Introduction

The molecular target of benzoylphenylurea (BPU) insecticides—a multi-billion-dollar class of insect growth regulators widely used in agriculture—remains one of the most significant unresolved controversies in insecticide toxicology [1]. Since their introduction over over four decades ago, BPUs have been known to catastrophically disrupt chitin formation, causing molting failure and larval death [2]. Despite this clear phenotypic effect, their direct molecular target has not been definitively identified, creating a persistent gap between their in vivo activity and their presumed mechanism of action.

The prevailing hypothesis posits that BPUs directly inhibit chitin synthase (CHS), the enzyme catalyzing the final polymerization of UDP-N-acetylglucosamine (UDP-GlcNAc) into chitin. This hypothesis is supported by indirect evidence, including the accumulation of UDP-GlcNAc upon BPUs treatment [3] and the genetic co-localization of BPU resistance mutations with the CHS locus in multiple insect species, most notably Plutella xylostella [1, 4, 5]. However, direct biochemical studies, beginning with the seminal work by Deul et al. in 1978 [6], have consistently failed to demonstrate CHS inhibition by BPUs in cell-free assays, even at supralethal concentrations [7-9]. This discrepancy between genetic association and enzymatic dataology argues against it—represents an unresolved paradox that has significant implications for understanding BPUs resistance mechanisms and for rational insecticide design.

A largely overlooked clue may help resolve this contradiction: the core acylurea structure of BPUs (Figure 1) closely resembles that of potent inhibitors of mammalian Glycogen Phosphorylase (GP) developed as therapeutic agents for type 2 diabetes [10, 11]. GP catalyzes the rate-limiting step of glycogenolysis, the phosphorolytic cleavage of glycogen to glucose-1-phosphate, which serves as the metabolic gateway controlling flux into the chitin biosynthetic pathway [12]. In insects, this pathway is particularly critical: GP-derived glucose is converted to trehalose (the primary hemolymph sugar), which is subsequently channeled through a series of enzymatic steps to generate UDP-GlcNAc, the substrate monomer for chitin synthesis [13]. GP also plays key roles in carbohydrate homeostasis, osmotic regulation, stress adaptation [14-17]. This structural similarity between BPUs and GP inhibitors (GPIs), combined with GP’s upstream regulatory position in chitin precursor synthesis, raises a compelling alternative hypothesis: BPUs may exert their insecticidal effects not by inhibiting CHS directly, but by targeting GP and thereby depleting chitin biosynthesis of its essential precursors.

Figure 1.

Here, we systematically test this hypothesis and evaluate GP’s potential as an insect control target. We cloned, expressed, and biochemically characterized GP from the globally significant agricultural pest P. xylostella (PxGP), measured its inhibition by the BPU—diflubenzuron (DFB) and a specific mammalian GPI, and assessed the physiological consequences of GP loss-of-function in vivo. Our findings definitively exclude GP as the target of DFB. More significantly, we identify a robust metabolic compensatory response—the upregulation of gluconeogenesis—that completely rescues insects from GP inhibition. This discovery not only explains why GP alone is not a viable insecticidal target but also reveals a principle of metabolic plasticity that must be considered in the design of metabolism-targeting pesticides.

Results

Recombinant P. xylostella glycogen phosphorylase (PxGP) is a functional, activatable enzyme

To investigate its biochemical properties and evaluate its potential as an insecticidal target, the full-length PxGP cDNA (predicted molecular mass of 98.1 kDa) was cloned and expressed in Sf9 cells using the Bac-to-Bac baculovirus expression system. Following purification via immobilized metal-affinity chromatography (IMAC) of the N-terminal 6×His tagged protein [18], SDS-PAGE analysis and Western blot analysis confirmed a highly pure protein band migrating at approximately 100 kDa, consistent with its theoretical mass (Figure 2A). A total yield of 7.5 mg of purified protein was obtained from 8.25 g of cell pellet (specific yield=0.91 mg/g cells).

Figure 2.

Glycogen phosphorylase activity is regulated by reversible phosphorylation, existing as a less-active dephosphorylated b form (PxGP-b) and a fully active phosphorylated a form (PxGP-a) [19]. Recombinant PxGP-b was activated by incubation with immobilized phosphorylase kinase (PhK) in the presence of ATP and Mg²⁺ for 2 h [20]. The activated enzyme showed robust catalytic activity in a coupled-enzyme spectrophotometric assay, with NADPH production (monitored by increase in A₃₄₀) increasing linearly for the first 15 min and reaching a plateau by approximately 20 min, consistent with substrate depletion and/or product accumulation. The specific activity of activated PxGP-a was 294.25 ± 3.45 U/mg protein (Figure S1A), confirming that the recombinant enzyme is catalytically competent for inhibitor screening.

GPI potently inhibits PxGP in vitro and in vivo

Given the structural similarity between insecticidal BPUs and mammalian GPIs [11] (Figure 1), we tested whether PxGP could be a target of acylurea-scaffold compounds. GPI exhibited potent, dose-dependent inhibition of recombinant PxGP-a, with a sigmoidal dose-response curve typical of competitive or allosteric inhibition (Figure 2B). Non-linear regression analysis yielded an IC₅₀ value of 2.96 nM (95% confidence interval [CI]: 1.66–5.35 nM, Hill slope = 0.74, R² = 0.877, n = 3). Near-complete inhibition (>95%) was achieved at 500 nM GPI, indicating saturable, high-affinity binding. The potency of GPI against PxGP is comparable to its effect on mammalian orthologs (see below) [11].

In striking contrast, the BPU insecticide Diflubenzuron (DFB), which shares the acylurea core with GPI (Figure 1), did not effectively inhibit PxGP-a. No standard sigmoidal dose-response curve was observed over 0.54 nM to 1700 nM (Figure 2C). Maximum inhibition observed did not exceed 61.96 ± 2.54% (mean ± SEM) even at 100 µM, with high variability at lower concentrations. These data indicate that DFB, unlike GPI, is not an effective inhibitor of PxGP, ruling out GP as the direct molecular target of this BPU insecticide.

Consistent with the recombinant enzyme results, GPI effectively inhibited native GP activity in crude lysates from third-instar larvae in a dose- and time-dependent manner (Figure S1B). After 45 min of pre-incubation, 10 µM GPI reduced native GP activity by 57.52 ± 1.88% (n = 3). DFB, however, showed no significant inhibitory effect on native GP activity at any concentration tested (10⁻⁷, 10⁻⁶, or 10⁻⁵ M) or time point (5, 10, 20, or 30 min), with activity remaining at 97-108% of control (Figure S1C).

Collectively, these results establish PxGP as a high-affinity molecular target for GPI but definitively exclude it as a target for DFB. This resolves the primary hypothesis by demonstrating that BPUs do not inhibit insect glycogen phosphorylase.

