Animal survival in fluctuating environments depends on the ability to modulate their developmental pace in response to nutrient availability, a phenomenon known as developmental plasticity. In Drosophila larvae, we uncover a critical endocrine mechanism that coordinates this process under conditions of amino acid restriction. We identify the peptide hormone Limostatin as an enterokine, produced by a small population of larval midgut enteroendocrine cells, that acts systemically to inhibit the expression and release of dIlp2, a major insulin-like peptide controlling developmental progression. Limostatin expression and secretion by enteroendocrine cells is triggered by reduced amino acid availability through an inter-organ relay involving the fat body and neuroendocrine insulin-producing cells in the brain. In turn, Limostatin participates in a feedback control loop that slows down developmental progression once systemic nutrient shortage is sensed. This bidirectional gut–brain axis enables larvae to preserve viability under nutritional stress. Our findings define the larval gut as a nutrient-sensitive endocrine organ and position Limostatin as a key regulator of developmental plasticity. Our work expands the concept of decretins to include developmental pace control, suggesting that enterokines that regulate IGF signaling, rather than insulin release per se, may represent an evolutionarily conserved or convergent strategy in regulating developmental plasticity.

Lst is a peptide hormone regulating developmental progression and survival under malnutrition

In our recent study [28], we performed transcriptomic profiling of dissected larval midguts from Drosophila reared on a low-yeast diet, a model of malnutrition that elicits developmental adaptation, characterized by reduced larval growth rate and delayed pupariation in wild-type (WT) animals. Given the known role of the microbiome in buffering nutritional stress, we compared germ-free larvae with animals mono-associated with Lactiplantibacillus plantarum^WJL (Lp^WJL, referred as Lp here), a well-characterized member of the Drosophila gut microbiota that recapitulates the growth-promoting effects of a complex microbiome under malnutrition [25]. Among the differentially regulated transcripts, we identified Limostatin (Lst) as significantly upregulated in germ-free larvae raised under malnutrition (Fig 1A) [28]. Lst was previously characterized in adult flies as a peptide hormone secreted by corpora cardiaca during fasting, where it suppresses dIlp production and release from the brain [27]. Our data suggest that Lst may also contribute to the adaptive growth response of germ-free larvae exposed to malnutrition. To test this hypothesis, we analyzed Lst loss-of-function mutants. In line with previous findings, heterozygous and homozygous Lst null mutant (Lst¹) animals showed normal developmental progression under standard nutritional conditions [27] (Fig 1B), but a severe developmental lethality was observed in homozygous mutants (Fig 1C). However, under malnourished conditions (low yeast diet, 7 g/L), both heterozygous and homozygous Lst null mutants exhibited significantly accelerated development (Fig 1D). Yet, heterozygous animals displayed reduced viability during development and homozygous animals had high lethality (Fig 1E) relative to their WT siblings and normal nutrition controls.

To determine whether Lst is specifically required for viability under malnutrition, we adopted a trans-heterozygous strategy by crossing the Lst¹ null allele (Lst^−/−) with a second allele carrying a CRIMIC insertion within the Lst locus, located at the beginning of the sole intron of the Lst gene (Lst^CR01853, hereafter referred to as Lst^CRIMIC). To bypass potential early embryonic lethality, we specifically selected first-instar (L1) larvae of each genotype and monitored larval-to-adult development and viability. Strikingly, upon normal nutrition, both Lst^−/− and Lst^−/CRIMIC larvae displayed normal developmental timing (Fig 1F) and survival rates (Fig 1G) comparable to controls once they enter the L1 stage after hatching. These results indicate that the reduced viability initially observed upon normal nutrition in Lst^−/− animals (Fig 1C) results from an embryonic lethality and/or reduced fertilization/oogenesis efficiency in this specific homozygous context. Importantly, under malnutrition conditions, all mutant genotypes (Lst^CRIMIC/+ - Lst^CRIMIC/− - Lst^−/−) exhibited significantly accelerated developmental timing (Fig 1H) coupled with an increased lethality (Fig 1I). These findings strengthen our conclusion that Lst is a critical regulator of adaptive larval growth, restraining developmental pace in order to preserve larval viability during nutrient stress.

These phenotypes were independently confirmed using two distinct Lst-targeting RNAi lines driven by Lst-GAL4 driver [27] (S1A–S1D Fig). Importantly, despite their accelerated developmental progression, Lst-RNAi animals did not exhibit increased food intake (S1E–S1F Fig), indicating that the enhanced developmental rate is not attributable to elevated feeding activity. Together, these findings establish Lst as a peptide hormone that coordinates developmental progression and survival in response to nutritional stress and microbiome depletion. Its expression under these conditions suggests a role as a critical effector in the endocrine network governing adaptive developmental timing.

Nutritional stress activates Lst in anterior midgut EECs

Prompted by the upregulation of Lst transcripts in dissected midguts from germ-free larvae raised under malnutrition (Fig 1A), we examined the tissue-specific expression of Lst using a Lst-GAL4 > GFP reporter line [27]. As previously reported in adults, Lst-GAL4 drives GFP expression in the larval corpora cardiaca (Figs 2A and S2A). However, we also detected additional GFP-positive cells in a discrete region of the anterior midgut, just anterior to the copper cell region (Figs 2B, 2C, and S2B), a signal we did not detect in adult midguts. Although we cannot exclude the possibility that the Lst-GAL4 driver does not fully recapitulate endogenous Lst expression, all GFP-positive cells in this region co-localized with Prospero, a marker of EECs (Fig 2B and 2C). Of note, GFP-labeled cells were basally located in the epithelium with no obvious apical projections to the intestinal lumen (S1 Movie). We observed less than 10 GFP-positive cells per midgut with no marked differences across diet or microbiome conditions (Fig 2D), indicating a restricted and stereotyped expression pattern. To further confirm endogenous Lst expression, we generated affinity-purified antibodies targeting a 15-amino acid peptide of the Lst protein. Immunostaining with these antibodies revealed specific signal in the same anterior EEC subset; notably, this signal was completely abolished in the midguts of Lst¹ animals, hereafter referred to as the Lst ⁺ EEC population (Figs 2B, 2C, and S3A).

