Caenorhabditis elegans p53: Role in Apoptosis, Meiosis, and Stress Resistance

See allHide authors and affiliations

Science  19 Oct 2001:
Vol. 294, Issue 5542, pp. 591-595
DOI: 10.1126/science.1065486


We have identified a homolog of the mammalian p53 tumor suppressor protein in the nematode Caenorhabditis elegans that is expressed ubiquitously in embryos. The gene encoding this protein,cep-1, promotes DNA damage–induced apoptosis and is required for normal meiotic chromosome segregation in the germ line. Moreover, although somatic apoptosis is unaffected, cep-1mutants show hypersensitivity to hypoxia-induced lethality and decreased longevity in response to starvation-induced stress. Overexpression of CEP-1 promotes widespread caspase-independent cell death, demonstrating the critical importance of regulating p53 function at appropriate levels. These findings show that C. elegansp53 mediates multiple stress responses in the soma, and mediates apoptosis and meiotic chromosome segregation in the germ line.

The p53 tumor suppressor is among the most frequently mutated genes in human cancer and plays a critical role in maintaining genomic stability by regulating cell cycle progression and apoptosis in response to DNA damage (1, 2). Analysis of the mechanisms through which p53 integrates the cellular response to stress and damage in vivo has been limited by the absence of a genetic system. Recently, a p53 homolog was shown to participate in apoptosis induced by genotoxic stress inDrosophila (3–5) on the basis of forced expression of dominant negative forms; however, the organism-wide role of the gene could not be assessed in these experiments.

Standard searches of the genomic sequence suggested that C. elegans does not have a p53-like gene (6). However, using additional algorithms, we identified a C. elegans gene encoding a protein with signature sequences common to the p53 family, including the residues most frequently mutated in human cancers (7). The cDNA sequence of this gene, cep-1 (denoting C. elegans p53-like–1), predicts a 429–amino acid protein that is similar to the human protein in the NH2-terminal transactivation domain and the highly conserved DNA binding domains (Fig. 1). CEP-1 appears to be the only p53 family member encoded in the C. elegans genome, which suggests that p53 paralogs (including p63 and p73) may have evolved from a single ancestor related to CEP-1.

Figure 1

Conservation of transactivation and DNA binding domains in C. elegans CEP-1. (A) Low-resolution three-dimensional model of CEP-1 DNA binding domain (residues 22 to 197) created with the program Modeler/Insight II 98.0 (33). The coordinates of residues 108 to 298 from the crystal structure of the human p53 DNA binding domain were used as the template (34). Conserved Arg residues that make contact with the consensus DNA binding site and that are the most frequently mutated residues in human cancer are shown in red. Amino acids in yellow represent conserved Cys and His residues that coordinate a Zn ion. Portions of the structure shown in magenta are the β strands of the core domain. The green rod indicates the H2 helix that makes contacts with the DNA. (B) Alignment of conserved domains in p53 family members. Single-letter abbreviations for 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. Black boxes indicate amino acids that are identical in at least four of the sequences; gray boxes indicate conservative substitutions. Several residues in the NH2-terminal transactivation domain (domain I) are conserved in CEP-1, including Leu38 and Trp39, which are necessary for transcriptional activation and for the physical interaction of Mdm-2 with human p53 (35, 36). The region of highest conservation lies in the DNA binding domain (domains II to V), where several amino acids have been shown to contact the major and minor grooves of the p53 binding site in the DNA-p53 cocrystal (34). These include four of the five most frequently mutated Arg residues in human cancer (asterisks), as well as Cys and His residues (carets) that make critical contacts with DNA in the three-dimensional structure of human p53. The fifth cancer “hot spot” Arg is conservatively substituted with a Lys in CEP-1 (diamond). The CEP-1 sequence corresponds to F52B5.5 reported by the C. elegans Sequencing Consortium (GenBank accession number CAA99857).

