Saccharomyces cerevisiae strains used in this study are listed in Table 1. Growth media used in this study included: YPD (2% glucose, 1% yeast extract, 2% peptone), YPGal (1% yeast extract, 2% peptone, 2% galactose, 0.1% dextrose), and YPG (3% glycerol, 1% yeast extract, 2% peptone), and were prepared as described (Burke et al, 2000). Synthetic Dextrose (SD)/Minimal medium consisted of 0.18% yeast nitrogen base without amino acids, 2% dextrose, 0.14% NaH₂PO₄, 0.5% (NH₄)₂SO₄, and amino acids added to final concentrations as described (Barkovich et al., 1997). Selective SD/Minimal medium lacking uracil (SD–Ura) and selective SD/Minimal medium lacking uracil and leucine (SD–Ura–Leu) were similarly prepared. Agar plate media were prepared as described above and included 2% Bacto agar (Fisher).

COQ8 was over-expressed in yeast with the p4HN4 plasmid (mcCOQ8). The COQ8 gene was cloned into pRS426, a multi-copy yeast shuttle vector, resulting in mcCOQ8 (Hsieh et al., 2004). To construct plasmids expressing human COQ9, we cloned human COQ9 into pQM (Hsu et al., 1996) and pRCM (Allan et al., 2013). These are respectively low- and multi-copy vectors that express ORFs fused to the yeast Coq3 amino terminal mitochondrial leader sequence (amino acids 1–34) under control of the yeast CYC1 promoter. Human COQ9 was amplified from pBGcoq9, which contains the human COQ9 ORF in YEpJB1-21-10 and is expressed from a constitutive PGK promoter (Duncan et al., 2009). The human COQ9 ORF was amplified with Taq polymerase and primers Hcoq9F (5′-ATCGATATGGCGGCGG CGGCGGTAT-3′ with a ClaI restriction site at the 5′ end) and HcoqR (5′-GGTACCTC ACCGACGCTGGTTTAGACCTGTCAAGTTCTTGAGC-3′ with a KpnI restriction site at the 5′ end). PCR products were inserted into the TOPO vector resulting in a plasmid named HCOQ9TOPO. HCOQ9TOPO was digested with the restriction enzymes ClaI and KpnI (New England BioLabs) and inserted in pQM or pRCM prepared with ClaI and KpnI, resulting in the plasmids scHCOQ9 and mcHCOQ9, respectively. The nucleotide sequence of the human COQ9 ORF in scHCOQ9 and mcHCOQ9 was confirmed by sequencing (UCLA sequencing core, Los Angeles).

A PCR product containing the KanMX4 gene was amplified with the genomic DNA isolated from BY4741Δcoq9 (used as template) with primers that annealed 100 bp upstream and downstream of the COQ9 ORF. The sequences of the primers utilized were: 5′-TTTGGGCCTACATAAGGTACTTC-3′ and 5′-CGCACAGACCAATAAATCTGCC-3′. The PCR product was then transformed into the yeast W3031B to create W303Δcoq9K. Transformants that grew on YPD + 200 μg/ml G418 (Geneticin) were selected. Proteins were extracted from these transformants as described (Yaffe and Schatz, 1984) and separated by SDS-PAGE with a 10% polyacrylamide gel. Proteins were transferred to an Immobilon-P transfer membrane (Millipore) and analyzed by immunoblotting as described (He et al., 2014). The primary antibody against Coq9 was used at a 1:1000 dilution and the secondary antibody, goat anti-rabbit IgG H&L chain specific peroxidase conjugate (Calbiochem), at a 1:10,000 dilution. The absence of Coq9 polypeptide confirmed that COQ9 was replaced with KanMX4.

