Date of this Version
The prevailing dogma in structural genomics is the existence of a strong correlation between protein sequence, structure, and biological function. Proteins with high sequence similarity typically have a similar, if not the same, structure and function. In many cases this logic can fail due to distantly related proteins having very low sequence similarity, a lack of a representative structure, structural novelty, or the absence of a characterized function. Further, the paradigm fails to account for dynamics, which have a significant effect on structural stability and enzymatic efficacy.
Nuclear magnetic resonance (NMR) spectroscopy is uniquely capable of solving the structure, assisting with annotation, and deriving the dynamics of previously unstudied proteins. Historically, NMR has been used to calculate structures and dynamics of small or disordered proteins, which could then be used with computational methods to predict function. Predicted annotations are then confirmed by further experimentation such as ligand screens or titrations. The combination of NMR and bioinformatics, therefore, works synergistically to yield significant results, which has the ability to characterize highly complex proteins and fill gaps in the sequence to structure to function paradigm.
This dissertation begins with work accomplished using the Comparison of Active Site Structures (CPASS) software to show the functional evolution of a class of cofactor dependent enzymes and also expands on the utility of CPASS with the implementation of a functional clustering of its database. Described next is an emphasis on protein and peptide structure and the relationship between the experimentally derived ensembles and biological function and dynamics. Recent improvements in the calculation of protein fast-timescale dynamics are then introduced before a final concluding chapter.
Advisor: Robert Powers