Photoinduced Reductive Elimination of H₂ from the Nitrogenase Dihydride (Janus) State Involves a FeMo-cofactor-H₂ Intermediate

Abstract

N₂ reduction by nitrogenase involves the accumulation of four reducing equivalents at the active site FeMo-cofactor to form a state with two [Fe-H-Fe] bridging hydrides (denoted E₄(4H), the Janus intermediate), and we recently demonstrated that the enzyme is activated to cleave the N≡N triple bond by the reductive elimination (re) of H₂ from this state. We are exploring a photochemical approach to obtaining atomic-level details of the re activation process. We have shown that when E₄(4H) at cryogenic temperatures is subjected to 450 nm irradiation in an EPR cavity, it cleanly undergoes photoinduced re of H₂ to give a reactive doubly-reduced intermediate, denoted E₄(2H)*, which corresponds to the intermediate that would form if thermal dissociative re loss of H₂ preceded N₂ binding. Experiments reported here establish that photoinduced re occurs in two steps. Photolysis of E₄(4H) generates an intermediate state that undergoes subsequent photoinduced conversion to [E₄(2H)* + H₂]. The experiments, supported by DFT calculation, indicate that the trapped intermediate is an H₂ complex on the ground adiabatic potential energy suface that connects E₄(4H) with [E₄(2H)* + H₂]. We suggest this complex, denoted E₄(H_2; 2H), is a thermally populated intermediate in the catalytically central re of H₂ by E₄(4H), and that N₂ reacts with this complex to complete the activated conversion of [E₄(4H) + N₂] into [E₄(2N2H) + H₂].

TOC image

Introduction

In recent years we have demonstrated^1–4 that N₂ reduction by nitrogenase involves the accumulation of four reducing equivalents at the active site FeMo-cofactor (FeMo-co) to form a state with two [Fe-H-Fe] bridging hydrides and two sulfur-bound protons (denoted E₄(4H), the Janus intermediate), and that breaking the N≡N triple bond requires the reductive elimination (re) of H₂ from E₄(4H). This process corresponds to the forward direction of the equilibrium in Fig 1, upper, whose reverse is the oxidative addition (oa) of H₂ with release of N₂.⁵ This re/oa mechanism was initially founded on the identification of a freeze-trapped intermediate of the α-70^Val→Ile substituted MoFe protein as E₄(4H).^6–8 Subsequently, EPR/ENDOR, and photophysical measurements showed that the identical Janus intermediate can be freeze-trapped during N₂ reduction by wild type (WT) MoFe protein, thereby confirming the intermediate in the α-70^Val→Ile variant as a reliable guide to mechanism, and establishing re/oa as a thermoneutral, kinetically accessible equilibrium process.⁴ The overall result of these several findings was to establish that H₂ release drives the reduction of the N≡N triple bond, and thus the mechanistic requirement for the formation of one H₂ per N₂. This in turn implies the limiting stoichiometry of eight electrons/protons for the reduction of N₂ to two NH₃ (eq 1).³

$${\mathrm{N}}_2+8{\mathrm{e}}^{-}+16\text{ATP}+8{\mathrm{H}}^{+}\to 2{\text{NH}}_3+{\mathrm{H}}_2+16\text{ADP}+16{\mathrm{P}}_{\mathrm{i}}$$ (1)

Figure 1.

Representations of the mechanistically central re/oa equilibrium that activates nitrogenase to break the N≡N triple bond (upper), and of the photoinduced re/oa equilibrium (lower). These cartoons represent the Fe 2,3,6,7 face of FeMo-co; but we emphasize they are not meant to be ‘anatomically’ precise.

