In cellular respiration, cytochrome c transfers electrons from the cytochrome bc1 complex to cytochrome c oxidase by transiently binding to the membrane proteins. The first X-ray structure of the yeast cytochrome bc1 complex with bound cytochrome c revealed the general architecture of the electron-transfer complex. The interface of the complex is small. The haem moieties are centrally located in a mainly non-polar contact site, which includes a cation–π interaction and is surrounded by complementary charged residues. Only one cytochrome c1-docking site of the dimeric complex is occupied with cytochrome c. The recent 1.9 Å (1 Å=0.1 nm) resolution structure of the complex showed that the interface is highly hydrated. With cytochrome c bound, a higher number of interfacial water molecules are present on the cytochrome c1 interface, whereas its protein surface is not affected. Remarkably, the dimer structure is slightly asymmetric. Univalent cytochrome c binding coincides with conformational changes of the Rieske head domain and subunit QCR6p. Pronounced hydration and a mobility mismatch at the interface with disordered charged residues on the cytochrome c side are favourable for transient binding. Comparison with a new structure of the complex with bound isoform-2 cytochrome c led to the definition of a core interface, which refers to four common interaction pairs including the cation–π interaction. They encircle the haem groups and are surrounded by variable interactions. The core interface may be a feature to gain specificity for formation of the reactive complex. The consistency in the binding interaction despite differences in primary sequence, redox state and crystal contacts, together with crystallization at physiological ionic strength, clearly suggest that the structures show the native bound state of the electron-transfer complex.
- cytochrome bc1 complex
- cytochrome c
- electron-transfer complex
- interfacial water molecule
- mobility mismatch
- protein–protein interface
Electron-transfer processes are of great importance in many metabolic pathways of living organisms. They are essential for photosynthesis and cellular respiration, in which small diffusible redox proteins facilitate electron transport between large membrane-embedded enzymes. The intermolecular electron transfer involves transient interactions of the reaction partners [1,2]. These interactions are of low affinity, with equilibrium dissociation constants in the micromolar to milliomolar range . To promote high turnover and efficiency of the energy-converting machinery, binding of the mobile electron-carrier proteins has to be not only transient, but also specific. Transient complexes are difficult to crystallize and their structures are poorly represented in the RCSB PDB database (http://www.rcsb.org). For the respiratory chain, only crystal structures of the QCR (cytochrome bc1 complex) with bound CYC (cytochrome c) from Saccharomyces cerevisiae have been determined to date [4,5]. In the present article, the structural information about this transient electron-transfer complex is reviewed, with a focus on the recent 1.9 Å (1 Å=0.1 nm) resolution structure .
The mitochondrial respiratory chain
Oxidative phosphorylation produces the majority of energy equivalents in eukaryotic cells. In this final step of nutrient catabolism, the mitochondrial respiratory chain couples electron transfer from reduction equivalents to molecular oxygen, with vectorial proton translocation, thereby generating the protonmotive force that drives ATP synthesis. Electron and proton transfer in the respiratory chain employs four multisubunit enzymes (complexes I–IV), which are embedded in the inner mitochondrial membrane. Electron transfer between complex III (QCR) and complex IV [COX (cytochrome c oxidase)] is facilitated by reversible binding of the soluble protein CYC . CYC interacts with QCR and COX in two spatially and temporally distinct events. These interactions are highly transient in nature, enabling turnover numbers of 150 s−1 and 600 s−1 for QCR and COX respectively [7,8], which is essential for the continuous flow of electrons across the different components of the respiratory chain. Three crystal structures of QCR with bound CYC from S. cerevisiae have been determined [4,5] The first structure revealed the general features of the interaction at a resolution of 2.97 Å. Improving the resolution to 1.9 Å recently permitted a detailed analysis of the electron-transfer complex and the interface, including the contributions of water molecules [4,5]. Under certain growth conditions, yeast cells express isoform-2 CYC that shares 83% sequence identity with the major isoform-1 . The structure of QCR with bound isoform-2 CYC has now also been determined at 2.5 Å resolution .
