Histone modification is the covalent alteration of histone proteins, principally through the addition or removal of chemical groups at particular amino acid residues. Histones organize DNA into nucleosomes, the repeating units of chromatin. Their modifications influence how genetic material is packaged, recognized, transcribed, replicated, and repaired. Major forms include acetylation, methylation, phosphorylation, and ubiquitination; numerous less abundant modifications have also been identified.
These changes operate through two broad mechanisms: they can alter the physical properties of histones and their interactions with DNA, or provide recognition sites for other proteins. Their effects depend on the modified residue, the chemical group involved, neighboring modifications, and the biological setting. Histone modification is central to gene expression and epigenetics, but a modification is not necessarily an inherited epigenetic signal. Many marks are transient consequences of ongoing cellular processes, whereas others participate in mechanisms that maintain chromatin states through cell division.
Molecular substrates and chemical nomenclature
A canonical nucleosome contains approximately 147 base pairs of DNA wrapped around an octamer comprising two copies each of histones H2A, H2B, H3, and H4. The amino-terminal tails of these proteins extend from the nucleosome and contain many modification sites. Modifications also occur within the folded histone core, on carboxy-terminal regions, and on linker histone H1. Their positions matter: a modification near a DNA-contacting surface can affect nucleosome stability differently from one on a flexible tail that recruits a regulatory protein.
The conventional notation identifies the histone, the residue, its position, and the modification. H3K27ac denotes acetylation of lysine 27 of histone H3; H3K4me3 denotes trimethylation of lysine 4; H3S10ph denotes phosphorylation of serine 10. The abbreviations me1, me2, and me3 distinguish mono-, di-, and trimethylated lysine. Arginine can be monomethylated or dimethylated, with dimethylation occurring in symmetric or asymmetric configurations. Histone residue numbering conventionally excludes the initiator methionine removed during protein maturation.
The chemistry of each group determines its immediate effects. Lysine acetylation converts a positively charged amino group into a neutral amide, potentially weakening electrostatic interactions with negatively charged DNA. Lysine methylation generally preserves the residue’s positive charge but changes its shape and recognition properties. Phosphorylation introduces a negatively charged phosphate group, whereas ubiquitination attaches a small protein rather than a simple chemical group. These differences make “histone modification” a functional umbrella rather than a single biochemical mechanism.
Histone variants add another layer of specificity. H3.3 differs from canonical H3 at several residues and is deposited independently of DNA replication in many settings. H2A.Z, the DNA-damage-associated variant H2A.X, and the centromeric H3 variant CENP-A occupy specialized chromosomal environments. Variant identity can influence enzyme recruitment and modification patterns, but replacement by a variant is not itself normally classified as a covalent histone modification.
A nucleosome has two copies of each core histone, so a modification need not be present symmetrically. One H3 tail may carry a mark that the other lacks, and adjacent nucleosomes may differ substantially. Population measurements therefore describe distributions rather than a single uniform molecular state. Even apparently simple labels such as “H3K27me3-rich chromatin” encompass mixtures of modified and unmodified histones, different nucleosome arrangements, and proteins that interpret those arrangements.
Writers, erasers, and readers
Histone modifications are often described using a writer–eraser–reader framework. Writers are enzymes that install modifications; erasers remove them; readers recognize modified residues and connect them to downstream activities. This terminology is useful but not exhaustive. Some modifications lack a clearly established dedicated eraser, some recognition proteins bind unmodified residues, and many enzymes act on both histone and nonhistone substrates.
Writers obtain chemical groups from metabolic donors. Lysine acetyltransferases use acetyl-coenzyme A, while most histone methyltransferases use S-adenosylmethionine. Protein kinases transfer phosphate from adenosine triphosphate. Ubiquitination requires an activating enzyme, a conjugating enzyme, and usually a substrate-selecting ligase. Writer specificity depends not only on an enzyme’s catalytic site but also on its assembly into multiprotein complexes, contact with nucleosomal DNA, and interactions with previously modified histones.
