p) basis set. To calculate the free energy of aqueous solvation, SPE calculations [B3LYP/6-31+G(2d,p)] were conducted using a polarizable continuum model (31, 32) and employed the SCFVAC keyword in Gaussian03 using the United Atom Topological Model for atomic radii. Structures of modeled compounds, in the MDL.mol format, are available as Supporting Information.

Results Theoretical Calculations on Possible Reaction Intermediates. To make high level ab initio calculations computationally

feasible, abbreviations to the structures in Scheme 1 were employed. These abbreviations resulted in the replacement of the C and D rings of all estrogen structures with trans-methyl groups placed on the 8- and 9-positions to mimic the trans ring junction of the B ring to the C ring (Figure 3). In addition, the deoxyribose moiety was simplified to a 2′-tetrahydrofuranyl ring system. Because the Michael addition of dG to the catechol estrogen-3,4-quinones involves only the A (and to some degree the B) ring, this abbreviation seemed reasonable. Prior calculations to the overall thermodynamic change in formation of the 4-OHE2(E1)-1-N7Gua adducts from E2(E1)-3,4-Q and Gua showed little difference between 4-OHE2 and 4-OHE1 (33). The modeled structures in Table 1 correspond to the structures in Scheme 1 by adding the suffix -m to the structure’s corresponding number in Scheme 1 (m stands for modeled structure). Even with these abbreviations, the inclusion of kinetic data by searching for transition state structures is not computationally feasible for all steps. While only exploring thermodynamic differences between intermediates at the exclusion of kinetic barriers is precarious at best, many of the mechanistic branches in Scheme 1 involve intermediates that are tautomeric, and tautomeric equilibriums in acidic medium are generally under thermodynamic control. The computational results presented here are mainly to investigate which intermediates are reasonable structures on the PES associated with this Michael addition. Scheme 1 indicates a need for both a proton source and a base for several of the hypothesized steps. We chose to use an acetic acid-acetate pair for these proton transfer reactions since acetic acid is present in the actual reaction shown in Figure 2. Figure 3 shows the energy changes, both gas phase and in solution, for the formation of 2-m from the starting quinone-m and dG-m. The first observation of note is that the canonical structure of 1 shown in Scheme 1 is not a likely intermediate along the PES. When one submits the Lewis structure of 1-m to density functional calculations starting with a typical C-N bond length, the geometry optimization finds a spot on the PES resembling a complex between quinone-m and dG-m. The lack of bonding between the N7 nitrogen of dG-m and the C1 position of quinone-m, 3.99 Å, indicates that 1-m is better represented by the canonical form shown on the right in Figure 3. Also, the smaller endergonic energy change in the gas phase, +5.67 kcal/mol, as compared to calculations involving solvent effects, +15.64 kcal/mol, also supports the less polar canonical form on the right. The lack of bond formation between neutral species is consistent with the observed experimental requirement

Michael Addition of dG to Catechol Estrogen-3,4-quinones

Chem. Res. Toxicol., Vol. 21, No. 7, 2008 1419

Table 1. Thermodynamic Data for Modeled Compounds: B3LYP/6-31+G(2d,p)//B3LYP/6-31G (d)* Theorya structure

SCF (Hartrees)b

free energy correction (Hartrees)c

Ggas (Hartrees)

∆Gsolvation (kcal/mol)d

Gaq (Hartrees)

quinone-m dG-m Gua acetic acid acetate 1-m 2-m 2-m-keto 3-m 4-m 5-m 6 7-m 8-m 2-AcTHF

-616.185291 -773.880770 -542.605548 -229.116168 -228.552552 -1390.072750 -1390.476573 -1390.465085 -1390.052511 -1158.812376 -1390.498661 -1158.291020 -1158.825418 -1158.250817 -460.379224

0.191689 0.170355 0.081716 0.033500 0.019728 0.377773 0.399744 0.407666 0.386719 0.299107 0.398755 0.285478 0.297844 0.284364 0.122408

-615.993602 -773.710414 -542.523832 -229.082668 -228.532824 -1389.694976 -1390.076828 -1390.057419 -1389.665791 -1158.513268 -1390.099905 -1158.005541 -1158.527573 -1157.966452 -460.256815

