Cytochrome c oxidase deficiency due to mutations ... - Semantic Scholar

© 2001 Oxford University Press

Human Molecular Genetics, 2001, Vol. 10, No. 26 3025–3035

Cytochrome c oxidase deficiency due to mutations in SCO2, encoding a mitochondrial copper-binding protein, is rescued by copper in human myoblasts Michaela Jaksch1,2,*, Claudia Paret3, Rolf Stucka4, Nina Horn5, Josef Müller-Höcker6, Rita Horvath1,2, Nadine Trepesch1,2, Gerhard Stecker1,2, Peter Freisinger1,7, Christian Thirion4, Juliane Müller4, Renate Lunkwitz8, Gerhard Rödel3, Eric A. Shoubridge9 and Hanns Lochmüller4 1Metabolic

Disease Centre Munich-Schwabing and 2Institute of Clinical Chemistry, Molecular Diagnostics and Mitochondrial Genetics, Koelner Platz 1, 80804 Munich, Germany, 3Institute of Genetics, TU Dresden, Germany, 4Friedrich-Baur-Institute, Department of Neurology and Gene Center, Ludwig-Maximilians-University, Munich, Germany, 5JFK Institute, Glostrup, Denmark, 6Institute of Pathology, LMU Munich, Germany, 7Childrens Hospital, TU, Munich, Germany, 8Institute of Analytical Chemistry, TU Dresden, Germany and 9Montreal Neurological Institute and Department of Human Genetics, McGill University, Montreal, Quebec, Canada Received August 22, 2001; Revised and Accepted October 23, 2001

Mutations in SCO2, a cytochrome c oxidase (COX) assembly gene, have been reported in nine infants with early onset fatal cardioencephalomyopathy and a severe COX deficiency in striated muscle. Studies on a yeast homolog have suggested that human Sco2 acts as a copper chaperone, transporting copper to the CuA site on the Cox II subunit, but the mechanism of action remains unclear. To investigate the molecular basis of pathogenesis of Sco2 defects in humans we performed genetic and biochemical studies on tissues, myoblasts and fibroblasts from affected patients, as well as on a recombinant human C-terminal Sco2 segment (22 kDa), bearing the putative CxxxC metal-binding motif. Recombinant Sco2 was shown to bind copper with a 1:1 stoichiometry and to form homomeric complexes in vitro, independent of the metal-binding motif. Immunohistochemistry using antibodies directed against different COX subunits showed a marked tissue-specific decrease in the Cox II/III subunits that form part of the catalytic core, consistent with the differential tissue involvement, but a more uniform distribution of Cox Vab, a nuclearencoded subunit. Sco2 was severely reduced in patient fibroblasts and myoblasts by immunoblot analysis. Patient fibroblasts showed increased 64Cu uptake but normal retention values and, consistent with this, the copper concentration was four times higher in Sco2-deficient myoblasts than in controls. COX activity in patient myoblasts was completely rescued by transduction with a retroviral vector

expressing the human SCO2 coding sequence, and more interestingly by addition of copper–histidine (300 µM) to the culture medium. Whether the latter is accomplished by the very low residual levels of Sco2 in the patient cells, direct addition of copper to the CuA site, or by another copper-binding protein remains unknown. Whatever the mechanism, this result suggests a possible therapy for the early treatment of this fatal infantile disease. INTRODUCTION Mammalian cytochrome c oxidase (COX) is a multimeric enzyme of the inner mitochondrial membrane whose subunits are encoded by both the mitochondrial and nuclear genomes. The enzyme catalyzes the reduction of molecular oxygen by reduced cytochrome c, the terminal step in the respiratory chain. Isolated COX deficiency is found in neonatal, infantile and late onset diseases that can present with a wide range of multisystemic, neurological and muscular symptoms (1). Lethal neonatal and infantile respiratory chain deficiency disorders are thought to occur in approximately one in 10 000 newborns, and COX deficiency is one of the most frequent biochemical diagnoses. In isolated COX deficiency, several pathogenic mutations have been reported in the three mtDNAencoded subunits that make up the catalytic core of the enzyme of the complex, but mutations have not been found in the remaining 10 nuclear encoded structural subunits so far (2,3). In contrast, mutations in the nuclear genes encoding accessory factors involved in the assembly of the holoenzyme COX have been found in association with several severe autosomalrecessive infantile COX deficiency disorders: SURF1 (Leigh

*To whom correspondence should be addressed at: Metabolic Disease Center Munich-Schwabing, Koelner Platz 1, 80804 Munich, Germany. Tel: +49 89 3068 2670; Fax: +49 89 3068 3911; Email: [email protected]

3026 Human Molecular Genetics, 2001, Vol. 10, No. 26

syndrome), SCO2 [fatal hypertrophic cardiomyopathy (HCMP) with encephalopathy], COX10 (hepatic failure) and SCO1 (tubulopathy with encephalomyopathy) (4–9). Human SCO2 (GenBank accession no. AF177385), which maps to chromosome 22, has so far been specifically associated with a fatal infantile COX deficiency disorder, the predominant symptoms being muscular hypotonia, HCMP and encephalopathy (6,7). The gene encodes a mitochondrial protein (10) that is predicted to contain a mitochondrial targeting sequence, a single transmembrane domain and a highly conserved CxxxC putative copper-binding motif. Pathogenic mutations in SCO2 have been described in a total of nine patients and all carry a G1541A mutation near the predicted CxxxC binding motif (E140K). The other allele in the most severe cases (early onset, fatal) is a stop or missense mutation (6,7). Patients who are homozygous for the E140K mutation have a comparatively mild phenotype (delayed onset, less progressive but still fatal) (11). Consistent with this, the common G1541A mutation did not result in a measurable phenotype when engineered in the analogous position of yeast Sco1p (12), which likely represents the functional homolog of human Sco2. Mutations in human SCO1 have so far been described in a single pedigree with a progressive neonatal disorder, but predominantly affecting the liver and not the heart (9). The basis for the tissue specificity resulting from SCO2 or SCO1 mutations is unknown. The function of the two human Sco proteins as mitochondrial chaperones has been partly inferred from yeast, in which two homologous proteins, Sco1p (13) and Sco2p (14), were identified as putative mitochondrial copper transporters (14). The current yeast model of mitochondrial copper metabolism suggests that copper enters the cytoplasm by a high-affinity transporter Ctr1p (15), binds to Cox17p, a small specific copper chaperone that can probably enter into the mitochondrial intermembrane space (16), and then transfer copper to one or both of the Sco proteins. The Sco proteins are thought to chaperone copper to the catalytic CuA site on the mitochondrially encoded Cox 2p subunit (12,17). Studies in a bacterial system suggest that addition of copper to the CuB site on Cox 1p may be mediated by Cox11p (18). The exact function of the two Sco proteins remains unknown, and in particular it is not known why two different, but closely related, proteins are necessary for mitochondrial copper delivery. Although both Sco1p and Sco2p can suppress a COX17 deletion in yeast, deletion of SCO1, but not SCO2, produces a pet phenotype. Deletion of yeast SCO1 cannot be rescued by increased medium copper or by overexpression of yeast Cox17p, suggesting that yeast Sco1p is an essential and specific chaperone for the delivery of copper to COX acting downstream of Cox17p (12,14). In support of this, a direct interaction of yeast Sco1p and yeast Cox2p has recently been demonstrated (19). However, specific copper binding or transport have not been reported for either of the Sco proteins in yeast or human and, it has been suggested, based on their sequence similarity to thiol:disulfide oxidoreductases, that the Sco proteins may have a catalytic rather than a copper transport function (20). To investigate the molecular basis of pathogenesis of Sco2 defects in humans we performed genetic and biochemical studies on tissues, myoblasts and fibroblasts from affected patients, as well as on a recombinant human C-terminal Sco2 segment (22 kDa), bearing the putative CxxxC metal-binding motif.

