B TAFContaining Complexes

Similar to TBP that forms SL1, TFIID, TFIIIB, and TAC complexes, some TAF components in TFIID are also present in distinct complexes such as TBP-free TAFn-containing complex (TFTC), Spt-Ada-Gcn5-acetyltransferase (SAGA), Spt3-TAFn31-GCN5L (i.e., the long form of GCN5) acetylase (STAGA), and polycomb repressive complex 1 (PRC1). These TBP-lacking TAF-containing complexes are involved in diverse aspects of pol II-dependent transcription. Except PRC1 whose role is mainly in gene silencing, the other TAF-containing complexes are mostly involved in activator-dependent transcription likely due to the HAT activity inherent to each complex.

TFTC, originally identified in HeLa cells using a monoclonal antibody against human TAF 10 (Wieczorek et al., 1998), contains TAF2, TAF5, TAF5L, TAF6L, TAF7, and some histone fold-containing TAFs, including TAF4, TAF6, TAF9, TAF 10, and TAF12. In addition, TFTC also has TRRAP (transformation-transactivation domain-associated protein), SAP130 (spliceosome-associated protein 130), GCN5L HAT enzyme, Ada3 adaptor protein, and histone fold-containing Spt3 (Cavusoglu et al., 2003). The three-dimensional structures of TFTC and TFIID, both resolved at 35 A resolution by electron microscopy and single-particle image analysis, resemble a macromolecular clamp consisting of five (for TFTC) or four (for TFIID) globular domains organized around a solvent-accessible groove that may accommodate a DNA duplex (Andel et al., 1999; Brand et al., 1999a).

This configuration suggests that TFTC may adapt a DNA-binding conformation similar to that exhibited by TFIID. Indeed, like TFIID, TFTC is able to support basal and Gal4-VP16-mediated transcription in vitro from TATA-containing promoters and also basal transcription from a TATA-less promoter, presumably via TAF recognition of the core promoter and interaction with other components of the general transcription machinery (Wieczorek et al., 1998). Moreover, TFTC has coactivator activity likely through direct protein-protein contacts between the activation domain of Gal4-VP16 and multiple subunits of TFTC including TAF5, TAF6, TAF 10, SAP 130, Spt3, GCN5L, and TRRAP (Hardy et al., 2002). This activator-TFTC interaction may also facilitate p300-mediated transcription from a Gal4-VP16-dependent chromatin template (Hardy et al., 2002), probably resulting from acetylation of nucleosomal core histones H3 and H4 by GCN5L (Brand et al., 1999b). Besides the HAT component, the presence of many histone fold-containing proteins, including Spt3, TAF4, TAF6, TAF9, TAF 10, and TAF 12, further contributes to the integrity of the TFTC complex.

Similar to TFTC, yeast SAGA also lacks TBP but contains some common subunits such as Tral, Ada3, TAF5, the short form of GCN5, and histone fold-containing proteins Spt3, TAF6, TAF9, TAF 10, and TAF 12. The unique components present in SAGA, not found in TFTC, include Tral, Spt7, Spt8, Spt20, Sgf29, Adal, and Ada2 (Grant et al., 1998). These SAGA components are organized into five distinct domains, analogous to the structure of human TFTC, when resolved by electron microscopy at 31 A resolution (Wu et al., 2004). The largest subunit of SAGA, Tral, is located in the outermost globular domain able to contact transcriptional activators, as evidenced by the fluorescence resonance energy transfer (FRET) technique measuring direct in vivo association between yeast Gal4 activator and Tral, but not with other subunits in the SAGA complex (Bhaumik et al., 2004). This is consistent with the finding that the human homologue of Tral, TRRAP (27.3% identity and 58.9% similarity with S. cerevisiae Tral at the amino acid level), also interacts with c-Myc and E2F-1 oncoproteins implicated in transformation and transactivation (McMahon et al., 1998) and then recruits GCN5 HAT activity to activator-recruited transcriptional complexes (McMahon et al., 2000). However, whether recruitment of TRRAP and GCN5 occurs in the context of human SAGA-like complexes, such as TFTC, PCAF and GCN5 complexes (Ogryzko et al., 1998), remains to be investigated. Interestingly, approximately 10% of yeast genes, which are mostly driven by TATA-containing promoters and are typically stress-induced, appear to be SAGA-dependent (Huisinga and Pugh, 2004). These TFIID-independent yeast promoters are recognized by the free form of TBP that in turn recruits SAGA to the targeted promoters via direct interaction between TBP and the Spt3 subunit of SAGA. Clearly, some components of the SAGA complex, such as Spt7, Spt20, Adal, TAF5, TAF 10, and TAF 12 are necessary for the structural integrity of SAGA, as mutations in these subunits result in disruption of the holo complex (Grant et al., 1998; Sterner et al., 1999; Durso et al., 2001; Kirschner et al., 2002).

