COX15 Codes for a Mitochondrial Protein Essential for the Assembly of Yeast Cytochrome Oxidase*

(Received for publication, April 28, 1997, and in revised form, May 29, 1997)

D. Moira Glerum Dagger , Ivor Muroff , Can Jin and Alexander Tzagoloff §

From the Department of Biological Sciences, Columbia University, New York, New York 10027

ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS AND DISCUSSION
FOOTNOTES
REFERENCES


ABSTRACT

The respiratory defect of Saccharomyces cerevisiae mutants assigned to complementation group G4 of a pet strain collection stems from their failure to synthesize cytochrome oxidase. The mutations do not affect expression of either the mitochondrially or nuclearly encoded subunits of the enzyme. The cytochrome oxidase deficiency also does not appear to be related to mitochondrial copper metabolism or heme a biosynthesis. These data suggest that the mutants are likely to be impaired in assembly of the enzyme. A gene designated COX15 has been cloned by transformation of mutants from complementation group G4. This gene is identical to reading frame YER141w on chromosome 5. To facilitate further studies, Cox15p has been expressed as a biotinylated protein. Biotinylated Cox15p fully restores cytochrome oxidase in cox15 mutants, indicating that the carboxyl-terminal sequence with biotin does not affect its function. Cox15p is a constituent of the mitochondrial inner membrane and, because of its resistance to proteolysis, probably is largely embedded in the phospholipid bilayer of the membrane. The present studies further emphasize the complexity of cytochrome oxidase assembly and report a new constituent of mitochondria involved in this process. The existence of COX15 homologs in Schizosaccharomyces pombe and Caenorhabditis elegans suggests that it may be widely distributed in eucaryotic organisms.


INTRODUCTION

The completion of the Saccharomyces cerevisiae genome sequence (1) has spawned new large scale projects designed to unravel the functions of the numerous unknown reading frames. A substantial fraction of the total information in yeast chromosomal DNA is comprised of PET genes that are essential for the biogenesis of respiratory competent mitochondria. Mutations in these genes were shown in the early 1950s (2, 3) to affect the ability of yeast to respire. Renewed efforts to mutationally saturate for this class of genes (4, 5) have helped to expand our knowledge of the extent and nature of the contribution made by the nucleus toward maintenance of respiring mitochondria.

Biochemical studies of pet mutants1 have revealed that the assembly of respiratory chain enzymes is governed by an unexpectedly large number of genes. For example, some three dozen complementation groups have been reported to consist of mutants displaying a selective deficiency in cytochrome oxidase (4, 6). In addition to mutations in the structural genes, these strains are also affected in: 1) processing of the mitochondrial cytochrome oxidase pre-mRNAs (7-9), 2) translation of the resultant mRNAs (10, 11), 3) heme a biosynthesis (12), 4) copper import and transfer to the apoenzyme (13, 14), and 5) as yet poorly understood events in the pathway leading to the functional enzyme (15-17). To learn more about the assembly of this heteroligomeric membrane complex, we have continued to screen for, and to analyze pet mutants with lesions in cytochrome oxidase. In this article, we report the properties of mutants from complementation group G4 whose cytochrome oxidase deficiency results from mutations in COX15 (identical to reading frame YER141w on chromosome 5). We show that mutations in this gene have no effect on the expression of the mitochondrial or nuclear cytochrome oxidase genes. Since cox15 mutants also do not appear to be impaired in heme a synthesis and copper metabolism, the product of this gene (Cox15p) is most likely involved in some aspect of the assembly process itself. We also present evidence that Cox15p is imported into mitochondria and is a constituent of the inner membrane.


MATERIALS AND METHODS

Yeast Strains and Media

The genotypes and sources of the strains of S. cerevisiae used in this study are listed in Table I. The compositions of the media for growth of yeast have been described elsewhere (21).

Table I. Genotypes and sources of yeast strains


Strain Genotype Source

CB11 a ade1 Ref. 18
W303-1A a ade2-1 his3-1,15 leu2-3,112 trp1-1 ura3-1  ---a
W303-1B  alpha ade2-1 his3-1,15 leu2-3,112 trp1-1 ura3-1  ---a
LL20  alpha his3 leu2 Ref. 19
N7-211  alpha met6 cox15 Ref. 20
aN7-211 a ade1 cox15 N7-211 × CB11
N7-211/L1  alpha leu2 cox15 aN7-211 × LL20
C4  alpha met6 cox15-1 Ref. 4
aE412 a ade1 cox15 This study
E412/L1  alpha leu2 cox15 aE412 × LL20
aW303Delta COX15 a ade2-1 his3-1,15 leu2-3,112 trp1-1 ura3-1 cox15::HIS3 This study
LL20Delta COX15  alpha his3-1,15 leu2-3,112 cox15::LEU2 This study
WLDelta COX15 a his3-1,15 leu2-3,112 ura3-1 cox15::LEU2 LL20Delta COX15 × W303-1A
WLDelta COX15/Int a his3-1,15 leu2-3,112 cox15::LEU2 ura3::COX15-BIO This study

a Dr. Rodney Rothstein, Department of Human Genetics, Columbia University.

