TATA protein tells RNA polymerase where to get started on a gene Show
The enzyme RNA polymerase performs the delicate task of unwinding the two strands of DNA and transcribing the genetic information into a strand of RNA. But how does it know where to start? Our cells contain 30,000 genes encoded in billions of nucleotides. For each gene, the cell must be able to start transcription at the right place and at the right time. Specialized DNA sequences next to genes, called promoters, define the proper start site and direction for transcription. Promoters vary in sequence and location from organism to organism. In bacteria, typical promoters contain two regions that interact with the sigma subunit of their RNA polymerase. The sigma subunit binds to these DNA sequences, assists the start of transcription, and then detaches from the polymerase as it continues transcription through the gene. Our cells have a far more complex promoter system, using dozens of different proteins to ensure that the proper RNA polymerase is targeted to each gene. The TATA-binding protein is the central element of this system. Our protein-coding genes have a characteristic sequence of nucleotides, termed the TATA box, in front of the start site of transcription. The typical sequence is something like T-A-T-A-a/t-A-a/t, where a/t refers to positions that can be either A or T. Surprisingly many variations on this theme also work, and one of the challenges in the study of transcription is discovering why some sequences work and others don't. The TATA-binding protein (sometimes referred to as TBP) recognizes this TATA sequence and binds to it, creating a landmark that marks the start site of transcription. When the first structures of TATA-binding protein were determined, researchers discovered that TATA-binding protein is not gentle when it binds to DNA. Instead, it grabs the TATA sequence and bends it sharply, as seen in PDB entries 1ytb , 1tgh and 1cdw .
TATA-binding protein works as part of a larger transcription factor, TFIID, that starts the process of transcription. After it binds to the promoter, it recruits additional transcription factors. TFIIB, shown at the top here from PDB entry 1vol , binds next. Then a string of other transcription factors bind, constructing a large protein complex that decides whether or not to start transcription. These may include transcription activators, such as TFIIA shown in the middle from PDB entry 1ytf , that promote the start of transcription. Other factors inhibit the start of transcription, such as the transcription regulator NC2 (negative cofactor 2), shown at the bottom from PDB entry 1jfi . In all of these pictures, TATA-binding protein is shown in blue, a small piece of DNA is shown in red and the transcription factor is shown in green. TATA-binding protein uses two types of interactions to recognize and hold the TATA sequence, as seen in this structure from PDB entry 1ytb . First, as shown at the top, it has a string of lysine and arginine amino acids (colored dark blue) that interact with the phosphate groups of the DNA (colored bright yellow and red). This glues the protein to the DNA. Second, the protein uses specially-placed amino acids to interact with DNA bases. As shown in the lower picture, four phenylalanine amino acids jam into the DNA minor groove and form the kinks that bend the DNA. There are also two symmetrical asparagine amino acids that form hydrogen bonds at the very center. The combination of the unusual flexibility of TATA DNA sequences and these specific hydrogen bonds allows TATA-binding protein to recognize the proper sequence. As you are looking at these structures yourself, notice that TATA-binding protein, even though it is composed of a single protein chain, is composed of two symmetrical halves. This symmetry is easily seen in the two pairs of phenylalanines and the two asparagines shown in the lower figure. It is thought that an ancient gene duplication created this protein by combining two copies of the same gene. For more information on TATA-binding protein from a genomics perspective, visit the Protein of the Month at the European Bioinformatics Institute. These pictures were created with RasMol. You can create similar pictures by clicking on the accession codes here and picking one of the options for 3D viewing. The phenylalanines shown above are numbers 99, 116, 190, and 207, and the asparagines are numbers 69 and 159.
Related PDB-101 ResourcesReferences
July 2005, David Goodsell
