Regulatory processes operating at the level of transcription initiation are the best documented, and probably the most common. Elaborate mechanisms have evolved to regulate the process of transcription initiation—before large amounts of cellular energy are invested in the production of mRNAs and their protein products. But however diverse, these control mechanisms are really just variations on a common theme and boil down to simple protein-protein and protein-DNA interactions. Indeed, regulation at the step of transcription initiation can be explained simply by changes in how RNA polymerase interacts with the DNA at promoter sequences.

Regulatory proteins that bind DNA can have profound effects on the affinity of RNA polymerase for a promoter. These effects can flow in either direction, either enhancing or preventing RNA polymerase function. Given both the energy expenditure of gene expression and the need for cells to be able to respond quickly to changes in their environment, one can only imagine the enormous evolutionary pressure placed on regulatory mechanisms. We provide here an overview of the transcriptional regulatory mechanisms used by cells. The most detailed information derives from studies in bacterial systems, but eukaryotic mechanisms of gene regulation, although more complex and using somewhat different strategies, can be explained by the same basic principles.

The most basic mechanism for regulation of transcription initiation is encoded in the DNA sequence of the promoter. RNA polymerase has different intrinsic affinities for promoters of different sequence. In the absence of other controls, these differences in promoter strength correlate with the efficiency with which the genes are transcribed. Genes for products that are required at all times, such as the enzymes of central metabolic pathways, are expressed at a nearly constant level. These genes are often referred to as housekeeping genes, and unvarying expression is called constitutive gene expression. Although housekeeping genes are expressed constitutively, the expression levels of different housekeeping genes vary widely. For these genes, the RNA polymerase–promoter interaction strongly influences the rate of transcription initiation; with differences in promoter sequences, the cell can synthesize the appropriate level of each housekeeping gene product.

When the level of a gene product rises and falls with a cell’s changing needs, this is known as regulated gene expression. Activation is an increase in expression and repression is a decrease in expression of a gene in response to a change in environmental conditions. The mechanisms of gene activation and repression, in both bacteria and eukaryotes, require the assistance of transcription factors (also called transcription regulators), proteins that alter the affinity of the RNA polymerase for the promoter. Transcription factors that enhance gene expression are called activators, and those that reduce expression are called repressors. As we will see later in this chapter, bacterial and eukaryotic transcription factors have many common structural and functional features. These regulators act by binding to specific DNA sequences known as regulatory sites.

A gene is said to be under positive regulation when binding of an activator protein promotes or increases expression of that gene. Conversely, a gene is under negative regulation when binding of a repressor protein prevents or decreases expression. Thus, positive and negative regulation refer to the type of regulatory protein involved: the bound protein either facilitates or inhibits transcription.

A repressor can lower the rate of gene transcription if its regulatory site overlaps the gene’s promoter and repressor binding sterically occludes binding of the RNA polymerase to the promoter (Figure 19-2a). Repressors can act in other ways as well. Some prevent transcription by binding a regulatory site that is near the promoter, but the binding itself does not block RNA polymerase binding. Recall that the RNA polymerase–promoter complex converts from the closed complex to the open complex, in which the DNA strands are melted locally, opening up the duplex for transcription (see Figures 15-13 and 15-23). Repressors can block this closed-to-open transition, thereby preventing transcription. Instead of steric occlusion, the underlying principle in these mechanisms is allostery: a conformational change in the RNA polymerase–DNA complex prevents formation of the open complex. Repressors of this type can act on the RNA polymerase or directly on the DNA to stabilize the closed complex over the open complex. Other types of repressors act by holding the RNA polymerase to the promoter site, preventing its escape from the promoter.

668

Activators provide a molecular counterpoint to repressors: they bind to DNA and enhance the activity of RNA polymerase at a promoter. For example, an activator may induce a conformational change in the polymerase that accelerates transition to the open complex. Alternatively, an activator may alter the torsion of DNA, making it more likely to unwind and form the open complex. Probably the most common way that activators function is through cooperativity (Figure 19-2b). In this case, the activator binds both the RNA polymerase and a DNA regulatory site next to the promoter, thereby increasing the affinity of the polymerase for the promoter; this activation process is referred to as recruitment.

