The reduction of substrates catalyzed by nitrogenase utilizes an electron transfer (ET) chain comprised of three metalloclusters distributed between the two component proteins, designated as the Fe protein and the MoFe protein. The flow of electrons through these three metalloclusters involves ET from the [4Fe-4S] cluster located within the Fe protein to an [8Fe-7S] cluster, called the P cluster, located within the MoFe protein and ET from the P cluster to the active site [7Fe-9S-X-Mo-homocitrate] cluster called FeMo-cofactor, also located within the MoFe protein. The order of these two electron transfer events, the relevant oxidation states of the P-cluster, and the role(s) of ATP, which is obligatory for ET, remain unknown. In the present work, the electron transfer process was examined by stopped-flow spectrophotometry using the wild-type MoFe protein and two variant MoFe proteins, one having the β-188(Ser) residue substituted by cysteine and the other having the β-153(Cys) residue deleted. The data support a "deficit-spending" model of electron transfer where the first event (rate constant 168 s(-1)) is ET from the P cluster to FeMo-cofactor and the second, "backfill", event is fast ET (rate constant >1700 s(-1)) from the Fe protein [4Fe-4S] cluster to the oxidized P cluster. Changes in osmotic pressure reveal that the first electron transfer is conformationally gated, whereas the second is not. The data for the β-153(Cys) deletion MoFe protein variant provide an argument against an alternative two-step "hopping" ET model that reverses the two ET steps, with the Fe protein first transferring an electron to the P cluster, which in turn transfers an electron to FeMo-cofactor. The roles for ATP binding and hydrolysis in controlling the ET reactions were examined using βγ-methylene-ATP as a prehydrolysis ATP analogue and ADP + AlF(4)(-) as a posthydrolysis analogue (a mimic of ADP + P(i)).