A lesson in sharing?

Article metrics

The TATA-binding protein (TBP) is pivotal to transcription — it recognizes a specific DNA sequence, the TATA element, found in the control region of many genes, and is required for other proteins to assemble there and initiate transcription. A topic of much discussion has been to what degree the TAFs (TBP-associated factors) are required for TBP to do this. Initially TAFs were thought to be essential, but this idea was challenged by the observation that, in yeast, some TAFs are not generally required for transcription1,2. Moreover, a subset of the TAFs also contribute to the activities of other transcriptional regulatory complexes3,4,5,6. Now, four papers in Molecular Cell7,8,9,10 and one report in Cell11 fuel this debate by showing that a sub-group of TAF proteins, histone-like TAFs, may have a much broader role in gene expression.

Considerable work over the past decade led to the conclusion that TFIID — an evolutionarily conserved complex consisting of TBP and eight to twelve associated TAFs — rather than TBP alone is required for activation of protein-coding genes. Consistent with this, almost all TAFs were found to be essential for yeast viability. However, this model seemed all but destroyed when experiments to inactivate or deplete certain TAFs in yeast cells did not lead to the expected broad loss of activated transcription1,2. Moreover, TAF145/130, which was thought to be central to the assembly of TFIID, was found to regulate only a subset of genes, further deflating enthusiasm for a model of general TAF function12.

Now, in one of the new papers, Holstege et al.11 describe a transcriptional analysis of most of the yeast genome. They find that TAF145 is required for transcription of about 16% of the genes studied. Surprisingly, parallel analyses of another TFIID component, TAF17 (a histone-like TAF), by Apone et al.7, Michel et al.8 and Moqtaderi et al.9 indicate that it is broadly required, regulating the expression of at least 67% of all yeast genes7,11.

So what distinguishes TAF17 from TAF145? Sequence analysis reveals some similarity between TAF17, TAF60 and TAF68/61, and the histones H3, H4 and H2B, respectively. Perhaps these histone-like TAFs are important structurally, in a complex, through mutual interactions of their histone-fold motifs. Michel et al.8 examined the importance of histone-like TAFs by inactivating each of these proteins in yeast. These studies, along with inactivation of TAF68 by Natarajan et al.10, point to very general defects in transcription when histone-like TAFs are lost in yeast cells, and provide genetic evidence for interactions between these TAFs8. Although previous studies indicate that TAF60 and TAF68 do not have such broad effects12, the new studies imply that the histone-like TAFs are critical for the expression of most protein-coding genes.

These observations raise a dilemma — why are certain TAFs required for the transcription of more genes than others, if they are all subunits of the same TFIID complex? A possible explanation has come from studies of the yeast SAGA histone acetyltransferase complex, which contains a subset of TAF proteins including all three histone-like TAFs3 (Fig. 1). This arrangement seems to be conserved in homologous human acetyltransferase complexes4,6. SAGA is a large complex of around 20 different proteins, and it interacts directly with gene-specific activators and TBP3,10,13,14,15. Thus, a logical explanation for the broader transcriptional defect when histone-like TAFs (such as TAF17) are lost, but not when others (such as TAF145) are inactivated, is that the histone-like TAFs occur in other transcriptional regulatory complexes as well as TFIID.

Figure 1: Transcriptional complexes bound to an array of nucleosomes.

Both SAGA (SPT-ADA-GCN5 acetyltransferase) and TFIID contain the histone-like TAFs (yellow) and two additional TAFs, TAF25 and TAF90. Each complex contains a subunit with histone acetyltransferase activity (GCN5 in SAGA and TAF145 in TFIID) that is thought to lead to acetylation of histones in promoter proximal nucleosomes (purple). TFIID also contains additional TAFs, whereas SAGA contains ADA and SPT proteins. Both SAGA and TFIID can interact with transcriptional activators, and they bind or contain the TATA-binding protein, TBP. TFIID is also crucial in the formation of transcription complexes, including additional general transcription factors (GTFs) and RNA polymerase II.

It seems unlikely that the broad transcriptional defect observed when TAF17 is inactivated is simply due to a loss of SAGA activity. For example, GCN5, the catalytic subunit of SAGA, is required for the expression of only about 5% of all yeast genes11. Moreover, inactivation of another component of SAGA does not cause a general decrease in gene expression8. However, it is possible that the functions of TFIID and SAGA in transcription are redundant. So, only when shared components are lost is there a deficiency in both complexes, resulting in broad transcriptional defects. Loss of subunits unique to either TFIID or SAGA would then affect only specific subsets of genes11. This ‘redundancy’ theory is intriguing, because both TAF145 and GCN5 contain histone acetyltransferase activity. However, histone-like TAFs could exist in a third, as-yet-unidentified complex that is distinct from SAGA and TFIID. Or, perhaps the histone-like TAFs have a specific, essential function that is not shared by the other components of SAGA and TFIID. It remains to be seen whether functional redundancy exists, or whether the other, equally plausible, hypotheses hold true12.


  1. 1

    Moqtaderi, Z., Bai, Y., Poon, D., Weil, P. A. & Struhl, K. Nature 383, 188–191 (1996).

  2. 2

    Walker, S. S., Reese, J. C., Apone, L. M. & Green, M. R. Nature 383, 185–188 (1996).

  3. 3

    Grant, P. A. et al. Cell 94, 45–53 (1998).

  4. 4

    Ogryzko, V. V. et al. Cell 94, 35–44 (1998).

  5. 5

    Wieczorek, E., Brand, M., Jacq, X. & Tora, L. Nature 393, 187–191 (1998).

  6. 6

    Martinez, E., Kundu, T. K., Fu, J. & Roeder, R. G. J. Biol. Chem. 273, 23781–23785 (1998).

  7. 7

    Apone, L. M. et al. Mol. Cell 2, 653–661 (1998).

  8. 8

    Michel, B., Komarnitsky, P. & Buratowski, S. Mol. Cell 2, 663–673 (1998).

  9. 9

    Moqtaderi, Z., Keaveney, M. & Struhl, K. Mol. Cell 2, 675–682 (1998).

  10. 10

    Natarajan, K., Jackson, B. M., Rhee, E. & Hinnebusch, A. G. Mol. Cell 2, 683–692 (1998).

  11. 11

    Holstege, F. C. P. et al. Cell 95, 717–728 (1998).

  12. 12

    Hahn, S. Cell 95, 579–582 (1998).

  13. 13

    Saleh, A., Lang, V., Cook, R. & Brandl, C. J. J. Biol. Chem. 272, 5571–5578 (1997).

  14. 14

    Utley, R. T. et al. Nature 394, 498–502 (1998).

  15. 15

    15.Sterner, D. E. et al. Cell. Biol. (in the press).

Download references

Author information

Rights and permissions

Reprints and Permissions

About this article

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.