The zinc-finger protein SEA-2 regulates larval developmental timing and adult lifespan in C. elegans

Developmental timing mechanisms are integrated with various signaling pathways to achieve the synchrony and succession of stage-specific programs during animal development. In C. elegans, a network of heterochronic genes has been identified that control developmental timing of diverse post-embryonic cell lineages, best characterized by the development of a row of lateral hypodermal seam cells, which undergo stage-specific developmental programs during each of the four larval stages and terminally differentiate at the late L4 larval stage (Ambros, 2000; Rougvie, 2005; Moss, 2007). Mutations in heterochronic genes cause skipping or reiteration of stage-specific programs, resulting in premature or delayed terminal differentiation of seam cells. Adult animals, the somatic cells of which are post-mitotic, undergo a progressive decline in pharyngeal pumping and body movement, and also experience cell and tissue deterioration. Several mechanisms, including caloric restriction, the insulin/insulin-like growth factor 1 (IGF-1) endocrine system and the steroid hormone system, regulate aging in C. elegans (Guarente and Kenyon, 2000; Kenyon, 2005; Antebi, 2007). For example, reduced activity of daf-2, which encodes an insulin/IGF-1 receptor, extends the lifespan (Kenyon, 2005). The steroid hormone pathway, acting via the nuclear receptor DAF-12, also impacts the succession of developmental events in larvae (Fielenbach and Antebi, 2008). Two components of the heterochronic circuit, miRNA lin-4 and its target lin-14 (which encodes a nuclear transcription factor), also have a modest effect on adult lifespan (Boehm and Slack, 2005). lin-4/lin-14 may influence lifespan by regulating metabolic outputs, and the insulin-like gene ins-33 is a direct target of LIN-14 (Hristova et al., 2005). Nevertheless, it remains largely unknown whether genes controlling larval developmental timing also function in the adult and whether signals that modulate aging, such as insulin/IGF-1, also act within the heterochronic circuit.

The following strains were used in this study: LGI, lin-28(n719), daf-16(mu86); LGII, lin-42(n1089), sea-1(gk799), sea-2(bp283), sea-2(tm4355), lin-4(e912), lin-29(n333); LGIII, daf-2(e1370); LGIV, lin-66(ku423), jcIs1(ajm-1::gfp), zIs356(daf-16::gfp); LGV, lin-46(bp312), lin-46(bp284), wIs51(scm::gfp); and LGX, daf-12(rh257), daf-12(rh61rh411), lin-14(n179), alg-1(gk214), ain-1(bp299). The genetic location for muIs84(sod-3::gfp), bpIs124(dcap-1::rfp) and bpIs145(lin-28::gfp::lin-28 3′UTR(ΔLCE)) was not determined.

lin-46(bp312) contains a glycine to stop codon mutation at amino acid 249. lin-46(bp284) was isolated in a screen for mutants that enhanced the retarded heterochronic phenotype of sea-2(bp283). lin-46(bp284) contains a leucine to phenylalanine mutation at amino acid 44 of LIN-46.

Animals carrying the scm::gfp reporter were mutagenized and F2 progeny were examined for the number of SCM::GFP-positive cells. From ~4000 genomes screened, 10 mutations were isolated that caused altered number of seam cells. The sea-2 mutants are cold sensitive. At 15°C, the retarded heterochronic defects are more severe. ccDf2/sea-2(bp283) mutants had an average of 22.5 seam cells (n=8) at 15°C, similar to 22.7 in sea-2(bp283) mutants. All experiments were performed at 20°C unless otherwise noted.

bp283 was mapped by three factor mapping. From the sqt-2 lin-31 +/+ + bp283 cross, 0 out of 32 Lin no Sqt recombinants carried sea-2(bp283). From the lin-31 clr-1 dpy-10 +/+ + + bp283 cross, 0 out of 14 Lin non Clr recombinants carried bp283. From the dpy-10 vab-19 +/+ + bp283 cross, 17 out of 39 Dpy non Vab recombinants carried bp283. Single nucleotide polymorphism (SNP) mapping further located bp283 between pkp2115 (−6.31) and pkp2051 (−4.92). Fosmids covering this region were used for transformation rescue experiments.

