Conformational Analysis of a 4-Hydroxyequilenin Guanine Adduct

648

Chem. Res. Toxicol. 2002, 15, 648-653

Conformational Analysis of a 4-Hydroxyequilenin Guanine Adduct Using Density Functional Theory Shixiang Yan,† Min Wu,† Shuang Ding,† Nicholas E. Geacintov,*,† and Suse Broyde*,‡ Department of Chemistry, New York University, 1001 Main Building, 31 Washington Place, New York, New York 10003, and Department of Biology, New York University, 1009 Main Building, 100 Washington Square East, New York, New York 10003 Received November 30, 2001

Equilenin, a component of the drug Premarin (Wyeth), can be metabolized to a quinonoid, 4-hydroxyequilenin (4-OHEN). 4-OHEN can react with 2′-deoxynucleosides to form unusual cyclic adducts, among which 4-hydroxyequilenin-2′-deoxyguanosine (4-OHEN-dG) is the major product under physiological conditions. The structure and stereochemistry of one stereoisomer, 4-OHEN-dG1, has been obtained previously using electrospray mass spectrometry and NMR methods [Shen et al. (1997) J. Am. Chem. Soc. 119, 11126-11127]; however, details of the conformations around the linkage site have not yet been investigated. The objective of this paper was to determine the conformation at the five-membered ring linkage site for this adduct. We have carried out a computational investigation involving high level quantum mechanical geometry optimization using density functional theory (DFT) for the 4-hydroxyequileninguanine adduct (4-OHEN-G1). Our results reveal that there are three conformational families which differ in the puckering of the five-membered ring at the linkage site and in the cyclohexene-type A ring conformation. The overall structures of all three families are “V”shaped; however, two are quite compact while the third is more open. The lowest energy structure contains a half chair-type cyclohexene A ring, while two structures whose energies are ∼3-4 kcal/mol higher are boat-type. Since the Watson-Crick hydrogen bonding edge of the modified guanine is obstructed by the formation of this bulky nonplanar adduct, it likely would reside in a groove of the DNA double helix.

Introduction The risk women run of developing breast or endometrial cancer is greater if they are exposed longer to estrogens, either through early menarche and late menopause, through estrogen replacement therapy, or both (1, 2). However, the nature of estrogen-induced carcinogenesis is complicated and controversial (3). Several possible mechanisms have been proposed, including both nongenotoxic cell proliferative effects, and genotoxic DNA damage effects (4-6). The latter mechanism involves the metabolism of estrogens (7). Estrogens can be metabolically activated through the major catechol pathway (8, 9), yielding electrophilically active or redox-active quinonoids which can react with cellular DNA (4, 7). It has been shown that estrogen quinonoids can induce cell transformation in vitro (10), and it has been postulated that highly reactive estrogen quinonoids can initiate cancer (4). The major reactions of estrogen quinonoids with DNA lead to DNA damage such as oxidized bases, DNA strands breaks, apurinic sites, and stable bulky DNA adducts via DNA alkylation (7). The formation of stable bulky DNA adducts in vivo is generally considered to be an initiating event in the multistage process of carcinogenesis (4). * To whom correspondence should be addressed. (N. E. G.) Telephone: (212) 998-8407. Fax: (212) 998-8421. E-mail: [email protected]. (S. B.) Telephone: (212) 998-8231. Fax: (212) 995-4015. E-mail: [email protected]. † Department of Chemistry, New York University. ‡ Department of Biology, New York University.

