HES-Mediated Repression of Pten in Caenorhabditis elegans

The hairy/enhancer-of-split (HES) group of transcription factors controls embryonic development, often by acting downstream of the Notch signaling pathway; however, little is known about postembryonic roles of these proteins. In Caenorhabditis elegans, the six proteins that make up the REF-1 family are considered to be HES orthologs that act in both Notch-dependent and Notch-independent pathways to regulate embryonic events. To further our understanding of how the REF-1 family works to coordinate postembryonic cellular events, we performed a functional characterization of the REF-1 family member, HLH-25. We show that, after embryogenesis, hlh-25 expression persists throughout every developmental stage, including dauer, into adulthood. Like animals that carry loss-of-function alleles in genes required for normal cell-cycle progression, the phenotypes of hlh-25 animals include reduced brood size, unfertilized oocytes, and abnormal gonad morphology. Using gene expression microarray, we show that the HLH-25 transcriptional network correlates with the phenotypes of hlh-25 animals and that the C. elegans Pten ortholog, daf-18, is one major hub in the network. Finally, we show that HLH-25 regulates C. elegans lifespan and dauer recovery, which correlates with a role in the transcriptional repression of daf-18 activity. Collectively, these data provide the first genetic evidence that HLH-25 may be a functional ortholog of mammalian HES1, which represses PTEN activity in mice and human cells.

the REF-1 family proteins fail to heterodimerize with other bHLH proteins and instead bind E-box-like sequences as homodimers (Grove et al. 2009;Powell et al. 2004). Of all the REF-1 family proteins, HLH-25 has the greatest degree of DNA-binding promiscuity and tightly binds to five E-box-like sequences. As a comparison, REF-1 binds tightly and HLH-26 binds weakly to the same E-box-like sequence, whereas HLH-29 only binds weakly to two sequences (Grove et al. 2009). Gene Ontology (GO) of the genes containing HLH-25 binding sites predicts that HLH-25 plays critical roles in cell division and development, including embryonic development, larval development, growth, and specification of cell fate (Grove et al. 2009).
HLH-25's putative role in regulating cell division and embryonic and postembryonic development supports previous studies suggesting that the C. elegans REF-1 family proteins are orthologs of the mammalian HES protein family (Neves et al. 2007). Like the other REF-1 family proteins, the sequence similarity between HLH-25 and the HES protein family lies within the bHLH domain. HLH-25 is more similar to HES1 than to the other mouse HES proteins but is most similar to the mouse Atonal homolog protein Math1 when amino acid residues flanking the bHLH domain are included in the analysis. Both Atonal1 and Hes1 play well-established roles in directing neuronal cell fate decisions (Akazawa et al. 1995;Ben-Arie et al. 2000;Ouji et al. 2013;Tomita et al. 2000;Wang et al. 2001), as do the respective, previously characterized C. elegans orthologs LIN-32 and LIN-22 (Miller and Portman 2011;Wrischnik and Kenyon 1997). Other studies suggest that the human and mouse HES1 proteins are oncogenic (Axelson 2004;Gao et al. 2014;Sang et al. 2010), underscoring the roles that these proteins play in regulating cell division and cell proliferation.
In this study, we describe a functional characterization of HLH-25 by using reverse genetics and molecular approaches. We show that hlh-25 is expressed postembryonically in body wall muscles and in neurons and that the phenotypes of hlh-25 animals include reduced brood size, unfertilized oocytes, and abnormal gonad morphology, all of which are consistent with defective cell-cycle or cell division events during embryogenesis. We show that the predicted HLH-25 transcriptional network, based on gene expression microarray, includes the Pten ortholog DAF-18 and other genes that mediate cell division, embryonic development, and larval development. Finally, we show that the ability of HLH-25 to facilitate dauer recovery and lifespan correlates with our mRNA studies suggesting that HLH-25 transcriptionally represses daf-18/Pten expression.

Microscopy
For imaging, animals were anesthetized either with 0.2% levamisole (Govindan et al. 2009) or with 10 mM NaN 3 , mounted on 2% agarose pads containing 10 mM NaN 3 , and imaged using either a 40· or a 60· water-immersion objective with the Nikon Eclipse 90i microscope equipped with a Nikon Coolsnap CCD camera. Z projections (sums of intensity) were obtained using NIS-Elements AR, version 4.0. For expression profiling, animals were cultured for at least two generations at 16°a nd then were synchronized and allowed to grow to the appropriate stage (L12L4). Forty animals per larval stage were examined for expression. For dauer stage expression, 10 starvation-induced dauer were imaged.
