Identification of PEDV inhibitors targeting 3CL protease

Abstract

Porcine epidemic diarrhea (PED), caused by porcine epidemic diarrhea virus (PEDV), is a highly contagious gastrointestinal disease characterized by vomiting, diarrhea, and dehydration, with mortality rates approaching 100% among suckling piglets. The PEDV 3C-like protease (3CLpro) is essential for viral replication and regarded as a critical target for antiviral inhibitor development. In this study, we aimed to identify small-molecule inhibitors of PEDV by targeting 3CLpro. Virtual screening of 1.6 million compounds from the ChemDiv library identified four potential candidates. Molecular dynamics simulations, specifically analyzing RMSD, RMSF, and Rg, demonstrated increased structural stability of the compound-protease complexes compared to the monomeric enzyme. All compounds had low cytotoxicity in Vero cells (CC₅₀ ​> ​200 ​μM). Fluorescence resonance energy transfer-based assays demonstrated dose-dependent inhibitory activity of the compounds against 3CLpro. Among the candidates, compound F366-0161 exhibited the weakest inhibition, with an IC₅₀ value of 151.5 ​μM. Two analogues, 3238-0395 (IC₅₀ of 121.4 ​μM) and L878-0493 (IC₅₀ of 123.6 ​μM), exhibited moderately enhanced activity. Y041-1672 was identified as the most effective inhibitor, with an IC₅₀ of 86.48 ​μM. In viral replication inhibition assays, Y041-1672 reduced PEDV replication, with an EC₅₀ of 17.97 ​μM and a selectivity index (SI) of 15.5 (CC₅₀/EC₅₀). These results were validated by RT-qPCR, plaque assays, immunofluorescence, and Western blot analyses. In vitro validation confirmed Y041-1672 as the optimal antiviral candidate, and time-of-addition experiments indicated that inhibition primarily occurred during viral replication. This study identifies scaffold molecules for PEDV antiviral drug development, providing strategic insights for PED treatment.

Highlights

• •Four PEDV 3CLpro inhibitors were obtained by virtual screening and in vitro validation methods.
• •All the inhibitors exhibit low cytotoxicity and dose-dependent inhibitory activities against 3CLpro.
• •Studies revealed Y041-1672 targets the catalytic dyad (His41/Cys144), specifically inhibiting viral replication stage.
• •Compound Y041-1672 has the highest selectivity index and efficiently suppress viral replication as antiviral candidate.

Introduction

Porcine epidemic diarrhea (PED), caused by the porcine epidemic diarrhea virus (PEDV), is a highly contagious enteric disease characterized by severe diarrhea, vomiting, and dehydration in pigs. While affecting all age groups, PED is particularly lethal to neonatal piglets, with mortality rates approaching 100% (Jung et al., 2020). PED was first identified in the United Kingdom and Belgium in the 1970s, subsequently spread across Europe and became endemic (Wood, 1977). In 2010, an outbreak of highly pathogenic PED in the United States has spread to China, with morbidity and mortality rates approaching 100%, resulting in a devastating impact on China’s pig industry (Li et al., 2012; Sun et al., 2012; Tian et al., 2014).

PEDV, a member tof the genus Alphacoronavirus within the family Coronaviridae, possesses a ∼28 ​kb single-stranded, positive-sense RNA genome. The genome replicase polyproteins (pp1a/pp1ab) are processed by 3CLpro (Nsp5) into mature nonstructural proteins (Nsp4–16), which encode a variety of enzymes required for replication, such as RdRp (Nsp12), helicase (Nsp13), exonuclease (Nsp14), endonuclease (Nsp15) and 2′-o-methyltransferase (Nsp16). These enzymes are essential for viral replication (Kocherhans et al., 2001; Lee, 2015). As a result, 3CLpro has emerged as a significant target for antiviral strategies aimed at inhibiting viral replication (Su et al., 2023; Wu et al., 2023).

Although maternal vaccination reduces mortality in suckling piglets (Jung and Saif, 2015; Langel et al., 2020), no specific antiviral therapy is currently available. PEDV causes rapid disease progression and high mortality rates in piglets, while vaccination remains the only available prevention strategy. However, as the development of a protective immune response requires temporal progression, there is an urgent need for prophylactic or therapeutic interventions with rapid efficacy. Pharmacological interventions present a distinct advantage over vaccines by exhibiting a rapid onset of action, thereby enabling immediate mortality reduction and mitigation of economic losses. Furthermore, small-molecule inhibitors exhibit the capacity to traverse both the blood-brain barrier and placental barrier, thereby enabling therapeutic interventions that remain inaccessible to larger pharmaceutical compounds (Wang et al., 2023). Pathak et al. (2023) screened a library of 100,000 natural compounds targeting 3CLpro and identified four small molecules with potent enzymatic inhibition via computational simulations; however, these findings lack validation through in vitro or in vivo studies (Pathak et al., 2023). Similarly, Li et al. (2024) obtained compounds capable of inhibiting viral replication targeting 3CLpro, through screening a flavonoid compound library (Li et al., 2024). Bahun et al. (2022) conducted a screening study targeting the 3CL protease to evaluate plant-derived polyphenols as potential inhibitors against SARS-CoV-2, identifying several natural compounds including quercetin, ellagic acid, and curcumin as candidates with antiviral activity (Bahun et al., 2022).

