Binding sites as dynamic hubs of coevolution encoded information flow

The relationship between evolution and allostery is inherently intertwined. While structural studies have elucidated allosteric mechanisms, the extent to which evolutionary sequence information encodes these dynamic networks remains underexplored. Here, we introduce Sequence-based GNM-TE, a framework that integrates coevolutionary information with Gaussian Network Model (GNM) based Transfer Entropy (TE) calculations. This approach translates evolutionary constraints, shaped by both structural and functional requirements, into functionally meaningful allosteric dynamic information. Across 29 cases, we identify binding sites as dynamic hubs of collective information transfer and reveal mode-dependent, directional information flows between these sites. These findings overall indicate that evolutionary constraints embed the architecture of allosteric regulation within protein sequences, preserving binding site communication within layered dynamic ensembles that enable robust functional control and adaptation.