Part I presents the computational tools used in this work: the comparative modeling and molecular docking approaches along with molecular dynamics. Part II presents structural predictions of Ca2+-binding domains in Ca2+-gated channels. A detailed description of the structure and function of these proteins can be found in the following Chapters. Chapter 4 focuses on human large conductance Ca2+- and voltage-gated potassium channel (hBKCa). Bioinformatics approaches and MD simulations were used to construct models of two domains important for Ca2+ binding and channel gating, namely the Regulator of Conductance for K+ (RCK1) and the so called calcium bowl. The relevance of these models for interpreting the available molecular biology data is then discussed. Chapter 5 deals with bestrophins, a recently discovered family of Cl− channels. Bestrophins feature a well conserved Asp-rich tract in their C-terminal part, which is homologous to Ca2+-binding motifs in calcium bowl of hBKCa. Based on these considerations, we constructed homology models of human bestrophin-1 Asp-rich domain. MD simulations and free energy calculations were used to identify Asp and Glu residues binding Ca2+ and to predict eects of their mutations to Ala. My work, performed in collaboration with C. Anselmi (SISSA/ISAS), was complemented by free energy calculations carried out by F. Pietrucci (SISSA/ISAS). Selected mutations were investigated by electrophysiological experiments performed by Prof. A. Menini, J. Rievaj, F. W. Grillo, and A. Boccaccio (SISSA/ISAS). The model of Asp-rich domain was then validated against experimental results. Part III is devoted to the prion protein. In this Part, Chapter 6 presents in vitro studies of D18scFv anti-prion effects performed by groups of Prof. C. Zurzolo (Institut Pasteur, Paris, France), Prof. G. Legname (SISSA/ISAS), L. Zentilin and M. Giacca (ICGEB, Trieste, Italy) and by Prof. S. B. Prusiner (Institute for Neurodegenerative Diseases, University of California San Francisco, U.S.A.) and structural prediction of a complex between the small antibody fragment (D18scFv) and PrPC. The complex was modeled using bioinformatics approaches. Initially, the D18scFv fragment alone was modeled based on a similar antibody-fragment template and then docked with prion protein. Based on this, interactions relevant for the recognition between the two proteins and for the mechanism of action of D18scFv are discussed. Chapter 7 describes a computational protocol for the design of ligands targeting cavity-less proteins, like most proteins involved in neurodegenerative diseases. Molecular docking methods are combined with MD simulations and free energy calculations using the metadynamics method [33, 34] to gain insights in ligand binding to such proteins, in our case to prion protein. We focused on a compound showing antiprion activity in vitro. Ligand-target interactions and ligand binding affinity as emerged by using our approach are compared with the available NMR data  and experimental constant of dissociation . In this work, also other two students and one postdoc were involved beside myself, namely S. Bongarzone, G. Rossetti and X. Biarnes (SISSA/ISAS). Finally, the conclusions are drawn in the last Chapter. The thesis closes with the List of publications and with the Acknowledgments.
|Titolo:||Predicting structural determinants and Ligand poses in proteins involved in neurological diseases: bioinformatics and molecular simulation studies|
|Data di pubblicazione:||10-feb-2009|
|Appare nelle tipologie:||8.1 PhD thesis|