Egg-laying assays L4-stage animals were selected to individual NGM plates and allowed to molt to adulthood at 20°. At the start of the egg-laying period, each animal was moved daily to a new plate, already seeded with OP50, until 3 d after eggs or unfertilized oocytes were no longer detected. Eggs and oocytes were counted on the day that the hermaphrodite was removed. The plate was then incubated at 20°for 24 hr, and the numbers of live progeny and unhatched eggs were counted. Eggs that did not hatch within the 24-hr period were scored as dead, whereas the number of live progeny was checked again in the next 24 hr and then removed. For rescue assays, the ratio of unfertilized oocytes to fertilized eggs was calculated daily.

Lifespan assays
Lifespan assays were performed as previously reported (Quach et al. 2013). Lifespan survival comparisons were done individually for each of three biological replicates using the log rank test in GraphPad Prism 6 (GraphPad Software, Inc) and were then repeated with the data from all biological replicates combined into one experiment. Combined data are presented as a graph in Figure 5. Individual data are shown in Table S2.

Dauer recovery assays
To induce dauer formation, embryos were collected by hypochlorite treatment, washed extensively in water, and then approximately 200 embryos were placed onto an NGM plate containing a sparse lawn of OP50 (~10 mL of saturated OP50, 24 hr before use). Embryos were allowed to hatch, and the dauer larva to develop, undisturbed at 27°for at least 96 hr. Cultures were visually examined to ensure that at least 90% of the larvae were arrested as dauer. Failure to include OP50 on the plate resulted in animals that arrested as starved L1 stage larvae. For dauer recovery, 25 dauers per strain were individually moved to a plate with a thick lawn of OP50 and incubated at 20°. Dauers were examined every 3 hr during a 24-hr period for fat accumulation, pharyngeal pumping, crescent formation, and then at 8-hr intervals for another 24 hr for oocyte/embryo formation, and onset of egg-laying. The rates of onset of pharyngeal pumping were plotted as survival from dauer, and were compared by log-rank test using GraphPad Prism 6. The timing of dauer recovery varied with the day of the assay, with the onset of pumping occurring in 50% of wild-type animals as early as 3 hr or as late as 8.5 hr. In all assays the trend was the same, and the data shown are representative of one experiment.
Total RNA isolation, cDNA synthesis, reverse transcription, and quantitative PCR (qPCR) Total RNA extraction and cDNA synthesis were carried out as previously described (Quach et al. 2013) except total RNA was extracted from synchronized L4 stage animals. After cDNA synthesis, qPCR assays were performed with Taqman Gene Expression Assays (Invitrogen) for detection of amplicon, using relative quantitation with normalization against the endogenous control gene pmp-3 (Hoogewijs et al. 2008). Values shown represent the averages of at least three different experiments. The ID numbers for the Taqman Assays used in this study are available upon request.

Gene expression microarray
Gene expression microarray (GEO accession #GSE65417), including probe preparation, hybridization, fluidics run and chip scan, was performed by Georgia State University DNA/Protein Core Facility. Global gene expression in synchronized populations of VC1220 animals was compared to expression in N2 (wild-type) animals using GeneChip C. elegans Genome Array (Affymetrix). Data collection was carried out using GCOS 1.4 software (Affymetrix). Data analysis was performed using GeneSpring GX 11 Software (Agilent, Palo Alto, CA), and probe intensity values were normalized using robust multichip average algorithm. The quality controls on samples and on probe sets were performed stepwise to detect the outlying samples and the poor probe sets. The principal components analysis score plot and hybridization controls plot were applied for sample detection. Those probe sets passing minimal detection cutoffs and quality control Figure 2 Brood size and unfertilized oocytes hlh-25 (ok1710) animals. (A) hlh-25(ok1710) animals have fewer live progeny and lay more unfertilized eggs than wild-type animals. Expression of the extrachromosomal array cmjEx32, which carries hlh-25 genomic sequences, rescues this phenotype in hlh-25(ok1710) animals. Graph shows mean of total numbers of live progeny (dark bars) and unfertilized eggs (light bars) produced during the entire lifespans of wild-type (n = 7), hlh-25(ok1710) (n = 12), and hlh-25(ok1710; cmjEx32) animals (n = 9). Ã P-value , 0.05, ÃÃ P-value , 0.005. (B) Average numbers of live progeny (solid lines) and unfertilized eggs (dashed lines) produced per day of egg laying by the same hlh-25(ok1710) animals (triangles) and wild-type animals (circles) as those shown in (A). measurements subsequently were used for statistical analysis. The Student's t-test was performed to find the candidates for differential expression, and genes with significant signal level between different conditions (P , 0.05) were collected. In addition, fold change analysis were performed on the genes with significant expression, and all genes showing greater than two-fold change were considered putative targets.
