Lecture1-2_08 - 01_TRE_001-046.qxd:TransRegEukaryotes.qxd...

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/24/08 11:40 AM Page 1 1 A Primer on Transcriptional Regulation in Mammalian Cells INTRODUCTION AND OVERVIEW INTRODUCTION AND OVERVIEW, 1 Brief summary of genome-wide methods, 2 The combination of proteomic, molecular, celCHROMATIN AND THE GENERAL lular, and genome-wide studies has revealed TRANSCRIPTION MACHINERY, 5 numerous insights into the mechanisms underChromatin structure and organization, 5 lying differential regulation of transcription in Chromatin modification, 8 eukaryotes. Yet the field is only beginning to Chromatin remodeling, 11 glimpse the complexity of regulation, even in The general transcription machinery, 15 organisms as simple as Saccharomyces cerevisiae. Organization of the regulatory region, 15 In humans, vast networks of transcription facBasal transcription complex assembly and tors control waves of transcription that result in initiation, 22 the differential transcription of approximately Mediator, 25 TFIID and TAFs, 26 25,000 genes during development, differentiation, and signaling. This transcription involves ACTIVATION AND REPRESSION, 28 the recruitment of RNA polymerase II (Pol II), a Gene activation, 28 coactivator termed Mediator, and six ancillary Chromatin modification and remodeling during transcription initiation, 29 factors termed TFIIA, TFIIB, TFIID, TFIIE, A model for recruitment of the general TFIIF, and TFIIH (see below). These factors go machinery, 29 through a defined cycle to permit initiation and Initial stages of Pol II elongation, 31 reinitiation of transcription (Fig. 1.1). Pol II encounters nucleosomes within the gene, 34 Chromosomes appear to occupy distinct Silencing/repression of transcription, 35 territories (chromosome territories, or CTs) in CONCLUDING PERSPECTIVE, 41 the nucleus, as measured by fluorescence in situ hybridization (FISH), and it has been hypothesized that interchromosomal communication occurs because of dynamic movements of large chromatin loops (Kadauke and Blobel 2008), which may be directed by actin or nuclear myosin I (Branco and Pombo 2006; Cremer et al. 2006). Transcription is thought to occur in any of 200– 2000 transcription factories present within and between CTs (Sexton et al. 2007; Carter et al. 2008; Mitchell and Fraser 2008). In all eukaryotes, the chromosomal DNA is packaged into chromatin. Chromatin can be subdivided into heterochromatin and euchromatin. Heterochromatin replicates late and is densely staining, compact, and often present at the nuclear periphery, near nucleoli or in chromocenters. Heterochromatin is generally considered transcriptionally silent, although recent observations in the field suggest that limited transcription occurs. Euchromatin is light staining, decondensed, and usually, but not exclusively, located toward the interior of a nucleus. Euchromatin replicates early and contains transcriptionally active regions of the genome. Facultative chromatin is heterochromatin that has the ability to become euchromatin or vice versa (Trojer and Reinberg 2007). The basic unit of chromatin is the nucleosome, which comprises 147 bp of DNA, two copies each of four core histones and, in some contexts, a linker histone (Horn and Peterson 2002; Luger and 1 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 2 2 / Chapter 1 Recruitment Pol II IIF IID IIE IIH Med Preinitiation complex (PIC) Med Pol II IIF IIH Open complex (unstable) ATP IIH Med Pol II IIF Initiation NTPs Elongation IIA IID A TAT IIA IID A TAT RNA IIB IIA IIE IIB IIE IIB Pol II IIB IIF + Promoter DNA Pol II Reinitiation Med IIB IIF IIH IIE IIA IID A TAT Scaffold complex FIGURE 1.1. Pathway of transcription initiation and reinitiation for Pol II. (Adapted, with permission of Macmillan Publishers Ltd, from Hahn 2004.) Hansen 2005; Woodcock 2006). Many of the histones have several variants and isoforms, which appear to play specific roles in gene regulation. Nucleosomes themselves are arranged into higherorder structures and are subject to numerous forms of regulation. The process of gene activation and silencing is orchestrated by sequence-specific DNA-binding proteins that recruit enzymes and structural proteins involved in altering the chromatin template and transcribing the gene. There are estimated to be approximately 1800–1900 sequence-specific binding proteins encoded in the human genome, far less than the number of genes or mRNA expression patterns. Therefore, even the simplest genes in humans are regulated combinatorially by several sequence-specific factors (Maston et al. 2006). Activators stimulate transcription in a synergistic manner such that the gene responds not to a single regulatory signal but to several converging signals that control different activators. This type of combinatorial regulation is probably one of the most important aspects of differential gene expression (Carey 1998) (see Chapter 12). This chapter is a basic overview of transcription and chromatin regulatory mechanisms. Specific topics are discussed further in the ensuing chapters. We have chosen to cite review articles where possible because the topics covered involve a considerable amount of primary literature, and a novice in the field would probably require a general introduction to each of the areas discussed herein. This chapter focuses solely on Pol II transcription. Excellent reviews have appeared on Pol I, which transcribes genes encoding the 45S precursor to the large ribosomal RNAs, and Pol III, which transcribes many small RNA genes including those for tRNA and 5S ribosomal RNA (Haeusler and Engelke 2006; Russell and Zomerdijk 2006; Chedin et al. 2007; Dieci et al. 2007; Cramer et al. 2008). In plants, a fourth RNA polymerase is present to synthesize microRNAs (miRNAs) (Vaughn and Martienssen 2005), whereas organelles such as mitochondria have their own RNA polymerase (Scarpulla 2008). Brief Summary of Genome-wide Methods A brief overview of methods may be helpful to understand some of the information being presented in this primer although we will also cover some of these approaches in Chapter 9. Genome-wide protein–DNA interactions are generally studied by chromatin immunoprecipitation (ChIP) of fragmented chromatin from formaldehyde-treated nuclei (Kim and Ren 2006). The formaldehyde crosslinks proteins to DNA, and the crosslinked protein–DNA complex is fragmented by sonication or nuclease treatment and is then subjected to immunoprecipitation with antibodies targeting the 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 3 A Primer on Transcriptional Regulation in Mammalian Cells / 3 protein or histone modification of interest. The precipitated DNA is isolated, crosslinks are reversed, and the DNA is amplified by polymerase chain reaction (PCR) using fluorescent nucleotides. The amplified DNA is then hybridized to chips bearing tiled DNA arrays representing promoters (promoter arrays) and interesting regions of the genome, including some promoters and enhancers (i.e., the ENCODE arrays [The ENCODE Project Consortium 2004]). Alternatively, custom-designed arrays can be constructed. The amount of hybridization as measured by the fluorescent signal is a measure of binding and is detected by a fluorescence reader and quantitated by using the accompanying software. These so-called ChIP-chip (or ChIP-on-chip) assays have been widely used in the field to decipher the regulatory circuitry controlling transcription (Fig. 1.2). There are numerous limitations to the ChIP-chip technology, including antibody quality, the expense of covering the entire genome comprehensively, and hence the bias in choosing the regions to be examined. Additionally, ChIP is not quantitative. It is difficult to ascribe occupancy measurements to ChIP. For example, for any given gene, it is currently not feasible to determine if all or only a small percentage of cells bearing that gene contain the bound protein or histone modification. In cases where two modifications or binding events are present at the same location, it is possible to re-ChIP a sample and determine if it contains the other modification. This manipulation is rarely performed in genome-wide studies. Additionally, ChIP is largely descriptive and does not yield mechanistic information. Nevertheless, the information generated can be revealing and allows an investigator to pose testable hypotheses. A more recent technique involves ChIP followed by high-throughput DNA sequencing (ChIPSeq, or ChIP Sequencing) of the precipitated DNA in microfluidic devices (Collas and Dahl 2008). This approach has the advantage of being unbiased, but it is limited by antibody quality and full genome coverage, which is still difficult but attainable. This technology and the ability to increase sequence coverage are rapidly improving. The technologies allow simple assays such as ChIP of transcription factors and histone modifications to be examined on a genome-wide basis. The ChIP-chip and ChIP-Seq techniques can be adapted for studying DNA methylation (Zilberman and Henikoff 2007), DNase I hypersensitivity (Crawford et al. 2006a,b; Xi et al. 2007), chromatin conformation capture (3C) (Simonis et al. 2007), and RNA Trap (Chakalova et al. 2004). DNA methylation can be measured by precipitating chromatin with an antibody to methyl CpG. DNase I hypersensitivity is a technique that measures accessibility of chromatin to DNase I. Open chromatin, characteristic of enhancers, promoters, and insulators, is more susceptible to cleavage. The DNase I–cleaved DNA is analyzed by Southern blotting for individual genes or ligated to adapter-primers, which are used to amplify the DNA for sequencing or ChIP. 3C is a technique in which formalehyde-crosslinked protein/DNA complexes are digested with restriction enzymes and diluted. DNA ligase is then added. If a protein has crosslinked to two distal pieces of DNA (i.e., an enhancer and promoter), then the two DNA fragments will be ligated because they are in proximity (i.e., crosslinked to the same protein). Specific primers or ligated adapters are then used to amplify the DNA fragment. The DNA fragment can be hybridized to a chip or sequenced. In the RNA Trap method, currently used by only a few laboratories, an oligonucleotide probe is designed complementary to an RNA molecule representing a region near the transcription start site (TSS). The probe is hybridized to RNAs on isolated chromatin containing elongating Pol II. The oligonucleotide probe is attached to digoxygenin, which binds antibody conjugated to horseradish peroxidase. Biotin tyramide is added, and the peroxidase generates free radicals that link biotin to nearby proteins. After fragmentation of the DNA, the biotinylated protein and DNA can be isolated on a streptavidin-conjugated resin. The captured DNA fragments can be amplified and sequenced, or hybridized to chips. Thus, if two distal pieces of DNA were in proximity to the 5′ end of the RNA, they will be isolated in the procedure. Another technique discussed below is FISH. In this method, which is usually performed on fixed nuclei or cells, a fluorescently tagged nucleotide probe is hybridized to RNA or DNA and fluorescence is monitored using a confocal microscope. FISH can be combined with immunofluorescence 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 4 4 / Chapter 1 A B FIGURE 1.2. Genome-wide ChIP-chip. In a typical experiment, DNA from a ChIP reaction is amplified by PCR, labeled with a fluorophore, and hybridized to microarrays. (A) A scan of an Agilent promoter microarray hybridized with a fluorescent ChIPed DNA probe from adenovirus-infected cells. The inset shows a magnification of a small region of the chip. These arrays consist of two slides, together containing probes for 17,054 annotated human promoters, covering a region from –5.5 to +2.5 kb relative to the known annotated transcription start sites (TSSs) of each promoter. Each microarray slide contains 244,000 oligonucleotide probes that are 60 bp long. The fluorescent intensity of each spot reflects the level of enrichment of the corresponding DNA fragment in a ChIP sample. The data obtained from the scan of the chip are normalized and analyzed by various statistical methods including clustering algorithms (see below). The illustration is in grayscale although typically the arrays are illustrated in green and red. (B) This panel illustrates heat maps in grayscale, although these are usually shown in green and red or yellow and blue. Heat maps are used as a visual representation of the level of enrichment at each locus according to a given scale. By convention, increasing shades of red or yellow reflect higher enrichment values, whereas increasing shades of green or blue reflect relative depletion. In the example shown, the data are arranged by K-means clustering into three groups based on the gene expression data and gene ontogeny classification (http://david.abcc.ncifcrf.gov/). The rows along the vertical axis each represent a separate gene and the columns illustrate the ChIP signal according to the scale bar at lower right. Each column represents a 500-bp interval along the length of the promoter from –5.5 to +2.5 kb. In the experiment shown, cells were infected with adenovirus and after 24 hours the mRNA levels were examined by gene expression microarrays, and the levels of H3K18 acetylation, H3K9 acetylation, and binding of the adenovirus small E1A (e1a) protein were examined by ChIP-chip. The panels illustrate from left to right gene expression data (Exp), the levels of acetylation at H3K18 and H3K9 upon adenovirus infection, and binding of E1A. Note the binding of E1A occurs predominantly around the TSS but its affect on histone acetylation extends upstream from the TSS and downstream to the transcribed region of genes. (A: Rendered by R. Ferrari; B: adapted, with permission of AAAS, from Ferrari et al. 2008.) 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 5 A Primer on Transcriptional Regulation in Mammalian Cells / 5 to compare locations of DNA and RNA relative to various protein-containing structures. This technique is used to localize chromosome territories and to identify the location of an mRNA relative to RNA Pol II–containing foci, called transcription factories, and can in some instances be used to study the location of a DNA element, such as a locus control region (LCR), relative to distally located actively transcribing genes. CHROMATIN AND THE GENERAL TRANSCRIPTION MACHINERY Chromatin Structure and Organization Chromatin condensation allows 2 meters of DNA (2 nm diameter, 3.4 nm per 10.4 bp turn) to fit into a 5–20-µm-diameter nucleus. FISH analyses of nuclei have revealed that in a typical mammalian interphase nucleus, each chromosome occupies a distinct territory called a CT (Branco and Pombo 2006; Cremer et al. 2006). The chromosomes within these CTs are assembled from the basic 10-nm-diameter nucleosomal fiber (beads on a string), with each nucleosome containing a core histone octamer (containing two molecules each of histones H2A, H2B, H3, and H4) and occasionally a linker histone, H1 or H5 (Horn and Peterson 2002; Luger and Hansen 2005; Woodcock 2006). The 10-nm nucleosome fiber is assembled into a higher-order structure referred to as the 30-nm fiber (Tremethick 2007; Wu et al. 2007), and the 30-nm fibers coil into 80–100nm chromonema fibers (Horn and Peterson 2002). It is likely that several additional levels of organization exist, including multiple forms of chromatin loops within and between chromosome territories, heterochromatic regions, and transcription factories (Kadauke and Blobel 2008). In addition to the histones, several abundant nonhistone proteins (NHPs), including various highmobility group (HMGA, HMGB, and HMGC) proteins (Hock et al. 2007), are commonly associated with chromatin and contribute to its organization along with the nuclear matrix attachment sites (Linnemann et al. 2007). The nuclear matrix itself is an insoluble group of proteins constituting a lattice of fibers attached to the nuclear lamina. The initial crystal structure of the nucleosome at high resolution was solved in 1997, and higher-resolution structures have since appeared along with structures containing several histone variants (see below) (Chakravarthy et al. 2005). Additionally, the structure of a tetranucleosome has been solved, providing insight into how higher-order nucleosome packing may occur (Schalch et al. 2005; Tremethick 2007). A typical nucleosome contains 146–147 bp of DNA wrapped 1.65 times in a toroidal supercoil around the histone octamer. The structure reveals that the nucleosome has a disk-like shape with dimensions approximating 6 x 10 nm. The octamer can be subdivided into a heterotetramer of H3/H4 and two H2A/H2B heterodimers (Fig. 1.3). Each histone contains a characteristic motif termed the histone fold. This motif comprises a long α-helix linked at either end to two shorter α-helices via short β-turns. The histones insert arginines into all 14 of the minor grooves of the DNA facing the surface of the octamer in a mononucleosome. The estimated number of specific contact points is about 120. These contacts provide most of the energy that drives the DNA–octamer interactions. Different DNA sequences have different propensities to fold into nucleosomes on the basis of deformability of individual dinucleotide base steps (Richmond 2006; Segal et al. 2006). Paradoxically, biophysical studies have shown that the DNA–protein interactions within nucleosomes are dynamic. At physiological salt concentrations, the entry and the exit points of the DNA within the nucleosome appear to transiently dissociate from the surface of the octamer even on a high-affinity nucleosome positioning sequence such as the artificial 601 sequence (Mihardja et al. 2006; Hoch et al. 2007). In addition, on lower-affinity sequences, the nucleosomes can slide along the DNA, an effect that can be facilitated by histone chaperones such as NAP1 (Park and Luger 2006) and by adenosine triphosphate (ATP)-dependent remodeling machines (Cairns 2005; Saha et al. 2006b). The linker histones H1 and H5 contain a winged helix domain flanked by an unstructured domain on the amino terminus, and a highly basic but unstructured region on the carboxyl terminus. The amino terminus is 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 6 6 / Chapter 1 H3 K27 K4 K14 K9 H3 K16 K20 H4 H3 H3 FIGURE 1.3. Two views of the nucleosome. The nucleosome was rendered using PyMOL and the PDB file 1KX5. The tails were artificially extended away from the nucleosome core. (Dark gray) H3/H4 tetramer; (light gray) H2A/H2B dimers. The following tail residues are highlighted: lysines 4, 9, 14, and 27 on H3 and lysines 16 and 20 on H4. Note that the H4 tail is shorter than H3. Thus, residues 16 and 20 are very close to the initial H4 α-helix and close to the DNA. (Rendered by J. Heiss, UCLA.) phosphorylated, and both amino and carboxyl termini, as well as the winged helix domain, are modified posttranslationally at several positions (Woodcock et al. 2006; Wisniewski et al. 2007). The linker histone is believed to bind the DNA at the entry and exit points of the nucleosome, but there are some questions about whether this is always true. Binding of the linker histone restricts the mobility of nucleosomes. All four histones contain lysine-rich amino-terminal tails. Additionally, the H2A histone has a similar lysine- and serine-rich carboxy-terminal tail that is not well-characterized. The histone tails extrude from the surface of the nucleosomes and do not play a major role in organizing or stabilizing nucleosomal DNA. However, these tails are subject to numerous posttranslational modifications that act as part of a signaling pathway controlling chromatin metabolism during transcription, replication, and repair. Some of the tails stabilize internucleosomal DNA interactions. For example, the H4 tail interacts with the exposed surface of an H2A/H2B dimer within the adjacent nucleosome. This surface is formed primarily with seven acidic residues from H2A. Evidence shows that the H4 tail is necessary for the formation of the 30-nm fiber, as acetylation of the H4 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 7 A Primer on Transcriptional Regulation in Mammalian Cells / 7 tail at K16 causes neutralization of the lysine basic charge, which loosens the contact and prevents formation of the 30-nm fiber (Shogren-Knaak et al. 2006). There are a number of histone variants that diverge from the major families of the histones (Jin et al. 2005; Bernstein and Hake 2006; Hake and Allis 2006; Workman 2006; Eirin-Lopez and Ausio 2007; Kusch and Workman 2007; Loyola and Almouzni 2007). These variants include H2AZ, H2AX, macroH2A, H2A.Bbd, H3.3, and CenH3. H2AZ is found in repressed regions of the genome in its unacetylated form and active regions (i.e., promoters) when acetylated. Phosphorylated H2AX occurs at DNA-damage sites. macroH2A is enriched in the female inactive X chromosome. H2A.Bbd is located in transcriptionally active regions of the chromosome and excluded from inactive regions (i.e., excluded from inactive X; Bbd denotes Barr body–deficient). H3.3 is incorporated during active transcription within the gene. CenH3 or CenpA is located within centromeric heterochromatin. Much can still be learned about the properties of these alternate histones, but their function is probably related to their localization. For example, histone variant H2AZ (~60% identity to H2A) is inserted into nucleosomes flanking S. cerevisiae and mammalian promoters by the SWR1 ATP remodeling enzyme (Korber and Horz 2004). H2AZ-containing nucleosomes contain an extended acidic patch that enhances formation of 30-nm fibers (McBryant et al. 2006; Tremethick 2007), but H2AZ-containing nucleosomes may be less stable than nucleosomes with H2A (Zhang et al. 2005). Nevertheless, structural analysis reveals that H2AZ is accommodated into the nucleosome without any significant distortions or loss of DNA contact (Chakravarthy et al. 2005). The structure of the 30-nm filament is currently the subject of much study because correlative evidence strongly suggests that genes within the filament must be “decondensed” to become competent for transcription. In the earlier electron microscopy (EM) studies on the 30-nm filament, the individual nucleosomes were believed to be in a solenoidal configuration, folding into what is referred to as a one-start helix. Evidence for this structure includes EM and crosslinking data. In contrast, more recent high-resolution data suggest an alternative form, termed the zigzag or twostart helix (Fig. 1.4) (Tremethick 2007; Wu et al. 2007). In this model, nucleosomes zigzag between each other, forming two rows with the linker DNA crossing between the two. For example, the first and third nucleosomes would be stacked in one row and attached by the linker DNAs to a second row, where the second and fourth nucleosomes are stacked. The crystal structure of a tetranucleosome formed in high Mg++ concentrations reveals two stacks of nucleosomes with the linker DNA zigzagging between them, suggesting a relationship to the 30-nm fiber, although modeling suggests it would not be 30 nm in diameter. Nucleosomes bearing H1 formed similar zigzag-type structures shown by EM. Incorporation of linker histones can dramatically stabilize 30-nm filaments. Finally, crosslinking data on short nucleosomal arrays support the zigzag structure. Solenoid one-start helix Zigzag two-start helix N7 N3 N2 N6 N3 N2 Nucleosome 1 (N1) N4 FIGURE 1.4. Models for structure of the 30-nm filament. Two models for the 30-nm filament are shown: the solenoidal one-start helix and the zigzag two-start helix. There are currently data in support of both models. (Adapted, with permission of Elsevier, from Trementhick 2007.) 