A significant difference in the transcription of eukaryotic and prokaryotic mRNAs is that initiation at a eukaryotic promoter involves a large number of factors that bind to a variety of cis-acting elements. The promoter is defined as the region containing all these binding sites, that is, which can support transcription at the normal efficiency and with the proper control. So the major feature defining the promoter for a eukaryotic mRNA is the location of binding sites for transcription factors. RNA polymerase itself binds around the startpoint, but does not directly contact the extended upstream region of the promoter. By contrast, the bacterial promoters discussed in 9 Transcription are largely defined in terms of the binding site for RNA polymerase in the immediate vicinity of the startpoint.

Accessory factors are needed for initiation, but are not required subsequently. In each case, the factors, rather than the enzymes themselves, are principally responsible for recognizing the promoter. This is different from bacterial RNA polymerase, where it is the enzyme that recognizes the promoter sequences.

• The basal factors are required for the mechanics of initiating RNA synthesis at all promoters. They join with RNA polymerase to form a complex surrounding the startpoint, and they determine the site of initiation. The basal factors together with RNA polymerase constitute the basal transcription apparatus.
• Activators are transcription factors that recognize specific short consensus elements located upstream of the startpoint. They act by increasing the efficiency with which the basal apparatus binds to the promoter. They therefore increase the frequency of transcription, , and are required for a promoter to function at an adequate level. Other activators function in the same general way as the upstream factors, but have a regulatory role. They are synthesized or activated at specific times or in specific tissues, and they are therefore responsible for the control of transcription patterns in time and space. The sequences that they bind are called response elements.
• Another group of factors necessary for efficient transcription do not themselves bind DNA. Instead, co-activators work by protein-protein interactions, forming bridges between activators and the basal transcription apparatus.

Sequence components of the promoter are defined operationally by the demand that they must be located in the general vicinity of the startpoint and are required for initiation. The enhancer is another type of site involved in initiation. It is identified by sequences that stimulate initiation, but that are located a considerable distance from the startpoint. Enhancer elements are often targets for tissue-specific or temporal regulation. Figure 20.1 illustrates the general properties of promoters and enhancers.

A eukaryotic transcription unit generally contains a single gene, and termination occurs beyond the end of the coding region. We should like to define the mechanism of termination, but it lacks the regulatory importance that applies in prokaryotic systems. RNA polymerases I and III terminate at discrete sequences in defined reactions, but the mode of termination by RNA polymerase II is not clear. However, the significant event in generating the 3 end of an mRNA is not the termination event itself, but instead results from a cleavage reaction in the primary transcript (see 22 Nuclear splicing and RNA processing).

Figure 20.1 A typical gene transcribed by RNA polymerase II has a promoter that extends upstream from the site where transcription is initiated. The promoter contains several short (<10 bp) sequence elements that bind transcription factors, dispersed over >200 bp. An enhancer containing a more closely packed array of elements that also bind transcription factors may be located several kb distant. (DNA may be coiled or otherwise rearranged so that transcription factors at the promoter and at the enhancer interact to form a large protein complex.)

The three eukaryotic RNA polymerases have different locations in the nucleus, corresponding with the genes that they transcribe (for review see 71).

All eukaryotic RNA polymerases are large proteins, appearing as aggregates of >500 kD. They typically have ~12 subunits. The purified enzyme can undertake template-dependent transcription of RNA, but is not able to initiate selectively at promoters. The general constitution of a eukaryotic RNA polymerase II enzyme as typified in S. cerevisiae is illustrated in Figure 20.2. The two largest subunits are homologous to the and subunits of bacterial RNA polymerase. Three of the remaining subunits are common to all the RNA polymerases, that is, they are also components of RNA polymerases I and III (for review see 222).

