Theoretical Prediction of the p53 Gene Mutagenic Mechanism Induced

Apr 18, 2007 - Hai Huang , Hao Wang , Nan Qi , Albena Kozekova , Carmelo J. Rizzo and Michael P. Stone. Journal of the American Chemical Society 2008 ...
0 downloads 0 Views 292KB Size
Download PDF
5362

J. Phys. Chem. B 2007, 111, 5362-5371

Theoretical Prediction of the p53 Gene Mutagenic Mechanism Induced by trans-4-Hydroxy-2-nonenal Dianxiang Xing,†,‡ Lixiang Sun,† Robert I. Cukier,§ and Yuxiang Bu*,†,§ Institute of Theoretical Chemistry, Shandong UniVersity, Jinan, 250100, People’s Republic of China, School of Chemical Engineering, Shandong Institute of Light Industry, Jinan, 250100, People’s Republic of China, and Department of Chemistry, Michigan State UniVersity, East Lansing, Michigan 48824 ReceiVed: NoVember 8, 2006; In Final Form: January 20, 2007

The reaction mechanism of guanine with trans-4-hydroxyl-2-nonenal (4-HNE) and the mutagenic mechanism induced by adducts have been theoretically predicted at a molecular level from the energy point of view. 4-HNE directly reacts with guanine via three steps, yielding eventually four main diastereoisomers: trans4-HNE-dG adducts. A concerted six-atom-centered transition state is proposed for the first step, while the last two steps are involved in four-membered-ring transition states. The third step is the rate-determining step. The studies of base pairing properties of trans-4-HNE-dG adducts with A, T, C, A*, and T* together with the relationship between the mutation and structure of trans-4-HNE-dG indicate that syn- and anticonformations of trans-4-HNE-dG around the glycosidic bond are favorable for pairing with A* and T*, respectively, in the parental generation. As a result, the GC f CG or GC f TA mutation may be generated from the syn-4-HNE-dGA* during replication. Nevertheless, anti-4-HNE-dGT* creates GC f TA mutation or nonmutagenesis. Moreover, syn-4-HNE-dGA* has a slightly higher probability to be generated than anti4-HNE-dGT* in the parental generation; therefore, the GC to TA transversion is predominant among the mutations. In addition, no correlation between the mutations and the stereochemistry of C6 and C8 of trans4-HNE-dG adducts was found in this work. Our mutational results have interpreted well a part of the discrete experimental observations, but the mutagenic process itself has not previously been characterized, through either computation or experiment.

Introduction The p53 tumor suppressor gene is one of the most frequently mutated genes in human cancer, and nearly 50% of all tumors are estimated to contain a mutational hotspot in the p53 gene, such as codon 249 in liver and lung cancer and codons 280 and 285 in bladder cancer.1-8 Racemic (R/S)-4-hydroxyl-2noneal (4-HNE) is one of the carcinogens that are associated with liver cancer and the mutations of the p53 gene.9,10 It is metabolized to highly reactive species,11 which can react with the thiol group and amino acid residue of protein and be involved in a number of human diseases.12-14 Also, 4-HNE could indirectly contribute to carcinogenesis by reaction with DNA-repair enzymes and lead to the loss of nucleotide excision repair.15 Additionally, 4-HNE can bind covalently to DNA bases and exhibits a potential mutagenesis.16 It has been reported that 4-HNE could induce mutation in V79 Chinese hamster ovary cells.17 Recently, a high frequency of GC to TA and AT mutation at codons 249 and 250 was observed when a wildtype p53 TK-6 lymphoblastoid cell line was exposed to 4-HNE.18 It is suggested that these mutations may be induced by 4-HNE modified deoxyguanosine. Studies in vivo and vitro revealed that the direct reaction of 4-HNE with deoxyguanosine gives four possible diastereomers of 6-(1-hydroxyhexyl)-8-hydroxy-1,N2-propano-2-deoxyguanosine (4-HNE-dG) * To whom correspondence should be addressed. E-mail: [email protected]. † Shandong University. ‡ Shandong Institute of Light Industry. § Michigan State University.

adducts, namely, (6R,8S,11R/S)- and (6S,8R,11R/S)-4-HNE-dG (Scheme 1).16,19,20 Site-specific mutagenicity of these adducts in mammalian cells revealed that two of the diastereomers [(6R,8S,11R)- and (6S,8R,11S)-4-HNE-dG] were significantly more mutagenic than the other two.21 In support of the possible role of these adducts in mutation of the p53 gene, a correlation has recently been observed between the preferential formation of 4-HNE-dG adducts at codon 249 and the high probability of GC to TA transversion at this coding region of the p53 gene.22 These results emphasize the potentially important role that 4-HNE and 4-HNE-dG adducts may play in tumor initiation and promotion. Although the mutagenicity and the spectrum of mutagenic events of 4-HNE-dG are well documented, the mutagenesis of 4-HNE and the reactivity of guanine with 4-HNE are not very clarified yet. Previous study discovered that 4-HNE is negative in bacterial mutagenic tests and lacks genotoxicity in lacI transgenic mice.23,24 It seems that the high toxicity masks its genotoxicity.25 Additionally, it is proposed that 4-HNE can be further epoxidized by biological oxidants, forming 2,3-expoxy4-hydroxynonenal in cells, which can react with DNA base and form exocyclic etheno-dG adducts and thus induce GC to TA transversion at the coding regions.26 However, epoxidation of 4-HNE is not known to occur in vivo metabolism.27 It is likely that 4-HNE-dG adducts are the major adducts formed from 4-HNE. Combined together, these results raise the important question if 4-HNE or 4-HNE-dG is an important etiological agent for liver cancer, which has a mutation at codon 249 of the p53 gene. So the reaction mechanisms of guanine with

10.1021/jp0673922 CCC: $37.00 © 2007 American Chemical Society Published on Web 04/18/2007

Prediction of the p53 Gene Mutagenic Mechanism

J. Phys. Chem. B, Vol. 111, No. 19, 2007 5363

SCHEME 1: Schematic Illustration of the Reaction of 4-HNE with Guaninea

a

The stereochemistry of C6 and C8 of the final adducts is trans. For simplicity, they are denoted by trans-4-HNE-dG.

