Identification of Licoflavone C as a cap-dependent endonuclease inhibitor against severe fever with thrombocytopenia syndrome virus

Abstract

Severe fever with thrombocytopenia syndrome virus (SFTSV) is a tick-borne virus with a high fatality rate. Currently no approved drugs or vaccines are available against it. Sharing a common replication mechanism with negative-stranded, segmented viruses (NSVs), SFTSV utilizes a cap-dependent endonuclease (CEN) domain of the L segment to execute the cap-snatching process upon genome transcription initiation. Given the crucial role of CEN in the life cycle of NSVs, it is considered a promising target for discovery of antiviral agents against SFTSV. In this study, we established a high-throughput FRET-based enzymatic screening system to discover inhibitors of SFTSV CEN from a chemical library containing 3467 natural compounds. Finally, three compounds, i.e., Licoflavone C, 3,4-dicaffeoylquinic acid, and oleanolic acid displayed exceptional antiviral effects and minimal cytotoxicity. Licoflavone C (EC₅₀ = 1.85 μM) was selected for further investigation. Administration of Licoflavone C (20 mg/kg, i.v.) significantly reduced tissue viral loads in SFTSV-challenged mouse model. We demonstrated that Licoflavone C did not directly bind to the active pocket of SFTSV CEN but disrupted its active conformation, resulting in substrate non-competitive inhibition. Licoflavone C also exhibited broad-spectrum inhibition on several NSV CENs (HRTV, GTV, and LCMV) besides SFTSV. Furthermore, 15 analogs of Licoflavone C sharing a typical flavonoid structure were verified for targeting SFTSV CEN and exhibiting antiviral activities. In conclusion, Licoflavone C is a promising inhibitor of SFTSV, offering insights into targeting CEN with flavonoids in drug discovery.

Introduction

Severe fever with thrombocytopenia syndrome virus (SFTSV) is a pathogen causing severe fever with thrombocytopenia syndrome (SFTS), which was first reported in Henan, China [1], subsequently renamed Dabie Bandavirus by the International Committee on Taxonomy of Viruses, and categorized into the genus Bandavirus, family Phenuiviridae, and order Bunyavirales [2]. It was later reported in Korea, Japan, Vietnam, and other places [3–6] and is currently endemic to East Asian countries [7]. Its infection in humans primarily causes high fever, thrombocytopenia, and leukopenia, accompanied by gastrointestinal symptoms and metabolic dysfunction [8], intravascular hemorrhage, multiple systemic organ failure, and death (in critical cases) [9, 10]. The fatality rates of SFTS vary among regions, ranging from 5.3% to 16.2% in China [11], 27% in Japan, and 23.3% in the Republic of Korea [12]. The mortality rate is higher in older patients with underlying immune system deficiency diseases [13]. SFTSV is mainly transmitted to humans and animals through direct tick bites, contact with virus-contaminated blood, body fluids, or secretions, and aerosolized respiratory transmission [14]. Close-contact transmission among cats and dogs has also been documented [15, 16]. Although SFTSV presents a significant public health threat, no vaccines or specific antiviral medications have been approved thus far. Therefore, the World Health Organization considers SFTS as a disease requiring urgent research and development [17].

As a negative-stranded, segmented RNA virus (NSV), SFTSV possesses a genome consisting of small (S), medium (M), and large (L) segments. The S segment encodes two proteins: a nucleoprotein (NP) responsible for viral RNA encapsulation and complex formation and a non-structural protein that aids in viral immune evasion [18]. The M segment encodes the glycoprotein precursor that is cleaved into the glycoprotein N and C subunits by intracellular proteases during translation [19], which are crucial for receptor binding and membrane fusion [11, 20, 21]. The L segment is essential for transcription and genome replication, containing an N-terminal cap-dependent endonuclease (CEN), an RNA-dependent RNA polymerase (RdRP), and a cap-binding domain (CBD). Replication depends on the synthesis of the L protein, whereas transcription of the viral genome relies on capturing short-capped RNAs from host cell RNAs for primer-directed synthesis through the cap-snatching mechanism, a process that involves CBD and CEN functions [22, 23]. Similar to the structures of other reported NSV CENs [24, 25], SFTSV CEN has a PD-D/EXK nuclease superfamily structure, as well as a similar catalytic core and cation-binding folded structure [23]. The endonuclease activity of SFTSV CEN depends on the cofactor of divalent manganese ions, and its catalytic core contains conserved residues, including H80, P111, D112, D126, and K145, in the family Phenuiviridae [26]. Therefore, SFTSV CEN is an attractive target for antiviral drug discovery [27].

Currently, although the advancement of nucleic acid vaccine technology has led to the development of SFTSV mRNA vaccines, which have shown effectiveness in protection against SFTSV infection [28–30], no vaccine candidates have proceeded to clinical trials for evaluating efficacy. The broad-spectrum RdRP inhibitor Favipiravir (T-705) was recently demonstrated to confer protection against SFTSV in animals and humans, yet it may raise concerns regarding teratogenic effects due to associated mutagenesis [31]. A previous study has also shown that Baloxavir (BXA), the only marketed drug targeting influenza virus CEN, exhibits in vitro antiviral effects against SFTSV by inhibiting SFTSV CEN activity [32]. Although numerous candidate vaccines and drugs have been evaluated, none have reached the clinic, thus emphasizing an urgent need for discovering novel alternative candidates against SFTSV infection.

In this study, we focused on identifying inhibitors of the SFTSV CEN (CENis) from natural products known to harbor diverse chemical structures and various general acids capable of chelating divalent metal ions. Through a set of high-throughput screening assays based on our previously established in vitro enzymatic activity system targeting SFTSV CEN, several natural products were identified as hits from a large natural product library. Licoflavone C was subsequently demonstrated as a promising candidate in view of its potent antiviral activity both in vitro and in vivo. Further study demonstrated that it binds outside the active pocket of SFTSV CEN to interfere with its active conformation, thus further impacting RNA cleavage. A series of flavonoids highly similar in chemical structure to Licoflavone C also inhibited both CEN activity and SFTSV infection, thus exhibiting potential as excellent candidates.

