Rapid Decay of the Native Intermediate in the Metallooxidase Fet3p Enables Controlled Fe^II Oxidation for Efficient Metabolism

Abstract

The multicopper oxidases (MCOs) couple four 1e⁻ oxidations of substrate to the 4e⁻ reduction of O₂ to H₂O. These divide into two groups: those that oxidize organic substrates with high turnover frequencies (TOFs) up to 560 s⁻¹ and those that oxidize metal ions with low TOFs, ~1 s⁻¹ or less. The catalytic mechanism of the organic oxidases has been elucidated, and the high TOF is achieved through rapid intramolecular electron transfer (IET) to the Native Intermediate (NI), which only slowly decays to the resting form. Here we uncover the factors that govern the low TOF in Fet3p, a prototypical metallooxidase, in the context of the MCO mechanism. We determine that the NI decays rapidly under optimal turnover conditions, and the mechanism thereby becomes rate-limited by slow IET to the resting enzyme. Development of a catalytic model leads to the important conclusions that proton delivery to the NI controls the mechanism and enables the slow turnover in Fet3p that is functionally significant in Fe metabolism enabling efficient ferroxidase activity while avoiding ROS generation.

Graphical Abstract

1. Introduction

Multicopper oxidases (MCOs) are a class of metallo-enzymes that catalyze the four electron reduction of dioxygen to water, and couple this to four single electron oxidations of various substrates.^1–5 MCOs are ubiquitous in Nature, being found in all taxonomic domains, though they can be broadly divided into two groups of enzymes based on substrate specificity.⁶ The metallooxidases such as Fet3p from Saccharomyces cerevisiae, selectively oxidize transition metals including Fe^II, Mn^II and Cu^I.^7–10 The Fet3p homologue, human ceruloplasmin (hCp), and the bacterial cuprous oxidase (CueO) are examples of this group of MCOs, and are important for the homeostasis of metal ions. The other group is composed of the organic oxidases including laccases and bilirubin oxidases (BODs) that have broad specificities for organic substrates, typically diphenols.^11,12 Examples of this group include two widely studied MCOs: Rhus vernicifera laccase (RvL) and Trametes versicolor laccase (TvL). (Note that the species R. vernicifera has been reclassified as Toxiconendron vernicifluum; however, to maintain consistency with previous MCO literature, the enzyme will be referred to as RvL.) These two enzymes and other laccases are important for either lignin formation or degradation, depending on the organism. In general, the metallooxidases have turnover frequencies (TOFs) on the order of 0.01–10 s⁻¹,^7,8,13,14 significantly slower than those of organic oxidases with TOFs of 100—560 s⁻¹.^12,15,16

Structurally, the MCOs contain a minimum of four copper atoms arranged as a mononuclear Type 1 (T1) Cu site, and a trinuclear Cu cluster (TNC)^17,18 comprised of spectroscopically distinct mononuclear Type 2 (T2) and binuclear Type 3 (T3) Cu sites. The TNC is situated approximately 13 Å from the T1 site and linked via a conserved T1-Cys-His-T3 protein pathway.^19–21 During turnover, electrons donated from substrate oxidation enter the catalytic site one at a time via the T1 Cu and are transferred over the Cys-His pathway to the TNC, which is the site of O₂ reduction to water.³

The individual copper sites can be distinguished based on their unique spectroscopic features¹ and these spectroscopic handles report on the oxidation states and ligation of each Cu at various stages of the catalytic cycle. Figure 1 shows the UV-Vis absorption (Abs) and electron paramagnetic resonance (EPR) spectra of the Resting Oxidized (RO) form of Fet3p that has all four Cu’s fully oxidized.²² The T1 Cu^II site contributes a strong absorption feature at 608 nm (ϵ = 5500 M⁻¹cm⁻¹) resulting in an intense blue color. Additionally the T1 Cu^II exhibits a small parallel hyperfine splitting (|A_‖| = 88×10⁻⁴ cm⁻¹) in the EPR spectrum. Both of these features reflect the highly covalent Cu^II-S(Cys) π bond.²³ In the oxidized form, the T3 Cu^II’s are antiferromagnetically coupled via a bridging μ₂-OH⁻ ligand, resulting in EPR silence and giving rise to intense μ-OH→Cu^II₂ charge transfer transitions at 330 nm in Abs (ϵ = 5000 M⁻¹cm⁻¹). The T2 Cu^II does not contribute strongly to the Abs spectrum but exhibits ”normal” axial Cu^II features in the EPR spectrum (|A_‖| = 190×10⁻⁴ cm⁻¹).²⁴

Figure 1:

(A) UV-Vis absorption and (B) EPR spectra of Resting Oxidized Fet3p.

Numerous studies^25–28 on MCOs have defined the consensus catalytic mechanism shown in Scheme 1. Beginning with the RO form of the enzyme (lower left in Scheme 1), four electrons are donated by substrate to generate the Fully Reduced (FR) enzyme (top left, Scheme 1). The conversion of RO to FR involves three intramolecular electron transfer (IET) steps where at least the first rate is slow relative to turnover in organic oxidases (<1 s⁻¹ in RvL and TvL).^29,30 The FR enzyme rapidly reacts with O₂, reducing it to H₂O via two, 2e⁻ steps, the first of which forms the Peroxide Intermediate (PI, Scheme 1, top right) at a bimolecular rate of 2×10⁶ M⁻¹s⁻¹.^27,28,31 The second 2e⁻ step forms the Native Intermediate (NI, Scheme 1, bottom right)^32–34 at a rate >350 s⁻¹. Detailed mechanistic investigations have demonstrated that in laccases and BODs, high TOFs are achieved due to the rapid reduction of NI back to FR to continue the catalytic cycle (Scheme 1, k_NIred, middle).^30,35 The high proton affinity of the μ₃-O in the center of the TNC in NI drives fast proton-coupled electron transfer (PCET) in the rapid reduction of NI.³⁶ In the absence of excess reductant following the O₂ reaction, NI decays slowly to the RO form (Scheme 1, k_NIdecay, bottom).^37,38 Rapid re-reduction of NI to FR is the key catalytic step in the high TOFs of the organic oxidase MCOs as this rate has been determined to outcompete the decay of NI to RO by 3–5 orders of magnitude (in RvL k_NIdecay = 0.007 s⁻¹ and k_NIred > 700 s⁻¹).³⁶

Scheme 1:

Mechanism of MCO catalysis.

Herein, we evaluate the metallooxidase mechanism based on Scheme 1. Using a combination of stopped-flow Abs (SF Abs) and rapid freeze quench (RFQ) techniques to trap and spectroscopically define the catalytic intermediates, we determine that the rate of decay of NI in Fet3p is very fast and results in the RO form being the fully oxidized state of the TNC in the catalytic cycle. We next elucidate the mechanism of RO reduction in Fet3p and find that the second IET to the RO form of the TNC is thermodynamically disfavored and therefore rate-limiting in turnover. Using this modification of the mechanism in Scheme 1, we construct a kinetic model of catalysis to evaluate the functional significance of slow reactivity. This enables controlled Fe metabolism without the generation of reactive oxygen species (ROS). Finally, we use a variant of Fet3p to confirm the proton-dependence of NI decay, and underscore the importance of the proton delivery mechanism in controlling the decay of NI and the PCET steps in its reduction in the catalytic cycle.

2. Methods

General

All chemicals were of reagent grade and used as received. Water was purified to a resistivity of >17 MΩ·cm with a Barnstead Nanopure Diamond filtration system. The following Good’s buffers were prepared at 50 mM at each of the following pH values: pH 6 - MES, pH 7–7.5 - MOPS, pH 8.5 - HEPBS, pH 9.5–10 - CHES.

Enzyme expression and purification

Wild-type and mutant FET3 alleles were constructed in plasmid pDY148 by site directed mutagenesis using the QuickChange kit from Agilent. Expression and purification of Fet3 proteins was performed as previously described.¹⁴ Purified protein was characterized by the 2,2-biquinoline assay³⁹ for total Cu content and by EPR spin quantitation for EPR active Cu to verify 4 Cu/enzyme and 2 EPR active Cu^II/enzyme using a 0.945 mM CuSO₄ · 5 H₂O standard containing 20% (v/v) glycerol.

UV-Vis Spectroscopy

UV-Vis Abs measurements were collected at room temperature and at 4 °C on an Agilent 8453A spectrophotometer. Measurements below room temperature were collected with a Unisoku CoolSpeK UV USP-203 cryostat.

