Functions as a Positive Regulator of Growth and Metabolism in

Summary

Size and growth of an organism are adjusted to nutritional conditions by a complex regulatory network involving the Insulin receptor and TOR signaling cascades. Drosophila melanogaster has been used in the past as a genetically tractable model to unravel the complex circuitry by genetic means. We have identified CycG as an important player in the regulation of TOR signaling. CycG mutants are underweight in the midst of food and show typical signs of TOR defects. We provide evidence that CycG acts at the level of Akt1 kinase that links the Insulin receptor and TOR signaling cascades. Molecular and genetic data point to an interplay of CycG and phosphatase PP2A, a well established negative regulator of Akt1 activity. Moreover, CycG may influence PP2A-Akt1 binding. We propose that CycG, by impeding PP2A-Akt1 interaction, acts as a positive regulator of growth in Drosophila.

Introduction

The growth of an organism is a highly coordinated process regulated by a wide range of different inputs. Members of the Insulin receptor (InR) and Target of rapamycin (TOR) signaling pathways are well established key players in the control of cell growth in higher eumetazoa. Studies in different organisms support the idea that this signaling network modulates cellular growth in response to nutrient availability, growth factor signaling, energy status as well as to diverse cellular stressors (for review [1], [2]). Drosophila has proven to be a powerful system for investigating the InR/TOR signaling network. Signaling through the InR pathway is triggered through the binding of Drosophila Insulin-like peptides (dILPs) to the single InR. Four of the eight known dILPs (dILP1, 2, 3 and 5) are expressed in neurosecretory cells of the brain, the so-called Insulin producing cells (IPCs), from which they are released to act systemically via haemolymph transport [3]-[7]. Activation of the InR triggers a phosphorylation cascade mediated by a relay of kinases. As one of the first steps, the lipid kinase Phosphoinositide 3-kinase (PI3K92E) is activated leading to the activation of the kinase Akt1 that in turn phosphorylates the small GTPase Rheb (Ras homologue enriched in brain), an activator of TOR (for review [2], [8]). The TSC1/2 (tuberous sclerosis complex) tumor suppressor complex inhibits the activity of the TOR kinase by negatively regulating Rheb [9]-[13]. In addition, phosphatases like PTEN and PP2A were identified as negative regulators of the InR/TOR signaling cascade [14]-[18].

Starvation, especially amino acid withdrawal, interferes with dILP secretion in larval brains: dILPs accumulate in IPCs, and larval growth is impaired [4]. Nutrient availability is sensed by the TOR network which serves as the central coordinator of cellular and organismal growth, aging and fertility [19]-[21]. The TOR kinase, central to the TOR pathway, exists in two distinct conserved complexes, TORC1 and TORC2. Like in mammals, TORC1 is the crucial regulator of cell size and organismal growth in Drosophila (for review [2], [22]). The best studied substrates for TORC1 are S6 kinase (S6K) and the eukaryotic translation initiation factor 4E binding protein (4E-BP), both serving the regulation of translation (for review [22]). Phosphorylation of either protein enhances translation efficiency, either by relief of translational repression as in the case of 4E-BP, or by enhancement of ribosome recruitment as in the case of S6K (for review [1], [23]). This spectrum of phenotypes conforms to the pivotal role of TOR signaling in the control of growth and maintenance of cellular homeostasis in synchrony with the actual nutrient conditions. In Drosophila the larval fat body, a functional equivalent of the vertebrate liver and white adipose tissue, acts as a nutrient sensor controlling dILP release in the brain [4], [21]. In accordance, reducing TOR signaling specifically in the fat body has a negative impact on the overall growth of the animal comparable to the effects observed in underfed larvae [4].

Here we identify Cyclin G (CycG) as a new regulator of InR/TOR signaling activity in Drosophila. Homozygous cycG mutant flies are viable, however females are sterile. Mutant eggs display dorso-ventral patterning defects in the eggshell due to an impaired EGFR-signaling activity. This phenotype was shown to be a consequence of compromised double strand break repair, assigning CycG a role in meiotic checkpoint control during oogenesis [24], [25]. Moreover, CycG was proposed to act as a negative regulator of cell growth and cell cycle progression based on the misregulation of CycG activity [26]. Here we report that the phenotypes of cycG mutants recapitulate defects in InR/TORC1 signaling. Our genetic and molecular data indicate that CycG acts at the level of Akt1 presumably via a regulation of PP2A-Akt1 binding. Altogether our genetic and molecular observations provide evidence that CycG is required for InR/TORC1 pathway members to tap their full potential in mediating growth and metabolism in Drosophila.

