View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector Cell, Vol. 99, 455–458, November 24, 1999, Copyright 1999 by Cell Press Repression: Targeting Minireview the Heart of the Matter

Edio Maldonado,* Michael Hampsey,‡ complex. This complex mediates the response to tran- and Danny Reinberg,‡†§ scriptional activators as well as repressors. Other co- *Instituto de Ciencias Biomedicas activators have also been identified. These include Programa de Biologia Celular y Molecular TFIIA, which stabilizes DNA-TBP binding, and factors Facultad de Medicina that affect nucleosome structure, including histone ace- Universidad de Chile tyltransferases and ATP-dependent chromatin remodel- Santiago, Chile ing complexes. † Howard Hughes Medical Institute Eukaryotic repressors, like activators, are typically ‡ Department of Biochemistry modular, consisting of either a single polypeptide with Division of Nucleic Acids Enzymology functionally distinct domains or multisubunit complexes University of Medicine and Dentistry of New Jersey with distinct functions distributed among the subunits Robert Wood Johnson Medical School (reviewed in Hanna-Rose and Hansen, 1996). These do- Piscataway, New Jersey 08854 mains (or polypeptides) can target different components of the transcription machinery to affect distinct steps in initiation (Figure 1). Thus, multiple mechanisms built Transcriptional repression is an essential mechanism in within a single repressor ensure that a gene can be the precise control of gene expression. Nearly 40 years silenced in an efficient manner. In this review we concen- ago, Jacob and Monod recognized the importance of trate on repressors that affect the formation of the initia- transcriptional repressor molecules in the regulation of tion complex. Repressors that target the transcriptional gene expression in bacteria. While these initial studies machinery subsequent to initiation have also been de- focused on regulation of the lactose operon of Esche- scribed, but are beyond the scope of this review. richia coli, it was soon realized that transcriptional re- The Relationship between Nucleosomes pression is a general mechanism affecting gene expres- and General Repressors sion in prokaryotes. Because the basic mechanism of The functional relevance of general transcriptional re- transcription in bacteria and eukaryotes is remarkably pressors was questioned initially, as eukaryotic DNA is well conserved, it came as no surprise that gene-specific packaged into chromatin, and the nucleosome repeat repressors were later identified in eukaryotes. unit of chromatin represses transcription by blocking Transcriptional repressor proteins associate with their the access of transcription factors to the DNA. Recently, target genes either directly through a DNA-binding do- however, genome-wide gene expression analysis was main or indirectly by interacting with other DNA-bound used to study the effects of nucleosome depletion on proteins. To inhibit transcription in a selective manner, gene expression in Saccharomyces cerevisiae (Wyrick a repressor protein can (1) mask a transcriptional activa- et al., 1999). These studies revealed that nucleosome tion domain, (2) block interaction of an activator with other components of the transcription machinery, or (3) displace an activator from the DNA (reviewed in John- son, 1995). Furthermore, DNA response elements can exert allosteric effects on transcriptional regulators, such that regulators may activate transcription in the context of one gene, yet repress transcription in another (reviewed in Lefstin and Yamamoto, 1998). Although the first repressor molecules identified were gene specific in function, transcriptional repressors that exert more general effects were identified subsequently. For example, the anti-␴ family of bacterial repressors regulate the transcription of many genes by binding to specific ␴ subunits and preventing assembly of holo- RNA polymerase (reviewed in Helmann, 1999). Similarly, general repressors in eukaryotes target components of the RNA polymerase II (RNAP II) core transcription ma- chinery to block the formation of initiation complexes. Figure 1. Schematic Representation of Repressor Targets during The RNAP II transcription preinitiation complex (PIC) Assembly of the Transcription Initiation Complex is composed of RNAP II and a set of general transcription The factors required for each step are indicated, along with the factors (GTFs) that includes the TATA-binding protein corresponding repressors. Ssn6 and Srb10/Srb11 are shown to af- (TBP), TFIIB, TFIIE, TFIIF, and TFIIH. RNAP II is recruited fect transcription at two different steps in the formation of the tran- to the promoter as a so-called holoenzyme that includes scription initiation complex. One of the targets of Srb10/Srb11 is a subset of the GTFs and the SRB/MED coactivator recruitment of RNAPII; however, we also speculate that Srb10/Srb11 may act at a later step by targeting TFIIH (for review, see Hampsey and Reinberg, 1999). Pink barrels, nucleosomes; orange dumbbells, § To whom correspondence should be addressed (e-mail: reinbedf@ activator proteins; GTFs, general transcription factors. The TATA umdnj.edu). motif is illustrated by a green box. Cell 456

