Purpose: In this study, two unreported estrogen antagonists were identified using a combination of computational screening and a simple bacterial estrogen sensor.
Methods: Molecules here presented were initially part of a group obtained from a library of over a half million chemical compounds, using the Shape Signatures method. The structures within this group were then clustered and compared to known antagonists based on their physico-chemical parameters, and possible binding modes of the compounds to the Estrogen Receptor alpha (ER alpha) were analyzed. Finally, thirteen candidate compounds were purchased, and two of them were shown to behave as potential subtype-selective estrogen antagonists using a set of bacterial estrogen biosensors, which included sensors for ER alpha, ER beta, and a negative control thyroid hormone beta biosensor. These activities were then analyzed using an ELISA assay against activated ER alpha in human MCF-7 cell extract.
Results: Two new estrogen receptor antagonists were detected using in silico Shape Signatures method with an engineered subtype-selective bacterial estrogen biosensor and commercially available ELISA assay. Additional thyroid biosensor control experiments confirmed no compounds interacted with human thyroid receptor beta.
Conclusions: This work demonstrates an effective combination of computational analysis and simple bacterial screens for rapid identification of potential hormone-like therapeutics.