Vol. 16, No. 20, pp. 2593-2620, October 15, 2002

REVIEW
Recruitment of RNA polymerase III to its target promoters

Howard Hughes Medical Institute, Cold Spring Harbor Laboratory, Cold Spring Harbor, New York 11724, USA

	Introduction

Top Introduction Structure of RNA polymerase... The assembly pathways directed... RNA polymerase III Composition of TFIIIB Functions of TFIIIB Recruitment factors: TFIIIA Recruitment factors: TFIIIC Assembly of a stable... Recruitment factors: SNAPc Assembly of a stable... Stepwise assembly versus... Human TFIIIC1 Termination and recycling: does... Conclusion References

A key step in retrieving the information stored in the complex genomes of eukaryotes involves the identification of transcription units and, more specifically, the recognition of promoter sequences by RNA polymerase. In eukaryotes, the task of recognizing nuclear gene promoters and then transcribing the genes is divided among three highly related enzymes, RNA polymerases I, II, and III. Each of these RNA polymerases is dedicated to the transcription of specific sets of genes, and each depends on accessory factors, the so-called transcription factors, to recognize its cognate promoter sequences.

RNA polymerase I is unique among the nuclear RNA polymerases in transcribing only one set of genes, the large, tandemly repeated, ribosomal RNA genes, and thus in having to recognize a single promoter structure. RNA polymerase II transcribes the protein-coding genes (mRNA genes) as well as some small nuclear RNA (snRNA) genes. The RNA polymerase II promoters can be divided into a core region, defined as the minimal region capable of directing transcription in vitro, and a regulatory region. The regulatory regions are highly varied in structure, reflecting the highly varied synthesis patterns of cellular proteins and the need for exquisite and complex regulation of these patterns. The core promoters themselves come in different types that, in mRNA-encoding genes, can contain a TATA box, an initiator, a downstream promoter element, or various combinations thereof. The assembly of a functional RNA polymerase II transcription complex on a promoter consisting of just a TATA box has been extensively studied. All the factors involved in the process have been identified, and much is known about how these factors interact with DNA and with each other to recruit, eventually, RNA polymerase II (for reviews, see Orphanides et al. 1996; Woychik and Hampsey 2002). How RNA polymerase II transcription complexes assemble on TATA-less promoters is, however, not as well understood.

RNA polymerase III is dedicated to the transcription of an eclectic collection of genes whose main common features are that they encode structural or catalytic RNAs and that they are, as a rule, shorter than 400 base pairs (bp). This length limit is consistent with the elongation properties of RNA polymerase III, which recognizes a simple run of T residues as a termination signal. The genes transcribed by RNA polymerase III encode RNA molecules involved in fundamental metabolic processes, specifically components of the protein synthesis apparatus and components of the splicing and tRNA processing apparatus, as well as RNAs of unknown function. The RNA polymerase III promoters are more varied in structure than the uniform RNA polymerase I promoters, and yet not as diverse as the RNA polymerase II promoters. They have been divided into three main types, two of which are gene-internal and generally TATA-less, and one of which is gene-external and contains a TATA box. Remarkably, we have a good, and in some cases a very detailed, understanding of how RNA polymerase III is recruited to each of these types of promoters. This provides a paradigm of how the same enzyme can be recruited to different promoter structures through different combinations of protein-DNA and protein-protein interactions. Here we summarize our present understanding of the various pathways leading to recruitment of RNA polymerase III. Other recent reviews on RNA polymerase III transcription include those by Geiduschek and Kassavetis (2001) and Huang and Maraia (2001). Reviews on the regulation of RNA polymerase III transcription, which is not covered here, include those by Ghavidel et al. (1999) and Brown et al. (2000).

	Structure of RNA polymerase III promoters

Top Introduction Structure of RNA polymerase... The assembly pathways directed... RNA polymerase III Composition of TFIIIB Functions of TFIIIB Recruitment factors: TFIIIA Recruitment factors: TFIIIC Assembly of a stable... Recruitment factors: SNAPc Assembly of a stable... Stepwise assembly versus... Human TFIIIC1 Termination and recycling: does... Conclusion References

The three types of RNA polymerase III promoters are called types 1-3. The first RNA polymerase III promoters characterized were those of the Xenopus laevis 5S RNA gene (Bogenhagen et al. 1980; Sakonju et al. 1980), which encodes the small ribosomal RNA, the Adenovirus 2 (Ad2) VAI gene (Fowlkes and Shenk 1980), and various tRNA genes from X. laevis and Drosophila melanogaster (Galli et al. 1981; Hofstetter et al. 1981; Sharp et al. 1981). The 5S promoter is the only example of a type 1 RNA polymerase III promoter, and the Ad2 VAI and tRNA promoters are typical type 2 promoters. As shown in Figure 1, these promoters are intragenic. The X. laevis 5S gene promoter consists of an A box, an intermediate element (IE), and a C box that is conserved in the 5S promoters of different species. Together, these elements constitute the internal control region (ICR; Bogenhagen 1985; Pieler et al. 1985a,b, 1987). In the Saccharomyces cerevisiae 5S genes, only the C box is required for transcription (Challice and Segall 1989).

View larger version (11K): [in this window] [in a new window]   Figure 1.   Different types of RNA polymerase III promoters. The type 1 promoter of the Xenopus laevis 5S RNA gene consists of an internal control region (ICR), which can be subdivided into A box (+50 to +60), intermediate element (IE, +67 to +72), and C box (+80 to +90). The type 2 promoter of the X. laevis tRNALeu gene consists of an A box (+8 to +19) and a B box (+52 to +62). The type 3 promoter of the Homo sapiens U6 snRNA gene consists of a distal sequence element (DSE, 215 to 240) that enhances transcription and a core promoter composed of a proximal sequence element (PSE, 65 to 48) and a TATA box (32 to 25). The Saccharomyces cerevisiae promoter is a hybrid promoter consisting of a TATA box (30 to 23), an A box (+21 to +31), and a B box located downstream of the U6 coding region (from +234 to +244 relative to the start site of transcription).

The Ad2 VAI and most tRNA promoters consist of an A box and a B box (Galli et al. 1981; Hofstetter et al. 1981; Sharp et al. 1981; Allison et al. 1983). These are well conserved in tRNA genes from various species, probably in part because they encode the tRNA D- and T-loops, which are required for tRNA function. The spacing between the A- and B-boxes varies greatly, however, in part to accommodate introns. The A-boxes of type 1 and 2 promoters are structurally related and are interchangeable in X. laevis (Ciliberto et al. 1983). However, this apparently reflects a similarity in sequence rather than a conserved function, because, as detailed below, the A-boxes of 5S and tRNA genes bind different transcription factors (Braun et al. 1992a).

