In subunit I (FixN) of the Bradyrhizobium japonicum cbb3-type oxidase, only five instead of the normal six strictly conserved histidines (H) could be unambiguously assigned as the putative heme or copper ligands. The ambiguity concerned H43 or H131 as the presumptive N-terminal ligands of the low-spin heme B. We report here that a H43A replacement had a wild-type phenotype, whereas the H131A mutant was defective in oxidase function and subunit assembly or stability, suggesting that H131 serves as the N-terminal low-spin heme ligand. Topological studies revealed that H131 resides on the periplasmic side of helix 2, where one of the low-spin heme ligands is normally found in conventional heme-copper oxidases.