The human cytomegalovirus immediate early proteins, IE72 and IE86 are believed to play a key role in the viral life cycle. Both proteins are able to activate individually and synergistically a number of viral and cellular promoters. However, the mechanisms of activation by the two proteins are not clearly defined despite the identification of a number of viral and cellular proteins with which both proteins can interact. In this study I have tried to identify novel factors which can interact directly with the viral immediate early proteins. The original approach taken was the screening of a λgt11 cDNA expression library with recombinant IE72 and IE86. The screening using IE72 was unsuccessful, however, I was able to identify a number of putative interaction partners for IE86. Amongst these were two proteins which are believed to play a role in gene expression in the context of chromatin, histone deacetylase 1 (hdac1) and human brahma (hbrm). I have demonstrated that both proteins are able to bind to IE86 in vitro and have identified regions of IE86 which are important for this binding. In addition I have shown that IE86 is able to interact with hbrm in transfection studies giving synergistic activation of a cellular and a viral promoter. To identify IE72 binding proteins, a large-scale affinity column analysis was carried out using HeLa cell extract. This yielded a protein of approximately 100kDa which was able to interact with IE72. Mass spectrometry analysis of fragments of this protein suggested that the protein may be elongation factor (EF2), however, IE72 had no effect on protein translation in a rabbit reticulocyte lysate assay and an antibody to EF2 did not react with the 100kDa protein. The identify of this 100kDa is therefore still not clear. During this analysis I also examined the interaction of IE72 with the viral DNA polymerase and found that IE72 was able to interact in vitro with HCMV DNA polymerase.

Two expression systems were utilised in an attempt to produce suitable quantities of the influenza polymerase proteins for protein purification. Using the methylotrophic yeast <I>Pichia pastoris, </I>both PB1 and PB2 were expressed at around 0.1-0.4 mg/l. However, it did not prove possible to purify the polypeptides which appeared to be aggregated. The P proteins and NP were also expressed in mammalian cells using recombinant Semliki Forest viruses. However, soluble protein did not accumulate at levels sufficient for large scale purification studies. A transient "reverse genetics" assay whereby the expression of a synthetic model viral RNA segment is driven by influenza P proteins expressed by recombinant vaccinia viruses and NP from a T7 driven plasmid, was used for the analysis of mutant influenza proteins in the processes of viral transcription and replication. Recombinant vaccinia viruses expressing mutant forms of PB1 were generated and their ability to support viral replication determined. One mutant which was shown not to bind PB2 retained reduced but signifciant replication activity. In contrast, mutants that were unable to bind PA were defective in viral RNA synthesis. Of the mutants that were defective in viral transcription, one acted as a <I>trans</I>-dominant inhibitor of wild type protein function. Mutants that were unable to bind RNA were unable to support viral transcription, whilst those with reduced RNA binding capacity were similarly impaired in transcription. The mutants that were unable to bind RNA were also shown to be transdominant inhibitors of replication, indicating that they were able to compete with wild type NP for essential components of the transcription machinery. Mutations in the N-terminal region of NP previously described as the karyopherin α binding site were also analysed for their ability to support viral transcription. Neither mutation of a residue described as essential for karyopherin α binding nor deletion of the karyopherin α binding site affected viral RNA synthesis, though these mutants localised in the cell differently to wild type NP.

