Initiation of Eukaryotic Messenger RNA Synthesis

The expression of a large fraction of all protein-coding genes is controlled at the transcriptional level through mechanisms involving the regulation of initiation. In eukaryotic cells, initia- tion of mRNA synthesis by RNA polymerase I1 is governed by DNA sequence elements comprising several functional classes. These include a core promoter element, which contains the binding site for RNA polymerase I1 and controls the location of the start site of transcription, and upstream promoter elements and enhancers, which regulate the rate at which RNA polymerase I1 initiates new rounds of transcription from the core promoter. These sequence elements direct the action of two classes of transcription factors: initiation factors, which are essential for initiation and which are sufficient to direct a basal level of transcription from the core promoter, and regulatory factors, which are not required for initiation but which mediate the action of upstream promoter elements and enhancers (1-3). Analysis of a large number of eukaryotic genes has revealed structural diversity among core promoters. Whereas many are composed of a TATA box, located a short distance upstream of the cap site, and an initiator element, which encompasses the cap site, others, such as the terminal deoxynucleotidyl transferase, dihydrofolate reductase (DHFR),’ and mouse ribosomal protein rpL30 promoters, lack discernible TATA boxes To date, biochemical studies of initiation from the core region of TATA box containing promoters have progressed most rapidly. Initia- tion factors required for selective transcription of these promoters are being purified and characterized, and a detailed picture of the roles these factors play in assembly of an active initiation complex is emerging. This minireview will summarize recent advances in our understanding of the mechanism by which RNA polymerase I1 locates and binds selectively to the core region of TATA box containing


Initiation of Eukaryotic Messenger RNA Synthesis
The expression of a large fraction of all protein-coding genes is controlled at the transcriptional level through mechanisms involving the regulation of initiation. In eukaryotic cells, initiation of mRNA synthesis by RNA polymerase I1 is governed by DNA sequence elements comprising several functional classes. These include a core promoter element, which contains the binding site for RNA polymerase I1 and controls the location of the start site of transcription, and upstream promoter elements and enhancers, which regulate the rate at which RNA polymerase I1 initiates new rounds of transcription from the core promoter. These sequence elements direct the action of two classes of transcription factors: initiation factors, which are essential for initiation and which are sufficient to direct a basal level of transcription from the core promoter, and regulatory factors, which are not required for initiation but which mediate the action of upstream promoter elements and enhancers (1-3).
Analysis of a large number of eukaryotic genes has revealed structural diversity among core promoters. Whereas many are composed of a TATA box, located a short distance upstream of the cap site, and an initiator element, which encompasses the cap site, others, such as the terminal deoxynucleotidyl transferase, dihydrofolate reductase (DHFR),' and mouse ribosomal protein rpL30 promoters, lack discernible TATA boxes (4-6). To date, biochemical studies of initiation from the core region of TATA box containing promoters have progressed most rapidly. Initiation factors required for selective transcription of these promoters are being purified and characterized, and a detailed picture of the roles these factors play in assembly of an active initiation complex is emerging. This minireview will summarize recent advances in our understanding of the mechanism by which RNA polymerase I1 locates and binds selectively to the core region of TATA box containing promoters to form a complex capable of initiating RNA synthesis accurately at the cap site.

RNA Polymerase 11
Studies of initiation by RNA polymerase I1 have lagged behind similar studies of bacterial polymerases. The low abundance of RNA polymerase I1 in most eukaryotic cells has made both its purification and its characterization difficult. Compared with bacterial RNA polymerases, which can be purified relatively easily in yields as high as 400 mg/kg of cell paste (7), RNA polymerase I1 is usually obtained in milligram or sub-milligram quantities from equivalent amounts of starting material (8). Until the late 1970s, the lack of DNA templates containing well defined eukaryotic promoters made biochemical studies of gene transcription by RNA polymerase I1 virtually impossible. Studies of the mechanism of selective transcription by bacterial RNA polymerases, on the other hand, were aided greatly by the availability of bacteriophage templates, such as T7 and X, which contained genetically defined promoters (9). These limitations have largely been overcome by the isolation and characterization of many eukaryotic promoters and by the development of improved meth- ods for purifying RNA polymerase 11.
Over the past several years, genes encoding many of the subunits of RNA polymerase I1 have been cloned (for reviews, see Refs. 1, 8, 10). In addition, a picture of the overall architecture of the enzyme is emerging from electron crystallographic studies. By analyzing two-dimensional crystals of yeast RNA polymerase 11, Darst et al. 111) have recently determined the structure of the enzyme to 16 A (Fig. 1).
Despite these advances, however, the most severe impediment to inquiries into the molecular mechanism of initiation has arisen from the nature of the polymerase itselE purified RNA polymerase I1 is incapable of binding selectively to its promoter and initiating transcription without the assistance of a set of initiation factors. Unlike bacterial RNA polymerases, which, in most cases, are purified as holoenzymes that include one or more tightly associated initiation factors, RNA polymerase I1 is separated from its initiation factors by even the gentlest purification procedures.

