Abstract
Normal differentiation of red cells, platelets and eosinophils from a myeloid progenitor requires expression of the transcription factor GATA1. Moreover, GATA1 expression level influences lineage output; higher levels promote erythromegakaryocytic differentiation and lower levels eosinophil maturation. Conversely, repression of GATA1 expression is required for monocyte/neutrophil development. GATA1 expression is principally controlled transcriptionally. Thus, dissecting the molecular basis of transcriptional control of GATA1 expression will be one important facet in understanding how myeloid lineages are specified. To address this question we sought to identify all DNA sequences important for GATA1 expression. Previous analysis identified 3 murine (m)Gata1 cis-elements (an upstream enhancer, mHS-3.5, a haematopoietic IE promoter and elements in a GATA1 intron, mHS+3.5) conserved in sequence between human(h) and mouse. These studies also suggested additional unidentified elements were required for erythroid and eosinophil GATA1 expression. We compared sequence, mapped DNase I hypersensitive sites (HS) and determined histone H3/H4 acetylation over ~120 kb flanking the hGATA1 locus and corresponding region in mouse to pinpoint cis-elements. Remarkably, despite lying in a ~10 MB conserved syntenic segment, the chromatin structures of both GATA1 loci are strikingly different. Two previously unidentified haematopoietic cis-elements, one in each species (mHS-25 and hHS+14), are not conserved in position and sequence and have enhancer activity in erythroid cells. Chromatin immunoprecipitation studies show both mHS-25 and hHS+14 are bound in vivo in red cells by the transcription factors GATA1, SCL, LMO2, Ldb1. These findings suggest that some cis-elements regulating human and mouse GATA1 genes differ. Further analysis of in vivo transcription factor occupancy at GATA1 cis-elements in primary mouse eosinophils and red cells, megakaryocytic cells (L8057) and control fibroblasts show lineage- and cis-element-specific patterns of regulator binding (see table below). In red cells and megakaryocytes, GATA1, SCL, LMO2 and Ldb1 bind at two regulatory elements (mhHS-25 and mHS-3.5). Interestingly, the megakaryocyte transcriptional regulator Fli1 factor binds to mHS+3.5 specifically in megakaryocytes. In eosinophils, a different pattern of DNase I HS and transcription factor binding is seen. GATA1, PU.1 and C/EBPe (all regulate eosinophil gene expression) bind IE promoter and/or mHS+3.5. Collectively, these results suggest lineage-specific GATA1 expession is dependent on combinations of cis-elements and haematopoietic trans-acting factors that are unique for each lineage.
DNase I Hypersensitive sites and transcription factor occupancy at mGATA1 cis-elements.
mHS-26/-25* mHS-3.5 mIE mHS+3.5 m: mouse, h: human, *: HS identified in this study, TF: transcription factor Primary erythroid cells HS present, GATA1, SCL, LMO2, Ldb1 HS present, GATA1, SCL, LMO2, Ldb1 HS present, GATA1 HS present, GATA1 Megakaryocytic cells HS present, GATA1, SCL, LMO2, Ldb1 HS present, GATA1, SCL, LMO2, Ldb1 HS present, GATA1 HS present, GATA1 and Fli1 Primary eosinophils HS absent HS present, No TF detected HS present, GATA1 and C/EBPε HS present, GATA1, C/EBP ε and PU.1 Fibroblasts HS absent HS absent HS absent HS absent
Large-scale identification of sequence variants influencing human transcription factor occupancy in vivo