Histone Modification at Core of Epigenetic Control

Modifications to Histones Affect Gene Expression and Cell Fate Decisions

One of the most intriguing and complicated tasks that the genomes of all organisms have to accomplish is to find a strategy to be packaged into spaces that are many orders of magnitude smaller. For example, the genome of a somatic human cell is almost two meters long, but it has to be condensed into a nucleus about ten microns in diameter.

A strategy to accomplish this compaction relies on the organization of chromatin into protein-DNA complexes. Histones are a critical component of these complexes, and their post-translational modifications impart versatility to this structural and regulatory platform.

“Histones have a lot of intrinsic disorder, and are part of the dark proteome,” says Peter E. Wright, Ph.D., professor of integrative structural and computational biology at The Scripps Research Institute. A major challenge in proteomics stems from the fact that approximately one-third of the proteome is constituted of intrinsically disordered proteins and disordered protein regions. Some proteins are completely disordered proteins, and a larger fraction of proteins contain both ordered domains and disordered regions.

Such proteins can be studied using traditional techniques such as X-ray crystallography to determine the structure of their globular domains in isolation, and nuclear magnetic resonance spectroscopy to look at the disordered domains. “However, we need to understand how to do this holistically, and not by the current reductionist approach,” says Dr. Wright.

Current Characterization Approaches

Current approaches to study disordered proteins include additional techniques such as electron microscopy, mass spectrometry, single-molecule fluorescence resonance energy transfer (smFRET), and small-angle X-ray scattering. “I wish I saw a new technology on the horizon to solve the whole problem, but at the moment, it is a matter of applying all these different technologies synergistically to understand how these complex disordered proteins function,” concedes Dr. Wright.

The disordered regions of histones are the ones that tend to mostly harbor post-translational modifications. “Phosphorylation has been shown to predominantly occur in disordered regions, and acetylation is commonly found in disordered regions as well,” continues Dr. Wright, who along with colleagues, recently solved the crystal structure of the CREB-binding protein (CBP) catalytic core.

CBP contains a globular catalytic subunit and an approximately 70 amino-acid disordered region that plays a regulatory role. The studies revealed that several lysine residues within the autoregulatory loop of the histone acetyltransferase (HAT) domain are autoacetylated by the HAT domain. One of these acetylated residues, acetyl-K1596, subsequently binds intramolecularly to the bromodomain and inhibits acetylation of histone H3 in the nucleosome. “It was a surprise to see that the disordered loop could regulate histone acetylation by an intramolecular mechanism,” explains Dr. Wright.

Major efforts in the field are channeled toward harnessing histone post-translational modification therapeutically. “Acetylation is heavily regulated in the cell and there have been attempts to manipulate it by using acetyltransferase inhibitors and bromodomain inhibitors,” says Dr. Wright. The acetyltransferases and bromodomains, which write and read acetylation marks, have emerged as promising drug targets. “But first, we need to know a lot more about what the regulatory machinery is doing and how to manipulate it,” notes Dr. Wright.

“The overall concept is that histone deacetylase complexes and their associated co-repressors are functioning like molecular brake pads on a car,” says Marcelo A. Wood, Ph.D., professor and chair of the department of neurobiology and behavior at the University of California Irvine. In neurons that encode information for memory, these molecular brake pads are engaged and strong signaling is required to release them. When they are released at least temporarily, the resulting gene-expression changes may lead to neuronal changes, says Dr. Wood, “and this gives rise to long lasting changes in behavior.”

Role for Epigenetics

Studies from several experimental systems point towards epigenetic mechanisms as the molecular events that can lead to long-lasting changes in gene function and phenotypes. Epigenetic mechanisms are known to influence cell fate decisions, including cellular memory. “We reasoned that in neurons, changes via epigenetic mechanisms should give rise to long-lasting changes in plasticity and behavior,” says Dr. Wood.

Previously, Dr. Wood and colleagues demonstrated that in animals that undergo a sub-threshold learning event (which refers to a learning event that would not be encoded in short- or long-term memory), the inhibition of histone deacetylase complexes (HDACs) releases gene expression. This initiates signaling that allows a learning event to occur, leading to the formation of robust long-term memory. “We are very selective in what we remember, and [the inhibition of HDACs] could be a mechanism by which the brain is prevented from encoding everything that could become memory,” explains Dr. Wood.

For example, while most people would not remember many details of their everyday lives days later, most of them will vividly remember details of a near-hit car accident, including the location, the day, the time of the day, the street, and the color of the car. “This is because stress hormones that were released instantly disengaged the HDAC molecular brake pads and allowed unusual dynamics of gene expression to change neuronal plasticity in a way that is very persistent,” says Dr. Wood. Under normal conditions, HDACs suppress gene expression that would have effects on neuronal plasticity and long-lasting memory, but if a sub-threshold learning event happens when the HDAC is taken offline, that learning event can lead to the formation of a persistent memory.

These mechanisms could be engaged or disengaged inappropriately in maladaptive aspects of neuronal function. For example, reflects Dr. Wood, “Perhaps we could take advantage of this mechanism in addiction, and extinguish cocaine-associated memory by engaging extinction mechanisms of memory and make them more persistent by taking HDACs offline.”

In a recent animal study, Dr. Wood and colleagues showed that by manipulating HDAC3 in a fear circuit in the amygdala, it was possible to transform a sub-threshold learning event into a very persistent and robust fear memory. “It could be that in individuals who are more susceptible to PTSD [post-traumatic stress disorder] and fear-related memories, HDAC mechanisms are not as tightly engaged as they should be, so they encode these things much more easily and retrieve these memories maladaptively,” says Dr. Wood.

