A Mononuclear, Nonheme Fe^II-Piloty’s Acid (PhSO₂NHOH) Adduct: An Intermediate in the Production of {FeNO}^7/8 Complexes from Piloty’s Acid

Abstract

Reaction of the mononuclear nonheme complex [Fe^II(CH₃CN)(N3PyS)]BF₄ (1) with an HNO donor, Piloty’s acid (PhSO₂NHOH, P.A.), at low temperature affords a high-spin (S = 2) Fe^II-P.A. intermediate (2), characterized by ⁵⁷Fe Mössbauer and Fe K- edge X-ray absorption (XAS) spectroscopies, with interpretation of both supported by DFT calculations. The combined methods indicate that P.A. anion binds as the N-deprotonated tautomer (PhSO₂NOH⁻) to [Fe^II(N3PyS)]⁺, leading to 2. Complex 2 is the first spectroscopically characterized example, to our knowledge, of P.A. anion bound to a redox-active metal center. Warming of 2 above −60 °C yields the stable {FeNO}⁷ complex [Fe(NO)(N3PyS)]BF₄ (4), as evidenced by ¹H NMR, ATR-IR, and Mössbauer spectroscopies. Isotope labeling experiments with ¹⁵N-labeled P.A. confirm that the nitrosyl ligand in 4 derives from P.A. In contrast, addition of a second equivalent of a strong base leads to S-N cleavage and production of an {FeNO}⁸ species, the deprotonated analog of an Fe-HNO complex. This work has implications for the targeted delivery of HNO/ NO⁻/NO‧ to nonheme Fe centers in biological and synthetic applications, and suggests a new role for nonheme Fe^II complexes in the assisted degradation of HNO donor molecules.

Graphical Abstract

INTRODUCTION

Nitroxyl (HNO/NO⁻, also known as azanone or nitrosyl hydride) is a critically important biomolecule whose endogenous production and pharmacology have yet to be resolved.^1–6 The HNO molecule is a positive cardiac inotrope used in the treatment of heart failure,⁷ and is also a potent inhibitor of aldehyde dehydrogenase,^8–10 showing promise as a therapeutic for alcohol addiction. It is currently thought that HNO participates in signal transduction in biology by coordinating to the heme center in soluble guanylate cyclase.^11–12 Metal-bound HNO species (Mⁿ⁺(HNO)) have been implicated as reactive intermediates in a number of catalytic cycles that process NO_x,^13–15 but only a few examples of Mⁿ⁺(HNO) complexes with first row metals have been reported to date.^16–26 HNO has also been used as a surrogate for O₂; for example substitution of O₂ by HNO in reactions with the enzyme manganese quercetin 2,3-dioxygenase led to substrate nitroxygenation (i.e. HNO incorporation).^27–28 Furthermore, there is considerable interest in the reduction/protonation of metal nitrosyl species (M(NO‧)) to produce viable HNO-releasing agents, as well as to examine the feasibility of HNO production from metal centers and NO‧ in vivo.^{24–25, 29–38}

In contrast to NO‧, which is relatively stable under anaerobic conditions, HNO has an intrinsically short lifetime due to rapid (8.0 × 10⁶ M⁻¹ s⁻¹)³⁹ dimerization to give hyponitrous acid, HON=NOH, which further decays to yield N₂O and H₂O (8.0 × 10⁻⁴ s⁻¹).⁴⁰ To deliver the short-lived HNO in situ for biochemical studies or medical applications, donor molecules that release HNO in response to external stimuli (e.g. pH, temperature, light) must be employed. Sulfohydroxamic acids (RSO₂NHOH, R = -alkyl, aryl) constitute one important class of HNO donors that release HNO mainly in response to pH,⁴¹ and in some cases, through photoactivation.^42–43 Piloty’s acid (N-hydroxybenzenesulfonamide; P.A.), first reported in 1896,⁴⁴ is the oldest member of this class, and releases HNO under basic conditions according to the mechanism shown in Scheme 1. However, donation of HNO by P.A. is slow (k_max ~ 10⁻³ to 10⁻⁴ s⁻¹ at 25 °C).⁴¹

Scheme 1.

Mechanism of HNO Release from Piloty’s Acid

Given that basic conditions are necessary to produce HNO, the P.A.⁻ anion is presumed to be the immediate precursor to HNO release. There are two possible tautomers of monodeprotonated P.A.⁻, either N- or O-deprotonated PHSO₂NOH or PHSO₂NHO, respectively. Given the relatively slow rates of HNO release from P.A., we hypothesized that the P.A.⁻ anion could be an important player in the reactivity with a cationic metal center. Precedent for this hypothesis comes from a high-resolution crystal structure of zinc carbonic anhydrase (human isoform II) inhibited by P.A.⁴⁵ This structure shows that P.A.⁻ anion binds to the active site by coordinating to the zinc(II) center through the sulfonamido N atom, i.e. the N-deprotonated tautomer. Alkali metal salts of P.A.⁻ anion (e.g. Li⁺, Na⁺) exist as the N-deprotonated tautomer in both the solid state and in dioxane solution, ^46–47 and there is computational evidence to suggest that this tautomer is energetically favored. ⁴⁸ Other data suggest that it is the O-deprotonated tautomer that leads to HNO release.⁴⁹

The interactions of redox-active, biologically relevant metal ions such as Fe, Mn, or Co, with HNO donors such as P.A. have been examined previously, and in some cases, were shown to accelerate the decomposition of P.A. and related RSO₂NHOH compounds.^{48, 50–51} However, there is still relatively little known about the mechanism of action between redox-active metal ions and RSO₂NHOH compounds. In particular, direct structural or spectroscopic evidence for metal-P.A. adducts is lacking. We have set out to examine nonheme iron complexes and their interaction with P.A. and other RSO₂NHOH donors in order to characterize the potential binding modes, mechanisms of action, and HNO-releasing properties of Fe/P.A. systems.

