In this thesis I extend a class of grammars called conjunctive grammars to a stochastic form called stochastic conjunctive grammars. This extension allows the grammars to predict pseudoknotted RNA secondary structure. Since observing sec- ondary structure is hard and expensive to do with today's technology, there is a need for computational solutions to this problem. A conjunctive grammar can handle pseudoknotted structure because of the way one sequence is generated by combining multiple parse trees. I create several grammars that are designed to predict pseudoknotted RNA sec- ondary structure. One grammar is designed to predict all types of pseudoknots and the others are made to only predict a pseudoknot called H-type. These grammars are trained and tested and the results are collected. I am able to obtain a sensitivity of over 75% and a speci city of over 89% on H-type pseudoknots

Ebola virus (EBOV) is an enveloped, negative-sense single-stranded RNA virus that causes severe hemorrhagic fever in humans and nonhuman primates. The EBOV glycoprotein (GP) gene encodes multiple transcripts due to RNA editing at a conserved editing site (ES) (a hepta-uridine stretche). The majority of GP gene transcript is unedited and encodes for a soluble glycoprotein (sGP); a defined function has not been assigned for sGP. In contrast, the transmembrane glycoprotein (GP1,2) dictates viral tropism and is expressed through RNA editing by insertion of a nontemplate adenosine (A) residue. Hypothetically, the insertion/deletion of a different number of A residues through RNA editing would result in another yet unidentified GP gene product, the small soluble glycoprotein (ssGP). I have shown that ssGP specific transcripts were indeed produced during EBOV infection. Detection of ssGP during infection was challenging due to the abundance of sGP over ssGP and the absence of distinguishing antibodies for ssGP. Optimized two- dimensional (2-D) gel electrophoresis verified the expression of ssGP during infection. Biophysical characterization revealed ssGP is a disulfide-linked homodimer that is exclusively N-glycosylated. Although ssGP appears to share similar structural properties with sGP, it does not have the same anti-inflammatory function. Using a new rapid transcript quantification assay (RTQA), I was able to demonstrate that RNA editing is an inherent feature of the genus Ebolavirus and all species of EBOV produce multiple GP gene products. A newly developed dual-reporter minigenome system was utilized to characterize EBOV RNA editing and determined the conserved ES sequence and cis-acting sequences as primary and secondary requirements for RNA editing, respectively. Viral protein (VP) 30, a transcription activator, was identified as a contributing factor of RNA editing— a proposed novel function for this largely uncharacterized viral protein. Finally, I could show that EBOV RNA editing is GP gene-specific because a similar sequence located in L gene did not serve as an ES, most likely due to the lack of the necessary cis-acting sequences. In conclusion, I identified a novel soluble protein of EBOV whose function needs further characterization. I also shed light into the mechanism of EBOV RNA editing, a potential novel target for intervention.

The use of ribonucleoside phosphoramidites protected at the 2$ sp prime$-hydroxyl with alkylsilyl ethers was developed for the solid phase synthesis of oligoribonucleotides. A series of pentadecameric oligomers of various base composition was synthesized using this methodology and a complete protocol for the deprotection of synthetic oligoribonucleotides was developed. The use of $ beta$-cyanoethyl phosphate protection was compared to that of methyl protection in the synthesis of pentadeca-uridylic acid sequences and no significant difference between the use of the two protecting groups was found. Concurrently a procedure for the derivatization of long chain alkylamine controlled pore glass beads with ribonucleosides was used to provide an improved solid phase support for the efficient automated synthesis of long oligomers. A side reaction on this support was identified and a method for its prevention developed. / A reaction occurring at the O$ sp6$-position of guanosine residues, which leads to chain cleavage, was identified. The basis of this reaction was characterized by both $ sp{31}$P nuclear magnetic resonance and the synthesis of oligoriboguanylate sequences. Three methods involving: protection of the O$ sp6$-position, the use of an alternate coupling cycle, and the use of a different phosphoramidite activator, were successfully applied to the resolution of this problem. The aforementioned procedures were utilized in the synthesis of a half E. Coli N- f - met tRNA molecule 43 units in length. / Finally the cumulative knowledge gained from the above studies was applied to the synthesis of an entire E. Coli N- f - met analogue 77 units in length along with a number of other very long sequences. Methods for the deprotection of these oligomers were investigated culminating in the isolation of oligoribonucleotides which were successfully characterized by polyacrylamide gel electrophoresis, enzyme degradation, enzymatic and chemical sequencing, terminal nucleotide analysis, and, in the case of the E. Coli f - met tRNA analogue, an aminoacylation assay.

No description available.

Transcription in mammalian and yeast mitochondria proceeds from a few well defined promoters, with processing of polycistronic transcripts producing the mature RNAs. The levels of different sequences in the steady-state RNA populations depend on differential promoter strengths, transcription attenuation and/or selective termination, and differential RNA stabilities. To gain insights into the processes governing transcription and RNA levels in plant mitochondria, a system using isolated maize mitochondria, which synthesize bona fide mitochondrial RNAs, was developed and partially characterized with respect to exogenous requirements and sensitivity to inhibitors of DNA-dependent RNA synthesis. / Although initiation and processing probably occur at reduced levels in isolated maize mitochondria, endogenous DNA templates are extensively transcribed at the same relative rates as in vivo. Isolated maize mitochondria were used to demonstrate that differential rates of both synthesis and turnover help determine the steady-state abundances of various mitochondrial RNA sequences and that mitochondria from certain lines possess an autonomously-replicating, RNA-based genetic system.

