A low-sugar diet enhances Drosophila body size in males and females via sex-specific mechanisms

In Drosophila, dietary nutrients impact larval growth to influence final body size. Nutrient quantity promotes growth during larval development, as nutrient-rich conditions favour larger body sizes (Edgar, 2006; Hietakangas and Cohen, 2009; Nijhout et al., 2014). Nutrient quality also regulates larval growth, as individual macronutrients differ in their body size effects. For example, although dietary protein promotes a larger body size across a wide concentration range (Britton and Edgar, 1998; Britton et al., 2002; Edgar, 2006; Shingleton et al., 2017), moderate or high levels of dietary sugar inhibit growth and reduce body size (Musselman et al., 2011; Pasco and Léopold, 2012; Reis, 2016). This suggests a complex relationship between individual macronutrients and body size.

One factor that influences the magnitude of nutrient-dependent changes to Drosophila body size is biological sex (McDonald et al., 2021; Millington et al., 2021a; Shingleton et al., 2017; Stillwell et al., 2010; Teder and Tammaru, 2005). For example, manipulating nutrient quantity by altering dietary protein and carbohydrates causes sex-biased trait size effects (Shingleton et al., 2017). Male and female phenotypic responses to nutrient quality also differ, as the magnitude of protein-dependent changes to body size are larger in females (Millington et al., 2021a). Owing to the widespread use of mixed-sex groups in larval growth studies, however, it remains unclear whether sex-specific body size responses to dietary protein extend to other macronutrients, such as sugar.

Our examination of larval development revealed that lowering dietary sugar augmented the rate of growth and increased body size in both males and females. Indeed, the largest body size in each sex was observed in a diet with no added sugar (0S). Despite an equivalent low-sugar-induced increase in body size, signalling pathway activation, transcriptional responses and metabolic changes were not fully shared between the sexes. In males, our data show that the low-sugar-induced changes to metabolism and body size were triggered by higher insulin/insulin-like growth factor signalling pathway (IIS) activity, where increased IIS was due to improved insulin sensitivity. In females, there was no change in IIS activity or insulin sensitivity in 0S. Instead, females showed transcriptional responses consistent with increased anabolic metabolism, and genetic studies indicated a role for the Target of rapamycin (TOR) pathway in regulating the metabolic and body size effects of the low-sugar diet. Together, our data suggest male and female larvae achieve a larger body size in 0S via distinct mechanisms. This highlights the importance of including both sexes in larval growth studies, as different mechanisms may underlie similar phenotypic responses.

To determine the body size effects of dietary sugar in each sex, we measured pupal volume in white¹¹¹⁸ (w; FBgn0003996) male and female larvae reared in diets with different quantities of sugar. Because dietary sugar represses growth in a mixed-sex larval group (Musselman et al., 2011; Pasco and Léopold, 2012), we started with a widely-used diet (1S) (Lewis, 1960) and removed sugar in a stepwise manner until none remained (0S). In w¹¹¹⁸ females, body size was significantly larger in larvae cultured on a diet with half (0.5S) or one-quarter (0.25S) the amount of sugar found in 1S (Fig. 1A). Interestingly, the largest body size was found in larvae reared in 0S (Fig. 1A). In w¹¹¹⁸ males, body size was similarly larger in larvae reared on 0.5S, 0.25S, and 0S compared with larvae raised on 1S (Fig. 1B). Given that the body size effect of a low-sugar diet was equivalent between the sexes (Fig. 1C; TableS1), the phenotypic responses to dietary sugar were not different between males and females. Although we reproduced this finding using adult weight (Fig. 1D), more studies are needed to understand why the effect of sugar on adult weight was smaller than on pupal volume. More work is also needed to determine whether these growth effects were mediated by dietary sugar alone – as removing cornmeal from 1S to match the caloric content of 0S did not affect body size (Millington et al., 2021a) – or by changing the protein-to-carbohydrate ratio, as both factors affect larval development (Kim et al., 2020; Matzkin et al., 2011; Musselman et al., 2011; Pasco and Léopold, 2012). Nevertheless, we extend previous knowledge by demonstrating an equivalent body size response between the sexes to a low-sugar diet, and by showing a non-sex-specific increase in the larval growth rate (Fig. 1E-H).

