Nitrogenase Cofactor Assembly: An Elemental Inventory - Accounts of

Oct 24, 2017 - The nomenclature of proteins and genes in this account follows the convention in the field, with the first letters of the names of prot...
0 downloads 11 Views 3MB Size
Article pubs.acs.org/accounts

Cite This: Acc. Chem. Res. 2017, 50, 2834-2841

Nitrogenase Cofactor Assembly: An Elemental Inventory Nathaniel S. Sickerman,† Markus W. Ribbe,*,†,‡ and Yilin Hu*,† †

Department of Molecular Biology and Biochemistry, University of California, Irvine, California 92697-3900, United States Department of Chemistry, University of California, Irvine, California 92697-2025, United States



CONSPECTUS: Nitrogenase is known for its remarkable ability to catalyze the reduction of N2 to NH3, and C1 substrates to short-chain hydrocarbon products, under ambient conditions. The best-studied Mo-nitrogenase utilizes a complex metallocofactor as the site of substrate binding and reduction. Designated the M-cluster, this [MoFe7S9C(Rhomocitrate)] cluster can be viewed as [MoFe3S3] and [Fe4S3] subclusters bridged by three μ2-sulfides and one μ6-interstitial carbide, with its Mo end further coordinated by an Rhomocitrate moiety. The unique cofactor has attracted considerable attention ever since its discovery; however, the complexity of its structure has hindered mechanistic understanding and chemical synthesis of this cofactor. Motivated by the pressing questions related to the structure and function of the nitrogenase cofactor, one major thrust of our research has been to unravel the key biosynthetic steps of this metallocluster to cultivate a deeper understanding of these reactions and their effects on functionalizing the cofactor. In this Account, we will discuss our recent work that provides insights into how simple Fe and S atoms, along with a single C atom, a heterometallic Mo atom and an organic homocitrate entity, are assembled into one of the most complex metalloclusters known in Nature. Combined biochemical, spectroscopic and structural studies have led us to a working model of M-cluster assembly, which starts with the sequential synthesis of small [Fe2S2] and [Fe4S4] units by NifS/U, followed by the coupling and rearrangement of two [Fe4S4] clusters on NifB concomitant with the insertion of an interstitial carbide and a “9th sulfur” that give rise to a [Fe8S9C] core that is nearly indistinguishable in structure to the M-cluster except for the absence of Mo and homocitrate. This 8Fe core is then matured into an M-cluster on NifEN upon substitution of a Mo-homocitrate conjugate for one terminal Fe atom of the cluster prior to transfer of the M-cluster to its target binding site in the catalytic component of Monitrogenase. Taking stock of the elemental inventory during the cofactor assembly process, the core Fe and S atoms are derived from modular fusion of FeS building blocks, going through 2Fe, 4Fe and 8Fe stages to generate an 8Fe core of the cofactor. However, such a flow of Fe/S along the biosynthetic pathway of the M-cluster is “intervened” by the insertion of C and Mo, which renders the cofactor unique in structure and reactivity. Insertion of C occurs through a novel, radical SAM-dependent mechanism, which involves SN2-type methyl transfer from SAM to a [Fe4S4] cluster pair, hydrogen abstraction of the transferred methyl group by a SAM-derived 5′-dA· radical, and further deprotonation of the resultant methylene radical concomitant with radical chemistry-based coupling and rearrangement of the [Fe4S4] cluster pair into an [Fe8S9C] core. Insertion of Mo, on the other hand, employs an ATPase-dependent mechanism that parallels metal trafficking in the biosynthesis of molybdopterin and CO dehydrogenase cofactors. These findings provide a nice framework for further exploration of the “black box” of nitrogenase cofactor assembly and function.



INTRODUCTION Nitrogenase is a metalloenzyme capable of catalyzing a number of challenging reductive reactions. While nitrogenase is evolutionarily optimized for the catalytic reduction of dinitrogen (N2) to ammonia (NH3),1−3 it can also reduce C1 substrates, such as carbon monoxide (CO) and carbon dioxide (CO2), to short-chain hydrocarbon products.4−6 These two main types of chemical transformations by nitrogenase represent biological equivalents to the industrial Haber− Bosch (HB) and Fischer−Tropsch (FT) processes.7,8 The global importance of the HB and FT processes for fertilizer and carbon fuel production, respectively, has spurred research into improving catalyst performance in these reactions. Both HB and FT processes use heterogeneous transition metal catalysts, with protons and electrons supplied by hydrogen (H2); and both of them require very high temperatures and pressures to operate effectively. In contrast, nitrogenase performs HB- and © 2017 American Chemical Society

FT-like reactions under ambient conditions using discrete metalloclusters as the catalysts, as well as protons and electrons in aqueous solutions as the reducing power. Insights into how these homogeneous reactions can be carried out under mild conditions will facilitate efforts toward harnessing this chemistry on a larger scale. In particular, understanding the biosynthesis and structure−function relationship of the catalytic metallocofactor of nitrogenase could enable future development of biomimetic catalysts that perform ambient HB- and FT-types of chemistry. Most of our current knowledge of nitrogenase is derived from the molybdenum (Mo)-dependent form of this enzyme. The Mo-nitrogenase consists of two component proteins. One, designated the Fe protein (or NifH),9 is a homodimer Received: August 27, 2017 Published: October 24, 2017 2834