-7.49 -23.9 -27.14 -7.39 -70.32 -21.42 -54.66 -58.40 -36.5 -18.6 -54.8 -63.12 -21.57 -67.38 -8.92

-616.005539 -773.748503 -542.567083 -229.094445 -228.644888 -1389.729112 -1390.163936 -1390.150487 -1389.723959 -1158.542910 -1390.187237 -1158.106131 -1158.561948 -1158.073831 -460.271031

C1′-N9 distance (Å)

N7 chargee

C1′ chargee

1.4822

-0.564 -0.518

0.251

1.4809 1.5011

-0.495 -0.324

0.250 0.242

1.4799

-0.303 -0.373 -0.321 -0.360 -0.374 -0.365

0.245

1.4995

0.242

a An expanded version of Table 1, which includes enthalpy correction, is available as Supporting Information. b Gas phase calculations done at B3LYP/6-31+G(2d,p)//B3LYP/6-31G(d). c Obtained from B3LYP/6-31G(d) frequency calculations on optimized geometries. d SCRF SPE calculations using gas-phase geometries; see the Experimental Procedures for further details. e Natural population analysis from SCRF calculations.

Figure 4. Conversion of 2-m to 3-m or 4-m: depurination vs rearomatization.

of acidic conditions for this Michael addition. Evidently, if a proton source is not present to sequester the accumulation of negative charge on O3, bond formation between the N7 of dG and the C1 of the quinone will not occur. This initial finding is what prompted our investigation to a possible catalytic cycle involving regeneration of the o-quinone. If an intermediate similar to 2-m were to lose a proton at O3, a reverse Michael addition could be possible. The role of the proton source and proton transfer in this first step will be expanded upon in the Discussion section. For 2-m to convert to 7-m (see unabbreviated structures in Scheme 1; the thermodynamic data for 7-m is included in Table 1), two events must occur, rearomatization of the A ring and glycosidic bond cleavage. The sequence of these events is crucial to determining whether a catalytic cycle or a stoichiometric process, with respect to the o-quinone, predominates. Thus, 2-m could first lose the proton at C1 to form the zwitterionic compound 3-m, or cleavage of the glycosidic bond could occur to form 4-m (Figure 4). Two stabilizing forces are in competi-

tion. The formation of 3-m leads to rearomatization of the A ring but produces a very polar intermediate. The formation of 4-m leads to a less polar compound but no aromatic stabilization. The polarity of the medium would be expected to affect the choice between the two processes. In the gas phase, 4-m is favored by 13.57 kcal/mol, while in solution 3-m is favored by 2.80 kcal/mol. The level of theory employed in these calculations, B3LYP/6-31+G(2d,p)//B3LYP/6-31G(d), has a standard deviation from experiment of 4.2 kcal/mol (34). Thus, the preference for 4-m over 3-m in an aqueous environment cannot be solely established by computation. Because density functional calculations could not rule out 4-m based solely on thermodynamic data, we sought to model the energy differences between the conversion of 4-m to 6-m vs the conversion of 4-m to 8-m (Figure 5). The former process leads to stoichiometric formation of 7-m, while the latter could result in a reverse Michael process. Figure 5 shows 6-m to be the more stable isomer under both gas-phase and aqueous

1420

Chem. Res. Toxicol., Vol. 21, No. 7, 2008

Stack et al.

Figure 5. Conversion of 4-m to 6-m or 8-m.

Figure 6. HPLC analysis of the reaction of E1-3,4-Q with dG.

calculations, -24.52 and -20.27 kcal/mol, respectively. These results are consistent with formation of the weaker phenoixde base, 6-m, as compared to the enolate anion, 8-m. A noteworthy observation in 8-m is the lengthening of the C1-N7 bond after

removal of the proton at O3 to 1.554 from 1.487 Å. While this C-N bond is weaker than the corresponding bond in either 4-m or 6-m, it still indicates significant bonding between the two atoms unlike structure 1-m (Figure 3).

Michael Addition of dG to Catechol Estrogen-3,4-quinones

Chem. Res. Toxicol., Vol. 21, No. 7, 2008 1421

Figure 7. HPLC/MS spectrum of intermediate A.

