2.1 Chemicals

The subject of this study was whole A. argyi, which was collected from Tangyin, Henan Province, China in May 2018. The plant was identified as A. argyi by Professor Wang Yifei of the College of Life Science and Technology, Jinan University.

95% EtOH (AR), petroleum ether (PE) (AR), ethyl acetate (EA) (AR), MTT powder, and DMSO were purchased from Zhiyuan Chemical Reagent Factory. Fetal bovine serum, DMEM high-glucose medium, PBS, and penicillin–streptomycin solution were purchased from Thermo Fisher Scientific; MeCN (HPLC grade) and MeOH (HPLC grade) were purchased from Oceanpak, Sweden. All other chemicals used were molecular biology grade.

2.2 Instrument conditions

High-performance liquid chromatography (HPLC) SHIMADZU LC 20AT/LC-6AD was purchased from Shimadzu (Japan), the Ultra Performance Liquid Chromatography Coupled with Q-Exactive Orbitrap Mass Spectrometry(UPLC-Q-Exactive-Orbitrap MS) was purchased from Thermo Fisher Scientific (China), Bruker amaZon SL low-resolution mass spectrometer was purchased from Brooke Dalton (USA), the rotary evaporator was purchased from EYELA (Tokyo, Japan), thin-layer chromatography silica gel plate was purchased from Merck (Germany), TD-low-temperature microextraction and concentration unit were purchased from Tianzhong Machinery Manufacturing Co., LTD (Wenzhou, China), and the silica gel column chromatography fillers (200–300 mesh) were purchased from Marine Chemical Works (Qingdao, China).

2.3 Extraction and isolation of sesquiterpenoids from A. argyi

A sample of A. argyi (2 kg) was weighed accurately and extracted with 95% EtOH three times for 24 h each at room temperature. The EtOH solution was filtered and concentrated in vacuum until dry, and the extract was dissolved in 1L of water. The extraction was repeated three times with an equal volume of PE and EA. The EA part was eluted with silica gel filler as the stationary phase and dichloromethane and methanol as the mobile phase according to the ratio of 100:0, 99:1, 74:1, 64:1, 49:1, 33:1, 14:1, 9:1, 4:1, 1:1. All eluted parts were obtained with TLC and HPLC analysis, 9 components were obtained by combining them (Figure S1). Finally, according to the results of HPLC analysis, AASs were mainly enriched in Fr.6 and Fr.8 (Figure S2).

2.4 System conditions for UPLC-Q Exactive Orbitrap-MS

The AASs extract was prepared to a concentration of 1 mg/mL using methanol supplemented by ultrasonic dissolution, which was then filtered using 0.22 μm organic microporous filters.

Fr.6 and Fr.8 of A. argyi were separated on a Hypersil GOLD C18 (100 × 2.1 mm, 1.9 μm) column at a flow rate of 0.3 mL/min, injection volume of 2 μL, and column temperature of 40° C. The mobile phase consisted of 0.1% formic acid aqueous solution (phase A) and methanol solution (phase B). Gradient elution: 0.01–10.00 min, 10–30% B; 10.00 18.00 min, 30–95% % B; 18.00 to 24.00 min, 95% B; 24.00 ∼ 26.00 min, 95–10% B; 26.00 to 30.00 min, 10% B.

The high-resolution MS source parameters used in this study comprised + 4.0kv and −3.0kv spray voltage, 30arb sheath gas, 10arb auxiliary gas, 0 purge gas, 320 °C ion transfer tube temperature, and 350 °C auxiliary gas heating temperature, and were detected in positive and negative ion modes. The molecular formula and exact molecular weight of the primary mass spectrum were determined using Compound Discoverer 3.2 software and matched with both the mzCloud network database and a local traditional Chinese medicine component database. The data were further analyzed using mass spectrometry software and compound database for comparison.

2.5 Extraction and isolation

Subsequently, the compounds were isolated by silica gel column chromatography, ODS column chromatography and preparative high performance liquid chromatography (HPLC) (Chen et al., 2022). 17 compounds were obtained with a purity of greater than 95% and dissolved in DMSO at 4 °C (Xia et al., 2023). The specific purification process is shown in Figure S2.

