Selective proteolysis allows the isolation of a heme-binding fragment spectrally similar to microsomal cytochrome b 5 from both baker's yeast flavocytochrome b 2 (a flavohemoprotein) and liver sulfite oxidase (a molybdoprotein). The amino acid sequences of these two fragments have been published separately (Guiard & Lederer, 1976,1979). We present in this paper an alignment of those sequences with that of microsomal cytochrome b 5. The structural consequences of the similarity between the three primary structures are discussed in the light of the cytochrome b 5 three-dimensional model (Mathews et al., 1971,1972,1975; Mathews & Czerwinski, 1976). It is concluded that the three heme-binding proteins are in all probability the products of a divergent evolution from a common ancestor and that they must present a basically similar backbone with some surface alterations. We propose to name this backbone the “cytochrome b 5 fold”. The comparison of the three proteins suggests hypotheses concerning the molecular surface areas involved in the recognition of cytochrome c (the common acceptor) and of the respective reductase (flavo- or molybdoprotein). In addition, our results suggest that at some point in evolution, several copies of an initial hemoprotein gene were formed in the cellular genome. Subsequently, one copy was fused with the gene for another function: a flavoreductase in yeast cells or a molybdoreductase in hepatic cells.