Elucidating the molecular basis of ATP synthase inhibition: A virtual screening framework for next-gen selective acaricides

Abstract

The rational design of pesticides that selectively target harmful species while sparing beneficial organisms remains a major challenge in sustainable agriculture. Progress in this field has been limited by an incomplete understanding of the molecular basis of selectivity and the lack of effective strategies to exploit specific targets. Here, we provide direct molecular evidence that naturally occurring polymorphisms in the CV-a subunit of mitochondrial adenosine 5′-triphosphate (ATP) synthase are the primary determinants of differential sensitivity between pest mites and predatory mites. Structural modeling and sequence analysis identified the key functional residues (Leu¹⁸⁶ in Tetranychus cinnabarinus and Ala²⁰⁰ in Neoseiulus barkeri) that modulate the binding affinity of the inhibitor. Using these mechanistic insights, we developed a structure-based virtual screening pipeline to identify selective small-molecule inhibitors, including Vepdegestrant (ARV-471) and ICG-001, that exploit binding-site variation to achieve robust species specificity. This target-centric discovery framework not only advances the mechanistic understanding of ATP synthase inhibition but also establishes a scalable strategy for the rational design of environmentally safe and precision-oriented acaricides.

INTRODUCTION

Efficient metabolic energy production is fundamental to all forms of eukaryotic life, with oxidative phosphorylation serving as the dominant pathway for adenosine 5′-triphosphate (ATP) generation in most species (1–3). At the heart of this process is the enzyme ATP synthase (complex V), which harnesses the proton motive force driving the phosphorylation of adenosine diphosphate, thereby supplying the majority of the cell’s ATP pool (4, 5). Because of its essential role in maintaining cellular energetics, even subtle perturbations to ATP synthase can trigger a broad metabolic collapse, with lethal effects (6–8). These attributes render complex V not only a compelling biomedical target but also an attractive point of intervention for agricultural pest control, where its selective disruption would debilitate harmful species while sparing beneficial organisms (2, 6, 8). Cryo–electron microscopy (Cryo-EM) has recently illuminated ATP synthase structures and their regulation. A landmark 2019 study resolved mammalian ATP synthase tetramers bound to their endogenous inhibitor IF1 (ATPase inhibitory factor 1) at near-atomic resolution, offering unprecedented insight into the modulation of its activity (9). More recently, the structural dissection of the Tetrahymena thermophila mitochondrial respiratory supercomplex uncovered lineage-specific adaptations linked to the formation of tubular cristae (10). Despite such structural breakthroughs, a substantial knowledge gap remains for arthropods, particularly in agriculturally relevant mites, where the molecular foundations of complex V function and regulation remain poorly understood. Regarding pest management, species selectivity remains a critical hurdle: An ideal inhibitor must incapacitate the pest species while preserving nontarget, ecologically indispensable taxa such as pollinators and predatory arthropods (11, 12).

Mites (Arachnida) are a diverse and ecologically pivotal group in agroecosystems, encompassing herbivory, detritivory, and predation (13, 14). Phytophagous species such as Tetranychus urticae, Tetranychus cinnabarinus, and Tetranychus truncatus are among the most devastating agricultural pests worldwide, causing substantial crop losses through rapid reproduction and severe plant damage (15, 16). Unfortunately, broad-spectrum acaricides often fail to discriminate between pests and beneficial predatory mites, thereby destabilizing ecological networks (17, 18). This challenge has driven efforts to identify species-specific targets. Acaricides targeting the mitochondrial electron transport chain, particularly succinate dehydrogenase inhibitors that act on complex II and ubiquinol-cytochrome c reductase inhibitors that target complex III, have emerged as promising candidates because of their unique modes of action and substantial selectivity (19, 20). While these agents provide effective mite control and valuable molecular targets for pesticide innovation (21, 22), their extensive use has triggered rapid resistance evolution. Acaricides sharing similar modes of action, such as cyflumetofen, cyetpyrafen, and cyenopyrafen, now exhibit a severe target site and cross-resistance in field populations (20–22), threatening sustainable pest management. Overcoming resistance through distinct selective targets has thus become critical for next-generation acaricide development. Here, we address this gap by investigating propargite, a selective complex V inhibitor whose mechanism of selectivity has remained elusive (23, 24). By elucidating the molecular basis of propargite selectivity, we establish a virtual screening (VS) framework that enables efficient target exploitation. This work reveals the previously unknown mode of action of propargite and provides a strategic paradigm for finding selective acaricides.

In this study, we systematically elucidated the molecular basis of propargite’s selectivity by integrating target sequence bioinformatics, structural modeling, molecular docking, and functional characterization of key amino acid residues. Using the pivotal CV-a subunit as a bait protein for VS, we identified highly selective small-molecule inhibitors, including ARV-471 and ICG-001. ICG-001 demonstrated a selectivity mechanism distinct from that of propargite, offering the potential to substantially mitigate the risk of target-mediated resistance. Our findings not only provide molecular insights into the selectivity mechanisms underlying mitochondrial complex–targeted agents but also establish a robust screening paradigm for the efficient exploitation of these targets and the discovery of next-generation lead compounds for selective pesticides.

RESULTS

Toxicity of propargite toward phytophagous and predatory mites

The acaricidal selectivity of propargite was evaluated in bioassays with phytophagous mites (T. cinnabarinus and T. urticae) and predatory mites (Neoseiulus barkeri and Phytoseiulus persimilis), species with diverse genetic backgrounds (Fig. 1A). The bioassays showed that the toxicity of propargite varied between pest and predatory mites. The LC₅₀ (median lethal concentration) values for laboratory-maintained susceptible strains of T. cinnabarinus and T. urticae were 217.88 and 272.78 mg/liter, respectively (table S1). In contrast, predatory mites demonstrated exceptional tolerance, with <20% mortality even at the maximum test concentration of 50,000 mg/liter. Quantitative analysis revealed a 228.7-fold selectivity of propargite, indicating high efficacy against phytophagous pests while being safe for beneficial predatory species (Fig. 1B). This selectivity underscores propargite’s potential as a targeted acaricide and highlights the importance of further mechanistic studies.

Fig. 1. Toxicity bioassay of propargite against phytophagous and predatory mites.

(A) Schematic diagram of the bioassay via spraying and film methods. The blue color in the centrifuge tube indicates the formation of a uniform film. (B) Dose-response curves for the effects of propargite on predatory and phytophagous mites. The different colors of solid and dashed lines represent the responses of phytophagous mites and predatory mites, respectively. Created in BioRender. He, L. (2026) https://BioRender.com/i9pc2o4.

Assessment of epidermal penetration, detoxification, metabolic capacity, and selective toxicity

High-performance liquid chromatography (HPLC) revealed equivalent propargite levels in T. cinnabarinus (0.66 μg per mite) and the predatory mite N. barkeri (0.68 μg per mite) at 4, 6, 8, and 12 hours postexposure, with no temporal dependence (Fig. 2, A and B). Although T. cinnabarinus showed higher specific activities of cytochrome P450 monooxygenases, glutathione S-transferases (GSTs), and carboxyl/cholinesterases (CCEs) than N. barkeri (Fig. 2, C to E), pharmacological inhibition of each enzyme family failed to erode propargite’s selectivity; the acaricide remained lethal to the spider mites yet relatively harmless to the predator (Fig. 2, F to H). Thus, neither differential cuticular penetration nor disparity in detoxification metabolism was responsible for propargite’s safety to the predator mite.

Fig. 2. Relationship between cuticular penetration, detoxification metabolism, and the selective toxicity of propargite.

