Extension of Life-Span in Caenorhabditis elegans by a Diet Lacking Coenzyme Q

See allHide authors and affiliations

Science  04 Jan 2002:
Vol. 295, Issue 5552, pp. 120-123
DOI: 10.1126/science.1064653


The isoprenylated benzoquinone coenzyme Q is a redox-active lipid essential for electron transport in aerobic respiration. Here, we show that withdrawal of coenzyme Q (Q) from the diet of wild-type nematodes extends adult life-span by ∼60%. The longevity of clk-1,daf-2, daf-12, and daf-16 mutants is also extended by a Q-less diet. These results establish the importance of Q in life-span determination. The findings suggest that Q and thedaf-2 pathway intersect at the mitochondria and imply that a concerted production coupled with enhanced scavenging of reactive oxygen species contributes to the substantial life-span extension.

Coenzyme Q (Q) functions as a carrier of electrons and protons in respiratory complexes I, II, and III. Electron transport by complexes I and III establishes a proton gradient across the inner mitochondrial membranes, which is coupled to synthesis of adenosine triphosphate (ATP) by the F0/F1 ATPase (complex V). Q also functions in cells as a lipid-soluble antioxidant, as a redox-active component of plasma-membrane electron transport, in uridine synthesis, and as a cofactor for the proton-pumping activity of mitochondrial uncoupling proteins (1, 2). Caenorhabditis elegans clk-1 mutants lack the endogenously synthesized isoform Q9 (where 9 designates the number of isoprene units in the polyisoprene tail) and instead rely on the Escherichia coliisoform Q8, which is supplied by their standard diet for growth (3). The clk-1 gene encodes a di-iron carboxylate enzyme responsible for the final hydroxylase step in the synthesis of Q (4). The clk-1 mutant animals accumulate the Q-biosynthetic intermediate demethoxy-Q9 (DMQ9) (5). These findings suggest that the increased life-span and slowed rates of development reported for clk-1 mutants result from a decreased level of Q.

To test whether a decrease in dietary Q alters life-span, we fed a diet lacking Q to wild-type worms during the adult phase. The growth and adulthood phases are separated in C. elegans by a molt from the fourth larval stage (L4) to adult worm. Therefore, to avoid altering development from standard conditions, we switched larvae from the standard diet of OP50, a Q8-replete E. coli, to a diet of GD1, a Q-lessE. coli, at the L4 stage. Wild-type adults fed the Q-less diet throughout adulthood had a median life-span 59% longer than those fed standard Q-replete E. coli (Fig. 1A and Table 1). A Q-less diet also extended life-span for fer-15(b26ts) nematodes, which carry a temperature-sensitive mutation that causes sterility. When either wild-type or fer-15(b26ts) nematodes were fed the GD1:pAHG E. coli strain (which is rescued for theubiG gene and thereby rendered Q8-replete), the life-span was similar to those of animals fed OP50. These results indicate that the short life-span does not result from the genetic background of the OP50 E. coli strain and is unrelated to the presence or absence of kanamycin in the growth medium. Rather, the short life-span of wild-type and fer-15 mutant animals appears to depend on a dietary source of Q.

Figure 1

Adult C. elegans life-span is extended by a diet of Q-less E. coli. (A) Survival curves of wild-type N2 (cross-hatched squares) andfer-15(b26ts) (filled circles) adults fed either Q-less or Q-replete E. coli throughout their adult life. The diets were standard Q8-replete E. coli OP50 (dashed lines); GD1, a Q-less strain of E. coli (solid lines); or rescued GD1:pAHG, a Q8-replete strain ofE. coli (dotted lines). All experiments with non-OP50E. coli contained kanamycin in the plate media to kill Q-replete OP50 that accompanied the worms upon transfer to the Q-less food. (B) Adult life-span survival curves of N2 fed a variety of Q-less E. coli diets. Diets from hatching and throughout adulthood were standard Q8-replete E. coli OP50 (cross-hatched squares) or the Q-less GD1 E. coli strain, harboring a mutation in ubiG (open circles). L4 larvae were transferred to a Q-less diet of GD1 (filled circles); JC7623Δ4-1, ubiE::Kan (triangles); or JCΔG, ubiG::Kan (crosses). (C) Survival curves of wild-type N2 (cross-hatched squares),clk-1(e2519) (filled squares),clk-1(qm30) (triangles), andclk-1(qm51) (circles) adults fed either Q-less or Q-replete bacteria throughout their adult life. The lines are as described in (A). The clk-1 mutations have been sequenced (23), and e2519 is predicted to affect a carboxylate ligand (E148K) (4) whereas the other alleles result in truncated polypeptides.

