Structure and Function of the Core Nucleosome Particle: The Molecular Basis of Chromatin Organization

Cassidy Crook Chromatin and Genome Organization All cells face the considerable challenge of compacting the genome to a small fraction of its original volume such that it can exist within the space of a cell. Prokaryotes as well as eukaryotes accomplish this feat through the combined effects of protein interactions and DNA supercoiling. However, dissimilar to eukaryotes, the genetic material of prokaryotes is maintained in an irregular entanglement known as the bacterial nucleoid. The fluid nature of the nucleoid is governed principally by enzyme-mediated (as well as spontaneous) topological modifications, with the association of structural proteins serving largely to stabilize supercoiled DNA states. In contrast, DNA compaction in eukaryotes is achieved through an extensively ordered genome architecture for which an organizational hierarchy is defined. The nucleosome represents the fundamental unit of eukaryotic genome organization (Bates & Maxwell, 2006). Roughly 200 bp segments of relaxed eukaryotic DNA are regularly coiled around octameric histone proteins to form an assembly of nucleosome particles with a DNA packing ratio of around seven (Lewin, 2004). Nucleosomes arranged along a span of duplex DNA constitute a 10 nm chromatin fiber and have the canonical appearance of beads-on-astring (Olins & Olins, 1974). Subsequently, nucleosomes aggregate to form a 30 nm second-order chromatin fiber with a helical super-structure and a packing ratio of around forty. The 30 nm fiber is arranged into 20100 kbp loop domains which associate with nuclear matrix proteins and scaffold proteins, including topoisomerases and RNA polymerases, to achieve a packing ratio exceeding 1000. Surprisingly, little is known about chromatin organization beyond the level of the 30 nm filament. Higher order structures appear to include looped rosette configurations (or chromomeres), 100-130 nm chomomena, a 200-250 nm fiber, and, ultimately, the eukaryotic chromosome (Nelson & Cox, 2008). Based on this model of DNA compaction in eukaryotes, it has been suggested that the composition of the bacterial nucleoid exemplifies a rudimentary chromosome. Still, a fundamental difference between chromosome organization and nucleoid composition is the primary role of DNA-histone interactions in the manifestation of DNA supercoiling in eukaryotes as opposed to topological regulation achieved cheifly by the competition of topoisomerase enzymes within the nucleoid. Specifically, prokaryotes rely on the essential bacterial enzyme DNA gyrase, a type II topoisomerase (topo), to introduce negative supercoils into DNA through an ATP-dependent mechanism , while negative supercoiling in eukaryotic DNA is achieved exclusively through the winding of DNA around nucleosomal histone octamers. Interestingly, eukaryotic type II topos are not able to induce negative supercoiling in DNA, yet they carry out ATP hydrolysis through a mechanism reminiscent of bacterial gyrase. The free energy of ATP hydrolysis in eukaryotic topoII is utilized in what has been described as a highly inefficient manner, to needlessly drive sub-equilibrium topological simplification of DNA (Bates & Maxwell, 2006). Thus, one