Paradoxical finding: potent PxGP inhibition by GPI lacks in vivo toxicity

Because GPI potently inhibit PxGP, the rate-limiting enzyme in glycogenolysis, which supplies precursors for chitin synthesis [21], we expected it to exhibit significant insecticidal activity. In leaf-dip bioassays with third-instar P. xylostella larvae continuously exposed to GPI (250 or 500 mg/L) or DFB (125 or 250 mg/L), GPI caused no significant toxicity or developmental disruption across the entire time course (24 to 120 h post-treatment) (Figure 3A). At 120 h, mortality with 500 mg/L GPI was only 12.78 ± 2.42% (n = 3), which was not significantly different from the solvent control (CK: 5.00 ± 5.00%, P = 0.21) (Figure 3B). Larvae exposed to 250 mg/L GPI exhibited even lower mortality (2.50 ± 2.50%).

Figure 3.

In stark contrast, DFB induced significant, dose-dependent mortality. At 120 h, DFB treatment at 125 mg/L caused 41.60 ± 4.85% mortality (P < 0.01 vs. control) and at 250 mg/L caused 61.36 ± 10.02% mortality (P < 0.01 vs. control). Probit analysis of the DFB dose-response data yielded calculated LC₃₀ and LC₅₀ values of 141.900 mg/L (95% CI: 67.893–206.576 mg/L) and 229.537 mg/L (95% CI: 145.287-331.784 mg/L), respectively, with an acceptable goodness-of-fit (χ² = 11.946, P = 0.683) (Table S2).

Furthermore, GPI-treated larvae showed no developmental abnormalities. Pupation rates were 64.17 ± 4.79% for the 500 mg/L GPI group and 69.72 ± 6.81% for the 250 mg/L group, which were statistically indistinguishable from the control group (64.72 ± 9.33%, P > 0.05 for both comparisons) (Figure 3B, middle panel). Similarly, adult emergence rates were 56.11 ± 5.64% and 56.39 ± 6.24% for the 500 and 250 mg/L GPI groups, respectively, compared to 54.10 ± 11.91 % for the control (P > 0.05, Figure 3B). Phenotypically, GPI-treated larvae molted normally, fed actively, and displayed no cuticle defects (Figure 3C). DFB-treated larvae exhibited the characteristic phenotypes of BPU intoxication: catastrophic molting failure, larval-pupal intermediates, and complete failure of adult emergence (Figure 3C and 3D). This striking discrepancy between biochemical potency and in vivo inefficiency presents a profound paradox that demands mechanistic investigation.

GPI exposure triggers a compensatory upregulation of PxGP transcription

To explore this discrepancy, we measured PxGP gene expression after GPI exposure. RT-qPCR revealed that 500 mg/L GPI rapidly upregulated PxGP mRNA transcription. As shown in Figure 4, its relative expression increased 3.24-fold at 24 h, peaked at 3.48-fold at 48 h, and remained elevated at 96 h. This indicates that the insect responds to the pharmacological inhibition by transcriptionally increasing the de novo synthesis of the target enzyme.

Figure 4.

Sequence homology and rationale for target site selection

To achieve effective gene knockdown coupled with functional ablation, precise selection of the interference target site is critical. Sequence analysis of PxGP (837 residues) revealed that residues 88-680 form the active site pocket of the GT35_Glycogen_Phosphorylase family, and the C-terminal region contains a phosphatase-pyridoxal phosphate linkage site (residues 672-684). Eleven residues within positions 42-318 are conserved and correspond to the AMP-binding site in human GP (Figure 5A) [22].

Figure 5.

Multiple sequence alignment of PxGP with GP orthologs from other key lepidopteran pests (Helicoverpa armigera HaGP, Spodoptera exigua SeGP, Ostrinia furnacalis OfGP), human (HsGP), and rabbit (OcGP) demonstrated high conservation across species (Figure 5B). PxGP shares 82.29% sequence identity with human GP (Figure S2). Since GP activity is allosterically regulated by AMP binding and phosphorylation at Ser14, triggering a conformational shift from the inactive T-state GPb (unphosphorylated) to the active R-state GPa (phosphorylated) [19], we targeted the core AMP-binding site region (gray box in Figure 5B) for RNAi. Theoretically, interference in this region should prevent AMP-mediated activation of existing GPb while degrading the entire mRNA pool, drastically reducing total GP protein synthesis and ensuring complete functional ablation.

RNAi confirms PxGP is non-essential for larval development

The compensatory upregulation of PxGP mRNA expression in response to pharmacological inhibition (Section 3.4) indicated that the gene is under active transcriptional regulation and suggested that increased enzyme synthesis might partially mitigate the effects of GPI. However, this compensation alone could not fully explain the complete absence of phenotype. To determine whether a more profound loss-of-function—achieved by directly suppressing PxGP gene expression—is lethal, we performed RNAi to knock down PxGP at the transcriptional level.

Firstly, we quantified the expression profile of PxGP across developmental stages in P. xylostella, PxGP was highly expressed in fourth-instar larvae (L4) and adults (Figure 6), stages that correlate with the energy demands for metamorphosis and flight, respectively [12, 23]. Given the inherent time lag required for systemic dsRNA delivery to achieve effective post-transcriptional knockdown [24], dsRNA was administered during the mid-third-instar larval stage. Mid-third-instar larvae were microinjected with dsRNA targeting a conserved 606 bp region of PxGP (dsGP) or with a non-specific control dsGFP. Three concentrations of dsGP (6,000, 10,000, and 14,000 ng/µL) were tested, and knockdown efficiency was assessed by RT-qPCR at 24, 48, 72, and 96 h post-injection (Figure 7). The highest concentration (14,000 ng/µL) achieved maximal suppression at 48 h, reducing PxGP transcript levels to 12.47 ± 1.68% of the dsGFP control (87.59% knockdown, P < 0.001) (Figure 7B). Significant knockdown persisted at 72 h (69.86% suppression,P < 0.05), with partial recover by 96 h, likely due to dsRNA degradation and/or compensatory transcriptional upregulation. The 10,000 ng/µL dose also produced significant knockdown at 48 h (65.76% suppression, P < 0.01), whereas the 6,000 ng/µL dose showed more modest effects (Figure 7B).

Figure 6.

Figure 7.

Despite this robust, transient silencing, no adverse phenotypic effects were observed. Developmental stage distribution and mortality rates in all dsGP-treated groups (6,000, 10,000, and 14,000 ng/µL) were statistically indistinguishable from the dsGFP control group across the entire 120 h observation period (P > 0.05 at all time point) (Figure 8A and 8B). At 120 h, cumulative mortality was 8.10 ± 4.23%, 9.82 ± 0.81%, and 3.33 ± 3.33% for the 6,000, 10,000, and 14,000 ng/µL dsGP groups, respectively, comparable to the dsGFP control (Figure 8B). Pupation and adult emergence also indistinguishable from controls (Figure 8B). Phenotypic examination revealed that dsGP-injected larvae developed normally, molted successfully, formed morphologically normal pupae, and emerged as viable adults without cuticle defects or developmental delays (Figure 8C and 8D). Thus, acute, profound suppression of PxGP expression is not lethal to P. xylostella during the critical larval-pupal transition, despite the enzyme’s established role in glycogenolysis and chitin precursor synthesis. This finding necessitates an investigation into the compensatory mechanisms that allow the insect to survive in the absence of functional GP.

