Abstract
During embryogenesis, an essential process known as dosage compensation is initiated to equalize gene expression from sex chromosomes. Although much is known about how dosage compensation is established, the consequences of modulating the stability of dosage compensation postembryonically are not known. Here we define a role for the Caenorhabditis elegans dosage compensation complex (DCC) in the regulation of DAF-2 insulin-like signaling. In a screen for dauer regulatory genes that control the activity of the FoxO transcription factor DAF-16, we isolated three mutant alleles of dpy-21, which encodes a conserved DCC component. Knockdown of multiple DCC components in hermaphrodite and male animals indicates that the dauer suppression phenotype of dpy-21 mutants is due to a defect in dosage compensation per se. In dpy-21 mutants, expression of several X-linked genes that promote dauer bypass is elevated, including four genes encoding components of the DAF-2 insulin-like pathway that antagonize DAF-16/FoxO activity. Accordingly, dpy-21 mutation reduced the expression of DAF-16/FoxO target genes by promoting the exclusion of DAF-16/FoxO from nuclei. Thus, dosage compensation enhances dauer arrest by repressing X-linked genes that promote reproductive development through the inhibition of DAF-16/FoxO nuclear translocation. This work is the first to establish a specific postembryonic function for dosage compensation in any organism. The influence of dosage compensation on dauer arrest, a larval developmental fate governed by the integration of multiple environmental inputs and signaling outputs, suggests that the dosage compensation machinery may respond to external cues by modulating signaling pathways through chromosome-wide regulation of gene expression.
IN the nematode Caenorhabditis elegans, the DAF-2 insulin-like growth factor receptor (IGFR) ortholog promotes reproductive development and aging by inhibiting the FoxO transcription factor DAF-16 through the AGE-1 phosphoinositide 3-kinase (PI3K) and the conserved kinases PDK-1, AKT-1, and AKT-2 (Fielenbach and Antebi 2008; Kenyon 2010). daf-2 mutants were first isolated in genetic screens for dauer-constitutive mutants (Riddle et al. 1981). In replete environments, hatched embryos develop reproductively by traversing four larval stages (L1–L4) prior to adulthood. Under conditions of increased population density, reduced food availability, or elevated temperature, L1 larvae enter a distinct developmental pathway that culminates in arrest as an alternative, long-lived, morphologically distinct third larval stage known as dauer (Riddle 1988). daf-2/IGFR, age-1/PI3K, pdk-1, and akt-1 loss-of-function mutants all have dauer-constitutive phenotypes; i.e., they undergo dauer arrest under conditions in which wild-type animals develop reproductively (Riddle et al. 1981; Vowels and Thomas 1992; Gottlieb and Ruvkun 1994; Morris et al. 1996; Kimura et al. 1997; Paradis et al. 1999; Ailion and Thomas 2003). The dauer-constitutive phenotype of these mutants requires DAF-16/FoxO, as loss of daf-16/FoxO function fully suppresses dauer arrest in daf-2/IGFR, age-1/PI3K, pdk-1, and akt-1 mutants (Vowels and Thomas 1992; Gottlieb and Ruvkun 1994; Larsen et al. 1995; Paradis et al. 1999; Ailion and Thomas 2003). Taken together, these data indicate that the DAF-2/AGE-1/PDK-1/AKT-1 pathway promotes reproductive development by inhibiting DAF-16/FoxO.
Two other conserved signaling pathways play important roles in dauer regulation. The transforming growth factor-β (TGFβ)-like ligand DAF-7 (Ren et al. 1996) inhibits dauer arrest in parallel to the DAF-2/IGFR pathway by signaling through the type I TGFβ receptor homolog DAF-1 (Georgi et al. 1990) and the type II receptor homolog DAF-4 (Estevez et al. 1993) to regulate the SnoN homolog DAF-5 (Da Graca et al. 2004; Tewari et al. 2004) and the SMAD homologs DAF-3, DAF-8, and DAF-14 (Patterson et al. 1997; Inoue and Thomas 2000; Park et al. 2010). Downstream of the DAF-2/IGFR and DAF-7/TGFβ pathways, a hormone biosynthetic pathway consisting of DAF-36 (Rottiers et al. 2006), DHS-16 (Wollam et al. 2012), and DAF-9 (Gerisch et al. 2001; Jia et al. 2002) makes Δ7-dafachronic acid (Δ7-DA), a steroid ligand that prevents dauer arrest by binding to the DAF-12 nuclear receptor (Motola et al. 2006).
Although insulin- and insulin-like ligand-induced inhibition of FoxO transcription factors through nuclear export and cytoplasmic sequestration is a well-established mechanism of FoxO regulation, nuclear translocation is not sufficient to fully activate FoxO (Lin et al. 2001; Tsai et al. 2003). In C. elegans, the EAK proteins comprise a conserved pathway that acts in parallel to AKT-1 to control the activity of nuclear DAF-16/FoxO (Williams et al. 2010). eak mutations, while causing a weak dauer-constitutive phenotype in isolation, strongly enhance the dauer-constitutive phenotype caused by akt-1 mutations (Hu et al. 2006; Zhang et al. 2008; Alam et al. 2010; Dumas et al. 2010). EAK-7, which is likely the most downstream component of the EAK pathway, is a conserved protein of unknown function that is expressed in the same tissues as DAF-16/FoxO. Although EAK-7 likely regulates the nuclear pool of DAF-16/FoxO, it is situated at the plasma membrane (Alam et al. 2010), suggesting that it controls DAF-16/FoxO activity via unknown intermediary molecules.
We conducted a genetic screen to identify new FoxO regulators that may mediate EAK-7 action. Herein we describe our initial findings, which reveal an unexpected role for dosage compensation in controlling dauer arrest, insulin-like signaling, and FoxO transcription factor activity.