depletion increases the expression of only a minority of genes, with the majority of genes being unaffected. Moreover, nucleosome depletion actually reduced the expression of a subset of genes. Position-dependent transcriptional repression (silenc- ing) has also been observed. Nucleosome depletion ex- periments demonstrated that genes located up to 20 kb from transcriptionally silent telomeres are derepressed upon nucleosome depletion. However, the SIR proteins, which maintain telomeric silencing, bind to a more telo- mere-proximal region of the DNA. These observations indicate that histones can make SIR-independent contri- butions to silencing and suggest that the repressive role of nucleosomes may be position-dependent rather than Figure 2. Model of the Structure of the TBP-TFIIA-TFIIB-Promoter global. Taken together, these observations suggest that DNA Complex the main function of nucleosomes is to compact DNA, The components are labeled as follows: TBP, gray; TFIIA, green; rather than to repress transcription. Whether these ef- TFIIB, blue; promoter DNA, red. The model was prepared by super- fects pertain in higher eukaryotes remains to be ad- imposing the crystallographic structures of the TBP-TFIIB-DNA and dressed. TBP-TFIIA-DNA complexes. Residues of TBP involved in interac- How then is transcriptional repression most often tions with Dr1/NC2 are highlighted in orange; residues of TBP ␣ helix 2 involved in interactions with Mot1, BRF, and TFIIA are high- achieved? A likely possibility is that, by interacting with lighted in yellow. The structure of the TFIIA-TBP-TATA complex components of the core transcription machinery, gen- does not show the interaction of TFIIA with amino acid residues in eral repressors, together with the nucleosomes, function TBP helix 2, although biochemical and genetic data implicate these globally to maintain genes in a repressed state. Indeed, residues in TFIIA-TBP interaction (Orphanides et al., 1996). This the theme that has emerged from studies of general discrepancy might be accounted for by the absence of the central repressors is that proper control of gene expression domain of the larger subunit of TFIIA in the structural studies. (The depends on the combinatorial action of gene-specific figure was generously provided by Richard H. Ebright, Rutgers Uni- versity.) activators and general negative regulators (see, for ex- ample, http://web.wi.mit.edu/young/expression/). Be- low we discuss the mechanisms by which various types and general repressors in the regulation of eukaryotic of general repressors target the core transcription ma- gene expression. chinery. The mechanism of Dr1-mediated repression might be The Dr1/Drap1 Complex loosely analogous to that of the anti-␴ family of tran- The human Dr1/Drap1 complex, also known as NC2, scription factors in bacteria. Dr1-mediated repression was isolated as an activity that repressed transcription requires Dr1 binding to TBP (or TFIID). The TBP-Dr1 by RNAP II in a reconstituted system independent of complex precludes the interaction between TBP (or activators. The complex is composed of two subunits, TFIID) and TFIIB in vitro and thus halts the formation of Dr1 (NC2␤), which participates directly in transcriptional the transcription initiation complex. These conclusions repression, and Drap1 (NC2␣), a regulatory subunit that are supported by experiments in yeast demonstrating enhances the activity of Dr1. Dr1 and Drap1 form a that overexpression of Dr1 diminishes mRNA accumula- heterodimer through a histone fold motif. Human Dr1 is tion and confers a slow growth phenotype that is sup- a modular protein with at least four independent func- tional domains. pressed by TBP overexpression (Kim et al., 1997). Fur- The Dr1/Drap1 repressor has been conserved through- thermore, altered forms of yeast TBP were identified out evolution. Yeast Dr1 and Drap1 homologs were iden- that suppress suc2⌬UAS, some of which are defective tified by sequence analysis of the yeast genome and in for interaction with the Dr1/Drap1 complex. These muta- genetic selections designed to identify negative regula- tions are clustered in