. Author manuscript; available in PMC: 2011 Feb 4.
Published in final edited form as:J Phys Chem A. 2010 Feb 4;114(4):2038–2044. doi:
10.1021/jp911376p AbstractNon-covalent complexes of a tumorigenic benzo[a]pyrene diol epoxide with the guanine-cytosine and adenine-thymine base pairs have been examined computationally. (+)-BaP DE-2 forms covalent adducts with DNA via nucleophilic attack on the (+)-BaP DE-2 epoxide. Computational results predict five thermodynamically accessible complexes of AT with (+)-BaP DE-2 that are compatible with intact DNA. Among these, two are expected to lead to adenine adducts. In the lowest energy AT...(+)-BaP DE-2 complex, which has a gas-phase interaction energy of −20.9 kcal mol−1, the exocyclic NH2 of adenine is positioned for backside epoxide attack and formation of a trans adduct. The most energetically favorable complex leading to formation of a cis ring-opened adduct lies only 0.6 kcal mol−1 higher in energy. For GC...(+)-BaP DE-2, there are only two thermodynamically accessible complexes. The higher-lying complex, bound in the gas phase by 24.4 kcal mol−1 relative to separated GC and (+)-BaP DE-2, would lead to a trans ring opened N2-guanine adduct. In the global minimum energy GC...(+)-BaP DE-2 complex, bound by 27.3 kcal mol−1, the exocyclic NH2 group of cytosine is positioned for cis epoxide addition. However, adducts of (+)-BaP DE-2 with cytosine are rarely observed experimentally. The paucity of cytosine adducts, despite the predicted thermodynamic stability of this GC...(+)-BaP DE-2 complex, is attributed to the electrostatic destabilization of the benzylic cation intermediate thought to precede cis addition.
I. IntroductionPolycyclic aromatic hydrocarbons (PAHs), a main component of soot, are widely distributed in the environment due primarily to the combustion of biomass and fossil fuels.1 These PAHs form in flames via well-characterized pathways starting with the formation of an initial aromatic ring (usually benzene),2-4 and are known to be carcinogenic.3-7 Similarly, components of tobacco smoke have been shown to cause cancer, though the complexity of the underlying mechanisms has rendered a complete atomic-level understanding of the process elusive. Among the carcinogenic and tumorigenic components of tobacco smoke, PAHs are known to play a major role.8 Benzo[a]pyrene (BaP) is one of the most thoroughly studied carcinogenic PAHs in tobacco smoke, and also one of the most abundant.9
In mammals, BaP is metabolized to a bay region diol epoxide, BaP DE. Four stereoisomers of BaP DE are formed (Fig. 1a), of which (+)-BaP DE-210 is both the most abundant and most tumorigenic.11 Intercalation of (+)-BaP DE-2 into DNA is followed by nucleophilic attack on the epoxide by the exocyclic NH2 of adenine or guanine.12 Guanine adducts predominate, although adenine adducts can also be formed in significant quantities.13-15 Yields of cytosine adducts are much smaller,13 and the structures of these adducts were not definitively characterized until recently.16 Nucleophilic attack on the epoxide leads to a covalently bound DNA adduct via either cis or trans ring opening.17 Cis addition is postulated to occur via a resonance-stabilized benzylic cation intermediate,18 while trans ring-opening can proceed either from this cationic intermediate or by direct backside attack on the epoxide. Covalent DNA adducts inhibit enzymes such as helicase19 and topoisomerase I.20 These DNA adducts exhibit diverse, sequence-dependent mutagenic behavior, a feature that has been attributed to the presence of different stable adduct conformations.21-26 Tobacco-smoke-related cancers are typically due to mutations in the p53 tumor suppressor gene arising from the preferential addition of (+)-BaP DE-2 to specific sequences in this gene.6
Figure 1.(a) Four stereoisomers of the benzo[a]pyrene diol epoxide; (b) major trans and cis guanine and adenine (+)-BaP DE-2 adducts
Most previous investigations of the effects of (+)-BaP DE-2 on DNA, both experimental6,27,28 and theoretical,29-34 have focused on the covalent adducts with DNA oligomers or single nucleotides. Of particular recent interest has been the characterization of the conformational behavior of (+)-BaP DE-2–DNA adducts. NMR solution data have confirmed the presence of multiple stable conformations depending on the flanking nucleobases.35,36 In 2004, the first crystal structures of BaP DE-2–DNA adducts were published27,37 providing, along with previously published NMR structures,37-40 invaluable details regarding the structure of these adducts. A crystal structure of an adduct with BaP DE bound to N2-deoxyguanosine enabled the assignment of absolute configurations to the four optically active BaP DE isomers and the eight associated dG adducts resulting from the cis and trans ring opening of the epoxide.37 A second crystal structure of a BaP DE-2–adenine adduct in a ternary complex with DNA polymerase was published shortly thereafter.27
