Chapter 8: Chromatin

Edited by Craig Peterson

8.1 Introduction

Chromatin has a compact organization in which most DNA sequences are structurally inaccessible and functionally inactive. Within this mass is the minority of active sequences. What is the general structure of chromatin, and what is the difference between active and inactive sequences? The fundamental subunit of chromatin has the same type of design in all eukaryotes. The nucleosome contains about 200 base pairs (bp) of DNA, organized by an octamer of small, basic proteins into a beadlike structure. The protein components are histones. They form an interior core; the DNA lies on the surface of the particle. Additional regions of the histones, known as the histone tails, extend from the surface. Nucleosomes are an invariant component of euchromatin and heterochromatin in the interphase nucleus and of mitotic chromosomes. The nucleosome provides the first level of organization, compacting the DNA about 6-fold over the length of naked DNA, resulting in a “beads-on-a-string” fiber of approximately 10 nm in diameter. Its components and structure are well characterized.

The secondary level of organization involves interactions between nucleosomes of the 10-nm fiber, leading to more condensed chromatin fibers. Biochemical studies have shown that nucleosomes can assemble into helical arrays that form a fiber of approximately 30 nm in diameter. The structure of this fiber requires the histone tails and is stabilized by linker histones. Whether the 30-nm fiber is a dominant feature of chromatin within cells remains a topic of debate.

The final, tertiary level of chromatin organization requires the further folding and compacting of chromatin fibers into the 3D structures of interphase chromatin or mitotic chromosomes. This results in about 1,000-fold linear compaction in euchromatin, cyclically interchangeable with packing into mitotic chromosomes to achieve an overall compaction of up to 10,000-fold. Heterochromatin generally maintains this approximately 10,000-fold compaction in both interphase and mitosis.

In this chapter, we describe the structure of and relationships between these levels of organization to characterize the events involved in cyclical packaging, replication, and transcription. Association with additional proteins, as well as modifications of existing chromosomal proteins, is involved in changing the structure of chromatin. Replication and transcription, and most DNA repair processes, require unwinding of DNA, and thus first involve an unfolding of the structure that allows the relevant enzymes to manipulate the DNA. This is likely to involve changes in all levels of organization.

When chromatin is replicated, the nucleosomes must be reproduced on both daughter duplex molecules. In addition to asking how the nucleosome itself is assembled, we must inquire what happens to other proteins present in chromatin. Replication disrupts the structure of chromatin, which indicates that it poses a problem for maintaining regions with specific structure but also offers an opportunity to change the structure.

The mass of chromatin contains up to twice as much protein as DNA. Approximately half of the protein mass is accounted for by the nucleosomes. The mass of RNA is less than 10% of the mass of DNA. Much of the RNA consists of nascent transcripts still associated with the template DNA.

The nonhistones include all the proteins found in chromatin except the histones. They are more variable between tissues and species, and they comprise a smaller proportion of the mass than the histones. They also comprise a much larger number of proteins, so that any individual protein is present in amounts much smaller than any histone. The functions of nonhistone proteins include control of gene expression and higher-order structure. Thus, RNA polymerase can be considered to be a prominent nonhistone. The high-mobility group (HMG) proteins comprise a discrete and well-defined subclass of nonhistones (at least some of which are transcription factors).

8.2 DNA Is Organized in Arrays of Nucleosomes

When interphase nuclei are suspended in a solution of low ionic strength, they swell and rupture to release fibers of chromatin. FIGURE 8.1 shows a lysed nucleus in which fibers are streaming out. In some regions, the fibers consist of tightly packed material, but in regions that have become stretched, they consist of discrete particles. These are the nucleosomes. In especially extended regions, individual nucleosomes are visibly connected by a fine thread, which is a free duplex of DNA. A continuous duplex thread of DNA runs through the series of particles.

FIGURE 8.1 Chromatin spilling out of lysed nuclei consists of a compactly organized series of particles. The bar is 100 nm.

Reprinted from: Oudet, P., et al. 1975. “Electron microscopic and biochemical evidence.” Cell, 4:281–300, with permission from Elsevier (http://www.sciencedirect.com/science/journal/00928674). Photo courtesy of Pierre Chambon, College of France.

Researchers can obtain individual nucleosomes by treating chromatin with the endonuclease micrococcal nuclease (MNase), which cuts the DNA between nucleosomes, a region known as linker DNA. Ongoing digestion with MNase releases groups of particles, and eventually single nucleosomes. FIGURE 8.2 shows individual nucleosomes as compact particles measuring about 10 nm in diameter.

FIGURE 8.2 Individual nucleosomes are released by digestion of chromatin with micrococcal nuclease. The bar is 100 nm.

Reprinted from: Oudet, P., et al. 1975. “Electron microscopic and biochemical evidence.” Cell, 4:281–300, with permission from Elsevier (http://www.sciencedirect.com/science/journal/00928674). Photo courtesy of Pierre Chambon, College of France.

When chromatin is digested with MNase, the DNA is cleaved into integral multiples of a unit length. Fractionation by gel electrophoresis reveals the “ladder” presented in FIGURE 8.3. Such ladders extend for multiple steps (about 10 are distinguishable in this figure), and the unit length, determined by the increments between successive steps, averages about 200 bp.

FIGURE 8.3 Micrococcal nuclease digests chromatin in nuclei into a multimeric series of DNA bands that can be separated by gel electrophoresis.

Photo courtesy of Markus Noll, Universität Zürich.

FIGURE 8.4 shows that the ladder is generated by groups of nucleosomes. When nucleosomes are fractionated on a sucrose gradient, they give a series of discrete peaks that correspond to monomers, dimers, trimers, and so on. When the DNA is extracted from the individual fractions and electrophoresed, each fraction yields a band of DNA whose size corresponds with a step on the micrococcal nuclease ladder. The monomeric nucleosome contains DNA of the unit length, the nucleosome dimer contains DNA of twice the unit length, and so on. More than 95% of nuclear DNA can be recovered in the form of the 200-bp ladder, indicating that almost all DNA must be organized in nucleosomes.

FIGURE 8.4 Each multimer of nucleosomes contains the appropriate number of unit lengths of DNA. In the photo, artificial bands simulate a DNA ladder that would be produced by MNase digestion. The image was constructed using PCR fragments with sizes corresponding to actual band sizes.

Photo courtesy of Jan Kieleczawa, Wyzer Biosciences.

The length of DNA present in the nucleosome can vary from the “typical” value of 200 bp. The chromatin of any particular cell type has a characteristic average value (±5 bp). The average most often is between 180 and 200, but there are extremes as low as 154 bp (in a fungus) or as high as 260 bp (in sea urchin sperm). The average value might be different in individual tissues of the adult organism, and there can be differences between different parts of the genome in a single cell type. Variations from the genome average often include tandemly repeated sequences, such as clusters of 5S RNA genes.

A common structure underlies the varying amount of DNA that is contained in nucleosomes of different sources. The association of DNA with the histone octamer forms a core particle containing 145–147 bp of DNA, irrespective of the total length of DNA in the nucleosome. The variation in total length of DNA per nucleosome is superimposed on this basic core structure.

The core particle is defined by the effects of MNase on the nucleosome monomer. The initial reaction of the enzyme is to cut the easily accessible DNA between nucleosomes, but if it is allowed to continue after monomers have been generated, it proceeds to digest some of the DNA of the individual nucleosome, as shown in FIGURE 8.5. Initial cleavage results in nucleosome monomers with (in this example) about 200 bp of DNA. After the first step, some monomers are found in which the length of DNA has been “trimmed” to about 165 bp. Finally, this is reduced to the length of the DNA of the core particle, 145–147 bp. After this, the core particle is resistant to further digestion by MNase.

FIGURE 8.5 Micrococcal nuclease initially cleaves between nucleosomes. Mononucleosomes typically have ~200 bp DNA. End-trimming reduces the length of DNA first to ~165 bp, and then generates core particles with 145–147 bp.

As a result of this type of analysis, nucleosomal DNA is functionally divided into two regions:

Core DNA has a length of 145–147 bp, the length of DNA needed to form a stable monomeric nucleosome, and is relatively resistant to digestion by nucleases.
Linker DNA comprises the rest of the repeating unit. Its length varies from as little as 7 bp to as many as 115 bp per nucleosome.

Core particles have properties similar to those of the nucleosomes themselves, although they are smaller. Their shape and size are similar to those of nucleosomes; this suggests that the essential geometry of the particle is established by the interactions between DNA and the protein octamer in the core particle. Core particles are readily obtained as a homogeneous population, and as a result they are often used for structural studies in preference to nucleosome preparations.

8.3 The Nucleosome Is the Subunit of All Chromatin

The 10-nm particles shown in Figure 8.2 represent the fundamental building block of all chromatin, the nucleosome. The nucleosome contains about 200 bp of DNA associated with a histone octamer that consists of two copies each of histones H2A, H2B, H3, and H4. These are known as the core histones. FIGURE 8.6 illustrates their association and dimensions diagrammatically.

FIGURE 8.6 The nucleosome consists of approximately equal masses of DNA and histones (including H1). The predicted mass of a nucleosome that contains H1 is 262 kD.

The histones are small, basic proteins (rich in arginine and lysine residues), resulting in a high affinity for DNA. Histones H3 and H4 are among the most conserved proteins known, and the core histones are responsible for DNA packaging in all eukaryotes. H2A and H2B are also conserved among eukaryotes, but show appreciable species-specific variation in sequence, particularly in the histone tails. The core regions of the histones are even conserved in archaea and appear to play a similar role in compaction of archaeal DNA.

The shape of the nucleosome corresponds to a flat disk or cylinder of diameter 11 nm and height 6 nm. The length of the DNA is roughly twice the 34-nm circumference of the particle. The DNA follows a symmetrical path around the octamer. FIGURE 8.7 shows the DNA path diagrammatically as a helical coil that makes about one and two-thirds turns around the cylindrical octamer. Note that the DNA “enters” and “exits” on one side of the nucleosome.

FIGURE 8.7 The nucleosome is a cylinder with DNA organized into ~one and two-thirds turns around the surface.

Viewing a cross section through the nucleosome in FIGURE 8.8, we see that the two circumferences made by the DNA lie close to each other. The height of the cylinder is 6 nm, of which 4 nm are occupied by the two turns of DNA (each of diameter 2 nm). The pattern of the two turns has a possible functional consequence. One turn around the nucleosome takes about 80 bp of DNA, so 2 points separated by 80 bp in the free double helix can actually be close on the nucleosome surface, as illustrated in FIGURE 8.9.

FIGURE 8.8 DNA occupies most of the outer surface of the nucleosome.

FIGURE 8.9 Sequences on the DNA that lie on different turns around the nucleosome may be close together.

The core histones tend to form two types of subcomplexes. H3 and H4 form a very stable tetramer in solution (H3₂-H4₂). H2A and H2B most typically form a dimer (H2A-H2B). A space-filling model of the structure of the histone octamer (from the crystal structure at 3.1 Å resolution) is shown in FIGURE 8.10. Tracing the paths of the individual polypeptide backbones in the crystal structure shows that the histones are not organized as individual globular proteins, but that each is interdigitated with its partner: H3 with H4, and H2A with H2B. Figure 8.10 emphasizes the H3₂-H4₂ tetramer (white) and the H2A-H2B dimer (blue) substructure of the nucleosome, but does not show individual histones.

FIGURE 8.10 The crystal structure of the histone core octamer is represented in a space-filling model with the H3₂-H4₂ tetramer shown in white and the H2A-H2B dimers shown in blue. Only one of the H2A-H2B dimers is visible in the top view, because the other is hidden underneath. The path of the DNA is modeled in green.

Photos courtesy of E. N. Moudrianakis, the Johns Hopkins University.

In the top view, you can see that the H3₂-H4₂ tetramer accounts for the diameter of the octamer. It forms the shape of a horseshoe. The H3₂-H4₂ tetramer alone can organize DNA in vitro into particles that display some of the properties of the core particle. The H2A-H2B pairs fit in as two dimers, but you can see only one in this view. In the side view, we can distinguish the responsibilities of the H3₂-H4₂ tetramer and of the separate H2A-H2B dimers. The protein forms a sort of spool, with a superhelical path that corresponds to the binding site for DNA, which is wound in about one and two-thirds turns in a nucleosome. The model displays twofold symmetry about an axis that would run perpendicular through the side view.

All four core histones show a similar type of structure in which three helices are connected by two loops. This highly conserved structure is called the histone fold, which you can see in FIGURE 8.11. These regions interact to form crescent-shaped heterodimers; each heterodimer binds 2.5 turns of the DNA double helix. Consistent with the need to package any DNA irrespective of sequence, binding is mostly to the phosphodiester backbone through a combination of salt links and hydrogen bonding interactions. In addition, an arginine side chain enters the minor groove of DNA at each of the 14 times it faces the octamer surface. FIGURE 8.12 shows a high-resolution view of the nucleosome (based on the crystal structure at 2.8 Å). The H3₂-H4₂ tetramer is formed by interactions between the two H3 subunits, as you can see at the top of the nucleosome (in green) in the left panel of Figure 8.12. The association of the two H2A-H2B dimers on opposite faces of the nucleosome is visible in the right panel (in turquoise and yellow).

FIGURE 8.11 The histone fold (a) consists of two short α-helices flanking a longer α-helix. Histone pairs (H3 + H4 and H2A + H2B) interact to form histone dimers (b).

Data from: Arents, G., et al. 1991. “Structures from Protein Data Bank 1HIO.” Proc Natl Acad Sci USA 88:10145–10152.

FIGURE 8.12 The crystal structure of the histone core octamer is represented in a ribbon model, including the 146-bp DNA phosphodiester backbones (orange and blue) and eight histone protein main chains (green: H3; purple: H4; turquoise: H2A; yellow: H2B).

Data from: Luger, K., et al. 1997. “Structures from Protein Data Bank 1AOI.” Nature 389:251–260.

