A Nonheme Iron(III) Superoxide Complex Leads to Sulfur Oxygenation

Abstract

A new alkylthiolate-ligated nonheme iron complex, Fe^II(BNPA^Me2S)Br (1), is reported. Reaction of 1 with O₂ at −40 °C, or reaction of the ferric form with O₂^•− at −80 °C, gives a rare iron(III)-superoxide intermediate, [Fe^III(O₂)(BNPA^Me2S)]⁺ (2), characterized by UV–vis, ⁵⁷Fe Mössbauer, ATR-FTIR, EPR, and CSIMS. Metastable 2 then converts to an S-oxygenated Fe^II(sulfinate) product via a sequential O atom transfer mechanism involving an iron-sulfenate intermediate. These results provide evidence of the feasibility of proposed intermediates in thiol dioxygenases.

The activation of dioxygen by nonheme iron enzymes is critical to a range of biological processes.^1–5 One such class of nonheme iron enzymes are the thiol dioxygenases (TDOs), which convert thiol substrates into S-oxygenated compounds and are found in a wide range of organisms, from bacteria,^6,7 to plants,^8,9 to mammals.^10,11 The TDOs play critical roles in biochemical processes,^12–16 and their misfunctioning has been implicated in neurodegenerative diseases,^17,18 autoimmune disorders,^19,20 and cancer.^21–23 The TDOs utilize a mononuclear iron center bound by an unusual 3-His structural motif, as opposed to the canonical 2-His-1-carboxylate iron site found in other nonheme iron enzymes^24–29. Despite their widespread biochemical significance, the mechanism of S-oxygenation for TDOs remains poorly understood.^30–32 For example, two hypothetical mechanisms have been proposed for cysteine dioxygenase (Scheme 1), a mammalian TDO that converts cysteine to cysteine sulfinate.^1,33–35 However, experimental evidence for either mechanism remains scarce.^36–40 In synthetic systems, significant efforts have been made to bind and activate O₂ with nonheme iron complexes.^41–49 These efforts have yielded only three examples of Fe(O₂) adducts, and none of these adducts give dioxygenated sulfinate products.^41,45,46

Scheme 1.

Possible Mechanisms for S-Oxygenation in CDO, an Example of the Broader Class of Enzymes Known as TDOs

Herein, a new tetradentate ligand, (bis((6-(neopentylamino)pyridinyl)methyl)amino)-2-methylpropane-2-thiol (BNPA^Me2SH) is reported, designed to provide three neutral nitrogen donors and one anionic thiolate donor, as found in the TDOs. The second coordination sphere is engineered to stabilize metal–oxygen intermediates through the inclusion of hydrogen bonding substituents. An iron(II) complex, Fe^II(BNPA^Me2S)(Br) (1), reacts with O₂ to give a new Fe^III(O₂^•−) species, which was trapped and characterized at low temperature. This ferric superoxide complex converts to an Fe^II(sulfinate) product upon warming, and mechanistic analyses indicate that a mono-oxygenated sulfenate intermediate must be on the pathway to the dioxygenated product.

An abbreviated synthesis of the new tetradentate ligand BNPA^Me2SH, together with the preparation of the nonheme iron complex Fe^II(BNPA^Me2S)Br (1), is shown in Scheme 2. The secondary amine was synthesized by a Fukuyama amine synthesis,⁵⁰ followed by the installation of an alkylthiolate donor from dimethylthiirane, and finally LAH reduction to the target ligand BNPA^Me2SH. Deprotonation of the thiol followed by metalation gives Fe^II(BNPA^Me2S)Br (1). Complex 1 was crystallized from pentane/THF at −35 °C, giving yellow blocks suitable for single-crystal X-ray diffraction. The structure of 1 (Figure 1) reveals a 5-coordinate, iron(II) complex with a bromide ligand occupying the fifth site. The neopentyl amine groups are oriented toward the bromide ligand, forming two intramolecular hydrogen bonds with N1(H)–Br = 3.404(5)°, N1–H···Br1 = 161(5)°, and N5(H)–Br = 3.361(5) Å, N5–H···Br1 = 154(6)°. The ¹H nuclear magnetic resonance (NMR) spectrum of 1 in CD₃CN reveals paramagnetically shifted peaks ranging from +100 to −20 ppm, consistent with a high-spin (S = 2) Fe^II complex. Mössbauer spectroscopy on ⁵⁷Fe-labeled 1 in THF at 80 K shows a sharp quadrupole doublet with δ = 0.96 and $\left|\Delta E_Q\right|=2.80$ mm s⁻¹, confirming a high-spin ferrous ion (Figure 2).⁵¹

Scheme 2.

