The acidic activation domains of the GCN4 and GAL4 proteins are not alpha helical but form beta sheets.

The most common class of activation domains, the so-called acidic activators, has been proposed either to adopt an amphipathic alpha-helical structure or to exist as unstructured "acid blobs." However, genetic analysis of an acidic activation domain in the yeast GAL4 protein has suggested that the structure of the activation region is a beta sheet. To distinguish between these models, we conducted a biophysical analysis of peptides corresponding to the yeast GAL4 and GCN4 acidic activation domains. Circular dichroism spectroscopy shows that the peptides are not alpha helical, but that they can undergo a transition to a structure that is almost 100% beta sheet in character in slightly acidic solution. We also show that the artificial acidic activator AH has structural properties that are markedly different from the natural GAL4 and GCN4 domains and does not adopt a beta-rich structure at reduced pH.

Spectroscopic studies of the DNA binding site of the GAL4 "zinc finger" protein.

The yeast GAL4 protein, a transcriptional activator of genes involved in galactose metabolism, binds as a dimer to several closely related seventeen base pair upstream activation sequences (UASGs) that are nearly symmetric about a central dT-dA base pair. A previous study of a GAL4-UASG complex (Carey, M., Kakidani, H., Leatherwood, J., Mostashari, F. and Ptashne, M. (1989) J. Mol. Biol. 209, 423-432) elucidated a pattern of contacts consistent with the protein partially wrapping itself around the helical cylinder, assuming a B-form conformation for the DNA. Alternatively, both monomers could sit on one face of the cylinder if the DNA exists in an underwound conformation such as A-form. Spectroscopic studies that distinguish between these models are reported here. Oligonucleotides containing the consensus UASG or a nine base pair "half site" both exhibit circular dichroism (CD) spectra characteristic of B-form DNA. Two-dimensional NMR studies of the half-site also indicate a B-form conformation. When a GAL4 protein fragment containing the entire DNA-binding and dimerization domains (amino acids 1-140) is bound to the UASG, the CD spectrum above 240 nm changes only slightly, and not in a manner consistent with DNA unwinding. Our studies suggest that the UASG does not adopt an unusual underwound conformation in the absence or presence of the GAL4 protein, and favor the model in which the dimer partially wraps around the helix cylinder.