Genetic experiments have identified two structurally similar nucleosomal domains, SIN and LRS, required for transcriptional repression at genes regulated by the SWI/SNF chromatin remodeling complex or for heterochromatic gene silencing, respectively. Each of these domains consists of histone H3 and H4 L1 and L2 loops that form a DNA-binding surface at either superhelical location (SHL) +/-2.5 (LRS) or SHL +/-0.5 (SIN). Here we show that alterations in the LRS domain do not result in Sin(-) phenotypes, nor does disruption of the SIN domain lead to loss of ribosomal DNA heterochromatic gene silencing (Lrs(-) phenotype). Furthermore, whereas disruption of the SIN domain eliminates intramolecular folding of nucleosomal arrays in vitro, alterations in the LRS domain have no effect on chromatin folding in vitro. In contrast to these dissimilarities, we find that the SIN and LRS domains are both required for recruitment of Sir2p and Sir4p to telomeric and silent mating type loci, suggesting that both surfaces can contribute to heterochromatin formation. Our study shows that structurally similar nucleosomal surfaces provide distinct functionalities in vivo and in vitro.

Post-translational modifications of histones influence both chromatin structure and the binding and function of chromatin-associated proteins. A major limitation to understanding these effects has been the inability to construct nucleosomes in vitro that harbor homogeneous and site-specific histone modifications. Here, we describe a native peptide ligation strategy for generating nucleosomal arrays that can harbor a wide range of desired histone modifications. As a first test of this method, we engineered model nucleosomal arrays in which each histone H3 contains a phosphorylated serine at position 10 and performed kinetic analyses of Gcn5-dependent histone acetyltransferase activities. Recombinant Gcn5 shows increased histone acetyltransferase activity on nucleosomal arrays harboring phosphorylated H3 serine 10 and is consistent with peptide studies. However, in contrast to analyses using peptide substrates, we find that the histone acetyltransferase activity of the Gcn5-containing SAGA complex is not stimulated by H3 phosphorylation in the context of nucleosomal arrays. This difference between peptide and array substrates suggests that the ability to generate specifically modified nucleosomal arrays should provide a powerful tool for understanding the effects of post-translational histone modifications.