Materials and Methods
C. elegans strains and maintenance
The following strains were used in this study: N2 Bristol, CB4856 (Wicks et al. 2001), TJ356 [DAF-16::GFP(zIs356) IV] (Henderson and Johnson 2001), BQ1 akt-1(mg306) V (Hu et al. 2006), RB759 akt-1(ok525) V (Hertweck et al. 2004), VC204 akt-2(ok393) X (Hertweck et al. 2004), DR40 daf-1(m40) IV (Georgi et al. 1990), DR1572 daf-2(e1368) III (Kimura et al. 1997), CB1393 daf-8(e1393) I (Park et al. 2010), AA86 daf-12(rh61rh411) X (Antebi et al. 2000), DR77 daf-14(m77) IV (Inoue and Thomas 2000), CF1038 daf-16(mu86) I (Lin et al. 1997), AA292 daf-36(k114) V (Rottiers et al. 2006), CB428 dpy-21(e428) V (Yonker and Meyer 2003), and TY148 dpy-28(y1) III (Meyer and Casson 1986). The following mutant alleles were used: daf-9(dh6) and daf-9(k182) (Gerisch et al. 2001), eak-7(tm3188) (Alam et al. 2010), and sdc-2(y46) (Nusbaum and Meyer 1989). Double, triple, and quadruple mutant animals were constructed by conventional methods. Strains generated in this study are listed in Supporting Information, Table S4. Animals were maintained on nematode growth media (NGM) plates seeded with Escherichia coli OP50.
Suppressor of eak-7;akt-1 (seak) screen
eak-7(tm3188) and akt-1(ok525) both harbor deletion mutations that are likely to be null alleles (Hertweck et al. 2004; Alam et al. 2010); these alleles were chosen to minimize the chances of isolating informational suppressors such as tRNA anticodon mutants or mutants defective in nonsense-mediated mRNA decay. eak-7(tm3188);akt-1(ok525) double mutant animals were mutagenized at the mid-L4 larval stage using N-ethyl-N-nitrosourea (ENU) at a concentration of 0.5 mM in M9 buffer for 4 hr at room temperature, with gentle agitation (De Stasio and Dorman 2001). Mutagenized animals were plated on NGM plates and were allowed to recover overnight at 20°. After recovery, three P0 animals were plated on each of 14 plates and allowed to lay eggs for 6 hr at 20°. After egg lay, P0 animals were removed, and F1 animals were grown to adulthood at 20°. Ten gravid F1 animals were plated on each of 60 plates and allowed to lay eggs for 6 hr at 20°. After egg lay, F1 animals were removed, and plates containing F2 generation eggs were shifted to 25°. After 60 hr at 25°, all non-dauer F2 animals were picked, pooled on one plate corresponding to their F1 plate of origin, and grown to adulthood. Once gravid, these F2 dauer bypassors were singled to new plates, with up to eight animals singled per F1 plate of origin. A total of 330 F2 dauer bypassors were singled, approximately one third of which were sterile. Egg lays followed by dauer bypassor selection at 25° were repeated with this generation, and strains that showed highly penetrant dauer suppression in four subsequent generations were pursued further. In the case of multiple suppressed strains deriving from the same F1 plate of origin, only one strain was pursued to exclude the possibility of isolating sibling mutant animals. This selection scheme resulted in the isolation of 16 independent mutant strains with a verified eak-7;akt-1 dauer suppression phenotype. Approximately 1200 haploid genomes were screened.
Whole-genome sequencing
Genomic DNA was isolated from each of the 16 seak strains and the nonmutagenized eak-7;akt-1 double-mutant strain by phenol-chloroform extraction and subjected to whole-genome sequencing using the Illumina HiSeq2000 platform. Approximately 30–40× genome coverage was obtained using 100-nucleotide paired-end reads.
Sequence analysis
Sequence reads were mapped to the WormBase reference genome build WS220 using the short-read aligner BWA (Li and Durbin 2009). Single-nucleotide variants (SNVs) were identified with the help of the SAMtools toolbox (Li et al. 2009). Each SNV was annotated with a custom-made Perl script and gene information available from WormBase v. WS220. Sequences of the seak mutant strains were compared to that of the nonmutagenized eak-7;akt-1 parental strain. All nonsynonymous changes and predicted splice junction mutations were considered for subsequent analysis. A total of 664 such SNVs were identified among the 16 seak strains.
Mapping
Single-nucleotide polymorphism mapping was performed using the polymorphic Hawaiian CB4856 strain and a set of 48 primer pairs distributed throughout the genome (eight per chromosome) that flank DraI restriction site polymorphisms (Davis et al. 2005). Since the mutagenesis was performed on eak-7;akt-1 double-mutant animals, we constructed a recombinant inbred strain for mapping (BQ29 dpIr1 [N2 → CB4856, eak-7(tm3188)] IV; [N2 → CB4856, akt-1(ok525)] V). BQ29 was generated by outcrossing the original eak-7;akt-1 double mutant 10 times with CB4856 and identifying eak-7;akt-1 double-mutant recombinants harboring CB4856 polymorphisms at sites flanking eak-7 and akt-1. BQ29 animals arrested as dauers to the same extent as eak-7(tm3188);akt-1(ok525) double mutants in the N2 Bristol background, indicating that the CB4856 background does not strongly influence dauer arrest in eak-7;akt-1 double-mutant animals.
Mapping was performed by crossing seak mutant hermaphrodites (eak-7;akt-1;seak) with BQ29 males and isolating F2 dauers and bypassors. After confirming the dauer-constitutive and suppressor phenotypes in the following generation, F2 dauers and bypassors were pooled and assayed for all 48 DraI SNPs (Davis et al. 2005).
