Abstract:
Anisotropic Network Model (ANM) guided Langevin Dynamics (LD) method (ANM-LD) is an enhanced sampling algorithm, in-house developed to study conformational changes between functional protein structures, that are not possible by traditional techniques. In this thesis, the applicability of ANM-LD was validated on various non-globular systems, then assessed on vitamin B12 importer BtuCD by means of experimental observations and comparison of computational outcomes with maltose importer MalFGK2 and lipid-linked oligosaccharide flippase PglK. ANM-LD succeeded to extract the mechanistic differences among these transporters while predicting fluctuations and allosteric couplings of BtuCD residues in agreement with previous experiments and observed FRET intensities. The dynamically key residues enabling the sampled transition were defined as functional residue networks and their estimated perturbation response were highly agreeable with the functional assays of the BtuCD mutants on these sites (25 out of 26 mutants functioned as predicted). Later ANM-LD algorithm was advanced to improve sampling, then tested on case systems c-Src kinase and BtuCD. In c-src kinase, these enhancements enabled to predict dynamically key sites that overlapped with known oncogenic mutation sites. In BtuCD, diversification of guiding modes resulted in alternative transition pathways. Consequently with its practicality and modularity; ANM-LD stands as an efficient tool to study protein dynamics and their working mechanisms and to extract allosteric communication networks, toward the aim of controlling protein function.|Keywords : Computational Sampling Methods, Protein Dynamics, Enhanced Sampling Methods, ANM-LD, Conformational Change, Elastic Network Model.