The de novo synthesis of Q₆ and Q₆-intermediates was tracked in yeast cells labeled with ¹³C₆-pABA or ¹³C₆-4HB followed by lipid analysis. Labeling media were prepared with 10 μg/ml ¹³C₆-pABA or ¹³C₆-4HB dissolved in ethanol (0.2% final concentration). After 12.5 h of labeling cells were collected (a total of 50 A_{600 nm}) as pellets by centrifugation. Q₄ was added (164 pmol) to each cell pellet to serve as an internal standard. Lipid extracts were analyzed by RP-HPLC-MS/MS (Xie et al., 2012). For liquid chromatography, a phenyl-hexyl column (Luna 5u, 100 × 4.60 mm, 5-μm, Phenomenex) was used. The mobile phase consisted of Solvent A (methanol/isopropanol, 95:5, with 2.5 mM ammonium formate) and Solvent B (isopropanol, 2.5 mM ammonium formate). Solvent B was increased linearly from 0 to 5% with the flow rate increased from 600 to 800 μl/min from 0 to 6 min. The flow rate and mobile phase were changed back to 600 μl/min and 100% Solvent A respectively at 7 min. Multiple reaction monitoring mode (MRM) analysis was performed with the 4000 QTRAP linear MS/MS spectrometer from Applied Biosystems (Foster City, CA). Data were processed with Analyst version 1.4.2 software (Applied Biosystems).

To quantify Q₆ content, the peak areas of ¹²C-Q₆ (sum of oxidized and reduced) and ¹³C₆-Q₆ (sum of oxidized and reduced) were normalized by the peak areas of Q₄ (sum of oxidized and reduced); the pmol amounts were then determined from the Q₆ standard curve. The pmol of ¹²C-Q₆ and ¹³C₆-Q₆ were further normalized by the yeast pellet wet weights. A chemical standard for DMQ₆ is not available. To quantify this intermediate, the peak areas (sum of oxidized and reduced DMQ₆) were normalized by the recovery of Q₄ (sum of oxidized and reduced peaks). Finally, calculated values were further normalized by the yeast pellet wet weights.

Yeast cultures were grown to 3–4 A_{600 nm}/ml in YPGal medium at different temperatures (W3031B, W303Δcoq9:TS19, and W303Δcoq9:TS19+mcHCOQ9 were grown at 25 and 37°C for 18.5 h; BY4741ΔCOQ9, W303ΔCOQ7, W303ΔCOQ4, W303ΔCOQ6 were grown at 30°C overnight). Crude mitochondria were isolated from 1 L of culture as described (Padilla-López et al., 2009). Mitochondria were further purified with an OptiPrep discontinuous iodixanol gradient as described (He et al., 2014). The bicinchoninic acid assay was used to measure the total protein concentration in purified mitochondria (ThermoFisher Scientific). Purified mitochondria were solubilized with digitonin as described (He et al., 2014), and 15 μg of mitochondria were separated by SDS-PAGE with 10% polyacrylamide gels. Proteins were transferred to Immobilon-P transfer membranes (Millipore) and immunoblot analyses were performed as described (He et al., 2014). The source and use of primary antibodies are described in Table 2. Goat anti-rabbit IgG H&L chain specific peroxidase conjugate (Calbiochem), 1:10,000 was used as the secondary antibody.

Purified mitochondrial proteins (13 mg; 2 mg/ml) were solubilized with 4 mg/ml digitonin as described earlier. The soluble digitonin extract was collected after 100,000 × g centrifugation (Optimax TLX). Co-precipitation was then performed on the solubilized mitochondria with Ni-NTA resin (Qiagen) as described in Allan et al. (2015). Briefly, Ni-NTA resin (800 μl bed volume) was equilibrated with two volumes of lysis buffer. Solubilized mitochondria and 8 ml of lysis buffer were added to 800 μl of pre-equilibrated Ni-NTA resin (bed volume) in a 15 ml Falcon tube and rotated for 1.5 h at 4°C. The resin-slurry was applied to a column and the flow-through was collected. The resin was washed twice with Ni-NTA wash buffer (W1 and W2) and eluted twice with Ni-NTA elution buffer (E1 and E2). Aliquots of each fraction were examined for presence of Coq polypeptides: 0.17% of the FT, 0.25% of W1, 0.25% of W2, 1% of E1, 0.5% of E2, and 1.25% of Ni-NTA resin were analyzed by SDS-PAGE followed by immunoblot analysis with antibodies against Coq9, human COQ9 (1:1,000), Coq6, 1:250 and Atp2, 1:4,000. Purified mitochondria (15 μg protein) from CNAP6: mcHCOQ9 were included as a control.

