Multi-omics reveals an ecdysone-activated Eip75B-FABP signaling axis coordinating nutrient metabolism for development in Hermetia illucens

Graphical abstract

Highlights

• •Time-resolved multi-omics elucidate the developmental characteristics of insects.
• •Eip75B transcriptionally activates FABP to maintain lipid homeostasis.
• •PPARγ agonist increases insect lipid storage by 14%, whereas antagonist decreases it by 18 %.
• •FABP deficiency reprograms metabolism toward oxidation and synthesis.
• •Ecdysone-Eip75B-FABP axis integrates hormones and nutrients for development.

Abstract

Introduction

Efficient nutrient storage is essential for insect development and energy homeostasis; however, the mechanisms coordinating nutrient allocation during ontogeny are not well understood. Elucidating these systems may yield valuable insights to insect metabolic adaptation.

Objectives

This study aimed to identify regulatory modules governing nutrient metabolism in insects, focusing on hormonal and metabolic interplay.

Methods

Multi-omics profiling (proteomics, phosphoproteomics, and transcriptomics) was conducted throughout the life cycle, from egg to adult, to identify metabolic regulators. RNAi was utilized for gene knockdown, followed by qRT-PCR and mitochondrial DNA quantification to evaluate knockdown efficiency and its metabolic implications. Assessments of nutrient metabolism were performed using assays for triglycerides, crude protein, and fatty acid synthase. EMSA and BODIPY staining examined transcriptional regulation and lipid droplet dynamics.

Results

Utilizing an integrative multi-omics approach, this study elucidates the temporal metabolic regulators in insects. A conserved regulatory module was identified in which the PPAR homolog, ecdysone-induced protein 75B (Eip75B), functions as a transcriptional activator of fatty acid binding protein (FABP), sustaining lipid metabolic homeostasis during the larval stage. PPARγ modulators (rosiglitazone and GW9662) alter lipid accumulation, along with the expression of Eip75B and FABP, which was measured by qRT-PCR. Furthermore, the deficiency of FABP may reprogram metabolic pathways by inhibiting lipid storage and promoting mitochondrial β-oxidation, as supported by increased mitochondrial DNA copy number, as well as enhancing protein synthesis. This metabolic change could be modulated by ecdysone signaling, as hormonal supplementation effectively rescued the lipid loss phenotype. Our results establish the ecdysone-Eip75B-FABP signaling axis as a central regulatory module that integrates hormonal and nutrient-sensing signals to control insect nutritional metabolism.

Conclusion

The ecdysone-Eip75B-FABP axis integrates hormonal and nutrient signals to regulate metabolic plasticity, underscoring a universal strategy for developmental energy allocation. The data also offer potential implications for research on metabolic disorders and bioenergy applications.

Introduction

Carbohydrates, proteins, and lipids constitute the three major energy sources for living organisms [1]. Insects accumulate these macronutrients during the feeding stage to facilitate subsequent metamorphosis, with lipids playing particularly crucial roles not only in developmental processes but also in enabling flight, migration, and diapause [2]. Furthermore, lipids are fundamental components of biological membranes, where they perform essential functions such as cellular compartmentalization and segregation, signal transduction, as well as membrane fission and fusion [3]. The fat body (FB) serves as an organ characterized by significant biosynthetic and metabolic activity [4]. Insects store energy reserves primarily in the form of triglycerides within adipocytes, which are the main cell type in the fat body [5].

The life cycle of holometabolous insects encompasses four distinct stages: egg, larva, pupa and adult. These transitions are based on tightly regulated gene expression at the transcriptional, epigenetic, and translational levels [6]. However, transcript levels serve as only a moderate predictor of protein expression, as they do not account for post-transcriptional processes, including translational regulation and protein stability [7]. This limitation is illustrated by the non-monotonic relationship observed between mRNA levels and their corresponding protein concentrations [8]. In a study conducted by Kolja Becker et al. (2018), embryonic tissues of Drosophila were analyzed at 14 time points post-egg laying, revealing that the dynamic expression of mRNAs did not directly correlate with their respective proteins [9]. Beyond regulating most biological processes at the protein level, post-translational modifications (PTMs) also play a crucial role in modulating metabolic and other cellular activities [10]. Protein phosphorylation, a reversible PTM, represents one of the most prevalent mechanisms for regulating protein activity and transmitting signals within cell biology [[11], [12], [13]]. Notably, the activation of Drosophila eggs entails not only proteome remodeling but also significant phosphorylation changes in hundreds of proteins [14]. Additionally, it has been observed that 34 % of yeast lipid metabolic enzymes undergo phosphorylation changes throughout their life cycle [15].

This study explores the mechanisms of nutritional metabolism in Hermetia illucens (black soldier fly, BSF), a holometabolous insect of considerable economic significance. BSF larvae (BSFL) metabolically adapt to convert organic waste into biomass rich in proteins and lipids [16]. The lipid fraction, in particular, serves as a promising biodiesel precursor, exhibiting favorable fuel properties such as high energy content, elevated cetane number, and strong compatibility with conventional diesel fuels [17,18]. However, a comprehensive understanding of the nutritional complexity of this insect is essential for realizing its diverse applications across various industries [19]. Recent studies on functional genes in other insect systems further illustrate both the evolutionary conservation of regulatory modules and the translational potential of targeting metabolic and physiological pathways. For example, Zhang et al. (2025) performed a functional characterization of the serine protease EfSP1 in the predatory bug Eocanthecona furcellata, showing that perturbation of a single, conserved gene can profoundly affect development, digestion and predation performance, thereby linking gene function to organismal physiology and ecological function [20]. Likewise, Li et al. (2025) identified a fungal ribonuclease (CeRNase T2) from Cunninghamella echinulata that disrupts detoxification pathways and causes behavioral inhibition and mortality in the invasive fire ant Solenopsis invicta, exemplifying how interventions targeting metabolic/detoxification networks can yield practical biocontrol outcomes [21]. Strengthening the link between gene function and application in BSF depends on establishing reliable genetic and genomic resources. In previous research, a high-quality 1.1 Gb genome was assembled and 16,770 gene models were annotated, establishing a genomic foundation for black soldier fly studies [22], followed by transcriptomic analysis that facilitated innovative genetic engineering strategies for insect biodiesel production [23]. In this work, we integrated a comprehensive multi-omics analysis throughout the entire BSF life cycle, quantifying 5527 proteins, 16,620 phosphorylation sites, and 8656 transcripts. We identified fatty acid-binding protein (FABP), a key molecular chaperone [24], as a central node in nutrient partitioning, whose deficiency induces a metabolic reprogramming towards mitochondrial β-oxidation and protein synthesis. The evolutionary conservation of PPAR-mediated lipid metabolic regulatory networks across species motivates us to investigate whether ecdysone-induced protein 75B (Eip75B), the insect ortholog of PPAR, modulates similar pathways of lipid homeostasis in Hermetia illucens (H. illucens). In vivo treatment with GW9662 (a PPARγ activator) significantly increased lipid deposition and upregulated Eip75B expression, while rosiglitazone (a PPARγ inhibitor) elicited opposing effects. Furthermore, we identified Eip75B as the upstream regulator, while EMSA assays confirmed its direct binding to the FABP promoter. Notably, supplementation with 20-hydroxyecdysone (20E) ameliorated lipid deficits induced by FABP deficiency and stabilized FABP expression. Our findings elucidate dynamic, evolutionarily conserved mechanisms of nutrient sensing in H. illucens and establish a framework for engineering this species to enhance its utility in bioconversion and circular bioeconomy applications.

