ound an octameric structure formed by core histones. Repeating nucleosomes assemble into higher-order structures stabilized by a linker histone. Chromatin modification involves the methylation of cytosine residues in DNA and the many ways in which the histone residues are modified. Histone modifications mostly targets the histone tails which extend from the core fold of the nucleosome and thus, are accessible for modifications and interactions. The modification pathways bring together the enzymes that catalyze the modification and those that catalyze its removal or change, known as 'writers' and 'erasers', respectively, the proteins whose particular domains recognize the modification, known as the 'readers' and the various other interactors or regulators. The remodeling of chromatin involves the ATP-dependent sliding, spacing and repositioning of entire nucleosomes, insertion, eviction or exchange of histones, chromatin assembly and disassembly. It is carried out by four protein complexes representing four remodeling pathways with distinct functions. The replication-dependent and independent assembly of nucleosomes is mediated by histone chaperone proteins. Epigenetic changes are heritable and independent of DNA sequence. However, DNA sequence variation can affect the function of players involved and thus, of epigenetic outcomes. An intimate dialog exists between the pathways of chromatin modification and remodeling to finely tune the epigenetic responses. The interaction of chromatin modifying enzymes and readers with non-coding RNA, particularly long non-coding RNA (lncRNA), further modulates the epigenetic landscape and the outcomes of regulated gene expression. Chromatin modifying enzymes and readers are target of drug development, primarily for cancer treatment. Several members are also found mutated in a number of human cancers. The histone modification pathway is presented here.

Histone modification
Post-translational modification (PMT) of histone tails involves a number of residues and is subject to a wide range of modification types. Primarily, the lysine residues are subject to some twelve types of modifications of which acetylation, methylation and ubiquitination are the better documented. The methylation of arginine and phosphorylation of several residues such as serine are also relatively well documented. Others PMT include sumoylation, crotorylation or ADP-ribosylation of lysines, deamination/citrullation of arginines, phosphorylation of tyrosines, threonines and histidines or glycosylation of serines and threonines. The recognition of modified residues by reader domains enables the recruitment of proteins and complexes resulting in changes of chromatin states. Only the better documented PMT types are described and shown on the diagram.

The histones
The histone families includes the four core histones and the histone linker; most families also contain variants. The core histones are represented by H2A, H2B, H3 and H4 which form the octameric structure wrapping DNA. The names are generic as multiple copies of highly similar genes are found clustered. Their expression is replication-dependent to assure the packaging of the newly synthesized genome during the S-phase of cell cycle. Histone variants present greater sequence variation but structurally adopt the same fold. The variants are replication-independent and they play important regulatory roles in chromosome segregation, DNA repair, meiotic recombination and transcription. The octameric structure involves a tetramer of H3-H4 flanked on either side by an H2A-H2B dimer. A histone linker, H1 binds the nucleosome to properly lock the DNA. Chaperones, found in the cytoplasm and nucleus regulate the flow and availability of histones and prevent non-specific interactions with cellular other proteins. They are also involved in modulating histone dynamics in response to DNA damage. Some chaperones have distinct specificity for binding to particular histones while others exhibit broader specificity.

Lysine acetylation, readers and deacetylases
Acetylation of lysine residues on histone tails neutralizes the positive charge of lysine and therefore weakens the electrostatic interaction with the negatively charged DNA promoting a more 'open' chromatin structure. Acetylation is carried out by acetyltansferases that fall into two classes: type-A which are nuclear and are further classified into three subclasses - GNAT, MYST and CPB/p300, and type-B that are cytoplasmic and modify free histones. The modified lysine are recognized by proteins with conserved binding motif known as bromodomain (BrD); most bromodomains are part of multi-domain proteins containing PHD (plant homeodomain), PWWP (conserved proline tryptophan) or WD40 repeats. The BrD fold is characterized by a four-helix bundle. Within the large BrD family some members are chromatin modifying enzymes or modulators, others are transcriptional regulators. Proteins with tandem PHD fingers or tandem PH (plekstrin homology) domains can also interact with acetylated lysines. A number of BrD proteins also interact with unmodified lysines. The deacetylases are grouped into four classes of which class I, II and IV are the classical, zinc-dependent enzymes and class III forms the NAD-dependent sirtuin family.

