Design and Use of Photoactive Ruthenium Complexes to Study Electron Transfer within Cytochrome bc₁ and from Cytochrome bc₁ to Cytochrome c

Abstract

The cytochrome bc₁ complex (ubiquinone:cytochrome c oxidoreductase) is the central integral membrane protein in the mitochondrial respiratory chain as well as the electron-transfer chains of many respiratory and photosynthetic prokaryotes. Based on X-ray crystallographic studies of cytochrome bc₁, a mechanism has been proposed in which the extrinsic domain of the iron-sulfur protein first binds to cytochrome b where it accepts an electron from ubiquinol in the Q_o site, and then rotates by 57^o to a position close to cytochrome c₁ where it transfers an electron to cytochrome c₁. This review describes the development of a ruthenium photooxidation technique to measure key electron transfer steps in cytochrome bc₁, including rapid electron transfer from the iron-sulfur protein to cytochrome c₁. It was discovered that this reaction is rate-limited by the rotational dynamics of the iron-sulfur protein rather than true electron transfer. A conformational linkage between the occupant of the Q_o ubiquinol binding site and the rotational dynamics of the iron-sulfur protein was discovered which could play a role in the bifurcated oxidation of ubiquinol. A ruthenium photoexcitation method is also described for the measurement of electron transfer from cytochrome c₁ to cytochrome c. This article is part of a special issue entitled: Respiratory Complex III.

1. Introduction

The cytochrome bc₁ complex (cyt bc₁) (ubiquinone:cytochrome c oxidoreductase) is the central integral membrane protein in the mitochondrial respiratory chain as well as the electron transfer chains of many respiratory and photosynthetic prokaryotes [1,2]. The overall net reaction catalyzed by cyt bc₁ involves the 2-electron oxidation of ubiquinol (QH₂) to ubiquinone (Q), and the reduction of two molecules of cytochrome c (Cc). The energy of electron transfer is coupled to the uptake of two protons from the inside of the membrane, and the release four protons to the outside of the membrane (equation 1):

$${\mathrm{QH}}_2+2{\mathrm{Cc}}^{3+}+2{\mathrm{H}}_{\text{in}}^{+}\to \mathrm{Q}+2{\mathrm{Cc}}^{2+}+4{\mathrm{H}}_{\text{out}}^{+}$$ (1)

Mitochondrial cyt bc₁ is a homodimer with 11 polypeptide chains, while prokaryotic cyt bc₁ complexes contain as few as 3 polypeptide chains [1,2]. Cyt bc₁ contains three redox proteins: cyt b with two b hemes (b_L and b_H), the Rieske iron-sulfur protein (ISP) containing a 2Fe2S cluster, and cyt c₁ with one c-type heme. In the widely accepted Q-cycle mechanism, QH₂ binds to the Q_o site near the outside of the membrane and transfers its first electron to the Rieske iron-sulfur center 2Fe2S, which is then transferred to cyt c₁ and finally to Cc [1-4]. The second electron is transferred from semiquinone in the Q_o site to cyt b_L, and then to cyt b_H and ubiquinone in the Q_i site to form semiquinone. The cycle is repeated to reduce semiquinone in the Q_i site to QH₂. X-ray crystallographic studies have shown that the conformation for the extrinsic domain of the Rieske iron-sulfur protein depends on the crystal form and the presence of Q_o site inhibitors [5-8]. In crystals of cyt bc₁ from all species grown in the presence of stigmatellin, the ISP is in a conformation with 2Fe2S close to the cyt b_L heme, called the b state [5-11] (Figure 1). In contrast, the ISP is in a conformation with 2Fe2S close to cyt c₁, called the c₁ state, in native chicken or beef P6₅22 crystals in the absence of inhibitors [6,7]. The intensity of the anomalous signal for 2Fe2S close to cyt b_L is small in bovine I4₁22 crystals in the absence of inhibitors, indicating that the ISP is conformationally mobile (5,8). A rotational shuttle mechanism involving the extrinsic domain of the ISP has been proposed that is supported by these structural studies (Figure 1, Scheme 1) [5-8]. QH₂ in the Q_o site transfers an electron to the oxidized 2Fe2S center of the ISP in the b state, followed by rotation of the ISP to the c₁ state and electron transfer from reduced 2Fe2S to cyt c₁. This mobile shuttle mechanism is supported by extensive mutational, cross-linking, and kinetic studies [12-28].

Figure 1.

X-ray crystal structure of chicken cyt bc₁ in b state in presence of stigmatellin and antimycin (PDB: 3H1I ) [6], and of beef P6₅ 22 crystals in c₁ state (PDB:1BE3) [7]. The ISP, cyt c₁ and cyt b subunits are colored blue, green, and gray, respectively. The hemes, 2Fe2S, stigmatellin, and antimycin are colored red, yellow, cyan, and green. The Rieske neck region residues 66-72 are colored orange, and the ef loop residues 252-268 are colored yellow.

Scheme 1.

An important goal is to measure the rate constants for all of the electron-transfer reactions in cyt bc₁, as well as the dynamics of the conformational changes of the extrinsic domain of the ISP. This has been a challenging goal, since many of the electron-transfer reactions are very rapid. This review describes the development and use of a ruthenium photooxidation technique to study electron transfer in cytochrome bc₁ with microsecond time resolution [13]. This technique has been used to study the key electron-transfer steps in the mechanism of cyt bc₁ as well as the dynamics of the ISP extrinsic domain rotation.

