Insights into how heme reduction potentials modulate enzymatic activities of a myoglobin-based functional oxidase

Abstract

Heme-copper oxidase (HCO) is a class of respiratory enzymes that use a heme-copper center to catalyze O₂ reduction to H₂O. While heme reduction potential (E°′) of different HCO types has been found to vary >500 mV, its impact on HCO activity remains poorly understood. Here, we use a set of myoglobin-based functional HCO models to investigate the mechanism by which heme E°′ modulates oxidase activity. Rapid stopped-flow kinetic measurements show that increasing heme E°′ by ca. 210 mV results in increases in electron transfer (ET) rates by 30-fold, rate of O₂ binding by 12-fold, O₂ dissociation by 35-fold, while decreasing O₂ affinity by 3-fold. Theoretical calculations reveal that E°′ modulation has significant implications on electronic charge of both heme iron and O₂, resulting in increased O₂ dissociation and reduced O₂ affinity at high E°′ values. This work suggests fine-tuning E°′ in HCOs and other heme enzymes can modulate their substrate affinity, ET rate and enzymatic activity.

TOC image

A four-electron reduction of O₂ to H₂O in heme-copper oxidase (HCO) requires an efficient control of electron transfer, O₂ binding/dissociation rates and O₂ affinity. By employing a functional model of HCO, we show that HCOs use their heme reduction potential to control these parameters, the electronics of bound O₂ and their overall enzymatic activity.

Heme proteins are a major class of metalloproteins that utilize the same cofactor, protoporphyrin IX, for a variety of functions including electron transfer (ET, as in cytochromes), storage and transfer of O₂ (as in myoglobin and hemoglobin), and O₂ activation (as in cytochrome P450).^[1–3] A prime example of heme proteins is heme-copper oxidase (HCO) that catalyzes the kinetically challenging reduction of O₂ to H₂O and generates a trans-membrane proton gradient driving ATP synthesis.^[4–7] The catalytic site of HCO, where O₂ reduction occurs, is a binuclear heme-copper center consisting of a high spin heme iron and a copper (Cu_B) coordinated to three histidines, one of which is cross-linked to a tyrosine residue. Despite this similarity, HCOs from different species use different heme types, such as heme a, b, and o at the catalytic heme center.^[7] How different heme types impact biochemical properties HCOs is not understood. One such property is the reduction potential (E°′) of the catalytic heme (Fe³⁺/Fe²⁺), which varies by ca. 500 mV in different HCO types (see Table S1 in SI).^[8] Thus, question arises as to what is the origin for such variations of heme E°′ and how does heme E°′ impact HCO function. An efficient way to answer these questions is by systematically tuning heme E°′ and probing resulting changes in functional activity. However, such manipulations are difficult in native HCOs due to their large size (~100–200 KDa), membranous nature and presence of multiple metal cofactors. To overcome these challenges, numerous small molecule models of HCOs have been designed,^[9–10] but none to our knowledge have attempted to investigate the importance of heme E°′ in regulating HCO function.

In an approach complementary to studying complex native enzymes and their small-molecule models, we use biosynthetic modelling, that utilizes smaller proteins/peptides as simpler synthetic models while retaining structural features of native enzymes.^[11] We have designed biosynthetic structural and functional models of HCO in a smaller (17.4 KDa), easy-to-purify and soluble protein, myoglobin (Mb). To accomplish the goal, we first created a Cu_B-binding site in the distal pocket of Mb analogous to that in HCOs through L29H and F43H mutations to introduce two histidine ligands, that along with H64 complete the Cu_B coordination sphere.^[12] We then introduced a tyrosine through F33Y mutation next to histidine ligands to model the conserved tyrosine in HCO.^[13] The resultant mutant named F33Y-Cu_BMb (Fig. 1A) mimicked HCOs functionally, as it could selectively reduce oxygen to water with hundreds of turnovers.^[14] We further modulated heme E°′ of F33Y-Cu_BMb by ~210 mV (Fig. 1B) via tuning of hydrogen bonding to heme iron (through S92A mutation) and using non-native heme cofactors with increased E°′, such as monoformyl (MF-) and diformyl (DF-) hemes.^[15] The F33Y-Cu_BMb variants, thus obtained namely, F33Y-Cu_BMb, S92A-F33Y-Cu_BMb, F33Y-Cu_BMb (MF-heme) and F33Y-Cu_BMb (DF-heme) exhibited systematic increase in E°′ values of 95 ± 2 mV, 123 ± 3 mV, 210 ± 6 mV and 320 ± 10 mV respectively (all E°′ reported in this work are vs. SHE). The increase in heme E°′ for F33Y-Cu_BMb variants correlated with increases in their O₂ reduction activity (Fig. 2B). In particular, F33Y-Cu_BMb (DF-heme) with highest heme E°′ displayed ca. 6-fold higher oxidase activity than parent F33Y-Cu_BMb.^[15] In this work, we investigate the mechanism through which heme E°′ impacts O₂ reduction activity of F33Y-Cu_BMb variants. Specifically, we focus on four key factors – ET rates, O₂ binding/dissociation rates and O₂ affinity – that can be modulated through tuning heme E°′ and affect oxidase activity. Our results suggest that while the ET rates and O₂ binding/dissociation rates increase with increasing heme E°′, the O₂ affinities decrease. Overall, the study shows that heme enzymes such as HCOs use heme E°′ to control their substrate binding, electron transfer and enzymatic activities.

