Klebsiella pneumoniae nitrogenase: formation and stability of putative beryllium fluoride-ADP transition state complexes.

Incubation of the MoFe protein (Kp1) and Fe protein (Kp2), the component proteins of Klebsiella pneumoniae nitrogenase, with BeF(3)(-) and MgADP resulted in a progressive inhibition of nitrogenase activity. We have shown that at high Kp2 to Kp1 molar ratios this inhibition is due to the formation of an inactive complex with a stoichiometry corresponding to Kp1.{Kp2.(MgADP.BeFx)2}2. At lower Kp2:Kp1 ratios, an equilibrium between this 2:1 complex, the partially active 1:1 Kp1.Kp2.(MgADP. BeFx)2 complex, and active nitrogenase components was demonstrated. The inhibition was reversible since incubation of the 1:1 complex in the absence of MgADP and beryllium resulted in complete restoration of activity over 30 h. Under pseudo-first-order conditions with regard to nitrogenase components and MgADP, the kinetics of the rate of inhibition with increasing concentrations of BeF(3)(-) showed a square dependence on [BeF(3)(-)], consistent with the binding of two Be atoms by Kp2 in the complex. Analytical fplc gel filtration profiles of Kp1.Kp2 incubation mixtures at equilibrium resolved the 2:1 complex and the 1:1 complex from free Kp1. Deconvolution of the equilibrium profiles gave concentrations of the components allowing constants for their formation of 2.1 x 10(6) and 5.6 x 10(5) M(-1) to be calculated for the 1:1 and 2:1 complexes, respectively. When the active site concentration of the different species was taken into account, values for the two constants were the same, indicating the two binding sites for Kp2 are the same for Kp1 with one or both sites unoccupied. The value for K(1) we obtain from this study is comparable with the value derived from pre-steady-state studies of nitrogenase. Analysis of the elution profile obtained on gel filtration of a 1:1 ratio incubation mixture containing 20 microM nitrogenase components showed 97% of the Kp2 present initially to be complexed. These data provide the first unequivocal demonstration that Fe protein preparations which may contain up to 50% of a species of Fe protein defective in electron transfer is nevertheless fully competent in complex formation with MoFe protein.     

Oxidation of nitrogenase iron protein by dioxygen without inactivation could contribute to high respiration rates of Azotobacter species and facilitate nitrogen fixation in other aerobic environments.

The kinetics of oxidation of the Fe proteins of nitrogenases from Klebsiella pneumoniae (Kp2) and Azotobacter chroococcum (Ac2) by O2 and H2O2 have been studied by stopped-flow spectrophotometry at 23 degrees C, pH 7.4. With excess O2, one-electron oxidation of Kp2 and Ac2 and their 2 MgATP or 2 MgADP bound forms occurs with rate constants (k) in the range 5.3 x 10(3) M-1.S-1 to 1.6 x 10(5) M-1.S-1. A linear correlation between log k and the mid-point potentials (Em) of these protein species indicates that the higher rates of electron transfer from the Ac2 species are due to the differences in Em of the 4Fe-4S cluster. The reaction of Ac2(MgADP)2 with O2 is sufficiently rapid for it to contribute significantly to the high respiration rate of Azotobacter under N2-fixing conditions and may represent a new respiratory pathway. Excess O2 rapidly inactivates Ac2(MgADP)2 and Kp2(MgADP)2; however, when these protein species are in greater than 4-fold molar excess over the concentration of O2, 4 equivalents of protein are oxidized with no loss of activity. The kinetics of this reaction suggest that H2O2 is an intermediate in the reduction of O2 to 2 H2O by nitrogenase Fe proteins and imply a role for catalase or peroxidase in the mechanism of protection of nitrogenase from O2-induced inactivation.     

The mechanism of Klebsiella pneumoniae nitrogenase action. The determination of rate constants required for the simulation of the kinetics of N2 reduction and H2 evolution.

