Antiviral host factors constitutively expressed to high levels can confer intrinsic immunity. Unlike, induced antiviral responses, these pre-existing intrinsic defences can protect cells against the initial stages of virus infection. However, there has been little investigation into the existence of such defences at mucosal surfaces, major portals of virus entry. Accordingly, a role for intrinsic immunity to influenza A virus (IAV) infection of the lung and respiratory tract is yet to be established. We therefore set out to investigate the hypothesis that constitutive expression of key antiviral host factors in cells of the lung and respiratory tract may confer intrinsic immunity to IAV infection. We focused on the TRIpartite Motif (TRIM) proteins, a family known to play a role in multiple aspects of host immunity to virus infection. By analysing a cell line of lung origin restrictive to IAV plaque formation, we identified constitutive expression of TRIM22, an antiviral effector and interferon stimulated gene (ISG) product reported to restrict a number of viruses as part of an innate immune response, including IAV. Analysis of open source data, respiratory tissues, and cultured cell lines, demonstrated that TRIM22 was constitutively expressed to high levels in cells of the respiratory epithelium, independently of IAV infection. By depleting TRIM22 in a cell culture model of constitutive expression, we investigated whether constitutive TRIM22 expression confers an intrinsic defence against IAV infection. We found that TRIM22 supresses IAV gene expression, inhibiting the initiating cycle of IAV replication. Reconstitution of TRIM22 expression in a cell line deficient in TRIM22 supressed the expression of foreign reporter genes in a SPRY domain-dependent manner, suggesting a potential mechanism by which TRIM22 may restrict the infection of multiple viruses. Together, our data demonstrate that TRIM22 confers intrinsic immunity to IAV infection in the respiratory tract, thereby establishing an important role for intrinsic immunity in the regulation of IAV infection.

Tuberculosis (TB) is the leading cause of infectious disease mortality in the world by a single bacterial pathogen, Mycobacterium tuberculosis. Current TB chemotherapy remains useful in treating susceptible M. tuberculosis strains, however, the emergence of MDR-TB and XDR-TB demand the development of new drugs. Enzymes involved in mycobactin biosynthesis, low molecular weight iron chelators, do not have mammalian homologues; therefore they are considered potential targets for the development of new anti-TB drugs. The aims of this study were to identify potential inhibitors and to investigate the function of the mbtG and AmbtE and AMbtF genes during mycobactin biosynthesis. The full length of mbtB and the ArCP domain were successfully cloned and post-translationally modified by MtaA, a broad phosphopantetheinyl transferase from Stigmatella aurantiaca, using Escherichia coli. Inhibitors identified by virtual screening as well as 13 chemically synthesised PAS analogues were initially investigated in whole-cell assay against Mycobacterium bovis BCG Pasteur. Seven of these compounds had interesting growth inhibition under ironsufficient conditions. The mbtA gene was cloned and expressed as soluble protein using Mycobacterium smegmatis mc2155. Preliminary in vitro MbtA assays provided hints of its activity, although, the KM for SAL and ATP have not been determined yet. The mbtG and ambtE genes have been cloned and expressed in E. coli to further investigate their biochemical function in mycobactin biosynthesis.