Materials and methods

Cells and viruses

Vero and BHK-21 cells were maintained in a Dulbecco’s modified Eagle’s medium (DMEM, Gibco, NY, USA) supplemented with 10% fetal bovine serum (FBS) (Gibco, NY, USA) and 1% antibiotic/antimycotic (Gibco, NY, USA) at 37 °C in a humidified 5% CO₂ incubator.

SFTSV (HBMC16), Heartland virus (HRTV; MO-4), Guertu virus (GTV; DXM), and lymphocytic choriomeningitis virus (LCMV; Armstrong) were obtained from the National Viral Resource Center (Wuhan, China). LCMV, SFTSV, HRTV, and GTV were propagated in Vero cells, and LCMV was propagated in BHK-21 cells.

Plasmid constructs and recombinant protein expression

The 210-amino acid (aa) sequence of the SFTSV L gene fragment (1–210 aa) was inserted into the pET42a expression vector with an N-terminal Glutathione-S-Transferase (GST) as a purification tag (pET42a-GST-SFTSV CEN). The constructed expression plasmid was transformed into BL21(DE3) cells, induced with 0.2 mM isopropyl-β-d-thiogalactoside (IPTG), and expressed for 18 h at 16 °C. Similarly, LCMV, Junin virus (JUNV), Guanarito virus (GTOV), Chapare virus (CHAPV), Sabia virus (SABV), and Lujo virus (LUJV) (1–196 aa of the L segment), Machupo virus (MACV) (1–201 aa of the L segment), Lassa virus (LASV) (1–169 aa of the L segment), Dandenong virus (1–171 aa of the L segment), Hantaan Virus (HTNV) (1–180 aa of the L segment), and La Crosse virus (LACV) (1–185 aa of the L segment) CENs were cloned into the pET28a-sumo expression vector with an N-terminal 6× His purification tag. The constructed expression plasmids were transformed into BL21 (DE3) cells (Beyotime, Shanghai, China). LCMV CEN expression was induced with 1.0 mM IPTG at 37 °C for 4 h. The expression of other CENs was induced with 0.5 mM IPTG at 16 °C for 20 h. Subsequently, bacterial cells were collected and lysed through ultrasonication, and the target proteins were purified from the soluble supernatants using a GST- or Ni-affinity column (Protein Ark, Sheffield, UK).

High-throughput screening based on fluorescence resonance energy transfer (FRET)

A library containing 3467 monomer natural products was purchased from MedChemExpress (MCE, Shanghai, China). Compounds were dissolved in dimethyl sulfoxide (DMSO, Yeasen, Shanghai, China) and stored at −80 °C until use. High-throughput screening of the natural product library was conducted in a 384-well plate using a 20 µL system. Briefly, SFTSV CEN (1 µM) and each compound (50 µM) were incubated for 0.5 h at 37 °C, followed by the addition of a substrate (5′-FAM-AGGAAGAUUAAUAAUUUUCCU-BHQ1-3′, 0.1 µM) (Sangon Biotech, Wuhan, China) dissolved with reaction buffer (50 mM HEPES pH 7.8, 150 mM KCl and 1 mM MnCl₂) to reach a 20 µL reaction system. The reactions were monitored by detecting fluorescence intensity (λ_ex/λ_em = 485/535 nm) every 2 min on a microplate reader (SYNERGY H1, BioTek, VT, USA). The inhibition of SFTSV CEN enzymatic activity by the compounds was calculated from the relative fluorescence units at saturation.

$$V=\left(F_2-F_1\right)/\left(t_2-t_1\right)$$ 1

where V is the enzyme reaction rate, t is the reaction duration, t₁ is the reaction start time, t₂ is the reaction saturation time, F₁ is the fluorescence intensity at t₁, and F₂ is the fluorescence intensity at t₂.

$$\mathrm{Inhibition}\left(\%\right)=100-\left(V_{\mathrm{experimental}}/V_{\mathrm{control}}\right)\times 100$$ 2

Cellular antiviral assay conducted through real-time quantitative PCR (RT-qPCR)

Vero cells were seeded in 48-well plates at 1 × 10⁵ cells/well in a total volume of 200 μL and incubated overnight at 37 °C. Monolayer cells were treated with the compounds at a fixed or gradient-diluted concentration for 1 h and infected with SFTSV, GTV, or HRTV at a multiplicity of infection (MOI) of 0.1. After 1 h, the supernatant was replaced with a fresh medium supplemented with the same compounds. After 36 h, the supernatant from each well was collected, and total nucleic acids were extracted using a DNA/RNA Extraction Kit (Vazyme, Nanjing, China), according to the manufacturer’s instructions. For vRNA quantification, a standard curve was established with a plasmid containing the full-length NP gene of the respective virus. RT-qPCR was performed using the HiScript II One Step RT-qPCR SYBR Green Kit (Vazyme, Nanjing, China) on a QuantStudio6Pro (Thermo Scientific, MA, USA). An antiviral assay against LCMV was performed using the BHK-21 cells. The remaining procedures were performed as previously described.