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 8 8 / Chapter 1 Chromatin Modification Modification of the histone tails is a major form of regulation during transcriptional activation and silencing (Kouzarides 2007). The tails, and in some cases the globular domains of histones, are subject to at least eight types of covalent marks: acetylation, methylation (lysine mono-me1, di-me2, and tri-me3; arginine mono and di; di can be symmetric and asymmetric), phosphorylation, ubiquitinylation, sumoylation, adenosine diphosphate (ADP)-ribosylation, deimination, and proline isomerization. These marks have context-dependent effects and enzymes exist to remove the modifications including histone deacetylases (HDACs), histone demethylases, and deubiquitinylases. Therefore, modifications are dynamic and are removed when a specific process must be reversed (i.e., during gene activation and repression/silencing). The enzymes catalyzing modifications function within different regulatory contexts and are involved in all aspects of chromatin metabolism. Dozens of modifications, some in intricate combinations, have been detected by mass spectrometry and ChIP. Chromatin modifications have two global roles. One role is to create binding sites for proteins containing a recognition domain for a particular modification. The protein recruited to the modification then elicits an effect. Several examples of these effects are discussed below. Alternatively, some modifications affect chromatin compaction as discussed above for H4K16 acetylation. Certain mutations in histones such as the Sin (SWI/SNF-independent) mutants also affect nucleosome stability and chromatin compaction. Table 1.1 lists the enzymes responsible for various modifications and Table 1.2 lists the types of domains involved in binding these modifications. A subset of these will be discussed in the sections below. Acetylation Acetylation is carried out by histone acetyltransferases (HATs) using acetyl–coenzyme A (CoA) as the donor for the transfer of acetyl groups onto lysine (Dyda et al. 2000; Peterson and Laniel 2004; Lee and Workman 2007). In this reaction, the positively charged NH3+ amino group is replaced by the uncharged NH-CO-CH3. A cursory examination of Table 1.1 reveals a striking observation about the specificity of various modifying enzymes. Many HATs are promiscuous and target numerous lysine residues, sometimes in different histones, for acetylation. In mammalian cells, for example, the highly related p300 and CREB-binding (CBP) proteins (generally termed p300/CBP), as discussed further in Chapter 10, modify residues on all four histone tails. The related mammalian GCN5 and PCAF enzymes, and S. cerevisiae Gcn5 (found in SAGA), modify several residues on H3. Note that several of the residues modified by p300/CBP and GCN5/PCAF in mammalian cells are found in active promoters as measured by individual studies and genomewide analyses as discussed below (Wang et al. 2008). A specialized domain of known structure, termed the bromodomain, recognizes the acetylation mark (Mujtaba et al. 2007). Bromodomains are found in the subunits of numerous HATs and ATPdependent remodeling complexes. Conversely, there are numerous bromodomain-containing factors that are not subunits of ATP-dependent machines, including bromodomain factor 1 (BDF1) in S. cerevisiae, which is believed to correspond to a missing segment found in the human (but not S. cerevisiae) TATA box–binding protein-associated factor 1 (TAF1; see below). Additionally, bromodomain protein 2 (BRD2) and BRD3 in mammalian cells bind to H4K5ac, H4K12ac, and H3K14ac and stimulate transcription of Pol II through chromatin in vitro, possibly by acting as histone chaperones (LeRoy et al. 2008). Methylation Histone methylation occurs on lysines and arginines (Cheng et al. 2005; Martin and Zhang 2005, 2007; Shilatifard 2006). On lysine, the methyltransferase (methylase) catalyzes the transfer of methyl group from S-adenosylmethionine onto lysine, replacing one, two, or three of the hydrogens on the NH3+ amino group (mono-, di-, and trimethylation). In contrast to acetylases, most histone methylases and the enzymes that remove methyl groups, the demethylases, are exquisitely specific. Six differ- 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/24/08 11:40 AM Page 9 A Primer on Transcriptional Regulation in Mammalian Cells / 9 TABLE 1.1 Covalent modification of chromatin Histone-modifying enzymes Acetyltransferase HAT1 CBP/P300 PCAF/GCN5 TIP60 HB01 (ScESA1, SpMST1) ScSAS3 ScSAS2 (SpMST2) ScRTT109 Lysine methyltransferase SUV39H1 SUV39H2 G9a ESET/SETDB1 EuHMTase/GLP CLL8 SpClr4 MLL1 MLL2 MLL3 MLL4 MLL5 SET1A SET1B ASH1 Sc/Sp SET1 SET2 (Sc/Sp SET2) NSD1 SYMD2 DOT1 Sc/Sp DOT1 Pr-SET 7/8 SUV4 20H1 Residues modified H4 (K5, K12) H3 (K14, K18) H4 (K5, K8) H2A (K5) H2B (K12, K15) H3 (K9, K14, K18) H4 (K5, K8, K12, K16) H3 K14 H4 (K5, K8, K12) H3 (K14, K23) H4 K16 H3 K56 Histone-modifying enzymes SUV4 20H2 SpSet 9 EZH2 RIZ1 Lysine demethylases LSD1/BHC110 JHDM1a JHDM1b JHDM2a JHDM2b JMJD2A/JHDM3A JMJD2B JMJD2C/GASC1 JMJD2D UTX JMJD3 Arginine methlytransferases CARM1 PRMT4 PRMT5 Serine/threonine kinases Haspin MSK1 MSK2 RSK2 CKII Mst1 Ubiquitinylase Bmi1/Ring1A/Ring2B RNF20/RNF40 Proline isomerases ScFPR4 Residues modified H4K20 H4K20 H3K27 H3K9 H3K4 H3K36 H3K36 H3K9 H3K9 H3K9, H3K36 H3K9 H3K9, H3K36 H3K9 H3K27 H3K27 H3 (R2, R17, R26) H4R3 H3R8, H4R3 H3T3 H3S28, H3S10 H3S28, H3S10 H3S10 H4S1 H2BS14 H2AK119 H2BK120 H3P30, H3P38 H3K9 H3K9 H3K9 H3K9 H3K9 H3K9 H3K9 H3K4 H3K4 H3K4 H3K4 H3K4 H3K4 H3K4 H3K4 H3K4 H3K36 H3K36 H3K36 H3K79 H3K79 H4K20 H4K20 Adapted, with permission of Elsevier, from Kouzarides (2007). Only enzymes with specificity for one or a few sites have been included, along with the sites they modify. Human and yeast enzymes are shown. The yeast enzymes are distinguished by the prefix Sc (Saccharomyces cerevisiae) or Sp (Schizosaccharomyces pombe). Enzymes that fall within the same family are grouped. Note that a new nomenclature has been agreed on for naming chromatin-modifying enzymes (Allis et al. 2007). ent lysines are mono-, di-, and trimethylated in mammals, including H3K4, H3K9, H3K27, H3K36, H3K79, and H4K20. In mammalian cells, SUV39H1 and SUV39H2 trimethylate H3K9, whereas G9a dimethylates H3K9 during the process of pericentric heterochromatin formation and euchromatic gene silencing, respectively. Other methylases involved in silencing include the SUV4-20H1 and H2 enzymes, which catalyze H4K20 trimethylation in heterochromatin, and the EZH2 enzyme, which catalyzes H3K27 trimethylation in facultative heterochromatin (i.e., over inactive Hox genes). The MLL1-5 family members and SET1A and SET1B are highly specific for catalyzing di- and trimethylation of H3K4 during the early stages of transcription initiation and elongation (Eissenberg and Shilatifard 2006; Shilatifard 2008). Similarly, during gene activation and elongation, H3K36 is methylated by SET2 associated with Pol II, and H3K79 is modified by DOT1. H3K36me3 appears to be localized downstream from H3K4me3, whereas H3K79me3 is localized throughout the gene. Except for DOT1, the other methylases are part of the SET domain family of methylases and share a structurally conserved catalytic domain (Li et al. 2007a). Arginine is methylated by protein arginine 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 10 10 / Chapter 1 TABLE 1.2. Modification recognition domains Reader module Royal Bromodomain Chromodomain Double chromodomain Chromo barrel Tudor Double/tandem tudor MBT PHD finger WD40 repeat 14-3-3 BRCT PTM mark Many histone Kac, (Kac) H3K9me2/3, H3K27me2/3 H3K4me1/2/3 H3K36me2/3 (Rme2s) H3K4me3, H4K20me3 H4K20me1/2, (Kme2) H4K20me1/2, H1K26me1/2 H3K4me1, H3K9me1/2 H3K4me3, H3K4me0 H3K9me3, H3K36me3 H3R2/K4me2, (R, Sph, Tph) H3S10ph, H3S28ph, (Sph, Tph) H2AX-S139ph, (Sph, Tph) Adapted, with permission of Macmillan Publishers Ltd, from Taverna et al. (2007). Shown grouped by domain family are known chromatin-associated modules and the histone marks they have been reported to bind. Parentheses denote examples where structural information is known about related family members and their interactions with nonhistone PTMs (posttranslational modifications). (R) methylases (PRMTs). These enzymes modify H3R2, R8, R17, R26, and H4R3 and possibly other positions (Cheng et al. 2005; Kouzarides 2007). Several types of domains recognize methylated histone tails (Taverna et al. 2007). The domains include members of the Royal family. These include the chromodomain, the double chromodomain, the chromo barrel domain, the Tudor domain, the double Tudor domain, and the malignant brain tumor (MBT) domain. Plant homeodomain (PHD) and WD40 repeats also are capable of binding specific methylated lysines or arginines. Proteins bearing these domains are targeted to the various modifications. For example, the chromodomain of HP1 directs the protein to bind H3K9me2 and H3K9me3, whereas the chromodomain of Polycomb protein (Pc or in humans HPC) binds H3K27me3. These and other examples are discussed in the ensuing sections. Demethylation is achieved by two classes of demethylases termed the jumonji and the amine oxidase families, the latter of which is represented by lysine-specific demethylase 1 (LSD1) (Anand and Marmorstein 2007; Shi and Whetstine 2007; Lan et al. 2008). Demethylases can be mono-, di-, or trimethyl-specific, although LSD1 is only capable of demethylation of di- and monomethylated histones because of its chemical cleavage mechanism. Demethylases are slightly more promiscuous than methylases and sometimes are able to demethylate two different residues. For example, the jumonji demethylases JMJD2A–D demethylate H3K9me2/me3 and H3K36me2/me3. A key issue in the field is how these and other histone-modification enzymes are recruited to genes. UTX and JMJD3 are H3K27me2 and H3K27me3 demethylases. These proteins are components of a complex of proteins containing the MLL2 and MLL3 H3K4 methylases (Lan et al. 2008). Hence, a complex bearing a methylase associated with gene activation contains an enzyme that removes a repressive mark. Similarly, the CoREST and CtBP corepressor complexes described below contain LSD1 as a subunit along with an HDAC. In this case, a complex recruited by a sequencespecific repressor contains at least two enzymes capable of removing modifications associated with transcription activation. Monoubiquitinylation Histone monoubiquitinylation (Ub) on lysines is another modification involved in gene regulation (Shilatifard 2006). It is associated with both gene activation and silencing. For example, 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 11 A Primer on Transcriptional Regulation in Mammalian Cells / 11 H2AK119 is ubiquitinylated by the Ring1B/Ring1A/Bmi-associated subunits of polycomb repressive complex 1 (PRC1) that bear the chromodomain-containing Pc or HPC protein (see below). In this case, one mark, H3K27me3, can recruit a complex that adds a second repressive mark involved in silencing. In contrast, H2BK120Ub (or H2B123Ub in S. cerevisiae) is catalyzed by the RNF20/40 and Ubc6 proteins in humans and by the Rad6/Bre1 complex in S. cerevisiae. As described below, this modification is somehow necessary for trimethylation of H3K4 during the initial phases of elongation in S. cerevisiae. H2B120Ub strongly stimulates Pol II elongation in vitro in mammalian chromatin transcription systems (Pavri et al. 2006). Paradoxically, in S. cerevisiae, the SAGA complex, bearing the Gcn5 HAT, contains a subunit termed Ubp8, which deubiquitinylates H2BK123. Because both the ubiquitinylase and deubiquitinylase are associated with gene activation, this would suggest that a series of sequential events are necessary for activation (Shilatifard 2006). Phosphorylation Histone phosphorylation also contributes to gene activation (Kouzarides 2007). Phosphorylation of H3S10 is a mark associated with mitosis and with activation of many rapid-response genes in mammals. The MSK1/2 and RSK2 kinases phosphorylate H3S10 in mammals; Snf1 does this in S. cerevisiae. H2AX S139 is phosphorylated by the ATM kinase in mammalian cells during DNA damage and generates what is referred to as γH2AX. Various 14-3-3 proteins recognize H3S10P, and γH2AX (phosphorylated) is recognized by the tandem BRCT1 domains of Mediator of DNAdamage checkpoint protein 1 (MDC1). In another example, the Aurora B kinase phosphorylates H3S10 and H3S28 during mitosis. In addition to recruiting other proteins, H3S10P can have an effect on its own. H3S10P is thought to be particularly important because it can block binding of the HP1 protein to H3K9me3. Thus, an activating mark alone can antagonize the binding of a nearby repressor. Chromatin Remodeling The mobility of nucleosomes on DNA and the eviction of histones are dependent on a class of ATP-dependent factors known as chromatin-remodeling enzymes (Cairns 2005; Saha et al. 2006a,b). This class of enzyme has diverse roles in chromatin metabolism and gene regulation, and, not surprisingly, several families of proteins have evolved. These families appear to have related mechanisms based on the structural similarities of the ATPase domain to DNA translocases. However, their functions have become highly specialized in part because of differences in the associated subunits and in part because of subtle differences in the function of the ATPase domain. There are currently five families: SWI/SNF (BAF), ISWI (imitation switch), INO80, NURD/Mi-2CHD1-9, and SWR1 (Table 1.3). These enzymes and their associated subunits use the energy of ATP to slide or evict histone octamers. Remember that the octamer contacts 14 of the DNA minor grooves via approximately 120 specific interactions. The energy required to break these contacts simultaneously, which has been estimated in the 12–14-kcal range for an octamer (Kd ~ 10–10), requires ATP hydrolysis. The SWI/SNF family includes the S. cerevisiae SWI/SNF and RSC, and two major forms of mammalian SWI/SNF termed human brahma-related gene–1 associated factors (BAFs) and PBAF (Mohrmann and Verrijzer 2005). S. cerevisiae SWI/SNF and RSC contain bromodomains that recognize acetylated lysines in histone tails. These bromodomains increase the binding affinity of RSC and SWI/SNF for chromatin and enhance their ability to remodel nucleosomes. In all SWI/SNF family members, the ATPase subunit contains one bromodomain located near the carboxyl terminus of the protein. However, additional subunits can contribute additional bromodomains. For example, RSC contains 8 of the 15 known S. cerevisiae bromodomains. Most of the domains are contributed by subunits other than the catalytic subunit Sth1. Mammals have several variants of the SWI/SNF enzyme (Mohrmann and Verrijzer 2005). The ATPase is contained within one of two related proteins termed brahma-related gene 1 (Brg1) or 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 12 12 / Chapter 1 TABLE 1.3. ATP remodeling machines and their subunits INO80 Species Complex Homologous subunitsa Yeast INO80 Ino80 Rvb1, Rvb2 Arp4,5,8, Act1 Taf14 Ies2 Ies6 Human INO80 hIno80 Tip49a, Tip49b BAF53a, Arp5,8 hIes2 hIes6 SWR1 Yeast SWR1 Swr1 Rvb1, Rvb2 Arp4,6, Act1 Yaf9 Human SRCAP SRCAP Tip49a, Tip49b BAF53a, Arp6 GAS41 Unique subunits Ies1, Ies3-5, Nhp10 Amida, NFRKB, MCRS1, FLJ90652, FLJ20309 Swc4/Eaf2 Swc2/Vps72 Bdf1 H2AZ, H2B Swc6/Vps71 Swc3,5,7 DMAP1 YL-1 H2AZ, H2B ZnF-HIT1 ISWI Subfamily Species Complex Homologous subunitsa Unique subunits ACF/CHRAC Yeast ISW1a Isw1 Ioc3 Yeast ISW1b Isw1 Ioc2, Ioc4 Yeast ISW2b Isw2 Itc1 Human ACFc hSNF2H WCRF180/hACF1 NURF Human NURF hSNF2L BPTF RbAP46, RbAP48 Mi-2/CHD Subfamily Species Complex Homologous subunitsa Unique subunits CHD1 Yeast CHD1 Chd1 Human CHD1 CHD1 Mi-2/CHD Human NuRDd Mi-2α/CHD3, Mi-2β/CHD4 MBD3; MTA1,2,3; HDAC1,2; RbAp46,48; p66α,β; DOC-1 SWI/SNF Species Complex Homologous subunitsa Yeast SWI/SNF Swi2/Snf2 Swi1/Adr6 Yeast RSC Sth1 Rsc1e Rsc2 Rsc4 Rsc8 Rsc6 Arp7, Arp9 Sfh1 Human BAFe BRG1 or hBRM BAF250/hOSA1 Human PBAF BRG1 BAF180 Swi3 Swp73 Arp7, Arp9 Snf5 BAF60a BAF53 hSNF5 BAF57 β-actin BAF155, BAF170 BAF60a BAF53 hSNF5 BAF57 β-actin Unique subunits Swp82, Taf14, Snf6, Snf11 Rsc3,5,7,9, 10,30, Htl1, Ldb7, and Rtt102 Adapted, with permission of Elsevier, from Bao and Shen (2007b). a First row indicates core ATPase. b A variant ISWI2 complex also contains Dls1 and Dpb4 in addition to Isw2 and Itc1. c Human CHRAC also contains hCHRAC15, hCHRAC17 in addition to hSNF2H and hACF1. hSNF2H has also been found in other human ISWI complexes: RSF, WICH, B-WICH, and SNF2h/NuRD/cohesion complexes. d Human MeCP1 complex contains MBD2 instead of MBD3; MEP50 and PRMT5 are also found in human MeCP1. e PBAF complex contains BAF180 and BAF200 rather than BAF250. 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 13 A Primer on Transcriptional Regulation in Mammalian Cells / 13 Brm1 (mammalian brahma). Mammals contain two SWI/SNF complexes termed BAF and PBAF. BAF utilizes either Brg1 or Brm1 as the ATPase, whereas PBAF uses Brg1. BAF and PBAF share numerous ancillary subunits and differ in that PBAF also contains at least one novel subunit termed BAF180, while the BAF complex contains a unique subunit termed BAF250. BAF180 or polybromo contains six bromodomains, which are thought to be somewhat conserved with various bromodomains in RSC. Hence, human PBAF is thought to be related to S. cerevisiae RSC. BAF250 contains a unique AT-rich interaction domain (ARID) that may bind DNA. Such a domain is found in SWI1, which is unique to the SWI/SNF complex in S. cerevisiae. Thus, BAF is thought to resemble SWI/SNF more closely than RSC. PBAP and BAP are the Drosophila melanogaster complexes similar to PBAF and BAF in humans. Unlike humans, D. melanogaster only contains a single ATPase named Brahma or Brm. Note that all of the ATP-dependent remodeling machines contain actin and actin-related proteins, or ARPs. The specific functions of the ARPs are unknown, although they do not appear to polymerize or hydrolyze ATP (Chen and Shen 2007). Several studies suggest that ARP subunits play a role in assembly of the remodeling complexes. ARP4 is found in many higher eukaryotic chromatin remodeling and modifying complexes in stoichiometric amounts with actin. In S. cerevisiae, ARP7 and ARP9 apparently substitute for ARP4 and actin in SWI/SNF and RSC. It has been hypothesized that actin polymerization may control movement of SWI/SNF, either alone or attached to chromatin, through the nucleus. For example, one could imagine a model in which actin would attach to SWI/SNF and move chromatin into a transcription factory. Another interesting feature of the SWI/SNF family is that it generates disordered nucleosomal arrays and can transfer H2A-H2B dimers or even the octamer to another DNA molecule or chaperone in vitro (i.e., eviction). SWI/SNF family members have been shown by genome-wide studies to bind overlapping genes. Genetic analysis, however, suggests the enzymes play nonoverlapping roles in vivo. SWI/SNF family members have been implicated in gene activation in vivo and can facilitate Pol II elongation in biochemical systems. SWI/SNF family members can also directly bind some activators (Peterson and Workman 2000). The EM structures of SWI/SNF and RSC have been solved, revealing a central cavity into which a nucleosome can be modeled. This result suggests that the ATPase complex envelopes the nucleosome as part of the remodeling mechanism (Smith et al. 2003; Leschziner et al. 2007). The current models also show that RSC seems to completely envelope the nucleosome (Fig. 1.5), whereas SWI/SNF only interacts with one gyre of DNA along one face of the nucleosome (Dechassa et al. 2008). Models for the human SWI/SNF more closely mimic the RSC model. A variety of data suggest that the enzymes directly contact DNA two turns from the nucleosome dyad and draw DNA toward the dyad, around the histone octamer, and out the other end, effectively moving the position of the octamer on the DNA (Saha et al. 2006a). A detailed model termed the wave-ratchet-wave model has been proposed (Saha et al. 2006a). The effect of the bromodomain–acetyl lysine interactions to the mechanism of remodeling is just beginning to be investigated, but early data suggest that the modifications not only increase affinity, but also alter the distribution of remodeled products and the ability of the enzyme to transfer the octamer to another DNA molecule (Ferreira et al. 2007). The ISWI chromatin remodeling enzymes share the ISWI ATPase, which does not contain a bromodomain, unlike SWI2 and STH1, the catalytic subunits of SWI/SNF and RSC, respectively. ISWI ATPases do, however, contain SANT and SLIDE domains, which help target the enzyme to chromatin (Mellor and Morillon 2004; Saha et al. 2006b). In S. cerevisiae and mammals, there are two ISWI