4-HNE and the mutagenic mechanism of the p53 gene induced by 4-HNE are of great interest. In the past years, there have been some experimental studies about the reaction of guanine (G) with R,β-unsaturated carbonyl.16,28 Yet no detailed reaction mechanisms are proposed. A number of NMR and X-ray crystallographic techniques have been used to detect the presence of ionized and wobble mispairs in duplex DNA. However, most of these structures were prepared by cocrystallizing short complementary oligonucleotides containing a single mismatched base pair.29 These conditions are far from those required for DNA replication.30 In addition, conformational DNA dislocations brought about by wobble base pairing could be more easily located and excised by the polymerase proofreading mechanisms than could mispairs involving rare tautomers.31 Recently, theoretical calculations based on various theories and methods have received increasing attention.32 What is more, the applications of theoretical calculations are important for those biological systems that are not amenable to experiments. The present work was done theoretically (i) to provide a possible mechanism of the conjugate addition of G to 4-HNE and (ii) to explore the pairing selectivity of 4-HNE-dG adducts in directing the incorporation of erroneous base from the energy’s viewpoint and thus to clarify one possible mutagenic mechanism of the p53 gene at a molecular level. For this purpose, the focus of the current work is on the base pairing properties of A, T, and C to each conformer of 4-HNE-dG adducts. In addition, extensive studies have confirmed that a series of gene alterations are closely related to the participation of the rare tautomer forms or/and the syn-conformation of nucleic acid bases.33,34 Therefore, besides the normal forms of A and T, the base pairing properties of the imino or enol forms of A and T (A* and T*) are also considered in this work to fully investigate the mutagenesis of 4-HNE-dG. We are aware that the imino (enol) forms of bases are rare and to date no experimental evidence has been given in support of this mechanism, but the mutagenic bypass may involve geometries

present in small proportions, which would be difficult to observe experimentally due to the detection limits of available experimental techniques. However, computational approaches can help elucidate structural possibilities in molecular detail, and thereby suggest how such rare but potentially harmful events may occur. Computational Details Since mutation of codon 249 and the changes in properties (conformation, stabilized energy, and formation free energy, etc.) of the p53 gene are mainly caused by 4-HNE-dG, 4-HNE-dG and the opposite bases were used as a unit of codon 249 to simplify the computational model. The second model employed in this work is that 4-hydroxyl-2-pentenal and guanine were used to model 4-HNE and deoxyguanosine, respectively, in the discussion of the conjugate addition mechanism of guanine to 4-HNE. Since there is no significant impact on the reaction, adducts of 4-hydroxyl-2-pentenal are still named 4-HNE-dG. To explore the relative stabilities of the anti- and synconformation of 4-HNE-dG around the glycosidic bond, the third model is used, in which the 3′ site of deoxyribose in 4-HNEdG employed a hydroxyl group, the 5′-hydroxyl hydrogen atom is mimicked as the methyl group, and the C6-hydroxyhexyl is modeled by C6-hydroxymethyl. The last model is that 4-HNEdG is further simplified by X on the basis of the third model (Figure 4), in which the deoxyribose moiety is substituted by a hydrogen atom. All structures were fully optimized by employing Density Functional Theory (DFT) method with the B3LYP exchange correlation functional and the 6-31+G* basis set.35-37 Frequency analyses have been made at the same level to ascertain the nature of the optimized structures. Relative energies were determined at the B3LYP/6-311++G** level. Preliminary calculations showed that activation energy (∆Ea) with ZPVE at the B3LYP/ 6-311++G** level is almost the same as that calculated at the B3LYP/6-311++G** level by using the B3LYP/6-31+G* geometries and ZPVE corrections. Therefore, all Ea values in

5364 J. Phys. Chem. B, Vol. 111, No. 19, 2007

Xing et al.

Figure 1. Energy profile for the reaction of S-4HNE with guanine. Energies of the transition states are calculated relative to their respective reactants, while those of stable structures are calculated relative to S-4-HNE plus G. Conformation around C6-C11 in I2-a1/I2-b1 is with antiorientation between C11-H and C6-H.

this work were computed at the B3LYP/6-311++G**//B3LYP/ 6-31+G* level with ZPVE corrections at the B3LYP/6-31+G* level for computational time, unless otherwise noted. The binding energy (∆E) and formation free energy (∆G) of the base pair are defined as the difference between the total energy of the base pair and the sum of the total energies of the isolated bases. All of the binding energies are corrected by the basis set superposition error (BSSE) corrections method, where the BSSEs have been evaluated with the Boys-Bernardi counterpoise technique.38 All formation free energies were corrected by the zero-point vibrational energy (ZPVE) calculated at the B3LYP/6-31+G* level. To verify the transition states proposed, the intrinsic reaction coordinates (IRC) tracing the steepest descent path from the transition states toward the reactants and products were evaluated. In determining IRC, the first step was taken from the transition state along the normal mode corresponding to the negative eigenvalue of the Hessian (second energy derivative) matrix. The IRC was computed in mass-weighted internal coordinates by the method of Gonzales and Schlegel,39 using a step size of 10, which is the default step size in the Gaussian 98 program. The effect of the solvent on the addition of guanine to 4-HNE was investigated with the Onsager model.40 As is known, PCM and COSMO are more accurate, but for our system, it is time costly. Thus, we choose the more economic solvent model to illustrate the solvent effect qualitatively. To model in vivo environment, protein with a dielectric constant of 4.3 was employed to compute the solvation effect of the addition reaction as suggested in ref 41. All the calculations were performed at 298.15 K and 1.0 atm with use of the Gaussian 98 suite of programs, and the SCF convergence criteria Tight was used throughout the calculations.42 Results and Discussions 1. A Possible Mechanism of the Conjugate Addition of Guanine to 4-HNE. Mechanistically, the cyclic deoxyguanosine adducts of 4-HNE may be formed via Michael addition at the olefinic carbon (β-C) of 4-HNE by either N1 or N2-NH2 of G followed by 1,2-addition at the aldehydic group by the N2-NH2 or N1 of G, respectively. But experimental studies found that the steric effect of the alkyl group of β-C precludes the addition of N1 to the substituted olefinic carbon of 4-HNE.16,26 This

implies that only Michael addition at β-C of 4-HNE by N2NH2 of G followed by 1,2-addition at the aldehydic group by N1 is possible. Thus, the overall reaction may be mainly outlined as a three-step process (Scheme 1), in which the first step is Michael addition at β-C of 4-HNE by N2-NH2 of G, forming the enol-form intermediate; the second step is the isomerization of the enol-form intermediate to its aldehyde; and the last step is 1,2-nucleophilic addition of the aldehydic group by N1 of G. The transition states between conformational isomers related by rotation about the single bond are not studied here since the barrier to rotation about the single bond is less than 5 kcal/ mol.43 They are not the main factors affecting the reaction. Thus, the three main steps mentioned above were investigated theoretically here in order to understand more fully the reaction of G with racemic (R/S)-4-HNE and to probe the possible intermediates and transition states in the reaction. All optimized structures, the corresponding energy profile, and the detailed energies for the reaction of R/S-4-HNE with G are presented in Figures S1-S4 and Table S1 of the Supporting Information, respectively. 1.1. Reaction of S-4-HNE with G. Step 1: The transition state of this step involves simultaneous nucleophilic attack of the N2-NH2 nitrogen atom of G at the olefinic carbon (β-C) and the transfer of a N2-NH2 proton to the carbonyl oxygen atom of S-4-HNE. The attack of nitrogen from a- and b-face of S-4HNE (Scheme 1) results in two transition states ts1-a and ts1-b with barriers of 27.4 and 24.7 kcal/mol, respectively (Figure 1). The higher energy barrier of ts1-a is presumably due to the larger steric hindrance of the a-face compared with the b-face. ts1-a as well as ts1-b has the nonplanar sixmembered ring with the forming C‚‚‚N bond lengths of 1.822 (ts1-a) and 1.773 Å (ts1-b), and the H‚‚‚O bond lengths of 1.569 and 1.603 Å, respectively. The CdO bond has a length between that of a double and a single bond (1.280 Å) at these two transition states. There is little proton transfer in the concerted transition states. The imaginary vibrational frequencies for ts1-a and ts1-b (404.8i and 323.3i cm-1, respectively) are linked to the attack of the N2-NH2 nitrogen atom toward the olefinic carbon and a proton transfer from the nitrogen atom of N2NH2 toward the carbonyl oxygen atom. This result is similar to that of the concerted mechanism observed experimentally of N-alkylhydroxylamine conjugate addition to alkenotes,28b where the reaction gives a five-atom-centered concerted transition state. No transition state associating only with nucleophilic attack of

