Design and debugging of ultrastable engineered genetic systems

Abstract

Engineered genetic systems in bacteria have tremendous potential for biotech applications ranging from living therapeutics to the controlled production of chemicals. Engineering such genetic systems is challenging as these genetic systems often consist of multiple genes and genetic parts (>4 genes or > 45 genetic parts) interacting with each other as intertwined networks. These intertwined networks are invisible, making the design and debugging of these genetic systems to be particularly challenging. Additionally, expressing a large number of genes creates burdens on the host cell and reduces the long-term stability of these genetic systems. Here we address these two problems by (i) adapting high-throughput RNA-seq to visualize the inner-workings of these engineered genetic systems and (ii) developing a robust and efficient genome engineering platform that enables the implementation of long-term stable engineered genetic systems on the genome.

First, we applied a high-throughput RNA-sequencing, RNA tag-seq, to analyze the behavior of engineered genetic systems. We analyzed two systems with RNA-seq: (i) a library of 84 refactored nitrogenase clusters where each cluster consists of six genes with varying levels of expression and (ii) a genetic circuit that consists of eight interacting genes. With this analysis, we studied the design parameters for these genetic systems and identified various unexpected failure modes. Swapping a troubling genetic part in RNA-seq profile allowed us to effectively debug unwanted circuit expression profiles. To reduce the cellular burden from expressing these genetic systems, we developed a reliable and efficient genome engineering platform on the E. coli MG1655 K-12 genome. We built three genome landing pads, each of which consists of an att (phage attachment sites) site insulated with ultra-strong bidirectional terminators.

Landing pads locations were determined by Tn5 transposon library screening by finding genomic locations that showed high gene expression levels without interfering endogenous gene expression. We also developed a set of plasmids that integrates genetic circuits into these landing pads via simple transformation. With these landing pads, seven orthogonal sensors and eight orthogonal TetR-homolog NOT gates were engineered on the genome to have up to 640-fold changes in output promoter activity upon induction. Utilizing these sensors and gates, we successfully implemented 3-input genome circuits that are stably maintained without antibiotics for more than two weeks in rich media with continuous daily ON/OFF state cycling. We expect this platform could facilitate the design and debugging of long-term stable engineered genetic systems.