a previously undefined domain of tors of transcription (reviewed in Hampsey, 1998). The TBP adjacent to the TFIIB-binding site (Figure 2) (Cang yeast Drap1 homolog was identified in two different ge- et al., 1999). Thus, Dr1 appears to repress transcription netic selections. In one screen, suppressors were se- by preventing the association of TFIIB with TBP. Inter- lected that increased transcription of the SUC2 gene in estingly, TFIIA competes with Dr1 for binding to TBP, the absence of its enhancer (UAS) element (suc2⌬UAS). yet TFIIA and TFIIB bind to diametrically opposed sur- One suppressor, bur6 (bypass of UAS requirement), not faces of TBP (Figure 2). Structural studies have shown only enhanced suc2⌬UAS expression, but conferred that the association of TFIIA with TBP does not alter the multiple pleiotropic phenotypes, consistent with the no- conformation of the TBP-DNA binary complex (reviewed tion that Drap1 plays a general role in transcriptional in Orphanides et al., 1996). We therefore suggest that repression. An independent screen for suppressors of Dr1 either induces a conformational change in TBP that a mutation in SRB4, which encodes a component of affects TFIIA binding or alters the structure of the TBP- the SRB/MED complex, identified the genes encoding DNA complex. Thus, Dr1-mediated repression is likely Drap1 and Dr1 (Lee et al., 1998). Moreover, mutations to involve two mechanisms, one blocking TFIIB associa- in these genes alleviate the global defects in mRNA tion and the other inducing a conformational change in synthesis observed in the srb4 mutant. These results TBP or DNA that alters TFIIA binding. confirm that Dr1/Drap1 is a general repressor and under- Because Dr1 functions through TBP, which is required score the interdependence of the RNAP II holoenzyme for transcription by all three eukaryotic RNAPs, it was Minireview 457

thought that Dr1 might also affect transcription by RNAP by all three RNAPs, Mot1 is specific for RNAP II tran- I and RNAP III. Yet Dr1 represses transcription by RNAP scription. Even transcription from the RNAP III U6 pro- III and not by RNAP I. The molecular mechanism by moter, which includes a functional TATA box, is unaf- which Dr1 inhibits RNAP III transcription is unknown, fected by Mot1. Therefore, a TATA box is not the sole but repression is mediated through TFIIIB. Yeast TFIIIB determinant of transcriptional regulation by Mot1, a con- comprises three subunits, TBP, BЈЈ, and BRF. BRF is clusion that is consistent with the above model of Mot1 composed of at least two functional domains: the N function. terminus is homologous to TFIIB, whereas the C termi- The Srb10/Srb11 Complex nus is unique to the BRF family of proteins. Surprisingly, Srb10 is a cyclin-dependent kinase regulated by the it is the C-terminal domain of BRF that interacts with Srb11 cyclin. The yeast Srb10/Srb11 complex is homolo- TBP. Furthermore, the TBP residues required for interac- gous to Cdk8/cyclin C in mammalian cells. Both com- tion with BRF are distinct from those that interact with plexes are components of their respective RNAP II holo- TFIIB, but overlap with those that interact with TFIIA enzymes, which is consistent with the identification of (Figure 2) (Colbert et al., 1998). By analogy to its effect srb10 and srb11 as suppressors of truncations in the on RNAP II transcription, Dr1 might repress RNAP III carboxy-terminal repeat domain (CTD) of the largest transcription by blocking the association of BЈЈ or BRF subunit of yeast RNAP II (reviewed in Hampsey and with TBP, or by inducing a TBP and/or DNA conforma- Reinberg, 1999). The Srb10/Srb11 complex is a negative tional change that precludes assembly of the TFIIIB regulator of transcription that can target the CTD. Phos- complex. phorylation of the CTD by the Srb10 kinase occurs prior Mot1 to the association of RNAP II with the transcription pre- Another transcription factor that targets TBP is Mot1. initiation complex and precludes complex assembly. In contrast to Dr1, Mot1 displaces TBP from DNA in vitro This is in contrast to CTD phosphorylation by TFIIH, in a manner dependent upon ATP hydrolysis. Consistent which occurs subsequent to complex assembly and is with its ability to displace TBP, Mot1 represses tran- a prerequisite for transcription. As might be expected scription in vitro. This effect is counteracted by TFIIA, based on their different functions, Cdk8/cyclin C and but in contrast to its effect on Dr1-mediated repression, TFIIH kinase complexes also differ in CTD substrate TFIIA competes with Mot1 for binding to the same sur- specificity. TFIIH phosphorylates exclusively Ser-5, while face of TBP (Figure 2). Yeast Mot1 exists in a complex Cdk8/cyclin C phosphorylates both Ser-2 and Ser-5 (re- with TBP that is distinct from the TFIID complex. Simi- viewed in Hampsey and Reinberg, 1999). The Srb10/ larly, a human homolog of Mot1 was identified as the Srb11 complex has not been shown to phosphorylate