Preparation of Yeast Mitochondria

Unless otherwise indicated, wild-type and mutant yeast were grown to stationary phase in YPGal (2% galactose, 1% yeast extract, and 2% peptone) and mitochondria were prepared by the procedure of Faye et al. (22), except that Glusulase was replaced by Zymolyase 20,000 (ICN Biomedicals, Inc.) to convert cells to spheroplasts. Difference spectra of reduced versus oxidized cytochromes of mitochondrial extracts were taken at room temperature (20). Cytochrome oxidase activity was measured as described previously (20).

Extraction and Separation of Mitochondrial Hemes

Mitochondria were suspended in 0.5 M sorbitol, 20 mM Tris-HCl, pH 7.5, 0.5 mM EDTA at a protein concentration of 15-20 mg/ml. Extraction of mitochondrial hemes was achieved by addition of mitochondria to acidified acetone such that the final concentration of protein was 3 mg/ml in 83% acetone (acetone:HCl = 97.5:2.5). The mixture was vortexed for 30 s and centrifuged for 5 min in a Beckman Microfuge. The clarified extract was collected and mixed with an equal volume of 50% acetonitrile. Insoluble material formed was removed by a second centrifugation. The extract was then neutralized with 1.65 M ammonium hydroxide. The neutralized extract equivalent of 1.5 mg of mitochondrial protein was applied to a Phenomenex (Torrance, CA) 3.9 × 300-mm C18 Bondclone column. The column was developed by a modification of the method of Sone and Fujiwara (23). Hemes were eluted at a flow rate of 1 ml/min using a 30-50% acetonitrile gradient over the first 5 ml, followed by a 50-75% linear acetonitrile gradient over the subsequent 35 ml. The acetonitrile solutions contained 0.05% trifluoroacetic acid. The elution times of protoheme, heme a, and heme o were determined prior to each set of runs with standards. The elution of heme compounds was monitored at 400 nm.

Cloning of the COX15

The wild-type COX15 gene was cloned by transformation of E412/L1 and N7-211/L1 with a yeast genomic library by the method of Beggs (24). The library used for the transformation was constructed from partial Sau3A fragments of nuclear DNA (averaging 7-10 kb) from the respiratory competent haploid strain S. cerevisiae D273-10B/A1. The purified DNA fragments were ligated to the BamHI site of YEp13 (25). E412/L1 or N7-211/L1 were grown in YPGal to early stationary phase and approximately 108 cells were transformed with 50 µg of plasmid DNA. Part of the transformation mixture (10%) was plated on minimal glucose to estimate the total number of plasmid-bearing transformants, and the rest of the cells were plated on rich glycerol medium (2% peptone, 1% yeast extract, 3% glycerol, 2% ethanol, 1.2 M sorbitol) to select respiratory competent clones. Sixteen transformants whose respiratory competency and leucine prototrophy co-segregated were used to isolate and characterize the complementing plasmids.

Construction of COX15-BIO Fusion Gene

A gene expressing biotinylated Cox15p was constructed by introducing a XbaI site 12 nucleotides before the termination codon. The polymerase chain reaction primer complementary to the sense strand was 5'-ttataatggtctagaggctaactt. The second polymerase chain reaction primer was identical to the sense strand except for two nucleotides. This primer (5'-tcgtttattaagcttcctattgta) introduced a HindIII site 72 nucleotides upstream of the COX15 initiation codon. The 1.6-kb fragment amplified from pG4/ST3 was transferred to YEp352-Bio7 (17). The resultant plasmid (pG4/ST10) contained an in-frame fusion of COX15 to a 270-nucleotide long sequence coding for the biotinylation site of bacterial transcarboxylase (26). The fusion gene was transferred from the multicopy plasmid to the integrative plasmid YIp352 (27). After linearization at the NcoI site in the plasmid-borne URA3 gene, the COX15 fusion gene sequence was integrated at the ura3 locus of the cox15 mutant aWLDelta COX15.

Disruption of COX15

COX15 was disrupted at the internal BglII site, either with a 1.8-kb BamHI fragment containing HIS3, or alternatively, with a 3-kb BglII fragment containing LEU2. In both cases, pG4/ST4 was digested with BglII and the linearized plasmid ligated with the respective disruptors. The mutant alleles were recovered as linear SacI fragments and used to transform either W303-1A (with cox15::HIS3) or LL20 (with cox15::LEU2) by the one-step gene replacement method (28). Respiratory defective transformants prototrophic for either of the markers were crossed to a cox15 mutant and to a rho o tester strain. Integration of the disrupted alleles in chromosomal DNA was verified genetically by the crosses to the cox15 mutants, and in the case of the cox15::LEU2 allele, also by Southern blot analysis of genomic DNA.

Miscellaneous Procedures

Standard procedures were used for the preparation and ligation of DNA fragments, and for transformation and recovery of plasmid DNA from Escherichia coli (29). The preparation of yeast nuclear DNA and the conditions for the Southern hybridizations, were as described by Myers et al. (30). DNA was sequenced by the method of Maxam and Gilbert (31). Proteins were separated by PAGE either in the buffer system of Laemmli (32) or by the method of Schagger and von Jagow (33). Immunodetection of proteins on Western blot was carried out with 125I-protein A (34). Protein concentrations were determined by the method of Lowry et al. (35).


