The Developmental Timing Regulator hbl-1 Modulates the Dauer Formation Decision in Caenorhabditis elegans
Xantha Karp, Victor Ambros

Abstract

Animals developing in the wild encounter a range of environmental conditions, and so developmental mechanisms have evolved that can accommodate different environmental contingencies. Harsh environmental conditions cause Caenorhabditis elegans larvae to arrest as stress-resistant “dauer” larvae after the second larval stage (L2), thereby indefinitely postponing L3 cell fates. HBL-1 is a key transcriptional regulator of L2 vs. L3 cell fate. Through the analysis of genetic interactions between mutations of hbl-1 and of genes encoding regulators of dauer larva formation, we find that hbl-1 can also modulate the dauer formation decision in a complex manner. We propose that dynamic interactions between genes that regulate stage-specific cell fate decisions and those that regulate dauer formation promote the robustness of developmental outcomes to changing environmental conditions.

ACHIEVING the correct outcome of developmental processes is essential for survival and fitness in animal species. Therefore, developmental outcomes must be robust to the range of environmental conditions experienced by animals developing in the wild. While much is understood about genetic pathways that regulate temporal and spatial cell fate specification, less is known about how these pathways are integrated with pathways that regulate the response to environmental cues. A dramatic example of a response to environmental cues is dauer diapause in Caenorhabditis elegans. In favorable environmental conditions, C. elegans larvae progress rapidly and continuously through four larval stages prior to adulthood (Sulston and Horvitz 1977). However, unfavorable environmental cues sensed in the first larval stage (L1) cause larvae to enter the predauer L2d stage, which is >50% longer than the rapid L2 stage (Golden and Riddle 1984). During the extended L2d stage, L3 cell fates are temporarily postponed while larvae continue to sense their environment and prepare for the possibility of continued harsh conditions. At the end of L2d, larvae make a choice between one of two distinct life histories: (1) to continue with developmental progression, molt to the L3 stage, and express L3 cell fates or (2) to interrupt development by entry to the stress-resistant dauer diapause, thereby indefinitely postponing L3 cell fates (Cassada and Russell 1975; Golden and Riddle 1984). Thus, the timing of expression of L3 cell fates is inextricably linked to the animal's decision of whether or not to enter the dauer diapause. This suggests that pathways regulating stage-specific cell fate decisions are coordinated with pathways that regulate dauer diapause.

The dauer formation decision occurs in response to environmental cues, including temperature, food supply, and population density. These cues are perceived by sensory neurons, resulting in up- or downregulation of two partially parallel signaling pathways, TGFβ and insulin-/insulin-growth-factor-1-like (IIS). These two pathways in turn regulate the activity of the third major dauer formation pathway, nuclear hormone receptor signaling (for review see Fielenbach and Antebi 2008). Favorable environmental cues lead to upregulation of the TGFβ encoded by daf-7 (Ren et al. 1996; Schackwitz et al. 1996). High DAF-7/TGFβ signaling opposes the activity of the downstream transcription factor DAF-3/SMAD that complexes with DAF-5/Sno-Ski to regulate target gene expression (Patterson et al.1997; Da Graca et al. 2004). Also in favorable environmental conditions, the insulin-like receptor (InsR) DAF-2 is activated by one or more of ∼40 insulin-related proteins (Kimura et al. 1997; Pierce et al. 2001). High DAF-2/InsR activity opposes the activity of the downstream transcription factor DAF-16/FOXO (Lin et al. 1997; Ogg et al. 1997). When active, both DAF-16 and DAF-3-DAF-5 promote dauer formation by affecting expression of downstream genes (Thatcher et al. 1999; Jensen et al. 2006), including daf-9 (Gerisch et al. 2001; Gerisch and Antebi 2004; Mak and Ruvkun 2004; Motola et al. 2006). daf-9 encodes a homolog of cytochrome P450 (Gerisch et al. 2001) and is required to produce dafachronic acid (DA), the ligand for the DAF-12 nuclear hormone receptor (Motola et al. 2006). Liganded DAF-12 promotes reproductive growth, whereas unliganded DAF-12 is absolutely required for dauer formation (Antebi et al. 1998; Antebi et al. 2000; Gerisch et al. 2001; Motola et al. 2006).

hbl-1 promotes dauer formation downstream or in parallel to daf-9:

