The COQ5 gene encodes a yeast mitochondrial protein necessary for ubiquinone biosynthesis and the assembly of the respiratory chain.

Saccharomyces cerevisiae is a facultative anaerobe capable of meeting its energy requirements by fermentation and is thus an ideal system for studying the biogenesis of respiring mitochondria. We have isolated a respiration-deficient mutant exhibiting a pleiotropic loss of the mitochondrial electron transport chain. The corresponding wild-type gene, COQ5, was cloned, sequenced, and able to restore respiratory growth. Deletion of the chromosomal COQ5 gene results in a respiration deficiency and reduced levels of respiratory protein components. Exogenously added decylubiquinone can partially restore electron transport chain function to mitochondrial membranes from the deletion mutant. The COQ5 nucleotide sequence predicts a polypeptide of 307 amino acids containing a mitochondrial targeting signal. COQ5p is 43% identical to the polypeptide predicted by the Escherichia coli open reading frame, o251 (1). The COQ5 gene, when introduced into E. coli, complements the respiratory deficiency of an ubiE mutant that maps near o251, suggesting that it is the yeast homolog of the ubiE gene product. We conclude that the COQ5 gene encodes the mitochondria-localized 2-hexaprenyl-6-methoxy-1,4-benzoquinone methyltransferase of the yeast ubiquinone biosynthetic pathway.

The yeast, Saccharomyces cerevisiae, is the ideal organism for investigating the processes necessary for the production and maintenance of a functional electron transport chain in the mitochondria because it is able to make sufficient ATP for growth by fermentation. Respiration-deficient mutants form smaller colonies than wild-type cells when grown on glucosecontaining media, and are therefore called petite. Such mutants can arise in two ways, by mutation of the nuclear genome giving rise to nuclear petites or by mutation of the mitochondrial genome giving rise to cytoplasmic petites. Thus, the biogenesis of the mitochondrial electron transport chain is a complex process involving the coordinate expression of both nuclear and mitochondrial genes.
Ubiquinone is a lipid component of the mitochondrial electron transport chain that serves to transport electrons from the NADH or succinate dehydrogenase complexes to the cytochrome bc 1 complex, where it functions in a Q 1 cycle to generate a proton motive force across the inner membrane (2). In addition to its mitochondrial location, ubiquinone is found in a variety of intracellular membranes and in lipoproteins, where it may serve as an antioxidant (3). Ubiquinone is also an electron carrier in the plasma membrane respiratory chains of prokaryotes.
We have been interested in the biogenesis of the succinate dehydrogenase complex of the electron transport chain and the tricarboxylic acid. This enzyme is composed of four nuclear encoded protein products, a covalent FAD cofactor, iron-sulfur clusters, and heme and donates the electrons derived from succinate to ubiquinone (12). To better understand the biogenesis of this enzyme, we have isolated nuclear petite mutants that are defective in succinate dehydrogenase. Mutations affecting the biogenesis of more than one enzyme will have pleiotropic effects on mitochondrial function. A large number of pleiotropic mutants have been isolated, but few have been studied (4). We have studied one such mutant, TCM7, that has lost several respiratory chain activities including succinate oxidase activity.
The TCM7 mutant was complemented to respiration proficiency by transformation with a genomic yeast DNA library. The complementing gene, identified as COQ5, has been cloned, and its nucleotide sequence has been determined. It corresponds to a hypothetical protein encoded on chromosome XIII, YM8339.09c, whose sequence is retrievable under the Gen-Bank TM accession number P49017. COQ5 encodes a protein of 34.7 kDa (307 amino acids), which has an amino terminus that resembles a mitochondrial targeting sequence. The COQ5 gene probably encodes a methyltransferase in the ubiquinone biosynthetic pathway, since its expression in E. coli can complement a ubiE mutant, which lacks the 2-octaprenyl-6-methoxy-1,4-benzoquinone methyltransferase. Furthermore, the succinate oxidase activities of cytoplasmic membranes from an E. coli ubiE mutant or of yeast mitochondrial membranes from a coq5 mutant can be partially restored in vitro by the addition of quinone analogs.