Steady-state turnover kinetics

Steady-state turnover kinetics were measured at 4 °C using (NH₄)₂Fe(SO₄)₂ as substrate by monitoring growth of the 315 nm band of Fe^III product (ϵ = 2200 M⁻¹cm⁻¹).^13,40 Ferroxidase reaction was initiated by adding 4 μL of freshly prepared 100 mM Fe^II solution to 2 mL samples with enzyme concentrations from 0.2–1 μM in air-saturated buffer (50 mM MES, pH 6). Linear fitting of the Fe^III formation rates vs. pH gave the steady-state TOF. Measurements were repeated three times to determine uncertainty.

Stopped-flow Abs

Stopped-flow Abs experiments were performed using an Applied Photo-physics SX.19 MV-R with a PDA1 Photodiode array detector. The SF instrument was installed in a custom purge glovebox (Cleatech Isolation Glove Box 2100) equipped with an oxygen sensor (Neutronics Model 1100) and maintained in an Ar atmosphere with < 0.05% O₂. Temperature control was maintained at 4 °C with a circulating water bath (Fisher Scientific Isotemp 3016). Enzyme samples were prepared separately in an anaerobic glovebox (Vacuum Atmospheres) maintained with a N₂ atmosphere at < 0.5 ppm O₂. All solutions were prepared anaerobically before being transferred via gas-tight syringes (Hamilton) to the SF. Before each experiment, the internal syringes and tubing of the SF instrument were made anaerobic using 10 mM solutions of sodium hydrosulfite, followed by extensive rinsing with deoxygenated buffer.

Concentrated Fet3p and RvL solutions were degassed using a N₂ purge, transferred into the glovebox, and allowed to equilibrate for at least 12 hours prior to any experiment. For Fet3p RO reduction SF experiments, the enzyme was in the resting oxidized form as verified by UV-Vis Abs and EPR. This solution was diluted to a concentration of 70 μM before 1:1 mixing in the SF with anaerobic solutions of reductant containing either 210 mM hydroquinone or freshly prepared 70 μM ferrous ammonium sulfate (dissolved in 100 μM H₂SO₄ to prevent hydrolysis). Fully reduced enzyme was prepared by reducing the resting enzyme with excess sodium hydrosulfite, followed by extensive buffer exchange using 30 kDa centrifugal filters (Amicon) to remove excess reductant. For SF experiments to monitor NI formation and decay, these solutions were used directly and mixed in the SF with O₂-saturated buffer. For SF experiments designed to monitor T1 reduction immediately following turnover with O₂, reductant was added to the anaerobic enzyme to match the concentrations in the RO reduction experiments (210 mM hydroquinone or 70 μM Fe^II) before 1:1 mixing with O₂ solutions. Pre-mixing concentrations of RvL SF experiments were 100 μM and hydroquinone concentrations were ~100 mM.

Rapid freeze quench EPR and MCD

RFQ EPR samples were prepared by mixing the reduced enzyme with O₂ saturated buffer using a Rapid Freeze Quench system (Kintek RFQ-3) and ejecting samples into cold isopentane maintained at −120 °C with an isopentane/liquid N₂ slurry. These samples were then tightly packed into an EPR tube and the isopentane was removed. MCD samples were prepared in deuterated buffer and both the enzyme and O₂ solutions were pre-mixed with 60% (v/v) glycerol as glassing agent. MCD samples were frozen by mixing directly into an assembled MCD cell and submerging in liquid N₂. The MCD cell consisted of two quartz discs separated by a 3 mm rubber spacer inside a specially designed Cu sample holder. RFQ EPR and MCD samples were prepared anaerobically following the same protocol as in SF Abs experiments but at higher enzyme concentrations (~800 μM). Internal syringes and tubing of the RFQ instrument were rinsed with sodium hydrosulfite and deoxygenated buffer prior to experiment.

EPR Spectroscopy

EPR spectra were measured at X- and Q-band using Bruker ER 041XG/ER 051QT microwave bridges equipped with ER4119HS or ER4116DM (X-band) and ER5106QT (Q-band) resonators. At 77 K, X-band EPR was collected with a liquid N₂ finger dewar, and at Q-band with a CF935 Dynamic Continuous Flow Cryostat (Oxford). Low temperature X-band EPR was collected using an ITC5 03 temperature controller with an ESR900 Continuous Flow Cryostat (Oxford). EPR simulations were performed using the EasySpin toolbox⁴¹ for MATLAB (v7.14.039).

CD and MCD spectroscopy

CD and MCD spectra were recorded on a Jasco J810 spectropolarimeter using an extended S-20 photomultiplier tube. CD measurements were recorded at 4 °C maintained with a circulating water bath (Fisher Scientific Isotemp 1006S). For MCD measurements, the sample was placed in an Oxford Instruments SM4000 7T superconducting magnetooptical dewar which cooled the sample and supplied the magnetic field.

SF Abs data fitting

Exponential fits to SF kinetic data were fit using single and biexponential fitting functions in OriginPro 2020 (v9.7.0.188). Simulations of data to Schemes 2–5, and Scheme S1 and estimations of kinetic rate constants were performed using the Copasi software.⁴² Single wavelength traces (608 nm) were used for most simulations. For NI formation and decay in wt Fet3p singular value decomposition (SVD) was first performed in MATLAB and these SVD traces were simultaneously fit to extract rates and projected spectra of each species (FR, NI, and RO) in the reaction.

Scheme 2:

T1 reduction followed by IET to the TNC

Scheme 5:

O₂ Reaction Scheme for Y133R Fet3p

Steady-state turnover rates and membrane-bound kinetic model

Copasi software was used to simulate steady-state turnover kinetics for wt and Y133R Fet3p under substrate saturating conditions. The membrane-bound model was simulated similarly with the following assumptions: a) the T1 self-exchange reaction was not possible, and therefore excluded, due to the localization of each Fet3p enzyme to a membrane surface; and b) [Fre1p] = [Fet3p] = 1μM on the cell surface. This surface concentration was chosen given an upper limit of 1.7 μM Fet3p calculated based on the average diameter of a yeast cell (6 μm)⁴³ and the dimensions of the Fet3p ferroxidase domain estimated from the crystal structure (surface area ~100 nm², PDB:1ZPU).

3. Results and Analysis

3.1. Kinetics of Reduction of Resting Oxidized versus After Single Turnover

3.1.1. Determination of Steady-State Turnover Frequency

Earlier studies of Fet3p reported k_cat for Fe^II oxidation to be 0.16 s⁻¹ (pH 6, 25 °C).¹⁴ For comparison with single turnover data presented here, a steady-state rate for Fe^II oxidation at lower temperature (pH 6, 4 °C) under saturating substrate conditions was measured to be 0.053(±0.003) (Fe^III/s/Fet3p) (Figure S1, Table S1). This decrease in the steady-state rate is consistent with the decreased temperature used in the present study.

3.1.2. Reduction with Hydroquinone

Analysis of the catalytic mechanism of Fet3p began with an appraisal of the relative rates of reduction of the T1 and the TNC during different steps in the catalytic cycle. These mirror experiments reported in our previous study³⁵ on RvL, in which there is a dramatic rate enhancement for intramolecular electron transfer (IET) between the T1 and the TNC in the reduction of the enzyme immediately following a single reoxidation with O₂ (i.e. the reduction of NI in Scheme 1) relative to the IET for reduction of the resting oxidized enzyme.

First, reduction of the RO form of Fet3p with a large excess of hydroquinone (~3000 enzyme equivalents) was monitored via SF Abs. Upon mixing RO enzyme with excess reductant, the absorption features at 608 nm and 330 nm (corresponding to oxidized T1 and T3 sites, respectively) both decayed within 30 sec (Figure 2A). These T1 and T3 Abs traces were used to evaluate IET rates between the T1 and the TNC (Figure 2B). The reduction trace of the T1 required a biexponential fit with two components each contributing 50% to the intensity lost with rates of 0.55 and 0.11 s⁻¹. This complicated behavior of the T1 feature is due to the fact that its reduction by excess hydroquinone is occurring at a comparable rate to the reoxidation of the T1 that results from the IET from the T1 to the TNC. The decay of the T3 intensity, on the other hand, can be fit to a single first order process with a rate of 0.11 s⁻¹. The lower reactivity of Fet3p with organic substrates (for T1 reduction by hydroquinone, $k_{red}^{\text{H2Q}}=20{\text{M}}^{-1}{\text{s}}^{-1}$)⁴⁴ results in these two reduction processes occurring at similar rates; however, the overall slower reduction of the T3 relative to the T1 in the RO reaction is a result of the slow IET between the T1 and the TNC in the first reduction step of RO with a rate of ~0.1 s⁻¹.