homologs (Isw1 and Isw2 in S. cerevisiae). ISWI associates with different subunits in different complexes, including RSF, NURF, CHRAC, and ACF. In NURF and ACF, bromodomains are found in the ACF and NURF301 subunits, respectively. Unlike SWI/SNF, the ISWI family seems to generate regularly spaced or ordered nucleosomal arrays in vitro. In yeast, the Isw2 ATPase complex has been implicated in repression and is targeted by repressors such as Ume6 along with the Rpd3/Sin3 complex (described further below) (Fazzio et al. 2001). It also is necessary for the nucleosome positioning function of the Ssn6/Tup1 corepressor (Fazzio et al. 2001). Interestingly, the Isw2 complex is required to suppress antisense transcription emanating from the 3′ end of 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 14 14 / Chapter 1 a. RSC nucleosome DNA entry/exit points Histone H3 tail Translocase binding site Dyad axis DNA entry/exit points b. SWI/SNF nucleosome FIGURE 1.5. Model of RSC-nucleosome (a) and SWI/SNF-nucleosome (b) EM structures. The ATPase envelopes the nucleosome completely in RSC and partially in SWI/SNF. The location of the DNA exit–entry points, the nucleosome dyad, and the contact points with the DNA translocase are shown. (RSC-nucleosome adapted from Leschziner et al. 2007, © National Academy of Sciences, U.S.A., and SWI/SNF-nucleosome adapted, with permission of the American Society for Microbiology, from Dechassa 2008.) some yeast genes (Whitehouse et al. 2007). ISWI has also been implicated in transcriptional elongation in S. cerevisiae (Mellor and Morillon 2004). The CHD or Mi-2 family of remodeling proteins consists of nine family members in mammals (Hall and Georgel, 2007). These members are characterized by a duplicated chromodomain located immediately amino terminal to an SWI/SNF-like helicase domain. The chromodomain appears to recognize H3K4me3 with high affinity. CHD1 is thought to be a component of the SAGA and related SLIK HATs in S. cerevisiae and to coordinate with these HATs in gene activation. CHD1 slides nucleosomes in vitro and supports the assembly of ordered arrays. It can also assemble nucleosomes from free octamers and chaperones in vitro and has been reported to insert H3.3 during gene activation in vivo (Konev et al. 2007). CHD3 (Mi-2a) and CHD4 (Mi-2b) are the ATPase subunits of the NuRD complex and associate with HDACs 1 and 2 along with RbAP46/48. NuRD has been implicated in gene silencing and interacts with several sequence-specific repressors in vitro. CHD7 is identical to the D. melanogaster Kismet protein, which has been broadly linked to gene activation and transcription elongation. Kismet is classified with the Trithorax factors important for maintenance of the activated state of transcription during D. melanogaster development. The 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 15 A Primer on Transcriptional Regulation in Mammalian Cells / 15 Trithorax group of proteins also includes SWI/SNF and the MLL/SET1 complexes. CHD8 is known to be associated with CTCF and may participate in boundary functions (described below). The CHD family members are typically associated with numerous ancillary subunits with the possible exception of Kismet. The INO80 complex is found in organisms from S. cerevisiae to human (Bao and Shen 2007a). It is unusual in that its ATPase domain displays a large insertion much greater in size than that found in other ATPases. It was originally classified with SWR1 because of the similarity in the socalled split ATPase domain, but it has been recently classified differently based on the unusual properties of SWR1 relative to INO80. INO80 is tightly associated with the sequence-specific protein YY1 in mammalian cells, which helps target it to its binding sites in vivo (Cai et al. 2007). INO80 also slides nucleosomes in vitro. Unlike INO80, SWR1 has the ability to specifically transfer H2AZ into nucleosomes bearing H2A (Korber and Horz 2004). Genetic studies also indicate that Swr1 and Htz (the H2AZ gene) mutants have similar phenotypes. Additionally, gene expression and ChIP studies show that SWR1 regulates genes where H2AZ is inserted (Bao and Shen 2007a). The General Transcription Machinery Mammalian gene regulation involves an interplay between activators, repressors, the general transcription machinery, and chromatin. The general transcription machinery consists of RNA Pol II, the general transcription factors (GTFs) TFIIA, TFIIB, TFIID, TFIIE, TFIIF, and TFIIH, and a major coactivator termed the Mediator, which links activators to Pol II and the GTFs (see Tables 12.2, 12.3, and 12.4 in Chapter 12 describing TFIID, Mediator, and GTFs) (Hahn 2004; Roeder 2005; Kornberg 2007). TFIID is a multisubunit protein containing the TATA box–binding protein (TBP) and TBPassociated factors, or TAFs (Albright and Tjian 2000). The TAFs bear coactivator functions, although a key role may be to recognize different core promoter elements, as discussed below. The structures of the GTFs, Mediator, Pol II, and TFIID are known from either crystallography or EM studies. One important feature of Pol II is the heptapeptide repeat constituting the carboxyl terminal domain (CTD) of the largest Pol II subunit, Rpb1. The heptapeptide (Y1S2P3T4S5P6S7) is repeated 26 times in S. cerevisiae and 52 times in mammals. This CTD is phosphorylated extensively by different kinases involved in transcription regulation, and these events regulate initiation, elongation, and various mRNA-processing events (discussed below). Biochemical studies show that the GTFs support basal transcription and carry out many of the catalytic functions required for initiation. Coactivators such as Mediator and the TAF subunits of TFIID bridge the activators and GTFs. It should be pointed out that Mediator also enhances basal transcription in vitro. A family of TBP-related factors, TRFs, is also found in eukaryotes and can substitute for TFIID in certain contexts (Reina and Hernandez 2007; Torres-Padilla and Tora 2007). TRF1 is found in D. melanogaster, TRF2 in all metazoans, and TRF3 in vertebrates. TRF3 associates with TAF3, TFIIA, and TFIIB and can bind TATA-containing promoters and support transcription in vitro and in vivo. The TRF3/TAF3 complex is thought to replace TFIID and support activation of genes involved in myogenesis and perhaps other processes. Several in vivo targets are known for the various TRF factors in addition to myogenesis genes (Reina and Hernandez 2007). Be aware that most GTFs, TAFs, and Mediator subunits have multiple splice variants called isoforms. Indeed, when all of the combinatorial possibilities of the isoforms are considered, the actual protein composition of the general machinery can vary considerably. Therefore, we use the term “general machinery” throughout this chapter with the understanding that the reader is aware that PICs containing different isoforms may exist in parallel in the same cell. Organization of the Regulatory Region A typical mammalian gene can be subdivided into several regions that contribute to transcriptional activity including a core promoter, a proximal promoter, and enhancers (Fig. 1.6) (Smale and Kadonaga 2003; Maston et al. 2006; Yang et al. 2007). Other composite elements, like LCRs, 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 16 16 / Chapter 1 –50 kb Insulator Enhancer –2 kb –10 kb +50 kb Insulator Proximal promoter Core promoter TSS Gene –300 bp –50 bp –37 bp +32 bp 5’ UTR –37 to –32 –31 to –26 –2 to +4 +28 to +32 BRE TFIIB recognition element G G G CGC C CC A TATA Box Inr Initiator +1 T Drosophila T C A G T C T T Mammals Py PyAN A PyPy DPE Downstream core promoter element AG A CG A G T TC T AT A A A A G T A FIGURE 1.6. Promoter region of a regulated gene. Various core and upstream promoter elements are shown. Note that the position of enhancers, LCRs, and insulators varies considerably from gene to gene. (Bottom: Adapted, with permission, from Smale and Kadonaga 2003, © by Annual Reviews, www.annualreviews.org.) regulate groups of related genes in a temporal or cell-specific manner (Dean 2006). The enhancer and proximal promoter along with other DNA elements are thought to form a higher-order structure termed an “active chromatin hub” or ACH (de Laat and Grosveld 2003). Genes that are regulated similarly can also be found in dense nuclear foci containing Pol II and the general transcription factors. These foci are termed “transcription factories” (Sexton et al. 2007; Carter et al. 2008). There are up to 2000 such factories in a fibroblast and approximately 200 in smaller primary cells; the number of transcription factories appears to be related to nuclear volume. Active genes or colinear gene sets are flanked by boundary elements, which are thought to tether to the nuclear matrix or other nuclear structures such as nucleoli (Zhao and Dean 2005; Valenzuela and Kamakaka 2006; Dorman et al. 2007). Thus, when considering how genes are regulated, many factors should be taken into account beyond the simple linear organization of a gene or a locus. We briefly elaborate on the various elements constituting a regulated locus and discuss the types of chromatin modifications that are associated with each. The Core Promoter The core promoter binds Pol II and the general transcription machinery and is positioned approximately 35–40 bp on either side of the TSS (Smale and Kadonaga 2003; Maston et al. 2006). The core promoter can be subdivided into numerous distinct sequence elements. These elements include a IIB recognition element (BRE), which recognizes the general factor TFIIB; a TATA box, which recognizes the TBP subunit of TFIID (described in detail below); the initiator element (Inr), which recognizes TAF1 and TAF2 of TFIID and encompasses the start site of transcription; and the downstream promoter element (DPE), found at higher prevalence in D. melanogaster, which recognizes TAF6 and TAF9 of TFIID (discussed below). Other downstream elements found in some promoters include the downstream core element and motif ten element (MTE) (Maston et al. 2006). 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 17 A Primer on Transcriptional Regulation in Mammalian Cells / 17 Only a subset of these elements is present in any individual promoter. For example, based on the positions of the annotated TSSs in mammals, only about 25% of mammalian genes contain a consensus TATA or an AT-rich sequence resembling it (Yang et al. 2007). The current view is that TATA boxes are typically located in regulated genes such as those that are cell or developmentally specific or respond to signals. The vast majority of so-called ubiquitously expressed “housekeeping” genes (>60%) lack a TATA box; the majority of these contain a 0.5–2-kbp unmethylated CGrich region (CpG islands) upstream of a range of TSSs located within an approximately 100-bp window. Despite the propensity of housekeeping genes to contain CpG-rich regions, some regulated genes also are rich in CpG. A large percentage of genes contain an Inr element versus a TATA box, but this amount is still less than 50%. Even fewer genes contain a DPE or BRE (see Yang et al. 2007, and references therein). Whether these genes require or bind TFIID in the manner predicted by the crystal structure (see below) remains to be determined. Genome-wide studies have begun to reveal some general principles of promoter and enhancer organization. Genome-wide DNase I hypersensitivity analyses have shown that 22% of the hypersensitive sites are conserved between cell types (in a limited study covering six distinct cell lines) (Xi et al. 2007). However, among these, 86% are CpG-rich promoters as defined by gene expression data and the presence of an annotated TSS. These studies also showed that TAF1, and by implication TFIID, is bound to a significant portion of CpG-rich promoters in the absence of a TATA box (Kim et al. 2005). In a separate study by the same group, genome-wide ChIP studies on the ENCODE arrays identified promoters based on the binding of TAF1 and Pol II along with the presence of H3, H3K9/K14 diacetylation, H4K5/8/12/16 acetylation, H3K4me, H3K4me2, and H3K4me3 (see below) (Heintzman et al. 2007). On the basis of the various modification patterns, these authors identified four classes of promoters with increasing expression levels (P1–P4), which were related, on average, to the amount of TAF1, Pol II, and H3/H4 acetylation (Kim et al. 2005). As in S. cerevisiae, the level of H3K4me3 peaked at the promoter, with H3K4me2 and H3K4me1 increasing downstream from the TSS. There was a bimodal distribution of acetylation surrounding the TSS and a general depletion of H3 at the TSS, suggesting that the nucleosome at the TSS is removed and the flanking nucleosomes are acetylated. The amount of depletion was related to the activity of the promoter based on transcript abundance. These data agree well with earlier ChIP studies of individual promoters. Thus, these authors concluded that hyperacetylation, H3K4me3, TAF1, and Pol II are indicative of active promoters. However, note that it is somewhat unclear in published studies whether H3 depletion is found primarily at CpG island promoters or all promoters. A different study identified what was referred to as a “common modification module” consisting of 17 different chromatin marks found at 25% of active promoters in CD4+ T cells (Wang et al. 2008). A subset was found upstream of the TSS, and a different subset was found downstream. These authors subdivided their data into three expression classes from low to high (I, II, and III). Class I was enriched in H3K27me3 and H2AZ among others but no acetylation, and it included genes involved in development and differentiation. Class II consisted of H3K36me3 and the common modification module. Class III included the common modification module plus H4K16ac, H4K20me1, and H3K79me1/2/3. Class III seems to largely define the housekeeping genes. Highresolution mapping showed that many of the modifications appear to colocalize to the same positions, although it is not clear if this represents a population effect or whether the modifications indeed colocalize to a single nucleosome. The roles of some of these modifications are discussed above and below, although much remains to be learned about the specific functions of most. It has been proposed that coregulated genes involved in the same pathways have similar histone acetylation patterns (Kurdistani et al. 2004). The Proximal Promoter Immediately upstream of the core promoter, usually within the first 250–300 bp, is a cluster of binding sites for sequence-specific transcription factors that is called the proximal promoter (Maston et 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 18 18 / Chapter 1 al. 2006). The boundaries of the proximal promoter can generally be defined by examination of phylogenetically conserved sequences upstream of the core promoter and TSS (see Chapters 5 and 6). The proximal promoter can bind some of the same factors bound at distal enhancers or it can bind unique factors. Many studies on individual genes have shown that the proximal promoter has a low level of activity on its own, but it is greatly augmented by distal enhancers. Conversely, distal enhancers do not function in the absence of a proximal promoter. It is very likely that enhancers and proximal promoters share some common coactivator such as Mediator, which allows the loop between the two to form and brings the enhancer into proximity to the general machinery. Genome-wide ChIP-chip studies typically examine proximal promoters and fixed amounts of sequence upstream or downstream from the TSS due to the uncertainty of locating distal enhancers and incorporating them into the tiled chip arrays. Therefore, in many studies where sequence-specific transcription factors are being studied, be aware that they are typically examining proximal promoters. ChIP-Seq does not have this limitation. Studies on individual genes suggest that the proximal promoter plays a key role in recruitment of the general machinery to the core promoter. Enhancers Enhancers can be positioned many kilobases upstream of the gene, in introns, and occasionally after the 3′ end of the coding region (de Laat and Grosveld 2003; Zhao and Dean 2005; Maston et al. 2006). The enhancers bind combinations of sequence-specific activators, sometimes in a cooperative manner, to form an enhanceosome (discussed in Chapter 11). Many genes contain multiple enhancers that bind different factors, which allow the gene to be activated combinatorially in different cell types or in response to unique signals. Enhancers combine with proximal promoters to form the active chromatin hub as described above. Several models describe how enhancers communicate with proximal promoters. The observation that an enhancer requires a proximal promoter to function is probably a central facet of this interaction and very likely the organizing principle underlying the active chromatin hub. There are numerous examples of enhancers that do not communicate with nearby promoters of nontarget genes. One reason is that many activators undergo specific, cooperative protein–protein contacts, which may facilitate enhancer–promoter interactions. Thus, a proximal promoter may bind sequence-specific factors that are not compatible with the enhancer. Alternatively, coactivators that bridge enhancers and proximal promoters may be compatible with only a particular configuration of activators. Moreover, insulator elements regulate the interaction of enhancers and proximal promoters via looping/tethering (see above) and formation of restricted interaction domains. There is some limited knowledge of how distal enhancers communicate with proximal promoters to cooperatively recruit the general machinery. For example, 3C studies have shown that the androgen receptor (AR) loops from the enhancer of the prostate-specific antigen (PSA) gene, positioned 4.2 kb from the TSS, to the proximal promoter during gene activation (Wang et al. 2005). Notably, the proximal promoter also contains AR-binding sites. Additionally, ChIP studies on PSA show that Pol II and CBP are located at both the promoter and enhancer sites. A plausible explanation for this observation is that Pol II is crosslinking to the enhancer because the enhancer is in the vicinity of the Pol II–bound promoter. In other cases, there appears to be a facilitated tracking mechanism, where the chromatin is acetylated unidirectionally and the enhancer complex appears to move along the acetylated chromatin until it reaches the promoter (i.e., HNF4α, human β-globin) (Hatzis and Talianidis 2002; Zhao and Dean 2005). In other instances, this long-range tracking requires Pol II and TBP and intergenic transcription to facilitate tracking of the enhancer complex to the promoter. One good example of this type of tracking is the β-globin locus (Zhao and Dean 2005). Genome-wide studies have not generally found long tracks of acetylation between the enhancer and promoter of genes and instead have demonstrated isolated clusters of hyperacetylation suggesting that tracking may be limited to a few enhancer–promoter interactions (Roh et al. 2007). As 3C technology is combined with mechanistic approaches, the issue of enhancer–promoter communication is likely to become clearer. 