Evidence That MgATP Accelerates Primary Electron Transfer in aClostridium pasteurianum Fe Protein-Azotobacter vinelandii MoFe Protein Nitrogenase Tight Complex*

The nitrogenase catalytic cycle involves binding of the iron (Fe) protein to the molybdenum-iron (MoFe) protein, transfer of a single electron from the Fe protein to the MoFe protein concomitant with the hydrolysis of at least two MgATP molecules, followed by dissociation of the two proteins. Earlier studies found that combining the Fe protein isolated from the bacteriumClostridium pasteurianum with the MoFe protein isolated from the bacterium Azotobacter vinelandii resulted in an inactive, nondissociating Fe protein-MoFe protein complex. In the present work, it is demonstrated that primary electron transfer occurs within this nitrogenase tight complex in the absence of MgATP (apparent first-order rate constant k = 0.007 s−1) and that MgATP accelerates this electron transfer reaction by more than 10,000-fold to rates comparable to those observed within homologous nitrogenase complexes (k = 100 s−1). Electron transfer reactions were confirmed by EPR spectroscopy. Finally, the midpoint potentials (Em ) for the Fe protein [4Fe-4S]2+/+ cluster and the MoFe protein P2+/N cluster were determined for both the uncomplexed and complexed proteins and with or without MgADP. Calculations from electron transfer theory indicate that the measured changes in Em are not likely to be sufficient to account for the observed nucleotide-dependent rate accelerations for electron transfer.

The metalloenzyme nitrogenase, which catalyzes the biological reduction of nitrogen (N 2 ) to ammonia (NH 3 ), is composed of two separable component proteins (1). One component, the molybdenum-iron (MoFe) protein, 1 is an ␣ 2 ␤ 2 tetramer containing two [7Fe-Mo-9S-homocitrate] cofactors (FeMoco), which are the site of substrate reduction, and two P or [8Fe-7S] clusters, which are believed to mediate electron transfer to FeMoco (2,3). The other component, the iron (Fe) protein, is a homodimer containing a single [4Fe-4S] cluster bridged between the subunits and two nucleotide binding sites (4,5). The current model for the catalytic mechanism of nitrogenase holds that the reduced Fe protein, with two bound MgATP molecules, binds to the MoFe protein, and a single electron is transferred from the Fe protein to the MoFe protein in a reaction that is somehow coupled to the hydrolysis of two MgATP molecules (6). The oxidized Fe protein, with two molecules of MgADP bound, then dissociates from the MoFe protein (7). Several cycles of component protein binding, MgATP hydrolysis, and intercomponent electron transfer are required to reduce substrates by multiple electrons (8,9).
MgATP appears to play several key roles in the reduction of substrates by nitrogenase (8). First, the binding of MgATP to the Fe protein induces conformational changes within the Fe protein (10). These conformational changes influence the binding of the Fe protein to the MoFe protein (11). Once the two proteins have associated, MgATP hydrolysis functions somehow to facilitate electron transfer from the Fe protein to the MoFe protein. Finally, the hydrolysis of MgATP to MgADP and P i is believed to be involved in stimulating the dissociation of the Fe protein from the MoFe protein (12)(13)(14).
The association constant (K a ) for the Cp2⅐Av1 complex was determined to be approximately 10-fold higher than that for the homologous A. vinelandii nitrogenase complex and 100-fold higher than that for the homologous C. pasteurianum nitrogenase complex (15). In addition, this protein-protein complex could form with or without nucleotides (17). Although the Cp2⅐Av1 complex is inactive in all substrate reduction activities, it was later found (18) that this complex could still hydrolyze MgATP to MgADP ϩ P i at low rates (36-fold lower than the homologous protein complexes).
In the present work, we present evidence that primary electron transfer, but not subsequent electron transfer reactions, occurs within the Cp2⅐Av1 tight complex. Primary electron transfer was found to occur without added nucleotides at low rates, but the addition of MgATP accelerated electron transfer by more than 10,000-fold to rates near those observed in the homologous protein complexes. How these results relate to our current understanding of the roles of MgATP in the nitrogenase mechanism is discussed.

