The fully oxidized form of the cytochrome bd quinol oxidase from E. coli does not participate in the catalytic cycle: direct evidence from rapid kinetics studies

Abstract

Cytochrome bd catalyzes the two-electron oxidation of either ubiquinol or menaquinol and the four-electron reduction of O₂ to H₂O. In the current work, the rates of reduction of the fully oxidized and oxoferryl forms of the enzyme by the 2-electron donor ubiquinol-1 and single electron donor TMPD have been examined by stopped-flow techniques. Reduction of the all-ferric form of the enzyme is 1000-fold slower than required for a step in the catalytic cycle, whereas the observed rates of reduction of the oxoferryl and singly-reduced forms of the cytochrome are consistent with the catalytic turnover. The data support models of the catalytic cycle which do not include the fully oxidized form of the enzyme as an intermediate.

1. Introduction

The cytochorme bd quinol oxidase is one of the two terminal oxidases in the respiratory chain of E. coli, and catalyzes the 2-electron oxidation of either menaquniol or ubiquinol and the 4-electron reduction of oxygen to water [1–5]. This enzyme is a member of a large family of respiratory oxidases that are found in many prokaryotes, both bacteria and archaea. There is no eukaryotic homologue, and the presence of bd-type oxidases in a number of human pathogens makes this enzyme a possible drug target. Cytochrome bd is a transmembrane heterodimer, with one copy each of two proteins coded by the genes CydA (subunit I; 58 kD) and CydB (subunit II; 43 kD)[6]. There is no sequence homology between cytochrome bd and the large superfamily of the heme-copper respiratory oxidases [7], and the bd-type oxidases do not contain copper[6]. During turnover, cytochrome bd does not pump protons but it generates a proton motive force due to the fact that the protons used in the reaction to make H₂O come from the bacterial cytoplasm, whereas the protons from the oxidation of QH₂, which is oxidized near the opposite side of the membrane, are released to the periplasm [8,9].

Cytochrome bd contains three heme prosthetic groups to transfer electrons from quinol to oxygen: heme b₅₅₈, heme b₅₉₅ and heme d. The low spin heme b₅₅₈ is believed to be the entry site of electrons from quinol and is located near the putative quinol binding site. Heme d is the site where oxygen binds and is reduced to water. The role of the high spin heme b₅₉₅ is still not clear, but heme b₅₉₅ is located about 10 Å from the high spin heme d [10] and minimally its role is to provide an electron to facilitate the reduction of oxygen bound to heme d. The membrane potential appears to be generated essentially by proton uptake accompanying the electron transfer from heme b₅₅₈ to the oxygenated heme b₅₉₅/heme d center, which generates a product with a high affinity for protons [8].

The isolated cytochrome bd appears to have a single binding site for quinols. Unlike the heme-copper oxidases that utilize ubiquinol as a substrate, such as cytochrome bo₃ from E. coli, there is no evidence for a tightly bound quinone which does not readily exchange with the quinone pool in the membrane. Hence, the isolated cytochrome bd has no bound quinone and the fully reduced form of the enzyme contains three electrons for redox chemistry, corresponding to the ferrous forms of each of the three hemes.

Four different states of heme d have been characterized spectroscopically in cytochrome bd. These are 1) ferric (Fe³⁺); 2) ferrous (Fe²⁺); 3) ferrous-oxy complex (Fe²⁺-O₂) and 4) oxoferryl (Fe⁴⁺=O²⁻). A putative peroxy complex (supposedly, Fe³⁺-OOH) has been observed recently as a transient intermediate during oxidation of the fully reduced enzyme by O₂ [11], but its structure requires further confirmation. Each of the other four forms of heme d can be generated as stable or metastable forms that can be examined and characterized. These forms suggest a sequence of events at heme d as O₂ is converted to water.

$$F{e}^{2+}\underset{+O_2}{\to }F{e}^{2+}{O}_2\underset{+e^{-}+H^{+}}{\to }F{e}^{3+}O-OH\underset{\begin{array}{l}+e^{-}+H^{+}\\ -H_{2}O\end{array}}{\to }F{e}^{4+}=O^{2-}\underset{\begin{array}{l}+e^{-}+2H^{+}\\ -H_{2}O\end{array}}{\to }F{e}^{3+}\underset{+e^{-}}{\to }F{e}^{2+}$$ (1)

The analysis of enzyme turnover, however, must take into account the redox status of the other two heme prosthetic groups, heme b₅₅₈ and heme b₅₉₅, which can exist in either the Fe²⁺ or Fe³⁺ states, and also the fact the ubiquinol is a 2-electron donor.