Figure 8.

PxGP knockdown coordinately suppresses the glycogenolysis-to-chitin pathway

To elucidate how larvae survive GP loss, we examined downstream genes linking glycogenolysis to chitin synthesis. Specifically, we measured the expression of trehalase (PxTre), which hydrolyzes circulating trehalose to glucose [12], and hexokinase (PxHex), which phosphorylates glucose to glucose-6-phosphate—a key branch point leading to UDP-GlcNAc synthesis for chitin production [25, 26].

RT-qPCR analysis revealed that suppression of PxGP triggered a coordinated transcriptional downregulation of both downstream genes (Figure 9A and 9B). At 48 h post-injection, coinciding with the peak of PxGP knockdown (87.59% suppression), PxTre expression was reduced to 30.46 ± 1.25% of the dsGFP control (69.54% suppression, P < 0.001,), and PxHex to 22.82 ± 3.39% of control (77.18% suppression, P < 0.001). This suppression persisted at 72 h (PxTre: 61.27 ± 2.21% suppression, P < 0.001; PxHex: 41.96 ±4.02% suppression, P < 0.001) and returned to near baseline levels by 96 h, mirroring the recovery kinetics of PxGP itself.

Figure 9.

This coordinated downregulation indicates that the insect responds to the loss of GP not by futilely attempting to maintain flux through a pathway that lacks its key upstream enzyme, but rather by implementing a broader transcriptional shutdown of the glycogenolysis-to-glucose-to-chitin pathway. This suggests a coherent metabolic response program, raising the question: if glycogenolysis is blocked, how does the insect obtain the glucose required for survival and chitin synthesis?

Biphasic metabolic response: transient depletion followed by gluconeogenic rescue

The absence of a lethal or developmental phenotype despite glycogenolysis pathway suppression suggested activation of alternative glucose synthesis pathway to compensate for GP loss. We therefore measured expression of key gluconeogenic enzymes and metabolite levels across the glycogeno-glucose axis.

Gene expression analysis reveals gluconeogenic activation

We assessed transcript levels of two rate-limiting gluconeogenic enzymes: phosphoenolpyruvate carboxykinase (PEPCK, encoded by PxPEPCK), which catalyzes the conversion of oxaloacetate to phosphoenolpyruvate; glucose-6-phosphatase (G-6-Pase, encoded by PxG-6-Pase), which catalyzes the dephosphorylation of glucose-6-phosphate to free glucose. These enzymes are reciprocally regulated relative to glycolysis/glycogenolysis and serve as molecular markers of gluconeogenic flux [27, 28].

At 96 h post-injection―a critical pre-pupal stage with high carbohydrate demand for pupal cuticle synthesis [29, 30]―both PxPEPCK and PxG-6-Pase were significantly upregulated in dsGP-treated larvae (3.95 ± 0.44-fold and 3.34 ± 0.20-fold versus dsGFP control, respectively; P < 0.001 for each) (Figure 9C and 9D). This indicates robust transcriptional activation of the gluconeogenesis pathway in response to GP loss.

Protein catabolism provides gluconeogenic substrate mobilization

To identify the carbon source for gluconeogenesis, we quantified total protein content in dsGP-treated larvae across the time course. Strikingly, a significant 29.39% reduction in protein was observed at 72 h post-injection (P < 0.01), coinciding precisely with the nadir of glucose insufficiency (Figure 9D).

This protein decline provides direct biochemical evidence for substrate mobilization: the degradation of cellular proteins liberates amino acids, which serve as the primary carbon source for gluconeogenesis in insects [27, 31-33]. The magnitude of protein loss (∼30%) is sufficient to provide substantial amino acid flux into the gluconeogenic pathway.

Metabolite profiling reveals a biphasic adaptation

Quantification of four key metabolites (glycogen, trehalose, glucose-6-phosphat [G6P], and free glucose) along the glycogen-glucose axis at 48, 72, and 96 h post-injection uncovered a dynamic, biphasic response (Figure 9E-9H).

1. Glycogen Content: Contrary to the expectation of accumulation upon GP blockade, glycogen levels remained unchanged relative to controls throughout the time course (P > 0.05) (Figure 9E). This paradox indicates that insects activate alternative, GP-independent pathways for glycogen catabolism―most likely via α-amylase or glycogen debranching enzymes [34-36]―to provide emergency glucose mobilization during the period of metabolic stress, suggesting that these compensatory pathways are sufficient for short-term energy needs but are eventually superseded by gluconeogenic glucose production.

2. Trehalose: Trehalose levels remained stable at 48-72 h but surged dramatically by 96 h (7.44-fold elevation, P < 0.001) (Figure 9F and 9G). This massive accumulation, coinciding precisely with the peak gluconeogenic gene (e.g., PEPCK and G-6-Pase) expression, indicates that newly synthesized glucose derived from gluconeogenesis is being actively channeled into trehalose synthesis. The magnitude of this accumulation indicates that gluconeogenic glucose production exceeds the immediate metabolic demands of the larva, creating a metabolic “overshoot” that ensures adequate carbohydrate reserves for the upcoming pupation [37].

3. G6P: G6P, which sits at the intersection of glycolysis, gluconeogenesis, the pentose phosphate pathway, and glycogen synthesis [38], exhibited a pattern mirroring that of trehalose (Figure 9G). G6P content declined gradually from 48 to 72 h, consistent with reduced glucose mobilization from glycogen, but then increased sharply at 96 h (6.49-fold elevation, P < 0.001). This surge directly reflects the massive influx of glucose generated via the gluconeogenic pathway, which enters the metabolic network at the G6P node. The coordination between G6P accumulation and trehalose accumulation confirms that gluconeogenesis-derived glucose is being efficiently converted into storage and transport carbohydrates.

4. Free Glucose: Free glucose levels showed the most subtle but perhaps most physiologically informative pattern (Figure 9H). At 48 h―coinciding with peak PxGP suppression―glucose content was statistically indistinguishable from controls (P > 0.05), indicating that initial homeostatic mechanisms (including GP-independent glycogen breakdown and dietary carbohydrate absorption) successfully buffer against acute GP loss. However, at 72 h, the nadir of the metabolic stress response, free glucose levels declined significantly, albeit modestly (18.15% reduction, P < 0.05). This transient dip reveals a critical window of metabolic insufficiency during which the insect’s compensatory mechanisms are insufficient to fully maintain glucose homeostasis. Importantly, by 96 h, glucose levels fully recovered to control values (P > 0.05), demonstrating that the gluconeogenic pathway successfully restores glucose homeostasis before critical developmental transitions occur.

