DNA-BINDING DOMAINS OF EUKARYOTIC TRANSCRIPTION FACTORS

Most of transcription factors recognize specific DNA sequences through small, discrete domains. In some cases, these domains can be interchanged between proteins showing that they are independently folded units. Mutational and three-dimensional structure analyses have been used to understand the means by which protein-DNA interactions are achieved. There are different kinds of DNA-binding domains. In many cases, this domain is fixed, but in others, it folds only on binding to DNA. Thus homeodomains, C2-H-2-class Zn fingers, steroid receptor-like C4-class Zn fingers, GAL4-like C6-class Zn fingers and E2 antiparallel beta-barrel have a << recognition helix >> that protrudes from a compact domain. By contrast, the basic region of bZIP-class proteins has a flexible segment that wraps around the major groove of DNA. The E2 protein, the steroid receptor-like C4-class Zn finger, the GAL4-like C6-class Zn finger, the bZIP-class and the bHLH-class of transcription factors form dimers. By opposition, most homeoproteins bind to DNA as monomers, and finally the C2-H-2-Class Zn finger proteins form covalent concatemers. The study of DNA-binding domains of eukaryotic transcription factors shows more than one way to present an alpha-helix in the major groove. In evolutionary terms, this implies that these proteins have arisen from different ancestors : consequently, the DNA-binding domain must have been << invented >> several times.