A complete deep mutational map of the lactose repressor and quantification of allosteric communication

The lactose repressor is a well-studied allosteric biosensor that has been used in myriad applications. Despite our deep understanding of the structure function relationship for this important transcription factor, we lack a full understanding of allosteric communication at a quantitative level. We posit that a deeper understanding of allosteric communication in said protein will improve our ability to rationally design new biosensors of similar topology with prescribed phenotype and dynamic range setpoints. In this study we have generated a complete deep mutational map of the wild-type lactose repressor composed of all single mutations and cognate phenotypes. We leveraged the resulting map to quantify allostery at the resolution of individual sites throughout the transcription factor. In addition, we generated partial deep mutational maps for two engineered antilac repressors to illustrate and quantify allosteric plasticity in a shared topology. Next, we deconvoluted the requirements for anti-repression in a collection of engineered antilac repressors to better reconcile our design heuristics. Finally, we leveraged this collection of data sets to construct highly accurate predictive models for single and double mutations, setting the stage for the rational design of cognate biosensors with prescribed phenotype and dynamic range setpoints.