Materials and methods

Insect rearing

The colony of black soldier fly (Hermetia illucens) was maintained in a greenhouse at the National Engineering Research Center of Microbial Pesticides, Huazhong Agricultural University (HZAU), Wuhan, Hubei, China (30.47°N, 114.35°E). The strain has been continuously reared under laboratory conditions for approximately 14 years, with about 125 generations. The rearing conditions for BSF were established at a temperature of 28 ± 2 ℃, with relative humidity maintained at 70 % ± 10 %, and a photoperiod of 14 L: 10 D light. Larvae were fed a diet composed of wheat bran and wheat middlings mixed at a 1:1 (w/w) ratio, moistened with water to achieve an appropriate texture. The ratio of larval number to substrate wet weight was maintained at approximately 1:1 (e.g., 100 larvae per 100 g of moistened substrate, including wheat bran, wheat middlings, and water).

Protein extraction in insects

The sample was ground using liquid nitrogen to obtain a cell powder, which was subsequently transferred to a 5-mL centrifuge tube. Four volumes of lysis buffer were then added to the cell powder, followed by sonication performed three times on ice using a high-intensity ultrasonic processor. The residual debris was eliminated through centrifugation at 12,000 g at 4 °C for 10 min. Finally, the supernatant was collected, and the protein concentration was quantified using a BCA Protein Assay Kit (Beyotime, Cat. No. PP102-01, China).

Sampling method of egg, larvae, prepupa, pupa, and adults

BSF eggs were collected over a three-hour period. Six sub-samples of the collected eggs were taken at 24, 48, and 72 h after egg deposition. Following the 72-h, the eggs were monitored for hatching. The appearance of the first neonate (1st instar) initiated the timing for larval age. Six replicates of the first neonates within the initial 24 h post-hatch were placed in 2 mL centrifuge tubes and immediately snap-frozen in liquid nitrogen. The 1-old larvae were then transferred to incubators for monitoring. At 3, 7, 11, and 15 d after hatching, six larvae (i.e., replicates) were collected. Observations of prepupae, which can be identified by their black integument, were recorded at 0 h, marking the first observed prepupae. Six replicates of prepupae were taken and immediately snap-frozen in liquid nitrogen. The remaining prepupae were transferred to the incubator for monitoring during their transition to the pupal stage. Once pupae were observed, six replicates were collected and immediately snap-frozen in liquid nitrogen. The remaining pupae were returned to the incubator for monitoring of adult emergence. Similar to other developmental stages, six adults (<24-h-old) were collected (three male and three female). A small amount of pure water was sprayed into the cages using a spray bottle to maintain the normal activity of the 1-day adults after sampling, and the culture was continued for three days.

Mass spectrometry measurement and label-free analysis

The tryptic peptides were dissolved in solvent A (0.1 % formic acid, 2 % acetonitrile/in water), and subsequently loaded onto a home-made reversed-phase analytical column (25-cm in length, 75/100 μm i.d.). The peptides were separated using a gradient from 6 % to 24 % solvent B (0.1 % formic acid in acetonitrile) over a duration of 70 min, followed by an increase from 24 % to 35 % over 14 min and finally ramping up to 80 % in 3 min, with a hold at 80 % for the final 3 min. This process was conducted at a constant flow rate of 450 nL/min on a nanoElute UHPLC system (Bruker Daltonics, Bremen, Germany). The peptides were then subjected to a capillary source before entering the timsTOF Pro mass spectrometer (Bruker Daltonics, Bremen, Germany). The electrospray voltage applied during the analysis was 1.60 kV. Precursors and fragments were analyzed using the TOF detector, with a MS/MS scan range from 100 to 1700 m/z. The timsTOF Pro was operated in parallel accumulation serial fragmentation (PASEF) mode, whereby precursors with charge states ranging from 0 to 5 were selected for fragmentation, and 10 PASEF-MS/MS scans were acquired per cycle. The dynamic exclusion was set to 30 s. The database utilized was Hermetia_illucens_343691_NCBI_20220719_seqkit.fasta (20722 sequences), with a negative database included to calculate the false discovery rate (FDR) attributed to random matches.

Weighted gene co-expression network analysis (WGCNA)

Weighted gene co-expression network analysis (WGCNA) was performed using the WGCNA R package (Version 1.72-1) [25] to identify protein modules with highly correlated expression patterns. Initially, all proteins containing missing values (NAs) were removed, yielding a total of 4,892 quantified proteins. Missing values were imputed using the K-nearest neighbor (KNN) method, followed by Z-score normalization across samples. The resulting topological overlap matrix (TOM) was used for hierarchical clustering, and protein modules were identified using the dynamic tree-cutting algorithm with the parameter minModuleSize = 300, while other parameters were kept at default settings. Modules were assigned unique colors for visualization, and the gray module represented proteins that could not be classified into any specific module. Module–trait relationships were further examined by calculating the Pearson correlation coefficients between module eigengenes and experimental traits.

Integrative gene set enrichment analysis (GSEA) for transcriptome and proteome data

Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway annotations were used for functional analysis. The correlation coefficients between the quantified proteins and corresponding transcripts were analyzed using GSEA (Version 4.0.3, https://software.broadinstitute.org/gsea/) to identify the major KEGG pathways underlying distinct regulatory relationships. Pathways with a nominal p-value < 0.05 were considered significantly enriched and selected for functional mapping.

Protein-protein interaction network

All differentially expressed protein database accession numbers or sequences were searched against the STRING database (Version 11.0, https://string-db.org/) to identify protein–protein interactions. Only interactions involving proteins within the searched dataset were selected, thereby excluding external candidates. STRING defines a metric known as the “confidence score” to quantify interaction confidence; interactions with a confidence score of ≥0.7 (high confidence) were retrieved. The interaction network form STRING was visualized using the R package “networkD3”.

Prediction of kinase-substrate relationships

Kinase–substrate relationship prediction was performed using GPS 5.0 software (Version 5.0, http://gps.biocuckoo.cn/) [26]. This software is grounded in the theory of short linear motifs (SLMs) surrounding phosphorylation sites (p-sites), which provide the primary specificity (table S1). For each phosphorylation site, GPS 5.0 predicted the potential upstream kinase families under a medium threshold setting. Subsequently, the predicted kinase families were aligned to the kinase sequences in the iEKPD 2.0 database (Version 2.0, http://iekpd.biocuckoo.org/) [27] using BLAST (Version 2.2.28, https://blast.ncbi.nlm.nih.gov/Blast.cgi) to identify the corresponding kinase proteins. To minimize false-positive interactions, the resulting kinase–substrate pairs were further filtered by protein–protein interaction (PPI) evidence.

Prediction of kinase activities

GSEA (Version 4.0.3, http://software.broadinstitute.org/gsea/) was applied to estimate kinase activities [28]. Log-transformed phosphorylation ratios were used as the ranking metric, while kinase–substrate relationships were formatted into a GMT file. The normalized enrichment scores (NES) derived from the enrichment results were interpreted as kinase activity scores. For each kinase, a positive prediction was made if the predominant change in substrates was characterized by an increase in phosphorylation, whereas a negative prediction was assigned in cases of predominant dephosphorylation.

