Guanine Alkylation by the Potent Carcinogen Aflatoxin B1: Quantum

Jul 13, 2007 - Guanine Alkylation by the Potent Carcinogen Aflatoxin B1: Quantum Chemical Calculations .... Journal of Chemical Information and Modeli...
13 downloads 0 Views 491KB Size
Download PDF
1134

Chem. Res. Toxicol. 2007, 20, 1134–1140

Guanine Alkylation by the Potent Carcinogen Aflatoxin B1: Quantum Chemical Calculations Urban Bren,*,† F. Peter Guengerich,‡ and Janez Mavri† National Institute of Chemistry, HajdrihoVa 19, SI-1001 Ljubljana, SloVenia, and Department of Biochemistry and Center in Molecular Toxicology, Vanderbilt UniVersity School of Medicine, NashVille, Tennessee 37232 ReceiVed March 7, 2007

We report a series of ab initio and density functional theory simulations of guanine alkylation by aflatoxin B1 exo-8,9-epoxide. This reaction represents an initial step of carcinogenesis associated with aflatoxin B1, a potent genotoxic fungal metabolite. Effects of hydration were considered in the framework of the Langevin dipoles solvation model and the solvent reaction field method of Tomasi and co-workers. In silico-calculated activation free energies are in good agreement with the experimental value of 15.1 kcal/mol. This agreement presents strong evidence in favor of the validity of the proposed SN2 reaction mechanism and points to the applicability of quantum chemical methods in studies of reactions associated with carcinogenesis. In addition, we predict that the preference of aflatoxin B1 exo-8,9-epoxide over the endo stereoisomer for the reaction with guanine exists in the aqueous solution and is only further amplified in the DNA duplex. Finally, through comparison with an analogous reaction between 3a,6a-dihydrofuro[2,3b]furan exo-4,5-epoxide and guanine, we show that the large planar body of aflatoxin B1 does not enhance its reactivity and related carcinogenicity. This explains why the planar region of related mycotoxins sterigmatocystin and aflatoxin G1 could have been evolutionarily optimized in a different way. 1. Introduction Aflatoxin B1 (AFB1)1 and related furofuran fungal metabolites are of great interest due to their role in the etiology of human liver cancer (1, 2). These mycotoxins are produced by the common molds Aspergillus flaVus, Aspergillus parasiticus, and Aspergillus nomius which infest agricultural commodities stored in hot moist conditions (3, 4). AFB1 is among the most potent mutagens implicated in human carcinogenesis (5, 6). Moreover, it is mutagenic in bacteria and carcinogenic in fish, and it is a hepatocarcinogen in rodents (7, 8). AFB1 has also been considered a biological/chemical weapon, e.g., in 1989 in Iraq (9). AFB1 is primarily metabolized in humans by cytochrome P450 3A4 to yield the ultimate carcinogen AFB1 exo-8,9epoxide (2, 10). This very reactive electrophile undergoes spontaneous hydrolysis, giving rise to 8,9-dihydro-8,9-dihydroxy AFB1 which under slightly basic conditions converts to Rhydroxydialdehyde AFB1 (5). This dialdehyde forms Schiff bases with primary amine groups leading to protein adducts (6). AFB1 exo-8,9-epoxide also reacts with DNA, evidently after intercalation (1). Reaction occurs with high regiospecificity at the N7 position of guanine (Gua) residues, yielding the trans8,9-dihydro-8-(N7-guanyl)-9-hydroxyaflatoxin adduct (4). The proposed reaction mechanism is depicted in Figure 1. The SN2 substitution represents the rate-limiting step of the reaction. Subsequent protonation is a fast process due to a proton-rich microenvironment surrounding DNA (5). The positively charged * To whom correspondence should be addressed. E-mail: [email protected]. † National Institute of Chemistry. ‡ Vanderbilt University School of Medicine. 1 Abbreviations: AFB1, aflatoxin B1; Gua, guanine; DFT, density functional theory; MO, molecular orbital; EVB, empirical valence bond; DHFF, 3a,6a-dihydrofuro[2,3-b]furan; BO, Born–Oppenheimer; HF, Hartree–Fock; SCRF, solvent reaction field; LD, Langevin dipoles.

imidazole ring of this cationic DNA adduct promotes depurination, resulting in the formation of an apurinic site (11). Alternatively, under slightly basic conditions, the imidazole ring opens to form the chemically and biologically stable AFB1 formamidopyrimidine (12, 13). The initial cationic DNA adduct and the two secondary DNA lesions, individually or collectively, promote GC f TA transversions (11). In more than 50% of human hepatocellular carcinoma cases in areas with AFB1 exposure, a characteristic G to T mutation is observed at the third position of codon 249 of the p53 tumor suppressor gene (14). Moreover, epidemiological as well as experimental data show synergistic effects of AFB1 and hepatitis B virus on hepatocellular carcinoma formation, because hepatitis B acts as a promoting agent in hepatocytes and thus causes manifestations of the initial genetic damage (15). In this article, the activation free energy ∆G+ of the reaction between AFB1 exo-8,9-epoxide and Gua, calculated at several ab initio, density functional theory (DFT), and semiempirical molecular orbital (MO) levels, is compared to the experimental free energy barrier. Its value of 15.1 kcal/mol was obtained by the transition state theory of Eyring from the first-order rate constant k of 42 s-1 determined experimentally (5) for the reaction between guanine residue of DNA and intercalated AFB1 exo-8,9-epoxide. This theory based on the assumption that reactants and transition states are at thermal equilibrium is formulated as

k)

