(620f) Toward Structure Prediction and Design of Protein Glycosylation

Glycosylation plays a critical role in protein folding, activity, stability, and processing. Glycosylation often underlies differences in protein properties when expressed in different organisms or under altered conditions such as in tumor environments or under metabolic stress. Three-dimensional models of the molecular structure of glycoproteins would be useful for identifying the molecular origins of these functional differences. However, few structural tools exist for identifying low-energy conformations of carbohydrates or glycoproteins. We have created a general tool for modeling the three-dimensional structures of oligosaccharides, polysaccharides, and glycoproteins as part of the Rosetta biomolecular modeling package. Our tool samples ring, backbone, and side chain conformations. We employ a library of sugar monomers and a methodology for chaining, branching, and modifying the monomers to create complex sugar polymers covering most biologically and synthetically relevant chemistries. Initial tests validate the accuracy of the models through prediction of ring conformation population frequencies (benchmarked to quantum calculations and x-ray crystal structure propensities) and docking of oligosaccharides (benchmarked to x-ray crystal complex structures). Some early applications include structure-determination of highly glycosylated protein structures and mechanisms of glycosylation effects on carboxylesterase activity. Finally, we present possibilities for computational design of glycosylations.