RESULTS SCO2 mutations result in tissue-specific COX assembly defects In order to investigate the basis for the tissue-specific effects of SCO2 mutations on COX assembly, we examined the steadystate levels of COX subunits in liver, kidney, heart and skeletal muscle of a patient with SCO2 mutations (E140K/R90X) by immunohistochemistry using polyclonal antibodies directed against the mitochondrially encoded subunits II/III, and the nuclearly encoded subunit Vab. Immunoreactivity for subunits II/III was similar to control in liver and kidney (Fig. 1G and H), but was virtually absent in heart and skeletal muscle (Fig. 1A and D) compared with age-matched controls (Fig. 1C and F). In contrast, immunoreactivity for subunit Vab was similar in all tissues examined in both patients and controls (Fig. 1B and E). Human Sco2 is a copper-binding protein that can oligomerize in vitro and it is not active as a thiol:disulfide oxidoreductase To characterize the structural and functional properties of human Sco2, we constructed plasmids coding for two different C-terminal Sco2 segments of 22 kDa: a wild-type protein (Sco2) and a mutant form lacking the CxxxC metal-binding site (Sco2∆cys) (Fig. 2A). To test whether Sco2 binds copper in vivo, the C-terminal portions of Sco2 and Sco2∆cys expressed in Escherichia coli grown in a medium containing 0.5 mM copper were affinity purified, and bound copper was quantified using atomic absorption spectrometry (AAS) or the more specific ion-coupled plasma atomic emission spectrometry (ICP–AES). The stoichiometry of copper binding was repeatedly shown to be in the range between 1 ± 0.2 µmol copper/µmol protein for Sco2 and 0.2 ± 0.05 µmol copper/ µmol protein for Sco2∆cys, demonstrating specific copper binding dependent on the CxxxC motif. To test whether Sco2 has the ability to form higher order structures, as has been suggested from the extent of the COX deficiency in patients with different combinations of mutant alleles (11), we investigated the ability of the C-terminal fragment of Sco2 to form homomeric complexes by affinity chromatography. Sco2–EGFP bound to glutathione S-transferase (GST)–hSco2-coupled Sepharose, but not to GST–Sepharose alone (Fig. 2B). In a further control, Sepharose-bound GST–Sco2 was incubated with cell lysates containing EGFP and no binding was observed (data not shown). To test whether the CxxxC motif is required for homomerization, we repeated the experiments with the Sco2∆cys mutant protein. Wild-type Sco2–GST was coupled to glutathione (GSH)–Sepharose and tested for binding of the mutant protein fused to EGFP. The mutant protein interacted with Sco2 in a manner similar to the wild-type protein (Fig. 2B). An alternative function for the Sco proteins was recently suggested based on sequence similarity to thiol:disulfide oxidoreductases (20). To test for this activity, the C-terminal recombinant Sco2, which contains the key conserved residues between the Sco proteins and the thiol:disulfide oxidoreductases, was assayed for its ability to reduce disulfide bridges in insulin in the presence of NADPH and human thioredoxin reductase, a classical thioredoxin assay. The

Human Molecular Genetics, 2001, Vol. 10, No. 26 3027

Figure 1. Steady-state levels of COX subunits in heart, skeletal muscle, liver and kidney of a patient with SCO2 mutations (140K/R90X). (A–C) Heart muscle. (A) Subunit II/III. Absence of subunit II/III in heart muscle tissue of a SCO2 patient. (B) Subunit Vab shows a normal distribution in the same patient. (C) Normal expression of subunit II/III in an age-matched control heart. (D–F) Skeletal muscle. (D) Absence of subunit II/III in the extrafusal fibres of the patient, the muscle spindle fibres being spared. (E) Normal expression of subunit Vab in skeletal muscle of the same patient. (F) Normal expression of subunit II/III in an age-matched control muscle. (G and H) Normal immunoreactivity of subunit II/III in kidney (G) and liver (H) of the patient. Magnification (A, B, D, E, G and H) 640× and (C and F) 960×.

results clearly show that this portion of Sco2 is not active as a disulfide reductase (Fig. 2C). SCO2 mutations cause a severe reduction of Sco2 protein and retrovirally mediated transfer of the SCO2 coding sequence into myoblasts restores COX activity On immunoblotting, primary human myoblasts and fibroblasts from healthy controls showed a specific band for Sco2 at the expected molecular weight of the processed protein (∼25 kDa) in mitochondrially enriched fractions. Primary fibroblasts and myoblasts from HCMP patients with previously described SCO2 mutations showed a severe reduction of Sco2 protein (Fig. 3). Extended exposure of the blots revealed weak signals at the expected molecular weight that may correspond to lowlevel residual Sco2 protein in mitochondria (data not shown). In contrast, fibroblasts from a patient with Leigh syndrome due to a SURF1 mutation, as well as myoblasts from a patient with infantile HCMP and a normal SCO2 sequence showed normal levels of Sco2 protein (Fig. 3A). Retrovirally mediated gene transfer of the SCO2 coding sequence into myoblasts from a

SCO2 patient resulted in overexpression of normal Sco2 protein (Fig. 3), and the restoration of COX activity to normal levels (Table 1). SCO2 mutations cause abnormal 64Cu uptake in fibroblasts and increased basal copper concentrations in myoblasts To test for possible effects of SCO2 mutations on cellular copper handling, we measured 64Cu uptake (ng/mg protein/20 h) and retention in fibroblasts from three unrelated patients with different SCO2 mutations. 64Cu uptake was significantly increased in repeated experiments in patients A, 1A and 2B1 (Fig. 4). The highest uptake was found in patient 2B1 (E140K/R90X), who had the most severe form of the disease with death at age 4 weeks, followed by patient 1A (E140K/R171W) with death at age 4 months and patient A (140K/140K) with death at age 13 months. The retention values, which are increased in Menkes patients with mutations in ATP7A, encoding the α subunit of a Cu(2+) transporting ATPase (21–23), were normal in all three patients.


