The Saccharomyces cerevisiae COQ10 Gene Encodes a START Domain Protein Required for Function of Coenzyme Q in Respiration*

Deletion of the Saccharomyces cerevisiae gene YOL008W, here referred to as COQ10, elicits a respiratory defect as a result of the inability of the mutant to oxidize NADH and succinate. Both activities are restored by exogenous coenzyme Q2. Respiration is also partially rescued by COQ2, COQ7, or COQ8/ABC1, when these genes are present in high copy. Unlike other coq mutants, all of which lack Q6, the coq10 mutant has near normal amounts of Q6 in mitochondria. Coq10p is widely distributed in bacteria and eukaryotes and is homologous to proteins of the “aromatic-rich protein family” Pfam03654 and to members of the START domain superfamily that have a hydrophobic tunnel implicated in binding lipophilic molecules such as cholesterol and polyketides. Analysis of coenzyme Q in polyhistidine-tagged Coq10p purified from mitochondria indicates the presence 0.032–0.034 mol of Q6/mol of protein. We propose that Coq10p is a Q6-binding protein and that in the coq10 mutant Q6 it is not able to act as an electron carrier, possibly because of improper localization.

diates of the pathway are not detected in coq mutants (6 -12). These observations indicate that the pathway is stringently regulated and/or that most of the intermediates are degraded when biosynthesis of Q 6 is arrested. COQ gene products are located in, or are peripherally associated with the inner membrane of mitochondria (9 -15) where they constitute a pathway that is similar to but diverges from the one in bacteria in at least two steps (16,17).
In the present study we report a novel phenotype displayed by mutants with a deletion of reading frame YOL008W, which we have named COQ10. The coq10 mutant is similar to other coq mutants in which optimal oxidation of NADH and succinate by isolated mitochondria depends on addition of coenzyme Q 2 . Unlike the coq mutants, which lack Q 6 , the coq10 mutant reported here has nearly normal concentrations of Q 6 . The product of COQ10 is a hydrophobic protein located in the inner membrane of mitochondria. It is a member of the large Pfam03654 family of proteins (aromatic-rich protein family) (18). The recent NMR structure of a bacterial member of this family (19) has revealed a structural similarity to the START domain family (20), which includes the cholesterol-binding protein StarD4 (21). Based on the presence in Coq10p of this domain and the detection of Q 6 in the purified protein, we propose that Coq10p is a Q 6 -binding protein, which functions in the delivery of Q 6 to its proper location for electron transport.
Cloning of COQ2, COQ7, COQ8, and COQ10-Recombinant plasmids containing COQ2 (pG10/T1) and COQ7 (pG64/T2) and COQ8 (pG75/T2) were cloned by transformation of W303⌬COQ10 with a yeast genomic library constructed from nuclear DNA partially digested with a combination of BamHI and BglII and cloned in the yeast/Escherichia coli shuttle plasmid YEp352 (22). Approximately 5 ϫ 10 8 cells were transformed with 50 g of the plasmid library by the method of Schiestl and Gietz (23). COQ10 was PCR amplified from yeast nuclear DNA with primers: 5Ј-cacacagaaagcttagaaatag and 5Ј-ggcaagcttatcctggatggcatgatc. The 1.3-kb fragment containing the COQ10 reading frame plus 700 bp of 5Ј and 50 bp of 3Ј-untranslated sequence was digested with HindIII and cloned in YEp352.
Disruption of COQ10-The COQ10 gene and ϳ500 bp of upstream and downstream sequence was amplified with primers: 5Ј-ggcggatccttcatgattagctatagtacg and 5Ј-ggcggatccgtcctcccagtgttcatctgg. This fragment was digested with BamHI and cloned in pUC19 (24). This plasmid was used to delete the COQ10 coding sequence with the two bidirectional primers 5Ј-ggcagatctccacggtatctttagccg and 5Ј-ggcagatcttccgtataatttcagataacaaagatc. The PCR product consisting of linear pUC19 plus the COQ10 flanking sequences was digested with BglII and ligated to a 1-kb BamHI fragment containing the yeast HIS3 gene (this fragment lacked the downstream PET56 gene). The linear BamHI fragment with the coq10::HIS3 allele was isolated from this plasmid and substituted for the wild type gene in W303-1A and W303-1B by the one-step gene replacement method (25).

