Direct selection for sequences encoding proteases of known specificity

Abstract

We have developed a simple genetic selection that could be used to isolate eukaryotic cDNAs encoding proteases that cleave within a defined amino acid sequence. The selection was developed by using the transcription factor GAL4 from Saccharomyces cerevisiae as a selectable marker, a cloned protease from tobacco etch virus (TEV), and an 18-amino acid TEV protease target sequence. In yeast, TEV protease cleaves its target even when the target is fused to internal regions of the GAL4 protein. This cleavage separates the DNA binding domain from the transcription activation domain of GAL4, rendering it transcriptionally inactive. The proteolytic cleavage can be detected phenotypically by the inability of cells to metabolize galactose. Cells expressing the TEV protease can also be selected on the suicide substrate 2-deoxygalactose. DNA binding studies show that the TEV protease decreases the activity of the GAL4/target fusion protein. Because another protease target sequence of 55 amino acids can be inserted into GAL4 without any loss of transcriptional activity, this assay offers the opportunity to use high-efficiency cDNA cloning and expression vectors to select coding sequences of other proteases from various species. The assay could also be used to help define both target specificities and functional domains of proteases.