Heme–Cu Binucleating Ligand Supports Heme/O₂ and Fe^II−Cu^I/O₂ Reactivity Providing High- and Low-Spin Fe^III−Peroxo−Cu^II Complexes

Abstract

The focus of this study is in the description of synthetic heme/copper/O₂ chemistry employing a hemecontaining binucleating ligand which provides a tridentate chelate for copper ion binding. The addition of O₂ (−80 °C, tetrahydrofuran (THF) solvent) to the reduced heme compound (P^ImH)Fe^II (1), gives the oxy-heme adduct, formally a heme–superoxide complex Fe^III−(O₂^•−) (2) (resonance Raman spectroscopy (rR): ν_O−O, 1171 cm⁻¹ (Δ¹⁸O₂, −61 cm⁻¹); ν_Fe−O, 575 cm⁻¹ (Δ¹⁸O₂, −24 cm⁻¹)). Simple warming of 2 to room temperature regenerates reduced complex 1; this reaction is reversible, as followed by UV–vis spectroscopy. Complex 2 is electron paramagnetic resonance (EPR)-silent and exhibits upfield-shifted pyrrole resonances (δ 9.12 ppm) in ²H NMR spectroscopy, indicative of a six-coordinate low-spin heme. The coordination of the tethered imidazolyl arm to the heme–superoxide complex as an axial base ligand is suggested. We also report the new fully reduced heme–copper complex [(P^ImH)Fe^IICu^I]⁺ (3), where the copper ion is bound to the tethered tridentate portion of P^ImH. This reacts with O₂ to give a distinctive low-temperature-stable, high-spin (S = 2, overall) peroxo-bridged complex [(P^ImH)Fe^III−(O₂²⁻)−Cu^II]⁺ (3a): λ_max, 420 (Soret), 545, 565 nm; δ_pyrr, 93 ppm; ν_O−O, 799 cm⁻¹ (Δ¹⁸O₂, −48 cm⁻¹); ν_Fe−O, 524 cm⁻¹ (Δ¹⁸O₂, −23 cm⁻¹). To 3a, the addition of dicyclohexylimidazole (DCHIm), which serves as a heme axial base, leads to low-spin (S = 0 overall) species complex [(DCHIm)(P^ImH)Fe^III−(O₂²⁻)−Cu^II]⁺ (3b): λ_max, 425 (Soret), 538 nm; δ_pyrr, 10.2 ppm; ν_O−O, 817 cm⁻¹ (Δ¹⁸O₂, −55 cm⁻¹); ν_Fe−O, 610 cm⁻¹ (Δ¹⁸O₂, −26 cm⁻¹). These investigations into the characterization of the O₂-adducts from (P^ImH)Fe^II (1) with/without additional copper chelation advance our understanding of the dioxygen reactivity of heme-only and heme/Cu-ligand heterobinuclear system, thus potentially relevant to O₂ reduction in heme–copper oxidases or fuel-cell chemistry.

Graphical Abstract

INTRODUCTION

Cytochrome c oxidase (CcO), the terminal enzyme of the respiratory chain in the mitochondrial inner membrane from mammalian and bacterial sources, reduces dioxygen to water, coupling the redox energy to membrane proton translocation utilized for ATP synthesis.^1–10 The enzyme possesses five key redox-active metal ion centers, dicopper complex Cu_A, which is a binuclear electron-transfer center, a low-spin heme a (cytochrome a), and the heterobinuclear (Fe_a3/Cu_B) active site, where dioxygen binding and reduction occurs.^1–10 The latter (Scheme 1) consists of a high-spin heme and a neighboring (~4.5–5.0 Å) copper ion possessing three histidine imidazole nitrogen ligands, one of which is cross-linked to a tyrosine (Tyr) residue biosynthesized via a posttranslational modification.^11–16 With a focus on the heme/Cu active site in varying redox or ligation states, structural aspects and the O₂ reduction mechanism have been studied extensively, utilizing many spectroscopic methods,^1–10 X-ray crystallography,^17,18 and computational approaches.¹⁹

Scheme 1.

Proposed O−O Reductive Cleavage Mechanism by Heme–Copper Oxidases for Cytochrome c Oxidase

The current state of our understanding of early stages in the CcO O₂ binding and reduction are given in Scheme 1. Molecular oxygen reacts with a fully reduced Fe^II⋯Cu^I binuclear active site, initially forming a detectable, transient, Fe−O₂ (“oxy-heme”) adduct, formally a Fe^III−(O₂^•−) (A) species possessing end-on superoxide binding. This assignment is consistent with resonance Raman (rR) spectroscopic properties, as being similar to those found in oxy-hemoglobin (ν_Fe−O = 567 cm⁻¹) and oxy-myoglobin (ν_Fe−O = 570 cm⁻¹).^20,21

