Background. Viruses strongly influence microbial population dynamics and ecosystem functions However, our ability to quantitatively evaluate those viral impads is limited to the few cultivated viruses and double-stranded DNA (dsDNA) viral genomes captured in quantitative viral metagenornes (vromes). This leaves the ecology of nondsDNA viruses nearly unlmovvn, including single-stranded DNA (ssDNA) viruses that have been frequently observed in viromes, but not quantified due to amplification biases in sequencing library preparations (Multiple Displacement Amplification, Linker Amplification or Tagmentation). Methods. Here we designed mock viral communities including both ssDNA and dsDNA viruses to evaluate the capability of a sequencing library preparation approach including an Adaptase step prior to Linker Amplification for quantitative amplification of both dsDNA and ssDNA templates. We then surveyed aquatic samples to provide first estimates of the abundance of ssDNA viruses. Results. Mock community experiments confirmed the biased nature of existing library preparation methods for ssDNA templates (either largely enriched or selected against) and showed that the protocol using Adaptase plus Linker Amplification yielded viromes that were 1.8-fold quantitative for ssDNA and dsDNA viruses. Application of this protocol to community virus DNA from three freshwater and three marine samples revealed that ssDNA viruses as a whole represent only a minor fraction (<5%) of DNA virus communities, though individual ssDNA genomes, both eukaryoteinfecting Circular Rep-Encoding Single-Stranded DNA (CRESS-DNA) viruses and bacteriophages from the Microviridae family, can be among the most abundant viral genomes in a sample. Discussion. Together these findings provide empirical data for a new virome library preparation protocol, and a first estimate of ssDNA virus abundance in aquatic systems.

The introduction of next-generation sequencing (NGS) technologies has dramatically changed the field of virology, with many significant discoveries of novel circular replication-associated protein (Rep) encoding single-stranded (CRESS) DNA viruses. Traditionally, most research into CRESS DNA viruses has often focused on investigating plant and animal pathogens that are of significant economic importance. This research has led to the discovery and establishment of three different CRESS DNA families including Geminiviridae, Nanoviridae and Circoviridae, which infect eukaryotes. CRESS DNA viruses can have single or multicomponent genomes, with the latter requiring all components for infection. CRESS DNA viruses have circular single-stranded DNA (ssDNA) genomes with at least one protein encoding a Rep which is responsible for viral replication. It has been shown that CRESS DNA viruses are able to evolve rapidly with nucleotide substitution rates that are similar to those observed in RNA viruses. The Rep gene has conserved regions known as motifs which are often used to determine relatedness between CRESS DNA virus. NGS has expanded our knowledge on the diversity of novel CRESS DNA viruses. Viral genomes are now routinely recovered from different sample types without any prior knowledge of the viral sequence. This has led to the development of the field of viral ecology. This field places an emphasis on viruses being one of the most abundant organisms on earth, and are therefore likely to play a major role in ecosystems. Environmental metagenomic studies have isolated CRESS DNA viruses from sea water, freshwater, faecal matter from various animals, soil, the atmosphere, sediments and sewage; dramatically increasing the known CRESS DNA viral genomes in the public domain. These studies are shedding light on the distribution of CRESS DNA viruses, as well as providing baseline data for future studies to examine virus-host interactions, community structure and ultimately viral evolution. Vector enable metagenomics (VEM) is another novel approach utilising NGS techniques for discovering CRESS DNA viruses. As many plant-infecting CRESS DNA viruses such as geminiviruses and nanoviruses are vectored by insects, this approach exploits this mechanism by using insect vectors as a surveillance tool to monitor and survey these viruses circulating in ecosystems. Recent studies have used these methods to identify known viral plant pathogens as well as novel viruses circulating in insect vectors such as whiteflies