Figure 2. Characterization of the human C-terminal portion of wild-type and mutant Sco2. (A) C-terminal portion of Sco2 (a) and of Sco2∆cys (b). Lane 1, supernatant after cell lysis prior to incubation with GSH–Sepharose; lane 2, GSH–Sepharose-bound GST-fusion protein; lane 3, purified protein after cleavage with thrombin. (B) GST–Sco2 was bound to Sepharose beads and incubated with lysate of HeLa cells expressing Sco2–EGFP (lane 1) or Sco2 ∆cys–EGFP (lane 2). In lane 3, GST was bound to Sepharose beads and incubated with lysate of HeLa cells expressing Sco2–EGFP. (C) Human thioredoxin (hThio) and Sco2 were tested for their ability to reduce insulin. Formation of insoluble insulin B chain was followed at 595 nm for 120 min.

We directly investigated intracellular copper concentrations in myoblasts from a patient with SCO2 mutations (E140K/R90X) and from a control after supplementation of the growth medium of myoblasts for 3 weeks. Basal copper values, measured by ICP–AES, were four times higher in patient versus control myoblasts (46.3 versus 10.3 ng/mg protein). The copper concentration increased with increasing copper–histidine (Cu–His) in the medium to 219.4 ng/mg protein in the patient cells and to 219.5 ng/mg protein in the control at highest concentrations (300 µM) of Cu–His (Table 2). Cu–His supplementation of SCO2 myoblasts (E140K/ R90X) rescues COX activity To test whether increased intracellular copper concentrations could rescue the COX deficiency, SCO2-mutant myoblasts (E140K/R90X) were incubated with Cu–His at different concentrations. COX activity was determined spectrophotometrically (Fig. 5A) and histochemically (data not shown). Increasing concentrations of Cu–His resulted in increased COX activity, and a complete rescue of COX activity, comparable to that in wild-type control myoblasts was obtained at 300 µM (Fig. 5A). Paralleling the increase in COX activity, increased levels of COX subunit II were detected with increasing copper concentrations on immunoblotting

(Fig. 5B). Control myoblasts do not show any significant change in COX activity in response to increased intracellular copper (Fig. 5A). Real time PCR shows stable transcriptional levels of COX17 and SCO1 after Cu–His supplementation in SCO2-mutant myoblasts To test whether the rescue of COX activity by increased copper might involve transcriptional upregulation of other components of the mitochondrial copper delivery pathway, expression of SCO1 and COX17 mRNA was quantified in patient myoblasts (E140K/R90X) supplemented with increasing concentrations of Cu–His (0, 75, 150 and 300 µM) using real-time RT–PCR. SCO1 and COX17 expression was standardized to β-actin (Table 3). There was no significant difference in SCO1 and COX17 mRNA expression in treated versus untreated myoblasts (E140K/R90X). Similar values were obtained for control myoblasts (data not shown). DISCUSSION Mutations in human SCO2 result in a distinct fatal infantile disorder with severe COX deficiency, neurodegeneration, muscular hypotonia and HCMP leading to early cardiac failure


Figure 3. Deficiency of Sco2 protein in fibroblasts and myoblasts of patients with SCO2 mutations. Immunoblotting of fibroblasts (A) and myoblasts (B) reveals absence or severe reduction of a Sco2-specific band at 25 kDa [(A) lane 4, E140K homozygous; lane 5, E140K homozygous; lane 7, E140K/R171W; (B) lane 4, E140K/R90X]. Sco2 is readily detected in patient cells (E140K/R90X) after retroviral gene transfer of SCO2 [(A and B) lane 2], in normal fibroblasts [(A) lane 3], normal myoblasts [(B) lane 3], and fibroblasts of a Leigh syndrome patient with SURF-1 mutations [(A) lane 6]. As a positive control 1 pg of a 22 kDa bacterially expressed C-terminal Sco2 fragment is used [(A and B) lane 1]. A minor, unspecific signal of ~22 kDa is seen in mitochondrial extracts of both patient and normal fibroblasts [(A) lanes 3, 4, 6, 7].

(6,7,11). Using a polyclonal antibody against human Sco2, we demonstrate severely decreased levels of Sco2 protein in fibroblasts and myoblasts from patients with different SCO2 mutations suggesting an unstable gene product. To directly prove the pathogenetic role of SCO2 mutations in human cells, we transduced a patient myoblast cell line (E140K/R90X), with a retroviral vector expressing the SCO2 coding sequence. The very low residual COX activity in the patients cells was completely rescued by the retroviral transduction, formally proving the causal nature of the mutation. Studies in yeast have identified homologs of the human Sco proteins; however, the different roles of Sco1 and Sco2 in mitochondrial copper metabolism remain unclear and direct evidence for binding to the CxxxC metal-binding site has not been demonstrated. To further elucidate the role of human Sco2 in copper metabolism and transport, we investigated a human recombinant C-terminal segment of the protein, containing the CxxxC motif.

We demonstrate that Sco2 binds copper with a 1:1 stoichiometry using two independent detection methods, and show that the cysteine residues of the conserved CxxxC motif are essential for this function. This is the first direct demonstration of copper binding of a member of the Sco protein family. Recently, Chinenov (20) reported the similarity of proteins of the Sco1p family to thiol:disulfide oxidoreductases. The overall similarity is low, but cysteine residues and hydrophobic amino acids in the active center are conserved, and the predicted secondary structure of yeast Sco1p is similar to that of peroxiredoxins. Moreover, a conserved histidine was proposed to activate a sulfhydryl group of the active center and indeed, yeast Sco1p carrying a mutation of this histidine was shown to be inactive (19). Bacterial thiol:disulfide oxidoreductases are involved in the attachment of heme and copper, respectively, to cytochrome c and COX by maintaining cysteine residues in a reduced state (24). Sco proteins could be required to maintain the cysteine residues of the CxxxC motif of Cox II in a reduced state to allow the insertion of copper. Using the purified recombinant C-terminus of Sco2, containing the conserved active center, we find no evidence for thioredoxin activity using a classical thioredoxin assay. These data suggest that Sco2 is functional as a copper chaperone. Using affinity chromatography we also show that the C-terminal portion of Sco2 can form homomeric complexes. We cannot rule out the possibility that Sco2 complex formation is mediated by other proteins present in the lysate; however, the interaction, whether direct or indirect, does not require the CxxxC motif, as shown by the formation of homomeric complexes by the Sco2∆cys mutant protein. Homo- and heterodimerization of copper-binding proteins have been reported between Cu–Zn superoxide dismutase (SOD) and its chaperone CCS (25). Interestingly, CCS has a high degree of similarity to SOD, reminiscent of the situation with Sco1 and Sco2. It will be interesting to test whether Sco2 can also form heteromeric complexes with Sco1. Our finding of oligomerization of Sco2 molecules supports our recent hypothesis on the effect of different SCO2 mutations on COX activity in muscle (11). All patients so far identified have at least one E140K allele and COX activity is most severely reduced when the other allele is a different missense mutation, and to a lesser extent when it is a stop mutation. This finding could be explained if the protein forms dimers or higher order oligomers (11). Patients homozygous for the E140K allele have the mildest COX deficiency, presumably because Sco2 in these patients has the highest residual copper-binding potential. As Sco2 protein levels appear very low on immunoblots, we

Table 1. COX activity after wild-type SCO2 cDNA transfection and overexpression in Sco2-deficient myoblasts E140K/R90X COX (U/g protein) Untransfected Transfected

E140K/R90X CS (U/g protein)

E140K/R90X normalized ratio: U COX/U CS

Controls (n = 10) normalized ratio: U COX/U CS

0, 5

33, 30

0, 0.05

0.7–2.7

41, 94

20, 32

2, 2.8

–

Control values are calculated from 10 controls (duplicates) and are given as 95% confidence interval. COX activity in myoblasts: activity is expressed as units (µmol/min) and grams of total protein (range 43–190 U/g total protein). CS activity range 20–106 U/g total protein. As an additional value the normalized ratio (COX/CS) is given.


