Isolation of a novel missense mutation in insulin receptor as a spontaneous revertant in ImpL2 mutants in Drosophila

The insulin/insulin-like growth factor (IGF) signaling (IIS) pathway is a key regulator of cellular growth and metabolism throughout the life cycle of several organisms. Dysregulation of IIS is linked to many human diseases including diabetes mellitus and cancer (Gallagher and LeRoith, 2020; Saltiel and Kahn, 2001). In addition, IIS regulates a wide variety of physiological phenomena including reproduction, aging, sleep and memory formation (Belfiore et al., 2017; Fontana et al., 2010; Nässel and Zandawala, 2020; Piper and Partridge, 2018; Tatar, 2021; Templeman and Murphy, 2018). The insulin receptor is a disulfide-linked (αβ)₂ homodimer that consists of an extracellular ligand-binding α subunit and the membrane-spanning β subunit, which harbors the intracellular kinase domain (Belfiore et al., 2017). Binding of insulin to the extracellular region induces conformational changes, leading to an elevation of intrinsic tyrosine kinase signaling.

The functions and components of IIS are remarkably conserved in invertebrates, including in the fruit fly Drosophila (Teleman, 2010; Wu and Brown, 2006). Impairment of IIS in Drosophila causes growth and metabolic defects, including elevation of circulating carbohydrates, as observed in mammals (Chatterjee and Perrimon, 2021; Hietakangas and Cohen, 2009; Mattila and Hietakangas, 2017; Texada et al., 2020). Drosophila insulin-like peptides (Ilps), mainly derived from brain insulin-producing cells (IPCs), act as systemic ligands that stimulate insulin receptor (InR) in the peripheral tissues (Ikeya et al., 2002; Rulifson et al., 2002). The cellular components of the Drosophila IIS include the IRS homolog Chico, phosphoinositide 3-kinase (PI3K), the phosphatase PTEN, and the serine/threonine protein kinase Akt (also known as PKB). The domain structures in Drosophila InR highly correspond with those in the human insulin receptor (hInsR) and IGF1 receptor (hIGF1R) (Fernandez et al., 1995). Importantly, mammalian insulin binds to Drosophila InR and induces PI3K-Akt signaling, as revealed in ex vivo culture experiments (Musselman et al., 2011; Pasco and Léopold, 2012). In contrast, recombinant Ilp5 binds to and activates mammalian InsR (Sajid et al., 2011), further indicating the evolutionary conservation of ligand-receptor interactions.

Nutritional signals tightly regulate the production and secretion of Ilps in IPCs through systemic factors derived from the fat body, an organ equivalent to the mammalian liver and adipocyte (Droujinine and Perrimon, 2016; Koyama et al., 2020; Meschi and Delanoue, 2021). Moreover, the function of circulating Ilps is regulated by their binding proteins, ImpL2 and the decoy receptor SDR (Honegger et al., 2008; Okamoto et al., 2013). IIS, in turn, regulates the metabolism and storage of ingested fuels in the fat body to ensure appropriate growth of the organisms (Britton et al., 2002). As increased IIS activity continues to maintain anabolic pathways, thereby limiting the amounts of circulating nutrients, the lack of either ImpL2 or SDR causes lethality under unfavorable nutritional conditions. Thus, proper regulation of the IIS is crucial for adaptation to nutrient fluctuations.

Drosophila genetics has provided powerful approaches to unravel the normal functions of gene products in development and physiology. In addition to null alleles that completely lack the protein function, missense mutations that partially lack protein function are also informative, particularly for modeling congenital diseases in humans. In this study, we identified a novel temperature-sensitive point mutation in InR as a spontaneous revertant that suppresses constitutive elevation of IIS.

The secreted protein ImpL2 is an antagonist of Ilps and its mutation results in an increase in peripheral IIS and promotes body growth (Fig. 1A) (Honegger et al., 2008). Consistently, homozygous ImpL2^Def20 and transheterozygous ImpL2^Def20 mutations over deficiency lines led to large-sized adults in both males and females (Fig. 1B; Fig. S1A). Unexpectedly, however, we found that homozygous ImpL2^Def42 mutants led to small-sized adults, whereas transheterozygotes of ImpL2^Def42 over deficiency led to large-sized adults, suggesting that second site mutation(s) mask the ImpL2^Def42 mutant phenotype regarding body growth. Because ImpL2 inhibits the function of Ilps, we speculated that second site mutation(s) might be involved in IIS on the same chromosome. We found that transheterozygotes of ImpL2^Def42 with two InR mutants, InR⁰⁵⁵⁴⁵ and InR^93dj4, but not Akt mutants, showed severe growth defects (Fig. 1C; Fig. S1B), suggesting that a second site mutation is located in the InR gene locus. Given that both ImpL2^Def20 and ImpL2^Def42 mutants were generated by imprecise excision of the same P-element (Honegger et al., 2008), we compared the entire coding sequence of the InR gene between the ImpL2^Def20 and ImpL2^Def42 alleles and found that the ImpL2^Def42 mutants, but not the ImpL2^Def20 mutants, retained a missense mutation c.2705A>G (p.Tyr902Cys) (Fig. 1D; Table S1). Tyr902 is in the C-terminal end of the first fibronectin type III domain (FnIII-1) in the InR extracellular region. This amino acid residue is not perfectly identical but functionally conserved through evolution (Fig. 1E), suggesting that this residue is vital for InR function. Homozygotes in backcrossed lines, which harbor the Y902C missense mutation but not the deletion in the ImpL2 locus, always led to small-sized adults (see below). A Y902C mutant allele in InR, hereafter named InR^Y902C, was established after backcrossing with the control, w⁻, seven times. Notably, a backcrossed ImpL2^Def42 allele, which did not retain the InR^Y902C mutation, led to large-sized adults, as expected (Fig. S1C).

