Bacillus subtilis partially inhibits African swine fever virus infection in vivo and in vitro based on its metabolites arctiin and genistein interfering with the function of viral topoisomerase II

ABSTRACT

African swine fever virus (ASFV) is an infectious disease with a mortality rate of nearly 100% in pigs; however, no safe commercial vaccines or antiviral drugs are currently available, which seriously threatens the global pig industry. Therefore, effective biologics against ASFV are urgently needed. Here, we screened 138 Bacillus subtilis strains, four of which evidently inhibited ASFV replication in vitro. Pigs fed with biologics of different types from the four B. subtilis strains showed reduced pathological changes and viral loads in tissues, with a survival rate of up to 100%. The antiviral activity of B. subtilis was attributed to small-molecule metabolites, rather than to secretory proteins. A total of 169 small molecules were obtained from the metabolites of B. subtilis using liquid chromatograph mass spectrometer/mass spectrometer (LC-MS/MS), arctiin, and genistein, which showed the highest inhibition efficiency, suppressing ASFV proliferation at the mid-stage of infection. These molecules acted as competitive inhibitors by complexing the ATP-binding domain of viral topoisomerase II, as demonstrated using molecular docking, biolayer interferometry binding, and a competitive decatenation assay, thereby disrupting the catalytic activity of the enzyme and inhibiting ASFV replication. Furthermore, pigs administered arctiin and genistein orally showed decreased mortality and tissue damage. Collectively, these results suggest that the four B. subtilis strains screened may be preventative biologics against ASFV infection. Our findings pave the way for ASFV prevention and control strategies in the pig industry to curb the economic losses caused by the disease.

IMPORTANCE

African swine fever virus (ASFV) is a highly fatal swine disease that severely affects the pig industry. Although ASFV has been prevalent for more than 100 years, effective vaccines or antiviral strategies are still lacking. In this study, we identified four Bacillus subtilis strains that inhibited ASFV proliferation in vitro. Pigs fed with liquid biologics or powders derived from four B. subtilis strains mixed with pellet feed showed reduced morbidity and mortality when challenged with ASFV. Further analysis showed that the antiviral activity of B. subtilis was based on its metabolites arctiin and genistein interfering with the function of viral topoisomerase II. Our findings offer a promising new strategy for the prevention and control of ASFV that may significantly alleviate the economic losses in the pig industry.

INTRODUCTION

African swine fever (ASF), an acute, highly contagious disease affecting domestic pigs and wild boars, presents a significant threat to the global swine industry owing to its high morbidity and mortality rates of up to 100% (1). The causative agent of ASF is African swine fever virus (ASFV), which belongs to the Asfarviridae family and is the only known DNA arbovirus (2). ASFV is an enveloped virus with a double-stranded genome between 170 and 190 kbp in length, encoding more than 150 open reading frames depending on the strain (3, 4). Although some ASFV proteins are involved in viral replication and immune evasion associated with the regulation of interferon (IFN)-I production, autophagy, inflammation, and apoptosis, the functions of more than half of the ASFV proteins remain unknown (5, 6). Some viral proteins that play a vital role in ASFV replication have been used as antiviral targets, such as cysteine protease pS273R, type II topoisomerase, dUTPase, DNA ligase, and DNA polymerase X (7 – 10). The function of ASFV type II topoisomerase is consistent with that of mammalian topoisomerase II (9). Topoisomerase II has the capability to modulate the topology of DNA molecules and allow the decatenation and proper segregation of intertwined DNA molecules following replication (11). Topoisomerase II is known to be essential in unicellular and multicellular organisms, and constitutes a target in antibacterial and anti-cancer treatments (12, 13). In recent years, it has been shown that the catalytic activity of ASFV topoisomerase II is disrupted by genistein, thereby inhibiting viral proliferation (14). In particular, viral type II topoisomerase possesses a domain that requires ATP to exert its catalytic activity for viral replication (15).

ASF was first reported in Kenya in 1921, before spreading beyond Africa to Europe, being reported in Portugal in 1960, and spreading throughout the entire Iberian Peninsula in 1995 (3). In 2007, ASFV was again reported outside Africa in the Caucasus region of Georgia, where it rapidly spread to neighboring countries; it then emerged in the Russian Federation in November 2007 (16, 17). ASFV was first recorded in China in August 2018 and was subsequently reported in other Asian countries, including Vietnam, Mongolia, and Cambodia (18, 19).

China accounts for 45% of global pork production; however, ASF has led to huge economic losses in China. Vaccines against ASFV, including subunit vaccines, live attenuated vaccines, inactivated vaccines, and virus-vectored vaccines, have been studied; however, safe and effective commercial vaccines have not been applied in the pig industry (8, 20). In this context, studies on antiviral drugs targeting conserved critical viral proteins are crucial for the prevention and control of ASFV (21, 22). Thus, new anti-ASFV strategies are urgently required.

Natural metabolites from plants, animals, and microbes exhibit excellent antiviral and anti-cancer activities. The N6-(Δ²-isopentenyl) adenosine of human gut microbiome-derived metabolites has been reported to prevent severe acute respiratory syndrome coronavirus 2 infection (23). Moreover, four matrine-based alkaloids from the seeds of Sophora alopecuroides suppress hepatitis B virus replication (24). Urolithin A from pomegranate inhibits cancer growth by activating mitochondrial autophagy (25). Meanwhile, the metabolites of Bacillus subtilis, a probiotic with excellent antiviral properties, have been shown to inhibit various viruses, including viral hemorrhagic septicemia virus, porcine epidemic diarrhea virus, transmissible gastroenteritis virus, and influenza virus (26 – 28). The advantages of B. subtilis as a probiotic are its low cost of preparation, ease of storage, and ability to produce metabolites, all of which make B. subtilis suitable for viral prevention and control.

In this study, we screened four B. subtilis strains and assessed their effects on ASFV infection using both in vitro and in vivo experiments. We found that arctiin and genistein, small-molecule metabolites of B. subtilis, compete for ATP binding to the ATP-binding domain of ASFV type II topoisomerase, thus suppressing the catalytic activity of type II topoisomerase and ASFV proliferation. Our data pave the way for the use of B. subtilis probiotics as an antiviral approach for preventing ASFV infection in the pig industry.

MATERIALS AND METHODS

Cells and viruses

Primary porcine alveolar macrophages (PAMs) were acquired from 1-month-old weanling pigs (negative for both ASFV antigen and antibody) by bronchoalveolar lavage, as previously described (29). Porcine peripheral blood mononuclear cells (PBMCs) were collected from EDTA-treated swine blood using a pig PBMC isolation kit (TBD Sciences, Tianjin, China). Primary bone marrow-derived macrophages (BMDMs) were obtained as described in previous research (30). PAMs, PBMCs, and BMDMs were cultured in RPMI 1640 culture medium (Gibco, Waltham, MA, USA) supplemented with 10% fetal bovine serum (Gibco), 100 µg/mL streptomycin, and 100 IU/mL penicillin at 37°C in a 5% CO₂ atmosphere. The ASFV SY-1 (GenBank accession number: OM161110) strain was propagated in PAMs for amplification and stored at −80°C until use.

LC-MS/MS analysis of B. subtilis metabolites

To analyze the small-molecule metabolites of B. subtilis that inhibit ASFV proliferation, LC-MS/MS was performed. The fermentation broths of four B. subtilis strains with the highest inhibition efficiency and six B. subtilis strains without inhibition efficiency were collected. For each B. subtilis strain, 10⁶ colony-forming units (CFUs) were inoculated into 5 mL of nutrient broth medium. After 24 h, 5 mL of the supernatant was collected and freeze-dried. The samples were resuspended in 1 mL of RPMI 1640 medium and 4 mL of pre-chilled methanol by vortexing, then incubated on ice for 5 min, and centrifuged at 15,000 × g at 4°C for 15 min. Some of the supernatant was diluted to a final concentration in 53% methanol using LC-MS-grade water. The samples were subsequently transferred to fresh Eppendorf tubes and centrifuged at 15,000 × g at 4°C for 15 min. Finally, the supernatant was injected into an LC-MS/MS system for analysis.

Ultra high performance liquid chromatography (UHPLC)-MS/MS analyses were performed using a Vanquish UHPLC system (Thermo Fisher Scientific, Waltham, MA, USA), coupled with an Orbitrap Q Exactive HF mass spectrometer (Thermo Fisher Scientific), at Novogene Co. Ltd (Beijing, China). Samples were injected onto a Hypersil Gold column (100 × 2.1 mm, 1.9 µm) using a 17-min linear gradient at a flow rate of 0.2 mL/min. The eluents for the positive polarity mode were eluents A (0.1% formaldehyde in water) and B (methanol), whereas those for the negative polarity mode were eluentsA (5 mM ammonium acetate, pH 9.0) and B (methanol). The solvent gradient was set as follows: 2% B for 1.5 min, 2%–100% B for 3 min, 100% B for 10 min, 100%–2% B for 10 min, and 2% B for 12 min. The Q Exactive HF mass spectrometer was operated in the positive/negative polarity mode, with a spray voltage of 3.5 kV, capillary temperature of 320°C, sheath gas flow rate of 35 psi, auxiliary gas flow rate of 10 L/min, S-lens RF level of 60, and Aux gas heater temperature of 350°C.