might speculate that ATP-hydrolysis in eukaryotic topoII is an evolutionary relic of an ancestral prokaryotic topoII once essential for the maintenance of genome compaction through negative supercoiling. A Historical Perspective of Chromatin and the Nucleosome The term chromatin was first introduced by the highly-renowned German cytogeneticist Walther Flemming in the early 1880s and was so-named to convey its refractive properties and its affinity for dyes. In 1884 Albrecht Kossel described histon, a proteoid isolated from the phosphrous-rich acidic nuclein extracts of avian erythrocyte nuclei. Jointly, these preliminary discoveries established the basis for our current understanding of eukaryotic DNA organization. Yet, approximately 80 years after histones were first described, the discovery of a basic subunit of chromatin structure in 1973-1974, coined the nucleosome shortly thereafter, revolutionized the field of chromosome organization and sparked what has been recognized as the nucleosome era (Olins & Olins, 2003). Prior to the earliest results indicating a discrete sub-structure repeated in chromatin, the most widely accepted model for chromatin structure was the Pardon-Wilkins super-helical model. Based on the results of low-angle x-ray diffraction, the super-helical model of chromatin compaction essentially described a nucleohistone coiled-coil in which the histone protein structure was that of a long fibrillous coil around which DNA was wrapped (Olins & Olins, 1972, 1974, 1997; Pardon & Wilkins, 1972). In the early 1970s the work of Olins and Olins (1974), Woodcock (1973), and Kornberg and Thomas (1974) independently utilized electron scattering to visualize chromatin filaments at low-resolution. Consistently, chromatin appeared to form irregularly distributed spherical bodies, designated nu-bodies, which exhibited a tendency to arrange in clusters (Olins & Olins, 1974; Woodcock, 1973; Kornberg & Thomas, 1974). Micrograph images of nucleosomes were paralleled by experimental evidence demonstrating a repeating structure in condensed DNA which facilitated protection of contiguous 200 bp segments in DNase digest assays (Hewish & Burgoyne, 1973). Furthermore, studies characterizing the properties of histone proteins made vital contributions to a unified theory of chromatin structure. Notably, such studies demonstrated the presence of one of each class of histone per 100 bp of DNA (excluding H1), the in vitro assembly of an (H3-H4)2 tetramer, and the vital role of histones H2A, H2B, H3, and H4 in the formation of chromatin beads (Kornberg, 1974). Thus, by the mid-1970s the sum of experimental data relating to chromatin was sufficient to formulate a cohesive theory for chromatin structure and so the nucleosome model was conceived. In a basic sense, the nucleosome model identifies the nucleosome unit as the basic repeating sub-structure of chromatin. Specifically, this model recognized, in considerable detail, the structure and organization of the nucleosome as comprised of a 200 bp segment of DNA oriented around a tripartite histone octamer composed of a (H3-H4)2 tetramer and two H2A-H2B dimers (see figs. 2, 3, & 4) (Kornberg, 1974). The Core Histone Architecture

Fundamentally, the histone octamer is composed of the four primary histone monomers, H2A, H2B, H3, and H4, each of which is structurally similar to the rest. In particular, the core histones are distinguished by a unique secondary structure known as the histone-fold. The histone-fold is a three-helix motif composed of an N-terminal 11-residue alpha-helix (I), a long central 27-residue helix (II), and a C-terminal 11-residue helix (III) consecutively joined by two short flexible linking domains each containing a single beta-sheet. The three helices of the histone-fold are arranged with respect to one another such that they create an overall Z-like form.The N and C terminal regions are described as helix-sheet-helix regions (HSH) and are designated HSH1 and HSH2 respectively (see fig. 1). Additionally, the amino acid residues of the histone-fold have been categorized into four classes: surface, self, pair, and interface residues. Of relevance, surface residues are positioned on the outside of the histones in the context of the nucleosome to facilitate histoneDNA interactions, pair residues are involved in histone dimerization, and interface residues contribute to the contacts formed between two dimers either between the H2A-H2B dimers or in the formation of the (H3H4)2 tetramer. Surface residues demonstrate sequence conservation of the basic amino acids Arg and Lys, which are critical in the binding of DNA to the histone octamer through ionic interactions. The high degree of structural similarity between the core histones is thought to be essential to the nature of DNA binding and chromatin organization in general. It has been suggested that transitions in chromatin are achieved through discrete changes in dimer-dimer and dimer-tetramer interactions. One mode of structural regulation may be an inherent susceptibility of the histones to subtle environmental changes at the interface regions between the three histone multimers. This idea is supported by the observation that interface residues are generally found to be less hydrophobic and more diversified than pair residues (Arents & Moudrianakis, 1995). The stability of the core histone octamer is DNA-dependent at physiological conditions; however, biochemical studies have demonstrated that the octamer is also stable in vitro under highly ionic conditions, particularly at high NaCl concentrations. This property of the octamer ultimately enabled the elucidation of high-resolution structures in the absence of DNA (Ramakrishnan, 1997). The first accurate crystal structure of the core histone complex was not solved until 1991 (Arents et al, 1991). The 1991 structure, solved at 3.1 , demonstrated that a previous structure, resolved at 3.3 , was critically flawed the unfortunate consequence of a subtle error in the coordinates for heavy atoms along the octamers two-fold axis of symmetry, resulting in an erroneous phase estimate (Ramakrishnan, 1997). Nonetheless, the 3.1 resolution structure of the nucleosome core histone complex represents a major leap in our understanding of the nucleosome and its role in chromatin structure and regulation. The nucleosomal core histone octamer is arranged in a tripartite left-handed superhelix made up of two H2A-H2B dimers and a central (H3-H4)2 tetramer. Core complex assembly is preceded by the binding of individual histone monomers in an antiparellel head-to-tail arrangement, forming a characteristic handshake motif. This motif describes the unique way in which the long helices of paired histone monomers intersect such that the HSH domains converge to create a structure that is distinctly similar to two hands clasped

