The critical myeloid transcription factor GATA1 can specify erythroid cells at expense of granulocytes/macrophages (GM) from the common myeloid progenitor (CMP). In red cells sustained GATA1 expression is required for terminal maturation. Conversely, GATA1 expression has to be extinguished to allow GM differentiation. Therefore, one component in dissecting myeloid specification will be to define how GATA1 expression is regulated. The level of GATA1 is mainly controlled transcriptionally. Here, we show that murine(m) GATA1 mRNA rises 6-fold as CMPs differentiate to MEPs (megakaryocyte-erythroid progenitor) and another 4-fold as MEPs differentiate to Ter119+ erythroid cells. As a step towards understanding the molecular basis of erythroid-specific GATA1 expression, we have been characterising GATA1 cis-elements. Previously, we and others showed that an upstream enhancer (mHS−3.5) is required to direct erythroid-specific GATA1 expression in cooperation with sequences near the mGata1 gene (IE promoter and intron element mHS+3.5) in transgenic mice. Though a mGata1 transgene regulated by mHS−3.5-IE-mHS+3.5 grossly rescues erythropoiesis in GATA1 knock out mice, germ line deletion of mHS−3.5 leaves red cell GATA1 expression unaffected. This suggests other cis-elements in the mGata1 locus can substitute for mHS−3.5 in red cells. Recently, we identified an erythroid-specific DNase I hypersensitive site (DHS), mHS-25, with enhancer activity in erythroid cell lines and where the chromatin associated with it is hyperacetylated at histone H3/H4. We now demonstrate by fine DHS mapping that mHS-25 is more complex, being composed of two adjacent DHSs ~500 bp apart (mHS-25 and mHS-26). These DHSs are present only in primary red cells and not other primary cells. Chromatin immunoprecipitation (ChiP) shows that H3 and H4 associated with both sites is hyperacetylated only in Ter119+ cells. We tested mHS25/6 function in mice transgenic for a mGata1-LacZ reporter construct regulated by mHS-25/6-IE-mHS+3.5. β-galactosidase expression was quantitated in myeloid lineages by FACS analysis using lineage-specific antibodies and the β-galactosidase substrate FDG. 6 out of 7 F0 transgenic embryos expressed β-galactosidase in 5–23% of fetal liver Ter119+ cells. FDG staining was not detected in CD61+Mac1- megakaryocytes or Mac-1+ macrophages. <0.2% of Ter119+ cells stained with FDG when mHS-25/6 was not present. Erythroid-specific reporter gene expression was confirmed in bone marrow samples in 2 independent lines of mice transgenic for this reporter construct. ChiP analysis demonstrated that GATA1, SCL, E2A, the SCL-interacting protein Ldb1 and LMO2 bind in vivo to mHS-25/6 only in Ter119+ cells (and not primary megakaryocytes or neutrophils). Previously, we and others have shown that these transcription factors exist in a multi-protein complex to activate gene expression. Taken together, these findings suggest mHS-25/6 is an erythroid-specific GATA1 enhancer in primary murine cells. Finally, to begin to understand the relative roles of mHS−3.5 and mHS-25/6 during erythroid differentiation, we show by DHS mapping and ChiP analysis that in the multi-potential myeloid cell line FDCP-mix, only mHS−3.5 is present (but not mHS-25/6) and that it binds GATA1/SCL/E2A/Ldb1/LMO2 in vivo. Later in Ter119+ cells both mHS−3.5 and mHS-25/6 are seen with GATA1/SCL/E2A/Ldb1/LMO2 detected at these sites. This suggests a hierarchical utilisation of these two mGata1 cis-elements during erythroid differentiation.