Regulation of nuclear-cytoplasmic partitioning by the lin-28-lin-46 pathway reinforces microRNA repression of HBL-1 to confer robust cell-fate progression in C. elegans Free

Precise and robust gene regulation is crucial for animal development. Optimal doses of developmental gene products expressed with spatiotemporal precision produce the wild-type body plan, whereas abnormally lower or higher doses or ectopic expression of developmental genes can result in morphological defects that reduce the fitness of the individual. The proper spatiotemporal activity of developmental gene products is ensured by elaborate gene regulatory mechanisms, which often involve collaboration across semi-redundant mechanisms controlling the gene activity at different levels – transcriptional, translational and post-translational.

Caenorhabditis elegans development consists of an invariant set of cell division and differentiation events that produces the stereotyped adult body plan (Sulston and Horvitz, 1977). C. elegans developmental regulators are identified by loss-of-function or gain-of-function mutations that cause developmental lethality, or evident morphological defects. One class of developmental defects in C. elegans stems from changes in the order and/or timing of larval developmental events, controlled by the heterochronic gene pathway (Ambros and Horvitz, 1984). In this pathway, three transcription factors (LIN-14, HBL-1, LIN-29) control cell-fate transitions from earlier to later stages, and are temporally regulated – directly or indirectly – by certain microRNAs. In particular, LIN-14 is regulated by lin-4 (Lee et al., 1993), HBL-1 (Hunchback-like-1) is regulated by three members of the let-7 family (also known as the let-7 sisters: mir-48, mir-84 and mir-241) (Abbott et al., 2005), and LIN-29 is post-transcriptionally regulated by the RNA-binding protein LIN-41 (Slack et al., 2000), which is in turn regulated by let-7 (Reinhart et al., 2000). These microRNAs are dynamically expressed during larval development and they ensure proper temporal downregulation of their targets, which is crucial for the proper program of stage-appropriate cell-fate transitions.

The conserved RNA-binding protein LIN-28 also plays key roles in the C. elegans heterochronic pathway. lin-28 regulates early cell fates upstream of lin-46 (Pepper et al., 2004) and in parallel to mir-48/84/241 (Abbott et al., 2005), and regulates late cell fates upstream of the conserved microRNA let-7 (Van Wynsberghe et al., 2011; Vadla et al., 2012). In larvae lacking lin-28, hypodermal stem cells (called seam cells) skip L2-stage proliferative cell fates and precociously express terminally differentiated adult cell fates (Ambros and Horvitz, 1984; Moss et al., 1997), whereas the rest of the tissues, e.g. the gonad, are still juvenile and developing. lin-46(lf) suppresses these lin-28(lf) phenotypes (Pepper et al., 2004): lin-28(lf);lin-46(lf) double mutants are wild type for the phenotypes observed in lin-28(lf) animals. lin-28(lf) suppresses the heterochronic phenotypes of mir-48/84/241 mutants (Abbott et al., 2005), and lin-46 is required for this suppression (Abbott et al., 2005).

Animals lacking lin-46 display weak heterochronic phenotypes that are enhanced when the larvae are cultured at low temperatures, such as 15°C, a condition that does not affect wild-type development (Pepper et al., 2004). Although wild-type larval development is similarly robust against other stresses, such as population density pheromones or starvation (Ilbay and Ambros, 2019), lin-46(lf) phenotypes are enhanced under these conditions (Ilbay and Ambros, 2019) and when animals develop through a temporary diapause in response to these stresses (Karp and Ambros, 2012). Interestingly, diapause-inducing conditions also repress let-7 family microRNA expression (Bethke et al., 2009; Hammell et al., 2009); which is thought to be important for an optimal diapause decision (Hammell et al., 2009) and for prolonging HBL-1 expression in coordination with the rate of developmental stage progression (Ilbay and Ambros, 2019). Therefore, LIN-46 activity is important for the proper downregulation of HBL-1 under physiological conditions where let-7 family microRNA levels are reduced.