Results

cycG mutants display defects in growth regulation

Homozygous cycG^HR7 null mutants are viable but female sterile [24]. They are, however, developmentally delayed and underrepresented with regard to their siblings (Fig 1A). In addition, cycG^HR7 mutant animals are smaller and slimmer than the controls (Fig 1B), and have a reduced body weight (Fig 1C). This finding was unexpected as CycG was reported to function as a negative regulator of growth and proliferation based on overexpression studies [26]. We also observed that the ubiquitous overexpression of CycG (da::CycG) resulted in a weight loss, however, was able to ameliorate the weight deficit of the cycG^HR7 homozygotes (Fig 1C). Because of its ability to bind to several cyclin dependent kinases [26]-[28], a strong CycG overexpression is likely to interfere with cell cycle regulation, which may explain these observations. In accordance, the subtle induction of a heat shock CycG construct (hs-CycG) at ambient temperature was sufficient to robustly rescue the observed growth and weight deficits in cycG^HR7 mutant animals (Fig 1B and 1D).

The size defect of the cycG^HR7 mutant animals was further studied in the wing. Here the reduced size was associated with a reduction in cell size and cell number (quantified via trichome density) (Fig 1E), pointing to a defect in InR/TORC1 signaling [29]. Next we induced mutant clones in the developing imaginal tissue by Flp/FRT mediated mitotic recombination [30]. By 72 hours of larval development, the majority of cycG^HR7 mutant cell clones was smaller than their wild type twin spots in eye-antennal as well as in wing discs (Fig 1F and 1F'), indicating a cell autonomous requirement of CycG for normal growth.

In order to exclude second site defects in the cycG^HR7 allele, a second independent cycG allele was generated by ‘ends out’ homologous recombination [31]: in the resultant allele cycG^eoC, nearly all of the coding region is deleted (S1A Fig). Like cycG^HR7, the cycG^eoC mutant as well as the transheterozygous cycG^HR7/cycG^eoC combinations behave also as protein null on western blots (S1B Fig) and display the same phenotypic characteristics as cycG^HR7, i.e. female sterility with defective egg patterning, developmental delay and a reduced body size and weight (S1C and S1D' Fig). Taken together, our results support a role for CycG as a positive effector of growth/weight control in the fly.

The fat metabolism is disturbed in cycG mutants

The distinct weight reduction of cycG mutant flies suggested a defect in metabolic homeostasis. Weight is a parameter that is directly correlated with food intake and metabolism. However, the ingestion of cycG mutant animals appeared normal as judged by the intake of colored yeast paste (Fig 2A). When exposed to starvation stress, cycG^HR7 mutant flies had a reduced life span compared to the wild type control (Fig 2A' and 2A''). Moreover, cycG^HR7 mutant larval fat bodies displayed an aggregation of lipid droplets similar to starved controls (Fig 2B–2B''). This phenotype has been described before as a result of amino acid withdrawal and likewise loss of TOR and can be taken as an early evidence of fat mobilization for energy consumption [20]. We therefore determined the proportion of lipid and protein in the cycG mutants compared to control larvae of the same developmental stage. We noted a significant shift in favor of the triacylglycerol (TAG) level in cycG mutant larvae, suggesting an elevated level of stored fat (Fig 2C and 2C' and S2A Fig). Fat accumulation was further confirmed with a buoyancy-based assay [32]. Whereas wild type larvae sink in a 10% sucrose solution, cycG mutant larvae float due to their higher fat content (Fig 2D and 2D' and S2B and S2B' Fig). The metabolic defect in the cycG^HR7 mutant larvae was rescued by low level expression of the hs-CycG construct, which on its own was indistinguishable from the control, emphasizing the specific requirement of CycG for a normal fat metabolism (Fig 2C and 2D').

