The reengineering of protein-small molecule interfaces represents a powerful tool of chemical biology. For many applications it is necessary to engineer receptors so that they do not interact with their endogenous ligands but are highly responsive to designed ligand analogues, which in turn do not interact with endogenous proteins. The chemical design strategy used to reengineer protein-small molecule interfaces is particularly challenging for interfaces involving relatively plastic receptor binding sites and therefore presents a unique challenge in molecular design. In this study we explore the scope and limitations of a new strategy for manipulating polar/charged residues across the ligand receptor interface of estradiol (E2) and the estrogen receptor (ER). Carboxylate-functionalized E2 analogues can activate ERalpha(Glu353-Ala) and ERbeta(Glu305-->Ala) with very large selectivites, demonstrating that this design strategy is extendable to other members of the steroid hormone receptor family. Neutral E2 analogues were found to complement ERalpha(E353A) with similar potencies but with generally lower selectivities. This suggests that the high selectivity observed with ligand-receptor pairs generated by exchanging charged residues across ligand-receptor interfaces is only due in part to their complementary shapes and that appropriate introduction of charged functionality on the ligand can provide substantial enhancement of selectivity by decreasing the engineered ligands affinity for the endogenous receptor. Attempts to modify the cationic residues by complementing Arg394-->Ala or Arg394-->Glu were not successful.