2.6 Target collection and network pharmacology analysis

The chemical components in this study were derived from the extraction and isolation of A. argyi. The target gene set of A. argyi was obtained by searching several databases: (1) CTD (https://ctdbase.com/); (2) Swissstargetprediction server (https://www.swiss-targetprediction.ch); (3) SuperPred (https://prediction.charite.de/). Target to obtain from GeneCards database (https://www.genecards.org/) to select 400 protein targets. Finally, Cytoscape will be used to collect the targets for “compaction-target-disease” mapping.

2.7 Bioinformatics analysis of AASs

To enhance the understanding of drug efficacy, the core targets of COVID-19 associated AASs were incorporated into STRING for Protein-protein interaction (PPI) analysis. Interaction networks were constructed using PPI data with a confidence score above 0.9, and hub genes were identified to perform functional enrichment through KEGG pathway and GO analyses, which were visualized using Metascape (https://metascape.org) and DAVID (https://david.ncifcrf.gov) combined with the R package cluster Profiler (version 4.2.1) (Wang et al., 2022). Only pathways or terms with an adjusted p-value < 0.05 and q-value < 0.01 were deemed significant.

2.8 SARS-CoV-2 M^Pro activity assay

The activity of SARS-CoV-2 M^pro was measured using a fluorescence resonance energy transfer assay in a 96-well black flat plate. The reaction volume was 100 μL per well and consisted of 90 μL of assay buffer, 1 μL of 2019-nCoV M^pro/3CL^pro, 4 μL of substrate (Dabcyl-KTSAVLQSGFRKME-Edans) (Beyotime Company, China), and 5 μL of AASs at various concentrations (dissolved in DMSO). The concentrations of Ebselen were 0.08, 0.16, 0.31, 0.63, 1.25, 2.50, 5.00, and 10.00 μM, while those of AASs were 2.5, 5, 10, and 20 μM. The blank control well contained 91 μL of assay buffer, 5 μL of DMSO, and 2 μL of substrate. The enzyme activity control well contained 92 μL of detection buffer, 1 μL of M^pro, 5 μL of DMSO, and 4 μL of substrate, and the sample well contained 90 μL of detection buffer, 1 μL of Mpro, 5 μL of DMSO, and 4 μL of substrate. The plate was incubated in darkness at 37 °C for 5 min, and the fluorescence signal was measured using a multi-scan spectroscopy (Thermo Fisher, Shanghai, China) with excitation/emission wavelengths of 325/393 nm. The results were quantified using Formula 1.

2.9 Surface plasmon resonance (SPR) assay

The RBD (Sino Biological, 40592-V05H) was immobilized on a CM5 sensor chip using an amino coupling reaction, with PBS-P + used as the running buffer. The RBD was diluted to a final concentration of 25 µg/mL using a sodium acetate solution (10 mM, pH 4.5) and the immobilization level was around 9700 RU (Response Units). To perform the binding assays, AASs were diluted in 5% DMSO PBS-P + running buffer to concentrations of 200 µM, 100 µM, 50 µM, 25 µM, 12.5 µM, 6.25 µM, and 3.125 µM, and then passed over the sensor chip at a flow rate of 30 µL/min for 60 s. The dissociation time was 60 s. The data were analyzed using the Biacore evaluation software, and the K_D values were calculated by either kinetic analysis or steady-state affinity method(Yi et al., 2022).

2.10 Cell culture

HEK293T, HEK293T/ACE2, RAW264.7, HepG2, A549 and Caco-2 cell lines were cultured in Dulbecco's Modified Eagle Medium (DMEM, Gibco, USA) supplemented with 10% fetal bovine serum (FBS, Capricorn scientific, Germany) and 1% penicillin (100 units/ml)/streptomycin (100 μg/mL) (Gibco, USA). 293T cells stably expressing human-ACE2 (293T/ACE2) were constructed by our laboratory.

2.11 Cell viability

HEK293T/ACE2, RAW264.7, HepG2, A549, and Caco-2 cells were seeded in 96-well plates at 5000 cells per well. After 24 h, the complete medium was removed, and 100 μL of DMEM solution containing AASs at different gradient concentrations was added, and the culture was continued for 48 h. Cell viability was examined by an MTT colorimetric assay. Briefly, 10 μL 0.5 mg/mL MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] (Sigma-Aldrich) was added to each well and incubated for 4 h in an incubator. The medium was then replaced with DMSO (Macklin, Shanghai, China) to dissolve the formazan crystals (Sa-Ngiamsuntorn et al., 2021). Absorbance was measured at a wavelength of 570 nm on a microplate reader (Thermo Fisher).