IIS has emerged as a key regulator of Drosophila nutrient-dependent growth (Gokhale and Shingleton, 2015; Grewal, 2009; Koyama and Mirth, 2018; Lecuit and Le Goff, 2007; Teleman, 2010). Indeed, high levels of IIS activity promote a larger body size (Böhni et al., 1999; Britton et al., 2002; Brogiolo et al., 2001; Chen et al., 1996; Fernandez et al., 1995; Ikeya et al., 2002; Patel et al., 2003; Poltilove et al., 2000). Given the larger body size of males and females cultured on 0S, we measured diet-induced changes to IIS activity. To quantify IIS activity, we measured mRNA levels of genes co-regulated by transcription factor Forkhead box, sub-group O (Foxo; FBgn0038197), for example Insulin receptor (InR; FBgn0283499), brummer (bmm; FBgn0036449), eukaryotic initiation factor 4E-binding protein (4E-BP; also known as Thor; FBgn0261560). The transcriptional activity of Foxo is normally repressed by IIS; thus, high IIS activity inhibits Foxo and reduces mRNA levels of Foxo target genes (Alic et al., 2011; Gershman et al., 2007; Jünger et al., 2003; Puig and Tjian, 2005; Zinke et al., 2002).

In w¹¹¹⁸ females, mRNA levels of Foxo target genes were not different between larvae reared in 1S and 0S (Fig. 2A). In contrast, mRNA levels of Foxo targets were significantly lower in w¹¹¹⁸ male larvae in 0S (Fig. 2B). Given that we confirmed the accuracy of our larval sorting by detecting the correct male- and female-specific isoforms of sex determination genes (Fig. S1A), these findings suggest that a low-sugar diet enhanced IIS activity in males but not females. To confirm this, we monitored the subcellular localization of green fluorescent protein (GFP) fused to a pleckstrin homology (PH) domain (GFP-PH), which shows increased membrane localization when IIS activity is high (Britton et al., 2002). No diet-dependent change in membrane GFP-PH localization occurred in w¹¹¹⁸ females (Fig. 2C); however, male larvae cultured in 0S had higher membrane GFP-PH than males raised on 1S (Fig. 2D). Because feeding behaviour was not different between the sexes (Fig. 2E), our data support a model in which the low-sugar diet augments IIS activity in male but not female larvae.

One potential reason for higher IIS activity in males reared on 0S is improved insulin sensitivity, as moderate levels of dietary sugar cause insulin resistance in a mixed-sex larval group (Musselman et al., 2011; Lourido et al., 2021; Pasco and Léopold, 2012). To test this, we monitored insulin sensitivity by quantifying GFP-PH membrane localization in larval fat bodies with and without human insulin stimulation (Pasco and Léopold, 2012). Although insulin stimulation significantly enhanced GFP-PH membrane localization in females reared on both diets (Fig. 2F), insulin stimulation only augmented GFP-PH membrane localization in males reared on 0S (Fig. 2G). This suggests that males reared on 0S had higher insulin sensitivity than males reared on 1S, whereas females were insulin sensitive in both contexts. Supporting this, we observed a significant downregulation of puckered (puc; FBgn0243512), a gene that is lower in insulin-sensitive larvae (Lourido et al., 2021; Pasco and Léopold, 2012), in males but not females reared on 0S (Fig. 2H). Thus, higher IIS activity in male larvae reared in 0S was caused by improved insulin sensitivity, an effect we did not observe in females. This extends our understanding of sex differences in the nutrient-dependent regulation of IIS (Millington et al., 2021a) and shows that, in flies as in mammals (Guerre-Millo et al., 1985; Mittendorfer, 2005; Macotela et al., 2009; Yki-Järvinen, 1984), females show higher insulin sensitivity than males in some contexts.