In the context of this stoichiometry, and the Lowe-Thorneley kinetic scheme that embodies it,^9–11 one sees that E₄(4H) sits at a transition in the N₂ reduction pathway, poised to relax to E₀ through sequential release of two H₂, but equally poised to bind and reduce N₂ through the accumulation of four more equivalents, hence the designation of E₄(4H) as the Janus intermediate.²

As a part of our efforts to (i) more deeply integrate the re/oa mechanism into the organometallic chemistry of hydrides and (ii) obtain atomic-level details of the re activation process we adopted a photochemical approach inspired by the properties of inorganic dihydride complexes.¹² The photolysis of transition metal dihydride complexes, primarily mononuclear centers with cis hydride ligands, commonly results in the release of H₂^13–22 with a decrease in metal-ion oxidation state by two, “a typical example of reductive elimination”, while the thermally induced reverse reaction “is the prototype example of an oxidative addition reaction.”¹⁴ EPR and ENDOR measurements showed that when the nitrogenase E₄(4H) intermediate is subjected to 450 nm irradiation in an EPR cavity, it cleanly undergoes photo-induced re of H₂ to give an intermediate which we denote E₄(2H)*, Figure 1, lower. It is a reactive version of the E₂(2H) state that has accumulated two [e⁻/H⁺], and corresponds to the intermediate that would form upon thermal dissociative re loss of H₂ prior to N₂ binding. The photoinduced re of H₂/D₂ from E₄(4H)/E₄(4D) in H₂O/D₂O is temperature invariant in both buffers in the range of 4 – 12 K, a variation in the thermal energy (proportional to T) by a factor of ~ 3, and the photoinduced decay of Janus intermediate shows a large KIE over this range, defined as the ratio of the median decay times for D₂O and H₂O buffers: KIE ~ 10. These observations together imply that photoinduced release of H₂ through re involves nuclear tunneling through a barrier, in contrast to the barrierless process inferred for mono-metallic metal complexes.^14,18

The photoinduced state E₄(2H)* relaxes to E₄(4H) by the thermally activated oa of H₂ to E₄(2H)* (Figure 1, lower) during annealing of the frozen solid at temperatures above T ≳ 175 K. The observation of a significant KIE plus a strong temperature dependence in the oa process revealed that it involves traversal of an energy barrier associated with H₂ binding and/or bond cleavage. Overall, the measurements led to proposed energy surfaces associated with photoinduced re/oa for the Janus intermediate that will be refined in the present report.

The reported results were consistent with the absence of intermediate state(s) during photoinduced re,²³ but they left open that possibility. As we noted: “whether the process involves an actual intermediate H₂ complex remains to be determined”. Experiments reported here not only establish the presence of an intermediate in the photoinduced re of H₂, but furthermore show that this intermediate undergoes photoinduced re to E₄(2H)* (Fig 1, lower). The experiments, supported by DFT calculation, indicate that the trapped intermediate is an H₂ complex and lead us to suggest that it functions as an intermediate in the thermal re of H₂ that activates FeMo-co to break the N≡N triple bond.

Materials and Methods

Materials and Protein purifications

All the reagents were obtained from SigmaAldrich (St. Louis, MO) or Fisher Scientific (Fair Lawn, NJ) and were used without further purification. Argon and N₂ were purchased from Air Liquide America Specialty Gases LLC (Plumsteadville, PA).

Remodeling the active site of MoFe protein by the α-70^Val→Ile mutation permits the freeze trapping of MoFe with high populations of E₄(4H)⁶ identical to that in WT enzyme.⁴ The protein was obtained from the corresponding Azotobacter vinelandii strains as described elsewhere.²⁴ The handling of all buffers and proteins were done anaerobically under Ar atmosphere or under Schlenk vacuum line unless stated otherwise.

EPR samples

The E₄(4H) intermediate and its deuterated analogue, E₄(4D), are prepared by turnover of the MoFe protein in H₂O and D₂O buffers, respectively. EPR samples were prepared in a dioxygen free buffer and freeze-trapped during turnover as described.¹²

EPR Measurements

X-band EPR spectra were recorded on a Bruker ESP 300 spectrometer equipped with an Oxford Instruments ESR 900 continuous He flow cryostat. All spectra were measured at the following conditions: temperature, 12 K; microwave frequency, 9.36 GHz; microwave power, 10 mW; modulation amplitude, 9 G; time constant, 160 ms; field sweep speed, 20 G/s. In situ photolysis of a sample held within the cryostat employed a Thorlabs Inc (Newton, New Jersey) PL450B, 450 nm, 80 mW Osram Laser Diode mounted on the cavity optical access port as described.¹² Light intensity was varied through controlling the laser current.²⁵ Data points for observed photolysis progress curves were obtained as well resolved g₁ amplitudes of corresponding EPR signals, with normalization of total concentration of the intermediate species to unity.

