Misregulation of transcription is a hallmark of a number of human diseases ranging from diverse classes of cancers to neurodegenerative diseases. Within the organism, the cell uses a number of cues to properly coordinate transcriptional events, including the recruitment of transcriptional coactivator complexes (reviewed in ). Many transcriptional coactivator complexes possess intrinsic enzymatic activities that allow for access to DNA for transcription. Chromatin remodeling complexes, such as ATPases, are a type of transcriptional coactivator complex that move nucleosomes, while histone acetyltransferases and histone methyltransferases are examples of complexes that modify histones post-translationally. The importance of these chromatin-modifying complexes is demonstrated by the fact that they are often conserved from yeast to humans (reviewed in [1–3]).
In yeast, one of the most well-characterized processes in regulating chromatin structure is the post-translational modification of histones. Originally identified over 40 years ago, histone acetylation has now become a model for understanding the role of histone modifications in modulating chromatin structure . To this end, the histone acetyltransferase Gcn5 and its associated complexes have been shown to regulate the transcription of up to 10% of genes in yeast . With the advent of high-throughput mass spectrometry, the array of proteins associated with Gcn5 continues to increase [2, 6–10]. There are currently at least four Gcn5 complexes: SAGA, SLiK(SALSA), ADA and HAT-A2 . SAGA, the best characterized of these, has 18 protein subunits with distinct functions organized in a modular fashion [1, 11, 12]. In addition to acetyltransferase activity, a second enzymatic activity has been attributed to SAGA involving the regulation of histone H2B ubiquitination [8, 9]. Specifically, the SAGA subunit Ubp8, a deubiquitinating enzyme, has been shown to deubiquitinate H2B, thus allowing for properly regulated transcription at a number of genes. Accentuating the crucial role of modularity within in the SAGA complex, Ubp8 has been shown to require another protein, Sgf11, for its activity . Additionally, Ubp8 and Sgf11 together regulate a subset of genes, and so have been revealed to function as a module in yeast . More recent work has demonstrated that another SAGA subunit, Sus1, is also important for maintaining proper H2B ubiquitination, and that, in fact, it is these three proteins that constitute the deubiquitination module of SAGA and the related SLiK(SALSA) complex .
During our study of SAGA we asked how this deubiquitination module is held within the complex. As the initial three-dimensional rendering of SAGA did not contain these proteins, we looked at other components of SAGA. We found that Sgf11 shared a conserved domain with the Sgf73 protein of SAGA, which is implicated in neurodegeneration in higher eukaryotes . Specifically, we discovered that both Sgf11 and Sgf73 contain a conserved non-canonical zinc finger. This zinc finger is termed the ataxin box as it is conserved in the known ataxin proteins . We therefore hypothesized that Sgf73 is related to Sgf11 in function, and found that, indeed, loss of Sgf73 resulted in an increase in global histone ubiquitination. During our studies on Sgf73, the Hurt group also reported the role this Ataxin-7 orthologue has in maintaining proper histone ubiquitination levels in yeast, as well as playing a role in gene gating and mRNA export by recruiting other factors to SAGA .
Interestingly, the SGF73 gene has been shown to be the Ataxin-7 homologue. Mutations in Ataxin-7 result in trinucleotide expansions and Spinocerebellur Ataxia-7 (SCA7) [16, 18]. SCA7 is one of the many types of spinocerebellar ataxia (SCA), including SCA types 1–3, 6–10, 12, and 17, caused by a genetic defect that results in a trinucleotide repeat expansion in the DNA-coding sequence [19, 20]. The fact that SGF73 links a histone-modifying complex to a neurodegenerative disorder is intriguing as it was shown that disease levels of trinucleotide expansions of SCA7 (36–306 CAG expansions) cause defects in the HAT activity of the SAGA complex in vitro [18, 21, 22]. As diseases of the trinucleotide repeat variety, including Huntington's disease and the SCA disorders, generally exhibit defects in transcription, it is important to note that deletion of SGF73 in yeast prevents formation of the transcription pre-initiation complex at a number of genes, and inhibits SAGA recruitment .
Here we demonstrate, similar to the Hurt group, that both Sgf11 and Sgf73 are related to the ataxin proteins in higher eukaryotes . Focusing our study on the Gcn5-containing HAT complexes we also looked at the effect of losing Sgf73 on the SAGA-related complex, SliK(SALSA). We found that the deletion of SGF73 from SliK(SALSA) also liberates the deubiquitination module from the complex, rendering it non-functional; a similar loss is observed by us and the Hurt group for the SAGA complex . We also observe that Gcn5 HAT activity in SAGA is independent of both Sgf73 and deubiquitination activity. We definitively show that the deubiquitination module of Ubp8, Sgf11 and Sus1 is necessary, but not sufficient to carry out or maintain proper levels of histone H2B ubiquitination in yeast. Additionally, we show that deletion of SGF73 in combination with GCN5 results in similar defects as the deletion of UBP8 and GCN5, confirming the importance of the deubiquitination module as part of SAGA for proper transcriptional regulation. Interestingly, the Hurt group showed that in the presence of a fragment of Sgf73 that can co-purify Ubp8, Sgf11 and SusS1, they can observe activity using ubiquitin-AMC hydrolysis . Therefore our data along with the published work from the Hurt lab point to the crucial role that Sgf73 plays in anchoring the deubiquitination module into both SAGA and SliK(SALSA).