Nitrogenase reactivity with P-cluster variants

Abstract

Nitrogenase is a multicomponent metalloenzyme that catalyzes the conversion of atmospheric dinitrogen to ammonia. For decades, it has been generally believed that the [8Fe-7S] P-cluster of nitrogenase component 1 is indispensable for nitrogenase activity. In this study, we identified two catalytically active P-cluster variants by activity assays, metal analysis, and EPR spectroscopic studies. Further, we showed that both P-cluster variants resemble [4Fe-4S]-like centers based on x-ray absorption spectroscopic experiments. We believe that our findings challenge the dogma that the standard P-cluster is the only cluster species capable of supporting substrate reduction at the FeMo cofactor and provide important insights into the general mechanism of nitrogenase catalysis and assembly.

Nitrogenase catalyzes one of the most remarkable chemical transformations in biological systems, the reduction of atmospheric dinitrogen to a bioavailable form, ammonia. Three classes of nitrogenase systems, namely, the Mo-, V-, and Fe-only nitrogenases, have been identified (for recent reviews see refs. 1–11).f Although encoded by different structural genes, all three nitrogenases are comprised of two essential component metalloproteins, component 1 (MoFe, VFe, or FeFe protein) and component 2 (Fe protein).g The homodimeric Fe protein of the Mo-nitrogenase has one [4Fe-4S] cluster bridged between the two subunits, whereas the α₂β₂-tetrameric MoFe protein contains two unique metal clusters per αβ-subunit: the [8Fe-7S] P-cluster (14), which is located at the αβ-interface, and the [Mo-7Fe-9S-X-homocitrate]h FeMo cofactor (FeMoco), which is situated within the α-subunit. ATP-dependent electron transfer is believed to proceed from the [4Fe-4S] cluster of the Fe protein to the P-cluster of the MoFe protein and finally to FeMoco where substrate reduction takes place. The Fe protein of the V-nitrogenase shares a high degree of homology with the Fe protein of the Mo-nitrogenase and, apart from the presence of an additional γ-subunit, the VFe protein is also believed to be highly homologous to the MoFe protein with respect not only to the structural genes encoding the α- and β-subunits, but also to the redox centers contained within the protein, designated the P-cluster and the FeV cofactor (FeVco), respectively, in this case (4, 10, 16). The structurally homologous heterometal centers, i.e., FeMoco and FeVco, are also categorically termed cofactors.

It is generally believed that the structural integrity of the P-cluster is indispensable for nitrogenase reactivity. However, the exact mechanism by which the P-cluster carries out its function in substrate reduction and, in particular, the oxidation states and structural conformations of the P-cluster involved in this process, remain largely a puzzle (17). Two types of cofactor-deficient MoFe or VFe protein variants, generated by nifH or nifB deletion,i respectively, have been used to study the features of P-clusters without the interference of the cofactor centers (17–25). Recent studies have shown that a MoFe protein variant generated by nifB deletion contains normal P-clusters (22), whereas a MoFe protein variant generated by nifH deletion has catalytically inactive P-cluster precursors comprised of [4Fe-4S]-like fragments (19). These results, apparently, further support the dogma in the field that a fully assembled P-cluster is essential for enzyme activity. Here, we report the observation of catalytic activity in two cofactor-deficient, P-cluster variant-containing species of component 1 upon FeMoco insertion. These two species are (i) ΔnifH Av1′, a MoFe protein variant isolated from a nifH-deletion Azotobacter vinelandii strain after preincubation with Fe protein and MgATP and (ii) ΔnifB Av1^V, a VFe protein variant isolated from a nifB-deletion A. vinelandii strain (for designation of strains and protein species related to this work see Table 1). We believe that our findings challenge the concept that the standard [8Fe-7S] P-cluster is the only species responsible for nitrogenase reactivity and provide insights into the general mechanisms of nitrogenase assembly and catalysis.