Integrated interpretation

Collectively, these metabolite and gene expression data reveal a sophisticated, biphasic metabolic adaptation strategy that fully explains the paradox of why profound GP loss (87.59% knockdown) causes no mortality or developmental defects:

1. Phase I (emergency compensation, 48-72 h): Metabolic stress and emergency compensation during the initial response to PxGP suppression, the insect experiences genuine metabolic stress, evidenced by the glycogen levels neither accumulated nor were depleted and the transient reduction in free glucose at 72 h. The static glycogen content indicates activation of GP-independent catabolism pathways, which provide emergency glucose but are insufficient for complete homeostatic maintenance.The modest glucose dip at 72 h demonstrates that this stress is real but tolerable, likely due to several factors: (i) the magnitude is modest (only 18.15% reduction), (ii) the duration is brief (∼24 h), (iii) the timing coincides with the mid-L4 larval stage when energy demands are moderate and active feeding allows dietary carbohydrate supplementation, and (iv) trehalose stores (which remain stable during this phase) can be mobilized to buffer hemolymph glucose.

2. Phase II (gluconeogenic rescue, 96 h): Robust transcriptional upregulation of PEPCK and G-6-Pase (3-4-fold increases) drives massive de novo glucose synthesis from non-carbohydrate precursors such as amino acids, lactate, and glycerol. The surge of gluconeogenesis-derived glucose is rapidly converted into trehalose and G6P, creating a metabolic overshoot that restores basal glucose levels. This overcompensation ensures adequate carbohydrate supply for the energetically demanding process of pupal cuticle synthesis, which begins shortly after 96 h.

This substrate-to-product chain for gluconeogenesis-mediated metabolic rescue—protein degradation → amino acid release → gluconeogenic conversion → glucose synthesis → trehalose storage—constitutes a complete compensatory mechanism that sustains development despite profound GP ablation.

Discussion

Resolving a long-standing controversy: BPUs do not target GP

A central challenge in mechanism-based insecticide development is distinguishing potent in vitro biochemical inhibitors from physiologically effective in vivo toxicants. This study addresses this challenge by investigating the debated molecular target of BPU insecticides. For over 40 years, the field has contented with conflicting evidence: while genetic studies implicated chitin synthase [5], direct biochemical assays consistently failed to validate it as a direct BPU target [6, 39, 40]. The structural similarity between BPUs and mammalian GPIs proposed an alternative hypothesis that BPUs might inhibit insect GP, thereby limiting glucose-derived precursors for chitin synthesis. Our systematic investigation demonstrates that while a human GPI potently inhibits PxGP, the BPU insecticide diflubenzuron does not. This discrepancy is attributable to key substitutions in its aromatic halogen substituents, which preclude effective binding to GP’s active or allosteric sites. This provides the first direct biochemical evidence excluding GP as a candidate molecular target for BPUs. Consequently, BPU lethality must arise from a different mechanism, likely involving interference with chitin assembly, cuticle organization, or CHS trafficking rather than direct enzymatic inhibition, a mechanism that remains to be conclusively identified [21, 39, 40].

Metabolic plasticity and gluconeogenic compensation underlie GP non-essentiality

The resolution of the BPU target question raised a more profound and unexpected puzzle: despite being a potent GP inhibitor in vitro, the GPI exhibited no insecticidal activity in vivo. This paradox—wherein a nanomolar-potency inhibitor of a putatively essential metabolic enzyme causes no phenotype—demanded mechanistic investigation.

Our investigation shows that this tolerance stems from a robust, developmentally-coordinated metabolic compensation mechanism. Even acute RNAi-mediated suppression of PxGP expression (87.59% knockdown) induced no discernible phenotype, as insects activated a compensatory gluconeogenic pathway. This was evidenced by the concerted suppression of downstream glycogenolytic enzymes and, crucially, the significant upregulation of the key gluconeogenic enzymes PEPCK (3.95-fold) and G-6-Pase (3.34-fold) at 96 h post-knockdown. This response provides a metabolic “escape route,” synthesizing glucose de novo from non-carbohydrate precursors to bypass the blocked glycogenolytic pathway and maintain homeostasis, particularly during the high carbohydrate demand preceding metamorphosis. Future work utilizing transcriptomic and proteomic approaches will be essential to delineate the signaling pathways (potentially involving nutrient sensors like AMPK and transcription factors such as FOXO) that link GP suppression to this precise reprogramming of metabolic gene expression [41, 42].

Direct metabolite quantification validated this compensation and revealed its sophisticated temporal dynamics. The most striking finding was the lack of glycogen accumulation upon GP knockdown, a paradox explained by the activation of GP-independent glycogen catabolism via alternative enzymes like α-amylase (which hydrolyzes internal α-1,4-bonds) and glycogen debranching enzymes (which hydrolyze α-1,6-branch points) [34-36]. Definitive metabolic rescue emerged at 96 h, characterized by a dramatic surge in trehalose and G6P, directly driven by the 3-4-fold upregulation of PEPCK and G-6-Pase, providing definitive biochemical proof that de novo glucose synthesis is the primary rescue mechanism. This biphasic response—initial metabolic stress followed by delayed (48-72 h), robust compensation—creates a temporal buffer (96 h), explaining the absence of phenotypes despite severe GP knockdown. This temporal coordination suggests a programmatically timed, possibly hormonally regulated response that anticipates developmental energy needs. Critically, it reveals a metabolic vulnerability window (∼72 h); strategies that disrupt both glycogenolysis and gluconeogenesis simultaneously could overcome this compensation.

Implications for insecticide target selection: metabolic redundancy and the dual-target strategy

Our findings offer direct and actionable insights for the rational design of metabolism-targeting insecticides. A key lesson is that the viability of a metabolic enzyme as an insecticidal target is determined not solely by its biochemical function or pathway position but by the presence of compensatory metabolic routes. GP fails as a standalone target due to compensation via gluconeogenesis. This contrasts sharply with other carbohydrate-metabolizing enzymes that have proven to be highly effective insecticidal targets. For example, trehalase inhibitors such as Validamycin A are potent insecticides [43, 44], suggests that inhibition of trehalase does not cause other pathways to initiate glucose compensation. This distinction informs a refined strategy: dual-target inhibition. Simultaneously blocking both glycogenolysis (via GP) and gluconeogenesis (e.g., via PEPCK or G-6-Pase) would create irreversible glucose starvation, a “metabolic pincer” attack that could overcome adaptive compensation—a strategy with precedent in cancer therapy targeting both glycolysis and oxidative phosphorylation [45, 46]. Ultimately, our work highlights the value of systems-level metabolic modeling for target prioritization. Enzymes at metabolic branch points with redundant outputs are less vulnerable than those in non-bypassable pathways (e.g., chitin synthase, trehalase). Prospective application of computational flux balance analysis and constraint-based modeling, as supported by existing literature, can systematically identify such non-compensatable nodes for effective target selection [47].

