Histone Variants

In addition to several loci encoding core histones, eukaryotes also contain genes that encode variants of the core histones. These variants often impart special properties to the chromatin where they are incorporated and may, in part, provide a mechanism for inheritance of chromatin transcription states (Henikoff et al., 2004). Variant histones are incorporated into previously assembled nucleosomes by specific histone chaperones during DNA replication and/or by a process that may be coupled to transcription, independently of DNA replication. In this manner, specific variants that denote transcriptional domains can be maintained through multiple rounds of replication.

Specialized functions of the variant histones illustrate the breadth of influence that chromatin has on nuclear processes. CENP-A is an H3-like, histone variant that is found only in nucleosomes at the centromere, a chromatin, structural organization that plays an essential role during cell division. H3.3, another H3-like histone, is incorporated into chromatin throughout the cell cycle, unlike canonical H3.1, which enters chromatin during active DNA replication in S-phase of the cell cycle. H3.3 is enriched in transcribed chromatin. Chaperone-mediated swapping of H3.3 for existing H3.1, within a nucleosome, serves as a means of erasing repressive H3.1 post-translation modifications even in the absence of DNA replication (Tagami et al., 2004). Likewise, variants of H2A perform specific roles. In S. pombe, H2A.Z is incorporated near the edges of silenced chromatin and inhibits spreading of higher-order structures into active chromatin by promoting formation of the 30 nm chromatin fiber (Fan et al., 2002). Finally, H2A.X is localized to the foci of DNA repair proteins at the sites of double-strand breaks and is phosphorylated, either to mark the damage site or as a result of the process of repair.

Higher Order Structures

Electron microscopy studies of extracted chromatin reveal a 30 nm chromatin fiber, which can be further compacted, ultimately forming the mitotic chromosome (Adolph, 1981). For decades intense research has not yet resolved the molecular structure of this fiber, nor of any higher order chromatin structure. Chromatin reconstituted in vitro can undergo salt-dependent compaction from the nucleosomal array (at 1-10 mM salt concentration) to the 30 nm fiber (100-200 mM salt concentration) and addition of HI increases the rate of compaction (Carruthers et al., 1998). X-ray and neutron-scattering experiments led to numerous models of chromatin structure beyond repeated nucleosomes or "beads-on-a-string", including the solenoid, continuous superhelix, twisted-ribbon, helical-ribbon, and cross-linker (Horn and Peterson, 2002). These can be divided into one-start and two-start models based on whether one strand wraps around itself (solenoid, continuous superhelix, and twisted-ribbon) or two strands wrap around each other (helical-ribbon and crossed-linker). Current, in vitro analysis of a twelve- nucleosome array favors the two-start class of models due to the ability of an endonuclease to separate the array into two six-nucleosome halves (Dorigo et al., 2004). Deletion experiments have shown that the histone H4 tail is essential for compaction and that it interacts with neighboring H2A/H2B (Davey et al., 2002). Substitution of H2A with variant histone H2A.Z results in a higher rate of compaction (Fan et al., 2004).

In order to view higher order chromatin structures in a physiological setting, microscopy studies were performed to visualize expressed Green Fluorescent Protein (GFP)-LacI fusion proteins interacting with Lacl-binding sites, inserted as tandem, multi-copy arrays within an endogenous, genomic locus in cells. Exogenous expression of GFP-LacI proteins showed GFP-LacI fluorescence at a single focus of compacted chromatin.

However, when a GFP-LacI-VP16 fusion protein was used to simulate binding of a transcription-activating protein to chromatin, the single focus dispersed and spread into a ribbon of GFP-fluorescence, 80 to 100 nm in diameter (Tumbar et al., 1999). This type of chromatin structure, which likely contains three 30 nm chromatin fibers, may be the basic unit of higher order chromatin structure that is competent for activated gene expression.

