Activated AKT/PKB signaling in C. elegans uncouples temporally distinct outputs of DAF-2/insulin-like signaling

Background In the nematode, Caenorhabditis elegans, a conserved insulin-like signaling pathway controls larval development, stress resistance and adult lifespan. AGE-1, a homolog of the p110 catalytic subunit of phosphoinositide 3-kinases (PI3K) comprises the major known effector pathway downstream of the insulin receptor, DAF-2. Phospholipid products of AGE-1/PI3K activate AKT/PKB kinase signaling via PDK-1. AKT/PKB signaling antagonizes nuclear translocation of the DAF-16/FOXO transcription factor. Reduced AGE-1/PI3K signaling permits DAF-16 to direct dauer larval arrest and promote long lifespan in adult animals. In order to study the downstream effectors of AGE-1/PI3K signaling in C. elegans, we conducted a genetic screen for mutations that suppress the constitutive dauer arrest phenotype of age-1(mg109) animals. Results This report describes mutations recovered in a screen for suppressors of the constitutive dauer arrest (daf-C) phenotype of age-1(mg109). Two mutations corresponded to alleles of daf-16. Two mutations were gain-of-function alleles in the genes, akt-1 and pdk-1, encoding phosphoinositide-dependent serine/threonine kinases. A fifth mutation, mg227, located on chromosome X, did not correspond to any known dauer genes, suggesting that mg227 may represent a new component of the insulin pathway. Genetic epistasis analysis by RNAi showed that reproductive development in age-1(mg109);akt-1(mg247) animals was dependent on the presence of pdk-1. Similarly, reproductive development in age-1(mg109);pdk-1(mg261) animals was dependent on akt-1. However, reproductive development in age-1(mg109); mg227 animals required only akt-1, and pdk-1 activity was dispensable in this background. Interestingly, while mg227 suppressed dauer arrest in age-1(mg109) animals, it enhanced the long lifespan phenotype. In contrast, akt-1(mg247) and pdk-1(mg261) did not affect lifespan or stress resistance, while both daf-16 alleles fully suppressed these phenotypes. Conclusion A screen for suppressors of PI3K mutant phenotypes identified activating mutations in two known pathway components, providing insights into their regulation. In particular, the interdependence of akt-1 and pdk-1, even in activated forms, supports the existence of AGE-1-independent pathways for these phospholipid-dependent kinases. Phenotypic analysis of these alleles shows that the larval and adult outputs of AGE-1/PI3K are fully separable in these mutants.


Background
In the nematode C. elegans, development, lifespan and stress resistance are regulated by a conserved insulin-like signaling pathway [1]. Under optimal conditions, C. elegans nematodes develop through four larval stages (L1-L4) to become reproductive adults. In adverse environmental conditions, such as low food abundance or high population density, larvae can enter diapause and arrest development as dauer larvae. Dauer larvae are non-feeding, long-lived, sexually immature and stress resistant animals. Morphologically they are thinner and darker than reproductive adults [2,3]. Once favorable conditions resume, dauer larvae can re-enter the reproductive developmental pathway to become fertile adults.
Most of the known daf-2 pathway components were identified in genetic screens on the basis of dauer arrest phenotype. In particular, a screen for suppressors of dauer larval arrest in animals with the null age-1(mg44) mutation uncovered hypermorphic alleles of akt-1 and pdk-1 [11,12]. In order to determine whether additional effectors of AGE-1/PI3K signaling could be recovered by this approach, we repeated this screen, using the missense age-1(mg109) allele [10].
Here, we describe five mutations identified in this screen, comprising two daf-16 alleles and new hypermorphic alleles of akt-1 and pdk-1 that appear to promote activation of AKT/PKB signaling in the absence of AGE-1/PI3K. The akt-1 and pdk-1 alleles also uncoupled the larval dauer arrest, adult longevity and stress resistance phenotypes of age-1(mg109) animals. We also describe an uncloned allele that has the unusual phenotype of enhancing the longevity of age-1(mg109) adult animals. These findings suggest that the larval and adult phenotypes of insulin signaling in C. elegans may be differentially affected by AKT/PKB signaling.