Binding sites for activators and repressors are often found at or near the promoter, particularly in bacteria. However, regulatory sites can also be found far away from the promoter. In fact, in eukaryotes, regulatory sites are sometimes thousands of base pairs upstream or downstream from a promoter. How do transcription factors exert their effects on RNA polymerase when their binding sites are so far away from the gene’s promoter? Experiments directed at understanding this “action at a distance” have demonstrated that distant regulatory sites can often be placed closer to or farther from the promoter and still retain their function. Not only is distance from the promoter of little consequence (provided it is not too close and overlaps the promoter), but the regulatory sites also retain their function regardless of experimental changes in their sequence orientation relative to the promoter.

When distant regulatory sites were first discovered, scientists imagined that the regulatory proteins might bind to these sites and then slide along the DNA until they reached RNA polymerase at the promoter. However, experiments revealed that this was not the case (see the How We Know section at the end of this chapter). Instead, the DNA between the regulatory site and the RNA polymerase–promoter complex loops out to bring the regulatory protein and RNA polymerase together (Figure 19-3). Looping can be directly observed using an electron microscope (Figure 19-4). DNA looping is facilitated by proteins called architectural regulators that bind to DNA sequences between the regulatory site and the promoter, bending the DNA (Figure 19-5). The use of DNA looping is common in eukaryotic gene regulation, and it is also used by some bacterial gene regulatory proteins. In eukaryotes, these regulatory sites that bind transcription factors and exert their regulatory effect on the promoter over long distances are called enhancers.

669

Gene regulation by DNA looping can result in either activation or repression. Activation can recruit RNA polymerase to the promoter through cooperativity, in much the same way as an activator that binds near the promoter. Recruitment of RNA polymerase through DNA looping can also be mediated by a protein “bridge” between the activator and the polymerase (Figure 19-6a). Proteins that act by bridging activators and RNA polymerase, but do not bind DNA directly, are called coactivators. For example, the eukaryotic protein complex Mediator acts as a bridge between RNA polymerase II and regulatory proteins bound to distant sites and is essential for transcription activation (see Chapter 15). Repression can also occur through proteins that do not bind the DNA directly but instead bind activator proteins and prevent the recruitment of RNA polymerase (Figure 19-6b). Repressors that act through protein-protein interaction rather than by binding DNA directly are called corepressors.

There could be an unintended consequence of gene regulation that uses DNA loops over large distances. A regulator meant to target a distant promoter could act instead on a different promoter located in the opposite direction. In eukaryotes, this problem is solved by the presence of insulators, short sequences of DNA that prevent inappropriate cross-signaling (Figure 19-7). Insulators are discussed further in Chapter 21.

670

Activators and repressors often function at the same promoter. The use of multiple transcription factors allows the expression of a gene to be affected by more than one environmental condition. Signal integration, occurring in both eukaryotes and bacteria, is the control of a gene by multiple regulators in response to more than one environmental signal. A simple example in bacteria is the regulation of genes that encode products responsible for metabolizing sugar, the main energy source for bacteria (Figure 19-8).

Bacteria can derive energy from many different sugars, and they have sets of genes for metabolizing each one. But it would be a waste of cellular resources to express all of these genes all the time, and systems of regulation have evolved in which the genes for metabolizing a given sugar are expressed only when that sugar is present in the environment. Take, for example, the lac operon, a set of genes for the metabolism of lactose (see Chapter 5 and Chapter 15; operons are more fully defined later in this section). When lactose is not present, the Lac repressor protein is bound to the operon DNA at a sequence called the operator and ensures that the genes for lactose metabolism are not transcribed. When lactose is present, the cell sends a signal for the Lac repressor to dissociate from the operator, allowing transcription of the genes encoding lactose-metabolizing enzymes.

671

Though bacteria can metabolize many different sugars, their best energy source is glucose. When both glucose and lactose are present in the environment, the cell preferentially metabolizes glucose. It would be a waste of energy to continue producing the lactose-metabolizing enzymes, but the presence of lactose causes dissociation of the Lac repressor from the DNA. And yet, under these conditions, the genes encoding lactose-metabolizing enzymes are not highly transcribed. How does the cell do this? This is where signal integration comes in. The lactose-metabolizing genes are also under the control of an activator protein needed for the efficient transcription of the lac operon genes, even in the absence of the Lac repressor (see Figure 19-8). When glucose is present, the activator protein is kept in a nonfunctional form. But in the absence of glucose, the activator becomes functional and, provided lactose is present (and thus the Lac repressor is not bound to the operator), the genes for lactose metabolism are expressed at a high level.

