Gastric inhibitory polypeptide (GIP) is susceptible to degradation, but only recently has dipeptidyl peptidase IV been identified as the enzyme responsible. Most RIAs recognize both intact GIP-(1-42) and the noninsulinotropic N-terminally truncated metabolite, GIP-(3-42), hampering measurement of plasma concentrations. The molecular nature of GIP was examined using high pressure liquid chromatography and a newly developed RIA specific for the intact N-terminus of human GIP. In healthy subjects after a mixed meal, intact GIP (N-terminal RIA) accounted for 37.0+/-2.5% of the total immunoreactivity determined by C-terminal assay. High pressure liquid chromatographic analysis of fasting samples by C-terminal assay revealed one major peak (73.8+/-2.9%) coeluting with GIP-(3-42). One hour postprandially, two major peaks were detected, corresponding to GIP-(3-42) and GIP-(1-42) (58.1+/-2.7% and 35.7+/-4.2%, respectively). GIP-(3-42) was not detected by N-terminal assay; the major peak coeluted with intact GIP (86.4+/-5.8% and 81.3+/-0.9%, 0 and 1 h, respectively). After iv infusion, intact GIP constituted 37.1+/-4.1% and 41.3+/-3.4% of the total immunoreactivity in healthy and type 2 diabetic subjects, respectively. The plasma t1/2 was shorter (P < 0.0001) when determined by N-terminal compared with C-terminal assay (7.3+/-1.0 vs. 16.8+/-1.6 and 5.2+/-0.6 vs. 12.9+/-0.9 min, healthy and diabetic subjects, respectively), and both t1/2 were shorter in the diabetic group (P < 0.05). We conclude that dipeptidyl peptidase IV is important in GIP metabolism in humans in vivo, and that an N-terminally directed assay is required for determination of plasma concentrations of biologically active GIP.