We analyzed the sequences required for transcription of the mouse P-major-globin gene by introducing deletion and linker scanning mutations into the 5'-flanking region and then studying the effects of these mutations on 0-globin gene transcription in a HeLa cell transient expression assay or after stable introduction into mouse erythroleukemia cells. Consistent with earlier studies, we found that three distinct regions upstream from the RNA capping site are required for efficient P-globin gene transcription in HeLa cells: the ATA box located 30 base pairs upstream from the mRNA capping site (-30), the CCAAT box located at -75, and the distal sequence element CCACACCC located at -90. In the ATA and CAAT box regions, the sequences necessary for efficient transcription extend beyond the limits of the canonical sequences. Mutations in the sequences located between the three transcriptional control elements do not significantly affect transcription in HeLa cells. Although the promoter defined in HeLa cell transfection experiments is also required for efficient transcription in mouse erythroleukemia cells, none of the mutations tested affects the regulation of 3-globin gene transcription during mouse erythroleukemia cell differentiation. Thus, DNA sequences downstream from the mRNA cap site appear to be sufficient for the regulation of ,-globin gene expression during the differentiation of mouse erythroleukemia cells in culture.A prerequisite for understanding the detailed mechanisms of P-globin gene regulation during erythroid cell differentiation is the identification of the cis-acting DNA sequences required for accurate and efficient initiation of transcription. Comparison of the DNA sequences immediately upstream from the mRNA capping site of a number of mammalian P-globin genes revealed three highly conserved regions (17,28,32). The first region is the ATA box, which is located approximately 30 base pairs (bp) upstream from the mRNA cap site. This sequence, which was initially noted in Drosophila histone genes (M. Goldberg, Ph.D. thesis, Stanford University, Stanford, Calif., 1979), is found at the same relative position in most eucaryotic genes transcribed by RNA polymerase II (7, 11). The second region is the CCAAT box, which is located approximately 80 bp upstream from the mRNA cap site (5,17). This sequence is also found in a similar position near many other eucaryotic genes (5). A third region, which is conserved among adult mammalian globin genes that are p-like but is not a common feature of other eucaryotic promoters, is the sequence CCA/TCACCCT located approximately 90 or 105 bp upstream from the mRNA cap site (14, 32).The functional significance of these highly conserved sequence elements has been assayed both in vitro (27) and in vivo (14,15,25,26) in the case of the rabbit P-globin gene.As with many other eucaryotic promoters (7), only the ATA box is required for in vitro transcription (27). Deletion or mutation of the CCAAT or CACCC box sequences has no * Corresponding author.