The purpose of the current work is to directly test the proposal based on the steady-state kinetics studies [12,13] that the fully oxidized ( $b_{558}^{3+}{b}_{595}^{3+}{d}^{3+}$) form of cytochrome bd may not be part of the catalytic cycle (Figure 1).

Figure 1. Catalytic cycle of ubiquinol oxidase activity of cytochrome bd.

The scheme is based on the work of Jünemann et al [12] as modified by Matsumoto et al [13].

The model shown in Figure 1 implies that the rate of reduction of the all-ferric form of the enzyme should be slow. Contrary to this proposal, fast reduction of heme d in the all-ferric enzyme was observed in pulse radiolysis studies [14]. The generation of a strong 1-electron reductant in this experiment resulted in very rapid reduction of heme b558, equilibration within 10 μs between heme b₅₅₈ and heme b₅₉₅, followed by electron transfer to heme d with the rate constants of about 3400 s⁻¹ and 590 s⁻¹, for the ferric and the oxoferryl forms, respectively. The 1-electron transfer to heme d in the ferrous-oxy form (Fe²⁺-O₂) was about 100-fold slower than observed with the ferric and oxoferryl forms. Taken at face value, these data suggest that the all-ferric form should be reconsidered as a possible intermediate during steady state analysis, and raises questions about the catalytic competence of the ferrous-oxy complex as an intermediate. An important difference between the steady-state turnover and pulse radiolysis experiments is that the former normally utilizes a 2-electron donor whereas the latter utilizes a 1-electron reductant. This was pointed out by Kobayashi et al [14], in reference to the slow single-electron transfer to the ferrous-oxy complex.

In the current work, a rapid mixing/diode array spectrophotometer was used to monitor the rate of reaction of UQH₂-1 and of ascorbate-reduced TMPD with the all-ferric form of cytochrome bd. The kinetics of these reactions were compared with the reactions with the oxoferryl form of the enzyme generated by transient oxidation of the fully-reduced cytochrome bd by a stoichiometric amount of oxygen. The results reveal very slow (seconds) reduction of the all-ferric enzyme by ubiquinol-1, whereas reduction of the oxoferryl form of cytochrome bd is at least 1000-fold more rapid and compatible with the turnover number of the enzyme. With a single-electron donor TMPD, the oxoferryl form of cytochrome bd is also a much better electron acceptor than the all-ferric form, however, the difference is much less pronounced than in the case of 2-electron reductant UQH₂-1. The data show directly that the all-ferric form of the enzyme is not likely to be part of the normal catalytic cycle.

2. Materials and Methods

Strains and Plasmids

E. coli strain GO105 (cyd AB::kan, cyo, recA), which lacks both the cytochrome bo₃ and cytochrome bd quinol oxidases [15], was used as the host strain for overexpressing the wild type cytochrome bd from plasmid pTK1 [15] which was introduced into the strain.

Cell Growth and Enzyme Purification

Large scale cell growth was carried out in 24 2-liter flasks shaking at 220 rpm 37 °C using an Innova 4330 incubator shaker (New Brunswick Scientific) [16]. Cytochrome bd was purified as described previously [17]. The pooled fractions were concentrated using an Amicon concentrator with a 50 kDa molecular weight cut-off filter and then dialyzed three times against 50 mM sodium phosphate buffer, pH 7.8, containing 5 mM EDTA, 0.05% N-lauroylsarcosine.