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 19 A Primer on Transcriptional Regulation in Mammalian Cells / 19 Genome-wide studies seem to disagree on the types of histone modifications associated with enhancers. All enhancers that have been studied individually and genome-wide typically display DNase I hypersensitivity and hyperacetylation. However, genome-wide studies suggest enhancers are more complex. One study showed association of enhancers with DNase I hypersensitive sites, an enrichment of p300 and associated modifications, and a general depletion of H3. In contrast to promoters, enhancers were shown to be enriched in H3K4me1 but not H3K4me3 (Heintzman et al. 2007). TAF1, Med1, and RNA Pol II were found at some enhancers but were less enriched relative to the levels found at promoters. Another study agreed that enhancers correspond to DNase I hypersensitivity and are heavily acetylated, but the authors argued that there are a variety of different modification patterns associated with enhancers with some enriched in H3K4me2 and H3K4me3 (Barski et al. 2007). This discordance, particularly with respect to H3K4 methylation levels, suggests that different enhancers may use distinct mechanisms of action. It may also reflect the difficulty of identifying enhancers on a global scale in vivo. Locus Control Regions LCRs are a combination of multiple enhancers, insulators, and other elements that control groups of genes in a temporal or cell-specific manner (Dean 2006; Mahajan et al. 2007). LCRs comprise multiple distinct DNase I hypersensitive sites. They modulate transcription by influencing chromatin structure through an extended DNA region and appear to induce and maintain accessibility of that region to transcription factors as measured by DNase I hypersensitivity. LCRs were identified during studies of the β-globin locus in transgenic mice. β-globin transgenes were expressed at low levels, and the expression level was strongly influenced by the site of insertion into the chromosome. However, when the transgene contained a specific DNA fragment from the distal end of the globin locus, which is now known to encompass the LCR, high levels of positionindependent (or integration site–independent) expression were observed. The defining characteristics of the LCR are (1) the ability to confer high levels of induced expression on linked genes in transgene experiments, (2) expression levels directly linked to copy number, (3) the ability to enhance transcription in a chromatin setting, (4) the ability to open chromatin over long distances, and (5) the ability to confer position-independent expression even in the presence of centromeric heterochromatin, which generally inactivates integrated transgenes. In general, LCRs share properties with cell-specific enhancers in that they coincide with DNase I hypersensitive sites and bind to typical transcription factors. Furthermore, they stimulate hyperacetylation of surrounding chromatin. Despite these similarities, LCRs are clearly distinct from enhancers, because, in many instances, they enhance transcription only when integrated into a chromosome. Moreover, typical enhancers alone do not impart high levels of position-independent expression in transgenic mice. The current view, based on 3C and RNA Trap studies, is that the LCR loops out and interacts with the proximal promoter and enhancer of linked genes. The frequency and duration of contact between an LCR and a linked gene are believed to control the period of time during which the gene is transcriptionally active. Deletions in the LCR still retain minimal function, although they can diminish the activity and decrease position independence. The paradigm for LCRs is the β-globin locus, which has been extensively studied in multiple organisms (Fig. 1.7). In mouse, the β globin locus consists of four genes termed ε, βh1, β major, and β minor. The LCR is positioned 6–60 kb distal to the ε and β minor genes, respectively. The ε and βh1 genes are expressed in the embryo, whereas β major and minor are expressed in the adult. Among the eight hypersensitive (HS) sites, three are constitutive and present in preerythroid cells (5′HS–60/–62, HS5, and 3′HS1) and five are erythroid-specific sites (HS1, 2, 3, 4, and 6). In preerythroid cells, the constitutive HSs appear to interact, as determined by 3C, forming a poised chromatin hub. In embryonic cells, the HS1–3 and the embryonic globin genes join to form an active chromatin hub; in adult cells, the embryonic genes are looped out and the adult genes join 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 20 20 / Chapter 1 LCR 5′HS–60/–62 OR 6 54 3 2 1 3′HS1 10 kb ε Y β h1 β maj β min OR 5′HS–60/–62 HS4–6 3′HS1 βmin 5′HS–60/–62 3′HS1 HS1–3 εY βh1 εY βmaj HS1–6 βh1 βmin βmaj Progenitor “poised” β-globin chromatin hub Adult stage β-globin active chromatin hub FIGURE 1.7. LCRs and active chromatin hubs. (Adapted, with permission of Elsevier, from Dean 2006.) The organization of the mouse LCR and globin locus is shown schematically. The folding of the various elements into a poised chromatin hub in preerythroid cells and an active chromatin hub in adult erythroid cells is shown. There is 3C evidence for the various loops. The positions of the DNase I hypersensitive sites are shown. the active chromatin hub. Erythroid-specific transcription factors and coactivators, including GATA-1, EKLF, and FOG-1, are required for the active chromatin hub stability (see Chapter 10). In mouse, the active genes attached to the hub are heavily acetylated but the inactive genes are not; in humans, there seems to be persistent acetylation of the embryonic genes. Numerous LCRs have been characterized in metazoan genomes, and the diversity of organization suggests multiple ways that LCRs can regulate target genes (Dean 2006). Some exciting examples of well-characterized LCRs include the growth hormone LCR, which differentially controls a cluster of genes. One of these genes is expressed in the pituitary gland and the others are expressed in placenta. The TH2 LCR controls interleukin (Il) -4, -5, and -13 gene expression during T-cell differentiation. The LCR is actually positioned within the Rad50 gene located between Il-5 and Il-13. Recent data also suggest that the cytokine genes function in higher-order chromosomal domains defined by the protein SATB1, which controls cytokine gene expression by binding to baseunpaired regions (Galande et al. 2007). Insulators Insulator DNA elements and their accompanying binding proteins play multiple roles in regulation of transcription (Zhao and Dean 2005; Valenzuela and Kamakaka 2006; Dorman et al. 2007; Maeda and Karch 2007; Wallace and Felsenfeld 2007). Two major roles for insulators are regulating the spread of heterochromatin into transcriptionally active regions and maintaining the gene specificity of enhancers by preventing them from acting on promoters other than their intended target. For example, the β-globin genes in mammals are located between two clusters of odorant receptor genes. Because the two types of genes are not coexpressed in the same tissues, there must be a mechanism in place to prevent their enhancers and promoters from interacting. Insulators apparently prevent this crosstalk. However, another consideration is the 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 21 A Primer on Transcriptional Regulation in Mammalian Cells / 21 absence of factors that bind to the proximal promoter of the odorant receptor genes in erythroid cells and the absence of erythroid factors in sensory neurons. Finally, as the human genome continues to be analyzed, many examples have begun to emerge in which regulatory elements of one gene are actually present in or separated by another gene (e.g., the cytokine LCR described above). As discussed above, in their simplest form, insulators play two roles; their functions can frequently but not always be distinguished by the proteins that bind to them. In one role, insulators block the access of heterochromatin into active regions of the genome. This is often referred to as a barrier function. In a second role, insulators block an enhancer from inappropriately communicating with the wrong promoter. This function is often referred to as enhancer blocking. The literature on insulators is very complicated, and the function of insulators is linked to higher-order chromosome organization and nuclear architecture. Barrier insulators are usually associated with sequence-specific transcription factors, chromatin modification, and remodeling. Enhancer-blocking insulators show a propensity to form loops and sometimes attach to nuclear structures such as the nucleolus and the nuclear lamina. The 5HS4 element of vertebrate β-globin insulator contains binding sites for the upstream stimulatory factor (USF) and the CCCTC-binding factor (CTCF). USF is a transcriptional activator that recruits chromatin acetylation and remodeling proteins. CTCF is an 11 zinc finger DNAbinding protein that recognizes a 20-bp GC-rich site with a known consensus from genome-wide analyses. CTCF appears to be the major vertebrate protein that mediates enhancer blocking. In isolation, USF displays clear barrier function, whereas CTCF displays only enhancer-blocking activity. Another role of insulators may be the organization of an active chromatin hub via the formation of stereospecific arrangements of chromatin loops, which position active and inactive elements into different subdomains. The contact points in the β-globin active chromatin hub as determined by 3C correlate best with the positions of CTCF, rather than the borders of the DNase I HS sites. Genome-wide studies have shown that 10% of the DNase I hypersensitive sites in a cell map to CTCF-binding sites (Barski et al. 2007; Kim et al. 2007; Xi et al. 2007). Recent genome-wide studies also suggest that CTCF colocalizes with and is necessary for cohesin binding (Parelho et al. 2008). As mentioned above, CTCF-binding sites frequently, but not always, flank individual genes, and in many cases, they flank clusters of related genes such as the odorant receptors. Typically, CTCF sites are located in gene-rich regions of the genome. They tend to be surrounded by about 20 positioned nucleosomes rich in H2AZ (Fu et al. 2008). The mean distance of a CTCF-binding site to a TSS is approximately 50 kb. CTCF appears to interact with the nucleophosmin protein located at the nucleolus and with the nuclear matrix. Because of these associations, it has been proposed that tethering to a structure is important for its activity. However, immunofluorescence studies have shown a wide range of positions for CTCF in the nucleus, which are not restricted to the nucleolus or nuclear periphery. In D. melanogaster, many insulators bound by suppressor of Hairy-wing [Su(Hw)] form aggregates at the nuclear periphery called insulator bodies (Maeda and Karch 2007). The insulator bodies are hypothesized to control formation of specific transcriptionally active and inactive domains. The insulator function of CTCF has been examined on the imprinted Igf2/H19 locus (Fig. 1.8) (Zhao and Dean 2005; Valenzuela and Kamakaka 2006; Wallace and Felsenfeld 2007). In this case, an enhancer located downstream from H19 can control either H19 or the Igf2 gene positioned upstream of H19. A boundary element known as the differentially methylated region (DMR) is positioned between Igf2 and H19. When the DMR is unmethylated, as in the maternal alleles, CTCF binds and prevents the action of the enhancer on Igf2. This restricts the enhancer to H19. In the paternal allele, the DMR is methylated preventing CTCF binding and allowing the enhancer to activate Igf2. 3C studies have shown that the presence of CTCF allows different loops to be made; in its presence, Igf2 is looped out, and in its absence, Igf2 forms a different loop with the enhancer. 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 22 22 / Chapter 1 CTCF Igf2 ––– OFF ICR H19 +++ ON E E Enhancers Me Igf2 +++ ON ICR H19 ––– OFF E E FIGURE 1.8. CTCF insulator at H19/Igf2 locus. H19/Igf2 is an example of an imprinted locus in which one allele is maternally inherited and active and the other allele is paternally inherited. In the maternal allele, CTCF binds to the DMR and restricts the enhancer to H19. In the paternal allele, the DMR is methylated, which inhibits CTCF binding. The enhancer now loops and activates Igf2. (Adapted, with permission of Elsevier, from Wallace and Felsenfeld 2007.) The Pol III transcription factor TFIIIC can also serve as a barrier element in S. pombe. It forms about ten nuclear bodies, where various chromatin loops are attached to the nuclear periphery (Valenzuela and Kamakaka 2006). These nuclear bodies have been proposed as organizing sites for the barrier activity of TFIIIC and may be related to the Su(Hw) insulator bodies in D. melanogaster. There is an extensive literature on insulators and their binding proteins in D. melanogaster and many elegant mechanistic experiments testing how barriers and blockers work. For recent reviews and a more comprehensive treatment of the subject, see Dorman et al. (2007) and Maeda and Karch (2007). Basal Transcription Complex Assembly and Initiation Purified GTFs and Pol II mediate a low level of basal transcription on a core promoter in vitro, but they cannot support activated transcription in the absence of coactivators (see below). The study of the basal transcription machinery was initiated before the discovery of coactivators, and, importantly, there is probably no such phenomenon as basal transcription in vivo. Nevertheless, the characterization of the basal reaction in vitro has revealed much about the catalytic events transpiring during Pol II initiation and escape from the promoter. In most of the original studies on preinitiation complex assembly, TBP was used in place of TFIID because TBP was small, it could support basal transcription in the presence of the other GTFs, and, finally, it could be subjected to electrophoretic mobility-shift analysis (EMSA) for detailed structure–function analyses. Furthermore, TFIID was not sufficiently purified at that time to analyze its role in preinitiation complex (PIC) formation. These early studies indicated that purified GTFs and TBP assembled into a transcription PIC on the DNA in a stepwise fashion (Hahn 2004; Kornberg 2007). The complex was nucleated by the binding of TBP to the TATA box, a reaction aided by TFIIA or TFIIB, which can bind in any order after TBP. The crystal structures of TBP and the TBP/TFIIB and TBP/TFIIA complexes with DNA were solved, revealing insights into the process of promoter recognition. (For detailed structures and information on binding, see Chapter 10.) Both TFIIA and TFIIB contact DNA and TBP, increasing the stability of TBP binding to the TATA box. After TFIIB binds to TBP, a complex of TFIIF in association with Pol II is recruited, followed by sequential binding of TFIIE and TFIIH. Exciting insights into the mechanisms of initiation and elongation emerged with a series of crystal structures of Pol II alone, elongating, and in complexes with TFIIS and the amino-terminal domain of TFIIB (Kornberg 2007; Cramer et al. 2008). Pol II comprises 12 subunits with a 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 23 A Primer on Transcriptional Regulation in Mammalian Cells / 23 RNA RPB1 DNA RNA:DNA hybrid RPB2 Bridge helix (striped) FIGURE 1.9. RNA Pol II elongation complex. The S. cerevisiae Pol II elongation complex from (PDB file 1R9S) was rendered using PyMOL. The DNA–RNA hybrid from the Thermus thermophilus structure (PDB 2O5I) was modeled in place of the original RNA–DNA hybrid from the yeast structure. The entering DNA and exiting RNA were artificially extended to better illustrate the paths. Coloring: (Light gray; on top) Rpb1; (medium gray; on bottom) Rpb2; (dark gray; on sides) remainder of Pol II. DNA and exiting RNA are labeled and are very dark gray. The bridge helix is striped. (Rendered by J. Heiss, UCLA.) combined molecular mass of 515 kD in S. cerevisiae. The largest two subunits, Rpb1 and Rpb2, form the catalytic cleft and have homologs among Pol I and Pol III and the β′ and β subunits of prokaryotic RNA polymerase. These subunits serve general roles in transcription initiation and catalysis (Hahn 2004). A large central cleft formed largely by Rpb1 and Rbp2 (Hahn 2004; Kornberg 2007) contains three key features termed the clamp, the jaw/lobe, and the wall. The clamp is extremely mobile and is in an open configuration in the absence of DNA but clamps down in the presence of DNA. The DNA enters through the cleft and proceeds into the catalytic site near the base of the wall (Fig. 1.9). The catalytic site contains a magnesium ion, which is important for phosphodiester bond formation. Two channels allow for the entry of nucleotides into and the exit of RNA from the catalytic site. An α-helix termed the “bridge helix” spans across the cleft and translocates the DNA through the catalytic site. Much is known about the mechanism of catalysis and the movement of various mobile elements within the catalytic domain. For molecular details about the process of elongation, see Kornberg (2007) and Cramer et al. (2008). The positioning of the TSS into the active cleft is a result of a series of protein–protein and protein–DNA interactions among TBP, TFIIB, and Pol II. (Chen and Hahn 2003, 2004; Miller and Hahn 2006). TBP binds in the minor groove of the TATA box and bends the DNA 80° toward the major groove. This bending is one of the steps necessary for positioning the start-site DNA within the cleft (Fig. 1.10). TFIIB consists of two domains: an amino-terminal zinc ribbon domain and a carboxy-terminal core domain separated by a flexible linker containing a conserved region termed the B finger. Chemical crosslinking combined with the crystal structures of TBP/TFIIB and TFIIB/Pol II reveals that the B finger of TFIIB enters from the RNA exit channel and binds within the cleft adjacent to the Pol II catalytic site. The carboxy-terminal domain of TFIIB binds both TBP and the cleft and wall area of Pol II. This binding directs the DNA that exits TBP toward the active cleft. At this stage, the details become murky because of a lack of crystal structures. Chemical crosslinking studies place the XPB helicase subunit of TFIIH downstream from the TSS. The XPB helicase causes negative superhelical strain on the DNA between the downstream position and the 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 24 24 / Chapter 1 TFIIB ribbon TFIIB core TBP TFIIB core RPB1 TFIIH crosslinking site Catalytic site TFIIB ribbon Cleft Cleft TBP RPB2 FIGURE 1.10. TFIIB and TBP position promoter DNA adjacent to Pol II cleft. (The coordinates are from S. Hahn and adapted from Miller and Hahn 2006.) Two views of the TBP TFIIB, Pol II, DNA complex are shown. TBP bends , the DNA toward the major groove and around Pol II. TFIIB core binds DNA adjacent to both TBP and Pol II. TBP and TFIIB direct DNA toward the Pol II cleft. (Thick, dark gray noodle) Position of the TFIIB amino-terminal domain. (Upper view) Bending by TBP and the positioning of the two domains of TFIIB core; (lower view) binding of TFIIB amino terminus and its positioning near the catalytic site. (Rendered by J. Heiss, UCLA.) TFIIB-binding site, leading to melting of 11–15 bp of DNA at the start site (Kim et al. 2000; Hahn 2004). This allows the template strand to be inserted into the catalytic cleft. Some evidence suggests that TFIIF and TFIIE play roles in positioning the nontemplate stand against the polymerase to form the stable transcription bubble referred to as the open complex (Kornberg 2007). Studies reveal that templates bearing a preformed bubble over the TSS bypass the need for TFIIH and TFIIE but not TFIIF (Holstege et al. 1995). After the formation of the first phosphodiester bond, the DNA begins to translocate, driven by the bending of the bridge helix, and moves sequentially through the active site. Modeling of the B finger in the structure of elongating Pol II reveals that beyond nucleotide 5 of the elongating RNA, there would appear to be a steric clash between TFIIB and the growing RNA chain. It has been hypothesized that TFIIB would be required to dissociate from the PIC for productive elongation to continue. This distance in the crystal structure is similar to what has been described as abortive initiation in which Pol II synthesizes short transcripts without leaving or escaping from the promoter. Thus, the exit of TFIIB is probably important for promoter escape by Pol II. 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 25 A Primer on Transcriptional Regulation in Mammalian Cells / 25 Studies in S. cerevisiae extracts have found that upon productive elongation, TFIIB is indeed released from the PIC along with the elongating Pol II and TFIIF, leaving behind the activator, Mediator, TFIID, TFIIE, and TFIIH (Hahn 2004). The set of factors remaining at the core promoter has been referred to as a reinitiation scaffold. The scaffold complex assists and accelerates the process of activator-dependent reinitiation. The crystal structure of Pol II suggests that the CTD sits immediately outside of the RNA exit channel. As the RNA emerges from the channel, it is capped with 7meGTP in a process requiring the phosphorylation of the Pol II CTD at Ser-5 by the Cdk7 subunit of TFIIH. This phosphorylation provides a binding site for the mRNA capping machinery. Mediator Most current studies support the idea that the Mediator is the major coactivator for the PIC in S. cerevisiae and mammalian cells (Biddick and Young 2005; Conaway et al. 2005; Kornberg 2005; Malik and Roeder 2005; Roeder 2005). The Mediator complex was first identified as a biochemical activity in S. cerevisiae extracts that permitted activators to stimulate the basal transcription reaction observed with pure GTFs and Pol II. Indeed, not only does the Mediator allow activated transcription, but it is also required for high-level basal transcription in vitro in the absence of activators. S. cerevisiae Mediator consists of 25 subunits, whereas mammalian Mediator contains greater than 30 (for a comprehensive evolutionary comparison, see Bourbon [2008]). One or more subunits of Mediator are necessary for transcription of most S. cerevisiae genes, and many Mediator subunit knockouts in mice are lethal. Both structural and genetic studies show that Mediator not only coactivates, but, depending on the context, also corepresses transcription (Ding et al. 2008) and may play additional postinitiation roles in gene regulation (Wang et al. 2005). The Mediator structure as revealed by EM consists of four modules termed middle, head, tail, and kinase (Chadick and Asturias 2005). The structures of the S. cerevisiae and mammalian Mediators are similar. The largest form of mammalian Mediator is approximately 42 x 18.5 nm at its longest and widest points. In contrast, Pol II is only 14 x 13.6 x 11 nm. It is clear that Mediator can basically envelop Pol II. The Mediator displays an elliptical shape, which in the presence of Pol II changes conformation to a crescent shape encompassing the more globular shaped Pol II (Fig. 1.11). Several studies have identified at least two forms of Mediator. One form contains the kinase module bearing CDK8 or CDK11 and very little Med26; the other form contains Med26 and nearstoichiometric amounts of Pol II but little of the kinase module (Paoletti et al. 2006). Some data suggest that the kinase form is repressive and the Med26/Pol II form is stimulatory. However, other studies show that the kinase activity contributes to gene activation in vitro and probably in vivo (Liu et al. 2004). Pol II appears to bind Mediator through both its CTD and regions of the core Exiting DNA Pol II Entering DNA Clamp Head CTD base Middle Tail Mediator FIGURE 1.11. Model of Mediator/Pol II structure. The CryoEM structure of Mediator encompassing Pol II. (Adapted, with permission of Macmillan Publishers Ltd, from Hahn 2004.) 