Combinatorial diversification of shared binding residues generates an angulin-1-selective angubindin-1 variant

iota-toxin Ib component and binds the tricellular tight junction proteins angulin-1 and angulin-3, thereby reversibly enhancing paracellular permeability to macromolecules. Because this intrinsic dual specificity limits its utility as a receptor-selective probe or targeted barrier modulator, we sought to uncouple recognition of the two receptors. We previously identified six angubindin-1 residues (i.e. L562, L598, E638, V640, Y643, and K644) that act cooperatively in angulin binding. We therefore examined whether combinatorial diversification of these shared binding determinants could preserve angulin-1 recognition while eliminating angulin-3 recognition. We constructed a phage-display library in which all six positions were randomized and subjected the library to both standard panning and a stringent selection scheme incorporating competition with angulin-3 and wild-type angubindin-1. Next-generation sequencing revealed convergence toward cationic, arginine-enriched variants in which the wild-type acidic and aromatic residues were depleted. Among 34 candidates, R121 showed selective binding to angulin-1, and flow cytometry and surface plasmon resonance confirmed preferential binding to angulin-1 over angulin-3. Functional selectivity was also observed in epithelial EpH4 monolayers: wild-type angubindin-1 reduced transepithelial electrical resistance (TER) in cells expressing either angulin-1 or angulin-3, whereas R121 reduced TER only in angulin-1-expressing cells. In human intestinal epithelial Caco-2 monolayers, R121 likewise reduced TER and increased FITC-dextran (4 kDa) permeability, but both barrier-modulating effects were abolished in angulin-1-knockout cells. These findings provide proof that the receptor selectivity of a multispecific toxin-derived binder can be reprogrammed through combinatorial diversification of a small set of shared binding residues.