Synthesis of 1

Figure 1.

Displacement ellipsoid plot (50% probability level) for 1 at 110(2) K. Hydrogen atoms (except for N–H) have been omitted for the sake of clarity.

Figure 2.

Zero-field ⁵⁷Fe Mössbauer spectra of (a) 1, 80 K, (b) 2, 80 K, (c) 2, 5 K, and (d) final reaction mixture after warming up 2 to 298 K in THF, 80 K. Isomer shift (δ) and quadrupole splitting ($\left|\Delta E_Q\right|$) values are given in mm s⁻¹. Solid lines represent the best fits.

The pale-yellow solution of 1 in THF exhibits a UV–vis absorption band at 330 nm (ε = 9600 M⁻¹ cm⁻¹). Exposure of 1 to O₂ in THF, 2-MeTHF, or CH₃CN at −40 °C generates a dark green species (2) that features an absorption band at 620 nm (ε = 1400 M⁻¹ cm⁻¹), which decays upon warming to 23 °C. Spectral titration of 1 with an O₂-saturated solution of THF at −40 °C shows an Fe/O₂ 1:1 binding stoichiometry, consistent with the rapid formation of a mononuclear Fe(O₂) adduct, 2 (Figure 3). Bubbling argon for 1 h through a solution of 2 results in no spectral change, and the formation of 2 cannot be reversed by repeated vacuum/purge cycles. Thus, the binding of the aqueous O₂ to 1 is irreversible at −40 °C.

Figure 3.

UV–vis spectra showing the conversion of 1 (0.5 mM, black line) to 2 (red line) after stoichiometric O₂ addition in THF at −40 °C.

Examination of ⁵⁷Fe-labeled 2 in THF by Mössbauer spectroscopy reveals a single quadruple doublet with parameters δ = 0.45 mm s⁻¹ and $\left|\Delta E_Q\right|=1.25$ mm s⁻¹, accounting for >90% of the total area (Figure 2b). Analysis at 5 K in the absence of a magnetic field shows a quadrupole doublet with the same parameters but narrower line widths (Figure 2c). The Mössbauer data are consistent with an integer spin ground state arising from a spin-coupled iron(III)-superoxide species.^52,53 The X-band EPR spectrum for 2 (15 K) is silent, providing further evidence of an integer spin ground state. Solution-phase magnetic measurements (Evans method) for 2 in CD₃CN indicate an apparent S = 2 spin ground state with μ_eff = 4.36 μ_B, assuming a concentration of 2 equal to the total iron content. This ground state can be explained by antiferromagnetic coupling between a high-spin Fe^III ion (S = 5/2) and an O₂^•− ion (S = 1/2).

Analysis of 2 by cryospray ionization mass spectrometry (CSI-MS) shows an intense cluster at 544.23 m/z (Figure S6), consistent with the addition of two oxygen atoms and loss of one bromine atom from 1. Upon addition of the isotopically pure ¹⁸O₂ (99% ¹⁸O) to 1, the center of the cluster shifts to 548.22 m/z, and fitting of the isotope pattern indicates 63% incorporation of ¹⁸O.

An alternative pathway to intermediate 2 involves potassium superoxide (Scheme 3). Reaction of 1 with the one-electron oxidant FcBAr^F₄ in THF at −80 °C leads to a new UV–vis feature at 710 nm (Figure S9), representative of the new species 3. Mössbauer spectroscopy on ⁵⁷Fe-labeled 3 (Figure S10) reveals a broad quadrupole doublet with δ = 0.39 mm s⁻¹ and $\left|\Delta E_Q\right|=1.04$ mm s⁻¹, typical for a high-spin iron(III) ion, and assigned to [Fe^III(BNPA^Me2S)(Br)]⁺. Addition of a slight excess of KO₂ and Kryptofix 222 in THF/CH₃CN (~20/1) to 3 at −80 °C gives a new UV–vis feature at 620 nm, which matches the intermediate 2 generated from 1/O₂. Mössbauer analysis following addition of KO₂ to ⁵⁷Fe-labeled 3 also matches that seen for 2.