2. Design of Photoactive Ruthenium Complexes

Ruthenium polypyridine complexes have a number of remarkable properties that make them excellent photoredox initiators [29]. They have a long-lived metal-to-ligand charge-transfer excited state that is both a strong oxidant and a strong reductant, and are very stable in both the ground state and the excited state. Ruthenium complexes can be used to rapidly photoreduce or photooxidize neighboring redox centers. In the photooxidation mechanism of scheme 2, the excited state Ru(II*) accepts an electron from Fe(II) to form Ru(I)---Fe(III) with rate constant k₃. A sacrificial electron acceptor A can oxidize Ru(I) to Ru(II) with rate constant k₈, preventing the k₄ back reaction to Ru(II)---Fe(II). All three chelating ligands of ruthenium can be altered to tune the redox potentials over a wide range, and optimize the rate and yield of photooxidation or photoreduction. The effects of different ligands on the redox potentials of the ruthenium complexes are shown in Table I. The Ru(bpz)₂(dmb) complex is particularly well-suited for photooxidation applications. We have introduced four different strategies for specifically labeling proteins with photoredox active ruthenium polypyridyl complexes [30-37]. The most useful method involves the formation of a thioether link between a protein cysteine residue and a ruthenium complex containing a bromomethyl group on the heterocyclic ring [34-37]. The location of the cysteine residue on the protein can be genetically engineered to address specific questions [30-37]. We have recently designed binuclear ruthenium complexes which bind non-covalently to CcO and cyt bc₁ and initiate photoinduced electron transfer [13,38, 39]. [Ru(bpy)₂]₂(qpy)⁴⁺ (Ru₂D) (Figure 2) has a net charge of +4 which allows it to bind strongly to the acidic domain on cyt bc₁ and photooxidize cyt c₁ with a yield of 25% according to Scheme 2 [13].

Scheme 2.

Table 1.

Standard reduction potentials (in V) of ruthenium complexes vs. normal hydrogen electrode

Complex	(II)/(III)	(II*)/(III)	(II)/(I)	(II*)/(I)
Ru(bpy)3	1.27	−0.87	−1.31	0.83
Ru(bpy)2(dmb)	1.27	−0.83	−1.36	0.79
Ru(bpz)2(dmb)	1.76	−0.25	−0.79	1.22
Ru(bpd)2(dmb)	1.49	−0.49	−1.00	0.98

Figure 2.

Modeled structure of Ru₂D. The carbons are colored grey, the nitrogens are blue, and the rutheniums are green.

Extensive studies on a wide range of ruthenium-labeled proteins have provided important information on the dependence of electron transfer on driving force, distance and pathway [40-49]. The theory developed by Marcus has revealed that three important factors control the rate of electron transfer, the free energy change ΔG^o’ of the redox reaction, the reorganization energy λ, and the electronic coupling between the redox centers [50]. The reorganization energy λ is a measure of the energy required to rearrange and repolarize the reactants and surrounding solvent before electron transfer can occur. Dutton and coworkers have reported that the rate constants in a broad range of biological systems can be described approximately by a simple exponential dependence on the distance between the redox centers, as originally proposed by Marcus [51]:

$${\mathrm{k}}_{\mathrm{et}}={\mathrm{k}}_0\exp [-\beta (r-r_{\mathrm{o}}\left)\right]\exp \left[\right(-{(\Delta {\mathrm{G}}^{\mathrm{o},}+\lambda )}^2∕4\lambda \mathrm{RT}\left)\right]$$ (2)

where r is the distance between the closest macrocycle atoms in the two redox centers, the van der Waals contact distance r_o = 3.6 Å, β = 1.4 Å⁻¹, and the nuclear frequency k_o = 10¹³ s⁻¹.

3. Kinetics of Electron Transfer within Cytochrome bc₁

The electron-transfer reactions of mitochondrial cyt bc₁ has been studied extensively by stopped-flow spectroscopy and rapid-mix/freeze-quench EPR [52,53]. In the absence of inhibitors, the reduction of heme b_H by ubiquinol in bovine cyt bc₁ is multiphasic. The fast phase is inhibited by antimycin, indicating that it occurs through the Q_i site. The reduction of heme b_H by 300 μM duroquinol through the Q_o site in the presence of antimycin occurred with a half-time of 25 ms, while cyt c₁ and the 2Fe2S center were also reduced with the same half-time, indicating that they are in rapid equilibrium [53]. A rapid technique using photo-releasable decylubiquinol was also used to study reduction of heme b_H and cyt c₁ [54]. In another study, it was found that the rate of reduction of yeast cyt bc₁ by menaquinol was significantly decreased by antimycin, suggesting that the redox states of heme b_L and heme b_H control the reduction of the 2Fe2S center [55]. The effects of Q_o and Q_i site inhibitors on the kinetics of yeast cyt bc₁ led Trumpower to propose an alternating, half-of-the-sites mechanism [56].

A photoexcitation method has been developed to study the kinetics of cyt bc₁ electron transfer in chromatophores of photosynthetic bacteria including Rhodobacter sphaeroides and R. capsulatus [57]. Photoexcitation of the photosynthetic reaction center rapidly oxidizes cyt c₂, which diffuses to cyt bc₁ and oxidizes cyt c₁ with a half-time of 150 μsec [58,59]. Cyt c₁ then oxidizes the 2Fe2S center, allowing bifurcated electron transfer from QH₂ in the Q_o site to 2Fe2S and heme b_L and heme b_H. The following rate constants, which are identified in Scheme 1, have been estimated for the cyt bc₁ reactions in R. sphaeroides chromatophores : k₁ > 10⁴ s⁻¹ for electron transfer from cyt c₁ to Cc, k₂ > 10⁵ s⁻¹ for electron transfer from 2Fe2S to cyt c₁, k₃ = 1650 s⁻¹ for electron transfer from QH₂ to 2Fe2S in the Q_o site, k₄ > 10⁹ s⁻¹ for electron transfer from Q^•− to heme b_L in the Q_o site, k₅ > k₃, and k₆ > k₃ [59]. Shinkarev et al. [60] have determined that electron transfer between cyt b_L and cyt b_H has a half-time of 0.1 ms using the transient electric field generated by excitation of the reaction center to initiate reverse electron transfer from cyt b_H to cyt b_L.

