Search

Long non-coding RNAs (lncRNAs) are present in many eukaryotes and have an important function in controlling transcription. In the Transcriptional Gene Silencing (TGS) pathway lncRNAs can act as scaffolds for RNA binding proteins to target specific loci and affect chromatin structure. In many organisms genetic studies of TGS are limited because lncRNAs are produced by RNA Polymerase II (Pol II), the same polymerase responsible for transcribing protein coding genes. This limitation is overcome by using the model organism Arabidopsis thaliana in which lncRNA is produced by Pol V. Pol V is a specialized RNA polymerase which recently has evolutionarily diverged from Pol II in plants. Knockout mutants of Pol V allow the study of pathways involving lncRNA without disrupting the essential Pol II machinery. This tool has allowed discovery of many mechanistic principles of TGS. In this pathway, ARGONAUTE 4 binds to small interfering RNA (siRNA) and to lncRNA for chromatin targeting. Other proteins, SUPRESSOR OF TY INSERTION 5-LIKE (SPT5L) and INVOLVED IN DE NOVO 2 (IDN2) also bind to Pol V transcripts and are important for TGS. Components of TGS are essential for directing histone modification and de novo DNA methylation to affect Pol II transcription at specific loci. While this was thought to mainly occur to silence transposons, this work shows that TGS components localize to promoters and function in controlling gene expression. I also demonstrate that TGS is important for positioning nucleosomes, and that AGO4, IDN2, and SPT5L are all important for this function. This is in contrast to directing DNA methylation where AGO4 plays a large role, while the contribution of SPT5L and IDN2 is minor. I also find that TGS can control gene expression long-range by inhibiting long distance chromatin interactions. Overall this work revolutionizes how we think of TGS and is a major contribution to the mechanistic knowledge on the functions of lncRNA.

Although every cell contains a full complement of genes, not every gene is expressed equally. It is through different types of regulation that each cell is able to differentiate into various cell types to generate a multi-cellular organism. One such type of regulation is at the transcriptional level. In my thesis work, I have gained a better understanding of how transcriptional regulation occurs in Arabidopsis thaliana through two divergent projects. First, I have characterized the molecular mechanism by which APETALA2 (AP2), a floral homeotic gene, acts in floral development, by identifying its elusive target binding sequence, TTTGTT and/or AACAAA. In addition, I have found that AP2 modulates the expression of AGAMOUS (AG), another floral homeotic gene involved in floral organ specification as well as stem cell maintenance, through this site in vivo. This finding provided a molecular link between AP2 and AG, thereby answering a long-standing question in the field of floral development. Second, using a forward chemical genetics screen for players involved in transcriptional gene silencing, I have identified a chemical, Camptothecin (CPT), which released a transcriptionally silenced luciferase-based reporter. CPT is a well-known anti-cancer compound known to target DNA Topoisomerase Iα. Topoisomerases are used to relieve torsional stress on the DNA double helix during replication. Here in my thesis work, I have found that in Arabidopsis, TOP1α is also involved in transcriptional gene silencing through several different modes: body methylation and RNA-directed DNA methylation. Addition of the chemical resulted in the release of not only DNA methylation at the reporter transgene, but also endogenous loci as well. These findings highlight the diverse functions of TOP1α in development and provide implications in the study of cancer biology.

A major finding of this study is that plant hosts methylate geminivirus chromatin as an antiviral epigenetic defense. The susceptibility of methylation-deficient mutant plants to geminiviruses of two distinct genera (Beet curly top virus, BCTV, a monopartite curtovirus, and Cabbage leaf curl virus, CaLCuV, a bipartite begomovirus) was evaluated. It was found that Arabidopsis thaliana plants with mutations in genes encoding cytosine or histone H3 lysine 9 (H3K9) methyltransferases, RNA-directed DNA methylation pathway components, or adenosine kinase, an enzyme required for maintaining the methyl cycle, show enhanced susceptibility to geminivirus infection. We found that cytosines in geminivirus DNA are methylated in the intergenic region, which contains the origin of replication and promoters for bidirectional transcription, and that methylation levels are reduced in mutants that show enhanced susceptibility. In addition, methylation-deficient mutant plants are unable to recover from geminivirus infection, and do not display symptom remission characteristic of wild-type plants. We concluded that viral genome methylation, leading to transcriptional gene silencing, is a strategy employed by plant hosts to silence invading geminiviruses. Genetic analysis revealed that genes important for methylation-based antiviral defense were also critical for repression of endogenous transposons. Thus, we proposed that geminiviruses may be used as a model to study chromatin methylation. In this investigation, we sought to ascertain which members of the Arabidopsis dsRNA binding protein family (DRB 2, 3, 4 or 5) are associated with DCL3 and AGO4, key components in the chromatin methylation pathway. We found that drb3 mutants are uniquely hypersusceptible to geminivirus infection, and that viral methylation levels are greatly reduced in the drb3 mutant. The drb3 mutant, like ago4 and dcl3, fails to recover from geminivirus disease. In addition, DRB3 physically interacts with DCL3 and AGO4 in distinct subnuclear locations. This uncovers a function for DRB3 in methylation, downstream of 24 nt siRNA biogenesis. It also suggests that geminivirus genomes may be used as convenient probes to identify additional pathway components and elucidate novel functions for known components in the chromatin methylation arm of RNA silencing. Finally, we expanded our analysis of host genes required for recovery from geminivirus disease in order to better define their roles in the methylation pathway. In this study, we discovered that non-CG methylation, conditioned by the cytosine methyltransferases CMT3 and DRM2, is required for recovery. Of the two plant-specific RNA polymerases involved in methylation, Pol V is required for recovery. Recovery was delayed in pol IV mutants suggesting that Pol II may compensate for Pol IV activity, and that Pol IV might also be involved in the spread of antiviral silencing. An investigation of DCL2/4-mediated post-transcriptional gene silencing (PTGS) revealed that the dcl2/4 mutants showed a mixture of recovered and non-recovered shoots suggesting that PTGS, although non-essential, plays a role in recovery. Interestingly, a dcl2/3/4 triple mutant was found to be methylation-competent, suggesting a possible new role for DCL1 in methylation. In summary, this work identifies methylation as a host defense against geminiviruses, and introduces the geminivirus model to study chromatin methylation.