To investigate the regulation of Lst expression, we performed RT-qPCR analyses on dissected midguts and corpora cardica from larvae raised under either conventional or malnutrition diets, in both germ-free and Lp-associated conditions (Fig 2E and 2F). These experiments revealed a robust and selective induction of Lst expression in the midgut of malnourished germ-free animals, while in contrast to adults [27], expression in the corpora cardiaca remained unchanged. We further found that severe dietary restriction alone was sufficient to trigger strong induction of Lst expression, even in the absence of microbiota alteration: Lst transcript levels were increased approximately 6-fold in the midguts of Lp-associated larvae raised on an extremely low-yeast diet (3 g/L) compared to those on a low-yeast diet (7 g/L) (Fig 2G).

To extend these findings at the protein level, we next performed complementary immunostaining analyses using anti-Lst antibody. We detected Lst protein in the relevant EEC population, and observed differences in steady-state intracellular staining across nutritional and microbiome conditions that paralleled the transcriptional changes (Fig 3A). We then sought to visualize intracellular accumulation of Lst by acutely blocking vesicle-mediated secretion. To this end, we expressed the temperature-sensitive dynamin allele shibire^ts [29] in Lst⁺ cells and shifted larvae to the restrictive temperature for 3 hours to inhibit secretion. Under these conditions, we observed a marked intracellular accumulation of Lst protein, which was substantially stronger in undernourished germ-free larvae than in normally fed controls (Fig 3B). Together, these experiments support the conclusion that the induction of Lst mRNA under malnutrition is accompanied by increased Lst protein production and secretion. Together, these results identify a previously unrecognized site of Lst expression during larval stage and demonstrate that Lst is specifically expressed and secreted by this discrete population of anterior midgut EECs in response to combined mild nutritional stress and microbiota depletion or to severe malnutrition alone.

Lst from EECs regulates adaptive growth and survival under malnutrition and microbiota depletion

We next sought to determine which pool of Lst—produced by the corpora cardiaca and/or by EECs—contributes to the regulation of adaptive growth under nutritional and microbiome stress. Consistent with earlier findings showing that ectopic Lst expression in the larval fat body can delay developmental progression [27], we found that targeted overexpression of Lst in either EECs (using an EEC-specific GAL4 driver; EEC-GAL4 [30]; Fig 4A) or corpora cardiaca cells (using the AKH-GAL4 driver [31]; Fig 4B) was sufficient to impair developmental progression in larvae mono-associated with Lp. These results indicate that Lst is sufficient to modulate maturation rates, independently of its tissue of origin.

To test which Lst source is necessary for adaptive growth regulation under nutritional stress, we used tissue-specific RNA interference to knock down Lst expression. We expressed two independent RNAi constructs targeting Lst in either EECs (EEC-GAL4, Figs 4C, 4D, S3B, S3C, S4A, S4B or Voilà-GAL4 [32], S4C) or in corpora cardiaca cells (AKH-GAL4, Figs 4E, 4F, S4D, and S4E). These experiments were performed in both germ-free and Lp-associated larvae, under either conventional or malnutrition diets (7 g/L low-yeast diet). Strikingly, Lst knockdown in EECs (Figs 4C, 4D, 4G, 4H, and S4A–S4C), but not in corpora cardiaca cells (Figs 4E, 4F, S4D, and S4E) was sufficient to phenocopy the altered adaptive maturation (Figs 4C, 4D, and S4A–S4C) and low viability (Fig 4G and 4H) observed in whole-animal Lst null heterozygous and trans-heterozygous mutants (Fig 1C–1E, 1H, and 1I) under germ-free malnutrition conditions. Importantly, despite their accelerated developmental progression, EEC > Lst-RNAi animals did not exhibit increased food intake (S4F and S4G Fig). These findings demonstrate that Lst derived from EECs is necessary for restraining developmental progression and ensuring viability when animals are challenged by both nutrient deprivation and microbiota loss.

Lst expression in EECs is sensitive to amino acid of dietary and microbial origins

Our findings indicate that Lst expression in EECs is modulated by fluctuations in both diet and microbiota. While Lst was previously shown to be regulated by dietary sugar levels in adult corpora cardiaca cells [27], it is unclear which nutrients influence its expression in the larval midgut. Notably, conventional inactivated yeast-based (oligidic) diets do not allow for a precise dissection of the nutritional components responsible for Lst regulation, as yeast provides a complex mixture of macro- and micronutrients. Although commonly referred to as “low-protein” diets, yeast-reduced regimens likely affect multiple nutritional inputs beyond amino acids availability. To directly test whether dietary amino acids regulate Lst expression and function, we employed the FLYAA holidic diet (HD) system, a chemically defined nutritional framework previously established for dissecting nutrient-microbiota–host interactions in Drosophila [23,33]. Within this system, we altered total amino acid content while maintaining their relative proportions. Specifically, we used a High Amino Acid Holidic Diet (HAAHD; 21.38 g/L total AAs) and a Low Amino Acid Holidic Diet (LAAHD; 8.55 g/L total AAs), the latter of which has been shown to impair developmental growth [23]. Using these HDs, we found elevated Lst transcripts levels in the midgut (Fig 5A) but not in corpora cardiaca (S5A Fig) under amino acid restriction and microbiota depletion, mirroring our previous observations in germ-free animals fed oligidic low-yeast diets. Functionally, we confirmed that Lst is required for coordinating development under these conditions: germ-free Lst¹ mutants (Fig 5B) and germ-free animals subjected to tissue-specific Lst knockdown in EECs (Figs 5C, 5D, S5B, and S5C), but not in corpora cardiaca cells (S5D–S5G Fig), exhibited altered adaptive maturation under the LAAHD regimen. Together, these findings demonstrate that Lst acts as a key hormonal regulator of adaptive developmental growth in response to dietary amino acid availability and microbiota status.