Subsequently, we performed sequence-based in silico prediction analysis to predict the effect of single amino acid variations on protein function and stability. Computational prediction with the program PROVEAN [Protein Variation Effect Analyzer; the J. Craig Venter Institute (JCVI)] predicts that a tyrosine to cysteine amino acid change at this position is ‘deleterious’ with a score of −6.735 (if the score is equal to or below −2.5, the variant is predicted to have a deleterious effect). The corresponding amino acids of Y902 of Drosophila InR are F565 in hInsR and Y580 in hIGF1R. Although no single nucleotide polymorphisms at these positions in human genes have been reported, missense mutations in these sites to cysteine were also predicted to be highly damaging in silico: PROVEAN score of −5.208 (in hInsR) and −4.302 (in hIGF1R). Thus, the Y902C mutation likely alters InR function.

In addition to insulin binding to the leucine-rich repeat (L1) domain, insulin independently binds to the FnIII-1 domain (Gutmann et al., 2020; Li et al., 2022; Uchikawa et al., 2019). To further understand the molecular nature of the Y902C mutation, we mapped it onto a three-dimensional structural model of hInsR (Fig. 1F). Y902 (F565 in hInsR) is located in the β-sheet that is laterally connected by two adjacent antiparallel strands (Fig. 1G). The side chain of F565 is located on the inside, but not on the surface, of the molecule and is in close proximity to the hydrophobic side chains of F506 and I585, both of which are functionally conserved in Drosophila (Fig. 1H; Fig. S2A). Thus, F565 is not directly involved in insulin binding and is thought to be important in stabilizing the overall secondary structure and conformational changes upon ligand binding.

It has been recently reported that the classical allele InR^E19 has a missense mutation p.Val811Asp (Yamamoto et al., 2021), which is located at the N-terminal end of the FnIII-1 domain (Fig. S2A). Interestingly, the hydrophobic side chain of the corresponding F475 in hInsR (V811 in Drosophila InR) is also inside the molecule and placed close to F565. However, there was no direct interaction between these two side chains (Fig. 1G; Fig. S2B,C).

To further characterize the InR^Y902C mutant phenotype on body growth and IIS, we examined body size. Pupal volume, adult weights and adult wing area in homozygous InR^Y902C mutants were significantly reduced in both males and females compared with the control w⁻ (Fig. 2A-C). Subsequently, we used InR⁰⁵⁵⁴⁵, a strong hypomorphic allele of InR, in which homozygotes were lethal mainly at the second instar larvae (Bolukbasi et al., 2012; Fernandez et al., 1995). InR⁰⁵⁵⁴⁵ has a P-element insertion upstream of the mature protein. Transheterozygotes of InR⁰⁵⁵⁴⁵ with InR^Y902C (InR^05545/Y902C) had severely reduced body size compared with InR^Y902C homozygotes (Fig. 2A-C), indicating that InR^Y902C is a weak hypomorphic allele. Next, we analyzed the expression levels of Foxo target genes. InR and 4E-BP (Thor) are upregulated by the reduction of IIS (Okamoto et al., 2013; Puig et al., 2003). Similarly, a Foxo target ilp6 is induced in the fat body upon starvation (Slaidina et al., 2009). We found that InR was expressed normally in InR^Y902C mutant larvae compared with control larvae under fed conditions (Fig. 2D), suggesting that the Y902C mutation does not compromise InR mRNA stability. InR^05545/Y902C mutants show slightly but significantly increased InR expression levels, suggesting that IIS is decreased in InR^05545/Y902C mutants. Consistent with the degree of body size reduction, the expression of 4E-BP and ilp6 increased in the InR^Y902C mutants and increased more significantly in the InR^05545/Y902C mutants under fed conditions. Starvation further increased the expression of InR and 4E-BP in InR^05545/Y902C mutant larvae. Furthermore, ilp2 and ilp5, major Ilps regulating body growth during the larval period, were significantly increased in the InR^Y902C and InR^05545/Y902C mutants (Fig. 2D). Interestingly, starvation completely ceased this upregulation, suggesting that ilp2 and ilp5 are induced to compensate for the reduction of IIS in a nutrient-dependent manner.

