The PEDV YN13 (GenBank accession No. KT021228), a PEDV variant strain, was isolated from the intestine of a piglet with severe diarrhea and was passaged for 13 generations. The DR13-GFP strain was a recombinant strain which was constructed by replacing the ORF3 gene of the DR13 strain (GenBank accession no. JQ023161) with the GFP gene using the reverse genetic technique. Vero and LLC-PK1 cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) containing 8% fetal bovine serum (FBS) and were incubated at 37°C with 5% CO₂. The flavonoid library containing 270 compounds (MedChemExpress, Shanghai, China) was used for screening for the inhibitors of PEDV 3CL^pro. Detailed information on the flavonoid library is shown in the Supplementary File S1.

The PEDV 3CL^pro expression vector was obtained by cloning the PEDV nsp5 gene into the Pet-28a (+) vector. This vector was then transformed into the E.coli BL21 (DE3) strain to create recombinant bacteria. The bacteria were cultured at 37°C in Luria Bertani medium until the optical density value at a wavelength of 600 nm (OD₆₀₀) reaching 0.6–0.8. After that, 0.8 mM isopropyl β-D-thiogalactoside (IPTG) was added to induce protein expression. The bacteria were harvested after incubation at 37°C for 24 h, resuspended in PBS and disrupted. The supernatant was filtered and loaded onto a nickel iminodiacetic acid NUPharose fast flow (Nuptec, Hangzhou, China) to purify the protein. Subsequently, the N and C terminal His-tagged protein was eluted using elution buffer. The protein eluent was desalted by dialysis in PBS buffer and was concentrated by ultrafiltration.

In order to evaluate the enzymatic activity of the purified protease, a peptide substrate whose sequence is Dabcyl-YNSTLQ↓AGLRKM-E-Edans was synthesised based on the cleavage site of PEDV 3CL^pro. The two fluorophores, namely Dabcyl and Edans, formed a quenching pair and exhibited FRET within the peptide. The enzyme reaction system (100 μL) used to measure PEDV 3CL^pro activity consisted of 10 μM FRET substrate, 20 mM Tris/HCl buffer (pH 7.5), and various concentrations of the purified 3CL^pro. Fluorescence emitted by the reactions was continuously monitored for a duration of 90 min, with excitation at 340 nm and emission at 485 nm, using fluorescence plate reader reactions with a multimode reader platform (Molecular Devices, San Jose, CA, United States). For the purpose of screening PEDV 3CL^pro inhibitors, reaction systems (100 μL) were prepared, utilizing 10 μM FRET substrate, 20 μg/mL 3CL^pro, 20 μM natural molecule, and 20 mM Tris/ HCl buffer (pH 7.5). Each natural molecule presented in the flavonoid library was initially incubated with 3CL^pro for a period of 30 min at a temperature of 37°C. Subsequently, the substrate was added to the mixture and a reaction system without any natural molecule was set up as a control group. The fluorescence of reactions was monitored according to the method mentioned above. The inhibition ratio was determined by employing the following equation: Percentage of inhibition (%) = 100× [1 − fluorescence of the experimental group (90–0 min)/fluorescence of the control group (90–0 min)].

The crystal structures of the chosen compounds (baicalein and baicalin) as well as PEDV 3CL^pro (PDB: 4XFQ) were downloaded from the website https://www.rcsb.org/. Autodock 4 software was used to generate the homology models of PEDV 3CL^pro with the selected molecules. Grid maps of 19 × 19 × 19 grid points (center: x = 20.254, y = 11.946, z = 5.239; spacing = 1.000) were produced, covering the active site and the binding pocket of PEDV 3CL^pro. To ensure accurate docking, the following parameters were set: a population size of 150, a maximum of energy evaluations of 5,000,000, a maximum of generations of 27,000, a mutation rate of 0.02, and a crossover rate of 0.8. Ultimately, a total of 10 docked conformations were produced and the conformation displaying the lowest binding energy was selected for the analysis of the interaction between the chosen compounds and PEDV 3CL^pro. PyMOL software was used to analyze the docking results to obtain the most likely binding pattern.

The MD simulations of PEDV 3CL^pro-baicalein and PEDV 3CL^pro-baicalin complexes were carried out at 100 ns using GROMACS version 2018.1 (Páll et al., 2020). PRODRG was utilized to generate the ligand topology, while the protein topology was created using pdb2gmx. Solvation was carried out utilizing the simple point charge water model. In order to produce the topologies for complexes formed by proteins and ligands, the topologies of the proteins and ligands were combined. Boxes with a cube shape were generated and the complexes were positioned inside them. The electroneutrality of the system was maintained by incorporating Na⁺ and Cl⁻ ions. Additionally, the steepest descent minimization algorithm was employed to accomplish energy minimization of both the protein and protein-ligand complexes. Equilibration of the systems was also achieved through NPT and NVT simulations. Finally, all systems underwent a simulation time of 100 ns, with coordinates saved at 2 fs intervals. The trajectories of the protein and protein-ligand complexes were assessed for structural stability analysis using root-mean-square deviation (RMSD).