DOI: 10.1021/acs.accounts.7b00417 Acc. Chem. Res. 2017, 50, 2834−2841

Article

Accounts of Chemical Research

major advances toward understanding these complex biosynthetic events, critical questions remain concerning the source and fate of some atomic and molecular species along the cofactor assembly pathway, as well as the order of events and operative species in many assembly steps. This Account will discuss work in the Ribbe and Hu laboratories that sheds light on some of the “black-box” events of M-cluster assembly, with an emphasis on providing an elemental inventory at each step of this process.

containing a subunit-bridging [Fe4S4] cluster and an MgATPbinding site in each subunit; the other, designated the MoFe protein (or NifDK), is a heterotetramer containing a P-cluster ([Fe8S7]) at the α/β-subunit interface and an M-cluster (or cofactor; [MoFe7S9C(R-homocitrate)]) within the α-subunit.10,11 Catalysis by Mo-nitrogenase involves repeated association and dissociation between NifH and NifDK, which permits interprotein, ATP-dependent electron transfer from the [Fe4S4] cluster of NifH, via the P-cluster, to the M-cluster of NifDK, where substrate reduction eventually occurs upon accumulation of a sufficient number of electrons (Figure 1).1,2,12 While in vitro methods have been developed to



GENERATION OF 4Fe UNITS ON NifS/U The initial FeS building blocks are proposed to be assembled through the concerted actions of NifS and NifU, which mobilize Fe and S for the synthesis of small FeS building blocks. This process utilizes the function of NifS as a pyridoxal phosphate-dependent cysteine desulfurase to form a proteinbound cysteine persulfide, which is subsequently donated to the assembly scaffold, NifU, for the synthesis and reductive coupling of [Fe2S2] fragments into [Fe4S4] clusters.15,16 Following this event, a pair of [Fe4S4] clusters are transferred to the next biosynthetic scaffold, NifB, where they are processed into an 8Fe core of the cofactor (designated Lcluster; Figure 2). While details of the reactions catalyzed by

Figure 1. Structure of the Mo-nitrogenase complex. Shown is 1/2 MgADP·AlF4−-stabilized complex comprising the dimeric NifH and one-half of the α2β2-tetrameric NifDK. In a process coupled to ATP hydrolysis, electrons flow from the [Fe4S4] cluster of NifH, via the Pcluster, to the M-cluster of NifDK, where substrate reduction takes place. Atoms are colored as follows: Fe, orange; S, yellow; C, gray; Mo, teal; O, red; N, blue; P, tangerine; Mg, green; Al, light blue. PDB ID: 1N2C.

perform substrate reduction under ATP-independent conditions (i.e., without NifH),13,14 the M-cluster is absolutely necessary for catalysis.1 Coordinated by only two amino acids (i.e., Cysα275 and Hisα442) to NifDK, this metallocofactor can be extracted intact into an organic solvent (e.g., Nmethylformamide) and inserted back into an inactive, cofactor-deficient form of NifDK to restore catalytic activity, demonstrating its essential role in the enzymatic function of nitrogenase.15−17 How the 18-atom core of the M-cluster, a motif with no parallel in the biological or synthetic realms, is assembled is a pressing question central to the advancement of both fundamental and application-based nitrogenase research. Combined genetic, biochemical, spectroscopic, and structure studies have allowed us to identify a minimum set of proteins required for cofactor biosynthesis, as well as key events that occur in this process.15,16 Based on these studies, the current model of M-cluster biosynthesis involves three major synthetic platforms: NifU, which interacts with NifS to mobilize Fe and S for the synthesis of small [Fe4S4] units; NifB, where two [Fe4S4] units are coupled and rearranged into a [Fe8S9C] core concomitant with the insertion of an interstitial carbide and a “9th sulfur”; and NifEN, where the [Fe8S9C] core is matured into an M-cluster upon replacement of the “8th Fe” of this core by Mo and homocitrate via the action of NifH.15,16 The molecular transformations that take place on these scaffolds are of particular interest, as these reactions involve the influx and efflux of single atoms and simple molecular fragments that are challenging processes in their own right. However, despite

Figure 2. Formation of an 8Fe core of the M-cluster. Shown are the transformation of cluster species (A), the scaffold proteins housing the cluster conversion (B), and the elemental inventory of cluster intermediates (C) during this process. The formation of an 8Fe core starts with the mobilization of Fe and S by NifS/U for the synthesis of [Fe4S4] units. This event is followed by transfer of a [Fe4S4] cluster pair (K-cluster) to NifB, where the two [Fe4S4] clusters are coupled and rearranged into an [Fe8S9C] core (L-cluster) concomitant with insertion of an interstitial C and a “9th” S. Atoms are colored as follows: Fe, orange; S, yellow; C, gray; O, red. HC, homocitrate.

NifS/U are yet to be fully elucidated, the elemental inventory of this initial stage of cofactor assembly is rather straightforward, with four Fe ions and four NifS-donated S atoms making up the composition of a [Fe4S4] cluster (Figure 2, left).