Given that InR^Y902C mutants responded to starvation, we next examined whether InR^Y902C could activate IIS in response to Ilps. To this end, we incubated the larval fat body with bovine insulin under in vitro organ culture conditions and analyzed pAkt levels by immunostaining. We found that insulin stimulation increased pAkt signals at the plasma membrane in the control fat body, whereas those in the nucleus did not change (Fig. 2E,F). Similarly, insulin strongly increased pAkt signals in InR^Y902C mutant, but not in homozygous InR⁰⁵⁵⁴⁵ mutant, fat body cells. From these results, we concluded that the InR Y902C mutation attenuates IIS but retains the binding ability to Ilps to induce PI3K-Akt signaling to some degree.

Missense mutations are sometimes associated with temperature sensitivity. We observed that the InR^Y902C mutation is a temperature-sensitive allele in which the mutant phenotype became severe at high temperatures (29°C) and, in turn, lost or reversed at low temperatures (18°C). The pupariation timing in InR^Y902C mutants was slightly delayed compared with that in the control larvae at 25°C (Fig. 3A). However, at 29°C, InR^Y902C mutants displayed a significant delay in the pupariation timing by ∼5 days and produced small-sized adults (Fig. 3A-D). Further analysis of adult wings revealed that the reduction in wing area in the InR^Y902C mutants was caused by a reduction in both cell size and cell number (Fig. 3E; Fig. S3A). Moreover, InR^Y902C mutants successfully underwent pupariation, but half of them failed to eclose at 29°C (Fig. 3F,G; Fig. S3B). At 18°C, the timing of pupariation and adult weights of InR^Y902C flies were indistinguishable from those of control flies (Fig. 3A,D). In contrast, the pupal volume and adult wing area in InR^Y902C were larger than those in the control at 18°C (Fig. 3B,E). The increase in the wing size of the InR^Y902C mutants at 18°C was caused solely by the increase in cell number but not by cell size (Fig. 3E; Fig. S3A). Together, these results indicate that InR^Y902C is a temperature-sensitive allele.

Developmental delay and growth defects observed in the InR^Y902C mutants at high temperatures imply a substantial reduction in IIS. To confirm the reduction in IIS, we next analyzed temperature-dependent expression changes in Ilp genes and Foxo-target genes. As expected, ilp2, ilp3 and ilp5 were strongly upregulated in InR^Y902C mutant larvae grown at 29°C (Fig. 3H) compared with those grown at 25°C (Fig. 2D). Moreover, InR^Y902C mutants showed a greatly increased expression of InR, 4E-BP and ilp6 at 29°C (Fig. 3I). The expression levels of these genes did not change between rearing temperatures in control larvae. Thus, the observed growth defects in the InR^Y902C mutants were caused by further reduction of IIS at high temperatures. Notably, a day of starvation at 29°C further increased InR and 4E-BP expression in InR^Y902C mutants, indicating that InR^Y902C mutants maintain a certain IIS to sustain larval development even at high temperatures. Interestingly, InR and ilp6 expression was reduced in InR^Y902C mutants compared with that in control larvae at 18°C. The relative upregulation in the IIS may explain the increased pupal volume and wing size at low temperatures.

Downregulation of IIS causes metabolic defects, such as increased stored triglycerides (TAG) and circulating sugars (Teleman, 2010). Consistent with previous observations, both InR^Y902C and InR^05545/Y902C mutants showed increased amounts of TAG, glycogen, trehalose and glucose in the late third instar larvae (Fig. 4A). To further understand the metabolic defects in InR mutants, a widely targeted metabolomic analysis was conducted using liquid chromatography with tandem mass spectrometry (LC-MS/MS). A principal component analysis (PCA) of 122 water-soluble metabolites revealed that the level of IIS depended on genotype and temperature according to the PC1 axis (Fig. 4B; Table S2). These results suggested that homozygous InR^Y902C mutants grown at 29°C were metabolically similar to InR^05545/Y902C mutants grown at 25°C, which is largely consistent with the degree of body size reduction (Figs 2B and 3D). The rearing temperature itself had a significant impact on the metabolic states in control larvae (Fig. S4A). Nevertheless, high temperatures caused a clear separation between the control and mutant larvae compared with the low-temperature condition (Fig. S4B). Notably, the positive loading for the metabolites that contributed maximally to PC1 included several glycolytic and TCA cycle metabolites (Fig. 4C; Table S3). In addition, the polyol pathway metabolites, sorbitol and fructose, were also increased in InR mutants (Fig. 4C,D; Table S3), suggesting that excess glucose in circulation is converted to fructose via the polyol pathway. Moreover, glucogenic amino acids, such as alanine, serine, proline and glutamine, significantly increased according to the reduction in IIS. These results suggested that mutations in InR lead to a substantial increase in glucose-related metabolic pathways.

In contrast, negative loadings contributed to the PC1 enriched ribonucleotides, particularly pyrimidine diphosphate (CDP and UDP) and triphosphate (CTP and UTP) (Table S3). These pyrimidine ribonucleotides were significantly reduced according to the reduction in IIS (Fig. 4D; Fig. S4C), suggesting that IIS directly or indirectly affected pyrimidine ribonucleotide levels. In contrast, ATP levels increased according to the reduction in the IIS (Fig. 4D). These results imply that growth defects in InR mutants cannot be attributed to a reduction in ATP levels. Overall, our metabolomics data demonstrate that a reduction in IIS leads to widespread effects on metabolic pathways, including amino acids and nucleotides.