During poor nutritional conditions TAG is mobilized from the larval fat body and the free fatty acids are delivered to the larval oenocytes [33]-[35]. Oenocytes are hepatocyte-like cells that are clustered underneath the lateral epidermis of the larva [33]. To test lipid release in cycG mutants, we examined lipid accumulation in the oenocytes under conditions of feeding and starvation (i.e. amino acid deprivation) [33]. Well fed control larvae show little lipid accumulation in the oenocytes (Fig 2E and 2H), yet an aggregation of lipid droplets is observed after a 14 hours fasting period (Fig 2E'). Oenocytes of cycG mutant larvae accumulated numerous lipid droplets already under normal feeding conditions (Fig 2F and S2C–S2F Fig), which was rescued by the hs-CycG background at ambient temperature (Fig 2G), but which was hardly increased under conditions of starvation (Fig 2F'). The observed differences were highly significant (S3 Fig). In summary, cycG mutants display a starvation phenotype even under normal feeding conditions suggesting a defect in the regulation of lipid mobilization from the fat body. The influence of TOR activity on lipid metabolism is well established [35]. For example, a downregulation of the TORC1 signal by overexpression of the negative regulator TSC1/2 in the fat body has been shown to provoke a marked lipid droplet accumulation in the oenocytes of well fed larvae [33] (Fig 2I and 2J'). We noted that the loss of CycG had a very similar effect as the overexpression of TSC1/2 (compare Fig 2F and 2J), further supporting a link between CycG and the InR/TOR signaling pathway.

cycG mutants display defects in InR/TOR signaling activity

The primary downstream targets of TORC1 are S6 kinase (S6K) and elF-4E binding protein (4E-BP) (for review [22], [23]). The growth defects of the cycG mutants prompted us to analyze the phosphorylation level of S6K and 4E-BP in protein extracts from cycG^HR7 and wild type control flies (Fig 3): as expected for a positive role of CycG in TORC1 signaling, the level of the phosphorylated isoform was each decreased in the mutant (Fig 3). We further addressed the phosphorylation status of the kinase Akt1, which is at the point of intersection between InR and TOR signaling cascades, activating the latter (for review [2], [8]). Interestingly, phosphorylation levels of Akt1 were also reduced in the cycG mutants (Fig 3 and S4A Fig), and rescued to normal in the hs-CycG background at ambient temperature, which displayed normal levels on its own (S4A Fig).

Due to its impact on energy homeostasis and cell growth, loss of TOR in Drosophila affects multiple tissues and organs. For example, TOR mutants display characteristic defects in endoreplication, e.g. of cells in the salivary glands, accompanied by a reduction of Cyclin E levels that regulate G1/S phase entry in mitotic and endoreplicative tissues [20]. In fact, the salivary glands and their polytene nuclei were smaller in cycG^HR7 mutant larvae compared to control, indicative of a reduced ploidy (S4B–S4D Fig). In addition, we also observed a reduced level of Cyclin E protein in the cycG^HR7 mutant compared to wild type larvae of the same developmental stage (S4E Fig). In summary, cycG mutants phenocopy a reduced TORC1 activity, which can be explained at the molecular level by a requirement of CycG for full Akt1 activity, and hence, TOR activation.

CycG influences dILP abundance in the brain

TORC1 activity in the larval fat body is required for the release of Drosophila Insulin-like peptides (dILPs) from specialized neurosecretory cells in the brain called insulin-producing cells (IPCs) [4]. Four of the eight known dILP genes (dILP1, 2, 3 and 5) are specifically expressed in IPCs [3], [5]. Interestingly, only IPC-derived transcription of dILP3 and dILP5 is sensitive to food deprivation, whereas dILP2 transcription remains unchanged [36]. Expression of dILP2 and dILP5 was monitored in the brains of cycG^HR7 mutant third instar larvae under fed conditions. In contrast to dILP2 (Fig 4A–4C), dILP5 expression was reduced in the cycG^HR7 mutant comparable to the level observed in starved control animals (Fig 4A'-4C'). Moreover, it was reported that the production and secretion of dILP2 and dILP5 peptides are also controlled post-transcriptionally by nutritional inputs [4]. Under normal fed conditions, both proteins are evenly distributed within the IPC cell body and its axons. Upon starvation, a strong accumulation of dILP protein is observed in the IPC cell body and the axonal termini [4] (see also Fig 4D and 4E for controls). Although this analysis does not allow discriminating between insulin production and secretion, it does reveal changes in insulin dynamics [37]. In accordance with the starvation phenotype of the cycG mutants, we detected a strong increase in dILP2 protein labeling similar to that of the IPCs in starved wild type larval brains (compare Fig 4E and 4F). Compared to well-fed larvae, the signal intensity was nearly doubled in starved wild type larvae as well as in well-fed cycG^HR7 mutant larvae (Fig 4G). Ablation of dILP producing cells is correlated with a reduced egg production, linking dILP activity directly to fertility [38]. As expected by the perturbed dILP2 abundance in IPCs, cycG^HR7 mutant females laid significantly less eggs per day, reaching only two thirds of control females (S4F Fig). This phenotype was normalized in the hs-CycG background at ambient temperature (S4F Fig). Together, these observations indicate that CycG is required for the endocrine control of dILP production or secretion that is regulated by TOR activity.

