Ferryl Protonation in Oxoiron(IV) Porphyrins and Its Role in Oxygen Transfer

Abstract

Ferryl porphyrins, P−Fe^IV=O, are central reactive intermediates in the catalytic cycles of numerous heme proteins and a variety of model systems. There has been considerable interest in elucidating factors, such as terminal oxo basicity, that may control ferryl reactivity. Here, the sulfonated, water-soluble ferryl porphyrin complexes tetramesitylporphyrin, oxoFe^IVTMPS (FeTMPS-II), its 2,6-dichlorophenyl analogue, oxoFe^IVTDClPS (FeTDClPS-II), and two other analogues are shown to be protonated under turnover conditions to produce the corresponding bis-aqua-iron(III) porphyrin cation radicals. The results reveal a novel internal electromeric equilibrium, P−Fe^IV=O ⇆ P⁺−Fe^III(OH₂)₂. Reversible pK^a values in the range of 4–6.3 have been measured for this process by pH-jump, UV–vis spectroscopy. Ferryl protonation has important ramifications for C–H bond cleavage reactions mediated by oxoiron(IV) porphyrin cation radicals in protic media. Both solvent O–H and substrate C–H deuterium kinetic isotope effects are observed for these reactions, indicating that hydrocarbon oxidation by these oxoiron(IV) porphyrin cation radicals occurs via a solvent proton-coupled hydrogen atom transfer from the substrate that has not been previously described. The effective FeO–H bond dissociation energies for FeTMPS-II and FeTDClPS-II were estimated from similar kinetic reactivities of the corresponding oxoFe^IVTMPS⁺ and oxoFe^IVTDClPS⁺ species to be ~92–94 kcal/mol. Similar values were calculated from the two-proton P⁺−Fe^III(OH₂)₂ pK^a_obs and the porphyrin oxidation potentials, despite a 230 mV range for the iron porphyrins examined. Thus, the iron porphyrin with the lower ring oxidation potential has a compensating higher basicity of the ferryl oxygen. The solvent-derived proton adds significantly to the driving force for C–H bond scission.

INTRODUCTION

Ferryl porphyrin species, P⁺−Fe^IV=O and P−Fe^IV=O, play pivotal roles in biological oxidations mediated by heme proteins.¹ These two reactive intermediates, historically referred to as compound I and compound II,² respectively, differ in the oxidation state of the porphyrin ring (P). Among the heme enzymes, only those bearing an axial thiolate ligand, as in the cytochromes P450 (CYP)³ and aromatic peroxygenases (APO),⁴ have the ability to perform the most challenging oxidation, the oxy-functionalization of unactivated sp³ C–H bonds. The consensus hydroxylation mechanism of CYP enzymes involves a hydrogen atom abstraction from a substrate C–H bond by P⁺−Fe^IV=O (P450-I) to afford a thiolateiron(IV) hydroxide⁵ (P450-II) and a substrate radical. The carbon-centered radical formed in this process then rebounds to the Fe^IV−OH hydroxyl, forming the product alcohol and reducing the heme center back to the ferric resting state. The strength of the FeO–H bond formed upon hydrogen atom abstraction is of considerable importance in determining the ability of a metal-oxo compound to abstract a hydrogen atom from a substrate,⁶ as is the electronic configuration.⁷ This bond strength, D(OH), can be dissected into a one-electron reduction potential of P⁺−Fe^IV=O (E⁰_{cpd I}) and the pK_a of P−Fe^IVO–H (eq 1).⁸For example, we have estimated the reduction potential of APO-I to be ~1.4 V and the FeO–H

$$D(\text{OH})=23.06{E^0}_{\text{cpdI}}+1.37\mathrm{p}{K}_{\text{a,cpdII}}+57\text{kcal}/\text{mol}$$ (1)

bond dissociation energy (BDE) of APO-II to be ~100 kcal/mol.^4b,c Likewise, a highly reactive synthetic iron porphyrin, oxoFe^IVTM-4-PyP⁺, also has a 100 kcal/mol FeO–H bond.⁹ A more basic FeO–H pK_a would lead to a stronger Fe^IVO–H bond. Green et al. have suggested that the axial thiolate ligand of P450 hydroxylases is an innovation used to engineer more basic ferryl states while sacrificing some redox potential.^5a The net result is a stronger Fe^IVO–H bond and the ability to oxidize unactivated aliphatic C–H bonds while protecting the protein from oxidative damage.^5,10 Support for this hypothesis derives from the recent determination that P450-II has a pK_a ≈ 12, existing as an iron(IV) hydroxide under turnover conditions.

While P450-II is protonated, there is still spirited debate concerning the ferryl states of heme enzymes lacking thiolate axial ligands. For example, while X-ray data have suggested an elongation of the Fe^IV−O bond in cytochrome c peroxidase compound II (CcP-II) and ascorbate oxidase compound II,¹¹ that effect has also been attributed to photoreduction of the sample in the X-ray beam.¹² A recent neutron diffraction study of CcP-I has shown that the ferryl is not protonated.¹³ Likewise, ferryl myoglobin resists protonation even below pH 3.¹⁴ There are no reports of basic ferryls in model porphyrin systems, although basic oxomanganese and non-heme iron complexes have been discussed,¹⁵ and dimethoxyiron(IV) porphyrin complexes are known.¹⁶

In this work we provide a direct measure of the basicity of sulfonated ferryl porphyrins (Figure 1) and the effects of ferryl basicity on their reactivity toward C–H bonds. Protonation of P−Fe^IV=O species using rapid-mixing, pH-jump, stopped-flow spectrophotometry reveals an electromeric equilibrium with measurable pK_a values for oxoFe^IV porphyrins and the corresponding iron(III) porphyrin cation radicals, P⁺−Fe^III(OH₂)₂. We also present evidence that the observed pK_a influences C–H bond scission by compound I models through synchronous solvent proton and substrate hydrogen atom transfer (HAT).

Figure 1.

Sulfonated ferryl porphyrins used in this study arranged, left to right in order of increasing electron-donation ability of the meso substituent, Ar. to right in order of increasing electron-donation ability of the meso substituent, Ar.