Figure 4. 64Cu uptake and retention in fibroblasts from controls, patients with Menkes disease and from patients with SCO2 mutations. 64Cu uptake and retention per milligram protein per 20 h are shown: triangles, patients with SCO2 mutations (patient A, E140K/E140K; patient 2B1, E140K/R90X; patient 1A, E140K/R171W); boxes, controls; diamonds, Menkes patients. 64Cu uptake was significantly increased in repeated experiments in: patient A, 45.0, 35.3; patient 2B1, 61.1, 59.6; patient 1A, 37.5, 45.6. Menkes patients (n = 47): median, 82.6; range, 42.6–127.6 ng. Controls (n = 18): median, 19.9; range, 8.2–30.8. The retention value was completely normal in: patient A, 26.1%; patient 2B1, 25.3%, 18.4%; patient 1A, 18.9%, 18.4%. Menkes patients: median, 65.8%; range, 40.3–94.5%. Controls: median, 16.3%; range, 9.7–22.7%.

Table 2. Copper concentration in myoblasts (ICP–AES) Control Basal [Cu]/ng/mg protein (no copper added to the medium) [Cu] medium: 75 µM (5 µg/ml)

(E140K/R90X)

10.3 ± 0.7

46.3 ± 2.3

27.8 ± 1.8

87.1 ± 3.7

[Cu] medium: 150 µM (10 µ g/ml)

105.4 ± 5.2

101 ± 5.7

[Cu] medium: 300 µM (20 µ g/ml)

219.4 ± 10.4

219.5 ± 11.2

Values (quadruplicates) [Cu]/ng/mg protein are given as mean ± SD.

have not been able to detect any differences in the steady-state levels of the protein with any of the different mutations. Skeletal and cardiac muscle of patients with SCO2 mutations show a severe decline in COX activity whereas fibroblasts exhibit a relatively high residual enzyme activity, despite the fact that SCO2 is ubiquitously expressed (6). To further characterize these tissue-specific effects, we investigated different tissues from a patient with SCO2 mutations (E140K/R90X) and found a clear absence of the mitochondrially encoded COX subunits II/III in heart and skeletal muscle, but a normal concentration of the nuclear encoded subunit Vab. In contrast, liver and kidney exhibited normal concentrations for all three subunits. These effects might reflect the high postnatal energy demand in human skeletal muscle, heart and other organs like CNS, the low levels of residual Sco2 being insufficient to provide mitochondrial copper for COX assembly. This would also explain the unremarkable fetal development and birth in all index patients so far described. In support of this hypothesis, it has been shown in normal fetal and adult organs, that heart and skeletal muscle COX transcripts increase between 2- and 20-fold from the fetal to the adult state, whereas in liver little changes are observed (26). Mutations in SURF1, another COX assembly

gene, produce systemic deficiencies in COX activities (15–25% residual activities), but the majority of cases are associated with Leigh syndrome, which rarely includes cardiac decompensation. Although this difference might be explained on the basis of a threshold for residual COX activity in the heart, which is never transgressed in the heart of SURF1 patients, it cannot explain why patients homozygous for the E140K mutation still develop cardiomyopathy with relatively high levels of COX activity (20–50% residual activities) (11). It remains possible that abnormal copper handling in the cells of SCO2 patients has additional pathological effects, other than just the impairment of COX activity. However, taking together all that we know about deficient assembly factors of COX (Surf1, Cox10, Sco1, Sco2) the high energy demand per se cannot explain the variable clinical involvement in the other instances. The genotype–phenotype correlation observed in patients carrying different combinations of SCO2 mutations (11) is also reflected by 64Cu uptake measurements in skin fibroblasts derived from three affected patients, which revealed higher uptake values in more severe forms (E140K/R90X and E140K/R173W) and a lower, but still significantly increased uptake, in the milder phenotype (E140K/E140K) (11). In all three cell lines the


Figure 5. Rescue of COX activity and increasing Cox II protein by Cu–His supplementation of SCO2-deficient myoblasts. (A) The increase in COX activity (black columns) after Cu–His supplementation to the medium is shown in relation to the activity of CS (white columns). Two independent assays are shown. Control myoblasts do not show significant changes in COX (black striped columns) or CS activity (gray columns). (B) Immunoblotting of myoblasts (E140K/R90X) reveals severe reduction of a Cox II-specific band at 24 kDa (lane 2, 25 µg mitochondrial extract). Cox II is readily detected in the patient’s myoblasts after retroviral gene transfer of SCO2 (lane 1, 10 µg mitochondrial extract). An increase of Cox II in the patient’s myoblasts is also seen with increasing doses of Cu–His supplementation (lane 3, 75 µM; lane 4, 150 µM; lane 5, 300 µM; 25 µg mitochondrial extract each).

Table 3. Real-time PCR on SCO1 and COX17 transcripts after copper supplementation Copper

0 µM

75 µM

150 µM

300 µM

SCO1

31.2 ± 0.2

27.8 ± 0.1

29.1 ± 0.1

30.7 ± 0.2

ACTB

25.3 ± 0.5

21.8 ± 0.3

23.1 ± 0.1

24.6 ± 0.1

SCO1/ACTB

1.23 ± 0.03

1.28 ± 0.02

1.26 ± 0.01

1.25 ± 0.01

COX17

35.5 ± 0.1

32.2 ± 0.1

33.2 ± 0.2

34.6 ± 1.6

ACTB

24.8 ± 0.2

21.0 ± 0.1

22.5 ± 0.1

24.7 ± 0.1

1.53 ± 0.01

1.48 ± 0.02

1.40 ± 0.03

COX17/ACTB 1.43 ± 0.02

Mean CT values (triplicates for each copper concentration) normalized to mean CT values for β-actin.

