Structure and dynamics of mucin-like glycopeptides. Examination of peptide chain expansion and peptide-carbohydrate interactions by stochastic dynamics simulations.

Mucins and other highly O-linked glycoproteins have been found to exist in random-coil conformations with peptide chain dimensions about 3-fold more expanded than found for deglycosylated mucins or denaturated proteins. We have examined the origin of the peptide chain expansion in mucins by stochastic dynamics simulations which include a treatment of solvation energy effects based on solvent-accessible surface area and polarizability [GB/SA; Still, C. W., et al. (1990) J. Am. Chem. Soc. 112, 6127]. The glycopeptides studied contained pairs of threonine residues (flanked by alanine residues) which were O-glycosylated by the di- and monooligosaccharide side chains alpha-NeuNAc(2-6)alpha-GalNAc and alpha-GalNAc. These glycopeptides serve as simple models for native and asialo ovine submaxillary mucin. Computer stochastic dynamic simulations show a significant decrease in end-to-end distance and radius of gyration (32% and 33%, respectively) upon complete removal of carbohydrate from the glycopeptide AAA(NeuNAc-(2-6)GalNAc)-T(NeuNAc(2-6)GalNAc)-TAAA. These changes are consistent with the extrapolations of the mucin chain dimension data to glycopeptides of this size. The simulations have identified two potentially strong peptide-carbohydrate hydrogen bonds that can influence the orientation of O-linked GalNAc. With two contiguous glycosylated sites, the lowest energy conformation obtained is characterized by a GalNAc amide proton hydrogen bond to the carbonyl of the peptide residue C-terminal to the site of glycosylation. This conformation differs from the glycopeptide conformations predicted for glycopeptides with single or widely spaced glycosylation sites. The results suggest that the experimentally determined mucin peptide chain dimensions can be fully accounted for by short-range (+/- 3 residue) intramolecular steric and hydrogen bond interactions resulting from the clustering of glycosylated residues.