Figure 1.

a) Active site of F33Y-Cu_BMb showing heme b and side chains of His29, His43, His64, His93 and Tyr33. b) The heme E°′ range of HCOs that varies from −59 mV to 460 mV is represented as a grey bar while the curly brackets show heme cofactor type present in HCOs displaying those heme E°′ values. The heme E°′ of F33Y-Cu_BMb variants are shown as dotted black lines.

Figure 2.

a) Spectra obtained for 6 μM F33Y-Cu_BMb starting from Fe³⁺ form (black) going to Fe²⁺ form (grey). Isosbestic points are indicated by a star (*). Inset shows the variation in absorbance at 433 nm and 407 nm. b) Variation in oxidase activity (grey) and ET rates (black) for F33Y-Cu_BMb variants.

F33Y-Cu_BMb variants were expressed and purified without a copper at the Cu_B site. No copper was added in this work, as previous studies have shown that the presence of copper has little influence on the oxidase activity of F33Y-Cu_BMb.^[14] To elucidate how E°′ impacts enzymatic activity, we first probed variation in ET since tuning of E°′ in metalloproteins is known to modulate their ET rates.^[16–18] We reasoned that increasing E°′ of heme such that it becomes higher than that of electron donor (N,N,N′,N′-tetramethyl-p-phenylenediamine, TMPD with E°′ = 276 mV) may increase the driving force for ET. Increase in ET rates may then translate to higher O₂ reduction activity as previous reports on Mb-based HCO models revealed ET as the rate-limiting step in both enzymatic^[19] and electrocatalytic^[20] O₂ reduction reactions. To assess the role of ET, we measured the rate of reduction of Fe³⁺ to Fe²⁺ forms of F33Y-Cu_BMb variants using ascorbate (E°′ = 90 mV) as a reductant and TMPD as a redox mediator. Owing to the strong and distinct spectroscopic signatures of heme iron in its Fe³⁺ and Fe²⁺ forms, we measured the rate of heme reduction by using stopped-flow absorption spectroscopy under strictly anaerobic conditions. Upon mixing 6 μM F33Y-Cu_BMb with 1500 eq. ascorbate and 150 eq. TMPD, we observed a rapid decrease in absorbance at 407 nm, 501 nm and 618 nm (corresponding to Fe³⁺ form) with a concomitant increase in absorbance at 434 nm and 556 nm (corresponding to Fe²⁺ form). The presence of isosbestic points in the spectra confirmed a clean transformation of Fe³⁺ to Fe²⁺ with no intermediate species (Fig. 2A). Fitting the absorbance change at 434 nm and 407 nm with time allowed us to determine the ET rate of F33Y-Cu_BMb as 0.10 ± 0.05/s. Similar experiments performed with other F33Y-Cu_BMb variants also displayed a clean transition from Fe³⁺ to Fe²⁺ form (Fig. S1). A plot of ET rates vs. heme E°′, shown in Fig. 2B, indicates that as the heme E°′ increases from 95 ± 2 mV to 123 ± 3 mV, 210 ± 6 mV and 320 ± 10 mV for F33Y-Cu_BMb, S92A-F33Y-Cu_BMb, F33Y-Cu_BMb (MF-heme) and F33Y-Cu_BMb (DF-heme) respectively, the ET rates increase correspondingly from 0.10 ± 0.005/s, to 0.19 ± 0.03/s, 0.76 ± 0.02/s, and 3.19 ± 0.07/s. This increase in ET is consistent with the Marcus theory of electron transfer in a regime where ΔE° is lower than reorganization energy.^[16] These results suggest that increasing heme E°′ in HCO mimics result in an increase in ET rates, which is potentially responsible for increase in O₂ reduction activity. However, the correlation between the two rates is not linear - While F33Y-Cu_BMb (MF-heme) and F33Y-Cu_BMb (DF-heme) exhibit 8-fold and 30-fold increase in ET rates as compared to F33Y-Cu_BMb, their oxidase activity increases only 4-fold and 5-fold respectively. Therefore, while fast ET is important to HCO activity, other factors such as O₂ association, dissociation rates and O₂ affinity can also play an important role in determining HCO activity.