Experimental Characterization of a Reactive Intermediate. HPLC analysis of a typical reaction between E1-3,4-Q and dG under conditions described in Figure 2 had indicated the possible presence of a reaction intermediate. Figure 6 shows HPLC analysis of the reaction taken at 2, 4, and 24 h intervals at room temperature. The peak at 10.6 min appears early in the reaction and starts to diminish as the reaction progresses. Concurrent with the disappearance of this peak is the increase in the final product peak, 7, at 11.8 min. When the peak at 10.6 min is collected from the HPLC and allowed to stand at room temperature for 24 h and then reinjected into the HPLC, it was shown to convert to 7, indicating it as some type of reaction intermediate. We refer to this compound as “intermediate A”. Further structural characterization of intermediate A was obtained by HPLC/MS analysis employing time-of-flight detection. The base peak at 552.23 m/z indicates that A contains the ribose moiety (Figure 7). Discrimination between the analogous structures, 2, 5, or 2-keto, in Scheme 1 is not possible since these structures are tautomers. The peak at 436.19 m/z matches the known parent peak of the 4-OHE1-1-N7Gua adduct and is the result of ribose loss from A. Because A could be the early intermediate 2, the HPLC and MS data do not rule out the possibility of a catalytic mechanism. However, the conversion of A to product 7 does represent cleavage of the glycosidic bond, and the spectroscopic investigation of this key step will be presented below. Assay of 2′-Deoxyribose Formation and 4-OHE1-1-N7Gua Adduct Formation. To determine if the reaction of dG with E1-3,4-Q is stoichiometric or catalytic, we assayed the produc-

Table 2. Assay of Both 4-OHE1-1-N7Gua Adduct Formation and 2′-Deoxyribose Formation in the Reaction of dG with E1-3,4-Q time (h)

mol % 4OHE1-1-N7Guaa

mol % D-deoxyribosea

1 9 24

9 28 44

6 20 41

a

On the basis of the starting molar amount of E1-3,4-Q.

tion of the adduct product, 4-OHE1-1-N7Gua, and the production of 2′-deoxyribose from the same reaction. Equal amounts of both would be consistent with a stoichiometric mechanism. A catalytic mechanism with respect to the quinone should result in the production of more 2′-deoxyribose when compared to the 4-OHE1-1-N7Gua product. Table 2 shows the results of removing two aliquots from a room temperature reaction of dG with E1-3,4-Q at 1, 9, and 24 h. One aliquot was assayed for 4-OHE1-1-N7Gua production via HPLC analysis using 4-nitroaniline as an internal standard. The second aliquot assayed 2′-deoxyribose production by forming the acylated aldononitrile derivative followed by GC/MS analysis (26). This process employed D-arabinose as an internal standard. Table 2 indicates that the formation of 2′-deoxyribose and the 4-OHE1-1-N7Gua adduct occurred in approximately a 1:1 mol ratio. These data strongly suggest a stoichiometric mechanism, at least under the conditions shown in Figure 2, where one o-quinone produces one estrogen DNA adduct. UV Measurement of Glycosidic Bond Cleavage. The HPLC separation of unknown intermediate A provides UV spectrophotometric analysis of this intermediate. Figure 8 shows that

1422

Chem. Res. Toxicol., Vol. 21, No. 7, 2008

Stack et al.

Figure 8. (I) UV spectrum of intermediate A vs UV spectrum of 4-OHE1-1-N7Gua. (II) Kinetic scanning of the conversion of A to 4-OHE1-1N7Gua. (III) Plot of ln [(262/292) - (262f/292f)] vs time (min). (IV) Arrhenius plot of ln(k) vs (1/K × 1000).