Equilenin, an equine estrogen, and its 17β-hydroxylated analogue, comprise ∼15% of the drug Premarin (Wyeth), the number one choice for estrogen replacement treatment and one of the most widely prescribed drugs in the United States (11). Equilenin can be metabolized by human breast cancer cells (12), as well as by hamster kidney or liver microsomes (13, 14), through the 4-hydroxylation pathway. Though little is known about the mechanism of equilenin carcinogenesis in humans, experimental evidence has shown a link between equilenin and cancer in hamsters (11). The quinonoids from the catechol metabolites of equilenin, 4-hydroxyequilenin (4OHEN),1 can react with 2′-deoxynucleosides at the DNA level to form unusual cyclic adducts, among which 4-hydroxyequilenin-2′-deoxyguanosine (4-OHEN-dG) is the major product under physiological condition (15, 16). The structure and stereochemistry of one stereoisomer, 4-OHEN-dG1, has been determined previously using electrospray mass spectrometry and NMR methods (15). However, conformational details around the linkage site were not investigated. These structural features of the linkage site may govern the conformations of DNA adducts, and hence their biological impact. To determine the conformation at the linkage site for this adduct, we carried out a computational investigation involving high 1 Abbreviations: 4-OHEN, 4-hydroxyequilenin; 4-OHEN-dG, 4-hydroxyequilenin-2′-deoxyguanosine; 4-OHEN-G, 4-hydroxyequileninguanine; DFT, density functional theory; DNA, deoxyribonucleic acid; MMFF94, Merck molecular force field; NMR, nuclear magnetic resonance.

10.1021/tx0101797 CCC: $22.00 © 2002 American Chemical Society Published on Web 05/03/2002

Conformation of 4-OHEN-G1 Adduct Using DFT

Chem. Res. Toxicol., Vol. 15, No. 5, 2002 649

Figure 1. Structure of the 4-OHEN-G1 adduct. Torsion angles δ′, ′, ζ′, and η′ are defined as δ′, N1(G)-C2-C3-N2; ′, C1-C2-C3-C4; ζ′, C3-C4-C5-C10; η′, C9-C10-C1-O1.

level quantum mechanical geometry optimization using density functional theory (DFT) for the 4-hydroxyequilenin-guanine adduct (4-OHEN-G1, Figure 1). DFT calculations are considered to be among the most reliable methodologies in geometry optimization and conformational energy determination (17). Our calculations revealed three possible conformational families with different stabilities; two of these have remarkably different overall shapes, which would produce quite different conformations of this adduct on the DNA level.

Computational Methodology On the basis of the structural characterization and stereochemistry of the 4-hydroxyequilenin-2′-deoxyguanosine adduct (4-OHEN-dG1) (15), we constructed four possible initial models for the 4-hydroxyequilenin-guanine adduct (4-OHEN-G1, Figure 1). These are representative of all the feasible conformational families for the 4-OHEN-G1 adduct, based on the geometric constraints and hybridization of the five-membered ring at the linkage site and the cyclohexene-type A ring of the equilenin moiety (see Results and Discussion). The geometry of the B, C, and D rings of the equilenin moiety (Figure 1) were taken from a crystal structure containing equilenin (18). We used SPARTAN 5.1 from Wavefunction, Inc. for this component of the work. Restrained energy minimizations, which preserved the critical dihedral angles governing the conformations of the linkage five-membered ring and the A ring (see Results and Discussion), were then carried out to obtain starting models using the Merck molecular force field (MMFF94) (19-23) implemented in SPARTAN 5.1. We then optimized the geometries of the four structures using the Becke3LYP functional and the 6-31G* basis set (24, 25) with analytical gradients (26), together with energy calculations. All the quantum mechanical calculations were carried out with Gaussian 98 (27).

Results and Discussion Model Building. Our goal in this part of the work was to survey the conformational space of the linkage site for the 4-OHEN-G1 adduct. We used SPARTAN 5.1 to build the four possible starting models that are representative of the feasible conformations of the 4-OHEN-G1 adduct. The stereochemistries of the five-membered ring at the linkage site and the adjacent cyclohexene-type A ring of the equilenin moiety (Figure 1), together with the assistance of a hand model, served as guides. As shown in Figure 1, the dihedral angle δ′ (N1(G)-C2-C3-N2) is defined to describe the conformation of the linkage fivemembered ring, ′ (C1-C2-C3-C4) and ζ′ (C3-C4-C5C10) define the A ring, and η′ (C9-C10-C1-O1) describes the orientation of the O1-HO1 hydroxyl group. Given the constrained nature of the linkage fivemembered ring and the adjacent cyclohexene-type A ring,