While information about (+)-BaP DE-2...DNA complexes can be inferred from the structures of these adducts, a complete understanding of the addition of (+)-BaP DE-2 to DNA will require explicit examination of the complexes that precede epoxide attack. Harvey and co-workers41 studied the formation of non-covalent intercalative complexes of (+)-BaP DE-2 and (−)-BaP DE-1 with DNA via kinetic flow linear dichroism experiments. Structural differences between these non-covalent complexes were highlighted that presumably underlie the different mutagenic and tumorigenic activities of these diastereomers. Computational methods are ideally suited to provide additional details regarding non-covalent (+)-BaP DE-2···DNA complexes, because it is possible to directly probe their structures and thermochemistry and to quantify the role of π-stacking and individual hydrogen bonding interactions. Previously, the inability of popular density functional theory (DFT) functionals42,43 to accurately describe the π-stacking interactions that drive the intercalation of (+)-BaP DE-2 into DNA hampered high-level computational studies of such systems. Recent advances in the development of DFT functionals44,45 have led to methods that may accurately describe π-stacking, opening the door for quantum mechanical studies of stacking phenomena in myriad biological contexts.
Several DNA intercalators have previously been examined by ab initio and DFT methods,46-48 with a focus on intercalators that are utilized in antitumor chemotherapy. In 2002, Hobza and co-workers47 studied the π-stacking interactions of ethidium, daunomycin, ellipticine, and 4,6′-diaminide-2-phenylindole with DNA base pairs. It was shown that in each of these cases the net attraction arises from the competing effects of electrostatic and dispersion interactions and short-range exchange repulsion. In 2006, Leszczynski and co-workers48 examined the nature of interactions between ethidium and proflavine with DNA bases starting from published crystal structures. A more recent study46 of the interaction of ellipticine and proflavine with DNA base pairs showed that stabilizing interactions are maximized when the main axis of the intercalator is nearly aligned with the main axis of the bases. In contrast to (+)-BaP DE-2, these previously studied intercalating agents interact with DNA solely through π-stacking interactions, rather than a combination of intermolecular hydrogen bonds and π-stacking.
Despite decades of study of the effects of BaP DE on DNA, there have been few detailed explorations of the non-covalent (+)-BaP DE-2...DNA complexes that precede adduct formation.41 Consequently, there are details concerning the complexation of (+)-BaP DE-2 with DNA base pairs that warrant further exploration. A lingering conundrum concerns the scarcity of observed cytosine adducts,16 despite the prevalence of adducts of cytosine with diol epoxides derived from dibenz[a,j]anthracene and benz[a]anthracene.49-51 Presumably, this could arise from disfavored non-covalent interactions between cytosine and (+)-BaP DE-2, an insurmountable free energy barrier for epoxide attack by cytosine, or some combination of these two factors. Similarly, the energetically favored orientation of (+)-BaP DE-2 relative to a given base pair prior to adduct formation has not been established. It has been proposed21-26 that the existence of multiple conformations of (+)-BaP DE-2–DNA adducts is responsible for their divergent mutagenic behavior. The geometry of the complex preceding adduct formation could presumably play a role in the formation of these different conformers. Finally, understanding the role of individual π-stacking and hydrogen bonding interactions in these complexes will help further unravel the unique features of (+)-BaP DE-2 that lead to its pronounced tumorigenicity. As a first step toward addressing these issues, we have examined the complexes of (+)-BaP DE-2 with the AT and GC base pairs to gain insight into the factors governing the addition of (+)-BaP DE-2 to DNA.