Animals were placed on a 2% agarose pad in 5 μl of M9 and the seam cell division pattern was analyzed by microscopy. After observation, each animal was rescued and placed on a plate to recover. This procedure was repeated every 2 to 4 hours.

The PCR templates used for synthesizing RNA were: sea-2 (K10G6, nt 9550-10438); fox-1 (T07D1, nt 29314-30084); and sex-1 (F44A6, nt 11213-12115). For RNA feeding of hbl-1, synchronized L1 worms were placed on RNAi plates and worms from the next generation were examined.

Reporters for sea-2 were constructed by a PCR fusion based approach. The fused PCR products were derived from two overlapping PCR fragments. One contained the promoter region and the entire ORF of sea-2 (K10G6, nt3055-16856), the other contained gfp and the unc-54 3′UTR or the sea-2 3′UTR. The PCR products were co-injected with pRF4 [rol-6(su1006)] into wild type animals and at least two stable transgenic lines were analyzed. The sea-2::gfp::unc-54 3′UTR reporter displayed the same expression pattern as the reporter containing the sea-2 3′UTR. We inserted a nuclear localization signal sequence (PKKKRK) at the N terminus of SEA-2 to determine whether SEA-2 acts in the nucleus or in the cytoplasm to specify the temporal fates of seam cells. However, the sea-2::NLS::sea-2::gfp transgene was expressed at a much lower level than sea-2::gfp, indicating the addition of the NLS destabilized SEA-2.

lin-28::gfp::lin-28 3′UTR(ΔLCE) contains the promoter and the entire ORF of lin-28 (F02E9, nt 3769-7599), gfp and the lin-28 3′UTR (F02E9, nt 3235-3765) with a deletion of the LCE (F02E9, nt 3424-3438). Animals carrying the lin-28::gfp::lin-28 3′UTR(ΔLCE) extrachromosomal array were γ-ray irradiated and the resulting stable integrated line (bpIs145) was outcrossed two times.

The lifespan assay was performed at 20°C. Animals that had just passed the final larval molt were transferred to new plates every 1-2 days until the end of reproduction and 2-4 days thereafter. Animals were scored as dead when they failed to response to gentle prodding. Worms with exploded vulva, or that had bagged (died from internal hatching) or crawled off the plate were excluded. Three independent assays were tested for each experiment. GraphPad Prism 5 was used for survival curves and statistical analysis.

The heat stress assay was performed at 32°C using 1-day-old adults. Animals were cultured and scored as described for lifespan assays. At least 100 animals were tested for each strain.

Synchronized L1 and L3 animals were collected and total RNA was extracted from about 500 animals using Trizol reagent (Sigma) according to the manufacturer's protocol. Total RNA (2 μg) was reverse transcribed using an Invitrogen Superscript III kit. Quantitative PCR reactions were carried out using a SYBR RT-PCR kit (TaKaRa) and a Mastercycler ep realplex machine (Eppendorf). eft-2 was used as an internal control (Bagga et al., 2005). The level of lin-28 mRNA was normalized to the level of wild type L1 worms, which was set to 1. Error bars indicate the standard deviation (s.d.) of three independent experiments.

Primers used were: lin-28 FW, TCGGAGTCTTGATGAAGGAG; lin-28 RW, GAGACAGCCTTCTTACGACC; eft-2 FW, ATGGTCAACTTCACGGTCGATG; eft-2 RW: GATGGTAATACAACGCTCCTGC.

Gut autofluorescence was photographed using a Zeiss Axioplan 2 imaging system and quantified by AxioVision Rel. 4.6. The same exposure time was used for different strains and also for animals of different ages.

Lysates from synchronized animals were prepared as previously described (Seggerson et al., 2002), and endogenous LIN-28 was detected with diluted anti-LIN-28 serum (1:2000) and HRP-conjugated goat anti-rabbit secondary antibody. Anti-actin monoclonal antibody (A3853, Sigma) was used as a gel loading control.