Expression of human COQ9 has previously failed to rescue yeast coq9 null mutant growth on non-fermentable carbon sources (Duncan et al., 2009; Hayashi et al., 2014). It seemed likely that the destabilization of other Coq polypeptides in the yeast coq9 null mutant might account for the inability of human COQ9 to rescue the mutant. To stabilize other Coq polypeptides in the coq9 null mutant, we co-expressed multi-copy COQ8 (mcCOQ8) with either single copy (scHCOQ9) or multi-copy human COQ9 (mcCOQ9) in a coq9 null yeast mutant (W303Δ9K) and tested its growth on YPG plate medium, containing glycerol as the sole non-fermentable carbon source. However, expression of mcCOQ8 alone did not enable human COQ9 to rescue the growth of the coq9 null mutant (Figure 1).

Next, we turned to the Δcoq9K:TS19 mutant, which retains detectable steady-state levels of the yeast Coq9-TS19 polypeptide and other Coq polypeptides, is able to grow on YPG at the permissive temperature (25°C) (He et al., 2015), but shows defective growth on YPG at the non-permissive temperature (37°C; Figure 1). Expression of mcCOQ8, scHCOQ9, or mcHCOQ9 were each able to rescue the growth of the Δcoq9:TS19 mutant on YPG at the non-permissive temperature. While expression of yeast COQ9 yielded the most robust rescue at the non-permissive temperature, expression of mcCOQ8 and mchCOQ9 also rescued growth of the coq9 null mutant, with schCOQ9 providing noticeably less rescue (Figure 1). At the permissive temperature it is not possible to distinguish the effects of single-copy human COQ9 or multi-copy COQ8. The yeast Δcoq9K:TS19 mutant still functions at 25°C, and its growth is similar to wild type. However, there is a slight increase of growth of Δcoq9K:TS19 harboring mcHCOQ9. Yeast cells were also plated on SD-Ura-Leu to confirm that W303Δ9K was successfully transformed with the two plasmids. The empty vector pRS426, which is the parent vector of mcCOQ8, was included as a control (EV). As expected, Δcoq9K:TS19 cannot be rescued by the empty vector (Figure 1). Therefore, the rescue effects were specific and depended on the expression of either human COQ9 or over-expression of yeast COQ8.

Yeast Coq9 is required to remove the nitrogen group from Q₆ intermediates derived from pABA (He et al., 2015). To assess whether this function is manifested by the expression of human COQ9 in yeast, we determined the de novo synthesis of ¹³C₆-Q₆ from either ¹³C₆-pABA or ¹³C₆-4HB. The yeast Δcoq9K:TS19 mutant was transformed with the designated plasmids and levels of Q₆ and Q₆ intermediates were determined at permissive and non-permissive temperatures. The presence of mcCOQ8 significantly increased the amount of ¹³C₆-Q₆ synthesized from ¹³C₆-pABA at the permissive temperature relative to the empty vector control (Figure 2A). At the non-permissive temperature, mcCOQ8 and both the scHCOQ9 and mcHCOQ9 plasmids increased the amount of ¹³C₆-Q₆ synthesized from ¹³C₆-pABA (Figure 2A). In contrast, mcCOQ8 and the human COQ9 homolog increased the amount of ¹³C₆-Q₆ synthesized from ¹³C₆-4HB at the permissive temperature, but only the human COQ9 homolog increased the amount of ¹³C₆-Q₆ synthesized from ¹³C₆-4HB at the non-permissive temperature (Figure 2D). There is a significant increase of ¹³C₆-DMQ₆ in Δcoq9K:TS19 with the expression of mcCOQ8 and both the scHCOQ9 and mcHCOQ9 (Figures 2B,E). Interestingly, mcHCOQ9 has the most dramatic effect on ¹³C₆-DMQ₆ levels when ¹³C₆-4HB was provided at non-permissive temperature (Figure 2E). These findings suggest that expression of human COQ9 rescues the Δcoq9K:TS19 mutant by increasing Q biosynthesis with 4HB as the precursor. The levels of ¹²C-Q₆ were elevated by mcCOQ8 and human COQ9 at both permissive and non-permissive temperatures (Figures 2C,F), perhaps indicating that mcCOQ8 and human COQ9 may enhance the utilization of other (unlabeled) ring precursors to increase Q content.