The molecular functions of LIN-46, and how these functions may relate to its role in the heterochronic pathway are not known. However, given the prominent involvement of microRNAs in the heterochronic pathway, in the context of regulating temporal cell fates, LIN-46 has been thought to act possibly by modulating (more precisely ‘boosting’) the activities of certain microRNAs, e.g. lin-4 or let-7 family microRNAs, perhaps by interacting with the miRISC (microRNA-induced silencing complex). By hypothetically boosting the activities of lin-4 and/or let-7 family microRNAs in response to environmental stress, LIN-46 could compensate for the reduced levels of these microRNAs. This model is supported by the correlated conservation of LIN-46 with the argonaute family proteins, suggesting a potential function for LIN-46 related to microRNA or small RNA pathways (Tabach et al., 2013). Alternatively, LIN-46 could regulate HBL-1 activity by a mechanism independent of microRNAs.

Here, we report that deleting a genomic region encompassing the let-7/lin-4 complementary sites (LCSs) in the hbl-1 3′UTR results in strong extra seam cell phenotypes, which is consistent with lack of let-7/lin-4 microRNA regulation and a gain-of-function of HBL-1. Importantly, we find that lin-28(lf) suppresses, and lin-46(lf) enhances, the extra seam cell phenotype of hbl-1(gf/ΔLCSs), indicating that regulation of HBL-1 by the lin-28-lin-46 pathway is independent of the LCSs. Moreover, HBL-1, which normally localizes to the nucleus, accumulates in the cytoplasm of hypodermal seam cells in lin-28(lf) and hbl-1(gf/ΔLCSs) animals, and lin-46 is required for this cytoplasmic accumulation of HBL-1. This cytoplasmic accumulation is accompanied by reduced nuclear accumulation of HBL-1, which also correlates with reduced HBL-1 activity. Lastly, we found that precocious expression of LIN-46 in L2-stage seam cells is sufficient to localize HBL-1 to the cytoplasm, reducing the nuclear accumulation of HBL-1, and thereby suppressing hbl-1 gain-of-function phenotypes in hbl-1(gf/ΔLCSs) mutants. Our results indicate that the C. elegans lin-28-lin-46 pathway regulates the temporal dynamics of nuclear accumulation of the HBL-1 transcription factor, acting in parallel with the translational repression exerted by the let-7-family microRNAs, to confer precision and robustness to the temporal downregulation of HBL-1 activity.

To explore whether lin-46 acts downstream of lin-28 in the heterochronic pathway by modulating the regulation of hbl-1 by let-7 family and/or lin-4 microRNA, we sought to generate hbl-1(gf) alleles free from regulation by these microRNAs. The hbl-1 3′UTR contains ten let-7 complementary sites (LeCSs) and a single lin-4 complementary site (LiCS), collectively abbreviated as LCSs. In order to abrogate let-7- and lin-4-mediated regulation of hbl-1, we deleted a genomic region encompassing all LCSs in the hbl-1 3′UTR (ma354; Fig. 1A; Table S1). We found that, similar to mir-48/84/241(0) mutants, hbl-1(ma354[ΔLCSs]) animals have retarded seam cell defects, wherein L2-stage fates are reiterated at later stages, resulting in extra seam cells in young adult animals (Fig. 1A).

Additionally, in our screens for large deletions in the progeny of CRISPR/Cas9-injected animals, we recovered smaller 3′UTR deletions of various sizes that removed several but not all LeCSs in the hbl-1 3′UTR (Fig. 1A; Table S1). We analyzed these smaller deletions along with the largest deletion (ma354), and found that most of these mutants also have extra seam cell phenotypes, although weaker than the ma354 deletion (Fig. 1A). 3′UTR deletions that removed more LeCSs resulted in stronger extra seam cell phenotypes, which is consistent with the idea that these LeCSs are functional and they act partially redundantly.