Functional analysis by GO GO analysis was performed via the Database for Annotation, Visualization and Integrated Discovery (i.e., DAVID), version 6.7, to cluster related target genes (Huang da et al. 2009a,b). Additional GO terms and functional information for putative targets were assigned based on Wormbase annotations (Harris et al. 2010). Interactions and connectivity between target gene products were identified using the Search Tool for the Retrieval of Interacting Genes/Proteins (Szklarczyk et al. 2011) and redrawn manually.

Mobility assays
For locomotion assays, individual L4 stage animals were placed by platinum wire onto an unseeded NGM plate, allowed to recover for 5 min, and then examined after the start of forward locomotion for 30 sec. For thrashing assays, individual animals were placed into 50 mL of M9 medium. Each lateral movement of the head was counted as one thrash, and movement was monitored for one 30-sec interval per animal. Each biological replica included at least 25 animals per strain, and each assay was repeated for three biological replicas.

Data availability
The Gene expression microarray (GEO accession #GSE65417) will release for public access on January 1, 2016.

RESULTS
Phenotypes of hlh-25(ok1710) animals and Phlh-25::GFP activity suggest roles for HLH-25 in reproduction Previous studies indicate that hlh-25 is expressed in eight-cell stage embryos in response to Notch signalling, and in later stage embryos in response to GATA signalling (Broitman-Maduro et al. 2005;Neves and Priess 2005). Using an hlh-25::GFP transcriptional reporter cmjEx31 (supporting information, Figure S1), we detected postembryonic expression of hlh-25 at all larval and adult stages, including dauer stage, in unidentified neurons of the head and tail, the anal depressor muscle, anterior and posterior intestinal cells, and the head and body wall muscles of transgenic animals ( Figure 1). We detected expression in the PDE sensory neuron pairs and rare, transient expression in migrating distal tip cells of L3/L4 stage animals. We also detected stage-specific hlh-25 expression in multiple neurons of the ventral nerve cord in starvation-induced dauer stage animals. We were able to use cell body location and commissure morphology to identify three of the motor neurons as VD1, DA3, and DA7. Expression in the body wall and head muscles and in the PDE neuron pair was not evident in dauer larvae.
The hlh-25(ok1710) allele is a null allele that was generated by ethyl methanesulfonate mutagenesis. It is a 1550-bp deletion that spans from 292 bp upstream of the initiator codon to 348 bp downstream of the terminator codon ( Figure S1). We outcrossed these animals 10 times, confirmed homozygosity for the ok1710 allele by PCR, and then examined them for mutant phenotypes. hlh-25(ok1710) animals produced fewer live progeny at 20°than wild-type animals (Figure 2A, P-value = 0.0027) whereas laying more unfertilized oocytes (P-value = 0.00014) throughout the egg-laying period.