Here, we combined structure-based virtual screening of 1.6 million synthetic compounds with molecular dynamics simulations to screen PEDV 3CLpro inhibitors. Lead candidates were further validated through enzymatic assays, cytotoxicity, and in vitro antiviral studies. Our work not only identifies novel scaffolds for PEDV drug development but also establishes a multidisciplinary framework for targeting coronaviral proteases.

Results

Virtual screening and binding energy calculationss

AutoDock Vina was employed to simulate compound binding to PEDV 3CLpro and to calculate their binding free energies. To identify the most promising compounds against PEDV 3CLpro, we applied the Molecular Mechanics/Generalized Born Surface Area (MM/GBSA) method for calculating receptor-ligand binding free energies. The compounds were ranked based on their binding energy scores, and the top 8 compounds with binding energies lower than −8.8 ​kcal/mol and the highest MM/GBSA values were selected. The results revealed MM/GBSA values ranging from −74.43 ​kcal/mol (most favorable) to −52.95 ​kcal/mol (least favorable) (Table 1). For further validation, four highest-scoring compounds were prioritized: chemdiv_56825-3238-0395 (MM/GBSA ​= ​−64.78 ​kcal/mol); chemdiv_566221-F366-0161 (MM/GBSA ​= ​−65.34 ​kcal/mol); chemdiv_906108-L878-0493 (MM/GBSA ​= ​−74.43 ​kcal/mol); chemdiv_1197684-Y041-1672 (MM/GBSA ​= ​−70.71 ​kcal/mol).

Table 1.

The top 8 compounds screened for PEDV 3CL^pro.

Compound number	Binding energy score	MMGBSA (Kcal/mol)	Compound structure
V026-1640	−9.6	−63.16
F366-0161	−9.1	−65.34
L878-0493	−8.8	−74.43
Y041-1672	−8.8	−70.71
S720-0109	−8.8	−65.48
3238-0395	−8.8	−64.78
4896-4227	−8.7	−60.24
F816-0106	−9.1	−52.95

Compound-protein interaction analysis

To elucidate the interaction mechanisms and dynamic binding behavior of small molecule compounds targeting the PEDV 3CLpro, we implemented molecular visualization techniques coupled with computational simulation analysis. The structural basis of ligand-protease interactions was systematically investigated through molecular dynamics simulations, with computational characterization of non-covalent interaction networks using PyMOL (https://pymol.org). Molecular dynamics simulations revealed that all four compounds formed non-covalent interaction networks with PEDV 3CLpro, including hydrogen bonding, π-π stacking, halogen bonding, and alkyl interactions (Fig. 1A and B). These interactions stabilized the ligand-protease complexes and influenced the enzyme’s catalytic activity by restricting substrate access to the active site.

Fig. 1.

Interaction diagram of PEDV 3CLpro amino acid residues with compounds. A 2D graphical representation of the chemical interactions between small molecule compounds and the amino acids of the substrate binding pocket. B 3D spatial structure of small molecule compounds bound to the substrate binding pocket.

Stability analysis of compound-protease complexes through MD simulations

Protein stability serves as a fundamental determinant of functional efficacy. To assess the structural preservation of compounds upon binding to the PEDV 3CLpro, molecular dynamics simulations were conducted over a 100-ns trajectory. System stability was evaluated using critical metrics: root mean square deviation (RMSD), root mean square fluctuation (RMSF), and radius of gyration (Rg), hydrogen bond occupancy, and free energy landscape (FEL). These analyses collectively elucidated the conformational dynamics and binding interactions governing compound-protease stability.

Over a 100-ns trajectory, all systems exhibited minimal structural deviations. The RMSD values of the complexes ranged between 0.18 and 0.35 ​nm, indicative of robust structural integrity throughout the simulation. Notably, the Y041-1672 complex displayed superior stability compared to others (Fig. 2A). RMSF analysis highlighted reduced residue flexibility near the catalytic dyad (Cys144-His41) in ligand-bound systems, suggesting steric hindrance of the active site (Fig. 2B). Hydrogen bond dynamics were quantified to evaluate intermolecular interactions, hydrogen bond analysis revealed persistent interactions for Y041-1672 (≤4 bonds), whereas L878-0493 exhibited transient peaks of five bonds between 60 and 80 ns (Fig. 2C). Rg analysis demonstrated structural compaction in ligand-bound systems (2.13–2.25 ​nm), with 3238-0395 showing the most pronounced effect (Fig. 2D). This trend implies that ligand binding induces structural tightening without disrupting native folding. FEL analysis corroborated these findings, as all complexes exhibited single dominant energy minima (blue regions, Fig. 2E–H), consistent with stable conformational ensembles. The absence of erratic energy clusters further supports the thermodynamic stability of ligand-bound states.

Fig. 2.

Molecular dynamics simulation stability analysis of PEDV 3CLpro in complex with compounds. Molecular dynamics simulations of the protein-small molecule complex were conducted using GROMACS. The resulting simulation data were processed to extract root mean square deviation (RMSD), root mean square fluctuation (RMSF), and radius of gyration (Rg), hydrogen bond occupancy, and free energy landscape (FEL). The data after molecular dynamics simulations were extracted, with RMSD as the X, Rg as the Y-axis, and the energy during the simulation process as the Z-axis. A RMSD analysis of the four small molecule compounds and their corresponding proteins. B RMSF analysis of the four small molecule compounds and their corresponding proteins. C Hydrogen bond analysis of the four small molecule compounds and their corresponding proteins. D Radius of gyration analysis of the four small molecule compounds and their corresponding proteins. E FEL of PEDV 3CLpro. F FEL of PEDV 3CLpro with F366-161. G FEL of PEDV 3CLpro with Y041-1672. H FEL of PEDV 3CLpro with L878-0493.