II. MethodsAccurate ab initio descriptions of π-stacking interactions require rigorous correlated theoretical methods paired with large basis sets.52 However, (+)-BaP DE-2···DNA base pair complexes, which comprise 66 and 67 atoms with GC and AT, respectively, are too large to treat with such rigorous approaches. Unfortunately, many DFT functionals, which are the most popular methods for computational investigations of systems of this size, fail to accurately describe the dispersion effects that underlie stacking interactions.42,43 However, the M05-2X and M06-2X functionals have been shown to provide accurate interaction energies for stacked dimers.44,53-56 Hohenstein, Chill, and Sherrill57 recently showed that M06-2X in particular performs well for a standard benchmark set of stacked complexes. Also, among DFT functionals including empirical dispersion corrections, PBE-D has been shown57,58 to yield stacking energies of comparable quality to M06-2X. Similarly, Gu et al.59 recently demonstrated that the M06-2X functional paired with a double zeta basis set yields interaction energies in good agreement with reliable CCSD(T) results for stacked nucleic acid bases.
Geometry optimizations were performed for two conformers of (+)-BaP DE-2 with each of the DNA base pairs using the M05-2X functional. The potential energy surface of each base pair with each conformer of (+)-BaP DE-2 was explored by executing optimizations from several initial geometries for each relative orientation of the stacked system. Starting structures were generated by varying the position of (+)-BaP DE-2 relative to the base pair in ~2 Å increments while keeping the molecular planes of the two species parallel, to ensure that all low-lying configurations were sampled. Preliminary optimizations were carried out using the M05-2X functional paired with the 3-21G basis set. Structures lying within three kcal mol−1 of the predicted global minimum were then further refined using the 6-31+G(d) basis set. To verify the M05-2X optimized structures, M06-2X/6-31+G(d) optimizations were also carried out on the minimum energy and second lowest-lying AT and GC complexes. Differences between the M05-2X and M06-2X geometries were minor. Presented gas-phase interaction energies are M06-2X/6-31+G(d) electronic energies evaluated at M05-2X/6-31+G(d) geometries, and are given relative to separated base pair and conformer I of (+)-BaP DE-2. PBE-D/aug-cc-pVDZ single point energies45 were also evaluated at M05-2X geometries for the two lowest lying AT and GC complexes. The PBE-D/aug-cc-pVDZ interaction energies for the AT complexes are similar to the M06-2X/6-31+G(d) data, differing by ±0.6 kcal mol−1. On the other hand, the PBE-D interaction energies for the two lowest-lying GC complexes are smaller than the M06-2X values by 3.5 and 1.9 kcal mol−1. In both cases, the energy ordering of the complexes is unchanged.
All computations were performed using NWChem.60,61 For the M05-2X and M06-2X computations a fine DFT integration grid was used, consisting of 70 radial shells and 590 angular points, since these functionals are known to be sensitive to the choice of integration grid.42,62-64 Standard atomic labels are utilized for the nucleobases, which are depicted in Fig. 2. Relevant atom designations for (+)-BaP DE-2 are shown in Fig. 1a and denoted by subscripts in the text.
Figure 2.Canonical atomic labels for GC and AT base pairs.
III. Results and Discussion A. Conformers of Free (+)-BaP DE-2In the gas phase, there are two low-lying conformers of (+)-BaP DE-2 (I and II, Fig. 3a). In the higher-lying conformer (II), the hydroxyl groups occupy pseudo-axial positions, whereas the OH groups in I are equatorial. The M06-2X/6-31+G(d)//M05-2X/6-31+G(d) predicted energy difference is 2.1 kcal mol−1, with a conformational barrier of 7.8 kcal mol−1. Associated with each of these conformers are other low-lying minima, connected to I and II via changes in OH orientations. Both conformers I and II are stabilized by intramolecular hydrogen bonds that must be broken during the change in ring conformation. Such interactions will be less important in an aqueous environment, and the conformational barriers will be smaller in solution. Most importantly, both conformers I and II will be thermodynamically accessible and rapidly inter-converting at biologically relevant temperatures.
Figure 3.(a) Two low-lying gas-phase conformers of (+)-BaP DE-2. At the M06-2X/6-31+G(d)//M05-2X/6-31+G(d) level of theory, conformer II lies 2.1 kcal mol−1 higher than I, separated by a conformational barrier of 7.8 kcal mol−1. (b) Low-lying complexes of (+)-BaP DE-2 with the GC base pair. (c) Lowest-lying GC complex compatible with cis guanine addition. In (b) and (c), gas-phase interaction energies relative to separated GC and conformer I of (+)-BaP DE-2 are given in kcal mol−1. Hydrogens connected to N1 and N9 are highlighted in green. Hydrogen bond distances are in Angstroms.
