Molecular Mechanisms of Superoxide Production by Complex III: A Bacterial versus Human Mitochondrial Comparative Case Study

Abstract

In this minireview, we briefly survey the molecular processes that lead to reactive oxygen species (ROS) production by the respiratory complex III (CIII or cytochrome bc₁). In particular, we discuss the “forward” and “reverse” electron transfer pathways that lead to superoxide generation at the quinol oxidation (Q_o) site of CIII, and the components that affect these reactions. We then describe and compare the properties of a bacterial (Rhodobacter capsulatus) mutant enzyme producing ROS with its mitochondrial (human cybrids) counterpart associated with a disease. The mutation under study is located at a highly conserved tyrosine residue of cytochrome b (Y302 in R. capsulatus and Y278 in human mitochondria) that is at the heart of the quinol oxidation (Q_o) site of CIII. Similarities of the major findings of bacterial and human mitochondrial cases, including decreased catalytic activity of CIII, enhanced ROS production and ensuing cellular responses and damages, are remarkable. This case illustrates the usefulness of undertaking parallel and complementary studies using biologically different yet evolutionarily related systems, such as α-proteobacteria and human mitochondria. It progresses our understanding of CIII mechanism of function and ROS production, and underlines the possible importance of supra molecular organization of bacterial and mitochondrial respiratory chains (i. e., respirasomes) and their potential disease-associated protective roles.

1. Introduction

1.1 Reactive Oxygen Species

Reactive oxygen species (ROS) are a group of radical or non-radical oxygen containing molecules that display high reactivity with lipids, proteins and nucleic acids. Depending on the concentration, location, and molecular context, ROS can be beneficial or harmful to cells. Increasing evidence indicates that homeostatic and physiological levels of ROS are indispensable in regulating diverse cellular processes, including ion channel/transporter function [1], Ca²⁺ spark production [2, 3], protein kinase/phosphatase activation and gene expression [4]. A current view is that low levels of ROS production contribute to many essential intracellular signaling processes ranging from cell metabolism to ischemia preconditioning in eukaryotic cells [4–6]. Conversely, excessive ROS generation often leads to apoptotic and necrotic cell death [7, 8] as well as to a panel of clinically distinct disorders, including neurodegeneration (e.g., Alzheimer’s disease), cardiomyopathies, diabetes and cancer [9–11]. Accumulative and systemic ROS damages also underlie cell senescence and aging [6, 12].

In quiescent cells, ROS are primarily produced as byproducts of mitochondrial respiration when electrons leak from the electron transport chain (ETC) [13] (Fig. 1). Superoxide anions (O₂^.−) are the primary ROS species generated by the ETC, and are converted to hydrogen peroxide (H₂O₂) either spontaneously or by the superoxide dismutase (SOD). In the presence of transition metals, O₂^.− can also be transformed into hydroxyl radicals (^.OH) that are considered to be more reactive and damaging. Recent studies documented that massive increases in localized ROS production could occur during metabolic stress [6, 14] and photostimulation [15]. Besides, excessive amounts of intracellular ROS ultimately contribute to necrotic or apoptotic cell death [16, 17].

Figure 1. Structures of CI, CIII and CIV complexes of respiratory ETC and ROS production.

Overall structures of individual respiratory complexes CI (PDB 3IAM + 3M9C in red and blue for membrane integral and membrane external parts, respectively), CIII₂ (PDB 1PP9 in shades of green for the two monomers) and CIV (PDB 1OCC in yellow) are shown together with the Q pool and the electron carrier cytochrome c in purple. In CIII₂, the iron sulfur clusters are positioned in the stigmatellin-bound form and depicted as yellow and orange balls. All heme cofactors are shown in white, red, and blue balls, except in the case of the electron carrier cytochrome c where heme is shown in red. The Q pool (made of oxidized Q and reduced QH₂) is indicated in black. The electron transfer and the proton uptake/release pathways are shown in red and blue arrows, respectively. The CI and CIII complexes of the mitochondrial respiratory ETC are generally considered to be the main sources of O₂^.− radicals, indicated with thick dark blue arrows. In CI, O₂^.− is thought to be produced primarily at the FMN (flavine mononucleotide) cofactor facing the matrix side, whereas in CIII it is generated at the ubiquinol oxidation site (Q_o site) facing the inter membrane space of mitochondria.

1.2 Oxidized and Reduced Quinones

Quinone molecules (oxidized Q/reduced QH₂) play a central role in O₂^.− generation because they act as direct electron donors/acceptors, or redox mediators, to reduce molecular oxygen (O₂). They are present in energy transducing membranes in large amounts as compared with other components of the respiratory chain, and serve as intermediates during electron transfer reactions connecting together ETC complexes (Fig. 1). The redox chemistry of Q/QH₂ is reversible, fast, and involves two consecutive “one-electron” reduction steps via a semiquinone radical (SQ^.−) intermediate. While this property of Q/QH₂ is important during energy transduction to store reducing equivalents, it also constitutes a liability for ROS generation, especially when Q/QH₂ catalysis is not well confined to specific niches in related proteins. This process is often referred to as the ‘leakage of electrons’ from the ETC complexes [18]. In addition to their redox functions, QH₂ molecules can also reduce O₂ via a “one-electron” reduction step leading to the formation of O₂^.−, which is the main source of ROS in cells. However, direct oxidation of QH₂ by O₂ is slow and spin-forbidden, whereas the redox reactions between SQ^.− and O₂ ultimately yielding O₂^.− can be very fast [19]. These reactions strongly depend on how well SQ^.− binds to its site of generation as well as the redox potential of the Q/SQ^.− couple, thus on the stability constant and local factors modulating its disproportionation [20]. As O₂ partitions favorably into the lipid phase, co-partitioning would allow O₂ and SQ^.− to react effectively unless the latter species is sequestered away from accessing O₂. The local concentration and stability of SQ^.− have been optimized during evolution in the case of the ETC complexes that perform Q/QH₂ redox chemistry. It is also noteworthy that the reaction leading to O₂^.− production is reversible, which allows Q to be used as a protective agent against oxidative damages by consuming O₂^.− radicals [21, 22].

1.3 Production of ROS by ETC

Although still subject to controversy, it is believed that ROS are produced in the cell mainly during perturbations of respiratory chain functions [6]. Increased supply of electrons to ETC, (e. g., excess of reducing equivalents) or enhanced membrane potential generation lead to an increase of SQ^.− content in membranes, and subsequently, to higher O₂^.− production. Moreover, kinetic constraints exerted downstream of the Q pool (e. g., blocking electron flow at the level of cytochrome c oxidase, CIV) further enhance production of SQ^.−. Conversely, moderate uncoupling of the membrane potential decreases the electron flux constraints across ETC therefore descreasing O₂^.− production by the respiratory complexes (Fig. 1) [23]. Both of these effects can also be observed upon modification of the respiratory chain by specific chemicals [24, 25]. For example, free fatty acids exert different effects on mitochondria in respect to ROS generation: they inhibit electron flux through the complex I (CI), and probably the complex III (CIII), by inducing an increase of ROS production as seen in isolated rat heart or liver mitochondria [26]. Fatty acids also induce an uncoupling effect, which seems responsible for a large decrease in ROS formation. Thus, ROS production is fine-tuned in response to changes in electron flux through ETC.

The respiratory complexes CI and CIII were well known for some time to be involved directly in superoxide production (Fig. 1). More recently, complex II (CII) is implicated in this process as well [27, 28]. It has been suggested that about 2–5% of O₂ consumed might lead to O₂^.− generation, of which roughly 70–80% is linked to the mechanism of function of CIII [29] based on quantitative data obtained using isolated mitochondria. However O₂^.− production measurements depend on the cell type, respiration steady-state, materials and techniques used to quantify the chemical free radicals. More conservative predictions estimate that O₂^.− production by ETC is perhaps only 0.1% of physiological respiratory rates [30, 31].

2. Complex III (CIII or cytochrome bc₁)

2.1 Structure and mechanism of function of CIII

CIII is a major multi-subunit, membrane-bound enzyme that is central to respiratory energy transduction pathways in many organisms [32]. In different species, it enzymatically converts various derivatives of QH₂ to Q, and reduces various mobile or membrane-anchored electron carriers, typically c-type cytochromes. CIII operates via the Q-cycle mechanism, and contributes to the formation of both the membrane potential and the proton gradient used for ATP production by ATP synthase (or complex V, CV) [33–35]. Depending on the species, mitochondrial CIII contains up to eleven subunits, of which eight are not essential for enzymatic activity (referred to as the “supernumerary subunits”). These subunits are absent in most bacterial CIII, which usually contain only three subunits as the essential catalytic core of the enzyme (Fig. 2a). Due to their structural simplicity and evolutionarily conserved sequence and structure, the facultative phototrophic bacteria (Rhodobacter capsulatus and Rhodobacter sphaeroides) and Paracoccus denitrificans are widely used organisms as CIII models for their mitochondrial counterparts [36–38]. The three universally conserved catalytic subunits of CIII are the cytochrome b, the Fe/S (also called the Rieske) protein and cytochrome c₁. Cytochrome b is an integral membrane protein, whereas the Fe/S protein and cytochrome c₁ are membrane-anchored by their amino- and carboxyl-terminal helices, respectively. These three subunits carry specific cofactors that are required for the catalytic activity of the enzyme. These cofactors are two b-type hemes (axially coordinated protophorphyrin IX-iron) with one low (b_L) and one high (b_H) E_m of cytochrome b, the [2Fe-2S] cluster with a high redox midpoint potential (E_m) of the Fe/S protein, and a high E_m c-type heme (covalently bound protophorphyrin IX-iron) of cytochrome c₁ (Fig. 2a). Most bacterial and mitochondrial purified CIII form dimers, and their three-dimensional structures depict these proteins as symmetrical homodimers [36–42]. In the case of R. capsulatus CIII, monomeric forms of the enzyme are neither active nor stable. However, tetrameric forms of bacterial CIII have been reported in at least two instances. In R. capsulatus, upon fusion of cytochrome c₁ with its physiological electron carrier cytochrome c_y, formation of active CIII tetramers was observed [43]. The P. denitrificans native enzyme forms tetramers, but elimination of a naturally present amino-terminal extension of cytochrome c₁ was reported to yield dimeric CIII [36].

Figure 2. Structure of CIII (cytochrome bc₁) and location of the conserved Tyr 302/278 of cytochrome b.

a, The three dimensional structure of bacterial dimeric CIII bound with stigmatellin that mimics QH₂ (PDB 1ZRT) is shown. The catalytic subunits cytochrome b, cytochrome c₁ and the Fe/S protein are shown in blue, green and purple, respectively. Hemes are shown as yellow sticks, the [2Fe-2S] cluster as yellow and orange spheres, and the Q_o site inhibitor stigmatellin in red sticks. b, Close-up views of bacterial (R. capsulatus, PDB 1ZRT) and bovine (PDB 2A06) and yeast (PDB3CX5) mitochondrial CIII Q_o site structures. The four conserved cysteine residues of the Fe/S protein (yellow and white sticks, R. capsulatus numbering Cys 133 and Cys 153 acting as ligands of the [2Fe-2S] cluster and Cys 138 and Cys 155 forming a disulfide bond) are shown together with the conserved Tyr 302/278/279 of cytochrome b substituted with Cys (gray stick, R. capsulatus Y302->C, bovine Y278C and yeast 279). In the yeast case, the C342 is also indicated c, Sequence alignment of R. capsulatus cytochrome b residues around the conserved tyrosine 302 (278 in human and bovine, 268 in Plasmodium and 279 in yeast) and the conserved PEWY motif shown in red, and green, respectively. Human, Bovine, Yeast, Plasmodium and R. capsulatus refer to Homo sapiens (NCBI protein database accession no: P00156), Bos taurus (P00157), Saccharomyces cerevisiae (P00163), Plasmodium yoelii (AAC25924), and Rhodobacter capsulatus (P08502) cytochrome bc₁, respectively.

Each monomer of the bacterial enzyme contains the three catalytic subunits in an unusual organization. Cytochrome b with its eight transmembrane helices forms the membrane-embedded core to which the other two subunits are bound. Facing the lipid layer on cytochrome b, two Q/QH₂ binding (QH₂ oxidation (Q_o) and Q reduction (Q_i)) sites are located on the positive (p) and negative (n) sides of the membrane. The carboxyl-terminal helix of cytochrome c₁ interacts closely with the fifth helix of cytochrome b to form a cytochrome b-c₁ core, which interacts with the mobile head domain of the Fe/S protein, leaving its amino terminal membrane helix (i. e., tail) associated with cytochrome b of the other monomer. In agreement with this organization, a stable dimeric cytochrome b-c₁ subcomplex has been purified from R. capsulatus. Reconstitution of this subcomplex into an active enzyme was achieved when a full-length Fe/S protein was used, but not with a truncated Fe/S protein lacking its amino-terminal tail domain [44]. The mobility of the Fe/S protein head domain between cytochrome b and cytochrome c₁ and its [2Fe-2S] cluster is essential for CIII activity [45, 46].

According to the proton-motive Q-cycle mechanism, upon the diffusion of a QH₂ molecule from the Q-pool to the Q_o site of CIII, the oxidized [2Fe-2S] cluster of the Fe/S protein oxidizes this QH₂ and conveys a single electron via its mobile head domain to oxidized cytochrome c₁. This electron is then transferred down the ETC to a terminal oxidase (e. g., CIV). The highly unstable SQ^.− radical thus produced at the Q_o site gives an electron to heme b_L of cytochrome b, which rapidly transfers it to heme b_H across the lipid bilayer, to generate membrane potential and a stable SQ^.− at the Q_i site. Completion of the catalytic turnover of CIII involves a second QH₂ oxidation at the Q_o site of the dimeric CIII, via the same sequence of events described above, converting SQ^.− at the Q_i site to a QH₂ to be released from the enzyme. The Q_o and Q_i sites of CIII are not identical with respect to their ability to interact with the SQ^.− species. While the SQ^.− at the Q_i site is well characterized by EPR spectroscopy [47–49], SQ^.− at the Q_o site is a subject of controversy. It is difficult to detect the latter species experimentally, and it was only seen under specific conditions at extremely low amounts [50, 51]. Moreover, no structural information is yet available about the exact position of Q/QH₂ at the Q_o site. Thus, detailed descriptions of the events that follow QH₂ oxidation by the [2Fe-2S] cluster of the Fe/S protein until the transfer of the second electron to heme b_L of cytochrome b remain unknown.

2.2 Superoxide production at the Q_o site

Importance of the structural integrity of bacterial CIII for maximal rate of catalysis and minimal rate of electron leakage to O₂ is known. Heat-inactivated or proteinase K digested CIII [52], and catalytically impaired mutants producing higher amounts of O₂^.− [53, 54] have been reported. Bifurcated electron transfer from QH₂ to the [2Fe-2S] cluster of the Fe/S protein and to the heme b_L of cytochrome b at the Q_o site infers that either preventing the formation of a SQ^.−, or entrapping it within CIII to avoid its interaction with O₂, should prevent O₂^.− production. Conditions favoring the generation of SQ^.− should enhance O₂^.− production as analyzed in detail by Oscyzka et al., [18, 53, 54]. Two different situations leads to SQ^.− generation: a semiforward electron pathway that produces a SQ^.− following oxidation of QH₂ by the Fe/S protein [20, 50, 51, 53–57], and a semireverse electron transfer pathway that involves electron transfer from reduced heme b_L of cytochrome b to a Q bound at the Q_o site to yield a SQ^.−(referred to as “forward” and “reverse” for simplicity) (Fig. 1) [53, 54, 58, 59].

2.2.1 Forward electron transfer for SQ^.− generation at the Q_o site of CIII

Earlier studies focused mainly on the production of O₂^.− at the Q_o site via the forward electron transfer pathway [20, 50, 51, 55–57]. Historically, the pioneering work of Chance proposed that the residual cytochrome c reduction activity seen when CIII is inhibited with antimycin A was closely associated with O₂^.− production [60, 61]. Accordingly, both electrons from QH₂ oxidation would be transferred to the electron carrier cytochrome c, but via two disparate pathways. One electron would be delivered to cytochrome c via the high-potential chain (i. e., the Fe/S protein and cytochrome c₁), while the other electron would be conveyed to O₂ to yield O₂^.− which would rapidly oxidize cytochrome c. Later on, occurrence of this process was supported by the fact that chemical destruction of the [2Fe-2S] cluster of the Fe/S protein [62] or maintenance of this cluster in a reduced state [20] inhibited O₂^.− formation. Similarly, inhibiting reduction of Q at the Q_i site (e. g., using antimycin A) significantly increased Q_o site mediated O₂^.− production. Studies of the effects of specific Q_o and Q_i site inhibitors on O₂^.− production described the complementary bypass reactions in details [20]. For instance, decreasing the rate of electron transfer between the hemes b_L and b_H of cytochrome b, or abolishing the subsequent oxidation of these hemes via the Q_i site inhibitor antimycin A, resulted in the accumulation of electrons on cytochrome b [18]. This led to the accumulation of SQ^.− at the Q_o site and to the leak of electrons to O₂ to generate O₂^.−[20]. In general, if a SQ^.− is formed at the Q_o site (i. e., via a non concerted electron bifurcation) during the normal turn over of a native CIII, O₂^.− production is expected to be quite low to minimize electron leakage and energy waste. However, O₂^.− production at the Q_o site might become significant under compromising conditions, such as a highly reduced Q pool, presence of antimycin A-like molecules inhibiting oxidation of reduced b hemes of cytochrome b, extremely high membrane potential, or specific Q_o site mutations (see section 3). Such conditions may occur in damaged CIII enzymes, or under extreme physiological situations (e. g., ischemia and reperfusion) [63, 64].

Recently, a variant of the forward electron transfer pathway was proposed for R. sphaeroides CIII. In contrast to the earlier studies, this model postulated that under physiological conditions, O₂^.− production is not the result of a bypass reaction during the Q-cycle, but is a regulatory step for enhancing Q_o site catalysis [52, 65]. The authors entertained the idea that O₂ might act as a redox mediator during oxidation of QH₂ and reduction of heme b_L of cytochrome b. Accordingly, O₂^.− formation and CIII activity would increase together as a function of O₂ concentration available during the assay conditions. However, the relevance and validity of the relatively mild effects (< 2 fold) observed on enzyme activity require additional investigations [65].

2.2.2 Reverse electron transfer for SQ^.− generation at the Q_o site of CIII

In recent years, several studies focused on the bypass reactions of the Q-cycle yielding O₂^.− production via a reverse electron transfer pathway. It appears that partial oxidation of the Q pool in a physiologically relevant scenario significantly increases the rates of O₂^.− production by antimycin A inhibited CIII [58]. Using submitochondrial particles, Dröse and Brandt observed that CIII mediated ROS production was higher when CII activity was partially inhibited by malonate (or oxaloacetate), linking the Q pool redox state to O₂^.− production via the Q_o site of CIII. They proposed that the O₂^.− thus generated was produced at the Q_o site by reverse electron transfer from reduced heme b_L of cytochrome b to O₂ via a SQ^.− intermediate acting as a redox mediator. Quinlan et al (2011) further supported this proposal showing that this effect might be directly driven by the redox state of hemes b_L and b_H of cytochrome b that are sensitive to the Q pool redox state and membrane potential [59]. Additional studies using bacterial CIII mutants indicated that O₂^.− production at the Q_o site also involved reverse electron transfer from reduced heme b_L of cytochrome b, and Osyczka’s group proposed a “kinetic” mechanism to account for its occurrence [54]. Accordingly, the movement of the Fe/S protein [2Fe-2S] cluster from the Q_o site increased O₂^.− generation, whereas its stagnation at the Q_o site decreased it. This observation correlated the production of ROS with the position of the Fe/S protein head domain on cytochrome b. Assuming that the movement of the reduced Fe/S protein is not obligatorily “concerted” with electron transfer from SQ^.− to cytochrome b_L heme, ROS generation could be rationalized as the result of a kinetic competition between the internal reactions involving the cofactors of CIII, the Q residing at the Q_o site, and the reaction of SQ^.− with O₂ [54].

3. Defective CIII catalysis and enhanced ROS production due to specific mutations

3.1 Bacterial CIII mutations and ROS production

Studies using bacterial CIII also highlighted some specific amino acid residues as key contributors for affecting ROS production. For instance, the M183K or M183L substitutions in R. capsulatus cytochrome c₁ drastically decreased the E_m of heme c₁,severely impeding electron flow kinetics through the high potential chain of CIII, and enhancing O₂^.− production during Q₀ site catalysis [54]. In a recent study, Lee et al (2011) described a different role played by some amino acid residues of cytochrome b in controlling O₂^.− production via the Q_o site of the bacterial CIII [66] (Fig. 2b). In R. capsulatus, substitution of the conserved Y302 of cytochrome b with any other amino acid residue decreased CIII activity. Concomitantly, it increased O₂^.− production independently of antimycin A inhibition or other treatments known to enhance this process [66]. These findings indicated that some cytochrome b residues are critical for suppressing ROS production at the Q_o site of the enzyme. Various structures have depicted this tyrosine side chain in slightly different H-bonding patterns, depending on the position of the Fe/S protein head domain and the occupant of the Q_o site [67]. Moreover, the hydroxyl group of this residue is within H-bonding distance from a cluster of H₂O molecules in a high resolution structure [68]. Although it is unclear how Y302X (X being any amino acid) mutations enhance mechanistically O₂^.− production, the finding that even the Y302F substitution increases O₂^.− production suggests that it may be linked to the loss of the fixed H₂O cluster coordinated in the native enzyme by the hydroxyl group of Y302 [66]. Accordingly, any mutant losing the hydroxyl group would exhibit decreased CIII activity due to the incorrect positioning of the Fe/S protein head domain. Concomitantly, it would also produce increased O₂^.− due to the uncoordinated mobility of the Fe/S protein head domain vis-a-vis the electron transfer from SQ^.− to cytochrome b_L, and the ensuing undesirable electron leakage to O₂ during Q_o site catalysis.

The counterparts of R. capsulatus Y302 in other species, in particular the malarial (Y268) [69, 70], yeast (Y279) [71–73] and human (Y278C) mitochondrial mutants [74, 75] were also studied with respect to decreased CIII catalysis and enhanced ROS production,. Decreased CIII activities were reported for all mutants, and enhanced O₂^.− production was described for several yeast and human mutants (see below). The overall findings indicate that a number of amino acid residues of cytochrome b at the Q_o site affect both OH catalysis and ROS production. Whether or not all catalytically defective Q_o site mutants always produce enhanced ROS, as a general property of the CIII enzyme, is unknown.

3.2 Bacterial cytochrome b Y302C mutation forms an inter subunit disulfide bond

The bacterial mutant carrying the cytochrome b Y302C mutation was studied in detail [66]. This mutant supported CHI-dependent anoxygenic photo synthetic growth of R. capsulatus. However, it progressively lost its CIII activity upon exposure to air due to slow oxidative disintegration of the [2Fe-2S] cluster in its Fe/S protein both in chromatophore membranes as well as in purified samples [66]. On the other hand, although the homologous yeast cytochrome b Y279C mutant also produced ROS [73], its Fe/S protein [2Fe-2S] cluster did not exhibit oxidative damage [71, 73]. In the case of R. capsulatus, oxidative disintegration of the [2Fe-2S] cluster required not only the presence of O₂, but also the catalytic activity of the Q_o site and the presence of a free thiol group at position 302. Strict anaerobiosis, highly reducing conditions, as well as use of Q_o site inhibitor stigmatellin or thiol-alkylating reagents (e. g., iodoacetamide or N-ethyl-maleimide) abolished the oxidative damage in the Y302C mutant [66]. Using the bacterial mutant CIII, mass spectrometry analyses revealed for the first time that the mutant cytochrome b and the Fe/S protein subunits of CIII were covalently cross-linked to each other by an inter subunit disulfide bond formed between the thiol groups of cytochrome b Y302C and the Fe/S protein C155. It was therefore proposed that the ROS-induced cysteine redox chemistry reduced the intra molecular disulfide bridge, which is naturally present in the Fe/S protein and stabilizes its [2Fe-2S], to render this cluster oxygen labile and the mutant CIII air-sensitive (Fig. 2b) [66].

The striking difference seen in respect to the stability of their Fe/S protein [2Fe-2S] clusters between the bacterial Y302C and its yeast counterpart Y279C is intriguing. Comparison of R. capsulatus and S. cerevisiae cytochrome b amino acid sequences show that while the yeast protein has several cysteine residues, the bacterial counterpart has none. In the former species, one of these cysteine residues (C342, yeast numbering) is structurally located nearby the Y279 (Fig. 2b). Whether the presence of additional cysteine residues counteracts the effect of Y279C mutation (for example by promoting an intra molecular disulfide bond within cytochrome b) is unknown. Mass spectrometry analyses of purified native, Y279F and Y279C yeast CIII enzymes were conducted in our group. In the case of Y279C mutant, the data indicated that the trypsin-gluC fragment encompassing Y279 (W²⁷³YLLPFX²⁷⁹AILR²⁸³, where X²⁷⁹ is Y, F or C in native, Y279F and Y279C mutants, respectively) is only detectable after dithiothreitol (DTT) reduction and iodoacetamide alkylation (unpublished data). This finding suggests that in the yeast mutant the cysteine residue at position 279 might also be modified by a DTT-cleavable chemical group of unknown identity. Inspection of the bovine (Fig. 2b) (and also human) cytochrome b sequence indicates that, although it also contains several cysteine residues, none of them is structurally located in the vicinity of Y278 (homologue of R. capsulatus Y302). Whether the oxidative disintegration of the Fe/S protein [2Fe-2S] observed with the bacterial CIII also occurs in mammalian CIII remains to be seen.

3.3 Human mitochondrial cytochrome b Y278C mutation and ROS production

Very recently, a human mitochondrial CIII produced by a homoplasmic cybrid line generated using fibroblasts of a patient bearing the m.l5579A>G (p.Y278C, i.e., the human homologue of R. capsulatus cytochrome b Y302C mutation) (Fig. 2c) heteroplasmic mutation became available [74, 75]. This mutation was identified in a patient with severe exercise intolerance and multisystem disorders [76], and provided a unique opportunity to extend the significance of the findings emanating from the bacterial case. Comparison of appropriate homoplasmic cybrids carrying either the wild type or the cytochrome b Y278C mutation showed increased intolerance to galactose (hallmark of defective oxidative phosphorylation), drastic loss of CIII activity (over ~ 90%) and highly decreased CHI-dependent oxidative phosphorylation in the mutant [75]. However, the enzymatic activities of the individual CI and CIII complexes, as well as the coupled activities of the CI+III and CII+III supercomplexes were little affected in the mutant. Moreover, no complete loss of the ETC driven membrane potential, or ATP synthesis was observed, suggesting that CI and CII were able to sustain some ATP production despite the low CIII activity. Excitingly, as observed with the bacterial CIII, enhanced O₂^.− production was seen in mitochondria isolated from mutant cybrids, compared to wild type cells. Moreover, an imbalance in homeostasis of the major intracellular antioxidant homeostasis, i. e., an increase of the ratio of oxidized (GSSG) versus reduced (GSH) glutathione ratio was observed, in agreement with increased oxidative stress in mutant cybrids [75]. Indeed, the CI, CIII and CI+III activities increased significantly when mitochondrial preparations were carried out in the presence of DTT. Due to material limitations, reliable detection of the Fe/S protein [2Fe-2S] cluster by EPR spectroscopy has not yet been achieved even with mitochondria from wild type human cybrids, leaving open the question of oxidative disintegration of the Fe/S protein [2Fe-2S] cluster via ROS production.

Similar to the bacterial case, no subunit assembly defect of CIII was seen with the human cytochrome b Y278C mutation as compared to wild type cybrids, based on SDS-PAGE/immunoblots. Wild type and mutant mitochondria contained similar amounts of CIII as revealed by BN-PAGE of dodecylmaltoside dispersed mitoplasts. However, BN-PAGE analyses of digitonin dispersed mitochondrial respirasomes showed decreased amounts of CIII dimers and CIII+IV supercomplexes, but slightly compensatory increased levels of CI₁III₂IV_n (n = number of monomers) supercomplexes [75]. The overall data suggest that supra molecular interactions between the respiratory complexes are important for maintaining basal respiratory ETC function in the Y278C mutant. An emerging hypothesis from this ongoing work is that CIII activity might be better protected against oxidative damages when the mutant CIII is part of the CI₁III₂IV_n supercomplexes (Fig. 3) [77]. Thus, comparative studies conducted for the first time with bacterial and human mitochondrial CIII bearing the same homologous mutation (Y302C and Y278C, respectively) would suggest a new protective role for supra molecular organization of respiratory complexes in membranes.

Figure 3. Organization of ETC complexes into supercomplexes.

A hypothetical higher level of organization of the respiratory complexes CI, CIII and CIV, modified from the recent singleparticle electron cryo-microscopy mediated structure (PDB 2YBB) of the supercomplex CI₁CIII₂CIV₁ from bovine CIII and CIV and bacterial CI [77], is shown. The higher-level structural organization into larger macromolecular entities is assumed to provide structural integrity to the individual ETC complexes by maximizing inter-complex electron flux, and minimizing possible oxidative damages associated with electron leakages. A plausible protective effect exerted at the Q_o site of CIII via its close association with CI is depicted for illustrative purposes. All of the other features shown are as described in the legend of Figure 1.

4. Perspectives

Based on structural, biochemical and clinical studies, an integrated model for ROS production in ETC, involving CI, CII and CIII complexes with their higher order of organizations and physiological regulations is emerging [54, 78]. In the case of the Q_o site of CIII, factors such as the membrane potential, availability of oxidized and reduced equivalents (Q and cytochrome c), redox state of the enzyme, and presence of critical amino acid residues located at specific locations, together control the rate and the amount of ROS production. As described here, recently initiated comparative studies of bacterial and human mitochondrial CIII suggest a new role for supra molecular organizations of the ETC complexes in membranes. Formation of respirasomes appears to improve not only the substrate/product channeling between the related enzymes [79], but might also endow them with higher degrees of stability and protection against oxidative damages by restricting ROS generation (Fig. 3) [80].

Current challenges lay in the development of new tools to improve specific detection and differentiation of various types of ROS free radicals (e. g., O₂^.−, ^.OH, ^.H) at very low concentrations, especially when using integrated systems like respirasomes, mitochondria or whole cells, to better define their role(s) in both signal transduction and oxidative stress. Undoubtedly, future studies will better address whether the organization into larger macromolecular entities protects the structural integrity of native and mutant ETC complexes, and minimizes oxidative damages associated with electron leakages. Emerging investigations of novel proteins promoting formation of supercomplexes [81–83], and determination of ROS-mediated damages on surrounding phospholipids promise to be very informative. The lipochaperone cardiolipin [84] (and also ornithine lipid in some species [85]) is already implicated in affecting the catalytic activity, formation, stability and reconstitution of supra molecular organization of the ETC complexes [86]. Lipidomic analyses using mass spectrometry will allow accurate evaluation of subtle changes in lipid profiles of intact mitochondria, isolated complexes and supercomplexes [84]. Hopefully, these studies will merge together to pave the way towards the development of novel and specific therapeutic interventions for patients with mitochondrial CIII related diseases.

Highlights.

• Bacterial CIII cytochrome b Y302C mutation reduces enzyme activity and induces ROS.

• Homologous human mitochondrial CIII cytochrome b Y278C mutation behaves similarly.

• Supercomplex formation between CI, CIII and CIV might protect against oxidative damages.

• Comparison of homologous bacterial and human mitochondrial mutations are informative.

Abbreviations

ROS

reactive oxygen species

ETC

electron transport chain

O₂^.−

superoxide radical

^.OH

hydroxyl radical

SQ^.−

semiquinone radical

H₂O₂

hydrogen peroxide

CI, CII, CIII, CIV and CV

respiratory complex I, complex II, complex III (cytochrome bc ₁), complex IV and complex V, respectively

SOD

superoxide dismutase