EXPERIMENTAL PROCEDURES
Expression and Purification of Nitrogenase Proteins-Nitrogenase Fe and MoFe proteins were expressed in A. vinelandii cells and purified to apparent homogeneity essentially as described previously (19,20). C. pasteurianum strain W5 (a gift from J.-S. Chen, Virginia Tech) was grown under N 2 -fixing conditions essentially as described (21), and the nitrogenase Fe protein was purified (22). Protein concentrations were determined by a modified biuret method (23), with bovine serum albumin as the standard, or by the visible absorption using absorption coefficients of 11.1 mM Ϫ1 ⅐cm Ϫ1 at 400 nm for reduced Av2 and Cp2 (24), and 62.3 mM Ϫ1 ⅐cm Ϫ1 at 400 nm for reduced Av1 (24,25). All of the nitrogenase proteins used in this study had specific activities of at least 1,800 nmol of acetylene reduced⅐min Ϫ1 ⅐mg of protein Ϫ1 . Protein manipulations were performed in the absence of O 2 in sealed serum vials under an argon atmosphere or in an argon-filled glovebox (Vacuum Atmospheres, Hawthorne CA).
UV-visible Absorption Spectra of Av2 and Cp2-Absorption spectra were recorded on a Hewlett-Packard 8452A diode array spectrophotometer in 2.0-ml quartz cuvettes fitted with serum stoppers. Anaerobicity was obtained by purging the cuvettes with O 2 -free argon for 6 min. The buffer used for these experiments was 100 mM MOPS, pH 7.0, with 2 mM dithionite. A MgATP-regenerating system was included, as indicated, during nitrogenase catalysis to prevent the formation of MgADP, which is a known inhibitor of MgATP binding to the Fe protein (26).
Oxidation of Av1-The Av1 P clusters were oxidized by 2 eq of electrons by the addition of an excess of oxidized indigo disulfonate (27). This was done by exchanging dithionite-reduced Av1 into anaerobic, dithionite-free 100 mM MOPS buffer, pH 7.0, with 250 mM NaCl, by passage through a Sephadex G-25 column in an anaerobic glovebox. Av1 was then oxidized to the P 2ϩ state by the addition of an excess of oxidized indigo disulfonate (28), and the mixture was allowed to incubate for 15 min. Indigo disulfonate was then separated from Av1 by passage through a Dowex-1 (Sigma) column equilibrated with 50 mM MOPS buffer, pH 7.0. The oxidized Av1 (P 2ϩ ) was collected, and the concentration was determined from the absorption spectrum and the known absorption coefficient of 73 mM Ϫ1 ⅐cm Ϫ1 at 400 nm (29). Reduced, dithionite-free Cp2 was prepared by passage of the protein through a Sephadex G-25 column equilibrated with 50 mM MOPS buffer, pH 7.0.
EPR Spectroscopy-EPR spectra were recorded on a Bruker ESP300E spectrometer equipped with a dual-mode cavity and an Oxford ESR 900 liquid helium cryostat. In all cases, 4-mm calibrated quartz EPR tubes (Wilmad, Buena, NJ) were used. All spectra were recorded at 12 K with 1,024 points/scan. The conversion time and time constant were 10.24 ms. All other parameters are noted in the figure legends.
Stopped-flow Spectrophotometry-Electron transfer from Cp2 to Av1 was monitored by the increase in the visible absorbance of Cp2 upon oxidation of the [4Fe-4S] cluster from the reduced (ϩ) to the oxidized (2ϩ) state (30). This electron transfer reaction was monitored in real time by use of a Hi-Tech SF61 stopped-flow spectrophotometer equipped with a data acquisition and curve fitting system (Salisbury, Wilts, U. K.). The SHU-61 sample handling unit was kept inside an anaerobic glovebox (Coy Products, Grass Lake, MI) with a gas atmosphere of 95% N 2 and 5% H 2 and an oxygen concentration less than 1 ppm oxygen. Reactant solutions were thermostatted to within Ϯ 0.1°C by means of an FC-200 Techne flow cooler attached to a closed circulation Techne C-85D water circulator (Techne Ltd., Duxford, Cambridge, U. K.). Data were collected at 430 nm for the oxidation/reduction of Fe protein. Earlier work demonstrated no significant change in the absorption coefficient of the MoFe protein at 430 nm resulting from electron transfer from the Fe protein (30). All reactions were carried out in 100 mM HEPES buffer, pH 7.4, with 2 mM dithionite. In all cases, reactions were initiated by rapidly mixing reactants contained in the two drive syringes of the stopped-flow instrument. The instrument mixing time was determined to be approximately 4 ms. Reaction conditions are noted in the appropriate figure legends.