RESULTS

Iron tetramesitylporphyrin octasulfonate (FeTMPS) is a well-behaved, water-soluble iron porphyrin system. Previous work has shown that oxidation of Fe^IIITMPS with mCPBA or oxone above pH 7 affords oxoFe^IVTMPS, λ_max = 425–431 nm (pH dependent as shown in Figure S1); ¹H NMR δ 3.1 (para-methyl), −1.2 (β-pyrrole).¹⁷ Similarly, when Fe^IIITMPS was treated with mCPBA at pH 5.0, a new intermediate was formed with a weak, blue-shifted λ_max centered at 400 nm, a broad Q-band absorbance at ~650 nm, and strongly shifted ¹H NMR resonances at δ 23, 25 (ortho-methyls) and δ −26 (β-pyrrole) (Figures S2 and S3). We have assigned this species as oxoiron(IV) TMPS porphyrin cation radical (oxoFe^IVTMPS⁺) due to the closely analogous literature spectra of the tetramesityl derivative, oxoFe^IVTMP⁺,¹⁸ and similar species.^9,18,19 Electrochemical measurements with (H₂O)₂Fe^III-TMPS showed a reversible, pH-independent oxidation wave at 940 mV vs Ag/AgCl between pH 2 and pH 4, while spectroelectrochemical oxidation at 1100 mV afforded the corresponding iron(III) porphyrin cation radical, Fe^IIITMPS⁺, with λ_max = 390 nm and a broad Q-band feature (Figures S4 and S5).^19d The UV–vis spectra of the various oxidation/protonation states of FeTMPS are compared in Figure 2. Oxidation of Fe^IIITDClPS with 1 equiv of mCPBA at pH 3.0 also produced a compound I analogue, oxoFe^IVTDClPS⁺, with a similar, characteristic UV–vis spectrum (Figure S6).

Figure 2.

UV–vis spectra of various oxidation states of FeTMPS: Fe^IIITMPS, λ_max = 397 nm (green); oxoFe^IVTMPS, λ_max = 425 nm at pH 8 (blue); Fe^IIITMPS⁺, λ_max = 390 nm (red); and oxoFe^IVTMPS⁺, λ_max = 400 nm (black).

The interconversion of oxoFe^IVTMPS and Fe^IIITMPS⁺ was investigated using pH-jump, stopped-flow experiments. OxoFe^IVTMPS was formed via the stoichiometric addition of mCPBA to an unbuffered solution of Fe^IIITMPS adjusted to pH 11.5. The position of the Soret λ_max shifted upon mixing from a broad absorbance at 416 nm for (HO)Fe^IIITMPS to a sharper, red-shifted band at 425 nm for oxoFe^IVTMPS.^17a The stability of oxoFe^IVTMPS at high pH allowed for bulk generation of this species for use in single-mixing, pH-jump, stopped-flow experiments. Significantly, jumping the pH from 11.5 to <4 cleanly converted oxoFe^IVTMPS to Fe^IIITMPS⁺, while jumping the pH to values >4 led to mixtures of oxoFe^IVTMPS and Fe^IIITMPS⁺ (Figure 3A). A clear titration curve was revealed when the absorbance of oxoFe^IVTMPS at 425 nm observed within 10 ms of mixing was plotted against the pH of the final mixed solution (Figure 3B, black line). Fitting these data yielded an apparent pK_a = 5.5 ± 0.07. Over longer periods, the mixture of oxoFe^IVTMPS and Fe^IIITMPS⁺ disproportionated to form oxoFe^IVTMPS⁺ and Fe^IIITMPS. The identities of the species present after disproportionation were determined via generation of authentic oxoFe^IVTMPS⁺, as shown in Figures 2 and S2. Similar measurements for oxoFe^IVTDPS, oxo-Fe^IVTSMP,^19c,d and oxoFe^IVTDClPS (Figures S7 and S8) and the corresponding iron(III) porphyrin cation radicals gave a range of pK_a values from 4.0 to 6.3 (Figures 3B and S9−S11). These systems were chosen for their resistance to forming µ-oxo Fe^III dimers that could complicate the analyses.²⁰ Reverse pH-jump experiments of acidic solutions of Fe^IIITMPS⁺ to higher pH produced oxoFe^IVTMPS (Figure 4 inset). In this reaction sequence, Fe^IIITMPS at pH 2.5 was oxidized with 1 equiv of mCPBA to oxoFe^IVTMPS⁺, which quickly converted to Fe^IIITMPS⁺ (λ_max = 390 nm) under these conditions. Jumping the pH to ~13 produced oxo-hydroxoFe^IVTMPS (λ_max = 431 nm). The pH dependence of the UV–vis spectra of oxoFe^IVTMPS under highly alkaline conditions indicates the presence of a second pK_a = 12.5 ± 0.06 with the formation of HO−Fe^IV=O (Figure S1), as has been observed for other ferryl porphyrins.²¹

Figure 3.

(A) UV–vis spectra observed within 50 ms after rapid dilution of oxoFe^IVTMPS (10 µM, pH 11.5) with 50 mM phosphate/acetate buffer at the target pH: oxoFe^IVTMPS at pH 8.2 (red trace), Fe^IIITMPS⁺ at pH 4.7 (black trace), and both species at pH 5.4 (blue trace). (B) Plots of oxoFe^IVporphyrin absorbance vs pH for oxoFe^IVTDClPS (green trace, pK_a = 4.0 ± 0.05), oxoFe^IVTMPS (black trace, pK_a = 5.5 ± 0.07), oxoFe^IVTSMP (blue trace, pK_a = 5.9 ± 0.04), and oxoFe^IVTDPS (red trace, pK_a = 6.3 ± 0.06).

Figure 4.

Reduction of oxoFe^IVTMPS⁺ with TEMPO at various pH values (10 µM porphyrin in 50 mM phosphate/acetate buffer). Inset: Conversion of (H₂O)₂Fe^IIITMPS⁺ (10 µM) in pH 2.5 buffer (100 mM acetate) at 5 °C to oxoFe^IVTMPS at pH 13 upon bulk addition of NaOH. All concentrations and pH measurements are final after mixing.

The same equilibrium mixtures of oxoFe^IVTMPS and Fe^IIITMPS⁺ electromers could be generated via one-electron reduction of oxoFe^IVTMPS⁺. As shown in Scheme 1, oxoFe^IV-TMPS⁺ was generated via stoichiometric oxidation of Fe^III-TMPS with mCPBA in the first push of a double-mixing, stopped-flow experiment. After an aging delay of 1 s, 1 equiv of TEMPO in buffer at the target pH was added as a reducing agent in the second push.²² The reduction of oxoFe^IVTMPS⁺ by TEMPO was complete within the dead time of the measurement (~1 ms). By contrast, further reduction of oxoFe^IVTMPS to Fe^IIITMPS under these conditions occurred much more slowly (k ≈ 10⁴ M⁻¹ s⁻¹, Figure S12). Reduction of oxoFe^IVTMPS⁺ below pH 6.5 afforded mixtures of oxoFe^IV-TMPS and Fe^IIITMPS⁺, with greater amounts of the latter at lower pH, in accord with a prototropic equilibrium with a pK_a = 5.5 (Figure 4A). A summary of spectral data, porphyrin electrochemistry, and measured pK^a values for all porphyrins is presented in Table 1.

Scheme 1.

Design of a Double-Mixing, Stopped-Flow Experiment in Which OxoFe^IVTMPS⁺ Was Generated Transiently at pH 5.0 (Unbuffered), pH-Jumped, and Reduced with TEMPO To Yield Mixtures of OxoFe^IVTMPS and (H₂O)₂Fe^IIITMPS⁺

Table 1.

