Introduction

The pandemic COVID-19 caused by SARS-CoV-2 has threatened public health all over the world for more than 3 years. As of May 10, 2023, more than 765 million cases have been confirmed, and over 6.92 million lethal cases have been reported by the World Health Organization (WHO) [1]. Due to the widespread practice of traditional Chinese medicine (TCM), the war against COVID-19 was equipped with the confident weapon to protect people’s health [2]. Numerous clinical evidences showed that TCMs relieve symptoms of COVID-19, block the worse condition, and reduce mortality [3, 4]. Xuanfei Baidu formula (XFBD), as one of the effective formulae in “three medicines and three prescriptions”, is recommended for treating COVID-19 patients in the 3rd–9th editions of the diagnosis and treatment protocols released by the National Health Commission of the People’s Republic of China. In March 2021, XFBD granules has been approved by National Medical Products Administration (NMPA) for treating COVID-19 patients in mild or moderate cases, which can relieve fever, cough, fatigue and other clinical symptoms, prevent the mild to the severe progression, and decrease mortality in clinical practice [5, 6]. Composed of 13 herbs (Additional file 1: Table S1), XFBD is modified from classical TCM formulae, such as Maxing Shigan decoction (MXSG), Maxing Yigan decoction (MXYG), Qianjin Weijing decoction (QJWJ), and Tingli Dazao Xiefei decoction (TLDZXF) [7], with indications of dispelling lung dampness, clearing heat and penetration evil, and purifying lungs and detoxification.

In previous studies, the network pharmacology analysis showed that the efficacy of XFBD are possibly originated from antiviral infection and recovery of lung injury [7, 8]. The pharmacological studies have verified that XFBD can suppress inflammatory response, and restore the balance of energy metabolism, which also can be effective in inhibition of macrophage infiltration and pulmonary fibrosis via IL-6/STAT3 signaling pathway [9], and reduction of acute lung injury by regulating infiltration of neutrophils and macrophages via PD-1/IL-17A pathway [10]. However, as an effective formula, the bioactive compounds in XFBD against SARS-CoV-2 were yet unveiled.

SARS-CoV-2 is a novel positive-sense single-stranded RNA virus belonging to the subfamily β-coronavirus. The virus genomes mainly encode four structural proteins and sixteen nonstructural proteins that control the process of virus invasion, replication, assembly, and release. As the key protease to orchestrate the replication and transcription of SARS-CoV-2, the main protease (Mpro) is about 33.8 kDa, which consists of a polypeptide chain of 306 amino acids [11]. Given its low homology with mammalian proteins, Mpro is a satisfactory target for the development of chemical inhibitors with broad-spectrum antiviral activity [12]. Up to now, Paxlovid, S-217622, and DC402234 were reported as SARS-CoV-2 Mpro inhibitors in clinical trial phases or market [13], showing the dawn for the era of new inhibitors to be discovered.

In this study, we focus on Mpro to explore the bioactive compounds and related antiviral mechanisms from the compounds library built from XFBD. To test the intermolecular interaction, the integrated methods have proved that acteoside, licochalcone B, licochalcone D, linoleic acid, and physcion were identified as the potential active compounds by fluorescence resonance energy transfer (FRET) and surface plasmon resonance (SPR). And the interaction of Mpro-acteoside was further confirmed by isothermal titration calorimetry (ITC) to determine the thermodynamic parameters. Omicron BA.1.1 and BA.2.3 variants were employed to demonstrate the antiviral activity of acteoside in vitro. Molecular docking and molecular dynamics (MD) simulation suggested that the hydrogen bond between acteoside and Mpro was considered as the key interaction, which was demonstrated by nuclear magnetic resonance (NMR). Our study elucidated and evaluated the active compounds in XFBD against COVID-19 via inhibiting SARS-Cov-2 Mpro activity, which provides insights for discovering the active compounds from the tested formula and unveiling corresponding antiviral mechanism (Fig. 1).

Fig. 1
figure 1

The flow chart for discovery and evaluation of active compounds against SARS-CoV-2 Mpro

Material and methods

Reagents and chemicals

Based on the preliminary screening results by network pharmacology analysis (Additional file 1: Fig. S1, S2, S3, S4, S5), the compounds library was built, containing 83 compounds, which were obtained from Shanghai Yuanye Biotechnology Co., Ltd. (Shanghai, China) with HPLC purity > 98% (Additional file 1: Table S2). PBS, Tris, DMSO, and imidazole were acquired from Solarbio (Beijing, China) and other commercial suppliers. The sensor Chip (CM5), acetate (pH 4.0), PBS-P, and the amine coupling kit were purchased from GE Healthcare Life Sciences (Boston, USA).

