Development of hydroxyl free radical chemistry for the surface mapping of proteins

In order to rapidly probe the structure of proteins in solution, a protocol for chemical modification of solvent accessible amino acid side chains was developed, and the sites of modification were determined by mass spectrometry to describe the surface of the protein. The knowledge gained about side chain solvent accessibility allowed the critical evaluation of structural models for the proteins examined, allowing incorrect models to be rejected and more likely models to be proposed. Methods were developed using either Fenton chemistry or photolysis of hydrogen peroxide to generate hydroxyl radicals in situ. The oxidation chemistry of these radicals with the side chains of various amino acids were exploited to label solvent accessible sites on several model proteins of known tertiary structure. The relative apparent rate of oxidation of the side chains was shown to be a function of the known solvent accessibility and the chemical reactivity of the amino acid. The known properties of hydroxyl radical oxidation of amino acid side chains allows hydroxyl radical surface mapping data to be used as biophysical constraints for evaluating structural models of proteins and protein-protein interactions. Computational models of the yeast ribonucleotide reductase inhibitor protein Sml 1 p were evaluated using surface mapping data of the functional C 14S Sml 1 p protein. Various full atom computational models were discredited based on the surface mapping data, and a manually adjusted computational model was generated that possessed low free energy, and agreed with surface mapping data, partial NMR data, and tryptophan anisotropy and quenching data. In addition, the interaction between peptides forming AB fibrils, implicated in Alzheimer's disease, were examined using hydroxyl radicals. This radical mapping suggests that the model proffered by Perutz et al, which states that the AB fibril is a solvent-filled nanotube, is incorrect. Overall, chemically-generated hydroxyl radicals have been developed as a general, multi-target labeling reagent for protein surfaces. Hydroxyl radical surface maps can be used to characterize protein tertiary and quaternary structure and to apply valuable biophysical constraints for structural modeling.