Using computation to identify catalytic features of enzymes and to engineer enzymes with enhanced or altered function

The catalytic properties of enzymes make them useful for wide application, such as in medicine, biofuel production, or sample analysis. However, we are currently restricted to the enzymes that nature has to offer, limiting their applicability. Unfortunately, most initial enzyme design attempts have resulted in very low activities or no activity at all. Long-term directed evolution experiments have increased enzyme activities, but these require years, high cost, and sometimes sophisticated equipment, that limit their development and use. A major issue is the sheer size of mutational space, as there are over 1.0x1048 possibilities for a protein with 200 residues. The goal of our research is to use computation to reduce the number of possibilities, identifying a smaller number of potential designs that are more likely to be successful. The Ondrechen Lab has previously developed the catalytic residue predictors THEMATICS and POOL, which are now being applied to build important chemical properties into designed enzymes. One application of the methods is on the class of de novo designed enzymes called Retro-aldolases, where we have identified residues that contribute to their catalytic abilities. Another case is a natural enzyme, a DNA polymerase from Thermus aquaticus, which already has activity but may not accurately copy damaged DNA. Through combination of our computational methods with others such as docking, we have identified positions to mutate to alter fidelity but still retain catalytic activity. High-throughput assays will be carried out to test our focused variants for activity and fidelity.