Adult hermaphrodites produce more oocytes than sperm and will continue to lay unfertilized eggs for a short period after depleting their sperm supply. Under conditions where food is plentiful, and when cultured at 20°, wild-type hermaphrodites lay eggs over the  first 5 d of adulthood (Ward and Carrel 1979;White et al. 2012). As previously described, we found that wild-type animals lay an increasing number of unfertilized eggs as the sperm supply diminishes. In our assays, wild-type animals continued to lay a small number of unfertilized eggs for one additional day after the active egg-laying period ended (egg-laying day 6), to produce an average total of 49 unfertilized oocytes ( Figure 2B, Table S1). hlh-25(ok1710) animals continued to produce unfertilized oocytes for 4 days after the active egg-laying period ended (egg-laying day 9), and produced an average total of 232 unfertilized oocytes. This phenotype has been recently named uno-o (Riesen et al. 2014), and is also seen in animals that carry null alleles of the REF-1 family protein HLH-29 (White et al. 2012). We were able to partially rescue the uno-o phenotype in hlh-25(ok1710) animals by expressing hlh-25 from an extrachromosomal array, cmjEx32, so that transgenic animals produced an average of 38% fewer unfertilized oocytes than hlh-25(ok1710) animals ( Figure 2A). We also examined the gonad arms of adult hermaphrodites. In wildtype animals, oocytes production occurs as progressive movement through two U-shaped gonad arms, starting with the production of germline nuclei at the distal end of each gonad arm, and ending with the ovulation of the oocyte into the sperm-filled spermatheca at the ends proximal to the uterus (Hubbard and Greenstein 2000). In the distal, syncytial region of each arm, the germline nuclei go through mitotic and meiotic divisions while being surrounded by a commonly shared cytoplasmic core. Nuclei start to compartmentalize into oocytes as they move progressively closer to a bend in the gonad arm, and are fully compartmentalized, or budded, as they leave the bend and proceed towards the spermatheca (Green et al. 2011). We found that organization of the gonad arm and germline progression were abnormal in 63% of hlh-25(ok1710) animals. As shown in Figure 3, and in Figure S2 and Figure S3, gonad architecture was variably abnormal and included constrictions in the distal and proximal gonad arms, increased apoptosis, increased oocyte number, delayed oocyte compartmentalization, abnormal chromatin condensation, and arrested embryos in the uterus.
Other phenotypes of hlh-25(ok1710) animals included decreased movement as monitored by thrashing and locomotion assays ( Figure S4), leading us to classify hlh-25 mutants as locomotion variants. This phenotype correlates with the expression of cmjEx31 in the body wall muscles and in head neurons, but we could not accurately assay for rescue of this phenotype by transgenic expression of hlh-25 because transgenic animals also carried the rol-6 marker, which causes a dominant locomotor defect. We did not pursue the locomotion phenotype further in this study.

DAF-18 is one major hub in the hlh-25 transcriptional network
We used gene expression microarray to compare the expression profiles of late L4/young-adult stage hlh-25(ok1710) animals to profiles of agematched, wild-type animals. We found that the expression of 630 genes, 506 of which were up-regulated and 124 of which were down-regulated, was altered by at least 2.0-fold (Table S3), and confirmed by reverse transcription (RT)-qPCR that the expression of 21 genes was significantly affected in L4-stage hlh-25(ok1710) animals (Table 1). As shown in Table 1, the direction of the change in expression was the same for 18 of the 21 genes. It is possible that the opposite, but significant change in the three remaining genes was due to differences in the developmental age of animals at the times of RNA extraction. Interestingly, 152 of these genes, indicated by asterisks in Table 1 and Table S3, contain predicted HLH-25 binding sites within their putative promoters (Grove ). Hypergeometric analysis, indicated that this overlap of 152 genes was significantly lower than expected when compared to an overlap between the list of genes identified by protein binding microarray to a randomly chosen gene set of the same size (P-value ,2.5 · 10 -05 ).
We clustered related target genes based on enriched GO terms. A total of 328 genes were organized into 10 annotation clusters that all fit within the broad category of "Development" (top eight clusters are shown in Figure 4A). Using the Search Tool for the Retrieval of Interacting Genes/Proteins (STRING 9.0) to identify predicted and known interactions between proteins encoded by the target genes (Szklarczyk et al. 2011), we found that the hlh-25 regulatory network has a high degree of overlap and interconnectivity. For example, 38 genes clustered under the GO term "Sexual Reproduction." Of those, 22 genes (58%) formed a single, highly interconnected, interaction network ( Figure 4B) and also were associated with at least two other clusters. This trend of high connectivity among genes within a cluster and significant overlap among genes between clusters was evident for the entire network. We also identified six genes that we considered to be major hubs (Table 2). These genes were associated with at least four of the GO clusters and formed predicted interactions with at least 35 other targets in the network. Importantly, the hub genes underscore the developmental influence of the hlh-25 regulatory network, encoding helicases, cyclins, and RNA-binding proteins with established roles in embryogenesis ( Table 2). The C. elegans homolog of the mammalian Pten tumor suppressor, DAF-18 , is one of the major hubs, forming connections with 62 proteins distributed across six annotation clusters ( Figure 4C, bold genes in Table S3).