Prediction of the p53 Gene Mutagenic Mechanism the amine nitrogen at β-C of S-4-HNE is found in this step. That is, only the concerted mechanism is possible for this step. The products corresponding to ts1-a and ts1-b are two enolform intermediates. They are I1-a1 and I1-b1, respectively, in which the proton of N2-NH2 is already transferred to the carbonyl oxygen atom and the C-N bond is formed (1.490 Å for C-N and 0.981 Å for O-H). Additionally, a new regular sp3 tetrahedral configuration, also the new chiral center (C6 in adducts) around the original β-C of S-4-HNE, is created. Namely, there are two chiral centers in each of I1-a1 and I1b1. One is from S-4-HNE, and the other is newly formed. The newly formed chiral centers in I1-a1 and I1-b1 are 6R- and 6S-, respectively. The conformation of the enol-intermediate around the C6-C11 bond adopts the staggered form with the anti-orientation between C6-H and C11-H. Moreover, I1-a1 and I1-b1 are characterized by one weak hydrogen bond with an H-bond length of ∼2.000 Å and a bond angle of ∼140.0°. The corresponding intermediates are 1.93 and -1.48 kcal/mol relative to S-4-HNE plus G, respectively. It can be conferred that the formation of I1-a1 is slightly thermodynamically unfavorable and more difficult than that of I1-b1 in conjunction with its respective energy barrier mentioned above. Step 2: The intermediates I1-a1 and I1-b1 isomerize into their respective conformers I1-a2 and I1-b2 by rotation about the C6-C7 bond, and then the tautomerization of I1-a2 and I1-b2 proceeds via four-membered-ring transition states ts2-a and ts2-b with a sizable barrier of ∼55 kcal/mol (Figure 1). These high activation barriers, resulting mainly from unfavorable proton-transfer geometries, discourage the reactions. Further consideration of the biological environment effect shows that the addition of a single water to the four-membered ring, expanding it to a six-membered one with less strain, significantly reduces the activation energies to 28.3 (ts2-aw) and 26.9 kcal/ mol (ts2-bw), respectively (Figure 1). The significant catalytic effect indicates that water or protein environment could accelerate the reaction and favor the reaction proceeding. The imaginary frequencies show that the transition states ts2-aw and ts2-bw consist of a proton transfer from the hydroxyl to the catalytic H2O and a simultaneous transfer of another proton from the catalyst to the olefinic carbon of I1. The products corresponding to I1-a2 and I1-b2 are I2-a and I2-b1, respectively. Subsequently, I2-a converts into the more stable and favorable conformer I2-a1 (∼4.8 kcal/mol more stable than I2-a) by rotation about the N5-C6 bond. Step 3: In this step, N9 of I2-a1 (I2-b1) attacks at the carbonyl carbon simultaneously with a proton transfer from the N9-H bond to carbonyl oxygen, resulting in two nonplanar four-membered rings ts3-1 and ts3-1′ (ts3-2 and ts3-2′) due to the upcoming chirality of C8. The C8‚‚‚N9, N9‚‚‚H, and O8‚‚‚H bonds in the third transition states are elongated. Additionally, it is observed that the imaginary frequencies of ts3-1 and ts3-2 (2030i and 1754i cm-1) are linked most to the proton transfer from N9 to the carbonyl oxygen atom, indicating that these two transition states have somewhat zwitterionic character. For ts3-1′ and ts3-2′, the proton almost transfers to the carbonyl oxygen atom, also indicating a zwitterionic character. These results may imply that the third step is manipulated by proton transfer. Moreover, such a zwitterionic transition state confirms that only a concerted mechanism is possible for this step. To verify this assumption, we try to find an intermediate corresponding to only proton-transferred structure but with the carbonyl carbon atom not linked to nitrogen. As expected, the attempts to obtain such an intermediate are

J. Phys. Chem. B, Vol. 111, No. 19, 2007 5365 all converged to the final adducts. Such converged results also confirm the assumption that proton transfer manipulates the third step. After these transition states, four products (1, 1′, 2, and 2′) are produced. In these adducts, the third chiral center (C8) appears, which are S (1) and R (1′), R (2) and S (2′), respectively. Thus, four products with the corresponding three chiral centers are 1 (6R,8S,11S), 1′ (6R,8R,11S), 2 (6S,8R,11S), and 2′ (6S,8S,11S), respectively (Scheme 1). The stereochemistry between C6 and C8 is trans in both 1 and 2, and thus are denoted by trans4-HNE-dG for simplicity, which is a pair of diastereoisomers. Conversely, 1′ and 2′ are denoted by cis-4-HNE-dG. Comparisons of the activation barriers and reaction heats (cf. Figure 1) indicate that trans-4-HNE-dG adducts (1 and 2) are two major adducts of the addition of G to S-4-HNE, in accord with the experimental observation.16,19 Among the three steps, the third step is the rate-limiting step because of a very high barrier. It can be predicted that the reaction rate is slow and the yield is very low. This prediction has been verified by relevant experimental facts that the yield of the addition reaction is only 0.89% after reacting for 16 h.16 In addition, the impact of the conformation around the C6C11 bond on the energy barrier of the rate-limiting step was examined in detail. The free rotation of I2-a1/I2-b1 about the C6-C11 bond creates another two-staggered conformers for each case. One is with an anti-orientation between C11-OH and C6-H, and another is with an anti-orientation between C11-CH3 and C6-H (Figure S3). The results reveal that trans4-HNE-dG adducts are the major ones in every conformer for more favorable energy barriers than those of the corresponding cis ones (Figures 1 and S1). Moreover, two kinds of transadducts ((6R,8S,11S)- and (6S,8R,11S)-adducts) have almost the same yield because of almost equivalent lowest energy barriers of the third step (∼48.8 and 48.3 kcal/mol, respectively). For the (6R,8S,11S)-adduct, the preferentially created conformer in the reaction is anti for C11-OH and C6-H, while for the (6S,8R,11S)-adduct, the preferred conformer is anti for C11-H and C6-H. For these two conformers, the largest barrier difference between trans- and cis-adduct is ∼5.5/5.3 kcal/mol. On the other hand, different conformers of trans-4-HNE-dG can interconvert and coexist because of a low barrier to rotation about the single bond (less than 5 kcal/mol43) and a less than 3.0 kcal/mol energy difference between different conformers. To mimic the real living system, the effect of the solvent surrounding on the third transition states was also examined on two pathways. The result shows that the barrier difference between ts3-1 and ts3-1′ (ts3-2 and ts3-2′) is decreased from ∼2.5 (5.5) to 1.0 (3.8) kcal/mol in the protein environment. That is, the difference between the barriers for trans- and cisadducts becomes small, but the major adducts of the reaction are still trans-4-HNE-dG adducts and a small quantity of cis4-HNE-dG adducts also may be created from the effect of the solvent (protein) surroundings. IRC Analysis: To verify the transition states proposed above, IRC analyses are employed for several characteristic transition states. As Figure 2A shows, the highest energy point corresponds to the transition state (ts1-a). As it move forward and backward, the total energy of each point along the forward and reverse direction is reduced, indicating that when moving toward these two directions, the geometries are becoming more stable. At the same time, the O‚‚‚H and C‚‚‚N bond lengths are gradually lengthened, while the distance of the N‚‚‚H bond is gradually shortened when moving forward. This indicates that the guanine and S-4-HNE are created following the reaction path from the

