Modifications to the ET(A/B) mixed type compounds 1 (Ro. 46-2005) and 2 (bosentan) were performed. Introduction of a pyrimidine group into 1 resulted in a dramatic increase in affinity for the ET(A) receptor, and the subsequent optimization of substituents on the pyrimidine ring led us to the discovery of N-(6-(2-((5-bromo-2-pyrimidinyl)oxy)ethoxy)-5-(4-methylphenyl)-4-pyrimidinyl)-4-tert-butylbenzenesulfonamide (7k), which showed an extremely high affinity for the human cloned ET(A) receptor (K(i) = 0.0042 +/- 0.0038 nM) and an ET(A/B) receptor selectivity up to 29 000 (K(i) = 130 +/- 50 nM for the human cloned ET(B) receptor). The compound was designed on the hypothesis that the hydrogen atom of the hydroxyl group in 1 and 2 played a role not as a proton donor but as an acceptor in the possible hydrogen bonding with Tyr129. Since the incorporation of a pyrimidinyl group into the hydroxyethoxy side chain of the nonselective antagonist (1) dramatically enhanced both the ET(A) receptor affinity and selectivity, and since similar results were obtained from the benzene analogues, we put forward the hypothesis that a "pyrimidine binding pocket" might exist in the ET(A) receptor.