Yeast CcO (cytochrome c oxidase) has been developed as a facile system for the production and analysis of mutants of a mitochondrial form of CcO for mechanistic studies. First, a 6H tag (His6 tag) was fused to the C-terminus of a nuclear-encoded subunit of CcO from yeast Saccharomyces cerevisiae. This allowed efficient purification of a WT (wild-type) mitochondrial CcO, 6H-WT (yeast CcO with a 6H tag on the nuclear-encoded Cox13 subunit), with a recovery yield of 45%. Its catalytic-centre activity [≈180 e·s−1 (electrons per s)], UV–visible signatures of oxidized and reduced states and ability to form the PM [‘peroxy’ (but actually a ferryl/radical state)] and F (ferryl) intermediates confirm normal functioning of the histidine-tagged protein. Point mutations were introduced into subunit I of the 6H-WT strain. All mutants were screened for their ability to assemble CcO and grow on respiratory substrate. One such mutant [6H-E243DI (the 6H-WT strain with an additional mutation of E243D in mitochondrial DNA-encoded subunit I)] was purified and showed ~50% of the 6H-WT catalytic-centre activity, consistent with the effects of the equivalent mutation in bacterial oxidases. Mutations in both the D and the H channels affect respiratory growth and these effects are discussed in terms of their putative roles in CcO mechanism.
- cytochrome c oxidase
- energy coupling
- proton channel
- subunit I
CcO (cytochrome c oxidase, also called Complex IV) is the terminal enzyme of many respiratory chains that catalyses transfer of electrons from reduced cytochrome c to molecular oxygen . It is a transmembrane complex embedded in the plasma membrane of Gram-negative bacteria or the inner membrane of eukaryotic mitochondria. In mammals it is a homodimer, with each monomer composed of 13 protein subunits with a combined molecular mass of 204 kDa . In contrast, bacterial homologues are composed of only three or four subunits [3,4]. All forms of CcO share conserved subunits I, II and III that constitute the catalytic core and which, in eukaryotes, are usually encoded by the mitochondrial genome. Subunit III has no prosthetic group and may provide a channel for substrate O2 to diffuse to the catalytic site. Subunit II houses a dinuclear copper centre, CuA, which is the first electron acceptor from cytochrome c. The subsequent electron acceptors, located in subunit I, are haem a and a BNC (binuclear centre) consisting of haem a3 and another copper atom, CuB, where the oxygen reduction chemistry takes place.
A full catalytic cycle of CcO consumes eight protons from the negative (matrix or cytoplasmic) side of the membrane. Four are substrate protons used to produce two water molecules, and four are translocated across the membrane, contributing to the protonmotive force that drives ATP synthesis. All protons must travel through the protein structure along channels composed of protonatable residues and associated water molecules. Two such possible channels were identified in the first bacterial CcO structure from Paracoccus denitrificans : a K channel leading to the BNC and a D channel leading to a conserved buried glutamic acid residue (Glu278 in P. denitrificans, Glu242 in bovine and Glu243 in yeast CcO). A third possible H channel was first identified in bovine CcO  and is more weakly evident in bacterial CcOs. In bovine CcO, this H channel, with breaks, spans the entire subunit I and includes residues that interact directly with haem a ring substituents.
A consensus view has yet to emerge on the roles of these hydrophilic channels in coupled proton transfer and on whether they might even have different roles in some types of CcOs [6,7]. Mutations in all three channel regions of bacterial CcOs have confirmed that the D and K, but not the H, channels have crucial proton-transfer roles [8,9]. Equivalent information in mammalian CcO is sparse because site-directed mutation of the mtDNA (mitochondrial DNA)-encoded core subunits is difficult. However, results with a hybrid system in which bovine subunit I with mutations in the H channel was incorporated into human CcO in an immortal cell line  have suggested that the H channel is the route for translocated protons. The yeast Saccharomyces cerevisiae provides a flexible alternative system for studies of mitochondrial forms of CcO since it is amenable to genetic transformation of both nuclear and mitochondrial genomes. Furthermore, all 11 subunits of yeast CcO are homologous with mammalian CcO subunits and share extensive sequence identities. A predicted model of its structure, on the basis of homology modelling with bovine CcO, suggests that most structural and functional features have been conserved between yeast and mammalian forms . Hence many structure/function aspects of yeast CcO are likely to be similar in mammalian CcOs. Point mutations have to date been successfully introduced in the D and K channels of yeast CcO [12,13], but not in the H channel.