intermediate A and the final estrogen DNA adduct, 4-OHE11-N7Gua, have different absorptivites in the UV spectrum. Intermediate A displays a λmax at 262 nm, while the known adduct has a λmax at 292 nm (Figure 8I). Samples of A were collected from the HPLC and intermediately subjected to UV scanning kinetic analysis every 5 min. The change in UV spectra over time is shown in Figure 8II. This change demonstrated a first order process, which represents unimolecular, glycosidic bond cleavage (Figure 8III). A plot of this first order process at 30 °C produced a unimolecular rate constant of 2.92 × 10-4 s-1. This corresponds to a half-life of 40 min at 30 °C. The rate constant was also measured as a function of temperature to obtain an Arrhenius plot. The rate of glycosidic bond cleavage was measured at 15, 20, 25, 30, and 35 °C (Figure 8IV). Each data point was the average of two separate kinetic scanning experiments. The resulting line equation provided values for ∆H+, ∆S+ and ∆G+ of 24.7 kcal/mol, 7.2 eu, and 26.8 kcal/mol, respectively. Our preliminary studies indicate that the rate of glycosidic bond cleavage changes little when the pH is lowered from 6.8 to 3.5 using acetate buffers. Changing buffer concentrations from 0.1 to 1 M also showed little effect on the conversion of A to 4-OHE1-1-N7Gua. However, when the pH was raised to 8.0 or higher, using tris buffers, the conversion of A to 4-OHE1-1-N7Gua was instant. These results will be discussed below.

Discussion When proposing the mechanistic possibilities for the reaction of dG with E2(E1)-3,4-Q, a logical start involves examination

of the Michael addition process and then assessment of bond changes needed to account for the final product 7 in Scheme 1. The initial assumption that Michael addition between the dG and the estrogen quinone would produce the zwitterionic compound 1 is not supported by density functional calculations. This particular Michael addition, between two neutral compounds, is analogous to the addition of nitrogen nucleophiles to carbonyls in that the first steps in Scheme 1 are a conjugated version of this well-studied reaction. When a neutral amine adds to a carbonyl, the eventual product, imine, enamine, etc., is dependent on the structure of the nitrogen nucleophile. However, all of these processes are thought to start by addition of the nitrogen nucleophile to the π-system. Jencks’ extensive investigations to the addition of nitrogen nucleophiles to carbonyls indicated that many of these reactions involve general acid catalysis as opposed to specific acid catalysis especially with weaker nitrogen nucleophiles (35). Thus, the formation of the nitrogen-carbon bond coincides with proton transfer to the oxygen of the carbonyl when forming the carbinolamine intermediate (Scheme 2). The basicity of carbonyls in the electrophilic o-quinone is insufficient to expect complete protonation of O3 prior to Michael addition (36) under the conditions shown in Figure 2 (pH 4.0-4.5), especially in the presence of dG. Complete alkylation prior to proton transfer would lead to the intermediate 1 (Scheme 1), which (1-m) is shown to be an unstable geometry as modeled in the gas phase. In more polar mediums, the zwitter ion 1-m may represent an intermediate on the PES, which would allow for alkylation to precede protonation. Conversely, proton transfer could slightly precede

Michael Addition of dG to Catechol Estrogen-3,4-quinones

Chem. Res. Toxicol., Vol. 21, No. 7, 2008 1423

Figure 9. Free energy differences between the tauomeric forms of intermediate A.

Figure 10. Proposed mechanism for the formation of 4-OHE2(E1)-1-N7Gua from the Michael addition of dG to E2(E1)-3,4-Q.

Scheme 2. Acid-Catalyzed Michael Addition of dG to the o-Quinone

alkylation in strongly acidic medium, but such a medium would sequester the nitrogen nucleophiles of dG. Thus, the Michael addition likely involves a transition state where proton transfer from acetic acid occurs in a somewhat concerted fashion, either slightly preceding or following N7-C1 bond formation. Attempts to model this transition state under the aforementioned conditions are ongoing. The revelation that N7-C1 bonding is not stable without protonation of O3 led to the idea that a reverse Michael process could occur if intermediates similar to 2 lost a proton from O3. If this occurred after glycosidic bond cleavage, a catalytic cycle, where one catechol estrogen quinone causes multiple liaisons in the DNA, could occur. Reverse Michael addition between pyridine nucleophiles and Michael acceptors has been shown to proceed via an E1cB mechanism, where proton removal is the rate-limiting step (37). A necessary but not sufficient requirement for a catalytic cycle would be the conversion of 2 to 4 as opposed to the conversion of 2 to 3 (Scheme 1). Computational models suggest that in an aqueous environment 3-m is slightly favored as compared to 4-m; however, the reaction of catechol estrogen quinones with DNA may occur in the less polar environment present inside the DNA polymer.

