The human estrogen receptor-alpha (hER) is a ligand-activated transcription factor that functions as a homodimer. We sought to further understand the molecular processes involved in dimerization, and to develop a reagent that may function as an antiestrogen independent of the ligand binding site. To this end, we designed a 16-residue 'dimer-interface' oligopeptide derived from the helical region of the hER which is directly involved in dimerization. This peptide, termed the I-box peptide, has a high helical propensity in aqueous solution. The I-box peptide blocks hER action by causing aggregation and precipitation of both the ligand-bound and apo-hER. This effect is dependent on the helical nature of the peptide. A single Ile to Pro mutation in the helical region of the I-box peptide significantly reduces the helical content and abolishes the precipitation activity. Furthermore, the peptide activity appears to be specific for the hER. The I-box peptide does not significantly affect other proteins or steroid receptors tested. A homologous peptide derived from the nuclear receptor RXRalpha dimer interface, and a LXXLL-containing peptide from the coactivator TIF2 have no detectable in vitro effect on hER function or solubility. Our data suggest that rationally designed molecules capable of affecting steroid receptor quaternary structures may be potential avenues for the development of specific inhibitors of this class of proteins.