COQ11 deletion mitigates respiratory deficiency caused by mutations in the gene encoding the coenzyme Q chaperone protein Coq10

Coenzyme Q (Qn) is a vital lipid component of the electron transport chain that functions in cellular energy metabolism and as a membrane antioxidant. In the yeast Saccharomyces cerevisiae, coq1–coq9 deletion mutants are respiratory-incompetent, sensitive to lipid peroxidation stress, and unable to synthesize Q6. The yeast coq10 deletion mutant is also respiratory-deficient and sensitive to lipid peroxidation, yet it continues to produce Q6 at an impaired rate. Thus, Coq10 is required for the function of Q6 in respiration and as an antioxidant and is believed to chaperone Q6 from its site of synthesis to the respiratory complexes. In several fungi, Coq10 is encoded as a fusion polypeptide with Coq11, a recently identified protein of unknown function required for efficient Q6 biosynthesis. Because “fused” proteins are often involved in similar biochemical pathways, here we examined the putative functional relationship between Coq10 and Coq11 in yeast. We used plate growth and Seahorse assays and LC-MS/MS analysis to show that COQ11 deletion rescues respiratory deficiency, sensitivity to lipid peroxidation, and decreased Q6 biosynthesis of the coq10Δ mutant. Additionally, immunoblotting indicated that yeast coq11Δ mutants accumulate increased amounts of certain Coq polypeptides and display a stabilized CoQ synthome. These effects suggest that Coq11 modulates Q6 biosynthesis and that its absence increases mitochondrial Q6 content in the coq10Δcoq11Δ double mutant. This augmented mitochondrial Q6 content counteracts the respiratory deficiency and lipid peroxidation sensitivity phenotypes of the coq10Δ mutant. This study further clarifies the intricate connection between Q6 biosynthesis, trafficking, and function in mitochondrial metabolism.


Coenzyme Q (Q n ) is a vital lipid component of the electron transport chain that functions in cellular energy metabolism
and as a membrane antioxidant. In the yeast Saccharomyces cerevisiae, coq1-coq9 deletion mutants are respiratory-incompetent, sensitive to lipid peroxidation stress, and unable to synthesize Q 6 . The yeast coq10 deletion mutant is also respiratorydeficient and sensitive to lipid peroxidation, yet it continues to produce Q 6 at an impaired rate. Thus, Coq10 is required for the function of Q 6 in respiration and as an antioxidant and is believed to chaperone Q 6 from its site of synthesis to the respiratory complexes. In several fungi, Coq10 is encoded as a fusion polypeptide with Coq11, a recently identified protein of unknown function required for efficient Q 6 biosynthesis. Because "fused" proteins are often involved in similar biochemical pathways, here we examined the putative functional relationship between Coq10 and Coq11 in yeast. We used plate growth and Seahorse assays and LC-MS/MS analysis to show that COQ11 deletion rescues respiratory deficiency, sensitivity to lipid peroxidation, and decreased Q 6 biosynthesis of the coq10⌬ mutant. Additionally, immunoblotting indicated that yeast coq11⌬ mutants accumulate increased amounts of certain Coq polypeptides and display a stabilized CoQ synthome. These effects suggest that Coq11 modulates Q 6 biosynthesis and that its absence increases mitochondrial Q 6 content in the coq10⌬coq11⌬ double mutant. This augmented mitochondrial Q 6 content counteracts the respiratory deficiency and lipid peroxidation sensitivity phenotypes of the coq10⌬ mutant. This study further clarifies the intricate connection between Q 6 biosynthesis, trafficking, and function in mitochondrial metabolism.
Coenzyme Q (ubiquinone or Q) 2 is a benzoquinone lipid that functions as an essential electron carrier within the electron transport chain (1). Because of its redox activities, Q is a versatile electron acceptor in biological pathways such as cellular respiration, oxidation of proline and sulfide, fatty acid ␤-oxidation, and pyrimidine biosynthesis (1)(2)(3). The reduced hydroquinone form of Q (ubiquinol or QH 2 ) also serves as an important chain-breaking antioxidant shown to alleviate lipid peroxidative damage in cellular membranes (4).
For proper functional localization, Q relies on its polyisoprenoid tail to remain anchored at the mid-plane of phospholipid bilayers. The number of isoprene units (n) that comprise the polyisoprenoid tail of Q n depends on a species-specific polyprenyl diphosphate synthase (5), with Q 10 representing the major isoform in humans (6). Patients unable to produce adequate levels of Q 10 display a wide variety of health issues that stem from mitochondrial dysfunction across tissues (7). Attempts to ameliorate the consequences of primary Q 10 deficiency by early Q 10 supplementation have been partially successful in some cases (8); however, many patients fail to demonstrate full recovery, which is related to inefficient uptake of orally-supplied Q 10 . Because of the striking homology between human COQ genes and those of Saccharomyces cerevisiae (7,9), studies of Q 6 biosynthesis in S. cerevisiae may provide insight into human Q 10 biosynthesis, leading to the discovery of potential therapeutic targets.
In S. cerevisiae, at least 14 nuclear-encoded mitochondrial proteins (Coq1-Coq11, Yah1, Arh1, and Hfd1) drive Q 6 biosynthesis (7,9). Many Coq polypeptides (Coq3-Coq9, and Coq11) are localized to the matrix side of the mitochondrial inner membrane, where they organize into a high-molecularweight multisubunit complex known as the "CoQ synthome" (7,9). Several lines of evidence suggest that correct assembly of the CoQ synthome is necessary for efficient Q 6 biosynthesis (9 -12). In fact, deletion of certain COQ genes results in decreased levels of other Coq polypeptides and contributes to a destabilized CoQ synthome in these mutants (12,13). Recently, a protein of unknown function encoded by the ORF YLR290C was identified to associate with the CoQ synthome, via proteomic analysis of tandem affinity-purified tagged Coq proteins (14). YLR290C copurified with Coq5, Coq7, and Coq9, in addition to Q 6 and late-stage Q 6 -intermediates (14). Furthermore, the ylr290c⌬ mutant exhibited impaired de novo Q 6 biosynthesis, despite preserving growth on a nonfermentable carbon source (14). Given its effects on Q 6 biosynthesis and involvement with the CoQ synthome, YLR290C was renamed Coq11 (14).
In several fungi, Coq11 and Coq10 have evolved as fusion proteins (14), suggesting that Coq11 may have a functional relationship with Coq10 (15). High-throughput genetic analyses found COQ11 to correlate with both COQ2 and COQ10 (16). Whereas the coq mutants generally lack Q 6 , the coq10⌬ mutant is different because it produces near WT amounts of Q 6 in stationary phase and only has decreased de novo Q 6 biosynthesis in log phase (17,18). Although Q 6 biosynthesis is only minimally decreased in the absence of COQ10, the coq10⌬ mutant has decreased NADH and succinate oxidase activity and displayed sickly growth on respiratory medium (18). The coq10⌬ mutant is sensitive to lipid peroxidation initiated by exogenously supplemented polyunsaturated fatty acids (PUFAs), indicating that the Coq10 polypeptide is also required for antioxidant protection by Q 6 (17,19).
The NMR structure of a Coq10 ortholog in Caulobacter crescentus was shown to possess a steroidogenic acute regulatory protein-related lipid transfer (START) domain (20) that can directly bind Q and late-stage Q-intermediates (17). Purified Coq10 from either S. cerevisiae or Schizosaccharomyces pombe eluted with the respective species' Q isoform (18,21). This observation has prompted speculation that Coq10 acts as a Q 6 chaperone protein required for delivery of Q 6 from its synthesis site to sites where Q 6 functions as an antioxidant and to the respiratory complexes, thereby bridging efficient de novo Q 6 biosynthesis with respiration (17). Recent studies have shown a spatial compartmentalization of the mitochondrial inner membrane with the identification of different sites, such as the inner boundary membrane, the cristae membrane, and the ER-mitochondrial contact sites (22)(23)(24)(25). Thus, for optimal respiratory competence, newly-synthesized Q 6 must move from its site of synthesis (i.e. the ER-mitochondrial contact sites (23,24)) to the cristae membrane where the respiratory complexes are concentrated (22). The presence of Coq10 -Coq11 fusions in fungal species indicates that Coq11 may have a functional association with the Coq10 chaperone to facilitate or regulate Q 6 transport for respiration in yeast.
In this work, the functional relationship between Coq10 and Coq11 was investigated using a series of single-and doubleknockout mutants. Deletion of COQ11 alleviated the coq10⌬ respiratory defect, increased Coq polypeptides and CoQ synthome stability, and partially rescued Q 6 production. Based on this evidence, we propose a novel function for Coq11 as a negative modulator of Q 6 biosynthesis in the mitochondria.

