As in bacteria, the basal level of eukaryotic transcription is determined by the effect of regulatory sequences on the function of RNA polymerase and its associated transcription factors. As we discussed in Chapter 19, the nature of the eukaryotic genome lends itself to different regulatory strategies from those used in bacteria. The eukaryotic genome is packaged in chromatin, which presents a physical block to RNA polymerases, and therefore the majority of eukaryotic genes are repressed in their default (ground) state and require protein activators to stimulate expression. Because of the large size of eukaryotic genomes and the need to guard against nonspecific protein-DNA interactions, the binding of multiple protein regulators is required to activate each gene. As a result, eukaryotic promoters are more complex than their bacterial counterparts and contain many more regulatory protein–binding sites. In reality, though, the additional complexity in eukaryotes is handled in strategic ways that are not as complicated as we might expect, given the overwhelming difference between a bacterium and an animal.

The genomic DNA of eukaryotes wraps around small basic proteins called histones to form nucleosomes, the building blocks of chromatin (see Figure 10-4). The transcription machinery must necessarily deal with chromatin structure in order to access particular genes. As a result, eukaryotic genes are generally expressed at low levels—or not at all—in the absence of regulatory proteins.

729

Chromatin structure is controlled and altered by at least three interrelated mechanisms: ATP-dependent changes in nucleosome positioning on the DNA, posttranslational chemical modifications of histone proteins, and substitution of specialized histone variants into chromatin. These mechanisms were discussed in detail in Chapter 10, and we recap briefly here. Nucleosome remodeling complexes use ATP to shift nucleosomes along the DNA. Active promoters contain open regions with nucleosomes positioned away from the promoter region, allowing access to transcription factors. Some posttranslational modifications of histones, including acetylation by histone acetyltransferases (HATs), result in the decondensing of chromatin and provide access to DNA-binding factors; proteins containing a bromodomain bind acetylated histones and facilitate opening of the chromatin structure. Alternatively, histone modifications cause chromatin to become tightly closed to transcription. For example, methylated histones are bound by proteins containing chromodomains, and these proteins help condense the chromatin. Chromatin structure is also modulated by several histone variants. These proteins are homologous to the common histones and can take their place in nucleosomes, but they also contain amino acid extensions that have a variety of functional consequences.

In the eukaryotic cell cycle, interphase chromosomes appear to be dispersed and amorphous. However, chromosomes are not uniform structures, and several different forms of chromatin can be found along each chromosome. About 10% of the chromatin in a typical eukaryotic cell is in a much more condensed form than the rest of the chromatin. This form, heterochromatin, is transcriptionally inactive. Although heterochromatin does not contain any genes (which is why it is inactive), the heterochromatin structure itself is known to be repressive, because, experimentally, a gene is silenced when it is placed in heterochromatin. Heterochromatin is often associated with particular chromosome structures, including centromeres and telomeres. The remaining, less condensed chromatin is called euchromatin (Figure 21-1).

Transcription of a eukaryotic gene is strongly repressed when its DNA is condensed within heterochromatin, but in euchromatin, some of the DNA is transcriptionally active. Regions of transcriptionally active DNA can be detected on the basis of their increased sensitivity to nuclease-mediated degradation. Nucleases such as DNase I tend to cleave the DNA of carefully isolated chromatin into fragments of multiples of about 200 bp, reflecting the regular repeating structure of the nucleosome (see Figure 10-1). However, in actively transcribed regions, the fragments produced by nuclease activity are smaller and more heterogeneous in size. Actively transcribed regions contain hypersensitive sites, sequences especially sensitive to DNase I, which are typically found in noncoding regions within 1,000 bp of the 5′ ends of transcribed genes. In some genes, hypersensitive sites are found farther from the 5′ end, or near the 3′ end, or even within the gene itself. The presence of hypersensitive sites suggests that DNA in that region is not packaged in the regular repeating nucleosomal structure.