Erasers employ distinct chemical strategies. Classical histone deacetylases hydrolyze acetyl groups through metal-dependent catalysis, whereas sirtuins consume nicotinamide adenine dinucleotide during deacylation. Lysine demethylases include flavin-dependent enzymes such as LSD1 and iron- and 2-oxoglutarate-dependent JmjC enzymes. Phosphatases remove phosphate, and deubiquitinases cleave the linkage between ubiquitin and its substrate. Removal rates can be as important as installation rates in determining a mark’s abundance.
Readers recognize modifications through specialized protein domains. Bromodomains commonly recognize acetylated lysine. Chromodomains, PHD fingers, Tudor domains, and related modules recognize particular methylated residues, often using aromatic side chains to accommodate methylated lysine or arginine. Specificity is contextual: a domain that recognizes H3K4me3 may also contact adjacent unmodified residues, and phosphorylation nearby can prevent binding. Recognition can depend on more than one domain engaging the same nucleosome.
These interactions allow self-reinforcing circuits. The EED subunit of Polycomb repressive complex 2 recognizes H3K27me3 and stimulates the complex’s methyltransferase activity. In many heterochromatic systems, HP1-family proteins recognize H3K9 methylation and help recruit additional silencing machinery. Such reader–writer coupling can stabilize a chromatin domain, although it does not establish unlimited spreading: boundaries, competing proteins, histone turnover, and demethylation constrain the process. Enzymes are also recruited by transcription factors, regulatory RNA, or DNA-binding complexes, so sequence-directed recruitment and modification-dependent feedback often work together.
Acetylation and other lysine acylations
Histone acetylation is widespread on the tails of H3 and H4 and is generally associated with accessible, transcriptionally active chromatin. Important acetyltransferase families include GNAT proteins such as GCN5, MYST-family enzymes such as MOF, and the related coactivators CBP and p300. These enzymes differ in substrate preferences and cellular functions. MOF, for example, is strongly associated with H4K16 acetylation, whereas CBP and p300 contribute extensively to acetylation at regulatory elements.
Acetylation can act directly on chromatin structure. H4K16 lies in a region of the H4 tail that participates in internucleosomal contacts. Its acetylation weakens particular interactions and can reduce the compaction of nucleosome arrays in biochemical experiments. This does not mean that every acetylated nucleosome becomes freely accessible: DNA sequence, nucleosome-remodeling enzymes, linker histones, and other chromatin proteins also determine accessibility.
A second mechanism is recruitment of acetyl-lysine readers. BET-family proteins, including BRD4, possess tandem bromodomains and can associate with acetylated chromatin. They connect chromatin recognition to transcriptional regulators, including factors involved in productive elongation. Acetylation can therefore both change local physical interactions and create binding opportunities. These mechanisms are complementary rather than mutually exclusive.
Specific acetylation marks have become useful genomic indicators. H3K27ac frequently distinguishes active enhancers from some inactive or poised regulatory regions. H3K9ac is common near active transcription start sites, while H4K16ac has prominent roles in chromatin architecture. None is an absolute activity label: regulatory elements can function with variable acetylation levels, and acetyltransferases may acetylate transcription factors or other proteins as part of the same response.
Lysine can also carry propionyl, butyryl, crotonyl, succinyl, malonyl, and lactyl groups. These modifications differ in size, hydrophobicity, geometry, and charge. Succinylation, for example, introduces a negatively charged group at a residue that was originally positively charged. Some established acetyltransferases can transfer alternative acyl groups under appropriate conditions, and some reader domains preferentially recognize particular acylations.
The connection to metabolism is important but requires careful interpretation. Acyl-coenzyme A availability can influence modification, yet whole-cell metabolite abundance does not necessarily represent the concentration accessible to a nuclear enzyme. Histone lactylation, reported in 2019, illustrates the expanding chemical diversity of the field; its enzymatic routes and functional significance must be established for each biological setting rather than inferred solely from the presence of lactate or a correlated transcriptional response.
Lysine and arginine methylation
Histone methylation differs fundamentally from a simple activating or repressing switch. Its consequences depend on the residue and methylation state. H3K4me3 is enriched near many active transcription start sites, H3K36me3 occurs across the bodies of actively transcribed genes, and H3K9me3 and H3K27me3 are associated with distinct forms of repression. H4K20 methylation contributes to several processes, including genome stability and the recognition of damaged chromatin.