RESULTS AND DISCUSSION

Phenotype of G4 Mutants

Approximately 70 independent mutants have been assigned to complementation group G4 of a pet collection (4). G4 mutants are respiratory deficient as a result of recessive mutations in a nuclear gene, causing a specific deficiency in cytochrome oxidase. The enzymatic properties of N7-211, a representative of this complementation group have already been reported in an earlier mutant screen (20). Based on assays of respiratory chain enzymes and spectral analysis of mitochondrial cytochromes, this mutant was shown to have a specific cytochrome oxidase deficiency (20).

To better understand the biochemical basis for this phenotype, we have carried out a more complete analysis of cytochrome oxidase proteins. The failure of G4 mutants to express cytochrome oxidase is not due to a block of transcription or processing of the mitochondrially encoded (36) pre-mRNAs for subunits 1-3, or translation of the resultant mRNA. This is borne out by in vivo pulse labeling of mitochondrially synthesized proteins in whole cells poisoned with cycloheximide to inhibit cytoplasmic protein synthesis. Under these conditions, incorporation of [35S]methionine into subunits 1, 2, and 3 is as efficient in the G4 mutant as in wild-type (Fig. 1). Even though synthesis of these endogenously produced proteins appears to be normal, their steady-state concentrations are dramatically lower in the mutants. This is evident from the immunoblot analysis of subunits 1 and 2 in mitochondria of aW303Delta COX15 (see below for construction of this mutant) shown in Fig. 2. Subunit 3 could not be tested because of a lack of a suitable antibody. Proteolytic degradation of these hydrophobic membrane proteins is commonly observed in mutants unable to complete assembly of the enzyme (13, 14, 17).


Fig. 1. In vivo labeling of mitochondrial translation products. The parental strain D273-10B/A1 and the G4 mutant C4 were labeled with [35S]methionine in the presence of cycloheximide (20). Mitochondria were prepared and samples representing 50 µg of protein were separated by PAGE on a 7.5-15% polyacrylamide gel (37). The radioactively labeled proteins are identified in the margin: Var1, cytochrome b (B), subunits 1, 2, and 3 of cytochrome oxidase (CO1, CO2, and CO3), subunits 6 and 9 of mitochondrial ATPase (ATP6 and ATP9).
[View Larger Version of this Image (28K GIF file)]


Fig. 2. Western analysis of cytochrome oxidase subunits in wild-type and a cox15 mutant. Mitochondria were prepared from the wild-type strain W303-1A and from the cox15 mutant aW303Delta COX15. In the upper panel total mitochondrial proteins (10 µg) were separated on 12% polyacrylamide gels. The proteins were transferred to nitrocellulose and probed with antisera which detects subunits 1 and 2 (CO1 and CO2), but not subunit 3. In the lower panel, wild-type and mutant mitochondria were disrupted by a brief sonic treatment and centrifuged at 300,000 × gav for 15 min to separate the soluble proteins from the membrane vesicles which were resuspended in the starting volume of buffer. Equal volumes of mitochondria (M), soluble extract (E), and membranes (P) were separated on a 16.5% polyacrylamide gel (33). Following electrophoretic transfer to nitrocellulose subunits 4-8 (CO4-CO8) were detected with an antibody against a mixture of the low molecular weight subunits. The cytochrome oxidase subunits are identified in the margin.
[View Larger Version of this Image (28K GIF file)]

Since the synthesis of cytochrome oxidase depends on the expression of six nuclearly encoded subunits, the steady-state concentrations of these constituents were also compared in mitochondria from aW303Delta COX15 and the parental strain W303-1A. These imported components of the enzyme were examined in intact mitochondria, in submitochondrial particles, and the soluble protein fraction released from sonically disrupted mitochondria. The antibodies used for these analyses detect subunits 4, 5, 6, and the smaller polypeptides (subunits 7, 8, and 8a). The latter, however, are not resolved in the gel and migrate as a single composite band. The results of the immunoblots shown in Fig. 2 indicate that the mitochondrial content of the imported subunits in the mutant is not significantly different from the wild-type (Fig. 2).

In other studies we have noted that a sizeable fraction of subunits 4 and 6 are recovered in the soluble protein fraction following sonic disruption of mitochondria.2 The soluble forms of subunits 4 and 6 probably represent a mitochondrial pool of unassembled subunits. The relative distribution of subunits 4, 5, and 6 in the soluble and membrane fractions revealed an increase in the amounts of unassembled subunits 4 and 6, but not of subunit 5 which co-fractionates exclusively with the submitochondrial vesicles. The increase of soluble subunit 4 in the mutant is explained by the depletion of subunit 1, with which it is associated in the enzyme complex (38). X-ray crystallographic data indicate that in the bovine enzyme, subunit 6 is complexed to subunit 5, the latter being almost completely embedded in the phospholipid bilayer (38). The increase in soluble subunit 6 is surprising in view of the normal concentration of subunit 5 in the mutant, all of which is in the membrane. The recovery of a larger fraction of subunit 6 in the soluble fraction suggests a more labile association of subunits 5 and 6 when the normal complex is not formed. The presence in the G4 mutant of subunits 4, 5, and 6 (this probably also applies for the low-molecular weight subunits that are not resolved by PAGE) as mature proteins, and at concentrations comparable to those seen in wild-type mitochondria, argues against an effect of the mutation on expression, transport, or processing of these polypeptides.