Collectively, RMSD, RMSF, Rg, and FEL analyses confirm that compound binding enhances the structural stability of PEDV 3CLpro relative to its unbound form. The observed stabilization mechanism—combining hydrogen bonding, hydrophobic interactions, and steric constraints—positions these compounds as promising candidates for therapeutic development. These results provide a robust framework for rational drug design and functional group optimization targeting viral proteases.

The protease activity of PEDV 3CLpro

To obtain active PEDV 3CLpro, the protein was expressed using the E. coli BL21 (DE3) expression system and subsequently purified via nickel (Ni) affinity chromatography. SDS-PAGE and Western blot analysis confirmed a single band corresponding to the expected molecular mass (∼35 ​kDa) (Fig. 3A and B). Enzymatic activity was assessed through fluorescence resonance energy transfer (FRET)-based assays. Substantial fluorescence emission was observed in reactions containing PEDV 3CLpro, demonstrating robust enzymatic activity (Fig. 3C). Dose-response experiments revealed a concentration-dependent increase in reaction rates, with fluorescence intensity exhibiting temporal progression across tested enzyme concentrations (Fig. 3D). These findings confirm preserved catalytic functionality of the purified enzyme. No significant fluorescence signal was detected in negative control reactions, eliminating nonspecific activity concerns.

Fig. 3.

Expression and enzymatic activity verification of PEDV 3CLpro. A Protein samples were separated by electrophoresis on a 12% SDS-PAGE gel. M: protein marker (8–200 ​kDa); Lane 1: pET-30a plasmid vector (negative control); Lane 2/3: pET-30a-PEDV 3CLpro. B Protein samples were separated by SDS-PAGE, transferred to PVDF membranes, and detected using a His-tag monoclonal antibody. C The enzymatic activity of PEDV 3CLpro was verified using peptide substrates. When determining the enzyme activity, 10 ​μL of purified PEDV 3CLpro diluted with buffer was added, and then 10 ​μL of the substrate was immediately added. After reacting at 37 ​°C for 1 ​h, the emission light of 480 ​nm under 355 ​nm excitation light was measured in the microplate reader. D Substrate decomposition efficiency at different 3CL concentrations. Different concentrations of protein were diluted 2-fold and then added to 10 ​μL to a 384-well plate, followed by 10 ​μL of substrate, and the excitation light at 355 ​nm emission light was continuously measured in a microplate reader at 480 ​nm. Statistical analyses were performed using Student’s t-test, ∗P ​< ​0.05; ∗∗P ​< ​0.01; ∗∗∗P ​< ​0.001.

Inhibition of 3CLpro activity

To assess the inhibitory effect on PEDV 3CLpro activity, IC₅₀ (half-maximal inhibitory concentration for enzyme activity) was determined. The 2-fold diluted compounds were prepared and then added to the 3CLpro reaction mixture. The inhibition of 3CLpro activity is significantly dependent on the concentration of the compound, and IC₅₀ calculation is performed using non-linear fit. FRET-based IC₅₀ assays demonstrated dose-dependent inhibition of 3CLpro by all four compounds. F366-0161 exhibited the weakest IC₅₀ of 151.5 ​μM, whereas 3238-0395 and L878-0493 respectively showed slightly reduced IC₅₀ of 121.4 ​μM and 123.6 ​μM. Significantly, Y041-1672 displayed the strongest inhibition IC₅₀ of 86.48 ​μM (Fig. 4A–D).

Fig. 4.

Evaluation of the dose-response inhibitory effects of compounds against 3CLpro and determination of IC₅₀ values. The PEDV 3CLpro was subjected FRET assays in the presence of different concentrations of the compound, and the fluorescence ratios were calculated by comparison with the untreated control. The inhibition curves were generated by regression. A F366-161; B L878-0493; C 3238-0395; D Y041-1672.

Antiviral efficacy in vitro

In order to verify whether the compounds have an inhibitory effect on the virus, antiviral efficacy was evaluated in virus-susceptible cells. Before evaluating the antiviral inhibitory effects of the compounds, we assessed the cytotoxicity of four compounds in Vero cells using the CCK-8 assay. The results demonstrated that cells maintained favorable morphology with minimal mortality at low concentrations, while a gradual reduction in cell viability was observed with increasing concentrations. Notably, all four compounds exhibited 50% cytotoxic concentration (CC₅₀) values exceeding 200 ​μM (Fig. 5A–D), collectively indicating their low cytotoxic profiles.

Fig. 5.

Determination of CC₅₀ of the compounds in Vero cells. Vero cells were cultured in presence of different concentrations of compounds for 36 ​h. Cell viability was tested using CCK-8 reagent and the percentage of viable cells was calculated by comparison with the non-treatment control. The inhibition curve was generated by nonlinear regression. A F366-161; B 3238-0395; C L878-0493; D Y041-1672.

