Homozygous mice carrying kd (kidney disease) mutations in the gene encoding prenyl diphosphate synthase subunit 2 (Pdss2kd/kd) develop interstitial nephritis and eventually die from end-stage renal disease. The PDSS2 polypeptide in concert with PDSS1 synthesizes the polyisoprenyl tail of coenzyme Q (Q or ubiquinone), a lipid quinone required for mitochondrial respiratory electron transport. We have shown that a deficiency in Q content is evident in Pdss2kd/kd mouse kidney lipid extracts by 40 days of age and thus precedes the onset of proteinuria and kidney disease by several weeks. The presence of the kd (V117M) mutation in PDSS2 does not prevent its association with PDSS1. However, heterologous expression of the kd mutant form of PDSS2 together with PDSS1 in Escherichia coli recapitulates the Q deficiency observed in the Pdss2kd/kd mouse. Dietary supplementation with Q10 provides a dramatic rescue of both proteinuria and interstitial nephritis in the Pdss2kd/kd mutant mice. The results presented suggest that Q may be acting as a potent lipid-soluble antioxidant, rather than by boosting kidney mitochondrial respiration. Such Q10 supplementation may have profound and beneficial effects in treatment of certain forms of focal segmental glomerulosclerosis that mirror the renal disease of the Pdss2kd/kd mouse.

  • ubiquinone
  • mitochondrial lipid metabolism
  • genetic renal disease
  • focal segmental glomerulosclerosis
  • antioxidant

coenzyme q (Q or ubiquinone) is an essential factor of aerobic respiration and oxidative phosphorylation. Q also functions as a lipid-soluble antioxidant in cellular biomembranes and scavenges reactive oxygen species (4, 7). Yeast strains not able to produce Q show sensitivity to oxidative stresses such as hydrogen peroxide and autooxidized polyunsaturated fatty acids (10, 44). Q is a dietary supplement and is also used in therapeutic treatments for neurodegenerative and cardiovascular diseases and for statin-induced symptoms of Q deficiency (11, 35).

Q is composed of a benzoquinone ring with a polyisoprenoid side chain of varying length. Living organisms possess distinct isoforms of Q depending on the length of the isoprenoid side chain (17). For example, humans produce Q10, mice produce predominantly Q9 and small amounts of Q10, Saccharomyces cerevisiae produce Q6, and Escherichia coli produce Q8. The difference in the tail length of Q is determined by the long-chain polyprenyl diphosphate synthase in each organism (30). Long-chain trans-prenyl diphosphate synthases catalyze the condensation step for the isoprenoid tails from farnesyl diphosphate (C15) or geranylgeranyl diphosphate (C20) and isopentenyl diphosphate (C5). Structural analyses and site-directed mutagenesis have shown that short- and long-chain polyprenyl diphosphate synthases contain seven conserved regions designated as domains I through VII (Fig. 1A) and two aspartate-rich motifs (DDXXD) that are substrate binding sites in association with Mg2+ (12, 18, 28, 29, 45). These analyses suggest that amino acid residues at the fourth and fifth positions before the aspartate-rich motif in domain II and two amino acid residues in domain II have an important role in determining the length of the isoprenoid tails.

Fig. 1.

Amino acid sequence alignment of mouse nonaprenyl diphosphate synthase subunit 1 (PDSS1/mSPS1) and subunit 2 (PDSS2/mDLP1) isoforms. A: typical long-chain polyisoprenyl diphosphate synthases, such as PDSS1 and PDSS2-L, have conserved domains I through VII. An arrow shows the location of the kd/kd mutation V117M within conserved domain I. B: alternative splicing produces PDSS2-Short (PDSS2-S) containing exons 1–3b and PDSS2-Long (PDSS2-L) containing exons 1–3a and 4–8. PDSS2-S contains only domains I through III. The predicted polypeptide produced from the Pdss2 BAC clone 256E1 is also shown. One transgenic line of B6.kd/kd mice harboring an insertion of this BAC clone 256E1 exhibits sporadic rescue from kidney disease.

Polyprenyl diphosphate synthases are classified into two types of homo- or heterocomplexes. Most of the short polyprenyl diphosphate synthases (C10–C25) in both prokaryotes and eukaryotes are homodimers (27). In contrast, long-chain (C30–C50) polyprenyl diphosphate synthases are more diversified. In E. coli, octaprenyl diphosphate synthase forms a homodimer necessary for synthesis of the octaprenyl diphosphate tail for both Q and menaquinone (16), whereas the enzymes that supply hexa- and heptaprenyl diphosphate for menaquinone in the genus Micrococcus and Bacillus are heterodimers (47). The Coq1 polypeptide in S. cerevisiae functions on its own to produce hexaprenyl diphosphate; however, when expressed in fission yeast, Schizosaccharomyces pombe, it functions as a partner protein to produce decaprenyl diphosphate (49). In S. pombe, mice, and humans, long-chain prenyl diphosphate synthases are composed of two subunits (PDSS1/Dps1 and PDSS2/Dlp1) and form heterotetramers (39, 40). PDSS2/Dlp1, which is subunit 2 of nonaprenyl or decaprenyl diphosphate synthase in fission yeast and mammals, possesses the conserved domains of a prenyl diphosphate synthase. However, the subunit 2 lacks aspartate-rich motifs in domains II and VI (39, 40). PDSS2 orthologs also are found in Drosophila, Xenopus, and mammals. Therefore, the heterotetramer conformation of long-chain prenyl diphosphate synthases might be standard in higher eukaryotes.