To measure the cytotoxicity of baicalein and baicalin on Vero and LLC-PK1 cells, a Cell Counting Kit-8 (Beyotime Biotechnol, Shanghai, China) was utilized. To initiate the experiment, cells were seeded in a 96-well plate and were allowed to incubate at a temperature of 37°C with 5% CO₂ until a complete monolayer was formed. Subsequently, the cells were treated with baicalein or baicalin at different concentrations at 37°C with 5% CO₂ for 48 h. One hundred microliter of 10% CCK-8 solution (10% CCK-8 reagent in DMEM) was added to each well after being washed three times with PBS. The blank wells with 10% CCK-8 solution were set as negative control group. After 2 h in the incubator, the optical density values in each well were measured using a microplate reader at a wavelength of 450 nm (OD₄₅₀). Cell viability was determined using the formula: cell viability = [As-Ab]/[Ac-Ab]. The OD₄₅₀ value of the wells in the experimental groups (containing baicalein or baicalin treated cells and CCK8 solution) was denoted as As. The OD₄₅₀ value of the mock wells (containing only CCK8 solution) was represented as Ab. Finally, the OD₄₅₀ value of the wells in control group (containing cells and CCK8 solution) was denoted as Ac.

For YN13 strain, the cells (Vero and LLC-PK1) were grown to approximately 90% confluence in 12-well culture plates. Then the cells were infected with PEDV YN13 strain at a multiplicity of infection (MOI) of 0.001 with 8 μg/mL trypsin in the presence of baicalein or baicalin of different concentrations (0 μM, 2.5 μM, 5 μM, 10 μM, 20 μM and 40 μM for the Vero cells; 0 μM, 5 μM, 10 μM and 20 μM for the LLC-PK1 cells) for 24 h; For the DR13-GFP strain, Vero cells were grown to approximately 90% confluence in 12-well culture plates. Then the cells were infected with the DR13-GFP strain at a MOI of 0.001 in the presence of baicalein or baicalin of different concentrations (0 μM, 1.25 μM, 2.5 μM, 5 μM, 10 μM, 20 μM and 40 μM) for 36 h.

Then, the cells were collected and subjected to extract RNA using RNAiso Plus Reagent (TakaRa, Tokyo, Japan) following the manufacturer’s instructions. The cDNA was obtained by RT-PCR using the Primescript™ RT reagent Kit With gDNA Eraser (TakaRa, Tokyo, Japan) according to the manufacturer’s instructions. The Real-Time RT-PCR was performed using the SYBR Premix ExTaq (TakaRa, Tokyo, Japan) on an Applied Biosystems 7,500 Fast Real-Time PCR System (Life technologies, Marsiling, Singapore). Porcine ribosomal protein L4 (RPL4) and homo sapiens hypoxanthine phosphoribosyltransferase 1 (HPRT1) were used as the housekeeping gene for LLC-PK1 and Vero cells, respectively. The change of PEDV M gene expression was calculated using the 2^−ΔΔCT method. Primers used in the present study were listed in Table 1.

For YN13 strain, the cells (Vero and LLC-PK1) were grown to approximately 90% confluence in 12-well culture plates. Then the cells were infected with PEDV YN13 strain at a MOI of 0.001 with 8 μg/mL trypsin in the presence of baicalein or baicalin of different concentrations (0 μM, 2.5 μM, 5 μM and 10 μM for the Vero cells; 0 μM, 5 μM, 10 μM and 20 μM for the LLC-PK1 cells) for 30 h; For DR13-GFP strain, Vero cells were grown to approximately 90% confluence in 12-well culture plates. Then the cells were infected with the DR13-GFP strain at a MOI of 0.001 in the presence of baicalein or baicalin of different concentrations (0 μM, 2.5 μM, 5 μM and 10 μM) for 36 h. Then the cells and the supernatant were collected and placed in −80°C. After repeated freezing and thawing three times, the supernatant was harvested by centrifuge at 12,000 rpm for 10 min at 4°C and then used for TCID₅₀ assays according to the following steps.

Vero cells seeded in 96-well plates were cultured incubated at 37°C with 5% CO₂ until the monolayer was covered. Virus samples were 10-times serially diluted before they were inoculated on the confluent cell monolayers. DR13-GFP strain samples were diluted with DMEM, while YN13 strain samples were diluted with DMEM containing 8 μg/mL of trypsin. Eight wells were inoculated with 100 μL of each dilution, and plates were incubated at 37°C with 5% CO₂ for 2 days. For the YN13 strain, wells with syncytium formation, the specific cytopathic effect (CPE) of YN13 infection, were identified as PEDV-positive. For the DR13-GFP strain, wells with green signal under a fluorescence microscope were considered as PEDV-positive. The titration was calculated in the form of TCID₅₀ using the Reed–Muench method established by Reed and Muench (1938).