The study of daf-12 provided the first indication of coordination between the regulation of L2 vs. L3 cell fates and the regulation of the dauer formation decision. This was based on the finding that certain daf-12 alleles cause the inappropriate reiteration of L2 cell fates in L3-staged larvae (Antebi et al. 1998). More recently, additional genes that regulate the L2 vs. L3 cell fate decision have been shown to influence the dauer formation decision (Hammell et al. 2009; Tennessen et al. 2010). hbl-1 promotes L2 cell fates and inhibits L3 cell fates, downstream from other known L2 vs. L3 cell fate regulators (Abrahante et al. 2003; Lin et al. 2003; Abbott et al. 2005; Bethke et al. 2009; Hammell et al. 2009). We therefore tested whether hbl-1 could similarly modulate the dauer formation decision, perhaps by interacting with daf-12. The daf-12(rh273) allele encodes a DAF-12 protein with a defective ligand binding domain (Antebi et al. 2000). Because this mutant protein does not bind ligand well, it causes a constitutive dauer formation phenotype (“Daf-c”) similar to the loss of DA in daf-9 mutants (Antebi et al. 1998; Gerisch et al. 2001; Motola et al. 2006). To test the effect of lowering hbl-1 activity in a daf-12 mutant background we constructed double mutant strains between daf-12(rh273) and either of two hypomorphic hbl-1 alleles: hbl-1(ve18), and hbl-1(mg285). [Null alleles of hbl-1 are not available. Furthermore, RNAi experiments indicate that loss of hbl-1 is likely to result in embryonic lethality (Fay et al. 1999)]. Both hbl-1(ve18) daf-12(rh273) and hbl-1(mg285) daf-12(rh273) strains showed moderate but statistically significant suppression of the daf-12(rh273) Daf-c phenotype (Figure 1A). That the same effect was observed in both hbl-1(ve18)- and hbl-1(mg285)-containing strains indicates that the suppression is likely to be due to reducing hbl-1 activity rather than genetic background (Harvey et al. 2008). The suppression of a Daf-c phenotype by reduction of hbl-1 activity suggests that hbl-1 has a dauer-promoting activity.

Figure 1.—

hbl-1 promotes dauer formation downstream or in parallel to Daf-9. Bar graphs indicate the mean percentage of dauer formation ± SEM. At least three independent trials were carried out for each experiment. For each trial, gravid adults of control and test strains were allowed to lay eggs for several hours on 60-mm NGM plates with the OP50 strain of Escherichia coli as a food source. (While hbl-1 mutant hermaphrodites are egg-laying defective if they have developed continuously, postdauer animals do lay eggs; Abrahante et al. 2003). The adults were then removed and plates were incubated side by side at 20° until nondauer animals were L4 or early adult stage (generally 3–4 days). All animals on each plate were scored as dauer or nondauer. As daf-12(rh273) and daf-9 mutant larvae form partial and/or transient dauers (Antebi et al. 1998; Gerisch et al. 2001), dauer larvae were identified by the presence of dauer alae on the cuticle. Since we were interested in the decision to form dauer, and not in the ability to fully execute or maintain dauer arrest, we counted animals with dauer alae anywhere along the length of the body as dauers. The total number of animals scored across all trials is indicated in parentheses. Asterisks indicate statistically significant differences (P < 0.05 by a paired student's t-test, where the pairs consisted of control and test strains grown side by side). (A) Reduction of hbl-1 by either of two hypomorphic alleles can suppress the Daf-c phenotype of the daf-12(rh273) mutation. Strains: AA87 daf-12(rh273), VT2010 hbl-1(ve18) daf-12(rh273), and VT2280 hbl-1(mg285) daf-12(rh273). (B) Endogenous dafachronic acid (DA) production was eliminated by the daf-9(dh6) mutation, resulting in 100% dauer formation. Exogenous DA [(25S),26 –3-keto-4-cholestenoic acid] added to the medium at the indicated concentration can rescue this Daf-c phenotype. This rescue is more efficient if hbl-1 activity is reduced. Dafachronic acid was provided by David Mangelsdorf (University of Texas Southwestern Medical Center, Dallas, TX). Experiments were performed as described (Hammell et al. 2009). Strains: AA199 daf-9(dh6); dhEx24[daf-9+, sur-5∷GFP] and VT1996 hbl-1(ve18) daf-9(dh6); dhEx24[daf-9+, sur-5∷GFP]. Only GFP animals were scored.