Figure 2:

Stopped-flow absorption spectra and wavelength traces after mixing RO Fet3p with excess hydroquinone. (A) Representative Abs spectra between 1 ms and 20 s after mixing showing reduction of the T1 and T3 Cu^II Abs bands. (B) Abs intensity traces monitored at 330 and 608 nm. Reduced Fet3p spectrum was subtracted from all spectral data. Colored arrows in A match traces in B and indicate direction of intensity change. Fits to the data are shown as gray dashed lines in panel B with rates indicated in the text. Post-mixing conditions: [RO-Fet3p] = 35 μM, [hydroquinone] = 103.5 mM, pH 6, 4 °C.

Next, the reduction of Fet3p immediately following reoxidation by O₂ was monitored by SF Abs. In this experiment FR enzyme in the presence of a large excess of hydroquinone was mixed with one enzyme equivalent of O₂. Because the reaction of Fet3p with O₂ is fast (10⁶ M⁻¹s⁻¹) relative to reduction by hydroquinone (20 M⁻¹s⁻¹), the enzyme is essentially fully reoxidized before its reduction by hydroquinone. Indeed, as shown in Figure 3A, the enzyme is rapidly reoxidized within the first 200 ms after mixing with O₂, based on the reappearance of the T1 and oxidized TNC/T3 features at 608 and 330 nm, respectively. In the presence of excess hydroquinone, these features then decay within 30 sec (Figure 3B) just as was observed in the reduction of RO. Fitting the T1 reduction kinetics (Figure 3C) required a similar biexponential fit to T1 reduction with two components of equal intensity contributions having rate constants of 0.71 and 0.14 s⁻¹. The reduction of the T3 intensity was fit with a single first order decay rate of 0.12 s⁻¹, which is within error of the rate observed in RO reduction. These data demonstrate that with hydroquinone as the electron source, the reduction of Fet3p immediately following turnover is kinetically identical to RO reduction. Importantly, under these conditions there is no rate enhancement of IET between the T1 and the TNC in contrast to the behavior observed in NI reduction of RvL and other organic oxidase MCOs (vide infra).

Figure 3:

Stopped-flow absorption spectra and wavelength traces after mixing FR Fet3p with O₂ in the presence of excess hydroquinone. (A) Representative Abs spectra between 1 ms and 200 ms showing growth of T1 and TNC Abs bands. (B) Abs spectra between 200 ms and 20 s showing decay of T1 and T3 Abs bands. (C) Specific wavelength traces of 330 nm and 608 nm SF Abs data. Reduced Fet3p spectrum was subtracted from all spectral data. Colored arrows in panels A and B match traces in C and indicate direction of growth or decay. Fits to the data are shown as gray dashed lines in panel C with rates indicated in the text. Post-mixing conditions: [FR-Fet3p] = 35 μM, [hydroquinone] = 103.5 mM, [O₂] = 35 μM, pH 6, 4 °C.

3.1.3. Comparison of Fet3p with R. vernicifera laccase

In our previous study, the high TOF in RvL was found to arise from rapid IET in NI reduction resulting from the high driving force of the all-bridged structure of the NI TNC.³⁶ For a direct comparison to the Fet3p reduction kinetics, analogous experiments to those described above were performed on RvL. Reaction of RO RvL with a large excess (~1000 enzyme equiv) of hydroquinone resulted in a slow reduction of the T3 site (0.2 s⁻¹) relative to the fast reduction of the T1 site, which reduced with biphasic kinetics where 70% of the intensity was lost at a rate of 30 s⁻¹ and 30% at a rate of 3 s⁻¹ (Figure S2). Importantly, under equivalent conditions of excess reductant, reduction after a single turnover with O₂ (i.e. NI reduction) results in a much more rapid loss of the TNC features (6.5 s⁻¹) and relatively little change to the T1 reduction rate (3 s⁻¹, Figure S3). These data reflect the 10⁵-fold increase in the T1 to TNC IET rate for reduction of NI relative to RO in RvL. Figure 4 highlights the differences in the SF Abs traces of the T1 (blue) and T3/TNC bands (red) in the reduction of these two enzymes. The significant increase in the NI reduction rate in RvL relative to RO reduction (compare 365 and 330 nm (red) traces in panels C and A in Figure 4) is clearly not observed in the reduction of Fet3p immediately after turnover (compare 330 nm (red) traces in panels D and B). This is consistent with RO (rather than NI) being the form of the enzyme relevant in the reduction of Fet3p following turnover with O₂.

Figure 4:

Stopped-flow absorption traces after mixing Fet3p and RvL with excess hydroquinone. (A) and (C) show loss of T1 and T3/TNC bands with ~1000 equiv hydroquinone reacting with RO RvL and NI RvL, respectively. (B) and (D) show loss of T1 and T3 bands with ~3000 equiv hydroquinone reacting with RO Fet3p and immediately after O₂ turnover. Fits to the data are shown as gray dashed lines.

3.1.4. Reduction with Fe^II

The poor selectivity of Fet3p for organic substrates (e.g. the hydroquinone used above) results in sluggish T1 reduction rates even in the presence of a large excess of substrate. In order to maximize the T1 reduction rate and increase the possibility of observing rapid IET to a putatively formed NI, the native substrate, Fe^II, was employed as it reduces the T1 Cu^II center at >1200 s⁻¹.⁴⁴ Stopped-flow experiments analogous to those with hydroquinone were performed using the fast Fe^II substrate as the reductant. Because the T1 is rapidly reduced by Fe^II and the resultant Fe^III has strong Abs at wavelengths below 400 nm, the IET between the T1 and the TNC was monitored via the recovery of the T1 Abs intensity following its reduction using stoichiometric amounts of Fe^II (1 Fe^II/enzyme). Reaction with either the RO form or immediately following turnover with O₂ results in the rapid reduction of the T1 (within several ms) followed by the return of the T1 Abs intensity as the electron is transferred to the TNC (Schemes 2 and 3).

Scheme 3:

Oxidation of FR Fet3p and re-reduction of the T1 Cu followed by IET to the TNC

The reaction of RO Fet3p with 1 equiv of Fe^II follows Scheme 2 where the resting enzyme that contains a fully oxidized T1 and oxidized TNC (T1oxTNCox) reacts rapidly with Fe^II to reduce the T1 Cu (T1redTNCox). The electron is then transferred to the TNC, which reoxidizes the T1 (T1oxTNCred1). Figure 5A shows the SF Abs spectra immediately following Fe^II reduction of RO. As anticipated, the rapid reduction of the T1 site was complete within the first 5 ms after mixing, followed by the slow return of the T1 Cu^II signal. This can be well fit to a first order process with a rate (k_{IET 1}) of 0.24(±0.03) s⁻¹.

Figure 5:

Stopped-flow absorption spectra and wavelength traces showing IET from T1 to the TNC in Fet3p after reduction by Fe^II. (A) SF Abs spectra over 20 sec after reaction of RO Fet3p with 1 equiv Fe^II. (B) SF Abs spectra over 20 sec of Fet3p reaction with 1 equiv Fe^II immediately following a single turnover with O₂. (C) Wavelength traces of 608 nm SF Abs data from panels A (dashed line, RO Fet3p) and B (solid line, Fet3p after turnover). Reduced Fet3p spectrum was subtracted from all spectral data. Colored arrows in panels A and B match wavelength traces in C and indicate direction of growth. Post-mixing conditions: [Fet3p] = [Fe^II] = 35 μM, [O₂] = 0 or 1000 μM, pH 6, 4 °C.

The parallel experiment monitoring the first IET after turnover was performed by mixing O₂ with fully reduced Fet3p pretreated with 1 additional equiv of Fe^II. In these experiments, excess O₂ was used to maximize the NI formation rate (k_ox). This reaction is described by Scheme 3 where a fully reduced enzyme (FR) is rapidly reoxidized by O₂ to form a fully oxidized enzyme (4×Cu^II, where TNCox could be either the NI or RO form of the TNC). The oxidized T1 site is then immediately re-reduced by the 1 equiv of Fe^II and the electron is subsequently transferred to the TNC. Monitoring the SF Abs we observe rapid reoxidation of the enzyme, based on the return of the Abs feature at 330 nm, while the T1 Cu^II intensity does not return due to its immediate re-reduction by Fe^II. Instead, the slow growth of the oxidized T1 Abs intensity follows a first order process with a rate of 0.27(±0.01) s⁻¹. Both the rate of reoxidation and the final intensity of the T1 signal are identical (within error) to the those observed following the 1e⁻ reduction of RO (Figure 5C). These data show that even under conditions that maximize the amount of reduced T1 Cu, the dominant IET process following turnover is the same as in the reduction of RO.

