PolyQ Diseases as Transcriptionopathies

When localized in the nucleus, polyQ-expanded proteins aberrantly interact with a variety of transcription factors, many of which contain a polyQ or glutamine-rich domain (Table 2). Certain transcription pathways, namely those involving the cyclic AMP response element (CRE)-binding protein (CREB) and specificity protein-1 (Sp1) have been implicated in the pathogenesis of multiple polyQ diseases. Interestingly, the cofactor TBP-associated factor 4 (TAF4; formerly TAFII130), which was independently identified in a yeast two-hybrid screen for nuclear proteins that interact with polyQ tracts (Shimohata et al. 2000), mediates transcriptional activation by both CREB and Sp1 (Fig. 3, top). TAF4 is a component of the general transcription factor TFIID, a multi-subunit complex that comprises TBP and at least 12 TAFs (Muller and Tora 2004). Different glutamine-rich subdomains in TAF4 facilitate its interaction with Sp1 and CREB (Saluja et al. 1998). Although CREmediated transcription is constitutive at a subset of promoters (Conkright et al. 2003), recruitment of the cofactor CBP (or the related protein p300), which is contingent on the phosphorylation CREB at a single serine residue, is generally a prerequisite for transcriptional activation (Johannessen et al. 2004).

Various members of the CREB and Sp1 transcription pathways have been reported to interact with soluble and/or aggregated, mutant polyQ proteins (Table 2). As these two possibilities are not mutually exclusive, both forms of polyQ protein could contribute to transcriptional deregulation. Moreover, the sequestration of a given transcription factor in NII would have the same consequence as a soluble interaction of increased affinity. In either case, the transcription factor would be effectively titrated from its cognate promoter binding site (Schaffar et al. 2004; S.H. Li et al. 2002; Dunah et al. 2002) (Fig. 3, bottom). Consistently, reporter assays carried out in cellular models of certain polyQ diseases indicate that expanded polyQ antagonizes both CREmediated (Shimohata et al. 2000; Nucifora et al. 2001) and Sp1-dependent transcription (S.H. Li et al. 2002; Dunah et al. 2002). Overexpression of either TAF4 (Shimohata et al. 2000) or CBP (Nucifora et al. 2001) can rescue CRE-mediated transcription, while overexpression of both Sp1 and TAF4 is required to attenuate the effects of mutant htt on Sp1-dependent reporter activity (Dunah et al. 2002). Downregulation of CRE-mediated transcription has been corroborated by expression profiling in both cellular (Wyttenbach et al. 2001) and mouse (Luthi-Carter et al. 2000) models of HD. Unexpectedly, upregulation of the same transcriptional pathway was observed upon

Table 2 A survey of transcription factors that bind polyQ disease proteins

Transcription factor Interacting polyQ Colocalizes with Binds soluble References or cofactor disease protein polyQ aggregatesa polyQ protein(s)b













Spl htt V

AR, ataxin-1, ataxin-3, TBP, ataxin-7

Atrophin-1, ataxin-3

Ataxin-7 htt

Atrophin-1 htt htt V

htt htt, TBP

htt, TBP V

Ataxin-3, TBP Ataxin-3, htt htt, TBP

Ataxin-1 y/

atrophin-1, ataxin-3

Steffan et al. (2000), McCampbell et al. (2000), Chai et al. (2001), Stenoien et al. (2002), Swope et al. (1996), La Spada et al. (2001)

Shimohata et al. (2000)

Chen et al. (2004) Kegel et al. (2002) Wood et al. (2000) Faber et al. (1998) Boutell et al. (1999) Boutell et al. (1999)

Takano and Gusella (2002), Schmitz et al. (1995) Steffan et al. (2000), Truant et al. (1993) Li et al. (2002), Swope et al. (1996) Steffan et al. (2001), Li et al. (2002b) Zuccato et al. (2003), Murai et al. (2004) Tsai et al. (2004)

Shimohata et al. (2000), Dunah et al. (2002), S.H. Li et al. (2002), Emili A (1994)

Table 2 (continued)

Transcription factor

Interacting polyQ

Colocalizes with

Binds soluble


or cofactor

disease protein

polyQ aggregates a

polyQ protein(s)b


htt, atrophin-1,



Shimohata et al. (2000), Dunah et al. (2002)

ataxin-2, ataxin-3




Yvert et al. 2001


htt, atrophin-1, ataxin-3

Huang et al. (1998), Schaffar et al. (2004),

Shimohata et al. (2000)

TBP interacts with various transcription factors in the context of transcriptional initiation by all three nuclear RNA polymerases. It is presently unclear which of these myriad interactions may be relevant to polyQ-mediated pathogenesis (in SCA17) and most have been excluded from this table. An asterisk denotes repressor or corepressor activity. Bold font indicates the presence of a polyQ or glutamine-rich domain in the transcription factor.

AR androgen receptor, CA150 coactivator 150, CBP cyclic AMP response element binding protein, CtBP C-terminal binding protein, Crx cone-rod homeobox containing gene, htt huntingtin, HYP-B htt-yeast partner, mSin3a mammalian Sin3 protein-A, MTG8 myeloid translocation gene on 8q22, NCoR nuclear receptor corepressor; NF-kB nuclear factor-KB, P/CAF p300/CBP-associated factor, REST/NRSF repressor element-1 transcription factor/neuron restrictive silencer factor, SMRT silencing mediator of retinoid and thyroid hormone receptors, Spl specificity protein-1, TAF4 TBP-associated factor 4, TAF10 TBP-associated factor 10

a As determined by double immunolabeling and microscopy or biochemical analysis of aggregate content b As determined by yeast two-hybrid, in vitro binding, or coimmunoprecipitation c A component of TFIID as well as other transcriptional complexes; formerly known as TAFnl30 d A component of TFIID as well as other transcriptional complexes; formerly known as TAFn30