The present paper describes the construction of a modified yeast CcO 6H-WT [yeast CcO with a 6H (His6) tag on the nuclear-encoded Cox13 subunit; WT is wild-type], the 6H tag allowing rapid and efficient purification. Further mutations were introduced in the vicinity of the D or H channels, and the ability of the resulting strains to assemble CcO and grow on respiratory substrate was assessed.
Yeast extract was purchased from Ohly and CO gas was from BOC. All other reagents were purchased from Sigma–Aldrich unless otherwise specified.
Introduction of 6H tags and point mutations in yeast CcO
Modified yeast strains were derived from strain W303-1B . The alleles of the COX4, COX5A, COX8 and COX13 nuclear genes with a 6H tag sequence on their 3′ ends at their chromosomal loci were produced as described previously . A short linker encoding the sequence GARGS was also inserted before the 6H tag of COX13. The sequence of the genes was checked and respiratory growth competence of the strains was monitored to verify that the introduction of the 6H tag did not induce a respiratory deficiency. Mutations in mtDNA-encoded subunit I were then introduced into the 6H-WT by biolistic transformation as described previously .
Respiratory growth competence
Respiratory growth competence was monitored from growth on agar plates with respiratory medium (1% yeast extract, 2% peptone and 2% glycerol). For this test, a strain with a low respiratory growth capacity was used: CKWT (mat a, leu1, kar1-1; derived from ). Because of the low respiratory growth capacity, a moderate defect in respiratory function results in a significant decrease in respiratory growth. The mitochondrial genomes carrying the COX1 mutations were transferred into the CKWT strain by cytoduction , resulting in isogenic strains that differed only in the point mutation in COX1.
Batch growth for protein purification
Strains were first grown in 5 ml of YPGal (1% yeast extract, 2% peptone and 2% galactose) medium for 24 h at 28°C with shaking at 150 rev./min. Cells were then transferred (1:100 dilution) into 50 ml of YPGal in a 1 litre flask and grown under the same conditions. Finally, cells were transferred (1:100 dilution) into 10×500 ml of YPGal in 2 litre high-aeration shake flasks and grown for 16 h at 28°C with shaking at 200 rev./min. Cells were harvested in late exponential phase by centrifugation at 6500 g for 5 min at 4°C. Cells were washed by resuspension/centrifugation cycles in 50 mM KPi (potassium phosphate), pH 7.0, until the supernatant was clear (usually three times). Typically, 16 and 10 g of wet cells/litre of culture were obtained for the 6H-WT and 6H-E243DI (the 6H-WT strain with an additional mutation of E243D in mtDNA-encoded subunit I) strains respectively.
Preparation of mitochondrial membranes
Wet cells were resuspended 1:1 (w/v) on ice in 650 mM D-mannitol, 50 mM KPi, pH 7.4, and 5 mM EDTA. For disruption, glass beads (425–600 μm diameter) were added and 90–95% of cells were broken by mechanical lysis using a bead-beater cell disruptor (BioSpec Products). Cell debris was removed by centrifugation at 5600 g for 20 min at 4°C and mitochondrial membranes were pelleted by centrifugation at 40000 rev./min for 1 h at 4°C using a T647.5 rotor. The membranes were resuspended in 50 mM KPi, pH 7.4, 100 mM KCl and 5 mM potassium EDTA and repelleted, followed by two washing cycles using 50 mM KPi, pH 7.4, and 2 mM EDTA by resuspension/centrifugation at 40000 rev./min for 30 min at 4°C using a T647.5 rotor. The final pellet was homogenized with 50 ml of 50 mM KPi, pH 8.0, and stored at −80°C.
