Thesis title: Genetic controllers address noise-associated limitations of synthetic gene networks in E. coli
This thesis focuses on the design, building and testing of controllers for synthetic gene networks in bacterium E. coli, to address noise-associated limitations of such systems. Synthetic gene networks, also known as genetic circuits, can be implemented into bacterial cells in the form of genetic material (usually circular DNA molecules) programming the gene expression and behavior of host cells. Applications of genetic circuits in bacteria range from energy and metabolite production to the manipulation of microbial consortia and cancer therapy. However, stochasticity at both the single-cell and the population level affects the reliability of synthetic gene networks, sometimes hindering the purpose they are designed for. In this thesis, we applied rational bioengineering principles to integrate into underlying synthetic networks additional genetic parts, acting as designed external controllers. The controllers aimed at reducing the impact of noise in two genetic circuits in E. coli, with applications in the fields of non-linear physics and cancer therapy, respectively.
The first system is the repressilator, a renowned genetic circuit first presented in a seminal paper from 2000. It demonstrated the first experimental realization of a synthetic gene network designed from the bottom-up to produce periodic patterns of gene expression in individual bacteria. However, phase drifts and loss of synchronization are well-known noise-associated limitations of the repressilator, caused by the stochasticity of gene expression. We addressed these limitations by integrating an optogenetic module enabling to reset, delay, or advance the phase of the underlying oscillator using optical inputs. Our optically controllable genetic clock was the first experimental realization of a synthetic gene network to achieve long-term globally synchronized states through entrainment with a light cycle, with a mechanism reminiscent of natural circadian clocks. The high spatiotemporal resolution of optical stimulation and the simple structure of our network make it an optimal system to further investigate the non-linear dynamics of entrainment, phase adjustment and detuning in oscillating transcriptional networks.
The second system I worked on falls within the field of engineered bacteria cancer therapy. The field aims at exploiting genetic circuits to maximize the potential of bacteria to deliver therapeutic payloads directly to tumors, with precise control over therapeutic release in space and time. We characterized non-pathogenic E. coli expressing the bacterial toxin Perfringolysin O (PFO), a potent cancer cell cytotoxin, and presented experimental evidence that expression of PFO causes lysis of bacteria in both batch culture and microfluidic systems, facilitating its efficient release. However, we demonstrated that a major source of noise hindering the efficacy of the therapy is the emergence of a non-lysing mutant population that limits therapeutic-releasing bacteria in a PFO-expression level dependent manner. We addressed this limitation by integrating a chemically controllable genetic module enabling to regulate the rate and duration of pfo expression. We developed a mathematical model describing the evolution of therapeutic-releasing and mutant bacteria populations revealing trade-offs between therapeutic load delivered and fraction of mutants that arise. Altogether, we demonstrated how dynamic modulation of gene expression through external controllers can address mutant takeovers giving rise to limitations in engineered bacteria for therapeutic applications.
The two systems described above resulted in published papers on the journals eLife and SPJ BioDesign Research, respectively. In the last part of the thesis, we discuss the future perspectives of this work and present the design and characterization of additional controllers for lysis in E. coli. Notably, we describe an optogenetic controller that allows the fine-tuning of lysis in time and space, in response to three orthogonal wavelengths of light. We will end the thesis by addressing another question that the reader of this thesis might have: how can bacterial therapies evolve, going from bench to bedside? We will suggest an answer to this question by reviewing further developments of bacterial therapies: on the methodological side, reporting the construction of a microfluidic platform for the co-culture of engineered bacteria with human cancer cells, to evaluate the pharmacokinetics of bacterial therapies in vitro; and on the applicative side, introducing bacteria-based immunotherapies for cancer, another promising class of therapeutics whose e fficacy could be enhanced by designed external controllers programming dynamic expression of the therapeutics.
Altogether, this thesis demonstrates that rationally-designed genetic controllers are an effective tool to counteract the deleterious impact of noise affecting synthetic gene networks in E. coli, which hinders the development of safer, more reliable and reproducible gene-based systems for biomedical applications and beyond.