Conversely, people in whom HDAC and co-repressors are more tightly engaged might have a decreased susceptibility to PTSD. This individual variability might be explained by the epigenome. Genomic information, which is inherited, does not change over time, but the epigenome and the chromatin change. “The epigenome is dependent on many signaling cascades that get engaged by experiences such as stress, sleep, diet, and we can view it as a signal transduction platform that affects gene-expression profiles and explains the individual variability that we see,” according to Dr. Wood.

Alternative Splicing

“Alternative splicing intrigued me, because one of the ways cells define their identity during development is by alternative isoform selection,” says Elizabeth A. Heller, Ph.D., assistant professor of pharmacology and member of the Penn Epigenetics Institute at The University of Pennsylvania. Progenitor cells that differentiate into various cell types during development may accomplish this by stably expressing particular isoforms. “This is a mechanism that the body already uses to make permanent decisions about cell identity, so it has something stable about it.”

Findings from different labs have revealed that major changes in alternative splicing occur after cocaine exposure, and that chromatin changes may recruit the spliceosome as part of cell fate decisions. “We put all these findings together, and thought that there could be one intriguing mechanism, in that a cell, once exposed to stress or drugs, could have a change at the histone level that can permanently alter isoform expression,” says Dr. Heller.

In an unbiased approach that examined the potential relationship between histone modifications and alternative spliceoform expression, Dr. Heller and colleagues interrogated published data from a study on alternative splicing in brain tissue. “The same modification, trimethylated H3K36, was more relevant for predicting splicing and for being associated with spliceoform activity,” she notes.

A key focus in Dr. Heller’s lab is to understand whether chromatin directs splicing in brain tissue. Several brain-specific splicing factors have been described, and their dysfunction has been connected to various neuropsychiatric conditions. “Splicing is already known to be functional in the brain, and the chromatin-directed splicing is a new concept,” says Dr. Heller. To further examine chromatin-directed splicing in the brain, Dr. Heller and colleagues are generating and analyzing new brain tissue datasets.

“The main challenge when studying splicing in the brain is that neuronal tissue is highly heterogeneous, with thousands of neuronal cell types that are intermingled,” says Dr. Heller. In addition, hundreds of types of non-neuronal or glial cells are another critical component of the nervous tissue. Separating these cell types, each having their own transcriptome and spliceosome, is technically difficult. “We are constantly trying to figure out ways to obtain more homogeneous cellular populations,” continues Dr. Heller.

Fluorescence-activated cell sorting is used to separate many cell types. “But sorting is a physical process and neurons, which have long axons and complicated dendritic trees, typically do not sort well,” she points out. A new approach involves sorting only the neuronal nuclei. “But if we sort only nuclei, we are losing the cytoplasmic RNA, which is where much of the spliced RNA is found.”

At the Penn Epigenetics Institute, a laboratory led by Elizabeth Heller, Ph.D., studies show how remodeling of the epigenome can lead to aberrant neuronal gene function. For example, the lab has developed engineered transcription factors (ZFP- or dCas9-SET2) to deliver histone modifications to genes relevant to reward-related behaviors. This image by Carissa J. Lim, who is part of the Heller Lab team, shows how H3K36me3 is deposited at an exon-intron boundary to induce alternative splicing.

The Function of Histone Acetylation

“…[the] acetylation of newly synthesized histone lysine residues [has a direct function],” states Mark R. Parthun, Ph.D., professor and chair of biological chemistry and pharmacology at The Ohio State University. Work in Dr. Parthun’s lab focuses on histone acetylation during chromatin assembly. In the mid-1970s, it was discovered that histones are acetylated immediately after being synthesized. In particular, histone H4 is acetylated in a very conserved pattern, and diacetylation of lysines 5 and 12 has been reported in all eukaryotes examined to date. Subsequently, the residues become de-acetylated when histones are packaged into chromatin. “People have been trying to figure out why those molecules would be acetylated,” says Dr. Parthun. About 20 years ago, Dr. Parthun and investigators in Dr. Rolf Sternglanz’s laboratory from Stony Brook University discovered the enzyme that performs this acetylation. “We first discovered it in yeast, and then noticed that the yeast does not develop a phenotype when the enzyme is deleted,” explains Dr. Parthun. The lack of a phenotype opened up difficulties in understanding the function of the enzymes. “This led us to move to the mouse model, which biochemically is a better model system.”

Dr. Parthun and colleagues performed a quantitative proteomics analysis of newly replicated chromatin in cell lines from wild type and histone acetyltranferase 1 (Hat1) deletion-mutant mouse embryos. “We noticed that there are several bromodomain-containing proteins that are specifically depleted from nascent chromatin in the absence of Hat1,” says Dr. Parthun. These depleted proteins included Brg1, Baz1A, and Brd3, and it is thought that their depletion could alter the topology of nascent chromatin.

“Our work is looking at chromatin replication and inheritance during cell division and therefore, being able to develop techniques that follow chromatin dynamics—from when it first starts to replicate until it is mature—is critical,” Dr. Parthun concludes.

Researchers at Ohio State University led by Mark R. Parthun, Ph.D., have identified multiple roles for histone acetyltransferase 1 (HAT1) in replication-coupled chromatin assembly. For example, the researchers developed a model for how Hat1 remains associated with H3/H4 throughout the process of nuclear import and histone deposition. In this model, depicted here, the bromo- domains of Brg1 and Brd3 bind to newly assembled histone H4, helping to determine the structure and topology of newly assembled chromatin.