Development and fertility in Caenorhabditis elegans clk-1 mutants depend upon transport of dietary coenzyme Q8 to mitochondria.

The Caenorhabditis elegans clk-1 mutants lack coenzyme Q(9) and instead accumulate the biosynthetic intermediate demethoxy-Q(9) (DMQ(9)). clk-1 animals grow to reproductive adults, albeit slowly, if supplied with Q(8)-containing Escherichia coli. However, if Q is withdrawn from the diet, clk-1 animals either arrest development as young larvae or become sterile adults depending upon the stage at the time of the withdrawal. To understand this stage-dependent response to a Q-less diet, the quinone content was determined during development of wild-type animals. The quinone content varies in the different developmental stages in wild-type fed Q(8)-replete E. coli. The amounts peak at the second larval stage, which coincides with the stage of arrest of clk-1 larvae fed a Q-less diet from hatching. Levels of the endogenously synthesized DMQ(9) are high in the clk-1(qm30)-arrested larvae and sterile adults fed Q-less food. Comparison of quinones from animals fed a Q-replete or a Q-less diet establishes that the Q(8) present is assimilated from the E. coli. Furthermore, this E. coli-specific Q(8) is present in mitochondria isolated from fertile clk-1(qm30) adults fed a Q-replete diet. These results suggest that the uptake and transport of dietary Q(8) to mitochondria prevent the arrest and sterility phenotypes of clk-1 mutants and that DMQ is not functionally equivalent to Q.

Ubiquinone, or coenzyme Q (Q), 1 is a polyprenylated benzoquinone lipid that is predominantly associated with the inner mitochondrial membranes of eukaryotes. The number of prenyl groups in the tail is species-specific. Escherichia coli contains Q 8 and Caenorhabditis elegans contains Q 9 , where the subscript designates the number of isoprene units in the tail. Q has an essential role in mitochondrial electron transport from complexes I and II to III (1). Q also acts as the following: a lipidsoluble antioxidant (2); a component of plasma membrane electron transport (3); a component of uridine synthesis (4); and as a component of the proton-pumping function of the uncoupling proteins UCP1, UCP2, and UCP3 (5,6). Q is synthesized within mitochondria in a series of steps that is initiated by linking the polyisoprene tail to 4-hydroxybenzoic acid. The head group is then modified in a series of hydroxylation, methylation, and decarboxylation steps to form the final product (7). A defect in hydroxylation, late in the biosynthetic pathway, causes accumulation of demethoxy-Q 8 (DMQ 8 ) in E. coli (8), and a similar function was attributed to the COQ7/CAT5 locus in Saccharomyces cerevisiae (9 -11). Studies on prokaryotic homologues of COQ7 suggest that the encoded polypeptide is a membranebound di-iron carboxylate protein that catalyzes the hydroxylation of DMQ to demethyl-Q (12).
Mutation of the clk-1 gene in C. elegans, a homologue of COQ7, leads to slow embryonic and post-embryonic development, slow adult behaviors, and reduced brood sizes under standard culture conditions (13,14). clk-1 mutants have a defect in Q biosynthesis, and they accumulate the biosynthetic intermediate demethoxy-Q 9 (DMQ 9 ) instead of producing Q 9 (15). However, mitochondrial function in clk-1 mutants fed a standard E. coli diet is only slightly diminished compared with wild-type when assayed by Rhodamine-6G uptake, NADH-cytochrome c reductase, and succinate-cytochrome c reductase activities (15,16). These results have been interpreted as indicating that DMQ 9 is a functional substitute for Q in the respiratory chain and that the Clk-1 phenotype results from some non-mitochondrial function of Q. These conclusions contrast with previous analyses of DMQ 8 function in E. coli in which partial function of complex I is observed, but DMQ 8 is inactive in succinate dehydrogenase (complex II) activity (17). Significantly, the C. elegans studies employed nematode cultures that were grown on media containing OP50 E. coli, which provides a dietary source of Q 8 . The clk-1 mutant animals fed a Q-less diet as hatchlings arrest development as L2 larvae (18). If the Q-less diet is fed at a later stage, the clk-1 mutant dauer larvae develop into sterile adults.