In the gas phase, 4-m is favored when compared to 3-m (Figure 4). Even if glycosidic bond cleavage were to precede rearomatization of the estrogen A ring, a catalytic cycle would require loss of proton from O3 from 4-m as opposed to proton loss from C1. Calculations predict that the latter process is thermodynamically favored by more than 20 kcal/mol, 6-m vs 8-m. While proton transfer from carbon acids is known to have a higher kinetic barrier when compared to heteroatom proton transfers (38), a process leading to 8, then catalytic reformation of the o-qunione seems unlikely in the current reactions of E2(E1)-3,4-Q with dG. The independent assay of 2′-deoxyribose and 4-OHE1-1-N7Gua adduct also indicates that a catalytic cycle is not likely since both compounds were formed in equimolar amounts (Table 2). The observation of a reaction intermediate, A, by HPLC analysis when dG and E1-3,4-Q are mixed in the presence of a proton source provides some insight to both the reaction mechanism and the kinetics of glycosidic bond cleavage. Mass spectrum analysis indicates the presence of the ribose moiety, but several tatuomeric forms are consistent with the observed fragmentation pattern. Enol form 2, catechol form 5, and keto form 2-keto are possible structures based on the LC/MS data (Figure 9). The relative stability of the three tautomers, modeled by ab initio calculations (in the same manner described previously), indicates a strong preference for structure 5. While catechol form 5 is predicted to be the more thermodynamically stable tautomer containing the ribose moiety, the UV spectrum of intermediate A is not consistent with a catechol structure. Catechols typically display a significant UV band between 280 and 295 nm. The absorbance of A shows a prominent band around 260 nm (Figure 8I). Also, the observation that glycosidic bond cleavage is greatly accelerated at high

1424

Chem. Res. Toxicol., Vol. 21, No. 7, 2008

pH (greater than 8.0) is also inconsistent with A having a catechol structure. Increasing the electron-withdrawing properties of adducts attached to the N7-position of dG is known to increase the rate of glycosidic bond cleavage (39, 40). At higher pH, 5 would be expected to lose a phenolic proton, rendering the catechol estrogen adduct less of an electron-withdrawing group (see Scheme 1, 5 vs 3). The instant loss of the of 2′deoxyribose moiety at high pH is more consistent with A having the structure of 2. While the tautomerization of 2 to 5 is expected to be exergonic by approximately 15 kcal/mol, this value is less than the experimentally measured kinetic barrier of 26.8 kcal/ mol obtained from the Arrhenius plot (Figure 8IV). The kinetic barrier between 2 and 5 is consistent with the known decrease in proton transfer rates of carbon acids when compared to proton transfers from oxygen or nitrogen (38). Thus, if loss of the proton at C1 is a necessary prequisite for glycosidic bond cleavage, an increase of the pH would be expected to accelerate sugar loss. In light of both computational analysis and experimental results present above, Figure 10 shows the proposed mechanism for the reaction of E2(E1)-3,4-Q with dG in the presence of acetic acid. The Michael addition step occurs closely with proton transfer to O3 to produce the R-ketoenol 2. The relatively slow proton loss from C1 restores the aromatic A ring, which then quickly undergoes glycosidic bond cleavage to form the estrogen-DNA adduct 7 stoichometrically. While the reaction of one 4-OHE2(E1)-1-N7Gua adduct would correlate to one apurinic site in the DNA when estrogen 3,4-quinones react at the N7-position of dG, correlation between formation of other estrogen DNA adducts and generation of apurinic sites cannot be made at this time. In several systems, it has been observed that depurination of 4-OHE2(E1)-1-N3Ade adducts occurs instantly in the DNA, while 4-OHE2(E1)-1-N7Gua adducts have a half-life of about 3 h in the DNA at 37 °C. Any steric or structural feature that increases the rate of glycosidic bond cleavage would make a catalytic process, as proposed in Scheme 1, more likely. Unfortunately, the lack of direct reaction of dA with catechol estrogen-3,4-quinones precludes mechanistic studies on the formation of 4-OHE2(E1)-1-N3Ade as conducted in the manner described above. Such investigations would have to be conducted directly on DNA under in vitro conditions. Acknowledgment. We are grateful for the help of Ron Cerny for obtaining the LC/MS spectrum of intermediate A at the Nebraska Center for Mass Spectrometry (University of NebraskasLincoln).

