Accelerating transcription factor-based biosensor development using cell-free gene expression
Allosteric transcription factor biosensors (TFBs) have shown great promise as synthetic biology tools and molecular diagnostics for environmental (e.g., pesticides, heavy metals, and fluoride) and human health (e.g., hormone) targets. Cell-free biosensing systems offer a particularly exciting possibility for on-demand detection, as cell-free biosensors require minimal equipment and can be freeze-dried for non-cold chain transportation. Unfortunately, TFBs often require engineering to achieve the necessary sensitivity, selectivity, dynamic range, and/or kinetics for application. Additionally, their allosteric nature limits the effectiveness of rational protein engineering methods. In my PhD, I first focused on developing a CFE screening workflow to engineer TFBs with the characteristics needed for application. The open reaction environment and scalability of CFE systems enabled rapid generation of positive and negative sequence-function data for hundreds of mutants at once in different ligand conditions. Then, as machine learning (ML)-assisted directed evolution methods have emerged to decrease experimental burden and accelerate exploration of protein sequence space, I integrated a ML model into my workflow to accelerate TFB development and tune multiple biosensor parameters simultaneously. In this dissertation, I describe my efforts to use the CFE screening workflow to engineer transcription factor PbrR to be a highly sensitive and selective point-of-use diagnostic for lead (Pb²⁺) contamination in water. Then, using this work as a foundation, I describe future directions that could be valuable for developing TFBs. Altogether, the innovations in this work will accelerate future TFB engineering and expand the availability of biosensors with desired diagnostic capabilities
硕士论文、博士论文, English, 2025
[Stanford University], [Stanford, California], 2025