Terminal erythroid maturation requires coordinated activation of erythroid marker genes and repression of genes associated with the undifferentiated state. These gene expression patterns are mediated by the concerted action of the erythroid transcription factor GATA-1 and its cofactor FOG-1 that can activate or repress transcription depending on promoter context. We and others showed previously that one mechanism by which FOG-1 functions is to facilitate GATA-1 association with certain DNA target sites in vivo. Using gene complementation studies of GATA-1-ablated erythroid cells, we show that at several GATA-1-repressed target genes (c-kit, c-myc and GATA-2) FOG-1 is dispensable for GATA-1 occupancy in vivo but essential for gene repression and histone deacetylation. To examine how FOG-1 functions as co-repressor we performed affinity chromatography, conventional protein purification and in vitro binding studies to identify proteins that bind FOG-1. We discovered that FOG-1 directly associates with the nucleosome remodeling and histone deacetylase complex NURD. This interaction is mediated by a small conserved domain at the N-terminus of FOG-1 and the MTA-1 subunit of NURD. Association of FOG-1 with NURD occurs in vivo and depends on an intact N-terminus of FOG-1. A series of point mutations across the N-terminus of FOG-1 revealed a tight correlation between NURD binding and transcriptional repression. In particular, a single point mutation at the N-terminus of FOG-1 that abrogated NURD binding also blocked gene repression by FOG-1. Finally, the ability of GATA-1 to repress transcription was impaired in erythroid cells expressing a mutant form of FOG-1 that is defective for NURD binding. Together, these studies show that FOG-1 and very likely other FOG proteins are bona fide co-repressors that link GATA proteins to histone deacetylation and nucleosome remodeling via a novel protein interaction module.

Author notes

Corresponding author

Sign in via your Institution