Cloning, protein expression and purification of SARS-CoV-2 M pro

The plasmid, which encodes the full-length gene of SARS-CoV-2 Mpro (NC_045512), was transferred into Escherichia coli BL21 (DE3), cultured in Luria broth medium with 100 µg/mL ampicillin at 37 ℃ for 6–8 h. Then, 500 µM IPTG was added into the cell culture flasks to induce the expression of Mpro for 16–20 h at 16 ℃ [14]. The cells were enriched and crushed through the high-pressure homogenizer, whose supernatant was collected and purified through Ni–NTA affinity column. The buffer at pH 8.0 (20 mM Tris–HCl, 150 mM NaCl, and 5% Glycerol) containing 20 or 300 mM imidazole was used to remove debris and elute Mpro, respectively. The C-terminal 6 × His tag of Mpro was removed by human rhinovirus 3C protease, which was further subjected to purification by size exclusion chromatography (Superdex 200 Increase 10/300 GL). The purified Mpro was checked by SDS-PAGE with purity > 95%. At last, Mpro was preserved at – 80 ℃ in buffer (10 mM Tris–HCl, pH 8.0) [15].

High throughput screening of active compounds against M pro

FRET screening was carried out for identifying the inhibitors of SARS-CoV-2 Mpro. The fluorogenic substrate of Mpro (MCA-AVLQSGFRKK-Lys (Dnp) -Lys-NH2) was purchased from GL Biochem Ltd. (Shanghai, China). The screening system was established by employing Mpro (200 nM), substrate (20 µM), and the tested compounds with the different concentrations from 0.039 to 160 µM. The excitation and emission wavelength were set at 320 and 405 nm, separately. The data was analyzed with Graphpad prism 9.0 and all experiments were performed in triplicates [16].

The ligand-binding affinity with M pro

SPR assay. Biacore T200 system was used for SPR assay. Mpro was conjugated to a sensor Chip (CM5) via amino coupling reaction [17]. The tested compounds were diluted with the running buffer (1.05 × PBS-P, 5% DMSO) to reach the working concentration of 0.977–500 µM, being injected into the channels at a flow rate of 30 µL/min. The coupling time and the dissociation time were respectively set at 60 s and 120 s or 180 s. The data were analyzed for fitting the equilibrium dissociation constant (KD) [18, 19].

ITC measurement. ITC is the standard assay for measuring affinity interactions, which determines the thermodynamic parameters with label-free outputs [20]. The tested compounds and Mpro were dissolved by the same buffer at the suitable concentration, respectively. The cell filled with water as the reference, and the sample solution of the tested compound (500 µM) was titrated to the sample cell filled with Mpro (40 µM) in temporal sequence. The relevant parameters were monitored and documented, such as the thermodynamic equilibrium constant (K), enthalpy variations (ΔH), entropy variation (ΔS), and binding stoichiometry (N) [21].

Cell culture and cytotoxicity assay

Vero E6 (ATCC, CRL-1586) cells (1 × 104 cells/well) were maintained in a 96-well plate for 12 h. All cells were cultured in complete medium composed of Dulbecco’s Modified Eagle Medium (DMEM) (Gibco, USA), 10% (v/v) serum, 100 U/mL penicillin, and 100 µg/mL streptomycin. The diluted solutions of the tested compounds (200 µL) at the concentrations of 100, 80, 60, 40, 20, 10, and 5 µM were added to wells and incubated for another 24 or 48 h, respectively. At the terminus period of incubation, 50 µL MTT (2.5 mg/mL) was added to each well and continuously incubated at 37 ℃ for 4 h. After disposal of the liquid supernatant, 150 µL DMSO was added to each well, whose absorbance was detected at 490 nm via Microplate Reader (TECAN SunriseTM, China). Cytotoxicity was calculated as follows: Cell viability (%) = (ODexperiment-ODBlank)/(ODcontrol-ODBlank) × 100%. The data were analyzed with GraphPad Prism 9.0. All experiments were performed in triplicates.