We previously described the reaction of an Fe^II-solvento complex, [Fe^II(CH₃CN)(N3PyS)BF₄ (1) with P.A. in the presence of KO^tBu/18-crown-6 (1 equiv) at −40 °C. This reaction gave a new, thermally unstable species 2, which was characterized by UV-vis spectroscopy (λ_max = 418, 501 nm).²⁹ Reaction of 2 with a further equivalent of base at −40 °C afforded the {FeNO}⁸ complex [Fe(NO)(N3PyS)] (3) (λ_max = 520, 720 nm), which was also obtained by one-electron reduction of an {FeNO}⁷ precursor, [Fe(NO)(N3PyS)]BF₄ (4), with decamethylcobaltocene (CoCp*₂). The {FeNO}⁸ complex 3 exhibited a rare S = 1 ground state,^29,52 and was shown to produce N₂O in 54% yield upon standing in solution at 23 °C. We reasoned that complex 2, the immediate precursor to the {FeNO}⁸ species, could be an Fe(HNO) or Fe-P.A. adduct, but no further spectroscopic characterization was obtained.

Herein we provide new characterization of 2 by ⁵⁷Fe Mössbauer and Fe K-edge X-ray absorption spectroscopies, which show that 2 is in fact a high-spin (S = 2) Fe^II-P.A. adduct. Density functional theory (DFT) calculations, including predicted Mössbauer parameters for N-bound and O-bound Fe-P.A. adducts, support the spectroscopic results and suggest that the P.A. ligand in 2 is coordinated through nitrogen. Using a combination of Mössbauer, paramagnetic ¹H NMR, and attenuated total reflectance infrared (ATR-IR) spectroscopies, we show that the thermal decay of 2 under strictly anaerobic conditions yields the {FeNO}⁷ complex 4. In contrast, the {FeNO}⁸ complex 3 is produced from adding more base, consistent with 2 being a precursor to Fe(HNO).

RESULTS AND DISCUSSION

Synthesis and solution state properties of [⁵⁷Fe^II(CH₃CN)(N3PyS)]BF₄ (l-⁵⁷Fe).

The Fe^II-thiolate starting material, [⁵⁷Fe^II(CH₃CN)(N3PyS)]BF₄ (1-⁵⁷Fe), was prepared from the isotopically enriched (⁵⁷Fe, 95%) Fe^II-thioether precursor, [⁵⁷Fe^II(CH₃CN)(N3PySEtCN)](BF₄)₂ (5-⁵⁷Fe) by using the method previously described for the normal abundance analogue (Scheme 2).²⁹ Treatment of 5-⁵⁷Fe with KO^tBu (1 equiv) in dry, deoxygenated CH₃CN at 23 °C produces an instantaneous color change from orange-red to wine-red, corresponding to removal of the base-labile 2-cyanoethyl protecting group from the sulfur donor. Pure 1-⁵⁷Fe, a dark red-brown solid, is isolated in 65% crystalline yield after addition of excess diethyl ether to a concentrated CH₃CN/toluene (1:1 v/v) solution of the product.

Scheme 2.

Preparation of Isotopically Labeled [Fe^II(CH₃CN)(N3PyS)]BF₄ (1-⁵⁷Fe)

The zero-field ⁵⁷Fe Mössbauer spectra of 5-⁵⁷Fe and 1-⁵⁷Fe, recorded at 80 K in frozen CH₃CN, are shown in Figure 1. Thioetherligated 5-⁵⁷Fe produces a sharp quadrupole doublet with δ = 0.44 and |ΔE_Q| = 0.55 mm s⁻¹, consistent with a low-spin (is, S = 0) Fe^II ground state. Thiolate-ligated 1-⁵⁷Fe exhibits parameters (δ = 0.43, |ΔE_Q| = 0.31 mm s⁻¹) that are also consistent with ls-Fe^II and accord well with the data previously collected at 5.4 K for crystalline 1 (δ = 0.41 and |ΔE_Q| = 0.26 mm s⁻¹).⁵³ This result provides strong evidence that 1 retains its mononuclear, six-coordinate, solid state structure upon dissolution in CH₃CN. The nearly identical isomer shifts observed for 5-⁵⁷Fe and 1-⁵⁷Fe (0.44 vs. 0.43 mm s⁻¹) make good sense in light of the fact that these complexes possess very similar first coordination sphere metrics by single crystal X-ray diffraction. In the solid state, complexes 1 and 5 exhibit average Fe-N_py distances of 1.9594(13) Å and 1.9659(13) and Fe-S distances of 2.2848(4) and 2.3018 (4) Å, respectively.^53–54 Despite their similar overall coordination environments, 5-⁵⁷Fe and 1-⁵⁷Fe have different |ΔE_Q| values (0.55 vs. 0.31 mm s·1)1 which likely reflects the difference in thioether versus thiolate ligation.

Figure 1.

⁵⁷Fe Mössbauer spectra (80 K in frozen CH₃CN) showing the conversion of 5-⁵⁷Fe and 1-⁵⁷Fe according to Scheme 2. Experimental data are plotted as black dots while best fits are overlaid as solid lines. (a) Spectrum of 5-⁵⁷Fe fit to a single doublet with δ = 0.44, |ΔE_Q| = 0.55, Γ_L=R = 0.29 mm s⁻¹. (b) Spectrum of 1-⁵⁷Fe fit to a single doublet with δ = 0.43, |ΔE_Q| = 0.31, Γ_L=R = 0.25 mm s⁻¹.