Recently, several diagnostic examples of inherited Q deficiency, caused by mutations in Pdss1, Pdss2, human COQ2 (hCOQ2), or CABC1/COQ8 have been reported (8, 19, 23, 25, 26, 36). These patients have severe Q10 deficiency and severe mitochondrial disease phenotypes such as encephalomyopathy, ataxia, renal diseases, cerebellar atrophy, and hyperlactatemia. Oxidative phosphorylation activities in fibroblasts (19, 23, 25, 26), muscle extracts (9, 23), or muscle mitochondria (25) from these patients harboring a point mutation on Pdss1, Pdss2, hCOQ2, or CABC1/COQ8 are significantly lower than in samples from healthy individuals. Addition of either decylubiquinone or Q2, two soluble Q analogs, restored oxidative phosphorylation activity in cultured skin fibroblasts (19, 26) and muscle extracts (23) of the affected patients. Therefore, the respiratory deficiency and the symptoms in the patients are attributed to primary Q deficiency. In some cases, prolonged Q supplementation relieves the symptoms caused by Q deficiency (37). The reasons for the different responses to Q supplementation are still unclear.

In this study we have examined the nature of the Q deficiency in the Pdss2kd/kd mouse, which harbors a V117M mutation in PDSS2 and develops kidney disease in early adulthood (33, 34). The V117 residue is invariant in PDSS2 polypeptides from fission yeast to humans (39). Based on immunohistochemical profiling and electron microscopy, the phenotype of the Pdss2kd/kd mouse resembles the collapsing glomerulopathy variant of focal segmental glomerulosclerosis (3). We have shown that a deficiency in Q content is evident in kidney lipid extracts by 40 days of age and that dietary supplementation with Q provides a dramatic rescue of the proteinuria and interstitial nephritis. The results suggest that if human renal patients are identified who have PDSS2 mutations similar to that of the Pdss2kd/kd mouse, Q supplementation may have profound and beneficial effects.



All animal studies have been approved by the University of California, Los Angeles (UCLA) and University of Pennsylvania institutional review boards. Mice harboring the Pdss2kd/kd missense mutation (V117M) on the B6 genetic background, line G transgenic mice expressing a truncated form of Pdss2, and B6.Alb/cre,Pdss2loxP/loxP mice, in which the Pdss2 gene is knocked out in hepatocytes, have been described previously (14, 33, 34). Assays of urine albumin were performed with ELISA kits from Bethyl Laboratories as described previously (33). When the mice were euthanized, hematoxylin and eosin (H&E)-stained sections of the kidneys were prepared. The sections were examined blindly and scored as follows: 0, no tubular dilatation and no mononuclear cell infiltrates; 1, small focal areas of cellular infiltration and tubular dilatation involving <10% of the cortex; 2, involvement of up to 25% of the cortex; 3, involvement of up to 50% of the cortex; 4, extensive damage involving >75% of the cortex.

Preparation of mouse kidney and liver homogenates.

Whole kidneys and lobes of livers were dissected from euthanized mice and stored at −80°C until homogenization. Samples were homogenized in 5 ml of 1× PBS, pH 7.4 (137 mM NaCl, 8 mM Na2HPO4, 2.7 mM KCl, and 1.5 mM KH2PO4) at 4°C, with a total of 10 strokes with a tight-fitting Teflon pestle rotating at maximal speed with a Fisher Scientific Lab Stirrer LR400A. The homogenate was centrifuged at 1,000 g for 5 min, supernatants were removed to fresh vials, and protein concentrations were measured using the bicinchoninic acid assay (Pierce, Rockford, IL). Aliquots of each homogenate were transferred to 50-ml glass tubes and stored at −80°C until extracted.

Q10 supplementation.

Mice were provided Q10 supplements (Tishcon LiQsorb, 100 mg/ml) in their drinking water. In a pilot experiment, three males and two females were given 100 mg Q10/400 ml water (0.25 mg/ml), beginning at 47 days of age. While on this dose, both females had litters. After weaning, the progeny received 0.50 mg Q10/ml water. Mice consumed ∼12–16 ml of water per day; thus at this dose, the mice received ∼200 mg Q10·kg body wt−1·day−1. Some of their offspring were given a dose of 400·kg body wt−1·day−1, beginning at the time they were weaned.

Expression of murine PDSS1 and PDSS2 prenyl diphosphate synthase subunits in E. coli.

