Favorite this paper
How to cite this paper?
Abstract

MHC class I molecules present antigenic peptides to T cell receptors through a conserved structural scaffold, known as the MHC fold, formed by two semi-parallel α-helices resting on a β-sheet platform. Beyond this canonical role, the same fold has been evolutionarily repurposed to bind lipids (CD1, EPCR, ZAG), small metabolites (MR1), and to mediate protein-protein interactions independent of ligand presentation (FcRn), giving rise to a broad family of MHC-like proteins. Despite decades of family-specific studies, a comprehensive structural characterization spanning this entire functional diversity has been lacking. We present a systematic, high-throughput structural analysis of the MHC fold across classical, non-classical, MHC-like, and viral mimetic families, combining structures mined from the PDB and AFDB via Foldseek with functional annotation and structural clustering. All four major superfamilies of human MHC homologs were represented in the dataset, with no additional MHC fold-containing proteins identified in AFDB beyond those already present in PDB. Structural clustering largely recapitulated evolutionary relationships: classical and non-classical MHC molecules formed adjacent clusters, MHC-like families (CD1, FcRn, MIC, ZAG, ULBP, EPCR) formed distinct, compact clusters reflecting strong intra-family conservation, and viral mimetics clustered with their closest host family rather than forming an independent group, indicating that these mimetics retain the structural features of specific host MHC homologs rather than diverging into a distinct structural class. Groove characterization revealed conserved α-helical content (~35–40%) across most families, with helix-to-helix distances and groove volumes varying by family but only weakly correlated with each other. CD1 stood out with unusually variable groove depth. Interface analysis showed that MHC recognition is achieved through structurally diverse binding motifs: coil-dominated (Ig-like CDRs in TCRs, antibodies, KIRs), helix-dominated (PfEMP1, TfR, albumin), and interfaces with substantial β-strand contribution, observed only in E3 and UL16. The presenting surface is generally enriched in charged and polar residues rather than hydrophobic ones, despite some MHC-like families binding hydrophobic ligands internally. These findings provide the first unified structural framework across the full known diversity of the MHC fold, revealing shared and family-specific recognition principles with implications for immunotherapy and binder engineering. 

Share your ideas or questions with the authors!

Did you know that the greatest stimulus in scientific and cultural development is curiosity? Leave your questions or suggestions to the author!

Sign in to interact

Have a question or suggestion? Share your feedback with the authors!

Institutions
  • 1 Universidade Federal de Pernambuco
  • 2 Brazilian Biosciences National Laboratory (LNBio), Brazilian Center for Research in Energy and Materials (CNPEM), Campinas, SP 13083-970, Brazil
Track
  • 18. Protein Structure and Conformation
Keywords
MHC fold
Fold mining
Structural bioinformatics
Protein-protein interfaces