Designed β-hairpin switches for controllable mechanical properties

Nature achieves extraordinary mechanical performance by precisely regulating β-structure formation in proteins such as fibroin, elastin, and resilin. Replicating this level of structural control remains a major challenge in protein engineering. Here, we integrate biomimetic design with deep learning–guided de novo protein engineering to create environmentally responsive β-hairpin peptides. Computational optimization enhanced β-hairpin propensity, hydrophilicity, and solvent accessibility while preserving high aqueous solubility. The peptides remain intrinsically disordered in solution but rapidly undergo a β-hairpin transition upon exposure to minimal concentrations of sodium dodecyl sulfate (SDS), a model amphiphilic trigger. This structural conversion drives assembly into mechanically reinforced materials exhibiting increased stiffness and hardness relative to the unfolded state. Our findings provide mechanistic insight into regulated β-structure formation and demonstrate a scalable strategy for programming environmentally triggered protein folding, hierarchical assembly, and mechanical function, opening new opportunities for the rational design of next-generation adaptive biomaterials.