The Daf-c phenotype of daf-12(rh273) is caused by an excess of the unliganded (dauer promoting) form of DAF-12, since the mutant DAF-12 protein encoded by daf-12(rh273) binds its ligand poorly (Motola et al. 2006). However, because ligand binding is not completely abrogated in this mutant, hbl-1 could potentially promote dauer formation either upstream or downstream of ligand production by DAF-9. To test whether hbl-1 could promote dauer formation independently of DA ligand production, we used a daf-9(0) mutation to eliminate endogenous DA (Motola et al. 2006). The daf-9(dh6) mutant forms dauers constitutively since all DAF-12 is unliganded in these animals (Gerisch et al. 2001; Motola et al. 2006). This Daf-c phenotype can be rescued by addition of exogenous DA (Motola et al. 2006). At low levels of DA, populations of daf-9(dh6) larvae contain some dauers and some nondauers. At these limiting DA levels, reduction of hbl-1 caused a moderate but statistically significant suppression of the Daf-c phenotype (Figure 1B). Since hbl-1 cannot be affecting daf-9 expression or activity in a daf-9(0) background, this suggests that the dauer-promoting activity of hbl-1 is downstream of or in parallel to daf-9 activity and ligand production. hbl-1 may promote the activity of unliganded DAF-12, oppose the activity of liganded DAF-12, and/or regulate the transcription of common dauer-promoting targets.

In the absence of DA, hbl-1(ve18) daf-9(dh6) double mutants form dauers constitutively, indicating that reduction of hbl-1 cannot bypass the need for daf-9 to produce DA ligand (Figure 1B). By contrast, 100 nM DA was sufficient to completely rescue the daf-9(dh6) Daf-c phenotype. At this high concentration of DA, reduction of hbl-1 had no effect (Figure 1B). The inability of reduction of hbl-1 to affect dauer formation at very low or very high DA concentrations suggests that the dauer-promoting activity of hbl-1 is modulatory and is most important in moderately stressful situations where environmental cues do not strongly drive the larvae to a particular life history (i.e., a percentage of larvae choose the dauer life history, whereas others choose the continuous life history). We also find that the suppression of daf-9(dh6) by reduction of hbl-1 is specific for the dauer formation phenotype because reducing hbl-1 activity in a daf-9(dh6) background has no effect on the gonad migration phenotype observed in these animals at limiting concentrations of DA (supporting information, Figure S1) (Motola et al. 2006).

hbl-1 opposes dauer-promoting signals from both the TGFβ and insulin-like pathways:

Having detected a modulatory activity for hbl-1 in promoting dauer formation in response to exogenous DA, we tested whether hbl-1 would influence dauer formation in genetic backgrounds impaired in signals upstream of DA production. Specifically, we tested whether reduction of hbl-1 could suppress the incompletely penetrant Daf-c phenotypes of components of the two major signaling pathways that operate upstream of daf-9: daf-7/TGFβ and daf-2/INSR (reviewed in Fielenbach and Antebi 2008). Again, we used two different hypomorphic alleles of hbl-1 to ensure that any interactions we saw between hbl-1 and daf-7 or daf-2 were likely due to reducing hbl-1 activity and not to background mutations that affect dauer formation (Harvey et al. 2008). Surprisingly, we found that instead of suppressing the Daf-c phenotype, reduction of hbl-1 enhanced the Daf-c phenotype of both daf-7(e1372) and daf-2(e1370) at semipermissive temperatures. Again, the effects were moderate but statistically significant (Figure 2, A and B). This suggests that hbl-1 has reciprocal roles—acting upstream of daf-9 to negatively modulate dauer-promoting signals and also acting downstream of daf-9 to positively modulate the dauer formation decision.

