Final Report Summary - TXNMEMGAL1 (Single-cell dynamics of Gal1 transcriptional memory inheritance in Saccharomyces cerevisiae)
Mechanistically how transcriptional and “epigenetic” states are inherited through cellular divisions is currently only poorly understood. However to define the heritability of epigenetic states within a population of cells is difficult due to cell heterogeneity in the level and stability of the underlying mechanisms. Therefore novel single-cell approaches to observe these processes in live cells, combined with high-throughput screening to identify factors that are involved are important for understanding how chromatin states and epigenetic modulators mediate “epigenetic” memory.
Currently it appears that the epigenetic modification of chromatin could be a general mechanism for transcriptional memory, conserved from complex to simple eukaryotic organisms. We chose to study the budding yeast Saccharomyces cerevisiae Gal1 gene, which is involved in galactose metabolism, as a model to investigate transcriptional memory. Transcriptional memory of Gal1 in S. cerevisiae was originally identified by the observation that naïve yeast (never been exposed to galactose), have a slower Gal1 transcriptional response than yeast cells that have been previously exposed to galactose. Deletions of chromatin-remodeling enzymes and histone modifiers were shown to affect the rate of Gal1 reinduction, suggesting that Gal1 transcriptional memory may be propagated via chromatin modification and/or structure.
Previous approaches to studying transcriptional memory have been mostly limited to detection of Gal1 mRNA transcription from a population of cells. However, it is unknown how an individual yeast cell (or its progeny) responds to repeated exposure to stimuli, and more interestingly, the mechanism by which an individual pre-exposed mother yeast passes on transcriptional memory to her daughters and for how many generations this memory is propagated. Towards this end, we have implemented a state-of-the-art microfluidics approach to address dynamics of Saccharomyces cerevisiae Gal1 transcriptional memory and in particular epigenetic inheritance at the level of individual yeast cells across multiple generations. Our setup has allowed us to specifically observe maintenance of memory within individual cells and more intriguingly the inheritance of memory in their naïve progeny.
The objectives of the research project during the fellowship were to dissect Gal1 transcriptional memory and inheritance in single yeast cells, identify novel factors involved in Gal1 memory, perform mechanistic studies on candidates identified from the screen, and evaluate the role of transcriptional memory in cellular survival. Specifically, our proposal included the following: single-cell analysis and pedigree mapping of wild-type yeast and selected mutants of transcriptional memory determinants using cell-tracking microfluidics, construction of a library of yeast strains with single-gene knockouts of chromatin-associated factors carrying a Gal1 fluorescent reporter, development of a high-throughput screen for regulators of Gal1 transcriptional memory, followed by validation and in-depth analyses of the effects of these mutants on inheritance and variability of memory.
For single-cell analysis and pedigree mapping of wild-type yeast and mutants of transcriptional memory determinants, we were interested in dissecting the stochasticity and variability of transcriptional memory between individual yeast cells within a population by determining the percentage of cells that show transcriptional memory and the variability in the reinduction rate and final expression level of Gal1, determining the extent to which transcriptional memory is inherited, applying pedigree mapping to compare transcriptional memory to progeny, and finally comparing transcriptional memory in wild type and selected chromatin-modifying/remodeling mutants that have previously been described to play a role in transcriptional memory. For this we made two yeast reporter strains, in which Gal1 is either fused or replaced by a fluorescent GFP reporter. We collected Gal1 expression data via the fluorescence reporters for these two strains using a single cell-tracking microfluidics setup, exposing the cells to repeated conditions of Gal1 induction and repression by changing medium conditions from galactose media to glucose media repeatedly. Data analysis of these two strains comparing the rates of Gal1 expression during the initial induction and reinduction to determine whether they display memory first validated the method and further yielded more intriguing results. We also collected data for mutant yeast strains of previously identified determinants of Gal1 transcriptional memory, including ΔSet1 and ΔSwi2 in both of the Gal1 reporter strains. We then processed the data collected for these strains for image segmentation and pedigree analysis.
