Disease-related mutations in cytochrome c oxidase studied in yeast and bacterial models Melyssa Bratton 1 , Denize Mills 2 , C. Kate Castleden 3 , Jonathan Hosler 1 and Brigitte Meunier 3 1 Department of Biochemistry, University of Mississippi Medical Center, USA; 2 Department of Biochemistry, Michigan State University, USA; 3 Wolfson Institute for Biomedical Research, University College London, England Mitochondrial cytochrome c oxidase is a key protonmotive component of the respiratory chain. Mutations in the mito- chondrially-encoded subunits of the complex have been reported in association with a range of diseases. In this work we used yeast and bacterial mutants to assess the effect of human mutations in subunit 1 (L196I) and subunit 3 (G78S, A200T, DF94–F98, F251L and W249Stop). While the stop mutation at the C-terminus of subunit 3 and the short deletion were highly deleterious and abolished the assembly of the mitochondrial enzyme, the four missense mutations caused little or no effect on the respiratory function. Detailed analysis of G78S, A200T and DF94–F98 in Rhodobacter sphaeroides confirmed and extended these observations. We show in this study that the combination of yeast and bac- terial models is a useful tool to elucidate the effect of muta- tions in the catalytic core of cytochrome oxidase. The yeast enzyme is highly similar to the human enzyme and provides a good model to assess the deleterious effect of reported mutations. The bacterial system allows detailed biochemical analysis of the effect of the mutations on the function and assembly of the catalytic core of the enzyme. Keywords: cytochrome oxidase; diseases; Rhodobacter; assembly. Cytochrome oxidase (complex IV), embedded in the inner mitochondrial membrane, is the terminal enzyme complex of the mitochondrial respiratory chain. It catalyses the reduction of oxygen to water and the translocation of protons across the mitochondrial membrane. This process contributes to the electrochemical proton gradient, which is then used to drive ATP synthesis by the ATP synthetase. Mitochondrial cytochrome oxidase is composed of up to 13 subunits. Ten subunits in mammals (eight in yeast) are encoded by the nuclear genome. Three subunits (subunits 1, 2 and 3) are encoded by the mitochondrial genome. They form the catalytic core of the eukaryotic complex and are homologous to the subunits of the aa 3 -type cytochrome c oxidases of Rhodobacter sphaeroides and Paracoccus deni- trificans. Subunit 2 forms part of the docking site for cytochrome c and binds Cu A , the first electron acceptor. Subunit 1 binds heme a and the binuclear centre (heme a 3 -Cu B ), the site of oxygen reduction. Subunit 3 has no redox centre; it has been shown to prevent suicide inactivation of the Rhodobacter aa 3 -type oxidase via long range interactions with the Cu B centre [1]. The role of the nuclear encoded subunits is unclear but many of them are likely to have an assembly or stability function. The assembly of the enzyme is a complex process and requires a large number of nuclear factors, which have been identified by extensive studies of yeast respiratory mutants, deficient in cytochrome oxidase (reviewed in [2]). Cytochrome oxidase deficiency in humans is associated with a wide range of clinical phenotypes. Most cytochrome oxidase deficiencies are autosomal recessive and usually show early onset and a fatal outcome. Mutations have been found in four of the nuclear-encoded assembly factors, namely, Surf1p, Cox10p, Sco1p and Sco2p [3–7]. Mutations in the nuclear-encoded structural subunits of the enzyme have not been observed but disease-associated mutations in the mitochondrially encoded subunits 1, 2 and 3 have been reported. These mutations are relatively rare and the disease symptoms usually present during late childhood to adulthood. They are associated with a variety of clinical presentations including LHON (Leber’s here- ditary optical neuropathy), MELAS (mitochondrial ence- phalomyopathy, lactic acidosis and stroke like episodes), AISA (acquired idiopathic sideroblastic anaemia) and encephalomyopathy. In addition to these mutations sus- pected to cause enzymatic dysfunctions, a number of nonpathogenic polymorphisms in the mitochondrial genes for subunits 1, 2, and 3 have been reported (www.mito map.org). In this paper, we re-examine the disease mutations in the mitochondrially-encoded subunits 1 and 3 by using yeast and R. sphaeroides mutants to explore their effect on the respiratory function. We have previously produced five yeast strains, which carry mutations found in patients, A223S, M273T, I280T and G317S in subunit 1, and a short in-frame deletion DF94–F98 (human sequence) in subunit 3. In this work, another five human mutations were studied in yeast: L196I in subunit 1, and G78S, Correspondence to B. Meunier, Wolfson Institute for Biomedical Research, University College London, Gower Street, London WC1E 6BT, UK. Fax: 44 20 79165994, E-mail: b.meunier@ucl.ac.uk Abbreviations: AISA, acquired idiopathic sideroblastic anaemia; CCCP, carbonyl cyanide m-chlorophenylhydrazone; Cox, genes encoding for cytochrome oxidase subunits; LHON, Leber’s hereditary optical neuropathy; MELAS, mitochondrial encephalomyopathy lactic acidosis and stroke like episodes. (Received 20 November 2002, revised 23 January 2003, accepted 27 January 2003) Eur. J. Biochem. 