HLH-25 functions upstream of DAF-18 to regulate dauer recovery
It is not clear whether the REF-1 proteins and the HES proteins arose from the same parent gene; however, conservation within the helixloop-helix domains and the ability to transduce Notch signals suggest that REF-1 proteins and HES proteins are members of the same functional family. Importantly, the identification of daf-18 as a major hub in the HLH-25 transcriptional network, together with the identification of predicted HLH-25 binding sites in the daf-18 promoter (Grove et al. 2009), underscores the functional homology between the HLH-25 and vertebrate HES1 protein. In the mouse thymus, HES1 acts downstream of Notch signalling to repress Pten activity during T-cell differentiation (Wong et al. 2012), and the resistance of T-cell acute lymphoblastic leukemia cells to Notch inhibition depends on functional HES1 binding sites in the Pten promoter (Palomero et al. 2007). Therefore, we further characterized the genetic interaction between the genes encoding the HES homolog HLH-25 and the PTEN homolog DAF-18.
Using RT-qPCR, we found that the levels of daf-18 mRNA increase greater than two-fold in embryos (data not shown) and in L4 stage animals (Table 1) compared with age-matched controls. The levels of daf-18 mRNA also increased in wild-type animals that were subjected to hlh-25 RNA interference ( Figure S5), suggesting that the increased daf-18 activity is due to loss of HLH-25 in hlh-25(ok1710) animals. We failed to detect differences in daf-18 transcriptional activity in wild-type and hlh-25 animals expressing Pdaf-18::GFP via a previously described, integrated transgene. These animals were generated in a daf-2 loss-offunction background, which increases the transcriptional activity of the daf-18 promoter . It is possible that loss of daf-2 masks the effects of HLH-25. Unfortunately, we have not yet been successful in generating stable transgenic lines expressing Pdaf-18:: GFP in animals that produce functional DAF-2.
In C. elegans, longevity and entry into dauer stage are controlled through the insulin signalling pathway, and depend on the insulin receptor DAF-2. Decreased insulin signalling, via loss-of-function mutations in daf-2, increases C. elegans lifespan and causes a dauerconstitutive phenotype. These phenotypes are suppressed in daf-18 animals: daf-18 animals have reduced lifespan  and animals that carry strong loss-of-function alleles are unable to undergo dauer development (Gil et al. 1999;Mihaylova et al. 1999;Rouault et al. 1999). The reduced lifespan of daf-18 animals can be rescued, and slightly extended beyond the wild-type lifespan, by the overexpression of daf-18 from the extra-chromosomal array quIs18 (Brisbin et al. 2009;Solari et al. 2005). We reasoned that since hlh-25(ok1710) animals have increased daf-18 activity, they should have an extended lifespan. As expected, the lifespan of hlh-25(ok1710) animals (median LS = 19 d) is significantly longer than the lifespan of wild-animals (median LS = 16 d, P-value ,0.0001) grown at 20°( Figure 5A, Table S2). This result is n comparable with the extended lifespan previously reported for C. elegans quIs18 females (Brisbin et al. 2009). Normally when wild-type, starvation-induced dauer larvae are moved to favorable growth conditions, they take approximately 60 minutes to commit to exiting the dauer stage at 25°, and they molt into the L4 stage at approximately 10 hr (Golden and Riddle 1982;Golden and Riddle 1984a,b,c). As animals begin to exit dauer stage and enter into the L4 molt, fat accumulation and pharyngeal pumping begin after approximately 60 min and 3 hr, respectively, of making the dauer recovery/exit decision (Proudfoot et al. 1993). The onset of vulva formation is detectable within 12 hr of the exit decision, and recovering dauers enter late L4/early adult stage within 24 hr of the decision to exit dauer stage.
