N-Glycan Structure Modeling and in Silico Glycosylation: Template-Based Structure Prediction of Carbohydrate Structures of Glycoconjugates

Abstract

Obtaining the crystal structure of glycoconjugate is challenging due to the flexibility of the carbohydrate chains. Alternatively, computational modeling, which combines the primary sequence information of glycans determined by the mass spectrometry and known N-glycan structure, is an appealing approach. Here we present a survey of N-glycan structures of 35 different glycan sequences in the PDB, showing that N-glycan structures found on homologous glycoproteins are significantly conserved compared to the random background. This suggests that N-glycan chains can be confidently modeled to a glycoprotein if there exists a template N-glycan structure whose parent glycoprotein shares sequence similarity. On the other hand, N-glycan structures found on non-homologous glycoproteins have not shown significant structure similarity. However, despite that the global N-glycan structures are different, the internal substructure of those N-glycans found on the non-homologous glycoproteins, particularly, the substructure that are closer to the protein, showed significantly similar structure. Increased interaction with protein might be responsible to the restricted conformational space of N-glycan chains. Our results so far suggest that computational structure prediction of N-glycan portion of glycoconjugate using structure database would be effective, but different approaches must be needed depending on the availability of template structure. In addition, we also present a database for PDB glycan structural fragments (substructures) as well as PDB glycan-protein database, which are useful for glycan structure modeling.