An elevated level of Akt1 ameliorates cycG mutant defects

As the growth and metabolic deficits in cycG mutants reveal an impairment of InR/TORC1 signaling activity, we tested the potential role of CycG in this network. One of the first effectors of the activated InR is PI3K92E whose activity was monitored with the tGPH reporter in cycG larval tissues [39]. This reporter expresses GFP fused to a pleckstrin-homology domain (GPH) that is recruited to the plasma membrane upon PI3K activation [39]. We compared fat body cells from early third instar wild type larvae under fed and starved conditions with those of fed cycG^HR7 mutant larvae (Fig 5A). In agreement with the published data, tGPH accumulated little along the fat body cell membranes in starved larvae (Fig 5A) [39]. In the cycG^HR7 mutants, the tGPH reporter highlighted the membranes as strongly as in the well-fed wild type, indicating a normal threshold of PI3K92E activity in the absence of CycG (Fig 5A). The next factor downstream of PI3K, Akt1 shows a reduced phosphorylation level in cycG^HR7 mutants (Fig 3). We hence studied the ability of Akt1 to rescue the growth defects of the cycG mutant by genetic epistasis experiments. To this end, Akt1 was specifically induced in the larval fat body of cycG^HR7 mutant larvae, resulting in normal weight animals (Fig 5B). Furthermore, lipid droplet accumulation in oenocytes of cycG^HR7 mutant larvae was considerably improved by the fat body specific expression of Akt1, commuting to a more wild type level (Fig 5C and S3 Fig). These data show that activation of InR/TORC1 signaling at the level of Akt1 is sufficient to counteract the starvation phenotype of cycG mutants.

Mutants in PP2A-B' subunit widerborst (wdb) rescue cycG mutant phenotypes

Drosophila Widerborst (Wdb) is a B' subunit of the protein phosphatase 2A (PP2A) (for review: [40]). Wdb acts as a negative regulator in the InR/TOR network by targeting PP2A to dephosphorylate Akt1 [18]. Both mammalian CycG homologues, CycG1 and CycG2, interact directly with several B' subunits of PP2A, acting as specificity factors [41], [42]. A likewise direct molecular interaction of CycG and Wdb has been predicted in Drosophila [27], [28], which we confirmed in a yeast two-hybrid assay, showing that it involves the conserved cyclin domains (Fig 6A). Moreover, Wdb and CycG were co-precipitated from embryonic extracts, indicating in vivo complexes including the two proteins (Fig 6B and S5A Fig).

To determine whether the growth and metabolic defects observed in cycG mutants might be due to a deregulated PP2A activity, we first assayed the size and weight of wdb cycG double mutant larvae and adults. To this end, the two wdb alleles wdb¹⁴ and wdb^dw were used, which are lethal in homozygosis but viable in the trans combination [43]. Each allele was recombined with the cycG^HR7 allele to generate the double mutant heteroallelic combination. We found that wdb cycG double mutant larvae and adults showed a nearly wild type size and weight (Fig 6C and 6D and S5B and S5B' Fig). Accordingly, lipid storage defects of the wdb cycG double mutant larvae were likewise normalized, i.e. TAG-levels, specific weight and lipid droplet accumulation in the oenocytes were similar to wild type (Fig 6E and 6F and S5C and S5C' Fig). Finally, the abundance of phosphorylated Akt1 in the wdb cycG double mutants was similar to the control and no longer diminished compared to the cycG^HR7 homozygotes (Fig 6G).

CycG negatively influences the binding of Wdb and Akt1

The remarkably diverse and cell type specific functions of Akt1 in the context of InR/TOR signaling have recently been attributed to the existence of different subcellular pools of activated Akt1 kinase that control different cellular processes [44]. For example, whereas activated Akt1 is predominantly found at the apical membrane of Drosophila eye tissue, it is mostly cytoplasmic in the Drosophila ovary, where it regulates the lipid metabolism in nurse and follicle cells [18], [44], [45]. This specific ovarian function of Akt1 is under the control of Wdb, selectively modulating the levels of cytoplasmic phosphorylated Akt1 and thereby lipid droplet size in ovarian cells [18]. Accordingly, Wdb and Akt1 physically interact in the ovary, whereas no interaction was observed in larval tissue [18].

