The interaction of the erythroid transcription factor, GATA-1 with DNA is a major focus of our research. Vertebrate GATA factors have two zinc fingers that comprise the DNA binding domain. The C-terminal finger is the main DNA binding finger, and the three dimensional structure of this finger of GATA-1 bound to DNA has been solved by NMR. The structure reveals a finger and helix that bind to the major groove of DNA and an adjacent basic arm that binds in the minor groove. However, the GATA-1 N-terminal finger also plays an important role in stabilizing DNA binding, especially at genomic locations that contain multiple closely spaced GATA sites. The N-finger is also essential for interaction with the critical cofactor FOG. Mutations in the N-finger of GATA-1 that interfere with its ability to bind to DNA, but do not affect its interaction with FOG are associated with X-linked thrombocytopenia and thalassemia in humans and anemia in transgenic mice. More severe forms of these disorders are associated with GATA-1 mutations that disrupt FOG binding without altering DNA binding. Wild-type GATA-1 can rescue GATA-1 knockdown mice from lethal anemia, but a GATA-1 molecule lacking the N-terminal zinc finger cannot. Some GATA-1 N-finger point mutations that disrupt FOG binding have less severe erythroid phenotypes than the mutant lacking the entire N-finger, again suggesting that the N-finger has a FOG independent role in erythropoiesis in addition to its FOG dependent role. We are interested in understanding the contribution of the N-finger to GATA-1 function. In collaboration with the Yamamoto lab we have generated a mouse model for thrombocytopenia by expressing the V205G N-finger GATA-1mutant that is impaired in FOG binding, in GATA-1 knockdown mice. These mice develop thrombocytopenia at a high rate and we hope to use them to understand the mechanism of the disease and the role played by the N-finger. We are currently investigating two new complex GATA sites that occur in the GATA-2 promoter and are involved in turning the gene on and off during development. Both of these sites require the N-finger for full binding to DNA. DNA recognition by the N-finger of GATA-3 is also important for regulation of some cytokine genes that are important in TH2 cell development. IL-4, -5, -10. and -13 are expressed in TH2 but not TH1 cells, and IL-4, -5 and 13 are at the same genetic locus in mice and humans. It was previously shown that the IL-5 gene is GATA-3 activated, and that the N-finger of GATA-3 is required for binding to a palindromic GATA binding site in the IL-5 promoter. We have shown that the IL13 gene is also GATA-3 regulated. Three GATA recognition sequences in the IL13 gene promoter form a high affinity-binding site for two molecules of GATA-3, and two of the sites have GATG as their core sequence. All three sites are necessary for full activity of this promoter at limiting GATA-3 concentrations, and the N-finger is involved in binding to these sites. We have shown that GATA-1 can adopt different conformations that are DNA binding site specific, and these variations can be detected by altered migration in electrophoretic mobility shift assays or by differential resistance to proteases. These variations are not due to DNA bending since we have established that GATA-1 bends DNA slightly in a binding site independent manner. We have also demonstrated that GATA-1 is unable to stimulate transcription when bound to some DNA sites. In addition to FOG, there are several other cofactors that interact with the zinc fingers of GATA-1, and their ability to bind to GATA-1 may be influenced by the conformation that GATA-1 adopts in response to DNA. Consequently, we are attempting to solve the structure of the linked GATA-1 zinc fingers on a number of double DNA binding sites by X-ray crystallography. Meanwhile we have taken a biochemical approach to study the N-terminal finger interactions with DNA. We have shown that the binding specificity of the N- and C-terminal zinc fingers is distinct. Binding site selection experiments using the GATA-1 C-finger or the GATA-1 N-finger fused to the basic arm of the C-finger show that both of these peptides prefer GATA containing binding sites. However, the N-finger of GATA-2 prefers sites containing GATC. Recently, it has been shown that the N-finger of GATA-1 also binds to DNA independently, but with such low affinity that a selection experiment could not be performed. However, conventional binding studies show that this finger also prefers the GATC site, unlike the fusion protein containing the N-finger and the basic arm of the C-finger mentioned above. These observations lead to the conclusion that the C-terminal basic arm can change the specificity of the GATA-1 N-finger. Fusing this same basic arm to the GATA-2 N-finger also changes its specificity from GATC to GATA, suggesting that the GATA-1 basic arm controls the specificity at the last base of the four base core binding site. A cluster of amino acids, QTNRK, within this arm of the C-terminal finger, is largely responsible for this preference. These amino acids can convert the specificity of the N-finger of GATA-1 from GATC to GATA when they replace five analogously positioned N-finger amino acids. Because some biologically important GATA binding sites contain a combination of a canonical and non-canonical sequence, the mode of DNA recognition at these sites is significant and the N-finger of the GATA proteins may be particularly important here. In collaboration with the Bougneres lab, we have identified a complex GATA site in the promoter of the p110 ? subunit of the P13 kinase gene that appears to be involved in regulating insulin resistance. The C genotype of a previously identified T/C polymorphism was found to correlate with increased sensitivity to insulin in two cohorts of obese non-diabetic children. This polymorphism creates a strong GATA binding site between two weaker sites in the p110 ? gene promoter. Lymphocytes from two cohorts of obese children homozygous for the C polymorphism have 1.5 fold higher p110 ? mRNA levels, and 1.7 fold higher p110? protein levels than cohort members with the T genotype. The levels of the p85, the other subunit of the PI3 kinase, are the same throughout the cohorts. These increases most likely occur through enhanced activation of the p110 ? subunit gene by GATA-3. The C promoter is more active than its T counterpart in transient assays in GATA-3 containing cells. The C promoter has a higher affinity than the T for GATA-2 and -3, both of which are involved in adipogenesis. While lymphoctes, which are not physiologically relevant to insulin resistance, were used in these studies, insulin responsive tissues could not be collected from this group of children.
