Effect of replacement of "zinc finger" zinc on estrogen receptor DNA interactions.

Exposure of bovine estrogen receptor to the metal chelators EDTA and 1,10-phenanthroline results in a loss of nonspecific DNA binding, presumably because of the removal of "zinc finger" zinc. Nonspecific DNA binding, as measured by a DNA-cellulose binding assay, can be restored by dialysis of the aporeceptor against buffer containing zinc, cadmium, and cobalt but not with buffer containing copper or nickel. More detailed studies were carried out using a bacterially expressed polypeptide encompassing the DNA binding domain of the human estrogen receptor. Apopolypeptide fails to bind DNA specifically, as measured by mobility shift assay using a consensus estrogen response element hexamer containing oligonucleotide, but DNA binding was restored by dialysis of the apopolypeptide against buffer containing zinc, cadmium, and cobalt but not with buffer containing copper or nickel. Dissociation constants of zinc- and cadmium-reconstituted polypeptide for the estrogen response element hexamer (66 and 48 nM, respectively) are virtually indistinguishable from native polypeptide (Kd = 48 nM) whereas cobalt-reconstituted polypeptide has a lower affinity (Kd = 720 nM). However, native, zinc-, cadmium-, and cobalt-reconstituted polypeptides gave identical results in a methylation interference assay. Competition experiments with zinc and copper or nickel suggest that copper and nickel are able to bind to zinc finger residues but do so nonproductively. The relative affinities copper greater than cadmium greater than zinc greater than cobalt greater than nickel for the polypeptide were determined by a zinc blot competition assay. The ability of cadmium and cobalt to substitute for zinc in the zinc fingers demonstrates a structural "flexibility" in the DNA binding domain as each of these metals has slightly different ionic radii. On the other hand, subtle differences in DNA binding affinity and/or specificity could exist, which may not be detectable here. Also, the ability of metals to substitute for zinc in the DNA binding domain suggests that metal substitution in these zinc fingers in vivo may be of relevance to the toxicity and/or carcinogenicity of some of these metals.