B. (+)-BaP DE-2...DNA Base Pair ComplexesGas-phase interaction energy surfaces for (+)-BaP DE-2 with both the GC and AT base pairs have been examined as a model for the interaction of (+)-BaP DE-2 with DNA. When (+)-BaP DE-2 approaches a given base pair,65 there are four possible relative orientations: the epoxide functionality can be directed toward or away from the base pair and the functionalized end of (+)-BaP DE-2 can extend into the minor or major groove of intact DNA. Additional variation arises from the complexation of (+)-BaP DE-2 with AT or TA and GC or CG. For each of these arrangements, multiple relative positions of the centers of mass of the base-pair and (+)-BaP DE-2 were considered to ensure that the minimum energy complex has been obtained for each relative orientation. In total, 48 distinct optimizations were carried out for each conformer.
The optimized structures (Figs. 3b and 4) are named according to the relative orientation of (+)-BaP DE-2 and the base pair as follows: the orientation of the epoxide towards or away from the base pair is indicated by E or e, respectively. M or m indicates that the functionalized end of (+)-BaP DE-2 extends into the major or minor groove, respectively. Finally, the ordering of the base pair (e.g.: GC versus CG) is explicitly given. For example, the global minimum GC...(+)-BaP DE-2 complex is labeled CG(I-EM), signifying that (+)-BaP DE-2 is complexed with the epoxide functionality directed towards cytosine-guanine, with the functionalized end extending into the major groove of DNA. It should be noted that complexes in which the epoxide is facing the base pair (denoted by E) can only lead to cis adducts, while structures in which the epoxide is facing away from the base pair (e) are only compatible with trans addition.
Figure 4.Low-lying complexes of (+)-BaP DE-2 with the AT base pair. Gas-phase interaction energies relative to separated AT and conformer I of (+)-BaP DE-2 are given in kcal mol−1. Hydrogens connected to N1 and N9 are highlighted in green. Hydrogen bond distances are in Angstroms.
One of the primary goals of the current work is to gain insight into the non-covalent complexes of (+)-BaP DE-2 with DNA that precede covalent adduct formation. Stacking of a single DNA base pair with (+)-BaP DE-2 serves as the simplest possible model of the interaction with DNA. Rather than constraining optimizations to be compatible with intact DNA, unconstrained optimizations were executed and final structures incompatible with DNA were eliminated. Since the 2-deoxyribose of DNA was replaced with a hydrogen atom in our model, eliminated structures include five complexes in which the purine N9 or pyrimidine N1 act as hydrogen bond donors. Also excluded were two structures in which the base pair undergoes significant distortions that would be improbable in intact DNA due to backbone constraints and the presence of flanking base pairs. One such structure [AT(I-eM)] is depicted in Fig. 5. Some of the excluded structures were otherwise competitive energetically with the global minimum complexes.
Figure 5.Two views of a low-lying complex of conformer I of (+)-BaP DE-2 with the AT base pair. This structure is not compatible with intact DNA, since the base-pair undergoes significant distortion in order to form a strong hydrogen bond with conformer I of (+)-BaP DE-2 while maintaining favorable stacking interactions with the pyrene. The gas-phase interaction energy relative to separated AT and (+)-BaP DE-2 is given in kcal mol−1. Hydrogens connected to N1 and N9 are highlighted in green.