The raw data files generated by UHPLC-MS/MS were processed using Compound Discoverer 3.1 (Thermo Fisher Scientific) to perform peak alignment, peak picking, and quantitation for each metabolite. The main parameters were set as follows: retention time tolerance of 0.2 min, actual mass tolerance of 5 ppm, signal intensity tolerance of 30%, and signal/noise ratio of 3.0 Peak intensities were normalized to the total spectral intensity. The normalized data were used to predict the molecular formula based on additive ions, molecular ion peaks, and fragment ions. The peaks were then matched with the mzCloud, mzVaultand, and MassList databases to obtain accurate qualitative and quantitative data. Statistical analyses were performed using the statistical software R (R version R-3.4.3), Python (Python 2.7.6 version), and CentOS (CentOS release 6.6). When data were not normally distributed, normal transformations were attempted using the area normalization method.

Screening metabolites from B. subtilis and small-molecule metabolites against ASFV infection

To evaluate the metabolites of diverse B. subtilis strains inhibiting ASFV replication, for each B. subtilis strain, 10⁶ CFUs were inoculated in 5 mL of nutrient broth medium. After 24 h, the supernatants were collected and freeze-dried. The freeze-dried samples were dissolved in RPMI 1640 medium at a final concentration of 1 µg/mL to assess the antiviral effect. The PAMs were pre-seeded in 24-well plates (2 × 10⁵ cells per well). PAMs were infected with a 0.1 multiplicity of infection (MOI) of ASFV for 1 h, washed, and cultured in fresh medium containing the metabolites of diverse B. subtilis strains for 72 h. The supernatant was collected and analyzed by quantitative real-time PCR (qPCR). The inhibition rate of each metabolite from B. subtilis was calculated using the copy number of ASFV genomic DNA from PAMs treated with metabolites compared with the vehicle.

To analyze the components of the metabolites from the four screened B. subtilis anti-ASFV infections, the proteins and small molecules in the metabolites were separated according to a previous protocol (31). Briefly, 10 mg of freeze-dried metabolites was dissolved in 1 mL of RPMI 1640 medium and 4 mL pre-chilled methanol by vortexing. The samples were then incubated on ice for 5 min and centrifuged at 15,000 × g at 4°C for 15 min. The supernatant without proteins was collected to assess its role in ASFV replication. To further investigate the effect of small molecules in the metabolites from B. subtilis against ASFV replication, PAMs were pre-seeded in a 24-well plate (2 × 10⁵ cells per well), infected with 0.1 MOI ASFV for 1 h, washed, and cultured with fresh medium containing 10 µmol/L small molecules for 72 h. The supernatant was collected and analyzed using qPCR. The inhibition rate of each small molecule in the metabolites was calculated using the copy number of ASFV genomic DNA from PAMs treated with small molecules compared with that of the vehicle.

Detection of virus loading using qPCR

To calculate the copy number of ASFV genomic DNA from swabs, cell supernatants, tissue homogenate samples, and EDTA-treated whole peripheral blood samples, qPCR was performed as previously described (32). Briefly, ASFV genomic DNA was extracted using the FastPure Viral DNA/RNA Mini Kit (Vazyme, Nanjing, China). The reaction system for qPCR contained 0.2 pmol/µL sense primer (5´-CTGCTCATGGTATCAATCTTATCGA-3´), 0.2 pmol/µL anti-sense primer (5´-GATACCACAAGATCAGCCGT-3´), 0.2 pmol/µL probe (5´-FAM-CCACGGGAGGAATACCAACCCAGTG-3´-BHQ1), 10 µL Perfectstart II Probe qPCR Supermix UDG (TransGen Biotech, Beijing), 2 µL extracted template, and nuclease-free water up to 20 µL. The qPCR was performed on a QuantStudio 6 system (Applied Biosystems, Waltham, MA, USA). The samples were amplified using the following conditions: 50°C for 2 min, one cycle; 94°C for 5 min, one cycle; then 94°C for 5 s and 58°C for 30 s, 40 cycles.

Cytotoxicity assay

The cytotoxicity of metabolites of B. subtilis and small molecules for PAMs was evaluated using a cell counting kit-8 (CCK-8) (GlpBio, Montclair, CA, USA), according to the manufacturer’s instructions. Briefly, PAMs were pre-seeded in 96-well cell culture plates (10⁴ cells per well) for 24 h, followed by treatment with 10 µmol/L of different small molecules dissolved in dimethyl sulfoxide (DMSO) (MP Biomedicals LLC, Santa Ana, CA, USA) or 1 µg/mL of diverse metabolites of B. subtilis dissolved in RPMI 1640 medium for 48 h in triplicate. Then, 10 µL of the CCK-8 reagent was added to the cells and incubated in the dark for 2 h at 37°C. The optical density (OD) of each well was measured at 450 nm using a plate reader (Bio-Tek, Winusky, VT, USA). The viability of PAMs after treatment was calculated according to the CCK-8 protocol.

Half-maximal inhibitory concentration (IC₅₀) quantification

The PAMs were pre-seeded in 24-well plates (2 × 10⁵ cells per well). Eight-point dose-response curves were generated with concentrations of small molecules ranging from 312.5 to 40,000 nmol/L. PAMs were infected with 0.1 MOI ASFV for 1 h. The PAMs were then washed and cultured in fresh medium containing small molecules or DMSO for 72 h. The supernatants were collected to determine the viral titer using a hemadsorption (HAD) assay. The concentration of each small molecule required to inhibit viral infection by 50% (IC₅₀) was calculated by comparison with the viral titer from PAMs treated with DMSO using GraphPad Prism 6.0 software (GraphPad, San Diego, CA, USA).

Western blot analysis and an indirect immunofluorescence assay (IFA)

Cells were separated using SDS-PAGE and transferred to a nitrocellulose filter membrane (GE Medical, Waukesha, WI, USA) in transfer buffer (100 mM Tris, 190 mM glycine, 10% methanol). The membrane was blocked with 1% bovine serum albumin (BSA) for 1 h at 37°C. The membrane was incubated with ASFV p72 monoclonal antibodies preserved in our laboratory at 37°C for 2 h, followed by washing three times with TBS plus Tween 20 (TBST). Finally, the membrane was incubated with horseradish peroxidase (HRP)-conjugated AffiniPure Goat Anti-Mouse IgG (H + L) (Proteintech, Wuhan, China) secondary antibody at 37°C for 1 h and washed three times with TBST. The specific bands on the membrane were visualized using western blot ECL reagent (Advansta, San Jose, CA, USA).

PAMs were pre-seeded in 24-well plates (2 × 10⁵ cells per well) and infected with ASFV at an MOI of 0.1, followed by treatment with different metabolites from B. subtilis and arctiin for 72 h. Next, PAMs were fixed with 4% paraformaldehyde and permeabilized with 0.1% Triton X-100. After blocking with 1% BSA for 1 h at 37°C, PAMs were incubated with an ASFV p72 monoclonal antibody preserved in our laboratory at 37°C for 2 h. After washing three times with phosphate buffer solution (PBS) for 5 min each, the PAMs were incubated with CoraLite488-conjugated Goat Anti-mouse IgG(H + L) (Proteintech) at 37°C for 1 h. Samples were visualized using the EVOS FL Auto system (Thermo Fisher Scientific).

HAD assay

The HAD assay was performed to detect the titer of ASFV in the cell supernatant and tissue homogenate samples, as previously described (33). Briefly, the PAMs were pre-seeded in 96-well plates (10⁴ cells per well). The virus samples were 10-fold serially diluted and then pipetted into the wells of 96-well plates with eight replicates in each gradient. Next, 30 µL of 1% porcine red blood cells was added to each well after 3 days of cultivation, and rosette formation, representing hemadsorption, was observed around the infected cells. HAD was observed until 7 days post-infection (dpi), and the 50% HAD dose (HAD₅₀) was calculated using the Reed–Muench method (34).