Table 1. Nomenclature of strains and proteins of A. vinelandii related to this work.

Strain	Strain designation	Source	Genotype	Designation of component 1*	Designation of component 2*
Mo-nitrogenase
WT		26	WT	Av1	Av2
DJ1143	AvDJ1143	21	nifB deletion, His-tagged MoFe protein	ΔnifB Av1
DJ1165	AvDJ1165	17	nifH deletion, His-tagged MoFe protein	ΔnifH Av1
DJ1165	AvDJ1165	This study	nifH deletion, His-tagged MoFe protein, preincubated with Av2 and MgATP	ΔnifHAv1′
V-nitrogenase
WT		27-29	WT	Av1v	Av2v
YM7A	AvYM7A	This study	nifB deletion, His-tagged VFe protein	ΔnifB Av1v

^*The component proteins of the Mo-nitrogenase are designated by the initials of the organism from which they are isolated and the Arabic numeral of the component. For example, the MoFe and Fe proteins of the Mo-nitrogenase of A. vinelandii are designated Av1 and Av2, respectively. This nomenclature is also applied to the V-nitrogenase, with the addition of a superscript V. Therefore, the VFe and Fe proteins of the V-nitrogenase of A. vinelandii are abbreviated as Av1^v and Av2^v, respectively.

Materials and Methods

Unless otherwise noted, all chemicals and reagents were obtained from Fisher Scientific, Baxter Scientific Products (McGaw Park, IL), or Sigma.

Construction of Variant A. vinelandii Strains. Construction of A. vinelandii nifB- and nifH-deletion strains DJ1143 (AvDJ1143) and DJ1165 (AvDJ1165), which produce His-tagged MoFe proteins designated ΔnifB Av1 and ΔnifH Av1, respectively, has been described (17, 21). A. vinelandii strain DJ1258 (AvDJ1258), which produces a His-tagged WT VFe protein, was a generous gift from Dennis Dean (Virginia Polytechnic Institute and State University, Blacksburg). A. vinelandii nifB-deletion strain YM7A (AvYM7A), which produces a His-tagged VFe protein designated ΔnifB Av1^V, was constructed by introducing a deletion/kanamycin resistance cassette insertion in the nifB gene loci of AvDJ1258 following procedures as described (30, 31). For A. vinelandii strains related to this work see Table 1.

Cell Growth and Protein Purification. All A. vinelandii strains were grown and harvested as described (17, 30, 31). Published methods were used for the purification of Av2 (32), ΔnifB Av1, and ΔnifB Av1^V (17, 30, 31). Preincubated ΔnifH Av1 (designated ΔnifH Av1′) was purified as described (17) except that the crude extract prepared from 500 g of cells of AvDJ1165 was incubated with 100 mg of purified Fe protein (Av2) and 100 ml of reaction mixture (0.8 mM ATP, 1.6 mM MgCl₂, 10 mM creatine phosphate and 80 units/ml of creatine phosphokinase) for 30 min at room temperature before being loaded on a Zn(II)-charged immobilized metal affinity chromatography column. For designation of protein species related to this work see Table 1.

EPR Spectroscopy. All EPR samples were prepared in a Vacuum Atmospheres (Hawthorne, CA) dry box with an oxygen level of <2 ppm. Unless noted otherwise, all samples were in 25 mM Tris·HCl (pH 8.0), 10% glycerol, and 2 mM Na₂S₂O₄. Protein samples were oxidized by incubation with excess indigo disulfonate (IDS) for 30 min. Subsequently, IDS was removed by a single passage over an anion-exchange column as described (33). Samples were either used as they were or concentrated in Centricon-30 (Amicon) in anaerobic centrifuge tubes outside of the dry box. All perpendicular- and parallel-mode EPR spectra were recorded with a Bruker ESP 300 E_z spectrophotometer, interfaced with an Oxford Instruments (Pleasanton, CA) ESR-9002 liquid helium continuous flow cryostat. All spectra were recorded at 10 K by using a microwave power of 50 mW, a gain of 5 × 10⁴, a modulation frequency of 100 kHz, and a modulation amplitude of 5 G. The microwave frequencies of 9.62 and 9.39 GHz were used for the perpendicular-mode (10 scans) and parallel-mode (144 scans) EPR spectra, respectively. Spin quantitation of the S = 1/2 perpendicular-mode EPR signal was carried out under nonsaturating conditions with Cu(II)EDTA as the standard, using the procedures described by Aasa and Vänngård (34). This method has been applied to and outlined for quantitation of similar spin systems in earlier nitrogenase studies (17, 20, 31). Quantitation numbers in this study represent the average of three independently purified protein batches.