Next, to test if the regulation of hbl-1 by the lin-28-lin-46 pathway was dependent on the let-7 and lin-4 complementary sites (LCSs) in the hbl-1 3′UTR, we generated compound mutants containing the ma354 deletion, together with null alleles of lin-28, and/or lin-46. We found that lin-28(lf) suppresses and lin-46(lf) substantially enhances the extra seam cell phenotype of hbl-1(ma354) (Fig. 1B), indicating that the regulation of hbl-1 by lin-28 does not require the LCSs in the hbl-1 3′UTR. We also found that lin-46 is required for the suppression of hbl-1(ma354) by lin-28(lf) (Fig. 1B). These results suggest that the lin-28-lin-46 pathway regulates HBL-1 amount or activity through a mechanism independent of the let-7/lin-4 regulation of HBL-1.

LIN-46 is expressed precociously in lin-28(lf) animals (Ilbay et al., 2019 preprint) and so suppression of hbl-1(gf) by lin-28(lf) could be solely due to precocious LIN-46, which is sufficient to inhibit L2 cell fates and promote transition to L3 and later cell fates (Ilbay et al., 2019 preprint). To determine whether precocious LIN-46 expression alone is also sufficient to suppress the extra seam cell phenotypes of hbl-1(ma354) animals, we employed a lin-46(gf) mutation, lin-46(ma467), that consists of a 12 bp deletion of lin-46 5′UTR sequences, and that results in precocious expression of LIN-46 and lin-28(lf)-like phenotypes (Ilbay et al., 2019 preprint). We generated double mutant animals carrying hbl-1(ma354) and lin-46(ma467), and found that the gain-of-function allele of lin-46 suppresses the extra seam cell phenotypes of hbl-1(ma354) (Fig. 1B), suggesting that precocious LIN-46 expression is sufficient to suppress the hbl-1(gf) phenotypes. This result supports the interpretation of the suppression of hbl-1(gf) by lin-28(lf) as resulting from precocious LIN-46 expression.

The suppression and enhancement of the extra seam cell phenotypes of hbl-1(ma354) by lin-28(lf) and lin-46(lf), respectively, could reflect changes in the level of HBL-1 protein. In order to test for changes in the levels of HBL-1 protein in lin-28(lf) or lin-46(lf) mutants, we tagged hbl-1 at the endogenous locus with mScarlet-I using CRISPR/Cas9 (Ilbay and Ambros, 2019). We observed that in wild-type L1 and L2 larvae HBL-1::mScarlet-I localizes exclusively to the nucleus (Fig. 2A; Fig. S1A), which is consistent with HBL-1 functioning as a transcription factor (Fay et al., 1999; Niwa et al., 2009). Moreover, consistent with previous reports, endogenously tagged HBL-1 was expressed in the wild type in the hypodermal seam and hyp7 cells of L1- and L2-stage larvae and was not detected in L3- and L4-stage larvae (Lin et al., 2003; Abrahante et al., 2003; Ilbay and Ambros, 2019) (Fig. 2A; Fig. S1A).

To examine the impact of the loss of the LCSs in the hbl-1 3′UTR on the expression pattern of HBL-1, we deleted a region in the hbl-1 3′UTR of the mScarlet-I-tagged hbl-1 allele, generating the ma430ma475 [hbl-1::mScarlet-I::Δ8LeCSs] allele. The ma475 deletion encompasses eight LeCSs and the LiCS, similar to and only three base pairs shorter than the ma293 deletion (Fig. 1A; Table S1). Consistent with previous reports that utilized GFP reporters fused with wild-type hbl-1 3′UTR (Abrahante et al., 2003; Abbott et al., 2005), absence of these LeCSs resulted in HBL-1 expression that persisted in the L3- and L4-stage hypodermal cells (Fig. 2B versus 2A). Interestingly, we observed that at the L3 and L4 stages, HBL-1 accumulates in the cytoplasm of the seam cells, which is accompanied by a reduction in the nuclear accumulation of HBL-1 (Fig. 2B; Fig. S1B). This observation suggested that perhaps the nuclear accumulation of HBL-1 is hindered (or cytoplasmic accumulation of HBL-1 is facilitated) in these L3/L4-stage seam cells.

lin-28(lf) animals skip L2-stage proliferative seam cell fates, suggesting that lin-28 is required to support the activity of HBL-1 at the L2 stage. Moreover, loss of lin-28 can suppress the extra seam cell phenotypes of hbl-1 gain-of-function (gf) mutants (Fig. 1B), indicating that lin-28 is also required to support high and/or prolonged expression of HBL-1. We observed that in double mutants containing hbl-1(ma430ma475) and lin-28(lf), HBL-1 accumulates primarily in the cytoplasm and is largely absent from the nucleus of the L2-stage seam cells (Fig. 3B versus 3A), which explains the lack of HBL-1 activity and the suppression of hbl-1(gf) extra seam cell phenotypes in larvae lacking lin-28.