Coq10 and Coq11 reside in similar compartments within the mitochondria
Previous phylogenetic analyses of numerous fungi revealed that Coq11-like proteins are fused to Coq10 (14). Protein fusions often indicate a functional relationship between corresponding homologs in other organisms, such as direct proteinprotein interaction or operation within the same biological pathway (15). Although Coq10 and Coq11 are not physically fused in yeast (Fig. 1A), we sought to investigate whether there is a functional link between the two proteins. Because protein localization is often associated with function, we first performed mitochondrial fractionation to localize both Coq10 and Coq11. Coq10 has been localized previously (18), but fractionation was re-performed here in the context of Coq11.
S. cerevisiae mitochondria were fractionated as described under "Experimental procedures." Purified mitochondria were incubated in hypotonic buffer to disrupt the outer membrane and release soluble components of the intermembrane space (IMS). The inner membrane was kept intact following hypotonic buffer treatment, protecting inner membrane and matrix proteins. Analysis of the fractions via immunoblot suggested that both Coq10 and Coq11 remained associated with the mitoplast fraction as opposed to colocalizing with the IMS marker cytochrome b 2 . Mitoplasts were further fractionated after sonication to separate soluble matrix components (supernatant, S) from membrane components (pellet, P). The soluble matrix marker Hsp60 was partially released into the supernatant by sonication as demonstrated in earlier work (26). Although Coq11 remained associated with the membrane fraction, Coq10 was partially dissociated in a similar manner to Hsp60 (Fig. 1B). Previous Coq10 colocalization following sonication demonstrated that Coq10 was solely associated with the membrane fraction (18). The detection of Coq10 in the supernatant shown in Fig. 1B may be due to increased sensitivity of the polyclonal antisera used in this study.
Alternatively, mitoplasts were subjected to alkaline carbonate extraction to separate peripheral membrane components (supernatant, S) from integral membrane and matrix components (pellet, P) (27). Coq10 and Coq11 were released into the supernatant following alkaline treatment (Fig. 1B), matching the peripheral inner membrane marker Atp2 (28). There was no colocalization with the pellet fraction, marked by the integral membrane protein Cyt 1 (29). These results indicate that Coq10 and Coq11 are both peripheral inner membrane proteins, and Coq10 has additional localization to the matrix. The localization of Coq10 to the inner membrane is consistent with Coq10 knockout phenotypes are rescued by deletion of COQ11 its putative role as a START domain protein possessing a hydrophobic cavity to bind and chaperone Q 6 from its site of synthesis to complex III for respiration (17,18). We hypothesize that dual mitochondrial matrix localization occurs when Coq10 is tightly bound to protein partners to decrease its hydrophobicity.
For better insight into the membrane association of Coq10 and Coq11, intact mitochondria or mitoplasts were treated with proteinase K in the absence or presence of two individual detergents (1% Triton X-100 or 0.5% SDS). Coq10 and Coq11 were both protected from protease treatment in purified mito-chondria and mitoplasts, as was the matrix marker Hsp60 (Fig.  1C). When protease was used in the presence of either detergent in mitochondria or mitoplasts, all proteins became sensitive to the protease and were degraded. Expanding on the subfractionation results, these data indicate that Coq10 and Coq11 polypeptides are peripherally associated with the inner membrane facing the matrix side in yeast mitochondria, and Coq10 is also found in the mitochondrial matrix itself. The mitochondrial peripheral membrane association of these two proteins is also in agreement with their submitochondrial localization previously identified in a study of the yeast mitochondrial proteome (30). associated with the mitochondrial inner membrane facing the matrix, and Coq10 is additionally found in the mitochondrial matrix. A, Coq10 and Coq11 are fused in multiple fungi, suggesting an evolutionarily functional relationship between these proteins, although they are not found fused in S. cerevisiae. B, S. cerevisiae mitochondria purified from yeast strains cultured on YPGal medium were subjected to hypotonic swelling and centrifugation to separate the IMS proteins from mitoplasts. The mitoplasts were alkaline-treated (Na 2 CO 3 ; pH 11.5) or sonicated and then separated by centrifugation (100,000 ϫ g for 1 h) into supernatant (S) or pellet (P) fractions. C, intact mitochondria or mitoplasts were treated with 100 g/ml proteinase K for 30 min on ice, with or without detergent. Mitochondrial polypeptide markers are as follows: Atp2, peripheral inner membrane protein; Cyb 2 , intermembrane space protein; Cyt 1 , integral inner membrane protein; and Hsp60, soluble matrix protein. Results are representative of two experiments.

Coq10 knockout phenotypes are rescued by deletion of COQ11 coq10⌬ respiratory defect is alleviated by deletion of COQ11
Based on similar mitochondria localization and genetic evolutionary evidence, a putative functional relationship between Coq10 and Coq11 was further probed using a series of coq10 and coq11 single-and double-knockout mutants. Strain descriptions are listed in Table 1. The Coq10 polypeptide is required for respiration in yeast, and mutants lacking coq10 have poor growth on nonfermentable carbon sources, including YPGlycerol, hereafter referred to as "YPG" (18). Unlike deletion of COQ10, coq11⌬ mutants are respiratory-capable and have comparable growth to WT on nonfermentable carbon sources (14). When COQ11 was deleted in a coq10⌬ mutant in two different yeast genetic backgrounds, the sickly growth of coq10⌬ on nonfermentable YPG was rescued ( Fig. 2A).
Quantitative respiratory capacity of each mutant was evaluated with an XF96 Extracellular Flux Analyzer (Fig. 2, B and C). Representative and normalized traces of oxygen consumption rates (OCR) of four independent experiments performed in nonrepressive medium (YPGal) are shown in Fig. 2B. Basal rates of OCR were measured prior to the addition of any small molecule inhibitors. Consistent with its slow growth on nonfermentable medium, the coq10⌬ mutant had a low rate of basal oxygen consumption compared with WT (p ϭ 0.052) (Fig. 2C). Basal OCR was rescued in the coq10⌬coq11⌬ double mutant (Fig. 2C). Following the addition of two sequential injections of FCCP, a mitochondrial oxidative phosphorylation uncoupler, maximal respiration was also quantified. The maximal respiration of coq10⌬coq11⌬ was rescued to that of WT (Fig. 2C). These results show that the deletion of COQ11 in a coq10⌬ mutant confers a beneficial effect, such that both growth on respiratory medium and OCR are rescued to WT.

Deletion of COQ11 rescues PUFA sensitivity of the coq10⌬ mutant
PUFA autoxidation is initiated by the radical-mediated abstraction of vulnerable hydrogen atoms at bis-allylic posi-tions (31). The ensuing carbon-centered radical adds to molecular oxygen to form a lipid peroxyl radical that propagates lipid peroxidation, with the resulting lipid hydroperoxides ultimately driving cellular toxicity (32). The coq10⌬ mutant is sensitive to treatment with exogenous PUFAs (Fig. 2D) (17,19), likely because the Q 6 chaperone function of Coq10 is required for the antioxidant function of Q 6 . Attenuated respiration in coq10⌬ is rescued in the coq10⌬coq11⌬ double knockout (Fig.  2, A-C), presumably through regained function of Q 6 in the electron transport chain. To test whether the antioxidant capability of Q 6 is also restored in the coq10⌬coq11⌬ mutant, yeast strains were evaluated for sensitivity to added PUFAs (Fig. 2D). As anticipated, all strains were resistant to treatment with the monounsaturated oleic acid (Fig. 2D). Q 6 -less coq9⌬ was sensitive to ␣-linolenic acid due to the lack of Q 6 antioxidant protection (Fig. 2D). Conversely, the Q 6 -replete yet respiratorydeficient cor1⌬ remained resistant to ␣-linolenic acid (Fig. 2D). Deletion of COQ11 rescued the ␣-linolenic acid sensitivity of the coq10⌬ mutant, suggesting that the double knockout has restored Q 6 antioxidant protection (Fig. 2D) despite the absence of Coq10 as a Q 6 chaperone.

Independent coq10 revertant with rescued growth on respiratory medium harbors a mutation within COQ11
Although the coq10⌬ mutant is unable to grow robustly on nonfermentable medium, an earlier study identified a spontaneous coq10 revertant (coq10rev) that arose when coq10⌬ yeast was cultured for several weeks on nonfermentable medium containing ethanol and glycerol as carbon sources (18). Characterization of this revertant revealed a suppressor mutation within the COQ11 ORF, resulting in a truncated Coq11 protein that is predicted to be nonfunctional (Fig. 3A). This mutation was further assessed for dominance to determine whether it was sufficient to explain the respiratory competence of coq10rev. A haploid coq10⌬ mutant crossed with haploid coq10rev produced diploid MB-10 (Table 1), which was capable of growth on respiratory medium (Fig. 3B). Illustrated growth

Coq10 knockout phenotypes are rescued by deletion of COQ11
patterns suggest that the coq11 truncated allele present in coq10rev is a dominant-negative mutation. Because the dominant mutation in coq10rev restores growth on respiratory medium via a functionally suppressive Coq11 truncation mutation, this mutant effectively validates the coq10⌬coq11⌬ phenotype in an independent system.

