AI-enabled structural bioinformatics identifies repositioned kinase inhibitor against Poxviridae kinases

Emerging zoonotic poxviruses such as lumpy skin disease virus (LSDV) and monkeypox pose significant threats to animal and human health, yet validated antiviral targets remain scarce. Variola virus, the causative agent of smallpox, also belongs to the Poxviridae family, emphasizing the importance of preparedness against potential re-emergence.

We developed an AI-Enabled structural bioinformatics pipeline integrating sequence conservation, phylogenetic analysis, AlphaFold2/ESMFold structure prediction, structural comparison, active site mapping, molecular docking, molecular dynamics simulations, and cheminformatics profiling.

Two conserved kinases encoded in the LSDV genome-a serine/threonine kinase (LSTK) and a tyrosine kinase (LYK)-were identified as viral proteins with druggable domains, with subsequent structural and inhibitor analyses primarily focused on LSTK. High-confidence structural models of LSTK (pLDDT >85, pTM ∼0.85-0.88) revealed conserved motifs and functional similarities to monkeypox MSTK and related poxvirus kinases. Virtual screening of 88 FDA-approved kinase inhibitors identified lapatinib as a potential competitive inhibitor of LSTK, exhibiting stable ATP-site occupancy and favorable binding free energies during molecular dynamics simulations. Principal component analysis of physicochemical descriptors demonstrated overlap between kinase inhibitors and approved cattle antivirals, supporting translational feasibility. Additionally, a combinatorial analog library was generated to support future synthesis and experimental validation studies.

This study illustrates the utility of AI-assisted protein structure prediction integrated with structural bioinformatics approaches for the rapid identification of repurposable kinase inhibitors against viral targets. The proposed workflow provides a generalizable framework for structure-guided antiviral drug discovery in veterinary and human infectious disease contexts.