It has been reported that a high sugar diet (HSD) induces insulin resistance in Drosophila larvae (Musselman et al., 2011; Pasco and Léopold, 2012). Therefore, we compared the growth and metabolic phenotypes in control and InR^Y902C mutant larvae that were fed an HSD at 25°C (Fig. 5A). Feeding HSD decreased the larval growth rate in control animals compared with those fed a normal diet (ND), resulting in small-sized pupae (Fig. 5B). Interestingly, pupal volume in InR^Y902C mutants did not differ between ND and HSD, suggesting that InR^Y902C mutants were largely insensitive to HSD in terms of body growth.

We further examined the metabolic profiles in control and InR^Y902C mutant larvae fed either ND or HSD at the late third instar (Table S4). PCA revealed that HSD caused a clear metabolic shift in control larvae but not in InR^Y902C mutant larvae (Fig. 5C). A volcano plot further indicated that fewer significant changes occurred in InR^Y902C mutants fed an HSD compared with control larvae (Fig. 5D). Many metabolites involved in glycolysis and the TCA cycle were increased by HSD in control larvae (Fig. 5E,F), whereas InR^Y902C mutants did not show a further elevation in many of those metabolites. In contrast, the polyol pathway metabolites, sorbitol and fructose, were significantly increased in both control and InR mutants under HSD (Fig. 5D-F). Unexpectedly, HSD-fed control larvae showed metabolic states different from those of InR^Y902C mutants with respect to amino acids and the ketone body 3-hydroxybutyrate. HSD increased the levels of 3-hydroxybutyrate, many glucogenic amino acids and branched-chain amino acids (BCAAs) in control larvae. In contrast, InR^Y902C mutants showed no increase in BCAAs, even when fed HSD (Fig. 5E,F). Moreover, InR^Y902C mutants did not show further increases in glucogenic amino acids when fed HSD. Interestingly, several vitamin Bs significantly decreased in InR^Y902C mutants that were fed HSD (Fig. 5D), suggesting that HSD-induced metabolic stress leads to higher consumption of vitamin Bs in InR^Y902C mutants, presumably through carbohydrate oxidation and energy release. Together, these results suggested that InR^Y902C mutants are less sensitive to HSD and partly phenocopy the metabolic abnormalities caused by HSD.

To further demonstrate the advantage of temperature-sensitive mutations in InR, we next analyzed fecundity after a temperature shift (Fig. 6A). Strong InR mutants have been reported to be sterile because of defects in vitellogenesis (Richard et al., 2005; Tatar et al., 2001). Similarly, chico heterozygous females have reduced fecundity, and homozygous null mutants are sterile (Böhni et al., 1999; Clancy et al., 2001). Consistent with these previous studies, we found that the average number of deposited eggs was significantly decreased in InR^Y902C mutant females at 25°C (Fig. 6B), although InR^Y902C mutants could be maintained in homozygotes. When small-sized adult females grown at 25°C were kept for 4 days at either 18°C or 29°C, fecundity further decreased. To exclude the possible effect of body size reduction on egg production capacity, we raised InR^Y902C mutants at 18°C until the adult stage and then assessed fecundity. When normal-sized InR^Y902C mutant females grown at 18°C were maintained for 4 days at 29°C, fecundity significantly decreased (Fig. 6C). Under these conditions, normal-sized InR^Y902C mutant females laid eggs comparable with those of control flies at 18°C and 25°C. These results supported the idea that the IIS plays a crucial role in egg production, irrespective of body size.

Because IIS in the ovaries regulates yolk protein uptake (Richard et al., 2005), we analyzed embryo size in InR^Y902C mutants grown at 25°C. We found that the embryo volume decreased in InR^Y902C mutants at 25°C and more significantly at 29°C compared with those in control flies (Fig. 6D). Furthermore, both embryo length and width decreased at 25°C and 29°C, whereas the aspect ratio (length/width) did not change (Fig. 6E), suggesting that embryo size was proportionally reduced by the systemic reduction in IIS. Importantly, the reduction in embryo size was not a result of body size reduction because small-sized InR^Y902C mutants produced normal-sized embryos when adult flies were kept at 18°C (Fig. 6D,E). Given that these flies produced fewer eggs (Fig. 6B), body size reduction per se appears to result in a reduction in egg number at the expense of producing normal-sized eggs. We further noticed that inter-individual variation in the aspect ratio significantly increased in InR^Y902C mutants, particularly at 29°C (Fig. 6F). This increase was mainly due to the increase in egg length variation compared with width variation. Taken together, these results suggested that IIS affects egg morphology.

In this study, we isolated a novel missense mutation in InR as a revertant. Homozygous InR^Y902C mutants developed into adults with mild growth defects and produced subsequent generations at 25°C, albeit with reduced fecundity. We showed that homozygous InR^Y902C mutants displayed several phenotypic features that have been shown previously in IIS mutants and HSD-induced type 2 diabetes (T2D) models. These include upregulation of Ilp gene expression, increased lipid and carbohydrate contents, and decreases in adult wing size owing to cell size and cell number reduction (Böhni et al., 1999; Brogiolo et al., 2001; Grönke et al., 2010; Ikeya et al., 2002; Musselman et al., 2011; Pasco and Léopold, 2012; Rulifson et al., 2002; Tatar et al., 2001).

