O₂ Activation by Bis(imino)pyridine Iron(II)-Thiolate Complexes

Abstract

The new iron(II)-thiolate complexes [(^iPrBIP)Fe^II(SPh)(Cl)] (1) and [(^iPrBIP)Fe^II(SPh)(OTf)] (2) (BIP = bis(iminopyridine)) were prepared as models for cysteine dioxygenase (CDO), which converts Cys to Cys-SO₂H at a (His)₃Fe^II center. Reaction of 1 and 2 with O₂ leads to Fe-oxygenation and S-oxygenation, respectively. For 1 + O₂, the spectroscopic and reactivity data, including ¹⁸O isotope studies, are consistent with an assignment of an iron(IV)-oxo complex as the product of oxygenation ([(^iPrBIP)Fe^IV(O)(Cl)] (3)). In contrast, 2 + O₂ results in direct S-oxygenation to give a sulfonato product (PhSO₃⁻). The positioning of the thiolate ligand in 1 versus 2 appears to play a critical role in determining the outcome of O₂ activation. The thiolate ligands in 1 and 2 are essential for O₂ reactivity, and exhibit an important influence over the Fe^III/Fe^II redox potential.

Determining the factors that govern the activation of dioxygen by both heme and nonheme iron metalloenzymes is of fundamental importance. Mononuclear nonheme iron oxygenases typically contain a 2-His-1-carboxylate ligand set bound to the catalytic iron center. An interesting exception is cysteine dioxygenase (CDO), which utilizes a (His)₃Fe^II(H₂O) center to activate O₂ and oxidize cysteine to sulfinic acid (CysSO₂H), a key metabolic process vital for human health.¹ Despite the importance of CDO from a health perspective, little is known about the mechanism of this dioxygenase.² In general, the oxidation of Cys to disulfide, sulfenic acid (Cys(O)H) and other oxidized products has been implicated in oxidative stress response.³ Thus understanding the fundamental mechanistic pathways of biologically relevant sulfur oxidations is of high current interest.⁴

Although many studies on iron(II) model complexes have yielded key insights into the reactivity of nonheme iron centers, relatively few have involved the use of O₂ as the oxidant, in part because of the inherent difficulties with activating and controlling O₂.⁵ In an earlier report, we described the synthesis of an N₃S(thiolate)Fe^II model complex of CDO, which contains the 3 neutral N binding motif found in the enzyme, and reacts with O₂ selectively to yield an S-oxygenated sulfonato product.⁶ The thiolate donor was covalently tethered to a bis(imino)pyridine (BIP) framework, in part to favor S-oxygenation as opposed to disulfide formation. To our knowledge, this reaction was the first example of an Fe^II-thiolate complex reacting with O₂ to give S-, as opposed to Fe-oxygenation (e.g., Fe^III-O-Fe^III species).⁷

Herein we report the synthesis of two new unsymmetrical Fe^II-thiolate BIP complexes, [(^iPrBIP)Fe^II(SPh)(Cl)] (1) and [(^iPrBIP)Fe^II(SPh)(OTf)] (2) (^iPrBIP = 2,6-(ArN=CMe)₂C₅H₃N), Ar = 2,6-ⁱPr₂C₆H₃), in which the thiolate ligands are not covalently tethered to the BIP framework. The reactivity of these complexes toward O₂ has been examined together with non-thiolate-ligated analogs. We show that coordination of the thiolate ligands is crucial for O₂ activation by (BIP)Fe^II. We also show that S-oxygenation is possible for a terminal thiolate, and furthermore that the positioning of the thiolate donor specifies the outcome of oxygenation at either sulfur or iron.

Significant efforts have gone into the synthesis and study of (BIP)Fe complexes for their use in N₂ activation and catalysis.⁸ However, unsymmetrical derivatives of formula [(BIP)Fe^II(X)(Y)] (X ≠ Y) are scarce. Careful control of stoichiometry, together with the appropriate conditions (solvent, temperature), allowed for the isolation of the monothiolato complexes 1 and 2 (Figure 1). The molecular structures of 1 and 2 reveal 5-coordinate Fe^II ions with the desired single terminal thiolate ligands bound to the iron. Bond distances and angles are consistent with high-spin Fe^II BIP complexes.^8a,d,e A distinguishing feature of the structures of 1 and 2 is the positioning of the thiolate ligand. In complex 1, the PhS⁻ group sits in a pseudo-axial position in relation to the N₃Cl plane, trans to the open coordination site that subtends the obtuse N1-Fe-N3 angle (141.2°). This positioning may be aided by a π-stacking interaction between the pyridine and PhS⁻ groups. In contrast, the PhS⁻ ligand of 2 is bound in a pseudo-equatorial arrangement with the ^iPrBIP ligand, and is cis to the open coordination site.

