Viruses are obligate intracellular parasites that replicate utilizing the resources of host cells. The replication of positive sense RNA viruses is coupled with alterations to host cell membranes. These viruses are believed to replicate efficiently by using co-opted membrane structures assembled from viral and host cell proteins and lipids. Poliovirus is a prototypical positive-sense RNA virus, however the topological details of viral RNA replication are not well understood. In this work we use electron cryotomography, among other methods, to examine the ultrastructure of fractionated poliovirus RNA replication factories that were formed within infected cells, and to investigate oligomeric interactions within a three dimensional crystal formed by a poliovirus polymerase point mutant. Investigation of the ultrastructure of isolated viral RNA replication factories shows that the low resolution features of cryopreserved membrane structures are essentially identical to previously observed structures within plastic sections of infected cells. Furthermore, greater detail visible using electron cryotomography reveals pore-like structures and other high energy membrane conformations within the replication factories. We see a mix of single, double, and multi-membrane structures that are arranged with openings that connect their interior lumenal space to the exterior environment. The lumen of some of these membranous structures contains a linear polymeric density thought to be RNA. We conclude that the RNA replication of poliovirus may occur on the lumenal surface of vesicular membranes with an opening to the cytoplasm for metabolite and product exchange. Within the poliovirus replication machinery, the principal component is the RNA polymerase 3Dpol. This prototypical RNA-dependent RNA-polymerase forms homo-oligomeric interactions that are key to its functions. To investigate these interactions, previous studies focused on hollow helical structures formed by wild-type polymerase. Here, we investigate the structure of small three-dimensional crystals formed by 3Dpol with a mutation of a single residue, lysine 314, to alanine. Using electron cryotomography and volume averaging, we demonstrate that the crystal packing within this point mutant does not include physiological polymerase-polymerase interactions. Elucidation of the topology of poliovirus replication machinery provides a basis for future development of antiviral therapeutics.

The universal traits of cellular expression and replication in modern life point to the existence of an ancient RNA world. Leading up to the origin of life, this stage of evolution utilized RNA as the genetic material, and as a catalyst in the form of ribozymes. Although it is expected that a polymerase ribozyme was required for the efficient replication of RNA, it is also likely that the earliest form of replication took place under non-enzymatic conditions. There are several problems with the current scenarios depicting non-enzymatic RNA replication, thus we aim to examine them in more detail using computational models. We first consider the relationship between the thermodynamics of RNA base pairing and non-enzymatic nucleotide addition in an attempt to model the rate of primer extension. Our predicted rates reveal the model parameters to be too simple to produce reliably accurate results. For now, we should simply use available experimental rate data, until we have access to more data and less unknown parameters. Nevertheless, the model indicates that the primer extension rate does depend on thermodynamics of base pairing, and a more accurate model can be of great use when creating realistic complex models of RNA world scenarios. In chapter 3, we investigate non-enzymatic RNA replication under temperature cycling using computer simulations. When starting with a diverse mixture of sequences, partially matching sequences can reanneal in configurations that allow continued strand growth. This is in contrast to the case of having multiple copies of matching sequences, where reannealing occurs quickly upon cooling. We find that, starting with short oligomers, strands can grow over multiple cycles to produce long sequences over 100 nucleotides in length. The small strand extension per cycle does not produce replicates of any one specific sequence. This relates to the work done in chapter 4, where we look for the presence of a virtual circular genome within our simulations. In a virtual circle, short overlapping RNA sequences will make up a mutually catalytic set. Within the diversity of our simulation, virtual circles are rare, and require a specific level of starting mixture diversity along with no input of new sequences. Continued replication of the diverse sequence mixture and emergence of long strands may eventually lead to the creation of rolling circles and ribozymes. / Thesis / Master of Science (MSc) / The origin of biological life can be traced back by looking at the common themes between modern cellular processes. The role of RNA polymers seems to be of great importance, making us believe that an RNA world existed leading up to life’s origin. During this time, RNA would act as both a genetic material and a catalyst. To examine this theory in more detail, we use computational modeling to recreate and explore the various potential chemistries and conditions on the early Earth. Specifically, we explore the problems that exist for the replication and production of RNA polymers. Our results can be used to guide future theoretical and experimental research of the RNA world.

