Machine learning-driven discovery of antimicrobial peptides targeting the GAPDH-TPI protein-protein interaction in...

Schistosomiasis, caused by Schistosoma mansoni, remains a significant public health burden, particularly in endemic regions with limited access to effective treatment. The emergence of resistance to praziquantel necessitates the urgent discovery of novel therapeutic targets. This study explores the potential of antimicrobial peptides (AMPs) as inhibitors of the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and triose phosphate isomerase (TPI) protein-protein interaction (PPI) in S. mansoni, a crucial glycolytic pathway component essential for parasite survival. A machine learning-driven approach was employed to filter 3306 AMPs from the Antimicrobial Peptide Database (APD) based on physicochemical properties and predicted binding affinities. Eighteen peptides were selected based on desirable physicochemical attributes and further subjected to molecular docking using HADDOCK 2.4. The results identified AP02590 (-103.5 ± 2.7 kcal/mol) and AP02754 (-87.8 ± 1.0 kcal/mol) as the most promising inhibitors, exhibiting strong binding affinities and stable complex formation compared to the native GAPDH-TPI complex (-77.8 ± 17.2 kcal/mol). Molecular dynamics (MD) simulations confirmed the stability of these complexes, with lower root mean square deviation (RMSD) values (AP02590: ∼2.5 Å, AP02754: ∼3.0 Å) and reduced root mean square fluctuation (RMSF) of key interacting residues. Radius of gyration (Rg) analysis further indicated compact structural stability. MMGBSA analysis validated these findings, showing favourable binding free energies for AP02590 (-50.80 ± 0.90 kcal/mol) and AP02754 (-46.31 ± 0.83 kcal/mol), reinforcing their potential as lead compounds for antischistosomal drug development. These findings provide a foundation for further experimental validation of peptide-based inhibitors targeting metabolic pathways in S. mansoni.