Stack et al.

(5) (6) (7) (8) (9)

(10)

(11)

(12) (13) (14)

(15)

(16)

(17)

(18)

(19)

Supporting Information Available: File containing MDL.mol structures of all modeled compounds and an expanded version of Table 1 containing enthalpy data. This material is available free of charge via the Internet at http://pubs.acs.org.

(20)

References

(22)

(1) Cavalieri, E. L., Stack, D. E., Devanesan, P. D., Todorovic, R., Dwivedy, I., Higginbotham, S., Johansson, S. L., Patil, K. D., and Gross, M. L. (1997) Molecular origin of cancer: Catechol estrogen3,4-quinones as endogenous tumor initiators. Proc. Natl. Acad. Sci. U.S.A. 94, 10937–10942. (2) Liehr, J. G., Fang, W. F., Sirbasku, D. A., and Ari-Ulubelen, A. (1986) Carcinogenicity of catechol estrogens in Syrian hamsters. J. Steroid Biochem. 24, 353–6. (3) Rogan, E. G., and Cavalieri, E. L. (2004) The role of specific estrogen metabolites in the initiation of breast and other human cancers. Int. J. Cancer PreV. 1, 243–249. (4) Li, K.-M., Todorovic, R., Devanesan, P., Higginbotham, S., Koefeler, H., Ramanathan, R., Gross, M. L., Rogan, E. G., and Cavalieri, E. L. (2004) Metabolism and DNA binding studies of 4-hydroxyestradiol

(21)

(23)

(24)

(25)

