Design and characterization of defined alpha-helix mini-proteins with intrinsic cell permeability

Proteins with intrinsic cell permeability that can access intracellular targets represent a promising strategy for novel drug development; however, a general design principle is still lacking. Here, we established a library of 46,678 de novo-designed mini-proteins and performed cell permeability screening via phage display. Analyses revealed a characteristic neighboring distribution of positive charges across helices among enriched mini-proteins of CPP7, CPP11, CPP55, CPP109 and CPP112. Compared with the state-of-the-art cell-penetrating mini-protein ZF5.3, the optimized mini-protein CPP11D36R exhibited a sevenfold increase in cell permeability. Endocytosis uptake and early endosome release are the key penetrating mechanisms. A machine learning model with high-throughput data achieved an F1 score of 0.41, significantly outperforming the previously reported CPP prediction models, including MLACP, CPPpred and CellPPD, by 41 %. Overall, our findings validate the effectiveness of a helical structure with a cationic distribution as a design principle on a large scale and present a robust approach for the development of cell-permeable mini-protein drugs.