Kinetic data for Klebsiella pneumoniae nitrogenase were used to determine the values of nine of the 17 rate constants that define the scheme for nitrogenase action described by Lowe & Thorneley [(1984) Biochem. J. 224, 877-886]. Stopped-flow spectrophotometric monitoring of the MgATP-induced oxidation of the Fe protein (Kp2) by the MoFe protein (Kp1) was used to determine the rates of association (k+1) and dissociation (k-1) of reduced Kp2(MgATP)2 with Kp1. The dependences of the apparent KNm2 on Fe protein/MoFe protein ratio and H2 partial pressure were used to determine the mutual displacement rates of N2 and H2 (k+10, k-10, k+11 and k-11). These data also allowed the rate constants for H2 evolution from progressively more reduced forms of Kp1 to be determined (k+7, k+8 and k+9). A mechanism for N2-dependent catalysis of 1H2H formation from 2H2 that requires H2 to be a competitive inhibitor of N2 reduction is also presented.     

Long-range interactions between the Fe protein binding sites of the MoFe protein of nitrogenase

We report the properties and reactivity of the catalytically active heterologous nitrogenase formed between the Fe protein from Clostridium pasteurianum (Cp2) and the MoFe protein from Klebsiella pneumoniae (Kp1). Under turnover conditions, in the presence of MgATP, a stable 2:1 (Cp2)2Kp1 electron transfer complex is formed, in which the [4Fe-4S]+ centre of Cp2 is protected from chelation by alpha,alpha'-bipyridyl. However, the two Fe protein-binding sites on Kp1 are not equivalent, since a 1:1 Cp2.Kp1 complex was isolated by gel filtration. The non-equivalence of the Fe protein binding sites was also indicated by the inhibition pattern of Klebsiella nitrogenase by Cp2. The EPR spectrum of the isolated 1:1 Cp2.Kp1 complex showed an S=1/2 signal characteristic of dithionite-reduced Cp2 and signals with g values of 4.27, 3.73, 2.01 and 4.32, 3.63, 2.00 characteristic of the high- and low-pH forms of the FeMoco centre of Kp1, respectively. The unoccupied binding site of Kp1 of the isolated 1:1 Cp2Kp1 complex was shown to be catalytically fully functional in combination with Kp2. In contrast to homologous nitrogenases, which require MgATP for detectable rates of electron transfer from the Fe protein, stopped-flow kinetic studies revealed that electron transfer from Cp2 to Kp1 occurred in the absence of MgATP with a rate constant of 0.065 s(-1). Subsequently, a slower transient decrease and restoration of absorption in the electronic spectrum in the 500-700 nm region was observed. These changes corresponded with those in the intensity of the S=3/2 EPR signal of the FeMoco centres of Kp1 and were consistent with the transient reduction of the FeMoco centre of Kp1 to an EPR-silent form, followed by restoration of the signal at longer reaction times. These changes were not associated with catalysis since no evolution of H2 was detectable.     

Klebsiella pneumoniae nitrogenase. Inhibition of hydrogen evolution by ethylene and the reduction of ethylene to ethane.

Ethylene (C2H4) inhibited H2 evolution by the Mo-containing nitrogenase of Klebsiella pneumoniae. The extent of inhibition depended on the electron flux determined by the ratio of Fe protein (Kp2) to MoFe protein (Kp1) with KiC2H4 = 409 kPa ([Kp2]/[Kp1] = 22:1) and KC2H4i = 88 kPa ([Kp1]/[Kp2] = 21:1) at 23 degrees C at pH 7.4. At [Kp2]/[Kp1] = 1:1, inhibition was minimal with C2H4 (101 kPa). Extrapolation of data obtained when C2H4 was varied from 60 to 290 kPa indicates that at infinite pressure of C2H4 total inhibition of H2 evolution should occur. C2H4 inhibited concomitant S2O4(2-) oxidation to the same extent that it inhibited H2 evolution. Although other inhibitors of total electron flux such as CN- and CH3NC uncouple MgATP hydrolysis from electron transfer, C2H4 did not affect the ATP/2e ratio. Inhibition of H2 evolution by C2H4 was not relieved by CO. C2H4 was reduced to C2H6 at [Kp2]/[Kp1] ratios greater than or equal to 5:1 in a reaction that accounted for no more than 1% of the total electron flux. These data are discussed in terms of the chemistry of alkyne and alkene reduction on transition-metal centres.     

