Biosensing for Multiplexed Genome Engineering: Applications in Renewable Chemical Production

Abstract

Engineered biological systems are increasingly used to produce fuels, pharmaceuticals and industrial chemicals. While transforming cells into renewable chemical factories presents an enormous opportunity, development timelines are long, costly and often uncertain. Engineering microbes for chemical production is accomplished through the biological design-build-test cycle: many designs are formulated, the corresponding organisms are constructed, and their ability to produce the desired chemical is evaluated. Designs that perform well become the starting point for the next round of the cycle. Faster design cycles result in shorter and less costly product development timelines. Advances in DNA sequencing, synthesis and genome engineering technologies have sped up the design and build steps of the design cycle by enabling billions of organism variants to be designed and constructed simultaneously. However, evaluation of the resulting designs continues to rely on low-throughput technologies with evaluation rates on the order of thousands per day. Because the engineering process is a cycle, it can only proceed at the rate of the slowest step. A high-throughput method for design evaluation would increase the throughput of the design cycle by up to a million-fold. This thesis describes an engineering framework that makes high-throughput design evaluation a reality. By programming cells to keep track of their own success in making a desired product, I enable screens and selections to be used for the optimization of metabolic pathways. I develop biosensors that maintain gene expression at a rate proportional to the concentration of several different chemical products and show that higher product concentration results in a higher fluorescent output. I then construct metabolic pathways for the production of the renewable plastic precursors 3-hydroxypropionate, acrylate, glucarate and muconate. I combine each pathway with the appropriate biosensor and use fluorescence to observe product formation in real-time. Next, I replace the fluorescent protein with an antibiotic resistance gene and link the level of product formation to the cell’s ability to survive an antibiotic challenge. I deploy the selection to optimize production of both glucarate and naringenin from glucose. I further develop the characterization of these new biosensors to promote their use as genetic switches for synthetic biological circuits. Finally, I develop a device called the fluorimostat that makes long-term closed-loop programmable control of gene expression a reality.