The Fur protein regulates the expression of a wide variety of iron-responsive genes; however, the interaction of this repressor with its cognate metal ion remains controversial. The iron-bound form of Fur has proved difficult to obtain, and conflicting results have been published using Mn(II) as a probe for in vitro DNA-binding studies. We report here that the purified protein contains tightly bound zinc and propose that Zn(II) is bound to the protein in vivo. Upon purification, Fur retains ca. 2.1 mol of Zn(II)/mol of Fur monomer (Zn2Fur). One zinc is easily removed by treatment of Zn2Fur with zinc chelating agents, resulting in Zn1Fur with ca. 0.9 mol of Zn(II)/mol of protein. The remaining zinc in Zn1Fur can only be removed under denaturing conditions to yield apo-Fur with ca. 0.1 mol of Zn(II)/mol of protein. Our results suggest that many literature descriptions of purified Fur protein do not correspond to the apo-protein, but to Zn1Fur or Zn2Fur. Dissociation constants (Kd) of protein-DNA complexes are ca. 20 nM for both Zn2Fur and Zn1Fur as determined by electrophoretic mobility shift assays and DNase I footprinting assays. The two metalated forms, however, show qualitative differences in the footprinting assays while apo-Fur does not bind specifically to the operator. The existence of these Zn(II) binding sites in Fur may resolve some discrepancies in the literature and have implications concerning Zur, a Fur homologue in E. coli that regulates zinc-responsive genes.