Dissecting Electronic-Structural Transitions in the Nitrogenase MoFe Protein P-Cluster during Reduction.

The [8Fe-7S] P-cluster of nitrogenase MoFe protein mediates electron transfer from nitrogenase Fe protein during the catalytic production of ammonia. The P-cluster transitions between three oxidation states, PN, P+, P2+ of which PN↔P+ is critical to electron exchange in the nitrogenase complex during turnover. To dissect the steps in formation of P+ during electron transfer, photochemical reduction of MoFe protein at 231-263 K was used to trap formation of P+ intermediates for analysis by EPR. In complexes with CdS nanocrystals, illumination of MoFe protein led to reduction of the P-cluster P2+ that was coincident with formation of three distinct EPR signals: S = 1/2 axial and rhombic signals, and a high-spin S = 7/2 signal. Under dark annealing the axial and high-spin signal intensities declined, which coincided with an increase in the rhombic signal intensity. A fit of the time-dependent changes of the axial and high-spin signals to a reaction model demonstrates they are intermediates in the formation of the P-cluster P+ resting state and defines how spin-state transitions are coupled to changes in P-cluster oxidation state in MoFe protein during electron transfer.

Nitrogenase is a two-component enzyme that catalyzes the conversion of dinitrogen to ammonia. Under ideal reaction conditions, the Mo-dependent form of nitrogenase, composed of Fe protein and MoFe protein, catalyzes N2 reduction to ammonia according to eq :[1]

During turnover, the electrons required for N2 reduction are transferred from Fe protein to MoFe protein, which is an α2β2 tetramer that coordinates two sets of unique metal clusters. The [8Fe-7S] P-cluster functions in electron transfer with Fe protein, and the [7Fe-9S-1Mo-C-Homocitrate] iron–molybdenum cofactor (FeMo-co) functions as the site of N2 reduction.[2]

One of the unique aspects of how nitrogenase catalyzes ammonia production is the electron transfer process.[3,4] In the catalytic cycle, the P-cluster forms a metastable intermediate oxidation state, P+ ([7FeIIFeIII-7S]+1), that is rapidly reduced (k > 1700 s–1)[3] during electron transfer. In addition to P+, the P-cluster forms two stable oxidation states, PN ([8FeII-7S]0 and P2+ ([6FeII2FeIII-7S]+2) (Figure ).[5] Transitions between P-cluster states involve extensive structural changes, including a switch in Fe-coordination of the central sulfide (S1) that bridges the two [4Fe-3S] subclusters, and amide nitrogen coordination to Fe5 by α-88 cysteine (α-88Cys) and β-188 serine (β-188Ser) oxygen coordination to Fe6. Recently, the X-ray structure of MoFe protein was solved with the P-cluster poised in the P+ state, with an intermediate structural arrangement between P2+ and PN (Figure ).[6] In the P+ state, the S1 sulfide is pentacoordinate and β-188Ser coordinates Fe6. The observation of structural changes in the MoFe protein P-cluster has been incorporated into conformational gating[7] and mechanical coupling[8,9] electron transfer models. The model predicts that motions near the P-cluster and β-188Ser are coupled to “switch regions” in the Fe protein that steer structural interactions within the nitrogenase complex to enable electron delivery.[9]

Figure 1

MoFe protein P-cluster oxidation state structures for P2+, P+, and PN. The α-88Cys (P2+↔P+) and β-188Ser (P+↔PN) ligands that undergo redox-coupled coordination changes to the P-cluster are shown. The P2+↔P+ transition involves exchange of the S1–Fe5 thiolate bond (gray arrow) and α-88Cys amide bond at Fe5 (blue arrow) and proceeds via proton-coupled electron transfer (PCET), where changes in bonding (box) may lead to different conformers during electron transfer. The P+↔PN transition involves exchange of the S1–Fe6 thiolate bond and β-188Ser serine hydroxylate bond at Fe6 (red arrow). Em = −309 mV at pH 8 for both transitions.[11] PDB Codes: P2+, 2MIN; P+, 6CDK; PN, 3MIN.