(

kBT ∆G+ exp h kBT

)

(1)

where kB represents the Boltzmann constant, h Planck’s constant, and T the thermodynamic temperature. The validity of the transition state theory in biocatalysis was proven experimentally by the development of catalytic antibodies and theoretically by the success of the empirical valence bond (EVB) method (16). We incorporated the solvation effects by using the solvent

10.1021/tx700073d CCC: $37.00  2007 American Chemical Society Published on Web 07/13/2007

Guanine Alkylation by Aflatoxin B1

Chem. Res. Toxicol., Vol. 20, No. 8, 2007 1135

Figure 1. Proposed mechanism of the reaction between AFB1 exo-8,9-epoxide and guanine giving rise to the trans-8,9-dihydro-8-(N7-guanyl)-9hydroxyaflatoxin adduct.

Figure 2. Proposed mechanism of the reaction between DHFF exo-4,5-epoxide and guanine.

reaction field method of Tomasi and co-workers (17) and the Langevin dipoles method of Florian and Warshel (18). In addition, the solution effects in conjunction with the semiempirical MO methods were studied at the AM1-SM1 and PM3SM3 levels (19). Furthermore, we simulated an analogous reaction between 3a,6a-dihydrofuro[2,3-b]furan (DHFF) exo4,5-epoxide and Gua presented in Figure 2. Since DHFF exo4,5-epoxide represents a truncated version of AFB1 exo-8,9epoxide, it can help in elucidating the contribution of the large planar body of AFB1 to its high reactivity and related carcinogenicity. This is a relevant question, because related mycotoxins sterigmatocystin and aflatoxin G1 differ from AFB1 in the structure of this planar region (20).

2. Computational Methods All calculations were performed at the National Institute of Chemistry in Ljubljana on the CROW 9 cluster (21, 22) consisting of 64 Linux-based personal computers running two AMD Opteron processors at 1.6 GHz. To obtain the Born–Oppenheimer (BO) hypersurface (and consequently the activation energy of the reactions depicted in Figures 1 and 2), we applied a series of ab initio, DFT, and semiempirical MO calculations encoded in the Gaussian03 suite of programs (23). For the reactants, a full geometry optimization was performed. The transition state structure was located with the Berny algorithm. The difference between the energies of the transition state and the reactants is the activation energy. Moreover, we performed vibrational analysis in the harmonic approximation and obtained only real frequencies for the reactants and a single imaginary frequency for the transition state at all levels of theory. Calculation of the Born–Oppenheimer surface for chemical reactions is not a trivial task. It is generally accepted that one needs relatively flexible basis sets and adequate treatment of the electron correlation. The ab initio calculations were performed on the

Hartree–Fock (HF) level of theory in conjunction with the 6-31G(d), 6-31+G(d,p), and 6-311++G(d,p) basis sets. In addition, we considered the DFT method B3LYP that has the Becke threeparameter hybrid gradient-corrected exchange functional (24) combined with the gradient-corrected correlation functional of Lee, Yang, and Parr (25). 6-31G(d), 6-31+G(d,p), and 6–311++G(d,p) basis sets were again used. We are aware of the significant empirical character of the DFT methods, but they do to some extent include the electron correlation. Finally, we also applied the semiempirical MO methods AM1 and PM3. These two methods were used because of their low CPU cost, which facilitates their application in QM/ MM methods in conjunction with thermal averaging and calculation of nuclear quantum effects (26). Solvation free energies of reactants and the transition state were calculated with two methods, the solvent reaction field (SCRF) of Tomasi and co-workers (17) and the Langevin dipoles model (LD) parametrized by Florian and Warshel (18). The SCRF method encoded in the Gaussian03 package was applied at all ab initio and DFT levels. Obtained Merz–Kollman partial atomic charges served as an input for the LD model built in the ChemSol program (27). The AM1-SM1 and PM3-SM3 calculations were performed with the AMSOL-5.4.1 program of Truhlar and co-workers (19).

3. Results and Discussion The calculated activation energies, zero-point energy corrections, and solvation free energies obtained by the SCRF method are collected in Table 1. Hydration free energies calculated using the LD model are listed in Table 2. The AM1-SM1- and PM3SM3-based solvation free energies are presented in Table 3. Tables 1-3 also include activation free energies calculated as

∆G+ ) ∆E+ + ∆ZPE + ∆∆Ghydr

(2)

where ∆E+ represents the activation energy, ∆ZPE denotes the relative zero-point energy of the transition state and the reactants,

1136 Chem. Res. Toxicol., Vol. 20, No. 8, 2007

Bren et al.

