Thonningianin A disrupts pA104R-DNA binding and inhibits African swine fever virus replication

Abstract

African swine fever is a highly lethal disease caused by the African swine fever virus (ASFV), posing a significant threat to the global pig industry, wherease no approved treatments are currently available. The ASFV DNA-binding protein, pA104R, plays a critical role in viral genome packaging and replication, making it a key target for drug discovery. Through structure-based virtual screening, we identified a polyphenolic compound, thonningianin A, which disrupts the pA104R-DNA binding and significantly inhibits ASFV replication. Mechanistic study revealed that thonningianin A binds to the DNA-binding region of pA104R, forming strong hydrogen bonds with H100 and occupying the vital DNA-binding residues K92, R94, and K97. In addition, we resolved the high-resolution (1.8 Å) structure of pA104R (PDB ID 9JS5), providing valuable insights for future drug screening. Together, these results demonstrate that thonningianin A holds great potential for the development of anti-ASFV drug, as a herb extract with favourable pharmacokinetic properties and safety.

Keywords: ASFV; DNA-binding protein; EMSA; pA104R; structure-based virtual screening; thonningianin A.

Graphical abstract

Figure 1.

Structure-based discovery of small molecules targeting the β-strand DNA binding region (BDR) of pA104R. a Flowchart of the structure-based hierarchical virtual screening process. The figure was generated using BioRender.com. b Diverse binding energy levels exhibited by compounds in chemical libraries. Bars represent numbers of compounds with predicted free energies of binding in the indicated 1 kcal/mol bins.

Figure 2.

Identification of small molecules that inhibit pA104R−DNA binding using the EMSA method. a The effects of the screened 50 candidate compounds on the inhibition of pA104R−DNA binding in EMSA. DMSO was added as negative control and SD4 as positive control. The percentage of pA104R−DNA binding was analyzed by ImageJ software. The data are presented as the mean ± SD from three independent experiments. b–c The dose-dependent effects of compound 1 and compound 46 on the inhibition of pA104R−DNA binding in EMSAs. The data are presented as the mean ± SEM from three independent experiments. P-values were calculated using a one-way ANOVA followed by pairwise comparisons with Dunnett’s multiple comparisons test, with DMSO as the control. ****P < 0.0001. For clarity, only significant correlations are shown. d–e BLI measurement of the binding between active compounds and pA104R. His-tagged WT pA104R was immobilized on NTA biosensors, and incubated with compounds diluted in a 2-fold series from 100 μM to 6.25 μM. Double reference subtraction method was processed to subtract the effect of baseline drift and non-specific binding. The kinetic parameters were calculated using a 1:1 global fit model, and analyzed using the Octet data analysis software (ForteBio, version 9.0). Data shown are representative of three independent experiments. f–g The 2D chemical structure of compound 1 and compound 46.

Figure 3.

Inhibition of ASFV replication by thonningianin A. The inhibitory effects of thonningianin A on ASFV replication, as reflected by reductions in ASFV DNA levels (a, b). DNA was extracted from treated cells, and the VP72 gene was amplified using qPCR. The relative DNA levels were determined by normalizing the viral copy numbers in the test wells to the mean viral copy number in the negative control wells (DMSO), which was set to 100%. Data are presented as the mean ± SEM from three independent experiments. **** p < 0.0001; *** p < 0.001; ** p < 0.01; ns = not significant, as determined by a one-way ANOVA followed by pairwise comparisons with Dunnett’s multiple comparisons test, with DMSO as the control. c Cytotoxicity evaluation of thonningianin A in PAMs. Data are presented as the mean ± SD from three independent experiments. d Schematic illustration of time-of-addition experiment. PAM cells were infected with ASFV (MOI: 0.1) for 2 h. Thonningianin A (4 µM) was introduced at different time points. The viral genome copies were determined at 36 hpi by qPCR. e The effects of thonningianin A added at different time points on the viral genome copies. Data are presented as the mean ± SEM from three independent experiments. Statistical analysis was performed with two-way ANOVA, followed by Sidak's multiple comparisons test for pairwise analysis. **** p < 0.0001; ns, no statistical difference.

Figure 4.