Figure 2.—

hbl-1 opposes dauer formation in parallel to insulin and TGFβ signaling. Bar graphs indicate the mean percentage of dauer formation ± SEM. At least three independent trials were carried out for each experiment. For each trial, gravid adults of control and test strains were allowed to lay eggs on 60-mm NGM plates for several hours. (While hbl-1 mutant hermaphrodites are egg-laying defective if they have developed continuously, postdauer animals do lay eggs; Abrahante et al. 2003). The adults were then removed and plates were incubated side by side at the indicated temperature until nondauer animals were L4 or early adult stage (generally 3–4 days at 20° or 5 days at 15°). All animals on each plate were scored as dauer, or nondauer. Any remaining L1 or L2 animals were not included in the percentage of dauer calculation (see Figure S2). In most strains, dauers and nondauers could be easily distinguished by morphology under the dissecting microscope. However, because daf-16 mutant larvae can form partial and/or transient dauers (Vowels and Thomas 1992), we did not rely on the dissecting microscope to score dauer formation in strains carrying these mutations. Instead, we examined the animals by Nomarski optics for the presence of dauer alae. Since we were interested in the decision to form dauer, and not in the ability to fully execute or maintain dauer arrest, we counted animals with dauer alae anywhere along the length of the body as dauers. The total number of animals scored across all trials is indicated in parentheses. Asterisks indicate statistically significant differences (P <0.05 by a paired student's t-test, where the pairs consisted of control and test strains grown side by side). (A) Reduction of hbl-1 enhances the Daf-c phenotype of daf-2(e1370). This enhancement remains even when TGFβ signaling is abrogated by a daf-5 mutation. Strains (grown at 20°): CB1370 daf-2(e1370), VT1773 daf-2(e1370); hbl-1(ve18), VT2200 daf-2(e1370); hbl-1(mg285), VT2311 daf-5(m512); daf-2(e1370), and VT2312 daf-5(m512); daf-2(e1370); hbl-1(ve18). (B) Reduction of hbl-1 enhances the Daf-c phenotype of daf-7(e1372). This enhancement remains even when IIS signaling is abrogated by a daf-16 mutation. Strains (grown at 15°): CB1372 daf-7(e1372), VT1796 daf-7(e1372); hbl-1(ve18), VT2316 daf-7(e1372); hbl-1(mg285), VT2317 daf-16(mgDf50); daf-7(e1372), and VT2318 daf-16(mgDf50); daf-7(e1372), hbl-1(ve18). (C) In a wild-type background, reduction of hbl-1 fails to display a convincing effect on dauer formation, as the P-value was just outside our threshold for significance, and the error bars overlap substantially. A synthetic component of dauer pheromone called daumone [(−)-(6R)-(3,5-dihydroxy-6-methyltetrahydropyran-2-yloxy) heptanoic acid; Jeong et al. 2005] was added to a final concentration of either 0 μm or 384 μm to ∼55° NG agar lacking peptone and with 30 μg/ml streptomycin. We poured ∼750 μl of this modified NGM into each well of 12-well culture plates. When cool, 8.3 μl 6× concentrated OP50 bacteria were added to each well. Plates were used within 1–2 days of pouring. As above, gravid adults were added to each well and allowed to lay ∼50 eggs and then removed. Plates were incubated at 20° for 3 days and then scored. Strains: N2 wild-type, VT1874 hbl-1(ve18).

Neither daf-7(e1372) nor daf-2(e1370) are null alleles. Therefore, the enhancement of the Daf-c phenotype of these alleles by reduction of hbl-1 does not distinguish between whether hbl-1 acts in the daf-7 pathway, the daf-2 pathway, or in both pathways. DAF-7/TGFβ signaling antagonizes the activity of the DAF-3/SMAD + DAF-5/Sno-Ski transcription complex, whereas signaling through DAF-2/InsR antagonizes the activity of the DAF-16/FoxO transcription factor (for review see Fielenbach and Antebi 2008). Loss of either daf-3 or daf-5 completely suppresses the Daf-c phenotype of daf-7 alleles, but has no effect on the Daf-c phenotype of daf-2 alleles. Conversely, loss of daf-16 completely suppresses the Daf-c phenotype of daf-2 alleles, but has no effect on the Daf-c phenotype of daf-7 alleles (Vowels and Thomas 1992; Larsen et al. 1995). We took advantage of these genetic relationships to test whether hbl-1 can oppose dauer formation via TGFβ or IIS pathways. First, we constructed a daf-5(0); daf-2(e1370) double mutant strain. At the semipermissive temperature, the Daf-c phenotype of this strain is as penetrant as a daf-5(+); daf-2(e1370) strain (Figure 2A). If the enhancement of the Daf-c phenotype of daf-2(e1370) by reduction of hbl-1 were due to an activity of hbl-1 in the TGFβ pathway, that activity should be abrogated in the daf-5(0); daf-2(e1370) strain. However, we found that reduction of hbl-1 in this background still resulted in a significant enhancement of the Daf-c phenotype (Figure 2A). Therefore, the enhancement of the Daf-c phenotype of daf-2(e1370) by reduction of hbl-1 activity is not due solely to a role for hbl-1 in the TGFβ pathway upstream of daf-5, because daf-5 is not required to obtain the enhancement.