To determine the level of lin-4 RNA, total RNA isolation and northern analyses were performed as described (Grishok et al., 2001). Total RNA was isolated from synchronized L2 and L3 worms.

The following sequences were used to drive the expression of genomic coding region of sea-2 and unc-54 3′UTR in various tissues: Pceh-16 (C13G5, nt 4341-6883), Pvha-6 (VW02B12L, nt 1023-2624), Pmyo-3 (WRM061aH08, nt 24240-26523) and Prgef-1(F25B3, nt 10721-14265). At least three independent transgenic lines for each construct were examined.

Ten seam cells, aligned on each side of the animal, undergo asymmetric cell division at each of four larval stages (L1 to L4) with only one daughter retaining the seam cell identity. Certain seam cells also undergo one round of symmetric division with both daughter cells maintaining the seam cell fate at the L2 larval stage, increasing seam cell number from 10 at hatching stage to 16 from L2 stage onwards. All seam cells terminally differentiate at the late L4 stage, including cessation of cell division, fusion with neighboring seam cells and synthesis of adult-specific cuticular structures, called alae (Fig. 1A-C,G). In a genetic screen to identify mutants with defective seam cell development, we isolated a mutation, bp283, that increased the number of seam cells in young adults from 16 in wild-type animals to 20 (Table 1, Fig. 1D). Adult-specific alae were not completely formed in bp283 mutant young adults, indicating a delay in terminal differentiation of seam cells (Fig. 1E,F). Analysis of seam cell lineages revealed that the L2 stage-specific proliferative division pattern was reiterated at the L3 larval stage in bp283 mutants and certain seam cells failed to fuse and continued to divide at the L4/adult switch, resulting in an increase in the seam cell number and discontinuous alae (Fig. 1H). The seam cell developmental defects in bp283 animals, as in other classic heterochronic gene mutants, were completely suppressed when animals developed through the alternate dauer larval stage (Table 1). Thus, bp283 causes retarded heterochronic defects.

Using transformation rescue we found that a transgene containing a single gene, previously named sea-2 (see below), rescued the heterochronic defects in bp283 mutants (Fig. 1I). We sequenced cDNAs and found that sea-2 encodes a 1727 amino acid protein. bp283 contains an alanine to aspartate mutation at codon 1267 (see Fig. S1 in the supplementary material). sea-2(tm4355), which deletes amino acids 294 to 590 of SEA-2, showed retarded heterochronic defects in sensitized genetic backgrounds (Table 1). Animals bearing sea-2(bp283) in trans to ccDf2, a deficiency that removes the sea-2 locus, had an average of 20.2 seam cells (n=10), compared with 20.4 in sea-2(bp283) mutants. sea-2(RNAi) also caused retarded heterochronic defects, but did not further elevate the defects in sea-2(bp283) mutants (Table 1). These results indicate that sea-2(bp283) is probably a strong loss-of-function allele.

Bioinformatic analysis revealed that SEA-2 contains four CCHC zinc fingers and one CCHH zinc finger (see Fig. S1 in the supplementary material). The zinc fingers Z1 and Z5 of SEA-2 strongly bound to single stranded (ss) and double stranded (ds) RNA in an EMSA assay (see Fig. S2 in the supplementary material). Transgenes expressing truncated SEA-2 with a deletion of the first zinc finger (amino acid 319-339) or the fifth zinc finger (amino acid 1429-1449) failed to rescue the retarded heterochronic defects in sea-2 mutants (see Table S1 and Fig. S3 in the supplementary material), suggesting that the RNA binding domains are important for SEA-2 function.

We constructed a reporter with gfp inserted at the C terminus of the sea-2-coding region to determine the expression pattern of sea-2. This translational reporter rescued the heterochronic defect in sea-2 mutants (data not shown). sea-2 was strongly expressed in various tissues, including seam cells, intestine cells, pharyngeal muscles and nerve ring neurons (Fig. 1J-N). SEA-2::GFP expression persisted into adulthood. SEA-2::GFP was diffusely localized in both cytoplasm and nucleus (Fig. 1J-M).