The type 3 core promoters were identified originally in mammalian U6 snRNA genes, which encode the U6 snRNA component of the spliceosome (Krol et al. 1987; Das et al. 1988; Kunkel and Pederson 1988), and in the human 7SK gene (Murphy et al. 1986), whose RNA product has been recently implicated in the regulation of the CDK9/cyclin T complex (Nguyen et al. 2001; Yang et al. 2001). They are also found in, for example, the H1 RNA gene, which encodes the RNA component of human RNase P (Baer et al. 1989), and the gene encoding the RNA component of human RNase MRP (Topper and Clayton 1990), as well as in genes encoding RNAs of unknown function. Their discovery came as a surprise because, unlike the then-characterized type 1 and 2 promoters, the type 3 core promoters turned out to be gene-external. As illustrated in Figure 1, they are located in the 5'-flanking region of the gene and consist of a proximal sequence element (PSE), which also constitutes, on its own, the core of RNA polymerase II snRNA promoters, and a TATA box located at a fixed distance downstream of the PSE (Hernandez and Lucito 1988; Mattaj et al. 1988; Kunkel and Pederson 1989; Lobo and Hernandez 1989). Strikingly, in the vertebrate snRNA promoters, RNA polymerase specificity can be switched from RNA polymerase III to RNA polymerase II and vice versa by abrogation or generation of the TATA box (Lobo and Hernandez 1989). Upstream of the PSE is an element referred to as the distal sequence element (DSE), which activates transcription from the core promoter.

Although the presence of a TATA box is the hallmark of type 3, gene-external, promoters, it is also found in the 5'-flanking regions of some genes with gene-internal promoter elements. Figure 1 shows an example of such a hybrid promoter, namely the S. cerevisiae U6 snRNA promoter. It consists of an A box, a B box located at an unusual position 120 bp downstream of the RNA coding region, and a TATA box located upstream of the transcription start site. All three of these promoter elements are required for efficient transcription in vivo (Brow and Guthrie 1990; Eschenlauer et al. 1993). Other examples include some A- and B-box-containing tRNA genes in plants (Yukawa et al. 2000), yeast (Dieci et al. 2000), and silkworm (Ouyang et al. 2000), in which TATA boxes present in the 5'-flanking region greatly contribute to transcription efficiency. More recently, an analysis in Schizosaccharomyces pombe has revealed that in this organism nearly all tRNA and 5S genes contain a TATA box upstream of the transcription start site that is required for transcription (Hamada et al. 2001). Strikingly, in vitro, artificial promoters consisting of just a TATA box can direct RNA polymerase III transcription, indicating that under these circumstances, the TATA box contains all necessary information to assemble an RNA polymerase III transcription initiation complex (Mitchell et al. 1992; Roberts et al. 1995; Wang and Stumph 1995; Whitehall et al. 1995; Huang et al. 1996).

	The assembly pathways directed by the different types of RNA polymerase III promoters converge on recruitment of TFIIIB and RNA polymerase III

Top Introduction Structure of RNA polymerase... The assembly pathways directed... RNA polymerase III Composition of TFIIIB Functions of TFIIIB Recruitment factors: TFIIIA Recruitment factors: TFIIIC Assembly of a stable... Recruitment factors: SNAPc Assembly of a stable... Stepwise assembly versus... Human TFIIIC1 Termination and recycling: does... Conclusion References

The characterization of RNA polymerase III transcription factors started with the fractionation of a HeLa cell extract over a phosphocellulose column into three fractions known as fractions A (the phosphocellulose 100 mM KCl flowthrough), B (a 100 mM-350 mM KCl step elution), and C (a 350 mM-600 mM KCl step elution), and the observation that transcription from type 2 promoters required fractions B and C, whereas transcription from type 1 promoters required the three fractions (Segall et al. 1980). After the type 3 promoters were discovered, they were shown to require the B and C fractions, or the B fraction and a D fraction eluted between 600 mM and 1000 mM KCl from the phosphocellulose column (Lobo et al. 1991). Most of the activities in these fractions required for RNA polymerase III transcription have now been characterized, both from yeast and human cells. Figure 2 shows, in a highly simplified manner, how these factors can assemble in an ordered fashion to recruit RNA polymerase III. The green arrows symbolize interactions of DNA-binding proteins with promoter elements, the blue arrows protein-protein contacts among various transcription factors, and the purple arrows protein-protein contacts between RNA polymerase III and transcription factors.

View larger version (18K): [in this window] [in a new window]   Figure 2.   Different pathways for recruitment of TFIIIB and RNA polymerase III. The initiation complexes formed on type 2, 1, and 3 promoters, as well as on an artificial promoter consisting of just a TATA box, are shown. The green arrows symbolize interactions of DNA-binding proteins with promoter elements, the blue arrows protein-protein contacts among various transcription factors, and the purple arrows protein-protein contacts between RNA polymerase III and transcription factors.

In type 2 promoters, the A and B boxes are recognized by a multisubunit complex in the C fraction called TFIIIC or TFIIIC2 (Lassar et al. 1983). This initial DNA-protein interaction then allows the recruitment of an activity in the B fraction called TFIIIB (Bieker et al. 1985; Setzer and Brown 1985). TFIIIB is composed of three polypeptides, one of which is the TATA-box-binding protein TBP. The binding of TFIIIB to the promoter in turn allows the recruitment of RNA polymerase III, mainly through protein-protein interactions with TFIIIB, although contacts with TFIIIC may also contribute (Fig. 2A). In type 1 promoters, the ICR is recognized by the activity present in the A fraction, a zinc finger protein referred to as TFIIIA (Engelke et al. 1980; Sakonju et al. 1981). Formation of the TFIIIA-DNA complex then allows for the binding of TFIIIC (Lassar et al. 1983). Thus, TFIIIA can be viewed as a specificity factor that alters the promoter-recognition properties of TFIIIC and targets it to the 5S promoter. After the binding of TFIIIC, the pathway to recruitment of the polymerase is similar to that in type 2 promoters, with the recruitment of TFIIIB and RNA polymerase III (Fig. 2B). In type 3 promoters, the PSE is recognized by a multisubunit complex variously called the PSE-binding protein (PBP), the PSE transcription factor (PTF), or the snRNA activating protein complex (SNAP_c), and the TATA box is recognized by the TBP component of a specialized TFIIIB-like activity (Waldschmidt et al. 1991; Murphy et al. 1992; Sadowski et al. 1993; Yoon et al. 1995; Schramm et al. 2000; Teichmann et al. 2000). These DNA-protein interactions are reinforced by protein-protein interactions between SNAP_c and TBP (Mittal and Hernandez 1997; Ma and Hernandez 2002). The binding of SNAP_c and the TFIIIB-like activity then lead to recruitment of RNA polymerase III (Sepehri Chong et al. 2001), probably through protein-protein contacts with the two DNA-bound factors, SNAP_c and the TFIIIB-like activity, although this has not yet been demonstrated (Fig. 2C).