Deletion of COQ11 fails to fully restore coq10⌬ Q 6 biosynthesis in whole cells
When a coq mutant displays anemic growth on respiratory medium, it is often indicative of inefficient Q 6 biosynthesis (7,9); yeast lacking COQ10 exhibit both poor growth on respiratory medium and decreased Q 6 biosynthesis in log phase whole cells (17,19). Although the coq11⌬ mutant retains the ability to grow on nonfermentable medium, it is also characterized by impaired Q 6 biosynthesis (14). Only a small amount of Q 6 is required for growth on respiratory medium, ϳ0.2-3% of the total Q 6 found in WT (9,33,34). Because the coq10⌬coq11⌬ double mutant has rescued respiration, we wanted to assess whether recovered growth was accompanied by increased Q 6 biosynthesis. Whole-cell de novo-synthesized [ 13 C 6 ]Q 6 and [ 12 C]Q 6 were measured in yeast by feeding the quinone ringlabeled precursor, [ 13 C 6 ]4HB, or EtOH vehicle control (Fig. 4). These analyses were performed in the fermentable, nonrepressive YPGal medium (35) to match the conditions of experiments involving purified mitochondria.
Next, we quantified the concentrations of key Q 6 -intermediates in the same whole-cell yeast pellets. As shown previously Figure 2. COQ11 deletion rescues the lack of growth on YPG, low-oxygen consumption rates, and lost Q 6 antioxidant protection in the coq10⌬ mutant. A, strains were grown overnight in 5 ml of YPD, diluted to an A 600 ϭ 0.2 with sterile PBS, and 2 l of 5-fold serial dilutions were spotted onto fermentable (YPDextrose, YPD) or respiratory (YPGlycerol, YPG) medium, corresponding to a final A 600 ϭ 0.2, 0.04, 0.008, 0.0016, and 0.00032. Plates were incubated at 30°C, and growth was captured after 2 or 3 days. B and C, quadruplicates of 25-ml cultures of WT, coq10⌬, coq11⌬, and coq10⌬coq11⌬ yeast were grown in YPGal until they reached A 600 ϳ4. Yeast were diluted to an A 600 ϭ 0.1 in fresh YPGal and collected by centrifugation on poly-D-lysine-coated Seahorse XF96 microplates to assess oxygen consumption. B, representative traces of OCR of yeast strains with the XF96 extracellular flux analyzer. FCCP and antimycin A (AA) were sequentially added to evaluate mitochondrial respiratory states. Measurements were taken approximately every 4 min, as represented by points and their respective error bars. Four independent experiments were performed (Fig. S1), and each group of average traces represents 8 -10 technical replicates. C, quantification of basal and maximal (maximal electron transport activity induced by the uncoupler FCCP) OCR as obtained from four independent experiments (Fig. S1). The data show the mean Ϯ S.D., and the statistical significance as compared with WT is represented by *, p Ͻ 0.05. D, deletion of COQ11 in the coq10⌬ rescues PUFA sensitivity. Results are representative of three experiments. (17,19), the coq10⌬ mutant contained lower amounts of the late-stage intermediate [ 4D). In contrast to coq10⌬, the coq11⌬ mutant mirrored WT production of both de novo-synthesized and unlabeled early-and late-stage intermediates ( Fig.  4, C and D), as shown previously (14). Q 6 -intermediate trends in coq10⌬coq11⌬ matched those of the coq10⌬ mutant rather than coq11⌬ (Fig. 4, C and D). The low Q 6 content and accumulation of early-stage Q 6 -intermediates in the coq10⌬coq11⌬ double knockout suggest the absence of COQ10 still produces a notable effect on Q 6 biosynthesis, although respiratory capacity is rescued.

coq10⌬coq11⌬ double mutant has increased mitochondrial Q 6 compared with the coq10⌬ single mutant
Although Q 6 biosynthesis solely occurs within mitochondria, it is found in all cellular membranes (9). Therefore, Q 6 was quantified in both whole cells and purified mitochondria from mutant and WT cells cultured under the same conditions (36). Whole-cell Q 6 determined under mitochondrial purification conditions matched those determined in Figs. 4 and 5A. The coq10⌬coq11⌬ double mutant made slightly more Q 6 than the coq10⌬ single mutant, but overall less Q 6 compared with the coq11⌬ single mutant. All mutants had lower whole-cell Q 6 amounts than WT (Fig. 5A).
Similarly, mitochondrial Q 6 content per microgram of mitochondrial protein was lower in coq11⌬ than WT (Fig. 5B). However, deletion of COQ11 in the coq10⌬ mutant increased the mitochondrial Q 6 5-fold (Fig. 5B). Despite these profound differences in mitochondrial Q 6 content, mitochondrial mass was consistent between strains as determined by three distinct assays (Fig. 5, C-E). Increased mitochondrial Q 6 in the coq10⌬coq11⌬ double mutant compared with coq10⌬ indicates that the absence of COQ11 in part rescues defective Q 6 synthesis in the coq10⌬ mutant.

Low Coq protein content and destabilized CoQ synthome of the coq10⌬ mutant are restored in the coq10⌬coq11⌬ double mutant
Proper formation of the CoQ synthome from component Coq polypeptides is required for efficient Q 6 biosynthesis in yeast (9,12,13). Deletion of COQ10 causes a decrease in several other Coq polypeptides, including Coq3-Coq7, Coq9, as well as overall CoQ synthome destabilization (12,13,19). These results were confirmed when purified mitochondria from coq10⌬ yeast were analyzed for each Coq polypeptide (Fig. 6A). The coq10⌬ mutant had significantly decreased amounts of Coq3,  Table 1). Strains were grown overnight in 8 ml of YPDextrose (YPD), diluted to an A 600 ϭ 0.2 with sterile PBS, and 2 l of 5-fold serial dilutions were spotted onto fermentable YPD or respiratory (YPEGlycerol (YPEG)) medium, corresponding to a final A 600 ϭ 0.2, 0.04, 0.008, 0.0016, and 0.00032. Plates were incubated at 30°C, and growth was captured after 2 or 3 days. Results are representative of three experiments.

Coq10 knockout phenotypes are rescued by deletion of COQ11
Coq4, Coq7, and Coq9 compared to and plotted as a percentage of WT (Fig. 6B).
In contrast to coq10⌬, the coq11⌬ single mutant had elevated Coq4, Coq6, Coq7, and Coq9 (Fig. 6A), with protein quantification shown in Fig. 6B. Furthermore, the coq10⌬coq11⌬ double mutant also had raised amounts of Coq4, Coq7, and Coq9 polypeptides compared with WT (Fig. 6, A and B). This increase in Coq proteins could not be explained by enhanced COQ transcription as there was no corresponding change in the concentration of the respective mRNAs, (Fig. 6C), although COQ4 mRNA was not detected.
CoQ synthome formation was probed using two-dimensional blue native/SDS-PAGE (2D-BN/SDS-PAGE) with Coq4 and Coq9 serving as sensitive indicators of a high-molecularweight complex (13). As expected, the CoQ synthome in WT yeast presented as a heterogeneous high-molecular-weight complex, spanning a range of ϳ140 kDa to Ͼ1 MDa for Coq4 (Fig. 7A) and from ϳ100 kDa to Ͼ1 MDa for Coq9 (Fig. 7B). Consistent with prior results (13,19), the coq10⌬ mutant displayed a highly-destabilized CoQ synthome, with a disappearance of large complexes that were replaced by lower-molecularweight subcomplexes less than ϳ440 kDa for Coq4 (Fig. 7A) and less than ϳ232 kDa for Coq9 (Fig. 7B). The coq11⌬ mutant had a stabilized CoQ synthome compared with WT, with highmolecular-weight complexes shifting to the left and collapsing into a more homogeneous complex spanning ϳ900 kDa to Ͼ1 MDa for both Coq4 (Fig. 7A) and Coq9 (Fig. 7B). When COQ11 was deleted in combination with COQ10, there was a substantial rescue of high-molecular-weight complex formation compared with the coq10⌬ single mutant (Fig. 7, A and B). The CoQ

Coq10 knockout phenotypes are rescued by deletion of COQ11
synthome of the coq10⌬coq11⌬ double mutant appeared similar to that of coq11⌬ complexes spanning ϳ66 kDa to Ͼ 1 MDa for Coq4 (Fig. 7A) and from ϳ66 kDa to Ͼ 1 MDa for Coq9 (Fig.  7B). However, the deletion of COQ11 in coq10⌬ does not negate the effect from the COQ10 deletion, as it does not restore small subcomplexes Ͻ140 kDa to higher molecular weights (Fig. 7, A and B). These CoQ synthome signals for coq11 mutants were complementary to the observed increased Coq polypeptides, indicating that the absence of COQ11 enhanced the Q 6 biosynthetic machinery.