Many hypersensitive sites correspond to binding sites for known regulatory proteins, and the relative absence of nucleosomes in these regions may allow the binding of these proteins. Nucleosomes are entirely absent in some regions that are very active in transcription, such as the rRNA genes. Transcriptionally active chromatin also tends to be deficient in histone H1, which binds the linker DNA between nucleosome particles.

Histones within transcriptionally active chromatin and heterochromatin also differ in their patterns of covalent modification. The C-terminal tails of the core histones are modified by the acetylation and methylation of Lys and Arg residues, phosphorylation of Ser or Thr residues, and ubiquitination or sumoylation (see Chapter 22). In particular, the acetylation-deacetylation of histones figures prominently in the processes that activate chromatin for transcription. The HAT-mediated acetylation of multiple Lys residues in the N-terminal domains of histones H3 and H4 can reduce the affinity of the entire nucleosome for DNA. Acetylation may also prevent or promote interactions with other proteins involved in regulating transcription. When transcription of a gene is no longer required, acetylation of nucleosomes in that vicinity is reduced by the activity of histone deacetylases (HDACs), resulting in condensation of the chromatin to reduce or inactivate gene transcription. HDACs often function through protein-protein interactions, such as by binding corepressors or acting as components of chromatin remodeling complexes.

730

A general model for deacetylation and gene inactivation is shown in Figure 21-2. In addition to the removal of certain acetyl groups, new covalent modifications of histones mark chromatin as transcriptionally inactive. For example, the Lys residue at position 9 in histone H3 is often methylated in heterochromatin.

Gene regulation through histone modifications is typically achieved through activation or inhibition of transcription initiation. However, an exciting finding has demonstrated that chromatin structure is also involved in the control of mRNA splicing for some genes (see this chapter’s Moment of Discovery). This may seem surprising, given that histones do not bind RNA. Yet Gcn5, a well-studied transcription factor with HAT activity, is now known to also affect RNA processing: loss of Gcn5 HAT activity prevents proper pre-mRNA splicing in yeast, because components of the splicing machinery fail to properly bind the pre-mRNA splice sites (Highlight 21-1). This example shows how different types of regulation (in this case, transcription and mRNA processing) can be interrelated. It seems likely that processes previously thought to be isolated and separable events may be interwoven in complex regulatory networks in the living cell.

Each of the three eukaryotic RNA polymerases has little or no intrinsic affinity for its promoters. Instead, initiation of transcription almost always requires activator proteins. An important reason for the apparent predominance of positive regulation is clear from the earlier discussion: chromatin structure effectively renders most promoters inaccessible, so genes are normally silent in the absence of other regulation. The structure of chromatin affects access to some promoters more than others, but repressor binding to DNA to block access of RNA polymerase (negative regulation) would often be simply redundant. Other factors are also at play in the use of positive regulation, however, and speculation generally centers on two: the large size of eukaryotic genomes and the greater efficiency of positive regulation.

Because eukaryotes have much larger genomes than bacteria, there is an increased likelihood that a specific binding sequence for a regulatory protein will occur randomly in other regions of the DNA. Recall that a sequence of n nucleotides will occur randomly every 4ⁿ bp. Thus, any single regulatory protein with a small binding site will probably bind nonspecifically to multiple places in the eukaryotic genome (see the How We Know section at the end of this chapter). Specific transcriptional activation of a gene through the binding of one regulatory protein to a small binding site, as often occurs in bacteria, would be ineffective in eukaryotes. In theory, specificity of regulator binding would be increased if the DNA sequences recognized by the proteins were longer. Yet eukaryotes did not evolve in this way: eukaryotic regulators do not bind longer DNA sequences than their bacterial counterparts. Instead, to achieve specific transcriptional activation of a gene, eukaryotes employ multiple regulatory proteins, or transcription factors, each of which binds a short sequence; successful gene activation occurs only when all the factors are bound at their individual sites. This “combination” of factors for activating one gene is used in combinatorial control (see Section 21.2).

731

To accommodate the binding of multiple transcription factors, eukaryotic promoters are necessarily more complicated than their bacterial counterparts (Figure 21-3). Take, for example, a typical promoter recognized by RNA polymerase II (Pol II), the enzyme responsible for mRNA synthesis. Many (but not all) Pol II promoters include the TATA box and Inr (initiator) sequences, with their standard spacing (see Figure 15-20). These sequences comprise the core promoter.