Reactivity of 1 with Piloty’s acid.

It was shown previously that complex 1 reacted with Piloty’s acid (P.A.) and KO^tBu/18-crown-6 in CH₃CN at −40 °C to give a new, thermally unstable intermediate 2, which was characterized by UV-vis spectroscopy (λ_max = 418, 501 nm).²⁹ This species is not observed when the reaction is run at 23 °C (vide infra). We hypothesized that this intermediate might be either an Fe(HNO) or Fe(P.A.) adduct, but no further spectroscopic characterization was obtained. With 1-⁵⁷Fe in hand, intermediate 2-⁵⁷Fe could be examined by Mössbauer spectroscopy. To increase the stability of 2-⁵⁷Fe, we changed to a mixed solvent system of n-butyronitrile/THF, which allowed for lower temperatures to be employed and for the solubility of KO^tBu/ 18-crown-6 to be maintained.

The generation of 2-⁵⁷Fe at ca. −60 °C in CiH1CN/THF is outlined in Scheme 3. Anaerobic, dropwise addition of a THF solution of KO^tBu/18-crown-6 (1 equiv) to a pre-cooled mixture of 1-⁵⁷Fe (5 mM) and P.A. (1 equiv) in C₃H₇CN produces a gradual color change from deep wine red to bright pink red over 2 h. This color change is due to loss of the intense charge-transfer bands for diamagnetic 1 (ε ~ 4000–5000 M⁻¹ cm⁻¹)⁵³ as intermediate 2 (ε ~ 1000 M⁻¹ cm⁻¹) is formed, as shown previously. An aliquot of the cold, deoxygenated reaction mixture can be quickly transferred to a Mössbauer sample holder using air-free techniques. Alternatively, the reaction mixture can be rapidly frozen by pouring into N₂(I), giving frozen pellets which then can be further pulverized to give a fine powder that is loaded into the Mössbauer sample holder.

Scheme 3.

Reaction of [Fe^II(CFFsCN)(N3PyS)]BF₄ and Piloty’s Acid at −60 °C

A representative Miissbauer spectrum for the reaction mixture is shown in Figure 2. The major species, which accounts for 75% of the fitted area, has parameters consistent with a hs-Fe^II species (δ = 1.05, |ΔE_Q| = 2.93 mm s⁻¹; dashed blue line). We assign this doublet to 2-⁵⁷Fe. A minor component corresponds to starting 1-⁵⁷Fe (δ = 0.42, |ΔE_Q| = 0.33 mm s⁻¹; solid orange line, 7% offit). A third, broad component with δ = 0.36 and |ΔE_Q| = 0.61 mm s⁻¹ (18%) is required to fit the remaining area and corresponds to an unidentified product. Longer reaction times (3–6 h), or lower temperature (−80 °C), did not significantly increase the spectroscopic yield of 2-⁵⁷Fe.

Figure 2.

⁵⁷Fe Mössbauer spectrum (80 K, C₃H₇CN/THF, 9:1 v/v) of 1-⁵⁷Fe + PA. (1 equiv) + KO^tBu/18-crown-6 (1 equiv) reacted at −60 °C for 2 h according to Scheme 3. Experimental data (black dots); best fit (red line); majority hs Fe^II species 2-⁵⁷Fe (dashed blue line, 75%; Voigt line shape); minor Is Fe^II species (solid orange line, 7%; Lorentzian line shape) corresponding to 1-⁵⁷Fe; minor species (solid gray hne, 18%) corresponding to an unidentified product.

Our group has shown that an analogous change from low-spin to high-spin Fe^II occurs when I is dissolved in methanol.^{53, 55} This hs Fe^II species was assigned as [Fe^II(CH₃OH)(N3PyS)]⁺ (1-CH₃OH). The 80 K Mössbauer spectrum for 1-⁵⁷Fe dissolved in methanol contains two quadrupole doublets, one arising from the acetonitrilebound species (60%), and the other arising from the proposed methanol adduct (δ = 1.05, |ΔE_Q| = 2.77 mm s⁻¹; 40%) (Figure S1). At room temperature, the CH₃CN in 1 is completely displaced by CH₃OH as evidenced by Evans method measurements,⁵³ and the Mössbauer data show that upon freezing the re-binding of CH₃CN is greatly favored. In contrast, the P.A.-derived ligand readily outcompetes either CH₃CN or the related C₃H₇CN solvent for the Fe binding site. Notably, the isomer shifts for 1-⁵⁷Fe in CH₃OH and 2-⁵⁷Fe are identical, which indicates that their overall coordination environments are similar. The combined observations suggest that binding of the reactive P.A. species to Fe is favored over CH₃CN, and maintains the [Fe^II(N3PyS)]⁺ core structure.

As shown in Scheme 1, HNO release from Piloty’s acid is accompanied by generation of benzenesulfinate anion (PhSO₂⁻), a possible ligand for [Fe^II(N3PyS)]⁺. To test whether 2-⁵⁷Fe is an Fe^II-PhSO₂⁻·adduct rather than a P.A. or HNO adduct, 1-⁵⁷Fe was reacted with Na⁺PhSO₂⁻ according to Scheme 4. This reaction yielded a new, thermally stable hs Fe^II species (6-⁵⁷Fe) with δ = 1.07 and |ΔE_Q| = 3.11 mm s⁻¹, which we assign to a coordinated Fe^II(O₂SPh) adduct. Importantly, these parameters are significantly different from those seen for 2-⁵⁷Fe. Complex 6-⁵⁷Fe was also produced in good yield as seen by Möissbauer spectroscopy (84%) when 1-⁵⁷Fe was treated with a previously reacted 1:1 mixture of P.A. and KO^tBu/ 18-crown-6 at 23 °C (Scheme 4 and Figure 3). The latter result confirms that a 1:1 mixture of P.A. and strong base in organic solvent produces PhSO₂⁻, as it does in water.⁴¹ Taken together, these experiments show that metastable 2 is not an Fe^II(PhSO₂⁻) adduct.