Heme Analysis

The heme b content of purified cytochrome bd was measured by the pyridine hemochromogen assay, using an extinction coefficient for the wavelength pair 556.5–540 nm, Δε_556.5–540=24 mM⁻¹cm⁻¹ [18]. The heme d content and enzyme concentration were determined from the absolute spectrum of the fully reduced enzyme, using the extinction coefficient Δε628–670 = 25 mM⁻¹cm₋₁ [8].

Ubiquinol-1 and TMPD Oxidase Activity Assays

The steady state oxidase activity assays were performed as described previously by Zhang et al. [17]. The purified enzyme had a turnover numbers of about 1100 s⁻¹ with 100 μM UQH₂-1 + 3 mM DTT and about 160 s⁻¹ with 10 mM ascorbate + 0.5 mM TMPD.

UV-visible Spectroscopic Measurements

Static absorbtion spectra were obtained with a UV-2101PC spectrophotometer (Shimadzu).

Stopped-flow Experiments. were performed at 20°C using an Applied Photophysics stopped flow with a photodiode array detector. The flow system was flushed thoroughly with Argon-saturated buffer (50 mM sodium phosphate, 5 mM EDTA and 0.05% N-lauroylsarcosine, pH 7.8) before the experiments. The whole system was kept anaerobic by flowing Argon during the experiments. The mixing ratio of the two solutions was 1:1. After mixing, 1600 spectra were collected in the 300–1200 nm range over a period ranging from 2s to 1 min. At least three runs were performed for each time scale. The concentrations reported are the initial concentrations before mixing, if not stated otherwise.

Generation of the oxoferryl form of the enzyme

5 μM enzyme was fully reduced by 200 μM ubiquinone-1 + 6 mM dithiothreitol or by 1 mM TMPD + 20 mM ascorbate in an argon-saturated buffer (50 mM sodium phosphate, 5 mM EDTA and 0.05% N-lauroylsarcosine, pH 7.8). The fully reduced enzyme ( $b_{558}^{2+}{b}_{595}^{2+}{d}^{2+}$) was then rapidly mixed in the stopped-flow apparatus with an equal volume of a solution containing approximately 5 μM O₂ in the same buffer. The reduced cytochrome bd was rapidly oxidized by O₂ yielding essentially the oxoferryl form of the enzyme ( $b_{558}^{3+}{b}_{595}^{3+}{d}^{4+}=O^{2-}$), which was subsequently reduced by the excess reductant present in the solution. The oxidation reaction is too fast to resolve, but the tail end of the reduction kinetics could be observed spectroscopically. Spectroscopic analysis of the product of the reaction of O₂ with the reduced enzyme showed that it contained approximately 80% oxoferryl and 20% ferrous-oxy complex (not shown).

Experiments with the All-ferric Form of Cytochrome bd

5 μM of the “as isolated” cytochrome bd, mainly present as ferrous-oxy complex $b_{558}^{3+}{b}_{595}^{3+}{d}^{2+}-O_2$, was incubated for 20 minutes with 16 μM tetrachlorobenzoquinone, a lipophilic strong oxidant [19], in argon-saturated buffer (50 mM sodium phosphate, 5 mM EDTA and 0.05% N-lauroylsarcosine, pH 7.8). The incubation resulted in conversion of the enzyme to the fully oxidized, all ferric form of the enzyme ( $b_{558}^{3+}{b}_{595}^{3+}{d}^{3+}$). This form of the enzyme was rapidly mixed in the stopped flow with an equal volume of 200 μM ubiquinone-1 + 6 mM DTT or, alternatively, 1 mM TMPD + 20 mM ascorbate, in order to monitor the rate of reduction of the all-ferric form of the enzyme. A strongly lipophilic tetrachlorobenzoquinone is expected to be an antagonist of UQH₂-1 ubiquinol. Indeed, turnover of cytochrome bd with 100 μM UQH₂-1 as electron donor in the presence of 16 μM tetrachlorobenzoquinone was decreased by 30–40% (to about 600–800 electrons s⁻¹). There was no effect of tetrachlorobenzoquinone on the ascorbate+TMPD-oxidase activity of the enzyme. Data Analysis. The data were first analyzed with the use of Pro-Kineticist software package “ProK for PC” (Applied Photophysics) and selected data were imported into Origin 7 (Microcal) for further analysis and preparation of the figures.