(A) Chromatogram from reversed-phase HPLC depicting propargite. (B) Levels of propargite in T. cinnabarinus and N. barkeri at various time points following exposure. h, hours. (C to E) Activities of CarEs, GSTs, and P450s in T. cinnabarinus and N. barkeri after treatment with a sublethal dose of propargite. (F to H) Mortality curves illustrating the effects of synergists (DEF, DEM, and PBO) on the activities of CCEs, GSTs, and P450s in T. cinnabarinus and N. barkeri. The bars represent the means ± SEM of three biological replicates. The asterisks indicate a significant difference by Student’s t test. *P < 0.05; **P < 0.01; ns, no significant difference.

Assessment of target enzyme inhibition and analysis of target biofunction

According to the Insecticide Resistance Action Committee classification criteria, propargite targets mitochondrial complex V (ATP synthase), the enzyme responsible for synthesizing ATP and providing cellular energy by using the electrochemical gradients generated from the electron transport chain (25). To explore the role of targets in mediating selectivity, we first evaluated propargite’s inhibitory effect on mitochondrial complex V by measuring the ATP levels in treated mites. The results showed that propargite (400 mg/liter) caused a persistent reduction in ATP levels in mites (T. cinnabarinus), with a marked drop of 60.78% in ATP levels after 7 hours. In contrast, propargite did not affect the ATP content of predatory mites (N. barkeri) (Fig. 3, A and B). In addition, we measured reactive oxygen species (ROS), another key bioindicator of respiratory chain effects that can trigger oxidative reactions, cause oxidative stress, and affect normal cellular functions. Fluorescent probe binding assays showed significant changes in ROS levels in T. cinnabarinus 6 to 8 hours after treatment with propargite, while the ROS levels in N. barkeri remained stable (Fig. 3, C and D, and fig. S1). Cryosection visualization indicated that ROS in T. cinnabarinus increased posttreatment, but there were no marked changes in ROS in N. barkeri (Fig. 3, F and G). The results suggest that propargite may modulate ATP and ROS levels by inhibiting mitochondrial complex V. Follow-up experiments were conducted to explore propargite’s impact on this molecular target. Mitochondria were isolated and subjected to in vitro activity assays, with complex V activity assessed by measuring the rate of increase in inorganic phosphate (P_i). The results demonstrated that the median inhibitory concentration (IC₅₀) for propargite’s inhibitory effect on N. barkeri (278.49 mM) mitochondrial complex V was 15814.31-fold higher than that on T. cinnabarinus (17.61 μM) (Fig. 3E). To provide in vivo functional evidence at the cellular level, we performed transmission electron microscopy analysis of midgut epithelial cells from female adult mites treated with propargite (400 mg/liter, 24 hours) and solvent controls. In T. cinnabarinus, propargite treatment induced severe mitochondrial pathology, including crista disruption, matrix vacuolization, and pronounced mitochondrial swelling (Fig. 3, H and I, and fig. S2). In contrast, mitochondria from propargite-treated N. barkeri remained structurally intact and indistinguishable from controls (Fig. 3, J and K, and fig. S3). These differential ultrastructural changes are consistent with specific ATP synthase inhibition in susceptible mites and provide direct visual evidence supporting drug-target engagement.

Fig. 3. Inhibition of target enzymes and the effects on key biochemical indicators of propargite in T. cinnabarinus and N. barkeri.

(A and B) ATP content in T. cinnabarinus and N. barkeri at different times following treatment with propargite. (C and D) Quantitative analysis of the average fluorescence intensity using confocal laser imaging of T. cinnabarinus and N. barkeri treated with sublethal doses of propargite at various time points. (E) Dose-response inhibition curves of mitochondrial complex V activity in T. cinnabarinus and N. barkeri exposed to propargite. (F and G) Time-dependent fluorescence pattern of change in ROS intensity in T. cinnabarinus and N. barkeri after treatment with propargite. The number of the test: Tc-CK, 17 females; Tc-Pro, 20 females; Nb-CK, 19 females; Nb-Pro, 21 females. Mean fluorescence intensity (Mean) = Sum of the fluorescence intensity in this region (IntDen)/Area of the region (Area). (H and I) Mitochondrial ultrastructure of T. cinnabarinus before and after propargite treatment, respectively. (J and K) Mitochondrial ultrastructure of N. barkeri before and after propargite treatment, respectively. The number of the test: 15 individuals for T. cinnabarinus and 20 individuals for N. barkeri. Red arrows indicate mitochondrial damage. The magnification ratios were 15,000× (15K), 20,000× (20K), 25,000× (25K), and 40,000× (40K). The bars represent the means ± SEM of three biological replicates. The asterisks indicate a statistically significant difference by Student’s t test. *P < 0.05; **P < 0.01; ns, no significant difference.

Molecular docking and structural prediction reveal key pockets in propargite

Given the structural complexity of mitochondrial complex V, the variation among species, and the lack of research on arthropod ATP synthase structure, we used AlphaFold3 to predict the structures of ATP synthase in the pest mite (T. cinnabarinus) and the predatory mite (N. barkeri), obtaining predictive models suitable for subsequent experiments [pTM (predicted TM-score) values >0.5] (fig. S4). The results indicated that the ATP synthases of T. cinnabarinus and N. barkeri were F-type ATP synthases (Fig. 4, A and D). We performed subsequent molecular docking of propargite with the ATP synthase structures of T. cinnabarinus and N. barkeri using AutoDock Vina. Molecular docking showed that propargite binds to the ATP synthase by entering the binding pocket formed by the Fo subunits (Fig. 4, B and E). Further analysis of the docking results showed that propargite had a binding free energy of −7.9 to −6.3 kcal/mol with the Fo of T. cinnabarinus compared to a binding free energy of −7.0 to −4.8 kcal/mol with N. barkeri. Propargite forms more interactions with T. cinnabarinus’s Fo subunit, resulting in a more stable binding (Fig. 4, C and F). These findings indicate a marked discrepancy in toxicity between T. cinnabarinus and N. barkeri, likely attributable to disparities in their target binding capabilities.

Fig. 4. Structural prediction of ATP synthase in T. cinnabarinus and N. barkeri and molecular docking with propargite.

(A) Structure of FoF1-ATP synthase in T. cinnabarinus. Different colors represent subunits. OSCP, oligomycin sensitivity-conferring protein. (B) Surface structure of ATP synthase in T. cinnabarinus and the binding pocket of ligand propargite. (C) Interaction sites of propargite with ATP synthase in T. cinnabarinus. (D) FoF1-ATP synthase structure of N. barkeri. Different colors represent different subunits. (E) Surface structure of ATP synthase in N. barkeri and the binding pocket of ligand propargite. (F) Interaction sites of propargite with ATP synthase in N. barkeri.

Biofiducial analysis for targeting core functional subunits and key functional sites

After identifying propargite’s target as Fo, we further analyzed the a, b, and c core subunits of the target protein. Molecular docking showed that propargite had a stronger binding affinity to T. cinnabarinus’s CV-a subunit (−6.9 to −5.8 kcal/mol) than to N. barkeri’s CV-a subunit (−5.6 to −4.8 kcal/mol) (Fig. 5, A and B). The binding affinities of CV-b and CV-c subunits to propargite were moderate and similar between the two mite species. In particular, T. cinnabarinus’s CV-b and CV-c subunits had binding affinities of −5.1 to −4.1 kcal/mol and −5.0 to −4.1 kcal/mol, respectively, while N. barkeri’s CV-b and CV-c subunits had binding affinities of −4.9 to −3.3 kcal/mol and −5.0 to −4.0 kcal/mol, respectively (Fig. 5, C to F). These results indicate that among the three subunits, CV-a had the strongest binding affinity to propargite. In contrast, CV-b and CV-c had lower binding stability, with no significant differences between the two mite species.