Scheme 4.

Reaction of [Fe^II(CH₃CN)(N3PyS)]BF⁴ and Benzenesulfinate Anion (PhSO₂⁻) Sources at 23 °C

Figure 3.

⁵⁷Fe Mössbauer spectmm (80 K, C₃F₇CN/THF, 9:1 v/v) for the reaction of 1-⁵⁷Fe with a previously reacted 1:1 mixture of PA and KO^tBu/18-crown-6 at 23 °C. Experimental data (black dots); best fit (red line); Fe^II(PhSO₂⁻) adduct 6-⁵⁷Fe (dashed green line, 84%); 1-⁵⁷Fe (solid orange line, 12%); minor species (solid gray line, 4%) corresponding to an unidentified product.

Next, we considered the possibility that 2 may be an Fe-HNO adduct ({Fe(HNO)}⁸ in the Enemark-Feltham notation). The only Mössbauer data on a proposed {Fe(HNO)}⁸ species comes from cryoreduction of the {FeNO}⁷ adduct of taurine dioxygenase (TauD).⁵⁶ This species could only be generated in low yield (17%), but gave Mössbauer parameters of δ = 0.80 and |ΔE_Q| = 1.64 mm s⁻¹ that were consistent with DFT calculations for a quintet (S = 2) {Fe(HNO)}⁸ structure. We turned to DFT methods to predict the Mössbauer parameters for [Fe(HNO)(N3PyS)]⁺. Optimized geometries of the S = 0, 1, and 2 spin states were obtained and employed for the Mössbauer calculations (see Supporting Information). The calculated isomer shifts range from 0.27 to 0.55 mm s⁻¹ (estimated maximum errors of ±0.1 mm s⁻¹),⁵⁷ which are clearly inconsistent with the experimental data for 2-⁵⁷Fe (δ = 1.05 mm s⁻¹). The calculated quadrupole splittings (estimated maximum errors of ±0.5 mm s⁻¹)⁵⁷ are also in poor agreement with 2-⁵⁷Fe. We conclude that 2 is not an Fe-HNO adduct.

Fe K-Edge X-ray Absorption Spectroscopy (XAS).

The normalized X-ray absorption near-edge structure (XANES) of complex 2 is shown in Figure 4, overlaid with the spectrum of starting 1 for comparison. The energies of the rising edge inflection points increase upon going from 1 (7120.7 eV) to 2 (7121.5 eV), consistent with back-donation of electron density from Fe^II to the N–O π* of P.A.

Figure 4.

Normalized Fe K-edge XANES of the hs Fe^II-P A. complex 2 (solid red line) and the Fe^II-solvento precursor 1 (solid blue line) (frozen C₃H₇CN/THF, 9:1 v/v).

The extended X-ray absorption fine structure (EXAFS) of 2 is shown in Figure 5a (solid green trace), along with an overlaid line of best fit (dashed black trace). The corresponding Fourier transforms of the EXAFS data and best fit are shown in Figure 5b. Relevant fitting parameters are provided in Table 1, and result from sequentially improved fits to the data (see Table S1). The best fit for complex 2 includes four N/O scatterers at 2.17 Å and one S scatterer at 2.66 Å, which correspond to the Fe(N3PyS) core. A fifth N/O scatterer at 1.95 Å, an Fe-O scatterer at 2.51 Å, and a second S scatterer at 3.25 can be isolated from the data. These structural features strongly support the assignment of 2 as a hs Fe^II-Piloty’s acid adduct.

Figure 5.

(a) Fe K-edge EXAFS of 2 plotted from k = 2 to 12 Å⁻¹ (solid green line) overlaid with the corresponding best fit (dashed black line), (b) Fourier transforms of the EXAFS data, fit, and residual (data minus fit; solid gray line).

Table 1.

Best Fit to FeK-edge EXAFS of Complex2 (F = 32.97%)^a,b,c

	CN	R (Å)	±	σ2	±
Fe-S	1	2.66	0.01	0.005	0.001
Fe-N	4	2.17	0.01	0.010	0.001
Fe-N(O)	1	1.95	0.01	0.005	0.001
Fe-S	1	3.25	0.01	0.005	0.001
Fe-C	8	4.74	0.01	0.004	0.001
Fe-C	16	3.84	0.01	0.002	0.001
Fe-0	1	2.51	0.02	0.001	0.002

^aSee Table SI for a list of sequentially improved fits.

^b’EXAFS data were fit with EXAFSPAK using paths calculated by FEFF7. Coordination numbers were held constant, whereas distances (R) and Debye-Waller factors (σ²) were allowed to float. Goodness of fit was measured with F, which was defined as ${\left[({\sum }_i^n[k_i^3{({\text{EXAFS}}_{\text{obs}}-{\text{EXAFS}}_{\text{calc}})}_i]{)}^2/n\right]}^{1/2}$. E₀ for the best fit was +3.64 eV.

^cTabulated errors (±) correspond to fitting errors. Expected errors for Rare ± 0.02 Å; expected errors in CN are ca. 20–25% (reference 58).

DFT-calcuIated structures for the Fe^II-Piloty’s acid adduct (2).

