Gene Regulatory Sequences And Protein Binding in Genome Sequences

National Library Of Medicine

Investigators

Linked publications, trials & patents

Abstract

The cooperative binding among large numbers of transcription factors. Pair wise cooperativity between transcription factors (TFs) have been studied extensively. Yet, nearly one thousand TFs are expressed in each cell and the cooperativity among large number of TFs in their natural environment have not been explored. Using ChIP-seq data of approximately 1000 TFs and co-factors in HepG2 and K562 cells, we have systematically characterized the co-binding among these proteins. Our analysis revealed the high promiscuity of the cooperative binding among all families of TFs and co-factors at the major regulatory regions, as well as the lack of sequence specificity for most of TFs at their associated genomic sites. However, a small number of proteins exhibit strong preferences for promoters/enhancers or cell-type specific/housekeeping genes, and the main enhancer sites are highly enriched in canonical motifs of cell-type specific TFs. These findings present a picture of transcription regulation carried out through both specific TF-DNA binding and promiscuous cooperativity among most DNA-associated proteins. Emerging evidence has shown the common occupancy of dozens to hundreds of transcription factors (TFs) on cis-regulatory elements (CREs), yet the underlying details are largely unknown. In this study, leveraging extensive collections of TF ChIP-seq data of more than 1000 TFs in human HepG2 and K562 cells, we located highly focused TF binding sites (FBSs) within CREs as single-nucleosome depleted regions, which accommodate the majority of the total TF binding events. Approximately 25,000 FBSs were identified in each cell type with high confidence. For more than 90% of TFs, including some pioneer factors such as GATA1 and JUN, their binding sites out of FBSs barely show nucleosome depletion. Essential cellular function related motifs and phenotypically causal variants are strongly enriched in the FBSs, but not in their immediate flanking regions within CREs. Most TFs bind to FBSs not containing their canonical motifs. Our study constructed high-resolution maps of chromatin accessibility at distal CREs in the two human cells. Moreover, we revealed the critical connection between highly focused TF binding and the nucleosome depleted status of DNA in vivo. We propose a model of TF co-binding in vivo and suggest that a short DNA residence time of most TFs underlies the requirement of a large number of TFs for sustained nucleosome depletion at CREs.