Kinetic evidence for the existence of an unstable intermediate in the trinitrophenylation-induced rhodanese inactivation reaction.

1. Rhodanese inactivation by 2,4,6-trinitrobenzenesulphonate, in the presence of n-butylamine in the reaction medium, has been studied by a kinetic analysis of the data, based on the assumption that enzyme inactivation is brought about by direct reaction of this with the modifying agent. 2. Initial reaction rates for rhodanese activity loss were determined by a mathematical analysis of the first three recorded values of rhodanese residual activity. 3. It was found that fractional rhodanese activity values, at infinite reaction time with 2,4,6-trinitrobenzenesulphonate (end-point values), were significantly lower than the values calculated on the assumption of rhodanese inactivation being entirely due to direct trinitrophenylation of enzyme protein. 4. Also, initial enzyme inactivation values were higher in the presence, rather than in the absence, of n-butylamine. 5. These results indicate that 2,4,6-trinitrobenzenesulphonate-induced rhodanese inactivation, in the presence of n-butylamine in the reaction medium, is due to the generation of a highly reactive, unstable intermediate, probably a free radical species.     

Nitrogenase from Klebsiella pneumoniae. An e.p.r. signal observed during enzyme turnover under ethylene is associated with the iron-molybdenum cofactor.

During turnover at 10 degrees C at pH 7.4 in the presence of ethylene, the MoFe protein of Klebsiella pneumoniae nitrogenase (Kp 1) exhibited an electron-paramagnetic-resonance signal with g-values at 2.12, 1.998 and 1.987. 57Fe isotopic substitution demonstrated that this signal arose from the Kp 1 FeMo-cofactor in an S = 1/2 spin state.     

Purification and characterization of the inactive MoFe protein (NifB-Kp1) of the nitrogenase from nifB mutants of Klebsiella pneumoniae.

The inactive MoFe protein of nitrogenase, NifB-Kp1, from two distinct nifB mutants of Klebsiella pneumoniae, Kp5058 (a nifB point mutant) and UNF1718 (a nifB, nifJ double mutant) has been purified and characterized. NifB-Kp1 can be activated by reaction with the iron-molybdenum cofactor, FeMoco, extracted from active MoFe protein. NifB-Kp1 purified from either source had similar properties and was contaminated with an approximately equimolar amount of protein of mol.wt. 21 000. Like active wild-type Kp1, it was an alpha 2 beta 2 tetramer, but it was far less stable than Kp1, deteriorating rapidly at temperatures above 8 degrees C or on mild oxidation. NifB-Kp1 preparations contained 0.4-0.9 Mo and 9.0 +/- 0.9 Fe atoms . mol-1 and, when activated by FeMoco, had a specific activity of approx. 500 units . mg-1. The Mo in our preparations was not associated with the e.p.r. signal normally observed from FeMoco. All preparations exhibited a weak gav. = 1.95 e.p.r. signal which was probably not associated with activatable protein.     

Nitrogenase of Klebsiella pneumoniae. Kinetics of the dissociation of oxidized iron protein from molybdenum-iron protein: identification of the rate-limiting step for substrate reduction.