Because both larval and adult phenotypes of cycG mutants depend on Wdb activity (Fig 6 and S5 Fig), we asked whether CycG may influence the physical interaction of Wdb and Akt1. Indeed we observed a robust interaction of Akt1 and Wdb in head extracts from cycG^HR7 mutant animals in contrast to control animals (Fig 7A), implying an involvement of CycG in the regulation of Akt1/Wdb binding. Both protein species of Wdb co-precipitated with Akt1 (Fig 7A), indicating that the binding of Akt1 is not restricted to the higher molecular weight species of Wdb as previously reported [18]. Accordingly, no influence of CycG on the relative abundance of the two protein species of Wdb was detected comparing wild type and cycG^HR7 protein extracts (Fig 7B). Overall our data point to a causal link between CycG and PP2A activity in the regulation of growth and metabolism in Drosophila at the level of Akt1 (Fig 8): the presence of CycG may disfavor binding of PP2A-B’ to Akt1, which is facilitated in its absence, resulting in a decrease of phosphorylated, i.e. activated Akt1.

Discussion

In this study, we have analyzed the role of Cyclin G in growth regulation and metabolism of Drosophila. We made use of two different cycG null mutant alleles, thereby allowing us to follow the developmental consequences resulting from the absence of cycG gene activity instead of drawing conclusions from overexpression or RNAi experiments. Misexpression studies initially raised the assumption that CycG negatively regulated cell growth and cell proliferation in Drosophila [26]. Our results now indicate that CycG is required for normal growth, affecting both cell size and cell number. In fact, clonal analysis revealed a cell autonomous requirement of CycG not only in the wing but also the eye anlagen (Fig 1F and 1F’). In addition, the cycG null mutants show signs of metabolic disorder. We provide evidence that CycG facilitates InR/TORC1 mediated growth regulation via PP2A, thereby helping to sustain nutrient dependent growth in Drosophila (Fig 8).

The multiple roles of CycG in Drosophila

Drosophila CycG appears to have extraordinarily diverse roles. It has been involved in epigenetic regulation of homeotic gene activity, in cell cycle regulation, developmental stability and in DNA repair [24], [26], [46], [47], and now also in metabolic homeostasis. Our current work confirmed molecular interactions between CycG and Wdb proteins in vivo that had been predicted from genome-wide proteome analyses in vitro [27], [28]. Interestingly, similar molecular interactions have been described before for mammalian CycG1 and CycG2: both proteins interact with several B' subunits, thereby mediating the recruitment of PP2A to its different substrates [41], [42]. In contrast to mammals, the genetic relationship between CycG and PP2A is antagonistic in Drosophila as a reduction of PP2A activity ameliorates the consequences of CycG loss. The cycG mutation could be formally explained by a gain of PP2A activity. It is tempting to speculate that the diversity of CycG functions results from a regulation of PP2A by CycG. PP2A affects a plethora of developmental and cellular processes, hence, pleiotropy is expected in case of its misregulation (for review [40], [48]). Most likely, this hypothesis is too simplified. For example, loss of cycG in the female germ line results in an increase of phosphorylated H2Av (gamma-H2Av) [24], a known target of PP2A activity [49]. One might have expected a reduced amount of gamma-H2Av if loss of CycG equated with a gain in PP2A activity. Instead, we have shown that CycG is found in a protein complex together with Rad9 and BRCA2 that primarily acts in the sensing of DNA double strand breaks [24]. The importance of Drosophila CycG in DNA double strand break repair is reminiscent of functions described for mammalian CycG proteins: albeit CycG1 and CycG2 mutant mice are viable and healthy, they are both sensitive to DNA damaging reagents [50], [51]. Moreover, upregulation of CycG2 was involved in the activation of Chk2 and in damage induced G₂/M cell cycle arrest, i.e. in DNA damage response in mammals as well [51]. Whether the other phenotypes and interactions reported for Drosophila CycG are linked to the regulation of PP2A remains to be addressed in more detail.

Drosophila CycG is required for InR/TOR-mediated growth and metabolism

The cycG mutants display several phenotypic characteristics of a diminished TORC1 signaling activity [20], [33], including weight reduction, a reduced egg laying rate, impaired endoreplication and a general increase in lipid mobilization.

