A VECTOR PROJECTION METHOD FOR PREDICTING THE SPECIFICITY OF GALNAC-TRANSFERASE

The specificity of UDP-GalNAc:polypeptide N-acetylgalactosaminytransfe rase (GalNAc-transferase) is consistent with the existence of an exten ded site composed of nine subsites, denoted by P-4, P-3, P-2, P-1, P-0 , P-1', P-2', P-3', and P-4', where the acceptor at P-0 is being eithe r Ser or Thr. To predict whether a peptide will react with the enzyme to form a Ser- or Thr-conjugated glycopeptide, a vector projection met hod is proposed which uses a training set of amino acid sequences surr ounding 90 Ser and 106 Thr O-glycosylation sites extracted from the Na tional Biomedical Research Foundation Protein Database. The model post ulates independent interactions of the 9 amino acid moieties with thei r respective binding sites. The high ratio of correct predictions vs, total predictions for the data in both the training and the testing se ts indicates that the method is self-consistent and efficient. It prov ides a rapid means for predicting O-glycosylation and designing effect ive inhibitors of GalNAc-transferase. (C) 1995 Wiley-liss,Inc.