On the Shoulders of Giants-Reaching for Nitrogenase

Abstract

Only a single enzyme system-nitrogenase-carries out the conversion of atmospheric N₂ into bioavailable ammonium, an essential prerequisite for all organismic life. The reduction of this inert substrate at ambient conditions poses unique catalytic challenges that strain our mechanistic understanding even after decades of intense research. Structural biology has added its part to this greater tapestry, and in this review, I provide a personal (and highly biased) summary of the parts of the story to which I had the privilege to contribute. It focuses on the crystallographic analysis of the three isoforms of nitrogenases at high resolution and the binding of ligands and inhibitors to the active-site cofactors of the enzyme. In conjunction with the wealth of available biochemical, biophysical, and spectroscopic data on the protein, this has led us to a mechanistic hypothesis based on an elementary mechanism of repetitive hydride formation and insertion.

Keywords: energy conversion; enzyme mechanism; molybdenum; nitrogen fixation; nitrogenase; vanadium.

Figure 2

Evolution of structural models for the FeMo cofactor of Mo-nitrogenase. (a) The original model by Kim and Rees showed the cofactor with one µ₂-bridging ligand unassigned [6]. (b) Improving the structure to 2.0 Å resolution, the bridging ligand was identified as a third bridging sulfide, S5A [18]. (c) Only with an atomic-resolution structure, the Fourier series termination artifacts caused by the high-symmetry structure of the cluster were overcome and a central light atom was revealed [19]. (d) Through a combination of HERFD-XAS, ESEEM, and high-resolution crystallography, the central light atom was identified as a carbide in 2011 [21,22].

Figure 1

Components of the nitrogenase system and their interaction. A low-potential ferredoxin of flavodoxin reduces the reductase component (or Fe protein), which then binds 2 ATP and undergoes a conformational change that allows for complex formation with the dinitrogenase component. This triggers ATP hydrolysis, upon which an electron is transferred from the [8Fe:7S] P-cluster to the active-site cofactor, a [Mo:7Fe:9S:C]:homocitrate cluster in the case of Mo-nitrogenase (red arrows). The reduction of a single N₂ molecule to 2 NH₄⁺ requires this cycle to be repeated eight times at the expense of 16 ATP.

Figure 3

The kinetic scheme for nitrogenase catalysis according to Lowe and Thorneley [37]. The 8-electron process that reduces N₂ to 2 NH₄⁺ and releases a stoichiometric H₂ molecule upon substrate binding is defined in eight states, E₀–E₇, that correspond to single-electron transfer events. The enzyme must reach the E₄ state to bind N₂ in exchange for H₂, which is interpreted as a reductive elimination that leaves the enzyme in a super-reduced state.

Figure 4

The three variants of nitrogenase. (a) Mo-nitrogenase is encoded by nif genes and consists of the Fe protein NifH₂ and the MoFe protein NifD₂K₂ (above). MoFe protein contains [8Fe:7S] P-clusters at the DK interfaces (middle) and an active-site FeMo cofactor (below). This cofactor is a pseudo-D₃₂ symmetric [Mo:7Fe:9S:C]:homocitrate cluster attached to the protein only through its apical metal ions. (b) In the alternative V-nitrogenase, encoded by vnf genes, the Fe protein VnfH₂ works in conjunction with the VFe protein VnfD₂K₂G₂ (above). The P-cluster corresponds fully to that of the MoFe protein (middle), and the active-site FeV cofactor had Mo replaced for V, as expected (below). What was unexpected was the replacement of one µ₂-sulfide, S3A, by a carbonate anion that is not exchanged during catalysis. (c) In Fe-nitrogenase, anf genes encode the Fe protein AnfH₂ and the FeFe protein Anf D₂K₂G₂ (above). Although FeFe and VFe proteins are closely related, their Fe proteins AnfH and VnfH are more distinct than VnfH and NifH. The P-cluster of FeFe protein is highly similar to those of the isoenzymes (middle), and the FeFe cofactor (below) is a truly D₃₂ symmetric [8Fe:9S:C]:homocitrate cluster that likely corresponds to a precursor (L-cluster) of the other sites.

Figure 5

CO inhibition of the nitrogenase cofactor. Using the FeV cofactor as an example, the resting state is reduced by two electrons, leading to the release of sulfide S2B and its replacement by a bridging hydride in the E₂ state. Here, CO can bind terminally to the t-site at Fe6. If the µ-hydride is accidentally protonated and lost as H₂, the CO ligand can migrate to the µ-site, but the enzyme is formally returned to the E₀ state, now in a CO-inhibited form. Pressurization of crystals of this low-CO state with CO gas led to the structure of the high-CO state that still is an off-pathway resting state E₀, but with two CO ligands that highlight the µ- and t-site for substrate/intermediate binding.

Figure 6

Proposed mechanistic framework for nitrogenase on the example of CO reduction to the methyl stage by V- and Fe-nitrogenase. Two-electron reduction from the resting E₀ state leads to the formation of a terminal hydride at the t-site, Fe6. This t-H inserts into the bound ligand, and the cycle is repeated, leading through formyl and hydroxymethyl adducts to C-O bond cleavage, water release, and a methyl group bound to the enzyme.

Figure 7

Proposed electronic structure of the resting state E₀ for the nitrogenase cofactors. Only metals are shown. Sulfides and the central carbide are omitted for clarity. The arrow in each metal indicates the spin orientation for the high-spin systems, following the BS7 coupling scheme. As a common principle, Fe2 and Fe6 emerge as the most oxidized sites, and Fe6 takes up the electron from the Fe protein via the P-cluster. In all clusters, three additional electrons are distributed across Fe1, Fe3, Fe4, Fe5, and Fe7, with delocalization between Fe3–4 and Fe5–7. Only the heterometals or Fe8 in the FeFe cofactor have an octahedral ligand field. In this proposal, the resting state configurations add up to the apparent spins of S = 3/2 for FeMo cofactor and integer spins for FeV and FeFe cofactors.