We next explored whether microbiota-mediated nutrient provisioning contributes to Lst repression. Given the role of amino acid availability in regulating Lst expression, and our prior demonstration that commensal bacteria, including Acetobacter pomorum (Ap) and Lp strains, can supplement amino acids to their juvenile host during malnutrition [23], we hypothesized that Lp represses Lst expression by supplying limiting amino acids. To test this hypothesis, we raised germ-free larvae on either HAAHD or LAAHD that were selectively depleted or supplemented with the specific amino acids that Lp is able to produce (His, Lys, Met, Phe, Thr and Asp. Fig 5E and 5F). Consistent with our model, Lst expression was significantly repressed and developmental progression improved when His, Lys, Met, Phe, Thr and Asp were added to the LAAHD (Fig 5E and 5F). Conversely, reducing the concentration of those same AAs from the HAAHD to their levels in LAAHD led to Lst induction and maturation delay (Fig 5E and 5F). Similarly, association of germ-free larvae raised on LAAHD with Ap, a commensal bacteria capable of producing all amino acids, antagonized Lst expression upon malnutrition (S5H Fig), indicating that microbial production of amino acids may contribute to Lst-dependent developmental growth regulation.

Cellular amino acid sensors are dispensable for Lst induction in EECs

We next sought to identify how amino acid restriction leads to Lst induction specifically in EECs. A straightforward hypothesis is that Lst⁺ EECs directly sense amino acid scarcity in the gut lumen through conserved intracellular nutrient-sensing pathways. Two major evolutionarily conserved signaling hubs govern cellular responses to amino acid availability: the Gcn2 kinase, which is activated by uncharged tRNAs and signals amino acid deprivation [34], and the target of rapamycin complex 1 (TORC1), a central regulator of growth and metabolism activated by intracellular amino acids and negatively regulated by the Tsc1/2 complex [35]. To assess the functional requirement for these pathways in Lst⁺ EECs, we used EEC-specific RNAi to knock-down Tsc1 or Gcn2 (Fig 6A, 6B, 6C, and 6D, respectively). In both cases, Lst induction (Fig 6A and 6C) and developmental maturation rates (Fig 6B and 6D) in undernourished germ-free larvae remained unaffected, indicating that neither Gcn2 nor repression of TORC1 signaling in EECs contribute to Lst expression under amino acid-restricted conditions. Together, these results suggest that Lst induction in response to amino acid limitation occurs independently of the canonical intracellular amino acid sensors Gcn2 and TORC1 in EECs.

Lst induction in EECs is controlled by reduced systemic insulin signaling via FoxO

Since Lst expression in EECs is not regulated by classical intracellular amino acid sensors such as Gcn2 or TORC1, we next explored whether nutrient-sensitive systemic endocrine signals might be involved. In Drosophila, the insulin/insulin-like growth factor signaling (IIS) pathway is a central hormonal regulator of nutrient-dependent growth, and its activity is markedly reduced under protein malnutrition. Decrease in IIS leads to the nuclear translocation of the transcription factor FoxO, which drives transcriptional responses to nutrient stress [36]. To test whether reduced IIS signaling and FoxO activity regulate Lst expression, we first monitored intestinal expression of InR, a transcriptional target of FoxO and a sensitive negative marker of IIS activity. InR transcripts were elevated in dissected midguts of germ-free larvae raised on a low-yeast diet compared to Lp-associated controls (Fig 7A), indicating elevated FoxO activity, hence reduced IIS signaling, in the intestinal epithelium of germ-free animals under malnutrition. Consistently, phosphorylated protein kinase B (p-Akt) immunoreactivity, a marker of IIS pathway activation [37], was increased in Lst⁺ EECs of Lst > GFP larvae, and likely also in some adult midgut precursors, in response to Lp association, whereas surrounding enterocytes showed no detectable changes (Fig 7B).

We next tested whether FoxO activity is required for Lst induction. Tissue- and cell type-specific RNA interference against FoxO in enterocytes (mex-GAL4; Fig 7C), EECs (EEC-GAL4; Fig 7D), and Lst⁺ cells (Lst-GAL4; S6A Fig) revealed that FoxO knockdown in EECs or in Lst⁺ cells, but not in enterocytes, reduced Lst expression in germ-free animals reared under malnutrition. Conversely, ectopic expression of FoxO in EECs under microbiota-replete conditions was sufficient to drive Lst expression (Fig 7E) and significantly repress larval growth (Fig 7F). These results establish that FoxO acts cell-autonomously in Lst⁺ EECs to mediate Lst induction in response to reduced systemic IIS activity. Furthermore, we analyzed Lst genomic locus sequence for potential FoxO response elements using the consensus binding sequences previously described by Spellberg and colleagues [38]. This analysis identified three putative FoxO-binding motifs within the 2 kb proximal promoter region of Lst, as well as an additional consensus motif located within the unique intronic region of the Lst gene (S7 Fig). While the presence of these motifs does not formally demonstrate direct binding, these sequence analyses are consistent with a model in which FoxO directly regulates Lst transcription in response to nutritional stress.