To investigate additional factors that determine HCO activity, we explored the role of heme E°′ in modulating rate constants of O₂ association (k_on), dissociation (k_off) and O₂ affinity. The k_on of O₂ binding to F33Y-Cu_BMb variants was measured using flow-flash technique wherein fully reduced CO-bound heme enzyme was mixed in a stopped-flow apparatus with oxygenated solution. The reaction was initiated by a short laser flash, breaking the photolabile Fe–CO bond, allowing binding of O₂ to be studied by time-resolved spectroscopy. We prepared CO-bound F33Y-Cu_BMb by reacting 5 μM of Fe²⁺ form mixed with 1.5 mM CO. The resulting complex exhibited UV-Vis signals at 422, 540 nm and 573 nm (Fig. S2), suggesting complete formation of the CO-adduct. The CO bound F33Y-Cu_BMb was then subjected to flash-photolysis and reacted with O₂ (Fig. 3A and S3). The reaction proceeded predominantly in a monophasic manner to reach heme-Fe(II)-O₂ state and subsequently studied as a function of O₂ concentration to obtain the second-order k_on for O₂ binding as 21 ± 2 mM⁻¹ s⁻¹ (Fig. 3B). Similar experiments when conducted with other F33Y-Cu_BMb variants obtained k_on as 48 ± 3 mM⁻¹ s⁻¹, 70 ± 10 mM⁻¹ s⁻¹ and 250 ± 70 mM⁻¹ s⁻¹ for S92A-, (MF-heme) and (DF-heme) variants, respectively. Thus, increasing heme E°′ by ca. 210 mV results in 12-fold increase in k_on of the Mb-based HCO models (Fig. 3C). In addition to k_on for O₂, the observed rate constants for CO binding also increased systematically by ca. 21-fold upon increasing heme E°′ (Table S2 and Fig. S8B). The reason behind this consistent increase in O₂/CO binding rates with increasing E°′ can be explained by considering the electron density of heme iron. As the E°′ increases, electron density on heme iron decreases, which favors binding of electron donating ligands like O₂/CO and results in an increase in O₂/CO binding rate constants. To provide further support to this hypothesis, we surveyed literature for k_on of different HCO types and found that R. sphaeroides cbb₃ oxidase with low catalytic heme E°′ of −59 mV exhibits 10-fold slower O₂ association rates (k_on = 11,000 mM⁻¹ s⁻¹)^[21,22] than R. sphaeroides aa₃ oxidase (k_on = 100,000 mM⁻¹ s⁻¹) with heme E°′ of 220 mV.^[22,23] Moreover, NO reductase (NOR from P. denitrificans) that also possess a low-potential heme b₃ (E°′ = 60 mV) and performs O₂-reduction cross-reactivity, binds O₂ with k_on of 25,000 mM⁻¹ s⁻¹ approximately 4-fold slower than R. sphaeroides aa₃ oxidase.^[24] Therefore, a correlation between increased heme E°′ values and increased k_on for O₂ not only exists for Mb-based HCO models but native HCOs as well.

Figure 3.

a) Kinetic difference spectra for F33Y-Cu_BMb when heme binds O₂ (black) and CO (grey). b) Observed rate constant for O₂ binding as a function of O₂ concentration; the slope of this plot is used to calculate k_on for O₂ binding and the intercept is k_off for O₂ dissociation. c) Variation in k_on (grey) and k_off (black) for F33Y-Cu_BMb variants with tuned heme E°′.