A laser flash photolysis method was developed to study electron transfer within the cyt bc₁ complex using the binuclear ruthenium complex Ru₂D [13] (Schemes 2, 3). Rb. sphaeroides cyt bc₁ is typically redox poised with cyt c₁ and 2Fe2S reduced, heme b_L oxidized, and heme b_H and quinol partially reduced (Scheme 3). Laser excitation of Ru₂D to the metal-to-ligand charge transfer state, Ru₂D*, a strong oxidant, leads to oxidation of cyt c₁ within 700 ns, as indicated by the rapid decrease in the reduced cyt c₁ absorbance band at 552 nm (Figure 3). A sacrificial electron acceptor A such as [Co(NH₃)₅Cl]²⁺ is present in the solution to oxidize Ru^II* and/or Ru^I and promote oxidation of cyt c₁ by either of the pathways shown in Scheme 2. The +4 charge on Ru₂D allows it to bind selectively to the negatively charged domain on the surface of cyt c₁ with a dissociation constant of 8 μM [13]. The photooxidation of cyt c₁ is complete within 670 ns, the lifetime of the Ru₂D excited state, indicating that Ru₂D must be very close to cyt c₁ at the time of the flash [13,61]. Ru₂D does not photooxidize other proteins that do not have a negatively charged domain, such as cytochrome c, and does not photooxidize the iron-sulfur center, cyt b_L, or cyt b_H in cyt bc₁ [61].

Scheme 3.

Figure 3.

Electron transfer within wild-type R. sphaeroides cyt bc₁ initiated by photooxidation of cyt c₁ [68]. The sample contained 5 μM cyt bc₁, 20 μM Ru₂D, 5 mM [Co(NH₃)₅Cl]²⁺, in 20 mM sodium borate, pH 9.0 with 0.01% dodecylmaltoside. Treatment of cyt bc₁ with 10 μM Q_oC₁₀BrH₂, 1 mM succinate, and 50 nM SCR completely reduced 2Fe2S and cyt c₁, and reduced cyt b_H by 30%. Cyt c₁ was photooxidized within 1 μs, and then reduced with rate constants of 80,000 s⁻¹ and 2,000 s⁻¹, as indicated in the 552 nm transient. The rate constant for the reduction of cyt b_H measured at 561 – 569 nm was 2,300 s⁻¹. (Bottom two traces) Addition of 30 μM famoxadone decreased the rate of reduction of cyt c to 5,400 s⁻¹ 1 and eliminated reduction of cyt b_H.

The photooxidation of cyt c₁ by Ru₂D* is followed by biphasic reduction of cyt c₁ with rate constants of 80,000 s⁻¹ and 2,000 s⁻¹ (Figure 3) [13]. The fast phase has been assigned to electron transfer from reduced 2Fe2S to photooxidized cyt c₁ with rate constant k₂, while the slow phase of cyt c₁ reduction is correlated with the oxidant-induced reduction of heme b_H, monitored at 561 - 569 nm (Figure 3). The oxidant-induced reduction of cyt b_H is rate-limited by transfer of the first electron from QH₂ to 2Fe2S with rate constant k₃. The subsequent transfer of the second electron from the semiquinone to heme b_L and heme b_H with rate constants k₄ and k₅ is much more rapid than k₃, and not rate-limiting. The kinetics of both electron transfer from 2Fe2S to cyt c₁ and from QH₂ to 2Fe2S can thus be resolved by this technique. Electron transfer in the bovine cyt bc₁ complex and Paracoccus denitrificans cyt bc₁ has also been studied using this technique [13,61]. The rate constants for bovine cyt bc₁ are k₂ = 16,000 s⁻¹ and k₃ = 250 s⁻¹, while for the P. denitrificans complex they are k₂ = 10,700 s⁻¹ and k₃ = 700 s⁻¹.

An important question is whether the rate constant k₂ for electron transfer from 2Fe2S to cyt c₁ is rate-limited by true electron transfer, proton gating, or conformational gating [21]. The most definitive way to discriminate between true electron transfer and conformational gating mechanisms is by changing the driving force of the reaction, since Marcus theory predicts a large dependence of the rate of true electron transfer on driving force. The redox potential of 2Fe2S is decreased significantly as the pH is increased from pH 7.0 to pH 10.0 due to the deprotonation of the 2Fe2S ligand His-161, leading to an increase in the driving force ΔG^o’ from −0.02 V to +0.115 V [62,63]. However, the rate constant k₂ for electron transfer from 2Fe2S to cyt c₁ in R. sphaeroides cyt bc₁ was independent of pH, even though Marcus theory predicts that the rate constant should increase 12-fold as the pH is increased from 7.0 to 10.0 [21]. This Marcus theory calculation using equation (2) is based on the assumption that λ = 1.0 V. Although λ has not been measured for this reaction, a value of 1.0 V is typical for cytochromes with a similar heme solvent exposure to that of cyt c₁ [21,43,49]. Moreover, the rate constant k₂ was not affected by the ISP mutations Y156W, S154A, and Y156F/ S154A which decrease the redox potential of 2Fe2S by 62 mV, 109 mV, and 159 mV, respectively (Table 2) [21]. Marcus theory, equation 2, predicts that the increase in the driving force in these mutants would increase the rate constant by up to 17-fold (Table 2), assuming λ = 1.0 V [21]. These studies indicate that the rate constant k₂ for electron-transfer from the 2Fe2S center to cyt c₁ is controlled by conformational gating rather than the rate of transfer of the electron. For this conformational gating mechanism to be valid, the fluctuations in the conformation of the Rieske iron-sulfur protein must be slow compared to electron transfer in the active c₁ state. A model for the active c₁ state is provided by the bovine P6₅22 crystal structure [6, 7], in which the 2Fe2S ligand His-161 forms a hydrogen bond with the heme c₁ propionate oxygen (Figure 4). There is a pathway for electron transfer from the 2Fe2S center to heme c₁ that has a distance of 7.8 Å from the His-161 nitrogen to the closest heme c₁ macrocycle atom C3D (Figure 4). Equation 1 predicts a rate constant k₂ ranging from 1.5 × 10⁶ to 3 × 10⁷ s⁻¹ for this pathway, assuming λ values between 1.0 and 0.7 eV, which are typical for cytochromes [43,49]. The observed rate constant k₂ for electron transfer from 2Fe2S to cyt c₁ is considerably smaller than the predicted value for the c₁ state, consistent with a conformational gating mechanism.