Eukaryotic genes also include regulatory sequences called enhancers in higher eukaryotes and upstream activator sequences (UASs) in yeast, to which transcription activators bind. These sequences cannot all be positioned adjacent to the promoter—there is simply not enough room to accommodate the binding of so many regulatory proteins. The binding sites for multiple transcription factors must be able to act at a distance. In fact, they can be surprisingly far from the promoter. A typical enhancer may be hundreds or even thousands of base pairs upstream from the transcription start site, or downstream from the gene, or even within the gene itself. When bound by the appropriate regulatory proteins, an enhancer increases transcription at nearby promoters, regardless of its orientation in the DNA. Yeast UASs function in a similar way, although generally they must be positioned upstream and within a few hundred base pairs of the transcription start site. An average Pol II promoter may be affected by half a dozen regulatory sequences of this type, and many promoters are even more complex. In contrast, bacteria have very few genes that use a distantly bound transcription activator.

The more complex the eukaryotic organism, the more complex its promoters are likely to be. For example, mammalian promoters are generally much more complex than yeast promoters (Figure 21-4).

732

The requirement for the binding of several transcription activators to several specific DNA sequences vastly reduces the probability of the random occurrence of a functional juxtaposition of all the necessary binding sites. In principle, a similar strategy could be used by multiple negative-regulatory elements. However, positive regulation is simply more efficient. From an energy standpoint, it makes more sense for the cell to synthesize several activators to promote transcription of the subset of genes needed at that time, rather than constantly synthesize one or more repressors for every gene in the genome to keep them turned off until needed. Positive regulation of transcription predominates in eukaryotes, although, as we will see, there are some examples of negative regulation.

To further conserve resources, differently regulated eukaryotic promoters often use some of the same protein activators, so diverse promoters can have some of the same binding sequences. However, only a specific combination of regulatory factors can unlock a given promoter and activate transcription of that gene. With this mechanism, the cell can achieve specificity of gene regulation with a smaller number of transcription activators than if each gene were regulated by a set of unique proteins (see Figure 19-14). Some regulatory proteins facilitate transcription at hundreds of promoters, whereas others are specific for only a few promoters. In addition, many transcription activators are sensitive to the binding of effector signal molecules, providing the capacity to activate or deactivate transcription in response to a changing cellular environment.

733

Successful binding of active Pol II holoenzyme at one of its promoters usually requires the action of three types of regulatory proteins: general transcription factors, DNA-binding transcription activators, and coactivators. General (basal) transcription factors are required at every Pol II promoter; DNA-binding transcription activators, or DNA-binding transactivators, bind to enhancers or UASs to facilitate transcription; and coactivators act indirectly—by binding other proteins rather than DNA—and are required for essential communication between the DNA-binding transactivators and the complex composed of Pol II and the general transcription factors (Figure 21-5a). Sometimes, a variety of repressor proteins can interfere with communication between Pol II and the DNA-binding transactivators, resulting in repression of transcription (Figure 21-5b). In fact, some proteins act as an activator or coactivator at one promoter and a repressor or corepressor at another promoter. Here we focus on the protein complexes shown in Figure 21-5a and how they interact to activate transcription.

For transcription to begin, the Pol II holoenzyme must be recruited to the promoter to form a preinitiation complex with the general transcription factors. Assembly of a preinitiation complex at a typical Pol II promoter begins with the binding of TATA-binding protein (TBP) to the TATA box. TBP, which is part of the larger transcription factor complex called TFIID, then recruits additional general transcription factors and Pol II (see Figure 15-24). The minimal preinitiation complex, however, is often insufficient for the initiation of transcription, and it generally does not form at all if the promoter is buried in chromatin. Positive regulation by transcription activators and coactivators is required. We now know that the basal Pol II machinery is not as uniform as originally thought; the individual components can vary with cell type. A well-documented example is muscle cells (see the How We Know section at the end of this chapter). Thus, different combinations of general transcription factors form a complex at a promoter and are acted on by specific activator and coactivator binding proteins, adding a further level of control to the regulation of that gene.