Table 1. Activation Free Energies Calculated for the Reactions of Guanine with AFB1 exo-8,9-Epoxide and DHFF exo-4,5-Epoxide via the SCRF Method method AFB1 exo-8,9-epoxide HF/6-31G(d) HF/6-31+G(d,p) HF/6-311++G(d,p) B3LYP/6-31G(d) B3LYP/6-31+G(d,p) B3LYP/6-311++G(d,p) DHFF exo-4,5-epoxide HF/6-31G(d) HF/6-31+G(d,p) HF/6-311++G(d,p) B3LYP/6-31G(d) B3LYP/6-31+G(d,p) B3LYP/6-311++G(d,p)

∆E+ (kcal/mol)a

∆ZPE (kcal/mol)b

SCRF ∆Ghydr (TS) (kcal/mol)c

SCRF ∆Ghydr (R) (kcal/mol)d

SCRF ∆∆Ghydr (kcal/mol)e

58.33 54.77 53.91 49.66 45.60 44.48

-1.12 -1.12 -1.19 -1.30 -1.27 -1.29

-39.15 -48.90 -47.52 -31.59 -43.90 -43.48

-11.18 -15.12 -13.70 -7.78 -13.76 -13.02

-27.97 -33.78 -33.82 -23.81 -30.14 -30.46

47.84 44.11 43.70 37.26 34.13 33.60

-0.71 -0.60 -0.59 -0.78 -0.74 -0.77

-43.93 -50.03 -49.12 -35.64 -44.55 -44.27

-20.12 -23.09 -22.33 -16.14 -20.75 -20.37

-23.81 -26.94 -26.79 -19.50 -23.80 -23.90

+ ∆GSCRF (kcal/mol)f

∆G+exp ) 15.1 29.24 19.77 18.90 24.55 14.19 12.73 23.32 16.57 16.32 16.98 9.59 8.93

a Classical activation energy. b The zero-point energy of the transition state minus the zero-point energy of the reactants. c Hydration free energy of the transition state obtained via the SCRF method. d Hydration free energy of the reactants calculated with the SCRF method. e The hydration free energy of the transition state minus the hydration free energy of the reactants. f Activation free energy.

Table 2. Activation Free Energies Calculated for the Reactions of Guanine with AFB1 exo-8,9-Epoxide and DHFF exo-4,5-Epoxide via the LD Method method AFB1 exo-8,9-epoxide HF/6-31G(d) HF/6-31+G(d,p) HF/6-311++G(d,p) B3LYP/6-31G(d) B3LYP/6-31+G(d,p) B3LYP/6-311++G(d,p) DHFF exo-4,5-epoxide HF/6-31G(d) HF/6-31+G(d,p) HF/6-311++G(d,p) B3LYP/6-31G(d) B3LYP/6-31+G(d,p) B3LYP/6-311++G(d,p)

∆E+ (kcal/mol)a

∆ZPE (kcal/mol)b

LD ∆Ghydr (TS) (kcal/mol)c

LD ∆Ghydr (R) (kcal/mol)d

LD ∆∆Ghydr (kcal/mol)e

+ ∆GLD (kcal/mol)f

58.33 54.77 53.91 49.66 45.60 44.48

-1.12 -1.12 -1.19 -1.30 -1.27 -1.29

-70.11 -81.30 -80.57 -63.27 -77.69 -78.20

-39.62 -42.98 -42.10 -37.76 -44.31 -43.43

-30.49 -38.32 -38.47 -25.51 -33.38 -34.77

∆G+exp ) 15.1 26.72 15.23 14.25 22.85 10.95 8.42

47.84 44.11 43.70 37.26 34.13 33.60

-0.71 -0.60 -0.59 -0.78 -0.74 -0.77

-56.12 -61.58 -61.02 -50.45 -59.71 -59.16

-29.69 -32.31 -31.77 -29.00 -32.76 -32.13

-26.43 -29.27 -29.25 -21.45 -26.95 -27.03

20.70 14.24 13.86 15.03 6.44 5.80

a Classical activation energy. b The zero-point energy of the transition state minus the zero-point energy of the reactants. c Hydration free energy of the transition state calculated with the LD method. d Hydration free energy of the reactants calculated with the LD method. e The hydration free energy of the transition state minus the hydration free energy of the reactants. f Activation free energy.

Table 3. Activation Free Energies Calculated for the Reaction of Guanine with AFB1 exo-8,9-Epoxide and DHFF exo-4,5-Epoxide at the AM1-SM1 and PM3-SM3 Levels of Theory method AFB1 exo-8,9-epoxide AM1-SM1 PM3-SM3 DHFF exo-4,5-epoxide AM1-SM1 PM3-SM3

∆E+ (kcal/mol)a

∆ZPE (kcal/mol)b

TS ∆Ghydr (kcal/mol)c

R ∆Ghydr (kcal/mol)d

∆∆Ghydr (kcal/mol)e

50.52 44.91

-2.25 -2.08

-62.30 -66.67

-37.31 -45.19

-24.99 -21.48

39.24 37.97

-1.29 -1.48

-48.19 -51.58

-29.62 -33.64

-18.57 -17.94

∆G+ (kcal/mol)f ∆G+exp ) 15.1 23.28 21.35 19.38 18.55

a Classical activation energy. b The zero-point energy of the transition state minus the zero-point energy of the reactants. c Hydration free energy of the transition state calculated with the AM1-SM1 and PM3-SM3 methods. d Hydration free energy of the reactants calculated with the AM1-SM1 and PM3-SM3 methods. e The hydration free energy of the transition state minus the hydration free energy of the reactants. f Activation free energy.