Ten subunits of the yeast RNA polymerase II have been located on the crystal structure, as shown in Figure 20.3 (1714). The catalytic site is formed by a cleft between the two large subunits (#1 and #2), which hold DNA downstream in "jaws". The structure is generally similar to that of bacterial RNA polymerase. This can be seen more clearly in the crystal structure of Figure 20.4. RNA polymerase surrounds the DNA, as seen in the view of Figure 20.5. A catalytic Mg²⁺ ion is found at the active site. The DNA is clamped in position at the active site by subunits 1, 2, and 6. Access of nucleotides to the active site from above (in the orientation of Figure 20.5) is probably blocked by the DNA, and they may come from below via pores through the structure. Figure 20.6 shows that DNA is forced to take a turn at the entrance to the site, because of adjacent wall of protein. Subunits 4 and 7 are missing from this structure; they form a subcomplex that dissociates from the complete enzyme.

The largest subunit in RNA polymerase II has a carboxy-terminal domain (CTD), which consists of multiple repeats of a consensus sequence of 7 amino acids. The sequence is unique to RNA polymerase II. There are ~26 repeats in yeast and ~50 in mammals. The number of repeats is important, because deletions that remove (typically) more than half of the repeats are lethal (in yeast). The CTD can be highly phosphorylated on serine or threonine residues; this is involved in the initiation reaction (see later).

A major practical distinction between the eukaryotic enzymes is drawn from their response to the bicyclic octapeptide -amanitin. In basically all eukaryotic cells the activity of RNA polymerase II is rapidly inhibited by low concentrations of -amanitin. RNA polymerase I is not inhibited. The response of RNA polymerase III to -amanitin is less well conserved; in animal cells it is inhibited by high levels, but in yeast and insects it is not inhibited.

Figure 20.2 Some subunits are common to all classes of eukaryotic RNA polymerases and some are related to bacterial RNA polymerase.

Figure 20.3 Ten subunits of RNA polymerase are placed in position from the crystal structure. The color of the subunits is the same as the color of the corresponding parts of the crystal structure. Note that movement is from right to left. The colors of the subunits are the same as in the crystal structures of the following figures. The numbers identify the individual subunits.

Figure 20.4 The side view of the crystal structure of RNA polymerase II from yeast shows that DNA is held downstream by a pair of jaws and is clamped in position in the active site, which contains an Mg⁺⁺ ion. Photograph kindly provided by Roger Kornberg. (See 1714.)

Figure 20.5 The end view of the crystal structure of RNA polymerase II from yeast shows that DNA is surrounded by ~270° of protein. Photograph kindly provided by Roger Kornberg. (See 1714.)

Figure 20.6 DNA is forced to make a turn out of the active site by a wall of protein. Nucleotides may enter the active site through a pore in the protein. (See 1714.)

• In the oocyte system, a DNA template is injected into the nucleus of the X. laevis oocyte. The RNA transcript can be recovered and analyzed. The main limitation of this system is that it is restricted to the conditions that prevail in the oocyte. It allows characterization of DNA sequences, but not of the factors that normally bind them.
• Transfection systems allow exogenous DNA to be introduced into a cultured cell and expressed. (The procedure is discussed in 17 Rearrangement of DNA.) The system is genuinely in vivo in the sense that transcription is accomplished by the same apparatus responsible for expressing the cell’s own genome. However, it differs from the natural situation because the template consists of a gene that would not usually be transcribed in the host cell. The usefulness of the system may be extended by using a variety of host cells.
• Transgenic systems involve the addition of a gene to the germline of an animal. Expression of the transgene can be followed in any or all of the tissues of the animal. Some common limitations apply to transgenic systems and to transfection: the additional gene often is present in multiple copies, and is integrated at a different location from the endogenous gene. Discrepancies between the expression of a gene in vitro and its expression as a transgene can yield important information about the role of the geneomic context of the gene.
• The in vitro system takes the classic approach of purifying all the components and manipulating conditions until faithful initiation is seen. "Faithful" initiation is defined as production of an RNA starting at the site corresponding to the 5 end of mRNA (or rRNA or tRNA precursors). Ultimately this allows us to characterize the individual sequence elements in the promoter and the transcription factors that bind to them.