FORMATION OF AN 8Fe CORE ON NifB The formation of an 8Fe core of the cofactor occurs on NifB, a radical S-adenosyl-L-methionine (SAM) enzyme that is highly 2835

DOI: 10.1021/acs.accounts.7b00417 Acc. Chem. Res. 2017, 50, 2834−2841

Article

Accounts of Chemical Research

Figure 3. Reactivity of NifB with SAM. The perpendicular-mode EPR spectra of reduced (A) and oxidized (B) NifBMa (①) and NifBMt (②) in the absence (black) and presence (red) of SAM, and (C) the HPLC traces of the standards containing SAM, SAH and 5′-dAH (⓪) and the reactions containing SAM and NifBMa (①) and NifBMt (②). NifBMa, NifB of Methanosarcina acetivorans; NifBMt, NifB of Methanothermobacter thermoautotrophicus. The g values are indicated.

dAH is detected together with SAH as the cleavage products.25 These cleavage profiles suggest that the initial steps catalyzed by NifB involve the action of two equivalents of SAM, with the first equivalent of SAM transferring a methyl group to the substrate and producing SAH as a byproduct, and the second equivalent of SAM undergoing homolytic cleavage and forming a highly reactive 5′-adenosyl radical (5′-dA·) species that abstracts a hydrogen atom from the methyl group to yield 5′dAH. This observation is interesting, as it suggests the SAM methyl group as a potential source of the interstitial carbide that is incorporated into the L-cluster via a novel, radical SAMdependent biosynthetic route. Radiolabeling experiments provide strong support for this proposal, showing accumulation of the 14C label on NifB upon incubation with [methyl-14C]SAM, and tracing of the radiolabel into the L-cluster upon extraction from NifB.20,25 M-cluster maturation assays further demonstrate transfer of the radiolabel to NifDK when the 14Clabeled L-cluster is transferred from NifB to NifEN, matured into an M-cluster and inserted into NifDK; and subsequent extraction of the 14C-labeled M-cluster from the radiolabeled NifDK provides the definitive proof that the methyl group of SAM gives rise to the interstitial carbide of the cofactor.20,25 The observation that 14 C label is absent from the polypeptides of NifB points to a direct transfer of the SAMderived methyl group to the K-cluster without routing through a protein-bound C intermediate species.25 Consistent with this suggestion, acid treatment of an incubation mixture containing NifB and either unlabeled SAM or [methyl-d3] SAM, followed by identification of the liberated gaseous products by gas chromatography−mass spectrometry (GC-MS), reveals the formation of methanethiol (CH3SH) or methane-d3-thiol (CD3SH) in these reactions, suggesting that the initial site of SAM-mediated methylation is a labile S atom of the K-cluster, despite the fact that the interstitial carbide is hexacoordinated by Fe atoms in the L-cluster (Figure 4).26 Substitution of Se for S in the FeS clusters of NifB results in the formation of methylselenol (CH3SeH) as a product of acid quenching, further supporting the assignment of a cluster-associated S (or Se) atom as the initial site of SAM methylation. The attachment of methyl to an acid labile S atom of the K-cluster is somewhat analogous to the initial methylation step in the process of C−C bond formation by Cfr and RlnM, which is proposed to occur via an SN2 reaction that attaches a methyl

conserved across organisms harboring the nitrogenase enzyme. Strains of Azotobacter vinelandii that lack the nif B gene express a cofactor-deficient NifDK (designated apo NifDK) that is incapable of nitrogen fixation,15,17 suggesting that NifB plays an essential role in nitrogenase cofactor assembly; yet, how exactly NifB couples two [Fe4S4] clusters with an interstitial C and a “9th S” to form a [Fe8S9C] core, a transformation found in no other biological system, has remained elusive for a long time. Analysis of the amino acid sequence of NifB provides hints of how cluster fusion might be accomplished, showing the presence of a CX3CX2C motif at the N-terminus of the protein that is characteristic of the radical SAM enzymes, as well as a number of conserved cysteine and histidine residues that could potentially accommodate the entire complement of Fe atoms of the cofactor.18 Consistent with this observation, a SAM motifbinding [Fe4S4] cluster (designated the SAM-cluster) and a pair of proximal [Fe4S4]-like clusters (designated the K-cluster) have been identified on NifB (Figure 2A and B, middle).18−20 The SAM- and K-clusters of the reduced NifB give rise to a composite S = 1/2 EPR signal, which disappears upon addition of SAM (Figure 3A). Interestingly, the disappearance of this signal corresponds to the appearance of a diagnostic, g = 1.94 signal in the oxidized, SAM-treated NifB sample (Figure 3B), which corresponds to the formation of a [Fe8S9C] cluster (8 Fe core or L-cluster) that closely resembles the core structure of a mature cofactor.21,22 Together, these observations suggest that NifB utilizes a SAM-dependent mechanism to convert two [Fe4S4] clusters into an [Fe8S9C] cluster;23,24 however, there are questions regarding the elemental inventory in this process: while the 8 Fe atoms of the L-cluster are accounted for by the two 4Fe units of the K-cluster, a pair of C and S atoms would need to be incorporated concomitant with the rearrangement of the two 4Fe units to complete the stoichiometry and structure of the L-cluster. The profile of cleavage products from the reaction of SAM with NifB provides the first insight into the origin of the C atom and how it may be incorporated into the L-cluster. Monitoring the reaction with high-performance liquid chromatography−mass spectrometry (HPLC-MS), S-adenosyl-homocysteine (SAH) and 5′-deoxyadenosine (5′-dAH) are detected as the SAM cleavage products upon incubation with NifB (Figure 3C).20,25,26 When NifB is treated with methyl-d3SAM, however, a mixture of monodeuterated 5′-dAD and 5′2836

