The reduction of substrates catalyzed by nitrogenase normally requires nucleotide-dependent Fe protein delivery of electrons to the MoFe protein, which contains the active site FeMo cofactor. Here, it is reported that independent substitution of three amino acids (β-98Tyr→His, α-64Tyr→His, and β-99Phe→His) located between the P cluster and FeMo cofactor within the MoFe protein endows it with the ability to reduce protons to H2, azide to ammonia, and hydrazine to ammonia without the need for Fe protein or ATP. Instead, electrons can be provided by the low-potential reductant polyaminocarboxylate-ligated Eu(II) (Em values of −1.1 to −0.84 V vs the normal hydrogen electrode). The crystal structure of the β-98Tyr→His variant MoFe protein was determined, revealing only small changes near the amino acid substitution that affect the solvent structure and the immediate vicinity between the P cluster and the FeMo cofactor, with no global conformational changes observed. Computational normal-mode analysis of the nitrogenase complex reveals coupling in the motions of the Fe protein and the region of the MoFe protein with these three amino acids, which suggests a possible mechanism for how Fe protein might communicate subtle changes deep within the MoFe protein that profoundly affect intramolecular electron transfer and substrate reduction.

Nitrogenase reduction of dinitrogen (N2) to ammonia (NH3) involves a sequence of events that occur upon the transient association of the reduced Fe protein containing two ATP molecules with the MoFe protein that includes electron transfer, ATP hydrolysis, Pi release, and dissociation of the oxidized, ADP-containing Fe protein from the reduced MoFe protein. Numerous kinetic studies using the nonphysiological electron donor dithionite have suggested that the rate-limiting step in this reaction cycle is the dissociation of the Fe protein from the MoFe protein. Here, we have established the rate constants for each of the key steps in the catalytic cycle using the physiological reductant flavodoxin protein in its hydroquinone state. The findings indicate that with this reductant, the rate-limiting step in the reaction cycle is not protein–protein dissociation or reduction of the oxidized Fe protein, but rather events associated with the Pi release step. Further, it is demonstrated that (i) Fe protein transfers only one electron to MoFe protein in each Fe protein cycle coupled with hydrolysis of two ATP molecules, (ii) the oxidized Fe protein is not reduced when bound to MoFe protein, and (iii) the Fe protein interacts with flavodoxin using the same binding interface that is used with the MoFe protein. These findings allow a revision of the rate-limiting step in the nitrogenase Fe protein cycle.