We next performed the reciprocal experiment: we asked whether daf-16 was required to observe the enhancement of the Daf-c phenotype of daf-7(e1372). If hbl-1 opposes dauer formation in the IIS pathway upstream of the DAF-16/FoxO transcription factor, removing daf-16 activity should abrogate the dauer opposing effect of hbl-1. First, we examined the control daf-16(0); daf-7(e1372) strain. Surprisingly, the mean percentage of dauer formation is higher in the daf-16(0); daf-7(e1372) strain than the daf-7(e1372) strain grown in parallel, although this difference is slightly outside our cutoff for statistical significance (Figure 2B). We next tested the effect of reducing hbl-1 activity using the hbl-1(ve18) allele in the daf-16(0); daf-7(e1372) background. We found that 90% of daf-16(0); daf-7(e1372); hbl-1(ve18) triple mutant larvae formed dauers, significantly more than either the daf-16(+); daf-7(e1372); hbl-1(+) or daf-16(0); daf-7(e1372); hbl-1(+) strains above (Figure 2B). Therefore, the enhancement of the Daf-c phenotype of daf-7(e1372) by reduction of hbl-1 activity is not due solely to a role for hbl-1 in the IIS pathway upstream of daf-16, because daf-16 is not required to obtain the enhancement.

These results indicate that hbl-1 opposes dauer formation in a fashion that does not involve a role only in the DAF-7/TGFβ pathway or only in the DAF-2/IIS pathway. Therefore, hbl-1 could modulate critical factors in both pathways and/or could act in a parallel pathway. If hbl-1 opposes dauer formation in a parallel pathway, then a Daf-c phenotype should still be evident when both TGFβ and IIS pathways are inactivated through simultaneous removal of daf-5 and daf-16. However, because daf-7 and daf-2 mutations do not produce a Daf-c phenotype in a daf-16(0); daf-5(0) background, we are unable to test for enhancement of their Daf-c phenotype by reduction of hbl-1. Therefore, the question of whether hbl-1 can oppose dauer formation in the absence of both TGFβ and IIS signaling can be addressed only if hbl-1 mutations cause a Daf-c phenotype independently of mutations in daf-7 and daf-2. However, we failed to observe a convincing hbl-1 Daf-c phenotype in the absence of daf-7 and daf-2 alleles, even in animals sensitized for dauer formation by the addition of daumone (Jeong et al. 2005) to the growth medium (Figure 2C). In an independent study, no Daf-c phenotype was observed in hbl-1 mutant animals grown at 27° (Tennessen et al. 2010). We interpret this result to indicate that, under the conditions that we culture our developing larvae, the opposing roles for hbl-1 in modulating dauer formation effectively cancel each other. We propose that the dauer-modulatory activities of hbl-1 are more critical under particular environmental or physiological conditions that are not well replicated in our laboratory conditions. The natural habitat for C. elegans is complex, including soil, decomposing organic material, and even in association with various invertebrate species (Barriere and Felix 2005). Additionally, one or the other of these opposing roles for hbl-1 could become critical when larvae develop in environmental conditions that change over time (see model in the next section).

Although we are unable to unambiguously determine whether hbl-1 acts through both TGFβ and IIS pathways or in parallel to these pathways, we favor the latter possibility for two reasons. First, it is the simplest model. Second, consistent with hbl-1 acting in parallel to daf-16, we observed that when hbl-1 activity is reduced in the absence of daf-16, one-third of larvae arrest in the L1 stage (Figure S2). This synthetic phenotype is not observed in the control strains, although it is possible that it is related to the larval lethality observed after reduction of hbl-1 activity by RNAi (Figure S2 and legend) (Fay et al. 1999).

There is evidence for the existence of dauer formation pathways in addition to TGFβ, IIS, and DAF-12/nuclear hormone pathways. One example is the daf-11 pathway that is thought to function upstream of TGFβ and IIS pathways. (daf-11 encodes a transmembrane guanylyl cylclase) (reviewed in Fielenbach and Antebi 2008). hbl-1 could act in the daf-11 pathway to oppose dauer formation. However we do not favor this possibility because both daf-5 and daf-16 mutations can partially suppress the Daf-c phenotype of daf-11(-) (Vowels and Thomas 1992; Thomas et al. 1993), whereas we see no evidence of such suppression of the hbl-1 Daf-c phenotype in our assay (Figure 2, A and B).

There is also some evidence for unknown pathways regulating dauer formation, because dauer pheromone can induce larvae to enter dauer diapause even in the absence of both daf-3/SMAD and daf-16/FOXO (Ogg et al. 1997). Therefore, even when TGFβ and IIS pathways are removed, unfavorable environmental cues are transduced into the decision to enter dauer diapause. Since unliganded DAF-12 is absolutely required for dauer formation (Thomas et al. 1993; Antebi et al. 1998), one interpretation is that this unknown dauer formation pathway is also upstream of daf-9 expression or activity. Since hbl-1 encodes a transcription factor, we hypothesize that the dauer opposing role for hbl-1 could reflect direct transcriptional regulation by HBL-1 of some of the transcriptional targets of the TGFβ and IIS pathways.

