High-Throughput Cell-Based Strategies for Optimizing and Generating New RNA Component Functions

Advances in RNA biology and engineering have directed significant research efforts toward developing RNA devices that are capable of user-programmed processing and integration of diverse biological signals to control cellular function.  Synthetic biology approaches have resulted in the design of modular frameworks that integrate functional components for the reliable construction of synthetic RNA-based gene regulatory devices. As one example, a modular composition framework for a ribozyme-based device platform was recently described. The device platform is composed of three functional components: an actuator component, comprising a hammerhead ribozyme, a sensor component, comprising an RNA aptamer, and a transmitter component, comprising a sequence that transmits information between the actuator and sensor components through a strand-displacement mechanism. The modular device platform enables the engineering of individual component activities without significant device redesign. However, often the device regulatory properties, such as the basal activity and dynamics range, must be tuned when the device is integrated into the design of a system to match activities of the system components. Therefore, a strategy that enables the rapid engineering of RNA component functions within cellular environments is highly desired.

We have developed a high-throughput in vivo screening strategy based on fluorescence-activated cell sorting (FACS) by linking device regulatory activity to cellular fluorescence levels. A dual-color plasmid system has been constructed as a device screening plasmid. One color in the system is used to measure device performance, whereas the other color is used as a reporter that enables normalization of gene expression variation due to plasmid copy number, ribosome loading, and other cellular factors. The budding yeast Saccharomyces cerevisiae was selected as the cellular chassis for its ease of culturing, genetic manipulation, and high-efficiency transformation. RNA devices libraries were generated by introducing mutations in the targeted RNA component and transformed into yeast through homologous recombination. Control experiments with an RNA sensor library with over 10⁶ different variants, demonstrated that two sorting cycles were sufficient to pull out the single functional variant in that library. These initial control experiments illustrate the highly efficiency enrichment capacity of our sort (~100%).

We applied the FACS-based strategy to the optimization of the performance characteristics of the ribozyme-based device platform. Our efforts focused on using our high-throughput FACS approach to search a greater sequence and structural space for sequences that would support lower basal expression levels and larger switch dynamic ranges within the cellular environment. An initial actuator library was generated by focusing mutations in key areas of the hammerhead ribozyme to select for modified ribozyme sequences with greater activity in the context of the device platform. A single negative sort was sufficient to identify several variants with increased activity, and we performed a systematic sequence analysis to identify two consensus sequences from our screen with greater cleavage activities. One of the consensus loop sequences exhibits a three-fold lower basal activity than the parent ribozyme while maintaining moderate fold-activation activities. We demonstrated that the optimized actuator sequence exhibits modular activity and can be paired with other transmitter and sensor components to construct new switches with lower basal activities, thereby demonstrating the power of the modular composition framework. We generated a second device library by randomizing sequences in the transmitter component of two ribozyme variants. We performed iterative rounds of sorting in the absence and presence of ligand to isolate members with greater fold-activation than the respective parents. Isolated transmitter components enable for greater dynamic ranges up to doubling the observed activation ratio over the parents. Current efforts are extending the approach to screen sensor libraries within the cellular environment to develop a strategy to generate new sensors to desired sequences that function within the device context and desired cellular chassis. A surface plasmon resonance (SPR)-based strategy has also been developed on a Biacore instrument to enable rapid screening and characterization of new sensor binding properties. Therefore, the FACS-based screening strategy and SPR-based characterization strategy provide a powerful integrated technology platform that enables direct generation and integration of component functions within the device framework and desired cellular environment and rapid quantitative characterization of binding properties of generated devices.