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 26 26 / Chapter 1 enzyme. In addition to interacting with Pol II, the Mediator also contacts TFIID (Johnson et al. 2002) within the context of promoter DNA. Many different activators contact the Mediator, and this interaction is probably central to recruitment of Mediator to a core promoter (Biddick and Young 2005; Chadick and Asturias 2005; Conaway et al. 2005; Kornberg 2005; Malik and Roeder 2005). This issue has been more extensively investigated with the mammalian complex (see Chapters 10 and 12). For example, nuclear receptors, including the thyroid receptor (TR) and vitamin D receptor (VDR), interact with Mediator via the LXXLL motifs in the MED1 subunit found in the middle module. MED15 in the tail module interacts with SMAD2/SMAD4 and with SREBP. The highly potent herpes simplex virus 1 activator VP16 interacts with MED25, and the adenovirus activator E1A and the cellular activator ELK1 interact with MED23. The location of MED23 and MED25 is unresolved, although MED23 is likely to be in the tail module based on the instability of MED16 and MED24 in MED23 knockout mouse embryonic fibroblasts (MEFs), and the instability of MED23 and MED16 in MED24 knockout MEFs. Thus, MED23 and MED24 appear to be linked to the tail module via MED16. Finally, the MED29 subunit of Mediator, located in the head domain, is identical to the D. melanogaster Intersex protein, which interacts directly with the sequence-specific activator Doublesex. These observations collectively point to the idea that all of the core Mediator modules, aside from the kinase module, can engage in interactions with activators. It is clear that the recruitment of Mediator by activators is a rate-limiting step in vitro and is likely to be in vivo. This does not preclude the idea that Mediator may also act at postinitiation steps (Struhl 2005). The conformations of the Mediator, as measured by EM, change dramatically with different activators and Pol II bound (Taatjes et al. 2004a,b). Neither the mechanism nor the functional consequences of the conformational changes are known, but it has been speculated that these affect Pol II activity or the activity of the PIC. Nevertheless, it is very clear from numerous ChIP and biochemical studies that activators recruit the Mediator in vivo and in vitro and that this recruitment adds both stability and enhanced function to the GTFs. TFIID and TAFs As described above, TFIID is the key factor tethering the Mediator and remaining GTFs to the core promoter via its binding to numerous sequence elements surrounding the TSS. The TFIID structure in S. cerevisiae and humans has been revealed by EM studies. TFIID is a three-lobed horseshoe shaped structure (20 nm x 13.5 nm x 11 nm) with the TBP positioned at the base of a central cavity through which DNA is believed to pass (Fig. 1.12) (Andel et al. 1999; Grob et al. 2006). The complex undergoes considerable conformational flexibility, likely due to the movement of hinges between the three lobes. The structure contains several channels and cavities, which are thought to be interaction surfaces for other proteins. RNA Pol II transcription requires the 13–14 TAFs present in TFIID (Hahn 1998; Albright and Tjian 2000; Chen and Hampsey 2002; Tora 2002; Davidson et al. 2005). Note that TBP is required for transcription by all three nuclear RNA polymerases, but a different set of TAFs is bound to TBP in the context of Pol I and Pol III transcripLobe C Lobe B Channel 1 Lobe A d2 d1 FIGURE 1.12. TFIID Cryo-EM structure. TFIID is a three-lobed structure (A,B,C) with considerable flexibility and numerous channels and cavities. The approximate location of TBP is in the main cavity at the base of lobe C. TFIIB would be binding in the neighborhood of channel 1 and against lobe B. TFIIA is between the lobe A and d1. (Adapted, with permission of Elsevier, from Grob et al. 2006; image provided by Eva Nogales.) 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 27 A Primer on Transcriptional Regulation in Mammalian Cells / 27 tion. TAFs also play roles outside of the TBP complexes. A subset of TAFs is found in chromatinmodification enzymes such as SAGA in S. cerevisiae and STAGA/TFTC in mammals. The TAFs are somewhat conserved from S. cerevisiae to mammals. Although the core TAFs appear to be ubiquitous, there are several examples of cell-specific TAFs, including five paralogs, which are specific for D. melanogaster male germ cells and are involved in spermatogenesis. Several other paralogs exist, some with tissue-restricted patterns of gene expression, and there are splicing isoforms of many of the TAFs. TAFs are involved in core promoter recognition, interaction with other GTFs, activated transcription, and chromatin recognition. We have already discussed their roles as core promoter recognition factors above and discuss here their roles in activated transcription and chromatin. The precise roles and physiological function of TAFs have been the subject of much study and debate. One of the earliest proposals was that TAFs were required for transcriptional activation. Indeed, biochemical and genetic studies in mammalian cells and D. melanogaster revealed that TAFs interacted directly with activators and mediated the action of these activators in vitro (Albright and Tjian 2000). In a few instances, this requirement for activation was confirmed in vivo. Studies in D. melanogaster, mammalian, and S. cerevisiae biochemical systems have shown that activators recruit complexes containing TFIID and TFIIA to a core promoter, and this effect requires TAFs (see, e.g., Chi et al. 1995). The biochemistry was buttressed by genetic experiments in D. melanogaster pointing to a role for TAFs in activated transcription during development (Albright and Tjian 2000). For example, the activator Dorsal interacts with TAF4 and TAF6 in vitro, and genetic studies have shown that these TAFs are required for Dorsal-mediated activation of the snail and twist genes in vivo. Other studies in S. cerevisiae showed that the TAFs are essential genes and that transient inactivation of TAFs by temperature-sensitive mutants abolished transcription. Genome-wide studies suggest that up to 84% of S. cerevisiae genes require one or more TAFs. Analyses of TAF-responsive promoters and TAF-independent promoters suggest that the upstream activating sequence (UAS) of a TAF-responsive gene determines the TAF dependence. For example, UASs from a TAFdependent promoter were fused to a core promoter from a TAF-independent gene. It was shown that on these fusion constructs, the UAS recruits TAFs to a TAF-independent promoter. These data suggest that activators at these UASs directly contact TAFs (Chen and Hampsey 2002). Furthermore, in numerous ChIP studies of TAFs in S. cerevisiae and mammalian cells, TAF recruitment correlates with the activated state of the gene, even in cases where TBP alone may be prebound in the inactive state. Nine of the TAFs contain histone-fold domains. The significant homology between TAFs and histones has led to the hypothesis that TFIID may mimic nucleosome function, perhaps as a means of stabilizing its binding to DNA or as a way of displacing nucleosomes during transcription complex assembly. Five pairs of TAFs bearing histone folds are formed (4-12, 6-9, 3-10, 10-8, and 11-13). The different TAF pairs are distributed, some in multiple copies, throughout the three lobes of TFIID (Leurent et al. 2002). Indeed, chemical crosslinking and DNA topological analysis of TFIID/DNA complexes suggest that the DNA may be wrapped around TFIID, perhaps like DNA is wrapped around an octamer (Hoffmann et al. 1996; Oelgeschlager et al. 1996). However, examination of the histone surfaces that interact with DNA within the nucleosome crystal structure suggests that residues mediating the critical contacts (largely arginines with the minor groove and phosphate backbone) are not well-conserved in the TAFs. Despite this analysis, there have been reports showing that various histone fold TAFs bind to DNA (i.e., TAF6-9 and TAF4b-12), suggesting that TAFs use alternative surfaces for nucleic acid recognition (Shao et al. 2005). Another view is that TAFs allow internucleosomal contacts as a means of allowing TFIID to dock with a nearby nucleosome. In addition to the remote possibly of TAFs mimicking the structure of an octamer, some TAFs have domains that recognize modified chromatin. For example, TAF1 contains a double bromodomain that can bind to acetylated H4 tails in vitro (Jacobson et al. 2000). Note, however, that some bromodomains also recognize acetylated residues within transcription factors. Thus, TAF1 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 28 28 / Chapter 1 also binds acetylated p53, providing a mechanism for recruitment of TFIID to a promoter (Li et al. 2007b). Similarly, TAF3 contains a PHD (plant homeodomain) finger that binds to H3K4me3, providing a link between TFIID and the MLL/SET1 complex, which is recruited to promoters during gene activation (Vermeulen et al. 2007) (see below). In conclusion, TAFs have many roles in transcription. They are necessary for promoter recognition, bind certain chromatin modifications, and directly interact with some activators. Additionally, TAF1 has been reported to contain HAT and kinase activity thought to be important for gene activation. ACTIVATION AND REPRESSION Gene Activation It should be emphasized that genes can be regulated at different steps including initiation, elongation, mRNA processing and stability, transport to the cytoplasm, and translation. We focus here on the process of gene activation at the levels of initiation and early elongation. Proteins termed activators stimulate all mRNA-encoding genes. Activators are typically small modular proteins that tether to specific sites in the proximal promoter and enhancer (Ptashne and Gann 1990). Because gene activation begins with an activator, much emphasis was placed in the 1980s and 1990s on the structure, modular design, and regulation of activators. Most activators contain separable DNA-binding and activation domains, where the DNA-binding domain is responsible for promoter/enhancer binding (see Chapters 10 and 11) and the activation domain contacts chromatin modification/remodeling proteins and coactivators (see Chapters 10, 12, and 13). All regulated genes respond to signals or cues that begin with the activator (Ptashne and Gann 1990). Activators are regulated at several different levels. Some activators are not synthesized, active, or localized to the nucleus until they receive a specific signal. For example, the glucocorticoid receptor (GR) binds a small steroid ligand that causes GR to dissociate from chaperones in the cytoplasm and localize to the nucleus (Lefstin and Yamamoto 1998). Other nuclear receptors such as RARRXR heterodimers remain bound to responsive promoters in a repressive form and are activated by the ligand (Xu et al. 1999). Another way to activate a DNA-bound activator is by posttranslational modification. The Elk-1 transactivator binds DNA, but it is not active until phosphorylation by mitogen-activated protein kinase (MAPK) (Treisman 1996). Some activators are cell- and developmentally specific. Whatever the circumstances, these DNA-binding proteins, probably because of their small size, do appear to have limited access to chromatin by binding either in the linker region between nucleosomes or to short stretches of exposed DNA on the surface of the nucleosome. It is also known that in some instances, activators bind cooperatively to form stable complexes at the promoter and enhancer; these are called enhanceosomes (reviewed in Chapter 11). Once active and bound, these proteins initiate a series of events that lead to gene activation. Although this process seems to be quite straightforward, we have limited knowledge of the precise mechanism by which any single gene is activated. Some of the most thoroughly studied examples have been in S. cerevisiae, where the promoter–enhancer structure is much simpler and the use of genetics and ChIP has been able to confirm the biochemistry. What is clear from these studies is that the activator must first alter chromatin at the enhancer and promoter before recruiting the Pol II machinery (see, e.g., Reinke and Horz 2003). Promoter regions in S. cerevisiae seem to have lower densities of nucleosomes, and nucleosomes are clearly evicted during gene activation (see below) (Li et al. 2007b). After activator-mediated recruitment of the Pol II machinery to the core promoter, via interactions with TFIID and Mediator, Pol II must initiate and elongate transcription on a chromatin template (Li et al. 2007a). Nucleosomes serve as an intrinsic block to Pol II elongation, and thus Pol II must direct remodeling of the impeding nucleosomes during transcription. Importantly, nucleosomes must be returned to their original position after transcription to prevent spurious transcription initiation from within the gene body. Finally, during the process of transcription, Pol II must 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 29 A Primer on Transcriptional Regulation in Mammalian Cells / 29 communicate with the RNA-processing machinery. Hence, the details that have emerged suggest a very complicated series of catalytic events involving numerous types of protein modifications. Chromatin Modification and Remodeling during Transcription Initiation In many cases, acetylation and remodeling of the promoter precede transcription (Workman 2006). As discussed above, HATs carry out acetylation and ATP-dependent enzymes carry out remodeling. These topics are discussed in more detail below. In cases in which the promoter is in an actively silenced state (i.e., bound by Polycomb or HP1, see below), demethylation and removal of the repressive marks occur concomitantly with histone hyperacetylation and H3K4 trimethylation. Typically, one of the first steps in gene activation is recruitment of various HATs to the gene by the activators (Fig. 1.13). These HATs include complexes containing GCN5 (i.e., STAGA and TFTC in mammals and SAGA and its variants in S. cerevisiae) or p300 and CBP, among others (Lee and Workman 2007). For example, the interferon-β (IFN-β) enhanceosome binds p300 and recruits it to a promoter (Merika and Thanos 2001). Indeed, p300 binding seems to stabilize the enhanceosome in vitro. Once bound, the HATs acetylate the histone tails as well as other proteins. The acetylated tails destabilize higher-order chromatin contacts (i.e., H4K16ac) and serve as recognition sites for other bromodomain-containing proteins (Taverna et al. 2007), including ATP-dependent remodeling proteins (Peterson and Workman 2000; Mohrmann and Verrijzer 2005; Li et al. 2007a). For example, the SWI/SNF complex in S. cerevisiae contains a bromodomain that recognizes acetylated lysines. The S. cerevisiae RSC chromatin remodeler contains eight bromodomains, which recognize various acetylated histone residues. In mammals, PBAF contains at least seven bromodomains, which are similar to those found in RSC subunits (Mohrmann and Verrijzer 2005). As described above, RSC and SWI/SNF recognize acetylated lysines with greater avidity, and both enzymes remodel and evict nucleosomes from the template. Numerous studies suggest that nucleosomes are evicted from S. cerevisiae promoters in vivo during gene activation and that this is a prerequisite for recruitment of the general machinery (Workman 2006). Although the model discussed above implies that HATs precede the appearance of ATP-dependent remodeling machines, there are cases where the opposite occurs (Cosma 2002). Activators can directly contact SWI/SNF in S. cerevisiae and mammalian cells, and it is plausible that the action of a remodeling protein is necessary to loosen higher-order chromatin to permit acetylation. It is likely that a combination of acetylation and direct contact with activators recruits the remodeling enzymes to promoters. The ultimate goal is to generate a platform for the recruitment of the general transcription machinery. A Model for Recruitment of the General Machinery When reviewing the literature, it is quite common to see the term coactivators used very broadly to indicate anything that binds directly to a sequence-specific DNA-binding protein. In some cases, these proteins are involved in chromatin remodeling and in other cases, the so-called coactivator might simply be the true activation domain. In this chapter, we restrict the term coactivator to the Mediator and TFIID complexes because their binding is rate-limiting for transcription initiation, and, as discussed above, TAFs and Mediator subunits are known to be direct targets of activators. In biochemical studies, Mediator and TFIID form complexes with activators on promoter DNA and recruit the remaining GTFs. These recruitment events do not require ATP and upon binding of the remaining GTFs, the preinitiation complex is in the closed form with the DNA unmelted. The addition of ATP signals several key events that allow Pol II to form the open complex, in which the start site is melted, and initiate transcription. DNA melting requires the hydrolysis of ATP by the XPB subunit of TFIIH as discussed above. There is some evidence that activators may affect this melting step. During or shortly after initiation, the CDK8 subunit of Mediator and the CDK7 subunit of TFIIH phosphorylate serines within the heptad repeat of the Pol II CTD (Phatnani and Greenleaf 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 30 30 / Chapter 1 HAT ac ac Nap1 ATPase BD ac ac ATPase BD ac ac ac ac Nucleosome a Act Asf1 • • • Histone acetylation by HATs Histone eviction by ATPase Binding of additional activators b Act Act Nap1 ac Asf1 ac • • Activators recruit coactivators: IID, Mediator Coactivators recruit GTFs IIF Mediator TFIID Pol II IIB IIE IIH c Act Act M d IID ed iator IIF IIB IIH IIE M e IID ed iator IIF IIB IIE IIH • • ATP, IIH lead to open complex CTD is phosphorylated at Ser5 Initiation Promoter escape, Pol II elongation (TFIIB leaves PIC) FIGURE 1.13. (See facing page for legend.) 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 31 A Primer on Transcriptional Regulation in Mammalian Cells / 31 2006). Phosphorylation at Ser-5 by CDK7 is required for two events—the recruitment of the mRNA capping machinery and the recruitment of the PAF1 complex, which in turn is necessary for recruitment of the SET1/MLL complex, which trimethylates H3K4 (Shilatifard 2008). At this juncture, a series of confusing observations seem to contradict the model described above. First, mammalian TFIID contains tandem bromodomains and a PHD domain, both of which bind modified chromatin. Hence, if these modules are indeed functional during gene activation, then it would imply that nucleosomes are still present at the promoter, which contradicts the simple notion that they are removed to make way for the general machinery. A second caveat is that the general machinery must be bound at the promoter to phosphorylate the Pol II CTD, necessary for the binding of the Set1/MLL complex, which trimethylates H3K4. This would seem to imply that H3K4 trimethylation occurs after TFIID binding. Hence, the idea that H3K4me3 assists in the recruitment of TFIID would seem unlikely and instead suggests a postinitiation role for this interaction (i.e., stable maintenance of binding). It is noteworthy that in S. cerevisiae, a separate protein termed Bdf1 encodes the bromodomains found in mammalian TAF1 (Matangkasombut et al. 2000; Taverna et al. 2007). More importantly, there is no PHD domain in the S. cerevisiae TAF3. Thus, the mechanisms and order of function of various complexes in S. cerevisiae and mammalian cells may differ. Unlike TAF1 and TAF3, CHD1 is conserved from S. cerevisiae to humans (Daniel et al. 2005; Hall and Georgel 2007). As described above, CHD1 is a chromodomain-containing ATPase of the SWI/SNF family. CHD1 also binds H3K4me3 via its chromodomain. Because H3K4 trimethylation is associated with the 5′ end of active genes, it is believed to play some early role in elongation. A second caveat to the above recruitment model is that some mammalian promoters appear to be prebound by one or more of the GTFs and TBP (Cosma 2002). The original cases involved genes that were activated during the lengthy process of liver cell differentiation. These genes include α1 antitrypsin (Soutoglou and Talianidis 2002) and hepatocyte nuclear factor-4α (Hatzis and Talianidis 2002). In these cases, a subset of activators and GTFs is already present at the promoter and the appearance of a new activator leads to recruitment of the TAFs, Mediator complex, and a full complement of GTFs, along with chromatin-modifying complexes and Pol II. Initial Stages of Pol II Elongation One of the major obstacles to efficient Pol II elongation is pausing ( Shilatifard et al. 2003; Peterlin and Price 2006; Saunders et al. 2006; Armstrong 2007; Svejstrup 2007a). After initiation and promoter escape, Pol II elongates to between +20 and +50 relative to the TSS and pauses (Fig. 1.14). This pausing is dependent on two factors termed the DRB sensitivity–inducing factor (DSIF) and negative elongation factor (NELF). The pause appears to be a critical checkpoint in the process of elongation, possibly to allow the Pol II to switch from binding to the mRNA capping machinery to binding to the splicing and transport machinery. The CDK9 subunit of the elongation factor pTEF-b (Ctk1 in S. cerevisiae) is required to release Pol II from this pause (Peterlin and Price 2006). Several lines of evidence suggest that activators recruit pTEF-b and influence this early stage FIGURE 1.13. Gene activation. The various steps in gene activation are illustrated schematically. (a) An activator (Act) binds to DNA sites in the proximal promoter and recruits ATP-dependent remodeling enzymes (ATPases) and histone acetyltransferases (HATs). HATs acetylate the histone tails (ac) on the nucleosomes. ATPases can also bind acetylated lysines via their bromodomains (BD). (b) Chaperones such as Nap1 and Asf1 serve as acceptors for histones evicted by the ATPase. Additional activators may bind after the histones are removed. (c) Activators then recruit coactivators like Mediator and TFIID. TFIID binds the TATA box with the help of TFIIA (TFIIA is not shown). (d) The coactivators recruit the general transcription factors and Pol II. TFIIH (IIH) binds downstream of the start site. (e) ATP hydrolysis by IIH leads to melting of the start site to form the open complex. The CTD of Pol II is phosphorylated at Ser5 by the Cdk7 subunit of IIH. Pol II initiates transcription and TFIIB leaves the complex along with elongating Pol II and TFIIF. As Pol II elongates about 8–10 nucleotides into the gene it escapes from the promoter. Additional details are discussed in the text. Please note that the step affected by enhancers is not well understood. The drawing is not to scale. 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 32 32 / Chapter 1 a. Promoter-proximal pausing by Pol II RNA Paused Pol II DSIF NELF Nucleosome CTD Ser5 PO4 DNA bubble Pol II +50 • Phosphorylation of DSIF, NELF, CTD at Ser2 by pTEF-b causes release from pause • NELF leaves and DSIF stays and stimulates elongation PO4 pTEFb PO4 (leaves) RNA PO4 NELF DSIF (stays) Pol II Released Pol II b. Methylation of H3K4 and remodeling at beginning of gene CTD Ser5 PO4 PAF-1 H3K4me3 Pol II c. Methylation, remodeling, deacetylation within gene mRNA H3K4me3 Pol II H3K36me3 Set2 MLL/Set1 HATs ATPase ac ac ac ac ac FACT ac Spt6 ub Bre1/Rad6 • Methylation of H3K4 CHD1 ATPase CD H3K4me3 Pol II Recruitment of and remodeling by CHD1 ac ac ac ac HATs ATPases ac ac • Set2 methylation of nucleosome behind Pol II • Binding of Rpd3S to methylated nucleosome • Deacetylation • Concurrent acetylation and eviction of histones in front of Pol II H3K4me3 ATPase ac ac H3K4me3 Pol II Pol II H3K36me3 H3K36me3 Set2 ac FACT Deacetylation CD ac Spt6 PHD Rpd3S ac ac FIGURE 1.14. (See facing page for legend.) 