in the epigenetic regulation of gene...

The expansion of eukaryotic genomes has been accompa-nied by the use of packaging proteins in the regulation of gene expression1. Octamers of four histones — H2A, H2B, H3 and H4 — densely package eukaryotic genomes into nucleosomal units that must be mobilized for genes to be expressed. In addition, histones and their modifications, along with transcription factors, DNA methylation and small interfering RNAs, have been implicated in epige-netic memory — the propagation of the activity of a gene from mother to daughter cells. In this capacity, histones and other chromatin components have acquired active roles in developmental regulation; mutations in chroma-tin components can cause loss of epigenetic memory2,3, and the genetic dissection of developmental memory has led to the identification of many chromatin regulators4. However, the mechanisms whereby chromatin-related processes affect epigenetic regulation are not well understood.

Most discussions on the role of chromatin in epige-netic processes have focused on histone modifications and their regulation5. However, there is a growing appre-ciation of roles for other chromatin-based processes in epigenetic regulation, including the deposition of histone variants and nucleosome remodelling6. The aim of this Review is to explore the possible relationship between nucleosome dynamics and epigenetic memory, and the

molecular mechanisms that might be responsible for the link between the two. In particular, recent progress in understanding the biochemical activities of genetically defined regulators of chromatin, combined with genome-wide mapping, has led to the realization that maintaining regulatory sites in an active chromatin configuration is crucial for epigenetic inheritance. It now appears that multiple processes contribute to the maintenance of active chromatin, including the nucleosome-positioning potential of regulatory DNA, the assembly and remodel-ling of nucleosomes, the incorporation of histone vari-ants, the modification of histone tails and the binding of chromatin-associated proteins7–10. A unifying theme based on recent studies of nucleosome stability and histone replacement is that epigenetic memory of gene expres-sion is maintained by the continuous destabilization of nucleosomes. Many observations that have been inter-preted in terms of stable chromatin signatures, including histone modifications and variants as well as chromatin- associated proteins, will be reconsidered in the light of dynamic processes that maintain active chromatin.

Role of chromatin in epigenetic regulationThe idea that regions of the genome are differentiated at the level of chromatin and that alternative chromatin states underlie epigenetic regulation originated with the

Howard Hughes Medical Institute, Basic Sciences Division, Fred Hutchinson Cancer Research Center, 1100 Fairview Avenue North, Seattle, Washington 98109, USA.e‑mail: [email protected]:10.1038/nrg2206Published online 4 December 2007

Nucleosome destabilization in the epigenetic regulation of gene expressionSteven Henikoff

Abstract | Assembly, mobilization and disassembly of nucleosomes can influence the regulation of gene expression and other processes that act on eukaryotic DNA. Distinct nucleosome-assembly pathways deposit dimeric subunits behind the replication fork or at sites of active processes that mobilize pre-existing nucleosomes. Replication-coupled nucleosome assembly appears to be the default process that maintains silent chromatin, counteracted by active processes that destabilize nucleosomes. Nucleosome stability is regulated by the combined effects of nucleosome-positioning sequences, histone chaperones, ATP-dependent nucleosome remodellers, post-translational modifications and histone variants. Recent studies suggest that histone turnover helps to maintain continuous access to sequence-specific DNA-binding proteins that regulate epigenetic inheritance, providing a dynamic alternative to histone-marking models for the propagation of active chromatin.

NATuRe RevIews | genetics vOluMe 9 | jANuARy 2008 | 15

REVIEWS

© 2008 Nature Publishing Group

http://ca.expasy.org/uniprot/Q96QV6

http://ca.expasy.org/uniprot/Q96A08

http://ca.expasy.org/uniprot/Q6NXT2

http://ca.expasy.org/uniprot/P62805

mailto:[email protected]

Position-effect variegationThe variable, heritable silencing of genes by their juxtaposition to heterochromatin, or by movement of a gene into a different nuclear domain or chromosomal context.

study of position-effect variegation (Pev) in Drosophila melanogaster11. In Pev, juxtaposition of a gene to hetero-chromatin leads to heritable silencing. The traditional view of Pev has been that genes that are silenced by heterochromatin are compacted; however, cytological studies using GFP to monitor the mobility of heterochro-matin have led to the realization that silent chromatin is dynamic on a short timescale, albeit with a consist-ently reduced accessibility to transcription factors12–14. Consequently, epigenetic silencing of gene expression might be understood in terms of factors that affect nucleosome mobility, such as nucleosome-remodelling complexes and histone modifications, variants and chaperones.

Compelling evidence that nucleosomes and their modifications are directly involved in epigenetic processes comes from studies of Polycomb group (PcG) proteins in D. melanogaster15. A large body of work has implicated these chromatin-associated proteins in regulating the on versus off states of homeotic genes, which are the master regulators of development along the anterior–posterior axis. The specific mechanisms by which PcGs maintain the silent state across cell divisions is becoming increas-ingly well understood, and there has been substantial progress in characterizing PcGs both at the biochemical level and in vivo (BOX 1).

silencing by PcGs during development is prevented by the action of proteins in the trithorax group (trxG), which maintain transcriptional activity10,15. The silent state appears to be the default for PcG-regulated genes, whereas trxG proteins are needed to antagonize this state to provide an epigenetic memory of gene activation16. Proteins belonging to the trxG include motif-specific DNA-binding proteins (for example, GAGA factor (GAF) and Zeste), nucleosome remodellers (for exam-ple, Brahma and Kismet), a histone H3 lysine 4 (H3K4) methyltransferase (Trithorax) and an H3K4 demethylase (lid)17,18. This diversity in trxG protein function also seems to extend to vertebrates, in which analogous DNA-binding proteins and homologous nucleosome remodellers, H3K4 methyltransferases and H3K4 demethylases, seem to have comparable roles in devel-opmental regulation10,19,20. Given the limits of genetic analysis, there might well be other classes of trxGs that have not been identified in traditional screens, but which are intimately involved in maintaining active chromatin. For example, a histone H3–H4 chaperone, anti-silencing function protein 1 (Asf1), is encoded by a suppressor of Pev and interacts with Brahma in vivo and in vitro21. Also, the histone variant H2Av behaves genetically as a PcG in D. melanogaster22. These diverse observa-tions provide intriguing hints as to how trxGs might counteract the silencing effects of PcGs (BOX 1).

PcG and trxG proteins have become paradigms for understanding epigenetic regulation throughout devel-opment, but questions remain as to how they act to promote ‘cellular memory’. As described below, studies of the histone substrates of chromatin regulators lead to a convergence of genetic, biochemical and genomic evi-dence that suggests that nucleosome dynamics underlies epigenetic regulation of eukaryotic genes.

Nucleosome eviction at gene regulatory sitesA key insight into the role of chromatin in gene regula-tion was the realization that transcription factors bind to sites that are largely cleared of nucleosomes. This was shown directly for the yeast PHO5 promoter23,24 and extended by genome-wide studies in yeast and multi-cellular eukaryotes25–27. These findings agree with the well-known correlation between DNase I hypersensitive sites (‘open’ chromatin) and gene regulatory sequences28,29 as reflecting a local depletion of nucleosomes at sites of transcription factor binding. Current evidence favours nucleosome eviction — rather than stepwise loss of histones or passive dilution at replication — as the main

Box 1 | Epigenetic regulation by Polycomb and trithorax group proteins

Polycomb group proteins (PcGs) are responsible for maintaining the silent state of homeotic genes, and trithorax group proteins (trxGs) for maintaining their active state. This is the case not only in insects, but also in vertebrates, where PcG proteins also have important roles in stem-cell differentiation10,15. Some PcGs and trxGs are general chromatin regulators that act throughout the genome. For example, PcGs act partly by methylating histone H3 on lysine 27 (H3K27) through the enzymatic action of the Enhancer-of-zeste subunit of the Polycomb repressive complex 2 (PRC2) and partly by the binding of the PRC1 and Pleiohomeotic (PHO) complexes. Although these complexes do not appear to remain stably bound to their sites of action during development110, H3K27 methylation persists, and this modification helps to maintain the silencing throughout development79,80. Plants and nematodes also use the PRC2 complex for maintaining the silencing of imprinted genes, but these organisms lack components of PRC1 or PHO111. This strongly suggests that developmental silencing is inherited at least in part by H3K27 methylation, even in organisms that have evolved multicellularity independently. Another histone modification, methylation of histone H4 lysine 20 (H4K20), also appears to be a universal component of epigenetic regulation during development112.

Although the important roles of H3K27 and H4K20 methylation in gene silencing are clear, whether these modifications are themselves involved in epigenetic inheritance of the silent state is not. Evidence from Drosophila melanogaster genetic studies indicates that the silent state is the default, and that this is counteracted by the action of trxGs at specific sites15,16. In the case of pericentric heterochromatin — which forms at tandem repeats — H3K9 methylation has a silencing role analogous to that of H3K27 methylation, although in the case of H3K9 silent chromatin seems to be the permanent default state113. It seems likely that these silencing modifications help to maintain the vast majority of the genome relatively inaccessible to DNA-binding proteins, thus facilitating gene regulation by reducing the amount of the genome that a transcription factor can interact with. Only sites that are made accessible by the action of trxGs would be available for transcription factor action; epigenetic regulation of gene expression would thus result from the early action of transcription factors to make specific sites available to trxGs. As is the case with PRC2, trxGs act by altering chromatin, and their association with DNase I hypersensitive sites suggests that they do so by making chromatin continuously accessible to transcription factors10,17.