Stopped-flow spectrophotometry and e.p.r. spectroscopy were used to study the kinetics of reduction by dithionite of the oxidized Fe protein of nitrogenase from Klebsiella pneumoniae (Kp2ox.) in the presence of MgADP at 23 degrees C at pH 7.4. The active reductant, SO2.-, produced by the predissociation of S2O4(2-) in equilibrium 2SO2.-, reacts with Kp2ox. (MgADP)2, with k4 = 3.0 X 10(6) +/- 0.4 X 10(6) M-1 X s-1. The inhibition of this reaction by the Mo-Fe protein (Kp1) has enabled the rate of dissociation of Kp2ox. (MgADP)2 from Kp1+ (the Kp2-binding site on Kp1) to be measured (k-3 = 6.4 +/- 0.8 s-1). Comparison with the steady-state rate of substrate reduction shows that the dissociation (k-3) of the complex Kp2ox. (MgADP)2-Kp1+, which is formed after MgATP-induced electron transfer from Kp2 to Kp1+, is the rate-limiting step in the catalytic cycle for substrate reduction.     

The inactive MoFe protein (NifB-Kp1) of the nitrogenase from nifB mutants of Klebsiella pneumoniae. Its interaction with FeMo-cofactor and the properties of the active MoFe protein formed.

The inactive MoFe protein (NifB-Kp1) of nitrogenase from nifB mutants of Klebsiella pneumoniae may be activated by addition of the iron-molybdenum cofactor (FeMoco) extracted from active MoFe protein (Kp1). However, when apparently saturated with FeMoco, our preparations of NifB-Kp1 yielded activated protein, Kp1-asm, with a specific activity that was at best only 40% of that expected. This was not due to degradation of Kp1-asm, NifB-Kp1 or FeMoco during the activation reaction. Nor could activation be enhanced by addition of other nif-gene products or other proteins. Whereas fully active Kp1 contains 2 FeMoco/molecule, apparent saturation of our NifB-Kp1 preparations required the binding of only 0.4-0.65 FeMoco/molecule. By using chromatography Kp1-asm could be largely resolved from NifB-Kp1 that had not been activated. However, we were unable to isolate fully active MoFe protein (i.e. Kp1-asm containing 2 FeMoco/molecule) from solutions of NifB-Kp1 activated with FeMoco. The maximum activity/ng-atom of total Mo obtained for our purified Kp1-asm was approximately half the maximum activity for FeMoco. Since all NifB-Kp1 preparations contained some Mo, we suggest that FeMoco activated only those NifB-Kp1 molecules already containing one atom of (non-FeMoco) Mo, thus forming Kp1-asm with 2 Mo but only 1 FeMoco/molecule. Kp1-asm was identical with normal Kp1 in terms of its Mr, stability, e.p.r. signals, pattern of substrate reductions, CO inhibition and ATP/2e ratio. In addition, for preparations of differing specific activity, there was a constant and identical relationship between the e.p.r. signal intensity (from FeMoco) and the activity of both Kp1 and Kp1-asm. Assuming the above hypothesis on the structure of Kp1-asm, these data demonstrate that the two FeMoco sites in wild-type Kp1 operate independently.     

Electron-transfer studies involving flavodoxin and a natural redox partner, the iron protein of nitrogenase. Conformational constraints on protein-protein interactions and the kinetics of electron transfer within the protein complex.

