The presence and biological significance of circulating glycated insulin has been evaluated by high-pressure liquid chromatography (HPLC), electrospray ionization mass spectrometry (ESI-MS), radioimmunoassay (RIA), receptor binding, and hyperinsulinemic-euglycemic clamp techniques. ESI-MS analysis of an HPLC-purified plasma pool from four male type 2 diabetic subjects (HbA(1c) 8.1 +/- 0.2%, plasma glucose 8.7 +/- 1.3 mmol/l [means +/- SE]) revealed two major insulin-like peaks with retention times of 14-16 min. After spectral averaging, the peak with retention time of 14.32 min exhibited a prominent triply charged (M+3H)(3+) species at 1,991.1 m/z, representing monoglycated insulin with an intact M(r) of 5,970.3 Da. The second peak (retention time 15.70 min) corresponded to native insulin (M(r) 5,807.6 Da), with the difference between the two peptides (162.7 Da) representing a single glucitol adduct (theoretical 164 Da). Measurement of glycated insulin in plasma of type 2 diabetic subjects by specific RIA gave circulating levels of 10.1 +/- 2.3 pmol/l, corresponding to approximately 9% total insulin. Biological activity of pure synthetic monoglycated insulin (insulin B-chain Phe(1)-glucitol adduct) was evaluated in seven overnight-fasted healthy nonobese male volunteers using two-step euglycemic-hyperinsulinemic clamps (2 h at 16.6 micro g x kg(-1) x min(-1), followed by 2 h at 83.0 micro g x kg(-1) x min(-1); corresponding to 0.4 and 2.0 mU x kg(-1) x min(-1)). At the lower dose, the exogenous glucose infusion rates required to maintain euglycemia during steady state were significantly lower with glycated insulin (P < 0.01) and approximately 70% more glycated insulin was required to induce a similar rate of insulin-mediated glucose uptake. Maximal responses at the higher rates of infusion were similar for glycated and control insulin. Inhibitory effects on endogenous glucose production, insulin secretion, and lipolysis, as indicated by measurements of C-peptide, nonesterified free fatty acids, and glycerol, were also similar. Receptor binding to CHO-T cells transfected with human insulin receptor and in vivo metabolic clearance revealed no differences between glycated and native insulin, suggesting that impaired biological activity is due to a postreceptor effect. The present demonstration of glycated insulin in human plasma and related impairment of physiological insulin-mediated glucose uptake suggests a role for glycated insulin in glucose toxicity and impaired insulin action in type 2 diabetes.