The MoFe Proteins

The MoFe proteins are all a2 2 tetramers of 220-240 kDa, the a and (3 subunits being encoded by the nifD and K genes, respectively. The proteins can be described as dimers of a(3 dimers. They contain two unique metallosulfur clusters the MoFeTSg homocitrate, FeMo-cofactors (FeMoco), and the FesSy, P clusters. Neither of these two types of cluster has been observed elsewhere in biology, nor have they been synthesized chemically. Each molecule of fully active MoFe protein contains two of each type of cluster 2-7). 

The X-ray crystallographic structure of the MoFe protein from Azo-tobacter vinelandii (Avl) was reported by Kim and Rees in 1992 (2). This 2.7-A structure of Avl was at too low a resolution to produce definitive structures for the clusters in the MoFe protein. Data are now available on Avl(2-4,6), Cpl(5,31) and Kpl(7). 

FeMoco can be extracted from the MoFe protein into A(-methylfor-mamide (NMF) solution 32) and has been analyzed extensively using a wide range of spectroscopic techniques both bound to the protein and in solution after extraction from it (33). The extracted FeMoco can be combined with the MoFe protein polypeptides, isolated from strains unable to synthesize the cofactor, to generate active protein. The structure of the FeMoco is now agreed 4, 5, 7) as MoFeTSg homocitrate as in Fig. 4. FeMoco is bound to the a subunit through residues Cys 275, to the terminal tetrahedral iron atom, and His 442 to the molybdenum atom (residue numbers refer to A. vinelandii). A number of other residues in its environment are hydrogen bonded to FeMoco and are essential to its activity (see Section V,E,2). The metal 

Homocitrate is bound to the molybdenum atom by its 2-carboxy and 2-hydroxy groups and projects down from the molybdenum atom of the cofactor toward the P clusters. This end of FeMoco is surrounded by several water molecules (5, 7), which has led to the suggestion that homocitrate might be involved in proton donation to the active site for substrate reduction. In contrast, the cysteine-ligated end of FeMoco is virtually anhydrous. 

The MoFe proteins exhibit complex redox properties. Each tetra-meric a2/32 molecule of MoFe protein contains two P clusters and two FeMoco centers and, as normally isolated in the presence of sodium dithionite, the FeMoco centers are EPR-active, exhibiting an S = spin state with g values near 4.3 and 3.7 and 2.01 (Fig. 6). The P clusters are EPR silent and there is a wealth of evidence (39) using a variety of techniques that indicates that the iron atoms in these clusters are all reduced to the Fe state. 

---

Fig. 4. Requirements, substrates, and products of Mo-nitrogenase catalysis, where I is the MoFe protein II the Fe protein and Pi is inorganic phosphate. The generating system is composed of creatine phosphate and creatine phosphokinase to recycle the inhibitory MgADP produced during catalysis to...

Structure of the MoFe Protein. Extensive spectroscopic studies of the MoEe proteia, the appHcation of cluster extmsion techniques (84,151), x-ray anomalous scattering, and x-ray diffraction (10,135—137,152) have shown that the MoEe proteia contains two types of prosthetic groups, ie, protein-bound metal clusters, each of which contains about 50% of the Ee and content. Sixteen of the 30 Ee atoms and 14—16 of the 32—34... 

The VFe protein also has the equivalent of P-cluster pairs which have similar properties to those found in the MoFe protein (159). No information is available on whether P-cluster pairs exist in the FeFe protein, but because of the relatively high sequence identity and the similar genetic basis of its biosynthesis, the occurrence seems highly likely. The catalytic role assigned to the P-cluster pair involves accepting electrons from the Fe protein for storage and future deUvery to the substrate via the FeMo-cofactor centers. As of this writing (ca early 1995), this role has yet to be proved. 

The nitrogenase proteins are generally characterized by two letters indicating the species and strains of bacteria and the numerals 1 for the MoFe protein and 2 for the Fe protein. Thus, the Fe protein from Azotobacter vinelandii is Av2 and the MoFe protein from Klebsiella pneumoniae is Kpl. 

As well as donating electrons to the MoFe protein, the Fe protein has at least two and possibly three other functions (see Section IV,C) It is involved in the biosynthesis of the iron molybdenum cofactor, FeMoco it is required for insertion of the FeMoco into the MoFe protein polypeptides and it has been implicated in the regulation of the biosynthesis of the alternative nitrogenases. 