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 33 A Primer on Transcriptional Regulation in Mammalian Cells / 33 of elongation. Upon recruitment, CDK9 phosphorylates DSIF (Spt4 and Spt5 in S. cerevisiae), NELF, and Ser2 of the CTD. Phosphorylated NELF leaves, but the phosphorylated DSIF remains and stimulates elongation. The Spt4 subunit of DSIF has been shown in S. cerevisiae to help the PAF1 complex bind Pol II (Qiu et al. 2006). This observation implies that the PAF1-dependent H3K4 trimethylation events described above may be more related to the initial stages of Pol II elongation than to initiation. There has been a considerable amount of interest in the Pol II paused just downstream from the start site (Margaritis and Holstege 2008). Many inactive genes also contain a paused Pol II. It has been proposed that this paused Pol II acts to keep the chromatin downstream from a gene open for rapid reactivation and, additionally, performs the first or pioneer round of transcription upon gene activation. It has also been proposed that this Pol II is simply an occasional and stochastic event that occurs on all genes but in a small percentage of the population. There are proponents of both views, but of interest is the observation that ChIP experiments show much more Pol II at the pause site in active versus inactive genes, which would not make sense if in 100% of cells an inactive gene is bound by a paused Pol II. Either way, the issue is of interest and the subject of much research. After escaping from the proximal pause site, Pol II elongates into the gene. Pol II apparently pauses frequently during transcription. The cause of the pausing is not entirely clear, but it may be due to certain DNA sequences, the presence of bound proteins, or the presence of nucleosomes. However, during pausing, the 3′ end of the RNA chain can become misaligned with the template. A number of different protein factors act to either prevent or reverse pausing at intrinsic pause sites (Shilatifard et al. 2003; Svejstrup 2007a). These factors include SII (TFIIS), TFIIF, Elongin, and ELL (Shilatifard et al. 2003). It is believed that once paused, Pol II backtracks and the misaligned 3′ end of the RNA is extruded through the pore, where nucleotides are believed to enter (Svejstrup 2007a). TFIIS actively reverses pausing by stimulating a transcript cleavage activity of Pol II. The structure of Pol II/TFIIS is known (Cramer et al. 2008). TFIIS binds in the Pol II pore and extends to the active site. TFIIS is believed to rearrange the active site and position a metal ion plus a water molecule to stimulate hydrolysis of the misaligned 3′ RNA. This event allows Pol II to continue transcription from the new properly aligned 3′ end of the RNA. If the pause cannot be passed, Pol II is thought to be ubiquitinylated and degraded (Svejstrup 2007a). ELL and Elongin greatly stimulate rates of elongation in vitro and are thought to possibly prevent the misalignment of the RNA that leads to pausing. Another complex found in S. cerevisiae and termed Elongator also stimulates elongation (Svejstrup 2007b). Elongator has HAT activity, which might be important for elongation on chromatin. Elongator is also found in the cytoplasm where it plays several different roles. FIGURE 1.14. Promoter escape, pausing, and elongation through the gene. The various steps in Pol II elongation are illustrated schematically. (a) After Pol II escapes from promoter it pauses at around 50 bp downstream (+50) from the transcription startsite. Pausing is caused by the action of two factors termed DSIF and NELF. The elongation factor pTEF-b phosphorylates (PO4) DSIF, NELF, and Ser2 of the CTD. NELF leaves the pause site while DSIF remains and stimulates elongation away from the pause site. (b) During the initial stages of elongation, probably concurrent with or prior to the step described in a, the PAF-1 complex is recruited to Pol II. PAF-1 recruits the MLL/Set1 complex (COMPASS in yeast), which trimethylates H3K4. This methylation (H3K4me3) is dependent upon ubiquitinylation (ub) of H2B by Bre1 and Rad6 (RNF20 and RNF40 in mammals). H3K4me3 recruits the CHD1 ATPase via its chromodomain (CD). CHD1 remodels the nucleosome near the start site. The nucleosomes downstream are acetylated by HATs and remodeled/evicted by ATPases. Little is known of how this step occurs. (c) During elongation nucleosomes are assembled behind Pol II by chaperones such as Spt6 and FACT. Histone acetylation is required for elongation by Pol II but the cell has developed a system for deacetylating the trailing nucleosomes to stop remodeling after Pol II has passed. In this pathway, the Pol II CTD is phosphorylated (PO4) at Ser2. This causes binding of Set2. Set2 trimethylates trailing nucleosomes at H3K36 (H3K36me3). This modification creates a binding site for the Rpd3S histone deacetylase complex, which binds H3K36me3 via its chromodomain (CD) and PHD domain. The Rpd3 deacetylase subunit then removes the acetyl groups and prevents remodeling of the trailing nucleosomes by ATPases. The deacetylation of trailing nucleosomes prevents cryptic initiation. Additional details are discussed in the text. The drawing is not to scale. 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 34 34 / Chapter 1 Pol II Encounters Nucleosomes within the Gene As Pol II elongates further into the gene, it encounters nucleosomes. Numerous studies have shown that the density of nucleosomes is higher in the body of the gene versus the promoter. Nevertheless, in S. cerevisiae, where genome-wide ChIP studies have been performed, there appears to be an inverse correlation between transcriptional activity and the nucleosome density within the gene (Workman 2006). This observation suggests that remodeling and some eviction occur within the gene during transcriptional elongation (Armstrong 2007; Li et al. 2007a). Turnover rates of histones measured by either pulse labeling or incorporation of newly synthesized epitope-tagged (see Chapter 10) histones suggest a very high level of H2A/H2B turnover and lower but significant rates of H3/H4 turnover in transcribed genes. In mammalian cells, H3.1, the predominant H3 species, is replaced with H3.3 during transcription, demonstrating that H3 turnover does indeed occur (Armstrong 2007). Other studies show that H3.3 turns over at a higher rate than H3.1. The idea that nucleosomes are disassembled during transcription is strengthened by studies of the heat shock genes. Heat shock genes in both S. cerevisiae and higher eukaryotes exhibit a significant loss of nucleosomes across the entire locus even before Pol II has begun transcribing into the gene (Weake and Workman 2008). The mechanism of histone displacement is unclear, but active genes are acetylated within the coding regions and ATP-dependent remodeling enzymes are necessary for transcription (Saunders et al. 2006; Workman 2006). Additionally, a diverse set of histone chaperones is necessary for the removal and replacement of histones during elongation. Mutations in several S. cerevisiae histone chaperones, including Asf1, Spt6, and FACT (Spt16 and Pob3 in S. cerevisiae), have specific effects on elongation (Saunders et al. 2006; Workman 2006; Armstrong 2007). The requirement for histone chaperones is linked to the fact that ATP-dependent remodeling proteins do not evict histones from DNA in the absence of an acceptor. This is because of the high affinity of the histone octamer for DNA. In biochemical experiments, acceptors can include another DNA molecule or a histone chaperone. For example, Nap1 can accept histones evicted by RSC (Lorch et al. 2006). In the absence of Nap1 or acceptor DNA, RSC will remodel the nucleosome (i.e., slide it to a new position) but not evict the octamer. Thus, in vivo, there must be chaperones that help remove or evict octamers, H3/H4 tetramers, or H2A/H2B dimers. FACT has been reported to disassemble nucleosomes in vitro to promote elongation in mammalian biochemical systems (Reinberg and Sims 2006). S. cerevisiae FACT, however, is a very strong nucleosome assembly factor in vitro. Similarly, S. cerevisiae Spt6 assembles nucleosomes from free histones in vitro but does not seem to evict octamer (Bortvin and Winston 1996). However, FACT along with NHP6A does appear to remodel the nucleosome significantly in the absence of an ATP-dependent remodeling protein (Rhoades et al. 2004). Additionally, the Nap1 chaperone can slide histones along the DNA and may help in eviction (Park and Luger 2006). The idea that Spt6 and FACT may be chromatin assembly factors is supported by the observation that in S. cerevisiae, mutations in the genes encoding these proteins cause a phenomenon known as cryptic or internal transcription initiation (Kaplan et al. 2003). These same mutations do not seem to severely inhibit elongation. One interpretation of this observation is that FACT and Spt6 reassemble nucleosomes after Pol II has passed, and mutations in these factors prevent reassembly, thereby allowing random transcription to occur within the body of the gene. This cryptic transcription appears to be dependent on Pol II, and there is limited evidence that it is due to PIC formation. It is unknown how the PICs assemble or whether the process requires activators. A sophisticated pathway operates during Pol II elongation to ensure the orderly reassembly of nucleosomes during transcription. This machinery, in addition to the chaperones cited above, includes an H3K36 trimethylase termed Set2 and a histone deacetylase termed Rpd3S (Li et al. 2007a), where S stands for small form; a larger version of Rpd3, termed Rpd3L, is found at the promoter of genes during repression, a point we discuss below. The Set2 enzyme binds the phosphorylated CTD of Pol II and travels along with it during transcription. The current, albeit unproven, model is that as Pol II transcribes through a region, the chaperones reassemble the nucleosome behind it. The source of the octamer is currently unclear, but it could be 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 35 A Primer on Transcriptional Regulation in Mammalian Cells / 35 either the octamer that originally occupied the position or a different octamer. Whatever the source, the octamer is likely acetylated or subject to acetylation, as Rpd3S is a deacetylase and mutations of the Rpd3S pathway lead to hyperacetylation of the coding region during transcription. Set2 is believed to trimethylate H3K36 behind Pol II, and this modification then recruits the Rpd3S complex via a combination of its chromodomain and PHD finger encoded by the Eaf3 and Rco1 subunits, respectively (Li et al. 2007c). Rpd3S then deacetylates the histones behind Pol II, effectively blocking the recruitment and action of bromodomain-containing ATP-dependent remodeling enzymes. A key question is why nucleosomes need to be rapidly reassembled after Pol II has passed. There are probably several reasons for this, including protection of genomic DNA from damage and blocking PIC assembly within the coding regions of genes. It is now known that noncoding RNAs play major roles in transcript degradation, translational control, and heterochromatin formation, among others. Therefore, preventing spurious transcription may simply be important for genome regulation. One might predict that highly transcribed genes with a high density of transcribing Pol II might be less dependent on the Set2/Rpd3S pathway because the Pol II occupancy will, in effect, substitute for the nucleosome and protect the gene (Li et al. 2007d). Indeed, much evidence suggests that this idea is correct. The data show that highly transcribed genes are not as dependent on the Set2/Rpd3S machinery. H3K36 methylation occurs within genes from S. cerevisiae to humans, but in mammals, another highly unusual mark exists in the coding regions of actively transcribing genes. The coding regions contain both H3K9me2 and H3K9me3 and appear to bind the γ member of the heterochromatin protein 1 (HP1) family (Eissenberg and Shilatifard 2006) (see below). HP1γ tends to be found throughout euchromatin, whereas HP1α and HP1β are typically concentrated in heterochromatin. Therefore, it has been suggested that HP1γ performs a function analogous to H3K36 trimethylation. Another modification associated with active chromatin is H3K79 methylation (Shilatifard 2006). This mark is catalyzed by the DOT1 methylase. The same methylase appears to be required to prevent spreading of heterochromatin into transcriptionally active regions in S. cerevisiae. In summary, nucleosomes must be modified, remodeled, evicted to some extent, and reassembled during elongation by Pol II. Although much is known about the machinery involved, far less is known of the precise mechanism. This subject remains an active area of investigation in the field. Silencing/Repression of Transcription Although removal of the activation signal is probably the initial step in gene inactivation, mechanisms exist to actively remove modifications necessary for gene activation. Typically, gene silencing/repression is associated with five types of systems: histone deacetylation, H3K9 di- and trimethylation, H3K27 trimethylation, H4K20 trimethylation, and DNA methylation at CpG residues. Each of these leads to specific consequences with respect to the binding of corepressors or silencing proteins and the assembly of chromatin into a structure that is incompatible with regulated Pol II transcription. Histone Deacetylation The concept of histone deacetylation–controlled repression was originally derived from the observation that the promoters of actively transcribed genes, and to a limited extent the coding regions, are hyperacetylated relative to nontranscribed genes. Hence, the idea emerged that histone acetylases are required for gene activation and that histone deacetylases are involved in repression. In fact, histone deacetylases deacetylate numerous proteins in addition to histones. In some cases, the deacetylation leads to activation of transcription and in others to silencing (Smith 2008). Additionally, we know that some histone deacetylation occurs within the body of a gene and is required for active transcription (Wang et al. 2002). Finally, not all acetylation of specific lysines positively correlates with transcription. Now that the caveats have been presented, we provide a general overview of histone deacetylation as it relates to transcriptional repression. 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 36 36 / Chapter 1 TABLE 1.4. HDAC complexes and their subunits S. cerevisiae Complex Sin3 Mammals HDAC1, HDAC2 RbAp46, RbAp48 Sin3A Sds3, BRMS1 RBP1 SAP30 SAP18 ING1/2 Rpd3L Rpd3S Rpd3 Ume1 Sin3 Sds3 Sap30 Pho23 Rco1 Eaf3 Mi-2/NuRD HDAC1, HDAC2 RbAp46, RbAp48 Mi-2α/β MTA1-3 MBD2, MBD3 p66α/β HDAC1, HDAC2 CoREST LSD1 BHC80 CtBP1 HDAC3 N-CoR/SMRT TBL1/TBLR1 GPS2 JMJD2A Kaiso ubiquitin fold PHD finger PHD finger chromodomain class I deacetylase WD40 repeat helicase SANT domain methyl CpG binding class I deacetylase SANT domain SWIRM, K-demethylase PHD finger dehydrogenase Class I deacetylase SANT domain WD40 repeat jumonji demethylase methyl CpG binding Rpd3 Ume1 Sin3 Protein domain class I deacetylase WD40 repeat PAH motifs coiled coil CoREST N-CoR/SMRT Adapted, with permission of Macmillan Publishing, from Yang and Seto (2008). There are four classes of HDACs in mammalian cells and S. cerevisiae (Yang and Seto 2008). They are classified based on the major S. cerevisiae deacetylases RPD3, HDA1, and SIR2. Class I in mammals includes HDAC1, 2, 3, and 8, which are related to Rpd3; class 2 includes HDAC4, 5, 6, 7, 9, and 10, which are related to HDA1. HDAC 11 shows features of both and is labeled class 4. Class 3 includes the sirtuins, which are relatives of the S. cerevisiae Sir2 protein involved in heterochromatin formation. In mammals, these include SIRT1–7. All sirtuins require NAD+ as a cofactor for deacetylation. Most HDACs exist as components of multisubunit complexes, which are recruited to a gene via interaction with one of the subunits of the HDAC complex (Table 1.4). In some cases, the DNAbinding proteins copurify with the complex and in other cases, they do not. Similarly, HDAC activity at various histones appears to be dependent in many cases on a subunit that provides the specific histone-targeting function. A well-characterized example of an HDAC complex is the Rpd3 complex, which contains key subunits that are largely conserved in organisms from S. cerevisiae to humans. In S. cerevisiae, this complex exists in a large form and small form. The large form consists of Rpd3, Sin3, Ume1, Pho23p, Sap30, Sds3, Cti6, Rxt2, Rxt3, and Dep1. Numerous sequence-specific repressors in S. cerevisiae can recruit the complex. For example, the sequence-specific repressors Ume6 and Ash1 directly recruit Rpd3L to target genes. As discussed above, Rpd3 also exists in a smaller complex, Rpd3S, involved in the Set2 pathway. In this case, the complex contains Rpd3, Sin3, Ume1, Rco1, and Eaf3. As discussed above, the Rpd3S complex is targeted by H3K36me3. The interaction of various sequence-specific DNA-binding proteins with the Rpd3 complex is frequently mediated by the Sin3 subunit, which is conserved between S. cerevisiae and mammals 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 37 A Primer on Transcriptional Regulation in Mammalian Cells / 37 (Silverstein and Ekwall 2005). In mammals, there are two related family members, termed Sin3a and Sin3b. Most interactions of DNA-binding proteins or corepressors with Sin3 are through the four conserved paired amphipathic helix (PAH) domains or a domain termed the histone deacetylase interaction domain. In mammalian cells, the Mad family of sequence-specific repressors (discussed in Chapter 10) contains a Sin3-interacting domain (SID), which directly contacts PAH2 and recruits a Sin3 complex containing HDACs 1 and 2 to the promoter. The S. cerevisiae Ume6 also contains a SID domain. Ume6 is both a repressor and activator of genes involved in meiosis. In the absence of a coactivator Ime1, Ume6 binds Sin3 and recruits Rpd3 to deacetylate histones and hence inactivate genes (Silverstein and Ekwall 2005). Conversely, in the presence of Ime1 and phosphorylation by the Rim11 kinase, Ume6 activates transcription. This type of switching of a DNA-binding factor between a repressor and activator is a common theme in transcription. Another example is the mammalian nuclear receptor family. Nuclear receptors (NRs) interact with an HDAC3-containing complex via intermediate factors termed corepressors (Silverstein and Ekwall 2005). In the unliganded state, many NRs interact with a corepressor termed N-CoR or a highly related protein termed SMRT (silencing mediator of retinoic acid and thyroid hormone receptor) (see Chapter 10). N-CoR and SMRT are in a complex with HDAC3 and several other proteins. In the absence of ligand, the NR is bound to its target site on DNA and the unliganded ligand-binding domain (LBD) binds N-CoR or SMRT along with HDAC3. Upon ligand binding, the LBD changes conformation and the corepressor can no longer bind. Instead, the LBD is in a conformation that can bind HATs such as p300/CBP and SRC family members, or the Mediator via the LXXLL motifs contained within the HATs and Mediator (see Chapter 10). Class 1 HDACs form a number of other well-characterized complexes involved in repression in numerous organisms. The Mi-2/NuRD complex contains HDACs 1 and 2, the retinoblastomaassociated proteins RbAP46 and RbAP48, the Mi-2 ATP-dependent remodeling subunit (also known as CHD3/4, see above), a subunit that recognizes methylated CpG (MBD2), and two other subunits, both of which bind histones (Yang and Seto 2008). NuRD has been associated with repression in many systems. CoREST is another complex containing HDACs 1 and 2 (Lakowski et al. 2006). CoREST binds to the REST-1 transcriptional repressor, a sequence-specific binding protein involved in shutting down genes involved in neuronal differentiation. CoREST also contains the amine oxidase LSD1, which can demethylate H3K4me2 and H3K4me1. Although we have focused on class 1 HDACs, class 2 HDACs also form multisubunit complexes, which can be targeted by DNA-binding proteins (Yang and Seto 2008). Some of these complexes are targeted by myocyte enhancer factor 2 (MEF2), a sequence-specific DNA-binding protein. Vertebrates contain four MEF2 family members involved in a variety of developmental and signaling processes, including heart development. HDACs are involved not only in the action of specific genes, but also in global processes. Rpd3 and Hda1 in S. cerevisiae are also involved in global histone deacetylation (Vogelauer et al. 2000). Additionally, HDACs are involved in heterochromatin assembly. The S. cerevisiae class 3 HDAC Sir2 interacts in a complex with Sir4, which is targeted to S. cerevisiae heterochromatin by the sequence-specific factor Rap1 (Grunstein 1997). Deacetylation creates a surface for Sir3 to bind and Sir3 then binds Sir2 and Sir4. The continued deacetylation of neighboring nucleosomes by Sir2 then allows spreading of the deacetylase across a region, forming a heterochromatic structure. H3K9 Methylation and HP1 One of the major lysine modifications of pericentric heterochromatin is trimethylation of H3K9 (Martin and Zhang 2005, 2007; Shilatifard 2006; Kouzarides 2007). As discussed above, Suv39H1 and Suv39H2 histone methyltransferases catalyze this modification. Suv39 was first isolated in D. melanogaster as a suppressor of position effect variegation, termed Su(Var)3-9. Heterochromatin is enriched in H3K9me3, and euchromatin is enriched in H3K9me2 but contains some H3K9me3. 