The work described in this thesis identifies some properties of the major polypeptides of pneumonia virus of mice (PVM) and of turkey rhinotracheitis (TRT) virus. The PVM glycoproteins have been studied in particular detail while the results obtained with TRT virus provide a preliminary description of the polypeptides of this virus. Twelve major PVM specific polypeptides designated L, G1, G2, F1, N, 39K, 35K, M, 20K, 19K, 16K and 12K were identified. In addition PVM specific polypeptides designated 25K, 24K, 23K, 18K and 17K were sometimes detected. Monoclonal antibodies directed against the G1/G2, 39K and M polypeptides were produced. The a~ility of a monoclonal antibody to precipitate G1 and G2 suggested that these two glycosylated proteins were related and this was confirmed by tryptic peptide mapping. G2 was shown to be derived from G1 in pulse chase experiments and a similar relationship between two higher mobility polypeptides synthesized in the presence of tunicamycin was observed. The G protein may have a precursor since G1 did not appear immediately following a pulse labelling. The precursor could not however be identified. An additional minor glycosylated polypeptide of 42K was found to be related to the G protein. The F1 protein appeared to be poorly glycosylated and a difference in mobility of the polypeptide synthesized in the presence of tunicamycin did not appear to be directly due to a lack of N-linked oligosaccharides. The polypeptide migrated more slowly under non-reducing conditions but no evidence of a small disulphide bonded polypeptide was found in contrast to the situation with other paramyxoviruses. This polypeptide appeared to be the major PVM protein expressed on the cell surface and was associated with G1 and G2 as the major protein in a particulate fraction of the infected cell supernatant. Tentative relationships were suggested between the 39K, 35K and 25K polypeptides, the M and 24K polypeptides and the 20K and 19K polypeptides. This together with the observation that the 12K polypeptide was not a primary gene product suggested that there may be about 11 PVM polypeptides. The N or 39K and the 20K or 19K polypeptides were observed to be phosphorylated. Twelve possible TRT virus specific polypeptides of 150K, 129K, 95K, 83K, 57K, 45K, 38K, 35K, 3DK, 23K, 19K and 15K were identified. The 150K, 95K, 83K, 57K, 45K and 15K polypeptides were glycosylated with the latter three polypeptides showing a similar relationship to the F1,2, F1 and F2 polypeptides of paramyxoviruses. A broad glycosylated band designated the 31K polypeptide was identified that was similar to a smeared band observed on prolonged exposure of immunoprecipitates of PVM polypeptides labelled with [3H]-glucosamine. The 35K and 19K polypeptides were observed to be phosphorylated. PVM may be more closely related to RS virus than TRT virus since anti-PVM serum irnmunoprecipitated the RS virus N polypeptide but not any TRT virus polypeptides. The PVM 39K polypeptide and the RS virus P protein were recognised by a monoclonal antibody providing further evidence of a relationship between PVM and RS virus.

Genome scale ordered RNA structure (GORS) is found throughout the genome of many single stranded positive sense RNA viruses, including plant viruses. It was hypothesised that GORS may function to help evade RNAi either by preventing the generation of siRNAs or by stopping RNAi-mediated cleavage of the target. This project used Potato Leafroll Virus (PLRV) to investigate the function of GORS in plant viruses. The RNA structure of a 750nt region of the genome was modified to have less, more or the same energy as the WT sequence. The physical structure of these sequences was shown to be different using two distinct methods. Viral infectivity was tested and although all four viruses were able to replicate and spread to distal leaves, the WT virus was always able to outcompete the variant viruses in competition assays. This suggests GORS provides a distinct selective advantage to the WT virus. The effect on the siRNA response was tested using a dedicated siRNA assay. In plants, this showed that the WT sequence was more resistant to degradation by siRNAs than the variant sequences in the presence of their specific inducers. The WT inducer was also not able to cause suppression of the other targets indicating that this inducer failed to produce siRNAs or that they were not effective. The siRNA populations generated during infections were sequenced and the profiles compared. This showed that all four viruses stimulated the production of siRNAs but the location of siRNA hotspots differed. It is therefore hypothesised that GORS may function to evade the RNAi response by directing the generation of less effective siRNAs. The data presented in this thesis not only informs current work on GORS and RNA structure in viral genomes but also has wider implications for research on siRNAs and food biosecurity.