Figure 2. Stereoviews of the four starting conformations of the 4-OHEN-G1 adduct. (a) Conformation I. (b) Conformation II. (c) Conformation III. (d) Conformation IV. Only the atoms on the linkage five-membered ring and those on the cyclohexenetype A ring are shown. Atoms are colored according to their types: carbon, green; nitrogen, blue; oxygen, red; hydrogen, white. All stereo images are constructed for viewing with a stereoviewer.

the sp2 hybridization at N1(G) and C2(G) as well as the primarily sp2 hybridization at N2, the possible conformations are severely restricted since N1(G), C2(G), and N2 cannot pucker. Consequently, we had to consider conformational flexibility only at C2 and C3, which produced four starting conformations for subsequent quantum mechanical geometry optimizations of the next stage. Specifically, in conformation I, the linkage five-membered ring adopts an envelope conformation with δ′ of -33.2°, and the A ring is forced to adopt a sofa conformation with only C2 out of plane (′ ) -30.1°) due to the stereochemistry (Figure 2a); in conformation II, the linkage fivemembered ring adopts an envelope conformation that is opposite to that of conformation I with δ′ of 27.2°; the cyclohexene-type A ring is therefore forced to adopt a conformation with an opposite ′ value of 35.3°, which has only C3 out of plane (Figure 2b). Besides the two envelope conformations that the linkage five-membered ring adopts in conformations I and II, this ring can also be planar; in this case, the A ring can have two possible opposite boat conformations with O1-HO1 axial (conformation III, ζ′ ) -37.4°, η′ ) 102.2°, Figure 2c) and O1-HO1 equatorial (conformation IV, ζ′ ) 30.4°, η′ ) 34.9°, Figure 2d). This produced the set of four starting geometries for the linkage five-membered ring and the associated A ring of the equilenin moiety (Figure 2). Table 1 gives the values of δ′, ′, ζ′, and η′ for the four starting models of the 4-OHEN-G1 adduct. Given the stereochemical assignment at the linkage site of this adduct (15) and

650

Chem. Res. Toxicol., Vol. 15, No. 5, 2002

Yan et al.

Table 1. Starting Conformations of the 4-OHEN-G1 Adduct conformation conformation conformation conformation I II III IV δ′ a ′ ζ′ η′ a

-33.2° -30.1° -2.2° 36.1°

27.2° 35.3° 14.9° 69.3°

-2.0° 2.4° -37.4° 102.2°

-5.7° -4.2° 30.4° 34.9°

The definitions of δ′, ′, ζ′, and η′ are in the caption to Figure

1.

Figure 3. Stereoviews of the three DFT optimized structural families for the 4-OHEN-G1 adduct. (a) Structure 1: the linkage five-membered ring is in a negative envelope conformation (δ′ ) -28.2°), and the cyclohexene-type A ring is in a half chair conformation. (b) Structure 2: the linkage five-membered ring is in a positive envelope conformation (δ′ ) 21.7°), and the A ring is in a distorted boat conformation with O1-HO1 axial. (c) Structure 3: the linkage five-membered ring is planar (δ′ ) -3.6°), and the A ring is in a boat conformation with O1-HO1 axial. Only the atoms on the linkage five-membered ring and those on the A ring are shown. Atoms are colored according to their types: carbon, green; nitrogen, blue; oxygen, red; hydrogen, white.

the hybridization of the linkage site atoms, these four geometries represent all the possible families. Quantum Mechanical Geometry Optimization and Stability Analysis. The geometries of the four starting conformations for the 4-OHEN-G1 adduct were fully optimized using density functional theory (DFT), a high level quantum mechanical treatment using the 6-31G* basis set and Becke3LYP functional. Our four SPARTAN generated starting models converged to three structural families (Figure 3, Table 2). Specifically, starting conformation I converged to structure 1 (Figure 3a). The linkage five-membered ring still adopts the envelope conformation with C3 out of plane (δ′ ) -28.2°), and the cyclohexene-type A ring becomes a half-chair conformation (′ ) -26.2°, ζ′ ) -8.0°). While structure 1 resembles conformation I to a certain extent, it differs in a significant way. Specifically, in conformation I the

