The endothelium-derived relaxing factor (EDRF) has recently been reported to be nitric oxide. This study was performed to determine whether the vasorelaxant activity of EDRF could be accounted for by nitric oxide released from cultured endothelial cells. Nitric oxide released from cultured bovine aortic endothelial cells was monitored using a chemiluminescence technique with and without reflux processing of the effluent with a strong reducing environment (1% sodium iodide in glacial acetic acid). In the presence of indomethacin, basally released EDRF was sufficient to relax a bioassay detector vessel (preconstricted with prostaglandin F2 alpha) by 41 +/- 8%. Bradykinin (0.01 microM) and A23187 (10 microM) produced relaxation of the bioassay detector vessel equal to 74 +/- 7 and 69 +/- 13%, respectively. In both the absence and the presence of reflux preprocessing of the cell effluent in a reducing environment, the amount of nitric oxide detected by chemiluminescence was 7-10-fold less than that required to account for the detector vessel relaxation (determined from responses of the detector vessel to standard infusions of authentic nitric oxide). These experiments contradict the view that nitric oxide is the sole or principal EDRF. It is likely that either other nonprostanoid vasodilator substances are released in addition to nitric oxide or that EDRF is a nitric oxide containing or forming compound that is substantially more potent than authentic nitric oxide.