When a promoter is analyzed, it is imporant that only the promoter sequence changes. Figure 20.7 shows that the same long upstream sequence is always placed next to the promoter to ensure that it is always in the same context, . Because termination does not occur properly in the in vitro systems, the template is cut at some distance from the promoter (usually ~500 bp downstream), to ensure that all polymerases "run off" at the same point, generating an identifiable transcript.

We start with a particular fragment of DNA that can initiate transcription in one of these systems. Then the boundaries of the sequence constituting the promoter can be determined by reducing the length of the fragment from either end, until at some point it ceases to be active, as illustrated in Figure 20.8. The boundary upstream can be identified by progressively removing material from this end until promoter function is lost. To test the boundary downstream, it is necessary to reconnect the shortened promoter to the sequence to be transcribed (since otherwise there is no product to assay).

Figure 20.7 A promoter is tested by modifying the sequence that is connected to a constant upstream sequence and a constant downstream transcription unit.

Figure 20.8 Promoter boundaries can be determined by making deletions that progressively remove more material from one side. When one deletion fails to prevent RNA synthesis but the next stops transcription, the boundary of the promoter must lie between them.

RNA polymerase I transcribes only the genes for ribosomal RNA, from a single type of promoter. The transcript includes the sequences of both large and small rRNAs, which are later released by cleavages and processing. There are many copies of the transcription unit, alternating with nontranscribed spacers, and organized in a cluster as discussed in 4 Clusters and repeats. The organization of the promoter, and the events involved in initiation, are illustrated in Figure 20.9.

The promoter consists of two separate regions. The core promoter surrounds the startpoint, extending from –45 to +20, and is sufficient for transcription to initiate. It is generally G·C-rich (unusual for a promoter) except for the only conseved sequence element, a short A·T-rich sequence around the startpoint called the RInr. However, its efficiency is very much increased by the upstream promoter element (UPE), another G·C-rich sequence, related to the core promoter sequence, and which extends from –180 to –107. This type of organization is common to pol I promoters in many species, although the actual sequences vary widely (for review see 1672).

For high frequency initiation, the factor UBF is required. This is a single polypeptide that binds to a G·C-rich element in the upstream promoter element. One indication of how UBF interacts with the core-binding factor is given by the importance of the spacing between the upstream promoter element and the core promoter. This can be changed by distances involving integral numbers of turns of DNA, but not by distances that introduce half turns. This implies that UBF and core-binding factor need to be bound on the same face of DNA in order to interact. In the presence of UBF, core-binding factor binds more efficiently to the core promoter (1676).

Figure 20.9 Transcription units for RNA polymerase I have a core promoter separated by ~70 bp from the upstream promoter element. UBF binding to the UPE increases the ability of core-binding factor to bind to the core promoter. Core-binding factor positions RNA polymerase I at the startpoint.

When the deletions continue into the gene, a product very similar in size to the usual 5S RNA continues to be synthesized so long as the deletion ends before base +55. Figure 20.10 shows that the first part of the RNA product corresponds to plasmid DNA; the second part represents the segment remaining of the usual 5S RNA sequence. But when the deletion extends past +55, transcription does not occur. So the promoter lies downstream of position +55, but causes RNA polymerase III to initiate transcription a more or less fixed distance away.

So the promoter for 5S RNA transcription lies between positions +55 and +80 within the gene. A fragment containing this region can sponsor initiation of any DNA in which it is placed, from a startpoint ~55 bp farther upstream. (The wild-type startpoint is unique; in deletions that lack it, transcription initiates at the purine base nearest to the position 55 bp upstream of the promoter. (641, 642))

The structures of three types of promoters for RNA polymerase III are summarized in Figure 20.11. There are two types of internal promoter. Each contains a bipartite structure, in which two short sequence elements are separated by a variable sequence. Type 1 consists of a boxA sequence separated from a boxC sequence (1673), and type 2 consists of a boxA sequence separated from a boxB sequence (1674). The distance between boxA and boxB in a type 2 promoter can vary quite extensively, but the boxes usually cannot be brought too close together without abolishing function. Type 3 promoters have 3 sequences elements all located upstream of the startpoint (1675).