Table 2. Stability and Optimized Geometries of the 4-OHEN-G1 Adduct ∆Ea δ′ b ′ ζ′ η′

structure 1

structure 2

structure 3

0 -28.2° -26.2° -8.0° 55.8°

3.76 21.7° 30.1° -19.5° 92.7°

2.99 -3.6° 6.8° -35.2° 111.7°

a All energies (kcal/mol) are relative to structure 1, which is the most stable conformation. b The definitions of δ′, ′, ζ′, and η′ are in the caption to Figure 1.

O1-HO1 group is pseudoequatorial with dihedral angle η′ of 36.1° (near 8° if truly equatorial). However, following the DFT optimization the O1-HO1 group has become closer to pseudoaxial with η′ of 55.8° (near 109° if fully axial). Thus, the allylic strain (28) between the O1-HO1 hydroxyl group and the C11-H2 methylene group, present to some extent in the starting conformation I, has been alleviated in the DFT optimized structure 1. Starting conformation II converged to structure 2 (Figure 3b), in which the linkage five-membered ring adopts an opposite C3 envelope conformation (δ′ ) 21.7°) compared to structure 1 (Figure 3a). The cyclohexene-type A ring in structure 2 has a distorted boat conformation (′ ) 30.1°, ζ′ ) -19.5°). Starting conformation III converges to structure 3 (Figure 3c), in which the linkage fivemembered ring is planar (δ′ ) -3.6°), and the A ring adopts a boat conformation with O1-HO1 hydroxyl group axial (′ ) 6.8°, ζ′ ) -35.2°, η′ ) 111.7°). Starting conformation IV also converges to structure 1, because of the allylic strain (28) between the equatorial O1-HO1 hydroxyl group (Figure 2d) and the C11-H2 methylene group in this conformation. As shown in Figure 3, the three optimized structures 1, 2, and 3 have a pseudoaxial or axial O1-HO1 hydroxyl group (dihedral angle η′ is 55.8°, 92.7°, and 111.7° for structures 1, 2, and 3, respectively). These conformations avoid the steric crowding due to allylic strain (28) involving the O1-HO1 hydroxyl on the A ring of the 4-OHEN-G1 adduct, when it is equatorial. The three optimized structural families of the 4-OHENG1 adduct are characterized by the linkage five-membered ring conformations: negative envelope (structure 1), positive envelope (structure 2), and planar (structure 3). Furthermore, the conformations of the linkage fivemembered ring and the cyclohexene-type A ring are correlated, as demonstrated by the correlation between δ′ and ′; this is a consequence of the severe stereochemical constraints at the linkage site of the 4-OHEN-G1 adduct. The energies of the three optimized structures, derived from the DFT calculations, are shown in Table 2. Structure 1 is more stable than structures 2 and 3 by ∼3-4 kcal/mol. As shown in Figure 4, structure 1 has a very flat “V”-shaped conformation; on the other hand, structures 2 and 3 have a sharp “V”-shape. The flat “V”shaped conformation in structure 1 stems from the halfchair conformation of the cyclohexene-type A ring, which is coupled with the negative envelope conformation of the linkage five-membered ring. This conformation sets the guanine and the equilenin moieties apart, so that they are far from each other. On the other hand, the boat conformation in structure 3 and the similar distorted boat conformation in structure 2 cause a sharp “V”-shaped conformation of this adduct, which, in turn, brings

Conformation of 4-OHEN-G1 Adduct Using DFT

Chem. Res. Toxicol., Vol. 15, No. 5, 2002 651

Figure 4. Overall shape (in stereo) of the three DFT optimized structural families for the 4-OHENG1 adduct. (a) Structure 1. (b) Structure 2. (c) Structure 3. Atoms are colored according to their types: carbon, green; nitrogen, blue; oxygen, red; hydrogen, white.

