Comparison of in vivo and in vitro HSF1 occupancy uncovers factors influencing binding site determination
A transcription factor’s (TF) propensity to bind to any given locus is driven by several influences, including DNA sequence, chromatin environment, surrounding transcription factor binding landscape, and local TF concentration. For many TFs, DNA sequence preference is well characterized, but is insufficient to determine TF binding patterns in vivo. To pinpoint contextual elements beyond DNA sequence that determine TF binding, I compared in vivo and in vitro binding profiles of human heat shock factor 1 (hHSF1), a highly conserved TF that binds to its target elements upon heat stress and regulates the rapid and concerted heat shock response. hHSF1 binds the conserved heat shock element (HSE) motif, which has hundreds of thousands of instances in the human genome, yet only is detected at around 1000 sites by ChIP-seq after heat shock stimuli. I have identified all possible genomic binding sites of hHSF1 in the absence of chromatin with PB-seq, a high-throughput in vitro assay that detects regions of hHSF1-bound naked DNA. Up to 50,000 hHSF1-bound binding sites were identified with PB-seq, a majority of which contained a detectable HSE. Yet, only a small fraction (~2-8%) of these regions are bound in K562 cells by ChIP-seq upon heat shock. I have found that chromatin accessibility is a major determinant of hHSF1 binding, as ATAC-seq and DNase I-hypersensitive regions are depleted at sites where hHSF1 only binds in vitro. To consider the impact of other TFs, chromatin remodelers, and histone marks on hHSF1 binding sitedetermination, I have investigated occupancy and motif prevalence of hundreds of chromatin-associated factors pre- heat shock by using available ENCODE ChIP-seq datasets and motif databases.