Cytochrome oxidase utilizes heme a and copper as obligatory electron carriers (39). In addition to their functional roles, these prosthetic groups are required for the assembly of a stable enzyme. Mutations interfering with heme a synthesis (40) or mitochondrial copper metabolism (13, 14) lead to a deficiency of cytochrome oxidase activity, concomitant with proteolysis of subunits 1 and 2 of the enzyme. The enhanced turnover of these proteins under copper or heme a limiting conditions is probably a consequence of a more open and less stable quaternary structure.

The cytochrome oxidase deficiency of some mutants impaired in copper delivery to mitochondria can be suppressed by high copper concentrations in the medium (13). The failure of copper to rescue G4 mutants tends to argue against, but does not exclude, a defect in mitochondrial copper transport or transfer to the apoenzyme. Similarly, the presence of heme a, albeit at low concentrations in the mutant mitochondria (data not shown) made it unlikely that the lesion is related to the synthesis of this prosthetic group.

Cloning and Sequencing of COX15

A gene capable of conferring wild-type growth to G4 mutants on non-fermentable carbon sources was cloned by transformation of N7-211/L1 and E412/L1 with a genomic library of yeast nuclear DNA constructed in the shuttle vector YEp13 (25). The plasmids recovered from the different respiratory competent yeast transformants had either identical or overlapping fragments of genomic DNA. The recombinant plasmid pG4/T2 (Fig. 3) was used to subclone and localize the gene. The insert of this plasmid was dissected at an internal BamHI site into two fragments. Of the two, only the shorter 2.3-kb fragment starting at the internal BamHI site, and extending to the distal end of the insert, complemented the mutants (Fig. 3). Further subcloning of this region indicated that the internal BglII and PstI sites were likely to be in the gene (see pG4/ST3-ST7 in Fig. 3).


Fig. 3. Restriction map of pG4/T2 and derivative plasmids. The restriction map of the nuclear DNA fragment in pG4/T2 is shown as a linear insert in the YEp13 shuttle vector. The locations of the HindIII (H), BamHI (B), PstI (P), EcoRI (E), BglII (G), and SacI (Sa) are shown. The unique SphI (Sp) site of YEp13 is marked for orientation. The location and direction of transcription of COX15 is depicted by the solid arrow. The different regions subcloned in YEp351 (27) are indicated by the bars above the map of pG4/T2. Complementation, or lack thereof, of cox15 mutants by the different plasmids is indicated by the plus and minus signs, respectively.
[View Larger Version of this Image (11K GIF file)]

The 2.3-kb insert of pG4/ST3 was sequenced by the method of Maxam and Gilbert (31). The nucleotide sequence revealed two open reading frames, one of which included the BglII and PstI sites inferred from the subcloning results to be in the gene. This open reading frame is located on chromosome 5 next to, and in an orientation opposite to, MAG (3-methyladenine DNA glycosylase) (41). In keeping with the nomenclature adopted to designate genes required for expression of cytochrome oxidase, this gene has been designated COX15 (GenBank accession L38643).

COX15 codes for a protein of Mr = 54,667. The amino-terminal 76 residues constitute a hydrophilic sequence with a preponderance of basic amino acids. This sequence, which is not conserved in two Cox15p homologs (see below), probably serves as a mitochondrial import signal. The amino acid composition and the primary sequence of the protein suggests a hydrophobic protein.

Two homologs of Cox15p were found in current protein data bases. One is an unidentified protein of the fission yeast Schizosaccharomyces pombe (GenBank accession Z70043); the other homolog is a protein of Caenorhabditis elegans (GenBank accession Z49130). An alignment of the three proteins indicates 28% identity between the three proteins in an 375-amino acid span (Fig. 4). The S. pombe protein contains a carboxyl-terminal extension of 150 residues that is absent in the other proteins. The homology of the different Cox15p's is also evident from their hydropathy profiles. The yeast and nematode proteins have 7-8 analogously spaced hydrophobic domains of sufficient lengths to span a phospholipid bilayer (shown for the S. cerevisiae Cox15p in Fig. 4).