together. In contrast to localized contacts formed when most proteins associate, histones form extensive interactions spanning the entire chains (Arents et al. 1991). Following H3-H4 dimerization, formation of the (H3-H4)2 tetramer occurs through the binding of a pair of H3-H4 heterodimers at the HSH2 domains of the H3 subunits. Tetramerization represents the initial step of nucleosome assembly. The H3-H3 C-terminal interaction is arranged in a 4-helical bundle and is facilitated in part by an essential His113 residue, which forms a hydrogen bond with Asp123 buried within the two helices of the adjacent chain (Luger et al. 1997). Docking of HSH2 domains to form the (H3-H4)2 complex results in a twisted crescent structure with a central hinge at the H3-H3 interface through which a two-fold axis of symmetry lies (see fig. 3) (Ramakrishnan, 1997). Moreover, each half-crescent of the tetramer is rotated approximately 15 away from the two-fold axis (Arents et al. 1991). Assembly of the core histone complex is completed through a final octamerization step in which H2A-H2B heterodimers bind to the (H3-H4)2 tetramer at its terminal regions. The H2B histone interacts with the H4 histone at their respective HSH2 domains. Due to the conserved structure of the HSH2 domain among the four histones the H2B-H4 interaction is remarkably similar to the association of the H3 histones during tetramerization. HSH2 domains at the H2B-H4 interface constitute the 4-helical bundle in which His74 of H4 inserts between the two adjacent helices to form a hydrogen bond with Glu90 on the H2B chain (Luger et al. 1997). Importantly, the H2B-H4 interaction is weaker than the tetrameric H3-H3 interaction, despite a more extensive contact interface between H2B and H4; thus, the association of H2B and H4 is more susceptible to changes in the surrounding solvent (Ramakrishnan, 1997). Importantly, the fully assembled histone octamer assumes the form of a left-handed protein superhelix. The radial axis of symmetry for the superhelical conformation of the octamer runs perpendicular to its two-fold axis of symmetry. When the complex is viewed down its radial axis it assumes the appearance of a disk with a diameter of around 65 ; this would be akin to viewing a vertically oriented cylinder from above or below. The outer surface of the protein complex is characterized by regularly arranged ridges and grooves that constitute an unambiguous path with a pitch of 28 that follows the left-handed superhelical form of the protein. This path creates an ideal binding site for DNA to wrap around the surface of the histone octamer like thread on a spool (see fig. 4; Arents et al. 1997). DNA Binding on the Histone Octamer Structural studies of the nucleosome have been largely facilitated by the use of micrococcal nuclease (MNase). Digestion of unraveled chromatin using MNase is initiated at the nucleosomal DNA linker regions to produces a size distribution of 10 nm chromatin fragments, described as oligonucleosomes. However, oligonucleosomes tend to exhibit poor stability in solution. Extended digests with MNase result in structures of progressively increased stability, such that particle stability is directly proportional to the length of the free DNA fixed to the nucleosome core. Of interest, particles consisting of a 165 bp DNA component can be obtained from digests and are unique in that they are often associated with the H1 histone. This More extensive nuclease digestion results in the isolation of the highly stable core nucleosomal particle. This

