Thesis (M.Phil.)--Hong Kong University of Science and Technology, 2007. / Includes bibliographical references (leaves 193-218). Also available in electronic version.

SWI/SNF, the founding member of ATP dependent chromatin remodelers and its paralog RSC in yeast perform similar yet distinct functions inside the cell. In vitro these complexes use ATP dependent DNA translocation to either mobilize or disassemble nucleosomes. However, how these complexes interact with nucleosomes and the mechanism by which chromatin remodeling is achieved is not fully understood. Further, it is not understood how they perform disparate roles in vivo despite their similar biochemical activities. To understand the fundamental differences between these complexes the substrate specificity of RSC and SWI/SNF and their interaction with different parts of the nucleosome were investigated. SWI/SNF and RSC exhibited almost identical nucleosome binding affinities (~7 nm) with a minimal requirement of 20 bp of extranucleosomal DNA for efficient binding. Hydroxical-radical footprinting of RSC-nucleosome complex showed that RSC, unlike SWI/SNF, interacts extensively with approximately 50 bp of extranucleosomal DNA near the nucleosome entry site. RSC also interacts, but not as strongly as SWI/SNF, with almost one gyre of nucleosomal DNA (SHL-2 to SHL-6) on the same side of the extranucleosomal DNA. Analogous to the previously observed SWI/SNF-footprint the second gyre of nucleosomal DNA had no protection and in fact enhanced cleavage was seen starting from 3-4 helical turns from the dyad axis up to the exit site where DNA leaves the nucleosome. The asymmetry of the DNA footprint pattern confirmed binding of RSC in one preferred orientation guided by the extranucleosomal DNA at one end of the nucleosome like ISW2 and also like SWI/SNF but only when recruited by transcription factors. DNA crosslinking revealed that most of the SWI/SNF contacts are with a small region spanning the DNA translocation start site near SHL2 and does not extend to the rest of the footprint. Further, the SWI/SNF contacts are primarily through its catalytic subunit Snf2 which is found to intercalate between the DNA gyre and the histone octamer at SHL2. Consistent with its DNA footprint, RSC however makes extensive contacts with both nucleosomal and extranucleosomal DNA through five major subunits Sth1, Rsc2, Rsc3, Rsc30 and Rsc4. Excepting the catalytic subunit Sth1 which is highly homologous to Snf2, the remaining four are unique to RSC. Sth1 contacts a much broader region in the nucleosomal DNA than Snf2 with the primary contact being at SHL2 where it wedges between the DNA and the histone octamer surface. The accessory subunits Rsc2, Rsc3 and Rsc30 mostly contribute to the extranucleosomal DNA contacts of RSC. These subunits also make a second major contact near the dyad, with those made by Rsc3 and Rsc30 being the strongest. The histone N-terminal tails that emanate out of the nucleosome structure are implicated in the regulation of chromatin remodeling, in general, and in the activation of several SWI/SNF dependent genes, in particular. Remodeling kinetics studies with tailless nucleosomes revealed that the histone H4 tail is required for nucleosome mobilization, H2A/H2B dimer displacement and nucleosome disassembly by both RSC and SWI/SNF. Further, the H4 tail modulates RSC and SWI/SNF remodeling without affecting ATP hydrolysis or nucleosome binding. These data suggest a similarity between SWI/SNF and ISWI class of chromatin remodelers based on their dependence on the H4 tail. Owing to the presence of acetyl-lysine binding bromodomains in these complexes and to a greater extent in RSC the differences in their remodeling activities, if any, were expected to be accentuated by histone acetylation. Studies with H3 and H4 tail acetylated nucleosomes provided evidence for two pathways that work synergistically to recruit SWI/SNF and RSC to chromatin. While one of the pathways involves transcription activators, the other pathway of SWI/SNF recruitment is dependent on covalent acetylation of histone H3 tail. Bromodomain mediated recognition of these acetyl marks not only facilitates SWI/SNF recruitment but also stimulates their catalytic activity to mobilize nucleosomes. Importantly, extensive conformational changes occur in SWI/SNF in response to H3 tail acetylation. Chromatin remodeling by SWI/SNF and RSC is also regulated to different degrees by H3 tail acetylation depending on the number of bromodomains. The higher responsiveness of RSC to H3 tail acetylation than SWI/SNF can provide additional regulatory mechanisms for RSC which might ultimately account for their different functional roles inside the cell. When these same acetyl marks are within the H3 globular core and reside near the dyad axis of symmetry they are found to act in synergy with RSC and SWI/SNF to facilitate nucleosome movement as well as nucleosome disassembly. Unlike H3 tail acetylation, the remodeling enhancement by H3 core acetylation occurs via an acetyl lysine-bromodomain recognition independent mechanism. Further, supporting this recognition-independent mechanism H3 core acetylation does not affect the recruitment of these complexes. These data illustrate how histone acetylation modulates RSC and SWI/SNF function, and provide a mechanistic insight into their collaborative efforts to remodel chromatin.