The inhibitory effects of the compounds on viral replication were investigated by RT-qPCR. The RT-qPCR analysis demonstrated that both Y041–1672 and F366-0161 induced significant reductions in viral RNA copy numbers compared to the vehicle control. Notably, Y041-1672 exhibited a concentration-dependent antiviral effect, with RNA copy numbers decreasing by 8.1 ​× ​10⁷ ​at 6.25 ​μM and 3.99 ​× ​10⁸ ​at 100 ​μM relative to the control group (7.38 ​× ​10⁸). In contrast, F366-0161 showed no measurable impact on viral replication at lower concentrations (≤ 25 ​μM). However, dose-responsive inhibition was observed at higher concentrations, with viral copy reductions of 1.82 ​× ​10⁸ ​at 50 ​μM and 3.29 ​× ​10⁸ ​at 100 ​μM (Fig. 6A).

Fig. 6.

The effect of the compounds on PEDV replication in Vero cells. Vero cells were cultured in 12-well plates for 12 ​h before being infected with PEDV (MOI = 0.01) and cultured for 30 ​h in the present of different concentrations of compound. After adding the compound and culturing for 30 ​h, the cells of different groups and the supernatant cultures were collected respectively. The collected cells were subjected to RNA extraction and the PEDV N gene was detected by RT-qPCR. The collected supernatant was continued to infect Vero cells at the same dose. After culturing for 30 ​h, the cells were fixed and stained with crystal violet for plaque counting. A Assessment of the effect of different concentrations of compounds on the viral copy numbers during replication. B Empty-spot assay to evaluate the impact of different small molecule compounds on viral replication. C EC₅₀ of Y041-1672. For the EC₅₀, 2 ​h after Vero cells were infected with the virus, different concentrations of Y041-1672 compound were added, and cell viability was assessed using the CCK-8 assay 48 ​h post-infection. D Immunofluorescence assay of viral replication inhibition. Scale bar, 200 μm. E Antiviral effects of Y041-1672 were detected by Western blotting using PEDV-N monoclonal antibody. Statistical analyses were performed using Student’s t-test. ∗P ​< ​0.05; ∗∗P ​< ​0.01; ∗∗∗P ​< ​0.001.

In plaque assay, both F366-0161 and Y041-1672 exhibited significant inhibitory effects on viral proliferation (Fig. 5B). Compared with the control group, F366-0161 reduced viral titers by approximately 2-fold at 25 ​μM and 1.7-fold at 12.5 ​μM. Notably, Y041-1672 demonstrated enhanced potency, achieving a 2.2-fold reduction in viral titer at 6.25 ​μM and a more pronounced 5-fold reduction at 25 ​μM (Fig. 6B). Y041-1672 exhibited the highest antiviral efficacy against PEDV, and we performed EC₅₀ assay of Y041-1672 on Vero cells using CCK-8, which showed an EC₅₀ of 17.97 ​μM, selectivity index (SI) of 15.5 (CC₅₀/EC₅₀) (Fig. 6C).

Western blotting and immunofluorescence assay (IFA) were used to detect the inhibition effect of Y041-1642 on PEDV. Following viral infection, the compounds were added to the culture medium at concentrations of 2 ​μM, 10 ​μM, and 50 ​μM, and intracellular viral detection was performed using a PEDV-N antibody at 30 ​h. The results demonstrated that, compared to the positive control group, cells treated with 50 ​μM of the compound exhibited negative viral signals, while the 10 ​μM concentration significantly reduced viral content within the cells. In contrast, the 2 ​μM concentration showed no significant impact on viral replication (Fig. 6D). Further analysis by Western blotting confirmed the expression of PEDV-N protein in infected cells. Notably, viral N protein was undetectable at the 50 ​μM concentration, indicating that viral replication was effectively suppressed at this dose (Fig. 6E). These results from RT-qPCR, plaque assays, and IFA collectively confirm that Y041-1672 effectively inhibits PEDV replication.

Stage-specific antiviral effects

To further elucidate how the compounds exert their anti-PEDV activity, RT-qPCR and Western blotting were used for the effects of PEDV adsorption, invasion, and value addition. Vero cells were treated with 50 ​μM compound before, during, or after PEDV infection (Fig. 7A). Viral replication was assessed 30 ​h post-infection via Western blotting (anti-PEDV-N antibody) and qRT-PCR (N gene target). Post-infection treatment significantly reduced viral RNA and nucleocapsid protein yield versus controls. No inhibitory effects were observed in pre- or co-treatment groups (Fig. 7B and C). The result indicating selective interference with viral replication rather than adsorption or entry, this aligns with the role of 3CLpro in polyprotein processing during the replication phase.

Fig. 7.

The impact of Y041-1672 on the PEDV life cycle. A Experimental design. Compound was added to cells before, during, and after infection (horizontal red lines). During the pre-infection, during-infection, and post-infection stages of Vero cells infected with PEDV (MOI = 0.01), the Y041-1672 compound (50 ​μM) was added respectively. After an adsorption period of 1.5 ​h for each addition, the culture medium was replaced with DMEM supplemented with trypsin, followed by a continuous cultivation for 30 ​h. B Assessment of the effect of different time of compound on the viral copy numbers during replication. Cells with added compounds at different time points were collected. After extracting RNA, the PEDV N gene was detected by RT-qPCR, and GAPDH was used as the internal reference for reference. C Western blotting of samples before and after addition of Y041-1672 compound using PEDV N monoclonal antibody. Statistical analyses were performed using Student’s t-test. ∗P ​< ​0.05; ∗∗P ​< ​0.01; ∗∗∗P ​< ​0.001.