Fig. 5. Molecular docking and bioinformatic analysis of propargite with the Fo subunits in T. cinnabarinus and N. barkeri.

(A and B) Overall structure and magnified binding sites of the ATP synthase a subunit from T. cinnabarinus and N. barkeri in complex with propargite. (C and D) Overall structure and magnified binding sites of the ATP synthase b subunit from T. cinnabarinus and N. barkeri in complex with propargite. (E and F) Overall structure and magnified binding sites of the ATP synthase c subunit from T. cinnabarinus and N. barkeri in complex with propargite. (G) Comparison of amino acid sequences of ATP synthase a subunit between the target and nontarget species. The red dashed box with asterisks indicates the key amino acid sites that specifically bind to propargite in both the target and nontarget species (phytophagous and predatory mites), with notable differences at the aligned sites Ile⁷³/Leu⁸⁵/, Ile⁸⁶/Leu⁹⁸, and Leu¹⁸⁶/Ala²⁰⁰.

We retrieved and compared the amino acid sequences of Fo subunits (a, b, and c) from T. cinnabarinus and N. barkeri with those of other species using the UniProt (www.uniprot.org/) and RCSB (www.rcsb.org/) databases. We analyzed the relevant amino acids by combining these results with the ligand interaction data from the molecular docking analysis. The CV-b and CV-c subunit sequences of the two mites were conserved compared to nontarget species, with no differences in amino acid binding sites for propargite (figs. S5 and S6). In comparing the CV-a subunits of the two mites, three key sites (Ile⁷³ in T. cinnabarinus versus Leu⁸⁵ in N. barkeri, Ile⁸⁶ versus Leu⁹⁸, and Leu¹⁸⁶ versus Ala²⁰⁰) were identified as key sites for propargite binding (Fig. 5G), and these sites were present in both the target and nontarget species. The results suggest that the significant difference in toxicity between harmful and predatory mites is likely due to differences in target binding.

Validating key amino acid action sites by microscale thermophoresis (MST) and isothermal titration calorimetry (ITC)

After identifying the key target subunits, we expressed and purified the CV-a, CV-b, and CV-c subunits from both mite species. Nb-CV-a from N. barkeri showed only trace expression in Escherichia coli, yielding ∼0.4 mg of purified protein, which appeared as a single band on SDS–polyacrylamide gel electrophoresis (SDS-PAGE) at the expected molecular weight of 24.5 kDa (fig. S7). By contrast, expression of Tc-CV-a from T. cinnabarinus failed in multiple heterologous systems, including E. coli BL21(DE3) and Rosetta (DE3), Pichia pastoris, human embryonic kidney (HEK) 293, and Spodoptera frugiperda Sf9 cells (fig. S8). Similarly, Nb-CV-a-mut from N. barkeri, in which key amino acids were substituted, could not be expressed. To further investigate the basis of these differential expression properties, we compared the amino acid motifs of Tc-CV-a and Nb-CV-a (fig. S9). Synthetic peptides of key amino acid segments were synthesized and analyzed. Each peptide was purified, yielding 4 mg with greater than 90% purity (table S2). Meanwhile, both mites’ CV-b and CV-c subunits were successfully expressed in E. coli. The Tc-CV-b subunit was expressed as inclusion bodies; after purification, 0.5 mg of protein was obtained for each subunit. The proteins migrated as single bands on SDS-PAGE gels, close to the predicted molecular weights (figs. S10 and S11). We assessed the binding affinity of propargite to the purified core subunits and synthesized peptides using microscale thermophoresis (MST), including those with key amino acid site mutations (Fig. 6, A and B). The wild-type peptides Tc-CV-a1, Tc-CV-a2, and Tc-CV-a3 bound propargite with affinities of 2.0, 58.9, and 55.2 μM, respectively. The mutated Tc-CV-a peptides (Tc-CV-a1-mut, Tc-CV-a2-mut, and Tc-CV-a3-mut) had reduced binding affinities of 7.2 μM, 176.9 μM, and 2.8 mM, respectively (Fig. 6, C to E), while the wild-type Nb-CV-a subunits showed no binding. Reverse mutations in Nb-CV-a peptides (Nb-CV-a1-mut, Nb-CV-a2-mut, and Nb-CV-a3-mut) resulted in binding affinities of 1.59 mM, 0.48 μM, and 110 μM, respectively, significant increases compared to the wild type (Fig. 6, F to H). Following the identification of the most critical amino acid sites, the screened key sites were simultaneously validated using molecular docking. The results demonstrated that point mutation (Leu¹⁸⁶/Ala²⁰⁰) significantly altered the binding affinity between T. cinnabarinus, N. barkeri, and propargite (fig. S12). Meanwhile, we performed isothermal titration calorimetry (ITC) to quantify the thermodynamic parameters governing this interaction. ITC analysis revealed that the Leu¹⁸⁶Ala (L186A) substitution profoundly attenuated propargite binding affinity in T. cinnabarinus. The wild-type protein exhibited a saturable binding isotherm with a dissociation constant (K_d) of 5.12 × 10⁻⁶ M, whereas the L186A mutant displayed a markedly compromised binding curve with a K_d of 7.91 × 10⁻⁶ M, representing a 1.54-fold reduction in affinity (Fig. 6, I and J). Conversely, the reciprocal Ala²⁰⁰Leu (A200L) substitutionin N. barkeri conferred robust propargite binding capability: The wild type showed no detectable binding, whereas the A200L mutant exhibited a saturable isotherm with a K_d of 5.61 × 10⁻⁵ M (Fig. 6, K and L). These results indicated that the Tc-CV-a subunits have a stronger binding affinity to propargite than the Nb-CV-a subunits. Three key amino acid sites influence propargite binding: Ile⁷³ in T. cinnabarinus versus Leu⁸⁵ in N. barkeri, Ile⁸⁶ versus Leu⁹⁸, and Leu¹⁸⁶ versus Ala²⁰⁰, with the Leu¹⁸⁶/Ala²⁰⁰ site showing the strongest effect.

Fig. 6. Validation of the interactions between Tc-CV-a (peptide), Nb-CV-a, Nb-CV-a (peptide), and propargite by MST and ITC.

(A and B) Peptide structural model of Tc-CV-a and Nb-CV-a and schematic diagram of point mutations at key amino acid sites, respectively. (C to E) Determination of the binding affinity to propargite before and after mutation. (F to H) Determination of the binding affinity to propargite of the peptide segments of the key amino acid sites (Leu⁸⁵, Leu⁹⁸, and Ala²⁰⁰ respectively) of N. barkeri before and after mutation. (I to L) Representative ITC binding isotherms for the interaction between propargite and CV-a peptide fragments: (I) wild-type T. cinnabarinus CV-a (Leu¹⁸⁶); (J) mutant T. cinnabarinus CV-a (Ala¹⁸⁶); (K) wild-type N. barkeri CV-a (Ala²⁰⁰), showing no binding; (L) mutant N. barkeri CV-a (Leu²⁰⁰). Binding parameters are indicated in each panel.