734

As noted above, binding sites for transcription activators are often located far from the promoters they regulate. Recall from Chapter 19 that the intervening DNA is looped so that the various protein complexes can interact, directly or indirectly. DNA looping is promoted by certain nonhistone proteins that are abundant in chromatin and bind nonspecifically to DNA. These high-mobility group (HMG) proteins play an important structural role in chromatin remodeling and transcriptional activation (“high mobility” refers to the proteins’ rapid electrophoretic mobility in polyacrylamide gels). A structure formed by an HMG-box domain in HMG proteins can bind directly to nucleosomes, leading to altered local chromatin structure. Figure 21-6 shows the high degree to which DNA is bent by the HMG-box DNA-binding domain of the protein HMG-D of Drosophila melanogaster, one of many DNA-interactive protein structures determined in the laboratory of Mair Churchill.

In addition to transcription activators, most transcription requires coactivator protein complexes. Some major regulatory protein complexes that interact with Pol II have been defined both genetically and biochemically. They act as intermediaries between the DNA-binding transactivators and the Pol II complex (Pol II and the general transcription factors). The best-characterized coactivator is TFIID (see Chapter 15). In eukaryotes, TFIID includes TBP and 10 or more TBP-associated factors (TAFs). Some TAFs resemble histones and may play a role in competing with and thus displacing nucleosomes during the activation of transcription. Many DNA-binding transactivators aid in transcription initiation by interacting with one or more TAFs.

735

Another important coactivator is the Mediator complex (see Figure 15-25), which consists of 20 core polypeptides that are highly conserved from fungi to humans. Mediator binds tightly to the C-terminal domain (CTD) of the largest Pol II subunit. The Mediator complex is required for both basal and regulated transcription at Pol II promoters, and it also stimulates phosphorylation of the Pol II CTD by the general transcription factor TFIIH. Phosphorylation of the CTD enhances the efficiency of Pol II. As with TFIID, some DNA-binding transactivators interact with one or more components of the Mediator complex. Some promoters require both Mediator and TFIID coactivators. The coactivator complexes function at or near the promoter’s TATA box.

We can now begin to piece together the sequence of transcriptional activation events at a typical Pol II promoter. Crucial remodeling of the chromatin takes place in stages. Some DNA-binding transactivators have significant affinity for their binding sites even when the sites are within condensed chromatin. Binding of one transactivator may facilitate the binding of others, gradually displacing some of the nucleosomes that previously obscured the relevant DNA.

The bound transcription activators may have HAT activity or may recruit HATs or enzyme complexes such as SWI/SNF, accelerating the remodeling of surrounding chromatin (Figure 21-7). In this way, transcription activator binding can lead to the stepwise assembly of components necessary for further chromatin remodeling, to permit the transcription of specific genes. The bound transactivators, acting through complexes such as TFIID or Mediator (or both), stabilize the binding of Pol II and its associated general transcription factors, greatly facilitating formation of the preinitiation complex. Complexity in these regulatory circuits is the rule rather than the exception, with multiple DNA-bound transactivators promoting transcription.

The script can change from one promoter to another, but most promoters seem to require a precisely ordered assembling of components to initiate transcription. The assembly process is not always fast: at some genes it may take minutes, but it can take days at certain genes in higher eukaryotes.

736

Although rarer, some eukaryotic regulatory proteins that bind Pol II promoters can act as repressors, inhibiting the formation of active preinitiation complexes. Some transcription activators can adopt different conformations, enabling them to serve as activators or repressors. For example, some steroid hormone receptors function in the nucleus as DNA-binding transactivators, stimulating transcription of certain genes when a particular steroid hormone signal is present (see Section 21.3). When the hormone is absent, the receptor proteins revert to a repressor conformation, preventing formation of preinitiation complexes. In some cases this repression involves interaction with HDACs and other proteins that help restore the surrounding chromatin to its transcriptionally inactive state.