discrete unit is defined by a DNA element of around 146 bp wound approximately 1.65 turns around the histone octamer in a left-handed superhelix (Bates & Maxwell, 2006). Due to its stability, the nucleosome core particle has provided the basis for x-ray diffraction studies and is, thus, central to our present understanding of nucleosomal DNA-histone surface interactions. Similar to the subunit organization of the histone octamer, binding of a continuous DNA segment on the surface of the histone complex is facilitated predominantly through the conserved intrinsic properties of the histone-fold motif. Yet, DNA-protein interactions are coordinated largely on the scale of the handshake motif that is, through surface contacts formed with the four histone pairs and irrespective of the tripartite nature of the octamer. Each dimer is directly associated with approximately 27-28 bp of DNA, while unbound 4-bp segments of duplex join the discrete regions of DNA-histone contacts. Furthermore, an invariable Lys residue occupying the second position of the C-terminal loop II doman (L2) of each histone monomer (see fig. 1) forms a salt bridge with a distal phosphate group of the DNA backbone such that Lys residues of adjacent dimers traverse one another and effectively form a dimer-DNA-dimer cross-linkage (see fig. 5; Ramakrishnan, 1997; Luger et al. 1997). Extruding L2 Lys residues of the H2A-H2B dimers at the superhelical ends of the histone octamer facilitate docking of the remainder of DNA bases in the formation of the 165 bp core particle. The histone-DNA contacts occur predominantly through basic amino acid residues bonding with the negatively charged phosphate groups of the DNA duplex to establish a sequenceindependent association. This general mode of ionic interaction is a recurrent theme in nucleoprotein complexes. However, the helical structure of DNA in the nucleosome is such that only two adjacent phosphate groups per DNA strand are within direct hydrogen bonding range of the histone-fold dimers. Despite this, it is thought that solvation of the individual macromolecules mediates nucleosome assembly through the formation of water-bridges which transiently stabilize indirect hydrogen bonding between the two structures (Davey et al. 2006). The primary DNA binding site for each histone dimer is positioned parallel to the central junction of the II helices, such that it is localized amid the N-termini of the 1 helices, which form arm-like protrusions converging on the adjacent DNA backbone phosphates. Thus, the positive charges of the two I dipoles, as well as main chain amide groups, and variable side chain interactions act cooperatively to create a stable interaction between the DNA and histone dimer (Luger et al. 1997). Nucleosome Phasing and Gene Expression Histone-DNA interactions within the nucleosome are achieved predominantly through non-specific histone contacts with the DNA phosphodiester backbone. Nevertheless, a preferential specificity for histone binding at certain DNA sequences has been observed. Early descriptions of nucleosome specificity include work by Simpson and coworkers (1983) in which nucleosome reconstitution was carried out in vitro using a 260 bp DNA fragment encoding the 5S rRNA from the sea urchin L. variegatus. A strong preference was identified for nucleosome localization within the rRNA coding sequence such that the transcription start site