Each of the core histones has a histone fold domain that contributes to the central protein mass of the nucleosome, sometimes referred to as the globular core. Each histone also has a flexible N-terminal tail (H2A and H2B have C-terminal tails, as well), which contains sites for covalent modification that are important in chromatin function. The tails, which account for about one-quarter of the protein mass, are too flexible to be visualized by X-ray crystallography; therefore, their positions in the nucleosome are not well defined, and they are generally depicted schematically, as shown in FIGURE 8.13. However, the points at which the tails exit the nucleosome core are known, and we can see the tails of both H3 and H2B passing between the turns of the DNA super-helix and extending out of the nucleosome, as shown in FIGURE 8.14. The tails of H4 and H2A extend from both faces of the nucleosome. When histone tails are crosslinked to DNA by UV irradiation, more products are obtained with nucleosomes compared to core particles, which could mean that the tails contact the linker DNA. The tail of H4 is able to contact an H2A-H2B dimer in an adjacent nucleosome, which might contribute to the formation of higher-order structures (see the section The Path of Nucleosomes in the Chromatin Fiber later in this chapter).

FIGURE 8.13 The histone fold domains of the histones are located in the core of the nucleosome. The N- and C-terminal tails, which carry many sites for modification, are flexible and their positions cannot be determined by crystallography.

FIGURE 8.14 The histone tails are disordered and exit from both faces of the nucleosome and between turns of the DNA. Note this figure shows only the first few amino acids of the tails, because the complete tails were not present in the crystal structure.

Data from: Luger, K., et al. 1997. “Structure from Protein Data Bank 1AOI.” Nature 389:251–260.

The linker histones also play an important role in the formation of higher-order chromatin structures. The linker histone family, typified by histone H1, comprises a set of closely related proteins that show appreciable variation among tissues and among species. The role of H1 is different from that of the core histones. H1 can be removed without affecting the structure of the nucleosome, consistent with a location external to the particle, and only a subset of nucleosomes is associated with linker histones in vivo. Nucleosomes that contain linker histones are sometimes referred to as chromatosomes.

The precise interaction of histone H1 with the nucleosome is somewhat controversial. H1 is retained on nucleosome monomers that have at least 165 bp of DNA, but does not bind to the 146-bp core particle. The binding of H1 to a nucleosome also facilitates the wrapping of two full turns of DNA. This is consistent with the localization of H1 in the region of the linker DNA immediately adjacent to the core DNA. Although the precise positioning of linker histones remains somewhat controversial, protein crosslinking and structural studies are consistent with a model whereby H1 interacts with either the entry or exit DNA in addition to the central turn of DNA on the nucleosome, as shown in FIGURE 8.15. In this position, H1 has the potential to influence the angle of DNA entry or exit, which might contribute to the formation of higher-order structures (see the section The Path of Nucleosomes in the Chromatin Fiber later in this chapter).

FIGURE 8.15 Possible model for the interaction of histone H1 with the nucleosome. H1 can interact with the central gyre of the DNA at the dyad axis, as well as with the linker DNA at either the entry or exit.

8.4 Nucleosomes Are Covalently Modified

All of the histones are subject to numerous covalent modifications, most of which occur in the histone tails. Researchers can modify all of the histones at numerous sites by methylation, acetylation, or phosphorylation, as shown schematically in FIGURE 8.16. Even though these modifications are relatively small, other, more dramatic modifications occur, as well, such as mono-ubiquitylation, sumoylation, and ADP-ribosylation. Although different histone modifications have known roles in replication, chromatin assembly, transcription, splicing, and DNA repair, researchers have yet to characterize functions of a number of specific modifications.

FIGURE 8.16 The histone tails can be acetylated, methylated, phosphorylated, and ubiquitylated at numerous sites. Not all possible modifications are shown.

Data from: The Scientist 17 (2003):p. 27.

Lysines in the histone tails are the most common targets of modification. Acetylation, methylation, ubiquitylation, and sumoylation all occur on the free epsilon (ε) amino group of lysine. As shown in FIGURE 8.17, acetylation neutralizes the positive charge that resides on the NH3 form of the ε-amino group. In contrast, lysine methylation retains the positive charge, and lysine can be mono-, di-, or trimethylated. Arginine can be mono- or dimethylated. Phosphorylation occurs on the hydroxyl group of serine and threonine. This introduces a negative charge in the form of the phosphate group.

FIGURE 8.17 The positive charge on lysine is neutralized upon acetylation, whereas methylated lysine and arginine retain their positive charges. Lysine can be mono-, di-, or triacetylated, whereas arginine can be mono- or diacetylated. Serine or threonine phosphorylation results in a negative charge.

All of these modifications are reversible, and a given modification might exist only transiently, or can be maintained stably through multiple cell divisions. Some modifications change the charge of the protein molecule, and, as a result, they are potentially able to change the functional properties of the octamers. For example, extensive lysine acetylation reduces the overall positive charge of the tails, leading to release of the tails from interactions with DNA on their own or other nucleosomes. Modification of histones is associated with structural changes that occur in chromatin at replication and transcription, and specific modifications also facilitate DNA repair. Modifications at specific positions on specific histones can define different functional states of chromatin. Newly synthesized core histones carry specific patterns of acetylation that are removed after the histones are assembled into chromatin, as shown in FIGURE 8.18. Other modifications are dynamically added and removed to regulate transcription, replication, repair, and chromosome condensation. These other modifications are usually added and removed from histones that are incorporated into chromatin, as depicted for acetylation in FIGURE 8.19.

FIGURE 8.18 Acetylation during replication occurs on specific sites on histones before they are incorporated into nucleosomes.

FIGURE 8.19 Acetylation associated with gene activation occurs by directly modifying specific sites on histones that are already incorporated into nucleosomes.

The specificity of the modifications is controlled by the fact that many of the modifying enzymes have individual target sites in specific histones. TABLE 8.1 summarizes the effects of some of the modifications that occur on histones H3 and H4. Many modified sites are subject to only a single type of modification in vivo, but others can be subject to alternative modification states (such as lysine 9 of histone H3, which is acetylated or methylated under different conditions). In some cases, modification of one site might activate or inhibit modification of another site. The idea that combinations of signals can be used to define chromatin function led to the idea of a histone code. Although the use of the word “code” has been controversial, this key hypothesis proposes that the collective impact of multiple modifications at particular sites defines the function of a chromatin domain. These modifications are not restricted to a single histone; the functional state of a region of chromatin is derived from all the modifications within a nucleosome or set of nucleosomes. Some modifications of particular histone residues can also prevent or promote other specific histone modification events (or even modification of nonhistone proteins); these “cross-talk” pathways add another level of complexity to signaling through chromatin.

TABLE 8.1 Most modified sites in histones have a single, specific type of modification, but some sites can have more than one type of modification. Individual functions can be associated with some of the modifications.

Histone	Site	Modification	Function
H3	K-4	Acetylation	Transcription activation
H3	K-9	Methylation	Transcription repression
	K-9	Methylation	Promotes DNA methylation
	K-9	Acetylation	Transcription activation
H3	S-10	Phosphorylation	Chromosome condensation
	S-10	Phosphorylation	Transcription activation
H3	K-14	Acetylation	Transcription activation
H3	K-36	Methylation	Transcription repression
H3	K-79	Methylation	Transcription activation
H3	K-27	Methylation	Transcription repression
H4	R-3	Methylation	Transcription activation
H4	K-5	Acetylation	Nucleosome assembly
H4	K-16	Acetylation	Chromatin fiber folding
	K-16	Acetylation	Transcription activation
H2A	K-119	Ubiquitination	Transcription repression

Whereas some histone modifications can directly alter the structure of chromatin, a major function of histone modification lies in the creation of binding sites for nonhistone proteins that change the properties of chromatin. In recent years, a number of protein domains have been identified that bind to specifically modified histone tails. A few examples are provided here.

The bromodomain is found in a variety of proteins that interact with chromatin. Bromodomains recognize acetylated lysine, and different bromodomain-containing proteins recognize different acetylated targets. The bromodomain itself recognizes only a very short sequence of four amino acids, including the acetylated lysine, so specificity for target recognition must depend on interactions involving other regions. FIGURE 8.20 shows the structure of a bromodomain bound to its acetylated lysine target. The bromodomain is found in a range of proteins that interact with chromatin, including components of the transcription apparatus and some of the enzymes that remodel or modify histones (discussed in the chapter titled Eukaryotic Transcription Regulation).

FIGURE 8.20 Bromodomains are protein motifs that bind acetyl-lysines. The bromodomain fold consists of a cluster of four α-helices with an acetyl-lysine binding pocket at one end. This figure shows the bromodomain of yeast Gcn5 bound to an H4K16ac peptide.

Data from: Owen, D. J., et al. 2000. “Structure from Protein Data Bank 1E6I.” EMBO J 19:6141–6149.

Methylated lysines (and arginines) are recognized by a number of different domains, which not only can recognize specific modified sites but also can distinguish between mono-, di-, or trimethylated lysines. The chromodomain is a common protein motif of 60 amino acids present in a number of chromatin-associated proteins. Researchers have identified a number of other methyl-lysine binding domains, as shown in FIGURE 8.21, such as the plant homeodomain (PHD) and the Tudor domain; the number of different motifs designed to recognize particular methylated sites emphasizes the importance and complexity of histone modifications.

FIGURE 8.21 Numerous protein motifs recognize methylated lysines. (a) The chromodomain of HP1 binds trimethylated K9 of histone H3. (b) The Tudor domain of JMJD2A binds trimethylated K4 of histone H3. Chromodomains and Tudor domains are members of the “royal superfamily,” which bind their targets via a partial β-barrel structure. (c) The PHD finger of BPTF also binds trimethylated K4 of histone H3, using a structure related to DNA-binding zinc finger domains.

(a) Data from: Jacobs, S. A., and Khorasanizadeh, S. 2002. “Structure from Protein Data Bank 1KNE.” Science 295:2080–2083.
(b) Data from: Y. Huang, et al. 2006. “Structure from Protein Data Bank 2GFA.” Science 12:748–751.
(c) Photo courtesy of Sean D. Taverna, the Johns Hopkins University School of Medicine, and Haitao Li, Memorial Sloan-Kettering Cancer Center. Additional information at: Taverna, S. D., et al., Nat Struct Mol Biol 14:1025–1040.

The idea that combinations of modifications are critical, as proposed in the histone code hypothesis, has been reinforced by discoveries of proteins or complexes that can recognize multiple sites of modification simultaneously. For example, some proteins have tandem bromodomains or chromodomains with particular spacing, which can promote binding to histones that are acetylated or methylated at two specific sites. There are also cases in which modification at one site can prevent a protein from recognizing its target modification at another site. It is clear that the effects of a single modification might not always be predictable, and the context of other modifications must be accounted for in order to assign a function to a region of chromatin.

8.5 Histone Variants Produce Alternative Nucleosomes

Whereas all nucleosomes share a related core structure, some nucleosomes exhibit subtle or dramatic differences resulting from the incorporation of histone variants. Histone variants comprise a large group of histones that are related to the histones we have already discussed, but have differences in sequence from the “canonical” histones. These sequence differences can be small (as few as four amino acid differences) or extensive (such as alternative tail sequences).

Variants have been identified for all core histones except histone H4. FIGURE 8.22 summarizes the best characterized histone variants. Most variants have significant differences between them, particularly in the N- and C-terminal tails. At one extreme, macroH2A is nearly three times larger than conventional H2A and contains a large C-terminal tail that is not related to any other histone. At the other end of the spectrum, canonical H3 (also known as H3.1) differs from the H3.3 variant at only four amino acid positions—three in the histone core and one in the N-terminal tail.

FIGURE 8.22 The major core histones contain a conserved histone-fold domain. In the histone H3.3 variant, the residues that differ from the major histone H3 (also known as H3.1) are highlighted in yellow. The centromeric histone CenH3 has a unique N terminus, which does not resemble other core histones. Most H2A variants contain alternative C-termini, except H2ABbd, which contains a distinct N terminus. The sperm-specific SpH2B has a long N-terminus. Proposed functions of the variants are listed.

Data from: Sarma, K., and Reinberg, D. 2005. Nat Rev Mol Cell Biol 6:139–149.

Histone variants have been implicated in a number of different functions, and their incorporation changes the nature of the chromatin containing the variant. We have previously discussed one type of histone variant, the centromeric H3 (or CenH3) histone, known as Cse4 in yeast. CenH3 histones are incorporated into specialized nucleosomes present at centromeres in all eukaryotes (see the chapter titled Chromosomes). There remains a spirited debate over the structure and composition of centromeric nucleosomes. In one model, CenH3 nucleosomes contain a normal octameric histone core, containing two copies of the CenH3. However, compelling evidence in budding yeast supports an alternative model in which centromeric nucleosomes consist of “hemisomes” containing one copy each of Cse4, H4, H2A, and H2B. Whether one or both models are correct will likely involve further investigation.

The other major H3 variant is histone H3.3. In multicellular eukaryotes, this variant is a minority component of the total H3 in the cell, but in yeast, the major H3 is actually of the H3.3 type. H3.3 is expressed throughout the cell cycle, in contrast to most histones that are expressed during S phase, when new chromatin assembly is required during DNA replication. As a result, H3.3 is available for assembly at any time in the cell cycle and is incorporated at sites of active transcription, where nucleosomes become disrupted. For this reason, H3.3 is often referred to as a “replacement” histone, in contrast to the “replicative” histone H3.1 (see the section Replication of Chromatin Requires Assembly of Nucleosomes later in this chapter).

The H2A variants are the largest and most diverse family of core histone variants, and have been implicated in a variety of distinct functions. One that has been extensively studied is the variant H2AX. The H2AX variant is normally present in only 10%–15% of the nucleosomes in multicellular eukaryotes, though again (like H3.3) this subtype is the major H2A present in yeast. It has a C-terminal tail that is distinct from the canonical H2A, characterized by a SQEL/Y motif at the end. This motif is the target of phosphorylation by ATM/ATR kinases, activated by DNA damage, and this histone variant is involved in DNA repair, particularly repair of double-strand breaks (see the chapter titled Repair Systems). H2AX phosphorylated at the SQEL/Y motif is sometimes referred to as “γ-H2AX” and is required to stabilize binding of various repair factors at DNA breaks and to maintain checkpoint arrest. γ-H2AX appears within moments at broken DNA ends, as demonstrated in FIGURE 8.23, which shows a cartoon of foci of γ-H2AX forming along the path of double-strand breaks induced by a laser.

FIGURE 8.23 γ-H2AX is detected by an antibody (yellow) and appears along the path traced by a laser that produces double-strand breaks (white line).

© Rogakou et al., 1999. Originally published in The Journal of Cell Biology, 146: 905-915. Photo courtesy of William M. Bonner, National Cancer Institute, NIH.

Other H2A variants have different roles. Researchers have shown the H2AZ variant, which has ~60% sequence identity with canonical H2A, to be important in several processes, such as gene activation, heterochromatin–euchromatin boundary formation, cell-cycle progression, and it can be enriched at the centromere, at least in some species. The vertebrate-specific macroH2A is named for its extremely long C-terminal tail, which contains a leucine-zipper dimerization motif that might mediate chromatin compaction by facilitating internucleosome interactions. Mammalian macroH2A is enriched in the inactive X chromosome in females, which is assembled into a silent, heterochromatic state. In contrast, the mammalian H2ABbd variant is excluded from the inactive X and forms a less stable nucleosome than canonical H2A; perhaps this histone is designed to be more easily displaced in transcriptionally active regions of euchromatin.

Still other variants are expressed in limited tissues, such as spH2B, which is present in sperm and required for chromatin compaction. The presence and distribution of histone variants shows that individual chromatin regions, entire chromosomes, or even specific tissues can have unique “flavors” of chromatin specialized for different functions. FIGURE 8.24 is a schematic illustrating some typical distribution patterns of some of the better characterized histone variants. In addition, the histone variants, like the canonical histones, are subject to numerous covalent modifications, adding levels of complexity to the roles chromatin plays in nuclear processes.

FIGURE 8.24 Some histone variants are spread throughout all or most of the chromosome, whereas others show specific distribution patterns. Characteristic patterns are shown for several histone variants on a cartoon autosome. Note that histone variant distributions can be dramatically different on dosage-compensated sex chromosomes (like the mammalian inactive X), in sperm chromatin, or other highly specialized chromatin states.

8.6 DNA Structure Varies on the Nucleosomal Surface

So far, we have focused on the protein components of the nucleosome. The DNA wrapped around these proteins is in an unusual conformation. The exposure of DNA on the surface of the nucleosome explains why it is accessible to cleavage by certain nucleases. The reaction with nucleases that attack single strands has been especially informative. The enzymes DNase I and DNase II make single-strand nicks in DNA; they cleave a bond in one strand, but the other strand remains intact. No effect is visible in linear double-stranded DNA, but when this DNA is denatured, shorter fragments are released instead of full-length single strands. If the DNA has been labeled at its ends, the end fragments can be identified by detection of the label, as summarized in FIGURE 8.25. When DNA is free in solution, it is nicked (relatively) at random. The DNA on nucleosomes can also be nicked by the enzymes, but only at regular intervals. When the points of cutting are determined by using end-labeled DNA and the DNA is denatured and electrophoresed, a ladder of the sort displayed in FIGURE 8.26 is obtained.

FIGURE 8.25 Nicks in double-stranded DNA are revealed by fragments when the DNA is denatured to give single strands. For example, if the DNA is labeled at the 5′ ends, only the 5′ fragments are visible by autoradiography. The size of the fragment identifies the distance of the nick from the labeled end.

FIGURE 8.26 Sites for nicking lie at regular intervals along core DNA, as seen in a DNase I digest of nuclei.

Photo courtesy of Leonard C. Lutter, Molecular Biology Research Program, Henry Ford Hospital.

The interval between successive steps on the ladder is 10–11 bases. The ladder extends for the full distance of core DNA. The cleavage sites are numbered as S1 through S12 (where S1 is 10–11 bases from the labeled 5′ end, S2 is about 20 bases from it, and so on). The enzymes DNase I and DNase II generate essentially the same ladder, and the same pattern is obtained by cleaving with a hydroxyl radical, which argues that the pattern reflects the structure of the DNA itself rather than any sequence preference. The sensitivity of nucleosomal DNA to nucleases is analogous to a footprinting experiment. Thus, we can assign the lack of reaction at particular target sites to the structure of the nucleosome, in which certain positions on DNA are rendered inaccessible.

There are two strands of DNA in the core particle, so in an end-labeling experiment both of the 5′ (or 3′) ends are labeled, one on each strand. Thus, the cutting pattern includes fragments derived from both strands. This is visible in Figure 8.25, in which each labeled fragment is derived from a different strand. The corollary is that, in an experiment, each labeled band might actually represent two fragments that are generated by cutting the same distance from either of the labeled ends.

How, then, should we interpret discrete preferences at particular sites? One view is that the path of DNA on the particle is symmetrical (about a horizontal axis through the nucleosome, as illustrated in Figure 8.7). If, for example, no 80-base fragment is generated by DNase I, this must mean that the position at 80 bases from the 5′ end of either strand is not susceptible to the enzyme.

When DNA is immobilized on a flat surface, sites are cut with a regular separation. FIGURE 8.27 shows that this reflects the recurrence of the exposed site with the helical periodicity of B-form DNA. The cutting periodicity (the spacing between cleavage points) coincides with—indeed, is a reflection of—the structural periodicity (the number of base pairs per turn of the double helix). Thus, the distance between the sites corresponds to the number of base pairs per turn. Measurements of this type yield the average value for double-helical B-type DNA of 10.5 bp/turn.

FIGURE 8.27 The most exposed positions on DNA recur with a periodicity that reflects the structure of the double helix. (For clarity, sites are shown for only one strand.)

A similar analysis of DNA on the surface of the nucleosome reveals striking variations in the structural periodicity at different points. At the ends of the DNA, the average distance between pairs of DNase I digestion sites is about 10.0 bases each, significantly less than the usual 10.5 bp/turn. In the center of the particle, the separation between cleavage sites averages 10.7 bases. This variation in cutting periodicity along the core DNA means that there is variation in the structural periodicity of core DNA. The DNA has more bp/turn than its solution value in the middle, but has fewer bp/turn at the ends. The average periodicity over the entire nucleosome is only 10.17 bp/turn, which is significantly less than the 10.5 bp/turn of DNA in solution.

The crystal structure of the core particle (Figure 8.12) shows that DNA is wound into a solenoidal (spring-shaped) supercoil, with 1.67 turns wound around the histone octamer. The pitch of the superhelix varies and has a discontinuity in the middle. Regions of high curvature are arranged symmetrically and are the sites least sensitive to DNase I.

The high-resolution structure of the nucleosome core shows in detail how the structure of DNA is distorted. Most of the supercoiling occurs in the central 129 bp, which are coiled into 1.59 left-handed superhelical turns with a diameter of 80 Å (only four times the diameter of the DNA duplex itself). The terminal sequences on either end make only a very small contribution to the overall curvature.

The central 129 bp are in the form of B-DNA, but with a substantial curvature that is needed to form the superhelix. The major groove is smoothly bent, but the minor groove has abrupt kinks, as shown in FIGURE 8.28. These conformational changes might explain why the central part of nucleosomal DNA is not usually a target for binding by regulatory proteins, which typically bind to the terminal parts of the core DNA or to the linker sequences.

FIGURE 8.28 DNA structure in nucleosomal DNA. (a) The trace of the DNA backbone in the nucleosome is shown in the absence of protein for clarity. (b) Regions of curvature in nucleosomal DNA. Actual structures (left) and schematic representations (right) show uniformity of curvature along the major groove (blue) and both smooth and kinked bending into the minor groove (orange). Also indicated are the DNA axes for the experimental (pink) and ideal (gray) superhelices.

(a) Data from: Muthurajan, U. M., et al. 2004. “Structures from Protein Data Bank: 1P34.” EMBO J 23:260–271.
(b) Data from: Richmond, T. J., and Davey, C. A. 2003. Nature 423:145–150.

Some insights into the structure of nucleosomal DNA emerge when we compare predictions for supercoiling in the path that DNA follows with actual measurements of supercoiling of nucleosomal DNA. Circular “minichromosomes” that are fully assembled into nucleosomes can be isolated from eukaryotic cells. Researchers can measure the degree of supercoiling on the individual nucleosomes of the minichromosome as illustrated in FIGURE 8.29. First, the free supercoils of the minichromosome itself are relaxed, so that the nucleosomes form a circular string with an unconstrained superhelical density of 0. Next, the histone octamers are extracted. This releases the DNA to follow a free path. Every negative supercoil that was present but constrained in the nucleosomes will appear in the deproteinized DNA as −1 turn. Now the total number of supercoils in the DNA is measured.

FIGURE 8.29 The supercoils of the SV40 minichromosome can be relaxed to generate a circular structure, whose loss of histones then generates supercoils in the free DNA.

The observed value is close to the number of nucleosomes. Thus, the DNA follows a path on the nucleosomal surface that generates about one negative supercoiled turn when the restraining protein is removed. The path that DNA follows on the nucleosome, however, corresponds to −1.67 superhelical turns. This discrepancy is sometimes called the linking number paradox.

The discrepancy is explained by the difference between the 10.17 average bp/turn of nucleosomal DNA and the 10.5 bp/turn of free DNA. In a nucleosome of 200 bp, there are 200/10.17 = 19.67 turns. When DNA is released from the nucleosome, it now has 200/10.5 = 19.0 turns. The path of the less tightly wound DNA on the nucleosome absorbs −0.67 turns, which explains the discrepancy between the physical path of −1.67 and the measurement of −1.0 superhelical turns. In effect, some of the torsional strain in nucleosomal DNA goes into increasing the number of bp/turn; only the rest is left to be measured as a supercoil.

8.7 The Path of Nucleosomes in the Chromatin Fiber

When chromatin is released from nuclei and examined with an electron microscope, we can see two types of fibers: the 10-nm fiber and the 30-nm fiber. They are described by the approximate diameter of the thread (that of the 30-nm fiber actually varies from around 25–30 nm). The 10-nm fiber is essentially a continuous string of nucleosomes and represents the least compacted level of chromatin structure. In fact, a stretched-out 10-nm fiber resembles a string of beads in which we can clearly distinguish nucleosomes connected by linker DNA, as demonstrated in FIGURE 8.30. The 10-nm fiber structure is obtained under conditions of low ionic strength and does not require the presence of histone H1. This means that it is a function strictly of the nucleosomes themselves. FIGURE 8.31 shows a depiction of the continuous series of nucleosomes in this fiber.

FIGURE 8.30 The 10-nm fiber in partially unwound state can be seen to consist of a string of nucleosomes.

Photo courtesy of Barbara Hamkalo, University of California, Irvine.

FIGURE 8.31 The 30-nm fiber is a two-start helix consisting of two rows of nucleosomes coiled into a solenoid.

Reprinted from Cell, vol. 128, D. J. Tremethick, Higher-order structure of chromatin …, pp. 651–654. Copyright 2007, with permission from Elsevier [http://www.sciencedirect.com/science/journal/00928674].

When chromatin is visualized in conditions of greater ionic strength, the 30-nm fiber is obtained. An example is given in FIGURE 8.32. You can see that the fiber has an underlying coiled structure. It has approximately 6 nucleosomes for every turn, which corresponds to a packing ratio of 40 (i.e., each mm along the axis of the fiber contains 40 mm of DNA). The formation of this fiber requires the histone tails, which are involved in internucleosomal contacts, and is facilitated by the presence of a linker histone such as H1.

FIGURE 8.32 The 30-nm fiber has a coiled structure.

Photo courtesy of Barbara Hamkalo, University of California, Irvine.

Nucleosomes are arranged into a helical array within the 30-nm fiber, with the linker DNA occupying the central cavity. The two main forms of this helical structure are a single start solenoid, which forms a linear array, and a two-start zigzag that in effect consists of a double row of nucleosomes. FIGURE 8.33 shows a two-start model suggested by crosslinking data identifying a double stack of nucleosomes in the 30-nm fiber. Although this model is also supported by the crystal structure of a tetranucleosome complex, recent studies suggest that the type of helical structure (e.g., one-start solenoid or two-start zigzag) is influenced by the length of linker DNA within the 10-nm fiber. Furthermore, biochemical studies suggest that 30-nm fibers might contain a heterogeneous mixture of one-start and two-start helical organizations, rather than a single, uniform structure.

FIGURE 8.33 The 10-nm fiber is a continuous string of nucleosomes.

Levels of folding beyond the 30-nm fiber are very poorly understood, but it has long been believed that the 40-fold compaction provided by the 30-nm fiber is still a long way from the levels of compaction required for interphase or mitotic packaging of chromosomes. Researchers have observed chromatin fibers with diameters of 60–300 nm (called chromonema fibers) by both light and electron microscopy. Such fibers were presumed to consist of folded 30-nm fibers and would represent a major level of compaction (a 30-nm fiber running just across the width of a 100-nm fiber would contain more than 10 kb of DNA), but the actual substructures of these large fibers remain unknown. Indeed, recent microscopy studies do not detect significant levels of 30-nm fibers within chromatin in situ, suggesting that 30-nm fibers might exist only in regions of low chromatin density (or maybe not at all!). In contrast, several studies have provided compelling evidence that even highly condensed mitotic chromatin might be composed of only 10-nm fibers, densely packed into an interdigitated “polymer melt” or “fractal globule.” This type of organization facilitates a dense packaging of DNA while preserving the ability to fold and unfold genomic loci. FIGURE 8.34 shows a hypothetical depiction of this higher-order folding model.

FIGURE 8.34 A model for higher order chromatin structure involving interdigitation of 10-nm chromatin fibers. The resulting fractal globule allows for reversible extrusion of individual fibers for nuclear functions such as transcription.

How can genomic DNA fit into the nuclear volume if organization into 10-nm fibers provides only a 6-fold compaction ratio? Historically, we have thought about DNA packaging into the nucleus from the point of view of linear compaction—if DNA is stretched end-to-end, it must be shortened by about 10,000-fold to form a mitotic chromosome. This led to the popular idea of hierarchical levels of chromatin folding (e.g., 10-nm → 30-nm → 60- to 300-nm fibers). However, if genomic DNA is modeled as a simple cylinder, the volume of DNA in a diploid mammalian nucleus is actually less than 6% of the nuclear volume. Wrapping DNA around histones actually takes up more space! In this view, the role of chromatin organization is not to compact linear DNA into the nuclear space, rather it is to help oppose the negative charge of DNA and facilitate the folding and bending of DNA on itself. In this view, the extended 10-nm fiber is highly flexible and can not only bend and kink but also self-associate to form dense networks that satisfy nuclear packaging requirements.

8.8 Replication of Chromatin Requires Assembly of Nucleosomes

Replication separates the strands of DNA and therefore must inevitably disrupt the structure of the nucleosome. However, this disruption is confined to the immediate vicinity of the replication fork. As soon as DNA has been replicated, nucleosomes are quickly generated on both of the duplicates. The transience of the replication event is a major difficulty in analyzing the structure of a particular region while it is being replicated.

Replication of chromatin does not involve any protracted period during which the DNA is free of histones. This point is illustrated by the electron micrograph of FIGURE 8.35, which shows a recently replicated stretch of DNA that is already packaged into nucleosomes on both daughter duplex segments.

FIGURE 8.35 Replicated DNA is immediately incorporated into nucleosomes.

Photo courtesy of Steven L. McKnight, UT Southwestern Medical Center at Dallas.

Biochemical analysis and visualization of the replication fork indicate that the disruption of nucleosome structure is limited to a short region immediately around the fork. Progress of the fork disrupts nucleosomes, but they form very rapidly on the daughter duplexes as the fork moves forward. In fact, the assembly of nucleosomes is directly linked to the replisome that is replicating DNA.

How do histones associate with DNA to generate nucleosomes? Do the histones preform a protein octamer around which the DNA is subsequently wrapped? Or, does the histone octamer assemble on DNA from free histones? Researchers can use either of these pathways in vitro to assemble nucleosomes, depending on the conditions that are employed. In one pathway, a preformed octamer binds to DNA. In the other pathway, a tetramer of H3₂-H4₂ binds first, and then two H2A-H2B dimers are added. This latter stepwise assembly is the pathway that is used in replication, shown in FIGURE 8.36.

FIGURE 8.36 During nucleosome assembly in vivo, H3-H4 tetramers form and bind DNA first, then two H2A-H2B dimers are added to form the complete nucleosome.

Accessory proteins are involved in assisting histones to associate with DNA. Accessory proteins can act as “molecular chaperones” that bind to the histones in order to release either individual histones or complexes (H3₂-H4₂ or H2A-H2B) to the DNA in a controlled manner. This could be necessary because the histones, as basic proteins, have a generally high affinity for DNA. Such interactions allow histones to form nucleosomes without becoming trapped in other kinetic intermediates (i.e., other complexes resulting from indiscreet binding of histones to DNA).

Researchers have identified numerous histone chaperones. Chromatin assembly factor (CAF)-1 and anti-silencing function 1 (ASF1) are two chaperones that function at the replication fork. CAF-1 is a conserved three-subunit complex that is directly recruited to the replication fork by proliferating cell nuclear antigen (PCNA), the processivity factor for DNA polymerase. ASF1 interacts with the replicative helicase that unwinds the replication fork. Furthermore, CAF-1 and ASF1 interact with each other. These interactions provide the link between replication and nucleosome assembly, ensuring that nucleosomes are assembled as soon as DNA has been replicated.

CAF-1 acts stoichiometrically, and functions by binding to newly synthesized H3 and H4. New nucleosomes form by assembling first the H3₂-H4₂ tetramer, and then adding the H2A-H2B dimers. ASF1 appears to play an important role in transfer of parental nucleosomes from ahead of the replication fork to the newly synthesized region behind the fork, although ASF1 can bind and assemble newly synthesized histones, as well.

The pattern of disassembly and reassembly has been difficult to characterize in detail, but a working model is illustrated in FIGURE 8.37. The replication fork displaces histone octamers, which then dissociate into H3₂-H4₂ tetramers and H2A-H2B dimers. These “old” tetramers and dimers enter a pool that also includes “new” tetramers and dimers, which are assembled from newly synthesized histones. Nucleosomes assemble ~600 bp behind the replication fork. Assembly is initiated when H3₂-H4₂ tetramers bind to each of the daughter duplexes, assisted by CAF-1 or ASF1. Two H2A-H2B dimers then bind to each H3₂-H4₂ tetramer to complete the histone octamer. The assembly of tetramers and dimers is random with respect to “old” and “new” subunits. It appears that nucleosomes are disrupted and reassembled in a similar way during transcription, though different histone chaperones are involved in this process (see the section Nucleosomes Are Displaced and Reassembled During Transcription later in this chapter).

FIGURE 8.37 Replication fork passage displaces histone octamers from DNA. They disassemble into H3-H4 tetramers and H2A-H2B dimers. H3-H4 tetramers (blue) are directly transferred behind the replication forks. Newly synthesized histones (orange) are assembled into H3-H4 tetramers and H2A-H2B dimers. The old and new tetramers and dimers are assembled with the aid of histone chaperones into new nucleosomes immediately behind the replication fork. H2A-H2B dimers are omitted from the figure for simplicity; chaperones responsible for dimer assembly have not been identified.

Data from: Rocha, W., and Verreault, A. 2008. FEBS Lett 582:1938–1949.

During S phase (the period of DNA replication) in a eukaryotic cell, the duplication of chromatin requires synthesis of sufficient histone proteins to package an entire genome—basically the same quantity of histones must be synthesized that are already contained in nucleosomes. The synthesis of histone mRNAs is controlled as part of the cell cycle, and increases enormously in S phase. The pathway for assembling chromatin from this equal mix of old and new histones during S phase is called the replication-coupled pathway.

Another pathway, called the replication-independent pathway, exists for assembling nucleosomes during other phases of the cell cycle, when DNA is not being synthesized. This might become necessary as the result of damage to DNA or because nucleosomes are displaced during transcription. The assembly process must necessarily have some differences from the replication-coupled pathway, because it cannot be linked to the replication apparatus. The replication-independent pathway uses the histone H3.3 variant, which was introduced earlier in the section Histone Variants Produce Alternative Nucleosomes.

The histone H3.3 variant differs from the highly conserved H3 histone at four amino acid positions (see Figure 8.20). H3.3 slowly replaces H3 in differentiating cells that do not have replication cycles. This happens as the result of assembly of new histone octamers to replace those that have been displaced from DNA for whatever reason. The mechanism that is used to ensure the use of H3.3 in the replication-independent pathway is different in two cases that have been investigated.

In the protozoan Tetrahymena, histone usage is determined exclusively by availability. Histone H3 is synthesized only during the cell cycle; the variant replacement histone is synthesized only in nonreplicating cells. In Drosophila, however, there is an active pathway that ensures the usage of H3.3 by the replication-independent pathway. New nucleosomes containing H3.3 assemble at sites of transcription, presumably replacing nucleosomes that were displaced by RNA polymerase. The assembly process discriminates between H3 and H3.3 on the basis of their sequences, specifically excluding H3 from being utilized. By contrast, replication-coupled assembly uses both types of H3 (although H3.3 is available at much lower levels than H3 and therefore enters only a small proportion of nucleosomes).

CAF-1 is not involved in replication-independent assembly. (There also are organisms such as yeast and Arabidopsis for which its gene is not essential, implying that alternative assembly processes can be used in replication-coupled assembly.) Instead, replication-independent assembly uses a factor called HIRA, named for histone cell cycle regulator (HIR), genes in yeast. Depletion of HIRA from in vitro systems for nucleosome assembly inhibits the formation of nucleosomes on nonreplicated DNA, but not on replicating DNA, which indicates that the pathways do indeed use different assembly mechanisms. Like CAF-1 and ASF1, HIRA functions as a chaperone to assist the incorporation of histones into nucleosomes. This pathway appears to be generally responsible for replication-independent assembly; for example, HIRA is required for the decondensation of the sperm nucleus, when protamines are replaced by histones, in order to generate chromatin that is competent to be replicated following fertilization.

As described earlier, assembly of nucleosomes containing an alternative to H3 also occurs at centromeres (see the Chromosomes chapter). Centromeric DNA replicates early during S phase. The incorporation of H3 at the centromeres is inhibited during replication; instead, a CenH3 variant is preferentially (though not exclusively) incorporated. Interestingly, new CenH3 is incorporated during early G1 in vertebrates, but in budding yeast the CenH3 is incorporated in S phase and is linked to replication. In both vertebrates and yeast, CenH3 incorporation requires a CenH3-specific chaperone, called HJURP (mammals) or Scm3 (budding yeast).

8.9 Do Nucleosomes Lie at Specific Positions?

Does a particular DNA sequence always lie in a certain position in vivo with regard to the topography of the nucleosome? Or, are nucleosomes arranged randomly on DNA so that a particular sequence can occur at any location—for example, in the core region in one copy of the genome and in the linker region in another?

To investigate this question, it is necessary to use a defined sequence of DNA; more precisely, we need to determine the position relative to the nucleosome of a defined point in the DNA. FIGURE 8.38 illustrates the principle of a procedure used to achieve this.

FIGURE 8.38 Nucleosome positioning places restriction sites at unique positions relative to the linker sites cleaved by micrococcal nuclease.

Suppose that the DNA sequence is organized into nucleosomes in only one particular configuration so that each site on the DNA always is located at a particular position on the nucleosome. This type of organization is called nucleosome positioning (or sometimes nucleosome phasing). In a series of positioned nucleosomes, the linker regions of DNA comprise unique sites.

Consider the consequences for just a single nucleosome. Cleavage with MNase generates a monomeric fragment that constitutes a specific sequence. If the DNA is isolated and cleaved with a restriction enzyme that has only one target site in this fragment, it should be cut at a unique point. This produces two fragments, each of unique size.

Researchers separate the products of the MNase/restriction enzyme double digest by gel electrophoresis. They then use a probe representing the sequence on one side of the restriction site to identify the corresponding fragment in the double digest. This technique is called indirect end labeling (because it is not possible to label the end of the nucleosomal DNA fragment itself, it must be detected indirectly with a probe).

Reversing the argument, the identification of a single sharp band demonstrates that the position of the restriction site is uniquely defined with respect to the end of the nucleosomal DNA (as defined by the MNase cut). Thus, the nucleosome has a unique sequence of DNA. If a given region contains an array of positioned nucleosomes, researchers can map the position of each by using this method. FIGURE 8.39 shows an example of a gene promoter containing an ordered array of nucleosomes. In this MNase map, numerous positioned nucleosomes can be identified, indicated by the ovals to the left. Note that the TATA box is covered by a nucleosome; in this example this gene is not transcriptionally active.

What happens if the nucleosomes do not lie at a single position? Now the linkers consist of different DNA sequences in each copy of the genome. Thus, the restriction site lies at a different position each time; in fact, it lies at all possible locations relative to the ends of the monomeric nucleosomal DNA. FIGURE 8.40 shows that the double cleavage then generates a broad smear, ranging from the smallest detectable fragment (~20 bases) to the length of the monomeric DNA. Although the indirect end-labeling method is appropriate for monitoring nucleosome positioning at individual loci, MNase digestion can also be combined with massively parallel DNA sequencing to define nucleosome locations on a genome-wide scale.

FIGURE 8.39 An MNase map of nucleosome positions in an inactive gene. The lanes from left to right have been treated with increasing amounts of MNase. The nucleosomes occupy the regions that lack cut sites (indicated by ovals) and are arranged in a well-ordered array. The position of the TATA box and the transcriptional start site (arrow) are indicated.

Figure courtesy of Dr. Jocelyn Krebs.

FIGURE 8.40 In the absence of nucleosome positioning, a restriction site can lie at any possible location in different copies of the genome. Fragments of all possible sizes are produced when a restriction enzyme cuts at a target site (red) and micrococcal nuclease cuts at the junctions between nucleosomes (green).

In discussing these experiments, we have treated MNase as an enzyme that cleaves DNA at the exposed linker regions without any sort of sequence specificity. MNase does have some sequence specificity, though, which is biased toward selection of A-T–rich sequences. Thus, we cannot assume that the existence of a specific band in the indirect end-labeling technique represents the distance from a restriction cut to the linker region. It could instead represent the distance from the restriction cut to a preferred micrococcal nuclease cleavage site.

This possibility is controlled by treating the naked DNA in exactly the same way as the chromatin. If there are preferred sites for MNase in the particular region, specific bands are found. Researchers can compare this pattern of bands with the pattern generated from chromatin.

A difference between the control DNA band pattern and the chromatin pattern provides evidence for nucleosome positioning. Some of the bands present in the control DNA digest might disappear from the nucleosome digest, indicating that preferentially cleaved positions are unavailable. New bands might appear in the nucleosome digest when new sites are rendered preferentially accessible by the nucleosomal organization.

Nucleosome positioning might be accomplished in either of two ways:

Intrinsic mechanisms: Nucleosomes are deposited specifically at particular DNA sequences, or are excluded by specific sequences. This modifies our view of the nucleosome as a subunit able to form between any sequence of DNA and a histone octamer.
Extrinsic mechanisms: The first nucleosome in a region is preferentially assembled at a particular site due to action of other protein(s). A preferential starting point for nucleosome positioning can result either from the exclusion of a nucleosome from a particular region (due to competition with another protein binding that region), or by specific deposition of a nucleosome at a particular site. The excluded region of the positioned nucleosome provides a boundary that restricts the positions available to the adjacent nucleosome. A series of nucleosomes can then be assembled sequentially, with a defined repeat length.

We know that the deposition of histone octamers on DNA is not random with regard to sequence. The pattern is intrinsic in cases in which it is determined by structural features in DNA. It is extrinsic in other cases, resulting from the interactions of other proteins with the DNA and/or histones.

Certain structural features of DNA affect placement of histone octamers. DNA has intrinsic tendencies to bend in one direction rather than another. For example, AT dinucleotides bend easily, and thus A-T–rich sequences are easier to wrap tightly in a nucleosome. A-T–rich regions locate so that the minor groove faces in toward the octamer, whereas G-C–rich regions are arranged so that the minor groove points outward. Long runs of dA-dT (>8 bp), in contrast, stiffen the DNA and avoid positioning in the central, tight, superhelical turn of the core. It is not yet possible to sum all of the relevant structural effects and thus entirely predict the location of a particular DNA sequence with regard to the nucleosome, although recently researchers have developed some predictive models that appear to match at least some in vivo positioning data. Sequences that cause DNA to take up more extreme structures have effects such as the exclusion of nucleosomes, and thus cause boundary effects or nucleosome-free regions.

Positioning of nucleosomes near boundaries is common. If there is some variability in the construction of nucleosomes—for example, if the length of the linker can vary by, say, 10 bp—the specificity of positioning would decline proceeding away from the first, defined nucleosome at the boundary. In this case, we might expect the positioning to be maintained rigorously only relatively near the boundary.

The location of DNA on nucleosomes can be described in two ways. FIGURE 8.41 shows that translational positioning describes the position of DNA with regard to the boundaries of the nucleosome. In particular, it determines which sequences are found in the linker regions. Shifting the DNA by 10 bp brings the next turn into a linker region. Thus, translational positioning determines which regions are more accessible (at least as judged by sensitivity to MNase).

FIGURE 8.41 Translational positioning describes the linear position of DNA relative to the histone octamer. Displacement of the DNA by 10 bp changes the sequences that are in the more exposed linker regions, but does not necessarily alter which face of DNA is protected by the histone surface and which is exposed to the exterior.

DNA lies on the outside of the histone octamer. As a result, one face of any particular sequence is obscured by the histones, whereas the other face is exposed on the surface of the nucleosome. Depending upon its positioning with regard to the nucleosome, a site in DNA that must be recognized by a regulatory protein could be inaccessible or available. The exact position of the histone octamer with respect to DNA sequence can therefore be important. FIGURE 8.42 shows the effect of rotational positioning of the double helix with regard to the octamer surface. If the DNA is moved by a partial number of turns (imagine the DNA as rotating relative to the protein surface), there is a change in the exposure of sequence to the outside.

FIGURE 8.42 Rotational positioning describes the exposure of DNA on the surface of the nucleosome. Any movement that differs from the helical repeat (~10.2 bp/turn) displaces DNA with reference to the histone surface. Nucleotides on the inside are more protected against nucleases than nucleotides on the outside.

Both translational and rotational positioning can be important in controlling access to DNA. The best characterized cases of positioning involve the specific placement of nucleosomes at promoters. Translational positioning and/or the exclusion of nucleosomes from a particular sequence might be necessary to allow a transcription complex to form. Some regulatory factors can bind to DNA only if a nucleosome is excluded to make the DNA freely accessible, and this creates a boundary for translational positioning. In other cases, regulatory factors can bind to DNA on the surface of the nucleosome, but rotational positioning is important to ensure that the face of DNA with the appropriate contact points is exposed.

We discuss the connection between nucleosomal organization and transcription in the chapter titled Eukaryotic Transcription Regulation, but note for now that promoters (and some other structures) often have short regions that exclude nucleosomes. These regions typically form a boundary next to which nucleosome positions are restricted. A survey of an extensive region in the Saccharomyces cerevisiae genome (mapping 2,278 nucleosomes over 482 kb of DNA) showed that in fact 60% of the nucleosomes have specific positions as the result of boundary effects, most often from promoters. Nucleosome positioning is a complex output of intrinsic and extrinsic positioning mechanisms. Thus, it has been difficult to predict nucleosome positioning based on sequence alone, though there have been some successes. Large-scale sequencing studies of isolated nucleosomal DNA have revealed intriguing sequence patterns found in positioned nucleosomes in vivo, and it is estimated that 50% or more of in vivo nucleosome positioning is the result of intrinsic sequence determinants encoded in the genomic DNA. It is also important to note that even when a dominant nucleosome position is detected experimentally, it is not likely to be completely invariant (i.e., the nucleosome is not in that exact position in every cell in a sample); instead, it represents the most common location for a nucleosome in that region out of larger set of related positions.

8.10 Nucleosomes Are Displaced and Reassembled During Transcription

Heavily transcribed chromatin adopts structures that are visibly too extended to still be contained in nucleosomes. In the intensively transcribed genes encoding rRNA shown in FIGURE 8.43, the extreme packing of RNA polymerases makes it difficult to see the DNA. Researchers cannot directly measure the lengths of the rRNA transcripts because the RNA is compacted by proteins, but we know (from the sequence of the rRNA) how long the transcript must be. The length of the transcribed DNA segment, which is measured by the length of the axis of the “Christmas tree” shape shown, is about 85% of the length of the pre-rRNA. This means that the DNA is almost completely extended.

FIGURE 8.43 Individual rDNA transcription units alternate with nontranscribed DNA segments.

Reproduced from: Miller, O. L., and BeattyB. R. 1969. Science 164:955–957. Photo courtesy of Oscar Miller.

On the other hand, Researchers can extract transcriptionally active complexes of SV40 minichromosomes from infected cells. They contain the usual complement of histones and display a beaded structure. Chains of RNA can extend from the minichromosome, as shown in FIGURE 8.44. This argues that transcription can proceed while the SV40 DNA is organized into nucleosomes. Of course, the SV40 minichromosome is transcribed less intensively than the rRNA genes.

FIGURE 8.44 An SV40 minichromosome is transcribed while maintaining a nucleosomal structure.

Reprinted from: Gariglio, P., et al. 1979. “The template of the isolated native.” J Mol Bio 131:75–105, with permission from Elsevier (http://www.sciencedirect.com/science/journal/00222836). Photo courtesy of Pierre Chambon, College of France.

Transcription involves the unwinding of DNA, thus it seems obvious that some “elbow room” must be needed for the process. In thinking about transcription, we must keep in mind the relative sizes of RNA polymerase and the nucleosome. Eukaryotic RNA polymerases are large multisubunit proteins, typically greater than 500 kilodaltons (kD). Compare this with the approximately 260 kD of the nucleosome. FIGURE 8.45 illustrates the relative sizes of RNA polymerase and the nucleosome. Consider the two turns that DNA makes around the nucleosome. Would RNA polymerase have sufficient access to DNA if the nucleic acid were confined to this path? During transcription, as RNA polymerase moves along the template, it binds tightly to a region of about 50 bp, including a locally unwound segment of about 12 bp. The need to unwind DNA makes it seem unlikely that the segment engaged by RNA polymerase could remain on the surface of the histone octamer.

FIGURE 8.45 RNA polymerase is nearly twice the size of the nucleosome and might encounter difficulties in following the DNA around the histone octamer.

Top photo courtesy of E. N. Moudrianakis, the Johns Hopkins University. Bottom photo courtesy of Roger Kornberg, Stanford University School of Medicine.

It therefore seems inevitable that transcription must involve a structural change. Thus, the first question to ask about the structure of active genes is whether DNA being transcribed remains organized in nucleosomes. Experiments to test whether an RNA polymerase can transcribe directly through a nucleosome suggest that the histone octamer is displaced by the act of transcription. FIGURE 8.46 shows what happens when the phage T7 RNA polymerase transcribes a short piece of DNA containing a single octamer core in vitro. The core remains associated with the DNA after the polymerase passes, but it is found in a different location. The core is most likely to rebind to the same DNA molecule from which it was displaced. Crosslinking the histones within the octamer does not create an obstacle to transcription, suggesting that (at least in vitro) transcription does not require dissociation of the octamer into its component histones.

FIGURE 8.46 An experiment to test the effect of transcription on nucleosomes shows that the histone octamer is displaced from DNA and rebinds at a new position.

Thus a small RNA polymerase can displace a single nucleosome, which reforms behind it, during transcription. Of course, the situation is more complex in a eukaryotic nucleus. Eukaryotic RNA polymerases are much larger, and the impediment to progress is a string of connected nucleosomes (which can also be folded into higher-order structures). Overcoming these obstacles requires additional factors that act on chromatin (discussed in the chapter Eukaryotic Transcription and in detail in the chapter Eukaryotic Transcription Regulation).

The organization of nucleosomes can be dramatically changed by transcription. This is easiest to observe in inducible genes that have distinct on and off states under different conditions. In many cases, before activation a gene might display a single dominant pattern of nucleosomes that are organized from the promoter and throughout the coding region. When the gene is activated, the nucleosomes become highly mobilized and adopt a number of alternative positions. One or a few nucleosomes might be displaced from the promoter region, but overall nucleosomes typically remain present at a similar density. (However they are no longer organized in phase.) The action of ATP-dependent chromatin remodelers and histone modifiers are typically required to alter the nucleosomal positioning (ATP-dependent chromatin remodelers use the energy of ATP hydrolysis to move or displace nucleosomes; this is discussed in the chapter titled Eukaryotic Transcription Regulation). When repression is reestablished, positioning reappears.

The unifying model is to suppose that RNA polymerase, with the assistance of chromatin remodelers, displaces histone octamers (either as a whole, or as dimers and tetramers) as transcription progresses. If the DNA behind the polymerase is available, the nucleosome is reassembled there. If the DNA is not available—for example, because another polymerase continues immediately behind the first—the octamer might be permanently displaced, and the DNA might persist in an extended form.

Other factors that are critical during transcription elongation, when nucleosomes are being rapidly displaced and reassembled, have been identified. The first of these to be characterized is a heterodimeric factor called FACT (facilitates chromatin transcription), which behaves like a transcription elongation factor. FACT is not part of RNA polymerase; however, it associates with it specifically during the elongation phase of transcription. FACT consists of two subunits that are well conserved in all eukaryotes, and it is associated with the chromatin of active genes.

When FACT is added to isolated nucleosomes, it causes them to lose H2A-H2B dimers. During transcription in vitro, it converts nucleosomes to “hexasomes” that have lost H2A-H2B dimers. This suggests that FACT is part of a mechanism for displacing octamers during transcription. FACT may also be involved in the reassembly of nucleosomes after transcription, because it assists formation of nucleosomes from core histones, thus acting like a histone chaperone. There is evidence in vivo that H2A-H2B dimers are displaced more readily during transcription than H3-H4 tetramers, suggesting that tetramers and dimers can be reassembled sequentially after transcription as they are after passage of a replication fork (see the section Replication of Chromatin Requires Assembly of Nucleosomes earlier in this chapter).

This suggests a model like that shown in FIGURE 8.47, in which FACT (or a similar factor) detaches H2A-H2B from a nucleosome in front of RNA polymerase and then helps to add it to a nucleosome that is reassembling behind the enzyme. Other factors are likely to be required to complete the process. FACT’s role might be more complex than this, because FACT has also been implicated in transcription initiation and replication elongation. Another intriguing model that has been proposed is that FACT stabilizes a “reorganized” nucleosome, in which the dimers and tetramer remain locally tethered via FACT but are not stably organized into a canonical nucleosome. The model presumes the H2A-H2B dimers are less stable in this reorganized state, and thus more easily displaced. In this state, the nucleosomal DNA is highly accessible, and the reorganized nucleosome can either revert to the stable canonical organization or be displaced as needed for transcription.

FIGURE 8.47 Histone octamers are disassembled ahead of transcription to remove nucleosomes. They re-form following transcription. Release of H2A-H2B dimers probably initiates the disassembly process.

Several other factors have been identified that play key roles in either nucleosome displacement or reassembly during transcription. These include the Spt6 protein, a factor involved in “resetting” chromatin structure after transcription. Spt6, like FACT, colocalizes with actively transcribed regions and can act as a histone chaperone to promote nucleosome assembly. Although CAF-1 is known to be involved only in replication-dependent histone deposition, one of CAF-1′s partners in replication might in fact play a role in transcription, as well. The CAF-1–associated protein Rtt106 is an H3-H4 chaperone that has recently been shown to play a role in H3 deposition during transcription.

8.11 DNase Sensitivity Detects Changes in Chromatin Structure

Numerous changes occur to chromatin in active or potentially active regions. These include distinctive structural changes that occur at specific sites associated with initiation of transcription or with certain structural features in DNA. These changes were first detected by the effects of digestion with very low concentrations of the enzyme DNase I.

When chromatin is digested with DNase I, the first effect is the introduction of breaks in the duplex at specific, hypersensitive sites. Susceptibility to DNase I reflects the availability of DNA in chromatin; thus, these sites represent chromatin regions in which the DNA is particularly exposed because it is not organized in the usual nucleosomal structure. A typical hypersensitive site is 100 times more sensitive to enzyme attack than bulk chromatin. These sites are also hypersensitive to other nucleases and to chemical agents.

Hypersensitive sites are created by the local structure of chromatin, which can be tissue specific. Researchers can determine their locations by the technique of indirect end labeling that we introduced earlier in the context of nucleosome positioning. This application of the technique is recapitulated in FIGURE 8.48. In this case, cleavage at the hypersensitive site by DNase I is used to generate one end of the fragment. Its distance is measured from the other end, which is generated by cleavage with a restriction enzyme.

FIGURE 8.48 Indirect end labeling identifies the distance of a DNase hypersensitive site from a restriction cleavage site. The existence of a particular cutting site for DNase I generates a discrete fragment, whose size indicates the distance of the DNase I hypersensitive site from the restriction site.

Many hypersensitive sites are related to gene expression. Every active gene has a hypersensitive site, or sometimes more than one, in the region of the promoter. Most hypersensitive sites are found only in chromatin of cells in which the associated gene is either being expressed or is poised for expression; they do not occur when the gene is inactive. The 5′ hypersensitive site(s) appear before transcription begins and occur in DNA sequences that are required for gene expression.

What is the structure of a hypersensitive site? Its preferential accessibility to nucleases indicates that it is not protected by histone octamers, but this does not necessarily imply that it is free of protein. A region of free DNA might be vulnerable to damage, and would be unable to exclude nucleosomes. In fact, hypersensitive sites typically result from the binding of specific regulatory proteins that exclude nucleosomes. It is very common to find pairs of hypersensitive sites that flank a nuclease-resistant core; the binding of nucleosome-excluding proteins is probably the basis for the existence of the protected region within the hypersensitive sites.

The proteins that generate hypersensitive sites are likely to be regulatory factors of various types, because hypersensitive sites are found associated with promoters and other elements that regulate transcription, origins of replication, centromeres, and sites with other structural significance. In some cases, they are associated with more extensive organization of chromatin structure. A hypersensitive site can provide a boundary for a series of positioned nucleosomes. Hypersensitive sites associated with transcription may be generated by transcription factors when they bind to the promoter as part of the process that makes it accessible to RNA polymerase.

In addition to detecting hypersensitive sites, researchers also can use DNase I digestion to assess the relative accessibility of a genomic region. A region of the genome that contains an active gene can have an altered overall structure, often typified by a general increase in overall DNase sensitivity, in addition to specific hypersensitive sites. The change in structure precedes, and is different from, the disruption of nucleosome structure that might be caused by the actual passage of RNA polymerase. DNase I sensitivity defines a chromosomal domain, which is a region of altered structure including at least one active transcription unit, and sometimes extending farther. (Note that use of the term domain does not imply any necessary connection with the structural domains identified by the loops of chromatin or chromosomes.)

When chromatin is extensively digested with DNase I, it is eventually degraded into very small fragments of DNA. The fate of individual genes can be followed by quantitating the amount of DNA that survives to react with a specific probe. The protocol is outlined in FIGURE 8.49. The principle is that the loss of a particular band indicates that the corresponding region of DNA has been degraded by the enzyme.

FIGURE 8.49 Sensitivity to DNase I can be measured by determining the rate of disappearance of the material hybridizing with a particular probe.

Studies using these methods reveal that the bulk of chromatin is relatively resistant to DNase I and contains nonexpressed genes (as well as other sequences). A gene becomes relatively susceptible to nuclease digestion specifically in the tissue(s) in which it is expressed or is poised to be expressed, and remains nuclease resistant in lineages in which the gene is silent.

What is the extent of a preferentially sensitive region? Researchers can determine this by using a series of probes representing the flanking regions and the transcription unit itself. The sensitive region always extends over the entire transcribed region; an additional region of several kb on either side might show an intermediate level of sensitivity (probably as the result of spreading effects).

The critical concept implicit in the description of the domain is that a region of high sensitivity to DNase I extends over a considerable distance. Often we think of regulation as residing in events that occur at a discrete site in DNA—for example, in the ability to initiate transcription at the promoter. Even if this is true, such regulation must determine, or must be accompanied by, a more wide-ranging change in structure.

8.12 An LCR Can Control a Domain

Every gene is controlled by its proximal promoter, and most genes also respond to enhancers (containing similar regulatory elements located farther away; see the chapter titled Eukaryotic Transcription). These local controls are not sufficient for all genes, though. In some cases, a gene lies within a domain of several genes, all of which are influenced by specialized regulatory elements that act on the whole domain. The existence of these elements was identified by the inability of a region of DNA including a gene and all its known regulatory elements to be properly expressed when introduced into an animal as a transgene.

The best-characterized example of a regulated gene cluster is provided by the mammalian β-globin genes. Recall from the chapter titled Genome Sequences and Evolution that the α- and β-globin genes in mammals each exist as clusters of related genes that are expressed at different times and in different tissues during embryonic and adult development. These genes are associated with a large number of regulatory elements, which have been analyzed in detail. In the case of the adult human β-globin gene, regulatory sequences are located both 5′ and 3′ to the gene. The regulatory sequences include positive and negative elements in the promoter region as well as additional positive elements within and downstream of the gene.

All of these control regions are not, however, sufficient for proper expression of the human β-globin gene in a transgenic mouse within an order of magnitude of wild-type levels. Some further regulatory sequence is required. Regions that provide the additional regulatory function are identified by DNase I hypersensitive sites that are found at the ends of the β-globin cluster. The map in FIGURE 8.50 shows that the 20 kb upstream of the ε gene contains a group of 5 hypersensitive sites, and that there is a single site 30 kb downstream of the β gene.

FIGURE 8.50 The β-globin locus is marked by hypersensitive sites at either end. The group of sites at the 5′ side constitutes the LCR and is essential for the function of all genes in the cluster.

The 5′ regulatory sites are the primary regulators, and the region containing the cluster of hypersensitive sites is called the locus control region (LCR). The role of the LCR is complex; in some ways it behaves as a “super enhancer” that poises the entire locus for transcription. The precise function of the 3′ hypersensitive site in the mammalian locus is not clear, but it is known to physically interact with the LCR. A 3′ hypersensitive site in the chicken β-globin locus acts as an insulator, as does a fifth 5′ site upstream of the mammalian LCR. The LCR is absolutely required for expression of each of the globin genes in the locus. Each gene is then further regulated by its own specific controls. Some of these controls are autonomous: Expression of the ε and γ genes appears intrinsic to those loci in conjunction with the LCR. Other controls appear to rely upon position in the cluster, which provides a suggestion that gene order in a cluster is important for regulation.

The entire region containing the globin genes, and extending well beyond them, constitutes a chromosomal domain. It shows increased sensitivity to digestion by DNase I. Deletion of the 5′ LCR restores normal resistance to DNase over the entire region. In addition to increases in the general accessibility of the locus, the LCR is also apparently required to directly activate the individual promoters. Researchers have not yet fully defined the exact nature of the sequential interactions between the LCR and the individual promoters, but it has recently become clear that the LCR contacts individual promoters directly, forming loops when these promoters are active. The domain controlled by the LCR also shows distinctive patterns of histone modifications (see the chapter titled Eukaryotic Transcription Regulation) that are dependent on LCR function.

This model appears to apply to other gene clusters, as well. The α-globin locus has a similar organization of genes that are expressed at different times, with a group of hypersensitive sites at one end of the cluster and increased sensitivity to DNase I throughout the region. So far, though, only a small number of other cases are known in which an LCR controls a group of genes.

One of these cases involves an LCR that controls genes on more than one chromosome. The T_H2 LCR coordinately regulates the T helper type 2 cytokine locus, a group of genes encoding a number of interleukins (important signaling molecules in the immune system). These genes are spread out over 120 kb on chromosome 11, and the T_H2 LCR controls them by interacting with their promoters. It also interacts with the promoter of the IFNγ gene on chromosome 10. The two types of interaction are alternatives that comprise two different cell fates; that is, in one group of cells the LCR causes expression of the genes on chromosome 11, whereas in the other group it causes the gene on chromosome 10 to be expressed.

Looping interactions are important for chromosome structure, and function was introduced in the chapter titled Chromosomes. New methods have been developed to begin to dissect the physical interactions between chromosomal loci in vivo, leading to fresh understanding of how these interactions result in regulatory functions. Direct interactions between the β-globin and T_H2 LCRs and their target loci have been mapped using a method known as chromosome conformation capture (3C). There are now many variations of this procedure; the basic method is outlined in the top panel of FIGURE 8.51. Interacting regions of chromatin in vivo are captured using formaldehyde treatment to crosslink to fix the DNA and proteins that are in close contact. Next, the chromatin is digested with a restriction enzyme and ligated under dilute conditions to favor intra-molecular ligation. This results in preferential ligation of DNA fragments that are held in close proximity as a result of crosslinking. Finally, the proteins are removed by reversing the crosslinking and the new ligated junctions are detected by PCR or sequencing.

FIGURE 8.51 3C is one method to detect physical interactions between regions of chromatin in vivo. Looping interactions controlled by the β-globin and T_H2 LCRs have been mapped by 3C and some of the known contacts are shown.

Adapted from: Miele, A., and Dekker, J. 2008. Mol Biosyst 4:1046–1057.

As shown in the lower part of the Figure 8.51, 3C and similar methods have allowed researchers to begin to unravel the complex and dynamic interactions that occur at loci regulated by LCRs. The β-globin LCR sequentially interacts with each globin gene at the developmental stage in which that gene is active; the figure shows the interactions that occur between the LCR, 3′ HS, and the γ-globin genes in the fetal stage. Interestingly, the T_H2 LCR appears to interact with all three of its target genes (Il3, −4, and −5) simultaneously. These interactions occur in all T-cells regardless of whether these genes are expressed, but the precise organization of loops alters upon activation of the interleukin genes. This reorganization, which depends on the protein SATB1 (special AT-rich binding protein), suggests that the T_H2 LCR brings all the genes together in a poised state in T cells, awaiting the trigger of specific transcription factors to activate the genes rapidly when needed.

8.13 Insulators Define Transcriptionally Independent Domains

KEY CONCEPTS

Mammalian chromosomes are organized as strings of topologically associated domains (TADs) that average about 1 megabase (Mb) in size.
TADs or TAD-like structures have been found in most eukaryotes.
Loci within a TAD interact frequently with each other, but less frequently with loci in an adjacent TAD.
TAD organization is fairly stable between cells, but interactions within TADs are highly dynamic.
Boundary regions between TADs contain insulator elements that are able to block passage of any activating or inactivating effects from enhancers, silencers, and other control elements.
Insulators can provide barriers against the spread of heterochromatin.
Insulators are specialized chromatin structures that typically contain hypersensitive sites.
Different insulators are bound by different factors and may use alternative mechanisms for enhancer blocking and/or heterochromatin barrier formation.

Different regions of the chromosome have different functions that are typically marked by specific chromatin structures or modification states. We have discussed LCRs that control gene transcription from very large distances (see also the chapter Eukaryotic Transcription), and that highly compacted heterochromatin (introduced in the chapter Chromosomes) can also spread over large distances (see the chapter Epigenetics I). The existence of these long-range interactions suggests that chromosomes must also contain functional elements that serve to partition chromosomes into domains that can be regulated independently of one another. Over the past several years, the 3C method (see Figure 8.51) has been coupled with massively parallel sequencing, resulting in comprehensive interaction maps that probe the three-dimensional architecture of whole genomes. The results indicate that mammalian and Drosophila genomes are organized as a string of TADs that are separated from one another by distinct borders or boundaries (FIGURE 8.52). TADs are characterized by frequent interactions between loci within a domain (e.g., the β-globin genes), but loci within different TADs interact rarely with one another. Thus, TADs might allow for the compartmentalization of chromosomal regions with distinct functions. TADs vary in size, but in mammalian cells they average about 1 Mb. Interestingly, more than half of all mammalian TADs appear conserved between different cell types and even between mouse and human. Other TADs appear to be more dynamic during development. TAD organization is a feature of interphase chromatin, as mitotic chromosomes appear to lack such organization. More recently, similar structures have also been identified in budding and fission yeasts, suggesting that they might be a conserved feature of eukaryotic genomes.

FIGURE 8.52 Organization of a mammalian genome into strings of TADs. The TADs are defined as regions of the genome that show a high frequency of interactions. TADs are separated by border regions that often contain insulator elements.

The border or boundary elements that separate TADs contain a class of elements called insulators that prevent inter-TAD interactions and block the passage of activating or inactivating effects. Insulators were originally defined as having either or both of two key properties:

When an insulator is located between an enhancer and a promoter, it prevents the enhancer from activating the promoter. FIGURE 8.53 shows this enhancer-blocking effect. This activity might explain how the action of an enhancer is limited to a particular promoter despite the ability of enhancers to activate promoters from long distances away (and the ability of enhancers to indiscriminately activate any promoter in the vicinity).
When an insulator is located between an active gene and heterochromatin, it provides a barrier that protects the gene against the inactivating effect that spreads from the heterochromatin. FIGURE 8.54 illustrates this barrier effect.

Some insulators possess both of these properties, but others have only one, or the blocking and barrier functions can be separated. Likewise, only some insulators function as borders between TADs, whereas others do not. Although both actions are likely to be mediated by changing chromatin structure, they can involve different effects. In either case, however, the insulator defines a limit for long-range effects. By restricting enhancers so they can act only on specific promoters, and preventing the inadvertent spreading of heterochromatin into active regions, insulators function as elements for increasing the precision of gene regulation.

FIGURE 8.53 An enhancer activates a promoter in its vicinity but can be blocked from doing so by an insulator located between them.

FIGURE 8.54 Heterochromatin may spread from a center and then block any promoters that it covers. An insulator might be a barrier to propagation of heterochromatin that allows the promoter to remain active.

Insulators were first discovered in the analysis of a region of the Drosophila melanogaster genome shown in FIGURE 8.55. Two genes for hsp (heat-shock protein) 70 lie within an 18-kb region that constitutes band 87A7. Researchers had noted that when subjected to heat shock, a puff forms at 87A7 in polytene chromosomes, and there is a distinct boundary between the decondensed and condensed regions of the chromosomes. Special structures, called scs and scs′ (specialized chromatin structures), are found at the ends of the band. Each element consists of a region that is highly resistant to degradation by DNase I, flanked on either side by hypersensitive sites that are spaced at about 100 bp. The cleavage pattern at these sites is altered when the genes are turned on by heat shock.

FIGURE 8.55 The 87A and 87C loci, containing heat-shock genes, expand upon heat shock in Drosophila polytene chromosomes. Specialized chromatin structures that include hypersensitive sites mark the ends of the 87A7 domain and insulate genes between them from the effects of surrounding sequences.

Photo courtesy of Victor G. Corces, Emory University.

The scs elements insulate the hsp70 genes from the effects of surrounding regions (and presumably also protect the surrounding regions from the effects of heat-shock activation at the hsp70 loci). In the first assay for insulator function, scs elements were tested for their ability to protect a reporter gene from “position effects.” In this experiment, scs elements were placed in constructs flanking the white gene, the gene responsible for producing red pigment in the Drosophila eye, and these constructs were randomly integrated into the fly genome. If the white gene is integrated without scs elements, its expression is subject to position effects; that is, the chromatin context in which the gene is inserted strongly influences whether the gene is transcribed. This can be detected as a variegated color phenotype in the fly eye, as shown in FIGURE 8.56. However, when scs elements are placed on either side of the white gene, the gene can function anywhere it is placed in the genome—even in sites where it would normally be repressed by context (such as in heterochromatic regions), resulting in uniformly red eyes.

FIGURE 8.56 Position effects are often observed when an inversion or other chromosome rearrangement repositions a gene normally in euchromatin to a new location in or near heterochromatin. In this example, an inversion in the X chromosome of Drosophila melanogaster repositions the wild-type allele of the white gene near heterochromatin. Differences in expression due to position effects on the w⁺ allele are observed as mottled red and white eyes.

The scs and scs′ elements, like many other insulators, do not themselves play positive or negative roles in controlling gene expression, but restrict effects from passing from one region to the next. Unexpectedly, the scs elements themselves are not responsible for controlling the precise boundary between the condensed and decondensed regions at the heat shock puff, but instead serve to prevent regulatory crosstalk between the hsp70 genes and the many other genes in the region.

The scs and scs′ elements have different structures, and each appears to have a different basis for its insulator activity. The key sequence in the scs element is a stretch of 24 bp that binds the product of the zw5 (zeste white 5) gene. The insulator property of scs′ resides in a series of CGATA repeats. The repeats bind a pair of related proteins (encoded by the same gene) called BEAF-32. BEAF-32 is localized to about 50% of the interbands on polytene chromosomes, suggesting that there are many BEAF-32–dependent insulators in the genome (though BEAF-32 may bind noninsulators, as well).

Another well-characterized insulator in Drosophila is found in the transposon gypsy. Some experiments that initially defined the behavior of this insulator were based on a series of gypsy insertions into the yellow (y) locus. Different insertions cause loss of y gene function in some tissues, but not in others. The reason is that the y locus is regulated by four enhancers, as shown in FIGURE 8.57. Wherever gypsy is inserted, it blocks expression of all enhancers that it separates from the promoter, but not those that lie on the other side. The sequence responsible for this effect is an insulator that lies at one end of the transposon. The insulator works irrespective of its orientation of insertion.

FIGURE 8.57 The insulator of the gypsy transposon blocks the action of an enhancer when it is placed between the enhancer and the promoter.

The function of the gypsy insulator depends on several proteins, including Su(Hw) (Suppressor of Hairy wing), CP190, mod(mdg4), and dTopors. Mutations in the su (Hw) gene completely abolish insulation; su (Hw) encodes a protein that binds 12 26-bp reiterated sites in the insulator and is necessary for its action. Su(Hw) has a zinc finger DNA-motif; mapping to polytene chromosomes shows that Su(Hw) is bound to hundreds of sites that include both gypsy insertions and non-gypsy sites. Manipulations show that the strength of the insulator is determined by the number of copies of the binding sequence. CP190 is a centrosomal protein that assists Su(Hw) in binding site recognition.

mod(mdg4) and dTopors have a specific role in the creation of “insulator bodies,” which appear to be clusters of Su(Hw)-bound insulators that can be observed in normal diploid nuclei. Despite the presence of >500 Su(Hw) binding sites in the Drosophila genome, visualization of Su(Hw) or mod(mdg4) shows that they are colocalized at about 25 discrete sites around the nuclear periphery. This suggests the model of FIGURE 8.58, in which Su(Hw) proteins bound at different sites on DNA are brought together by binding to mod(mdg4). The Su(Hw)/mod(mdg4) complex is localized at the nuclear periphery. The DNA bound to it is organized into loops. An average complex might have 20 such loops. Enhancer–promoter actions can occur only within a loop, and cannot propagate between them. This model is supported by “insulator bypass” experiments, in which placing a pair of insulators between an enhancer and promoter actually eliminates insulator activity—somehow the two insulators cancel out each other. This could be explained by the formation of a minidomain between the duplicated insulator (perhaps too small to create an anchored loop), which would essentially result in what should have been two adjacent loops fused into one. Not all insulators can be bypassed in this way, however; this and other evidence suggests that there are multiple mechanisms for insulator function.

FIGURE 8.58 Su(Hw)/mod(mdg4) complexes are found in clusters at the nuclear periphery. They can organize DNA into loops that limit enhancer–promoter interactions.

The complexity of insulators and their roles is indicated by the behavior of another Drosophila insulator: the Fab-7 element found in the bithorax locus (BX-C). This locus contains a series of cis-acting regulatory elements that control the activities of three homeotic genes (Ubx, abd-A, and Abd-B), which are differentially expressed along the anterior–posterior axis of the Drosophila embryo. The locus also contains at least three insulators that are not interchangeable; Fab-7 is the best studied of these. FIGURE 8.59 shows the relevant part of the locus. The regulatory elements iab-6 and iab-7 control expression of the adjacent gene Abd-B in successive regions of the embryo (segments A6 and A7). A deletion of Fab-7 causes A6 to develop like A7, resulting in two “A7-like” segments (this is known as a homeotic transformation). This is a dominant effect, which suggests that iab-7 has taken over control from iab-6. We can interpret this in molecular terms by supposing that Fab-7 provides a boundary that prevents iab-7 from acting when iab-6 is usually active. In fact, in the absence of Fab-7, it appears that iab-6 and iab-7 fuse into a single regulatory domain, which shows different behavior depending on the position along the AP axis. The insulator activity of Fab-7 is also developmentally regulated, with a protein called Elba (Early boundary activity) responsible for Fab-7’s blocking function early in development, but not later in development or in the adult. Fab-7 is also associated with the Drosophila homolog of the CTCF protein, a mammalian insulator-binding protein that shows regulated binding to its targets (see the chapter titled Epigenetics II). In mammalian cells, CTCF is a key component of insulators that form borders between many TADs. Finally, both Fab-7 and a nearby insulator (Fab-8) are known to lie near “anti-insulator elements” (also called promoter-targeting sequences or PTS elements), which may allow an enhancer to overcome the blocking effects of an insulator.

FIGURE 8.59 Fab-7 is a boundary element that is necessary for the independence of regulatory elements iab-6 and iab-7.

The diversity of insulator behaviors and of the factors responsible for insulator function makes it impossible to propose a single model to explain the behavior of all insulators. Instead, it is clear that the term “insulator” refers to a variety of elements that use a number of distinct mechanisms to achieve similar (but not identical) functions. Notably, the mechanisms used to block enhancers can be very different from those used to block the spread of heterochromatin. There is also a diversity of proteins that bind to insulator elements, and the general term “architectural proteins” has been used to describe this group of factors. Furthermore, the density of architectural protein binding sites appears to correlate well with different types of insulator activities, with high-density regions corresponding to insulators that function as borders between TAD domains, and lower-density sites regulating intradomain interactions.

Summary

All eukaryotic chromatin consists of nucleosomes. A nucleosome contains a characteristic length of DNA, usually about 200 bp, which is wrapped around an octamer containing two copies each of histones H2A, H2B, H3, and H4. A single H1 (or other linker histone) might associate with a nucleosome. Virtually all genomic DNA is organized into nucleosomes. Treatment with micrococcal nuclease shows that the DNA packaged into each nucleosome can be divided operationally into two regions. The linker region is digested rapidly by the nuclease; the core region of 145–147 bp is resistant to digestion. Histones H3 and H4 are the most highly conserved, and an H3₂-H4₂ tetramer accounts for the diameter of the particle. Histones H2A and H2B are organized as two H2A-H2B dimers. Octamers are assembled by the successive addition of two H2A-H2B dimers to the H3₂-H4₂ tetramer. A large number of histone variants exist that can also be incorporated into nucleosomes; different variants perform different functions in chromatin and some are cell-type specific.
The path of DNA around the histone octamer creates −1.67 supercoils. The DNA “enters” and “exits” the nucleosome on the same side, and the entry or exit angle could be altered by histone H1. Removal of the core histones releases −1.0 supercoils. We can largely explain this difference by a change in the helical pitch of DNA, from an average of 10.2 bp/turn in nucleosomal form to 10.5 bp/turn when free in solution. There is variation in the structure of DNA from a periodicity of 10.0 bp/turn at the nucleosome ends to 10.7 bp/turn in the center. There are kinks in the path of DNA on the nucleosome.
Nucleosomes are organized into long fibers with a 10-nm diameter that has a linear packing ratio of 6. Linker histone H1, histone tails, and increased ionic strength promote intrafiber and interfiber interactions that form more condensed secondary structures, such as the 30-nm fiber or self-associated networks of 10-nm filaments. The 30-nm fiber probably consists of the 10-nm fiber wound into a heterogeneous mixture of one-start solenoids and two-start zigzag helices. The 10-nm fiber is the basic constituent of both euchromatin and heterochromatin; nonhistone proteins facilitate further organization of the fiber into chromatin or chromosome ultrastructure.
There are two pathways for nucleosome assembly. In the replication-coupled pathway, the PCNA processivity subunit of the replisome recruits CAF-1, which is a nucleosome assembly factor or histone “chaperone.” CAF-1 assists the deposition of H3₂-H4₂ tetramers onto the daughter duplexes resulting from replication. The tetramers can be produced either by disruption of existing nucleosomes by the replication fork or as the result of assembly from newly synthesized histones. CAF-1 assembles newly synthesized tetramers, whereas the ASF1 chaperone also assists with deposition of H3₂-H4₂ tetramers that have been displaced by the replication fork. Similar sources provide the H2A-H2B dimers that then assemble with the H3₂-H4₂ tetramer to complete the nucleosome. The H3₂-H4₂ tetramer and the H2A-H2B dimers assemble at random, so the new nucleosomes might include both preexisting and newly synthesized histones. Nucleosome placement is not random throughout the genome, but is controlled by a combination of intrinsic (DNA sequence–dependent) and extrinsic (dependent on trans-factors) mechanisms that result in specific patterns of nucleosome deposition.
RNA polymerase displaces histone octamers during transcription. Nucleosomes reform on DNA after the polymerase has passed, unless transcription is very intensive (such as in rDNA) when they can be displaced completely. The replication-independent pathway for nucleosome assembly is responsible for replacing histone octamers that have been displaced by transcription. It uses the histone variant H3.3 instead of H3. A similar pathway, with another alternative to H3, is used for assembling nucleosomes at centromeric DNA sequences.
Two types of changes in sensitivity to nucleases are associated with gene activity. Chromatin capable of being transcribed has a generally increased sensitivity to DNase I, reflecting a change in structure over an extensive region that can be defined as a domain containing active or potentially active genes. Hypersensitive sites in DNA occur at discrete locations and are identified by greatly increased sensitivity to DNase I. A hypersensitive site consists of a sequence of typically more than 200 bp from which nucleosomes are excluded by the presence of other proteins. A hypersensitive site forms a boundary that can cause adjacent nucleosomes to be restricted in position. Nucleosome positioning might be important in controlling access of regulatory proteins to DNA.
Hypersensitive sites occur at several types of regulators. Those that regulate transcription include promoters, enhancers, and LCRs. Other sites include insulators, origins of replication, and centromeres. A promoter or enhancer typically acts on a single gene, whereas an LCR contains a group of hypersensitive sites and may regulate a domain containing several genes.
LCRs function at a distance and might be required for any and all genes in a domain to be expressed. When a domain has an LCR, its function is essential for all genes in the domain, but LCRs do not seem to be common. LCRs contain enhancer-like hypersensitive site(s) that are needed for the full activity of promoter(s) within the domain and to create a general domain of DNase sensitivity. LCRs also act by creating loops between LCR sequences and the promoters of active genes within the domain.
Eukaryotic genomes are generally organized into discrete regions called TADs. Loci within a TAD interact frequently with each other (likely by looping), but interactions between different TADs are rare. TADs are separated by boundary or border regions that contain hypersensitive sites. These border regions also contain elements called insulators that can block the transmission of activating or inactivating effects in chromatin. An insulator that is located between an enhancer and a promoter prevents the enhancer from activating the promoter. Two insulators define the region between them as a regulatory domain (sometimes equivalent to a TAD); regulatory interactions within the domain are limited to it, and the domain is insulated from outside effects. Most insulators block regulatory effects from passing in either direction, but some are directional. Insulators usually can block both activating effects (enhancer–promoter interactions) and inactivating effects (mediated by spread of heterochromatin), but some are limited to one or the other. Insulators are thought to act via changing higher order chromatin structure, but the details are not certain.

References

8.3 The Nucleosome Is the Subunit of All Chromatin

Reviews

Izzo, A., Kamieniarz, K., and Schneider, R. (2008). The histone H1 family: specific members, specific functions? Biol. Chem. 389, 333–343.
Kornberg, R. D. (1977). Structure of chromatin. Annu. Rev. Biochem. 46, 931–954.
McGhee, J. D., and Felsenfeld, G. (1980). Nucleosome structure. Annu. Rev. Biochem. 49, 1115–1156.