Figure 3 shows the three-dimensional structure of the MoFe protein from Klebsiella pneumoniae, Kpl, obtained at 1.65-A resolution (7). The overall structure of the polypeptides is frilly consistent with that reported earlier for Avl (3). The a and /8 subunits exhibit similar polypeptide folds with three domains of parallel /3 sheet/a helical type. At the interface between the three domains in the a subunit is a wide shallow cleft with the FeMoco at the bottom of the cleft about 10 A from the solvent. FeMoco is enclosed within the a subunit. The P cluster, however, is buried within the protein at the interface between the a and /8 subunits, being bound by cysteine residues from each subunit. A pseudo-twofold rotation axis passes between the two halves of the P cluster and relates the a and (3 subunits. Each af3 pair of subunits contains one FeMoco and one P cluster and thus appears... 

Fig. 3. The tetrameric structure of the MoFe protein (Kpl) from Klebsiella pneumoniae (7). The two FeMoco clusters and the P clusters are depicted by space-filling models and the polypeptides by ribbons diagrams (MOLSCRIPT (196)). The FeMoco clusters are bound only to the a subunits, whereas the P clusters span the interface of the a and j8 subunits.

During oxidation of the MoFe protein the P clusters are the first to be oxidized at about -340 mV. This redox potential was first measured (40) using Mossbauer spectroscopy and exhibited a Nemst curve consistent with a two-electron oxidation process. It is possibly low enough for this redox process to be involved in enzyme turnover (see Section V). No additional EPR signal was observed from this oxidized form at this time. However, later a weak signal near g = 12 was detected and was finally confirmed, using parallel mode EPR... 

C. Substrate Interactions with the MoFe Protein 1. Interactions with Reducible Substrates... 

The significance of the observed interactions between MoFe proteins and nucleotides is not easy to determine. However, as noted in Section IV, ATP is involved in the maturation of the MoFe protein and the biosynthesis of FeMoco. It is therefore possible that the nucleotide interactions noted here are associated with the maturation of the MoFe protein and not with its catalytic activity. 

The biosynthesis of the MoFe protein is extremely complex. Here we will first describe the biosynthesis of FeMoco, then that of the apo-MoFe protein, encoded by the nifD and nifK genes and containing the P clusters, and finally we will summarize what is known about the combination of FeMoco with the apo-MoFe protein to form active MoFe protein. 

It is clear from these data that homocitrate is intimately involved in the mechanism of substrate reduction and that close homologs such as citrate cannot entirely fill this role. Rationalization of this phenomenon is difficult, but comparisons of the reactivity of extracted FeMoco from the MoFe protein from wild type and NifV strains have led to an intellectually satisfying explanation (see Section V,E,2). 

The term apo-MoFe protein will be used here to denote the MoFe protein lacking FeMoco, although this protein is not devoid of prosthetic groups since it has bound P clusters. 

The purified preparations of the apo-MoFe proteins from both organisms included a small additional polypeptide of around 20 kDa. This was shown to be the nifY product in K. pneumoniae (94) and a non-nif protein denoted y in A. vinelandii (95). These proteins are apparently essential for effective reaction with FeMoco and are associated with the MoFe protein polypeptide through its interaction with the Fe protein and MgATP (96, 97) (see Section IV,C,3). These observations demonstrate a third role for the Fe protein in generating a form of apo-MoFe protein that is capable of accepting FeMoco. 

A comprehensive description of the mechanism of molybdenum nitrogenase has been provided by the Lowe-Thorneley scheme 102) (Figs. 8 and 9). In this scheme the Fe protein (with MgATP) functions as a single electron donor to the MoFe protein in the Fe protein cycle (Fig. 8), which is broken down into four discrete steps, each of which may be a composite of several reactions ... 

The reduced Fe protein MgATP complex forms a complex with the MoFe protein... 

An electron is transferred from the Fe protein to the MoFe protein with concomitant hydrolysis of MgATP to MgADP and Pi... 

Fig. 8. The Fe protein cycle of molybdenum nitrogenase. This cycle describes the transfer of one electron from the Fe protein (F) to one afi half of the MoFe protein (M) with the accompEmying hydrolysis of 2MgATP to 2MgADP + 2Pf. The rate-determining step is the dissociation of F (MgADP)2 from M,rf. Subscript red = reduced and ox = oxidized.

Fig. 9. The MoFe protein cycle of molybdenum nitrogenase. This cycle depicts a plausible sequence of events in the reduction of N2 to 2NH3 + H2. The scheme is based on well-characterized model chemistry (15, 105) and on the pre-steady-state kinetics of product formation by nitrogenase (102). The enzymic process has not been chsiracter-ized beyond M5 because the chemicals used to quench the reactions hydrolyze metal nitrides. As in Fig. 8, M represents an aji half of the MoFe protein. Subscripts 0-7 indicate the number of electrons trsmsferred to M from the Fe protein via the cycle of Fig. 8.

---