Vero and HUVEC cells were obtained from American Type Culture Collection (ATCC), Huh7 and BSR-T7 cells were obtained from China Center for Type Culture Collection. All cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM; Gibco) supplemented with 10% fetal bovine serum (Gibco) and 1% antibiotics (Gibco) in a humidified incubator of 5% CO₂ at 37°C. The expression of T7 polymerase in BSR-T7 cells were selected by the addition of 1 mg/ml G418 (Thermo Fisher Scientific).

The SFTSV isolate HBMC16 (GenBank accession number: KY440775.1, KY440776.1 and KY440777.1) were obtained from China Centre for General Virus Culture Collection and was propagated in Vero cells. Viral titer was determined by focus-forming assay on Vero cell as described previously (Li et al., 2020). The SARS-CoV-2 (IVCAS 6.7512) was propagated and titered in Vero E6 cells (Zhou et al., 2020). Rift Valley fever virus (RVFV) was generated by using the infectious clones of RVFV BJ01 based on a T7 polymerase dependent system as described previously (Li et al., 2019b). All infection experiments with SARS-CoV-2 and RVFV were performed in a biosafety level 3 (BSL-3) facilities.

A small molecular compound library of 1,058 compounds derived from natural products, including Melia Toosendanin Sieb. et Zucc and pomegranate rind, was purchased from Weikeqi Biotech (Si Chuan, China). The compounds were purified with HPLC, the purity of compounds was ≥98%. Vero cells pre-seeded in 96-well plates at a density of 10,000 cells per well were treated with each compound at a final concentration of 10 μM for 1 h and infected with SFTSV at an MOI of 0.125. Vehicle (DMSO) was used as a negative control and benidipine (10 μM) was used as a positive control based on our previous study (Li et al., 2019a). The percentage of DMSO in vehicle group was 1‰, same as the percentage of DMSO in compound group. At 36 h post infection (p.i.), cells were fixed with 4% paraformaldehyde (PFA), followed by permeabilization with 0.2% (v/v) Triton X-100 and blocking with 3% (w/v) bovine serum albumin (BSA). Then cells were incubated with rabbit anti-NP antibody, followed by anti-rabbit Alexa488 (ThermoFisher Scientific) and DAPI (Beyotime). The images were captured by an Operetta high-content imaging system (PerkinElmer) and analyzed by Harmony 3.5 software (Li et al., 2019a). The infection rate of DMSO control was set as 100%, and the infection rate of each drug was calculated by normalizing to DMSO control. The percentage inhibition was equal to 100% minus infection rate in the drug treated samples. The survival rate of DMSO control was set as 100%, and the survival rate of each drug was calculated by normalizing to DMSO control.

Cells pre-seeded in 96-well plates were treated with different concentrations of drugs for 36 h. Cell viability was measured by Cell Counting kit-8 (CCK-8; Dojindo Molecular Technologies) according to the manufacturer’s instructions.

Total RNAs in infected cells were extracted with FastPure Cell/Tissue Total RNA Isolation Kit (Vazyme, China) following the manufacturer’s instructions. Quantitative reverse transcription PCR (qRT-PCR) was carried out with a two-step procedure. First, the cDNA of viral genome was generated by using the HiScript III 1st Strand cDNA Synthesis Kit (Vazyme, China), and then quantified by using ChamQ Universal SYBR qPCR Master Mix (Vazyme, China). Primer sequences are provided in Supplementary Table S1.

Infected cells lysates were subjected to 12% SDS-polyacrylamide gel electrophoresis (PAGE), and then transferred to polyvinylidene difluoride (PVDF) membranes (Millipore). The proteins were immunoblotted with antibodies against viral nucleoproteins (Li et al., 2019a), Tubulin (Polyclonal rabbit, Proteintech, 11224-1-AP) or GAPDH (Monoclonal mouse, Proteintech, 60004-1-Ig) then HRP-conjugated goat anti-Rabbit IgG (Proteintech, SA00001-2) and anti-Mouse IgG (Proteintech, SA00001-1). Prestained protein ladder was purchased from ThermoFisher (26616). Bands were detected by a Chemiluminescence Analyzer (Chemiscope600pro) using an enhanced chemiluminescence (ECL) kit (Millipore).