5366 J. Phys. Chem. B, Vol. 111, No. 19, 2007

Xing et al. TABLE 1: Free Energy Changes (∆G) and the Forward (Ea+) and Reverse (Ea-) Activation Energies for the Tautomerism from A to A* and T to T* a 6-31+G* A f A* A-w f A*-w A-2w f A2-2w T f T* T-w f T*-w

∆G

Ea+(ts)

Ea-(ts)

12.89 9.97 8.42 13.97 10.88

47.29 17.86 15.08 42.82 17.57

34.51 8.67 7.15 29.66 7.10

a ∆G is calculated with zero-point correction. Ea+ and Ea- are corrected with zero-point vibrational frequency; all energies are in kcal/ mol.

Figure 2. IRC calculations initiated from ts1-a (A) and ts3-1 (B) transition states. The highest energy point corresponds to the transition states (ts1-a and ts3-1, respectively). The path is followed in both directions from this point. Twelve points (A) and 9 points (B) are chosen for each forward and reverse direction, respectively. The changes of total energy (b) and rO‚‚‚H (0), rN‚‚‚H (1), and rC‚‚‚N (2) bond lengths of each point are given along the reaction path. All energies are in au, and all bond lengths are in Å.

transition state (ts1-a) to the reactant. However, when moving backward, the changes of the above three bond lengths go in the opposite direction. This means that guanine and S-4-HNE proceed via ts1-a to create the intermediate I2-a1. Additionally, similar results are also obtained in the intrinsic reaction coordinate (IRC) calculations initiated from the transition state ts3-1 as shown in Figure 2B. IRC calculations for other transition states are not performed, since similar results can be deduced from the above IRC analyses. 1.2. Reaction of R-4-HNE with G. The conjugate addition of G to R-4-HNE experiences a very similar process to the reaction of S-4-HNE with G, and four products are formed in this process. Also, the results show that the third step is the ratedetermining step (Table S1) and two kinds of trans-4-HNE-dG adducts, (6S,8R,11R)- and (6R,8S,11R)-4-HNE-dG (a pair of diastereoisomers), are the major products (Figures S2 and S4). Summarily, the addition of G to R-4-HNE proceeds mainly via three steps and the third is the rate-limiting step. Two pairs of diastereoisomers (trans-4-HNE-dG adducts) are the major products created in the lesion process, where the stereochemistry of C6 and C8 is trans with the C6-group occupying the equatorial positions. However, in the following discussion, only the mutagenesis of trans-4-HNE-dG was investigated in order to compare with a part of the experimental observations. For simplicity, 4-HNE-dG is used below to denote the trans-4-HNEdG.

2. Mutagenic Specificity of 4-HNE-dG Adducts. As mentioned above, 4-HNE-dG adducts have three chiral centers (Scheme 1). In the following calculations, the chiral center (C11) arising from 4-HNE was neglected for the computational time. That is, X was used as a model of 4-HNE-dG, in which C6hydroxyhexyl is modeled by C6-hydroxymethyl (Figure 4). Previously, Topal and Fresco33a hypothesized that a rare tautomer form of nucleic acid base also might be involved in mutations, and suggested that guanine in the syn-configuration, instead of the normal anti-configuration, pairs specifically with an imino (enol) form of adenine or guanine (A* or G*) and thus yielded the GC f TA or GC f CG transversions. Also, Eisenstadt et al.33b suggested that BPDE-modified guanine in the syn-conformation pairs with A* or G* to account for the forementioned transversion, where BPDE denotes a polycyclic aromatic hydrocarbon. These hypotheses have received qualitative support from experimental or theoretical studies.33c,34 Hence, the tautomeric reaction of A (T) to A* (T*) and the base pairing properties of X with A* and T* besides C, A, and T are analyzed below to judge which base is preferentially placed opposite X during replication. 2.1. Tautomeric Reaction from A to A* and T to T*. To explore the possibility of the tautomeric reaction from A to A* and T to T*, the relative free energies and the activation energies of the isolated base tautomerism from the imino (enol) form to the keto form are listed in the top part of Table 1. These results are calculated at the B3LYP/6-31+G* level, which shows clearly that the keto form is the most stable tautomer in the gas phase. This is in agreement with our common understanding. In fact, many structural features that are necessary for the biological functions of nucleic acids depend on the interactions with surrounding water. For this reason, we also investigate the role of water in the tautomerism process. One or two water molecules located in the vicinities of the adenine and thymine are taken into consideration (Figure 3). All the calculation results are also listed in Table 1. As presented in Figure 3, in the tautomerism from A-w to A*-w, water accepts the hydrogen atom of the N-H bond from A, and at the same time, it donates its hydrogen atom to A*, which is accepted by N6. In the tautomerism from A-2w to A*-2w, the proton transfers via a water chain from the N-H bond of A to N6 of A*. It can be seen from Table 1 that one and two water molecules significantly decrease the activation energy of tautomerism from 47.29 kcal/mol to 17.86 kcal/mol and 15.08 kcal/mol, respectively. These results indicate that water can catalyze the tautomerism from A to A* and the catalysis of two water molecules is more prominent than that of single water molecules. On the other hand, the corresponding energy barriers for the simultaneous tautomerism from A* to A amounts to 8.67 and 7.15 kcal/mol.

Prediction of the p53 Gene Mutagenic Mechanism

Figure 3. Optimized structures of A-w, A*-w, A-2w, A*-2w, T-w, and T*-w, and their respective transition states TS1, TS2, and TS3.

