Oncogene (2007) 26, 5373–5384 & 2007 Nature Publishing Group All rights reserved 0950-9232/07 $30.00 www.nature.com/onc REVIEW MYST opportunities for growth control: yeast genes illuminate human cancer gene functions A Lafon, CS Chang, EM Scott, SJ Jacobson and L Pillus Section of Molecular Biology, Division of Biological Sciences, UCSD Moores Cancer Center, University of California, San Diego, La Jolla, CA, USA The MYST family of histone acetyltransferases (HATs) genes were identified with potential roles in chromatin- was initially defined by human genes with disease mediated gene control, even modest degrees of similarity connections and by yeast genes identified for their role to the acetyl-CoA-binding regions were viewed with in epigenetic transcriptional silencing. Since then, many special interest as candidate HATs. new MYST genes have been discovered through genetic Such interest was particularly the case when mutants and genomic approaches. Characterization of the com- of SAS2 (something about silencing) were discovered as plexes through which MYST proteins act, regions of enhancers of mutations in the epigenetic transcriptional the genome to which they are targeted and biological silencer factor Sir1 (Reifsnyder et al., 1996) and consequences when they are disrupted, all deepen the suppressors of defects in cis-regulatory sequences for connections of MYST proteins to development, growth silent chromatin (Reifsnyder et al., 1996; Ehrenhofer- control and human cancers. Many of the insights into Murray et al., 1997). The observation that SAS2 and the MYST family function have come from studies in model closely related SAS3 yeast genes shared similarity to organisms. Herein, we review functions of two of the acetyl-CoA-binding domains, and even more significant founding MYST genes, yeast SAS2 and SAS3, and the similarity to two human genes, MOZ (Borrow et al., essential yeast MYST ESA1. Analysis of these genes in 1996) and TIP60 (Kamine et al., 1996), quickly led to yeast has defined roles for MYST proteins in transcrip- the recognition that together these genes defined a tional activation and silencing, and chromatin-mediated broadly conserved family that was named MYST for its boundary formation. They have further roles in DNA founding members (MOZ-YBF2/SAS3-SAS2-TIP60) damage repair and nuclear integrity. The observation that (Borrow et al., 1996; Reifsnyder et al., 1996). ESA1 MYST protein complexes share subunits with other HATs, was later identified as the third yeast MYST gene (Smith histone deacetylases and other key nuclear proteins, et al., 1998a; Clarke et al., 1999) with diverse functions: many with connections to human cancers, strengthens the it is essential for viability, and plays roles in transcrip- idea that coordinating distinct chromatin modifications tional activation, silencing, cell cycle progression and is critical for regulation. DNA damage repair. Oncogene (2007) 26, 5373–5384; doi:10.1038/sj.onc.1210606 The MYST domain – defined as the most conserved region in MYST family proteins – comprises a region Keywords: chromatin; transcription; Saccharomyces that includes both the sequences surrounding the acetyl- cerevisiae; silencing CoA-binding domain and in most enzymes an atypical C2HC zinc finger. Whereas the sequence of the MYST core domain is well conserved in Sas3 and Sas2, Esa1 lacks three of the four amino acids for zinc binding. Overview However, the crystal structure of Esa1 revealed that this region forms a classical TFIIIA-type zinc finger fold in Among the many chromatin modifications known, the absence of zinc suggesting a structurally conserved histone acetylation received particularly prominent early core domain with other MYST HATs (Yan et al., 2000). attention with the molecular identification of HAT1 and Besides the MYST domain, Sas3 and Esa1 contain GCN5 as genes encoding histone acetyltransferases additional conserved regions that are only found in (HATs) (Kleff et al., 1995; Brownell et al., 1996; subsets of the MYST family proteins, described below. Parthun et al., 1996). Although these two HATs have The early link between human MOZ, a recurrent different substrate specificities, initial sequence compar- translocation partner in several acute leukemias, and the isons revealed significant yet restricted similarity in three yeast genes implicated in chromatin-mediated regions predicted to bind acetyl-CoA. When additional transcriptional silencing raised the possibility that fundamentally conserved mechanisms might inform basic understanding between silencing mechanisms and Correspondence: Dr L Pillus, Section of Molecular Biology, Division of Biological Sciences, UCSD Moores Cancer Center, University of human disease. As further genomic and genetic analyses California, San