2.12 SARS-CoV-2 pseudovirus (PsV) infection assay

The pCDNA3.1(+)-2019-HnCoV-Spike(SARS-Cov-2), pVSV-G, and pNL4-3-Luc-R-E- plasmids (MiaoLingBio, Wuhan, China) were co-transfected into HEK293T cells to package SARS-CoV-2 PsV (Xia et al., 2020, Li et al., 2021). The supernatant containing PsV was collected at 48 h after transfection, centrifuged at 3000 g for 10 min, filtered and stored at -80 °C. To evaluate the inhibitory activity of AASs against PsV infection, 293T/ACE2 cells (5000 cells / well) were seeded into 96-well plates and cultured at 37 °C in 5% CO₂. After 24 h, 50 μL of gradient concentrations of AASs compounds and 50 μL of PsV supernatant were added to each well and mixed and incubated with the cells for 72 h. Remove medium, PBS washing, and use with 100 mM potassium phosphate buffer (pH 7.8), 0.2% of Triton X - 100, and 1 mM DTT splitting cells (Wu et al., 2019). After full lysis, 20 μL of cell lysate was added to a white 96-well plate, and 80 μL of reaction substrate (1 mM ATP, 5 mM MgSO₄ and 0.5 mM D-Luciferin) was added, and the luminescence reaction was quantitatively measured in a 96-well microplate photometer. The inhibition rate of PsV was calculated as shown in Formula 2.

2.13 Determination of nitric oxide content in RAW 264.7 cells induced by LPS

In order to study the anti-inflammatory activity of A. argyi, the anti-inflammatory activity of 17 compounds was evaluated in lipopolysaccharide (LPS) -induced RAW264.7 cell inflammation model in vitro (Masih et al., 2021). Cells were cultured in 24-well plates at 2 × 10⁵ cells per well for 24 h. The medium was removed and 400 μL of various concentrations of AASs were added. After 2 h of AASs treatment, 4 μL of 0.1 mg/mL LPS solution (dissolved in PBS) was added to the model group and the experimental group, and 4 μL of PBS was added to the blank group for 24 h. 50 μL of the supernatant from each well of a 24-well plate was transferred to a 96-well plate and the Griess kit (Beyotime Company, China) was added to determine the NO content (Xue et al., 2021). Absorbance values were measured at a wavelength of 562 nm using a microplate reader, and nitric oxide concentration was calculated from a standard curve, as shown in Formulae 3.

2.14 Molecular docking

Through SPR, enzyme inhibition experiments and anti-inflammatory activity screening, it was found that AASs could closely bind to the RBD of SARS-CoV-2, effectively inhibit the enzyme activity of M^pro, and effectively inhibit the production of NO by inflammatory cells. However, the structure–activity relationships between these compounds and their corresponding proteins are not yet.

ChemOffice 2019 was utilized to draw the structures of AASs, while Autodock 4.2.6 was employed to optimize these compounds. The protein structures of SARS-CoV-2 RBD (PDB ID: 6m0j), COVID-19 M^pro (PDB ID: 6LU7), and iNOS (PDB ID: 4CX7) were processed using Pymol and imported into Autodock for docking. The binding sites between small molecules and proteins were determined based on previous studies, and for RBD, the binding site was located between RBD and ACE2. A grid box with dimensions of 60 Å × 60 Å × 60 Å and grid spacing of 0.375 Å was centered at (x = -34.527, y = 20.545, z = 10.113). For COVID-19 M^Pro, the grid box was set according to the location of its receptor, with dimensions of 40 Å × 40 Å × 40 Å, coordinates of (x, y, z) of −9.253, 16.459, 65.388, and grid spacing of 0.375 Å (Ye et al., 2021). The iNOS binding site was defined by a grid box centered on the co-crystallization inhibitor (x = -11.594, y = -60.255; z = 15.846) with dimensions of 40 Å × 40 Å × 40 Å and mesh spacing of 0.375 Å.

The processed ligands and receptors were imported into Autodock 4.2.6 and molecular docking was performed using default parameters. Finally, the docking results were analyzed and processed using PyMol 2.5, ligplot⁺ 2.2.5 and Chimera 1.16.

2.15 Data statistics

Experimental data are presented as mean ± SD, and all grouped data were statistically analyzed using Graphpad Prism 8.4.3 software. When P < 0.05 was considered statistically significant. Significant differences between data from each group were assessed by ANOVA.

