Structural and Functional Conservation of the Caenorhabditis elegans Timing Gene clk-1

See allHide authors and affiliations

Science  14 Feb 1997:
Vol. 275, Issue 5302, pp. 980-983
DOI: 10.1126/science.275.5302.980


Mutations in the Caenorhabditis elegans gene clk-1 affect biological timing and extend longevity. The gene clk-1 was identified, and the cloned gene complemented the clk-1 phenotypes and restored normal longevity. The CLK-1 protein was found to be conserved among eukaryotes, including humans, and structurally similar to the yeast metabolic regulator Cat5p (also called Coq7p). These proteins contain a tandem duplication of a core 82-residue domain. clk-1 complemented the phenotype of cat5/coq7 null mutants, demonstrating that clk-1 and CAT5/COQ7 share biochemical function and that clk-1 acts at the level of cellular physiology.

The activity of clk-1 in the nematode Caenorhabditis elegans controls the rate of the worms' development, the pace of their behaviors, and when they die (1). In clk-1 mutants, the timing of a wide range of physiological processes is deregulated (1). This leads to an average lengthening of such diverse processes as the worms' early cell cycles, their embryonic and postembryonic development, and the period of rhythmic adult behaviors, such as swimming, pharyngeal pumping, and defecation. clk-1 mutants also have an extended life-span. This pleiotropic alteration of developmental and behavioral timing is also exhibited by worms with mutations in any one of the genes clk-2, clk-3, or gro-1 (1, 2). Mutations in these four genes interact genetically to affect developmental rate and longevity (3).

Mutations of clk-1 exhibit a maternal effect: homozygous mutant (clk-1/clk-1) progeny from a heterozygous hermaphrodite (clk-1/+) are phenotypically wild type; only homozygous mutants from a homozygous mother exhibit a Clock (Clk) phenotype. The maternal rescue not only influences early events, such as embryonic development, but it extends to adult phenotypes, such as defecation and longevity (1). This and other evidence [(1), reviewed in (4)] suggests the existence of a pervasive timing mechanism whose intrinsic rate can be set early in development and that influences diverse timed processes throughout the worm's life.

The clk-1 gene lies on linkage group III, between dpy-17 and lon-1 (Fig. 1). Overlapping cosmids from the candidate region were assayed for their ability to rescue the mutant phenotype by microinjection into clk-1 mutants. The cosmid ZC400 rescued the mutant phenotype of both strong (qm30) and weaker (e2519) clk-1 alleles and fully recapitulated the maternal effect. Over a number of generations, ZC400-containing extrachromosomal arrays lost their ability to rescue fully the slow-growth phenotype of clk-1 mutants, but the rescue of slow defecation persisted. This loss of rescue of the developmental phenotype probably reflects a ubiquitous phenomenon in C. elegans, in which transgenic arrays undergo transcriptional silencing in the germ line and in early embryos, possibly because of their complex repeated structure (5). If this is so, it implies that later zygotic expression of clk-1(+) is sufficient to rescue adult behavioral defects, but maternal or early zygotic expression is needed to rescue slow development.

Fig. 1.

Cloning of clk-1. The top line shows the genetic map in the clk-1 region (7). clk-1 had been previously mapped to this region (1, 2), but we have refined its location (25). Some cosmids tested for rescuing activity (5) are shown. Stable rescue that persisted for many generations was obtained only with ZC400. A plus sign indicates rescue of all phenotypes; a plus-minus sign indicates rescue of just the defecation phenotype; and a minus sign indicates no rescue. pRA41 is a derivative of ZC400 with an internal Sac I deletion. The insert in pRA41 contains three predicted genes, ZC395.10, ZC395.3, and ZC395.2 (7). A deletion in ZC395.2 eliminates clk-1 rescuing activity, but pRA40 rescues clk-1, indicating that ZC395.2 is the clk-1 gene. S, Sac I; X, Xba I; E, Eco RI. The nematode expressed sequence tag CEESX93F (7) matches the 3′ end of toc-1. We resequenced clk-1 from the mutant alleles (26). In allele e2519, a G→A transition introduces a new Hind III site and causes a Glu→Lys (E→K) missense alteration in an absolutely conserved residue (Fig. 3). Allele qm30 results in a 590-base pair deletion, starting 12 nucleotides 3′ of the lesion in e2519 and encompassing the entire last exon. Allele qm51 alters the absolutely conserved terminal G in the intron 2 splice acceptor.

Injection of ZC400 subclones localized clk-1 to a 1.9-kb Eco RI fragment. Extrachromosomal arrays containing this fragment restored developmental and behavioral rates to wild-type speed for at least one generation. Again, stable rescue over many generations was only obtained for the behavioral phenotypes such as defecation (6). This fragment is predicted to contain a single gene, ZC395.2 (7), which is altered in three clk-1 alleles, thereby confirming the identity of this gene as clk-1 (Fig. 1). The weaker clk-1 allele is a missense mutation; the stronger ones involve the disruption of entire exons.