Diego, La Jolla, CA 92093-0347, USA. proceeded, it became clear that the MYST family was E-mail: [email protected] broadly conserved. As of early 2007, several hundred Yeast MYST HATs and human cancer A Lafon et al 5374 related genomic sequences sharing significant sequence Here, their structural and functional distinctions are similarity over an B150 amino acid domain encompass- briefly highlighted in three categories: (1) definitions ing the acetyl-CoA-binding site have been annotated as of histone substrate specificity À of particular interest MYST genes (http://www.ncbi.nlm.nih.gov/entrez/query. are disease-relevant non-histone substrates such as fcgi?db ¼ Protein). Among MYST proteins of particular p53 in humans (Sykes et al., 2006; Tang et al., 2006), note are the Drosophila MOF gene that is essential only in (2) the MYST multi-protein complex(es) that modulate males for dosage compensation (Hilfiker et al., 1997) and enzymatic activity and promote coregulation with other the human MOF, MORF, MOZ and HBO1 genes that chromatin remodeling complexes through shared sub- have been implicated in multiple aspects of human units and (3) the gene-specific versus global acetylation development and disease (Avvakumov and Coˆ te´ ,2007; patterns that are of critical importance for disease and Rea et al., 2007). assessing potential success and side effect of therapeutic Many insights into MYST gene functions came initi- agents. These aspects are considered in more detail in ally from studies in model organisms, including Droso- other recent reviews (Utley and Coˆ te´ , 2003; Yang, 2004; phila, Caenorhabditis elegans, mouse, and the yeasts Altaf et al., 2007). We go on to address yeast MYSTs in Saccharomyces pombe and Saccharomyces cerevisiae the context of human disease, illustrated by mutant (described elsewhere in this issue and reviewed in Yang, phenotypes and conserved complex subunits. Further, 2004). This progress has come from the open-ended we consider what has been learned from genomic and nature of genetic screens, targeted gene deletion tech- proteomic surveys in yeast and speculate on what may niques and biochemical approaches combined with be learned from future studies that will be relevant for complete genome sequences and resources. understanding human health and disease. S. cerevisiae is a particularly rapid and informative genetic system for dissecting human protein function. For example, two-hybrid screening identified an inter- SAS3: critical functional links to GCN5 and H3 action of the MYST protein HBO1 with progesterone modifications receptor (PR) (Georgiakaki et al., 2006), a member of the family of steroid receptors. This observation SAS3 is arguably the least well-understood MYST gene prompted further analysis in human cells, which showed in yeast. In large part, this is because of the lack of that HBO1 enhancement of PR-dependent transcription extreme phenotypes observed upon deletion; indeed, required steroid receptor coactivator SRC-1 (Georgia- SAS3 was originally reported to have only minor kaki et al., 2006). Steroid receptors and their coactiva- functions in silencing at mating-type loci (Reifsnyder tors are key regulators of gene expression (Lonard and et al., 1996). O’Malley, 2006) and participate in signaling pathways Sas3 contains extensive stretches of acidic aspartate susceptible to misregulation in cancer cells. Yeast assays and glutamate residues in its C-terminus region, also can be designed also to study mechanisms of coactivator found in the human MYST HATs MOZ and MORF. function. For example, phenotypes conferred upon Although the functional significance of this domain in expression of hSRC-3 (Brown et al., 2003), GRIP1 the human proteins remains unclear, the acidic domain and SRC-1 (Anafi et al., 2000) revealed a role for Gcn5 of Sas3 is required for its physical interaction with the HAT complex components. These and other ingenious essential transcriptional elongation factor Spt16 through yeast assays will lend further insight into mechanisms two-hybrid screening (John et al., 2000). This suggests a underlying human MYST HAT function. The clinical function of Sas3 in transcriptional regulation mediated implications of these observations can be further tested by RNA polymerase II. in yeast screens to identify chemicals that interact with Although much remains to be learned of the biological steroid hormones and coactivators to modulate their processes regulated by SAS3, the determinants of its transcriptional output (Nishikawa et al., 1999) and molecular function have been well characterized, correlate mutations in key cancer-associated proteins including the multi-protein complex containing Sas3 with mutations that are clinically relevant, as was and the histone residues it targets. Sas3