Database search

The resultant MS/MS data were processed utilizing the MaxQuant software (Version 1.6.15.0, Max Planck Institute of Biochemistry, Martinsried, Germany). Tandem mass spectra were aligned against the human SwissProt database, consisting of 20,422 entries, in conjugation with a reverse decoy database. Trypsin/P was designated as the cleavage enzyme, permitting up to 2 missed cleavages. The mass tolerance for precursor ions was established at 20 ppm during the initial search and 5 ppm during the main search, while the mass tolerance for fragment ions was set at 0.02 Da. Carbamidomethylation of cysteine was specified as a fixed modification, whereas acetylation of protein N-terminal and oxidation on methionine were designated as variable modifications. FDR was adjusted to <1 %.

RNA interference

The mRNA sequences of L15605, Eip75B, and CPT1 were retrieved from the National Center for Biotechnology Information (NCBI), and a specific primer incorporating the T7 promoter (TAATACGACTCACTATAGGG) at the 5′ end was designed within the CDS region (table S2). The DNA fragments corresponding to L15605 and Eip75B were amplified via PCR using cDNA derived from the BSF. Double-stranded RNA (dsRNA) was synthesized in vitro using the T7 RNAi Transcription Kit (Vazyme, Cat. No. TR102, China), following the manufacturer’s instructions. To enhance RNA interference efficiency, dsRNA was encapsulated within chitosan-based nanoparticles prepared according to a modified ionic gelation method [29,30]. Briefly, a chitosan-sodium acetate-acetic acid (CS-NaAC-HAC) solution (pH adjusted to 4.5) was prepared by dissolving 0.02 g of chitosan in 50 mL of 0.1 M acetic acid and mixing with 30 mL of 0.1 M sodium acetate solution. Separately, 50 μg of dsRNA was dissolved in 100 μL of 0.1 M sodium sulfate solution, followed by the addition of 100 μL of the CS–NaAC–HAC solution. The mixture was heated at 55 °C for 1 min, vortexed for 30 s, and immediately cooled to room temperature to form dsRNA–chitosan nanoparticles. The resulting nanoparticle complexes were thoroughly mixed into the basal artificial diet and fed to healthy 5-day-old BSF larvae for 4/6 days. The feeding experiments were conducted at 28 ℃ with 75 ± 5 % relative humidity. A control group was provided with feed containing eGFP dsRNA nanoparticles. Each treatment was performed in triplicate, with five larvae sampled per replicate. Total RNA was extracted using the RNAeasy kit (Beyotime, Cat. No. R0026, China) to assess gene knockdown efficiency and the relative expression levels of target genes by qRT-PCR.

qRT-PCR and mtDNA measurement

cDNA was synthesized from 1 μg of total RNA using the ToloScript All-in-one RT EasyMix for qPCR kit (TOLOBIO, Cat. No. 22107, China). qPCR was conducted on a QuantStudio 3 Real-Time PCR System (Applied Biosystems, Thermo Fisher Scientific, USA) using SYBR Green Master Mix (Vazyme, Cat. No. Q711-02, China). Relative fold differences in the expression levels of target genes were calculated as ratios to the mean of the reference genes RP49, using the ΔΔC_t method. Primer sequences are provided in table S2. For mtDNA measurement, DNA was extracted using the kit (TIANGEN, Cat. No. DP304-02, China). The copy numbers of both mitochondrial and genomic DNA were determined through qPCR. mtDNA was analyzed by amplifying the mitochondrial genes mt:CoI and mt:ND5. Nuclear DNA (nDNA) was assessed by amplifying the genomic genes GAPDH and RP49.

Lipid droplet staining and imaging

Dissect the larvae of the BSF to obtain fresh fat body tissues. Subsequently, rinse the tissues gently with 1 × PBS three times, each rinse lasting five min, and fix them in 4 % paraformaldehyde for 30 min. Following fixation, perform three additional gentle rinses with 1 × PBS (5 min per rinse) before transferring the tissues to a 1 mg/mL BODIPY staining solution. Incubate the samples in the dark for 10 min. After incubation, rinse the adipose tissue gently with 1 × PBS three times (5 min per rinse), and transfer the tissue into a DAPI staining solution at a working concentration of 2 μg/mL. Incubate in the dark for another 10 min. The tissue should then undergo three gentle rinses with 1 × PBS (5 min/time). Place the tissue on a slide under dark conditions, add 10 μL of antifade mounting medium, and seal the slide with nail polish. The slide can then be observed using an Olympus FV1000. In this analysis, the blue signal indicates the presence of nuclei, while the green signal denotes lipid droplets. the laser wavelengths utilized are 405 and 561 nm. It is noteworthy that lipids are mainly distributed in the form of lipid droplets within fat bodies, with the lipid content exhibiting a positive correlation with both the diameter of lipid droplets and fluorescence intensity [31].

Electrophoretic mobility shift assay (EMSA)

Initially, three biotin-labeled DNA probes corresponding to predicted Eip75B binding motifs within the L15605 promoter region (−2000 bp upstream of the start codon) were synthesized. These probes correspond to positions −833 to −783 bp (Probe 1), −1207 to −1157 bp (Probe 2), and −1662 to −1612 bp relative to the L15605 translation start codon (ATG), as predicted by the Berkeley Drosophila Genome Project (BDGP) transcription factor binding site database. The binding reaction was conducted by incubating purified Eip75B protein in a buffer composed of 10 mM Tris-HCl (pH 7.5), 50 mM KCl, 1 mM DTT, 5 % glycerol, and 0.5 mM EDTA. The reactions were maintained at room temperature for a duration of 30 min, with appropriate controls that included the omission of Eip75B. Subsequently, samples were resolved on a 6 % native polyacrylamide gel in 0.5 × TBE buffer at 100 V for 1.5 h. Following electrophoresis, DNA was transferred to a positively charged nylon membrane via wet transfer at 380 mA for 1 h, crosslinked using UV light, and probed with a streptavidin-HRP conjugate.

Treatment with GW9662, rosiglitazone, and 20-hydroxyecdysone

In the GW9662 and rosiglitazone treatment groups, a total of 100 six-day-old larvae were randomly selected and subjected to a 12 h starvation period prior to the initiation of the experiment. Subsequently, these larvae were provided with an artificial diet supplemented with either rosiglitazone (0.2 mg), GW9662 (0.5 mg), or an equivalent volume of DMSO as a control. Freshly prepared medicated feed was administered daily for six consecutive days. The selected doses were determined through preliminary dose–response assays combined with literature-based estimation. Specifically, three concentration gradients (20, 50, and 100 mg·kg⁻¹ diet) were initially tested for each compound to evaluate larval survival, feeding behavior, triglyceride content, and expression of lipid-metabolism marker genes. Based on these pilot results, the final working doses were chosen as they produced consistent metabolic responses without excessive vehicle-related mortality. This design followed established methodologies for dietary administration of small molecules in insects [[32], [33], [34], [35]]. For the treatment involving 20-hydroxyecdysone (20E), an additional cohort of 50 larvae was randomly selected and also underwent a 24-hour starvation period. These larvae were then fed an artificial diet containing varying concentrations of 20E, with the hormone being replenished daily over the same six-day period. The design of this experiment was informed by previous studies in Drosophila demonstrating dietary 20E supplementation to modulate steroid signaling [36].

Statistical analyses

Analysis and plotting of data were performed using GraphPad Prism version 8.3.1 (GraphPad Software) and expressed as mean ± SD. The statistical significance of differences among two groups was assessed by two-tailed Student’s t-test. The results were considered significantly different when *P < 0.05, **P < 0.01, and ***P < 0.001, with ns denoting not significant (P > 0.05).