Fig. 4. Homology of yeast and nematode Cox15ps. The sequences of the unknown reading frames from S. pombe and C. elegans have been aligned with the S. cerevisiae Cox15p by the motif recognition algorithm of Vingron and Argos (42). Only the most conserved regions of the three proteins are shown. Identical amino acids in the alignment have been boxed. The hydropathy profile of the S. cerevisiae Cox15p is shown in the lower part of the figure. The method of Kyte and Doolittle (43) was used to calculate the hydrophobicity. The putative transmembrane sectors are shadowed.
[View Larger Version of this Image (62K GIF file)]

In Situ Disruption of COX15

COX15 was disrupted at the internal BamHI site, either with a 3-kb BglII fragment containing LEU2 or with a 1.8-kb BamHI fragment containing HIS3. The LEU2 and HIS3 disrupted alleles were substituted for the normal chromosomal copies of the gene in the respiratory competent haploid strains LL20 (20) and W303-1A, respectively. Respiratory deficient transformants were verified to be complemented by a rho o tester strain but not by G4 mutants, confirming the presence of the disruption alleles. In the case of the disruption in LL20, this was also shown directly by Southern analysis of genomic DNA.

The spectral properties (Fig. 5) and enzyme activities of mitochondria indicated the mutant constructs to lack cytochrome oxidase but not other components of the respiratory chain. The cytochrome oxidase deficiency of the cox15 mutant constructs and their failure to be complemented by G4 mutants argues that the latter have mutations in COX15.


Fig. 5. Spectra of mitochondrial cytochromes in mutants and transformants. Mitochondria were prepared from the respiratory competent strain W303-1A, from two cox15 mutants (aW303Delta COX15 and WLDelta COX15), and from the transformant WLDelta COX15/Int, which harbors a chromosomally integrated copy of COX15 fused to a bacterial sequence encoding a biotinylation signal. Mitochondria at a protein concentration of 5-7 mg/ml were extracted with potassium deoxycholate under conditions that quantitatively solubilize all the cytochromes (20). Difference spectra of extracts reduced with sodium hydrosulfite and oxidized with potassium ferricyanide were recorded at room temperature. The alpha  absorption bands corresponding to cytochromes a and a3 have maxima at 603 nm. The corresponding maximum for cytochrome b is 560 nm and for cytochrome c, 550 nm.
[View Larger Version of this Image (20K GIF file)]

Localization of Cox15p

The subcellular location of Cox15p was studied by expressing it as a biotinylated protein from constructs in which the coding sequence of COX15 was fused at the 3' end to a bacterial biotinylation signal sequence (see "Materials and Methods"). Transformation of the cox15 mutant WLDelta COX15 with the fusion gene either on a multicopy plasmid (pG4/ST10) or in an integrative vector (pG4/ST11) yielded respiratory competent clones (WLDelta COX15/ST10 and WLDelta COX15/Int, respectively) that grew on non-fermentable carbon sources with rates comparable to wild-type yeast. The COX15-BIO gene construct on multicopy and in single copy also restored a full complement of cytochrome oxidase (Fig. 5).

Wild-type yeast as well as the transformants harboring the fusion gene were fractionated to obtain mitochondria and total post-mitochondrial supernatant proteins. The biotinylated proteins in the two fractions were visualized on Western blots with a peroxidase-conjugated avidin detection system. Mitochondria from both transformants, but not from the wild-type strain, exhibit a unique biotin-containing protein with an apparent size of 48 kDa (Fig. 6). This protein is present in higher abundance in the mutant transformed with the fusion gene on a multicopy plasmid. The biotinylated protein is not detected in the post-mitochondrial supernatant fraction of the wild-type strain or the two transformants. These results indicate that the hybrid protein expressed from the fusion gene is biotinylated in vivo and is localized in mitochondria. The ability of yeast to biotinylate both cytoplasmic and mitochondrial proteins has been demonstrated previously (44-46). Whether biotinylation of mitochondrial proteins occurs in the cytoplasm or subsequent to import has not been investigated and is not currently known. The presence in mitochondria of biotin-dependent enzymes, however, tends to argue that there may be a separate organellar biotinylation system.


Fig. 6. Expression and mitochondrial localization of biotinylated Cox15p. Panel A, mitochondria and post-mitochondrial supernatant fractions were prepared from the wild-type strain W303-1A (WT) as well as from the respiratory competent transformants WLDelta COX15/ST10 (ST10) and WLDelta COX15/Int (Int) harboring the COX15-BIO fusion gene on a multicopy plasmid and as a single copy gene integrated at the chromosomal ura3 locus, respectively. Mitochondria from WLDelta COX15/Int were also sonically irradiated and separated into soluble proteins (Matrix) and submitochondrial membrane vesicles (SMP). The different fractions (20 µg of protein) were separated on a 12% polyacrylamide gel (32). Following transfer to nitrocellulose, biotinylated Cox15p was detected as described previously (17). The arrow identifies biotinylated Cox15p which, in this gel system, migrates slightly slower than the 45-kDa marker. Panel B, submitochondrial particles from WLDelta COX15/Int were treated with 0.5 M NaCl in the presence of the indicated concentrations of deoxycholate (DOC). Solubilized proteins (extract) were separated from the insoluble proteins (pellet) by centrifugation at 105,000 × gav for 20 min. The insoluble fractions were resuspended in the starting volumes of buffer. Equivalent volumes of detergent-solubilized and insoluble proteins were separated on a 12% polyacrylamide gel and analyzed as in panel A.
[View Larger Version of this Image (33K GIF file)]

The hydrophobic nature of the Cox15p sequence suggested it was likely to be a membrane component. This was ascertained by subfractionation of mitochondria from WLDelta COX15/Int, expressing chromosomally integrated COX15-BIO. Mitochondria were disrupted by sonic irradiation and separated into submitochondrial vesicles and soluble intermembrane and matrix proteins. Western analysis of the two fractions disclosed an exclusive association of biotinylated Cox15p with the membranes. Cox15p behaves as an integral membrane protein based on its solubility properties. As demonstrated in Fig. 6, extraction of the protein requires at least 0.25% deoxycholate and 0.5 M NaCl, conditions indicative of an integral membrane component.