Purification of CcO
The membrane preparation was diluted to 1.8 mg of protein/ml in 50 mM KPi, pH 8.0, and solubilized by incubation on ice for 30 min with 2% (w/v) DDM (n-dodecyl β-D-maltoside; Melford Laboratories). The solubilized material was centrifuged at 40000 rev./min for 35 min at 4°C using a T647.5 rotor. Imidazole at 5 mM was added to the supernatant before it was incubated for 1 h with Ni2+-iminodiacetic acid resin (His-bind® resin, Novagen) previously equilibrated with 50 mM KPi, pH 8.0. The resin was loaded into a column and washed with 5 column vol. of 50 mM KPi, pH 8.0, 150 mM KCl, 5 mM imidazole and 0.009% DDM followed by 3 column vol. of 50 mM KPi, pH 8.0, 10 mM imidazole and 0.009% DDM. The protein was eluted by increasing the imidazole to 100 mM and was directly loaded on to a DEAE Sepharose CL-6B column previously equilibrated with 20 column vol. of 50 mM KPi, pH 8.0, and 0.015% DDM. The column was washed with 2 column vol. of 50 mM KPi, pH 8.0, 50 mM NaCl and 0.01% DDM. The purified protein was eluted by increasing the NaCl to 250 mM and concentrated in a pressure cell with a 100 kDa cut-off membrane (YM100, Millipore) to 10–20 μM. CcO concentration was estimated from a dithionite-reduced minus oxidized visible absorption difference spectrum using an assumed Δϵ603–621nm of 26 mM−1·cm−1 . SDS/PAGE was performed in denaturating conditions with a 16.5% acrylamide gel containing 6 M urea as described previously [18,19].
O2-consumption rates and catalytic-centre activities
CcO was diluted to 10–20 nM in 67 mM KPi, pH 6.2, containing 40 μM TMPD (N,N,N′,N′-tetramethyl-p-phenylenediamine), 50 μM horse heart cytochrome c and 0.05% DDM in the chamber of a Clark-type oxygen electrode at 24°C. The reaction was initiated by addition of 20 mM sodium ascorbate. Catalytic-centre activities [expressed in e·s−1 (electrons per s)] were calculated from the steady-state rates of oxygen consumption.
UV–visible absorption spectroscopy
UV–visible absorption spectra were recorded with a single beam spectrometer constructed in-house and equipped with a stepped dispersive monochromator and photomultiplier detector. All spectra were recorded in a standard 1 cm-path-length quartz cuvette in 50 mM KPi, pH 8.5, and containing 0.009% DDM. Starting from the oxidized (O) state, the following CcO species were generated: FR (fully reduced), with excess of sodium dithionite; FR-CO compound, by saturation of a FR sample with CO gas; PM [‘peroxy’ intermediate (but actually a ferryl/radical state)], by bubbling with CO under aerobic conditions; and F (ferryl intermediate), by addition of 500 μM H2O2. The extent of PM or F formation was assessed from ΔA610–633 or ΔA580–633 using assumed molar absorption coefficients of 10.1 and 4.3 mM−1·cm−1 .
Photolysis of the FR-CO compound was achieved with a single flash from a Xenon flashlamp (20 J/flash; 6 μs duration at half maximum) that was filtered through a 520 nm-cut-off filter. The photomultiplier was screened with a 570 nm (OG570) cut-on filter. Photolysis and subsequent recombination kinetics at room temperature (24°C) were monitored over 85 ms at 590 minus 610 nm, and 50 individual transients were averaged to improve signal/noise. Averaged data were fitted to a single exponential decay using OriginPro 8.1 (OriginLab Corporation).
