Heme binding to rabbit hemopexin and its domain I, obtained by proteolytic cleavage of intact hemopexin, was studied by EPR, MCD and H-1-NMR spectroscopies. The data obtained support the proposal that the heme Fe(III) is coordinated by two histidine ligands (Morgan et al. (1988) J. Biol. Chem. 263, 8220-8225; Muster et al. (1991) J. Protein Chem. 10, 123-128) and are inconsistent with recently reported mutagenesis studies indicating that bis-histidine ligation is unlikely (Satoh et al. (1994) Proc. Natl. Acad. Sci. USA 91, 8423-8427). Although the MCD data are consistent with both bis-histidine and histidine/lysine ligation, the EPR spectra are typical of bis-histidine ligation. Overall the magneto-optical spectra are characteristic for bis-histidine ligation. The EPR and NMR data indicate that there is a difference in the heme environments of the intact hemopexin and its domain I but overall the spectroscopic information suggests heme bound to domain I has the same ligands as intact hemopexin. The H-1-NMR studies indicate that heme binding to domain I perturbs at least 4 of the 5 histidines. This is consistent with axial ligation of the heme by two histidines, and a conformational change induced by heme binding affecting two more. Interestingly, resonances of the carbohydrate bound to intact hemopexin and domain I were also perturbed by heme binding, pH dependence studies showed that heme remained bound to intact hemopexin over the pH range 6.5-10.0 without any major change in the ligation or environment of the heme.