To assess the in vivo function of cep-1, we isolated a chromosomal rearrangement, cep-1(w40) (8). This mutant strain contains an intact copy of cep-1 at its normal genomic location; the cep-1(w40) mutant gene, which encodes a truncated protein lacking the DNA binding domain, is translocated elsewhere in the genome. Although they exhibit impenetrant (∼2%) embryonic lethality, cep-1(w40) mutants are generally viable and fertile. Moreover, depleting cep-1 function by RNA interference (RNAi) (9) similarly leads to impenetrant embryonic lethality (Table 1). It is likely that RNAi results in a strong loss-of-function phenotype, as it eliminates detectable expression of a CEP-1::GFP (green fluorescent protein) reporter (10). We found that bothcep-1(w40) and cep-1(RNAi) embryos undergo a normal pattern of somatic apoptosis, suggesting that CEP-1 is not required for developmental programmed cell death in the soma (10).

Table 1

Elimination of cep-1 function causes meiotic X chromosome nondisjunction.

View this table:

Unlike somatic cells, which have a fixed cell division program, germ-line nuclei in C. elegans undergo indeterminate rounds of division and are subject to checkpoint control and apoptosis in response to genotoxic stresses (11); they also undergo developmentally programmed “physiological” cell death, which appears to be regulated by distinct signaling pathways upstream of the core apoptotic machinery (12). DNA damage activates germ cell apoptosis through a conserved checkpoint pathway that includes therad-5 and mrt-2 genes and the gene altered by theop241 mutation; however, none of these genes is required for physiological germ cell death (11). Because p53 coordinates cellular responses to DNA damage, we hypothesized that cep-1might regulate apoptosis in the germ line in response to genotoxic stress. Indeed, cep-1(w40) hermaphrodites are resistant to ionizing radiation (IR)–induced apoptosis of germ cells (Fig. 2), and cep-1(RNAi) phenocopies this effect ofw40 (Fig. 2E). This block in activation of the germ-line cell death program may be general to DNA damage becausecep-1(w40) mutants, like rad-5, mrt-2, and op241 mutants (11), also fail to undergo germ cell death induced by the DNA modifying compoundN-ethyl-N-nitrosourea (10).

Figure 2

Requirement of cep-1 for normal activation of germ cell apoptosis in response to DNA damage. Shown are wild-type (A and B) and cep-1(w40)adults (C and D) observed by differential interference contrast (DIC) microscopy 12 hours after the L4 stage, either without radiation [(A) and (C)] or after exposure to 60-Gy IR [(B) and (D)]. Arrowheads point to germ cell corpses in a single focal plane. (E) Quantification of germ cell corpses with increasing doses of IR in wild-type (•), cep-1(w40) (▴), and cep-1(RNAi) adults (□). (F) Dominance ofcep-1(w40) allele in suppressing DNA damage–induced germ cell apoptosis. Data are shown for wild type (solid bars),cep-1(w40)/+ heterozygotes (hatched bars), andcep-1(w40) homozygotes (open bars) in the absence versus presence of 120-Gy IR. L4-stage hermaphrodites were irradiated with gamma rays from a 137Cs source, and after 24 hours the number of apoptotic germ cells per gonad arm was determined in 10 to 15 animals. Error bars are SEM.

Our observations suggest that the truncated CEP-1(w40) protein interferes with the proapoptotic activity of wild-type CEP-1. Both a heterozygous w40 mutation and overexpression of thecep-1(w40) gene from a heat shock promoter in a wild-type background confer resistance to IR-induced germ cell apoptosis, confirming that w40 dominantly attenuates wild-typecep-1 function (Fig. 2F) (10).

Unlike rad-5, mrt-2, and op241mutants, which are defective in both germ cell apoptosis and cell cycle checkpoint arrest induced by DNA damage, cep-1(w40) andcep-1(RNAi) germ cells undergo a transient cell cycle arrest in response to IR that is indistinguishable from that of the wild type (10). Furthermore, ectopic expression of CEP-1 in early embryos fails to cause cell division arrest. This ability to activate apoptosis but not arrest the cell cycle is a property shared by Drosophila p53, but not vertebrate homologs (3, 4), possibly revealing a primordial role for p53 proteins in apoptosis specifically.