TAFII170 subunit of B-TFIID, an alternative form of TFIID Ser-2; however, Srb10/Srb11 exists in a complex with that includes TBP (reviewed in Hampsey, 1998). Srb8 and Srb9. Interestingly, the SRB9 gene was identi- Yeast Mot1 was also identified genetically, initially in fied in a genetic screen as a suppressor of a Ser-2 a screen for mutants that enhanced gene expression in to Ala replacement in the CTD, suggesting that Srb10/ the absence of an activator, and subsequently in the Srb11 also targets Ser-2. screen for bur mutants mentioned above. In both cases, The CTD of RNAP II is unlikely to be the sole target mot1 mutations stimulate basal, rather than activated of the Cdk8/cyclin C complex. Distinct RNAP II holoen- transcription. However, mot1 mutations have also been zyme complexes have been isolated. These complexes reported to diminish gene expression in vivo, specifically are composed of various modules that, depending upon from TATA-less promoters. their subunit composition, can have either positive or How can these seemingly disparate results be recon- negative effects upon transcription. Whether these mod- ciled? Recently, Mot1 was found to act as either a re- ules activate or repress transcription is determined in pressor or activator in vitro, depending on the relative part by Cdk8/cyclin C. In the absence of Cdk8/cyclin C, concentrations of Mot1, TBP, and DNA. Transcriptional these modules function as coactivators, whereas in the repression is observed in the presence of high concen- presence of Cdk8/cyclin C they repress transcription. trations of either Mot1 or DNA, presumably as a conse- Importantly, these effects are independent of the CTD, quence of displacing or sequestering TBP, respectively. suggesting that Cdk8/cyclin C targets additional com- Mot1 stimulates transcription in vitro only when present ponents of the transcriptional machinery (reviewed in in amounts that are stoichiometric with TBP. These re- Hampsey and Reinberg, 1999). sults suggest that Mot1 regulates the distribution of TBP The Not Complexes between promoter and nonpromoter sites (Muldrow et The HIS3 promoter includes two functionally distinct al., 1999). This premise is also consistent with the syn- TATA elements, TC and TR, that are utilized differentially. thetic lethal interaction between mutations in Mot1 and TC directs constitutive HIS3 expression, whereas TR is TFIIA (Collart, 1996; Madison and Winston, 1997). Pre- required for activation in response to Gcn4. Mutations sumably, viability requires either TFIIA to increase the in five different yeast genes, designated NOT1 through stability of the DNA-TBP complex, or Mot1 to increase NOT5, were identified in a genetic screen for mutants the TBP concentration at functional promoters. Accord- that specifically enhance expression from the TC ele- ingly, Mot1 is not a bona fide transcriptional repressor, ment. These mutants showed enhanced expression of but is instead a unique factor that functions by releasing several functionally distinct genes, suggesting that the TBP—a limiting factor in the cell—from nonproductive NOT proteins function as general repressors (reviewed binding sites. These results suggest that Mot1 primarily in Hampsey, 1998). affects promoters whose mechanism of TBP recruitment The NOT proteins are components of multisubunit is weak, including TATA-less promoters. complexes. Consistent with a general role for NOT pro- Although it might be expected to affect transcription teins in transcription, Not1 interacts directly with TBP Cell 458

(Lee et al., 1998). Mutations in the NOT genes were also have revealed that subsets of repressors overlap in their identified as suppressors of srb4. The same selection spectrum of affected genes. Biochemical studies have uncovered the genes encoding Dr1, Drap1, and Mot1, defined mechanisms by which the general repressors all of which function through TBP (Lee et al., 1998). function and genetic experiments have demonstrated These observations suggest that the NOT proteins are their relevance in vivo. The pattern that has begun to global transcriptional repressors that, like Dr1 and Mot1, emerge from these studies is that general repressors target the general transcription machinery. Furthermore, interact to regulate multiple steps in gene expression these results define a coordinated and balanced interac- from chromatin unfolding to transcription. We anticipate tion among different