The relationship between the withdrawal of dietary Q 8 and the severe clk-1 mutant phenotypes of sterility and larval arrest is not understood. To investigate this question, we quantified the amounts of accumulating quinones at different developmental stages of C. elegans. These experiments establish the normal developmental pattern of Q accumulation in wild-type nematodes and provide a basis for the analysis of defects present in the clk-1 mutants. The quinone content of N2 and clk-1(qm30) animals at two different stages of development fed either Q-replete (OP50) or Q-less E. coli was determined. We found that Q 8 is present in the lipid extracts of OP50-fed nematodes and the levels diminish greatly when they are moved to a Q-less diet. This decline in Q 8 levels is observed when either hatchlings or dauer larvae are fed Q-less E. coli. In the first case, Q 8 is only supplied maternally in the egg, whereas in the second the Q 8 derives from a Q-replete diet fed prior to dauer formation. In both situations, the clk-1(qm30) mutants fed the Q-less diet contain large amounts of DMQ 9 yet are developmentally arrested in the first case or are sterile adults in the second. This suggests that DMQ 9 is inadequate to sustain larval growth or germ line development. The clk-1(qm30) mutants complete development to fertile adults only when adequate levels of Q 8 are attained. Finally, we found that mitochondria isolated from both N2 and clk-1(qm30) animals fed OP50 contain diet-derived Q 8 . Therefore, uptake of dietary Q 8 and its delivery to mitochondria appear to be essential for both development and fertility of clk-1 mutants.
Isolation of Nematode Developmental Stages-Cultures of gravid adults were treated with alkaline hypochlorite solution to obtain eggs (19). The collected eggs were allowed to hatch without food to obtain synchronous L1 larvae. For the N2 developmental stages, L1 larvae were fed OP50 and grown to the desired stage. To confirm the developmental stages, cells in the gonadal or vulval lineage were scored using differential interference optics (21,22). The N2 strain used in the study of the above developmental stages was obtained from the Caenorhabditis Genetics Center. All subsequent analyses were done with N2 animals obtained from Dr. Hekimi (McGill University, Canada).
For the timed larval feeding experiments, starved N2 and clk-1(qm30) L1 larvae were fed either Q-replete or Q-less E. coli for 8 h at 20°C and then harvested. In addition, a sample of arrested clk-1(qm30) larvae was harvested after 30 h of feeding Q-less E. coli. For samples of post-dauer 2 (PD2) and young adult animals, dauer larvae were isolated with 1% SDS treatment, washed, and fed either Q-less or Q-replete E. coli. The samples were harvested when the desired stage was obtained, rather than after a certain number of hours, because the timing of development varied by the nematode genotype and E. coli fed. All nematode samples were collected by centrifugation, separated from bacteria and debris by sucrose flotation, rinsed in M9 buffer (17 mM potassium phosphate, 42 mM sodium phosphate, 8.6 mM sodium chloride, 1 mM magnesium sulfate) for 30 min, and frozen by dripping into liquid nitrogen.
Isolation of Nematodes after Long-term Feeding-N2 dauer larvae were isolated by treatment with 1% SDS and then fed either Q-replete or Q-less E. coli. The larvae were allowed to develop and reproduce at 20°C. After 4 days, a small sample was removed and transferred to fresh media. After another 4 days, a small sample was taken and again placed in fresh media. These cultures were allowed to develop into dauer larvae. These dauer larvae were isolated with 1% SDS and cleaned and frozen as above.
Cultivation at Different Temperatures-To study the effect of temperature on quinone content, dauer larvae were isolated with 1% SDS, fed OP50, and grown to reproductive adults at either 16.5°C or 25°C. The successive F1 generation was allowed to enter dauer stage. These animals were diluted and allowed to recover at either 16.5°C or 25°C. When gravid, the adults were subjected to alkaline hypochlorite solution, and the eggs were collected. These eggs were allowed to hatch in S medium overnight, and subsequently the starved larvae were fed and allowed to grow to young adulthood at either 16.5°C or 25°C. These adults were harvested, cleaned, and frozen as above.