STAGA, the human counterpart of yeast SAGA initially isolated from HeLa nuclear extracts using polyclonal antibodies against human TAF9 (Martinez et al., 1998), contains many homologues found in yeast SAGA, including TRRAP, Adal, Ada2, Ada3, Spt3, Spt7, and histone fold-containing TAFs (TAF9, TAF 10, and TAF 12). While additional components of STAGA, TAF5L, TAF6L, SAP130, and GCN5L are also present in TFTC, the identities of the other subunits, such as STAF36, STAF42, STAF46, STAF55, STAF60, and STAF65y remain to be characterized (Martinez et al., 2001). The conservation of histone fold-containing Spt3, TAF9, TAF 10, and TAF 12 in both SAGA and STAGA suggests that these subunits are important for the structural integrity of these distinct HAT complexes. As seen with yeast SAGA, the presence of TRRAP, GCN5L and adaptor proteins Adal, Ada2, and Ada3 in STAGA likely accounts for the coactivating activity of STAGA in supporting Gal4-VP16-mediated activation from a Gal4-driven chromatin template (Martinez et al.,

2001). Clearly, Ada proteins facilitate access of GCN5L to the chromatin template and thus allow acetylation on nucleosomal core histones (Balasubramanian et al.,

2002). That adaptor proteins are additionally required for activator-facilitated chromatin targeting of GCN5 is further supported by the observation that recombinant GCN5 protein is unable to potentiate activator-dependent transcription from preassembled chromatin templates (Thomas and Chiang, 2005). Based on our understanding of the yeast SAGA complex, it is likely that the Spt3 component of STAGA may contact TBP to facilitate transcription from STAGA-dependent promoters.

Unlike HAT-containing TFTC, SAGA and STAGA complexes involved in gene activation, the TAF-containing Drosophila PRC1 complex is implicated in repression of homeotic genes that govern body segmentation and the developmental process. This complex, initially isolated from Drosophila embryos using a monoclonal antibody against epitope-tagged polyhomeotic (PH) or posterior sex comb (PSC) protein (Shao et al., 1999), contains approximately 30 subunits, including TAF1, TAF4, TAF5, TAF6, TAF9, TAF11, PH, PSC, PC (polycomb), RING1, Zeste, HSC4, SMRTER, Mi-2, Sin3A, Rpd3, p55, Sbfl, DRE4/Sptl6, p90, HSC3, Modulo, Reptin, DNA Topoisomerase II, pi 10, Tubulin, Actin, Ribosome RS2, and Ribosome RL10 (Saurin et al., 2001). Although this holo complex has histone deacetylase (HDAC) Rpd3, chromatin remodeling ATPase Mi-2, and other corepressor components (SMRTER, Sin3A, and p55) likely contributing to repression of transcription and inhibition of chromatin remodeling, it is surprising to see that a PRC1 core complex (PCC) containing only PH, PSC, PC, and RING1 is sufficient for transcriptional silencing (King et al., 2002) and also blocking SWI/SNF-mediated mobilization of a nucleosomal array (Francis et al., 2001). This finding plus the fact that a sequence-specific transcription factor Zeste is present in PRC1 suggests that PRC1 may use different mechanisms to target PRC 1-regulated gene transcription, irrespective of the presence or absence of a Zeste-binding element (Mulholland et al., 2003). The roles of TAFs, especially the enzymatic activities of TAF1, and the other subunits constituting PRC1 await further investigation.