DOI: 10.1021/acs.accounts.7b00417 Acc. Chem. Res. 2017, 50, 2834−2841

Article

Accounts of Chemical Research

undergo homolytic cleavage to generate 5′-dA· for the subsequent hydrogen atom abstraction. By analogy, the fact that allylation can occur without hydrogen atom abstraction indicates that methylation of the K-cluster precedes the radicalmediated steps that convert the methyl group into an interstitial carbide. A mechanism can be proposed for carbide insertion based on these observations (Figure 4), which involves SN2-type transfer of the methyl group of SAM to a K-cluster-associated S atom, followed by formation of a putative methylene radical (S− CH2·) species upon abstraction of a hydrogen atom from the transferred methyl group by 5′-dA·.25 The reactive S−CH2· unit, which is attached to one [Fe4S4] module of the K-cluster, then binds to an Fe atom of the second [Fe4S4] module of the K-cluster to initiate cluster fusion and rearrangement. The Lcluster architecture is completed by double deprotonation of the methylene intermediate, reorganization of the cluster via ligand exchange to embed the central C atom, and addition of a “9th S” atom. It is plausible that deprotonation of the SAMderived methylene unit is coupled with protonation of a pair of His or Arg residues that are proposed to coordinate the cluster in NifB,24 thereby labilizing the newly formed L-cluster for its transfer from NifB to NifEN. However, details of this process are yet to be elucidated. Likewise, the stage at which the “9th S” enters the L-cluster formation pathway, as well as the mechanism of the radical-based cluster rearrangement that occurs concomitant with the formation of carbide, has not yet been determined. In particular, the incorporation of the “9th S” not only plays a key role in the formation of the 8Fe core of the cofactor, but also completes the elemental inventory of the biosynthetic events that occur on NifB. Tracking the source of the “9th S” atom has proven to be a challenge. Drawing analogy to similar systems, the S atom may be derived from the S2− ions of the SAM-motif cluster, similar to the proposed mechanism of the biotin synthase protein BioB that requires the sacrificial SAM-motif cluster to be reassembled after each turnover.28 Alternatively, the S atom could also come from a persulfide donor, perhaps from the cysteine desulfurase NifS,29 and any one of the conserved cysteine residues on NifB could potentially serve as a reservoir for the S0 atoms of persulfide.30 It is interesting to note that the in vitro preparations of NifB use dithionite as reductant, which is known to decompose into a number of S-based compounds, including thiosulfate (S2O32−) and sulfite (SO32−).31 These breakdown products could potentially donate their S atoms at some point along the L-cluster formation pathway in the in vitro assay and, given that some of these breakdown products (e.g., SO32−) represent central hubs of sulfur metabolism in cells, they could serve as a source of the “9th S” during the in vivo cofactor assembly process. Identification of the origin and the insertion mechanism of the “9th S”, therefore, will further refine the cofactor assembly pathway while establishing a potential link between nitrogen fixation and sulfur metabolism.

Figure 4. Proposed mechanism of carbide insertion. Two equivalents of SAM facilitate transformation of the K-cluster to an L-cluster on NifB: the first equivalent donates a methyl group to an S atom of the K-cluster, and the second gives rise to a 5′-dA· radical that abstracts an H atom from the methyl group. These events lead to the formation of a K-cluster associated CH2· radical. Subsequent deprotonation of the CH2· group occurs concomitant with radical-based coupling and rearrangement of the cluster, and addition of a “9th S” atom, which results in the formation of a [Fe8S9C] L-cluster with a μ6-coordinated carbide ion in the center of the cluster.

group to a S atom of a cysteine residue.27 However, in the case of NifB, methyl transfer only occurs in the presence of a reductant, suggesting that the K-cluster needs to be poised at a certain redox potential to render its S atom more nucleophilic for the transfer of methyl group from SAM via an SN2-type nucleophilic substitution. Use of an allyl SAM analogue in which the methyl group is replaced by an allyl (CH2CH−CH2−) moiety sheds light on the sequence of events between methyl transfer and hydrogen abstraction during carbide insertion.26 After treating the reduced NifEN-B with allyl-SAM, the resulting HPLC-MS trace shows that only SAH is generated, with no detectable trace of 5′-dAH. This result is complemented by GC-MS headspace analysis of the acid-quenched reaction mixture, which shows that allylthiol (CH2CH−CH2−SH) is liberated following such a treatment. Together, these important results delineate the sequence of SAM reactivity within NifB, suggesting allylation of a labile S atom of the K-cluster by the first allyl-SAM equivalent, followed by binding of a second allylSAM equivalent to the SAM-motif cluster that is unable to



MATURATION OF THE 8Fe CORE ON NifEN Upon completion on NifB, the 8Fe core (or L-cluster) was transferred to NifEN, where it is matured into an M-cluster upon interaction with NifH, prior to delivery of the M-cluster to its target binding site in NifDK (Figure 5). Sequence alignment indicates that NifEN and NifDK are highly homologous to each other. NifEN possesses four out of the six cysteine ligands associated with the P-cluster binding site in NifDK, accommodating a simple [Fe4S4] cluster instead of the 2837

DOI: 10.1021/acs.accounts.7b00417 Acc. Chem. Res. 2017, 50, 2834−2841

Article

Accounts of Chemical Research

Figure 5. Maturation of the 8Fe core into an M-cluster. Shown are the transformation of cluster species (A), the scaffold proteins housing the cluster conversion (B), and the elemental inventory of cluster intermediates (C) during this process. The maturation of an 8Fe core starts with the transfer of the L-cluster from NifB to NifEN. This event is followed by replacement of one terminal Fe atom by Mo and homocitrate by NifH in an ATPdependent process, and the subsequent transfer of the resulting M-cluster to its binding site in NifDK. Atoms are colored as follows: Fe, orange; S, yellow; C, gray; Mo, teal; O, red. Dashed lines indicate interprotein cluster transfer events, and solid lines indicate intraprotein cluster transformation processes. HC, homocitrate.

