Directed enzyme evolution guided by multidimensional analysis of substrate-activity space

The directed evolution of protein function frequently involves identification of mutants with improved properties from a population of variants obtained by mutagenesis. The selection of clones to parent the subsequent generation is crucial to the continued creation of superior progeny. In the present study, multivariate analysis guided the evolution of human glutathione transferase (GST) T1-1 to 65-fold enhanced alkyltransferase activity. Six alternative substrates monitored the substrate-activity space that characterized a mutant library of enzymes, obtained by recombination of DNA and heterologous expression in Escherichia coli. A subset of mutants was identified by their proximity in the targeted region of six-dimensional factor space. DNA from these mutants was recombined to create a new generation of GST variants from which an improved enzyme was isolated. The multidimensional cluster analysis is applicable to quantitative properties in any population of molecules undergoing evolution and can guide the tailoring of proteins, nucleic acids and other chemical structures to novel and improved functions.