To determine whether sea-2 acts cell-autonomously in controlling the stage specific fates of seam cells, we expressed sea-2 using a seam cell-specific ceh-16 promoter (Huang et al., 2009). sea-2 mutants carrying a ceh-16::sea-2 transgene had an average of 16.3 seam cells (see Table S1 in the supplementary material). Expression of sea-2 in the intestine, body wall muscle cells or neurons failed to rescue the increased number of seam cells in sea-2 mutants (see Table S1 and Fig. S3 in the supplementary material).

To further characterize the role of sea-2 in specifying the L2/L3 progression, we examined the genetic interaction between sea-2 and other heterochronic mutants. Other mutants that cause reiteration of the L2 stage fate at the L3 stage and an incomplete defect in the larval/adult switch phenotype include the recessive gain of function (rh257) or null allele (rh61 rh411) of daf-12 and loss of function of lin-46 (the gephyrin homolog), lin-66 (a novel protein), alg-1 (the Argonaute homolog), ain-1 (the GW182 homolog) and let-7 family miRNAs (Moss, 2007). sea-2(bp283) in combination with a mutation in each of these retarded heterochronic genes caused a dramatic increase in seam cell numbers and a much more complete terminal differentiation defect (Table 1, Fig. 2A-D). For example, sea-2; lin-66 double mutants had an average of 90 seam cells at the young adult stage, compared with 20 in sea-2 and 37 in lin-66 single mutants. In sea-2; lin-66 mutants the L2 stage program was reiterated at both the L3 and L4 stages (Fig. 2E). The mutant alleles used are strong loss of function or null. The genetic interactions suggest that sea-2 probably functions in parallel to these genes in specifying the L2/L3 switch.

We further examined the relationship between sea-2 and other retarded heterochronic mutants. The lin-4(e912) null mutation leads to the reiteration of the L1 stage fate at subsequent larval stages. lin-29 functions downstream in the heterochronic pathway in specifying the larval/adult switch (Ambros and Horvitz, 1984). We found that as in lin-4 and lin-29 single mutants, no alae were generated at the young adult stage in sea-2 lin-4 and sea-2 lin-29 double mutants (see Table S2 in the supplementary material). This is consistent with the hypothesis that sea-2 functions upstream of lin-29.

To place sea-2 in the heterochronic pathway, we examined the phenotype of sea-2 mutants combined with precocious heterochronic mutants. lin-28, which encodes a protein with a cold shock domain and a CCHC zinc finger, specifies the L2 program (Moss et al., 1997). In lin-28 mutants, L2-stage events are skipped and the larval/adult switch takes place at the L3 stage (Fig. 3A,B) (Ambros and Horvitz, 1984). The retarded heterochronic defects in sea-2 mutants were completely suppressed by lin-28. lin-28; sea-2 double mutants showed the same seam cell development phenotype as lin-28 single mutants (Table 1; Fig. 3C,D). hbl-1 (the Hunchback homolog) regulates the L2 fate and also the L4/adult switch. Loss of function of hbl-1 causes a precocious heterochronic phenotype and also suppresses the retarded heterochronic defect in let-7 family miRNAs mutants (Abrahante et al., 2003; Lin et al., 2003; Abbott et al., 2005). We found that sea-2(bp283) partially suppressed the precocious heterochronic defect in hbl-1(RNAi)animals (Table 1). The sea-2 mutation did not affect temporal expression of hbl-1 (see Fig. S4 in the supplementary material). sea-2 mutants also partially suppressed other precocious mutants, including lin-14 and lin-42 (see Table S2 in the supplementary material). These genetic analyses suggest that sea-2 functions through lin-28 in specifying the L2/L3 progression.