Antiviral activities assay

Vero E6 (ATCC, CRL-1586) cells were seeded in a 96-well plate at 2 × 104 cells/well in a fully humidified cultivation chamber at 37 ℃ containing 5% CO2 and bacteria-free condition. Omicron BA.1.1 and Omicron BA.2.3 variants were respectively added and propagated in Vero E6 cells. After inoculation for 3 days, a cytopathic effect (CPE) was scored by Celigo Image Cytometer, and the Reed-Muench formula was used to calculate the 50% tissue culture infectious dose (TCID50) as the reference.

The solutions of the tested compound with the different concentrations were respectively mixed with Omicron BA.1.1 and BA.2.3 variants (MOI = 0.01). Then, 200 μL mixture was added to a monolayer Vero E6 cells. After incubation for 48 h, CPE was scored, from which the value of half-maximal effective concentration (EC50) and the antiviral activity of the tested compounds were evaluated [22]. All experiments were performed at BSL-3 level laboratory in Guangzhou Customs Inspection and Quarantine Technology Center (IQTC).

Molecular docking

The crystal structure of SARS-CoV-2 Mpro (PDB ID: 6LU7) [23] and Omicron Mpro (PDB ID: 7TLL) was respectively downloaded from the Protein Data Bank database (http://www.rcsb.org/pdb/), which is acquired from the crystal complex structure of Mpro-N3 and Mpro-nirmatrelvir. The 3D structures of the tested natural compounds in our compounds library were downloaded from ChemSpider (https://www.chemspider.com/) and PubChem (https://pubchem.ncbi.nlm.nih.gov/) databases. The ligands and the proteins were prepared with Discovery Studio 2020. Energies minimization and the ensembles of ligands’ conformations were optimized. The structure of Mpro was prepared by removing the water molecules, adding hydrogen, fixing the missing residues or loops, and defining the binding site of original ligand as the active center. Then, the tested compounds were docked to Mpro by CDOCKER module [24], from which the CDOCKER interaction energy score was ranked to predict the interaction capacity.

Molecular dynamics (MD) simulation

MD simulation was performed with GROMACS 2018. The topologies of Mpro and ligands were prepared based on the CHARMM36 force field and the ChARMM-GUI server, respectively [25]. The complexes were solvated via the TIP3P water model in a dodecahedral periodic box with nearly 2.5 nm in all directions, which was balanced in the intracellular environment by adding 0.15 M NaCl. The 50,000 steps of steepest descent algorithm were used for minimizing energy [26]. The canonical ensemble, constant-pressure, and constant-temperature equilibration simulation in systems were carried out about 300 ns. The electrostatic interactions were described using the particle mesh Ewald (PME) algorithm with a cutoff of 1.2 nm. The LINear Constraint Solver (LINCS) algorithm [27] was used to constrain the bonds. The system pressure was maintained semi-isotopically at 1 bar in the x, y, and z-directions via the Berendsen barostat. The system temperature was maintained at 300 K via the Berendsen thermostat. The time of MD simulation was set at 2 fs. The analysis was carried out by the Gromacs software package, including GMX RMS, GMX RMSF, GMX gyration, GMX COVAR, and GMX ANAeig modules [28].

Analysis of NMR spectra

NMR spectra were employed to demonstrate the formation of hydrogen bond between the tested compound and Mpro, including saturation transfer difference (STD) and water-ligand observed via gradient spectroscopy (WaterLOGSY). The tested compound (acteoside) and Mpro were respectively dissolved in D2O, which were employed to prepare the sample solutions for STD analysis (Cacteoside: CMpro at about 100: 1) and WaterLOGSY analysis (Cacteoside: CMpro at about 10: 1) [29]. The spectra were recorded by 256 scans with 16 ppm spectral width at 600 MHz. The temperature was fixed at 298 K. The data were analyzed by mestrenova software (Version 9.0).

