Target characterization and ligand design directed at deoxyuridine 5'-triphospate nucleotidohydrolase

Abstract

dUTPase is an important enzyme in DNA metabolism. It can be thought of as a gate keeper for apoptosis and is therefore an attractive target when trying to kill malign cells which cause disease in the human body. dUTPase has been found to be an important drug target in diseases such as cancer and malaria, to mention two. Due to problems with drug resistance in existing treatments, the search for new and more efficient inhibitors against dUTPase is very relevant in present-day drug design. Computational methods play an important role in the development of novel inhibitors and can reduce the time and cost of the drug design process. These methods are applied to sample the binding modes of ligands to the receptor. To make the search base for ligands smaller, one must be able to rank the ligands with respect to the binding affinities. However, in order to integrate computational methods into the drug design process, suitable computational tools are needed, that can efficiently carry out and conduct ligand binding free energy calculations. As a part of this work, a graphical user interface for the simulation tools was developed. Molecular dynamics simulations and docking have been combined with the free energy calculations to predict the binding affinity for inhibitors against the human dUTPase using Qgui. A LIE model was first constructed to reproduce the experimental binding free energies for a training set of ligands. Judged by the coefficient of determination (R2), good agreement between the experimental and theoretical binding affinities was obtained. The model was tested for selectivity with a set of ligands targeted at Plasmodium falsiparum dUTPase showing the pivotal importance of good starting structure. The binding free energies of two novel ligands were measured with the LIE model and the calculations resulted in moderate binding affinities.