To understand how sea-2 regulates lin-28 activity, we examined expression of the translational fusion reporter lin-28::gfp::lin-28 3′UTR, which contains the lin-28 coding and regulatory region. This transgene rescues the mutant phenotype of lin-28(n719) animals (Moss et al., 1997). lin-28::gfp is expressed in diverse cell types, including cells in the head, tail, muscles and seam cells. In wild-type animals, lin-28::gfp is expressed in L1 larvae, is detectable but diminished in L2 larvae and is almost undetectable from the L3 stage onwards (Fig. 4A,D; data not shown) (Moss et al., 1997). However, we found that in sea-2 mutants high levels of LIN-28::GFP persisted in the head and tail at the L3 and L4 larval stages (Fig. 4B-D). The reporter also showed expression in seam cells in 87.5% of sea-2 mutant L3 larvae (n=16) (Fig. 4E,F), whereas its expression was not detected in seam cells in wild type L3 larvae (n=15).

We further performed an immunoblot assay to examine the levels of endogenous LIN-28 protein using an anti-LIN-28 antibody. In wild-type animals, LIN-28 protein was present at the L1 larval stage, but was greatly reduced at the L3 larval stage (Fig. 4G) (Seggerson et al., 2002; Morita and Han, 2006). In sea-2 mutants, levels of LIN-28 remained high in L3 larvae (Fig. 4G). Upregulation of LIN-28 in sea-2 mutants was more prominent than that in daf-12 mutants (Fig. 4G). We conclude that mutations in sea-2 cause ectopic expression of lin-28.

We next determined at what level sea-2 regulates lin-28 expression. No upregulation of lin-28 mRNA was observed in L1 and L3 larvae in sea-2 mutants (Fig. 4H); rather, levels of lin-28 transcripts were even lower than wild type (Fig. 4H). This could be because wild-type SEA-2 affects the transcription or stability of lin-28 or because high levels of LIN-28 in sea-2 mutants negatively regulate the lin-28 mRNA levels. To explore whether sea-2 regulates lin-28 expression at the post-transcriptional level via its 3′UTR, we examined expression of the col-10::lacZ::lin-28 3′UTR reporter (pKM50), in which the expression of lacZ is driven by the hypodermal cell-specific promoter col-10 (Fig. 4I) (Morita and Han, 2006). In wild-type animals, expression of this reporter was strong in L1 larvae but absent in adults (Fig. 4I,J and data not shown). High expression persisted, however, in sea-2 mutant adults (Fig. 4I,J). The col-10::lacZ::unc-54 3′UTR reporter, in which the unc-54 3′UTR was used instead of lin-28 3′UTR, was highly expressed in both wild type and sea-2 mutants at the young adult stage (pKM53, Fig. 4J), suggesting that sea-2 represses lin-28 expression through its 3′UTR. To identify the SEA-2 response element, we examined the expression of a series of reporters with deletions of discrete regulatory elements in the lin-28 3′UTR. sea-2(bp283) dramatically increased expression of a reporter with a mutation in the putative DAF-12 response element (pKM63) (Fig. 4J) (Morita and Han, 2006). However, sea-2(bp283) did not increase the expression of a col-10::lacZ::lin-28 transgene lacking the lin-4 binding site (pKM55) (Fig. 4J). A reporter lacking the lin-4 complementary element (LCE), lin-28::gfp::lin-28 3′UTR(ΔLCE), is strongly expressed at late larval stages and also causes retarded heterochronic defects (see Fig. S5 in the supplementary material) (Moss et al., 1997). sea-2(bp283) did not further elevate reporter expression or the retarded heterochronic defects in animals carrying lin-28::gfp::lin-28 3′UTR(ΔLCE) (Table 1 and data not shown). Compared with wild-type animals, levels of lin-4 remained unchanged at the L2 and L3 larval stage in sea-2 mutants (see Fig. S5 in the supplementary material). These results indicate that sea-2 probably acts through the LCE to regulate lin-28 expression at the post-transcriptional level.