Figure 5. Conformational differences for structures 1, 2, and 3 of the 4-OHEN-G1 adduct. All the structures are aligned to have both the plane of guanine and the mean plane through the equilenin moiety edge on, and the angle between these two planes are estimated. Atoms are colored according to their types: carbon, green; nitrogen, blue; oxygen, red; hydrogen, white.

the guanine and the equilenin moieties closer to each other. The 3-4 kcal/mol energy difference between structure 1 versus 2 and 3, favoring structure 1, probably stems from the difference between the half-chair conformation of the cyclohexene-type A ring in structure 1 and the boat-type conformation in structures 2 and 3. Both theoretical and experimental studies on cyclohexene (29-

32) have shown that the ideal half-chair conformation is more stable than the ideal boat conformation by ∼6 kcal/ mol. However, our half-chair conformation of the A ring in structure 1 is deviated from an ideal half-chair conformation, because of the coupled stereochemical constraints due to the linked five-membered ring. The dihedral angles ′ and ζ′ of structure 1 are -26.2° and -8.0°, respectively, while, in an ideal half-chair cyclo-

652

Chem. Res. Toxicol., Vol. 15, No. 5, 2002

hexene as revealed by X-ray studies (33-35), they are -62.2° and -14.5°, respectively. Thus, the cyclohexenetype A ring in structure 1 is in a higher energy conformational state compared to the ideal cyclohexene structure. Therefore, the energy difference between the conformational states of the 4-OHEN-G1 adduct comes largely from the stability difference between the halfchair and boat conformations of the A ring in these structures. Moreover, the boat conformation of the A ring in structure 3 is more ideal than the distorted boat conformation of the A ring in structure 2; this likely accounts for the slightly higher energy of structure 2 and, therefore, renders this conformation the least stable one in the three structural families of the 4-OHEN-G1 adduct. Implications for DNA Structure. Bulky DNA adducts have been of great interest because of their possible role in mutagenesis and cancer initiation. A number of different types of such adducts have been under investigation, including those derived from polycyclic aromatic hydrocarbons and amines. These contain two flexible torsion angles that permit a variety orientations of the bulky adduct with respect to the DNA (36, 37). Certain other bulky DNA adducts derived from the aflatoxin family have less flexibility with only a single variable torsion angle governing the linkage site (38). Cyclic adducts, such as the etheno and propano adducts (39, 40), which modify the base through formation of an additional ring, are not very bulky. The equilenin-derived family discussed here represents a novel conformational paradigm, with both highly restricted conformational flexibility and great bulk. These features are of interest on the chemical level as well as in relation to their role in governing adduct structure in DNA. Because of the structure of the 4-OHEN-G1 base adduct (Figure 1), it is evident that the Watson-Crick hydrogen bonding edge of the guanine, involving N1(G) and the N2 amino group, is obstructed by the formation of this unusual cyclic adduct. Therefore, in the context of the B-DNA duplex, the modified guanine residue cannot adopt a normal Watson-Crick conformation as in certain other covalent adducts, even though the O6 of guanine is not substituted. The “V”-shape of the adduct and the bulk of the equilenin moiety of the 4-OHEN-G1 adduct, which has only one aromatic ring (the B ring) and a C18 methyl group sticking out of the plane of the equilenin moiety, make it unlikely to intercalate into the DNA double helix; there is no room inside the DNA structure to accommodate such a large ring system dominated by aliphatic carbons without a massive local disruption of the DNA duplex. It is likely that this adduct will reside in a groove of the DNA double helix, probably with the adducted guanine syn. The structure 1 conformation is a very different conformation from those of structures 2 and 3, which are quite similar. As shown in Figure 5, the angle between the plane of the guanine and the mean plane through the equilenin moiety is ∼122° in structure 1, ∼85° in structure 2, and ∼84° in structure 3. Therefore, the equilenin moiety in structure 1 would have to rotate ∼37° toward the guanine moiety to be transformed into a structure 2- or 3-type form. This striking difference between these two types of conformations could play an important role in determining the conformation of this adduct in the context of a DNA duplex. The energy difference between structures 1 and 2, or 1 and 3 is only

Yan et al.

