DNA Binding Diversity Achieved Through the Interaction of GATA1 N-Finger and GATA Motif Is Important for Embryonic Erythropoiesis

Abstract 3441

Transcription factor GATA1 regulates a set of genes essential for the erythroid and megakaryocytic cell differentiation through the interaction with GATA motifs (consensus sequence: A/TGATAA/T). Two zinc fingers within GATA1 have been identified to be important in the DNA binding of GATA1, which are referred to as C-finger (CF) and N-finger (NF) domains. It has been shown that transactivation activity of GATA1 is completely abolished upon deletion of the CF domain, indicating that the CF domain is a requisite for the DNA binding of GATA1. While conventional reporter transactivation analyses hardly clarified the importance of the NF domain for the DNA binding, substitution mutations on 216^th arginine (R²¹⁶) located in the DNA-interacting surface of the NF domain have been identified to cause familial diseases of thrombocytopenia, thalassemia, and porphyria. As a consequence of the substitution of R²¹⁶ to glutamine (Q) or tryptophan (W), DNA binding activity of GATA1 to a palindromic configuration of two GATA motifs (palindromic GATA) was largely diminished, while that to a single GATA motif was maintained.

In this study we have examined the DNA binding diversity of GATA1 caused by the difference in the configuration of GATA motifs. We performed surface plasmon resonance (SPR) analyses of GATA1 to a single GATA, a palindromic GATA, and a repeating configuration of two GATA motifs (tandem GATA). We found that GATA1 binds to the palindromic GATA motif in a bivalent way, while it binds to the single GATA motif in a monovalent mode. We also found that a double quantity of GATA1 is associated with the tandem GATA motif and GATA1 lacking the NF domain binds to any configurations of GATA motif in a monovalent way.

To further investigate contribution of the NF domain to the binding mode of GATA1, we have constructed two types of GATA1 mutants; one type was the substitution mutations on R²¹⁶ (R216Q and R216W) that were mouse homologues of the human mutations, while the other type was the alanine substitution mutation on three lysine residues (K²⁴⁵, K²⁴⁶ and K³¹²; referred to as 3KA mutant), whereby dimerization potential of GATA1 was reduced to trace level similar to the case for GATA1 lacking the NF domain. Impotantly, R216Q and R216W mutants bind the palindromic GATA motif in a monovalent way, while these mutants bind normally to the other configuration of GATA motifs. In contrast, we found that one molecule of 3KA mutant bound to the tandem GATA motif and this observation seems to explain well the fact that dimerization potential of GATA1 is an important requisite for the full-function of GATA1 in embryos. The binding modes of this 3KA mutant to the other configurations were not influenced.

These results thus demonstrate that the both NF and CF domains recognize the multiple configurations of GATA motifs and specify the binding modes of GATA1. Importantly, GATA1-deficient mice rescued with R216Q were lethal during late gestation period due to abnormality in erythroid differentiation, indicating that the contribution of the NF domain to the recognition of the palindromic GATA motif configuration indeed functions in vivo. These results thus support our contention that the NF domain acts to regulate a proper spatio-temporal gene expression of a subset of GATA1 target genes utilizing the variations in the GATA motif configuration.