The Establishment, Maintenance and Epigenetic Inheritance of Silencing
Sir-based gene silencing is a molecular process by which the transcription of genes at specific locations within the genome are repressed. Yeast use this mechanism to repress extra copies of their mating type information and thus maintain their cell identity, and to regulate the expression of genes near telomeres. The molecular mechanism of Sir-based silencing in Saccharomyces cerevisiae is responsible for a level of repression that is perhaps more profound than any other known mechanism of transcriptional repression. The difference in expression level between the silenced and unsilenced state is at least 1000-fold and may be as much as two orders of magnitude larger. Nevertheless the mechanism of gene silencing has remained elusive. We established in the early 1990's that part of the mechanism involves a steric occlusion of DNA sequences from their recognition proteins, but more recent work indicates that there must be more than one mechanism to account for the strength of silencing. We are dissecting the mechanism of silencing by using a variety of molecular biology assays that interrogate the nature of a silenced locus and the contribution of individual promoters, activators, DNA sequence, and nucleosome position to silencing.
An outstanding conundrum in the biology of Sir2 and the sirtuin family in general is the role of the catalytic activity of Sir2. Sir2 is an NAD-dependent histone deacetylase that is thought to deacetylate lysine 16 on histone H4. This catalytic activity of Sir2 is necessary for silencing, even in cells that lack the acetyl mark that Sir2 is designed to remove. Clearly both the catalytic activity of Sir2 and a suitable substrate is required for silencing. A possible explanation for this paradoxical situation could be the required removal of an additional acetylated lysine, or possibly a small molecule by-product of Sir2's catalytic reaction. Interestingly, Sir2's deacetylation reaction converts the high-energy cofactor NAD+ into a small molecule, 2,3-O-acetyl-ADP-Ribose, and nicotinamide. 2,3-O-acetyl-ADP-Ribose has been implicated in silencing. We are investigating Sir2's catalytic activity through a forward genetics screen which has already turned up tantalizing new insights into the mechanisms of silencing.
Additionally, we have the chromatin architecture of Sir-based silencing. Previous methods of interrogating Sir enrichment across the silenced locus indicated that spreading of Sir proteins from recruitment sites at silencers across the silenced locus is necessary for silencing. Indeed, the model for the spreading of Sir proteins was paradigmatic for other "spreading" phenomena. However, new technologies have provided means to interrogate the chromatin architecture at higher resolution. By using comprehensive Chromatin Immunoprecipitation followed by high-throughput sequencing, a higher-resolution picture of Sir based silencing is revealing many nuances and intricacies of the chromatin architecture.
The great unsolved problem for most epigenetic phenomena is the mechanism by which a state of gene expression that is established at one time in a subset of otherwise genetically identical cells is then inherited efficiently by all the descendants of those cells. We have developed two new assays that should break this problem open. In one assay, we can detect all the descendants of those rare cells that spontaneously lose and then re-establish the silenced state. This assay has put mechanisms of establishing silencing within range of solution. The second assay provides a quantitatively rigorous test of the role of nucleosomes themselves and the proteins they carry in the inheritance of the silenced state.
Evo Devo in Fungi
There are many species that are considered to be model organisms due to various features that lend themselves to laboratory investigation of different aspects of biology. However, the recent completion of the genome sequences of many different species of the Saccharomyces genus has made this collection the first model genus for comparative genetic studies (D. Scannell, O. Zill et al, 2011). We have enjoyed great success in comparative studies of two different problems in S. cerevisiae and S. bayanus. We chose these two species for our focus as they are approximately as diverged as mouse and human, and hence could illuminate how different the solutions could be to the same problem in species this diverged.
Our initial investigations focused on the control of cell type and the mechanism of silencing in the two species. Both investigations lead to unanticipated insights that would have been impossible to discover in either species alone. In the case of cell type control, the studies in S. bayanus forced the first significant revision to the model for how cell type is controlled in 25 years (Zill and Rine 2007). In the case of silencing, the comparative studies have illustrated a new principle in the co-evolution of regulatory sites and the proteins that bind them (Zill et al, 2010), and uncovered a proliferation of the SIR1 gene family which jointly contribute to silencing in S. bayanus (Gallagher et al, 2009).
The question of how organisms adapt is fundamental to biology and is particularly relevant to issues of biomedical concern, such as the evolution of drug-resistant pathogens, or the progression of cancer. We are expanding these early studies into the molecular basis of adaptation using rapidly evolving genes in the Saccharomyces sensu stricto genus. We have systematically studied the evolutionary history of every gene in the genomes of the Saccharomyces clade, and have found the 100 most rapidly evolving genes in the clade. Among these we expect to find the points at which evolution is diversifying the biology of species in the clade. Specifically, we are investigating how both the physical and genetic interactions of rapidly evolving genes change across the sensu stricto species by looking at rapidly evolving proteins in the two most divergent species in this genus, S. cerevisiae and S. bayanus. By using physical and genetic interaction data from both of these species for several rapidly evolving genes, we hope to elucidate common patterns of evolution, enabling us to predict the most likely evolutionary paths in the sensu stricto genus and beyond.
The evolutionary history of many yeasts of the Saccharomyces complex, including S. cerevisiae, is marked by a critical whole-genome duplication. A direct result of this event was the duplication and subsequent diversification of many genes, two of which gave rise to SIR2 and SIR3, members of the SIR-silencing complex. Gene silencing in most species is accomplished by mechanisms involving RNAi, whereas Saccharomyces accomplishes gene silencing through formation of heterochromatin at specific loci using the Sir proteins. We are using Torulaspora delbrueckii as a pre-whole genome duplication model to understand how silencing evolved since this species has both the Sir proteins and the RNAi pathway's Argonaut and Dicer. T. delbrueckii may serve as the perfect model to reveal potential connections between RNAi, which is a more evolutionarily conserved silencing mechanism, and silencing mediated by the Sir proteins.