Activity Assays, Metal Analysis, and Visible Region Spectroscopy. All nitrogenase activity and FeMoco insertion assays were carried out as described (26, 31, 35). The products H₂ and C₂H₄ were analyzed as described (36). Ammonium was determined by a HPLC fluorescence method (37). Molybdenum (38) and iron (39) were determined as described. Vanadium was analyzed by inductively coupled plasma-MS at the chemical analysis facility of the University of Georgia (Athens). Visible region absorption experiments were carried out as described (17). Acid-labile sulfur was analyzed as described (40, 41).

X-Ray Absorption Spectroscopy (XAS) Data Collection and Analysis. Samples for XAS [60 μl of 33 mg/ml of ΔnifB Av1, 19 mg/ml of ΔnifH Av1, 46 mg/ml of ΔnifH Av1′, and 57 mg/ml of ΔnifB Av1^V in 25 mM Tris·HCl (pH 8.0), 200 mM imidazole, 500 mM NaCl, 2 mM dithionite, and 50% glycerol] were loaded into 70 μl of Lucite cells with Kapton tape windows and flash-frozen in a pentane/liquid nitrogen slush. XAS data were measured at the Stanford Synchrotron Radiation Laboratory under 3-GeV, 70- to 100-mA beam conditions by using beam line 9-3, with a Si(220) double-crystal monochromator. During data collection, the samples were maintained at 10 K in an Oxford Instruments CF1208 continuous-flow liquid He cryostat. Data were collected as Fe K_a fluorescence by using a Canberra (Meriden, CT) 30-element solid-state Ge array detector. The x-ray energy was calibrated to the inflection point at 7,111.3 eV of a standard Fe foil measured concurrent with the samples. Seventeen (ΔnifH Av1′), 22 (ΔnifB Av1 and ΔnifB Av1^V), or 26 (ΔnifH Av1′) scans to k of 16.5 Å^–1 were collected. The edge position at ≈7,119 eV was monitored for evidence of photoreduction; no changes were observed in any sample. The XAS data were normalized by using the program xfit (42) by first subtracting a polynomial background absorbance that was fit to the preedge region and extended over the postedge with weighted control points, followed by fitting a three-segment polynomial spline over the extended x-ray absorption fine structure (EXAFS) region, and finally by normalizing the resulting spectra to an edge-jump of 1.0 at 7,130 eV. Fourier transforms of the EXAFS data in the k-range from 2 to 15.5 Å^–1 were calculated with the program exafspak (43).