3. Results

Figure 2A compares the kinetics of reduction of heme d in the oxoferryl and all-ferric forms of cytochrome bd. From the spectra/time surface collected, the kinetics at a wavelength pair 628 nm and 719 nm (ΔA628–719) was extracted to monitor the reduction of heme d. The kinetics of the reduction of the all-ferric form of the enzyme by UQH₂-1 is extremely slow and non-homogeneous, reaching only about 25% of the expected maximum in 2 s, the maximal observation period in the experiment shown. This is in contrast to the extremely rapid reduction of heme d in the all-ferric enzyme (k~ 3,400 s⁻¹), reported in the pulsed radiolysis experiments of Kobayashi et al [14]. Reduction of hemes b₅₅₈ and b₅₉₅ in the same experiment was even slower that reduction of heme d, with either DTT+ubiquinol or ascorbate+TMPD as electron donors (data not included) and took about about 60 s for completion, whereas heme d reduction was complete in 20 to 30 s. Hence, it is likely that it is electron entry into the enzyme (reduction of b₅₅₈), rather than intramolecular electron transfer, that limits the rate of reduction of ferric heme d in the all-ferric enzyme.

Figure 2. Kinetics of reduction of heme d in the oxoferryl and the all-ferric forms of cytochrome bd. Panel A. Reduction by UQH₂-1.

Heme d in the oxoferryl form is reduced within about 10 ms, compatible with the turnover of the enzyme. No significant reduction of heme d in the all-ferric enzyme is observed in the same time range. Panel B. Reduction by TMPD: The reduction of the all-ferric form is significantly slower than the reduction of heme d in the oxoferryl form of the enzyme. Heme d in the oxoferryl form is reduced by TMPD much more slowly than with UQH₂-1 See Materials and Methods for details.

In order to examine the reactivity of the oxoferryl state of cytochrome bd, this form of the enzyme was generated in situ in the stopped flow mixing chamber by oxidation of the fully reduced cytochrome bd with stoichiometric amount of O₂.

$$b_{558}^{2+}{b}_{595}^{2+}{d}^{2+}+{\text{O}}_2\rightleftarrows b_{558}^{2+}{b}_{595}^{2+}{d}^{2+}-O_2$$ (2)

$$2H^{+}+(b_{558}^{2+}{b}_{595}^{2+}{d}^{2+}-O_2)\to (b_{558}^{3+}{b}_{595}^{3+}{d}^{4+}=O^{2-})+H_{2}O$$ (3)

Formation of the oxycomplex (reaction 2) and subsequent oxidation of the reduced cytochrome bd by bound oxygen (reaction 3) are complete within the mixing time, in agreement with the the known rate constants. At 2.5 μM oxygen (final concentration after mixing), the expected rate for the reaction with O₂ is about 2·10⁹ M⁻¹cm⁻¹ 2.5 μM O₂ = 5 × 103 s⁻¹, τ~200 μs (see [20]), and the oxidation yields largely the oxoferryl state of the enzyme (at least 80%). This was most clearly seen in the spectra from the experiments in which reduced TMPD was used as the reductant (data not shown), since the subsequent reduction of the enzyme in this case is much slower than with UQH₂-1.

The oxoferryl form of the enzyme is reduced quite rapidly by UQH₂-1, and the reaction is completed in about 10 ms (Figure 2A). During this time, up to 12 time-resolved spectra could be collected. The major part of the transiently oxidized cytochrome bd is actually reduced within the dead time of the mixing apparatus and only the tail end of this rapid reduction could be monitored with the current instrumentation. Global analysis of the spectra/time surface for the resolved part of the reduction process gave a rate constant of ~200–300 s⁻¹, which is close to enzyme turnover rate (~1000 electrons s⁻¹) when multiplied by 4 electrons required to convert enzyme from the oxoferryl to the all ferrous state.