Results

High throughput screening of M pro inhibitors from XFBD by FRET assay

As a high throughput screening method, FRET is frequently employed to identify the potentially active compounds by measuring the interaction between the tested compounds and target proteins. By employing Mpro with purity above 95% purified in our laboratory (Additional file 1: Fig. S6), a total of 83 natural compounds were screened by FRET to identify the potential active compounds in XFBD. Linoleic acid, acteoside, licochalcone D, licochalcone B, and physcion exhibited satisfactory activities with the inhibition rate above 90% at 40 µM (Additional file 1: Fig. S7), which showed Mpro inhibition activity with IC50 ranging from 1.93 to 42.96 µM (Fig. 2). It was reported that up to 19% of small molecules preferentially form aggregates at 30 µM, which could nonspecifically inhibit the protease activity [30]. As a detergent, triton X-100 was added to the tested system to reach 0.01% (m: v) for preventing the molecule aggregation. Interestingly, licochalcone D obviously fluctuated in IC50. When without triton X-100, the inhibition activity of licochalcone D was drastically elevated from 29.46 to 5.54 µM (Fig. 2d), which possibly derived from the nonspecific inhibition on Mpro caused by self-aggregation of licochalcone D, resulting in the diverged inhibition activity on Mpro [31]. Therefore, acteoside and licochalcone B were suggested as the potential Mpro inhibitors.

Fig. 2
figure 2

Inhibition activity of the tested compounds on SARS-CoV-2 Mpro (n = 5)

The binding affinities between the tested compounds and M pro determined by SPR

SPR is employed to measure the ligand-Mpro interaction in real-time, which is widely accepted in determination of the binding affinity, including dissociation rate constant (Kd), association rate constant (Ka), and dissociation constant (KD, KD = Kd/Ka).

In our study, the five focused compounds by FRET were passed over the sensor’s surface to evaluate their affinities to Mpro. The binding responses in resonance units (RU) were continuously recorded. As shown in Fig. 3a–f, the tested compounds showed the increased SPR response in a dose-dependent manner. Acteoside, licochalcone B, and licochalcone D showed good binding affinities with KD values from 4.67 to 5.64 µM. Linoleic acid showed a moderate binding affinity with KD value at 10.8 µM. Physcion exhibited a weak response with KD value at 167 µM. In general, acteoside and licochalcone B were employed as potential active compounds by taking the screening results of FRET and SPR into consideration.

Fig. 3
figure 3

The SPR affinity curves of acteoside (a), licochalcone B (b), licochalcone D (c) and linoleic acid (d) fitted by kinetic mode, and physcion fitted by affinity mode (e and f). The calorimetric titration curve of acteoside-Mpro complex at 298 K (g)

With superior solubility, acteoside was further studied by ITC to obtain the thermodynamic signatures with the thermodynamic equilibrium constant (K), enthalpy variations (ΔH), entropy variations (ΔS), and binding stoichiometry (N) [32]. The calorimetric titration profiles of acteoside to Mpro was exhibited in Fig. 3g. The K of acteoside bound with Mpro was 0.129 µM at 298 K, ΔH was − 1.55 ± 0.094 kcal/mol, and ΔS was 27.3 cal/mol/deg. Negative values of ΔH and positive value of ΔS suggested that the binding of acteoside with Mpro was preferentially driven by ΔH, which displayed the electrostatic attraction and hydrophobic interaction as the key modes of interaction in the formation of acteoside-Mpro complex [33, 34]. The N value shows 0.48, indicating that acteoside binds to Mpro complex with a 1:0.48 stoichiometry, which possibly cause Mpro dimerization [19].

Antiviral activities on Omicron BA.1.1 and BA.2.3 variants in vitro

By MTT assay, acteoside showed no inhibitive effect on Vero E6 cells proliferation up to 100 µM for 24 or 48 h, which gave the half-cytotoxic concentration (CC50) > 100 µM (Fig. 4a and b). In antiviral activities assay, S-217622 was applied as a positive control to evaluate the reliability of the experimental system, which is the first oral noncovalent and nonpeptidic SARS-CoV-2 Mpro inhibitor as a clinical candidate [35]. The positive drug and the tested compound solutions with the different concentrations separately mixed with virus were added to the medium of Vero E6 cells for infection in 48 h, respectively. Then, the CPE score was used to evaluate antiviral activity by determining EC50 of acteoside. As shown in Fig. 4c and d, EC50 values of positive drug was 0.1205 μM on Omicron BA.1.1 and 0.2354 μM on Omicron BA.2.3. And EC50 values of acteoside were 41.29 µM for Omicron BA.1.1 and 70.37 µM for BA.2.3. As a focused active compound in XFBD, acteoside exhibited promising inhibitory effects against both Omicron BA.1.1 and BA.2.3 variants via inhibition of Mpro activity, which is consistent with the reported inhibition activity on SARS-CoV-2 Mpro by in-silico approach [36]. Given that the dominant spread of Omicron variants BA.1.1 and BA.2.3, the antiviral activity of acteoside may be meaningful in the clinical application for treating COVID-19 patients. Also, as a main compound from verbenae herba in XFBD, acteoside performs anti-inflammatory, anti-oxidant, cardiovascular protective, and immunoregulation [37], conducing to the recovery of COVID-19 patients.