Results

Using the fast one-step purification method described (17, 30, 31), up to ≈750 mg of His-tagged ΔnifH Av1′ and ΔnifB Av1^V were purified from 250 g of cells of AvDJ1165 and AvYM7A, respectively. Like ΔnifB Av1 and ΔnifH Av1 (Fig. 1A, lanes 2 and 3), both ΔnifH Av1′ and ΔnifB Av1^V (Fig. 1 A, lanes 4 and 5) are composed of α-subunits (≈56 and ≈54 kDa, respectively) and β-subunits (≈59 and ≈55 kDa, respectively).j The molecular masses of ΔnifH Av1′ and ΔnifB Av1^V are ≈220 and ≈160 kDa, respectively, based on their elution profiles on a gel filtration Sephacryl S-200 HR column (data not shown). This observation, in combination with results from quantitative analysis of the SDS/PAGE and N-terminal amino acid sequencing (data not shown), indicates a subunit composition of α₂β₂ for ΔnifH Av1′ and αβ₂k for ΔnifB Av1^V (Table 2). Additional small subunits, encoded by nafY and vnfG, have been reported to be associated with various species of Av1 and Av1^V, respectively (23, 27, 45, 46). However, no additional subunits are detectable in ΔnifH Av1′ or ΔnifB Av1^V based on SDS/PAGE (Fig. 1 A, lanes 4 and 5) and quantitative N-terminal amino acid sequencing (data not shown). The absence of additional subunits can also be observed in the cases of ΔnifB Av1 and ΔnifH Av1 (Fig. 1 A, lanes 2 and 3) and may be a common trait of the His-tagged versions of these proteins. Interestingly, the lack of the nafY-encoded γ-subunit from ΔnifB Av1 does not affect the catalytic or structural properties of the protein (21, 22), which strongly suggests that γ is not essential for enzymatic activity or protein assembly in this species.

Fig. 1.

SDS/PAGE, color, and visible region spectra of ΔnifB Av1, ΔnifH Av1, ΔnifH Av1′, and ΔnifB Av1^V.(A) Coomassie-stained 10–20% gradient SDS/PAGE. Lane 1, 10 μg of protein standard; lane 2, 15 μg of purified ΔnifB Av1; lane 3, 15 μg of purified ΔnifH Av1; lane 4, 15 μg of purified ΔnifH Av1′; and lane 5, 15 μg of purified ΔnifB Av1^V.(B) Anaerobic cuvettes filled with dithionite-reduced ΔnifB Av1 (part 1), ΔnifH Av1 (part 2), ΔnifH Av1′ (part 3), and ΔnifB Av1^V (part 4). The protein concentrations of all samples were 5 mg/ml. (C) Visible region spectra of samples shown in B. Spectra of dithionite-reduced ΔnifB Av1 (line 1), ΔnifH Av1 (line 2), ΔnifH Av1′ (line 3), and ΔnifB Av1^V (line 4) are shown between 350 and 750 nm. The samples were prepared as described in Materials and Methods.

Table 2. Subunit compositions, metal contents, and S = 1/2 EPR signal features.

		Metal content			S = 1/2 EPR signal
Protein	Subunit composition	Fe, mol metal/mol protein	Mo,* mol metal/mol protein	V,* mol metal/mol protein	g values			Spin integration, mol spin/mol αβ-dimer
ΔnifB Av1	α2β2†	15.4 ± 0.8	<0.01	<0.01	None
ΔnifH Av1	α2β2†	15.5 ± 0.9	<0.01	<0.01	2.06	1.92	1.89	0.4 ± 0.1
ΔnifH Av1′	α2β2†	15.4 ± 0.4	<0.01	<0.01	2.06	1.92	1.89	1.1 ± 0.2
ΔnifB Av1v	αβ2‡	8.5 ± 0.4	<0.01	<0.01	2.08	1.92	1.89	0.9 ± 0.1

^*WT Av1 and WT Av1^v protein standards show metal contents of 2.0 ± 0.1 Mo and 1.8 ± 0.2 V, respectively.

^†The α- and β-subunits are encoded by nifD and nifK, respectively. Determination of the subunit composition of the protein is based on quantitative analysis of the SDS/PAGE (Fig. 1A) and the elution profiles on the Sephacryl S-200 HR gel filtration column (data not shown).

^‡The α- and β-subunits are encoded by vnfD and vnfK, respectively. Determination of the subunit composition of the protein is based on quantitative analysis of the SDS/PAGE (Fig. 1A), quantitative analysis of the N-terminal amino acid sequence (data not shown), and the elution profiles on the Sephacryl S-200 HR gel filtration column (data not shown).