RNA interference is a phenomenon by which double-stranded RNA is processed into small interfering RNAs (siRNAs) that can cause gene silencing in plants, yeast, Drosophila (fruit fly) and human cells. Small interfering double stranded RNAs (siRNAs) can induce gene silencing via two pathways: post-transcriptional gene silencing (PTGS) and transcriptional gene silencing (TGS). PTGS involves siRNA triggering of sequence- specific degradation of mRNA takes place in the cell's cytoplasm. In contrast, siRNAs, which induce TGS have sequence complementary to regions within a gene's promoter and the action happens in the nucleus and is associated with de novo methylation of CpG sites and histone methylation within the promoter region. Suziki et al. utilised siRNAs to target regions within 5'-long-terminal repeat (5'LTR) promoter region of human immunodeficiency virus-1 (HIV-1) and subsequently induce silencing of HIV-1 in HeLa cells. Here I recapitulate the previous findings in HIV-1 by showing that certain promoter-targeted siRNAs can also induce silencing of simian immunodeficiency virus (SIV) replication by inducing CpG methylation and epigenetic changes. A similar silencing effect was observed by using shRNA and miRNA mimics based on the most effective siRNAs. The shRNAs and miRNA mimics after conversion gained a PTGS effect in addition of TGS effect. The dual effect appeared to enhance the silencing efficiency. Here I show that the co-localisation of Ago1 and Ago2 with a TGS inducing siRNA (si2A) in nucleus and at the rim of the nucleus respectively in SIV infected cells but not in non-infected cells. The P-body protein, GW182 complexed with si2A was found in both the nucleus and nuclear envelope and was consistent with the location Ago1/si2A and Ago2/si2A complexes. This indicated a role for this protein in TGS together with Ago proteins. Using SIV 5'-LTR luciferase reporter cells, complexes containing Ago1/Ago2/si2A were showed to work in combination and provide the optimal suppression.

Modulation of gene transcription by agents composed of nucleic acid provides a novel approach for the sequence-specific control of gene activity. Transcriptional gene silencing or activation has been reported in human cells with molecules complementary to sequences found on either strand of chromosomal DNA within the promoter element of genes. Such molecules include nucleic acid analogues PNA and LNA, and in recent times dsRNA. In human cells, transcriptional control by small dsRNA complementary to the sense strand of DNA has been reported to occur through RNA-mediated epigenetic modifications, whereas dsRNA targeting the antisense strand are proposed to function by steric blocking of transcription, a model supported by promoter studies involving PNA and LNA. The aim of this thesis was to critically evaluate the potential for nucleic acid-based agents to modulate the transcription of genes in human cells. Furthermore, this work examined the mechanisms for nucleic acid agents found to influence promoter activity and determine their target-specificity in addition to sequence-specificity. To this end, dsRNA, PNA or LNA were directed to target genomic DNA at selected sites within the promoter region of the human VEGF gene. VEGF was considered as an appropriate gene model system as its expression and regulation have been well characterised. Evaluation of PNAs was frustrated by problems associated with solubility and cellular uptake, while LNA oligomers were determined to be cytotoxic. Transient assays in human cell lines revealed suppression of endogenous VEGF production by siRNA targeted to either strand of VEGF promoter DNA. However, subsequent experiments involving a VEGF promoter-expressed reporter plasmid engineered to contain an altered siRNA target site revealed that this suppression was due to an off-target phenomenon. A microarray and bioinformatic approach that was developed to test the hypothesis that the suppression of unintended genes was responsible for off-target silencing identified a known regulator of VEGF expression as a possible unintended off-target. Hence, this study is the first to investigate the possible targeting of mammalian promoter sequences using nucleic acid molecules that has included the examination of target-specificity. The findings in this thesis underscore the need for such studies, particularly in the investigation of such a novel area as the transcriptional regulation of genes by nucleic acid agents.