Report

TAF-Containing and TAF-Independent Forms of Transcriptionally Active TBP in Vivo

See allHide authors and affiliations

Science  19 May 2000:
Vol. 288, Issue 5469, pp. 1244-1248
DOI: 10.1126/science.288.5469.1244

Abstract

Transcriptional activity in yeast strongly correlates with promoter occupancy by general factors such as TATA binding protein (TBP), TFIIA, and TFIIB, but not with occupancy by TBP-associated factors (TAFs). Thus, TBP exists in at least two transcriptionally active forms in vivo. The TAF-containing form corresponds to the TFIID complex, whereas the form lacking TAFs corresponds to TBP itself or to some other TBP complex. Heat shock treatment altered the relative utilization of these TBP forms, with TFIID being favored. Promoter-specific variations in the association of these distinct forms of TBP may explain why only some yeast genes require TFIID for transcriptional activity in vivo.

Eukaryotic RNA polymerase II (Pol II) requires auxiliary factors to recognize promoters. The primary promoter recognition factor is TFIID, a complex that consists of TBP and about 10 TAFs (1). TBP binds TATA elements, which are found in most promoters, and it interacts with general transcription factors TFIIA and TFIIB (2). In yeast, TBP is generally required for Pol II transcription (3), and the level of TBP occupancy of promoters is correlated with transcriptional activity (4, 5). In the context of TFIID, certain TAFs directly contact initiator or downstream promoter elements (6). For this reason, TAFs are important for transcription in vitro from promoters lacking TATA elements, although they are dispensable for basal TATA-dependent transcription.

Analysis of the physiological functions of the TAF subunits of TFIID is complicated by the presence of certain TAFs in the SAGA histone acetylase complex (7). Many studies suggest that, in the context of TFIID, TAF subunits are not generally required for transcription because depletion or inactivation of individual TFIID-specific TAFs affects only a subset of genes (8–12). These results suggest that, with respect to functions within TFIID, there are distinct TAF-dependent and TAF-independent promoters, although this view has been challenged (13). In contrast to the TFIID-specific TAFs, TAFs that are present in both the TFIID and SAGA complexes are broadly required for transcription (12,14).

If TFIID is present at both TAF-dependent and TAF-independent promoters, the TAF/TBP occupancy ratio should be constant at all promoters. To investigate this issue, we used chromatin immunoprecipitation to measure promoter occupancy by three TFIID-specific TAFs (TAF130, TAF150, and TAF40); TAF17, which is present in both TFIID and SAGA (15); and TBP. This approach permits analysis of TAFs in wild-type cells under physiological conditions.

The TAF/TBP occupancy ratio varies considerably among promoters (Fig. 1). When normalized to levels of TBP occupancy, the TAF-dependent TRP3 (8,12) and ribosomal protein gene promoters RPS8A,RPL9A, and RPL5 (9,16) had six to seven times higher levels of TAFs than TAF-independent promoters such as ADH1, PGK, and PYK. TAF occupancy at the ACT1 andEFT2 promoters was half that observed on the ribosomal protein gene promoters. When normalized to TBP occupancy levels, the relative levels of all four TAFs tested were similar. Thus, these TAFs, and hence TFIID, were underrepresented at TAF-independent promoters. Similar observations for these and three additional TAFs have been obtained independently (17).

Figure 1

TAF and TBP occupancy at selected promoters. Crosslinked chromatin preparations from strains containing triple hemagglutinin (HA3)-tagged or untagged TAF130 (A), TAF40 (B), TAF150 (C), and TAF17 (D) were immunoprecipitated with monoclonal antibody to HA or polyclonal antibody to TBP (4). Polymerase chain reaction products corresponding to the indicated Pol II promoters, the tRNAArg (ACG) Pol III promoter, and the POL1structural gene were generated from total chromatin or immunoprecipitated DNA. For each promoter, the relative TAF/TBP occupancy ratio is indicated in terms of the percent of the maximal observed ratio, which is arbitrarily defined as 1.0 (21).

Because TBP occupancy is correlated with transcriptional activity (4, 5), the underrepresentation of TAFs at certain promoters suggests that there is a TAF-independent form of transcriptionally active TBP in vivo. To test this, we examined the occupancy of TFIIB and TFIIA at the same promoters (Fig. 2) and found that the TFIIA/TBP occupancy ratios were constant (within an experimental error of ±30%). Thus, associations of TBP, TFIIA, and TFIIB were very strongly correlated with each other, whereas the relationship with TAF association was much more variable. This observation indicates that recruitment of TAFs to promoters does not necessarily coincide with recruitment of TFIIA and TFIIB.