Foot-and-mouth disease (FMD) is a highly infectious disease of wild and domestic cloven–hoofed animals such as cattle, swine and deer. It is caused by one of the most contagious animal diseases known; FMD virus (FMDV). Since the disease is endemic in many countries, transmission by international travel/trade presents an on-going potential threat to the UK. Very little is known at the molecular level about how FMDV replicates within host cells. In this study, FMDV replicons have been used to investigate FMDV RNA replication and to improve our understanding of the viral life cycle: a process which will aid in the production of a new generation of live-attenuated vaccine candidate strains. Sequences encoding the capsid coding region of the genome have been replaced with green fluorescent protein (GFP) such that replication can be monitored in live cells via GFP fluorescence. This provides a rapid, non-invasive screen for replicative fitness that can be used outwith high disease security facilities. Differences between replicating and non-replicating forms could easily be distinguished, highlighting the potential of this system to screen for attenuated genomes. The FMDV replicon system was improved through a series of construct modifications until an optimal system was produced. A range of different methods were used to attenuate the replication of these genomes. Of major significance is the finding that increasing dinucleotide frequencies were shown to decrease growth kinetics of Echovirus 7 – as opposed to altering the codon-pair bias - and the application of this finding to construction of further replicon systems (and RNA viruses in general) is described.

Earth’s crust and primordial ocean formed more than 4 billion years ago and life is believed to have originated on earth at least 3.6 billion years ago. This suggests that primitive cellular life must have evolved from non-living matter during that period of several hundred million years. To study the transition from chemistry to biology, a simple vesicular system called a protocell is an ideal model that is self-organized and contains informational or metabolic materials. This thesis starts by exploring the replication of a model genetic material under plausible prebiotic conditions. The non-enzymatic copying of RNA was found to be catalyzed by Fe2+, which used to be abundant in aqueous environments on the early anoxic earth. Fe2+ was found to be a better catalyst of non-enzymatic RNA copying and ligation in slightly acidic to neutral pH conditions than Mg2+, the divalent cation used to catalyze these reactions in previous studies. This finding suggests that ferrous iron could have facilitated the replication and evolution of RNA on the prebiotic earth. To gain a better understanding of the properties of protocell membranes, the impact of membrane composition and multi-bilayer structure on non-enzymatic and enzymatic biochemical reactions was studied. A fatty acid/phospholipid hybrid membrane system was proposed as a potential intermediate state in protocellular evolution. This membrane composition was investigated for its stability and permeability, two fundamental features of functional protocells. The system proved stable in the presence of divalent cations and retained permeability to small building block molecule. Vesicles with this composition were shown to host faster non-enzymatic RNA copying, and to enable enzymatic protein synthesis. To study the effects of multi-lamellarity, giant multilamellar vesicles (GMVs) were prepared by an extrusion-dialysis method. Compared with small unilamellar vesicles (SUVs), GMVs show slightly better ability to retain encapsulated RNA, while maintaining good permeability for small charged molecules. The multilamellar structure also promotes non-enzymatic RNA copying, providing preliminary evidence that membranes could also mediate catalytic functions as well as acting as a compartment. / 2020-07-02T00:00:00Z

The 3' terminal stem loop of the WNV minus-strand [WNV3'(-) SL] RNA was previously shown to bind the cell protein, T-cell intracellular antigen-1 (TIA-1), and the related protein, TIAR. These two proteins are known to bind AU-rich sequences in the 3' UTRs of some cellular mRNAs. AU stretches are located in three single-stranded loops (L1, L2, and L3) of the WNV3'(-) SL RNA. The RNA binding activity of both proteins was reduced when L1 or L2, but not L3, AU sequences were deleted or substituted with Cs. Deletion or substitution with Cs of the entire AU-rich sequence in either L1 or L2 in a WNV infectious clone was lethal for the virus while mutation of some of these nt decreased the efficiency of virus replication. Mutant viral RNAs with small plaque or lethal phenotypes had similar translational efficiencies to wildtype RNA, but showed decreased levels of plus-strand RNA synthesis. These results correlated well with the efficiency of TIA-1 and/or TIAR binding in in vitro assays. In normal cells, TIA-1 and TIAR are evenly distributed in the cytoplasm and nucleus. Between 6 and 24 hr after WNV infection, TIAR concentrated in the perinuclear region and TIA-1 localization to this region began by 24 hr. Similar observations were made in DV2 infected cells but at later times after infection. In infected cells, both proteins colocalized with dsRNA, a marker for viral replication complexes, and with viral non-structural proteins. Anti-TIAR or anti-TIA-1 antibody coimmunoprecipitated viral NS3 and possibly other viral nonstructural proteins. In response to different types stress, TIA-1 and TIAR recruit cell mRNA poly(A)+ into cytoplasmic stress granules (SG) leading to general translational arrest in these cells. SG were not induced by flavivirus infection and cells became increasingly resistant to arsenite induction of SG with time after infection. Processing Body (PB) assembly was also decreased beginning at 24 hr. These data suggest that the sequestration of first TIAR and then TIA-1 via their interaction with viral components in flavivirus infected cells inhibits SG formation and prevents the shutoff of host translation.