Histones are Required for Life

While biochemists were analyzing the structure of chromatin and the nucleosome, and developing means of assembling chromatin in vitro (see below), geneticists were studying the impact that chromatin structure had on gene expression in vivo. These studies verified that histones, and thus chromatin, were essential for life itself. Deletion of the genes encoding histone proteins, using the yeast S. cerevisiae as a model organism, led to death of the yeast cells. Histone proteins are encoded by multiple copies of each gene; cells are viable with just one copy of each, but die in the complete absence of any one histone. Plasmid shuffle techniques were used where exogenously introduced plasmid DNA, encoding a single copy of a histone gene, could replace the endogenous copy. This methodology allows the regulated expression of specific histone proteins, as well as introduction of mutations or deletions within a histone gene, and analysis of their effect on gene expression in vivo. This approach, employing a genetically tractable model organism, revealed that individual regions of histones have specific functions in gene regulation (Grunstein, 1990). The amino-terminal tails of histones play major roles in transcription activation and repression. Deletions of these tails do not kill cells but the life cycle of yeast and expression of many genes, both activation and repression, are dramatically affected. Further deletions more C-terminal, into the hydrophobic cores of the histones, led to cell death. These findings show that histones perform essential functions in gene expression, both activating and repressing, and in maintenance of viability.

Heterochromatin and Euchromatin

Arguably, the most important function of chromatin may be repression and silencing of gene expression. While repressed genes may share some of the characteristics of continuously silenced chromatin, defined as heterochromatin, they retain the potential for active transcription and are placed in the subdivision of chromatin known as euchromatin. The term "silenced"

is used here with regards to a maintained absence of transcriptional activity, and "repression" for short-term or active down-regulation of transcription. Silenced chromatin (versus repressed chromatin) can be further subdivided into constitutive, primarily at centromeric and telomeric regions, and facultative heterochromatin, such as the inactivated X-chromosome (Maison et al., 2002; Richards and Elgin, 2002). Briefly, both types of heterochromatin are present within highly condensed and pericentric regions of chromosomes, often consisting of long stretches of repeated elements. They also share the properties of widespread histone hypoacetylation, extended areas of methylated histone H3 at lysine 9 (metH3K9) and DNA methylation. Additionally, recent results support the participation of unique, non-coding siRNAs and dedicated enzymatic complexes in the silencing of both constitutive and facultative heterochromatin (Grewal and Elgin, 2002), as will be discussed further in a later chapter.

Each heterochromatin subgroup may have its own specific, associated histone methyltransferases (HMT) and variant histones. The lack of acetylation at H3K9, which permits its modification by HMT's to form metH3K9, is characteristic of histone tails present in heterochromatin. The association of proteins such as Heterochromatin Protein 1 (HP1), which is usually concentrated in pericentric heterochromatin and telomeres, promotes spreading of silenced chromatin along the repetitive DNA characteristic of constitutive heterochromatin. HP1 interacts with metH3K9 via HPl's N-terminal chromo domain and further with HMT's through its C-terminal chromo-shadow domain. Thus, by interacting with both the enzyme and its substrate, HP1 ensures establishment and maintenance of a stable silenced state (Kouzarides, 2002; Richards and Elgin, 2002).

Histone deacetylation, metH3K9 modification and HP1 binding are associated with gene inactivation but are not unique to heterochromatin, as they also occur at repressed euchromatin (Nguyen et al., 2005; Nielsen et al., 2001; Schultz et al., 2002). Hypoacetylation, induced by histone deacetylases (HDAC's), and histone methylation, by HMT's, can spread along a region of silenced chromatin. The best-studied example of histone deacetylation/methylation linked to silencing is its propagation over a large region at the silent mating type HML/HMR loci of S. cerevisiae. Although not cytologically distinct as heterochromatin, silent mating type chromatin is never expressed and is heterochromatin-like. Sequence-specific, silencer-element binding proteins, such as Rapl, along with Ku70 and Ku80, recruit the multi-component, nucleosome-binding Sir (1-4)

complex in yeast. Sir2, a unique histone deacetylase, is dependent on the cofactor nicotinamide adenine dinucleotide (NAD) for its enzymatic activity (Tanny et al., 1999), catalyzes deacetylation of histone H3/H4 tails, which then facilitates the binding of Sir3 and Sir4. Interactions between the Sir proteins enable propagation of the silenced state (Gottschling, 2000; Shore, 2000). Interestingly, silencing and Sir2 activity, in particular, are linked to the process of ageing, which has generated considerable interest in this facet of chromatin function (Denu, 2003; Guarente, 2000).