[Fe8S7] P-cluster; it also contains a region that approximates the active site of NifDK, preserving one M-cluster-binding residue with a Cys residue and replacing the other by an Asn residue. The substantial homology between NifEN and NifDK has led to the proposal that NifEN is involved in cofactor biosynthesis; however, how this protein functions in this process has remained unclear for a long time. The first insight into the role of NifEN in cofactor maturation is obtained through the accumulation of L-clusters on NifEN (designated NifENL) in a nif HDK-deletion background, where NifH (an essential protein for maturing the Lcluster into an M-cluster) and NifDK (the downstream acceptor for the M-cluster) are absent.32 Fe K-edge XAS/ EXAFS and crystallographic analyses demonstrate a nearly indistinguishable core structure of this cluster from that of the M-cluster,22,33 whereas biochemical and Fe Kβ XES studies further reveal the presence of an interstitial carbide in this [Fe8S9C] cluster.34 Moreover, incubation of NifENL with dithionite, NifH, MgATP, molybdate, and homocitrate results in a form of NifEN (designated NifENM) that exhibits spectroscopic signals indicative of a mature M-cluster. The S = 3/2 EPR signal of NifENM resembles that displayed by the M-cluster of NifDK, although it is broader in line-shape and smaller in magnitude (Figure 6A).35 Likewise, the Fe K-edge XAS/EXAFS features of NifENM are highly similar to those of NifDK (Figure 6B), despite some differences between their Mo K-edge XAS spectra (Figure 6C).35 The spectral deviations between NifENM and NifDK are attributed to differences in the respective protein environments of the cofactors, especially for the Mo atom in NifENM, which has Asn in place of the Mobinding His residue of NifDK. Consistent with this argument, the M-cluster extracted from NifENM displays EPR and XAS/ EXAFS features that are nearly indistinguishable from those of the M-cluster extracted from NifDK, further demonstrating that

Figure 6. Transformation of L-cluster on NifEN. The perpendicularmode EPR (A), Fe K-edge XAS (B), and Mo K-edge XAS (C) spectra of NifENM (red) and NifDK (black), and comparison of total (red) and chelated (striped) Fe atoms of apo NifEN, NifENL, and NifENM (D). Apo NifEN, L/M-cluster-depleted NifEN; NifENL, L-clusterbound NifEN; NifENM, M-cluster-bound NifEN.

the M-cluster is fully assembled on NifEN prior to its transfer to NifDK.36 Maturation of the L-cluster into an M-cluster seems to trigger a conformational change of NifEN. This proposal is supported by chemical chelation experiments of NifENL and NifENM (Figure 6D), which suggests a conformational change that embeds the nascent cofactor within a less solventaccessible region of the α-subunit.35 Interestingly, an analogous conformational change also occurs in NifDK, as structural and small-angle X-ray scattering (SAXS) studies reveal a compact2838

DOI: 10.1021/acs.accounts.7b00417 Acc. Chem. Res. 2017, 50, 2834−2841

Article

Accounts of Chemical Research

scenario. Rather, one possibility for Fe sequestration could be associated with the proposed “16th Fe” site found in NifDK, which coodinates an extra Fe atom in addition to the 15 metallocluster-derived Fe atoms (i.e., 8Fe from the P-cluster, and 7 Fe from the M-cluster) within one αβ-dimer of NifDK (Figure 7A). This octahedral metal site was originally attributed