Construction of Hybrid COQ10 Genes Expressing Coq10p with a Carboxyl-terminal Polyhistidine Extension and of a trpE Fusion Protein for
Coq10p Antibody Production-The two PCR primers 5Ј-ggcagatctatataaaatggttttgataataaggccc and 5Ј-ggcaagctttcagtgatggtgatgatggtgcggagagccttctttagaag were used to amplify COQ10 with a 18-nucleotide insertion between the last codon and the termination codon. The PCR amplified product was digested with a combination of BglII and HindIII and was cloned into the BamHI and HindIII sites of YEp352/ GAL, a YEp352 (22) derivative with the URA3 selectable marker and the GAL10 inducible promoter immediately upstream of the multiple cloning sequence. This construct was used to transform the coq10 null mutant aW303⌬COQ10, to yield aW303⌬COQ10/ST13, which expressed the protein with six carboxyl-terminal histidines (Coq10p-his).
The trpE/COQ10 hybrid gene was obtained by cloning a 600-bp KpnI-HindIII fragment containing all but the first 35 codons of COQ10 in pATH20 (26) so as to effect an in-frame fusion to E. coli trpE coding for the amino-terminal half of anthranilate synthase component I. A 57-kDa hybrid protein was isolated from E. coli transformed with the plasmid. The insoluble protein was dissociated in SDS and was partially purified on a Bio-Gel A-0.5 column (26) and used to obtain a polyclonal antibody in rabbits.
Construction of a Plasmid with the Human Homologue of COQ10-The human homologue of COQ10 was PCR-amplified with primers 5Ј-ggcgagctcatgaggtttctgacctcctgc and 5-ggcggatcctcaagtctggtgcacctc with the Image clone 5299314 serving as the template. The 627-bp product was digested with a combination of SacI and BamHI and cloned into pMGL4, a multicopy copy yeast/E. coli shuttle vector with the ADH1 promoter and terminator (Dr. K. M. Glerum, University of Alberta, Edmonton, Canada).
Lipid Extraction and Quinone Identification by Reverse Phase HPLC with Electrochemical Quantification and Mass Spectrometry-Lipid extracts were prepared from 1 mg of yeast mitochondrial protein and quinones were detected and quantified by HPLC/ECD as described previously (4).
The mitochondrial lipid extracts were also examined by flow injection analysis with atmospheric pressure ionization and MS/MS product ion spectra. Aliquots of 20 -50 l of extracted calibration curve samples or standards were passed directly into an atmospheric pressure chemical ionization source (APCI, nebulizing gas, "zero"-grade air produced by a Zero Air Generator (Peak Scientific, Chicago, IL) at 100 -200 l/min, nebulizer 400 -450°C, orifice 50 volts) attached to a PerkinElmer Sciex (Thornhill, Canada) API III triple quadrupole mass spectrometer operated in the tandem mass spectrometric (MS/MS) mode (99.999% argon collision gas at an instrumental GCT setting of 200). Spectral data were collected by MS/MS methods (precursor or product scanning): Q 6 , [M ϩ H] ϩ ϭ 591.4. Authentic standards of Q 6 (Sigma) were fragmented and used to confirm the identity of the ions.
Miscellaneous Procedures-Standard methods were used for plasmid manipulations (27). Yeast mitochondria were prepared from cells grown in YPGal to early stationary phase by the method of Faye et al. (28) except that Zymolyase 20T (ICN Laboratories, Aurora, OH) instead of Glusulase was used to obtain spheroplasts. Spectral analyses of mitochondrial cytochromes and measurements of respiratory enzymes were performed as described previously (5). Protein concentrations were determined by the method of Lowry et al. (29).

RESULTS
BY⌬COQ10 and W303⌬COQ10 Phenotype-BY⌬COQ10 is a respiratory deficient mutant of the Genome Deletion Strain Collection. It was made by replacing reading frame YOL008W/COQ10 with the kanamycin-resistance cassette in strain BY4741. aW303⌬COQ10 and W303⌬COQ10 were obtained by replacing YOL008W with HIS3 in the wild type haploid W303-1A and W303-1B, respectively (see "Experimental Procedures"). The respiratory defect of both mutants is complemented by crosses to a o tester indicating that null mutations in this gene are recessive. In rich glucose medium, an early stationary culture of W303⌬COQ10 consists of ϳ50% o/Ϫ derivatives. On selective medium (YPEG) the mutant spontaneously reverts as a result of mutations in nuclear DNA.