We previously demonstrated that all phenotypes of clk-1 mutants can be fully maternally rescued, that all alleles exhibit the same pattern of phenotypes, and that all alleles fail to complement each other for all phenotypes (1, 2). Given this, the molecular evidence presented here unequivocally establishes that mutations in the clk-1 gene are responsible for all of the phenotypes seen in clk-1 mutant worms.

The clk-1 gene lies just downstream of the predicted gene, ZC395.3 (Fig. 1). Using reverse transcription polymerase chain reaction (RT-PCR) (8), we established the 5′ and 3′ ends of both genes and their splicing patterns. We found that clk-1 is exclusively trans-spliced to the splice leader SL2 at its 5′ end, and the upstream gene is trans-spliced to SL1, in both cases immediately upstream of the initiator AUG codon. The 3′ end of the upstream gene lies 105 base pairs from the 5′ end of clk-1. The pattern of trans-splicing and the intergenic distance are typical of genes organized into operons in C. elegans (9), suggesting that the two genes share a promoter 5′ of the upstream gene. The introns in clk-1 were correctly predicted, but the real product of the upstream gene (Fig. 1) lacks the first predicted exon. This gene potentially encodes a protein that has similarity to a family of divalent metal ion transporters, so we have named it toc-1, for transporter-like protein in an operon with clk-1. There is no obvious functional relation between clk-1 (see below) and toc-1.

To demonstrate that the lesion in clk-1(e2519) was responsible for the mutant worms' extended longevity, rather than it resulting from a difference in genetic background (10), we assayed the longevity of fully rescued e2519 mutant worms carrying the ZC400-containing extrachromosomal array qmEx109, as well as those carrying qmEx96, an extrachromosomal array containing the clk-1 1.9-kb Eco RI fragment. As adults, e2519; qmEx96 worms showed full rescue of their defecation cycle, even though their development was slow. We found that the presence of either qmEx96 or qmEx109 restored a wild-type life-span to e2519 mutant worms. The life-spans of both e2519; qmEx96 and e2519; qmEx109 worms were indistinguishable from that of N2 worms (Fig. 2).

Fig. 2.

The extension of life-span exhibited by clk-1(e2519) is rescued by the presence of extrachromosomal arrays containing the wild-type clk-1 gene. The graph shows the percentage of worms alive on a given day after being laid as eggs during a 2.5-hour period on day 0 for N2 (□), clk-1(e2519) (•), and e2519; qmEx109 (▴). For e2519; qmEx96 (▵) a 6-hour limited hatching was used. The mean life-spans, with standard errors, are 20.4 ± 0.8, 28.1 ± 1.4, 20.2 ± 0.9, and 20.4 ± 0.7 days, respectively. The worms were maintained at 18°C throughout and their longevity scored as described (3). The sample size was 50 worms of each genotype, except for e2519; qmEx109 which was 48.

The clk-1 gene is predicted to encode a 187-residue protein (CLK-1) that is similar to the product of the Saccharomyces cerevisiae gene CAT5/COQ7 (Cat5p/Coq7p) (11, 12) (Fig. 3A). A rat homolog of Cat5p/Coq7p has also been described (13). Using the sequence of the rat gene, we identified and partially sequenced murine and human homologs of clk-1 (14). The three full-length proteins are 33% identical over 177 residues, although their NH2-termini show no similarity, either in length or composition (Fig. 3A). Over the available predicted sequences of 43 and 126 amino acids, the human and mouse proteins are 93 and 97% identical to the rat protein, respectively (Fig. 3). These five proteins are unrelated to any other known sequence, and there are few indications as to their biochemical function (15, 16). The protein sequences can each be split and aligned to reveal the presence of an 82-residue tandemly repeated core domain, which we call the TRC domain, for tandemly repeated in CLK-1 (or Cat5p/Coq7p or rat COQ7) (Fig. 3B). For all the repeats, residues are absolutely conserved at eight positions, and at an additional 12 positions all the residues are similar. For each protein, its two TRC domains are juxtaposed without any linking sequence. For each domain, there appears to be only a single point at which insertions are tolerated, flanked by regions predicted to be helical (Fig. 3B). Within these helical regions (residues 34 through 56 and 116 through 144 for CLK-1), the spacing of conserved hydrophobic residues is suggestive of an interface for protein-protein interaction, such as a surface for dimerization (17) (Fig. 3B). The proteins' two-domain primary structure likely reflects two equivalent domains at the level of tertiary structure (17). Given the functional complementation of yeast cat5/coq7 mutants by the rat homolog (13), the level of identity between rat and human sequences strongly supports the notion of functional conservation of clk-1 from yeast to human.

Fig. 3.