High-copy COQ8 does not restore Q 6 content in the coq10⌬coq11⌬ single mutant
The Coq8 polypeptide is a member of an ancient atypical kinase family (37), with several conserved kinase motifs that are essential for Q biosynthesis (13,38). Prior studies have demonstrated that overexpression of Coq8 in a coq10⌬ mutant increased the otherwise low amounts of several key Coq polypeptides and stabilized CoQ synthome formation (13). This is similar to the phenotype observed when COQ11 was deleted in the coq10⌬ mutant. Furthermore, Coq8 overexpression has also been shown to influence Q 6 biosynthesis, including the restoration of late-stage Q 6 -intermediates in coq5-coq9 null mutants (39). Although the CoQ synthome of the coq10⌬ mutant was stabilized by deletion of COQ11, Q 6 and late-stage Q 6 -intermediates remained lower compared with WT and the coq11⌬ single mutant (Fig. 4C). We hypothesized that the overexpression of Coq8 in the coq10⌬ and coq10⌬coq11⌬ mutant may restore Q 6 content in both mutants.
Each strain was grown in minimal selection medium to maintain plasmid expression and was analyzed for Q 6 biosynthesis following metabolic labeling with the ring-labeled Q 6 precursor, [ 13 C 6 ]HB (Fig. 8, B and C). Changing the growth medium from rich (i.e. YPGal) to minimal synthetic (i.e. SD and dropout dextrose media (DOD)) changed the relative amounts of Q 6 content among the mutants (Figs. 4B versus 8C). Although WT [ 13 C 6 ]Q 6 and total Q 6 content is similar in Figs. 4B and 8C, the values for coq11⌬ and the double mutant are quite different. When grown in YPGal, coq10⌬ had the lowest Q 6 content, followed by the double mutant coq10⌬coq11⌬, with coq11⌬ hav- Figure 5. Deletion of COQ11 in the coq10⌬ mutant enhances mitochondrial Q 6 content. Triplicates of 30-ml cultures of WT, coq10⌬, coq11⌬, and coq10⌬coq11⌬ yeast were grown in YPGal until they reached A 600 ϳ4. A, 5 ml of whole cells from each culture were harvested, lipid-extracted, and analyzed by LC-MS/MS for Q 6 content. Alternatively, WT, coq10⌬, coq11⌬, and coq10⌬coq11⌬ yeasts were grown in YPGal until they reached A 600 ϳ4 and were subjected to mitochondrial preparation. B, lipids from triplicates of purified mitochondria (100 g) were analyzed by LC-MS/MS for Q 6 content. C-E, mitochondrial mass was estimated using three distinct methods. C, relative mitochondrial DNA to actin was quantified by qPCR. D, porin protein amounts were quantified by hand using ImageStudioLite following immunoblot and normalized to total protein levels evaluated by Ponceau stain. E, citrate synthase activity was determined using a colorimetric assay as outlined under "Experimental procedures." Values are the mean of three replicates. The data show the mean Ϯ S.D., and the statistical significance as compared with WT is represented by *, p Ͻ 0.05; ***, p Ͻ 0.001; and ****, p Ͻ 0.0001.

Coq10 knockout phenotypes are rescued by deletion of COQ11
ing the closest Q 6 content to WT (Figs. 4, A and B, and 5A and Fig. S2, A and B). In contrast, when these strains are cultured in SDϪUra (Fig. 8, B and C), the double mutant coq10⌬coq11⌬ had the lowest Q 6 content, as compared with either the coq10⌬ or coq11⌬ single mutant strains. Growth on minimal dextrose medium in the absence of plasmid selection produced similar trends (Fig. S2, C and D, and Fig. S3, A and B).
Upon Coq8 overexpression, coq10⌬ had increased de novosynthesized [ 13 C 6 ]Q 6 and [ 12 C]Q 6 (Fig. 8B), and total Q 6 concentrations ([ 13 C 6 ]Q 6 ϩ [ 12 C]Q 6 ) were restored to those of WT (Fig. 8C). This finding is consistent with previous results, which also indicated that hcCOQ8 restored Q biosynthesis and amounts of Coq polypeptides and the CoQ synthome in the coq10⌬ mutant (13, 17, 18). Intriguingly, expression of hcCOQ8 had no effect on de novo or unlabeled Q 6 content in either the coq11⌬ single mutant or the coq10⌬coq11⌬ double mutant (Fig. 8B). Total Q 6 contents of both coq11⌬ and coq10⌬coq11⌬ remained significantly decreased compared with WT (Fig. 8C). Together, these results show that the rescue of the coq10⌬ mutant mediated by Coq8 overexpression requires Coq11.

Expression of low-copy COQ11 rescues only some of the phenotypes of the coq10⌬coq11⌬ mutant
The functional complementation of coq11⌬ single and coq10⌬coq11⌬ double mutants with low-copy COQ11 was assessed (lcCOQ11, Table 2). As expected, the coq10⌬ mutant, coq10⌬ with empty vector, and coq10⌬ complemented with lcCOQ11 showed slow growth on the nonfermentable carbon Figure 6. Several Coq polypeptides have increased abundance in coq11⌬ and coq10⌬coq11⌬ mutants compared with WT. A, aliquots of purified mitochondria (25 g) from WT, coq10⌬, coq11⌬, and coq10⌬coq11⌬ yeasts were subjected to 10 or 12% Tris-glycine SDS-PAGE. Mitochondrial malate dehydrogenase (Mdh1) was included as a loading control, with a representative blot shown. The Coq5 protein also serves as a qualitative loading control, because the Coq5 polypeptide amounts remain unchanged across the panel of coq1-coq4 and coq6 -coq10 deletion mutants (12). Aliquots of purified coq⌬ mitochondria (coq1⌬-coq11⌬) were included as negative controls for immunoblotting with antisera to each of the Coq polypeptides. Black arrows highlight the location of each protein on the membrane. B, ImageStudioLite was used to quantify triplicates of each Coq protein band's intensity by hand, which were normalized to Mdh1 and plotted as a percentage of WT. The data show mean Ϯ S.D., and the statistical significance is as compared with WT is represented by *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001; and ****, p Ͻ 0.0001. C, qPCR was used to determine COQ gene expression from whole-cell cultures of WT, coq10⌬, coq11⌬, or coq10⌬coq11⌬, and data were normalized to actin. COQ RNA levels remain unchanged in the coq10⌬, coq11⌬, or coq10⌬coq11⌬ mutants as compared with WT.

Coq10 knockout phenotypes are rescued by deletion of COQ11
source YPGlycerol (Fig. 9A). Because yeast lacking COQ11 retain respiratory capacity (Fig. 2) (14), coq11⌬ complemented with lcCOQ11 had no detectable change in growth phenotype compared with either the coq11⌬ mutant or coq11⌬ with empty vector (Fig. 9A). Intriguingly, when lcCOQ11 was expressed in the coq10⌬coq11⌬ double mutant, there was no repression of growth on YPG compared with that of the coq10⌬ mutant (Fig. 9A).
This observation suggests that Q 6 biosynthesis in coq10⌬ coq11⌬ may not be affected by lcCOQ11. To determine the effect of lcCOQ11 expression on mutant Q 6 biosynthesis, yeast was grown in selection medium to maintain plasmid expression. We tested whether lcCOQ11 expression rescued Q 6 content in the coq11⌬ mutant. Expression of lcCOQ11 in coq11⌬ efficiently rescued total Q 6 ([ 13 C 6 ]Q 6 ϩ [ 12 C]Q 6 ) to WT amounts (Fig. 9B). Aliquots (100 g) of purified mitochondria from WT, coq10⌬, coq11⌬, and coq10⌬coq11⌬ yeasts cultured in YPGal were solubilized with digitonin and separated with two-dimensional BN/SDS-PAGE. Following transfer of proteins to membranes, the CoQ synthome was visualized as a heterogeneous signal from ϳ66 to ϳ669 kDa in WT control with antibodies to A, Coq4, or B, Coq9. Intact mitochondria (25 g) from each designated strain was included as a loading control (M). Aliquots of coq4⌬-or coq9⌬-purified mitochondria (25 g) were included as a negative control for the antisera to Coq4 and Coq9, as the Coq9 polypeptide is absent from the coq4⌬ mutant. Red arrowheads and brackets indicate distinct complexes.

Coq10 knockout phenotypes are rescued by deletion of COQ11
Finally, whole-cell steady-state Q 6 concentrations were evaluated in all mutants. Even though lcCOQ11 complementation did not suppress coq10⌬coq11⌬ growth on YPGlycerol (Fig.  9A), Q 6 concentrations were increased in coq10⌬coq11⌬ to a level comparable with that of coq10⌬ (Fig. 9C). This implies that Coq11's role in Q 6 biosynthesis is not effective when the COQ11 ORF is expressed on a single-copy plasmid in the absence of COQ10. Perhaps Coq11 expression from a plasmid does not account for multiple levels of regulation that occur through endogenous expression. Alternatively, the coq10⌬coq11⌬ double mutant may have slightly lower amounts of the Coq11 polypeptide compared with coq11⌬ when both are complemented by lcCOQ11 (Fig. 9D), and these lower levels may not be sufficient to suppress respiration (Fig. 9A).

