A Nonheme Mononuclear {FeNO}⁷ Complex that produces N₂O in the Absence of an Exogenous Reductant

Abstract

A new nonheme iron(II) complex, Fe^II(Me₃TACN)((OSi^Ph2)₂O) (1), is reported. Reaction of 1 with NO_(g) gives a stable mononitrosyl complex Fe(NO)(Me₃TACN)((OSi^Ph2)₂O) (2), which was characterized by Mössbauer (δ = 0.52 mm s⁻¹, |ΔE_Q| = 0.80 mm s⁻¹), EPR (S = 3/2), resonance Raman (RR) and Fe K-edge X-ray absorption spectroscopies. The data show that 2 is an {FeNO}⁷ complex with an S = 3/2 spin ground state. The RR spectrum (λ_exc = 458 nm) of 2 combined with isotopic labeling (¹⁵N, ¹⁸O) reveals ν(N-O) = 1680 cm⁻¹, which is highly activated, and is a nearly identical match to that seen for the reactive mononitrosyl intermediate in the nonheme iron enzyme FDPnor (ν(NO) = 1681 cm⁻¹). Complex 2 reacts rapidly with H₂O in THF to produce the N-N coupled product N₂O, providing the first example of a mononuclear nonheme iron complex that is capable of converting NO to N₂O in the absence of an exogenous reductant.

Graphical Abstract

We demonstrate the first example of a nonheme, mononuclear {FeNO}⁷ complex that mediates the conversion of NO to N₂O in the absence of an external reductant. A new mononuclear, 5-coordinate Fe^II complex is synthesized, which binds nitric oxide to form the {FeNO}⁷ complex. This complex exhibits a highly activated NO ligand as seen by RR spectroscopy, and generates N₂O in THF upon addition of H₂O.

Introduction

Nitric oxide (NO) is produced in mammals in nanomolar concentrations, and participates in a number of critical biological processes, from cell signaling to blood pressure regulation. During pathogenic invasion, the physiological concentration of NO in mammals increases to micromolar levels as part of the immune response. Microbes such as cyanobacteria, archaea and protozoa express enzymes to combat this influx of NO, including a diiron protein that contains a flavin mononucleotide cofactor (FMNH₂), and reduces NO to N₂O. The overall reaction catalyzed by flavodiiron protein NO reductase (FDPnor) is given in eq 1. ^[1]

$$2\text{NO}+2{\text{e}}^{-}+2{\text{H}}^{+}\to {\text{N}}_2\text{O}+{\text{H}}_2\text{O}$$ (1)

FDPnor contains a nonheme diiron active site with two five-coordinate Fe centers bound to the protein through 2 N(His)/1 O(Asp or Glu) donor ligands, and bridged by carboxylate and water/hydroxo ligands (Figure 1). The sixth coordination site on each Fe remains open for binding of NO, and is surrounded by non-coordinating water molecules.^[2] The balanced reaction in eq 1 shows that two reducing equivalents are required for conversion of NO to N₂O,^[3] but the timing and mechanism of the reducing events during the catalytic reduction of NO by FDPnor remain poorly understood. The enzymatic reaction begins with the initial binding of NO to the diferrous (Fe₂^II) state, and spectroscopic evidence shows the successive formation of mononitrosyl {FeNO}⁷ (Enemark-Feltham notation)^[4] [Fe^II{FeNO}⁷] and dinitrosyl [{FeNO}⁷]₂ species.^[5] An important question is whether the reducing electrons provided by the flavin cofactor are transferred to the diiron active site before or after the critical N–N bond formation step. These mechanistic possibilities are outlined in Figure 1. Mechanism A involves reduction of the dinitrosyl [{FeNO}⁷]₂ site, by either one or two electrons, to give {FeNO}⁸ species prior to N–N coupling, while mechanism B involves N–N bond formation at the {FeNO}⁷ stage, followed by the transfer of the reducing equivalents to the final diferric state to regenerate the Fe^II₂ active state. In the latter mechanism, N–N bond formation may occur via coupling of two {FeNO}⁷ units, or possibly by direct attack of free NO on a single {FeNO}⁷ site. Evidence for mechanism B was obtained from examination of a deflavinated form of FDPnor, which showed that the {FeNO}⁷ species was competent to produce N₂O in a single turnover reaction.^{[5a, 6]} However, the mechanism could be significantly different under catalytic conditions in the presence of FMNH₂ as compared to in the absence of FMNH₂. For either mechanism A or mechanism B to be operative, it is assumed that the FeNO moiety must be sufficiently activated and/or oriented toward another NO/FeNO to lead to productive N–N coupling.