Given the known role of IIS in regulating body size, we asked whether the male-specific increase in IIS activity between 1S and 0S was required for the low-sugar-induced increase in body size. To test this, we measured body size in male and female larvae carrying mutations in IIS that support normal growth, but which blunt high levels of IIS activation (InR^E19/+) (Chen et al., 1996; Millington et al., 2021a; Rideout et al., 2015). Although body size was larger in w¹¹¹⁸ control males reared on 0S (Fig. 3A), 100% of the low-sugar-induced increase in body size was blocked in InR^E19/+ male larvae (Fig. 3A; genotype:diet interaction P<0.0001). This suggests the low-sugar induced increase in IIS activity was required in males to achieve a larger body size. In females, only 48% of the low-sugar-induced increase in body size was blocked in InR^E19/+ larvae (Fig. 3B). Given that the magnitude of genotype effects on the body size response to diet was larger in males than in females (sex:diet:genotype interaction P=0.0114), our data indicates a male-biased phenotypic response to reduced IIS function, an effect we replicated across independent experiments despite modest interexperiment variation in the magnitude of the low-sugar-induced increase in body size (Fig. S2A-D).

We further reproduced this male-biased body size effect in larvae lacking the coding sequences for Drosophila insulin-like peptide 3 (dilp3; Ilp3; Fbgn0044050) (Fig. 3C,D; sex:diet:genotype interaction P=0.0003), Drosophila insulin-like peptide 2 (dilp2; Ilp2; Fbgn0036046) (Fig. 3C,D; sex:diet:genotype interaction P=0.0627) and in flies lacking the coding sequences for dilp2, dilp3 and Drosophila insulin-like peptide 5 (dilp5; Ilp5; Fbgn0044038) (genotype dilp2-3,5) (Fig. S3A,B; sex:diet:genotype interaction P<0.0001); however, loss of dilp5 had no effect on the low-sugar-induced increase in body size in either sex (Fig. 3C,D). Together with the male-specific increase in IIS activity between 1S and 0S, the male-biased body size effect of reduced IIS in a low-sugar context confirms that the signalling and genetic mechanisms that promote larval growth in 0S are not fully shared between the sexes.

IIS influences body size by triggering profound changes in gene expression (Alic et al., 2011; Bülow et al., 2010; Grewal, 2009; Guertin et al., 2006; Li et al., 2010; Musselman and Kühnlein, 2018; Teleman et al., 2008; Tiebe et al., 2015; Webb et al., 2016; Zinke et al., 2002). Because we observed sex-specific regulation of IIS in a low-sugar diet, we performed an unbiased analysis of transcriptional changes in males and females reared on 1S and 0S. We found significant differences in the transcriptional response to a low-sugar diet between the sexes (Fig. 4A), and show that diet affects sexual dimorphism in gene expression (Fig. 4B), consistent with previous reports (Camus et al., 2019; Jaime et al., 2017). Interestingly, the proportion of genes differentially regulated by a low-sugar diet was higher in males (Table S2): 298 (8.2%) of differentially expressed genes were unique to females and 1832 (50.3%) unique to males (Fig. 4C). A low-sugar diet therefore causes a distinct transcriptional response in each sex (Fig. S4A,B). Indeed, the majority (58%) of enriched Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways were different between males and females (P_adj<0.05; Fig. 4D). For example, genes in the ‘FoxO signaling pathway’ category were only downregulated in males in 0S (Fig. 4D, bold text). Given that Foxo is repressed when IIS activity is high, this further supports a male-specific IIS increase in 0S.