Animal experiments

Large white piglets were used for animal experiments. Before ASFV challenge, the pigs were tested to ensure that they were negative for porcine circovirus (PCV), porcine reproductive and respiratory syndrome virus (PRRSV), pseudorabies virus (PRV), and classical swine fever virus (CSFV) infection. Pigs were fed with B. subtilis biologics of different forms and small molecules throughout the experimental period, including 10 days before infection and 28 days after infection. Four-week-old piglets were randomly divided into different groups during each experiment and placed in a negative pressure isolator at the animal biosafety level 3 laboratory of Huazhong Agricultural University. The different groups included an experimental group, a positive control group, and a negative control group. Bacillus subtilis was converted to powder form by spray-drying. The pigs in the experimental group were fed with liquid biologics of B. subtilis, B. subtilis powder mixed feed, granular feed containing B. subtilis, or small molecules. Pigs in the experimental groups and positive control groups were challenged with 500 HAD₅₀ ASFV by oral administration. Each pig was observed daily for signs of disease, including anorexia, depression, fever, purple skin discoloration, staggering gait, diarrhea, and cough, and body temperature changes were recorded throughout the experiment. Nasal, oral, and rectal swabs were collected every day after infection and viral shedding was detected. Blood samples were collected at 0, 7, 14, 21, and 28 dpi for viral detection. Serum samples were collected at 5 and 10 days before infection and 7 and 14 days after infection to detect cytokines. The pigs were dissected immediately after death. Surviving pigs were euthanized at 28 dpi. Tissue samples were observed for pathological changes and retained from each necropsied pig for further analysis, including histopathology and immunohistochemistry (IHC). The viral titers of the samples were determined using qPCR and HAD assays.

Compound docking

The ASFV topoisomerase II model was constructed using AlphaFold-Multimer. AlphaFold2 assesses the model using Predict TM-SCORE (PTM) and Internet PTM (IPTM) (35). The weighted values of PTM and IPTM were used as confidence values. The formula used was as follows:

$${\displaystyle \mathrm{C}\mathrm{o}\mathrm{n}\mathrm{f}\mathrm{i}\mathrm{d}\mathrm{e}\mathrm{n}\mathrm{c}\mathrm{e}=0.8\ast \mathrm{I}\mathrm{P}\mathrm{T}\mathrm{M}+0.2\ast \mathrm{P}\mathrm{T}\mathrm{M}}$$

The full Casp14 database was used for dimer protein modeling (36, 37). Twenty-five models were obtained, and the model with the highest confidence score was selected as the final structure (confidence = 0.9075). According to the existing type II topoisomerase structure (PDB: 4GFH), there are two Mg²⁺ ions in type II topoisomerase. Mg²⁺ ions were docked to the protein structure, the best pose was selected by the affinity scoring value and visual analysis, the protein and Mg²⁺ ions were merged using PyMOL, and the complexes were then relaxed using maestro (38). The SDF structure files of small molecules in the PubChem database were downloaded, Open Babel (3.1.1) was used to convert SDF to PDB format, and then standard docking processes were conducted in AutoDock Tools (1.5.7) (39 – 41). The specific parameters were as follows: 16 amino acids in the ATP combination domain were selected as the grid; the grid center positions were −22.0, 65.0, and 40.7 for x, y, and z, respectively; the sizes for the docking of compounds were 22, 30, and 32; and the number of GaRuns was 200. All results were analyzed using PyMOL and Discovery Studio Visualizer (42).

Protein expression and purification

To obtain the wild-type (WT) P1192R, CP204L, NP419L, C129R, and P1192R of the deleted ATP-binding domain (1–247 amino acids) (ΔATP-P1192R) proteins, cDNA-encoding residues were synthesized and codon-optimized for expression in Escherichia coli. The gene was added with a His₆ tag-encoding sequence and a Strep-tag II at its 5´- and 3´-termini, respectively, and then cloned into the Nco I and Xho I sites of the pET-28a vector (Invitrogen). These recombinant proteins were expressed in E. coli strain BL21 (DE3) as soluble proteins after induction with isopropyl-beta-D-thiogalactopyranoside (IPTG) (0.5 mM) at an OD₆₀₀ of 0.6–0.8 at 16°C for 12 h. The cells were harvested and lysed using high-pressure homogenization in lysis buffer (20 mM Tris and 500 mM NaCl, pH 8.0). After centrifugation, the supernatants were purified by metal affinity chromatography using StrepTrap XP 5 mL columns (Cytiva). Partially purified proteins were eluted with elution buffer (20 mM Tris, 500 mM NaCl, 50 mM biotin, pH 8.0) and further purified by gel filtration chromatography using a Superdex 200 Increase column (Cytiva) equilibrated with binding buffer (20 mM Tris, 500 mM NaCl, pH 8.0).

Biolayer interferometry (BLI) binding assay

The binding kinetics between the WT P1192R, CP204L, NP419L, and C129R proteins and arctiin or ΔATP-P1192R proteins and arctiin, genistein, and ATP were determined using the Octet Red 96 instrument (Pall ForteBio). The entire experiment was performed in plates incubated at 30°C with shaking at 1,000 rpm. All proteins used in the BLI assay were exchanged with BLI buffer (phosphate-buffered saline, 0.1% BSA, and 0.02% Tween). The streptavidin (SA) sensors were loaded with 50 µg/mL WT P1192R, CP204L, NP419L, C129R, or ΔATP-P1192R proteins for 180 s and then exposed to serially diluted small molecules. The data traces shown in the graphs were obtained after the sensors were exposed to small molecules using the data analysis software v8.1 (ForteBio). The processed data were plotted using GraphPad Prism 6.0.

Competitive decatenation assays

To investigate whether arctiin and genistein affect the catalytic activity of viral topoisomerase II, a topoisomerase II drug screening kit (kDNA-based) (TopoGEN, USA) was used according to the manufacturer’s instructions. Novobiocin was used as a positive control as it can block the ATP-binding domain (43). Briefly, the 20 µL system included 4 µL of 5× complete buffer, 150 ng of kDNA, 10 μΜ of arctiin, genistein, or 50 μΜ of novobiocin, 50 µg/mL of viral topoisomerase II, and ddH₂O up to the final volume. The solution was incubated from 30 min at 37°C and the reaction was stopped by the addition of 2 µL of 10% SDS. Proteinase K was added to 50 µg/mL and incubated at 37°C for 15 min. Loading buffer (0.1 vol) was added and samples were loaded directly onto a 1% agarose gel containing 0.5 µg/mL ethidium bromide in the gel and buffer. The gel was de-stained in water for 15 min and then the results were photo-documented using a gel imaging system (Bio-Rad).

Flow cytometry and biochemical detection

Porcine PBMCs were collected from EDTA-treated swine blood using a pig PBMC isolation kit (TBD Sciences, Tianjin, China). A total of 10⁶ PBMCs were stained for surface markers and fixed in accordance with the manufacturer’s instructions. Briefly, PBMCs were stained using antibodies, including fluorescein isothiocyanate (FITC) mouse anti-pig CD8a (Cat. No. 551303; BD Biosciences), phycoerythrin (PE)-Cy 7 mouse anti-pig CD3ε (Cat. No. 561477; BD Biosciences), and PE mouse anti-pig CD4a (Cat. No. 559586; BD Biosciences) at 37°C for 1 h. After three washes, PBMCs were fixed in 4% buffered formalin solution at room temperature for 10 min. The stained cells were washed three times and analyzed using a CytoFlex LX flow cytometer (Beckman Coulter, Indianapolis, IN, USA). The data were analyzed using FlowJo software (Tree Star, Ashland, OR, USA).

Serum samples from pigs fed arctiin and genistein by oral administration were collected and used to analyze the toxic effects of the drug on the liver, kidney, heart, and pancreatic islets. The biochemical indices of the serum samples were detected using an automatic biochemical analyzer according to the manufacturer’s instructions (Rayto, Shenzhen, China). The resulting data were plotted using GraphPad Prism 6.0.

Enzyme-linked immunosorbent assay

The levels of cytokines including IFN-β, IFN-γ, IFN-α, interleukin (IL)-4, IL-1β, IL-6, tumor necrosis factor (TNF)-α, and IL-8 in porcine serum samples and the cell supernatant were detected using the porcine IFN-β ELISA kit (Cat. No. SEKP-0046; Solarbio), porcine IFN-γ ELISA kit (Cat. No. SEKP-0010; Solarbio), porcine IFN-α ELISA kit (Cat. No. SEKP-0045; Solarbio), porcine IL-4 ELISA kit (Cat. No. SEKP-0003; Solarbio), porcine IL-1β ELISA kit (Cat. No. SEKP-0001; Solarbio), porcine IL-6 ELISA kit (Cat. No. SEKP-0004; Solarbio), porcine TNF-α ELISA kit (Cat. No. SEKP-0009; Solarbio), and porcine IL-8 ELISA kit (Cat. No. SEKP-0005; Solarbio) according to the manufacturer’s instructions. Briefly, the microplate was washed three times before the assay using phosphate-buffered solution plus Tween 20 (PBST). Diluted standards (100 µL) or serum samples were added to the plates and incubated at 37°C for 1 h. After four washes with PBST, 100 µL of working solution including biotin-conjugated anti-porcine IFN-β, IFN-γ, IFN-α, IL-4, IL-1β, IL-6, TNF-α, or IL-8 antibody was added to each well at 37°C for 1 h. The microplate was washed four times with PBST, and 100 µL of streptavidin-HRP working solution was added to each well and incubated at 37°C for 20 min. After five washes with PBST, 100 µL of substrate solution was added to the plate and incubated at 37°C for 20 min, followed by the addition of 50 µL of stop solution to each well. The absorbance was measured at 450 nm. The cytokine levels were calculated using the corresponding standard curve.

