Rosetta FunFolDes - A general framework for the computational design of functional proteins

摘要

Author summary The ability to use computational tools to manipulate the structure and function of proteins has the potential to impact many facets of fundamental and translational science. Due to our limited understanding of the principles that govern protein function and structure, the computational design of functional proteins remains challenging. We developed a computational protocol (Rosetta FunFolDes) to facilitate the insertion of functional motifs into heterologous proteins. We performed extensive in silico benchmarks, and found that when the design of function is required the global energy minima may not be the optimal solution, in line with previously reported experimental studies. Further, we used FunFolDes to design two novel functional proteins, displaying two viral epitopes that can be of interest for vaccine development. The designed proteins were experimentally characterized, showing that functionalization was successfully achieved. These results highlight the capability of FunFolDes to address common challenges on the design of functional proteins. In particular, the reduced structural compatibility between functional sites and host scaffolds, effectively enabling the repurposing of old protein folds for new functions. Overall, FunFolDes provides new means to accomplish the challenging task of functionalizing computationally designed proteins.