Given that InR^Y902C suppresses the Imp-L2^Def42 mutant phenotype in body growth, it appears that a spontaneously occurring Y902C mutation in InR is fixed in the population during maintenance as an intergenic suppressor mutation. Consistent with this idea, homozygotes of the original Imp-L2^Def42 stock with the InR^Y902C mutation were healthy compared with homozygotes of the backcrossed Imp-L2^Def42 stock under our maintenance conditions. These results imply that suppressor mutations may occur in individuals with constitutively elevated IIS levels. In many studies on ImpL2, homozygous mutants of Imp-L2^Def20 or transheterozygous mutants of Imp-L2^Def20 over Imp-L2^Def42 have been used (Figueroa-Clarevega and Bilder, 2015; Honegger et al., 2008; Kwon et al., 2015; Owusu-Ansah et al., 2013; Sarraf-Zadeh et al., 2013). However, some studies have used Imp-L2^Def42 homozygotes for phenotypic analyses (Kapali et al., 2022; Lee et al., 2018). Therefore, it may be worthwhile to re-evaluate the phenotypes of Imp-L2^Def42 homozygous mutants.

Drosophila is a powerful genetic model system for studying obesity and diabetes (Chatterjee and Perrimon, 2021). However, in contrast to the genetic models of obesity, useful models of T2D are relatively limited. Instead, the HSD approach has been widely used to disrupt carbohydrate metabolism and induce insulin resistance in Drosophila. This may be because many of the available mutants in the components of IIS, including InR, cause a severe phenotype in body growth and larval lethality in homozygotes (Brogiolo et al., 2001; Chen et al., 1996; Fernandez et al., 1995; Tatar et al., 2001). Heterozygous InR mutants reduce InR kinase activity and exhibit increased Ilp secretion, similar to diabetic hyperinsulinemia (Park et al., 2014; Tatar et al., 2001). However, unlike HSD-induced T2D models, metabolic changes such as elevated lipid stores and circulating sugars are not observed in the heterozygous mutants. Consistently, we found that heterozygous InR mutants (InR⁰⁵⁵⁴⁵/+) showed mild defects in body growth but did not exhibit metabolic abnormalities. Therefore, it is worth investigating and comparing the effects of HSD and the genetic disruption of IIS using phenotypic and metabolomic approaches. To this end, we first conducted a comparative metabolomics analysis and found a strong correlation between IIS downregulation and metabolic defects. InR mutants showed significantly increased polyol pathway metabolites: sorbitol and fructose. The polyol pathway is a two-step process in which glucose is reduced to sorbitol by aldose reductase and then converted to fructose by sorbitol dehydrogenase (Sodh). In humans, increased flux of glucose into the polyol pathway leads to diabetic complications through oxidative and osmotic stresses (Brownlee, 2001; Jannapureddy et al., 2021; Tang et al., 2012). Consistently, increased intracellular sorbitol levels owing to the loss of Sodh cause age-dependent synaptic degeneration and motor impairment in Drosophila (Cortese et al., 2020). Therefore, it would be interesting to further examine tissue-specific oxidative and osmotic stresses caused by abnormal polyol metabolites in InR mutants.

We also showed that insulin resistance caused by HSD and a genetic mutation in InR exhibit similar metabolic abnormalities in glycolysis and the TCA cycle. However, these two conditions have different effects on the metabolism of ketone bodies and BCAAs. In humans, dysregulation of amino acid metabolism appears to be an early event in the progression to more severe insulin resistance and T2D (Mahendran et al., 2013; Morze et al., 2022). Insulin resistance is correlated with increased levels of ketone bodies and serum amino acids, including BCAAs, alanine and glutamate. Although we did not detect increased leucine and isoleucine levels in the InR mutants, HSD-fed control larvae showed apparent increases in these amino acids. Further studies will be required to explore the mechanistic basis underlying the amino acid changes between HSD-induced T2D models and InR mutants. Nevertheless, Drosophila InR mutants partly phenocopy perturbations in amino acid homeostasis observed in obese and insulin-resistant mammals. Further molecular characterization may expand the applicable models of insulin resistance and diabetic complications, for which Drosophila genetic models of T2D could be employed to genetically dissect the disease mechanisms.

We also found that IIS controls egg size and shape, in addition to fecundity. During oogenesis, IIS regulates germline stem cell proliferation, cyst growth, epithelial cell cycle progression, border cell migration and vitellogenesis (Ghiglione et al., 2018; Jouandin et al., 2014; LaFever and Drummond-Barbosa, 2005; Richard et al., 2005; Sharma et al., 2018; Tatar et al., 2001). The reduction in embryo volume observed in the InR^Y902C mutants may be caused by a reduction in yolk protein uptake. However, defects in vitellogenesis selectively reduce egg length but not width (Andersen and Horne-Badovinac, 2016). Thus, the reduced egg volume in InR^Y902C mutants cannot solely be attributed to the reduction in vitellogenesis. During oogenesis, spherical egg chambers within the ovary elongate along the anterior-posterior axis to create an elliptical egg shape. This elongation is thought to occur by the constraining force of the follicular epithelium (Cetera et al., 2014; Haigo and Bilder, 2011). Egg elongation in InR^Y902C mutants may be incomplete because IIS is required for stretch-induced follicle cell proliferation in the egg chambers (Borreguero-Muñoz et al., 2019). Another force that promotes elongation is the contraction of the muscle sheaths that surround the developing egg chambers (Andersen and Horne-Badovinac, 2016; Valer et al., 2022). Owing to its pleiotropic functions, it is reasonable to assume that the IIS in multiple tissues plays a role in generating a normal egg shape.