Intermediate A rapidly undergoes O−O cleavage to yield P_M.²² A total of four electrons are required to fully reductively cleave the O−O bond of dioxygen. Three electrons are provided by the binuclear site (two from iron (Fe^II → Fe^IV) and one from Cu_B (Cu^I → Cu^II)). It is now accepted that the fourth electron derives from the phenol residue of the nearby His–Tyr cross-link. This phenol residue serves as an overall hydrogen atom donor (H•), leaving P_M formulated as a Fe^IV=O/Cu^II−OH/Tyr^• moiety, i.e., thus with a ferryl-oxo (Fe_a3^IV=O; ν(Fe−O) = 804 cm⁻¹), cupric-hydroxide (Cu_B^II−OH) and a tyrosyl radical (Scheme 1).^6,8,10 Also, evidence suggests that the His–Tyr cross-link enables enzyme Cu_B incorporation and stabilizes the precise geometry of the binuclear active site.^11,16

On the basis of biochemical experiments,²³ computational studies,^24–27 and synthetic modeling,^5,10 a peroxo-bridged heterobinuclear Fe^III−(O₂²⁻)−Cu^II (I_p) has been predicted as an intermediate species occurring on the pathway from A to P_M. This peroxo moiety critically precedes O−O bond cleavage via initial H-bonding or protonation,¹⁹ possibly through a water molecule bridge, as shown in Scheme 1.

Design and construction of heme/copper synthetic models provide critical information about structure, spectroscopy and reactivity relevant to CcO enzyme mechanism, in the usual manner that models contribute to scientific advances, allowing a breakdown and attention to particular aspects of the protein structure/function. With respect to dioxygen reactivity, reduced heme–copper assemblies (or components) have been shown to generate heme−Fe^III−superoxo entities (like intermediate A, Scheme 1; see complexes from Collman²⁸ and Naruta²⁹ and their co-workers, Figure 1). Other studies from Naruta³⁰ and our laboratories^31,32 have revealed that peroxo-bridged heme−Fe^III−(O₂²⁻)−Cu^II(ligand) complexes can be generated possessing three structural types¹⁰ (high-spin (HS) μ-η²:η¹-peroxo (C, D),^30,31,33 HS μ-η²:η²-peroxo (E),^31,32 and low-spin (LS) μ−1,2-peroxo (F))^10,34 (Figure 1). Collman’s investigations involving elegant binucleating (for Fe and Cu) ligands demonstrated efficient electrocatalytic four-electron four-proton O₂ reduction to water (ORR)³⁵ and stoichiometric studies where a phenol H atom donor could reductively cleave the O−O bond of a heme−O₂−Cu construct.³⁶ We have developed similar model systems with simpler compounds, i.e., complex D (Figure 1), which effect efficient electrocatalytic ORR chemistry.^37,38

Figure 1.

Various complexes of heme/copper synthetic models: superoxide heme−Fe^III−(O₂^•−)⋯Cu^I(ligand) complexes, (A, Collman’s group; B, Naruta’s group). Heme−Fe^III−(O₂²⁻)−Cu^II(ligand) complexes: X-ray crystal structure of Naruta’s HS μ-η²:η¹-peroxobridged complex (C: adapted from ref 30); Karlin group analog (D); μ-η²:η²-peroxo compound (E); LS heme peroxo adducts with a μ-1,2-peroxo structure type, LS-3DCHIm (F); DFT-calculated structure for [LS-4DCHIm (ArOH)] (G: adapted from ref 34), a phenol adduct.

In stoichiometric reactivity studies, detailed synthetic and theoretical/computation insights have been obtained concerning elements critical for reductive cleavage of heme–Cu bound O₂ (as peroxide) intermediate. However, many details and fundamental aspects have not been elucidated or fully understood. These include the exact structural requirements for the Fe−O−O−Cu moiety, tridentate versus tetradentate Cu chelation, factors such as H-bonding to a peroxide O atoms (which one?) versus actual protonation, the origin and pK_a of the proton donor, the timing of electron versus proton injection, and so on. Thus far, for heme−Cu−O₂ constructs, very small changes in the local (ligand) environment may lead to absolutely varying chemistry, and there is a need to understand these effects in obtaining fundamental insights. For example, LS-3DCHIm (F) and LS-4DCHIm (G) (Figure 1), which differ by only one DCHIm Cu-ligand, behave completely differently toward proton–electron sources. The former is completely unreactive, while the latter, in the presence of an H-bonding phenol, undergoes clean peroxide O−O reductive cleavage to give water.^10,19,34,39

From these perspectives, considerable further studies and advancements are needed. These include a continued effort to define the dioxygen reactivity of reduced heme–copper assemblies, while also if possible, elucidating the heme-only and (ligand)–Cu O₂-chemistry. For further elaborated/new heme−O₂−Cu species, subsequent defining of O−O reductive cleavage chemistry is a longer-term goal, in the contexts mentioned above.

In the present study, we report on an advanced CcO active site model system with binucleating ligand P^ImH; this employs a new tridentate ligand which includes a histamine moiety appended to the periphery of a fluorinated tetraphenylporphyrin. The Fe^II−porphyrinate complex (P^ImH)Fe^II (1), is shown in Chart 1.⁴⁰ Herein, we report the oxygenation chemistry of 1, whose chemistry mimics the initial O₂ binding in CcO, as well as formation of oxy-hemoglobin and -myoglobin. The generation of a heme–copper derivative of 1 has been accomplished, and oxygenation of the reduced Fe^II−Cu^I form gives a high-spin heme–peroxo–copper product. Lastly, addition of DCHIm as a heme axial ligand to the high-spin solution leads to low-spin complex. All complexes have been spectroscopically characterized by UV–vis, ²H NMR, and electron paramagnetic resonance (EPR) spectroscopies; for the O₂-derived complexes, vibrational data has been obtained using rR.