Additionally, we predicted that A* is 12.78 kcal/mol less stable than A, but A*-w and A*-2w are 9.19 and 7.93 kcal/mol less stable than A-w and A-2w, respectively. That is to say, water makes A* more stable than isolated A*. With use of Boltzmann statistics, the energy difference implies a 6.7 × 10-7-5 × 10-8 ratio (equilibrium constant) between the imino form of adenine and adenine. It has been roughly estimated that an energy difference smaller than ∼13 kcal/mol between the canonical and rare tautomers is required to introduce the appreciable genetic instability induced by the enol/imino form tautomers. In addition, the barriers for the forward and reverse tautomerization should be in the 28-3 kcal/mol range.44 Just as mentioned above, our energy differences and activation energies fall in the appropriate range. At the same time, using the transition-state theory and calculated barrier heights, we roughly estimate that the rate constant for the tautomerism at 298.15 K is in the range of 2-2 × 104 s-1. This rate is high enough for A* to be formed during the lifetime of a cell. Logically, as Table 1 shows, this is also the case for the tautomerism from T to T*. 2.2. Base Paring of 4-HNE-dG Adducts. X has four halfchair conformers (1X, 2X, and their enantiomers), and the stereochemistry for C6 and C8 is (6S,8R) in 1X and 2X, and (6R,8S) in their enantiomers, respectively. It is known that each enantiomer has the same properties (including base pairing properties) as the other one except for optical activity. Therefore, hereafter only properties of 1X and 2X are discussed. Both structures are characterized by one weak intramolecular H-bond (O8-H‚‚‚O10) (2.217 Å for 1X; 2.231 Å for 2X). The geometric difference between 1X and 2X lies in the orientation of the O-H bond of the C6-hydroxymethyl. The difference in energy between 1X and 2X is only ∼0.2 kcal/mol, which could be compensated by interaction with the DNA bases consequently. Hence, 1X and 2X have almost the same possibility to be incorporated in the DNA duplex during replication. As far as the base pairing is concerned, each of 1X and 2X has three H-bonding active atoms that could form stable H-bonds with other bases, viz., O8H, O10, and O11H as shown in Figure 4. Due to the half-chair structure of 1X and 2X, only two adjacent proton-donor and proton-acceptor sites may

J. Phys. Chem. B, Vol. 111, No. 19, 2007 5367

Figure 4. Possible H-bonding sites of 1X, 2X, and five isolated bases. 1X has two binding site. For simplicity, they are denoted by 1X1 and 1X2, respectively; 2X has one binding site, denoted by 2X1. Binding site 1 and site 2 of C are denoted by C1 and C2, respectively; binding site 1 and site 2 of T are denoted by T1 and T2, respectively; binding site 1 and site 2 of T* are denoted by T*1 and T*2, respectively; either A or A* has only one binding site. The normal orientations of the bases shown in this figure are denoted by A, T, C, A*, and T*, while the flipped orientation will be denoted by adding a subscript f, such as Tf. Additionally, symbols X1Y or X2Y (Y stands for different nucleic acid bases opposite X1 or X2) were used denoting the names of a set of base pairs.

participate in forming the H-bond with another base. Thus, the relative positions of O8 (H), O10, and O11H have shown two possible binding sites. One site (site 1) is O8H and O10 as proton donor and acceptor, and the other (site 2) is O11H and O8 as proton donor and acceptor, respectively, as shown in Figure 4. Thus, three possible binding modes, 1X1, 1X2, and 2X1, are possible for 1X and 2X. In reference to the conformation of the modified guanine with respect to the sugar moiety, 4-HNE-dG may be in the anti- or syn-conformation. To further confirm our assumption, the relative stabilities of the anti- and syn-conformations corresponding to 1X have been calculated by using the third model. In these two optimized structures, the 3′ site of deoxyribose is a hydroxyl group while the 5′-hydroxyl hydrogen atom is substituted by the methyl group. The energy of the synconformation is only 2.81 kcal/mol higher than that of the anticonformation (Figure S5), indicating that its syn-conformation is slightly less stable than the anti-conformation. But the energy difference is small, which also may be compensated by interaction with the DNA bases consequently. Thus, unlike the unmodified guanine, 4-HNE-dG could adopt the anti- or synorientation to form base pairs. We did not calculate the relative stabilities of the anti- and syn-conformation corresponding to 2X since the relative stabilities of these systems can be extrapolated from the data calculated for the relative stability of the anti- and syn-conformations above. In conjunction with the binding sites analyzed above, it can be easily observed that X could adopt the syn- or anticonformation around the glycosidic bond when paired with other bases by site 1. However, when X pairs with another base by site 2, it could only adopt the anti-conformation. According to the possible binding sites of X and five bases (A, T, C, A*, and T*) as presented in Figure 4, the combination of X with five bases results in 18 base pairs including three XC, three XA*, three XT*, three XA, and six XT base pairs, respectively. These pairs are divided into X1 (12 base pairs) and X2 (6 base pairs) families according to the binding sites of X, as displayed in Figure S6. In the X1 family, the anticonformation of 1X and 2X could pair with A, site 1 of C (C1),

5368 J. Phys. Chem. B, Vol. 111, No. 19, 2007

Xing et al.

TABLE 2: Binding Energies (∆E), Free Energy Changes (∆G), and RN-N Values of Base Pairs in the X1 Family (energy in kcal/mol and RN-N in Å)a pairs 1X1sA* 1X1T*1 1X1C1 1X1A 1X1sT1 1X1T2

∆E

∆G

RNsN

-16.59 -4.27 10.28 -13.37 -1.34 9.84 -13.27 -0.60 10.32 -11.79 1.50 10.87 -9.54 2.20 8.49 -9.45 2.37 8.29

∆E

pairs 2X1sA* 2X1T*1 2X1C1 2X1A 2X1sT1 2X1T2

∆G

RNsN

-16.38 -4.12 10.20 -12.73 -1.29 8.93 -12.62 0.08 10.27 -11.69 1.64 10.86 -9.70 2.18 8.49 -9.35 2.45 8.25

a ∆E displayed here includes the BSSE correction and ∆G includes the zero-point energy correction. RN-N denotes distances between two nitrogen atoms of base pairs, which are connected with the deoxyribose unit in DNA. RN-N of GC is 9.05 Å. The positive ∆G means that the formation of base pairing is more difficult and nonspontaneous.

TABLE 3: Binding Energies (∆E), Free Energy Changes (∆G), and RN-N Values of Base Pairs in the X2 Family (energy in kcal/mol and RN-N in Å) ∆E ∆G RN-N

1X2A*f

1X2T*2

1X2C2

1X2A

1X2T1f

1X2T2

-13.02 0.23 11.92

-15.72 -1.84 10.33

-14.69 -0.60 10.34

-9.64 4.36 12.09

-10.10 2.48 10.40

-9.83 2.81 10.28

TABLE 4: Gibbs Free Energy Changes (∆G), Equilibrium Constants (K), and Activation Energies (Ea) of the Tautomerization for CA* f C*A and GT* f G*T (kcal/mol)a CA* f C*A GT* f G*T

∆G

K

Ea

-3.50 -1.11

367.84 6.55

3.87 2.05

a E corrected with zero-point vibrational energy calculated at a B3LYP/6-311++G**.

