The existence of two rather than one estrogen receptor, today characterized as estrogen receptor alpha (ERalpha) and estrogen receptor beta (ERbeta), indicates that the mechanism of action of 17beta-estradiol and related synthetic drugs is more complex than previously thought. Because the homology of amino acid residues in the ligand-binding domain (LBD) of ERbeta is high compared with those amino acid residues in ERalpha LBD, previously shown to line the ligand binding cavity or to make direct contacts with ligands, it is not surprising that many ligands have a similar affinity for both receptor subtypes. We report that 17alpha-ethynyl, 17beta-estradiol, for example, has an ERalpha-selective agonist potency and that 16beta,17alpha-epiestriol has an ERbeta-selective agonist potency. We also report that genistein has an ERbeta-selective affinity and potency but an ERalpha-selective efficacy. Furthermore, we show that tamoxifen, 4-OH-tamoxifen, raloxifene, and ICI 164,384 have an ERalpha-selective partial agonist/antagonist function but a pure antagonist effect through ERbeta. In addition, raloxifene displayed an ERalpha-selective antagonist potency, in agreement with its ERalpha-selective affinity. However, although ICI 164,384 showed an ERbeta-selective affinity, it had a similar potency to antagonize the effect of 17beta-estradiol in the ERalpha- and ERbeta-specific reporter cell lines, respectively. In conclusion, our data indicate that the ligand binding cavity of ERbeta is probably more different from that of ERalpha than can be anticipated from the primary sequences of the two ER subtypes and that it will be possible to develop receptor-specific ligands that may form the basis of novel pharmaceuticals with better in vivo efficacy and side effect profile than current available drugs.