Discussion
This work investigated a putative functional relationship between Coq10 and Coq11 within the S. cerevisiae Q 6 biosynthetic pathway. The presence of Coq10 -Coq11 fusions in several Ustilaginaceae species suggests that these proteins may directly interact or participate in the same biological pathway in yeast (Fig. 1A) (14). Yeast Coq10 and its orthologs were previously shown to be required for efficient de novo Q biosynthesis and respiration (17,18). We were surprised to discover that the yeast coq10⌬ growth defect on nonfermentable medium and oxygen consumption rates were rescued upon deletion of COQ11 (Fig. 2). Moreover, spontaneous revertants isolated from coq10⌬ yeast were previously found to exhibit growth on Figure 8. Overexpression of the CoQ synthome stabilizer, COQ8, has no effect on Q 6 synthesis in the coq10⌬coq11⌬ mutant. WT, coq10⌬, coq11⌬, and coq10⌬coq11⌬ mutants were transformed with high-copy COQ8 (hcCOQ8) or empty vector (pRS426) plasmids. A, strains were grown overnight in 5 ml of selection medium, diluted to an A 600 ϭ 0.2 with sterile PBS, and 2 l of 5-fold serial dilutions were spotted onto YPD, YPG, or selection medium (SDϪUra), corresponding to a final A 600 ϭ 0.2, 0.04, 0.008, 0.0016, and 0.00032. Plates were incubated at 30°C, and growth was captured after 2 or 3 days. Triplicates of 5 ml of culture in selection medium were labeled with 5 g/ml [ 13 C 6 ]4HB, collected after 4 h, lipid-extracted, and analyzed by LC-MS/MS. B, [ 12 C]Q 6 (white) and de novo [ 13 C 6 ]Q 6 (blue). C, total amount of Q 6 was also plotted from the sum of [ 13 C 6 ]Q 6 and [ 12 C]Q 6 . The values are the means of three replicates. The data show mean Ϯ S.D., and the statistical significance as compared with WT is represented by *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001; and ****, p Ͻ 0.0001. The ns signifies that values are not significantly different from WT.

Coq10 knockout phenotypes are rescued by deletion of COQ11
nonfermentable medium (18). We have shown that this reversion is due to a dominant base pair deletion within the COQ11 gene, likely resulting in a nonfunctional, truncated Coq11 protein (Fig. 3). Mutants lacking both COQ10 and COQ11 when cultured on YPGal have increased de novo Q 6 production ( Fig.  4A) in addition to a 5-fold increase in mitochondrial Q 6 content compared with the coq10⌬ single knockout (Fig. 5B). Therefore, we have demonstrated that deletion of the Coq11 polypeptide in a coq10⌬ mutant confers a beneficial effect on both respiration and Q 6 biosynthesis (Fig. 10).
Enhanced Q 6 content in the coq10⌬coq11⌬ double mutant compared with the coq10⌬ single mutant may be partially due to increased amounts of several key Coq polypeptides (Fig. 6) and CoQ synthome stabilization (Fig. 7). The ring-modifying enzymes within the Q 6 biosynthetic pathway colocalize to numerous distinct "CoQ domains" in vivo, and proper assembly of the CoQ synthome components is required for the presence of these CoQ domains (23). Two recent studies demonstrated that mitochondria isolated from yeast lacking COQ10 have a reduced number of CoQ domain puncta (23,24). This is likely due to lower levels of certain Coq polypeptides and partial CoQ synthome destabilization in the coq10⌬ mutant (17,19), which was confirmed in this work (Figs. 6 and 7). In contrast, coq11⌬ yeast displayed significantly higher amounts of Coq4, Coq6, Coq7, and Coq9 polypeptides (Fig. 6). The CoQ synthome was likewise shifted to a higher molecular weight in coq11⌬ mitochondria compared with WT (Fig. 7). When Coq9 -yEGFP was used as a marker for CoQ domains, coq11⌬ had increased CoQ Figure 9. Low-copy COQ11 rescues only some of the phenotypes of the coq10⌬coq11⌬ double mutant. A low-copy plasmid expressing COQ11 (lcCOQ11) and an empty vector control (pRS316) were transformed into WT, coq10⌬, coq11⌬, and coq10⌬coq11⌬. A, strains were grown overnight in 5 ml of YPDextrose (YPD) and diluted to an A 600 ϭ 0.2 with sterile PBS, and 2 l of 5-fold serial dilutions were spotted onto YPDextrose, YPGlycerol (YPG), or selection medium (SDϪUra), corresponding to a final A 600 ϭ 0.2, 0.04, 0.008, 0.0016, and 0.00032. Plates were incubated at 30°C, and growth was captured after 3 or 4 days. B, rescue of mutant de novo and unlabeled Q 6 production was initially demonstrated in coq11⌬. Triplicates of 6-ml cultures in selection medium were labeled with 5 g/ml [ 13 C 6 ]4HB, and 5 ml of each culture was collected after 4 h, lipid-extracted, and analyzed by LC-MS/MS. Total amount of Q 6 was plotted from the sum of de novo [ 13 C 6 ]Q 6 (blue) and unlabeled [ 12 C]Q 6 . Values are the mean of three replicates. C, rescue of mutant Q 6 content was evaluated in each mutant strain. Triplicates of 6-ml cultures in selection medium were grown until A 600 ϳ4. Lipid extracts from 5 ml of each culture were analyzed by LC-MS/MS. The data show the means Ϯ S.D., and the statistical significance as compared with WT is represented by *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001; and ****, p Ͻ 0.0001. ns signifies that values are not significantly different from WT. D, aliquots of purified mitochondria (25 g) from WT and mutant yeast containing empty vector or lcCOQ11 were isolated in YPGal medium and were separated on 10% Tris-glycine SDS-polyacrylamide gels to determine Coq11 protein expression. Proteins stained with Ponceau stain were used as loading control.

Coq10 knockout phenotypes are rescued by deletion of COQ11
domain intensity that stemmed from amplified expression of Coq9 -yEGFP, although the number of domains was similar to WT. Mutants lacking essential Coq polypeptides coq1-coq9, or coq10, displayed significantly less Coq9-labeled domains (24). The CoQ synthome stabilization seen in coq11⌬ via 2D-BN/ SDS-PAGE analyses performed in this work (Fig. 7) agrees with the observation of increased CoQ domains and argues that the CoQ synthome is truly stabilized upon deletion of COQ11, as opposed to inducing a greater number of domains.
These observations are consistent with the biochemical data that led to the notion of a CoQ synthome whose formation relies on the presence of prenylated Q-intermediates (13,40,41). The coq10⌬ mutant produced more early-stage intermediates (HHB) and had less late-stage intermediates (DMQ 6 ) compared with WT (Fig. 4, C and D), resulting in less CoQ syn-thome formation (Fig. 7). Because coq11⌬ yeast displayed comparable amounts of early-and late-stage Q 6 -intermediates to WT (Fig. 4, C and D), this mutant retained the ability to fully form the CoQ synthome (Fig. 7). The double knockout synthesized varying amounts of early-and late-stage Q 6 -intermediates that were largely in-between those of the single knockouts (Fig. 4, C and D). The CoQ synthome is thus able to form in coq10⌬coq11⌬ yeast, albeit not to the efficiency of the coq11⌬ single mutant (Fig. 7). We suspect that the accumulation of Coq polypeptides and restoration of the CoQ synthome in the coq10⌬coq11⌬ double mutant are sufficient to allow for Q 6 to escape its site of synthesis and reach the respiratory complexes, despite an absence of the Coq10 Q 6 chaperone protein and lower Q 6 in this strain (Fig. 10). How this occurs is presently not known. One possible explanation could be that Coq11 inhibits Under homeostasis, Coq11 associates with the CoQ synthome and acts as a modest negative regulator of Q 6 synthesis via Coq10. In the absence of COQ11, several Coq polypeptides are increased (dark shading), and the CoQ synthome is stabilized compared with WT cells, despite a slight decrease in Q 6 content. In contrast, the coq10⌬ mutant is missing the Q 6 chaperone protein, resulting in a decreased amount (light shading) of Coq3, Coq4, Coq7, and Coq9, a destabilized CoQ synthome, substantially decreased Q 6 concentrations, and a lack of respiration. The deletion of COQ11 in the coq10⌬ mutant counterbalances the destabilized CoQ synthome and decreased Q 6 content phenotype of the coq10⌬ mutant, allowing the coq10⌬coq11⌬ double mutant to grow on YPG and respire. Expression of hcCOQ8 in the coq10⌬ mutant produces many similar phenotypes to COQ11 deletion in coq10⌬ cells, including increased Coq polypeptides and a stabilized CoQ synthome (13), resulting in restored growth on YPG. Unlike the coq10⌬coq11⌬ double mutant, coq10⌬ ϩ hcCOQ8 has rescued Q 6 content pointing to an additional role of Coq11 in Q 6 biosynthesis, redox regulation, or transportation.