Apparent first-order rate constants for electron transfer (k obs ) were determined from nonlinear, least squares fits of the absorbance versus time traces to the equation for a single exponential. In all cases, the absorbance versus time traces represented the average of three consecutive experiments.
Redox Titrations-Potentiometric redox titrations were performed essentially as described previously (31) in 50 mM Tricine buffer, pH 8.0, with 150 mM NaCl and a series of redox mediators. For titrations of the Cp2 [4Fe-4S] 2ϩ/ϩ couple, the mediators included a 50 M concentration each of flavin mononucleotide (E m ϭ Ϫ172 mV and Ϫ238 mV), benzyl viologen (E m ϭ Ϫ361 mV), methyl viologen (E m ϭ Ϫ440 mV), and N,NЈ-propane-2,2Ј-dipyridinium (E m ϭ Ϫ590 mV; a gift from Dr. Ver-non Parker, Utah State University). For titrations of the Av1 P 2ϩ/N couple, the mediators included a 100 M concentration each of flavin mononucleotide, benzyl viologen, and methyl viologen. The reduction potential of the mediator and protein solution was adjusted by the addition of a 5 mM dithionite (Na 2 S 2 O 4 ) solution or a 25 mM oxidized indigo disulfonate solution. For titrations below Ϫ500 mV, reduction was accomplished in a standard H-cell using a Hewlett-Packard 6212C constant power supply with a gold wire working electrode and a platinum mesh counter electrode (2 ϫ 1 cm). In all cases, the reference electrode was an Ag/AgCl microelectrode that was calibrated against a standard calomel electrode. All potentials are reported relative to the normal hydrogen electrode. At defined potentials, 250-l aliquots were removed and were frozen in calibrated quartz EPR tubes (Wilmad, Buena, NJ). The relative concentrations of the reduced and oxidized states for each metal center were determined by the peak-to-peak height of the appropriate EPR signal. Plots of the fraction of maximum signal intensity versus potential were fit to the Nernst equation (Equation 2) using the nonlinear, least squares fitting program Igor Pro (Wavemetrics, Lake Oswego, OR) where E is the measured potential, E m is the midpoint potential, R is the gas constant, T is the temperature, n is the number of electrons transferred, F is Faraday's constant, [red] is the concentration of the reduced species, and [ox] is the concentration of the oxidized species (32).
For titrations of the Cp2 [4Fe-4S] 2ϩ/ϩ couple, EPR spectra were recorded at 12 K, with a 9.64-GHz microwave frequency, a 10-mW microwave power, a 100-kHz modulation frequency, and a modulation amplitude of 5.028 G. For titrations of the Av1 P 2ϩ/N couple, EPR spectra were recorded at 12 K, with a 9.38-GHz microwave frequency, a 50-mW microwave power, a 100-kHz modulation frequency, and a modulation amplitude of 12.63 G. Midpoint potentials have an estimated error of Ϯ 10 mV.

Electron Transfer in the Absence of Nucleotides Monitored by UV-visible Absorption-Electron transfer from the Fe protein to
the MoFe protein can be monitored by the increase in the electronic absorption spectrum in the 430 nm region as a result of the oxidation of the Fe protein [4Fe-4S] cluster from the 1ϩ to the 2ϩ oxidation state (30). When dithionite-reduced Cp2 ([4Fe-4S] ϩ ) was combined with dithionite-reduced Av1 in the absence of any nucleotides, an increase in the absorbance in the 430 nm region was observed over 10 min as shown in Fig. 1A. This is consistent with the oxidation of the [4Fe-4S] cluster of Cp2 to the 2ϩ state resulting from electron transfer to the MoFe protein in the absence of nucleotides. In contrast, when A. vinelandii Fe protein (Av2) was mixed with Av1 under identical conditions, no changes were observed in the 430 nm region of the absorbance spectrum even after a 10-min incubation (Fig. 1B, traces 1 and 2). Only upon the addition of MgATP to the Av2⅐Av1-containing solution did the absorbance in the 430 nm region increase (Fig. 1B, trace 5), consistent with the dependence of electron transfer on MgATP for the homologous complex.