Even though PcGs and trxGs promote opposite outcomes, they bind to the same sites, called PREs or PRE/TREs (Polycomb response elements and trithorax response elements). PREs can lie tens of kilobases away from the homeotic gene promoters that they regulate4 — evidently by looping114 — to maintain the active or silent state of reporter genes throughout D. melanogaster development. Binding of PREs to PcGs and trxGs occurs regardless of whether the homeotic gene promoter that is regulated is active or silent80. The fact that well-mapped PREs span only a few nucleosomes challenges models for epigenetic inheritance that are based on distributing nucleosomes at replication5. With random partitioning, epigenetic inheritance would be completely lost over a span of three nucleosomes with a probability of (1/2)3 — that is, from one in eight progeny cells — and non-random partitioning would cause loss of epigenetic memory even more frequently74. An alternative possibility that nucleosomes split at replication has been extensively tested and yet no compelling evidence could be found73.

R E V I E W S

16 | jANuARy 2008 | vOluMe 9 www.nature.com/reviews/genetics


http://ca.expasy.org/uniprot/Q08605



http://ca.expasy.org/uniprot/Q9Y1L3


http://ca.expasy.org/uniprot/Q9VMJ7

http://ca.expasy.org/uniprot/Q9V464


Satellite DNA Various classes of highly repetitive DNA that are tandemly repeated and are most often associated with centromeric or pericentromeric regions of the genome; α-satellite DNA is the primate centromere-specific satellite in which the monomeric unit is 171 bp.

mechanism by which nucleosomes are depleted6,30–32, and multiple factors participate in the eviction process by destabilizing nucleosomes.

DNA-sequence determinants of nucleosome stability. DNA sequence is likely to be an important factor in favouring or disfavouring nucleosome eviction, because nucleosomes show distinct DNA-sequence preferences for stable assembly33,34. using in vitro selection, it has been possible to evolve sequences that position nucleo-somes far more stably than in nature, which implies that nucleosome positions have evolved to be metast-able35. Nucleosomes manifest DNA compositional preferences that are especially evident at promoters. For example, high-resolution mapping of nucleosomes sur-rounding yeast promoters has detected preferred posi-tions, with DNA-binding proteins just on the edge of the first nucleosome downstream of the transcriptional start site36. yeast promoters tend to lie adjacent to sites that are predicted to stably position nucleosomes, which would place them in regions of low nucleosome occupancy7. sequence preferences determined in vitro have been used to predict nucleosome positions with a success rate of approximately 50%, compared with around 30% expected by chance7. These observations suggest that nucleosome stability is a major driving force in the evolution of genomes, much in the way that translational efficiency underlies codon biases. It seems likely that nucleosome-positioning sequences can dominate nucleosome spac-ing in particular situations, because nucleosomes are sometimes found to be spaced in satellite DNA with lengths that are identical to the satellite repeat unit length37. However, at regulatory elements, nucleosome- positioning sequences would only contribute to the probability that a site will be occupied by a nucleosome; processes governed by the action of DNA-binding proteins and nucleosome-remodelling complexes are likely to be more important for nucleosome retention38.

Chromatin-assembly proteins that influence nucleosome stability. The stability of nucleosomes can be influenced by processes that occur during their assembly or remod-elling. Nucleosomes are initially positioned during the first step of replication-coupled nucleosome assembly. Assembly is carried out by the chromatin assembly fac-tor 1 (CAF1) chaperone complex, which is tethered to the replication processivity clamp (PCNA)2. Two H3–H4 dimers are deposited behind the replication fork by CAF1, presumably without sequence preference39. In budding yeast, the Isw2 ATP-dependent nucleosome-remodelling complex seems to move nucleosomes from their energetically favoured positions to positions that can impede transcriptional activation40. The in vitro activities of this class of nucleosome remodellers suggest that they have a general role in repositioning nucleo-somes to restrict promoter activity33,41. Nucleosome positions, and therefore stability, are also influenced by whether or not the H1 linker histone is deposited, with a decrease in the average nucleosome spacing when H1 is depleted42. H1 affects the higher-order packag-ing of nucleosomes, although is not known whether

this role influences nucleosome stability. Other linker proteins, such as high mobility group (HMG) proteins and poly-ADP ribose polymerase 1 (PARP1), might also affect nucleosome stability through their roles in spacing nucleosomes43,44.

Recent structural and biochemical studies have elucidated the roles of the histone escort protein Asf1 in nucleosome stability45–47. Binding of Asf1 to the C-terminal domain of H3 forms a soluble, trimolecular Asf1–H3–H4 complex, suggesting that one role of Asf1 is to escort newly synthesized H3–H4 dimers to sites of assembly. Binding of Asf1 is thought to prevent pre-mature self-association of soluble H3–H4 into tetramers by directly blocking C-terminal residues that form the stable H3–H3 four-helix bundle in nucleosomes. yeast Asf1 also acts as a nucleosome disassembly factor in vitro and in vivo to evict nucleosomes from promoters for gene activation48,49. A model for disassembly catalysed by Asf1 is that, after H2A–H2B dimers are removed, Asf1 binds to the exposed H4 C-terminus and under-goes a conformational transition that effectively pries the H3–H4 dimer off the DNA, causing it to break the H3–H3 four-helix bundle that holds together the two halves of the nucleosome46 (FIG. 1).

In yeast, Asf1 also facilitates the acetylation of histone H3 lysine 56 (H3K56) by the unusual histone acetyl-transferase Rtt109 (ReF. 50). The soluble pre-assembly complex of Asf1–H3–H4–Rtt109 transiently acetylates H3K56 for replication-coupled nucleosome assembly51 and is associated with elongating RNA polymerase52. The link between H3K56 acetylation and Asf1-dependent processes suggests a direct role for this modification in assembly–disassembly reactions. Indeed, in yeast, Asf1-dependent H3K56 acetylation is tightly correlated with replication-independent histone replacement genome-wide32. An attractive possibility for how this modifica-tion influences nucleosomal stability is suggested by the location of H3K56 on the nucleosomal surface where the DNA gyres exit: the acetate group would interfere with the wrapping of DNA around the nucleosomal octamer53, thus destabilizing the nucleosome.

The proposed Asf1-catalysed disassembly reaction is expected to be reversible, and it is likely that ATP is hydrolysed to prevent immediate reassembly. The best candidates for providing the energy for this process are ATP-dependent nucleosome remodellers54. In vitro, these machines can slide nucleosomes and, in the pres-ence of a histone chaperone, can sometimes evict them. For example, the yeast RsC chromatin-remodelling complex can evict nucleosomes in the presence of the nucleosome-assembly protein 1 (NAP1) histone chaper-one55. In vivo, RsC is highly abundant, as are a number of animal nucleosome remodellers including Brahma, a trxG protein that is homologous with RsC56 and inter-acts with Asf1 (ReF. 21). An attractive possibility is that these abundant nucleosome remodellers, together with Asf1 and perhaps other chaperones, catalyse the eviction of nucleosomes, thereby allowing transcription factors to gain access to their binding sites55,57,58.

Another H3–H4 chaperone that has been impli-cated in epigenetic processes is RbAp48 (also known as

R E V I E W S






MsI1), a component of diverse nucleosome-assembly complexes as well as the Polycomb repressive complex 2 (PRC2) (BOX 1), the NuRF nucleosome remodeller and other chromatin-associated complexes3. Remarkably, reduced levels of Arabidopsis thaliana MsI1 cause progressive loss of memory of the activation state of genes involved in floral meristematic identity. Because RbAp48 binds a part of H4 helix 1 that is not accessible in the nucleosome, it is possible that RbAp48-containing complexes act on partially disassembled nucleosomes during dynamic processes such as replication and transcription. Although the mechanism of action of RbAp48 is unknown, it might mediate the continuous turnover of histones at sites of active chromatin, as described below.

Histone replacement in epigenetic regulationDifferentiation of chromatin by histone variants. Nucleosome stability can be affected by the incorpora-tion of histone variants, which differentiate chromatin into regions with distinct properties9 (TABLe 1). The most conspicuous distinction is that between the kine-tochore and chromosome arms, which depends on the centromere-specific histone H3 variant, CenH3. Nucleosomes containing CenH3 are epigenetically maintained, which raises the question of whether its counterpart on chromosome arms — the universal replacement variant, H3.3 — is the central player in maintaining epigenetic inheritance at genes59.

H3.3 is deposited throughout the cell cycle, unlike H3, which is the substrate for CAF1 and is incorporated at the replication fork. Only four amino acids distinguish H3.3 from H3, of which three core residues prevent H3 from being deposited outside of replication. These core residues are exposed in the soluble Asf1–H3–H4 escort complex, which presents H3–H4 and H3.3–H4 dimers to the CAF1 and HirA nucleosome-assembly complexes, respectively47. Importantly, deposition of H3.3 occurs primarily in transcriptionally active chromatin and gene regulatory sites25,58,60–63 (FIGS 2,3).