In humans, defects in the INSR gene cause a broad spectrum of congenital diseases associated with insulin resistance (Ardon et al., 2014; Hosoe et al., 2017). Disease-associated missense mutations have been mapped to the entire coding region of INSR, including the extracellular FnIII domain. The severity of the disease varies considerably and is generally correlated with the function of INSR; it can result in infant mortality, such as leprechaunism, or manifest only later in adult life, such as in insulin-resistant diabetes mellitus. Structural analysis revealed that insulin binding induces a large conformational change in hInsR (Gutmann et al., 2020; Li et al., 2022; Uchikawa et al., 2019). Binding of multiple insulin molecules to two distinct sites on each hInsR dimer breaks the auto-inhibited state and induces overall conformational changes, resulting in trans-autophosphorylation and activation. Y902C mutation in Drosophila InR may form an aberrant disulfide bond leading to misfolding of the extracellular domain. However, given the relatively mild defects in IIS and insulin responsiveness, the Y902C mutation likely affects insulin-binding affinity or conformational changes upon ligand binding for full activation, particularly at high temperatures. In contrast, the Y902C mutation may facilitate insulin binding and/or autophosphorylation at low temperatures, which explains the larger wing size. Interestingly, the InR^E19 allele, having V811D mutation in the FnIII-1 domain, also shows a temperature-sensitive growth and metabolic phenotype (Shingleton et al., 2005; Yamamoto et al., 2021). Further biochemical and structural analyses will advance our understanding of the underlying temperature-sensitive characteristics caused by a missense mutation in the FnIII-1 domain.

In summary, given the nature of the temperature-sensitive allele, InR^Y902C mutants will be a useful tool for further understanding IIS-related physiological phenomena in Drosophila. Detailed molecular and metabolic analyses of Drosophila may facilitate a better understanding of disease susceptibility, including adult-onset diabetes and age-related physiological decline.

The following Drosophila melanogaster strains were used: w¹¹¹⁸ (used as a control), Imp-L2^Def20 and Imp-L2^Def42 (Honegger et al., 2008). The following stocks were obtained from the Bloomington Drosophila Stock Center (BDSC): InR^93Dj-4 (BDSC9554), P{PZ}InR⁰⁵⁵⁴⁵ (BDSC11661), P{PZ}Akt⁰⁴²²⁶ (BDSC11627), Df(3L)GN19 (BDSC416) and Df(3L)BSC370 (BDSC24394). ImpL2^Def20 and ImpL2^Def42 were used in w⁻ background.

Genomic DNA was extracted from homozygous Imp-L2^Def20 and Imp-L2^Def42 stocks, and the coding regions of the N-terminal and C-terminal portions of InR were separately amplified by PCR using the following primers: InR-Ex_F1 (GGAATTCATGTTCAATATGCCACGGGGAG) and InR-Ex_R (GGTCGACCTTGTGCAGGTAACAGACATAG), InR-Ex_F2 (GGCGCAGACTCGAATGGAAAC) and InR-cyto_R (CCTCGAGCGCCTCCCTTCCGATGAATC). The amplified PCR products were used for Sanger sequencing and compared with the reference sequences. Detected missense mutations were further checked using the Drosophila Genetic Reference Panel (DGRP2) inbred lines derived from a natural population (Mackay et al., 2012). A summary of missense mutations is listed in Table S1. We used the amino acid numbering of the reference sequence in the FlyBase ID, FBgn0283499 (NP_001138093.1).

To isolate the InR^Y902C allele, Imp-L2^Def42 mutant males were crossed with w⁻ virgin females, and the resulting F1 heterozygous males were backcrossed with w⁻ virgin females. Around ten F2 females were singly crossed with w⁻ males and used for genotyping to identify females carrying the deletion in the Imp-L2 locus and InR^Y902C mutation. Single InR^Y902C-positive, Imp-L2^Def42-negative F3 females were further crossed with w⁻ males. This process was repeated seven times in total, after which the InR^Y902C allele was re-isolated by crossing to a chromosome III balancer strain that crossed into a w⁻ background. Similarly, a single InR^Y902C-negative, Imp-L2^Def42-positive F3 female was crossed with w⁻ males to generate the backcrossed Imp-L2^Def42 allele. Imp-L2^Def42 genotyping was performed using the primers ImpL2-Def_F1 (GATCTATCATCGTCCAACGTTGGG) and ImpL2-Def_R1 (AGTCATAAGTGCTTCTTGGGCTGC). InR^Y902C genotyping was performed using the primers: InR_F1 (GGCGCAGACTCGAATGGAAAC) and InR_R1 (GTTGGAGGACAAGTCTTCTTCGTC). Amplified PCR products were used for Sanger sequencing.

