How many amino acids of cytochrome c in chimpanzee are different from humans?

Respiring mitochondria require many interactions between nuclear and mitochondrial genomes. Although mitochondrial DNA (mtDNA) from the gorilla and the chimpanzee are able to restore oxidative phosphorylation in a human cell, mtDNAs from more distant primate species are functionally incompatible with human nuclear genes. Using microcell-mediated chromosome and mitochondria transfer, we introduced and maintained a functional orangutan mtDNA in a human nuclear background. However, partial oxidative phosphorylation function was restored only in the presence of most orangutan chromosomes, suggesting that human oxidative phosphorylation-related nuclear-coded genes are not able to replace many orangutan ones. The respiratory capacity of these hybrids was decreased by 65%–80%, and cytochrome c oxidase (COX) activity was decreased by 85%–95%. The function of other respiratory complexes was not significantly altered. The translation of mtDNA-coded COX subunits was normal, but their steady-state levels were ∼10% of normal ones. Nuclear-coded COX subunits were loosely associated with mitochondrial membranes, a characteristic of COX assembly-defective mutants. Our results suggest that many human nuclear-coded genes not only cannot replace the orangutan counterparts, but also exert a specific interference at the level of COX assembly. This cellular model underscores the precision of COX assembly in mammals and sheds light on the nature of nuclear-mtDNA coevolutionary constraints.

Introduction

The exact extent and nature of the contribution made by the nucleus to the maintenance and function of respiring mitochondria in mammalian cells is not yet fully understood. Nevertheless, it is clear that the nuclear and mitochondrial genomes coevolve to optimize protein-protein, protein–mitochondrial RNA, and protein–mitochondrial DNA (mtDNA) interactions, assuring an efficient oxidative phosphorylation system. The rate of mtDNA evolution in primates is 5 to 10 times as high as that of nuclear DNA, but functionally important mutations fixed in mtDNA may have to be compensated by alterations in nuclear-coded genes (Brown et al. 1979 ). Moreover, studies of rodent × rodent and rodent × human interspecific hybrids suggested that the coevolution of the two genomes has made these interactions species-specific (Clayton et al. 1971 ; De Francesco, Attardi, and Croce 1980 ; Giles, Stroynowski, and Wallace 1980 ; Ziegler and Davidson 1981 ; Hayashi et al. 1983 ). However, our laboratory has shown that the respiratory capacity of a human cell line lacking mtDNA (ρ°) could be essentially restored by introducing mtDNA from the genus Pan (common and pigmy chimpanzees) or Gorilla (Kenyon and Moraes 1997 ), which diverged from humans 6–15 MYA (Arnason et al. 1996 ). In these successful human xenomitochondrial cybrids (HXCs), the mtDNA was efficiently translated, and oxidative phosphorylation was restored. However, HXC had a respiratory capacity limited to ∼80% of the normal capacity due to a specific ∼40% decrease in mitochondrial complex I activity. This defect was attributed to defective interactions between the nuclear DNA–coded and the mtDNA-coded subunits of the enzyme (Barrientos, Kenyon, and Moraes 1998 ). The threshold in the evolutionary distance from humans for this functional complementation was abrupt. MtDNA from species that diverged from humans more than approximately 18–25 MYA (e.g., from the genera Pongo [orangutan] and species representative of Old World monkeys) could not restore oxidative phosphorylation in human cells. This abrupt incompatibility is probably the result of excessive (quantitative and qualitative) suboptimal interactions between nuclear and mitochondrial genomes, involving only one factor or several factors. The present work showed a very specific and strong interference between species-specific nuclear- and mitochondrial-coded subunits of complex IV in human × orangutan hybrids harboring exclusively orangutan mtDNA.

Materials and Methods

Cell lines and Culture Conditions

Orangutan (Pongo pygmaeus) and gorilla (Gorilla gorilla) skin fibroblasts were obtained from the Coriell Institute for Medical Research Repository. Cells were cultured in high-glucose Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum and 100 μg/ml sodium pyruvate. Human fibroblasts were obtained from a patient who was suffering from a nonmitochondrial disease. The human osteosarcoma-derived cell line 143B (TK−) and its mtDNA-less derivative, 143B/206 ρ°, were a gift from Dr. Michael P. King (Thomas Jefferson University, Philadelphia, Pa.) and were cultured as described elsewhere (King and Attardi 1989 ).

Transfer of Chromosomes and mtDNA from Orangutan to Human ρ° Cells

Microcell Fusions

Microcell isolation and fusion with ρ° cells were performed as described (Barrientos and Moraes 1998 ). After fusion, hybrid cells were allowed to grow overnight and then selected for the presence of oxidative phosphorylation (OXPHOS) function with a medium lacking uridine and supplemented with 10% dialyzed serum (only cells with at least partial OXPHOS function can grow in the absence of uridine). These hybrid clones, denoted M#, were maintained continuously in selective medium.

Whole-Cell Fusions

Orangutan fibroblasts were fused with the human 143B/206 ρ° cell line, following a procedure similar to that used for microcell fusion. The hybrid cells were also selected for the presence of orangutan mtDNA with a medium lacking uridine. Whole-cell hybrid clones, denoted H#, were maintained continuously in selective medium. To reduce the number of chromosomes to the minimum necessary to maintain a functional orangutan mtDNA in the hybrid cells, two of these hybrid clones (H1 and H2) were incubated with 50 ng/ml colcemide (Gibco-BRL) for 24 or 48 h. Colcemide arrests the cells in mitosis, which can result in chromosomal loss (Tsutsui et al. 1990 ). Cells were allowed to recover in two different media selective for respiratory function: a medium rich in glucose but lacking uridine, and a more stringent one in which glucose was substituted with galactose as a carbon source. Those clones were denoted H#c.

Quantitation of Fusion Products

Human, gorilla, and orangutan fibroblasts, as well as a human ρ° cell line (used as a negative control), were fused with the human 143B/206 ρ° cell line stably transfected with the plasmid pSV2-Neo (Southern and Berg 1982 ) containing a gene for neomycin resistance. Four independent fusions were made for each species. One day after fusion, cells were split into twenty 60-mm2 dishes. Cells were selected in 450 μg/ml of G418 (Gibco-BRL) and, for the presence of functional mitochondria, with a medium lacking uridine. The number of independent colonies formed was determined in 50% of the dishes after 13 days by staining the dishes with toluidine blue. After 25 days, the second half was counted to estimate the number and stability of the respiratory competent hybrids formed.