3.1 Analysis results of UPLC-Q-Exactive-Orbitrap-MS

Fr.6 and Fr.8 in the positive ion mode of the sagitta sesquiterpene extract are depicted in Fig. 1, which display the total ion current signals. By comparing the primary mass spectrum information, fragmentation, and molecular formula of compounds reported in relevant literature, and combining the matching compounds from mzCloud and mzVault databases, a total of 28 and 42 terpenoids were preliminarily identified in Fr.6 and Fr.8, respectively (Table 1 and Table 2). The structure diagram is illustrated in Fig. 2. The identified sesquiterpenes included 16 guaiacane, 30 eucalane, 13 gemmarane, and 8 other sesquiterpenes, as well as 2 sesquiterpene dimers and one monoterpenoid.

Close

Fig. 1

The total ion chromatograms in positive ion modes of Fr.6 (A) and Fr.8 (B) in A. argyi.

Export to PPT

Close

Fig. 2

Structure of sesquiterpenoids in A. argyi.

Export to PPT

3.2 Structural identification of A. argyi compounds

17 compounds were isolated from A. argyi by various chromatographic methods, and their structures were determined as Tanaphilin (1), Loliolide (2), 3α-hydroxyryanosin (3), 3-method-tanaparthole (4), 14-deoxyylactin (5), Artecalin(6), Chlorofluotzchin (7), Chrysanthemulide I (8), Artemisianins A (9), Achillinin C (10), Isotanciloide (11), Dehydromatricarin A (12), epi-Baynol A (13), 3, 3-Dimethyl-γ-methylene-2-(3-oxobutyl) cyclobutanoic acid (14), Argyin C(15), Ilicic alcohol(16), Argyin A(17), through extensive spectra analysis (Fig. 3).

Close

Fig. 3

Compounds isolated from Fr.6 and Fr.8 in A. argyi.

Export to PPT

3.3 Target collection of active ingredients of AASs

17 sesquiterpenoids isolated from A. argyi were identified using CTD, Swissstargetprediction server and SuperPred, and 806 protein targets were predicted. In addition, 400 targets of “COVID-19″ and ”SARS-CoV-2″ were obtained from GeneCard database, and 275 targets of SARS-CoV-2 were selected. 91 of these targets overlapped with AASs. Cytoscape was used to generate a drug-target-disease map (Table S1). The “drug-target-disease” diagram was generated using Cytoscape (Fig. 4). Compounds 10, 9, 16, 6, 4 and 12 showed more than 35 cross-protein targets with SARS-CoV-2, suggesting their potential therapeutic effects. Notably, AASs targeted key proteins such as NFKB1, ACACA, CNR2, KDM1A, TLR4 and PIK3R1 in treatment.

Close

Fig. 4

Compound-target network of A. argyi sesquiterpenoids formula constructed using Cytoscape.

Export to PPT

The PPI of intersection targets was obtained by String database and visualized by Cytoscape (Fig. 5). The average degree of each protein was 12.84, with 29 protein targets having nodes greater than the average, while STAT3, PIK3R1, AKT1, MAPK1, PTPN11, EGFR, and JUN had nodes greater than 30, indicating that these 7 targets are the core targets of AASs for the treatment of COVID-19.

Close

Fig. 5

PPI of AASs in treating COVID-19.

Export to PPT

3.4 GO biological function annotation and KEGG pathway enrichment analysis

Functional enrichment analysis of GO and KEGG was performed to investigate the potential mechanism of AASs in the treatment of SARS-CoV-2. For example, molecular functions mainly include ATP binding and viral entry into host cell, AASs involve cellular components including cytoplasm, and biological processes include the response to lipopolysaccharide, inflammatory response, and so on (Fig. 6(A)). The top 20 results obtained by KEGG pathway analysis are shown in Fig. 6(B), indicating the proportion of target genes belonging to a certain pathway to all annotated genes in the pathway, with dot size representing the number of target genes. The analysis revealed that a majority of the diseases associated with the identified targets were viral-related, with pathways in cancer being the most prevalent. This suggests that AASs may have potential involvement in multiple cancer-related pathways for treatment purposes. Furthermore, the NOD-like receptor and HIF-1 signaling pathways were identified as the primary signaling pathways targeted by AASs in relation to COVID-19.