Figure 1.

Plausible mechanisms for N₂O formation by FDPnor.

Only a small number of synthetic nonheme iron complexes are known to mediate the reduction of NO to N₂O.^[7] Two dinuclear dinitrosyl iron ({FeNO⁷})₂ complexes were shown to produce N₂O upon 1e^– or 2e^– reduction.^{[7c, 7i, 7o]} It was also shown that a dinuclear mononitrosyl iron (Fe^II{FeNO}⁷) complex can be reduced chemically or electrochemically to produce N₂O.^[7h] These studies lend support to mechanism A, in which the {FeNO}⁷ intermediate(s) are reduced prior to N–N coupling and N₂O production. The first example of a nonheme iron model complex that was capable of carrying out the direct reduction of NO to N₂O in the absence of an exogenous reductant was reported in 2014.^[8] In this study, a dinuclear ({FeNO}⁷)₂ complex, [Fe₂(N-Et-HPTB)(O₂CPh)(NO)₂](BF₄)₂, was formed by addition of NO to a diiron precursor with 5-coordinate Fe^II centers, and produced N₂O upon illumination with white light. Interestingly, the spectroscopic (infrared and resonance Raman (RR)) data provided strong evidence for the photoproduction of a mononitrosyl [Fe^II{FeNO}⁷) intermediate, which then underwent coupling with a second NO molecule to give N₂O. A later study in 2018 showed that a dinuclear Fe^II₂ complex was also capable of converting NO to N₂O in the absence of exogenous reductant.^[9] In this latter study, the saturated, 6-coordinate Fe^II centers were speculated to undergo dissociation of a carboxylate bridge prior to formation of a putative, reactive [{FeNO}⁷]₂ intermediate. Although low-temperature spectroscopic studies revealed an [{FeNO}⁷]₂ species, it was found to be inactive toward N₂O production. It was proposed that the latter low-temperature species was trapped in an unproductive conformation for N–N bond formation. A similar observation was made for an [{FeNO}⁷]₂ complex in which the anti-conformation of the two NO units negated the possibility of N₂O release upon 1e^– reduction.^[10]

We were interested in determining the conditions necessary to carry out NO reduction by nonheme iron centers. It is our hypothesis that establishing these minimal requirements in well-characterized, synthetic complexes will provide knowledge of the critical elements needed for the catalytic, enzymatic process, and will also help support or refute proposed mechanistic pathways.^[11] With these goals in mind, we focused our efforts on mononuclear nonheme iron complexes that can mediate the reduction of NO to N₂O. We speculated that a mononuclear nonheme iron center, with an appropriate ligand environment, might have the minimal attributes needed to activate NO toward N–N bond formation and N–O cleavage to give N₂O, circumventing the need for a dinuclear system by taking advantage of the diffusional reactivity accessible to small molecules. Our initial efforts led to a sulfur-ligated mononuclear iron complex ([Fe^II(N3PyS)]⁺) that supported the binding of NO in a rare series of redox-interconvertible {FeNO}⁶⁻⁸ species with identical ligand environments.^[12] Reversible one-electron reduction of the {FeNO}⁷ complex gave an {FeNO}⁸ complex, which then decayed to produce N₂O in sub-stoichiometric yield (54%).^[13] This result lent support to the two-electron reduced pathway in mechanism A (Figure 1), which requires reduction to {FeNO}⁸ prior to N₂O formation. It also provided initial evidence that a dinuclear scaffold was not required for reduction of NO to N₂O by nonheme iron in a small-molecule system.