Study limitations and future directions

We acknowledge several limitations that point to productive future investigations. First, the transient nature of RNAi-mediated gene silencing in Lepidoptera precluded assessment of long-term or stage-specific consequences of GP loss; its role in adult flight or reproduction warrants investigation via CRISPR/Cas9 knockout or advanced dsRNA delivery systems [48-50]. Second, the generalizability of the this gluconeogenic compensation mechanism across insect orders (e.g., Coleoptera, Diptera) remains to be tested, as metabolic regulation can vary with ecology and life history. Third, while our multi-layered evidence—from transcriptional upregulation to phenotypic rescue—provides robust qualitative support for the activation of gluconeogenesis, future studies employing ¹³C-isotope tracer analysis would be valuable to quantify the precise metabolic rates identify primary carbon precursors, and further refine potential metabolic targets for insecticide development [51].

Concluding remarks

In conclusion, our study definitively rules out GP as the target of BPU insecticides, resolving a long-standing controversy. More importantly, it uncovers a fundamental principle of insect metabolic biology: robust compensatory pathways, specifically gluconeogenesis, can buffer against the loss of key catabolic enzymes. This metabolic plasticity renders GP non-viable as a standalone insecticidal target but illuminates a strategic path forward: the development of dual-target inhibitors that simultaneously block glycogenolysis and gluconeogenesis. Our findings underscore the necessity of adopting a systems-level perspective in target-based insecticide discovery. Biochemical potency in vitro is necessary but not sufficient for in vivo efficacy; the latter requires consideration of metabolic network architecture, pathway redundancy, and compensatory mechanisms. As the pest management community confronts the twin challenges of insecticide resistance and environmental sustainability, this mechanistic understanding of metabolic robustness will be essential for designing the next generation of selective, effective and durable insect control agents.

Materials and Methods

Insects and rearing conditions

A susceptible laboratory strain of P. xylostella (L.), maintained for multiple generations at the College of Plant Protection, Northwest A&F University, without prior insecticide exposure, was used for all experiments. Larvae were reared on radish (Raphanus sativus) seedlings grown in vermiculite. Adults were provisioned with a 10% (w/v) honey solution. All insect stages were maintained in a controlled environment at 25 ± 1°C, 60 ± 5% relative humidity (RH), and under a 16:8 h (light:dark) photoperiod.

Cell culture and maintenance

Spodoptera frugiperda (Sf9) cells (Gibco, CA, USA) were cultured in Sf-900™ III SFM medium (Gibco) supplemented with 10% fetal bovine serum (FBS; Gibco) and 1% (v/v) Antibiotic-Antimycotic solution (Gibco, Cat# 15240062). Adherent monolayer cultures were maintained in T-flasks at 27°C. For large-scale protein expression, cells were adapted to suspension culture in baffled flasks, agitated at 120 rpm at 27°C. Cells were harvested by centrifugation (4,000 rpm, 10 min, 4°C) when the density reached 5.0 × 10⁶ cells/mL. Cell pellets were flash-frozen in liquid nitrogen and stored at -80°C until use.

Chemicals and reagents

Diflubenzuron (DFB, 1-(4-chlorophenyl)-3-(2,6-difluorobenzoyl) urea, ≥98% purity) was purchased from Shanghai Topscience Co., Ltd. (Shanghai, China, Cat# T0109).

The Glycogen phosphorylase-IN-1 (GPI), 1-(3-(3-(2-Chloro-4,5-difluorobenzoyl)ureido)-4-methoxyphenyl)-3-methylurea (also known as compound 42 in ref [11]), was purchased from TargetMol (Shanghai, China, Cat# T37577, ≥98% purity).

Glycogen (from bovine liver, Type III), β-glycerophosphate, HEPES, imidazole, EDTA, DTT, ATP, MgCl₂, protease inhibitor cocktail components (AEBSF, Pepstatin A, Leupeptin), and all other biochemical reagents were purchased from Sigma-Aldrich (Shanghai, China) or Sangon Biotech (Shanghai, China) and were of molecular biology grade (≥95% purity) unless otherwise specified.

Recombinant PxGP expression and purification

The full-length coding sequence of PxGP (2,511 bp, encoding 837 amino acids; GenBank Accession No. XM_038119125.2 was amplified from P. xylostella third-instar larval cDNA by PCR using gene-specific primers (Table S1), and cloned into the pFastBac1 vector (Invitrogen, Carlsbad, CA, USA) using homology-based cloning (ClonExpress II One Step Cloning Kit, Vazyme, Nanjing, China). The construct was engineered to express PxGP with an N-terminal 6×His tag (HHHHHH-) immediately following the initiating methionine, under the control of the polyhedrin (PH) promoter. The sequence fidelity of the construct was verified by Sanger sequencing (Tsingke Biotechnology, Beijing, China).

The sequence-verified pFastBac1-PxGP construct was transformed into DH10Bac™ competent cells (Invitrogen) to generate recombinant bacmid DNA via Tn7-mediated transposition, following the manufacturer’s Bac-to-Bac® protocol. Recombinant bacmid DNA was isolated, verified by PCR using M13 forward and reverse primers (Table S1), and transfected into Sf9 cells using Cellfectin® II Reagent (Invitrogen) to generate P0 viral stock. The viral titer was amplified through two additional passages (P1, P2) to obtain high-titer P3 viral stock.

For large-scale protein production, Sf9 cells in suspension culture (2.0 × 10⁶ cells/mL, 500 mL culture volume in a 2-L baffled Erlenmeyer flask) were infected with P3 viral stock at a Multiplicity of Infection (MOI) of 5. Infected cultures were maintained at 27°C with agitation at 120 rpm. Cells were harvested 72 h post-infection by centrifugation (4,000 rpm, 10 min, 4°C), yielding approximately 8.25 g of cell pellet. Cell pellets were flash-frozen in liquid nitrogen and stored at -80°C until use.

For protein purification, frozen cell pellets were thawed on ice and resuspended in ice-cold Lysis Buffer (50 mM β-glycerophosphate [pH 7.5], 150 mM NaCl, 0.2 mM DTT, 0.5 mM EDTA, and 1× protease inhibitor cocktail [1 mM AEBSF, 15 µM Pepstatin A, 20 µM Leupeptin]) at a ratio of 4 mL buffer per gram of cell pellet. Cells were lysed by sonication on ice using a Scientz-IID sonicator (Ningbo Scientz Biotechnology, China) with the following parameters: 6 cycles of 30 s ON / 60 s OFF at 40% amplitude. The lysate was clarified by centrifugation (16,000 × g, 30 min, 4°C).