Foot-and-mouth disease virus (FMDV) causes an acute vesicular disease of domesticated and wild ruminants, and pigs. The virus is highly variable with 7 serotypes and numerous subtypes. A method was developed to sequence the complete genome of the Pan Asia O serotype FMDV that caused the UK 2001 epidemic. The sequence data was used to quantify the genetic diversity of FMDV that arose over the course of this epidemic. 197 nucleotide substitutions were observed at 191 different sites across the genome, with a ratio of non-synonymous to synonymous change of 0.09. It was estimated that the date at which FMDV first infected livestock in the UK was the 7th February 2001 (95% CI Jan 20th – Feb 19th) using coalescent methods, which is in close agreement to that generated on the basis of lesion aging at the first infected premises. The rate of nucleotide substitution during the epidemic was estimated to be 2.26 x 10-5 per site per day (95% CI 1.75-2.80 x 10-5). This was a sufficiently high rate that detailed histories of transmission pathways could be reliably reconstructed and motivated a comprehensive study of genetic changes that arose between infected premises in the Darlington area. This study highlighted the potential of tracing virus transmission between farms using genetic data. A maximum likelihood methodology was proposed to combine epidemiological data (describing the infection profile of individual farms) with the genetic data. Integration of genetic and epidemiological data reduced the number of transmission trees (describing infection transmission within this cluster) that were consistent with the genetic data from 41,472, to 1,728, of which 9 represented 95% of the total likelihood calculated. An average of 4.3 (S.D. 2.1) variant nucleotides within the complete genome, were observed between consecutive farm infections. Difficulties in identifying direct transmission events in this analysis arose mainly as a result of limited data on the extent of genetic variation on a single premises (complicated by variation in animal types, and farm sizes), and the manner in which mutations become fixed within the consensus sequence (i.e. upon replication and selection, and transmission through a ‘bottleneck’). This encouraged an analysis of the minimum mutant frequency of Pan Asia O FMDV from a single lesion sampled from naturally infected hosts which led to an estimate of 3.06 x 10-4 mutations per nucleotide sequenced (this parameter has previously only been measured in cell culture). In addition an experiment was initiated to measure the spontaneous mutation rate of O1BFS FMDV. The relevance of this analysis of the evolution of FMDV during the UK 2001 epidemic was demonstrated during the outbreak of O1BFS FMDV in Surrey in 2007, where complete genome sequencing was used in real time to trace virus spread. Genetic analysis of complete viral genomes generated in real-time revealed a chain of transmission events, predicting undisclosed infected premises, and connecting a second cluster of outbreaks in September to the initial outbreaks in August. Thus, this thesis has pioneered the use of complete genome sequencing for fine scale molecular epidemiology.

The objective. of this project was to determine the protein specificity of the cytotoxic T- lymphocyte (CTL) response to rotavirus infection in mice and to assess the rotavirus serotype/strain independent nature of this response. Previous work, involving the rotavirus group at Warwick, had shown that the outer shell glycoprotein VP7 is a major target antigen for a CTL response that is virus serotype-independent. However, that work did not cover all twelve rotavirus proteins, was confined to one strain of adult mice (C57BL/6, H- 2b) and covered only two of the fourteen VP7 serotypes (serotypes 3 and 6) (Offit et al., 1994). Recombinant vaccinia viruses expressing individual rotavirus UKtc proteins VP2, VP3, NS26 and NS12 were constructed to complete a set of recombinant vaccinia viruses covering the full complement of rotavirus proteins from the bovine UKtc strain. These were used to define rotavirus proteins eliciting a CTL response in three different mouse haplotypes. UKtc NS53 and UKtc VP7 stimulated a strong CTL response only in the H-2b MHC class I haplotype (UKtc NS53 and UKtc VP7 were restricted at H-2Db and H-2K b, respectively). Conversely, UKtc VP3 stimulated a strong CTL response in the H-2d and H-2k (but not in H-2b) MHC class I haplotypes. Work using congenic mouse strains was used to verify that the VP7 protein specific CTL response is restricted solely by the MHC class I antigens. Rotavirus RRV NS53 was not only found to elicit a CTL response in the H-2b MHC class I haplotype, similar to UKtc NS53, but also in the H-2d MHC class I haplotype. Thus, the individual rotavirus protein that elicits a CTL response not only depends on the MHC class I haplotype, but also on the actual rotavirus strain being tested. Many of the previous studies looking at the CTL response-to individual rotavirus proteins have, unlike this study, used several different rotavirus strains and, therefore, may have given an inaccurate representation of the rotavirus proteins that elicit a CTL response. Recombinant vaccinia viruses were also used to examine the serotype/strain independent nature of the CTL response against the major target antigens. The analysis was extended to cover VP3 from two different strains, NS53 from three different strains and VP7 from seven of the fourteen serotypes. VP3 and NS53 were found to elicit a strain- dependent response whereas the serotype-independent nature of the CTL response to VP7 was confirmed. Since the serotype-independent nature of the rotavirus VP7-specific CTL response was found to cross-protect between half of the VP7 serotypes, irrespective of the immunising. serotype, it would be reasonable to speculate that the CTL response is serotype-independent between all the VP7 serotypes. Finally, recombinant vaccinia viruses were used to locate CTL epitopes on NS53 and VP7. Recombinant vaccinia virus expressing a UKtc NS53 deletant mutant (P9DM5) showed there to be at least one strain specific epitope in the first 150 amino acids of the UKtc NS53 protein. Recombinant vaccinia viruses expressing four different UKtc VP7 fragments spanning 46% of this protein were examined. It was found that the fragment spanning the restriction enzyme sites at nucleotide 90 (CIaI) and nucleotide 196 (Hhal), i. e. between amino acids 13 and 48 of the mature UKtc VP7 protein, contained a serotype- dependent CTL epitope. The finding suggests that the immunodominant epitope identified in the same region of VP7 by Franco et al. (1993) was not the serotype-independent epitope.