Tombusviruses are single, positive strand RNA viruses of plants, often associated with parasitic defective interfering (DI) RNAs. Two viral- coded gene products, namely p33 and p92, are required for tombusvirus replication. The overlapping domains of p33 and p92 contain an arginine/proline-rich (RPR) RNA binding motif. In this study, the role of RPR motif and viral RNA in tombusvirus replication and recombination, as well as involvement of viral RNA in tombusvirus replicase assembly was examined. Using site-directed mutagenesis I generated a series of RPR mutants of Cucumber necrosis tombusvirus (CNV). Analysis of RPR mutants defined that wild type RPR motif, especially two of the four arginines, were required for efficient RNA binding in vitro, for replication of tombusviruses, their associated DI RNAs, subgenomic (sg)RNA synthesis and DI RNA recombination in vivo. Experiments using a two-component tombusvirus replication system showed that RPR motif is critical for functions of both p33 and p92 in replication, but its role in these proteins might not be identical. Recombination studies using a novel tombusvirus three-component system revealed that mutations in RPR motif of p33 replicase protein resulted in an altered viral RNA recombination rate. Identified DI RNA recombinants were mostly imprecise, with recombination sites clustered around a replication enchancer and an additional putative cis-acting element that might facilitate the template switching events by the tombusvirus replicase. To study the role of RNA during the assembly of functional tombusvirus replicase, recombinant CNV replicase that showed similar properties to plant-derived CNV replicase was purified from Saccharomyces cerevisiae. When in addition to p33 and p92 proteins DI RNA was co-expressed in yeast cells, the isolated replicase activity was increased ~40 fold. Further studies defined RNA motifs within two short DI RNA regions that enhanced active CNV replicase formation. In summary, this study showed that the conserved RNA binding motif of the tombusvirus replicase proteins and viral RNA are involved in replicase assembly, viral RNA replication, subgenomic RNA synthesis and RNA recombination. This data shed new light on the complex roles of the viral elements in replication, and will help future studies aimed at interfering with viral infections.

Positive-strand RNA virus group are the most abundant among viruses affecting plants and animals. To successfully achieve replication, these viruses usurp or co-opt host proteins. To facilitate the discovery of host factors involved in Tomato bushy stunt virus (TBSV), yeast has been developed as a surrogate model host. Genome-wide approaches covering 95% of yeast genes, has revealed approximately hundred factors that could affect virus replication. Among the identified host factors, there are fourteen yeast genes, which affect/regulate lipid metabolism of the host. One of the identified host gene is ERG25, which is an important factor for sterol biosynthesis pathway, affecting viral replication. Sterols present in eukaryotes affect the lipid composition of membranes, where tombusviruses, similar to other plus-strand viruses of tobacco, replicate. Since potent inhibitors of sterol synthesis are known, I have tested their effects on tombusvirus replication. We demonstrated that these sterolsynthesis inhibitors reduced virus replication in tobacco protoplasts. Virus replication is resumed to the wild type level by providing phytosterols in tobacco protoplasts confirming the role of sterols in RNA virus replication in tobacco. We have also identified INO2, a transcription factor for many phospholipid biosynthetic genes, reduces virus replication in its deletion background. When we provided this gene product in the mutant background, viral replication was back to normal, confirming the role of Ino2p in tombusvirus replication. Further biochemical assays showed that the viral inhibition is because of alteration in the formation of the viral replicase complex. Using confocal microscopy, we showed that the viral replication protein, termed p33, is forming large and few punctate structures rather than the small and many by overexpressing Ino2p in the wild type yeast cells. Over-expression of Opi1, an inhibitor of Ino2p led to greatly reduced viral replication, further supporting the roles of the phospholipid pathway in tombusvirus replication. One of the phospholipid, which is regulated by this pathway, is cardiolipin an important component of the mitochondrial as well as peroxisomal membranes. We further characterized how cardiolipin is playing an important role for tombusvirus replication by using different biochemical approaches.