Given the availability of EXAFS data for 2, a constrained geometry optimization approach was employed. Our computational studies commenced with quintet [Fe^II(κ¹-N-PhSO₂NOH)(N3PyS)] (⁵A, Figure 6 and Table S3) which contains an N-bound, monodeprotonated P.A. ligand in the sixth coordination site. From EXAFS, the Fe-S(N3PyS), Fe-S(P.A.), Fe-N(O), and Fe▪▪▪O(N) distances were fixed at 2.66,3.25,1.95, and 2.51 Å, respectively. The four Fe-N(N3PyS) distances, which typically have a large degree of asymmetry, were chosen such that their average would equal 2.17 A (i.e., the EXAFS value for four Fe-N scatterers), while all remaining distances were allowed to vary. Using this spectroscopically constrained structure led to the optimized geometry shown in Figure 6. Calculation of the ⁵⁷Fe Mӧssbauer parameters for this geometry gave δ = 0.96 and |ΔE_Q| = 2.81 mm s⁻¹, both of which accord well with the experimental data for 2-⁵⁷Fe (6 = 1.05 and |ΔE_Q| = 2.93 mm s⁻¹). Similarly, when the four Fe-N(N3PyS) distances were allowed to optimize without constraints (structure ⁵B, Table S3), the calculated Mössbauer parameters were δ = 1.06 and |ΔE_Q| = 2.84 mm s⁻¹. An increase in the average Fe-N(N3PyS) distance for ⁵B (2.28 Å) translates to an increase in isomer shift of ~0.1 mm s⁻¹, but has no impact on the calculated quadrupole splitting. Thus, EXAFS-derived structures for 2 lead to calculated Mössbauer parameters that are fully consistent with the experimental data.

Figure 6.

DFT structure of ⁵[Fe^II(κ¹-N-PhSO₂NOH)(N3PySH)]] (⁵A) obtained by optimizing with constraints from the EXAFS data for 2 (Table 1). Calculated ⁵⁷Fe Mӧssbauer parameters are shown in green. C and H atoms are depicted as brown and white spheres, respectively.

Unconstrained geometry optimizations for quintet Fe^II(κ¹-N-PhSO₂NOH)(N3PyS)] produce shorter Fe-S(N3PyS) distances of 2.4–2.5 Å, depending on the choice of functional or basis set. These distances are not consistent with the EXAFS results. We considered the possibility that the N3PyS sulfur donor might be protonated, which could lead to elongation of the Fe-S bond. Accordingly, we investigated [Fe^II(κ¹-N-PhSO₂NOH)(N3PySH)]; a tautomer of [Fe^II(κ¹-N-PhSO₂NOH)(N3PySH)] that could be obtained by internal proton transfer from P.A. to N3PyS. Although this structure exhibits an Fe-S(N3PySH) distance of 2.60 Å, which is similar to the EXAFS-derived distance, it is ~20 kcal mol⁻¹ higher in energy than S-deprotonated structures obtained through fully relaxed optimizations. Thus, we conclude that ⁵A is the most reasonable structure for 2. However, we cannot exclude the possibility that hydrogen bonding or other electrostatic effects (such as an NSPyS▪▪▪K⁺ interaction) could contribute to elongation of the Fe-S bond in solution.

Generation of an {FeNO}⁷ complex from 2.

A rapid color change from bright red to dark brown is observed when samples of 2 are warmed to 23 °C. These changes correspond to loss of the UVvis bands for 2 and the appearance of a species with λ_max = 350, 450,and 550 nm (Figure 7). This spectrum is consistent with Fe(NO)(N3PyS)]BF₄ (4), an {FeNO}⁷ complex previously synthesized.⁵⁹ A representative ¹H NMR spectrum of final reaction mixtures in CD₃CN is shown in Figure 8 (top), and contains a number of well-separated, paramagnetically shifted peaks between 35 and −5ppm. These signals match the spectrum seen for 4 (Figure 8, bottom). No other paramagnetically shifted ¹H resonances are observed across the acquisition window (+130 to −50 ppm).

Figure 7.

UV-vis spectra (CH₃CN, 0.1 mM Fe, −40 °C) comparing Fe^II-solvento starting material 1 (solid black line), Fe^II-P.A. complex 2 (dashed red line), and the thermal decay product(s) of 2 (solid brown line).

Figure 8.

Paramagnetic ¹H-NMR spectra (CD₃CN, 23 °C) showing the thermal decay product of 2 (top) and an authentic sample of the {FeNO}⁷ complex 4 (bottom).

The generation of 4 is further evidenced by attenuated total reflectance infrared (ATR-IR) spectroscopy with ¹⁵N-labeled Piloty’s acid, which was synthesized according to Scheme 5. Samples of 2-⁵⁷Fe were prepared with ¹⁴N-P.A. and ¹⁵N-P.A. according to the standard protocol, yielding identical Mӧssbauer spectra which established the same purity level for both samples (Figures S3). Warming of these samples to 23 °C similarly yielded identical ¹H NMR spectra consistent with the {FeNO}⁷ complex 4. These samples gave the ATR-IR spectra shown in Figure 9 (¹⁴N: black line; ¹⁵N: red line). The ¹⁵N minus ¹⁴N difference spectrum (blue line) contains a prominent band at 1642 cm⁻¹ (positive intensity) that shifts to 1613 cm⁻¹ (negative intensity) upon ¹⁵N substitution. This band is assigned to the v(N-O) mode of 4 and exhibits a downshift of 29 cm⁻¹ that is in excellent agreement with the harmonic oscillator prediction for an isolated N-O stretch (−30 cm⁻¹). A second isotopically sensitive band is observed at 1735 cm⁻¹ (¹⁵N: 1703 cm⁻¹), also arising from 4. These two v(N-O) modes belong to the S = ½ and S = 3/2 spin states of 4, which are both populated at 23 °C (Figure 9, green line).^{55, 59} No other ¹⁴N/¹⁵N shifts are evident in the ATR-IR data. Importantly, these results establish that the nitrosyl ligand in 4 derives from Piloty’s acid.