Moreover, CycG activity promotes phosphorylation of the primary TORC1 targets, i.e. S6K and 4E-BP. In contrast to TOR mutants, however, cycG mutants are viable, implying that CycG facilitates InR/TOR signaling rather than being an essential factor. Overall, cycG mutant flies show typical signs of nutritional starvation distress even under normal food conditions, suggesting a problem in their capacity to take up food and/or to sense and utilize the food. This defect is not due to a general inability of the animal to grasp the feed, but instead reflects a defect in coordinating the energy status with the regulation of systemic growth. As dILP accumulation in the brain is altered in cycG mutants, we know that the signals transmitted from the nutritional sensor fat body must be disturbed. The fact, that we can strongly ameliorate the growth defects of cycG mutants by an induction of Akt1 specifically in the fat body rules out a function of CycG in the endocrine signal emanating from the fat body. Instead, all of our data indicate that CycG acts genetically at the level of Akt1, thereby controlling TOR signaling activity (Fig 8).

Akt1 is negatively regulated by PP2A (for review [48]), supporting a model whereby CycG exerts its positive input on Akt1 via an inhibition of PP2A. In accordance, mutations in wdb efficiently rescue the growth and metabolic defects observed in cycG mutants. Likewise a downregulation of Wdb ameliorates the weight deficits resulting from a loss of Akt1 activity [18]. In Drosophila, Wdb acts as a tissue-specific negative regulator of Akt1: it modulates lipid metabolism in the ovary as a result of a direct interaction with Akt1, whereas no such influence was seen in eye tissue [18]. We have shown that Wdb-Akt1 binding in the adult head is favored in the absence of CycG, i.e. CycG is able to influence the interaction between Wdb and Akt1 presumably by its direct binding to Wdb. A consequence of CycG loss may be the enhanced binding of PP2A to Akt1 and an enforced dephosphorylation of Akt1, resulting in the inhibition of downstream TOR signaling activity and affecting lipid metabolism and growth (Fig 8). Moreover, the second B'-subunit of Drosophila PP2A (also called Well rounded, Wrd) is involved in the negative regulation of the S6K [15]. Assuming a molecular interaction of Wrd and CycG, a likewise regulatory input of CycG on PP2A containing the Wrd B'-subunit is conceivable. In this case, CycG might influence S6K activity as well, having a regulatory input on InR/TOR signaling also downstream of TORC1. This scenario is complicated by the negative feed back regulation of InR signaling by S6K and of Akt1 by TORC1 [52]. Circular regulation of InR/TOR signaling has been described at several levels, implementing a tight control of dietary signals and growth but complicating genetic analyses [52] (for review [53]).

In conclusion, the identification of CycG as a novel regulator of InR/TOR signaling in Drosophila highlights the importance of studying the regulatory network at the Akt1—PP2A nexus. Based on the high conservation of the InR/TOR signaling pathway and its regulation by PP2A, mammalian fat homeostasis is likely to involve similar regulatory control mechanisms to those we have uncovered in Drosophila. Our work raises the possibility of an involvement of CycG in InR/TOR-associated diseases that might be modulated by PP2A. A better understanding of the underlying mechanisms could therefore open up avenues for new strategies to fight InR/TOR-associated disorders in the future.

Materials and Methods

Fly strains, genetics and mitotic cell clone induction

The cycG^HR7 null allele and the pUASp-cycG transgene have been previously described [24]. The generation and verification of the cycG^eoC allele is described in the supplementary experimental procedures. To generate the hs-CycG construct, cycG cDNA was cloned as 2.4 kb EcoRI/KpnI fragment into the pCasper-hsRX vector [54] and several independent lines were established by P-element mediated germline transformation [55]. For rescue assays an insertion on the second chromosome was used and combined with the cycG^HR7 mutant.

Fat metabolism and growth defects were analyzed with the following fly strains (BL strains were from Bloomington stock center): Adh-Gal4 (gift from R. Kühnlein) [39]), da-Gal4 (BL8641), Lsp2-Gal4 (BL6357); UAS-Akt1 (BL8191), UAS-GFP (BL4776), UAS-lacZ (BL8529), UAS-TSC1 UAS-TSC2 [56], the tGPH-reporter (gift from B. Edgar) [39] and the wdb mutant alleles wdb¹⁴/TM6B and wdb^dw/TM6B (both obtained from C. Wilson) [18], [43]. Oregon-R was used as wild type control. Flies were raised at 25°C under non-crowded conditions on standard cornmeal/molasses/agar medium. For amino acid deprivation, third instar larvae were kept on a sugar only diet [20% sucrose, 1% agar in PBS] for 14 hours.

