Previous studies investigating homocysteine and vascular disease have relied on total plasma homocysteine as the sole index of homocysteine status. We examined the dynamic relationship between vascular endothelial function and concentrations of total, protein-bound oxidized, free oxidized, and reduced homocysteine to identify the homocysteine form associated with endothelial dysfunction in humans. We investigated 14 healthy volunteers (10 men, 4 women). Brachial artery flow-mediated dilatation was measured at baseline and at 30, 60, 120, 240, and 360 minutes after oral (1) L-methionine (50 mg/kg), (2) L-homocysteine (5 mg/kg), and (3) placebo. Plasma concentrations of total, protein-bound oxidized, free oxidized, and reduced homocysteine were measured at each time point, and nitroglycerin-induced dilatation at was assessed at 0, 120, and 360 minutes. Flow-mediated dilatation fell, and concentrations of total, protein-bound oxidized, free oxidized, and reduced homocysteine increased after oral homocysteine and oral methionine (all P<0.05 for difference in time course compared with placebo). Flow-mediated dilatation showed a reciprocal relationship with reduced homocysteine during both homocysteine and methionine loading. In both loading studies, peak reduction in flow-mediated dilatation coincided with maximal reduced homocysteine concentrations. In contrast, there was no consistent relationship between flow-mediated dilatation and free oxidized homocysteine, protein-bound oxidized homocysteine, or related species. Nitroglycerin-induced dilatation was unchanged by oral homocysteine and oral methionine (P>0.10 compared with placebo). Reduced homocysteine is closely associated with endothelial dysfunction during oral methionine and oral homocysteine loading. Our observations support the hypothesis that reduced homocysteine is the deleterious form of homocysteine for vascular function in vivo and suggest a less important role for other homocysteine species.