The glucocorticoid receptor (GR) regulates numerous genes governing physiological processes including inflammation, metabolism, and development that may be influenced by the circadian clock, which itself can be entrained by GR. The molecular mechanism of this entrainment by GR has not been well described. Among the circadian genes whose transcriptional oscillation is affected by GR are the Period genes PER1, PER2 and PER3. To investigate the mechanism of clock entrainment, as well as gain further insight into gene-specific regulation by GR, we have examined the transcriptional dynamics and physiological effects of Period gene regulation by GR in both human and mouse primary cells and cell lines. Per2, a core clock protein which is transcriptionally induced by GR, has an effect on glucose homeostasis in a mouse model. Per3, which is transcriptionally repressed by GR, has an effect on adipogenesis in a mouse model. PER1 encodes a clock protein essential for maintenance of the transcriptional feedback loop that constitutes the basis of the circadian program. We found that PER1 is the circadian gene induced most rapidly and with the greatest magnitude upon addition of the GR agonist dexamethasone, and we found that PER1 induction appears to be necessary for circadian entrainment by GR.

The glucocorticoid receptor (GR) binds to genomic response elements and regulates gene transcription with exquisite cell- and gene-specificity. Within a response element, the precise sequence to which GR binds has been implicated in directing receptor structure and activity. Here, we use NMR to show that different binding sequences affect the conformation of distinct regions of the GR DNA binding domain. Chemical shift difference mapping and analysis of a GR mutant suggest that an allosteric pathway links the DNA-binding surface, the dimerization interface and the associated dimer partner. We show that disrupting this pathway alters sequence-specific conformations, DNA-binding kinetics and gene-specific transcriptional activity. Our study provides insight into mechanisms by which binding sequences may impart multiple activities to a single factor through gene-specific structural changes.

Metazoan transcriptional regulation is governed by combinatorial control, in which distinct multifactor regulatory complexes composed of various combinations of factors assemble at different genomic sites. The vast number of unique combinations enables, in principle, a virtually inexhaustible roster of specific transcription patterns. However, it is unclear what directs combinatorial assembly of regulatory complexes that may have a hundred or more different proteins and what distinguishes different complexes in both composition and function. To begin addressing this problem, we probed pairs of factors that reside together in subsets of functionally distinguishable complexes. We tested whether we could use this "paired analysis" of the occupancy and activities of just two factors to functionally classify large complexes. Specifically, we analyzed the composition, activities, and determinants of multifactor complexes containing the glucocorticoid receptor or NFkappaB and the coregulator Brm at differentially regulated genes. Our results showcase the complexity of combinatorial regulation and open pathways to further analysis and understanding.

Regulation of gene expression is a fundamental biological process required to adapt the full set of hereditary information (i.e., the genome) to the varied environments that any organism encounters. Here, we elucidate two distinct forms of gene regulation – of endogenous genes by binding of transcription factors to information-containing genomic sequences and of selfish genes (“transposons”) by targeting of small RNAs to repetitive genomic sequences – using a wide array of approaches.

To study regulation by transcription factors, we used glucocorticoid receptor (GR), a hormone-activated, DNA-binding protein that controls inflammation, metabolism, stress responses and other physiological processes. In vitro, GR binds as an inverted dimer to two imperfectly palindromic “half sites” separated by a “spacer”. Moreover, GR binds different sequences with distinct conformations, as demonstrated by nuclear magnetic resonance spectroscopy (NMR) and other biophysical methods.

In vivo, GR employs different functional surfaces when regulating different genes. We investigated whether sequences bound by GR in vivo might be a composite of several motifs, each biased toward utilization of a particular pattern of functional surfaces of GR. Using microarrays and deep sequencing, we characterized gene expression and genomic occupancy by GR, with and without glucocorticoid treatment, of cells expressing GR alleles bearing differences in three known functional surfaces. We found a “sub-motif”, the GR “half site”, that relates to utilization of the dimerization interface and directs genomic binding by GR in a distinct conformation.

To study repression of tranposons, we characterized the production and function of small RNAs in the yeast Cryptococcus neoformans. We found that target transcripts are distinguished by suboptimal introns and inefficient splicing. We identified a complex, SCANR, required for synthesis of small RNAs and demonstrate that it physically associates with the spliceosome. We propose that recognition of gene products by SCANR is in kinetic competition with splicing, thereby further promoting small RNA production from target transcripts.

To achieve these results, we developed new bioinformatics tools: twobitreader, a small Python package for efficient extraction of genomic sequences; scripter, a flexible back-end for easily creating scripts and pipeline; and seriesoftubes, a pipeline built upon scripter for the analysis of deep sequencing data.