The eukaryotic RNA polymerase (RNAP) II and RNAP III transcriptional machineries are evolutionarily related. Indeed, some of the polypeptides utilized by the two RNA polymerases are identical. For example, the TATA box-binding protein (TBP) is required for both RNAP II and RNAP III transcription as a subunit of the polymerase-specific factors TFIID and TFIIIB respectively [reviewed in (1,2)]. Moreover, at least five polypeptide subunits are shared in common by the RNAP II and RNAP III enzymes of yeast (3,4). A number of homologous yet non-identical polypeptides are also characteristic of the two systems. Examples include the RNAP II-specific factor TFIIB and the homologous RNAP III-specific factor BRF (5–8), as well as a number of evolutionarily related integral subunits of the RNA polymerases themselves (3). Although the TATA box was originally identified as a promoter element upstream of many mRNA genes, it is also capable of directing transcription by RNAP III (9–16). Interestingly, work from our lab recently demonstrated that, in the absence of other cis-acting elements, the RNA polymerase specificity of a promoter could be determined by the orientation of a certain TATA box sequence (17). In an unfractionated Drosophila nuclear extract, the canonical TATA sequence TATAAAAA specifically promoted RNAP II transcription, whereas the same sequence when flipped into the reverse orientation (TTTTTATA) specifically promoted RNAP III transcription. Moreover, in the same reaction mixture, RNAP II initiated transcription ∼22–23 base pairs (bp) downstream of the forward TATA box, whereas RNAP III initiated a similar distance from the TATA box but in the opposite ‘upstream’ direction (17). Those results indicated that the RNAP II and RNAP III initiation complexes preferentially assembled in opposite orientations on the chosen TATA box sequence. The sequence TATAAAAA was selected for the above-described studies based upon its high degree of sequence asymmetry and upon a previous observation that it acted as a strong RNAP II promoter in the Drosophila in vitro transcription system (18). We now report related studies using a number of different TATA box sequence variants. The data have enabled us to rank the TATA boxes in terms of their abilities to promote RNAP III transcription in the unfractionated Drosophila nuclear extract. The results of these studies indicate that the original sequence TTTTTATA may represent an optimal RNAP III TATA box and that Ts at positions 2 and 4 are important determinants of RNAP III specificity in this system. Materials and MethodsPlasmid templatesPlasmid templates H/D ‘forward’ TATA and H/D ‘reverse’ TATA have been previously described (17). Each contains the synthetic sequence TATAAAAA inserted between the KpnI and BamHI sites of pUC18, but oriented in opposite directions. All the other H/D templates were constructed by replacing the TATA sequence between the KpnI and BamHI sites of the H/D ‘forward’ TATA plasmid with other synthetic TATA variants. This generated a family of plasmids that were identical to each other except for the 8 bp TATA box sequence. The wild-type Drosophila U1 gene template with 391 bp of 5′-flanking DNA has been previously described (18). The wild-type U6 gene template was prepared starting from a plasmid obtained from Ram Reddy (Baylor College of Medicine) that contained the wild-type Drosophila U6-2 gene (19). An EcoRI-HaeIII restriction fragment was isolated that contained 59 bp of U6-2 gene coding sequence and 409 bp of 5′-flanking DNA. This fragment was then inserted between the EcoRI and HincII sites of pUC18 to use as a template for in vitro transcription assays. All plasmids were grown in E.coli TOP10 cells (Invitrogen), purified using Qiagen Plasmid Maxi kits, and sequenced to confirm the identity of the promoter region. In vitro transcriptionTranscription reactions, purification of transcription products, and analysis by primer extension were as described previously (17). The 32P-labeled primers were complementary to pUC18 DNA on either side of the polylinker (New England Biolabs no. 1233 or a primer [1211z] similar to New England Biolabs 1211, but extended by 7 nucleotides at its 5′-end). Primer extension products were separated in 10% denaturing polyacrylamide gels that were dried prior to autoradiography and/or PhosphorImager quantitation. Each template was transcribed in a minimum of five different experiments. ResultsStudies on ‘canonical’ TATA boxes: effect of sequence and orientation on promoter strength and RNA polymerase specificityIn an initial experiment, several