Scheme 5.

Synthesis of ¹⁵N-Labeled Piloty’s Acid

Figure 9.

ATR-IR spectra following the thermal decay of 2. Top: authentic 4-¹⁴N (green line). Middle: thermal decay of 2-¹⁴N (black line) and 2-¹⁵N (red line), overlaid and normalized relative to the 1467 and 1444 cm⁻¹ N3PyS ligand vibrations (denoted by*). The 1351 cm⁻¹ mode originates from 18-crown-6. Bottom: difference spectrum (¹⁵N minus ¹⁴N, blue line) showing isotopically sensitive modes at 1735 and 1642 cm⁻¹ corresponding to 4.

Mӧssbauer data for the thermal decay of 2-⁵⁷Fe. is presented in Figure 10. Given that the spectrum of 4-⁵⁷Fe. is well-resolved at 80 K in CH₃CN (but not in mixtures of C₃H₇CN/THF), final reaction mixtures in C₃H₇CN/THF (9:1 v/v) were concentrated to dryness under high vacuum before adding CH₃CN. The major species with δ = 0.31 and |ΔE_Q| = 0.57 mm s⁻¹ (78% offit) corresponds to 4-⁵⁷Fe. A minor species with δ = 1.09 and |ΔE_Q| = 3.12 mm s⁻¹ (22% of fit) is also observed, which we tentatively attribute to coordination of the PhSO₂⁻, by-product at Fe. Thus, the combined data (Mӧssbauer, NMR, ATR-IR, and UV-vis) confirm that the major product obtained from warming 2 at −60 °C to 23 °C is the {FeNO}⁷ complex 4.

Figure 10.

⁵⁷Fe Mössbauer spectrum (80 Kin frozen CH3CN) showing the thermal decay of 2-⁵⁷Fe. Experimental data (black dots); best fit (red line); majority {FeNO}⁷ product 4-⁵⁷Fe (dashed purple line, 78%); minor hs Fe^II co-product (solid green line, 22%).

Mechanism.

A mechanism for the formation and decay of Fe-P.A. complex 2 is shown in Scheme 6. Deprotonation of P.A with KO^tBu/18-crown-6 (1 equiv) at-60 °C affords the P.A. monoanion (PhSO₂NOH) Our data suggest that this species displaces CH₃CN from 1, yielding P.A.-bound 2, before HNO release can occur. The combined Mӧssbauer and XAS data, supported by DFT calculations, indicate that the P.A ligand in 2 is coordinated via nitrogen. Addition of a second equivalent of KO^tBu/l8-crown-6 leads to the {FeNO}⁸ complex 3. In the absence of excess base, thermal decay of 2 from −60 °C to 23 °C induces S-N bond cleavage, leading to the {FeNO}⁷ complex 4 (78%). There is precedent for the production of {FeNO}⁷ complexes from Piloty’s acid and related RSO₂NHOH compounds;^60,61 however, these complexes were obtained from iron(III) precursors via HNO /NO⁻ transfer. The production of an {FeNO}⁷ complex from Piloty’s acid, base, and an iron(II) precursor, as described here, is consistent with formation of a highly reactive {Fe(HNO)}⁸ species, which is then oxidized by an as yet unidentified pathway to give a stable {FeNO}⁷ complex. Efforts are ongoing to trap the elusive nonheme {Fe(HNO)}⁸ species.

Scheme 6.

Possible Mechanism of Formation of the Fe^II-PiIoty’s Acid Adduct and Reaction Pathways Leading to {FeNO}⁷ and {FeNO}⁸ Products

CONCLUSIONS

This work describes the characterization of a rare mononuclear, nonheme Fe^II-Piloty’s acid adduct by UV-vis and ⁵⁷Fe Mӧssbauer spectroscopy, Fe K-edge XAS, and DFT. This adduct can be formed at low temperatures from an Fe(II) precursor, P.A., and one equivalent of a strong alkoxide base. Piloty’s acid and related RSO₂NHOH compounds are synthetic precursors to HNO/NO⁻, and are known to react with Fe(III) centers to give {FeNO}⁷ species.^60–61 Much less is known about the reactions of RSO₂NHOH donors with Fe(II) complexes, where the anticipated products are less stable, one-electron reduced {FeNO}⁸ or {Fe(HNO)}⁸ species. In this regard, we have provided strong evidence for an unprecedented reaction where the P.A. donor, rather than free HNO, binds to a nonheme Fe(II) complex, which then ultimately facilitates nitroxyl production on the metal. Our results indicate that binding of monoanionic P.A. to the Fe(II) center is favored over spontaneous S-N bond cleavage to give HNO and benzenesulfinate anion (PhSO₂⁻)

Our findings also have implications for the chemistry of HNO with nonheme iron-thiolate metalloenzymes. The Fe^II precursor used in this study, [Fe^II(CH₃CN)(N₃PyS]BF4, was previously shown to be a structural and functional mimic of the thiol dioxygenases.^{53, 59} It is well known that HNO reacts with free thiols to give sulfonamide (RS(O)NH₂) and (RSSR) products, ¹² but considerably less is known about the reaction of HNO with metal-bound thiols/thiolates. In the present example, we have shown that coordination of P.A. to Fe can lead to either sulfur-ligated {FeNO}⁸ or {FeNO}⁷ species, while the sulfur donor remains unmodified. Thus, formation of a metal-bound adduct might protect sulfur sites from rreversible modification by preventing formation of free HNO from P.A. or similar HNO donors.