RESULTS AND DISCUSSION
Construction of yeast CcO with a 6H tag and mutations in subunit I
A 6H tag was fused to the C-terminus of the nuclear-encoded subunits Cox4, Cox5A, Cox8 and Cox13 as described in the Experimental section. The presence of the 6H tag on any of these subunits had no significant effect on the respiratory growth competence, indicating that it did not have a major impact on CcO assembly and/or activity in vivo. Steady-state levels of expression of the 6H-tagged subunits were assessed by Western blotting (Figure 1). The levels of immunodetected 6H-tagged Cox4 and Cox8 were significantly lower than those of 6H-tagged Cox5A and Cox13. The lower levels could have resulted from faster degradation of the tagged proteins or from proteolysis of the 6H tag from the subunits. Since the highest level of expression was that of 6H-tagged Cox13, and because an alternative isoform, Cox5B, can replace Cox5A under some conditions , this strain (6H-WT) was selected for further work. Mutations in mtDNA-encoded subunit I were then introduced into 6H-WT by biolistic transformation as described previously . The resulting strains are isogenic to 6H-WT, differing only in the point mutation.
Purification of 6H-tagged CcOs
The 6H-WT strain was grown in 5 litres of medium, yielding approximately 80 g wet mass of cells containing 160 nmol of CcO (assessed from a dithionite-reduced minus oxidized spectrum of whole resuspended cells; results not shown). Mitochondrial membranes were isolated after cell disruption, solubilized with DDM and the CcO purified by Ni2+-chelating affinity and DEAE Sepharose CL-6B ion-exchange chromatographies (see the Experimental section). Typically, 70 nmol of purified CcO was obtained, giving a purification yield of 45%. SDS/PAGE of the subunits of a representative 6H-WT CcO preparation is shown in Supplementary Figure S1 at http://www.BiochemJ.org/bj/444/bj4440199add.htm, where bands were tentatively assigned by comparison with previous results . The same procedure was used for the 6H-E243DI strain containing the E243D mutation in subunit I. In this case, the same growth conditions yielded 50 g wet mass of cells (40% less than 6H-WT). However, these cells contained 100 nmol of CcO, the same relative amount of CcO per g wet mass of cells.
The purity of each CcO preparation was routinely assessed by UV–visible spectroscopy from the absorbance ratio of protein at 280 nm/haem Soret band at 425 nm; this was between 3 and 4, depending on the quality of the preparation. Catalytic activities were assessed using a Clark-type oxygen electrode in 67 mM KPi at pH 6.2, conditions previously reported to give optimal activities for WT yeast CcO . The 6H-WT protein had a catalytic-centre activity under these conditions of 180 e·s−1, similar to the value reported by Taanman and Capaldi  for purified WT yeast CcO. This confirmed that addition of the 6H tag to the Cox13 C-terminus did not alter the catalytic-centre activity. However, the additional introduction of the E243D mutation in subunit I reduced the catalytic-centre activity to 93 e·s−1, representing 52% of the 6H-WT value under the same conditions. Similarly decreased activities have been reported for equivalent glutamate/aspartate mutants in CcO of P. denitrificans (60% activity relative to WT ), Rhodobacter sphaeroides (50% ) and the bo-type oxidase of Escherichia coli (31% ).
UV–visible absorption spectroscopy
Figure 2(A) shows absorption spectra of purified 6H-WT CcO. The O form (Figure 2A, continuous line) has a Soret band maximum at 425 nm, consistent with that of WT yeast CcO [19,25]. Addition of dithionite led to the characteristic 444 nm band of the FR state (Figure 2A, dotted line). Subsequent bubbling with CO induced the shift characteristic of formation of the FR-CO compound (Figure 2A, dashed line); assuming a molar absorption coefficient of 10.5 mM−1·cm−1 at 590–610 nm , its extent indicated essentially full formation of the CO compound. The 6H-E243DI mutant displayed equivalent spectra (results not shown), confirming that the UV–visible properties of the two haem centres were not affected. The kinetics of CO recombination after flash-photolysis of the FR-CO compound of the 6H-WT could be fitted to a single exponential with a rate constant of 82 s−1 at 24°C (Supplementary Figure S2 at http://www.BiochemJ.org/bj/444/bj4440199add.htm), consistent with previous measurements on the WT enzyme . A similar rate constant for CO recombination was measured for the 6H-E243DI mutant, also consistent with previous measurements on isolated mitochondria .
