| De novo protein design addresses the determination of the amino acid sequences that will fold into a given 3-dimensional template structure. The protein design problem exhibits degeneracy due to the fact that many amino acid sequences fold into a given template. It is therefore important to examine all the possible sequences for a given template and rank them based upon specific design properties (activity, specificity, stability, affinity, etc.). This thesis presents a new framework for de novo protein design with a focus on targeting various diseases. The framework contains a novel approximate binding affinity ranking metric that can be used to design peptidic drugs of specific target proteins.;The framework consists of two stages: a sequence selection stage and a validation stage. The sequence selection stage produces a rank-ordered list of amino acid sequences with lowest energies using integer linear optimization. The validation stage re-ranks the sequences using fold specificities or approximate binding affinities. Fold specificity measures how likely a given sequence will fold into the design template structure. Approximate binding affinity determines the likelihood a given sequence will bind to a target protein.;The framework was applied to 14 different proteins that are linked to complement diseases, AIDS, cancer, and autoimmune diseases. Inhibitors and agonists of C3 (human, rat, and mouse), C3a, and C5a (complement diseases), HIV-1 gp41 and gp120 (AIDS), Bcl-2, Bcl-xL, ERK2, EZH2, and LSD1/LSD2 (cancer), and HLA-DR1 and PNP (autoimmune diseases) were computationally designed and many were experimentally validated. Three inhibitors of C3, two inhibitors of C3a, two agonists of C3a, four inhibitors of gp41, ten inhibitors of EZH2, 17 inhibitors of LSD1, and 18 inhibitors of LSD2 were experimentally verified to be potent inhibitors/agonists and many were the most potent discovered thus far. The experimental validation of five of the protein systems demonstrates the success of the de novo protein design framework and highlights its broad applicability in drug discovery. |