Scheme 3.

Formation of 2 from 1 and 3

Characterization of 2 by resonance Raman spectroscopy was attempted with both red and blue excitations, but no ¹⁶O/¹⁸O sensitive bands were observed. However, we were able to successfully isolate 2 in the solid state at a low temperature, which allowed for interrogation by ATR-FTIR spectroscopy. Precipitation of 2 at −80 °C as a green solid, followed by rapid transfer to the ATR stage under cold N₂ gas led to the spectrum in Figure 4. This spectrum revealed intense bands in the 1300–1500 cm⁻¹ region for the pyridyl groups,⁵⁴ as well as an isotopically sensitive peak at 1061 cm⁻¹, which shifts to 1008 cm⁻¹ with ¹⁸O₂ (99%). The band at 1061 cm⁻¹ is assigned to an O–O stretch which downshifts by 53 cm⁻¹ with ¹⁸O substitution, in good agreement with a harmonic oscillator. The frequency of the O–O stretch matches that for a metal-superoxide species (Table S3).^55–57

Figure 4.

ATR-FTIR spectra of 1 (dotted line), 2(¹⁶O₂ natural abundance) (black line), 2(¹⁸O₂ (99%)) (red line), difference spectrum for 2(¹⁶O₂ – ¹⁸O₂) (blue line); 4(¹⁶O₂ natural abundance) (green line), and 4(¹⁸O₂ (99%)) (purple line).

The density functional theory (DFT) optimized geometry for 1 matches the crystal structure and was employed as the starting point for calculations on 2. The η²-O₂ (side-on) and η¹-O₂ (end-on) structures were evaluated, both in an S = 2 ground state, as seen experimentally. The optimized end-on structure is 6-coordinate, with the bromide ion coordinated in the open site, whereas for the side-on geometry, the bromide ion is expelled from the coordination sphere during optimization. If the bromide ligand is removed from the starting structure with the η¹-O₂ binding mode, then the structure converges to the side-on geometry. Mössbauer parameters calculated for the side-on structure give a significantly better match with experimental results (Figure 5). Time-dependent DFT calculations predict a thiolate-to-Fe(O₂) charge transfer transition at 627 nm for the side-on structure, as opposed to 500 nm for the end-on structure, providing further support for an η² binding mode (Table S4). These calculations support the formation of a side-on iron(III)-superoxide intermediate; however, further experimental evidence is needed to confirm the binding mode.

Figure 5.

DFT geometry optimized structures and Mössbauer parameters calculations (δ, $\left|\Delta E_Q\right|\left(mm{s}^{-1}\right))$ for side-on vs end-on binding modes of 2 in the S = 2 spin state.

Complex 2 in THF is stable at −40 °C, but upon warming to 23 °C, this species decays into two new iron products, as observed by Mössbauer spectroscopy (Figure 2d). Fitting of the spectrum is achieved with two quadrupole doublets: one with δ = 0.47 mm s⁻¹, $\left|\Delta E_Q\right|=0.77$ mm s⁻¹, consistent with high-spin Fe^III, and one with δ = 1.23 mm s⁻¹, $\left|\Delta E_Q\right|=2.96$ mm s⁻¹, consistent with high-spin Fe^II. One of the products was identified by crystallization, Fe^III(BNPA^Me2S)(OH)(Br) (5) (Figure S20, Scheme 4). Mössbauer spectroscopy on crystalline 5 in THF gives a single quadrupole doublet that matches the ferric component of the final reaction mixture. The EPR spectrum for 5 also matches that of high-spin iron(III) (Figure S21). A possible pathway for the formation of 5 involves release of O₂ upon warming, and then reaction of the iron(II) complex with O₂ at room temperature leads to the Fe^III(OH) product, as seen in a previous system.^58–60 Addition of O₂ to 1 at room temperature leads to the Fe^III(OH) product in approximately 60% yield, as well as an unidentified iron(II) species that has similar Mössbauer parameters to 4, but does not show any S═O vibrational modes in FTIR (Figure S23).

