Computational design of protein crystals: Controlling macroscopic molecular orientation through protein engineering

Computational protein design algorithms rapidly identify amino acid sequences consistent with a target protein fold. Given protein function is dictated by fold, through protein design, it is possible to engineer protein functionality, structure and assembly. Despite many advances in protein design methodology, the de novo design of a protein crystal lattice has yet to be conducted. Protein crystallography is the centerpiece of structural biology, however realization of diffraction quality crystals can be difficult to almost impossible in some cases. In addition to an inherent importance in structure determination, protein crystallization has the potential to be leveraged into materials with well-defined order. As a result, ab initio protein crystal design has the potential to alleviate difficulties associated with protein crystallization and develop new nanostructured materials. In this thesis a computational energy landscape approach for designing protein crystals was developed and experimentally tested. A three-helix coiled-coil protein was designed de novo to form a polar three-dimension crystal consistent with the honey-comb P6 space group. A high-resolution x-ray structure of the designed protein crystal revealed remarkable precision compared to the computational model. The approach has been expanded to more complex single-chain four-helix bundle protein, that has been previously designed to selectively bind a non-biological, non-linear optical cofactor. The ultimate goal of the project is to crystal this protein into a highly ordered, polar space group. Subsequently, orientating the cofactor in a high-density, polar, three-dimensional arrangement and achieving a nonlinear optical biomaterial