Most lysine methyltransferases contain a SET catalytic domain, named after the founding proteins Su(var)3-9, Enhancer-of-zeste, and Trithorax. DOT1L is a major exception: it methylates H3K79 within the globular region of H3 through a structurally different catalytic domain. Mammalian SET1 and KMT2 complexes methylate H3K4, SETD2 produces H3K36me3, and SUV39-family enzymes contribute to H3K9me3. Their activity depends on nucleosome context and associated subunits, not merely recognition of a short histone peptide.
H3K27 methylation is catalyzed by Polycomb repressive complex 2, whose catalytic subunit is EZH1 or EZH2. The complex helps maintain repression of developmental regulators. H3K9 methylation is especially prominent in constitutive heterochromatin, including many repetitive regions. These systems are mechanistically distinct: H3K27me3 is strongly associated with regulated developmental silencing, whereas H3K9 methylation is frequently involved in repeat repression and specialized chromosome domains, although their distributions and functions can overlap.
Demethylation is also site- and state-specific. LSD1, identified as a histone demethylase in 2004, can remove mono- and dimethyl groups from H3K4 in appropriate complexes but cannot directly demethylate trimethyllysine through its flavin-dependent mechanism. JmjC-family demethylases use oxygen, iron, and 2-oxoglutarate; members of this family can remove trimethyl groups. Consequently, methylation dynamics can be sensitive to oxygen availability and metabolic conditions.
Arginine methylation is catalyzed by protein arginine methyltransferases. Type I enzymes produce asymmetric dimethylarginine, while type II enzymes, including PRMT5, produce symmetric dimethylarginine. Histone arginine methylation can influence transcription, nucleosome interactions, and recruitment of other proteins, but a universal activating or repressive classification is not applicable.
Methylation also illustrates modification cross-talk. Ubiquitination of H2B promotes efficient H3K4 and H3K79 methylation in well-studied systems. Conversely, H3K4me3 and H3K36 methylation can inhibit PRC2 activity on the relevant nucleosomal substrate. Such interactions help distinguish chromatin environments and prevent incompatible regulatory activities from occupying the same molecular substrate indiscriminately.
Phosphorylation, ubiquitination, and less common modifications
Histone phosphorylation often provides a rapid response to cell-cycle progression or DNA damage. H3S10 phosphorylation rises prominently during mitosis, and Aurora B kinase contributes to its deposition on mitotic chromosomes. The same residue can also be phosphorylated locally during stimulus-responsive transcription. Thus, a single chemical mark can occur in highly condensed chromosomes and at activated genes; its interpretation requires information about location, timing, and associated proteins.
A particularly well-characterized phosphorylation event is formation of γH2AX. After DNA double-strand breaks, kinases including ATM, ATR, and DNA-PK can phosphorylate serine 139 of the mammalian H2A.X variant. γH2AX develops across chromatin surrounding damage and recruits proteins such as MDC1, helping organize a larger signaling and repair domain. It is a sensitive indicator of DNA-damage signaling, but it is not a direct one-to-one count of double-strand breaks because replication stress and other conditions can also produce the signal.
Ubiquitination adds ubiquitin, a 76-amino-acid protein, usually through an isopeptide bond to a lysine side chain. Unlike many polyubiquitin signals that direct protein degradation, prominent histone monoubiquitination events regulate chromatin without destroying the modified histone. In mammals, H2AK119 ubiquitination is a major activity of Polycomb repressive complex 1 and contributes to Polycomb-associated repression and communication with PRC2.
H2BK120 ubiquitination in mammals, corresponding to H2BK123 in budding yeast, is associated with transcription and supports other histone modifications. During the DNA-damage response, RNF8- and RNF168-dependent ubiquitin signaling creates recruitment sites for repair factors. The exact lysine and ubiquitin arrangement matter: different ubiquitinated forms of H2A are not interchangeable simply because they carry the same modifier.
Other modifications broaden the chemical repertoire. SUMOylation attaches small ubiquitin-like modifiers and participates in several chromatin-associated pathways. ADP-ribosylation transfers ADP-ribose from NAD to proteins; histones are among the substrates of DNA-damage-responsive ADP-ribosylation systems. In mammals, HPF1 helps direct PARP1- and PARP2-dependent modification toward serine residues. Deimination by peptidylarginine deiminases converts arginine to citrulline, removing positive charge and changing interactions.
