The study design was reviewed and approved by the Animal Ethics Committee of Ningxia University (permit No. 24-F-051). The procedures involving animals were carried out in accordance with the Animal Ethics Procedures and Guidelines of the People’s Republic of China. All efforts were made to minimize suffering and to reduce the number of sheep used in the experiment.

The IVM-sensitive and resistant strains of H. contortus were provided by the Inner Mongolia Academy of Agricultural and Animal Husbandry Sciences (IMAAAHS). The sensitive strain (HC-S) was isolated from Yuci, Shanxi, China, passaged, and maintained in sheep at IMAAAHS for 6 years. The resistant strain (HC-R) originally came from the Moredun Research Institute in the United Kingdom and was passaged and maintained in sheep at IMAAAHS for 9 years.

Since the three-dimensional structure of HC-Pgp has not been experimentally resolved, we constructed its structural model using homology modeling. The target protein preparation involved retrieving the full-length amino acid sequence of HC-Pgp (UniProt ID: A0A7I4YQ55) from the UniProt database for homology modeling using SWISS-MODEL.¹ The constructed protein model underwent rigorous validation through Ramachandran plot evaluation to ensure its suitability for subsequent molecular docking studies. SWISS-MODEL and validation Ramachandran plot ensure structural accuracy of HC-Pgp, enabling reliable virtual screening. Concurrently, Potential ligand binding sites were identified using Schrödinger software (Maestro12.8) suite. Schrödinger’s SiteMap and Glide modules identify ligand-binding pockets, foundational for docking studies. Preprocessing was performed using the Protein Preparation Wizard module. The SiteMap algorithm was applied to detect potential ligand binding sites with a 1.0 Å probe radius, and at least 15 characteristic probes were generated per site. Based on the prediction results, the Glide module was employed to construct a docking grid centered on the primary binding pocket, laying a solid computational foundation for virtual screening.

The predominant alkaloids in S. alopecuroides L. were retrieved from PubChem and subsequently subjected to computational preparation using the LigPrep 5.0 module of Schrödinger Suite. LigPrep adjusts protonation states, removes salts, and generates tautomers/stereoisomers, ensuring ligands mimic biological conditions. To enhance the reliability of docking results, ligand conformations were systematically optimized to mimic interactions under physiological conditions, thereby reducing potential computational false positives. Molecular processing included: (1) adjustment of protonation states under physiological pH conditions (7.0 ± 2.0), (2) removal of inorganic salts and counterions, (3) addition of hydrogen atoms with stereochemical optimization, (4) generation of biologically relevant tautomers and stereoisomers (maximum isomer limit = 32), and (5) conformational optimization through ring-constrained energy minimization employing the OPLS4 force field. OPLS4 force field minimization improves conformational accuracy, enhancing docking reliability. All computational parameters strictly adhered to the module’s default recommended settings to ensure methodological reproducibility and standardization.

Molecular docking were employed to prioritize alkaloids with optimal HC-Pgp binding stability, forming a robust theoretical framework for downstream in vitro studies. Initial ADMET filtration of database molecules was performed using the Schrödinger QuickProp module (23), retaining only compounds compliant with Lipinski’s rule of five and devoid of reactive functional groups for subsequent docking studies. Flexible docking was then conducted via the Schrödinger platform to prioritize alkaloids with the highest predicted binding affinities based on docking scores. Interaction analysis of the top five candidate compounds was performed through molecular visualization using Maestro, Discovery Studio (Dassault Systèmes BIOVIA, 2019), and PyMOL to identify key amino acid residues mediating protein-ligand interactions.