Fig. 4
figure 4

Cytotoxicity studies of acteoside by MTT assay (a and b) and antiviral activities on Omicron BA.1.1 and BA.2.3 variants in Vero E6 cells (c and d)

Unveilment of the antivirus mechanism for acteoside by in-silico approach

A homodimer is regarded as the functional unit of Mpro composed of two monomers. Domain I, domain II, and domain III are the three major domains of Mpro. The Cys145-His41 catalytic dyad between domain I and domain II is set as the active site. Based on the model of the crystal complex structure of SARS-CoV-2 Mpro-N3, we docked acteoside into the active pocket to estimate the binding modes by the CDOCKER interaction conformation with minimum binding free energy. As shown in Fig. 5a, acteoside and N3 (representation in stick) are respectively aligned in the pocket of Mpro, highlighting their location and distance to the residues. One conformation of acteoside fits perfectly in the binding pocket of Mpro, which is similar to N3. The CDOCKER interaction energy of N3-Mpro is − 72.88 kcal/mol and acteoside-Mpro is − 68.78 kcal/mol. By docking acteoside to Mpro, it showed that hydrogen bonds were formed with Cys1453.9 and Glu1663.6 as side chain residues of Mpro, which is essential for exerting its inhibition activity.

Fig. 5
figure 5

The predicted structure of the acteoside-Mpro and N3-Mpro complexes (a) (Mpro is colored in blue, N3 is colored in grey and acteoside is colored in salmon). The results of RMSD (b), RMSF (c) and Rg analysis (d), and count of hydrogen bonds between acteoside and Mpro (e)

According to the results of root mean square deviation (RMSD) and root mean square fluctuation (RMSF) analysis (Fig. 5b and c), the stability of acteoside-Mpro complex was predicted to be satisfactory with the minor deviation in RMSD and the stable intermolecular interactions in the simulated trajectory [38]. As shown in Fig. 5d, close to Mpro, the radius of gyration (Rg) of acteoside-Mpro complex shows a better compactness, indicating the perfect interaction between acteoside and Mpro. Hydrogen bonds between acteoside and residues of Mpro were also observed at the early phase of the simulation and tended to be stable during the entire simulation period, which were considered as the primary interaction mode between acteoside and Mpro [39].

Uncovering the intermolecular interaction between acteoside and M pro by NMR spectroscopy

Based on the results of the molecular docking and molecular dynamic simulations, the STD and WaterLOGSY NMR spectra are employed to identify the ligand binding epitopes. In the STD NMR testing, the identified hydrogen receives the most intense magnetization transfer [40], which is suggested as the most intimate contact between the tested compound and protein [41, 42]. Presented in the Fig. 6 are the 1H NMR spectra of acteoside (a), acteoside-Mpro complex (b), Mpro (c), and STD (d). The STD NMR analysis showed that acteoside was bound to Mpro by H-atom at δH = 1.923 ppm, which was identified at 7-H of acteoside.

Fig.6
figure 6

1H NMR spectra of acteoside (a), acteoside-Mpro complex (b), Mpro (c), STD (d), and WaterLOGSY NMR spectra of acteoside-Mpro complex (e)

WaterLOGSY NMR spectrum also based on magnetization transfer, which is another method for evaluating the binding affinity between acteoside and Mpro to evidently support the STD NMR results. In WaterLOGSY NMR spectrum, the detected signals represents the transfer of magnetization among water molecules, proteins, and the tested compound via the nuclear overhauser effect (NOE) and chemical exchange [43]. As the rotational correlation time increases, the signal of the interacted hydrogens with the target protein from the tested compound turns to be negative. As shown in Fig. 6e, the signal is found to be negative at δH = 1.923 ppm, suggesting that the H-bond interaction occurs between acteoside and Mpro by 7-H, which is in line with result of STD. Accordingly, acteoside inhibit Mpro activity via the formation of hydrogen bond at 7-H of acteoside.