Results and Discussion

EPR Spectra and Low-temperature Photoinduced re

As recently shown,¹² intra cavity photolysis with 450 nm light of the S = ½ E₄(4H) intermediate freeze-trapped during turnover of MoFe protein causes photoinduced conversion of the E₄(4H) signal, g = [2.15, 2.007, 1.965], to the E₄(2H)* signal with g = [2.098, 2.0, 1.956], and E₄(2H)* completely relaxes back to E₄(4H) by oa of H₂ with annealing in the solid state at temperatures above ~ 175 K.

To simultaneously obtain the time course for both the photoinduced loss of the E₄(4H) (A) and the appearance of the E₄(2H)* (C) signals, rather than continuously monitoring the occupancy of E₄(4H) during photolysis, we collected complete spectra at multiple times during 12 K and 50 K irradiation and measured the intensities of the two species as a function of time, Fig 2. If photoexcitation of A led directly to the A (E₄(4H)) → C (E₄(2H)*) conversion in a single kinetic step without buildup of a spectroscopic intermediate state, then at every time the C occupancy should equal the loss of the A occupancy, (1-A) as normalized to the value A at t = 0, and the rate parameters for the rise of C should equal those for the loss of A. However, as shown in Fig 2 (and see below), at low temperatures the rise of C lags the loss of A, which implies that photoexcitation of A creates an intermediate, B, which accumulates during the course of its conversion to C. The existence of such an on-path state B, is supported below in the characterization of B, kinetic fits, and in considerations of a potential-energy surfaces that captures the reversible [photoinduced-re/thermal-oa] process of Fig 1 lower.

Figure 2.

Comparison of normalized 12 K and 50 K time courses of E₄(4H) (red) and E₄(2H)* (black) states during irradiation of α-70^Val→Ile H₂O turnover sample with 450 nm laser diode light. Solid lines are to guide the eye; dotted lines show the lag in E₄(2H)* formation at 12 K.

Photolysis Intermediate

To observe the B intermediate directly we carried out a variant of the photolysis cryoannealing experiment. E₄(4H) was photolyzed for a fixed interval (25 min) at successively higher temperatures, then cooled back to 12 K for examination. In each step the sample was then annealed at 77 K for 2 min then restored to 12 K for another round of EPR examination. After such a process following photolysis at a given temperature, the sample was annealed at ~ 215 K, at which temperature E₄(2H)* relaxes to E₄(4H) within 2 minutes, thus restoring the sample to its initial state for photolysis at the next temperature (Fig 3).

Figure 3.

Results of 77 K annealing of α-70^Val→Ile turnover irradiated at various temperatures. (*)-labeled features of EPR spectra indicate recovery of dihydride intermediate state E₄(4H) upon annealing (left) in parallel with disappearance of photoinduced species (right).

Consider the results for this annealing process as applied to the samples photolyzed at 12 K (Figs 3, 4). Photolysis has converted the majority of E₄(4H) to E₄(2H)*, which does not relax during 77 K annealing. However, this annealing causes a small recovery of the E₄(4H) signal without increase in the E₄(2H)* signal. This implies another state, B, has been formed by photolysis, that B is stable in the dark at temperatures, T ≲ 30 K, and that it relaxes to E₄(4H) by 77 K, even though E₄(2H)* does not relax at this temperature. Essentially the same recovery of E₄(4H) upon annealing is observed with 20 K photolysis, but by photolysis at 30 K, the recovery upon 77 K annealing is clearly diminished, implying a lesser accumulation of B. When photolysis is carried out above 30 K, there is no recovery of E₄(4H) with 77 K annealing, indicating that B does relax at 30 K and above.

Figure 4.

Subtraction-elimination of E₄(4H) and E₄(2H)* signals from EPR spectrum of α-70^Val→Ile turnover irradiated at 12 K (red) with use of combination of spectra recorded before irradiation (black) and after 77 K annealing, which followed the irradiation (green). Result presents otherwise unresolved new photoinduced signal B (blue) unstable at 77K.