was positioned precisely at the center of the nucleosomal DNA superhelix (i.e. proximal to the H3-H3 junction) (Simpson & Stafford, 1983). In addition to preferential nucleosome binding in L. variegatus, this intrinsic property of nucleosomes has been distinguished in various other contexts, including the 5S rRNA locus of Drosophila and Xenopus, as well as at discrete sites in the SV40 genome (Ramakrishnan, 1997). The phenomenon of sequence-dependent nucleosome positioning is broadly referred to as nucleosome phasing. In terms of function, it has been suggested that nucleosome phasing is integral to the formation of higherorder chromatin structures by directly impacting linker DNA length and, thus, the flexibility of the 10 nm fiber. Moreover, sequence-dependent phasing suggests a unique role for the nucleosome in the suppression of transcription by physically obstructing RNA polymerases during initiation and elongation of RNA synthesis; thus, nucleosomes (and perhaps higher order chromatin structures as well) possess crucial generegulatory significance (Blank & Becker, 1996). However, prior to considering the myriad cellular consequences of nucleosome phasing, it is critical to resolve the molecular underpinnings of nucleosomal sequence specificity. Nucleosome phasing principally implies that electrostatic interactions cannot be the sole factor governing the way in which DNA is wrapped around a histone octamer. Rather, the process of nucleosome formation is inherently multipartite and, accordingly, far more dynamic than first suggested by structural studies. Demonstrating remarkable intuition, studies by Drew and Travers (1985) provided early insights to the molecular basis of sequence specificity in nucleosome positioning. Such work united two distinct emerging fields, nucleosome phasing and intrinsic DNA bending, to consider the former in terms of the latter. Electrophoretic fractionation studies of tightly compacted DNA minicircles in the trypanosome L. tarentolae led to the discovery of DNA bending at phased adenine/thymine (AT) tracts due to intrinsic deformation of the DNA helix, characterized by minor groove compression (Marini et al. 1982; Koo et al. 1986). It was found that intrinsically bent DNA is configured such that the minor groove at AT tracts is positioned on the inner face of the DNA bend, whereas intermittent GC clusters appear to be distinguished by a minor groove oriented outwards. Nucleosome reconstitution experiments using DNA oligonucleotides with sequence-directed curvature demonstrated that the helical conformation of intrinsic bending was conserved in the arrangement of DNA in the nucleosome (i.e. the rotational setting of the nucleosome). Consequently, it was suggested that intrinsically bent DNA may favor histone binding, while inherently rigid DNA sequences may prohibit nucleosome assembly (Drew & Travers, 1985). Further studies comparing the structure of DNA in the nucleosome to that of free DNA have bolstered the idea that the inherent flexibility of a DNA sequence conveys preference in nucleosome positioning (Hayes et al. 1990). Importantly, histone octamerization is accomplished in a DNA-dependent manner, wherein the (H3-H4)2 histone tetramer binds to DNA before the two H2A-H2B dimers are incorporated in the core histone complex. If one considers the general curved architecture of the histone tetramer (see Fig. x) it is apparent that this structure provides a suitable scaffold for preferential DNA binding on the basis of intrinsic flexibility (Dong & Holde, 1991).

Although the bent shape of the histone tetramer may underlie nucleosome phasing, it is critical to conceptualize phasing in terms of the kinetic and thermodynamic challenges to nucleosome formation. Namely, DNA bending by the (H3-H4)2 tetramer is a thermodynamically unfavorable process due to the structural distortions imposed on native B-form DNA. However, AT rich sequences are relatively more flexible than GC rich sequences due to the fact that A-T base pairing forms two hydrogen bonds, whereas GC base pairing involves three. Thus, the free energy required to induce deformation of an AT rich sequence is comparatively lesser than that needed to deform GC rich DNA. Likewise, DNA that possesses intrinsic curvature due to the presence of A/T tracts in phase with the helical periodicity presents an optimally suited binding site for the (H3-H4)2 tetramer. Such sequences with highly favorable rotational settings minimize the energetic requirements of nucleosome formation and, consequently, facilitate preferential nucleosome binding (Ramakrishnan, 1997). Therefore, while nucleosome positioning is certainly sequence-directed, it is not precisely sequence-specific; rather, a nucleosome will bind over a contiguous range of sites about a preferred sequence due to the comparable rotational settings of adjacent DNA regions. Consistent with this idea, nucleosome positioning has been found to be characterized by major and minor sites at a given locus (,). This property of nucleosome phasing reflects a limited role for DNA sequence in nucleosome binding. Major and minor nucleosome binding sites are thought to reflect local fluctuations in nucleosome position, wherein the complex transiently dissociates and reassembles at an adjacent site. This kinetic hallmark of the nucleosome is referred to as nucleosome mobility and is dependent on the degree of stability that a given complex possesses. Accordingly, nucleosomes are found to occupy major and minor sites at an equilibrium distribution (,). Mobility is a fundamental characteristic of nucleosomes because complex stability directly establishes the ease with which RNA polymerase (RNAP) can transcribe through nucleosomal DNA. It is thought that many RNAPs, including phage RNAP, bacterial RNAP and eukaryotic RNA polymerases II and III (Pol II and PolIII), adopt a paused conformation upstream of a nucleosome, such that the histone complex may effectively step around the transiently immobilized RNAP by means of a DNA loop (Workman & Kingston, 1998; Hodges et al. 2009; Workman, 2006). Of primary significance, nucleosome displacement and histone transfer during transcription-elongation are explicitly dependent on the mobile capacity of a given nucleosome complex. A seeming quandary then arises: while AT rich DNA sequences favor nucleosome formation, nucleosomes with an AT rich DNA moiety (including intrinsically bent DNA) are so-favored because they possess optimal thermodynamic stabilization and, thus, preclude transcription by RNAP. This problematic relationship between nucleosome positioning and nucleosome mobility during transcriptionelongation appears to be resolved through nucleosome (de)stabilization factors and epigenetic histone modifications that regulate the strength with which a histone octamer binds DNA (Henikoff, 2008). Posttranslational [histone] modifications (PTMs) are localized to the C-terminal domains (CTDs), which are characteristically short tails, less than 40 amino acids, extending from the globular body of the histone