Thesis (Ed.M.)--Boston University

The histone variant H2A.Z is an important regulator of transcription. One unsolved mystery is that why H2A.Z can have both activating and repressive effects on gene expression. By examining both coding and non-coding RNA transcripts in S.cerevisiae, we established that H2A.Z is present at both coding and non-coding promoters and have positive effects on the level of transcripts. The repressive effect of H2A.Z can be partially explained by the sense transcripts of gene being antagonised by H2A.Z-activated antisense transcripts. We also established that H2A.Z-associated non-coding transcripts are predominantly located at bidirectional promoters. The sense and antisense pairs produced from bidirectional promoters show high degrees of coregulation (especially co-activation) during stress response. Surprisingly, we found that the non-coding RNA co-activated with stress-response genes tend to spread the activation signal to the neighbouring gene further upstream, indicating their potential functions in gene regulation. In addition, we also observed that accumulation of H2A.Z at gene promoters is associated with slower recovery from gene induction, which could be related to the Ino80 pathway. In general, our results confirmed the interleaved nature of regulatory system in eukaryotes and highlighted the importance of taking both coding and non-coding transcripts into account while studying the transcriptional regulation.

CHAPTER I. The ultrastructure of transcription in cultured cells of T.c. carnifea was compared with, that of cultured cells from X. laevis using the "Miller spreading technique". Primary RHA transcripts of nonribosomal origin, up to 10 mum in length, and arranged in long transcription complexes were observed in T.c. camifex; whereas the maximum length of primary transcripts in S. laevis was about 2 mum. In both cell lines transcription was sparse and predominantly visualised as isolated RNP fibrils. CHAPTER II. Patterns of transcription in liver cells of X, laevis. T.c. carnifex and N. maculosus were compared, using the "Miller" technique. Chromatin had a nucleosomal structure and transcription complexes occurred on beaded chromatin. N. maculosus with the largest C-value (52 pg) synthesised the longest primary transcripts. Transcription complexes of two or more RNP fibrils were rare in the chromatin of these species. The percentages of different types of transcription complex (l RNP fibril, 2 RNP fibrils, 3 or more RNP fibrils) were similar for the three species. CHAPTER III. The ultrastructure of transcription in T.c. carnifex neurula cells was investigated, using the Miller technique. Transcription occurred predominantly in the form of isolated RNP fibrils whose median length was greater than for T.c. carnifex liver cells or cultured cells, in that order. The percentage of transcription complexes of two or more RNP fibrils was greater in T.c. carnifex liver (33.8%) than in T.c. carnifex embryos (17.3%). CHAPTER IV. The Miller technique was used to investigate the ultrastructure of transcription in X. laevis and T.c. carnifex' culture cells treated with cortisol or thyrozine. After cortisol treatment of X. laevis the median lengths of isolated nonrihosomal RNA fibrils, and the terminal fibrils of multifibril complexes, increased twofold, as did the percentage of multifibril transcription completes (from 5.6% to 28.1%), The DNA packing ratio of transcriptionally inactive regions of the chromatin was slightly less than that of untreated cells spread under the same conditions. There was also a decrease in the DNA packing ratio of transcriptionally active regions of the chromatin of treated cells as compared to the value for inactive regions. The results were inconclusive but were not sufficient to eliminate the possibility that an increase in the amount of transcription causes an increase in the size of the nucleus. The results were similarly inconclusive for T.c. carnifex cultured cells treated with cortisol.

No description available.