Coq10 knockout phenotypes are rescued by deletion of COQ11
a currently unidentified Q 6 chaperone with lower efficiency than Coq10 that is able to rescue respiration only in the absence of both Coq10 and Coq11.
Overexpression of the Q 6 -biosynthetic protein Coq8 also rescued coq10⌬ mutant growth on nonfermentable medium (Fig. 8A) and de novo Q 6 biosynthesis (Fig. 8, B and C) (17). Coq8 has been implicated in the partial extraction of Q 6 -intermediates out of the mitochondrial inner membrane for enzymatic modification by other Coq proteins, allowing for appropriate Q 6 biosynthesis (25). Prior investigations also revealed that Coq8 overexpression in coq10⌬ yeast increased Coq4 and Coq9 polypeptides and stabilized the CoQ synthome (13,24). Despite the substantial benefit of Coq8 overexpression in a coq10⌬ single mutant, Coq8 overexpression failed to enhance Q 6 biosynthesis in the coq10⌬coq11⌬ double knockout (Figs. 8,  B and C, and 10). The absence of COQ11 in the double mutant is sufficient to restore Coq polypeptides and CoQ synthome formation (Figs. 6, A and B, 7, and 10). Coq11 and Coq8 may therefore work by different mechanisms to serve opposing functions for CoQ synthome and Coq polypeptide stabilization. Furthermore, it is clear that Coq11 is required to perform an additional function to Coq8, as hcCOQ8 requires the presence of Coq11 to restore Q 6 biosynthesis (Figs. 8, B and C, and 10).
Proper CoQ synthome formation is not only required for efficient Q 6 biosynthesis, but it is also vital for the establishment of ER-mitochondrial contact sites mediated by the ER-mitochondrial encounter structure (ERMES) complex (23,24). The ERMES complex is essential for lipid exchange between the ER and mitochondria (42). Specifically, ERMES null mutants have irregular Q 6 cellular distribution and a destabilized CoQ synthome (23). When COQ10 was deleted in yeast expressing Coq6 -GFP, there was a significant decrease in Coq6 -GFP puncta colocalization with Mdm34 -mCherry, a component of the ERMES complex (23). These results indicate that CoQ synthome positioning next to the ERMES complex, and subsequent Q 6 distribution from the mitochondria, depends on Coq10. Because COQ11 deletion stabilizes the CoQ synthome (Figs. 6, A and B, and 7), it is possible that coq11⌬ mutants have more ER-mitochondrial contacts through the ERMES complex and improved transfer of lipids between these organelles. Therefore, one possible role of Coq11 may be an auxiliary protein mediating lipid transport between the ER and mitochondria.
In a recent study, Coq11 was named Mrx2, as part of the mitochondrial organization of gene expression (MIOREX) complex involved in the mitochondrial genetic expression system (43). Considering the proposed regulatory function of Coq11 in CoQ synthome assembly, it is tempting to speculate that Coq11 offers a mechanism to couple Q 6 synthesis with the assembly of the respiratory complexes. When the synthesis of the respiratory complexes is more active, Coq11 is associated to the MIOREX complex, and Q 6 synthesis at ER junctions is stimulated. Coq11 dual localization in the mitochondrial inner membrane, to the MIOREX complex and the CoQ synthome, would also explain the sensitivity of yeast cells to the number of COQ11 copies, as well as present another example of a loop system control for balanced expression of mitochondrial products (44).
Another potential explanation for the phenotypes induced by the knockout of Coq11 relates to its structural connection with the short-chain dehydrogenase/reductase (SDR) superfamily of NAD(P)(H)-dependent oxidoreductases (14,45). These enzymes catalyze an assortment of reactions, including isomerization, decarboxylation, epimerization, imine reduction, and carbonyl-alcohol oxidoreduction (45). SDR superfamily proteins contain a conserved protein structural motif known as a Rossmann fold, a feature used in the binding of nucleotide cofactors such as FAD, FMN, and NAD(P) (46). The crystal structure of the Pseudomonas aeruginosa gene UbiX, which catalyzes the decarboxylation step in Q 9 biosynthesis, revealed a Rossmann fold with a bound FMN (47). Thus, Coq11 may use its Rossmann fold in conjunction with a nucleotide cofactor to perform similar redox chemistry in S. cerevisiae Q 6 biosynthesis. The ratio of QH 2 /Q serves as a metabolic sensor for electron transport chain efficiency (48). High QH 2 /Q ratios induce respiratory complex I-mediated reverse electron transport (RET) under physiological conditions in both Drosophila and mammalian cell lines (48,49). Superoxide and secondary reactive species produced specifically through complex I RET extended Drosophila lifespan and improved mitochondrial function in a model of Parkinson's disease (49). RET induced by over-reduction of the Q pool presumably generates a superoxide-dependent signal essential for homeostasis, such that manipulation of the Q redox state is beneficial for mitochondrial function (48,49). Mitochondrial phenotypes in the absence of COQ11, including restored respiration in coq10⌬ coq11⌬ and up-regulated Q 6 machinery (Figs. 2, A-C, 6, and 7), seem to correlate well with the aforementioned effects of Q 6 H 2 accumulation. Yeast coq11⌬ and coq10⌬coq11⌬ mutants retain antioxidant protection by Q 6 H 2 , demonstrated by their resistance to treatment with exogenously-added PUFAs (Fig. 2D). Because cells lacking the Coq11 polypeptide maintain Q 6 H 2 as an antioxidant, it follows that Coq11 could be involved in the oxidation of Q 6 H 2 to Q 6 .
The phenotypes of Coq10 and Coq11 seen in this work are similar to those in both fungi and mammalian hosts. Several fungi use Coq11 or Coq11-like proteins as NAD-dependent epimerases/dehydratases, NADH-ubiquinone oxidoreductases, and NADH dehydrogenase subunits (14). Coq11 orthologs are commonly found in plant and algae genomes, including the chloroplast-localized flavin reductase protein At1g32220 from the land plant Arabidopsis thaliana, which is thought to be involved in plastoquinone biosynthesis and storage (14,50). The closest but distinct higher eukaryotic Coq11-like protein is the SDR subfamily protein NDUFA9 (14), an auxiliary subunit of complex I in humans (51). Patients with decreased NDUFA9 expression are unable to properly assemble complex I and may develop a degenerative infancy respiratory disorder known as Leigh syndrome (52). Although yeast cells do not possess complex I, this evidence indicates that Coq11 may play a crucial role in respiratory regulation or function, supporting the observations of this study.
The function of Coq10 is widely conserved across different organisms. Expression of the Coq10 homolog from C. crescentus (CC1736) rescues the impaired respiration and antioxidant function of Q 6 in coq10 yeast mutants (17). The NMR structure

Coq10 knockout phenotypes are rescued by deletion of COQ11
of CC1736 reveals a START domain, which is known to bind lipids via a hydrophobic tunnel (20). Studies of S. pombe Coq10 demonstrate that it is able to bind Q 10 (21). One proposed function of Q binding by CC1736 and Coq10 from S. pombe may be to regulate Q delivery to its proper sites in the respiratory complexes. Humans have two distinct homologs of yeast Coq10: COQ10A and COQ10B. Expression of either human protein rescues the coq10⌬ respiratory deficiency and sensitivity to oxidative stress, and it restores the amounts of Coq polypeptides to WT (19). The conserved function of yeast Coq10 with human COQ10A and COQ10B suggests that the findings of this work will shed light on the role of Coq10 as a chaperone in humans, leading to a better understanding of the pathobiology of Q 10 diseases.
In summary, this work reveals that Coq11 plays a regulatory role to maintain Q 6 homeostasis in concert with Coq10 in S. cerevisiae (Fig. 10). The absence of COQ11 caused an augmentation of Q 6 production and respiration in the coq10⌬ mutant, indicating that Coq11 confers a negative effect on the CoQ synthome. Coq11 may be crucial for Q 6 function in addition to Q 6 biosynthesis, as total whole-cell and mitochondrial Q 6 content remained lower than WT in coq11⌬ and coq10⌬coq11⌬ mutants.

Experimental procedures
All reagents were obtained commercially from Thermo Fisher Scientific unless otherwise specified
COQ11 was disrupted by the one-step gene replacement method (56). The LEU2 gene from pRS305 was amplified by polymerase chain reaction (PCR), with COQ11 upstream and downstream flanking sequences 5Ј-GGGAAATATGTATCG-TATACAAAAATACAGCTAAAGCTTGAACTG and 3Ј-GTACTTAACTATATACAGCTTGGTATAATTTTAAAA-TGGTAATAAC. Transformations of PCR products into yeast cells were performed using the Li-acetate method (57). The double coq10⌬coq11⌬ mutant was constructed via disruption of COQ10 within the coq11⌬ strain. The HIS3 gene from pRS313 was amplified by PCR with COQ10 upstream and downstream flanking sequences 5Ј-GGATAAGGAGCCAAA-CAATAAACGGCTAAAGATACCGTGG and 3Ј-CAGATA-ACAAAGATCATGCCATCCAGGATAAGCGTATGCA, and transformation was performed as for the COQ11 disruption.