Detection of the inhibitory effect of compounds on delta coronavirus

To verify whether compound Y041-1642 has the same inhibitory effect on other coronaviruses, we conducted a dose-added inhibition experiment in porcine delta coronavirus (PDCoV). The inhibitory effects of the compounds on viral replication were evaluated by adding them to the culture medium at concentrations of 10 ​μM, 50 ​μM, and 100 ​μM post-viral infection. Intracellular viruses were detected using a PDCoV-N antibody at 30 ​h post-infection. The results revealed that, compared to the positive control group, cells treated with 100 ​μM of the compound exhibited weakly negative viral signals, while concentrations at or below 50 ​μM showed no reduction in intracellular viral content. Further Western blot analysis confirmed the expression of PEDV-N protein in infected cells. Notably, compared to the positive control, a diminished signal was only observed at 100 ​μM, whereas no reduction was detected in the other two concentration groups (Fig. 8A and B). This suggests that the compound has limited cross-inhibitory activity against other coronaviruses.

Fig. 8.

Inhibition of PDCoV by compound Y041-1672. LLC-PK1 cells were infected with PDCoV at a multiplicity of infection (MOI) of 0.01 for 2 ​h, after which the culture medium was replaced with DMEM supplemented with varying concentrations of the compound. The cells were subsequently incubated for an additional 24 ​h. Viral replication was assessed by IFA and Western blot analysis using a monoclonal antibody specific to the PDCoV-N protein. A IFA experiment of compound inhibiting virus proliferation. Scale bar, 400 μm. B The antiviral effects of the Y041-1672 compound were detected by Western blot using PDCoV-N monoclonal antibody.

Discussion

Since its identification in the 1970s, PEDV has caused recurrent epizootics with near-100% mortality in neonatal piglets, despite partial resistance in adult swine (Moon et al., 1973). The viral main protease (3CLpro/Nsp5), essential for polyprotein cleavage and replication, represents a prime therapeutic target (Resnick et al., 2021). While previous studies identified natural compounds including tomatidine (Wang et al., 2020), hypericin (Zhang et al., 2021) and quercetin (Li et al., 2020) as 3CLpro inhibitors, their limited availability and high costs hinder scalable development. Small-molecule drugs offer distinct advantages due to their pharmacokinetic properties, such as blood-brain barrier-penetration (Miao et al., 2022). To overcome these limitations, we performed structure-based virtual screening of 1.6 million synthetic compounds targeting PEDV 3CLpro. Lead candidates were rigorously evaluated through molecular dynamics simulations, FEL analysis and binding energy calculations, culminating in the selection of four prioritized inhibitors for experimental validation.

Enzymatic activity modulation by small molecules depends on their binding specificity to catalytic residues and non-covalent interaction networks. For coronaviral 3CLpro, the conserved Cys-His dyad underpins proteolytic function (Niesor et al., 2021). Inhibitors typically employ dual mechanisms: (i) competitive occupation of the active site or (ii) steric occlusion via peripheral binding. Beyond classical hydrogen bonds, non-canonical interactions—including C–H···π contacts and halogen bonding—critically stabilize ligand-protease complexes (Bowling et al., 2021; Fargher et al., 2022; Jablonski, 2023). For instance, F366-0161 anchors to the catalytic pocket through hydrogen bonding with Cys144 (3.7 ​Å) and π-π stacking with His41, synergistically restricting substrate access. Y041-1672 illustrates a unique strategy: the C144 of PEDV 3CL acts as a nucleophilic group, and its thiol group (-SH) directly attacks the carbonyl carbon of the substrate peptide bond. H41, as a broad base, its nucleophilicity is enhanced by deprotonating the thiol group (- SH → - S ^-) of CYS144. Y041-1672 inhibits deprotonation of C144 through multivalent interactions with the His41 active residue through amide π stacking, while also hindering the active site through hydrogen bonding and hydrophobic contact with Asn141 (2.4 ​Å)/Gly142 (2.8 ​Å). These spatially distributed interactions collectively enhance binding stability, underscoring the versatility of non-covalent forces in antiviral compound design.

FRET has emerged as a pivotal tool for probing protein interactions and drug discovery, as exemplified by its applications in characterizing enzyme kinetics (Jo et al., 2020) and quantifying ligand binding affinities (Algar et al., 2019). In this study, FRET-based profiling identified Y041-1672 as the most potent inhibitor of PEDV 3CLpro, exhibiting an IC₅₀ of 86.48 ​μM, EC₅₀ of 17.97 ​μM, and CC₅₀ of 279.2 ​μM. Viral replication assays confirmed its efficacy, with RT-qPCR and plaque assays revealing significant suppression of PEDV RNA (1.4-fold reduction at 12.5 ​μM) and complete inhibition of viral plaques at 50 ​μM. Despite moderate enzymatic potency, Y041-1672’s modular scaffold permits structural optimization—a strategy validated in Zika virus studies where analogous compounds disrupted replication-stage polyprotein processing (Miao et al., 2022). Time-of-addition experiments corroborated its replication-specific mechanism, as no inhibition occurred during viral adsorption or entry. Cross-coronavirus assays further demonstrated partial efficacy against PDCoV (≥100 ​μM), likely constrained by divergent 3CLpro substrate pockets. These findings highlight the compound’s dual potential as a PEDV-specific therapeutic lead and a template for pan-coronavirus inhibitor design.

