Solution structural studies of the Saccharomyces cerevisiae TATA binding protein (TBP)

The intrinsic fluorescence of the six tyrosines located within the C-terminal domain of the Saccharomyces cerevisiae TATA binding protein (TBP) and the single tryptophan located in the N-terminal domain has been used to separately probe the structural changes associated with each domain upon DNA binding or oligomerization of the protein. The unusually short-wavelength maximum of TBP fluorescence is shown to reflect the unusually high quantum yield of the tyrosine residues in TBP and not to result from unusual tryptophan fluorescence. The anisotropy of the C-terminal tyrosines is very high in monomeric, octameric, and DNA-complexed TBP and comparable to that observed in much larger proteins. The tyrosines have low accessibility to an external fluorescence quencher. The anisotropy of the single tryptophan located within the N-terminal domain of TBP is much lower than that of the tyrosines and is accessible to an external fluorescence quencher. Tyrosine, but not tryptophan, fluorescence is quenched upon TBP-DNA complex formation. Only the tryptophan fluorescence is shifted to longer wavelengths in the protein-DNA complex. In addition, the accessibility of the tryptophan residue to the external quencher and the internal motion of the tryptophan residue increase upon DNA binding by TBP. These results show the following: (i) The structure of the C-terminal domain structure is unchanged upon TBP oligomerization, in contrast to the N-terminal domain [Daugherty, M. A., Brenowitz, M., and Fried, M. G. (2000) Biochemistry 39, 4869-4880]. (ii) The environment of the tyrosine residues within the C-terminal domain of TBP is structurally rigid and unaffected by oligomerization or DNA binding. (iii) The C-terminal domain of TBP is uniformly in close proximity to bound DNA. (iv) While the N-terminal domain unfolds upon DNA binding by TBP, its increased correlation time shows that the overall structure of the protein is more rigid when complexed to DNA. A model that reconciles these results is proposed.