Results and discussion
Genetic identification of age-1(mg109) suppressors In C. elegans, both lifespan and dauer formation are regulated by age-1 activity [10,27]. Animals with maternal, but not zygotic, age-1 activity do not arrest as dauer larvae, but grow into reproductive adults that are stress resistant and long-lived. In the absence of both maternal and zygotic age-1 activity, animals arrest development as dauer larvae. In order to identify genetic modifiers of age-1, we conducted a screen for mutations that could suppress the dauer-constitutive (daf-C) phenotype of age-1(mg109), which carries a missense mutation in the lipid kinase domain of AGE-1/PI3K [10]. Previously, a similar genetic screen identified the AKT-1 and PDK-1 effectors of AGE-1/ PI3K [11,12]. We repeated this screen in an effort to identify additional mutations capable of activating signaling in the absence of AGE-1/PI3K. From a screen of approximately 20,000 haploid genomes, we identified 40 alleles that could suppress the dauer-constitutive phenotype of age-1(mg109) animals (see Methods). Five of these alleles were chosen for further study, based on high penetrance and strong suppression of age-1(mg109) constitutive dauer arrest. In particular, four of these alleles, mg242, mg255, mg261 and mg227, strongly suppressed the age-1(mg109) dauer-constitutive phenotype ( Table 1). The mg247 allele partially suppressed this larval phenotype, as age-1(mg109);mg247 animals bypassed dauer arrest but then developed into sterile adults (Table 1). Previous studies have shown that adult sterility is correlated with modest decrements in AGE-1/PI3K signaling [27,28]. In addition to the adult sterility phenotype, mg247 animals also exhibited a slight developmental delay, which has also been previously observed in insulin pathway mutants [19,27,28]. Thus, we conclude that the mg247 allele only partially restores phospholipid signaling in age-1(mg109) animals.
Using standard SNP snip-mapping protocols, we mapped four of these alleles to genetic regions that contained known components of the age-1 pathway [29]. Molecular lesions in candidate genes were identified by direct DNA sequencing (see Methods). Two alleles, mg242 and mg255, corresponded to nonsense mutations in daf-16 ( Figure 1A). daf-16(mg242) is a Trp220Amb nonsense mutation in the forkhead DNA binding domain of DAF-16 and truncates both major daf-16 splice forms, daf-16a and daf-16b, resulting in a predicted null phenotype. daf-16(mg255) carries a Trp144Amb nonsense mutation in the forkhead domain of the daf-16a splice form only (Figure 1A). Two other suppressor alleles corresponded to mutations in the serine/threonine kinases that are regulated by phosphoinositide (PIP) products of AGE-1/PI3K. The mg247 allele corresponded to a missense mutation in akt-1 (Ala149Thr). This mutation changes a nonconserved residue in the linker region between the pleckstrin homology (PH) and kinase domains of AKT-1 and is predicted to affect both akt-1a and akt-1b splice forms ( Figure 1B). The mg261 allele is the result of two mutations in different regions of PDK-1 of which both are missense mutations. The first corresponds to a nonconserved residue in the linker region between the kinase domain and the PH domain (Ala447Val) and the second in a conserved residue in the PH domain (Glu530Trp) ( Figure 1C). A fifth mutation, mg227, was mapped to a region of chromosome X that does not contain any known components of the daf-2 pathway. Two known dauer pathway genes, sgk-1 (X 23.70) and daf-12 (X 2.39), are also on chromosome X but are on the outside of the candidate region and thus do not correspond to the gene mutated by mg227 (see Methods) [7,8].

Dauer arrest suppressors differentially affect stress resistance and FIRE response
Reductions in AGE-1/PI3K signaling are associated with increased resistance to oxidative and thermal stresses [22,23]. We therefore examined the effects of dauer arrest suppressors on survival of age-1(mg109) animals under conditions of oxidative and thermal stress. Both daf-16 alleles, mg242 and mg255, completely suppressed oxidative stress resistance of age-1(mg109) animals tested by treatment with 10 mM paraquat, an intracellular free radical generator (Figure 5a). However, neither akt-1(mg247), pdk-1(mg261) or mg227 suppressed oxidative stress resistance of age-1(mg109) animals to paraquat. Double mutant age-1(mg109) animals carrying any of these suppressor alleles appeared to have increased survival in the presence of 10 mM paraquat, compared to parental age-1(mg109) animals (Figure 5a. Similar results were observed for suppression of thermotolerance of age-1(mg109) adults, as tested by survival at the stressful temperature of 35°C (Figure 5b).
