Estudos das propriedades de inosina em DNA através do modelo Peyrard-Bishop e análise dos parâmetros termodinâmicos utilizados na predição de estruturas secundárias de RNA

Abstract

In this thesis we present different studies on the properties of nucleic acids regarding both structure and their composition. After an brief overview on deoxyribonucleic acid (DNA), ribonucleic acid (RNA), and the inosine nucleotide, given in the introduction, we start to present our results. In the first part we have the thermodynamic analysis of the physical properties of inosine pairing occurring when present in DNA. Using experimental data we were able to determine the Peyrard-Bishop parameters that describe its hydrogen bonding configurations with each type of DNA base, and the stacking interactions between inosine mismatches and regular neighbour Watson-Crick pairs. Our findings indicate two hydrogen bonds in all situations, which was also indicated in molecular dynamics studies of oligomers of hypoxantine¿the inosine nucleobase. In addition, we found that the stacking interactions between inosine mismatches are comparable with canonical interactions, with the exception of IG bases which seem to distort the DNA double helix to cause either a very strong or very weak coupling. The second work we present is a study on the stability of RNA sequences through analysis of their secondary structures, and the analysis of the influence of experimental deviation uncertainty of nearest-neighbour model parameters measurements on the prediction quality of common softwares used for secondary structure determination of RNA sequences. Using reverse engineering we were able to use secondary structure data from NMR and X-ray crystallography to determine the best set of parameters to optimize secondary structure prediction. We have found that for longer sequences, up to 800 nucleotides, the software does not guarantee a realistic prediction. We found thousands of structural variations are equally valid within the experimental thermodynamic parameters used as arguments in for these software packages. Yet most of these structures differ from the experimentally determined structures. Therefore, we hope that our study may draw the attention to the proper limits for the use of these secondary structure software packages.