270, 1222–1230 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03482.x A200T, F251L and W249Stop in subunit 3 (Fig. 1, Table 1). The subunit 3 mutations G78S, A200T and DF94–F98 were also introduced in R. sphaeroides.This allowed us to compare the effects of the mutations in two different systems. The yeast enzyme is highly similar to the human enzyme (there is, for instance, 57% and 44% amino acid identity between subunits 1 and 3, respectively) and therefore provides a good model to assess the deleterious effect of reported mutations. The structure of the R. sphaeroides cytochrome oxidase is nearly identical to that of the catalytic core of the mammalian enzyme, as predicted by their sequence identities (54% and 49% between subunits 1 and 3, respectively). The bacterial system allows detailed biochemical analysis of the effect of the mutations on the function and assembly of the catalytic core of cytochrome oxidase. Table 1. Disease-related mutations in the mitochondrially encoded subunits of cytochrome oxidase. Mutations Disease and references Yeast sequence Effect of the mutation in yeast References Subunit 1 L196I Epilepsy [23] L197 No effect This work A223S Multisystem disorder [24] A224 No effect [9] Y260H Myopathy [35] T261 Not studied M273T AISA [36] M273 Mildly deleterious [8] I280T AISA [36] I280 Mildly deleterious [8] G317S Polymorphism [25] G317 No effect [8] Ter514K,Q,K,Ter LHON [26] Not studied McArdle’s disease [27] Subunit2 M1T Encephalomyopathy [22] M1 Not studied M29K Myopathy [37] F45 Not studied A41T Multisystem disorder [38] M57 Not studied Subunit 3 G78S LHON [28] G86 No effect This work A178T LHON [29] Y186 Not studied A200T LHON [28] G208 No effect This work F251L MELAS [32] F259 No effect This work DF94–F98 Myoglobinuria [34] [20] F102–F106 Deleterious [8] W249stop Encephalopathy [33] W257 Deleterious This work Fig. 1. Location of disease-related mutations in the catalytic core of cytochrome oxidase. The structure has been drawn from the co-ordi- nates of the bovine enzyme [17]. Subunits 1 and 3 are in light and dark blue, respectively. The mutations are in yellow. The hemes a and a 3 are in red. The numbering is according to the human sequence, which is highly identical to the bovine sequence (approximately 90%). Ó FEBS 2003 Disease-mutations in cytochrome oxidase (Eur. J. Biochem. 270) 1223 Materials and methods Introduction of site-directed mutations in the mitochondrially encoded genes in yeast The yeast strains and media used in this study were described previously [8]. Site-directed mutagenesis, biolistic transformation of mitochondria, screening of the trans- formants and replacement of the wild-type cox1 and cox3 genes by the mutated form in the mitochondrial genome were performed as described in [8,9]. Spectrophotometric measurements of yeast cells Spectra were generated by scanning cell suspensions reduced by dithionite with a single beam instrument built in-house. The cells, grown on 1% yeast extract/2% peptone/3% glucose plates for 48 h, were resuspended in 5% Ficoll at a concentration of approximately 200 mg of cells per ml and reduced by dithionite. A quadratic baseline compensation was carried out on the data as described in [10] to remove the distortion of the baseline. Production of the Rhodobacter sphaeroides mutants In order to create the subunit 3 mutants G78S, A200T (A205 in R. sphaeroides)andDF94–F98 in the R. sphaero- ides aa 3 -type oxidase, plasmid pMB301 [11], containing only coxIII (the gene for subunit 3) of R. sphaeroides,was mutagenized using the QuikChange site-directed mutage- nesis system (Stratagene). Presence of the correct mutation was verified by DNA sequencing of the altered coxIII genes. A 956-bp SmaI fragment was restricted from each of the mutated pMB301 plasmids and cloned into pMB307 [11], yielding pUC-based plasmids that contained coxI6Xhis on one strand and the coxII-III operon (coxII, cox10, cox11, coxIII) on the other. An EcoR1–HindIII fragment that contained all of these cox genes was restricted from the pMB307 plasmids and cloned into pRK415 [12] in order to create plasmids capable of replicating and expressing in R. sphaeroides. These three pRK415-based plasmids, pMAG78S, pMAA205T and pMADF94–F98, were each conjugated into R. sphaeroides YZ200, a strain with a deletion of the genomic copy of the coxII-III operon [13], by established methods [14]. Preparation of bacterial cytochrome c oxidase, measurement of oxygen consumption and proton pumping activity, and determination of subunit 3 content R. sphaeroides cells were grown in minimal medium to late exponential phase and cytochrome c oxidase was purified from cytoplasmic membranes solubilized in N-dodecyl-b- D -maltoside by chromatography on Ni 2+ -nitrilotriacetic acid agarose (Qiagen) as previously described [13]. Oxygen reduction assays were as described in [1] and proton pumping was measured in a stopped-flow apparatus as described in [15]. The content of subunit 3 was determined by densitometry of Coomassie-stained SDS-urea gels [16] using a Personal Densitometer and ImageQuant software (Molecular Dynamics). Subunit 3/subunit 2 density ratios were determined for several different loadings of each oxidase and compared to the same ratio for the wild-type oxidase purified by the same method and run on the same gel. Thus, the staining intensity of subunit 2 serves as an internal control in each lane and the wild-type oxidase provides the baseline value of subunit 3 content on each gel. Results and discussion Generation of the yeast and Rhodobacter sphaeroides mutants The residue