Cytotoxicity assay

Vero cells seeded in 96-well plates were treated with gradient-diluted compounds for 24 h, and DMSO (1%) was used as a control. Subsequently, the culture medium was replaced with a fresh medium containing Cell Counting Kit-8 (CCK8) reagent (Vazyme, Nanjing, China). Next, the cells were incubated for an additional 2 h at 37 °C in the dark, and the absorbance at OD₄₅₀ was determined using a microplate reader (SYNERGY H1, BioTek, VT, USA). Cell availability was calculated as:

$$\mathrm{Cell\; availability}\left(\%\right)=\left({OD}_{\mathrm{experiment\; group}}-{OD}_{\mathrm{blank}}\right)/\left({OD}_{\mathrm{DMSO\; group}}-{OD}_{\mathrm{blank}}\right)\times 100.$$

Time of addition experiment

Vero cells seeded overnight in 48-well plates were infected with SFTSV (MOI = 0.1) for 1 h (0–1 h) in all groups, and four treatments conditions with the experimental compounds (Licoflavone C and T-705) were applied. These included pretreatment with the viruses (a: viruses were pretreated with compounds for 1 h and used to infect the cells), pre-treatment of the cells (b: −1 to 0 h), post-infection treatment (c: 1–36 h), and full-time treatment (d: −1 to 36 h). DMSO (0.1%) and T-705 served as the controls. After treatments, the cellular supernatant and cells were collected for viral copy detection using RT-qPCR and immunofluorescence.

Immunofluorescence assay (IFA)

Vero cells were fixed in 4% paraformaldehyde and permeabilized with 0.5% Triton X-100. Subsequently, the cells were blocked with 2.5% nonfat milk and stained with rabbit anti-NP polyclonal antibodies prepared in our laboratory, followed by treatment with anti-rabbit 488-conjugated secondary antibodies (Abcam, UK). The cell nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI) (Sigma-Aldrich, Shanghai, China). Images were captured using a Cytation Image Reader (BioTek, VT, USA).

Circular dichroism (CD) assay

SFTSV CEN protein was diluted to 0.2 mg/mL. Licoflavone C was co-incubated with SFTSV CEN in vitro for 0.5 h at final concentrations of 0, 50, and 100 µM. The samples in 0.1 cm cuvettes were measured at 190–260 nm wavelength using Chirascan V100 (Applied Photophysics, Surrey, UK) to collect CD signals with a 0.5 s interval. After determining the CD signal of the background solution, the data were analyzed using the CDNN 2.1 software (Applied Photophysics, Surrey, UK) and presented using GraphPad Prism 9.0.

Acrylamide-urea gel electrophoresis

SFTSV CEN (1 µM) was incubated with Licoflavone C at concentrations of 0, 50, 100, and 200 µM for 30 min. Subsequently, a single-stranded RNA substrate (5′-FAM-AGGAAGAUUAAUAAUUUUCCU-3′) diluted to 0.1 µM and dissolved with a reaction buffer (50 mM HEPES pH 7.8, 150 mM KCl, and 1 mM MnCl₂) was introduced. After incubation for 5, 10, and 20 min, the reaction products were subjected to acrylamide-urea gel electrophoresis to detect the cleavage of the RNA substrate by SFTSV CEN. The acrylamide-urea gel was formulated with 4.5 mL of 40% polyacrylamide solution, 1 mL of 10× Tris-Borate-EDTA (TBE) solution, 4.2 g of urea, 0.05 mL of 10% ammonium persulphate, and 0.005 mL of coagulant promoter. Diethyl pyrocarbonate (DEPC) water was added to a final volume of 10 mL. Samples from each time point were loaded into the gel wells, and electrophoresis was conducted in a TBE buffer for 20 min at 200 V. Images were captured using an Amersham Typhoon (Cytiva, MA, USA) and analyzed using ImageJ (Fiji).

Enzyme kinetic analysis

First, the initial reaction rates of SFTSV CEN at different concentrations (0, 0.025, 0.05, 0.1, and 0.2 µM) in the presence of Licoflavone C (0, 100, and 200 µM) were determined by fixing a constant concentration of an RNA substrate (0.1 µM) in the reaction system. The correlation between enzyme concentration and reaction rate was fitted using GraphPad Prism 9.0. Subsequently, the effect of Licoflavone C at different concentrations (0, 100, and 200 µM) on the initial reaction rate of SFTSV CEN in the presence of an RNA substrate at different concentrations (0, 0.25, 0.5, 1, 2, and 4 µM) was determined in a system where SFTSV CEN was fixed at a constant concentration of 0.2 µM. The correlation between the substrate concentration and reaction rate was fitted using GraphPad Prism 9.0.

Surface plasmon resonance (SPR)

SFTSV, LCMV, MACV, and LASV CENs were immobilized on a CM5 chip using an amino coupling kit (Cytiva, MA, USA) with a loading capacity of approximately 10,000 RU. Subsequently, Licoflavone C was diluted to 12.5, 6.25, 3.125, 1.5625, and 0 µM and flowed over the chip at 30 µL/min for 120 s each of association and dissociation on the Biacore 1 K system (Cytiva, MA, USA) at 25 °C. Data were analyzed using the Biacore evaluation software (version 1 K) fitted to a 1:1 binding model, and curves were drawn using GraphPad Prism 9.0.

Molecular docking

The 3D structure of Licoflavone C was downloaded in the SDF format from the PubChem database, and the SDF file was subsequently converted to the PDB format. The protein PDB structures of SFTSV CEN (6NTV) were obtained from the RCSB PDB. The compound and protein PDB files were converted to PDBQT files using the AutoDock Vina software. Active pockets of CENs were established for docking, and conformations with relatively low binding energies were analyzed using the online PLIP tool and mapped using PyMOL.

In vivo study

All animal experimental procedures were performed according to animal ethical guidelines and approved by the Animal Care Committee of the Wuhan Institute of Virology (WIVA25202403). Female C57BL/6 J mice (6–8 weeks of age) were purchased from GemPharmatech (Nanjing, China) and maintained in a specific-pathogen-free animal facility before use.

For in vivo screening of the three compounds, each group (n = 4) was infected with 1 × 10⁵ focus-forming units (FFUs) of SFTSV, followed by the administration of 20 mg/kg of 3,4-Dicaffeoylquinic acid and Licoflavone C via gavage (i.g.) and tail intravenously (i.v.), respectively, 20 and 50 mg/kg of Oleanolic acid via i.g. 2 h post-infection. T-705 (300 mg/kg, i.g.) served as a positive control. The compounds were administered once daily for three consecutive days. Viral loads were determined in the liver, spleen, kidney, and blood samples.