Collected Parameters for the Fe Porphyrin Model Systems Used in This Study

	FeIIITDPS	FeIIITSMP	FeIIITMPS	FeIIITDCIPS
meso substituent Ar =
pKa P-FeIII(OH2)2	6.65 (phos/acetate)	6.619d,19e (unbuffered)	7.5 (phos/acetate)17 6.8 (HEPES, CHES, MES)	7.85 (unbuffered)20,23
pKa P+FeIII(OH2)2/P-FeIV=O	pKaobs= 6.3	pKaobs= 5.9	pKaobs= 5.5	pKaobs= 4.0
E1/2 (vs. Ag/AgCl) ZnP/ZnP+	558 mV (pH 4.0)	630 mV (pH 4.0)19e	596 mV (pH 4.0)	790 mV (pH 4.0)
oxoFeIVP UV-vis (pH 11.5)	423 nm	424 nm19d	425 nm17	420 nm20,23
(H2O)2FeIIIP+UV-vis	394 nm	388 nm19d	390 nm17b	387 nm

Both oxoFe^IVTMPS⁺ and oxoFe^IVTDClPS⁺ were reduced to the ferric form by dihydroanthracene (DHA). Bimolecular rate constants for DHA oxidation by oxoFe^IVTMPS⁺ and oxoFe^IV-TDClPS⁺, determined at pH 5 and pH 3, respectively, in single-turnover, stopped-flow experiments were 1023 ± 53 and 5015 ± 170 M⁻¹ s⁻¹ (Figures S13 and S14). OxoFe^IVTDClPS⁺ was shown to oxidize xanthene with a bimolecular rate constant of (2.47 ± 0.15) × 10⁴ M⁻¹ s⁻¹ (R² = 0.99), while for oxoFe^IVTMPS⁺ background decay obscured this reaction. Use of xanthene-d₂ resulted in a slower rate constant of (1.37 ± 0.18) × 10⁴ M⁻¹ s⁻¹ (R² = 0.95) and, accordingly, a substrate kinetic isotope effect (KIE) of 1.80. This reaction was also shown to have a solvent KIE by oxidizing proteo xanthene in buffer made with D₂O (pD 3.0). The bimolecular oxidation rate constant in this case was (1.50 ± 0.11) × 10⁴ M⁻¹ s⁻¹ (R² = 0.99), yielding a solvent KIE of 1.65. Under identical conditions, xanthene-d₂ in D₂O afforded a bimolecular rate of (1.12 ± 0.09) × 10⁴ M⁻¹ s⁻¹ (R² = 0.98), indicating a combined substrate−solvent isotope effect of 2.2.

DISCUSSION

A unique feature of heme iron enzymes and iron porphyrin model systems is the non-innocence of the porphyrin ring. Iron(III)/iron(IV) and porphyrin ring redox potentials are strongly intermingled, with both processes occurring at ~1 V. The rapid pH-jump experiments we have described here show that the terminal oxo ligands of oxoiron(IV) porphyrin complexes can be protonated in mildly acidic media, exhibiting electromeric equilibria between the ferryl species, P−Fe^IV=O, and the corresponding bis-aqua-iron(III) porphyrin cation radicals, P⁺−Fe^III(OH₂)₂ (Figure 1). Apparent pK_a values from 4.0 to 6.3 were observed that vary with the redox potential of the porphyrin. Re-formation of P−Fe^IV=O upon deprotonation of P⁺−Fe^III(OH₂)₂ with base has shown that this equilibrium is reversible. Further, one-electron reduction of the corresponding compound I analogues, P⁺−Fe^IV=O, with TEMPO over the same range of pH produced the same mixtures of electromers and the same pK_a values. In the following discussion we analyze these interconversions and their implications for C–H oxidations mediated by these P⁺−Fe^IV=O species.

Analysis of the P−Fe^IV=O ⇆ P⁺−Fe^III−OH₂ Equilibrium

An important aspect of this study was differentiating among the various species involved. Fortunately, the UV–vis spectrum of Fe^IIITMPS⁺ is distinct from those of Fe^IIITMPS, oxoFe^IVTMPS⁺, and oxoFe^IVTMPS (Figure 2). In particular, Fe^IIITMPS⁺ exhibits a strong blue-shifted Soret band (λ_max = 390 nm) and broad, elevated Q-bands that are hallmarks of a porphyrin cation radical.^19d By contrast, the Soret band of oxoFe^IVTMPS⁺ is much weaker and centered at 400 nm, while the broad cation radical Q-bands were centered at longer wavelengths (Figure 2, black trace). These spectral features were particularly useful in observing Fe^IIITMPS⁺ before the appearance of oxoFe^IVTMPS⁺ via disproportionation. Electrochemical generation of Fe^IIITMPS⁺ from Fe^IIITMPS was shown to occur through a one-electron, no proton process from the lack of a pH dependence of the oxidation wave below pH 4. This observation indicates that the axial ligation of Fe^IIITMPS⁺ is bis-aqua, (H₂O)₂Fe^IIITMPS⁺, as is the starting Fe^IIITMPS.^17b Similar behavior has been described by Su et al. for a closely related porphyrin, Fe^IIITSMP.^19d

Rapid pH-jumping of preformed basic solutions of oxoFe^IV-TMPS to lower pH revealed the instantaneous formation of (H₂O)₂Fe^IIITMPS⁺. When the pH was jumped to intermediate values, both oxoFe^IVTMPS and (H₂O)₂ Fe^IIITMPS⁺ were observed in solution, with the ratio of these two species favoring oxoFe^IVTMPS as the pH increased (Figure S15). The isosbestic nature of this change and singular value decomposition (SVD) analysis of the data indicate that the only species present in solution just after mixing are oxoFe^IVTMPS and (H₂O)₂Fe^IIITMPS⁺. The rapid establishment of these mixtures (~1 ms) indicates that a fast and reversible equilibrium exists between them. A clear, pK_a^obs = 5.5 was observed for (H₂O)₂Fe^IIITMPS⁺ by plotting the ratio of species present vs the final pH (Figure 3).²⁴ For comparison, the first pK_a of the ferric derivative, (H₂O)₂Fe^IIITMPS, is 7.5 under these conditions,^17b thus providing a measure of the effect of porphyrin ring oxidation of the acidity of the ligated water protons. Interestingly, DFT computations have predicted six relatively low-lying electromers of HRP compound II, of which the hydroxoFe^III and aquaFe^III porphyrin cation radicals were indicated as the most stable isomers.²⁵ Two other pertinent examples of electromeric equilibria in iron(IV) porphyrin systems are apparent in CcP-II²⁶ and bis-azido-iron(IV) porphyrins.²⁷