E. coli strain DH5α was used for general plasmid construction and protein expression (42). E. coli cells were grown in Luria Bertani (LB) medium with appropriate antibiotics such as 100 μg/ml carbenicillin, 100 μg/ml ampicillin, and 50 μg/ml kanamycin. The pGEX-KG (GE Healthcare Bio-Sciences, Piscataway, NJ) and pET-28c (Novagen, Madison, WI) plasmids served as vectors.

cDNAs prepared from either B6 or B6.Pdss2kd/kd liver (34) were used as template DNA to PCR amplify the short and long isoforms of Pdss2. The 0.8-kb amplicon containing the Pdss2-Short isoform cDNA was generated with B6 liver cDNA as template and two oligonucleotide primers, 5-EcoRI-mDLP1 5′-CCGAATTCGAATGAGCTCCGGCAG-3′ (creates an EcoRI site) and 3-SalI-mDLP1-Short, 5′-ACGCGTCGACTTACTTCATGTTGTCACTGCC-3′ (creates a SalI site). The amplified 0.8-kb fragment was digested with EcoRI and SalI enzymes and inserted into the same sites of pET28c to yield pET28c-PDSS2-S. The pET28c-PDSS2-S plasmid was digested with XbaI and SalI, and the resulting fragment was cloned into the same sites of pGEX-mSPS1 (39) to construct pM12S. Glutathione S-transferase (GST)-tagged-PDSS1 and His6-tagged-PDSS2-Short are coexpressed from the Tac promoter of pM12S.

A similar scheme was used to generate pM12L and pM12Lkd. The 1.2-kb amplicon containing the Pdss2-Long isoform cDNA was generated with B6 liver cDNA as template and the oligonucleotide primers, 3-SalI-mDLP1-L, 5′-ACGCGTCGACTCAAGAAAATCTGGTCACAGC-3′ (creates a SalI site) and 5-EcoRI-mDLP1. The amplified 1.2-kb fragment was digested with EcoRI and SalI enzymes and inserted into the same sites of pET28c to yield pET28c-PDSS2-L. The pET28c-PDSS2-L plasmid was digested with XbaI and SalI, and the resulting fragment was cloned into the same sites of pGEX-mSPS1 to construct pM12L. The same sequence of steps was used to construct pM12Lkd, except that cDNA prepared from B6.Pdss2kd/kd mouse liver was used as template DNA. pM12L and pM12Lkd drive expression of GST-tagged-PDSS1 and either His6-tagged-PDSS2-Long form or His6-tagged-PDSS2-Long form harboring the kd missense mutation (V117M).

DH5α harboring pM12S, pM12LB6, pM12Lkd, or the empty vector pGEX-KG were inoculated in 5 ml of LB medium with ampicillin and cultured at 30°C, 250 rpm, for 18 h. Aliquots of each bacterial culture (500 μl) were used to inoculate 50 ml of fresh LB medium with carbenicillin and incubated at 30°C, 250 rpm, until the cell density reached an optical density (OD600) of 0.4. Isopropyl-β-d-thiogalactoside (IPTG; Fisher Biotech) was added to a final concentration of 0.5 mM, and cells were incubated at 30°C, 250 rpm, for 4 h. Cells were collected and stored at −20°C for lipid extraction.

Q measurements: electrochemical and UV detection.

A known amount of the internal standard Q6 was added to all samples and calibration curve standards before lipid extraction. The same amount of Q6 was also added to the Q9 and Q10 standards prepared at different concentrations to generate an extracted standard curve. Lipid extractions were then performed on Q9 and Q10 standards, mouse tissue homogenates, and E. coli cell pellets by using a standard method. Each sample or standard was mixed with 0.5 ml of water, 9 ml of methanol, and 6 ml of petroleum ether. The mixtures were vortexed for 1 min and centrifuged at 910 g, and the top layer of petroleum ether was removed and transferred to a separate vial. Fresh petroleum ether (6 ml) was added to the remaining aqueous phase and vortexed for 1 min. The vials were subjected to centrifugation as before, and the second petroleum ether layer was removed. The process was repeated once more, and the combined upper organic phase was dried under N2 and resuspended in 200 μl of methanol. The quinones were then separated and quantified by reverse-phase (RP)-HPLC connected to an electrochemical detector as described previously (15), with the following exceptions: the precolumn electrode was set at +650 mV to oxidize all hydroquinones, and a Gilson 118 UV/Vis detector was utilized to detect quinones (275 nm) as they eluted from the column. The amounts of Q9 and Q10 in the standards and samples were calculated via the calibration curve.

Q measurements: RP-HPLC-multiple reaction monitoring.

