A family of transcription activators called the basic region-leucine zipper (bZIP) proteins is distinguished by an unusually simple motif which interacts specifically with a duplex DNA promoter. The DNA binding domain, called the bZIP element, mediates DNA recognition with an element of protein secondary structure that is not stabilized by tertiary structure. The bZIP element, which can fold and bind DNA autonomously, is distinct from an amino-terminal acidic region of the protein that is involved in transcriptional activation function. The bZIP element contains three segments, each of which plays a role in DNA binding. The leucine zipper segment is defined by its conserved heptad repeat of leucine residues that facilitates protein dimerization by formation of a parallel coiled coil. The basic region is largely unstructured in the absence of DNA, but forms an α-helical structure with its conserved basic and hydrophobic residues upon binding the DNA target site. These two regions are separated by a non-conserved six residue spacer segment. The bZIP element when bound to duplex DNA is a dimer of two continuous a-helical structures with its leucine zipper-coiled coil perpendicular to the DNA target site, and its two splayed basic regions specifically contacting the major groove of the DNA.
The bZIP family proteins are ideal candidates for detailed study because they employ an unusually simple DNA binding element that is able to recognize a diverse set of 9-10 base-pair DNA target sites. Based on their target site specificities, bZIP proteins are divided into subfamilies, including the activator protein 1 (AP1), cyclic-AMP response element binding protein/ activating transcription factor (CREB/ATF), and CAAT enhancer binding protein (C/EBP) families. C/EBP family members prefer a consensus 10 base pair sequence which contains two ATTG half sites in an inverted pair separated by two base pairs (ATTGCGCAAT). CREB/ATF proteins recognize the cAMP response element (CRE) which maintains similar half site spacing as the C/EBP site, but contains an ATGA half site (ATGACGTCAT). Members of the AP1 family recognize the same half site sequence as do members of the CREB/ATF family, ATGA, but in this case, each half sites is separated by only one base pair (ATGACTCAT). Interestingly, the yeast bZIP protein, GCN4, cannot formally be categorized into any family since it binds the AP1 and CRE target sites with almost equal affinity.
A number of studies have focused on identifying the region and the specific amino acid residues involved in determining sequence and half site spacing specificity in bZIP proteins. Although the zipper region is critical for dimerization, only residues in the basic and spacer regions are known to make specific contacts with the DNA target site. Crystallographic data of the GCN4 homodimer in complex with the AP1 and CRE target sites, as well as the AP1 complex (FOS-JUN heterodimer) bound to the AP1 target site, identified a quintet of residues in the basic region, N--AA--SR, that make specific DNA base contacts. The quintet of residues is highly conserved among bZIP proteins implying that, regardless of DNA target site preference, the same DNA base pair contacts are maintained by all bZIP proteins. The fact that the quintet is highly conserved suggests that the remaining residues within the basic and spacer region are involved in determining DNA target site preference.
Although GCN4, FOS, and JUN bind with high specificity to the AP1 target site, a comparison of their crystal structures reveals that differential phosphate contacts may be involved in achieving specificity; unlike GCN4, FOS and JUN each maintain one phosphate contact amino-terminal to the conserved quintet of residues. In addition to crystallographic data, gel electrophoresis experiments with mutants of GCN4 suggest that specific residues at the junction of the basic and spacer regions influence the half site spacing preference of bZIP proteins.