To further define the regulatory cascade, we manipulated upstream IIS components. RNAi-mediated knockdown of Pten in EECs (EEC-GAL4; Fig 7G and 7H) or Lst⁺ cells (Lst-GAL4; S6B and S6C Fig) blunted Lst induction and accelerated larval development in germ-free animals raised on poor diet confirming that enhanced IIS represses Lst expression. Conversely, reducing InR activity by expressing InR RNAi constructs (Figs 7I, 7J, S6D, and S6E) or a dominant-negative form of InR (InR^DN [39]; Figs 7K, 7L, S6F, and S6G) in EECs or in Lst⁺ cells robustly induced Lst expression and delayed developmental progression (Fig 7I and 7L). We next performed genetic epistasis experiments whereby blocking insulin signaling in the EECs—via either InR-RNAi or overexpression of a dominant-negative InR (InR^DN)—and Lst expression resulted in a growth phenotype that mirrors the loss of Lst per se (Fig 7M and 7N). These results demonstrate that Lst interference is genetically dominant over InR loss of function, hence indicate that Lst acts downstream of InR in this context. Taken together, these results demonstrate that Lst expression in Lst⁺ EECs is negatively controlled by IIS and positively by FoxO, suggesting a direct endocrine link between systemic nutrient sensing and Lst production in the gut.

Lst induction in EECs is controlled by an inter-organ nutrient sensing regulatory loop

The IIS-dependent control of Lst raised the question of how nutrient availability modulated IIS in Lst⁺ EECs. In Drosophila, the fat body serves as a central nutrient-sensing organ that communicates systemic nutritional status to IPCs in the brain. Circulating amino acid uptake via the transporter slimfast (slif) and activation of the nutrient-responsive TORC1 pathway in the fat body promote the secretion of dIlps, particularly dIlp2, by IPCs [17]. Based on this, we hypothesized that Lst repression in EECs is not simply a local response to nutrient availability but is rather a gut response to low dIlp2 release by IPCs under nutritional stress. To test this, we impaired fat body nutrient sensing by overexpressing Tsc1/2, or by RNAi-mediated knockdown of slif or Raptor (Raptor being a TORC1-specific component [40]) using Lpp-GAL4 [41]. Each manipulation, known to suppress fat body TORC1 signaling, further delayed development under malnutrition (S8A–S8C Fig) and led to modest induction of Lst in the midgut despite the presence of microbial AAs (Fig 8A–8C). These findings suggest that TORC1 activity in the fat body is required to repress Lst expression in EECs in response to systemic cues. Because dIlp2 secretion by IPCs is a direct output of fat body nutrient sensing [17], we next asked whether reduced dIlp2 production in IPCs is sufficient to trigger Lst expression. Pan-neuronal knockdown of dIlp2 using Elav-GAL4 [42] significantly elevated Lst expression in the midgut (Fig 8D) and delayed larval maturation into pupae (Fig 8E), phenocopying the effects of fat body TORC1 inhibition.

Together, these results define a systemic nutrient-sensing loop linking dietary and microbial amino acid availability to Lst expression. In this loop, circulating amino acid are sensed by TORC1 in the fat body to promote dIlp2 release from IPCs, which maintains high IIS activity and suppresses Lst production in EECs. When circulating amino acid levels drop, reduced fat body TORC1 activity decreases dIlp2 release by the IPCs, leading to FoxO activation in Lst ⁺ EECs and induction of Lst. This inter-organ circuit thus connects systemic amino acid sensing to gut endocrine output, coordinating developmental pace with environmental nutrient supply.

Lst is an enterokine involved in an inter-organ feed-forward regulatory loop impacting circulating levels of dIlp2 in response to amino acids of dietary and microbial origins

Our data so far position Lst downstream of an inter-organ nutrient-sensing axis that links dietary and microbial amino acid availability to insulin signaling activity in the gut. However, the molecular consequence of Lst production and release by EECs under amino acid restriction remains unresolved. Given its endocrine nature and its previously described role in modulating dIlps release by IPCs in adult flies [27], we hypothesized that Lst might act as an enterokine feeding back into the systemic nutrient-sensing circuit. Specifically, we asked whether Lst expression from EECs could influence the activity of IPCs in the brain, thereby modulating the production and release of dIlp2 in response to amino acids.

To test this, we examined dIlp2 expression at the transcript and protein levels in the central nervous system following targeted Lst knockdown in EECs. This led to a clear increase in dIlp2 transcript levels in the brain (Fig 8F) and a marked decrease in dIlp2 staining in IPCs (Fig 8G), suggesting that Lst from EECs negatively regulates dIlp2 expression in the brain and stimulates dIlp2 retention in IPCs. We next assessed circulating dIlp2 levels under similar conditions. Strikingly, repression of Lst, either in Lst⁺ cells (S8D Fig) or in EECs (Fig 8H) of malnourished germ-free animals, resulted in a significant increase in hemolymph dIlp2 concentration with levels approaching those observed in malnourished animals mono-associated with Lp (S8E Fig). Accordingly, the transcript levels of InR, a well-established (repressed) transcriptional readout of IIS showed a trend to a reduction in the carcass (Fig 8I) and significantly downregulated in the fat body (Fig 8J), upon Lst knockdown in EECs, further confirming that the knockdown of gut-derived Lst leads to elevated systemic insulin activity across peripheral tissues.

Taken together, these findings establish Lst as an amino acid-responsive enterokine that contributes to an inter-organ regulatory loop engaged under malnutrition. By modulating dIlp2 production and release from brain IPCs, Lst helps adjust developmental progression at the systemic level in response to fluctuations in dietary and microbial amino acid supplies.