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 38 38 / Chapter 1 Zigzag 30-nm filament HP1 HP1 RNAi or sequence-specific factor G9a or Suv39 H3K9me2 or me3 HP1 HP1 HP1 HP1 HP1 CD HP1 HP1 HP1 HP1 FIGURE 1.15. HP1 silencing mechanism. RNAi or sequence-specific DNA-binding proteins recruit either Suv39 in heterochromatin or G9a in euchromatin. Suv39 trimethylates H3K9 and G9a dimethylates H3K9. This modification provides a binding site for HP1 chromodomain (CD). HP1 binds di- and trimethylated H3K9. Note that there are two modified residues per nucleosome. The HP1 chromoshadow domain dimerizes with chromoshadow domains on adjacent nucleosomes allowing for the formation of the condensed zigzag two-start helix. A heterodimer of the highly related G9a and GLP proteins is the major euchromatic H3K9 dimethylase. SetDB or ESET is a H3K9 trimethylase found in heterochromatin and euchromatin and is targeted to specific zinc finger genes in a form of autoregulation, as many zinc finger repressors recruit SetDB via a corepressor termed KAP-1 (see Chapter 10). H3K9me2 and HEK9me3 serve as binding sites for the heterochromatin protein 1 (HP1) family of proteins (Fig. 1.15) (Lomberk et al. 2006). Three HP1 family members, termed α, β, and γ, exist in mammalian cells. HP1 is found in Schizosaccharomyces pombe, where it is called Swi6, and in many other eukaryotes, although not in S. cerevisiae, which lacks H3K9 methylation. HP1 contains three conserved regions: an amino-terminal chromodomain, a variable linker, and a carboxyterminal chromoshadow domain. The chromodomain binds directly to H3K9me2 and H3K9me3 via a hydrophobic pocket (Taverna et al. 2007). The chromoshadow domain is structurally similar to the chromodomain, but it lacks residues essential for binding the methylated histone. Instead, the chromoshadow domain allows HP1 family members to dimerize, which is important for chromatin binding by HP1. The dimerized chromoshadow domain also interacts with a number of proteins involved in heterochromatin assembly, including SUV39. Indeed, the SUV39–HP1 interaction is thought to provide a rudimentary mechanism for heterochromatic spreading similar to the Sir2/4 interactions in S. cerevisiae (Martin and Zhang 2005). HP1 also interacts with corepressors such as KAP1/TIF1β, which binds HP1 via a PXVXL motif. Remarkably, the chromoshadow also interacts with TAF4, a TFIID-specific TAF. The linker region is the most evolutionarily divergent segment of HP1 and interacts with DNA and RNA (a point we discuss below). HP1 binds H3K9me2 and H3K9me3 in the micromolar concentration range in vitro, and binding to heterochromatin in vivo is highly dynamic, as measured by fluorescence recovery after 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 39 A Primer on Transcriptional Regulation in Mammalian Cells / 39 photobleaching (FRAP) (Schmiedeberg et al. 2004). HP1α and HP1β localize primarily to pericentric heterochromatin and HP1γ localizes to both heterochromatin and euchromatin. That said, a number of genes are thought to be specifically regulated by HP1α and HP1β in euchromatin (Martin and Zhang 2005). HP1γ, as described above, is found in actively transcribing genes. HP1 prefers to bind to nucleosomal arrays containing H2AZ in vitro and helps those arrays form the zigzag structure characteristic of condensed chromatin. HP1 dimerization is necessary for formation of condensed structures. It is not entirely clear how H3K9 methylation is targeted in mammalian cells. Knockout of G9a or GLP leads to a dramatic reduction of H3K9me2 in silent regions of euchromatin (Rice et al. 2003). Hence, it has been proposed that G9a and GLP are recruited by sequence-specific binding proteins to euchromatic genes during repression, and a few examples of such recruitment are known. In contrast, many zinc finger DNA-binding proteins contain a KRAB domain, which interacts with KAP-1. KAP-1, in turn, binds SetDB1 (see Chapter 10). This event leads to trimethylation of H3K9 and recruitment of HP1. Some insights have emerged for how H3K9 trimethylation is targeted to heterochromatin in S. pombe and D. melanogaster (Grewal and Elgin 2007; Grewal and Jia 2007). The studies in S. pombe reveal that the RNA interference (RNAi) pathway is necessary for HP1 binding and heterochromatin formation, and correlative evidence suggests the same may be true for D. melanogaster. Paradoxically, the studies in these organisms imply that active transcription of DNA, usually repeats, is necessary for silencing. Two complexes termed the RNA-induced transcriptional silencing complex (RITS) and the RNA-directed RNA polymerase (RDRP) are key to H3K9 trimethylation in S. pombe. RITS is bound by small interfering RNA (siRNA) via its argonaute subunit ago1. These siRNAs presumably direct the complex to small transcripts originating from repeats in centromeric heterochromatin, and processing of these transcripts into siRNA and amplification of the transcripts by RDRP appear to be essential. It is unknown how RITS recruits Clr4, which is homologous to Suv39. However, RITS, Clr4, and Swi6(HP1) colocalize to most heterochromatin in S. pombe. Mammalian cells lack RDRP but contain small transcripts originating from heterochromatic repeats. In addition to RNAi, certain sequence-specific proteins that bind repetitive elements, mating-type regions, and centromeres have been associated with HP1 recruitment. H3K27 Methylation and Polycomb Trimethylation of H3K27 by the Polycomb protein EZH2 is yet another mechanism of silencing, which is estimated to operate on about 4% or less of human genes (Cheng et al. 2005; Daniel et al. 2005; Martin and Zhang 2005; Shilatifard 2006; Martin and Zhang 2007). EZH2 was first characterized in D. melanogaster as a protein necessary for stable silencing of Hox genes during embryonic stages of development. EZH2 is one of the founding members of the SET family of histone methylases. EZH2 is a subunit of the Polycomb repressive complex 2 (PRC2) along with EED, SUZ12, RbAP46, and RbAP48. Another PRC complex in D. melanogaster and mammals is called PRC1, which contains Ring1B, Ring1A, HPC1-3 (human Polycomb, or human Pc), HPH1-3, Bmi1, Mel18, and ScmH1-2. HPC contains a chromodomain with significant similarity to the chromodomain of HP1. However, structural studies show that subtle differences allow it to recognize H3K27me3 but not H3K9me3. In the simplest models of repression, PRC2 is targeted to chromatin by a DNA-binding protein, where it trimethylates H3K27. This modification, in turn, recruits PRC1, although it is likely that other factors contribute to binding. PRC1 is known to form a stable repressive structure with nucleosomes in vitro, which blocks access to the SWI/SNF remodeling complex. The PRC1 subunits Ring1A and Ring1B together with Bmi1 possess histone ubiquitinylation activity targeted to H2AK119. In D. melanogaster, Polycomb repressive elements (PREs) recruit PRC complexes to a gene (Schuettengruber et al. 2007). At this stage, no known PREs exist in mammals, although the field is 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 40 40 / Chapter 1 rapidly evolving; if they exist, they should be identified shortly. A number of sequence-specific binding proteins have been postulated to recruit PRC2 in D. melanogaster. The sequence-specific binding protein Pho binds and recruits PRC2 to a PRE controlling the Ubx gene in D. melanogaster. However, it also recruits PRC1 independently. In contrast to what one might expect, PREs in D. melanogaster are frequently several kilobases away from the promoter and thus the mechanism of repression is unclear, but it may involve chromatin loop formation. Additionally, PREs are thought to have few nucleosomes, and thus the concept of PRC2 modifications recruiting PRC1 via the HPC subunit seems to be less likely in some cases. The RNAi machinery along with noncoding RNAs such as hotair and Xist have been suggested to participate in recruitment of PRC2 along with several other sequence-specific DNAbinding proteins such as the GAGA factor and Pho. Genome-wide studies show large genomic regions that are trimethylated at H3K27 and a negative correlation with the presence of Pol II (Ringrose 2007; Ringrose and Paro 2007), although some inactive genes contain paused Pol II. Finally, genome-wide studies in mammals have shown considerable overlap between PRC1 and PRC2 complexes at silenced genes, suggesting that the two complexes work in concert in support of a model where PRC2 modifies the chromatin for PRC1 to bind. However, they may bind independently at some locations. DNA Methylation at CpG DNA in mammals, but not D. melanogaster or S. cerevisiae (at least to any significant extent), can be methylated at the C5 position of cytosine within the major groove of the sequence CpG (Klose and Bird 2006; Martin and Zhang 2007; Miranda and Jones 2007). CpG methylation plays a role in silencing. CpG methylation is catalyzed by three major DNA methyltransferases termed DNMT1, DNMT3A, and DNMT3B that use S-adenosylmethionine (SAM) as a cofactor to donate the methyl group to cytosine. DNMT3L binds DNMT3A and DNMT3B and stimulates their activities. There are two classes of DNMTs, maintenance and de novo. DNMT1 is thought to be the major maintenance methylase and converts hemimethylated DNA to fully methylated DNA at the DNA replication fork. DNMT3A and DNMT3B are de novo methyltransferases and are recruited to genes to enact gene silencing. Promoters bearing CpG islands are not normally methylated, but they become methylated in some disease states such as cancer (Miranda and Jones 2007). Aberrant methylation is one way in which tumor suppressor genes become silenced during cancer progression. Additionally, CpG-rich promoters are methylated on the inactive X chromosome. Because DNA methylation can be inherited, it is thought to be a more efficient mechanism of silencing than repressive histone methylation. Indeed, repressive histone methylation appears to be linked in many cases with DNA methylation, suggesting a functional interrelationship in which proteins binding to histone modifications such as HP1 and PRC1 might recruit DNMTs and vice versa. Some factors such as Myc and the promyelocytic leukemia–retinoic acid receptor-α (PML–RARα) fusion are thought to recruit DNMT3 directly. It is also believed that the RNAi machinery may help recruit DNMTs. Several models describe how DNA methylation leads to repression (Klose and Bird 2006). In one model, the methylation inhibits transcription factor binding to the major groove of the factor’s binding site. This has been shown in vitro and in vivo for a small set of factors, including for the insulator protein CTCF at the differentially methylated region (DMR) in the Igf2/H19 imprinted locus. In the second model, CpG methylation is recognized by methyl-binding proteins, which recruit HDAC complexes (Fatemi and Wade 2006; Clouaire and Stancheva 2008). There are currently five proteins containing a conserved 75-amino-acid methyl-binding domain (MBD) (Fatemi and Wade 2006; Clouaire and Stancheva 2008). These proteins are termed MeCP2, MBD1, MBD2, MBD3, and MBD4. In mammals, MBD3 contains a mutation that prevents binding to methylated CpG, but this mutation is not present in other vertebrates (i.e., Xenopus). In addition to the MBD-containing proteins, a second family of proteins—the Kaiso family—apparently recognizes methyl CpG via a different motif. There are three members of this family, all of which contain 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 41 A Primer on Transcriptional Regulation in Mammalian Cells / 41 three C2H2 zinc fingers, two of which are required for binding methylated DNA. However, unmethylated sequences that bind Kaiso have also been found, suggesting that this family has dual roles. A large portion of CpGs are methylated in mammals, raising the question of whether the MBDs display any specificity. The phenotypes of knockout mice, while subtle, are distinct. ChIP-Seq of DNA bound to MBD2 and MeCP2 in vivo revealed nonoverlapping specificities. Although this could result from unique protein–protein interactions of MBD2 and MeCP2, both proteins contain additional domains that might recognize DNA and confer selectivity. For example, MeCP2 has two AT-hook domains, which bind AT-rich regions and are found in chromatin-remodeling proteins and HMGA family proteins. Indeed, MeCP2 seems to bind methylated CpG adjacent to AT-rich DNA. The solution structures of several MBDs are known (Clouaire and Stancheva 2008), and the crystal structure of MeCP2 with methylated DNA has recently been solved (Ho et al. 2008). Key questions include how these methyl-binding proteins repress transcription and/or whether they are involved in other processes. MeCP2 has probably been the most extensively studied member of the family. Different studies have led to conflicting results as to whether MeCP2 is part of a larger complex. Either way, however, MeCP2 has been clearly associated with recruitment of the Sin3A/HDAC2 complexes and the Brm1 form of mammalian SWI/SNF. As described above, MBD2 is a subunit of the Mi-2/NuRD complex. Additionally, SETDB1 and SUV39 have been found to be associated with MDB1 and Kaiso copurifies with NCoR. Thus, there is certainly a relationship of methyl CpG-binding proteins with corepressor complexes. Finally, MeCP2 is highly effective at chromatin compaction in vitro. It binds to the entry and exit points of a nucleosome in vitro and apparently multimerizes to form a condensed structure distinct from the zigzag structure promoted by HP1. Paradoxically, the idea that MeCP2 is solely a repressor was challenged recently when MeCP2 was found at actively transcribed genes in mouse hypothalamus, genes whose activity decreases in MeCP2 knockout mice (Cohen et al. 2008). CONCLUDING PERSPECTIVE Technology advancements during the past decade have led to an explosion of information on eukaryotic gene regulation. It should be apparent from the synopsis above that we know much about the machinery and modifications; however, the precise mechanisms, with the exception of some basic events such as Pol II initiation, are unclear. An amazing amount of descriptive information has emerged about the events linking chromatin and gene control. However, it is important to realize that these observations represent a means to frame a hypothesis. The spirit of the ensuing chapters is to provide a student in the field with the strategies and concepts to begin rigorously testing ideas using the scientific method. REFERENCES Albright S.R. and Tjian R. 2000. TAFs revisited: More data reveal new twists and confirm old ideas. Gene 242: 1–13. Allis C.D., Berger S.L., Cote J., Dent S., Jenuwien T., Kouzarides T., Pillus L., Reinberg D., Shi Y., Shiekhattar R., et al. 2007. New nomenclature for chromatin-modifying enzymes. Cell 131: 633–636. Anand R. and Marmorstein R. 2007. Structure and mechanism of lysine-specific demethylase enzymes. J. Biol. Chem. 282: 35425– 35429. Andel F., 3rd, Ladurner A.G., Inouye C., Tjian R., and Nogales E. 1999. Three-dimensional structure of the human TFIID-IIA-IIB complex. Science 286: 2153–2156. Armstrong J.A. 2007. Negotiating the nucleosome: Factors that allow RNA polymerase II to elongate through chromatin. Biochem. Cell Biol. 85: 426–434. Bao Y. and Shen X. 2007a. INO80 subfamily of chromatin remodeling complexes. Mutat. Res. 618: 18–29. Bao Y. and Shen X. 2007b. SnapShot: Chromatin remodeling complexes. Cell 129: 632. Barski A., Cuddapah S., Cui K., Roh T.Y., Schones D.E., Wang Z., Wei G., Chepelev I., and Zhao K. 2007. High-resolution profiling of histone methylations in the human genome. Cell 129: 823–837. Bernstein E. and Hake S.B. 2006. The nucleosome: A little variation goes a long way. Biochem. Cell Biol. 84: 505–517. Biddick R. and Young E.T. 2005. Yeast mediator and its role in transcriptional regulation. C.R. Biol. 328: 773–782. Bortvin A. and Winston F. 1996. Evidence that Spt6p controls chromatin structure by a direct interaction with histones. Science 272: 1473–1476. Bourbon H.M. 2008. Comparative genomics supports a deep evolu- 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 42 42 / Chapter 1 tionary origin for the large, four-module transcriptional mediator complex. Nucleic Acids Res. 36: 3993–4008. Branco M.R. and Pombo A. 2006. Intermingling of chromosome territories in interphase suggests role in translocations and transcription-dependent associations. PLoS Biol. 4: e138. Cai Y., Jin J., Yao T., Gottschalk A.J., Swanson S.K., Wu S., Shi Y., Washburn M.P., Florens L., Conaway R.C., and Conaway J.W. 2007. YY1 functions with INO80 to activate transcription. Nat. Struct. Mol. Biol. 14: 872–874. Cairns B.R. 2005. Chromatin remodeling complexes: Strength in diversity, precision through specialization. Curr. Opin. Genet. Dev. 15: 185–190. Carey M. 1998. The enhanceosome and transcriptional synergy. Cell 92: 5–8. Carter D.R., Eskiw C., and Cook P.R. 2008. Transcription factories. Biochem. Soc. Trans. 36: 585–589. Chadick J.Z. and Asturias F.J. 2005. Structure of eukaryotic Mediator complexes. Trends Biochem. Sci. 30: 264–271. Chakalova L., Carter D., and Fraser P. 2004. RNA fluorescence in situ hybridization tagging and recovery of associated proteins to analyze in vivo chromatin interactions. Methods Enzymol. 375: 479–493. Chakravarthy S., Park Y.J., Chodaparambil J., Edayathumangalam R.S., and Luger K. 2005. Structure and dynamic properties of nucleosome core particles. FEBS Lett. 579: 895–898. Chédin S., Laferté A., Hoang T., Lafontaine D L., Riva M., and Carles C. 2007. Is ribosome synthesis controlled by Pol I transcription? Cell Cycle 6: 11–15. Chen B.S. and Hampsey M. 2002. Transcription activation: Unveiling the essential nature of TFIID. Curr. Biol. 12: R620–R622. Chen H.T. and Hahn S. 2003. Binding of TFIIB to RNA polymerase II: Mapping the binding site for the TFIIB zinc ribbon domain within the preinitiation complex. Mol. Cell 12: 437–447. Chen H.T. and Hahn S. 2004. Mapping the location of TFIIB within the RNA polymerase II transcription preinitiation complex: A model for the structure of the PIC. Cell 119: 169–180. Chen M. and Shen X. 2007. Nuclear actin and actin-related proteins in chromatin dynamics. Curr. Opin. Cell Biol. 19: 326–330. Cheng X., Collins R.E., and Zhang X. 2005. Structural and sequence motifs of protein (histone) methylation enzymes. Annu. Rev. Biophys. Biomol. Struct. 34: 267–294. Chi T., Lieberman P., Ellwood K., and Carey M. 1995. A general mechanism for transcriptional synergy by eukaryotic activators. Nature 377: 254–257. Clouaire T. and Stancheva I. 2008. Methyl-CpG binding proteins: Specialized transcriptional repressors or structural components of chromatin? Cell. Mol. Life Sci. 65: 1509–1522. Cohen S., Zhou Z., and Greenberg M.E. 2008. Medicine. Activating a repressor. Science 320: 1172–1173. Collas P. and Dahl J.A. 2008. Chop it, ChIP it, check it: The current status of chromatin immunoprecipitation. Front. Biosci. 13: 929–943. Conaway R.C., Sato S., Tomomori-Sato C., Yao T., and Conaway J.W. 2005. The mammalian Mediator complex and its role in transcriptional regulation. Trends Biochem. Sci. 30: 250–255. Cosma M.P. 2002. Ordered recruitment: Gene-specific mechanism of transcription activation. Mol. Cell 10: 227–236. Cramer P., Armache K.J., Baumli S., Benkert S., Brueckner F., Buchen C., Damsma G.E., Dengl S., Geiger S.R., Jasiak A.J., et al. 2008. Structure of eukaryotic RNA polymerases. Annu. Rev. Biophys. 37: 337–352. Crawford G.E., Davis S., Scacheri P.C., Renaud G., Halawi M.J., Erdos M.R., Green R., Meltzer P.S., Wolfsberg T.G., and Collins F.S. 2006a. DNase-chip: A high-resolution method to identify DNase I hypersensitive sites using tiled microarrays. Nat. Methods 3: 503–509. Crawford G.E., Holt I.E., Whittle J., Webb B.D., Tai D., Davis S., Margulies E.H., Chen Y., Bernat J.A., Ginsburg D., et al. 2006b. Genome-wide mapping of DNase hypersensitive sites using massively parallel signature sequencing (MPSS). Genome Res. 16: 123–131. Cremer T., Cremer M., Dietzel S., Muller S., Solovei I., and Fakan S. 2006. Chromosome territories—A functional nuclear landscape. Curr. Opin. Cell Biol. 18: 307–316. Daniel J.A., Pray-Grant M.G., and Grant P.A. 2005. Effector proteins for methylated histones: An expanding family. Cell Cycle 4: 919– 926. Davidson I., Kobi D., Fadloun A., and Mengus G. 2005. New insights into TAFs as regulators of cell cycle and signaling pathways. Cell Cycle 4: 1486–1490. Dean A. 2006. On a chromosome far, far away: LCRs and gene expression. Trends Genet. 22: 38–45. Dechassa M.L., Zhang B., Horowitz-Scherer R., Persinger J., Woodcock C.L., Peterson C.L., and Bartholomew B. 2008. Architecture of the SWI/SNF-nucleosome complex. Mol. Cell Biol. 28: 6010–6012. de Laat W. and Grosveld F. 2003. Spatial organization of gene expression: The active chromatin hub. Chromosome Res. 11: 447–459. Dieci G., Fiorino G., Castelnuovo M., Teichmann M., and Pagano A. 2007. The expanding RNA polymerase III transcriptome. Trends Genet. 23: 614–622. Ding N., Zhou H., Esteve P.O., Chin H.G., Kim S., Xu X., Joseph S.M., Friez M.J., Schwartz C.E., Pradhan S., and Boyer T.G. 2008. Mediator links epigenetic silencing of neuronal gene expression with X-linked mental retardation. Mol. Cell 31: 347–359. Dorman E.R., Bushey A.M., and Corces V.G. 2007. The role of insulator elements in large-scale chromatin structure in interphase. Semin. Cell Dev. Biol. 18: 682–690. Dyda F., Klein D.C., and Hickman A.B. 2000. GCN5-related N-acetyltransferases: A structural overview. Annu. Rev. Biophys. Biomol. Struct. 29: 81–103. Eirín-López J. and Ausió J. 2007. H2A.Z-mediated genome-wide chromatin specialization. Curr. Genomics 8: 59–66. Eissenberg J.C. and Shilatifard A. 2006. Leaving a mark: The many footprints of the elongating RNA polymerase II. Curr. Opin. Genet. Dev. 16: 184–190. The ENCODE Project Consortium. 2004. The ENCODE (ENCyclopedia Of DNA Elements) Project. Science 306: 636– 640. Fatemi M. and Wade P.A. 2006. MBD family proteins: Reading the epigenetic code. J. Cell Sci. 119: 3033–3037. Fazzio T.G., Kooperberg C., Goldmark J.P., Neal C., Basom R., Delrow J., and Tsukiyama T. 2001. Widespread collaboration of Isw2 and Sin3-Rpd3 chromatin remodeling complexes in transcriptional repression. Mol. Cell. Biol. 21: 6450–6460. Ferrari R., Pellegrini M., Horwitz G.A., Xie W., Berk A.J., and Kurdistani S.K. 2008. Epigenetic reprogramming by adenovirus e1a. Science 321: 1086–1088. Ferreira H., Flaus A., and Owen-Hughes T. 2007. Histone modifications influence the action of Snf2 family remodelling enzymes by different mechanisms. J. Mol. Biol. 374: 563–579. Fu Y., Sinha M., Peterson C.L., and Weng Z. 2008. The insulator binding protein CTCF positions 20 nucleosomes around its binding sites across the human genome. PLoS Genet. 