and ∆∆Ghydr stands for the corresponding relative hydration free energy. The entropic contribution is frequently calculated as the sum of translational, rotational, and vibrational entropies obtained using the ideal gas, rigid rotor, and harmonic oscillator approximations (28). However, it has been argued that the use of these approximations often leads to a dramatic overestimation of the entropy term for reactions in solution (29). The most important reason for the observed overestimation lies in the fact that the harmonic approximation underestimates the entropy contribution from the low-frequency modes that are more abundant in larger solutes. Therefore, solvation entropies included in relative solvation free energies present the correct

treatment of entropic contribution for reactions in solution (30). Our simulated reactions were initiated from a close-contact reactant configuration trapped within a cage of implicit solvent. Such reactions are unimolecular in nature, and the calculated activation free energies can be, therefore, directly compared to the experimentally measured activation free energy for the firstorder reaction between a guanine residue of DNA and intercalated AFB1 exo-8,9-epoxide. If we wanted to calculate the activation free energy for a bimolecular process (i.e., reaction of deoxyguanosine with AFB1 exo-8,9-epoxide), an additional entropy term to account for the formation of a close-contact reactant configuration within a cage of solvent would have to

Guanine Alkylation by Aflatoxin B1

Chem. Res. Toxicol., Vol. 20, No. 8, 2007 1137

Figure 3. Structures of the reactants and the transition state of the chemical reaction between AFB1 exo-8,9-epoxide and guanine obtained at the B3LYP/6-31G(d) level of theory. Oxygens are colored red, carbons orange, nitrogens blue, and hydrogens white.

be appended to eq 2. This entropy penalty accounts for the process of bringing the reactants from the 1 M standard state to the bulk water concentration of 55 M and could be expressed as –RT ln(1/55) which increases the reaction barrier by 2.4 kcal/ mol (31). Our in silico calculations focus only on the first part of the reactions depicted in Figures 1 and 2 as this SN2 substitution represents the rate-limiting step of both reactions. It leads to the formation of the relatively unstable zwitterionic intermediate. The activation free energy is defined as a free energy difference between the transition state and the reactants. To obtain the activation free energy of this first step, we therefore need only to consider its reactants (AFB1 exo-8,9-epoxide + Gua in Figure 1 and DHFF exo-4,5-epoxide + Gua in Figure 2) and its transition state, i.e., the saddle point (characterized by a single imaginary frequency) on the potential energy surface connecting the reactants and the zwitterionic intermediate. The subsequent protonation step is a fast process due to a proton-rich microenvironment surrounding DNA (5). The predicted activation energy at the HF level for the reaction of AFB1 exo-8,9-epoxide with Gua lies between 53 and 59 kcal/mol (Table 1). For the reaction between DHFF exo4,5-epoxide and Gua, the calculated energy barrier at the same theory level is significantly lower (between 43 and 48 kcal/ mol). DFT calculations maintain this relation, although they significantly reduce the activation energy for both studied reactions. The simulations of the reaction between AFB1 exo8,9-epoxide and Gua were very demanding in terms of CPU time due to large size of the studied system which precluded the application of a post-HF methodology. Using the semiempirical PM3 MO method, which was shown to yield reasonable energetics for the reaction catalyzed by xylose isomerase (26), an activation energy similar to the DFT level was obtained for both reactions (Table 3). The zero-point vibrational energy correction of the reaction barrier is almost negligible for both reactions (Table 1). The difference in zero-point vibrational energies presented in the Supporting Information reflects the difference in the size of both reactive systems. In addition, it should be noted that the DFTand semiempirical MO-calculated BO surfaces are shallower than the HF-calculated ones. A single imaginary vibrational frequency was obtained for the transition state structure at all theory levels for both studied reactions. Values of all corresponding imaginary frequencies are collected in the Supporting Information. To check whether the correct transition state structure was found, we performed a visualization of the