Figure 1.

Synthetic scheme and displacement ellipsoid plots (50% probability level) for 1 and 2 at 110 K. H atoms are omitted for clarity.

Both 1 and 2 exhibit relatively sharp, paramagnetically shifted peaks in the ¹H NMR spectrum (CD₂Cl₂) typical of high-spin (BIP)FeX₂ complexes, and these spectra are consistent with their solid-state structures. Magnetic susceptibility for 1 measured by Evan’s method in CD₂Cl₂ gives μ_eff = 5.2 μ_B, close to the spin-only value for a high-spin Fe^II (S = 2) ion.

Reaction of 1 (10 – 20 mM) with a slight excess of dry O₂ (5 equiv) leads to a color change from dark blue to green over the course of 1 h. A decrease of the λ_max for 1 at 715 nm (ε ~ 4000 M⁻¹ cm⁻¹) is observed, and a new band for the green species appears at λ_max 690 nm (ε ~ 1500 M⁻¹ cm⁻¹) (Figure 2; for time-dependence see Fig. S5). This spectrum is similar to that reported for a closely related bis(imino)pyridine iron(IV)-oxo complex (λ_max 660 nm, ε ~ 1200 M⁻¹ cm⁻¹).⁹ Analysis by laser-desorption ionization mass spectrometry (LDIMS(+)) reveals a dominant isotopic cluster at m/z = 588, whose isotope and fragmentation pattern (Figs. 2, S8 and S9) are consistent with an Fe^IV(O) complex, [(^iPrBIP)Fe^IV(O)(Cl)]⁺ (3). The thiolate ligand is oxidized to disulfide during the production of 3, as determined by ¹H NMR (PhS-SPh, 85%). Introduction of ¹⁸O₂ in place of ¹⁶O₂ causes a shift of two mass units for the LDIMS of 3, giving m/z = 590 (80% ¹⁸O incorporation). Finally, green 3 is EPR-silent (X-band, 15 K). These data are consistent with the assignment of 3 as an Fe^IV(O) species.

Figure 2.

a) UV-vis spectral changes for the reaction of 1 (715 nm, 0.370 mM) with excess O₂ in CH₂Cl₂, leading to formation of 3 (690 nm). b) LDIMS of 3 formed in the reaction of 1 + O₂. Peaks at m/z of 588 and 572 correspond to [(^iPrBIP)Fe^IV(O)(Cl)]⁺ and [(^iPrBIP)Fe^II(Cl)]⁺, respectively. Inset: isotopic cluster for 3 prepared from ¹⁶O₂ (top) and ¹⁸O₂ (bottom).

If the reaction of 1 with excess O₂ in CH₂Cl₂ is carried out in the presence of PPh₃ (5 equiv), OPPh₃ is produced in good yield (70%, ³¹P NMR). Alternatively, formation of green 3, followed by removal of O₂ under vacuum and addition of PPh₃ (50 – 300 equiv) under Ar, results in the smooth decay of the peak for 3 at 690 nm (Figure S6). This decay follows good pseudo-first-order behavior, and the rate constants (k_obs) thus obtained were found to increase linearly with [PPh₃], yielding a second-order rate constant of k₂ = 3.6 ± 0.3 × 10⁻³ M⁻¹ s⁻¹ for oxygen-atom-transfer from 3 to PPh₃ (Figure S7). This relatively slow reactivity may in part be due to the steric encumbrance imposed by the 2,6-iPr₂-C₆H₃ substituents. The ¹⁸O-labeled 3 produces ¹⁸OPPh₃ with modest isotopic incorporation (¹⁶O:¹⁸O 85:15). However, addition of excess H₂ ¹⁸O to the reaction of 3-¹⁶O and PPh₃ results in a significant increase in the isotopically labeled product ¹⁸OPPh₃ (50% ¹⁸O) (eq 1). These data indicate that the O atom in 3 undergoes facile exchange with exogenous H₂O, as seen for other iron terminal oxo species.¹⁰ Although further spectroscopic studies are needed to definitively characterize the structure of 3, all of the spectroscopic data and reactivity presented here strongly

(1)

support the formulation of 3 as a terminal iron-oxo complex generated from 1 + O₂, with the PhS⁻ ligand undergoing concomitant oxidation to disulfide.

