A Nonheme Sulfur-ligated {FeNO}⁶ Complex and Comparison with Redox-Interconvertible {FeNO}⁷ and {FeNO}⁸ Analogs

Abstract

A nonheme {FeNO}⁶ complex, [Fe(NO)(N3PyS)]²⁺, was synthesized by reversible, one-electron oxidation of an {FeNO}⁷ analog. This complex completes the first known series of sulfur-ligated {FeNO}⁶⁻⁸ complexes. All three {FeNO}⁶⁻⁸ complexes are readily interconverted by one-electron oxidation/reduction. A comparison of spectroscopic data (UV-vis, NMR, IR, Mössbauer, X-ray absorption) provides a complete picture of the electronic and structural changes that occur upon interconversion between {FeNO}⁶ – {FeNO}⁸. Dissociation of NO• from the new {FeNO}⁶ complex is shown to be controlled by solvent, temperature and photolysis, which is rare for a sulfur-ligated {FeNO}⁶ species.

Graphical Abstract

A nonheme thiolate-ligated {FeNO}⁶ complex is synthesized by oxidation of its {FeNO}⁷ analog, which completes the first sulfur-ligated redox interconvertibile {FeNO}^6-8 series. The series exhibits a unique change from low spin to high spin upon stepwise reduction of the {FeNO}⁶ throug {FeNO}⁸. The thiolate-ligated {FeNO}⁶ shows photolabililty of the nitrosyl ligand without S-oxygenation.

Sulfur ligation plays a significant role in the stability and reactivity of biological iron-nitrosyl (FeNO) species, including both in heme and nonheme metalloenzymes.^[1] For example, thiolate ligation to the Fe-centers in proteins is known to influence the relative binding affinity of NO• to the metal.^[2] Nonenzymatic, sulfur-bound iron regulatory proteins contain [Fe–S] clusters which also react with NO• to modulate gene expression.^[3] The Fe(NO) species observed in biology, including sulfur-ligated species, generally span three redox states abbreviated as {FeNO}⁶⁻⁸ in the Enemark-Feltham notation.^[4] Shuttling between these different redox states by the same FeNO complex may be an essential mechanistic requirement for certain biochemical pathways, such as the reduction of NO• to N₂O mediated by heme or nonheme iron sites in bacterial or fungal nitric oxide reductases (NORs).^[5] Meanwhile, oxidation of NH₂OH, essential to energy transduction by NH₃-oxidizing bacteria, requires facile {FeNO}^7/6 interconversion.^[6]

Given the importance of {FeNO}⁶⁻⁸ centers in biological systems, significant efforts have been made to synthesize well-defined {FeNO}⁶⁻⁸ complexes and examine their structural, spectroscopic and reactivity properties. However, it is challenging to construct ligands that will stabilize an FeNO unit in all three redox states. Examples of {FeNO}⁶⁻⁸ complexes that maintain identical ligand environments are scarce, and to our knowledge, a sulfur-ligated series of such complexes has not been reported to date.^[7] Herein we describe the synthesis of a thiolate-ligated {FeNO}⁶ complex, [Fe(NO)(N3PyS)]²⁺ (1), providing us with the first complete series of sulfur-bound {FeNO}⁶⁻⁸ complexes (Scheme 1). Complex 1 is prepared via a rare one-electron oxidation of an {FeNO}⁷ precursor, [Fe(NO)(N3PyS)](BF₄) (2),^[8] whereas normally {FeNO}⁶ complexes are prepared by reaction of NO• with an Fe^III precursor with a few exceptions.^{[7b, 7d, 9]} The one-electron reduction of 2 was reported to give the {FeNO}⁸ complex [Fe(NO)(N3PyS)]⁰ (3), although no structural information was available.^[10] Herein X-ray absorption spectroscopy (XAS) of the complexes 1–3 confirms their structural integrity in solution, and allows for a comparison of iron-ligand bond lengths across the series. Finally, it is shown that low-intensity irradiation of 1 causes the facile release of NO•, in contrast to previous reports suggesting that thiolate-ligated nonheme {FeNO}⁶ species do not release NO• under photolysis conditions.^[11]

Scheme 1.