Unlike most coq mutants, which show a complete growth defect on glycerol/ethanol (5)(6)(7)(8)(9)(10)(11)(12), aW303⌬COQ10 and the other coq10 mutants grow slowly on this medium with a measurable generation time of 11.8 h (TABLE TWO). Addition of Q 6 to growth medium has been shown to promote growth of coq mutants on non-fermentable carbon sources (12). This is confirmed by the partial rescue of the coq5 and coq9 mutants in media supplemented with Q 6 . The generation time of aW303⌬COQ10 is reduced to 8.5 h in media supplemented with Q 6 (TABLE TWO).
Spectra of W303⌬COQ10 mitochondria show decreased concentrations of "a" and "b" type cytochromes (Fig. 1A). Some of the decrease, however, is because of the large proportion (50%) of o/Ϫ mutants in the culture. The tendency of nuclear pet mutants to accumulate large deletions in mitochondrial DNA is common when the mutations are in genes coding for components of the mitochondrial translation machinery (32). In vivo assays of mitochondrial protein synthesis in the coq10 null mutant revealed a normal pattern of mitochondrial translation products indicating that the excessive loss of the wild type mitochondrial genome was not caused by a defect in this translation system (Fig.  1B). The somewhat lower incorporation of [ 35 S]methionine in the mutant is consistent with the percentage of o/Ϫ cells in the cultures used for the assays.
The reductions of cytochromes a, a 3 , and b in the coq10 null mutant is also evident in the lower NADH-cytochrome c reductase and cytochrome oxidase activities of mutant mitochondria (TABLE THREE). In both mutants examined, the NADH-cytochrome c reductase activity was ϳ10% of that measured in wild type even though cytochrome oxidase was reduced by only 40%. The lesion, therefore, was likely to be in a component of the respiratory chain preceding cytochrome oxidase. This activity was confirmed by the observation that the NADH-cytochrome c reductase activity of the coq10 mutants could be restored to ϳ50% of the wild type by addition of exogenous coenzyme Q 2 to the assay (TABLE THREE). The Q 2 dependent activity is commensurate with complete restoration of NADH-cytochrome c reductase because half of the culture used for the preparation of mitochondria consisted of o/Ϫ mutants and therefore lacked the bc 1 complex. Coenzyme Q 2 also restored NADH and succinate oxidation in mitochondria of the coq10 mutant (Fig. 2). Measurements of NADH oxidase activity as a function of Q 2 added to the assay indicated that the response of the coq10 mutant was similar to another mutant (coq9) that lacks Q 6 (Fig. 3).
The coq10 Mutant Has Almost Normal Concentrations of Coenzyme Q 6 in Mitochondria-The biochemical phenotype of the coq10 mutant, particularly the requirement of exogenous coenzyme Q 2 for the NADH and succinate oxidase activities, are very similar to the properties of other coq mutants unable to synthesize this obligatory electron carrier (5). The concentration of Q 6 in the mutant was determined by HPLC separation and quantification of mitochondrial lipid extracts on a reverse phase column with an electrochemical detector set for optimal detection of Q 6 . Samples were analyzed against authentic standards of known concentration. The HPLC-ECD elution profile indicated the presence of a redox-active compound with a retention time consistent with that of coenzyme Q 6 ( Fig. 4). The identity of this compound as Q 6 was confirmed by mass spectrometry (Fig. 5), which revealed the presence in lipid extracts of the wild type and coq10 null mutant of the correct molecular ion with a mass/charge ratio (m/z) of Q 6 at 591.4. This ion produced characteristic fragment ions including the predominant base peak and tropyllium ion at m/z 197, a definitive breakdown product of Q 6 (35). The concentration of Q 6 in the mutant was estimated to be 3.32 g/mg of mitochondrial protein (Fig. 3B). The value for the wild type parent was 3.33 g/mg protein.