To ascertain whether Cox15p is located in the outer or inner membrane, mitochondria were lysed under hypotonic conditions and sonically irradiated to fragment the outer and inner membranes (47). The mixture was separated on an isopycnic sucrose gradient and the distribution of biotinylated Cox15p was compared with inner (subunit 5 of cytochrome oxidase) and outer membrane (porin) markers. Even though the method achieves a separation of outer membranes relatively free of inner membrane contamination in the upper part of the gradient, the denser inner membranes contain substantial amounts of outer membrane. This is seen in the bimodal distribution of porin (Fig. 7). Cox15p-Bio fractionates as a single peak with a distribution similar to cytochrome oxidase subunit 5, indicating that it is an inner membrane protein.


Fig. 7. Co-localization of Cox15p with the inner membrane. Mitochondria prepared from WLDelta COX15/Int were resolved into outer and inner membranes by the method of Pon et al. (47). Mitochondria, suspended in 0.5 M sorbitol, 20 mM Tris-HCl, pH 7.5, 0.5 mM EDTA at a protein concentration of 22 mg/ml, were diluted with 10 volumes of 20 mM Tris-HCl, pH 7.5, 0.5 µM EDTA, 2 mM MgCl2, 5 mM phenylmethylsulfonyl fluoride. After incubation on ice for 30 min, the mixture was adjusted to 0.45 M sucrose and incubated for an additional 10 min on ice. The suspension was then sonically irradiated for a total of 55 s and centrifuged at 10,000 × gav for 15 min. The supernatant was collected and centrifuged at 267,000 × gav for 45 min. The pelleted membranes from the second centrifugation were resuspended in 5 mM Tris-HCl, pH 7.5, 10 mM KCl, 2 mM MgCl2, applied to a 5-ml column of a 28-55% linear sucrose gradient, and centrifuged for 15 h at 40,000 rpm. The gradient was fractionated into 18 equal fractions and analyzed for the distributions of cytochrome oxidase subunit 5, porin, and biotinylated Cox15p (Cox15p-Bio) as described in the legends to Figs. 2 and 6. The first lane (S) contains the extract applied to the gradient.
[View Larger Version of this Image (37K GIF file)]

Treatment of mitochondria and mitoplasts with proteinase K did not elicit loss of the protein. Under the same conditions an inner membrane protein containing an extramembrane domain facing the intermembrane space underwent proteolysis in mitoplasts but not in mitochondria (data not shown). This suggests that most of the Cox15p sequence must be embedded in the phospholipid bilayer and whatever domains exist outside the membrane are protected from attack by proteinase K. A largely membrane localization of Cox15p is supported by the 7-8 putative transmembrane domains making up most of the protein. Assuming Cox15p has 8 transmembrane segments and insertion into the membrane is coupled to transport, the biotinylated carboxyl terminus would be expected to face the matrix side of the membrane. By the same token, 7 transmembrane segments would dictate a localization of the carboxyl terminus in the intermembrane space. As indicated above, biotinylated Cox15p is resistant to proteinase K in mitoplasts. The protein was also resistant to proteolysis when mitochondria were sonically disrupted in the presence of proteinase K. These observations indicate that, independent of its location with respect to the two faces of the inner membrane, the carboxyl-terminal bacterial sequence with covalently bound biotin is not available to the protease, perhaps because it is masked by some other protein. Treatment of mitochondria with 0.1% deoxycholate prior to addition of proteinase K, however, led to a quantitative loss of Cox15p, thereby excluding the possibility that the protein is inherently resistant to proteolysis.

An estimate of the molecular weight of biotinylatd Cox15p was obtained by sedimentation analysis of a detergent-solubilized extract of WLDelta COX15/Int mitochondria in a sucrose gradient. The sedimentation of Cox15p-Bio relative to known size markers suggests a native molecular weight approximately two times that of the monomer (Fig. 8). Several other proteins involved in cytochrome oxidase assembly (17)3 have also been found to be larger than indicated by their monomer molecular weight, possibly because they may be part of larger heteroligomeric complexes. Attempts to obtain evidence for the existence of a complex of Cox15p with other proteins by biochemical or genetic means, however, were unsuccessful.