Quantification of Q Levels-Lipid extraction of animals with a Q 6 internal standard and quantification by HPLC linked with electrochemical detection (ECD) was performed as previously described (18), with the following exceptions: the precolumn electrode potential was E ϩ650 mV, which completely oxidizes Q and DMQ. The electrode potentials E1 Ϫ650 mV and E2 ϩ600 mV were used to obtain a plateau for the current response, thus achieving maximum signals for Q and DMQ. However, rhodoquinone (RQ) is still within the linear range on a current/voltage curve and therefore results in more variability in measurement of this compound. DMQ 8 isolated from the E. coli strain AN78 (Table I) was used as an external standard to quantify the DMQ 9 levels by ECD analysis of clk-1(qm30) subcellular fractions. Cultures of AN78 were grown in LB with high aeration overnight. Total lipid extracts were isolated from the bacterial pellet as in Ref. 18 and separated by HPLC, and the peak fraction corresponding to DMQ 8 was collected and dried down. The sample was differentially extracted by MeOH/petroleum ether, and the petroleum ether layer was removed, dried, and resuspended in 100 l of 9:1 MeOH/EtOH. The concentration of DMQ 8 was calculated by recording the absorbance of an ethanol solution at 271 nm and by using the extinction coefficient E M 271 ϭ 14,500 (15). The chromatographic areas recorded for DMQ 8 were used to extrapolate the amount of DMQ 9 in clk-1(qm30) animals, because the mole/area ratios are identical. The relative ratios comparing milligrams of wet weight to milligrams of dry weight and milligrams of wet weight to milligrams of protein were determined. Wet weights were measured for L2, L4, and young adult stages, and then these samples were dried overnight in a Speedvac. Dry weights were determined. Finally, the pellets were resuspended in 1 N NaOH and assayed for protein concentration by the bicinchoninic acid assay. The ratios of milligrams of wet weight to milligrams of dry weight were 6.6 Ϯ 0.6, and the milligrams of wet weight to milligrams of protein ratios were 11.8 Ϯ 0.5. This suggests that for these various stages, the animals have similar water content.
Mitochondrial Isolation-New cultures were inoculated at a concentration of 2000 dauer larvae/ml S medium, fed OP50, and incubated at 20°C until the animals reached adulthood. The animals were cleaned by sucrose flotation and washed thoroughly. The nematodes were resuspended in Isolation Buffer (IB: 210 mM mannitol, 70 mM sucrose, 0.1 mM EDTA, 5 mM Tris, pH 7.4) with 1 mM PMSF at 10 ml per 5 g of nematodes. The samples were kept cold on ice throughout the fractionation procedure. The animals were homogenized with a Kontes ground glass tissue grinder (Fisher catalogue number K885450-0025) using 15 strokes. The volume was increased to 25 ml with IB ϩ 1 mM PMSF and centrifuged at 750 ϫ g for 10 min. The supernatants were saved, and another 10 ml of IB ϩ 1 mM PMSF was added to the pellets. The pellets were resuspended and homogenized again with another 15 strokes, and the volumes were increased to 25 ml and centrifuged at 750 ϫ g for 10 min. The homogenates were microscopically examined to verify disruption. An aliquot of the combined supernatants was saved as total lysate. The supernatants were centrifuged at 12,000 ϫ g for 10 min. An aliquot of the resulting supernatant was saved as post-mitochondrial supernatant. The mitochondrial pellets were gently resuspended in 12 ml of IB. The mixture was centrifuged at 750 ϫ g for 10 min. The supernatants were collected, avoiding the pellets, and were centrifuged at 12,000 ϫ g for 10 min. The final mitochondrial pellets were resuspended in IB. All subcellular fractions were stored at Ϫ20°C until use.
Western Analysis-Fractions were assayed for protein concentration by the bicinchoninic acid assay. Western analysis of 40 g of protein from the total lysates, the post-mitochondrial supernatants, and the mitochondrial fractions from N2 and clk-1(qm30) and 2.5 g and 1 g samples of protein from an OP50 lysate were performed by electrophoresis on 12% Tris-glycine gels, followed by transfer to Hybond ECL nitrocellulose. The primary antibody to yeast F 1 ␤-ATPase was used at a 1:5000 dilution, and the primary antibody to E. coli cytochrome o oxidase was used at a 1:200 dilution. Although generated against the holo-enzyme cytochrome oxidase, this antibody only reacted well against subunit II, at 35 kDa. 2 Horseradish peroxidase-linked secondary antibodies to rabbit IgG were used in a 1:2000 dilution.
Identification of Quinones by Atmospheric Pressure Chemical Ionization Mass Spectrometry-A PerkinElmer Life Sciences Sciex (Thornhill, Canada) API III triple quadrupole mass spectrometer was tuned and calibrated as previously described with an Ion Spray TM source using a solution of polypropylene glycol (23). The manufacturer-supplied atmospheric pressure chemical ionization (APCI) source consisting of a heated nebulizer and a corona discharge needle was then installed and operated at 450°C in the flow injection mode using methanol as the flowing solvent (0.2 ml/min). Dried HPLC fractions were redissolved in MeOH/water/petroleum ether (1/0.5/1, all v/v), vigorously mixed, and then centrifuged, and the petroleum ether layer was removed to a clean 2 K. Matsushita and H. R. Kaback, UCLA, personal communication.