The measurements in Figs 3 showed the way to acquiring the spectrum of B through subtraction. The use of three spectra – before irradiation, after irradiation, after 77 K annealing – allows a subtraction-elimination of signals from all species except B. The procedure is shown in Fig S1, and as can be seen in Fig 4, reveals an S = ½ signal for B with g ~ [2.02, 2.00, 1.97]. Despite the need for this cumbersome subtraction procedure, the time course of B could actually be followed during photolysis at 12 K, as well as those of the initial A = E₄(4H) and final C = E₄(2H)* states to which it is kinetically linked by the kinetic scheme of eq 2.

The finding that B persists indefinitely (hours) at low temperature (T < 30 K) in the absence of the actinic light, yet converts to C during photolysis at these temperatures implies that both the A→ B and B → C conversions are photoinduced: the photoconversion of A to C thus involves two steps, each of which requires the absorption of a photon, as incorporated in the scheme of eq 2:

$$\mathbf{A}\stackrel{hv}{\to }\mathbf{B}\stackrel{hv}{\to }\mathbf{C}$$ (2)

Though surprising, this result nonetheless can be analogized to the textbook photocleavage of a methyl C-H bond during photolysis of toluene, which likewise requires two consecutive photons.^26–28

This finding does, however, require that we revisit our initial assignment of the photolysis product C = E₄(2H)*, Fig 1, lower. In that earlier report, in addition to re mechanism of photoinduced re of H₂ to yield C = E₄(2H)*, now shown to be a 2-step process, eq 2, Fig 5, upper, we considered two other 2-step mechanisms that might eliminate the hydride bridges of E₄(4H), each mechanism having a different assignment of the photoproduct, C, Fig 5.

Fig 5. Alternative 2-step mechanisms for photoinduced reaction.

(Top) re; assignment of intermediate discussed below; (middle, bottom) alternative mechanisms, see text.

One of those alternative interpretations postulates two steps of photoinduced hydride protonolysis (hp; Fig 5, middle), each with loss of H₂. This mechanism can be dismissed, for it would not lead to a photoproduct C that can regenerate E₄(4H) during cryoannealing, as observed. Instead it would lead to FeMo-co at the E₀ resting state level, which could not react with two released H₂ to regenerate E₄(4H) during cryoannealing: neither E₀ nor E₂(2H) react with H₂.²⁹

The second alternative shown in Fig 5, lower posits two steps of photoinduced re of a single hydride, each with transfer of the proton ‘released’ to a bridging sulfide. Instead of resulting in the release of H₂, the final result of such a process would be creation of an E₄(4H) ‘isomer’ with four protonated sulfur and no hydrides. Simple considerations of the expected relative stabilities of the experimentally observed B and C states rule out this mechanism. B is dark-stable at low temperatures and reverts to E₄(4H) at T > 30 K, while C is more stable than B, relaxing to A only at T > 175 K. The greater stability of C is understandable within the re mechanism. The product state, C = E₄(2H)*, undergoes activated ‘second-order’ binding of H₂ on the ground potential-energy surface in the frozen matrix, to B as a higher-energy intermediate, which undergoes oa of H₂ to generate the bridging hydrides of A = E₄(4H). In contrast, in the L mechanism B is the destabilized E₄(L₁) ‘isomer’ of E₄(4H) (Fig 5), and C would be a yet more unstable E₄(L₂) isomer. But this would be expected to relax to E₄(L₁) at a temperature lower than T ~ 30 K where B reverts to A, not at a temperature ~145 K higher.

These qualitative ideas are supported by exploratory DFT calculations (SI) on a realistic model of E₄ and its local environment (Fig S4), which show that any structure with more than two protonated sulfur atoms is very high in energy relative to E₄(2H)* and other relevant E₄-level states, with E₄(L₂) much higher in energy than E₄(L₁) (Fig S5). Thus, the greater stability of C than B is indeed incompatible with the L mechanism.

Potential-Energy Landscape

The above results can be captured by several alternative heuristic potential-energy landscapes. The most attractive of such qualitative conceptualizations is shown in Fig 6, which represents a projection of multi-dimensional potential energy surfaces that incorporate the experimental observations, along with features of re/oa of H₂ associated with conventional inorganic complexes.