Fig. 8. Sedimentation properties of biotinylated Cox15p. Biotinylated Cox15p was extracted from WLDelta COX15/Int mitochondria (7 mg of protein) in the presence of 0.5 M NaCl and 0.5% potassium deoxycholate. After addition of 5 mg of hemoglobin and 0.25 mg of lactate dehydrogenase to the extract, the mixture was applied to a linear 7-20% sucrose gradient, and centrifuged for 6.5 h at 65,000 rpm in a Beckman SW65 rotor. The gradient was collected in 14 fractions and the distribution of biotinylated Cox15p (Cox15p-Bio in the upper part of the figure) was determined as described in the legend to Fig. 6. The migration of hemoglobin (×) and lactate dehydrogenase (open circle ) is shown in the lower part of the figure. Cox15p-Bio sediments slightly slower than lactate dehydrogenase.
[View Larger Version of this Image (24K GIF file)]

The production of respiratory chain enzymes, particularly cytochrome oxidase, has been shown to depend on numerous nuclear genes, not all of whose functions are understood at present (4-6). The products of these genes act at all stages of assembly, including maturation of mitochondrial cytochrome oxidase transcripts and steps unique to cytochrome oxidase, such as addition of the heme a and copper prosthetic groups. COX15 is representative of a large group of genes that catalyze events occurring late in the assembly pathway. Even though the precise function of this protein is not clear at present, our results show that it acts subsequent to the synthesis of the endogenous and imported subunits. Of the more extensively studied pet genes known to affect cytochrome oxidase, all except COX17 (13) appear to code for mitochondrial membrane proteins (15-17); and as demonstrated in this study for Cox15p, are components of the inner membrane. The co-localization of Cox15p and cytochrome oxidase in the same compartment of mitochondria is consistent with the notion that it may be directly involved in the assembly process.


FOOTNOTES

*   This work was supported in part by Research Grant GM50187 from the National Institutes of Health, United States Public Health Service.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Dagger    Recipient of a Medical Research Council of Canada Post-doctoral Fellowship. Current address: Dept. of Medical Genetics, University of Alberta.
§   To whom correspondence should be addressed. Tel.: 212-854-2920; Fax: 212-865-8246.
1   The abbreviations used are: pet mutant, nuclear respiratory deficient mutant of yeast; rho o mutant, cytoplasmic petite mutant lacking mitochondrial DNA; PAGE, polyacrylamide gel electrophoresis; kb, kilobase pair.
2   D. M. Glerum, I. Muroff, C. Jin, and A. Tzagoloff, unpublished data.
3   D. M. Glerum, I. Muroff, C. Jin, and A. Tzagoloff, unpublished observations.