copper excretion pump showed a normal function, as estimated by retention time. Higher uptake with normal retention implies an increased concentration of cellular copper, and this is supported by the findings of a 4-fold increased basal copper concentration in myoblasts from patients with Sco2 mutations (E140K/R90X) as compared to controls. Although a different impairment of copper metabolism is present in Menkes patients, similarly high basal copper concentrations (three to five times that of control) have been found in earlier studies in cultured human fibroblasts from these patients (27,28). Interestingly, increasing Cu–His supplementation resulted in elevated intracellular copper concentration both in control cells and in Sco2-deficient cells, but the increase in the latter was

only 5-fold compared to 20-fold in the former. As a result, at a Cu–His concentration of 300 µM the intracellular copper concentration was identical in both cell types. Such findings have also been obtained in cells from patients with Menkes disease with copper levels three times that at basal medium and 19 times in normal cells, when the medium was supplemented with 15–30 µg/ml copper (28). Although the basis for the abnormal copper handling in Sco2-deficient cells is currently unknown, it likely represents a compensating mechanism of the cell to overcome impaired Sco2 function. The complete restoration of COX activity after Cu–His supplementation clearly indicates that Sco2 plays a major role in mitochondrial copper transport to COX in muscle cells. Cu–His is a physiological compound in human cells with less toxic side effects than those described for other copper salts (29). There are several reports of clinical efficacy with this compound in the early treatment of Menkes patients (30–32). In our study, a primary toxic effect (necrosis, vacuolization, etc.) on cultured myoblasts was not observed even at the highest concentrations during a 3 week exposure (300 µM), and proliferation rates did not differ from those with lower or no copper supplementation. A long-term incubation (300 µM) over 6 weeks similarly resulted in normal cell growth with normalized COX values, indicating no visible side effects in vitro. The fact that increased intracellular copper can rescue the COX deficiency in Sco2-deficient human myoblasts contrasts with results obtained on Sco1 in yeast, where it was shown that addition of copper to the medium could not suppress the pet phenotype resulting from a SCO1 deletion (14). What might account the difference between the yeast and human cells? There are at least three possible explanations. First, it may be that the effective copper concentration in the mitochondrial intermembrane space of human myoblasts is higher than in yeast and sufficient to bind the CuA site (which protrudes into this compartment) directly. Secondly, it may be that a very low residual amount of Sco2 in human cells is sufficient to confer normal function, but only in the presence of high copper concentrations. A complete deletion might completely abolish respiratory chain function, resulting in embryonic lethality. Such a situation has been observed for frataxin in mice, whose deficiency in humans causes Friedreich ataxia (33). Finally, it is possible that another copper-binding protein with a partially overlapping function, perhaps Sco1, can substitute for Sco2, but only when the copper concentration is high. If this in fact occurs it apparently does not involve transcriptional upregulation of SCO1 as indicated by the real-time RT–PCR results. Whatever the mechanism of rescue of the biochemical defect, the copper supplementation experiments suggest a possible therapy. We treated two unrelated infants (both carriers of E140K/E140K) with Cu–His, but unfortunately in a late stage of the disease. No beneficial effects were observed in these cases, perhaps due to irreversible damage in the postmitotic CNS, heart and skeletal muscle. It is possible, that early treatment, as recommended in Menkes disease, could result in beneficial effects in patients with SCO2 mutations. The crucial point would be to recognize affected infants early. We have tested the occurrence of the common mutation in 400 agematched controls and found one heterozygous carrier without an additional mutation in SCO2 (data not shown). Given a frequency of 1:160 000 multiplied with the probability of 25% (autosomal-recessive trait) for the constellation E140K/E140K


and a still lower frequency for compound mutations, one can estimate a total frequency of ∼1:500 000. The fact that all patients so far identified carry the common G1541A mutation should be considered for a fast screening approach in all infants presenting with muscular hypotonia at or soon after birth. MATERIALS AND METHODS Patients, cell lines: fibroblasts, myoblasts and HeLa cells Skin fibroblast cell lines were established from three index patients. Patients 1A and 2B1, harboring mutations E140K/R171W (1A) and E140K/R90X (2B1), were described by Jaksch et al. (7) and patient A (E140K/E140K) was described by Jaksch et al. (11). One myoblast cell line was obtained from patient 2B2 (E140K/R90X), described by Jaksch et al. (7). Primary skin fibroblast cell lines from the four index patients and HeLa cells were grown in high glucose Dulbecco’s modified Eagle’s medium (DMEM) (Gibco) supplemented with 10% fetal bovine serum (FBS) (Gibco). One primary myoblast cell line from patient 2B2 and myoblast cell lines from 10 controls were grown in skeletal muscle growth medium (PromoCell), supplemented with 10% FBS (Gibco). Immunohistochemistry of different tissues (E140K/R90X) with polyclonal antibodies against Cox subunits II/III and Vab Autopsy specimens of skeletal muscle, heart, liver and kidney fixed in formalin and embedded in paraffin were obtained from patient 2B1 (see above) and used for immunohistochemical studies. For immunohistochemistry the ABC-Elite kit (DAKO) was used. Cox subunits II/III and Vab were detected as previously described by Müller-Höcker et al. (34). Construction of plasmids A plasmid containing enhanced green fluorescence protein (EGFP; Clontech) fused to the SCO2 gene (pEGFP–Sco2) was recently described by Paret et al. (10). The mutant gene coding for Sco2∆cys, in which the cysteines at positions 133 and 137 are changed to alanines, was obtained by overlap extension PCR with plasmid pEGFP–Sco2 as template and using primer 5′-tatataaagcttccaccatgctgctgctgactcggagcccc-3′ carrying a HindIII restriction site, primer 5′-agtataccgcggagacaggacactgcggaaagccgc-3′ carrying a SacII restriction site and the overlapping primers 5′-gatgtcaggggcgtgagtgaagcc-3′ and 5′-actcacgcccctgacatcgccccagacgagctgg-3′. The overlapping regions are underlined and the introduced mutations are shown in bold. The PCR products were cloned into the vector pEGFP–N1 to create pEGFP–Sco2∆cys. Plasmid pEGFP–Sco2 and pEGFP–Sco2∆cys were used for PCR amplification of the C-terminal portion (amino acids 79–266) of SCO2 gene using primer 5′-tatataggatccagggctgagaaggagaggctg-3′ with a BamHI restriction site and primer 5′-tataggatccgtcgactcaagacaggacactgcgg-3′ with a SalI restriction site. The products were ligated into the GST-fusion expression vector pGEX-4T-3 (Pharmacia) to create pGEX-4T-3 plasmid coding for GST–Sco2 and GST–Sco2∆cys.