and other higher order insects such a mosquitoes and dragonflies. These approaches successfully demonstrated that VEM can be used as a unique method, with the first mastrevirus discovered in the new world being recovered from dragonfly species Erythrodiplax fusca using this approach. The research in this thesis uses metagenomics to survey CRESS DNA viral diversity in different organisms and environments. Two hundred and sixty eight novel CRESS DNA viruses were recovered and verified in this study from a range of sample types (adult Odonata, Odonata larvae, Mollusca, benthic sediment, water, Oligochaeta and Chironomidae) collected in the United States of America, Australia and New Zealand. All viral genomes isolated had two major proteins encoding for a putative Rep and coat protein (CP), with major Rep motifs identified in most Reps. Phylogenetic analysis of the Reps encoded by the viral genomes highlighted that most were extremely diverse falling outside of the previously described ssDNA viral families. A top-down approach was implemented to recover CRESS DNA viruses and possible viral pathogens from Odonata and their larvae. Thirty six viral genomes were recovered from terrestrial adult dragonflies as well as the twenty four from aquatic larvae. Dragonfly cycloviruses were isolated from the some adult Odonata species which were closely related to the isolates previously described by Rosario et al. (2012). The viruses isolated in the aquatic and terrestrial ecosystems differed substantially indicating that different CRESS DNA viromes exist in both land and water. The diversity of CRESS DNA viruses in seven different mollusc species (Amphibola crenata, Austrolvenus stutchburyi, Paphies subtriangulata, Musculium novazelandiae, Potamopyrgus antipodarum, Physella acuta and Echyridella menziesi) from Lake Sarah and the Avon-Heathcote estuary both in New Zealand, were also investigated. One hundred and forty nine novel viral genomes were recovered. Two CRESS DNA genomes were recovered from molluscs which have Rep-like sequences most closely related to those found in some bacterial genomes. Sclerotinia sclerotiorum hypovirulence-associated DNA virus 1 (SsHADV-1) was originally isolated from fungal species Sclerotinia sclerotiorum in china and was later found in benthic sediments in New Zealand. As part of this study, SsHADV-1 was recovered from dragonflies (Erythemis simplicicollis, Ischnura ramburii and Pantala hymenaea) collected in Arizona and Oklahoma, USA suggesting a larger distribution of these viruses and not surprising given the near global distribution of S. sclerotiorum. Dragonfly larvae-associated circular DNA viruses (DflaCVs) that were originally isolated in Odonata larvae samples from three New Zealand lakes were later recovered from water, benthic sediment, worms and molluscs from one of the lakes initially sampled, suggesting that these viruses are ubiquitous in freshwater environments. This study has attempted to generate baseline data of CRESS DNA viruses in certain environments using NGS-informed approaches. This data was used to try and establish whether viral distribution in different samples types can potentially be explained by the food web interactions between different samples types. Although the analysis did not show any significant relationships between sample type interactions and viral distribution a few common associations between Odonata larvae and benthic sediment were evident. This was expected as the larvae live within the sediment so it could be assumed that they potentially have similar CRESS DNA viral distribution. Although the distribution of viruses varied across sample types, molluscs proved the best sampling tool for isolating largest numbers of CRESS DNA viruses in an ecosystem with extensive diversity. Overall, this research demonstrates the applications of NGS for investigating the diversity of CRESS DNA viruses. It demonstrates that some sample types such as Odonata in terrestrial systems and molluscs in aquatic environments, can be used as effective sampling tool to determine the diversity of CRESS DNA viruses in different environments as well as detecting previously isolated viruses. The CRESS DNA viruses isolated in this body of work provides baseline data that can potentially be used in future research to investigate hosts of these viruses and their interactions with hosts and potential flow in their environments.