Although few encoded differences distinguish H3.3 from H3, there are major differences in their modifi-cation patterns in vivo. Both plant and animal H3.3 proteins are enriched in modifications associated with active chromatin, including di- and trimethylation of K4 (H3K4me2 and H3K4me3), dimethylation of K79 (H3K79me2) and acetylation of multiple lysines, and are depleted for ‘silent’ modifications, such as H3K9me2 (ReFS 64–67). These global differences are consistent with the distinctive localization patterns seen for H3.3 relative to H3; active genes and regulatory regions are enriched in both H3.3 and active histone modifications, whereas silent chromatin is depleted for both25,60,62,68–70 (FIG. 2a). Removal of the N-terminal tail has no discernable effect on H3.3 incorporation25,59, which indicates that tail modifications are not responsible for localization of H3.3. Rather, as described below, dynamic processes that act on nucleosomes can lead to incorporation of H3.3 and may contribute to the enrichment of active tail modifications.

Rapid histone turnover at yeast regulatory elements. Histone replacement implies turnover, because inser-tion of new histones on a chromatinized template requires that pre-existing histones are lost. Because the dyad axis of the nucleosome lies at the H3 dimeriza-tion interface, replacement of an H3 histone involves unravelling of the DNA from at least halfway around the nucleosome core. The actual rate of turnover that results from unravelling and histone replacement could not be determined in the animal cell systems used in studies that profiled H3.3 and H3 levels at active genes and regulatory elements. However, turnover could be measured directly in yeast using two different tagged H3 constructs, one produced constitutively and the other by transcriptional activation71. By measuring their rela-tive concentrations over a time course using chromatin immunoprecipitation, turnover rate profiles could be determined. Promoters are ‘hot’ in that they have a high rate of turnover and genes become progressively ‘colder’

H4

Nature Reviews | Genetics

Asf1

b

a

H3 CAF1

H4Asf1

H3

Figure 1 | Asf1isinvolvedinbothnucleosomeassemblyanddisassembly.a | Anti-silencing function protein 1 (Asf1) is a nucleosome escort and disassembly factor. Asf1 functions in nucleosome assembly by depositing an H3–H4 histone dimer into DNA. b |In nucleosome disassembly, Asf1 pries H3–H4 dimers off DNA. CAF1, chromatin assembly factor 1.

R E V I E W S




http://www.expasy.org/uniprot/Q8RVQ9

http://www.expasy.org/uniprot/P84243

downstream30–32 (FIG. 3a). In D. melanogaster, a similar 5′ to 3′ gradient of H3.3 (and RNA polymerase II) is seen over gene bodies, such that the 3′ ends of long genes are colder on average than non-genic regions25. Highly transcribed genes in yeast also show enhanced levels of turnover30, consistent with the enhanced levels of H3.3 seen for highly transcribed genes in D. melanogaster 25,68 (FIG. 2b,c). yeast use both the CAF1 and HirA assembly pathways, but they have only one type of H3, which is phylogenetically classified as an H3.3 — ascomycetes have evidently lost H3 after divergence from basidi-omycetes59. Therefore, the concordance between animal and yeast studies implies that the difference between conserved assembly pathways, not between H3.3 and H3 variants, results in hot versus cold nucleosomes.

How is active chromatin inherited? High histone turnover at promoters of active genes might help to explain how an active chromatin state can be propagated from mother to daughter cells. More than 30 years ago it was proposed that the nucleosome splits along the H3–H3 dimerization interface at replication, and the resulting half-nucleosomes template the addition of new half-nucleosomes72. However, a substantial body of work has failed to find evidence that nucleosomes ever split in half at replication; instead, (H3–H4)2 tetramers are found to be inherited intact73,74. Furthermore, no mechanism is known that can replicate a modification or conformational state between halves of the nucleosome75. Nevertheless, semi-conservative replication of nucleosomes remains a possi-bility for the small subset of regulatory sites that transmit epigenetic memory45,76.

An alternative possibility is that a process of histone turnover at regulatory elements perpetuates itself, thus maintaining chromatin in a constitutively active state74. The silent state would be the default: CAF1, together with Asf1, would assemble H3 nucleosomes at replication that are enriched in silent modifications and deficient in active modifications. Conversely, replication-independent incorporation of H3.3 by other chaperones, such as HirA,

and disassembly by Asf1 would occur at sites of tran-scriptionally active chromatin and regulatory elements, a process that is set in motion by the action of transcription factors. Over the course of a cell cycle, actively modified H3.3 would accumulate at active genes and regulatory ele-ments, and the random partitioning to daughter chroma-tids would favour perpetuation of the active state. efficient CAF1-dependent assembly behind the replication fork would help to perpetuate the silent state over broad regions, whereas the local turnover process that results in histone replacement would cause H3K27 methylation to be lost, thus counteracting silent chromatin.

Turnover at promoters and regulatory elements is most likely driven by ATP-dependent nucleosome remodel-lers57,58. Disruption during remodelling would occasion-ally evict nucleosomes, transiently exposing DNA and allowing DNA-binding proteins and PcGs to find their sites (FIG. 4). The continued local presence of nucleosome remodellers would result in another cycle of remodelling, nucleosome depletion and histone replacement at the site. This dynamic process was proposed to account for the surprisingly short residence times of transcription factors in vivo77,78. It could result in reduced nucleosome density and DNase I hypersensitivity58, especially if replacement of an evicted nucleosome is a slow process. Consistent with this proposal, Polycomb response elements (PRes) are deficient in nucleosomes relative to flanking regions79,80, and this relationship holds for sites of PcG and trxG pro-tein binding genome-wide58 (FIG. 3b). Thus, peaks of H3.3 correspond on average to dips in nucleosome occupancy, a relationship that would be difficult to explain unless histones are continuously turning over at sites of regula-tory protein binding. In this way, a continuous process of histone turnover would maintain PRes, promoters and other gene regulatory elements in an accessible configura-tion by continually mobilizing nucleosomes. several lines of evidence support this mode of epigenetic inheritance.

First, one expectation for any hypothesis that accounts for epigenetic inheritance is that the proposed proc-ess should take place at sites that are responsible for

Table 1 | Histone variants: functions and assembly

Histone Description/functions escortand/orchaperonecomplex

Refs

H3, H4 Canonical core histones encoded by replication-coupled genes CAF1, Asf1 115, 48

CenH3 Centromere-specific histone 3 variant RbAp48 116

H3.3 Replacement histone 3 variant found in active chromatin HirA, Asf1 76

H2A, H2B Canonical core histones encoded by replication-coupled genes FACT 96

H2A.X H2A form with a C-terminal motif that becomes serine phosphorylated at sites of double-strand breaks

INO80? 117, 118

H2A.Z Diverged form of H2A enriched around gene promoters SWR1 119

H2Av A single Drosophila melanogaster variant that is both H2A.X and H2A.Z Tip60 120

MacroH2A Vertebrate specific H2A variant associated with silent chromatin that has an additional globular domain

Unknown –

H2ABbd Vertebrate-specific H2A variant associated with active chromatin Unknown –

Sperm histones Variants that have evolved to tightly package sperm or pollen Unknown –

Asf1, anti-silencing function protein 1; CAF1, chromatin assembly factor 1; FACT, a histone chaperone complex; HirA, a nucleosome assembly complex; INO80, Inositol-requiring protein 80; RbAp48, retinoblastoma-associated p48; SWR1, Swi2/Snf2-related adenosine triphosphate complex; Tip60, Tat interactive p60.

R E V I E W S



transmitting epigenetic memory. It is not known whether nucleosomes split during replication at PRes, and addressing this possibility experimentally would provide a crucial test of the semi-conservative nucleo-some replication model. However, histone replacement is prominent at PRes, which are conspicuously enriched for H3.3 relative to immediately surrounding regions58

(FIG. 3c). H3.3 enrichment at PRes is seen regardless of whether the homeotic gene that is regulated is active (FIG. 3b), suggesting that histone replacement occurs continuously at sites responsible for epigenetic memory, whereas gene expression depends on the trxG-dependent activation of promoters downstream80. similarly, H3.3 is enriched at the promoter of the inducible chicken folate receptor gene regardless of whether the gene is active or silent60 (FIG. 3d), and over D. melanogaster HsP70 ‘poised’ promoters in the absence of heat shock58. It thus appears that constitutive histone replacement is a general feature of regulatory elements that are able to switch between active and silent configurations.

Many of the unusual features of sites associated with PRes might be ultimately attributable to the turnover process that continually replaces histones and erases histone modifications30. when mobilized to ectopic sites, PRes cause pairing-dependent silencing of reporter genes and frequently ‘home’ to the vicinity of their endogenous sites in the germ line81; perhaps exposure of PRe DNA sequences caused by turnover facilitates these homologous interactions. PRes corre-spond to boundary elements that separate cis-regulatory domains of segment identity 82, and yeast chromatin boundaries are also sites of enhanced histone replace-ment30 (FIG. 3a). Boundary elements and insulators have also been inferred from studies in mammalian cells, and these correspond to CTCF binding sites; D. melanogaster CTCF is found at all PRes, except for the GAF-regulated Fab‑7 PRe63, and its predicted binding sites are enriched in H3.3 (ReF. 83 and s.H., unpublished observations). A broad correspondence between H3.3 and DNA- and chromatin-binding proteins has been seen for transcrip-tion factor ‘hotspots’84, suggesting that histone turnover is a general mechanism for proteins to gain access to chromosomal targets.

Another prediction of the turnover model for main-taining active chromatin is that trxGs and proteins with similar in vivo properties will correspond to mediators of the process. In agreement with this prediction, both GAF and Zeste are found to be enriched at sites of his-tone replacement genome-wide58 (FIG. 3c). Importantly, mutations in GAF and in the FACT histone chaperone complex cause reduced H3.3 levels specifically at sites of GAF binding, including the Fab‑7 PRe63. Nucleosome remodellers, such as Brahma and Kismet, are expected to drive replacement10, although exactly which ones might be involved in replacement at PRes is not known.