Reversible one-electron redox interconversion of [Fe(NO)(N3PyS)]ⁿ⁺ species

Complex 2 was previously characterized by cyclic voltammetry (CV), leading to the observation of a reversible reduction wave at E_1/2 = −1.18 V and an oxidation wave at E_1/2 = 0.013 V (vs Fc⁺/Fc) in CH₃CN at 25 °C.^[8] We have recently reported that the one-electron reduction of 2 yields an {FeNO}⁸ complex.^[10] From the CV results, the one-electron-oxidized {FeNO}⁶ complex should also be chemically accessible. Upon introduction of one equiv of acetylferrocenium tetrafluoroborate ( E_1/2 = +0.27 V vs Fc⁺/Fc in CH₂Cl₂) to a dark brown solution of 2 in MeCN at 25 °C, a rapid color change to green is observed. Analysis of this reaction by UV-visible spectroscopy showed isosbestic conversion of 2 (350 nm, 440 nm, 540 nm) to a new species with peaks at 342 (ε ≈ 12200 M⁻¹ cm⁻¹) and 636 nm (ε ≈ 820 M⁻¹ cm⁻¹). However, decay of the new peak at 342 nm is observed within several minutes at 25 °C, along with a concomitant formation of a broad absorbance centered at 880 nm. This low energy band corresponds to the previously characterized [Fe^III(N3PyS)(MeCN)]²⁺ (4), which lacks the NO ligand.^[12] The use of other one-electron oxidants, such as tris(4-bromophenyl)ammoniumyl tetrafluoroborate or thianthrenium tetrafluoroborate, leads to similar results. Oxidation of 2 in CH₃CN at −40 °C gives the same oxidation product, but now with good stability for up to 8 h (Figure 1). Changing the solvent to CH₂Cl₂ further enhances the stability of the product such that there is no apparent decay by UV-vis at 25 °C for 24 h. However, precipitation occurs in CH₂Cl₂ at higher concentrations, making CD₃CN more useful for NMR studies (vide infra). The formation of the oxidation product is fully reversible in either CH₃CN at –40 °C or CH₂Cl₂ at 25 °C, as shown by the quantitative conversion back to 2 with the addition of the one-electron reductant decamethylferrocene (FeCp*₂; E_1/2 = 0.013 V vs Fc⁺/Fc in CH₂Cl₂). Reversible oxidation of thiolate-ligated iron complexes is challenging because of anticipated side-reactions such as disulfide formation, or generation of sulfur-bridged species.^[13] However, the latter results indicate that the thiolate-ligated {FeNO}⁷ complex 2 is quantitatively and reversibly oxidized to a new species, assigned as the {FeNO}⁶ complex 1 (Scheme 2). Monitoring the oxidation of 2 by ¹H-NMR spectroscopy in CD3CN at −40 °C reveals the disappearance of the paramagnetic spectrum seen for 2 and the generation of a new set of sharp peaks in the diamagnetic region (0 – 10 ppm) assigned to 1. This result indicates that 1 exhibits an S = 0 ground state, similar to other {FeNO}⁶ complexes. ^{[7a, 7b, 7f, 11, 14, 15]} Addition of one equiv of FeCp*₂ restores the paramagnetic spectrum for 2, confirming the reversible transformation between 1 and 2. Complex 2 exhibits temperature-dependent spin-crossover behavior between high-spin (S = 3/2) and low-spin (S = 1/2) states.^[12] It is of interest to note that reduction of 2 leads to the high-spin (S = 1) {FeNO}⁸ complex 3, ^[10] whereas oxidation of 2 leads to a low-spin {FeNO}⁶ complex (Scheme 1). The few other series of {FeNO}⁶⁻⁸ complexes do not show such spin-state variability, remaining either high- or low-spin throughout. These results indicate that the N3PyS ligand is versatile and can support a range of redox and spin states, and provides a ligand field that appears to be on the border between strong and weak-field systems.

Figure 1.

UV-vis spectra for the oxidation of 2 (solid) to 1 (dashed) in CH₃CN at −40 °C.

Scheme 2.

Reversible one electron oxidation of [Fe(NO)N3PyS]⁺ (2⁺)

The IR spectrum for 2 in CD₃CN at 25 °C was reported to show peaks at 1649 cm⁻¹ and 1737 cm⁻¹, assigned as the ν(NO) stretches for S = 1/2 and S = 3/2 spin states, respectively.^[8] Upon one-electron oxidation, these bands disappear and a peak at 1909 cm⁻¹ is observed in the low-temperature IR spectrum for 1 (Figure S13, vide infra), which is at the high end of the range reported for {FeNO}⁶ complexes (1780 – 1927 cm⁻¹).^{[7a, 7b, 7f, 11a, 11b, 11d, 14–15]} This relatively high energy ν(NO) stretch suggests that 1 may have significant Fe^II−NO⁺ character. Geometry optimization by density functional theory (DFT) calculations at the BP86/def2-TZVP level of theory, together with frequency calculations for the singlet state (¹1), produce an intense ν(NO) mode at 1928 cm⁻¹, a good match with the experimental NO stretch. Complex 3 exhibits ν(NO) as 1588 cm⁻¹.^[10] The ν(NO) stretch decreases from {FeNO}⁶ to {FeNO}⁸ (Table 1), as may be anticipated from sequential reduction across the series.