sea-2 was previously identified as an autosomal signal element (ASE), loss of function of which suppresses the XX lethality phenotype caused by simultaneous depletion of the X signal elements (XSEs) sex-1 and fox-1 (Meyer, 2005) (P. Nix and B. Meyer, personal communication). ASEs (also known as denominator elements) function with XSEs (also known as numerator elements) to communicate the ratio of X chromosomes to sets of autosomes (X:A signal) to determine sexual fate and to equalize expression of X-linked genes between hermaphrodites (XX) and males (XO) (Meyer, 2005). Thus, we determined whether sea-2 also functions through lin-28 to suppress the lethality of fox-1 sex-1 mutants in XX animals. Simultaneously depleting the activity of lin-28 in sea-2; fox-1 sex-1 hermaphrodites caused XX animals to arrest during embryogenesis (Fig. 4K). lin-28 has no effect on the suppression of the lethality of fox-1 sex-1 by loss of function of another ASE, sea-1 (Powell et al., 2005) (Fig. 4K). Mutations in other heterochronic genes, including lin-14, daf-12 and hbl-1 on the X chromosome, had no effect on the ASE function of sea-2 (see Table S3 in the supplementary material). Loss of function of fox-1 sex-1 still caused embryonic lethality in animals carrying the lin-28::gfp::lin-28 3′UTR(ΔLCE) transgene (Fig. 4K), indicating that lin-28 is necessary but not sufficient to mediate the role of sea-2 in the sex determination and dosage compensation pathways.

We next examined whether sea-2 functions in adult animals. Compared with wild-type animals, sea-2(bp283) mutants had a significant increase in lifespan (Fig. 5A; see Table S4 in the supplementary material). To determine whether sea-2 mutants had a slower aging process, we examined two age-related markers, the accumulation of lipofuscin fluorescence in intestine cells (Garigan et al., 2002) and the accumulation of DCAP-1 (mRNA decapping enzyme)-labeled cytoplasmic processing bodies (P bodies). P bodies contain a variety of ribonucleoproteins and serve as sites for mRNA turnover and storage (Parker and Sheth, 2007). We found that DCAP-1 bodies gradually increased in hypodermal and muscle cells in aged wild-type animals (see Fig. S6 in the supplementary material). The accumulation of DCAP-1 bodies was slightly accelerated in short-lived daf-16 mutants, and greatly decreased in long-lived daf-2 mutants (Fig. 5B-D). sea-2 mutants accumulated DCAP-1 bodies and lipofuscin fluorescence more slowly than wild-type animals (Fig. 5E-K). Many long-lived C. elegans mutants are resistant to heat stress (Kenyon, 2005). After heat shock treatment, sea-2 mutants also survived longer than wild-type animals (Fig. 5L). The decreased rate of age-dependent reporter accumulation and elevated heat stress resistance confirm that loss of function of sea-2 slows the rate of aging and extends the adult lifespan.

DAF-16, a FOXO family transcription factor, is a master regulator of adult lifespan that integrates multiple inputs, including insulin/IGF-1 signaling, increased dosage of sir-2.1 (the C. elegansSIR2 NAD⁺-dependent protein deacetylase homolog) and reduced activity of lin-14 (Boehm and Slack, 2005; Lin et al., 1997; Ogg et al., 1997; Tissenbaum and Guarente, 2001). We thus investigated whether sea-2 modulates lifespan through daf-16. daf-16; sea-2 double mutants had the same lifespan as daf-16 single mutants (Fig. 5A; see Table S4 in the supplementary material). The heat stress resistance of sea-2 mutants was also abolished by the daf-16 mutation (Fig. 5L). Thus, loss of activity of daf-16 suppressed the longevity of sea-2 mutants. Compared with sea-2 and daf-2 single mutants, the lifespan was further extended in sea-2; daf-2 double mutants (Fig. 6A). The long-lived phenotype of sea-2; daf-2 double mutants was also completely suppressed by loss of function of daf-16 (Fig. 6A; see Table S4 in the supplementary material). Thus, sea-2 probably acts in parallel to daf-2 signaling and converges on daf-16 in regulating lifespan.