Detection of pA104R−DNA binding by Proximity Ligation Assay in the presence of the compounds. (left) Duolink proximity ligation assay for pA104R−DNA binding in Vero cells. The cells were treated with SD4 (2 μM) or compound 46 (2 μM), and then a Duolink assay was performed as described, with untreated cells serving as negative controls. Association of pA104R with DNA was detected as red puncta within cells. Representative images are shown, with experiments performed in triplicate. Nuclei were counterstained with DAPI (blue). Bars, 10 μm. (right) Red puncta were quantified in randomly selected cells (n = 20) using Image-Pro Plus 7.0 software (Media Cybernetics, USA). Data are presented as the mean of 20 cells per condition, collected from three independent biological replicates (n = 3), with error bars representing ± SEM. Statistical significance was determined using nonparametric Kruskal–Wallis one-way ANOVA on ranks, followed by Dunn’s multiple comparison test. ***p < 0.001; ****p < 0.0001.

Figure 5.

Computational analysis of the binding mode of thonningianin A to pA104R. a Representative structure of pA104R–thonningianin A obtained from clustering analysis of the MD trajectories. The pA104R protein is displayed in a cartoon representation, with the backbone atoms of chain A depicted in green and those of chain B in magenta. The ligand is displayed in yellow sticks. b RMSF of Cα atoms relative to the average structure of pA104R. c The RMSD values for thonningianin A-bound pA104R system along the MD simulation time. Black curves represent the RMSD values of pA104R backbone atoms (excluding the highly flexible terminal tail residues 66-85), whereas the yellow lines represent the RMSD curve of compound 46 heavy atoms. d Predicted binding pose of thonningianin A in pA104R. compund and the amino acid residues involved the protein-ligand interactions are shown as sticks. Hydrogen bonds are depicted as black dotted lines. e The 2D ligand-protein interaction diagrams. Residues within 5 Å of thonningianin A are displayed. Hydrogen bonds are represented by magenta arrows, and π-π stacking is marked in green. Dark blue indicates positive charge, red indicates negative charge, light blue for polar, and green for hydrophobic. f Distance of H-bond trajectories during MD simulation. g Residue-wise energy decomposition analysis. The diagram displays residues with a total energy contribution of ≤ −0.05 kcal/mol. The top ten amino acids contributing most significantly to the ligand binding are labelled with red asterisks (*). h Multiple sequence alignment of pA104R across different ASFV genotypes. The alignment was performed using ClustalW, and conservation levels were visualized with ESPript 3.0. The predicted critical binding residues of thonningianin A are highlighted by blue arrows.

Figure 6.

Validation of the potential binding sites of thonningianin A through hotspot mutations and EMSA analysis. (a–f) Effects of thonningianin A on the inhibition of DNA binding to K92D (a), R94D (b), K97D (c), H100D (d), and H100A (e) mutants in EMSAs. The wild-type pA104R was used as a control. The data are presented as the mean ± SEM from three independent experiments. P-values were calculated using a two-way ANOVA followed by Sidak's multiple comparisons test for pairwise analysis. ****P < 0.0001. For clarity, only the comparisons at the highest concentration are shown. (f–j) The binding dynamics between K92D (f), R94D (g), K97D (h), H100D (i), and H100A (j) and compound46, respectively. Double reference subtraction method was processed to subtract the effect of baseline drift and non-specific binding. The kinetic parameters were calculated using a 1:1 global fit model, with the Rmax unlinked by sensors, analyzed using the Octet data analysis software (ForteBio, version 9.0). Data shown are representative of three independent experiments.

Figure 7.

The x-ray structure of pA104R. a The overall structure of dimeric pA104R. b One protomer of the dimeric pA104R. c Structural superposition of for the pA104 (PDB ID 9JS5) with the previously reported structures of pA104R both in its ligand-free form (grey, PDB ID 6LMH) and in complex with DNA (light blue, PDB ID 6LMJ). d Surface electrostatic potential of pA104R. The electrostatic potential was calculated using the Adaptive Poisson-Boltzmann Solver (APBS) programme within the PyMOL environment. The calculations of the electrostatic potential were based on a pH of 7.5, with values ranging from −5 kT (red) to 5 kT (blue). e The sulfate ions in the BRD region of the crystal structure and their interacting residues. Sulfate ions are displayed as spheres, while amino acids are shown using the stick model. The PDB file was generated using molecular replacement, with residue numbering inherited from the template structure (PDB ID: 6LMH), which is offset by +6 relative to the NCBI reference sequence (GenBank: AYW34006.1). For consistency with previous studies, residue numbering in the figures follows the NCBI reference sequence. The visualization and analysis of structures were conducted using the PyMOL Molecular Graphics System (version 3.0.0).