Electron Transfer Monitored by EPR Spectroscopy-To confirm that the increase in absorbance observed by UV-visible absorption spectroscopy upon incubation of Cp2 with Av1 was caused by the oxidation of the [4Fe-4S] cluster of Cp2, EPR spectroscopy was used to monitor the oxidation state of the [4Fe-4S] cluster. The Cp2 [4Fe-4S] ϩ cluster is an S ϭ 1/2 system with a rhombic EPR spectrum centered at g ϭ 1.93 (Fig.  2, trace 2) (33)(34)(35). Upon oxidation to [4Fe-4S] 2ϩ , the cluster goes to an S ϭ 0 state and is EPR-silent (36). Dithionitereduced Av1 demonstrates low field EPR signals (g ϭ 4.3 and 3.6) (not shown) and a signal at g ϭ 2.0 (Fig. 2, trace 1), which arise from FeMoco (37)(38)(39). The Fe protein spectrum can be added to the MoFe protein spectrum to give a predicted spectrum that would arise from a nonreacting mixture of Cp2 and Av1 (Fig. 2, trace 3) in dithionite. An EPR spectrum was recorded for the mixture of Cp2 and Av1 after 1 h of incubation (Fig. 2, trace 4). As shown, the intensity of the reduced Cp2 signal centered at g ϭ 1.93 decreased, consistent with the one-electron oxidation of the [4Fe-4S] cluster of Cp2 to the EPR-silent 2ϩ oxidation state. This result confirms the transfer of an electron from Cp2 supposed from the changes in the absorbance spectra.
No changes were seen in the intensity or line shapes of the EPR signals attributed to FeMoco of Av1 after incubation with Cp2. Likewise, EPR signals ascribed to the P clusters did not appear upon incubation of Av1 with Cp2. To confirm that Cp2 could transfer an electron to the P cluster, which is the putative electron acceptor, changes in the EPR signals of more oxidized states of the P clusters were monitored. One-electron oxidation of the P cluster from the P N state results in an S ϭ 1/2 and 5/2 mixed spin state (designated as P ϩ ) that gives rise to perpendicular mode EPR signals in the g ϭ 2 and g ϭ 5 regions (40). Two-electron oxidation of the P cluster from the P N state results in an S Ն 3 spin system (designated as P 2ϩ ) that gives rise to a parallel mode EPR signal at g ϭ 11.8 (28). Therefore, it is possible to use EPR to monitor electron transfer from the Fe protein to the P cluster of the MoFe protein by monitoring the reduction of the P 2ϩ oxidation state to the P ϩ oxidation state (41). Fig. 3 (trace 1) shows the perpendicular mode EPR spec-trum of Av1 oxidized by four electrons (each P cluster oxidized by two electrons to the P 2ϩ state). The existence of the P 2ϩ oxidation state was evident by a parallel mode EPR signal at g ϭ 11.8 (data not shown). The predicted EPR spectrum (obtained from the sum of spectra 1 and 2) for the nonreacting mixture of Av1 in the P 2ϩ oxidation state and reduced Cp2 is shown in Fig. 3 (trace 3). When the two proteins were actually allowed to react, however, the spectrum shown in Fig. 3, trace 4, was obtained. The disappearance of the g ϭ 1.93 signal ascribed to Cp2 and the appearance of signals with g values at 5.28, 2.04, 1.94, and 1.80 are consistent with the oxidation of the [4Fe-4S] ϩ cluster to the 2ϩ state and the reduction of each P cluster from the P 2ϩ to the P ϩ state. In addition, the parallel mode EPR signal for P 2ϩ disappeared after Cp2 and Av1 were mixed (data not shown). In control incubations, when reduced, dithionite-free Av2 was incubated with oxidized Av1 (P 2ϩ state) in the absence of MgATP, no changes in the EPR spectrum were observed, indicative of a lack of electron transfer in the absence of nucleotides for the Av2-Av1 mixture.