In wild-type young adult animals, DAF-16 is diffusely localized in the cytoplasm (Fig. 6B). Reduced activity of insulin/IGF-l signaling promotes nuclear translocation of DAF-16, which subsequently activates expression of genes involved in modulating stress resistance and aging (Lin et al., 1997; Ogg et al., 1997; Lee et al., 2001; Lee et al., 2003). In sea-2 mutants, prominent nuclear localization of DAF-16 was observed in cells of the intestine (Fig. 6C), a tissue that is important for mediating the effect of DAF-16 on aging (Libina et al., 2003). Weak nuclear localization of DAF-16 was also evident in hypodermal and muscle cells (Fig. 6C). We further examined the expression of sod-3, a well-characterized target of DAF-16, in sea-2 mutants (Libina et al., 2003). In wild-type young adult animals, sod-3::gfp was weakly expressed in intestinal, hypodermal and pharyngeal cells (Fig. 6D,E). Expression of sod-3::gfp was dramatically elevated in sea-2 mutants (Fig. 6F,G), but this upregulation was completely abolished by reduced activity of daf-16 (Fig. 6H,I). Therefore, loss of function of sea-2 results in nuclear translocation of DAF-16 and activation of DAF-16 targets.

To investigate whether lifespan extension in sea-2 mutants results from the retarded heterochronic defects at larval stages or is determined at the adult stage, we measured the lifespan of animals with reduced sea-2 activity at different ages. RNAi inactivation of sea-2 at early larval stages caused retarded heterochronic defects and also increased the lifespan (Table 1; data not shown). RNAi inactivation of sea-2 in young adults had no effect on the temporal fate of seam cells, but still extended the lifespan (Fig. 6J; see Table S4 in the supplementary material). Therefore, sea-2 regulates lifespan independent of its role in specifying developmental timing in larvae.

We next expressed sea-2 in a tissue-specific fashion to determine whether sea-2 activity in any single tissue was sufficient to affect the lifespan. sea-2 was specifically expressed in seam cells, neurons, body wall muscles or the intestine by fusing it with the ceh-16, rgef-1, myo-3 or vha-6 promoters, respectively (see Fig. S3 in the supplementary material). The lifespan of sea-2 mutants carrying the transgene was measured. We found that expression of sea-2 in the intestine, but not in other tissues, rescued the extended lifespan phenotype in sea-2 mutants (Fig. 6K; see Table S5 in the supplementary material). Therefore, sea-2 appears to act in the intestine to regulate adult lifespan.

Finally, we investigated whether insulin/IGF-1 signaling plays a role in the heterochronic circuit in larvae. daf-2 and daf-16 single mutants had the wild-type number of 16 seam cells and displayed no evident heterochronic defects (Table 1). However, daf-2(e1370) greatly enhanced the retarded heterochronic defects in sea-2, daf-12 or alg-1 mutants (Table 1). For example, the average number of seam cells was increased from 20 in sea-2 mutants to 28 in sea-2; daf-2 mutants and 42% of sea-2; daf-2 double mutants showed defective terminal differentiation of seam cells, compared with 6% of sea-2 single mutants. Loss of function of daf-16 did not suppress the heterochronic defects in sea-2 mutants (Table 1). However, daf-16 completely suppressed the daf-2-dependent enhancement of the heterochronic defects in sea-2 mutants (Table 1). Thus, daf-16 also functions downstream of daf-2 in controlling developmental timing in larvae. Expression of the lin-28::gfp reporter in sea-2 mutants was not further elevated by simultaneous depletion of daf-2 (see Fig. S7 in the supplementary material). The hbl-1::gfp reporter was also unchanged in sea-2; daf-2 mutants (see Fig. S7 in the supplementary material). Thus, it remains to be determined whether daf-2 signaling converges on lin-28 or hbl-1 in the heterochronic circuit.