Chart 1.

RESULTS AND DISCUSSION

Dioxygen Reactivity of (P^ImH)Fe^II (1).

Complex 1 reacted with dioxygen to form a low-temperature stable superoxide level intermediate, which was characterized by UV–vis, ²H NMR, rR, and EPR spectroscopies. Bubbling dry O₂ through a solution of (P^ImH)Fe^II (1) {λ_max = 419, 526 and 552(sh) nm} in THF at −80 °C led to new features at 424 and 541 nm which were ascribed to a (P^ImH)Fe^III−(O₂^•−) (2) (Figure 2). Further characterization of the 1/O₂ reaction product comes from low-temperature ²H NMR spectroscopic investigation, employing a deuterated (β-pyrrolic hydrogens) analog of 2 (2-d₈) in THF solvent (Figure 3). It confirms the proposed formation of a myoglobin-like Fe^III−superoxide complex. Complex 2 gives rise to a diamagnetic spectrum (δ = 9.12 ppm), consistent with the formation of (P^ImH)Fe^III−(O₂^•−) (2-d₈) and suggesting a low-spin (S = 0) six-coordinate iron.⁴¹ As expected for a diamagnetic species, complex 2 is EPR-silent. The literature on hemoglobin and model compounds has generally proposed that the unpaired electron of the low-spin Fe^III ion is antiferromagnetically coupled with the unpaired electron of the superoxide radical anion, making the complex diamagnetic.⁴¹ Recent resonant inelastic X-ray scattering (RIXS) analyses have shown that the electronic structure more closely fits to an Fe^II—(O₂) description for a model system (the “oxy picket-fence porphyrin”), while oxyhemoglobin has more Fe^III—(O₂^•−) character (although not fully to the low-spin ferric level).⁴² Nevertheless, the latter notation is used here for clarity.

Figure 2.

(A) Reaction scheme showing reversible ferric heme superoxide (2) formation by bubbling O_2(g) into a solution of fully reduced complex 1. (B) UV–vis spectra (THF at −80 °C) of the deoxygenation reaction of 1. Shown are spectra of complex 1 (black) and 2 (red).

Figure 3.

²H NMR spectra of (A) d₈-(P^ImH)Fe^II (1-d₈) and (B) d₈-(THF)(P^ImH)Fe^III−(O₂^•−) (2-d₈) in THF at −80 °C. The sharp peaks at δ 3.58 and 1.73 ppm correspond to solvent THF.

We postulate that the sixth axial ligand present in 2 is the imidazolyl group of the tridentate tether in P^ImH (see Chart 1). We base this conclusion on our previous X-ray crystallographic finding that in a six-coordinate iron(II) complex with P^ImH, also containing a strong isocyanide donor ligand DIMPI (= 2,6-dimethylphenyl isocyanide), [(P^ImH)Fe^II−DIMPI], the free imidazole group binds strongly (Fe−N_Im = 2.024 Å) as the axial ligand trans to DIMPI (see diagram below).⁴⁰

The O₂ binding to 1 was reversible, as we have found for some other heme–Fe(II) O₂-adducts with porphyrinates possessing electron-withdrawing fluorine substituents,⁴³ where warming a solution of 2 to room temperature in a closed system allows for release of dioxygen and regeneration of (P^ImH)Fe^II (1), to the extent of ~80%. Cooling of this solution (without further addition of O₂) to −80 °C leads to close to full regeneration of (P^ImH)Fe^III−(O₂^•−) (2). Dioxygen binding cycling in this manner illustrated in the Figure S1. Several cycles of the reversible O₂-binding process could be monitored by UV–vis spectroscopy.

Complex (P^ImH)Fe^III−(O₂^•−) (2) was further characterized by rR spectroscopy (Figure 4). In the lower-frequency region, a ν_Fe−O stretch is detected at 575 cm⁻¹ which is similar to that of other known “oxy-heme” Fe^III−(O₂^•−) synthetic models with a nitrogen axial ligand.^10,41,44 This Fe−O stretch shifts down by 24 cm⁻¹ upon ¹⁸O₂ substitution. An oxygen isotopesensitive feature corresponding to the O−O stretch was detected at ν_O−_O = 1171 cm⁻¹, which shifted to 1110 cm⁻¹ with ¹⁸O₂. With the hemes that our research group has utilized, having 3 or 4 of the meso-aryl groups possessing 2,6-difluoro substituents, we have observed O−O stretches at similar frequencies (Table 1). Also consistent with the assignments is our finding that the ν_Fe−O and ν_O−O of ¹⁶O/¹⁸O isotope shifts (vide supra) closely match those calculated using the harmonic oscillator model: Δ_Fe−O,calc (¹⁶O₂/¹⁸O₂) = −26 cm⁻¹, Δ_O−O,calc (¹⁶O₂/¹⁸O₂) = −67 cm⁻¹. However, it is quite common that the O−O vibrational mode of dioxygen-bound heme complexes is not observable.^10,28,45 Table 1 lists examples from our own investigations, along with several protein and synthetic oxy–heme (Fe^III−superoxide) complexes, where ν_O−O is observed. A more complete listing can be found elsewhere.¹⁰

Figure 4.