site 2 of T (T2), and site 1 of T* (T*1). However, rotations of 1X and 2X around the glycosidic bond from the normal antito syn-conformation would allow the pairing with A* and site 1 of T (T1). In the X2 family, A, C2, T*2, and T2 adopt the normal Watson-Crick face when paired with the anticonformation of 1X2, while A* and T1 should be flipped to facilitate the binding site 2 of 1X. It is noted that all base pairs in X1 and X2 families are nonplanar and include two H-bonds. Two H-bond lengths and the corresponding angles for every base pair are listed in Table S2 of the Supporting Information. Hopefully, they may be helpful in further experimental investigations on these structures. Tables 1 and 2 show that the size expansion of guanine increases RN-N of XC (not mispairing in the sizeexpanded case) relative to that of natural GC. In the X1 family, the RN-N value of X1sA* is proximal to that of X1C1 (Table 1), whereas those of other base pairs have a large difference from that of X1C1. The largest difference is ∼2.00 Å (X1T2), while the smallest one is ∼0.60 Å (X1A). In the X2 family, RN-N of 1X2T*2, 1X2T1f, and 1X2T2 are very close to that of 1X2C2 (within (0.06 Å). However, RN-N of 1X2A*f and 1X2A relative to 1X2C2 are larger (>1.5 Å, see Table 2). Thus, X1sA*, X2T*, and X2T have similar RN-N values to XC (not mispairing) in the respective family, while other base pairs display a large difference in RN-N. As indication of the stability of a pair and the magnitude of the driving force of a pair formation, the predicted binding energies (∆E) and formation free energies (∆G) are compared to judge the stability and the formation ability of different base pairs. ∆E and ∆G have similar orders in the same family. In the X1 family, X1sA* with the largest binding energy is the most stable pair, while the next most stable one is X1T*1 at 3.54 kcal/ mol of binding energy smaller than X1sA*, being only slightly

Figure 5. Optimized geometry and RN-N of the 1X1sA* and 2X1sA* calculated at b3lyp/6-31+G*. The values on the dashed line refer to the computed RN-N values (in Å) of the base pairs (distances between two nitrogen atoms of base pairs, which are connected with the deoxyribose unit in DNA). Subscripts (digit and s) stand for the binding sites and conformation around the glycosidic bond. For example, 1X1sA* indicates that 1X adopts the syn-conformation to pair with the normal Watson-Crick face of A* by site 1.

more stable than X1C1 (not mispairing). In other words, replacement of C in X1C1 by A* leads to an increase in stability by 3.54 kcal/mol, while replacement of C in X1C1 by T* results in a relatively minor change in stability. However, replacement of C in X1C1 by A or T destabilizes the base pair by 1.21 or 3.44 kcal/mol, respectively. The same stability can be observed from the ∆G analysis for these base pairs, in which only X1sA*, X1T*1, and X1C1 could be formed spontaneously, and X1sA* is the easiest one to form in five types of base pairs. Additionally, as mentioned above, RN-N of X1sA* is much closer to that of X1C1 than that of the other base pairs. Overall, the possible bases to form base pairs with X1 are A*, T*, and C during replication. Similarly, in the X2 family, ∆E and ∆G analyses (Table 2) reveal that the most stable base pairs are 1X2T*2 and 1X2C2. The relative stability of 1X2T*2 is slightly larger than those for 1X2C2, but the relative stability and formation ability of the other three base pairs (1X2A*, 1X2T, and 1X2A) are far lower or more difficult than those of 1X2C2. Clearly, the higher stability of 1X2T*2 may be due to its stronger H-bonds relative to other pairs. Overall, one possible mechanism for mutation induction at GC sites by 4-HNE might involve the covalent binding of 4-HNE to G in the template strand, and a subsequent rotation of the modified guanine around its glycosidic bond from the normal anti- to the syn-conformation. This could allow the preferred pairing of A* with the modified guanine (Figure 5). This result also theoretically confirmed the hypothesis presented by Eisenstadt et al. that bulky BPDE-modified G in the synconformation pairs with the imino of adenine.34a Another possible mechanism for mutation at GC sites might involve a preferred pairing of the modified G in the anti-conformation with T* (Figure 6). Such pairing does not need the rotation of the modified G around the glycosidic bond in the template strand. Though no direct experimental evidence has been found for the predicted X1sA* and X2T*2 so far, it is believed that the existence of X1sA* and X2T*2 should be detected with the advancement of experimental technique. 2.3. Mutagenic Mechanism. As discussed above, the synconformation of X around the glycosidic bond is favorable for pairing with A* by the binding sites 1, respectively, and the corresponding anti-conformation is favorable for pairing with T* by the binding site 2 among all bases considered during replication. It is rational that 4-HNE-dG has similar pairing ability to X. First, if the 4-HNE-dG adduct rotates around the

Prediction of the p53 Gene Mutagenic Mechanism

Figure 6. Optimized geometry and RN-N of 1X1sT* calculated at the B3LYP/6-31+G* level. Notations in this figure represent a similar meaning to those in Figure 3.

Figure 7. The sketch of mutation induced by 4-HNE-dG.

glycosidic bond from the normal anti- to the syn-conformation in the template strand, it is preferred to pair with A* (Figure 7, 1a). When this pair is introduced into the DNA duplex in the parental generation, C will be incorporated opposite A* in the first generation45 (Figure 7, 1a f 1b). Then in the second generation, C will pair with G and thereby a CG pair is generated (Figure 7, 1b f 1c). Therefore, the possible mutation at GC sites induced by the syn-conformation of 4-HNE-dG is the GC f CG transversion. However, If A* tautomerizes to A in the first generation, T will be placed opposite A in the second generation (Figure 7, 1d f 1e). Actually, the barrier of CA* f C*A is only 3.87 kcal/mol, indicating that the tautomerization is a very rapid process (Table 2, geometries in Figure S7 of the Supporting Information). Additionally, the equilibrium constant and negative free energy in Table 2 show that C*A is more stable and the probability of obtaining C*A is much higher than that of CA* in the first generation (Figure 7, 1b f 1d). As a result, 4-HNE-dG adducts induce the GC f TA transversion. The GC f TA transversion is one of the most common mutations in human cancer.45,46 As discussed above, the CA* mismatch has a high possibility to tautomerize to the C*A mismatch, so the frequency of GC f CG transversion in the second generation is lower than that of the GC f TA transversion. Briefly, the pairing between the modified G in the syn-conformation and A* yields a high frequency of GC f TA transversion and, to a lesser extent, the GC f CG transversion. However, if 4-HNE-dG is in the anti-conformation during replication, T* is slightly preferential to be placed oppositely in the parental generation (Figure 7, 2a). Thus, if this pair is introduced into the DNA duplex in the parental generation, G is placed opposite T* in the first generation45 (Figure 7, 2a f 2b), and then C will be placed opposite G in the second generation (Figure 7, 2b f 2c). Finally, the mismatch of DNA is repaired to canonical GC (Figure 7, 2c). Similarly, if T* tautomerizes to T in the first generation, A will pair with T in the second generation (Figure 7, 2b f 2d, 2d f 2e). Thus, the