and estradiol-3,4-quinone in vitro and in female ACI rat mammary gland in vivo. Carcinogenesis 25, 289–297. Zhu, B. T., and Conney, A. H. (1998) Functional role of estrogen metabolism in target cells: Review and perspectives. Carcinogenesis 19, 1–27. Martucci, C. P., and Fishman, J. (1993) P450 enzymes of estrogen metabolism. Pharmacol. Ther. 57, 237–57. Guengerich, F. P. (1989) Characterization of human microsomal cytochrome P-450 enzymes. Annu. ReV. Pharmaco. Toxicol. 29, 241– 64. Liehr, J. G., and Ricci, M. J. (1996) 4-Hydroxylation of estrogens as marker of human mammary tumors. Proc. Natl. Acad. Sci. U.S.A. 93, 3294–6. Murray, G. I., Taylor, M. C., Mcfadyen, M. C. E., Mckay, J. A., Greenlee, W. F., Burke, M. D., and Melvin, W. T. (1997) Tumorspecific expression of cytochrome P450 CYP1B1. Cancer Res. 57, 3026–3031. Rogan, E. G., Badawi, A. F., Devanesan, P. D., Meza, J. L., Edney, J. A., West, W. W., Higginbotham, S. M., and Cavalieri, E. L. (2003) Relative imbalances in estrogen metabolism and conjugation in breast tissue of women with carcinoma: Potential biomarkers of susceptibility to cancer. Carcinogenesis 24, 697–702. Singh, S., Chakravarti, D., Edney, J. A., Hollins, R. R., Johnson, P. J., West, W. W., Higginbotham, S. M., Cavalieri, E. L., and Rogan, E. G. (2005) Relative imbalances in the expression of estrogen-metabolizing enzymes in the breast tissue of women with breast carcinoma. Oncol. Rep. 14, 1091–1096. Li, J. J., and Li, S. A. (1987) Estrogen carcinogenesis in Syrian hamster tissues: Role of metabolism. Fed. Proc. 46, 1858–63. Newbold, R. R., and Liehr, J. G. (2000) Induction of uterine adenocarcinoma in CD-1 mice by catechol estrogens. Cancer Res. 60, 235–237. Stack, D. E., Byun, J., Gross, M. L., Rogan, E. G., and Cavalieri, E. L. (1996) Molecular characteristics of catechol estrogen quinones in reactions with deoxyribonucleosides. Chem. Res. Toxicol. 9, 851– 859. Chakravarti, D., Pelling, J. C., Cavalieri, E. L., and Rogan, E. G. (1995) Relating aromatic hydrocarbon-induced DNA adducts and c-H-ras mutations in mouse skin papillomas: The role of apurinic sites. Proc. Natl. Acad. Sci. U.S.A. 92, 10422–6. Chakravarti, D., Mailander, P. C., Cavalieri, E. L., and Rogan, E. G. (2000) Evidence that error-prone DNA repair converts dibenzo[a,l]pyrene-induced depurinating lesions into mutations: Formation, clonal proliferation and regression of initiated cells carrying H-ras oncogene mutations in early preneoplasia. Mutat. Res. 456, 17–32. Chakravarti, D., Mailander, P. C., Li, K.-M., Higginbotham, S., Zhang, H. L., Gross, M. L., Meza, J. L., Cavalieri, E. L., and Rogan, E. G. (2001) Evidence that a burst of DNA depurination in SENCAR mouse skin induces error-prone repair and forms mutations in the H-ras gene. Oncogene 20, 7945–7953. Markushin, Y., Zhong, W., Cavalieri, E. L., Rogan, E. G., Small, G. J., Yeung, E. S., and Jankowiak, R. (2003) Spectral characterization of catechol estrogen quinone (CEQ)-derived DNA adducts and their identification in human breast tissue extract. Chem. Res. Toxicol. 16, 1107–1117. Chakravarti, D., Mailander, P. C., Higginbotham, S., Cavalieri, E. L., and Rogan, E. G. (2003) The catechol estrogen-3,4-quinone metabolite induces mutations in the mammary gland of ACI rats. Proc. Am. Assoc. Cancer Res. 44, 180. Sriram, M., andWang, A. H.-J. (1996) In Bioorganic Chemistry: Nucleic Acids (Hecht, S. M., Ed.) Oxford University Press, New York. Ramakrishna, N. V. S., Padmavathi, N. S., Cavalieri, E. L., Rogan, E. G., Cerny, R. L., and Gross, M. L. (1993) Synthesis and structure determination of the adducts formed by electrochemical oxidation of the potent carcinogen dibenzo[a,l]pyrene in the presence of nucleosides. Chem. Res. Toxicol. 6, 554–60. Mulder, P. P. J., Chen, L., Sekhar, B. C., George, M., Gross, M. L., Rogan, E. G., and Cavalieri, E. L. (1996) Synthesis and structure determination of the adducts formed by electrochemical oxidation of 1,2,3,4-tetrahydro-7,12-dimethylbenz[a]anthracene in the presence of deoxyribonucleosides or adenine. Chem. Res. Toxicol. 9, 1264–1277. Hanson, A. A., Rogan, E. G., and Cavalieri, E. L. (1998) Synthesis of adducts formed by iodine oxidation of aromatic hydrocarbons in the presence of deoxyribonucleosides and nucleobases. Chem. Res. Toxicol. 11, 1201–1208. Li, K.-M., Byun, J., Gross, M. L., Zamzow, D., Jankowiak, R., Rogan, E. G., and Cavalieri, E. L. (1999) Synthesis and structure determination of the adducts formed by electrochemical oxidation of dibenzo[a,l]pyrene in the presence of adenine. Chem. Res. Toxicol. 12, 749– 757. Pezzella, A., Lista, L., Napolitano, A., and D’ischia, M. (2005) An expedient one-pot entry to catecholestrogens and other catechol

Michael Addition of dG to Catechol Estrogen-3,4-quinones

(26) (27)

(28)

(29) (30)