Vero, Huh-7 and HUVEC cells pre-treated with different concentrations of notoginsenoside Ft1, punicalin, and toosendanin were infected with SFTSV for 24 h. The viral RNA in the supernatant and infected cells was quantified with qRT-PCR. Infectious virus in supernatant was determined by focus-forming assay. The immunofluorescence assay was performed and analyzed as described above.

For the determination of inactivation effect of punicalin and toosendanin. SFTSV was incubated with punicalin and toosendanin at 37°C for 1 h, and then diluted 10,000-fold for titer determination. Time-of-addition assay was performed as described previously with adjustment (Wang et al., 2017a). HUVEC cells were infected with SFTSV or RVFV at an MOI of 2 (0 h), the initial inoculum was removed at 2 h post infection. For the pre group, cells were pre-treated with toosendanin (1 μM) for 1 h, and incubated with virus for 2 h, and then drug and virus were removed. For the post group, cells were infected with SFTSV for 2 h, and toosendanin was added after the removal of virus. For the whole group, cells were treated with toosendanin from 1 h pre-infection to 12 h post infection. Infected cells were collected at 12 h post infection and subjected to qRT-PCR and western blot. Vero-E6 cells were infected with SARS-CoV-2 at an MOI of 0.01 for 1 h, and treated with toosendanin (1 μM) as described above. Infected cells were collected at 24 h post infection for qRT-PCR and western blot.

HUVEC cells were pretreated with different concentrations of toosendanin or vehicle (DMSO) at 37°C for 1 h. Then the cells were chilled at 4°C for 15 min, and then incubated with SFTSV (MOI = 5) at 4°C for 2 h. The unbound virions were removed by wash with pre-cooled PBS, cells were harvested and lysed and the relative level of bound virions was quantified by qRT-PCR.

HUVEC cells pre-treated with different concentrations of toosendanin or vehicle (DMSO) for 1 h were chilled at 4°C for 15 min and incubated with SFTSV (MOI = 5) for 1 h. Cells were washed with pre-cooled PBS to eliminate the unbound virions and incubated at 37°C for 3 h in the presence of ammonium chloride (20 mM), then washed with PBS and treated with trypsin to remove the un-internalized virions. The relative level of internalized virions was determined by qRT-PCR.

For the visualizing assay, HUVEC cells pre-treated with toosendanin (2 μM) or vehicle (DMSO) were infected with SFTSV (MOI = 100) for 2 h in the presence of ammonium chloride (20 mM). Cells were washed with PBS and treated with trypsin, then fixed and subjected to immunofluorescence assay. Images were capture by a confocal microscope (Andor Dragonfly 202), and internalized viral particles were analyzed and quantitated by Imaris software (Imaris V9.2.1, Bitplane Inc.).

HUVEC cells were infected with SFTSV (MOI = 5) for 24 h and treated with indicated concentrations of toosendanin or vehicle (DMSO) for 1 h, followed by incubation of citric acid-sodium citrate buffer (0.1 M citric acid, 0.1 M sodium citrate, pH 5.0) for 20 min. Cells were fixed by methanol and stained by Giemsa solution (Sigma). Syncytium formation was imaged by an epifluorescence microscope (Olympus IX73).

T7 mini-genome-eGFP plasmid and viral proteins (L and NP) expression plasmids were co-transfected to BSR-T7 cells using lipofectamine 3000 (Thermo Fisher Scientific). Different concentrations of toosendanin or vehicle (DMSO) were added 6 h post transfection, and benidipine (20 μM) was used as a positive control. The eGFP signal was captured by an epifluorescence microscope (Olympus IX73) at 24 h post transfection and analyzed with Fiji.