Table 2.

Kinetic properties of R. sphaeroides cyt bc₁ mutants [21]. Enzymatic activity in μmol cyt c reduced/min/μmol cyt b at 25 °C. ΔE_m is the difference in redox potential between 2Fe2S and cyt c₁ at pH 8.0, 25 °C. k₂ is the experimental rate constant for electron transfer from 2Fe2S to cyt c₁ at pH 8.0, 25 °C. Theoretical rate constant for electron transfer is calculated from equation 2 with r = 9.9 Å, λ = 1.0 eV, and ΔG^o calculated from the ΔE_m value of the mutant cyt bc₁.

Mutant	Enzymatic activity	ΔEm (mV)	k2 (104 s−1) (experimental)	k2 (104 s−1) (theory)
Wild-type	2.5	0	8.0	8.0
Y156W	0.58	−62	15.0	26.0
S154A	0.23	−109	7.8	60.0
S154A/Y156F	0.03	−159	9.0	140.0

Figure 4.

Structure of bovine cyt bc₁ P6₅22 crystals in the c₁ state (PDB:1BE3) [7]. The Rieske and cyt c₁ subunits are colored dark blue and light blue, respectively, the 2Fe2S center is shown as a CPK model colored red/yellow, and heme c₁ is colored by element. His-161, Ser-163,Tyr-165, and Cys-139 are shown as sticks. The hydrogen bond between the Nε2 nitrogen of His-161 and the heme c₁ propionate oxygen is shown with a line. The distance of 7.8 Å from the His-161 nitrogen to the closest heme c₁ macrocycle atom C3D is indicated by the red arrow. Tyr-165 and Ser-163 in the bovine ISP are homologous to Tyr-156 and Ser-154 in the R. sphaeroides ISP.

A central question about the bifurcated reaction at the Q_o site is how QH₂ can deliver 2 electrons sequentially to the high and low potential chains, while avoiding short–circuit and bypass reactions. It has been proposed that the controlled motion of the ISP extrinsic domain may play a role in a gating mechanism for the electron bifurcation reaction at the Q_o site [10,12,28]. X-ray crystallography studies have shown that inhibitors bound at the Q_o site have a significant influence on the position and mobility of the ISP. Binding P_m type inhibitors such as myxothioazol to the Q_o site lead to the release of ISP from the b-state to a disordered “mobile” state not detected in the crystal, while structures complexed with P_f type inhibitors such as stigmatellin show that ISP binds to the b subunit in a “fixed” state [8, 10, 25, 28]. EPR studies have also provided valuable information on the effect of Q_o site inhibitors on the position and orientation of the ISP [17,22,26,27]. Havens et al. [61] have studied the effects of six different Q_o site inhibitors on electron transfer from ISP to cyt c₁ in P. denitrificans cyt bc₁ using the Ru₂D photooxidation technique. Binding any of the P_m inhibitors MOA-stilbene, myxothiazol, or azoxystrobin to cyt bc₁ increased the rate and extent of electron transfer from ISP to cyt c₁, consistent with release of ISP from the b state which causes a linked decrease in the redox potential and increase in the mobility of the ISP (Table 3). Binding the P_f inhibitor stigmatellin completely prevented electron transfer from ISP to cyt c₁, consistent with X-ray crystallography studies showing that stigmatellin locks the ISP in the b-state with a hydrogen bond between a carbonyl group of the inhibitor and His-161, a ligand of the 2Fe2S cluster [8,10,64-66]. In contrast, binding the P_f type inhibitors JG-144 and famoxadone decreased the rate constant by 5 to 10-fold, and increased the amplitude over 2-fold. These inhibitors therefore do not lock the ISP in the b state, but rather decrease the rate of its release from the b state and rotation to the c₁ state. Binding reduced QH₂ leads to a two-fold increase in the amplitude of the fast phase, A_1f. These results indicate that the species occupying the Q_o site has a significant effect on the dynamics of the ISP domain rotation.

Table 3.

Effects of Q_o site inhibitors on electron transfer in P.denitricans cyt bc₁ [61]. k_2f and A_2f are the rate constant and amplitude of the fast phase of electron transfer between 2Fe2S and cyt c₁, respectively. Solutions contained 5 μM cyt bc₁, 20 μM Ru₂D, 1 mM ascorbate, 4 μM TMPD, and 5 mM [Co(NH₃)₅Cl]²⁺ in 20 mM TRIS-HCl pH 8.0. Inhibitor concentrations were 25 μM.

inhibitor/substrate	type	k2f (s−1)	A2f
none	--	6,300	10 %
MOAS	Pm	9,900	33 %
myxothiazol	Pm	8,900	30 %
azoxystrobin	Pm	8,000	28 %
stigmatellin	Pf	0	0
JG-144	Pf	1,300	16 %
famoxadone	Pf	600	26 %
Q	--	5,300	10 %
QH2	--	10,700	18 %

X-ray crystallography studies of bovine cyt bc₁ have shown that binding P_f type inhibitors such as famoxadone displace the cd1 helix and the ef helix away from each other to widen the Q_o pocket and form a binding crater for the capture of the ISP in the b-state (Figures 5, 6) [28]. Photoactivated ruthenium kinetic studies have shown that famoxadone binding does not completely immobilize the ISP in the b state, but rather slows down the rate of its release and rotation to the c₁ state. Famoxadone binding to bovine cyt bc₁ decreased k₂ from 16,000 s⁻¹ to 1,500 s⁻¹, while in R. sphaeroides cyt bc₁ k₂ was decreased from 60,000 s⁻¹ to 5,400 s⁻¹ [67] (Figure 3). A series of mutants at residues in the ef loop were constructed to explore the role of the ef loop in regulating the dynamics of the ISP [68] (Table 4; Figure 5). The mutation Y280A caused a decrease in k₂ from 60,000 s⁻¹ to 7,900 s⁻¹, but famoxadone binding only decreased k₂ to 3,200 s⁻¹. Similarly, the I292A mutation decreased k₂ to 4,400 s⁻¹, but famoxadone binding only decreased it to 3,000 s⁻¹. These mutations might cause a conformational change similar to that of famoxadone, limiting the additional effect of famoxadone binding. The I292A mutation caused a decrease in the rate constant k₃ for electron transfer from QH₂ to 2Fe2S from 2,300 s⁻¹ to 350 s⁻¹, indicating an effect on the conformation of the QH₂ reaction site. Mutation of L286A at the tip of the ef loop had an interesting effect, decreasing k₂ to 33,000 s⁻¹ and k₃ to 740 s⁻¹. However, famoxadone binding does not lead to any further decrease in k₂, suggesting that this mutation might block the famoxadone-induced conformational change in the wild-type protein. Darrouzet et al. have also carried out experiments indicating the importance of L286 [18,23].