Mitochondria isolation from BY4742 WT and mutant yeast
Yeast cultures of BY4742, coq10⌬, coq11⌬, and coq10⌬ coq11⌬ were grown overnight in 5 ml of YPD. Yeast-containing plasmids were grown overnight in 5 ml of selection medium (SDϪUra). All pre-cultures were back-diluted with YPGal and grown overnight with shaking (30°C, 250 rpm) until cell density reached an A 600 ϳ4. Spheroplasts were prepared with Zymolyase-20T (MP Biomedicals) and fractionated as described previously (36), in the presence of cOmplete TM EDTA-free protease inhibitor mixture tablets (Roche Applied Science), phosphatase inhibitor mixture set I (Sigma-Aldrich), phosphatase inhibitor mixture set II (Sigma-Aldrich), and phenylmethylsulfonyl fluoride (Thermo Fisher Scientific). Nycodenz (Sigma-Aldrich) density gradient purified mitochondria were frozen in liquid nitrogen, aliquoted, and stored at Ϫ80°C until further use. Protein concentration of mitochondria was measured by the bicinchoninic acid (BCA) assay (Thermo Fisher Scientific).

Submitochondrial localization of Coq10 and Coq11 polypeptides
Purified mitochondria from BY4742 yeast (3 mg of protein, 150 l) were subfractionated, as described previously (13). Proteinase K treatment of purified BY4742 mitochondria was also performed as described previously (13). Proteinase K-treated mitoplasts and control samples were resuspended in SDS sample buffer (50 mM Tris, pH 6.8, 10% glycerol, 2% SDS, 0.1% bromphenol blue, and 1.33% ␤-mercaptoethanol); equal aliquots were separated by SDS-gel electrophoresis on 10 or 12% Tris-glycine polyacrylamide gels as detailed below. Several mitochondrial compartment markers and proteins of interest, Coq10 and Coq11, were detected with rabbit polyclonal antibodies prepared in blocking buffer at dilutions listed in Table S1.

Oxygen consumption evaluation by Seahorse
Mitochondrial function was assessed using the XF96 extracellular flux analyzer (Seahorse Bioscience, Agilent Technologies). Seahorse plates were coated with 50 g/ml poly-D-lysine (Sigma-Aldrich), diluted 1:1 in UltraPure distilled water (Thermo Fisher Scientific). Volumes of 25 l were added to each well for 30 min at room temperature and then aspirated before plates were dried overnight at room temperature. The Seahorse XF96 sensor cartridge was hydrated with Seahorse XF calibrant solution (Agilent) and was incubated overnight at room temperature.
Yeast cultures of BY4742, coq10⌬, coq11⌬, and coq10⌬coq11⌬ were grown overnight in 25 ml of YPGal medium. On the day of measurement, all cultures were diluted to seed an A 600 ϭ 0.1 cells/well of BY4742, coq10⌬, coq11⌬, and coq10⌬coq11⌬ into a Seahorse XF96 microplate in a total volume of 175 l YPGal. Four wells containing only medium were included for background measurement. The loaded plate was centrifuged at 500 ϫ g for 3 min at room temperature (with no brakes). Following centrifugation, the loaded plate was incubated for 30

Coq10 knockout phenotypes are rescued by deletion of COQ11
min at 37°C with no CO 2 to aid in the transitioning of the plate into the Seahorse machine's temperature. Cells were stimulated sequentially with two injections of 4 M FCCP in ports A and B (optimized for maximum oxygen consumption rate) (Enzo Life Sciences) and 2.5 M antimycin A in port C (Enzo Life Sciences), delivered in YPGal. Mix, wait, and measure times were 2 min, 30 s, and 2 min, respectively. Basal respiration included four measurements, and then following each injection three measurements were made. All OCR were subtracted for nonmitochondrial respiration and normalized to A 600 ϭ 0.1. Basal respiration was calculated as an average of OCR prior to the first FCCP addition. Maximal respiration was calculated as an average of OCR following the second FCCP addition. Nonmitochondrial respiration was measured as average OCR following antimycin A addition.

Fatty acid sensitivity assay
Sensitivity of yeast cells to PUFA-induced oxidative stress was performed as described previously (19,58,59), with some modifications. Briefly, BY4742 WT, cor1⌬, coq9⌬, coq10⌬, coq11⌬, and coq10⌬coq11⌬ were inoculated in 5 ml of YPD medium and incubated overnight at 30°C, 250 rpm. Cultures were subinoculated to an A 600 ϭ 0.25 in 15 ml of fresh YPD medium and incubated at 30°C, 250 rpm until they reached an A 600 ϳ1. Cells were harvested, washed twice with 10 ml of sterile H 2 O, and diluted in 0.1 M phosphate buffer with 0.2% dextrose, pH 6.2, to an A 600 ϭ 0.2. This cell suspension was divided into 5-ml aliquots and treated with an ethanol vehicle control (final concentration 0.1% v/v), ethanol-diluted oleic acid (Nu-Check Prep), or ␣-linolenic acid (Nu-Check Prep) to a final concentration of 200 M. Fatty acid-treated cultures were incubated for 4 h at 30°C, 250 rpm, after which cell viability was assessed via plate dilutions. Cell viability prior to the addition of fatty acids was determined via plate dilutions, represented in the 0-h plate.

Analysis of Q 6 and Q 6 -intermediates
Standards of Q 6 were obtained from Avanti Polar Lipids, and Q 4 was from Sigma-Aldrich. Yeast cultures were grown overnight in 30 ml of YPGal, or selection medium (SD-complete or SDϪUra) for strains harboring plasmids. Cultures were diluted into triplicates of 6 ml of fresh medium to A 600 ϭ 0.5, and 5 ml of medium was harvested by centrifugation once they reached A 600 ϳ4. Cell pellets were stored at Ϫ20°C. Following collection, frozen cell pellets were lipid-extracted in the presence of internal standard Q 4 and analyzed for Q 6 and Q 6 -intermediates by LC-MS/MS as described previously (19).

Stable isotope labeling for determination of de novo Q 6 and Q 6 -intermediates
Yeast cultures were grown overnight in 30 ml of YPGal and diluted in triplicates of 6 ml of fresh medium to an A 600 ϭ 0.1. Cultures were incubated until they reached an A 600 ϳ1, at which point ethanol vehicle control (0.1% v/v) or 5 g/ml of the stable isotope [ 13 C 6 ]4HB (Cambridge Isotope Laboratories, Inc.) was added. Cultures were allowed to grow for an additional 4 h when 5 ml of each culture was harvested by centrifu-gation and stored at Ϫ20°C. Cell pellets were lipid extracted and analyzed by LC-MS/MS as described previously (19).

Mitochondrial DNA determination by qPCR analysis
DNA was extracted from yeast cells as follows. Yeast pellets (10 ml) grown in YPGal were collected at an A 600 ϳ4 by centrifugation at 3000 ϫ g for 5 min, washed with 5 ml of H 2 O, and transferred to 2-ml screw-cap tubes. Pellets were frozen at Ϫ80°C until DNA extraction was carried out. Cell pellets were resuspended in 200 l of lysis buffer (10 mM Tris-Cl, pH 8.0, 2% (v/v) Triton X-100, 1 mM EDTA, 100 mM NaCl, 1% SDS), and 200 l of acid-washed glass beads with 200 l of phenol/chloroform/isoamyl alcohol (25:24:1) were added to the cell suspension. Cells were lysed using a bead-beater (Precellys 24; Bertin Technologies) three times for 10 s at 6500 rpm with a 45-s break between rounds at 4°C. Tris-EDTA (TE, 200 l) was added, and the cell suspension was centrifuged at 13,000 ϫ g for 5 min at room temperature. The aqueous layer was removed to a new tube containing 200 l of chloroform, mixed by inversion, and centrifuged at 13,000 ϫ g for 5 min at room temperature. This was repeated once more. The aqueous layer was then transferred to a 2-ml screw-cap tube containing 1 ml of 95% EtOH, mixed by inversion, and centrifuged at 13,000 ϫ g for 2 min at room temperature. The resulting pellet was resuspended in 400 l of TE containing 30 g of RNase A and incubated at 37°C for 30 min. Then, 10 l of 3 M sodium acetate and 1 ml of 95% EtOH was added, mixed by inversion, and incubated at Ϫ20°C for 1 h. After 1 h at Ϫ20°C, the suspension was centrifuged at 13,000 ϫ g for 5 min, and the pellet was washed twice with 70% EtOH and air-dried. The dried pellet was resuspended in 25 l of TE; concentration was measured by Nanodrop (Thermo Fisher Scientific), and the pellet was stored at Ϫ20°C until use.
qPCR was performed on a CFX384 instrument (Bio-Rad) using the SensiFAST TM SYBR NO ROX kit (Bioline) as per the manufacturer's instructions. Each sample was run in duplicate with 150 ng of total DNA used per reaction using the following thermocycling protocol (95°C for 2 min, 95°C for 5 s, 60°C for 10 s, and 72°C for 20 s, plate read and cycle repeated ϫ40, melt curve 40 -92°C with plate read and 40°C for 10 s). Meltingcurve analysis confirmed that all PCRs produced a single product. mtDNA-specific primers (forward (13,999), 5Ј-GTG CGT ATA TTT CGT TGA TGC GT-3Ј; reverse (14,297), 5Ј-TTC ACA CTG CCT GTG CTA TCT AA-3Ј (60) and actin-specific primers (forward, 5Ј-GAA TTG AGA GTT GCC CCA GA-3Ј; reverse, 5Ј-ATC ACC GGA ATC CAA AAC AA-3) were used. The relative level of gene expression of mitochondrial DNA was normalized to the expression level of actin as described previously (61).