Not all modification cycles consist of simple addition and enzymatic removal. Histones can also undergo proteolytic clipping, in which a tail segment is removed, eliminating multiple potential modification sites at once. Such events occur in specific biological contexts and are not a general replacement for reversible modification. The diversity of these reactions makes residue-level identification essential: global labels such as “phosphorylated histones” conceal molecular signals with very different functions.
Transcription and chromatin organization
Histone modifications help coordinate transcription at promoters, enhancers, and gene bodies. Around an active promoter, acetylated histones and H3K4me3 frequently accompany transcription-factor binding and transcription-initiation machinery. Enhancers often show H3K4me1 and, when active, H3K27ac. These patterns are useful for identifying regulatory regions, but they are statistical associations rather than complete definitions of promoter or enhancer function.
Recruitment usually involves several mutually reinforcing steps. A sequence-specific transcription factor can bind DNA and recruit an acetyltransferase. Acetylated histones then help retain reader proteins and remodeling complexes, increasing opportunities for transcriptional machinery to engage. ATP-dependent chromatin remodelers reposition or remove nucleosomes; they alter chromatin organization without necessarily installing a covalent histone mark. Histone modification and remodeling are therefore distinct activities that frequently cooperate.
Modification continues during elongation. In mammals, SETD2 associates with elongating RNA polymerase II and deposits H3K36me3 across transcribed gene bodies. This mark can recruit proteins involved in chromatin restoration, RNA processing, and genome maintenance. In budding yeast, the related Set2 pathway recruits a histone deacetylase complex that suppresses inappropriate initiation within coding regions. A mark associated with active genes can therefore support local repression of unwanted transcription.
Repressive chromatin also has multiple architectures. H3K9-methylated regions recruit HP1-family proteins and commonly participate in the organization of repetitive DNA and chromosome-associated heterochromatin. Polycomb systems repress many developmental genes through H3K27 methylation, H2A ubiquitination, and additional activities of their protein complexes. Repression cannot be attributed to a single modification alone: some Polycomb effects depend on structural or recruitment functions distinct from catalytic activity.
Modification patterns influence, and are influenced by, three-dimensional chromosome organization. Repressed domains may interact preferentially with similar domains, while active regulatory elements associate with transcriptional machinery. However, chemical marks do not independently specify every chromosomal contact. DNA-binding architectural proteins, nuclear positioning, transcription, and nucleosome density also contribute.
A central interpretive distinction is whether a mark precedes transcription, reinforces it, or results from it. H3K36me3 is often deposited through transcription-coupled mechanisms, while acetylation can facilitate activation and also increase after regulatory complexes arrive. Genome-wide correlations alone cannot establish direction of causation. Mechanistic explanations therefore distinguish recruitment, catalytic action, reader binding, and transcriptional outcome rather than treating a chromatin signature as a self-sufficient instruction.
DNA replication, repair, and inheritance of chromatin states
During DNA replication, nucleosomes ahead of the replication fork are disrupted and reassembled on daughter DNA. Parental histones are redistributed, while newly synthesized histones fill the additional chromatin required after genome duplication. Histone chaperones and replication-associated proteins coordinate these processes. As a result, the density of many parental modifications initially falls, and chromatin maturation restores characteristic patterns over time.
New histones carry deposition-associated modifications. H4 acetylation at lysines 5 and 12 is a well-established feature of newly synthesized histones in many eukaryotic systems. H3K56 acetylation has prominent functions in newly assembled chromatin in budding yeast, although its abundance and biological importance differ substantially among organisms. These examples emphasize that histone modifications are not exclusively transcriptional signals: some indicate histone synthesis, transport, or assembly.
Parental marks can contribute to local restoration. If modified nucleosomes are redeposited near their former positions, reader–writer complexes can recognize them and modify neighboring new histones. This provides a mechanism for maintaining aspects of chromatin identity despite replication-associated dilution. Nevertheless, restoration also depends on continuing recruitment by DNA sequences, DNA-binding proteins, RNA, and other chromosomal features. Histone marks need not function as an autonomous copying system.