Analysis of animals depleted for cep-1 function also uncovered a meiotic role in the absence of genotoxic stress. Nondisjunction of the X chromosome at meiosis I in the hermaphrodite germ line generates nullo-X gametes, leading to XO male progeny (13). We found that depletion of cep-1 function by RNAi causes an increase in production of males (the Him phenotype, for high incidence of males) under normal growth conditions (Table 1). This effect was observed uniformly among broods of individual hermaphrodites, which implies that cep-1 is required for chromosome segregation during meiosis rather than during the mitoses preceding meiosis. Mitotic proliferation of nuclei missing an X chromosome would be expected to produce much more variable broods, some with very high numbers of males. The cep-1(w40) mutant does not show a Him phenotype, which suggests that the truncated protein does not interfere with the meiotic chromosome segregation activity of CEP-1.

The low frequency of embryonic lethality in cep-1 mutants (Table 1) might result from autosomal meiotic nondisjunction or could reflect an essential function during normal embryogenesis. Consistent with the latter notion, we found that zygotic expression of a CEP-1::GFP fusion reporter is first detected at the ∼50-cell stage and appears to be ubiquitous throughout embryonic development (14) (Fig. 3, A to D). Near the end of embryogenesis, GFP fluorescence decreases; after hatching, expression is restricted to a subset of pharynx cells, becoming concentrated in nucleoli (Fig. 3E).

Figure 3

Expression and requirement ofcep-1 in somatic cells. (A to E) Zygotic expression pattern of a CEP-1::GFP fusion reporter in embryos and larvae. Shown are DIC (A and C) and fluorescence (B and D) images of embryos at ∼50-cell (A and B) and pretzel (C and D) stages. Similar expression patterns were observed in six independent lines (10). Scale bar, ∼10 μm. (E) Overlay of GFP and DIC images of CEP-1 expression in pharynx after hatching. Anterior is to the right. Arrows point to nucleolar localization of CEP-1::GFP in anterior m2 muscle cells and other pharyngeal muscle and neurons of an L3-stage hermaphrodite. (F) Lethality of wild-type (solid bars) and cep-1(w40) embryos (hatched bars) under normoxic (21% O2) and hypoxic (0.5% O2) conditions. Early embryos were placed in chambers maintained with a constant atmosphere at the indicated oxygen concentration, as measured with a Systech oxygen analyzer. Lethality (percent ± SEM) was scored by quantifying the number of surviving adults arising from a known number of embryos. (G) Effect of prolonged L1 starvation on survival to adulthood of cep-1(w40) (▪) and wild-type larvae (•). Embryos were collected from gravid adults by hypochlorite treatment and hatched in M9 buffer with cholesterol (10 μg/ml) at 20°C in the absence of food. Aliquots of arrested L1 larvae were taken every 72 hours and grown on NGM plates with OP50 bacteria. The fraction surviving to adulthood was determined after 3 days of growth at 20°C. We observed a slight increase in the number of wild-type surviving adults between 6 and 9 days; this likely reflects sticking of some larvae to the culture tube at earlier time points. (H) Quantification of apoptotic death throughout embryonic stages after overexpression of wild-type CEP-1 (hatched bars) compared with overexpression of CEP-1(w40) (solid bars) by heat shock. Embryos between the 50- and 100-cell stage were collected from gravid adults and heat-shocked at 34°C for 15 min; cell corpses were quantified as the embryos developed. Error bars are SEM.

Although little is known about the role of p53 in embryogenesis, knockout mice have revealed a role in normal development (15). p53 is also highly expressed embryonically in mice and frogs; however, its precise role during embryogenesis remains unclear (16–18). The high levels of ubiquitous CEP-1 expression in C. elegans might serve a protective function during embryogenesis, when cell division is rapid and replication errors are likely to occur at a higher frequency. However,cep-1(w40) embryos and larvae are not resistant to IR, the intensity and pattern of CEP-1::GFP expression does not change in response to this treatment, and the pattern of apoptosis incep-1(w40) or cep-1(RNAi) embryos is indistinguishable from that of the wild type (10). Thus, the proapoptotic function of CEP-1 may be restricted to germ-line cells. Because somatic cells in C. elegans cannot generally be replaced if damaged, and arise by a determinate number of cell divisions (and hence are less likely to become tumorous), damage-induced apoptosis in the soma could be detrimental to the animal. In contrast, the germ line contains an excess of germ cells that are not used in self-fertilizing hermaphrodites, and damaged germ cells that are not eliminated could result in defective progeny, making it desirable to eliminate these expendable cells.