Collectively, these experiments demonstrate that rapid reduction of NI is not part of the catalytic cycle in Fet3p. In the context of the MCO mechanism, this behavior could be consistent with two possibilities: either (a) NI is not formed in Fet3p or (b) NI is formed but rapidly decays at a rate much greater than the IET rate from the T1 to the NI TNC. These possibilities were tested by first determining whether NI could be trapped and spectroscopically characterized in Fet3p.

3.2. Formation, Spectroscopy, and Decay of the Native Intermediate in Fet3p

3.2.1. Native Intermediate Spectroscopy

SF Abs was used to investigate the formation and decay of NI in the reaction of FR with O₂ in the absence of reductant. The MCO mechanism in Scheme 1 (clockwise from FR) dictates that this should result in the rapid formation of NI that then decays to RO. The Native Intermediate has previously been shown to be stabilized at elevated pH.³⁷ Indeed, upon reaction of FR Fet3p with O₂ at elevated pH (pH 8.5, whereas the studies in 3.1 were done at pH 6), there is rapid growth (within 100 ms) of broad Abs intensity between 300–475 nm as well as the growth of the T1 feature at 608 nm (Figure 6A). This is followed by a decay that is evident based on the decay of Abs intensity at 420 nm (Figure 6B). The position, intensity and kinetic profile of the Abs features observed in the SF are consistent with the formation and decay of an NI species that has been observed in other MCOs.^{13,30,32,33,45}

Figure 6:

Stopped-flow absorption spectra and wavelength traces after mixing FR Fet3p with O₂. (A) Abs spectra between 1 ms and 100 ms showing growth of T1 and NI Abs bands. (B) Abs spectra between 100 ms and 10 s showing decay of NI Abs bands. (C) Specific wavelength traces of SF Abs data. Colored arrows in panels A and B match traces in C and indicate direction of growth or decay. Conditions: [Fet3p] = 35 μM, [O₂] = 1 mM, pH 8.5, 4 °C.

To verify the assignment of this O₂ reaction intermediate as NI, a rapid freeze-quench (RFQ) EPR sample was prepared to trap the intermediate at 80 ms after mixing. The EPR signals of NI are unique and have been well characterized in studies on several other MCOs.^30,33 At 77 K, its EPR spectrum has the same T1 Cu^II signal as in RO, but lacks the T2 Cu^II intensity despite having a fully oxidized TNC, just as in the RO form. This is because the NI form of the TNC contains a μ₃-oxo bridging all three TNC Cu^II’s as well as a μ₂-hydroxo bridge between the T3 Cu atoms; these oxo and hydroxo ligands are derived from O-O cleavage.³⁸ This unique bridging geometry of the TNC leads to a spin frustrated ground state with a fast relaxing S_tot = 1/2 (which is not observable by EPR at 77 K) rather than the isolated T2 Cu^II EPR signal.³⁴ Alternatively, at lower temperature (<20 K) this S_tot = 1/2 NI ground state exhibits a broad EPR signal with g < 2 due to the field-induced mixing with a low lying doublet excited state of the all bridged 3×Cu^II NI TNC (Scheme 1, bottom right). Figure 7A shows both the 77 K EPR spectrum with the lack of the T2 Cu^II signal and the 5 K spectrum with a broad negative feature at g = 1.89. This g-value is higher than that in RvL (g = 1.65), but at similar values to the NI seen in both TvL³⁰ and hCp.¹³

Figure 7:

(A) EPR spectra of RO (blue, top) compared with NI Fet3p at 77 K, 10 mW (black, middle), and 5 K, 100 mW (black, bottom). (B) RFQ-MCD spectrum of NI Fet3p measured at 4 K, 7 T

Shifting the reaction conditions to pH 10 allowed for the decay of NI to be slowed sufficiently (vide infra) to obtain RFQ samples suitable for magnetic circular dichroism (MCD) spectroscopy. The MCD spectrum of NI in Fet3p exhibits an intense derivative-shaped, pseudo-A term (two MCD C terms with opposing signed intensity) centered at ~30,000 cm⁻¹. This pseudo-A term is diagnostic of the all-bridged structure of the TNC of NI.^33,46 Also present in the MCD spectrum are oxidized T1 features between 13000–19000 cm⁻¹.⁴⁴ Simultaneous fitting of the NI MCD and Abs spectra (extracted from fitting of SF Abs data, vide infra) revealed that the band positions and intensities of the NI in Fet3p are nearly identical to those of the NI in RvL (Figure S4 and Table S2).³³

3.2.2. pH Dependence of the Decay of the Native Intermediate

SF Abs experiments to monitor NI formation and decay were performed between pH 6 and pH 10. At each pH, singular value decomposition (SVD) analysis of the time evolution of the SF Abs spectra was performed to fit the NI decay rates and extract RO, NI, and FR Abs spectra (except pH 6, where no NI formation/decay was required to fit the data). The extracted spectra from the SVD analysis agreed with known Abs spectra of NI in RvL³³ and TvL³⁰ and with those of FR and RO Fet3p (Figures S5–S10). The fitted NI decay rates (Table S3) are plotted in Figure 8 as log(k_NIdecay) versus pH. The linear relationship of these data with a slope of −0.96(±0.04) indicates that the decay of NI is first order in [H⁺]. The lack of a pK_a for NI decay rates, in contrast to the pK_a of 3.7 observed in TvL,³⁰ implies that the source of the proton in Fet3p NI decay is bulk solvent rather than proton transfer being mediated by a nearby protonatable residue. Extrapolating the best fit line to the data yields a NI decay rate of 750(±300) s⁻¹ at pH 6. This rate of NI decay represents a 10⁵-fold increase over the rate of NI decay in RvL and a 10³-fold increase over NI decay in TvL and hCp (at optimal pH for turnover for each enzyme).

Figure 8:

Plot of log(k_NIdecay) vs pH from SF Abs fitting.

The decay of NI at a rate of 750 s⁻¹ at pH 6 is consistent with the lack of any observable NI feature in the SF Abs data at this pH (Figure S5B). To experimentally determine a lower limit on the NI decay rate at pH 6, simulations were performed of the 420 nm intensity following the O₂ reaction at pH 6 (Figure S11). The simulations included NI formation (using the O₂ reaction rate of 1×10⁶ M⁻¹s⁻¹) and varying the NI decay rate from 1–1000 s⁻¹. The 420 nm extinction coefficient of NI (1200 M⁻¹cm⁻¹) was taken from experiments at higher pH. In order to adequately fit the pH 6 SF data, the NI decay rate would need to be > 500 s⁻¹. This lower limit on the NI decay rate agrees with the extrapolated NI decay rate of 750 s⁻¹.

The experiments outlined above establish that NI forms in Fet3p, but decays too rapidly to be efficiently re-reduced as part of the catalytic cycle. Thus, the catalytic cycle of Fet3p is contingent upon the reduction of RO, not NI in the mechanism (Scheme 1, thin arrows). We therefore focused on the mechanistic steps in RO reduction to understand the nature of the slow turnover in Fet3p.

3.3. Kinetic Model of Fet3p Catalysis

3.3.1. Thermodynamics of Reduction of Resting Oxidized

Having demonstrated that RO is in the catalytic cycle of Fet3p, the development of a kinetic model of Fe^II turnover required investigation of the individual mechanistic steps of RO reduction. In Section 3.1.4, we established the kinetic profile of the first IET process of RO reduction. Assessment of subsequent IET steps began with an anaerobic titration using additions of stoichoimetric amounts of Fe^II to the resting enzyme (Figure 9). After each addition, the T1 Cu^II was rapidly reduced, leading to an initial loss of the 608 nm band, followed by its slow reappearance over tens of seconds as the electron was transferred to the TNC. For the addition of the first equiv of Fe^II, the kinetic profile of the 608 nm band matched that measured in the SF (Section 3.1.4) with 90% of the T1 intensity reappearing at a unimolecular rate of ~0.2 s⁻¹ (= k_{IET 1,obs}). Subsequent additions of a second and third equiv of Fe^II resulted in slower return of the T1 signal and to a lower maximum T1 intensity at equilibrium. While the growth rates of the T1 signal were multiphasic, approximate fits with first order rate constants yielded k_{IET 2,obs} = 0.05 s⁻¹ and k_{IET 3,obs} = 0.01 s⁻¹.

Figure 9:

T1 Abs intensity versus time for multiple Fe^II additions. Inset shows spectra after equilibration following each Fe^II addition.