canonical variants of the TATA sequence and their reverse-orientation counterparts were tested for their ability to promote transcription by either RNAP II and/or RNAP III. These synthetic 8 bp TATA sequences were cloned within the pUC18 vector polylinker region (see top of Fig. 1). Toward the right (or ‘downstream’) of the TATA element each template contained synthetic sequence that corresponds to a combination of sequences surrounding the Drosophila U1 and U6 gene transcription start sites. Toward the left of the TATA box (in the ‘upstream’ direction), the templates contained sequence resembling a combination of the human U1 and U6 gene transcription start sites. Knowing that U1 genes are transcribed by RNAP II and U6 genes by RNAP III, we chose this combination of U1 and U6 sequences with the goal of optimizing the context for initiation of transcription for each of the RNA polymerases. A second important consideration in choosing these sequences was the knowledge that both the U1 and U6 genes have external promoters that reside entirely upstream of position −20; thus there are no known promoter elements within the synthetic U1/U6-like sequences that were employed to optimize the transcription start sites. The DNA templates were transcribed in vitro using a soluble nuclear fraction (SNF) prepared from Drosophila embryos (20,21). The drug α-amanitin was used as a specific inhibitor of RNAP II. Because insect RNAP III is resistant to inhibition by even a high concentration of α-amanitin (22,23), tagetitoxin was employed as a specific RNAP III inhibitor (24). Transcription products were assayed by primer extension analysis. Figure 1, lanes 1–32, shows typical results using an oligonucleotide primer (1211z) that detects transcription initiating in the ‘Drosophila’ sequences and proceeding in the rightward direction. The sequence TATAAAAA and the cognate reverse TATA sequence, TTTTTATA, produced results identical to previously reported findings (17). Both TATA sequences supported transcription in the Drosophila SNF with similar efficiencies (Fig. 1, compare lane 1 with 18, and lane 19 with 32). However, the canonical forward sequence TATAAAAA promoted RNAP II-specific transcription (inhibited by α-amanitin but not by tagetitoxin, lanes 2 and 3) whereas the cognate reverse TATA sequence TTTTTATA specifically promoted transcription by RNAP III (not inhibited by α-amanitin but inhibited by tagetitoxin, lanes 20 and 21). Transcription was TATA box-dependent, since no products were detectable when a sequence with three G/C substitutions was inserted in either orientation (lanes 17 and 31). As expected, two other canonical forward TATA sequences, TATATAAA and TATAAATA, overwhelmingly supported RNAP II transcription (Fig. 1, lanes 5–12). However, when these two TATA sequences were placed in the reverse orientation, a mixture of both RNAP II and RNAP III transcription was observed, and the overall level of transcription was somewhat decreased (Fig. 1, lanes 23–30). Transcription promoted by the sequence TTTATATA was more thoroughly inhibited by tagetitoxin than by α-amanitin, indicating that 60–70% of the transcription was due to RNAP III (lanes 23–26). The second sequence, TATTTATA, promoted RNAP II and RNAP III transcription with approximately equal efficiencies (lanes 27–30). Somewhat surprisingly, the completely symmetrical TATA box, TATATATA, very efficiently promoted transcription by RNAP II but worked poorly for RNAP III transcription (Fig. 1, lanes 13–16). Indeed, by PhosphorImager analysis, the RNAP II signal was >15-fold greater than the RNAP III signal. Together, these results suggest that increasing the number of alternating TpA residues disfavors transcription by RNAP III relative to RNAP II. Next, transcription initiating in the upstream direction (in the synthetic ‘human’ DNA) was measured for each of these constructs by using the 1233 oligonucleotide for the primer extension assays (Fig. 1, lanes 33–64). Since transcription was assayed in the opposite direction, the TATA boxes can be considered to be reversed in orientation relative to the experiments shown in lanes 1–32. In accord with previous results (17), transcription initiating in the reverse direction from TATAAAAA was predominantly due to RNAP III (lanes 33–36). However, the data shown in lanes 37–44 was somewhat unexpected because transcription promoted by these TATA boxes in the leftward direction was almost entirely due to RNAP II. These results should be compared with those shown in lanes 23–30, in which