C. Guanine-Cytosine...(+)-BaP DE-2 ComplexesThe present results indicate that (+)-BaP DE-2 will preferentially form complexes with the guanine-cytosine base pair; the global minimum GC...(+)-BaP DE-2 complex (Fig. 3b) lies 6.4 kcal mol−1 lower in energy than the global minimum energy AT...(+)-BaP DE-2 complex (Fig. 4) at the M06-2X/6-31+G(d) level of theory. PBE-D/aug-cc-pVDZ predicts a smaller energy difference of 2.4 kcal mol−1. In the global minimum GC structure [CG(I-EM)], conformer I of (+)-BaP DE-2 is complexed with the epoxide functionality directed towards CG. The gas-phase interaction energy relative to separated GC and conformer I of (+)-BaP DE-2 is 27.3 kcal mol−1. This complex is stabilized by two somewhat strained hydrogen bonds joining the epoxide oxygen with the exocyclic NH2 of cytosine and the C7 hydroxyl group with O6 on guanine. In order to maintain these intermolecular hydrogen bonds and the favorable stacking interaction of cytosine with the aromatic core of (+)-BaP DE-2, the base pair is distorted slightly from planarity. In intact DNA there will be an associated energetic cost due to interactions with flanking base pairs and this structure will lie higher in energy in a more complete DNA model. The oxidized end of (+)-BaP DE-2 interacts with the major groove; as a result, formation of a covalent guanine adduct is impossible, since the exocyclic NH2 of guanine is directed away from the epoxide. Instead, this complex is compatible with formation of an N4-cytosine adduct via cis ring opening of the epoxide, which only form in small quantities.16
There is only one additional thermodynamically accessible GC...(+)-BaP DE-2 complex [CG(II-em)], lying 2.9 kcal mol−1 above the global minimum. In this case, it is the higher-lying conformer (II) that is complexed with CG, with the epoxide functionality directed away from the base-pair and the functionalized end in the minor groove. As such, this complex is poised for back-side nucleophilic attack and formation of the frequently observed trans N2-guanine adducts. This complex is stabilized by two cooperative hydrogen bonds involving the C7 OH group and the exocyclic NH2 and N7 of guanine.
D. Adenine-Thymine...(+)-BaP DE-2 ComplexesEven though the minimum-energy AT...(+)-BaP DE-2 complex lies 6.4 kcal mol−1 higher in energy than the global minimum GC...(+)-BaP DE-2 complex, it is important to consider the AT case since covalent adenine adducts frequently form. In contrast to the GC...(+)-BaP DE-2 complexes, for which there was only one structure within three kcal mol−1 of the global minimum, for AT...(+)-BaP DE-2 there are five thermodynamically accessible complexes that are compatible with intact DNA.
The two lowest-lying AT...(+)-BaP DE-2 complexes, AT(II-eM) and TA(II-em), involve conformer II of (+)-BaP DE-2 with the epoxide functionality directed away from the base pair. These complexes are bound by 20.9 and 20.4 kcal mol−1 in the gas phase, respectively. In AT(II-eM), (+)-BaP DE-2 is complexed with AT with the functionalized end directed into the major groove, while in TA(II-em) the complex involves TA and the functionalized end of (+)-BaP DE-2 is directed towards the minor groove. Consequently, in AT(II-eM) the exocyclic NH2 is positioned for backside attack and trans adduct formation, while no covalent adduct could form from TA(II-em).
A complex of conformer I with AT was also optimized [AT(I-eM)], and it is 0.5 kcal mol−1 lower than AT(II-eM). However, as seen in Fig. 5, when complexed with I, the AT base pair is drastically distorted from planarity to maintain the cyclic hydrogen bonding arrangement between the C7 OH group on (+)-BaP DE-2 and atoms N6 and N7 of adenine and stacking interaction between thymine and the pyrene. As mentioned above, these extreme distortions of the AT base pair are incompatible with the structure of intact DNA and such complexes are not expected to occur in intact DNA.
There are three other structures that are within 3 kcal mol−1 of AT(II-eM), with the most favorable of these [AT(I-EM)] higher in energy by only 0.6 kcal mol−1. This is the lowest-lying AT complex with conformer I that is compatible with intact DNA. In AT(I-EM), (+)-BaP DE-2 is oriented with the epoxide facing the base pair and the functionalized end extending into the major groove. In this complex, both N6 and N7 are poised for front-side attack on the epoxide, leading to either cis N6-adenine adduct or the less frequently observed N7 adduct. The next higher-lying structure, TA(II-EM), is 2.2 kcal mol−1 above the global minimum and features (+)-BaP DE-2 bound with the epoxide towards TA and the functionalized end extending into the major groove. This complex is not expected to lead to adduct formation, since N6 is not near the epoxide carbon. The final structure, TA(I-Em), also involves conformer I, this time complexed with TA with the epoxide towards the base pair and the functionalized end in the minor groove. No adduct can be formed from TA(I-Em).