EXPERIMENTAL SECTION

General Procedures.

All syntheses and manipulations were conducted in an N₂-filled glovebox (Vacuum Atmospheres, O₂ < 0.2 ppm, H₂O < 0.5 ppm) or using standard Schlenk techniques under an atmosphere of dry Ar unless otherwise noted. ⁵⁷Fe powder (95.5 atom %; 98% purity) was purchased from Cambridge Isotope Laboratories. Piloty’s acid (N-hydroxybenzenesulfonamide) and ¹⁵N-labeled Piloty’s acid were synthesized according to a literature procedure.⁶² Potassium tert-butoxide (KO^tBu, sublimed grade, 99.99% trace metals basis) was purchased from Sigma-Aldrich and stored in the glovebox. 18-crown-6 was purchased from Sigma-Aldrich and purified by recrystallization from hot acetonitrile. The 18-crown-6-acetonitrile solvate thus obtained was dried under highvacuum at 35 °C for several hours to afford pure 18-crown-6. Acetonitrile, acetonitrile-d₃, and toluene were distilled from CaH₂. Tetrahydrofuran was distilled from a purple sodium/benzophenone ketyl solution. Butyronitrile was distilled from Na₂CO₃/KMnO₄ according to a literature procedure.⁶³ Diethyl ether was obtained from a PureSolv solvent purification system (SPS) and used without further purification. All solvents were degassed by a minimum of three freeze-pump-thaw cycles and stored over freshly activated 3 Å molecular sieves in the glove box.

Instrumentation.

UV-visible spectra were recorded on a Varian Cary 50 Bio spectrophotometer integrated with a Unisoku USP-203 cryostat. ¹H NMR spectra were recorded on a Bruker Avance 400 MHz FT-NMR spectrometer at 298 K and referenced against residual solvent proton signals. Attenuated total reflectance (ATR) infrared spectra were obtained with a Golden Gate Reflectance diamond cell in a Nexus 670 Thermo-Nicolet FTIR spectrometer. ⁵⁷Fe Mӧssbauer spectra were collected on a spectrometer from SEE Co. (Science Engineering & Education Co., MN) integrated with a Janis SVT-400-MOSS LHe/LN₂ cryostat. The spectrometer was operated in the constant acceleration mode in a transmission geometry.

[⁵⁷Fe^II(CH₃CN)(N3PySEtCN)](BF₄)₂ (5-⁵⁷Fe).

The title compound was prepared with a ⁵⁷Fe enrichment level of approximately 80%. A three-neck, 100 mL round-bottom flask equipped with a stir bar was fitted with a reflux condenser. Amounts of ⁵⁷Fe powder (33 mg, 0.58 mmol) and ⁵⁶Fe powder (11 mg, 0.20 mmol) were added. A solution of HBF₄ (48 wt. % in H₂O, 0.3 mL, 2.9 equiv) was added via syringe, and the mixture was refluxed for 45 min until no Fe powder remained. The resulting aqueous solution of ⁵⁷Fe-enriched ⁵⁷Fe(BF₄)₂ was cooled to 23 °C and then diluted with CH₃CN (5 mL). The N3PySEtCN ligand (355.1 mg, 0.786 mmol) was dissolved in CH₃CN (10 mL), and the ⁵⁷Fe(BF₄)₂ solution was added at 23 °C, producing an immediate color change from pale orange to dark red-orange. After 2 h, excess Et₂O (20 mL) was added, resulting in precipitation of a brown residue. The supernatant was decanted by cannula, and the residue was dried in vacuo over P₄O₁₀. Slow vapor diffusion of Et₂O into a concentrated CH₃CN solution of the product furnished large red crystals of 5-⁵⁷Fe after 24 h (413 mg, 73% yield).

[⁵⁷Fe^II(CH₃CN)(N3PyS)]BF₄ (1-⁵⁷Fe).

An amount of crystal-line 5-⁵⁷Fe (81 mg, 0.11 mmol) was dissolved with vigorous stirring in CH₃CN mL). A slurry of KO^tBu (13 mg, 0.11 mmol, 1 equiv) in CH³CN (4 mL) was prepared and added dropwise at 23 °C to the solution of 5-⁵⁷Fe, resulting in an immediate color change from dark red-orange to dark wine red. The mixture was stirred for 2.5 h, filtered through a short plug of Celite, and concentrated under reduced pressure. The dark red-brown residue was triturated with Et₂O (3×5 mL) and dried to yield crude 1-⁵⁷Fe. The crude product was dissolved in a mixture of CH₃CN and toluene (1:1 v/v, 4 mL) and filtered through Celite. Excess Et₂O (12 mL) was added, and the turbid mixture was placed in a −35 °C freezer overnight, resulting in precipitation of a red-brown microcrystalline solid. Removal of the yellow-brown supernatant and further drying under reduced pressure yielded pure 1-⁵⁷Fe (42 mg, 65% yield).

⁵⁷Fe Mössbauer Spectroscopy.

⁵⁷Fe Mössbauer spectra were collected on frozen solution samples at 80 K in the absence of an applied magnetic field. ⁵⁷Fe-enriched samples were prepared in custom Delrin cups (10.0 mm x 12.4 mm OD) equipped with tight-fitting Delrin inserts (5.0 mm x 11.0 mm). All samples were stored under LN₂ prior to analysis. The zero velocity of each Mössbauer spectrum refers to the centroid of a 25 μm, metallic iron (α-Fe) foil collected at room temperature. The WMOSS program (SEE Co., formerly WEB Research Co., Edina, MN) was used to analyze the Mӧssbauer data. Data were fit to quadrupole doublets with Lorentzian or Voigt line shapes as specified.