For in vivo investigation of Licoflavone C, mice were randomly divided into five groups (n = 5). Subsequently, 2 h after an intraperitoneal injection (i.p.) with 1 × 10⁴ FFUs of SFTSV, the mice were subjected to different treatments. The mice in the vehicle group received the solvent, those in the positive control group were orally administered 150 mg/kg T-705 every 12 h, and those in the experimental groups were intravenously administered a single dose of 5, 10, or 20 mg/kg Licoflavone C via tail vein injection every 12 h. Furthermore, 2 days later, the mice were euthanized to determine viral infection in their livers, spleens, kidneys, and blood. The harvested tissues were subjected to viral copy detection using RT-qPCR or immunofluorescence.

Tissue IFA

Animal tissues were fixed with 8% formaldehyde overnight, dehydrated with gradient ethanol, and frozen with an embedding agent at −20 °C. The paraffin blocks were sectioned into slices and soaked in 0.01 M citrate buffer (pH 6.0). The slices were permeabilized with 0.5% Triton X-100 for 30 min at room temperature and blocked with FBS. Finally, the slices were incubated with polyclonal antibodies against SFTSV NP and probed with secondary antibodies conjugated with Cy5. The cell nuclei were stained with DAPI. Images were captured using a fluorescence microscope (Panoramic MIDI, 3D HISTECH), and the fluorescence signals were analyzed using ImageJ (Fiji) based on multi-views from each mouse.

Pharmacokinetic experiment

Six female rats were evenly divided into two groups. Animals in the i.v. group had ad libitum access to food and water, while those in the i.g. group were fasted and did not receive water prior to administration, with normal feeding and water intake resumed two hours post-administration. Licoflavone C was administered via both i.g. and i.v. at a dose of 20 mg/kg, and blood samples were collected at 0, 5, 15, 20, and 30 min and at 1, 2, 3, 4, 5, 6, and 24 h post-administration. Plasma samples (10 µL) were mixed with an internal standard (25 µL), followed by the addition of acetonitrile (300 µL), vortexed for 10 min, and centrifuged at 1700 × g for 15 min in a refrigerated high-speed centrifuge at 4 °C. The supernatant (50 µL) was mixed with acetonitrile and water (2:8 ratio) containing 100 µg/mL VC, vortexed for 5 min, and prepared for LC-MS/MS analysis using the TRIPLE QUAD 6500 + AB SCIEX system (Sciex, Framingham, MA, USA). Pharmacokinetic parameters were calculated using the WinNonlin software (Version 8.3) by applying the statistical method of moments.

Statistical analysis

Plots and histograms were generated and visualized using GraphPad Prism 9.0. Statistical analyses were conducted using the t-test, with a P value of <0.05 considered statistically significant. Statistical significance is indicated as follows: ns, P > 0.05; *, P < 0.05; **, P < 0.01; ***, P < 0.001; and ****, P < 0.0001.

Results

High-throughput screening of SFTSV CEN inhibitors from natural compound library

Considering that natural products have advantages in chemical structural diversity and general acid abundance, we established a high-throughput FRET-based enzymatic screening system to discover inhibitors of SFTSV CEN from a natural product library and preliminarily validated these at the cellular level (Fig. 1a). First, 150 compounds were screened out from 3467 natural products that exhibited over 70% of enzymatic activity inhibition (Fig. 1b). Subsequently, 24 out of the 150 compounds were verified with antiviral activity against SFTSV through cellular assays, with inhibition on SFTSV infection exceeding 85% at 10 μM (Fig. 1c). Finally, three compounds out of the 24 were selected for further investigation due to their exceptional antiviral effects and minimal cytotoxicity thus resulting in selectivity indices (SIs) exceeding 100 (Table 1). Among the 24 compounds, Licoflavone C, 3,4-Dicaffeoylquinic acid, and Oleanolic acid were evaluated on Vero cells to determine their 50% effective concentrations (EC₅₀s) and 50% cytotoxic concentrations (CC₅₀s). The results revealed that 3,4-Dicaffeoylquinic acid exhibited an EC₅₀ of 1.39 μM with the highest SI greater than 216, Licoflavone C had an EC₅₀ of 1.85 μM with SI exceeding 162, and Oleanolic acid had an EC₅₀ of 2.95 μM with an SI surpassing 101 (Fig. 1d).

Fig. 1. Screening of natural products against severe fever with thrombocytopenia syndrome virus (SFTSV) in vitro.

a Flowchart of the screening process for discovering antiviral inhibitors against SFTSV. b Screening of 3467 natural products through fluorescence resonance energy transfer (FRET)-based high-throughput system against SFTSV cap-dependent endonuclease (CEN), 150 compounds with inhibition rate of >70% were marked with purple. c The antiviral activity of 150 compounds was tested with live SFTSV on Vero cells through RT-qPCR. The top 30 compounds with an inhibition rate of >80% were marked with red. d Structural formulae of 3,4-Dicaffeoylquinic acid, Licoflavone C, and Oleanolic acid and their curves of cytotoxicity (CC₅₀) and activity (EC₅₀) were tested with live SFTSV on Vero cells.

Table 1.

A list of toxicity (CC₅₀) and activity (EC₅₀) of 30 compounds in the order of selective index (SI).