Rates of Primary Electron Transfer from Cp2 to Av1-To obtain kinetic constants for the observed electron transfer from Cp2 to Av1, the changes in the electronic spectrum at 430 nm were monitored by stopped-flow spectroscopy upon mixing reduced Cp2 with Av1 in the absence of nucleotides. A first-order increase in the absorbance at 430 nm was recorded with an apparent first-order rate constant of 0.007 s Ϫ1 (Fig. 4A, trace 1). When MgADP was included, again an apparent first-order increase in the absorbance was observed, with an apparent first-order rate constant of 0.018 s Ϫ1 (Fig. 4A, trace 2). The addition of MgATP significantly increased the rate of oxidation of Cp2 as evident in Fig. 4B. An apparent first-order rate constant for electron transfer in the presence of MgATP was measured to be 100 s Ϫ1 . This apparent rate constant for pri-

FIG. 2. Electron transfer from the Cp2 [4Fe-4S] cluster to Av1
monitored by EPR spectroscopy. EPR spectra are shown for the as-isolated state of Av1 (trace 1), the reduced state of Cp2 (trace 2), the mathematically additive spectrum of traces 1 and 2 (trace 3), and the mixture after reaction of Cp2 and Av1 (trace 4). In all cases, the Cp2 concentration was 100 M, the Av1 concentration was 75 M, the buffer was 100 mM HEPES, pH 7.4, with 2 mM dithionite, and the samples were incubated for 1 h prior to freezing in liquid nitrogen. All EPR spectra were recorded at 12 K, with a 9.64-GHz microwave frequency, a 6.36-mW microwave power, a 100-kHz modulation frequency, and a modulation amplitude of 5.028 G. mary electron transfer from Cp2 to Av1 in the presence of MgATP represents more than a 10 4 -fold increase from the nucleotide free condition and approaches that observed for the MgATP-dependent rate constants found for the homologous A. vinelandii nitrogenase complex (42).
Redox Titrations-To gain insights into how nucleotides might accelerate the rate of electron transfer in the Cp2⅐Av1 complex, the midpoint potentials (E m ) of the [4Fe-4S] 2ϩ/ϩ couple of Cp2 and the P 2ϩ/N couple of Av1 were determined for the individual proteins alone and for the two proteins when complexed together. According to electron transfer theory (43), the difference between the E m values (or the free energy change ⌬G) for the electron donor and the electron acceptor will affect the rate of electron transfer (44). E m values for the metal centers were also determined in the presence of MgADP. Similar experiments in the presence of MgATP are not possible because of MgATP hydrolysis by this nitrogenase complex. Fig.  5A presents redox titrations for the [4Fe-4S] 2ϩ/ϩ cluster couple of Cp2 either in the uncomplexed state or when complexed with Av1. From fits of the data to the Nernst equation (Equation 1), E m values of Ϫ300 and Ϫ510 mV were determined in the absence of nucleotides for the uncomplexed and complexed states, respectively. When MgADP was included, the E m for the [4Fe-4S] 2ϩ/ϩ couple of free Cp2 was shifted to Ϫ380 mV, consistent with shifts seen earlier (45). When MgADP was added to Cp2 complexed with Av1, the calculated E m of the [4Fe-4S] 2ϩ/ϩ cluster couple was Ϫ510 mV, the same as the value calculated for the complexed Cp2 couple in the absence of nucleotides. Fig. 5B presents redox titrations for the P 2ϩ/N couple of Av1 in the uncomplexed state and when complexed to Cp2, either in the absence of nucleotides or in the presence of MgADP. Within the error of the experiment, no changes in the E m for the P 2ϩ/N couple were observed as a result of complex formation with Cp2 (approximately Ϫ310 mV). Likewise, the addition of MgADP to the Cp2⅐Av1 complex did not result in detectable changes in the E m for the P clusters.

DISCUSSION
The recent x-ray structures for two nitrogenase Fe protein-MoFe protein complexes have provided important insights into the interactions between these two proteins (46,47). However, some major questions remain, including how the affinity between the two component proteins is changed during the catalytic cycle, how MgATP binding and hydrolysis are involved in this process, and how MgATP is coupled to intercomponent electron transfer. Recent investigations of the three known tight nitrogenase complexes, namely the Cp2⅐Av1 complex (18), the MgADP⅐AlF 4 Ϫ -stabilized Av2⅐Av1 complex (48), and the L127⌬ Av2⅐Av1complex (49,50), have begun to provide some insights into these outstanding questions. For example, the earlier observation that MgATP could be hydrolyzed at considerable rates by the Cp2⅐Av1 complex (18) 3. Electron transfer from the Cp2 [4Fe-4S] cluster to the Av1 P cluster monitored by EPR spectroscopy. Av1, with each P cluster oxidized by two electrons (P 2ϩ state), and reduced, dithionitefree Cp2 were prepared as outlined under "Experimental Procedures." Perpendicular mode EPR spectra are shown for the oxidized (P 2ϩ ) state of Av1 (trace 1), the reduced state of Cp2 (trace 2), the mathematically additive spectrum of traces 1 and 2 (trace 3), and the mixture after reaction of Cp2 and the P 2ϩ state of Av1 (trace 4). In all cases, the Cp2 concentration was 70 M, the Av1 concentration was 33 M, the buffer was 50 mM MOPS, pH 7.0, and the samples were incubated for 3 min prior to freezing in liquid nitrogen. All EPR spectra were recorded at 12 K, with a 9.65-GHz microwave frequency, a 10-mW microwave power, a 100-kHz modulation frequency, and a modulation amplitude of 7.969 G. a, resonances assigned to the P ϩ state of Av1; b, resonances assigned to the as-isolated, dithionite-reduced state of FeMoco of Av1. revealed that primary electron transfer can occur within the complex without MgATP hydrolysis. It was concluded, contrary to longstanding models, that MgATP hydrolysis was not absolutely required for electron transfer between the proteins. The present results for the Cp2⅐Av1 nitrogenase complex further define a role for MgATP in the electron transfer process by revealing that one important function is to accelerate the primary electron transfer rate.
It is of interest to consider how nucleotides might function to accelerate primary electron transfer within the complex. Perhaps the simplest explanation would be that nucleotides change the thermodynamic driving force for the electron transfer reaction. This could be accomplished by nucleotide-induced changes in the E m values for either the Fe protein [4Fe-4S] cluster or the MoFe protein P cluster, thereby increasing the difference between the two E m values (⌬E m ). This explanation is tempting considering earlier work that demonstrated that MgATP or MgADP binding to the free Fe protein results in about a Ϫ120 mV shift in the E m for the [4Fe-4S] 2ϩ/ϩ couple (10,45). More recently, it has also been shown that association of the L127⌬ Fe protein with the MoFe protein to form a tight complex results in shifts in the E m values of two of the three nitrogenase metal centers to favor electron transfer (50). The E m for the L127⌬ Fe protein [4Fe-4S] 2ϩ/ϩ couple was observed to shift by more than Ϫ200 mV upon complex formation to about Ϫ600 mV, and the E m for the MoFe protein P 2ϩ/N couple shifted by Ϫ80 mV to Ϫ390 mV. Negative shifts in the E m values of the P cluster and the [4Fe-4S] cluster were also observed in the Av2⅐Av1⅐MgADP⅐AlF 4 Ϫ nitrogenase complex (48). In the present work, the E m values for the metal centers of free Cp2 and free Av1 and for the Cp2⅐Av1 complex were determined. E m values observed for the free proteins were consistent with previously reported values (28,45,51). In the complex, only the E m of the [4Fe-4S] cluster changed (from Ϫ300 to Ϫ510 mV). Unlike the other two complexes described above, no change in the E m of the P cluster was observed in the Cp2⅐Av1 complex. Despite the particular differences that complex formation has on the E m values for each metal cluster in the three tight complexes, a theme that emerges is that the energy of complex formation is coupled to changes in the E m values to favor electron transfer.
Although MgADP is observed to accelerate the electron transfer rate in the Cp2⅐Av1 complex by 2.5-fold, MgADP resulted in no additional changes in the E m values for any of the metal centers within the complex. This observation suggests that the rate acceleration induced by MgADP is not the result of changes in the driving force (i.e. E m values). Although it is not possible to measure the effects of MgATP on the E m values in the complex because of the hydrolysis reaction, electron transfer theory can be used to make the argument that, similar to MgADP, the rate acceleration in electron transfer induced by MgATP cannot be accounted for solely by changes in driving force. It should be noted that the electron transfer reactions in the Cp2⅐Av1 complex may not necessarily follow electron transfer theory, but for the purposes of this argument this assumption will be made. A simplified version of the Marcus equation which has evolved from analysis of many protein electron transfer reactions (44) where k et is the electron transfer rate constant (s Ϫ1 ), R is the distance between the donor and acceptor (Å), ⌬G is the free energy difference (eV), and is the reorganization energy (eV). Using this equation, values for the primary electron transfer rate constants with and without MgATP, a value for R of 14 Å (46), and a range for from 1.8 to 2.4 eV (52,53), it was possible to estimate changes in the free energy (⌬⌬G) that would be required for this rate acceleration. From such a calculation, and using the standard relationship between ⌬G and ⌬E m , a ⌬⌬E m from Ϫ620 to Ϫ720 mV is estimated. Thus, for a change in the driving force to account fully for the observed change in the rate constant for electron transfer, MgATP binding or hydrolysis in the complex would have to change the ⌬E m more than Ϫ600 mV. Such large changes in E m values seem unreasonable given the previously observed effects of MgATP on nitrogenase. In addition, changes in E m in redox sites in proteins of this magnitude would be unprecedented. Whereas it is reasonable to assume that part of the mechanism for the MgATPdependent rate acceleration of electron transfer might involve changes in E m values, this alone is not likely to account for the entire observed rate enhancement. was monitored from the intensity of the S ϭ 1/2 EPR signal. The signal peak-to-peak height was normalized to the maximum height