The eight negative-strand RNA genome segments of fowl plague virus, an influenza A virus, were polyadenylated in vitro at their 3' termini using an enzyme, poly(A)polymerase, which was isolated from E. coli. The subsequent use of oligo(dT) to prime reverse transcription of this RNA, and of mRNA isolated from fowl plague virus-infected chick embryo fibroblasts (which contains poly(A)), allowed the production of full-length DNA copies of both classes of influenza virus RNA. These single-strand cDNAs were annealed together to produce individual dsDNAs corresponding to each segment of the influenza virus genome. dsDNAs produced by this approach were used for molecular cloning in plasmid pBR322, and a recombinant plasmid containing influenza virus-specific sequences derived from genome segment 7 (the matrix gene) was isolated. Direct sequence analysis of this plasmid demonstrated the presence of additional, non-virus coded sequences at the 5' end of matrix gene mRNA. A specific restriction fragment derived from this plasmid was used for sequence analysis of the 5' region of a population of matrix gene mRNAs by the dideoxy-chain termination method of DNA sequencing, demonstrating that the additional host-derived region is heterogeneous in both sequence and length, and covers a size range of approximately 9-15 nucleotides. In addition, these analyses indicated that during transcription of matrix gene mRNA, the 3' terminal nucleotide present in virion RNA is not transcribed.

Translation of picornavirus RNA takes place by internal initiation, determined by the presence of an internal ribosome entry site (IRES) in the 5'-untranslated region of the genomic RNA. Efficient translation from the human rhinovirus-2 (HRV-2) IRES is dependent on host cell <I>trans</I>-acting factors. These include unr, p38 and polypyrimidine tract binding protein (PTB). This thesis details the investigation into how these factors act to promote translation from the HRV-2 IRES. Unr, an RNA-binding protein with five cold-shock domains (CSDs), binds to the HRV-2 IRES and this interaction was studied by crude and then fine mapping of the binding sites of unr on the IRES. The functions of the CSDs of unr were investigated by point mutation of each of the CSDs and testing the ability of these mutants to bind the IRES and stimulate translation from it. p38, a WD-motif protein with no RNA-binding activity, was expressed using recombinant baculovirus-infected insect cells. An <I>in vivo</I> interaction between unr and p38 was demonstrated, and the effect of p38 on unr's binding to the HRV-2 IRES was tested <I>in vivo. </I>After gaining insight into the complexes of unr and p38 that form on the IRES, the function of p38 in translation from the HRV-2 IRES was demonstrated. Unr and PTB were also used as tools to compare the factor requirements of the HRV-2 and poliovirus IRESs for efficient translation. Finally, an investigation was made into the cellular role of unr, in terms of the cellular mRNAs that unr binds, and those whose translation it stimulates.

The aim of this project was to investigate the role of PA in the virus life cycle. Firstly, four archived A/FPV/Rostock/34 (FPV) temperature sensitive (ts) mutants were analysed. Duplicate cDNA clones of each PA gene were sequenced and all but one had single amino-acid mutations compared to wild-type PA. One clone of ts30 PA (ts30.2) had two amino acid changes, the other only one. Two viruses (ts45 and tsmN4) were defective in plaque formation at the non-permissive temperature (NPT), while two (ts30 and ts29) were not, possibly due to reversion. All viruses synthesised normal amounts of protein at the NPT, but ts45 did not shut off host cell protein synthesis effectively. Mutants ts45 and tsmN4 failed to synthesise vRNA at the NPT. Recombinant RNPs containing ts45, mN4 or ts30.2 PA were ts for viral gene expression; both mutations seen in ts30.2 were necessary to confer this defect. In a recombinant system, ts45 displayed a specific defect in synthesis of vRNA from a cRNA template, consistent with its failure to accumulate vRNA in the authentic viral background. However, tsmN4 was unable to synthesis either vRNA or mRNA in this system. Overall these data show that PA is an essential component of the polymerase but do not support the hypothesis that it is only required for genome replication. We propose the novel hypothesis that PA induces RNAse activity that contributes to host-cell shut off and interferon antagonism.

