(164ae) Biosensor Engineering for Reliable Production Detection | AIChE

(164ae) Biosensor Engineering for Reliable Production Detection


Prather, K., Massachusetts Institute of Technology
Metabolic engineering has shown that it is possible to improve microbial cell factories for higher productivities, titers, and yields of useful chemical products. However, there exists a finite knowledge of metabolism, meaning that there is only so much rational improvement to be made to organisms and pathways. To bridge this gap, the use of directed evolution can push both the host organism and the desired enzymatic pathways within these hosts to yield industrially relevant levels. Directed evolution relies on the ability to create large and diverse genetic libraries, and traditional quantification methods fail to achieve the throughput necessary to search through a significant amount of the library diversity to find improved genetic variants. To reach the throughputs needed to access this library sequence space, it is possible to use transcription-factor-based biosensors. These genetic tools allow for the transformation of the relative levels of desired metabolite into a proportional and detectable output, such as fluorescence. Here, we have developed a biosensor for application in the directed evolution of myo-inositol oxygenase for improved titers of glucaric acid from glucose. In response to glucuronic acid, the precursor to glucaric acid and the product of the myo-inositol enzyme, the biosensor outputs GFP, allowing for the use of fluorescence-activated cell sorting to detect cells with improved enzyme variants. In such screens there is a challenge to detect true improved variants over false positives, making a reliable biosensor the key to a successful screening campaign. Thus, to develop this biosensor for application in screening we have focused on a two-pronged tuning of our biosensor characteristics; both developing the sensor in global and specific contexts.

Desirable global characteristics for biosensors are: dynamic range, sensitivity, and operating range of the sensor. We tuned these characteristics by modifying the promoter and verified these changes for in vivo detection of product. The specific characteristics we chose were those that were screen dependent. In a system seeking improved producers, the product transport was a key element to tune. By preventing the importation of desired product while maintaining product export, we were able to obtain a system that was sensitive only to intracellular production. Another screen-specific challenge to the system was the lack of orthogonality between the production and biosensor vectors which impacted fluorescence. By creating a single plasmid system, we were able to see that the fluorescence output was more consistent and reliable for system application. By focusing on both global and specific contexts of the biosensor we were able to create a system that has qualities that increase the likelihood to detect true improved mutants amidst the false positives.