vibration mode of this imaginary frequency using MOLDEN (32) because it should correspond to the reaction coordinate of the first step of the reaction mechanisms depicted in Figures 1 and 2 (33). For the reaction between AFB1 exo-8,9-epoxide and Gua, the vibration mode of this imaginary frequency at all theory levels coincided with the formation of a chemical bond between the N7 atom of Gua and the C8 atom of the epoxide and with the cleavage of the chemical bond connecting this C8 atom to the epoxide oxygen, thus confirming the allocation of the correct transition state structure. An analogous procedure was applied to show that the correct transition state geometry was also allocated for the reaction of DHFF exo-4,5-epoxide with Gua. The structures of the reactants and the transition state calculated at the B3LYP/6–31G(d) level of theory for the reactions of AFB1 exo-8,9-epoxide with Gua and DHFF exo-4,5-epoxide with Gua are presented in Figures 3 and 4, respectively. In Table 1, we also list the hydration free energies calculated by the SCRF method of Tomasi and co-workers (17). The solvent accelerates both studied reactions because the transition state is better solvated than the reactants. This finding reflects the formation of the zwitterionic intermediate in the first step of the reactions depicted in Figures 1 and 2. Reduction of the reaction barrier in terms of the relative hydration free energies is larger for HF methods than for DFT methods. All in all, the SCRF model in conjunction with flexible basis sets 6-31+G(d,p) and 6-311++G(d,p) yields reasonable agreement with the experimental activation free energy for the reaction between AFB1 exo-8,9-epoxide and Gua (HF methods somewhat overestimate it, while DFT methods slightly underestimate it). Moreover, it should be noted that although the reaction of AFB1 exo-8,9-epoxide with Gua is accelerated more by the solvent, our results at all theory levels still predict lower free energy barriers for the reaction of DHFF exo-4,5-epoxide with Gua which should consequently proceed faster. Hydration free energies obtained by the LD solvation model of Florian and Warshel are listed in Table 2 (18). Reduction of the reaction barrier, in terms of the hydration free energies, is larger relative to the corresponding SCRF values for both studied reactions. Again, a smaller reduction is obtained for DFT methods than for HF methods. HF calculations in conjunction with the LD solvation model and flexible basis sets 6-31+G(d,p) and 6-311++G(d,p) give activation free energies for the reaction between AFB1 exo-8,9-epoxide and Gua that are in very good agreement with the experimental value of 15.1 kcal/mol. DFT simulations applying identical basis sets significantly underestimate the reaction barrier. Again, although the reaction of

1138 Chem. Res. Toxicol., Vol. 20, No. 8, 2007

Bren et al.

Figure 4. Structures of the reactants and the transition state of the chemical reaction between DHFF exo-4,5-epoxide and guanine obtained at the B3LYP/6-31G(d) level of theory. Oxygens are colored red, carbons orange, nitrogens blue, and hydrogens white.

AFB1 exo-8,9-epoxide with Gua is accelerated more by the solvent, our results at all theory levels predict lower free energy barriers for the reaction of DHFF exo-4,5-epoxide with Gua which should consequently proceed faster. All in all, the B3LYP functional in conjunction with flexible basis sets 6-31+G(d,p) and 6-311++G(d,p) regardless of the applied solvation model systematically underestimates the activation free energy of the reaction between AFB1 exo-8,9epoxide and Gua. It appears that for the studied system the LD solvation model outperforms the SCRF method, though there are no experimental data available on hydration free energies of the studied species. The good performance of the LD solvation model could be attributed to the fact that it does to some extent include thermal averaging. Moreover, the discrepancy between the solvation free energies obtained by both models could be rationalized by the fact that there were no epoxy species or zwitterions used in their parametrization sets (18, 34). To check for the stereoselectivity of the reaction between AFB1 8,9-epoxide and Gua in aqueous solution, we recalculated its activation free energy for the endo stereoisomer of AFB1 8,9-epoxide. Calculations at the HF level of theory in conjunction with flexible basis set 6-31+G(d,p) and the LD solvation model (which yielded good agreement with the experimental free energy barrier) gave an activation energy of 58.23 kcal/ mol, a zero-point energy correction of -1.17 kcal/mol, a hydration free energy correction of -41.34 kcal/mol, and consequently the activation free energy of 15.72 kcal/mol for the endo stereoisomer of epoxide. Therefore, reaction of Gua with AFB1 exo-8,9-epoxide is preferred by 0.49 kcal/mol in aqueous solution, which could be explained by the steric hindrance of the nucleophilic attack of the N7 atom of Gua due to the H6a and H9a atoms of AFB1 8,9-epoxide in the case of the endo stereoisomer (Figure 5). This is in agreement with a previous computational study of the reaction between ammonia and AFB1 8,9-epoxide coordinated with a few explicit water molecules, where the exo stereoisomer was found to be more reactive as well (35). Such a preference is further enhanced in the DNA duplex, where precovalent intercalation on the 5′ face of Gua places AFB1 exo-8,9-epoxide in the proximity of and in the proper orientation for the SN2 substitution (1). The results of AM1-SM1 and PM3-SM3 methods (Table 3) often represent a good compromise in terms of the required CPU time and the quality of the results. However, in our case, both semiempirical MO methods significantly overestimate the activation free energy of the reaction between AFB1 exo-8,9-

Figure 5. Higher reactivity of AFB1 exo-8,9-epoxide than of AFB1 endo-8,9-epoxide for the reaction with guanine in aqueous solution could be explained by the steric hindrance of the nucleophilic attack of the N7 atom of guanine due to the H6a and H9a atoms of AFB1 8,9-epoxide in case of the endo stereoisomer.

epoxide and Gua. On the other hand, the basic features of previous calculations are retained. Again, although the reaction of AFB1 exo-8,9-epoxide with Gua is accelerated more by the solvent, our results at both theory levels predict lower free energy barriers for the reaction of DHFF exo-4,5-epoxide with Gua which should consequently proceed faster. In our study, we applied only some of the available quantum chemical methods. There is still room for the use of QM/MM methodology with an all-atom representation of the polar environment and application of thermal averaging (36–38). In addition, it would be a challenge to test novel DFT functionals (39) or to reparameterize the semiempirical methods as reported by Truhlar and co-workers (40). Finally, solvation or activation free energies could be calculated with explicit treatment of the polar environment using the free energy perturbation (41) or EVB method (16), respectively.