Suppressors of coq10 Mutants-BY⌬COQ10 and aW303⌬COQ10 give rise to revertants when placed on a selective medium containing ethanol and glycerol as carbon sources. Diploid cells issued from back crosses of o derivatives of several revertants to the mutant had the revertant phenotype, indicating that the suppressor mutations were nuclear in origin. The nuclear DNA of a BY⌬COQ10 revertant was used to construct a genomic plasmid library in the high-copy plasmid YEp352 (22). Transformation of aW303⌬COQ10 with this library yielded a number of respiratory competent clones, which had plasmids containing COQ2, COQ7, or COQ8. Of the three genes the most effective in restoring respiration was COQ8, followed by COQ7, whereas the

Growth properties of coq mutants
Overnight cultures of wild type and mutant strains were inoculated in liquid YPEG medium, and grown at 30°C. Coenzyme Q 6 was added to a final concentration of 2 or 15 M. Growth was monitored by absorbance at 600 nm. Both the coq5 (8) and coq9 (4) mutants are depleted of Q 6 . N.D., not done Strain Generation time (hs) 11.8 8.6 8.5 least effective was COQ2 (Fig. 6). The NADH oxidase activity measured in isolated mitochondria was also highest in the clone with the COQ8-(80% of wild type) and lowest in the clone with the COQ2-bearing plasmid (11% of wild type) (not shown).    Transformants harboring the different suppressors either on a multicopy plasmid or integrated into nuclear DNA did not grow on the non-fermentable carbon sources as well as the revertant (Fig. 6). This suggested that the three COQ genes did not have the suppressor mutation conferring respiratory competence to the revertant. Growth of the mutant on non-fermentable substrates was substantially improved when COQ2, COQ7, and COQ8 were introduced on a high-copy episomal plasmid rather than integrated into chromosomal DNA in single copy, suggesting that the rescue is because of a higher gene dosage allowing for more efficient synthesis of Q 6 (Fig. 6). This is supported by the increased Q 6 content (150 -220% of wild type; Fig. 4) in mitochondria of the transformant with COQ8 on the high-copy plasmid. This strain also exhibited a greater stability of mitochondrial DNA, as stationary cultures of the coq10 null mutant transformed with COQ8 had very few secondary o/Ϫ clones.
Complementation of the Yeast coq10 Mutant by a Human COQ10 Homologue-A search of the current protein data base indicates that Coq10p homologues exist in diverse phylogenetic groups spanning bacteria and humans (Fig. 7A). Coq10p is a member of the large START superfamily (36) that is composed of polyketide cyclases/lipid transport proteins, aromatases, RNases, and other proteins of unknown function such as members of the ARPF/Pfam03654 family ("aromatic-rich protein family"). START proteins are defined as having a structurally similar, although not necessarily homologous, ligand binding pocket of the helix-grip type (36), formed from seven anti-parallel ␤ strands apposed to a C-terminal helix. Many START proteins bind lipophilic compounds and some have been shown to be involved in lipid and sterol transport and delivery to membranes (37,38).
To confirm that the human gene codes for a functional analogue of yeast Coq10p, the human cDNA was cloned in a yeast multicopy plasmid and was tested for complementation of the yeast coq10 null mutant. The functional identity of the yeast and human COQ10 (hCOQ10) was confirmed by rescue of growth of the yeast mutant on glycerol/ethanol by human cDNA (Fig. 7B). FIGURE 6. Suppression of aW303⌬COQ10 by COQ2, COQ7, and COQ8. Serial dilutions of the wild type W303-1A, the coq10 null mutant aW303⌬COQ10, the revertant aW303⌬COQ10/R1, and of the null mutant transformed with the COQ genes either in a multicopy plasmid (ϩCOQ2, ϩCOQ7, and ϩCOQ8) or integrated into nuclear DNA (ϩiCOQ2, ϩiCOQ7, and ϩiCOQ8) were spotted on rich glucose medium (YPD) and on rich medium containing glycerol plus ethanol as carbon sources (YPEG). The photograph was taken after the plates had been incubated at 30°C for 2 days. FIGURE 7. Alignment of putative Coq10p homologues and complementation of the yeast coq10 mutant by the human COQ10 cDNA. A, the C. crescentus CC1736 protein is a member of the Pfam03654 family. Identical residues are highlighted in black. The six residues in CC1736 inferred from structural considerations to be important for ligand binding (19) are marked by asterisks. B, serial dilutions of the wild type W303-1A, the coq10 null mutant aW303⌬COQ10, and the mutant transformed with the human COQ10 cDNA (hCOQ10) were spotted on rich glucose (YPD) and rich ethanol/glycerol (YPEG) plates and incubated at 30°C for 48 h.

Localization and Sizing of Coq10p-A polyclonal antibody against
Coq10p was used to detect the protein in yeast. This antibody recognized a protein of ϳ23-25 kDa in wild type mitochondria. This protein was absent in the coq10 null mutant and was restituted in the mutant containing a copy of COQ10 integrated at the leu2 locus (Fig. 8A), thereby confirming the identity of the protein as Coq10p. Coq10p with a carboxyl-terminal polyhistidine tag was purified from yeast mitochondria (see below). The sequence of this protein starts with the phenylalanine at residue 31, indicating that the amino-terminal 30 residues of the primary translation product constitute a mitochondrial targeting sequence that is cleaved during transport.