Figure 2

TFIIB, TFIIA, and TBP occupancy at selected promoters. Crosslinked chromatin preparations from strains containing HA3-tagged or untagged TFIIB were immunoprecipitated with monoclonal antibody to HA or polyclonal antibody to TBP or TFIIA. Experiments were done as in Fig. 1. For each promoter, the relative TFIIB/TBP (A) or TFIIA/TBP (B) occupancy ratio is indicated; a value of 1.0 is arbitrarily defined as the average ratio. ND = not determined.

We next examined activator-dependent recruitment of TFIID by monitoring TAF occupancy at promoters whose transcription is induced by heat shock factor or the Msn2 and Msn4 activators (Fig. 3, A to C). In all cases tested, TAF occupancy increased upon heat shock, which indicates that these activators increase recruitment of TFIID. The TAF/TBP occupancy ratios suggest that TFIID is underrepresented at heat shock–inducible promoters (18). Occupancy by TAF17 was two times higher than occupancy by the other TAFs, which suggests the possibility of activator-dependent recruitment of the SAGA complex (19). Interestingly, heat shock caused a two- to threefold increase in the TAF/TBP occupancy ratio at several promoters whose levels of transcription were unaffected (Fig. 3, D to F). This increased TAF occupancy was not due to more efficient crosslinking at high temperature because it was not observed at the uninducibleTRP3 and ARF1 promoters, which have inherently high TAF/TBP occupancy ratios (18).

Figure 3

TAF and TBP occupancy at heat shock–inducible and uninducible promoters in response to a transient heat shock. Crosslinked chromatin preparations from normally growing or heat shocked (15 min at 39°C) cells containing HA3-tagged or untagged TAF130 (A and D), TAF150 (Band E), and TAF17 (C and F) were immunoprecipitated with monoclonal antibody to HA or polyclonal antibody to TBP. Experiments were performed as in Fig. 1. (A to C) Heat shock–inducible promoters. Values for RPL9 under heat shock conditions are in parentheses to indicate very low occupancy levels due to transcriptional inhibition (24). (D to F) Promoters not inducible by heat shock. The fold increases in the TAF/TBP ratio as a consequence of heat shock are also indicated. ND = not determined.

Several lines of evidence indicate that our experiments provide quantitative measurements of promoter occupancy and are not influenced by conformational changes in proteins or DNA that affect crosslinking efficiency. First, associations of TBP, TFIIB, and TFIIA with promoters are remarkably well correlated with each other, even though individual promoter sequences are typically unrelated. Second, the relative occupancies of the four TAFs tested are strongly correlated with each other. Thus, any promoter-specific conformational difference that affects TAF crosslinking would have to affect all TAFs in a quantitatively similar manner. Third, the absolute level of crosslinking is comparable for all four TAFs, even though homologs of TAF17, TAF130, and TAF150 contact promoter DNA to various extents (6), and there is no evidence that homologs of TAF40 contact DNA. Similarly, chromatin immunoprecipitation has been applied successfully to many proteins that do not directly contact DNA (20). Thus, protein-protein crosslinks that occur in the vicinity of the promoter contribute significantly (and perhaps predominantly) to the observed crosslinking to DNA, and it is unlikely that protein-protein interactions within TFIID will change when it is bound to different promoters. For these reasons, the reduced levels of multiple TAFs at certain promoters is almost certainly due to the absence of these TAFs, and hence TFIID, at these promoters.

Our results provide strong evidence for at least two transcriptionally active forms of TBP in vivo (Fig. 4). One form corresponds to TFIID and is defined here by the association of the four TAFs tested with promoters and by a TAF/TBP ratio of 1.0 (21). The other form lacks the four TAFs tested (and probably other TAFs present in the TFIID complex) and is defined by promoters whose TAF/TBP occupancy ratio is significantly less than 1. This TAF-independent form might correspond to TBP itself, or it might be another TBP complex. For simplicity, we consider the TAF-independent form to be a single entity, although multiple TAF-independent forms of transcriptionally active TBP are possible. In addition, our results do not exclude the possibility of transcriptionally active forms of TBP containing a subset of TAFs.