Figure 5.

X-ray crystal structure of bovine cyt bc₁ in the presence of famoxadone (PDB:1LOL) [58]. Famoxadone is colored cyan, the cyt c₁ and cyt b_L hemes are red, and the 2Fe2S center is represented by a CPK model. The ISP is blue, and cyt b is grey. Residues 252-268 in the ef loop are colored orange while residues 269-283 in the PEWY sequence and the ef helix are red. Residues 136-152 in the cd1 helix are green and residues 163-171 in the neck-contacting domain are colored yellow. Residues of interest are indicated by sticks, and labeled with R. sphaeroides sequence numbering.

Figure 6.

X-ray crystal structure of ISP bound to the docking crater on cyt b with stigmatellin bound to bovine cyt bc₁ (PDB:1SQX) [28]. View is parallel to membrane, showing ef loop and helix (gray), cd1 helix (green), and ISP (orange). The H-bond between stigmatellin and the His-152 ligand to 2Fe2S is shown with a line. Residues on cyt b near the Q_o site or interacting with the ISP are shown as sticks, and labeled with R. sphaeroides sequence numbering.

Table 4.

Effect of mutations on electron transfer within R sphaeroides. cyt bc₁ [68]. The rate constant k₂ for electron transfer from 2Fe2S to cyt c₁ was measured in a solution containing 5 μM cyt bc₁, 20 μM Ru₂D, 5 mM [Co(NH₃)₅Cl]²⁺, in 20 mM sodium borate, pH 9.0, 0.01% dodecylmaltoside. The cyt bc₁ was treated with 10 μM Q_oC₁₀BrH₂, 1 mM succinate, and 50 nM SCR to completely reduce 2Fe2S and cyt c₁, and reduce cyt b_H by about 30%. Famoxadone (30 μM) was added where indicated. The rate constant k₃ for electron transfer from QH₂ in the Q_o site to 2Fe2S was measured without famoxadone.

Mutant	Activity	k2 (s−1)	k2 (s−1) with famoxadone	k3 (s−1)
WT	2.35	60,000	5,400	2,300
Y280A (b)	1.34	7,900	3,200	2,800
L286A (b)	0.78	33,000	35,000	740
I292A (b)	0.81	4,400	3,000	350
P150C (ISP)	0.23	50,000	2,000	4
G153C (ISP)	0.78	13,000	1,470	450

Extensive research has been carried out on the bifurcated electron transfer reaction at the Q_o site where QH₂ transfers the first electron to the ISP and cyt c₁, and the resulting semiquinone transfers the second electron to cyt b_L and cyt b_H [28,69-79]. Potential mechanisms must be consistent with the reversibility of the reaction, and prevent short-circuit reactions, including the delivery of both electrons from ubiquinol to 2Fe2S and cyt c₁ in the high potential chain. Double gating mechanisms have been proposed in which QH₂ can only react if b-state ISP and cyt b_L are both initially oxidized [69,76-79]. Crofts and colleagues have proposed a coulombic gating mechanism in which the semiquinone anion moves from the distal site near the ISP to a site near oxidized cyt b_L in a process controlled by electrostatics [71,77]. Another possibility is simultaneous transfer of two electrons from QH₂ to the ISP and cyt b_L in a concerted reaction without the formation of a semiquinone intermediate [69,76,78-80]. However, a semiquinone radical at the Q_o site has been detected at the Q_o site by two different methods [81,82], and a semiquinone is also thought to be required in the bypass reaction during formation of superoxide [83].

Unfortunately, it has not been possible to experimentally detect QH₂ or Q in the Q_o binding pocket by X-ray crystallography. However, the effects of Q_o site inhibitors on the structural linkage between the conformations of cyt b and the ISP have led to a proposal for the mechanism for bifurcated electron transfer [12,28]. It was proposed that binding QH₂ to the Q_o site widens the Q_o pocket between the cd1 helix and ef helices forming a crater to bind the ISP in the b-state (Figure 6) [12,28]. This would promote the formation of a hydrogen bond between QH₂ and His-161 and lead to proton-coupled electron transfer from QH₂ to oxidized 2Fe2S. After the second electron was transferred from semiquinone to cyt b_L and cyt b_H, the resulting oxidized Q would leave the distal Q_o binding pocket, triggering the cd1 and ef helices to come closer together and release the ISP from the docking crater, allowing it to rotate to the c₁ position and transfer an electron to cyt c₁ [12,28]. The effects of P_m and P_f inhibitors on the rapid kinetics of bovine, R. sphaeroides, andP. denitiricans cyt bc₁ provide evidence that the conformations of the Q_o site, the ISP docking crater, and the ISP extrinsic domain orientation and dynamics are tightly linked. It is suggested that binding P_f inhibitors to the Q_o site leads to a conformation similar to that of the active QH₂ – oxidized ISP complex, while binding P_m inhibitors leads to a conformation in which the ISP is released from the b-state to a mobile state. The linkage between the Q_o site and the ISP conformation and dynamics may play an important role in gating the electron transfer bifurcation reaction in the Q_o site to minimize short-circuit and bypass reactions. Other factors are also likely to be involved in gating, including coulombic gating of the motion of the semiquinone [72,77], and the conformations of water or amino acid side chains in the Q_o pocket [69,75-79]. There is also evidence that events at the Q_i site and the b_L and b_H hemes might be linked to turnover at the Q_o site [26,27,84-89]. In addition, conformational interactions and electron transfer between the two monomers of the cyt bc₁ dimer might play a role in the mechanism of the reaction at the Q_o site [85,89-92].