EXPERIMENTAL PROCEDURES
Strains and Media-The yeast strains used in this study are listed in Table I. Yeast media have been described (13,14). Yeast transformation was with lithium acetate (15). The E. coli strain, AN70 (Hfr, metB, Str R ,ubiE-401), was kindly provided by Dr. Ian G. Young (16).
Mutant Isolation-MH125 was grown to stationary phase in YPD (1% yeast extract, 2% peptone, 2% dextrose), diluted 10-fold in water, and irradiated with ultraviolet light at 254 nm (1.4 J/m 2 s) for 9 min to achieve about 95% killing. petite strains that stained red with 2,3,5triphenyltetrazolium chloride and secreted acid were identified as described and were candidate succinate dehydrogenase mutants (14). Mitochondrial membranes were isolated and assayed for the NADHand the succinate-dependent reduction of cytochrome c as described (13), and strains, such as TCM7, displaying an apparently selective loss of the succinate-dependent activity were further analyzed.
Cloning and Identification of the COQ5 Gene-The mutant, TCM7, was transformed with a yeast genomic DNA library (17) and transformants were selected for uracil prototrophy on SD (0.67% yeast nitrogen base without amino acids, 2% dextrose plates. Approximately 6,000 transformants were replica-plated to SG (0.67% yeast nitrogen base without amino acids, 3% glycerol) plates to test for complementation to respiration proficiency. The plasmids were recovered and analyzed by restriction mapping; this resulted in the isolation of two complementing plasmids with overlapping inserts. A 3.0-kb restriction fragment from a partial Sau3AI digest was cloned into the vector, YCplac33 (18), to give the complementing plasmid, p7S6. Exonuclease digestions and the production of nested deletions were performed as described by the supplier (Stratagene, La Jolla, CA). All deletions affecting the COQ5 open reading frame resulted in the loss of complementation. The COQ5 gene sequence was obtained for both strands using an Applied Biosystems 373A DNA sequencer.
Construction of a Chromosomal Disruption Mutant-The plasmid, pSK7-TRP (described in the legend to Fig. 1), was linearized and used to transform MH125 and MH123 to tryptophan prototrophy by selection on SD medium supplemented with histidine, uracil, and leucine. At least 30 stably transformed, MH125-derived tryptophan prototrophs had also lost the ability to grow on YPG (1% yeast extract, 2% peptone, 3% glycerol), and 4 were further examined. Polymerase chain reactions (19) with the oligonucleotides 5Ј-TGGAAACTAGCTTCCGCATT-3Ј and 5Ј-CTAACAGTAATCTCGCAGTT-3Ј were used to amplify the COQ5 region and verify the presence of the disruption allele, coq5⌬-1::TRP1, in the transformants.
Construction and Expression of a Tagged COQ5 Protein-A polymerase chain reaction with the following oligonucleotides 5Ј-CATAGAT-TGTGGAAGGACCA-3Ј and 5Ј-CTTCGTCGACTTACAAGTCTTCTT-CAGAAATAAGCTTTTGTTCACCTCCAACTTTAATGCCCCAATGG-3Ј was performed to add the 10-residue c-myc protooncogene sequence recognized by the monoclonal antibody, 9E10, to the carboxyl terminus of COQ5p (20). The latter oligonucleotide also incorporates two glycine residues between the COQ5 and the c-myc sequences, a stop codon, and a SalI restriction site for cloning. The carboxyl-terminal 280 base pairs of COQ5, now tagged with the myc sequence were amplified, digested with XbaI and SalI, and cloned into likewise digested p7S6 to generate p7S6myc encoding the tagged protein. The amplified region was sequenced to confirm the construction. For overexpression of the tagged COQ5p, the 1.0-kb MunI to SalI fragment encoding the entire gene was cloned into the vector, pUCu, under control of the copper-inducible CUP1 promoter (20) to generate pUCu-7myc. This plasmid was introduced into MH125, the transformants were grown on lactate medium (21) and inoculated into lactate medium supplemented with 20 g/ml histidine, leucine, and tryptophan containing 0.5 mM CuSO 4 . Samples for Western blot analysis were prepared and transferred to nitrocellu-lose as described (13) and detected with the monoclonal antibody, 9E10 (Chemicon International, Inc., Temecula, CA), goat anti-mouse secondary antibody (Bio-Rad Laboratories Ltd., Hercules, CA), and the ECL detection system (Amersham Canada Ltd., Oakville, Ontario, Canada).