Screening of selective small-molecule inhibitors

VS for selective ATP synthase inhibitors

We advanced our investigation through an integrated, target-directed VS strategy. This approach enables the rational identification of small molecules with preferential binding affinities at precisely defined target sites, thereby establishing a robust framework for the development of next-generation selective acaricides. Initially, a diverse compound library comprising ∼50,000 molecules, including a curated subset of known bioactive inhibitors, was subjected to molecular docking against the T. cinnabarinus CV-a subunit, our validated selective target (Fig. 7A). A multitiered pipeline combining high-throughput VS with systematic refinement identified the top 100 candidates that exhibited favorable binding energies, predominantly ranging from −7 to −10 kcal/mol. The top 20 candidate compounds were then prioritized on the basis of predicted bioactivity scores and synthetic accessibility. Following a comprehensive evaluation of the binding differences of the molecules in question, in addition to the access routes, costs, and development potential, it was decided that ARV-471 (780757-88-2_log) and ICG-001 (134562533_log) would be selected for further biochemical and biophysical analysis (table S3).

Fig. 7. VS and functional validation of selective inhibitors.

(A) Schematic of the VS pipeline based on the target structure. (B) Survival curves for ARV-471 in different pest mite populations (T. cinnabarinus, T. urticae, P. citri, and Tetranychus truncates) and predatory mite populations (N. barkeri, P. persimilis, N. californicus, and A. swirskii). (C) Concentration-response inhibition curves of ARV-471 on the target enzymes of T. cinnabarinus and N. barkeri. (D) Survival curves for ICG-001 in different pest mite populations (T. cinnabarinus, T. urticae, P. citri, and T. truncates) and predatory mite populations (N. barkeri, P. persimilis, N. californicus, and A. swirskii). (E) Concentration-response inhibition curves of ICG-001 on the target enzymes of T. cinnabarinus and N. barkeri. (F and G) Results of molecular docking of ARV-471 with the CV-a subunits of T. cinnabarinus and N. barkeri. (H and I) Molecular docking of ICG-001 with the CV-a subunit of T. cinnabarinus and N. barkeri. (J) Alignment of key amino acid residues mediating the binding of ICG-001 to the target and nontarget species, with critical sites highlighted in red and marked with asterisks. (K and L) MST assays verifying the differences in binding affinity of ARV-471 in T. cinnabarinus and N. barkeri upon mutation of key amino acid residues (Leu¹⁸⁶/Ala²⁰⁰). (M and N) MST assays verifying the differences in binding affinity of ICG-001 with T. cinnabarinus and N. barkeri after mutation of key amino acid residue corresponding to Phe⁷⁷ in T. cinnabarinus and Pro⁸⁹ in N. barkeri. (O to R) Representative ITC binding isotherms for the interaction between ICG-001 and CV-a peptide fragments: (O) wild-type T. cinnabarinus CV-a (Phe⁷⁷); (P) mutant T. cinnabarinus CV-a (Pro⁷⁷); (Q) wild-type N. barkeri CV-a (Pro⁸⁹), showing no binding; (R) Mutant N. barkeri CV-a (Phe⁸⁹). Binding parameters are indicated in each panel. (S) Survival curves of propargite-, bifenazate-, fenpropathrin-, and acynonapyr-resistant T. cinnabarinus strains after treatment with ICG-001. Created in BioRender. He, L. (2026) https://BioRender.com/o3gzlac.

In vitro and in vivo validation of ARV-471 and ICG-001

Bioassays demonstrated that both ARV-471 and ICG-001 had potent toxic effects against pest mites, achieving ∼50% mortality within 72 hours, while having minimal adverse impacts on predatory mite species. Concentration-response inhibition assays further revealed IC₅₀ values of 143.96 μM for ARV-471 and 63.18 μM for ICG-001 against the target enzyme in pest mites, whereas both compounds exhibited negligible inhibitory activity toward the homologous enzyme in predatory mites (Fig. 7, B to E). We then conducted docking studies to determine the molecular basis of the selectivity. ARV-471 showed binding affinities ranging from −11.9 to −10.2 kcal/mol with the CV-a subunit of T. cinnabarinus, but it did not bind to the homologous subunit of N. barkeri. Similarly, ICG-001 demonstrated significantly higher binding affinities (−10.2 to −8.5 kcal/mol) with T. cinnabarinus CV-a compared to the N. barkeri subunit (−7.6 to −6.7 kcal/mol) (Fig. 7, F to I). These results confirmed the compounds’ selectivity at the molecular level.

Guided by our VS platform designed around the key residue pair Leu¹⁸⁶ in T. cinnabarinus and Ala²⁰⁰ in N. barkeri—previously demonstrated to be critical for propargite’s activity—we proceeded to experimentally validate the computationally selected compounds through MST assays. The Leu¹⁸⁶Ala (L186A) substitution in the pest mite ATP synthase from L to A triggered a pronounced loss of binding affinity for ARV-471, from 292.94 μM in the wild type to undetectable levels in the mutant variant (Fig. 7K and fig. S13, A and B). In the reciprocal experiment for the predatory mite, the Ala²⁰⁰Leu (A200L) substitution resulted in a marked gain of binding affinity, increasing from a nonbinding baseline value to ∼1.95 mM (Fig. 7L and fig. S13, C and D). In contrast, mutational analyses involving ICG-001 revealed an entirely different interaction profile. Alterations at the propargite-defined selectivity positions had minimal influence on ICG-001 binding in either mite species (fig. S14), indicating that this compound’s selective activity is not mediated via the same residue pair. This prompted further exploration into alternative structural determinants. An expanded binding-site analysis uncovered distinct clusters of critical residues associated with ICG-001 specificity, particularly within regions unique to the nontarget species that are highlighted and marked with asterisks in structural overlays (Fig. 7J). A point mutation molecular docking strategy (fig. S15) and mutational validation showed that substituting Phe⁷⁷ with Pro in the pest mite’s ATP synthase significantly decreased ICG-001 binding affinity, from 28.01 μM in the wild type to undetectable levels in the mutant variant (Fig. 7M and fig. S15, A and B), whereas the reciprocal substitution of Pro⁸⁹ with Phe in the predatory mite led to a nearly 10-fold increase in affinity (Fig. 7N and fig. S15, C and D). Furthermore, ITC experiments revealed that mutation of this critical site—corresponding to Phe⁷⁷ in T. cinnabarinus and Pro⁸⁹ in N. barkeri- significantly altered the binding affinity of ICG-001 in both mite species (Fig. 7, O to R), providing additional support for our conclusion that ICG-001 targets a distinct site from propargite. Subsequent bioassays showed that ICG-001 exhibited comparable efficacy against field-collected resistant populations of T. cinnabarinus, including propargite-resistant strains, as against susceptible strains (table S4 and Fig. 7S), suggesting that it enabled effective resistance management of propargite resistance. These findings indicated that the sites governing ICG-001’s selectivity are distinct from those associated with propargite, supporting the concept of divergent binding epitopes. Collectively, the results confirm that VS based on molecular targets can identify selective small-molecule inhibitors.

DISCUSSION

Achieving species-specific selectivity remains one of the most intractable challenges in agrochemical development (26). Conventional pesticide design has seldom incorporated intrinsic structural variation at the molecular level, often resulting in unintended toxicity to nontarget organisms and accelerating the evolution of resistance (27, 28). Here, we demonstrated that naturally occurring amino acid polymorphisms within the CV-a subunit of mitochondrial ATP synthase constitute the principal molecular determinants of selective toxicity between pest and predatory mites. These mechanistic insights enabled us to establish a target-guided VS framework that transforms the historically underexploited concept of molecular target diversity into a practical strategy for the discovery of selective pesticide candidates. This approach not only illuminates the structural underpinnings of selectivity but also provides a scalable template for the rational design of next-generation, environmentally sustainable agrochemicals.