Figure 2 also shows a recruitment pathway in which TFIIIB is directly recruited to a TATA box without the help of protein-protein contacts with either TFIIIC or SNAP_c (Fig. 2D). This pathway can be observed in vitro with S. cerevisiae TFIIIB, and, although it is not observed in vivo, it reveals a profound aspect of RNA polymerase III transcription, namely, that TFIIIB is sufficient for RNA polymerase recruitment. TFIIIB was first identified as the key RNA polymerase III transcription factor by a series of experiments in which S. cerevisiae TFIIIB was first recruited to either a 5S promoter through prior binding of TFIIIA and TFIIIC, or a tRNA promoter through prior binding of TFIIIC (Kassavetis et al. 1990). TFIIIA and/or TFIIIC were then stripped from the DNA by treatment with heparin or high concentrations of salt. Under these conditions, functional TFIIIA and TFIIIC were released from the templates, but remarkably, TFIIIB remained bound to the DNA, generating the same footprint upstream of the transcription start site as it did in the presence of TFIIIA and/or TFIIIC (Kassavetis et al. 1989, 1990). These stripped templates were able to support several rounds of properly initiated RNA polymerase III transcription. This suggested that, at least in yeast, TFIIIB was sufficient to recruit RNA polymerase III and direct several rounds of transcription, and therefore that the main function of TFIIIA and TFIIIC was to recruit TFIIIB to the DNA (Kassavetis et al. 1990). With the observation that just a TATA box could direct RNA polymerase III transcription in vitro and with the availability of recombinant TFIIIB, it then became possible to confirm that a TATA box could direct several rounds of RNA polymerase III transcription with just recombinant TFIIIB and highly purified RNA polymerase III (Kassavetis et al. 1995; Rüth et al. 1996). Thus, TFIIIA, TFIIIC, and SNAP_c can be viewed as recruitment factors whose main function is to recruit TFIIIB to promoters of various structures, which then allows the recruitment of RNA polymerase III.

The different RNA polymerase III promoters offer a unique system in which we know how different accessory factors combine to recruit, ultimately, a TFIIIB activity and RNA polymerase III. Below, we first describe briefly the subunit composition of RNA polymerase III. For a discussion of the likely three-dimensional structure of RNA polymerase III, see Geiduschek and Kassavetis (2001). We then describe the characterization of its key transcription factor, TFIIIB, both yeast and human, and we summarize our present understanding of how this factor bridges DNA and RNA polymerase III. We then summarize what is known about the various factors that, in vivo, mediate the recruitment of TFIIIB on most, if not all, promoters, namely, TFIIIA, TFIIIC, and SNAP_c. We end with a description of some factors that have been implicated in termination and recycling of RNA polymerase III.

	RNA polymerase III

Top Introduction Structure of RNA polymerase... The assembly pathways directed... RNA polymerase III Composition of TFIIIB Functions of TFIIIB Recruitment factors: TFIIIA Recruitment factors: TFIIIC Assembly of a stable... Recruitment factors: SNAPc Assembly of a stable... Stepwise assembly versus... Human TFIIIC1 Termination and recycling: does... Conclusion References

RNA polymerase III is well defined in S. cerevisiae, consisting of 17 subunits, as shown in Table 1. All the corresponding genes except for RPC37 have been disrupted and shown to be essential (for review, see Chedin et al. 1998). Of the 17 subunits, 10 are unique to RNA polymerase III and are designated the C subunits, two are common to RNA polymerases I and III and are designated AC subunits, and five are common to the three RNA polymerases and are designated ABC subunits. The common subunits have different names in RNA polymerases I and II, as indicated in Table 1 for RNA polymerase II. C160, C128, AC40, AC19, and ABC23 are evolutionarily related to the core subunits of Escherichia coli RNA polymerase, as indicated in parentheses in the table. Of the C subunits, five, indicated in bold in Table 1, are specific to RNA polymerase III.

Human RNA polymerase III has been purified both by conventional chromatography (Wang and Roeder 1996) and from cell lines expressing tagged Homo sapiens (Hs) RPC4/RPC53/BN51 (Wang and Roeder 1997), but until recently, only five of its subunits had been characterized: HsRPC4/RPC53 (Ittmann et al. 1993; Jackson et al. 1995), HsRPC1/RPC155 (Sepehri and Hernandez 1997), HsRPC3/RPC62, HsRPC6/RPC39, and HsRPC7/RPC32 (Wang and Roeder 1997). Human RNA polymerase III has now been purified from a stable cell line expressing a doubly tagged HsRPC4 subunit, and its subunits have been identified by mass spectrometry (Hu et al. 2002). This analysis has resulted in the identification of orthologs of all of the yeast RNA polymerase III subunits except for ABC10, which was not detected probably because of its small size (7 kD). The newly described human subunits were named according to the guide shown in the fourth column of Table 1, in which the yeast C, AC, and ABC subunits were numbered separately in order of decreasing apparent molecular weight. Such a nomenclature would provide the same name for orthologs from different species, as shown in the fourth and sixth column in Table 1.

The characterization of human RPC8 and RPC9 brought an unexpected result. BLAST searches revealed that HsRPC8 is related to the RNA polymerase II subunit RPB7, as noted earlier for the S. cerevisiae HsRPC8 ortholog C25 (Sadhale and Woychik 1994). In addition, however, HsRPC9 is related to RPB4, and like RPB4 and RPB7, which associate with each other and form a dimer detachable from the rest of RNA polymerase II (Edwards et al. 1991; Khazak et al. 1998), HsRPC8 and HsRPC9 associate with each other (Hu et al. 2002). This strongly suggests that HsRPC8 and HsRPC9 are paralogs of RPB7 and RPB4, as indicated in Table 1, and that the corresponding S. cerevisiae RNA polymerase III subunits C25 and C17 can similarly associate with each other.