Although IET in RO reduction became slower as more equivalents were added, it was not possible to parse individual kinetic rates for this mixture of IET processes. Instead, a description of the thermodynamics of RO reduction was sought to elucidate these IET steps. An anaerobic titration was performed with addition of Fe^II into RO Fet3p in 0.5 equiv steps and the UV-Vis Abs and EPR spectra were recorded. In both sets of data, the T1, T2, and T3 sites were found to reduce at different rates. The Abs spectra are shown in Figure S12 and agree with the data from the above titration experiment (Figure 9). In addition to the anticipated loss of T1 and T2 Cu^II EPR signals as the enzyme was reduced, there was an appearance of a new Cu^II signal that gained and then lost intensity as more electron equivalents were added (Figures S13–S15).

In order to determine the speciation of the EPR active species throughout the titration it was necessary to determine the EPR parameters of the new Cu^II signal. X- and Q-band EPR spectra were measured for a sample treated with three equiv Fe^II and simulations of these data (Figure S16) yielded the parameters of this new Cu^II species (Table S4). This species showed an increased g_‖-value and a decreased |A_‖| relative to a typical T2 Cu^II in addition to a rhombic splitting with g_z > g_y > g_x and significant Cu hyperfine coupling in the lowest g-value. This EPR signal is similar to, though more rhombic than, that previously characterized as the half-reduced T3 site in the alternative resting form of the TNC in two MCOs with high potential T1 sites from Magnaporthe oryzae⁴⁷ and Podospora anserina.⁴⁸ This new Cu^II species is thus assigned as a half-reduced T3 site (T3-HR, Figure 10).

Figure 10:

EPR spectra of RO Fet3p overlaid with sample containing 2.5 equiv Fe^II. Hyper-fine features of g_‖ values for T1, T2, and the half-reduced T3 (T3-HR) signals are indicated. Inset: enlarged g_‖ region.

With the parameters of the new Cu^II species established, the speciation of the EPR titration data consisting of oxidized T1, T2, and T3-HR was determined (Figure S17). Notably, the intensity of the T1 signal in the 77 K EPR data was substantially lower than the T1 Abs at 4 °C, and this difference matched the intensity of the T3-HR signal. This indicated that upon lowering the temperature from 277 K to 77 K, a fully reduced T3 site transfers one electron back to an oxidized T1 Cu. Accounting for this temperature-dependent redis-tribution of electrons, a stoichiometric analysis of the Fe^II titration data yielded the total amount of oxidized T1, T2, and T3 sites under the kinetically relevant (4 °C) conditions (Table S5). Importantly, these data require the T3 center to either be fully oxidized or fully reduced at 4 °C and therefore this site behaves as a two-electron acceptor, consistent with previous MCO literature.^48,49

These data alone, however, do not address the question of whether the T2 and T3 sites are reduced independently, or if reduction of the TNC occurs in a stepwise fashion. To illustrate this distinction, consider the 1e⁻ reduced RO sample (Table S5) where 47% of the T2 sites are reduced and 21% of the binuclear T3 sites are fully reduced. If these sites are reduced independently we would anticipate (0.47 × 0.21 =) 11% of the molecules to have a fully reduced TNC (i.e. both T2 and T3 reduced). Alternatively, in a stepwise reduction mechanism (where the T2 must be reduced before the T3) these data would correspond to 21% of the molecules with fully reduced TNC Cu’s and 26% having TNCs where only the T2 was reduced. To distinguish between these two possibilities, a circular dichroism (CD) experiment was designed wherein a partially reduced sample of Fet3p was exposed to O₂. Based on our previous studies only a fully reduced TNC will react with O₂.^24,50,51 This 3×Cu^I TNC will first form PI (2×Cu^II 1×Cu^I), and then, when the T1 site is reduced, go on to form NI (see Scheme 1). Any PI formed in the absence of a reduced T1 site would not be competent to perform the next 2e⁻ reduction step of PI to generate NI, but would instead decay slowly (0.001 s⁻¹ at pH 6 from Type 1-depleted (T1D) Fet3p studies)⁵¹ to the RO form. In contrast, any partially reduced TNC (i.e. with only T2 or T3 reduced) would be unreactive with O₂. Therefore, if the T2 and T3 sites are reduced independently there would be up to 11% PI formation, whereas if the TNC Cu’s are reduced sequentially this should form 21% PI. Upon exposure of a sample of 1e⁻ reduced Fet3p to O₂, PI formation was confirmed by CD spectroscopy (compared with the PI spectrum in T1D Fet3p, Figure S18) and its decay matched the expected rate of 0.001 s⁻¹ (Figure 11). The amount of PI formed was 20(±3)% of the total enzyme (based on the known intensity of the PI CD signal of Fet3p, Δϵ = 4 M⁻¹cm⁻¹ at 28,000 cm⁻¹). This agrees with the amount of reduced T3 (20%) determined from Abs and EPR for this sample prior to the O₂ reaction, indicating that the reduction of the TNC occurs sequentially, with the first electron reducing the T2 Cu, and the second and third electrons together reducing the T3 Cu center.

$$\text{T1red}\text{+}\text{TNCox}\rightleftharpoons \text{T1ox}\text{+}\text{TNCred1}$$ (1)

$$\text{T1red}\text{TNCred1}\rightleftharpoons \text{T1ox}\text{+}\text{TNCred2}$$ (2)

$$\text{T1red}\text{+}\text{TNCred2}\rightleftharpoons \text{T1ox}\text{+}\text{TNCred3}$$ (3)

Based on these findings the speciation data of T1, T2, and T3 sites could be mapped onto a model of RO reduction wherein the T1 Cu sites are in equilibrium with 1, 2, or 3 electron reduced TNC sites as in Equations 1–3, each having a specific 1e⁻ reduction potential (Table 1). Simulation of RO reduction using this thermodynamic model yielded good agreement with the titration data (Figure 12). Previous studies have established the reduction potential of the T1 site in Fet3p to be 434 mV.⁵² At 4 °C, the first electron into the TNC is thermodynamically favorable with a potential of 488 mV while the second has an unfavorable potential relative to the T1 of 340 mV. The third electron, however, has a high reduction potential of 615 mV resulting in an effective two electron reduction process to fully reduce the T3 site in the presence of a reduced T2.

Figure 11:

CD spectra recorded after O₂ reaction of 1 equiv Fe^II-reduced Fet3p sample. Spectra shown are at 80, 250, 520, 870, 1490, and 2820 s after mixing with O₂. The CD spectrum of RO has been subtracted from all spectra. Inset shows log plot of intensity at 27173 cm⁻¹ feature vs time. Arrows indicate direction of change of major CD bands.

Table 1:

Reduction Potentials of Electron Accepting Sites in Fet3p

Electron Acceptor	E° (mV)
T1	434
TNCox	488
TNCredl	340
TNCred2	615

Figure 12:

Cu^II speciation of Fet3p during reduction titration. Oxidized T1 (blue squares), T2 (green circles), and T3 (red triangles) sites determined from UV-Vis and EPR measurements. Dotted lines are simulation to the data from sequential reduction model.

3.3.2. Kinetic Model of Reduction of Resting Oxidized

With an understanding of the thermodynamics of the reduction of RO, we were able to construct a kinetic model for RO reduction consistent with turnover. The minimal model for RO reduction involves four T1 reduction steps and three reversible IET steps between the T1 and the TNC reduced by 0–3 electrons (Scheme 4). The bimolecular rate of the T1 reduction by Fe^II ($k_{\mathit{\text{red}}}^{\mathit{\text{Fe}}}$) was estimated to be > 1 × 10⁷ M⁻¹s⁻¹ based on SF Abs data in Section 3.1.4 where the T1 was reduced within 5 ms, which is consistent with previous literature values for T1 reduction by Fe^II.⁴⁴ The reduction potentials in Table 1 enabled calculation of equilibrium constants (K^IET) as well as forward rates (k_IET) from Marcus Theory.

