UC San Diego

Genome-scale technologies have transformed our understanding of the biomolecular signaling networks that underpin cellular function and adaptation. Omics-level analysis has cemented the view that biological signal processing is not the result of linear pathways but an emergent property of complex networks whose functions and dynamics we now seek to understand. In model organisms such as E. coli, biomolecular networks are often elucidated by observing how gene expression patterns change in reaction to experimentally-induced perturbations. However, the high-throughput experimental techniques traditionally used for this purpose are inherently destructive and only offer snapshots of a cell's state. As such, these technologies do not fully capture the information encoded in the dynamics of biomolecular networks, which are complex, time-dependent signals.

In the past twenty years, microfluidic technology combined with fluorescence microscopy has established itself as a powerful tool to study time-dependent biological processes while precisely controlling the cellular environment. This thesis focuses on bridging the gap between genome-wide assays and microfluidics-based dynamic perturbation experiments. Here I report the development of a high-throughput microfluidic platform capable of culturing 2176 unique microbial microcolonies in parallel and monitoring the changes in expression of fluorescent proteins in each strain. By loading the platform with some of the readily available libraries of fluorescent transcriptional reporters and dynamically tuning the growth media, I show that we can measure microbial gene expression dynamics in response to environmental inputs in vivo and genome-wide.

Chapter 1 provides an overview of the role of high-throughput microfluidics in systems and synthetic biology research. Chapter 2 describes the design of a highly multiplexed microfluidic platform for monitoring gene expression in GFP-tagged E. coli with both industrial and research applications. Chapter 3 illustrates the platform's applicability as an environmental biosensor that uses the dynamics of 2000 E. coli GFP-promoter strains coupled with machine learning algorithms to detect the presence of heavy metals in drinking water in real-time. Chapter 4 further demonstrates the potential of microfluidics-based biosensing by reporting the use of devices loaded with diverse engineered microbes to detect pollutants in seawater. Finally, in Chapter 5, I use the platform to probe the dynamics of the S. cerevisiae proteome in response to the drug metformin and lay the foundations for a new type of dynamics-based chemogenetic screen. The overarching aim of this research is the capture of microbial gene expression dynamics in response to environmental stimuli on a genome-wide scale with applications in biosensing and the characterization of drug targets.