Investigating the contributions to cell function of the different Swi-Snf complex subunits in Saccharomyces cerevisiae

Abstract

The eukaryotic genome is packaged as a DNA-protein structure known as chromatin. The basic subunit of chromatin is the nucleosome, which contains two copies of each of the core histone proteins H2A, H2B, H3, and H4, around which is wrapped 146 - 147bp of DNA. This structure is generally repressive for processes such as gene transcription that require access to the DNA. However, chromatin is dynamic. ATP dependent chromatin remodelling complexes can act to open or close this structure to enable or deny access to the DNA as and when required. The first ATP-dependent chromatin remodelling complex discovered was the Swi-Snf complex found in the yeast, Saccharomyces cerevisiae. This multi-subunit complex is best characterized as an activator of transcription. However, the precise role of each of the Swi-Snf subunits in transcription is not well understood. The first goal of the study was to investigate the role of each Swi-Snf complex subunit in cell function by examining the phenotypes in various Swi-Snf subunit mutants. Specifically, I examined snf2, swi3, snf5, snf6, snf11, snf12, taf14, and swp82 deletion mutants in addition to a snf2K798A catalytically dead mutant. Unfortunately, I was unable to obtain a SWI1 mutant to include in my analysis. My repeated gene deletion attempts failed which I hypothesised was due to the swi1 deletion mutant being either too sick to recover or being inviable. Also, my attempt to create a Swi1p anchor-away strain was not successful. Considering there are reports of swi1 mutants in the literature, I would suggest that the ability to obtain a viable swi1 mutant is strain specific and was incompatible with the strain backgrounds I used. The phenotypic analysis revealed that some Swi-Snf subunit mutations significantly affected doubling time and growth, while others had a minimal impact. For example, the greatest growth defect was found in snf2 and snf5 mutant strains. The snf2 and snf2K798A mutants were not always similar and displayed many distinct phenotypes. III When cell morphology was examined, the data revealed that the snf6 mutant displayed an elongated cell morphology, while snf2, snf2K798A, and snf5 strains had a clumpy cell morphology. According to spot test results using different reagents and different temperatures in liquid and solid media, the mutants again showed differences in their response to DNA-damaging reagents and cell wall stress. Together, the phenotypic tests indicated that the different subunits can possess diverse and distinct activities in multiple cellular processes, such as controlling cell size and proliferation, as well as influencing the cellular response to DNA damage. RNA-Seq analysis was used to compare the transcriptomes of each mutant strain to wt and to each other. Since the Swi-Snf complex functions as a co-activator of gene transcription, I expected most genes in each mutant to be downregulated. Surprisingly, this was not the case. For example, snf2 mutants, which should completely cripple the remodelling activity of the complex, had a similar number of genes up-regulated as were downregulated. This suggests SNF2 plays an equal role in the negative regulation of transcription as it does in positively regulating gene transcription. Furthermore, the swi3 deletion mutant showed almost twice as many genes were upregulated suggesting the major role of SWI3 is in the repression of transcription. Thus, the data suggests that the Swi-Snf co-activator complex can be considered to equally be a co repressor of transcription. The transcriptome data also revealed that even among Swi-Snf subunits within the same module, there were large variations in the number of genes regulated by each subunit, as well as differences in how these functionally related subunits contribute to the transcription of these genes. For example, whereas one subunit might be required for activation of a gene, another subunit could play a role in repression of the same gene. Most strikingly, there was only a minimal overlap (57 genes) in the transcriptomes of snf2, swi3, snf6 and snf5 mutants which, between them, are subunits representative of each of the Swi-Snf submodules. Together, this suggests each subunit can exert widespread distinct regulatory effects upon positive and negative gene IV transcription when functioning either within the Swi-Snf complex or, most intriguingly, when functioning outwith the complex. Analysis of published global sites of occupancy across the genome for the various Swi Snf subunits revealed that each subunit occupied a distinctly different number of sites. Swi3p occupied the greatest number of sites compared to other subunits (980 sites), whereas Taf14p had the most unique occupancy