Many genes that influence dauer formation also influence lifespan, including daf-2, daf-7, daf-9, and daf-12 (Kenyon et al. 1993; Larsen et al. 1995; Gerisch et al. 2007; Shaw et al. 2007). In particular, the longevity observed in strains carrying daf-2 mutations has been long studied, and the role of the IIS pathway in aging is conserved to mammals (for review see Broughton and Partridge 2009). Since reduction of hbl-1 can enhance the dauer formation phenotype of daf-2(e1370) (Figure 2A), we asked whether reduction of hbl-1 activity might similarly enhance the longevity phenotype of the daf-2(e1370) allele. We found no effect on the longevity of daf-2(e1370) animals when hbl-1 was reduced using either the hbl-1(ve18) or the hbl-1(mg285) alleles (Figure 3). Furthermore, we found no effect on lifespan by either hbl-1 allele in a wild-type background (Figure 3). Since these are not null alleles of hbl-1, this does not rule out a role for hbl-1 in aging. However, this observation suggests that the ability of hbl-1(−) to enhance the phenotype of daf-2(e1370) is specific for the dauer formation decision. This is consistent with hbl-1 affecting the dauer formation decision independently from daf-16, since negative regulation of daf-16 by daf-2 (and daf-7) is the cause of the longevity phenotype observed in these mutant strains (Kenyon et al. 1993; Shaw et al. 2007).

Figure 3.—

hbl-1 does not affect lifespan. Survival curves show the mean percentage of survival over two independent trials. Error bars indicate maximum and minimum percentage of survival in the two trials. Neither hypomorphic allele of hbl-1 affected lifespan either in wild-type or daf-2(e1370) backgrounds. Approximately 100 synchronized L4 larvae that had grown on standard NGM plates with plenty of food at 15° were transferred to two NGM plates + 50 μg/ml FUDR (5-fluorodeoxyuridine, Sigma). Plates were incubated at 20° and examined every 2 days for living or dead animals. Animals that failed to respond to stimulation by touch were counted as dead and removed. Animals that crawled off the plate or exploded through the vulva were censored. All strains were grown in parallel. Strains: N2 wild-type, VT1874 hbl-1(ve18), CT11 hbl-1(mg285), CB1370 daf-2(e1370), VT1773 daf-2(e1370); hbl-1(ve18), and VT2200 daf-2(e1370); hbl-1(mg285).

Model for the integration of hbl-1 into dauer formation pathways:

hbl-1 expression during continuous development is negatively regulated by let-7-family microRNAs, such that hbl-1 levels are high in the L1 stage, but low in the L3 stage (Abrahante et al. 2003; Lin et al. 2003; Abbott et al. 2005). We monitored hbl-1 expression during dauer-interrupted development using the same reporter transgenes as in the above studies and found a similar expression pattern: hbl-1∷GFP levels are high in L1 and L2d stages, but downregulated at the L2d to dauer molt (Figure 4, A–D). hbl-1∷GFP remains off during dauer diapause (Figure 4E). The downregulation at the L2d to dauer molt appears to be partially 3′-UTR independent, because a hbl-1∷GFP transgene that lacks the hbl-1 3′-UTR is largely downregulated at that stage (Figure 4I). However, expression of this transgene returns during dauer diapause, indicating that the continued lack of hbl-1 expression during dauer diapause requires the hbl-1 3′-UTR (Figure 4J). While there are many possible models for how hbl-1 could modulate the dauer formation decision, below and in Figure 5 we discuss a model in which the temporal regulation of hbl-1 expression is important for its effect on dauer formation.