Purification of a 22 kDa recombinant C-terminal Sco2 and Sco2∆cys The GST-fusion proteins were expressed in E.coli strain BL21 by induction with 0.1 mM isopropyl-1-thio-β-D-galactopyranoside (Sigma) for 30 min at 30°C. For protein analyses using AAS, 0.5 mM CuSO4 was added at the time of induction. Cells were lysed two times using a French press at 1000 p.s.i. in phosphate-buffered saline (PBS) containing 1 mM dithiothreitol (DTT) pH 7.3 with 10% glycerol and 1 mM phenylmethylsulfonyl fluoride. Cell debris was removed by centrifugation at 20 000 g at 4°C for 1 h. The GST-fusion proteins were affinity purified using GSH–Sepharose 4B beads (Pharmacia). Bound fusion proteins were cleaved on the matrix with thrombin (Sigma) in PBS–DTT to liberate the C-terminal portion of Sco2 or Sco2∆cys from the GST moiety. The proteins were analyzed by SDS–PAGE and stained with Coomassie blue G-250. Protein concentration was determined according to the method of Bradford. Analysis of the copper-binding potential of the C-terminal portion of Sco2 or Sco2∆cys using AAS and ICP–AES AAS and ICP–AES of the purified Sco2 segments isolated from E.coli grown in the presence of 0.5 mM CuSO 4 (see above) were performed with the SpectrAA GTA-96 (AAS, Varian; ICP–AES, Varian-Vista). Thioredoxin assay Thioredoxin activity of recombinant Sco2 (79–266) was assayed by its ability to reduce disulfide bridges in insulin in the presence of NADPH and thioredoxin reductase (35). The reaction mixture contained 500 µM NADPH, 55 µM insulin, 15 µM purified Sco2 (79–266) or 5 µM human thioredoxin as positive control (IMCO) in 100 mM potassium phosphate, 2 mM EDTA pH 7.4. The reaction was started by adding 80 nM human thioredoxin reductase (IMCO). The increase in turbidity as a result of the formation of the insoluble insulin B chain was followed at 595 nm. In vitro interaction assay of Sco2 Sco2–GST fusion proteins and GST, respectively, were immobilized on GSH–Sepharose beads and allowed to interact for 2 h at 4°C in the presence of 1 mM DTT with whole cell lysate from HeLa transformants expressing Sco2–EGFP, Sco2∆cys–EGFP or EGFP. Transfection of HeLa cells was performed using liposomes (Tfx-20; Promega) as described by the manufacturer. Lysis of cells was performed with 1% NP40, 150 mM KCl, 10 mM Tris–HCl pH 7.6, 20 mM MgCl2, 1 µg/ml chymostatin, 1 µg/ml pepstatin, 1 µg/ml antipain, 1 µg/ml aprotinin and 1 µg/ml leupeptin for 1 h on ice. After centrifugation at 4°C at 20 000 g for 20 min the supernatant was collected and diluted 4-fold with PBS–DTT. Following binding, the GSH–Sepharose beads were washed extensively with PBS–DTT containing 0.3% NP-40 and proteins were released by addition of sample buffer with β-mercaptoethanol and by heating at 100°C for 5 min. Samples were separated by SDS–PAGE, transferred to a PVDF membrane (Immobilon-P; Millipore), probed for 1 h with anti-GFP monoclonal antibody (Roche) and 30 min with a secondary anti-mouse antibody (Amersham Pharmacia Biotech). Antigen–antibody complexes


were visualized by enhanced chemiluminescence (ECL plus; Amersham Pharmacia Biotech) according to the manufacturer’s instructions.

selection for retroviral integration was performed in growth medium containing 500 µg/ml geneticin (G418) (Gibco) and continued for at least 7 days. The assay was done twice on different days with separately cultivated cells.

Immunoblotting of human cells with antibodies to Sco2 A polyclonal antiserum against the C-terminal part of human Sco2 (GRSRSAEQISDSVRRHMAAFRSVLS, conjugated with KLH) was raised in chicken. Eggs were harvested after three consecutive immunizations and total IgY was purified. The polyclonal antiserum to human Sco2 was affinity-purified from IgY with the respective peptide packed column. Mitochondria were purified using differential centrifugation. Twenty-five micrograms of protein from the mitochondrial fractions was loaded on 10% polyacrylamide gels. Electrophoresis and western blotting were done on the Bio-RAD mini-blot apparatus. Membranes (protran nitrocellulose; Schleicher & Schuell) were incubated overnight at 4°C with affinity-purified anti-Sco2 (1/100) and using a 1:10 000 dilution of a secondary anti-chicken IgY coupled to HRP (Jackson Immunoresearch). Signals were visualized with enhanced chemiluminescence (Super Signal West Femto/Pico Kit; Pierce) and X-ray film (CL-XPosure film; Pierce) according to the manufacturer’s recommendations. Bacterially expressed and purified C-terminal Sco2 protein (see above) served as positive control on Sco2 immunoblots at a concentration of 1 pg. Immunoblotting of copper-supplemented myoblasts with antibodies against Cox II was done as described for anti-Sco2 (see above), except that anti-Cox II (Molecular Probes) was used at a dilution of 1/5000 in TBST with 10% skim milk, followed by secondary anti-mouse IgG at a dilution of 1:10 000 coupled to HRP (Dako). Retrovirally mediated gene transfer of SCO2 into COX-deficient myoblasts Human SCO2 was amplified by PCR from genomic DNA of a healthy control subject using Pfu-polymerase and primers SCO2 (sense) 5′-taagcttagatccatgctgctgctga-3′ and SCO2 (antisense) 5′-tggatccggcccagactgcagtggct-3′, resulting in a PCR fragment of 839 bp containing the entire coding sequence of SCO2 (exon 2). To facilitate subcloning, a HindIII restriction site was included in the sense primers and a BamHI in the antisense primer. The fragment was shown not to harbor any alterations compared to the reported SCO2 sequence by direct sequencing. The fragment was cloned into the retroviral vector pLXSN (GenBank accession no. M28248; a kind gift by Dr A.D.Miller, Fred Hutchinson Cancer Research Center, Seattle, WA). The preparation of retroviral stocks (RV-SCO2) and infection of human myoblasts was carried out essentially as described previously (36,37). In brief, packaging cells were grown to near confluence in DMEM containing 10% FBS; 24 h later the medium containing retroviral particles was harvested and filtered through a 0.4 µm filter. Retroviral titers were determined by infecting NIH/3T3 cells. The 24 h supernatant (1 or 2 ml) was mixed with 2 or 3 ml of DMEM containing 10% FBS and polybrene (4 µg/ml). Human myoblasts (approximately 105) in 100 mm tissue culture dishes were incubated with this medium for 2 h after which an additional 5 ml of medium was added to the dish and the incubation was continued from 5 h to overnight. The cells were washed once in fresh medium to remove the polybrene, and 48 h later

Biochemical and histochemical determination of COX activity in myoblasts COX activity in non-transfected and transfected myoblasts (E140K/R90X), in control myoblasts, in myoblasts with and without Cu–His supplementation was determined spectrophotometrically (38) and histochemically (38,39) as described. Biochemical results were normalized to citrate synthase (CS), a mitochondrial marker enzyme. The reference ranges of COX and CS activity in myoblasts was calculated using control cells from 10 controls without a history of neuromuscular disease and without respiratory chain deficiency in skeletal muscle. Copper uptake of cultured fibroblasts with different SCO2 mutations Copper uptake and retention were measured using JFK’s standard protocol (40) in fibroblast cell lines from the three patients described above. The cells were pulsed for 20 h in 10 µM 64CuCl2 added to medium F12 supplemented with 20% FBS followed by a 24 h pulse chase in non-labeled medium. As reference values were evaluated in fibroblasts only, measurements in myoblasts were not performed. Determination of copper concentration in myoblasts using ICP–AES Copper concentrations were determined using ICP–AES in homogenates of myoblasts (E140K/R90X) and from a healthy control with and without Cu–His supplementation (final concentrations: 0, 15, 75, 150 and 300 µM) and related to g protein/ml. To prevent copper contamination from the medium, the cells were washed three times in Tris–EDTA buffer prior to homogenization. Cu–His supplementation of COX-deficient myoblasts (E140K/R90X) and control cells Several final concentrations of copper added (15, 75, 150 and 300 µM) as a Cu–His compound were made up in skeletal muscle growth medium (Promocell), added to the patient’s primary myoblast cell line (E140K/R90X) as well as to control myoblasts and incubated for 20 days with a medium change every third day. For each concentration two dishes were incubated and duplicate analyses were performed. Cu–His was prepared according to the protocol of Kreuder et al. (30) and stored at –80°C at a stock concentration of 2 mg Cu–His/ml/vial. These preparations were used for a maximum of three months. Expression of COX17/SCO1 in copper-treated myoblasts using real-time PCR Total RNA was isolated from Cu–His supplemented myoblasts (E140K/R90X) (Qiagen) and dissolved according to the manufacturer’s recommendation (Roche). RT was performed using hexamer primer (pN6; Roche) and 200 IU murine leukemia virus reverse transcriptase (Gibco Life Technologies). Quantification of mRNA abundance was performed by real-time PCR using an