Figure 20.12 summarizes the stages of reaction at type 2 internal promoters. TFIIIC binds to both boxA and boxB. This enables TFIIIB to bind at the startpoint. Then RNA polymerase III can bind.

The difference at type 1 internal promoters is that TFIIIA must bind at boxA to enable TFIIIC to bind at boxC. Figure 20.13 shows that, once TFIIIC has bound, events follow the same course as at type 2 promoters, with TFIIIB binding at the startpoint, and RNA polymerase III joining the complex. Type 1 promoters are found only in the genes for 5S rRNA.

TFIIIA and TFIIIC are assembly factors, whose sole role is to assist the binding of TFIIIB at the right location. Once TFIIIB has bound, TFIIIA and TFIIIC can be removed from the promoter (by high salt concentration in vitro) without affecting the initiation reaction. TFIIIB remains bound in the vicinity of the startpoint and its presence is sufficient to allow RNA polymerase III to bind at the startpoint. So TFIIIB is the only true initiation factor required by RNA polymerase III (643). This sequence of events explains how the promoter boxes downstream can cause RNA polymerase to bind at the startpoint, farther upstream. Although the ability to transcribe these genes is conferred by the internal promoter, changes in the region immediately upstream of the startpoint can alter the efficiency of transcription.

TFIIIC is a large protein complex (>500 kD), comparable in size to RNA polymerase itself, and containing 6 subunits. TFIIIA is a member of an interesting class of zinc finger proteins that we discuss later. The positioning factor, TFIIIIB, consists of three subunits. It includes the same protein, TBP, that is present in the core-binding factor for pol I promoters, and also in the corresponding transcription factor (TFIID) for RNA polymerase II (1677). It also contains Brf, which is related to the factor TFIIB that is used by RNA polymerase II. The third subunit is called B"; it is dispensable if the DNA duplex is partially melted, which suggests that its function is to initiate the transcription bubble (945). The role of B" may be comparable to the role played by sigma factor in bacterial RNA polymerase (see 9.15 Substitution of sigma factors may control initiation).

The upstream region has a conventional role in the third class of polymerase III promoters. In the example shown in Figure 20.11, there are three upstream elements. These elements are also found in promoters for snRNA genes that are transcribed by RNA polymerase II. (Genes for some snRNAs are transcribed by RNA polymerase II, while others are transcribed by RNA polymerase III.) The upstream elements function in a similar manner in promoters for both polymerases II and III.

The factors work in the same way for both types of promoters for RNA polymerase III. The factors bind at the promoter before RNA polymerase itself can bind. They form a preinitiation complex that directs binding of the RNA polymerase. RNA polymerase III does not itself recognizes the promoter sequence, but binds adjacent to factors that are themselves bound just upstream of the startpoint. For the type 1 and type 2 internal promoters, the assembly factors ensure that TFIIIB (which includes TBP) is bound just upstream of the startpoint, to provide the positioning information. For the upstream promoters, transcription factors that directly recognize the upstream sites form a complex (including TBP) that is recognized by RNA polymerase III. So irrespective of the location of the promoter sequences, factor(s) are bound close to the startpoint in order to direct binding of RNA polymerase III (for review see 220).

This section updated 4-30-2001

Figure 20.10 Deletion analysis shows that the promoter for 5S RNA genes is internal; initiation occurs a fixed distance (~55 bp) upstream of the promoter.

Figure 20.11 Promoters for RNA polymerase III may consist of bipartite sequences downstream of the startpoint, with boxA separated from either boxC or boxB. Or they may consist of separated sequences upstream of the startpoint (Oct, PSE, TATA).

Figure 20.12 Internal type 2 pol III promoters use binding of TFIIIC to boxA and boxB sequences to recruit the positioning factor factor TFIIIB, which recruits RNA polymerase III.

Figure 20.13 Internal type 1 pol III promoters use the assembly factors TFIIIA and TFIIIC, at boxA and boxC, to recruit the positioning factor factor TFIIIB, which recruits RNA polymerase III.