We proceeded to analyze the single-cell data for the stochasticity, variability, and heritability of Gal1 transcriptional memory in single cells and their progeny. Initially, we defined Gal1 memory in single cells by a cell that shows faster reinduction of Gal1 after an initial induction. Using single-cell trace data of the wild-type yeast strain, we observed memory for individual cells. In comparison, mutants of Gal1 and the chromatin remodeler Swi2 also had memory but was impaired. We separated both time of induction and intensity of induction as two separate parameters to define the rate of reinduction, with the idea that perhaps separate memory mechanisms affect each of these processes. The heritability of transcriptional memory was measured by lineage mapping and correlations of daughter-cell expression during reinduction with mother cells in the first and second induction. We saw a clear correlation between daughter cell expression levels with their mother cell in the wild-type strain. We found that the percentage of cells within a colony that express Gal1 differs among mutants, thus our single-cell microscopy experiments also led us to adapt a third parameter to identify memory – the timing by which 10% and 50% of cells in the population have begun to express Gal1. We then correlated these memory parameters in several combinations, for example between the time delay that the cells take to start expressing Gal1 and the final intensity of Gal1 expression during reinduction. In this example, in the wild-type strain, we observed a negative correlation for time delay with intensity, suggesting that these two parameters are related, but not completely dependent on each other. We are in the process of modeling these parameters.
Towards our goal of identifying novel factors involved in transcriptional memory, we applied a high-throughput microfluidics approach in combination with genome-wide screening to identify factors that are potentially implicated in the memory of transcriptional and/or chromatin signatures. For this, we constructed a library of yeast strains that each contain a reporter for Gal1 expression and a single gene knockout. We focused our studies on ~550 mutants from the yeast knockout collection that encompass all nonessential chromatin-related factors. These strains were mated with two reporter strains for Gal1 using Synthetic Genetic Array (SGA) methodology for high-throughput strain construction. This resulted in our comprehensive chromatin-associated factor yeast mutant library.
We used a high-throughput microfluidics setup to screen all ~1100 strains simultaneously during memory experiments. We were able to obtain single-cell information during memory experiments for each of the mutants in our library. With this single-cell data, we analyzed the time that it takes for 50% of the cells to actively express Gal1, and also the intensity of Gal1 expression throughout the media changes. We have also been able to address stochasticity and variability/noise questions in high throughput with our screen data, though we were not able to obtain lineage information due to lack of time resolution amongst other reasons. From these experiments, however, we were able to choose ~30 candidates to further test in our cell-tracking microfluidics setup. This allowed us to extend the project profile to also analyze mutants for general Gal1 induction at the single-cell level, and specifically at the effect on the inheritance in memory in these mutant candidates, amongst several other interesting avenues.
Currently we are studying the mechanistic contribution of several factors identified in the screen to transcriptional and epigenetic memory with transcription assays and classic chromatin techniques. We were able to show that indeed some of these mutants affect the reinduction of Gal1 at the transcript level, and through chromatin biochemistry, we have unraveled the mechanistic basis behind these regulators. Finally, our single cell-tracking microfluidics setup has allowed us to probe the specific effects on the inheritability of these states. With these data, we hope to provide a comprehensive model of Gal1 transcriptional memory.
Currently it appears that the epigenetic modification of chromatin could be a general mechanism for transcriptional memory, conserved from complex to simple eukaryotic organisms. We chose to study the budding yeast Saccharomyces cerevisiae Gal1 gene, which is involved in galactose metabolism, as a model to investigate transcriptional memory. Transcriptional memory of Gal1 in S. cerevisiae was originally identified by the observation that naïve yeast (never been exposed to galactose), have a slower Gal1 transcriptional response than yeast cells that have been previously exposed to galactose. Deletions of chromatin-remodeling enzymes and histone modifiers were shown to affect the rate of Gal1 reinduction, suggesting that Gal1 transcriptional memory may be propagated via chromatin modification and/or structure.
Previous approaches to studying transcriptional memory have been mostly limited to detection of Gal1 mRNA transcription from a population of cells. However, it is unknown how an individual yeast cell (or its progeny) responds to repeated exposure to stimuli, and more interestingly, the mechanism by which an individual pre-exposed mother yeast passes on transcriptional memory to her daughters and for how many generations this memory is propagated. Towards this end, we have implemented a state-of-the-art microfluidics approach to address dynamics of Saccharomyces cerevisiae Gal1 transcriptional memory and in particular epigenetic inheritance at the level of individual yeast cells across multiple generations. Our setup has allowed us to specifically observe maintenance of memory within individual cells and more intriguingly the inheritance of memory in their naïve progeny.
The objectives of the research project during the fellowship were to dissect Gal1 transcriptional memory and inheritance in single yeast cells, identify novel factors involved in Gal1 memory, perform mechanistic studies on candidates identified from the screen, and evaluate the role of transcriptional memory in cellular survival. Specifically, our proposal included the following: single-cell analysis and pedigree mapping of wild-type yeast and selected mutants of transcriptional memory determinants using cell-tracking microfluidics, construction of a library of yeast strains with single-gene knockouts of chromatin-associated factors carrying a Gal1 fluorescent reporter, development of a high-throughput screen for regulators of Gal1 transcriptional memory, followed by validation and in-depth analyses of the effects of these mutants on inheritance and variability of memory.