One overrepresented biological process in both sexes was metabolic regulation: 80.7% of genes differentially expressed in a low-sugar context were linked with metabolism (Fig. 4D). We therefore examined several metabolic parameters in male and female larvae in 1S and 0S to determine the physiological significance of these sex-specific gene expression responses. Although we found non-sex-specific changes to whole-body triglyceride and glucose levels on a low-sugar diet (Fig. S5A,B), we observed sex-specific effects on whole-body protein, glycogen and trehalose levels (Fig. 4E-G). Whole-body protein levels were significantly higher only in w¹¹¹⁸ females reared in 0S (Fig. 4E), whereas whole-body glycogen and trehalose were significantly higher only in males reared on 0S (Fig. 4F,G). Importantly, these metabolic changes cannot be attributed to sex differences in the relationship between organ and body size (Fig. S6A).

Instead, the male-specific increase in glycogen was likely due to increased IIS activity in 0S, as the low-sugar-induced increase in glycogen was blunted in InR^E19/+ males (Fig. 4H), aligning with the known role of IIS in regulating carbohydrate metabolism (Mattila and Hietakangas, 2017). Thus in males, the low-sugar diet augments IIS activity due to improved insulin sensitivity, leading to higher whole-body glycogen levels and a larger body size. In females, these mechanisms are not fully shared: the low-sugar diet caused no change in insulin sensitivity, and reduced IIS function did not block the female-specific increase in body size or whole-body protein levels (Fig. S7A).

To gain deeper insight into the phenotypic effects of a low-sugar diet in females, we focused on female-specific metabolic and transcriptional changes. The main metabolic phenotype was increased whole-body protein (Fig. 4E), and female-specific transcriptional changes included higher mRNA levels of genes related to glycolysis and gluconeogenesis, folate biosynthesis, biosynthesis of amino acids and the pentose phosphate pathway, and lower mRNA levels of genes linked with lysosomes, arachidonic acid metabolism and fatty acid degradation (down). Together, these changes indicate a general upregulation of anabolic processes in 0S. This anabolic regulation is unlikely to depend on Mondo (FBgn0032940) and bigmax (FBgn0039509), two key regulators of sugar sensing and tolerance (Mattila et al., 2015; Havula and Hietakangas, 2012; Havula et al., 2013; Havula and Hietakangas, 2018), as we observed non-sex-specific changes to known Mondo-bigmax target genes (Fig. S8A-D; Table S2). Another key regulator of anabolic metabolism is the TOR pathway (Wullschleger et al., 2006). In multiple organisms, TOR promotes mRNA translation and protein synthesis, the pentose phosphate pathway and glycolytic flux (Dunlop and Tee, 2009; LaPlante and Sabatini, 2013; Wullschleger et al., 2006), and represses catabolic processes such as autophagy and lysosomal biogenesis (Puertollano, 2014).

Given that the female-specific metabolic and transcriptional changes in 0S are consistent with a potential role for TOR, we asked whether TOR function was required in females for the metabolic and body size responses to a low-sugar diet. In female larvae with reduced TOR function (Target of rapamycin, Tor; FBgn0021796; genotype Tor^ΔP/+), the low-sugar-induced increase in whole-body protein was blocked (Fig. 4I). This provides genetic evidence that TOR plays a role in regulating the diet-dependent increase in protein levels, aligning with the well-known effect of TOR on protein synthesis (Wullschleger et al., 2006). Because TOR controls growth in part due to regulation of protein synthesis (Killip and Grewal, 2012; Marshall et al., 2012; Rideout et al., 2012; Ghosh et al., 2014; Terada et al., 1995; Barbet et al., 1996; Zhang et al., 2000), we monitored the low-sugar-induced increase in body size in larvae with reduced TOR function. In w¹¹¹⁸ females, larvae reared on 0S were significantly larger than genotype-matched larvae raised on 1S (Fig. 4J); however, 100% of the low-sugar-induced increase in body size was blocked in Tor^ΔP/+ female larvae (Fig. 4J; genotype:diet interaction P<0.0001). This suggests that TOR function in females contributes to their larger body size in 0S. In males, although the low-sugar-induced increase in body size was 50% blocked in Tor^ΔP/+ larvae (Fig. 4K; genotype:diet P<0.0001), the magnitude of genotype effects on the diet-induced increase in body size was greater in females than in males (sex:diet:genotype interaction P=0.0303). This female-biased effect was robust, as we reproduced the sex:diet:genotype interaction in Tor^ΔP/+ larvae across multiple biological replicates (Fig. S9A-D). These data provide genetic evidence that TOR plays a female-biased role in regulating the metabolic and body size responses to a low-sugar diet.