Histopathology and IHC

For histopathological observations, hematoxylin-eosin staining was performed according to the conventional procedure. Tissue samples were collected from large white piglets infected with ASFV (positive control group), large white piglets infected with ASFV along with intervention with different biologics (experimental group), and large white piglets infected without ASFV (negative control group), and were fixed in 4% buffered formalin solution, embedded in paraffin, and sectioned at 4 µm thickness. Next, the sections were washed in dimethylbenzene for 20 min twice, absolute ethanol for 5 min twice, 75% absolute ethanol for 5 min, then sterile water for 1 min. The sections were then stained with hematoxylin for 5 min, followed by washing with sterile water for 1 min. Finally, the sections were dehydrated using 85%–95% ethanol absolute for 5 min, stained with eosin for 5 min, and washed with sterile water for 1 min. Finally, sections were dehydrated using absolute ethanol for 10 min and sealed with neutral balsam.

IHC was performed as described previously (44). Briefly, tissue samples from large white piglets infected with ASFV (positive control group), large white piglets infected with ASFV along with intervention with different biologics (experimental group), and large white piglets infected without ASFV (negative control group) were fixed in 4% buffered formalin solution, embedded in paraffin, and sectioned at 4 µm thickness. After hydration, the tissue sections were incubated in trypsin solution (0.1%) in the presence of calcium chloride dihydrate (3 M) for 20 min at 37°C. Slides were incubated with mouse monoclonal antibodies (1:1,000) against ASFV p72 protein preserved in our laboratory at 37°C for 2 h, followed by washing with TBST three times for 5 min each. The sections were further incubated with HRP-conjugated goat anti-mouse IgG (H + L) (Cat. No. ab6789; Abcam) followed by washing with TBST three times for 5 min each. diaminobenzidine (DAB) chromogenic solution was used to dye the sections for 5 min, followed by washing with TBST three times for 5 min each.

Time-course inhibition test

PAMs were pre-seeded in 24-well plates and infected with 0.1 MOI ASFV for 1 h at 37°C to estimate the effect of arctiin on the ASFV replication cycle. Arctiin was added 1 h before (pre), simultaneously with (co), or 1 h after (post) ASFV infection. For the pre-treatment experiment, PAMs were incubated with arctiin for 1 h at 37°C, washed three times with sterile PBS, and then infected with ASFV for 1 h. For the co-treatment assay, PAMs were simultaneously incubated with ASFV and arctiin. After 1 h, the mixture was removed, and the cells were washed three times with sterile PBS before fresh medium was added. For the post-treatment assay, cells were first infected with ASFV for 1 h, washed three times with sterile PBS, and then incubated with medium containing arctiin for 1 h. The cells were then washed three times with sterile PBS and cultured in fresh medium for 72 h. The cell supernatant was collected, and the amount of virus was detected using qPCR and HAD assays.

Statistical analysis

Statistical significance was analyzed using an unpaired Student’s t-test and was set at P < 0.05. Statistical analysis was performed using GraphPad Prism software (version 6.0). The specific details of the statistical tests are described in the figure legends.

RESULTS

Screening of B. subtilis metabolites against ASFV infection in vitro

When the working solution of the metabolites was 1 µg/mL, no cytotoxicity against PAMs was observed (Fig. S1A). A total of 138 B. subtilis metabolites were detected, and they were used to analyze anti-ASFV activity (Fig. 1A and B). ASFV replication was significantly inhibited by the metabolites of four B. subtilis strains, with an inhibition rate greater than 90%. These strains were designated ZF-1, GLSZ-1, GLSZ-2, and LD-2-1. B. subtilis LD-2-1 had the highest inhibition rate at 96.37%. ASFV genomic DNA was detected to further estimate the anti-ASFV effect of the four B. subtilis strains in PBMCs and BMDMs. Their metabolites markedly inhibited ASFV replication, and the virions of ASFV from the B. subtilis LD-2-1-treated group showed the highest inhibitory effect in PBMCs (10^5.44 copies/mL vs 10^1.61 copies/mL) and BMDMs (10^6.12 copies/mL vs 10^1.69 copies/mL), compared with the vehicle group (Fig. S1C and F). Western blot analysis was used to explore the function of these metabolites in suppressing ASFV proliferation. The metabolites of the B. subtilis LD-2-1 strain showed evident inhibition of proliferation (Fig. 1Ci), and we detected the late protein p72 of ASFV, whose band intensity decreased by 78% compared with vehicle treatment (Fig. 1Cii). Similar results were confirmed in PBMCs and BMDMs, which showed band intensity decreases of 52% and 55%, respectively, compared with vehicle treatment (Fig. S1D and G). The viral titers in PAMs treated with metabolites of the B. subtilis LD-2-1, GLSZ-1, GLSZ-2, and ZF-1 strains were reduced by 10^3.36-, 10^2.55-, 10^3.05-, and 10^3.01-fold, respectively, compared with those treated with the vehicle (Fig. 1D). The decrease in the viral titers in PBMCs and BMDMs after treatment with the metabolites of four B. subtilis strains was also confirmed using the HAD assay. The viral titers from the B. subtilis LD-2-1-treated group with the highest inhibition efficiency compared with the vehicle-treated group were reduced 10^4.09- and 10^4.59-fold in PBMCs and BMDMs, respectively (Fig. S1B and E). The effect of these metabolites on ASFV replication was verified using an IFA. We confirmed that LD-2-1 had the highest inhibition efficiency among the four B. subtilis strains, and the mean fluorescence intensity was reduced 4.68-fold compared with that of the vehicle-treated control (Fig. 1E), whereas the mean fluorescence intensity of PAMs treated with metabolites from the other three B. subtilis strains was reduced by 2.39-, 3.32-, and 3.21-fold, respectively. These results indicate that the metabolites of the four screened B. subtilis strains efficiently inhibited ASFV replication in vitro.

Fig 1.

Metabolites derived from Bacillus subtilis against African swine fever virus evaluated in PAMs. (A) Schema of the experimental design for the screening in PAMs. (B) The role of metabolites of B. subtilis in ASFV infection. The dot represents the mean, the black dotted line represents the 90% inhibition ratio, the red dots represent the metabolites of the four B. subtilis strains with the highest inhibition efficiency against ASFV proliferation. (C) The function of the metabolites of the four B. subtilis strains assessed using western blot analysis. (i) ASFV p72 and porcine glyceraldehyde-3-phosphate dehydrogenase (GAPDH) proteins were detected. (ii) The band intensity of ASFV p72 proteins. (D) The metabolites from the four B. subtilis strains showing the highest inhibition efficiency against ASFV proliferation were estimated using a hemadsorption assay. (E) The metabolites from the four B. subtilis strains showing the highest viral inhibition efficiency were estimated by immunofluorescence staining. ASFV p72 proteins were stained with CoraLite488-conjugated Goat Anti-mouse IgG (green); nuclei were stained with 2-(4-amidinophenyl)-6-indolecarbamidine dihydrochloride (blue). (i) Stained cells were examined using the EVOS FL Auto system. (ii) Mean fluorescence intensity in PAMs treated with different media. AU, arbitrary units. Data were analyzed using a two-tailed Student’s t-test (**P < 0.01, ***P < 0.001, ****P < 0.0001).

