Evolutionarily conserved Ets family members display distinct DNA binding specificities [published erratum appears in J Exp Med 1993 Sep 1;178(3):1133]

Members of the Ets family of proto-oncogenes encode sequence-specific transcription factors that bind to a purine-rich motif centered around a conserved GGA trinucleotide. Ets binding sites have been identified in the transcriptional regulatory regions of multiple T cell genes including the T cell receptor alpha and beta (TCR-alpha and -beta) enhancers and the IL-2 enhancer, as well as in the enhancers of several T cell-trophic viruses including Maloney sarcoma virus, human leukemia virus type 1, and human immunodeficiency virus-2. T cells express multiple members of the Ets gene family including Ets-1, Ets-2, GABP alpha, Elf-1, and Fli-1. The different patterns of expression and protein-protein interactions of these different Ets family members undoubtedly contribute to their ability to specifically regulate distinct sets of T cell genes. However, previous studies have suggested that different Ets family members might also display distinct DNA binding specificities. In this report, we have examined the DNA binding characteristics of two Ets family members, Ets-1 and Elf-1, that are highly expressed in T cells. The results demonstrate that the minimal DNA binding domain of these proteins consists of adjacent basic and putative alpha-helical regions that are conserved in all of the known Ets family members. Both regions are required for DNA binding activity. In vitro binding studies demonstrated that Ets-1 and Elf-1 display distinct DNA binding specificities, and, thereby interact preferentially with different naturally occurring Ets binding sites. A comparison of known Ets binding sites identified three nucleotides at the 3' end of these sequences that control the differential binding of the Ets-1 and Elf-1 proteins. These results are consistent with a model in which different Ets family members regulate the expression of different T cell genes by binding preferentially to purine-rich sequences that share a GGA core motif, but contain distinct flanking sequences.