RPB7, but not RPB4, is essential for yeast cell viability (Woychik and Young 1989; McKune et al. 1993). RPB4 is, however, essential for cellular responses to stress (Choder and Young 1993) and thus in vivo, the requirement for the RPB4 subunit may be promoter-specific. The RPB4/RPB7 complex is thought to stabilize the open promoter complex and perhaps the early transcribing complex prior to promoter escape by binding to nascent RNA or to single-stranded DNA in the transcription bubble (Orlicky et al. 2001; Todone et al. 2001). The S. cerevisiae RNA polymerase III paralogs of RBP7 and RPB4, C25 and C17, are both essential for viability in yeast (Sadhale and Woychik 1994; Ferri et al. 2000). Two-hybrid and coimmunoprecipitation experiments indicate that C17 interacts with the transcription initiation factor Brf1 and with the RNA polymerase III C31 subunit (Ferri et al. 2000), which, as described below, is itself required for transcription initiation (Werner et al. 1992, 1993; Wang and Roeder 1997). Thus, the RNA polymerase III paralogs of RPB4 and RPB7 may also be involved in transcription initiation, but in this case both subunits are essential for yeast cell viability, perhaps because most RNA polymerase III genes encode components essential for cell metabolism.

The human RNA polymerase III subunits are in general quite similar to their yeast counterparts with the notable exception of the subunits with no paralogs in RNA polymerase II (Jackson et al. 1995; Wang and Roeder 1997; Hu et al. 2002). For example, the human ortholog of yeast C37 is an 80-kD protein, HsRPC5, whose similarity to the yeast protein is confined to its N-terminal fourth, which shows 26% identity with C37 (Hu et al. 2002). Nevertheless, like yeast C37, which associates with the yeast C53 subunit (Flores et al. 1999), HsRPC5 associates with HsRPC4/RPC53, the human ortholog of yeast C53, and this association is through the HsRPC5 and HsRPC4/RPC53 domains conserved in their yeast counterparts (Hu et al. 2002). Interestingly, at least some of the subunits with no paralogs in RNA polymerase II seem to be involved in promoter recognition. The C82, C34, and C31 subunits (bold and underlined in Table 1) dissociate from a yeast enzyme carrying a mutation within the zinc finger domain of the largest subunit, and each associates with the two others in a yeast two-hybrid assay, suggesting that these three subunits form a subcomplex detachable from the rest of the enzyme (Werner et al. 1992, 1993). In the human enzyme, such a subcomplex could be demonstrated directly by sucrose gradient centrifugation under partially denaturing conditions and by reconstitution of the subcomplex from recombinant subunits (Wang and Roeder 1997). The subunits in the subcomplex are not required for efficient elongation and termination, but are required for specific initiation (Thuillier et al. 1995; Brun et al. 1997; Wang and Roeder 1997). Consistent with this observation and as detailed further below, the C34 subunit and its human counterpart interact directly with TFIIIB subunits (Werner et al. 1993; Khoo et al. 1994; Wang and Roeder 1997).

	Composition of TFIIIB

Top Introduction Structure of RNA polymerase... The assembly pathways directed... RNA polymerase III Composition of TFIIIB Functions of TFIIIB Recruitment factors: TFIIIA Recruitment factors: TFIIIC Assembly of a stable... Recruitment factors: SNAPc Assembly of a stable... Stepwise assembly versus... Human TFIIIC1 Termination and recycling: does... Conclusion References

TBP is required for transcription by RNA polymerase III both in S. cerevisiae and human cells

Although the presence of an RNA polymerase III transcription activity in the phosphocellulose B fraction was recognized in the early 1980s, the composition of this activity remained a mystery for the next 10 years. By the late 1980s, however, the concept that the TATA-box-binding protein TBP was a factor uniquely dedicated to transcription by RNA polymerase II began to change with the finding that an essential element of the U6 promoter was an A/T-rich region, that is, a potential binding site for TBP. Biochemical fractionation and reconstitution experiments then identified TBP as a factor required for transcription of both the yeast and human U6 snRNA genes (Lobo et al. 1991; Margottin et al. 1991; Simmen et al. 1991), whose binding to wild-type and mutant U6 TATA boxes correlated with transcription activity (Lobo et al. 1991). These findings established that TATA boxes are part of at least some RNA polymerase III promoters, and that they act by recruiting TBP. They also raised the possibility that TBP might be required for RNA polymerase III transcription in general. Indeed, in vitro competition experiments with TATA-containing oligonucleotides then indicated that a TATA-box-binding factor was required for transcription of the VAI and tRNA genes (White et al. 1992), and inactivation of TBP in yeast was shown to lead to defects in transcription by all three RNA polymerases (Cormack and Struhl 1992; Schultz et al. 1992). The remaining question was how to place TBP in what was then known about RNA polymerase III transcription factors.

Identification of S. cerevisiae Brf1 and Bdp1

Yeast TFIIIB had been shown to consist of two chromatographically separable activities, named B` and B", which contained polypeptides of 70 and 90 kD, respectively, that could be cross-linked to the DNA (Bartholomew et al. 1991; Kassavetis et al. 1991). A major step in the complete characterization of TFIIIB came with the cloning of the gene encoding the 70-kD polypeptide, now referred to as Brf1 (TFIIB-related factor 1; for a description of a universal nomenclature of TFIIIB components, see Willis 2002). The gene was cloned as a suppressor of a tRNA gene A-box mutation and called PCF4 (López-De-León et al. 1992). It was also cloned, however, as an allele-specific high-copy suppressor of certain mutations in TBP and called BRF1 (Colbert and Hahn 1992) or TDS4 (Buratowski and Zhou 1992). This ability to suppress mutations in TBP suggested that Brf1 might be associated with TBP, and thus that TBP might be part of the TFIIIB activity. Indeed, TBP was shown to be part of both the yeast and mammalian TFIIIB activity by biochemical methods (Margottin et al. 1991; Huet and Sentenac 1992; Kassavetis et al. 1992; Lobo et al. 1992; Taggart et al. 1992; White and Jackson 1992; Chiang et al. 1993; Meyers and Sharp 1993), and to constitute a previously unrecognized, Brf1-associated, component of the B` activity (Kassavetis et al. 1992). The cloning of the gene encoding yeast B" (B", Kassavetis et al. 1995; TFIIIB90, Roberts et al. 1996; TFC7p, Rüth et al. 1996), now referred to as Bdp1 (B double prime 1, Willis 2002), then completed the characterization of S. cerevisiae TFIIIB.