Bacillus subtilis metabolites slowed down ASFV infection in pigs

A schematic of the experimental workflow is shown in Fig. 2A. Biologics of the four B. subtilis strains were orally administered to pigs for 38 days, including 10 days before and 28 days after infection. The pigs were challenged with 500 HAD₅₀ ASFV by oral administration on day 0. The body temperature of pigs orally administered with biologics did not exceed 40°C after ASFV infection; however, that of pigs orally administered with the vehicle gradually increased over time after ASFV infection until death, with the highest body temperature recorded at 41.78°C (Fig. 2B). All pigs orally administered with the vehicle died within 14 dpi, whereas those orally administered with biologics survived after ASFV infection (Fig. 2C). Nasal, oral, and rectal swab samples were collected and detected every day after infection. The amount of ASFV genomic DNA from pigs orally administered with the vehicle increased over time. The highest viral load was detected from nasal, oral, and rectal swab samples, reaching 5.49 × 10⁶ copies/mL at 12 dpi, 4.07 × 10⁶ copies/mL at 14 dpi, and 5.12 × 10⁶ copies/mL at 13 dpi, respectively (Fig. 2D through F). The amount of ASFV genomic DNA from pigs orally administered with biologics varied, and the viral load in rectal swab samples of pigs was not detected from 26 dpi to the end of the observation period. The highest viral load was observed in the nasal, oral, and rectal swab samples, reaching 2.39 × 10³ copies/mL at 13 dpi, 1.62 × 10³ copies/mL at 13 dpi, and 8.31 × 10² copies/mL at 14 dpi, respectively (Fig. 2D through F). The amount of ASFV genomic DNA in the blood samples of pigs orally administered with vehicle gradually increased with prolonged infection time, and the highest viral load reached 5.24 × 10⁸ copies/mL at 14 dpi. However, the amount of ASFV genomic DNA in the blood samples of pigs orally administered with biologics showed an increase at 0–14 dpi and a decrease at 14–21 dpi. The highest viral load reached 1.81 × 10⁴ copies/mL at 14 dpi (Fig. 2G). Although the amount of ASFV genomic DNA, which is indicative of viral shedding, was almost undetectable in the swab samples at 28 dpi, viral titers and ASFV genomic DNA were detected in the organs, including the liver, heart, lung, spleen, kidney, mesenteric lymph nodes, submaxillary lymph nodes, and inguinal lymph nodes from pigs orally administered with biologics (Fig. 2H and I). In the spleen of pigs orally administered with the vehicle, the number of lymphocytes decreased sharply with necrosis; the glomeruli of the kidney were clearly atrophic and necrotic; and the renal tubular epithelial cells were edematous, exfoliative, and necrotic. However, in ASFV-infected pigs orally administered with biologics and ASFV-uninfected pigs, pathological changes in the spleen, kidney, and inguinal lymph nodes were not observed (Fig. 2J). An IHC assay clearly showed ASFV virions in the spleen, kidney, and inguinal lymph nodes when comparing ASFV-infected pigs orally administered with the vehicle to ASFV-infected pigs orally administered biologics (Fig. 2K).

Fig 2.

Oral administration of four B. subtilis strains with the highest inhibition efficiency prevents ASFV occurrence in pigs. (A) Schema of the experimental design for assessing the activity of B. subtilis strains against ASFV infection in pigs. The biologics of four B. subtilis strains and fermentation broth were orally administered throughout the experimental cycle for 38 days, including 10 days before and 28 days after infection. The biologics consisted of 10⁹ CFUs of the four B. subtilis strains (2.5 × 10⁸ CFU each) and 8 mL of metabolites from the four B. subtilis isolates (2 mL each) for each pig daily. The pigs were challenged with 500 HAD₅₀ ASFV by oral administration on day 0. Daily observations and swab collection were conducted after ASFV infection. (B) Daily body temperature of pigs after ASFV infection. (C) Survival rate of pigs after ASFV infection. Virus shedding after ASFV infection. Daily viral genomic DNA copies were determined from nasal (D), oral (E), and rectal (F) swabs of pigs from different groups after infection. (G) Viral genomic DNA copies in the blood of pigs at 0, 7, 14, 21, and 28 dpi. (H) Viral genomic DNA copies in the heart, liver, spleen, lung, kidney, SLN, ILN, and MLN of pigs at death or up to 28 dpi. (I) Viral titer in the heart, liver, spleen, lung, kidney, SLN, ILN, and MLN of pigs at death or up to 28 dpi. (J) Hematoxylin and eosin staining assay showing subtle pathological changes in the spleen, kidney, and ILN of pigs. Red arrows indicate lesions in tissue samples. (K) Immunohistochemistry using antibodies to ASFV p72 proteins. SLN, submaxillary lymph node; ILN, inguinal lymph node; MLN, mesenteric lymph node. Data were analyzed using a two-tailed Student’s t-test (****P < 0.0001).

The expression levels of IL-4, IL-1β, IL-8, IFN-β, IFN-γ, and IFN-α, and CD4⁺ and CD8⁺ T lymphocytes were detected at 0 and 5 days before and 7 and 14 days after infection between pigs orally administered with biologics and those with vehicle. The expression levels of cytokines did not differ between the two groups before ASFV infection (Fig. 3A through G). However, the expression levels of IL-4, IL-1β, IL-8, IFN-β, IFN-γ, and IFN-α in pigs after oral administration of biologics were lower than those in pigs after oral administration of vehicle at 14 dpi, which were reduced 2.85-, 2.37-, 3.98-, 2.83-, 2.35-, and 2.59-fold, respectively; the expression of CD4⁺ and CD8⁺ T lymphocytes did not differ (Fig. 3H through N). These results indicated that metabolites of B. subtilis do not regulate the expression of CD4⁺ and CD8⁺ T lymphocytes, IL-4, IL-1β, IL-8, IFN-β, IFN-γ, and IFN-α to exert anti-ASFV effects.

Fig 3.

Cellular and humoral immunity before and after ASFV infection. (A) Ratio of porcine positive CD4⁺ and CD8⁺ T lymphocytes at 5 and 10 days after oral administration with B. subtilis or vehicle in pigs without ASFV infection. Porcine IL-4 (B), IL-1β (C), IL-8 (D), IFN-β (E), IFN-γ (F), and IFN-α (G) levels. (H) Ratio of porcine positive CD4⁺ and CD8⁺ T lymphocytes at 7 and 14 days after oral administration with B. subtilis or vehicle in pigs infected with ASFV. Porcine IL-4 (I), IL-1β (J), IL-8 (K), IFN-β (L), IFN-γ (M), and IFN-α (N) levels. Data were analyzed using a two-tailed Student’s t-test (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001).

We further evaluated the inhibitory function of powders derived from the four B. subtilis strains mixed with pig feed or prepared mixed pellet feed. A schematic of the experimental workflow is shown in Fig. 4A. Pigs were fed with powders derived from four B. subtilis strains mixed with feed for 38 days, including 10 days before infection and 28 days after infection. The pigs were challenged with 500 HAD₅₀ ASFV via oral administration on day 0. The body temperature of pigs fed the B. subtilis powders did not exceed 40°C during the entire infection period, and no deaths were reported (Fig. 4B and C). Virus shedding in the nasal, oral, and rectal swab samples from pigs fed with the powders of four B. subtilis strains mixed with feed was lower than that in the positive control group, and ASFV genomic DNA was not detected in rectal swabs from 26 dpi to the end of the observation period in those fed powders mixed with feed (Fig. 4D through F). The changes in the amount of viral genomic DNA in the blood and organs and the pathological changes in the organs of pigs fed with powders mixed with feed were highly coincident with those of pigs fed with biologics by oral administration (Fig. 4G through K). A schematic of the experimental workflow is shown in Fig. S2A. Pigs were fed with powders derived from the strains in pellet feed for 38 days, including 10 days before and 28 days after infection. After 10 days of feeding, pigs were challenged with 500 HAD₅₀ ASFV. The body temperature of pigs administered the powder in pellet feed did not exceed 40°C and no deaths were recorded after ASFV infection (Fig. S2B and C). The changes in the amount of viral genomic DNA from swabs, blood, and organs and the pathological changes in the organs of pigs fed with preparative pellet feed powder were highly consistent with those of pigs fed biologics by oral administration (Fig. S2D through K). These results indicate that the four B. subtilis strains used in combination slowed down ASFV infection in vivo.

Fig 4.

Administration of B. subtilis mixed with pig feed reduces ASFV infection and prevents disease occurrence in pigs. (A) Schema of the experimental design. Powders of B. subtilis contain 2 × 10⁹ CFU from the four B. subtilis strains (5 × 10⁸ CFU each) administered to each pig daily. Mixed pig feed was administered throughout the experimental cycle for 38 days, including 10 days before infection and 28 days after infection. The pigs were challenged with 500 HAD₅₀ ASFV via oral administration on day 0. (B) Daily body temperature of pigs after ASFV infection. (C) Survival rate of pigs after ASFV infection. Virus shedding was detected after ASFV infection. Daily viral genomic copies from nasal (D), oral (E), and rectal (F) swabs of pigs after infection. (G) Viral genomic copies in the blood samples of pigs at 0, 7, 14, 21, and 28 dpi. (H) Viral genomic copies in the heart, liver, spleen, lung, kidney, SLN, ILN, and MLN of pigs at death or up to 28 dpi. (I) Viral titer in the heart, liver, spleen, lung, kidney, SLN, ILN, and MLN of pigs at death or up to 28 dpi. (J) Hematoxylin and eosin staining of the tissue samples of the spleen, kidney, and ILN of pigs. Red arrows indicate lesions in tissue samples. (K) Immunohistochemistry of the antibodies of ASFV p72 proteins. SLN, submaxillary lymph node; ILN, inguinal lymph node; MLN, mesenteric lymph node. Data were analyzed using a two-tailed Student’s t-test (****P < 0.0001).