Typescript. / Thesis (Ph. D.)--University of Hawaii, 1968. / Bibliography: leaves 91-103. / ix, 103 l graphs, tables

Typescript. / Thesis (Ph. D.)--University of Hawaii, 1969. / Bibliography: leaves 136-144. / xii, 144 l

The human hepatitis B virus (HBV) is a small hepatotropic virus, which affects approximately 350 million chronic sufferers worldwide. It has a compact 3.2 kbp dsDNA genome encoding four major overlapping genes namely core, polymerase, surface and X required for its replication. The virus synthesises a pregenomic RNA (pgRNA) which functions both as an RNA intermediate for reverse transcription into the DNA genome and as the mRNA for the translation of the core (C) and polymerase (P) proteins. The core overlaps the polymerase gene and is translated at a 10 to 1 ratio. The polymerase gene translated from the P AUG codon is preceded by at least 4 upstream AUG codons (uAUGs), namely C AUG, C1 AUG, J AUG and C2 AUG. Various mechanisms have been implicated in the synthesis of the polymerase protein. This led to the currently accepted model which involves leaky scanning and a reinitiation mechanism in polymerase synthesis. However, multiple sequence alignment of the pgRNA revealed a short upstream open reading frame (uORF) highly conserved at the nucleotide level in all HBV subtypes and mammalian hepadnaviruses. This previously unreported uORF, designated as C0 ORF in this study is also conserved in its position and length. Past studies have either omitted this uORF in their test constructs or ignored its potential role. The C0 ORF has a conserved weak initiation context and is located within the epsilon structure within the 5� leader of the pgRNA, required for viral encapsidation. Importantly, the C0 ORF precedes and overlaps the core ORF, which may suggest an alternative model in which the core and polymerase may be translated and coordinately regulated. Fusion of the C0 ORF to luciferase showed for the first time that this uORF is translated through the detection of reporter activity (~20% of C) and also visualisation of the fusion protein via western analysis using anti-C0 and anti-luciferase antibodies. Subsequent removal of the C0 ORF implicated a role in repressing downstream core fusion protein synthesis in HepG2 cells. A similar repression was observed on J expression. To study the effect of C0 on downstream polymerase translation, a pgRNA-like DNA construct was made and subsequent mutations introduced. Mutation of the C0 AUG led to an increase in initiation at the downstream P AUG. Alteration of the existing weak initiation context to an optimal context which favours stronger initiation consistently showed a potential role for C0 ORF in facilitating reinitiation at certain downstream initiation codons including P AUG. Mutations of other uAUGs preceding the P AUG were also done to better understand their roles in regulating polymerase synthesis. The removal of the C AUG markedly increased expression from the P AUG. This study revealed other internal uAUGs in-frame to the C AUG, namely the C1 and C2 AUGs are also effectively translated, further reducing availability of translating ribosomes to downstream P AUG. Indeed the removal of the C1 and C2 AUGs led to a corresponding increase in initiation from the P AUG. Initiation at the internal J AUG was also reported and its removal showed a significant decrease in expression from the P AUG, consistent with the previous model implicating reinitiation at the P initiation site after translation of the short J ORF. The inhibitory role of the 5 uAUGs prior to the P AUG were confirmed when all were removed, giving rise to translation almost equal to that at C AUG. Taken together, these results suggest a new model in which the HBV C0 ORF plays a key role in controlling core and polymerase synthesis by repressing core translation and making available more ribosomes to downstream AUGs possibly facilitating translation reinitiation. In addition, the translation of the C0 ORF across the [epsilon] region may also preclude encapsidation, potentially acting as a switch discriminating the pgRNA template between encapsidation and translation. Therefore, the highly conserved [epsilon] region and C0 ORF present an excellent target for molecular based antiviral drugs (antisense oligonucleotides, aptamers, ribozymes) potentially providing new anti HBV drugs.

