Deep learning-based dipeptidyl peptidase IV inhibitor screening, experimental validation, and GaMD/LiGaMD analysis

Area of Science:

• Computational chemistry
• Drug discovery
• Biochemistry

Background:

• Dipeptidyl peptidase-4 (DPP4) is a key target enzyme in type 2 diabetes (T2D) management.
• DPP4 inhibitors enhance glucagon-like peptide-1 (GLP-1) levels, crucial for glycemic control.

Purpose of the Study:

• To develop a novel and accurate screening strategy for identifying DPP4 inhibitors.
• To investigate the binding and dissociation mechanisms of DPP4 inhibitors.

Main Methods:

• Integrated receptor-based ConPLex, ligand-based KPGT, and molecular docking for enhanced screening accuracy.
• Utilized Gaussian accelerated Molecular Dynamics (GaMD) and ligand Gaussian accelerated Molecular Dynamics (LiGaMD) for mechanism analysis.
• Developed DPP4META server and pymd Python toolkit for inhibitor prediction and binding analysis.

Main Results:

• Achieved a 100% hit rate in identifying potential DPP4 inhibitors from the FDA database.
• Identified Isavuconazonium as the most potent inhibitor with an IC50 of 6.60 µM.
• Demonstrated strong correlations between molecular dynamics simulations and IC50 values.

Conclusions:

• The developed screening approach provides a robust method for DPP4 inhibitor discovery.
• Offers valuable insights into protein-compound binding and dissociation mechanisms.
• Facilitates the development of effective T2D therapeutics.