Computational approaches to study mechanisms of regulation and inhibition of enzymes involved in phospho-transfer reactions

Protein kinases are the enzymes in the cell that catalyze phosphorylation reactions. They are essential for almost all cellular processes and many of them are considered promising pharmaceutical targets since they are involved in a large number of tumorigenic functions such as proliferation, immune evasion, anti-apoptosis, metastasis and angiogenesis. The progress in high-resolution structure determination techniques has contributed enormously to a better understanding of the structural basis of kinase regulation and the associated structural plasticity. However, because of the high sequence and structural conservation across the kinome, new efforts are required that combine a variety of methodologies, which in particular exploit the differential dynamical behaviour of kinases. In the following doctoral thesis different computational methodologies are employed to study three topics related to phosphorylation: 1.Understanding the reaction mechanism of phosphorylation and dephosphorylation: -Using PKA and GSK3β as model kinases we perform molecular dynamics simulations and carry out hybrid quantum mechanics/molecular mechanics (QM/MM) calculations on the evolution of the Michaelis complexes formed between these kinases and their bona fide substrates towards the respective phosphorylated products and characterize each step of the phosphorylation reactions in atomic detail paying particular attention to the roles and fates of the catalytic metal ions . -We analyse the dephosphorylation reaction catalyzed by the SHIP2 inositol phosphatase. Models of the two substrates, PI(4,5)P3 and IP4, in complex with SHIP2 phosphatase are built to understand the reaction mechanism in atomic detail . In addition, Principal Component Analysis and molecular dynamics simulations are used to study the allosteric role of the C2 domain and to propose and test different mutants with a view to confirming or rejecting our hypothesis. 2.Analysis of conformational changes involved in the activation of two prototypical kinases: -Free energy calculations using umbrella sampling and metadynamics are applied to validate the energetic profiles of the opening and closing of the activation loop in non-receptor Abelson tyrosine kinase (Abl) codificated in the protooncogene ABL1 and to characterize the differences between the phosphorylated and the unphosphorylated forms of this pharmacologically important enzyme. -Molecular dynamics simulations and normal mode analysis are performed on focal adhesion kinase (FAK), another non-receptor tyrosine kinase involved in cancer, in the presence or absence of ATP/Mg2+ in order to understand the allosteric effect of ATP on the conformational and dynamic properties of the enzyme. 3.Computational search for specific protein kinase inhibitors: -We perform extensive molecular dynamics simulations of the apo enzymes to identify transient and potentially targetable allosteric pockets. -We calculate molecular interaction fields and putative hotspots on the active site and regulatory domains of these kinases to characterize the potential ligand-binding sites. -We make use of a variety of docking tools to identify new potential hits present in chemical libraries and/or fragment databases (large-scale virtual screening)