Identification of human Brf1 and Brf2

In S. cerevisiae, all RNA polymerase III promoters recruit the same TFIIIB factor (Joazeiro et al. 1994). In higher eukaryotes, however, the situation is more complex, consistent with the need to transcribe much more complex genomes. Thus, the initial characterization of mammalian TFIIIB not only indicated that TBP was part of the activity (Lobo et al. 1992; Taggart et al. 1992; White and Jackson 1992), but also that type 1 and 2 promoters used different components in the TFIIIB fraction than type 3 promoters. Type 1 and 2 promoters were shown to require a TBP-containing complex (Lobo et al. 1992; Teichmann and Seifart 1995) consisting of TBP and a homolog of yeast Brf1 (Wang and Roeder 1995; Mital et al. 1996) referred to as HsBrf1 (Homo sapiens Brf1). Depletion of extracts with antibodies directed against the C-terminal half of HsBrf1 debilitated transcription from the type 2 VAI promoter, as expected, but had no effect on transcription from the type 3 human U6 snRNA promoter (Mital et al. 1996; Henry et al. 1998a). On the other hand, depletion of extracts with antibodies raised against full-length HsBrf1 or against a peptide derived from the N-terminal portion of the protein inhibited transcription from all types of RNA polymerase III promoters, although only transcription from type 1 and 2 promoters could be reconstituted by addition of recombinant HsBrf1 (Wang and Roeder 1995; Schramm et al. 2000). These observations suggested that type 3 promoters use a protein related to Brf1 in its N-terminal but not its C-terminal region, and led to the characterization of a new protein, originally called BRFU (Schramm et al. 2000) or TFIIIB50 (Teichmann et al. 2000), and now referred to as HsBrf2 (Willis 2002). Thus, S. cerevisiae Brf1 has at least two homologs in human cells, HsBrf1 and HsBrf2.

Figure 3 shows the structure of TFIIB and various Brf proteins. H. sapiens and S. cerevisiae (Sc) Brf1 as well as HsBrf2 contain, like TFIIB, an N-terminal zinc-binding domain (green box) and a "core domain" consisting of two imperfect repeats (blue box). In addition, the Brf1 and Brf2 proteins contain C-terminal domains absent in TFIIB. Within the C-terminal segment of Brf1, three regions, designated regions I, II, and III, are conserved in the yeasts Candida albicans, Kluyveromyces lactis, S. pombe, and S. cerevisiae (Khoo et al. 1994). Regions II and III are also conserved in the human Brf1 protein (Mital et al. 1996; Andrau et al. 1999). Consistent with the antibody depletion data, the C-terminal domain of HsBrf2 shows very little, if any, homology with Brf1.

View larger version (18K): [in this window] [in a new window]   Figure 3.   TFIIB, Brf1, and Brf2 form a family of related transcription factors. The location of the structured zinc ribbon as modeled in ScTFIIB and ScBrf1 (Hahn and Roberts 2000) based on the NMR structure of PfTFIIB (Zhu et al. 1996) and that of the corresponding region in HsBrf1 and HsBrf2 is indicated in green. The location of the structured core domain of TFIIB (Bagby et al. 1995; Nikolov et al. 1995) and that of the corresponding regions in the other proteins is indicated in blue. The percentages below the sequences indicate percent identities between HsBrf2 and HsTFIIB, ScTFIIB, ScBrf1, and HsBrf1 within the region of highest conservation (bracketed by the stippled lines) in pairwise alignments performed with BLAST. The purple boxes in the C-terminal regions of ScBrf1 and HsBrf1 indicate conserved regions I, II, and III (Mital et al. 1996). In HsBrf1_v2, the blue region is identical to the corresponding HsBrf1 region.

HsBrf2 was isolated through a database search for proteins related to TFIIB and to the TFIIB-related segment of Brf1 (Schramm et al. 2000), as well as through biochemical purification of a complex, consisting of HsBrf2 and four associated proteins, required for transcription from type 3 promoters (Teichmann et al. 2000). It is clear that HsBrf2 itself is specifically required for transcription from type 3, but not types 1 and 2, promoters, but the exact role of the HsBrf2-associated factors remains to be determined. Although in one case, U6 transcription in HsBrf2-depleted extracts could be restored only by addition of the HsBrf2-containing complex immunopurified from HeLa cells expressing tagged HsBrf2 (Teichmann et al. 2000), in another case it could be restored by addition of just HsBrf2 synthesized in E. coli (Schramm et al. 2000). This last observation suggests that the HsBrf2-associated polypeptides may not be absolutely required for U6 transcription but may contribute to the efficiency of the reaction.

Figure 3 also illustrates the structure of HsBrf1_v2 (originally named BRF2), a factor encoded by one of at least four alternatively spliced BRF1 pre-mRNAs (McCulloch et al. 2000). HsBrf1_v2 lacks the zinc finger domain and the first repeat that are present in Brf1 and conserved in the other proteins of the TFIIB family, as well as the C-terminal region present in Brf1. Although HsBrf1 is not involved in human U6 transcription, HsBrf1_v2 has been implicated in U6 transcription because when antibodies recognizing all HsBrf1 variants were used to deplete extracts, U6 transcription was lost and could be specifically restored by addition of material immunopurified from cells expressing tagged HsBrf1_v2 (McCulloch et al. 2000). It will be necessary to define the composition of this immunopurified fraction to confirm the role of HsBrf1_v2 in U6 transcription.

Identification of human Bdp1

Figure 4 shows the structure of S. cerevisiae Bdp1. It contains a domain related to a Myb repeat, identified in the SWI-SNF and ADA complexes, the transcriptional corepressor N-Cor, and yeast TFIIIB Bdp1, and therefore referred to as the SANT domain (Aasland et al. 1996). The SANT domain is absolutely required for TFIIIC-dependent (but not TFIIIC-independent, see below) RNA polymerase III transcription (Kumar et al. 1997). In addition, a region upstream of the SANT domain (indicated in orange in Fig. 4) is required for transcription from linear, but not supercoiled, templates (Kassavetis et al. 1998a).

View larger version (10K): [in this window] [in a new window]   Figure 4.   Comparison of the ScBdp1 and HsBdp1 polypeptides. The proteins contain a conserved SANT domain (brown box). The regions upstream and downstream of the SANT domain are also quite conserved, especially a segment upstream of the SANT domain (indicated in orange) that is required for transcription from linear, but not supercoiled, templates. The percentages indicate amino acid identities between ScBdp1 and HsBdp1 in the regions bracketed by dotted lines. HsBdp1, HsBdp1_v2, and HsBdp1_v3 are identical in the colored regions. The repeats extend from amino acids 822 to 1338. HsBdp1 and HsBdp1_v2 diverge after amino acid 1353. HsBdp1 and HsBdp1_v3 diverge after amino acid 684.

