Abstract
The human glycome comprises a vast untapped repository of 3D-structural information that holds the key to glycan recognition and a new era of rationally designed mimetic chemical probes, drugs, and biomaterials. Toward routine prediction of oligosaccharide conformational populations and exchange rates at thermodynamic equilibrium, we apply hardware-accelerated aqueous molecular dynamics to model μs motions in N-glycans that underpin inflammation and immunity. In 10μs simulations, conformational equilibria of mannosyl cores, sialyl Lewis (sLe) antennae, and constituent sub-sequences agreed with prior refinements (X-ray and NMR). Glycosidic linkage and pyranose ring flexing were affected by branching, linkage position, and secondary structure, implicating sequence dependent motions in glycomic functional diversity. Linkage and ring conformational transitions that have eluded precise quantification by experiment and conventional (ns) simulations were predicted to occur on μs timescales. All rings populated non-chair shapes and the stacked galactose and fucose pyranoses of sLe(a) and sLe(x) were rigidified, suggesting an exploitable 3D-signature of cell adhesion protein binding. Analyses of sLe(x) dynamics over 25μs revealed that only 10μs were sufficient to explore all aqueous conformers. This simulation protocol, which yields conformational ensembles that are independent of initial 3D-structure, is proposed as a route to understanding oligosaccharide recognition and structure-activity relationships, toward development of carbohydrate-based novel chemical entities.