Heterochromatin of higher eukaryotes is generally marked by methylation of DNA. Methylated DNA can serve as a recruiting center for enzyme complexes that promote chromatin inactivation, which can further support additional DNA methylation. DNA methylases, such as Dnmt, can methylate DNA following directions from histone modification codes rather than the DNA sequence itself. Recent evidence links histone methylation and DNA methylation through specific HMT-mediated methylation of H3K9 and recruitment of DNA methyltransferases. These HMT's, e.g, Neurospora crassa Dim-5 and Krypotonite from A. thaliana, through their enzymatic function at K9 of H3, recruit HP 1-like proteins that in turn attract specific DNA methyltransferases (Lindroth et al., 2004; Naumann et al., 2005; Tamaru and Selker, 2001). Cytosine-methylated DNA is bound by MeCP/MBD (methyl cytosine-binding protein/methyl binding domain) proteins, which act as adaptors between DNA methylases and histone deacetylases. There is also evidence for direct interaction between DNA methylase Dnmt and HDACs, which is independent of Dnmt enzymatic activity. Again, similar to histone methylation, DNA methylation-mediated recruitment of HDACs via MBDs may not be limited to heterochromatin silencing (Fuks et al., 2003; Jaenisch and Bird, 2003).

Silencing of gene expression must be tightly regulated, just as activation of gene expression is. Aberrant methylation of DNA at gene sequences enriched in C-G nucleotide pairs, CpG islands, has been reported in a number of tumor-derived cells. In this example, proximal promoter sequences of tumor suppressor genes are methylated at CpG sites. This DNA methylation and chromatin silencing disrupts regulatory protein binding, and the tumor suppressor gene cannot be activated to perform its protective functions when cellular homeostasis is threatened. Tumor cells may express abnormally high levels of active Dnmt's or other factors, normally protective of the CpG regulatory sequences found enriched in the promoters of many genes, are not active. The link between increased HMT activity, by a number of family members bearing conserved SET domains, and aberrant DNA methylation (Marmorstein, 2003) has led to combined therapeutic approaches whereby 5-azacytidine, an inhibitor of Dnmt's, and HDAC inhibitors, which can maintain acetylated H3K9 and inhibit its methylation, are used in an attempt to re-activate aberrantly silenced tumor suppressor function (Baylin et al., 2001; Jones, 2002).

Questions of accessibility

As structural knowledge of chromatin increased, the natural question arose of how DNA, wrapped around a nucleosome and condensed into fibers, can be accessed for transcription. It was apparent to microscopists that chromatin structure was different somehow when actively transcribed. Stained images of polytene chromosomes revealed an unevenly spaced banding pattern and physical distortions along the length of the chromosomal fibers. Polytene chromosomes, consist of thousands of copies of the same chromosomal DNA, which are amplified, align one alongside the other and are highly enriched in the giant cells of Drosophila salivary glands. These amplified chromosomes are visible by light microscopy and can be stained or probed for specific RNA expression, by in situ hybridization, or the presence of specific proteins, using immunological or histochemical methods (Fig.5.3). Regions of polytene chromosomes that are actively transcribing RNA are expanded or "puffed", while regions that are silent and not transcribing are more condensed. These patterns of RNA expression accounted for the distorted shape of the chromosome fiber, observed under the microscope. What causes the expanded regions to puff? Is it the act of RNA polymerase moving along the chromatin and pushing its way through nucleosome after nucleosome? Are there mechanisms at work that open up the chromatin structure to allow passage to a polymerase, and others that close up the chromatin behind it? What marks a certain section of a chromosome for activation at a specific time in a particular cell? How does RNA polymerase find this specific site when chromatin is so highly structured?

Biochemists attempted to study these problems much the same way as reconstruction of basal transcription proceeded, by in vitro assembly of the process using purified components. The additional challenge, beyond the essential need to purify and reconstitute active RNA polymerase and basal transcription factors, as discussed in the previous chapter, was how to build chromatin in vitro that bore any resemblance to chromatin in real life. Bringing pieces of chromatin from living cells into the test tube and attempting to transcribe them was not a successful approach. Most of the chromatin in any given cell is silent or heterochromatic, and isolation of chromatin in the act of gene expression is difficult. The problem of how to build a nucleosomal substrate for RNA polymerase to copy needed to be addressed first.

0 0

Post a comment