The formation of nonheme Fe^IV(O) complexes from Fe^II and O₂ can be induced by the addition of external co-reductants (e.g. cyclohexene or NADH).^5b,c,e In the case of 1, the thiolate ligand functions as a built-in co-reductant to assist in the activation of O₂. In comparison, the covalently-tethered thiolate complex [Fe^II(N₃S(thiolate))(OTf)] (4) also serves to activate dioxygen, but in this case participation from sulfur leads to direct oxygenation of the S atom.⁶

To our surprise, the addition of stoichiometric amounts of O₂ to the triflate complex 2 follows a very different oxidation pathway than followed by the chloro analog 1. An immediate color change from dark blue to brown is noted upon addition of O₂, but LDIMS reveals a cluster at m/z 694 corresponding to S-oxygenated [Fe^II(^iPrBIP)(PhSO₃)]⁺. Attempts to crystallize [Fe^II(BIP)(PhSO₃)]⁺ have led thus far only to the crystallization of the known Fe^II(^iPrBIP)(OTf)₂ complex, but the production of benzenesulfonic acid was readily confirmed by ¹H NMR, and quantitation by reverse-phase HPLC after hydrolytic workup gave a yield of 30% for PhSO₃H (based on total Fe). The use of labeled ¹⁸O₂ results in ~90% incorporation of ¹⁸O into the PhSO₃⁻ ligand. Despite the fact that the thiolate donor in 2 is not part of a chelate ring, sulfur oxygenation does occur, as seen for the covalently tethered 4. In contrast, no evidence for PhSO₃H was detected by LDIMS or HPLC for 1 + O₂ in control experiments.

The reactivity of the related non-thiolate-ligated complexes Fe(^iPrBIP)Cl₂ (5) and Fe(^iPrBIP)(OTf)₂ (6) was next examined for comparison with 1 and 2. These complexes are completely inert toward O₂ in both solution (e.g. CH₂Cl₂, CH₃CN) and the solid-state (eq 2). Addition of PPh₃ to oxygenated solutions of 5 and 6 showed no formation of OPPh₃. The incorporation of a thiolate donor thus clearly plays a critical role in the activation of O₂ by these nonheme iron(II) complexes.

$$\begin{array}{cc}\hfill & \left[{\mathrm{Fe}\left(\mathrm{BIP}\right)\left(X\right)}_2\right]\underset{{\mathrm{CH}}_2{\mathrm{Cl}}_2}{\overset{{\mathrm{O}}_2}{\to }}\text{no reaction}\hfill \\ \hfill & \mathrm{X}=\mathrm{Cl},5;\mathrm{X}=\mathrm{OTf},6\hfill \end{array}$$ (2)

The redox potentials of 1, 2, 5, and 6 are compared in Table 1. The thiolate-ligated complexes exhibit significantly lower redox potentials than the non-thiolate analogs, correlating nicely with their relative O₂ reactivities. A similar correlation was made for [Fe^II(TMC)(OTf)₂] (TMC = 1,4,8,11-tetramethyl-1,4,8,11-tetraazacyclotetradecane), which exhibits a solvent-dependent redox potential, and reacts with O₂ to give an Fe^IV(O) complex only in solvents where E_1/2(Fe^III/II) < −0.1 V (e.g. THF).^5a Similarly, nonheme iron(II) complexes with more positive E_1/2 values fail to react with O₂ to give oxoiron(IV) species. From Table 1, an E_1/2(Fe^III/II) < −0.1 V appears to be a prerequisite for O₂ activation in nonheme iron(II) complexes, and inclusion of a single thiolate donor is sufficient to lower the redox potential of (^iPrBIP)Fe^II complexes into this range. It should be noted that the E_1/2 values for 1 and 2 remain more than 1 V above the one-electron reduction potential for the O₂/O₂⁻ couple in organic solvents,^5h ruling out an outer-sphere mechanism of O₂ activation.