Figure 6.

Idealized energy landscape for the low-temperature photoinduced re/oa of the Janus intermediate, E₄(4H). Excited-state surface: the proposed identifications of the states A†, B, and B† are discussed in the text. Yellow arrows, photoexcitation; black arrows, relaxation, red arrow, nuclear tunneling.

The low-temperature steady-state photoinduced A → C re of H₂ on the landscape of Fig 6 is visualized as following a pathway in which photoexcitation of A produces A^†, which tunnels through a barrier to form an excited H₂ complex, B^†, which undergoes rapid relaxation to the observed S = ½, intermediate B on the ground surface. The B → C conversion then occurs by photoexcitation of B to B^†, which undergoes tunneling/barrier crossing on its way to the re of H₂ in state C; in the absence of direct evidence for nuclear tunneling by B^†, tunneling is not indicated.

Although we as yet have no direct evidence as to the structure of the observed B intermediate (Fig 4), as visualized in Fig 6 the behavior of inorganic H₂ complexes^30–32 indicates that this intermediate is an H₂-bound state on the ground adiabatic potential-energy surface, a possibility first noted thirty years ago.³³ Firstly, thermally induced re/oa of inorganic H₂ complexes generally proceed through such an intermediate,²³ indicating an expectation that such a state is present on the ground adiabatic surface as shown in Fig 6. Secondly, the assignment of B as such a complex is supported by the finding that the B → C re of H₂ is photoinduced: such behavior has been established for inorganic-H₂ complexes.^13,34–37 In support of this assignment, the DFT calculations (Fig S5 suggest that A = E₄(4H) and C = E₄(2H)* (+ H₂) are the most stable E₄ states and are very close in energy, and that they likely are connected by B = E₄(H_2; 2H) at a slightly higher energy.

As a final note, in SI support is provided for the assignment of B as on-path. There we show that an alternative assignment of the observed S = ½ intermediate, Figs 3, 4, to an off-path state, D, Fig S3 is inconsistent with experiment. It is shown that in general, if an off-path intermediate that both forms and decays photochemically were to be observed, then it would have to steadily accumulate during low-temperature photolysis, rather than approach a steady-state population as seen in detailed kinetic measurements presented below.

Kinetic Scheme

In the context of the landscape of Fig 6, the finding that the photoconversion of A to C involves the dark-stable intermediate B and the absorption of two photons consecutively, with B reverting solely to A for T > 30 K, implies the minimum kinetics of Scheme 1. During steady-state photoexcitation of A = E₄(4H), one photon forms the excited state, A^† = E₄(4H)^†. This can decay back to A or tunnel through the energy barrier to form B^†, without return at low temperature (k_b ≈ 0). This excited state branches: it can proceed directly with the re of H₂ to form C, but, primarily decays to B (k_d’ > k_B†C). B is dark-stable at low temperatures (k_BA ≈ 0), so would accumulate, but reaches a steady-state population because photoexcitation of B re-forms B^†, and cycling through this process ultimately allows A to convert entirely to C during prolonged steady-state photolysis. At higher temperatures, T ≳ 30 K, B is no longer dark-stable, which implies that the A^†⥨ B^† reaction becomes reversible and/or direct reversion of B to A becomes operative: [k_b and/or k_BA] > 0. At temperatures, below ~ 175 K, C is indefinitely stable, k_CB ≈ 0. Upon warming the frozen solid to T ≳ 175 K, C relaxes to A, [k_CB, k_BA] > 0, namely E₄(2H)* undergoes oa of H₂ to regenerate E₄(4H), passing through the H₂ intermediate B on the ground adiabatic surface of Fig 6.

Scheme 1.

