The intrinsically disordered oncoprotein c-MYC is a master regulator of transcription whose activity is mediated by six conserved sequence motifs known as MYC boxes (MBs). These short linear motifs, embedded within disordered regions, act as modular hubs for protein-protein interactions and enable c-MYC to recognize a broad spectrum of binding partners. Here, we combined large-scale sequence analysis, machine learning-based structure predictions, and molecular dynamics simulations to investigate how individual MYC boxes encode interaction specificity. Our analyses revealed distinct chemical signatures across MBs: MB0 is enriched in acidic and aromatic residues; MBI is acidic and proline rich; MBII and MBIIIA are predominantly hydrophobic; MBIIIB is strongly acidic; and MBIV is enriched in basic residues. To elucidate structural determinants of recognition, we modelled MB regions in complex with experimentally reported partners. These models revealed that conserved residues facilitate interactions by forming persistent hydrogen bonds, salt bridges, or hydrophobic contacts, while flanking disordered segments remain conformationally dynamic. Simulations further showed that sequence composition biases the conformational ensembles sampled by each MB, effectively predisposing them toward partner compatible states. Together, these findings delineate mechanistic principles by which chemically diverse MYC boxes achieve selective partner recognition despite intrinsic disorder, and highlight potential avenues that can be leveraged for therapeutic intervention in cancer.
Krishnakumar et al. (2026) studied this question.