Table 1.

Redox potentials for (^iPrBIP)Fe^II and related nonheme Fe^II complexes.

Compound	E1/2 (ΔEpp)a	O2 reactivity
[(iPrBIP)FeII(SPh)(Cl)] 1	−0.173b (0.114) (r)	yes
[(iPrBIP)FeII(SPh)(OTf)] 2	−0.372b (0.149) (r)	yes
[(iPrBIP)FeII(Cl)2] 5	0.025b (0.153) (r)	no
[(iPrBIP)FeII(OTf)2] 6	0.613b,c (ir)	no
[(TMC)FeII(OTf)2]d	−0.14e (qr)	yes
[(TMC)FeII(OTf)2]d	0.02f (r)	no
[(TPA)FeII]2+, d,g	0.36h (r)	no

^aV vs Fc⁺/Fc; ΔE_pp = peak-to-peak separation; r = reversible, ir = irreversible, qr = quasi-reversible.

^bIn CH₂Cl₂, scan rate of 0.1 V/s.

^cAnodic peak potential.

^dRef. 5a.

^eIn MeCN/THF (1:1).

^fIn MeCN/CH₂Cl₂ (1:1).

^gTPA: tris(2-pyridylmethyl)amine.

^hIn neat MeCN.

Given the structural and electronic similarities between the two thiolate-ligated complexes 1 and 2, why does their reactivity with O₂ follow such dramatically different paths? Scrutiny of the structures of 1 and 2 appears to hold the key. The PhS⁻ ligand in 1 is bound trans to the open site available for O₂ binding, whereas it is bound cis in 2. A plausible mechanism for O₂ activation in 1 thus begins with coordination of O₂ to the open site trans to the thiolate donor, followed by electron-transfer from both the iron and sulfur centers to the bound O₂ (Scheme 1a). In this case, intramolecular attack of an Fe-O₂ intermediate on the sulfur donor would be strongly disfavored by the trans orientation of the PhS⁻ ligand. In contrast, the analogous Fe-(O₂) intermediate in 2 would be generated cis to the thiolate ligand, providing a facile pathway for intramolecular S-oxygenation (Scheme 1b). Similarly, the thiolate donor in the covalently-tethered 4 is also found cis to the open coordinate on site.

Scheme 1.

Proposed mechanisms of O₂ activation by 1 and 2.

This hypothesis depends upon the feasibility of attaining a 6-coordinate structure with the sterically encumbered BIP ligand in 1 and 2. For less bulky BIP analogs, where Ar =2,6-Me₂-C₆H₃, 6-coordinate Fe^II complexes are known,^8d but to our knowledge there are no examples with Ar = 2,6-ⁱPr₂-C₆H₃. Thus we were pleased to isolate [Fe^II(^iPrBIP)(H₂O)₂(NCCH₃)](OTf)₂ (7) as a product from reactions of 2 + O₂, whose molecular structure is given in Figure 3. Despite the large steric encumbrance provided by the flanking 2,6-ⁱPr₂C₆H₃ substituents, a 6-coordinate geometry is clearly attainable in 7.

Figure 3.

Displacement ellipsoid plots (left, 50% probability level) and molecular structure (right) of 7. H atoms, except for those attached to the water molecules, and the OTf⁻ ions have been omitted for clarity.

In summary, we have demonstrated that a thiolate donor is essential for the activation of O₂ by nonheme iron (BIP)Fe^II complexes, and can serve as either a co-reductant or as a site for O-capture. The relative positioning of the PhS⁻ ligand in relation to the potential O₂ binding site appears to play a critical role in determining whether oxygenation occurs at iron or sulfur.¹¹ It is also shown that S-oxygenation can occur for terminal, iron-bound thiolates, contrary to established precedent. It has been proposed that the Cys substrate in CDO coordinates to the Fe center through a chelate ring involving sulfur and the amino group.¹ The findings presented here suggest that this unusual binding mode for Cys is not required for S-oxygenation to occur.