One-electron reduction of oxoFe^IVTMPS⁺ near the pK_a^obs of (H₂O)₂Fe^IIITMPS⁺ was explored as an alternative approach to the equilibrium mixture of oxoFe^IVTMPS and (H₂O)₂Fe^III-TMPS⁺. We found that the nitroxyl radical, TEMPO (E⁰ = 760 mV), reacted rapidly with oxoFe^IVTMPS⁺ to provide the ferryl species, oxoFe^IVTMPS.²² Thus, stopped-flow mixing of oxoFe^IVTMPS⁺ with 1 equiv of TEMPO over a range of pH values resulted in the transient formation of mixtures of oxoFe^IVTMPS and (H₂O)₂Fe^IIITMPS⁺ (Figure 4), as was observed with the pH-jump experiments above (Figure 3). Further reduction of oxoFe^IVTMPS with TEMPO was found to occur at a much slower rate, thus allowing for the observation of both (H₂O)₂Fe^IIITMPS⁺ and oxoFe^IVTMPS without an appreciable accumulation of the ferric porphyrin. The apparently instantaneous establishment of the equilibrium mixture via TEMPO reduction is another indication that the interconversion of (H₂O)₂Fe^IIITMPS⁺ and oxoFe^IVTMPS is fast and reversible and that the product of reduction (which electromer or mixture of electromers) depends upon the pH. This reversibility was further confirmed by pH-jumping an acidic solution of (H₂O)₂Fe^IIITMPS⁺ to high pH and observing the clean formation of oxoFe^IVTMPS.

To elucidate the effect of ligand electronics on the observed aquaFe^IIIP⁺/oxoFe^IV porphyrin pK_a^obs, we examined several other sulfonated iron porphyrins: oxoFe^IVTSMP, oxoFe^IV-TDClPS, and oxoFe^IVTDPS. Electrochemical analysis of the zinc derivatives showed that the porphyrin ring oxidation potentials of these ligands spanned 230 mV (Table 1), allowing for examination of this parameter on the ferryl pK_a^obs. As was observed with oxoFe^IVTMPS, rapid pH-jump experiments with each of these oxoFe^IV porphyrins revealed an apparent pK_a^obs (Figure 3B). The most electron-donating durenyl sulfonate meso substituent produced the most basic pK_a^obs = 6.3 for (H₂O)₂Fe^IIITDPS⁺, while the most electron-withdrawing 2,6-dichlorophenyl sulfonate substituent led to the lowest pK_a^obs = 4 for (H₂O)₂Fe^IIITDClPS⁺. SVD and evolving factor analyses for each case indicated that only two species were present in significant concentrations in these equilibrium mixtures (Figures S17−S30). Each of these Fe^III porphyrin cation radical complexes exhibited a broader, blue-shifted Soret λ_max than that of the oxoFe^IV porphyrin upon jumping to acidic pH regimes. Additionally, the Q-band regions of these new species were all typical of porphyrin cation radical spectra. In the case of oxoFe^IVTSMP, the UV–vis spectrum of the protonated species matched that of electrochemically generated bis-aqua-Fe^IIITSMP porphyrin cation radical, which has been previously characterized.^19d These results, coupled with the evidence that protonation of oxoFe^IVTMPS produces (H₂O)₂Fe^IIITMPS⁺, indicate that the protonation of oxoFe^IV porphyrin systems to their corresponding Fe^III porphyrin cation radicals is a general phenomenon.

Comparing the reduction potential of the corresponding Zn porphyrin cation radicals to the (H₂O)₂Fe^III pK_a^obs revealed an inverse relationship between these two parameters. The lower the redox potential of the porphyrin ring, the higher the observed ferryl pK_a^obs. This trend is important for several reasons. First, the (H₂O)₂Fe^IIIP⁺/ferryl pK_a^obs appears to be very sensitive to changes in the oxidation potential of the porphyrin ring. When comparing the two extremes of the series, (H₂O)₂Fe^IIITDClPS⁺ is more than 500-fold more acidic than (H₂O)₂Fe^IIITDPS⁺, while 230 mV separates the oxidation potentials of their porphyrin rings. Thus, stabilizing the bisaqua-Fe^III porphyrin cation radical through a lower porphyrin redox potential shifts the PFeO-H₂ pK_a^obs to a higher value, creating an effectively more basic ferryl oxygen.

Discussion of ferryl heme basicity has generally focused on the role of the trans-axial ligand (L) as it affects the L−Fe^IV−OH ⇆ L−Fe^IV=O equilibrium. Specifically, the ability of a cysteine thiolate to allow for basic ferryl hemes as a way for heme-thiolate proteins such as cytochrome P450 and APO to oxidize strong C–H bonds in the presence of an oxidizable protein superstructure.^4b,c,5a,c Spectroscopic data indicate that the iron centers of CPO-II and P450-II exist as iron(IV) hydroxides.^5a,c The differences in structure and pK_a between the protonated CPO-II or P450-II and the protonated model systems used here can be attributed to the thiolate axial ligand in the active site. The anionic, strongly donating thiolate ligand is able to stabilize the iron(IV) center upon oxo protonation to a greater extent than the analogous aqua ligand present in FeTMPS-II,¹⁰ although a recent case involving an amidate oxyanion ligand apparently does not.²⁸ The relative importance of various parameters, such as the nature of the anionic ligand, Coulombic effects, σ- and π-trans-axial ligand effects, and field effects of other charges in the active sites of these proteins, has yet to be evaluated. The axial ligand is only one factor in determining the basicity of oxoFe^IV porphyrins, however. As can be seen from the results, ligand electronics represent another way to modulate ferryl basicity, which in this case corresponds to a two-proton Brønsted acidity of P⁺−Fe^III(OH₂)₂ and the accompanying electromeric equilibrium. As such, it appears likely that the compound II of some heme enzymes lacking a thiolate axial ligand could be protonated.^11a

An important aspect of the current study is the fast establishment of the electromeric equilibria determining the observed pK_a^obs values before the subsequent disproportionations. Disproportionation of iron(IV) porphyrins to oxoiron-(IV) porphyrin cation radicals and iron(III) species has long been known.^16,19f The tetrasulfonated iron porphyrin, oxo-Fe^IVTSMP, has been reported to disproportionate to a mixture of oxoFe^IVTSMP⁺ and Fe^IIITSMP in 4:1 methanol−water.^19c While a similar disproportionation reaction was observed in the current study in aqueous buffer over a longer time regime (>50 ms), the electromeric equilibrium was clearly established prior to this. One explanation for the difference in observed behavior between the tetra- and octasulfonated systems may be the different solvents. When FeTMPS was oxidized in basic methanol, oxidation to dimethoxyFe^IVTMPS was observed, with its distinctive UV–vis spectrum (Q-bands at 550 and 580 nm), rather than oxoFe^IVTMPS (Figure S16).^17b DimethoxyFe^IV porphyrins have been shown by us to disproportionate to oxoFe^IV porphyrin cation radicals and Fe^III porphyrins upon addition of acid.¹⁶

Analysis of Solvent Proton-Coupled Substrate Hydrogen Atom Transfer

Now we consider the reduction of P⁺−Fe^IV=O via HAT, which can be seen as an alternative way to access the P−Fe^IV = O/P⁺−Fe^III−OH₂ mixtures. Overall, the process is similar to the one-electron reduction of P⁺−Fe^IV=O with TEMPO described above, except that the electron and one proton derive from the same C–H bond. The observed equilibrium between P−Fe^IV=O and P⁺−Fe^III−OH₂ (pK_a^obs = 5.5 for FeTMPS) allows an estimate of the effective FeO–H BDE for hydrogen abstraction by P⁺−Fe^IV=O and the construction of an energy landscape for this process. We propose a mechanism for C–H bond scission that involves a simultaneous transfer of a proton from the medium to the ferryl oxygen (Scheme 2).