Table 1

Life-span analysis of C. elegans fed different E. coli diets. Independent experiments are in groups beginning with N2.

View this table:

The GD1 E. coli ubiG mutation causes accumulation of 2-octaprenyl phenol, an intermediate specific to the Q-biosynthetic pathway in prokaryotes (6). To test whether the increased longevity depended on this intermediate rather than on the absence of Q, we used a Q-less E. coli ubiEmutant strain that accumulates 2-polyprenyl-6-methoxy-1, 4-benzoquinol as food for wild-type worms. Because this ubiE mutation was present in JC7623, a different E. coli genetic background than that of GD1, we included JCΔG, a strain harboring theubiG mutation in the JC7623 genetic background, as a control (6). An increase in nematode life-span was observed for each of the mutant E. coli strains, suggesting that the observed life-span extension was independent of the type of Q-biosynthetic intermediate that accumulated (Fig. 1B and Table 1). These results imply that a lack of Q8 extends life-span.

Wild-type animals were fed the Q-less diet from egg to old age so that we could investigate the effect of the absence of Q during development on adult aging. A small fraction of this adult population died earlier than did animals raised on the Q-replete diet (Fig. 1B). Although most of the remaining nematodes raised and maintained on the Q-less diet survived longer than those raised and maintained on the Q-replete diet, they had decreased median survival when compared with nematodes switched as L4 larvae to the Q-less food. Of the nematodes raised on Q-less food, 45 out of 165 had reproductive defects, whereas of the nematodes raised on Q-replete food and switched to the Q-less diet as L4 larvae, only 21 out of 181 had reproductive defects (7). These defects are presumed to be developmental, because that is the period during which the reproductive structures are formed. This suggests that dietary Q is beneficial during development and engenders formation of robust adults. These results suggest that, although development proceeds more reliably with Q in the diet, adult consumption of Q8 shortens nematode life-span.

The short life-span of wild-type animals fed OP50 is considered normal for C. elegans. Animals with mutations in previously described life-span–determination genes (clk-1and daf) were tested for a long-lived (Age) phenotype when fed Q-less food. clk-1(qm30,qm51, and e2519) mutant animals were transferred to the Q-less diet as L4 larvae, and adult life-span was determined. Mutant strains fed the Q-less diet showed increases in median life-span relative to the same genotype fed the Q-replete diet (Fig. 1C and Table 1). However, insufficient Q appears to be detrimental because, regardless of diet, nearly all the clk-1 animals, which do not synthesize Q, died at younger ages than did the wild-type animals on a Q-less diet. The clk-1–mutant animals fed the Q-replete diet showed increased survival when compared with wild-type animals only at the oldest ages (Fig. 1C) (7). The median life-span for the clk-1–mutant population fed the Q-replete diet was decreased relative to that of the wild type. This result differs from a reported 7 to 20% increase in mean life-span forclk-1 (8, 9). These published analyses describe animals that started from eggs and include the increased developmental time of the clk-1 mutants in the number of days lived, whereas our experiments start on the first day of adulthood. A decrease in Q levels, whether due to a Q-less diet or a mutation in the clk-1 gene, results in longevity.

