Biosynthesis of the Iron-Molybdenum Cofactor of Nitrogenase*

The iron-molybdenum cofactor (the M-cluster) serves as the active site of molybdenum nitrogenase. Arguably one of the most complex metal cofactors in biological systems, the M-cluster is assembled through the formation of an 8Fe core prior to the insertion of molybdenum and homocitrate into this core. Here, we review the recent progress in the research area of M-cluster assembly, with an emphasis on our work that provides useful insights into the mechanistic details of this process.

The iron-molybdenum cofactor (the M-cluster) serves as the active site of molybdenum nitrogenase. Arguably one of the most complex metal cofactors in biological systems, the M-cluster is assembled through the formation of an 8Fe core prior to the insertion of molybdenum and homocitrate into this core. Here, we review the recent progress in the research area of M-cluster assembly, with an emphasis on our work that provides useful insights into the mechanistic details of this process.
Nitrogenase catalyzes the reduction of nitrogen (N 2 ) under ambient conditions (1)(2)(3). The molybdenum-dependent nitrogenase consists of a reductase component (NifH) and a catalytic component (NifDK). The active site of molybdenum nitrogenase is called the iron-molybdenum (FeMo) cofactor or the M-cluster (supplemental Fig. S1A). Located within NifDK (an ␣ 2 ␤ 2 -tetramer), the M-cluster receives electrons from NifH (a ␥ 2 -dimer) in an ATP-assisted process and subsequently serves as the site for substrate reduction upon accumulation of sufficient electrons (supplemental Fig. S1B) (4). Arguably one of the most complex metal cofactors found in nature, the M-cluster can be viewed as [Fe 4 S 3 ] and [MoFe 3 S 3 ] subclusters bridged by three sulfide atoms (5,6). Additionally, it has a homocitrate moiety attached to its molybdenum end, as well as a carbide atom coordinated in the center of its structure (6 -8). The M-cluster is ligated to the ␣-subunit of NifDK by Cys ␣275 at its iron end and His ␣442 at its molybdenum end, with Lys ␣426 providing an additional anchor for its homocitrate entity (5,6).
The structural complexity and biological importance of the M-cluster have prompted vigorous research on the assembly mechanism of this metal cofactor, as knowledge in this regard is not only important for understanding the structure-function relationship of nitrogenase but is also instrumental in developing future synthetic strategies for nitrogenase-based biomimetic catalysts. Previous studies have established that the M-cluster is synthesized stepwise on a number of scaffold proteins before it is delivered to its destined location in NifDK and that the minimum set of factors required for this process includes the gene products of nifS, nifU, nifB, nifE, nifN, and nifH (9). Here, we briefly review the recent progress in this research area, highlighting our work on the molybdenum nitrogenase from Azotobacter vinelandii that provides meaningful insights into the biosynthetic pathway of its M-cluster. An alternative view of some aspects of M-cluster assembly can be found elsewhere (10).

Synthesis of a 4Fe Cluster Pair
Biosynthesis of the M-cluster is launched by NifS and NifU, which mobilize iron and sulfur for the generation of small FeS fragments (Fig. 1). It is believed that NifS, a pyridoxal phosphate-dependent cysteine desulfurase, forms a protein-bound cysteine persulfide, which is subsequently donated to NifU for the sequential formation of [Fe 2 S 2 ] and [Fe 4 S 4 ] units (11). The [Fe 4 S 4 ] clusters are then delivered from NifU to NifB and further processed into a large FeS core (Fig. 1). This hypothesis is supported by the identification of a pair of [Fe 4 S 4 ] clusters (designated the K-cluster) on NifB, which can be easily accommodated by a sufficient amount of ligands in NifB (9). Metal analysis further reveals the presence of additional [Fe 4 S 4 ] clusters on NifB, which are associated with the conserved CxxxCxxC motifs (a characteristic feature of radical S-adenosyl-L-methionine (SAM) 3 enzymes) in this protein (9). Collectively, the K-cluster and the SAM motif-associated cluster (designated the SAM cluster) give rise to an S ϭ 1/2 EPR signal at g ϭ 2.02, 1.95, and 1.90 (12). More excitingly, this composite S ϭ 1/2 signal disappears upon addition of SAM, suggesting that the K-cluster and the SAM cluster are located in close proximity to each other (12). Such a cluster arrangement is important, as it facilitates the subsequent coupling of the two 4Fe units of K-clusters via a radical SAM-dependent route (see below).