Macrophages are multifaceted innate immune cells, able to adapt their phenotype to respond to a myriad of conditions, engaging in tissue-specific functions and mediating either inflammatory or anti-inflammatory responses depending on the encountered stimuli. They conduct key roles in the orchestration of immune responses; from pathogen recognition through sterilising inflammation to resolution and repair. The Udalova laboratory has previously demonstrated that IRF5 promotes a pro- inflammatory macrophage phenotype, leading to the secretion of TNF, IL-12, and IL-23, enhancing Th1/Th17-mediated immune responses, and described the cooperation between IRF5 and the transcription factor RelA, which mediate the production of pro-inflammatory genes. The aim of this thesis is to further characterise the activity of IRF5 in macrophage inflammatory responses. I demonstrate that IRF5 not only regulates the transcription of cytokines and chemokines in response to bacterial stimuli, but also anti-microbial peptides, whilst simultaneously down-regulating homeostatic and resolving macrophage functions. My data also suggests that IRF5 plays a role in enforcing monocyte to macrophage differentiation by up-regulating the transcription of key macrophages markers and repressing dendritic cell identity genes. To further characterise the mechanisms of the inflammatory response mounted by macrophages I used an unbiased approach; combining twenty-three transcription factor ChIP-seq data sets with chromatin accessibility information from ATAC-seq, uncovering RUNX1 as a novel partner of IRF5 that binds co-operatively to clusters of enhancers, which control the transcription of pro-inflammatory genes in a signal-dependent manner. This is the first study demonstrating a critical role for RUNX1 in activity of inflammatory macrophages.

Doctor of Philosophy / Biochemistry and Molecular Biophysics Interdepartmental Program / Kathrin Schrick / Class IV homeodomain leucine-zipper transcription factors (HD-Zip TFs) are master regulators of cell-type differentiation in the plant epidermis. These transcription factors contain a putative START (STeroidogenic Acute Regulatory (StAR)-related lipid Transfer) lipid/sterolbinding domain that is hypothesized to link metabolism to gene expression in plant development. This study is focused on two class IV family members that serve as models in many of the experiments: GLABRA2 (GL2) is a key regulator of differentiation in hair cells called trichomes as well as other epidermal cell types in various plant tissues. The second member addressed in this study is PROTODERMAL FACTOR2 (PDF2), which plays a crucial role in epidermal cell specification in shoots. A leading hypothesis is that the START domain, by binding a ligand, controls transcription factor function, analogously to nuclear receptors from mammals. Domain swap experiments indicated that the START domain from both plants and mammals is a conserved ligand-binding motif that is required for transcription factor activity. To further address its function in ligand binding, mutational analysis of the START domain of GL2 was performed. Several of the mutations remove charged residues in the predicted ligand-binding pocket and resulted in loss-of-function phenotypes, suggesting that ligand binding is critical for HD-Zip TF activity. Chromatin immunoprecipitation–based sequencing (ChIP-seq) revealed that the START domain is dispensable for transcription factor binding to DNA. Using a high throughput thermal shift assay to screen a library of pure natural compounds, specific secondary metabolites were identified as putative START domain ligands for PDF2. Experiments in both yeast and N. benthamiana demonstrated that the START domain is required for homodimerization of GL2 through its Zip domain. It was also found that the START domains physically interact with RHAMNOSE SYNTHASE I (RHM1). Further, this work provided evidence for a previously elusive redundancy between GL2 and another class IV HD-Zip TF, and unveils a positive feedback loop in the maintenance of the GL2 activity during trichome differentiation. Taken together, these findings support the premise that START domains are central players in metabolic regulatory networks that can modulate transcription factor activity by binding ligands and mediating protein-protein interactions.

Pour s’adaptater à l’environnement et répondre aux besoins énergétiques, tous les organismes doivent pouvoir contrôler l’expression de certains gènes. La trancription constitue le premier échelon de cette régulation en déterminant le niveau d’ARNm produit au cours du temps. Chez les procaryotes, une seule ARNp est responsable de la synthèse de l’ensemble des ARN de la cellule et celle-ci est soumise à différents types de régulation. Ce contrôle peut s’effectuer à tous les niveaux par des processus faisant intervenir bon nombre de facteurs externes ou intrinsèques à la transcription comme les pauses de l’ARNp. Les riborégulateurs sont des ARN structurés majoritairement retrouvés dans la région 5’ non traduite des ARNm bactériens qui vont réguler l’expression du gène situé en aval. Suite à la liaison d’un métabolite particulier, appelé ligand, le riborégulateur change de conformation induisant une réponse directe qui déterminera si un gène est exprimé ou non. Au cours de mes travaux j’ai établi le lien qu’il existait entre les pauses de l’ARNp au cours de la transcription et la structure des riborégulateurs. En prenant pour modèles deux riborégulateurs liant le TPP, j’ai identifié une conformation des riborégulateurs qui est réfractaire à la liaison du ligand. Cette structure appelée «Anti-P1» se forme lorsque l’ARNp pause à la fin du riborégulateur, ce qui détermine une fenêtre de liaison cotranscriptionnelle du ligand.

No description available.