Dauer arrest assays
Dauer arrest assays were performed at the indicated temperatures in Percival I-36NL incubators (Percival Scientific, Inc., Perry, IA) as described (Hu et al. 2006). In assays involving daf-9(k182) and daf-36(k114) mutants, supplemental cholesterol was not added to the NGM assay plates (Gerisch et al. 2001; Rottiers et al. 2006). In assays involving daf-9(dh6) mutants, animals were propagated on NGM plates supplemented with 10 nM Δ7-DA to rescue the constitutive dauer phenotype of daf-9(dh6) (Sharma et al. 2009). Gravid animals raised on Δ7-DA were transferred to NGM plates without Δ7-DA and allowed to lay eggs for the dauer assay. Raw data and statistical analysis of all dauer assays are presented in Table S1.
RNAi
Feeding RNAi was performed using variations of standard procedures (Kamath et al. 2001). For dauer assays, NGM plates containing 5 mM IPTG and 25 μg/ml carbenicillin were seeded with 500 μl of overnight culture of E. coli HT115 harboring either control L4440 vector or indicated RNAi plasmid. Gravid animals cultured on NGM plates seeded with E. coli OP50 were picked to assay plates for 6-hr egg lays at 20°. Dauers were scored after progeny had been incubated at the assay temperature for 48–60 hr.
Male dauer assay
eak-7;akt-1 males were crossed with eak-7;akt-1 hermaphrodites. Mated hermaphrodites were picked for egg lays on RNAi plates. The dauer arrest phenotype of male and hermaphrodite cross progeny was scored. The sex of dauers was determined by allowing the animals to recover to adulthood.
RNA preparation and real-time quantitative PCR
A total of 100–200 adult animals were allowed to lay eggs for 6 hr at 20° on NGM plates seeded with E. coli OP50. After egg lay, adults were removed and eggs were shifted to 25°. L2 larvae were harvested 24 hr later, washed twice in M9 buffer, and then frozen in TRIzol reagent (Invitrogen, catalog no. 15596-026). RNA was extracted using TRIzol, cleaned using the RNeasy Mini Kit (Qiagen, catalog no. 74104), and treated with DNAse (Qiagen, RNase-Free DNase Set, catalog no. 79254). mRNA was reverse transcribed using the SuperScript First-Strand Synthesis System for RT–PCR and oligo(dT) primers (Invitrogen, catalog no. 11904-018). cDNA, 10 ng, was used as template for real-time quantitative PCR amplification using the Absolute Blue QPCR SYBR Green Mix (Thermo Scientific, catalog no. AB-4166/B) in a 15-μl reaction volume. Reactions were run on an Eppendorf realplex Mastercycler. Primer sequences are presented in Table S2. Relative expression levels of each gene were determined using the ΔΔ2Ct method (Nolan et al. 2006). Gene expression levels were normalized relative to actin (act-1) in the same sample, and then relative to the levels of the same gene in the control sample (wild-type/N2 Bristol or eak-7;akt-1 siblings, as indicated). Measurements were performed in triplicate, with three or four biological replicates for each condition.
DAF-16A:GFP localization
daf-16(mu86);zIs356;akt-1(mg306) animals (akt-1 mutants harboring a DAF-16A::GFP transgene as the only source of DAF-16) were cultured for two generations on dpy-21 RNAi or vector control at 20°. L2 or L3 larvae were mounted on glass slides using a thin pad of 2% agarose and 10 mM sodium azide to paralyze the animals. Images were captured using an Olympus BX61 epifluorescence compound microscope equipped with a Hamamatsu ORCA ER camera and Slidebook 4.0.1 digital microscopy software (Intelligent Imaging Innovations) and processed using ImageJ software. Images were captured within 20 min of mounting animals on slides to prevent variations in DAF-16A::GFP localization due to prolonged stress induced by mounting. Images were then blinded and animals sorted into categories 1–5 (Figure S2) based on the extent of nuclear localization of DAF-16A::GFP: 1 indicates that all cells have only cytoplasmic GFP, and 5 indicates that all cells have exclusively nuclear GFP.
Results
A genetic screen for novel DAF-16/FoxO regulators
We pursued a genetic approach to identify molecules that mediate EAK-7 regulation of DAF-16/FoxO. The strong dauer-constitutive phenotype of eak-7;akt-1 double mutants is fully suppressed by daf-16/FoxO loss-of-function mutations (Alam et al. 2010). We reasoned that screening for loss-of-function mutations that suppress the dauer-constitutive phenotype of eak-7;akt-1 double mutants might identify genes encoding novel DAF-16/FoxO activators. Therefore, we mutagenized eak-7(null);akt-1(null) double mutant animals with ENU and screened for rare F2 suppressors of the eak-7;akt-1 dauer-constitutive phenotype (seak mutants). Sixteen independent mutant strains were validated as bonafide seak mutants based on consistent high penetrance of the seak phenotype in subsequent generations.
To identify molecular lesions in these mutant strains, we isolated and sequenced genomic DNA isolated from nonoutcrossed seak mutants and the original nonmutagenized eak-7;akt-1 double mutant. Subsequently, each seak mutant was subjected to low-resolution single nucleotide polymorphism (SNP) mapping with the BQ29 recombinant inbred strain (Wicks et al. 2001; Davis et al. 2005).