Motivated by our success with the 1e^– reduced, mononuclear {FeNO}⁸ complex, we wondered whether NO reduction might be achieved by an appropriate mononuclear {FeNO}⁷ complex, in the absence of any exogenous reductants. Our criteria for an appropriate system included 1) a five-coordinate Fe^II precursor with N/O donors in the primary coordinating sphere, 2) enough steric encumbrance to stabilize mononuclear FeNO species, and 3) an electron-donating ligand environment to help activate bound NO toward N–N coupling and N-O cleavage. These criteria were fulfilled by [Fe^II(Me₃TACN)((OSi^Ph2)₂O)] (1), which we have employed to prepare the new {FeNO}⁷ complex [Fe(NO)(Me₃TACN)((OSi^Ph2)₂O)] (2). Importantly, complex 2 also rapidly produces N₂O in THF upon addition of H₂O, providing the first example of a mononuclear, nonheme {FeNO}⁷ complex that generates N₂O in the absence of external reductants. In addition, vibrational analysis by RR reveals that 2 has a dramatically weakened N–O stretch for an {FeNO}⁷ (S = 3/2) species, and the N-O stretching frequency is nearly identical to that seen for the activated {FeNO}⁷ unit in the mononitrosyl intermediate of FDPnor.^[14]

Results and Discussion

Synthesis and characterization of [Fe^II(Me₃TACN)((OSi^Ph2)₂O)] (1).

The synthesis of [Fe^II(Me₃TACN)((OSi^Ph2)₂O)] (1) is shown in Scheme 1. The disiloxide ligand^[15] was selected because of the similarities with the anionic O-donors of FDPnor, and because it provides a stable chelate with steric protection by the Ph groups. We hypothesized that the strong σ-donating ability of the disiloxide donors would significantly activate an {FeNO}⁷ species toward N–N bond-forming reactivity. Mixing of K₂[(OSi^Ph2)₂O] in THF with an equimolar purple solution of [Fe^II(Me₃TACN)(CH₃CN)₃](OTf)₂^[16] in CH₃CN was followed by workup and crystallization in pentane/THF to give pale blue crystals of 1 (35%).

Scheme 1.

Syntheses of Complexes 1 and 2

X-ray diffraction (XRD) revealed that 1 is a 5-coordinate, mononuclear Fe^II complex bound by the Me₃TACN and [(OSi^Ph2)₂O]^2- ligands in a distorted square pyramidal geometry (τ= 0.29).^[17] The two Fe–O distances are nearly identical, and the Fe–N_TACN distances are similar to those seen for other Fe^II(Me₃TACN) complexes (Figure 2).^[18] The cyclic voltammogram (CV) of 1 in MeCN shows a quasi-reversible wave at E_1/2 = –0.90 V (vs Fc⁺/Fc; ΔE_pp = 0.12 mV) (Figure S3), assigned to the Fe^III/II redox potential and is consistent with an electron-rich metal center for 1. The zero-field Mössbauer spectrum of a frozen sample of ⁵⁷Fe-1 in THF is shown in Figure 2. A sharp quadrupole doublet with δ = 0.97 mm s⁻¹, |ΔE_Q| = 1.98 mm s⁻¹ is observed, consistent with a high-spin Fe^II species.

Figure 2.

(Left) Displacement ellipsoid plot (50% probability level) for 1 at 110(2) K. H atoms are omitted for clarity. Selected bond distances (Å): Fe1–O1, 1.9402(16), Fe1–O2, 1.9458(13), Fe1–N1, 2.265(6), Fe1–N2, 2.280(6), Fe1–N3, 2.212(6). (Top-right) Mössbauer spectrum of ⁵⁷Fe-1 (80 K). Data: black circles; best fit: red line.

Reactivity of NO_(g) with 1 and characterization of [Fe(NO)(Me₃TACN)((OSi^Ph2)₂O)] (2).