Molecular dynamics simulations were conducted using the GROMACS 2022.6 GPU version to confirm dynamic stability and binding mode persistence (24, 25). The initial structure for the simulations was derived from molecular docking results. The ligand topology file was generated employing the AMBER force field through the ACPYPE script, specifically using the AMBER99SB-ILDN force field for the protein topology file. The molecular dynamics simulation protocol included the definition of the simulation box, solvation with water molecules, and the addition of counter ions. Subsequently, energy minimization was performed, followed by equilibration in both the canonical ensemble system (NVT) and the constant-pressure system (NPT). The final phase of the simulation involved 100 ns molecular dynamics run under constant temperature and pressure conditions. Upon completion of the simulation, the last 5 ns of the trajectory were extracted for further analysis. The gmx_mmpbsa tool was utilized for the calculation of free energy to gain insights into the energetics of the molecular system.

The development inhibition effect was assessed using a modified larval development inhibition test (LDT) (26). In a 24-well plate, 265 μL of physiological saline (Nacl, 0.9%), 70 μL of Earle’s balanced salt solution (Sigma–Aldrich Corporation, United States), 20 μL of H. contortus eggs (About 100 eggs), and 5 μL of 5 mg/mLwater-soluble amphotericin B (Solarbio, Beijing, China) were added to each well. The mixture was incubated at 27°C in a biochemical incubator for 24 h. Subsequently, 80 μL of ALO (Baoji Fangsheng Biological Development Co., Ltd., Baoji, China), 80 μL of IVM (Shanghai Yuanye Bio-Technology Co., Ltd., Shanghai, China), a combination of 40 μL ALO + 40 μL IVM, and 1% DMSO (a negative control) was added to the system in individual wells. The final concentrations of both ALO and IVM were prepared through two-fold serial dilutions spanning 0 to 400 ng/mL. Three independent biological replicates were set up for each experimental group. After 6–7 d, L3 larvae were counted under an inverted microscope (OLYMPUS IX73), and the median lethal dose (LD₅₀) was calculated using the Karber method, its basic arithmetic formula is:lgLD₅₀ = ∑1/2 (Xi + Xi + 1) (Pi + 1 − Pi). Where Xi is the logarithm of the dose Pi is the rate of developmental inhibition.

Feces from lambs infected with H. contortus were collected and incubated in aerobic conditions at 24°C with 80–85% humidity for 5–10 d. L3 larvae were isolated from the feces using the Baermann funnel system (27), and the migration ability was assessed using a modified larval migration inhibition test (LMIT) (28). In each well, approximately 100 L3 larvae were added to 1,710 μL of sterile water. Various treatment solutions were then added to individual larval suspensions: 90 μL of ALO (final concentration 0.39–400 μg/mL), 90 μL of IVM (final concentration 0.39–400 μg/mL), 45 μL of IVM + 45 μL of ALO, and 1% DMSO (a negative control). Three replicate experiments were conducted for each treatment.

After 24 h of incubation at 24°C, the liquid from each well was transferred to migration plates containing 0.125% agar and incubated for an additional 24 h. The solutions at the bottom of the migration plates were transferred to a new 24-well plate. The number of migrating larvae was assessed under an inverted microscope, and the migration rate was calculated. Using GraphPad Prism 8.0.1, a regression curve was fitted with the logarithm of each treatment concentration and the larval migration inhibition rate on the X-and Y-axes, respectively. The half-maximal effective concentration (EC₅₀) was calculated using the equation: Y = Bottom + (Top − Bottom)/(1 + 10^[(LogEC50 – X) × HillSlope]).

The exsheathed L3 larvae were exposed to 25 μg/mL ALO, 25 μg/mL IVM, or a combination of 25 μg/mL ALO + 25 μg/mL IVM for 24 h, with 0.1% DMSO serving as the vehicle control. Post-treatment, the specimens were fixed with 2.5% glutaraldehyde and morphologically analyzed using a scanning electron microscope (SU8100) at an accelerating voltage of 3.0 kV.