CLK-1 is highly conserved between nematodes, yeast, rodents, and humans. (A) Alignment of the nematode CLK-1, rat COQ7, and yeast Cat5p/Coq7p sequences, together with the partial sequence of murine and human CLK-1 homologs (14). Over the length of the rat protein, the identity between CLK-1 and its yeast and rat homologs is 42 and 53%, respectively. Introduced gaps are marked by dashes. Reduction-of-function alleles are indicated by arrows in (A) and are boxed in (B); G→D in coq7-1 (12) and E→K in e2519. The rat sequence (GenBank accession number U46149) appears to contain sequencing errors in the vicinity of residues 82 through 84 and 151 through 154 (marked by dots). It is presumably a coincidence that the tripeptide motif AGE (residues 28 through 30 in CLK-1) is present in all three proteins. (B) Duplication within the CLK-1 sequence and its homologs. Each of the sequences shown in (A) can be split and aligned to reveal the presence of a tandemly repeated TRC domain. There is a single site of insertions for both NH2- and COOH-terminal domains; these have been removed for this alignment, as marked by the small black dots. Those residues identical in more than half of the domains are shown in black lettering; those that are similar in more than half are shown in dark gray. A bar is drawn under those residues predicted to be in a helical conformation [(16), with input being the alignment shown here]. Within the extended helical regions, positions where there is absolute conservation of a hydrophobic residue are marked by the symbol ϕ. Single-letter abbreviations for the amino acid residues are as follows: A, Ala; C, Cys; D, Asp; E, Glu; F, Phe; G, Gly; H, His; I, Ile; K, Lys; L, Leu; M, Met; N, Asn; P, Pro; Q, Gln; R, Arg; S, Ser; T, Thr; V, Val; W, Trp; and Y, Tyr.

CAT5/COQ7 is required for the derepression of PCK1, which encodes the gluconeogenic enzyme phosphoenolpyruvate carboxykinase, that accompanies the transfer of yeast from glucose to a nonfermentable carbon source such as glycerol or ethanol (11). This derepression is mediated by transcriptional activation (18), and CAT5/COQ7 is required for the formation of a specific transcriptional activation complex (11). CAT5/COQ7 also appears to be involved in the control of expression of other enzymes of gluconeogenesis and those of respiration and the glyoxylate cycle (11). But its role in all these processes appears to be indirect, because Cat5p/Coq7p is not part of the transcriptional complex (11). In addition, the expression of CAT5/COQ7 appears to be highly regulated, by glucose and by itself (11).

cat5/coq7 mutants do not synthesize ubiquinone (coenzyme Q), a lipid-soluble two-electron carrier, which is essential for respiration and consequently for nonfermentative growth (12). CAT5/COQ7 appears to control ubiquinone biosynthesis at two or more steps, although its mode of action is obscure (11, 12). The pleiotropic effects of mutation of CAT5/COQ7 have led to the proposal that there is a co-regulation of respiratory chain components, the biogenesis of mitochondria, and gluconeogenesis, with CAT5/COQ7 being a likely link connecting glucose derepression with respiration (11). Thus, CAT5/COQ7 appears to be important in the regulation of multiple parallel processes of metabolism, consistent with our view that clk-1 regulates many disparate physiological and metabolic processes in C. elegans.

To test whether the structural similarity extends to functional equivalence, we constructed a cat5/coq7 deletion yeast strain (Δcat5/coq7), which, as expected (11, 12), did not grow on glycerol (19). Introduction of a multicopy plasmid containing the C. elegans clk-1 coding sequence, within an expression cassette with the constitutive promoter and 3′ sequence of the ADH1 gene, conferred to the Δcat5/coq7 strain the ability to grow on glycerol (20). This functional complementation of the Δcat5/coq7 null mutation by clk-1 is consistent with the functional complementation of the yeast mutant by the rat homolog (13) and is indicative of a common biochemical function for these three genes. In spite of their common biochemical function and roles in regulatory mechanisms, one must be cautious in attempting to understand the physiological defects seen in clk-1 worms in terms of the phenotypic defects of cat5/coq7 mutant strains because yeast have some highly specialized systems of metabolic regulation not seen in most other eukaryotes (21). Nevertheless, the interspecific functional complementation suggests that a central mechanism of metabolic coordination, which regulates many distinct downstream regulators, is conserved in all eukaryotes, including humans.

The action of clk-1 is highly pleiotropic; it acts at the level of single-celled eggs and subsequently apparently affects all tissues of the worm at all stages of life (1). The clk-1 homolog CAT5/COQ7 affects cellular physiology in yeast. These observations suggest that clk-1 increases animal longevity by slowing down general cellular aging. It remains possible, however, that the longer life of clk-1 mutants reflects a decreased rate of deterioration of the cells of a particular organ critically required for life, rather than a reflection of the health of the whole of the animal's cells.

In conclusion, clk-1 is structurally and functionally homologous to a yeast central metabolic regulator. This is consistent with our previous speculation that the long life of clk-1 mutants might be a consequence of slower cellular metabolism, with an attendant reduction in the rate of production of detrimental by-products (3). Our findings also lend support to the idea that multicellular organisms age because their cells age (22).


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
View Abstract

Stay Connected to Science

Navigate This Article