Lysine methylation, readers and demethylases
The lysine of histone tails can be mono- di- and tri-methylated. It is the most extensive modification type and provides an essential epigenetic mark. The methyltransferases (KMTs) possess a 130 amino acid enzymatic SET domain present in the larger of the two classes of enzymes while the second class, which does not have a SET domain, only has one member. Both classes utilize S-adenosylmethionine (SAM) intermediate generated in the methionine cycle metabolic pathway as the methyl group donor. Methylation of lysines K4, K9, K29 and K36 on H3 involves several protein complexes. COMPASS di- and tri-methylates K4 with activating functions and consists of enzymes of the Trithorax group (TrxG) and additional core and accessory subunits (six complexes in mammals). H3K4 trimethylation is dependent upon H2B monoubiquitination. The repressive polycomb complex 2 (PRC2) modifies K27 for which a mutually exclusive acetylation is reported. Phosphorylation of an adjacent serine displaces PRC2, promoting recruitment of activators. Other complexes modify K9 with mostly repressive outcomes and K36 that may affect splicing events. In addition to histone methylation, trithorax genes are also involved in ATP-dependent chromatin remodeling and sequence-dependent DNA binding. They antagonize the silencing action of polycomb group proteins (PcG). PcG also include the members of complex 1 (PRC1) which ubiquitinates H2A. Most of the TrxG and PcG proteins are conserved in mammals. Several protein domains are readers of methylated lysines and they include Tudor, chromo and PWWP (conserved proline tryptophan) which together are part of the 'Royal family', the PHD (plant homeodomain) finger and BAH (bromo-adjacent homology) domain, the MBT (malignant brain tumor) and Ankyrin repeats. In many cases, they combine several domains such as for instance, Tudor and PHD fingers. Of the first demethylases (KDMs) identified, Kdm1a was shown to have a flavin adenine dinucleotide (FAD)- dependent amine oxidase domain involved in the enzymatic activity. Other KDM classes rely on molecular oxygen, Fe(II) and alpha-ketoglutarate for activity and they contain the JmjC domain (Jumonji); in some cases the JmjN with JmjC is also observed. The KMTs and KDMs enzymes are highly specific in their reactions. Both classes of enzymes are represented by large families with over 30 KMT and over 15 KDM members, respectively. The specific methylation of lysine residues, the degree of methylation and the genomic distribution have important consequences and the patterns of histone methylation are described as 'chromatin states'. In humans, up to 51 such states are observed, sometimes in conjunction with other modifications and they describe promoter, exon, intron or splicing marks, activating or silencing nucleosomes as well as bivalent. Both enzyme types are themselves targets of post-translational modification to assure the proper control of their localization and activity. There is a complex interplay between histone lysine methylation and DNA methylation. Histone and modifying enzymes bind readers of methylated DNA. Some methylation enzymes, as members of ZF-CxxC (cysteine-rich zinc finger-CxxC) domain protein, are readers of non-methylated DNA.

Lysine ubiquitination and deubiquitinases
The mono-ubiquitination of lysine 119 of H2A is an abundant histone modification and is carried out by the Polycomb repressive complex 1 (PRC1). Polycomb group proteins (PcG) are essential epigenetic regulators with important roles in development, senescence and cancer. PcG complexes modulate chromatin structure and silence the expression of target genes. PRC1, together with the PRC2 complex involved in H3K4 methylation, promotes chromatin compaction; PRC1 also favors the formation of chromatin loops. The canonical PRC1 contains the RING1/2 E3 ligases, the polycomb (CBX), polycomb group factor (PCGF) and human polyhomeotic homolog (PHC known as HPH) proteins. Due to the presence of multiple isoforms - five CBX, six PCGF and three HPH, in addition to the two RING proteins, a diversity of complexes can be generated. The non-canonical complexes do not contain CBX proteins, instead they contain RYBP proteins that are mutually exclusive with CBX in PRC1 binding. The canonical complex is shown here. The identification of proteins that would interact with the modified H2AK119 continues to be a challenge. The deubiquitinases that have been identified include five members of the ubiquitin specific protease family (USP3, 12, 16, 21 and 44), ubiquitin C-terminal hydrolase BAP1 and metalloproteases MYSM1 and BRCC36.

Arginine methylation and readers
he methylation of arginine residues on histone tails includes three main types: monomethylation (MMA) and symmetric (SDMA) and asymmetric (ADMA) dimethylation. The mammalian arginine methyltransferases (PRMTs) are classified as type I, II and III. Type I and II catalyze the formation of an MMA intermediate which is further processed to ADMA by type I or to SDMA by type II PRMTs. Type III catalyze the formation of monomethylated arginines. A type IV enzyme has only been reported in yeast. The methyl mark on arginines is recognized by proteins with Tudor domains, others include proteins with chromo domains, WD40 repeats and PHD fingers; their effector function can be either repressive or activating. Many Tudor domain partners have been identified; many are non-histone proteins and the identity of interacting partners for the methylated arginines of histones remains to be unraveled. Less is known about the arginine demethylases but Jmjd6 (not shown) - a Jumonji domain-containing enzyme, has been reported to catalyze the reaction on dimethylated arginine residues and to carry out the reaction on both symmetric and asymmetric substrates. However, studies also show that Jmjd6 is in fact a lysine hydroxylase and the issue of arginine demethylases continues to be opened.

Phosphorylated histone readers
Serine, threonine and tyrosine residues on histone tails can be phosphorylated; several kinases are considered as possible modifiers which include those downstream of MAPK where they serve as a last, additional tier and possibly the protein kinase A or members of protein kinase C. At least for S28 on H3, adjacent to K27 which is tri-methylated by PRC2 complex and whose phosphorylation promotes PRC2 displacement and recruitment of activators, the kinases are MSK1 and MSK2 (not shown). The identity of phosphatase is even more elusive than that of kinases. However, several reader domains of phosphorylated histone residues have been identified and they include members of the 14-3-3 protein family and proteins with tandem BRCT (tumor breast cancer susceptibility) and BIR (baculovirus IAP repeat) domains.

To see the ontology report for annotations, GViewer and download, clock here...(less)