Chairs Comments Overview Of The Session

B.K. Burgess

Department of Molecular Biology & Biochemistry, University of California, Irvine, Irvine, CA 92697-3900, USA

This session brought together chemists and biochemists working on nitrogenase and related systems. Four of the talks focused on the conventional molybdenum nitrogenase which is composed of two separately purified proteins. The Fe protein (NifH) is a 60,000 molecular weight dimer of identical subunits connected through a single [4Fe-4S] cluster. It serves as an electron donor for the MoFe protein in a reaction somehow coupled to MgATP hydrolysis. The MoFe protein is often viewed as being composed of two identical halves that do not communicate with each other. Each half has one subunit (NifD) and one subunit (NifK), one [MoFe7St) homocitrate] cluster designated FeMo-co, one [8Fe-7S] cluster designated the P-cluster, and one binding site for the Fe protein. The P-clusters appear to be involved in electron transfer from the Fe protein to FeMo-co which serves as the site of substrate binding and reduction. Although the structures of both proteins and their metal centers have been available for several years the chemical synthesis of P-cluster and FeMo-co analogs has not yet been accomplished and the specific binding site for substrates on FeMo-co has not been determined.

The first talk (R. Holm) focused on the chemical synthesis problem. The current emphasis is on viewing both the P-clusters and FeMo-co as being composed of two discrete fragments that can be individually synthesized followed by attempts to covalently bridge the fragments. Another recent strategy is to use designed peptides as scaffolds for the construction of the covalent bridges. These strategies have been successful in the synthesis of a carbon monoxide dehydrogenase model that is spectroscopically identical to the active site cluster. In the case of the P-clusters a topological analog has been successfully constructed that includes the hexa-coordinate sulfide and has the composition [Mo2Fe6Sv]. The third talk (P. Ludden) approached the synthesis problem from the biological side where progress continues to be made on defining what gene products are required for FeMo-co biosynthesis and on determining the sequence of events. It has been known for many years that the NifH gene product was somehow required for the initial biosynthesis of FeMo-co but this was the first report that an Mo and Fe containing FeMo-co precursor actually accumulates on NifH in an MgATP dependent process. This was also the first report that under certain conditions Mo, as well as Fe, accumulates on NifNE. For many years a protein called gamma was believed to play a critical role in FeMo-co biosynthesis but the gene encoding the protein had not been cloned, precluding detailed characterization of its role in the process. That problem has now been solved and the gene has been isolated and designated nafY. The NafY protein is homologous to NifY, it is a dimer both in the presence and absence of FeMo-co and its function is currently proposed to be stabilization of the FeMo-co-deficient protein in a state that can accept FeMo-co.

In R. Thorneley's talk, electron transfer from the [4Fe-4S]1+ cluster of the Fe protein to the MoFe protein was examined with emphasis on, different conformational states, the role that MgATP plays in the process and on the MgATP binding site. Three different methods were employed including studies with deoxy ATP where it was shown that removing the H-bond from ribose 2'-OH did not affect the cluster but did influence the rate of electron transfer. In a second series of experiments Asp43, which is H-bonded to H2O coordinated to Mg in the complex structure was modified to Glu and Asn, yielding Fe proteins that no longer underwent the normal MgATP induced conformational change and had very little activity. A partially active [4Fe-4Se] form of the Fe protein was also constructed and characterized. The same talk considered the exciting progress that is being made in using stopped-flow FTIR to directly look for intermediates in multi-electron substrate reduction with current efforts focusing on azide reduction. Perhaps the biggest breakthrough in the session came in the talk of D. Dean who used novel approaches to determine where substrates bind on FeMo-co. In earlier studies the authors had used a genetic approach to identify MoFe protein residue Gly69 as being critical for the interaction of acetylene with the FeMo-co site. In this study, alterations of Gly69, Val70, and Arg96 were characterized, with the latter two residues capping a specific [4Fe-4S] face of FeMo-co, remote from the Mo site. Introducing smaller residues at position 70 allowed efficient reduction of larger substrates than can be efficiently reduced by the native enzyme and led to the elimination of proton reduction during acetylene reduction. These experiments provide compelling evidence in support of the proposed model that all substrates bind to a specific 4Fe-4S face of FeMo-co.

