Toward a molecular understanding of yeast silent chromatin : roles for H4K16 acetylation and the Sir3 C-terminus

Discrete regions of the eukaryotic genome assume a heritable chromatin structure that is refractory to gene expression. In budding yeast, silent chromatin is characterized by the loading of the Silent Information Regulatory (Sir) proteins (Sir2, Sir3 and Sir4) onto unmodified nucleosomes. This requires the deacetylase activity of Sir2, extensive contacts between Sir3 and the nucleosome, as well as interactions between Sir proteins forming the Sir2-3-4 complex. During my PhD thesis I sought to advance our understanding of these phenomena from a molecular perspective.
Previous studies of Sir-chromatin interactions made use of histone peptides and recombinant Sir protein fragments. This gave us an idea of possible interactions, but could not elucidate the role of histone modifications in the assembly of silent chromatin. This required that we examine nucleosomal arrays exposed to full length Sir proteins or the holo Sir complex. In Chapter 2, I made use of an in vitro reconstitution system, that allows the loading of Sir proteins (Sir3, Sir2-4 or Sir2-3-4) onto arrays of regularly spaced nucleosomes, to examine the impact of specific histone modifications (methylation of H3K79, acetylation of H3K56 and H4K16) on Sir protein binding and linker DNA accessibility. The “active” H4K16ac mark is thought to limit the loading of the Sir proteins to silent domain thus favoring the formation of silent regions indirectly by increasing Sir concentration locally. Strikingly, I found that the Sir2-4 subcomplex, unlike Sir3, has a slight higher affinity for H4K16ac-containing chromatin in vitro, consistent with H4K16ac being a substrate for Sir2. In addition the NAD-dependent deacetylation of H4K16ac promotes the binding of the holo Sir complex to chromatin beyond generating hypoacetylated histone tails. We conclude that the Sir2-dependent turnover of the “active” H4K16ac mark directly helps to seed repression.
The tight association of the holo Sir complex within silent domains relies on the ability of Sir3 to bind unmodified nucleosomes. In addition, Sir3 dimerization is thought to reinforce and propagate silent domains. However, no Sir3 mutants that fail to dimerize were characterized to date. It was unclear which domain of Sir3 mediates dimerization in vivo. In Chapter 3, we present the X-ray crystal structure of the Sir3 extreme C-terminus (aa 840-978), which folds into a variant winged helix-turn-helix (Sir3 wH) and forms a stable homodimer through a large hydrophobic interface. Loss of wH homodimerization impairs holo Sir3 dimerization in vitro showing that the Sir3 wH module is key to Sir3-Sir3 interaction. Homodimerization mediated by the wH domain can be fully recapitulated by an unrelated bacterial homodimerization domain and is essential for stable association of the Sir2-3-4 complex with chromatin and the formation of silent chromatin in vivo.