The difference spectrum of the rapid phase of the reduction, as resolved by global analysis, is shown in Figure 3. The spectrum shows clearly simultaneous disappearance of the oxoferryl state (trough around 670 nm – 680 nm) and appearance of reduced heme d (peak at ~630 nm), as well as reduced heme b₅₅₈ and heme b₅₉₅ (the peaks at 595 nm, 560 nm and 530 nm in the visible, and an asymmetric Soret band with a maximum at 431 nm and a shoulder around 438 nm). These data imply that the reaction monitored involves reduction of the enzyme by two equivalents of UQH₂-1, since four electrons are required to generate the fully reduced form of the enzyme ( $b_{558}^{2+}{b}_{595}^{2+}{d}^{2+}$) from the oxoferryl state ( $b_{558}^{3+}{b}_{595}^{3+}{d}^{4+}=O^{2-}$). The first equivalent of UQH₂-1 produces a partially reduced form of the enzyme ( $b_{558}^{3+}{b}_{595}^{3+}{d}^{2+}$), so these results imply that the rate of reduction by the second equivalent of UQH₂-1 of the partially reduced enzyme ( $b_{558}^{3+}{b}_{595}^{3+}{d}^{2+}$) is about as fast as the reaction of the oxoferryl species ( $b_{558}^{3+}{b}_{595}^{3+}{d}^{4+}=O^{2-}$) with the first equivalent of UQH₂-1.

Figure 3. Difference spectrum of the rapidly reduced component of cytochrome bd in the reaction of the oxoferryl form reduced by UQH₂-1 found by global analysis of the spectra/time surface.

The spectrum is that of major product of the reaction formed in the first 50 ms minus the initial spectrum, which is primarily the oxoferryl form of the enzyme. The conditions correspond to the trace shown in Figure 2A.

The reduction kinetics of the enzyme with a single electron donor TMPD reduced by excess ascorbate was also examined ( Figure 2B). Reduction of the all-ferric enzyme in this case is also very slow. As in the case of ubiquinol, the reaction of reduced TMPD with the oxoferryl form is more rapid than with the all-ferric form of the enzyme, although it is still much slower than observed with UQH₂-1. The reduction of the all-ferric cytochrome bd by reduced TMPD occurs on a time scale of 0.1–0.5 s, not in milliseconds. It cannot be excluded that the reaction of cytochrome bd with reduced TMPD is rate-limited by the oxidation of TMPD, i.e., TMPD is a much poorer electron donor than UQH₂-1.

4. Discussion

The data clearly show that the reduction of heme d in the oxidized (all-ferric) form of cytochrome bd by its natural 2-electron substrate, ubiquinol, is much too slow to be part of the catalytic cycle. In contast, the reaction of the oxoferryl form of cytochrome bd with ubiquinol is rapid and results in formation of the fully reduced enzyme. Presumably, the reaction proceeds in 2 sequential two-electron steps which are not time-resolved under the conditions of our experiments at a UQH₂-1 concentration of 100 μM (the final concentration after mixing).

$$UQ{H}_2-1+(b_{558}^{3+}{b}_{595}^{3+}{d}^{4+}=O^{2-})\rightleftarrows UQ-1+H_{2}O+(b_{558}^{3+}{b}_{595}^{3+}{d}^{2+})$$ (4)

$$UQ{H}_2-1+(b_{558}^{3+}{b}_{595}^{3+}{d}^{2+})\rightleftarrows UQ-1+(b_{558}^{2+}{b}_{595}^{2+}{d}^{2+})$$ (5)

The rate of reaction 5 must be at least as fast as reaction 4 under the conditions of these experiments, since the appearance of reduced heme d and the reduced hemes b occur in the same time window. However, in the presence of excess of O₂, the very rapid rate of binding of O₂ with the 1-electron reduced enzyme (microseconds) is expected to result in formation of the 1-electron ferrous-oxy complex.

$$(b_{558}^{3+}{b}_{595}^{3+}{d}^{2+})+O_2\rightleftarrows (b_{558}^{3+}{b}_{595}^{3+}{d}^{2+}-O_2)$$ (6)

Hence, when the concentration of O₂ is high, reaction 6 would prevail, followed by reaction 7.

$$(b_{558}^{3+}{b}_{595}^{3+}{d}^{2+}-O_2)+UQ{H}_2-1\rightleftarrows (b_{558}^{2+}{b}_{595}^{2+}{d}^{2+}-O_2)$$ (7)

The product of reaction 7 has sufficient electrons to split the O-O bond (reaction 3), provided protons are available.