Scheme 2. Proposed Mechanism for Solvent Proton-Assisted C–H Bond Scissiona.

^aThe solvent proton is shown approaching the ferryl oxygen lone pair, while the substrate proton is engaging the ferryl π* SOMO orbital.

A significant aspect of the behavior of this system is that pK_a^obs represents a two-proton equilibrium. As mentioned above, SVD analysis of the titration data (Figures S17−S30) indicates that the main two components, P−Fe^IV=O and P⁺−Fe^III−OH₂, account for >98% of all species present. A small amount of disproportionation during the time of the measurement is expected from the observed kinetics that could account for the residual. There could also be ~1% of a monoprotonated intermediate, although no such species was observed. Accordingly, we adopt a two-state model, P−Fe^IV=O ⇆ P⁺−Fe^III−OH₂, with no other intermediates accumulating at any pH over the range of the measurements (Figure S18), simplifying the analysis, although interactions of P⁺−Fe^III−OH₂ with the phosphate buffer are certainly also occurring.²⁴ We estimate that pK_a1 for TMPS⁺−Fe^III−OH₂ should be at least 7. A value of 6.5 or less would significantly increase the amount of monoprotonated intermediates predicted at pH 5.5, which is not supported by the data or SVD analysis. This value also follows from the measured first pK_a for the ferric complex, P−Fe^III−OH₂, which is 7.5 for FeTMPS under these conditions.^17b Thus, pK_a1 for P⁺−Fe^III−OH₂ should be similar. With pK_a1 estimated to be ~7, and pK_a^obs measured at 5.5, we can calculate pK_a2 of P⁺−Fe^III−OH₂ to be ~4. The ferryl O-protonation event and the resulting electromeric equilibrium between P−Fe^IV and P⁺−Fe^III are contained in pK_a2. This unusual situation for a dibasic acid (pK_a2 < pK_a1) will be discussed further below. Finally, we estimate E′ for P⁺−Fe^IV=O/P−Fe^IV=O to be 1.06 V (vs NHE) from the measured porphyrin ring oxidation for FeTSMP of 1.1 V under these conditions and the very clear ring oxidations observed for the zinc porphyrin analogues (Table 1).^19d,e Similar analyses were performed for each iron porphyrin.

The results show that the effective ferryl basicity reflected in the P−Fe^IV=O ⇆ P⁺−Fe^III−OH₂ equilibrium has a measurable effect on reactivity in this aqueous model system. OxoFe^IVTMPS⁺ and oxoFe^IVTDClPS⁺ react with DHA with similar rate constants, 1075 ± 52 and 5015 ± 170 M⁻¹ s⁻¹, respectively. Estimates of the effective FeO-H BDE formed during substrate HAT by these two oxidants were ~92 and ~94 kcal/mol, respectively, determined by comparing the rate constants for other oxidants using the Brønsted−Evans−Polyani relationship for the same substrate in the usual way (Figure 6).²⁹

Figure 6.

Rate constants for reduction of oxoFe^IVTMPS⁺ and oxoFe^IVTDClPS⁺ by DHA mapped onto the Brønsted−Evans−Polyani relationship. Estimated FeO-H BDEs for FeTM-4-PyP and AaeAPO, determined by similar kinetic means using xanthene, are shown for comparison (cf. refs 4c and 9).

The accuracy of these BDE estimates could be confirmed electrochemically in this case. To account for the two-proton equibrium discussed above, we define a new FeO-H BDE parameter, D(OH₂), in a modified Bordwell equation (eq 2) that reflects both the homolytic FeO-H BDE for the hydrogen atom in flight from the substrate and pK_a1 for the proton derived from the medium.³⁰

$$D({\text{OH}}_2)=23.06E{\prime }_{\text{por}}+1.37[2(\mathrm{p}{K_{\mathrm{a}}}^{\text{obs}})]-1.37\text{pH}+57=90-93\text{kcal}/\text{mol}$$ (2)

Thus, D(OH₂), which is pH dependent, can be calculated to be 90–93 kcal/mol from the two measured quantities, E′ = 1.06 V and pK_a^obs = 5.5, over the pH range of the kinetic measurements (pH 3–5) (Scheme 3). It is satisfying that D(OH₂) determined in this way is the same as the estimate determined kinetically in Figure 6. As can be seen in eq 2, there is a considerable increase in the driving force for C–H bond scission due to the large contribution from the two-proton pK_a^obs (15 kcal/mol for oxoFe^IVTMPS⁺). Notably, a ferryl basicity corresponding to pK_a = 11 for P−Fe^IV−OH would be required to obtain this much of a driving force enhancement in a single-proton event. Since the C–H BDE of DHA is 76 kcal/mol, the overall driving force for C–H bond homolysis is 16.5 kcal/mol if the solvent protonation is included, while it is only 11 kcal/mol without solvent assistance. The energetics for this concerted, solvent proton-assisted HAT are compared to a stepwise process in Figure 7.

Scheme 3. Two-Proton Equilibrium and HAT Driving Force for OxoFe^IVTMPS⁺a.

^a Simultaneous hydrogen atom abstraction and ferryl protonation from the medium increase the overall driving force for C–H bond scission.

Figure 7.

Energy landscape for C–H bond scission in dihydroanthracene by oxoFe^IVTMPS⁺. A concerted process involving simultaneous hydrogen atom abstraction and ferryl protonation has a driving force of 16.5 kcal/mol and D(OH₂) = 92.5 kcal/mol at pH 3, while a stepwise hydrogen atom transfer has only an 11 kcal/mol driving force and D(OH) = 87 kcal/mol.

