Transcription: Gene Regulation in Eukaryotes an Overview

Eukaryotes face the same basic tasks of coordinating gene expression as do prokaryotes but in a much more intricate way. Some genes have to respond to changes in physiological conditions. Many others are parts of developmentally triggered genetic circuits that organize cells into tissues and tissues into an entire organism (except for unicellular eukaryotes). In these cases, the signals controlling gene expression are the products of developmental regulatory genes, rather than signals from the external environment.

Most eukaryotic genes are controlled at the level of transcription, and the mechanisms are similar in concept to those found for bacteria. Trans-acting regulatory proteins work through sequence-specific DNA binding to their cis-acting regulatory target sequences. Because of the much more complex regulation that is required to coordinate proper gene activity throughout the lifetime of a multicellular organism, there are some considerable novelties as well. We shall see examples of these novelties in this and subsequent chapters.

Typically, eukaryotes have many more genes than do prokaryotes, sometimes by several orders of magnitude. The genes of higher organisms also tend to be larger, owing to the facts that cis-acting sequences on the DNA can be located tens of thousands of base pairs away from the transcription start site and that a battery of regulatory factors is sometimes needed to bring about proper regulation of certain genes.

As mentioned in Chapter 3, eukaryotes have three different classes of RNA polymerases (distinguished by roman numerals I, II, and III). All mRNA molecules are synthesized by RNA polymerase II, and the rest of this chapter will focus on the transcription of mRNAs. To achieve maximal rates of transcription by RNA polymerase II, the cooperation of multiple cis-acting regulatory elements is required. We can distinguish three classes of elements on the basis of their relative locations. Near the transcription initiation site are the core promoter (the RNA polymerase IIbinding region) and promoter-proximal cis-acting sequences that bind to proteins that in turn assist in the binding of RNA polymerase II to its promoter. Additional cis-acting sequence elements can act at considerable distancethese elements are termed enhancers and silencers. Often, an enhancer or silencer element will act only in one or a few cell types in a multicellular eukaryote. The promoters, promoter-proximal elements, and distance-independent elements are all targets for binding by different trans-acting DNA-binding proteins.

Figure 14-19 is a schematic view of the core promoter and promoter-proximal sequence elements. The core promoter usually refers to the region from the transcription start site including the TATA box, which resides approximately 30 bp upstream of the transcription initiation site. This core promoter is unable to mediate efficient transcription by itself. Some important elements near the promoter, the promoter-proximal elements, are found within 100200 bp of the transcription initiation site. The CCAAT box functions as one of these promoter-proximal cis-acting sequences, and a GC-rich segment often functions as another. An example of the consequences of mutating these sequence elements on transcription rates is shown in Figure 14-20.

In eukaryotes, we distinguish between two classes of cis-acting elements that can exert their effects at considerable distance from the promoter. Enhancers are cis-acting sequences that can greatly increase transcription rates from promoters on the same DNA molecule; thus, they act to activate, or positively regulate, transcription. Silencers have the opposite effect. Silencers are cis-acting sequences that are bound by repressors, thereby inhibiting activators and reducing transcription. Enhancers and silencers are similar to promoter-proximal regions in that they are organized as a series of cisacting sequences that are bound by trans-acting regulatory proteins. However, they are distinguished from promoter-proximal elements by being able to act at a distance, sometimes 50 kb or more, and by being able to operate either upstream or downstream from the promoter they control. Enhancer and silencer elements are intricately structured. Figure 14-21 shows the DNA sequence of the SV40 (simian virus 40) enhancer, which is required for high-level expression of SV40 transcripts. Within the enhancer, there are five sequence elements required for maximal enhancement of transcription. Enhancers that are themselves composed of multiple copies of a DNA-binding element are common. Different DNA sequences serve as target-recognition sites for specific trans-acting regulatory proteins.

How do enhancer and silencer elements many thousands of base pairs away regulate transcription? Most models for such action at a distance include some type of DNA looping. Figure 14-22 details a DNA-looping model for activation of the initiation complex (see Figure 14-23 on page 452). In this model, a DNA loop brings activator proteins bound to distant enhancer elements into proximity to protein complexes associated with promoter-proximal cis-acting sequences.

A large number of trans-acting regulatory proteins have now been identified in eukaryotic cells. Like their counterparts in prokaryotes, these regulatory proteins act by binding to specific target DNA sequences.

