see designations in Table 5

vibrational modes.

Mutations in DNA: A Novel Quantum-Chemical Insight into the Classical Understanding 81

Base pair -ΔEint ΔΔE ΔΔG Ade·Thy 14.92 - - Ade\*·Thy\* 33.80 0.11 -1.01 Gua·Cyt 29.28 - - Gua\*·Cyt\* 22.94 5.15 1.49 Ade·Cyt\* 15.73 - - Ade\*·Cyt 23.50 6.44 3.75 Gua·Thy\* 33.39 - - Gua\*·Thy 19.82 4.16 1.17

Table 6. Electronic and Gibbs free energies (in kcal/mol) (T=298.15 K) of base pairs obtained

Fig. 6. Interconversion of Ade·Cyt\*↔Ade\*·Cyt and Gua\*·Thy↔Gua·Thy\* mispairs involving mutagenic tautomers of DNA bases. Relative Gibbs free energies (T=298.15 K, in vacuum) are obtained at the MP2/6-311++G(2df,pd)//B3LYP/6-311++G(d,p) level of theory and reported near each structure in kcal/mol. The dotted lines indicate H-bonds AH…B (their lengths H…B are presented in angstroms), while continuous lines show covalent bonds

The obtained Gibbs free energies of interaction indicate that Gua\*·Thy and Ade·Cyt\* are more favorable than Gua·Thy\* and Ade\*·Cyt. It was established that the Ade\* Cyt and Gua\*·Cyt\* base pairs are metastable and easily (i.e., without facing significant barrier) "slip" into the energetically more favorable Ade Cyt\* and Gua·Cyt base pairs, respectively. The comparison of reverse electronic barriers of interconversion with the zero-point energies of competent vibrational modes (Table 7) of the tautomerized complexes allows concluding that Ade\*Thy\* and GuaThy\* complexes are dynamically unstabletheir electronic barriers of the reverse transition are noticeably lower than zero-point energy of corresponding

at the MP2/6-311++G(2df,pd)//B3LYP/6-311++G(d,p) level of theory in vacuum#

those of the Watson–Crick base pairs. These values for the irregular base pair as distinguished from the Watson–Crick base pairs reflect the distortion of double helix conformation and can be factor taking into account the recognition of the structural invariants of the sugar-phosphate backbone by the polymerase.

Detailed study of the geometric characteristics for the optimized mutagenic and Watson– Crick base pairs leads to the following results. The distance between the bonds joining the bases to the deoxyribose groups in the Gua\*·Thy and Gua·Thy\* mutagenic base pairs is close to the corresponding canonical distance in the Gua·Cyt base pair, and the corresponding distance in the Ade\*·Cyt and Ade·Cyt\* base pairs is close to that in the Ade·Thy base pair. Moreover, in each pair of stereoisomers (Gua\*·Thy, Gua·Thy\* and Ade\*·Cyt, Ade·Cyt\*), the N9–C1-C1 and N1–C1–C1 glycosidic angles are close to the corresponding value in one of the Watson–Crick canonical base pairs. Analogous conclusions were made earlier by Topal and Fresco (Topal & Fresco, 1976) and Danilov et al. (Danilov et al., 2005), who studied each of the above-mentioned mutagenic base pairs by model building and by ab initio methods, respectively, and showed that these pairs are sterically compatible with the Watson–Crick base pairs.

Finally, according to the molecular mechanism of recognition of the complementary base pairs of nucleic acids by DNA polymerase (Li & Waksman, 2001), the key role in the selection of the correct substrate is the interactions of the certain amino acid residues in the recognition site of DNA polymerase with the invariant arrangement of the N3 purine and O2 pyrimidine atoms (Beard & Wilson, 1998, 2003; Poltev et al., 1998). These hydrogenbonding interactions may provide a means of detecting misincorporation at this position. Our data show that the structural invariants of the mutagenic nucleotide pairs are very close to those of the correct nucleotide pairs. In other words, the mutual position of the atoms and atomic groups is practically the same both for the correct and the irregular pairs, so that the DNA polymerase (more exactly its recognizing site) can play the role of additional matrix under the inclusion of the nucleotides. Therefore, we conclude that the formation of the DNA mutagenic base pairs satisfies the geometric constraints of the standard double helical DNA. If these mutagenic base pairs would be incorporated into a standard Watson–Crick double helix, the helix would not likely experience significant distortion and its stability would not be greatly deteriorated.

The comparison of the formation energies of the canonical and mutagenic base pairs (Table 6) shows that the Löwdin's Ade\*·Thy\* base pair, which electronic formation energy is -33.80 kcal/mol, is the most stable among all the studied base pairs. At the same time, the formation of the Gua\*Thy and Ade\*·Cyt mispairs is more favorable than that of the AdeThy canonical base pair, GuaThy\* and Ade·Cyt\* mispairs which have -14.92; -33.39 and -23.50 kcal/mol formation energy, respectively (Table 6). From the other point of view, it may evidence that dissociation of the Gua\*Thy and Ade\*·Cyt mispairs will be complicated during the strand separation. These data therefore confirm that Ade·Cyt\* and Gua\*·Thy mispairs are suitable candidates for the spontaneous point mutations arising in DNA (Fig. 6). The Ade\*·Cyt and Gua·Thy\* lifetimes (3.4910-11 s and 3.5910-13 s, accordingly) are too short comparably with the time of one base pair dissociation during the enzymatic DNA replication (10-9 s). This means that these mispairs will "slip away" from replication machinery: they transfer to Ade·Cyt\* and Gua\*·Thy accordingly (Fig. 6). In this way Ade\*·Cyt and Gua·Thy\* mispairs act as intermediates in this reaction.