Table 1:

Spin states, N-O stretching frequencies, and metrical parameters obtained from EXAFS for the {FeNO}⁶⁻⁸ series 1–3.

Complex	{FeNO}6 (1)	{FeNO}7 (2)		{FeNO}8 (3)
Spin state	S = 0	S = 1/2	S = 3/2	S = 1
ν(NO) (in cm−1)	1909	1640	1753	1588
Fe–Navg (Å)	1.99	2.00	-	2.11
Fe–S (Å)	2.30	2.28	-	2.39
Fe–NO (Å)	1.69	1.75	-	1.68

A Mössbauer spectrum for 2–⁵⁷Fe in CH₃CN at 80 K revealed a doublet with isomer shift δ = 0.32 mm s⁻¹ and quadrupole splitting |ΔEQ| = 0.48 mm s⁻¹, similar to that obtained for 2 in the solid state (Figure 2).^[8] Generation of 1-⁵⁷Fe in CH₃CN at −40 °C leads to a significant decrease in isomer shift to 0.03 mm s⁻¹ and an increase in quadrupole splitting to 1.7 mm s⁻¹ (Figure 2). Although it is tempting to attribute the decrease in isomer shift to an increase in formal metal oxidation state, 1 complies more with an Fe^II−NO⁺ description in line with the high ν(NO) stretch. Other nonheme {FeNO}⁶ (δ = −0.16 – 0.05 mm s⁻¹) species have been similarly described as Fe^II−NO⁺, without a metal centered oxidation.^{[7b, 7f, 11b, 16]} Similar parameters have been reported for heme {FeNO}⁶ complexes as well.^[17]

Figure 2.

Zero-field Mössbauer spectra for 2 (top, solid line = best fit), and 1 (bottom, solid line = best fit), in CH₃CN at 80 K.

Mössbauer parameters were calculated for ¹1 with the help of an experimentally calibrated data set,^[18] giving δ_(calc) = −0.04 mm s⁻¹, |ΔEQ|_(calc) = 1.83 mm s⁻¹, that match closely with experiment. The frontier orbitals for ¹1 (Figures S24–25) indicate that the best description of the electronic structure is Fe^II(NO⁺), with π-donation indicated via filled Fe(d_π)-NO(p_π*) interactions. An analogous {FeNO}⁶ center in biology with mixed N/S ligation comes from the nonheme iron enzyme nitrile hydratase (NHase_dark) which exhibits δ = 0.03 mm s⁻¹; |ΔEQ| = 1.47 mm s⁻¹, and was described as having an electronic configuration in its inactive form similar to 1-⁵⁷Fe.^[19]

The molecular structure of 2 was previously obtained by single crystal X-ray diffraction (XRD), but efforts to grow single crystals of 1 and 3 for XRD were unsuccessful. Thus we used X-ray absorption spectroscopy (XAS) for structural characterization of 1 – 3. The Fe K-edge X-ray absorption spectra for 1–3 are given in Figure 3. The energies of the rising edge inflection points increase upon sequential oxidation from 3 (7121.7 eV) to 2 (7122.6 eV) to 1 (7125.7 eV). Smaller shifts are found for the weaker pre-edge feature assigned using TDDFT (Figures S2,S4,S6) as excitations from Fe 1s to valence vacancies comprising variable contributions of Fe 3d, NO π*, and N3PyS orbitals: the band for 3 is at 7112.8 eV, while 2 and 1 are isoenergetic at 7113.6 eV. The intensity of the pre-edge feature is diminished in 1, consistent with increased centrosymmetry of Fe with a low-spin, effectively d⁶ configuration and linear Fe–NO interaction.

Figure 3.

Fe-K edge XAS spectra for 1 (short dashed), 2 (solid) and 3 (long dashed).

Fitting of the extended X-ray absorption fine structure (EXAFS) for 1-3 (Figure S1, S3, S5) support their expected overall structural features. The best fits for all three complexes included four N/O scatterers, one S scatterer, and one short N/O scatterer corresponding to the NO moiety. The average Fe–N distance for the four N ligands was 2.00 Å in 1-2, but 2.11 Å in 3. A similar trend was seen for the Fe–S distance, in which 1 and 2 exhibit the same length of 2.30 Å, while 3 shows an elongated Fe–S = 2.40 Å (Table 1). The Fe–NO distance for 2 (1.75 Å) matches that obtained from XRD, and decreases significantly upon oxidation to 1 (1.69 Å), as would be expected for oxidation of an {FeNO}⁷ complex.^{[7b, 7e, 7f, 20]} The Fe-NO distance for 1 compares with low-spin heme {FeNO}⁶ complexes,^[15] but also matches that seen for a nonheme high-spin {FeNO}⁶ complex (1.68 Å).^[7e] The short Fe–NO distance obtained from EXAFS for 1 and the ν(NO) stretch of 1909 cm⁻¹ accord with a linear FeNO unit. Moreover, no satisfactory EXAFS fit could be obtained for an Fe–O scatterer.The parameters are reproduced well by the optimized geometry for ¹1 (Fe–NO = 1.637 Å). A similar decrease in the Fe–NO bond distance is also seen upon reduction to 3 (1.68 Å), as has been observed upon reduction of other {FeNO}⁷ complexes.^{[7f, 21]} DFT calculations on the triplet ground state for 3 (³3) lead to a geometry in good agreement with that seen by EXAFS, with calculated Fe–NO = 1.684 Å. Thus the XAS data fully support the six-coordinate structural assignments and electronic configurations shown in Scheme 1.