4: e1000138. Galande S., Purbey P.K., Notani D., and Kumar P.P. 2007. The third dimension of gene regulation: Organization of dynamic chromatin loopscape by SATB1. Curr. Opin. Genet. Dev. 17: 408–414. Grewal S.I. and Elgin S.C. 2007. Transcription and RNA interference in the formation of heterochromatin. Nature 447: 399–406. Grewal S.I. and Jia S. 2007. Heterochromatin revisited. Nat. Rev. Genet. 8: 35–46. Grob P., Cruse M.J., Inouye C., Peris M., Penczek P.A., Tjian R., and Nogales E. 2006. Cryo-electron microscopy studies of human TFIID: Conformational breathing in the integration of gene regulatory cues. Structure 14: 511–520. Grunstein M. 1997. Molecular model for telomeric heterochromatin in yeast. Curr. Opin. Cell Biol. 9: 383–387. Haeusler R.A. and Engelke D.R. 2006. Spatial organization of transcription by RNA polymerase III. Nucleic Acids Res. 34: 4826–4836. Hahn S. 1998. The role of TAFs in RNA polymerase II transcription. 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 43 A Primer on Transcriptional Regulation in Mammalian Cells / 43 Cell 95: 579–582. Hahn S. 2004. Structure and mechanism of the RNA polymerase II transcription machinery. Nat. Struct. Mol. Biol. 11: 394–403. Hake S.B. and Allis C.D. 2006. Histone H3 variants and their potential role in indexing mammalian genomes: The “H3 barcode hypothesis.” Proc. Natl. Acad. Sci. 103: 6428–6435. Hall J.A. and Georgel P.T. 2007. CHD proteins: A diverse family with strong ties. Biochem. Cell Biol. 85: 463–476. Hatzis P. and Talianidis I. 2002. Dynamics of enhancer–promoter communication during differentiation-induced gene activation. Mol. Cell 10: 1467–1477. Heintzman N.D., Stuart R.K., Hon G., Fu Y., Ching C.W., Hawkins R.D., Barrera L.O., Van Calcar S., Qu C., Ching K.A., et al. 2007. Distinct and predictive chromatin signatures of transcriptional promoters and enhancers in the human genome. Nat. Genet. 39: 311–318. Ho K.L., McNae I.W., Schmiedeberg L., Klose R.J., Bird A.P., and Walkinshaw M.D. 2008. MeCP2 binding to DNA depends upon hydration at methyl-CpG. Mol. Cell 29: 525–531. Hoch D.A., Stratton J.J., and Gloss L.M. 2007. Protein-protein Forster resonance energy transfer analysis of nucleosome core particles containing H2A and H2A.Z. J. Mol. Biol. 371: 971–988. Hock R., Furusawa T., Ueda T., and Bustin M. 2007. HMG chromosomal proteins in development and disease. Trends Cell Biol. 17: 72–79. Hoffmann A., Chiang C.M., Oelgeschlager T., Xie X., Burley S.K., Nakatani Y., and Roeder R.G. 1996. A histone octamer-like structure within TFIID. Nature 380: 356–359. Holstege F.C., Tantin D., Carey M., van der Vliet P.C., and Timmers H.T. 1995. The requirement for the basal transcription factor IIE is determined by the helical stability of promoter DNA. EMBO J. 14: 810–819. Horn P.J. and Peterson C.L. 2002. Molecular biology. Chromatin higher order folding—Wrapping up transcription. Science 297: 1824–1827. Jacobson R.H., Ladurner A.G., King D.S., and Tjian R. 2000. Structure and function of a human TAFII250 double bromodomain module. Science 288: 1422–1425. Jin J., Cai Y., Li B., Conaway R.C., Workman J.L., Conaway J.W., and Kusch T. 2005. In and out: Histone variant exchange in chromatin. Trends Biochem. Sci. 30: 680–687. Johnson K.M., Wang J., Smallwood A., Arayata C., and Carey M. 2002. TFIID and human mediator coactivator complexes assemble cooperatively on promoter DNA. Genes Dev. 16: 1852–1863. Kadauke S. and Blobel G.A. 2008. Chromatin loops in gene regulation. Biochim. Biophys. Acta (in press). Kaplan C.D., Laprade L., and Winston F. 2003. Transcription elongation factors repress transcription initiation from cryptic sites. Science 301: 1096–1099. Kim T.H. and Ren B. 2006. Genome-wide analysis of protein-DNA interactions. Annu. Rev. Genomics Hum. Genet. 7: 81–102. Kim T.K., Ebright R.H., and Reinberg D. 2000. Mechanism of ATPdependent promoter melting by transcription factor IIH. Science 288: 1418–1422. Kim T.H., Abdullaev Z.K., Smith A.D., Ching K.A., Loukinov D.I., Green R.D., Zhang M.Q., Lobanenkov V.V., and Ren B. 2007. Analysis of the vertebrate insulator protein CTCF-binding sites in the human genome. Cell 128: 1231–1245. Kim T.H., Barrera L.O., Zheng M., Qu C., Singer M.A., Richmond T.A., Wu Y., Green R.D., and Ren B. 2005. A high-resolution map of active promoters in the human genome. Nature 436: 876–880. Klose R.J. and Bird A.P. 2006. Genomic DNA methylation: the mark and its mediators. Trends Biochem. Sci. 31: 89–97. Konev A.Y., Tribus M., Park S.Y., Podhraski V., Lim C.Y., Emelyanov A.V., Vershilova E., Pirrotta V., Kadonaga J.T., Lusser A., and Fyodorov D.V. 2007. CHD1 motor protein is required for deposition of histone variant H3.3 into chromatin in vivo. Science 317: 1087–1090. Korber P. and Hörz W. 2004. SWRred not shaken: Mixing the histones. Cell 117: 5–7. Kornberg R.D. 2005. Mediator and the mechanism of transcriptional activation. Trends Biochem. Sci. 30: 235–239. Kornberg R.D. 2007. The molecular basis of eukaryotic transcription. Proc. Natl. Acad. Sci. 104: 12955–12961. Kouzarides T. 2007. Chromatin modifications and their function. Cell 128: 693–705. Kurdistani S.K., Tavazoie S., and Grunstein M. 2004. Mapping global histone acetylation patterns to gene expression. Cell 117: 721–733. Kusch T. and Workman J.L. 2007. Histone variants and complexes involved in their exchange. Subcell. Biochem. 41: 91–109. Lakowski B., Roelens I., and Jacob S. 2006. CoREST-like complexes regulate chromatin modification and neuronal gene expression. J. Mol. Neurosci. 29: 227–239. Lan F., Nottke A.C., and Shi Y. 2008. Mechanisms involved in the regulation of histone lysine demethylases. Curr. Opin. Cell Biol. 20: 316–325. Lee K.K. and Workman J.L. 2007. Histone acetyltransferase complexes: One size doesn’t fit all. Nat. Rev. Mol. Cell Biol. 8: 284–295. Lefstin J.A. and Yamamoto K.R. 1998. Allosteric effects of DNA on transcriptional regulators. Nature 392: 885–888. LeRoy G., Rickards B., and Flint S.J. 2008. The double bromodomain proteins Brd2 and Brd3 couple histone acetylation to transcription. Mol. Cell 30: 51–60. Leschziner A.E., Saha A., Wittmeyer J., Zhang Y., Bustamante C., Cairns B.R., and Nogales E. 2007. Conformational flexibility in the chromatin remodeler RSC observed by electron microscopy and the orthogonal tilt reconstruction method. Proc. Natl. Acad. Sci. 104: 4913–4918. Leurent C., Sanders S., Ruhlmann C., Mallouh V., Weil P.A., Kirschner D.B., Tora L., and Schultz P. 2002. Mapping histone fold TAFs within yeast TFIID. EMBO J. 21: 3424–3433. Li B., Carey M., and Workman J.L. 2007a. The role of chromatin during transcription. Cell 128: 707–719. Li A.G., Piluso L.G., Cai X., Gadd B.J., Ladurner A.G., and Liu X. 2007b. An acetylation switch in p53 mediates holo-TFIID recruitment. Mol. Cell 28: 408–421. Li B., Gogol M., Carey M., Lee D., Seidel C., and Workman J.L. 2007c. Combined action of PHD and chromo domains directs the Rpd3S HDAC to transcribed chromatin. Science 316: 1050–1054. Li B., Gogol M., Carey M., Pattenden S.G., Seidel C., and Workman J.L. 2007d. Infrequently transcribed long genes depend on the Set2/Rpd3S pathway for accurate transcription. Genes Dev. 21: 1422–1430. Linnemann A.K., Platts A.E., Doggett N., Gluch A., Bode J., and Krawetz S.A. 2007. Genomewide identification of nuclear matrix attachment regions: An analysis of methods. Biochem. Soc. Trans. 35: 612–617. Liu Y., Kung C., Fishburn J., Ansari A.Z., Shokat K.M., and Hahn S. 2004. Two cyclin-dependent kinases promote RNA polymerase II transcription and formation of the scaffold complex. Mol. Cell Biol. 24: 1721–1735. Lomberk G., Wallrath L., and Urrutia R. 2006. The heterochromatin protein 1 family. Genome Biol. 7: 228. Lorch Y., Maier-Davis B., and Kornberg R.D. 2006. Chromatin remodeling by nucleosome disassembly in vitro. Proc. Natl. Acad. Sci. 103: 3090–3093. Loyola A. and Almouzni G. 2007. Marking histone H3 variants: How, when and why? Trends Biochem. Sci. 32: 425–433. Luger K. and Hansen J.C. 2005. Nucleosome and chromatin fiber dynamics. Curr. Opin. Struct. Biol. 15: 188–196. Maeda R.K. and Karch F. 2007. Making connections: Boundaries and insulators in Drosophila. Curr. Opin. Genet. Dev. 17: 394–399. Mahajan M.C., Karmakar S., and Weissman S.M. 2007. Control of β globin genes. J. Cell. Biochem. 102: 801–810. Malik S. and Roeder R.G. 2005. Dynamic regulation of pol II transcription by the mammalian Mediator complex. Trends Biochem. Sci. 30: 256–263. Margaritis T. and Holstege F.C. 2008. Poised RNA polymerase II gives 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 44 44 / Chapter 1 pause for thought. Cell 133: 581–584. Martin C. and Zhang Y. 2005. The diverse functions of histone lysine methylation. Nat. Rev. Mol. Cell Biol. 6: 838–849. Martin C. and Zhang Y. 2007. Mechanisms of epigenetic inheritance. Curr. Opin. Cell Biol. 19: 266–272. Maston G.A., Evans S.K., and Green M.R. 2006. Transcriptional regulatory elements in the human genome. Annu. Rev. Genomics Hum. Genet. 7: 29–59. Matangkasombut O., Buratowski R.M., Swilling N.W., and Buratowski S. 2000. Bromodomain factor 1 corresponds to a missing piece of yeast TFIID. Genes Dev. 14: 951–962. McBryant S.J., Adams V.H., and Hansen J.C. 2006. Chromatin architectural proteins. Chromosome Res. 14: 39–51. Mellor J. and Morillon A. 2004. ISWI complexes in Saccharomyces cerevisiae. Biochim. Biophys. Acta 1677: 100–112. Merika M. and Thanos D. 2001. Enhanceosomes. Curr. Opin. Genet. Dev. 11: 205–208. Mihardja S., Spakowitz A.J., Zhang Y., and Bustamante C. 2006. Effect of force on mononucleosomal dynamics. Proc. Natl. Acad. Sci. 103: 15871–15876. Miller G. and Hahn S. 2006. A DNA-tethered cleavage probe reveals the path for promoter DNA in the yeast preinitiation complex. Nat. Struct. Mol. Biol. 13: 603–610. Miranda T.B. and Jones P.A. 2007. DNA methylation: The nuts and bolts of repression. J. Cell Physiol. 213: 384–390. Mitchell J.A. and Fraser P. 2008. Transcription factories are nuclear subcompartments that remain in the absence of transcription. Genes Dev. 22: 20–25. Mohrmann L. and Verrijzer C.P. 2005. Composition and functional specificity of SWI2/SNF2 class chromatin remodeling complexes. Biochim. Biophys. Acta 1681: 59–73. Mujtaba S., Zeng L., and Zhou M.M. 2007. Structure and acetyllysine recognition of the bromodomain. Oncogene 26: 5521– 5527. Oelgeschlager T., Chiang C.M., and Roeder R.G. 1996. Topology and reorganization of a human TFIID-promoter complex. Nature 382: 735–738. Paoletti A.C., Parmely T.J., Tomomori-Sato C., Sato S., Zhu D., Conaway R C., Conaway J.W., Florens L., and Washburn M.P. 2006. Quantitative proteomic analysis of distinct mammalian Mediator complexes using normalized spectral abundance factors. Proc. Natl. Acad. Sci. 103: 18928–18933. Parelho V., Hadjur S., Spivakov M., Leleu M., Sauer S., Gregson H.C., Jarmuz A., Canzonetta C., Webster Z., Nesterova T., et al. 2008. Cohesins functionally associate with CTCF on mammalian chromosome arms. Cell 132: 422–433. Park Y.J. and Luger K. 2006. Structure and function of nucleosome assembly proteins. Biochem. Cell Biol. 84: 549–558. Pavri R., Zhu B., Li G., Trojer P., Mandal S., Shilatifard A., and Reinberg D. 2006. Histone H2B monoubiquitination functions cooperatively with FACT to regulate elongation by RNA polymerase II. Cell 125: 703–717. Peterlin B.M. and Price D.H. 2006. Controlling the elongation phase of transcription with P-TEFb. Mol. Cell 23: 297–305. Peterson C.L. and Laniel M.A. 2004. Histones and histone modifications. Curr. Biol. 14: R546–R551. Peterson C.L. and Workman J.L. 2000. Promoter targeting and chromatin remodeling by the SWI/SNF complex. Curr. Opin. Genet. Dev. 10: 187–192. Phatnani H.P. and Greenleaf A.L. 2006. Phosphorylation and functions of the RNA polymerase II CTD. Genes Dev. 20: 2922–2936. Ptashne M. and Gann A.A. 1990. Activators and targets. Nature 346: 329–331. Qiu H., Hu C., Wong C.M., and Hinnebusch A.G. 2006. The Spt4p subunit of yeast DSIF stimulates association of the Paf1 complex with elongating RNA polymerase II. Mol. Cell. Biol. 26: 3135–3148. Reina J.H. and Hernandez N. 2007. On a roll for new TRF targets. Genes Dev. 21: 2855–2860. Reinberg D. and Sims R.J., 3rd. 2006. de FACTo nucleosome dynamics. J. Biol. Chem. 281: 23297–23301. Reinke H. and Horz W. 2003. Histones are first hyperacetylated and then lose contact with the activated PHO5 promoter. Mol. Cell 11: 1599–1607. Rhoades A.R., Ruone S., and Formosa T. 2004. Structural features of nucleosomes reorganized by yeast FACT and its HMG box component, Nhp6. Mol. Cell. Biol. 24: 3907–3917. Rice J.C., Briggs S.D., Ueberheide B., Barber C.M., Shabanowitz J., Hunt D.F., Shinkai Y., and Allis C.D. 2003. Histone methyltransferases direct different degrees of methylation to define distinct chromatin domains. Mol. Cell 12: 1591–1598. Richmond T.J. 2006. Genomics: Predictable packaging. Nature 442: 750–752. Ringrose L. 2007. Polycomb comes of age: Genome-wide profiling of target sites. Curr. Opin. Cell Biol. 19: 290–297. Ringrose L. and Paro R. 2007. Polycomb/Trithorax response elements and epigenetic memory of cell identity. Development 134: 223–232. Roeder R.G. 2005. Transcriptional regulation and the role of diverse coactivators in animal cells. FEBS Lett. 579: 909–915. Roh T.Y., Wei G., Farrell C.M., and Zhao K. 2007. Genome-wide prediction of conserved and nonconserved enhancers by histone acetylation patterns. Genome Res. 17: 74–81. Russell J. and Zomerdijk J.C. 2006. The RNA polymerase I transcription machinery. Biochem. Soc. Symp. 73: 203–216. Saha A., Wittmeyer J., and Cairns B.R. 2006a. Chromatin remodelling: The industrial revolution of DNA around histones. Nat. Rev. Mol. Cell Biol. 7: 437–447. Saha A., Wittmeyer J., and Cairns B.R. 2006b. Mechanisms for nucleosome movement by ATP-dependent chromatin remodeling complexes. Results Probl. Cell Differ. 41: 127–148. Saunders A., Core L.J., and Lis J.T. 2006. Breaking barriers to transcription elongation. Nat. Rev. Mol. Cell Biol. 7: 557–567. Scarpulla R.C. 2008. Transcriptional paradigms in mammalian mitochondrial biogenesis and function. Physiol. Rev. 88: 611–638. Schalch T., Duda S., Sargent D.F., and Richmond T.J. 2005. X-ray structure of a tetranucleosome and its implications for the chromatin fibre. Nature 436: 138–141. Schmiedeberg L., Weisshart K., Diekmann S., Meyer Zu Hoerste G., and Hemmerich P. 2004. High- and low-mobility populations of HP1 in heterochromatin of mammalian cells. Mol. Biol. Cell 15: 2819–2833. Schuettengruber B., Chourrout D., Vervoort M., Leblanc B., and Cavalli G. 2007. Genome regulation by polycomb and trithorax proteins. Cell 128: 735–745. Segal E., Fondufe-Mittendorf Y., Chen L., Thastrom A., Field Y., Moore, I.K., Wang J.P., and Widom J. 2006. A genomic code for nucleosome positioning. Nature 442: 772–778. Sexton T., Umlauf D., Kurukuti S., and Fraser P. 2007. The role of transcription factories in large-scale structure and dynamics of interphase chromatin. Semin. Cell Dev. Biol. 18: 691–697. Shao H., Revach M., Moshonov S., Tzuman Y., Gazit K., Albeck S., Unger T., and Dikstein R. 2005. Core promoter binding by histone-like TAF complexes. Mol. Cell. Biol. 25: 206–219. Shi Y. and Whetstine J.R. 2007. Dynamic regulation of histone lysine methylation by demethylases. Mol. Cell 25: 1–14. Shilatifard A. 2006. Chromatin modifications by methylation and ubiquitination: Implications in the regulation of gene expression. Annu. Rev. Biochem. 75: 243–269. Shilatifard A. 2008. Molecular implementation and physiological roles for histone H3 lysine 4 (H3K4) methylation. Curr. Opin. Cell Biol. 20: 341–348. Shilatifard A., Conaway R.C., and Conaway J.W. 2003. The RNA polymerase II elongation complex. Annu. Rev. Biochem. 72: 693–715. Shogren-Knaak M., Ishii H., Sun J.M., Pazin M.J., Davie J.R., and Peterson C.L. 2006. Histone H4-K16 acetylation controls chromatin structure and protein interactions. Science 311: 844–847. 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 45 A Primer on Transcriptional Regulation in Mammalian Cells / 45 Silverstein R.A. and Ekwall K. 2005. Sin3: A flexible regulator of global gene expression and genome stability. Curr. Genet. 47: 1–17. Simonis M., Kooren J., and de Laat W. 2007. An evaluation of 3C-based methods to capture DNA interactions. Nat. Methods 4: 895–901. Smale S.T. and Kadonaga J.T. 2003. The RNA polymerase II core promoter. Annu. Rev. Biochem. 72: 449–479. Smith C.L. 2008. A shifting paradigm: Histone deacetylases and transcriptional activation. Bioessays 30: 15–24. Smith C.L., Horowitz-Scherer R., Flanagan J.F., Woodcock C.L., and Peterson C.L. 2003. Structural analysis of the yeast SWI/SNF chromatin remodeling complex. Nat. Struct. Biol. 10: 141–145. Soutoglou E. and Talianidis I. 2002. Coordination of PIC assembly and chromatin remodeling during differentiation-induced gene activation. Science 295: 1901–1904. Struhl K. 2005. Transcriptional activation: Mediator can act after preinitiation complex formation. Mol. Cell 17: 752–754. Svejstrup J.Q. 2007a. Contending with transcriptional arrest during RNAPII transcript elongation. Trends Biochem. Sci. 32: 165–171. Svejstrup J.Q. 2007b. Elongator complex: How many roles does it play? Curr. Opin. Cell Biol. 19: 331–336. Taatjes D.J., Marr M.T., and Tjian R. 2004a. Regulatory diversity among metazoan coactivator complexes. Nat. Rev. Mol. Cell Biol. 5: 403–410. Taatjes D.J., Schneider-Poetsch T., and Tjian R. 2004b. Distinct conformational states of nuclear receptor-bound CRSP-Med complexes. Nat. Struct. Mol. Biol. 11: 664–671. Taverna S.D., Li H., Ruthenburg A.J., Allis C.D., and Patel D.J. 2007. How chromatin-binding modules interpret histone modifications: Lessons from professional pocket pickers. Nat. Struct. Mol. Biol. 14: 1025–1040. Tora L. 2002. A unified nomenclature for TATA box binding protein (TBP)-associated factors (TAFs) involved in RNA polymerase II transcription. Genes Dev. 16: 673–675. Torres-Padilla M.E. and Tora L. 2007. TBP homologues in embryo transcription: Who does what? EMBO Rep. 8: 1016–1018. Treisman R. 1996. Regulation of transcription by MAP kinase cascades. Curr. Opin. Cell Biol. 8: 205–215. Tremethick D.J. 2007. Higher-order structures of chromatin: The elusive 30 nm fiber. Cell 128: 651–654. Trojer P. and Reinberg D. 2007. Facultative heterochromatin: Is there a distinctive molecular signature? Mol. Cell 28: 1–13. Valenzuela L. and Kamakaka R.T. 2006. Chromatin insulators. Annu. Rev. Genet. 40: 107–138. Vaughn M.W. and Martienssen R.A. 2005. Finding the right template: RNA Pol IV, a plant-specific RNA polymerase. Mol. Cell 17: 754– 756. Vermeulen M., Mulder K.W., Denissov S., Pijnappel W.W., van Schaik F.M., Varier R.A., Baltissen M.P., Stunnenberg H.G., Mann M., and Timmers H.T. 2007. Selective anchoring of TFIID to nucleosomes by trimethylation of histone H3 lysine 4. Cell 131: 58–69. Vogelauer M., Wu J., Suka N., and Grunstein M. 2000. Global histone acetylation and deacetylation in yeast. Nature 408: 495–498. Wallace J.A. and Felsenfeld G. 2007. We gather together: Insulators and genome organization. Curr. Opin. Genet. Dev. 17: 400–407. Wang A., Kurdistani S.K., and Grunstein M. 2002. Requirement of Hos2 histone deacetylase for gene activity in yeast. Science 298: 1412–1414. Wang G., Balamotis M.A., Stevens J.L., Yamaguchi Y., Handa H., and Berk A.J. 2005. Mediator requirement for both recruitment and postrecruitment steps in transcription initiation. Mol. Cell 17: 683-694. Wang Q., Carroll J.S., and Brown M. 2005. Spatial and temporal recruitment of androgen receptor and its coactivators involves chromosomal looping and polymerase tracking. Mol. Cell 19: 631–642. Wang Z., Zang C., Rosenfeld J.A., Schones D.E., Barski A., Cuddapah S., Cui K., Roh T.Y., Peng W., Zhang M.Q., and Zhao K. 2008. Combinatorial patterns of histone acetylations and methylations in the human genome. Nat. Genet. 40: 897–903. Weake V.M. and Workman J.L. 2008. Clearing the way for unpaused polymerases. Cell 134: 16–18. Whitehouse I., Rando O.J., Delrow J., and Tsukiyama T. 2007. Chromatin remodelling at promoters suppresses antisense transcription. Nature 450: 1031–1035. Wisniewski J.R., Zougman A., Kruger S., and Mann M. 2007. Mass spectrometric mapping of linker histone H1 variants reveals multiple acetylations, methylations, and phosphorylation as well as differences between cell culture and tissue. Mol. Cell. Proteomics 6: 72–87. Woodcock C.L. 2006. Chromatin architecture. Curr. Opin. Struct. Biol. 16: 213–220. Woodcock C.L., Skoultchi A.I., and Fan Y. 2006. Role of linker histone in chromatin structure and function: H1 stoichiometry and nucleosome repeat length. Chromosome Res. 14: 17–25. Workman J.L. 2006. Nucleosome displacement in transcription. Genes Dev. 20: 2009–2017. Wu C., Bassett A., and Travers A. 2007. A variable topology for the 30nm chromatin fibre. EMBO Rep. 8: 1129–1134. Xi H., Shulha H.P., Lin J.M., Vales T.R., Fu Y., Bodine D.M., McKay R.D., Chenoweth J.G., Tesar P.J., Furey T.S., et al. 2007. Identification and characterization of cell type–specific and ubiquitous chromatin regulatory structures in the human genome. PLoS Genet. 3: e136. Xu L., Glass C.K., and Rosenfeld M.G. 1999. Coactivator and corepressor complexes in nuclear receptor function. Curr. Opin. Genet. Dev. 9: 140–147. Yang C., Bolotin E., Jiang T., Sladek F.M., and Martinez E. 2007. Prevalence of the initiator over the TATA box in human and yeast genes and identification of DNA motifs enriched in human TATAless core promoters. Gene 389: 52–65. Yang X.J. and Seto E. 2008. The Rpd3/Hda1 family of lysine deacetylases: From bacteria and yeast to mice and men. Nat. Rev. Mol. Cell Biol. 9: 206–218. Zhang H., Roberts D.N., and Cairns B.R. 2005. Genome-wide dynamics of Htz1, a histone H2A variant that poises repressed/ basal promoters for activation through histone loss. Cell 123: 219–231. Zhao H. and Dean A. 2005. Organizing the genome: Enhancers and insulators. Biochem. Cell Biol. 83: 516–524. Zilberman D. and Henikoff S. 2007. Genome-wide analysis of DNA methylation patterns. Development 134: 3959–3965. 01_TRE_001-046.qxd:TransRegEukaryotes.qxd 10/23/08 2:13 PM Page 46 ...
View Full Document

This document was uploaded on 04/12/2010.

Ask a homework question - tutors are online