Based on these metabolic and body size effects, we next examined TOR regulation in males and females in 1S and 0S. Our RNA-seq data showed that mRNA levels of two TOR-responsive genes, unkempt (unk; FBgn0004395) and cabut (cbt; FBgn0043364) (Guertin et al., 2006; Ingaramo et al., 2020; Tiebe et al., 2015), were differentially regulated in a low-sugar diet in females but not in males (Fig. 4L). The low-sugar diet-induced decrease in both unk and cbt mRNA is consistent with increased TOR activity (Guertin et al., 2006; Ingaramo et al., 2020; Tiebe et al., 2015), suggesting that a low-sugar diet augments TOR in females but not males. When we examined levels of phosphorylated Ribosomal protein S6 kinase (S6k; FBgn0283472) (p-S6k), a downstream target of TOR related to ribosomal biogenesis (Miron et al., 2003; Radimerski et al., 2002; Rintelen et al., 2001; Wullschleger et al., 2006; Zhang et al., 2000), we found a trend toward increased p-S6k levels in both sexes (P=0.0625) potentially related to an increase in total S6k levels, as the trend was abolished when we monitored the ratio of p-S6k/t-S6k (Fig. S10A-C). Although this S6k regulation did not align with the sex-biased regulation of TOR transcriptional targets or our genetic data supporting a female-biased role for TOR in a low-sugar context, TOR has many downstream targets that mediate its effects on diverse cellular processes. Thus, future studies will need to investigate TOR targets beyond S6k for sex-biased regulation, and test whether additional targets mediate the female-biased changes to metabolism and body size in a low-sugar context. This highlights the importance of studying IIS and TOR biology in both sexes to gain a deeper understanding of how these key nutrient-sensing pathways couple metabolic regulation with diet.

Overall, our findings support a clear role for the upregulation of IIS as a key mechanism underlying the metabolic and body size responses in male larvae reared on a low-sugar diet. In females, TOR had female-biased effects on the low-sugar-induced changes to metabolism and body size. Together with the differential transcriptional responses in 0S between males and females, these data suggest that the mechanisms underlying the low-sugar-induced increase in body size are not fully shared between the sexes. This highlights the importance of including both males and females in larval growth studies, as not all mechanisms will be shared between the sexes even in contexts where the phenotypic response is equivalent.

Our 1S diet consists of 20.5 g/l sucrose, 70.9 g/l D-glucose, 48.5 g/l cornmeal, 45.3 g/l yeast, 4.55 g/l agar, 0.5 g CaCl₂·2H₂O, 0.5 g MgSO₄·7H₂O, 11.77 ml acid mix (propionic acid/phosphoric acid). Our 0S diet consists of 48.5 g/l cornmeal, 45.3 g/l yeast, 4.55 g/l agar, 0.5 g CaCl₂·2H₂O, 0.5 g MgSO₄·7H₂O, 11.77 ml acid mix. Details of 0.75S, 0.5S and 0.25S diets can be found in Table S4. Larvae were raised at a density of 50 animals per 10 ml food at 25°C, and sexed by gonad size as mid third instar larvae at the time of collection for gene expression or metabolic experiments [108 h after egg-laying (AEL)]. For pupal volume experiments, animals were separated by sex between 0-12 h after puparium formation. Adult flies were maintained at a density of 20 flies per vial in single-sex groups. For all metabolic, gene expression, imaging, insulin stimulation and western blotting experiments, larvae were collected at 108 h AEL.