profile in terms of its presence at sites other than protein coding genes. Interestingly, Snf2p was found located most frequently across gene coding regions compared to its occupancy at promoters. This was surprising, as the best characterized role for Swi-Snf has been in remodelling nucleosomes at gene promoters to enable transcription initiation. This latter data might suggest a primary role for Snf2p dependent chromatin remodelling by Swi-Snf is during transcription elongation. Importantly, there was only a small overlap in the number of sites occupied by all the Swi-Snf subunits analysed. This is in support of Swi Snf subunits being able to function independently from each other and outwith the confines of the Swi-Snf complex. Swi-Snf subunit occupancy data was also correlated with the lists of genes up and downregulated in each mutant. The goal was to determine the specific genes that are directly controlled by each Swi-Snf subunit, and to assess whether Swi-Snf subunits could be working independently from each other within or outwith the complex. The data revealed that the correlation between Swi-Snf subunit occupancy and genes shown to be influenced by the particular subunits was poor. This could indicate that the major impact upon transcription is indirect or works over long-range distances. Conversely, the results could be due to limitations of the ChIP technique in fully identifying the sites of protein occupancy, or the sites of occupancy of the subunits are not occupied under the glucose-grown conditions used in this study. Indeed, performing ChIP for the subunits under multiple growth and stress conditions might be needed to identify the full suite of sites occupied by Swi-Snf as its recruitment at certain target sites will be cell signal specific. V The second part of my project examined the interplay between the Swi3p subunit of the Swi-Snf co-activator complex and the Cyc8p subunit of the Tup1-Cyc8 co-repressor complex. This interaction was chosen for analysis because previous unpublished data, which was confirmed in this study, revealed that Cyc8p levels were undetectable by Western blotting in a swi3 deletion mutant. Importantly, since CYC8 transcription levels were unaffected in the swi3 mutant, this suggests that the impact of the SWI3 gene deletion upon Cyc8p levels was occurring at the level of the Cyc8 protein. Furthermore, the SWI3 gene was shown to be responsible for repressing almost twice as many genes as it activates, thus implicating Swi3p as having a large role in mediating gene repression. Firstly, to investigate this potential interplay between SWI3 and CYC8 the swi3 transcriptome was compared to that of the swi3 cyc8 double mutant. This was to test the prediction that the swi3 mutant, in which Cyc8p was potentially absent, would have a transcriptome similar to that in a swi3 cyc8 double mutant. However, the results showed that this was not the case. The data showed that CYC8 could still negatively influence transcription of genes in the swi3 mutant suggesting that despite the abundance of Cyc8p in this mutant being below the detection threshold of western blotting, Cyc8p was still present at a level that could influence gene transcription. Secondly, by comparing the transcriptomes of swi3, cyc8 and swi3 cyc8 mutants, the data revealed that gene repression by SWI3 could function both independently of CYC8, and synergistically with CYC8. These findings therefore showed that the Swi3p subunit significantly contributes to the repression of gene transcription via Cyc8p-dependent and independent mechanisms. Finally, co-immunoprecipitation was used to confirm a direct interaction between Swi3p and Cyc8p. The results also suggested that the Swi3p and Cyc8p interaction might be mediated by Swi3p and Cyc8p whilst these subunits are residing within their respective Swi-Snf and Tup1-Cyc8 complexes. VI Overall, these data suggest wide-ranging unique roles for the individual subunits of the Swi-Snf complex that may be functioning when these proteins are residing within or out with the complex. The data also implicates the Swi-Snf co-activator complex as having an almost equal role in mediating gene repression, and that Swi3p plays the greatest role in this negative regulation of transcription. Finally, I have revealed evidence of a direct physical interaction occurring between the Swi-Snf and Tup1-Cyc8 complexes mediated by the Swi3p and Cyc8p subunits. Together, these data oppose the generally accepted consideration of these complexes independently acting as predominant activators or repressors of transcription. Instead, the data might suggest these complexes should be considered as intimate partners which work together to ensure gene transcription is accurately and appropriately positively and negatively regulated in response to the changing environment.