Name	EC50 (μM)	CC50 (μM)	SI	Name	EC50 (μM)	CC50 (μM)	SI
3,4-Dicaffeoylquinic acid	1.39	>300.0	>215.83	1,4-Dicaffeoylquinic acid	15.25	>300.00	>19.67
Licoflavone C	1.85	>300.00	>162.16	Proanthocyanidins	16.88	202.25	11.98
Oleanolic acid	2.95	>300.00	>101.69	Tannic acid	3.77	39.09	10.37
Demethylzeylasteral	0.49	41.42	84.53	Salvianolic acid A	37.10	296.35	7.99
Naringin	3.69	>300.00	>81.3	Gamma-Mangostin	1.53	10.59	6.92
3′-Hydroxypterostilbene	1.50	96.93	64.62	Tripterin	0.46	2.43	5.28
Vanillin	4.67	>300.00	>64.24	Punicalagin	63.48	>300.00	>4.73
Hypericin	0.49	31.40	64.08	Thonningianin A	2.65	11.25	4.25
Luteolin	2.56	129.12	50.44	Tellimagrandin II	4.49	18.11	4.03
Marein	6.28	>300.00	>47.77	Licochalcone D	10.95	42.43	3.87
2′-O-Galloylhyperin	6.93	>300.00	>43.29	Glycyrrhisoflavone	18.09	47.02	2.60
7,8-Dihydroxyflavone	8.85	>300	>33.90	Favipiravir	28.20	>300.00	>10.64
Rottlerin	0.52	12.80	24.62

Favipiravir (T-705) was used as a positive control.

Licoflavone C was confirmed as a promising candidate to inhibit SFTSV infection in vivo

To confirm the antiviral activity of the three hits from the above screen, 3,4-Dicaffeoylquinic acid, Licoflavone C, and Oleanolic acid were evaluated in an SFTSV-challenged C57BL/6 J mouse model. As described in the schematic diagram (Fig. 2a), mice were treated with 3,4-Dicaffeoylquinic acid, Licoflavone C, and Oleanolic acid via i.g. or i.v. 2 h after the SFTSV challenge. Mice body weights were recorded daily, and the liver, spleen, kidney, and blood samples were collected 3 days post-infection (dpi) for viral copy number quantification. From the body weight curves, we found that SFTSV challenge led to mild weight loss in mice on the first day, which rapidly recovered in the subsequent two days, whereas treatments with 3,4-Dicaffeoylquinic acid, Licoflavone C, Oleanolic acid, and the positive control T-705 had little influence on the treated mice (Fig. 2b). The viral load results showed that T-705 exhibits a significant antiviral effect, reducing viral copies in the liver, spleen, kidney, and blood, when administered at 300 mg/kg via i.g.. In comparison, 3,4-Dicaffeoylquinic acid, Licoflavone C, and Oleanolic acid elicited no obvious reduction in viral copies when administered at 20 mg/kg via i.g.. However, the i.v. administration route resulted in a greater effect than i.g., especially for Licoflavone C, which decreased the viral loads in all analyzed tissues (Fig. 2c). This was corroborated by the pharmacokinetic results showing a poor oral bioavailability and a short half-life when Licoflavone C was administered via i.g. (Fig. 2d).

Fig. 2. Antiviral efficacy of 3,4-Dicaffeoylquinic acid, Licoflavone C, and Oleanolic acid against SFTSV in C57BL/6 J mice.

a Schematic diagram of animal experiments. Female C57BL/6 J mice (n = 4) were injected intraperitoneally with 1 × 10⁵ FFUs of SFTSV. Subsequently, 2 h post-infection, 3,4-Dicaffeoylquinic acid and Licoflavone C were administered i.g. and i.v. at 20 mg/kg. Oleanolic acid was administered i.g. at 20 and 50 mg/kg. Viral copy numbers in the liver, spleen, kidney, and blood were measured on the third day post-infection. b Changes in the body weights of mice challenged with SFTSV after treatments with 3,4-Dicaffeoylquinic acid, Licoflavone C, and Oleanolic acid. T-705 was used as a positive control. c Viral copies in the liver, spleen, kidney, and blood were determined through RT-qPCR analysis of the nucleoprotein (NP) gene of SFTSV. d Pharmacokinetic data for Licoflavone C in rats following administration via both i.v. and i.g. at a dose of 20 mg/kg. ns, P > 0.05; and *, P < 0.05.

Licoflavone C was thus subjected to further in vivo analysis. Prior to the careful efficacy study, we preliminarily investigated its toxicity in mice. From a two-week observation after i.v. administration, we found that a single dose lower than 20 mg/kg had minimal influence on mice body weights or behaviors (Supplementary Fig. S1), implying that 20 mg/kg is a safe dose in mice. Mice were then challenged with 1 × 10⁴ FFUs of SFTSV, and viral loads were detected at dpi 2. Licoflavone C was administered at optimal doses as 5, 10, or 20 mg/kg via i.v., and T-705 was administered at 150 mg/kg twice a day via i.g. (Fig. 3a). SFTSV challenge or drug administration had minimal influence on mouse body weight in two days (Fig. 3b). As for the viral loads, the positive control T-705 significantly reduced SFTSV infection in the liver, spleen, kidney, and blood. Similarly, Licoflavone C decreased the viral loads in these target tissues in a dose-dependent manner (Fig. 3c). Correspondingly, SFTSV infection in the spleen visualized through IFA revealed the same tendency: T-705 and Licoflavone C decreased the fluorescence intensity represented for SFTSV infection (Fig. 3d, e). These results collectively confirmed that Licoflavone C is a promising candidate for further studies.

Fig. 3. Evaluation of the antiviral activity of Licoflavone C in vivo.

a C57BL/6 J mice (n = 5) were challenged with 1 × 10⁴ FFUs of SFTSV via i.p.. Subsequently, 2 h post-challenge, Licoflavone C was administered twice daily via i.v. at 5, 10, and 20 mg/kg. Viral copies in the liver, spleen, kidney, and blood were determined 2 days post-challenge. T-705 was used as a positive control and was administered twice daily via i.g. at 150 mg/kg. b Body weight changes in mice over 2 days. c Quantification of SFTSV viral copy numbers in mouse liver, spleen, kidney, and blood via RT-qPCR. d Immunofluorescence analysis of SFTSV infection in the spleen after treatments with Licoflavone C or T-705. Red indicates SFTSV NP, and blue indicates the nucleus. e The red fluorescent signals in mice from each group were analyzed. ns, P > 0.05; *, P < 0.05; and **, P < 0.01.