ing effect caused by incorporation of the M-cluster, presumably locking the cofactor within the α-subunit.37 Biochemical experiments demonstrate that NifEN is only capable of forming a complex with the cofactor-deficient, apo NifDK upon conversion of NifENL to NifENM, suggesting that the conformational change of NifEN upon maturation of the Mcluster is necessary for NifEN to dock on NifDK to enable delivery of the cofactor from the former protein to the latter protein.15,16 Transfer of the M-cluster between these two structurally homologous proteins is likely facilitated by the preservation of only one of the two M-cluster-binding residues of NifDK (i.e., Cysα275) in NifEN (i.e., Cysα250), as well as the absence of several residues from NifEN that presumably lock the M-cluster within the active site of NifDK, which favors the release of the cofactor from NifEN once its assembly is completed. Upon release, the cofactor could travel from the binding site within the α-subunit of NifEN to the surface of the protein, where it is “relayed” to the surface of NifDK through the coordinated actions of cysteinyl ligands, such as the Cysα25 in NifEN and the corresponding Cysα45 in NifDK, prior to being inserted into its final binding site within the α-subunit of NifDK.21 The cluster maturation process on NifEN finalizes the elemental inventory of the cofactor via the replacement of one terminal Fe atom of the L-cluster by a heterometal, Mo, concomitant with the attachment of an organic compound, homocitrate, to Mo (Figure 5C). The ATP-dependent nature of this process suggests a key role of NifH, an ATPase, in maturing the L-cluster to an M-cluster. Consistent with this argument, reisolation of NifH after incubation with NifEN, MgATP, dithionite, molybdate and homocitrate results in a form of NifH that is loaded with Mo and homocitrate and capable of maturing the L-cluster on NifEN, suggesting that NifH serves as a donor of these two missing components for the maturation of the L-cluster via interaction with NifEN.38 Mo K-edge XAS analysis indicates that, upon association with NifH, the Mo atom in molybdate experiences a change in its coordination environment and a reduction in its formal oxidation state, which aligns well with the previously reported activity of NifH in reducing molybdate in the presence of ATP. Interestingly, biochemical experiments suggest an obligate comobilization of Mo and homocitrate by NifH36,38 for the maturation of L-cluster on NifEN. EPR analysis further reveals an intermediary line-shape of the Mo/homocitrate-bound NifH between those of the ADP- and ATP-bound forms of NifH,38 which coincides with binding of Mo at a position that approximates the γ-phosphate of ATP in a conformation of MgADP-bound NifH that is crystallized in the presence of molybdate.39 Together, these observations point to an initial attachment of the Mo atom to the nucleotide, a process that parallels the appearance of an adenylylated molybdate intermediate during the assembly of molybdopterin cofactors,40 or the formation of a nucleotide-metal conjugate by CooC during the process of nickel insertion into the C cluster of the carbon monoxide dehydrogenase (CODH).41 On the other hand, participation of the NifH-associated [Fe4S4] cluster in this process cannot be excluded, which would involve a potential swap between one cluster Fe atom and a Mo atom that is further attached to a homocitrate entity. The fate of the “8th Fe” that is removed from L-cluster concomitant with the insertion of Mo is unknown. Since free Fe concentrations in cells are tightly regulated, a random “ejection” of the Fe atom into the bulk solvent is an unlikely

Figure 7. Possible fate of the “8th” Fe of the L-cluster. (A) Proposed binding site for the “16th Fe” in NifDK. (B) Plausible binding site for the “8th Fe” in NifEN based on homology between NifEN and NifDK. (C) Partial sequence alignment of NifK and NifN indicating conserved residues in the “extra Fe” binding site. Atoms are colored as follows: Fe, orange; C, gray; O, red; N, blue.

to either Ca2+ or Na+, but XAS studies suggest occupancy of this site by a ferrous center. While the role of this single Fe site is unclear, the residues that bind this Fe center in NifDK are completely conserved in NifEN (Figure 7B and C). Based on the structural homology between NifEN and NifDK, it is plausible that an analogous single Fe site becomes occupied in NifEN during the conversion of L- to M-cluster; however, a high-resolution structure of NifENM would be required to evaluate this idea. Other potential destinations for the “8th Fe atom” include NifH, which facilitates the handoff of Mo and homocitrate to NifEN. Alternatively, some other Fe sequestration proteins, such as NifU or ferritin, may participate in this process under in vivo conditions.



SUMMARY AND OUTLOOK The assembly of the M-cluster follows a modular pattern in the formation of the “FeS backbone”, sequentially generating Fe2S2, Fe4S4 and 2xFe4S4 fragments along the assembly pathway that easily account for the inventory of the core Fe and S elements in this process. However, such an Fe/S flow is “intercepted” by the influx and efflux of certain elements, including the incorporation of a central C and a “9th S” concomitant with the coupling and rearrangement of two [Fe4S4] clusters into an [Fe8S9C] core, and further maturation of this core into a [MoFe7S9C(R-homocitrate)] M-cluster upon removal of an “8th Fe” concomitant with the insertion of Mo and homocitrate. Significant progress in the past decade has uncovered many facets of the composition and assembly of the nitrogenase cofactor; in particular, it has unveiled a novel biosynthetic route to bridged, high-nuclearity metalloclusters that relies on radical SAM chemistry (i.e., the carbide insertion pathway), as well as an ATPase dependent mechanism for metal trafficking (i.e., the NifH-mediated Mo insertion). While the elemental inventory of M-cluster biosynthesis has been clarified substantially, many open questions still remain to be answered, such as what is the source of the “9th S” in L-cluster 2839

