Computational design of soluble analogues of integral membrane protein structures

De novo design of complex protein folds using solely computational means remains a significant challenge. Here, we use a robust deep learning pipeline to design complex folds and soluble analogues of integral membrane proteins. Unique membrane topologies, such as the GPCRs, are not found in the soluble proteome and we demonstrate that their structural features can be recapitulated in solution. Biophysical analyses reveal high thermal stability of the designs and experimental structures show remarkable design accuracy. Transplantation of native structural motifs demonstrates their potential for further functionalization as novel small molecule binders and new approaches to drug discovery. In summary, our workflow enables the design of complex protein topologies with high experimental success rates and low sequence similarity to natural proteins, leading to a de facto expansion of the soluble fold space.