Hfr, metB, ubiF (8) clk-1 Development and Fertility Requires Q in Mitochondria container and dried in a stream of nitrogen. The dried sample was redissolved in an appropriate volume of methanol/ethanol (9/1, v/v, typically 100 l for the putative Q 8 fraction and 400 l for the putative DMQ 9 fraction) and injected into the methanol stream that was flowing into the heated nebulizer (20 l/injection). Positive ion spectra were collected

RESULTS
Quinone Levels Vary with Development and Temperature-To establish the normal pattern of quinone accumulation throughout development, N2 populations were fed standard Q-replete OP50 and harvested at various life stages, lipids were extracted, and quinone content was quantified. C. elegans larvae proceed through four molts, designated as the four larval stages L1, L2, L3, and L4, before reaching young adulthood. Fig. 1A shows the quinone content of N2 harvested at the stages of L1, L2, L4, and young adulthood. When expressed as picomoles of quinone per mg wet weight, there was an overall increase in the levels of all the quinones at the L2 larval stage, followed by a decrease at L4. Quinone content is higher in young adults than in L4. Analyzed on a separate day, Fig. 1B shows that the quinone content in L1 larvae feeding for 4 h is higher than that in eggs. The absolute response from the HPLC/ECD system for the same standard samples varied from day to day, as is typical (24). However, within one working day the response variation was smaller, as indicated by the error bars. Consequently, samples were analyzed in batches that could be accommodated within one working day and are presented together in the graphs.
Quinone amounts were found to be affected by temperature.
N2 animals were reared at either 16.5°C or 25°C for multiple generations to avoid a maternal effect (Fig. 1C). Q 8  Long-term Feeding of Q-less Diet Alters Quinone Profiles of N2 Dauer Larvae-To investigate the extent to which the diet contributes to Q 8 content, wild-type animals were fed Q-less E. coli for several generations. N2 were divided into cultures containing either Q-replete OP50, or the Q-less foods GD1 or JCG⌬-1. Small populations were expanded two times sequentially in Q-less food to avoid lingering effects from the Q-replete food. The lipids were extracted from dauer larvae harvested from these cultures, and the quinones were quantified by HPLC/ECD. Representative data from animals fed either OP50, GD1, or JCG⌬-1 are displayed in Fig. 2A. In the presence of Q-less food, Q 8 levels dropped to 0.24 Ϯ 0.02 and 0.34 Ϯ 0.05 pmol/mg wet weight for the GD1 or JCG⌬-1 fed strains, respectively. In contrast, RQ 9 and Q 9 amounts increased dramatically as compared with those fed OP50. For instance, the OP50-fed animals contained 23.56 Ϯ 0.65 pmol Q 9 /mg wet weight, while N2 fed GD1 or JCG⌬-1 accumulated 32.25 Ϯ 0.70 pmol Q 9 /mg wet weight and 32.77 Ϯ 1.64 pmol Q 9 /mg wet weight, respectively. Interestingly, if one compares the combined pmol Q 8 ϩ Q 9 totals, there is no difference in total amount of Q in these animals (Fig. 2B). There may be a compensatory response to the lack of dietary Q 8 by which wild-type animals increase endogenous Q 9 production.
The clk-1 Developmental Arrest Results from Insufficient Q-All of the clk-1 mutant alleles fail to grow beyond the L2 stage when fed Q-less E. coli from hatching (18). To determine the types and levels of Q isoforms present at the point of growth arrest, synchronized starved L1 larvae were transferred to either Q-replete or Q-less food and analyzed as a function of FIG. 1. Analysis of quinone content in wild-type C. elegans through development. A and B, HPLC/ECD was used to quantify the quinone levels in N2 cultured at 20°C and fed OP50 at five stages: eggs, L1, L2, L4, and young adulthood. C, N2 young adults reared at either 16.5°C or 25°C were harvested and analyzed for quinone content as above. Q 7 and menaquinone (MK 8 ), derived from the E. coli food, and Q 10 were also detected in the lipid extracts, based on coelution with known quinone standards, at levels Ͻ1 pmol/mg wet weight. Error bars represent standard deviation from three separate injections of the same sample, and data shown are representative of six independent experiments for panel A, and three independent experiments in panels B and C.
FIG. 2. Long-term feeding of Q-less food to wild-type populations. A, N2 animals were cultivated in media containing either OP50, GD1 (ubiG Ϫ ), or JCG⌬-1 (ubiG Ϫ ) for at least three generations, as under "Experimental Procedures." Dauer larvae were harvested, and the lipid contents were analyzed by HPLC/ECD as in Fig. 1. B, combined picomoles of Q 8 ϩ Q 9 levels from panel A show similar total Q amounts regardless of the E. coli fed.