4. Conclusions In this article, we present quantum chemical simulations of the rate-limiting step of the reaction between AFB1 exo-8,9epoxide and Gua giving rise to the trans-8,9-dihydro-8-(N7-

Guanine Alkylation by Aflatoxin B1

guanyl)-9-hydroxyaflatoxin adduct, the primary mutagenic DNA lesion that is formed. We demonstrated that the Hartree–Fock level of theory in conjunction with flexible basis sets and the Langevin dipoles solvation model gives a good agreement with the experimental activation free energy of 15.1 kcal/mol. This agreement presents strong evidence in favor of the validity of the proposed reaction mechanism and points to the applicability of quantum chemical methods in studies of other reactions associated with carcinogenesis, for which the activation free energies (and the corresponding rate constants) or reaction mechanisms have not yet been experimentally determined. Moreover, we calculated a free energy preference of 0.49 kcal/ mol for AFB1 exo-8,9-epoxide over the endo stereoisomer for the reaction with Gua. Thus, the stereoselectivity of this reaction exists already in the aqueous solution and is only further enhanced in the DNA duplex. Furthermore, calculations at all theory levels predict a lower activation free energy for the analogous reaction between DHFF exo-4,5-epoxide (a truncated version of AFB1 exo-8,9-epoxide) and Gua. Therefore, the large planar body of AFB1 exo-8,9-epoxide does not contribute to its reactivity and related carcinogenicity, which explains why this region could have been evolutionarily optimized in a different way in related mycotoxins sterigmatocystin and aflatoxin G1. The role of the planar part of these micotoxins is probably to strengthen their intercalation in the DNA duplex through stacking interactions, although this assumption would need to be elaborated in future binding calculations with the methodology that is developed and ready to be used (42). Carcinogenesis is a complex pathological process, in which normal cells become neoplastic. In many of the experimental cases, this process is associated with chemical modifications of DNA (43, 44). However, computer modeling of chemical reactions relevant for carcinogenesis has been [with rare exceptions (45–49)] overlooked and can contribute to our understanding, prevention, and treatment of cancer. Acknowledgment. We thank Ana Bergant (University of Ljubljana, Slovenia) for many helpful discussions. Financial support from the Slovenian Ministry of Science and Education through Grant P1-0012 is gratefully acknowledged along with U.S. Public Health Service Grants R01 ES10546 and P30 ES00267. Supporting Information Available: Zero-point energies and imaginary frequencies of the transition state structures for the reactions of guanine with AFB1 exo-8,9-epoxide and DHFF exo4,5-epoxide (Table S1). This material is available free of charge via the Internet at http://pubs.acs.org.

References (1) Iyer, R. S., Coles, B. F., Raney, K. D., Thier, R., Guengerich, F. P., and Harris, T. M. (1994) DNA adduction by the potent carcinogen aflatoxin B1: Mechanistic studies. J. Am. Chem. Soc. 116, 1603–1609. (2) Ueng, Y.-F., Shimada, T., Yamazaki, H., and Guengerich, F. P. (1995) Oxidation of aflatoxin B1 by bacterial recombinant human cytochrome P450 Enzymes. Chem. Res. Toxicol. 8, 218–225. (3) Jones, W. R., Johnston, D. S., and Stone, M. P. (1998) Refined structure of the doubly intercalated d(TATAFBGCATA)2 aflatoxin B1 adduct. Chem. Res. Toxicol. 11, 873–881. (4) Giri, I., Jenkins, M. D., Schnetz-Boutaud, N. C., and Stone, M. P. (2002) Structural refinement of the 8,9-dihydro-8-(N7-guanyl)-9hydroxy-aflatoxin B1 adduct in a 5′-CpAFBG-3′ sequence. Chem. Res. Toxicol. 15, 638–647. (5) Johnson, W. W., and Guengerich, F. P. (1997) Reaction of aflatoxin B1 exo-8,9-epoxide with DNA: Kinetic analysis of covalent binding and DNA-induced hydrolysis. Proc. Natl. Acad. Sci. U.S.A. 94, 6121– 6125.