Licoflavone C inhibits SFTSV infection at the post-infection stage

To investigate the mechanism of action of Licoflavone C, a time-of-addition experiment was conducted, assessing the disruption of virion stability, viral entry blockade, or post-entry effect (Fig. 4a). Licoflavone C had a minimal effect on virion stability or viral entry, but significantly inhibited SFTSV infection in the post-infection and full-time treatment modes, similar to the pattern of T-705, which is known to act at the viral replication stage [33–35] (Fig. 4b). These results were further confirmed by detecting intracellular viral proteins using IFA (Fig. 4c). Moreover, we determined the antiviral activity of Licoflavone C in the four treatment modes (Fig. 4a) and found that the post-infection treatment mode exhibited the most similar EC₅₀ to that of the full-time treatment mode, whereas pre-treatment of the viruses or cells had a larger or unmeasured EC₅₀ under the same concentration range (Fig. 4d). Furthermore, we evaluated its inhibitory effect at various timepoints of addition post viral infection (Fig. 4e). The results showed that adding Licoflavone C (5, 20, or 50 μM) at 1, 2, 4, 8, and 12 h post-infection did not influence inhibition significantly. In contrast, there was a notable decrease in inhibition when Licoflavone C was added at 24 h or later (Fig. 4f). These results indicate that Licoflavone C effectively impeded the viral proliferation of SFTSV at the post-entry stage.

Fig. 4. Licoflavone C inhibits SFTSV at the post-infection stage.

a Schematic representation of four treatment modes: pretreatment of virus, pretreatment of cells, treatment post-infection, and full-time treatment. The red line represents the duration of SFTSV infection, whereas the black line represents the duration of Licoflavone C treatment. b The inhibition of Licoflavone C on SFTSV under four treatment modes was determined through RT-qPCR, with T-705 as a positive control. c The inhibition of SFTSV by Licoflavone C under the four treatment modes was assessed through IFA. Green indicates SFTSV NP, and blue indicates the nucleus. d The EC₅₀ values of Licoflavone C against SFTSV under four treatment modes. e The adding patterns of Licoflavone C during 48 h. f The inhibition of Licoflavone C at 5, 10, and 20 µM when added at 0, 1, 2, 4, 8, 12, 24, and 36 h after SFTSV infection. ns, P > 0.05; **, P < 0.01; and ***, P < 0.001; and ****, P < 0.0001.

Licoflavone C targets SFTSV CEN for reversible, non-competitive inhibition

As Licoflavone C was screened out against SFTSV CEN and was shown to act at the post-infection stage, further experiments were performed to investigate the interaction between Licoflavone C and SFTSV CEN. The inhibitory effect of Licoflavone C on the enzymatic activity of SFTSV CEN was examined using an in vitro enzymatic activity detection system based on gel electrophoresis. As shown in Fig. 5a and the corresponding analysis in Fig. 5b, Licoflavone C inhibited the cleavage of the RNA substrate by SFTSV CEN in a concentration- and time-dependent manner. Moreover, its 50% inhibitory concentration (IC₅₀) against SFTSV CEN was determined as 35.5 µM (Fig. 5c). Subsequently, the binding of Licoflavone C to SFTSV CEN protein was evaluated using SPR, which revealed that Licoflavone C exhibited fast association and dissociation with SFTSV CEN protein, with a K_D of 9.53 µM (Fig. 5d).

Fig. 5. Licoflavone C binds to SFTSV CEN in a substrate non-competitive manner.

a Efficiency of Licoflavone C at 0, 50, 100, and 200 µM in inhibiting the RNA cleavage activity of SFTSV CEN was detected via electrophoresis at 5, 10, and 20 min. Ethylenediaminetetraacetic acid (EDTA) was used as a positive control. b Gray analysis of the bands in (a). c The IC₅₀ of Licoflavone C on SFTSV CEN detected through FRET. Baloxavir (BXA) was used as a control. d The affinity of Licoflavone C for SFTSV CEN was measured through SPR. e The reaction rates of SFTSV CEN in a system with varied Licoflavone C and fixed substrate concentrations. f The reaction rates of SFTSV CEN in a system with varied Licoflavone C and fixed enzyme concentrations. g Molecular docking of the interaction between Licoflavone C and SFTSV CEN. The blue dashed lines indicate hydrogen-bonding interactions; the gray dashed lines indicate hydrophobic interactions. h The impact of Licoflavone C on the secondary structures of SFTSV CEN at varying concentrations was detected using CD. *, P < 0.05; **, P < 0.01; and ***, P < 0.001; and ****, P < 0.0001.

Subsequently, we investigated how Licoflavone C interacts with SFTSV CEN to inhibit its enzymatic activity by implementing a set of enzymatic kinetic experiments by adjusting the concentrations of the enzyme, substrate, and compound. The results showed that Licoflavone C reversibly targets SFTSV CEN (Fig. 5e), consistent with the results shown in Fig. 5d. Furthermore, it was shown to act in a substrate-non-competitive mode (Fig. 5f). To verify its interaction with SFTSV CEN, we simulated the binding between Licoflavone C and SFTSV CEN through molecular docking. The analysis revealed that Licoflavone C engaged with CEN primarily at N64, T69, and R194 through hydrogen bonds, participating in hydrophobic interactions at N64, I70, I78, D81, and F82 in the lowest binding energy mode (Fig. 5g). Notably, these amino acids were not positioned at the active center pocket of SFTSV CEN, which was coincidently confirmed by Licoflavone C targeting SFTSV CEN in a substrate non-competitive manner. Hence, we hypothesized that Licoflavone C influences the enzymatic activity of SFTSV CEN by altering its active conformation. Through a set of CD assays to identify the secondary structure, we found that the addition of Licoflavone C interfered with the conformation of SFTSV CEN by interrupting partial parallel β-sheets (Fig. 5h). These results verified that Licoflavone C targets SFTSV CEN in a reversible, non-competitive manner to exert inhibitory effects on enzymatic activity and viral infection.