Mitochondrial Protein Import-The 1.0-kb MunI to SalI fragment from p7S6myc was cloned into pBluescript II KSϪ under the control of the T7 promoter. Template mRNA was produced by in vitro transcription using T7 RNA polymerase (Life Technologies Inc., Burlington, Ontario, Canada) and translated in rabbit reticulocyte lysate as described by the supplier (Promega Corp., Madison, WI) using Tran 35 S-Label (ICN Biomedicals, St. Laurent, Quebec). Import reactions were performed as described (22).

RESULTS
TCM7 Is a Nuclear petite Mutant-To better understand the mechanisms by which complex membrane proteins are assembled into the mitochondrial inner membrane, we have screened for mutants that are defective in the activity of our model protein, succinate dehydrogenase. Succinate dehydrogenase mutants are unable to grow on glycerol as a sole carbon source but still possess a functional electron transport chain linking NADH or glycerol-1-phosphate oxidation to the reduction of oxygen (13,27). In our screen, we took advantage of the observation that strains unable to grow on glycerol form small or petite colonies on YPDG (1% yeast extract, 2% peptone, 0.1% dextrose, 3% glycerol) plates (28). 0 or Ϫ strains that are completely lacking an electron transport chain due to defects in their mitochondrial genomes were identified by their inability to reduce tetrazolium dye and eliminated. The secretion of acid, as monitored by the indicator dye, bromcresol purple, is taken as an indication of tricarboxylic acid cycle mutants (14). Tetrazolium staining, acid-secreting petites were grown on liquid YPD-0.6% (1% yeast extract, 1% peptone, 0.6% glucose, 0.1% KH 2 PO 4 , 0.12% (NH 4 ) 2 SO 4 , pH 6.2), harvested, and lysed in a French pressure cell to isolate submitochondrial membranes (13). Succinate-and NADH-cytochrome c reductase activities were measured; loss of the former and retention of the latter activity was taken as a preliminary indication of a succinate dehydrogenase mutant. In MH125, a substantial amount of antimycin A-insensitive NADH-cytochrome c reductase activity was induced, a phenomenon that has been reported previously (29,30). This resulted in the isolation of putative succinate dehydrogenase mutants, which upon closer examination were revealed to have pleiotropic losses of mitochondrial respiratory functions. The mutant, TCM7, is thus a nuclear petite with a pleiotropic loss of most respiratory chain proteins, including the cytochrome bc 1 complex and succinate dehydrogenase. It defines one of four complementation groups isolated. When the mutant is mated and the resulting diploid is sporulated, the respiration deficiency in TCM7 segregates 2:2 in over 20 tetrads as expected of a single-gene, nuclear mutation.
Construction of a TCM7 Null Allele-To more fully characterize the function of the TCM7 gene in a nonmutagenized background, we decided to clone the wild-type gene (see "Experimental Procedures") and to create a null allele by a onestep gene disruption (Fig. 1). The plasmid, pSK7-TRP, was transformed into the haploid, MH125, and the diploid, MH123, and tryptophan prototrophs were selected. Verification of the presence of the tcm7⌬-1::TRP1 allele in both haploid and diploid transformants was shown by polymerase chain reactions using genomic DNA as template (Fig. 2, lanes 2 and 4, respectively). The successful isolation of disruption mutants in the haploid, MH125, indicates that TCM7 is not essential for viability (lane 4). The disruption mutant, ED7, like TCM7, is respiration-deficient; it does not grow on YPG and has a low growth yield on a medium containing low levels of glucose (not shown). When mated to the tcm7-1 mutant, TCM7-2A5A, ED7 was not complemented, indicating the tcm7⌬-1::TRP1 allele belongs to the same complementation group as the tcm7-1 mutation in TCM7-2A5A and TCM7. ED7 is also rescued to respiration proficiency by the plasmid, p7S6. When a diploid carrying a single copy of the tcm7⌬-1::TRP1 disruption is sporulated, respiration deficiency and tryptophan prototrophy cosegregated in six dissected tetrads.