The following fly strains from the Bloomington Drosophila Stock Center were used: w¹¹¹⁸ (#3605), InR^E19 (#9646), Tor^ΔP (#7014). Additional fly strains include: dilp2, dilp3, dilp5 and dilp2-3,5 (Grönke et al., 2010). All fly strains were backcrossed for at least six generations, in addition to extensive prior backcrossing (Grönke et al., 2010; Millington et al., 2021a,b).

Pupal volume and adult weight were measured as previously described (Delanoue et al., 2010; Millington et al., 2021a,b; Rideout et al., 2015). For pupal volume, one biological replicate consists of one pupa; for adult weight, one biological replicate consists of one tube of ten flies.

Feeding behaviour was quantified as number of mouth-hook contractions per 30 s. One biological replicate represents one larva.

Time to pupariation was measured as previously described (Millington et al., 2021a). Time to 50% pupariation was calculated per replicate and used for quantification and statistical analysis. One biological replicate consists of 50 animals in one vial, with pupae sexed at 12 h intervals to determine pupariation in both sexes.

One biological replicate consists of ten female or male larvae. Larvae were frozen on dry ice and homogenized in appropriate buffers to measure whole-body lipid, protein, glucose, glycogen and trehalose levels. All assays were performed as described in Tennessen et al. (2014) and Wat et al. (2020) and according to manufacturer's instructions.

RNA extraction and cDNA synthesis were performed as previously described (Marshall et al., 2012; Rideout et al., 2012, 2015; Wat et al., 2020). Briefly, each biological replicate consists of ten w¹¹¹⁸ larvae frozen on dry ice and stored at −80°C. Each experiment contained 3-4 biological replicates per sex, and each experiment was performed at least twice. RNA was extracted using 500 µl Trizol (Thermo Fisher Scientific, 15596018) and precipitated using isopropanol and 75% ethanol. Pelleted RNA was resuspended in 200 µl molecular biology grade water (Corning, 46-000-CV) and stored at −80°C until use. For cDNA synthesis, an equal volume of RNA per reaction was DNase-treated and reverse transcribed using the QuantiTect Reverse Transcription Kit (Qiagen, 205314).

qPCR was performed as previously described (Marshall et al., 2012; Rideout et al., 2012, 2015; Wat et al., 2020). Each biological replicate consists of ten larvae. Values displayed in each graph represent the fold change for the mRNA of a gene, normalized to Act5c and β-tub, housekeeping genes that were not differentially regulated between the sexes or between 1S and 0S. A complete primer list is included in Table S5.

Samples were generated as previously described (Millington et al., 2021a). We loaded 20 µg of protein per lane, separated on a 12% SDS-PAGE gel in SDS running buffer, and transferred onto a nitrocellulose membrane (Bio-Rad) for 2 h at 40 V on ice. Membranes were incubated for 24 h in blocking buffer at 4°C [5% milk or 5% bovine serum albumin in Tris buffered saline, 0.1% Tween 20 (TBST)] and subsequently incubated with primary antibodies overnight at 4°C. Anti-pS6k (#9209, Cell Signaling Technology), anti-tS6k (gift from A. Teleman, German Cancer Research Center and Heidelberg University, Germany) and anti-Actin (#8432, Santa Cruz Biotechnology) were used at 1:1000. After 3×2 min washes in 0.1% TBST, HRP-conjugated secondary antibodies were used at 1:5000 for pS6k (#65-6120, Invitrogen) and 1:3000 for actin (#7076, Cell Signaling Technology). Membranes were washed (3×2 min, 2×15 min) in 0.1% TBST, washed 1×5 min in TBS, and finally Pierce ECL was applied according to the manufacturer's instructions (#32134, Thermo Fisher Scientific). To measure the concentration of all proteins loaded in a sample, we used the stain-free labelling system from Bio-Rad (#1610185) according to manufacturer's instructions.