Multiple Initiation Factors Direct Assembly of an Active Preinitiation Complex
The finding that RNA polymerase I1 lacks the ability to initiate transcription accurately from promoters prompted a search for factors that mediate this process. Fractionation of transcriptionally active extracts from cultured cells revealed that selective transcription from such TATA box-containing promoters as the AdML, conalbumin, and 8-globin promoters requires the action of multiple initiation factors (12)(13)(14). Purification of these factors, however, proved to be a formidable undertaking because of their low abundance in most eukaryotic cells. In recent years, the use of more plentiful sources, such as rat liver (15) and yeast (16,17), have combined with improved cell culture and protein purification techniques to spur rapid progress in elucidating the structures and functions of the initiation factors. Initiation by RNA polymerase I1 from the core region of TATA box-containing promoters is a multistage process requiring the action of at least five initiation factors and an ATP cofactor. In the first stage, which we designate site selection, RNA polymerase I1 locates and binds selectively to the core promoter. Site selection is a compound process requiring assembly of a nucleoprotein recognition site for polymerase at the core promoter and binding of polymerase at this site. In this stage, an initiation factor, referred to as the TATA factor, first binds stably to the core promoter to form an Initial Complex. This complex serves as at least part of the recognition site for polymerase. Following site selection, additional initiation factor(s) promote formation of the complete, but inactive, preinitiation complex. Finally, in a reversible, ATP-dependent step, this intermediate is converted to an activated complex capable of initiating transcription rapidly and accurately from the core promoter.

Site Selection
Initial Complex Formation-The first committed step in assembly of the preinitiation complex is binding of the TATA factor to the core promoter to form the Initial Complex (Fig. 2). TATA factors have been identified in cell extracts from a wide variety of species including yeast, Drosophila, rat, and man. Although purification of a TATA factor from higher eukaryotes has not yet been reported, a yeast TATA factor, designated yTFIID or BTFlY, has been purified, and its gene has been cloned (18-25). Both native and recombinant yeast TATA factors substitute for partially purified preparations of native TATA factors from higher eukaryotes in reconstituted transcription reactions in vitro ( While studies performed with recombinant TATA factors have provided considerable insight into the mechanism of transcription initiation by RNA polymerase 11, it is doubtful that these small proteins will provide ideal models of native TATA factors from higher eukaryotes. A growing body of evidence suggests that the p-y (RAP30/74 (TFIIF)) and are likely to play an important role in site selection.
Several lines of evidence support the model that P-y (RAP30/ 74 (TFIIF)) promotes site selection through a direct interaction with RNA polymerase I1 by decreasing the affinity of the enzyme for free DNA and, in concert with a (TFIIB), increasing the affinity of the enzyme for the Initial Complex. First, binding of p-y (RAP30/74 (TFIIF)) from HeLa cells to RNA polymerase I1 has been well characterized, This factor was initially purified by its ability to bind to an RNA polymerase 11-Sepharose affinity column and was subsequently shown to co-sediment with polymerase in sucrose gradients (51). Second, rat liver P-y (RAP30/ 74(TFIIF)) markedly inhibits nonselective binding of RNA polymerase I1 to free DNA, much as Escherichia coli u7" inhibits nonspecific binding of RNA polymerase to nonpromoter sites in DNA (52). In addition, in the absence of LY (TFIIB), P-y (RAP30/ 74 (TFIIF)) inhibits nonspecific binding of RNA polymerase I1 to templates containing preassembled Initial Complexes (32