Scheme 4.

Thermal Decay of 2

Further analysis of the final products by ATR-FTIR shows the loss of ν(O–O) at 1061 cm⁻¹ and the appearance of two new absorption bands at 1233 and 1207 cm⁻¹, which fall in the range of metal-sulfinate stretches (Figure 4).^61–69 Warm up of the ¹⁸O₂-labeled 2 leads to downshifts in these vibrations to 1215 and 1189 cm⁻¹. These shifts are consistent with the DFT calculated shifts for RSO₂–M vibrations.⁴⁴ The Mössbauer and IR data together point to the formation of an Fe^II(sulfinate) complex as one of the final products following oxygenation. Laser desorption ionization mass spectrometry (LDI-MS) of the final reaction mixture revealed a prominent isotopic cluster centered at 544.5 m/z, which matches that of [Fe^II(BNPA^Me2SO₂)]⁺ (4). Addition of ¹⁸O₂ leads to a shift in the isotopic cluster, and the best simulations of the isotopic pattern involve a mixture of S¹⁶O₂ (40%), S¹⁶O¹⁸O (40%), and S¹⁸O₂ (20%). These results suggest that O₂ is a source of O atoms for the S-oxygenated product, but there is significant exchange with exogenous H₂O prior to formation of the final product. Addition of H₂¹⁸O to the reaction with natural abundance O₂ leads to S¹⁸O₂ (85%) and S¹⁶O¹⁸O (13%), while addition of H₂¹⁸O to the final Fe(sulfinate) product does not lead to incorporation of any ¹⁸O isotope. Addition of H₂¹⁸O also does not lead to any ¹⁸O-labeling in the Fe^III(O₂^•−) species. These data support a mechanism which involves the formation of one or more intermediates after Fe^III(O₂^•−) that are susceptible to H₂¹⁸O exchange.

A proposed mechanism for S-oxygenation is shown in Scheme 4 and follows similar proposals for the TDOs. After Fe^III(O₂^•−) formation, S–O bond formation and O–O homolytic cleavage occur, giving a putative Fe^IV(O)(sulfenate) intermediate. This intermediate can be susceptible to exchange with H₂¹⁸O; both Fe^IV(O) and sulfenate species^70,71 are known to exchange with H₂O in synthetic systems.^72,73 Some evidence for CDO indicates H₂¹⁸O does not lead to isotopic labeling in the sulfinate product, but in this case, the nature of the enclosed active site may prevent this exchange.^38,74 To test this mechanism further, we added the known sulfenate trap dimedone. LDI-MS analysis shows a peak at 650.4 m/z, corresponding to the expected dimedone adduct (Figure S17). Addition of 4-chloro-7-nitrobenzofurazan (NBD-Cl), another sulfenate trap,⁷⁵ gives a peak at 727.5 m/z, matching with the S-oxygenated NBD adduct, which is further confirmed with ATR-FTIR experiments (Figure S19).^67,76 Control reactions show that these sulfenate adducts only form during the reaction of 1 with O₂ and can be specifically labeled using ¹⁸O₂ gas (see Supporting Information).

In conclusion, a rare sulfur-ligated Fe(O₂) adduct can be obtained from either 1) Fe^II/O₂ or 2) Fe^III/O₂^•−. The spectroscopic and computational results for this species satisfyingly converge, leading to the conclusion that this complex is a mononuclear iron(III)-superoxide species. Warming of this metastable intermediate leads to the production of the S-oxygenated Fe^II(sulfinate) product, in direct analogy with the reactivity seen in the TDOs. A second intermediate containing a mono-oxygenated sulfenate moiety was also captured and, combined with the superoxide complex, provides the first full set of experimental data for a pathway 1 mechanism (Scheme 1), in either synthetic or enzymatic S-oxygenation. Future studies will focus on this and other mechanistic questions.