Here, we demonstrated that the RNA-binding protein SEA-2 regulates the expression of lin-28 in the heterochronic circuit, probably via the miRNA lin-4 complementary element in the 3′UTR. SEA-2 may modulate the function of the miRNA-induced silencing complex (miRISC) at the lin-28 3′UTR. SEA-2 is not generally involved in modulating miRNA activity, because loss of function of sea-2 has no effect on expression of two other miRNA targets, hbl-1 (regulated by let-7 family miRNAs) and cog-1 (regulated by lsy-6 miRNA) (see Fig. S4 in the supplementary material and data not shown). The role of sea-2 in suppressing the XX lethality of fox-1 sex-1 animals also requires the activity of lin-28. Mammalian Lin28 functions as a post-transcriptional repressor of the biogenesis of let-7 family miRNAs (Heo et al., 2008), raising the possibility that LIN-28 mediates the biogenesis of miRNAs (e.g. let-7) that in turn play a role in sex determination and dose compensation. We also found that sea-2 acts in parallel to daf-2 insulin signaling to activate daf-16. Whether sea-2 regulates adult lifespan through lin-28 remains uncertain. lin-28 mutants are short-lived (data not shown), which could be due to pleiotropic defects caused by loss of function of lin-28, including defective egg-laying. Animals carrying the lin-28::gfp::lin-28 3′UTR(ΔLCE) transgene that showed elevated levels of lin-28 are also short-lived (data not shown). Overexpression of sir-2.1 extends the adult lifespan in a manner dependent on DAF-16. However, sea-2; sir-2.1(O/E) animals are short-lived, which may be because double mutants display multiple defects such as bursting vulva (data not shown). Thus, the way in which sea-2 converges on daf-16 in regulating lifespan has yet to be determined. Consistent with a study showing that dauer arrest is decoupled from lifespan regulation, even though both processes are controlled by IGF-1 signaling (Dillin et al., 2002), sea-2 mutation alone has no effect on dauer formation nor does it potentiate dauer formation in daf-2 mutants (data not shown). Loss of function of daf-16 had no effect on the role of sea-2 either in the heterochronic pathway or in suppressing fox-1 sex-1 lethality (see Table S3 in the supplementary material). Therefore, sea-2 acts through distinct effectors to regulate developmental timing, dose compensation and lifespan (Fig. 6L). The multiple functions of sea-2 during embryonic and larval development, such as specifying the temporal fate of seam cells and functioning as an ASE in dose compensation, indicate that wild-type sea-2 has beneficial effects early in animal development. Thus, sea-2 is favored by selection during evolution, even though sea-2 mutants are long-lived. This provides support for the antagonistic pleiotropy theory of aging (Hughes and Reynolds, 2005).

Our study further supports the presence of an intrinsic timing mechanism that temporally initiates a program of aging in the adult, consistent with a coordinate shift in gene expression pattern early in adulthood in C. elegans and Drosophila (McCarroll et al., 2004). Here, we have revealed a novel function of daf-2 insulin/IGF-1 signaling in controlling developmental timing at larval stages. Reduced activity of daf-2 dramatically enhances retarded heterochronic defects by causing reiteration of the L2 stage fate at late larval stages. On the other hand, several components of the heterochronic circuit also influence the rate of aging. Reducing lin-14 activity or overexpressing lin-4 modestly extends lifespan in a daf-16-dependent fashion (Boehm and Slack, 2005). However, there is no correlation between the degree or type of heterochronic defect and the rate of aging. lin-14 specifies the L1/L2 transition, while sea-2 and daf-2 control the L2/L3 switch. Loss of function of lin-14 causes precocious, whereas sea-2 and daf-12 mutations cause retarded, heterochronic defects. lin-14 and sea-2 mutants have an extended lifespan, whereas null daf-12 mutants have slightly shortened lifespans (Fisher and Lithgow, 2006). sea-2, lin-14 and daf-2 influence aging during adulthood in a way that is temporally separable from their role in determining developmental timing (Boehm and Slack, 2005; Dillin et al., 2002). Thus, the timing program that modulates the aging process in adults shares a subset of genes with the heterochronic circuit that functions at larval stages.

We thank Dr Barbara Meyer for comments on the manuscript, Drs Victor Ambros and Eric Moss for the lin-28::gfp reporter and anti-LIN-28 antibody, Dr David Fay for the hbl-1::gfp reporter, Dr Min Han for col-10::lacZ::lin-28 plasmids, Dr Shohei Mitani for sea-2(tm4355), and Dr Isabel Hanson for proofreading services. Some strains used in this work were received from the Caenorhabditis Genetics Center. This work was supported by a Ministry of Science and Technology grant (2010CB835201) to H.Z.