Energetic consequences of structural features and dynamics changes upon nucleotide binding to ribonuclease SA: molecular basis for nucleotide binding specificity

Proteins are often able to distinguish between closely related ligands, thus achieving specificity. A major goal in biophysical chemistry is to understand the molecular basis for protein-ligand recognition. This level of understanding is necessary for developing methods to accurately predict protein-ligand binding energetics from structural data. The goal of this thesis was to identify features of protein-ligand interactions that may not be adequately accounted for in structure energetics calculations in order to improve our ability to predict binding energetics for these interactions. Specifically, the features of protein-nucleotide binding were studied using the small, guanine-specific ribonuclease, RNase Sa binding to two closely related nucleotide inhibitors, guanosine-3'-monophosphate (3'GMP) and inosine-3'-monophosphate (3'IMP) as a model system. Comparing the binding of these two inhibitors using isothermal titration calorimetry (ITC), x-ray crystallography, NMR and molecular dynamics (MD) simulations has revealed important determinants of guanine base recognition by proteins, specifically the role of the exocyclic amino group (N2) of the guanine base. Importantly, due to the high conservation of guanine binding sites in proteins, the observations for RNase Sa can potentially be extended to other systems. In addition, RNase Sa has provided a well-defined system for the investigation of changes in heat capacity and changes in backbone dynamics upon ligand binding. All of the data presented here support the idea that fluctuations in protein structure can contribute significantly to protein-nucleotide binding energetics even for an apparently rigid-body interaction. These fluctuations make a significant contribution to the enthalpy, entropy, and heat capacity changes associated with the RNase Sa-nucleotide interaction. This implies that fast time-scale motions must be accounted for to optimize structure-based calculations for protein-nucleotide binding. The use of molecular dynamics simulations is presented as a promising way to account for these motions for surface area based calculations.