This exquisite control, achieved by two different transcription factors working together, is an example of signal integration. The cell can adjust its energy resources by taking into account more than one environmental condition (the availability of glucose and of lactose).

The regulation of gene expression usually operates as a feedback circuit. This is easier to explain in bacteria than in eukaryotes, although similar principles apply in both. Recall that genes for the metabolism of lactose are controlled by multiple transcription factors. The repressors and activators either bind DNA or do not, depending on signals received from the environment. The binding of a repressor or activator to DNA is often regulated by a molecular signal called an effector, usually a small molecule or another protein that binds the activator or repressor and causes a conformational change that results in an increase or decrease in transcription.

Repressors can be activated or inactivated by effectors. In one scenario, the effector binds to the repressor and induces a conformational change that results in dissociation of the repressor from its binding site on the DNA, allowing transcription to proceed (Figure 19-9a). Alternatively, the interaction of an inactive repressor and a signal molecule could cause the repressor to bind to DNA, shutting down transcription (Figure 19-9b).

The same considerations apply to activators. Some activators bind DNA and enhance transcription until dissociation of the activator is triggered by the binding of a signal molecule (Figure 19-10a). In other cases, the activator binds to DNA only after interaction with a signal molecule (Figure 19-10b). Signal molecules that bind activators can therefore increase or decrease transcription, depending on how they affect the activator.

672

Given the allosteric control of activators and repressors, we can understand how a regulatory feedback loop functions. In the bacterial lac operon, Lac repressor binds DNA in the absence of an effector, preventing the expression of genes required for the metabolism of lactose. The effector for the Lac repressor is allolactose, a minor byproduct of lactose metabolism. Therefore, when lactose is present in the environment, the signal molecule is formed and binds the Lac repressor, causing it to dissociate from the DNA. This gives RNA polymerase access to the promoter of the lac operon for a low, basal level of transcription. Transcription of the operon is greatly enhanced by binding of the activator cAMP receptor protein (CRP). CRP does not bind its regulatory site when glucose is available. In the absence of glucose, however, cells produce cAMP (cyclic AMP), which is an allosteric effector of CRP, producing a conformational change that enables CRP to bind its regulatory site. The bound activator then recruits RNA polymerase and boosts gene expression from the lac operon. A second level of control, called inducer exclusion, occurs when the lactose transporter is blocked by the glucose permease. The glucose permease is usually phosphorylated, and it becomes dephosphorylated on the transfer of phosphate to glucose during glucose transport. The dephosphorylated form of the permease binds and directly blocks the lactose and maltose transporters. When glucose is absent, the glucose permease exists mainly in its phosphorylated form and no longer inhibits the transporters of other sugars.

When lactose is depleted, allolactose is also depleted, and in the absence of this effector the Lac repressor again binds the operator site, preventing RNA polymerase from transcribing the lac operon. Likewise, when glucose becomes available, cAMP levels diminish and CRP no longer binds DNA. Regulatory feedback loops like these are common in all cells.

Bacterial promoters are often positioned upstream from several genes that operate in a common metabolic pathway. Transcription produces a long polycistronic mRNA that contains multiple genes in one transcript. The single promoter that initiates transcription of the cluster is the site of regulation for all the genes in the polycistronic message. The polycistronic DNA, its promoter, and all the additional sequences that function together in regulating its transcription are called an operon (Figure 19-11). Most operons contain 2 to 6 genes, but some have more than 20 genes.

The organization of bacterial genes into operons allows small sets of genes that function together to be regulated together. But there are also instances in which multiple operons are controlled in a coordinated fashion. A group of operons with a common regulator is called a regulon. This arrangement allows shifts in cellular functions that can require the action of hundreds of genes—a major theme in the regulated expression of dispersed networks of genes in bacteria. Eukaryotes also exhibit global regulation of genes; genes that function together are dispersed over different chromosomes, yet are typically controlled in a coordinated way through common control elements and transcription factors. Figure 19-12 shows a generalized view of global regulation, in which multiple genes may be turned on by the presence of the same activator or by the removal of a common repressor. Mechanisms of global transcriptional gene regulation in bacteria and eukaryotes are described in detail in Chapters 20 and 21.