J. Phys. Chem. B, Vol. 111, No. 19, 2007 5369 mutation from GC to AT transition induced by 4-HNE-dG adducts is generated. This result shows that the barrier for GT* f G*T is only 2.05 kcal/mol (Table 2, geometries in Figure S7 of the Supporting Information), indicating that the tautomerization is a very rapid process. The equilibrium constant and negative free energy show that GT* is more stable and the probability of obtaining GT* is higher than that of G*T in the first generation, so it can be conferred that the GC to AT transition has a higher frequency of being generated than the canonical GC pair in the second generation. Altogether, GC to TA transversion, GC to AT transition, and GC to CG transversion could be observed in the mutagenic study of the 4-HNE-dG adducts. Since A* has a slightly higher probability to be placed opposite the syn-conformation of 4-HNE-dG than T* opposite the anti-conformation of 4-HNEdG, the mutation induced by the XA* mismatch is higher than that by XT* and thus the GC to TA transversion is the most frequently observed one among the mutations induced by 4-HNE-dG. Notably, all the mutations derived from our predicted mutagenic mechanism have been evidenced experimentally by Feng et al.22b and these pairing modes are consistent with the hypothesis that the syn- or anti-conformation of the base around the glycosidic bond is associated with mutagenesis.21 In addition to these mutations, a high frequency of nonmutagenesis GC is generated for the pairing between the anti-conformation of X and T*. Fernandes et al. also hypothesized that the ring-opened forms of 4-HNE-dG adducts are responsible for the high number of nonmutagenic substitutions observed with four stereoisomers.21 In contrast, our results found that the ring-closed forms could also contribute to the high frequency of nonmutagenesis. Moreover, no correlation between mutation and the stereochemistry of C6 and C8 was found in the current study. 4-HNE could inhibit nucleotide excision repair in human cells through the direct interaction with proteins involved in DNA repair, and the inhibitory effect on DNA repair induced by 4-HNE may inhibit repair of various kinds of DNA damage induced either endogenously or exogenously.15 This also may be a contributory factor in the occurrence of mismatch and subsequent mutation. Concluding Remarks In this work, the chemical lesion mechanism of G and the subsequent mutagenesis were explored theoretically. Several conclusions can be drawn as follows: (1) The reaction process of 4-HNE and guanine mainly involves three steps except for the inversion between conformers related by rotation about the single bond, eventually giving four major adducts (trans-4-HNE-dG adducts) with an equatorial C6group. In the first step, the conjugate addition of N2-NH2 of G to β-C of 4-HNE proceeds via a concerted six-membered-ring transition state to form enol-form geometry. Consequently, the enol form tautomerizes to its keto form via a four-membered ring transition state. Then the guanine N1 addition to the aldehydic group of the keto form proceeds via a four-memberedring transition state to the final adducts. The second step can be catalyzed by a single water and the activation energy is reduced significantly. The third step is the rate-determining step because it has the highest energy barrier (∼50 kcal/mol) among the three steps. Also, this may be the main reason for their low experimental yields, which will further affect the subsequent mutation frequency. (2) One possible mechanism for the mutation induction at the GC site of the p53 gene by 4-HNE might involve the rotation of adducts around its glycosidic bond from the anti- to the syn-

5370 J. Phys. Chem. B, Vol. 111, No. 19, 2007 conformation. This conformation would allow preferential pairing with an imino tautomer of adenine (A*). However, the anti-conformation would allow preferential pairing with the imino tautomer of Thymine (T*). As a result, syn-4-HNE-dGA* alignment generates GC to TA and GC to CG transversions, while anti-4-HNE-dGT* eventually generates GC to AT transition, and also nonmutagenesis. (3) No correlation between the mutations and the stereochemistry of C6 and C8 of 4-HNE-dG adducts was found in this study. The results reported here provide the first insight into the detailed reaction mechanism of G with 4-HNE, the base pairing ability of 4-HNE-dG, as well as the possible mutagenic mechanism induced by 4-HNE-dG adducts. The results have explained some discrete experimental observations regarding the mutation observed by Feng et al. and that 4-HNE-dG adducts may contribute to GC to TA transversion at codon 249 of the p53 gene.22 It also can be conferred that the low yield of adducts and the subsequent pairing with the rare tautomer of A and T (A* and T*) will induce a small frequency of mutation of the p53 gene. Moreover, our results also theoretically verified and detailed the hypothesis about the transversion induced by BPDE modified guanine.33 Though a direct experimental proof of the predicted mutagenic mechanism is missing, we believe that the plausibility of the obtained results will be warranted with the advancement in experimental technique. Additionally, it should be noted that the impact of the stereochemistry of C11 in 4-HNE-dG adducts on the mutation and the impact of the base pairs formed by 4-HNE-dG with A*, T*, C, A, and T on the properties and the dynamic behaviors of duplex as well as why 4-HNE preferentially adducted at the third base of codon 249 (AGG) were not studied in this work. Therefore, further work should be performed to investigate fully the mutation induced by 4-HNE. Acknowledgment. This work is supported by NSFC (20573063,20633060), NIH (Grant No. GM62790), NCET, and shandong-NSF (Z2003B01). Support from Virt Lab Comput Chem of CNIC and Supercomputing Center of CNIC-CAS is also acknowledged. Portions of the calculations were performed on the CBM and the MCBILIN clusters at Michigan State University, and HPCC at Shandong University. Y.B. also thanks Prof. Yi Hu for his help in calculations. Supporting Information Available: Computational detail, reaction mechanism, and various possible base pairs, etc. This material is available free of charge via the Internet at http:// pubs.acs.org. References and Notes (1) Chen, J. X.; Zheng, Y.; West, M.; Tang, M. S. Cancer Res. 1998, 58, 2070-2075. (2) Greenbaltt, M. S.; Bennett, W. P.; Hollstein, M.; Harris C. C. Cancer Res. 1994, 54, 4855-4878. (3) Hollstein, M.; Shomer, B.; Greenblatt, M.; Soussi, T.; Hovig, E.; Montesano, R.; Harris, C. C. Nucleic Acids Res. 1996, 24, 141-146. (4) Pfeifer, G. P.; Denissenko, M. F. EnViron. Mol. Mutagen. 1998, 31, 197-205. (5) Hainaut, P.; Pfeifer, G. P. Carcinogenesis 2001, 22, 367-374. (6) Denissenko, M. F.; Pao, A.; Tang, M. S.; Pfeifer, G. P. Science (Washington, D.C.) 1996, 274, 430-432. (7) Smith, L. E.; Denissenko, M. F.; Bennett, W. P.; Li, H.; Amin, S.; Tang, M.; Pfeifer, G. P. J. Natl. Cancer Inst. 2000, 92, 803-811. (8) Denissenko, M. F.; Chen, J. X.; Tang, M. S.; Pfeifer, G. P. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 3893-3898. (9) Esterbauer, H.; Schaur, R. J.; Zollner, H. Free Radical Biol. Med. 1991, 11, 81-128.

