Toward high throughput directed evolution of protease specificity using fluorescence activated cell sorting

Directed evolution by high throughput screening is used for improving protein properties and functions. In our laboratory, a FRET peptide probe was used with a Fluorescence Activated Cell Sorting (FACS)-based screening method for OmpT, a native outer membrane protease of E. coli. This earlier FRET probe enabled the isolation of an OmpT variant with altered specificity, although wild-type Arg-Arg cleavage activity was still present. Newly devised protease probes that operate via electrostatic capture on the bacterial cell surface allowed a very powerful “twocolor” type of FACS sorting, in which new catalytic activities are selected while simultaneously deselecting unwanted activities. Newly designed autoquenching peptide substrates with two identical fluorophores (BODIPY) were used for post screening of new variants on 96 well plates. Using the two-color FACS sorting approach, we were able to isolate OmpT variants 1.2.19 and 1.3.19, displaying a million-fold conversion of specificity (high Ala-Arg cleavage and very low wildtype Arg-Arg cleavage). This study demonstrates that the two-color sorting system for protein evolution can be used for pinpointing protein functions such as protease specificity and scFv antibody. To apply Periplasmic Expression with Cytometric Screening (PECS) for high throughput directed evolution of exogenous proteases, a permeabilization study of the E. coli outer membrane was carried out using FACS with cells expressing periplasmic scFv antibodies. Probes (>1000 Da) hardly gave practical signals for FACS sorting without a permeabilizer. However, outer membrane permeability toward a peptide probe (~ 2800 Da) was significantly enhanced by adding polymixin B nonapeptide (PMBN) or using a “heptoseless” LPS (lipopolysaccharide) E. coli mutant, D21f2. Based on the PECS system, the active form of chymotrypsin, potentially toxic to its hosting cell, was successfully expressed in the periplasm, and its activity was detected using a fluorescent probe and FACS. However, when positive and negative cells were mixed, the FACS signal was not specific to the positive cells. We believe this is due to enzyme leakage to the medium prior to substrate cleavage. Although further FACS screen assays are required to improve signal resolution of positive and negative cells, this work provides the foundation for high throughput directed evolution study of exogenous proteases.