numbers used throughout are those of human cytochrome c oxidase (unless otherwise indicated). As the catalytic core of human enzyme is essentially identical (90%) to that of the bovine enzyme, the structure of the bovine oxidase can be used to predict some of the conse- quences of the human mutations. Most of the mutations discussed below occur in residues that are completely conserved in humans, yeast and R. sphaeroides;theremain- ing residues are conservative replacement within conserved regions. Six human mutations reported in patients suffering from a range of disorders were studied here: G78S, DF94– F98, A200T, F251L and W249Stop in subunit 3, and L196I in subunit 1 (Table 1). As shown in Fig. 1, the mutations are located far from the redox centres of cytochrome oxidase. They were therefore unlikely to directly affect the catalytic activity of the enzyme but they might alter its assembly. L196I, G78S and DF94–F98 are located at the interface between subunits 1 and 3 and might weaken the assembly of these two subunits. A200T is close to residue S195, which is required for enzyme assembly [9]. F251L and W249Stop, located at the C-terminal end of subunit 3 might alter its folding. In order to assess the effect of the mutations on the respiratory function, and in particular on the assembly and/or stability of the complex, the mutations were introduced into yeast mitochondrial genome using biolistic transformation methods as described in [8,9]. The yeast mutants were then used to monitor the effect of the mutations on the respiratory growth and on the cytochrome oxidase content. Three mutations, G78S, A200T and DF94– F98 in subunit 3 were chosen for more detailed analysis and introduced in R. sphaeroides (Materials and methods). The effect of the mutations on oxygen consumption, proton- pumping activity, and on the binding of subunit 3 were examined. W249Stop and DF94–F98 in subunit 3 alter the assembly of cytochrome oxidase The W249Stop mutation, which is predicted to result in the loss of the last 13 amino acids of subunit 3, had a dramatic effect on respiratory function in yeast. The cells were unable to grow on respiratory medium (Fig. 2B). The mutation abolished the assembly of the complex as no cytochrome oxidase signal was detected by optical spectroscopy (Fig. 2A). That seems to indicate that the well-conserved C-terminal end of subunit 3 is required for the correct folding and assembly of the subunit into the oxidase complex. This severe effect on enzyme assembly was identical to that induced by the short deletion DF94–F98. The deletion of F94–F98 in helix 3 is likely to severely 1224 M. Bratton et al.(Eur. J. Biochem. 270) Ó FEBS 2003 compromise the ability of subunit 3 to bind to subunit 1, particularly as one of the principal contacts between the two subunits is an ion pair of H103 (subunit 3)-D227(subunit 1) located one turn above F98 [17,18]. Shortening of helix 3 should disrupt this interaction. The W249Stop mutation may cause a similar disruption as the region of helix 7 from W249 onwards comes close to the F94–F98 region of helix 3. So, like DF94–F98, loss of the C-terminus of subunit 3 could weaken the H103 salt bridge at the top of helix 3 and disrupt assembly. Because in yeast, subunit 3 is required for the assembly or stability of the other subunits and unassembled subunits are rapidly degraded by the AAA proteases (ATPases associated with diverse cellular activities) of mitochondria, no further analysis could be performed in yeast. In order to study in more detail the assembly defect caused by DF94–F98, the short-deletion was introduced in R. sphaeroides as described in Materials and methods. Contrary to the yeast enzyme, the bacterial subcom- plex containing only subunits 1 and 2 is stable in the absence of subunit 3, which allows further analysis. As expected, the mutant enzyme contained little subunit 3 (Fig. 3), but the remaining subunits 1–2 oxidase was active (V max ¼ 900 s )1 ). However, as a result of the loss of subunit 3, the mutant enzyme underwent rapid suicide inactivation with catalytic turnover (Fig. 4) and it pumped protons with reduced efficiency (Fig. 5). The loss of subunit 3 could be due to weaker binding to subunit 1, as suggested above, and/or to structural insta- bility and degradation of the mutant subunit in vivo. When DF94–F98 was isolated from R. sphaeroides cells grown only to mid-log phase, the purified enzyme contained 20– 25% of the normal amount of subunit 3 (Fig. 3). However, what seemed to be proteolytic fragments were apparent below the subunit 3 band. Indeed, when the mutant oxidase was isolated from bacterial cells grown to late stationary phase, where protein degradation is highly active, subunit 3 was completely absent in the purified product. These data suggest that in the bacterial cell, the short deletion in helix 3 affects the folding and stability of subunit 3. In contrast, the wild-type oxidase was purified from stationary phase cells with normal amounts of subunit 3. Bound subunit 3 was not removed from the DF94–F98 mutant enzyme by Fig. 3. Subunit 3 content of the R. sphaeroides mutants. Cytochrome oxidase samples were separated on SDS-polyacrylamide gels contain- ing urea as described in [16] and stained with Coomassie Blue. The location of subunit 1 (M r ¼ 48), subunit 2 (M r ¼ 37 and 35) and subunit 3 (M r ¼ 20) are indicated by the arrows. Subunit 2 runs as a doublet due to incomplete C-terminal processing [13]. Lanes A and G contain