TATA boxes with the same sequences (relative to the direction of transcription) promoted a mixture of RNAP II and RNAP III transcription. Thus, there appears to be an intrinsic bias in favor of RNAP II over RNAP III transcription in the leftward direction relative to transcription in the rightward direction, but the underlying cause of this bias has not been investigated. Apparently a strong RNAP III TATA box (TTTTTATA) can overcome this bias (lanes 33–36), whereas TATA boxes that are less selective are unable to do so (lanes 37–44). The symmetrical TATA sequence TATATATA acted as a very strong RNAP II-specific promoter in the leftward direction (lanes 45–48 in comparison to lanes 13–16), lending further support to the concept that there is a bias favoring RNAP II transcription in the leftward direction. As expected, the TATA boxes in the canonical forward orientation (relative to the leftward direction of transcription being assayed) were highly specific for RNAP II (lanes 51–62). Primer extension analysis of in vitro transcription products synthesized from H/D templates that contain various TATA box sequences in ‘forward’ and ‘reverse’ orientations. Transcription was carried out using an unfractionated Drosophila nuclear extract. Templates were constructed containing the DNA sequences shown at the top of the figure inserted between the EcoRI and XbaI sites of pUC18. Italics represent plasmid-derived DNA. The bold Xs indicate the position where the variant TATA box sequences were positioned. The actual TATA sequences utilized in each construction are shown above the corresponding lanes of the autoradiograms. The arrows above or below the TATA box sequences represent the direction of transcription being assayed. Plus signs above the individual lanes indicate inclusion of the RNAP II-specific inhibitor α-amanitin (2 µg/ml) or the RNAP III-specific inhibitor tagetitoxin (0.4 U/ml, Epicentre Technologies). Transcription start sites (indicated by arrows above or below the sequence at the top of the figure) were mapped by running the primer extension products alongside sequencing ladders (data not shown). Bands on the autoradiograms are indicated that correspond to reverse transcription products of 76–77 (1211z primer) or 71–72 nucleotides (1233 primer). The position of a 54mer oligonucleotide that was added to each reaction as a recovery standard is also shown. Each panel represents a different set of transcriptions, but reactions with the original forward TATA (TATAAAAA) and reverse TATA (TTTTTATA) templates were included in each panel to permit a comparison of intensities from set to set. Effect of an A to G mutation at position 2 of the TATA boxThe substitution of a G residue for the canonical A residue at position 2 of the TATA box has been shown to be highly deleterious for TATA-mediated RNAP II transcription (25) and more recently for TATA-mediated RNAP III transcription (11,26). Constructs containing this mutation (TGTAAAAA) were therefore analyzed for their ability to support transcription by either RNAP II or RNAP III (Fig. 2). A comparison of lanes 1 and 2 (Fig. 2) shows that the G substitution abolished nearly all transcription (due to RNAP II) in the forward (rightward) direction. Interestingly, the mutant TATA box, when reversed in orientation, was able to support RNAP III-specific transcription in the rightward direction (lanes 3–6), although at an ∼5-fold reduced level compared to the non-mutant reverse TATA box (lane 7). When transcription was assayed in the leftward direction (lanes 8–14), similar results were obtained. Whereas RNAP II transcription was reduced over 30-fold by the base substitution (compare lanes 8 and 9), RNAP III transcription promoted by the mutated reverse TATA box was reduced only 5-fold relative to the original reverse TATA sequence (lanes 10–14). Primer extension analysis of in vitro transcription products from H/D ‘forward’ and ‘reverse’ TGTAAAAA templates. Conditions and explanation of symbols are the same as described in the legend to Figure 1. Determinants of RNAP III specificityOne interpretation of the above experiments is that a run of five Ts on the non-template strand in the upstream portion of the TATA box (e.g. TTTTTATA or TTTTTACA) represents a sufficient determinant for RNAP III transcription. Therefore, in the next series of experiments (Fig. 3), we tested the ability of a number of variant TATA sequences to promote transcription by RNAP III. To avoid any variation that might arise from the contribution of the flanking