In order to test the possibility of an effect of the linker or the mScarlet-I tag on the localization of HBL-1, we also tagged another transcription factor, daf-12 (nuclear accumulation of which is not affected by lin-28(lf); Fig. S2A), at its endogenous locus with the same linker and mScarlet-I and determined whether the localization of DAF-12 changed in lin-28(lf) animals (Fig. S2). We observed that, unlike HBL-1, linker-mScarlet-I tagged DAF-12 did not accumulate in the cytoplasm of L2-stage seam cells in lin-28(lf) animals (Fig. S2B), suggesting that the linker-mScarlet-I tag could not be the cause of cytoplasmic accumulation of HBL-1.

HBL-1 accumulated both in the nucleus and the cytoplasm of L3- and L4-stage seam cells in hbl-1(ma430ma475[mScarlet-I::Δ8LCSs]) larvae (Fig. 2B). When we combined hbl-1(ma430ma475[mScarlet-I::Δ8LCSs]) with lin-46(lf) we did not observe cytoplasmic accumulation of HBL-1 at the L3 and L4 stages, rather HBL-1 accumulated in the nucleus at all stages (Fig. 2C; Fig. S1C), indicating that the L3/L4-stage cytoplasmic accumulation of HBL-1 requires lin-46 activity.

HBL-1 accumulated primarily in the cytoplasm of L2-stage seam cells in lin-28(lf) larvae (Fig. 3B). By contrast, in L2 larvae lacking both lin-28 and lin-46 HBL-1 no longer accumulated in the cytoplasm of L2 or later stage seam cells, rather HBL-1 accumulated in the nucleus of the seam cells at all stages (Fig. 3C; Fig. S1F).

In brief, lin-46(lf) resulted in the loss of cytoplasmic HBL-1 accumulation at all stages, accompanied by restoration of nuclear accumulation of HBL-1 (Fig. S1).

Lastly, we found that precocious LIN-46 expression, which is sufficient to suppress the extra seam cell phenotypes of hbl-1(ma354) (Fig. 1B), is also sufficient to reduce the nuclear accumulation of HBL-1 (Fig. 3D).

These results, together with those presented above show that the suppression of hbl-1(ma354) phenotypes by lin-28(lf) or lin-46(gf) is accompanied by increased cytoplasmic accumulation of HBL-1 and a reduction in the nuclear accumulation of HBL-1. These findings suggest that LIN-46 negatively regulates HBL-1 activity by hindering its nuclear accumulation.

Endogenously tagged lin-46 is expressed in the seam cells of L3- and L4-stage larvae but not in the seam cells of L1- and L2-stage larvae (Ilbay et al., 2019 preprint). Therefore, the onset of LIN-46 expression coincides with the onset of cytoplasmic accumulation of HBL-1 in hbl-1(ma430ma475[mScarlet-I::Δ8LCSs]) animals (which are, unlike the wild type, not capable of properly downregulating HBL-1 at the end of the L2 stage owing to the lack of LCSs) (Fig. S1). In lin-28(lf) animals, LIN-46 is expressed precociously (Ilbay et al., 2019 preprint), starting in mid L1-stage seam cells. This precocious onset of LIN-46 expression in lin-28(lf) animals also coincides with the precocious onset of the cytoplasmic accumulation of HBL-1 in lin-28(lf) animals (Fig. S1).