Addition of NO_(g) to 1 in THF or 2-MeTHF forms a new species (2) with UV-vis absorption peaks at 446 nm (ε ~ 850 M^–1cm^–1), 570 nm (ε ~ 350 M^–1cm^–1), (Figure 3), and a ¹H-NMR spectrum with paramagnetically shifted resonances distinct from those of 1. (Figure S4). In contrast to 1, the CV of 2 shows only irreversible redox events at negative potentials (Figure S5). Complex 2 is stable to sparging with Ar(g), indicating that the binding of NO to Fe is not reversible. Complex 2 immediately decays upon exposure to aerobic conditions, and adequate care was thus taken for further characterization.

Figure 3.

UV-vis spectra for 1 (1.4 mM) (black line) after addition of NO(g) to form 2 (green line) in THF. Inset: EPR spectrum of 2 in 2-MeTHF recorded at 8 K, 9.21 GHz.

The EPR spectrum obtained for 2 in 2-MeTHF exhibits g-values at g = [4.08, 3.93, 1.99], which are typical of {FeNO}⁷ (S = 3/2) centers (Figure 3 inset).^{[19],[11a, 12]} Mössbauer spectroscopy for the reaction of 1 with NO(g) shows a broad doublet with δ = 0.52 mm s⁻¹, |ΔE_Q| = 0.80 mm s⁻¹ (Figure 4).^[20],[21] The asymmetry in the intensity is due to vibrational anisotropy observed commonly in low symmetry systems.^[22] DFT calculations (vide infra) yield theoretical values of δ = 0.44 mm s⁻¹, |ΔE_Q| = 0.99 mm s⁻¹ for the {FeNO}⁷ complex 2, in reasonable agreement with experiment. The isomer shift for complex 2 is lower than that observed for the {FeNO}⁷ unit in the deflavo-(FDP)NO (δ = 0.68 mm s⁻¹) suggesting a stronger polarization of the Fe–NO center towards Fe^III-NO^–.^[23]

Figure 4.

⁵⁷Fe Mössbauer spectrum (80 K in frozen THF) of 2. Experimental data are plotted as black dots and best fits are overlaid as solid lines. The major species (δ = 0.52 mm s⁻¹, |ΔE_Q| = 0.80 mm s⁻¹, 90 %) is shown as a green line. A second broad sub-component with δ = 0.6 mm s⁻¹, |ΔE_Q|= 1.00 mm s⁻¹, Г_R= Г_L = 6.0 (blue line) was added to represent the intermediate relaxation of the doublet at 80 K.

A series of RR spectra obtained for 2 are shown in Figure 5. Frozen solutions of 2 in THF reveal a RR band at 1680 cm^–1 which shifts to 1648 (Δν = –32 cm^–1) and 1609 cm^–1 (Δν = –71 cm⁻¹) upon labeling with ¹⁵NO and ¹⁵N¹⁸O, respectively. These downshifts are in good agreement with a simple harmonic oscillator model for ν(NO), and thus, the 1680 cm⁻¹ band is assigned as the ν(NO) mode. Complex 2 displays one of the lowest energy ν(NO) vibrations observed for a high-spin (S = 3/2) {FeNO}⁷ complex, which typically range between 1700 and 1840 cm⁻¹ (Table 1).^{[21], [24]} The low ν(NO) indicates that 2 contains a highly activated nitrosyl ligand for an {FeNO}⁷ (S = 3/2) species, and is consistent with our hypothesis that the strongly donating ancillary ligands of the disiloxide chelate would lead to potent activation of the FeNO unit.^{[14, 25]} Complex 2 also provides a relatively rare example of an {FeNO}⁷ species for which ν(Fe–NO) is observed. The Fe–NO stretch is seen at 460 cm⁻¹ and shifts to 453 and 447 cm⁻¹ with ¹⁵NO and ¹⁵N¹⁸O, respectively. A weaker signal at 915 cm⁻¹, that shifts −40 cm⁻¹ with ¹⁵N¹⁸O, is assigned to a combination mode of the stretching and bending vibrations indicating that the fundamental δ(Fe-N-O) occurs at 455 cm^-1. Similar detection of [ν(Fe–NO)+δ(Fe-N-O)] combination mode was reported previously in Fe(NO)(EDTA) and several nonheme diiron proteins including deflavo-FDP(NO)₂.^{[5a, 8, 26]}

Figure 5.