Next, we probed the impact of heme E°′ on O₂ dissociation of F33Y-Cu_BMb variants. The rates of O₂ dissociation (k_off) was extracted from the plot of observed rate constants for O₂ binding at different O₂ concentration using protocols reported previously (Fig. 3A, S5).²⁴ The k_off values were found to be 14 ± 1 mM⁻¹ s⁻¹, 11 ± 2 mM⁻¹ s⁻¹, 160 ± 10 mM⁻¹ s⁻¹ and 500 ± 40 mM⁻¹ s⁻¹ for F33Y-Cu_BMb, S92A-, (MF-heme) and (DF-heme) variants respectively. Thus, k_off values for O₂ also increased with increasing E°′ values (Fig. 3B). Specifically, the HCO model with highest E°′ (F33Y-Cu_BMb(DF-heme)) displayed ca. 35-fold enhanced k_off than that of F33Y-Cu_BMb. To explain these observations, we probed the properties of O₂ bound to heme center using density functional theory (DFT). Quantum chemical DFT calculations were performed on three O₂₋bound heme models: heme b, MF-heme and DF-heme, in which all porphyrin substituents were kept the same as in the native enzyme-based systems, except that the propionate group was replaced by methyl to facilitate the calculations. Since, O₂-bound heme is known to exist in resonance between its ferrous-oxy (Fe(II)-O₂) and ferric-superoxy (Fe(III)-O₂⁻) forms, we investigated the electronic charge, spin densities and energy of both the structures. A comparison of charge densities within the three heme types showed that O₂ molecule becomes less negative as the heme E°′ increases – The negative charge on O₂ of superoxy decreases from native heme (−0.152 e), MF-heme (−0.143 e) to DF-heme (−0.135 e). Even the charge on the O₂ molecule of oxy form became less negative with the addition of electron-withdrawing formyl groups (Table 1). These results suggest that electron-withdrawing substituents on porphyrin, as evident by higher E°′, withdraw negative charge from O₂ fragment back to iron porphyrin as also shown by less positive iron charges in Table 1. This phenomenon makes O₂ closer to a neutral state for higher E°′ values and thus, more prepared for faster dissociation. This computational trend is consistent with and explains the observed experimental O₂ dissociation rates for different F33Y-Cu_BMb variants.

Table 1.

Variation in electronic charge on iron and O₂ molecule for heme-O₂ variants namely, Fe(II)-O₂ and Fe(III)-O₂⁻ forms

Sample	QFe (e)	QO2 (e)
heme-b-Fe2+-O2	1.950	−0.051
heme-b-Fe3+-O2−	1.953	−0.152
MF-heme-Fe2+-O2	1.943	−0.044
MF-heme-Fe3+-O2−	1.946	−0.143
DF-heme-Fe2+-O2	1.934	−0.034
DF-heme-Fe3+-O2−	1.938	−0.135

Since the variation in heme E°′ affects both k_on and k_off rates, we looked at its impact on O₂ affinity (K_d) of F33Y-Cu_BMb variants. F33Y-Cu_BMb was found to exhibit a K_d of 0.7 ±0.08 mM, which is 5-fold weaker than that of WTMb (K_d =0.14 mM). Thus, adding hydrophilic residues H43, H29 and Y33 close to catalytic heme reduced its affinity for non-polar hydrophobic oxygen. Similarly, increasing hydrophobicity by addition of S92A residue close to catalytic heme iron in S92A-F33Y-Cu_BMb (K_d =0.2 ±0.04 mM) increases the O₂ affinity of heme by 3.5-fold. Finally, F33Y-Cu_BMb (MF-heme) and F33Y-Cu_BMb (DF-heme) exhibit rather weak K_d values of 2.3 ±0.4 mM and 2.0±0.6 mM respectively. Thus, the two variants with high heme E°′ reveal 3-fold lower O₂ affinity than parent F33Y-Cu_BMb explaining why an increase in their ET rates does not translate directly to an increased O₂ reduction rates. Overall, these results indicate that increasing heme E°′values leads to a decrease in O₂ affinity consistent with previous studies on Mb models that show electron-withdrawing fluoro-substituted hemes exhibiting low O₂ affinity values.²⁸ This observation is further corroborated with R. sphaeroides cbb₃ oxidase that exhibits the lowest heme E°′ value of −59 mV and also displays the lowest K_m for O₂ (7 nM) among all oxidases.²⁹ This apparent high O₂ affinity of cbb₃ oxidase helps them cope with extremely low concentration of O₂ (3–22 nM) in root legumes. Thus, tuning heme E°′ is an efficient method for HCOs to adapt to environmental constraints such as low O₂ concentration.

The complete reduction of O₂ to H₂O requires an efficient control of its ET, O₂ binding/dissociation rates and O₂ affinity. By employing a functional model of HCO with systematically tuned E°′, we show that enzymes like HCOs use their heme E°′ to control these parameters as well as the electronics of bound O₂. These results not only have significant impact in bioenergetics but also help understand how nature has fine-tuned E°′ for various metalloproteins for their optimal function. In particular, heme proteins exhibit a wide variety of heme E°′ (see Fig. S7 for few examples), and understanding the reason for this variation and associated implications on their enzymatic activity will help better understand the structure and reaction mechanism of these proteins.

Experimental Section

Experimental details pertaining to expression and purification of proteins, kinetic and computational measurements are detailed in the Supporting Information.