The rotavirus replication cycle has not been fully characterised, one vital stage of virus replication involves large cytoplasmic occlusion bodies termed viroplasms. These are the sites of synthesis and replication of dsRNA, packaging of viral RNA into newly synthesized cores and the formation of double-shelled previrions. The detailed mechanism by which these events occur is poorly understood but is thought to be mediated by the non-structural proteins localised to these structures. Rotavirus gene segment 11 expresses two proteins NSP5 and NSP6 which are found in alternate open reading frames. NSP5 exists is several isoforms which differ on their level of phosphorylation. It has been shown to be essential for virus replication and localises to the viroplasms. The smaller NSP6 protein is the most uncharacterised of all of the rotavirus proteins. It has however been shown to interact with NSP5 and has been tentatively suggested to be localised to the viroplasms. To further investigate these two proteins the pET expression system was utilised to obtain purified protein which was subsequently used to generate mono specific polyclonal antisera. Studies into the function and localisation of these proteins found that both localised to the viroplasms and their relative distributions within these structures were defined. NSP6 was found to be expressed at a low level throughout the course of a rotavirus infection and in contrast to other non-structural proteins, to have a high rate of turnover. The RNA binding ability of both NSP5 and NSP6 was investigated using quantitative filter binding assays and these showed both have sequence independent nucleic acid binding ability. Studies were also conducted into the mechanism of NSP6 expression from the second open reading frame of gene 11, the results obtained being consistent with a leaky scanning mechanism of expression.

The study achieved a better understanding of the epidemiology and ecology of Banana streak virus (BSV) in Uganda, generating information useful in designing management strategies to limit the spread of BSV. An identification key for the mealybugs on Musa in Africa was constructed. Screenhouse transmission experiments identified Dysmicoccus brevipes, Planococcus citri and a Pseudococcus sp. from bananas as mealybugs able to transmit the virus. Field observations also demonstrated that mealybugs are able to spread from plant to plant. Monitoring spread of BSV in new fields revealed that both onset and rate of spread was site specific; however, it was not clear why this was the case. The first incidence of BSV infection in Rakai was 28 months after planting (MAP), but in Ntungamo it was only 6 MAP. BSV incidence then increased to 28%, 72MAP at a rate of 0.10 new infections/infected plant/month and 43%, 28MAP at a rate of 0.23 new infections/infected plant/ month in Rakai and Ntungamo respectively. BSV spreads slowly in some locations, and is therefore probably amenable to control by phytosanitary measures. These would comprise primarily of the use of virus-free planting material and roguing infected plants. BSV spread predominantly within-field once initial infections were established. New infections were more likely to occur within 10 rows/columns from an old infection. More spread occurred to plants bordering established infected fields in Ntungamo. Spread of BSV into the new fields from the surroundings suggests need for separation of new fields from old infected fields in order to delay onset of BSV and reduce BSV incidence.

