Human O-GlcNAc transferase substrate recognition via MALDI-TOF MS quantification of peptide glycosylation

Human O-GlcNAc transferase (hOGT) modifies serine/threonine residues on cytosolic and nuclear proteins, but predictive rules for substrate recognition are incomplete. Here, we combine MALDI-TOF MS quantification of intact peptides, conversion-based sequence logos, and molecular docking to elucidate recognition patterns. Testing a panel of synthetic peptides (7-25 residues) yielded O-GlcNAcylation from 0 % to 100 %. Stereoselective glycosylation of EA2-R produced β-O-GlcNAc (hOGT/UDP-GlcNAc), α-O-GlcNAc, and α-O-GalNAc (ppGalNAc-T2/UDP-GlcNAc or UDP-GalNAc), enabling comparative studies. Tiered analysis (≥30 %, <30 %, 0 % conversion) highlighted trends: efficient substrates prefer Pro at -2/-3, Val/Ser at -1, Ser at +1, Ala at +2, and Ser acceptors; lower groups feature reduced Pro/Val, increased acids; non-substrates show Gly at -2/-1 and poor +1/+2 residues. Targeted variants tuned reactivity (e.g., Val near -3 achieved 100 % for CKII; distal edits affected CRYAA). Docking revealed productive contacts in reactive peptides. These findings provide design rules for O-GlcNAc assays and a framework for peptide substrate engineering.