Structures CG(II-em) and AT(II-eM) provide a demonstration of the differential stacking avidities of (+)-BaP DE-2 with the GC and AT base pairs. These two structures exhibit similar hydrogen bonding interactions. Both include a cyclic hydrogen bonding arrangement of the C7 OH group on (+)-BaP DE-2 with the exocyclic NH2 and ring nitrogen of the purine base. The hydrogen bonds in the GC complex are slightly weaker, yet the GC complex is more strongly bound by 3.5 kcal mol−1.66 This difference is attributed to a stronger π-stacking interaction with GC over AT, which is consistent with previous findings for proflavin.46 This difference can be understood qualitatively in terms of simple electrostatic effects. Electrostatic potential surfaces for conformer II of (+)-BaP DE-2 and the GC and AT base pairs are shown in Fig. 6. These plots provide a simple tool for understanding the electrostatic component of non-covalent interactions.67-69 The primary difference in ESPs between the AT and GC base pairs is the sign of the ESP surrounding cytosine-N4 compared to O4 on thymine. In AT(II-eM) there will be an unfavorable electrostatic interaction between the negative ESP surrounding O4 on thymine and the negative ESP above the pyrene. The corresponding electrostatic interaction with N4 on cytosine will be favorable. In other words, the AT...(+)-BaP DE-2 complex is destabilized by a direct electrostatic interaction70 between thymine-O4 and the pyrene, while the analogous direct interaction with the cytosine NH2 stabilizes CG(II-em).
Figure 6.Electrostatic potentials of conformer II of (+)-BaP DE-2, guanine-cytosine, and adenine-thymine, mapped onto electron density isosurfaces (ρ = 0.001 e/au3).
E. Implications for DNA Adduct FormationThe present computations employ a simple model of the interactions of BaP DE-2 with DNA consisting of gas-phase complexes between (+)-BaP DE-2 and a single base pair. Regardless, some insight into the formation of covalent adducts between (+)-BaP DE-2 and DNA can be gleaned from the computed structures. The present results predict that the higher-lying conformer of (+)-BaP DE-2 (conformer II) leads to formation of trans adducts with both guanine and adenine, via complexes CG(II-em) and AT(II-eM), respectively. Conformer I is predicted to undergo cis addition to yield an adenine adduct via structure AT(I-EM). The lowest-lying GC complex leading to cis addition of guanine (Fig. 3b) lies 6.4 kcal mol−1 above the global minimum, which is isoenergetic with the minimum energy AT complex. The global minimum GC structure [CG(I-EM)] will not lead to guanine adduct formation, and is separated from the next lowest-lying structure by 2.9 kcal mol−1. This is in contrast to the AT complexes, for which both the lowest-lying structure [AT(II-eM)] and a second complex 0.6 kcal mol−1 higher in energy [AT(I-EM)] are expected to both lead to covalent adducts.
The prediction that the thermodynamically preferred complex, CG(I-EM), is positioned for formation of an N4-cytosine adduct is inconsistent with the paucity of observed cytosine adducts.16 As noted above, in this structure there is some distortion of the base pair from planarity and it would likely lie slightly higher in energy in a more complete model of DNA. Additionally, the effects of solvent, the sugar-phosphate backbone, or the presence of a sandwiching base pair could all alter the predicted energetic ordering in a more realistic model of DNA. On the other hand, there could be some mechanistic origin preventing cytosine attack. One possibility arises from differences in the electrostatic potential surrounding cytosine compared to guanine or adenine. The ESP surrounding cytosine is mostly positive (see Fig. 6), in contrast to the more negative ESP surrounding the purine bases. Front-side epoxide attack is postulated to occur through a fleeting benzylic cation intermediate.18 The positive ESP surrounding cytosine will destabilize this incipient cation, raising the energy of this intermediate and associated reaction barriers and preventing cytosine addition to (+)-BaP DE-2.
The two low-lying conformers of (+)-BaP DE-2 lead to qualitatively different intermolecular hydrogen bonds with the GC and AT base pairs. More importantly, the complexes leading to trans adduct formation [CG(II-em) and AT(II-eM)] involve conformer II, while conformer I is present in the complexes compatible with cis adenine and guanine addition [AT(I-EM) and CG(I-Em)]. These latter complexes feature a hydrogen bond between the exocyclic NH2 of the purine base and the epoxide oxygen. Additional stabilization is achieved via hydrogen bond contacts with the OH connected to C7, which are only possible for conformer I. Conversely, for structures leading to trans adducts, the exocyclic NH2 is hydrogen bonded to the C7 hydroxyl group. Forming this hydrogen bond while simultaneously maintaining favorable stacking interactions with the pyrene and a planar base pair arrangement requires conformer II.