It is of interest to consider the extent to which the equilibrium P−Fe^IV=O ⇆ P⁺−Fe^III−OH₂, as well as the solvent proton-coupled C–H bond scission process in Scheme 3, are concerted, two-proton events. Since the reactions are carried out in water, solvent hydrogen bonds are expected to be in place at the terminal oxo ligand. The electromeric equilibrium P−Fe^IV=O ⇆ P⁺−Fe^III−OH₂ creates an interesting spring-loaded situation wherein pK_a2 is significantly more acidic than pK_a1. Thus, the second ferryl protonation in the P−Fe^IV=O ⇆ P⁺−Fe^III−OH₂ equilibrium may be synchronous with the first because the unseen P⁺−Fe^III−OH is more basic than P−Fe^IV=O. In this context, the trigger for this increased Fe−OH basicity is the electron transfer from the porphyrin ligand. Concerted two-proton transfers have been observed in other proton relays³¹ and in photoexcited hydrogen-bonded pairs.³² Winkler and Gray have shown that laser flash oxidation of ferric microperoxidase-8 affords an aquairon(III) porphyrin cation radical below pH 6 and the corresponding ferryl species at pH 7.³³ Loss of the two protons in P⁺−Fe^III−OH₂ in that case was too fast to be resolved (>10⁴ s⁻¹). Simultaneous two-proton transfers, with an attendant increase in driving force, also seem likely for the one-electron reductions for typical enzymatic ferryl species, such as HRP-II, for which the corresponding ferric-aqua pK_a values are known to be in the range of 7–11,³⁴ and for non-heme ferryl species, for which the Fe^III−OH₂ pK_a is in that same range.

Applied to the HAT event from the substrate mediated by P⁺−Fe^IV=O, a concerted ferryl protonation/hydrogen abstraction would have the advantage of the additional 5.5 kcal/mol driving force expressed in pK_a1 for oxoFe^IVTMPS⁺ (Figure 7). The experimental observation of both solvent and substrate isotope effects can be interpreted to result from this synchronicity, although a protonation pre-equilibrium leading to the HAT transition state could also give this result.

Interestingly, D(OH₂) values for TDPS⁺Fe^III−OH₂, TSMP⁺Fe^III−OH₂, and TDClPS⁺Fe^III−OH₂ were determined to be 91, 91.5, and 90.5 kcal/mol, respectively (SI), almost identical to that of oxoFe^IVTMPS⁺, despite a 230 mV range for the porphyrin ring oxidation potentials, corresponding to ~5 kcal/mol in expected driving force. Thus, the D(OH₂) values for all four porphyrins studied are essentially the same and identical to the effective FeO–H BDE estimated from the rates of C–H bond scission. Clearly, the porphyrin redox potential difference is not reflected in the observed rates of DHA oxidation by FeTMPS-I and FeTDClPS-I. If the FeO-H BDE is indeed determined via eq 2, the ferryl heme basicity reflected in pK_a^obs values may be compensating for the lower redox potential to lead to such similar C–H cleavage rates. In this regard, it is interesting to compare the much higher rates of reaction (>100-fold) observed with oxoFe^IVTM-4-PyP⁺.⁹ The ring oxidation potential for this cationic porphyrin has been measured at 1.22 V vs NHE, about 200 mV higher than that of FeTDClPS,³⁵ about the same as the redox potential difference between FeTMPS and FeTDClPS. Why should this be? The results suggest that the relatively high pK_a^obs of (H₂O)₂Fe^IIITMPS⁺ (5.5, and the correspondingly high pK_a1 ≈ 7) is sufficient to contribute to the overall driving force and lead to a higher reactivity than expected. Conversely, the pK_a of (H₂O)₂Fe^IIITDClPS⁺ is only 4, and that of (H₂O)₂FeTM-4-PyP⁺, while unknown, is certainly lower. Thus, the hydrogen abstraction rates for these low-pK_a iron porphyrins are largely determined by their relative oxidation potentials, with reduced contributions from the ferryl oxygen basicity. Obviously, it would be useful to have synthetic, biomimetic C–H oxygenation catalysts that have high C–H scission rates at relatively low oxidation potentials.^15e

The idea that protonation or hydrogen-bonding of a ferryl oxo can influence the reactivity of such complexes has been suggested for non-heme oxoFe^IV complexes. Ferryl basicity has been discussed recently in connection with differences between C–H and O–H or N−H bond scission.^15b Also, addition of strong acids to the non-heme oxoFe^IV complex [(N₄Py)-Fe^IV(O)]²⁺ resulted in an increased reduction potential and increased rates for C–H oxidation.³⁶ Concomitant with this increase in rate, however, was the loss of the substrate KIE and the appearance of an inverse solvent isotope effect, suggesting that the rate-determining step is electron transfer rather than H-atom abstraction. By contrast, oxoFe^IVTDClPS⁺ displays both substrate and solvent KIEs, suggesting that the solvent proton is assisting in the HAT from the substrate as opposed to changing the mechanism from H-atom abstraction to electron transfer. Brønsted acid catalysis of substrate C–H bond cleavage may also be occurring in heme proteins such as cytochrome P450 and APO. The proton relay channel of cytochrome P450 has been well studied and is usually invoked to catalyze the O−O bond heterolysis in a ferric hydroperoxo precursor (compound 0) of the reactive oxoiron(IV) porphyrin cation radical and a water molecule.^13,37 The same proton relay channel could use the newly formed water molecule to activate the ferryl oxygen through protonation in a similar manner as in these iron porphyrin model systems. More generally, ferryl protonation could play a significant role in the reactivity and selectivity of non-heme iron proteins such as the α-ketoglutarate-dependent halogenase SyrB2³⁸ and in (S)-2-hydroxypropyl-phosphonic acid epoxidase (HppE),³⁹ which mediates the unusual epoxide formation in the biosynthesis of fosfomycin. In these cases, simultaneous ferryl protonation and substrate hydrogen abstraction could both facilitate C–H bond scission and disfavor hydroxylation of the incipient substrate radical.

SUMMARY AND CONCLUSIONS

We have shown that the oxo groups of sulfonated oxoFe^IV porphyrin model systems can be protonated under mild conditions, with apparent pK_a values for the P⁺−Fe^III(OH₂)₂ ⇆h P−Fe^IV=O interconversion between 4.0 and 6.3, depending on the porphyrin meso substituent. The protonated form of these oxoFe^IV porphyrins is a bis-aqua-Fe^III porphyrin cation radical, or a buffer-bound equivalent. The reversible protonation of this family of ferryl porphyrins reveals a novel electromeric equilibrium, which, while predicted by theory,²⁵ has not been previously described. Hydrogen abstraction by the compound I analogues, oxoFe^IVTMPS⁺ and oxoFe^IVTDClPS⁺, leads to protonated rebound intermediates when the pH of the medium is below the pK_a^obs of P⁺−Fe^III(OH₂)₂. The rates of reaction of oxoFe^IVTMPS⁺ and oxoFe^IVTDClPS⁺ with dihydroanthracene are within a factor of 5 of each other, despite a 200 mV difference in the ring oxidation potential. The observed pK_a^obs values for (H₂O)₂Fe^IIITMPS⁺ and (H₂O)₂Fe^IIITDClPS⁺ differ by 1.5 pH units but in a compensating direction. Using kinetic and electrochemical estimations, the effective FeO–H BDEs of FeTMPS-II and FeTDClPS-II, D(OH₂), have been estimated to be ~90–93 kcal/mol, involving both the hydrogen atom donated by the substrate and a proton delivered simultaneously from the medium. The presence of solvent O–H and substrate C–H deuterium KIEs indicate that hydrocarbon oxidation by these anionic oxoFe^IV porphyrin cation radicals occurs via a novel solvent protoncoupled hydrogen atom transfer process. Thus, hydrogen atom abstraction from the substrate by the ferryl species produces a basic Fe−OH intermediate, the protonation of which can increase the overall driving force for C–H bond scission by at least 5 kcal/mol.

