Minimal components of the RNA polymerase II transcription apparatus determine the consensus TATA box

Abstract

In Saccharomyces cerevisiae, multiple approaches have arrived at a consensus TATA box sequence of TATA(T/A)A(A/T)(A/G). TATA-binding protein (TBP) affinity alone does not determine TATA box function. To discover how a minimal set of factors required for basal and activated transcription contributed to the sequence requirements for a functional TATA box, we performed transcription reactions using highly purified proteins and CYC1 promoter TATA box mutants. The TATA box consensus sequence is a good predictor of promoter activity. However, several nonconsensus sequences are almost fully functional, indicating that mechanistic requirements are not the only selective pressure on the TATA box. We also found that the effect of a mutation at a certain position is often dependent on other bases within a particular TATA box. Although activators and coactivators strongly influence TBP recruitment and stability at promoters, neither Mediator, the activator Gal4-V16, nor TFIID specifically compensate for the low transcription levels of the weak TATA boxes. The addition of Mediator to purified transcription reactions did, however, increase the functional selectivity for certain consensus TATA sequences. Transcription in whole-cell extracts or in vivo with these TATA box mutants indicated that factors, other than those in our purified system, may help initiate transcription from weak TATA boxes.

Figure 1.

The CYC1 promoter of the transcription templates used in this work. (A) Above is shown the double-stranded sequence of the CYC1 promoter fragment fused to the G-less cassette in the original pJJ470 plasmid (40). Potential transcriptional start sites (41,65,66) located in the G-less cassette are indicated by dots. Below is the double-stranded sequence of the modified CYC1 promoter from the pJJ470(MA) plasmid. Mutations in the TATA-like sequences are shown in bold. (B) List of the wild-type and all 23 TATA box mutants that were used in this study. Whether the sequence was a consensus TATA box and/or TBP sensitive were based on a previous study (4,53). TBP-sensitive promoters show decreased transcription in yeast strains with mutations in the DNA-binding surface of TBP (53).

Figure 2.

Transcription activities of CYC1 TATA box mutants in a purified system in vitro. Mutations that represent another consensus base (4) at that position are underlined. Mutants that are sensitive to mutations in the DNA-binding surface of TBP (4) are marked with an asterisk. Transcription levels were measured in a G-less cassette assay. Basal transcription was measured in the presence of the highly purified general transcription factors RNA Pol II, TBP, TFIIB, TFIIE, TFIIF and TFIIH. Enhanced transcription was measured in the presence of purified GTFs and Mediator. Activated transcription was measured in the presence of purified GTFs, Mediator and Gal4-VP16. Shown is the quantification of transcription signals, expressed as a percentage of the signal from a wild-type TATA box template. Representative primary data are shown in Figures 3–5. The CYC1 core promoter/G-less cassette has two major start sites for transcription. We chose only the upstream start site (longer transcript) for quantification. Most columns represent the mean of at least three independent experiments. Reactions with signals that were almost undetectable were only measured twice. Error bars indicate the SEM. RNA recovery controls were run for most reactions and showed a consistent recovery over all samples checked.

Figure 3.

Certain nonconsensus TATA boxes facilitate high rates of transcription. Shown are basal transcription activities of wild-type and mutant CYC1 TATA boxes in a purified system in vitro. Mutated bases are shown in bold. Lanes 2 and 4 show nonconsensus TATA boxes with high transcription signals. Lanes 3 and 5 serve as controls and show nonconsensus TATA boxes that give low transcription signals. Basal transcription was measured in a G-less cassette assay that was performed in the presence of highly purified RNA Pol II, TBP, TFIIB, TFIIE, TFIIF and TFIIH.

Figure 4.

Substituting a G for the A in the eighth position of the TATA box suppresses the transcription defect in the TATA consensus sequence TATAAAAA. Shown are basal, enhanced basal and activated transcription activities of CYC1 TATA box mutants and wild type in a purified system in vitro. Lanes 10–12 show the transcription with the wild-type TATA box. Lanes 1–3 show the transcription decrease with a T5A mutation in the TATA box. Placing a G in the eighth position compensates for the transcription defect in the TATAAAAA sequence (lanes 4–6). Placing a G in the eighth position of the wild-type TATA box, however, leads to no increase in transcription (lanes 7–9). G-less cassette assays were performed in the presence of highly purified RNA Pol II, TBP, TFIIB, TFIIE, TFIIF and TFIIH, in addition to the factors indicated. Salt concentration was 165 mM KOAc and reaction time was 60 min for all reactions. Mutated bases are shown in bold.

Figure 5.

Mediator-enhanced basal transcription is disabled by substitution of a T in the eighth position of the CYC1 TATA box, but is functional with either A, C or G. Shown are basal, enhanced basal and activated transcription activities of CYC1 TATA box mutants and wild type in a purified system in vitro. Substitution of a T in the eighth position of the CYC1 TATA box specifically disables Mediator enhanced basal transcription (lanes 10–12 versus lanes 1–3), while substitution with a C (lanes 4–6) or G (lanes 7–9) has no effect. G-less cassette assays were performed in the presence of highly purified RNA Pol II, TBP, TFIIB, TFIIE, TFIIF and TFIIH, in addition to the factors indicated. Salt concentration was 165 mM KOAc and reaction time was 60 min for all reactions. Mutated bases are shown in bold.

Figure 6.

TFIID does not lead to greater enhancement of transcription from weak CYC1 TATA boxes than from strong TATA boxes in a purified system. Transcription was measured in a G-less cassette assay that was performed in the presence of highly purified RNA Pol II, TBP, TFIIB, TFIIE, TFIIF, TFIIH, TFIIA and Mediator. The reactions were performed in the absence (A) or presence (B) of the activator Gal4-VP16. Quantitation of the fold-enhancement by TFIID is shown for each TATA box. Mutated bases are shown in bold.

Figure 7.

Transcription activities of wild-type and mutant CYC1 TATA boxes in an extract-based system in vitro, and in a reporter assay in vivo. Consensus mutants (4) are underlined. Mutants that are sensitive to mutations in the DNA-binding surface of TBP (4) are marked with an asterisk. Activated transcription levels with purified factors are from Figure 2 and are repeated here for ease of comparison. The transcription levels in an extract-based system are results from G-less cassette assays performed in the presence of a whole-cell extract supplemented with the activator Gal4-VP16. The final concentration of TBP was about 10 times lower in the extract-based system than in the purified system. RNA recovery controls were run for the purified and extract-based transcription reactions and showed a consistent recovery over all samples checked. Shown is the quantification of transcription signals, expressed as a percentage of the wild-type signal. Most columns represent the mean of at least three independent experiments. Some reactions whose signals were almost undetectable were only measured twice. Transcription levels in vivo were measured in a β-galactosidase reporter assay in a yeast strain expressing the Gal4-VP16 (F456A) fusion protein. Shown is the quantification of β-galactosidase activity units expressed as a percentage of the wild type. The columns represent the mean of three independent experiments. Error bars indicate the SEM.