Samples and standards were extracted as described above, except with 0.05 ml of water, 0.9 ml of methanol, and 0.6 ml of petroleum ether. Samples were resuspended in 200 μl of ethanol. Rhodoquinone-9 (RQ9) isolated from Ascaris suum and rhodoquinone-10 (RQ10) isolated from Rhodospirillum rubrum were used as internal standards. RP-HPLC-multiple reaction monitoring (MS/MS) was performed on a 4000 Q Trap-Hybrid triple-quad linear ion trap analyzer (Applied Biosystems/MDS Sciex) with an autosampler (CTC Analytics/LEAP Technologies, HTC PAL; 10-μl injections, 20-μl loop) and a Turbo-V source with the ESI probe inserted. Reverse-phase sample separation was performed via a pentafluorophenyl propyl column (5 μm, 4.6 × 100 mm; Phenomenex, Torrance, CA). The mobile phase was 95:5 acetonitrile-water containing 0.1% formic acid (2 ml/min). The mass spectrometer was set to the positive ion mode with the following parameters: declustering potential, 61 V for RQ9, 91 V for Q9 and RQ10, and 126 V for Q10; entrance potential, 10 V for all quinones; collision energy, 39 V for RQ9, 37 V for Q9, 45 V for RQ10, and 51 V for Q10; and collision exit potential, 12 V for RQ9 and 10 V for Q9, RQ10, and Q10. The samples were analyzed in the multiple reaction monitoring (MRM) mode: 210-ms dwell; MRM transitions: m/z 781.5/182.1 for RQ9, m/z 795.6/197.1 for Q9, m/z 848.7/182.1 for RQ10, and m/z 863.7/197.0 for Q10. Quantification was via embedded Analyst software version 1.4.2 (Applied Biosystems/MDS Sciex).

Coprecipitation of PDSS1 and PDSS2 polypeptides.

To evaluate whether PDSS2-Long-kd interacts with PDSS1, we adapted an affinity purification to isolate GST-PDSS1, GST, His-PDSS2-L, and His-PDSS2-L kd. GST-PDSS1 and GST were purified using glutathione Sepharose 4B (GE Healthcare BioSciences), and His-PDSS2-L and His-PDSS2-L kd were purified using Ni-NTA Superflow (Qiagen) according to the manufacturer's instructions. Briefly, BL21 (DE3) E. coli cells harboring pET28c-Pdss2-L, pET28c-Pdss2-L kd, pGEX-mSPS1 (39), and pGEX-KG empty vector grown in 50 ml of LB medium plus appropriate antibiotics were induced by 0.5 mM IPTG for 3 h at 30°C. After centrifugation to collect cells, the pellets were sonicated for 12 cycles of 10-s sonication and 10-s intervals on ice in buffer 1, composed of 50 mM sodium phosphate, 0.5 M sodium chloride, 5% (wt/vol) sucrose, 1% (vol/vol) glycerol, and 1% (vol/vol) Tween 20 (pH 7.0). After addition of protease inhibitors (Complete, EDTA-free; Roche), the sonicated samples were rotated gently for 1 h at 4°C. The samples were centrifuged at 18,000 g for 20 min at 4°C, and supernatants were subsequently transferred to new tubes.

To purify the GST-PDSS1 and GST polypeptides, the supernatants containing the solubilized proteins were incubated with glutathione Sepharose 4B resin for 1 h at room temperature and subjected to three washes with buffer 1. The bound proteins were subsequently eluted by 10 mM reduced glutathione in buffer 1. To purify the His-PDSS2-L and His-PDSS2-L kd, the supernatant with 10 mM imidazole was incubated with Ni-NTA Superflow resin for 1 h at room temperature, washed three times in buffer 1 containing 20 mM imidazole, and eluted in buffer 1 with 200 mM imidazole.

The purified proteins with each combination of PDSS1 and B6- and kd-isoforms of PDSS2-L were suspended in 1× PBS with 0.1% Tween 20 and 5 mM MgCl2. His-PDSS2-L B6 or His-PDSS2-L kd protein were incubated with GST-PDSS1 for 30 min and subsequently precipitated by Ni-NTA Superflow resin. A coprecipitation assay of GST and His-PDSS2-L B6 also was performed as a negative control at the same time. After the precipitation step, the resins were washed three times with 1× PBS with 0.1% Tween 20, 5 mM MgCl2, and 20 mM imidazole, and then precipitated proteins were analyzed by Western blotting.

Generation of polyclonal antisera to the PDSS2 polypeptides.

Polyclonal antisera were obtained from rabbits by using a standard immunization protocol (Cocalico Biologicals, Reamstown, PA). To generate antibody against PDSS2, the pET28c-PDSS2-L plasmid expressing the full-length of PDSS2 with a 6-His tag at the NH2 terminus was used to transform E. coli strain BL21 (DE3) (Novagen). The fusion protein was induced by 0.5 mM IPTG (Fisher Biotech) at 30°C for 3 h and purified from the bacterial lysate under denaturing conditions (8 M urea in 1× PBS, adjusted to pH 7.4) by Ni-NTA Superflow resin (Qiagen). Elution fractions containing partially purified 6× histidine-fused PDSS2 were concentrated by Amicon Ultra-10,000 MWCO (Millipore). These procedures were repeated until a total of 2 mg of the PDSS2 protein was recovered. The concentrated sample was subjected to SDS-PAGE on an 11% gel, 1.5 mm thick. After staining of the gel with Coomassie blue, the band corresponding to PDSS2 was excised and used as antigen. The specificity of the antisera was determined by immunoblotting with E. coli-produced PDSS2 polypeptide without any epitope tags. The antisera were affinity purified as described previously (32).


