A Microfluidics Approach to the High‐Throughput Generation of Quantifiable Cell‐Free Protein Synthesis Systems

Abstract

Synthetic biologists construct parts, devices and systems to engineer cells for applications in medicine, biofuels, chemical commodities, and the environment. A common problem with projects that focus on the synthesis of natural and engineered proteins is unpredictable translation directed by ribosomal binding sites (RBS). Researchers have hypothesized that base pairing interactions between RBSs and downstream coding sequences form secondary structures that affect the efficiency of translation initiation. Bicistronic designs (BCDs) were developed in 2013 by Drew Endy and his colleagues to enable more predictable and reliable in vivo translation. BCDs have an RBS that leads to the production of a leader polypeptide with no functionality and a second RBS that directs the production of a protein of interest. BCDs are better at producing predictable protein levels in bacterial cells than simple RBSs, thereby improving the ability of BCDs to function reliably and predictably as classic synthetic biology parts. In a companion study, we used GFP expression to compare the translational efficiencies of 19 BCDs during cell‐free protein synthesis (CFPS) to their efficiencies in vivo and found a rank correlation of 0.88. This report describes our efforts to expand the toolkit for CFPS protein production to microfluidic droplets. We used a microfluidic flow control system to rapidly produce nanoliter‐scale droplets and characterized them with microscopy. Image analysis of the droplet fluorescence intensities allowed us to quantify and compare BCD‐directed protein synthesis within droplets. We automated this process for many images to accelerate the workflow. We focused on three BCDs for CFPS of GFP in droplets and documented significant increases in observed fluorescence compared to batch CFPS outside of droplets. We also found that the rank order function for the three BCDs was the same in droplets as it was in batch CFPS. Our results support the use of BCDs to gain predictability of protein production in CFPS. These results set the stage for partitioning large libraries of gene regulatory combinations into microfluidic droplets to find the best combinations suited for a given synthetic biology application.Support or Funding InformationNSF RUI MCB‐1613281 to Missouri Western State University, NSF RUI MCB‐1613203 to Davidson College