DOI: 10.1021/acs.accounts.7b00417 Acc. Chem. Res. 2017, 50, 2834−2841

Article

Accounts of Chemical Research

(9) The nomenclature of proteins and genes in this account follows the convention in the field, with the first letters of the names of proteins capitalized and the names of genes italicized in lower-case letters. (10) Lancaster, K. M.; Roemelt, M.; Ettenhuber, P.; Hu, Y.; Ribbe, M. W.; Neese, F.; Bergmann, U.; DeBeer, S. X-ray emission spectroscopy evidences a central carbon in the nitrogenase ironmolybdenum cofactor. Science 2011, 334, 974−977. (11) Spatzal, T.; Aksoyoglu, M.; Zhang, L.; Andrade, S. L.; Schleicher, E.; Weber, S.; Rees, D. C.; Einsle, O. Evidence for interstitial carbon in nitrogenase FeMo cofactor. Science 2011, 334, 940. (12) Schindelin, H.; Kisker, C.; Schlessman, J. L.; Howard, J. B.; Rees, D. C. 1997. Structure of ADP x AIF4–stabilized nitrogenase complex and its implications for signal transduction. Nature 1997, 387, 370− 376. (13) Roth, L. E.; Nguyen, J. C.; Tezcan, F. A. ATP- and iron-proteinindependent activation of nitrogenase catalysis by light. J. Am. Chem. Soc. 2010, 132, 13672−13674. (14) Danyal, K.; Inglet, B. S.; Vincent, K. A.; Barney, B. M.; Hoffman, B. M.; Armstrong, F. A.; Dean, D. R.; Seefeldt, L. C. Uncoupling nitrogenase: catalytic reduction of hydrazine to ammonia by a MoFe protein in the absence of Fe protein-ATP. J. Am. Chem. Soc. 2010, 132, 13197−131979. (15) Ribbe, M. W.; Hu, Y.; Hodgson, K. O.; Hedman, B. Biosynthesis of nitrogenase metalloclusters. Chem. Rev. 2014, 114, 4063−4080. (16) Hu, Y.; Ribbe, M. W. Biosynthesis of the Metalloclusters of Nitrogenases. Annu. Rev. Biochem. 2016, 85, 455−483. (17) Schmid, B.; Ribbe, M. W.; Einsle, O.; Yoshida, M.; Thomas, L. M.; Dean, D. R.; Rees, D. C.; Burgess, B. K. Structure of a cofactordeficient nitrogenase MoFe protein. Science 2002, 296, 352−356. (18) Hu, Y.; Ribbe, M. W. Maturation of nitrogenase cofactor  the role of a class E radical SAM methyltransferase NifB. Curr. Opin. Chem. Biol. 2016, 31, 188−194. (19) Wiig, J. A.; Hu, Y.; Ribbe, M. W. NifEN-B complex of Azotobacter vinelandii is fully functional in nitrogenase FeMo cofactor assembly. Proc. Natl. Acad. Sci. U. S. A. 2011, 108, 8623−8627. (20) Fay, A. W.; Wiig, J. A.; Lee, C. C.; Hu, Y. Identification and characterization of functional homologs of nitrogenase cofactor biosynthesis protein NifB from methanogens. Proc. Natl. Acad. Sci. U. S. A. 2015, 112, 14829−14833. (21) Kaiser, J. T.; Hu, Y.; Wiig, J. A.; Rees, D. C.; Ribbe, M. W. Structure of precursor-bound NifEN: a nitrogenase FeMo cofactor maturase/insertase. Science 2011, 331, 91−94. (22) Fay, A. W.; Blank, M. A.; Lee, C. C.; Hu, Y.; Hodgson, K. O.; Hedman, B.; Ribbe, M. W. Spectroscopic characterization of the isolated iron-molybdenum cofactor (FeMoco) precursor from the protein NifEN. Angew. Chem., Int. Ed. 2011, 50, 7787−7790. (23) A NifB-co comprising six to eight Fe atoms has been reported and proposed to be an intermediate state of cluster assembly on NifB,24 although our data clearly show direct formation of an 8Fe Lcluster from two 4Fe modules on NifB.15,16 (24) Wilcoxen, J.; Arragain, S.; Scandurra, A. A.; Jimenez-Vicente, E.; Echavarri-Erasun, C.; Pollmann, S.; Britt, R. D.; Rubio, L. M. Electron paramagnetic resonance characterization of three iron-sulfur clusters present in the nitrogenase cofactor maturase NifB from Methanocaldococcus infernus. J. Am. Chem. Soc. 2016, 138, 7468−7471. (25) Wiig, J. A.; Hu, Y.; Lee, C. C.; Ribbe, M. W. Radical SAMdependent carbon insertion into the nitrogenase M-cluster. Science 2012, 337, 1672−1675. (26) Wiig, J. A.; Hu, Y.; Ribbe, M. W. Refining the pathway of carbide insertion into the nitrogenase M-cluster. Nat. Commun. 2015, 6, 8034. (27) Bauerle, M. R.; Schwalm, E. L.; Booker, S. J. Mechanistic diversity of radical S-adenosylmethionine (SAM)-dependent methylation. J. Biol. Chem. 2015, 290, 3995−4002. (28) Fugate, C. J.; Jarrett, J. T. Biotin synthase: Insights into radicalmediated carbon−sulfur bond formation. Biochim. Biophys. Acta, Proteins Proteomics 2012, 1824, 1213−1222.

formation on NifB, how does NifH process and deliver Mo and homocitrate to NifEN during M-cluster maturation, and what is the fate of the L-cluster Fe atom after M-cluster formation? We aim to supply these answers in the coming years in hopes of finalizing the elemental inventory of nitrogenase cofactor assembly and furthering our understanding of the structure− function relationship of the unique nitrogenase enzyme.



AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. ORCID

Markus W. Ribbe: 0000-0002-7366-1526 Yilin Hu: 0000-0002-9088-2865 Author Contributions

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. Funding

This work was supported by NIH Grant R01 GM67626 (M.W.R. and Y.H.). Notes

The authors declare no competing financial interest. Biographies Nathaniel S. Sickerman received B.A. and B.S. degrees from the University of Florida and a Ph.D. from the University of California, Irvine, where he is currently a postdoctoral fellow. Markus W. Ribbe received his B.S., M.S., and Ph.D. degrees from the University of Bayreuth, Germany. He was a postdoctoral fellow and is now Chancellor’s Professor of Molecular Biology and Biochemistry at Univsersity of California, Irvine. Yilin Hu received her B.S. degree from FuDan University, China, and a Ph.D. from Loma Linda University. She was a postdoctoral fellow and is now Assistant Professor of Molecular Biology and Biochemistry at University of California, Irvine.