The kinetics of electron-transfer reactions involving flavodoxins from Klebsiella pneumoniae (KpFld), Azotobacter chroococcum (AcFld), Anacystis nidulans (AnFld) and Megasphaera elsdenii (MeFld), the free, MgADP-bound and MgATP-bound forms of the Fe protein component of nitrogenase from K. pneumoniae [Kp2, Kp2(MgADP)2 and Kp2(MgATP)2] and Na2S2O4 were studied by stopped-flow spectrophotometry. Kinetic evidence was obtained for the formation of binary protein complexes involving KpFldSQ (semiquinone) with either Kp2(MgADP)2 (KD = 49 microM) or Kp2(MgATP)2 (KD = 13 microM) but not with Kp2 (KD greater than 730 microM). The binding of 2MgATP or 2MgADP to Kp2 therefore not only shifts the midpoint potential (Em) of the [4Fe-4S] centre from -200 mV to -320 mV or -350 mV respectively but also changes the affinity of Kp2 for KpFldSQ. Thermodynamically unfavourable electron from Kp2(MgADP)2 and Kp2(MgATP)2 to KpFldSQ occurs within the protein complexes with k = 1.2 s-1 (delta E = -72 mV) and 0.5 s-1 (delta E = -120 mV) respectively. Although AcFldSQ is reduced by Kp2, Kp2(MgADP)2 and Kp2(MgATP)2 (k = 8 x 10(3), 2.4 x 10(3) and 9 x 10(2) M-1.s-1 respectively), protein-complex formation is weak in each case (KD greater than 700 microM). Electron transfer in the physiologically important and thermodynamically favourable direction from Kp2FldHQ (hydroquinone) and AcFldHQ to Kp2ox.(MgADP)2 (the state of Kp2 that accepts electrons from FldHQ in the catalytic cycle of nitrogenase) is rapid (k greater than 10(6) M-1.s-1). The second-order rate constants for the reduction of KpFldSQ, AcFldSQ, AnFldSQ and MeFldSQ by SO2.- (active reductant formed by the predissociation of S2O4(2-) ion) exhibited the linear free-energy relationship predicted by the Marcus theory of electron transfer.     

Nitrogenase of Klebsiella pneumoniae nifV mutants.

The MoFe protein of nitrogenase from Klebsiella pneumoniae nifV mutants, NifV- Kp1 protein, in combination with the Fe protein from wild-type cells, catalysed CO-sensitive H2 evolution, in contrast with the CO-insensitive reaction catalysed by the wild-type enzyme. The decrease in H2 production was accompanied by a stoicheiometric decrease in dithionite (reductant) utilization, implying that CO was not reduced. However, CO did not affect the rate of phosphate release from ATP. Therefore the ATP/2e ratio increased, indicating futile cycling of electrons between the Fe protein and the MoFe protein. The inhibition of H2 evolution by CO was partial; it increased from 40% at pH6.3 to 82% at pH 8.6. Inhibition at pH7.4 (maximum 73%) was half-maximal at 3.1 Pa (0.031 matm) CO. The pH optimum of the mutant enzyme was lower in the presence of CO. Steady-state kinetic analysis of acetylene reduction indicated that CO was a linear, intersecting, non-competitive inhibitor of acetylene reduction with Kii = 2.5 Pa and Kis = 9.5 Pa. This may indicate that a single high-affinity CO-binding site in the NifV- Kp1 protein can cause both partial inhibition of H2 evolution and total elimination of acetylene reduction. Various models to explain the data are discussed.     

Studies on fatty acid-binding proteins. The diurnal variation shown by rat liver fatty acid-binding protein.

The concentration of fatty acid-binding protein in rat liver was examined by SDS/polyacrylamide-gel electrophoresis, by Western blotting and by quantifying the fluorescence enhancement achieved on the binding of the fluorescent probe 11-(dansylamino)undecanoic acid. A 2-3-fold increase in the concentration of this protein produced by treatment of rats with the peroxisome proliferator tiadenol was readily detected; however, only a small variation in the concentration of the protein due to a diurnal rhythm was observed. This result contradicts the 7-10-fold variation previously reported for this protein [Hargis, Olson, Clarke & Dempsey (1986) J. Biol. Chem. 261, 1988-1991].     

The mechanism of replication of phiX174 single-stranded DNA. X. Distribution of the gaps in nascent phiX174 replicative form DNA

When the enzyme rhodanese (EC 2. 8. 1. 1) is digested with trypsin under controlled conditions, the parent protein is converted from a polypeptide of molecular weight 32,600 to a polypeptide of molecular weight 28,800. This proteolytic conversion occurs with no loss of rhodanese activity. In fact, preliminary results indicate that the polypeptide produced by proteolysis has higher sulfur transferase activity than the native rhodanese.     