Prior to the commencement of this project, few plant viruses had been identified from Vietnam despite virus-like symptoms being commonly observed on many crops and weeds. In limited surveys in the late 1990's, preliminary evidence was obtained indicating that potyviruses and geminiviruses were causing significant diseases. As a result, this study was aimed at developing generic PCR-based methods for the rapid detection of viruses belonging to viruses in the families Potyviridae and Geminiviridae in plant samples collected from Vietnam, and to characterise the viruses at the molecular level. Novel degenerate PCR primers were developed for the identification of begomoviruses. Using these primers, 17 begomoviruses species infecting seven crop and nine weed species in Vietnam were identified and characterised. Sequence analyses showed that ten of the viruses (six monopartite and four bipartite) were new species. Of the seven previously characterized begomoviruses, five were identified in Vietnam for the first time. Additionally, eight DNA-ß and three nanovirus-like DNA-1 molecules were also found associated with the monopartite viruses. Five of the DNA-β molecules were putatively novel. Two novel bipartite begomoviruses, named Corchorus yellow vein virus (CoYVV) and Corchorus golden mosaic virus (CoGMV), were isolated from jute plants. Analysis of these viruses showed that they were more similar to New World begomoviruses than to viruses from the Old World. This was based on the absence of an AV2 open reading frame, the presence of an N-terminal PWRLMAGT motif in the coat protein and phylogenetic analysis of the DNA A and DNA B nucleotide and deduced amino acid sequences. This is the first known occurrence of Old World viruses bearing features of New World viruses, and their presence in Vietnam suggests the presence of a &quotNew World" virus in the Old World prior to Gondwana separation. Other interesting features relating to begomoviruses identified in Vietnam were; (i) the detection of several recombination events, particularly between the newly identified Tomato yellow leaf curl Vietnam virus (TYLCVNV), and the previously characterised, Tomato leaf curl Vietnam virus (ToLCVV), (ii) the identification of new natural hosts of Sida leaf curl virus (SiLCV), Papaya leaf curl China virus (PaLCuCNV) and Alternanthera yellow vein virus (AlYVV), (iii) the first report of variation in the geminivirus stem-loop nonanucleotide sequence (CoGMV sequence was TATTATTAC rather than TAATATTAC) and (iv) the first report of different stem sequences in the stem-loop structure of two genomic components from a bipartite begomovirus, Kudzu mosaic virus (KuMV). The sequence and phylogenetic analyses of the begomoviruses and begomovirus-associated DNAs identified in this study suggested that South East Asia, and Vietnam in particular, may be a centre of begomovirus diversity. Two pairs of degenerate primers, designed in the CI gene (CIFor/CIRev) and HC-Pro gene (HPFo/HPRev), were developed for the detection of viruses in the genus Potyvirus. Using these primers, three novel potyviruses from Vietnam were detected, namely Telosma mosaic virus (TelMV) infecting telosma (Telosma cordata), Peace lily mosaic virus (PeLMV) infecting peace lily (Spathiphyllum patinii) and Wild tomato mosaic virus (WTMV) infecting wild tomato (Solanum torvum). The fragments amplified by the two sets of primers enabled additional PCR and complete genomic sequencing of these three viruses and a Banana bract mosaic virus (BBrMV) isolate from the Philippines. All four viruses shared genomic features typical of potyviruses. Sequence comparisons and phylogenetic analyses indicated that WTMV was most closely related to Chilli veinal mottle virus (ChiVMV) and Pepper veinal mottle virus (PVMV) while PeLMV, TelMV were related to different extents with members of the BCMV subgroup. The incidence of potyviruses infecting plants in Vietnam was investigated using the potyvirus-specific primers. Fifty two virus isolates from 13 distinct potyvirus species infecting a broad range of crops were identified in Vietnam by PCR and sequence analysis of the 3' region of the genome. The viruses were Bean common mosaic virus (BCMV), Potato virus Y (PVY), Sugarcane mosaic virus (SCMV), Sorghum mosaic virus (SrMV), Chilli veinal mottle virus (ChiVMV), Zucchini yellow mosaic virus (ZYMV), Leek yellow stripe virus (LYMV), Shallot yellow stripe virus (SYSV), Onion yellow dwarf virus (OYDV), Turnip mosaic virus (TuMV), Dasheen mosaic virus (DsMV), Sweet potato feathery mottle virus (SPFMV) and a novel potyvirus infecting chilli, which was tentatively named Chilli ringspot virus (ChiRSV). With the exception of BCMV and PVY, this is first report of these viruses in Vietnam. Further, rabbit bell (Crotalaria anagyroides) and typhonia (Typhonium trilobatum) were identified as new natural hosts of the Peanut stunt virus (PStV) strain of BCMV and of DsMV, respectively. Sequence and phylogenetic analyses, based on the nucleotide sequence of the entire CP-coding region of all 52 virus isolates, revealed considerable variability in BCMV, SCMV, PVY, ZYMV and DsMV. The phylogenetic analyses also suggested the possible presence of ancestral groups of BCMV, SCMV and ZYMV in Vietnam.