Fig 9. Schematic model illustrating the gut–brain axis regulating larval adaptation to nutritional stress.

Compared with normal fed conditions, undernourished germ-free (GF) status or severe malnutrition associated with Lp^WJL induces Lst expression in a small population of midgut EECs. Once induced, gut-derived Lst acts systemically to suppress the circulating levels of dIlp2. This EEC-specific Lst up-regulation is triggered by reduced amino acid availability through a multi-organ relay involving the fat body and IPCs in the brain. Consequently, Lst orchestrates a feedback control loop that temporarily attenuates developmental progression upon sensing systemic nutrient shortages. This bidirectional gut–brain axis ultimately enables Drosophila larvae to preserve viability under severe nutritional stress. Created in BioRender. Leulier, F. (2026) https://BioRender.com/.

https://doi.org/10.1371/journal.pbio.3003911.g009

Fly stocks

All stocks are referenced in Flybase [45]. y,w (gift from Bruno Lemaitre), AKH-GAL4 (BDSC#25684), EEC-GAL4 (BDSC#84614), voilà-GAL4 (BDSC#84276), mex-GAL4 (gift from Irene Miguel-Aliaga), Lst-GAL4 (gift from Seung K Kim), Lpp-GAL4 (gift from Alexandre Djiane), Elav-GAL4 (BDSC#458), Lst¹ (gift from Seung K Kim), Lst^CRIMIC (BDSC#91330), dIlp2¹-gd2HF (gift from Pierre Léopold), UAS-GFP (cytoplasmic GFP; gift from Jonathan Enriquez), Lst RNAi^KK (VDRC#106861), Lst RNAi^TRiP (BDSC#60,400), Gcn2 RNAi^TRiP (BDSC#67215), FoxO RNAi^KK (VDRC#107786), slif RNAi^GD (VDRC#45587), Tsc1 RNAi^KK (VDRC#110811), InR RNAi^GD (VDRC#992), Raptor RNAi^KK (VDRC#106491), Pten RNAi^KK (VDRC#101475), dIlp2 RNAi^KK (VDRC#102158), PK1-R RNAi^KK (VDRC#101115), PK1-R RNAi^TRiP (BDSC#27539), PK1-R RNAi^SH (VDRC#330176), GD control (VDRC#60000), KK control 1 (VDRC#60102), KK control 2 (designated as KK control) (VDRC#60103), TRiP control 1 (BDSC #36304), TRiP control 2 (designated as TRiP control) (BDSC#35785), SH control (VDRC#60201), UAS-Lst (gift from Seung K Kim), UAS-FoxO (BL#9575), UAS-InR^DN (BL#8253), OE control (BL#35786). To eliminate potential genetic background confounding, all experimental lines (RNAi or overexpression) and their respective controls were crossed to the identical GAL4 driver line, ensuring a highly consistent genetic architecture across all conditions.

Fly husbandry

Conventional fly stocks were raised on a yeast/cornmeal medium (7.14 g/L of agar, 50 g/L of inactivated yeast, 80 g/L of cornmeal, 4 mL/L of propionic acid, and 5.2 g/L of Methyl 4-hydroxybenzoate sodium salt) at 25 °C under a 12:12-hour light-dark cycle. Germ-free flies were generated as previously described [46] and raised on a yeast/cornmeal diet supplemented with four antibiotics (50 µg/mL of ampicillin, 5 µg/mL of erythromycin, 50 µg/mL of kanamycin, and 10 µg/mL of tetracycline). For the experimental malnutrition condition, the concentration of inactivated yeast in the medium was reduced to either 7 or 3 g/L (malnutrition or severe malnutrition, respectively).

Holidic diets

HDs were prepared based on previously described FLYAA recipes [33], with minor modifications to the amino acid composition. Specifically, the HAAHD contained a total amino acid concentration of 21.38 g/L, while the LAAHD contained 8.55 g/L (representing a 60% reduction in the concentration of all AAs compared to the HAAHD). More specific HD compositions were generated by further manipulating the amino acid content of the initial LAAHD and HAAHD formulations. Specifically, the six essential amino acids produced by Lp (histidine, lysine, methionine, phenylalanine, threonine, and asparagine) were either reduced by 60% from the HAAHD level or increased by 2.5 times the LAAHD concentration. To prevent the inhibition of bacterial growth in gnotobiotic experiments, chemical preservatives such as propionic acid or nipagin were intentionally omitted from all HD formulations. To ensure sterile conditions for experiments, all tubes were UV-treated, egg-laying cages were autoclaved, and all diet solutions were either filter-sterilized or autoclaved as appropriate. Freshly prepared HDs were stored at 4 °C and used within one week of preparation.

Bacterial culture

Lp^WJL was cultured in MRS Broth (Carl Roth) overnight without shaking at 37 °C. Ap^WJL was cultured in 10 mL of Mannitol Broth—comprising 3 g/L Bacto peptone, 5 g/L yeast extract (both from Becton Dickinson), and 25 g/L D-mannitol (Carl Roth)—in a 50-mL flask at 30 °C with shaking at 180 rpm for 24 hour. After centrifugation, the supernatant was discarded, and the bacterial pellet was resuspended in sterile PBS for fly food inoculation at a final concentration of 10⁸ CFU/tube for Lp^WJL and 10⁵ CFU/tube for Ap^WJL.