Acid pH-induced conformational changes in bovine liver rhodanese.

The enzyme rhodanese is greatly stabilized in the range pH 4-6, and samples at pH 5 are fully active after several days at 23 degrees C. This is very different from results at pH greater than 7, where there is significant loss of activity within 1 h. A pH-dependent conformational change occurs below pH 4 in a transition centered around pH 3.25 that leads slowly to inactive rhodanese at pH 3 (t 1/2 = 22 min at pH3). The inactive rhodanese can be reactivated by incubation under conditions required for detergent-assisted refolding of denatured rhodanese. The inactive enzyme at pH 3 has the maximum of its intrinsic fluorescence spectrum shifted to 345 nm from 335 nm, which is characteristic of native rhodanese at pH greater than 4. At pH 3, rhodanese shows increased exposure of organized hydrophobic surfaces as measured by 1,1'-bis(4-anilino)naphthalene-5,5'-disulfonic acid binding. The secondary structure is maintained over the entire pH range studied (pH 2-7). Fluorescence anisotropy measurements of the intrinsic fluorescence provide evidence suggesting that the pH transition produces a state that does not display greatly increased average flexibility at tryptophan residues. Pepsin digestibility of rhodanese follows the pH dependence of conformational changes reported by activity and physical methods. Rhodanese is resistant to proteolysis above pH 4 but becomes increasingly susceptible as the pH is lowered. The form of the enzyme at pH 3 is cleaved at discrete sites to produce a few large fragments. It appears that pepsin initially cleaves close to one end of the protein and then clips at additional sites to produce species of a size expected for the individual domains into which rhodanese is folded. Overall, it appears that in the pH range between pH 3 and 4, titration of groups on rhodanese leads to opening of the structure to produce a conformation resembling, but more rigid than, the molten globule state that is observed as an intermediate during reversible unfolding of rhodanese.     

Expression of nitrogen fixation genes in foreign hosts. Assembly of nitrogenase Fe protein in Escherichia coli and in yeast

In Klebsiella pneumoniae, the nifH gene encodes the Fe protein (Kp2) polypeptide that is assembled into a homodimer responsible for the reduction of nitrogenase. Escherichia coli or the yeast Saccharomyces cerevisiae, transformed with the K. pneumoniae nifH gene in suitable expression vectors, synthesize the Fe protein polypeptide. This study examines the assembly of the nifH gene product into its characteristic dimeric structure in E. coli and in yeast. Immunoblotting methods, as well as 55Fe2- labeling of K. pneumoniae were employed to detect native nitrogenase components in cell lysates. E. coli and yeast transformants contained a protein similar to native Kp2 in its immunoreactivity, apparent molecular weight, and lability in the presence of oxygen or MgATP. While in E. coli the co-introduction of nifH and nifM resulted in enhanced levels of the nifH product, it appears that the nifH gene product alone is sufficient for the assembly of an Fe protein-like structure in foreign prokaryotic and eukaryotic hosts.     

A study of the single polypeptide nature of rhodanese. A comparison of different preparations.

The enzyme rhodanese (EC 2.8.1.1) appears as a single polypeptide chain protein on sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The molecular weight of this species is approx. 33 000. This contrasts with previous reports that rhodanese behaves on gel filtration chromatography as a rapidly equilibrating monomer-dimer system composed of identical subunits with a molecular weight of 18 500. We have investigated this apparent discrepancy by isolating the enzyme by the two different preparative procedures used in the above investigations. The two crystalline samples were subjected to gel filtration chromatography under a wide variety of conditions and to sodium dodecyl sulfate disc gel electrophoresis. The two preparations yielded rhodanese which behaved identically and no evidence for the monomeric species was obtained under any experimental condition tested. Thin-layer gel chromatography of clarified liver homogenates gave no evidence of rhodanese species other than that present in the purified samples. The variation in molecular weights observed in gel filtration chromatography may be a reflection of the conformational mobility of the enzyme leading to solvent-dependent changes in Stokes radius. If rhodanese is dimeric, special interactions must stabilize it under the conditions tested here.     