clk-1 Development and Fertility Requires Q in Mitochondria
time. Concurrent cultures of clk-1(qm30) and control N2 eggs were collected and allowed to hatch overnight in the absence of food. In N2 animals, the level of Q 8 dropped to 0.09 Ϯ 0.02 pmol/mg wet weight and the amount of RQ 9 and Q 9 increased in these starved L1 larvae relative to the amounts in the eggs (Fig. 3A). The RQ 9 levels increased from 0.28 Ϯ 0.02 to 6.82 Ϯ 0.15 pmol/mg wet weight while the Q 9 amounts increased from 7.12 Ϯ 0.17 to 27.33 Ϯ 0.72 pmol/mg wet weight. Upon feeding with OP50 for 8 h, Q 8 levels increased to 1.38 Ϯ 0.18 pmol/mg wet weight while Q 9 levels decreased to 11.87 Ϯ 1.41 pmol/mg wet weight. N2 fed either of two Q-less E. coli strains, GD1 (ubiG Ϫ ) or MU1227 (ubiA Ϫ ), showed slightly greater amounts of Q 9 than those fed OP50 (20.25 Ϯ 0.15 and 14.53 Ϯ 0.85 pmol/mg wet weight, respectively), while the Q 8 levels remained below the limit of detection. clk-1(qm30) mutants displayed a similar pattern, but produced DMQ 9 rather than Q 9 (Fig. 3B). When the clk-1(qm30) larvae remained in the Q-less food beyond 8 h, they arrested at the L2 stage, as previously described (18), which allowed a later time point to be collected. Quinone measurements of the arrested larvae that had been fed Q-less E. coli for 30 h showed a distinct increase in RQ 9 and DMQ 9 levels, from 4.25 Ϯ 0.39 pmol RQ 9 /mg wet weight and 9.98 Ϯ 0.03 pmol DMQ 9 /mg wet weight at the 8-h time point to 9.05 Ϯ 0.19 pmol RQ 9 /mg wet weight and 15.85 Ϯ 0.06 pmol DMQ 9 /mg wet weight after 30 h. The Q 8 levels on the ubiG Ϫ Q-less food were 0.15 Ϯ 0.05 and 0.12 Ϯ 0.04 pmol/mg wet weight for the 8-and 30-h time points, respectively. Q-less E. coli strains with defects in other genes necessary for E. coli biosynthesis were fed to the clk-1(qm30) larvae, because these E. coli accumulated different biosynthetic intermediates. The clk-1(qm30) larvae fed MU1227 (ubiA Ϫ ) and JC7623⌬-1 (ubiE Ϫ ) showed similar RQ 9 and DMQ 9 increases at the 30-h time point (data not shown). In contrast, clk-1(qm30) mutants fed Q 8 -replete E. coli contained 0.58 Ϯ 0.02 pmol Q 8 /mg wet weight and less RQ 9 (2.82 Ϯ 0.26 pmol/mg wet weight) and DMQ 9 (7.02 Ϯ 0.11 pmol/mg wet weight) at the 8-h time point and were able to progress through development. Therefore, the critical difference that enables growth of clk-1 mutants appears to be the uptake of sufficient Q 8 from the E. coli diet.
Sterility of clk-1 Adults Results from Inadequate Q-clk-1 mutants will form dauer larvae under crowded culture conditions if fed OP50. If these dauer larvae are fed Q-less food, they will recover and grow to sterile adults (18). The quinone levels of these sterile adults were compared with quinone levels present in N2 dauer larvae fed either Q-replete or Q-less E. coli during recovery. Fig. 4 shows data from animals harvested at PD2 and young adulthood. PD2 morphologically resemble L4 (25). The quinone contents of N2 animals fed either Q-less E. coli (GD1 (ubiG Ϫ ), JC7623⌬-1 (ubiE Ϫ ), or JCG⌬1 (ubiG Ϫ )) or OP50 were similar, except for the 10-fold decline in Q 8 accumulation in nematodes cultured on the Q-less foods (Fig. 4 A and B) and clk-1(qm30) (C and D) dauer larvae were allowed to recover in either Q-replete (OP50) or Q-less (GD1 (ubiG Ϫ ), JCG⌬-1 (ubiG Ϫ ), JC7623⌬-1 (ubiE Ϫ )) food and were harvested at the PD2 and young adult stages. Their lipids were extracted, and the quinones were quantified by HPLC/ECD as in Fig. 1.