Determination of Q content in kidney and liver of B6 wild-type and B6.Pdss2kd/kd mutant mice.

Previous analyses indicated that the content of Q9 and Q10 in kidneys of 123- and 151-day-old B6.Pdss2kd/kd mice were only 15–20% of that present in similarly aged B6 wild-type littermates (33). Young B6.Pdss2kd/kd mice appear to grow and mature normally and onset of kidney disease is variable but usually is evident no earlier than 8 wk of age (14, 34). Given this relatively late onset of kidney disease, it is important to determine the age at which low Q levels manifest. As shown in Fig. 2 and Table 1, the content of Q9 and Q10 in kidney lipid extracts of B6.Pdss2kd/kd mice failed to increase with age, an increase that is dramatic in the B6 wild-type littermates. The content of Q9 and Q10 in kidney lipid extracts of B6 wild-type littermates 75 days of age or older was significantly higher compared with that of the B6.Pdss2kd/kd mice. To delineate the timing of the increase in quinone content, kidneys were obtained from B6 wild-type mice at 20, 30, 40, 50, and 60 days of age, and the content of Q9 and Q10 were determined by RP-HPLC/MS-MS as described. As shown in Table 2, the content of Q9 and Q10 was already elevated in kidney lipid extracts prepared from 40-day-old mice. Thus the B6.Pdss2kd/kd mice develop kidney disease several weeks after the normal increase in Q9 and Q10 content fails to occur.

Fig. 2.

Kidney Q9 and Q10 levels are significantly lower in B6.Pdss2kd/kd mice than in B6 wild-type mice. Wild-type B6, heterozygous B6.+/kd, and mutant B6.Pdss2kd/kd mice were killed at the ages shown, and Q levels were measured in kidney homogenates prepared from each mouse. Error bars indicate the standard deviation in 3 separate measurements of the same lipid extract. Levels of Q9 and Q10 in 123- and 151-day-old B6.Pdss2kd/kd mice and in 115- and 149-day-old B6 mice were previously reported (33).

View this table:
Table 1.

Content of Q9 and Q10 in kidney extracts of B6.Pdss2kd/kd and B6 mice

View this table:
Table 2.

Content of Q9 and Q10 in kidney extracts of young B6 mice

A similar trend was observed in liver, except that the increase in B6 wild-type liver Q9 content occurred later (Fig. 3 and Table 3). Hence, a lower content of liver Q9 in the B6.Pdss2kd/kd mice was evident only after 149 days of age.

Fig. 3.

Differences in liver Q9 content in B6.Pdss2kd/kd mice and B6 wild-type mice are evident by 149 days of age. Wild-type B6 and mutant B6.Pdss2kd/kd mice were killed at the ages shown, and Q levels were measured in liver homogenates prepared from each mouse. Error bars indicate the standard deviation in 3 separate measurements of the same lipid extract.

View this table:
Table 3.

Content of Q9 in liver extracts of B6. Pdss2kd/kd and B6 mice

Sporadic rescue of Q content B6.Pdss2kd/kd mutant mice by a partial Pdss2 transgene.

The BAC clone 256E1 contains exons 1–5 of the Pdss2 gene (see Fig. 1B) and was used to establish a rescued line of B6.Pdss2kd/kd mice (line G) (34). The B6.Pdss2kd/kd-line G mice were the only one of 13 transgenic lines generated that expressed the Pdss2 transgene at a high level. Mouse kidney lipid extracts prepared from B6.Pdss2kd/kd and B6.Pdss2kd/kd-line G transgenic mice were analyzed for Q9 and Q10 content as described in Fig. 2. Remarkably, rescue of Q content in the line G mice was sporadic. Intriguingly, a low content of Q9 and Q10 in the transgene-positive mice was found to correlate with elevated levels of albumin in the urine (Fig. 4 and Table 4).

Fig. 4.

Expression of a truncated form of Pdss2 in B6.Pdss2kd/kd mice restores content of Q in kidney extracts to wild-type levels in most cases. Wild-type B6, mutant B6.Pdss2kd/kd, and line G B6.Pdss2kd/kd transgenic mice (indicated by asterisks) were killed at the ages shown, and Q9 and Q10 levels were measured in kidney homogenates taken from each mouse. Error bars indicate the standard deviation in 3 separate measurements of the same lipid extract.

View this table:
Table 4.

Content of Q9 and Q10 in kidney extracts of B6.Pdss2kd/kd, B6.Pdss2kd/kd-line G, and B6 mice

Functional analysis of PDSS2-Long, PDSS2-LongV177M, and PDSS2-Short isoforms of prenyl diphosphate synthase.