∼3-4 kcal/mol, which could be compensated by interactions with the DNA; consequently, all these conformations could manifest themselves on the DNA level.

Conclusions The cyclic bulky adduct to guanine, formed with the equilenin quinonoid metabolite 4-OHEN, has unusual structural features, due to the conformationally restricted five-membered ring linkage site. Nonetheless, conformational flexibility involving puckering of this ring leads to strikingly different possible structural families. The adduct structure on the double helix level would be governed by these limited conformational choices.

Acknowledgment. This research is supported by NIH Grants CA-75449 and CA-28038 as well as DOE Grant DE-FG02-90ER60931 to S.B. and NIH Grant CA76660 to N.E.G. Computations were carried out at the Academic Computing Services at New York University, and our own SGI workstations were used for visualization, modeling, and data analysis.

References (1) Liehr, J. G. (1990) Genotoxic effects of estrogens. Mutat. Res. 238, 269-276. (2) Colditz, G. A., Hankinson, S. E., Hunter, D. J., Willett, W. C., Manson, J. E., Stampfer, M. J., Hennekens, C., Rosner, B., and Speizer, F. E. (1995) The use of estrogens and progestins and the risk of breast cancer in postmenopausal women. New Engl. J. Med. 332, 1589-1593. (3) Service, R. F. (1998) New role for estrogen in cancer? Science 279, 1631-1633. (4) Yager, J. D., and Liehr, J. G. (1996) Molecular mechanisms of estrogen carcinogenesis. Annu. Rev. Pharmacol. Toxicol. 36, 203232. (5) Zhu, B. T., and Conney, A. H. (1998) Functional role of estrogen metabolism in target cells: review and perspectives. Carcinogenesis 19, 1-27. (6) Li, J. J., Li, S. A., Gustafsson, J. A., Nandi, S., and Sekely, L. I., Eds. (1996) Hormonal Carcinogenesis II, Springer-Verlag, New York. (7) Bolton, J. L., Pisha, E., Zhang, F., and Qiu, S. (1998) Role of quinoids in estrogen carcinogenesis. Chem. Res. Toxicol. 11, 1113-1127. (8) Maclusky, N. J., Naftolin, F., Krey, L. C., and Franks, S. (1981) The catechol estrogens. J. Steroid Biochem. 15, 111-124. (9) Fishman, J. (1983) Aromatic hydroxylation of estrogens. Annu. Rev. Physiol. 45, 61-72. (10) Pisha, E., Lui, X., Constantinou, A. I., and Bolton, J. L. (2001) Evidence that a metabolite of equine estrogens, 4-hydroxyequilenin, induces cellular transformation in vitro. Chem. Res. Toxicol. 14, 82-90. (11) Li, J. J., Li, S. A., Oberley, T. D., and Parsons, J. A. (1995) Carcinogenic activities of various steroidal and nonsteroidal estrogens in the hamster kidney: relation to hormonal activity and cell proliferation. Cancer Res. 55, 4347-4351. (12) Spink, D. C., Zhang, F., Hussain, M. M., Katz, B. H., Liu, X., Hilker, D. R., and Bolton, J. L. (2001) Metabolism of equilenin in MCF-7 and MDA-MB-231 human breast cancer cells. Chem. Res. Toxicol. 14, 572-581. (13) Li, S. A., Klicka, J. K., and Li, J. J. (1985) Estrogen 2- and 4-hydroxylase activity, catechol estrogen formation, and implications for estrogen carcinogenesis in the hamster kidney. Cancer Res. 45, 181-185. (14) Sarabia, S. F., Zhu, B. T., Kurosawa, T., Tohma, M., and Liehr, J. G. (1997) Mechanism of cytochrome P450-catalyzed aromatic hydroxylation of estrogens. Chem. Res. Toxicol. 10, 767-771. (15) Shen, L., Qiu, S., van Breemen, R. B., Zhang, F., Chen, Y., and Bolton, J. L. (1997) Reaction of the Premarin metabolite 4-hydroxyequilenin semiquinone radical with 2′-deoxyguanosine: formation of unusual cyclic adducts. J. Am. Chem. Soc. 119, 1112611127. (16) Shen, L., Qiu, S., Chen, Y., Zhang, F., van Breemen, R. B., Nikolic, D., and Bolton, J. L. (1998) Alkylation of 2′-deoxynucleosides and DNA by the Premarin metabolite 4-hydroxyequilenin semiquinone radical. Chem. Res. Toxicol. 11, 94-101.