MATERIALS AND METHODS

H₂TSMP, H₂TMP, and H₂TDClP were purchased from Frontier Scientific and used without further purification. Ferrous ammonium sulfate, Dowex AG-50W X8 ion-exchange resin (Na form), Sephadex LH-20 size exclusion gel, anhydrous sodium phosphate monobasic, acetic acid, perchloric acid, potassium hydroxide, TEMPO, and methanol were purchased from Sigma-Aldrich and used as received. mCPBA was obtained from Sigma-Aldrich and was purified by washing with pH 7.0 phosphate buffer (100 mM) to remove mCBA. D₂O and methanol-d₄ were purchased from Cambridge Isotope Laboratories, Inc. Water was distilled and deionized (Millipore, Milli-Q). Buffer solutions were prepared from a stock solution of 100 mM sodium phosphate/acetic acid and were corrected using concentrated KOH or perchloric acid solutions. pH values of the ACN/buffer mixtures were determined using the method of Bosch and co-workers.⁴⁰ pH was measured using a sympHony meter that had been calibrated prior to use (pH 4.0, 7.0, and 10.0).

Porphyrin Synthesis

H₂TMPS was prepared as previously reported from commercially available H₂TMP (Frontier Scientific).⁴¹ H₂TDClPS was prepared from commercially available H₂TDClP (Frontier Scientific) using the method of Gonsalves et al.⁴² H₂TDPS was prepared from H₂TDP using the method of Badger et al.⁴³

H₂TDPS-Na₄

The sulfonation of H₂TDP was performed using the method reported by Gonsalves et al.⁴² An oven-dried 25 mL round-bottom flask equipped with a stir bar and a rubber septum was charged with H₂TDP (0.12 mmol) and ~6 mL of neat chlorosulfonic acid. The mixture was allowed to stir for 1 h at room temperature, at which point it was quenched by slowly pouring onto 100 g of ground ice (Caution: this mixture will violently splatter!). The chlorosulfonyl porphyrin intermediate was extracted with chloroform (3 × 50 mL). The combined organic washings were then washed with saturated NaHCO₃ (50 mL) and brine (50 mL) and dried over MgSO₄. The purple solution was dried in vacuo to yield a shiny purple solid. The solid was then resuspended in water (50 mL) and refluxed overnight to yield the corresponding tetrasulfonic acid. The compound was purified via ion exchange (Dowex AG-50W X8, Na form) to yield H₂TDPS-Na₄ as a purple solid (101 mg, 69% of theoretical). UV–vis (λ, nm; ε, M⁻¹ cm⁻¹): 397 (27 000), 416 (140 000), 516 (6500), 553 (3300), 583 (3300), 591 (2700) (100 mM PBS, pH 7.4). ¹H NMR (D₂O, 500 MHz, δ): 8.47 (s, 8H, β-pyrrole), 2.77 (s, 24H, meta-methyl), 1.71 (s, 24H, ortho-methyl). ¹³C NMR (D₂O, 500 MHz, ppm): 140.8, 139.1, 136.8, 131.5, 126.5, 120.0, 20.2, 19.0.

Iron Porphyrins

The insertion of iron into the porphyrin ligands was performed according to literature methods.^19d Briefly, sulfonated porphyrin ligand (0.04 mmol) was dissolved in 15 mL of unbuffered water and pH corrected to pH 4.0 with sulfuric acid. The solution was added to a 25 mL round-bottom flask equipped with a reflux condenser and an argon inlet. The mixture was brought to and held at reflux for 30 min under argon. (NH₄)₂Fe(SO₄)₂ (100 mg, 0.26 mmol) was then quickly added to the refluxing solution under argon. The mixture was allowed to stir at reflux overnight, at which point the argon inlet was removed and the mixture refluxed open to air for 1 h. NaOH was added to the mixture to precipitate out the excess iron salts in solution. The mixture was then filtered through a bed of Celite and concentrated in vacuo. The resulting solid was purified via ion exchange (Dowex AG-50W X8, Na form) using water as the eluent. Finally, the mixture was desalted via size exclusion chromatography (Sephadex LH-20), eluting with 20% aqueous methanol, yielding the desired Fe complex. FeTMPS was prepared and characterized as we have previously described.^17a

FeTDPS-HSO₄-Na₄

13.9 mg, 25% of theoretical. UV–vis (λ, nm; ε, M⁻¹ cm⁻¹): 328 (14 000), 416 (43 500), 505 (4000), 638 (1500). HR-ESI-MS (flying as a formate salt) for C₆₁H₅₈FeN₄O₁₄S₄³⁻: expected 418.073 m/z, observed 418.073 m/z.

Zinc Porphyrins

ZnTMPS-Na₈

H₂TMPS (100 mg, 0.063 mmol) was added to a 50 mL round-bottom flask equipped with a stir bar and a reflux condenser. A minimal amount of water was added to dissolve the solid. The solution was brought to reflux, followed by addition of ZnOAc (77.5 mg, 0.35 mmol). The reaction was monitored by UV–visible spectroscopy, and an increasing amount of Zn was added until the reaction was complete. Dowex AG 50W X8 (sodium form, 1 g) was then added and the mixture stirred for 1 h. The mixture was concentrated in vacuo and redissolved in a minimal amount of 50:50 methanol/water. This solution was then chromatographed on Sephadex LH-20, eluting with 50:50 methanol/water, to yield ZnTMPS as a shiny purple sodium salt (36.2 mg, 30%). UV–vis (λ, nm; ε, M⁻¹ cm⁻¹): 405 (47 000), 426 (578 000), 562 (19 500), 601 (7300) (100 mM PBS, pH 7.4). ¹H NMR (D₂O, 500 MHz, δ): 8.73 (s, 8H, β-pyrrole), 3.17 (s, 12H, para-methyl), 2.11 (s, 24H, ortho-methyl). HR-ESI-MS for C₅₆H₄₉N₄O₂₄S₈Zn³⁻: expected 493.660 m/z, observed 493.659 m/z.