Animal experiment was approved by the Animal Care Committee of the Wuhan Institute of Virology (permit number: WIVA38202007) and was performed in accordance with the ethical guidelines.

All statistical analyses were performed in GraphPad Prism version 8.4.3, as defined in the text and figure legends.

In order to identify potent inhibitors against SFTSV infection, we performed a high-throughput screening (HTS) of a small molecular compound library consisting of 1,058 natural extracts (Figure 1A). Briefly, Vero cells pre-treated with the compounds (10 μM) using vehicle (DMSO) as a negative control and benidipine as a positive control were infected with SFTSV for 36 h. The infection rate was determined with the immunofluorescence analysis (IFA) assay using antibody against the viral nucleoprotein (NP) (Li et al., 2019a). The correlation between two replicates was evaluated by Pearson correlation coefficient analysis (Figures 1B,C). Nine compounds were identified as candidates with inhibition rate >80% and the treated cell survival rate >60% (Figure 1D). Another screen was carried out to identify compounds with anti-viral effect in a dose-dependent manner. Three compounds, notoginsenoside Ft1, punicalin, and toosendanin, were identified as hits showing an anti-viral effect in a dose-dependent manner and with the selective index >10 (SI, 50% cytotoxic concentration [CC₅₀]/50% inhibitory concentration [IC₅₀]) (Figure 1E).

Virus titration assay revealed that all three compounds inhibited the production of infectious virions in a dose dependent manner (Figure 2A). Similar inhibitory effects were observed on the intracellular and extracellular vRNA levels and the production of the viral nucleoprotein under drug treatment (Supplementary Figure S1). The inhibitory activity of these selected compounds on SFTSV infection was further validated in additional cell lines, including Huh7 and HUVEC cells, which are susceptible to SFTSV infection (Sun et al., 2014). In consistent with the result on Vero cells, anti-SFTSV effects were observed in a dose dependent manner (Figures 2B,C). HUVEC cells are vascular endothelial cells, which might represent the vascular endothelial injury induced by SFTSV infection (Li et al., 2018), thus the subsequent experiments were performed mainly on HUVEC cells. The SI of punicalin and toosendanin on HUVEC cells were determined to be >100 and 4,800, respectively (Figure 2D). These two compounds were subjected to further investigation.

Next, we determined the inhibition mechanisms of punicalin and toosendanin. To investigate whether punicalin and toosendanin have virucidal effect on SFTSV virions, SFTSV was incubated with the compounds at the indicated concentrations for 1 h, and then the infectivity of the treated viruses were analyzed (Figure 3A). As shown in Figure 3B, punicalin directly inactivated virions in a dose dependent manner and the treatment of toosendanin demonstrated no virucidal effect on SFTSV.

A time-of-addition assay was performed to identify the SFTSV replication stage that was inhibited by toosendanin (Figure 3C). Toosendanin was added pre or post SFTSV infection and the infected cells were harvested at 12 hpi. Intracellular level of vRNA and expression level of viral protein were determined by quantitative RT-PCR and western blot, respectively. As shown in Figures 3D,E, vRNA and viral protein expression decreased significantly when toosendanin was added prior to but not post SFTSV infection, indicating that the drug inhibited SFTSV during the entry step. Similar results were observed when using a serial concentrations of drug (0.25–2 μM), confirming that toosendanin inhibits the entry rather than the viral replication stage (Figures 3F–K). Furthermore, the mini-genome (MG) assay revealed that toosendanin treatment did not affect the viral RNA replication and transcription (Supplementary Figure S2A).