No Coq10p was detected in the post-mitochondrial supernatant fractions consisting of cytosolic proteins (data not shown). Coq10p is a hydrophobic protein that is recovered in the membrane fraction following disruption of mitochondria by sonic irradiation (Fig. 8B), or by treatment with 0.2 M sodium carbonate (Fig. 8C) but is solubilized by dodecyl maltoside suggesting that it is an integral membrane protein (not shown). This is also supported by its hydrophobicity profile, which indicates at least four putative transmembrane domains.
Coq10p is an inner membrane protein facing the matrix compartment. This is evident from its lack of sensitivity to proteinase K in mitochondria and in mitoplasts, the latter lacking an intact outer mem-brane (Fig. 8D). Conversion of mitochondria to mitoplasts by hypotonic shock was evident from the substantial loss of the soluble intermembrane marker cytochrome b 2. This is also supported by the decrease of Sco1p as a result of the proteinase K treatment of mitoplasts but not mitochondria. Sco1p is an inner membrane protein facing the intermembrane space (40). The matrix protein ␣-ketoglutarate dehydrogenase was protected against proteinase K in mitochondria and mitoplasts confirming the intactness of the inner membrane in the mitoplasts.
Sucrose gradient sedimentation indicates that native Coq10p is considerably larger than the monomer. When extracted from mitochondria with 0.5% dodecyl maltoside the protein sediments slightly slower than lactate dehydrogenase, which has a molecular weight of 140,000 (Fig.  9A). At present we have no data to indicate whether Coq10p is a homoor hetero-oligomer. Coq10p does not co-sediment with either Coq3p or Coq5p but does overlap with Coq4p, which is broadly distributed indicating heterogeneity in size.
Steady-state Levels of Coenzyme Q Biosynthetic Enzymes in the coq10 Mutant-Mutations that block coenzyme Q biosynthesis lower the steady-state concentrations of a number of enzymes of this pathway (15,41). To determine whether these proteins are also reduced in the coq10 mutant, total mitochondrial proteins were separated by SDS-PAGE, and Western blots were probed with antibodies against Coq3p, Coq4p, and Coq5p. The results indicated that the coq10 mutant has signifi- FIGURE 9. Sedimentation properties and steady-state levels of Coq10p. A, mitochondria (10 mg/ml) were adjusted to 0.5% dodecyl maltoside and centrifuged at 150,000 ϫ g av for 10 min. To 0.5 ml of the clear supernatant was added 50 l of a 3% solution of hemoglobin and 15 l of an ammonium sulfate suspension containing 100 units of bovine lactate dehydrogenase. The mixture was applied on top of a 4.6-ml linear 10 -25% sucrose gradient containing 10 mM Tris-Cl, pH 7.5, and 0.05% Triton X-100. The gradient was centrifuged at 65,000 rpm in a Beckman SW65 rotor for 6 h. Fractions 3-12 of the 13 fractions collected were separated by SDS-PAGE, on a 12% polyacrylamide gel, transferred to nitrocellulose, and reacted with a rabbit antibody against Coq10p followed by a secondary reaction anti-rabbit IgG. The gradient fractions were also analyzed for the distribution of hemoglobin (A 410 ) and lactate dehydrogenase (LDH) by measuring the pyruvate-dependent oxidation of NADH at 340 nm. B, mitochondria from the wild type W303-1A, from the coq10 null mutant aW303⌬COQ10 (⌬COQ10), and from the null mutant transformed with COQ8 on a high copy plasmid (⌬COQ10 ϩ COQ8), with COQ8 integrated into nuclear DNA (⌬COQ10 ϩ iCOQ8), or COQ2 on a high copy plasmid (⌬COQ10 ϩ COQ2), were separated by SDS-PAGE and reacted with antibodies against Coq3p, Coq4p, and Coq5p as in A.

FIGURE 8. Localization of Coq10p.