Two thirds of plant viruses are positive-strand RNA viruses including the family Tombusviridae. One of the best-studied members of this family is Tomato bushy stunt virus (TBSV). Like many other viruses, TBSV has much fewer genes when compared to its hosts’ genome. Nevertheless, TBSV utilizes its genome very judiciously. To compensate for a lack of many proteins of its own, it codes for multi-functional replication protein p33 and also co-opts host factors to facilitate its replication. By using recombinant replication proteins p33 and p92 containing single amino acid changes in protein-protein interaction domains (S1 and S2), I demonstrated that the replication proteins are required in sequential steps during virus replication. The in vitro cell-free extract(CFE) based TBSV replication assays revealed that mutations in S1 and S2 domains affected RNA template selection, recruitment and assembly of replicase complex. TBSV replicates on the cytosolic surface of peroxisomal membranes. To identify the host factor involved in this process of transporting viral replication proteins to peroxisome, I tested the peroxisomal transporter proteins for their ability to bind to p33 in vitro, which led to the discovery of Pex19p. Pull-down and co-purification experiments revealed transient nature of p33-Pex19p binding as expected from a transporter. When pex19p was retargeted to mitochondria, a large fraction of p33 was also re-distributed to the mitochondria validating the importance of Pex19p in p33 localization. TBSV also utilizes its genomic RNA for non-template activities during its replication. Accordingly, TBSV RNA serves as a platform for the assembly of replicase complex. To further characterize the regulatory cis-elements involved in this process, I utilized CFE and different TBSV RNA mutants together with recombinant p33 and p92 in vitro replication assays. These experiments revealed the role of RNA recruitment element [RIISL(+)] and 3’ non-coding regions as minimal cis-elements required to assemble functional replicase complex. The experiments also indicated that the RIISL(+) and 3’ non coding regions could be physically separated on two different RNA molecules to assemble TBSV replicase, suggesting insights into viral evolution.

Cellular T-cell intracellular antigen-1 related protein (TIAR) binds to the 3' terminal stem-loop of the West Nile virus minus-strand RNA [WNV 3'(-) SL RNA]. TIAR binding sites were previously mapped on loop 1 (L1) and loop 2 (L2) of the 3' (-) SL RNA and mutations of these sites in a WNV infectious clone inhibited virus replication. In the present study, data from in vitro binding assays suggested that multiple TIAR proteins bind to each WNV 3′ (-) SL RNA in a positively cooperative manner. The tertiary structure of WNV 3′ (-) SL RNA was predicted and it was suggested that L2 forms an exposed loop while L1 forms an embedded loop. We propose that TIAR binds first to L2 and that this interaction facilitates the binding of a second TIAR molecule to L1. Data from in vitro assays also showed that TIAR binds specifically to the WNV 3' (-) SL RNA but not to the complementary WNV 5' (+) SL RNA and that the C-terminal prion domain of TIAR contributes to RNA binding specificity. Immunoprecipitation experiments indicated that TIAR interacts with the WNV 3' (-) SL RNA in cells. Colocalization of TIAR and viral dsRNA in the perinuclear region of WNV-infected cells was visualized using a proximity ligation assay. In WNV-infected, TIAR-overexpressing cells, increased extracellular virus yields, intracellular viral protein and RNA levels, and an increased ratio of viral plus-strand RNA to minus-strand RNA were observed. These data suggest that TIAR enhances WNV plus-strand RNA synthesis from the minus-strand template. WNV infections induce small TIAR foci formation in primate cells but not rodent cells. The TIAR foci are located in the perinuclear region and differ in size and location from arsenite-induced stress granules (SGs). However, the small TIAR foci contain many SG components, such as G3BP, PABP, and eIF3A, but not HuR. Arsenite-induced SG formation is still inhibited by WNV infection in these cells. eIF2a phosphorylation was observed in some infected cells that contained WNV-induced TIAR foci but viral NS3 protein accumulation was not inhibited. The data suggest that WNV-induced TIAR foci in primate cells are not canonical SGs.

Kyoto University (京都大学) / 0048 / 新制・課程博士 / 博士(農学) / 甲第18333号 / 農博第2058号 / 新制||農||1023(附属図書館) / 学位論文||H26||N4840(農学部図書室) / 31191 / 京都大学大学院農学研究科応用生物科学専攻 / (主査)教授 奥野 哲郎, 教授 佐久間 正幸, 准教授 吉田 天士 / 学位規則第4条第1項該当