Considering in more detail the predictions of Scheme 1 in the steady-state photolysis measurements for temperatures T < 175 K, where C accumulates (k_CB ≈ 0), the rate constant for formation of the photoexcited A^† state is the product of the temperature-independent quantum yield, ϕ_A (taken to include the extinction coefficient) with the light intensity, I₀, and k_d is the rate constant for direct relaxation of A^† back to ground. It can be presumed that k_d is by far the largest rate constant in this scheme, and as a result, A^† will enter a photostationary state during the intra-cavity photolysis, with instantaneous population proportional to the instantaneous concentration of A, with a temperature-independent proportionality constant determined by the photophysical constants,

$$A^{†}=\frac{{\varphi }_A{I}_o}{k_d}A$$ (3)

and analogously for B^†. As a result, Scheme 1 maps onto the far-simpler Scheme 2, an elaboration of eq 2, with a two photoinduced reaction steps involving observable states, but with provision for the first step to become reversible at higher temperatures. The net rate constant for the A → B conversion in Scheme 2 is given by

$$k_1=\frac{{\varphi }_A{I}_o}{k_d}{k}_A^{†}$$ (4)

where k_A† is the (tunneling) rate constant for A^† → B^† conversion. The rate constant for the second step, k₂, takes an analogous form; we show shortly that this is augmented by an activated, thermally induced contribution.

Scheme 2.

Kinetic Measurements

Progress curves for photolysis at 12 K and 50 K were collected for the frozen sample in H₂O buffer, Fig 7, and D₂O buffer, Fig S2. A fit to Scheme 2 using stretched exponentials, a decay function written, exp(-(t/τ)^m), with median time constant, τ and ‘stretch parameter’, 0 < m ≤ 1, which decreases with deviations from simple exponential (m = 1). This description is needed to incorporate variation in I₀ over the sample tube caused by light-scattering by the ‘frozen snow’ samples,¹² and yields rate parameters listed on Fig 7. As shown in this figure, at 12 K the occupancy of B quickly rises to a value of roughly 20% occupancy that persists over the time of the experiment. Within the errors of determining the occupancy of intermediate B by the subtraction procedure described above, and of using stretched exponentials to model the distribution in apparent rate constants caused by light scattering, the measured time course of B at 12 K is in satisfactory agreement with the time course predicted by an ‘omit-fit’, namely a fit of the progress curves of A and C alone to eq 2, the low-temperature limit of the more elaborate Scheme 2.

Figure 7.

Normalized time courses of E₄(4H) (red) and E₄(2H)* (black) states measured during photolysis at 12 K (upper) and 50 K (lower) are shown fitted as previously described⁷ in accordance with corresponding low and high temperature kinetic schemes. Data points for photoinduced intermediate B at 12 K were obtained with subtraction procedure described in Figs 4, S1, and are shown in comparison with the kinetic scheme prediction (blue trace). Low accumulation of B state at 50 K is due to back reaction step with time constant τ₃ = 160 sec.

In discussing the stretched-exponential rate parameters that describe the progress curves in Fig 7, there are different possible ways to define the correspondence of a rate constant defined in Scheme 2 with the stretched-exponential time-constant τ and ‘stretch parameter, 0 < m ≤ 1³⁸ For present purposes it is adequate to discuss the results as though rate constant and time constant are inverses: k ~ 1/τ.

Considering the 12 K measurements, B is stable at this temperature, which implies that at this temperature the rate constants for B^† → A^† and direct B → A reversion are vanishingly small, k₋₁ ~ 0. As a result, the photoinduced decay of A as measured by steady-state intracavity photolysis corresponds directly to k₁ (eq 4). The fit to the present measurements at 12 K shown in Fig 7 confirm the previously reported low-temperature kinetic parameters for the A → B as well as the large isotope effect associated with k₁, KIE ~ 13 (Fig S2). It was shown in the earlier measurements from 3.8 – 12 K that this process exhibits A^† → B tunneling associated with the factor, k_A†, that is incorporated into k₁ (eq 4). Importantly, experiments in which the intensity of the actinic light is varied we find that k₁ exhibits the expected linear dependence, eq 4. In the fits to Scheme 2 the time constant for A → B is unchanged by warming from 12 → 50 K (Fig 7), consistent with a persistence of the A^† → B tunneling and continuation of the temperature invariance of the tunneling rate constant previously seen between 4 and 12K.¹² However, the overall reversion of B to A (k₋₁) has become significant by 50 K.