The correlation between H3.3 and H3K4me is consist-ent with a role for the Trithorax H3K4 methyltransferase in histone replacement, but is also consistent with H3K4me directly facilitating gene expression. If H3K4me does have a direct role, it would seem counterintuitive that lid (a jumonji C-domain-containing H3K4 demethylase) is also


B

C

Ca

Cb Cd

Cc

E74 E75 E74 E75

E74 E75 E74 E75


Histone H3.3DNA

Histone H3K4me2

A

AutosomesX chromosomes

AutosomesX chromosomes

Figure 2 |HistoneH3.3marksactivechromatin.A| During pachytene of Caenorhabditis elegans gametogenesis, the meiotically silent X chromosome is deficient in both H3.3–GFP (left) and histone H3 dimethylated on lysine 4 (H3K4me2) (right). In both cases, DNA is labelled in red, and either H3.3 or H3K4me2 is labelled in green. Lack of H3.3 or H3K4me2 results in a red X chromosome. B| Staining of Drosophila melanogaster larval salivary gland chromosomes reveals interbands, which are transcriptionally active regions, enriched in H3.3–GFP (green). 4′,6-diamidino-2-phenylindole (DAPI) staining (red) shows DNA-rich bands. Arrows are reference markers. c| During D. melanogaster development, salivary gland chromosome puffs show enrichment of both RNA polymerase II and H3.3 (cb), with depletion of H3 (cd).In the top panels, vertical lines indicate polytene chromosome divisions spanning the E74 and E75 ecdysone-inducible puff sites before induction. Lower panels show incorporation of RNA polymerase II (blue) and H3.3 or H3 (green) after ecdysone induction. DNA is labelled in red. Part A reproduced with permission from ReF. 70 (2006) Public Library of Science. Parts B and c reproduced with permission from ReF. 68 (2005) Cold Spring Harbour Laboratory Press.

R E V I E W S




http://www.ncbi.nlm.nih.gov/sites/entrez?Db=gene&Cmd=ShowDetailView&TermToSearch=5657585&ordinalpos=1&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSum

Boundary element A genetic element that separates independent cis-acting regulatory domains, or separates active from silent chromatin, preventing them from ‘spreading’ into one another.

Insulator A segment of DNA that prevents silencing of a reporter gene by adjacent heterochromatin. Some insulators have been demonstrated to be boundary elements in their native context.

CTCF A highly conserved DNA-binding protein with 11 zinc fingers that binds to insulators and boundaries in mammalian genomes.

a trxG; this class of proteins should silence transcription and so counteract any positive effects of the Trithorax H3K4 methyltransferase on transcription18. However, this apparent contradiction would be explained by continuous histone turnover maintaining active chromatin at PRes. Methylation of H3K4 might serve to promote disassem-bly and its removal to promote assembly, in a manner

similar to that proposed for dynamic acetylation–deacetylation cycles85. evidence for such roles of histone modifications in modulating histone turnover is discussed in a later section. The ability of histone turnover to ration-alize the diverse activities of trxGs makes this process an especially attractive candidate for mediating epigenetic memory at PRes and other regulatory elements.


0.8

0.6

0.4

1.0

0.2

0.0

Condensed chromatin

β-globin domain

3′HS HSA HS4 HS′ H

S3

HS2

H

S1

βH2 enhancer COR3FR gene

5 kb

d

b

Distance from binding sites (kb) 5 0 5

H3.3 (Zeste sites)H3.3 (EZ+PSC sites)

ORFs SNT1 FEN1

iab-3 iab-4 iab-5 iab-6 iab-7 iab-8 10 kb

Abd-B Mcp Fab-7 Fab-8

EZ PSC

Log

(ratio

)

a c

0.7

–0.3

0

0

–1

Nor

mal

ized

fold

diff

eren

ce

Log(

H3.

3/H

3)

0 5 10 15 20 30 35 40 45 50 55 25

FR on

FR off

Nucleosome occupancy (Zeste sites)Nucleosome occupancy (EZ+PSC sites)

Figure 3 | Histonereplacementatregulatoryelements.a| Histone turnover measurements at two yeast genes. Saccharomyces cerevisiae nucleosomes show high levels of turnover around the nucleosome-deficient promoter regions of SNT1 and FEN1. ‘Hot’ nucleosomes, which have a high rate of turnover, are depicted as red balls, ‘cold’ nucleosomes, which have a low rate of turnover, are coloured green and those that turn over at an intermediate rate are coloured black. Nucleosome positions of the SNT1 and FEN1 genes are hot at promoters, becoming colder downstream. b| Histone H3.3 is enriched and nucleosome occupancy is reduced genome-wide at sites of Drosophila melanogaster Polycomb group protein (PcG; PSC+EZ) binding and trithorax group protein (trxG; Zeste) binding58. PSC, posterior-sex-comb protein, a component of Polycomb repressive complex 1 (PRC1); EZ, enhancer-of-zeste protein, a component of PRC2.c| H3.3 is enriched at boundaries of cis-regulatory domains and sites of PcG binding in D. melanogaster cells. Part of the abdominal regulatory region is shown, with boundaries of cis-regulatory domains indicated by dashed arrows and well-mapped PREs shown in magenta. Note that Fab‑7 and Fab‑8 are enriched in H3.3 but are not bound by EZ and PSC PcG proteins79. iab, infraabdominal. d| The chicken folate receptor (FR) promoter region is enriched in H3.3 regardless of whether FR is on or off. Normalized fold difference is a measure of the excess of epitope-tagged H3.3 over H3. The map below shows DNase I hypersensitive (HS) sites in the region that includes the FR gene, with strong peaks of histone replacement over the FR promoter region at 6–8 kb, the β-globin locus at 20–50 kb and the olfactory receptor 3 (COR3) gene at 52–53 kb. Part a modified with permission from ReF. 30 (2007) American Association for the Advancement of Science. Data for part b from ReF. 58. Part c modified with permission from ReF. 58 (2007) American Association for the Advancement of Science. Data for part d from ReF. 60.

R E V I E W S



Histone modifications at sites of replacement. evidence that histones turn over rapidly at gene regulatory sites calls for a reconsideration of observations that have been widely interpreted in terms of static nucleosomes. For example, a large body of evidence shows that H3K4 methylation correlates with active chromatin, leading to the popular idea that di- and trimethylation of H3K4 are themselves epigenetic marks86. However, these modifications are most abundant on H3.3 nucleosomes, which are turning over. Therefore, H3K4 methylation might instead serve as a marker for histone replacement with a regulatory func-tion, not itself an epigenetic mark that would allow for inheritance of active chromatin. For example, patterns of H3K4me2 during Caenorhabditis elegans gametogenesis and early development correspond broadly to patterns of H3.3 incorporation70 (FIG. 2a). Furthermore, the 5′ to 3′ gradient of tri- to di- to monomethylation that is seen for

H3K4 (ReF. 87) might simply reflect the longer exposure of H3.3 or H3 tails to the set1 H3K4 methyltransferase bound to the initiation form of RNA polymerase II88, which itself shows a 5′ to 3′ gradient27.

Histone turnover might also underlie the observa-tion of ‘bivalent domains’ of H3K4me3 and H3K27me3, which are regions in which these two components of active and silent chromatin, respectively, co-localize in pluripotent cells89. During differentiation, H3K4me3 is retained at active sites, and H3K27me3 at silent sites. H3K4me3 would mark H3.3 nucleosomes that are turn-ing over and H3K27me3 would mark H3 nucleosomes that are not. Although bivalency is seen on small pieces of fragmented chromatin, it is not known whether both modifications are found on the same histone H3 tail. Resolution of this issue could provide a test of the turnover hypothesis, which predicts monovalent, but not necessarily bivalent, histone H3 tails. High-resolution mapping of bivalent domains has shown that they corre-spond to short stretches of H3K4me3 surrounded by long regions of H3K27me3 (ReFS 27,90). This is reminiscent of H3.3 peaks over genetically defined PRes surrounded by regions of exceptionally low replacement58. It will be interesting to determine whether H3K4me3 peaks at bivalent domains correspond to sites of H3.3 enrichment. Although no PRes have been identified in mammalian cells so far, it is possible that they correspond to bivalent domains and/or sites of histone replacement.

Modulating nucleosome stabilityObservations of histone turnover at gene regulatory sites have implications for how differences in nucleo-some stability might contribute to the propagation of epigenetic memory.