ABI PRISM 7700 sequence detector (PE Biosystems) and SybrGreen as a double-stranded DNA-specific fluorescent dye. Amplification primers were designed using the software Primer Express (PE Biosystems). Each assay included triplicates of cDNA primed separately for the two genes of interest, SCO1 (accession no. NM_004589) and COX17 (accession no. L77701), for the reference gene ACTB (accession no. XM_037239) and a negative control (no template). The following primers were used: ACTB (forward) 5′-accccagccatgtacgtagc-3′ and (reverse) 5′-gtgtgggtgaccccgtctc-3′; SCO1 (forward) 5′-gctgcttcaattgccacaca-3′ and (reverse) 5′-actgcatccagcaacactgc-3′; COX17 (forward) 5′-cgatgcgtgtatcatcgagaa-3′ and (reverse) 5′-tcatgcattccttgtgggc-3′. Four dilutions of cDNA (1:25, 1:50, 1:100 and 1:200) were used to calculate the corresponding amplification efficiency (E = 10–(1/b) – 1; b = regression coefficient). The parameter CT (threshold cycle) is defined as the cycle number at which fluorescence intensity exceeds a fixed threshold. Relative mRNA expression for each cDNA of interest (I) was calculated using the formula: (1 + E(I))–CT(I) / (1 + E(β-actin))–CT(β-actin). ACKNOWLEDGEMENTS We greatly appreciate the excellent technical assistance of A.Zimmermann, U.Klutzny, K.Schulte and H.Hartl and we thank the pharmacologist M.Wurzer for providing us with Cu–His. We also thank A.Höflich for help and advice with the real-time RT–PCR analysis. We are grateful to the families whose collaboration made this study possible. Human myoblast cultures were obtained from the Muscle Tissue Culture Collection at the Friedrich-Baur-Institute. The Muscle Tissue Culture Collection is supported by grants of the Deutsche Gesellschaft für Muskelkranke (Freiburg, Germany) and the Association Francaise contre les Myopathies (Paris, France). This work was supported by grants from the Deutsche Forschungsgemeinschaft (Ja 802/2-1) (M.J.), the Friedrich-Baur-Stiftung (M.J., H.L.), the Ernst und Berta Grimmke Stiftung (M.J., R.H.), the CIHR (E.A.S.) and the March of Dimes Birth Defects Association (E.A.S.). E.A.S. is an MNI Killam Scholar. REFERENCES 1. Shoubridge, E.A. (2001) Cytochrome c oxidase deficiency. Am. J. Med. Genet. (Semin. Med. Genet.), 106, 46–52. 2. Adams, P.L., Lightowlers, R.N. and Turnbull, D.M. (1997) Molecular analysis of cytochrome c oxidase deficiency in Leigh’s syndrome. Ann. Neurol., 41, 268–270. 3. Jaksch, M., Hofmann, S., Kleinle, S., Liechti-Gallati, S., Pongratz, D.E., Muller-Hocker, J., Jedele, K.B., Meitinger, T. and Gerbitz, K.D. (1998) A systematic mutation screen of 10 nuclear and 25 mitochondrial candidate genes in 21 patients with cytochrome c oxidase (COX) deficiency shows tRNA(Ser) (UCN) mutations in a subgroup with syndromal encephalopathy. J. Med. Genet., 35, 895–900. 4. Zhu, Z., Yao, J., Johns, T., Fu, K., de Bie, I., Macmillan, C., Cuthbert, A.P., Newbold, R.F., Wang, J.-C., Chevrette, M. et al. (1998) Surf1, a factor involved in the biogenesis of cytochrome c oxidase, is mutated in Leigh Syndrome. Nat. Genet., 20, 337–343. 5. Tiranti, V., Hoertnagel, K., Carrozzo, R., Galimberti, C., Munaro, M., Granatiero, M., Zelante, L., Gasparini, P., Marzella, R., Rocchi, M. et al. (1998) Mutations of SURF-1 in Leigh disease associated with cytochrome c oxidase deficiency. Am. J. Hum. Genet., 63, 1609–1621. 6. Papadopoulou, L.C., Sue, C.M., Davidson, M., Tanji, K., Ichizo, N., Sadlock, J., Selby, J., Glerum, D.M., Van Coster, R., Lyon, G. et al. (1999) Fatal infantile cardioencephalomyopathy with cytochrome c oxidase (COX)

7.

8.

9.

10.

11.

12.

13.

14.

15.

