Early efforts to genetically engineer biological systems were restricted to relatively crude cut and paste operations, with the sequence of even model organisms' genomes inaccessible for detailed study. Exponential improvement in our ability to both read and write DNA sequences has dramatically altered the landscape of genetic engineering, rendering de novo synthesis of entire genomes possible. However, our ability to design functional systems has not kept pace with our ability to fabricate them. The field of synthetic biology seeks to address this shortcoming and provide a theoretical framework for rationally manipulating biological systems. Of particular interest is exploration of how core engineering principles such as modularity, abstraction, and standardization can be applied to the development of genetic systems. Despite considerable recent progress toward answering that question, imperfect and incomplete knowledge about biological systems necessitates exploration of numerous variants to craft even relatively simple systems. In this work, we first explore development of biological systems that make engineering biology easier, then pursue development of a complex system for a demanding application to probe the limits of our design capability, and conclude with a discussion of key challenges in designing biological systems and how to address them.
Motivated by the observation that tuning the expression of system components is often essential to achieving desired functions, we begin by describing development of a system for rapidly prototyping genetic constructs to determine an optimal expression level. Although several methods exist for manipulating the transcription or translation of a desired gene, they require fabrication of numerous distinct variants. To work around this limitation, we instead explore manipulation of plasmid copy number and develop a set of strains that are capable of maintaining the same plasmid at a desired level in the range of 1-250 copies per cell. We demonstrate that this system can be applied to rapidly find the optimal expression level for a model biosynthetic pathway, regardless of the transcriptional activity of the pathway. Recognizing that fabricating and assaying distinct variants is in some cases essential, we then examined the problem of enabling high-throughput manipulation of DNA. Existing methods are limited by difficult to automate operations, such as centrifugation, application of a vacuum, or gel electrophoresis. Development of alternative methods that only require liquid handling operations would enable utilization of very high throughput automation platforms. We therefore engineer a P1 phage-based system for transferring plasmids between cells using only liquid handling operations. After establishing exogenous control of phage lysis through addition of a small molecule inducer, we incorporate phage cis elements that enable desired DNA to be transferred 1600-fold more efficiently than other cellular DNA. The system is capable of transferring large libraries of DNA up to 25 kilobases in length at small volumes. This provides an important first step toward development of a highly automatable suite of tools for biological engineering.
Biological engineers ultimately want to manipulate biological systems to solve human specified problems. Although a number of potential applications of synthetic biology have been described, most projects rely on relatively simple designs to achieve a desired outcome. To identify bottlenecks in the development of more complex systems, we investigate engineering a laboratory strain of E. coli to localize to, invade, and kill cancer cells. After attempting to transfer capsular polysaccharide synthesis clusters to a laboratory strain to improve bacterial survival in an animal model, we generate a strain capable of delivering a cytotoxic ribonuclease to and subsequently killing cultured cancer cells. Although the strain did not inhibit tumor growth in an animal model, the lessons learned during the execution of this and other projects guided our thinking about current bottlenecks and the path forward. We conclude with a discussion of the potential for modular design to enable routine development of complex biological systems, and identify six failure modes that hinder current efforts. By architecting software to codify existing biological knowledge and investigating the basis of important phenomena such as load and stress, we can harness the power of improved DNA sequencing and synthesis capabilities to build novel, useful biological systems.