In DNA repair, modifications help expose damaged DNA and recruit pathway-specific factors. The protein 53BP1 recognizes a combination that includes H4K20me2 and damage-induced H2AK15 ubiquitination. Newly deposited histones after replication initially lack the same mature H4K20 methylation pattern, providing information about the replication state of chromatin. BRCA1–BARD1-associated mechanisms can use such information in the regulation of double-strand-break repair, linking histone age and modification status to repair-pathway choice.
Repair also requires restoration after the lesion is resolved. Phosphatases, deubiquitinases, histone exchange, and chromatin assembly remove or reset damage-associated signals. Persistence of a modification may reflect unresolved damage or delayed chromatin recovery rather than a stable memory intended for inheritance.
The term “epigenetic inheritance” therefore requires a defined timescale and experimental demonstration. Persistence through mitotic divisions differs from transmission through meiosis and across generations. Extensive chromatin remodeling during gamete formation and early embryogenesis limits direct extrapolation from somatic memory to transgenerational inheritance. Some organisms provide strong examples of inherited chromatin states, but the responsible molecular carriers and their dependence on continuing recruitment must be established separately for each system.
Development, cellular identity, and disease
Histone-modifying systems contribute to the stabilization and alteration of cellular identity during development. Cells sharing essentially the same genome express different sets of genes partly because transcription factors establish distinct chromatin environments. Histone modifications participate in these regulatory networks by helping maintain active programs, suppress alternative developmental programs, and coordinate responses to signaling. They do not replace sequence-specific regulation but provide additional control over its persistence and accessibility.
A notable developmental pattern is the “bivalent” domain, described in mammalian embryonic stem cells in 2006. These regions show enrichment for both H3K4me3 and H3K27me3, often around promoters of developmental regulators. Bivalency has been interpreted as a configuration compatible with repression while retaining regulatory flexibility. However, overlapping signals in a cell population do not automatically prove that both marks occupy the same nucleosome or the same allele in every cell. Molecular arrangement and functional significance require finer-resolution analysis.
Human disease can arise from alterations in histones themselves, their modifying enzymes, or their readers. Recurrent histone H3 mutations in some cancers demonstrate that a small change in a histone substrate can perturb an extensive regulatory system. H3K27M, found in a defined group of diffuse midline gliomas, is associated with disruption of PRC2-dependent H3K27 methylation. H3K36M mutations, characteristic of many chondroblastomas, interfere with H3K36 methylation pathways. Their effects extend beyond the small fraction of histones carrying the mutation.
Mutations in regulatory enzymes also show context-dependent consequences. Activating EZH2 mutations occur in subsets of B-cell lymphomas, whereas loss-of-function alterations in PRC2 components occur in other malignancies. Disruption of KMT2-family methyltransferases, acetyltransferases such as CREBBP and EP300, and chromatin-reader proteins can affect differentiation and transcription. Developmental disorders likewise result from pathogenic variants in genes encoding chromatin regulators, including the association of KMT2D or KDM6A variants with Kabuki syndrome.
These findings have supported pharmacological targeting of histone-regulatory machinery. Histone deacetylase inhibitors such as vorinostat and romidepsin have established uses in particular hematologic cancers. Other drug classes target methyltransferases or acetyl-lysine readers. Their biological effects cannot be assigned solely to histones, because many targeted enzymes also modify nonhistone proteins and participate in multiprotein complexes.
Disease-associated chromatin patterns can also reflect altered cell composition, proliferation, inflammation, or metabolism rather than a primary lesion in histone regulation. Distinguishing initiating mechanisms from downstream changes is essential when interpreting tumor profiles or developmental phenotypes. Genetic perturbation, biochemical characterization, and measurements in relevant cell populations are therefore complementary to descriptive modification maps.
Experimental measurement and causal testing
Histone modifications are studied through complementary biochemical, genomic, imaging, and structural methods. Mass spectrometry identifies modified peptides by their mass and fragmentation patterns and can estimate modification abundance. Because histones are rich in lysine and arginine, standard digestion can produce peptides too short for convenient analysis; chemical derivatization and alternative digestion strategies help overcome this problem. Accurate analysis must distinguish modifications with similar masses and localize them to the correct residue.
Bottom-up mass spectrometry analyzes short peptides, whereas middle-down and top-down approaches retain larger histone fragments or intact proteins. The latter approaches can reveal combinations of modifications on the same molecule that are lost when proteins are fragmented into short peptides. Their analytical complexity is greater, particularly when many closely related histone variants and modification states coexist.