Scheme 4:

Fet3p Reduction Model. Equilibrium constants are related to forward and reverse IET rates by ${\text{K}}^{\text{IET}n}=\frac{k_{\mathit{\text{IETn}}}}{k_{\mathit{\text{IETn}}}^{-1}}$

Using the Marcus equation,⁵³ the ratio of two IET rates, k₁ and k₂, can be rearranged into Equation 4.⁵⁴ The rate for the first IET from the T1 to the TNC was determined from the SF Abs data in Section 3.1.4 to be k_{IET 1} = 0.24 s⁻¹ (Figure 5C), Previous work on RvL estimated the reorganization energy (λ) of the 1st IET to RO to be 1.5 eV.³⁵ If the subsequent IET steps are assumed to have the same reorganization energy, then the reduction potentials in Table 1 for RO reduction (ΔG° in Eqn 4) yield rates of 0.01 s⁻¹ and 2.6 s⁻¹ for the second and third IET steps, respectively. Thus, the second electron into the TNC is rate determining.

$$\frac{k_1}{k_2}=\sqrt{\frac{{\lambda }_1}{{\lambda }_2}}\text{exp}\left(\frac{-{\left(\mathrm{\Delta }{G}_1^{\circ }+{\lambda }_1\right)}^2}{4{\lambda }_1{k}_{B}T}\right)/\text{exp}\left(\frac{-{\left(\mathrm{\Delta }{G}_2^{\circ }+{\lambda }_2\right)}^2}{4{\lambda }_2{k}_{B}T}\right)$$ (4)

Also included in the RO reduction model is the T1 Cu self exchange interaction, which is necessary to mediate the equilibrium between buried TNC sites. Inclusion of this T1 Cu self exchange rate with a bimolecular rate constant of 10⁵ M⁻¹s⁻¹ is required to accurately model the IET behavior observed in the SF. Self exchange rates are reported to vary from 10³–10⁶ M⁻¹s⁻¹, largely based on the proximity of the T1 Cu to the enzyme surface.^55,56

Simulation of the SF Abs data for the first IET to the RO TNC (from Section 3.1.4) using the parameters in Table 2 gives good agreement to the data (Figure 13A). The full anaerobic reduction titration can also be simulated and gives good agreement with the observed rates of each IET step (k_IET,obs = 0.2, 0.05, and 0.01 s⁻¹) and equilibrium concentrations of T1 Cu^II (Figure S19). This RO reduction model was therefore used as part of a complete model of Fet3p catalysis.

Table 2:

Kinetic Rates for Catalytic Cycle of Fet3p

Rate	Value
kIET1	0.24 s−1
kIET2	0.01 s−1
kIET3	2.6 s−1
$k_{red}^{Fe}$	1 × 107 M−1s−1
KIET1	9.6
KIET2	0.01
KIET3	1965
kT1self-exchange	105 M−1s−1
kNIdecay	750 s−1
kNIred	100 s−1

Figure 13:

T1 Abs trace and simulation of SF Abs data of (A) reoxidation following 1 equiv Fe^II reduction of RO and (B) reoxidation following 1 equiv Fe^II reduction of turnover Fet3p.

3.3.3. Kinetic Model of Fet3p Catalysis

The final requirement for the description of the catalytic cycle (Scheme 1) and to complete the kinetic model, was an estimate of the IET rate between the T1 and the NI form of the TNC generated by O₂ reduction. As noted above, following a single turnover, the 1st IET is dominated by the same slow process as the reduction of RO (0.24 s⁻¹). However, previous work estimated that the 1st IET rate to the NI form of the TNC was >700 s⁻¹ (in RvL³⁵), which is comparable to the NI decay rate determined here for Fet3p (k_NIdecay = 750 s⁻¹ at pH 6). This suggested that some amount of NI reduction could occur at a rate competitive with the decay of NI to RO. With a working model of RO reduction, the rate of IET to NI was estimated using the SF Abs data for the 1e⁻ reduction following turnover (Section 3.1.4, Figure 5). The RO reduction model was expanded to include NI formation and decay steps using rates determined in Section 3.2.2 for the O₂ reaction of a fully reduced enzyme. Once NI was formed, the model was expanded to include a step where the the T1 site in NI could be rapidly reduced by Fe^II prior to decay, and this afforded a potentially rapid IET step to reduce NI (k_NIred, Scheme S1). Such a rapid IET to NI would appear in the T1 SF Abs data as a significantly faster formation of the T1 Cu^II signal at early times. Simulations of time-dependence of the T1 Abs intensity for the first IET in the reduction of Fet3p following turnover with O₂ (data from Figure 5B) were performed with IET rates to NI (k_NIred) varying from 1 to 700 s⁻¹ (Figure S20). The best fit to the data gives an IET rate to NI of 100 s⁻¹ (Figure 13B).

This completes the mechanistic model of the Fet3p catalytic cycle. The overall rate-limiting step is the second IET to RO, which is necessarily in the catalytic cycle based on the rapid decay of NI. Given the derived rate constants (Table 2), under substrate saturating conditions the steady-state TOF is predicted to be 0.04 (Fe^II/s/Fet3p), which is in good agreement with the measured rate of 0.053(±0.003) (Fe^II/s/Fet3p). Because the rates of IET to NI and NI decay are so similar, the model predicts that under substrate saturating conditions 10% of the reaction proceeds through NI reduction, while 90% of the reaction involves the RO reduction pathway.

3.4. NI in the Y133R Fet3p Variant

3.4.1. Formation and Decay Kinetics and Spectroscopy of the Native Intermediate in Y133R Fet3p

Based on the proton dependent mechanism of NI decay proposed³⁸ by Yoon et al., and in light of its 1st order [H⁺] dependence in Fet3p (Figure 8), we hypothesized that a positive charge near the TNC might stabilize NI. This could result in slower NI decay and potentially increase the fraction of the enzyme proceeding through NI reduction during turnover. Therefore, a variant of Fet3p was expressed in which a Tyr near the T3 edge of the TNC (Y133, Figure S21) was mutated to an Arg. This enzyme expressed with a full complement of 4 Cu per enzyme and exhibited Abs and EPR features of RO similar to those of the wild-type enzyme (Figure S22). SF Abs experiments monitoring the reaction of the fully reduced enzyme with O₂ were performed between pH 6–9.5. The O₂ reaction occurred at rates comparable to wild-type (wt) Fet3p, but, unlike in the wild-type enzyme, the formation and decay of NI and an earlier O₂ intermediate (i.e. PI) were observed in the variant at all pH values tested. At pH 7.5, the reaction of fully reduced Y133R Fet3p with O₂ resulted in the rapid growth of Abs features during the first 7 ms that correspond to oxidized T1 and TNC sites (Figure 14A). Between 7 ms and 300 ms, the fast decay of the first intermediate was observed based on the loss of intensity at 480 nm concomitant with the continued growth of the T1 Abs band, resulting in an isosbestic point at 515 nm (Figure 14B). Importantly, this first intermediate (that maximizes at 7 ms) cannot be NI since this would require a fully oxidized T1³³ that has not yet formed, but instead increases as the first intermediate decays. Following the complete formation of the T1 Abs by 300 ms, the slow decay of a feature at 365 nm was observed at a rate of ~0.15 s⁻¹ (Figure 14C). (Note that the Y133R Fet3p variant was sensitive to photo-bleaching in the SF as evidenced by the slow loss of T1 intensity (<0.01 s⁻¹) in Figure 14C.)

Figure 14:

SF Abs data for O₂ reaction of Y133R Fet3p showing the two O₂ intermediates during (A) the first 20 ms, (B) 20 ms-1 sec, and (C) from 1 sec-200 sec. Arrows indicate direction of growth or decay of Abs intensity described in text. Post-mixing conditions: [Y133R-Fet3p] = 35 μM, [O₂] = 1 mM, pH 7.5, 4 °C

The second intermediate (that maximizes at 300 ms) was assigned as NI based on its intensity at 365 nm that decays over 20 sec. This assignment was confirmed by RFQ EPR and MCD spectra (at pH 6) where the unique spectroscopic features of NI were observed (Figures S24 and S25 and Table S7) and found to be nearly identical to those in wtFet3p at elevated pH (see Section 3.2.1). The first intermediate of the O₂ reaction (Figure 14A) was then assigned as PI (Scheme 1), based on its decay leading to NI formation and the similarity of its Abs spectrum to the known PI Abs spectrum (i.e. intensity at wavelengths < 600 nm and lack of T1 Cu^II intensity).²⁶

At all pH values between 6–9.5, the O₂ reaction of fully reduced (FR) Y133R-Fet3p produced the same spectral features, and similar kinetic rates of these PI and NI intermediates. The SF Abs data were simulated (Figure S23) using the model shown in Scheme 5 and rates were determined for both the PI to NI ($k_{\mathit{\text{PIdecay}}}^{\text{Y133R}}$) and the NI to RO ($k_{\mathit{\text{NIdecay}}}^{\text{Y133R}}$) conversions (Table S6). The PI decay rates showed a small pH dependence, slowing by less than one order of magnitude from 110 to 23 s⁻¹ between pH 6 and 9.5, while the NI decay rates remained unchanged over this pH range with an average rate of 0.13(±0.05) s⁻¹.