A, mitochondria were prepared from the wild type W303-1A, from the coq10 null mutant aW303⌬COQ10 (⌬COQ10), and from the null mutant with the wild type gene integrated at the LEU2 locus, aW303⌬COQ10/ST9 (⌬COQ10/ST9). Mitochondria (40 g) were separated by SDS-PAGE (30) on a 12% polyacrylamide gel and transferred to nitrocellulose. The Western blot was reacted with a rabbit polyclonal antibody against Coq10p. Following a secondary reaction with peroxidase-conjugated anti-rabbit IgG (Sigma), proteins were visualized with the SuperSignal chemiluminescent substrate kit (Pierce). B, mitochondria at a protein concentration of 10 mg/ml were sonically irradiated for 5 s with a Branson microtip sonic probe. The sample was centrifuged at 100,000 ϫ g av for 30 min and the clear part of the supernatant was collected. The packed part of the pellet was resuspended in the starting volume of 0.6 M sorbitol, 10 mM Hepes, pH 7.5. An intermediate oily layer consisting of very small membrane fragments was discarded. The mitochondria (Mit) representing 40 g of protein and equivalent volumes of the supernatant (Sup) and pellet (SMP) were separated on a 12% polyacrylamide gel and processed as in A. C, wild type mitochondria were converted to submitochondrial particles as in B. The submitochondrial particles at a protein concentration of 10 mg/ml were mixed with an equal volume of 0.2 M sodium carbonate and incubated on ice for 30 min. The extracted membranes were separated from the solubilized proteins by centrifugation at 100,000 ϫ g av for 20 min. The mitochondrial (Mit) and equivalent volumes of the extracted membranes (C-pellet), and carbonate soluble proteins (C-sup) were separated on a 12% polyacrylamide gel. D, wild type mitochondria (Mit) were converted to mitoplasts (Mpl) by the method of Glick (39). Both mitochondria and mitoplasts were incubated on ice for 1 h in the absence or presence of 0.1 mg/ml proteinase K. Following addition of phenylmethylsulfonyl fluoride mitochondria were isolated by centrifugation for 10 min at 14,000 ϫ g. Equivalent amounts (40 g) of mitochondria and mitoplasts were separated by SDS-PAGE and further processed as in panel A. cantly reduced concentrations of Coq4p but not of Coq5p (Fig. 9B). Coq3p was also reduced in the mutant but not to the same degree as Coq4p. This pattern is very similar to that observed in mutants from the eight complementation groups previously shown to be Q 6 deficient (15). In the transformant harboring COQ8 on a high copy, both Coq3p and Coq4p were restored to levels even higher than those observed in wild type. There was also a significant restoration of Coq3p and Coq4p in the strain in which COQ8 was integrated into the nuclear genome. Transformation of the mutant with COQ2 on a high copy plasmid, however, did not affect the level of either Coq3p or Coq4p (Fig. 8B).
Analysis of Q6 in Purified Coq10p-his-The presence of a START domain in Coq10p made it of interest to determine whether it is a Q 6 -binding protein. Coq10p was purified from mitochondria of a coq10 mutant transformed with a COQ10 gene modified so as to express the protein with six histidine residues at the carboxyl terminus. This construct complemented the coq10 mutant when expressed from an episomal multicopy plasmid or integrated in single copy into chromosomal DNA. Complementation by the single copy integrated gene indicates that the presence of the histidine tag does not affect the activity of the protein. The histidine-tagged protein (Coq10p-his) was purified from the COQ10 overexpressing strain aW303⌬COQ10/ST13 after extraction of mitochondria with 1% dodecyl maltoside in the presence of salt (Fig. 10A). Most of the Coq10p-his in the extract was adsorbed on Ni-NTA and was eluted at high concentration of imidazole. A highly purified preparation of Coq10p-his was obtained by chromatography of the Ni-NTA eluate on a Mono S column (Fig. 10B). A mock run was done with the same Mono S column developed under identical conditions but without prior application of protein. Two independent preparations of the protein (Fig. 10B) and analogous mock column fractions were analyzed for Q 6 by three sensitive, quantitative mass spectrometric analyses. A robust Q 6 signal was present in column fractions with Coq10p-his as compared with the corresponding fractions from a mock column (Fig. 10D). The amount of Q 6 co-purified with Coq10p-his was calculated to be 0.032-0.034 mol of Q 6 /mol of Coq10p-his.
A comparison of the Western signals of Coq10p in mitochondria to known amounts of purified Coq10p-his indicated that there is ϳ0.1 g or 5 pmol of Coq10p/mg of mitochondrial protein (Fig. 10C). This value is some 3 orders of magnitude lower than the mitochondrial concentration of Q 6 . In aW303⌬COQ10/ST13 the comparable value for Coq10phis was 0.5 nmol/mg of mitochondrial protein. This corresponds to a 100-fold increase when the protein is expressed from the high copy plasmid (not shown).