Screening of small-molecule metabolites against ASFV infection

To further explore the main components of the metabolites of the four B. subtilis against ASFV infection, the metabolites were treated with methanol to eliminate the effect of proteins (Fig. 5A). The 1 µg/mL treated and untreated metabolites of B. subtilis significantly inhibited ASFV infection compared with the positive control group (Fig. 5B), indicating that the antiviral property is attributed to small molecules rather than proteins. To identify these small molecules, the metabolites of four B. subtilis strains (GLSZ-1, GLSZ-2, LD-2-1, and ZF-1) with the highest inhibition efficiency and six B. subtilis strains (BE-13, BLL-1, EMB, JSY-1, XYY-4, and XHYB-2) without inhibition efficiency were collected for LC-MS/MS analysis. The metabolomics data were filtered according to the following thresholds—P value <0.05, VIP >1, and |log₂ (fold change)| >2—in the positive and negative ion modes, respectively (GLSZ-1, GLSZ-2, LD-2-1, or ZF-1 vs BE-13, BLL-1, EMB, JSY-1, XYY-4, or XHYB-2) (Fig. 6A through H). Under these criteria, a total of 652 and 615 differentially expressed small molecules in the positive and negative ion modes were detected, respectively (Fig. 6A through H). However, there were differentially expressed small molecules that were detected in both the positive ion mode (Fig. 6A through D) and the negative ion mode (Fig. 6E through H). After integration, a total of 250 unique differential small molecules were identified including 143 and 107 differentially expressed small molecules in the positive and negative ion modes, respectively (Fig. 6I), 169 of which are commercially available and were used to evaluate anti-ASFV proliferation activity (Fig. 5C). Detailed information on the 169 metabolites is shown in Table S1.

Fig 5.

Screening of small-molecule metabolites from B. subtilis for their activity against ASFV replication. (A) Schema of the experimental design of protein precipitation of the fermentation broth of four B. subtilis strains. (B) Viral genomic copies were detected for the assessment of the anti-ASFV activity of untreated and treated fermentation broth from four B. subtilis strains. (C) Schematic diagram of the study design for screening small-molecule metabolites. (D) The activity of small-molecule metabolites of B. subtilis against ASFV infection. The dot represents the mean, the black dotted line represents the 90% inhibition ratio, and the red dots represent the eight small molecules with the highest inhibition efficiency against ASFV proliferation. (E) The function of the eight small molecules against ASFV replication was assessed using western blot analysis. (i) ASFV p72 and porcine GAPDH proteins were detected. (ii) The band intensity of ASFV p72 proteins. (F) Viral titer was detected for evaluating the function of eight different small molecules against ASFV infection. (G) Measurement of the half-maximal inhibitory concentration of the eight small molecules. The black dotted line represents the 50% inhibition ratio. (H) Biological characterization of the eight small-molecule metabolites from B. subtilis strains. Data were analyzed using a two-tailed Student’s t-test (**P < 0.01, ***P < 0.001, ****P < 0.0001).

Fig 6.

Screening of differential small-molecule metabolites of B. subtilis. Identification of small molecules in positive ion mode by comparison of B. subtilis GLSZ-1 strains (A), GLSZ-2 strains (B), LD-2-1 strains (C), ZF-1 strains (D), and six B. subtilis strains without inhibition efficiency (P < 0.05, VIP >1, and |log₂ (fold change)| >2). Identification of small molecules in negative ion mode by comparison of B. subtilis GLSZ-1 strains (E), GLSZ-2 strains (F), LD-2-1 strains (G), ZF-1 strains (H), and six B. subtilis strains without inhibition efficiency (P < 0.05, VIP >1, and |log₂ (fold change)| >2). (I) All differential small molecules in positive and negative ion modes after integration.

The cytotoxicity of 10 µM of available small molecules was assessed using the CCK-8 assay, and this concentration did not induce cytotoxicity in PAMs (Fig. S3A). The eight small molecules reduced the amount of viral DNA, with an inhibition rate higher than 90% (Fig. 5D). We found that the amount of ASFV genomic DNA in the group treated with the eight small molecules was significantly decreased compared with that in the positive control groups in PBMCs and BMDMs (Fig. S3B and E). Compared with the vehicle group, the ASFV copies from the arctiin-treated group showed the highest inhibitory effect in PBMCs (10^5.84 copies/mL vs 10^1.51 copies/mL) and BMDMs (10^6.39 copies/mL vs 10^1.63 copies/mL). ASFV p72 proteins were detected, and the intensity of the p72 protein band for PAMs treated with arctiin, genistein, deoxycholic acid, stearic acid, catechol, Fmoc-L-isoleucine, soyasaponin Bb, and piceatannol was decreased by 72.4%, 63.8%, 42.2%, 41.4%, 45.6%, 46.7%, 53.4%, and 52.6%, respectively, compared with the positive control (Fig. 5E). The results of western blotting following treatment with small molecules in PBMCs and BMDMs were significantly different from those of the positive control group (Fig. S3D and G). The HAD assay showed that the viral titer of PAMs treated with arctiin, genistein, deoxycholic acid, stearic acid, catechol, Fmoc-L-isoleucine, soyasaponin Bb, and piceatannol was reduced by 75.3%, 72.5%, 60.5%, 52.1%, 60.7%, 61.2%, 62.7%, and 61.5%, respectively, compared with the positive control (Fig. 5F). Moreover, their viral titers were also significantly reduced compared with the positive control in PBMCs and BMDMs (Fig. S3C and F). Compared with the vehicle-treated group, the viral titers from the arctiin-treated group with the highest inhibition efficiency were reduced 10^4.26- and 10^4.83-fold in PBMCs and BMDMs, respectively. The 50% maximal inhibitory concentration of the eight small molecules ranged from 4.139 μM to 8.825 µM in PAMs (Fig. 5G). The characteristics of the eight small molecules are shown in Fig. 5H. Altogether, these results indicated that the eight small-molecule metabolites inhibited ASFV proliferation, with arctiin showing the highest inhibition efficiency.

Arctiin inhibited ASFV replication during the mid-stage of infection

Because arctiin had the highest inhibitory effect on viral proliferation among the eight molecules, arctiin was subjected to further analysis. According to a previous study, arctiin exerts an antiviral effect by inhibiting the inflammatory response (45). However, arctiin did not inhibit ASFV proliferation by suppressing the inflammatory response in our research (Fig. S4). To explore the timing of the antiviral mechanism of arctiin during infection, PAMs were treated with arctiin 1 h before (pre), simultaneously with (co), or 1 h after (post) ASFV inoculation (Fig. 7A). Arctiin showed no effect following pre-treatment or co-treatment, whereas post-treatment led to a dramatic reduction in the amount of viral genomic DNA compared with the vehicle group (10^5.02 copies/mL vs 10^2.12 copies/mL), and viral titers reduced 10^3.37-fold (Fig. 7B and C). Similarly, ASFV genomic DNA and viral titers were detected in PBMCs and BMDMs. Therefore, arctiin exerted a role against ASFV infection post-treatment (Fig. S5A, B, D, and E). Post-treatment, arctiin caused a significant reduction in the amount of viral p72 protein of 45% in PAMs, 51% in PBMCs, and 46% in BMDMs (Fig. 7D; Fig. S5C and F). This indicates that arctiin may exert its maximal antiviral effect after ASFV infection. Arctiin was also added to PAMs, PBMCs, and BMDMs at 2, 4, 8, 12, or 16 h post-infection (hpi). Interestingly, the amounts of ASFV genomic DNA were markedly decreased upon arctiin treatment at 8 hpi compared with the vehicle group in PAMs (10^5.18 copies/mL vs 10^1.39 copies/mL), PBMCs (10^5.89 copies/mL vs 10^1.71 copies/mL), and BMDMs (10^6.38 copies/mL vs 10^1.81 copies/mL), and the viral titers were reduced 10^4.01-, 10^4.45-, and 10^4.59-fold (Fig. 7E and G; Fig. S5G, H, J, and K). Compared with the vehicle group, the level of viral p72 proteins was evidently reduced in PAMs by 61%, in PBMCs by 41%, and in BMDMs by 59% following treatment with arctiin at 8 hpi (Fig. 7F; Fig. S5I and L). The timing of arctiin addition after infection that inhibited ASFV replication was verified by an IFA. Consistent with our previous findings, PAMs treated with arctiin at 8 hpi showed the highest inhibition efficiency, with a reduction in mean fluorescence intensity of 9.38-fold (Fig. 7H). Overall, these results indicate that arctiin may exert its maximal antiviral effect after viral infection by targeting viral proteins that are expressed in the mid-stage of ASFV replication, rather than by regulating the host’s antiviral processes.

Fig 7.

