View article

Our ability to routinely engineer genetic networks for applications is limited by the scarcity of highly specific and non–cross-reacting (orthogonal) gene regulators with predictable behavior. Though antisense RNAs are attractive contenders for this purpose, quantitative understanding of their specificity and sequence-function relationship sufficient for their design has been limited. Here, we use rationally designed variants of the RNA-IN–RNA-OUT antisense RNA–mediated translation system from the insertion sequence IS10 to quantify >500 RNA-RNA interactions in Escherichia coli and integrate the data set with sequence-activity modeling to identify the thermodynamic stability of the duplex and the seed region as the key determinants of specificity. Applying this model, we predict the performance of an additional ∼2,600 antisense-regulator pairs, forecast the possibility of large families of orthogonal mutants, and …