The effect of ALO on HC-Pgp activity was determined by monitoring rhodamine 123 (Rhe123) accumulation in H. contortus (29). Approximately 30,000 exsheathed L3 larvae were exposed to ALO, IVM, or their combination for 24 h (as described previously), followed by incubation in Rhe123 solution (1 mL of 1.5 μM) under dark conditions at room temperature using a constant temperature shaker for 30 min. The worms were subsequently centrifuged (3,000 × g, 3 min) and washed twice with 1 mL distilled water. The pellet was resuspended in 1 mL distilled water and maintained in darkness for 60 min. After final centrifugation (3,000 × g, 3 min), the supernatant was collected and stored protected from light for 60 min prior to analysis. Rhe123 concentration in the supernatant was quantified using a multifunctional microplate reader (Agilent Technologies Inc., United States) with specific fluorescence parameters (excitation wavelength λex = 495 nm, emission wavelength λem = 525 nm). Sample concentrations were calculated against a standard curve. Three independent experimental replicates were performed. Statistical analysis was conducted using GraphPad Prism 8.0.1, with statistical significance defined as p < 0.05.

Using the primary amino acid sequence of HC-Pgp (accession number: A0A7I4YQ55) as the target, a BLASTP and SWISS-MODEL template search identified 4f4c.1. A (X-ray, 3.40 Å) as the best template with a sequence identity of 55.83% with HC-Pgp, which is sufficient to construct a reliable model. The sequence alignment between HC-Pgp and 4f4c.1. A is illustrated in Supplementary Figure S1. The model generated using SWISS-MODEL was validated through a Ramachandran plot (Supplementary Figure S2). The analysis revealed that 93.3% of the residues were in the most favored regions and 6.6% in the additional allowed regions, while 0.1% of the amino acids were in the disallowed regions. These data validated the modeled structure, making it suitable for docking studies (30).

Structure-based virtual screening (SBVS) in drug discovery offers multiple advantages, including the capability to efficiently screen large compound libraries, reduced costs and time compared to experimental screening, as well as the potential to explore broad chemical spaces. In this study, preliminary ADMET filtering of database molecules was performed using the QuickProp module, retaining molecules compliant with Lipinski’s rule of five (Ro5) and devoid of reactive fragments. The results demonstrate that all 13 major alkaloids from S. alopecuroides L. exhibited physicochemical parameters within acceptable ranges for hydrogen bond donors/acceptors, molecular weight, and lipophilicity (Table 1), suggesting their potential as promising drug candidates. Subsequent molecular docking screening revealed that ALO exhibited the highest binding affinity (−6.83 kcal/mol) toward HC-Pgp (Supplementary Table S1; Supplementary Figure S3). As shown in Figures 1A,B, ALO interacts with HC-Pgp through binding interactions predominantly mediated by hydrogen bonding and hydrophobic forces at the Glu556 residue.

Molecular dynamics simulations spanning 100 ns were conducted to evaluate the interaction between ALO and HC-Pgp. As shown in Figure 2A, the ALO-HC-Pgp complex achieved equilibrium after 20 ns of simulation. RMSF analysis, which quantifies atomic deviations from average positions, revealed relatively low values (<1 Å) across both ligand and receptor domains (Figure 2B), indicating sustained structural stability. The Rg, calculated as the mass-weighted root mean square distance of atoms from the system’s centroid, demonstrated minimal variation (ΔRg ≈ 0.05 nm) within the range of 3.94–3.99 nm throughout the trajectory (Figure 2C), further confirming conformational stability. Protein-ligand hydrogen bonding patterns, critical for binding affinity and complex stabilization, were dynamically monitored (Figure 2D). Additionally, Figure 2E depicts the binding mode of the ALO-HC-Pgp complex at 0, 20, 40, 60, 80, and 100 ns, illustrating the initial movement of the ALO around the binding pocket of the HC-Pgp. These multi-parameter analyses demonstrate that ligands achieve specific regulation of HC-Pgp function by stabilizing active site conformations and forming high-strength hydrogen bond networks.