■ ■

ABBREVIATIONS NifDK, nitrogenase MoFe protein; NifH, nitrogenase Fe protein REFERENCES

(1) Burgess, B. K.; Lowe, D. J. Mechanism of molybdenum nitrogenase. Chem. Rev. 1996, 96, 2983−3011. (2) Howard, J. B.; Rees, D. C. Structural basis of biological nitrogen fixation. Chem. Rev. 1996, 96, 2965−2982. (3) Eady, R. R. Structure-function relationships of alternative nitrogenases. Chem. Rev. 1996, 96, 3013−3030. (4) Lee, C. C.; Hu, Y.; Ribbe, M. W. Vanadium nitrogenase reduces CO. Science 2010, 329, 642−642. (5) Hu, Y.; Lee, C. C.; Ribbe, M. W. Extending the carbon chain: hydrocarbon formation catalyzed by vanadium/molybdenum nitrogenases. Science 2011, 333, 753−755. (6) Rebelein, J. G.; Hu, Y.; Ribbe, M. W. Widening the product profile of carbon dioxide reduction by vanadium nitrogenase. ChemBioChem 2015, 16, 1993−1996. (7) Schlögl, R. Catalytic synthesis of ammonia-a ″never-ending story″? Angew. Chem., Int. Ed. 2003, 42, 2004−2008. (8) Rofer-DePoorter, C. K. A comprehensive mechanism for the Fischer−Tropsch synthesis. Chem. Rev. 1981, 81, 447−474. 2840

DOI: 10.1021/acs.accounts.7b00417 Acc. Chem. Res. 2017, 50, 2834−2841

Article

Accounts of Chemical Research (29) Zheng, L.; White, R. H.; Cash, V. L.; Dean, D. R. Mechanism for the desulfurization of L-cysteine catalyzed by the nifS gene product. Biochemistry 1994, 33, 4714−4720. (30) Mueller, E. G. Trafficking in persulfides: delivering sulfur in biosynthetic pathways. Nat. Chem. Biol. 2006, 2, 185−194. (31) Lister, M. W.; Garvie, R. C. Sodium dithionite, decomposition in aqueous solution and in the solid state. Can. J. Chem. 1959, 37, 1567−1574. (32) Hu, Y.; Fay, A. W.; Ribbe, M. W. Identification of a nitrogenase FeMo cofactor precursor on NifEN complex. Proc. Natl. Acad. Sci. U. S. A. 2005, 102, 3236−3241. (33) Corbett, M. C.; Hu, Y.; Fay, A. W.; Ribbe, M. W.; Hedman, B.; Hodgson, K. O. 2006. Structural insights into a protein-bound ironmolybdenum cofactor precursor. Proc. Natl. Acad. Sci. U. S. A. 2006, 103, 1238−1243. (34) Lancaster, K. M.; Hu, Y.; Bergmann, U.; Ribbe, M. W.; Debeer, S. X-ray spectroscopic observation of an interstitial carbide in NifENbound FeMoco precursor. J. Am. Chem. Soc. 2013, 135, 610−612. (35) Fay, A. W.; Blank, M. A.; Rebelein, J. G.; Lee, C. C.; Ribbe, M. W.; Hedman, B.; Hodgson, K. O.; Hu, Y. Assembly scaffold NifEN: A structural and functional homolog of the nitrogenase catalytic component. Proc. Natl. Acad. Sci. U. S. A. 2016, 113, 9504−9508. (36) Yoshizawa, J. M.; Blank, M. A.; Fay, A. W.; Lee, C. C.; Wiig, J. A.; Hu, Y.; Hodgson, K. O.; Hedman, B.; Ribbe, M. W. Optimization of FeMoco maturation on NifEN. J. Am. Chem. Soc. 2009, 131, 9321− 9325. (37) Corbett, M. C.; Hu, Y.; Fay, A. W.; Tsuruta, H.; Ribbe, M. W.; Hodgson, K. O.; Hedman, B. Conformational differences between Azotobacter vinelandii nitrogenase MoFe proteins as studied by smallangle X-ray scattering. Biochemistry 2007, 46, 8066−8074. (38) Hu, Y.; Corbett, M. C.; Fay, A. W.; Webber, J. A.; Hodgson, K. O.; Hedman, B.; Ribbe, M. W. Nitrogenase Fe protein: a molybdate/ homocitrate insertase. Proc. Natl. Acad. Sci. U. S. A. 2006, 103, 17125− 17130. (39) Georgiadis, M. M.; Komiya, H.; Chakrabarti, P.; Woo, D.; Kornuc, J. J.; Rees, D. C. Crystallographic structure of the nitrogenase iron protein from Azotobacter vinelandii. Science 1992, 257, 1653− 1659. (40) Schwarz, G.; Mendel, R. R.; Ribbe, M. W. Molybdenum cofactors, enzymes and pathways. Nature 2009, 460, 839−847. (41) Jeoung, J. H.; Giese, T.; Grünwald, M.; Dobbek, H. CooC1 from Carboxydothermus hydrogenoformans is a nickel-binding ATPase. Biochemistry 2009, 48, 11505−11513.

2841

DOI: 10.1021/acs.accounts.7b00417 Acc. Chem. Res. 2017, 50, 2834−2841