In C. elegans, the daf genes regulate both development and life-span (10). This longevity pathway includes the daf-2 and age-1 genes that increase life-span when mutated. Participation of these gene products in an insulin-like signaling pathway is inferred from their molecular identity with mammalian homologs (11, 12). Traditionally, suppression analysis is used to determine the gene products required for a particular phenotype. For example, the longevity phenotype of daf-2(e1370) is considered to be suppressed by daf-16 because the double-mutantdaf-16(m26); daf-2(e1370) is short- lived. Hence, daf-2 mutant longevity requires wild-type DAF-16 activity (10, 13). By similar criteria, the life-span extension by laser ablation of germ line precursor cells requires wild-type DAF-16 and DAF-12 activities (14). We performed suppression tests for the Age phenotype due to withdrawal of dietary Q. As previously observed, when fed a Q-replete diet, the daf-16(m26) anddaf-12(m20) adults have shorter life-spans than wild-type animals (Fig. 2A and Table 1). The life-spans of the daf-16 and daf-12 mutants fed a Q-less diet were longer than those of wild-type animals fed a Q-replete diet. Neither the daf-12(m20) nor thedaf-16(m26) mutation suppressed the life-span extension generated by the Q-less diet. Thus, thedaf-12 and daf-16 gene products are not required for the longevity effect mediated by the diet of Q-less bacteria.

Figure 2

Life-span of daf mutant animals is extended by withdrawal of dietary Q. (A) Survival curves of N2 (cross-hatched squares), daf-12(m20) (circles), and daf-16(m26) (triangles) adults fed either Q-less or Q-replete E. coli throughout their adult life. The diets were standard Q-replete E. coli OP50 (dashed lines) or GD1, a Q-less strain of E. coli (solid lines). (B) Survival curves of N2 (cross-hatched squares),clk-1(qm30) (filled squares),daf-2(m41), a mutation in the ligand-binding domain (24) (circles), and daf-2(e1370), a mutation in the kinase domain (11) (triangles) adults. The lines are as described in (A). Animals were raised at 15°C and fed OP50 until reaching L4. L4 larvae were transferred to the restrictive temperature (25°C) and were fed either Q-less or Q-replete E. coli throughout their adult life. (C) Survival curves of N2 (cross-hatched squares),daf-2 (e1370); daf-12(m20) (filled squares), daf-2(e1370)clk-1(qm30) (circles), anddaf-12(m20) (triangles) adults. The lines are as described in (A). The animals were handled as described in (B).

We examined the effect of a Q-less diet on the life-span of long-lived daf-2 mutants. The restrictive temperature, 25°C, was used to obtain the largest increase in life-span relative to that obtained with the wild type (10). Under these conditions, both daf-2(e1370) anddaf-2(m41) adults displayed longer median and maximum life-spans when fed a Q-less as compared with a Q-replete diet (Fig. 2B and Table 1). The C. elegans strains with the longest life-spans are double mutants of daf-2(e1370) with mutations in either the daf-12 or clk-1 gene (8, 10). The daf-2(e1370) and clk-1 longevities are additive, whereasdaf-2(e1370) and daf-12 are synergistic (8, 10). The long-lived double mutants were tested along with the single-mutant controls on Q-less and Q-replete diets. There was no increase fordaf-2(e1370); daf-12(m20) fed the Q-less food (Fig. 2C and Table 1). There was a small increase for daf-12(m20), and perhaps it is this phenotype that we observed in the daf-2(e1370); daf-12(m20) fed a Q-less diet for which a 2-day increase of an already long life-span was not significant. The increased median life-span of daf-2(e1370)clk-1(qm30) animals was further extended when they were fed a diet lacking Q. Thus, the longevity mechanisms are additive. There are two possible interpretations with regard to mechanism. Either DAF-2 signaling and decreased Q constitute separate parallel pathways, both needed for a later step, or they act serially in a single pathway, each with a partial effect.

For wild-type nematodes, the percent life-span extension resulting from the Q-less diet relative to the Q-replete diet was smaller at 25°C than it was at 20°C (Table 1). The N2 genotype had increased oxygen consumption at 25°C (15), and the number of mitochondria in muscle cells was nearly doubled in animals raised at 25°C compared with animals raised at 15°C (16). Thus, increased respiration and mitochondrial biogenesis at 25°C may require higher levels of Q. These observations suggest that the site of action of the Q-less food on longevity is the mitochondria.