Rat brain fatty acid-binding protein during development.

Cytosolic fatty acid-binding proteins (FABPs) have been described in rat and bovine whole brain. In the present study we investigated the distribution of FABP among white matter and gray matter as well as its changes during development. Fatty acid binding activity was similar in white and gray matter up to 40 days of age. In white matter it showed an age dependent increase thereafter, while in gray matter it remained constant throughout. Gel filtration (Sephadex G-75) of white matter cytosol of adult female rats resolved the fatty acid-binding activity in two peaks: A (Vo) and B (12-14 KDa; FABP). The specific binding activity in the FABP fraction was 10.4 pmol/micrograms of protein. The activity in peak A showed an age-dependent increase which paralleled myelin deposition. In contrast, the activity in the FABP fraction (peak B) remained undetectable up to 40 days of age, increasing thereafter. The differential distribution of cellular brain proteins with the capacity to bind fatty acids in gray matter and white matter suggests that this activity could be related to glial cells or to cell related structures such as myelin.     

Transition state complexes of the Klebsiella pneumoniae nitrogenase proteins. Spectroscopic properties of aluminium fluoride-stabilized and beryllium fluoride-stabilized MgADP complexes reveal conformational differences of the Fe protein.

Stable inactive 2 : 1 complexes of the Klebsiella pneumoniae nitrogenase components (Kp2/Kp1) were prepared with ADP or the fluorescent ADP analogue, 2'(3')-O-[N-methylanthraniloyl] ADP and AlF(4)(-) or BeF(3)(-) ions. By analogy with published crystallographic data [Schindelin et al. (1997) Nature 387, 370-376)], we suggest that the metal fluoride ions replaced phosphate at the two ATP-binding sites of the iron protein, Kp2. The beryllium (BeF(x)) and aluminium (AlF(4)(-)) containing complexes are proposed to correspond to the ATP-bound state and the hydrolytic transition states, respectively, by analogy with the equivalent complexes of myosin [Fisher et al. (1995) Biochemistry 34, 8960-8972]. (31)P NMR spectroscopy showed that during the initial stages of complex formation, MgADP bound to the complexed Kp2 in a manner similar to that reported for isolated Kp2. This process was followed by a second step that caused broadening of the (31)P NMR signals and, in the case of the AlF4- complex, slow hydrolysis of some of the excess ADP to AMP and inorganic phosphate. The purified BeFx complex contained 3.8 +/- 0.1 MgADP per mol Kp1. With the AlF(4)(-) complex, MgAMP and adenosine (from MgAMP hydrolysis) replaced part of the bound MgADP although four AlF(4)(-) ions were retained, demonstrating that full occupancy by MgADP is not required for the stability of the complex. The fluorescence emission maximum of 2'(3')-O-[N-methylanthraniloyl] ADP was blue-shifted by 6-8 nm in both metal fluoride complexes and polarization was 6-9 times that of the free analogue. The fluorescence yield of bound 2'(3')-O-[N-methylanthraniloyl] ADP was enhanced by 40% in the AlF(4)(-) complex relative to the solvent but no increase in fluorescence was observed in the BeFx complex. Resonance energy transfer from conserved tyrosine residues located in proximity to the Kp2 nucleotide-binding pocket was marked in the AlF(4)(-) complex but minimal in the BeFx fluoride complex, illustrating a clear conformational difference in the Fe protein of the two complexes. Our data indicate that complex formation during the nitrogenase catalytic cycle is a multistep process involving at least four conformational states of Kp2: similar to the free Fe protein; as initially complexed with detectable (31)P NMR; as detected in mature complexes with no detectable (31)P NMR; in the AlF(4)(-) complex in which an altered tyrosine interaction permits resonance energy transfer with 2'(3')-O-[N-methylanthraniloyl] ADP.