Combining Computational And Experimental Approaches To Study Disordered And Aggregation Prone Proteins

Over the past two decades disordered proteins have become more widely recognized, challenging the canonical structure-function paradigm associated with proteins. These highly dynamic proteins have been identified across a wide range of species and play a variety of functional roles. Furthermore, the structural plasticity of these proteins gives way to their increased aggregation susceptibility, compared to canonical, well-folded proteins, placing disordered proteins at the center of many neurodegenerative diseases. Despite the increased recognition of the abundance and complexity of disordered proteins, their structural features and the mechanisms by which they transit between functional and pathological roles remains elusive. The efforts described herein focus on leveraging both experimental and computational approaches to study the structure and dynamics of these proteins. Fluorescence-based experiment have proven useful for studying these systems as the intrinsic heterogeneity of this class of proteins, which precludes the use of many traditional structural biochemistry techniques, can be accommodated. Therefore, initial efforts focused on developing new minimally perturbing fluorescence probes and coupling these probes with site-selective labeling strategies. Subsequent efforts focused on identifying methods which could predict where these probes would be tolerated to boost protein yield and avoid structural perturbation. These and other fluorescence probes were employed in Förster Resonance Energy Transfer (FRET) experiments, to study the conformational ensemble of α-synuclein, a disordered protein whose aggregation is implicated in Parkinson’s Disease pathogenesis. Experimental FRET data was paired with molecular modeling in PyRosetta to simulate the conformational ensembles of α-synuclein in the presence and absence of 2 M TMAO. The accuracy of the resultant ensembles was corroborated by comparison to other experimental data. Following this initial success using experimentally constrained simulations, attention was directed towards the development of algorithms capable of generating accurate structural representations of both disordered and ordered proteins de novo. Lastly, this work showcases the utility of a high-throughput in-silico screening approach in identifying a compound that binds selectively to α-synuclein fibrils with nanomolar affinity. Overall this work highlights several computational and experimental approaches which are broadly applicable to the study of disordered and aggregation prone protein