DISCUSSION
Deletion of reading frame YOL008W/COQ10 of S. cerevisiae chromosome XV confers a respiration defect. Here we show that mitochondrial NADH and succinate oxidase activities are decreased in coq10 mutants but that both activities are reconstituted by addition of Q 2 to the assay. Q 2 -dependent respiration is a hallmark of mutants blocked in Q 6 synthesis (5). Unlike other coq mutants, which are deficient in Q 6 (4, 6 -16), the coq10 null mutant reported in this study contains almost FIGURE 10. Purification of Coq10p-his and analysis of its Q 6 content. A, extraction and affinity purification of Coq10p-his on Ni-NTA. Mitochondria from aW303⌬COQ10/ST13 grown on rich 2% galactose medium (YPGal) were sonically irradiated for 30 s with a Branson sonifier. The submitochondrial particles obtained by centrifugation of the disrupted mitochondria at 100,000 ϫ g av for 30 min were suspended in 0.6 M sorbitol, 20 mM Tris-HCl, pH 7.5, 0.5 mM EDTA and adjusted to 1% dodecyl maltoside and 1 M NaCl at a final protein concentration of 10 mg/ml. The extract obtained after centrifugation at 100,000 ϫ g av was adjusted to pH 8 with Tris base, mixed with Ni-NTA beads (Qiagen) at a ratio of 0.2 ml of beads per 10-ml extract, and mixed by rotation for 2 h at 4°C. The slurry was poured into a column and the unadsorbed proteins followed by 5 column volumes of 10 mM imidazole was collected as a single fraction (Flow-tru) by gravity flow. The beads were then washed with 5 volumes of 30 mM imidazole followed by 50 mM imidazole in 50 mM NaCl, 10 mM Tris-HCl, pH 7.5. The bulk of the protein was eluted in a final wash with 3 column volumes of 250 mM imidazole in 50 mM NaCl, 10 mM Tris-HCl, pH 7.5. This fraction was diluted with an equal volume of 10 mM Tris-HCl, pH 7.5, and applied to a 1-ml Pharmacia Mono S column. The column was developed sequentially with 10 ml of 50 mM NaCl, 5 ml of 300 mM NaCl, and 1 M NaCl in 10 mM Tris-HCl, pH 7.5. Most of the protein eluted as a trailing peak in the 300 mM NaCl fraction. Mock column fractions were obtained by replicating the Mono S column purification without protein. The submitochondrial particles (SMP), dodecyl maltoside (LM) extract, and the different imidazole washes of the Ni-NTA column were separated by SDS-PAGE and the Western blot was probed with the antibody against Coq10p as in the legend to Fig. 8A. Each fraction was adjusted for volume except that three times as much of the 250 mM imidazole elute was applied to the gel. B, silver-stained gel of Coq10p-his eluted in the 300 mM NaCl fraction from the Mono S column. The two different preparations of the protein used for the Q 6 analyses are shown. The faster migrating band seen in the gel of preparation 1 is detected by the Coq10p antibody and is probably a proteolytic product. C, the indicated amounts of purified Coq10p-his and mitochondrial proteins were separated by SDS-PAGE on a 12% polyacrylamide gel. The Western blot was processed as described in the legend to Fig. 8A. D, three separate lipid extractions and quantitative mass spectral measurements were performed on the lipid extract obtained from two preparations of Coq10p-his shown in B, mock column fractions and calibration curve samples. All extractions contained Q 4 as a internal standard, and the amounts of Q 6 standards ranged from 6 fmol/l to 2 pmol/l. The spectra shown are the relative intensities of the molecular ion for Q 6 DECEMBER 30, 2005 • VOLUME 280 • NUMBER 52 normal concentrations of Q 6 in mitochondria. This excludes Coq10p from being required for Q 6 biosynthesis directly.

COQ10 Is Required for Expression of Functional Coenzyme Q
Coq10p is a membrane protein located in the inner membrane. The 30 amino-terminal residues of the protein encoded by COQ10 are absent in the mature protein. This sequence is probably a mitochondrial targeting signal that is cleaved during transport. Coq10p is resistant to proteinase K in mitoplasts suggesting that it faces the matrix compartment. This localization is similar to all the other Coq polypeptides that function in Q 6 biosynthesis in yeast (9 -11, 15). Other features of coq mutants are also shared by the coq10 mutant. Coq3p (6), Coq4p (9), and Coq5p (8) have previously been shown to be required for Q 6 synthesis. Coq3p and Coq5p function, respectively, as O-and C-methyltransferases (see Ref. 11 for the biosynthetic pathway). Western analysis of these proteins indicates a large decrease of Coq4p and some reduction of Coq3p but not Coq5p in the coq10 mutant. This pattern is very similar to what has been reported for mutants in the eight coq genes (15). The pleiotropic effects of mutations in a single coq gene on the steady-state concentrations of other coq gene products provide genetic evidence that stability of these proteins may depend either on their association into complexes or on the presence of one or more biosynthetic intermediates of the pathway (15). Recent evidence suggests that Coq4p may be a structural component of a Q-biosynthetic polypeptide complex (42).