Conformation of 4-OHEN-G1 Adduct Using DFT (17) Hehre, W. J., Radom, L., Schleyer, P. R., and Pople, J. A. (1986) Ab initio Molecular Orbital Theory, John Wiley & Sons, New York. (18) Cho, H.-S., Ha, N.-C., Choi, G., Kim, H.-J., Lee, D., Oh, K. S., Kim, K. S., Lee, W., Choi, K. Y., and Oh, B.-H. (1999) Crystal structure of ∆5-3-ketosteroid isomerase from Pseudomonas testosteroni in complex with equilenin settles the correct hydrogen bonding scheme for transition state stabilization. J. Biol. Chem. 274, 32863-32868. (19) Halgren, T. A. (1996) Merck molecular force field. I. Basic, form, scope, parameterization, and performance of MMFF94. J. Comput. Chem. 17, 490-519. (20) Halgren, T. A. (1996) Merck molecular force field. II. MMFF94 van der Waals and electrostatic parameters for intermolecular interactions. J. Comput. Chem. 17, 520-552. (21) Halgren, T. A. (1996) Merck molecular force field. III. Molecular geometries and vibrational frequencies for MMFF94. J. Comput. Chem. 17, 553-586. (22) Halgren, T. A., and Nachbar, R. B. (1996) Merck molecular force field. IV. Conformational energies and geometries for MMFF94. J. Comput. Chem. 17, 587-615. (23) Halgren, T. A. (1996) Merck molecular force field. V. Extension of MMFF94 using experimental data, additional computational data, and empirical rules. J. Comput. Chem. 17, 616-641. (24) Lee, C., Yang, W., and Parr, R. G. (1998) Development of the ColleSalvetti correlation-energy formula into a functional of the electron density. Phys. Rev. B 37, 785-789. (25) Becke, A. D. (1993) Density-functional thermochemistry. III. The role of exact exchange. J. Chem. Phys. 98, 5648-5652. (26) Peng, C., Ayala, P. Y., Schlegel, H. B., and Frisch, M. J. (1996) Using redundant internal coordinates to optimize geometries and transition states. J. Comput. Chem. 17, 49-56. (27) Frisch, J. M., Trucks, W. G., Schlegel, B. H., Scuseria, E. G., Robb, A. M., Cheeseman, R. J., Zakrzewski, G. V., Montgomery, A. J., Stratmann, E. R., Burant, C. J., Dapprich, S., Millam, M. J., Daniels, D. A., Kudin, N. K., Strain, C. M., Farkas, O., Tomasi, J., Barone, V., Cossi, M., Cammi, R., Mennucci, B., Pomelli, C., Adamo, C., Clifford, S., Ochterski, J., Petersson, A. G., Ayala, Y. P., Cui, Q., Morokuma, K., Malick, K. D., Rabuck, D. A., Raghavachari, K., Foresman, B. J., Cioslowski, J., Ortiz, V. J., Baboul, G. A., Stefanov, B. B., Liu, G., Liashenko, A., Piskorz, P., Komaromi, I., Gomperts, R., Martin, L. R., Fox, J. D., Keith, T., Al-Laham, A. M., Peng, Y. C., Nanayakkara, A., Gonzalez, C., Challacombe, M., Gill, W. M. P., Johnson, B., Chen, W., Wong, W. M., Andres, L. J., Gonzalez, C., Head-Gordon, M., Replogle, S. E., and Pople, A. J. (1998) Gaussian 98 (Revision A.7), Gaussian, Inc. Pittsburgh, PA. (28) Johnson, F. (1968) Allylic strain in six-membered rings. Chem. Rev. 68, 375-413.