We note that LIN-46 expression is detected in the seam cells but not in the hyp7 cells (Ilbay et al., 2019 preprint), and the presence/absence of LIN-46 expression in these two hypodermal cell types correlates with the presence/absence of cytoplasmic accumulation of HBL-1. For example, whereas HBL-1 accumulates in both the cytoplasm and the nucleus of seam cells L3/L4-stage hbl-1(ma430ma475[mScarlet-I::Δ8LCSs]) animals (Fig. 2B), HBL-1 accumulates only in the nuclei of the hyp7 cells (Fig. 2B). Similarly, whereas HBL-1 accumulates primarily in the cytoplasm of L2-stage seam cells of lin-28(lf) animals, HBL-1 still accumulates in the nuclei of hyp7 cells in these animals (Fig. 3B, black arrow versus white arrowhead).

In brief, cytoplasmic accumulation of HBL-1 was observed in the hypodermal cells where (seam) and when [by the L3 stage in wild type and by the late L1 stage in lin-28(lf)] LIN-46 is expressed (Fig. S1).

Our results suggest that the C. elegans lin-28-lin-46 pathway regulates the nuclear accumulation of HBL-1, a transcription factor that specifies L2-stage proliferative cell fates and opposes the progression to L3-stage self-renewal cell fates during C. elegans development. lin-28 is required for the nuclear accumulation of HBL-1 in hypodermal seam cells at the L2 stage, which is, in turn, necessary for the execution of L2-stage proliferative cell fates. The lin-28 target lin-46 is responsible for preventing the nuclear accumulation of HBL-1 in lin-28(lf) animals, and in wild-type animals lin-28-mediated repression of the LIN-46 expression at the L1 and L2 stages (Ilbay et al., 2019 preprint) allows the nuclear accumulation of HBL-1 at those early larval stages. Using a lin-46 5′UTR mutation that renders LIN-46 expression poorly repressed by lin-28, we show that precocious LIN-46 expression in the seam cells of the L1/L2-stage larvae is sufficient to reduce the nuclear accumulation of HBL-1. Furthermore, using hbl-1 gain-of-function mutations with let-7 and lin-4 sites deleted from the hbl-1 3′UTR, we show that the lin-28-lin-46 pathway acts in parallel with let-7 family microRNAs. Hence, these two parallel pathways – the microRNA pathway controlling the rate of synthesis of HBL-1 through repression of hbl-1 mRNA translation, and the lin-28-lin-46 pathway controlling the nuclear accumulation of HBL-1 – function together to ensure the precision and robustness of stage-specific HBL-1 downregulation (Fig. 4).

In wild-type animals, HBL-1 and LIN-46 are expressed at temporally distinct stages: HBL-1 is expressed at the L1 and L2 stages whereas LIN-46 is expressed at the L3 and L4 stages. In larvae of certain mutants, such as hbl-1(gf) L3 and L4 larvae, or lin-28(0) L1 and L2 larvae, LIN-46 and HBL-1 expression overlap, and cytoplasmic accumulation of HBL-1 is observed, accompanied by a reduction in the nuclear accumulation of HBL-1. Our data further show that the nucleus-to-cytoplasm displacement of HBL-1 in these contexts depends on lin-46 activity. Therefore, one might have expected to observe cytoplasmic accumulation of HBL-1 after the L2-to-L3 transition in wild-type larvae, when LIN-46 begins to accumulate. Curiously, in wild-type animals, cytoplasmic accumulation of HBL-1 is not evident at any stage, despite the presence of LIN-46 in L3-adult animals. If LIN-46 causes cytoplasmic accumulation of HBL-1 in the wild type, why is HBL-1 not detected in L3 and L4 larvae? One explanation could be that in wild-type larvae, the post-translational repression of HBL-1 activity by LIN-46 (via cytoplasmic localization) functions semi-redundantly with the translational repression of HBL-1 by let-7 family microRNAs. In this scenario, the microRNA pathway could exert the lion's share of HBL-1 downregulation at the L2-to-L3 transition, and the upregulation of LIN-46 at the L3 stage could play a secondary role to inhibit the nuclear accumulation of any residual HBL-1 protein. Indeed, in support of this idea, a low level of HBL-1 expression persists in the nuclei of L3-stage seam cells in lin-46(lf) animals (Fig. S1). This scenario is also consistent with the differing strengths of the weaker retarded phenotypes of lin-46(lf) animals compared with hbl-1(gf) animals under standard culture conditions (Fig. 1B).