Crystallization and structure of the QCR–CYC electron-transfer complex
Mitochondrial QCR is a homodimeric multisubunit membrane protein complex with a molecular mass close to 500 kDa. Each monomer contains three redox-active subunits: cytochrome b with two b-type haem groups, CYT1 (cytochrome c1) with a c-type haem, and the Rieske protein containing a [2Fe–2S] cluster. The enzyme operates via a mechanism called Q cycle, in which it couples electron transfer from ubiquinol to CYC, with the net translocation of protons across the membrane . Key features of the mechanism are: (i) the bifurcated electron transfer upon ubiquinol oxidation at the Qo site, (ii) a spatially separated second catalytic site for ubiquinone reduction (Qi site) , and (iii) the large-scale domain movement of the Rieske protein, which facilitates electron transfer from the Qo site to subunit CYT1 . CYC docks on to the latter subunit to accept the electron.
For all three QCR–CYC structures, independently purified reaction partners were mixed with CYC in excess. The complexes were crystallized at physiological ionic strength of 110–120 mM and with an antibody fragment bound to QCR [4,5,13]. Binding of the antibody fragment to the Rieske protein (Figure 1) does not interfere with QCR activity . X-ray diffraction data for the first structure were collected at 277 K . Subsequently, growth of crystals in the presence of sucrose controlled the excessive nucleation and permitted transfer of the crystals to a cryo-buffer, so that high-resolution data could be collected at 100 K .
In the available structures, a single CYC molecule is bound to the dimeric QCR (Figure 1). The antibody fragments were advantageous for crystallization as they contribute major contacts in the crystal lattice. In addition, they also provide space for CYC which is bound to its docking site on CYT1 without steric hindrance.
The CYT1–CYC interface of the electron-transfer complex
The general features of the interface are a small non-polar contact area including a cation–π interaction, with the haem cofactors in the centre and complementary charged residues in the periphery. This is consistent with the two-step model of electron-transfer complex formation, which involves the initial pre-orientation of the molecule via electrostatic steering and a final bound form with specific hydrophobic interactions . The QCR–CYC structures represent the bound state of the electron-transfer complex.
The interfaces in the medium- and high-resolution structures with bound isoform-1 CYC cover 956 Å2 and 957 Å2 respectively, and 1139 Å2 in the complex with isoform-2 CYC. This is considerably smaller than an average protein–protein interface of 1600±400 Å2 , but is similar in size compared with other electron-transfer complexes, e.g. 1150 Å2 for cytochrome c peroxidase and CYC from yeast, 1270 Å2 for the reaction centre and cytochrome c2 from Rhodobacter sphaeroides, and 1200 Å2 for plastocyanin and cytochrome f from Phormidium laminosum [1,2]. Small interfaces are favourable for transient protein complexes.
The relative orientation of the haem groups and their distance is highly similar in all three structures, supporting the physiological nature of the complex. Electron transfer in the bound state of the complex is likely to be a direct and thus rapid passage from haem c1 to haem c. The exposed edges of the two haem groups are very close, with the neighbouring carbon atoms 4 Å apart and a 9 Å edge-to-edge distance between the porphyrin rings. Calculated electron tunnelling rates  range from 1.0×106 to 2.6×107 s−1 for the three structures. The differences between the structures are minor, and the range is mainly affected by the reorganization energy, which was either set to λ=0.7 or to λ=1.0 eV.