Resonance Raman spectra of ferric superoxide complex (P^ImH)Fe^III−(O₂^•−) (2) in frozen THF obtained at 77 K with 413 nm excitation: Fe−O and O−O stretching frequencies for the complex generated with ¹⁶O₂ (blue) or ¹⁸O₂ (orange). The ¹⁶O₂−¹⁸O₂ difference spectrum is shown in green. Other peaks which seem to be visible are thought to arise from heme core vibrations or weak coupling of intraheme bands with O₂ motion. The seemingly less well-defined (based on the green difference spectrum shown here) ¹⁶O₂ and ¹⁸O₂ O−O stretches are readily assignable, based on analysis of the original data, in the context of extensive experience and literature.

Table 1.

rR Stretching Frequencies (cm⁻¹) of Heme−Fe^III−Superoxide Complexes

complexa	vO-O (Δ18O2)	vFe-O (Δ18O2)	ref
Mb WT	1103	578 (−29)	46
Cyt P450 WT	1139 (−66)	546 (−31)	47
Cyt P450 D251N	1136 (−66)	537 (−30)	48
(THF)(F8)FeIII−(O2•−)	1178 (−64)	568 (−24)	49
(6L)FeIII−(O2•−)	1176 (−64)	572 (−24)	33
(PIm)FeIII−(O2•−)	1180 (−56)	575 (−23)	50
(PImH)FeIII−(O2•−) (2)	1171 (−61)	575 (−24)	this work
[(α4Fe(CO2Me)4)O2•−]	1004(−53)	581 (−22)	51
	967 (−56)
[(FeIII(OPhP))−O2•−]	1147 (−59)	570 (−22)	44

^aMb WT, wild-type myoglobin; Cyt P450 WT, wild-type cytochrome P-450 monooxygenase (P450); P450 Asp-251 mutated to Asn. See Figure S2 for diagrams of the other synthetic Fe^III−(O₂^•−) complexes.

Generation of a Reduced Heme–Cu Complex Using P^ImH.

Addition of 1 equiv of [Cu^I(CH₃CN)₄](B(C₆F₅)₄) to the reduced complex (P^ImH)Fe^II (1) led to the desired reduced heterobinuclear Fe^II/Cu^I compound [(P^ImH)Fe^IICu^I]⁺ (3), which had prominent UV–vis absorptions at 423 and 542 nm at low temperature (Figure 5). An asymmetry in pyrrole resonances of reduced complex 1 is observed in ¹H NMR spectroscopy, where different chemical shifts for inequivalent porphyrin phenyl substituents leads to observation of two peaks (at δ 50 and 59 ppm; Figure 6). Addition of the copper(I) ion leads to disappearance of upfield-shifted resonances (<0 ppm) which we assign as being derived from methylene hydrogen atoms of the tethered histamine group. Since the copper ion takes up, i.e., binds to, the tridentate chelate from this complex, the imidazolyl group of the histamine arm is no longer bound to the iron center. The pyrrolic protons in ²H NMR spectroscopy at −80 °C resonate at δ 89.4 and 102.8 ppm, typical of a five-coordinate high-spin ferrous heme (Figure 7A). Complex 3 is EPR-silent.

Figure 5.

(A) Overall scheme for reactivity of [(P^ImH)Fe^IICu^I]⁺ (3) toward O_2(g) at −80 °C in THF to yield a HS peroxo species [(P^ImH)Fe^III−(O₂²⁻)−Cu^II]⁺ (3a) and LS peroxo complex [(DCHIm)(P^ImH)Fe^III−(O₂²⁻)−Cu^II]⁺ (3b) following addition of DCHIm. (B) UV–vis spectra (−80 °C) for in situ generated HS 3a (red) and LS 3b (blue) starting following oxygenation of 3 (black).

Figure 6.

¹H NMR spectrum of (top) complex (P^ImH)Fe^II (1) and (bottom) (P^ImH)Fe^IICu^I (3) at room temperature.

Figure 7.

²H NMR spectra of (A) in situ generation of d₈-[(P^ImH)Fe^IICu^I]⁺ (3-d₈); (B) d₈-[(P^ImH)Fe^III−(O₂²⁻)−Cu^II]⁺ (3a-d₈) generated by bubbling O₂; (C) d₈-[(DCHIm)(P^ImH)Fe^III−(O₂²⁻)−Cu^II]⁺(3b-d₈) generated by the reaction of 3a-d₈ with DCHIm. The strong sharp peaks at δ 1.73 and 3.58 ppm correspond to solvent THF.

Generation of the HS Heme–Peroxo–Cu Complex.

