Atomistic Simulations of Model Amyloid Beta Aggregates, Water Networks and their Optical Properties

Based on the amyloid hypothesis, amyloid oligomers and the fibrils that they aggregate into, have been implicated in neurodegenerative diseases. Most of these amyloid proteins live in a solvent environment. The role of solvent in modulating the structural and dynamical properties of amyloid proteins remains poorly understood. In this thesis, computer simulations are used to reveal the structural properties of the amyloid protein and the coupling between protein and water using model systems. After assessing the validity of the force fields by comparison with high-level quantum chemistry calculations, we examine further the conformational free energy landscape of an amyloid protein. Different conformations characterized in the free energy surface are driven by internal protein interactions as well as interactions between protein and water, resulting in the collective reorganization of protein and water hydrogen bond networks. We show that these proteins are surrounded by water wires that add a roughness to the free energy surface. To better understand the water hydrogen bond network and particularly the water wires around protein, we used data-science algorithms allowing for the dimensionality and free energy landscape of different water coordinates to be determined. These results confirm that using water wire coordinates encodes more information on the underlying secondary structure of the protein. Finally, ab initio calculations are used to investigate the optical properties of amyloid proteins to help rationalize recent experiments suggesting the intrinsic fluorescence in fibrils that can occur without aromatic residues.