The FRET-based screening identified Y041-1672 as a small-molecule inhibitor targeting the PEDV 3CLpro, representing a promising scaffold for antiviral development. However, its moderate inhibitory potency and the observed discrepancy between enzymatic and antiviral efficacy (IC₅₀ ​> ​EC₅₀) necessitate structural optimization. The compound is characterized by a methyl-substituted benzofuran core, a bicyclic ester moiety, and a chiral side chain containing a benzyl-protected carbamate group. Structural modifications could enhance interactions with the 3CLpro substrate-binding pocket: 1) Introducing bulkier substituents (halogens or trifluoromethyl) at the C1 methyl and methoxycarbonyl (OC O) positions may strengthen hydrophobic interactions, as evidenced in SARS-CoV-2 3CL inhibitors (Bai et al., 2022). 2) Replacing the labile ester group with more stable amides or heterocycles could improve metabolic stability while preserving hydrogen-bonding capacity (Leuner and Dressman, 2000). 3) Incorporating polar groups into the benzyl carbamate may enhance aqueous solubility and cellular permeability, whereas intracellular esterase-mediated cleavage of carbamate/ester moieties might generate active metabolites with improved 3CLpro affinity (Zhang et al., 2020).

The superior antiviral efficacy (EC₅₀) compared to enzymatic inhibition (IC₅₀) suggests additional antiviral mechanisms beyond 3CLpro inhibition, potentially involving multi-targeted viral inhibition or immune-modulatory pathways. In addition to its primary role as a viral RNA-dependent RNA polymerase (RdRp) inhibitor, remdesivir has been demonstrated to potentiate host antiviral immunity through the induction of interferon-beta (IFN-β) secretion (Martinez, 2020). This dual mechanism—direct suppression of viral replication coupled with enhanced cytokine-mediated immune activation—positions remdesivir as a multifunctional therapeutic agent in the management of RNA virus infections. As a nucleoside analog, ribavirin inhibits viral RdRp to interfere with viral nucleic acid synthesis; concurrently, it modulates the Th1/Th2 balance by promoting the secretion of Th1-type cytokines, such as interleukin-2 (IL-2), interferon-gamma (IFN-γ), and tumor necrosis factor-alpha (TNF-α), thereby enhancing T cell-mediated viral clearance (Werner et al., 2014). In addition, RAF265 has been verified to effectively inhibit the replication of PEDV both in vivo and in vitro. This compound inhibits cytoskeletal rearrangement and cellular translation. This dual inhibitory approach suggests that it has multiple targets (Wang et al., 2022). While this compound provides a foundational scaffold for PEDV therapeutics, iterative optimization cycles are required for drug development. The discrepancy between IC₅₀ and EC₅₀ values highlights the multifactorial nature of antiviral efficacy, where polypharmacology (multi-target effects) may contribute to efficacy. Balancing target specificity with off-target benefits will be crucial for advancing this chemotype.

Structural biology and bioinformatics are integral to modern scientific research, but studying these fields solely through virtual screening is insufficient for comprehensive drug discovery. In our study, we employed both virtual screening and subsequent validation techniques to identify small molecule compounds capable of targeting the 3CLpro of PEDV. These compounds effectively inhibited the replication of PEDV and provided valuable clues for the development of drugs for the treatment of PED, as well as ideas for the design and development of other coronavirus drugs.

Conclusions

In summary, our research systematically investigated the inhibitory mechanisms of small-molecule compounds against the 3CLpro of PEDV. Virtual screening of 1.6 million organic compounds successfully identified potential inhibitors, with Y041-1672 exhibiting the most potent antiviral activity (IC₅₀ of 86.48 ​μM, EC₅₀ of 17.97 ​μM) and low cytotoxicity (CC₅₀ of 279.2 ​μM), suppressing viral replication by targeting the catalytic dyad (Cys144-His41) via amide-π stacking and steric hindrance of the substrate-binding pocket. Molecular dynamics simulations confirmed enhanced structural stability of the inhibitor-protease complex. The compound specifically blocked polyprotein processing during viral replication, without affecting adsorption or entry. The structural conservation of coronaviral 3CLpro underscores the broader applicability of our screening strategy and optimization framework. While moderate potency limits immediate clinical utility, Y041-1672 provides a foundational template for rational optimization, such as introducing halogen substituents or stabilizing labile groups. This study establishes a multidisciplinary approach integrating virtual screening, molecular dynamics simulations, and biological validation, offering critical insights for coronavirus protease inhibition and antiviral drug development.