Chem. Res. Toxicol., Vol. 15, No. 5, 2002 653 (29) Anet, F. A. L., and Haq, M. Z. (1965) Ring inversion in cyclohexene. J. Am. Chem. Soc. 87, 3147-3150. (30) Rabideau, P. W., Ed. (1989) The Conformational Analysis of Cyclohexenes, Cyclohexadienes, and Related Hydroaromatic Compounds, VCH Publishers, New York. (31) Anet, F. A. L., Freedberg, D. I., Storer, J. W., and Houk, K. N. (1992) On the potential energy surface for ring inversion in cyclohexene and related molecules. J. Am. Chem. Soc. 114, 10969-10971. (32) Jensen, F. R., and Bushweller, C. H. (1969) Conformational preference and interconversion barriers in cyclohexene and derivatives. J. Am. Chem. Soc. 91, 5774-5782. (33) Pedone, C., Benedetti, E., Immirzi, A., and Allegra, G. (1970) Solid state conformation of N-substituted amides. 1. Crystal structure of 4-diethylcarbamoyl-1-cyclohexene-5-carboxylic acid. J. Am. Chem. Soc. 92, 3549-3552. (34) Holbrook, S. R., and van der Helm, D. (1975) Crystal and molecular-structure of δ-1,1′-dicyclo-hexenyl ketone. Acta Crystallogr., Sect. B 31, 1689-1694. (35) Thompson, J. S., and Whitney, J. F. (1984) Copper(i)-olefin complexes-structure of 1-chloro-2-(1-2-η-cyclohexene)-µ-hydrotri(1-pyrazolyl)borato-N-N′-N′-dicopper(i)-C15H20BClCu2N6. Acta Crystallogr., Sect. C 40, 756-759. (36) Geacintov, N. E., Cosman, M., Hingerty, B. E., Amin, S., Broyde, S., and Patel, D. J. (1997) NMR solution structures of stereoisometric covalent polycyclic aromatic carcinogen-DNA adduct: principles, patterns, and diversity. Chem. Res. Toxicol. 10, 111146. (37) Patel, D. J., Mao, B., Gu, Z., Hingerty, B. E., Gorin, A., Basu, A. K., and Broyde, S. (1998) Nuclear magnetic resonance solution structures of covalent aromatic amine-DNA adducts and their mutagenic relevance. Chem. Res. Toxicol. 11, 391-407. (38) Johnston, D. S., and Stone, M. P. (1995) Refined solution structure of 8,9-dihydro-8-(N7-guanyl)-9-hydroxyaflatoxin B1 opposite CpA in the complementary strand of an oligodeoxynucleotide duplex as determined by 1H NMR. Biochemistry 34, 14037-14050. (39) Cullinan, D., Johnson, F., Grollman, A. P., Eisenberg, M., and de los Santos, C. (1997) Solution structure of a DNA duplex containing the exocyclic lesion 3,N4-etheno-2′-deoxycytidine opposite 2′deoxyguanosine. Biochemistry 36, 11933-11943. (40) Weisenseel, J. P., Moe, J. G., Reddy, G. R., Marnett, L. J., and Stone, M. P. (1995) Structure of a duplex oligodeoxynucleotide containing propanodeoxyguanosine opposite a two-base deletion in the (CpG)3 frameshift hotspot of the Salmonella typhimurium hisD3052 determined by 1H NMR and restrained molecular dynamics. Biochemistry 34, 50-64.

TX0101797