Theoretical studies of the structure and energies of base-paired nucleotides and the dissociation kinetics of a proflavine-dinucleotide complex.

We presented calculations of base-paired dinucleoside phosphates and hexanucleoside pentaphosphates of varying compositions. Complete energy minimizations were performed for (a) the ten base-pair combinations of dinucleoside phosphates, starting from a B-DNA conformation, (b) six hexanucleoside pentaphosphates--base-paired CGCGCG, GCGCGC, G6-C6, TATATA, ATATAT, and A6-T6--starting with a B-DNA geometry, and (c) the four hexanucleoside pentaphosphates that have alternating pyrimidine-purine sequences, starting with a Z-DNA geometry. In addition, we studied the proflavine-base-paired CpG complex, using both complete energy minimization and energetic constraints to force the drug to dissociate from the dinucleoside phosphate. In many of these calculations, we examined the dependence of the calculated energies and structures on the potential function, focusing mainly on the effect of nonbonded potentials, the effective dielectric constant, and the role of counterions. These calculations allow us to explain why pur-(3',5')-pyr sequence isomers are more stable than pyr-(3'-5')-pur isomers. Both base-base and base-backbone energies are important in this differentiation, with the former being mainly van der Waals attraction and the latter mainly electrostatic energies. The calculations also allow us to understand the differences in double helical stabilities found by Wells et al. These differences, caused by electrostatic interactions between those bases not Watson-Crick hydrogen bonded, allow us to explain the following experimental data: poly(dG-dC) melts 12 degrees C higher than poly dG-poly dC, poly(dA-dT) melts 6 degrees C lower than poly dA-poly dT, and poly(dA-dG)-poly(dC-dT) melts 6 degrees C lower than poly(dA-dC)-poly(dT-dG). These results have interesting implications for drug binding: they imply that simple intercalators, such as ethidium, will exhibit a greater affinity for hetero- than for homopolymers and that this preference will be greater in the AT polymers than it is in the GC polymers. Our calculations allow us to explain the fact that Z-DNA is more stable than B-DNA under high salt conditions and to suggest some sequence dependence for the Z to B transition. We found that the activation energy for proflavine dissociating from dCpG is almost equal to the dissociation energy.