Materials and methods

Small molecule processing

The small molecule compound database employed in this study was the commercial ChemDiv (ChemDiv Inc., USA), which contained 1,613,879 small molecule compounds. The preprocessing of the database typically involves three main steps as following. First, the small molecule libraries were converted from SDF to PDB format using the Raccoon program in MGLTools software (version 1.5.7). Then, batch energy optimization of the small molecules in PDB format using the MMFF94 force field via OpenBabelGUI software. Finally, hydrogen atoms were added to the energy-optimized molecules, followed by the identification of torsional bonds using OpenBabel GUI, ultimately yielding files in PDBQT format. The resulting PDBQT files can be directly used for docking in AutoDockVina. Ligand-protein interaction presentation was performed using PyMol software (version 2.5.2). Small molecule format conversion was performed using OpenBabel GUI software (version 3.1.1). AutoDock Vina was implemented on a cloud-based computational platform. The target protein structure and a natural product small-molecule library were uploaded to the server. A predefined docking pocket configuration was established with the following spatial parameters: center coordinates (x ​= ​18.02, y ​= ​22.827, z ​= ​−2.394) and search space dimensions (x ​= ​23.04, y ​= ​29.39, z ​= ​25.42). Additional docking parameters included an exhaustiveness value of 10. Virtual screening was executed through the command-line interface using the vina_screen_local.sh script.

Molecular dynamics (MD) simulation

Molecular dynamics (MD) simulations were conducted to evaluate the stability of the docking complex between PEDV 3CLpro (PDB: 5HYO) and small molecule inhibitors in ionic solvents. All-atom simulations were performed using GROMACS 2022.2 (Groningen Machine for Chemical Simulations) (Van Der Spoel et al., 2005). Prior to the MD simulations, the CHARMM-GUI (www.charmm-gui.org) web server was used to generate the initial input parameters (Jo et al., 2008), applying the CHARMM36 force field (Brooks et al., 2009; Huang and MacKerell, 2013; Lee et al., 2016; Ziada et al., 2022; Faris et al., 2024). Simulations were conducted at pH 7.0.

Each complex was placed in a rectangular simulation box, solvated with transferable interatomic potential with three points model (TI3P) water molecules, and neutralized with Na⁺ and Cl⁻ ions to achieve a salt concentration of 0.15 ​M via Monte Carlo ion displacement. Energy minimization was performed using the steepest descent algorithm, with a maximum of 50,000 steps and a maximum force threshold of 10.0 ​kJ/mol. The temperature and pressure were set to 310 ​K and 1.0 ​bar, respectively. NVT (constant volume and temperature) and NPT (constant pressure and temperature) equilibration protocols were conducted over two stages of 10 ns each. Following equilibration, production MD simulations were performed for 100 ns.

To assess the structural stability of the complexes, the trajectories were analyzed using various metrics, including the root mean square deviation (RMSD), radius of gyration (Rg), and root mean square fluctuation (RMSF).

Expression of PEDV 3CLpro

E coli. BL21 (DE3) was transformed with pET30a (+) carrying the 3CLpro gene. The bacteria were incubated at 27 ​°C for 8 ​h with 1 ​mM IPTG. The cells were broken by high-pressure ultrasonication and centrifuged to collect the supernatant. The protein was purified by nickel affinity chromatography and frozen at −80 ​°C after adding 5% glycerol. The samples were then resolved with SDS-PAGE and transferred to polyvinylidene difluoride membranes (Millipore, USA) for Western blotting. An anti-His antibody (Proteintech, China) was used to analyze the expression of PEDV 3CLpro in Western blotting.

Enzymatic activity assay of 3CLpro

A fluorescence resonance energy transfer (FRET) approach was utilized to conduct the 3CLpro activity assay. A peptide substrate encompassing the cleavage site of PEDV 3CLpro was synthesized. The fluorescent donor (Edans) and quenching acceptor (Dabcyl) moieties were respectively appended to the N-terminus and C-terminus of the synthesized peptide (Dabcyl-YNSTLQ↓AGLRKME-Edans, the arrow position shows the 3CLpro cleavage site).

To evaluate the enzyme activity, 10 ​μL of PEDV 3CLpro with a concentration of 2.74 ​mg/mL and 10 ​μL of the peptide substrate with a concentration of 1 ​mg/mL were diluted with the substrate buffer (20 ​mM HEPES, 50 ​mM NaCl, 0.4 ​mM EDTA, 4 ​mM DTT, 30% glycerol, pH 8.0) and then mixed in a black, opaque 384-well plate, resulting in a final concentration of 10 ​μM for each component. A negative control was prepared by excluding the protease, and each sample was set up with three replicates. The reaction mixture was incubated at 37 ​°C for 1 ​h, and the fluorescence was measured by a fluorescence plate reader, with excitation at 355 ​nm and emission monitored at 480 ​nm.

Inhibition assay of the compound inhibitors on PEDV 3CLpro

Compounds which purchased from Chemdiv Company were dissolved in DMSO to obtain a 10 ​mM stock solution. Serial dilutions were then prepared in DMSO, ranging from 100 ​μM to 0.78125 ​μM. Different concentrations of compounds (2 ​μL) were incubated with 8 ​μL of PEDV 3CLpro, which was adjusted to a final concentration of 5 ​μM. After incubating at room temperature for 30 ​min, 10 ​μL of the substrate was added in to the mixture to a final concentration of 5 ​μM. The emission of 480 ​nm with an excitation of 340 ​nm was immediately measured as F0. After 1 ​h incubation at 37 ​°C, the fluorescent value was measured again and recorded as F1. The inhibition rate was calculated as [1−(F1–F0)/(F1nc−F0nc)] ​× ​100%, where the F1nc and F0nc was represent the values of non-compound control. The IC₅₀ of each compound was calculated by a nonlinear fitting calculation.