Prenyl diphosphate synthase activity has been shown to require both PDSS1 and PDSS2 subunits (39). The plasmid construct pM12L allowed for coexpression of mouse PDSS1 and PDSS2-Long in E. coli (Fig. 5A). Transformation of an ispB prenyl diphosphate synthase mutant (E. coli strain KO229) with this construct rescued aerobic growth and converted the quinone synthesized in E. coli cells from Q8 to Q9 (Fig. 5B and Table 5) (39). Coexpression of PDSS1 and PDSS2-Long-V117M (from plasmid pM12Lkd) failed to rescue the E. coli ispB mutant (data not shown). However, transformation of wild-type DH5α E. coli with pM12Lkd allowed for the expression of Q9 in addition to Q8. E. coli DH5α harboring either empty vector (pGEX-KG) or pM12S (directing coexpression of PDSS1 and PDSS2-Short) failed to generate appreciable levels of Q9, indicating that the PDSS2-Short form fails to support prenyl diphosphate synthase activity. The Q9 content in DH5α E. coli cells harboring pM12Lkd (34.8 ± 7.6 pmol Q9/mg wet weight cell pellet) was significantly lower than in cells harboring pM12L (93.7 ± 6.0 pmol Q9/mg wet weight cell pellet). These results indicate that heterologous expression of the kd mutant form of PDSS2 in E. coli recapitulates the deficiency observed in the B6.Pdss2kd/kd mice. Capture of GST-tagged PDSS1 coprecipitated PDSS2-Long and PDSS2-Long-kd (Fig. 6). Thus the V117M mutation in domain I of PDSS2-Long-kd does not prevent its association with PDSS1.

Fig. 5.

Heterologous expression of the mutant kd form of Pdss2 in Escherichia coli recapitulates the deficiency in Q biosynthesis observed in B6.Pdss2kd/kd mice. A: pGEX-KG (empty vector) produces just the glutathione S-transferase (GST) polypeptide. pM12S contains the open reading frame (ORF) of Pdss1 inserted in frame after the carboxyl terminus of GST (GST-PDSS1) and Pdss2-S (Short form) inserted in frame after a six-Histidine tag (His6-PDSS2-S). pM12L expresses GST-PDSS1 and His6-PDSS2-L (Long form). pM12Lkd expresses GST-PDSS1 and His6-PDSS2-L-V117M (Long form harboring the kd mutation). B, BamHI; E, EcoRI; S, SalI, and Xb, XbaI. An asterisk shows the kd/kd mutation position V117M, and “rbs” indicates a ribosomal binding site from pET-28c to allow polycistronic expression of the mouse polypeptides. B: Q8, normally produced in E. coli, and Q9, generated by the heterologous expression of PDSS1 and PDSS2, were measured in lipid extracts prepared from E. coli DH5α harboring each of the constructs designated in A. E. coli strain KO229 harboring a deletion of the ispB prenyl diphosphate synthase gene was rescued by pM12L. Error bars indicate the standard deviation in 3 separate measurements of the same lipid extract.

Fig. 6.

Coprecipitation of PDSS1 and PDSS2 polypeptides. His-PDSS2-L or His-PDSS2-L-V117M polypeptides were incubated with GST-PDSS1 for 30 min and subsequently precipitated by Ni-NTA Superflow resin (lanes 2 and 3). A coprecipitation assay of GST and His-PDSS2-L also was performed as a negative control at the same time (lane 1). Antibodies to GST (A) and the 6-His tag (B) were utilized to recognize GST-PDSS1 and His-PDSS2 polypeptides, respectively. The asterisks indicate nonspecific bands recognized by the GST antibody. GST*(truncated), observed in lanes 2, 3, 5, and 6 in A, bottom, is likely to result from proteolysis of the GST-PDSS1 fusion protein. Lanes 4, 5, and 6 show the relative amounts of input protein used for each coprecipitation assay and correspond to lanes 1, 2, and 3, respectively. Similar amounts of GST-PDSS1 were coprecipitated with both His-PDSS2-L and His-PDSS2-Lkd fusion proteins (A, lanes 2 and 3, top). In contrast, GST did not coprecipitate with the His-PDSS1-L fusion protein (A, lane 1, bottom).

View this table:
Table 5.

Content of Q8 and Q9 in extracts of Eschericia coli expressing Pdss1 and isoforms of Pdss2

Steady-state polypeptide levels of PDSS2-Long and PDSS2-Short were examined in liver mitochondria isolated from B6 wild-type, B6.Pdss2kd/kd, and B6.Alb/cre,Pdss2loxP/loxP mice. Both PDSS2-Long and PDSS2-Short polypeptides were readily detected in liver mitochondria isolated from B6 wild-type mice (Fig. 7). Steady-state levels of both polypeptides were lower in the liver mitochondria of B6.Pdss2kd/kd mice, suggesting that the lower Q9 and Q10 content in liver of the B6.Pdss2kd/kd mice may result from a combination of impaired polyprenyl diphosphate activity and lower steady-state levels of the PDSS2 polypeptide. Only small amounts of the PDSS2-Long polypeptide were detectable in liver mitochondria isolated from B6.Alb/cre,Pdss2loxP/loxP mice, where the Pdss2 gene is knocked out in hepatocytes (33). A drastic decrease in liver Q9 content was evident in these mice; the content of Q9 measured in two different animals (95 days old) using the RP-HPLC-MS/MS assay (see materials and methods) was 8.9 ± 2.4 and 33.5 ± 6.7 pmol Q9/mg liver protein, whereas normal Q9 content in age matched controls was ∼10 times higher.