wild-type oxidase; lane B, G78S; lane C, A200T; lane D, DF94– F98 isolated from cells grown to mid-log phase; lane E, as D but repurified on Ni 2+ -nitrilotriacetic acid agarose; lane F, DF94–F98 isolated from cells in stationary phase. Based on densitometry meas- urements (see Materials and methods) G78S and A200T contain as much subunit 3 as the wild-type oxidase and DF94–F98 in lanes D and Econtains 25% of the normal amount of subunit 3. Note that subunit 2 rather than subunit 1 is used as the reference in order to determine the amount of subunit 3. In the absence of subunit 3, sig- nificant amounts of a free form of subunit 1 (termed subunit Ia,see[19]) accumulate in the membrane. As subunit Ia contains a histidine tag, it is isolated along with the subunit 1–2 oxidase and leads to an apparent overabundance of subunit I in the absence of subunit 3 (lane F). Fig. 2. Respiratory growth and optical spectra of the yeast strain harbouring human mutations in cytochrome oxidase subunits 1 and 3. (A) Optical spectra of reduced cell suspensions (see Materials and methods). (B) Respiratory growth. The cells were incubated on glycerol medium for 4 days at 28 °C. Ó FEBS 2003 Disease-mutations in cytochrome oxidase (Eur. J. Biochem. 270) 1225 increasing the detergent concentration or by multiple chromatographic separations on Ni 2+ -nitrilotriacetic acid agarose (Fig. 3). Thus, while weaker binding of the mutant subunit 3 seems likely, it could not be demonstrated by this method. The results indicate that the short deletion affects the assembly of the oxidase at some point after the assembly of subunits 1 and 2. This is consistent with previous studies showing that subunit 3 is not required for assembly of the redox centres in subunits 1 and 2, nor is subunit 3 required for the association of subunits 1 and 2 [11,19]. Assuming that the assembly of the catalytic core is not largely different in mitochondria and bacteria, the same effect wouldbeexpectedinyeastandhumancells.However, while subunits 1 and 2 form a stable subcomplex in the bacterial membrane, the analogous subcomplex fails to accumulate in the inner mitochondrial membrane in the absence of normal subunit 3 [20]. If subunit 3 is not directly necessary for the association of subunits 1 and 2 in mitochondria, it seems likely that the binding of subunit 3 is a necessary prerequisite for the binding of critical, nuclear-encoded subunits that stabilize the growing com- plex. Thus, while the binding of subunit 3 seems to be an end-stage event in the assembly of the bacterial oxidase, it may be an indispensable middle step in the assembly of the mitochondrial enzyme. The missense mutations G78S, A200T, F251L in subunit 3, and L196I in subunit 1 have no effect on the respiratory function The mutations G78S, A200T, F251L in subunit 3, and L196I in subunit 1 were first studied in yeast. In contrast to the severe effects of W249Stop and DF94–F98, the four missense mutations had no or little effect on respiratory growth competence at 28 °C (Fig. 2B). The doubling time of the mutant and control strains in respiratory medium was approximately four hours. Cytochrome oxidase con- tent was monitored and, as expected, no difference in enzyme content was observed (Fig. 2A). As it has been observed that mutants with defective assembly of cyto- chrome oxidase were more affected at higher temperature [9], we monitored the respiratory growth and cytochrome oxidase content at 35 °C. Again, the mutations had no effect (data not shown). We have previously observed that a relatively small decrease in cytochrome oxidase content or activity strongly affects the respiratory growth compet- ence of yeast cells [9]. It seems that there is limited buffering capacity in respiratory function in the strains used in our studies. Monitoring the respiratory growth of the mutant strains is therefore a sensitive test to assess the deleterious effect of the mutations. The results indicate that the missense mutations had little effect on cytochrome oxidase content or activity in yeast. For the L196I and F251L mutations, these results were not unexpected. Leucine196 is located in transmembrane helix 5 of subunit 1, which is close to subunit 3. However the replacement of leucine by isoleucine should effect only a minor steric change and the substitution does not modify the polarity of the residue. It seems likely that the enzyme can accommodate a slightly larger sidechain at that position as the closest residues in subunit 3 are 5–6 A ˚ away. F251 is located in helix 7 at the C-terminal part of subunit 3. In the bovine enzyme, its sidechain extends into the lipid bilayer and has no obvious contact with other subunits. It is likely that the replacement of a phenylalanine by a leucine does not distort the helix and alter the folding of the subunit. On the basis of the structure, it might have been expected that G78S and A200T could hinder, at least slightly, the folding or assembly of the enzyme. G78 is located at the base of helix 3 of subunit 3. The alpha carbon of G78 comes close to the ring of F94 of subunit 1 (4 A ˚ ), which makes hydrophobic contact with PE9, one of the two phospholipids specifically bound in the cleft of subunit 3 [18]. Alteration of F94 to alanine appears to slightly weaken the interaction between subunits 3 and 1, as evidenced by  20% reduction in the content of subunit 3 in the purified R. sphaeroides oxidase (Hosler, unpublished results). There- fore, a polar or bulky group