Development of tunable synthetic circuits for precise control of gene expression in mammalian cells

Abstract

Precise control of gene expression is not only crucial for the study of gene functions but also pivotal to the development of sophisticated synthetic gene circuits. A few inducible expression systems for the quantitative control of gene expression in mammalian cells have been reported. However, it is still a challenge to develop synthetic systems with good tunability and predictability. The ability to modify the characteristics of the response, attenuate cell-to-cell heterogeneity (gene expression noise), and adjust output value through multi-input signals is essential for the development of synthetic circuits with good controllability. Moreover, quantitative study involving experimental characterization and computational modeling would be a promising way to overcome the current challenges.

In an attempt to construct reliable gene expression circuits for precise control of gene of interest, I first used experimental approaches as well as computational modeling to investigate how to tune the response curve and harness gene expression noise, quantitatively. Next, I established multi-signal processing modules, which allow us to control output gene expression values via various input signals, including chemical molecules and blue light.

Specifically, I constructed and optimized tet-repressor (TetR)-based negative autoregulation circuits, which displayed graded dose-response and minimized cell-to-cell variation. Then, I generated TetR-repressible promoter variants that showed similar mean expression levels but different levels of noise. These modified promoters were integrated into negative feedback circuits, which exhibited gradual changes in mean expression but had distinct noise profiles. The data suggested that negative feedback circuits harboring relatively noisy promoters showed a U-shaped noise curve. In other words, the noise attenuation effect of negative feedback was moderate at low and high expression phases, but optimal at the medium expression phase. The computational simulation results were in good agreement with the experimental data. Further parameter sensitivity analysis was performed to investigate the effects of various parameters on response curve and noise of negative autoregulation. The results suggested that the response curve was sensitive to varying of parameters, while noise in the circuits was non-sensitive to the parameters that were examined.

In addition, I constructed combinatory genetic circuits responding to light and chemical signals, simultaneously. I first generated a dual input circuit converting different light intensity into varying of response thresholds to a chemical inducer (doxycycline). Next, I generated a ternary input circuit composed of LightOn, Cumate-switch, and TetR-inducible system. It allowed us to use different combinations of blue light and the two chemical inducers to generate gradual output values over two orders of magnitude. Therefore, we suggested that this circuit could provide a “digital-to-analog” converting function in mammalian cells.

Overall, this study provides insights into the quantitative control of gene of interest and design of future bio-computation systems in mammalian cells.