clk-1 Mitochondria Contain Q 8 -N2 and clk-1(qm30) adults fed OP50 were isolated, homogenized, and fractionated to separate a mitochondria-enriched fraction from the total nematode lysate. Antibodies generated to the yeast Atp2 polypeptide (the ␤ subunit of F 1 -ATPase) were previously found to recognize the homologous ATP2 polypeptide of C. elegans (26). Western analysis showed that the mitochondrial protein F 1 ␤-ATPase present in the total lysate was absent from the post-mitochondrial supernatant and was enriched in the mitochondrial fraction (Fig. 5A). In addition, the nematode subcellular fractions did not show detectable levels of the E. coli marker cytochrome o oxidase, indicating that the samples are substantially free of E. coli contamination. Hence, the quinones present in these samples do not derive from E. coli cells either present in the gut or adhering to the surface of the nematodes. Lipids from each fraction were then separated by HPLC and analyzed by ECD for its quinone content per mg protein (Fig. 5, B and C). The mitochondrial fraction contained the majority of the quinones, with Q 9 levels in N2 reaching 3.2 nmol/mg of protein and DMQ 9 amounts in clk-1(qm30) at 2.9 nmol/mg. The Q 8 levels in the mitochondria are substantial, at 0.65 nmol/mg of protein in N2 and 0.45 nmol/mg of protein in clk-1(qm30). Therefore, not only is dietary Q 8 present in total lipid extracts of whole animals, but Q 8 is also present in isolated mitochondria.
The structural assignments of DMQ 9 and Q 8 isolated from clk-1(qm30) mitochondrial lipid extracts were verified by APCI mass spectrometry and tandem mass spectrometry. The APCI mass spectra of authentic Q 9 (Sigma) and DMQ 8 (purified from AN78) yielded intense signals at m/z 795.6 and 697.5, respectively, corresponding to their protonated molecular ions (calculated 795.6291 and 697.5560 Da, respectively, Fig. 6, A and B). MS/MS on these parent ions (Fig. 6, A and B, insets) yielded intense diagnostic fragment ions at m/z 197.0 and 167.0, respectively, corresponding to tropylium ions of the polar head groups (calculated 197.0814 and 167.0708, respectively). The APCI mass spectra of the putative Q 8 and DMQ 9 samples (Fig.  6, C and D) isolated from clk-1(qm30) mitochondrial lipid extracts showed the analogous ions at m/z 727.5 and 765.6, respectively, corresponding to the predicted protonated molecules (calculated 727.5665 and 765.6186 Da, respectively). Furthermore, MS/MS on these parent ions (Fig. 6, C and D,  insets) yielded the expected fragment ions at m/z 197.3 and 167.0, respectively, in agreement with the predicted corresponding head groups. DISCUSSION Knowledge of quinone production and uptake throughout C. elegans development is necessary for understanding the phenotypic differences of clk-1 mutants fed a diet lacking or containing a source of Q. Here we present the temporal analysis of the quinone types and amounts in wild-type animals from FIG. 5. Analysis of mitochondrial quinones isolated from N2 and clk-1(qm30) adults. A, C. elegans mitochondrial fractions are enriched in the mitochondrial inner membrane protein F 1 ␤-ATPase as assayed by Western analysis with antibodies generated against yeast F 1 ␤-ATPase. Antibodies to E. coli cytochrome o oxidase, of which subunit II at 35 kDa is immunodominant, were used to assess the degree of E. coli contamination. Each lane contains 40 g of protein from the total lysates (tl), post-mitochondrial supernatants (pms), and mitochondrial fractions (m) from N2 and clk-1(qm30) adults. 2.5-and 1-g samples of OP50 lysate are also analyzed. B, lipid extracts of total lysates, postmitochondrial supernatants, and mitochondria from N2 adults were analyzed by HPLC/ECD as in Fig. 1. Q 7 , MK 8 , and Q 10 were also found in the extracts at levels Ͻ8 pmol/mg of protein for the total lysate and post-mitochondrial supernatant and Ͻ85 pmol/mg of protein in the mitochondrial fraction. C, lipid extracts of total lysates, post-mitochondrial supernatants, and mitochondria from clk-1(qm30) adults were analyzed by HPLC/ECD as in Fig. 1. The levels of pmol/mg of protein of Q 8 , RQ 9 , and DMQ 9 are displayed. Q 7 , MK 8 , and Q 10 were also identified, but at levels Ͻ7 pmol/mg of protein in the total lysate and postmitochondrial supernatant and Ͻ40 pmol/mg of protein in the mitochondrial fraction.