Results

Temporal profiling of the proteome and phosphoproteome across different life stages of BSF

We conducted proteomic and phosphoproteomic analyses of BSF across five developmental stages (Fig. 1A). The sampling time points encompassed the egg stage (12, 24, 48, and 72 h), larval stage (1, 4, 8, and 12 days), prepupal stage (1 and 4 days), pupal stage (1, 3, and 6 days), and adult stage (1 and 4 days). All samples were collected in biological triplicates and analyzed using 4D label-free quantitation (LFQ) technology. The majority of identified peptides ranged from 7 to 20 amino acids, adhering to stringent quality control criteria (Fig. S1A). By leveraging MS (mass spectrometer)-based phosphoproteomics and proteomics workflows, we quantified a total of 5527 proteins and 16,620 phosphorylation sites (Figs. 1B and C, S1B). Furthermore, Pearson's correlation coefficient (PCC) analysis confirmed the high reproducibility of the data (Figs. 1C and S1C). Principal component analysis (PCA) of relative protein abundance demonstrated distinct time-dependent clustering, indicating gradual compositional changes from the egg to adult stages (Fig. 1D). At each time point, over 4600 proteins were expressed, with a peak of 5157 proteins detected at the 4-day larval stage (Fig. 1E). To further elucidate protein functions, we predicted subcellular localization, revealing a predominance of cytoplasmic proteins and a nuclear enrichment among phosphorylated proteins (Fig. 1F).

Fig. 1.

Temporal proteome, phosphoproteome, and transcriptome profiling of Hermetia illucens. (A) Schematic representation illustrating the temporal collection points across the four major metamorphic stages of H. illucens. (B) Summary of the identification processes for proteins and phosphoproteins. Protein identification required at least one unique peptide, with a false discovery rate (FDR) of <1 %. (C) Heat map displaying log₂ LFQ expression values for 5527 developmentally regulated protein groups, n = 3. (D) Results of quantitative principal component analysis of concerning proteins and phosphoproteins, n = 3. (E) Statistical analysis of the total number of proteins and phosphorylated proteins across five developmental stages. (F) Identification of subcellular localization characteristics within the proteome and phosphoproteome. (G) PTR distribution of genes for selected developmental stages, with genes categorized as having high or low PTRs defined as those falling two standard deviations outside the median PTR distribution. Representative images are shown from three independent experiments.

We delineated a core proteome by systematically categorizing proteins that are consistently expressed throughout all developmental stages, identifying a total of 4677 proteins and 3122 phosphorylated proteins present across the entire life cycle (Fig. S1D). KEGG enrichment analysis indicated that these core proteins were significantly associated with ribosomal pathways, whereas the core phosphoproteome exhibited enrichment in proteasome and AMPK signaling pathways (Fig. S1E). Notably, pathways related to fatty acid degradation and amino acid metabolism were enriched in both the proteome and phosphoproteome. The stage-specific protein analysis revealed 1860 differentially regulated protein groups, with the egg stage exhibiting the highest number of unique proteins (Fig. S1F).

Multi-omics integration reveals coordinated regulation of gene and protein expression

Gene expression in organisms is regulated at multiple levels; therefore we conducted a comprehensive analysis of the proteome alongside previously published transcriptome data [23]. A total of 5527 proteins were mapped to their corresponding transcript IDs, with 5335 genes quantified across both transcriptomic and proteomic levels (Fig. S2A). The overall correlation between RNA and protein levels across five developmental stages was modest, with an average Pearson correlation coefficient of 0.252 (Fig. S2B). Notably, the transcript–protein correlation demonstrated dynamic, stage-dependent variations (Fig. S2C). We also calculated protein-to-mRNA ratios (PTRs), which likely reflect differences in translation efficiency, transcript stability, and/or protein stability [37]. Significant variations in PTRs were observed between the larval and adult stages, suggesting stage-specific regulation at either the transcriptional or translational level (Fig. 1G). KEGG pathway enrichment analysis of genes exhibiting high PTRs consistently revealed enrichment of the PPAR signaling pathway across multiple stages (Fig. S2D). Gene set enrichment analysis (GSEA) further elucidated the biological processes associated with the identified proteins and transcripts (Fig. S2E). The significant enrichment of the retrograde endocannabinoid signaling pathway, suggesting a coordinated up-regulation, provides further insight into the lipid metabolism of BSF, given the role of endocannabinoids as important lipogenic neuroactive substances [38].

Identification of stage-specific protein modules through clustering and co-expression mapping

Although the overall developmental process appears to undergo gradual changes, distinct critical events signify the presence of significant transformations throughout development [39]. To elucidate the representative alterations across developmental stages, we conducted hierarchical clustering of all identified proteins (Fig. 2A). Proteins within Cluster 2 were associated with oxidative phosphorylation and thermogenesis, indicating that adults necessitate substantial energy to sustain their activities. Cluster 4 was enriched in pathways related to the degradation of valine, leucine, and isoleucine, as well as the PPAR signaling pathway, which is implicated in lipid metabolism. Considering the importance of BSFL, we illustrated the dynamic expression patterns of four key proteins within these two signaling pathways (Fig. 2B). FABP modulates lipid deposition through the regulation of the PPAR signaling pathway, while short-chain acyl-coenzyme A dehydrogenase (ACADSB) participates in both amino acid and lipid metabolism. Both proteins demonstrated a marked increase on the fourth day of the larval stage, maintaining elevated levels throughout larval development.

Fig. 2.

Analysis of proteome enrichment cluster and network dynamics of protein expression throughout development. (A) A heatmap illustrating the results of hierarchical clustering of differentially expressed proteins. The right panel present the top two most significantly enriched functional terms for each cluster, comprising Gene Ontology Biological Process (GO-BP) and KEGG pathways. Complete enrichment statistics for all clusters are provided in Supplementary Table S3. (B) Relative quantification of proteins associated with the PPAR signaling pathway (indicated in yellow) and the degradation of valine, leucine, and isoleucine (indicated in blue). (C) The hierarchical clustering tree and overview of the identified modules. (D) Network connections among proteins within the yellow and pink modules. (E) The primary KEGG pathways associated with the proteins within the various color-coded modules.

Subsequently, we applied weighted gene co-expression network analysis (WGCNA) to systematically characterize eight highly correlated protein expression patterns, each represented by a distinct color (Fig. 2C). Proteins expressed during the egg stage were primarily classified into the black and turquoise modules, whereas those expressed during the larval stage were grouped into the purple, orange, and pink modules. To visualize protein–protein interactions (PPIs), we constructed a network (Figs. 2D and S3). The orange module was notably enriched in ribosomal proteins, highlighting its significance in protein synthesis during early larval development (Fig. 2E). Importantly, the core proteins within the pink module included FABP and ACADSB, further underscoring their crucial roles in larval metabolism. KEGG pathway analysis of the pink module indicated significant enrichment in pentose and glucuronate interconversions, as well as in the PPAR signaling pathway. Furthermore, the gray module exhibited an interaction between cuticle protein 7-like (CP7L) and the ubiquitin-like modifier-activating enzyme ATG7 (Fig. S3).

Motif enrichment and phosphosignaling dynamics during BSF development

Bioinformatic analysis of flanking sequences has the potential to elucidate significant characteristics of modification sites. Consequently, we conducted an investigation into the sequence context surrounding these modified sites (Fig. 3A). Our analysis demonstrated that, during the phosphorylation of serine residues by protein kinases transferring a phosphate group from ATP, glutamate (E) is frequently enriched at positions +2 to +6, while proline (P) is commonly observed at positions ranging from −6 to +6. Conversely, phosphorylation is less likely to occur in the presence of cysteine (C) between positions −6 and +6, as well as lysine (K) at positions +1 and +2. For substrates containing threonine, phosphorylation is most favored when tyrosine (Y) is located at position +6. In total, we identified 77 distinct motifs, and we present their sequence logos and scores to further elucidate the substrate preferences of the enzyme (Fig. 3B and Table S1).