The clarified supernatant was loaded onto a 5-mL HisTrap™ HP column (GE Healthcare, Uppsala, Sweden) pre-equilibrated with Lysis Buffer containing 10 mM imidazole. The column was first washed with 10 column volumes (CV) of Equilibration Buffer (Lysis Buffer + 10 mM imidazole), followed by 5 CV of Wash Buffer (Lysis Buffer + 30 mM imidazole). Bound 6×His-tagged PxGP protein was eluted with a 15 CV linear gradient from 30 to 200 mM imidazole in Lysis Buffer. Fractions (2 mL each) were collected and analyzed by 13% SDS-PAGE with Coomassie Brilliant Blue R-250 staining. Fractions containing PxGP, identified as a prominent band at ∼100 kDa, were pooled and dialyzed overnight at 4°C against Storage Buffer (50 mM HEPES [pH 7.4], 150 mM NaCl, 1 mM DTT, 10% [v/v] glycerol) using a 10 kDa molecular weight cut-off dialysis membrane (Spectrum Labs, Rancho Dominguez, CA, USA), with three buffer changes. Protein concentration was determined using the Pierce™ BCA Protein Assay Kit (Thermo Fisher Scientific, Waltham, MA, USA) with bovine serum albumin (BSA) as the standard. Aliquots of purified PxGP were flash-frozen in liquid nitrogen and stored at -80°C. The typical yield was 7.5 mg of purified protein from 8.25 g of cell pellet.

In vitro activation of recombinant PxGP

Purified recombinant PxGP exists predominantly in the inactive dephosphorylated b form (PxGP-b) and requires phosphorylation at a conserved serine residue (corresponding to Ser14 in mammalian GP) for conversion to the active a form (PxGP-a) [19]. Activation was performed in vitro using immobilized Phosphorylase Kinase (PhK) (Cat# HY-P2757, MedChemExpress, Monmouth Junction, NJ, USA).

Immobilized PhK beads (100 µL slurry) were washed three times with 1 mL of Kinase Buffer (25 mM β-glycerophosphate, pH 7.5, 0.5 mM EDTA, 0.25 mM DTT, 1× protease inhibitor cocktail) and resuspended in 500 µL of dialyzed PxGP-b solution (final protein concentration: 1 mg/mL). The activation reaction was initiated by adding ATP (sodium salt) to a final concentration of 7 mM and MgCl₂ to 22 mM. The suspension was incubated for 2 h at room temperature (22-25°C) with gentle end-over-end rotation.

Following incubation, the activated PxGP-a (present in the supernatant) was separated from the immobilized PhK beads using a magnetic separation stand. The supernatant was passed through a 0.22 µm syringe filter (Millipore) to remove any residual beads and was used immediately for enzymatic assays. The extent of activation was assessed by measuring specific activity (see Section 2.7). Activated PxGP-a was used fresh and not stored, as phosphorylated GP is subject to dephosphorylation by endogenous phosphatases over time.

Western blot analysis

Protein samples were separated by 13% SDS-PAGE (120 V, 90 min) and electrophoretically transferred to a nitrocellulose (NC) membrane (0.45 µm pore size, Millipore, Burlington, MA, USA) at 300 mA for 90 min at 4°C using a Bio-Rad Mini Trans-Blot® system. The membrane was blocked for 1 h at room temperature in 5% (w/v) non-fat dry milk in TBST (10 mM Tris-HCl, pH 7.4, 150 mM NaCl, 0.1% [v/v] Tween-20).

The membrane was incubated overnight at 4°C with an anti-His mouse monoclonal antibody (1:1500 dilution in blocking buffer; Cat# ABT2050, Abbkine, Wuhan, China). After three washes with TBST (10 min each), the membrane was incubated for 1 h at room temperature with an alkaline phosphatase (AP)-conjugated goat anti-mouse IgG secondary antibody (1:4000 dilution in blocking buffer; Cat# A0208, Beyotime Biotechnology, Shanghai, China). Following three additional TBST washes, immunoreactive bands were visualized using BM Purple AP Substrate (Roche, Basel, Switzerland) according to the manufacturer’s instructions. Images were captured using a ChemiDoc™ XRS+ imaging system (Bio-Rad, Hercules, CA, USA).

Glycogen phosphorylase activity assay

GP enzymatic activity was measured using a coupled-enzyme spectrophotometric assay kit (Cat# BC3345, Solarbio Life Science, Beijing, China). This assay quantifies GP activity by measuring the GP-catalyzed phosphorolysis of glycogen to glucose-1-phosphate (G-1-P). The product G-1-P is enzymatically converted to glucose-6-phosphate (G-6-P) by phosphoglucomutase (PGM), and G-6-P is subsequently oxidized by glucose-6-phosphate dehydrogenase (G6PDH), which stoichiometrically reduces NADP⁺ to NADPH. The rate of NADPH formation, which is directly proportional to GP activity, was monitored by measuring the increase in absorbance at 340 nm (A₃₄₀) using a Infinite 200 PRO microplate reader (Tecan, Männedorf, Switzerland).

The reaction mixture (200 µL total volume per well in a 96-well UV-transparent microplate) contained the following components (as supplied in the kit): 160 µL of Reaction Buffer (containing glycogen [2 mg/mL final], inorganic phosphate [10 mM], AMP [1 mM], NADP⁺ [0.4 mM], PGM, G6PDH, MgCl₂, and appropriate buffer), 30 µL of deionized water, and 10 µL of enzyme sample (purified PxGP-a or crude larval lysate). The reaction was initiated by adding the enzyme sample, and A₃₄₀ was recorded continuously at 15 s intervals for 30 min at 25°C.

Specific activity was calculated using the following formula:

where:

ΔA₃₄₀ represents the change in absorbance at 340 nm,

ε = molar extinction coefficient of NADPH at 340 nm (6.22 mM⁻¹ · cm⁻¹),

l = light path length (0.6 cm for 200 µL in a 96-well microplate),

[Protein] = protein concentration in the assay (mg/mL),

Vt = total reaction volume,

Vs = sample volume added.

One Unit (U) of GP activity is defined as the amount of enzyme that catalyzes the formation of 1 µmol of NADPH per minute under the assay conditions.

Inhibitor screening and IC₅₀ determination

For in vitro enzyme inhibition assays, purified activated PxGP-a was diluted to 10 µg/mL in Assay Dilution Buffer (30 mM HEPES, pH 7.2, 60 mM KCl, 1.5 mM EDTA, 1.5 mM MgCl₂. Test compounds (GPI, DFB) were prepared as 100× stock solutions in DMSO and serially diluted in Assay Dilution Buffer to achieve final assay concentrations ranging from 0.16 nM to 500 nM (for GPI) and 0.54 nM to 1700 nM (for DFB), respectively (final DMSO concentration ≤ 1% [v/v] in all wells, including vehicle controls).

The enzyme (10 µL of 10 µg/mL PxGP-a, final concentration 10 µg/mL in the assay) was pre-incubated with 10 µL of inhibitor solution (or vehicle control) for 15 min at room temperature in a 96-well microplate. The reaction was then initiated by adding 180 µL of the Reaction Buffer (as described in Section 2.7), and A₃₄₀ was monitored continuously for 30 min at 25°C.