Stabilizing nucleosomes via histone tails. If nucleosome instability is ultimately responsible for maintaining regulatory sites in an accessible state, then the diversity of post-translational histone modifications might have evolved to modulate nucleosome stability8. One proposed mechanism is that histone tail modifications, especially those on the (H3–H4)2 tetramer, strengthen or weaken nucleosome anchoring via their interactions with linker DNA or other histones8,91,92 (FIG. 5a). The H3 tail exits from between the gyres on either side of the dyad axis where linker DNA and H1 are present, and the H4 tail makes contact with adjacent nucleosomal surfaces93. lysine acetylation is thought to interfere with such contacts by charge neutralization, and this would result in increased nucleosome mobility85. lysine methylation and acetyla-tion and arginine methylation are also thought to stabilize binding of non-histone proteins to H3 tails94. This would either reduce or increase anchoring depending on the res-idue that is methylated and on the chromatin-associated protein that binds8. For example, binding of PRC1 to methylated H3K27 and binding of heterochromatin pro-tein 1 (HP1) to methylated H3K9 would help to anchor the H3 tail and stabilize nucleosomes (FIG. 5b), whereas binding of the Brahma ATP-dependent nucleosome remodeller to acetylated H4 tails would destabilize them. The possibility that the role of histone tail modifications


Factors are bound

Nucleosome remodellingreleases bound factors

Factors (re)bind to accessible DNA

ZBS ZBS

ZBSZBS

ZBSZBS

Zeste Zeste

Zeste

Zeste

Zeste

Zeste

ASF1

PcGs

HIRA

ASF1

ASF1

Figure 4 | Modelforconstitutiveaccessibilityatacis-regulatorysitemediatedbyhistoneturnover.A model for nucleosome dynamics at transcription factor binding sites57 has been adapted to account for the maintenance of accessibility at Polycomb response elements (PREs)58 and other constitutive regulatory sites60, regardless of the activity state of the gene that is regulated. A PRE that controls a homeotic gene (the gene itself is not shown here) includes binding sites for the trithorax group protein (trxG) Zeste (ZBS). Binding motifs for other DNA-binding proteins are also likely to be present, and other PREs would include different clusters of sites. The top part of the figure shows binding of transcription factors to their sites, such as binding of Zeste, which occurs at gaps between H3-containing nucleosomes (represented as green disks). The middle part of the figure shows the nucleosome translocation activity of a trxG remodelling complex (represented as purple ovals), together with the nucleosome disassembly activity of Anti-silencing function protein 1 (ASF1), which results in the dynamic release of Zeste from its binding sites. After release, these factors would become locally available for rapid rebinding to exposed DNA behind the complex (shown in the bottom part of the figure). The exposed DNA also would become transiently available to other diffusible factors, including Polycomb group (PcG) proteins. The continued local presence of a nucleosome remodeller would result in another cycle of remodelling and more nucleosome depletion, nuclease hypersensitivity and factor rebinding at the site. A similar cycle would occur for other motif-specific DNA-binding proteins, such as CTCF83. This process would lead to occasional histone replacement by HIRA–ASF1, resulting in the deposition of H3.3–H4 dimers (represented as orange balls)76, and continue through replication and replication-coupled assembly (not shown), which deposits H3-containing nucleosomes. This dynamic cycle would keep regulatory regions relatively free of nucleosomes over time and after passage of the replication fork, thus maintaining DNA accessibility through cell generations.

R E V I E W S



is to alter tetramer stability provides a simple alternative to the histone code hypothesis, in which combinations of modifications are proposed to specify a diversity of distinct chromatin states94.

Differential stability of subnucleosomal components. The subunit organization of the nucleosome core, with the (H3–H4)2 tetramer flanked by dimers of H2A–H2B, means that DNA unravelling from around the core will first release an H2A–H2B dimer, yielding a hexamer. Further unravelling past the H3–H3 four-helix bundle at the dyad axis can lead to eviction. One way to prevent nucleosome eviction would be to actively retain H2A–H2B

dimers during disruptive processes such as replication and transcription; the FACT chaperone complex seems to be responsible for dimer stability95,96. sometimes dim-ers are lost, in which case rapid replacement would be expected to help prevent eviction. In yeast, H2A–H2B dimer replacement appears to be too rapid to measure in the same type of pulse experiment that can readily detect hot versus cold nucleosomes using tagged H3 (ReF. 31). similarly, dimer replacement in Physarum polycephalum is rapid relative to that of (H3–H4)2 tetramers97. Dimer replacement and turnover is much slower on a global level in mammalian cells, where retention times on the order of hours have been measured for H2A and


HAT

Unstable Stable

H2A.Z–H2Bdimer

H2A–H2B dimer

HATHAT

HP1 HP1

SWR1

SWR1 SWR1

a

b

Nucleosome octamer(containing H3.3 + H2A)

c

H2A

H3

H4

H2B

HAT

Acetyl group

Methyl group

Figure 5 | Modulationofnucleosomestability.a| Histone tail modifications might affect nucleosome stability8. A histone acetyltransferase (HAT) puts acetyl groups onto histone H3 tails. The presence of many acetylated lysines on histone tails reduces interactions with other chromatin components, destabilizing nucleosomes. b| The effect of chromatin-associated proteins on nucleosome stability8. A heterochromatin protein 1 (HP1) dimer that binds to two histone H3 tails that are methylated at lysine 9 (H3K9) might bridge two tails on the same nucleosome or on two nearby nucleosomes and reduce mobility, thus preventing nucleosome eviction. c| Histone variants modulate nucleosome stability. The ATP-dependent dimer-replacement protein SWR1 mediates the replacement of an H2A–H2B dimer with an H2A.Z–H2B dimer. H3.3 but not H3 nucleosomes become destabilized after replacement of H2A with H2A.Z, perhaps owing in part to a steric clash between H2A.Z and H2A121.

R E V I E W S



H2B, and are essentially unmeasurable for H3 and H4 (ReF. 98). This global quiescence probably reflects the fact that only a small proportion of a vertebrate genome is actively transcribed, and only regions that are tran-scribed at a high level are likely to lose nucleosomes at a measurable rate99. Thus, the modulation of nucleosome stability would be a general mechanism for restricting the accessibility of large genomes.

Modulation of nucleosome stability by histone variants. Dimer replacement can also lead to the incorporation of H2A variants, and over the past few years the process whereby H2A.Z replaces canonical H2A has been inten-sively studied in yeast100. The swR1 ATP-dependent chromatin-remodelling complex deposits H2A.Z at sites that immediately flank most yeast promoters. yeast H2A.Z also behaves as an anti-silencer101, and nucleosomes that contain it are unstable relative to H2A nucleosomes in vitro102. In addition, H2A.Z is necessary for yeast to maintain recently repressed genes in an environment at the nuclear periphery that is favourable for their rapid reactivation103. In mammalian cells, H2A.Z localizes to active chromatin27,104. These observations have sug-gested that H2A.Z nucleosomes are prone to eviction, and that anti-silencing results from the erasure of silenc-ing modifications that occurs with the loss of a nucleo-some30. However, other studies suggest that H2A.Z nucleosomes are more stable than H2A nucleosomes105. This view is supported by the presence of mammalian H2A.Z in pericentric heterochromatin106 and by the genetic behaviour of D. melanogaster H2Av as a PcG pro-tein22. Therefore, different studies of H2A.Z have led to contradictory conclusions.

A feasible resolution of this enigma comes from the discovery that H2A.Z destabilizes H3.3-containing nucleosomes107 (FIG. 5c). Destabilization is so strong that crosslinking is necessary to detect the two variants in the same nucleosome core particle, even when purified under low salt conditions. Because yeast H3 is actually an H3.3, the anti-silencing properties of H2A.Z might be understood in terms of instability and eviction of H3.3 nucleosomes, whereas its structural stability, its presence in mammalian heterochromatin and its behaviour as a PcG protein might reflect resistance to eviction when present in an H3 nucleosome. Relative to H3 nucleo-somes, H3.3 nucleosome core particles are inherently unstable to high ionic strength buffers107. This is consist-ent with earlier evidence that overexpression of H3.3 enhances and overexpression of H3 represses reporter-gene activity60. These surprising findings indicate that H3.3 and H3 nucleosomes are profoundly different in stability, and H2A.Z has evolved to amplify this differ-ence. Another H2A variant, H2ABbd, resists assembly into nucleosomes in vitro when coupled with H3.3 relative to H3 (ReF. 108), suggesting that this mammalian-specific H2A variant has also evolved to amplify stability dif-ferences between H3.3 and H3. It remains to be seen whether stability differences between H3.3 and H3 are attributable to encoded amino-acid differences or to a destabilizing histone modification, such as acetlyated K56 (ReFS 32,53). Regardless of the exact mechanisms that lead

to differential stability of nucleosomes containing differ-ent histone variants, it is becoming increasingly clear that modulation of nucleosome stability is a key mechanism for epigenetic regulation.

Perspectiveepigenetic regulation of gene expression is typically described as a chromatin-based phenomenon, but it has been well documented in organisms that lack histones. For example, in Escherichia coli, the epigenetic proc-esses of lysogeny and phase variation do not require packaging proteins at all, but only sequence-specific DNA-binding proteins (lysogeny)75, or a combination of DNA methylation and a transcription factor (phase variation)109. From the evidence described in this Review, chromatin regulation appears to involve similar dynamics, with DNA-binding proteins rebinding after replication to maintain the active state, except that they would do so in large part by mobilizing or excluding nucleosomes, which otherwise promote transcriptional silence.

The cataloguing and mapping of chromatin compo-nents has suggested a fixed template that is scanned and acted upon by histone-modifying enzymes that attach moieties to which chromatin-associated proteins bind. However, chromatin is far more dynamic than expected, with RNA polymerases and nucleosome remodellers providing the driving forces that disrupt chromatin and result in diverse patterns of histone modifications and variants38,74. In addition, increasing evidence suggests that rather than being instructive for gene expression, histone modifications that correlate with gene expression and that are important for gene regulation might mainly reflect the consequences of active processes75. Meanwhile, genetic, biochemical and genomic studies of PcGs, trxGs and systems of gene regulation in various model systems have confirmed that chromatin processes are intimately involved in gene regulation and epigenetic inheritance4,15. However, even though we are gaining a fairly complete list of the players, how they interact with the sequence-specific regulatory proteins that control development remains speculative.