The animals were reared on fly food (ND) that contained 8 g agar, 100 g glucose, 45 g dry yeast, 40 g corn flour, 4 ml propionic acid and 0.45 g butylparaben (in ethanol) per liter (1× recipe). Yeast paste was not added to the fly tubes in any of the experiments. All the experiments were conducted under non-crowded conditions at 25°C, unless otherwise indicated. Developmental staging was performed as previously described (Okamoto and Nishimura, 2015; Yamada et al., 2020). For the transient starvation experiments, staged larvae were washed in phosphate buffered saline (PBS) to remove all traces of food and transferred to a vial containing an agar-only diet for defined time periods.

Adult flies, pupae and embryos were photographed using a Zeiss Stemi 2000-C stereomicroscope equipped with a Canon PowerShot G15 digital camera. Length and width were measured manually in ImageJ, and the volumes of pupae and embryos were determined using the formula 4/3π(L/2)(l/2)², where L is the length and l is the width. Given that the sex ratios of InR^Y902C mutants at the wandering stage and newly eclosed adults were almost equal, an equal ratio of sexes were assumed to exist in the pupal populations. The weights of adult flies were measured in 1.5 ml tubes using an analytical balance (XS104, Mettler-Toledo Co.).

Adult wings were dissected from each fly using micro-forceps and mounted on a glass slide using 25% glycerol/25% isopropanol. Digital images were acquired using a Zeiss Primo Star microscope equipped with an AxioCam ERc. The wing area was manually traced and measured using either ImageJ or AxioVision software (Zeiss). Cell size and cell number in the adult wing were measured as previously described (Matsushita and Nishimura, 2020). Cell density was manually analyzed by counting the number of wing hairs in a 0.01 mm² area at the third posterior cell of the wing, and cell size and approximate cell number per wing were calculated.

Three virgin females of each genotype raised at either 25°C or 18°C were placed with three control males raised at 25°C in triplicate. Mated females were used for the experiments 4 days after the temperature shift. Fecundity was assessed as the number of eggs laid daily during the following 5 days.

Amino acid sequence alignment was performed using the ClustalW program with default parameters (https://www.genome.jp/tools-bin/clustalw). The following published reference sequences were used: HsInsR, NP_000199.2; MmInsR, NP_034698.2; GgInsR, XP_040509734.1; HsIGF1R, NP_000866.1; MmIGF1R, NP_034643.2; GgIGF1R, NP_990363.1; DmInR, NP_001138093.1; DmSDR, NP_001287331.1; CeDaf-2, NP_497650.4.

In silico prediction analysis was performed using PROVEAN (http://provean.jcvi.org/seq_submit.php) (Choi et al., 2012). If the score is equal to or below a predefined threshold (−2.5), the protein variant is predicted to have a ‘deleterious’ effect. If the score is above the threshold, the variant is predicted to have a ‘neutral’ effect.

Molecular graphics images were prepared using PyMOL version 2.5.2 (https://pymol.org/2/). The mutated residue of Drosophila InR was mapped onto the three-dimensional structure of the extracellular domain of human InsR [Protein Data Bank (PDB) code: 6sof] (Gutmann et al., 2020).

For insulin stimulation, the dissected fat bodies of the early second instar larvae were incubated in Schneider's Drosophila medium (Sigma-Aldrich) supplemented with 0.1 µM bovine insulin (Sigma-Aldrich) for 30 min at room temperature. The samples were fixed for 10 min in 3.7% formaldehyde in PBS containing 0.2% Triton X-100 and processed as previously described (Okamoto and Nishimura, 2015). The following primary and secondary antibodies were used: rabbit anti-pAkt-S505 antibody (1/200, 4054, Cell Signaling Technology) and Alexa Fluor 555-conjugated goat anti-rabbit IgG antibody (1/500, A-21428, Thermo Fisher Scientific). The nuclei were stained with Hoechst 33342 (Thermo Fisher Scientific). Images were acquired with a Zeiss LSM800 confocal microscope and were processed using Fiji and Photoshop CS6 (Adobe Systems). Single confocal sections are shown. For quantification of pAkt signals, mean fluorescence intensity at the cell-cell boundaries was manually analyzed using Fiji.

qRT-PCR analysis was performed as previously described (Okamoto et al., 2013; Okamoto and Nishimura, 2015). Total RNA was prepared using the RNeasy mini kit with an RNase-Free DNase Set (Qiagen), and reverse transcription was performed using PrimeScript RT Master Mix (Takara Bio). qRT-PCR was performed on an ABI PRISM 7500 or StepOnePlus Real-Time PCR system (Thermo Fisher Scientific) using TB Green Premix Ex TaqII (Takara Bio). Transcript levels of the target mRNA were normalized to rp49 (RpL32) levels in the same samples. The primers used in this study are listed in Table S5. Notably, the primer set of InR targets the 5′ region upstream of the P-element insertion of InR⁰⁵⁵⁴⁵.

