Nitrogenase consists of two metalloproteins (Fe protein and MoFe protein) which are assumed to associate and dissociate to transfer a single electron to the substrates. This cycle, called the Fe protein cycle, is driven by MgATP hydrolysis and is repeated until the substrates are completely reduced. The rate-limiting step of the cycle, and substrate reduction, is suggested to be the dissociation of the Fe protein-MoFe protein complex which is obligatory for the reduction of the Fe protein [Thorneley, R. N. F., and Lowe, D. J. (1983) Biochem. J. 215, 393-403]. This hypothesis is based on experiments with dithionite as the reductant. We also tested besides dithionite flavodoxin hydroquinone, a physiological reductant. Two models could describe the experimental data of the reduction by dithionite. The first model, with no reduction of Fe protein bound to MoFe protein, predicts a rate of dissociation of the protein complex of 8.1 s-1. This rate is too high to be the rate-limiting step of the Fe protein cycle (kobs = 3.0 s-1). The second model, with reduction of the Fe protein in the nitrogenase complex, predicts a rate of dissociation of the protein complex of 2.3 s-1, which in combination with reduction of the nitrogenase complex can account for the observed turnover rate of the Fe protein cycle. When flavodoxin hydroquinone (155 microM) was the reductant, the rate of reduction of oxidized Fe protein in the nitrogenase complex (kobs approximately 400 s-1) was 100 times faster than the turnover rate of the cycle with flavodoxin as the reductant (4 s-1). Pre-steady-state electron uptake experiments from flavodoxin hydroquinone indicate that before and after reduction of the nitrogenase complex relative slow reactions take place, which limits the rate of the Fe protein cycle. These results are discussed in the context of the kinetic models of the Fe protein cycle of nitrogenase.