This situation is likely to change soon. Technological advances in genomics have made it possible for afford-able genome-wide studies of chromatin landscapes to be carried out in individual laboratories. especially exciting is the mapping of nucleosomes using deep sequencing technologies, which have begun to uncover differences in nucleosome mobility36. Advances in imaging have resulted in the routine visualization of chromatin processes in three dimensions and in real-time to measure nucleosome dynamics. High-resolution three-dimensional structures are now available for most classes of chromatin proteins involved in nucleosome stabilization, and the availability of so many genome sequences has greatly increased the power of using protein-sequence conservation to guide one’s thinking about molecular mechanisms. Meanwhile, the internet has made biocomputing tools and databases readily accessible. These powerful resources should make it possible to explore nucleosome dynamics in molecular detail to help us to understand how chromatin participates in the epigenetic regulation of gene expression.

R E V I E W S



1. Felsenfeld, G. & Groudine, M. Controlling the double helix. Nature 421, 448–453 (2003).

2. Zhang, Z., Shibahara, K. I. & Stillman, B. PCNA connects DNA replication to epigenetic inheritance in yeast. Nature 408, 221–225 (2000).

3. Hennig, L., Bouveret, R. & Gruissem, W. MSI1-like proteins: an escort service for chromatin assembly and remodeling complexes. Trends Cell Biol. 15, 295–302 (2005).

4. Ringrose, L. & Paro, R. Polycomb/trithorax response elements and epigenetic memory of cell identity. Development 134, 223–232 (2007).

5. Turner, B. M. Cellular memory and the histone code. Cell 111, 285–291 (2002).

6. Li, B., Carey, M. & Workman, J. L. The role of chromatin during transcription. Cell 128, 707–719 (2007).

7. Segal, E. et al. A genomic code for nucleosome positioning. Nature 442, 772–778 (2006).Many yeast promoter sequences have evolved to be unfavourable for wrapping nucleosomes, which would facilitate eviction.

8. Cosgrove, M. S., Boeke, J. D. & Wolberger, C. Regulated nucleosome mobility and the histone code. Nature Struct. Mol. Biol. 11, 1037–1043 (2004).

9. Henikoff, S. & Ahmad, K. Assembly of variant histones into chromatin. Ann. Rev. Cell Dev. Biol. 21, 133–153 (2005).

10. Kingston, R. E. & Tamkun, J. in Epigenetics (eds Allis, C. D., Jenuwein, T., Reinberg, D. & Caparros, M.-L.) 231–248 (Cold Spring Harbor Laboratory Press, Cold Spring Harbor, 2006).

11. Spofford, J. B. in Genetics and Biology of Drosophila Vol. 1c (eds Ashburner, M. & Novitski, E.) 955–1019 (Academic, London, 1976).

12. Cheutin, T. et al. Maintenance of stable heterochromatin domains by dynamic HP1 binding. Science 299, 721–725 (2003).

13. Festenstein, R. et al. Modulation of heterochromatin protein 1 dynamics in primary mammalian cells. Science 299, 719–721 (2003).

14. Ahmad, K. & Henikoff, S. Modulation of a transcription factor counteracts heterochromatic gene silencing in Drosophila. Cell 104, 839–847 (2001).

15. Schwartz, Y. B. & Pirrotta, V. Polycomb silencing mechanisms and the management of genomic programmes. Nature Rev. Genet. 8, 9–22 (2007).

16. Klymenko, T. & Muller, J. The histone methyltransferases Trithorax and Ash1 prevent transcriptional silencing by Polycomb group proteins. EMBO Rep. 5, 373–377 (2004).

17. Brock, H. W. & Fisher, C. L. Maintenance of gene expression patterns. Dev. Dyn. 232, 633–655 (2005).

18. Secombe, J., Li, L., Carlos, L. & Eisenman, R. N. The trithorax group protein Lid is a trimethyl histone H3K4 demethylase required for dMyc-induced cell growth. Genes Dev. 21, 537–551 (2007).

19. Secombe, J. & Eisenman, R. N. The function and regulation of the JARID1 family of histone H3 lysine 4 demethylases: the MYC connection. Cell Cycle 6, 1324–1328 (2007).

20. Sparmann, A. & van Lohuizen, M. Polycomb silencers control cell fate, development and cancer. Nature Rev. Cancer 6, 846–856 (2006).

21. Moshkin, Y. M. et al. Histone chaperone Asf1 cooperates with the Brahma chromatin-remodelling machinery. Genes Dev. 16, 2621–2626 (2002).

22. Swaminathan, J., Baxter, E. M. & Corces, V. G. The role of histone H2Av variant replacement and histone H4 acetylation in the establishment of Drosophila heterochromatin. Genes Dev. 19, 65–76 (2005).

23. Reinke, H. & Horz, W. Histones are first hyperacetylated and then lose contact with the activated PHO5 promoter. Mol. Cell 1, 1599–1607 (2003).The authors use chromatin immunoprecipitation to show that nucleosomes are evicted at the PHO5 promoter.

24. Boeger, H., Griesenbeck, J., Strattan, J. S. & Kornberg, R. D. Nucleosomes unfold completely at a transcriptionally active promoter. Mol. Cell 11, 1587–1598 (2003).

25. Mito, Y., Henikoff, J. & Henikoff, S. Genome-scale profiling of histone H3.3 replacement patterns. Nature Genet. 37, 1090–1097 (2005).

26. Yuan, G. C. et al. Genome-scale identification of nucleosome positions in S. cerevisiae. Science 309, 626–630 (2005).

27. Barski, A. et al. High-resolution profiling of histone methylations in the human genome. Cell 129, 823–837 (2007).

28. Weintraub, H. & Groudine, M. Chromosomal subunits in active genes have an altered conformation. Science 193, 848–856 (1976).

29. Sabo, P. J. et al. Genome-scale mapping of DNase I sensitivity in vivo using tiling DNA microarrays. Nature Methods 3, 511–518 (2006).

30. Dion, M., Kaplan, T., Friedman, N. & Rando, O. J. Dynamics of replication-independent histone turnover in budding yeast. Science 315, 1405–1408 (2007).Direct measurements of histone turnover rates reveal that promoters and boundaries are hot, which implies that measurements of histone modifications at regulatory sites miss intermediates.

31. Jamai, A., Imoberdorf, R. M. & Strubin, M. Continuous histone H2B and transcription-dependent histone H3 exchange in yeast cells outside of replication. Mol. Cell 25, 345–355 (2007).The authors show that most newly synthesized yeast H3 turns over at promoters over the course of the cell cycle, whereas H2B turnover in genic and intergenic regions appears to be too rapid to allow differences to be measured.

32. Rufiange, A., Jacques, P. E., Bhat, W., Robert, F. & Nourani, A. Genome-wide replication-independent histone H3 exchange occurs predominantly at promoters and implicates H3 K56 acetylation and Asf1. Mol. Cell 27, 393–405 (2007).Measurements of histone turnover reveal that replication-independent replacement correlates closely with Asf1-dependent H3K56 acetylation genome-wide in yeast, suggesting that nucleosomes are destabilized by a key H3 core modification.

33. Stockdale, C., Flaus, A., Ferreira, H. & Owen-Hughes, T. Analysis of nucleosome repositioning by yeast ISWI and Chd1 chromatin remodeling complexes. J. Biol. Chem. 281, 16279–16288 (2006).

34. Drew, H. R. & Travers, A. A. DNA bending and its relation to nucleosome positioning. J. Mol. Biol. 186, 773–790 (1985).Rules for nucleosome positioning based on AT–GC composition are deduced 20 years before they are confirmed in genome-wide studies.

35. Lowary, P. T. & Widom, J. New DNA sequence rules for high affinity binding to histone octamer and sequence-directed nucleosome positioning. J. Mol. Biol. 276, 19–42 (1998).

36. Albert, I. et al. Translational and rotational settings of H2A.Z nucleosomes across the Saccharomyces cerevisiae genome. Nature 446, 572–576 (2007).

37. Doshi, P., Kaushal, S., Benyajati, C. & Wu, C. I. Molecular analysis of the responder satellite DNA in Drosophila melanogaster: DNA bending, nucleosome structure, and Rsp-binding proteins. Mol. Biol. Evol. 8, 721–741 (1991).

38. Rando, O. J. & Ahmad, K. Rules and regulation in the primary structure of chromatin. Curr. Opin. Cell Biol. 19, 250–256 (2007).

39. Loyola, A. & Almouzni, G. Histone chaperones, a supporting role in the limelight. Biochim. Biophys. Acta 1677, 3–11 (2004).

40. Whitehouse, I. & Tsukiyama, T. Antagonistic forces that position nucleosomes in vivo. Nature Struct. Mol. Biol. 13, 633–640 (2006).The ISW2 nucleosome-remodelling complex can slide nucleosomes from their default positions to energetically unfavourable positions in vivo.

41. Zofall, M., Persinger, J., Kassabov, S. R. & Bartholomew, B. Chromatin remodeling by ISW2 and SWI/SNF requires DNA translocation inside the nucleosome. Nature Struct. Mol. Biol. 13, 339–346 (2006).

