Computational structural analysis sheds light on the molecular grammar of gap junction proteins.

Gap junctions are multimeric intercellular channels that permit ions and small molecules to pass directly from one cell to another. Despite being fundamental to multicellular life, these channels are formed by distantly related protein families: innexins in invertebrates and connexins in vertebrates. Vertebrates also express pannexins, distant homologues of primordial innexins, which only form hemichannels. While these families have diverse sequences and different oligomeric states, their monomeric structures are highly similar. We generated structure-guided sequence alignments to establish equivalent residues across innexins, connexins, and pannexins. Further, computational approaches for determining protein-protein interaction hotspots, residue conservation, accessible surface area and local conformations of residues, provide insights into the relationships between residue positions and channel properties. We find that exposed transmembrane residues of TM1 and TM2 are more conserved than those in TM3 and TM4, especially in connexins and pannexins. Moreover, we see that residues in the extracellular extended hairpins of pannexins show more conformational flexibility, in variable protein blocks, than equivalent residues in connexins. This hints that the rigidity of this element could be a prerequisite for hemichannel docking. Finally, we identify inter- and intra-hemichannel interface hotspots that are positionally conserved across the families, implying their role in hemichannel and, ultimately, gap junction formation. Such analyses reveal a molecular grammar that underlies gap junction design and offer a basis for targeted perturbation of channel properties.