4. Reaction between Cytochrome bc₁ and Cytochrome c

Cytochrome c is a hydrophilic heme protein with a molecular weight of 12,500 Da that transfers electrons from the cyt bc₁ complex to cytochrome c oxidase by a diffusional shuttle mechanism. The electron-transfer reaction from the cyt bc₁ complex to Cc involves three steps: a) formation of a 1:1 reactant complex between reduced cyt bc₁ and Cc³⁺, b) intracomplex electron transfer from cyt c₁²⁺ to Cc³⁺, and c) release of the product Cc²⁺. The overall rate of the reaction is optimized when the interaction between Cc and cyt bc₁ stabilizes a reactant complex which allows rapid electron transfer, and also promotes rapid reactant complex formation and product complex dissociation. The steady-state reaction rate decreases with increasing ionic strength, indicating the involvement of electrostatic interactions between the two proteins [93-95]. Extensive chemical modification studies have demonstrated that six lysine amino groups surrounding the heme crevice of Cc are involved in binding to cyt bc₁ [93-96]. Chemical modification and cross-linking studies have shown that acidic residues on cyt c₁ and subunit 8 in bovine cyt bc₁ are involved in binding Cc [97, 98]. X-ray crystal structures of beef, chicken, yeast, and R. sphaeroides cyt bc₁ reveal that the cyt c₁ heme edge on the cytoplasmic surface is surrounded by acidic residues that could form a binding site for Cc [5-9]. Most importantly, the X-ray crystal structure of the complex between yeast Cc and yeast cyt bc₁ revealed that it is stabilized by non-polar interactions at the center of the binding domain, including a planar stacking interaction between yCc Arg-13 and Phe-230 of cyt c₁ (Figure 7) [99, 100]. There are only two direct polar interactions in the binding domain, but additional charged residues around the periphery of the binding domain may contribute to the electrostatic interaction. The distance between the edges of the heme c and heme c₁ groups is 9.4 Å.

Figure 7.

X-ray crystal structure of the complex between yeast cyt bc₁ and yCc [99] (PDB: 1KYO). yCc is colored light blue, Cyt c₁ is gray, the heme groups are red, basic residues on yCc are blue, acidic residues on cyt c₁ are red, and Phe-230 is purple. The ruthenium complex on Cys-39 (green) was attached to the crystal structure by molecular modeling.

Stopped-flow spectroscopy has been used to measure the second-order electron transfer reaction between Cc and bovine cyt c₁ at high ionic strength, but the reaction becomes too fast to resolve by this technique below 200 mM ionic strength [52]. In R. sphaeroides chromatophores, the reaction is rate-limited by the diffusion of photooxidized cyt c₂ from the reaction center to the cyt bc complex with an apparent rate constant of 5000 s⁻¹ 1 [58, 59]. These techniques have provided valuable information about the reaction between cyt c₁ and Cc, but it was not possible to measure the intracomplex rate constant.

It was necessary to design a new ruthenium-labeled Cc derivative to study rapid electron transfer from cyt bc₁ to Cc in the forward, physiological direction. A brominated Ru(bpz)₂(dmb) reagent was used to label the Cys-39 sulfhydryl group on yeast H39C,C102T Cc to form Ru_z-39-Cc (Figure 7) [101]. The ruthenium complex is on the surface opposite from the heme crevice of Cc, and does not affect the interaction with yeast cyt bc₁. There is an efficient pathway for electron transfer between the heme and the ruthenium complex consisting of 13 covalent bonds and one hydrogen bond, with a distance of 12.6 Å. The new Ru(bpz)₂(dmb) complex has a reduction potential of 1.22 V for the Ru(II*)/Ru(I) transition. The driving force of 1.0 volt for the Ru(II*)-Fe(II) → Ru(I)-Fe(III) reaction is close to the expected reorganization energy λ of 0.8 V, which should allow optimal photooxidation of the reduced heme c according to Scheme 2. The photoinitiated electron transfer from heme c Fe(II) to Ru(II*) in Ru_z-39-Cc occurred with a rate constant of k₃ = 1.5 × 10⁶ s⁻¹, followed by back electron transfer from Ru(I) to Fe(III) with a rate constant of k₄ = 7000 s⁻¹ [101]. The back reaction is prevented in the presence of atmospheric oxygen, which rapidly oxidizes Ru(I). The yield of photooxidized heme c is 20% in a single flash.

Laser excitation of reduced yeast Ru_z-39-Cc and yeast cyt bc₁ at 250 mM ionic strength led to rapid photooxidation of heme c, followed by electron transfer from cyt c₁ to oxidized heme c with a rate constant of 3900 s⁻¹, as monitored at 546 nm (Figure 8) [101]. The oxidation of cyt c₁ was observed directly at 557 nm, which is an isobestic point for Cc. A fast phase with a rate constant of 14,000 s⁻¹ was observed at ionic strengths below 150 mM due to electron transfer in a preformed Ru-39-Cc:cyt bc₁ complex (Figure 9). The rate constant of this intracomplex electron transfer reaction was independent of ionic strength from 5 mM to 120 mM, indicating that the complex does not change its configuration. Equation 1 was used to calculate a theoretical rate constant for electron transfer from cyt c₁ to Cc based on the X-ray crystal structure of the complex between yeast iso-1-Cc and yeast cyt bc₁ [99] (Figure 7). The calculated rate constant is between 1.8 × 10⁵ s⁻¹ and 3.3 × 10⁶ s⁻¹, assuming a reorganization energy λ between 0.7 and 1.0 V, and an edge-to-edge separation of 9.4 Å between the heme c and heme c₁ groups as given in the crystal structure. Although the theoretical value is larger than the experimental value of 1.4 × 10⁴ s⁻¹, the through-water jump of 4.5 Å between the two hemes in the crystallographic complex could give a large barrier to electron transfer that is not accounted for by equation (2).