The trans arrangement in (+)-BaP DE-2 enables this OH group to position the AT base pair with the NH2 group ideally oriented for backside attack on the epoxide. Overall, the tendency for formation of a strong hydrogen bond with the this OH group combined with the maximization of π-stacking interactions between the pyrene moiety and the DNA base pairs drive the complexes to adopt arrangements that are pre-organized for epoxide attack and DNA-adduct formation.
IV. Summary and Conclusions(+)-BaP DE-2 is a tumorigenic metabolite of benzo[a]pyrene, a polycyclic aromatic hydrocarbon found in soot and tobacco smoke. (+)-BaP DE-2 forms covalent DNA adducts by intercalation into DNA followed by the nucleophilic attack of the (+)-BaP DE-2 epoxide by the exocyclic NH2 group of either adenine or guanine. These covalent adducts interfere with key DNA processes, leading to sequence-specific mutations.
Complexes of the AT and GC DNA base pairs with two low-lying conformers of (+)-BaP DE-2 have been examined using DFT methods to gain insight into the factors governing adduct formation. Formation of complexes with the GC base pair is favored over AT, due to stronger π-stacking interactions between the pyrene core of (+)-BaP DE-2 and guanine-cytosine. Only one GC...(+)-BaP DE-2 complex is predicted to lie within 3 kcal mol−1 of the global minimum energy GC...(+)-BaP DE-2 structure, which is bound in the gas phase by 27.3 kcal mol−1. In this global minimum complex [CG(I-EM)], the exocyclic NH2 group of cytosine is positioned for cis addition to the epoxide of (+)-BaP DE-2. The scarcity of experimentally observed cytosine adducts arising from this thermodynamically favorable complex are explained based on a simple electrostatic model; the region surrounding cytosine in the GC base pair has a positive electrostatic potential, which will destabilize the cationic intermediate leading to cis addition to the epoxide. The second lowest lying GC complex is compatible with backside attack on the epoxide by guanine, leading to the observed trans ring-opened N2-guanine adducts. We predict five thermodynamically accessible AT...(+)-BaP DE-2 complexes that are compatible with intact DNA. The most energetically favored complex [AT(II-eM)], lies 6.4 kcal mol−1 higher in energy than the global minimum GC...(+)-BaP DE-2 complex. Among these low-lying AT complexes, two structures are compatible with adenine adduct formation, accounting for both the cis and trans adducts.
There are two stable conformers of free (+)-BaP DE-2, and when bound to either the AT or GC base pair these are competitive energetically. Complexes involving the lower-lying conformer (I) lead to cis ring-opened adducts, while trans adducts are predicted to form from complexes featuring the higher-lying conformer (II).
Stable non-covalent complexes between DNA base pairs and (+)-BaP DE-2 arise from a combination of intermolecular hydrogen bonds and π-stacking interactions. Because of the trans arrangement of the C7 OH group and epoxide on (+)-BaP DE-2, in many of the low-lying complexes the base pair is oriented with the exocyclic NH2 group pre-organized for epoxide attack. The energetic tendency to maximize π-stacking interactions further orients the base pair for epoxide attack.
Supplementary Material1_si_001
AcknowledgementsThis work was supported by the Department of Energy Combustion Program and the National Science Foundation (J.C.H. and H.F.S.), National Institute of General medical Sciences (NIGMS), National Institutes of Health (NIH) grant 1R01GM075962 (K.N.H), and NIGMS, NIH grant 5F32GM082114-02 (S.E.W.). S.E.W. would also like to thank E. C. P. Hohenstein for stimulating discussions and J. M. Sayer for helpful suggestions regarding cytosine adducts. J.C.H. thanks K. Morris for comments on the manuscript. Molecular figures were generated using CYLview.71 Computer time was provided in part by the National Energy Research Scientific Computing Center (NERSC), which is supported by the Office of Science of the U.S. Department of Energy under Contract No. DE-AC02-05CH11231, and the UCLA Institute for Digital Research and Education (IDRE).
FootnotesSupporting Information Available. Cartesian coordinates; absolute energies.
ReferencesThis section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials1_si_001
RetroSearch is an open source project built by @garambo | Open a GitHub Issue
Search and Browse the WWW like it's 1997 | Search results from DuckDuckGo
HTML:
3.2
| Encoding:
UTF-8
| Version:
0.7.4