16. 17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

deficiency and mutations in SCO2, a human COX assembly gene. Nat. Genet., 23, 333–337. Jaksch, M., Ogilvie, I., Kortenhaus, G., Bresser, H.G., Gerbitz, K.D. and Shoubridge, E.A. (2000) Mutations in SCO2 are associated with a distinct form of hypertrophic cardiomyopathy and cytochrome c oxidase deficiency. Hum. Mol. Genet., 9, 795–801. Valnot, I., von Kleist-Retzow, J.C., Barrientos, A., Gorbatyuk, M., Taanman, J.W., Mehaye, B., Rustin, P., Tzagoloff, A., Munnich, A. and Rotig, A. (2000) A mutation in the human heme A:farnesyltransferase gene (COX10) causes cytochrome c oxidase deficiency. Hum. Mol. Genet., 9, 1245–1249. Valnot, I., Osmond, S., Gigarel, N., Mehaye, B., Amiel, J., Cormier-Daire, V., Munnich, A., Bonnefont, J.P., Rustin, P. and Rotig, A. (2000) Mutations of the SCO1 gene in mitochondrial cytochrome c oxidase (COX) deficiency with neonatal-onset hepatic failure and encephalopathy. Am. J. Hum. Genet., 67, 1104–1109. Paret, C., Ostermann, K., Krause-Buchholz, U., Rentzsch, A. and Rodel, G. (1999) Human members of the SCO1 gene family: complementation analysis in yeast and intracellular localization. FEBS Lett., 447, 65–70. Jaksch, M., Horvath, R., Horn, N., Auer, D.P., Macmillan, C., Peters, J., Gerbitz, K.-D., Shoubridge, E.A., Krägeloh-Mann, I., Lochmüller, H. and Freisinger, P. (2001) Homozygosity (E140K) in SCO2 causes delayed infantile onset of cardiomyopathy and neuropathy. Neurology, 57, 1440–1446. Dickinson, E.K., Adams, D.L., Schon, E.A. and Glerum, D.M. (2000) A human SCO2 mutation helps define the role of Sco1p in the cytochrome oxidase assembly pathway. J. Biol. Chem., 275, 26780–26785. Schulze, M. and Rödel, G. (1988) SCO1, a yeast nuclear gene essential for accumulation of mitochondrial cytochrome c oxidase subunit II. Mol. Gen. Genet., 211, 492–498. Glerum, D.M., Shtanko, A. and Tzagoloff, A. (1996) SCO1 and SCO2 act as high copy suppressors of a mitochondrial copper recruitment defect in Saccharomyces cerevisiae. J. Biol. Chem., 271, 20531–20535. Dancis, A., Yuan, D.S., Haile, D., Askwith, C., Eide, D., Moehle, C., Kaplan, J. and Klausner, R.D. (1994) Molecular characterization of a copper transport protein in S.cerevisiae: an unexpected role for copper in iron transport. Cell, 76, 393–402. Neupert, W. (1997) Protein import into mitochondria. Annu. Rev. Biochem., 66, 863–917. Mattatall, N.R., Jazairi, J. and Hill, B.C. (2000) Characterization of YpmQ, an accessory protein required for the expression of cytochrome c oxidase in Bacillus subtilis. J. Biol. Chem., 275, 28802–28809. Hiser, L., Di Valentin, M., Hamer, A.G. and Hosler, J.P. (2000) Cox11p is required for stable formation of the Cu(B) and magnesium centers of cytochrome c oxidase. J. Biol. Chem., 275, 619–623. Lode, A., Kuschel, M., Paret, C. and Rodel, G. (2000) Mitochondrial copper metabolism in yeast: interaction between Sco1p and Cox2p. FEBS Lett., 485, 19–24. Chinenov, Y.V. (2000) Cytochrome c oxidase assembly factors with a thioredoxin fold are conserved among prokaryotes and eukaryotes. J. Mol. Med., 78, 239–242. Chelly, J., Tumer, Z., Tonnesen, T., Petterson, A., Ishikawa-Brush, Y., Tommerup, N., Horn, N. and Monaco, A.P. (1993) Isolation of a candidate gene for Menkes disease that encodes a potential heavy metal binding protein. Nat. Genet., 3, 3–4. Vulpe, C., Levinson, B., Whitney, S., Packman, S. and Gitschier, J. (1993) Isolation of a candidate gene for Menkes disease and evidence that it encodes a copper-transporting ATPase. Nat. Genet., 3, 7–13. Mercer, J.F., Livingston, J., Hall, B., Paynter, J.A., Begy, C., Chandrasekharappa, S., Lockhart, P., Grimes, A., Bhave, M., Siemienak, D. et al. (1993) Isolation of a partial candidate gene for Menkes disease by positional cloning. Nat. Genet., 3, 20–25. Loferer, H., Bott, M. and Hennecke, H. (1993) Bradyrhizobium japonicum TlpA, a novel membrane anchored thioredoxin-like protein involved in the biogenesis of cytochrome aa3 and development of symbiosis. EMBO J., 12, 3373–3383. Casareno, R.L.B., Waggoner, D. and Gitlin, J.D. (1998) The copper chaperone CCS directly interacts with copper/zinc superoxide dismutase. J. Biol. Chem., 273, 23625–23628. Bonne, G., Seibel, P., Possekel, S., Marsac, C. and Kadenbach, B. (1993) Expression of human cytochrome c oxidase subunits during fetal development. Eur. J. Biochem., 217, 1099–1107. Goka, T.J., Stevenson, R.E., Hefferan, P.M. and Howell, R.R. (1976) Menkes disease: a biochemical abnormality in cultured human fibroblasts. Proc. Natl Acad. Sci. USA, 73, 604–606.


28. Chan, W.Y., Garnica, A.D. and Rennert, O.M. (1978) Cell culture studies of Menkes kinky hair disease. Clin. Chim. Acta, 88, 495–507. 29. Ferruzza, S., Scarino, M.L., Rotilio, G., Ciriolo, M.R., Santaroni, P., Muda, A.O. and Sambuy, Y. (1999) Copper treatment alters the permeability of tight junctions in cultured human intestinal Caco-2 cells. Am. J. Phys., 277, G1138–G1148. 30. Kreuder, J., Otten, A., Fuder, H., Tumer, Z., Tonnessen, T., Horn, N. and Dralle, D. (1993) Clinical and biochemical consequences of copper-histidine therapy in Menkes disease. Eur. J. Pediatr., 152, 828–832. 31. Kaler, S.G. (1998) Diagnosis and therapy of Menkes syndrome, a genetic form of copper deficiency. Am. J. Clin. Nutr., 67, 1029S–1034S. 32. Christodoulou, J., Danks, D.M., Sarkar, B., Baerlocher, K.E., Casey, R., Horn, N., Tumer, Z. and Clarke, J.T. (1998) Early treatment of Menkes disease with parenteral copper-histidine: long-term follow-up of four treated patients. Am. J. Med. Genet., 76, 154–164. 33. Puccio, H., Simon, D., Cossee, M., Criqui-Filipe, P., Tiziano, F., Melki, J., Hindelang, C., Matyas, R., Rustin, P. and Koenig, M. (2001) Mouse models for Friedreich ataxia exhibit cardiomyopathy, sensor nerve defect and Fe-S enzyme deficiency followed by intramitochondrial iron deposits. Nat. Genet., 27, 181–186.

34. Müller-Höcker, J., Droste, M., Kadenbach, B., Pongratz, D. and Hübner, G. (1989) Fatal mitochondrial myopathy with cytochrome-c-oxidase deficiency and subunit-restricted reduction of enzyme protein in two siblings: an autopsy-immunocytochemical study. Hum. Pathol., 20, 666–672. 35. Holmgren, A. (1979) Reduction of disulfides by thioredoxin. Exceptional reactivity of insulin and suggested functions of thioredoxin in mechanism of hormone action. J. Biol. Chem., 254, 9113–9119. 36. Miller, A.D. and Rosman, G.J. (1989) Improved retroviral vectors for gene transfer and expression. Biotechniques, 7, 980–990. 37. Lochmuller, H., Johns, T. and Shoubridge, E.A. (1999) Expression of the E6 and E7 genes of human papillomavirus (HPV16) extends the life span of human myoblasts. Exp. Cell Res., 248, 186–193. 38. Capaldi, R.A., Marusich, M.F. and Taanman, J.W. (1995). Mammalian cytochrome c oxidase: characterization of enzyme and immunological detection of subunits in tissue extracts and whole cells. Methods Enzymol., 260, 117–132. 39. Seligmann, A.M., Karnovsky, M.J., Wasserkrug, H.L. and Hanker, J.S. (1968) Nondroplet ultrastructural demonstration of cytochrome c oxidase activity with a polymerizing osmiophilic reagent, diaminobenzidine (DAB). J. Cell Biol., 38, 1–14. 40. Tumer, Z. and Horn, N. (1997) Menkes disease: recent advances and new aspects. J. Med. Genet., 34, 265–274.