42. Woodcock, C. L., Skoultchi, A. I. & Fan, Y. Role of linker histone in chromatin structure and function: H1 stoichiometry and nucleosome repeat length. Chromosome Res. 14, 17–25 (2006).

43. Ragab, A. & Travers, A. HMG-D and histone H1 alter the local accessibility of nucleosomal DNA. Nucleic Acids Res. 31, 7083–7089 (2003).

44. Kim, M. Y., Mauro, S., Gevry, N., Lis, J. T. & Kraus, W. L. NAD+-dependent modulation of chromatin structure and transcription by nucleosome binding properties of PARP-1. Cell 119, 803–814 (2004).

45. Natsume, R. et al. Structure and function of the histone chaperone CIA/ASF1 complexed with histones H3 and H4. Nature 446, 338–341 (2007).

46. English, C. M., Adkins, M. W., Carson, J. J., Churchill, M. E. & Tyler, J. K. Structural basis for the histone chaperone activity of Asf1. Cell 127, 495–508 (2006).

47. Antczak, A. J., Tsubota, T., Kaufman, P. D. & Berger, J. M. Structure of the yeast histone H3–ASF1 interaction: implications for chaperone mechanism, species-specific interactions, and epigenetics. BMC Struct. Biol. 6, 26 (2006).

48. Adkins, M. W., Howar, S. R. & Tyler, J. K. Chromatin disassembly mediated by the histone chaperone Asf1 is essential for transcriptional activation of the yeast PHO5 and PHO8 genes. Mol. Cell 14, 657–666 (2004).In vivo evidence that the H3–H4 assembly protein, Asf1, mediates nucleosome disassembly.

49. Boeger, H., Griesenbeck, J., Strattan, J. S. & Kornberg, R. D. Removal of promoter nucleosomes by disassembly rather than sliding in vivo. Mol. Cell 14, 667–673 (2004).The authors use chromatin circles to show that nucleosomes are evicted from the yeast PHO5 promoter.

50. Recht, J. et al. Histone chaperone Asf1 is required for histone H3 lysine 56 acetylation, a modification associated with S phase in mitosis and meiosis. Proc. Natl Acad. Sci. USA 103, 6988–6993 (2006).

51. Han, J. et al. Rtt109 acetylates histone H3 lysine 56 and functions in DNA replication. Science 315, 653–655 (2007).

52. Schneider, J., Bajwa, P., Johnson, F. C., Bhaumik, S. R. & Shilatifard, A. Rtt109 is required for proper H3K56 acetylation: a chromatin mark associated with the elongating RNA polymerase II. J. Biol. Chem. 281, 37270–37274 (2006).

53. Xu, F., Zhang, K. & Grunstein, M. Acetylation in histone H3 globular domain regulates gene expression in yeast. Cell 121, 375–385 (2005).

54. Smith, C. L. & Peterson, C. L. ATP-dependent chromatin remodeling. Curr. Top. Dev. Biol. 65, 115–148 (2005).

55. Lorch, Y., Maier-Davis, B. & Kornberg, R. D. Chromatin remodeling by nucleosome disassembly in vitro. Proc. Natl Acad. Sci. USA 103, 3090–3093 (2006).

56. Armstrong, J. A. et al. The Drosophila BRM complex facilitates global transcription by RNA polymerase II. EMBO J. 21, 5245–5254 (2002).

57. Nagaich, A. K., Walker, D. A., Wolford, R. & Hager, G. L. Rapid periodic binding and displacement of the glucocorticoid receptor during chromatin remodeling. Mol. Cell 14, 163–174 (2004).Interplay between a transcription factor and an ATP-dependent nucleosome remodeller in vitro leads to a dynamic model to account for the instability of DNA-binding proteins in vivo.

58. Mito, Y., Henikoff, J. & Henikoff, S. Histone replacement marks the boundaries of cis-regulatory domains. Science 315, 1408–1411 (2007).PREs are both enriched in H3.3 and depleted of nucleosomes relative to surrounding regions regardless of the on–off state of the gene, indicating that histone turnover is a constitutive feature of sites that propagate cellular memory.

59. Ahmad, K. & Henikoff, S. The histone variant H3.3 marks active chromatin by replication-independent nucleosome assembly. Mol. Cell 9, 1191–1200 (2002).Replication-independent assembly of H3.3 is shown to be a distinct pathway from replication-coupled assembly of H3, and its deposition marks active chromatin.

60. Jin, C. & Felsenfeld, G. Distribution of histone H3.3 in hematopoietic cell lineages. Proc. Natl Acad. Sci. USA 103, 574–579 (2006).The authors show that H3.3 is enriched at a vertebrate promoter whether or not it is active, and that expression is increased by high levels of H3.3 and decreased by high levels of H3.

61. Zhang, R., Chen, W. & Adams, P. D. Molecular dissection of formation of senescence-associated heterochromatin foci. Mol. Cell. Biol. 27, 2343–2358 (2007).

62. Chow, C. M. et al. Variant histone H3.3 marks promoters of transcriptionally active genes during mammalian cell division. EMBO Rep. 6, 354–360 (2005).

63. Nakayama, T., Nishioka, K., Dong, Y. X., Shimojima, T. & Hirose, S. Drosophila GAGA factor directs histone H3.3 replacement that prevents the heterochromatin spreading. Genes Dev. 21, 552–561 (2007).Histone replacement at regulatory sites, including the Fab‑7 PRE, depends on GAF, a DNA-binding protein in the trithorax group.

R E V I E W S



64. McKittrick, E., Gafken, P. R., Ahmad, K. & Henikoff, S. Histone H3.3 is enriched in covalent modifications associated with active chromatin. Proc. Natl Acad. Sci. USA 101, 1525–1530 (2004).

65. Hake, S. B. et al. Expression patterns and post-translational modifications associated with mammalian histone H3 variants. J. Biol. Chem. 281, 559–568 (2006).

66. Johnson, L. et al. Mass spectrometry analysis of Arabidopsis histone H3 reveals distinct combinations of post-translational modifications. Nucleic Acids Res. 32, 6511–6518 (2004).

67. Waterborg, J. H. Sequence analysis of acetylation and methylation in two histone H3 variants of alfalfa. J. Biol. Chem. 265, 17157–17161 (1990).The enrichment of active lysine modifications on the replication-independent histone 3 variant and of silent modifications on its replication-coupled counterpart is demonstrated years before it was generally recognized that these methylated and acetylated lysines have roles in epigenetic regulation.

68. Schwartz, B. E. & Ahmad, K. Transcriptional activation triggers deposition and removal of the histone variant H3.3. Genes Dev. 19, 804–814 (2005).Induction of transcription leads to rapid loss of histone H3 and replacement with H3.3, which turns over during transcriptional elongation, and becomes stable when transcription shuts down.

69. Wirbelauer, C., Bell, O. & Schubeler, D. Variant histone H3.3 is deposited at sites of nucleosomal displacement throughout transcribed genes while active histone modifications show a promoter-proximal bias. Genes Dev. 19, 1761–1766 (2005).

70. Ooi, S., Priess, J. & Henikoff, S. Histone H3.3 variant dynamics in the germline of Caenorhabditis elegans. PLoS Genet. 2, e97 (2006).

71. Schermer, U. J., Korber, P. & Horz, W. Histones are incorporated in trans during reassembly of the yeast PHO5 promoter. Mol. Cell 19, 279–285 (2005).

72. Weintraub, H., Worcel, A. & Alberts, B. A model for chromatin based upon two symmetrically paired half-nucleosomes. Cell 9, 409–417 (1976).

73. Annunziato, A. T. Split decision: what happens to nucleosomes during DNA replication? J. Biol. Chem. 280, 12065–12068 (2005).

74. Henikoff, S., Furuyama, T. & Ahmad, A. Histone variants, nucleosome assembly and epigenetic inheritance. Trends Genet. 20, 320–326 (2004).

75. Ptashne, M. On the use of the word ‘epigenetic’. Curr. Biol. 17, R233–R236 (2007).

76. Tagami, H., Ray-Gallet, D., Almouzni, G. & Nakatani, Y. Histone H3.1 and H3.3 complexes mediate nucleosome assembly pathways dependent or independent of DNA synthesis. Cell 116, 51–61 (2004).Distinct histone chaperone complexes are responsible for replication-coupled and replication-independent assembly of histone H3 variants.

77. McNally, J. G., Muller, W. G., Walker, D., Wolford, R. & Hager, G. L. The glucocorticoid receptor: rapid exchange with regulatory sites in living cells. Science 287, 1262–1265 (2000).

78. Bosisio, D. et al. A hyper-dynamic equilibrium between promoter-bound and nucleoplasmic dimers controls NF-κB-dependent gene activity. EMBO J. 25, 798–810 (2006).

79. Schwartz, Y. B. et al. Genome-wide analysis of Polycomb targets in Drosophila melanogaster. Nature Genet. 38, 700–705 (2006).

80. Papp, B. & Muller, J. Histone trimethylation and the maintenance of transcriptional ON and OFF states by trxG and PcG proteins. Genes Dev. 20, 2041–2054 (2006).

81. Kassis, J. A. Pairing-sensitive silencing, Polycomb group response elements, and transposon homing in Drosophila. Adv. Genet. 46, 421–438 (2002).

82. Maeda, R. K. & Karch, F. The ABC of the BX-C: the bithorax complex explained. Development 133, 1413–1422 (2006).

83. Holohan, E. E. et al. CTCF genomic binding sites in Drosophila and the organisation of the bithorax complex. PLoS Genet. 3, e112 (2007).