To analyze the molecular determinants of propargite selectivity, we integrated comparative bioinformatics, molecular docking, VS, and functional mutagenesis. We first evaluated non-target factors, including cuticular penetration and metabolic detoxification. Isotopically labeled compounds provide a powerful means to directly quantify pesticide penetration, as shown in studies of penetration resistance in mosquitoes (29), but establishing radiotracer-based protocols for propargite in Tetranychidae would require substantial methodological development. To complement such radiotracer approaches, we therefore used validated HPLC-based penetration assays as an equally informative and feasible method (19, 20). The result showed that propargite penetration was quantitatively equivalent between N. barkeri and T. cinnabarinus despite the former having a more rigid dorsal cuticle (30–32). Moreover, T. cinnabarinus exhibited substantially higher basal activities of detoxification enzymes, consistent with previous acaricide studies (19, 20). These findings excluded nontarget physiological traits as the primary basis of selectivity, redirecting our attention to the target-site structure. Mitochondrial function inhibitors often cause marked structural damage, including matrix swelling, vacuolization, and crista disruption (33). Consistent with this pattern, propargite rapidly damaged mitochondria in T. cinnabarinus, producing these characteristic morphological changes, but had no obvious effect on the mitochondrial structure in N. barkeri. This difference in inhibitory capacity was further reflected in two key physiological indicators, ATP and ROS. Propargite treatment caused a rapid 40% decline in ATP levels and a concomitant increase in ROS production in T. cinnabarinus, whereas N. barkeri remained largely unaffected. Consistent with these cellular responses, in vitro enzyme inhibition assays revealed a 15,814.31-fold difference in propargite sensitivity between complex V preparations from the two species, identifying ATP synthase as the primary determinant of selectivity. ATP synthase presents exceptional technical challenges: Its membrane-embedded nature and extreme hydrophobicity render crystallization and cryo-EM reconstruction notoriously difficult (34–36). Compounding these constraints, no high-resolution experimental structures from Tetranychidae or related Acari species currently exist in the Protein Data Bank (37, 38). Under these circumstances, AlphaFold prediction represents the only viable avenue for structural investigation in our target organism. This approach has proven effective in analogous nonmodel systems, including Drosophila and Caenorhabditis ATP synthase complexes (39–42). Our study revealed that both species harbored canonical FoF1-type ATP synthases (Fig. 3). Molecular docking pinpointed the propargite-binding site as a hydrophobic pocket formed by the Fo subunits a, b, and c. Sequence alignment highlighted distinct conservation patterns: The pest mites had highly conserved sequence motifs, while predatory mites showed unique substitutions, particularly within the CV-a subunit (Fig. 5G). These sequence variants were predicted to modulate propargite binding affinity, thereby conferring selectivity at the molecular level. This conclusion aligns with analogous findings in other agrochemical systems, including fluopicolide’s interaction with fungal V-ATPase subunit a (43), pyrethroid selectivity conferred by minute differences in insect sodium channels (44), amitraz selectivity through divergence in octopamine receptors (45), the selective mechanisms of cyflumetofen and bifenazate (19, 20), and succinate dehydrogenase inhibitor resistance mediated by a single point mutation in T. urticae (46). Collectively, these studies demonstrate that the target-site structure is the principal determinant of pesticide selectivity.

Direct assessment of binding between a chemical agent and its target protein is an important strategy for validating target-based selectivity, but such analyses typically require a soluble target protein. In the present study, multiple heterologous expression systems failed to yield soluble Tc-CV-a, consistent with the well-known challenges of membrane protein expression (47). Therefore, we synthesized peptides corresponding to the key binding region to further analyze the binding capacity between the pesticide and its target. Although synthetic peptides cannot fully reproduce the tertiary structure, membrane environment, or local folding context of the native protein, this strategy represents a reliable means of circumventing technical barriers associated with membrane protein expression (48–50). Moreover, our objective was not to resolve global conformational dynamics or allosteric regulation of the entire Fo complex but rather to identify key residues involved in the direct interaction between propargite and the CV-a subunit. Because the predicted propargite-binding pocket is confined to a discrete region of Tc-CV-a, synthetic peptides spanning this region provide a reasonable surrogate for probing local residue-ligand interactions, which are likely to be largely preserved in such fragments (49). This approach is also consistent with established practices in structure-based drug discovery, in which peptide fragments corresponding to functional regions are used to validate binding events and identify hotspot residues (50). Thus, although peptide-based assays do not fully recapitulate the native membrane-protein context, they offer a practical system for interrogating residue-level interactions in the absence of soluble full-length Tc-CV-a. Results from site-directed mutagenesis, MST, and ITC further support the conclusion that residue variation within the target-binding site is a major determinant of propargite selectivity. In both biophysical assays, substitution of Leu¹⁸⁶ with alanine in T. cinnabarinus markedly weakened propargite binding, whereas the reciprocal substitution in N. barkeri introduced detectable affinity. MST showed the expected directional changes in binding strength, with L186A in T. cinnabarinus substantially reducing affinity and A200L in N. barkeri generating measurable binding (Fig. 6, E and H). ITC revealed the same pattern, with weakened saturable binding in the T. cinnabarinus mutant and a clear gain of binding in the reciprocal N. barkeri mutant (Fig. 6, I to L). The concordance between these independent assays indicates that the A/L polymorphism in CV-a is a key determinant of propargite recognition. More broadly, these results show that even a single amino acid substitution can reshape the local binding environment and alter inhibitor sensitivity between the two mite species. On the basis of our phylogenetic analysis of the CV-a subunit (fig. S16), significant evolutionary divergence was observed between pest mites (Tetranychidae) and predatory mites (Phytoseiidae), indicating substantial phylogenetic distance between these two groups. This divergence leads us to hypothesize that functional diversification at key amino acid residues in CV-a may reflect adaptive evolution in predatory mites, potentially enhancing their metabolic plasticity and environmental adaptability, which is consistent with previous studies on the evolution of resistance and selectivity mechanisms (46, 51).

Current insecticide and acaricide research and development remain largely driven by phenotypic screening coupled with empirical structural optimization. This traditional approach has long served as the cornerstone of agrochemical discovery, yielding numerous important and innovative pesticide scaffolds (52, 53). However, this route often suffers from low hit rates and limited structural diversity, and in many cases, the mode of action and molecular targets of candidate molecules remain unclear, requiring extensive effort and experimentation to resolve. For example, the insecticide flupyradifurone, launched in 2003, required nearly two decades before its molecular target, nicotinamidase (Naam), was identified (54), illustrating the mechanistic lag inherent to traditional screening. By contrast, target-based drug design, a mature framework in pharmaceutical research, is now being adapted to agrochemical discovery, where it serves to complement rather than replace phenotypic strategies. Target-based drug design has enabled the development of landmark therapeutics such as tyrosine kinase inhibitors and HIV reverse transcriptase inhibitors (55, 56). It provides a conceptual and methodological basis for the rational design of new insecticides and acaricides. Rapid advances in cryo-EM and x-ray crystallography have recently provided high-resolution structures for numerous pest-control targets, laying the groundwork for true “target-to-drug” design (57, 58). Structural dissection of chitin synthase has furnished templates for substrate-competitive inhibitor design (59), while detailed analysis of interactions between mitochondrial electron transport complex III has guided the creation of highly selective mitochondrial inhibitors such as YF23694 and informed rational optimization of succinate dehydrogenase modulators (60, 61). Collectively, these studies underscore that target-centered VS is emerging as a powerful strategy to accelerate acaricide lead identification while maintaining a balance between potency and selectivity.

