UC Berkeley

Eukaryotic cells package designated regions of their DNA into a condensed structure referred to as heterochromatin. By doing so, they preserve integrity of both the genome and the epigenome, as heterochromatin represses recombination and transcription at loci where such activities would be undesirable. In Saccharomyces cerevisiae, heterochromatin represses transcription at the silent mating-type loci, HML and HMR. Although the heterochromatin structure is dynamic and subject to perturbations such as DNA replication, all analyses to date have classified HML and HMR as transcriptionally inert. Therefore, either the mechanism of silencing compensates for fluctuations in the heterochromatin structure and flawlessly reassembles each cell cycle, or there are transient losses of silencing that underlie a hidden, dynamic dimension to heterochromatic repression.

To test whether RNA polymerase ever gains access to the silent mating-type loci, I used Cre-loxP recombination technology to devise a genetic assay whereby transient transcription of a cre reporter at HML would trigger a permanent, recombination-based switch from RFP expression to GFP expression. By recording short-lived events of HML::cre expression with the Cre-Reported Altered States of Heterochromatin (CRASH) assay, I found that approximately 1/1000 cells lost silencing per cell division. Consistent with this observation, measurements I made by single-molecule RNA FISH indicated that transcription of HML was rare, yet detectable, and limited in the number of RNA molecules that could be synthesized before silencing was re-established. These approaches revealed dynamics of silencing that had escaped detection by all previous measurements.

I used the CRASH assay to identify several genetic and environmental factors that modify the dynamics of heterochromatic repression. Whereas Sir2 is the only member of the sirtuin family of NAD+-dependent deacetylases previously shown to have a role in silencing at HML, I uncovered roles for two additional sirtuins. I showed that one of these sirtuins, Hst3, helped stabilize silencing through the deacetylation of acetylated lysine 56 on histone H3. In addition, I identified effects of histone gene dosage, SIR gene dosage, ploidy, and various environmental conditions on the stability of silencing. Quantitative analyses of these phenotypes were streamlined through the development of software that detects and compares patterns of differential gene expression in yeast colonies.

The sensitivity of the CRASH assay also revealed that the stability of silencing at HML differed between the two mating types of haploid cells. This mating-type effect depended on the activity of the recombination enhancer, a DNA element centromere-proximal to HML that was previously described for its role in directing the pattern of mating-type switching. I showed that this locus acted in cis to destabilize silencing at HML in a mating-type specific manner. Thus, the recombination enhancer was moonlighting as a long-range regulator of gene expression, the first such element identified in yeast.