Structural and Computational Analysis of African Swine Fever Virus Capsid Proteins for Antiviral Drug Design
1. Introduction
African swine fever virus (ASFV) is the etiological agent of African swine fever, a hemorrhagic disease of domestic and wild suids that poses a significant threat to global swine production. The virus is a large, enveloped, double-stranded DNA virus classified within the family Asfarviridae and genus Asfivirus. ASFV possesses a multilayered icosahedral capsid that encloses a dense core genome. The capsid is a critical structural component responsible for genome protection, cellular entry, and assembly. Understanding the atomic architecture of the ASFV capsid is a prerequisite for rational antiviral drug design. This review focuses on the integration of cryo-electron microscopy (cryo-EM), X-ray crystallography, homology modeling, molecular docking, and molecular dynamics (MD) simulations to characterize capsid protein structures and identify potential druggable pockets. The computational strategies discussed here are directly applicable to the development of small-molecule inhibitors that disrupt capsid assembly or stability. For broader context on ASFV biology and detection, readers are referred to the article on African Swine Fever Virus and the diagnostic approaches described in CRISPR-Cas12a-Based Biosensor for Rapid Detection of African Swine Fever Virus: From Assay Design to Field Deployment.
2. Architecture of the ASFV Capsid
The ASFV capsid is an icosahedral shell with a triangulation number T=277, one of the largest capsid geometries known among viruses. The capsid is composed primarily of the major capsid protein p72 (also designated B646L), along with minor capsid proteins such as p17 (D117L), p49 (B438L), and the penton protein (E120R). Cryo-EM reconstructions at near-atomic resolution have revealed that p72 forms trimeric capsomers arranged on a pseudo-hexagonal lattice. Each p72 monomer consists of a jelly-roll fold composed of two beta-barrel domains, a hallmark of many icosahedral viruses. The minor proteins stabilize the capsid by occupying specific positions at the vertices and edges of the icosahedral facets. The penton protein forms the fivefold symmetry axes, while p17 and p49 contribute to the capsid floor and inter-capsomer contacts. The structural integrity of this shell is essential for viral infectivity, as disruption of capsid assembly leads to non-infectious particles.
3. Experimental Structure Determination: Cryo-EM and X-ray Crystallography
High-resolution structural data for ASFV capsid proteins have been obtained primarily through single-particle cryo-EM and X-ray crystallography. Cryo-EM has been instrumental in determining the overall architecture of the intact capsid at resolutions approaching 3.5 to 4.0 angstroms. These reconstructions have allowed the unambiguous placement of p72 trimers and the identification of minor capsid components. X-ray crystallography has provided atomic models of individual capsid proteins, including the p72 core domain and the p17 membrane-associated protein. The combination of these techniques yields a comprehensive view of capsid organization. For example, the crystal structure of p72 revealed a conserved beta-sandwich fold with extended loops that mediate inter-trimer contacts. Cryo-EM maps further showed that these loops adopt distinct conformations in the assembled capsid compared to the isolated protein, indicating conformational flexibility that may be targeted by small molecules. The integration of cryo-EM density maps with crystallographic models is a standard workflow in structural virology, as detailed in the article on Cryo-EM Density Map Interpretation and Computational Structure Fitting.
4. Homology Modeling and Comparative Structural Analysis
For capsid proteins that have not been crystallized, homology modeling provides a reliable alternative for generating three-dimensional structures. This approach relies on the identification of template proteins with known structures and sequence similarity to the target ASFV protein. The p49 and E120R proteins, for instance, have been modeled using templates from other large DNA viruses such as poxviruses and iridoviruses. The accuracy of homology models depends on sequence identity and the quality of the alignment. Models are typically refined using energy minimization and validated through Ramachandran plots and MolProbity scores. Comparative structural analysis of ASFV capsid proteins with those of other nucleocytoplasmic large DNA viruses (NCLDVs) has revealed conserved folds and potential functional motifs. Such comparisons can highlight conserved pockets that are essential for capsid assembly and are therefore attractive targets for broad-spectrum antiviral agents. The principles of homology modeling are further explored in the article on Structural Prediction of Viral Envelope Glycoproteins Using AlphaFold2: Implications for Host Receptor Binding and Vaccine Design.
