Carbohydrates are infamously challenging to model, yet they affect protein structure, stability, and activity. We are developing accurate and fast methods of modeling and designing carbohydrates for applications in glycobiology. We have built a framework within the Rosetta structure and design suite for modeling saccharide ligands and complex glycoconjugates. Our intuitive and efficient data structures allow access to all torsion angles (φ, ψ, ω, and χ) and Cremer–Pople parameters for sampling ring forms and capture the high degree of flexibility, stereochemistry, and branching in carbohydrates. Rosetta's flexibility and speed enable modeling of any glycan-containing molecule in docking and refinement protocols through exploration of this vast torsional and ring-conformational diversity.Rosetta's residue-centric approach, coupled with combinatorial “patching” of standard residues with specific functional groups, allows for design algorithms that sample alternative saccharide units, enabling high-throughput screening of thousands of protein variants and glycoforms in a search for stable or functional molecules.Here, we will report three studies benchmarking monosaccharide ring conformations, oligosaccharide structure prediction, and bound–bound protein–oligosaccharide docking. We explore the relative energy surfaces of the ring forms of all D-aldohexopyranoses; we examine predicted structures of two LewisX oligosaccharides; and we compare the docking predictions of eleven antibody–glycoantigen pairs with known structures. These studies will allow us to rigorously test the Rosetta scoring (energy) function, to adapt it to the unique chemical effects of sugars.We will also present preliminary real-world applications in antibody accessibility for glycosylation enzymes, x-ray crystal refinement of an extensively glycosylated HIV-1 envelope protein trimer, and the activity of glycosylated carboxylesterases.These new approaches will provide glycobiologists and glycoengineers a new computational toolbox, further the understanding of the biomolecular mechanisms of disease, and create opportunities for a wide range of previously intractable studies.
Read full abstract