Spectra of the PM and F reaction intermediates  of the 6H-WT CcO were also recorded (Figure 3). PM was formed by reaction of CO with the O state in aerobic buffer. The PM minus O difference spectrum at pH 8.5 (Figure 3A) is characterized by a trough/peak in the Soret region at 414/438 nm, a broad band at 567 nm and an α-band at 610 nm (which is red-shifted in comparison with the 607 nm peak of PM of bovine CcO). As observed in bovine CcO , the extent of PM formation was pH-dependent (results not shown) with approximately 26% (ΔA610–633=0.008) formed at pH 8.5. The kinetics of reaction with CO to form PM were also markedly biphasic, with the slow phase most probably caused by a significant fraction of the oxidized enzyme being in the chloride-ligated form .
Figure 3(B) shows the difference spectrum induced by addition of 500 μM H2O2 to oxidized 6H-WT at pH 8.5, a reaction that in bovine CcO results in transient formation of PM followed by full conversion into F . In the case of the yeast 6H-WT, although the same O→PM→F sequence occurs, a significant amount of PM remains in the steady state. Nevertheless, the spectrum has a Soret band that is slightly blue-shifted in comparison with PM, together with a broad band centred at 580 nm that can be attributed to the F form. The extent of F formation from O was estimated to be 12% from the ΔA580–633.
Minor traces of an additional haem could be present to a variable extent, appearing as peaks at 422 and 554 nm in FR minus O difference spectra. In Figure 2(B) (lower trace), this species is marked with ‘*’. The same species can be seen in previously reported yeast CcO preparations . It might be a contaminating haemoprotein or, possibly, a partially assembled/disassembled form of CcO.
Effect of mutations in the D and H channels on respiratory growth
A range of mutations were introduced into subunit I of the 6H-WT strain. These mutations are located in the vicinity of either the D or H channels (Figure 4). Figure 5 shows the growth profiles of all mutant strains. All three mutations in the D channel (E243D, N99D and I67N) affected respiratory growth. In the case of the E243D mutation, this correlates with a 50% lower O2-consumption rate of the purified CcOs (see above). I67N had previously been generated by random mutagenesis. It severely inhibited CcO turnover  and it was proposed that this might be caused by an interaction with Glu243 that prevented it from changing configuration. N99D was chosen to mimic the equivalent mutations of N131D in P. denitrificans and N139D in R. sphaeroides which result in 62 and 270% of WT activity respectively [9,31]. Figure 5 shows that this mutation in yeast greatly diminishes the ability of the strain to grow on respiratory medium, suggesting an inhibitory effect. Taken together, these results confirm the importance of the D channel in the catalytic cycle of yeast CcO.
Of the strains with mutations in the vicinity of the H channel, two (R37M and A446D) failed to assemble CcO. X-ray data of bovine CcO showed that Arg38 is in hydrogen-bond interaction with the formyl group of haem a. Mutation of the equivalent residue in R. sphaeroides led to either loss of activity (R52A/Q) or a decrease in catalytic-centre activity to 65% of the WT level (R52K) . In all cases the redox spectra of assembled CcOs with Arg52 mutations showed a shift of the α-band maximum, confirming its interaction with haem a. Its replacement with a methionine residue in yeast could destabilize binding of haem a and preclude enzyme assembly. A446D, which had been identified as causing respiratory deficiency by random mutagenesis, is predicted to be located at the interface between subunits I and II (Figure 4). Hence it is likely that the subunit I–II interaction is destabilized by the replacement of a small apolar alanine residue with a negatively charged aspartic acid. Replacement of Arg37 and Ala446 with alternative residues that may result in assembled CcO will be investigated in future work.