at position 78 could potentially disrupt the contact between F94 of subunit 1 and PE9 of subunit 3. A200 is located toward the top of the helix 6 of subunit 3. Its methyl group extends into the centre of the five helix bundle, into a locally hydrophobic area containing the sidechains of F94, L252 and I256 of subunit 3 [17,18]. The introduction of the longer and more polar sidechain of threonine into this region could potentially destabilize the five-helix bundle and hinder the subunit folding. A200 is also close to residue S195, which is involved in assembly or Fig. 4. O 2 reduction activity of cytochrome c oxidases of R. sphaeroides mutants. A200T (5 pmol), DF94–F98 (9 pmol) and Subunit 3 (–), the oxidase containing only subunits 1 and 2 (6 pmol) were assayed for O 2 reduction (as O 2 uptake) as in [1]. O 2 uptake was initiated by the addition of 40 l M horse heart cytochrome c. The inactivation shown by Subunit 3 (–) and DF94–F98 is irreversible. The O 2 reduction activity of G78S and the wild-type oxidase was essentially identical to that shown for A200T. 1226 M. Bratton et al.(Eur. J. Biochem. 270) Ó FEBS 2003 stability of the enzyme complex in yeast [9]. In addition, it has been suggested that A200T affected the proton transfer activity of the enzyme [21] in Paracoccus denitrificans. In order to study further their effects, G78S and A200T were introduced in R. sphaeroides.Bothmutants showed wild-type levels of O 2 reduction activity (V max of G78S ¼ 1600 s )1 ; V max of A200T ¼ 1860 s )1 )withno indication of suicide inactivation (Fig. 4), normal proton pumping (Fig. 5) and no loss of subunit 3 (Fig. 3). These results show that any disruption of the F94–PE9 interaction caused by G78S is not sufficient to weaken the interaction between subunits 1 and 3 to the point where subunit 3 fails to bind, and that the introduction of a threonine in position 200 does not compromise folding and binding of subunit 3. In addition, the normal O 2 reduction and proton pumping activity of the A200T mutant argues against a proposed role for this region of subunit 3 as an exit pathway for protons [21]. Mutations in mitochondrially encoded subunits of cytochrome oxidase in humans: correlating pathogenicity to the biochemistry elucidated in yeast and bacterial models Mitochondrial genes are present in hundreds of copies in human cells. Heteroplasmy is a common feature for mitochondrial genome mutations. The severity of the respiratory defects and of the disease depends on the load of mutated genes, which varies between tissues. A few nonsense and frameshift mutations in the mitochondrially- encoded subunits of cytochrome oxidase have been found. These mutations should result in truncated subunits, which abolish complex assembly and thereby cause a respiratory defect in the patients. Similarly, the mutation M1T in subunit 2 causes a severe decrease of the level of subunit 2 and a low enzyme content [22]. Several other mutations whose deleterious effects are more difficult to predict have Fig. 5. Proton pumping activity of the R. sphaeroides mutants. Cytochrome c oxidase was reconstituted into asolectin vesicles as previously described [15]. Measurements of the absorbance of phenol red dye (100 l M ) were made in an Olis-RSM stopped-flow spectrophotometer and kinetic traces (average of at least three data sets) were taken at the isosbestic point for horse-heart cytochrome c (557 nm). Wild-type and G78S vesicles were measured using 0.1 l M enzyme and 6.5 l M cytochrome c 2+ , A200T with 0.15 l M enzyme and 5.5 l M cytochrome c 2+ ,andDF94–F98 with 0.08 l M enzyme and 3 l M cytochrome c 2+ , all in 50 l M Hepes/KOH pH 7.4, 45 m M KCl, 44 m M sucrose and 2 l M valinomycin to relieve the membrane potential. The bottom panel in each figure depicts the decrease in absorbance (acidification of the outside) showing the extent of proton pumping. The top panel in each figure is the alkalinization seen in the presence of 5 l M carbonyl cyanide m-chlorophenylhydrazone (CCCP). In the presence of this concentration of uncoupler pumped protons are not observed and the alkalinization is due to the net consumption of protons in the synthesis of H 2 O. In these experiments the H + /e – value for the wild-type oxidase averages 0.9 ± 0.2. The H + /e – values for G78S and A200T are, within error, the same as that of the wild-type enzyme, while the H + /e – value for DF94–F98 is significantly lower. Ó FEBS 2003 Disease-mutations in cytochrome oxidase (Eur. J. Biochem. 270) 1227 been reported: 13 missense mutations, one short in-frame deletion and one stop mutation at the C-terminal of the polypeptide (Table 1). Some of these mutations affect well- conserved regions of the enzyme and have been re-examined using yeast and bacterial mutants. It is interesting to note that in addition to the suspected pathological mutations, over 35 nonpathological amino acid replacements (poly- morphisms) have been observed in humans (www.mito map.org). Subunit 1 L196I has been reported in a patient with epilepsia partilis continua [23]. As the mutation has been transmitted through the germline, and is present at high levels in asymptomatic relatives of the patient, it is likely that the mutation has only very mild effects. In yeast, the same mutation has no effect as shown in this work. A223S has been observed in a family with diverse clinical features ranging from myopathy to a multisystem disorder [24]. However this same change has also been listed as