MM-PBSA calculations were performed using the final 5 ns of molecular dynamics simulations to quantify the binding affinities between ligand and its receptor. Van der Waals energy, electrostatic energy, electrostatic contribution to solvation-free energy from the Poisson-Boltzmann model, nonpolar contribution, and entropy were calculated for each complex. The results demonstrated that ALO exhibits strong binding affinity with HC-Pgp, as evidenced by a calculated binding free energy of −88.16 kcal/mol (Table 2). Furthermore, we analyzed the residue binding energy contribution of each amino acid in the receptor after ligand binding, which is a critical approach for identifying hotspot residues that play major roles in ligand binding during simulations. The energy decomposition results revealed that LEU555 and GLU556 in the active site and catalytic residues made significant contributions, while residues such as LEU536, GLN552, THR573, and GLU548 also played important roles in energy contributions (Figure 3).

A comparison of the median lethal dose between the experimental and control groups in the LDT revealed an inhibitory effect of ALO on larval development to a lesser extent than that of IVM. The ALO and IVM combination exerted a significant inhibitory effect on H. contortus development that was approximately 2-to 3-fold of the effect of ALO treatment alone (Table 3; Figure 4). In the HC-R strain, the LD₅₀ for the ALO + IVM group was 17.72 ng/mL, while the LD₅₀ values for the IVM and ALO groups were 45.02 and 50.90 ng/mL, respectively, indicating that the ALO + IVM combination had a significantly higher developmental inhibition effect on HC-R than that of either agent alone.

Similarly, in the HC-S strain, ALO exhibited an inhibitory effect on H. contortus development; ALO and IVM combination showed a significantly higher inhibitory effect on HC-S development than ALO or IVM monotherapy.

At the L3 stage of larvae development, the LMIT method was used to evaluate the effect of the ALO and IVM combination on the migration ability of H. contortus (Table 4; Figure 5). The EC₅₀ values for the resistant strain of H. contortus were 25.43 μg/mL for ALO, 18.77 μg/mL for IVM, and 7.50 μg/mL for the combination of ALO and IVM. For the sensitive strain, the EC₅₀ values were 5.33 μg/mL for ALO, 5.42 μg/mL for IVM, and 4.34 μg/mL for the combination of ALO and IVM. These results indicate that ALO can inhibit the migration ability of both IVM-resistant and sensitive strains of H. contortus. Furthermore, the ALO and IVM combination showed a significantly greater inhibitory effect on the migration ability of the HC-R strain than ALO or IVM alone, although this effect was not significant for the HC-S strain. This differential effect appears to be directly associated with P-gp, as evidenced by the Rhe123 accumulation assay, which revealed significantly higher activity in HC-R compared to HC-S (p < 0.05) (Supplementary Figure S4).

The nematode cuticle, a complex extracellular matrix composed of proteins, lipids, and carbohydrates, protects nematodes from environmental and pathogenic damage (31). To evaluate the effects of ALO on the cuticle of H. contortus L3, we observed H. contortus L3 exposed to 25 μg/mL ALO under a scanning electron microscope. ALO-induced changes in the cuticle and led to severe cuticle shrinkage when combined with IVM (Figure 6). This is consistent with LDT and LMIT results, indicating that the ALO and IVM combination has a greater effect on the cuticle of H. contortus than either agent individually, suggesting that ALO has a synergistic effect with IVM.

The intracellular accumulation of Rhe 123 among different treatment groups was analyzed to evaluate the effects of ALO on HC-Pgp activity (Figure 7). Compared with the control group (R + 1%DMSO), the R + 25 μg/mL IVM group exhibited a significantly higher intracellular Rhe 123 concentration (p < 0.05), indicating reduced efflux activity or enhanced retention. In contrast, the R + 5 μg/mL ALO group displayed a significantly lower Rhe 123 concentration relative to the control (p < 0.05). Notably, the combination treatment group (R + 5 μg/mL ALO + 25 μg/mL IVM) showed a pronounced synergistic interaction, with Rhe 123 concentrations being significantly lower than those in the R + IVM group alone (p < 0.05). These results suggest that ALO effectively modulates HC-Pgp-mediated efflux activity, potentially reducing substrate efflux.