Chem. Res. Toxicol., Vol. 20, No. 8, 2007 1139 (6) Johnson, W. W., Harris, T. M., and Guengerich, F. P. (1996) Kinetics and mechanism of hydrolysis of aflatoxin B1 exo-8,9-epoxide and rearrangement of the dihydrodiol. J. Am. Chem. Soc. 118, 8213–8220. (7) Giri, I., Johnston, D. S., and Stone, M. P. (2002) Mispairing of the 8,9-dihydro-8-(N7-guanyl)-9-hydroxy-aflatoxin B1 adduct with deoxyadenosine results in extrusion of the mismatched dA toward the major groove. Biochemistry 41, 5462–5472. (8) Giri, I., and Stone, M. P. (2003) Wobble dC·dA pairing 5′ to the cationic guanine N7 8,9-dihydro-8-(N7-guanyl)-9-hydroxyaflatoxin B1 adduct: Implications for nontargeted AFB1 mutagenesis. Biochemistry 42, 7023–7034. (9) Zilinskas, R. A. (1997) Iraq’s biological weapons: The past as future? J. Am. Med. Assoc. 278, 418–424. (10) Johnston, D. S., and Stone, M. P. (1995) Refined solution structure of 8,9-dihydro-(N7-guanyl)-9-hydroxyaflatoxin B1 opposite CpA in the complementary strand of an oligodeoxynucleotide duplex as determined by 1H NMR. Biochemistry 34, 14037–14050. (11) Bailey, E. A., Iyer, R. S., Stone, M. P., Harris, T. M., and Essigmann, J. M. (1996) Mutational properties of the primary aflatoxin B1-DNA adduct. Proc. Natl. Acad. Sci. U.S.A. 93, 1535–1539. (12) Smela, M. E., Hamm, M. L., Henderson, P. T., Harris, C. M., Harris, T. M., and Essigmann, J. M. (2002) The aflatoxin B1 formamidopyrimidine adduct plays a major role in causing the types of mutations observed in human hepatocellular carcinoma. Proc. Natl. Acad. Sci. U.S.A. 99, 6655–6660. (13) Mao, H., Deng, Z., Wang, F., Harris, T. M., and Stone, M. P. (1998) An intercalated and thermally stable FAPY adduct of aflatoxin B1 in a DNA duplex: Structural refinement from 1H NMR. Biochemistry 37, 4374–4387. (14) Smela, M. E., Currier, S. S., Bailey, E. A., and Essigmann, J. M. (2001) The chemistry and biology of aflatoxin B1: From mutational spectrometry to carcinogenesis. Carcinogenesis 22, 535–545. (15) Williams, J. H., Phillips, T. D., Jolly, P. E., Stiles, J. K., Jolly, C. M., and Aggarwal, D. (2004) Human aflatoxicosis in developing countries: A review of toxicology, exposure, potential health consequences, and interventions. Am. J. Clin. Nutr. 80, 1106–1122. (16) Warshel, A. (2003) Computer Modeling of Chemical Reactions in Enzymes and Solutions, Wiley, Weinheim, Germany. (17) Miertus, S., Scrocco, E., and Tomasi, J. (1981) Electrostatic interaction of a solute with a continuum. A direct utilization of ab initio molecular potentials for the prevision of solvent effects. Chem. Phys. 55, 117– 129. (18) Florian, J., and Warshel, A. (1997) Langevin dipoles model for ab initio calculations of chemical processes in solution: Parametrization and application to hydration free energies of neutral and ionic solutes and conformational analysis in aqueous solution. J. Phys. Chem. B 101, 5583–5595. (19) Hawkins, G. D., Lynch, G. C., Giesen, D. J., Rossi, I., Storer, J. W., Liotard, D. A., Cramer, C. J., and Truhlar, D. G. (1996) AMSOL, version 5.4.1, University of Minnesota, Minneapolis, MN. (20) Raney, K. D., Shimada, T., Kim, D.-H., Groopman, J. D., Harris, T. M., and Guengerich, F. P. (1992) Oxidation of aflatoxins and sterigmatocystin by human liver microsomes: Significance of aflatoxin Q1 as a detoxication product of aflatoxin B1. Chem. Res. Toxicol. 5, 202– 210. (21) Borštnik, U., Hodošcˇek, M., and Janežicˇ, D. (2004) Improving the performance of molecular dynamics simulations on parallel clusters. J. Chem. Inf. Comput. Sci. 44, 359–364. (22) Borštnik, U., and Janežicˇ, D. (2005) Symplectic molecular dynamics simulations on specially designed parallel computers. J. Chem. Inf. Model. 45, 1600–1604. (23) Frisch, M. J., Trucks, G. W., Schlegel, H. B., Scuseria, G. E., Robb, M. A., Cheeseman, J. R., Montgomery, J. A., 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., 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-Lahram, M. A., Peng, C. Y., Nanayakkara, A., Challachombe, M., Gill, P. M. W., Johnson, B., Chen, W., Wong, M. W., Gonzales, C., and Pople, J. A. (1996) Gaussian 03, revision C02, Gaussian Inc., Pittsburgh, PA. (24) Becke, A. D. (1993) Density-functional thermochemistry. III. The role of exact exchange. J. Chem. Phys. 98, 5648–5652.