Developmental timing experiments

To establish germ-free and mono-associated conditions, germ-free adult flies were allowed to lay eggs overnight in sterile breeding cages containing fly food medium. Axenic embryos were collected and transferred to individual fly food tubes. Each tube was then inoculated with 200 µL of either sterile PBS for the germ-free condition or 200 µL of bacteria resuspended in sterile PBS, for mono-association experiments. Unless otherwise specified, all fly lines were reared at a standard temperature of 25 °C under a 12:12-hour light-dark cycle. Larval development and growth rate were assessed by monitoring pupariation time. The number of pupae in each tube was counted daily until the emergence of the last pupa. The median time required for larvae to reach pupariation after egg laying was used as a quantitative measure of the larval growth rate. Each data point represents the median time to pupariation for one tube of pupae (40 ± 15 pupae).

Larval size measurements

For larval size measurements, Drosophila larvae were collected and washed with PBS. The larvae were then transferred to a beaker containing PBS and euthanized using a brief heat treatment (60 °C). Subsequently, the larvae were positioned for imaging and photographed using a Leica M205FA Stereomicroscope. Larval longitudinal size was quantitatively measured from the resulting images using the Fiji software (Version 1.54r).

Food intake experiments

Stage-(L3) and Size-(3.5 mm) matched larvae were collected and placed on the low-yeast diet plates supplemented with 0.8% Erioglaucine disodium salt (Sigma-Aldrich, #861146) for either 1 or 2 hours. Following the feeding period, larvae were washed with distilled PBS to remove residual surface dye. Seven larvae were then pooled in a 2 mL Eppendorf tube containing beads and 500 µL PBS. Larvae were homogenized using a Precellys 24 tissue homogenizer (Bertin Technologies) with two cycles of 30 s at 6,000 rpm, separated by a 30 s pause. Homogenates were then centrifuged for 60 s at 10,000g. The absorbance of the resulting supernatant was measured at 629 nm using a microplate reader (BMG LABTECH).

Quantitative real-time PCR

Total RNA was extracted from dissected tissues of stage- and size-matched larvae using the NucleoSpin RNA (MACHEREY-NAGEL, #740955) or NucleoSpin RNA XS (MACHEREY-NAGEL, #740902). RNA was reverse-transcribed into cDNA using the SensiFAST cDNA Synthesis Kit (Bioline, #BIO-65054). Quantitative real-time PCR (qRT-PCR) was performed using the No ROX SYBR 2× MasterMix Blue dTTP (Takyon, #UF-NSMT-B0701) on a CFX96 Real-Time System (BIO-RAD). All expression levels were normalized to two reference genes, RpL32 and Tubulin using delta-delta-Ct method. The following primers (5′–3′) were used for qPCR assays: Tubulin qF TGTCGCGTGTGAAACACTTC, Tubulin qR AGCAGGCGTTTCCAATCTG; Rpl-32 qF ATGCTAAGCTGTCGCACAAATG, Rpl-32 qR GTTCGATCCGTAACCGATGT; Lst qF ATAGTCTGCGATCCAAGCCG, Lst qR CTGCAGACGGTTCAGATCGT; InR qF AAGCGTGGGAAAATTAAGATGGA, InR qR GGCTGTCAACTGCTTCTACTG; dIlp2 qF CGAGGTGCTGAGTATGGTGTG, dIlp2 qR CCCCCAAGATAGCTCCCAGGA.

Hemolymph collection

A 20 μL filtered pipette tip and a 10 μL filtered pipette tip were each cut at an angle. The modified 20 μL tip was placed inside a 2 mL Eppendorf tube, and the modified 10 μL tip was then inserted into the wider opening of the 20 μL tip, creating a nested collection system. Stage- and Size-matched Drosophila larvae were washed with PBS, and excess surface moisture was carefully removed using a paper towel. Individual larvae were transferred to the inner wall of the 10 μL pipette tip, where they were gently torn open from the anterior end to release hemolymph. Hemolymph was collected into the 2 mL Eppendorf tube by centrifuging the assembly at 10,000g for 1 min at 4 °C. The collected hemolymph was immediately flash-frozen in liquid nitrogen and subsequently stored at −80 °C. To minimize protein degradation, all steps of the hemolymph collection procedure were performed on ice.

dIlp2-HF ELISA

ELISA was performed using Nunc-Immuno modules (ThermoFisher Scientific, #468667). Each well was initially coated with 100 μL of anti-FLAG antibody (Sigma-Aldrich, #F1804), diluted to a concentration of 5 μg/mL in 0.2 M sodium carbonate/bicarbonate buffer (pH 9.4). The modules were incubated overnight at 4 °C to allow for antibody binding. Following the coating step, the plate was washed three times with 0.2% PBST (PBS containing 0.2% Tween 20). Non-specific binding sites were then blocked by adding 350 μL of blocking buffer (PBS containing 2% BSA (ThermoFisher Scientific, #37525) to each well and incubating for 2 hours at room temperature. After blocking, the plate was washed three times with 0.2% PBST. Subsequently, 1 μL of either collected hemolymph sample or synthetic FLAG(GS)HA peptide standard was mixed with 50 μL of detection antibody solution. The detection antibody solution consisted of anti-HA-Peroxidase 3F10 antibody (Roche, #12013819001) at a concentration of 25 ng/mL in 0.1% PBSTx (PBS containing 0.1% Triton X-100). This mixture was then added to the appropriate wells. The wells were sealed, and the plate was incubated overnight at 4 °C. The solution was removed from the wells by aspiration, and the plate was washed thoroughly six times with 0.2% PBST. For color development, 100 μL of 1-Step Ultra TMB ELISA Substrate (ThermoFisher Scientific, #34028) was added to each well. The plate was incubated on a rotary shaker for 30 min at room temperature to facilitate the enzymatic color reaction. The reaction was terminated by adding 100 μL of 2 M sulfuric acid to each well. The absorbance at 450 nm was immediately measured using a microplate reader (BMG LABTECH).