clk-1 Development and Fertility Requires Q in Mitochondria
different developmental stages and from young adults grown at different temperatures. The results show Q content is regulated by both developmental stage and temperature. Q levels were highest during the L2 larval and young adult stages. The Q levels in young adult N2 reared at 25°C were twice the amount as those reared at 16.5°C (Fig. 1C). The increase in Q content in young adult animals and the higher temperature are probably related to the increased number of mitochondria. In wild-type animals, mitochondrial DNA (mtDNA) copy number is developmentally regulated. mtDNA copy numbers showed a 5-fold increase from the L3 to L4 stage, and an additional 6-fold increase from the L4 stage to reproductive adulthood (27). Separately, quantitation of fluorescently labeled mitochondria in muscle cells showed that the mitochondrial number almost doubled in animals reared at 25°C versus 15°C (28). In both situations, increased numbers of mitochondria would require more Q.
The sterility of clk-1 dauer larvae recovered and grown to adulthood on Q-less food may result from both limited Q stores and an increased demand for mitochondria. Germ cell production occurs during the L4 and adult stages in C. elegans (21). clk-1 mutant dauer larvae derived from Q 8 -replete culture conditions and subsequently fed Q-less food have sufficient Q 8 for somatic development to adulthood. However, the Q 8 reserve is insufficient for development of functional germ cells. These sterile clk-1 adult animals contain DMQ 9 and RQ 9 , but the Q 8 content is 10-fold lower than that present in fertile clk-1 adults fed a Q 8 -replete diet (Fig. 4, C and D). The demand for increased mtDNA has been tied to germ line development and function with a small increase in L4 for sperm production and a large increase in adult hermaphrodites for oocyte production (27). Thus, the sterility phenotype of the clk-1 mutants emerging from dauer larvae on Q-less food coincides with the demand for increased mtDNA copy number and with increased levels of Q in response to production of functional germ cells, especially oocytes.
The L2 developmental arrest of clk-1 mutants fed Q-less food from hatching also coincides with increased Q production in wild-type, but there is no corresponding increase in mitochondria reported for this stage. Energy metabolism in wild-type C. elegans is regulated during larval development, with a key transition after the L1 stage (29). There is high activity of the glyoxylate pathway during embryogenesis, but this decreases during the L1 stage (30). Upon maturation to L2 larvae, there is an increased reliance on trichloroacetic acid cycle metabolism, which continues through the L3 and L4 stages (29). The arrest at the L2 larval stage of the clk-1 mutant fed Q-less food from hatching coincides temporally with this metabolic shift. The L2 developmental arrest, with L2-like gonads, is not a unique phenotype for the Q-deficient clk-1 mutants, because similar arrests are seen in animals harboring null mutations in genes encoding a subunit of complex I, nuo-1(ua1), or of complex V, atp-2(ua2) (26). Other phenotypic similarities for clk-1, nuo-1, and atp-2 mutant animals include slowed movements, slowed pharyngeal pumping, and slowed defecation. In addition, a maternal effect has been described for each of these mutants. Maternal contributions allow survival through embryogenesis for the nuo-1 and atp-2 mutants, because the phenotype produced by RNA-mediated interference of these genes is embryonic lethality (26). Thus, clk-1 mutant animals reared on Q-less food contain DMQ 9 and RQ 9 and behave as if they are respiratory-deficient.
However, other investigators have offered a different interpretation. The assertion has been made that DMQ 9 supports mitochondrial respiration in clk-1 mutant mitochondria (15,31). These studies demonstrated the function (or partial function) of DMQ 9 in complexes I, II, and III and emphasized that oxygen consumption rates are nearly normal in mitochondria isolated from clk-1 mutant nematodes. The rescue of growth arrest of clk-1 mutant nematodes by a Q-replete diet is attributed to the inability of DMQ to replace Q at a non-mitochondrial site, because Q derived from the diet was considered incapable of being transported into mitochondria (31,32). Although activity of respiratory complexes in mitochondria isolated from clk-1 mutant nematodes was found to be nearly normal (15,16), these animals had been fed a standard diet containing Q 8 , and neither the possible contribution of dietary Q 8 nor endogenously produced RQ 9 to assays of respiration was considered. Assays performed with mclk-1 mutant ES mouse cell extracts (which contain DMQ 9 but no RQ 9 ) showed profound defects in complex II ϩ III activity (32). Homozygous mclk-1 mutant animals display an embryonic lethal phenotype (32,33).
