Automated Design of Genetically-Encoded Protein-Sensing Riboswitches in a Cell-Free Expression System

A diverse array of genetically-encoded sensor devices have been implemented in cell-free systems, enabling the detection of toxins like fluoride and pathogens like Zika virus in a portable and cost-effective platform. However, protein detection has lagged behind. Here, we use RNA aptamer sensing elements that detect specific proteins to design riboswitches, using a previously-developed model of riboswitch translational regulation, the Riboswitch Calculator. We designed 35 activating and repressing riboswitches, detecting the phage MS2 coat protein, human monomeric C-reactive protein, and human IL32γ, and used a TXTL cell-free expression system as a test platform to determine riboswitch function. The best ON riboswitch activated the expression of a fluorescent reporter up to 15.9-fold, while the best OFF riboswitch repressed reporter production down to 13.6-fold. We discovered that, while the MS2 and IL32γ riboswitches functioned largely due to inducing structural changes in the ribosome binding site, the function of the CRP riboswitches was also affected by direct CRP-mediated repression. We further demonstrate the importance of accurate structural knowledge of aptamers for the precise design of highly-functional riboswitches. Overall, we expand the genetically-encoded sensor toolbox, and demonstrate the validity of automated computational design of protein-detecting RNA structural switches.