Chromosomal Identification by Microsatellite Markers

Human microsatellite markers have been used for analysis of genetic variation in apes (Coote and Bruford 1996 ). We selected 33 microsatellite primer pairs for potential use in distinguishing chromosomes from human 143B/206 ρ° cells and orangutan fibroblasts. We used MAPPAIRS primers (Research Genetics, Huntsville, Ala.) references D1S551, D1S552, D2S2739, D2S427, D3S3038, D3S1535, D4S2639, D4S2431, D5S1470, D5S820, D6S1056, D7S820, D7S1820, D8S1113, D9S934, D10S677, D11S1985, D11S2369 D12S391, D13S788, D14S587, D15S642, D16S753, D16S2621, D17S1293, D17S1290, D18S851, D19S246, D20S481, D21S1437, D21S1270, D22S683, and DXS6797. MAPPAIRS primer references D7S820, D16S753, and D21S1437 were not informative and were not considered in the study. Basically, the forward primer of each pair was end-labeled with [γ-32P]dATP using T4-polynucleotide kinase (New England Biolabs), the loci were PCR amplified, and 4 μl of the product was loaded onto a 6% denaturing polyacrylamide gel, followed by autoradiography.

Karyotype Analysis by Fluorescent In Situ Hybridization (FISH)

By cross-species color segmenting (Müller et al. 1998 ), also termed RxFISH, a chromosome bar code on human and great ape chromosomes was generated, which assisted in the identification of human and orangutan chromosomes. The respective FISH probe set was composed of gibbon chromosome–specific painting probes, which were labeled with three fluors in various combinations. Gibbon chromosomes differ from all other great ape homologs by multiple translocations and inversions. For this reason, the RxFISH probe set produces a multicolored banding pattern on both human and orangutan homologs. Since the majority of both species' homologs are evolutionary conserved and cannot be distinguished by chromosome morphology and banding pattern alone, the interspecies comparative genomic hybridization (iCGH) technique was employed simultaneously. Originally, CGH was developed in order to visualize genomic imbalances in differentially labeled tumor versus control genomic DNA samples, particularly deletions and amplifications (Kallioniemi et al. 1992 ). Adaptation of the experimental setup allowed the differential display of human and orangutan chromosomes, taking advantage of the overall sequence divergence between both genomes: human genomic probe DNA preferentially hybridizes to human chromosomes and vice versa. Both the RxFISH and the iCGH probe sets were hybridized simultaneously using M-FISH technology and five fluors (Speicher, Gwyn Ballard, and Ward 1996 ). Approximately 400 ng of each genomic orangutan DNA (biotin-dUTP labeled) and human genomic DNA (digoxigenin-dUTP labeled) were ethanol-precipitated and subsequently resuspended in 10 μl commercially available RxFISH probe mixture (labeled with FITC-dUTP, Cy3-dUTP, and Cy5-dUTP; Applied Imaging Corp.). Hybridization in situ to metaphase preparations of human × orangutan hybrid cell lines was performed for 72 h at 37°C. Post-hybridization washes included 2 × 5 min 50% formamide/1 × SSC, 45°C; 2 × 5 min 2 × SSC, 45°C; and 2 × 5 min 0,1 × SSC, 60°C. Biotinylated probe was detected by avidin-Cy3.5, and digoxigenin labeled probe by mouse-anti-digoxigenin-Cy5.5 and visualized by a microscopic setup described in Speicher, Gwyn Ballard, and Ward (1996).

Mitochondrial DNA Identification and Quantification

To insure that the only mtDNA present in the human ρ° × orangutan ρ+ (Hρ° × Or) hybrids were from orangutan, a Southern blot analysis was performed using total DNA digested with PvuII and a mixture of a 956 bp human mtDNA (nt 3305–4261) and its homologous orangutan mtDNA fragment [α-32P]dCTP-labeled probes. PvuII restriction patterns differ in human mtDNA (a fragment of 16.5 kb) and orangutan mtDNA (two fragments of 8 and 8.3 kb, only one of them detected with the partial mtDNA probe used). To determine the level of repopulation by orangutan mitochondria in Hρ° × Or hybrid clones, the mtDNA content relative to the nuclear DNA (nDNA) was quantified by a slot blot experiment using the same mixture of mtDNA probes plus a 5.8-kb 18S rDNA nuclear (Wilson et al. 1978 ) [α-32P]dCTP-labeled probe. For each probe, we used three slots, containing 100, 200, and 300 ng of total DNA. After laser-scanning the autoradiograms, band signals were quantified using NIH Image 1.6 software. The mtDNA/nDNA ratio was considered the division of the arbitrary densitometrical values of the signals using each probe.

Mitochondrial Function Studies

Exponentially growing cells were collected by trypsinization, pelleted, and resuspended in cold phosphate-buffered saline (PBS) medium. The KCN-sensitive endogenous cell respiration in intact cells was measured polarographically as described (Barrientos, Kenyon, and Moraes 1998 ). Mitochondria were isolated from cells and resuspended in a medium consisting of 20 mM Tris (pH 7.2), 0.25 M sucrose, 40 mM KCl, 2 mM EGTA, and 1 mg/ml BSA. The measurement of the specific activity of the individual complexes of the electron transport chain was performed spectrophotometrically as described (Barrientos, Kenyon, and Moraes 1998 ). The electron transport chain enzyme activities were normalized both by milligrams of protein and by citrate synthase activity. The protein content in the cell and mitochondria samples was determined according to Bradford's (1976) method. Complex IV activity was also assayed cytochemically on cells grown on coverslips, as described (Seligman et al. 1968 ). Complex IV inhibition by KCN was titrated by measuring the reduction in cytochrome c oxidation in the presence of increasing concentrations (up to 20 mM) of the specific inhibitor KCN. All experiments were performed at least in triplicate.