The electronic structure of the NI in this variant was not perturbed relative to NI in the wild-type enzyme based on the similarity of their EPR and Abs/MCD spectra; however, the decrease in both the PI and NI decay rates in the variant indicate that proton transfer has been inhibited to these two forms of the TNC during the O₂ reduction reaction. The NI decay rate was decreased by a factor of 10³ (at pH 6) and the pH dependence (relative to the wt enzyme) was eliminated (Figure 15) indicating that the mechanism of proton transfer in the decay of NI has been greatly impacted. The PI decay rate in Y133R Fet3p was also substantially slower than in wtFet3p, where the PI to NI conversion is not observable in the data, and has been estimated to be >350 s⁻¹.²⁸ Previous work in our lab has also demonstrated that this PI to NI decay involves O-O cleavage with a proton transfer to one of the O₂-derived ligands bridging the TNC Cu’s (Scheme 1, right), and a computational model has identified both a proton-assisted and a proton-unassisted pathway for the PI to NI conversion.³¹ The dramatic decrease in the PI decay rate and its pH dependence are consistent with either a dramatically lower proton affinity or decreased proton accessibility to the active site.

Figure 15:

NI Decay Rates vs pH for wt (black circles) and Y133R Fet3p (blue squares).

3.4.2. Reduction of the Resting Oxidized and Native Intermediate Forms of Y133R Fet3p

SF Abs experiments analogous to those in Section 3.1.4 were performed on Y133R Fet3p, wherein 1 equiv of Fe^II was reacted with either the RO enzyme (Scheme 2) or with the enzyme immediately after turnover (Scheme 3). Monitoring the return of the T1 Abs band after its reduction by 1 equiv of Fe^II in RO yielded a rate of 0.5(±0.2) s⁻¹ for the 1st IET from the T1 to the TNC (Figure 16), which is comparable to that of wtFet3p. The analogous experiment for the first IET in the reduction of NI in this variant gives a rate of 1.5(±0.2) s⁻¹ (Figure 17). Thus, by slowing down the rate of NI decay, the first IET to NI in this variant has become 10× faster than the rate of NI decay (0.13 s⁻¹); however, this also represents a 10²-fold decrease relative to the estimated IET rate for NI reduction in wtFet3p (100 s⁻¹). Observation of this slower rate of IET to the NI in Y133R Fet3p, which is a PCET process, is consistent with the decrease in proton affinity or solvent accessibility observed in the slow decay of NI in Y133R Fet3p.

Figure 16:

SF Abs spectra (A) and T1 absorption intensity (B) for IET following RO reduction by 1 equiv Fe^II. A fitted curve to the T1 Abs trace is shown as a gray dashed line in B.

Figure 17:

SF Abs spectra (A) and T1 absorption intensity (B) for IET following NI reduction by 1 equiv Fe^II. A fitted curve to the T1 Abs trace is shown as a gray dashed line in B.

3.4.3. Steady-State Fe^II Oxidation by Y133R Fet3p

The above single turnover experiments in the Y133R Fet3p variant demonstrated that the rates for NI decay and for the first IET to NI were decreased by factors of 10³ and 10², respectively, relative to the wild-type enzyme. Thus in the variant, the NI decay rate was 10× slower than the rate of IET to NI (0.13 vs 1.5 s⁻¹), whereas, in wt, the NI decay rate was ~ 10× faster than IET to NI (750 vs 100 s⁻¹). Because the mechanism for fast turnover in the organic oxidases involves the rapid reduction of the long-lived NI, we investigated whether this stabilization of NI in the Y133R Fet3p variant would lead to faster turnover.

Measurement of the steady-state rate of Fe^II oxidation in the variant gave a TOF of 0.09(±0.02) (Fe^III/s/Fet3p). This is consistent with the application of the general model of Fet3p catalysis developed in Section 3.3.3 with the following assumptions: a) the rates of the second and third IET steps in RO reduction are equal to those of wtFet3p; and b) the entire NI reduction process (a 3e⁻ process) is rate limited by the first IET as measured in SF. This yields a TOF for the variant of 0.1 (Fe^III/s/Fet3p), which is in line with the measured value of steady state turnover. However, it is important to emphasize that the flux of the reaction proceeding through NI versus RO reduction has inverted relative to wild-type, with 90% now proceeding through NI reduction and 10% through RO reduction (versus 10 and 90% in wt). Although the TOF of 0.09(±0.02) (Fe^II/s/Fet3p) is slightly faster than the wild-type turnover rate (0.053 s⁻¹, see Table S1), it remains 10³× slower than turnover in RvL (560 s⁻¹).¹² Thus, the Y to R mutation was successful in stabilizing NI and favoring NI reduction over RO reduction during turnover in Fet3p, but it also slowed the rate of IET to NI and resulted in only a modest increase in the TOF.

4. Discussion

Metallooxidases typically exhibit TOFs > 10²-fold slower than the TOFs of organic oxidases.^{7,8,11–15,57} Indeed, the TOF of Fet3p for Fe^II oxidation is 10³× slower than oxidation of diphenols by RvL (0.16 vs. 560 s⁻¹ at 23 °C), and was measured here to be 0.053 s⁻¹ at 4 °C. The key to high turnover frequencies achieved by laccases and BODs is the rapid re-reduction of the NI form of the enzyme to the O₂-reactive FR enzyme, bypassing the slow-reduction of the RO form (Scheme 1 and Figure 18, left, dark red).³⁵ Reduction of RO is slower than NI reduction because it lacks the μ₃-oxo of the TNC of NI. This bridging μ₃-oxo ligand provides a large driving force for PCET due to its high proton affinity upon reduction.³⁶ In this study, we have determined that Fet3p rapidly generates the same NI in the O₂ reduction steps as present in the fast MCOs,³³ but exhibits very slow turnover due to the rapid decay of NI to RO (750 s⁻¹). This rate of NI decay is fast enough to outcompete the NI reduction process (from single turnover SF), where the first IET to NI rate is estimated to be 100 s⁻¹. This results in RO being the relevant fully oxidized form of the enzyme in the catalytic cycle of the metallooxidases and leads to the overall slow turnover of the enzyme (Figure 18, right).

Figure 18:

Comparison of the catalytic cycles in organic oxidases vs metallooxidases. The key steps of fast NI reduction in organic oxidases vs fast NI decay in metallooxidases that govern TOFs in these two groups of enzymes are highlighted in yellow.

We have elucidated the mechanism of RO reduction in Fet3p, a process that requires a total of four single electron reductions of the T1 Cu and three IET steps from the T1 to the TNC to reach the fully reduced form (Scheme 4). Reduction of the T1 occurs rapidly in the presence of Fe^II (> 1 × 10⁷ M⁻¹s⁻¹) and would not be rate-limiting except under severe Fe deprivation. Reduction of the RO TNC occurs sequentially with the T2 being the first to reduce in Fet3p, followed by reduction of the T3 Cu’s. The first electron reduces the T2 Cu of the TNC that has a reduction potential of 488 mV, enabling its thermodynamically favorable reduction by the T1 Cu (E° = 434 mV). The second IET is thermodynamically disfavored with a reduction potential of 340 mV, and is the rate-limiting step in RO reduction. The high driving force of the third electron into the TNC (615 mV) results in the T3 site behaving as an effective two-electron acceptor. This is consistent with the model of RO reduction proposed by Kjaergaard et al. for reduction of P. anserina laccase.⁴⁸ In that study, a computational model showed that the hydrogen bonding network imposed by protein constraints prevented the 2e⁻ reduced TNC (where the T2 was reduced and 1e⁻ was shared between the T3 Cu’s) from relaxing to the thermodynamically favored alternative resting form in which the bridging μ₂-OH bond is broken. The barrier to this relaxation effectively destabilized the first electron into the T3 center, allowing the second electron into the T3 to drive the reduction, resulting in a concerted two-electron process.

The catalytic model developed here for Fet3p is similar to that of recent work on hCp where the rates of NI reduction and NI decay were comparable, leading to turnover being dominated by the slow process of RO reduction.¹³ This is likely common to the metallooxidases and responsible for their slow turnover. By tuning the relative rates of NI reduction and NI decay, a MCO can control the TOF over a range spanning four orders of magnitude (~0.05–560 s⁻¹, in hCp and RvL, respectively). This effectively functions as a switch to funnel the reaction through the rapid NI reduction process for a high TOF (Figure 18, left) or through the slow RO reduction leading to a low TOF (Figure 18, right).The evolutionary advantage to the organism favoring fast or slow turnover would then depend on the role of the enzyme in its metabolic process. In Fet3p, the slow turnover is advantageous based on its function in Fe homeostasis.