We also analyzed the intestinal response to fasting in these strains by monitoring the FIRE response (Fasting-Induced Redistribution of Esterase activity) [32]. After a 6hour fast, intestinal esterase activity in wildtype adults redistributes from cytoplasmic to nuclear localization RNAi epistasis analysis of age-1(mg109) suppressors reveals interdependence of AKT-1 and PDK-1 Figure 3 RNAi epistasis analysis of age-1(mg109) suppressors reveals interdependence of AKT-1 and PDK-1. To establish whether akt-1(mg247), pdk-1(mg261) and mg227 mutations suppressed dauer arrest through the AGE-1/PI3 kinase pathway we carried out a series of genetic epistasis experiments using RNAi. Bars show the percentage of animals that bypassed dauer arrest, developing into L4 larvae or adults (grey), and those that arrested as dauers (black). Numbers on the side of the bars represent the total number of animals examined. RNAi experiments were diluted with control L4440 RNAi to match the mixed akt-1 and akt-2 RNAi dose. (A) age-1(mg109);akt-1(mg247) animals arrested as dauer larvae on pdk-1 RNAi. Similarly, age-1(mg109);pdk-1(mg261) animals arrested as dauer larvae on akt RNAi. (B) Epistasis analysis on age-1(mg109);mg227 animals suggests mg227 requires akt-1 but not akt-2, sgk-1 or pdk-1 to bypass dauer larval arrest.
In C. elegans, signaling by the DAF-2/insulin receptor, via AGE-1/PI3K, regulates developmental arrest at the dauer larval stage, as well as longevity and stress resistance in adult animals. Here, we report the characterization of new mutations identified through a genetic screen designed to identify components of the AGE-1/PI3K signaling pathway. This screen identified 2 mutations inactivating one or two isoforms of daf-16, which encodes the FOXO transcription factor that is the ultimate output of AGE-1/PI3K signaling. Both of the daf-16 alleles suppressed all the age-1(mg109) dauer and adult phenotypes examined, consistent with the conclusion that daf-16 activity is necessary for both the larval and adult phenotypes of age-1 mutant animals. Additionally, we identified 3 mutations that uncoupled the larval and adult phenotypes of AGE-1/PI3K signaling. These mutations appear to activate the AKT/ PKB effector pathway of AGE-1/PI3K signaling, as evidenced by the dependence of these 3 alleles on wildtype akt-1 and/or pdk-1 activity. Previous reports also identified activating mutations in akt-1 and pdk-1, yet ours is the first report to probe in depth the effects of this class of mutations on adult AGE-1/PI3K functions [11,12]. In addition, we report the phenotypic characterization of a new mutation, mg227, which has the unusual phenotype of enhancing adult longevity in age-1(mg109) animals in addition to suppressing larval dauer arrest.