Figure 4

Distinct forms of transcriptionally active TBP. Two forms of TBP are indicated, a TAF-containing form (presumably TFIID) and a TAF-independent form (depicted as free TBP, but it could contain non-TAF subunits). As indicated by the thickness of the arrows, TFIID is the predominant form at TAF-dependent promoters (e.g.,TRP3 and ribosomal protein gene promoters), whereas the non-TAF form predominates at TAF-independent promoters (e.g.,PGK, PYK, ADH1). Heat shock results in preferential utilization of TFIID (free TBP shown by smaller thinner arrows) even at promoters whose transcriptional activity and TBP occupancy are unaffected by heat shock. Heat shock could directly affect the relative activities or levels of TFIID and the TAF-independent form of TBP, or it could affect some other component(s) of the Pol II machinery that indirectly influences utilization of these TBP forms.

TFIID and the TAF-independent form(s) of TBP have distinct promoter selectivities. At one extreme, TFIID is the predominant (and perhaps exclusive) form of TBP at the TAF-dependent promoters. At the other extreme, the TAF-independent form predominates at the TAF-independent promoters, although TFIID may represent 10 to 20% of the transcriptionally active TBP. For the other promoters tested, the relative association of the two forms falls along a continuum between these extremes. The fact that heat shock–inducible promoters have intermediate TAF/TBP occupancy ratios suggests that heat shock factor and the Msn2 and Msn4 activators increase recruitment of both TFIID and the TAF-independent form of TBP.

The relative association of TFIID and the TAF-independent form of TBP is also affected by environmental conditions. Specifically, heat shock causes an increased TAF/TBP occupancy ratio at five of seven promoters that are transcriptionally unaffected; the two exceptions,TRP3 and ARF1, have the maximal TAF/TBP occupancy ratio of 1 even under normal growth conditions. The simplest interpretation of these results is that heat shock differentially affects the activity or amount of the two TBP forms, so that the relative utilization of TFIID is increased genome-wide, except for promoters where TFIID already predominates (Fig. 4).

The existence of a TAF-independent form of TBP in wild-type strains provides a simple explanation for the observation that transcription of many genes is unaffected upon destruction of TFIID by depletion of TFIID-specific TAFs (8–12). The low TAF occupancy at many promoters argues that the broad transcriptional effects reported to occur upon inactivation of the TFIID-specific TAF40 are indirect (13). In contrast, TAF-dependent promoters have high levels of TAF (and hence TFIID) occupancy, presumably because the TAF-independent form of TBP is not stably associated and hence not transcriptionally active. Because TFIID-specific TAFs contact DNA in the core promoter region (6) and have core-specific functions in vivo (8, 10, 12), differential occupancy by the distinct forms of TBP may reflect promoter-specific variations in the requirement for TAFs to stabilize TBP association.

The presence of TFIID-specific TAFs at all promoters tested suggests that TFIID contributes to transcription of most, and perhaps all, genes. However, at many promoters, this contribution is small compared with that of the TAF-independent form of TBP. Thus, the broad decrease in transcription observed upon depletion of TAFs present in TFIID and SAGA (12, 14) cannot be explained simply by destruction or inactivation of TFIID. Because SAGA-specific components are not essential for growth (22), widespread effects caused by depletion of certain TAFs are likely due to the simultaneous inactivation of TFIID and SAGA (and perhaps other TAF-containing complexes). Interestingly, TAF17, which is present in TFIID and SAGA and is broadly required for transcription, is underrepresented at many promoters to the same extent as the TFIID-specific TAFs. We suspect that, in the context of SAGA, TAF17 associates only transiently with promoters and/or is crosslinked with low efficiency.

The core RNA polymerases from bacteria and eukaryotes do not bind specific sequences. In bacteria, promoter recognition is provided by multiple σ factors that interact with the core RNA polymerase and direct the enzyme to specific classes of promoters. Many eukaryotes have distinct TBP-like proteins that show promoter specificity in vivo (23). In contrast to these structurally distinct promoter-recognition factors, yeast TBP exists in at least two distinct forms that differentially associate with promoters in vivo. These two forms may have distinct sequence recognition properties per se, and they may differ with respect to their ability to functionally interact with activators, repressors, or other transcriptional regulatory proteins.

  • * To whom correspondence should be addressed. E-mail: kevin{at}hms.harvard.edu

REFERENCES AND NOTES

View Abstract

Navigate This Article