complex (Campos & Reinberg, 2009). In this way, histone CTD tails are thought to act as platforms for covalent modifications, including ubiquitination, phosphorylation, methylation, acetylation, and SUMOylation (i.e. the addition of small ubiquitin-like modifier proteins). Collectively, histone modifications comprise a highly orchestrated epigenetic histone code, which appears to specifically regulate the strength and efficiency of intra-nucleosomal contacts, as well as extra-nucleosomal interactions (e.g. nucleosome-nucleosome contacts in the 30 nm fiber) by inducing conformational changes in the nucleosome complex. In addition to directly altering intra- and extra-nucleosomal histone contacts, PTMs can be read by effector proteins through particular domains, such as chromodomains, which recognize methylated sites, and bromodomains, which distinguish acetylated residues (Campos & Reinberg, 2009). The tremendous complexity of the histone code is established through the diverse classes of PTMs, the various residues to which PTMs can be made, and the fact that a single amino acid can acquire several modifications (Berger, 2007). Ongoing research aims to further decipher the histone code and improve our current understanding of nucleosome remodeling. Nevertheless, it is apparent that such phenomena contribute significantly to global chromatin organization and the large-scale regulation of gene expression. In the past forty years our understanding of chromatin organization and its functional implications has expanded exponentially. Since the nucleosome model of chromatin organization was first described by Roger Kornberg in 1974, our perception of chromatin and its fundamental structure, the nucleosome, has undergone an ideological metamorphosis. No longer is the nucleosome simply the primary unit of chromatin by which genome compaction is achieved in eukaryotes. Rather, we now recognize the nucleosome, as well as chromatin in general, in terms of the paradigms of gene expression, and thus cellular diversity, which they mutually establish. Most remarkable, this monumental function of chromatin organization is communicated through discrete molecular modifications to the primary structure of the nucleosomal histone core; thereby, specifying nucleosome structure as central to the vast epigenetic language of our cells.

Loop II HSH1

I N term.

III II

C term.

HSH2
Loop I

Figure 1. Histone H2A of X. laevis resolved at 2.8 resolution the histone fold. N and C terminal helix-sheet-helix motifs are displayed at helices I and II. Also, note the overall inverted Z-form characteristic of the fold.

Figure 2. The H2A-H2B heterodimer of X. laevis resolved at 2.8 resolution.

HSH1

HSH1

HSH2

HSH2

H3-H3 interface

Figure 3. H3-H4 tetramer of X. laevis resolved at 2.8 resolution. The general form of the tetramer is that of a crescent. The tetramer is formed through contacts between the H3 histones of the two dimers via binding at the HSH1 domains to form a four helical bundle composed of helices I and II.