In this work, we identified an activating mutation in the AKT-1 linker region between the PH and kinase domains. Interestingly, the previously identified mg144 allele also contained a linker region mutation (Ala183Thr) [12]. The linker region in mammalian AKT/PKB proteins is considerably shorter than in the C. elegans protein, so it is not clear whether linker region function is conserved in the C. elegans and human proteins or whether the C. elegans linker region possesses additional functions. The structure of the PH domain from human AKT has been determined [33]. This AKT fragment contained part of the linker region which appeared to face away from the residues that mediate phospholipid binding and probably does not play a significant role in determining lipid binding specificity. Instead, linker region mutations have been postulated to relax autoinhibition of the kinase and PH domains in the inactive AKT protein [12]. Interestingly, our epistasis analysis showed that the gain-of-function phenotype of akt-1(mg247) required the presence of wild type pdk-1. This suggests that the mutant protein encoded by akt-1(mg247) may be more easily phosphorylated by PDK-1 in the absence of AGE-1/PI3K-generated phospholipids We also identified an activating allele of pdk-1, encoding the kinase that phosphorylates and activates phospholipid-bound AKT/PKB. The mg261 mutant contains mutations of both a nonconserved residue in the linker region  [27]. c Although the akt-1(mg247) allele increased mean lifespan compared to age-1(mg109) control animals, the median lifespan and maximum lifespan was not significantly affected (age-1(mg109) a, b , 25 days median, 32 days maximum; age-1(mg109);akt-1(mg247) a , 24 days median, 33 days maximum, Figure 4) d Wilcoxan P value = 0.2370. f Lifespan assays performed at 20°C. g P vs age-1(mg109) a, b SE, Standard error. n indicates the number of animals scored in at least three independent trials. and a conserved residue in the PH domain which may mediate binding to the phospholipid products of AGE-1/ PI3K [34,35]. A previously identified activated pdk-1 allele, mg142, corresponded to an Ala303Val substitution in the PDK-1 kinase domain and was shown to increase the basal level of PDK-1 kinase activity [11]. The regulation of PDK-1 signaling is complex and depends on the downstream target in question [36]. In C. elegans, the major targets of PDK-1 are believed to be AKT-1, -2 and SGK-1 [8,11,12]. One possibility is that the mg261 mutation in the PH domain could alter the specificity of PDK-1 phospholipid binding, enhancing PDK-1 activity in the absence of AGE-1/PI3K phospholipid products. However, PDK-1 is active in age-1(mg109) as evidenced by age-1(mg109);akt-1(mg247) mutants. Therefore it is not entirely clear how the pdk-1(mg261) allele is activating AKT-1.
daf-16 activity is required for both the larval dauer arrest and adult longevity and stress resistance phenotypes that result from genetic deficiencies in the age-1 pathway [21,27]. Consistent with previous findings, the daf-16 alleles described here suppressed both larval and adult phenotypes of age-1(mg109) animals. In contrast, the akt-1(mg247), pdk-1(mg261) and mg227 alleles did not affect the age-1(mg109) adult phenotypes. Thus, the larval and adult outputs of AGE-1/PI3K signaling are differentially affected by akt-1(mg247), pdk-1(mg261) and mg227 alleles. It was previously shown that pdk-1 activity is necessary for wildtype adult lifespan, demonstrating that PDK-1 mediates the adult outputs of AGE-1/PI3K signaling [11]. Thus, we propose that the differential suppression by pdk-1(mg261) of the age-1(mg109) larval versus adult phenotypes results from insufficient activation of PDK-1 activity for adult, but not larval, phenotypes.
pathways that converge with AGE-1/PI3K signaling to regulate DAF-16 activity and, ultimately, lifespan in adult animals.
One possible explanation for the different effects of these mutations on the larval versus adult phenotypes of age-1(mg109) animals is that they partially restore insulin pathway signaling and negative regulation of DAF-16. According to this model, wildtype levels of insulin pathway signaling may strongly inhibit DAF-16 and promote reproductive development and adult lifespan. Whereas, low insulin signaling promotes both dauer larval arrest and adult longevity [19][20][21]. The mutations activating AKT signaling appear to promote intermediate levels of insulin pathway signaling, allowing reproductive development, but prolonging adult lifespan. This phenotypic gradient could reflect a gradient of DAF-16 activity in these strains, in which the highest levels of DAF-16 activity causes dauer larval arrest (Figure 8). Thus the mutations activating AKT signaling may allow an intermediate level of DAF-16 activity, which is insufficient to promote dauer larval arrest, but does promote adult longevity. This is similar to a DAF-16 gradient model that has been previously proposed [8].

Conclusion
In this study we identified mutations capable of activating signaling in the absence of AGE-1/PI3K. We describe five mutations including two mutations inactivating one or two isoforms daf-16 and two new hypermorphic alleles of akt-1 and pdk-1. The daf-16 alleles described in this work suppress age-1(mg109) dauer and adult phenotypes examined consistent with previous reports. The akt-1 and pdk-1 alleles uncouple the dauer arrest, adult longevity and stress resistance phenotypes of age-1(mg109) mutants. Furthermore, genetic epistasis analysis suggests these hypermorphic alleles are interdependent. These findings suggest larval and adult phenotypes of insulin signaling are fully separable in these mutants.