sequences, these experiments were done only with the 1211z primer to assay transcription only in the rightward direction. A run of eight consecutive T residues (Fig. 3, lanes 1–4) supported RNAP III transcription but with only ∼10% of the efficiency of the sequence TTTTTATA (compare with lane 18). The sequences TTTTTAAA, TTTTAAAA and TTTAAAAA also exhibited a high degree of selectivity in favor of RNAP III versus RNAP II transcription (Fig. 3, lanes 5–16). The sequence TTTAAATA preferentially supported RNAP III transcription, but in this case a significant RNAP II component (∼30% of the total transcription signal) was clearly detectable (Fig. 3, lanes 19–22). The sequence CTTTTATA, which is the reverse orientation of the adenovirus major late canonical TATA box, was a reasonably strong and highly selective promoter for RNAP III (Fig. 3, lanes 23–26). Two variants with changes at the second position (TCTTTATA and TGTTTATA) supported very low levels of transcription, although the low level that remained was RNAP III-specific. (In this particular experiment, a shorter product apparently resulting from downstream initiation was detected from these two templates, but this shorter transcript was not consistently observed in other experiments.) Primer extension analysis of in vitro transcription products synthesized from H/D templates containing a variety of TATA sequences. All reactions in this figure utilized the 1211z oligonucleotide for primer extension. Other conditions and explanation of symbols are the same as described in the legend to Figure 1. Relative strength of various TATA box sequences as promoters for RNAP III transcriptionIn order to rank the various TATA box variants according to their ability to promote RNAP III transcription in vitro, 14 of the constructs were transcribed in the presence of α-amanitin and the products were run together in the same gel (Fig. 4A). Band intensities from five experiments similar to that shown in Figure 4A were quantified by PhosphorImager analysis, and the levels of transcription were determined relative to that obtained with the sequence TTTTTATA. In Figure 4B, the TATA sequences are ordered according to their relative strengths as promoters for RNAP III transcription. From these data, it is evident that the sequence TTTTTATA represents the optimal RNAP III TATA sequence among those tested. Moreover, there is a wide range of RNAP III transcription efficiencies that is dependent upon the specific TATA box sequence. These results will be further addressed in the Discussion section. The synthetic templates are transcribed in vitro with efficiencies comparable to native U1 and U6 promotersA final experiment was performed to assess the physiological relevance of the in vitro transcriptional levels obtained from the artificial templates relative to the transcriptional activities of natural promoters recognized by RNAP II and RNAP III. Figure 5 compares the activity of the H/D ‘forward’ TATA template (lanes 1–4) with the activity of the native Drosophila U1 gene promoter (lanes 5–8). The synthetic template was transcribed ∼3-fold more efficiently in vitro by RNAP II than the template that contained the wild-type U1 gene promoter. (A) Autoradiogram showing the relative efficiencies of RNAP III transcription promoted by various TATA box sequences. All transcription reactions were performed in the presence of 2 µg/ml α-amanitin. The TATA sequences driving transcription in each lane are indicated in the first and second columns of part (B). (B) Results from several experiments similar to the one shown in (A) were quantified by PhosphorImager analysis. The TATA sequences are arranged in descending order of RNAP III transcription efficiency relative to the sequence TTTTTATA. Errors shown are the standard deviation of the mean. Figure 5 also presents data for RNAP III-specific transcription. The relative transcription levels obtained from the H/D ‘reverse’ TATA template (lanes 9–12) were compared with levels obtained from the wild-type U6 gene promoter (lanes 13–16). The U6 promoter was transcribed ∼2-fold more efficiently than the reverse TATA promoter. In conclusion, transcription from the synthetic templates in vitro occurred at levels that were reasonably comparable to those obtained from the natural U1 and U6 gene templates. DiscussionThe TATA box as a promoter element for RNAP III transcriptionWe have examined the ability of various TATA box sequences to selectively promote RNAP II or RNAP III transcription in a Drosophila nuclear extract. Our data