In light of this, the present study focuses on the validated selective target CV-a as a starting point to apply a structure-based VS strategy for the discovery of novel inhibitors. VS has proven to be highly effective for the high-throughput identification of bioactive molecules, as illustrated by the identification of esculetin in Phytophthora-host interactions (62) and isoginkgetin as an ACLY (ATP citrate lyase) inhibitor in atherosclerosis models (63). Using the well-characterized ATP synthase subunit CV-a as a molecular target, we constructed a chemically diverse library of ∼50,000 small molecules, including synthetic inhibitors, natural products, and emerging drug candidates. Screening of this library yielded two candidates, ARV-471 and ICG-001. The sources, market supervision, and safety information of the two small-molecule compounds are detailed in table S5. Both compounds originate from small-molecule inhibitor libraries, and their bioactivities are primarily mediated through the modulation of protein-protein interactions. Given that protein-protein interaction interfaces often exhibit substantial interspecies divergence, such an approach holds promise for achieving pest insect specificity. Neither compound had been previously investigated in the context of pest management, highlighting exploratory avenues for agricultural applications (64, 65). Both compounds exhibited certain inhibitory activity against the ATP synthase of T. cinnabarinus, with IC₅₀ values of 143.96 and 63.18 μM. In contrast, their activity against N. barkeri was negligible. Nevertheless, while acknowledging that ARV-471 and ICG-001 exhibit potency below commercial acaricide standards, these compounds function as effective chemical probes to validate the druggability and tractability of CV-a as a molecular target, distinct from end-use agricultural products. This further demonstrates that inhibitors identified through this approach can selectively target ATP synthase. Molecular docking, MST, and site-directed mutagenesis further revealed the structural determinants of the activity. ARV-471 exhibited selectivity patterns similar to those of propargite, mediated by key targets Leu¹⁸⁶ and Ala²⁰⁰, whereas ICG-001 activity was governed by different residues (Phe⁷⁷ and Pro⁸⁹). The molecular binding site of ICG-001 is distinct from that of conventional mitochondrial complex V acaricides (e.g., propargite), suggesting that small-molecule compounds acting on the same target but at different binding sites may enable effective overcoming of the resistance to traditional acaricides (Fig. 7S). The results presented herein serve merely as a foundation for the validation and discovery of the target platform; it is anticipated that systematic medicinal chemistry studies (including structural modification and structure-activity relationship optimization) will be necessary to enhance the efficacy of the compounds to the level required by commercial acaricides. Accordingly, target-based VS strategies not only enable the identification of compounds acting at previously recognized binding sites but also facilitate the discovery of small-molecule inhibitors capable of interacting with novel sites. Such strategies, while enhancing the efficiency of target utilization, also present avenues for obtaining new compounds, thereby providing a broader framework for the development of innovative acaricides.

This study establishes a rational blueprint for acaricide development by leveraging species-specific variation within the ATP synthase CV-a subunit to achieve high selectivity (Fig. 8). By integrating structural modeling, target-based VS, and functional validation, we identified inhibitors capable of discriminating between pest and beneficial species. These findings not only enhance our mechanistic understanding of ATP synthase–mediated toxicity but also lay the groundwork for precision agrochemical strategies that incorporate selectivity mechanisms, resistance management, and environmental sustainability.

Fig. 8. Schematic diagram illustrating the targeting of ATP synthase for next-generation selective acaricide development.

This schematic diagram was created in BioRender: L. He (2026), https://biorender.com/c6x45qz.

MATERIALS AND METHODS

Mites

T. cinnabarinus-CQ and T. urticae-CQ were collected in 1998 from Beibei District, Chongqing, China. The T. cinnabarinus-YN strains were sourced from fields in Kunming, Yunnan, China, in 2015. All three strains were maintained on fresh cowpea (Vigna sinensis) leaves without exposure to pesticides. The T. truncatus strain was collected from kidney bean fields in Zhangye, Gansu, China, in 2020 and has since been maintained on cowpea, with no exposure to pesticides during this period. Panonychus citri was provided by Pan et al. (66). The N. barkeri-BJ, Amblyseius swirskii, Neoseiulus californicus, and P. persimilis were obtained in 2022 from Beijing Kuoye Tianyuan Biochemical Co., Ltd. (Beijing, China). The Tc-Pro–resistant strain was originally collected from Yuxi City, Yunnan Province, China (101.97°E, 23.62°N), and the Tc-Bif–resistant strain was collected from Chuxiong Yi Autonomous Prefecture, Yunnan Province, China (101.24°E, 25.60°N). The Tc-AcyR and Tc-FenR strains were derived from a subline of the YN-SS strain that was selected under laboratory conditions (32). The experimental environment was maintained at a constant temperature of 26 ± 1°C, with a photoperiod of 16:8 hours (light: dark) and a relative humidity of 50 to 70%. The predatory mites were reared under controlled conditions of 22°C, a photoperiod of 16:8 hours (light: dark), and 60% humidity.

Bioassay

Residual contact vial bioassay and spraying methods were used to conduct biological assays for the phytophagous and predatory mites, as the test solution precipitated and clogged the nozzle of the Potter spray tower at concentrations above 10,000 mg/liter. The experimental techniques were adapted from a study by Feng et al. (32) with several modifications. Propargite was dissolved in acetone, and then the test mites were introduced after a uniform film had formed in the centrifuge tube. The concentrations for predatory mite bioassays were set at 50,000, 25,000, 20000, 10,000, 5000, 2000, and 1000 mg/liter. The spray method used a Potter spray tower using concentrations of 25, 50, 100, 200, 300, and 400 mg/liter to determine the toxicity of acaricides against phytophagous mites. Three biological replicates were used for each treatment. After a 24-hour posttreatment period, the mite bodies were examined using a high-magnification microscope. The criterion for ascertaining mortality was defined as the lack of any response to delicate tactile stimulation using a brush or the manifestation of minimal tremors in fewer than three limbs.

Enzyme activity and synergism experiment

We used a pre-exposure method to investigate the potential role of differences in detoxification enzyme activity in selective toxicity. The procedure was adapted from the method described by Van Pottelberge (12) and was conducted as follows. Healthy female adults of T. cinnabarinus and N. barkeri, aged 3 to 5 days, were carefully selected and positioned on leaf discs. The three synergists diethyl maleate (DEM), S,S,S-tributylphosphorotrithioate (DEF), and piperonyl butoxide (PBO) were individually diluted to their designated concentrations with DEF and PBO at 2000 mg/liter and with DEM at 4000 mg/liter, the lowest doses previously shown to inhibit detoxification enzymes in mites (67, 68).

The pest and predatory mites were then subjected to pretreatment using a Potter spray tower at a pressure of 1 bar. Following a 24-hour treatment period, the mites were separately collected for detoxifying enzyme activity assays and transferred to pre-prepared insecticide films for bioassays. Each concentration was replicated three times in the bioassays. The control group was treated with a corresponding proportion of an acetone-Tween 80 solution (12). A BCA Protein Assay Kit (P0012BCA, Beyotime Biotechnology, Shanghai, China) was used to measure the total protein content of the samples using bovine serum albumin as the standard. The activities of mixed-function oxidase, GST, and CCEs were assessed following the reagent preparation and detection protocols described by Wen et al. (69).

