Exploring the role of conformational dynamics in the regulation of tyrosine kinases.

Tyrosine kinases (TKs) are a family of signalling proteins of great pharmaceutical im- portance, as they are involved in the regulation of most cellular pathways. TKs catalytic activity is strictly regulated by conformational changes and post-translational modifi- cations, and their deregulation is involved in numerous human diseases, ranging from cancer to autoimmune diseases. Among tyrosine kinases, Abl and Src are of particular interest for cancer research. The Abl domain in the BCR-Abl fusion protein is the main cause of chronic myeloid leukemia, and it was the target of the first successful anti- leukemic therapy, the powerful kinase inhibitor imatinib. We now know that imatinib effectively inhibits BCR-Abl, as well as Kit and Lck kinases, by binding to a specific inactive state, in which the conserved Asp-Phe-Gly motif (DFG) assumes a peculiar "out" conformation. Still, there are many questions on its mode of action. For instance, other TKs with an extended identity with Abl (such as Src, which has 45% sequence identity) bind much less strongly to imatinib, in spite of very similar binding mode. Moreover, the mode of action of drug-resistant mutations that induce imatinib resis- tance and cause an increasing number of relapses in patients under treatment, is still poorly understood. Understanding the molecular mechanisms responsible for the ob- served differences in imatinib activity, is essential for the development of new selective anticancer drugs. In this thesis, by using computational and experimental approaches, I have investigated the reasons leading to drug resistance and the differential binding affinity in homologous TKs. A combination of enhanced sampling molecular dynam- ics simulations (such as parallel tempering metadynamics or PTmetaD) were used to reconstruct and compare the free energy landscape associated with the relevant con- formational changes. Mutagenesis and isothermal titration calorimetry were used to validate the computational results.