Structural studies of proteins using computational methods
2014-12-15T10:35:40Z (GMT) by
Protein function is inextricably linked with protein three-dimensional (3D) structure. Therefore, a greater understanding of the 3D structure of proteins leads to a deeper understanding of how and why a particular protein functions in a specific way. Computer-based methods have been used to study the relationship between protein structure and function. The methods applied to the work presented in this thesis are: structure prediction through comparative modelling, both by homology (i.e. sequence similarity) with, and by analogy (i.e. threading) to, proteins with known 3D structure generation using distance geometry and simulated annealing calculations on NMR derived data assessment of generated structures and molecular interaction of 3D structures, particularly studies on electrostatic surface potential, ligand binding and molecular docking. These modelling methods have been applied as follows: Comparative modelling of human NADPH cytochrome P450 reductase, including an investigation into the interaction of its component domains. Specific residues were identified as possibly being involved in electron transfer from the FAD to the FMN prosthetic groups within the enzyme. Comparative modelling of ubiquitin conjugating enzyme 9, glutathione transferase and neurocalcin delta for the purpose of the 1996 CASP2 modelling assessment. The models compared favourably with models from other contributing groups, and predicted well the errors in the atomic positions. The solution structure of the GDP-bound form of the G-protein, Cdc42Hs, was determined using experimentally derived NMR data. Chemical shift changes (obtained by collaborators) between the GTP-bound and GDP-bound forms indicate that conformational change between the two states, facilitating interaction with effector and regulatory proteins, are associated with a contiguous region on the surface of the protein. Docking a ligand into the crystal structure of the third PDZ domain of the human homologue of Drosophila discs-large tumour suppressor protein thus providing a structural hypothesis for ligand specificity. Modelling the role of the enzyme trimethylamine dehydrogenase in electron transfer, substrate access and substrate specificity. Channels characterised within the enzyme indicate possible additional regions for substrate access to the binding site. A number of residues were identified as having possible roles in specificity for trimethylamine over dimethylamine as substrate.