Figure 4. Complete nucleosome complex (top- side view, bottom-birds-eye.). The DNA follows the left-handed superhelical organization of the histone complex (above). H2AH2B dimers are displayed in yellow and cyan, while H3-H4 dimers are shown is blue and magenta. The structure was obtained from X. laevis and resolved at 2.8 resolution by Luger et al.

Figure 5. Loop II Lys residues at the H3-H3 tetramer interface facilitate cross-linkage between adjacent histone dimers through contact with distal phosphate groups on the DNA backbone.

References Arents, G. et al. (1991). The nucleosomal core histone octamer at 3.1 resolution: a tripartite protein assembly and a left-handed superhelix. PNAS. 88: 10148-52. Arents, G. & Moudrianakis, E.N. (1995). The histone fold: a ubiquitous architectural motif utilized in DNA compaction and protein dimerization. PNAS. 92: 11170-74. Bates A. D. & Maxwell A. (2006) DNA topology. New York, NY: Oxford University Press, Inc. Berger (2007). The complex language of chromatin regulation during transcription. Nature. 447: 407-412. Blank, T.A. & Becker, P.B. (1996). The effect of nucleosome phasing sequences and DNA topology on nucleosome spacing. J. Mol. Biol. 260: 1-8. Campos & Reinberg (2009). Histones: annotating chromatin. Annu. Rev. Genet. 43: 559-599. Davey, C.A. et al. (2006). Solvent mediated interactions in the structure of the nucleosome core particle at 1.9 resolution. J. Mol. Biol. 319: 1097-1113. Dong, F. & Holde, K.E. (1991). Nucleosome positioning is determined by the (H3-H4)2 tetramer. PNAS. 88: 10596-10600. Drew & Travers (1985). DNA bending & its relation to nucleosome positioning. J. Mol. Biol. 186: 773-90. Hayes, J.J. et al (1990). The structure of DNA in a nucleosome. PNAS. 87: 7405-09. Hodges et al (2009). Nucleosomal fluctuations govern the transcription dynamics of RNA polymerase II. Science. 325: 626-628. Koo, H. et al (1986). DNA bending at adeninethymine tracts. Nature 320, 501506. Lewin, B. (2004) Genes VIII. Upper Saddle River, NJ: Pearson Prentice Hall Luger, K. et al. (1997). Crystal structure of the nucleosome core particle at 2.8 resolution. Nature. 389: 25160. Marini, J.C. et al. (1982). Bent helical structure in kinetoplast DNA. Proc. Natl. Acad. Sci. USA 79, 76647668. Nelson, D.L. & Cox, M.M (2008) Principles of Biochemistry. New York, NY: W.H. Freeman and co. Olins, D.E. & Olins A.L. (1974). Spheroid chromatin units (nu bodies). Science. 183 : 330-32. Olins, D.E. & Olins A.L. (2003). Chromatin history : our view from the bridge. Nature. 4 : 809-14. Olins, D.E. & Olins A.L. (1972). Physical Studies of isolated eukaryotic nuclei. J Cell Biol. 53 : 715-36. Pardon, J.F. & Wilkens, M.H. (1972). A super-coil model for nucleohistone. J. Mol. Biol. 68: 115-24. Ramakrishnan (1997). Histone structure and the organization of the nucleosome. Annu. Rev. Biophys. Biomol. & Struct. Biol. 26: 83-112. Simpson, R.T. & Stafford, D.W. (1983). Structural features of a phased nucleosome core particle. PNAS. 80: 50-51. Thomas, J.O. & Kornberg, R.D. (1974). An octamer of histones in chromatin and free in solution. 72: 262630. Woodcock, C.L.F. (1973). Ultrastructure of inactive chromatin. J. Cell. Biol. 59: 368.

Workman & Kingston (1998). Alteration of nucleosome structure as a mechanism of transcriptional regulation. Annu. Rev. Biochem. 67: 545-79. Workman (2006). Nucleosome displacement in transcription. Genes Dev. 20: 2009-17. Henikoff, S. (2008). Nucleosome destabilization in the epigenetic regulation of gene expression. Nature. 9: 1526.