Fig. 7.

Western blot analysis of PDSS2 polypeptides in mice liver mitochondria. An anti-PDSS2 antibody was utilized to detect PDSS2 polypeptides in 120 μg of mouse mitochondria protein (A). An existing amount of PDSS2-L and -S in B6 mouse mitochondria (lane 1) was significantly higher than in kd/kd and alb/cre mutant mice (lanes 2 and 3, respectively). PDSS2-S is present in relatively smaller amounts than PDSS2-L. An anti-yeast F1βATP synthase subunit that cross-reacts with mouse F1βATP synthase subunit provided a loading control (B).

Rescue of renal disease by Q10 supplementation.

In a pilot experiment, three male and two female B6.Pdss2kd/kd mice were supplemented with 0.25 mg Q10/ml drinking water and compared with their untreated littermates. When the mice were tested at 132 days of age, the mutants had less urine albumin than their untreated littermates (7.2 ± 5.8 vs. 13.5 ± 8.3 mg/24 h, respectively). However, this difference was not significant as indicated by Student's t-test (P = 0.288). Therefore, beginning at weaning, the offspring produced by the B6.Pdss2kd/kd-supplemented mice received double the dose (0.50 mg Q10/ml drinking water), equivalent to ∼200 mg Q10·kg body wt−1·day−1 (see materials and methods). After mice reached at least 130 days of age, urine albumin was determined. Supplemented mice had on average a fivefold reduction in the urine albumin content compared with unsupplemented, age-matched B6.Pdss2kd/kd mice, and their nephritis scores were significantly lower (P = 0.004, Student's t-test) than those of untreated B6.Pdss2kd/kd controls (Table 6). However, the content of Q9 and Q10 of supplemented animals (n = 7; 267.2 ± 86.7 pmol Q9/mg kidney protein and 37.1 ± 16.1 pmol Q10/mg kidney protein) was not significantly different from the content in the untreated B6.Pdss2kd/kd controls (n = 3; 353.0 ± 40.5 pmol Q9/mg kidney protein and 45.2 ± 6.0 pmol Q10/mg kidney protein). Doubling the dose to 400 mg·kg body wt−1·day−1 did not improve the effectiveness of the treatment (Table 6). Examples of H&E-stained sections from the kidneys of Q10-treated and untreated Pdss2kd/kd mice are shown in Fig. 8, A and B, respectively.

Fig. 8.

Hematoxylin- and eosin-stained sections of kidneys from Q10-treated and untreated B6.Pdss2kd/kd mice. A: no histological evidence of disease in a 162-day-old female treated with Q10. The amount of urine albumin measured was 3.7 mg/24 h. B: severe interstitial nephritis in an untreated 128-day-old female. The amount of urine albumin measured was 57.34 mg/24 h. Note the severe interstitial inflammation, tubular dilatation, and glomerular crescents. Magnification, ×100.

View this table:
Table 6.

Effect of adding Q10 to the drinking water of mutant mice


Mice harboring the homozygous kd/kd mutation in Pdss2 show normal growth and fertility but develop kidney disease as young adults. Severe tubulointerstitial nephritis and tubular dilation are accompanied by morphological disorders of mitochondria in kidney and liver (34). The kd/kd mutation (V117M) occurs within the conserved domain I of PDSS2 (Fig. 1A). A decreased Q content was observed in adult B6.Pdss2kd/kd mice (33). In this study we have shown that Q9 and Q10 content in kidney lipid extracts of young B6.Pdss2kd/kd mice and their B6 wild-type littermates (20 days of age) are very similar. However, Q9 and Q10 content in kidney lipid extracts of B6 wild-type mice 40 days of age or older are significantly higher compared with that in B6.Pdss2kd/kd mice. The normal increase at 30–40 days of age may reflect unidentified physiological changes that occur at weaning, which normally occurs soon after day 20. The failure to increase the content of Q in kidneys of the B6.Pdss2kd/kd mutant mice at 40 days of age precedes by at least several weeks the renal disease that develops at about 8–10 wk of age. In contrast, Q content in liver from B6.Pdss2kd/kd mice remains similar to wild-type mice until later in adulthood (∼150 days of age). This profound deficiency in Q content in the kidney early in life is likely responsible for the renal disease phenotype.

