Although it is now possible to create new enzyme active sites with immunological methods or by redesigning existing proteins, the chemical efficiency of these catalysts is typically considerably lower than that of naturally occurring enzymes. Genetic selection is a potentially general method for evolving the properties of these first-generation molecules. To test this notion, we have engineered strains of Saccharomyces cerevisiae and Escherichia coli that lack the genes for chorismate mutase, an enzyme that is essential for the biosynthesis of the aromatic amino acids tryosine and phenylalanine (Scheme 3). We have shown that antibody 1F7, which possesses modest chorismate mutase activity, is able to complement the metabolic defect of the chorismate mutase-deficient yeast strain when induced at high levels in the cytoplasm. This result demonstrates the feasibility of using catalytic antibodies in vivo to effect vital biochemical reactions and establishes the growth selection assay needed to improve the properties of the abzyme. We are now exploiting multiple rounds of mutagenesis and selection to identify versions of the antibody with enhanced catalytic efficiency.
Scheme 1: The shikimic acid pathway.
In parallel with our experiments in yeast, we are using combinatorial mutagenesis and selection to probe the mechanism and structure of natural chorismate mutases. For example, we have successfully examined electrostatic effects in the Bacillus subtilis enzyme (BsCM) in our bacterial complementation system. Our results show that stabilization of charge separation in the transition state by a cation in the vicinity of Arg90 is a critical feature of catalysis (see, for example, Fig. 2). We have similarly investigated the role of interhelical turns in the all a-helical AroQ type of chorismate mutase and have used directed evolution to rationally alter enzyme topology. Genetic selection in living organisms thus represents a powerful tool for analyzing and (re)designing protein structure and function which may help bring us a step closer to our goal of creating enzyme-like catalysts for regulating cellular function, altering metabolism, and destroying toxins in a rational fashion.
Fig. 1: Schematic view of the active site of BsCM (left) and growth phenotypes of active clones from randomized BsCM libraries (right). Growth rate on selective agar plates lacking tyrosine and phenylalanine is ranked from 0 (no growth) to 5 (wild-type levels). The results shown are for randomized position 90 alone and in combination with position 88.
Biocatalysis, like many of today's most interesting scientific questions, demands study from an integrated, multidisciplinary vantage point. To that end, our group at the ETH-Zürich is applying the tools of chemistry, immunology, molecular biology and genetics to the design, preparation and study of new enzymes. We anticipate that effective implementation of these technologies will lead to a general understanding of how enzymes work and evolve, how protein function is related to structure, and how entirely new enzymatic activities can be created for use in research, industry and medicine.