Formation of an 8Fe Core
The K-cluster on NifB can be converted to an 8Fe core in the presence of SAM ( Fig. 2A). The concomitant disappearance of the K-cluster-and SAM cluster-originated S ϭ 1/2 signal upon addition of SAM is accompanied by the appearance of a unique g ϭ 1.94 signal, which has been attributed to an [Fe 8 S 9 ] cluster (designated the L-cluster) (12). Iron K-edge x-ray absorption spectroscopy (XAS)/extended x-ray absorption fine structure and crystallographic studies show that the L-cluster closely resembles the core structure of the mature M-cluster, except that the molybdenum and homocitrate components at one end of the cluster are replaced by an iron atom (13)(14)(15)(16). Iron K␤ x-ray emission spectroscopy analysis further demonstrates the presence of a carbide atom in the center of the L-cluster (17). Together, these observations imply that the insertion of carbide may occur simultaneously with the transformation of a K-cluster into an L-cluster. A novel synthetic route to the 8Fe L-cluster can be postulated, one that inserts the carbon atom via the radical chemistry at the SAM domain of NifB while attaching an additional sulfur atom and coupling/rearranging the two 4Fe * This work was supported, in whole or in part, by National Institutes of Health Grant GM67626 (to M. W. R.). This is the second article in the Thematic Minireview Series on Metals in Biology 2013. □ S This article contains supplemental Fig. S1. 1 To whom correspondence may be addressed. E-mail: yilinh@uci.edu. 2 To whom correspondence may be addressed. E-mail: mribbe@uci.edu.
units of the K-cluster into an [Fe 8 S 9 ] L-cluster at the same time (Fig. 2B).
The cleavage pattern of SAM provides initial insights into the role of radical SAM chemistry in this process (18). In the presence of NifB, SAM is cleaved into two products: 5Ј-deoxyadenosine (5Ј-dA) and S-adenosyl-L-homocysteine. Upon substitution of [methyl-d 3 ]SAM for unlabeled SAM, however, a mixture of deuterium-enriched and unlabeled 5Ј-dA can be detected along with S-adenosyl-L-homocysteine as products of SAM cleavage (18). Interestingly, two radical SAM RNA methylases, RlmN and Cfr, display the same SAM cleavage and deuterium substitution patterns as those of NifB, which have been accounted for by an S N 2-type of methyl transfer from one equivalent of SAM, followed by the generation of 5Ј-dA ⅐ from a second equivalent of SAM and the subsequent hydrogen atom abstraction from the methyl group by 5Ј-dA ⅐ (19,20). The striking parallelism between NifB and the two RNA methylases suggests that NifB utilizes a similar mechanism to mobilize the methyl group of SAM for carbon insertion. Radiolabeling experiments provide convincing evidence for this proposal, showing that the incubation of NifB with [methyl-14 C]SAM leads to an accumulation of 14 C label on NifB, which can be further traced to the L-cluster in this protein (18).
Two mechanisms can be proposed for the formation of the L-cluster on NifB (18). One involves the transfer of the methyl group of SAM to a sulfide atom of the K-cluster via an S N 2 mechanism, followed by the formation of a methylene radical through the abstraction of a hydrogen atom from this methyl group (Fig. 2C, left); the other involves the formation of a methyl radical via the reductive cleavage of SAM, followed by the transfer of this radical to an iron atom of the K-cluster and the subsequent rearrangement of this radical into a methylene radical (Fig. 2C, right). In either scenario, the transfer of the carbon intermediate to the K-cluster may play a pivotal role in coupling the two 4Fe units of the K-cluster into an 8Fe L-cluster, initiating radical chemistry for the initial bond rearrangement that is required for the subsequent restructuring of cluster units. Regardless of the type of the initial carbon intermediate, it must undergo additional deprotonation and/or dehydrogenation steps until a carbide atom is generated in the center of the L-cluster (Fig. 2C).

