(675b) Modeling, Optimization, and Control of Bioprocesses Using Optogenetics

Microbial cell cultures can be used to produce powerful anti-cancer drugs, precursors for sustainable bioplastics, biofuels for drop-in gasoline substitutes, and many other valuable products. Yet, due partly to process complexity, and partly to the lack of available tools for process optimization (either computational or experimental), these bioprocesses frequently have low yields and limited economic viability. One challenge is balancing the trade-off between essential mechanisms for cell growth with the demand of producing desired products. Managing this trade-off requires dynamic control of cellular metabolism. Dynamic control is sometimes implemented using chemical inducers to control metabolism; however, these chemicals can persist in the system for long times, making the control action slow to reverse and hard to tune. Alternatively, optogenetics, which uses light-sensitive proteins to control metabolism, provides a promising tool for improved real-time control with faster responses and higher yields.

Optogenetic circuits can be constructed so that light will activate or inhibit a set of genes or metabolic pathways, usually by regulating transcription. Our group has developed a suite of optogenetic circuits for the yeast S. cerevisiae that respond to blue light. These circuits, depending on design, will either activate a gene when blue light is present, or activate a gene when blue light is turned off. Because the circuits are governed by multistep biochemical reaction pathways with different time scales, predicting the circuit’s dynamic response can be challenging. We developed semi-empirical mechanistic models of each of the circuits that enable us to examine the effects of light scheduling on the circuits. Through different light schedules, we can manipulate the expression level of whichever gene is controlled by the circuit. Different light doses, duty cycles, and waveforms were explored to investigate the dynamic behavior of the circuits, helping us uncover design criteria to select control policies for particular applications.

With optogenetic circuits in place for regulating metabolism, a new challenge is the development of an optimal control strategy for higher yields. To address this challenge, we developed a multiscale ordinary differential equation model describing cell growth, consumption of nutrients, production of desired compounds, and the state of the light-actuated metabolic pathway. We analyzed the model and sought forcing strategies that maximized the yield of desired products. In our case study—continuous isobutanol production in yeast—a periodic forcing function was determined to be effective at substantially increasing isobutanol production over any reachable steady-state. This result provides a compelling case for applying optogenetics in metabolic engineering.