Recently, a knockout mutation in the nematode gene coq-3 was observed to be maternal effect larval lethal (31). coq-3 homozygotes, from a heterozygous mother, develop slowly, are small, and most are sterile. The sterility of coq-3 mutants, and of clk-1 mutants fed Q-less E. coli, may be due to a similar mechanism that we believe to be related to biogenesis of germ line mitochondria. The dietary contribution of Q 8 does not rescue the lethality of the coq-3 null mutant from homozygous mothers (31). The difference in severity between coq-3 and clk-1 null mutants may derive from the biosynthetic step affected. Coq3p has been shown to be responsible for both Omethylation steps in Q biosynthesis (34) and results in the production of a very early non-redox active intermediate. Additionally, it is likely that these coq-3 mutants lack not only Q 9 and DMQ 9 , but also RQ 9 , because the biosynthetic pathways of Q 9 and RQ 9 probably share most early steps (35). Given the potential complete lack of quinones, the coq-3 mutant larvae that do hatch may lack energy to power muscle contraction of the pharynx and thereby fail to take up sufficient Q 8 from the diet due to an inability to feed.
As shown here, arrested clk-1 mutants fed a Q-less diet contain DMQ 9 and RQ 9 , and levels of both quinones continue to rise over time (Fig. 4). It is possible that RQ 9 could partially compensate for the defect in complex II by operating fumarate reductase in reverse (36). Normally, those eukaryotes that can survive periods of anoxia utilize RQ and fumarate reductase as an essential step of malate dismutation (37). It is presumably the operation of malate dismutation that enables wild-type C. elegans to survive 1-2 days of anaerobic treatment (38). Indeed, it is likely that survival of the clk-1 mutants up to and during the L2 larval arrest depends upon RQ 9 and fumarate reductase functioning in both its intended forward direction in malate dismutation and in reverse to produce fumarate from succinate. Importantly, such RQ 9 function is still unable to support development of the clk-1 mutant larvae on Q-less diets past the L2 stage or to sustain reproduction. Hence, despite increasing levels of RQ 9 , the clk-1 mutant larvae remain arrested as L2 larvae until Q 8 is supplied. Therefore, although dietary Q 8 may indeed serve functions at non-mitochondrial sites, it is reasonable to expect that it also serves crucial functions within the mitochondria.
Q is involved in numerous cellular processes both within and outside of the mitochondria (39), and it is possible that DMQ fails to functionally replace Q at multiple sites. In addition to a role in complexes I, II, and III of respiratory electron transport, Q is required for the operation of glycerol-3-phosphate dehydrogenase, fatty acid ␤-oxidation, pyrimidine synthesis at the dihydroorotate dehydrogenase step (4), and the detoxification of sulfide by sulfide dehydrogenase (40), and it is a substrate of lipoamide dehydrogenase, a component of ␣-keto-acid dehydrogenase complexes that oxidize pyruvate, ␣-ketoglutarate, and branched chain ␣-keto acids (41). The studies presented here show dietary Q 8 is present in the mitochondria isolated from either clk-1 mutants or wild-type nematodes (Fig. 5) and would be expected to participate in these mitochondrial functions.
There is precedence for the uptake and transport of dietary Q to the mitochondria in species other than C. elegans. When dietary Q 10 is supplied to human patients with Q 10 deficiencies (42) or to old rats (43,44), mitochondrial Q 10 levels increase and respiratory function is enhanced. Q-deficient yeast coq null mutants, including a coq7 null mutant, readily take up Q 6 from their media (11,45,46), and such Q 6 supplementation restores growth on non-fermentable carbon sources. The uptake of Q 6 results in a substantial increase in Q 6 levels in the mitochon-dria, in levels of succinate cytochrome c reductase enzyme activity, and in levels of the cytochrome c 1 polypeptide of the bc 1 complex (46). These results indicate that the critical function of COQ7/clk-1 is in Q biosynthesis, because such defects can be rescued by dietary supplementation with Q.
In C. elegans, it is not known whether dietary Q is assimilated by all cells or just in certain cells, although it is clearly delivered to oocytes as evidenced by the presence of Q 8 in eggs. It is also unknown whether the E. coli Q isoform (Q 8 ) behaves differently from the de novo synthesized Q 9 in nematodes. Previous work shows that mammalian mitochondrial respiratory complexes can use Q 8 as well as Q 9 . Respiratory chain activity is restored upon addition of various Q isoforms (Q 1 -Q 10 ) to mitochondria depleted of endogenous Q by pentane extraction (47). The exogenous Q isoform has the potential to influence non-mitochondrial as well as numerous mitochondrial functions. The studies presented here indicate that Q content in C. elegans is regulated during development, in response to temperature of cultivation and in response to diet. Dietary Q 8 decreased adult lifespan in C. elegans (48), and the model presented relied upon uptake and delivery of Q 8 to the mitochondria, which has been substantiated here. The key aspects of metabolism that are altered with the environmental and genetic manipulations that alter Q content remain to be defined in development and aging.