Ôpoly- morphismÕ (www.mitomap.org). In addition, the mutation in yeast has no effect [9], which seems to indicate that the A223S is indeed a silent mutation. Another silent mutation, G317S, has been found in fibroblasts from a patient presenting with lactic acidaemia and cytochrome oxidase deficiency. Residue G317 is a highly conserved residue located next to T316, which is part of the K-channel. Thus, it might have been expected that the replacement of G317 by serine could affect the catalytic activity of the enzyme. However, it was shown that the mutation had no effect on human enzyme and that the disease was caused by a mutation in the nuclear SURF1 gene [25]. Consistent with this, G317S had no effect in yeast [8]. Two other mutations in the region of the K-channel, I280T and M273T have been observed in hematopoietic cells of patients suffering from acquired idiopathic sideroblastic anaemia (AISA). In yeast the mutations caused identical effects to those reported in human cells. They were mildly deleterious, showing a twofold decrease in cytochrome oxidase activity and perturbed binuclear centre properties [8]. These changes are likely due to altered K-channel function as both M273 and I280 are closely associated with two key residues of the K-channel, K319 and T316 [17,18]. The sulfur and carbonyl oxygen of M273 are within 3.5 A ˚ and 3.8 A ˚ , respectively, of the terminal nitrogen of K319, while the sidechain of I280 is within 4.1 A ˚ of the sidechain hydroxyl group of T316. Substitution of threonine at these positions is likely to force some rearrangement of K319 or T316. In addition, I280T will place another hydroxyl group close to that of T316. As these mutations do not eliminate cytochrome oxidase activity in yeast their pathogenic significance is not clear. It is possible, however, that the high energy demands of hematopoietic cells could not be fully met by mitochondria having even mildly decreased respiratory function. Two other mutations have been reported but were not studied in yeast: Y260H and Ter514K,Q,K,Ter. Residue 260 is not conserved between species. A histidine is found in some sequences. The replacement of the stop codon by lysine extends the polypeptide by three residues. This mutation has been reported in patients with LHON [26] but also in a patient suffering from McArdle’s disease, caused by a mutation of the myophosphorylase gene [27]. As the C-terminal end of subunit 1 is not conserved between species (in yeast, for instance, it is 21 amino acids longer) it is difficult to estimate the deleterious impact of this change, if any. Subunit 2 In addition to M1T, two pathological missense mutations located in subunit 2 have been described: M29K and A41T. These two residues are not conserved and were not studied in the yeast model. Subunit 3 Three mutations in subunit 3 have been associated with LHON: G78S [28], A178T [29] and A200T [28]. The pathogenicity of G78S is controversial [30,31], as some reports list the mutation as primary while others suggest that the change is not pathologic but accidentally present in patients. When introduced into yeast or R. sphaeroides this mutation has no effect on respiratory competence or cytochrome oxidase function. Therefore, it seems unlikely that G78S is a primary disease mutation. Residue 78 is located in the helix 3 of subunit 3 at the interface with subunit 1, as discussed above. The introduction of a polar residue can probably be compensated by re-arrangement of the solvent. Likewise, the mutation A200T, located in the upper region of the five-helix bundle, has no effect on oxidase activity or assembly in yeast or R. sphaeroides. Thus, it is also unlikely to be a primary disease mutation. Residue A178 is not conserved between species and is replaced by a tyrosine in yeast. Therefore the mutation was not studied in yeast. F251L has been observed in a patient with MELAS [32]. As described above, on the basis of the structure, it was not expected that the mutation could severely hinder the assembly of the enzyme. Indeed the mutation did not cause any respiratory dysfunction in yeast. No decrease in cytochrome oxidase content could be detected. In contrast, the stop mutation at codon 249 (W249stop), which has been found in a patient with encephalopathy [33], inhibited the respiratory growth of yeast and prevented the assembly of cytochrome oxidase. Similarly, cytochrome oxidase activity was severely decreased in the patient. The deletion of five residues (DF94–F98) in a conserved region of subunit 3, observed in a patient with myoglobinuria [20,34], severely affects oxidase assembly both in human and yeast cells and leads to the loss of subunit 3 in R. sphaero- ides due to instability and protein degradation. In conclusion, on the basis of the yeast and bacterial models, the human mutations could be placed into three classes. (a) Two mutations in subunit 3, DF94–F98, W249stop, are highly deleterious and abolish enzyme assembly. The consequences on energy production by the cells must be dramatic, depending on the load of mutations in the cells. These are clearly pathogenic mutations. (b) Two mutations in subunit 1, I280T and M273T, have a signifi- cant but lesser effect on cytochrome oxidase activity that is likely to partly compromise cellular energy production. Cells with a high demand in energy may be affected by these mutations, leading to disease. (c) Several mutations, L196I, A223S and G317S in subunit 1, and G78S, A200T and 1228 M. Bratton et al.