Environmental Factor, August 2010, National Institute of Environmental Health Sciences
Chromatin Remodeling and the Glucocorticoid Receptor
By Sophie Bolick
Geneticist Gordon Hager, Ph.D. (http://ccr.cancer.gov/staff/staff.asp?profileid=5817) , chief of the Laboratory of Receptor Biology and Gene Expression at the National Cancer Institute, shared his recent research findings on epigenetic modification of chromatin in an afternoon seminar July 7 at NIEHS titled "Rapid Dynamics and Gene Regulation by Nuclear Receptors."
Hosted by Linda Birnbaum, Ph.D., John Cidlowski, Ph.D., and Ken Korach, Ph.D., Hager discussed his work on how nuclear receptors interact with chromatin and the effect of this interaction on hormone function and the transcription mechanisms involved in many disease processes, including cancer development.
Chromatin status is important for receptor action
As Hager and his group (http://rex.nci.nih.gov/RESEARCH/basic/lrbge/hager.htm) investigated the role of SWI/SNF protein remodeling complex glucocorticoid receptor (GR) action, they observed patterns emerging. "What we began to see is a layer of organization at the level of chromatin structure that dictates whether a receptor can gain access and act at those specific sites," explained Hager. The finding led him to speculate on several specific mechanisms that may be involved.
For one, if a specific remodeling system is necessary, then its presence is critical for the ability of the receptor to work at the site in question. A second possibility is that epigenetic marks on histones or DNA determine chromatin conformation at a given site. Finally, the presence of specific factors that recruit remodeling systems is also needed, he said, because "if the receptor can't open those sites on its own, then another protein is apparently able to open those sites for them."
Mechanisms governing access to GR binding sites
The data generated by his group's experiments (see text box) correlated well with the earlier data on GR binding sites. There is a universal overlap between GR binding and hypersensitivity sites. According to Hager, the vast majority of these GR binding events occur at pre-existing DNase hypersensitivity sites, which are already open due to some other process. But, if the chromatin isn't open and accessible, the GR cannot initiate that activity on its own.
As Hager explained, it is thought that there are cell-specific transcription factors that prepare sites for subsequent binding by receptors. Identifying these is an arduous informatics exercise, utilizing de novo motif discovery. One example of this is the AP-1 motif present in mammary cells. A c-Jun binding event is equivalent to local DNase hypersensitivity activity in mammary cells. In fact, there is approximately 40 percent overlap between c-Jun binding and GR binding in mammary cells. This is supported by functional evidence using dominant negative forms of AP-1.
Rapid dynamics of chromatin remodeling processes
The residence time of a receptor on a binding site is very short, involving seconds, rather than minutes or hours. "We feel this [rapid exchange] is very important to the way transcription biology works," emphasized Hager. This line of thinking led his group to propose a "hit-and-run" model, in which binding events are seen as very transient. When receptor is activated, the result is more frequent events, or hits. When the receptor is inactive, there are infrequent, non-productive hits. GR is not even in the nucleus in the absence of ligand, so very few of these events occur. "We argue 'hit-and-run' is a required mechanism for how GR-regulated transcriptional processes work," Hager concluded.
(Sophie Bolick, Ph.D., is a postdoctoral fellow with the Molecular and Genetic Epidemiology Group in the Laboratory of Molecular Carcinogenesis.)
Cell-specific Mechanism of GR Action
Hager's group set out to determine whether the SWI/SNF protein-remodeling complex was important for glucocorticoid receptor (GR) action. Surprisingly, knockdown of SWI/SNF activity completely obliterated constitutive hypersensitive sites in a given region, but the GR-induced locus was refractory to SWI/SNF knockdown. These surprising results led to the idea there are either GR-dependent de novo sites or constitutive sites, induced in the absence of GR. It was becoming clear that chromatin status of response elements is a major contributor to cell selective GR action at specific genes.
The researchers also used genome-wide mapping of DNase I hypersensitivity sites - a technical advance that has allowed for localized chromatin transitions to be mapped on a much larger scale than before. This is important, Hager explained, because it "gives you a powerful window on what's going on in terms of regulation with that particular cell type."
Data generated using this technology shows the receptor always binds at regions of hypersensitivity. When a mammary cell is compared to a pituitary cell, the hypersensitivity sites where receptor binding occurs are constitutively open in both cell types. However, in the instance of a gene active in a mammary cell line, a de novo event where chromatin is open on its own is present, creating a hypersensitivity site. This does not occur in the pituitary cell, so the receptor is unable to bind to chromatin.