Citrate synthase activity
The measurement of citrate synthase activity in cells was carried out as described previously (62). Briefly, yeast pellets (10 ml) grown in YPGal were collected at an A 600 ϳ4 by centrifugation at 3000 ϫ g for 5 min, washed with 5 ml of H 2 O, and transferred to 2-ml screw-cap tubes. Cell pellets were resuspended in 200 l of lysis buffer (100 mM Tris-Cl, pH 7.4, 1% (v/v) Triton X-100, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 1ϫ cOmplete TM Protease Inhibitor Mixture (Roche

Coq10 knockout phenotypes are rescued by deletion of COQ11
Applied Science)), and then 200 l of acid-washed glass beads were added. Cells were lysed using a bead-beater (Precellys 24; Bertin Technologies) three times for 10 s at 6500 rpm with a 45-s break between rounds at 4°C. The clarified cell lysate was collected after centrifugation at 16,000 ϫ g for 10 min at 4°C. The concentration of protein was determined with the BCA assay (Thermo Fisher Scientific). Cell lysates were normalized to 0.05 g/l protein. The colorimetric citrate synthase assay was carried out using a VersaMax plate reader (Molecular Devices) and a flat-bottom 96-well plate. First, 40 l of 500 mM Tris-Cl, pH 7.4, 2 l of 30 mM acetyl-CoA, 8 l of 2.5 mM 5,5Ј-dithiobis(2-nitrobenzoic acid), 90 l of H 2 O, and 50 l cell lysate (2.5 g total protein) were added into each well. Then, 10 l of 10 mM oxaloacetic acid were added per well and mixed by pipetting up and down. A 412 was measured every 30 s at 25°C. The initial slope was calculated by using data from the first 10 min and used to determine the enzyme reaction rate using the extinction coefficient for 2-nitro-5-thiobenzoate of 14.15 mM Ϫ1 cm Ϫ1 (63).

Porin quantification
Porin content was quantified via immunoblot of yeast WT and mutant whole cells. Protein extraction from whole cells was performed (64), and 25 g of each sample was separated by SDS-gel electrophoresis as described below. Three replicates of the immunoblots were performed and quantified by hand using ImageStudioLite software normalized to Ponceau total protein staining.

Quantitative real-time PCR (qRT-PCR)
Total RNA was isolated from yeast cells using TRIzol reagent (Invitrogen). DNA contamination from the resulting RNA was removed using the DNase TURBO kit as per the manufacturer's instructions (Invitrogen). RNA concentration was measured by Nanodrop (Thermo Fisher Scientific), and RNA was stored at Ϫ20°C. Reverse transcription was carried out using the Superscript III first strand synthesis system using random hexamer primers (Invitrogen). cDNA was stored at Ϫ20°C until qPCR analyses were carried out. Quantitative real-time PCR was performed on a CFX384 instrument (Bio-Rad) using the Sensi-FAST TM SYBR NO-ROX kit (Bioline) in duplicate. The relative levels of gene expression were normalized to the expression level of actin. Melting curve analysis confirmed that all PCRs produced a single product. The primers (forward/reverse) used in real-time PCR were designed using Primer3 on line (RRID: SCR_003139). Primers used are given in Table S2.

SDS-PAGE and immunoblot analysis
Purified mitochondria (25 g) were resuspended in SDS sample buffer and separated by SDS-gel electrophoresis on 10 or 12% Tris-glycine polyacrylamide gels. Proteins were transferred to a 0.45-m polyvinylidene difluoride membrane (Millipore) and blocked with blocking buffer (0.5% BSA, 0.1% Tween 20, 0.02% SDS in PBS). Representative Coq polypeptides and loading control mitochondrial malate dehydrogenase (Mdh1) were probed with rabbit polyclonal antibodies prepared in blocking buffer at dilutions listed in Table S1. IRDye 680LT goat anti-rabbit IgG secondary antibody (LiCOR) was used at a dilution of 1:10,000. Proteins were visualized using a LiCOR Odyssey IR Scanner (LiCOR). Immunoblots are representative of three replicates and were quantified by hand using ImageStudioLite software normalized to Mdh1.

Two-dimensional Blue Native/SDS-PAGE immunoblot analysis of high-molecular-weight complexes
2D-BN/SDS-PAGE was performed as described previously (13,65,66). Briefly, 200 g of purified mitochondria were solubilized at 4 mg/ml for 1 h on ice with 16 mg/ml digitonin (Biosynth) in the presence of the protease and phosphatase inhibitors used during mitochondrial isolation. Protein concentration of solubilized mitochondria was determined by BCA assay. NativePAGE 5% G-250 sample additive (Thermo Fisher Scientific) was added to a final concentration of 0.25%. Solubilized mitochondria (100 g) were separated on NativePAGE 4 -16% BisTris gels (Thermo Fisher Scientific) in the first dimension, and native gel slices were further separated on 12% Tris-glycine polyacrylamide gel in the second dimension. Following the second-dimension separation, immunoblot analyses were performed as described above, using antibodies against Coq4 and Coq9 at the dilutions indicated in Table S1. Molecular weight standards for BN gel electrophoresis and SDS gel electrophoresis were obtained from GE Healthcare (Sigma-Aldrich) and Bio-Rad, respectively.

Construction of low-copy COQ11 yeast expression vectors
Plasmids used in this study are described in Table 2. A lowcopy COQ11-containing plasmid was constructed using the pRS316 low-copy empty vector. The COQ11 ORF and regions corresponding to 842 bp upstream and 256 bp downstream were cloned into pRS316 using Gibson Assembly (New England Biolabs). Clones were sequenced by Laragen, and successful clones were transformed into WT and mutant yeast, along with the corresponding empty vector (pRS316) control as described above.

Revertant isolation
As reported previously, coq10 mutant growth deficiency on nonfermentable carbon sources (YPEG) spontaneously revert due to nuclear suppression mutations (18). Here, W303 coq10⌬ yeast was grown on glucose to stationary phase, and ϳ10 million cells were plated on YPEG. After several weeks, a colony began to appear on this medium. The colony was purified, and its genome was sequenced.

Genome sequencing
The Wizard genomic purification kit (Promega) was used to extract total DNA from the parental respiratory-deficient mutant W303 coq10⌬ and from the spontaneous revertant W303 coq10rev. The DNA was quantified using the QUBIT DNA TM high-sensitivity assay, and 1 ng of the normalized DNA was tagged by the Nextera XTTM (Illumina) protocol. The libraries were amplified and pooled as described (67). The pooled libraries were subjected to sequencing with the Next-Seq TM (Illumina) equipment in the Genome Investigation and Analysis Laboratory of the Institute of Biomedical Sciences at the University of Sao Paulo. The BWA Aligner tool, Coq10 knockout phenotypes are rescued by deletion of COQ11 version 1.1.4 (Base Space Labs-Illumina), was used to align ϳ23,000,000 reads obtained from each strain with the reference genomes of S. cerevisiae. The alignments were compared using the Integrative Genomics Viewer (Base Space Labs Illumina).

Statistical analyses
All data sets were tested for normality using GraphPad Prism 7 with the Shapiro-Wilk normality test. Because a majority of sets passed the normality test (␣ ϭ 0.5), statistical analyses were performed using GraphPad Prism 7 with parametric one-way analysis of variance correcting for multiple comparisons using Tukey's test, comparing the mean of each sample to the mean of its corresponding WT or empty vector control. The data show the means Ϯ S.D., and the statistical significance as compared with WT or empty vector control is represented by the following: *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001; and ****, p Ͻ 0.0001. The denotation ns indicates values with "not significant" differences from the corresponding control.

Data availability
The MS source data for determination of Q 6 and Q 6 intermediates will be shared upon request. Please contact Catherine Clarke at cathy@chem.ucla.edu. All remaining data are contained within the article.