Fig. 3.

Analysis of motif and dynamics of phosphoproteome during the development of Hermetia illucens. (A) Characterization of the frequency of amino acids adjacent to modification sites. (B) Motif analysis of flanking sequence preferences for phosphorylation sites, with letter size corresponding to the frequency of each amino acid residue at that specific position. (C) Fuzzy c-means clustering revealed distinct temporal patterns in the expression modification of phosphorylated proteins. The right panel illustrates the top two most significantly enriched functional terms for each cluster, covering the three Gene Ontology categories (Biological Process: GO-BP, Cellular Component: GO-CC, Molecular Function: GO-MF) and KEGG pathways. Complete statistical data for the enrichment analysis are presented in Supplementary Table S4. (D) The observed changes in the phosphoproteome at various developmental time points were mapped onto the TOR signaling network.

We further clustered alterations in protein phosphorylation levels (Fig. 3C), resulting in four distinct expression profiles. Proteins within Cluster 3 exhibited low phosphorylation levels during the egg stage, which rapidly increased upon transition to the larval stage. Proteins in Cluster 3 were enriched in pathways related to arginine biosynthesis and glycolysis/gluconeogenesis, while proteins in Cluster 1 were associated with the TOR and insulin signaling pathways. TOR, a serine/threonine kinase, functions as a conserved cellular nutrient sensor across species from yeasts to mammals [40]. We constructed a phosphorylation-level network of the TOR signaling pathway to elucidate its role in the growth of BSF larval (Fig. 3D). Our findings indicated that phosphorylation levels of IRS1 and PDK1 increased from 72 h eggs to 1-day-old larvae. The increased phosphorylation of IRS1 and PDK1 suggests an upregulation of upstream insulin signaling. This signaling cascade, likely mediated through the canonical PI3K–AKT intermediaries, converges on TOR activation and provides a plausible explanation for the enhanced protein and lipid synthesis observed during larval growth [23,41]. Similar phosphorylation dynamics have been associated with TOR activation and anabolic metabolism in Drosophila and other insect species [42].

Landscape of kinase activity and regulatory inference derived from phosphoproteomic profiling

The regulatory interactions between kinases and their substrates throughout development exhibit a high degree of complexity. We identified a total of 2066 regulatory interactions involving 184 protein kinases and 908 phosphorylation sites across 432 substrate proteins. Employing gene set enrichment analysis (GSEA) [28], we evaluated kinase activity at various developmental stages, categorizing kinases as either activated or inhibited. Notably, activated kinases were predominantly observed at 3 h and 24 h post-oviposition and in adult specimens, whereas the majority of kinases were found to be suppressed during the pupal stage (Fig. S4A). Specifically, CDK1, CDK6, and CDK14 exhibited activation during the egg stage, while PAK and STK kinases were suppressed (Fig. S4B).

We conducted a detailed analysis of kinases across 17 subphases to clarify the relationship between specific phosphorylation events and the activity of signaling pathways. By concentrating on kinases with identifiable substrate motifs, we identified 26 kinases from the CMGC, AGC, atypical, and STE families (Fig. 4A). Each developmental stage presented both unique and shared phospho-site motifs that can be attributed to the activities of various kinases and kinase groups [15]. For example, CMGC kinases, including CDK1, p38beta-like, ERK-A, and PK shaggy, were consistently active throughout development, indicating their essential roles in cell cycle regulation. These kinases were often associated with motifs containing S/T-P and T-P-x-R. Notably, cyclin-dependent kinase 1 isoform X2 (CDK1 X2) exhibited the highest number of substrate motifs (4 motifs, 111,133,102 substrates).

Fig. 4.

Statistical analysis of kinase regulatory states across different developmental stages. (A) An enrichment analysis of p-site motifs and kinase prediction for regulated p-peptides is presented. The column labeled “N” indicates the number of p-peptide sequences that correspond to a specific motif. (B and C) Statistical data illustrating the relationship between kinase and their substrates are provided. Red represents activated kinases, while pink indicates differentially upregulated phosphorylation sites.

PI3K, a critical upstream regulator of the TOR signaling pathway, is instrumental in the integration of growth factor and nutritional signals [43]. Analysis focused on motif enrichment and kinase prediction have demonstrated that PI3KC3 (phosphatidylinositol 3-kinase catalytic subunit type 3), a key regulator of autophagy and vesicular trafficking, exhibits significant enrichment during the late larval and adult stages (Fig. 4A). Further examination of the phospho-kinase–substrate regulatory network revealed stage and sex-specific differences in kinase activity (Fig. 4B and C). Notably, TOR kinase was positively regulated in 1-day-old adult females compared to 6-day-old pupae. This enhancement is likely to augment TORC1 activity, thereby facilitating protein synthesis and growth in adult females. Conversely, PI3KC3 was found to be upregulated in 4-day-old adult females relative to their male counterparts, with notable increases in the phosphorylation of ATG4B at S338 and Vam6/Vps39 at S665.

Impact of FABP and Eip75B on lipid and protein metabolism in BSFL

BSFL are characterized by a high content of fats and proteins, with the former serving as a valuable feedstock for biodiesel production and the latter constituting an important source of sustainable protein. Among the key proteins implicated in lipid metabolism, fatty acid-binding protein (XP_037924714.1) has been identified as a central protein interacting with other lipid metabolism-related proteins (Fig. 2D). Notably, the FABP isoform is expressed at significantly elevated levels in BSFL compared to other developmental stages (Fig. 5A). Furthermore, FABP (XP_037924714.1) demonstrated a significant transcriptional correlation with its mRNA abundance (r > 0.8). Quantitative analysis indicates that the transcriptional level of FABP reaches its peak at the 8-day-old larval stage (Fig. 5B). Based on its open reading frame (ORF) annotation as HERILL_LOCUS15605 in the UniProt database, we provisionally designated the gene encoding this protein as L15605. Expression analysis revealed that L15605, along with genes involved in fatty acid synthesis such as acetyl coenzyme A carboxylase (ACC) and fatty acid synthetase (FAS), exhibited significantly higher expression levels in the fat body compared to the gut (Fig. S5A). Interference with L15605 significantly reduced its mRNA expression in both the gut and fat body (Fig. S5B and C), while also downregulating ACC and Eip75B expression levels. Consistently, the fat storage capacity of both the gut and the fat body exhibited a marked decrease (Fig. 5C). It is important to note that the interference with L15605 did not alter the levels of free fatty acid (Fig. 5D), although it did result in a significant increase in both acetyl-CoA synthetase (ACS) activity and crude protein content (Figs. 5E and S5D). Considering that in industrial production, the substrate is generally subjected to continuous bioconversion for 6 days, we extended the dsRNA feeding interference duration from 4 days to 6 days. Under this prolonged interference, both L15605 expression and triglyceride content were significantly reduced (Fig. S5E and F), while crude protein content continued to rise (Fig. S5G).

Fig. 5.

