Ab initio studies of targets for pharmaceutical intervention

In this thesis we further explore the capability of first principle methods to provide insights on drug/target interactions in different contexts. In the first part of this work, we address the issue whether OFT methods can be used as a potential tool for drug-screening. First principle calculations are particularly interesting for screening the energetics of drug/target interactions, as they do not involve the painstaking procedure of developing each set of new parameters for each novel drug. In this context, we use ab initio_ methods as a novel tool to determine a scoring function in a series of prodrug I target (herpes simplex type 1 thyimidine kinase) complexes for gene-therapy based anticancer approaches. This work, accompanied by experimental data provided by Prof. Folkers' Lab (ETH, Zurich) provides a new, very simple, ab initiobased approach to the construction of scoring functions for drug-screening. In the second part of the thesis we investigate the capability of OFT to describe non trivial interactions which are encountered in several inhibitor/enzyme complexes of pharmaceutical interest. Clearly, the description of these non-trivial phenomena might require the use of electronic structure methods. Here we present an example of cation-n interaction found in the human immunodeficiency virus reverse transcriptase (HIV-1 RT), one of the major targets for anti-AIDS therapy(Furman et al., 2000)). Furthermore, we provide a description of the hydroxyl-n interactions in the active site of μ-glutathione S-transferase(Xiao et al., 1996) (μ-GST), whose differential expression has been implicated in the development of cancers as well as their resistance to chemotherapeutic drugs ((Mccallum et al., 2000) and reference therein). Finally we present a classic problem treated by quantum-chemical methods: the simulation of an enzymatic reaction. We focus on a class of cysteine proteases, the caspases. These enzymes are extremely important targets for pharmaceutical intervention in therapies against Alzheimer's and other neurodegenerative processes, yet very few inhibitors have been so far designed. Since an important class of inhibitors is the given by the transition state analogs, it is of importance to fully understand the · enzymatic reaction, that is the hydrolysis of peptides. Because of the crucial importance of temperature and environment(Karplus, 2000; Glennon and Warshel, 1998; Varnai and Warshel, 2000; Villa et al., 2000) effects for enzymatic catalysis, we use here a hybrid Car-Parrinello Molecular dynamics I Molecular mechanics approach recently developed in the Lab of Prof. U. Roethlisberger (Laio et al., 2001 ). This technique allows to evaluate the intermolecular interactions at the active site from electronic structure calculations as the simulation proceeds(Car and Parrinello, 1985). Steric and electrostatic effects of the protein scaffold on the quantum region are included using classical MD approach on the rest of the system. The free energy of the process is calculated using a thermodynamic integration approach(Ciccotti et al., 1989; Carloni et al., 2000; Piana et al., 2001). This thesis is organized as follows. The first chapter provides a description of the used computational techniques. The following chapter describes the systems investigated here and summarizes our findings. The subsequent three chapters are devoted to a - detailed description of my thesis work. In a final chapter we draw some conclusions and provide a perspective for possible future applications, which could follow this work.