We have previously shown that engineered chimeric enzymes that combine nuclear hormone receptor ligand-binding domains with a highly sensitive thymidylate synthase (TS) reporter yield simple sensors for hormone-like compounds based on bacterial growth. This work describes the construction of an optimized estrogen sensor in Escherichia coli with dramatically enhanced sensitivity and extraordinary reliability in identifying diverse estrogenic compounds and recognizing their agonistic/antagonistic pharmacological character. The ability of this system to assist discovery of new estrogen-mimicking compounds was validated by the screening of a small compound library, which led to the identification of two structurally novel estrogen receptor modulators and the accurate prediction of their agonistic/antagonistic effects in human cells. Strong evidence is presented here that the ability of our engineered sensor to detect ligand binding and discriminate between hormone agonists and antagonists occurs due to the allosteric communication between the artificially combined protein domains, where distinct conformational changes that occur in the receptor upon ligand binding are transmitted to the TS domain and alter its enzymic efficiency. A mechanism based on the repositioning of the hormone binding-domain helix 12 upon ligand binding and its effect on the intein-TS fusion junction will be presented to partially explain the changes in phenotype. Further, the use of the thymidylate synthase reporter system allows the rapid selection of new hormone sensors by providing a method for selection of improved mutants in liquid culture at large scale. Because the mechanism for allosteric regulation is likely to be general to nuclear hormone receptors, we have anticipated that additional sensors might be generated by simple substitutions of alternate ligand binding domains. This hypothesis has now been supported through the generation of a sensor for insect ecdysone hormone agonists and antagonists. The ability to generate selectable phenotypes in simple bacterial cells based on the modulation of activity of human drug targets, combined with a large-scale selection method in liquid media, represents the beginning of a potential path for the evolution of human therapeutics through cellular and metabolic engineering in simple unicellular organisms.

Extended Abstract Status: Not Uploaded