Figure 8.

Photoinduced electron transfer between yeast Ru-39-Cc and yeast cyt bc₁ [101]. The solution contained 5.2 μM yeast Ru_z-39-Cc and 4.4 μM yeast cyt bc₁ in 5 mM sodium phosphate, pH 7.0, 250 mM NaCl, and 0.1% lauryl maltoside. It was treated with 2 μM TMDP and 10 μM ascorbate to reduce the c₁ and c hemes. The 550 nm transient shows the photooxidation and reduction of Ru-39-Cc, while the 557 nm transient shows the oxidation of cyt c₁. The transients at both wavelengths indicate electron transfer from cyt c₁ to Cc with a rate constant of 3900 ± 600 s⁻¹.

Figure 9.

Ionic strength dependence of the rate constant for photoinduced electron transfer between yeast Ru_z-39-Cc and yeast cyt bc₁. The conditions were the same as in Figure 8 except that 0 to 800 mM NaCl was present. The square root of ionic strength is in units of [M]^1/2.

Since both intracomplex and bimolecular phases are observed at 110 mM ionic strength, the bimolecular reaction involves formation of a 1:1 complex with rate constant k_f = 2.0 × 10⁹ M⁻¹s⁻¹, intracomplex electron transfer with k_et = 14,000 s⁻¹, and complex dissociation with k_d = 1.7 × 10³ s⁻¹ [101]. The fast intracomplex phase disappears and the rate constant of the bimolecular phase increases to a maximum at 200 mM ionic strength (corresponding to (I)^1/2 = 0.45 in (M)^1/2 units), indicating an increase in k_d (Figure 9). The second-order rate constant decreases with increasing ionic strength above 250 mM ionic strength, consistent with a reaction between oppositely charged proteins (Figure 9). Rajagukguk et al. [102] prepared a series of yeast Ru_z-39-Cc mutants containing mutations of residues at the binding domain in order to characterize the interaction with cyt bc₁. The rate constants were measured using the ruthenium photooxidation technique at 250 mM, where the reaction is bimolecular (Table 5). The largest effect was observed for the R13A mutant, where the rate constant was decreased from 3500 s⁻¹ to 153 s⁻¹. This indicates that the π-cation interaction between yCc Arg-13 and Phe-230 of cyt c₁ observed in the crystal structure of the complex [99,100] is important for the reaction in solution. A substantial decrease in rate constant to 1090 s⁻¹ and 190 s⁻¹ for the K86A and K86D mutants, respectively, demonstrates that the charge-pair interaction between yCc Lys 86 and Glu 235 of cyt c₁ in the crystal structure is also important for the reaction in solution. Mutation of other yCc lysines to alanine, including 11, 72, 73, 79, and 87, also led to significant decreases in the rate constant (Table 5). Although there is only one electrostatic charge-pair interaction in the binding domain of the yCc:ybc₁ crystallographic complex, the 5 lysine amino groups on yCc and 5 carboxylate groups on cyt bc₁ immediately surrounding the interaction domain could guide Cc to the binding site and contribute to complex formation [99,100].

Table 5.

Reaction of yeast Ru-39-Cc mutants with yeast cyt bc₁ [102]. The reaction solution contained 5 μM yeast Ru-39-Cc mutant and 5 μM yeast cyt bc₁ in 5 mM sodium phosphate, pH 7.0 with 250 mM NaCl and 0.1% lauryl maltoside. The c₁ and c hemes were reduced with 10 μM ascorbate. The sequence numbering of horse Cc is used.

Mutant	k (s−1)
None	3500
K11A	1480
T12A	2540
R13A	153
V28A	2620
K72A	1960
K73A	2190
K79A	1530
A81G	2400
K86A	1090
K86D	190
K87A	1120

Sarewicz et al. [103] carried out EPR studies indicating that the lifetime of the complex between Rhodobacter capsulatus cyt c₂ and cyt bc₁ was longer than 100 μs at low ionic strength, decreasing to less than 400 ns at ionic strengths above 125 mM. Their results are consistent with a mechanism in which cyt c₂ binds rapidly to cyt bc₁ at low ionic strength allowing efficient intracomplex electron transfer, but product complex dissociation is slow, limiting enzyme turnover. At high ionic strength complex formation is slow and complex dissociation is so rapid that most collisions do not result in electron transfer. At intermediate physiological ionic strength, complex formation and dissociation are both moderately rapid, and there is a moderate ratio of electron transfer per collision. The rapid kinetic studies discussed above [101] are also consistent with this mechanism. Sarewicz et al. [104] carried out EPR studies indicating that the dipole moment of cyt c₂ plays an important role in orienting the molecule for efficient electron transfer during the collision process.