Chromatin immunoprecipitation followed by sequencing, or ChIP-seq, maps DNA associated with a selected histone modification. An antibody enriches chromatin carrying the target epitope, and DNA sequencing identifies the associated genomic regions. Interpretation depends on antibody specificity, chromatin preparation, nucleosome abundance, and normalization. A rise in signal may reflect more modified histones, more nucleosomes, or a change in the proportions of cell types.
CUT&RUN uses an antibody-directed nuclease to release targeted chromatin fragments, whereas CUT&Tag uses tethered transposase activity to generate sequencing-ready fragments near the target. These methods can operate with relatively small samples and have single-cell adaptations. They nevertheless share an important limitation with ChIP-based methods: recognition reagents and sample processing can bias what is detected. Conventional population maps also do not directly reveal every modification combination present on individual nucleosomes.
Causality is tested by perturbing writers, erasers, readers, or histone residues. Acute protein degradation and selective chemical inhibition can separate immediate effects from long-term adaptation. CRISPR-based recruitment systems can direct modifying enzymes to selected loci, but artificial recruitment may act through associated proteins or nonhistone substrates as well as through the intended histone mark.
Residue substitutions require particular caution. Lysine-to-arginine substitutions prevent several lysine modifications while preserving positive charge, but they do not reproduce unmodified lysine exactly. Lysine-to-glutamine is often used as an acetylation mimic, yet it lacks the precise chemistry and reader-binding properties of acetyllysine. Semisynthetic histones and defined nucleosomes containing authentic modifications provide stronger tests of direct molecular effects. Structural methods, including X-ray crystallography and cryo-electron microscopy, then reveal how enzymes and readers engage those substrates.
Historical development and limits of the histone-code model
The modern study of histone modification grew from biochemical work on nuclear proteins rather than from genome-wide sequencing. In 1964, Vincent Allfrey, Alfred Mirsky, and colleagues reported histone acetylation and methylation and proposed a relationship between acetylation and the regulation of RNA synthesis. This work established an influential connection between chemical changes in histones and gene activity before the modern nucleosome model had been fully developed.
The identification of histone-modifying enzymes in the 1990s connected chromatin biochemistry directly to transcriptional regulation. In 1996, work on the Tetrahymena acetyltransferase p55 linked it to the yeast transcriptional regulator Gcn5. In the same period, identification of a mammalian histone deacetylase related to the yeast repression factor Rpd3 showed that both acetylation and deacetylation were tied to established gene-regulatory machinery. These discoveries helped replace the view of histones as largely passive packaging proteins with a mechanistic account of regulated chromatin.
Further breakthroughs included identification of histone lysine methyltransferases and modification-recognition domains. Discovery of LSD1-mediated demethylation in 2004 demonstrated that lysine methylation could be enzymatically reversible, overturning a widespread assumption that removal largely required histone replacement. Genome-wide mapping subsequently revealed reproducible distributions of modifications across promoters, enhancers, gene bodies, and repressed domains.
The histone-code hypothesis, prominently articulated around 2000–2001 by Brian Strahl, C. David Allis, Thomas Jenuwein, and others, proposed that combinations of histone modifications are interpreted by binding proteins to produce distinct biological outcomes. Its lasting contribution was to emphasize combinatorial recognition rather than one-mark–one-function explanations. Reader domains and modification cross-talk supplied concrete molecular mechanisms for parts of this proposal.
The word “code,” however, does not imply a fixed dictionary comparable to the genetic code. The same mark can have different consequences in different organisms, cell types, or genomic locations. Many combinations are constrained by enzyme specificity or mutual exclusion: a single lysine cannot be both acetylated and methylated at the same time, although those states can occur on different histone copies within the same region.
Current interpretation therefore combines chemical specificity with cellular context. A modification can be a causal regulator, a permissive feature, a feedback signal, or a consequence of another process. Distinguishing these roles requires attention to timing, molecular occupancy, enzyme recruitment, and the properties of the relevant reader complexes. Increasingly precise measurements address not only where marks occur, but which combinations coexist on individual nucleosomes and which molecular events those combinations actually control.