Allowing 1 to warm from −40 °C to 25 °C in CH₃CN leads to the gradual formation of 4 by UV-vis, implicating thermal release of NO• (Scheme 3). In contrast, solutions of 1 in the non-coordinating solvent CH₂Cl₂ are stable for at least 12 h at 25 °C. However, addition of CH₃CN to 1 in CH₂Cl₂ at 25 °C leads to formation of 4 (Figure S14). Addition of azide anion to 1 in CH₂Cl₂ also leads to displacement of the NO• ligand in 1, as seen by UV-vis. These results show that the release of NO• from 1 can be controlled by solvent, temperature, or addition of exogenous ligands.

Scheme 3.

Dissociation of NO• from 1. Conditions: A) warming from −40 °C to 25 °C in CH₃CN, B) addition of CH₃CN to 1 in CH₂Cl₂ at 25 °C, or C) irradiation of 1 (λ > 400 nm) in CH₃CN at −40 °C.

Although the NO• in 1 does not appear to be released in the absence of a coordinating ligand such as CH₃CN or N₃⁻, the stoichiometric addition of Co(TPP) (TPP = tetraphenylporphyrin) to a solution of 1 in CH₂Cl₂ leads to formation of Co(NO)(TPP) as seen by UV-vis, ¹H-NMR, and IR spectroscopies (Scheme 4).^[22] The transfer of NO• from 1 to Co(TPP) either could involve bimolecular attack of Co(TPP) on the NO ligand of 1 or capture of free NO• generated by Fe–NO bond cleavage in 1. Further studies are needed to sort out the details of the NO• transfer mechanism.

Scheme 4.

NO• transfer from 1 to Co(TPP).

Given that 1 releases NO• under different thermal conditions, we sought to determine if 1 would exhibit photodissociation of NO•. Irradiation of 1 with visible light (λ > 400 nm) in CH₃CN at −40 °C led to the rapid decay of 1 and the isosbestic appearance of 4 (Figure 4). In contrast, irradiation of 1 in CH₂Cl₂ over 30 min at 25 °C or −40 °C did not lead to any detectable loss of 1. However, the lack of apparent Fe–NO photolysis in CH₂Cl₂ could be due to the rapid rebinding of NO• in the non-coordinating solvent. Irradiation of 1 at cryogenic temperatures in a frozen matrix of CH₂Cl₂ confirmed this analysis. FTIR spectra were recorded at 15 K before (dark) and after irradiation (illuminated) of a sample of 1-⁵⁷Fe in CH₂Cl₂ (the ⁵⁷Fe-label was used to confirm the purity of the samples by Mössbauer spectroscopy). The “dark” minus “illuminated” difference spectra show peaks at 1909 and 1828 cm⁻¹ for samples prepared with ¹⁴N¹⁶O and ¹⁵N¹⁸O, respectively, in good agreement with an N–O harmonic oscillator (Figure S14).

Figure 4.

UV-Vis spectrum for formation of 4 (black dashed) upon photolysis of 1 (solid) in MeCN at −40 °C.

The few known nonheme thiolate-ligated {FeNO}⁶ complexes^{[11a–d, 23]} do not appear reactive toward light for NO• release.^[24] Similar solvent-dependent NO• lability was seen for a bis(carboxamido-phenylthiolate)iron complex, but photolability of the bound NO• was not observed.^[11d] In fact, it was suggested that S-oxygenation, as seen in the sulfenate/sulfinate ligation of NHase, was likely necessary to induce NO• photodissociation.^[11d] It is clear from the findings presented here that S-oxygenation is not necessary for facile photodissociation of 1.

In summary, we have synthesized a new nonheme {FeNO}⁶ complex, completing a series of isostructural {FeNO}⁶⁻⁸ species. This series is the first to contain a sulfur donor, and demonstrates the versatility of the N3PyS scaffold to stabilize the FeNO unit in multiple redox and spin states.