Because daf-18(ok480) animals do not form dauers under starvation conditions, we used heat stress to induce dauer formation via an IISindependent pathway in wild-type, hlh-25(ok1710), daf-18(ok480), and quIs18 animals, and compared their dauer recovery phenotypes. We found that at least 80% of wild-type animals (n = 1199) reached early L4 stage within 12 hr after the switch to 20°. As expected, dauer recovery was similar to wild-type in daf-18(ok480) animals (n = 485) but was longer in quIs18 animals (n = 955). At 12 hr, approximately 65% of the daf-18(ok480) animals, but only 10% of the quIs18 animals, reached early L4. Interestingly, the dauer recovery timing of the hlh-25(ok1710) animals (n = 738) was very similar to the quIs18 animals: only 21% of hlh-25(ok1710) animals reached L4 stage within 12 hr, and only 45% reached L4 stage within 24 hr. This delayed dauer recovery phenotype in hlh-25(ok1710) animals was partially rescued by transgenic expression of hlh-25 (cmjEx32): 70% of transgenic animals (n = 300) reached L4 within 24 hr. For a more quantitative approach, we monitored the onset of pharyngeal pumping during dauer recovery. Dauer recovery in daf-18(ok480) animals was not significantly different from recovery in wild-type animals; however, it took 1.5 times longer for the onset of pharyngeal pumping to occur in 50% of hlh-25(ok1710) animals (P-value ,0.0001) than in wild-type animals ( Figure 5B). This delay was rescued by by transgenic expression of hlh-25 ( Figure 5C). These results, together with our gene expression data, suggest that HLH-25 acts upstream of daf-18 to negatively regulate lifespan and dauer recovery.

DISCUSSION
This study is a functional characterization of HLH-25, a member of the C. elegans REF-1 family of transcriptional regulators. Previous studies have shown that the gene hlh-25 is transcriptionally activated through Notch-signaling in cells that eventually give rise to the pharynx and nervous system, and through the Notch-independent GATA family of transcription factors in tissues that give rise to the pharynx, muscle, and somatic gonad. Here we show that expression of hlh-25 persists throughout the C. elegans life cycle. Although we identified a number of HLH-25 targets that are germline expressed and whose mRNAs are stored in oocytes before fertilization, we were unable to recover transgenic lines that report germline expression of hlh-25. Future studies may further define the role of HLH-25 in embryogenesis, which includes determining whether the maternal contribution of HLH-25 is required for proper embryogenesis.
Our results show that the phenotypes of hlh-25 animals include reduced brood size, unfertilized oocytes in the uterus, abnormal gonad morphology/architecture, extended lifespan, and delayed dauer recovery. In previous studies, functional annotation of candidate target genes suggested overlapping roles in cell signaling events for all of the REF-1 family proteins, and HES-like roles for HLH-25 and HLH-29 in regulating embryonic and larval development, growth, cell fate specification, and reproductive behaviors (Grove et al. 2009). Interestingly, this animals (n = 282) have longer than lifespans than wild-type animals (n = 264, P-value , 0.0001). (B) Representative experiment showing the percentage of animals with pharyngeal pumping during dauer recovery at 20°. P-value , 0.0001 and P-value = 0.938 when hlh-25(ok1710) animals and daf-18(ok480) animals, respectively, were compared with wild-type. (C) Representative experiment showing the percentage of animals with pharyngeal pumping during dauer recovery at 20°. P-value , 0.0001 and P-value = 0.2771 when hlh-25(ok1710) and hlh-25(ok1710;cmjEx32) animals, respectively, were compared with wild-type. functional overlap is suggested, but not clearly evident, in the phenotypes of REF-1 family mutants that have been characterized to date. Both ref-1 and hlh-29 animals, for example, show varying degrees of embryonic lethality that would ultimately lead to reduced brood size (Alper and Kenyon 2001;McMiller et al. 2007;White et al. 2012); however, the reduced brood size in hlh-25 animals in not due to lethality during embryogenesis, but to abnormalities that occur before fertilization. Likewise, mutations in ref-1, hlh-29, and hlh-25 cause morphologic defects that affect sexual reproduction. Loss of ref-1 affects cell fusion events during vulva morphogenesis (Alper and Kenyon 2001) and ray lineage decisions during morphogenesis of the male tail (Ross et al. 2005). Loss of either hlh-29 or hlh-25 results in unfertilized oocytes, abnormal gonad morphologies, and the exploded through vulva phenotype; however, the underlying mechanisms of these phenotypes appears to differ in hlh-25 and hlh-29 animals. The unfertilized oocyte phenotype in hlh-29 animals, for example, is caused in part by altered expression of genes required for inositol 1, 4, 5-triphosphate (IP 3 ) signaling, which affects the ability of oocytes and eggs to move through the spermatheca (White et al. 2012). On the basis of our observations of independent ovulation events, however, spermatheca function in hlh-25 animals appears normal, and genes required for IP 3 signaling are not affected by loss of HLH-25. These data suggest that the REF-1 family proteins do not act within a single transcriptional cascade to regulate specific events, and that each family member plays unique, stage-specific roles in C. elegans development.