Long ncRNAs (non-protein coding transcripts generally considered longer than 200 nucleotides to be distinguished from classes of small RNAs) are abundantly transcribed from the mammalian genome. Despite their abundance, little is known about these transcripts. Although several individual long ncRNAs have been well-characterised and ascribed important cellular functions, there remains considerable controversy as to whether long ncRNAs are, in the main, functional. Indeed, their abundance has prompted many people to argue that long ncRNAs are simply transcriptional ‘noise’ generated by spurious transcription initiation events resulting from low RNA polymerase II fidelity. This thesis demonstrates that large numbers of long ncRNAs are specifically expressed along both temporal and spatial axes of mouse development in a manner consistent with a biological function. Custom-designed microarrays were employed to analyse the expression profiles of large numbers of long ncRNAs, along with protein-coding genes, in two models of cellular differentiation; the differentiation of mouse embryonic stem (ES) cells from pluripotency to differentiation along a hemopoietic lineage; and the commitment and differentiation of neural stem cells to oligodendrocytes. The core networks that include gene expression, transcription factor binding sites and chromatin domains that regulate ES cell pluripotency and lineage specification have been the subject of considerable attention and provide a detailed context in which to analyse ncRNA expression. Of those ncRNAs examined, 945 (26% of total) ncRNAs were expressed during the differentiation of ES to embryoid body (EB), of which 174 were significantly differentially expressed. Many of these ncRNAs were transcribed from genomic locations that overlapped modified chromatin domains, and in two further studied cases directly engaged with epigenetic machinery. Similarly, 332 long ncRNAs (9% of those examined) were expressed during processes of neuronal-glial fate switching, neurogenesis and oligodendrocyte progressive differentiation and termination, of which around half were also significantly differentially expressed. Furthermore, many of these ncRNAs exhibited expression profiles that coincided with pivotal events during the commitment and differentiation of neural stem cells (NSC) to mature myelinating oligodendrocytes. Consideration of the genomic context revealed many long ncRNAs were expressed from diverse places including intergenic, intronic, and imprinted loci and may overlap with, or are transcribed antisense to, protein-coding genes with previously described roles in either ES or NSC pluripotency and differentiation. This association also extended to expression profiles, where a comparative analysis often showed complex relationships of expression between ncRNAs and associated protein coding genes, suggesting a potential role for ncRNAs in regulating the expression of associated gene loci. The complexity and specificity of the long ncRNAs expression was illustrated by analysis of the in situ hybridisation (ISH) data conducted in collaboration with the Allen Brain Atlas. Of 1328 long ncRNAs, 849 (64%) were expressed in the mouse brain, 623 (47%) of which exhibited specific expression profiles associated with distinct neuroanatomical regions, cell types, or subcellular compartments. Again, examination of their genomic context revealed long ncRNAs were often associated with protein-coding genes of neurological importance and this association often extended to include linked expression profiles in the mouse brain. The comparative analysis of protein-coding gene expression relative to associated noncoding transcription also revealed an additional level of complexity in gene structure and genomic architecture. Analysis of both microarray and ISH data show 3’UTRs can exhibit discordant expression profiles relative to their associated protein coding genes, often in a tissue- and developmentally-specific manner. Indeed, a genome-wide analysis showed that the independent expression of 3’UTR transcripts is prevalent throughout the mouse genome where they may function intrinsically as long ncRNAs during development. Together, these genome-wide analyses indicate a large proportion of long ncRNAs exhibit specific expression profiles that are inconsistent with the notion they are meaningless transcriptional noise. Taken together with numerous studies published in recent years, this thesis provides evidence to support the emergence of long ncRNAs as a major functional component of the regulatory network that underpins differentiation and development in mammals and other complex organisms.

Chronic Myeloid Leukaemia (CML) is a clonal haemopoietic stem cell (HSC) disorder characterised by the presence of a disease-specific gene, BCR-ABL, which leads to the production of a bcr-abl mRNA transcript. CML is an ideal candidate for gene therapy using ribozymes (Rz), catalytic RNA molecules that cleave and inactivate target RNA in a sequence specific manner. Limited data is available on the activity of ribozymes in human CML cells. In this study, hammerhead ribozyme sequences directed against the b3a2 bcr-abl mRNA sequence (Rz6-Rz10) were cloned into several retroviral vectors. Initial experiments using MSCVHSA based retroviral constructs failed to express the sequences in cell lines. Rz cDNA fragments were then cloned into an LNL6 based retroviral vector (LGL1) encoding a GFP reporter gene and stable LGLRz constructs produced. Using cell sorting, high-titre PA317 producer cell line clones were isolated. Transcriptional silencing of the LGLRz6 producer cell line occurred with prolonged culture, with partial reversal on treatment with the demethylating agent 5' azacytidine. To assess the activity of these constructs in human cells, CD34+ HSC were isolated from newly diagnosed b3a2 Ph+ CML patients. Cells were transduced with either control LGL vector or the LGLRz6 construct. Transduced human cells were sorted based on GFP expression and placed into long-term HSC culture (LTC-IC assays). Using a common cDNA, RT-PCR was performed to detect the expression of both the transgene and bcr-abl in individual colonies derived from the LTCIC assay at various time points, allowing assessment of the effect of transgene expression on bcr-abl expression. LGLRz transgene expression was detectable for up to 6 weeks in culture. Colony RT-PCR results from 3 patients showed that expression of the LGLRz6 construct was associated with decreased bcr-abl expression. It also appeared that the reduced bcr-abl expression decreased the proliferation of Ph+ cells leading to their loss from culture. In summary, these results appear to show an effect of a retroviral vector containing a bcr-abl Rz sequence on human CML HSC. Targeting of bcr-abl remains a valid therapeutic goal in the Imatinib era, particularly if problems related to effective ribozyme delivery and targeting can be overcome.