Low-temperature RR spectra of 2-NO (black), 2-¹⁵NO (red), and 2-¹⁵N¹⁸O (blue) in THF obtained with a 458-nm laser excitation. The difference spectra are shown in green.

Table 1.

Vibrational Data for Selected Six-Coordinate Nonheme {FeNO}⁷ (S = 3/2) Complexes.

Complex	ν(NO) (cm−1)	ν(Fe–NO) (cm−1)	References
[Fe(NO)(Me3TACN)(OSiPh2)2O)]	1680	460	This work
Deflavo-FDP(NO)	1681	451	[14]
[Fe(NO)(SMe2N4(tren)]+	1685	-	[24b]
[Fe(NO)(Me3TACN)(N3)2]	1690	436	[19]
[Fe(NO)(BMPA-Pr)(Cl)][a]	1726	484	[28]
Deflavo-FDP(NO)2	1749	459	[14]
[Fe(NO)N3PyS]BF4	1768	-	[11a, 12, 29]
[Fe2(BPMP)(OPr)(NO)2]2+[e]	1768	-	[7i]
[Fe(NO)(N3PySEtCN2Ph)](BF4)2	1776	-	[11c]
[Fe2(N-Et-HPTB)(NO)2(O2CPh)](BF4)2 [f]	1784	492	[8]
[Fe2(N-EtHPTB)(NO)2(DMF)2](BF4)3	1798	-	[7o]
[Fe2(N-Et-HPTB)(NO)(DMF)3](BF4)3	1810	492	[7h, 8]
[Fe(NO)(TPA)(CH3CN)](ClO4)2	1810	502	[28]
Fe(T1Et4iPrIP)(THF)(NO)(OTf)](OTf)	1831	-	[30]

^[a]BMPA-Pr = N-propano-ate-N,N-bis(2-pyridylmethyl)amine

^[b]T1Et4iPrIP T1Et4iPrIP = Tris(1-ethyl-4-isopropyl-imidazolyl)phosphine

^[c]TPA = tris(2-pyridylmethyl)amine and BF = benzoylformate

^[d]TLA = tris((6-methyl-2-pyridyl)methyl)amine

^[e]BPMP = 2,6-bis[[bis(2-pyridylmethyl)amino] methyl]-4-methylphenol

^[f]Et-HPTB = N,N,N′,N′-tetrakis(2-(l-ethylbenzimidazolyI))-2-hydroxy-1,3-diaminopropane.

Calculations performed on FDPnor, as well as on synthetic models, suggest that a decrease in the energy barrier for N–N bond formation is expected upon enhancement of the nitroxyl (NO^–) character of an FeNO center.^[27] The weakened N–O vibration of 1681 cm^–1 seen for the mononitrosyl form of FDPnor suggests a fairly polarized Fe^III-NO^– unit.^[14] This enhanced nitroxyl character was partly attributed to an interaction of the bound NO group with the second Fe^II site. The ν(NO) and ν(Fe–NO) values for 2 are in excellent agreement with ν(NO) = 1681 cm^–1 and ν(Fe–NO) = 451 cm^–1 for deflavo-FDP(NO),^[14] and the ν(N–O) mode for 2 is indicative of a same degree of N-O bond activation. We surmise that the mononuclear complex 2 does not require a second metal site to activate NO to the same extent as the protein because of the strongly donating anionic oxygen ligands in the primary coordination sphere. Electronic donation to the Fe center from the ancillary ligands is known to lower the ability of Fe to accept π-donation from π*(NO),^[28] which leads to an increase in the π*-antibonding population and weakening of the N–O bond. The decrease in the π*-antibonding donation to the iron also results in a weak Fe–NO bond.^{[21, 28]} Our strategy to use electronically-enriched ligand environments for NO activation by nonheme iron is supported by the vibration data.