dpy-21 inactivation suppresses dauer arrest in eak-7;akt-1 double mutants
Three independent seak strains harbored distinct nonsynonymous mutations in the conserved gene dpy-21 (alleles dp253, dp579 and dp381; Figure 1A). Each allele is predicted to be a missense mutation affecting a conserved residue in the conserved C-terminal region of DPY-21 (Figure 1A; Yonker and Meyer 2003). Low-resolution SNP mapping indicated that the genomic region containing dpy-21 is linked to the dauer suppression phenotype in each of these three strains. dpy-21 RNAi also suppressed the dauer-constitutive phenotype of eak-7;akt-1 double mutants (Figure 1B). Moreover, the dpy-21(e428) null allele [hereafter referred to as “dpy-21(null)”] (Yonker and Meyer 2003), as well as dpy-21(dp253), isolated from other mutagenesis-induced SNVs present in the original seak strain, also strongly suppressed the dauer-constitutive phenotype of eak-7;akt-1 double mutants (Figure 1C). Taken together, these data indicate that dpy-21 inactivation suppresses the dauer-constitutive phenotype of eak-7;akt-1 double mutants. Therefore, the three mutant alleles that emerged from our screen (Figure 1A) are likely dpy-21 loss-of-function mutations.
dpy-21 inactivation suppresses the dauer-constitutive phenotype of eak-7;akt-1 double mutants. (A) Genomic structure of the dpy-21 locus and locations of the e428 null allele and three alleles isolated in this study. Exons are denoted by boxes and introns by lines. (B and C) Effect of reduction of dpy-21 activity on the dauer-constitutive phenotype of eak-7;akt-1 double mutants at 25°. Error bars indicate SEM. Raw data and statistics are presented in Table S1. (B) dpy-21 RNAi suppresses the dauer-constitutive phenotype of eak-7;akt-1 mutants (13.4% mean dauer arrest in animals exposed to dpy-21 RNAi compared to 75.6% in animals exposed to control vector, P = 0.0004 by two-sided t-test). Data are from one representative experiment with >250 animals assayed per condition. (C) dpy-21(e428) and dpy-21(dp253) suppress the dauer-constitutive phenotype of eak-7;akt-1 mutants (0% mean dauer arrest in eak-7;akt-1 dpy-21(e428) triple mutants compared to 94.2% in eak-7;akt-1 double mutants, P < 0.0001; 0% mean dauer arrest in eak-7;akt-1 dpy-21(dp253) triple mutants, P < 0.0001). Data are pooled from three replicate experiments with at least 1000 animals assayed per genotype.
DPY-21 is a general regulator of dauer arrest
To determine whether DPY-21 influences dauer arrest generally as opposed to specifically in the context of eak-7 and akt-1 mutation, we determined the effect of reducing DPY-21 activity on dauer arrest in the background of other dauer-constitutive mutations using both dpy-21 RNAi and dpy-21(null) genetic mutation. Dauer arrest is controlled by three major signal transduction pathways; in addition to the DAF-2/IGFR pathway, the DAF-7 TGFβ-like pathway and the DAF-9 DA pathway also promote reproductive development by inhibiting dauer arrest (Fielenbach and Antebi 2008). We tested the effect of dpy-21 inactivation on dauer arrest in animals harboring mutations in each of these three pathways that confer a dauer-constitutive phenotype.
Reducing the activity of dpy-21 suppressed dauer arrest in animals with reduced DAF-2/IGFR signaling. dpy-21 RNAi modestly but significantly suppressed the 25° dauer-constitutive phenotype of animals harboring daf-2(e1368), which contains a missense mutation in the ligand-binding domain of DAF-2/IGFR (Kimura et al. 1997) (Figure 2A), and dpy-21(null) strongly suppressed daf-2(e1368) dauer arrest (Figure 2B). Additionally, dpy-21(null) suppressed the 27° dauer arrest phenotype of akt-1(ok525), a likely null allele harboring a 1251-bp deletion eliminating the kinase domain (Hertweck et al. 2004) (Figure S1A). These data indicate that DPY-21 promotes dauer arrest in the context of reduced DAF-2/IGFR signaling.
DPY-21 is a general regulator of dauer arrest. Effects of reducing dpy-21 activity on the dauer-constitutive phenotypes of DAF-2/IGFR pathway, DAF-7/TGFβ-like pathway, and dafachronic acid pathway mutants are shown. (A) dpy-21 RNAi suppresses the dauer-constitutive phenotype of daf-2(e1368) mutants (91.7% mean dauer arrest in animals exposed to dpy-21 RNAi compared to 95.2% in animals exposed to control vector, P = 0.0275 by two-sided t-test). Data are pooled from three replicate experiments with at least 1000 animals assayed per genotype. (B) dpy-21(null) suppresses the dauer-constitutive phenotype of daf-2(e1368) mutants (17.2% mean dauer arrest in daf-2;dpy-21 compared to 96.5% in daf-2, P < 0.0001). Data are pooled from three replicate experiments with at least 1300 animals assayed per genotype. (C) Influence of dpy-21 RNAi on the dauer-constitutive phenotype of mutants with reduced DAF-7/TGFβ-like pathway signaling (91.0% mean dauer arrest in daf-1(m40) mutant animals exposed to control vector compared to 89.3% in animals exposed to dpy-21 RNAi, P = 0.7064; 16.0% mean dauer arrest in daf-14(m77) mutant animals exposed to dpy-21 RNAi compared to 33.8% in animals exposed to control vector, P = 0.0321). Data are pooled from two replicate experiments with at least 550 animals assayed per condition. (D) dpy-21(null) suppresses the dauer-constitutive phenotype of daf-1(m40) and daf-14(m77) mutants (74.5% mean dauer arrest in daf-1;dpy-21 compared to 94.3% in daf-1, P = 0.01; 20.0% mean dauer arrest in daf-14;dpy-21 compared to 42.5% in daf-14, P = 0.0037). Data are pooled from three replicate experiments with at least 750 animals assayed per genotype. (E and F) Neither dpy-21 RNAi nor dpy-21(null) suppresses the dauer-constitutive phenotype of daf-9(dh6) mutant animals. (E) P = 0.2153 for dpy-21 RNAi compared to vector control. Data are pooled from three replicate experiments with at least 800 animals per condition. (F) P = 0.9503 for dpy-21(null);daf-9(dh6) compared to daf-9(dh6). Data are from one representative experiment with at least 250 animals per genotype. Error bars indicate SEM. Raw data and statistics are presented in Table S1.