The rate constant, k₂, for the photoinduced B → C conversion is much greater than k₁ at 12 and 50 K (Fig 7). As with k₁, at 12 K it exhibits the expected linear dependence on actinic light intensity and a large isotope effect, KIE ≳ 10. The significant KIE plus an observed increase with increasing temperature dependence (Fig 7) is consistent with the B^† → C with loss of H₂ involving traversal of or tunneling through an energy barrier, rather than excitation of B to a dissociative state.

On the Structure of A^†, and the mechanism of A^†→B conversion

Regarding the nature of the excited state A^†, as previously,¹² we suggest that photoinduced re is initiated by excitation of E₄(4H) into a (σ*) antibonding ligand-field excited state of one of the two bridging hydrides, as in the orbital sketch of Fig 8, which leads to the cartoon representation of A^† in Fig 6. We suggest that this anti-bonding state could undergo reductive elimination, liberating a proton that tunnels the short distance to the second hydride bridge and protonates that hydride, thereby generating the intermediate B^†, the excited state of H₂-bound FeMo-co (B).

Figure 8.

Antibonding excited state of bridging hydride.

Conclusions

This report describes a detailed study of the photoinduced re of H₂ by E₄(4H) to form the activated E₄(2H)* state, and the thermal regeneration of E₄(4H) by the oa of H₂ to E₄(2H)*, Fig 1, lower. We have examined the progress curves for loss of E₄(4H) and appearance of E₄(2H)* at 12 K and 50 K with samples prepared in H₂O and D₂O buffer. The key advance of these measurements is the discovery and characterization of a photogenerated intermediate, B, whose properties assign it as an H₂ complex on the pathway for thermal oa of H₂ by E₄(2H)*, as visualized in Fig 6. This state accumulates during 12 K photolysis, and is dark-stable at this temperature, but can undergo photoinduced re to release H₂.

The present study furthers our effort to obtain atomic-level details of the central mechanistic step in nitrogen fixation, re of H₂ by E₄(4H) coupled to the binding/reduction of N₂, by which the nitrogenase MoFe protein is activated to carry out one of the most challenging chemical transformation in biology, the reduction of the N≡N triple bond. In the present work we extend the use of a photochemical approach inspired by the photoinduced re of H₂ by inorganic dihydride complexes. Our initial report showed that E₄(4H) exhibits the photoinduced re/oa equilibrium of Fig 1,¹² and we subsequently showed that the E₄(4H) in the α-70^Val→Ile variant and the WT enzyme are identical in properties and behavior.⁴

The initial study of E₄(4H) photolysis¹² did not detect an intermediate state during photoinduced re of H₂ from E₄(4H). However, the more detailed experiments reported here now establish the presence of an intermediate, and furthermore show that this photogenerated intermediate can proceed to form E₄(2H)* through subsequent photoinduced re with loss of H₂. These findings, supported by DFT calculations, lead us to conclude that the trapped intermediate is an H₂ complex on the ground adiabatic potential energy suface that connects E₄(4H) with [E₄(2H)* + H₂], Fig 6. As such, we further suggest that this H₂ complex is a thermally populated intermediate in the catalytically central re of H₂ by E₄(4H), and that N₂ reacts with this complex to complete the conversion of [E₄(4H) + N₂] into [E₄(2N2H) + H₂], Fig 9. Whether there are additional intermediates in this catalytic conversion remains to be determined. We do not at this point favor the idea that E₄(2H)* itself is such an intermediate. If H₂ were thermally released from FeMo-co before N₂ binds, forming E₄(2H)*, then it is hard to see why D₂ would not undergo oa by this FeMo-co state, contrary to the observation that D₂ does not react with MoFe protein during turnover in the absence of N₂.⁹

Figure 9.

Cartoon version of suggested catalytic pathway for re/oa activation of FeMo-co for N₂ reduction.

Synopsis.

N₂ reduction by nitrogenase involves the accumulation of four reducing equivalents at the active site FeMo-cofactor to form a state with two [Fe-H-Fe] bridging hydrides (denoted E₄(4H), the Janus intermediate), which is activated to cleave the N≡N triple bond by the reductive elimination (re) of H₂. We report that photolysis of E₄(4H) induces re of H₂ in two steps, each photoinduced. Results indicate the intermediate state in this process is an H₂ complex of FeMo-cofactor.