Disruption of Hsp70.14–BAG2 Protein–Protein interactions using deep Learning–Driven peptide design and molecular simulations

Protein-protein interactions (PPIS) are critical in proteostasis, stress response, and disease progression. Targeting the interaction between Hsp70.14 and BAG2, a co-chaperone implicated in oncogenic survival, offers a promising therapeutic approach. This study employed a comprehensive in silico framework to identify bioactive antimicrobial peptides (AMPS) capable of disrupting the Hsp70.14–BAG2 interaction. In this study, we present an integrated in silico pipeline combining deep learning-based peptide screening, molecular docking, molecular dynamics (MD) simulations, and MM-GBSA free energy analysis to identify antimicrobial peptides (AMPS) capable of disrupting the Hsp70.14–BAG2 interaction. Peptides were shortlisted from an extensive public database using stringent physicochemical and safety filters, yielding candidates with high therapeutic potential. DBAASPS_19370 and DBAASPS_17167 exhibited more favourable binding free energies and lower dissociation constants than the reference Hsp70.14–BAG2 complex. Molecular dynamics simulations revealed that one lead peptide demonstrated superior complex stability, characterised by compact structure, reduced flexibility, and solvent exposure. MM/GBSA calculations further validated these observations, revealing the most favourable free energy profile for DBAASPS_19370 compared to the other complexes. Interface analysis showed improved atomic packing, enhanced residue participation, and better solvation energetics in the selected peptide complexes. These findings highlight DBAASPS_19370 as a potent AMP candidate capable of competitively inhibiting BAG2 and disrupting its interaction with Hsp70.14, offering a rational avenue for chaperone-targeted therapeutic development. Future work may explore in vitro validation and structural optimization of this peptide to support its translational potential.