Anti-Lst antibody generation

A rabbit polyclonal anti-Lst serum was produced by Pacific Immunology. A 15-amino acid peptide (Cys-AIVFRPLFVYKQQEI) was synthesized and used for rabbit immunization. Lst antibody was purified by affinity chromatography from a 25 mL serum pool collected from two rabbits.

Immunofluorescence

Fresh tissues from stage- and size-matched larvae were dissected in cold PBS and fixed in 4% paraformaldehyde (PFA) (16% PFA (ThermoFisher Scientific, #28906) diluted in PBS) for 60 min with gentle agitation. After fixation, the tissues were permeabilized by washing 3 times with 0.1% PBSTx and incubated with 3% goat serum (Sigma-Aldrich, #G9023) in 0.1% PBSTx at RT with shaking. The tissues were then incubated with primary antibodies overnight at 4 °C. After the primary antibody solution was removed, the tissues were washed 5 times with 0.1% PBSTx and further incubated with secondary antibodies diluted in 0.1% PBSTx for 2 hour at RT. After 5 washes with 0.1% PBSTx, the samples were mounted in ROTI Mount FluorCare DAPI (Carl Roth, #HP20.1). To reduce the strong background of Lst staining, the Lst antibody was pre-cleared using posterior larval midguts and re-collected for the experiments. Details of the antibodies used in this study are listed below: chicken anti-GFP (Abcam, #ab13970, 1:1,000), mouse anti-Prospero (MR1A) (DSHB, 1:10), rabbit anti-Lst (Custom-made, 1:50), rabbit anti-dIlp2 (gift from Jan A Veenstra, 1:500), rabbit anti-Phospho-Drosophila Akt (Ser505) (Cell Signaling Technology, #4054, 1:200), Alexa Fluor 488-conjugated goat anti-chicken (ThermoFisher Scientific, #A11039, 1:200), Alexa Fluor 555-conjugated goat anti-rabbit (ThermoFisher Scientific, #A21428, 1:200), and Alexa Fluor 647-conjugated goat anti-mouse (ThermoFisher Scientific, #A21235, 1:200).

Samples were imaged using a Leica SP8 confocal microscope. To compare Lst, dIlp2, and pAKT protein levels, all samples were imaged using identical acquisition settings. Image processing was performed using Fiji software, and 3D reconstructions were generated using the ImageJ 3D Viewer plugin (Version 1.54r).

Statistical analyses

All statistical analyses were conducted using GraphPad Prism 10 software (Version 10.6.1). Detailed statistical methods are indicated in the figure legends. Normality and lognormality tests were conducted. Statistical significance was analyzed using t tests, one-way ANOVA, and two-way ANOVA (employing the Geisser–Greenhouse correction). Detailed statistical methods are indicated in the figure legends.

Declaration of generative AI and AI-assisted technologies in the writing process

During the preparation of this work, the authors used ChatGPT in order to polish the language of the manuscript. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.

S1 Fig. Related to Fig 1. Knockdown of Lst accelerates larval systemic growth specifically under malnutrition.

(A–D) Median time to pupariation measured for Lst-GAL4 mediated Lst RNAi larvae and their respective controls. Assessments were performed under malnutrition (A, B) or normal fed conditions (C, D), utilizing either the KK RNAi library (A, C) or the TRiP RNAi library (B, D) (n ≥ 9). (E, F) Food intake quantification for undernourished germ-free larvae with Lst-GAL4 driven Lst knockdown (Lst > UAS-Lst RNAi) and their corresponding KK control larvae after 1 hour (E) and 2 hours (F) of feeding (n = 8). Statistical significance was determined using a two-way ANOVA with Tukey’s multiple comparisons test (A–D) or a one-way ANOVA with Tukey’s multiple comparisons test (E, F). Exact P values are indicated on the respective panels. The data underlying the graphs can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003911.s002

(EPS)

S2 Fig. Related to Fig 2. Lst expression patterns in Drosophila larvae.

(A, B) Spatial distribution of Lst expression revealed by the Lst-GAL4 > GFP reporter line under germ-free and malnourished conditions. (A) The reporter line driving GFP signals in corpora cardiaca (CC); the ring gland is outlined by dashed lines, and the corpora cardiaca is localized at the base of the ring gland structure. Anti-HRP (horseradish peroxidase) staining was utilized to label the secretory neurons. (B) GFP-positive EECs localized in the midgut region anterior to the acidic zone.

https://doi.org/10.1371/journal.pbio.3003911.s003

(EPS)

S3 Fig. Related to Fig 2. Validation of Lst antibody specificity and genetic tools.

(A–C) Midgut anti-Lst immunostaining in germ-free (GF) malnourished larvae. (A) The Lst peptide was visualized in the midguts of yw control larvae and Lst^−/− null mutant larvae. (B, C) Validation of Lst RNAi lines via Lst immunostaining of Lst-GAL4-mediated Lst RNAi larvae, utilizing either the KK RNAi library (B) or the TRiP RNAi library (C). Scale bar: 50 μm.

https://doi.org/10.1371/journal.pbio.3003911.s004

(EPS)

S4 Fig. Related to Fig 4. Tissue-specific Lst knockdown reveals tissue-dependent roles in larval developmental timing.