Arctiin inhibits ASFV replication at the mid-stage of infection. (A) Schematic diagram of the study design. (B) Viral genomic copies at different stages of infection after pre-treatment, co-treatment, and post-treatment with arctiin. (C) Viral titers were measured using the hemadsorption assay. (D) Western blot analysis. (i) ASFV p72 and porcine GAPDH proteins were detected. (ii) The band intensity of ASFV p72 protein. (E) Viral genomic copies were detected at the time of addition of arctiin after infection. (F) Western blot analysis of proteins at the time of addition of arctiin after infection. (i) ASFV p72 and porcine GAPDH proteins were detected. (ii) The band intensity of ASFV p72 protein. (G) Viral titers were detected at the time of addition of arctiin after infection. (H) Antiviral activities at the time of addition of arctiin after infection were analyzed by immunofluorescence staining. The ASFV p72 proteins were stained with CoraLite488-conjugated Goat Anti-mouse IgG (green); nuclei were stained with 2-(4-amidinophenyl)-6-indolecarbamidine dihydrochloride (blue). (i) Stained cells were examined using the EVOS FL Auto system. (ii) Mean fluorescence intensity in PAMs. AU, arbitrary units. Data were analyzed using a two-tailed Student’s t-test (**P < 0.01, ***P < 0.001, ****P < 0.0001).

Arctiin targets ASFV type II topoisomerase

There was no significant difference in cellular and humoral immunity, including the expression of CD4⁺ and CD8⁺ T lymphocytes, and the levels of IL-4, IL-1β, IL-8, IFN-β, IFN-γ, and IFN-α, between pigs orally administered with and without B. subtilis in the in vivo experiment (Fig. 3). Meanwhile, arctiin, a metabolite from B. subtilis, showed the strongest antiviral effect among small molecules and this effect was most evident at 8 hpi. Therefore, the antiviral mechanism of B. subtilis against ASFV infection is probably due to small-molecule metabolites targeting viral proteins in the middle stages of infection process.

Native-PAGE followed by silver staining was used to confirm the interaction between lysed proteins from PAMs infected with ASFV at 8 hpi and arctiin, and demonstrated that three protein bands exhibited a mobility shift due to arctiin binding (Fig. 8A). The three differential protein bands were further analyzed by LC-MS/MS, and four ASFV proteins, namely P1192R, CP204L, NP419L, and C129R, were identified. The characteristics of these four viral proteins are described in Fig. 8A. The BLI assay was performed to validate the interaction between arctiin and P1192R, CP204L, NP419L, and C129R, revealing that arctiin at 200 µM interacts with P1192R rather than with CP204L, NP419L, and C129R (Fig. 8B). Interestingly, genistein also showed inhibition rates higher than 90% in our research; in previous research, it has been proven to target ASFV P1192R encoding the type II topoisomerase (14). Type II topoisomerase contains an ATP-binding site and requires ATP to function (15). Therefore, we further used the BLI assay to validate the interaction between P1192R and arctiin, genistein, and ATP, which showed that the lowest K _D value (11.48 µM) was obtained between arctiin and viral type II topoisomerase, whereas the binding affinities of genistein and ATP to viral type II topoisomerase were 16.82 µM and 49.37 µM, respectively (Fig. 8C through E). Molecular docking analysis showed that the binding affinity of arctiin to the ATP-binding site of type II topoisomerase (−10.15 kcal/moL) was higher than that of genistein (−8.33 kcal/moL), which was higher than that of ATP (−6.98 kcal/moL) (Fig. 8F through H). The BLI assay results were consistent with those of molecular docking, demonstrating that the interaction between arctiin and viral type II topoisomerase was the strongest. We investigated whether arctiin and genistein bind to the ATP-binding site of viral type II topoisomerase and interfere with its function, and found that P1192R proteins lacking the ATP-binding domain (1–247 amino acids; ΔATP-P1192R) did not interact with arctiin, genistein, or ATP (Fig. 8I). As shown in Fig. 8J, under normal circumstances, ASFV topoisomerase II enters a breakage/resealing cycle that favors the resealed products including nicked circular and relaxed circular DNA. However, when novobiocin (positive control), arctiin, or genistein was added, which disturbs ATP binding to ASFV topoisomerase II, this led to a reduction in resealed products. Arctiin showed the best efficiency in inhibiting enzyme activity, followed by genistein (Fig. 8Jii). Therefore, we concluded that arctiin competed for the ATP-binding site of viral type II topoisomerase, disturbing the function of viral type II topoisomerase and suppressing ASFV replication.

Fig 8.

Arctiin inhibits ASFV infection by targeting the ATP-binding domain of viral topoisomerase II. (A) Native-PAGE assay to explore the interaction between arctiin and viral proteins. (i) Native-PAGE followed by silver staining to analyze protein bands showing a mobility shift between the lysed proteins from PAMs infected with ASFV at 8 hpi and arctiin. (ii) The results obtained for arctiin combined with four viral proteins by LC-MS/MS analysis. (B) Binding dynamics of the four viral proteins, P1192R, CP204L, NP419L, and C129R, to arctiin as assessed by a BLI binding assay. (C) Binding dynamics between arctiin and viral topoisomerase II. (D) Binding dynamics between ATP and viral topoisomerase II. (E) Binding dynamics between genistein and viral topoisomerase II. (F) Arctiin combined with the ATP-binding site of viral topoisomerase II, as assessed using molecular docking analysis. (G) ATP combined with the ATP-binding site of viral topoisomerase II, as assessed using molecular docking. (H) Genistein combined with the ATP-binding site of viral topoisomerase II, as assessed using molecular docking. (I) Binding dynamics of ATP, arctiin, and genistein with deletion of the ATP-binding domain (1–247 amino acids) of viral topoisomerase II, as assessed using a BLI binding assay. (J) Competitive decatenation assays performed to validate the effect of arctiin and genistein on the catalytic activity of viral topoisomerase II. (i) The decatenation products are detected. (ii) The band intensity of decatenation products. Novobiocin was used as a positive control.

Arctiin and genistein slowed down ASFV infection in vivo

The role of arctiin in ASFV was explored further in vivo. The inhibition of ASFV by genistein was assessed. A schematic of the experimental workflow is shown in Fig. 9A. Compared with the pigs in the control group, when pigs were fed with 2 mg/kg arctiin, 2 mg/kg genistein, or 2 mg/kg arctiin and 2 mg/kg genistein by oral administration, no toxic effects were observed (Fig. S6). The body temperature of pigs orally administered with genistein exceeded 40°C at 11 dpi, that of pigs orally administered with arctiin exceeded 40°C at 8 dpi, that of pigs orally administered with arctiin and genistein exceeded 40°C at 15 dpi, and that of pigs orally administered with the vehicle exceeded 40°C at 5 dpi (Fig. 9B). All pigs treated with the vehicle died within 17 dpi, whereas five pigs orally administered with arctiin died at 10–22 dpi, five pigs orally administered with genistein died at 15–21 dpi, and three pigs orally administered with both arctiin and genistein died at 21–24 dpi, and the corresponding survival rates reached 50%, 50%, and 70%, respectively (Fig. 9C). The amount of ASFV genomic DNA from pigs orally administered with the vehicle showed an upward trend over time. The highest viral loads from nasal, oral, and rectal swab samples reached 2.34 × 10⁶ copies/mL at 16 dpi, 1.51 × 10⁶ copies/mL at 16 dpi, and 1.73 × 10⁶ copies/mL at 12 dpi (Fig. 9D through F). The viral loads from nasal, oral, and rectal swab samples of surviving pigs fed with arctiin, genistein, and arctiin and genistein were significantly lower than those of the positive controls (Fig. 9D through F). The viral load in blood samples of pigs fed orally with vehicle gradually increased with a prolonged infection time, and the highest viral load reached 2.45 × 10⁷ copies/mL at 14 dpi (Fig. 9G). The viral load in blood samples from surviving pigs fed orally with genistein, arctiin, or arctiin and genistein reached high levels of 3.09 × 10⁴ copies/mL, 9.54 × 10⁴ copies/mL, and 1.91 × 10⁴ copies/mL, respectively (Fig. 9G). Several pigs fed orally with drugs survived to 28 dpi; however, ASFV genomic DNA and viral titers were detected in various organs, including the liver, heart, lung, spleen, kidney, mesenteric lymph nodes, submaxillary lymph nodes, and inguinal lymph nodes (Fig. 9H and I).

Fig 9.