Human Bdp1 cDNAs were isolated through a combination of database searches for sequences similar to the yeast Bdp1 SANT domain and library screening (Schramm et al. 2000). The structure of the protein encoded by one of these cDNAs (HsBdp1) is shown in Figure 4. It is highly related to the yeast protein within the SANT domain (43% identity) as well as both immediately upstream, in a region that encompasses the segment required for transcription from linear DNA templates (21% identity), and downstream (17% identity). Outside of these regions, the two proteins are not conserved, and the human protein differs from the yeast protein by a striking C-terminal extension containing a number of repeats with potential phosphorylation sites. A number of alternatively spliced BDP1 cDNAs have been isolated (Kelter et al. 2000; Schramm et al. 2000). Two of these encode strikingly different proteins, which are also shown in Figure 4. The longest protein (labeled HsBdp1_v2 in the figure) is identical to Bdp1 except that the last few amino acids are replaced by a 901-amino-acid extension, giving a protein of 2254 amino acids. Another cDNA encodes a 725-amino-acid protein (Bdp1_v3), which contains Bdp1 sequences up to amino acid 684, followed by a divergent 47-amino-acid extension (Kelter et al. 2000).

Which of the alternatively spliced forms of human Bdp1 are involved in RNA polymerase III transcription in vivo is not clear at present. Depletions of extracts with antibodies directed against regions both upstream and downstream of the SANT domain within the N-terminal half of human Bdp1 (Schramm et al. 2000), as well as against the repeat region (L. Schramm and N. Hernandez, unpubl.), debilitate transcription from both type 2 and 3 promoters in vitro, and transcription can be restored by addition of recombinant human Bdp1, either full-length or truncated downstream of the SANT domain. This suggests that HsBdp1 is generally required for RNA polymerase III transcription, and that the C-terminal repeats are not required for basal in vitro transcription from naked DNA templates. However, the functional protein present in HeLa cell extracts probably contains the repeat region, because it can be depleted by antibodies directed against this region. Perhaps the repeat region performs a regulatory role not scored in the in vitro transcription assay.

The characterization of human TFIIIB has revealed that unlike in S. cerevisiae, where type 3 promoters apparently do not exist and one form of TFIIIB serves all RNA polymerase III promoters (Joazeiro et al. 1994), there are at least two forms of TFIIIB in human cells. As shown in Figure 5, one of them consists of HsTBP, HsBrf1, and HsBdp1 and is used by type 2 (and probably type 1) promoters. The other consists of HsTBP, HsBrf2, and HsBdp1, and is used by type 3 promoters. Future work may reveal that different spliced variants of Bdp1 are recruited to different RNA polymerase III promoters in vivo. Furthermore, in D. melanogaster cells, the TBP in TFIIIB is replaced by a TBP-related factor called TRF1 (Takada et al. 2000). Thus, there may be a wide range of TFIIIB activities in different species containing variants of each of the three TFIIIB components.

View larger version (27K): [in this window] [in a new window]   Figure 5.   Promoter-selective TFIIIB activities. The TFIIIB components required by different classes of promoters in Homo sapiens and Drosophila melanogaster are depicted. Strong (red bars) and weak (blue bars) direct protein-protein associations in solution are indicated. Stippled lines indicate that a direct protein-protein contact has not been demonstrated. DmBdp1 has not been characterized, but a candidate gene has been identified (Schramm et al. 2000).

In S. cerevisiae, H. sapiens, and D. melanogaster, Brf1 is tightly associated with TBP or TRF1 in solution, as symbolized by red bars in Figure 5. On the other hand, Bdp1 is weakly associated with the TBP-Brf1 complex in S. cerevisiae, and very weakly, if at all, in human cells (Kassavetis et al. 1991; Wang and Roeder 1995; Mital et al. 1996; Schramm et al. 2000). Indeed, an association between HsBdp1 and HsTBP can only be detected in GST pull-downs (blue bars in Fig. 5; Cabart and Murphy 2002). Similarly, although HsBrf2 can be shown to associate with HsTBP in GST pull-downs (Cabart and Murphy 2001, 2002), it is not strongly associated with TBP in HeLa cell extracts (Schramm et al. 2000). Thus, the TFIIIB components do not always form a stable complex off the DNA.

	Functions of TFIIIB

Top Introduction Structure of RNA polymerase... The assembly pathways directed... RNA polymerase III Composition of TFIIIB Functions of TFIIIB Recruitment factors: TFIIIA Recruitment factors: TFIIIC Assembly of a stable... Recruitment factors: SNAPc Assembly of a stable... Stepwise assembly versus... Human TFIIIC1 Termination and recycling: does... Conclusion References

In RNA polymerase II transcription, the opening of the transcription bubble that occurs after recruitment of the polymerase is dependent on TFIIE and an ATP-dependent helicase activity of TFIIH (Holstege et al. 1996; Tirode et al. 1999). In contrast, in RNA polymerase III transcription, the opening of the transcription bubble occurs in an ATP-independent manner after recruitment of RNA polymerase III by TFIIIB (Kassavetis et al. 1990, 1992). The availability, in yeast, of both recombinant TFIIIB and a transcription system independent of TFIIIC, that is, a system in which a TATA box can recruit TFIIIB directly, has allowed detailed analyses of the functions of the TFIIIB subunits. These studies have given a detailed picture of how TFIIIB recognizes the TATA box and how it recruits RNA polymerase III. They have also revealed that, remarkably, TFIIIB not only functions to recruit RNA polymerase III but also participates in opening of the transcription bubble.

Binding of TFIIIB to the TATA box

In promoters consisting of just a TATA box, S. cerevisiae TFIIIB binds to the DNA through recognition of the TATA box by its TBP subunit. Indeed, a mutation in the TATA box that debilitates RNA polymerase III transcription can be compensated by a mutation in TBP that alters the DNA-binding specificity of the protein and allows it to bind to the mutated TATA box (Strubin and Struhl 1992; Whitehall et al. 1995). The TBP-TATA-box complex is then recognized by Brf1 or, in the case of the human U6 promoter, by Brf2. The similarity of both Brf1 and Brf2 to TFIIB is very striking, and immediately suggests that the conserved domains of these proteins may perform equivalent functions during assembly of RNA polymerase II and III transcription initiation complexes. The reality, however, is more complex. In TFIIB, the core domain is sufficient for association with the TATA-box-TBP complex. However, recruitment of RNA polymerase II and TFIIF to the TATA-box-TBP-TFIIB complex requires the TFIIB zinc-binding domain (Barberis et al. 1993; Ha et al. 1993; Hisatake et al. 1993; Yamashita et al. 1993; Pardee et al. 1998).