Close

Fig. 6

(A) The top 5 pathways for GO enrichment analysis of the targets of AASs. (B) The top 20 pathways for KEGG enrichment analysis of the targets of AASs.

Export to PPT

3.5 Analysis of anti-pseudovirus infection

To screen for potential inhibitors of SARS-CoV-2 entry in A. argyi, pseudovirus coated with SARS-CoV-2 Spike protein was used to infect HEK293T/ACE2, and chloroquine was selected as the positive drug. As shown in Fig. 7, the three compounds 5, 10 and 12 in A. argyi had a good inhibition effect on pseudovirus invasion within 20 μM, and the inhibition rate reached 54.2385%, 53.7423% and 63.964% at 20 μM, and the cell survival rate was higher than 80% at this concentration. The anti-pseudovirus results of other compounds are shown in Figure S4.

Close

Fig. 7

Inhibitory activity of the positive drugs chloroquine, Compound 5, 10 and 12 against pseudovirus infection in HEK293T/ACE2 cells.

Export to PPT

In order to explore the most important domain RBD binding to ACE2 in Spike protein. By SPR analysis of 17 compounds, only C10 and C12 showed strong binding to the RBD of SARS-CoV-2 spike protein, while other sesquiterpenoids such as C3, C8, C13 and C16 showed only weak binding (Figure S5, S6). The K_D values for C10 and C12 binding to SARS-CoV-2-RBD were 181.9 μM and 164.4 μM, respectively. Molecular docking analysis indicated that Compounds 10 and 12 formed hydrogen bonds with Arg403, which led to tight binding to RBD. The α, β-unsaturated γ-lactone structure of C10 and C12 interacted with Glu406 and Arg403 via their ester groups (Fig. 8 (C, D)). Therefore, the unsaturated lactone structure of sesquiterpenes plays a crucial role in their binding conformation to the RBD protein, with Arg403 serving as the key residue for binding to the spike RBD protein.

Close

Fig. 8

The molecular interaction of RBD and C10(A) and C12(B). The molecular interaction of RBD and C10(C) and C12(D).

Export to PPT

3.6 Inhibitory activity of AASs on SARS-CoV-2 M^pro

Main protease (M^pro) is a crucial protease for mediating viral replication and transcription in SARS-CoV-2 (Marinho et al., 2020), making it a potential target for developing drugs against the virus (Fu et al., 2022). To investigate the efficacy of AASs in treating SARS-CoV-2-induced pneumonia, the inhibitory activity of 17 sesquiterpenoids from A. argyi against M^pro was analyzed using Ebselen as a positive control drug (Figure S7). The inhibition rate of M^pro on the test compounds showed a clear dose–effect relationship, with the inhibition rate increasing with increasing concentration (Fig. 9(C)). IC₅₀ values for Compounds 1, 8, 10, 13, 14, 15, 16 and 17 in A. argyi were found to be 24.64 μM, 21.46 μM, 23.17 μM, 35.58 μM, 33.45 μM, 26.34 μM, 35.27 μM, and 8.737 μM.

Close

Fig. 9

(A) Protease activities of SARS-CoV-2 M^pro in the presence of inhibitors were measured by FRET-based enzymatic activity assay. IC₅₀ of Ebselen and Compounds 1, 8, 10, 13, 14,15, 16 and 17.

Export to PPT

Based on the inhibition of M^pro activity by sesquiterpenes, the relationship between chemical structure and activity was further analyzed by molecular docking of several sesquiterpenes with strong inhibitory activity (Fig. 10). From the docking results, it is obvious that compounds 15 and 17 are guaiacane-type sesquiterpene lactones, and the O-diol hydroxyl groups at C-3 and C-4 positions of their five-membered rings form a stable hydrogen bond with residue Ser46; Compound 8 also contains the same structure, and its five-membered ring of guaiacane sesquiterpene lactone also has O-diol hydroxyl group, and forms a stable combination with Asn119, so it is considered that this structure is beneficial to interact with M^pro to achieve lower concentration to inhibit its activity. From other docking results, it is also found that residues Thr24 and Thr26 are important groups for sesquiterpenes to bind with M^Pro, and unsaturated ketones in sesquiterpenes can form hydrogen bonds with these residues to achieve stability (Fig. 10 (A, C)).

Close

Fig. 10

(A-H) The molecular interaction of SARS-CoV-2 M^pro and Compounds 1, 8, 10, 13, 14, 15, 16 and 17, respectively.