Fig. 3 Transcriptional dysregulation in the presence of nuclear, polyQ-expanded proteins. Under normal circumstances (top), transcriptional activators bind upstream promoter elements and interact (curved arrow) with various components of the RNA polymerase II (Pol II) preinitiation complex (PIC). These interactions, which are important for PIC recruitment to class II promoters, are often facilitated by polyQ or glutamine-rich domains that are present in many activator proteins and some general transcription factors. However, in the polyQ diseases (bottom), both the soluble and the aggregated versions of mutant polyQ can act as a sink for the same transcriptional activators, effectively titrating the latter from their cognate DNA binding sites (arrows). It should be noted that many changes in gene expression are disease-specific and may also involve aberrant interactions with transcriptional repressors or altered chromatin acetylation. Moreover, targeting of expanded polyQ to the core promoter, as occurs in SCA17 but in none of the other polyQ diseases, may have a unique transcriptional impact. IIA, IIB, IIE, IIF, and IIH represent TFIIA, TFIIB, TFIIE, TFIIF, and TFIIH, respectively. CRE cyclic AMP response element, CREB CRE binding protein, CTD carboxy-terminal domain of Pol II, P — phosphorylation, Sp1 specificity protein-1, TAF4 TATA-box binding protein associated factor 4

introduction of a CRE-regulated reporter transgene into an HD mouse model (R6/2) (Obrietan and Hoyt 2004). The basis of this discrepancy is currently unclear.

Microarray experiments, utilizing brain messenger RNAs (mRNAs) from various polyQ mouse models, have revealed some overlap in the expression changes induced by the different polyQ disease proteins (Sugars and Ru-binsztein 2003). Moreover, a comparison of cerebellar mRNAs derived from a DRPLA (At-65Q) and an HD (N171-82Q) mouse model demonstrated that the transcriptional impact of expanded polyQ is partly context independent (Luthi-Carter et al. 2002). Despite a difference in the length of the polyQ tract, the two models have an identical genetic background and the respective transgenes are driven by the same promoter. These data support the idea that particular transcription pathways may be disrupted in polyQ disease.

In addition to changes in activated transcription, there is some evidence that basal gene expression may also be dysregulated by expanded polyQ. Aberrant interactions involving TBP, which is required for transcription by all three nuclear RNA polymerases, have been documented in several of the polyQ diseases (Table 2). TBP has been reported to colocalize in HD (Schaf-far et al. 2004), DRPLA, and SCA3 aggregates (Shimohata et al. 2000), and the transcription factor also preferentially coimmunoprecipitates with soluble, mutant htt (Schaffar et al. 2004). Functional deactivation of TBP in the presence of mutant htt has been demonstrated in vitro (Schaffar et al. 2004); however, it is unlikely that a polyQ expansion in TBP, which is causative for SCA17, abrogates its function. Heterozygous TBP knockout mice are pheno-typically normal, but nullizygous embryos do not develop beyond the blas-tocyst stage (Martianov et al. 2002). Consistently, below a certain pathogenic repeat threshold, polyQ-expanded TBP upregulates a CRE-regulated reporter gene in a cellular model of SCA17 (Reid et al. 2003).

Although the relevance of general transcriptional repression to polyQ-mediated pathogenesis has not been firmly established, it is clear that histone acetylation is disrupted in the presence of mutant polyQ (Bodai et al. 2003). Three histone acetyltransferases (HATs), including CBP, p300, and p300/CBP-associated factor (P/CAF), interact directly with ataxin-3 (F. Li et al. 2002). htt exon 1 protein also binds CBP and P/CAF (Steffan et al. 2001). CBP physically interacts with the soluble and aggregated forms of various polyQ disease proteins (Table 2). However, since p300 contains only a short polyQ tract and P/CAF lacks this domain entirely, these interactions are not contingent on the association of independent polyQ domains. Rather, mutant htt has been reported to impair the HAT activity of CBP and P/CAF by binding to their acetyltransferase domains (Steffan et al. 2001). Hypoacetylation of his-tones H3 and H4, which has been documented in multiple polyQ disease models (McCampbell et al. 2001; Steffan et al. 2001), could potentially have widespread effects on transcription. It is well established that acetylation of particular histone residues is associated with euchromatin and active genes. Either local or global changes in acetylation could significantly impact the expression of genes that are important for cell function and viability. Interestingly, acetylation of two lysine residues on histone H3, namely, K9 and K14, may be necessary for the recruitment of TFIID to promoters (Agalioti et al. 2002). Accordingly, polyQ-induced histone deacetylation might antagonize this crucial step in preinitiation complex assembly on certain RNA polymerase II promoters. Alternatively, it is noteworthy that HATs also modu late the function of various non-histone proteins, like p53, by domain-specific acetylation (Guand and Roeder 1997).

Although the outcome is not surprising, it is not entirely clear how tran-scriptional dysregulation leads to neuronal cell death. Interestingly, mice that lack CREB in the postnatal forebrain, as a result of conditional disruption, display neurodegeneration in the hippocampus and striatum (Mantamadi-otis et al. 2002). Nevertheless, changes in gene expression can be detected in early symptomatic (Luthi-Carter et al. 2000) as well as presymptomatic (Lin et al. 2000; Serra et al. 2004) transgenic polyQ mouse models, long before any evidence of neurodegeneration has emerged. Thus, in many cases, neuronal dysfunction and not neuronal death may be responsible for the polyQ-induced phenotype (Hientz and Zoghbi 2000).

Was this article helpful?

0 0

Post a comment