The focus of this proposal is to develop automated design techniques for the structure-based design and engineering of function in proteins. We will accomplish this by the introduction of new ligand-binding sites into proteins of known structure, and the redesign of the properties of existing ligand-binding sites. We propose to extend our previous work in which we have successfully used a structure-based computational approach, the DEZYMER program, to build metal centers with a variety of different activities in the hydrophobic core of a model protein. Here we propose to develop novel computational techniques that introduce further important structural degrees of freedom into the design algorithm. We also propose to develop novel algorithms for the structure-based redesign of ligand binding sites. This procedures will be tested experimentally by optimizing the stabilities and reactivities of a series of rationally designed metal centers; by structure-based design of metalloenzyme active sites in several different protein scaffolds, including thioredoxin, fluorescent green protein and fatty acid binding protein; and by the structure-based redesign of the specificity of a bacterial receptor protein. Central to this approach is the design cycle in which modeling and experimental analysis alternate, thereby iteratively testing and improving both the designs and the computational methods. We propose to extend this approach by introducing a directed evolution stage to improve the designs by experimental selection and screening methods. The development of methodologies for structure-based redesign of protein structure and function have important industrial and biomedical applications such as the development of new catalysts, biosensors, novel biomaterials, and protein-based pharmaceuticals.
| Benson, David E; Haddy, Alice E; Hellinga, Homme W (2002) Converting a maltose receptor into a nascent binuclear copper oxygenase by computational design. Biochemistry 41:3262-9 |
| Looger, L L; Hellinga, H W (2001) Generalized dead-end elimination algorithms make large-scale protein side-chain structure prediction tractable: implications for protein design and structural genomics. J Mol Biol 307:429-45 |