Genetics Center at the University of Minnesota. All strains were cultivated at 15°C on NGM agar plates seeded with the E. coli strain OP50 following standard protocols [37].

RNA interference
A 500 bp region spanning the largest exon in both akt-2 and pdk-1 was amplified from N2 DNA and cloned into the RNAi feeding vector L4440 [38]. akt-1 (RNAi) and daf-16 (RNAi) was obtained from the RNAi feeding library [30]. LB broth supplemented with 100 μg/mL of ampicillin was inoculated with RNAi clones and grown overnight at 37°C without shaking. RNAi experiments were typically carried out on 6-well plate NGM media supplemented with 50 μg/mL ampicillin and 1 mM IPTG. Overnight bacterial cultures were seeded on duplicate wells and allowed to grow at room temperature for a further 12 -16 hr. G 1 and G 2 progeny of fertile hermaphrodites were collected in a 5-hour egg lay on RNAi bacteria. The G 2 progeny were allowed to grow at 20°C for 3 days after which any phenotypes were scored. Plates were reexamined 24 hr later for changes in phenotypes if any. Each experiment was done in triplicate.

Phenotypic assays
To assay dauer larval arrest, gravid adult hermaphrodites were allowed to lay eggs at room temperature for 4 hours to produce a synchronized population. Animals were then shifted to 25°C or 20°C and left to develop for a further 72 or 96 hours after which the number of dauers and adults (sterile or fertile) was counted. Sterility was scored as the absence of embryos in the uterus in young adults (day 1-2).
For lifespan analysis, larvae were raised at 15°C until the first day of adulthood (Day 0) prior to being shifted to 25°C on standard NGM plates supplemented with FUDR (50 μg/mL). Survival analyses were always performed in the presence of FUDR, since some of these strains have an egg-laying defect that causes eggs to accumulate within the uterus and hatch, thereupon killing the parent hermaphrodite. FUDR treatment helps to remedy this problem as it is an inhibitor of DNA synthesis and prevents embryonic development. Animals were scored every two days for death by failure to respond to gentle prodding on the head and tail. All aging experiments were repeated in at least three separate trials and cumulative results were analyzed. Lifespan statistical analysis was performed using the JMP software package (version 5.1).
For paraquat and thermotolerance survival, young adults were obtained as for lifespan assays. For paraquat treatments, day 0 adults were shifted to standard NGM plates supplemented with 10 mM paraquat at 20°C. Survival of Model for regulation of insulin pathway phenotypes by AKT-1 and PDK-1 activation Figure 8 Model for regulation of insulin pathway phenotypes by AKT-1 and PDK-1 activation. Activated signaling components in the absence of AKT-1 signaling in age-1(mg109) mutant background. See text for details.
animals was scored at 2 or 3 day intervals by failure to respond to gentle prodding. For thermotolerance assays, synchronized young gravid adults were placed on bacteria-free NGM plates pre-heated to 35°C. All plates were wrapped in Parafilm to prevent dessication and placed in a 35°C incubator for 16 hours. Survival was scored by failure to respond to gentle prodding after 16 hours. FIRE response was assayed by scoring cytoplasmic versus nuclear intestinal esterase activity, as described previously [32]. Briefly, stress due to food deprivation was achieved by transferring bacteria-free young adult animals to NGM agar with 100 μg/mL ampicillin and incubating at 25°C for 6 hours. Fasted adults were fixed with methanol (-20°C) for 10 minutes and incubated overnight at room temperature in staining solution described previously [32]. Stained specimens were mounted on 2 % agarose pads and viewed on a Nikon E800 microscope with Nomarski optics. Images were collected using a Hamamatsu ORCA-ER CCD camera using OpenLab software (Improvision, Lexington, MA).

DAF-16:GFP localization
A gfpdaf-16 cDNA fusion was constructed using an intestinal gly-19 promoter as previously described [32]. In brief, the DAF-16:GFP plasmid was constructed by inserting a gfp coding sequence in-frame with the daf-16a full-length cDNA. The intestinal promoter was cloned upstream of gfpdaf-16 using unique SphI and KpnI sites. Plasmid were injected into adult hermaphrodites at a final concentration of 25-100 μg/mL [32].