In the TGFβ-like pathway, a ligand (DAF-7, Ren et al. 1996) and its receptors [the type I and type II TGFβ receptor-like molecules DAF-1 (Georgi et al. 1990) and DAF-4 (Estevez et al. 1993), respectively] promote reproductive development by regulating the activity of three SMAD-like molecules [DAF-3 (Patterson et al. 1997), DAF-8 (Park et al. 2010), and DAF-14 (Inoue and Thomas 2000)] and a Sno transcription factor homolog [DAF-5 (Da Graca et al. 2004; Tewari et al. 2004)]. dpy-21 inactivation also suppresses the dauer-constitutive phenotypes of mutants in the DAF-7/TGFβ-like pathway. Reducing dpy-21 activity modestly suppresses dauer arrest caused by the daf-1(m40) and daf-14(m77) nonsense mutations (Georgi et al. 1990; Inoue and Thomas 2000) (Figures 2, C and D), indicating that DPY-21 also promotes dauer arrest in the context of reduced DAF-7 TGFβ-like signaling. The effect of dpy-21 knockdown on the dauer-constitutive phenotype caused by the daf-8(e1393) missense allele is unclear; dpy-21 RNAi suppresses daf-8(e1393) dauer arrest, but dpy-21(null) slightly enhances daf-8(e1393) dauer arrest (Figure S1B).
In the dafachronic acid (DA) pathway, multiple biosynthetic enzymes [DAF-36 (Rottiers et al. 2006), DHS-16 (Wollam et al. 2012), and DAF-9 (Gerisch et al. 2001; Jia et al. 2002)] participate in the biosynthesis of Δ7-DA, a steroid ligand that promotes reproductive development by binding to the DAF-12 nuclear receptor (Motola et al. 2006). DAF-9 and DAF-12 act downstream of the DAF-2/IGFR and DAF-7/ TGFβ-like pathways to control dauer arrest (Gerisch et al. 2001; Jia et al. 2002; Gerisch and Antebi 2004; Mak and Ruvkun 2004). We tested the effect of dpy-21 inactivation on the dauer-constitutive phenotype of daf-9(dh6) animals, which harbor a null mutation in daf-9 that completely abrogates DA biosynthesis (Gerisch et al. 2001; Motola et al. 2006) [hereafter referred to as “daf-9(null)”]. Neither dpy-21 RNAi (Figure 2E) nor dpy-21(null) (Figure 2F) influenced the dauer-constitutive phenotype of daf-9(null) mutants at 25°. In contrast, dpy-21(null) modestly suppressed the 27° dauer-constitutive phenotypes caused by the hypomorphic daf-9(k182) mutation (Gerisch et al. 2001) and the daf-36(k114) null mutation (Rottiers et al. 2006), which reduces the biosynthesis of dafachronic acids (Wollam et al. 2011) (Figure S1C).
Overall, these data indicate that DPY-21 is important for dauer arrest in the context of reduced DAF-2/IGFR signaling and plays a modest role in controlling DAF-7/TGFβ-like and perhaps DA hormonal signaling.
Dosage compensation influences dauer arrest
DPY-21 was originally identified as one of 10 components of the C. elegans dosage compensation complex (DCC) (Yonker and Meyer 2003). This multiprotein complex mediates dosage compensation by assembling on both hermaphrodite X chromosomes to reduce X-linked gene expression by ∼50%, thereby equating X-linked gene expression between XX hermaphrodites and XO males (Meyer 2010). Based on its established role in dosage compensation, we sought to determine whether the suppression of dauer arrest by DPY-21 inactivation was a consequence of reduced dosage compensation per se as opposed to impairment of a DPY-21 activity that is independent of dosage compensation. To test this, we determined the effect of RNAi inactivation of each of the nine other DCC components on the dauer-constitutive phenotype of eak-7;akt-1 double mutants. Strikingly, we found that reducing the activity of most DCC components by RNAi suppresses the dauer-constitutive phenotype of eak-7;akt-1 double mutants (Figure 3A). It is noteworthy that, whereas loss-of-function mutations in most DCC components results in lethality (Plenefisch et al. 1989), RNAi-based inactivation of individual DCC components RNAi for multiple generations is not lethal (data not shown), suggesting that RNAi directed against these genes does not completely abrogate their function. Thus, the differential penetrance of dauer suppression observed after RNAi-based knockdown of DCC components could be a consequence of variable efficacy of RNAi against each targeted gene.
Dosage compensation influences dauer arrest. (A) Effect of RNAi of individual dosage compensation complex (DCC) components on the dauer-constitutive phenotype of eak-7;akt-1 hermaphrodites (black bars) and males (red bars). RNAi of most DCC components suppresses the dauer-constitutive phenotype of eak-7;akt-1 hermaphrodites (P < 0.05 for all DCC components except for sdc-1 and dpy-30, which were not significantly different from vector, as calculated by one-way ANOVA). Dauer arrest in male animals subjected to DCC component RNAi was not statistically different from vector control for any RNAi clone tested. In contrast, daf-16 and daf-12 RNAi caused significant suppression of dauer arrest in both hermaphrodite and male animals (P < 0.001 for comparison between daf-16 or daf-12 RNAi and vector control). Data are from a single experiment, representative of three replicates, with 100–800 animals scored per RNAi condition, per sex. (B) dpy-28(y1) suppresses the dauer-constitutive phenotype of eak-7;akt-1 double mutants (0.6% mean dauer arrest in dpy-28;eak-7;akt-1 compared to 99.5% in eak-7;akt-1, P < 0.0001). Data are pooled from three replicate experiments with at least 200 animals assayed per genotype. (C) sdc-2(y46) suppresses the dauer-constitutive phenotype of eak-7;akt-1 double mutants (0.1% mean dauer arrest in eak-7;akt-1;sdc-2 compared to 99.3% in eak-7;akt-1, P < 0.0001). Data are pooled from three replicate experiments with at least 700 animals assayed per genotype. Error bars indicate SEM. Raw data and statistics are presented in Table S1.
