THE PROTEIN REPAIR SHOP: HIGHLY EXPRESSED AND STABLE PROTEIN VARIANTS BY COMPUTATIONAL PROTEIN DESIGN

Abstract: Upon over expression, many proteins suffer from misfolding and aggregation thus resulting in low yields of functional protein, hampering their use in research, biotechnology and pharmaceutics. Human Acetylcholinesterase (hAChE), a key enzyme in neuro-signaling and the target of nerve agents, represents a typical case of a human protein that can only be expressed in very low yields in bacterial strains, thus forcing the use of mammalian expression systems. In this talk I will present a novel, fully automated computational strategy to design stable protein variants. Using this method, we designed an hAChE variant encoding 51 substitutions relative to wild type, including improvements in core packing, surface polarity, and loop rigidity; this variant expressed with 250-fold higher yields in E. coli than the w/t protein without sacrificing catalytic activity. We demonstrate the generality of our approach by applying it to three other challenging enzymes, in each case enhancing the yields of soluble and active protein in E. coli.