Mitochondrial Protein Synthesis

Mitochondrial protein synthesis was determined by pulse-labeling cell cultures in the presence of emetine as described (Chomyn 1996 ). Four Hρ° × Or hybrids (M1, M2, H1, and H1c) were used for this experiment. Orangutan fibroblasts and human 143B cells were used as a control for the mtDNA-coded protein pattern in both species, and the 143B derivative 206 ρ° was used as a negative control. Semiconfluent cells were labeled and processed as described (Chomyn 1996 ). Approximately 45 μg of total protein was resolved by electrophoresis in a 15%–20% exponential gradient polyacrylamide gel (Chomyn 1996 ). Gels were fixed in a 30% methanol/10% acetic acid solution and treated with Fluoro-Enhance (Research Products International), dried, and exposed to an X-ray film at −80°C.

Detection of Mitochondrial Respiratory Chain Enzyme Subunits

Immunoblottings were performed using monoclonal antibodies against the human succinate dehydrogenase flavoprotein subunit (SDH(Fp)), core-1 of complex III, COX I, COX II, COX IV, COX Va, and subunit α of ATPase (Marusich et al. 1997 ) (a gift from Dr. R. Capaldi, University of Oregon) and a polyclonal antibody against the human ND1 (a gift from Dr. A. Lombes, Groupe Hospitalier Pitiè-Salpètrière, Paris). A mitochondria-enriched pellet was prepared as described (Barrientos, Kenyon, and Moraes 1998 ) from orangutan fibroblasts, from 143B and 143B/206 ρ° cells, and from several Hρ° × Or hybrid clones (M1–M4, H1, H2, H1c, and H2c). Ten to twenty micrograms of mitochondrial proteins were separated onto 15% SDS-PAGE gels and transferred to PVDP membranes (Immobilon, Bio-Rad). Membranes were incubated for 1 h with 10% milk in PBS with 0.05% Tween 20 and with antibodies against different mtDNA- or nDNA-coded respiratory chain subunits for 14 h at 4°C. The chemiluminescent detection of the proteins was performed with the Phototope-HRP Western blot detection kit using an anti-rabbit or anti-mouse IgG secondary antibody, which was HRP-linked (New England Biolabs, Beverly, Mass.) following the manufacturer's recommendations.

Results

Many Human Nuclear-Coded Factors Do Not Interact Productively with Orangutan mtDNA-Coded Factors

We have shown that human cells lacking mtDNA (ρ°) do not regain respiratory function through the introduction of orangutan mitochondria (hence, mtDNA) into their cytoplasm (Kenyon and Moraes 1997 ). The exact causes of this incompatibility are unknown, and it is a challenge not only at the conceptual level, but also at the technical level. We tried to identify the factor(s) involved in orangutan mtDNA–human nuclear DNA incompatibilities by supplementing the human background not only with orangutan mtDNA, but also with a limited set of orangutan chromosomes. Mitochondrial DNA and chromosomes from orangutan were transferred in a single step to human cells by microcell-mediated transfer or by whole-cell hybridization. Our previous experience with this technique assured us that we could transfer a limited number of chromosomes simultaneously with mtDNA (Barrientos and Moraes 1998 ). More than 10 independent microfusions were performed, ∼50% of which produced a few surviving clones when selected in uridineless medium. Cells with a complete lack of respiratory function cannot grow in the absence of uridine, because the pyrimidine biosynthetic enzyme dihydroorotate dehydrogenase is located in the mitochondrial inner membrane and requires electron transfer for its function (King and Attardi 1989 ). This low efficiency in the generation of uridine-independent clones suggested that the transfer of orangutan mtDNA plus a very limited number of chromosomes was not sufficient to restore OXPHOS in a human nuclear background. To observe the effects of supplementing human ρ° cells with a complete set of orangutan chromosomes, three independent fusions were performed between human ρ° and intact orangutan fibroblasts. The efficiency of these whole-cell fusions in producing clones after 25 days of selection in uridineless medium was ∼7% that observed in a human ρ° × human fibroblast fusion, suggesting that there is a dominant detrimental effect on OXPHOS when human nuclear genes interact with orangutan nDNA or mtDNA. The presence of mtDNA in the hybrids was ascertained by Southern blot analysis, and its orangutan origin was confirmed by its specific PvuII restriction pattern (fig. 1A) . To estimate the levels of orangutan mtDNA in the human ρ° × orangutan ρ+ hybrid clones (M1–M8 [derived from microcell fusions] and H1 to H8 [derived from whole-cell fusions]), the mtDNA content relative to an nDNA gene was assessed by slot blot. All hybrids had mtDNA/nDNA ratios in the range of those found in the orangutan fibroblast and 143B cells (fig. 1B ).

The chromosome content of Hρ°–Or hybrid cell lines that could grow in the absence of uridine (i.e., had “some” respiratory function) was analyzed by multicolor FISH. In order to discriminate between human and orangutan homologous chromosomes and to simultaneously identify individual chromosomes, we integrated the recently introduced iCGH and RxFISH (Müller et al. 1998 ) techniques. These techniques are explained in details in the Materials and Methods section. This combined approach was obligatory, since chromosome morphologies in humans and orangutans are essentially identical. The majority of both species homologs cannot be distinguished by chromosome banding techniques. The iCGH protocol, which uses genomic DNA (either human or orangutan) as a probe, proved to be robust and reliable in discriminating DNA sequences between humans and orangutans. Despite an approximate overall divergence of less than 4%, chromosomes of both species were hybridized preferentially by genomic DNA of the same species, and hybridization of the other species was reciprocally suppressed.

The chromosome content analyzed in four hybrids produced by orangutan microcell fusions displayed many orangutan chromosomes, as well as several chromosomal abnormalities, including frequent human-orangutan translocations and dicentric (same or different species) chromosomes (fig. 2A, summarized in table 1 ). These made very difficult, if not impossible, any genotype-phenotype association. The total number of chromosomes in the hybrids able to grow in uridineless medium after fusions between human ρ° × orangutan microcells was close to 100, and approximately 30% of them were from orangutan (table 1 ).

We also selected 33 microsatellite primer pairs to further study the chromosomal composition in these cells. In all hybrids, the microsatellite analysis showed the presence of all of the human and most of the orangutan loci tested (table 2 and fig. 2B ). The difference between the numbers of microsatellite markers detected in hybrids produced by microcell transfer and those produced by whole-cell hybridization was not significant. In complete hybrids, the maximum number of orangutan loci missing was two. This number increased to five (e.g., clone M1) in the hybrids obtained by the microcell transfer system. Among all of the clones obtained, the one containing fewer orangutan chromosomal markers, while still maintaining some OXPHOS function, was clone M10, which showed loss of heterozygosity for seven loci and the absence of three other loci (table 2 ). The orangutan locus D19S246 was lost in 8 out of 10 hybrids tested (table 2 ). Cells treated with colcemide to induce chromosome loss further lost some of the orangutan alleles (table 2 ), but a dramatic reduction in the number of orangutan chromosome markers was never obtained. The presence of a large number of orangutan chromosomes in hybrids produced by microcell-mediated fusion explained the very low efficiency in obtaining OXPHOS-competent cells, as microcells containing large numbers of chromosomes are very rare (Barrientos and Moraes 1998 ). Nevertheless, hybrids made with these microcells seem to be the only ones that can survive in selective media for OXPHOS function (i.e., no uridine).

Human ρ° × Orangutan ρ+ Hybrids Have a Specific Mitochondrial Complex IV Deficiency

Although human ρ° × orangutan ρ+ hybrids could grow in the absence of uridine, it was not clear if their OXPHOS function was only minimally restored. In fact, the low efficiency in producing full hybrids under no uridine selection suggested that human chromosomes were associated with a dominant negative effect in the respiration of orangutan cells. Therefore, we tested the growth properties of these hybrids in galactose media. Respiratory competent cells grow exponentially in medium containing glucose as a carbon source and at a slightly reduced rate in galactose-containing medium (Robinson 1996 ; Barrientos and Moraes 1999 ). Cells lacking mtDNA cannot grow in galactose medium, since they are totally dependent on glycolysis due to a nonfunctional mitochondrial respiratory chain, and galactose cannot be efficiently used for glycolysis (Robinson 1996 ). The ρ° cells are auxotrophic for pyrimidines and must be cultured in glucose-containing medium supplemented with uridine (King and Attardi 1989 ). Selection in uridineless medium, however, is not as stringent as galactose selection, and cells with severely defective OXPHOS can still grow in the absence of uridine (Hao and Moraes 1997 ). As described in the previous section, Hρ° × Or hybrids were selected in the absence of uridine, but these hybrid clones were unable to grow in galactose medium. We studied this phenotype further in Hρ° × Or hybrids using somatic cell manipulations and biochemical techniques. Several strategies were attempted to select galactose-resistant hybrid clones. First, different hybrid clones (108 cells) growing in uridineless medium were switched to galactose-containing medium in an attempt to isolate revertants. No resistant clones were observed after this procedure. A progressive substitution of the carbon source to facilitate an adaptation to the new metabolic conditions was also unable to produce revertants. It seems that the defective phenotype of Hρ° × Or hybrids could not be reverted by metabolic pressure. Finally, hybrids were treated with colcemide to induce chromosomal losses, followed by galactose selection. However, we still could not obtain growing clones.

The apparent OXPHOS deficiency, suggested by the growth behavior of Hρ° × Or hybrids in the different selective media used, prompted us to perform polarographic analyses of the hybrids' respiratory capacity. Intact cell oxygen consumption in the hybrids was decreased by 68%–80% as compared with the human parental cell line (143B), and by 62%–75% as compared with the orangutan parental line (fibroblast) (P < 0.001 in both cases). Figure 3A summarizes the oxygen consumption data on intact cells. Mitochondrial respiratory chain enzyme activities, measured in isolated mitochondria, were normalized to citrate synthase (CS) activity (fig. 3B ). Cell respiration in the hybrids was always less than 40% of that in controls regardless of the selection strategy used. Complex IV/CS ratios were decreased by 87%–95 % as compared with the human cell line, and by 85%–94% as compared with the orangutan cell line (P < 0.001 in both cases). Complex I, II+III, and III activities were not significantly altered in Hρ° × Or hybrids (fig. 3B ). These data are consistent with a specific complex IV or cytochrome c oxidase (COX) deficiency in these cell lines. To investigate whether the enzyme deficiency was associated with structural changes that would affect binding to inhibitors, we studied the kinetics of KCN inhibition on COX activity (fig. 3C ). In all cell lines studied, the kinetics of COX inhibition was similar. The Ki of KCN in each case was obtained by plotting the concentration of KCN versus the ratio [KCN]/% inhibition (fig. 3C ). The Ki values of KCN inhibition were similar between the parental cells (1.67 ± 0.16 for 143B, and 1.48 ± 0.25 for orangutan fibroblasts) and the Hρ° × Or hybrids (1.73 ± 0.39), indicating that there were no changes in KCN affinity for its binding site in complex IV, but, rather, a different amount of either enzyme or activity susceptible to inhibition by KCN. Therefore, the biochemical basis of the galactose growth failure in our human ρ° × orangutan ρ+ hybrids seemed to consist in a specific decrease in COX activity that is otherwise similar to the activity observed in orangutan cells.

Mitochondrial Complex IV Deficiency in Human ρ° × Orangutan ρ+ Hybrids Is Associated with a Defect in COX Assembly

We analyzed some of the putative factors responsible for the mitochondrial complex IV deficiency observed in Hρ° × Or hybrids. Pulse-labeling of mitochondrial translation products in whole cells in the presence of emetine (an inhibitor of cytosolic protein synthesis) showed that the incorporation of [35S] methionine into the mtDNA-coded COX subunits was as efficient in the hybrids harboring orangutan mtDNA as in the orangutan fibroblasts (fig. 4A ). This indicated that the defects found in the Hρ° × Or hybrids could not be explained by a defect in either transcription or translation of orangutan mitochondrial genes. However, the steady-state levels of the mtDNA-coded COX subunits I and II were drastically reduced (∼90%) in Hρ° × Or hybrids when compared with the human and orangutan parental cell lines, as determined by immunoblot analyses (fig. 4B ). The nuclear-coded subunit COX Va was also normally transported and processed. In the hybrids, this subunit was detected at levels comparable to those of the parental lines (fig. 4B ). The monoclonal antibody used for the COX IV subunit recognizes the orangutan form of this protein very weakly, and therefore it was possible to estimate only the levels of the human form. Immunodetectable COX IV protein levels were lower in the hybrids than in the human ρ°, but the total content of this protein in the hybrids was likely to be similar to that of the parental cell lines.

When the cytochrome oxidase complex is not well assembled, the stability of its nuclear-coded subunits can be compromised. Mitochondria isolated from 143B cells and from two of the Hρ° × Or hybrids were disrupted by brief sonication, to release proteins that are either soluble or loosely associated with membranes, and then subjected to ultracentrifugation. Immunoblots of the starting mitochondria, pellets, and supernatants showed that in control cells, COX nuclear-coded proteins were detected almost exclusively in the pellet (fig. 4C ). A very small percentage of COX Va, but not COX IV, subunit, was present in the soluble fraction, and it may represent a small pool of unassembled subunits. In human ρ° × orangutan ρ+ hybrids, the subunit COX Va was partially recovered (∼30%) in the soluble protein fraction (fig. 4C ), indicating a labile association of subunits Va and IV (the latter being almost completely embedded in the phospholipid bilayer in intact complexes [Tsukihara et al. 1996] ). A similar analysis of protein subunits from the other four mitochondrial OXPHOS complexes showed no differences between hybrids and controls (fig. 4C ). The instability of nuclear-coded COX subunits in the hybrids seems to reflect a defective assembly of the complex subunits.

Mitochondrial Complex IV Deficiency in Human ρ° × Orangutan ρ+ Hybrids Is Not a Common Feature of Human ρ°–Nonhuman Ape Interspecific Hybrids

To determine if the interference between nuclear-coded factors was also present in other ape hybrids, human, gorilla, and orangutan fibroblasts were fused with the neomycin-resistant human 143B/206NEO ρ° cell line. Cells were selected in G418 (a neomycin analog) and uridine-lacking medium. After 13 days of selection, the efficiency of the fusions was determined by the number of growing clones and expressed as a ratio (%) to the clones obtained in the human ρ° × human fibroblast hybrid fusion. The efficiency for human ρ° × gorilla fibroblast fusions was 38.5 ± 10.9, and that for human ρ° × orangutan was 13.7 ± 5.0. After 25 days of selection, the percentages of clones which survived with respect to the number of colonies obtained from human ρ° × human fibroblast fusions dropped (28.2 ± 7.3 for human ρ° × gorilla, and 6.9 ± 3.8 for human ρ° × orangutan). These results indicate that a significant proportion of hybrid clones were unstable and died under prolonged culture conditions (∼27% for human ρ° × gorilla and ∼50% for human ρ° × orangutan, respectively). In contrast to Hρ° × Or hybrids, hybrids generated with human fibroblasts, and also with gorilla fibroblasts, exhibited endogenous cell respiration (fig. 3A ), steady-state levels of mtDNA coded complex IV subunits (fig. 4D ), and cytochrome oxidase activity (semiquantitatively determined by cytochemistry, not shown) which were comparable to those of the parental human cell line.

Discussion

Many Orangutan Chromosomes are Required to Restore OXPHOS Function in a Human Cell Harboring Orangutan mtDNA

Orangutan mtDNA and human nuclear genomes encode incompatible oxidative phosphorylation components (Kenyon and Moraes 1997 ). In the present report, the creation and characterization of human-orangutan somatic hybrids containing mtDNA only from orangutans, helped us to define the multifactorial and dominant nature of this incompatibility.

We tried to identify the factor(s) involved in orangutan mtDNA–human nuclear DNA incompatibilities by functional complementation. Since OXPHOS cannot be restored in a cell system consisting of a pure human nuclear background coexisting with orangutan mtDNA, it was reasonable to think that the respiratory capacity in that system could be reestablished if the human nucleus was supplemented with one or a few orangutan chromosomes. The rationale was to provide the system with the orangutan homolog(s) of the human factor(s) unable to interact properly with orangutan mtDNA or its products. We took advantage of our capacity to transfer in a single step mtDNA and a limited number of chromosomes from orangutan to human cells to create hybrid cell lines which could be selected in a medium selective for respiratory function. In half of our attempts, no respiring clones were obtained. This unusually low yield, compared with our previous experience with this technique (Barrientos and Moraes 1998 ), suggested that the orangutan chromosomal supplement necessary to maintain functional orangutan mtDNA in a human ρ° was more than one chromosome and that only in rare cases, when the infrequent microcells containing many chromosomes were fused to the human ρ° cells, were the right combination of chromosomes achieved. The FISH and microsatellite analyses of the hybrids obtained supported this assumption. In all clones with some respiratory function, the presence of a high number of orangutan chromosomes was observed. The generation of these hybrids, resembling complete hybrids constructed by fusing whole orangutan fibroblasts with human ρ° cells, would occur only when a large piece of nucleus containing several chromosomes or a whole nucleus escaped the several filtration steps. We speculate that when these large portions of nucleus surrounded by a thin ring of cytoplasm containing mitochondria were fused to a human ρ° cell, a hybrid with some respiratory capacity was created, probably explaining the low efficiency observed. In addition, the low level of respiration in the hybrids obtained suggests interference between nuclear-coded subunits of orangutan and human origins, creating new interactions that are less stable or efficient in the assembly and function of the multimeric respiratory enzymes. Our attempts to reduce the total number of chromosomes in the hybrids by treating the cells with the antimitotic agent colcemide failed to generate galactose-resistant clones with a significantly reduced number of orangutan chromosomes. Our results suggest that orangutan mtDNA requires interactions with several nuclear-coded factors that cannot be replaced by human genes. However, we cannot rule out the possibility that the right combination of chromosomes was not achieved in our experiments. In any case, our experiments showed that human chromosomes present in the hybrids have a dominant negative effect on the orangutan COX function.

Why Is Complex IV (COX) Deficient in Cells Containing Orangutan mtDNA and Chromosomes from Both Human and Orangutan?

Why was COX activity specifically affected in our hybrids? MtDNA-coded COX subunits are among the most conserved polypeptides in mitochondria, and the sequence similarity between human and orangutan subunits is higher than 90%. The rate of mtDNA evolution varies for each mitochondrial gene (Holmes 1991 ; Jukes 1994 ): for example, COX II has evolved at a high rate in primates (Adkins, Honeycutt, and Disotell 1996 ). Nevertheless, some COX nuclear-coded subunits, like subunit IV (Wu et al. 1997 ) and subunit VII (Schmidt, Goodman, and Grossman 1999 ), also had high rates of variation, illustrating the coevolutionary pressure associated with nuclear-mitochondrial interactions. Although at this point we cannot determine which specific factor(s) causes the COX assembly defect in the Hρ° × Or hybrid system, our results showed that this incompatibility can result from competition between nuclear-coded proteins.

Cytochrome c oxidase (a multimeric enzyme complex) resides in the mitochondrial inner membrane and catalyzes the final step of electron transfer through the respiratory chain. In eukaryotic organisms, COX is composed of up to 13 subunits encoded by both the mitochondrial (subunits I, II, and III, which form the catalytic core of the enzyme; Tsukihara et al. 1996 ) and the nuclear genomes. The enzyme comprises four redox-active metal centers (two heme and two copper centers), and proteins implicated in heme biosynthesis (Poyton and McEwen 1996 ) or copper homeostasis (Glerum, Shtanko, and Tzagoloff 1996 ) are also nuclear-coded. In addition, some nuclear-coded proteins are implicated in the transport of nuclear-coded subunits across mitochondrial membranes, and others mediate the assembly or stability of the enzyme (reviewed by Poyton 1998 ). COX assembly is not completely understood, and new proteins have recently been associated with this function, such as surf-1 (Mashkevich et al. 1997 ), a protein implicated in a human mitochondrial disease (Tiranti et al. 1998 ; Zhu et al. 1998 ). The complex appears to be initiated by the assembly of subunits I and IV, which is followed by the addition of most other subunits (Nijtmans et al. 1998 ).

In Hρ° × Or hybrids, mtDNA-coded COX subunits (COX I and COX II) were efficiently translated, but their steady-state levels were greatly reduced. The import into the mitochondria and expression of nuclear-coded COX subunits were not affected. These characteristics, together with a loose association observed between COX nuclear-coded proteins and mitochondrial membranes, are the hallmarks of cytochrome oxidase assembly-defective mutants, as has been described for yeast (Glerum and Tzagoloff 1997 ). In humans, a pathogenic missense mutation in COX II was shown to disturb the assembly of the holoenzyme (Rahman et al. 1999 ). Nonsense mutations in the COX I gene also disrupt COX assembly (Bruno et al. 1999 ; Comi et al. 1998 ). Moreover, most patients with undefined COX deficiency also show a decrease in the steady-state levels of several subunits, suggesting an assembly defect (Taanman et al. 1996 ).

In Hρ° × Or hybrids, the mitochondrial membranes could have several versions of complex IV: (1) formed exclusively by the orangutan subunits, (2) constructed with human nuclear subunits and the orangutan mtDNA-coded subunits, and (3) built with nuclear-coded subunits of both orangutan and human origins. Probably, most of the complex IV present in the hybrids falls into the last category. It is possible that the residual COX activity measured in the Hρ° × Or hybrids is derived from the first class of enzyme (orangutan enzyme) or even some from the third class with mostly orangutan subunits. The nuclear subunits cross-interact at different regions of the complex (e.g., COX IV interacts with COX Va, which interacts with COX VIc), and if those interactions are defective, the assembly of the enzyme can be compromised, as has been shown for yeast (Glerum and Tzagoloff 1998 ). We can speculate that most of the assembly-defective complexes are associated with defective interactions caused by the dominant negative effect of human nuclear-coded subunits. Mitochondrial DNA from humans, gorillas, and chimpanzees were able to produce HXC with normal COX activity. Amino acid changes in mtDNA-coded COX subunits in gorillas or chimpanzees have to be considered neutral polymorphisms, as they do not affect the activity of the enzyme. Therefore, changes in the orangutan COX subunits that can be considered potentially deleterious for COX assembly are those differing from the other three ape species (table 3 ). COX I is located mainly in the transmembrane domain of the enzyme complex, consisting of 12 transmembrane helices and without any large extramembrane portion (Tsukihara et al. 1996 ). There are 17 amino acid changes between COX I from the orangutan and its human, chimpanzee, or gorilla homologs, 9 of which are conservative changes which do not affect hydrophobicity and most probably do not affect interactions between COX I and other subunits. There are no changes in the first transmembrane helix of COX I, which interacts with COX VIIc and COX VIII, and no significant changes in transmembrane helix IX, which interacts with COX V, or in helix XII, which interacts with COX IV. More stringent changes are those from Y to H (in amino acid 260; helix VI), from L to P (in amino acid 483), from E to Q (in amino acid 487), and from S to P (in amino acid 513) in the CH2 terminal extramembrane part after helix XII, which is located in the matrix side. COX II contains two transmembrane helices. Helix I interacts with COX VIc. The single change observed in this helix (from I to V in amino acid 21) is conservative. The large extramembrane domain of COX II is located above COX I in the cytosolic side, having a barrel structure that holds the CuA site. The A-to-T change in amino acid 164 probably will not alter the binding of the metal ion per se, but it could affect the interaction between COX II and other nuclear-coded proteins implicated in the delivery of Cu++ to the COX (e.g., Sco1). COX III contains seven transmembrane helices without extensive extramembrane domains. COX VIIa cross-interacts with its helices I and II. Two changes occurred in helix II (from M/T to L in amino acid 44 and from L to T in amino acid 45), which could well be the result of an inversion of those two amino acids. There is a single but less conservative change (from S to A in amino acid 135) in helix IV of COX III, which interacts with COX VIa. The NH2 terminal of COX VIa is in an extended conformation in the transmembrane region that makes contact with helices V and VII of COX I from the other monomer. These contacts seem likely to stabilize the dimeric structure (Tsukihara et al. 1996 ). Changes in COX III (the one in helix IV) and several stringent changes in helix VI (table 3 ) could affect the interaction with COX VIa and the stability of the dimeric structure of the enzyme.

Human × primate hybrids generated with human and gorilla fibroblasts did not show COX deficiency. This is not surprising due to the high similarity between human and gorilla mtDNA-coded COX subunits (table 3 ). Essentially all human-orangutan differences between amino acids in these subunits were also observed in the gorilla-orangutan comparison (table 3 ). However, the efficiency of human-gorilla hybrid formation under selection for OXPHOS function was less than half that of a human-human control, suggesting that a milder functional interference may also take place in some step involved in the biogenesis or function of OXPHOS components. We were somehow surprised that complex I was not affected in the hybrids, as it was partially deficient in HXC harboring chimpanzee or gorilla mtDNAs (Barrientos, Kenyon, and Moraes 1998 ). We can only speculate that human nuclear-coded subunits of complex I that do not interact well with other primates' mtDNA-coded counterparts are unable to assemble in human × orangutan hybrids without preventing complex formation of compatible subunits.

In conclusion, the cellular hybrid system presented here showed that COX assembly in mammals is very sensitive to small changes in amino acid sequences not involved in catalytic function and that even minimal variations in mtDNA-coded subunits must be compensated by a change in nuclear-coded subunits.

Table 1 Molecular Cytogenic Characterization of Human {ρ}°–Orangutan Hybrids M1, M22, M23, and H4 by RxFISH-iCGH

Table 1 Molecular Cytogenic Characterization of Human {ρ}°–Orangutan Hybrids M1, M22, M23, and H4 by RxFISH-iCGH

Table 2 Identification of Chromosomes from Orangutan Retained in Human {ρ}° × Orangutan {ρ}+ Hybrids

Table 2 Identification of Chromosomes from Orangutan Retained in Human {ρ}° × Orangutan {ρ}+ Hybrids

Table 3 Comparison of Amino Acid (AA) Variations of the Three mtDNA-Coded COX Subunits Between Orangutans (Orang) and Humans (H), Chimpanzees (C), and Gorillas (G)

Table 3 Comparison of Amino Acid (AA) Variations of the Three mtDNA-Coded COX Subunits Between Orangutans (Orang) and Humans (H), Chimpanzees (C), and Gorillas (G)

Fig. 1.—Features of the mtDNA present in human ρ° × orangutan ρ+ hybrids. A, The origin of the mtDNA in the hybrids was ascertained by Southern blot analysis using total DNA digested with PvuII. The PvuII restriction patterns differ between human mtDNA (a fragment of 16.5 kb) and orangutan mtDNA (two fragments of 8 and 8.3 kb, only one of them detected with the partial mtDNA probe used). In all hybrids, mtDNA was of orangutan origin. B, To determine the orangutan mtDNA levels in Hρ° × Or hybrid clones, the mtDNA content relative to the nDNA (18S gene) was quantified by a slot blot experiment (see Materials and Methods). All hybrids had mtDNA/nDNA ratios between those observed in orangutan fibroblast and 143B

> >Fig. 2.—Chromosomal origin identification in human ρ° × orangutan ρ+ hybrids. A, Molecular cytogenetic analysis by RxFISH-iCGH (see Materials and Methods for details); binary display of a single representative metaphase. The RxFISH multicolor banding pattern allowed the identification of approximately 80% of all individual chromosomes (left, h = human, o = orangutan, dic = dicentric chromosome). B, The iCGH hybridization pattern used to determine the species origin (orangutan chromosomes are shown in red, human chromosomes in green). C, Informative human microsatellite markers were used to distinguish chromosomes from human ρ° and orangutan fibroblasts and to identify the chromosomal content of hybrid cells (see Materials and Methods). The figure shows an example for the analyses of loci D9S934 (top) and D18S851 (bottom)

Fig. 3.—Mitochondrial respiratory chain functional studies on human ρ° × orangutan ρ+ hybrids. A, Polarographic assay of intact cell respiration. The bars represent percentage of oxygen consumption rate of the 143B parental cell line. Cell respiration was significantly reduced in Hρ° × Or hybrids compared with the parental cell lines. B, Spectrophotometrical assays of mitochondrial respiratory chain enzyme activities in isolated mitochondria. Specific activities were normalized to the citrate synthase activity (CS). Assays included complex I (CI), complex II+III (CII+III), and complex IV (CIV). * Values statistically different from the control values, with P < 0.050, according to Student's t-test for equality of means for independent samples. Error bars represent standard deviation (n ≥ 3). C, KCN titration of complex IV activity. Data are represented in an Eadie plot, in which the x-intercept represents the −Ki value. The Ki values were similar for all cell lines

Fig. 4.—Expression of COX subunits in human ρ° × orangutan ρ+ hybrids. A, Pulse-labeling of mitochondrial translation products in whole cells in the presence of emetine. B, Western analysis of COX subunits in mitochondrial samples using antibodies against different mtDNA-coded (COX I and COX II) or nDNA-coded (COX IV and COX Va) COX subunits and against subunit 1 of the NADH dehydrogenase (ND1). C, Immunoblotting analysis of COX subunits in mitochondria (M) isolated from 143B cells and from one of the Hρ° × Or hybrids, disrupted by sonication. After ultracentrifugation, a pellet (P-100) and a supernatant (S-100) were obtained and analyzed. In the Hρ° × Or hybrids, subunit COX Va was partially recovered in the soluble protein fraction. The proteins are identified on the right. D, Expression of COX subunits. Only Hρ° × Or hybrids, but not the ones constructed with fibroblasts from humans or gorillas, showed decreased expression of mtDNA-coded COX subunits

This work was supported by National Institutes of Health grant GM55766. We thank Habbib Chaudhary for his technical assistance with the KCN titration experiments. We are indebted to Drs. Roderick Capaldi (Institute of Molecular Biology, University of Oregon, Eugene) for several antibodies and critical comments, Dr. Giovanni Manfredi (Cornell University) for critical comments, and Anne Lombes for the anti-ND1 antibody.

literature cited

Adkins, R. M., R. L. Honeycutt, and T. R. Disotell.

. Evolution of eutherian cytochrome c oxidase subunit II: heterogeneous rates of protein evolution and altered interaction with cytochrome c. :–1404

Arnason, U., A. Gullberg, A. Janke, and X. Xu.

. Pattern and timing of evolutionary divergences among hominoids based on analyses of complete mtDNAs. :–661

Barrientos, A., L. Kenyon, and C. T. Moraes.

. Human xenomitochondrial cybrids.

Cellular models of mitochondrial complex I deficiency. J. Biol. Chem.

:–14217

Barrientos, A., and C. T. Moraes.

. Simultaneous transfer of mitochondrial DNA and single chromosomes in somatic cells: a novel approach for the study of defects in nuclear-mitochondrial communication. :–1808

———.

. Titrating the effects of mitochondrial complex I impairment in the cell physiology. :–16197

Bradford, M. M.

. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. :–254

Brown, W. M., M. George Jr., and A. C. Wilson.

. Rapid evolution of animal mitochondrial DNA. Proc. Natl. Acad. Sci. USA 76:1967–1971

Bruno, C., A. Martinuzzi, Y. Tang et al. (11 co-authors).