Residual enzyme activity at each inhibitor concentration was calculated as a percentage of the vehicle control (1% DMSO alone). Dose-response curves were generated by plotting percent residual activity versus log[inhibitor concentration]. IC₅₀ values (the concentration of inhibitor required to reduce enzyme activity by 50%), Hill slopes, and 95% confidence intervals (CI) were calculated by non-linear regression analysis using a four-parameter logistic equation in GraphPad Prism 9.0 (GraphPad Software Inc., San Diego, CA, USA):

where Y is percent residual activity, X is log[inhibitor concentration], Top and Bottom are the upper and lower plateaus of the curve (constrained to 100% and 0%, respectively), and Hill slope describes the steepness of the curve.

For each compound, at least three independent experiments were performed, each with triplicate technical replicates.

Preparation of larval crude lysate for in vivo activity

To assess the inhibitory effects of GPI and DFB on native GP activity in vivo, crude enzyme extracts were prepared from P. xylostella larvae. Healthy third-instar (L3) larvae (n = 30 per biological replicate, three replicates total) were collected, briefly washed in ice-cold phosphate-buffered saline (PBS, pH 7.4) to remove surface contaminants, and gently blotted dry on filter paper. Larvae were then homogenized on ice in 3 volumes (w/v) of ice-cold Homogenization Buffer (30 mM HEPES, pH 7.2, 60 mM KCl, 1.5 mM EDTA, 1.5 mM MgCl₂, 1× protease inhibitor cocktail).

The homogenate was centrifuged at 15,000 × g for 20 min at 4°C, and the supernatant (crude enzyme extract) was carefully collected, avoiding the lipid layer and pellet. Protein concentration was determined using the BCA assay, and extracts were either used immediately for activity assays or aliquoted and stored at -80°C (activity remained >90% for up to 2 weeks).

GP activity in crude lysates was measured using the same coupled-enzyme assay described in Section 2.7, with the following modifications: crude lysate (diluted to 2 mg/mL total protein) was used in place of purified enzyme, and pre-incubation with inhibitors was performed for 5 to 45 min to assess time-dependent inhibition. Final inhibitor concentrations tested were 10⁻⁵, 10⁻⁶ or 10⁻⁷ M.

Sequence alignment and bioinformatic analysis

The amino acid sequence of PxGP (XP_037975053.2) was obtained from the NCBI database. Multiple sequence alignment with orthologs from Homo sapiens (NP_002854.3), Oryctolagus cuniculus (NP_001075653), Helicoverpa armigera (XP_021190210.1), Spodoptera exigua (ACN78408.1), and Ostrinia furnacalis (AFO54708.2) was performed using ClustalX. Conserved domains and motifs were analyzed using the ScanProsite and NCBI CD-Search web servers. Protein pairwise similarity matrix was visualized as a heatmap using the “HeatMap” module in TBtools-Ⅱ (v2.390) [52].

Larval bioassays

Larval bioassays were conducted using a standard leaf-dip method adapted from established protocols [53]. Stock solutions of GPI and DFB were prepared in HPLC-grade acetone at 10,000 mg/L. Serial dilutions were prepared in deionized water containing 0.1% (v/v) Triton X-100 as a wetting agent to achieve final test concentrations of 250 and 500 mg/L for GPI, and 125 and 250 mg/L for DFB. The final acetone concentration in all treatment solutions was 1% (v/v). A solvent control group (CK) consisted of 1% acetone and 0.1% Triton X-100 in water without any test compound.

Fresh cabbage (Brassica oleracea var. capitata) leaf discs (5 cm diameter) were cut using a cork borer, dipped in the respective treatment solutions for 6-8 s with gentle agitation, and then air-dried for 30 min at room temperature in a fume hood. Once dry, each leaf disc was placed in a sterile 90 mm petri dish lined with moistened filter paper.

Ten synchronized third-instar larvae of P. xylostella (24-36 h post-molt to L3) were transferred to each treated leaf disc using a soft brush. Each treatment consisted of three biological replicates (i.e., three Petri dishes, 30 larvae total per treatment). Petri dishes were sealed with Parafilm® and maintained under standard rearing conditions (25 ± 1°C, 60 ± 5% RH, 16:8 h L:D photoperiod). Leaf discs were replaced with freshly treated discs every 12 h to ensure continuous exposure.

Mortality was recorded every 24 h for 120 h (5 days). Larvae that failed to respond to gentle prodding or showed moribund behavior (inability to right themselves or move coordinately) were considered dead. Corrected mortality was calculated using Abbott’s formula [54]:

Pupation rates (number of pupae formed / initial number of larvae × 100) and adult emergence rates (number of emerged adults / number of pupae formed × 100) were recorded at 8 and 10 days post-treatment, respectively.

Lethal concentration values (LC₃₀ and LC₅₀) and 95% confidence intervals for DFB were calculated using probit analysis in POLO-Plus software (LeOra Software, Berkeley, CA, USA).

GPI exposure for gene expression analysis

For time-course gene expression analysis, synchronized third-instar larvae were exposed to 500 mg/L GPI (the highest concentration tested in bioassays) or solvent control (1% acetone + 0.1% Triton X-100) using the leaf-dip method described in Section 2.11.

Surviving larvae were collected at 24, 48, 72, and 96 h post-treatment. At each time point, three biological replicates (10 pooled larvae each) were flash-frozen in liquid nitrogen and stored at -80°C until RNA extraction. This sampling scheme ensured sufficient RNA yield while minimizing inter-individual variability.

Quantitative real-time PCR (RT-qPCR) analysis

Total RNA was extracted from frozen samples using RNAiso Plus reagent (TaKaRa, Kusatsu, Japan). RNA concentration and purity were assessed by NanoDrop 2000c spectrophotometer (Thermo Fisher Scientific). Only samples with A₂₆₀/A₂₈₀ ratios between 1.9 and 2.1 and A₂₆₀/A₂₃₀ ratios >1.8 were used for downstream analysis. RNA integrity was confirmed by agarose gel electrophoresis (sharp 28S/18S rRNA bands, ratio ≈ 2:1). cDNA was synthesized from 1 µg RNA using PrimeScript™ RT Reagent Kit with gDNA Eraser (TaKaRa).

Gene-specific primers for PxGP, trehalase (PxTre), hexokinase (PxHex), phosphoenolpyruvate carboxykinase (PxPEPCK), PxG-6-Pase, and the reference gene PxRPS13 (ribosomal protein S13) were designed using Primer-BLAST (NCBI) and are listed in Table S1. PxRPS13 was selected based on its stable expression across developmental stages and treatment conditions (geNorm M value < 0.5) [55]. All primers were validated for specificity (single peak in melt-curve), efficiency (90-110%, R² > 0.98), and absence of primer-dimers (no-template controls).

RT-qPCR reactions were performed on a QuantStudio™ 5 Real-Time PCR System (Thermo Fisher Scientific) using TB Green® Premix Ex Taq™ II (Tli RNaseH Plus) (2×) (TaKaRa) in 20 µL reactions. The thermal cycling conditions were as follows: initial denaturation at 95°C for 30 s; 40 cycles of 95°C for 15 s (denaturation) and 60°C for 30 s (annealing/extension); followed by melt-curve analysis. Each sample was analyzed in technical triplicate, and each biological replicate consisted of three independent samples. Relative gene expression was calculated using the 2^-ΔΔCt method [56].