indicate that canonical forward TATA box sequences (e.g. TATAAAAA, TATATAAA or TATAAATA) are strongly selective for RNAP II transcription. The cognate reverse TATA box, in one case, was highly selective for RNAP III (TTTTTATA), and in the other two cases (TTTATATA and TATTTATA) RNAP II and/or RNAP III transcription was preferentially supported depending upon context (i.e. polymerase specificity was dependent upon direction of transcription assayed). On the other hand, the symmetrical sequence TATATATA strongly favored RNAP II transcription in both directions of transcription. These results argue that T residues at positions 2 and 4 in the TATA box act as important determinants for RNAP III specificity in the crude Drosophila nuclear extract. Comparison of the in vitro transcription activities of synthetic H/D templates with the activities of natural U1 and U6 gene promoters. All analyses were carried out using the 1211z oligonucleotide for primer extension. Products arising from transcription of the wild-type U1, wild-type U6 and H/D templates are indicated. Other conditions and explanation of symbols are described in the legend to Figure 1. Several TATA box variants were found to be preferentially selective for RNAP III. However, none of these was as efficient for promoting RNAP III transcription as the original sequence first studied in earlier experiments (TTTTTATA) (17). Although the library of mutations tested is certainly not exhaustive, several conclusions can be drawn from the available data and can be summarized as follows. Position 1. Substitution of C for T (CTTTTATA) decreased RNAP III-specific transcription ∼3-fold (Fig. 4), yet the selectivity for RNAP III versus RNAP II remained very high (Fig. 3). Position 2. Data for all four nucleotides were obtained at the second position and the results indicate the following order of nucleotide preference for effective RNAP III transcription: T > A > C > G (compare TTTTTATA, TATTTATA, TCTTTATA and TGTTTATA in Fig. 4). Position 3. No comparative data is available since all TATA variants tested contained T at this position. Positions 4 and 5. A TpT dinucleotide appears to be optimal; RNAP III transcription from templates with either an ApT or TpA dinucleotide was significantly less efficient. Interestingly an ApA double substitution partly restored a higher level of expression (compare TTTTTATA versus TTTATATA versus TTTAAATA, and compare TTTTTAAA versus TTTTAAAA versus TTTAAAAA in Fig. 4). Positions 6 and 8. The functional role of the nucleotides at these positions is addressed by a double point substitution (TTTTTTTT versus TTTTTATA). In this case, RNAP III transcription was reduced 10-fold, indicating that transcriptional efficiency by RNAP III can be drastically affected by the specific nucleotides at positions 6 and/or 8 of the TATA box. Position 7. The importance of this position is revealed by the fact that an A or C substitution resulted respectively in either a 3- or 6-fold decrease in RNAP III transcription (TTTTTATA versus TTTTTAAA versus TTTTTACA in Fig. 4). This result is consistent with the recent observation that a seventh-position C substitution (in the TATA box context TATAAACA or TATTTACA) was extremely detrimental to TATA-mediated RNAP III transcription in a reconstituted yeast system (26). In the Drosophila SNF, the effect appears to be not quite as severe; alternatively, perhaps the five Ts in the upstream portion of our sequence (TTTTTACA) produce a stronger RNAP III TATA box and partially overcome the detrimental effect of the C at position 7. The contextual (flanking sequence) contribution to RNA polymerase specificity was not examined in this study. The templates were designed to minimize known RNAP II or RNAP III promoter elements other than the TATA box, but it is possible that cryptic elements may affect RNA polymerase specificity in a particular direction. This is evidenced by the fact that the sequences TTTATATA and TATTTATA promoted a mixture of RNAP II and III transcription in the rightward direction in our experiments, but when inverted favored nearly exclusive RNAP II transcription in the leftward direction (Fig. 1). This is not unreasonable in that the RNAP II and RNAP III basal transcription machineries contain polypeptide subunits in addition to TBP that directly contact the DNA in the pre-initiation complex, and in some instances these interactions exhibit sequence-specificity. For example, Drosophila TFIID interacts with both a consensus initiation site sequence and with a downstream promoter