(1224-A) Automated Directed Evolution: A PicoShell Approach

Abstract: Directed evolution is a powerful approach for enhancing the function of proteins and developing novel biomolecules. Unfortunately, current methods to screen large libraries of proteins based on complex functions face bottlenecks in automation. While microfluidic microwell and droplet technologies have increased throughputs to hundreds of thousands of variants, the inability to easily swap out solutions to reversibly measure biomolecular function in complex environments limits functional testing. A new screening platform that both enables high-throughputs and is capable of easily changing the environment is necessary to accelerate the development of biomolecules with enhanced functions.

To overcome these limitations, we have developed novel PicoShell technology that allows scientists to screen millions of variants, easily swap out the soluble environment the biomolecules are exposed to, and place them in conditions that more readily model behavior in the environment for their eventual use. PicoShells are spherical microcompartments with a hollow center and porous polyethylene glycol (PEG) outer shell that act as replacements for wells in a microtiter plate. The pore size of the outer shell can be tuned to have a 15 kDa molecular weight cutoff (MWCO) such that most proteins and DNA material encoding for each protein can be retained within the shell while still allowing diffusion of desired analytes in solutions.

PicoShells can be generated at a rate of 2000 particles per second and are compatible with screening of hundreds of thousands of variants per day using standard fluorescent-activated cell sorters (FACS). Using standard FACs instrumentation we have demonstrated that PicoShells can be used to select protein-based calcium biosensors (GCaMP) expressed within clonal colonies of E. coli. We performed a screen of GCaMP protein variants exhibiting the highest fluorescence in calcium-rich media and the lowest fluorescence in calcium-depleted media to select for the dynamic range of the sensor. This effectively automates colony-picking processes followed by manual well-plate-based functional assays.

However, taking advantage of this throughput requires streamlined automated processes for downstream sequencing. Coupling PicoShell-FACS index sorting into well plates with an automated lab bench and Tecan liquid handling, we can effectively perform automated molecular labeling and sequencing from the highest-performing variants in a given library. Here we demonstrate an automated process to selectively label DNA fragments from screened variants via PCR and perform pooled sequencing using next-generation sequencing (NGS). The continuous capabilities of our lab automation workflow enable a range of PicoShell screening and sorting approaches to ensure an entire workflow that maintains high throughput for the directed evolution of industrially and scientifically useful proteins.

Overall, PicoShell technology coupled with lab automation can significantly accelerate protein engineering and increase the range of selection criteria that can be used to guide automated directed evolution efforts for novel biomolecule discovery.