FABP and Eip75B positively mediate lipid synthesis in BSFL. (A) Alterations in the levels of fatty acid binding protein across all developmental stages, n = 3. (B) The temporal expression profile of the FABP gene, n = 4. (C) Variations in triglyceride content following the modulation of L15605 expression, n = 3. (D) levels of free fatty acid content, n = 6. (E) The enzymatic activity of acetyl-CoA synthetase (ACS), n = 3. (F) Representative images of BSFL subsequent to the modulation of Eip75B expression. (G) Changes in gene expression levels, n = 3. (H) Alterations in triglyceride content following the modulation of Eip75B expression, n = 3. (I) LSCM visualization of lipid droplets (LDs) in the BSFL fat body, which were permeabilized and stained with BODIPY. Scale bars: 50 μm. (J) Fluorescence intensity of LDs in each group after BODIPY staining. RU: Relative unit, n = 3. Data are measured as means ± SD. Statistical analysis was conducted using a two-tailed Student’s t-test, with n ≥ 3 biological replicates. Statistical significance is indicated as *P < 0.05, **P < 0.01, and ***P < 0.001, with ns denoting not significant (P > 0.05).

Eip75B, a homolog of PPARγ, which is a well-established regulator of FABP [44], has an unelucidated specific role in insect lipid metabolism. Silencing of Eip75B induced larval mortality, with death onset occurring from day 3 and a final survival rate of 92 % (Figs. 5F and S5H). Notably, larvae that died following Eip75B knockdown exhibited a significantly smaller body size compared with the surviving controls (Fig. S5I). This partially lethal phenotype, together with the observed growth retardation, supports the conclusion that Eip75B is indispensable for maintaining energy homeostasis and viability during larval development. Although reduced expression of Eip75B did not significantly alter the levels of L15605 in the gut, it caused a marked downregulation in the fat body (Figs. 5G and S5J). Correspondingly, silencing of Eip75B resulted in decreased triglyceride content in both the gut and fat body (Fig. 5H). Laser scanning confocal microscopy (LSCM) further revealed that the knockdown of both L15605 and Eip75B led to a reduction in both the number and size of lipid droplets (LDs) (Figs. 5I, J and S5K).

Eip75B-mediated transcriptional regulation underlies ecdysone-driven lipogenesis in Hermetia illucens

As the PPAR-related homolog in Drosophila, Eip75B displays a conserved domain structure in H. illucens (Fig. 6A). Treatment with GW9662 (PPARγ antagonist) resulted in a significant reduction in lipid accumulation and concurrently suppressed the expression of Eip75B as well as key lipogenic genes. Conversely, rosiglitazone (RO, PPARγ agonist) was found to enhance lipid production and upregulate these genes (Fig. S6A–D). Staining of the fat body indicated that GW9662 decreased lipid droplet diameter, while RO promoted lipid droplet accumulation (Figs. 6B and S6E, F). While no significant change in lipid droplet diameter was observed upon RO treatment, the significant increase in integrated density suggests a potential role in modulating total lipid content or accumulation. Subsequently, we purified the H. illucens Eip75B protein utilizing an Escherichia coli (E. coli) expression system (BL21 strain) and confirmed its direct binding to the L15605 promoter through EMSA (Fig. 6C), thereby establishing a transcriptional link between Eip75B and L15605.

Fig. 6.

Functional characterization of the ecdysone-Eip75B-FABP regulatory axis in lipid metabolism. (A) Comparative analysis of the domains of Eip75B in Drosophila and H. illucens. (B) Effect of rosiglitazone and GW9662 on LDs formation. (C) EMSA analysis of the transcription factor Eip75B binding to the promoter region of L15605. Probe 1 (−833 to − 783 bp), Probe 2 (−1207 to − 1157 bp), and Probe 3 (−1662 to − 1612 bp) correspond to predicted Eip75B-binding motifs within the 2-kb upstream promoter region of L15605. (D) The changes in gene expression levels following ecdysone supplementation, n = 4. (E) The changes in triglyceride content following ecdysone supplementation, n = 3. (F) Effects of ecdysone supplementation during RNAi-mediated silencing of L15605 on gene expression profiles and triglyceride concentrations, n = 4. (G) Effects of ecdysone supplementation during RNAi-mediated silencing of L15605 on gene expression profiles and triglyceride concentrations, n = 3. The data is presented as means ± SD. Statistical analysis was conducted using a two-tailed Student’s t-test, with n ≥ 3 biological replicates. Statistical significance is indicated as *P < 0.05, **P < 0.01, and ***P < 0.001, with ns denoting not significant (P > 0.05).

Subsequently, we examined the effects of ecdysone on the regulation of lipid metabolism. Dietary supplementation with 0.5 μM and 1.0 μM ecdysone in 50 BSFL resulted in a significant upregulation of the expression of the Eip75B, L15605, ACC, and FAS genes (Fig. 6D). Concurrently, triglyceride content exhibited the significant increase in both the gut and fat body (Fig. 6E), suggesting that ecdysone facilitates lipogenesis through this pathway. To further elucidate the regulatory hierarchy, we induced a deficiency of L15605 in the presence of 1.0 μM ecdysone. qRT-PCR analysis indicated that the expression of Eip75B remained suppressed, the inhibitory effect on L15605 expression was alleviated (Fig. 6F). Notably, triglyceride levels in both tissues were restored to control levels (Fig. 6G).

Enhancing fatty acid β-oxidation reprograms metabolic allocation in BSFL

Consistent with the observed increase in protein content, the expression levels of genes that positively regulate protein synthesis were significantly upregulated (Fig. 7A). In accordance with prior reports indicating that FABP5 deficiency enhances fatty acid β-oxidation (FAO) [45], the silencing of L15605 resulted in a notable upregulation of key FAO genes (Fig. 7B). Additionally, mitochondrial mass, as evidenced by an increase in mitochondrial DNA (mtDNA) copy number, was elevated following the knockdown of L15605 (Fig. 7C). These findings suggest that the previously observed absence of significant alterations in free fatty acid levels (Fig. 5D) may be indicative of a compensatory increase in FAO, serving as a homeostatic mechanism to mitigate excess free fatty acid accumulation.

Fig. 7.

Investigation of the relationship between fatty acid oxidation and protein synthesis. (A) The relative mRNA expression levels of genes associated with the TOR signaling pathway after interfering L15605 expression, n = 5. (B) The relative mRNA expression levels of genes involved in FA β-oxidation. Specifically, ACSL4 (long-chain-fatty-acid-CoA ligase 4), ACSL6, CPT1 (carnitine palmitoyltransferase 1), and CPT2, n = 4. (C) analysis of the mtDNA/nDNA ratio subsequent to the inhibtion of L15605 expression, determined from the relative levels of mitochondrial genes mt:CoI and mt:ND5 in comparison to genomic genes GAPDH and RP49, n = 3. (D and E) Assessment of changes in triglyceride and crude protein content following the addition of L-carnitine, n = 3. (F) Analysis of the mtDNA/nDNA ratio subsequent to the addition of L-carnitine, n = 3. (G) evaluation of changes in mRNA expression levels of CPT1, n = 4. (H) Assessment of changes in mRNA expression levels of genes related to FA β-oxidation after interfering the expression of CPT1, n = 4. (I) Examination of changes in triglyceride content after interfering the expression of CPT1, n = 4. Data are presented as means ± SD. Two-tailed Student’s t-test, with n ≥ 3 biological replicates. Statistical significance is indicated as *P < 0.05, **P < 0.01, and ***P < 0.001, with ns denoting not significant (P > 0.05).