observed at the lowest potential. The solid lines represent nonlinear least squares fits of the data to the Nernst equation where n ϭ 1, and the calculated midpoint potentials were found to be Ϫ300, Ϫ380, Ϫ510, and Ϫ510 mV, respectively. Panel B, electrochemical redox titrations are presented for the P 2ϩ/N couple of Av1 (E), of Av1 complexed to Cp2 (f), and of Av1 complexed to Cp2 in the presence of MgADP (OE). The P 2ϩ oxidation state of the P cluster was monitored by the intensity of the S Ն 3 parallel mode EPR signal. The signal peak height was normalized to the maximum peak height observed at the most positive potential. The solid lines represent nonlinear least squares fits of the data to the Nernst equation with n ϭ 2, and the calculated midpoint potentials were found to be Ϫ310, Ϫ320, and Ϫ300 mV, respectively. Potentials are relative to the normal hydrogen electrode.
Another possible way to increase the rate of electron transfer could be through changes in the pathway(s) for electron transfer. Changes in electron transfer pathways in other proteins have been observed to have large effects on the electron transfer rate (54). Possible electron transfer pathways from the Fe protein [4Fe-4S] cluster to the MoFe protein P cluster can be suggested from the structures of the two nitrogenase complexes (46,47). In the two likely pathways, at least one H-bond (i.e. noncovalent bond) jump is predicted. Such jumps have been suggested to be less efficient in electron transfer (55,56). One possibility is that MgATP binding or hydrolysis in the Cp2⅐Av1 complex could result in changes in the electron transfer pathway, possibly including changes in the distance or orientation of the H-bond jump. In this case, some of the energy associated with MgATP binding or hydrolysis would be coupled to changes in protein conformation within the complex that would alter the electron transfer pathway.
Finally, two additional observations from the present work deserve comment as they provide additional insights into the roles of MgATP in the nitrogenase reaction. First, it is evident that for the Cp2⅐Av1 nitrogenase complex, the acceleration of the electron transfer rate is at least partially uncoupled from the rates of MgATP hydrolysis. Although MgATP addition accelerated the primary electron transfer to rates near those observed for the homologous nitrogenase complexes, the rates of MgATP hydrolysis for the Cp2⅐Av1 complex are 36-fold lower than for the homologous complex (18). It has been known for some time that MgATP hydrolysis can occur in nitrogenase complexes without electron transfer, thus indicating that hydrolysis can be uncoupled from electron transfer. The present results indicate that electron transfer rates can also be uncoupled from MgATP hydrolysis rates. This latter point is supported by the observations of electron transfer without MgATP in other nitrogenase tight complexes (48,49). Finally, an important observation from the present work is that Cp2 is unable to transfer more than one electron into the resting state of the MoFe protein even when MgATP is present. After electron transfer, the [4Fe-4S] cluster of Cp2 can be reduced in the complex, yet no further electron transfer (i.e. oxidation of the [4Fe-4S] cluster) is observed. This would explain why the Cp2⅐Av1 complex is inactive because at least two electrons would be required to reduce even the simplest substrates (57). The reason that the second electron cannot be transferred in the Cp2⅐Av1 complex is not clear. It would appear from the current results that the limitation is not the hydrolysis of MgATP because MgATP hydrolysis continues in the Cp2⅐Av1 complex before and after the primary electron transfer event. Instead, the lack of transfer of the second electron into the MoFe protein may have to do with the proper coupling of MgATP binding or hydrolysis to changes in the Fe protein or MoFe protein structures, or the oxidation state of the electron acceptors in the MoFe protein. The lower rates of MgATP hydrolysis by the Cp2⅐Av1 complex may also account for the lack of transfer of a second electron. Additionally, dissociation of the Fe protein from the MoFe protein, which does not occur readily in the Cp2⅐Av1 complex, may somehow be required for subsequent electron transfer reactions. Studies to begin to understand these open questions are currently under way.