Few transcription factors (TFs) have been studied in the context of an integrative analysis incorporating genomic datasets from diverse genome regulatory mechanisms. Such an analysis allows the testing of specific regulatory associations in an unbiased and comprehensive manner. The promoter binding TF complex NF-Y regulates a diverse set of constitutive, inducible, developmental, oncogenic and tissue-specific genes. Using cancer models, ChIP-Seq, shRNA, and genomics, I have undertaken a genome-wide study of NF-Y. NF-Y binds to not only promoters but also extensively to enhancers, select classes of repetitive elements, inactive chromatin domains and insulators. NF-Y is a “pioneer”-like factor able to access its motif within closed, transcriptionally inactive chromatin domains. NF-Y pervasively associates with FOS, usually in the absence of JUN and the AP-1 motif, and with a group of growth controlling oncogenic TFs. I also show that NF-Y asymmetrically binds to its motif and stereo-aligns with specific TFs and their motifs. My results indicate that NF-Y is not merely a commonly-used, proximal promoter TF, but rather functions at a more diverse set of genomic elements. The dynamics of TF occupancy, cis-regulatory element (CRE) usage and their linkage to gene expression during a differentiation process, from a genome-wide perspective, is poorly understood and is critical to the understanding of fundamental aspects of development and disease. I utilize a model of inflammation-mediated oncogenic transformation, siRNA, ChIP-Seq, FAIRE-Seq, and microarrays to study the genomic aspects of transformation driven by Src-mediated activation of the inflammatory TF STAT3. I show that CRE usage is static, even in the presence of induced STAT3 activity, and large-scale transcriptional and phenotypic changes. STAT3 induced occupancy is tightly associated with FOS, pre-existing CREs, and does not create CREs de novo. I also highlight a putative role of TSC22D3 in inhibiting an epigenetic switch and in STAT3 and AP-1 factors driving the embryonic-like and bone-like phenotypes of breast cancer. The research presented here suggests that phenotypic alterations occurring during disease are not driven by large-scale perturbations of CRE usage. Overall, this dissertation provides an invaluable resource of genome-scale datasets within cancer models that will assist in future endeavors of scientific discovery.
| Identifer | oai:union.ndltd.org:harvard.edu/oai:dash.harvard.edu:1/9920185 |
| Date | 19 November 2012 |
| Creators | Fleming, Joseph |
| Contributors | Struhl, Kevin |
| Publisher | Harvard University |
| Source Sets | Harvard University |
| Language | en_US |
| Detected Language | English |
| Type | Thesis or Dissertation |
| Rights | open |
Page generated in 0.0028 seconds