Mice express two polypeptide isoforms of PDSS2 due to alternative splicing, PDSS2-Long (exons 1–8) and PDSS2-Short (exons 1–3b) (Fig. 1B) (34). Previous work has shown that PDSS2 forms a heterotetramer with PDSS1 and that both PDSS1 and PDSS2-Long are required for polyprenyl diphosphate synthesis (39). In this study we have shown that coexpression of the short isoform of PDSS2 with PDSS1 in E. coli fails to reconstitute synthesis of Q9. We hypothesize that the PDSS2-Short form may repress nonaprenyl diphosphate synthase activity by forming an unproductive complex with PDSS1. E. coli harboring a construct coexpressing PDSS1 and PDSS2-Long-V117M (the kd mutant form) had significantly lower Q9 content than did cells expressing PDSS1 and PDSS2-Long (wild type). These results recapitulate the low Q content observed in the B6.Pdss2kd/kd mice and suggest that the V117M mutation in Pdss2 occurs in a conserved domain of polyprenyl diphosphate synthase and partially impairs synthesis of nonaprenyl and decaprenyl diphosphate. The mutation may affect enzymatic activity and/or heterotetramer formation. However, capture of PDSS1 coprecipitates the PDSS2-Long-V117M polypeptide, suggesting that the V117M mutation does not prevent interaction of the two subunits.

Given that the short form of PDSS2 (exons 1–3b) failed to restore Q biosynthesis, it is curious that the BAC clone 256E1, which contains only a segment of the Pdss2 gene, established a rescued line of B6.Pdss2kd/kd mice (line G) (34). However, this clone includes exons 1–5 and in most cases rescued both kidney disease and Q levels. Remarkably, rescue of kidney disease and Q content in line G mice was sporadic. Intriguingly, a low Q content in transgene-positive mice was found to correlate exactly with elevated levels of protein in urine.

The data suggest that the renal disease phenotypes in Pdss2kd/kd homozygous mice are caused by a decreased content of Q in the kidney. The renal disease could result from either inadequate respiratory function due to low levels of Q or increased oxidative stress because QH2 is a potent antioxidant. The partial rescue by dietary supplementation with Q10 is consistent with both ideas. Several lines of evidence suggest that Q/QH2 is serving a crucial antioxidant role. Conditional knockout mice with deletions of Pdss2 in the glomerular podocytes (B6.Podocin/cre,Pdss2loxP/loxP) fully recapitulate the kd/kd phenotype (33). However, kidney disease does not develop in the B6.PEPCK/cre,Pdss2loxP/loxP mice, in which Pdss2 is deleted in renal tubular epithelial cells. Podocytes are not thought to have an energetic requirement as high as that of renal tubular epithelium, supporting the hypothesis that susceptibility to oxidative damage is the critical factor in the Pdss2kd/kd model of kidney disease. The kidney content of Q10 was not enhanced by supplementation with Q10, a finding consistent with other similar Q10 supplementation studies (21). However, longer periods of supplementation (e.g., 17 mo) do elevate both Q9 and Q10 kidney content (43). It is well established that short-term supplementation does impact levels of Q10 in the liver and blood plasma (24). Hence, it seems likely that the benefits of Q10 supplementation may be exerted on the podocytes via their filtration of blood plasma components (46). Although in this study we have not evaluated oxidative stress or antioxidant status, other evidence that the Pdss2kd/kd mice may be subject to oxidative stress is the finding that their susceptibility to renal disease is markedly diminished by either dietary restriction or a germ-free environment (13). Dietary restriction is associated with decreased oxidative stress (48) and has been shown to decrease lipid peroxidation products and apoptosis in aged rat kidneys (22). Similarly, mice reared in a germ-free environment recapitulate many aspects of diet-restricted animals and show enhanced AMP-activated protein kinase activity (1), a hallmark of the response to dietary restriction in yeast and worms (5).

It is possible that renal disease in Pdss2kd/kd mice results from a defect in an isoprenylated product other than Q. The PDSS2 polypeptide produces nona- and decaprenyl diphosphate, and studies of mevalonate labeling of mice and rats suggest these long-chain prenyl groups may be utilized in kidney protein prenylation (6, 31), similar to the better characterized protein posttranslational modification with farnesyl or geranylgeranyl groups (20). However, the fact that disease susceptibility is ameliorated by Q10 supplementation suggests that Q itself is the most essential substrate with regard to this phenotype. Patients with mutations in the COQ2 gene show renal disease with collapsing glomerulopathy (2, 9), further indicating the causative role is a deficiency in Q itself.

Several human Q10-deficient patients have responded dramatically to Q10 supplementation (36–38, 41). If their response is primarily based on the antioxidant function of Q10 rather than its role in mitochondrial respiration, then antioxidants that are more readily taken up by the target tissues could be more effective therapies than Q supplementation. Experiments are in progress to determine whether other antioxidants may be more effective than Q10 in preventing kidney disease in Pdss2kd/kd mice.


This work was supported in part by National Institutes of Health (NIH) Grants GM45952 and AG19777 (to C. F. Clarke) and DK55852 (to D. L. Gasser). A. L. Lunceford received support from the Ruth L. Kirschstein National Research Service Award issued from NIH Training Grant GM007185. The LC-MS/MS determination of Q content was supported by National Center for Research Resources Grant S10 RR024605. Histological preparations were done in the Morphology Core for Molecular Studies in Digestive and Liver Diseases at the University of Pennsylvania, which is supported by NIH Grant P30 DK50306.


  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.


View Abstract