The Paracocus denitrificans cyt bc₁ complex contains just three subunits, the b subunit with heme b_L and heme b_H, the Rieske iron-sulfur protein (ISP), and cyt c₁ [105]. The cyt c₁ subunit has a tripartite domain structure consisting of a unique N-terminal acidic domain of 150 amino acids, a periplasmically oriented core domain containing the covalently attached heme c, and a C-terminal membrane anchor. The acidic domain may be analogous to the small acidic subunits of eukaryotic cyt bc₁, including the hinge protein of bovine cyt bc₁ and subunit 6 of yeast cyt bc₁ [106,107]. An usual feature of P. denitrificans cyt bc₁ is that it has a “dimer of dimers” quaternary structure rather than the dimeric structure found in other cyt bc₁ complexes [108]. Cyt bc₁ transfers electrons to membrane-bound cyt c₅₅₂ [109]. Kinetic studies have been carried out on genetically engineered soluble modules of both redox partners [110]. The soluble cytochrome c₁ core fragment (cyt c_1CF) consists of only the central core domain, without the acidic domain and the membrane anchor. The soluble cytochrome c₅₅₂ fragment (cyt c_552F) contains only the C-terminal hydrophilic heme domain without the N-terminal membrane anchor and linker region. A new ruthenium cyt c_552F derivative (Ru_z-23-c_552F) was designed to measure rapid electron transfer with cyt c_1CF using the ruthenium photooxidation technique [110]. The bimolecular rate constant k₁₂ decreased rapidly with increasing ionic strength above 40 mM, indicating that electrostatic interactions were important for the reaction between the two proteins. However, k₁₂ was rapid, 3 × 10⁹ M⁻¹s⁻¹, and nearly independent of ionic strength below 35 mM. These results are consistent with a two-step process involving very rapid formation of an initial complex guided by long-range electrostatic interactions, followed by short-range diffusion along the protein surfaces guided by hydrophobic interactions. No intracomplex electron transfer between Ru_z-23-c_552F and c_1CF was observed even at the lowest ionic strength, indicating a low-affinity complex. In contrast, yeast Ru_z-39-Cc formed a tight 1:1 complex with cyt c_1CF at ionic strengths below 60 mM with an intracomplex electron transfer rate constant of 50,000 s⁻¹. A group of cyt c_1CF mutants in the presumed docking site were generated based on information from the yeast cyt bc₁/cyt c co-crystal structure. Kinetic analysis of cyt c_1CF mutants located near the heme crevice provided preliminary identification of the interaction site for cyt c_552F, and suggest that formation of the encounter complex is guided primarily by the overall electrostatic surface potential rather than by defined ion pairs [110]. Ruthenium kinetic studies have shown that the acidic domain does not play a significant role in the reaction of cyt c_552F with P. denitrificans cyt bc₁[111]. The reaction between a ruthenium horse Cc derivatve and R. sphaeroides cyt bc₁ was found to have an intracomplex rate constant of 60,000 s⁻¹ [112]. Mutagenesis studies indicated that acidic residues near the heme crevice of cyt c₁ are important in guiding Cc to the binding domain [112].

Conclusions and Future Prospects

Extensive structural, spectroscopic, and kinetics studies have provided considerable insight into the mechanism of the cytochrome bc₁ complex. Most of the rate constants for the key electron transfer reactions have been measured for R. sphaeroides cyt bc₁ as indicated in Scheme 1 and Table 6. X-ray crystallographic and EPR studies have shown that inhibitors bound at the Q_o site affect the position and mobility of the ISP [8,10,17, 25-28]. Kinetic studies have revealed that the reduction of cyt c₁ by 2Fe2S is rate-limited by the rotational diffusion of the ISP from the b-state to the c₁ state, and that the rate of this rotation is controlled by the type of inhibitor bound in the Q_o site [13,21,61,67-68]. On the basis of these studies it has been proposed that binding QH₂ to the Q_o site induces a conformational change in the ef and cd₁ loops leading to capture of the ISP in the b state and proton-coupled electron transfer from 2Fe2S to cyt c₁[12,28]. Following electron transfer to cyt b_L and cyt b_H, the ISP is released from the b state, rotates to the c₁ state, and transfers an electron to cyt c₁. The detailed mechanism of how this bifurcated electron transfer reaction avoids short-circuit and bypass reactions remains enigmatic, however. A number of possibilities have been proposed, including control of the rotation of the ISP by the release of Q from the Q_o site or the reduction of cyt b_L and cyt b_H [12,28], double gating mechanisms [69,75-79], or coulombic gating of the motion of the semiquinone [72,77]. It has also been proposed that the reaction of Q at the Q_i site could affect the bifurcated reduction of QH₂ at the Q_o site and the motion of the ISP [26,27,84-89]. Moreover, experiments from several different laboratories indicate that electrons can be transferred between the two cyt b_L hemes in the homodimeric enzyme, suggesting that electrons can be distributed between the four quinone sites in the dimer[85,89-92]. However, it is not clear whether cross-monomer communication regulates reactions at the Q_o and Q_i sites. Future experiments are needed to address these important questions.

Table 6.

Rate constants of reactions in R. sphaeroides cyt bc₁. Rate constants for the reactions shown in Scheme 1 were measured under the conditions reported in the references.

Reaction	Rate constant	References
c1 → c	k1 = 60,000 s−1	[101,112]
2Fe2S → c1	k2 = 80,000 s−1	[13,21,68]
QH2 → 2Fe2S	k3 = 2000 s−1	[13,21,59,68]
Q•- → bL	k4 > 109 s−1	[59]
bL → bH	k5 = 10,000 s−1	[60]
bH → Q	k6 > k3	[59]

Highlights.

• Photoactive ruthenium polypyridine complexes have been developed to study rapid electron transfer reactions in the cytochrome bc₁ complex at a microsecond time scale.

• A ruthenium photooxidation technique has been developed to measure the rate constant for electron transfer from the iron-sulfur protein to cytochrome c₁.

• The dynamics of rotation of the iron-sulfur protein from the b state to the c₁ state has been measured.

• A conformational linkage between the occupant of the Q_o ubiquinol binding site and the rotational dynamics of the iron-sulfur protein was investigated which could play a role in the bifurcated oxidation of ubiquinol.

• A ruthenium photoexcitation method is described for the measurement of electron transfer from cytochrome c₁ to cytochrome c.

Abbreviations

Cc

cytochrome c

yCc

yeast Cc

cyt bc₁

cytochrome bc₁

CcO

cytochrome c oxidase

2Fe2S

Rieske iron-sulfur center

ISP

iron-sulfur protein

bpy

2,2′-bipyridine

dmb

4,4′-dimethyl-2,2′-bipridine

bpz

2,2′-bipyrazine

bpd

3,3′-bipyridazine

qpy

2,2′:4′,4″:2″,2″′-quaterpyridine

Ru₂D

[Ru(bpy)₂]₂qpy⁴⁺

R.sphaeroides

Rhodobacter sphaeroides

MOAS

methoxyacrylate stilbene

JG144

S-3-anilino-5-methyl-5-(4,6-difluorophenyl)-1,3-oxazolidine-2,4-dione

Q

ubiquionone

QH₂

ubiquinol

Q_o

outside ubiquinol binding site

Q_i

inside ubiquinone binding site

P_f

Q_o site inhibitor which fixes ISP in b state

P_m

Q_o site inhibitor which promotes mobile state of ISP