(383a) Computational Design of Peptides As Biomarkers, Sensors, Detectors and Drugs

We describe our efforts to develop an efficient computational algorithm that searches for peptides that bind strongly and selectively to specific biomolecular targets, and to use that algorithm in the design of peptide-based sensors, drugs and affinity chromatography ligands. The algorithm is an iterative procedure that involves as many as 50,000 sequence mutation moves and/or peptide backbone conformation moves to arrive at the peptide sequence and conformation that has the lowest binding energy to the target. The top scoring peptides are then further evaluated by performing explicit-solvent atomistic simulations of the peptide–target complex to determine their binding free energies. The biomarker chosen for initial study was Cardiac Troponin I, a 210-amino acid protein that is elevated in the blood of heart attack victims. The top-scoring peptides were synthesized by collaborators at the Air Force Research Laboratory who then used a variety of experimental techniques to assess the binding capabilities of the top candidate peptides. The measured binding affinity of the designed variant, P2, for cTnI was extremely strong (K_D = 0.27 nM) and comparable to that of the natural antibody (K_D = 0.12 nM). We further describe two other projects: (1) design of peptide affinity ligands for bioseparating monoclonal antibody, IgG, from the cell cultures in which they are synthesized, and (2) design of peptides to block the action of the toxins secreted by C- difficile bacteria in the large intestine.