Figure 4.—

hbl-1 expression in the hypodermis of dauer and predauer larvae. Gravid adults carrying hbl-1 reporter transgenes were allowed to lay eggs on 35-mm NG agar plates containing 50 μl crude dauer pheromone and seeded with 20 μl 6× concentrated OP50 as a food source (Vowels and Thomas 1994). Mothers were removed after several hours, and the progeny were incubated at 20° and monitored for GFP expression throughout predauer developmental stages. Developmental stages were determined by the extent of gonad development, and molting larvae were identified by the presence of an extra cuticle. Dauer larvae were harvested from crowded and starved plates incubated at 20°. Dauer larvae and larvae at the L2d to Dauer molt were identified by the presence of dauer alae on the cuticle. Developmental stages are indicated underneath the images. GFP images appear in the top panels, and the corresponding DIC images appear beneath them. Images were obtained on a Zeiss Axio Imager D1 with an AxioCam MRm camera, and a X-Cite 120Q light source (EXFO Photonic Solutions) using a ×63 objective. Anterior is to the left. (A–E) Larvae of the strain BW1932 ctIs39[hbl-1∷GFP∷hbl-1, rol-6(su1006)] (Fay et al. 1999). (F–J) Larvae of the strain BW1891 ctIs37[hbl-1∷GFP∷unc-54, rol-6(su1006)] (Fay et al. 1999). GFP images of BW1932 were taken using a 100-ms exposure time, and GFP images of BW1891 were taken using a 40-ms exposure time. Note that ctIs39 and ctIs37 contain the identical regulatory sequence, apart from the 3′-UTR. BW1932 contains the hbl-1 3′-UTR, whereas ctIs37 contains the unregulated unc-54 3′-UTR. Numbers indicate the number of larvae at that developmental stage that display unambiguous GFP expression in hypodermal nuclei at the indicated exposure time, out of the total number examined.

Figure 5.—

A model for integration of hbl-1 activity with the regulation of dauer formation. Early in larval development, environmental cues signal through TGFβ and IIS pathways to direct progression to either the rapid L2 that leads to continuous development or the extended L2d that can lead to dauer diapause. Downstream transcription factors DAF-3 (complexed with DAF-5) and DAF-16 regulate target genes that in turn regulate dauer formation. One consequence of this regulation is daf-9 expression in favorable environmental conditions. High levels of HBL-1 at this stage promote L2 cell fates and also oppose dauer formation, perhaps by regulating common targets with DAF-3-DAF-5 and DAF-16 (blue arrow). By the time of the L2d-to-dauer molt, HBL-1 levels have been reduced (Figure 4). Lower levels of HBL-1 oppose L3 cell fates and also promote dauer formation, perhaps by regulating common targets with DAF-12. At the same time, a feedback loop between DAF-12 and let-7-family microRNAs is operating (Hammell et al. 2009), and LIN-42/Period modulates the dauer formation decision by opposing the activity of unliganded DAF-12 (Tennessen et al. 2010). This complex set of interactions may coordinate stage-specific cell fates with the dauer formation decision.

In the L1 stage, larvae assess their environment and decide whether to enter L2 or L2d, a decision involving DAF-7/TGFβ signaling (Golden and Riddle 1984; Vowels and Thomas 1992). Both DAF-7/TGFβ and DAF-28/insulin-like are expressed in sensory neurons in L1 (and later) stage larvae experiencing favorable environmental conditions, but downregulated in L1 (and later) stage larvae experiencing unfavorable environmental conditions (Ren et al. 1996; Schackwitz et al. 1996; Li et al. 2003). Indeed, the critical periods for the assessment of environmental stimuli through daf-7 and daf-2 occur within the L1 or early L2 stage (Swanson and Riddle 1981; Golden and Riddle 1984). We propose that the early, high expression of HBL-1 at this stage has an activity distinct from the later, low expression of HBL-1. For example, high levels of HBL-1 could be required to bind to the promoter of certain target genes, whereas low HBL-1 levels could be sufficient for others. It is also possible that HBL-1 could regulate the same targets early and late in development, but that the regulation could be in opposite directions. The related Drosophila Hunchback transcription factor has been shown to repress a target gene at high concentrations, but activate it at low concentrations (Schulz and Tautz 1994; Schulz and Tautz 1995; Papatsenko and Levine 2008). We hypothesize that during the L1 and early L2d stage, high HBL-1 levels promote L2 cell fates and prevent the precocious expression of late-stage programs, including not just L3 cell fates (Abrahante et al. 2003; Lin et al. 2003; Abbott et al. 2005) but also dauer formation (Figure 2). The novel dauer-opposing activity described here may occur at the level of integration between DAF-7/TGFβ and DAF-2/IIS signaling in response to environmental cues (blue arrow in Figure 5). If environmental conditions are mildly stressful at this time, hbl-1 activity may modulate the sensitivity of the larva to dauer-promoting signals to delay expression of the dauer program and help to keep development progressing rapidly.