Vero cells were grown to approximately 90% confluence in 48-well culture plates. Then the cells were infected with PEDV YN13 strain at a MOI of 0.001 with 8 μg/mL trypsin in the presence of baicalein or baicalin of different concentrations (0 μM, 1.25 μM, 2.5 μM and 5 μM) for 24 h. After that, cells were gently rinsed with PBS and then fixed with 4% paraformaldehyde at room temperature for a duration of 10 min. The fixed cells underwent three subsequent washes with PBS, followed by treatment with pre-chilled methanol at a temperature of −20°C for a period of 15 min. Blocking was carried out by treating the cells with a PBS buffer that contained 5% bovine serum albumin (BSA) for 30 min at room temperature. The cells were then subjected to an incubation period of 1 h at 37°C with the PEDV S protein monoclonal antibody (diluted 1:500 in PBS), followed by three PBS washes. Afterwards, the cells were incubated at 37°C for 45 min with the Alexa 488-labeled anti-mouse antibody (Antgene, Wuhan, China). DAPI dihydrochloride (Beyotime Biotechnol, Shanghai, China) solutions were added to each well to stain the cell nuclei at room temperature for 5 min. After being washed three times with PBS, the results of IFA were observed and acquired using a fluorescence microscope.

Vero cells were grown to approximately 90% confluence in 48-well culture plates. Then the cells were infected with PEDV DR13-GFP strain at a MOI of 0.001 in the presence of baicalein or baicalin of different concentrations (0 μM, 2.5 μM, 5 μM and 10 μM) for 36 h. After that, cells were gently rinsed with PBS and then incubated with hoechst 33258 (Beyotime Biotechnol, Shanghai, China) solution at 37°C for 5 min to stain the cell nucleus. After being washed three times with PBS, the GFP-expression were observed and acquired using a fluorescence microscope.

Vero cells were infected with the PEDV YN13 strain at a MOI of 0.001. Then, 20 μM of baicalein and baicalin were added into the culture medium at different time periods of PEDV infection: −2–24 h, 0–24 h, 2–24 h, 4–24 h and 6–24 h. At 24 h post infection, RNA was extracted from the cells and PEDV M gene was detected using real-time quantitative RT-PCR.

The half maximal inhibitory concentration (IC₅₀) was determined using SPSS 17.0 software. The experiments were conducted in triplicate, and the results were presented as means ± standard deviation (SD). To assess the statistical significance, student’s t-test was performed using SPSS 17.0 software. *, p < 0.05 (considered significant compared with control group); **, p < 0.01 (considered highly significant); ***, p < 0.001 (considered extremely significant).

PEDV 3CL^pro was expressed using a prokaryotic expression system and was purified through affinity chromatography using a nickel column purification column. As shown in Figure 1A, PEDV 3CL^pro was expressed in both the supernatant and the inclusion bodies. The supernatant was then subjected to affinity chromatography, resulting in the production of a relatively pure protein after purification. To assess the cleavage activity of the purified protein, a fluorogenic peptide substrate was designed based on the conserved cleavage site recognized by PEDV 3CL^pro and was introduced in the FRET assays. The fluorescence increased in a time and dose-dependent manner, indicating that the purified protein effectively cleaved the fluorogenic peptide substrate (Figure 1B). These findings suggest that the constructed FRET assay can be utilized to test the enzymatic activity and to screen for the inhibitors of PEDV 3CL^pro. A flavonoid library was employed for screening PEDV 3CL^pro inhibitors, and the molecules with inhibitory rates above 30% are presented in Table 2.

Since our aim was to search for natural molecules to control PED in the pig industry, the relatively low-cost molecules were selected to be used in the following studies. As a result, baicalein and baicalin were chosen. These two flavonoids have structural similarities, with baicalin consisting of a baicalein group and a glucuronic acid group (Figure 2A). To confirm their inhibitory effects on PEDV 3CL^pro, different concentrations of the molecules were added to the FRET systems for detecting PEDV 3CL^pro enzymatic activity. The results indicated that baicalein and baicalin could inhibit PEDV 3CL^pro enzymatic activity in a dose-dependent manner (Figure 2B). The IC₅₀ for baicalein and baicalin were found to be 9.50 ± 1.02 μM and 65.80 ± 6.57 μM, respectively.

The interaction between each selected molecule and PEDV 3CL^pro was analyzed using molecular simulation techniques. As shown in Figure 3, baicalein could interact with His162, Ile140, Gly142, Ala143, and His41 of PEDV 3CL^pro, while baicalin could only interact with Asn141 and Glu165.