To elucidate the intricate relationship between fatty acid utilization and protein synthesis, we employed two complementary approaches: enhancement of FAO through L-carnitine supplementation in the feed of BSFL and impairment of FAO via silencing of CPT1 expression. Notably, the incorporation of 10 mg of L-carnitine into the feed for every 50 larvae resulted in a marked reduction in fat accumulation, accompanied by a significant increase in crude protein content (Fig. 7D and E). Furthermore, the mtDNA/nDNA ratio exhibited an increase, suggesting enhanced mitochondrial biogenesis (Fig. 7F). At the molecular level, qRT-PCR analysis revealed elevated expression levels of CPT1 in the 2 mg group (Fig. 7G), alongside altered expression patterns of L15605 and CPT2 (Fig. S6G). RNAi-mediated silencing of CPT1 resulted in the downregulation of long-chain acyl-CoA synthetase4 (ACSL4) and ACSL6 expression (Fig. 7H), which was associated with a significant increase in triglyceride content (Fig. 7I). However, in contrast to the effects of L-carnitine supplementation, CPT1 knockdown did not influence crude protein levels (Fig. S6H), thereby reinforcing the notion that modulation of FAO can shift metabolic allocation in BSFL. These findings suggest a complex, dose-sensitive regulatory network that links fatty acid metabolism to protein biosynthesis.

Discussion

Efficient nutrient storage is essential for insect development and energy homeostasis; however, the functional regulatory systems governing nutrient allocation during ontogeny remain inadequately understood. This study aims to address this knowledge gap through an integrative multi-omics approach across the entire life cycle of H. illucens to elucidate the temporal dynamics of metabolic regulation. This systems-level analysis demonstrated that development across diverse taxa is orchestrated by layered transcriptional, translational, and metabolic networks [46]. In holometabolous insects, such as H. illucens, development is characterized by significant phenotypic and metabolic reprogramming across distinct life stages. In this research, we generated a comprehensive proteomic and phosphoproteomic dataset covering five developmental stages. These proteomic data were integrated with previously sequenced transcriptomes to elucidate the regulatory architecture of gene expression in BSF [23]. Our integrated analysis underscores the necessity of combining transcriptomics with proteomics and phosphoproteomics to resolve complex gene function and regulatory dynamics [47,48]. By linking these molecular profiles to functional experiments, we identified stage-specific regulatory networks that coordinate growth, nutrient metabolism, and developmental transitions.

Distinct patterns of protein expression and phosphorylation were correlated with developmental progression. Protein abundance reached its zenith during the larval stage, particularly on day 4 (Fig. 1E), coinciding with a period of rapid biomass accumulation and elevated metabolic activity [49]. Hierarchical clustering analysis revealed functional stage specificity: clusters enriched in larval proteins were associated with lipid metabolism (PPAR signaling) and ribosome-related protein synthesis, whereas clusters enriched in adult proteins were linked to oxidative phosphorylation and thermogenesis (Fig. 2A and E). Additionally, phosphoproteomic data indicated that the TOR and PI3KC3 signaling pathways serve as crucial regulators of nutrient sensing and metabolic adaptation. The phosphorylation of TOR substrates (S6K, 4EBP2) increased throughout larval development (Fig. 3D), consistent with enhanced protein synthesis and cellular growth. Concurrently, PI3KC3 activity peaked in late larvae and adults (Fig. 4A), correlating with autophagy-dependent nutrient recycling as the BSF transitions from feeding to non-feeding stages. Notably, sex-specific phosphorylation patterns of PI3KC3 substrates in adults (Fig. 4B) suggest divergent metabolic strategies between sexes, potentially linked to reproductive investment [[49], [50], [51]].

Insect lipids, particularly those derived from BSF, are increasingly recognized as sustainable feedstocks for biodiesel and bio-lubricants [52,53]. The biomass of H. illucens demonstrates advantageous biofuel properties, including elevated energy content and compatibility with diesel fuels [17,54]. Our multi-omics approach has elucidated key lipid regulators, including fatty acid-binding proteins (FABPs), which are known to facilitate the transport of fatty acids and modulate PPAR signaling [55,56]. In this study, we identified a highly expressed FABP gene in BSFL that is upregulated during periods of peak lipid accumulation (Fig. 5). Mitochondria is the key organelle involved in lipid metabolism, and maintaining its structural and functional integrity is essential for energy balance and lipid homeostasis [57]. The increase in mitochondrial DNA copy number following FABP knockdown suggests an attempt at mitochondrial biogenesis, which is consistent with the metabolic shift toward fatty acid β-oxidation. Loss of FABP function reprogrammed metabolism from lipid storage toward β-oxidation and mitochondrial biogenesis, an adaptive metabolic shift characteristic of insects to preserve energy balance when lipid trafficking is compromised [5]. Importantly, while prior studies have indicated specific accumulation of free long-chain unsaturated fatty acids in FABP5-deficient macrophages [45], our experimental findings illustrate that FABP impairment in insects does not result in significant alterations in free fatty acid levels. This discrepancy may be due to the unique regulatory characteristics of insect lipid metabolism, which operates through integrated hormonal and transcriptional networks that maintain systemic energy balance [58]. Insects possess coordinated, whole-body mechanisms for lipid management that are primarily mediated by the fat body. This system, modulated by insulin-like peptides and adipokinetic hormones, enables dynamic yet stable adjustments in lipid synthesis, storage, and mobilization under varying physiological and environmental conditions [5]. Such multilayered regulation provides the metabolic flexibility necessary for insects to maintain lipid homeostasis under fluctuating nutrient and environmental stress [59]. In recent years, omega-3 polyunsaturated fatty acids (PUFAs) have gained attention for their health benefits and strong anti-inflammatory [[60], [61], [62]]. Beyond identifying key regulators of efficient lipid synthesis in H. illucens, future studies could focus on the targeted and efficient biosynthesis of specific fatty acids, such as omega-3 PUFAs, to enhance the nutritional and functional value of insect-derived lipids.

In the regulation of lipogenesis and lipolysis, endocrine hormones are undoubtedly among the most critical regulators [63]. Among these hormones, neuropeptides such as adipokinetic hormone (AKH) and insulin-like peptides, which modulate lipid metabolism, have been extensively characterized in prior research [58,64]. In contrast, investigations into the regulatory effects of insect hormones, including the sterol hormone 20-hydroxyecdysone (20E) and juvenile hormone (JH), on lipid metabolism remain relatively limited. Ecdysone-induced protein 75 (E75), a nuclear receptor homologous to PPARγ, is known to influence lipid metabolism and aging [[65], [66], [67], [68]]. E75 is a classic early-response gene activated by the ecdysone receptor EcR/USP complex during Drosophila developmental transitions, and it is essential for coordinating organismal physiology with steroid hormone pulses [69]. Our results demonstrate that this core function is preserved in BSF, where Eip75B is indispensable for larval energy homeostasis and viability. However, we also observe evidence of functional divergence. In Drosophila, a prominent and specific role of E75 is its function as a heme-binding gas sensor within its HOLI domain that regulates circadian rhythms and juvenile hormone titers, creating a complex interface between metabolism, circadian biology, and hormone cross-talk [70]. In contrast, our study in BSF delineates a direct and central role for Eip75B in a larval-specific lipogenic program by identifying a specific FABP gene as its key transcriptional target. This suggests that while the overarching role of E75/Eip75B as a developmental and metabolic integrator is deeply conserved, its immediate regulatory targets and primary physiological outputs may have been tailored in the H. illucens to support its unique life history, which is dominated by the imperative for rapid larval lipid accumulation as a prepupal energy reserve. The precise evolutionary relationship and the full extent of functional divergence between these orthologs warrant further investigation, for instance, by determining whether the gas-sensing capability is preserved in the HOLI domain of H. illucens Eip75B and how it might influence its newly identified role in lipid metabolism.