. A stop-codon mutation in the human mtDNA cytochrome c oxidase I gene disrupts the functional structure of complex IV. :–620

Chomyn, A.

. In vivo labeling and analysis of human mitochondrial translation products. :–211

Clayton, D. A., R. L. Teplitz, M. Nabholz, H. Dovey, and W. Bodmer.

. Mitochondrial DNA of human-mouse cell hybrids. Nature 234:560–562

Comi, G. P., A. Bordoni, S. Salani et al. (14 co-authors).

. Cytochrome c oxidase subunit I microdeletion in a patient with motor neuron disease. :–116

Coote, T., and M. W. Bruford.

. Human microsatellites applicable for analysis of genetic variation in apes and Old World monkeys. :–410

De Francesco, L., G. Attardi, and C. M. Croce.

. Uniparental propagation of mitochondrial DNA in mouse-human cell hybrids. Proc. Natl. Acad. Sci. USA 77:4079–4083

Giles, R. E., I. Stroynowski, and D. C. Wallace.

. Characterization of mitochondrial DNA in chloramphenicol-resistant interspecific hybrids and a cybrid. :–554

Glerum, D. M., A. Shtanko, and A. Tzagoloff.

. Characterization of COX17, a yeast gene involved in copper metabolism and assembly of cytochrome oxidase. :–14509

Glerum, D. M., and A. Tzagoloff.

. Submitochondrial distributions and stabilities of subunits 4, 5, and 6 of yeast cytochrome oxidase in assembly defective mutants. :–414

———.

. Affinity purification of yeast cytochrome oxidase with biotinylated subunits 4, 5, or 6. :–44

Hao, H., and C. T. Moraes.

. A disease-associated G5703A mutation in human mitochondrial DNA causes a conformational change and a marked decrease in steady-state levels of mitochondrial tRNA(Asn). :–6837

Hayashi, J., Y. Tagashira, M. C. Yoshida, K. Ajiro, and T. Sekiguchi.

. Two distinct types of mitochondrial DNA segregation in mouse-rat hybrid cells.

Stochastic segregation and chromosome-dependent segregation. Exp. Cell Res.

:–61

Holmes, E. C.

. Different rates of substitution may produce different phylogenies of the eutherian mammals. :–215

Jukes, T. H.

. Divergent proteins and views [letter]. Nature 371:734

Kallioniemi, A., O. P. Kallioniemi, D. Sudar, D. Rutovitz, J. W. Gray, F. Waldman, and D. Pinkel.

. Comparative genomic hybridization for molecular cytogenetic analysis of solid tumors. Science 258:818–821

Kenyon, L., and C. T. Moraes.

. Expanding the functional human mitochondrial DNA database by the establishment of primate xenomitochondrial cybrids. Proc. Natl. Acad. Sci. USA 94:9131–9135

King, M. P., and G. Attardi.

. Human cells lacking mtDNA: repopulation with exogenous mitochondria by complementation. Science 246:500–503

Marusich, M. F., B. H. Robinson, J. W. Taanman, S. J. Kim, R. Schillace, J. L. Smith, and R. A. Capaldi.

. Expression of mtDNA and nDNA encoded respiratory chain proteins in chemically and genetically-derived Rho0 human fibroblasts: a comparison of subunit proteins in normal fibroblasts treated with ethidium bromide and fibroblasts from a patient with mtDNA depletion syndrome. Biochim. Biophys. Acta 1362:145–159

Mashkevich, G., B. Repetto, D. M. Glerum, C. Jin, and A. Tzagoloff.

. SHY1, the yeast homolog of the mammalian SURF-1 gene, encodes a mitochondrial protein required for respiration. :–14364

Müller, S., P. C. O'Brien, M. A. Ferguson-Smith, and J. Wienberg.

. Cross-species colour segmenting: a novel tool in human karyotype analysis. Cytometry 33:445–452

Nijtmans, L. G., J. W. Taanman, A. O. Muijsers, D. Speijer, and C. Van den Bogert.

. Assembly of cytochrome-c oxidase in cultured human cells. :–394

Poyton, R. O.

. Assembling a time bomb—cytochrome c oxidase and disease [news; comment]. :–317

Poyton, R. O., and J. E. McEwen.

. Crosstalk between nuclear and mitochondrial genomes. :–607

Rahman, S., J. W. Taanman, J. M. Cooper et al. (12 co-authors).

. A missense mutation of cytochrome oxidase subunit II causes defective assembly and myopathy. :–1039

Robinson, B. H.

. Use of fibroblast and lymphoblast cultures for detection of respiratory chain defects. :–464

Schmidt, T. R., M. Goodman, and L. I. Grossman.

. Molecular evolution of the COX7A gene family in primates. :–626

Seligman, A. M., M. J. Karnovsky, H. L. Wasserkrug, and J. S. Hanker.

. Nondroplet ultrastructural demonstration of cytochrome oxidase activity with a polymerizing osmiophilic reagent, diaminobenzidine (DAB). :–14

Southern, P. J., and P. Berg.

. Transformation of mammalian cells to antibiotic resistance with a bacterial gene under control of the SV40 early region promoter. :–341

Speicher, M. R., S. Gwyn Ballard, and D. C. Ward.

. Karyotyping human chromosomes by combinatorial multi-fluor FISH. :–375

Taanman, J. W., M. D. Burton, M. F. Marusich, N. G. Kennaway, and R. A. Capaldi.

. Subunit specific monoclonal antibodies show different steady-state levels of various cytochrome-c oxidase subunits in chronic progressive external ophthalmoplegia. Biochim. Biophys. Acta 1315:199–207

Tiranti, V., K. Hoertnagel, R. Carrozzo et al. (18 co-authors).

. Mutations of SURF-1 in Leigh disease associated with cytochrome c oxidase deficiency. :–1621

Tsukihara, T., H. Aoyama, E. Yamashita, T. Tomizaki, H. Yamaguchi, K. Shinzawa-Itoh, R. Nakashima, R. Yaono, and S. Yoshikawa.

How many amino acid differences were there between the human and the monkey?

How many amino acids does a chimpanzee have?

How many genes are different between humans and chimps?

How many differences in amino acids of the section of cytochrome c are found between humans and?