Addition of dry O_2(g) to the THF solution of fully reduced heterobinuclear heme–copper complex [(P^ImH)Fe^IICu^I]⁺ (3) at low-temperature resulted in the formation of new species [(P^ImH)Fe^III−(O₂²⁻)−Cu^II]⁺(3a) with UV–vis spectral features at 420, 545, and 565 nm (Figure 5B), notably with two Q-band features (500–600 nm spectral region). These characteristics are the same as those known for other previously well-characterized high-spin heme–peroxo–Cu complexes also with tridentate ligands on the copper ion. (Figure 5A).^{10,31,32,49,52,53}

To provide more detail concerning the structure of [(P^ImH)Fe^III−(O₂²⁻)−Cu^II]⁺(3a), we carried out rR spectroscopy (Figure 8) and found that the peroxo ligand ν_O−O vibration is observed at 799 cm⁻¹; this shifts to 751 cm⁻¹ upon ¹⁸O₂ substitution. In addition, a band at 524 cm⁻¹ (Δ¹⁸O₂ = −23 cm⁻¹) is assigned to ν_Fe−O for this peroxide complex. Again, as discussed above for (P^ImH)Fe^III−(O₂^•−) (2), observed isotope frequency shifts when using ¹⁸O₂ match well with values calculated using the harmonic oscillator model (Δ_O−O,calc (¹⁶O₂/¹⁸O₂) = −46 cm⁻¹, Δ_Fe−O,calc (¹⁶O₂/¹⁸O₂) = −23 cm⁻¹), corroborating our assignments. While the UV–vis features for 3a are most like high-spin heme–peroxo complexes with tridentate chelates bound to copper(II) (vide supra), we note that the ν_O−O value observed for 3a (799 cm⁻¹) is much higher in energy than that found for the three other examples we have previously studied, which have tridentate ligands on the Cu(II) ion and where side-on μ-η²:η²−O₂²⁻ binding to both Fe(III) and Cu(II) occurs, ν_O−O = 747–767 cm⁻¹.^31,32,49,54 In copper–dioxygen coordination chemistry, ligand denticity induces different O₂ (as peroxide, O₂²⁻) binding modes;^55–57 Cu^I chelates with tridentate ligands prefer a side-on binding mode with a resulting peroxide ligand, giving [(L)Cu^II−(μ-η²:η²-O₂²⁻)−Cu^II(L)]²⁺ binuclear complexes whereas with tetradentate ligands (L)Cu^I complexes react with O₂ to give end-on binding, i.e., [(L)Cu^II−(μ−1,2-O₂²⁻)−Cu^II(L)]²⁺ species.

Figure 8.

rRaman data (¹⁶O₂−¹⁸O₂ difference) collected at 413 nm excitation and 77 K for HS 3a and LS 3b. Complex 2 is present as an impurity, observed as a set of ¹⁶O₂/¹⁸O₂ peaks at 575/550 cm⁻¹ (marked with an asterisk).

However, these 3a vibrations (ν_Fe−O = 524 cm⁻¹; ν_O−O = 799 cm⁻¹; Figure 8) are very close to those previously reported for the HS-TMPA heme–peroxo–copper complex (Chart 2). For this complex with tetradentate chelate bound to copper, ν_O−O = 804 cm⁻¹, and the peroxo unit is bound side-on to the high-spin Fe^III, but end-on to the Cu^II, for an overall μ-η²:η¹ coordination mode that has been supported by EXAFS spectroscopy and DFT calculations.^39,58 We posit that the P^ImH tridentate chelate induces considerable geometric distortion in the Cu(II) coordination sphere, which results in the peroxo binding to Cu in an essentially η¹ fashion. Thus, we postulate that 3a possesses a μ-η²:η¹-peroxo coordination, formulated as [(P^ImH)Fe^III−(μ-η²:η¹-O₂²⁻)−Cu^II]⁺, as shown in Figure 5A.

Chart 2.

On the basis of past extensive NMR spectroscopic studies on [(F₈)Fe^III−X−Cu^II(TMPA)]⁺ (X = O²⁻, OH⁻; TMPA = tris(2-pyridyl)methylamine),⁵⁹ the peroxo analog (X = O₂²⁻),^58,60 and [(²L)Fe^III−(μ-η²:η²-O₂²⁻)−Cu^II]^+32,61 (shown as E, Figure 1), the further analog with tridentate chelates for Cu ion,^49,59,61,62 [(P^ImH)Fe^III−(O₂²⁻)−Cu^II]⁺ (3a), has an overall electronic structure where S = 2, derived from antiferromagnetic coupling of high-spin Fe^III (S = 5/2) and Cu^II (S = 1/2) centers. This assignment is further supported by its EPR-silent behavior (12 K, in THF frozen solution). Further supporting the S = 2 electronic structure assignment, 3a displays a pyrrole signal in the paramagnetic region at 93.0 ppm (−80 °C, THF; see Figure 7B), very similar to that observed for other HS complexes which are listed in Table 2. The 93.0 ppm pyrrole resonance is in the range expected for HS ferric iron.

Table 2.

²H NMR Data and Pyrrole Chemical Shifts of Heme/Copper Dioxygen Adducts or Heme–Peroxo–Copper Complexes^a

complex	δpyrrole	solvent	temp (K)	ref
D	92	acetone	193	63
(F8)FeIIIO2CuII(LMe2N)	105	CH2Cl2/6% EtCN	178	49
E	106	THF	183	61
(F8)FeIIIO2CuII(LN4OH)	83	THF	213	64
HS-TMPA	68	MeCN	233	65
HS-AN	95	acetone	193	53
3a	93	THF	193	this work

^aSee the Figure S3 for figures of (F8)Fe^IIIO₂Cu^II(L^Me2N), (F₈)Fe^IIIO₂Cu^II(L^N4−OH), and HS-AN and Chart 2 for HS-TMPA.

