| Phosphorylation is essential for mediating switch-like changes in protein function and cellular behavior. In eukaryotes, protein kinases catalyze phosphorylation of the hydroxyl groups of serine, threonine, and tyrosine, adding a -2 charge to otherwise neutral residues. This large electrostatic perturbation can alter protein structure, leading to switch-like changes in protein function. While mass spectrometry has identified many thousands of phosphorylation sites, few have been functionally characterized, and the structural effects of phosphorylation have been elucidated for even fewer. It is particularly difficult to obtain atomic-level resolution of structural changes due to phosphorylation. The use of atomic-level computational methods is an attractive candidate to augment low-resolution structural and functional studies in probing phosphorylation-induced conformational changes.;This thesis presents work using all-atom computational techniques to study phosphorylation-induced conformational changes. We studied phosphorylation-induced conformational changes in protein loops and helices for several diverse proteins. We present a case study examining the structural effects of activation loop phosphorylation in cyclin-dependent kinase 2 and the underlying energetic basis, highlighting the important balance between electrostatics and desolvation. From observations of phosphorylationinduced changes in the orientation of protein kinase lobes, we devised the beginnings of a method for the prediction of the orientation of protein domains, using antibodies as a model system. This method involved extensive sampling of side-chain conformations at the interface of the antibody heavy and light chain variable domains, and was successful at reconstructing existing antibody crystal structures. In addition, this energy-based method for the prediction of variable domain orientation, when coupled with a filter based on sequence identity, was more successful than using sequence identity alone when predicting the variable domain orientation in antibody homology models. Finally, we used all-atom molecular dynamics simulations to prospectively study conformational changes in the Arp2/3 complex upon phosphorylation of the Arp2 subunit. We find that Arp2 phosphorylation induces changes in the interactions in the highly charged Arp2-Arp3-ARPC4 interface, leading to a breakage of contacts between Arp2 and Arp3. We posit a model by which phosphorylation relieves the auto-inhibitory interactions at this interface, leading to activating conformational changes that allow full activation upon binding of other co-activators. |