element (DPE) present in many TATA-less Drosophila mRNA-encoding genes (27). The RNAP II-specific factor TFIIB contacts DNA on each side of the TATA box (28,29); by analogy, the evolutionarily homologous RNAP III-specific factor BRF may make similar contacts with the DNA. The B″ subunit of TFIIIB contains a putative DNA-binding domain (30), and X-ray crystallography has revealed that TFIIA directly contacts DNA (31,32). Thus, multiple subunits of the basal transcription machinery may contribute to specific DNA recognition. In spite of this, transcription of our constructs by both RNAP II and RNAP III was highly TATA box-dependent since certain TATA mutations (e.g. TGTAAAAA, CATGGAAA and TTTCCATG) reduced transcription to essentially non-detectable levels for both RNA polymerases. Moreover, the quantitative comparisons of various TATA sequences as promoters for RNAP III transcription (Figs 3 and 4) were performed by assaying transcription in only the rightward direction. Since the TATA variants were present in identical environments in these constructions, the differences in the activities of these templates reflect the net effects of the specific TATA sequences that they contain. RNAP III versus RNAP II specificityIn this report we have examined the ability of a number of TATA box variants to preferentially support either RNAP II or RNAP III transcription in an unfractionated soluble nuclear extract that contains both polymerase activities. In general, our results suggest that runs of T residues in the first five positions of the TATA box favor RNAP III transcription, whereas alternating T-A residues favor RNAP II transcription. In an earlier paper, we interpreted these data to indicate that a ‘forward’ TATA box favored RNAP II transcription, and a ‘reverse’ TATA box RNAP III transcription (17). One mechanism to explain this phenomenon is that TBP binds in opposite orientations on RNAP II and RNAP III promoters (15,17). Alternatively, different TBP associated factors (TAFs) in TFIID and TFIIIB may differentially modulate the DNA binding specificity of TBP or themselves contribute to TATA box recognition. In this case, the sequence of the TATA box, rather than the orientation per se, may be the feature responsible for polymerase selection. Indeed, when complexed with TAFs in the RNAP I-specific factor SL1, TBP is unable to recognize a TATA box (33). Additional examples in which the DNA binding specificity of TBP may be modified by associated factors have been noted by other workers (26,34,35). Moreover, Whitehall et al. (26), using a yeast recombinant RNAP III system, have recently presented evidence that TBP binds in the same orientation on the yeast U6 RNAP III promoter as it does on RNAP II promoters. The data presented here are compatible with either model of TBP binding. Whether the preferences for RNAP II or RNAP III transcription observed in our experiments arise from TBP interacting in opposite orientations or alternatively arise from the differential action of TAFs in different TBP complexes can only be determined by further experimentation. A definitive answer will undoubtedly require high-resolution cross-linking experiments or the comparative topographical mapping of polypeptides in partially assembled RNAP II and RNAP III pre-initiation complexes. Significantly, however, our results provide an insight into the preferred functional DNA binding specificities of TFIIIB relative to that of TFIID. AcknowledgementsWe thank Cathy Tankersley for the construction of the wild-type Drosophila U6 gene template. This work was supported by Grant GM-33512 from the National Institutes of Health and in part by the California Metabolic Research Foundation. Y.W. is a recipient of the Arne N. Wick Predoctoral Research Fellowship from the California Metabolic Research Foundation. References1 . , , , vol. (pg. -) 3 , , , , , , , , , , , , . , . , Transcriptional Regulation. ,Cold Spring Harbor, Plainview, NY Cold Spring Harbor Laboratory Press (pg. -)8 , . , Proc. Natl. Acad. Sci. USA , , vol. (pg. -)9 , , . , , , vol. (pg. -) 10 , , , , . , , , vol. (pg. -) 11 , , , . , , , vol. (pg. -) 14 , , . , , , vol. (pg. -) 15 , . , , , vol. (pg. -) 17 , . , Proc. Natl. Acad. Sci. USA , , vol. (pg. -)18 19 , , . , , , vol. (pg. -) 20 , , . , Proc. Natl. Acad. Sci. USA , , vol. (pg. -)21 , . , , , vol. (pg. -) 22 , , , . , , , vol. (pg. -) 24 , , , . , , , vol. (pg. -) 29 , , , , , , , . , , , vol. (pg. -) 30 , , . , , , vol. (pg. -) 34 , , , , . , , , vol. (pg. -) 35 , , . , , , vol. (pg. -) |