To measure IIS activity in the fat body, late third instar larvae were collected at 108 h AEL and washed in PBS. Twelve to fifteen larvae were collected for each batch and inverted in PBS; inverted carcasses were fixed in 4% paraformaldehyde in PBS on a rocking platform at room temperature for 40 min, then washed twice in PBS. Fixed fat bodies were mounted in SlowFade Diamond mounting media (S36972, Thermo Fisher Scientific) before imaging. To measure insulin sensitivity, late third instar larvae were collected at 108 h AEL and washed in PBS once. The larvae were then moved into a dish with ice-cold Schneider's Drosophila Media (21720024, Thermo Fisher Scientific) and inverted. Inverted carcasses (with fat body) were immediately moved to a separate dish with fresh Schneider's. Once all dissections were complete, Schneider's Media was replaced with 2 ml of fresh Schneider's Drosophila Media containing a final concentration of 0.5 μM recombinant human insulin (I2643, Sigma-Aldrich) and incubated at room temperature for 20 min. Carcasses were then washed once with PBS and fixed with 4% paraformaldehyde for 40 min, washed twice with PBS, and mounted in SlowFade Diamond mounting media (S36972, Thermo Fisher Scientific). Images were acquired on a Leica SP8 confocal microscope with 20× objective and quantified using Fiji (Schindelin et al., 2012). For each cell, the average GFP intensity of a short stretch of cell membrane and a small stretch of adjacent were quantified, and the ratio of membrane to cytosolic GFP was calculated.

RNA from w¹¹¹⁸ males and females reared on either the 1S or 0S diet was sequenced to determine gene expression. Quality control of samples was performed using the Agilent 2100 Bioanalyzer to confirm sample quality. Sequencing was performed using a NextSeq 500 sequencer. Sequence quality was assessed using FastQC on Illumina BaseSpace. Reads were aligned to Drosophila melanogaster genome (UCSC dm3) using STAR aligner with default settings. All raw sequencing data have been deposited and are publicly available through the NCBI Gene Expression Omnibus database (accession number GSE198435). For differential expression analysis, fragments per kilobase of transcript per million mapped reads were calculated using the DESeq2 normalization method. Log fold changes and significance for differential expression were determined and a significance threshold of 0.05 was selected for adjusted P-values determined with Bonferroni correction. Differentially expressed genes are presented in Table S2.

GraphPad Prism (GraphPad Prism version 8.4.3 for Mac OS X) was used for all statistical tests and for figure preparation. For normally distributed data with comparisons between two experimental groups, an unpaired two-tailed Student's t-test was used to determine significance. For normally distributed data with comparisons between more than two experimental groups, one-way ANOVAs were used to determine significance. In order to determine significant interactions between sex, diet and genotype, multivariate ANOVAs (two-way and three-way ANOVAs) were used on normally distributed data. For non-normally distributed data, we used Wilcoxon signed rank tests for comparisons between two groups. A complete list of P-values and full details of statistical tests is provided in Table S1. Raw values for all data collected and displayed in this paper are available in Table S3.

We thank Dr Linda Partridge for sharing the dilp2, dilp3, dilp5 and dilp2-3,5 mutant strains. We thank the Bloomington Drosophila Stock Center (NIH P40OD018537) for stocks used in this study. We thank Aurelio Teleman for the kind gift of the total S6k antibody. We acknowledge critical resources and information provided by FlyBase (Thurmond et al., 2018); FlyBase is supported by a grant from the National Human Genome Research Institute at the US National Institutes of Health (U41 HG000739) and by the British Medical Research Council (MR/N030117/1). We acknowledge that our research takes place on the traditional, ancestral and unceded territory of the Musqueam people; a privilege for which we are deeply grateful.

The peer review history is available online at https://journals.biologists.com/dev/article-lookup/doi/10.1242/dev.200491.