Estimating the vertical ionization potential of single-stranded DNA molecules

Abstract

AbstractThe electronic properties of DNA molecules, defined by the sequence-dependent ionization potentials of nucleobases, enable long-range charge transport along the DNA stacks. This has been linked to a range of key physiological processes in the cells and to the triggering of nucleobase substitutions, some of which may cause diseases. To gain molecular-level understanding of the sequence dependence of these phenomena, we estimated the vertical ionization potential (vIP) of all possible nucleobase stacks in B-conformation, containing one to four Gua, Ade, Thy, Cyt or methylated Cyt. To do this, we used quantum chemistry calculations and more precisely the second-order Møller-Plesset perturbation theory (MP2) and three double-hybrid density functional theory (DFT) methods, combined with several basis sets for describing atomic orbitals. The calculated vIP of single nucleobases were compared to experimental data and those of nucleobase pairs, triplets and quadruplets, to observed mutability frequencies in the human genome, reported to be correlated with vIP values. This comparison selected MP2 with the 6-31G* basis set as the best of the tested calculation levels. These results were exploited to set up a recursive model, called vIPer, which estimates the vIP of all possible single-stranded DNA sequences of any length based on the calculated vIPs of overlapping quadruplets. vIPer’s vIP values correlate well with oxidation potentials measured by cyclic voltammetry and activities obtained through photoinduced DNA cleavage experiments, further validating our approach. vIPer is freely available on thegithub.com/3BioCompBio/vIPerrepository.