Research in the Unit on Chromatin and Transcription is devoted to understanding the mechanisms by which entire regions of the genome are rendered inaccessible to transcription and recombination. Using the yeast Saccharomyces cerevisiae as a model organism, genetic analysis is coupled with biochemical fractionation and reconstitution experiments to explore the issues of genome accessibility. Silencing of genomic domains requires a complex series of interactions between inactivation centers called silencers and numerous repressor proteins. The silencers recruit repressor protein complexes composed of the Sir proteins that interact with histones in nucleosomes to form a chromatin domain that is inaccessible and inert to various cellular processes. We focus on the Sir proteins and their interactions with the histones to understand, in molecular detail, the mechanism by which these proteins effect silencing.

Histone Variants and Silencing
Dhillon, Kelman
It has been previously demonstrated that RAP1, ORC, and SIR1 are necessary for the establishment of the silent state. In the simplest model, Sir1p acts as an intermediary molecule between the proteins bound to the silencer element (Rap1p and Orc) and the proteins involved in maintaining the repressed state, such as the remaining Sir proteins and histones. One of the events during the establishment of the silenced state is therefore most likely to be the recruitment of Sir1p to the silencer.

To understand this event, we generated mutants in the Sir1 proteins that were unable to silence even after Sir1 was recruited to the silencer. Following the characterization of these mutants, we performed a suppressor screen to isolate proteins that, when overexpressed, lead to suppression of the mating defect of Sir1p point mutations. The isolation of distinct point mutations in Sir1p suggests a weakening of its interactions with other proteins involved in silencing. Overproducing such proteins in the sir1 mutant should compensate for the weakened interaction. After screening multicopy plasmid libraries in the Sir1p mutant for genes whose elevated dosage overproduces a protein that restores silencing, we isolated and identified four suppressors. Given that we are mainly interested in novel factors involved in establishment and inheritance, the laboratory further characterized two of the newly identified genes (Dhillon and Kamakaka, 2000). One of these genes, Esc2p, appears to function at the silencer and requires Sir1 for function while the second gene is a histone variant Htz1p that is required for the maintenance of silencing.

The role of histone variants in gene expression is poorly understood, and our results provide a unique opportunity to understand fully the role of these variants in gene expression. Since the protein Htz1 has been shown to affect both gene activation and repression, we are interested in understanding precisely how it functions. Classical molecular genetic and biochemical experiments are in progress to ascertain the domains of Htz1 involved in activation and repression as well as to determine the proteins that interact with Htz1 to function in the cell. The availability of the entire DNA sequence of S. cerevisiae also allows us to investigate the distribution of Htz1 throughout the genome and, using a novel chromatin immunoprecipitation scheme coupled with genomic microarrays, to determine the distribution changes under different growth conditions.

Identification and Characterization of Sir Protein Complexes Involved in Silencing
Ghidelli
Another focus of our research is to determine how the Sir proteins function in Sir-mediated silencing. Genetic studies have revealed that distinct combinations of the Sir protein complexes transcriptionally repress multiple loci. Silencing at all of these loci requires Sir2p, which has recently been shown to possess enzymatic activities. In addition, Sir2p is the only Sir protein to have additional homologs in yeast (Hst1-4p) and the only Sir protein conserved throughout evolution. We have started to purify Sir2p-containing protein complexes from yeast cells and to identify and characterize the individual components within these complexes and their associated enzymatic activities. Using these purified complexes and histones in nucleosomes, we have also begun studies on the reconstitution of silenced chromatin (Ghidelli et al., 2001). If it is biochemically possible to reconstitute the silenced state, our studies will provide an important index of our current understanding of transcriptional silencing, since mechanisms are rarely established by genetic means alone.

Further studies will aim at determining the molecular structure of the complexes and their associated enzymatic activities. We will use reverse genetics to determine the function of any novel polypeptides associated with the complexes. Additional experiments will analyze the nature of the interactions between Sir protein complexes and histones in nucleosomes. The analysis will involve binding studies with positioned nucleosomes and purified Sir complexes followed by DNaseI foot-printing analysis as well as protein-protein cross-linking and sedimentation analysis. Long-term goals will include studies on the regulation of these enzymes within the cell and in vitro studies aimed at the eventual development of specific inhibitors for enzymes that may have therapeutic value.

Role of Nuclear Architecture in Silencing
Donze, Oki, Kamakaka
We are also interested in understanding the mechanism by which the silenced chromatin domains are restricted to specific regions along the DNA fiber. Eukaryotic chromosomes are organized into discrete domains delimited by domain boundaries. The boundaries define units of gene activity by insulating a locus from the influence of flanking chromatin. The observation that the silenced chromatin domain at HMR extends beyond the two silencers begged the question of the mechanism by which the silenced chromatin domain is restricted to a specific region of the genome. We have demonstrated that a specific tRNA gene mediates barrier functions (Donze et al., 1999). The proteins that are required to prevent the spread of heterochromatin into neighboring euchromatin have also been identified (Donze and Kamakaka, 2000). We have demonstrated that specific elements act as barriers to the continuous spread of the silenced chromatin, and our results suggest that barrier activity may arise from an underlying competition between chromatin remodeling and silencing activities at the interface of euchromatin and heterochromatin. Further studies call for genetic screens to isolate, first, other DNA elements from yeast and other eukaryotes that act as barrier elements and, second, additional proteins from yeast and other eukaryotes that possess barrier activity. The studies will help delineate general principles of barrier activity in all eukaryotes.