Design of self-organizing peptide chassis materials for synthetic cells by machine learning, molecular modeling, and cell-free protein synthesis.

Synthetic cells are engineered cell-like compartments of interest for molecular encapsulation, reaction, and transport in non-biological environments. The fragility of cell-sized lipid vesicles outside of biological environments has motivated the exploration of alternative 'chassis' materials for synthetic cells. Elastin-like polypeptides (ELP) have been recently shown to form robust vesicular micro-compartments. This has opened up opportunities for the design of novel ELPs from which to self-assemble synthetic cells that can survive in harsher environments than lipid membranes and with greater biocompatibility and biofunctionalization than synthetic polymersome materials. In this study, we combine molecular simulation, machine learning, and cell-free protein synthesis to build predictive models for the discovery and design of ELP sequences and identify a small number of promising candidates for experimental testing.