RNA interference (RNAi)

A conserved 606 bp region of PxGP (nucleotides 66-671, part of the AMP-binding domain) was selected for RNAi. The fragment was amplified from larval cDNA using primers flanked by T7 promoters (Table S2). Double-stranded RNA (dsRNA) was synthesized with the T7 RiboMax™ Express RNAi System (Promega, Madison, WI, USA). The dsRNA of green fluorescent protein (dsGFP, 541 bp) was synthesized using a pGFP plasmid template as a non-specific negative control. Following synthesis, dsRNA was treated with RNase-free DNase I (30 min at 37°C), purified by isopropanol precipitation, and integrity verified by 1.5% agarose gel electrophoresis. Concentration and purity (A₂₆₀/A₂₈₀=1.9-2.1) was measured using a NanoDrop spectrophotometer. Purified dsRNA was resuspended in nuclease-free water at concentrations of 6,000, 10,000, or 14,000 ng/µL and stored in single-use aliquots at -80°C.

For microinjection, synchronized early third-instar larvae (6-12 h post-molt) were briefly chilled on ice for 2-3 min and injected laterally using a DMP-300 Injector (Micrology Precision Instruments, Ltd, Wuhan, China). After recovery on fresh cabbage leaves for 30 min at room temperature, larvae were maintained under standard rearing conditions. Phenotypic parameters (mortality, molting success, pupation rate, adult emergence) were monitored daily for 10 days. Larvae were collected at 24, 48, 72, and 96 h post-injection for RT-qPCR analysis to confirm target gene knockdown (n = 3 biological replicates, each containing 10 pooled larvae per time point).

Metabolite quantification

To directly assess metabolic flux alterations following PxGP knockdown, we quantified four key metabolites in the glycogen-glucose metabolic axis using commercial enzymatic assay kits: glycogen and glucose (Glycogen Assay Kit, Grace Biotechnology, Suzhou, China, Cat# G0590W96), trehalose (Trehalose Assay Kit, Grace Biotechnology, Cat# G0553W96), and glucose-6-phosphate (G6P Assay Kit, Beyotime Biotechnology, Shanghai, China, Cat# S0185). All kits utilize colorimetric or fluorometric enzymatic reactions and were performed according to the manufacturers’ protocols with minor modifications as described below.

Third-instar larvae injected with dsGP (14,000 ng/µL, the highest dose achieving maximal knockdown) or dsGFP control were collected at 48, 72, and 96 h post-injection. At each time point, three biological replicates were collected, each consisting of 10 pooled larvae. Samples were immediately flash-frozen in liquid nitrogen and stored at -80°C until analysis.

For metabolite extraction, frozen larvae were homogenized on ice in ice-cold extraction buffer specific to each metabolite assay (typically 100 µL buffer per 10 mg tissue). For glycogen extraction, samples were homogenized in deionized water and immediately boiled for 3 min to inactivate endogenous glycogenolytic enzymes. For trehalose, glucose, and G6P extraction, samples were homogenized in the respective assay buffers provided in the kits. Homogenates were centrifuged at 12,000 × g for 10 min at 4°C, and supernatants were immediately used for metabolite quantification or stored at -80°C for no more than 24 h.

Metabolite quantification was performed in 96-well microplates using Infinite 200 PRO microplate reader (Tecan, Männedorf, Switzerland) at the wavelengths specified in each kit’s protocol. Sample metabolite concentrations were calculated by comparison with the absorbance of the kit’s standard. Metabolite concentrations were normalized to tissue protein content to account for variation in tissue input. Protein concentration in the crude lysates was determined using the Protein Quantification Kit (BCA Assay) (Abbkine Scientific) with bovine serum albumin as the standard. All samples were analyzed in technical triplicate. Final metabolite levels were expressed as: Glycogen: mg/mg protein; Glucose: µmol/mg protein; Trehalose: µmol/mg protein; G6P: µmo/mg protein.

Data are presented as mean ± standard error of the mean (SEM). Statistical significance between dsGP and dsGFP groups at different time point was assessed using two-way ANOVA (GraphPad Prism 8.0), with Sidak’s multiple comparisons test [57]. Statistical significance was set at P < 0.05 (*), P < 0.01 (**) or P < 0.001 (***).

Data availability

All data generated or analyzed during this study arewithin the paper and its Supporting information file; source data files have been provided for all figures.

Supplementary Figures and Tables

Figure S1.

Figure S2.

Table S1.

Table S2.

Additional information

Author contribution

Conceptualization: Zhen Tian, Yalin Zhang and Jiyuan Liu.

Data curation: Yifei Zhou, Yanqi Kang, and Yan Liu.

Formal analysis: Yifei Zhou.

Funding acquisition: Zhen Tian, Yalin Zhang and Jiyuan Liu.

Investigation: Yifei Zhou, Yanqi Kang, Yan Liu, Ruichi Li, Dongliang Wang, Chong Yi, Yifan Li, Yalin Zhang, Zhen Tian, and Jiyuan Liu.

Methodology: Yifei Zhou, Yanqi Kang, Yifan Li, and Jiyuan Liu.

Resources: Jiyuan Liu, Zhen Tian, and Yalin Zhang.

Supervision: Jiyuan Liu, Zhen Tian, and Yalin Zhang.

Validation: Yifei Zhou, Yanqi Kang, Yan Liu, and Jiyuan Liu.

Visualization: Yifei Zhou, Zhen Tian, and Jiyuan Liu.

Writing – original draft: Yifei Zhou.

Writing–review & editing: Yifei Zhou, Zhen Tian, and Jiyuan Liu.

Funding

This research was supported by the Guangdong Basic and Applied Basic Research Foundation (https://gdstc.gd.gov.cn/) grant 2024A1515010156 (to JYL), the National Natural Science Foundation of China (https://www.nsfc.gov.cn/) grants 32172452 (to JYL) and 32372626 (to ZT), the Shenzhen Basic Research Program (https://stic.sz.gov.cn/) grant JCYJ20240813152006009 (to JYL), and Chinese Universities Scientific Fund (https://www.nwafu.edu.cn/) grants 2452022348 (to YLZ), 2452021151 (to YLZ).

Funding

GDSTC | Basic and Applied Basic Research Foundation of Guangdong Province (2024A1515010156)

• Jiyuan Liu

National Natural Science Foundation of China (NSFC) (32172452)

• Jiyuan Liu

MOST | National Natural Science Foundation of China (NSFC) (32372626)

• Zhen Tian

Shenzhen Municipal Fundamental Research Program (JCYJ20240813152006009)

• Jiyuan Liu

MOE | Chinese Universities Scientific Fund (2452022348)

• Yalin Zhang

MOE | Chinese Universities Scientific Fund (2452021151)

• Yalin Zhang

Additional files

Supplementary Data S1. Contains values used for data presentation of Figures 2-4, 6-9, and S1.