Tight and specific interactions of the interface are mainly mediated by non-polar forces with contact distances of 3–4 Å. In the higher-resolution structure, hydrophobic interactions and van der Waals contacts are contributed by seven CYT1/CYC residue pairs: Ala103/Ala87, Ala103/trimethyl-Lys78, Met233/Gly89, Phe230/Thr18, Ala168/Val34, Gln170/Gln22 and Gln170/Lys33 (Figure 2). Phe230/Arg19 of CYT1/CYC form the cation–π interaction. For the high-resolution structure, it contributes a calculated interaction energy of −7.1 kcal/mol (1 kcal=4.184 kJ) comparable with the energy of a hydrogen bond [5,17]. Cation–π interactions are also present at the interfaces of the photosynthetic reaction centre–cytochrome c2 complex  and the amicyanin–cytochrome c-551i complex from Paracoccus denitrificans . It appears to be an important feature of CYC binding and contributes specificity to the binding interaction, supporting the assumption that there is a distinct bound state with one defined orientation of the reactions partners.
CYC reduction by QCR strongly depends on ionic strength [13,20], clearly indicating an electrostatic contribution to the interaction. Pairs of complementary charged residues form a semicircle on the periphery of the interface (Figure 2). With distances of 5–9 Å, none of them has atoms close enough to form salt bridges or intermolecular hydrogen bonds for the structures with isoform-1 CYC. The higher-resolution structure also showed that there are no water-molecule-mediated hydrogen bonds present . Thus the electrostatic interactions between CYT1 and CYC are weak and favourable for transient binding. In addition, a mobility mismatch of the CYT1 and CYC surfaces, with wellordered negatively charged and mobile positively charged residue side chains respectively, has been suggested to aid transient interaction .
The electrostatic component is thought to accelerate association by limiting diffusion space. The semicircular arrangement indicates that pre-orientation involves not only a parallel alignment of the respective surfaces, but also a defined orientation along the haem–haem axis, with a 55° interplanar angle between the haem cofactors. A similar interplanar angle was observed for the structure of cytochrome c peroxidase with bound CYC . A small interface of hydrophobic interactions surrounded by charged residue pairs is characteristic of transient electron-transfer complexes and was also described for the cyt c peroxidase–CYC and the bacterial reaction centre–cytochrome c2 complexes [18,21].
The core interface
Four contact pairs contribute to the interface of all three structures: Ala103/Ala87, Ala168/Val34, Phe230/Thr18 and the cation–π pair Phe230/Arg19 of CYT1/CYC (Figure 2). The interface residues are identical between the two CYC isoforms, except for Gly89 of isoform-1 which is replaced by homologous Ala93 in isoform-2. The common contact pairs define an area around the haem cleft, which has been defined as the core interface . Furthermore, variable hydrophobic interactions are present peripheral to the core (Figure 2). This coincides with small shifts of the CYC position, while the haem–haem distance and geometry are kept constant.
One should note that the QCR–CYC structures were obtained at different redox states of the complex. Ascorbate-reduced crystals were cryo-trapped for the high-resolution structure, so that Rieske protein, CYT1 and CYC were reduced . For the other two structures, the proteins and the crystals were used as isolated with the complexes in a mixed redox state. Even full reduction does not interfere with the bound state of the complex. Redox-dependent differences may be subtle and may occur especially in the zone of variable interactions. However, structural information for a defined oxidized complex is required to discriminate between redox-specific contributions.
The highly hydrated interface
The high resolution of the structure allowed, for the first time, the analysis of interfacial water molecules. Although the CYT1–CYC interface is approx. 70% non-polar, it is not sealed to the bulk solvent, and a relatively low surface complementarity provides space for hydration (Figure 3). A total of 30 structural water molecules are present in an interface area of 886 Å2, whereas an average protein–protein interface has 1.5-fold fewer water molecules . The water distribution is not equal among the two partners: 24 water molecules are bound to CYT1 and four to CYC. Only two water molecules bridge a hydrogen bond between the two partners. This is correlated with the observed mobility mismatch of the surface-exposed residues of CYC and CYT1.
Uniquely, the univalent CYC binding permitted a direct comparison of free and occupied docking sites in the same crystal structure . Whereas the side chains of CYT1 residues are not affected by CYC binding, their hydration pattern changes significantly. Five additional water molecules and overall ten water molecules in different positions are present on the CYT1 surface when CYC is bound as compared with the vacant binding site. The major rearrangement is in the non-polar region of the interface.
Univalent binding of CYC to QCR
The distance between the two CYT1 docking sites (Figure 3) and the space in the crystal lattice would permit binding of a second CYC molecule. The two CYT1 molecules do not differ in mobility, as indicated by the same average B factor of 42 Å2, and the conformation of CYT1 interface residues is highly similar with an overall RMSD (root mean square deviation) ranging from 0.07 to 0.27 Å, close to the overall co-ordinate error of the structure of 0.13 Å. Besides the considerable rearrangement of water molecules at the interface, univalent CYC binding also coincides with a subtle asymmetry in the dimeric QCR as revealed by the high-resolution structure .
Conformational changes were described for the Rieske head domain and subunit QCR6p. The latter subunit has been implicated in CYC binding . Its N-terminus contains 79% negatively charged residues. In X-ray structures of mitochondrial QCRs, a 40-residue acidic fragment at the N-terminus could not be located, implying its high flexibility and mobility [11,23–25]. In the yeast QCR–CYC complex, the first structurally resolved residue of the N-terminus is located close to the highly conserved residues Lys92 and Lys95 of CYC (Figure 1). The N-terminal peptide might be involved in steering CYC in a reactive conformation . The differences for the Rieske protein are surprising, as the [2Fe–2S] cluster-bearing portion of the extrinsic domain is fixed in both monomers in an identical position by the Qo site inhibitor stigmatellin. However, the Rieske proteins are intertwined between the two monomers, and they have been suggested as regulatory elements in the mechanistic alternating-site model, in which only one monomer of QCR is active at a time . There are also subtle differences in ubiquinone occupancy of the Qi site. In the medium-resolution structure, the latter was higher for the monomer with CYC bound, so that a co-ordinated binding of the two electron acceptors was suggested . This tendency is similar in the high-resolution structure, but the substrate occupancy is generally low and prevents a quantitative analysis . The physiological relevance of the single-site binding observed in the crystal structures has to be addressed by functional analysis and binding studies in solution.
The structures of the QCR–CYC electron-transfer complex obtained at physiological ionic strength have a highly similar interface despite variations in redox state, data collection temperature, amino acid sequence and weak crystal contacts of CYC. The consistency suggests that there is a single defined bound state for the complex. The structural features facilitating the highly transient interaction are a small non-polar interface, a high level of hydration and the mobility mismatch. CYC is rolling on water when bound to the rigid unmodified CYT1 surface. The core interface may create sufficient specificity, providing an optimal geometry of the complex for transient interaction and efficient electron transfer.
This work was supported by the DFG (Deutsche Forschungsgemeinschaft) (SFB 472) and the Center of Excellence “Macromolecular Complexes” at the Goethe University Frankfurt (DFG Project EXC 115).
Integration of Structures, Spectroscopies and Mechanisms: Second Joint German/British Bioenergetics Conference, a Biochemical Society Focused Meeting held at University of Edinburgh, U.K., 2–4 April 2008. Organized by Ulrich Brandt (Frankfurt, Germany), Steve Chapman (Edinburgh, U.K.), Peter Heathcoate (Queen Mary, University of London, U.K.), John Ingledew (St Andrews, U.K.), Mike Jones (Bristol, U.K.), Bernd Ludwig (Frankfurt, Germany), Fraser MacMillan (University of East Anglia, Norwich, U.K.), Hartmut Michel (Max-Planck-Institute for Biophysics, Frankfurt am Main, Germany), Peter Rich (University College London, U.K.) and John Walker (MRC Dunn Human Nutrition Unit, Cambridge, U.K.). Edited by Ulrich Brandt and Peter Rich.
Abbreviations: COX, cytochrome c oxidase; CYC, cytochrome c; CYT1, cytochrome c1; QCR, cytochrome bc1 complex
- © The Authors Journal compilation © 2008 Biochemical Society