1140 Chem. Res. Toxicol., Vol. 20, No. 8, 2007 (25) 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 37, 785–789. (26) Garcia-Viloca, M., Alhambra, C., Truhlar, D. G., and Gao, J. (2003) Hydride transfer catalyzed by xylose isomerase: Mechanism and quantum effects. J. Comput. Chem. 24, 177–190. (27) Florian, J., and Warshel, A. (1999) Calculations of hydration entropies of hydrophobic, polar, and ionic solutes in the framework of the Langevin dipoles solvation model. J. Phys. Chem. B 103, 10282– 10288. (28) Kuhn, B., and Kollman, P. A. (2000) QM-FE and molecular dynamics calculations on catechol O-methyltransferase: Free energy of activation in the enzyme and in aqueous solution and regioselectivity of the enzyme-catalyzed reaction. J. Am. Chem. Soc. 122, 2586–2596. (29) Strajbl, M., Florian, J., and Warshel, A. (2001) Ab initio evaluation of the free energy surfaces for the general base/acid catalyzed thiolysis of formamide and the hydrolysis of methyl thiolformate: A reference solution reaction for studies of cysteine proteases. J. Phys. Chem. B 105, 4471–4484. (30) Klahn, M., Rosta, E., and Warshel, A. (2006) On the mechanism of hydrolysis of phosphate monoesters dianions in solutions and proteins. J. Am. Chem. Soc. 128, 15310–15323. (31) Jencks, W. P. (1987) Catalysis in Chemistry and Enzymology, Dover Publications, New York. (32) Schaftenaar, G., and Noordik, J. H. (2000) Molden: A pre- and postprocessing program for molecular and electronic structures. J. Comput.Aided Mol. Des. 14, 123–134. (33) Okajima, T. (2001) Ab initio MO investigation on the reactivity for electrophilic substitution of phenolics with oxirane and aziridine, as the model compounds of binding site of mutagen. THEOCHEM 536, 73–82. (34) Barone, V., Cossi, M., and Tomasi, J. (1997) A new definition of cavities for the computation of solvation free energies by the polarizable continuum model. J. Chem. Phys. 107, 3210–3221. (35) Okajima, T., and Hashikawa, A. (2000) Ab initio MO study on the difference of solvent effect between exo and endo stereoisomers of aflatoxin B1 8,9-oxide for SN2 type nucleophilic ring opening. THEOCHEM 532, 205–212. (36) Carloni, P., and Alber, F., Eds. (2003) Quantum Medicinal Chemistry, Wiley, Weinheim, Germany. (37) Spiegel, K., Rothlisberger, U., and Carloni, P. (2004) Cisplatin binding to DNA oligomers from hybrid car-parrinello/molecular dynamics simulations. J. Phys. Chem. B 108, 2699–2707.

Bren et al. (38) Bren, U., Hodošcˇek, M., and Koller, J. (2005) Development and validation of empirical force field parameters for netropsin. J. Chem. Inf. Model. 45, 1546–1552. (39) Zhao, Y., and Truhlar, D. G. (2005) Infinite-basis calculations of binding energies for the hydrogen bonded and stacked tetramers of formic acid and formamide and their use for validation of hybrid DFT and ab initio methods. J. Phys. Chem. A 109, 6624–6627. (40) Pu, J., Ma, S., Gao, J., and Truhlar, D. G. (2005) Small temperature dependence of the kinetic isotope effect for the hydride transfer reaction catalyzed by Escherichia coli dihydrofolate reductase. J. Phys. Chem. B 109, 8551–8556. (41) Bren, U., Martinek, V., and Florian, J. (2006) Decomposition of the solvation free energies of deoxyribonucleoside triphosphates using the free energy perturbation method. J. Phys. Chem. B 110, 12782–12788. (42) Bren, U., Martinek, V., and Florian, J. (2006) Free energy simulations of uncatalyzed DNA replication fidelity: Structure and stability of T-G and dTTP-G terminal DNA mismatches flanked by a single dangling nucleotide. J. Phys. Chem. B 110, 10557–10566. (43) Einolf, H. J., Story, W. T., Marcus, C. B., Larsen, M. C., Jefcoate, C. R., Greenlee, W. F., Yagi, H., Jerina, D. M., Amin, S., Park, S. S., Gelboin, H. V., and Baird, W. M. (1997) Role of cytochrome P450 enzyme induction in the metabolic activation of benzo[c]phenanthrene in human cell lines and mouse epidermis. Chem. Res. Toxicol. 10, 609–617. (44) Yakovleva, L., Handy, C. J., Yagi, H., Sayer, J. M., Jerina, D. M., and Shuman, S. (2006) Intercalating polycyclic aromatic hydrocarbonDNA adducts poison DNA religation by vaccinia topoisomerase and act as roadblocks to digestion by exonuclease III. Biochemistry 45, 7644–7653. (45) Politzer, P., Trefonas, P., Politzer, I. R., and Elfman, B. (1981) Molecular properties of the chlorinated ethylenes and their epoxide metabolites. Ann. N.Y. Acad. Sci. 367, 478–492. (46) Ford, G. P., and Scribner, J. D. (1990) Prediction of nucleosidecarcinogen reactivity. Alkylation of adenine, cytosine, guanine, and thymine and their deoxynucleosides by alkanediazonium ions. Chem. Res. Toxicol. 3, 219–230. (47) Huetz, P., Mavaddat, N., and Mavri, J. (2005) Reaction between ellagic acid and an ultimate carcinogen. J. Chem. Inf. Model. 45, 1564–1570. (48) Kranjc, A., and Mavri, J. (2006) Guanine alkylation by ethylene oxide: Calculation of chemical reactivity. J. Phys. Chem. A 110, 5740–5744. (49) Bren, U., Zupan, M., Guengerich, F. P., and Mavri, J. (2006) Chemical reactivity as a tool to study carcinogenicity: Reaction between chloroethylene oxide and guanine. J. Org. Chem. 71, 4078–4084.

TX700073D