Energy landscapes of peptide-MHC binding.

Area of Science:

• Immunology
• Computational Biology
• Biophysics

Background:

• Major Histocompatibility Complex (MHC) molecules present peptide fragments to T cells, initiating immune responses.
• MHC-I and MHC-II present peptides from intracellular and extracellular proteins, respectively, playing distinct roles in immunity.
• Understanding MHC-peptide binding is critical for deciphering immune recognition and developing targeted therapies.

Purpose of the Study:

• To investigate the sequence-dependent energy landscapes of MHC-peptide binding and their encoded nonlinearities (epistasis).
• To compare the epistasis characteristics of MHC-I and MHC-II and their implications for immune recognition and evolution.
• To assess how epistasis affects the predictability of binding energies and the impact of mutations on epitopes.

Main Methods:

• Analysis of sequence-dependent energy landscapes of MHC-peptide binding.
• Development of a matrix model for MHC-I binding energies.
• Comparison of model performance with machine learning approaches.
• Investigation of epistasis effects on epitope mutations.

Main Results:

• MHC-I exhibits a smooth energy landscape with global epistasis, enhancing discrimination of strong-binding peptides.
• MHC-II displays a rugged landscape with idiosyncratic epistasis, where binding depends on specific amino acid combinations.
• A simple matrix model for MHC-I outperforms complex machine learning models.
• MHC-II's complexity hinders learning by simple regression methods.
• Epistasis influences the location and magnitude of mutations affecting epitope presentation and immune escape.

Conclusions:

• The distinct epistasis landscapes of MHC-I and MHC-II shape the dynamics of T cell immunity and immune escape.
• MHC-I's global epistasis simplifies binding energy prediction and mutation impact analysis.
• MHC-II's idiosyncratic epistasis creates a broader target for immune escape mutations, influencing adaptive immune responses.