(Eur. J. Biochem. 270) Ó FEBS 2003 F251L in subunit 3, have no effect on respiratory compet- ence in yeast. G78S and A200T also have no effect on cytochrome oxidase activity or assembly in R. sphaeroides. If any of these mutations were the primary source of disease in humans, it would indicate that the human enzyme is more constrained in its structure than its yeast or bacterial counterparts. However, a large number of nonpathogenic residue replacements have already been described, suggest- ing that the human enzyme has significant flexibility in its structure and can accommodate changes. We cannot exclude that the mutations have very mild effects in human cells that are below detection with the yeast or bacterial models. Their pathogenicity is difficult to understand. It may be that some, or all, of this latter group of mutations are also silent in humans and should more properly be listed as polymorphisms. Acknowledgements This work has been supported by a Medical Research Fellowship and a BBSRC grant to BM, and by NIH Grant R01-GM56824 to J.P.H.; D.M is supported by NIH Grant R37-GM26916 to S. Ferguson- Miller. References 1. Bratton, M., Pressler, M. & Hosler, J. (1999) Suicide inactivation of cytochrome c oxidase. Catalytic turnover absence subunit III alters active site. Biochem. 38, 16236–16245. 2. Barrientos, A., Barros, M.H., Valnot, I., Rotig, A., Rustin, P. & Tzagoloff, A. (2002) Cytochrome oxidase in health and disease. Gene 286, 53–63. 3. Zhu,Z.,Yao,J.,Johns,T.,Fu,K.,DeBie,I.,Macmillan,C., Cuthbert, A.P., Newbold, R.F., Wang, J., Chevrette, M., Brown, G.K., Brown, R.M. & Shoubridge, E.A. (1998) SURF1, encoding a factor involved in the biogenesis of cytochrome c oxidase, is mutatedinLeighsyndrome.Nature Genet. 20, 337–343. 4. Tiranti, V., Hoertnagel, K., Carrozzo, R., Galimberti, C., Munaro,M.,Granatiero,M.,Zelante,L.,Gasparini,P.,Marzella,R., Rocchi, M., Bayona-Bafaluy, M.P., Enriquez, J A., Uziel, G., Bertini, E., Dionisi-Vici, C., Franco, B., Meitinger, T. & Zeviani, M. (1998) Mutations of Surf-1 in Leigh disease associated with cytochrome c oxidase deficiency. Am.J.Hum.Genet.63, 1609– 1621. 5. Valnot, I., von Kleist-Retzow, J C., Barrientos, A., Gorbatyuk, M., Taanman, J.W., Mehaye, B., Rustin, P., Tzagoloff, A., Munich, A. & Rotig, A. (2000) A mutation in the human heme A: farnesyltransferase gene (COX10) causes cytochrome c oxidase deficiency. Hum. Mol. Genet. 9, 1245–1249. 6. Valnot, I., Osmond, S., Gigarel, N., Mehaye, B., Amiel, J., Cormier-Daire, V., Munich, A., Bonnefont, J P., Rustin, P. & Rotig, A. (2000) Mutations of the SCO1 gene in mitochondrial cytochrome oxidase (COX) deficiecy with neonatal-onset hepatic failure and encephalopathy. Am.J.Hum.Genet.67, 1104–1109. 7. Papadopoulou, L.C., Sue, C.M., Davidson, M.M., Tanji, K., Nishino, I., Sadlock, J.E., Krishna, S., Walker, W., Selby, J., Glerum,D.M.,VanCoster,R.,Lyon,G.,Scalais,E.,Lebel,R., Kaplan, P., Shanske, S., De Vivo, D.C., Bonilla, E., Hirano, M., DiMauro, S. & Schon, E.A. (1999) Fatal infantile cardioence- phalomyopathy with COX deficiency and mutations in SCO2, a COX assembly gene. Nature Genet. 23, 333–337. 8. Meunier, B. (2001) Site-direct mutations in the mitochondrially- encoded subunits I and III of yeast cytochrome oxidase. Biochem. J. 354, 407–412. 9. Meunier, B. & Taanman, J W. (2002) Mutations of cytochrome c oxidase subunits 1 and 3 in Saccharomyces cerevisiae: assembly defect and compensation. Biochim. Biophys. Acta 1554, 101–107. 10. Brown, S., Colson, A M., Meunier, B. & Rich, P.R. (1993) Rapid screening of cytochromes of respiratory mutants of Saccharo- myces cerevisiae – application to the selection of strains containing novel forms of cytochrome c oxidase. Eur. J. Biochem. 213, 137–145. 11. Hiser, L. & Hosler, J.P. (2001) Heme A is not essential for assembly of the subunits of cytochrome c oxidase in Rhodobacter sphaeroides. J. Biol. Chem. 276, 45403–45407. 12. Keen, N.T., Tamaki, S., Kobayashi, D. & Trollinger, D. (1988) Improved broad-host-range plasmids for DNA cloning in Gram- negative bacteria. Gene 70, 191–197. 13. Zhen, Y., Follmann, K., Qian, J., Hayward, T., Nilsson, T., Dahn, M., Hamer, A.G., Hosler, J.P. & Ferguson-Miller, S. (1998) Over- expression and purification of cytochrome c oxidase from Rhodo- bacter sphaeroides. Protein Expression Purification 13, 326–336. 14. Donohue, T.J. & Kaplan, S. (1991) Genetic techniques in Rhodo- spirillaceae. Meth. Enzymol. 204, 459–485. 15. Mills, D.A., Schmidt, B., Hiser, C., Westley, E. & Ferguson- Miller, S. (2002) Membrane potential-controlled inhibition of cytochrome c oxidase by zinc. J. Biol. Chem. 277, 14894–14901. 16. Hosler, J.P., Fetter, J., Tecklenburg, M.M.J., Espe, M., Lerma, C. & Ferguson-Miller, S. (1992) Cytochrome aa 3 of Rhodobacter sphaeroides as a model for mitochondrial cytochrome c oxidase. Purification, kinetics, proton pumping, and spectral analysis. J. Biol. Chem. 267, 24264–24272. 17. Tsukihara, T., Aoyama, H., Yamashita, E., Tomizaki, T., Yamaguchi,H.,Shinzawa-Itoh,K.,Nakashima,R.,Yaono,R.& Yoshikawa, S. (1996) The whole structure of the 13-subunit oxi- dized cytochrome c oxidase at 2.8 A ˚ . Science 272, 1136–1144. 18. Svensson-Ek, M., Abramson, J., Larsson, G., Tornroth, S., Brzezinski, P. & Iwata, S. (2002) The X-ray crystal structure of wild type and EQ (I-286) mutant cytochrome c oxidase from Rhodobacter sphaeroides. J. Membr. Biol. 321, 329–339. 