Glucocorticoids are steroid hormones that control multiple aspects of mammalian physiology. Glucocorticoids act through the glucocorticoid receptor (GR), a ligand-regulated transcriptional regulatory factor that is ubiquitously expressed. It is still unknown how the structurally simple glucocorticoid hormones generate diverse physiological outputs across multiple target tissues. To examine this question we characterized the activities of synthetic glucocorticoid ligands across a panel of GR-induced biological activity profiles. Our studies demonstrated that ligand structure is an important determinant of GR function. We found that arylpyrazole glucocorticoid compounds supported GR activities related to adipogenesis, inhibition of cell proliferation, inhibition of asthma marker gene expression and anti-inflammation, but did not inhibit osteoblast differentiation in cell culture. Additionally, arylpyrazole GR ligands displayed gene-specific transcriptional effects.

To investigate the therapeutic effects of glucocorticoids in lung epithelial cells during asthma treatment we developed a cell culture-based model for gene regulation. We found that dexamethasone and budesonide repressed the IL-13-induced asthma marker genes POSTN, SERPINB2 and CLCA1 in primary human bronchoepithelial cells. We found that the kinetics of regulation of asthma marker genes was slower when compared than that of other IL-13 regulated genes. In addition, the gene SERPINB2 was regulated by dexamethasone as part of a cluster of SERPINB genes that were all also induced by IL-13 treatment.

Finally, we discovered that glucocorticoids induce a novel state during adipogenesis in cell culture. Dex-primed preadipocytes are an intermediate cellular state between preadipocytes and adipocytes that differentiate into adipocytes when exposed to cAMP signaling. Importantly, we found that the mode of differentiation influences the metabolic properties of adipocytes. In summary, our results suggest that the diversity of glucocorticoid activities is generated by the combined actions of signals induced by the ligand itself, the cellular context and additional signaling pathways that affect glucocorticoid activity and the associated phenotypes.

The proper localization of chromatin modifications is critical to their function as signals for DNA compaction and transcription regulation. This localization is controlled by the activities of chromatin modifying enzymes, which “write” and “erase” chromatin modifications. While the chromatin field has identified the chromatin modifiers responsible for the deposition and removal of chromatin modifications, little is known about how the activities of chromatin modifiers are controlled to localize these modifications in the genome. In this thesis, we investigate the mechanisms that direct and control the activity of chromatin modifiers. We hypothesize that the identified mechanisms determine the localization of chromatin modifications and the biological roles of chromatin modifiers in the nucleus. To answer these questions we use a model family of chromatin modifiers, the human KDM4 histone demethylases, specifically KDM4C. Our approach is to reconstitute KDM4C and its substrates to probe the mechanisms which regulate demethylase activity. Moreover, we identify a transcriptional network reliant on KDM4C as a means to validate and discover novel regulatory mechanisms important to its role in controlling gene expression. In this thesis, we demonstrate that KDM4C can be regulated by chromatin modifications auxiliary to its substrates, methylated histone H3 lysine 9 and methylated histone H3 lysine 36. We show that auxiliary chromatin modifications can affect activity through directly disrupting interactions between the histone substrate and the catalytic domain. Alternatively, we show that recognition by KDM4C “reader” domains, the tandem tudor domain and the plant homeodomains, regulate the activity of the demethylase. For example, the recognition by the tandem tudor domain drastically stimulates the activity of KDM4C toward its substrate. Moreover, we have generated hypotheses about novel functions of the plant homeodomains that may regulate demethylase activity and localization. Thus, our work has identified many mechanisms fundamental to the regulation of KDM4C. We hypothesize that these modifications will inform future work on other members of the KDM4 demethylase family and broader classes of chromatin modifications. This work will also contribute to the broader understanding of how chromatin modifications are established and maintained in order to properly package the genome and regulate gene expression.

In response to glucocorticoids, the glucocorticoid receptor induces a cell type specific transcriptional program through both direct and indirect interactions with glucocorticoid response elements (GREs). The association between GREs and regulated genes is often unclear, because a GRE can be tens of thousands of base pairs from the gene it regulates. To help understand the mechanisms driving GR-DNA binding and transcription, we looked at the genome-wide binding and transcriptional response to glucocorticoids in three cell lines. Consistent with previous reports, we found that binding is more closely associated with upregulation than with downregulation. Interestingly, we found that GR binding regions (GRBRs) that contain a canonical GR motif, called a GR binding sequence (GBS), are better predictors of upregulation than those lacking a GBS, while the reverse is true of downregulation. To help understand the determinants of cell type specific binding, we compared the sequences of the GRBRs in the three different cell types and found that binding motifs for both the glucocorticoid receptor and for other transcription factors are overrepresented to different degrees in each cell type. The canonical GR motif is present in 66% of the U2OS sites, verses 14% of the Nalm6 sites; similarly, the AP-1 binding motif is present in 22% of A549 sites and 6% of Nalm6 sites. This difference in the sequence composition of binding sites suggests that the determinants of binding may be encompassed in the local sequence of a binding site. Supporting this, we found that the cell type specificity of binding could be recapitulated in reporters containing approximately 400 base pairs of DNA.