Export to PPT

3.7 Inhibition of NO induced by LPS in RAW264.7 cells

Before investigating the inhibitory effect of AASs on NO release in inflammatory RAW264.7 cells in vitro, the inhibitory effect of these compounds on cell viability was first studied to screen sesquiterpenes that could exert good anti-inflammatory activity without affecting normal cell growth. Therefore, the cytotoxicity of 17 sesquiterpenoids at 50 μM was studied (Fig. 11 (A)). Compounds 4, 10, 12, 15, 16, 17 showed significant inhibitory effects at the concentration of 50 μM, but Compounds 4, 12, 15, 17 still inhibited normal cell growth at the concentration of 25 μM (Fig. 11 (B)).

Close

Fig. 11

MTT assay of 17 compounds (50 μM). (B) MTT assay of Compounds 4, 10, 13, 15, 16 and 17 (25 μM). (C) Inhibitory effects of Compounds 2, 5, 9, 10, 13, 16 on NO production in RAW264.7. Results presented as the mean ± SD of three replicated tests. The GraphPad Prism was used to analyze the results using one-way ANOVA. Note: *P < 0. 05, ^**P < 0.01, ^***P < 0.001, ^****P < 0.0001.

Export to PPT

The results showed that Compounds 2, 4, 5, 9, 10, 12, 13, 15, 16, and 17 could significantly inhibit the release of NO in RAW264.7 inflammatory model in a concentration-dependent manner, while Compounds 1, 3, 6, 7, 8, 11, and 14 showed a weak inhibitory effect (Figure S6). For example, there are many hydroxyl groups in the structure of C3, C8 and C11, which weaken the lipid solubility of compounds and lead to their weaker inhibitory ability on NO release in RAW264.7 cells than other sesquiterpenes.

Sesquiterpene dimers showed stronger ability to inhibit cellular NO secretion levels compared to monosesquiterpenoids. C10, formed by the polymerization of C1 with another sesquiterpene, showed better inhibition than C1 due to the presence of two active groups despite the destruction of one lactone outer double bond structure. Compound 2, 5, 9, 10, 13, and 16 did not inhibit the growth of RAW264.7 cells at a concentration of 25 μM (Figure S8), while their IC₅₀ values for inhibiting the level of NO in RAW264.7 cells were 11.03, 4.415, 6.718, 3.656, 2.605, and 4.189 μM (Fig. 11 (C)), respectively, indicating good activity.

Molecular docking results indicated that C9 and C10 had similar structures, both possessing α, β-unsaturated γ-lactone and α, β-unsaturated ketone functionalities, and displayed comparable binding affinity towards iNOS protein (Fig. 12). However, C10, which had an additional hydroxyl group at the C-8′ position, formed a hydrogen bond with amino acid residue Glu263, resulting in a stronger binding force and lower binding energy. Thus, the unsaturated lactone ring and unsaturated ketone functionalities were crucial for the binding conformation of AASs and iNOS protein, while the hydroxyl group in AASs contributed to the binding affinity by reducing the binding energy.

Close

Fig. 12

Interactions of compounds 2, 5, 9, 10, 13, and 16 with iNOS protein, respectively.

Export to PPT

Close

Fig. 13

Cytotoxicity of AASs on three cell lines representing major human organs was assessed by MTT assay after 48 h treatment with (A) liver (HepG2), (B) lung (Calu-3), and (C) intestine (Caco-2) cell lines. All experiments were performed in three biological replicates. The GraphPad Prism was used to analyze the results using one-way ANOVA. Note: *P < 0. 05, ^**P < 0.01, ^***P < 0.001, ^****P < 0.0001.

Export to PPT

3.8 Cytotoxic profiles of the 17 compounds

One of the major problems in the development of medicinal plant-derived drugs is the possible damage to different organs of the human body caused by herbs. To address this issue, in this study, we used human cell lines from three major organs, including liver (HepG2), lung (A549), and intestine (Caco-2), to evaluate the cytotoxicity of these 17 compounds by MTT assay. The results showed that the same compound showed different cytotoxicity in different cell lines. Among them, Compound 10, which had the best combination of anti-Pseudovirus and anti-inflammatory effect, showed obvious toxicity at the concentration of 50 μM, but did not show great toxicity to the three cell lines when the concentration was reduced to 25 μM.