Following the L1 stage, larvae either commit to continuous development by entering the rapid L2 stage or leave open the possibility for either continuous or dauer interrupted development by entering L2d (Golden and Riddle 1984). Although the critical period for daf-12 activity in the dauer formation decision is not known, one possibility is that daf-12 acts after the L2 vs. L2d decision, to regulate the L3 vs. dauer decision within the L2d stage. This possibility is suggested by the observation that, in contrast to daf-3 and daf-5 mutants that are defective in entry to L2d, dauer-defective daf-12 mutants enter L2d normally (Vowels and Thomas 1992). This idea is consistent with the model that TGFβ and IIS pathways affect daf-9 expression, which eventually translates into altered levels of DA and an effect on DAF-12 activity (Gerisch et al. 2001; Motola et al. 2006). During the L2d stage, the let-7-family microRNAs and LIN-42/Period fine-tune DAF-12 expression and activity, respectively, permitting high sensitivity to environmental conditions (Hammell et al. 2009; Tennessen et al. 2010). In wild-type larvae, the final decision to commit to dauer diapause is made during the early L2d molt (Golden and Riddle 1984). We hypothesize that at the late L2d or L2d molt stage, the lower levels of HBL-1 that remain act to oppose L3 cell fates, thereby postponing their expression. Along with opposing L3 cell fates during L2d, we suggest that HBL-1 would simultaneously promote dauer formation, perhaps by assisting with DAF-12–dependent dauer gene expression (red arrow, Figure 5). If this dauer-promoting activity of HBL-1 indeed occurs only when HBL-1 levels have been sufficiently lowered, that could help to link dauer commitment to a particular time in the L2d stage. Furthermore, the potential ability of HBL-1 activity to change from dauer opposing to dauer promoting over time could be important in the context of the rapidly changing environmental conditions that may be experienced in the wild.

The discovery that hbl-1 can modulate the dauer formation decision adds an important dimension to our understanding of the coordination between the regulation of dauer formation and the regulation of stage-specific cell fate decisions. As mentioned above, several other regulators of the L2 vs. L3 cell fate decision can also influence the dauer formation decision, including lin-42, let-7-family microRNAs, and the core dauer pathway component daf-12 (Antebi et al. 1998; Hammell et al. 2009; Tennessen et al. 2010). Interestingly, the effects of these genes on dauer formation are not necessarily predictable from their roles in regulating stage-specific cell fates. For example, while hbl-1 and lin-42 both promote L2 cell fates and oppose L3 cell fates, they appear to act in opposition with respect to modulation of daf-12 activity. Specifically, hbl-1 promotes dauer formation downstream or in parallel to daf-9 (Figure 1), whereas lin-42 opposes dauer formation downstream of daf-9 (Tennessen et al. 2010).

Additionally, lin-4, lin-14, and lin-28 influence different aspects of dauer formation (Liu and Ambros 1989). lin-14 encodes a transcription factor that regulates L1 vs. L2 and L2 vs. L3 cell fate decisions, lin-28 encodes an RNA binding protein that regulates the L2 vs. L3 cell fate decision, and lin-4 encodes a microRNA that targets both lin-14 and lin-28 (Ambros and Horvitz 1984, 1987; Lee et al. 1993; Moss et al. 1997; Hristova et al. 2005). lin-4(0) larvae are unable to enter dauer diapause, a phenotype that is due to persistent lin-14 expression. By contrast, no Daf-c or Daf-d phenotypes have been reported to result from loss of lin-14 or lin-28. However, lin-14 acts to restrict the entry into dauer diapause to the proper stage (following the L2 molt), and both lin-14 and lin-28 are required for proper dauer morphogenesis (Liu and Ambros 1989). The genetic relationships between lin-4, lin-14, and lin-28 are different for the dauer formation phenotypes than they are for stage-specific cell fate decisions (Ambros 1989; Liu and Ambros 1989). Therefore, while the modulation of the dauer formation decision by regulators of stage-specific cell fates is a common theme that is emerging, there are unique effects on dauer formation by specific genes. The complexity of the interactions described here may underscore the need for coordination between pathways that regulate developmental timing and pathways that respond to environmental conditions for development to occur robustly in changing environments.

Acknowledgments

We are grateful to Jason Tennessen and Ann Rougvie at the University of Minnesota for discussing their data prior to publication. We also thank Molly Hammell for statistical advice and Iva Greenwald for the use of reagents and equipment. This work was supported by National Institutes of Health (NIH) grants GM30428 (to V.A.) and F32 GM73307 (to X.K.). Some nematode strains used in this work were provided by the Caenorhabditis Genetics Center, which is funded by the NIH National Center for Research Resources. V.A. is a member of the UMass DERC (DK32520), which provided Core Resources.

Footnotes

  • Received May 17, 2010.
  • Accepted October 12, 2010.

References

View Abstract