Fe K-Edge X-ray Absorption Spectroscopy.

Fe K-edge XAS and extended X-ray absorption fine structure (EXAFS) data were collected at the 16-pole, 2-T wiggler Beamline 9–3 at the Stanford Synchrotron Radiation Lightsource (SSRL) under ring conditions of 500 mA and 3 GeV. All samples were ⁵⁷Fe-enriched in order to assess purity by Mӧssbauer spectroscopy, and were prepared in custom screw-top Delrin Mössbauer cups. The base of each cup contained a slit (1 × 4 mm) that was covered in 38 μM Kapton to provide an X-ray transparent window, thus allowing for sequential Mӧssbauer and XAS measurements. Samples were maintained at 10 K in an Oxford liquid He flow cryostat. A Si(220) double-crystal monochromator was used for energy selection. A Rh-coated mirror (set to an energy cutoff of 9 key) was used for harmonic rejection. Internal incident energy calibration was achieved by assigning the first inflection points of a Fe foil placed downstream of the sample to 7111.2 eV. Data were collected in fluorescence mode using a Canberra 30-element Ge array detector windowed on Fe Kα emission. Elastic scatter into the detector was attenuated using a Soller slit with an upstream Co filter. Multiple scans were measured and averaged with SIXPACK⁶⁴ software package. No spectral changes due to photo-damage were observed after multiple scans for these complexes. Data were normalized to post-edge jumps of 1.0 at 7130 eV in SIXPACK by applying a Gaussian normalization for the pre-edge and a three-region cubic spline to model the smooth background above the edge. EXAFS data were fitted using the OPT module of EXAFSPAK⁶⁵ using initial scattering paths calculated using FEFF7^66–67.

Generation of 2-⁵⁷Fe.

A dark red solution of 1-⁵⁷Fe (5 mg, 9 μmol) in C₃H₇CN (1.6 mL) was combined with a stir bar in a 10 mL Schlenk flask. The flask was cooled to −60 °C in a slurry of dry ice and chloroform. A stock solution of Piloty’s acid (9 mg, 53 μmol in C₃H₇CN (1.15 mL) was prepared, and an aliquot (200 μL, 9 μmol, 1 equiv per Fe) was loaded into a gas-tight syringe. A stock solution of KO^tBu (6 mg, 50 timol) and 18-crown-6 (13 mg, 50 μmol in THF (1.1 mL) was prepared, and an aliquot (200 μL, 9 μmol, 1 equiv per Fe) was loaded into a separate gas-tight syringe. The Piloty’s acid solution was added dropwise over 10 min to the solution of 1-⁵⁷Fe at −60 °C, and stirring was continued for an additional 5 min. Next, the KO^tBu/18-crown-6 solution was added dropwise over 10 min. Stirring was continued for 2 h, over which time the color of the reaction mixture turned from dark wine red to bright pink red. For Mӧssbauer spectroscopy, the reaction mixture was poured into N₂(1), and the resulting frozen powder was packed into a sample cup under N₂(1). Alternatively, the reaction mixture was further cooled to −98°C, and transferred by pre-cooled syringe to a pre-cooled sample cup. Samples for Fe K-edge X-ray absorption spectroscopy were prepared in an identical manner.

Reaction of 1-⁵⁷Fe with Previously Reacted 1:1 Mixture of Piloty’s Acid and KO^tBu/18-crown-6 at 23 °C.

Piloty’s acid (15 mg, 87 μmol) was dissolved in C₃H₇CN (1.0 mL). Amounts of KO^tBu (11 mg, 94 μmol) and 18-crown-6 (24 mg, 90 μmol) were combined and dissolved in THF (1.0 mL). The solution of KO^tBu/18-crown-6 was added to the solution of Piloty’s acid with stirring, resulting in a dark yellow mixture. The color of the reaction mixture rapidly bleached within ⁓5 s. Stirring was continued for 48 h at 23 °C, at which point the mixture appeared colorless and slightly turbid.

An amount of 1-⁵⁷Fe (2 mg, 3 μmol) was dissolved in C₃H₇CN (600 μL). An aliquot of the previously reacted P.A./KO^tBu/18-crown-6 mixture (80 μL; 3 μmol, 1 equiv per Fe based on initial P.A.) was added dropwise to 1-⁵⁷Fe, resulting in an immediate color change from dark wine-red to bright pink-red. For Mössbauer spectroscopy, an aliquot of the reaction mixture (400μL) was transferred to a sample cup and frozen in N₂(1).

Reaction of 1-⁵⁷Fe with NaPhSO₂. at 23 °C

An amount of 1-⁵⁷Fe (2 mg, 3 Etmol) was dissolved in a mixture of C₃H₇CN (500 μL) and THF (50 μL). Sodium benzenesulfinate (11 mg, 66 μmol) and 18-crown-6 (18 mg, 70 μmol) were combined with C₃H₇CN (1.0 mL), and the resulting slurry was vigorously stirred for 1 h. An amount of the slurry (60 μL; 4 μmol Na⁺PhSO₂⁻) was added dropwise to 1-⁵⁷Fe with continued stirring, producing an immediate color change from dark wine-red to bright pink-red. For Mӧssbauer spectroscopy, an aliquot of the reaction mixture (400 μL) was transferred to a sample cup and frozen in N₂(1).