Chromatin: The Bigger Picture
Given the importance of nucleosome architecture to the regulation of gene expression, we also study proteins which modulate the structure and occupancy of nucleosomes, such as the SWR1 Complex and histone chaperones. Active areas of research include the structural and biochemical studies of H3/H4 chaperone Rtt106, the isolation of the AAA-ATPase Yta7 as a novel post-transcriptional regulator of histone H3, and the functional intersection of these two chromatin regulators.
Among chaperones, Rtt106 is particularly interesting as it is responsible for introducing newly synthesized H3-H4 dimers carrying an acetylated lysine 56 on histone H3, which is primarily an S-phase mark. A recent collaboration with the lab of James Berger involved a structure function analysis of the double pleckstrin homology domain and showing that the role of this chaperone in replication and silencing are inseparable, as well as mapping the determinants of Rtt106's preference for H3 carrying the K56 acetylation. Our more recent work involves a study of how histone proteins control their own synthesis through the function of histone chaperones at the promoters of histone genes (R. Zunder, A. Antzak, J.M. Berger and J. Rine, 2011 PNAS In Press).
Transcription, and especially high-level transcription, places special challenges on chromatin structure, requiring the movement of nucleosomes from in front of RNA polymerase to behind RNA polymerase. We have found that the Yta7 AAA ATPase of yeast, previously found to function to create chromatin boundaries, also plays a critical role in controlling nucleosome density, and through that effect, levels of gene expression (L. Lombardi, A. Ellahi and J. Rine, PNAS 2011 in press). More recent work is focusing on the post-transcriptional regulation of histone levels by Yta7, and a collaboration between Yta7 and Rtt106 in histone gene regulation.
Nutrition and Heritable Effects on Gene Expression
We are interested in how variation in the environment - in particular, nutritional deficiency - can affect chromatin state. Research has shown that variation in chromatin state can be inherited across cell and organismal generations, suggesting that adverse environmental conditions could affect an individual and its offspring even after the stress has been lifted. So far, we have focused on the effect of folate deficiency on histone methylation in S. cerevisiae and human cells in culture. Folate is required for the synthesis of s-adenosyl-methionine (SAM), the methyl donor for all histone methylation reactions. In yeast cells, limiting folate can result in the reduction in methylation of histone H3 at two different positions. Genetic experiments have allowed us to conclude that this is a direct effect of limiting the methyl donor pool rather than an indirect effect of altered transcription. These studies have been complemented by biochemical and genetic studies that are revealing new dimensions to the complexity of methionine regulation.
In related studies we are looking at the nutritional impact of folate limitation on the epigenetic landscape of the human genome. Unlike the genome whose information remains which remains static through generations, the epigenome is dynamic and can accumulate changes over time as a function of the environment, causing even individuals with identical genome to exhibit diverse phenotypes. In humans, diet is potentially one of the most influential environmental factors in epigenetic changes. Moreover, people with mutations in genes affecting metabolism could in principle be more epigenetically susceptible to their diet. In fact, nine percent of humans are homozygous for a mutation in the folate metabolic pathway that has been implicated in a variety of diseases including cancer and cardiovascular disease. We hypothesize that these people are potentially epigenetically labile. We are evaluating the human genome in cells from people with or without this mutation for its epigenetic lability as a function of folate and methionine, to determine whether most epigenetic marks are equally stable in response to changes in the environment, or whether some are more labile than others. Our results may provide insights into how genetic variation and nutrition can, in combination, affect the human epigenetic status and offer the potential to address disease states with an epigenetic component.
Human Nutritional Genetics
Humans differ from each other at the single nucleotide level at about one out of every thousand base pairs, which amounts to about 3 million positions in the genome. Given that humans have on the order of 640 different enzymes that require a vitamin or a mineral cofactor, it stands to reason that each of us is likely to have one or more vitamin-dependent enzymes that differ in the level of vitamin or mineral they need for optimum function. Using yeast as a surrogate host and assay for the function of human cDNA clones of genes for metabolic enzymes, we have already characterized well over 100 human SNPs that affect the function of the gene in which they reside, and a surprisingly large fraction of them have function restored when the level of vitamins in the medium is raised. In principle, this result points to the possibility of understanding our own individual nutritional requirements for optimum performance which would be a foundation of a truly personalized medicine.
Nutritional Genetic Variation and Birth Defects
There is a direct link between folic acid and the incidence of spina bifida, one of the most common and devastating birth defects in humans affecting between 1/1000 and 3.5 /1000 live births, depending upon the population. Folate supplementation of all food grains was mandated in 1998 in the US based upon the well-demonstrated ability of folic acid supplementation to reduce the frequency of children born with spina bifida.
We reasoned that spina bifida may be caused in part by genetic variation in one or more of the approximately 30 genes involved in folate metabolism. Therefore we performed a comprehensive analysis of the genetic variation in these genes in hundreds of children born with spina bifida and in case-matched controls. We recently published the results of this study, the largest of its kind, which identifies genetic signatures in these genes made of up collections of mutations that have a large effect on the risk of spina bifida (N. Marini et al, 2011 PLOS ONE). We are continuing to refine the risk mechanism, as well as extending these studies into the genetic basis of cleft palate, another common birth defect whose frequency also responds to folate levels.
updated November, 2011