TCM7 Is a coq5 Mutant-As discussed below, we discovered that the cloned TCM7 gene could complement an E. coli ubiquinone biosynthetic mutant. This raised the possibility that the TCM7 gene encodes an enzyme of the yeast ubiquinone biosynthetic pathway. When TCM7 or ED7 are mated to the known coq5 mutant, CH83-B3, the diploids remain respiration-deficient, demonstrating that the mutations are allelic. As a control, we ensured that all haploids were ϩ by mating them to the 0 tester strains and verifying the respiration competency of these diploids. Furthermore, we could complement the coq5 mutants, CH83-B1 and CH83-B3, with the plasmid p7S6. From these data, we conclude that TCM7 and ED7 are coq5 mutants, and we refer to the TCM7 gene as COQ5. We could not restore respiratory growth to TCM7 or ED7 by supplementing the medium with coenzyme Q 2 , but this property is not universal to ubiquinone biosynthetic mutants (31).
COQ5 Encodes a Mitochondrial Protein-The nucleotide sequence of the COQ5 gene was determined and is identical to part of the sequence found in GenBank TM (accession number Z49210). The predicted protein sequence suggests that COQ5p is synthesized as a precursor protein that is imported into and cleaved by mitochondria, since the amino-terminal 31 amino acids are rich in basic and hydroxylated residues and devoid of acidic ones. To investigate the localization of COQ5p, we constructed a fusion protein containing the entire COQ5p, two glycine residues as a spacer domain, and the 10-residue c-myc sequence that serves as an antibody recognition site. The myctagged COQ5 produced a functional protein, since it could complement the mutants, TCM7 and ED7. When the fusion protein was expressed under the control of the inducible CUP1 promoter and the cells were fractionated, it was primarily detected in the mitochondrial fraction by Western blot analysis (Fig. 3A, lane 3). Ready detection of the tagged COQ5p could be seen when expression was under the control of the uninduced CUP1 promoter but not when expression was mediated by the COQ5 promoter (not shown). This suggests that COQ5p is not normally a highly expressed protein. Immunofluorescence analysis of cells overexpressing the myc-tagged COQ5p produced a punctate staining pattern characteristic of mitochondrial proteins (not shown). When mitochondria were further fractionated, COQ5p was predominantly found in the matrix fraction (Fig. 3B, lane 2 of c-myc panel). Small amounts were also found associated with the membrane fractions (lane 1) or the intermembrane space (lane 3), but these levels reflect crosscontamination of fractions as judged by the membrane (porin), matrix (hsp60), and intermembrane space (cytochrome b 2 ) marker proteins. Our mitochondrial fractions contain contaminating organelles that co-purify, and hence we cannot exclude the possibility that COQ5p is located in other organelles beside mitochondria.