The Flp/FRT system was used to generate cycG^HR7 mutant clones by mitotic recombination [30]. To this end cycG^HR7 was recombined with the FRT82B bearing chromosome (BL2050). Females of the genotype yw hsFlp; FRT82B cycG/TM6B were mated with FRT82B arm-lacZ/TM6C (BL7369) males. 24–36 hours after egg laying, the animals were subjected to a 30 minute heat shock at 37°C to induce recombination with low frequency. Control clones were generated in parallel with only the FRT82B bearing chromosome. Mutant cell clones are characterized by the loss of lacZ and were analyzed in imaginal discs of third instar larvae. A total number of 30 discs was assayed and compared with control clones.

Documentation of adults, determination of cell size and number

Adult males, three to five days old, were used for analysis. Dehydrated wings were mounted in Euparal (Roth; Karlsruhe, Germany). Wing area was measured using Image J software (oval selection for total wing size in two measurements that were sampled). Cell number was determined by counting the individual trichoma on the wing blade in three defined 10.000 μm² squares localized in the L4/L5 field next to the posterior cross vein. Subsequently, the total cell number for the determined wing area was calculated, as was cell size. Pictures were assembled using Corel Draw and Corel Photo Paint. Flies and wings were photographed with an ES120 camera (Optronics; Goleta CA, USA) using Pixera Viewfinder software, version 2.0.

To investigate developmental timing, offspring of six parallel inter se crosses with the genotype w¹¹¹⁸; cycG^HR7/+ was counted at days 9 to 20. Since the cycG^HR7 mutant carries a mini-white⁺ gene [24], the homozygous cycG^HR7 and the heterozygous cycG^HR7/+ flies could be distinguished by red and orange eye color, respectively, from the white-eyed control siblings.

Ingestion, starvation assay and amino acid deprivation

As an assay for ingestion, blue-colored yeast paste was offered to larvae as a food source. Food uptake in the gut was visualized by illuminating larvae from the side and taking pictures with a Pixera camera coupled to a Leica stereo-microscope. For the starvation assay, triplicate batches of 15 three days old males of each genotype were transferred to vials containing 1% agarose in PBS only (wet starvation). Mortality rate was determined by counting the number of dead flies every two hours. Third instar larvae were likewise amino acid deprived for two days before dissection.

Weight measurement, TAG and protein assay, buoyancy test

Newly hatched male flies were transferred to fresh food vials and maintained at 25°C for three days before measurement. Body weight of 100 flies of each genotype was measured individually with a precision scale. Organismal triacylglycerol (TAG) and protein content was quantified using the Pierce BCA Protein determination Kit (Thermo Fisher Scientific; Rockford IL, USA) and the protocol procedure B of the Triglycerid-Assay Kit (Sigma-Aldrich; St. Louis MO, USA). For each genotype, batches of five third instar larvae were homogenized in 150 μl 0.01% Tween in PBS. After incubation at 70°C for 5 minutes the samples were centrifuged for 1 minute at 5000 rpm. The supernatant was transferred to a fresh microtube and after additional centrifugation for 3 minutes at 14000 rpm the cleared lysate was applied in the appropriate assay. A minimum of three independent experiments was performed for each genotype and the results were sampled. A simplified version of the buoyancy-based screen protocol was used [32]: 10 larvae of each genotype were placed in 3 ml of 10% sucrose solution. After gentle mixing and five minutes without agitation, the number of larvae floating at the surface was counted and documented.

Statistical evaluation

Statistical significance of probes was determined according to Student's T-test (http://www.physics.csbsju.edu/stats/t-test.html) and p-value was scaled accordingly: p>0.05 (not significant, n.s.); p<0.05 (weakly significant; *); p<0.01 (significant; **); p<0.001 (highly significant; ***).