Attempts to isolate 2 in a pure crystalline form were unsuccessful. We turned to X-ray absorption spectroscopy (XAS) for geometric and electronic structural characterization of 2. Fitting of the extended X-ray absorption fine structure (EXAFS) for 2 is consistent with a six-coordinate Fe center with three N/O-scatterers at 2.32 Å and two N/O-scatterers at 1.97 Å (Table S1). An additional Fe–N/O scattering path with a distance of 1.76 Å accords with the Fe–NO distance expected for an S=3/2 {FeNO}⁷ complex. The pre-edge of 2 is blue-shifted relative to 1 consistent with a strengthened ligand field upon increase of coordination number from 5 to 6. The blue shift of the rising edge for 2 relative to 1 suggests diminished electron density at Fe consistent with delocalization of electron density from Fe to NO (Figure 6). Taken together, the XAS data confirm the structural integrity of 2 in solution as a mononuclear {FeNO}⁷ complex.

Figure 6.

Fe K-edge X-ray absorption near edge (XANES) spectra of [Fe^II(Me₃TACN)((OSi^Ph2)₂O)] (1, black line) and [Fe(NO)(Me₃TACN)((OSi^Ph2)₂O)] (2, red line).

Geometry optimization of 2 by DFT (BP86/6–311g*) yielded an intact six-coordinate structure around the iron center, with three nitrogen atoms, two oxygen atoms, and an axial nitrosyl group, as supported by the experimental results from EXAFS. Frequency calculation predicts a low value of 1629 cm⁻¹ for the NO vibration, and a shift to 1598 cm⁻¹ for the ¹⁵N¹⁸O isotopomer, in agreement with a simple harmonic oscillator model. The Fe–NO and N–O bond lengths were calculated to be 1.738 Å and 1.199 Å respectively. The Fe–N–O angle (=144.16°) was found to be significantly bent, which accords with the low N–O vibration. A higher negative spin density was observed on the nitrogen atom as compared to the oxygen atom of the NO group.^{[7j, 7l]} The absorption spectrum for 2 was simulated by using the TDDFT method with the CAM-B3LYP range-separated functional (19–65% Hartree-Fock exchange) and an Ahlrichs triple zeta basis set (def2-TZVP).^[31] A correction of the simulated spectra by 0.56 eV was required to match the experimental spectral features, to account for the well-established over-estimation of the absolute transition frequency by TDDFT calculations.^[32] Transition difference density plots were analyzed for interpretation of the vertical transitions under the major absorption features as shown in Figure 7. The experimental and simulated spectrum are found to be in good agreement with each other. The intense band at 446 nm could be assigned to excitations within the β- spin manifold. However, the less intense band at 570 nm contains a group of excitations occurring within both the α- and β-spin manifolds which coalesce into one due to their closeness in energy. For both cases, the major contribution arises from a ligand-to-metal charge transfer (LMCT) originating from the NO-(π*) to an admixture of Fe(III)-d_xz/yz orbitals. The band at 570 nm has additional contributions due to the ligand-ligand charge transfer excitations arising from the O(siloxide)/ N(Me₃TACN) donors to the NO(π*) within the α-manifold.

Figure 7.

TD-DFT simulated UV-vis absorption spectrum for 2 (dashed purple line) overlaid with the experimental spectrum (red line). The computed spectral excitations are shown as black sticks. Electron difference density map for the computed transition (λ_max = 447 nm) is shown with the green and red regions indicating gain and loss of electron density respectively.

Formation of N₂O from 2.