The virus entry can be further divided into binding, internalizing and fusion events. We next investigated the specific viral entry event that is targeted by toosendanin. First, HUVEC cells pre-treated with toosendanin or vehicle (DMSO) were incubated with SFTSV at 4°C for 1 h to synchronize the binding process of virus. The un-bound virions were removed by three times of washing, and the relative level of virions bound to cells was analyzed. As shown in Figure 4A, vRNA level in toosendanin treated group was comparable to the vehicle control, suggesting that toosendanin did not inhibit SFTSV binding. To evaluate the effect of toosendanin on viral internalization, HUVEC cells were infected with SFTSV in the presence of toosendanin or vehicle, and the ammonium chloride was added to prevent low-pH-dependent membrane fusion to block viral entry at the step of fusion. The relative intracellular vRNA was significantly reduced in the toosendanin treated cells, compared with the vehicle control (Figure 4B). Furthermore, toosendanin treatment showed a dose dependent inhibition effect on accumulation of vRNA (Figure 4C). Virus internalization was further visualized with microscopy analysis. Internalized viral particles were stained with polyclonal antibody against the viral protein NP and images of the cross-section of infected cells were acquired. The intracellular viral particles were analyzed and quantified. As shown in Figures 4D,E, toosendanin treatment significantly reduced the amount of internalized viral particles. To evaluate whether toosendanin treatment inhibits viral fusion, syncytium formation assay was performed. Briefly, infected cells were treated with low pH medium, followed by addition of different concentrations of toosendanin and the syncytium formation was monitored. No significant difference of syncytium formation was observed between toosendanin treated and vehicle treated control (Supplementary Figure S2B), indicating that toosendanin did not inhibit the viral fusion step. These results suggested that toosendanin inhibited the internalization event of SFTSV infection.

The anti-SFTSV effect of toosendanin was further evaluated on C57BL/6 mice, an established mouse model of SFTSV infection (Jin et al., 2012). Eight-week old female C57BL/6 mice were pretreated with toosendanin through intraperitoneal administration for 3 days and infected intraperitoneally with SFTSV. Toosendanin or vehicle was administered on a daily basis. Toosendanin treatment and/or SFTSV infection did not lead to weight loss in this infected mouse model (Supplementary Figure S3). Spleen samples were collected from mice treated with toosendanin (n = 5) or vehicle (n = 5) and the viral antigen was determined by immunohistochemistry staining, less viral protein positive cells was detected in the spleen sections of mice treated with toosendanin (Figure 5A). The infection of SFTSV lead to an increase of megakaryocytes in the spleen of infected mice, which is correlated to the compensatory phenomenon of circulating platelet (PLT) depletion induced by SFTSV infection (Jin et al., 2012). The hematoxylin and eosin (H&E) staining of spleen sections revealed that toosendanin treatment reduced the megakaryocytes accumulation in the spleens of infected mice compared with the vehicle control (Figures 5B,C). The results indicated that toosendanin treatment can inhibit SFTSV replication and reduce histopathological changes in the infected C57BL/6 mouse model. Similar results were also observed in an anti-IFNAR1 IgG treated C57BL/6 mouse model, in which the administration of toosendanin reduced the antigen of SFTSV nucleoprotein in the spleen sections of infected mice (Figure 5D). Furthermore, viral load in the serum of toosendanin treated mice was significantly lower than that of the mice in the vehicle control group (Figure 5E).

In order to determine whether toosendanin has a broad anti-viral effect, we analyzed the potential inhibitory effect of this compound against another bunyavirus, Rift Valley fever virus (RVFV) and the recently emerging SARS-CoV-2. Toosendanin showed potent inhibition on the infection of RVFV (IC₅₀ = 0.13 μM, CC₅₀ > 450 μM, SI > 3,461 μM) and SARS-CoV-2 (IC₅₀ = 0.24 μM, CC₅₀ > 450 μM, SI > 1875 μM) (Figures 6A,D). Time-of-addition assay indicated that toosendanin inhibited the entry step of RVFV (Figures 6B,C), which is similar with the inhibition mechanism against SFTSV infection. Differently, in the case of SARS-CoV-2, toosendanin revealed inhibitory effect when added 1 h post infection (Figures 6E,F), suggesting that it targets the viral replication event.