Signal integration is important to gene regulation in both bacteria and eukaryotes. However, eukaryotic promoters for Pol II, the RNA polymerase that transcribes protein-coding genes, typically contain more regulatory binding sites than do bacterial promoters (Figure 19-13). The use of more transcription factors in eukaryotic gene control reflects the greater need for gene regulation in a more complex organism with a larger genome. For example, nonspecific DNA binding of regulatory proteins could become a problem in the much larger genomes of higher eukaryotes, because the chance that a specific binding sequence will occur randomly at an inappropriate site increases with genome size. Indeed, the number of transcription factor–binding sites in eukaryotic promoter regions varies with the complexity of the organism. Genes in single-celled yeasts have only a few regulator sites and are not much more complicated than bacterial genes, whereas the promoters in multicellular organisms can have 10 or more regulator-binding sites spaced over long distances, 50 kbp or more away from the transcription start site. Specificity for transcriptional regulation is improved by multiple regulatory proteins that must bind DNA and form a multiprotein complex to become active. This multiprotein requirement vastly reduces the probability of random gene activation or repression.

673

Using multiple transcription factors for every gene in a genome would be energetically costly if every gene required unique regulators, but the use of different combinations of a limited set of transcription factors to differentially regulate many genes provides an opportunity for efficiency. This is made possible by combinatorial control—the need for specific combinations of factors to unlock each particular gene (Figure 19-14). Consider the hypothetical genes A, B, and C, each of which requires five transcription factors. If each factor were distinct, the cell would require 15 transcription factors to control the expression of these three genes. But if genes A and B used three of the same factors, and a combination of the factors for genes A and B is used to regulate gene C, then differential regulation of these three genes would require only 7 different proteins instead of 15.

674

Combinatorial control occurs in bacteria as well as in eukaryotes, and we have already seen an example in bacteria in the case of the two regulatory elements of the genes involved in lactose metabolism. Recall that these genes are controlled by a repressor that senses the presence of lactose and by an activator that senses the presence of glucose. The genes encoding proteins for the metabolism of other sugars have their own repressors, but they use the same activator. For instance, the digestion of galactose requires removal of the galactose repressor from the DNA, and this occurs only when galactose is present. However, as with the lactose genes, high expression of the galactose genes is achieved only when glucose is absent from the environment. The same protein activator used at the lac operon also regulates the galactose genes: CRP, which becomes functional by binding its effector molecule cAMP when glucose is not present. The regulation of the genes for different sugar-metabolizing pathways by a common activator is an example of combinatorial control.

In eukaryotes, transcription initiation almost always depends on the action of activator proteins. One important reason for the apparent predominance of positive regulation seems obvious: packaging of DNA into chromatin renders most promoters inaccessible, and thus their associated genes are silent. Chromatin structure affects access to some promoters more than others, but generally, repressors that prevent the access of RNA polymerase to DNA would be redundant. Therefore, eukaryotic genes are constitutively repressed and require activation in order to be transcribed. Recall from Chapter 10 that transcription is regulated by different types of change in chromatin structure. The chromatin state can be either open or closed. Open chromatin is often (but not always) associated with acetylation of nucleosomes, whereas closed chromatin is associated with methylation of nucleosomes. Thus eukaryotic activators and repressors can act through modification of nucleosomes that alter chromatin structure, rather than by recruiting RNA polymerase or preventing polymerase binding to DNA.

Bacterial RNA polymerase generally has access to every promoter, and most bacterial genes are controlled by specific repressors. In eukaryotes, however, general repression by nucleosomes, combined with the use of activators to regulate transcription, is more efficient than the use of specific repressors. If the 20,000 to 25,000 genes in the human genome were negatively regulated, each cell would have to constantly synthesize specific repressors to prevent the transcription of a great many genes. Instead, the nucleosomes that function to condense DNA also repress most genes, and the cell has to synthesize only the activators needed to promote transcription of the subset of genes required at a particular time. These arguments notwithstanding, there are examples of negative regulation in eukaryotes, from yeast to humans.