All other H channel mutant strains successfully assembled CcO. Replacement of Gln413 (equivalent to bovine His413 and after which the channel was named) with a leucine residue did not affect respiratory growth, in contrast with the dramatic effect of the nearby mutation Q411L. Mutation S382A, a residue in hydrogen-bond interaction with haem a farnesyl, had a similar effect on respiratory growth as E243D. Its replacement with an alanine residue in R. sphaeroides decreased the CcO catalytic-centre activity to 67% of the WT level . S458A had an even more pronounced effect on respiratory growth. This residue is also present in bovine CcO, but is an alanine in P. denitrificans and R. sphaeroides CcOs. In bovine CcO it is in hydrogen-bonding distance with two crystallographically resolved water molecules and might also interact with the farnesyl OH of haem a in a redox-dependent manner . In contrast, mutation S455A is not expected to interact with haem a (Figure 4) and had little effect on respiratory growth, in agreement with effects of the equivalent mutation in P. denitrificans . A multiple mutant of Q411L/Q413L/S458A/S455A was also constructed (4Hmut, Figure 5). Growth on respiratory medium was barely detectable, suggesting that effects of some individual single point mutations could be somewhat additive. Mutation S52D was also introduced at the ‘top’ of the H channel to mimic the equivalent bovine CcO Asp51, a residue whose conformation is redox-linked and proposed to provide the proton exit route, or valve, for proton translocation through the H channel. Its mutagenesis to an asparagine residue in bovine subunit I results in loss of coupled proton translocation without inhibiting electron transfer . In yeast CcO this mutation had no effect on respiratory growth, and in bacteria a glycine is present in the equivalent position. Mutation D445E, first identified by random mutagenesis to inhibit respiration , was also introduced in the 6H-WT strain. The homology model suggests that it is at the interface with Cox5A, a potential regulatory supernumerary subunit . Finally, the mutation E39Q, equivalent to bovine Glu40 that forms the Ca2+/Na+-binding site near the ‘top’ of the H channel, also markedly affected respiratory growth. Overall, these H channel mutants indicate that this structure is playing a function of some kind in yeast CcO. Whether this is one of coupled proton translocation , of providing a dielectric channel  or, possibly, other control functions can now be addressed through purification and mechanistic studies of these and related mutations.
Brigitte Meunier, Amandine Maréchal and Peter Rich designed the research and wrote the paper. Brigitte Meunier produced and characterized all histidine-tagged and mutant CcO strains. Amandine Maréchal grew the yeast cells, purified the CcOs and carried out the biochemical and biophysical analyses.
This work was supported by the Agence Nationale de la Recherche [grant number ANR-07-BLAN-0360-02 (to B.M.)] and the Biotechnology and Biological Sciences Research Council U.K. [grant number BB/H000097/1 (to P.R.R.)].
We thank Thomas Warelow and Talha Arooz for technical assistance.
Abbreviations: BNC, binuclear centre; CcO, cytochrome c oxidase; DDM, n-dodecyl β-D-maltoside; F, ferryl intermediate; FR, fully reduced; 6H tag, His6 tag; 6H-WT, yeast CcO with a 6H tag on the nuclear-encoded Cox13 subunit; 6H-E243DI, the 6H-WT strain with an additional mutation of E243D in mtDNA-encoded subunit I; KPi, potassium phosphate; mtDNA, mitochondrial DNA; PM, ‘peroxy’ intermediate (but actually a ferryl/radical state); WT, wild-type; YPGal, 1% yeast extract, 2% peptone and 2% galactose
- © The Authors Journal compilation © 2012 Biochemical Society