To verify the findings of the docking results and gain a deeper understanding of the stability of the ligand-protein complex, MD simulations were conducted for each complex formed by PEDV 3CL^pro and its inhibitors (baicalein and baicalin). The results of the MD simulations were analyzed based on the RMSD. Figure 4 clearly demonstrated that both two systems maintained equilibrium throughout the entire simulation duration, with particular stability observed in the PEDV 3CL^pro-baicalein complex. The average RMSD values for the PEDV 3CL^pro-baicalein and PEDV 3CL^pro-baicalin complexes were calculated as 0.1829 nm and 0.1824 nm, respectively. Since the fluctuation amplitude of the RMSD curve was inversely related to the stability of the protein, the above results indicated that the complexes formed by PEDV 3CL^pro with baicalein and baicalin had high stability.

In order to investigate the effects of baicalein and baicalin on PEDV infection, it was necessary to determine the safe concentrations of them for the host cells. The viability of Vero or LLC-PK1 cells was assessed after treatment with different concentrations of baicalein or baicalin. As shown in Figure 5, the cell viability was above 95% when treated Vero and LLC-PK1 cells with baicalein and baicalin at a concentration of no more than 200 μM.

Real-time RT-PCR was employed to evaluate the antiviral properties of baicalein and baicalin on two PEDV strains. Baicalein displayed a dose-dependent inhibition of viral gene production in the YN13 and DR13-GFP strains of PEDV, with an IC₅₀ of 3.28 ± 0.24 and 1.65 ± 0.16 μM, respectively (Figure 6A). Similarly, baicalin also demonstrated a dose-dependent inhibitory impact on viral gene production in the YN13 and DR13-GFP strains of PEDV, with an IC₅₀ value of 9.65 ± 1.51 and 11.53 ± 1.45 μM, respectively (Figure 6A). Nonetheless, baicalein exhibited a stronger inhibitory effect than baicalin. The selectivity index (SI) of baicalein was found to be greater than 100 for both PEDV YN13 strain and DR13-GFP strain, while baicalin exhibited a selectivity index greater than 30 for both strains (Supplementary File S2). These findings suggest the potential of baicalein and baicalin as promising anti-PEDV agents.

Confirmatory evidence of the anti-PEDV effect of baicalein and baicalin in Vero cells was obtained using a TCID₅₀ assay. Figure 6B depicts the significant reductions in titer of the YN13 and DR13-GFP strain that was achieved through the administration of baicalein or baicalin, in a dose-dependent manner. Moreover, the IFA was conducted to manifest the inhibitory effects of these two compounds on the quantities of YN13-infected cells. The IFA results demonstrated a dose-dependent reduction in the number of YN13-infected cells upon treatment with both baicalein and baicalin. Furthermore, the baicalein and baicalin-treated group exhibited a decrease in the size and number of syncytia, which is the characteristic CPE of YN13-infected cells (Figure 6C). This provided additional substantiation of their inhibitory effects on YN13 infection. DR13-GFP, a lab-engineered recombinant virus designed by substituting the ORF3 gene of DR13 strains with the GFP gene via the reverse genetic technique, allowed for the expression of GFP as an indicator of the inhibitory effects of anti-PEDV compounds on the quantities of viral-infected cells. Consistent with the IFA results, the quantities of the DR13-GFP strain infected cells decreased with increasing concentrations of baicalein and baicalin (Figure 6C).

To further confirm the anti-PEDV effects of baicalein and baicalin, another cell line, LLC-PK1, was selected. The results in the LLC-PK1 cells were similar to those in Vero cells. Both baicalein and baicalin were able to effectively suppress the replication of the YN13 gene and reduce the viral titers in a dose-dependent manner (Figure 7). Based on the data presented above, it can be concluded that baicalein and baicalin are effective inhibitors of PEDV infection in vitro.

Natural products seem to exhibit their antiviral effects through various mechanisms. To further explore the potential mechanisms of baicalein and baicalin in inhibiting PEDV infection in vitro, a time-of-addition experiment was conducted. Figure 8 demonstrates a significant difference between the −2-24 h group and the 0–24 h group, suggesting that pre-incubation with baicalein and baicalin prior to PEDV infection enhances their antiviral effects in vitro. It is widely recognized that viral attachment and invasion primarily occur within the first 2 h after PEDV infection. Notably, the baicalein treated group exhibited a significant difference between the 0–24 h group and the 2–24 h group, indicating that baicalein may influence the attachment or invasion step of PEDV. Conversely, baicalin treatment did not appear to affect PEDV attachment or invasion. Furthermore, the addition of baicalein and baicalin after PEDV infection (2–24 h, 4–24 h, and 6–24 h) also exhibited significant anti-PEDV effects, suggesting that these two molecules may exert their antiviral effects by interfering with viral replication within the host cells.