84. Moorman, C. et al. Hotspots of transcription factor colocalization in the genome of Drosophila melanogaster. Proc. Natl Acad. Sci. USA 103, 12027–12032 (2006).

85. Waterborg, J. H. Dynamics of histone acetylation in vivo. A function for acetylation turnover? Biochem. Cell Biol. 80, 363–378 (2002).

86. Ruthenburg, A. J., Allis, C. D. & Wysocka, J. Methylation of lysine 4 on histone H3: intricacy of writing and reading a single epigenetic mark. Mol. Cell 25, 15–30 (2007).

87. Rando, O. J. Global patterns of histone modifications. Curr. Opin. Genet. Dev. 17, 94–99 (2007).

88. Krogan, N. J. et al. The Paf1 complex is required for histone H3 methylation by COMPASS and Dot1p: linking transcriptional elongation to histone methylation. Mol. Cell 11, 721–729 (2003).

89. Bernstein, B. E. et al. A bivalent chromatin structure marks key developmental genes in embryonic stem cells. Cell 125, 315–326 (2006).

90. Mikkelsen, T. S. et al. Genome-wide maps of chromatin state in pluripotent and lineage-committed cells. Nature 448, 553–560 (2007).

91. Hamiche, A., Kang, J. G., Dennis, C., Xiao, H. & Wu, C. Histone tails modulate nucleosome mobility and regulate ATP-dependent nucleosome sliding by NURF. Proc. Natl Acad. Sci. USA 98, 14316–14321 (2001).

92. Ferreira, H., Somers, J., Webster, R., Flaus, A. & Owen-Hughes, T. Histone tails and the H3 αN helix regulate nucleosome mobility and stability. Mol. Cell. Biol. 27, 4037–4048 (2007).

93. Luger, K., Mader, A. W., Richmond, R. K., Sargent, D. F. & Richmond, T. J. Crystal structure of the nucleosome core particle at 2.8 Å resolution. Nature 389, 251–260 (1997).

94. Strahl, B. D. & Allis, C. D. The language of covalent histone modifications. Nature 403, 41–45 (2000).

95. Formosa, T. et al. Defects in SPT16 or POB3 (yFACT) in Saccharomyces cerevisiae cause dependence on the Hir/Hpc pathway: polymerase passage may degrade chromatin structure. Genetics 162, 1557–1571 (2002).

96. Belotserkovskaya, R. et al. FACT facilitates transcription-dependent nucleosome alteration. Science 301, 1090–1093 (2003).

97. Thiriet, C. & Hayes, J. J. Replication-independent core histone dynamics at transcriptionally active loci in vivo. Genes Dev. 19, 677–682 (2005).

98. Kimura, H. & Cook, P. R. Kinetics of core histones in living human cells: little exchange of H3 and H4 and some rapid exchange of H2B. J. Cell Biol. 153, 1341–1353 (2001).

99. Kulaeva, O. I., Gaykalova, D. A. & Studitsky, V. M. Transcription through chromatin by RNA polymerase II: histone displacement and exchange. Mutat. Res. 618, 116–129 (2007).

100. Raisner, R. M. & Madhani, H. D. Patterning chromatin: form and function for H2A.Z variant nucleosomes. Curr. Opin. Genet. Dev. 16, 119–124 (2006).

101. Meneghini, M. D., Wu, M. & Madhani, H. D. Conserved histone variant H2A.Z protects euchromatin from the ectopic spread of silent chromatin. Cell 112, 725–736 (2003).

102. Zhang, H., Roberts, D. N. & Cairns, B. R. Genome-wide dynamics of Htz1, a histone H2A variant that poises repressed/basal promoters for activation through histone loss. Cell 123, 219–231 (2005).

103. Brickner, D. G. et al. H2A.Z-mediated localization of genes at the nuclear periphery confers epigenetic memory of previous transcriptional state. PLoS Biol. 5, e81 (2007).

104. Farris, S. D. et al. Transcription-induced chromatin remodeling at the c-myc gene involves the local exchange of histone H2A.Z. J. Biol. Chem. 280, 25298–25303 (2005).

105. Park, Y. J., Dyer, P. N., Tremethick, D. J. & Luger, K. A new fluorescence resonance energy transfer approach demonstrates that the histone variant H2AZ stabilizes the histone octamer within the nucleosome. J. Biol. Chem. 279, 24274–24282 (2004).

106. Fan, J. Y., Rangasamy, D., Luger, K. & Tremethick, D. J. H2A.Z alters the nucleosome surface to promote hp1α-mediated chromatin fiber folding. Mol. Cell 16, 655–661 (2004).

107. Jin, C. & Felsenfeld, G. Nucleosome stability mediated by histone variants H3.3 and H2A.Z. Genes Dev. 21, 1519–1529 (2007).Nucleosome stability differs depending on the variant, with H3.3 nucleosome core particles becoming unstable under ionic conditions in which H3 nucleosomes are stable, becoming especially unstable when H2A.Z is also present.

108. Okuwaki, M., Kato, K., Shimahara, H., Tate, S. & Nagata, K. Assembly and disassembly of nucleosome core particles containing histone variants by human nucleosome assembly protein I. Mol. Cell. Biol. 25, 10639–10651 (2005).

109. Casadesus, J. & Low, D. Epigenetic gene regulation in the bacterial world. Microbiol. Mol. Biol. Rev. 70, 830–856 (2006).

110. Ficz, G., Heintzmann, R. & Arndt-Jovin, D. J. Polycomb group protein complexes exchange rapidly in living Drosophila. Development 132, 3963–3976 (2005).

111. Pien, S. & Grossniklaus, U. Polycomb group and trithorax group proteins in Arabidopsis. Biochim. Biophys. Acta 1769, 375–382 (2007).

112. Karachentsev, D., Sarma, K., Reinberg, D. & Steward, R. PR-Set7-dependent methylation of histone H4 Lys 20 functions in repression of gene expression and is essential for mitosis. Genes Dev. 19, 431–435 (2005).

113. Richards, E. J. & Elgin, S. C. Epigenetic codes for heterochromatin formation and silencing: rounding up the usual suspects. Cell 108, 489–500 (2002).

114. Cleard, F., Moshkin, Y., Karch, F. & Maeda, R. K. Probing long-distance regulatory interactions in the Drosophila melanogaster bithorax complex using Dam identification. Nature Genet. 38, 931–935 (2006).

115. Smith, S. & Stillman, B. Purification and characterization of CAF-I, a human cell factor required for chromatin assembly during DNA replication in vitro. Cell 58, 15–25 (1989).

116. Furuyama, T., Dalal, Y. & Henikoff, S. Chaperone-mediated assembly of centromeric chromatin in vitro. Proc. Natl Acad. Sci. USA 103, 6172–6177 (2006).

117. van Attikum, H., Fritsch, O., Hohn, B. & Gasser, S. M. Recruitment of the INO80 complex by H2A phosphorylation links ATP-dependent chromatin remodeling with DNA double-strand break repair. Cell 119, 777–788 (2004).

118. Morrison, A. J. et al. INO80 and γ-H2AX interaction links ATP-dependent chromatin remodeling to DNA damage repair. Cell 119, 767–775 (2004).

119. Mizuguchi, G. et al. ATP-driven exchange of histone H2AZ variant catalyzed by SWR1 chromatin remodeling complex. Science 303, 343–348 (2004).The SWR1 complex replaces H2A with the H2A.Z variant, demonstrating a direct connection between a histone variant and an ATP-dependent nucleosome remodeller.

120. Kusch, T. et al. Acetylation by Tip60 is required for selective histone variant exchange at DNA lesions. Science 306, 2084–2087 (2004).

121. Suto, R. K., Clarkson, M. J., Tremethick, D. J. & Luger, K. Crystal structure of a nucleosome core particle containing the variant histone H2A. Z. Nature Struct. Biol. 7, 1121–1124 (2000).

122. Talbert, P. B. & Henikoff, S. Spreading of silent chromatin: inaction at a distance. Nature Rev. Genet. 7, 793–803 (2006).

AcknowledgementsI thank K. Ahmad, Y. Mito, T. Furuyama and other past and present members of my laboratory for the many stimulating discussions and ideas that have contributed to this synthesis.

DATABASESEntrez Gene: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=geneFab‑7UniProtKB: http://ca.expasy.org/sprotAsf1 | Brahma | CenH3 | GAF | H2A | H2Av | H2B | H3 | H4 | HSP70 | Kismet | Lid | MSI1 | NAP1 | PARP1 | Rtt109 | Trithorax | Zeste

AlllinksAReActiveintHeonlinepDf

R E V I E W S



http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=gene

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=gene

http://www.ncbi.nlm.nih.gov/sites/entrez?Db=gene&Cmd=ShowDetailView&TermToSearch=5657585&ordinalpos=1&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSum

http://ca.expasy.org/sprot

http://ca.expasy.org/uniprot/Q9V464


http://www.expasy.org/uniprot/Q8RVQ9


http://ca.expasy.org/uniprot/Q96QV6


http://ca.expasy.org/uniprot/Q96A08

http://ca.expasy.org/uniprot/Q6NXT2



http://ca.expasy.org/uniprot/Q9Y1L3

http://ca.expasy.org/uniprot/Q9VMJ7







in the epigenetic regulation of gene...

Documents