Cytotoxicity assay

African green monkey kidney (Vero) cells (ATCC, CCL-81) were seeded on a 96-well plate and cultured for 24 ​h. The compounds were serially diluted from 1 ​mM to 1.953 ​μM and then added to the cells for a further incubation period of 36 ​h. After the incubation, the culture supernatant was removed, and the cells were washed three times with PBS. CCK-8 (Life-iLab, China) reagent, diluted with serum-free DMEM (Gibco, USA), was added to the cells, and the cells were incubated at 37 ​°C for 1 ​h. The absorbance of OD₄₅₀ nm was measured using Victor Nivo 3S. The survival rate was calculated as (OD_positive ​− ​OD_blank)/(OD_control ​− ​OD_blank) ​× ​100%, and the CC₅₀ (cytotoxic concentration) was calculated by non-linear fit method.

Inhibition of the compound on virus replication

To evaluate the antiviral efficacy, the compound was serially diluted (100–6.25 ​μM) in serum-free medium. Vero cells were cultured overnight in 12-well plates and infected with PEDV (MOI = 0.01). After 1 ​h adsorption, unbound virus was removed by PBS washing, and compound-supplemented maintenance medium were treated for 24 ​h. The viral RNA in lytic cells was analyzed by RT-qPCR using the specific primers for PEDV (qPCR-N-F: GAGGGTGTTTTCTGGGTTG; qPCR-N-R: CGTGAAGTAGGAGGTGTGTTAG).

Plaque assay was employed to quantify infectious particles. Cells were cultured overnight in 12-well plates and infected with PEDV (MOI = 0.01). After adsorption for 1 ​h, unbound virus was removed by PBS washing, and compound-supplemented maintenance medium were added. After 30 ​h incubation, supernatants were harvested, centrifuged, and viral titers quantified via plaque assay (PFU/mL).

The addition time was determined to measure the antiviral effect of the compounds. Vero cells were cultured in a 12-well plate and infected with PEDV, the compound was added to cells at a final concentration of 50 ​μM ​at different times with respect to infection (Fig. 7A). For time-course analysis, cells at multiple infection stages were harvested at 30 ​h post-infection and the levels of viral RNA was quantitative by RT-qPCR. The relative expression of each designated gene was calculated by the 2^−△△Ct method and normalized to GAPDH (qPCR-GAPDH-F: CAGCCTCAAGATCGTCAGCA; qPCR-GAPDH-R: CGTGGACTGTGGTCATGAGT). PEDV-N monoclonal antibody was used to quantitatively detect the expression of the virus at the protein level.

In order to detect the concentration for 50% of maximal effect (EC₅₀) of the compound, the compound was added after infection with the virus and cell viability was detected with CCK-8. Vero cells were cultured overnight in 96-well plates and infected with PEDV for 1 ​h. After washing for 3 times, culture medium containing compounds of different concentrations was added and continued for 48 ​h. CCK-8 was added and incubated at 37 ​°C for 1 ​h, and the absorbance at OD₄₅₀nm was measured with Victor Nivo 3S. The survival rate was calculated as (OD_positive ​− ​OD_blank)/(OD_control ​− ​OD_blank) ​× ​100%, and the EC₅₀ was calculated by non-linear fit method.

For indirect immunofluorescence analysis, Vero cells infected with PEDV were treated with varying concentrations of test compounds and maintained in culture for 30 ​h. Cells were triple-washed with phosphate-buffered saline (PBS, pH 7.4), followed by blocking with 5% (w/v) skim milk at room temperature for 1 ​h. After three additional PBS washes, cells were incubated with a monoclonal anti-PEDV-N primary antibody (1:1000) for 1 ​h at room temperature. Unbound antibodies were removed by repeated PBS washing, and samples were subsequently exposed to fluorescein isothiocyanate (FITC)-conjugated sheep anti-mouse IgG secondary antibody (1:1000) for 40 ​min under light-protected conditions. Nuclear counterstaining was performed using 4′,6-diamidino-2-phenylindole (DAPI) for 5 ​min.

Inhibition to other coronaviruses

To verify the inhibitory activity against other coronaviruses, LLC-PK1 cells (ATCC, CL-101) were infected with porcine deltacoronavirus (PDCoV) (MOI = 0.01). The medium was replaced with DMEM containing compound Y041-1672 after 2 ​h of infection. After culture for another 24 ​h, cells were collected and subjected to Western blotting and IFA using PDCoV-N monoclonal antibody as primary antibody.

Statistical analysis

Each test was conducted in triplicate for all experiments. The data are expressed as mean ​± ​standard deviation. Statistical analyses were performed using Student’s t-test. ∗P < 0.05; ∗∗P < 0.01; ∗∗∗P < 0.001.

Data availability

All data relevant to the study are included in the article.

Ethics statement

This article does not contain any studies with human or animal subjects performed by any of the authors.

Author contributions

Ang Tian: conceptualization; data curation; formal analysis; investigation; methodology; writing-original draft. Shutong Shi: validation; formal analysis. Siying Zou: data curation. Shuaiyin Guan: software; methodology. Hao Wu: supervision; investigation. Zhen Li: data curation. Huanchun Chen: technology and resources; supervision. Yunfeng Song: funding acquisition; writing-review and editing; supervision; project administration.

Conflict of interest

The authors declare that there are no conflicts of interest.