Assembly and Activation of the Complete Preinitiation Complex
The Complete Preinitiation Complex-Once RNA polymerase I1 has bound selectively to the Initial Complex, additional initiation factors promote formation of the fully assembled, or complete, preinitiation complex. Two factors from rat liver, designated t and 6, act at this stage. Competition binding experiments suggest that both factors ultimately become integral components of the complete preinitiation complex (49).' t and 6 appear to enter the preinitiation complex by different routes: c only after RNA polymerase I1 has bound to the Initial Complex and 6 either before or after RNA polymerase I1 has bound (48).' Restriction site protection experiments indicate that 6 and 6 promote formation of stable protein-DNA contacts that anchor the transcription apparatus to promoter sequences near the cap site (48). These protein-DNA contacts may result from direct interactions of t, 6, or both factors with promoter DNA; alternatively, c or 6 may stabilize interactions of RNA polymerase I1 and the other initiation factors with the promoter. t is a heterodimer of 34-and 58-kDa polypeptides; both polypeptides are essential for transcription activity (33). A similar factor, designated TFIIE, has been purified from HeLa cells and, like t, appears to function late in assembly of the preinitiation complex. TFIIE is composed of 34-and 57-kDa polypeptides; unlike c, however, the large subunit of TFIIE exhibits significant transcription activity in the absence of the small subunit (56,57). A chromatographic fraction containing both TFIIF and TFIIE promotes formation of stable protein-DNA contacts near the cap site of the AdML promoter during the final step in preinitiation complex formation (40). TFIIE appears to interact with RNA polymerase I1 in solution (58), but the functional relevance of this interaction has not yet been demonstrated. 6 has been purified to near-homogeneity from rat liver nuclear extracts (59).' The purified factor exhibits a native molecular mass of approximately 390 kDa and has an associated DNAdependent ATPase activity. ATPase and transcription activities co-purify with a set of eight polypeptides ranging in size from approximately 90 to 35 kDa. Reconstitution of 6 from isolated ' R. C. Conaway and J. W. Conaway, unpublished results polypeptides has not yet been achieved; thus, it remains to be determined whether each of these polypeptides is essential for transcription or whether some are derived from others by proteolysis. 6 has no obvious counterparts in other transcription systems.
Formation of the Activated Preinitiation Complex-Transcription initiation from the core promoter has a strict requirement for an adenine nucleoside triphosphate cofactor (60-62). Several lines of evidence argue that this cofactor is required for conversion of the complete preinitiation complex to an active configuration. First, assembly of the complete preinitiation complex does not require ATP. Second, ATP is not required after synthesis of the first nine nucleotides of transcripts initiated at the AdML promoter (61). Finally, work in our laboratory (62) provided evidence that ATP functions immediately prior to RNA synthesis. We observed that inhibition of transcription by ATP-yS, a potent inhibitor of the ATP-dependent step, could be prevented by brief incubation of complete preinitiation complexes with ATP or ATP analogs prior to addition of ATP-yS and the ribonucleoside triphosphates needed for RNA synthesis. Further experiments indicated that the ATP-dependent step is rapidly reversible, suggesting that, in the presence of ATP, active and inactive forms of the preinitiation complex are in dynamic equilibrium.
It is not yet known how ATP activates the preinitiation complex or which components of the transcription system mediate activation. ATP may, for example, be hydrolyzed to provide energy for some crucial step in initiation or to serve as a phosphate donor in a phosphorylation reaction. On the other hand, ATP may simply bind one or more components of the transcription system to promote transition of the complete preinitiation complex to a transcriptionally active conformation.
A role for HeLa cell transcription factor RAP30/74 (TFIIF) in activation of the preinitiation complex has been proposed. Sopta and co-workers (63) reported that RAP30/74 (TFIIF) possesses an associated ATP(dATP)-dependent DNA helicase activity and suggested that this helicase might unwind the promoter to facilitate formation of an active preinitiation complex analogous to the "open" complex formed by E. coli RNA polymerase (86). Others (47), however, have reported that highly purified preparations of this factor lack detectable helicase activity.
The possibility that a protein kinase is involved in activation of the preinitiation complex has been investigated in light of evidence suggesting that phosphorylation of the carboxyl-terminal domain (CTD) of the largest subunit of RNA polymerase I1 plays a role in transcription. The CTD is rich in serine and threonine residues, which are extensively phosphorylated in vivo (64,65). UV cross-linking studies indicate that, in isolated nuclei as well as in crude transcription extracts, the CTDs of polymerase molecules actively engaged in transcription are highly phosphorylated (66, 67). Recently, several groups have obtained evidence that RNA polymerase 11, bound in preinitiation complexes at the AdML promoter, can be phosphorylated in the CTD by a template-associated protein kinase, immediately before or during initiation (68,69). The template-associated protein kinase(s) has not been isolated, and a requirement for it in transcription has not yet been demonstrated directly. Moreover, interpretation of these results is complicated by the observation that RNA polymerase I1 lacking all or most of the CTD is capable of catalyzing accurate initiation from the AdML promoter in vitro (70, 71). In addition, it has been observed that, in the absence of ATP, GTP can serve as a substrate for a template-associated kinase, raising the possibility that phosphorylation of the CTD is not required for activation of the preinitiation complex but, instead, plays some role in modulating the activity of RNA polymerase I1 (68).
Finally, several lines of evidence suggest that, in the rat liver system, transcription factor 6 could mediate activation of the preinitiation complex (59). Highly purified 6 has a closely associated DNA-dependent ATPase activity, specific for ATP or dATP. The associated ATPase of 6 is stimulated more strongly by DNA fragments containing the AdML or mouse interleukin-3 core promoters than by fragments containing non-promoter sequences, consistent with the notion that 6 may function through a direct interaction with promoter sequences. Of the rat liver factors required for assembly of the complete preinitiation corn-26. Muhich, M. L., Iida, C. T., Horikoshi, M., Roeder, R. G., and Parker, C. s.