A new approach to isolate genomic control regions. Application to the GATA transcription factor family.

Abstract

We have designed a new strategy to isolate unknown DNA regions interacting with one or several related regulatory proteins. It involves trapping such DNAs by their cognate binding proteins followed by PCR amplification, as described previously [Kinzler, K. & Vogelstein, B. (1989) Nucleic Acids Res. 17, 3645-3653]. To overcome the inability of such a procedure to discriminate between functional and non-functional binding sites as well as to specifically trap short DNA motifs from the whole higher eukaryotic genome, we have used as starting material DNA isolated from transcriptionally competent chromatin fractions, instead of total genomic DNA. To test our strategy, we selected human DNA sequences that bind members of the GATA family, known to recognize similar WGATAR motifs. These proteins are expressed in different cell types in which they regulate the transcription of different sets of genes; thus, transcriptionally active chromatin containing GATA motifs should differ according to the cell type. We have trapped and analyzed DNA fragments isolated from an active chromatin fraction, from erythroid cells and lymphoid cells, using GATA-1 and GATA-3 proteins, respectively. We show that regulatory GATA sequences known to be in open chromatin in erythroid cells (typified by the HSIII fragment of the beta-globin locus control region) or in lymphoid cells (typified by a fragment of the CD2 locus control region) are dramatically enriched in a cell-specific manner, demonstrating the potency of the method. The sequences of the erythroid or lymphoid DNA fragments isolated through the procedure described here were determined and display subset-site preference for GATA-1 and GATA-3.