19. Bratton, M.R., Hiser, L., Antholine, W.E., Hoganson, C. & Hosler, J. (2000) Identification of the structural subunits required for formation of the metal centers in subunit I of cytochrome c oxi- dase of Rhodobacter sphaeroides. Biochemistry 39, 12989–12995. 20. Hoffbuhr, K.C., Davidson, E., Filiano, B.A., Davidson, M., Kennaway, N.G. & King, M.P. (2000) A pathogenic 15-base pair deletion in mitochondrial DNA-encoded cytochrome c oxidase subunit III results in the absence of functional cytochrome c oxidase. J. Biol. Chem. 275, 13994–14003. 21. Mather, M.W. & Rottenberg, H. (1998) Intrinsic uncoupling of cytochrome c oxidase may cause the maternally inherited mito- chondrial diseases MELAS and LHON. FEBS Lett. 433, 93–97. 22. Clark, K.M., Taylor, R.W., Johnson, M.A., Chinnery, P.F., Chrzanowska-Lightowlers, Z., Andrews, R.M., Nelson, I.P., Wood, N.W., Lamont, P.J., Hanna, M.G., Lightowlers, R.N. & Turnbull, D.M. (1999) An mtDNA mutation in the initiation codon of the cytochrome c oxidase subunit II gene results in lower levels of the protein and a mitochondrial encephalomyopathy. Am.J.Hum.Genet.64, 1330–1339. 23. Varlamov, D.A., Kudin, A.P., Vielhaber, S., Schro ¨ der, R., Sassen, R., Becker, A., Kunz, D., Haug, K., Rebstock, J., Heils, A., Elger, C.E. & Kunz, W.S. (2002) Metabolic consequences of a novel missense mutation of the mtDNA COI gene. Hum. Mol. Genet. 11, 1–9. 24. Akman, C.I., Sue, C.M., Musumaci, O., Tanji, K., Sims, K., Rowland, L.P., Bonilla, E. & DiMauro, S. (2000) Maternally inherited mitochondrial encephalomyopathy due to a novel missense mutation in the mtDNA cytochrome c oxidase. Ann. Neurol. 48, S46, A69. Ó FEBS 2003 Disease-mutations in cytochrome oxidase (Eur. J. Biochem. 270) 1229 25. Williams, S.L., Scholte, H.R., Gray, R.G.F., Leonard, J.V., Schapira, A.H.V. & Taanman, J W. (2001) Immunoligical phe- notyping of fibroblast cultures from patients with a mitochondrial respiratory chain deficit. Laboratory Invest. 81, 1069–1077. 26. Brown, M.D., Yang, C C., Trounce, I., Torroni, A., Lott, M.T. & Wallace, D.C. (1992) A mitochondrial DNA variant, identified in Leber Hereditary Optic Neuropathy patients, which extends the amino acid sequence of cytochrome c oxidase subunit I. Am. J. Hum. Genet. 51, 378–385. 27.Aguilera,I.,Garcia-Lozano,J.R.,Munoz,A.,Arenas,J., Campos, Y., Chinchon, I., Roldan, A.N. & Bautista, J. (2001) Mitochondrial DNA point mutation in COI gene in a patient with McArdle’s disease. J. Neurol. Sci. 192, 81–84. 28. Johns, D.R. & Neufeld, M.J. (1993) Cytochrome c oxidase mutations in Leber hereditary optic neuropathy. Biochem. Bio- phys. Res. Commun. 196, 810–815. 29. Johns,D.R.,Neufeld,M.J.&Hedges,T.R.(1994)Mitochondrial DNA mutations in Cuban optic and peripheral neuropathy. J. Neuro-Ophtalmology 14, 135–140. 30. Brown, M.D., Torroni, A., Huoponen, K., Chen, Y S., Lott, M.T. & Wallace, D.C. (1994) Pathological significance of the mtDNA COXIII mutation at nucleotide pair 9438 in Leber Hereditary Optical Neuropathy. Am.J.Hum.Genet.55,410. 31. Oostra, R.J., Van den Bogert, C., Nijtmans, L.G.J., Van Galen, M.J.M., Zwart, R., Bolhuis, P.A. & Bleeker-Wagemakers, E.M. (1995) Simultaneous occurence of the 11778 (ND4) and the 9438 (COXIII) mtDNA mutations in Leber Hereditory Optic Neuro- pathy: Molecular, biochemical and clinical findings. Am. J. Hum. Genet. 57, 954–957. 32. Manfredi, G., Schon, E.A., Moroaes, C.T., Bonilla, E., Berry, G.T., Sladky, J.T. & DiMauro, S. (1995) A new mutation asso- ciated with MELAS is located in a mitochondrial DNA poly- peptide-coding gene. Neuromusc. Disord. 5, 391–398. 33. Hanna, M.G., Nelson, I.P., Rahman, S., Lane, R.J.M., Heales, S., Cooper, M.J., Shapira, A.H.V., Morgan-Hughes, J.A. & Wood, N.W. (1998) Cytochome c oxidase deficiency associated with the first stop-codon point mutation in human mtDNA. Am.J.Hum. Genet. 63, 29–36. 34. Keightley, J.A., Hoffbuhr, K.C., Burton, M.D., Salas, V.M., Johnston, W.S.W., Penn, A.M.W., Buist, N.R.M. & Kennaway, N.G. (1996) A microdeletion in cytochrome c oxidase (COX) subunit III associated with COX deficiency and recurrent myo- globinuria. Nature Genet. 12, 410–416. 35. De Vries, D., de Coo, R., Buddiger, P., Ruitenbeek, W., Albrecht, R. & Van Oost, B. (1994) Isolated respiratory chain enzyme deficiency in patients with a mitochondrial (encephalo-) myo- pathy: sequence analysis of the mitochondrial complex I and IV genes. Am.J.Hum.Genet.55, 2099. 36. Bro ¨ ker, S., Meunier, B., Rich, P.R., Gattermann, N. & Hofhaus, G. (1998) MtDNA mutations associated with sideroblastic anae- mia cause a defect of mitochondrial cytochrome c oxidase. Eur. J. Biochem. 258, 132–138. 37. Rahman, S., Taanman, J W., Cooper, J.M., Nelson, I., Harg- reaves, I., Meunier, B., Hanna, M.G., Garcia, J., Capaldi, R.A., Lake, B.D., Leonard, J.V. & Shapira, A.H.V. (1999) A mis-sense mutation of cytochrome oxidase subunit II causes defective assembly and myopathy. Am.J.Hum.Genet.65, 1030–1039. 38. Finnila, S., Uusimaa, J., Vainionpa ¨ a ¨ ,L.,Herva,R.,Rantala,H., Hassinen, I.E. & Majamaa, K. (2001) A mutation in mitochon- drial DNA encoded cytochrome c oxidaseIIgeneinachildwith Alpers–Huttenlocher–like syndrome. Am.J.Hum.Genet.69, 1803. 1230 M. Bratton et al.(Eur. J. Biochem. 270) Ó FEBS 2003 . Rhodobacter sphaeroides and Paracoccus deni- trificans. Subunit 2 forms part of the docking site for cytochrome c and binds Cu A , the first electron acceptor. Subunit. Disease-related mutations in cytochrome c oxidase studied in yeast and bacterial models Melyssa Bratton 1 , Denize Mills 2 , C. Kate Castleden 3 ,