A predictive multiscale framework for post-translational modification-dependent peptide-MHC class I binding.

The specific binding of antigenic peptides to major histocompatibility complex class I (MHC-I) molecules is a pivotal step in adaptive immune responses. Post-translational modifications (PTMs) have been shown to profoundly regulate this process and thereby modulate T-cell recognition; however, their atomic-level mechanisms remain insufficiently understood. To address this gap, we employed all-atom molecular dynamics simulations combined with multidimensional energetic and dynamic analyses to systematically dissect PTM-dependent regulatory mechanisms across diverse antigen peptide-MHC-I (pMHC) systems. Representative viral (SARS-CoV-2 spike protein), model (ovalbumin), autoimmune-associated (MBP), and tumor-associated (TVF and RSP) antigen peptides were examined, encompassing acetylation, phosphorylation, citrullination, methylation, hydroxylation, and succinylation modifications. Our results demonstrate that PTM effects are highly context-dependent and governed by both the modified site and the physicochemical nature of the introduced functional group. Charge-altering modifications at critical anchoring positions-such as N-terminal acetylation and phosphorylation-substantially weaken pMHC binding by disrupting electrostatic complementarity, reorganizing hydrogen-bond networks, accompanied by altered collective motions, and expanding the MHC α1/α2 binding groove. In contrast, conservative modifications located in solvent-exposed regions (e.g., lysine methylation) exert minimal structural and energetic perturbations. Notably, citrullination in disease-associated antigens enhances binding affinity through strengthened hydrophobic interactions, optimized hydrogen-bond rearrangements, accompanied by increased dynamic cooperativity, and contraction of the binding groove, providing a mechanistic basis for its immunological consequences. Across all systems, PTMs regulate pMHC recognition through multiscale coupling mechanisms that integrate residue-level energetic redistribution, cooperative motion reprogramming, and global groove geometry remodeling. Importantly, the simulation-derived binding trends are consistent with available experimental observations, supporting the reliability of the computational framework. Collectively, this study establishes a unified structure-energy-dynamics model explaining how PTMs function as atomic-level chemical switches in antigen presentation. Beyond mechanistic insight, the demonstrated agreement with experimental data suggests that this computational strategy possesses predictive potential for estimating PTM-dependent pMHC binding behaviors across diverse immunological contexts.