negative regulators and positive that future studies using multicellular organisms will un- cofactors to control gene expression. derscore the critical nature of repressors in cell growth Multifunctional Repressors and development. The repressors described above target exclusively com- ponents of the general transcriptional machinery. How- Selected Reading ever, general repressors have been described that target components of both the general machinery and chroma- Brehm, A., Miska, E.A., McCance, D.J., Reid, J.L., Bannister, A.J., and Kouzarides, T. (1998). Nature 391, 597–601. tin. Below we discuss representative members of this Cang, Y., Auble, D.T., and Prelich, G. (1999). EMBO J., in press. class of transcriptional repressor proteins. Colbert, T., Lee, S., Schimmack, G., and Hahn, S. (1998). Mol. Cell. RBP (CBF1) is a DNA sequence-specific repressor Biol. 18, 1682–1691. that can also function by recruiting factors that activate Collart, M.A. (1996). Mol. Cell. Biol. 16, 6668–6676. transcription. RBP represses transcription of the adeno- Hampsey, M. (1998). Microbiol. Mol. Biol. Rev. 62, 465–503. viral pIX promoter by interacting with the transcriptional coactivators TAF 110 and TFIIA. This effect does not Hampsey, M., and Reinberg, D. (1999). Curr. Opin. Genet. Dev. 9, II 132–139. preclude the association of the transcriptional activator Hanna-Rose, W., and Hansen, U. (1996). Trends Genet. 12, 229–234. Sp1 with TAFII110, but impairs the TAFII110–TFIIA inter- action that is essential for activation (Olave et al., 1998). Helmann, J.D. (1999). Curr. Opin. Microbiol. 2, 135–141. This is the first example of a repressor that targets the Johnson, A.D. (1995). Cell 81, 655–658. coactivator machinery to repress transcription. The Rb Kao, H.Y., Ordentlich, P., Koyano-Nakagawa, N., Tang, Z., Downes, M., Kintner, C.R., Evans, R.M., and Kadesch, T. (1998). Genes Dev. tumor suppressor protein functions in a similar manner 12, 2269–2277. by blocking E2F-mediated recruitment of TFIID and Kim, S., Na, J.G., Hampsey, M., and Reinberg, D. (1997). Proc. Natl. TFIIA (Ross et al., 1999). Acad. Sci. USA 94, 820–825. RBP and Rb can also repress transcription by tar- Knoepfler, P.S., and Eisenman, R.N. (1999). Cell 99, this issue, geting histone deacetylases and presumably altering 447–450. chromatin structure. RBP interacts with the SMRT (si- Lee, T.I., Wyrick, J.J., Koh, S.S., Jennings, E.G., Gadbois, E.L., and lencing mediator of retinoid and thyroid hormone recep- Young, R.A. (1998). Mol. Cell. Biol. 18, 4455–4462. tors) protein, which recruits the Sin3 histone deacetylase Lefstin, J.A., and Yamamoto, K.R. (1998). Nature 392, 885–888. complex (Kao et al., 1998). Rb also associates with the Madison, J.M., and Winston, F. (1997). Mol. Cell. Biol. 17, 287–295. Sin3 complex through the Rb pocket domain (Brehm et Muldrow, T.A., Campbell, A.M., Weil, P.A., and Auble, D.T. (1999). al., 1998). In this case, the mechanism of transcriptional Mol. Cell. Biol. 19, 2835–2845. repression involves recruitment by E2F of a complex Olave, I., Reinberg, D., and Vales, L.D. (1998). Genes Dev. 12, 1621– that includes the hypophosphorylated form of Rb in as- 1637. sociation with the Sin3 complex. The Rb-Sin3 complex Orphanides, G., LaGrange, T., and Reinberg, D. (1996). Genes Dev. presumably alters the local chromatin structure to re- 10, 2657–2683. press transcription. These examples define Sin3 as a Ross, J.F., Liu, X., and Dynlacht, B.D. (1999). Mol. Cell 3, 195–205. general repressor that functions via histone deacety- Wyrick, J.J., Holstege, F.C.P., Jennings, E.G., Causton, H., Shore, lases. However, Sin3 also appears to mediate repression D., Grunstein, M., Lander, E.S., and Young, R.A. (1999). Nature 402, independently of deacetylation, as mutations in mam- 418–421. malian Sin3 that abolish its interaction with deacetylases do not impede its ability to repress transcription (see Knoepfler and Eisenman, 1999 [this issue of Cell]). Taken together, these observations suggest that repressor complexes can affect transcription by multiple mecha- nisms. Furthermore, repressors that target histone de- acetylases and the general transcription machinery are not restricted to mammalian cells. A comparable exam- ple is the yeast Ssn6-Tup1 repressor, which has been reported to target both chromatin and the general tran- scription machinery of S. cerevisiae (reviewed in Hamp- sey, 1998). Concluding Remarks This review analyzes several functionally distinct, gen- eral repressors of transcription. Recent whole-genome, microarray analyses of gene expression in yeast cells that have been independently depleted of specific gen- eral transcriptional repressors, including nucleosomes,