Our gene expression microarray analysis suggests that there are six major hubs in the HLH-25 transcriptional network, including the C. elegans PTEN homolog, daf-18. We demonstrated by RT-qPCR that daf-18 expression is normally repressed by HLH-25. This result is the first direct correlation between a HES protein target gene in mammals and a REF-1 family target gene in C. elegans. Hes1 directs cell division, in part, by repressing Pten expression in mouse (Wong et al. 2012), in human large cell neuroendocrine carcinoma (Nasgashio et al. 2011), and in regenerating Clara cells of wild-type lungs (Xing et al. 2012). In C. elegans, DAF-18 is required during embryogenesis to maintain G2 arrest in germline precursor cells (Fukuyama et al. 2006) via inhibition of Tor complex 1 signaling (Fukuyama et al. 2012). We did not examine embryonic or L1-stage phenotypes of hlh-25 animals; however, we show that, like animals that overexpress daf-18 (Brisbin et al. 2009;Liu and Chin-Sang 2015), hlh-25 animals have an extended lifespan and take longer to exit dauer stage. Further examination of both the embryonic and postembryonic regulation of daf-18 by HLH-25 will further our understanding of the mechanisms and biological significance of HES-dependent activation and repression of PTEN.
C. elegans enter the dauer stage in response to adverse growth conditions, maintain this state while conditions remain undesirable, and then exit, or recover from dauer stage when environmental conditions are more favorable (Hu 2007;Ouellet et al. 2008). The decision and ability to enter into dauer stage requires the integration of dietary and environmental signals through multiple molecular pathways, including insulin/insulin-like growth factor, transforming growth factor-b, and guanylyl cyclase, and the molecular details of dauer entry in response to starvation, heat stress, and overcrowding have been well established (reviewed in Fielenbach and Antebi 2008). Less is known about the molecular requirements for dauer maintenance and dauer recovery. Studies suggest that GLP-12dependent Notch activation in chemosensory neurons is required for dauer maintenance (Ouellet et al. 2008) and that metabotropic acetylcholine signaling is required for the induction of an insulin-like signal during dauer recovery (Tissenbaum et al. 2000). Previous studies have also shown that dauer recovery requires LIN-122dependent Notch activation, though the downstream target of this signaling event has not been identified (Ouellet et al. 2008). Future studies will determine whether HLH-25 acts downstream of Notch to modulate dauer recovery by repressing daf-18 activity.
Finally, GO and protein interaction analysis of the HLH-25 network suggests that HLH-25 affects genes required for three highly connected processes: cell division, cell-cycle regulation, and reproduction, including gamete formation. The connectivity between the genes in these processes suggests that the HLH-25 regulatory network is amendable to change without the corresponding disruption in overall functionality. This feature is illustrated by the hub gene daf-18, and the hub genes cgh-1, cyb-2.1, cyb-3, mcm-3, and puf-3, which encode proteins required for early embryonic cytokinesis, meiosis II, G2/mitosis transition, licensing of DNA replication, and spindle positioning, respectively (Audhya et al. 2005;Boag et al. 2005;Hubstenberger et al. 2012;Kaitna et al. 2002;Sonneville et al. 2012;van der Voet et al. 2009). Although mutations in any of the hub genes, or in any of the nodes connected to them, result in severe phenotypes, including sterility, mutant animals are viable and have functional cell cycles. Future work will include further defining and validating the HLH-25 transcriptional network and correlating this network with the postembryonic functions of HLH-25.
Taken together, our results demonstrate that the C. elegans HES homolog, HLH-25, functions postembryonically to transcriptionally regulate genes required for cell division, cell-cycle regulation, and sexual reproduction, including the gene encoding the C. elegans Pten homolog, DAF-18.