Under microaerophilic growth conditions, in which cytochrome bd is normally expressed in E. coli[21], the reactions of the one-electron reduced enzyme ( $b_{558}^{3+}{b}_{595}^{3+}{d}^{2+}$) with O₂ (reaction 6) and with UQH₂-1 (reaction 5) might occur in the reverse order, which, however, makes no difference to the final outcome of the reaction cycle. It is interesting that the heme d ferrous-oxy complex ( $b_{558}^{3+}{b}_{595}^{3+}{d}^{2+}-O_2$) is very stable, being the major form of the enzyme as it is isolated [22]. The dissociation of oxygen in the form of superoxide ( $O_2^{-}$) which would be deleterious to the cell, is negligible. On the other hand, 2-electron reduction of the all-ferric cytochrome bd by UQH₂ and its subsequent oxidation by oxygen could give rise to an unstable peroxide adduct of the ferric enzyme that could dissociate releasing hydrogen peroxide, or produce hydroxyl radicals in case of homolytic scission of the O-O bond. Hence, the intermediate states of cytochrome bd with an odd number of electrons relative to the all-ferric form appear to be preferred in the catalytic cycle of the enzyme. These catalytic intermediate forms contain +3, +1 and −1 electrons relative to the all-ferric enzyme: a) +3 electrons = all-ferrous ( $b_{558}^{+2}{b}_{595}^{+2}{d}^{+2}$) ; b) +1 electron = oxycomplex ( $b_{558}^{+3}{b}_{595}^{+3}{d}^{+2}-O_2$); c) − 1 = oxoferryl ( $b_{558}^{+3}{b}_{595}^{+3}{d}^{+4}=O^{2-}$).

The expectation from the previous work of Jünemann et al [12] and of Matsumoto et al [13], based on steady state kinetics, that the all-ferric form of cytochrome bd is not part of the catalytic cycle of the enzyme is directly verified by pre-steady state kinetics. It remains to be seen why the one-electron reductant formed during the pulsed radiolysis experiments by Kobayashi et al [14] reacts with the all-ferric enzyme rapidly. The explanation might be simply that the N-methylnicotinamide radical and solvated electron generated under the conditions of pulse radiolysis are much stronger reductants than the one-electron donor used in the current work, TMPD ( $E_m^{\prime o}=260mV$) well as than the ubiquinol/ubisemiquinone couple that serves as an initial reductant for cytochrome bd during the reaction of the enzyme with ubiquinol. One might also note, that the all-ferric forms of cytochrome bd were obtained in this work and in ref. [14] by different methods, but it is highly unlikely that a variance in the preparations can account for the ~1000-fold difference in reactivity of heme d.

The observation that UQH₂-1 is a poor reductant for the all-ferric form of the enzyme, but a good reductant for the oxoferryl (b₅₅₈³⁺b₅₉₅³⁺ d⁴⁺ - O²⁻) and one-electron reduced enzyme ( $b_{558}^{3+}{b}_{595}^{3+}{d}^{2+}$) suggests that the redox state of heme d plays a role in determining the rate of reduction of heme b₅₅₈ (and, perhaps, of heme b₅₉₅ as well). Further work is needed to clarify the mechanism of this effect. Recent work has shown that the redox state of heme b₅₅₈ and heme b₅₉₅ regulate the accessibility of heme d by O₂ [23], thus, reciprocal regulation of the reactivity of the two hemes b by the redox and/or ligand-binding state of heme d may be a reasonable speculation. Evidently, the interactions between the three hemes play a critical role in determining the sequence of reactions in the catalytic cycle.

abbreviations

TMPD

N,N,N′,N′-tetramethyl-p-phenylendiamine

DTT

dithiothreitol

EDTA

ethylenediaminetetraacetic acid

UQH₂-1

ubiquinol-1

NMA

N-methylnicotinamide