Overall, dietary Q is not necessary for survival, but the animals' dietary needs differed, depending on their stage of development. Dietary Q is essential during larval stages for theclk-1–mutant animals to develop into fertile adults (3), but in its absence the arrested larvae survived for weeks with only the maternal contribution of Q. In adulthood, the postmitotic animals of all genotypes tested survived with or without exogenous Q. A Q-less diet may trigger an adaptive response with physiological and biochemical changes that benefit longevity. Our data suggest a model in which dietary Q and insulin-like signaling are parallel pathways that converge to influence the mitochondrial production of reactive oxygen species (ROS) (Fig. 3). Both dietary Q and Q produced by de novo biosynthesis would be expected to function within the mitochondria and have a direct effect on respiratory-chain efficiency (Fig. 3, A to D). Q is also an essential cofactor for the H+ pumping of each of the UCP1, UCP2, and UCP3 uncoupling proteins (2), which may play a role in controlling ROS production by mitochondria (17). In our model, decreased Q levels result in reduced generation of oxidative damage, which is predicted to increase life-span. The changes downstream of the daf-2 insulin-like receptor are predicted to control respiration by a variety of indirect mechanisms (Fig. 3D). Intermediary metabolism is altered in dauer larvae and daf-2mutant adults (18–20) such that limitation of oxidizable substrates would control respiration rates. DAF-2 signaling would regulate transcription of genes involved in respiratory or fermentative metabolism. Finally, the daf-2/age-1 pathway mutant longevity is associated with increased ROS scavenging ability in the long-lived mutants (20–22). Although mitochondria are a common focal point for the Q-less food and daf-2 mutant longevity mechanism proposed, there are metabolic alterations unique to each condition. For example, the clk-1 mutants have higher levels of RQ9 (3), and the increase in RQ9 would be expected to shift metabolism toward anaerobic fermentation, which would be expected to decrease ROS. Thus, the combination of reduced generation and increased scavenging mechanisms are predicted to result in a substantial decrease in the total cellular ROS and thereby allow for an extended life-span.

Figure 3

A model relating the withdrawal of dietary Q8 to life-span extension in three C. elegans genotypes. Q within the mitochondria is derived from either de novo synthesis (Q9) or from the diet (Q8). The nematode respiratory chain (complexes I, II, III, and IV) is shown. Red arrows depict aerobic respiration (electron transport to oxygen). The associated production of ROS is designated by red bolts, and the amount of ROS is represented by the number and size of the bolts. Blue arrows depict nematode anaerobic respiration with RQ9, an amino quinone that functions in complex I; instead of carrying electrons to complex III, RQ9H2 is oxidized by fumarate reductase, providing a bypass of O2 as a terminal electron acceptor. (A) Under standard growth conditions, wild-type animals contain Q9, RQ9, and Q8(3). (B) N2 animals transferred to a Q-less diet produce Q9 and RQ9. When the animals are transferred to the Q-less diet, withdrawal of Q8 from the diet is proposed to lead to decreased ROS production. (C)clk-1 mutants transferred to the Q-less diet as L4 larvae and maintained on this diet. The quinones present probably would consist only of RQ9 and DMQ9. Based on functional studies of DMQ8 inE. coli (25), it seems likely that DMQ9 may retain partial function in nematode complex I but is inactive in complex II. Withdrawal of Q8 from the diet is proposed to lead to decreased ROS production. (D) In addition to the mitochondrion, the insulin-like signal transduction pathway is depicted on the right side of the panel (26). The dashed lines signify that the daf-2 mutations cause reduced signaling, rather than the absence of signaling. The T-bar represents inhibition of mitochondrial respiratory activity after an as yet unknown number of steps. In this situation, the mutation is proposed to lead to increased ROS scavenging and slightly decreased ROS production relative to that of the wild type [in (A)]. Withdrawal of Q8 from the diet is proposed to lead to decreased ROS production, and this is additive to the increased ROS scavenging and slightly decreased ROS production that result from mutation of thedaf-2 gene.

  • * To whom correspondence should be addressed. E-mail: larsen{at}


View Abstract

Stay Connected to Science

Navigate This Article