Tissue staining protocols

Larval brains were dissected in PBS and fixed for 20 minutes in 4% paraformaldehyde. After several washes with PBS plus 0.3% Triton X100 followed by a preincubation step with 4% normal goat serum, rabbit anti-dILP2 antibody (1:800; gift from P. Léopold) [4] was added and incubated over night at 4°C. Imaginal discs were likewise treated, and stained with anti beta-Galactosidase antibodies (1:50; DSHB; Iowa, USA). Secondary antibodies coupled to DTAF or Cy3 were purchased from Jackson ImmunoResearch (Dianova; Hamburg, Germany). The Oil Red O staining of oenocytes was performed exactly as described previously [33]. Larval fat body was dissected from well fed or amino acid deprived third instar larvae in PBS, fixed in 4% paraformaldehyde for 15 min and stained with Nile Red (Sigma-Aldrich; St. Louis MO, USA) at a concentration of 10 μg/ml. As a measurement of PI3K92E activity in vivo, the tGPH reporter [39] was used and the degree of membrane tGPH localization analyzed in early third larval tissues by confocal microscopy. Fluorescently labeled tissues were mounted in Vectashield. Larval brains were analyzed with a Zeiss-ApoTome Axio Imager using AxioVision Software. Imaginal discs and fat bodies were documented with a Bio-Rad MRC1024 confocal microscope coupled to a Zeiss Axiophot using Laser Sharp 2000 software (Carl Zeiss AG; Oberkochen, Germany). Pictures of fat bodies were inversed for better visibility.

In situ hybridization

In situ hybridization on larval brains was performed with digoxygenin labeled DNA probes of dILP2 and dILP5 according to standard protocols [57]. The templates for the probes were generated by PCR using the following primer sets: UP dILP2: GAT CGT AAA GCA ACC TAA GCA GTA A; LP dILP2: ATT CGT AAA GAG TAA CAT GCA ACA A; UP dILP5: GAT CCC AGT TCT CCT GTT CCT GAT C; LP dILP5: TTT CAA GTT TCA AAG CCG TGC ATA T.

Immunoblots

For each genotype 100 adult heads were homogenized in RIPA I buffer [50 mM Tris-HCl pH 7.5, 150 mM NaCl, 1% Triton X-100, 0.1% SDS, protease inhibitor cocktail (Roche Diagnostic; Basel, Switzerland)] on ice. After centrifugation, loading buffer was added to the supernatant and the probes were loaded on a SDS-PAGE followed by Western blotting. The extracts were probed with guinea pig anti-CycG (1:400) [46]. As loading controls we used anti-beta-Tubulin (1:50) (E7 DSHB; developed by M. Klymkowsky) and anti-Erk1/2 antibodies (1:1000) (Cell Signaling Technology; Danvers MA, USA), as tubulin levels might change when InR signaling is influenced [15]. The amount of total and phosphorylated protein was determined with rabbit anti-4E-BP (1:100, gift from G. Tettweiler) [58], rabbit anti-Akt1 (1:250), rabbit anti-p70 S6 kinase (1:100), rabbit anti-Phospho-Akt (1:250), rabbit anti-Phospho-p70 S6 kinase (1:100) and rabbit anti-Phospho-4E-BP (1:100) (all from Cell Signaling Technology; Danvers MA, USA).

Yeast two-hybrid studies and co-immunoprecipitations

Full length wdb was PCR amplified from cDNA (LD34343, obtained from DGRC, Bloomington IN, USA) and cloned as BglII/NotI fragment into pEG and VP16 (BamHI/NotI) vectors. pJG-CycG (1–566), GST-CycG (1–215) and GST-CycG (215–566) DNA was a gift from F. Peronnet, France [46]. The CycG subdivision constructs were PCR-amplified using the pJG/GST-CycG constructs as template and further subcloned as EcoRI/XhoI fragments in either pEG, pJG or VP16 vectors [59], [60]. Protein-protein interaction assays were done according to standard protocols using the Brent two-hybrid system [59]. Protein expression in yeast cells (EGY40: Mata, ura3, his3, trp1, leu2, GAL) was verified either with mouse anti-HA (1:1000; St. Louis MO, USA), mouse anti-VP16 (1:100; Santa Cruz Biotechnology, Dallas, USA) or rabbit anti-LexA antibodies (1:1000; Bio Acadamia, Osaka, Japan).

CycG or Wdb protein was immuno-precipitated from about 500 embryos (0–24h) using either anti-CycG antibodies or anti-Wdb antibodies (see supporting materials and methods) as described before [46]. Akt1 and Wdb complexes were co-immunoprecipitated from 150 heads each of either wild type or cycG^HR7 homozygous mutant animals using rabbit anti-Akt1 (1:50; Cell Signaling Technology; Danvers MA, USA) and detected with rabbit anti-Akt1 or rat anti-Wdb (see supporting materials and methods).

Supporting Information

Attachment 1

Attachment 2

Attachment 3

Attachment 4

Attachment 5

Attachment 6