To independently confirm that COQ5p is a mitochondrial protein, in vitro synthesized precursor protein was imported into isolated yeast mitochondria (Fig. 4). COQ5p bound to mitochondria (lanes 2 and 4) and could be imported in an energy-dependent manner (lanes 4 and 5) to a protease-protected location (lane 3) that is made accessible by the addition of detergent (lane 6). Import is accompanied by a proteolytic cleavage event to a faster migrating species (lanes 2 and 3). The COQ5p presequence contains the sequence RC2FTQA-HRAC2 which resembles the consensus sequence for twice cleaved precursor proteins, RX2(F/L/I)XX(T/S/G)XXXX2 (32). This motif is found in many proteins required for respiratory function (32) 1. Construction of a chromosomal deletion mutant. The 3-kb Sau3AI fragment from the plasmid p7S6 that encodes the COQ5 gene was transferred to the vector pBluescript II SKϩ as a HindIII to SstI fragment. The 1.11-kb MunI to EcoRI fragment carrying the entire COQ5 gene was replaced with a 0.85-kb EcoRI fragment encoding the selectable marker, TRP1, creating the plasmid pSK7-TRP. The TRP1 and the COQ5 open reading frames are represented as boxes, while flanking DNA is represented as lines. E. coli chromosome (1). Interestingly, three mutations that lead to respiration deficiency in E. coli had also been mapped to this region but not assigned to any open reading frame; these are ubiB, ubiD, and ubiE (33). These mutations lead to defects in ubiquinone biosynthesis that are manifested as an inability to grow on the nonfermentable carbon source, succinate. The ubiB and ubiD gene products are part of a prokaryotic specific ubiquinone biosynthetic pathway and are believed to catalyze the conversion of intermediates II to III and I to II, respectively (Fig. 5). The ubiE gene product was proposed to catalyze a reaction common to the eukaryotic and the prokaryotic pathways, the methylation of intermediate VII to intermediate VIII, because an ubiE mutant accumulates 2-octaprenyl-6-methoxy-1,4-benzoquinone (intermediate VII; Ref. 5). To determine whether the COQ5p could be the yeast homolog of the ubiE gene product, we transformed a ubiE mutant, AN70, with the plasmid p7S6. COQ5 is able to complement AN70 and restore growth on succinate. Furthermore, when the quinone pools are extracted from AN70, quinones are not detectable by thin layer chromatography in either chloroform/benzene (1:1) or in chloroform/methanol (95:5) solvent systems (not shown). In contrast, the quinone pool from AN70 expressing COQ5 contained compounds that comigrate with a Q 0 standard in both solvent systems used (R F (relative mobilities) ϭ 0.2 and 0.85, respectively). This result suggests strongly that COQ5 complements AN70 by restoring ubiquinone biosynthesis.
Membrane vesicles prepared from AN70 have no detectable succinate oxidase activity (Fig. 6). The addition of increasing concentrations of Q 0 restored succinate oxidase activity. Membrane vesicles from AN70 transformed with p7S6 displayed a high level of succinate oxidase activity in the absence of added quinone; this level of oxidase activity is similar to that found in a related wild-type strain of E. coli (34). The addition of 200 M coenzyme Q 0 to COQ5-complemented bacterial membranes only slightly increased respiratory activity, indicating that the yeast COQ5p had restored adequate quinone levels.
Exogenous Quinone Stimulates Respiration in ED7 Mitochondrial Membranes-Submitochondrial membranes were prepared from the yeast disruption mutant, ED7, and its wildtype parent, MH125, and analyzed for respiratory chain activities (Table II). ED7 membranes showed a dramatic loss of NADH oxidase and NADH-cytochrome c reductase activities. Even more dramatically, succinate oxidase and succinate-cytochrome c reductase activities were undetectable, although a small amount of succinate-dependent phenazine methosulfatemediated reduction of the artificial electron acceptor, dichlorophenol indophenol, a ubiquinone-independent activity, remained. The addition of 200 M decylubiquinone, Q 2 , led to the severalfold stimulation of some of the respiratory activities, with NADH oxidase activity being most affected. Higher concentrations of quinone could not further increase activity levels. The ED7 culture used in this experiment had remained at least 75% ϩ ; the less than full recovery of respiratory chain activities with added quinone cannot be accounted for by mitochondrial DNA mutations.
