UC Berkeley

Single-celled microbes have been the darlings of scientific study for many decades. The facility with which they can be grown in large batch culture has secured their historic popularity in biochemical and molecular biology studies. With ease, a researcher can harvest 10⁷ yeast cells, grind them into a pulp, and survey their collective mRNA content or protein populations. We have learned much from this practice as any cursory glance through a textbook on modern biology will attest. But what have we missed?

In this dissertation, I describe my single-cell approach to understanding gene expression changes in the yeast Saccharomyces cerevisiae. Gene expression in this organism is influenced by a number of activating and repressive processes, one of which is transcriptional silencing, a regional form of repression. Silencing, though known to be constitutive in S. cereivsiae , can be induced into a facultative state thereby serving as a valuable model system for the more dynamic heterochromatin of higher eukaryotes. For my dissertation, I took the yeast out of the test tube, out of large populations, and studied the process of silencing establishment at the individual cell level.

To better understand single-cell dynamics of silencing establishment, I developed a phenotype-based assay of functional silencing in individual cells. This technique, called the pedigree assay, was used to measure the speed of silencing establishment resulting in a phenotypic change – an alteration of cell identity, or mating-type. In this context, I discovered that silencing can occur within just two cell divisions, a much shorter timeline than previously inferred from batch culture studies. In addition, I noted variation among individuals and discovered that a cell's history influences silencing kinetics. That is, daughter cells are slightly more apt to establish silencing prior to their mother. In addition, I found that cells lacking specific histone methyltransferase enzymes are expeditious in silencing establishment compared to wild-type cells. This finding bolstered the hypothesis that removal of the histone methylation marks associated with euchromatin is a fundamental step in the process of silencing establishment.

To better understand the mechanism by which a key methyltransferase, Dot1, impacts the rate of silencing establishment, I developed a second, complementary assay for surveying expression dynamics in single cells using a destabilized green fluorescent protein (GFP) marker housed at a silencing-sensitive locus. By measuring the GFP fluorescence intensity of cells as they established silencing, I tested two hypotheses explaining the mechanism by which Dot1 may antagonize silencing establishment. Dot1 is known to methylate histone H3 lysine 79 (H3 K79), a mark associated with euchromatin that is thought to prevent or reduce Sir protein binding within transcribed genes. However, recently Dot1 protein was also shown to compete with a critical silencing protein, Sir3, for a binding site on histone H4. Therefore, Dot1 may impact silent chromatin formation via two routes. Using a series of dot1 mutants and histone mutants, I found that Dot1's impact on one silencing-sensitive locus was solely dependent on H3 K79 methyl status and was independent of the overall concentrations of the protein itself. This result is important because it illustrated how Dot1impacted silent chromatin formation and also implies a difference between Dot1's effet at the mating-type locus (I measured) and at telomeres (previously reported).

An advantage of the GFP reporter system is that it can easily be coupled with automated microscopy techniques to constantly monitor gene expression and silencing establishment over time. By doing so, I noticed considerable variability in gene expression at silencing-sensitive loci under unsilenced conditions. I ruled out the possibility that such variability is cell-cycle-dependent. Therefore, it is likely that variability of expression at this locus is due to microenvironmental response and/or to stochasticity in transcription. I also confirmed that daughter cells are slightly more likely to establish silencing prior to mother cells. These two findings can serve as a starting point for future studies aimed at better understanding expression variation, mechanisms of silent chromatin replication, and the asymmetrical partitioning of chromatin between the mother cell and the developing bud.

Our current understanding of silencing is littered with players – proteins, small molecules, and post-translational modifications that influence silencing chromatin formation. However, our model is still incomplete as indicated by a long-standing conundrum in the field. It is known that the catalytically active member of the silencing complex, Sir2, is required to deacetylate a key residue on histone H4 (H4 K16). However, pre-emptive removal of this mark does not rescue silencing in a cell containing a catalytically inactive sir2 (sir2-345 ) despite the fact that sir2-345 can associate with other complex members. It is highly possible that Sir2 has other substrates or that a small molecule by-product of Sir2 catalysis is also required for silencing. I sought to discover missing pieces of our model using both targeted and unbiased approaches. Through this effort, I discovered that histone acetylation on H3 K56 catalyzed by Rtt109 plays a role in antagonizing silencing formation, particularly in cells that lack H4 K16 acetylation (due to sas2 Δ) and contain the sir2-345 mutation. Moreover, I organized a screen aimed at identifying novel genome-wide mutations capable of restoring silencing in the sir2-345 sas2 Δ background. This unbiased approach was successful in identifying a collection of mutants capable of recovering silencing in that background. The study of these mutants will continue in the lab, and it will be interesting to learn their causal mutations as their identities are likely to bridge our understanding of silencing establishment and Sir2 biology.