Epidermal penetration experiment

HPLC was performed to detect propargite in the mite bodies. Adult female T. cinnabarinus and N. barkeri were placed on leaf discs, ∼5 mg per disc. A propargite solution (200 mg/liter) was sprayed onto the mites using a Potter spray tower at 1-bar pressure, applying 1 ml per leaf disc. The treated mites were kept under rearing conditions for 4, 6, 8, and 12 hours and then collected into centrifuge tubes. The mites were washed five times with acetone to remove residual pesticide, and the eluate was analyzed by HPLC to confirm the complete removal of cuticular residues. The washed mites were homogenized in 0.1 M tris-HCl buffer, vortexed with ethyl acetate for 2 min to extract propargite (repeated twice), and centrifuged at 4°C and 10,000 rpm for 10 min. The supernatant was evaporated to near dryness at 40°C, dissolved in 1 ml of acetonitrile, filtered through a 0.22-μm filter, and transferred to an HPLC vial for analysis. A propargite standard solution (400 mg/liter) was prepared by dissolving a precise amount of propargite in acetonitrile. This solution was serially diluted to concentrations of 200, 100, 50, 25, and 12.5 mg/liter. HPLC was used to measure the peak areas of propargite at these concentrations. A standard curve was generated by plotting the propargite concentration (x axis) against the peak area (y axis). The HPLC conditions were as follows: a C18 column (Kromasil, 4.6 mm by 250 mm, 5 μm) maintained at 35°C, with a mobile phase of 90% acetonitrile (A) and 10% ultrapure water (B). The flow rate was 0.8 ml/min, and the detection wavelength was 254 nm.

ATP determination

T. cinnabarinus and N. barkeri mites were exposed to a propargite solution (400 mg/liter) (serves to assess the toxic risk of the agent to natural enemies while simultaneously elucidating the physiological response mechanisms under high-concentration stress) and subsequently collected at 1, 3, 5, and 7 hours posttreatment. The ATP content was measured using an Enhanced ATP Assay Kit (S0027, Beyotime Biotechnology, Shanghai, China) according to the manufacturer’s instructions. The mites were lysed by adding 50 μl of lysis buffer per 5 mg of mite tissue, followed by homogenization using an electric grinder. After complete lysis, the samples were centrifuged at 12,000g for 5 min at 4°C, and the supernatants were collected for subsequent analyses. To eliminate errors arising from differences in protein content during sample preparation, the protein concentration in the samples was determined using a BCA protein assay kit. Each treatment group consisted of 5 mg of mites, with three biological replicates.

Determination and visualization of ROS

To investigate the effects of propargite on T. cinnabarinus and N. barkeri, each mite species was treated with a propargite solution (400 mg/liter) via spray application, with an acetone-treated group serving as the control. At 4, 6, 8, and 12 hours posttreatment, the surviving T. cinnabarinus mites were examined under a microscope and collected for subsequent experiments. Approximately 200 individuals of each mite species were collected separately, mixed with 1 ml of prechilled phosphate-buffered saline (PBS) buffer, and homogenized on ice using a handheld electric homogenizer. The homogenates were centrifuged at 10,000g for 10 min at 4°C. The supernatants were divided into two portions: one for protein content determination and the other for measurement of the ROS content. Dihydroethidium (DHE) (Yeasen Biotech Co., Ltd., Shanghai, China) was added to the supernatant to a final concentration of 30 μM and incubated in the dark at 37°C for 1 hour. The fluorescence intensity was measured using a multifunctional microplate reader, with excitation and emission wavelengths set at 488 to 535 nm and 610 nm, respectively. To visualize ROS accumulation in the mites after application of the acaricide, mites treated with propargite were stained with the ROS probe DHE and subjected to cryosectioning. The sections were observed and photographed under a confocal microscope, and the fluorescence intensity was quantified using ImageJ software.

Transmission electron microscopy detection

The transmission electron microscopy protocol used in this study was adapted from Feng et al. (32), with modifications. To ensure developmental consistency, synchronized 5-day-old female adult mites of T. cinnabarinus and N. barkeri were treated with propargite (400 mg/liter) via spray application, with 1% Tween-acetone solution serving as the solvent control. At 24 hours posttreatment, surviving mites were collected (30 individuals per tube, three tubes per treatment) and fixed in 4% glutaraldehyde for 36 hours. Samples were washed three times with 0.01 M PBS (pH 7.4), postfixed in 1% osmium tetroxide for 1 hour at room temperature, and dehydrated through a graded ethanol series (30 to 100%) followed by 100% acetone. Samples were embedded in Spurr resin and polymerized at 40°C (12 hours), 50°C (12 hours), 60°C (12 hours), and 70°C (24 hours). Ultrathin sections (70 nm) were prepared using an ultramicrotome (Leica, Germany), mounted on copper grids (200-mesh), stained with 2% uranyl acetate, and air dried. The mitochondrial ultrastructure was examined using a transmission electron microscope (Hitachi, Japan) operated at 80 kV.

Inhibitory effect of propargite on mitochondrial complex V

A 10 mM propargite stock solution was prepared in dimethyl sulfoxide (DMSO) and serially diluted to concentrations ranging from 5 mM to 1 nM. The solutions were used to assess the inhibitory effects of propargite on complex V activity across different species. Complex V activity was measured using a Mitochondrial Respiratory Chain Complex V/ATP Synthase Activity Assay Kit (Beijing Solarbio Science & Technology Co., Ltd.) that quantified the rate of P_i produced from the hydrolysis of ATP. For each assay, 100 mg of pest mites and predatory mites was separately homogenized in 100 μl of extraction solution using an ice bath homogenizer. The homogenates were centrifuged at 600g for 10 min at 4°C, and the supernatant was transferred to a fresh tube. Further centrifugation at 11,000g for 15 min at 4°C separated the mitochondrial pellet, which was then resuspended in 600 μl of reagent 1 and sonicated (200 W, 5 s on, 10 s off, 12 cycles) to determine enzyme activity and protein content. The protein concentration was quantified using a BCA Protein Assay Kit, and the absorbance was measured at 660 nm using a multifunctional microplate reader. Each 100 mg of mites represented one biological replicate, with each treatment being repeated three times.

Bioinformatic analysis and molecular docking

To elucidate the molecular basis of propargite’s interaction with mitochondrial complex V (ATP synthase), we used AlphaFold3 to model the FoF1 complexes of the pest and predatory mites. Structure reliability was assessed using pLDDT (predicted Local Distance Difference Test) and PAE (predicted aligned error) metrics, and the predicted structures were converted to PDB format for further analysis. For molecular docking, small-molecule ligands were prepared in PDB or MOL2 formats and docked using AutoDock Vina. The spatial boundaries for molecular docking calculations were rigorously defined to encompass the entire ligand-binding cavity of ATP synthase. For the T. cinnabarinus system, the rectangular search grid was centered at (x, y, z) = (−0.364, 9.0, 2.348) Å with axis-aligned dimensions of 122.0 Å (width) by 118.0 Å (height) by 124.0 Å (depth). The N. barkeri counterpart used a cubic search grid of 125.3 Å per dimension, centered at (−2.124, 8.712, −42.622) Å. These parameters were optimized to ensure that the grid volume exceeded the geometric extent of the binding site by a minimum margin of 10 Å in each direction, thereby precluding edge effects during conformational sampling. Binding modes and energies were evaluated from the resulting PDBQT files. Amino acid sequences of FoF1-ATPase core subunits were obtained from UniProt (www.uniprot.org/) and the RCSB Protein Data Bank (www.rcsb.org/). Multiple sequence alignments of the Fo component’s three core subunits (CV-a, CV-b, and CV-c) were conducted, including sequences from nontarget organisms (Lemur catta, Oryctolagus cuniculus, Rattus norvegicus, Drosophila simulans, and N. californicus) and target organisms (Tetranychus pueraricola, T. urticae, T. truncatus, and P. citri).