The COQ5 Mutant Has Lowered Levels of Electron Transport Chain Components-The failure of added quinone to fully restore respiratory function to ED7 mitochondrial membranes might be due to lower levels of assembled respiratory chain components. We examined mitochondrial membranes for the presence of respiratory chain components by Western blot analysis and normalized mutant and wild-type samples by loading equivalent amounts of the outer membrane protein, porin (Fig.  7A). We reasoned that respiratory deficiency should not affect the import of mitochondrial outer membrane proteins, since their insertion is energy-independent (35). We detected low levels (less than 10%) of succinate dehydrogenase with an antibody against the iron-sulfur subunit (panel B), low levels of ATP synthase with antibodies against the ␤-subunit (panel C), and low levels of cytochrome c oxidase with antibodies against subunit IV (panel D), consistent with the reduced enzymatic activities reported in Table II. Therefore, a ubiquinone deficiency as present in the COQ5 null mutant leads to lowered steady state levels of electron transport chain components. DISCUSSION In this study, we characterize a coq5 mutant, isolate the corresponding wild-type COQ5 gene, and provide evidence that COQ5p is involved in ubiquinone biosynthesis in S. cerevisiae. Loss of COQ5 function and the resultant loss of ubiquinone synthesis should result in a respiration-deficient phenotype that is manifested as an inability to grow on nonfermentable carbon sources. The original mutant, TCM7, however, is at FIG. 3. COQ5p is a mitochondrial protein. Samples of mitochondrial proteins from lactate-grown cells, MH125 (lanes 2, 4, and 6) and MH125, with the plasmid pUCu-7myc (lanes 1, 3, and 5) were run on a 10% acrylamide gel and transferred to a nitrocellulose membrane for Western blot analysis with anti-myc antibodies. A, lanes 1 and 2, 100 g of cytosolic fractions; lanes 3 and 4, 100 g of mitochondrial fractions; lanes 5 and 6, 100 g of cell homogenates. The position of the myctagged COQ5p protein is indicated by the arrow. The positions of size markers in kilodaltons are indicated on the left. B, 400 g of mitochondria from MH125 carrying the plasmid pUCu-7myc were fractionated into membrane (lane 1), matrix (lane 2), and intermembrane space (lane 3) fractions as described except that 2 M urea was included in the mitoplast resuspension buffer (40). Identical samples of 10% of each fraction were resolved on 12% acrylamide gels, transferred to nitrocellulose, and probed with anti-myc, anti-heat shock protein 60 (hsp60), anti-porin, and anti-cytochrome b 2 antibodies. All antibodies were diluted 10,000-fold. least partially respiration-proficient in that it is able to reduce the tetrazolium dye used in our mutant screen, suggesting that the coq5-1 allele is leaky. Mutations that completely block ubiquinone biosynthesis might not have been detected by our screening procedure. The secretion of acid by TCM7 and the low succinate dehydrogenase activity in the COQ5 disruption mutant (Table II) suggest that the tricarboxylic acid cycle is also impaired. We have not measured the levels of other tricarboxylic acid cycle enzymes.
The E. coli ubiE mutant, AN70, is characterized by an in-ability to grow on nonfermentable carbon sources such as succinate and by low oxidase activities (33). Expression of the yeast COQ5 gene restores the bacterium's ability to grow with succinate as the carbon source and the succinate oxidase activity of isolated bacterial membranes. The COQ5 gene also rescues the ubiquinone synthesis deficiency of the ubiE mutation; although we did not characterize the exact nature of the products, the COQ5 gene results in the production of quinone compounds with a mobility by thin layer chromatography similar to Q 0 in a chloroform/methanol solvent system. The ubiE mutant accumulates intermediate VII or 2-octaprenyl-6-methoxy-1,4-benzoquinone ( Fig. 5; Ref. 16). Restoration of ubiquinone biosynthesis in AN70 by the yeast COQ5 strongly suggests that COQ5p is the methyltransferase that modifies position 3 of the benzoquinone ring. We have not attempted to show methyltransferase activity in vitro nor have we attempted to isolate the intermediates that accumulate in ED7. There are two other methyltransferases, both O-methyltransferases, in the ubiquinone biosynthetic pathway. The yeast COQ3 gene encodes the 3,4-dihydroxy-5-hexaprenylbenzoate methyltransferase, which converts intermediate IV to V. It has been cloned, sequenced, and characterized and is distinct from the COQ5 gene (7). The second O-methyltransferase catalyzes the final step in the pathway, converting intermediate IX to ubiquinone (X) and has not been characterized in yeast (9). In E. coli, the second O-methyltransferase is encoded by the E. coli ubiG gene. The ubiG gene has recently been shown to complement a yeast coq3 mutant, strongly suggesting that the ubiG protein catalyzes both O-methyltransferase reactions in E. coli and leading to speculation that a bifunctional O-methyltransferase may also exist in yeast (9). The ubiG gene has   been sequenced and is not related in sequence to the yeast COQ5 gene except in the proposed S-adenosylmethionine binding domains ( Fig. 8; Ref. 6). The simplest explanation for our results is that COQ5 encodes the yeast 2-hexaprenyl-6-methoxy-1,4-benzoquinone methyltransferase, which can functionally complement a mutation in the E. coli homolog encoded by the ubiE gene.