Generation of a LS Heme–Peroxo–Cu Complex with P^ImH.

A low-spin analog to [(P^ImH)Fe^III−(O₂²⁻)−Cu^II]⁺ (3a) was generated by the addition of axial base DCHIm in THF at −80 °C (Figure 5A). This resulted in a drastic change in the UV–vis features (Figure 5B). Relatively intense low-energy features observed at ~800–900 nm (assigned as peroxo-to-Fe LMCT bands) are characteristic of LS μ−1,2-peroxo species, as previously demonstrated for F and G (Figure 1) and in a number of other cases (see Figure S4 for 4-DCHIm and LS-AN and Chart 2 for LS-TMPA).^34,39,66 The addition of DCHIm resulted in clear Soret and Q-band shifts (from those observed in 3a) to 425 and 538 nm, respectively (Figure 5B). ²H NMR spectroscopy confirms the spin state change from high-spin 3a (pyrrole signal at 93.0 ppm) to low-spin 3b (pyrrole signal at 10.2 ppm, i.e., in the diamagnetic region; see Figure 7). This behavior suggests that in 3b there exists antiferromagnetic coupling between the low-spin (S = 1/2) six-coordinate Fe^III and the d⁹ Cu^II (also S = 1/2) centers, through the peroxo bridge. Complex 3b is EPR-silent (12K, THF), also consistent with antiferromagnetic coupling and a resulting S = 0 electronic structure formulation.

Resonance Raman spectroscopic interrogation of LS compound [(DCHIm)(P^ImH)Fe^III−(O₂²⁻)−Cu^II]⁺ (3b) reveals the expected differences in ν_Fe−O and ν_O−O stretching vibrations in comparison to HS complex [(P^ImH)Fe^III−(O₂²⁻)−Cu^II]⁺ (3a) (Figure 8). Coordination of DCHIm to generate 3b produced two isotope-sensitive peaks: an O−O vibration at 817 cm⁻¹ (Δ¹⁸O₂ = −55 cm⁻¹) and an Fe−O stretch at 610 cm⁻¹ (Δ¹⁸O₂ = −26 cm⁻¹). This observed ν_Fe−O (¹⁶O₂/¹⁸O₂) shift is close to expected value (Δ_Fe−O,calc (¹⁶O₂/¹⁸O₂) = −27 cm⁻¹) in the harmonic oscillator approximation. The magnitude of the isotope shift for ν_O−O (= −55 cm⁻¹) is a bit larger than the calculated value (Δ_O−O,calc (¹⁶O₂/¹⁸O₂) = −47 cm⁻¹); however, it is still consistent with the assignment of this stretch to that of a peroxide ligand. Changing from HS (3a) to LS (3b) dramatically increases ν_Fe−O by 86 cm⁻¹ (Figure 5A), while a smaller increase (18 cm⁻¹) is observed for ν_O−O. Similar effects have been observed recently^39,67 for DCHIm binding and indicate a change in Fe spin state and Fe–peroxo coordination geometry. The relatively high ν_O−O stretch for the peroxo group in 3b (compared to that for 3a) indicates a LS μ−1,2-peroxo coordination mode in the former. Thus, 3b is formulated as [(DCHIm)(P^ImH)Fe^III−(μ−1,2-peroxo−O₂²⁻)−Cu^II]⁺ (shown in Figure 5A), based on our previous investigation and resulting established structure of LS-TMPA (Chart 2), that has been supported by DFT calculations.³⁹

CONCLUSION

In this paper, we investigated the dioxygen reactivity of reduced heme (P^ImH)Fe^II (1) in the absence and the presence of copper ion. Reaction of 1 and O₂ results in O₂-bound ferrous heme compound, (P^ImH)Fe^III−(O₂^•−) (2), such as that found in oxygenated myoglobin and hemoglobin. ²H NMR suggests that 2 is low-spin six-coordinate iron. The evidence points to the imidazolyl group of the appended tridentate chelate as serving as an axial ligand base. Addition of Cu(I) ion leads to a new reduced heme–copper complex (P^ImH)Fe^IICu^I (3), an advanced CcO active site model system due to (i) its tridentate ligand Cu^I-binding and (ii) the imidazolyl ligand has a free N−H group (see below). At low temperature, dioxygen reaction with 3 gives a high-spin species [(P^ImH)Fe^III−(μ-η²:η¹O₂²⁻)−Cu^II]⁺ (3a), and addition of DCHIm as an axial ligand changes the peroxo coordination to end-on, resulting in formation of LS species with end-on peroxo coordination, [(DCHIm)(P^ImH)Fe^III−(μ−1,2-O₂²⁻)−Cu^II]⁺(3b).

This oxygenation chemistry with (P^ImH)Fe^II (1) lays the groundwork for a better understanding of Fe^II/Cu^I/O₂ reactions chemistry. Our ongoing studies are focused on investigating the reductive O−O cleavage relevant to CcO enzyme active site chemistry of heme−(O₂²⁻)−Cu complexes by employing systematically varied H⁺/e⁻ sources. Also, since P^ImH has a built-in imidazole and tridentate system in copper chelation as occurs in CcO, another goal of our research is to mimic the post-translational modification which occurs in CcO, that being formation of a histidine-tyrosine cross-link which is functionally critical.¹⁶

EXPERIMENTAL SECTION

Materials and Methods.

