A database analysis of potential glycosylating Asn-X-Ser/Thr consensus sequences.

An analysis of the frequency of occurrence of various residues at position X was carried out on the consensus glycosylating sequence Asn-X-Ser/Thr using the PDB three-dimensional database. 488 non-homologous proteins bearing 696 Asn-X-Ser/Thr (X not equal Pro) sequences were analysed. More than 65% of Asn residues, when they occur as part of the consensus sequence, lie on the surface of the protein, implying a potentiality for glycosylation. A deviation parameter (DP) was calculated as a measure of preferential (positive) or non-preferential (negative) selection. At the X position in the consensus-sequence segment, the amino acids Gly, Asn and Phe have statistically significant positive DP values. The high value of DP for Asn is a consequence of the preferential occurrence of homodoublets, while for Phe it may be a consequence of the stacking interaction of the aromatic ring with the glycan. Gly at the X position in the consensus glycosylating sequence may be functionally significant owing to its preference and its high percentage of occurrence in proteins. The Ramachandran (Phi,Psi) angles around Gly in the consensus sequence show clustering in the region which is disallowed for non-glycyl residues. In this region, a hydrogen bond between the side chain of Asn and the peptide backbone/side chain of Ser/Thr is possible, reflecting a positional as well as a conformational role in the consensus glycosylating sequence. For the 44 confirmed N-glycosylating sequences, an in-depth analysis of the (Psi(N), Phi(X), Psi(X), Phi(S/T)) dihedral angles, which position the side chains of Asn and Ser/Thr, shows that these can be grouped into nine conformational states. In most cases, a direct or water-mediated hydrogen bond between OD1 of Asn and OG of Ser/Thr is possible, reflecting the possible importance of this hydrogen bonding in the glycosylation process.