Azide binds to fast cytochrome bo with a stoichiometry of 1:1, the dissociation constant for this reaction being approximately 2 x 10(-5) M. The changes induced in the electronic absorption are very slight and are consistent with heme o remaining hexacoordinate high-spin, an observation confirmed by room temperature MCD spectroscopy in the region 350-2000 nm. X-band EPR spectroscopy of the azide-bound form shows heme o remains coupled to CuB, but that the integer spin signal (g = 3.7) that we have previously reported to be associated with the binuclear center of fast cytochrome bo [Watmough et al. (1993) FEBS Lett. 319, 151-154], is shifted to higher field. The kinetics of azide binding are an order of magnitude faster than those observed for the binding of cyanide. Unlike cyanide, the observed rate constants do not saturate in the range 0.05-25 mM. The value of Kon shows a marked dependence on pH, indicating that the active species is hydrazoic acid. It is argued that these data are consistent with the binding of azide ion as a terminal ligand to CuB yielding a binuclear center in the form FeIII-OH2:: CuBII-N3. The binding of azide in heme-copper oxidases may cause displacement of another nitrogenous ligand from CuB which might explain the absence of electron density associated with histidine-325 in the structure of the Paracoccus denitrificans CCO [Iwata et al. (1995) Nature 376, 660-669]. Formate appears to act as a bidentate ligand to the binuclear center-, blocking not only the binding of azide to CuB but also the binding of cyanide to heme o.