HsBrf2 resembles TFIIB in that it recognizes the TATA-box/TBP complex through its TFIIB-related core domain (Cabart and Murphy 2001). In contrast, for S. cerevisiae Brf1, the task of recognizing the TBP-TATA-box complex is performed by two regions of the protein, the TFIIB-related N-terminal half as well as the Brf1-specific C-terminal half, with the latter playing the major role. Thus, as summarized in Figure 6, a truncated ScBrf1 protein retaining just the zinc-binding domain and the core associates only very weakly with a TATA-box-TBP complex. Indeed, the association is so weak that it is only detected by methods such as photochemical cross-linking (Kassavetis et al. 1997, 1998b; Colbert et al. 1998). This weak association appears to involve a TBP surface that overlaps or lies near the TFIIB-interacting surface in the "stirrup" of the second TBP repeat, because a triple-amino-acid change in TBP that disrupts the TFIIB interaction suppresses cross-linking of the N-terminal half of ScBrf1 to DNA and thus probably association with the TBP-TATA-box complex (Kassavetis et al. 1998b). On the other hand, a 110-amino-acid region encompassing conserved region II within the C-terminal half of the protein is sufficient for stable association with a TATA-box-TBP complex as well as for recruitment of ScBdp1. Moreover, the hydroxyl radical footprint observed with just the C-terminal domain of ScBrf1 is identical to that observed with the full-length protein (Colbert et al. 1998). Thus, despite the strong conservation of the core domains in TFIIB and Brf1, it appears that in ScBrf1, the function of recognizing the TBP-TATA-box complex has been largely transferred to the C-terminal half of the protein and in particular to conserved region II. This region of ScBrf1 binds the opposite face of the TBP-TATA-box complex from TFIIB and recognizes a TBP surface that overlaps that recognized by TFIIA (Colbert et al. 1998; Kassavetis et al. 1998b; Shen et al. 1998; for models of the structure of the TBP-DNA-ScBrf1 complex, see Colbert et al. 1998; Geiduschek and Kassavetis 2001). It will be important to contrast HsBrf1 and HsBrf2 with ScBrf1 to determine how these TBP-association activities have been conserved among the human Brf1 and Brf2 proteins.

View larger version (8K): [in this window] [in a new window]   Figure 6.   Functional domains of Saccharomyces cerevisiae Brf1. S. cerevisiae Brf1 is depicted, with the locations of the zinc domain, direct repeats in the core, and conserved regions I, II, and III. The brackets below indicate regions of the proteins sufficient for association with the TBP-TATA-box complex (Kassavetis et al. 1998b), TBP alone (Khoo et al. 1994), and the C34 (Khoo et al. 1994; Andrau et al. 1999) and C17 (Ferri et al. 2000) subunits of RNA polymerase III. The black boxes indicate regions where mutations or deletions have a strong negative effect on the associations. The stippled line indicates an association detected only by UV cross-linking. The upstream boundary of the Brf1 region sufficient for interaction with C17 is not precisely defined.

ScBdp1 can associate with a preassembled TBP-ScBrf1-TATA-box complex but not with a complex lacking ScBrf1, and this confers on the yeast TFIIIB-DNA complex its striking resistance to salt and heparin. ScBdp1 contacts not only ScBrf1 but also TBP, because at least one mutation in yeast TBP prevents association of ScBdp1 without affecting association of ScBrf1 (Colbert et al. 1998). ScBdp1 also contacts DNA because its assembly onto the TATA-box-TBP-ScBrf1 complex both requires DNA, and extends the DNA footprint, upstream of the TATA box (Colbert et al. 1998; Shah et al. 1999). Moreover, ScBdp1 can be cross-linked to the DNA at sites upstream of the TATA box (Shah et al. 1999). The binding of ScBdp1 to the TBP-ScBrf1-TATA-box complex induces a bend in the DNA between the TATA box and the transcription start site, which is in phase with the bend imposed by TBP on the TATA box (Leveillard et al. 1991; Braun et al. 1992b; Grove et al. 1999). This bending of the DNA has been postulated to contribute to the ScBdp1-dependent stabilization of the TFIIIIB-DNA complex by helping impede sliding of the DNA out of the complex (Grove et al. 1999), a hypothesis consistent with thermodynamic and kinetic data indicating ScBdp1-dependent kinetic trapping of the DNA (Cloutier et al. 2001). In the human system, HsBdp1 has been shown to assemble, albeit inefficiently, on a preformed TATA-box-TBP-HsBrf2 complex (Cabart and Murphy 2002).

RNA polymerase recruitment by TFIIIB

Figure 7 shows the known protein-protein contacts between TFIIIB and RNA polymerase III subunits with arrows for contacts identified with human (solid) or yeast (hatched) subunits, respectively. Eight RNA polymerase III subunits can be cross-linked to DNA in a transcription initiation complex (Bartholomew et al. 1993). Of these, C34, which is part of the three-subunit subcomplex that is required for transcription initiation (Werner et al. 1993; Wang and Roeder 1997), maps the furthest upstream and can be localized between positions 17 and +6 relative to the transcription start site, in close proximity to TFIIIB (Bartholomew et al. 1993). ScBrf1 interacts in vivo and in vitro with C34, and human Brf1 associates with the human homolog of C34, HsRPC39, in vitro (Werner et al. 1993; Khoo et al. 1994; Wang and Roeder 1997). As shown in Figure 6, ScBrf1 appears to contact C34 through three regions: regions II and III within the Brf1-specific C-terminal domain (Andrau et al. 1999), and another region, identified by GST pull-down assays, located within the core region in the TFIIB-related N-terminal half of the protein (Khoo et al. 1994). ScBrf1 also contacts the recently identified RNA polymerase III subunit C17 through the C-terminal half of its core region (Ferri et al. 2000). Notably, unlike the zinc-binding domain of TFIIB, the zinc-binding domain of ScBrf1 is not required for RNA polymerase recruitment (Kassavetis et al. 1997; Hahn and Roberts 2000). How HsBrf2 contacts RNA polymerase III is not known. It will be highly interesting to determine further which parts of the protein are required for assembly with TBP and SNAP_c onto the human U6 promoter and for recruitment of RNA polymerase III. Contacts between Bdp1 and RNA polymerase III subunits have not been described, but as shown in Figure 7, human TBP associates with the HsRPC39 RNA polymerase III subunit in vitro (Wang and Roeder 1997).

View larger version (56K): [in this window] [in a new window]   Figure 7.   Protein-protein contacts between TFIIIB components and RNA polymerase III subunits. The solid arrows represent contacts identified with human subunits, the stippled arrows depict contacts identified with Saccharomyces cerevisiae subunits.