Xing et al. (10) (a) Pandya, G. A.; Moriya, M. Biochemistry 1996, 35, 1148711492. (b) Fink, S. P.; Reddy, G. R.; Marnett, L. J. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 8652-8657. (11) Schneider, C.; Tallman, K. A.; Porter, N. A.; Brash, A. R. J. Biol. Chem. 2001, 276, 20831-20838. (12) Uchida, K.; Szweda, L. I.; Chae, H. Z.; Stadtmanm, E. R. Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 8742-8746. (13) Sayre, L. M.; Zelasko, D. A.; Harris, P. L.; Perry, G.; Salomon, R. G.; Smith, M. A. J. Neurochem. 1997, 68, 2092-2097. (14) Yoritaka, A.; Hattori, N.; Uchida, K.; Tanaka, M.; Stadtman, E. R.; Mizuno, Y. Proc Natl. Acad. Sci. U.S.A. 1996, 93, 2696-2701. (15) Feng, Z. H.; Hu, W. W.; Tang, M.-S. Proc Natl. Acad. Sci. U.S.A. 2004, 101, 8598-8602. (16) Winter, C. K.; Segal, H. J.; Haddon, W. F. Cancer Res. 1986, 46, 5682-5686. (17) Cajelli, E.; Ferraris, A.; Brambilla, G. Mutat. Res. 1987, 190, 169. (18) Hussain, S. P.; Raja, K.; Amstad, P. A.; Sawyer, M.; Trudel, L. J.; Wogan, G. N.; Hofseth, L. J.; Shields, P. G.; Billiar, T. R.; Trautwein, C.; Ho¨hler, T.; Galle, P. R.; Phillips, D. H.; Markin, R.; Marrogi, A. J.; Harris, C. C. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 12770-12775. (19) (a) Douki, T.; Ames, B. N. Chem. Res. Toxicol. 1994, 7, 511518. (b) Yi, P.; Zhan, D.; Samokyszyh, V. M.; Doerge, D. R.; Fu, P. P. Chem. Res. Toxicol. 1997, 10, 1259-1265. (20) (a) Wacker, M.; Schuler, D.; Wanek, P.; Eder, E. Chem. Res. Toxicol. 2000, 13, 1165-1173. (b) Chung, F. L.; Nath, R. G.; Ocando, J.; Nishikawa, A.; Zhang, L. Cancer Res. 2000, 60, 1507-1511. (21) Fernandes, P. H.; Wang, H.; Rizzo, C. J.; Lloyd, R. S. EnViron. Mol. Mutagen. 2003, 42, 68-74. (22) (a) Hu, W. W.; Feng, Z. H.; Eveleigh, J.; Iyer, G.; Pan, J.; Amin, S.; Chung, F. L.; Tang, M. S. Carcinogenesis 2002, 23, 1781-1789. (b) Feng, Z. H.; Hu, W. W.; Amin, S.; Tang, M. S. Biochemistry 2003, 42, 7848-7854. (23) Chung, F.-L.; Chen, H.-J. C.; Guttenplan, J. B.; Nishikawa, A.; Hard, G. C. Carcinogenesis 1993, 14, 2073-2077. (24) Nishikawa, A.; Furukawa, F.; Kasahara, K.; Ikezaki, S.; Itoh, T.; Suzuki, T.; Uchida, K.; Kurihara, M.; Hayashi, M.; Miyata, N.; Hirose, M. Cancer Lett. 2000, 148, 81-86. (25) Wang, H.; Kozekov, I. D.; Harris, T. M.; Rizzo, C. J. J. Am. Chem. Soc. 2003, 125, 5687-5700. (26) Chung, F. L.; Chen, H. J.; Nath, R. G. Carcinogenesis 1996, 17, 2105-2111. (27) Alary, J.; Bravais, F.; Cravedi, J. P.; Debrauwer, L.; Rao, D.; Bories, G. Chem. Res. Toxicol. 1995, 8, 34-39. (28) (a) Chung, F. L.; Roy, K. R.; Hecht, S. S. J. Org. Chem. 1985, 53, 14-17. (b) Niu, D. Q.; Zhao. K. J. Am. Chem. Soc. 1999, 121, 24562459. (29) (a) Echols, H.; Goodman, M. F. Annu. Rev. Biochem. 1991, 60, 477. (b) Grollman, A. P. Proc. Am. Assoc. Cancer Res. 1989, 30, 682 and references therein. (30) Kornberg, A.; Baker, T. A. DNA Replication; W. H. Freeman: New York, 1992. (31) Strazewski, P.; Tamm, C. Angew. Chem., Int. Ed. Engl. 1990, 29, 36. (32) (a) Chen, X. L.; Kelso, C.; Hornak, V.; Santos, C.; Grollman, A. P.; Simmerling, C. J. Am. Chem. Soc. 2005, 127, 13906-13918. (b) Hu, X. B.; Li, H. R.; Ding, J. Y.; Han, S. J. Biochemistry 2004, 43, 63616369. (33) (a) Topal, M. D.; Fresco, J. Nature (London) 1976, 263, 285289. (b) Eistenstadt, E.; Warren, A. J.; Porter, J.; Atkins, D.; Miller, J. H. Proc. Natl. Acad. Sci. U.S.A. 1982, 79, 1945-1949. (c) Fersht, A. R.; KnillJones, J. W. Proc. Natl. Acad. Sci. U.S.A. 1981, 78, 4251-4255. (34) (a) Sinha, N. K.; Haimes, M. D. J. Biol. Chem. 1981, 256, 1067110683. (b) Hu, X. B.; Li, H. R.; Ding, J. Y.; Han, S. J. Biochemistry 2004, 43, 6361-6368. (35) Becke, A. D. J. Chem. Phys. 1993, 98, 5648-5652. (36) Becke, A. D. J. Chem. Phys. 1993, 98, 1372-1377. (37) (a) Stephens, P. J.; Devlin, F. J.; Chabalowski, C. F.; Frisch, M. J. J. Phys. Chem. 1994, 98, 11623-11627. (b) Lee, C.; Wang, W.; Parr, R. G. Phys. ReV. B 1987, 37, 785-789. (38) (a) Boys, S. F.; Bernardi, F. Mol. Phys. 1970, 19, 553-575. (b) Xantheas, S. S. J. Chem. Phys. 1996, 104, 8821-8824. (39) (a) Gonzalez, C.; Schlegel, H. B. J. Chem. Phys. 1989, 90, 2154. (b) Gonzalez, C.; Schlegel, H. B. J. Phys. Chem. 1990, 94, 5523. (40) Onsager, L. J. Am. Chem. Soc. 1936, 58, 1486-1493. (41) Dragan, M.; Popovic, J. Q.; Alexei, A. S. J. Phys. Chem. B 2005, 109, 3616-3626. (42) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A.; Stratmann, R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick, D. K.;

Prediction of the p53 Gene Mutagenic Mechanism Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz, J. V.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M. W.; Johnson, B. G.; Chen, W.; Wong, M. W.; Andres, J. L.; Head-Gordon, M.; Replogle, E. S.; Pople, J. A. GAUSSIAN 98; Gaussian, Inc.: Pittsburgh, PA, 1998. (43) Carey, F. A. Organic Chemistry, 4th ed.; McGraw-Hill: Boston, MA, 2000; pp 93-106.

J. Phys. Chem. B, Vol. 111, No. 19, 2007 5371 (44) Floria´n, J.; Leszczynski, J. J. Am. Chem. Soc. 1996, 118, 30103017 and references therein. (45) Goodman, M. F. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 1049310495. (46) Hainaut, P.; Hernandez, T.; Robinson, A.; Rodriguez-Tome, P.; Flores, T.; Hollstein, M.; Harris, C.; Montesano, R. Nucleic Acids Res. 1993, 26, 205-213. (47) Hollstein, M.; Greenblatt, M.; Soussi, T.; Hovig, E.; Montesano, R.; Harris, C. C. Nucleic Acids Res. 1996, 24, 141-146.