The {FeNO}⁷ complex 2 is stable at room temperature (23 °C) in dry solvent (e.g. THF, CH₃CN) for at least 2 h, as seen by UV-vis spectroscopy. However, addition of a small amount of water to a THF solution of 2 (1:10 H₂O:THF v:v) results in the rapid decay of 2, as seen by the loss of absorbance at 570 nm (Figure S6). Headspace analysis of the reaction mixture after addition of H₂O was carried out by gas chromatography (GC), and revealed the formation of N₂O. Quantitation against an N₂O calibration curve showed that N₂O was formed in 30% yield based on a 2:1 Fe:N₂O ratio (Scheme 2). Formation of N₂O was observed in similar yield when 2 is generated by the controlled addition of NO_(g) pre-dissolved in THF. The controlled addition of solubilized NO in THF helped in limiting excess NO(g) in the reaction mixture, allowing only the minimum amount required to attain full formation of 2. The production of N₂O was the same under these conditions, showing that excess NO_(g) is not required for N₂O formation. In the absence of H₂O, no N₂O formation was observed, and similarly, no N₂O was observed in the absence of the metal complex. We have made efforts to pinpoint the role of water by testing other Bronsted acids (e.g. HCl, HBF₄, CF₃CO₂H, CF₃SO₂H), but the addition of these acids led to rapid bleaching of the solution. Weaker acids (e.g. HOAc, MeOH) did not give N₂O formation. Addition of isotopically labeled ¹⁵NO (98% ¹⁵N) to give ¹⁵NO-labeled 2 and examination of the same reaction by GC mass spectrometry (GC-MS) revealed a major peak at 46 m/z, corresponding to ¹⁵N₂O. The final reaction mixture was tested for nitrite ion via a Griess assay, which was shown to be negative. This result, along with other control reactions in the absence of iron, provide good evidence that disproportionation of NO to N₂O and NO₂^- may not be occurring. Taken together, the data show that 1 directly converts NO to N₂O without the addition of an exogenous reductant. There is only one other example of a mononuclear Fe^II species in a heme-based environment that produces N₂O in a similar manner.^[7j]

Scheme 2.

N₂O formation from 2.

The ¹H-NMR spectrum of the final reaction mixture showed disappearance of 2, and no other well-resolved species could be observed. The Mössbauer spectrum is poorly resolved and appears to contain a mixture of species (Figure S11). The EPR spectrum of the reaction mixture also suggests a potential mixture of products (Figure S12). Attempts to analyze the final reaction mixture by RR spectroscopy was unsuccessful due to high background signals. Thus we are unable to identify the final iron decay products in the reaction mixture, but the production of N₂O is reproducible under these conditions.

Formation of N₂O from NO in eq 1 requires the presence of H⁺, consistent with the need for the addition of H₂O to trigger production of N₂O from 2. The exogenous water may also help facilitate the reaction by activating the FeNO moiety, or enhancing either the formation or decay of related intermediates through hydrogen bonding. Mutation of a Tyr to Phe residue in the active site of FDPnor resulted in loss of catalytic activity, and DFT calculations implicated the Tyr as providing a key hydrogen bond to an N–N coupled hyponitrite (ONNO^2-) intermediate during NO reduction.^{[27c, 33]} Hydrogen bonding between the solvent and NO-derived intermediates has also been implicated as an important factor in other synthetic complexes.^{[7g, 7n]}

Conclusion

A new 5-coordinate Fe^II complex, 1, was synthesized. It was shown that 1 reacts with NO to yield a 6-coordinate, S = 3/2, {FeNO}⁷ complex 2, which contains a highly activated NO ligand. Complex 2 has been characterized by UV-vis, ¹H-NMR, resonance Raman, EPR, Mössbauer, XAS, and DFT studies. The N–O and Fe–NO vibrations for 2 are observed at 1680 cm^–1 and 460 cm^–1 respectively, close to the ν(NO) = 1681 cm^–1 and ν(Fe-NO) = 451 cm^–1 observed for the mononitrosyl deflavo-FDP(NO), and indicates a similar degree of NO activation to that observed in the protein. Complex 2 reacts rapidly with excess water in THF to yield N₂O, and thus is a rare example of a mononuclear {FeNO}⁷ complex that can mediate the conversion of NO to N₂O without an exogenous reductant. Given that 2 is stable in solution even under vigorous Ar(g) sparging, it is reasonable to assume that NO dissociation from 2 is negligible. The production of N₂O is also unaffected by the presence of excess NO(g). Although the mechanism of N₂O formation has yet to be studied in detail, these observations suggest the possibility of a bimolecular reaction between intact {FeNO}⁷ monomers to give N₂O. However, further studies will be required to support this proposed mechanism. This work shows that an appropriately designed, electron-rich ligand environment for nonheme iron can promote substantial activation of small molecules like NO.