Because the DNA damage checkpoint function of CEP-1 is apparently restricted to the germ line, we reasoned that somatic CEP-1 might instead activate a response to other stresses. In vertebrates, p53 is activated by diverse stress signals, including hypoxia, which leads to stabilization of the protein (19, 20). As a soil-dwelling nematode, C. elegans is likely to encounter hypoxic environments frequently. Indeed, we found that cep-1(w40)mutants are hypersensitive to the lethal effects of hypoxia (Fig. 3F).

Under conditions of starvation stress, C. elegansfirst-stage (L1) larvae undergo developmental arrest until food is available. We found that the life-span of cep-1(w40) larvae is reduced relative to the wild type when they were starved at the L1 stage (Fig. 3G). Wild-type survival was reduced by 50% after ∼14 days, whereas survival of cep-1(w40) larvae was reduced by the same magnitude after only ∼9 days (Fig. 3G). In contrast, we found that the life-span of mutant animals during normal growth was unaffected (10). The effect of starvation- and hypoxia-induced stress on cep-1 mutants suggests that CEP-1 can modulate responses to both genotoxic stress in the germ line and environmental stress in the soma.

To address the importance of maintaining proper CEP-1 levels during development, we overexpressed CEP-1 from a heat shock–inducible promoter in 50- to 100-cell-stage embryos (21). The resultant embryos often arrested before hatching and showed severe morphological abnormalities. These embryos did not undergo cell cycle arrest, but they showed a significant increase in the number of cell corpses that accumulated throughout embryogenesis; some terminally arrested embryos contained as many as 40 cell corpses (Fig. 3H) at a time when wild-type animals contain virtually none. No apoptotic corpses were observed when CEP-1 was overexpressed in a mutant lacking CED-3 caspase function (10), which is required for all developmentally programmed cell deaths (22). CEP-1–overexpressing embryos that underwent apparently normal development, and that did not show significantly elevated numbers of cell corpses, nevertheless invariably succumbed, arresting before hatching or as L1 larvae with widespread signs of necrosis. Indeed, overexpression of CEP-1 at all larval stages and during adulthood also caused penetrant lethality and widespread necrotic cell death, independent of CED-3 caspase function. All larvae overexpressing the protein became uncoordinated within 8 hours after induction ofcep-1 overexpression and eventually degenerated.

The lethality of overexpressed CEP-1 appears to be a specific effect, as it requires an intact DNA binding domain; overexpression of the truncated cep-1(w40) allele resulted in virtually no effect on viability. Moreover, we found that expression of human p53 results in similar degenerative phenotypes in C. elegans embryos and larvae (10), which suggests that human p53 and nematode CEP-1 can perform similar biochemical functions. The lethality of overexpressed cep-1 does not appear to result from activation of the core apoptotic machinery, because mutations inced-3 or ced-4 (22) did not block these effects (10). However, dying animals contained large numbers of nuclei that stained positive for acridine orange, generally regarded as a marker of apoptosis (23). Therefore, high levels of CEP-1 may override the requirement for the CED-3 caspase and activate a caspase-independent cell death program, perhaps analogous to the caspase-independent apoptosis observed recently in other systems, which is revealed when caspase function is blocked in cells otherwise programmed to die (24).

We find that C. elegans p53 functions both during normal development (e.g., to ensure proper meiotic chromosome segregation) and under conditions of cellular and genotoxic stress (e.g., in response to DNA damage, hypoxia, or starvation). Although it is expressed ubiquitously in embryos, cep-1 must be carefully regulated because elevated levels of CEP-1 protein are invariably lethal. It should now be possible to use C. elegans as a genetic system to screen for modifiers of the cep-1 mutant phenotype, allowing a comprehensive dissection of the pathways through which p53-like proteins function to mediate stress response, to activate germ-line apoptosis, and to regulate meiotic chromosome segregation.

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


View Abstract

Navigate This Article