Arctiin and genistein inhibit ASFV infection and reduce mortality in pigs. (A) Schematic diagram of the study design. Arctiin (2 mg/kg), genistein (2 mg/kg), and arctiin (2 mg/kg) and genistein (2 mg/kg) were orally administered for 38 days, from 10 days before infection to 28 days after infection for each pig, daily. The pigs were challenged with 500 HAD₅₀ ASFV by oral administration on day 0, and daily observations and swab collection were conducted after ASFV infection. (B) Daily body temperature of pigs after ASFV infection. (C) Survival rate of pigs after ASFV infection. Virus shedding after ASFV infection. Daily viral genomic DNA copies from the nasal (D), oral (E), and rectal (F) swabs of pigs from different groups after infection. (G) Viral genomic DNA copies in the blood of pigs at 0, 7, 14, 21, and 28 dpi. (H) Viral genomic DNA copies in the heart, liver, spleen, lung, kidney, SLN, ILN, and MLN of pigs at death or up to 28 dpi. (I) Viral titer in the heart, liver, spleen, lung, kidney, SLN, ILN, and MLN of pigs at death or up to 28 dpi. (J) Hematoxylin and eosin staining of the tissue samples of the spleen, kidney, and ILN of pigs. Red arrows indicate lesions in tissue samples. (K) Immunohistochemistry of the antibodies of ASFV p72 proteins. SLN, submaxillary lymph node; ILN, inguinal lymph node; MLN, mesenteric lymph node. Data were analyzed using a two-tailed Student’s t-test (*P < 0.05, **P < 0.01, ***P < 0.001).

We next analyzed the pathological changes in the organs of ASFV-infected pigs orally administered with drugs or vehicle and ASFV-uninfected pigs. The findings revealed that the spleen of pigs orally administered with the vehicle had significantly decreased numbers of lymphocytes in the white pulp area, most lymphocytes were necrotic, a small number of renal glomeruli showed evident atrophy and necrosis, some renal tubular epithelial cells showed edema, and lymphocytes showed necrosis and were decreased in number in inguinal lymph nodes (Figure 9J). The pathological changes were significantly reduced in ASFV-infected pigs orally administered with drugs, to almost the same level as that observed after the oral administration of vehicle alone (Figure 9J). ASFV virions were clearly observed in the spleen, kidney, and inguinal lymph nodes by an IHC assay of ASFV-infected pigs orally administered with the vehicle. ASFV virions from ASFV-infected pigs orally administered with arctiin and genistein were less abundant than those fed with arctiin or genistein in the spleen, kidney, and inguinal lymph nodes (Fig. 9K).

DISCUSSION

In this study, we successfully screened four B. subtilis strains against ASFV in vitro and estimated their antiviral effect using different preparation types in vivo. We found that the survival rate of pigs fed with biologics was up to 100% compared with pigs infected with ASFV alone. Furthermore, arctiin and genistein, metabolites of B. subtilis, target the ATP-binding site of viral type II topoisomerase, thereby hindering its catalytic activity and disrupting viral DNA synthesis.

B. subtilis, as a probiotic, has antiviral functions. It has been reported that the small double-stranded RNA from B. subtilis MTCC5480 exhibits anti-HIV activity (46). In addition, one of the secretions of B. subtilis, surfactin, can inhibit viral hemorrhagic septicemia virus (VHSV) infection in olive flounder intestinal epithelial cells and the infection of internal organs (26). Our results also showed that the four B. subtilis strains screened in this study exerted anti-ASFV activity both in vitro and in vivo. Although B. subtilis slowed down ASFV infection, ASFV genomic DNA and viral titers were still detectable in the organs at 28 dpi. Pigs are the only target animal of ASFV infection, and virulent ASFV can rapidly proliferate in primary PAMs, PBMCs, and BMDMs. Therefore, ASFV is likely to rebound in the absence of intervention strategies after 28 dpi. The infected pigs subjected to treatment would be expected to show chronic clinical signs of ASF if the experiments were extended. Therefore, antiviral agents that are effective over the longer term are worthy of future investigation.

Bacillus subtilis is also used as a vector to express and display the transmissible gasteroenteritis virus (TGEV) spike protein. Recombinant B. subtilis can recruit dendritic cells, migrate to mesenteric lymph nodes, and induce immune responses (47). However, there was no significant difference in the percentage of positive CD4⁺ and CD8⁺ T lymphocytes between pigs fed with and without B. subtilis by oral administration before and after ASFV challenge (Fig. 3A and H). Similarly, there was no significant difference in cytokines, including the expression of IL-4, IL-8, IL-1β, IFN-β, IFN-γ, and IFN-α before and at 7 dpi with ASFV; however, the difference was significant at 14 dpi (Fig. 3I through N). This may be because during the late stage of ASFV infection in pigs, which is characterized by a massive increase in virions in vivo, a cytokine storm can be induced (33).

Critical ASFV proteins are used as antiviral targets. For example, the viral pS273R protease can cleave polyproteins pp62 and pp220, which thereby become mature viral proteins for viral assembly (48). Peptidomimetic aldehyde compounds and myricetin target the pS273R protease, exerting prominent inhibitory effects (7, 49). ASFV pA104R, a nucleoid-associated histone-like protein, is indispensable for successful viral replication (50). Liu et al. reported that stilbene derivatives, SD1 and SD4, could disrupt pA104R-DNA binding and inhibit ASFV replication in swine macrophages (2). Type II topoisomerase encoded by ASFV P1192R modulates the topological state of viral DNA during replication and transcription, which causes overwinding and/or underwinding of the DNA and requires ATP to provide energy for catalytic activity (15). Genistein, as a competitive inhibitor of ATP, can bind to the ATP-binding site of type II topoisomerase, hampering viral DNA synthesis (14). In this study, we similarly found that genistein, a metabolite of B. subtilis, exhibited anti-ASFV effects. Meanwhile, a new small molecule, arctiin, was detected among the metabolites of B. subtilis that can target the ATP-binding site of type II topoisomerase, thus inhibiting ASFV infection (Fig. 8). Arctiin and genistein also exerted anti-inflammatory effects in previous reports (51, 52). Zhou et al. reported that arctiin treatment reduced H9N2 virus-triggered proinflammatory cytokines, including IL-6 and TNF-α, and evidently reversed the H9N2 virus-induced reduction of Nrf2, increasing the nuclear translocation of Nrf2 (45). Furthermore, arctiin suppresses activation of the NOD-like receptor thermal protein domain associated protein 3 (NLRP3) inflammasome through the Toll-like receptor 4 (TLR4)/myeloid differentiation factor 88 (Myd88)/nuclear factor-k-gene binding (NF-κB) pathway, inhibiting silicosis (53). Xu et al. found that arctiin ameliorates depression by inhibiting the activation of microglia and inflammation via the high mobility group box protein 1 (HMGB1)/TLR4 and TNF-α/TNFR1 signaling pathways (54). However, arctiin did not inhibit ASFV proliferation by suppressing the inflammatory response in our research (Fig. S4). The survival rates of pigs fed arctiin, genistein, and B. subtilis were different (Fig. 2 and 9). It is worth noting that we only analyzed two small molecules, and other small molecules with significant antiviral effects might have remained undetected. Commercially available small molecules are limited, and there are also small molecules that could not be characterized using LC-MS/MS.

Viremia has been observed in pigs infected with ASFV (55). The small molecules can enter the blood and host cells targeting ASFV, thereby inhibiting viral proliferation, which is not possible to achieve with vaccines. Pigs fed with B. subtilis showed an increase in the abundance of beneficial bacteria and a decrease in potential pathogenic bacteria (56). Metabolites of beneficial bacteria are also likely to play an antiviral role in ASFV infection. Changes in the composition of the intestinal microbiota in response to B. subtilis require further investigation.

In summary, our data demonstrated that B. subtilis suppresses ASFV replication both in vitro and in vivo. Arctiin and genistein, competitive inhibitors of metabolites from B. subtilis, can compete with ATP, deactivating the catalytic function of ASFV type II topoisomerase, thus disrupting viral DNA synthesis and reducing the number of lesions caused by ASFV infection in pigs. To the best of our knowledge, this is the first study to evaluate the anti-ASFV effects of B. subtilis. In the absence of effective strategies for the prevention and control of ASFV, our findings shed new light on potential drugs to combat ASFV.

DATA AVAILABILITY

The metabolomic data generated and analyzed in the current study are available in the MetaboLights repository (accession number MTBLS6586). All data and materials supporting the findings of this study are available from the corresponding author upon reasonable request.

ETHICAL STATEMENT

All experiments on live ASFV manipulation and animal infections were carried out in the animal biosafety level 3 (ABSL-3) laboratory of Huazhong Agricultural University and were approved by the Ministry of Agriculture and Rural Affairs of China. All animal studies were performed in accordance with the Care and Use of Laboratory Animals of the Research Ethics Committee, Huazhong Agriculture University (HZAUSW-2022–0030).

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/jvi.00719-23.

Whole WBlot images. jvi.00719-23-s0001.docx.

Whole WBlot images associated with figures in the article.

Supplemental figures and tables. jvi.00719-23-s0002.pdf.

Figure S1 to S6 and Table S1.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.