When theory meets experiment: the PD-1 challenge.

Area of Science:

• Computational biology and structural bioinformatics.
• Drug discovery and molecular modeling.

Background:

• Atomistic computational modeling is highly successful for predicting drug-receptor interactions.
• Predicting protein-protein interactions and designing their regulators remains a significant challenge for computational tools.
• The programmed cell death protein 1 (PD-1) and its ligand PD-L1 form a critical complex in immune regulation.

Purpose of the Study:

• To evaluate the accuracy of advanced computational simulations in predicting complex protein-protein interactions.
• To validate a computational workflow for protein-protein interaction prediction using the human PD-1/PD-L1 complex as a case study.
• To compare simulation-derived models with experimental crystal structures.

Main Methods:

• Utilized state-of-the-art computer simulations guided by experimental data.
• Revisited and analyzed a previously predicted model of the human PD-1/PD-L1 complex.
• Performed a side-by-side comparison of the computational model with the recently published crystal structure of the complex.

Main Results:

• The computational simulations demonstrated outstanding agreement with the experimental crystal structure of the human PD-1/PD-L1 complex.
• This represents a rare instance where sophisticated protein-protein interactions were correctly predicted computationally.
• The findings validate the accuracy of the developed protein-protein prediction workflow.

Conclusions:

• The study successfully validated a computational workflow for predicting complex protein-protein interactions.
• The validated workflow shows potential for rational design of regulators targeting similar protein-protein interactions.
• This approach could advance drug discovery efforts by enabling accurate prediction of protein complex formation.