Troponin C is a regulatory protein of the myofilament which binds to calcium to trigger the process of contraction. This protein exists in two isoforms, skeletal and cardiac, which are spatially and temporally regulated. Work in this project builds the primary stage of a long-term project, for using the gene transfer method to overexpress the skeletal isoform of Troponin C in cardiomyocytes. The long-term aim is to achieve complete or partial substitution of the native cardiac isoform and study the effects on contractile force produced, in normal and ischemic cardiomyocytes, both in vitro and in vivo. This project has involved designing, constructing and analyzing expression of adenoviral gene transfer vectors overexpressing the sTnC isoform. Several adenoviral vectors were generated with the wild type sTnC gene under the control of muscle-specific promoters. To facilitate analysis of protein expression and its subcellular localization, the sTnC protein was tagged with epitope tags and adenovirus generated, with this gene under the control of constitutive (CMV) and cardiac-specific (HCA) promoters. Epitope-tagged adenoviruses were expressed in vitro using mouse fibroblast (NIH3T3) cells and analyzed by western blot analysis, showing successful constitutive expression. Recombinant adenoviruses containing epitope-tagged-sTnC under the control of the human cardiac actin promoter showed cardiac-specific expression in cultured cardiomyocytes, in situ, using immunocytochemistry. The constitutively-expressing sTnC adenoviral vector showed successful expression in cardiomyocytes in culture, using northern blot analysis. A range of adenoviral vectors have been successfully generated, and constitutive and tissue-specific expression has been established for some of these vectors. Successes attained in this project have established the initial requirements to achieve the long-term goal to alter calcium sensitivity of myofilaments, by overexpression of sTnC isoform in cardiomyocytes, both in vitro and in vivo.

Staphylococcal pathogenicity islands (SaPIs), the prototype members of the family of phage inducible chromosomal islands (PICIs), are extremely mobile phage satellites, which are transferred between bacterial hosts after their induction by a helper phage. The intimate relationship between SaPIs and their helper phages is one of the most studied examples of virus satellite interactions in prokaryotic cells. SaPIs encode and disseminate virulence and fitness factors, representing a driving force for bacterial adaptation and pathogenesis. Many SaPIs encode a conserved morphogenetic operon, including a core set of genes whose function allows them to parasitize and exploit the phage life cycle. One of the central mechanisms of this molecular piracy is the specific packaging of the SaPI genomes into reduced sized capsid structures derived from phage proteins. Pac phages were classically thought to be the only phages involved in the mobilisation of phage-mediated virulence genes, including the transfer of SaPIs within related and non-related bacteria. This study presents the involvement of S. aureus cos phages in the intra- and intergeneric transfer of cos SaPIs for the first time. A novel example of molecular parasitism is shown, by which this newly characterised group of cos SaPIs uses two distinct and complementary mechanisms to take over the helper phage packaging machinery for their own reproduction. SaPIbov5, the prototype of the cos SaPIs, does not encode the characteristic morphogenetic operon found in pac SaPIs. However, cos SaPIs features both pac and cos phage cleavage sequences in their genome, ensuring SaPI packaging in small- and full-sized phage particles, depending on the helper phage. Moreover, cos-site packaging in S. aureus was shown to require the activity of a phage HNH nuclease. The HNH protein functions together with the large terminase subunit, triggering cleavage and melting of the cos-site sequence. In addition, a novel piracy strategy, severely interfering with the helper phage reproduction, was identified in cos SaPIs and characterised. This mechanism of piracy depends on the cos SaPI-encoded ccm gene, which encodes a capsid protein involved in the formation of small phage particles, modifying the assembling process via a scaffolding mechanism. This strategy resembles the ones described for pac SaPIs and represents a remarkable example of convergent evolution. A further convergent mechanism of capsid size-reduction was identified and characterised for the Enterococcus faecalis EfCIV583 pathogenicity island, another member of the PICI family. In this case, the self-encoded CpmE conducts this molecular piracy through a putative scaffolding function. Similar to cos SaPIs, EfCIV583 carries the helper phage cleavage sequence in its genome enabling its mobilisation by the phage terminase complex. The results presented in this thesis show how two examples of non-related members of the PICI family follow the same evolutionary convergent strategy to interfere with their helper phage. These findings could indicate that the described strategies might be widespread among PICIs and implicate a significant impact of PICIs mediated-virulence gene transfer in bacterial evolution and the emergence of pathogenic bacteria.