A common gene variant in the heparin-binding domain (HBD) of extracellular superoxide dismutase (ECSOD) may predispose human carriers to ischemic heart disease. We have demonstrated that the HBD of ECSOD is important for ECSOD to restore vascular dysfunction produced by endotoxin. The purpose of this study was to determine whether the gene variant in the HBD of ECSOD (ECSOD_R213G) protects against endothelial dysfunction in a model of inflammation. We constructed a recombinant adenovirus that expresses ECSOD_R213G. Adenoviral vectors expressing ECSOD, ECSOD_R213G or -galactosidase (LacZ, a control) were injected i.v. in mice. After 3 days, at which time the plasma SOD activity is maximal, vehicle or endotoxin (lipopolysaccharide or LPS, 40 mg/kg) was injected i.p. Vasomotor function of aorta in vitro was examined 1 day later. Maximal relaxation to sodium nitroprusside was similar in aorta from normal and LPS-treated mice. Maximal relaxation to acetylcholine (10^-5) was impaired after LPS and LacZ (63±3%, mean±SE) compared to normal vessels (83±3%) (p<0.05). Gene transfer of ECSOD improved (p<0.05) relaxation in response to acetylcholine (76±5%) after LPS, whereas gene transfer of ECSOD_R213G had no effect (65±4%). Superoxide was increased in aorta (measured using lucigenin and hydroethidine) after LPS, and levels of superoxide were significantly reduced following ECSOD but not ECSOD_R213G. Thus, ECSOD reduces superoxide and improves relaxation to acetylcholine in the aorta after LPS, while the ECSOD variant R213G had minimal effect. These findings suggest that, in contrast to ECSOD, the common human gene variant of ECSOD fails to protect against endothelial dysfunction produced by an inflammatory stimulus.