Prediction of glycation sites: new insights from protein structural analysis

Glycation of proteins is a nonenzymatic process in which proteins react with reducing sugar molecules. This process takes place at epsilon-amino (epsilon-NH+) groups of lysine or hydroxylysine residues as well as alpha-amino groups of N-terminal residues. In the present study glycation of epsilon-NH+ groups of lysines was computationally analyzed for 26 proteins based on their 3D structures. We determined the spatial relationship with acidic or basic residues and correlated them with the glycation prediction algorithm Netglycate 1.0 software, which employed primary structure exclusively for glycation site prediction. Of the lysines from 19 of the 20 proteins employed to build the Netglycate 1.0 algorithm 87.80% depicted a spatial relationship with acidic or basic residues. For the remaining seven proteins that were not included in the algorithm, 95.23% of the lysines exhibited a spatial relationship with acidic or basic residues. For these seven proteins, Netglycate 1.0 predicted only 52.38% of the lysines with a previously reported experimental glycation as potential glycation sites. In all cases, distances between residues were less than or equal to 9.78 angstrom. These results suggest that it is the spatial relationship of lysines with acidic or basic residues in the 3D conformation of a protein that determines the glycation target site, rather than a specific sequence of the primary structure.