Like ΔnifB Av1 and ΔnifH Av1, ΔnifH Av1′ and ΔnifB Av1^V do not contain cofactors. This observation is consistent with the absence of molybdenum or vanadium from the proteins (Table 2), as well as the absence of the typical S = 3/2 EPR signals arising from the FeMoco or FeVco centers of Av1 and Av1^V, respectively (data not shown). The α₂β₂-tetrameric ΔnifH Av1′, like ΔnifB Av1 or ΔnifH Av1, contains ≈16 mol Fe/mol protein, whereas the αβ₂-trimeric ΔnifB Av1^V contains ≈8 mol Fe/mol protein (Table 2). Therefore, it can be assumed that all four protein species have eight Fe per complete αβ-subunit pair.

Despite having presumably the same Fe content per αβ-dimer, the colors of ΔnifH Av1′ and ΔnifB Av1^V are similar to that of ΔnifH Av1 and are clearly distinguishable from that of ΔnifB Av1. In contrast to the reddish-brown color of ΔnifB Av1, which is consistent with earlier reports (17, 21), ΔnifH Av1′ or ΔnifB Av1^V, like ΔnifH Av1 (17), is light brown (Fig. 1B). The apparent difference in color is further supported by the visible range absorption spectra of these proteins. The spectrum of ΔnifB Av1 shows a broad shoulder in the ≈475- to 575-nm range (17, 21), whereas the spectrum of ΔnifH Av1′ or ΔnifB Av1^V, like that of ΔnifH Av1 (17), is essentially featureless throughout the entire visible region (Fig. 1C). These observations indicate that ΔnifH Av1′ and ΔnifB Av1^V, like ΔnifH Av1 (19), contain metal centers different from the P-clusters in ΔnifB Av1.

It has been established that the P^N state of the P-cluster can be two-electron-oxidized by dyes like IDS to the P²⁺ state (21), which is then recognized by a g = 11.8 parallel-mode EPR signal (21, 47, 48). Such a signal is observed in the case of the IDS-oxidized ΔnifB Av1 (Fig. 2A, line 1), yet absent in the case of the IDS-oxidized ΔnifH Av1, ΔnifH Av1′, and ΔnifB Av1^V, even at a protein concentration 10-fold higher than that of ΔnifB Av1 (Fig. 2 A, lines 2–4). Meanwhile, ΔnifH Av1, ΔnifH Av1′, and ΔnifB Av1^V exhibit an S = 1/2 perpendicular-mode EPR signal in the dithionite-reduced state (Fig. 2B, lines 2–4), which integrates to 0.4, 1.1, or 0.9 spin per αβ-dimer, respectively (Table 2). This signal is absent in the case of ΔnifB Av1 because the dithionite-reduced P^N state of the P-cluster is EPR silent in this region (Fig. 2B, line 1). These results strongly suggest that ΔnifH Av1′ and ΔnifB Av1^V do not contain normal P-clusters.

Fig. 2.

EPR spectroscopy of ΔnifB Av1, ΔnifH Av1, ΔnifH Av1′, and ΔnifB Av1^V.(A) Parallel-mode EPR spectra of IDS-oxidized ΔnifB Av1 (line 1), ΔnifH Av1 (line 2), ΔnifH Av1′ (line 3), and ΔnifB Av1^V (line 4). (B) Perpendicular-mode EPR spectra of dithionite-reduced ΔnifB Av1 (line 1), ΔnifH Av1 (line 2), ΔnifH Av1′ (line 3), and ΔnifB Av1^V (line 4). (C) Parallel-mode EPR spectra of IDS-oxidized, FeMoco-reconstituted ΔnifB Av1 (line 1), ΔnifH Av1 (line 2), ΔnifH Av1′ (line 3), and ΔnifB Av1^V (line 4). The protein concentrations of parallel-mode EPR samples were 5, 50, 50, and 50 mg/ml for ΔnifB Av1, ΔnifH Av1, ΔnifH Av1′, and ΔnifB Av1^V, respectively. The protein concentrations of all perpendicular-mode EPR samples were 10 mg/ml. All spectra were measured as described in Materials and Methods. The g values are indicated.