To investigate the mechanisms responsible for the rescue of the Δcoq9K:TS19 mutant by human COQ9, we determined whether human COQ9 stabilizes other Coq polypeptide levels. Wild type (WT), Δcoq9K:TS19 (Δ9K:TS19) and Δ9K:TS19+ mcHCOQ9 yeast were cultured in YPGal for 18.5 h at either 25 or 37°C, and mitochondria were isolated. Steady-state levels of Coq4, Coq6, Coq7, Coq9, and human COQ9 were analyzed by immunoblotting (Figure 3). The levels of Atp2 were analyzed as a loading control. At the non-permissive temperature, the expression of Coq9-ts19 resulted in lower steady-state levels of the other yeast Coq polypeptides and human COQ9. In contrast, in wild-type yeast, Coq4, Coq6, Coq7, and Coq9 levels were increased at 37°C (Figure 3). In the Δcoq9K:TS19 mutant, Coq9-ts19 levels were increased at 37°C (Figure 3A), while steady-state levels of the Coq4, Coq6, and Coq7 polypeptides were decreased at 37°C (Figure 3). The changes of Coq polypeptides in either wild type or the temperature-sensitive mutant at different temperatures did not result from changes in the corresponding COQ RNA levels (He et al., 2015). When human COQ9 was expressed in Δ9K:TS19 (Δ9K:TS19+ mcHCOQ9), the steady state levels of Coq4, Coq6, Coq7, and Coq9 were significantly increased at the permissive temperature (Figure 3). The results suggest that at permissive temperature the expression of human COQ9 stabilizes certain yeast Coq polypeptides. Two distinct polypeptides were detected in the mitochondria of Δ9K:TS19+mcHCOQ9 with the antibody against human COQ9. Based on the mass of the polypeptides, it seems likely that the larger polypeptide corresponds to unprocessed human COQ9 with the mitochondrial leader (39 kDa), and the smaller polypeptide represents processed human COQ9 (30.5 kDa). At non-permissive temperature, steady-state levels of human COQ9 were also decreased (Figure 3A). The observed decrease is specific to human COQ9 and Coq proteins of the CoQ synthome because the steady state levels of Atp2, the beta subunit of the F1 sector of the mitochondrial F₁F₀ ATP synthase, did not change at higher temperature. These results suggested that human COQ9 may be associated with the CoQ synthome.

To determine whether the human COQ9 polypeptide might interact with yeast components of the CoQ synthome (He et al., 2014), we expressed human COQ9 in the yeast strain Coq6-CNAP (CNAP6). We chose Coq6-CNAP as the bait protein because Coq9 is known to be important for the Coq6 hydroxylation step (Xie et al., 2012). A consecutive non-denaturing tag containing a His₁₀ tag and protein C epitope was integrated at the C-terminus of yeast Coq6, resulting in Coq6-CNAP (Allan et al., 2015). The CNAP6 strain has normal levels of Coq6 and Q₆ and co-precipitates other Coq polypeptides in the CoQ synthome (Coq4, Coq5, Coq7, Coq8, and Coq9) (Allan et al., 2015). Mitochondria purified from CNAP6:mcHCOQ9 were solubilized with digitonin and subjected to co-precipitation over Ni-NTA resin. Fractions corresponding to flow through (FT), washes (W1 and W2), eluates (E1 and E2), and beads after elution were analyzed by SDS-PAGE and Western blot. We found that unprocessed human COQ9 and yeast Coq9 co-purified with Coq6-CNAP, (Figure 4). As a negative control, we also blotted with an antibody against Atp2. As expected, Atp2 did not co-purify with Coq6-CNAP (Figure 4). The majority of human COQ9 was detected in the flow through and wash fractions, indicating the interaction between human COQ9 and the Coq6-containing complex is weak or that only a small fraction of the over-expressed human COQ9 interacted with the yeast CoQ synthome (which is not over-expressed).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.