Maturation of the 8Fe Core
Once formed, the L-cluster is transferred from NifB to NifEN, where it is matured into an M-cluster upon the insertion of molybdenum and homocitrate (Fig. 3A). Such a conversion can be achieved in vitro by incubating L-cluster-bound NifEN with NifH, MgATP, dithionite, molybdate (MoO 4 2Ϫ ) and homocitrate, generating an NifEN-bound M-cluster (Fig. 3B) that can be used for the subsequent reconstitution and activation of apo-NifDK (21,22). Maturation of the cluster species on NifEN is reflected by the disappearance of the L-cluster-specific g ϭ 1.94 signal and the concurrent appearance of a small M-cluster-like signal at g ϭ 4.45, 3.96, 3.60, and 2.03 (21,22). Iron and molybdenum K-edge XAS/extended x-ray absorption fine structure analyses further confirm that the structure of the M-cluster on NifEN (Fig. 3B, right) is nearly identical to that of the M-cluster on NifDK (supplemental Fig. S1A), except for a somewhat asymmetric coordination of molybdenum that arises from a different ligand environment in NifEN (21,22).
Interestingly, biochemical experiments suggest that NifEN undergoes a conformational change upon the conversion of the L-cluster to the M-cluster, as only the M-cluster-bound form of NifEN is capable of complex formation with NifDK (23). Consistent with this suggestion, crystallographic analysis of the L-cluster-bound form of NifEN reveals an unusual, nearly surface-exposed location of the L-cluster (16). This observation is unexpected, as the highly homologous NifEN and NifDK proteins should have homologous cluster-binding sites that are buried within their respective polypeptides. It is conceivable that the surface location of L-cluster permits an easier access of molybdenum and homocitrate for cluster maturation, whereas a concurrent conformational change of NifEN enables subsequent relocation of the matured cluster from the surface to the binding site within the protein (Fig. 3A).
The conformational change of NifEN is likely induced by its interaction with NifH upon cluster maturation, which parallels the interaction between NifDK and NifH upon substrate turnover. Indeed, NifH exhibits a strict dependence on ATP hydrolysis and redox potential to carry out its function both in catalysis (1-3) and in assembly (21,22). However, the role of this ATPase in assembly appears more complicated, as NifH can be "loaded" with molybdenum and homocitrate and subsequently used as a donor of these two missing components to the NifENbound L-cluster (24). Biochemical analysis shows a mandatory co-mobilization of molybdenum and homocitrate by NifH, whereas molybdenum K-edge XAS analysis reveals that the molybdenum species is "processed" upon binding to NifH, dis- playing a change in its formal oxidation state and/or the ligation pattern (24). More interestingly, EPR analysis suggests that NifH binds nucleotide along with molybdenum and homocitrate, showing that the loaded NifH protein gives rise to a signal that has an intermediary line shape between those of the ADP-and ATP-bound forms of NifH (24). This observation coincides with the outcome of the initial crystallographic analysis of an ADP-bound form of NifH, which places molybdenum at a position that corresponds to the ␥-phosphate of ATP (25). A binding pattern of NifH/ADP-molybdenum-homocitrate can be  proposed based on these results, which could represent the initial step of molybdenum mobilization by NifH (Fig. 3C). Such a binding pattern parallels the formation of a possible adenylylated molybdate intermediate during the biosynthesis of molybdopterin cofactors (9). Moreover, another ATPase, CooC, has been implicated in the insertion of nickel into the C-cluster of the carbon monoxide dehydrogenase from Rhodospirillum rubrum (26). Thus, the mobilization of molybdenum by NifH may exemplify a general scheme for metal trafficking in biological systems.

Delivery of the M-cluster to Its Target Location
The final "touch-up" of the structure of the M-cluster is achieved upon delivery of this cluster from NifEN to NifDK. The completion of M-cluster assembly on NifEN signals the cluster transfer from NifEN to apo-NifDK, which begins with the complex formation between these two homologous proteins (23). Sequence analysis suggests the presence of homologous M-cluster-binding sites in NifEN and NifDK; however, certain residues that either provide a covalent ligand to the M-cluster or tightly pack the M-cluster within the polypeptide matrix of NifDK cannot be found in the sequence of NifEN (9). Thus, it can be postulated that the M-cluster-binding sites in NifEN and NifDK are "lined up" upon complex formation, which enables the "diffusion" of the M-cluster from its transient binding site in NifEN ("low affinity site") to its final binding site in NifDK ("high affinity site") (Fig. 4A). Given the structural homology between NifEN and NifDK, the process of M-cluster insertion into apo-NifDK could be depicted by three available "snapshots" of NifDK and/or NifEN: (i) an apo conformation, which contains a positively charged cluster insertion funnel; (ii) an intermediate conformation, which is generated upon docking of the cluster at the entrance of the insertion funnel; and (iii) a holo conformation, which has the cluster secured at its final binding site (Fig. 4B). The negative charge of homocitrate is crucial for the insertion of the M-cluster along the positively charged funnel (23,27), and the M-cluster interacts with a number of NifDK residues en route to its target binding site (28 -30), where the final structure of the M-cluster is achieved through proper ligand coordination.

Conclusion
The M-cluster is assembled on a number of scaffold proteins prior to its delivery to its target location in NifDK. This process involves two key steps: the stepwise formation of an 8Fe core and the subsequent maturation of this core upon incorporation of molybdenum and homocitrate. The former utilizes radical SAM chemistry for the insertion of interstitial carbide, which likely occurs concomitantly with the coupling and restructuring of two 4Fe units, whereas the latter employs a nucleotidebinding protein for the mobilization of molybdenum and homocitrate, which may represent a common strategy for metal trafficking. Future studies will focus on the mechanistic investigation of radical SAM-dependent carbon insertion, as well as the structural details of the NifH/ATPase-based molybdenum mobilization.