Discussion

Since the outbreak of COVID-19 in December 2019, it caused the number of infections and enormous pressure of intensive care for medical organizations. Lots of conventional antiviral drugs were used for screening the specific anti-SARS-CoV-2 candidates. To our disappointment, multiple drugs show the poor antiviral effects. Until now, molnupiravir and paxlovid (PF-07321332) are the two oral anti-SARS-CoV-2 drugs, which has been approved for marketing via FDA. However, these two drugs are not suitable for universal drugs for treating COVID-19 patients due to occurrence of high-risk groups by its side effects or drug resistance [41]. The vaccines, as the first line against the virus, exhibited the unsatisfactorily protective effect when patients were exposed to mutant strains, whose safety and efficacy remain to be elucidated. Whether fighting against SARS-CoV-2 or Omicron strains, the TCMs show an evidently therapeutic effect, which may be a gold mine for antiviral drugs discovery [44, 45]. XFBD is recommended for the treatment of COVID-19 in China and also approved in Canada, Uzbekistan, and so on. However, the antiviral mechanisms are less understood.

Mpro is a well-known target for drug discovery and designing, which plays the indispensable role in the replication of SARS-CoV-2 [46]. Focusing on Mpro, we screened and evaluated a series of compounds from XFBD to identify the antiviral bioactive compounds and clarify the potential mechanism by means of virtual screening and experimental verification. In the end, acteoside, as a representative active compound, was confirmed to be an effective inhibitor on wild-type Mpro (WT Mpro). The substrate binding pocket on SARS-CoV-2 Mpro is highly conserved among all coronaviruses. Compared to WT Mpro His132, mutation in Pro132 is observed in Mpro P132H of Omicron variants, which decreases thermal stability without compromising catalysis or small molecule drugs inhibition [47]. As shown in Fig. 7, P132H doesn’t change the conformation of the Cys145-His41 catalytic pocket, which shows the comparable catalytic activity compared to WT Mpro His132 [48]. In order to identify the binding form, the complex of nirmatrelvir-Mpro P132H (PDB ID: 7TLL) was employed to dock acteoside to Omicron Mpro (Additional file 1: Fig. S8). The CDOCKER interaction energy is − 65.44 kcal/mol for nirmatrelvir-Mpro P132H and − 62.40 kcal/mol for acteoside-Mpro P132H, showing the excellent activity of acteoside against Omicron Mpro. Moreover, the limitation of observable differences was observed in conformation between acteoside-WT Mpro and acteoside-Omicron Mpro. Therefore, whatever WT Mpro or Omicron Mpro, they could be employed as the screening target to identify antiviral candidates. With the continuous mutation of SARS-CoV-2, WT strain is rarely popular and has been instead of Omicron strains. Omicron variants are more invasive to immune system, resulting in higher transmissibility. Amazingly, acteoside inhibited omicron BA.1.1 and BA.2.3 in a dose-dependent manner.

Fig. 7
figure 7

The X-ray crystal structure of WT Mpro (SARS-CoV-2 Mpro, PDB ID: 6LU7, gray) and Omicron Mpro (SARS-CoV-2 Mpro P132H, PDB ID: 7TLL, green). The residues of Omicron Mpro (magenta) superposed on the WT Mpro (blue)

Acteoside is a phenylpropanoid glycoside, which was identified in XFBD via high-resolution MS (2021) [49]. In previous study, it was proved to enhance IFN-γ and then selectively activate T cells in mouse lymphocytes [50]. In vitro, acteoside has been demonstrated to have antivesicular stomatitis virus (VSV) activity. Moreover, acteoside can effectively inhibit influenza infection in vivo. As a natural product, acteoside is observed to be safe up to 800 mg/kg in mice [51]. Our finding provides a new avenue for antiviral applications of acteoside through inhibiting Mpro activity to block the replication of SARS-CoV-2.

Conclusions

As a representative compound, acteoside showed excellent inhibition on Mpro activity by the intermolecular interaction. Acteoside was successfully proved to display anti-Omicron BA.1.1 and anti-Omicron BA.2.3 activities in vitro. By computational modeling and NMR testing, we unveiled the intermolecular hydrogen bond formed by 7-H of acteoside and Mpro, resulting in inhibition of Mpro activity, which was speculated to be the antiviral mechanism of acteoside. In general, our study provides the strategy and procedures for rapid and effective identification of antiviral compounds and elucidation of its mechanism, conducing to scientifical application of formula in treating COVID-19 patients.