The COQ5p sequence is weakly related to other methyltransferases; it does contain a closely related version of a consensus amino acid sequence common to methyltransferase enzymes that use S-adenosylmethionine as a donor ( Fig. 8; Refs. 6 and 36). As in the human glycine methyltransferase sequence, the first of the conserved glycines in the motif is replaced by an alanine in COQ5p, but the remainder of the adjacent COQ5p sequence conforms to the consensus motif (Fig. 8A). There is a second region of sequence similarity found in a smaller number of methyltransferases that is also found in COQ5p ( Fig. 8B; Ref. 36). These sequence similarities are consistent with the use of S-adenosylmethionine as the donor for all the methyl groups in the ubiquinone biosynthetic pathway (37).
We localized COQ5p to the mitochondrial matrix by Western blot analysis of a tagged, overexpressed protein (Fig. 3) and by in vitro import into isolated mitochondria (Fig. 4). This localization is consistent with the proposed role for Coq5p in ubiquinone biosynthesis, which is generally believed to be compartmentalized within mitochondria (7). Complementation of a yeast coq3 mutant by the E. coli ubiG gene required the addition of a mitochondrial targeting signal to the ubiG gene, indicating a need for organellar targeting of some proteins in the ubiquinone biosynthetic pathway (9). The predicted protein sequence does not contain any putative membrane-spanning domains, so COQ5p was expected to be soluble. In E. coli, the enzymes of the ubiquinol biosynthetic pathway that can convert 2-octaprenylphenol (intermediate II) to ubiquinone-8 can be released from membranes as a complex without detergent despite the hydrophobic nature of their substrates (37). Ubiquinone's primary role is in electron transfer to the cytochrome c reductase complex in the mitochondrial respiratory chain, consistent with a mitochondrial site of synthesis, but it is also found in other eukaryotic organelles where its function is less well defined.
The addition of an ubiquinone analog to ED7 ubiquinonedeficient membranes only partially restores respiratory activities, particularly succinate oxidase activity. In contrast, the ubiE mutant, AN70, which is similarly blocked in ubiquinone synthesis, shows a more marked recovery of succinate oxidase activity with added Q 0 (Fig. 6). We have shown that the partial recovery of respiratory activities is likely attributable to the substantially lower levels of respiratory chain components in ED7 membranes ( Fig. 7 and Table II). Why does the absence of COQ5p lead to low levels of respiratory chain components in yeast, whereas some other nuclear petite mutations do not? For example, sdh1 or sdh4 mutants have normal levels of NADH oxidase activities in their mitochondrial membranes, despite their respiration deficiencies due to losses of succinate dehydrogenase subunits (13,27). A yeast coq3 mutant also has normal levels of succinate or NADH oxidase activities if a quinone analog is added to the assays (30).
We do not believe ubiquinone deficiency is affecting the translocation of proteins into the organelle. Even cytoplasmic petite mutants that are devoid of respiratory activity are still able to import and assemble normal F 1 -ATP synthase (38). One possibility is that a lack of ubiquinone destabilizes the respiratory chain complexes and leads to their rapid degradation and low steady state levels. Destabilization could be manifested by a change in the structural properties of the bilayer or of the proteins themselves. Alternatively, the low levels of respiratory components may arise from the failure of the cells to adapt to respiratory conditions and activate the expression of mitochondrial proteins upon the depletion of available glucose. The production of respiratory chain components is normally accelerated in stationary phase cells that we used for our studies (39). 2 We have not been able to test whether the failure to recover from catabolite repression is responsible for the low levels of respiratory components because our coq5 mutants do not grow under nonrepressing conditions with galactose as a carbon source. An explanation for this phenomenon awaits further investigations.