All reagents and solvents purchased and used were of commercially available quality except as noted. Inhibitor-free tetrahydrofuran (THF) was distilled over Na/benzophenone under argon and deoxygenated with argon before use. The preparation and handling of air-sensitive compounds were performed under a MBraun Labmaster 130 inert atmosphere (<1 ppm of O₂ and <1 ppm of H₂O) glovebox filled with nitrogen. Dioxygen gas purchased from Airgas and dried by passing it through Drierite. ¹⁸O₂ gas was purchased from ICON (Summit, NJ), and ¹⁶O₂ gas was purchased from BOC gases (Murray Hill, NJ).

All UV–vis measurements were carried out using a Hewlett-Packard 8453 diode array spectrophotometer with HP Chemstation software and a Unisoku thermostated cell holder for low-temperature experiments. A 10 mm path length quartz cell cuvette modified with an extended glass neck with a female 14/19 joint and stopcock was used to perform all UV–vis experiments, as previously described.^{33,50,68 1}H and ²H NMR spectra were measured on a Bruker 300-MHz NMR spectrometer at ambient or low temperatures. Chemical shifts were reported as δ (ppm) values relative to an internal standard (tetramethylsilane) and the residual solvent proton peaks. EPR spectra were recorded with a Bruker EMX spectrometer equipped with a Bruker ER 041 × G microwave bridge and a continuous flow liquid helium cryostat (ESR900) coupled to an Oxford Instruments TC503 temperature controller. Spectra were obtained at 8 K under nonsaturating microwave power conditions (ν = 9.428 GHz, microwave power = 0.201 mW, modulation amplitude = 10 G, microwave frequency = 100 kHz, and receiver gain = 5.02 × 10³).

The compounds (P^ImH)Fe^II,³⁰ the pyrrole deuterated derivative d₈(P^ImH)Fe^II,³⁰ and Cu^I(CH₃CN)₄(B(C₆F₅)₄)³⁵ were synthesized as previously described.

UV−Vis Spectroscopy.

[(P^ImH)Fe^III−(O₂^•−)] (2). Complex [(P^ImH)Fe^III−(O₂^•−)] (2) was generated in THF solution by preparing 0.01 mM (for Soret band monitoring) or 0.1 mM solutions of (P^ImH)Fe^II (1) in a 10 mm path length quartz Schlenk cuvette, which was sealed with a rubber septum in the glovebox. The cuvette was then cooled to −80 °C, and the solution was bubbled with O₂ to generate the superoxide compound 2. UV−vis: λ_max = 424 and 541 nm.

[(P^ImH)Fe^IICu^I]⁺ (3), [(P^ImH)Fe^III−(O₂²⁻)−Cu^II]⁺ (3a), and [(DCHIm)(P^ImH)Fe^III−(O₂²⁻)−Cu^II]⁺ (3b). In a similar way, in the glovebox, 1 equiv of [Cu^I(CH₃CN)₄](B(C₆F₅)₄) (from a 2 mM stock solution (THF)) was added to the 10 mm path length quartz Schlenk cuvette containing the reduced complex (P^ImH)Fe^II (1) (125 μL of a 2 mM solution) to generate complex [(P^ImH)Fe^IICu^I]⁺ (3); the cuvette was then filled with THF up to a total volume of 2.5 mL. This cuvette was then cooled to −80 °C, and the solution was bubbled with O₂ to generate the complex [(P^ImH)Fe^III−(O₂²⁻)−Cu^II]⁺ (3a). Subsequently, addition of 1.5 equiv of DCHIm to the same cuvette results in LS complex [(DCHIm)(P^ImH)Fe^III−(O₂²⁻)−Cu^II]⁺ (3b). UV–vis: [(P^ImH)Fe^IICu^I]+ (3) λmax = 423, 542 nm; [(P^ImH)Fe^III−(O₂²⁻)−Cu^II]⁺ (3a) λ_max = 420, 545, 565 nm; [(DCHIm)(P^ImH)Fe^III−(O₂²⁻)−Cu^II]⁺ (3b) λ_max = 425, 538 nm.

EPR Spectroscopy.

(P^ImH)Fe^II (1) and [(P^ImH)Fe^III−(O₂^•−)] (2). In a glovebox, 1.0 mg of (P^ImH)Fe^II (1) was dissolved in 0.5 mL of deoxygenated THF in an EPR tube. Outside the glovebox, the solution was cooled to −80 °C (acetone–dry-ice bath) and bubbled with O₂ to generate the superoxide compound 2. Then, the EPR spectrum was recorded at 12 K. Both 1 and 2 were found to be EPR-silent.

