Novel Fluorescence-Assisted Whole-Cell Assay for Engineering and Characterization of Proteases and Their Substrates

We have developed a sensitive and highly efficient whole-cell methodology for quantitative analysis and screening of protease activity in vivo. The method is based on the ability of a genetically encoded protease to rescue a coexpressed short-lived fluorescent substrate reporter from cytoplasmic degradation and thereby confer increased whole-cell fluorescence in proportion to the protease''s apparent activity in the Escherichia coli cytoplasm. We demonstrated that this system can reveal differences in the efficiency with which tobacco etch virus (TEV) protease processes different substrate peptides. In addition, when analyzing E. coli cells expressing TEV protease variants that differed in terms of their in vivo solubility, cells containing the most-soluble protease variant exhibited the highest fluorescence intensity. Furthermore, flow cytometry screening allowed for enrichment and subsequent identification of an optimal substrate peptide and protease variant from a large excess of cells expressing suboptimal variants (1:100,000). Two rounds of cell sorting resulted in a 69,000-fold enrichment and a 22,000-fold enrichment of the superior substrate peptide and protease variant, respectively. Our approach presents a new promising path forward for high-throughput substrate profiling of proteases, engineering of novel protease variants with desired properties (e.g., altered substrate specificity and improved solubility and activity), and identification of protease inhibitors.Proteases constitute a group of enzymes that irreversibly catalyze the cleavage of peptide bonds and represent approximately 2% of all protein-encoding genes in living organisms (39). Besides acting as virulence factors for many pathogens (16), proteases are crucial for the regulation of numerous biological processes that influence the life and death of a cell (4). These enzymes also underlie several pathological conditions, such as cancer (13) and neurodegenerative (20) and cardiovascular (8) diseases. A key issue for increasing our knowledge about such complex biological processes, and thereby hopefully also providing possibilities for new therapeutic strategies, is to deduce the proteases’ substrate repertoires. Consequently, a lot of efforts around the world are dedicated to the characterization of proteases and their substrates (2, 31). In addition to their biological importance, proteases have attracted much interest in several biotechnological and industrial applications, such as removal of “fusion tags” from recombinant target proteins (38), as supplements in dishwashing and laundry detergents, or for bating of hides and skin in the leather industry (41, 44). Sometimes, however, their use is hindered due to limitations inherent to a specific protease: for example, low solubility, poor enzyme stability and specificity, or limited activity. It would therefore be of great aid to have powerful and straightforward methods available that facilitate the engineering of novel protease variants not suffering from such limitations.Traditionally, protease substrate specificity has been studied by comparison and alignment of naturally occurring substrate peptide sequences (7) or through biochemical analysis of cleavage products with synthetic peptides (47). More recent and powerful methods instead rely on the use of combinatorial substrate libraries, which can either be chemically or biologically generated (6, 15). Although all of these methods have proven useful in determining protease function, many suffer from being laborious and of limited throughput capacity, having an insufficient dynamic range, and resulting in limited information on the substrate profile. Moreover, only a small fraction of all proteases have been studied to date, and there is a need for novel approaches that allow for determination of protease specificity in a rapid, accurate, and quantitative manner.Concerning the engineering of enzymes toward novel desired properties, like altered substrate specificity and improved activity, solubility, and stability; researchers have relied on the use of rational design and/or directed-evolution methods in combination with appropriate screening and selection procedures (1, 10, 11, 22). For instance, various mutagenesis procedures and subsequent screening via assays that report on the successful folding of a protein of interest (9, 32, 45) have been used to engineer protein variants exhibiting improved solubility (35, 37, 46). Despite the obvious success of using such folding reporters in solubility/folding engineering projects, there is a risk that the engineered protein may lose its inherent activity since these screening procedures in general do not select for retained activity but only improved solubility/folding. Therefore, as in the case of a protease, it would be advantageous to establish a screening or selection system that has the ability to simultaneously address traits such as improved folding/solubility without loss of proteolytic activity. However, directed evolution of desired catalytic properties has proven quite a challenge. A popular strategy has been to use phage display technologies, often in combination with transition state analogues (18) or mechanism-based suicide inhibitors, for selection (30). Although successful, the enrichment conferred by these methods is generally based on binding rather than catalysis. Georgiou and coworkers circumvented this potential problem by developing an interesting system that actually enables function-based isolation of novel protease variants from large libraries (34, 42, 43). However, their methodology is dependent on the use of cell surface-displayed proteases, which is not applicable to all proteases and therefore may limit its usefulness.Herein, we present a novel, function-based, and highly efficient fluorescence-assisted whole-cell assay for characterization and engineering of proteases and their cognate substrate peptides. The method takes advantage of genetically encoded short-lived fluorescent substrates that upon coexpression of a substrate-specific protease result in a fluorescence signal, which can easily be monitored on a flow cytometer. Cells having a desired fluorescence profile can then be collected through sorting and sequenced to identify the protease-sensitive substrate peptide or protease capable of processing a particular peptide. Using this approach, we show that it is possible to analyze the efficiency with which the highly sequence-specific tobacco etch virus protease (TEVp) processes different substrate peptides and in model experiments also identify and enrich cells expressing the most favorable substrate peptide or protease from a large background of cells harboring less-efficient variants.