The structural rearrangements of the P-cluster in different oxidation states also coincide with changes in spin states and EPR properties. P2+ is an integer spin state, likely S = 4, and gives rise to an EPR signal at g = 11.8,[10,11] whereas PN is an S = 0 spin state and EPR-silent. The P+ oxidation state has a rhombic, S = 1/2 EPR signal at g = 2.05, 1.94, 1.81 that shifts to g = 2.03, 1.97, 1.93 when β-188Ser is substituted by Cys.[12,13] Additional magnetic signals associated with the P+ oxidation state include an S = 1/2 signal with g = 2.00 and 1.89,[13,14] and low-field S = 5/2 signals[12,13] (Table S1). Variations in P+ magnetic states have been observed in MoFe protein under different redox titration conditions (Table S2).[13−15] Whether these states have a functional role in electron transfer in MoFe protein remains unclear. Recently, the structural and magnetic configurations of the P-cluster oxidation states were shown to coincide with profound differences in the density of low lying electronic states, implying there is a deeper relationship between the electronic-structural properties of the P-cluster and its function in electron transfer.[10]

Resolving the relationship between the magnetism and structure of the P-cluster, most notably for the metastable P+ state, is important for elucidating a complete mechanistic understanding of the P-cluster in the nitrogenase electron transfer cycle. Herein, we address this goal by combining light-controlled reduction of MoFe protein in complexes with cadmium sulfide nanocrystals (CdS)[16,17] with EPR to resolve magnetic changes in the P-cluster during electron transfer that arise from discrete electronic-structural intermediates in the reduction of P2+ to P+.

An oxidized sample of nitrogenase MoFe protein was mixed with mercaptopropionic acid capped CdS quantum dots (Figure S1; see Supporting Information for details), and the P-cluster P2+ oxidation state was verified by EPR (Figure S2). The CdS:MoFe protein complexes were illuminated with a 405 nm LED at either 231 K or 263 K and then allowed to anneal in the dark at 236 K or 263 K, respectively, to prevent further light-driven reduction. By illuminating at subambient temperatures, the light-driven redox (i.e., electron transfer) process is decoupled from temperature sensitive chemical (i.e., ligand switching) steps during electron transfer and P-cluster conversion from P2+↔P+. As shown in Figure A, illumination at 263 K for 12 min resulted in reduction of the P-cluster exemplified by loss of P2+ intensity (36%, Figure S2 and Table S1). This change coincided with the appearance of an S = 1/2 rhombic signal at g = 2.05, 1.94, 1.81 (P+1.81, Figure A)[13,15,18] and an S = 1/2 axial signal with g = 2.006, 1.89 (assigned to P+1.89, Figure B).[13,14] Dark annealing at 263 K for 20 min (Figure A, green trace) led to complete loss of the P+1.89 signal, and an increase in amplitude of the P+1.81 (21%) and P2+ (7%) signals (Table S1).

Figure 2

Illumination and EPR spectra of CdS:MoFe protein complexes at 263 K and 231 K. (A) T = 263 K. Blue trace, oxidized CdS:MoFe protein complexes. Red trace, after 12 min illumination at 263 K. Green trace, spectrum after incubation in the dark at 263 K for 20 min. (B) Illuminated (red trace) minus dark (green trace) difference spectrum. Simulation (black dashed trace) using g = 2.006, 1.89 assigned to the S = 1/2, P+1.89 signal. Buffer pH = 7; EPR conditions, T = 12 K, microwave power = 1 mW. (C) T = 231 K. Blue trace, oxidized CdS:MoFe protein complexes. Red trace, after 15.5 min illumination at 231 K. (D) Low-field EPR spectrum showing the high-spin, S = 7/2, g = 6.54 signal assigned as P+6.54. Sample pH = 7. EPR conditions: (C) T = 12 K, microwave power = 1 mW, (D) T = 18 K, microwave power = 25 mW. Populations of EPR signals are summarized in Table S1.

When illuminated at a lower temperature of 231 K for 15.5 min, reduction of the MoFe protein P-cluster led to a decrease in the P2+ signal intensity (19%) and appearance of the P+1.89 signal, whereas formation of the P+1.81 signal was suppressed. Rather, a low-field inflection at g = 6.54 (Figure D) appeared, resembling other high-spin signals observed for MoFe protein (Table S2, Figure S3). Rhombogram analysis and simulation of the g = 6.54 signal (referred to as P+6.54) indicates that it originates from a S = 7/2 spin system with E/D ≈ 0.024 (D = −3.2 cm–1) of the reduced P-cluster, where E and D[19] are the zero-field splitting parameters (see Figure S3 for details).

Dark annealing at 236 K of the CdS:MoFe protein complexes illuminated at 231 K (Figure C) was used to monitor relative intensities of P+ intermediates following light-driven electron transfer to MoFe protein (Figure S4). EPR spectra of CdS or MoFe protein alone, before and after illumination and annealing, or of CdS:MoFe protein prior to illumination, did not produce any detectable signal changes (Figure S5). Simulations of the low-field regions using singular value decomposition (SVD, Figure S6) and the high-field region using EasySpin[20] (Table S3, Figure S7) enabled time-dependent changes in signal intensities of P+ intermediates (P+1.89, and P+1.81 and P+6.54) to be fit to reaction models (Tables S4 and S5, Figure S8). The P+ signal intensity versus annealing time best fit to a three-step reaction model as summarized in eqs and 3:

The fit shown in Figure gave relative values for rate constants where k2 > k1 > k3 and predicts the high-spin P+6.54 P-cluster intermediate originates together with P+1.89 under photochemical reduction of the P2+ state (Figure ). In the dark, P+6.54 partitions rapidly between P+1.89 or P+1.81. Therefore, a lack of P+6.54 under illumination at 263 K (Figure ) is likely due to more rapid conversion to either P+1.81 or P+1.89 (k1 and k2 > k3) than at 231 K.