Regulatory proteins that bind the core promoter and promoter-proximal elements help RNA polymerase II to initiate transcription, and together with the polymerase they form an initiation complex, as pictured in Figure 14-23a. Several different transcription factor complexes (TFII complexes) interact with RNA polymerase II. For example, the TFIID complex consists of a TATA boxbinding protein (TBP) and more than eight additional subunits (TAFs). The TFII complexes are often referred to as basal or general transcription factors, because they are the minimal requirement for RNA polymerase II to initiate transcription (usually very weakly) at a promoter. Figure 14-23b shows the structure of the TATA boxbinding protein binding to DNA. CCAAT and GC boxes are recognized by additional DNA-binding proteins.

Some of the proteins that bind distance-independent elements also have been identified. The protein encoded by the yeast GCN4 gene is an example of a trans-acting enhancer-binding protein. It operates on enhancers called upstream activating sequences (UASs). GCN4 activates the transcription of many yeast genes that encode enzymes of amino acid biosynthetic pathways. In response to amino acid starvation, the level of GCN4 protein rises and, in turn, increases the expression levels of the amino acid biosynthetic genes. The UASs recognized by GCN4 contain the principal recognition sequence element ATGACTCAT.

Many enhancer elements in higher eukaryotes activate transcription in a tissue-specific mannerthat is, they induce expression of a gene in one or a few cell types. For example, antibody genes are flanked by powerful enhancers that operate only in the B lymphocytes of the immune system. Many enhancers are integral components of complex tissue-specific genetic circuits that underlie complex events in development in higher eukaryotes. Tissue specificity is conferred in one of two ways. An enhancer can act in a tissue-specific manner if the activator that binds to it is present in only some types of cells. Alternatively, a tissue-specific repressor can bind to a silencer element located very near the enhancer element, making the enhancer inaccessible to its transcription factor.

In some genes, regulation can be controlled by simple sets of enhancers. For example, in Drosophila, vitellogenins are large egg yolk proteins made in the female adult's ovary and fat body (an organ that is essentially the fly's liver) and transported into the developing oocyte. Two distinct enhancers located within a few hundred base pairs of the promoter regulate the vitellogenin gene, one driving expression in the ovaries and the other in the fat body.

The array of enhancers for a gene can be quite complex, controlling similarly complex patterns of gene expression. The dpp (decapentaplegic) gene in Drosophila, for example, encodes a protein that mediates signals between cells (see Chapter 15). It contains numerous enhancers, perhaps numbering in the tens or hundreds, dispersed along a 50-kb interval of DNA. Some of these enhancers are located 5 (upstream) of the transcription initiation site of dpp, others are downstream of the promoter, some are in introns, and still others are 3 of the polyadenylation site of the gene. Each of these enhancers regulates the expression of dpp in a different site in the developing animal. Some of the better-characterized dpp enhancers are shown in Figure 14-24.

The requirement for multiple enhancer elements to regulate tissue-specific expression helps to explain the large size of genes in higher eukaryotes. The tissue-specific regulation of a gene may be quite complex, requiring the action of numerous, distantly located enhancer elements.

An important part of modern genetics is the identification and characterization of distantly located regulatory elements by means of transgenic constructs, in which recombinant DNA molecules are inserted into the genome of an organism (see Chapter 12). In these constructs, isolated pieces of a gene are incorporated to determine what tissue and temporal regulatory patterns it can control. By means of such slicing and dicing experiments, it is possible to home in on the locations of specific regulatory elements. We can also exploit these regulatory elements to develop new ways to identify genes of interest. In the following sections, we shall see how such experiments are carried out, by using studies of transcriptional enhancers as examples. Remember, however, that the same techniques and logic can be applied to any other classes of regulatory elements, some of which are considered in Chapter 16.

Enhancers of a cloned gene are typically identified by means of transgenic reporter genes. In reporter-gene constructs, pieces of cis-regulatory DNA are fused (usually by restriction-enzyme-based "cutting and pasting" of recombinant DNA molecules) near a transcription unit that can express a reporter proteinthat is, a protein whose presence can be monitored. Our old friend, the E. coli b-galactosidase enzyme encoded by the lacZ gene, is a very popular reporter protein, because it is very easy to detect the presence of b-galactosidase histochemically by adding a synthetic substrate, X-gal, to the medium and observing which tissues turn blue. The reporter-gene-construct transcription unit contains a "weak" promoterone that cannot initiate transcription without the assistance of an enhancer (Figure 14-25). The construct is then introduced by DNA transformation into the germ line of a host organism, and appropriate cells are histochemically assayed for the presence of the reporter protein. Two examples of reporter-gene expression are shown, one for enhancers of the dpp gene in Drosophila (Figure 14-26) and the other for a mouse enhancer expressed in muscle precursor cells (Figure 14-27).