Licoflavone C elicits broad-spectrum inhibition on CENs of negative-strand, segmented RNA viruses

Next, we investigated the broad-spectrum antiviral activity of Licoflavone C on NSVs other than SFTSV. Thus, representative NSVs, such as HRTV and GTV, which are evolutionarily close to SFTSV within the family Phenuiviridae, and LCMV within the family Arenaviridae, were investigated. The cellular antiviral assays showed that Licoflavone C exhibited antiviral activity against these NSVs with EC₅₀s of 9.21, 4.21, and 2.61 µM against HRTV, GTV, and LCMV, respectively (Fig. 6a). From the enzymatic assays, we found that Licoflavone C inhibits the enzymatic activity of LCMV, MACV, and LASV CENs within the Arenaviridae at IC₅₀s of 135.8, 177.7, and 547.6 µM, respectively (Fig. 6b). Correspondingly, we determined the affinity of Licoflavone C for LCMV, MACV, and LASV CENs through SPR, and found that there were moderate interactions between Licoflavone C and LCMV and MACV CENs, with K_D values of 59.1 and 82.7 µM, respectively, while the affinity of Licoflavone C for LASV CEN was undetectable at the used concentrations (Fig. 6c).

Fig. 6. Licoflavone C has a broad-spectrum inhibitory activity against negative-strand, segmented RNA viruses.

a The EC₅₀ values of Licoflavone C against HRTV, GTV, and LCMV were determined through RT-qPCR. b The IC₅₀ values of Licoflavone C on LCMV, MACV, and LASV CEN were measured through FRET. c The binding affinity of Licoflavone C for LCMV, MACV, and LASV CEN, as determined through SPR.

To further expand the investigation scope of the broad-spectrum effect of Licoflavone C, we assessed its IC₅₀ on various CENs from pathogens from the family Arenaviridae, including JUNV, GTOV, CHAPV, SABV, LUJV, and Dandenong, HTNV from the family Hantaviridae, LACV from the family Peribunyaviridae, and CCHFV from the family Nairoviridae. Licoflavone C exhibited extensive inhibitory effects on CENs of these NSVs, being more active against pathogens, such as SFTSV, and the New World arenaviruses, including MACV, JUNV, GTOV, CHAPV, and SABV (Table 2 and Supplementary Fig. S2). Taken together, these data demonstrated that Licoflavone C has potential broad-spectrum antiviral activity against NSVs besides SFTSV, implying a wide application value of Licoflavone C in antiviral studies.

Table 2.

The IC₅₀s of Licoflavone C against CENs of pathogens from the families Arenaviridae, Hantaviridae, Peribunyaviridae, and Nairoviridae.

CEN	Phenuiviridae			Hantaviridae			Peribunyaviridae			Nairoviridae
Licoflavone C (µM)	SFTSV			HTNV			LACV			CCHFV
	35.5			278.1			>1000			118.8

	Arenaviridae
	MACV	JUNV	GTOV		CHAPV	SABV		LCMV	LASV		LUJV	Dandenong
	177.7	64.2	85.2		32.2	57.8		135.8	547.6		>1000	109.0

Natural analogs of Licoflavone C exert extensive inhibition on SFTSV CEN

Finally, 15 compounds that are structurally similar to Licoflavone C were selected to explore the structure-function relationship between flavonoids and their effects on SFTSV CEN enzymatic activity. Through a set of in vitro enzymatic and cellular assays, the IC₅₀, EC₅₀, and CC₅₀ of the 15 compounds were determined (Table 3). These flavonoid analogs exhibited inhibitory effects on SFTSV infection with comparable or higher EC₅₀ values than Licoflavone C while having smaller CC₅₀ values, resulting in unacceptable SIs.

Table 3.

Fifteen analogs of Licoflavone C were tested with their inhibitory effects on both CEN activity and SFTSV infection.

Name	Structure	IC50 (µM)	EC50 (µM)	CC50 (µM)	SI	Name	Structure	IC50 (µM)	EC50 (µM)	CC50 (µM)	SI
Licoflavone C		35.50	1.85	>300.00	>162.16	Licoisoflavone A		94.15	5.00	40.60	8.12
Sophoraflavanone G		11.20	0.96	6.74	7.02	Sophoranone		122.60	4.00	15.80	3.95
Licoflavonol		16.36	6.86	42.97	6.26	Luteolin		133.20	4.32	90.80	21.02
Mulberrin		18.80	1.53	31.80	20.78	Glabrol		162.30	1.69	28.65	16.95
Kushenol A		19.63	2.34	13.62	5.82	Isobavachin		221.10	8.00	102.80	12.85
Glycyrrhisoflavone		46.45	32.00	124.40	3.89	Leachianone G		300.20	6.90	99.60	14.43
Licoflavone A		65.76	2.48	46.90	18.91	Broussoflavonol F		318.80	10.70	164.80	15.40
Licoflavone B		81.04	1.22	17.90	14.67	Isoxanthohumol		394.90	9.60	192.50	20.05

Their cytotoxicity was tested on Vero cells. IC₅₀s, EC₅₀s, and CC₅₀s were listed in the order of IC_50.