To elucidate the mechanistic basis by which Eip75B regulates lipogenesis, we integrated our multi-omics data to construct a model of the ecdysone–Eip75B–FABP regulatory axis. Although Eip75B was not detected in the proteome, phosphoproteomics data indicated a significant 90 % decrease in phosphorylation levels from the 4th to the 8th day of larval development, suggesting a temporal regulatory transition. Phosphorylation is a critical post-translational modification controlling the activity, localization, and transcriptional specificity of nuclear receptors [71]. Thus, this pronounced dephosphorylation likely reflects the conversion of Eip75B from an active, DNA-binding state to an inactive conformation. We propose that Eip75B undergoes an early, transient activation phase, during which its phosphorylated form promotes the transcription of FABP, initiating lipogenic gene expression. Such transient activation events are characteristic of ecdysone-responsive nuclear receptors, which exhibit pulse-like dynamics to temporally coordinate developmental and metabolic transitions [72]. This model aligns with the broader principle that brief signaling events can trigger long-lasting transcriptional responses, leading to stable physiological changes [73]. As development proceeds, the gradual dephosphorylation of Eip75B protein may reflect a self-limiting feedback mechanism, similar to the ligand-dependent downregulation in other nuclear receptors such as PPARγ [74]. Collectively, these findings delineate a temporal regulatory hierarchy, in which a phosphoproteomic-defined switch in Eip75B activity orchestrates the downstream transcriptomic and proteomic remodeling of lipid metabolism in H. illucens.

Furthermore, we noted Eip75B knockdown reduced L15605 (FABP) expression in the fat body but not in the gut. This observation underscores the metabolic compartmentalization within insects, where the fat body, as the primary organ for systemic energy metabolism, is highly dependent on Eip75B for transcriptional control of lipid-related genes [75]. In contrast, the gut, specialized for nutrient absorption and immune defense, may rely on distinct transcriptional programs. Potential compensatory mechanisms could involve other nuclear receptors, such as DHR96, a key regulator of midgut lipid metabolism, or transcription factors like the GATA-family member Serpent [76,77]. In addition to its tissue-specific regulatory role, Eip75B also exhibits conserved PPARγ-like functionality in lipid metabolism (Fig. 6). Rosiglitazone treatment enhanced total lipid accumulation, as evidenced by increased integrated staining density, no corresponding changes in the average diameter of lipid droplets [78,79]. This pattern of lipid metabolic activation, characterized by elevated integrated intensity without changes in mean diameter, is consistent with established models of lipid storage regulation [80,81]. Collectively, our findings validate Eip75B as a central regulator of lipid metabolism in H. illucens and delineate an ecdysone–Eip75B–FABP regulatory axis that integrates hormonal signaling with transcriptional control of lipid homeostasis (Fig. 8). Contrary to the prevailing view of 20E as an inducer of lipid catabolism [[82], [83], [84]], our data demonstrate that appropriate concentrations of 20E significantly stimulate lipogenic gene expression and promote triglyceride accumulation. Although 20E treatment alone robustly induced Eip75B, the induction was attenuated when L15605 (FABP) was simultaneously silenced. This pattern is consistent with L15605 participating in lipid-mediated positive feedback that sustains nuclear-receptor activity. FABPs are intracellular lipid chaperones that facilitate ligand delivery and can enhance ligand-dependent nuclear receptor activation [[85], [86], [87]]. Loss of FABP function may therefore reduce local ligand or coactivator availability required for full Eip75B transcriptional activation, attenuating the hormone-driven response. Similar integration of ecdysone and metabolic signals by E75 family receptors has been documented in Drosophila, supporting the plausibility of this model [88]. In addition to lipids, insect-derived proteins have recently garnered considerable attention as a sustainable and eco-friendly protein source [89]. Our experimental findings indicate that L-carnitine supplementation enhances mitochondrial fatty acid oxidation (FAO), promotes protein accumulation, and reduces lipid storage. Conversely, silencing CPT1 impaired FAO and increased triglyceride accumulation without affecting protein levels (Fig. 7), suggesting a metabolic prioritization towards protein maintenance during periods of suppressed lipid catabolism. The metabolic reprogramming of H. illucens to enhance lipid catabolism represents a potential strategy to improve feed utilization efficiency and augment protein synthesis and deposition, thereby facilitating sustainable waste management.

Fig. 8.

The ecdysone-Eip75B-FABP axis regulates insect lipid-nutrient homeostasis. The deficiency of FABP reprograms metabolism toward mitochondrial β-oxidation and protein synthesis. Ecdysone induced protein 75B (Eip75B), which serve as the insect PPAR ortholog, directly interact with the FABP promoter to enhance its expression, as confirmed by EMSA. Pharmacological investigations indicate that the PPARγ agonist GW9662 enhances lipid deposition through Eip75B, while the antagonist rosiglitazone produces contrasting effects. Furthermore, 20-Hydroxyecdysone (20E) rescues lipid deficits associated with FABP deficiency by stabilizing the activity of the Eip75B-FABP axis. This finding establishes a conserved hormonal pathway wherein the 20E-Eip75B-FABP axis regulates lipid homeostasis and nutrient partitioning in insects.

By synthesizing proteomic, phosphoproteomic, and transcriptomic datasets, we construct dynamic regulatory networks controlling nutrient partitioning and metabolic transitions during insect ontogeny. We identify the ecdysone-Eip75B-FABP axis as an evolutionarily conserved regulatory hub that integrates hormonal and nutrient signaling in insect. This axis is responsible for driving larval lipid storage through Eip75B-mediated transcriptional activation of FABP, while its disruption initiates metabolic reprogramming toward mitochondrial β-oxidation and protein synthesis. We delineate a hierarchical regulatory cascade that links ecdysone signaling to lipid chaperone function. This axis regulates lipid-protein allocation, presenting actionable targets for enhancing feed efficiency and protein yield through metabolic engineering. Finally, it is important to acknowledge that the reliance on whole-organism RNAi, while informative, may mask further tissue-specific nuances of this axis. Future studies employing tissue-specific techniques will be essential to fully dissect its spatial functions and regulatory complexity. These findings address a significant gap in our understanding of insect nutrient regulatory networks, establish the ecdysone-Eip75B-FABP axis as a target for sustainable bioresource engineering, and connect mechanistic biology with circular economy objectives.

Conclusion

The nutritional metabolism during the larval stage of insects undergoes dynamic regulation to support rapid growth and developmental transformation. Ecdysone signaling activates Eip75B, which transcriptionally upregulates FABP expression to maintain lipid homeostasis, while disruption of this axis reprograms metabolism toward mitochondrial β-oxidation and protein synthesis. The ecdysone-Eip75B-FABP regulatory cascade plays a central role in coordinating larval nutrient allocation and metabolic plasticity. Our findings reveal a novel mechanism through which steroid hormone signaling integrates with nutrient-sensing pathways to regulate FABP-mediated lipid metabolism in insects. Given its crucial role in lipid storage, energy mobilization and protein synthesis, the ecdysone-Eip75B-FABP axis and its downstream effectors represent promising targets for metabolic engineering approaches aimed at optimizing BSF for bioconversion applications, either by enhancing lipid yields for biofuel production or increasing protein accumulation for sustainable feed.

Compliance with ethics requirements

This article does not contain any studies with human or animal subjects.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following are the Supplementary data to this article: