1.

1 Nucleosomes: the Building Blocks of Chromatin

Chromatin is the complex of DNA and cellular proteins which form eukaryotic

chromosomes. It is composed of an elementary repeating unit called the nucleosome,

which is the major factor of DNA packaging in eukaryotic genomes (Figure 1.1).

Figure 1.1: A hierarchical view of chromatin structure. Reproduced figure (Hartl &
Jones, 1998).

Nucleosomes are DNA-protein complexes, which are comprised of a core

particle of 1.6 left-handed turns of DNA (roughly 146 bp) wound around a protein

complex called the histone octamer (Figure 1.1(B)). The histone octamer is a set of 8

basic proteins, which are among the most well conserved proteins known in

eukaryotes. It is comprised of a central tetramer, (H3/H4)2, flanked by two H2A/H2B

1-2
dimers (Figure 1.2). The structure of a single histone molecule includes three major α

helices with positively-charged loops protruding at the N-terminals.

Figure 1.2: Top-level view of a nucleosome. Cylinders indicate alpha-helices; white

The DNA wrapped around the histone octamer is called the core DNA and the

DNA joining adjacent nucleosomes is called linker DNA. Unlike core DNA, linker

DNA exhibits great variability in length: anywhere between 8 to 200 bp. This

variation in the length of linker DNA may be important for the diversity of gene

regulation; however, chromatin structure formation is independent of the length of

linker DNA (Kornberg & Lorch, 1999).

The constraint of the nucleosome on the DNA path forms the first level of

higher-order packing, compacting DNA by a factor of ~6 (Lewin, 2000). An extra

histone H1 (also called the linker histone) may also be present, clamping the DNA at

the position at which it enters and leaves the histone core (Karrer & VanNuland,

1999; Satchwell & Travers, 1989; Widlund et al., 2000).

The series of nucleosomes along a DNA sequence then coil into a helical array

forming a fibre of ~30 nm (Figure 1.1(C)); this results in further compaction by a

factor of ~40. In the recent crystal structure of the nucleosome (Luger et al., 1997),

1-3
it had been reported that the basic tail of H4 protrudes extensively and makes contacts

with acidic patches of H2A and H2B on neighbouring octamers; this implies a role for

H4 in stabilizing higher level structures. Histone H1 is thought to appear mainly

towards the middle of the 30 nm fibre where it may play a role in stabilizing

chromatin interactions (Staynov, 2000). Specialised nucleosomes are also known, for

example the centromere-specific nucleosomes, which contain a variant of histone H3

called CENP-A; these occur in a range of organisms from yeast to human (Smith,

2002). Many non-histone chromatin proteins also interact with histones to enable

formation of higher-order structures. The fibre itself undergoes further levels of

packaging resulting in compaction by a factor of ~1000 in interphase euchromatin and

~10,000 in heterochromatin (Figure 1.1(D-F)).

The structure of chromatin is dynamic. It exists in a number of distinct

functional states which can often be characterised by the level of transcriptional

activity. The dynamic transitions between these states occur through a range of post-

translational modifications of the histone tails which includes acetylation and

phosphorylation (Jenuwein & Allis, 2001). This forms the basis of the “histone code

hypothesis” which states that the combinatorial nature of these modifications results

in the generation of altered chromatin structures that mediate specific biological

responses (Turner, 2000).

1-4
1.2 DNA-Protein Interactions in the Nucleosome Core

Particle

The earliest concepts for the association of DNA and histones in the core particle

came from image reconstruction analysis using electron micrographs (Klug et al.,

1980). At 20 Ǻ resolution, a left handed helical ramp was apparent on the octamer

surface and proposals were made for how the DNA-protein interactions might occur.

Since then, X-ray crystallography has helped to advance understanding of the DNA-

protein interactions involved in the nucleosome core particle. Milestones included the

solving of the nucleosome structure at 7 Ǻ resolution (Uberbacher & Bunick, 1985),

which reconfirmed the initially inferred arrangement of histones and DNA. This led

to the highest resolution structures of the nucleosome core particle to date at 2.8 Ǻ

(Luger et al., 1997) and 1.9 Ǻ (Davey et al., 2002).

The high-resolution structure of the core-particle firstly revealed that the core

particle had a pseudo-dyad1 axis of symmetry: 1 bp sat on the dyad axis of the

octamer. It further revealed in fine detail that the histone-DNA interactions were

confined towards the phosphodiester backbone of the DNA strand (Luger et al.,

1997). Arginine/lysine-rich tails, protruding from the core histones, made “hook-like”

contacts every 10 bp where the minor groove of the double-helix faced inwards. The

histone-DNA contacts were non-base-specific and included predominantly salt-

bridges and H-bonds as well as non-polar contacts with DNA sugars.

The 10 periodic contact feature of the DNA backbone was suggested much

earlier. It was suggested, for example, when 10 bp-phased digestion patterns were

observed upon using the enzyme DNase I2 to cut nucleosome-bound DNA (Wang,

1
The central axis of the histone octamer is herein referred to as the dyad axis.
2
DNase I is an endonuclease, which breaks phosphodiester bonds within DNA.

1-5
1982). The observed cutting periodicity of 10 bp, which is “in phase” with the helical

periodicity of DNA, forms the basis of many computational approaches aimed at

finding nucleosome rotational positioning signals (Section 1.9).

The helical periodicity of DNA is not constant as it traverses around the

histone octamer. For example, experiments using hydroxyl-radical cleavage of

nucleosome-bound DNA showed that the helical periodicity was 10.0 bp/turn in the

vicinity of the dyad axis and 10.7 bp/turn towards the ends of the nucleosome (Puhl &

Behe, 1993). Most experimental evidence for B-DNA in solution suggests that it has

a helical periodicity of 10.5–10.6 bp in solution (Wolffe, 1998). This variation in

DNA periodicity along the core particle is thought to be a consequence of local

histone-DNA interactions.

1-6
1.3 The Concept of Nucleosome Positioning

Nucleosome positioning has been proposed to be a potential mechanism for regulating

gene expression, providing the view that nucleosomes could play important roles in

addition to organizing higher order chromatin structures in eukaryotic cells. The term

‘positioning’ refers to a pre-determined organization of nucleosomes on a DNA

sequence. In contrast, in a random arrangement of nucleosomes, all DNA sequences

will have an equal probability of binding histones (Sinden, 1994). This gives rise to

the idea that the local DNA structure, which is affected by the underlying DNA

sequence, may play a role in positioning nucleosomes.

Two kinds of DNA structural patterns may thus be envisioned to direct

nucleosome positioning: those that strongly favour nucleosome formation and those

that strongly obstruct it. Nucleosome positioning can help to either selectively expose

functionally important DNA sequences by constraining their locations to the linker

region or impede accessibility to functionally important sequences by constraining

their location to within the core particle. This can impose another level of regulation

in gene expression, for instance, by controlling the accessibility of binding sites

available to RNA polymerases or specific transcription factors. Two kinds of DNA

structure-based nucleosome positioning have been described previously and these will

be discussed next (Sections 1.4, 1.5).

1-7