Moving away from conventional nitrogenase, Bauer's presentation considered the fact that nifHDK homologs encode proteins with completely different functions. In light-independent chlorophyll/bacteriochlorophyll biosynthesis the double bond of ring D of protochlorophyllide is reduced stereospecifically by an enzyme (DPOR) consisting of three subunits. These three subunits BchL, BchN and BchB are similar to NifH, NifD and NifK respectively. Now for the first time the enzyme has been purified and detailed biochemical characterization has begun. The DPOR catalyzed reaction has the same requirements as the nitrogenase reaction for all three subunits, dithionite and MgATP, however preliminary characterization of the purified BchNB complex and sequence comparisons reveal that it does not have the conserved Cys and His residues required for FeMo-co ligation nor does it appear to contain Mo or V. It appears to resemble NifNE (which has a [4Fe-4S] cluster in place of the P-clusters) more than it does the MoFe protein. Finally, D. Gadkari's talk continued to emphasize biological diversity by considering that one organism, Streptomyces thermoautotrophicus fixes atmospheric dinitrogen using an enzyme with no sequence similarity to conventional nitrogenase and with very different biochemical properties. In spite of these differences, new information presented showed that, like conventional nitrogenase the S. thermoautotrophicus system was capable of reduction of both azide and cyanide.

LOCALIZATION OF THE NITROGENASE SUBSTRATE BINDING SITE

S.M. Mayer1, J. Christiansen1, P.C. Dos Santos1, W.G. Niehaus1, P. Benton2,

'Department of Biochemistry, Virginia Tech, Blacksburg, Virginia, 24061, USA

department of Chemistry & Biochemistry, Utah State University, Logan, Utah, 84322, USA

1. Introduction

Nitrogenase is the two-component metalloenzyme that catalyzes the MgATP-dependent reduction of dinitrogen. For the Mo-dependent nitrogenase, the homodimeric Fe protein (component II) delivers electrons to the C/2P2 MoFe protein (component I) within which is located two substrate reduction sites (see Christiansen et al. 2001; Rees, Howard 2000; and Burgess, Lowe 1996 for recent reviews on the catalytic features of nitrogenase). The path of electron transfer during nitrogenase catalysis is believed to proceed from a [4Fe-4S] cluster contained within the Fe protein to an [8Fe-7S] P cluster bridged between an individual ap-subunit pair of the MoFe protein and finally to a substrate reduction site located within a MoFe protein a-subunit. One nucleotide-binding site is located within each Fe protein subunit and each of these is separated from an individual substrate reduction site by more than 30 angstroms.

Substrate binding and reduction takes place at a complex metallocluster called FeMo-cofactor. The framework of the FeMo-cofactor is constructed from 4Fe-3S and 3Fe-3S-Mo sub-fragments bridged by three sulfides. An organic constituent, homocitrate, is also attached to the Mo atom through its 2-hydroxy and 2-carboxyl groups. It has not yet been possible to identify the metal-sulfur surface of FeMo-cofactor that provides the substrate interaction site by simple inspection of the FeMo-cofactor structure, nor by kinetic and spectroscopic studies. Nevertheless, a variety of different substrate reduction sites have been proposed. For example, FeMo-cofactor contains three geometrically identical 4Fe-4S faces, and any of these faces is a reasonable candidate for providing a substrate interaction site because each of them contains four coordinately unsaturated Fe atoms. The Mo atom has also been proposed as a site to which substrates might bind during the catalytic cycle.

2. Isolation of an Acetylene-resistant Nitrogenase

In addition to the natural substrate, dinitrogen, nitrogenase is able to reduce a variety of other substrates, including acetylene. The Km for acetylene reduction (-0.005 atm) is much lower than the

Km for dinitrogen reduction (-0.10 atm). Thus, under conditions of low electron flux, acetylene is a very effective physiological inhibitor of dinitrogen reduction. As an approach to identify the nitrogenase substrate interaction site, we developed a genetic selection for the isolation of mutant strains that are able to effectively reduce dinitrogen, but are significantly impaired in their ability to reduce acetylene. Mutants resistant to acetylene were obtained by plating approximately 108 cells of a strain defective in electron flux on a minimal medium containing no fixed nitrogen source and incubating these cells under an ambient atmosphere that also contains 0.025 atm of acetylene. Under these conditions, no immediate growth was observed. However, after about two weeks, a small number of colonies appeared. These colonies were picked and streaked again under a 0.025 acetylene atm to establish that each strain was indeed stably resistant to acetylene. Examination of the growth of these strains without acetylene, showed that they are not impaired in their ability to fix dinitrogen. Subsequent DNA sequence analysis for several independently isolated acetylene-resistant strains revealed that each has the MoFe protein a-subunit 69° y residue codon substituted by an a-69Ser codon. Site-directed mutagenesis and gene replacement experiments were also used to independently confirm that substitution of a-69Gly by a-69Ser results in the acetylene-resistant phenotype.

An altered nitrogenase a-69Ser MoFe protein was isolated and shown to have the following properties: (1) It is unaffected in its S=3/2 resting state EPR spectrum. (2) It exhibits a 20-fold increase in Km for acetylene reduction but is not affected in its VmaK for acetylene reduction. (3) It is not significantly impaired in either Km or Vnmx for dinitrogen reduction. (4) Acetylene is converted from a non-competitive to a competitive inhibitor of dinitrogen reduction. (5) The inhibitor CO is converted from a non-competitive inhibitor to a competitive inhibitor of acetylene and dinitrogen reduction. Our interpretation of these results is that there are two functional acetylene-binding sites located within the MoFe protein. One of these is a high-affinity acetylene-binding site and is the one most affected by the a-69Ser substitution. The other is a low-affinity acetylene-binding site that is also the same site as the dinitrogen-binding site. An issue that emerges from this interpretation is the nature of the structural relationship between the high-affinity and low-affinity acetylene-binding sites. We suggest that both of these sites are actually the same with respect to their general location within the metal-sulfur surface of FeMo-cofactor but that they are manifested by different redox states of the MoFe protein. The low-affinity acetylene-binding site/state is also proposed to correspond to the dinitrogen-binding site/state.

Where is the actual substrate interaction site located? Inspection of the MoFe protein crystal structure shows that the a- 69aly residue is located on a short helix that extends from the P-cluster coordinating residue a- 62Cys to a-70Val. The a-70Val side chain approaches a specific 4Fe-4S face of FeMo-cofactor. Furthermore, the backbone carbonyl oxygen of the a- 69aly residue is located within hydrogen bonding distance to a guanido NH-group of a-96Arg, which also approaches the same 4Fe-4S face. The a- 96Arg residue is likewise connected to the P cluster through a short helix extending from the P cluster coordinating a-88Cys residue. Based on these considerations we suggest that redox-dependent structural rearrangements known to occur within the P cluster could result in the communication along the connecting helices and the subsequent movement of either or both a-70Val and a-96Arg during catalysis. Thus, our interpretation of the physiological and kinetic consequences of the a-69Ser substitution is that the redox-dependent movement of either or both a-70Val and a-96Arg is affected.

3. Reduction of Short-chain Alkynes by Altered Nitrogenases

If our model that the 4Fe-4S face of FeMo-cofactor capped by a-70Val and oe-96Arg provides the substrate interaction site is correct, then the side chains of either or both of these residues could have an effect on the ability of certain substrates to interact with the substrate binding site. To test this possibility we substituted the a- 70Val residue by a-70Ala and by a-70Gly. The effect of these substitutions on the ability of the altered MoFe proteins to catalyze the reduction of short-chain alkynes that are not effectively reduced by the wild type enzyme was then examined. Our initial approach towards this end was to determine if the addition of propargyl alcohol (C3H4O) to the growth medium affects the diazotrophic growth of the strain having the a-70Ala substitution. We found that this strain grows well under normal diazotrophic conditions but that addition of 6 mM propargyl alcohol to the growth medium completely inhibits diazotrophic growth. Propargyl alcohol had no effect on diazotrophic growth of the wild type strain and did not affect growth of the a-70Ala-substituted strain if a source of fixed nitrogen was also added to the growth medium. Kinetic characterization of the a- 70Ala MoFe protein revealed that propargyl alcohol is a competitive inhibitor of acetylene reduction having a Kt of approximately 2 mM.

Because of the difficulty in measuring the product of the reduction of short-chain alkyne alcohols, we turned to an examination of the ability of the a- 70Aia- and a-70Gly-substituted MoFe proteins to reduce propyne and butyne. The reduction products of both of these substrates can be easily and accurately quantified by gas chromatography. For the wild type enzyme, propyne reduction is extremely poor and butyne reduction, if it occurs, is below levels that we are able to accurately detect. In contrast, both the a-70Ala- and a-70Gly-substituted MoFe proteins are able to reduce both propyne and butyne, but at different levels depending on the respective substituted residue. For example, both the a-70AIa- and a-70Gly-substituted MoFe proteins are able to reduce both propyne and butyne at readily detectable levels but reduction of these alkynes to yield the corresponding alkene is more effective for the a-70Gly MoFe protein than for the a- 70Ala MoFe protein. Also, for both substituted MoFe proteins, propyne reduction is much more effective than butyne reduction. These results establish a direct correlation between a decrease in side-chain length at the a-70 residue position with the capability of the MoFe protein to reduce short-chain alkynes of increasing length. Thus, reduction of short-chain alkynes does occur at the 4Fe-4S face of FeMo-cofactor we have targeted.

Another interesting feature of the a- 70Ala- and a-70Gly-substituted MoFe proteins is that acetylene is able to effectively suppress proton reduction catalyzed by these altered MoFe proteins (>98% suppression by 0.10 atm acetylene). This level of H2 suppression has previously been predicted on the basis of the extrapolation of kinetic data that suggest acetylene should be able to completely suppress proton reduction. However, it has not been possible to prove this hypothesis experimentally using wild type nitrogenase because acetylene concentrations greater that -0.20 atm have an adverse effect on electron flux. Our interpretation of the result that 0.1 atm acetylene virtually eliminates proton reduction catalyzed by the a-70AIa- and a-70Gly-substituted MoFe proteins is that proton reduction and acetylene reduction occur at the same 4Fe-4S face. In other words, constraints on acetylene binding in these altered MoFe proteins is sufficiently relaxed so that proton reduction is nearly eliminated when >0.10 atm acetylene is present.

4. The a-96Arg Side-chain Moves to Accommodate Substrate Interaction

A caveat to our suggestion for the substrate interaction site, as well as with every other model that has been proposed, is that there is not sufficient space to accommodate substrate interaction unless certain amino acid side-chains move away from the FeMo-cofactor during catalysis. We have already suggested that the side-chains of either or both a-70Val and a-96Arg move during catalysis. Once again, inspection of the structural model provides clues about how such movement might occur. Close examination of the structural model reveals that there is little room to permit movement of a-70Val unless there are major structural rearrangements within the MoFe protein during catalysis, a possibility we consider unlikely. However, there is sufficient room to accommodate movement of the a-96Arg side-chain without the requirement of other severe structural perturbations. Based on this observation, we are considering a model where the a- 96Arg side-chain becomes repositioned during catalysis when compared to its location in the resting state of the MoFe protein. An attractive possibility is that the a- 96Arg side-chain moves back and forth during turnover - first to expose a substrate binding site and then, perhaps, to trap bound substrate at an iron-sulfur surface of FeMo-cofactor. Such movement of the a-96Arg side-chain could also serve as a shuttle to facilitate proton transfer. The proposed dynamic movement of the a-96Arg side-chain during catalysis is also in line with the observation that this side-chain occupies three slightly different conformations in each of the crystallographically determined structures for MoFe protein from Azotobacter vinelandii, Clostridium pasteurianum, and Klebsiella pneumoniae.

It does not appear that any substrate or inhibitor can interact with FeMo-cofactor in the as-isolated resting state of the MoFe protein. The basis for this conclusion is that the S=3/2 EPR signal observed for the as-isolated MoFe protein, which originates from FeMo-cofactor, is not perturbed by the addition of any substrate or inhibitor. The usual explanation offered for this situation is that electrons must first be delivered to the active site in order to prime it for substrate interaction and reduction. Although this is almost certainly the case for many substrates and inhibitors, there is no prima faciae chemical reason why all substrates and inhibitors should be unable to interact with

FeMo-cofactor contained in the MoFe protein in its resting state, unless these molecules are denied access to the FeMo-cofactor. Indeed, it has been shown that addition of cyanide to isolated FeMo-cofactor, which contains an S=3/2 EPR spectrum very similar to that of the protein-bound species, does result in the appearance of a new S=3/2 EPR signal.

As an approach to investigate whether or not the a-96Arg side-chain might move during catalysis, we placed a leucine substitution at this position and asked if the altered MoFe protein is now able to interact with substrates in the resting state under non-turnover conditions. The logic of the a-96Leu substitution was to shorten the side-chain at this position as a way to expose the substrate interaction site in the resting state. Incubation of acetylene with the a-96Leu MoFe protein in the as-isolated resting state decreased the intensity of the normal S=3/2 FeMo-cofactor signal with appearance of a new EPR signal having inflections at g=4.50 and 3.50. Similarly, incubation of cyanide with the a- 96Leu MoFe protein also decreased the normal FeMo-cofactor signal with concomitant appearance of a new EPR signal having an inflection at g=4.06. No such EPR changes were elicited from the wild type MoFe protein under the same conditions or for the a- 96Leu MoFe protein when incubated in the presence of dinitrogen. These results support a model where effective interaction of dinitrogen with FeMo-cofactor occurs as a consequence of both increased reactivity and accessibility of FeMo-cofactor under turnover conditions. For example, the reason that dinitrogen does not interact with the a-96Leu MoFe protein in the resting state, but acetylene does, is that a lower redox state is required for dinitrogen binding.

5. All Substrates are Reduced at the Same 4Fe-4S Face of FeMo-cofactor

We have now suggested that both the high-affinity and low-affinity acetylene-binding sites, as well as the proton reduction site(s), are located at the same 4Fe-4S face of FeMo-cofactor approached by the a- 70Val and a-96Arg residue side-chains. Where is the dinitrogen-binding site? If all substrates, including dinitrogen, interact with the same 4Fe-4S face of FeMo-cofactor, it should, in principle, be possible to isolate an altered MoFe protein having a single amino acid substitution which severely impairs the access of all substrates to the reduction site. However, it is experimentally difficult to distinguish effects on substrate reduction that arise from disruption of electron or proton delivery to the active site (flux effects) from alterations that specifically affect the accessibility of substrates to the active site. As a way to circumvent this problem we reasoned that, because protons are so small, it might be possible to isolate an altered MoFe protein that is not altered in proton reduction but is ineffective in the reduction of larger nitrogenase substrates. Namely, an altered MoFe protein that is not altered in its rate of proton reduction is, by definition, not affected in flux.

The next challenge involved selecting the position and choice of amino acid substitution. If there is only one major substrate entry pathway to the 4Fe-4S face we have targeted, then examination of the MoFe protein structure reveals that the most reasonable pathway for access to this site is located between the short arm of homocitrate and the a-69GIy residue. We considered this an attractive possibility because substitution of a-69Ser at this position has a dramatic affect on the ability of the MoFe protein to reduce acetylene. It therefore seemed reasonable to expect that substitution of a residue having an even bulkier side-chain at this position could affect the accessibility of all substrates, with the possible exception of protons. Although we have now placed a number of different amino acid substitutions at this position, the most instructive results obtained so far are with the a-69Val MoFe protein.

Kinetic characterization of the a-69Val MoFe reveals that it is able to reduce protons at the same rate as the wild type protein and, therefore, is not perturbed with respect to electron flux. However, the a-69Val MoFe protein reduces acetylene and dinitrogen at rates of only about 2% when compared to the wild type MoFe protein. In separate experiments, it was also found that the interaction of CO with the a- 69Val MoFe protein under turnover conditions is severely impaired when compared to the wild type protein. We, therefore, conclude that most, and perhaps all, nitrogenase substrates and inhibitors interact with the same 4Fe-4S face of FeMo-cofactor and that there is a single major access pathway to that face.

6. Conclusions

Based on the worked described here, we make the following conclusions. First, molybdenum does not provide the site for substrate binding. Second, the 4Fe-4S face of FeMo-cofactor capped by a-70Val and a-96Arg provides the substrate-binding site. Third, all substrates bind at the same 4Fe-4S face of FeMo-cofactor. Fourth, all substrates, with the possible exception of protons, gain access to the substrate-binding -site primarily through the same pathway. Fifth, the side-chain of a-96Alg moves during catalysis in order to make the binding site available to substrate. The a- 96Arg side-chain might also serve as a proton shuttle during catalysis. Sixth, although we propose that all substrates access the same metal-sulfur surface during catalysis, it is expected that different substrates will interact with this surface in different ways. Seventh, it also seems reasonable to anticipate that an individual substrate might interact with the metal-sulfur surface of FeMo-cofactor in different geometric orientations depending upon the particular redox-state of the enzyme.

We believe that isolation of an altered MoFe protein that interacts with certain substrates in the resting state now makes it possible to crystallographically determine exactly where and how these substrates initially interact with FeMo-cofactor. This approach, as well as the continued mapping of the active site environment by using amino acid substitutions and substrate analogs, will remain the emphasis of our studies aimed at determining the nitrogenase catalytic mechanism.

7. References

Burgess BK, Lowe DJ (1996) Chem. Rev. 96, 2983-3011

Christiansen JC et al. (2001) Ann. Rev. Plant Physiol. Plant Mol. Biol. 52, 269-295 Rees DC, Howard JB (2000) Curr. Opin. Chem. Biol. 4, 559-566

BIOSYNTHESIS OF THE IRON-MOLYBDENUM COFACTOR OF NITROGENASE: ROLES OF DINITROGENASE REDUCTASE AND NAFY

P.W. Ludden, P. Rangaraj, L. Rubio, C. Ruttimann-Johnson, D. Dyer, J. Cheng, V.K. Shah

Dept of Biochemistry, University of Wisconsin-Madison, Madison, WI 53706, USA

1. Introduction

The iron molybdenum cofactor (FeMo-co) constitutes the active site of the wz/-encoded, molybdenum nitrogenase and its analogs, FeV-co and FeFe-co comprise the actives sites of the vnf-encoded, vanadium nitrogenase and the ««/-encoded, iron only nitrogenase, respectively. While the structure for FeMo-co is known (Figure 1), FeV-co and FeFe-co are thought to have very similar structures based on their ability to substitute for FeMo-co in the «¿/DAT-encoded nitrogenase protein. Furthermore, all three systems share a requirement for functional nifB gene; NifB generates the Fe and S precursor to FeMo-co.

Figure 1. The structure of FeMo-co (adapted from Kim, Rees 1992).

An in vitro FeMo-co synthesis system has been developed and each of the components of that system is now available in a highly purified form. The system includes NifB-co (the metabolic product of NifB), NifNE, NifX, NifH, NafY (aka Gamma), homocitrate (the metabolic product of NifV), MgATP, reductant in the form of sodium dithionite, molybdate, and a FeMo-co-deficient form of dinitrogenase (apodinitrogenase in its OC2P2Y2 form). In this system, FeMo-co is synthesized and inserted into apodinitrogenase and is capable of substrate reduction. To date, it has not been possible to synthesize FeV-co in vitro starting with a source of V that has not been processed biologically. However, it is possible to convert a protein bound form of V into a cofactor that can be inserted into «(/-encoded apodinitrogenase and function in acetylene reduction (Ruttimann-Johnson etal. 2001).

The two main points of this chapter are: (i) that NifH (dinitrogenase reductase) plays a catalytic role in FeMo-co synthesis; and (ii) that there is a family of proteins (the X-proteins) whose members are able to bind FeMo-co/FeV-co and/or their precursors.

2. Dinitrogenase Reductase is a "Moonlighting Protein"

Moonlighting proteins are proteins that have more than one role in the cell (Jeffery 1999). The roles can be related, as they are in the case of dinitrogenase reductase, or seemingly unrelated, as in the case of phosphoglucose isomerase, which serves as both a glycolytic enzyme and an autocrine factor (Jeffery et al. 2000). Dinitrogenase reductase is a supermoonlighting protein in that it has at least three essential roles in nitrogen fixation (Figure 2).

®Substrate reduction (e- transfer to dinitrogenase)

NifH is a "rriqon lighting" protein

©Regulation by

ADP-ribosylation

(DApodinitrogenase Maturation

®FeMo-co biosynthesis

Figure 2. NifH is a "moon lighting" protein

First, dinitrogenase reductase is the unique and specific electron donor to dinitrogenase and was first identified for this role (Bulen, LeComte 1966). No other protein is known to be able to reduce dinitrogenase to a state able to reduce substrates. As the electron donor, it is the target for regulation of this process in some organisms. Dinitrogenase reductase from Rhodospirillum rubrum is regulated by reversible ADP-ribosylation at an arginine residue near the Fe4S4 cluster that bridges the two subunits (Pope et al. 1985).

Second, dinitrogenase reductase is a biosynthetic enzyme. Filler and Smith first noted that non-polar nifH mutants of Klebsiella pneumoniae accumulated FeMo-co-deficient dinitrogenase (Filler et al. 1986). Subsequent experiments showed that dinitrogenase reductase is required for in vitro synthesis and that 55Fe from NifB-co accumulated on dinitrogenase reductase during in vitro FeMo-co synthesis (Rangaraj et al. 2001). The accumulation of 55Fe from 55Fe-NifB-co required NifNE and MgATP. In this presentation, it is demonstrated that 99Mo can also accumulate on dinitrogenase reductase during FeMo-co synthesis and that this accumulation is NifB-co, NifNE and MgATP-dependent. VnfH can substitute for NifH and can accumulate 99Mo. The role of dinitrogenase reductase in FeMo-co synthesis is not to provide reductant, because a form of dinitrogenase reductase that is devoid of Fe4S4 clusters and the ability to transfer electrons to dinitrogenase is fully functional in FeMo-co synthesis (Rangaraj etal. 1997). The amount of "Mo accumulated on dinitrogenase reductase is significant, approximately 0.5 moles of Mo per mole of dimeric protein. It is unlikely that Mo is accumulating at the site observed by Georgiadis and Rees in the first crystal structure of dinitrogenase from Azotobacter vinelandii (Georgiadis et al. 1992) because that molecule of M0O42" occupied the binding site for the terminal phosphoryl group of ATP. We observe "Mo accumulation only in the presence of MgATP when that site would be expected to be occupied.

Dinitrogenase reductase is proposed to accept Fe and S from NifNE and add Mo to the complex. Neither homocitrate nor NifX are required for the accumulation of "Mo on dinitrogenase reductase, so dinitrogenase reductase is proposed to play its role before the addition of the organic acid ligand to Mo. While dinitrogenase reductase is proposed to serve as the site of entry of the heterometal of the FeMo-co, and by analogy, to FeV-co and FeFe-co, it does not appear to be the site of specificity for heterometal selection. For example, VnfH, the dinitrogenase reductase from the vanadium system, will substitute for NifH and accumulate 99Mo during in vitro FeMo-co synthesis (Figure 3).

Finally, dinitrogenase reductase plays an as yet undefined role in FeMo-co insertion that is distinct from its role in FeMo-co synthesis (Allen et al. 1993; Robinson et al. 1987). Dinitrogenase from cells that lack dinitrogenase reductase accumulate FeMo-co-deficient apodinitrogenase and are unable to insert FeMo-co. Once the FeMo-co-deficient apodinitrogenase has been treated with dinitrogenase reductase, it gains the ability to accept FeMo-co and become active. It also gains the ability to bind NafY (gamma protein); NafY is also a FeMo-co-binding protein (see below). A. vinelandii cells with lesions in FeMo-co-biosynthetic genes other than nifH accumulate a form of apodinitrogenase with NafY tightly bound. A. vinelandii nifH cells accumulate an NafY-free form of dinitrogenase reductase. Now that nafY mutants are available, it is known that NafY is not essential for FeMo-co insertion in vivo; apodinitrogenase from, for example, nifBnafY mutants can be activated by the addition of purified FeMo-co in the absence of NafY. Once again, Fe4S4-deficient dinitrogenase reductase is able to perform its role in dinitrogenase maturation required for FeMo-co insertion, thus that role does not require electron transfer to or from dinitrogenase reductase (Rangaraj et al. 1997).

It is interesting that the nifH gene product plays multiple roles in the overall process of nitrogen fixation and that it is such an amazingly conserved protein across nature. Perhaps the high degree of sequence identity for this protein across the many genera of bacteria and the arachea reflects that fact that it has several domains with different function, each of which must be tightly conserved.

3. NafY is a Member of a Family of FeX-co-Binding Proteins

FeMo-co-deficient dinitrogenase from A. vinelandii was discovered to have a third type of subunit (Paustian et al. 1990). This protein was termed the "gamma" subunit of apodinitrogenase and it was shown to bind FeMo-co (Homer et al. 1995). The gene for this protein has been obtained and designated nafY (Nitrogenase Accessory Factor). The deduced sequence for NafY shows that it belongs to a small group of proteins that are known to bind FeMo-co/FeV-co or their precursors. This family of proteins includes NafY, VnfX, NifX, NifY and VnfY (Ruttimann-Johnson et al. 2000, 2001). The nafY gene is immediately downstream of rnfGEH in A. vinelandii, but is not co-transcribed with these genes. NafY has been overexpressed and shown to exist as a dimer in both free and FeMo-co-bound forms. NafY binds a single molecule of FeMo-co per dimer with a Kd of 1 |j.M. The ability to fix nitrogen by nafY mutants is affected only at high temperatures and our current model is that NafY plays a role in stabilizing apodinitrogenase. nafYnifB mutants accumulate normal levels of apodinitrogenase, but the apodinitrogenase in extracts of these mutants can only be activated to 50% of the level achieved in a nifB mutant. The role of the FeMo-co binding site is unknown. NafY will also bind the FeMo-co precursor, NifB-co. Crystals of NafY have been obtained.

Was this article helpful?

0 0

Post a comment