The protein, erythropoietin, is a primary regulator of erythropoiesis. The binding of erythropoietin to the surface of hematopoietic progenitor cells is followed by receptor mediated endocytosis of erythropoietin. We have recently cloned the human erythropoietin gene which extends 2 Kb 5' of the canonical start site. The 5' flanking region does not contain TATA or CCAAT sequences usually associated with proximal promoters of cellular genes. The minimal promoter contains an SP1 binding consensus sequence located at minus 19 in addition to the binding consensus sequence for the erythroid specific transcription factor, GATA-1, located at -47. In gel mobility assays, nuclear extracts from erythroid and non-erythroid cell lines provide similar protein-DNA binding patterns to the proximal promoter. An additional band was observed in erythroid cell lines, likely corresponding to GATA-1 binding. In DNase I footprinting, we observed extensive protein binding to the region about -19, the location of the SP1 site, and the GATA-1 binding site at -47. Other protein binding regions are also detected within the proximal promoter including AP2 and another region with no known consensus sequences. Truncated deletion mutations of the regions extending beyond the cap site were linked to a luciferase reporter gene and assayed in OCIM1 erythroid cells by transient transfection. Deleting the unknown footprinted region at -151 in the promoter fragment resulted in an increase of activity to about 1.8 of the activity observed with the longer construct extending to 200 bp 5' indicating that the unknown footprint region may have a negative regulatory effect on the erythropoietin receptor promoter. Deleting the AP2 site reduced transcriptional activity to levels observed with the reference fragment indicating that the AP2 site may have an enhancing activity or a positive regulatory effect on the promoter activity. Surprisingly, deletion of the GATA-1 site at -47 in a promoter fragment reduced the transcriptional activity to only about 80% of the reference sequence. This suggests that while the GATA-1 site is able to enhance activity of the promoter, it is not absolutely required for promoter activity. This construct extending to only -29 leaves an SP1 site at -19 intact, suggesting that the SP1 site is sufficient to obtain at least minimal promoter activity. In contrast, deletion of these 5' sequences to +3 of the cap site markedly reduces luciferase activity to about 20% of the reference construct. These results suggest that SP1 is required for promoter activity and that GATA-1 acts to further activate transcription.