To further evaluate the role of other DCC components in dauer regulation, we reduced the activity of two DCC components via genetic mutation. y1 is a temperature-sensitive missense allele of dpy-28 (Plenefisch et al. 1989), which encodes a core DCC subunit (Tsai et al. 2008), and y46 is a weak allele of sdc-2 (Nusbaum and Meyer 1989), which encodes a hermaphrodite-specific protein that targets the DCC to X chromosomes (Dawes et al. 1999). In accordance with our RNAi-based findings, both dpy-28(y1) and sdc-2(y46) completely suppress the dauer-constitutive phenotype of eak-7;akt-1 double mutant animals (Figures 3B and 3C, respectively). Together, these results implicate the DCC itself in the control of dauer arrest.
Reducing DCC activity could suppress dauer arrest due to a reduction in dosage compensation. Alternatively, the dauer suppression phenotype could be a consequence of perturbing a novel, dosage-compensation-independent function of the DCC. To distinguish between these two possibilities, we took advantage of the observation that mutations that disrupt dosage compensation, while causing severe phenotypes in XX hermaphrodite animals, are phenotypically silent in XO male animals owing to the fact that the DCC does not repress X-linked gene expression in males (Meyer 2010). Therefore, we compared the effect of DCC component RNAi on the dauer-constitutive phenotype of eak-7;akt-1 hermaphrodites and males. In contrast to what was observed in eak-7;akt-1 hermaphrodites, DCC component RNAi had no effect on the dauer-constitutive phenotype of eak-7;akt-1 male siblings. In comparison, daf-16/FoxO and daf-12 RNAi fully suppressed dauer arrest in eak-7;akt-1 hermaphrodites and males (Figure 3A). Because DCC-mediated repression does not occur in males, these data suggest that the DCC controls dauer arrest through dosage compensation per se.
X-linked dauer inhibitory genes are upregulated in dpy-21 mutants
Dosage compensation could potentially influence dauer arrest by controlling the expression of X-linked genes that normally function to inhibit dauer arrest. C. elegans has nine known X-linked dauer regulatory genes, most of which encode proteins that inhibit dauer arrest: mrp-1 (Yabe et al. 2005), daf-3 (Patterson et al. 1997), pdk-1 (Paradis et al. 1999), ncr-1 (Li et al. 2004), daf-9 (Gerisch et al. 2001; Jia et al. 2002), ist-1 (Wolkow et al. 2002), daf-12 (Antebi et al. 2000), ftt-2 (Li et al. 2007), and akt-2 (Paradis and Ruvkun 1998). Six of these genes have previously been shown to be upregulated approximately twofold in embryos defective in dosage compensation (Jans et al. 2009) (Table S3). Notably, four of these genes, ist-1, pdk-1, akt-2, and ftt-2, encode components of the DAF-2/IGFR pathway, and the products of all four genes inhibit DAF-16/FoxO activity (Paradis et al. 1999; Paradis and Ruvkun 1998; Wolkow et al. 2002; Li et al. 2007).
We quantified expression of these nine dauer regulatory genes in eak-7;akt-1 dpy-21 triple mutant animals and their eak-7;akt-1 double-mutant siblings in animals raised at 25° for 24 hr after hatching (this time point corresponds to the L2 transition, following which animals commit to either dauer arrest or reproductive development). Expression of most of the X-linked dauer regulatory genes was increased approximately twofold in eak-7;akt-1 dpy-21 triple mutants compared to eak-7;akt-1 double-mutant siblings, consistent with their modulation by dosage compensation [Figure 4A, left, and Figure S3; (Jans et al. 2009)]. All four genes encoding DAF-2/IGFR pathway components that inhibit DAF-16/FoxO were upregulated in the context of dpy-21 mutation (ist-1, 2.87-fold increase in eak-7;akt-1 dpy-21 compared to eak-7;akt-1; pdk-1, increased 1.77-fold; akt-2, increased 3.46-fold; ftt-2, increased 1.57-fold). An increase in expression of these proteins would be predicted to inhibit DAF-16/FoxO activity by promoting its phosphorylation, nuclear export, and/or retention in the cytoplasm.
DPY-21 activates DAF-16/FoxO. (A) Effect of dpy-21(null) on the expression of X-linked and autosomal dauer inhibitory genes (blue bars) in larvae grown at 25° 24 hr after hatching. Data are normalized to expression levels of actin (act-1, orange bars), which is not X-linked. Expression in eak-7;akt-1 dpy-21 relative to expression in eak-7;akt-1 is shown. The dashed line indicates relative expression of one. Data are from a single representative experiment. Error bars indicate 95% confidence interval. Replicate data are presented in Figure S3. (B) DPY-21 promotes DAF-16/FoxO nuclear localization in an akt-1(null) background. Localization of DAF-16A::GFP was assessed in L2 or L3 daf-16(null);akt-1(null) double mutant animals after growth on E. coli harboring either control plasmid or dpy-21 RNAi plasmid for two generations. DAF-16A::GFP localization was scored in a blinded fashion on a 1–5 cytoplasmic/nuclear scale: 1, all cells in the animal have completely cytoplasmic localization; 2, most cells in the animal have completely cytoplasmic localization; 3, cells have both nuclear and cytoplasmic localization; 4, most cells in the animal have completely nuclear localization; 5, all cells in the animal have completely nuclear localization. See Figure S2 for representative images. Data represent three replicate experiments of ∼30 animals per condition. Error bars indicate SEM. (C) DPY-21 promotes DAF-16/FoxO target gene expression. sod-3, mtl-1, and dod-3 expression were quantified in larvae grown at 25° 24 hr after hatching. Data are normalized to expression levels of actin (act-1) and expressed relative to wild-type N2 Bristol expression. eak-7;akt-1(ok525) animals are siblings of eak-7;akt-1 dpy-21 animals. Data are from a single experiment. Error bars indicate 95% confidence interval. Replicate data are presented in Figure S4. (D) Reducing akt-2 gene dosage induces dauer arrest in eak-7;akt-1 dpy-21(null) triple mutants. Dauer arrest was assayed in the progeny of eak-7;akt-1 dpy-21(null);akt-2/+ parents compared to eak-7;akt-1 dpy-21(null) sibling controls. Progeny of eak-7;akt-1 dpy-21(null);akt-2/+ parents exhibited 14.72% dauer arrest. The dashed line indicates 25%. Data are pooled from two replicate experiments with at least 500 animals assayed per parental genotype. Error bars indicate SEM. Raw data and statistics are presented in Table S1.