Interestingly, conditions such as diapause-inducing stress signals that enhance lin-46(lf) phenotypes (Ilbay and Ambros, 2019) also result in a reduction in the expression of let-7 family microRNAs (Hammell et al., 2009), resulting in a shift from primarily microRNA-mediated regulation of HBL-1 to primarily LIN-46-mediated regulation (Ilbay and Ambros, 2019). In this context, where wild-type larvae experience diapause-inducing stress signals, and hence LIN-46-mediated cytoplasmic localization becomes the primary mode of HBL-1 downregulation, we expected to observe cytoplasmic accumulation of HBL-1 after the L2d-to-L3 transition. However, we could not detect any cytoplasmic HBL-1::mScarlet-I fluorescence in wild-type L3 larvae under L2d-promoting conditions. It is possible that the cytoplasmic HBL-1 in L3 larvae is unstable and/or dispersed such that the HBL-1::mScarlet-I signal is below the limit of detection in our fluorescence microscopy assays.

Sequence homology places LIN-46 into a conserved protein family, members of which include bacterial MOEA as well as human GPHN (gephyrin), which are implicated in molybdenum co-factor (MoCo) biosynthesis (Schwarz et al., 2009). GPHN has also been reported to function as a scaffold protein that is required for clustering of neurotransmitter receptors (Feng et al., 1998; Kneussel et al., 1999), and has been shown to physically interact with several other proteins (Fritschy et al., 2008), including tubulin (Kirsch et al., 1991), dynein (Fuhrmann et al., 2002) and mTOR (Sabatini et al., 1999). It is not known whether LIN-46 possesses MOEA-related enzymatic activity and/or has scaffolding functions similar to GPHN, and, if so, how such activities could (directly or indirectly) contribute to inhibition of the nuclear accumulation of a transcription factor such as HBL-1.

Analysis of the HBL-1 amino acid sequence does not reveal a predicted nuclear localization signal (NLS) that could mediate HBL-1 nuclear transport. If HBL-1 has an unconventional or ‘weak’ NLS, it is possible that other unknown factors may be required to efficiently couple HBL-1 to the nuclear import machinery. LIN-46 might inhibit HBL-1 nuclear accumulation by binding or modifying a factor that is crucial for HBL-1 nuclear transport. Alternatively, LIN-46 could bind or modify HBL-1 directly in order to prevent its association with the nuclear import machinery. It is also possible that LIN-46 could act not by directly preventing nuclear import of HBL-1, but by causing HBL-1 to be trapped in the cytoplasm, for example through the formation of LIN-46–HBL-1 complexes in association with a cytoplasmic compartment.

Regulation of nuclear accumulation in the context of temporal cell-fate specification during C. elegans development has not previously been reported. Other transcription factors, including LIN-14, DAF-12 (Antebi et al., 2000) and LIN-29, play key roles in regulating temporal cell fates during C. elegans development. These other transcription factors are also regulated by microRNAs, like HBL-1, and regulation of their temporal abundances is important for the proper execution of stage-specific cell fates. Although we have no evidence that LIN-46 may also regulate the nuclear/cytoplasmic partitioning of other heterochronic pathway proteins (except for DAF-12; Fig. S2), our findings suggest that similar post-translational mechanisms might be in place to function in parallel with the microRNA-mediated regulation and hence promote the robust temporal regulation of key developmental regulators.

Many let-7 targets, as well as many targets of other microRNAs, in worms, flies and mammals are transcription factors (Ambros, 2004; Bartel, 2004; Enright et al., 2004; John et al., 2004). Therefore, similar mechanisms, whereby a transcription factor is regulated both by a microRNA and in parallel by a gene product that controls the nuclear accumulation of the same transcription factor, may be common. Additionally, the regulation of the let-7 microRNA by LIN-28 is widely conserved. Targets of LIN-28 in other species may have roles, similar to LIN-46, in regulation of let-7 targets, controlling their nuclear accumulation in particular and their activities by means of post-translational interventions in general. A dual control of a gene product – its synthesis rate by microRNAs and activity by post-translational regulators – would allow more precise and/or more robust transitions between active and inactive states (Fig. 4B).