Results presented here suggest that Coq10p exists as a homo-oligomer or is complexed with other proteins. This is supported by the five times larger size of native Coq10p than the monomer. Coq10p sediments differently than Coq3p or Coq5p on sucrose gradients. Coq4p is broadly distributed in the gradient suggestive of several sedimenting species. The peak of Coq10p does coincide with that of Coq4p. Whether this is coincidental or indicative of an association of the two proteins cannot be concluded at present. The presence of normal amounts of Q 6 despite the large reduction of Coq4p suggests that the mitochondrial concentration of the latter protein is substantially greater than what is needed for Q 6 biosynthesis.
COQ2, COQ7, and COQ8 are high copy suppressors of the coq10 mutant. COQ8, when present on a multicopy plasmid, is the most effective of the three. The efficiency of suppression, judged by growth on glycerol/ethanol, correlates with an increase in the mitochondrial concentrations of Coq4p and Coq3p and a greater stability of mitochondrial DNA. Overexpression of COQ8 also caused a 150 -220% increase in the mitochondrial concentration of Q 6 . These observations indicate that the absence of Coq10p can be compensated to some extent by higher than normal concentrations of Q 6 in mitochondria. When adjusted for the 50% of o/Ϫ cells in cultures of the mutant, the NADH and succinate oxidation rates in the mutant correspond to 10% of wild type. This indicates that the absence of Coq10p does not totally abolish the ability of the mitochondria to oxidize and reduce endogenous Q 6 and helps to explain the observed partial rescue of the mutant phenotype by supplementation of growth media with Q 6 , or by the COQ8-induced higher mitochondrial concentration of Q 6 .
In addition to the three suppressors identified from transformations with the plasmid library, we also tested COQ1, COQ5, and COQ6. These genes did not improve growth of the coq10 mutant on glycerol/ethanol, suggesting that the encoded enzymes are not rate-limiting in Q 6 biosynthesis.
Coq10p is homologous to members of the Pfam03654 protein family, including the 18-kDa Caulobacter crescentus protein, C1736, whose solution structure was recently reported (19). This bacterial protein of unknown function has a structure similar to the mouse StarD4 cholesterol transport protein (21). The latter has a hydrophobic tunnel capable of accommodating one cholesterol molecule. A comparison of the CC1736 and the StarD4 structures has helped to identify amino acid residues likely to confer ligand specificity to the binding pocket (19,21). It is striking that four of the six residues proposed to be important for ligand recognition are conserved in CC1736 and Coq10p (Fig. 7A). The presence of this putative binding site in Coq10p suggests that this protein and its homologues may bind coenzyme Q. This is supported by the presence of Q 6 in purified preparations of Coq10p-his. The amount of Q 6 detected, however, is not stoichiometric with the protein. Because the extraction and subsequent fractionation of Coq10p-his entailed the use of detergent, it is possible that most of the Q 6 was lost during the purification. The substoichiometic concentration of Q 6 associated with Coq10p-his could also be because of the fact that the protein was purified from a highly overexpressing strain of yeast.
Because Coq10p is not a subunit of the bc 1 complex it is unlikely to function in the catalytic mechanism of this respiratory complex. The same argument applies to the succinate-and NADH-coenzyme Q reductase complexes of yeast, both of which have known subunit compositions. Coq10p may be a chaperone for Q 6 within a newly identified high molecular weight complex containing other Coq polypeptides (42). Alternatively, Coq10p may function in transporting Q 6 from its site of synthesis to the catalytic sites of the respiratory chain complexes where it is used. Another function of Coq10p could be to shuttle coenzyme Q from the dehydrogenase complexes to the bc 1 complex during electron transport. This, however, seems highly unlikely in view of the mitochondrial Coq10p concentration, which we estimate to be 3 orders of magnitude lower than that of Q 6 and at least 2 orders of magnitude lower than most components of the respiratory chain.