Strikingly, the expression of daf-9, the cytochrome P450 family member that catalyzes the final step in DA biosynthesis (Motola et al. 2006), was increased in the context of dpy-21 mutation to a much greater extent than would be expected based on a reduction in dosage compensation. In four independent biological replicates, daf-9 expression increased ∼8- to 30-fold in eak-7;akt-1 dpy-21 triple mutants compared to eak-7;akt-1 siblings. This may be a consequence of suppression of the eak-7;akt-1 dauer-constitutive phenotype by dpy-21(null), as hypodermal daf-9 expression is induced in reproductively growing daf-2 mutant larvae but repressed in daf-2 mutant dauers (Gerisch and Antebi 2004). Notably, although many X-linked genes are upregulated approximately twofold in mutants defective in dosage compensation, the influence of the DCC on the expression of individual X-linked genes is highly variable (Jans et al. 2009).
To determine whether changes in the expression of autosomal dauer inhibitory genes contribute to suppression of eak-7;akt-1 dauer arrest by dpy-21 inactivation, we quantified the expression of 11 autosomal dauer inhibitory genes in eak-7;akt-1 dpy-21 triple mutants and their eak-7;akt-1 double mutant siblings (Figure 4A, right, and Figure S3). In contrast to X-linked dauer regulatory genes, expression of autosomal dauer inhibitory genes either did not change or increased only modestly in eak-7;akt-1 dpy-21 triple mutants compared to eak-7;akt-1 sibling controls (Figures 4A and Figure S3, compare left and right). This finding is consistent with a model whereby dpy-21 mutation suppresses dauer arrest via the perturbation of dosage compensation.
DPY-21 promotes DAF-16/FoxO nuclear localization
Our genetic analysis suggests that DPY-21 controls dauer arrest primarily by influencing DAF-2/IGFR signaling (Figure 2). The upregulation of four X-linked genes encoding DAF-2/IGFR pathway components that inhibit DAF-16/FoxO activity in the context of dpy-21 mutation (Figure 4A) is predicted to result in DAF-16/FoxO inhibition due to increased nuclear export and cytoplasmic sequestration of DAF-16/FoxO. If this is the major mechanism through which DPY-21 controls dauer arrest, then dpy-21 inactivation should promote the nuclear export of DAF-16/FoxO. To test this model, we determined the effect of dpy-21 RNAi on the subcellular localization of a functional DAF-16A::GFP fusion protein (Henderson and Johnson 2001b) in daf-16(null);akt-1(null) double-mutant animals (Figures 4B and Figure S2). As previously shown (Zhang et al. 2008; Alam et al. 2010; Dumas et al. 2010), DAF-16A::GFP was nuclear in many akt-1(null) animals cultured on E. coli containing a control RNAi construct. Exposure of animals of the same genotype to dpy-21 RNAi promoted the nuclear export and cytoplasmic retention of DAF-16A::GFP. Therefore, DPY-21 promotes the nuclear localization of DAF-16A::GFP.
DPY-21 activates DAF-16/FoxO
To determine the impact of dpy-21 inactivation on DAF-16/FoxO activity in dauer regulation, we quantified the expression of the DAF-16/FoxO target genes sod-3, mtl-1, and dod-3 (Murphy et al. 2003; Oh et al. 2006) in eak-7;akt-1 dpy-21 triple mutants and their eak-7;akt-1 double-mutant siblings grown at 25° for 24 hr after hatching. As previously shown (Alam et al. 2010), the expression of all three genes is increased in a daf-16/FoxO-dependent manner in eak-7;akt-1 double mutants (Figure 4C). dpy-21 null mutation strongly reduced the expression of at least two of the three genes in the eak-7;akt-1 double mutant background, although not to the same extent as a daf-16 null mutation. These results are consistent with a model whereby DPY-21 activates DAF-16/FoxO, as dpy-21 mutation results in significant but incomplete inhibition of DAF-16/FoxO activity.
The X-linked gene akt-2 is required for suppression of the dauer-constitutive phenotype of eak-7;akt-1 double mutants by dpy-21 mutation
If suppression of the dauer-constitutive phenotype of eak-7;akt-1 double mutants by dpy-21 mutation (Figure 1C) is due to DAF-16/FoxO inhibition secondary to increases in the expression of the X-linked genes ist-1, ftt-2, pdk-1, and akt-2 (Figure 4), then a reduction in the expression of ist-1, ftt-2, pdk-1, and/or akt-2 in eak-7;akt-1 dpy-21 triple mutants should restore the dauer-constitutive phenotype. To test this, we determined the effect of reducing akt-2 gene dosage on the dauer-constitutive phenotype of eak-7;akt-1 dpy-21 triple mutants.