5. Identification of Druggable Binding Sites
The identification of small-molecule binding sites on capsid proteins is a central goal of structure-based drug design. Computational solvent mapping algorithms, such as FTMap and SiteMap, are used to identify pockets on the protein surface that are geometrically and chemically favorable for ligand binding. These algorithms place small organic probe molecules across the protein surface and cluster the binding hotspots. For the ASFV p72 trimer, several conserved pockets have been identified at the inter-monomer interfaces. These pockets are lined with hydrophobic residues and contain hydrogen bond donors and acceptors, making them suitable for binding small molecules. Additionally, the penton protein E120R contains a central cavity that is essential for vertex formation. Virtual screening of compound libraries against these pockets can identify lead molecules that inhibit capsid assembly. The workflow for pocket identification and virtual screening is described in Structure-Based Drug Design in Bioinformatics: Computational Pipelines, Active Site Grid Mapping, and Virtual Screening Workflows.
6. Molecular Docking of Antiviral Compounds
Molecular docking is a computational technique that predicts the binding orientation and affinity of a small molecule within a target binding site. Docking programs such as AutoDock Vina, Glide, and GOLD use scoring functions that estimate the free energy of binding based on van der Waals interactions, electrostatic forces, and desolvation penalties. For ASFV capsid proteins, docking studies have focused on the p72 inter-trimer interface and the p17 membrane-binding region. Libraries of natural products, FDA-approved drugs, and custom-designed small molecules have been screened in silico. Top-ranked compounds are selected based on docking scores, ligand efficiency, and predicted pharmacokinetic properties. It is critical to validate docking results through MD simulations to assess the stability of the predicted complexes. False positives arising from scoring function inaccuracies are common, and consensus docking approaches that combine multiple scoring functions can improve reliability. The methodology of molecular docking is covered in detail in Computational Modeling of Protein-Ligand Docking and Alphafold Protein Ligand Docking: Structural Analysis and Computational Methodologies in Bioinformatics.
7. Molecular Dynamics Simulations of Capsid Stability and Ligand Binding
Molecular dynamics simulations provide a dynamic view of protein-ligand interactions and capsid assembly. All-atom MD simulations of ASFV capsid fragments, such as p72 trimers or penton complexes, are performed using force fields like CHARMM36 or AMBER ff14SB. Simulations are typically run for hundreds of nanoseconds to microseconds to capture conformational changes and ligand binding events. The root-mean-square deviation (RMSD) and root-mean-square fluctuation (RMSF) of protein backbone atoms are monitored to assess structural stability. Binding free energies are calculated using the molecular mechanics generalized Born surface area (MM-GBSA) or molecular mechanics Poisson-Boltzmann surface area (MM-PBSA) methods. These calculations can rank ligand affinities and identify key residues contributing to binding. For ASFV, MD simulations have shown that the p72 trimer undergoes significant conformational breathing at the inter-monomer interfaces, which may be exploited by inhibitors that lock the protein in a non-productive state. Simulations of the full capsid shell, though computationally expensive, are becoming feasible with coarse-grained models and enhanced sampling techniques. The application of MD to viral capsids is reviewed in Molecular Dynamics of Viral Capsid Assembly and Stability and Molecular Dynamics Simulations of Proteins and Force Fields.
8. Computational Workflow for Capsid-Targeted Drug Design
The following Mermaid diagram outlines a typical computational workflow for identifying and validating small-molecule inhibitors against ASFV capsid proteins.
flowchart TD
A[Experimental Structures: Cryo-EM / X-ray], > B[Structure Refinement & Validation]
C[Target Sequence], > D[Homology Modeling]
B, > E[Binding Site Identification: Solvent Mapping]
D, > E
E, > F[Virtual Screening: Molecular Docking]
F, > G[Hit Selection: Docking Score & Ligand Efficiency]
G, > H[Molecular Dynamics Simulations: Stability & Binding Free Energy]
H, > I[Lead Optimization: Medicinal Chemistry]
I, > J[In Vitro Antiviral Assays]
J, > K[In Vivo Efficacy Studies]
This pipeline integrates experimental and computational methods to reduce the time and cost of drug discovery. Each step involves iterative feedback, with MD results informing new docking rounds and experimental validation guiding model refinement.
