Spectroscopic studies of wild-type and mutant "zinc finger" peptides: determinants of domain folding and structure.

The "zinc finger" model [Miller, J., McLachlan, A. D. & Klug, A. (1985) EMBO J. 4, 1609-1614; Brown, R. S., Sander, C. & Argos, P. (1985) FEBS Lett. 186, 271-274] makes both specific structural and specific functional predictions about zinc finger consensus sequences that can be tested with a combination of genetic, molecular biological, and biophysical techniques. The yeast transcription factor ADR1 contains two adjacent zinc finger domains; genetic and deletion analyses showed that amino acid substitutions and deletions in the zinc finger domains resulted in the loss of protein activity. To test the structural and folding predictions of the zinc finger model, peptides encompassing each of the ADR1 fingers were synthesized (ADR1a and ADR1b) as well as a mutant finger peptide (del138) deleted for a single amino acid residue. The folding and metal-binding characteristics of these were assessed by 1H nuclear magnetic resonance (NMR) and visible spectroscopy. While a single unique conformational species was detected for the two wild-type peptides upon tetrahedral binding of zinc, the deletion peptide did not bind zinc with tetrahedral geometry, nor did it fold into a zinc finger domain. The metal-binding and folding results found with the mutant peptide were similar to those obtained when thiol alkylation or imidazole protonation of the wild-type peptides was performed. These data indicate that ligand spacing and both thiol and imidazole participation in zinc binding are specific and necessary requirements for zinc finger folding, which provides direct support for the initial predictions of the model.