Glutamine:fructose 6-phosphate amidotransferase (GFA) is rate-limiting for hexosamine biosynthesis, while a UDP-GlcNAc beta-N-acetylglucosaminyltransferase (O-GlcNAc transferase) catalyses final O-linked attachment of GlcNAc to serine and threonine residues on intracellular proteins. Increased activity of the hexosamine pathway is a putative mediator of glucose-induced insulin resistance but the mechanisms are unclear. We determined whether O-GlcNAc transferase is found in insulin-sensitive tissues and compared its activity to that of GFA in rat tissues. We also determined whether non-insulin-dependent diabetes mellitus (NIDDM) or acute hyperinsulinaemia alters O-GlcNAc transferase activity in human skeletal muscle. O-GlcNAc transferase was measured using 3H-UDP-GlcNAc and a synthetic cationic peptide substrate containing serine and threonine residues, and GFA was determined by measuring a fluorescent derivative of GlcN6P by HPLC. O-GlcNAc transferase activities were 2-4 fold higher in skeletal muscles and the heart than in the liver, which had the lowest activity, while GFA activity was 14-36-fold higher in submandibular gland and 5-18 fold higher in the liver than in skeletal muscles or the heart. In patients with NIDDM (n = 11), basal O-GlcNAc transferase in skeletal muscle averaged 3.8 +/- 0.3 nmol/mg.min, which was not different from that in normal subjects (3.3 +/- 0.4 nmol/mg.min). A 180-min intravenous insulin infusion (40 mU/m2.min) did not change muscle O-GlcNAc transferase activity in either group. We conclude that O-GlcNAc transferase is widely distributed in insulin-sensitive tissues in the rat and is also found in human skeletal muscle. These findings suggest the possibility that O-linked glycosylation of intracellular proteins is involved in mediating glucose toxicity. O-GlcNAc transferase does not, however, appear to be regulated by either NIDDM or acute hyperinsulinaemia, suggesting that mass action effects determine the extent of O-linked glycosylation under hyperglycaemic conditions.