In addition, all 15 flavonoid analogs manifested inhibitory effects on the enzymatic activity of SFTSV CEN, among which Sophoraflavanone G, Licoflavonol, Mulberrin, and Kushenol A showed IC₅₀ values superior to that of Licoflavone C, whereas Isoxanthohumol did the contrary. Notably, comparing the chemical structures of Sophoraflavanone G, Licoflavonol, Mulberrin, and Kushenol A to that of Licoflavone C (Table 4), we could conclude that extending the existing hydrophobic side chain or adding hydrophobic side chains based on the flavonoid structure to enhance the fat solubility of Licoflavone C and facilitate the formation of tetrahydro-4H-pyran-4-one may improve the inhibitory effect on SFTSV CEN, while also enhancing its cytotoxicity. However, comparing the chemical structure of Isoxanthohumol to that of Licoflavone C, destroying the phenol hydroxyl group of the flavonoid structure may impair its inhibitory effect on SFTSV CEN, which may be caused by the removal of hydrogen bonding interactions with the T69 site within CEN (Fig. 5g). Hence, flavonoids, including Licoflavone C, have extensive inhibitory effects on SFTSV and its CEN, and optimization on the hydrophobic side chain of the flavonoid scaffold may be a feasible way to enhance its efficacy.

Table 4.

Chemical structure analysis of the top four analogs and the last one from Table 3.

		R1	R2	R3	R4	R5
Licoflavone C		−H			−H	−OH
Sophoraflavanone G		−H2			−H	−OH
Licoflavonol		−OH	−H			−OH
Mulberrin					−H	−OH
Kushenol A		−H2			−H	−OH
Isoxanthohumol		−H2			−H	−OCH3

Discussion

NSVs, including representative viruses such as influenza virus, bunyavirus, and arenavirus, utilize their encoded endonucleases to initiate viral genome transcription by cleaving the 5′ cap structure from the host’s RNAs in a process known as cap-snatching [36]. Given the pivotal role of CEN in the SFTSV life cycle, efforts towards CEN offers a promising avenue for antiviral drug discovery and design [37]. The only commercially available CEN-targeted drug is baloxavir (BXA), which was developed to specifically target the influenza virus PA_N [38, 39]. BXA demonstrates potent antiviral efficacy against influenza virus [39] and exhibits some broad-spectrum antiviral activity against other NSVs [32, 40, 41]. At present, potent drug candidates targeting the SFTSV CEN are insufficient, whereas the success of BXA against influenza provides a basis for developing CENis for SFTSV and other NSVs.

Currently, three primary methodologies have been reported for the discovery of CENi: target structure-based drug design, ligand-based drug design, and target-based drug screening [42, 43]. Structure and ligand-based drug design are characterized by target specificity and structural innovation; however, they are intricate, time-consuming, and associated with a low success rate. In contrast, target-based drug screening offers efficient and extensive screening capabilities, although it is challenged by compound resource limitation [44–46]. In this study, a target-based drug screening approach was applied for high-throughput screening of candidate compounds from natural product libraries. Natural products are known to possess significant medicinal value; the successful application of Artemisinin in malaria is a good example [47]. Additionally, we previously reported the antiviral properties of natural products, such as Salvianolic C and Tanshinone I, by targeting viral proteins to exhibit anti-SARS-CoV-2 and anti-NSVs activities, respectively [48, 49]. Based on the recognition of the inherent antiviral properties of natural products, we identified 3,4-Dicaffeoylquinic acid, Licoflavone C, and Oleanolic acid from 3467 natural compounds, confirming the antiviral efficacy of Licoflavone C against SFTSV in vitro and in vivo. Furthermore, we validated the broad-spectrum inhibitory effect of Licoflavone C on the NSV CENs. We propose Licoflavone C as a promising lead compound for SFTSV, representing a feasible approach for the discovery of CENi from natural products.

Among the reported SFTSV CENis, BXA exhibits an EC₅₀ value of 250 nM against SFTSV [32]. L-742,001 displayed inhibitory potency against SFTSV CEN with an IC₅₀ of 2.1 μM, in addition to limited antiviral effectiveness against SFTSV with an EC₅₀ exceeding 50 μM [32]. Tanshinone I, as reported in our previous study, exhibited broad-spectrum effects at micromolar levels against NSVs, such as SFTSV, LCMV, and influenza A virus [49]. Herein, Licoflavone C was reported to inhibit SFTSV at the micromolar level and has selective broad-spectrum antiviral activity, particularly against viruses such as SFTSV and the New World arenaviruses. Nevertheless, our study revealed that Licoflavone C demonstrates inferior inhibition of CEN activity when compared with BXA. Animal studies also demonstrated that Licoflavone C may have a narrow dose range with limited administration routes. Therefore, we posited that an optimization of Licoflavone C through side chain replacement on the flavonoid scaffold may improve its activity, toxicity profile, and oral availability.

Notably, our study revealed that Licoflavone C does not directly bind to the active pocket of SFTSV CEN. Rather, it disrupts the active conformation of SFTSV CEN, resulting in substrate non-competitive inhibition, which differs from the mechanisms of most previously reported CENis that predominantly function as metal ion-chelating agents. Although chelating metal ions is an efficient approach to deactivating metal ion enzymes such as CENs, it may also introduce a foreseeable risk of widespread toxicity owing to non-specific chelation. In contrast, specific binding to CEN and interference with its active conformation may aid in reducing the side effects of chelating agents. Additionally, targeting sites outside the active pocket to develop novel CENis may reduce the possibility of drug resistance. Therefore, we report a CENi acting through a novel mechanism, thus providing clues for discovering CENis that act through a different approach.

In conclusion, we identified Licoflavone C as an inhibitor of SFTSV CEN and validated its antiviral activity against SFTSV in vitro and in vivo. Moreover, this study provides a basis for the development of novel CENis from natural products, particularly flavonoids.

Author contributions

XG and XXH designed and performed most of the biochemical, cellular, and animal experiments and drafted the manuscript; XRZ helped with the docking analysis; YW, JL, XLC and CSZ assisted with plasmid construction and protein expression and purification; HYL and ZFZ helped with data analysis; SWL and GFX coordinated the project; XYP contributed to the interpretation of the results, revised, and approved the manuscript.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41401-025-01533-7.