Research

Angelov, D., et al. (2001). Preferential interaction of the core histone tail domains with linker DNA. Proc. Natl. Acad. Sci. USA 98, 6599–6604.
Arents, G., et al. (1991). The nucleosomal core histone octamer at 31 Å resolution: a tripartite protein assembly and a left-handed superhelix. Proc. Natl. Acad. Sci. USA 88, 10148–10152.
Finch, J. T., et al. (1977). Structure of nucleosome core particles of chromatin. Nature 269, 29–36.
Kornberg, R. D. (1974). Chromatin structure: a repeating unit of histones and DNA. Science 184, 868–871.
Luger, K., et al. (1997). Crystal structure of the nucleosome core particle at 28 Å resolution. Nature 389, 251–260.
Richmond, T. J., et al. (1984). Structure of the nucleosome core particle at 7 Å resolution. Nature 311, 532–537.
Shen, X., et al. (1995). Linker histones are not essential and affect chromatin condensation in vitro. Cell 82, 47–56.

8.4 Nucleosomes Are Covalently Modified

Review

Gardner, K. E., et al. (2011). Operating on chromatin, a colorful language where context matters. J. Mo. Biol. 409, 36–46.
Suganuma, T., and Workman, J. L. (2011). Signals and combinatorial functions of histone modifications. Ann. Rev. Biochem. 80, 473–499.

Research

Shogren-Knaak, et al. (2006). Histone H4-K16 acetylation controls chromatin structure and protein interactions. Science 311, 844–847.

8.5 Histone Variants Produce Alternative Nucleosomes

Reviews

Maze, I., et al. (2014). Every amino acid matters: essential contributions of histone variants to mammalian development and disease. Nat. Rev. Genet. 4, 259–271.

8.6 DNA Structure Varies on the Nucleosomal Surface

Reviews

Luger, K., and Richmond, T. J. (1998). DNA binding within the nucleosome core. Curr. Opin. Struct. Biol. 8, 33–40.
Travers, A. A., and Klug, A. (1987). The bending of DNA in nucleosomes and its wider implications. Philos. Trans. R. Soc. Lond. B. Biol. Sci. 317, 537–561.
Wang, J. (1982). The path of DNA in the nucleosome. Cell 29, 724–726.

Research

Richmond, T. J., and Davey, C. A. (2003). The structure of DNA in the nucleosome core. Nature 423, 145–150.

8.7 The Path of Nucleosomes in the Chromatin Fiber

Reviews

Fussner, E., et al. (2011). Living without 30 nm chromatin fibers. Trends Biochem. Sci. 36, 1–6.
Maeshima, K., et al. (2015). Chromatin as dynamic 10-nm fibers. Chromosoma 123, 225–237.
Tremethick, D. J. (2007). Higher-order structures of chromatin: the elusive 30 nm fiber. Cell 128, 651–654.

Research

Dorigo, B., et al. (2004). Nucleosome arrays reveal the two-start organization of the chromatin fiber. Science 306, 1571–1573.
Nishino, Y., et al. (2012). Human mitotic chromosomes consist predominantly of irregularly folded nucleosome fibers without a 30-nm chromatin structure. EMBO J. 31, 1644–1653.
Schalch, T., et al. (2005). X-ray structure of a tetranucleosome and its implications for the chromatin fibre. Nature 436,138–141.
Scheffer, M. P., et al. (2011). Evidence for short-range helical order in the 30-nm chromatin fibers of erythrocyte nuclei. Proc. Natl. Acad. Sci. 108, 16992–16997.

8.8 Replication of Chromatin Requires Assembly of Nucleosomes

Reviews

Corpet, A., and Almouzni, G. (2008). Making copies of chromatin: the challenge of nucleosomal organization and epigenetic information. Trends Cell Biol. 19, 29–41.
Eitoku, M., et al. (2008). Histone chaperones: 30 years from isolation to elucidation of the mechanisms of nucleosome assembly and disassembly. Cell Mol. Life Sci. 65, 414–444.
Osley, M. A. (1991). The regulation of histone synthesis in the cell cycle. Annu. Rev. Biochem. 60, 827–861.
Steiner, F. A., and Henikoff, S. (2015). Diversity in the organization of centromeric chromatin. Curr. Opin. Genet. Dev. 31, 28–35.
Verreault, A. (2000). De novo nucleosome assembly: new pieces in an old puzzle. Genes Dev. 14, 1430–1438.

Research

Ahmad, K., and Henikoff, S. (2001). Centromeres are specialized replication domains in heterochromatin. J. Cell Biol. 153, 101–110.
Ahmad, K., and Henikoff, S. (2002). The histone variant H3.3 marks active chromatin by replication-independent nucleosome assembly. Mol. Cell 9, 1191–1200.
Gruss, C., et al. (1993). Disruption of the nucleosomes at the replication fork. EMBO J. 12, 4533–4545.
Loppin, B., et al. (2005). The histone H3.3 chaperone HIRA is essential for chromatin assembly in the male pronucleus. Nature 437, 1386–1390.
Ray-Gallet, et al. (2002). HIRA is critical for a nucleosome assembly pathway independent of DNA synthesis. Mol. Cell 9, 1091–1100.
Shibahara, K., and Stillman, B. (1999). Replication-dependent marking of DNA by PCNA facilitates CAF-1-coupled inheritance of chromatin. Cell 96, 575–585.
Smith, S., and Stillman, B. (1989). Purification and characterization of CAF-I, a human cell factor required for chromatin assembly during DNA replication in vitro. Cell 58, 15–25.
Smith, S., and Stillman, B. (1991). Stepwise assembly of chromatin during DNA replication in vitro. EMBO J. 10, 971–980.
Tagami, H., et al. (2004). Histone H3.1 and H3.3 complexes mediate nucleosome assembly pathways dependent or independent of DNA synthesis. Cell 116, 51–61.
Yu, L., and Gorovsky, M. A. (1997). Constitutive expression, not a particular primary sequence, is the important feature of the H3 replacement variant hv2 in Tetrahymena thermophila. Mol. Cell Biol. 17, 6303–6310.

8.9 Do Nucleosomes Lie in Specific Positions?

Research

Chung, H. R., and Vingron, M. (2009). Sequence-dependent nucleosome positioning. J. Mol. Biol. 386, 1411–1422.
Field, Y., et al. (2008). Distinct modes of regulation by chromatin encoded through nucleosome positioning signals. PLoS Comput. Biol. 4(11), e1000216.
Peckham, H. E., et al. (2007). Nucleosome positioning signals in genomic DNA. Genome Res. 17, 1170–1177.
Segal, E., et al. (2006). A genomic code for nucleosome positioning. Nature 442, 772–778.
Yuan, G. C., et al. (2005). Genome-scale identification of nucleosome positions in S. cerevisiae. Science 309, 626–630.
Zhang, Z., et al. (2011). A packing mechanism for nucleosome organization reconstituted across a eukaryotic genome. Science 332, 977–980.

8.10 Nucleosomes Are Displaced and Reassembled During Transcription

Reviews

Formosa, T. (2008). FACT and the reorganized nucleosome. Mol. BioSyst. 4, 1085–1093.
Kornberg, R. D., and Lorch, Y. (1992). Chromatin structure and transcription. Annu. Rev. Cell Biol. 8, 563–587.
Kulaeva, O. I., and Studitsky, V. M. (2007). Transcription through chromatin by RNA polymerase II: histone displacement and exchange. Mutat. Res. 618, 116–129.
Thiriet, C., and Hayes, J. J. (2006). Histone dynamics during transcription: exchange of H2A/H2B dimers and H3/H4 tetramers during pol II elongation. Results Probl. Cell Differ. 41, 77–90.
Workman, J. L. (2006). Nucleosome displacement during transcription. Genes Dev 20, 2507–2512.

Research

Belotserkovskaya, R., et al. (2003). FACT facilitates transcription-dependent nucleosome alteration. Science 301, 1090–1093.
Bortvin, A., and Winston, F. (1996). Evidence that Spt6p controls chromatin structure by a direct interaction with histones. Science 272, 1473–1476.
Cavalli, G., and Thoma, F. (1993). Chromatin transitions during activation and repression of galactose-regulated genes in yeast. EMBO J. 12, 4603–4613.
Imbeault, D., et al. (2008). The Rtt106 histone chaperone is functionally linked to transcription elongation and is involved in the regulation of spurious transcription from cryptic promoters in yeast. J. Biol. Chem. 283, 27350–27354.
Saunders, A., et al. (2003). Tracking FACT and the RNA polymerase II elongation complex through chromatin in vivo. Science 301, 1094–1096.
Studitsky, V. M., et al. (1994). A histone octamer can step around a transcribing polymerase without leaving the template. Cell 76, 371–382.

8.11 DNase Sensitivity Detects Changes in Chromatin Structure

Reviews

Gross, D. S., and Garrard, W. T. (1988). Nuclease hypersensitive sites in chromatin. Annu. Rev. Biochem. 57, 159–197.
Krebs, J. E., and Peterson, C. L. (2000). Understanding “active” chromatin: a historical perspective of chromatin remodeling. Crit. Rev. Eukaryot Gene Expr. 10, 1–12.

Research

Groudine, M., and Weintraub, H. (1982). Propagation of globin DNAase I-hypersensitive sites in absence of factors required for induction: a possible mechanism for determination. Cell 30, 131–139.
Stalder, J., et al. (1980). Tissue-specific DNA cleavage in the globin chromatin domain introduced by DNAase I. Cell 20, 451–460.

8.12 An LCR May Control a Domain

Reviews

Bulger, M., and Groudine, M. (1999). Looping versus linking: toward a model for long-distance gene activation. Genes Dev. 13, 2465–2477.
Grosveld, F., et al. (1993). The regulation of human globin gene switching. Philos. Trans. R. Soc. Lond. B. Biol. Sci. 339, 183–191.
Miele, A., and Dekker, J. (2008). Long-range chromosomal interactions and gene regulation. Mol. BioSyst. 4, 1046–1057.

Research

Cai, S., et al. (2006). SATB1 packages densely looped, transcriptionally active chromatin for coordinated expression of cytokine genes. Nat. Genet. 38, 1278–1288.
Gribnau, J., et al. (1998). Chromatin interaction mechanism of transcriptional control in vitro. EMBO. J. 17, 6020–6027.
Spilianakis, C. G., et al. (2005). Inter-chromosomal associations between alternatively expressed loci. Nature 435, 637–645.
van Assendelft, G. B., et al. (1989). The β-globin dominant control region activates homologous and heterologous promoters in a tissue-specific manner. Cell 56, 969–977.

8.13 Insulators Define Transcriptionally Independent Domains

Reviews

Bushey, A. M., et al. (2008). Chromatin insulators: regulatory mechanisms and epigenetic inheritance. Mol. Cell 32, 1–9.
Gaszner, M., and Felsenfeld, G. (2006). Insulators: exploiting transcriptional and epigenetic mechanisms. Nat. Rev. Genet. 7, 703–713.
Gibcus, J. H., and Dekker, J. (2013). The hierarchy of the 3D genome. Molec. Cell 49, 773–782.
Gomez-Diaz, E., and Corces, V. G. (2014). Architectural proteins: regulators of 3D genome organization in cell fate. Trends Cell Biol. 24, 703–711.
Maeda, R. K., and Karch, F. (2007). Making connections: boundaries and insulators in Drosophila. Curr. Opin. Genet. Dev. 17, 394–399.
Valenzuela, L., and Kamakaka, R. T. (2006). Chromatin insulators. Annu. Rev. Genet. 40, 107–138.
West, A. G., et al. (2002). Insulators: many functions, many mechanisms. Genes Dev. 16, 271–288.

Research

Aoki, T., et al. (2008). A stage-specific factor confers Fab-7 boundary activity during early embryogenesis in Drosophila. Mol. Cell Biol. 28, 1047–1060.
Chung, J. H., et al. (1993). A 5′ element of the chicken β-globin domain serves as an insulator in human erythroid cells and protects against position effect in Drosophila. Cell 74, 505–514.
Cuvier, O., et al. (1998). Identification of a class of chromatin boundary elements. Mol. Cell Biol. 18, 7478–7486.
Dixon, J. R., et al. (2012). Topological domains in mammalian genomes identified by analysis of chromatin interactions. Nature 485, 376–380.
Gaszner, M., et al. (1999). The Zw5 protein, a component of the scs chromatin domain boundary, is able to block enhancer–promoter interaction. Genes Dev. 13, 2098–2107.
Gerasimova, T. I., et al. (2000). A chromatin insulator determines the nuclear localization of DNA. Mol. Cell 6, 1025–1035.
Hagstrom, K., et al. (1996). Fab-7 functions as a chromatin domain boundary to ensure proper segment specification by the Drosophila bithorax complex. Genes Dev. 10, 3202–3215.
Harrison, D. A., et al. (1993). A leucine zipper domain of the suppressor of hairy-wing protein mediates its repressive effect on enhancer function. Genes Dev. 7, 1966–1978.
Kellum, R., and Schedl, P. (1991). A position-effect assay for boundaries of higher order chromosomal domains. Cell 64, 941–950.
Kuhn, E. J., et al. (2004). Studies of the role of the Drosophila scs and scs′ insulators in defining boundaries of a chromosome puff. Mol. Cell Biol. 24, 1470–1480.
Mihaly, J., et al. (1997). In situ dissection of the Fab-7 region of the bithorax complex into a chromatin domain boundary and a polycomb-response element. Development 124, 1809–1820.
Pikaart, M. J., et al. (1998). Loss of transcriptional activity of a transgene is accompanied by DNA methylation and histone deacetylation and is prevented by insulators. Genes Dev. 12, 2852–2862.
Roseman, R. R., et al. (1993). The su(Hw) protein insulates expression of the D. melanogaster white gene from chromosomal position-effects. EMBO J. 12, 435–442.
Zhao, K., et al. (1995). Visualization of chromosomal domains with boundary element-associated factor BEAF-32. Cell 81, 879–889.
Zhou, J., and Levine, M. (1999). A novel cis-regulatory element, the PTS, mediates an anti-insulator activity in the Drosophila embryo. Cell 99, 567–575.