C. elegans strains used in this study and corresponding figures in the paper are listed in Table S2. C. elegans strains were maintained at 20°C on nematode growth media (NGM) and fed with the Escherichia coli HB101 strain.

The worms were scored at the young adult stage (determined by the gonad development) for the number of seam cells using fluorescence microscopy with the help of the maIs105 [pCol-19::gfp] transgene, which marks the lateral hypodermal cell nuclei, or the wIs51[pScm::gfp] transgene, which marks the seam cell nuclei.

Each circle on the genotype versus number of seam cells plots shows the observed number of seam cells on one side of a single young adult worm. Twenty worms (except for wild type, n=12) for each genotype were analyzed and the average number of seam cells are denoted by vertical bars in the genotype versus number of seam cell plots. Student's t-test was used to calculate statistical significance when comparing different genotypes. GraphPad Prism 8 software was used to plot the graphs and for statistical analysis.

All differential interference contrast and fluorescent images were obtained using a ZEISS Imager Z1 equipped with a ZEISS Axiocam 503 mono camera, and ZEN Blue software. Prior to imaging, worms were anesthetized with 0.2 mM levamisole in M9 buffer and mounted on 2% agarose pads. ImageJ Fiji software was used to adjust the brightness and contrast of the images to enhance the visualization of the fluorescent signal. All images were taken using the same microscopy settings and a standard exposure time for all larval stages and genetic background, but because the brightness and contrast of the individual images were enhanced separately, the signal intensities do not represent the relative expression levels and cannot be used to compare expression levels across different larval stages of genetic backgrounds.

CRISPR/Cas9 genome editing tools were used to generate the hbl-1 3′UTR deletion alleles, the lin-46 open-reading frame (ORF) deletion allele, and to tag the daf-12 gene with GFP and mScarlet-I at its endogenous locus.

For the hbl-1 3′UTR deletions and the lin-46 ORF deletion (for alleles, see Table S2), a mixture of plasmids encoding SpCas9 (pOI90), and a pair of single guide RNAs (sgRNAs, expressed from pOI83; Ilbay and Ambros, 2019) targeting both sites of interest (for primers, see Table S3) and the unc-22 gene (pOI91) as co-CRISPR marker (Kim et al., 2014), and a rol-6(su1006)-containing plasmid (pOI124) as co-injection marker was injected into the gonads of young adult worms. F1 roller and/or twitcher animals (∼50 or more worms until the desired allele was detected) were cloned and screened by PCR amplification (for primers, see Table S3) for the presence of the expected size PCR product consistent with deletion of the genomic region spanning between the sites targeted by the pair of guides.

To tag daf-12 at the endogenous locus with the same linker and mScarlet-I sequence as the hbl-1(ma430) allele, a homologous recombination (HR) donor plasmid (pOI193) and sgRNA plasmid (pOI93) were included in the CRISPR mix, which contained plasmids pOI90 (spCas9), pOI91 (unc-22 guide) and pOI124 (rol-6). The HR plasmid pOI193 contains the C-terminal end of the daf-12 gene fused in-frame with the linker and mScarlet-I sequence, subcloned from pOI191, which was used to tag hbl-1 to generate the ma430 allele. To tag daf-12 with GFP, instead of pOI193, an HR donor plasmid (pOI122) that contained the GFP sequence flanked by HR sequences was included in the CRISPR mix.

In all new CRISPR alleles, genomic regions spanning the deletion site or the HR arms and the tags introduced were sequenced using Sanger sequencing. For each allele, a single worm with a precise (HR) edited locus was cloned and backcrossed twice before being used in the experiments.

We thank the members of the Mello (UMass Medical) and Ambros laboratories for helpful discussions and the members of the Ambros laboratory for critical reading of the manuscript.