TAG, trehalose, glycogen and glucose levels were measured as previously described (Nishimura, 2020; Yamada et al., 2020). In brief, samples were homogenized using a pellet pestle in 100 µl of cold PBS containing 0.1% Triton X-100, immediately heat-inactivated at 80°C for 10 min and then cooled to room temperature. The homogenate samples were used to determine the TAG and glycogen levels, and the cleared samples after centrifugation at 12,000 rpm (13,000 g) for 10 min at room temperature were used to determine the trehalose, glucose and protein levels.

To determine TAG levels, the homogenate was mixed with triglyceride reagent (Sigma-Aldrich), incubated at 37°C for at least 30 min, and then cleared by centrifugation. The supernatant was used for the measurement of TAG by using the free glycerol reagent (Sigma-Aldrich). A triolein equivalent glycerol standard (Sigma-Aldrich) was used as the standard. The amount of TAG per fly was normalized to the total protein level per fly, as described below.

To determine sugar levels, the samples were incubated with PBS containing amyloglucosidase (Roche) or bacterially produced recombinant His-tagged Drosophila cTreh (Yoshida et al., 2016) at 37°C overnight. The sample was incubated with PBS without enzymes in parallel to determine glucose levels. The amount of glucose was determined using a glucose assay kit (Sigma-Aldrich) with a serial dilution of glucose as the standard. The trehalose and glycogen concentrations in each sample were determined by subtracting the values of free glucose in the untreated samples. The amounts of trehalose, glycogen and glucose were normalized to the total protein levels.

To determine total protein levels in larval samples, the homogenates were mixed with two volumes of 0.2 N NaOH, vortexed for 10 min at room temperature and heated at 95°C for 15 min to solubilize proteins. The cleared samples after centrifugation at 12,000 rpm (13,000 g) for 10 min at room temperature were used to quantify protein using a BCA protein assay kit (Thermo Fisher Scientific).

Widely targeted metabolomics analysis was performed as previously described (Nishimura, 2020; Yamada et al., 2020). In brief, frozen samples in 1.5 ml plastic tubes were homogenized in 300 µl cold methanol with a 3 mm diameter zirconia bead using a freeze crusher (Tokken Inc.) at 41.6 Hz for 2 min. The homogenates were mixed with 200 µl of methanol, 200 µl of H₂O and 200 µl of CHCl₃ and vortexed for 20 min at room temperature. The samples were then centrifuged at 15,000 rpm (20,000 g) for 15 min at 4°C. The insoluble pellets were used to quantify the total protein using a BCA protein assay kit (Thermo Fisher Scientific). The supernatant was mixed with 350 µl of H₂O, vortexed for 10 min at room temperature and centrifuged at 15,000 rpm for 15 min at 4°C. The aqueous phase was collected and dried down using a vacuum concentrator. The samples were re-dissolved in 2 mM ammonium bicarbonate (pH 8.0) and analyzed by LC-MS/MS.

Chromatographic separations in an Acquity UPLC H-Class System (Waters) were carried out under reverse-phase conditions using an Acquity UPLC HSS T3 column (100 mm×2.1 mm, 1.8 µm particles, Waters) and under hydrophilic interaction liquid chromatography conditions using an Acquity UPLC BEH Amide column (100 mm×2.1 mm, 1.8 µm particles, Waters). Ionized compounds were detected using a Xevo TQD triple quadrupole mass spectrometer coupled with an electrospray ionization source (Waters). The peak area of a target metabolite was analyzed using MassLynx 4.1 software (Waters). The metabolite signals were normalized to the total protein level of the corresponding sample after subtracting the values from the blank sample. P-values were calculated using unpaired two-tailed Welch's t-test in Microsoft Excel. Further statistical analyses, including PCA, were performed in MetaboAnalyst 5.0 (https://www.metaboanalyst.ca/home.xhtml) (Pang et al., 2022). The data were normalized to the median per sample.

Experiments were replicated at least twice using independently reared populations to ensure reproducibility. Alternatively, samples were collected from populations that were independently reared. The experiments were not randomized, and the investigators were not blinded to the fly genotypes during the experiments. The sample sizes were chosen based on the number of independent experiments required for statistical significance and technical feasibility. The sample size for each experiment is indicated in the figures and figure legends. Statistical tests were performed using Microsoft Excel or GraphPad Prism 7. The statistical tests used are described in the figure legends. Statistical significance is presented as follows, except in Figs 4C, 5E, and Fig. S4C: *P<0.05, **P<0.01, ***P<0.001; n.s., not significant. Source data are provided in Table S6.

We thank Hugo Stocker for fly stocks and discussions, the Bloomington Drosophila Stock Center for fly stocks, members of fly laboratories at RIKEN BDR for their valuable support and discussions, Ryota Matsushita and Masayuki Tobo for technical assistance, and the genome resource and analysis unit at RIKEN BDR for their assistance with Sanger sequencing. We are grateful to Naoki Okamoto and Yuto Yoshinari for their comments on the manuscript.

The peer review history is available online at https://journals.biologists.com/dev/lookup/doi/10.1242/dev.201248.reviewer-comments.pdf