In higher eukaryotes, a variety of proteins are post-translationally modified by adding O-linked N-acetylglucosamine (GlcNAc) residue to serine or threonine residues. Misregulation of O-GlcNAcylation is linked to a wide variety of diseases, such as diabetes, cancer, and neurodegenerative diseases, including Alzheimer's disease. GlcNAc transfer is catalyzed by an inverting glycosyltransferase O-GlcNAc transferase (uridine diphospho-N-acetylglucosamine:polypeptide β-N-acetylaminyltransferase, OGT) that belongs to the GT-B superfamily. The catalytic mechanism of this metal-independent glycosyltransferase is of primary importance and is investigated here using QM(DFT)/MM methods. The structural model of the reaction site used in this paper is based on the crystal structures of OGT. The entire enzyme-substrate system was partitioned into two different subsystems: the QM subsystem containing 198 atoms, and the MM region containing 11,326 atoms. The catalytic mechanism was monitored by means of three two-dimensional potential energy maps calculated as a function of three predefined reaction coordinates at different levels of theory. These potential energy surfaces revealed the existence of a concerted S(N)2-like mechanism, in which a nucleophilic attack by O(Ser), facilitated by proton transfer to the catalytic base, and the dissociation of the leaving group occur almost simultaneously. The transition state for the proposed reaction mechanism at the MPW1K level was located at C1-O(Ser) = 1.92 Å and C1-O1 = 3.11 Å. The activation energy for this passage was estimated to be ~20 kcal mol(-1). These calculations also identified, for the first time for glycosyltransferases, the substrate-assisted mechanism in which the N-acetamino group of the donor participates in the catalytic mechanism.