compounds via IBX-mediated phenolic oxygenation. Tetrahedron Lett. 46, 3541–3544. Guerrant, G. O., and Moss, C. W. (1984) Determination of monosaccharides as aldononitrile, O-methyloxime, alditol, and cyclitol acetate derivatives by gas chromatography. Anal. Chem. 56, 633–8. Frisch, M. J., Trucks, G. W., Schlegel, H. B., Scuseria, G. E., Robb, M. A., Cheeseman, J. R., Montgomery, J. A., Jr., Vreven, T., Kudin, K. N., Burant, J. C., Millam, J. M., Iyengar, S. S., Tomasi, J., Barone, V., Mennucci, B., Cossi, M., Scalmani, G., Rega, N., Petersson, G. A., Nakatsuji, H., Hada, M., Ehara, M., Toyota, K., Fukuda, R., Hasegawa, J., Ishida, M., Nakajima, T., Honda, Y., Kitao, O., Nakai, H., Klene, M., Li, X., Knox, J. E., Hratchian, H. P., Cross, J. B., Bakken, V., Adamo, C., Jaramillo, J., Gomperts, R., Stratmann, R. E., Yazyev, O., Austin, A. J., Cammi, R., Pomelli, C., Ochterski, J. W., Ayala, P. Y., Morokuma, K., Voth, G. A., Salvador, P., Dannenberg, J. J., Zakrzewski, V. G., Dapprich, S., Daniels, A. D., Strain, M. C., Farkas, O., Malick, D. K., Rabuck, A. D., Raghavachari, K., Foresman, J. B., Ortiz, J. V., Cui, Q., Baboul, A. G., Clifford, S., Cioslowski, J., Stefanov, B. B., Liu, G., Liashenko, A., Piskorz, P., Komaromi, I., Martin, R. L., Fox, D. J., Keith, T., Al-Laham, M. A., Peng, C. Y., Nanayakkara, A., Challacombe, M., Gill, P. M. W., Johnson, B., Chen, W., Wong, M. W., Gonzalez, C., and Pople, J. A. (2004) Gaussian 03, revision D01, Gaussian Inc., Pittsburgh, PA. Osawa, E., Goto, H., and Yamato, M. An efficient algorithm for finding low-energy conformations of chain molecules. Comput. Aided Innovation New Mater. 2 Proc. Int. Conf. Exhib. Comput. Appl. Mater. Mol. Sci. Eng., 2nd, 1993, 845-848. Lee, C., Yang, W., and Parr, R. G. (1988) Development of the ColleSalvetti correlation-energy formula into a functional of the electron density. Phys. ReV. B: Condens. Matter Mater. Phys. 37, 785–789. Becke, A. D. (1993) Density-functional thermochemistry. III. The role of exact exchange. J. Chem. Phys. 98, 5648–5652.

Chem. Res. Toxicol., Vol. 21, No. 7, 2008 1425 (31) Cances, E., Mennucci, B., and Tomasi, J. (1997) A new integral equation formalism for the polarizable continuum model: Theoretical background and applications to isotropic and anisotropic dielectrics. J. Chem. Phys. 107, 3032–3041. (32) Mennucci, B., and Tomasi, J. (1997) Continuum solvation models: A new approach to the problem of solute’s charge distribution and cavity boundaries. J. Chem. Phys. 106, 5151–5158. (33) Huetz, P., Kamarulzaman, E. E., Wahab, H. A., and Mavri, J. (2004) Chemical reactivity as a tool to study carcinogenicity: Reaction between estradiol and estrone 3,4-quinones ultimate carcinogens and guanine. J. Chem. Inf. Comput. Sci. 44, 310–314. (34) Foresman, J. B., and Frisch, Æ. (1996) Exploring Chemistry with Electronic Structure Methods, 2nd ed., Gaussian, Inc., Pittsburgh. (35) Cordes, E. H., and Jencks, W. P. (1962) General acid cataclysis of semicarbazone formation. J. Am. Chem. Soc. 84, 4319–4328. (36) Ettlinger, E. G. (1950) Hydroxynaphthoquinones. II. Cyclization and the basicity and interconversion of ortho and para quinones. J. Am. Chem. Soc. 72, 3090–3095. (37) Heo, C. K. M., and Bunting, J. W. (1992) Rate-determining steps in Michael-type additions and E1cb reactions in aqueous solution. J. Org. Chem. 57, 3570–3578. (38) Wirz, J. (1998) Kinetics of proton transfer reactions involving carbon. Pure Appl. Chem. 70, 2221–2232. (39) Moschel, R. C., Hudgins, W. R., and Dipple, A. (1984) Substituentinduced effects on the stability of benzylated guanosines: model systems for the factors influencing the stability of carcinogen-modified nucleic acids. J. Org. Chem. 49, 363–372. (40) Mueller, N., and Eisenbrand, G. (1985) The influence of N7 substituents on the stability of N7-alkylated guanosines. Chem.-Biol. Interact. 53, 173–181.

TX800071U