9. Targeting Capsid Assembly and Disassembly
Beyond direct inhibition of a single protein, computational approaches can target the assembly and disassembly processes of the capsid. The ASFV capsid assembles from soluble p72 trimers and minor proteins at the viral factory membranes. Disruption of protein-protein interactions (PPIs) at the capsid interfaces can prevent the formation of a stable shell. Computational alanine scanning and free energy decomposition methods identify hot spot residues that contribute disproportionately to binding affinity. These hot spots are prime targets for small-molecule PPI inhibitors. For example, the interface between p72 and p17 has been shown to contain several conserved hydrophobic hot spots. Fragment-based drug design, where small chemical fragments are docked to these hot spots and then linked, is a promising strategy. The design of PPI inhibitors is discussed in Protein-Protein Interface Design and Binding Energy Prediction and Computational Design of Antiviral Peptides Targeting Viral Envelope Proteins.
10. Integration with Machine Learning and Deep Learning
Recent advances in machine learning (ML) and deep learning have enhanced the accuracy of structure prediction and binding affinity estimation. AlphaFold2 and RoseTTAFold have been used to predict the structures of ASFV capsid proteins that lack experimental templates, achieving high confidence scores for globular domains. These predicted structures can serve as starting points for docking and MD simulations. Graph neural networks (GNNs) trained on protein-ligand complexes can predict binding affinities with improved accuracy over classical scoring functions. Additionally, generative models such as variational autoencoders (VAEs) and generative adversarial networks (GANs) are being explored to design novel ligands tailored to ASFV capsid pockets. The integration of these tools into the drug discovery pipeline is covered in Protein Language Models in Drug Discovery: Embeddings, Variant Effect Prediction, and Binder Prioritization and The Bioinformatics Revolution in Structural Proteomics and Computational Drug Discovery: A Unified Paradigm.
11. Challenges and Future Directions
Despite significant progress, several challenges remain in the computational analysis of ASFV capsid proteins. The large size and complexity of the capsid limit the application of all-atom MD simulations to sub-complexes. Coarse-grained and multiscale modeling approaches are needed to simulate entire capsid assembly pathways. The conformational flexibility of capsid proteins, particularly in loop regions, complicates docking studies. Ensemble docking, where multiple protein conformations are used, can partially address this issue. Another challenge is the lack of high-resolution structures for all capsid components, particularly the minor proteins. Continued investment in cryo-EM and crystallography is essential. Finally, the translation of computational hits into antiviral drugs requires rigorous experimental validation, including cell-based infection assays and animal models. The computational modeling of ASFV spread in wild boar populations, as described in African Swine Fever: Computational Models for Early Detection and Spread Prediction in Wild Boar Populations, can complement drug development by identifying high-risk areas for field trials.
12. Conclusion
The structural and computational analysis of ASFV capsid proteins provides a robust foundation for antiviral drug design. Cryo-EM and X-ray crystallography have elucidated the atomic architecture of the capsid, while homology modeling extends structural coverage to less characterized proteins. Molecular docking and MD simulations enable the identification and optimization of small-molecule inhibitors that target capsid assembly and stability. The integration of machine learning further accelerates the discovery process. A multidisciplinary approach that combines experimental structural biology with computational chemistry and virology is essential for developing effective therapeutics against African swine fever.
References
- Standard veterinary virology textbooks and general knowledge of ASFV structure and life cycle. (No specific journal citations available from the provided literature context.)
Disclaimer: This article is for educational and informational purposes only. It is not intended to substitute for professional veterinary advice, diagnosis, treatment, or regulatory guidance. Always consult a licensed veterinarian or qualified specialist regarding animal health, disease diagnosis, and therapeutic decisions.