[(P^ImH)Fe^IICu^I]⁺ (3), [(P^ImH)Fe^III−(O₂²⁻)−Cu^II]⁺ (3a), and [(DCHIm)(P^ImH)Fe^III−(O₂²⁻)−Cu^II]⁺ (3b). In a similar way, in a glovebox, 1.0 mg of (P^ImH)Fe^II and 0.9 mg of Cu^I(CH₃CN)₄(B(C₆F₅)₄) were dissolved in 0.5 mL of deoxygenated THF in an EPR tube and sealed properly. Outside the glovebox, the solution was cooled to −80 °C (acetone–dry-ice bath) and bubbled with O₂ using a syringe to generate the heme–peroxo–Cu complex [(P^ImH)Fe^III−(O₂²⁻)−Cu^II]⁺ (3a). A total of 1.5 equiv of DCHIm (3.5 mg of DCHIm in 0.5 mL of THF) was added to the reaction mixture. Then, the EPR spectrum was recorded at 12 K. Both 3a and 3b were found to be EPR-silent.

²H NMR Spectroscopy.

[d₈-(P^ImH)Fe^III−(O₂^•−)] (2-d₈). In the glovebox, 2.6 mg of d₈-(P^ImH)Fe^II (1-d₈) was dissolved in 0.5 mL of deoxygenated THF in a NMR tube. The cold THF solution (at −80 °C, acetone–dry-ice bath) of 1-d₈ was bubbled with O₂. ²H NMR (300 MHz, THF) 2-d₈: δ_pyrr 9.12 ppm.

[d₈-(P^ImH)Fe^IICu^I]+ (3-d₈), [d₈-(P^ImH)Fe^III−(O₂²⁻)−Cu^II]+ (3a-d₈), and [d₈-(DCHIm)(P^ImH)Fe^III−(O₂²⁻)−Cu^II]⁺ (3b-d₈). In the glovebox, a 1:1 mixture of 2.6 mg of d₈-(P^ImH)Fe^II (1-d₈) and 2.3 mg of [Cu^I(CH₃CN)₄](B(C₆F₅)₄) was dissolved in 0.5 mL of deoxygenated THF in a NMR tube to generate complex [d₈-(P^ImH)Fe^IICu^I]⁺ (3-d₈). The cold THF solution (at −80 °C, acetone–dry-ice bath) of 3-d₈ was bubbled with O₂. In the same NMR tube, 1.5 equiv of DCHIm was added to generate LS peroxo [d₈-(DCHIm)(P^ImH)Fe^III−(O₂²⁻)−Cu^II]⁺ (3b-d₈). ²H NMR (300 MHz, THF): Complex 3-d₈, δ_pyrr 89.4, 102.8 ppm; 3a-d₈, 93.0 ppm; 3b-d₈, 10.2 ppm.

Resonance Raman Spectroscopy.

[(P^ImH)Fe^III−(O₂^•−)] (2). In the glovebox, 2 mM solutions of an (P^ImH)Fe^II in THF were prepared and transferred to rR tubes and capped with tight-fitting septa. The sample tubes were placed in a cold bath (dry ice/acetone) and oxygenated using ¹⁶O₂ or ¹⁸O₂ gases. The oxygenated samples were set in a cold bath for 10 min, after which the sample tubes were frozen in liquid N₂ and sealed by flame. rR samples were excited at 413 nm, using either a Coherent I90C–K Kr⁺ ion laser as the sample was immersed in a liquid-nitrogen-cooled (77 K) EPR finger Dewar (Wilmad). Power was ~2 mW at the sample, which was continuously rotated to minimize photodecomposition. The spectra were recorded using a Spex 1877 CP triple monochromator, and detected by an Andor Newton CCD cooled to −80 °C. rR [Fe^III−(O₂^•−)] (2): ν_O−O, 1171 cm⁻¹ (Δ¹⁸O₂, −61 cm⁻¹); ν_Fe−O, 575 cm⁻¹ (Δ¹⁸O₂, −24 cm⁻¹).

[(P^ImH)Fe^IICu^I]⁺ (3), [(P^ImH)Fe^III−(O₂²⁻)−Cu^II]⁺ (3a), and [(DCHIm)(P^ImH)Fe^III−(O₂²⁻)−Cu^II]⁺ (3b). In the glovebox, 5 mM solutions of an equimolar mixture of (P^ImH)Fe^II and [Cu^I(CH₃CN)₄](B(C₆F₅)₄) in THF were prepared and transferred to rR tubes and capped with tight-fitting septa. The sample tubes were placed in a cold bath (dry ice/acetone) and oxygenated using ¹⁶O₂ and ¹⁸O₂. The labeled gases were cooled in dry ice for 5 min and injected through the solution by using a Hamilton gastight syringe. The oxygenated samples were set in a cold bath for 10 min, after which the sample tubes were frozen in liquid N₂ and sealed by flame. To the cold THF solution (at −80 °C, acetone–dry-ice bath) of 3a was added 1.5 equiv of DCHIm (3.5 mg of DCHIm in 0.5 mL of THF) for complex 3b. rR [(P^ImH)Fe^III−(O₂²⁻)−Cu^II]⁺ (3a): ν_O−O, 799 cm⁻¹ (Δ¹⁸O₂, −48 cm⁻¹); ν_Fe−O, 524 cm⁻¹ (Δ¹⁸O₂, −23 cm⁻¹). [(DCHIm)(P^ImH)Fe^III−(O₂²⁻)−Cu^II]+ (3b): ν_O−O, 817 cm⁻¹ (Δ¹⁸O₂, −55 cm⁻¹); ν_Fe−O, 610 cm⁻¹ (Δ¹⁸O₂, −26 cm⁻¹).