Post-RNA polymerase III recruitment roles for Brf1 and Bdp1

In in vitro transcription assays with supercoiled templates, the activities of S. cerevisiae Brf1 and Bdp1 are surprisingly resistant to deletions. The N-terminal half of ScBrf1 forms an unstable TFIIIB complex but is nevertheless capable of directing TFIIIC-independent transcription from a TATA box if high amounts of ScBdp1 are supplied (Kassavetis et al. 1997). The C-terminal half on its own shows little or no transcription activity, but when the N-terminal half of ScBrf1 is added in trans, peptides encompassing region II mediate high levels of transcription. Perhaps most surprising, an ScBrf1 protein lacking the first 164 amino acids including the zinc-binding domain and the first TFIIB-related repeat retains up to 25% of the activity of full-length ScBrf1 for TFIIIC-independent transcription from supercoiled templates in vitro (Kassavetis et al. 1997). Importantly, however, none of these ScBrf1 truncations function in vivo or for TFIIIC-dependent transcription in vitro. Moreover, they do not function for TFIIIC-independent transcription in vitro from a linear template, suggesting that they are somehow defective in promoter opening. Indeed, with the ScBrf1 protein lacking the first 164 amino acids, RNA polymerase III is recruited on a linear template, but the transcription bubble does not form (Kassavetis et al. 1998a). This is probably caused at least in part by the absence of the ScBrf1 zinc ribbon, because point mutations within the zinc domain show defects in promoter opening as determined by sensitivity to potassium permanganate (Hahn and Roberts 2000). Therefore, in ScBrf1, the zinc ribbon, which is not required for polymerase recruitment, plays a role at a later stage, during promoter opening.

The ScBdp1 TFIIIB subunit also plays a post-RNA polymerase III recruitment role. Thus, ScBdp1 molecules lacking the conserved region upstream of the SANT domain (see Fig. 4) can direct somewhat reduced levels of transcription from supercoiled templates but are inactive with linear templates and fail to generate permanganate sensitivity around the transcription start site (Kassavetis et al. 1998a). Moreover, ScBdp1 is dispensable for transcription altogether under conditions in which promoter opening is not required (Kassavetis et al. 1999). Upon recruitment of RNA polymerase III, the SUP4 tRNA gene promoter opens in two segments, one surrounding the transcription start site and the other located ~7 bp upstream (Kassavetis et al. 1992). With templates containing preformed bubbles extending from 9 to 5, TBP and ScBrf1 alone are sufficient to recruit RNA polymerase III and direct multiple rounds of transcription, although the efficiency is only 5% to 10% of that observed with the complete TFIIIB complex. Thus, ScBdp1 plays an essential role in promoter opening.

	Recruitment factors: TFIIIA

Top Introduction Structure of RNA polymerase... The assembly pathways directed... RNA polymerase III Composition of TFIIIB Functions of TFIIIB Recruitment factors: TFIIIA Recruitment factors: TFIIIC Assembly of a stable... Recruitment factors: SNAPc Assembly of a stable... Stepwise assembly versus... Human TFIIIC1 Termination and recycling: does... Conclusion References

In an in vitro system in which TFIIIB can be recruited directly to a TATA box, TFIIIB on its own is sufficient to recruit RNA polymerase III. In natural RNA polymerase III promoters, however, TFIIIB is recruited to the DNA in large part through protein-protein contacts with promoter-bound recruitment factors, specifically TFIIIC or SNAP_c. The type 1 5S promoters and type 2 promoters both use TFIIIC, but on the 5S promoters TFIIIC is recruited through the specificity factor TFIIIA. TFIIIA is the founding member of the C₂H₂ zinc finger family of DNA-binding proteins (Miller et al. 1985) and contains nine C₂H₂ zinc fingers. In S. cerevisiae, the only essential role of TFIIIA is in the transcription of the 5S RNA genes, because strains engineered to express the 5S rRNA from a tRNA-type promoter and lacking TFIIIA are viable (Camier et al. 1995). This may explain in part the rapid evolution of TFIIIA: TFIIIA sequences from various organisms are poorly conserved, even among vertebrates. As an example, human and X. laevis TFIIIAs share 61% identity over a 264-amino-acid regionof 423 and 344 amino acids for the human (Arakawa et al. 1995) and X. laevis (Ginsberg et al. 1984) proteins, respectivelywhereas the RNA polymerase II transcription factor TFIIB is 94% identical in the two species over its entire length. TFIIIA binds directly to the ICR of type 1 promoters. TFIIIA also binds to 5S RNA to form the 7S storage ribonucleoprotein particle (Pelham and Brown 1980). It is present in massive amounts in immature X. laevis oocytes, because they accumulate 5S RNA for later use during oogenesis and the first rounds of embryonic cell division, which occur at a rapid pace in the absence of transcription. This allowed early on the purification of TFIIIA to near homogeneity; indeed, X. laevis TFIIIA was the first eukaryotic transcription factor to be purified (Engelke et al. 1980) and the first for which a corresponding cDNA was isolated (Ginsberg et al. 1984).

Upon binding of X. laevis TFIIIA to the 5S gene, the TFIIIA zinc fingers are aligned over the length of the ICR with the C-terminal zinc finger in proximity of the 5' end, and the N-terminal finger in proximity of the 3' end, of the ICR (for references, see Paule and White 2000). Zinc fingers 1-3, which contact the C box, have been reported to contribute most of the binding energy of the entire protein (Clemens et al. 1992; Liao et al. 1992). Interestingly, however, like TFIIIA fragments containing fingers 1-3, fragments containing fingers 4-9 bind, in this case to the A box and intermediate element, with affinities approaching that of the full-length protein (Liao et al. 1992; Kehres et al. 1997). This observation, as well as the binding behavior of full-length proteins with zinc fingers mutated either singly or in pairs, suggest that simultaneous binding by all nine TFIIIA zinc fingers to DNA requires energetically unfavorable distortions, either in the DNA, the protein, or both. Thus, there is negative cooperativity between certain zinc fingers such that loss of binding by a subset of zinc fingers has only a small negative effect on the overall stability of the complex (Kehres et al. 1997). Although TFIIIA on its own is displaced from DNA upon passage of RNA polymerase III, the unusual TFIIIA binding properties may contribute to the resilience of the complete 5S transcription complex to repeated passage of the RNA polymerase (Bogenhagen et al. 1982; Setzer and Brown 1985; Wolffe et al. 1986; Darby et al. 1988; Kehres et al. 1997).

Surprisingly little is known about how TFIIIA recruits TFIIIC to the DNA. X. laevis TFIIIA contains a 14-amino-acid domain located C-terminal of the ninth zinc finger, and thus located toward the 5' end of the ICR in the TFIIIA/5S gene complex, that is dispensable for DNA binding but essential for transcription (Mao and Darby 1993). In S. cerevisiae TFIIIA, a hydrophobic segment within an 84-amino-acid region located between zinc fingers 8 and 9 is similarly required for cell viability and transcription but not for DNA binding (Rowland and Segall 1998). These protein domains may play a role in the recruitment of TFIIIC.

	Recruitment factors: TFIIIC

Top