(A, B) Median time to pupariation was measured under normal fed conditions in larvae with Lst knockdown specifically in EECs (EEC > UAS-Lst RNAi), compared to their respective controls (n ≥ 9). Data are shown for the KK RNAi line (A) and the TRiP RNAi line (B). (C) Median time to pupariation was measured under malnutrition in larvae with EEC-specific Lst knockdown driven by the voilà-GAL4 (voilà>UAS-Lst RNAi), compared to KK control larvae. (n = 10). (D, E) Median time to pupariation was measured under normal fed conditions in larvae with Lst knockdown in the corpora cardiaca (AKH > UAS-Lst RNAi), compared to their respective controls (n ≥ 7). Data are shown for the KK RNAi line (D) and the TRiP RNAi line (E). (F, G) Food intake was quantified over 1 hour (F) and 2 hours (G) in undernourished germ-free larvae with EEC-specific Lst knockdown (EEC > UAS-Lst RNAi), compared to KK control larvae. (n = 8). Statistical significance was determined using a two-way ANOVA with Tukey’s multiple comparisons test (A–E) or a one-way ANOVA with Tukey’s multiple comparisons test (F, G). Exact P values are indicated on the respective panels. The data underlying the graphs can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003911.s005

(EPS)

S5 Fig. Related to Fig 5. The expression of corpora cardiaca-derived Lst is insensitive to dietary AAs at the larval stage.

(A, H) Relative transcript levels of corpora cardica-derived Lst (A) or gut-derived Lst (H) analyzed by qRT-PCR in GF, Lp-associated (A), or Ap-associated (H) larvae reared under HAAHD or LAAHD conditions (n ≥ 5). (B, C) Median time to pupariation was measured under in larvae reared under HAAHD conditions with EEC-specific Lst knockdown (EEC > UAS-Lst RNAi), compared to their respective controls (n ≥ 9). Data are shown for the KK RNAi line (B) and the TRIP RNAi line (C). (D–G) Median time to pupariation was measured in larvae with corpora cardiaca-specific Lst knockdown (AKH > UAS-Lst RNAi), compared to their respective controls (n ≥ 7), reared either under LAAHD conditions (D, E) or under HAAHD conditions (F, G). Data are shown for the KK RNAi line (D, F) and the TRiP RNAi line (E, G). In (A, H), data are presented as mean ± SEM. Statistical significance was determined using a two-way ANOVA with Tukey’s multiple comparisons test (A–H). Exact P values are indicated on the respective panels. The data underlying the graphs can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003911.s006

(EPS)

S6 Fig. Related to Fig 7. Lst induction in EECs is controlled by reduced systemic insulin signaling via FoxO.

(A) Under malnutrition, relative transcript levels of gut-derived Lst analyzed by qRT-PCR in germ free larvae with FoxO knockdown in Lst-expressing cells (Lst > UAS-FoxO RNAi), compared to their respective KK controls (n = 6). (B, C) Relative transcript levels of gut-derived Lst analyzed by qRT-PCR (B) and median time to pupariation measured (C) under malnutrition for germ-free larvae with Pten knockdown in Lst-expressing cells (Lst > UAS-Pten RNAi), compared to their respective KK controls. n = 6 (B), n = 12 (C). (D–G) Under malnutrition, relative transcript levels of gut-derived Lst analyzed by qRT-PCR (D, F) and median time to pupariation measured (E, G) for Lp-associated larvae with InR knockdown in Lst-expressing cells (Lst > UAS-InR RNAi) and their respective GD controls (D, E), or with specific overexpression of dominant-negative InR in Lst-expressing cells (Lst > UAS-InR^DN) and their corresponding overexpression controls (F, G). n = 6 (D), n ≥ 13 (E, G), n = 5 (F). In (A, B, D, F), data are presented as mean ± SEM. Statistical significance was determined using a two-tailed unpaired t test (A–C, E, F) and two-tailed Mann-Whitney U test (D, G). Exact P values are indicated on the respective panels. The data underlying the graphs can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003911.s007

(EPS)

S7 Fig. Related to Fig 7. Identification of putative FoxO-binding motifs within the Lst promoter region.

(A) Schematic representation of putative FoxO-binding motifs within the 2 kb proximal promoter region of Lst, indicated by yellow and green boxes. For the sequence nomenclature: bold letters denote the Lst genomic sequence; uppercase letters represent exons; lowercase letters represent introns; pink text indicates the coding sequence (CDS); and lowercase letters in pink signify alternative splicing events.

https://doi.org/10.1371/journal.pbio.3003911.s008

(EPS)

S8 Fig. Related to Fig 8. Systemic TORC1 and Lst peptide coordinately regulate developmental timing and circulating insulin-like peptides.

(A–C) Under malnutrition, the median time to pupariation was measured in Lp-associated larvae with fat body-specific genetic manipulations, including: Overexpression of Tsc1/2 (Lpp > UAS-Tsc1/2) compared to overexpression controls (A); knockdown of slif (Lpp > UAS-slif RNAi) compared to GD controls (B); or knockdown of Raptor (Lpp > UAS-Raptor RNAi) compared to KK controls (C). n ≥ 9. (D) Under malnutrition, circulating dIlp2(HF) in the hemolymph of germ-free larvae with Lst knockdown in Lst-expressing cells (Lst > UAS-Lst RNAi) and their respective KK controls were quantified by ELISA. n = 12. (E) The effect of Lp-association was assessed by measuring circulating dIlp2(HF) levels in the hemolymph of undernourished heterozygous Ilp2¹ gd2HF larvae. n = 16. (F–H) Larval sizes were measured at Day 6 AEL in undernourished germ-free larvae reared with Pk1-R knockdown in IPCs (dIlp2 > UAS-Pk1-R RNAi), compared to their respective controls. Results are shown for the KK RNAi line (F), the SH RNAi line (G), and the TRiP RNAi line (H). n ≥ 67. Statistical significance was determined using a two-tailed unpaired t test (A, C, H), a two-tailed Welch’s t test (B), a one-way ANOVA with Tukey’s multiple comparisons test (D), a Lognormal t test (E), or a two-tailed Mann-Whitney U test (F, G). Exact P values are indicated on the respective panels. The data underlying the graphs can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003911.s009

(EPS)