For single-cell analysis and pedigree mapping of wild-type yeast and mutants of transcriptional memory determinants, we were interested in dissecting the stochasticity and variability of transcriptional memory between individual yeast cells within a population by determining the percentage of cells that show transcriptional memory and the variability in the reinduction rate and final expression level of Gal1, determining the extent to which transcriptional memory is inherited, applying pedigree mapping to compare transcriptional memory to progeny, and finally comparing transcriptional memory in wild type and selected chromatin-modifying/remodeling mutants that have previously been described to play a role in transcriptional memory. For this we made two yeast reporter strains, in which Gal1 is either fused or replaced by a fluorescent GFP reporter. We collected Gal1 expression data via the fluorescence reporters for these two strains using a single cell-tracking microfluidics setup, exposing the cells to repeated conditions of Gal1 induction and repression by changing medium conditions from galactose media to glucose media repeatedly. Data analysis of these two strains comparing the rates of Gal1 expression during the initial induction and reinduction to determine whether they display memory first validated the method and further yielded more intriguing results. We also collected data for mutant yeast strains of previously identified determinants of Gal1 transcriptional memory, including ΔSet1 and ΔSwi2 in both of the Gal1 reporter strains. We then processed the data collected for these strains for image segmentation and pedigree analysis.
We proceeded to analyze the single-cell data for the stochasticity, variability, and heritability of Gal1 transcriptional memory in single cells and their progeny. Initially, we defined Gal1 memory in single cells by a cell that shows faster reinduction of Gal1 after an initial induction. Using single-cell trace data of the wild-type yeast strain, we observed memory for individual cells. In comparison, mutants of Gal1 and the chromatin remodeler Swi2 also had memory but was impaired. We separated both time of induction and intensity of induction as two separate parameters to define the rate of reinduction, with the idea that perhaps separate memory mechanisms affect each of these processes. The heritability of transcriptional memory was measured by lineage mapping and correlations of daughter-cell expression during reinduction with mother cells in the first and second induction. We saw a clear correlation between daughter cell expression levels with their mother cell in the wild-type strain. We found that the percentage of cells within a colony that express Gal1 differs among mutants, thus our single-cell microscopy experiments also led us to adapt a third parameter to identify memory – the timing by which 10% and 50% of cells in the population have begun to express Gal1. We then correlated these memory parameters in several combinations, for example between the time delay that the cells take to start expressing Gal1 and the final intensity of Gal1 expression during reinduction. In this example, in the wild-type strain, we observed a negative correlation for time delay with intensity, suggesting that these two parameters are related, but not completely dependent on each other. We are in the process of modeling these parameters.
Towards our goal of identifying novel factors involved in transcriptional memory, we applied a high-throughput microfluidics approach in combination with genome-wide screening to identify factors that are potentially implicated in the memory of transcriptional and/or chromatin signatures. For this, we constructed a library of yeast strains that each contain a reporter for Gal1 expression and a single gene knockout. We focused our studies on ~550 mutants from the yeast knockout collection that encompass all nonessential chromatin-related factors. These strains were mated with two reporter strains for Gal1 using Synthetic Genetic Array (SGA) methodology for high-throughput strain construction. This resulted in our comprehensive chromatin-associated factor yeast mutant library.
We used a high-throughput microfluidics setup to screen all ~1100 strains simultaneously during memory experiments. We were able to obtain single-cell information during memory experiments for each of the mutants in our library. With this single-cell data, we analyzed the time that it takes for 50% of the cells to actively express Gal1, and also the intensity of Gal1 expression throughout the media changes. We have also been able to address stochasticity and variability/noise questions in high throughput with our screen data, though we were not able to obtain lineage information due to lack of time resolution amongst other reasons. From these experiments, however, we were able to choose ~30 candidates to further test in our cell-tracking microfluidics setup. This allowed us to extend the project profile to also analyze mutants for general Gal1 induction at the single-cell level, and specifically at the effect on the inheritance in memory in these mutant candidates, amongst several other interesting avenues.
Currently we are studying the mechanistic contribution of several factors identified in the screen to transcriptional and epigenetic memory with transcription assays and classic chromatin techniques. We were able to show that indeed some of these mutants affect the reinduction of Gal1 at the transcript level, and through chromatin biochemistry, we have unraveled the mechanistic basis behind these regulators. Finally, our single cell-tracking microfluidics setup has allowed us to probe the specific effects on the inheritability of these states. With these data, we hope to provide a comprehensive model of Gal1 transcriptional memory.