The pleiotropic effect of the sickle gene suggests that factors in addition to polymerization of the mutant gene product might be involved in sickle disease pathobiology. We have examined rates of heme transfer to hemopexin from hemoglobin in dilute aqueous solution (0.5 mg of Hb per ml) at 37 degrees C. HbO2 S loses heme 1.7 times faster than HbO2 A, with apparent rate constants of 0.024 hr-1 and 0.014 hr-1, respectively. In contrast, Hb A and Hb S behave identically in their MetHb forms (very rapid heme loss) and their HbCO forms (zero heme loss). This indicates that the faster heme loss from HbO2 S is due to accelerated autoxidation (HbO2----MetHb) rather than to some other type of instability inherent in the relationship of sickle heme to its pocket in globin. This interpretation is supported by spectrophotometric measurement of initial rates of MetHb formation during incubation at 37 degrees C. This directly shows 1.7 times faster autoxidation, with apparent rate constants of 0.050 hr-1 for HbO2 S and 0.029 hr-1 for HbO2 A. While the participation of this process in the cellular pathobiology of sickle erythrocytes remains unproven, the present data are consistent with, and perhaps help explain, two prior observations: the excessive spontaneous generation of superoxide by sickle erythrocytes; and the abnormal deposition of heme and heme proteins on membranes of sickle erythrocytes.