Reporter protein expression in a tissue indicates the presence of one or more enhancers within the tested piece of DNA. When a piece of DNA has been found to act as an enhancer, the enhancer can be further localized by testing smaller and smaller subfragments of the original DNA segment, by using the same reporter-gene assay.

Ultimately, the DNA sequence of the enhancer can be identified by whittling down the piece of cis-regulatory DNA. With this sequence known, the next question of importance is the identity of the transcription-factor proteins that bind to the enhancer. Methods now exist to identify enhancer-binding proteins and to clone the genes that encode these proteins. When these genes have been cloned, they can be characterized by genetic and molecular techniques. With the use of these approaches, it is possible to build detailed circuit diagrams of the genetic pathways that regulate gene expression in eukaryotes.

The properties of regulatory elements help us to understand certain classes of dominant mutations. We can divide dominant mutations into two general classes. For some dominant mutations, inactivation of one of the two copies of a gene reduces the gene product below some critical threshold for producing a normal phenotype; we can think of these mutations as loss-of-function dominant mutations (referred to as haploinsufficient mutations in earlier chapters). In other cases, the dominant phenotype is due to some new property of the mutant gene, not to a reduction in its normal activity; this class comprises gain-of-function dominant mutations.

Commonly, gain-of-function dominant mutations arise through the fusion of parts of two genes to each other. (Be aware that there are also other mechanisms for producing a gain-of-function dominant mutation.) Such fusions can occur at the breakpoints of chromosomal rearrangements such as inversions, translocations, duplications, or deletions (see Chapter 8). Because enhancers can act at long distance and can activate many different promoters, misregulation of a gene can occur if a chromosomal rearrangement juxtaposes enhancers of one gene and a transcription unit of another gene. In such cases, the enhancers of the gene at one breakpoint can now regulate the transcription of a gene near the other breakpoint. Often, this misregulation leads to the misexpression of the mRNA encoded by the transcription unit in question. Such fusions can lead to gain-of-function dominant mutant phenotypes, depending on the nature of the protein product of the misexpressed mRNA and the tissues in which it is misexpressed.

The classic Bar dominant mutation in Drosophila is an example of such misregulation through gene fusion. In the Bar mutation, cis-regulatory elements that promote expression in the developing eye are fused to a gene that is ordinarily not expressed in the eye. This latter gene encodes a transcription factor, and misexpression of that transcription factor in the developing eye leads to the death of many cells of the developing eye and thus to the small eye Bar phenotype.

In a few cases, the basis for the misexpression in such gene fusions is quite well understood. One such case is the Tab (Transabdominal) mutation in Drosophila. Tab causes part of the thorax of the adult fly to develop instead as tissue normally characteristic of the sixth abdominal segment (A6) (Figure 14-28). Tab is associated with a chromosomal inversion. One breakpoint of the inversion is within an enhancer region of a different gene, the sr (striped) gene. These enhancers of the sr gene induce gene expression in certain parts of the thorax of the fly. The other breakpoint is near the transcription unit of the Abd-B (Abdominal-B) gene. The Abd-B gene encodes a transcription factor that normally is expressed only in posterior regions of the animal, and this Abd-B transcription factor is responsible for conferring an abdominal phenotype on any tissues that express it. (We shall have more to say about genes such as Abd-B in the treatment of homeotic genes in Chapter 16.) In the Tab inversion, the sr enhancer elements controlling thoracic expression are juxtaposed to the Abd-B transcription unit, causing the Abd-B gene to be activated in exactly those parts of the thorax where sr would ordinarily be expressed (Figure 14-29). Because of the function of the Abd-B transcription factor, its activation in these thoracic cells changes their fate to that of posterior abdomen. In this way, we can understand the molecular basis of a dominant mutation.

Gene fusions are an extremely important source of genetic variation. Through chromosomal rearrangements, novel patterns of gene expression can be generated. In fact, we can imagine that such fusions might play an important role in the shifts of gene expression pattern in the divergence and evolution of species. In addition to affecting development (discussed in Chapter 16), such mutations can play a pivotal role in the formation and progression of many cancers (discussed in Chapter 15).