ZnTDClPS-Na₄

Synthesized as above; 19.3 mg, 30%. UV–vis (λ, nm; ε, M⁻¹ cm⁻¹): 402 (46 500), 423 (535 000), 557 (20 000), 592 (22 000) (100 mM PBS, pH 7.4). ¹H NMR (D₂O, 500 MHz, δ): 8.74 (s, 8H, β-pyrrole), 8.32 (m, 4H), 7.89 (m, 4H). HR-ESI-MS for C₄₄H₁₆Cl₈N₄O₁₂S₄Zn⁴⁻: expected 316.910 m/z, observed 316.911 m/z.

ZnTDPS-Na₄

Synthesized as above, starting with H₂TDPS (0.038 mmol, 43.9 mg); 24.6 mg, 42%. UV–vis (λ, nm; ε, M⁻¹cm⁻¹): 404 (52,000), 423 (400,000), 560 (19,000), 599 (7500). ¹H NMR (MeOD, 300 MHz, δ) 8.47 (s, 8H, β-pyrrole), 2.84 (s, 24H), 1.80 (s, 24H). HR-ESI-MS for C₆₀H₅₆ZnN₄O₁₂S₄⁴⁻ expected 304.052 m/z, observed 304.051 m/z.

Instrumentation

UV–vis spectra were recorded using either a Hewlett-Packard 8453 diode array spectrophotometer or a Varian Cary Bio-300 double-beam spectrophotometer with temperature control. Rapid-mixing stopped-flow experiments were obtained using a thermostated Hi-Tech SF-61 DX2 double-mixing instrument with a 1 cm or 2 mm path length cuvette equipped with a diode array detector. NMR spectra were recorded using either a Bruker Avance (500 MHz) or a Bruker Ultra-Shield (300 MHz) spectrometer. HR-MS data were recorded using an Agilent 6220 accurate-mass LC-TOF mass spectrometer operating in negative mode. Electrochemical measurements were obtained using a BAS 100B/W electrochemical workstation.

Electrochemical Measurements

Cyclic voltammograms were obtained using a three-electrode cell: a glassy carbon electrode (3 mm diameter) as the working electrode, a platinum wire as the counter electrode, and a saturated Ag/AgCl reference electrode purchased from BASi. The electrochemical measurements were performed in 50 mM phosphate/acetate solution, which served as both buffer and supporting electrolyte.

Rapid-Mixing pH-Jump Experiments/pK_a Determination

All kinetic data were obtained at 14.5 °C unless otherwise noted using a pH-jump technique similar to that reported by Yosca.^5a Each experiment was carried out in runs composed of six individual tests, which were then averaged and duplicated at least twice. Reactions that were completed in under 1 min were measured using a thermostated Hi-Tech SF-61 DX2 double-mixing stopped-flow instrument with a 1 cm path length cuvette in single-mixing mode or in double-mixing mode with the reported aging time. Reactions taking longer than 1 min were measured using a Varian Carey Bio-300 double-beam spectrophotometer operating in single-wavelength mode. pH-jump experiments for oxoFe^IVTMPS, TSMP, and TDClPS were performed in single-mixing mode by the 1:1 rapid mixing (and dilution) of a high-pH oxoFe^IVporphyrin solution and 100 mM buffer (phosphate acetate). pH-jump experiments for oxoFe^IVTDPS were performed in double-mixing mode, where the first push is used to generate oxoFe^IVTDPS and the second push to jump the pH of the newly formed oxoFe^IV species. Ionic strength was not compensated and varied between 0.055 and 0.175 M, depending upon the pH of the phosphate−acetate buffer after the pH jump. The pH of the resulting solutions was measured for each run. Spectrophotometric titrations were made using the first scan of the experiment after the instrumental dead time (~1 ms).

Formation of OxoFe^IVporphyrin Complexes

OxoFe^IVTSMP and oxoFe^IVTMPS were formed in bulk by the addition of stoichiometric mCPBA (100 mM stock solution in ACN) to a solution of Fe^IIITSMP or Fe^IIITMPS (20 µM, pH 11.5, unbuffered), respectively. OxoFe^IVTDClPS was formed by the bulk addition of 4 equiv of mCPBA (100 mM stock solution in ACN) to a solution of Fe^IIITDClPS (20 µM, pH 11.5 unbuffered). OxoFe^IVTDPS was formed under rapid-mixing stopped-flow conditions by the 1:1 mixing of Fe^IIITDPS (40 mM, pH 11.5, unbuffered) and mCPBA (40 mM, pH 11.5, unbuffered).

Kinetic Measurements

All kinetic experiments were performed at 14.5 °C in 50 mM phosphate acetate buffer. Oxidations of hydrocarbon substrates were performed in 20% ACN buffer mixtures to increase solubility. Each experiment was carried out in runs composed of six individual shots, which were then averaged and duplicated at least twice. The concentrations presented are final after mixing. Time-resolved UV–vis spectra are recorded using Hi-Tech SF-61 DX2 double mixing instrument with a 1 cm path length cuvette equipped with diode array detector. Substrate reactions with xanthene and dihydroanthracene were performed in double-mixing mode where the oxoFe^IV porphyrin cation radical species were generated by mixing Fe^IIIporphyrin complexes and mCPBA in the first push. The substrate was then introduced in the second push after the appropriate aging time. Values of k_obs were obtained by fitting the kinetic profile to a single-exponential equation using Hi-Tech KinetAsyst 2.28 software. Bimolecular reaction rate constants were then obtained from a linear fit of k_obs versus substrate concentration. Global spectral analyses, evolving factor analyses and singular value decomposition analyses⁴⁴ were performed using the ReactLab Kinetics 1.1 software package.

Figure 5.

Oxidation of xanthene by oxoFe^IVTDClPS⁺ under pseudofirst-order conditions. Reaction with proteo xanthene in proteo buffer (pH 3.0) yielded a bimolecular rate constant of (2.47 ± 0.15) × 10⁴ M⁻¹ s⁻¹ (R² = 0.99) (red). Xanthene-d₂ showed a slower bimolecular rate constant of (1.37 ± 0.18) × 10⁴ M⁻¹ s⁻¹ (R² = 0.95), which yields a substrate KIE of 1.80 (blue). Reaction of oxoFe^IVTDClPS⁺ with proteo xanthene in deutero buffer (pD 3.0) also showed a slower bimolecular rate constant than in the proteo buffer case, (1.50 ± 0.11) × 10⁴ M⁻¹ s⁻¹ (R₂ = 0.99), which yields a solvent KIE of 1.65 (black). Xanthene-d₂ in D₂O afforded a bimolecular rate of (1.12 ± 0.09) × 10⁴ M⁻¹ s⁻¹ (R₂ = 0.98), indicating a combined substrate−solvent isotope effect of 2.2 (purple).