REFERENCES

  1. Goffeau, A., Barrell, B. G., Bussey, H., Davis, R. W., Dujon, B., Feldmann, H., Galibert, F., Hoheisel, J. D., Jacq, C., Johnson, M., Louis, E. J., Mewes, H. W., Murakami, Y., Philippsen, P., Tettelin, H., and Oliver, S. G. (1996) Science 274, 546-567 [Abstract/Free Full Text]
  2. Sherman, F. (1963) Genetics 48, 375-385 [Free Full Text]
  3. Sherman, F., and Slonimski, P. P. (1964) Biochim. Biophys. Acta 90, 1-15
  4. Tzagoloff, A., and Dieckmann, C. L. (1990) Microbiol. Rev. 54, 211-225
  5. Burkl, G., Demmer, W., Holzner, H., and Schweizer, E. (1976) in Genetics Biogenesis and Bioenergetics of Mitochondria (Bandlow, W., Schweyen, R. J., Thomas, D. Y., Wolf, K., and Kaudewitz, E. F., eds), pp. 39-48, Walter de Gruyter, Berlin
  6. McEwen, J. E., Cameron, V. L., and Poyton, R. O. (1985) J. Bacteriol. 161, 831-835 [Medline] [Order article via Infotrieve]
  7. McEwen, J. E., Ko, C., Kloeckener-Gruissem, B., and Poyton, R. O. (1986) J. Biol. Chem. 261, 11872-11879 [Abstract/Free Full Text]
  8. Pel, H. J., Tzagoloff, A., and Grivell, L. A. (1992) Curr. Genet. 21, 139-146 [Medline] [Order article via Infotrieve]
  9. Seraphin, B., Simon, M., and Faye, G. (1988) EMBO J. 7, 1455-1464 [Abstract]
  10. Costanzo, M. C., Seaver, E. C., and Fox, T. D. (1986) EMBO J. 5, 3637-3641 [Abstract]
  11. Poutre, C. G., and Fox, T. D. (1987) Genetics 115, 637-647 [Abstract/Free Full Text]
  12. Tzagoloff, A., Nobrega, M., Gorman, N., and Sinclair, P. (1993) Biochem. Mol. Biol. Int. 31, 593-598 [Medline] [Order article via Infotrieve]
  13. Glerum, D. M., Shtanko, A., and Tzagoloff, A. (1996) J. Biol. Chem. 271, 14504-14509 [Abstract/Free Full Text]
  14. Glerum, D. M., Shtanko, A., and Tzagoloff, A. (1996) J. Biol. Chem. 271, 20531-20535 [Abstract/Free Full Text]
  15. McEwen, J. E., Hong, K. H., Park, S., and Preciado, G. T. (1993) Curr. Genet. 23, 9-14 [CrossRef][Medline] [Order article via Infotrieve]
  16. Bonnefoy, N., Chalvet, F., Hamel, P., Slonimski, P. P., and Dujardin, G. (1994) J. Mol. Biol. 239, 201-212 [CrossRef][Medline] [Order article via Infotrieve]
  17. Glerum, D. M., Koerner, T. J., and Tzagoloff, A. (1995) J. Biol. Chem. 270, 15585-15590 [Abstract/Free Full Text]
  18. ten Berge, A. M. A., Zoutewelle, G., and Needleman, R. B. (1974) Mol. Gen. Genet. 131, 113-121 [Medline] [Order article via Infotrieve]
  19. Orr-Weaver, T. L., Szostak, J. W., and Rothstein, R. J. (1981) Proc. Natl. Acad. Sci. U. S. A. 78, 6354-6358 [Abstract]
  20. Tzagoloff, A., Akai, A., and Needleman, R. B. (1975) J. Biol. Chem. 250, 8228-8235 [Abstract]
  21. Tzagoloff, A., Akai, A., and Foury, F. (1976) FEBS Lett. 65, 391-396 [CrossRef][Medline] [Order article via Infotrieve]
  22. Faye, G., Kujawa, C., and Fukuhara, H. (1974) J. Mol. Biol. 88, 185-203 [Medline] [Order article via Infotrieve]
  23. Sone, N., and Fujiwara, Y. (1991) J. Biochem. (Tokyo) 110, 1016-1021 [Abstract]
  24. Beggs, J. D. (1978) Nature 275, 104-109 [Medline] [Order article via Infotrieve]
  25. Broach, J. D., Strathern, J. N., and Hicks, J. B. (1979) Gene (Amst.) 8, 121-133 [CrossRef][Medline] [Order article via Infotrieve]
  26. Murtif, V. L., Bahler, C. R., and Samols, D. (1985) Proc. Natl. Acad. Sci. U. S. A. 82, 5617-5621 [Abstract]
  27. Hill, J. E., Myers, A. M., Koerner, T. J., and Tzagoloff, A. (1986) Yeast 2, 163-167 [Medline] [Order article via Infotrieve]
  28. Rothstein, R. J. (1983) Methods Enzymol. 101, 201-211
  29. Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982) Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY
  30. Myers, A. M., Pape, K. L., and Tzagoloff, A. (1985) EMBO J. 4, 2087-2092 [Abstract]
  31. Maxam, A. M., and Gilbert, W. (1977) Proc. Natl. Acad. Sci. U. S. A. 74, 560-564 [Abstract]
  32. Laemmli, U. K. (1970) Nature 227, 680-685 [Medline] [Order article via Infotrieve]
  33. Schagger, H., and von Jagow, G. (1987) Anal. Biochem. 166, 368-379 [Medline] [Order article via Infotrieve]
  34. Schmidt, R. J., Myers, A. M., Gillham, N. W., and Boynton, J. E. (1984) Mol. Biol. Evol. 1, 317-334 [Abstract]
  35. Lowry, O. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951) J. Biol. Chem. 193, 265-275 [Free Full Text]
  36. Dujon, B. (1981) in The Molecular Biology of the Yeast Saccharomyces cerevisiae (Strathern, J. N., Jones, E., and Broach, eds), pp. 506-635, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY
  37. Chua, N. H., and Bennoun, P. (1975) Proc. Natl. Acad. Sci. U. S. A. 72, 2173-2179
  38. Tomitake, T., Aoyama, H., Yamashita, E., Tomizaki, T., Yamaguchi, H., Shinzawa-Itoh, K, Nakashima, R., Yaono, R., and Yoshikawa, S. (1996) Science 272, 1136-1144 [Abstract]
  39. Saraste, M. (1990) Q. Rev. Biophys. 23, 331-360 [Medline] [Order article via Infotrieve]
  40. Nobrega, M. P., Nobrega, F. G., and Tzagoloff, A. (1990) J. Biol. Chem. 265, 14220-14226 [Abstract/Free Full Text]
  41. Chen, J., Derfler, B., and Samson, L. (1990) EMBO J. 9, 4569-4575 [Abstract]
  42. Vingron, M., and Argos, P. (1991) J. Mol. Biol. 218, 33-43 [CrossRef][Medline] [Order article via Infotrieve]
  43. Kyte, J., and Doolittle, R. F. (1982) J. Mol. Evol. 157, 105-132
  44. Cronan, J. E., Jr. (1990) J. Biol. Chem. 265, 10327-10333 [Abstract/Free Full Text]
  45. Bowman, S., Ackerman, S. H., Griffiths, D. E., and Tzagoloff, A. (1991) J. Biol. Chem. 266, 7517-7523 [Abstract/Free Full Text]
  46. Nobrega, F. G., Nobrega, M. P., and Tzagoloff, A. (1992) EMBO J. 11, 3821-3829 [Abstract]
  47. Pon, L., Moll, T., Vestweber, K., Marshallsay, B., and Schatz, G. (1989) J. Cell Biol. 109, 2603-2616 [Abstract]

©1997 by The American Society for Biochemistry and Molecular Biology, Inc.