Step‐Wise Directed Evolution of an Epoxide Hydrolase Against Progressively Larger Non‐Natural Substrates

Trichothecenes are a class of mycotoxins that function as inhibitors of eukaryotic protein synthesis. Deoxynivalenol (DON), a.k.a. “vomitoxin”, is a trichothecene that contaminates crops and is stable enough to withstand high temperature and concussive pressure as a chemical weapon. Therefore it poses a health risk to both livestock in its natural state and to humans as a chemical weapon. The epoxide portion of trichothecenes is responsible for biological activity and de‐epoxidation leads to detoxification. We are in the process of using directed evolution to re‐engineer an epoxidase hydrolase to catalyze the epoxide ring opening reaction that will neutralize the DON mycotoxin. The directed evolution process will be greatly enhanced through the use of a fluorescence‐based Bacterial Surface Display (BSD) system used to display the C ystic Fibrosis Transmembrane Conductance Regulator I nhibitory F actor, or Cif. The Cif enzyme is a well‐characterized epoxide hydrolase that is secreted by Pseudomonas aeruginosa , it exhibits robust folding properties, and its structure has been solved to high‐resolution. The BSD system consists of a series of fused protein elements that 1) direct the fusion protein to the E. coli outer membrane, 2) contain a fluorescent protein (mCherry) for expression optimization, and 3) a transmembrane protein that functions to anchor the fusion protein in the E. coli outer membrane. A number of proteins and enzymes have already been expressed and successfully characterized using this display system. In fact, we have already used the system to express wild‐type Cif on the surface of E. coli . Catalytic analysis confirmed that BSD Cif is just as active against a standard substrate ( i.e., epibromohydrin) as soluble Cif. In silico evolution (machine learning) is being used to reduce and focus the sequence search space needed for step‐wise in vivo directed evolution. The results of the in silico re‐design of the Cif active‐site (select mutations) will be incorporated into the BSD system. Instead of attempting to re‐design Cif to hydrolyze the DON epoxide in just one step, we plan to re‐design its active‐site against increasing larger synthetic substrates that contain a fluorescent moiety. Fluorescence‐activated cell sorting (FACS) will be tested to identify successful mutants. In other cases the synthetic substrates will be linked to biotin and used to capture E. coli on immobilized streptavidin beads. These Cif mutants will incorporate an active‐site mutation (H297F) that results in the formation of a covalent bond between the enzyme and synthetic substrate. The captured variants will be sequenced and identified mutations will be incorporated into the Cif variant that contains wild‐type catalytic residues. In subsequent steps, mutations designed to enlarge the Cif active‐site will be incorporated into the BSD libraries, as well as mutations that are driven via random mutagenesis methods. All resulting libraries will be incorporated into the BSD system and incubated against synthetic substrates. These methods will be employed in multiple rounds of computational design and directed evolution against progressively larger epoxide substrates. The goal is to increase the volume of the active‐site in a step‐wise fashion and ultimately generate a Cif variant that effectively degrades deoxynivalenol.