19.2: Epigenetic Information in Nucleosomes

In order to fit two meters of DNA into a 5-20 μm diameter cell nucleus and arrange the DNA for easy access to transcriptional machinery, DNA is packaged into chromatin. Nucleosomes form the unit of this packaging. A nucleosome is composed of DNA approximately 150-200 bp long wrapped around an octamer consisting of two copies each of histone proteins H2A, H2B, H3, and H4 (and occasionally a linker histone H1 or H5). While the structure and importance of higher-level packaging of nucleosomes is less known, the lower-level arrangement and modification of nucleosomes is very important to transcriptional regulation and the development of different cell types. Histone proteins H3 and H4 are the most highly conserved proteins in the eukaryotic domain of life.

Nucleosomes encode epigenetic information in two main ways: chromatin accessibility and histone mod- ifications.

First, the nucleosomes’ positions on the DNA determine which parts of DNA are accessible. Nucleosomes are often positioned at the promoters of inactive genes. To initiate transcription of a gene, transcription factors (TFs) and the RNA polymerase complex have to bind to its promoter. Therefore, when a gene becomes active, the nucleosomes located at its promoter are often removed from the promoter to allow RNA polymerase to initiate transcription. Hence, nucleosome positioning on the DNA is stable, yet mutable. This property of stability and mutability is a prerequisite for any form of epigenetic information because cells need to maintain the identity of a particular cell type, yet still be able to change their epigenetic state to respond to environmental circumstances.

Chromatin accessibility can also be modulated by transcribed RNA (specifically, “enhancer RNA,” or eRNA) floating around the nucleus. In particular, Mousavi et al. found in 2013 that eRNAs, which are tran- scribed at extragenic enhancer regions, enhance RNA pol II occupancy (which is rate-limited by chromatin accessibility) and deployment of other transcriptional machinery, leading to enhanced expression of distal target genes [9].

Second, histones contain unstructured tails protruding from the globular core domains that comprise the nucleosome octamer. These tails can undergo post-translational modification such as methylation, acetyla- tion and phosphorylation, each of which affect gene expression. Some proteins involved in transcriptional regulation bind specifically to particular histone modifications or combinations of modifications, and recruit yet more transcription factors which enhance or repress expression of nearby genes. Thus, the “histone code hypothesis” posits that different combinations of histone modifications at specific genomic loci encode bio- logical function via differential transcriptional regulation. In this model, histone modifications are analogous to different readers marking sections of a book with different-colored post-it notes – histone modifications allow the same genome to be interpreted (i.e., transcribed) differently at different times and in different tissues. There are over 100 distinct histone modifications that have been found experimentally. Six of the most well-characterized histone modifications, along with the typical signature widths of their appearances in the genome and their putative associated regulatory elements, are listed in Table 19.1. Note that all of these modifications are on lysines in H3 and H4. Modifications of H3 and H4 are most well-characterized be- cause H3 and H4 are the most highly conserved histones (making modifications of those histones more likely to have conserved regulatory function) and because good antibodies exist for all of the commonly-observed modifications of those histones.

Histone modifications are so commonly-referenced that a shorthand has been developed to identify them. This shorthand consists of the name of the histone protein, the amino acid residue on its tail that has been modified and the type of modification made to this residue. To illustrate, the fourth residue from the N-terminus of histone H3, lysine, is often methylated at the promoters of active genes. This modification is described as H3K4me3 (if methylated thrice). The first part of the shorthand corresponds to the histone protein, in this case H3; K4 corresponds to the 4th residue from the end, in this case a lysine, and me3 corresponds to the actual modification, the addition of 3 methyl groups in this case.

Histone Modification	Signature	Associated Regulatory Element
H3K4me1	(wide) focal	active promoters / enhancers
H3K4me3	(wide) focal	active promoters / enhancers
H3K9me3	wide	repressed regions
H3K27ac	focal	active promoters / enhancers
H3K27me3	wide	repressed regions
H3K36me3	wide	transcribed regions

Table 19.1: Six of the most well-characterized histone modifications along with their typical signature widths and putative associated regulatory elements. “Focal” indicates that each instance of the histone modification has a relatively narrow signature in the genome (peak width < 5kb) whereas “wide” indicates wide signatures.

One example of epigenetic modifications influencing biological function is often seen in enhancer regions of the genome. Often these enhancer regions are far away from the genes and promoters that they regulate. The enhancer is able to come into contact with a specific promoter by histone modification (acetylation and methylation). This causes the DNA to fold upon itself to bring the promoter, enhancer, and recruited transcription factors into contact, activating the previously repressed promoter. This system can be very dynamic such that less than a minute after histone modification the cell will show signs of epigenetic influence, while other modifications (mainly those during development) will show themselves in a slower manner. This is also an example how how certain types of modifications of the histones can help us to predict enhancer regions.

It is possible for more than one histone modification to be present at a given genomic locus, and histone modifications thereby can act cooperatively and competitively. It is even possible for the two copies of a given histone protein within the same nucleosome to have different modifications (though usually the histone modification “writers” will localize together, thereby creating the same modification on both copies within the nucleosome). Thus, it is necessary to simultaneously take into account all histone modifications in a genomic region in order to accurately call the chromatin state of that region. As described in Section , with the completion of the Roadmap Epigenome Project in 2015, a robust hidden Markov model (with histone modifications as emissions and chromatin states as hidden states) can be used to do so.

Did You Know?

The simplest organisms that have epigenetic modifications are yeasts. Yeast is a single celled organism; thus, epigenetic modifications are not responsible for cell differentiation. As organisms become more complex they tend to have more epigenetic modifications.

Epigenetic Inheritance

The extent to which epigenetic/epigenomic features are heritable is poorly understood and is therefore the subject of much debate and ongoing investigation. In organisms that reproduce sexually, most epige- netic modifications are lost during meiosis and/or at fertilization, but some modifications are sometimes maintained. Additionally, biases exist in the ways in which paternal versus maternal epigenetic marks are removed or remodeled during this process. In particular, maternal DNA methylation is often retained at fertilization, whereas paternal DNA is almost always completely demethylated. Furthermore, for unknown reasons, some genomic elements, such as centromeric satellites, are more likely to evade epigenetic reset. In cases where epigenomic erasure does not occur completely at meiosis and fertilization, trans-generational epigenetic inheritance can occur. See generally [4].

Another mechanism likely to be governed by epigenetic inheritance is the phenomenon of parental imprinting. In parental imprinting, certain autosomal genes are expressed if and only if they are inherited from an individual’s mother, and other automsomal genes are expressed if and only if they are inherited from an individual’s father. Examples are the Igf2 gene in mice (only expressed if inherited from the father) and the H19 gene in mice (only expressed if inherited from the mother). There are no changes in the DNA sequence of these genes, but extra methyl groups are observed on certain nucleotides within the inactivated copy of the gene. The mechanisms and causality of this imprinting are poorly understood.