Figure 3

Time-dependent changes of the P+ EPR signals intensities in CdS:MoFe protein complexes. Initial (time = 0 min) P+ signal intensities were collected at 231 K. Changes are plotted versus time under dark annealing at 236 K. P+ signal intensities were determined using EasySpin and SVD analysis (see Supporting Information, Figures S6 and S7).[20] Solid lines are fits of the experimental data to differential equations; dP+1.89/dt = k[P+6.54] – k[P+1.89], dP+1.81/dt = k[P+6.54] + k[P+1.89], and dP+6.54/dt = −(k + k)[P+6.54] (Table S4). Green, P+6.54; blue, P+1.89; red, P+1.81.

Figure 4

Schematic representation of the P2+ to P+ conversion in low temperature photochemical reduction of the MoFe protein P-cluster. Photoexcitation at 231 K of CdS:MoFe protein complexes poised in P2+ (left) leads to electron injection into the P-cluster and reduction to a mixed population of P+ states; the S = 7/2 P+6.54 and S = 1/2 P+1.89, which are based on the reaction model (Figure ), correspond to distinct conformers (inset). Dark annealing at 236 K results in the conversion of P+6.54 to either P+1.89 (faster) or P+1.81 (slower), and conversion of P+1.89 to P+1.81. The rate constants of the conversion of P+ states are obtained from fits shown in Figure to a reaction model in Table S4.

Dark annealing was performed over a range of 231 K to 245 K to obtain the temperature-dependence of k3 for the P+1.89↔P+1.81 step (eq , Figure S9, Table S6). An Arrhenius plot of ln k3 vs 1/T gave a value of Ea = 24 ± 8.3 kcal mol–1. The value suggests the P+1.89↔P+1.81 involves structural changes in MoFe protein at the P-cluster. For example, reductive formation of P+1.81 from P2+ at 298 K is pH-dependent (Figure ) and is favored at pH ≈ 6 and nearly undetectable at basic pH (>8).[15] Likewise, chemical oxidation of MoFe protein P-cluster from PN↔P+ at 298 K led to formation of both P+1.81 and P+1.89,[13] with P+1.89 intensity being maximal at pH 8.4.[14] The two results are consistent with formation of P+1.89 being reversible and both pH- and temperature-dependent.

In addition to analysis of the P+1.89 intermediate, photochemical reduction of MoFe protein at 231 K also enabled assignment of the P+6.54 high-spin state to a unique electron transfer intermediate (Figure ). In the P2+↔P+ reduction step, the high-spin P+6.54 state has two possible fates: direct conversion to P+1.81 (eq ) where k2 > k3 or rapid conversion to P+1.89 followed by slow P+1.89↔P+1.81 conversion (eq ). Thus, the reaction model for the P-cluster P2+↔P+ conversion, summarized in Figure , involves two spin-state isomer intermediates. The observation of multiple electronic intermediates associated with a redox step in the P-cluster is similar to the observation of low-spin and high-spin S2 states of the PSII oxygen evolving complex that arise from valence isomerism in Mn–O–Mn coordination from different S2 conformers that function in the catalytic cycle of water oxidation.[21−25] The interconversion of P+6.54↔P+1.89 may likewise arise from conformational isomerism in Fe-coordination to S1 (see Figure ) that guide formation of P+ with surrounding structural changes. Overall, the results from combining low temperature photochemical reduction of the MoFe protein with dark annealing reveal that formation of the metastable P+ state, P+1.81, involves intermediate spin states and electronic configurations that occur with changes in P-cluster coordination.

As established by kinetic and theoretical studies, correlated motions within the nitrogenase complex during turnover have an important function in enabling P-cluster mediated electron transfer to be integrated with catalysis.[7−9,26] As shown here, during electron transfer, there are also discrete changes in P-cluster magnetic structure that are linked to changes in oxidation state. The EPR analysis and kinetic model are most consistent with these magnetic states originating from different P-cluster conformers during electron transfer and reduction of P2+ to P+, which may function in the electron transfer mechanism within the nitrogenase complex during ammonia production.[27]