We assayed the progeny of eak-7;akt-1 dpy-21 animals heterozygous for the akt-2(ok393) null mutation (eak-7;akt-1 dpy-21;akt-2/+) for dauer arrest at 25°. Approximately one-quarter of these animals should be eak-7;akt-1 dpy-21 triple mutants that are wild type at the akt-2 locus and should not undergo dauer arrest (Figure 1C). Half of the progeny should be eak-7;akt-1 dpy-21;akt-2/+, and the remaining quarter should be eak-7;akt-1 dpy-21;akt-2 quadruple mutants. Therefore, any evidence of dauer arrest in the progeny of eak-7;akt-1 dpy-21;akt-2/+ animals would indicate that suppression of eak-7;akt-1 dauer arrest by dpy-21 mutation requires akt-2.
Slightly less than one-quarter of the progeny of eak-7;akt-1 dpy-21;akt-2/+ animals underwent dauer arrest at 25° (Figure 4D). These dauers did not recover after several days of incubation at 15°, suggesting that they were all eak-7;akt-1 dpy-21;akt-2 quadruple mutant animals. Therefore, akt-2 activity is essential for the suppression of eak-7;akt-1 dauer arrest by dpy-21 null mutation. Reducing akt-2 gene dosage by twofold appears to be insufficient to induce dauer arrest in this context.
Discussion
Although much is known about the molecular components of the dosage compensation machinery and its influence on X-linked gene expression in C. elegans and other organisms, the physiologic consequences of dosage compensation are poorly understood. We have discovered a new function for dosage compensation in the control of dauer arrest, DAF-2/IGFR signaling, and DAF-16/FoxO activity.
We have shown that dpy-21 mutation suppresses dauer arrest due to a defect in dosage compensation (Figures 1, 2, and 3). Impaired compensation of X-linked gene expression results in the increased expression of dauer inhibitory genes, including four genes encoding DAF-2/IGFR signaling components that inhibit DAF-16/FoxO activity by promoting its nuclear export and cytoplasmic retention (Figure 4A). Accordingly, dpy-21 inactivation induces the relocalization of DAF-16/FoxO from nuclei to cytoplasm (Figure 4B) and inhibits the expression of DAF-16/FoxO target genes (Figure 4C). akt-2 is required for the suppression of eak-7;akt-1 dauer arrest by dpy-21 mutation (Figure 4D), indicating that dpy-21 inactivation suppresses the eak-7;akt-1 dauer-constitutive phenotype at least in part by promoting DAF-16/FoxO inhibition via AKT-2. Our results support a model in which DPY-21 (and presumably the DCC) controls dauer arrest and DAF-16/FoxO activity by modulating the expression of DAF-2/IGFR pathway components that influence the nucleocytoplasmic trafficking of DAF-16/FoxO (Figure 5). At this point, we cannot exclude the possibility that other X-linked and autosomal dauer inhibitory genes, the expression of some of which is increased in dpy-21 mutants (Figure 4A), may contribute to the suppression of eak-7;akt-1 dauer arrest by dpy-21 inactivation. This could explain the modest suppressive effect of dpy-21 inactivation on the dauer-constitutive phenotype of DAF-7/TGFβ-like pathway mutants (Figures 2C and 2D).
Model of DAF-2/IGFR pathway regulation by DPY-21. In eak-7;akt-1 double mutant animals, activated DAF-16/FoxO promotes dauer arrest (left). In eak-7;akt-1 dpy-21 triple mutant animals, increased expression of ist-1, pdk-1, akt-2, and ftt-2 contributes to suppression of dauer arrest by promoting the inhibition of DAF-16/FoxO (right).
The establishment of a role for dosage compensation in the control of dauer arrest and insulin-like signaling may serve as a platform for investigations into other postembryonic processes that might be influenced by dosage compensation, whether dosage compensation is physiologically regulated, and whether it is dysregulated in human disease. Indeed, a recent study in mice suggests that sex chromosome dosage per se can influence metabolic phenotypes independently of gonadal sex (Chen et al. 2012). Furthermore, skewing of X chromosome inactivation increases with age in human females and is attenuated in cohorts of female centenarians (Gentilini et al. 2012), suggesting a correlation between dysregulation of dosage compensation and the aging process. To date, studies on dosage compensation have largely focused on the establishment of dosage compensation during embryogenesis; mechanisms governing the postembryonic stability of dosage compensation remain poorly understood. Further inquiry into the function and regulation of dosage compensation in postembryonic contexts has the potential to illuminate new functions for dosage compensation and provide insights into the pathogenesis of human disease.
Acknowledgments
We thank Robert Lyons and Brendan Tarrier at the University of Michigan DNA Sequencing Core for their assistance with whole genome sequencing, Frank Schroeder for Δ7-DA, Michael Wells for helpful discussions, John Kim for comments on the manuscript, and Chris Webster and Alex Soukas for sharing data prior to publication. Some strains were provided by the Caenorhabditis Genetics Center (CGC), which is funded by the National Institutes of Health (NIH) Office of Research Infrastructure Programs (P40 OD010440). This work was supported by the Biology of Aging Training Grant (T32AG000114) awarded to the University of Michigan Geriatrics Center by the National Institute on Aging (K.J.D.), the Canadian Institute for Health Research (D.G.M.), Research Scholar Grant DDC-119640 from the American Cancer Society (P.J.H.), and a Kimmel Scholar Award from the Sidney Kimmel Foundation for Cancer Research (P.J.H.). D.G.M. is a senior fellow of the Canadian Institute for Advanced Research.
Footnotes
Communicating editor: D. I. Greenstein
- Received January 30, 2013.
- Accepted April 18, 2013.
- Copyright © 2013 by the Genetics Society of America
Available freely online through the author-supported open access option.