The Fet3 protein is part of the high affinity Fe^III trafficking complex in fungi, and is required for respiratory growth under low Fe conditions.⁵⁸ The role of Fet3p in S. cerevisiae is to oxidize Fe^II, provided by a membrane bound reductase, Fre1p, to Fe^III.⁵⁹ This ferric ion is then efficiently transferred to Ftr1p, a metal permease responsible for shuttling Fe^III through the cell membrane. Structurally, Fet3p has a typical three domain MCO architecture,²⁰ but with an additional transmembrane domain that serves to tether the enzyme to the plasma membrane and contains the carboxy-terminus important for the interactions that ensure membrane co-targeting with Ftr1p.⁶⁰ The genes encoding this Fe transport pathway are upregulated only under limiting Fe conditions, and the substrate of Fet3p, Fe^II, would be even more scarce in an aerobic environment. Enzyme turnover is therefore necessarily limited by substrate availability, which is provided by the reductase Fre1p with a reported V_max of 0.7 s⁻¹.²²

Under these conditions, Fet3p is ideally suited to its role in this Fe transport pathway for several reasons. First, the oxidation of Fe^II by the T1 Cu is very rapid (>1200 s⁻¹),⁴⁴ preventing loss either to the environment or to competition from other organisms for labile Fe^II generated by the action of the reductase. Second, the rate-limiting step in turnover is the slow IET to the TNC. Given the V_max of Fre1p (0.7 s⁻¹), the IET rate in Fet3p (0.01 s⁻¹ at 4 °C) is similar to, but slower than, the expected flux of Fe^II available to the enzyme. This slow reduction of the TNC lowers the fraction of enzymes having a fully reduced TNC in the presence of an oxidized T1 Cu. As demonstrated above, this form of the enzyme (T1 oxidized, TNC fully reduced) would react with O₂ to form PI, that would decay to produce reactive oxygen species (ROS) that are toxic to cells.⁶¹ Finally, previous studies have demonstrated the importance of the Fet3p-Ftr1p interactions for proper Fe homeostasis in yeast.^60,62 The Fe^III trafficked into the cell via Ftr1p is provided by Fet3p, with tight Fe binding and a slow rate of transport (V_max < 1 × 10⁻³ s⁻¹).²² Slow turnover by Fet3p is therefore preferable, since rapid accumulation of extracellular Fe^III could overwhelm the capacity of the permease, leading to loss of Fe^III rather than transport into the cell. This would be metabolically costly to the organism.

The model of Fet3p catalysis developed here highlights the functional significance of slow turnover in fungi. Simulations of a membrane-bound model of Fet3p catalysis was performed using an Fe^III concentration of 1 μM and steady state reduction to Fe^II by Fre1p. The model predicts minimal formation of PI in Fet3p under these conditions, and therefore low accumulation of ROS from PI decay ([ROS] production was < 0.5 nM/hr). If the rates of NI decay and NI reduction were altered in the model to match those of $R\upsilon \text{L}\left(k_{\mathit{\text{NIdecay}}}^{R\upsilon \text{L}}=0.01{\text{s}}^{-1},k_{\mathit{\text{NIred}}}^{R\upsilon \text{L}}=700{\text{s}}^{-1}\right)$ with all other parameters held constant, the rate of ROS generated increases by three orders of magnitude (50 μM/hr). This demonstrates the importance of the slow turnover of Fet3p in avoiding ROS generation during Fe metabolism. The rapid decay of NI, therefore, is an important feature in Fet3p function.

A computational model of NI decay was developed in a previous study³⁸ where protonation of the μ₃-oxo leads to H₂O being extruded from the center of the TNC. This process of water extrusion would occur after proton transfer to NI. Using the same model for NI decay in Fet3p, the increase in this decay rate with increasing [H⁺] indicates that proton transfer to NI, rather than water extrusion, is the rate limiting step in its decay. The NI decay rate in Fet3p is faster by 4–5 orders of magnitude than in other MCOs^13,30,37 (at their optimal pH for turnover), and is first order in proton concentration. In contrast, the NI decay rate in TvL increases only slightly from 0.1 to 0.6 s⁻¹ between high and low pH with an apparent pK_a of 3.7.³⁰ In RvL, the rate increases from 0.05 to 2.1 s⁻¹ between pH 7.4 and 4 (at 23 °C),³⁷ though a similar pK_a was not established. In neither enzyme do the data agree with a first order dependence on [H⁺], nor are the decay rates accelerated as greatly as in Fet3p. This is evidence of a mechanistic difference for NI decay in Fet3p that relies on proton transfer to NI from bulk solvent, rather than proton transfer being mediated by a nearby protonatable residue.

Comparison of the structures of TvL and Fet3p do not reveal obvious structural differences that could account for the difference in proton accessibility of the TNCs in these two enzymes. The first coordination spheres of the TNC Cu’s are conserved as are nearly all of the second sphere residues. This includes the Y133 residue (Fet3p numbering), and the carboxylic acids determined from multiple MCO studies^62–64 to be relevant for proton transfer steps in the O₂ reaction (D94 and E487 in Fet3p), which overlay well between these two structures. One possible explanation for the difference in proton availability between Fet3p and TvL may relate to the the two solvent channels that lead to the TNC. Residues along these solvent channels are also mostly conserved between these two enzymes, but subtle changes could impact the hydrogen bonding network and influence μ₃-oxo protonation via a Grotthuss mechanism.⁶⁵ One potential perturbation that could influence proton transfer is on the outside of the solvent channel leading to the T3 edge of the TNC, where a Leu and Ser in TvL on the lip of this channel are substituted with a Thr and Asp in Fet3p (Figure S26). The side chains of these residues are ~4 Å from the crystallographic waters in the channel but are on a flexible loop and could move closer in solution.

The reactivity of the Y133R Fet3p variant highlights important factors controlling proton transfer steps in the catalytic mechanism of MCOs. The elimination of the pH dependence for the NI decay rate in the Y133R variant (Figure 15) indicates that the mechanism has changed relative to wild-type and that the proton leading to the decay of NI is no longer donated by bulk solvent. The rate of NI decay also decreased by a factor of 10³ relative to wild-type (at pH 6). Similarly, the rates of IET to NI and of the PI to NI conversion, both requiring protons, were slower than in wild-type by factors of 100 and ~10, respectively. As noted in Section 3.4.1, these diminished rates are consistent with either a decrease in proton affinity of O₂ intermediates of the TNC due to the positive charge of the Arg, or an increase in the kinetic barrier for proton transfer due to a structural rearrangement in the Y to R variant. Examination of the Fet3p crystal structure reveals that the Y133 residue is nearby, though not directly involved in the network of water molecules leading from the TNC through the solvent channel to the exterior of the protein discussed above (Figure S27). The indirect involvement of this residue in a putative proton transfer pathway disfavors a kinetic argument for slow proton transfer to the TNC in the variant. Rather, it is likely that introducing the positive charge of the Arg leads to a lower proton affinity of the TNC decreasing the decay of NI and its PCET.

Interestingly, however, the NI decay rate in RvL is slower than in wtFet3p by a factor of 10⁵, despite the rate of IET to NI increasing by 10-fold. This underscores the importance of controlled proton delivery in MCO turnover and, in contrast to the Y133R Fet3p variant, is more consistent with a kinetic barrier to proton transfer in NI decay. In the fast MCOs, including RvL, to ensure a slow decay of NI, protons must be transferred slowly, but also available for rapid proton coupled electron transfer steps, with a PCET barrier being lower than the barrier for proton transfer alon. These steps include the rapid cleavage of the O-O bond in the PI to NI conversion and in the PCET to rapidly reduce the μ₃-oxo ligand of NI.

Overall, the catalytic model developed for Fet3p illustrates the importance of the NI decay rate on the mechanism of MCO turnover. The controlled proton transfer is responsible for either slow or fast decay of NI. By balancing the rates of NI reduction versus decay, enzyme turnover can be finely tuned to match its specific biological function.

5. Conclusion

The slow turnover of metallooxidases relative to organic oxidases was investigated in the ferroxidase Fet3p, and found to be controlled by the rapid decay of the NI to the RO form. This results in the RO form being the fully oxidized state of the TNC in the catalytic cycle. The balance between fast and slow NI decay via proton transfer thereby enables tuning of the TOF of MCOs. Slow reduction of RO is a result of the low driving force for the first electron transfer to reduce the T3 site, whereas the high driving force for the subsequent second electron enables the T3 to behave as a two-electron acceptor. A complete model of the catalytic cycle of this metallooxidase demonstrates that slow turnover in Fet3p is functionally significant for avoiding ROS generation in Fe metabolism. Finally, the kinetic results using a Fet3p variant demonstrate a mechanistic difference relative to MCOs with high TOFs that highlights the necessity of controlled proton delivery to intermediates in the MCO catalytic cycle.