Protein expression and purification

The CV-a subunit from N. barkeri was cloned into the pET-28b vector and expressed in E. coli BL21 (DE3) cells at 16°C for 16 hours after induction with 1 mM isopropyl-β-d-thiogalactopyranoside. Cells were harvested by centrifugation at 10,000g for 1 min and lysed in buffer containing 50 mM tris (pH 8.0), 150 mM NaCl, 10% glycerol, 1 mM phenylmethylsulfonyl fluoride, and 1% n-dodecyl-β-d-maltoside using an ultrasonic processor for 20 min. The lysate was centrifuged, and the soluble fraction was collected for protein purification. The recombinant protein was then purified, dialyzed, and quantified. Protein purity and expression were verified by SDS-PAGE and Western blotting. Western blot analysis was performed using a 6×His Tag monoclonal antibody (HIS.H8; cat. no. MA1–21315) and a horseradish peroxidase–conjugated goat anti-mouse IgG (H + L) secondary antibody (cat. no. 31430), both from Thermo Fisher Scientific. Similarly, the CV-b and CV-c subunits from T. cinnabarinus and N. barkeri were cloned into the pET-28a(+)-SUMO and pET-32a(+) vectors, respectively, and expressed in Rosetta (DE3) cells at 22°C for 18 hours following induction with 0.5 mM isopropyl-β-d-thiogalactopyranoside. Cells were lysed in 0.01 M PBS (pH 6.9) containing 0.1 M phenylmethylsulfonyl fluoride and 10 mM dithiothreitol. Recombinant proteins were purified and desalted using Millipore ultrafiltration devices with 30- to 50-kDa molecular weight cutoffs. Protein concentrations were determined using bovine serum albumin as the standard (69). Primer sequences are listed in table S6.

For eukaryotic expression, the coding sequence of the CV-a subunit from T. cinnabarinus was cloned into the pcDNA-N1 vector for transient expression in HEK293 cells and into the pFastBac vector for baculovirus-mediated expression in Sf9 cells. All constructs were verified by Sanger sequencing. HEK293 cells were cultured in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum at 37°C under 5% CO₂ and transfected with the pcDNA-N1-CV-a construct using Lipo8000 Transfection Reagent (Beyotime Biotechnology) according to the manufacturer’s instructions. Cells were harvested 24 to 48 hours after transfection for subsequent analysis. Sf9 cells were maintained in serum-free medium at 27°C. Recombinant bacmid DNA was generated using the Bac-to-Bac system and transfected into Sf9 cells with Lipo8000 reagent to produce recombinant baculovirus. Viral stocks were further amplified and used to infect Sf9 cells for protein expression, and infected cells were collected 48 to 72 hours later. Cell morphology and transfection-associated signals were examined using an Axio Observer imaging system (Zeiss). Expression of CV-a in both HEK293 and Sf9 cells was further confirmed by Western blotting using a specific antibody against the CV-a subunit. Detailed methods for Tc-CV-a expression in P. pastoris have been described previously in a study concerning bifenazate (19).

Peptide synthesis and amino acid point mutation

We performed a motif analysis of the a-subunit amino acid sequence of T. cinnabarinus using the MEME Suite (http://meme-suite.org/index.html) and integrated these results with molecular docking and sequence alignment data. On the basis of these analyses, we commissioned GenScript Biotech Corporation in Nanjing to synthesize the required peptides. The peptides were designed with an FITC-Ahx tag at the N terminus to facilitate detection and were solubilized in analytical-grade DMSO. In addition, the a-subunit sequences of T. cinnabarinus and N. barkeri with specific point mutations were synthesized. Detailed information on the synthesized peptides is provided in table S2.

MST assay

The purified protein was labeled using the Monolith Protein His-Tag Labeling Kit RED Tris-NTA 2nd Generation (Notemp Technology Co., Ltd., Beijing). The labeling reaction was performed by mixing 4 μM purified protein with 50 nM dye and incubating the mixture at 37°C for 30 min. A stock solution of the acaricide propargite was prepared at a concentration of 2 mM in DMSO (final concentration of 5%). This stock solution was then serially diluted 1:1 with phosphate-buffered saline containing Tween 20 (PBST) to generate a total of 16 different concentrations, each with a volume of 10 μl. Equal volumes (10 μl) of the labeled protein were added to each concentration of the acaricide solution and mixed thoroughly. The resulting mixtures were aspirated into capillaries and subjected to MST analysis. The binding affinity was measured at 40% light-emitting diode/excitation power and medium MST power, with all other parameters set to the system defaults. The binding affinity of the protein to propargite was analyzed using MO Affinity Analysis software (version 2.3.0).

Isothermal titration calorimetry

Binding affinities between propargite and the CV-a subunit (wild type and 186A/L mutants) from T. cinnabarinus and N. barkeri were measured using a Nano-ITC calorimeter. Proteins were purified and dialyzed overnight at 4°C against ITC buffer [20 mM Hepes, pH 7.5; 150 mM NaCl; and 1 mM tris(2-carboxyethyl)phosphine (TCEP)]. Propargite was dissolved in DMSO and diluted with the same dialysis buffer to a final DMSO concentration of 2%. The DMSO concentration was matched between the protein and ligand solutions to minimize baseline drift. The sample cell was loaded with a 20 μM protein solution, and the syringe was filled with 200 μM propargite. Titrations were performed at 25°C with a stirring speed of 750 rpm. An initial injection of 0.5 μl was followed by 19 injections of 2.0 μl each, with 150-s intervals between injections. Control experiments were conducted by titrating propargite into the buffer to determine the heat of dilution, which was subtracted from the binding data. The resulting thermograms were analyzed using PEAQ-ITC analysis software with a single-site binding model. The fitted parameters included K_d, stoichiometry (n), entropy change (ΔS), and enthalpy change (ΔH). All experiments were performed in triplicate, and data are presented as the means ± SEM.

Virtual screening

To identify additional small molecules with species-selective activity and validate our target-centric screening strategy, we conducted a structure-based VS campaign. The sheer scale of existing commercial compound libraries poses a substantial challenge for rapid target-focused screening. To address this, we constructed a focused library of ∼50,000 compounds tailored to our research objectives, encompassing synthetic inhibitors, natural products and their derivatives, and clinically approved drugs. This collection was enriched for compounds targeting mitochondrial function, including propargite, a miticide that disrupts the mitochondrial electron transport chain and oxidative phosphorylation (23, 24), thereby enabling mechanism-guided prioritization of candidate molecules.

The three-dimensional structure of the bait protein was modeled using AlphaFold3 (https://alphafoldserver.com/). Molecular docking with small molecules was performed using Autodock Vina (version 4.2.6) with the following parameters. The grid center, representing the center of the docking space, was set at coordinates (10.078, 6.834, 9.385). The grid size, indicating the edge length of the search space, was set at 20 Å in the x, y, and z directions. The grid space, representing the interval between each grid point, was set to 0.375 Å. The exhaustiveness parameter was set to 128 to ensure a thorough search of the docking space. Subsequently, we used a multitiered screening approach, including high-throughput screening, VS, and refined filtering, to identify small molecules with favorable binding affinities ranging from −7.0 to −10.0 kcal/mol. This systematic selection process focused on commercially available compounds that demonstrated promising binding characteristics, paving the way for further analysis and validation.

Statistical analysis

Data are expressed as the means ± SEM. Statistical analysis was performed with GraphPad Prism software (version 9.0). Student’s t test was used to assess the differences between groups. A one-way analysis of variance (ANOVA) was used to assess the differences among multiple groups. A P value less than 0.05 was considered statistically significant.

Supplementary Materials

This PDF file includes:

Figs. S1 to S16

Tables S1 to S6