Consistent with the EPR spectroscopic data, XAS provides further evidence of the absence of normal P-clusters in either ΔnifH Av1′ or ΔnifB Av1^V. In an earlier study, the absence of a distinct Fe-Fe scattering peak in the Fourier transform of ΔnifB Av1 was identified as being a signature of the all-ferrous P-cluster structure (19). The presence of this peak at ≈2.4 Å in the Fourier transforms of ΔnifH Av1′ and ΔnifB Av1^V indicates that these proteins do not contain standard P-clusters (Fig. 3B). Based on the similarity of the EXAFS (Fig. 3A) and Fourier transforms (Fig. 3B) of ΔnifH Av1′ and ΔnifB Av1^V to those of ΔnifH Av1, the structures of the clusters in ΔnifH Av1′ and ΔnifB Av1^V are expected to be similar to those described previously for ΔnifH Av1, which was shown to contain [4Fe-4S]-like centers (19). The shift in the edge position of the absorption spectra of ΔnifH Av1′ and ΔnifB Av1^V, by 0.4 and 0.6 eV, respectively, relative to that of ΔnifH Av1, indicates that, although structurally similar, the clusters of ΔnifH Av1′ and ΔnifB Av1^V are in a more reduced state than that of ΔnifH Av1 (Fig. 3C). Despite being shifted relative to ΔnifH Av1, the absorption edges of ΔnifH Av1′ and ΔnifB Av1^V are still higher in energy than that of ΔnifB Av1, which contains a standard all-ferrous P-cluster. Taken together, the x-ray absorption edge and EXAFS data provide clear evidence that ΔnifH Av1′ and ΔnifB Av1^V contain P-cluster variants, presumably comprising [4Fe-4S]-like fragments.

Fig. 3.

Iron K-edge EXAFS (A), nonphase-shifted EXAFS Fourier transforms (B), and x-ray absorption edge spectra (C) for ΔnifB Av1 (black), ΔnifH Av1 (green), ΔnifH Av1′ (blue), and ΔnifB Av1^V (red). The samples were prepared and measured as described in Materials and Methods.

Surprisingly, despite the fact that ΔnifH Av1′ and ΔnifB Av1^V do not contain normal P-clusters, these proteins are active in terms of H₂ evolution, C₂H₂ reduction, and N₂ fixation upon FeMoco insertion (Table 3).l With all tested substrates, ΔnifH Av1′ shows consistent activities of ≈10% of the ΔnifB Av1 standard, whereas ΔnifB Av1^V shows H₂ evolution activities of up to ≈79% of the ΔnifB Av1 standard and relatively lower N₂ and C₂H₂ reduction activities, 3% and 6%, respectively (Table 3). Both proteins show significant substrate reducing activities in patterns that closely resemble those of their respective WT counterparts (Av1^V proteins show lower activities for all substrates except H₂ compared with Av1), strongly suggesting that the metal centers contained within these proteins, which are distinct from P-clusters, are able to function in electron transfer in nitrogenase reactions and that these P-cluster variant-containing species may bear physiological relevance. Note that consistent with earlier reports (17), ΔnifH Av1 cannot be activated upon FeMoco insertion (Table 3), indicating that despite their apparent structural similarity, the cluster species in ΔnifH Av1 is different from those in ΔnifH Av1′ and ΔnifB Av1^V.

Table 3. Substrate reducing activities upon FeMoco addition.

	Activities
	C2H4 formation under C2H2/Ar		H2 formation under Ar		NH3 formation under N2		H2 formation under N2
Protein	Per mg of protein	Per nmol of αβ-dimer	Per mg of protein	Per nmol of αβ-dimer	Per mg of protein	Per nmol of αβ-dimer	Per mg of protein	Per nmol of αβ-dimer
ΔnifB Av1*	1,623 ± 15	173 ± 17 (100)	1,723 ± 82	184 ± 9 (100)	640 ± 23	68 ± 2 (100)	357 ± 43	38 ± 2 (100)
ΔnifH Av1	0	0 (0)	0	0 (0)	0	0 (0)	0	0 (0)
ΔnifH Av1′	148 ± 32	16 ± 3 (9)	172 ± 7	18 ± 1 (10)	85 ± 25	9 ± 3 (13)	26 ± 6	3 ± 1 (8)
ΔnifB Av1v†‡	66 ± 4	11 ± 1 (6)	183 ± 11	29 ± 2 (16)	15 ± 1	2 ± 0.2 (3)	188 ± 23	30 ± 2 (79)

It needs to be noted that IDS-oxidized, FeMoco-reconstituted ΔnifH Av1′ or ΔnifB Av1^v does not show the g = 11.8 parallel-mode EPR signal (Fig. 2C), which is typical for the IDS-oxidized P²⁺ state of the P-cluster. This observation indicates the presence of active P-cluster variants in ΔnifH Av1′ and ΔnifB Av1^v after FeMoco reconstitution. Activities are expressed in nmol per min per mg of protein and nmol per min per nmol of complete αβ-dimer. The percentages of activities relative to those per nmol of αβ-dimer of ΔnifB Av1 are given in parentheses. The lower detection limits were 0.01, 0.02. 0.001, and 0.02 nmol per min per mg of protein for C₂H₄ formation under C₂H₂/Ar, H₂ formation under Ar, NH₃ formation under N₂, and H₂ formation under N₂, respectively. ΔnifB Av1, ΔnifH Av1, ΔnifH Av1′, and ΔnifB Av1^v do not show any substrate reducing activities without the addition of isolated FeMoco.

^*Specific activities of FeMoco-reconstituted ΔnifB Av1 are not appreciably different from those of WT Av1 expressed by AvDJ1141 (data not shown).

^†Isolation of FeVco from WT Av1^v on a scale that is required for our experiments has not been reported so far. Therefore, it is not possible at this point to provide any data of substrate reducing activities of FeVco-reconstituted ΔnifB Av1^v. Nonetheless, a FeMoco-containing ΔnifB Av1^v “hybrid” could be formed, and its specific activities were compared with the activities of fully FeMoco-reconstituted ΔnifB Av1. In this context it should be noted that the FeMoco-containing Av1 shows considerably higher substrate reducing activities than FeVco-containing Av1^v (26, 27). Therefore, the FeMoco-reconstituted ΔnifB Av1^v will show higher percentages of activation if Av1^v is used as a standard and may represent an almost fully activated FeMoco-Av1^v hybrid. However, a conclusive assessment of the percentages of activation of ΔnifB Av1^v is only feasible when a sufficient amount of isolated FeVco is available.

^‡The FeMoco reconstitution assay contained only purified αβ₂-trimeric ΔnifB Av1^v and isolated FeMoco. An αβ₂-trimeric subunit composition of FeMoco-reconstituted ΔnifB Av1^v was confirmed by molecular weight determination of the protein on the Sephacryl S-200 HR gel filtration column (data not shown).

Discussion

For decades it has been assumed that a fully assembled P-cluster is necessary to support nitrogenase reactivity (1–11). Here, we show that certain P-cluster variants can also carry out the enzymatic activity. Both ΔnifH Av1′ and ΔnifB Av1^V contain catalytically active P-cluster variants presumably composed of [4Fe-4S]-like centers that are clearly distinct from the normal P-clusters. Interestingly, the S = 1/2 EPR signal exhibited by the P-cluster variant in ΔnifB Av1^V is also observed in WT Av1^V and thought to be associated with the activity of the protein, indicating that the same P-cluster variant is present in WT Av1^V and may represent a physiologically relevant species during the process of nitrogenase assembly and/or turnover (4, 27). The S = 1/2 EPR signal of ΔnifH Av1′ is also observed in ΔnifH Av1, the protein species isolated from the same nifH-deletion A. vinelandii strain without preincubation with Fe protein and MgATP (17). The P-cluster species in ΔnifH Av1, although structurally similar to that in ΔnifH Av1′, is catalytically inactive and cannot be activated by FeMoco insertion (17, 19, 21). Combined EPR and EXAFS data show that the P-cluster variant in ΔnifH Av1′ is more reduced than ΔnifH Av1, indicating that further reduction and/or subtle structural rearrangement of the P-cluster precursor might occur during its preincubation with Fe protein and MgATP.m The association of catalytic activity with this putatively more reduced form of ΔnifH Av1 may indicate that the P-cluster variant in ΔnifH Av1′ represents a physiologically relevant intermediate during the assembly process of MoFe protein.

Conflicting results have been reported earlier on whether a FeMoco-containing Av1^V hybrid (in contrast to the FeVco-containing Av1^V) is active in terms of N₂ reduction (23, 50). Our data clearly show that such a hybrid is indeed able to reduce N₂ and other tested substrates (Table 3). More interestingly, our FeMoco-containing Av1^V hybrid has a substrate reduction profile similar to that of the WT FeVco-containing Av1^V, rather than that of the WT FeMoco-containing Av1, most notably with respect to the following two aspects: (i) C₂H₂ is a relatively poor substrate and (ii) H₂ evolution in the presence of N₂ occurs at a high rate. This observation is consistent with the general belief that the enzymology of substrate reduction in A. vinelandii depends on the immediate protein environment surrounding the cofactor.

The incomplete αβ₂-trimeric subunit composition has been reported earlier for an Av1^V species generated during the time-consuming purification of the highly unstable WT Av1^V (27). The same subunit composition of αβ₂ is observed here in ΔnifB Av1^V, with the notable absence of the small vnfG-encoded γ-subunit. It appears that the lack of cofactor/γ-subunit, which is a direct/indirect result of the deletion of nifB gene, has a significant impact on the structural integrity of the protein. Nevertheless, the intact αβ dimer of ΔnifB Av1^V is catalytically active upon insertion of FeMoco (Table 3). This observation is contradictory to some earlier reports (23) and indicates that the vnfG-encoded γ-subunit is not essential for FeMoco insertion and protein activity; rather, it might be required for the overall stability of the protein. Recently, we showed that the two αβ-dimers of Av1 of the Mo-nitrogenase are assembled in a stepwise fashion and that the formation of the second αβ-dimer requires additional factors (31). The incomplete αβ₂-trimeric subunit composition of the ΔnifB Av1^V, which is a result of the disassembly of the unstable protein, might reflect, in an opposite sense, the fact that the Av1^V, like Av1, is assembled stepwise and that the vnfG-encoded γ-subunit might be involved in the assembly of the second αβ-dimer of Av1^V and contribute to the stability of the fully assembled protein. It needs to be noted, however, that our study uses FeMoco for the activation experiments and not FeVco, the native cofactor of Av1^V. Therefore, the role of vnfG-encoded γ-subunit in the assembly process needs to be addressed further when a sufficient amount of isolated FeVco is available.

In summary, we show that nitrogenase reactivity can be achieved by metalloclusters other than P-clusters. Although the physiological relevance, the detailed structures, and the oxidation states of these species await further investigation, the dogma that “nitrogen is reduced when P-clusters are produced” is shaken, giving way to additional perspectives on how this fascinating enzyme works and what can be learned about it.