C: TAF Variants During the Developmental Process

Some components of TFIID are present in a substoichiometric ratio relative to the other TAFs. These TAFs, such as TAF4b, TAF5L, and TAF7L, are often found in a tissue-specific manner and likely confers TFIID unique properties functioning in a specialized environment. TAF4b, a paralogue of TAF4 initially identified in TFIID purified from B cells and later found expressed at low levels in every cell types but specifically enriched in the testes and ovary, appears to function in a gonad-specific manner, as knockout of this TAF variant severely affects ovarian development in female mice (Freiman et al., 2001) and also spermatogenesis in male mice (Falender et al., 2005). Likewise, TAF5L and TAF7L, paralogues of TAF5 and TAF7, respectively, are implicated in male gametogenesis (Hiller et al., 2001; Pointud et al., 2003). The presence of these tissue-specific TAF variants likely enables TFIID to work in conjunction with germ cell-specific transcription factors, cofactors, or other components of the general transcription machinery, such as TFIIAaP-like factor (see Chapter by Hou and Chiang). Related to this, a unique TFIID subunit, TAF8, is induced during adipocyte differentiation (Guermah et al., 2003). Other than the tissue-specific expression pattern observed with these TAF variants, functional inactivation of TAF1 (Hisatake et al., 1993; Ruppert et al., 1993), which is present in TFIID and PRC1, and of TAF 10 (Metzger et al., 1999), found in

TFTC, STAGA and SAGA-like complexes, causes cell cycle arrest, indicating a general role of selective mammalian TAFs in modulating cell growth. Undoubtedly, the presence of TAF variants further expands the general properties of TFIID to specialized needs in differentiated tissues.


Mediator, which represents the second class of general cofactors that transmit the regulatory signals from gene-specific transcription factors to the general transcription machinery, was first identified in yeast and found to consist of more than 20 polypeptides (Kim et al., 1994), of which 11 are essential for yeast viability (Rgrl, Rox3, Srb4, Srb6, Srb7, Med4, Med6, Med7, Med8, Medl0/Nut2, and Medll; Myers and Kornberg, 2000). While nine of the Mediator components were originally defined by genetic screens as proteins interacting with the CTD of pol II (i.e., Srb2 and Srb4-ll for different dominant suppressors of RNA polymerase B mutations; Myers and Kornberg, 2000), later biochemical purification of Mediator complexes from various species has identified additional conserved as well as species-specific subunits, for which a unified nomenclature has been proposed (Table 4.2) (Bourbon et al, 2004).

A: Isolation of Mediator Complexes

Human Mediator, first purified from HeLa cells as a protein complex that associates with the thyroid hormone receptor a (TR a) in a ligand-dependent manner, was able to potentiate TRa-mediated transcription in vitro (Fondell et al., 1996). This TR-associated protein complex (TRAP) contains many protein subunits subsequently found also present in other coactivator complexes, such as SRB/MED-containing cofactor complex (SMCC; Ito et al., 1999), vitamin D receptor-interacting protein complex (DRIP; Rachez et al., 1998), activator-recruited complex (ARC; Naar et al., 1999), positive cofactor 2 (PC2; Malik et al., 2000), cofactor required for Spl activation (CRSP; Ryu et al., 1999), and negative regulator of activated transcription (NAT; Sun et al., 1998). In humans, at least two forms of Mediator complexes, Mediator-P.5 and Mediator-P.85 isolated individually from 0.5 M and 0.85 M KC1 fractions of the Pll phosphocellulose ion-exchange column, have been identified and demonstrated to enhance activator-dependent and basal transcription, respectively (Wu et al., 2003). Mediator-P.5 represents a class of larger Mediator complexes, including also TRAP/SMCC and ARC/DRIP, which contain a dissociable

Table 4.2 Mammalian and yeast Mediator complexes.


S. cerevisiae



0 0

Post a comment