Rotaviruses are a leading cause of acute gastroenteritis in infants and young children worldwide and possess a genome of 11 double-stranded (ds) RNA segments. Early morphogenesis of RV particles and viral RNA replication occur in cytoplasmic inclusion bodies called ‘viroplasms’, of which the viral non-structural proteins NSP2 and NSP5 are essential components. Using confocal microscopy (CM), we demonstrated association of viroplasms with lipids and lipid droplet (LD)-associated proteins (perilipin A, ADRP). LD-associated proteins were found to co-localise with viroplasm-like structures (VLS) in cells co-transfected with NSP2- and NSP5- expressing plasmids in the absence of rotavirus infection, as well as viroplasms containing NSP5-EGFP. Close spatial proximity between NSP5-EGFP and perilipin A was demonstrated by Fluorescence Resonance Energy Transfer (FRET). Time course CM studies showed increasing recruitment of perilipin A to viroplasms during the progression of rotavirus infection. Separation of RV-infected cell extracts on iodixanol gradients demonstrated co-localisation of rotavirus dsRNA, NSP5 and perilipin A in low density fractions. Chemical compounds interfering with LD formation or homeostasis (isoproterenol + IBMX, triacsin C, C75) decreased the number and size of viroplasms in rotavirus-infected cells, dsRNA replication and the production of infectious progeny virus, whilst significantly protecting cells from cytopathicity caused by rotavirus infection. These data contribute to a novel understanding of the role of LDs in rotavirus replication. Using a cell line stably expressing NSP5-EGFP, the dynamics of viroplasm formation in relation to dsRNA replication and production of infectious virus was investigated.

Murine gammaherpesvirus 68 (MHV-68) is a virus of wild rodents which provides a laboratory model for many aspects of gammaherpesvirus infection. Gammaherpesviruses have a colinear arrangement of conserved genes interspersed with unique sequences, many of which have roles in latency or pathogenesis. The aim of this project has been to characterise a locus of MHV-68 which is positionally analogous to divergent loci of other gammaherpesviruses, in order to determine whether MHV-68 encodes either analogous or dissimilar genes. 6.2kb of MHV-68 unique DNA sequence at the left end of the genome was determined. 8 sequences were identified with primary and secondary structural characteristics of transfer RNAs (tRNAs), the first report of such sequences encoded by a virus of eukaryotes. Northern analysis of lytic viral infection indicated the abundant expression of small RNAs from a DNA segment encoding only tRNA-like sequences (vtRNAs). To confirm that such lytic cycle transcripts corresponded to vtRNAs, S1 analysis was performed, demonstrating that vtRNAs 5 and 6 were transcribed into mature tRNA-like molecules. Further, vtRNAs 4 and 6 were shown to be post-transcriptionally CCA-modified in a similar way to cellular tRNAs. Investigation of vtRNA function by acidic northern analysis showed that during lytic infection in cell culture, vtRNAs 3,4,5 and 6 were not significantly aminoacylated, indicating that the vtRNAs do not function conventionally in protein synthesis. In addition, two open reading frames and a partial coding sequence were identified which had no close relatives amongst published sequences: ORF1 has low level sequence similarity to both poxvirus serpins and the complete sequence of the truncated ORF3 derived from the recently published MHV-68 genomic sequence, however ORF3 is not significantly homologous to serpin sequences. Alignments indicated that the serpin active site was not conserved in ORF1 or ORF3, suggesting functional divergence. ORF2 has no detectable homologues amongst published sequences.

The work in this thesis draws on a prior bioinformatics survey, in which it was shown that the existence of <i>cis-</i>acting RNA sequences within the coding regions of influenza A could be deduced from regions of low codon variability. To identify and characterise sequences with a putative role in packaging, synonymous mutations were introduced into the terminal coding sequences of segments 7 and 5 of influenza A/Puerto Rico/8/34 virus, targeting both conserved codons and non-conserved codons. Mutating conserved codons in segment 7 of the virus caused significant growth defects, reducing the output of plaque-forming virus by 10- to 1000-fold. The mutations were shown to affect various <i>cis-</i>acting functions, including, in all cases, disruption of genome packaging. This manifested differently in MDCK cells, where the mutations reduced the total amount of virus produced, and in embryonated chicken eggs, where virus was released in quantities similar to wild-type but with reduced incorporation of all eight genomic segments. In the latter case, it was shown that the defective mutations caused virions to be released with incomplete genomes. In contrast, the introduction of similar mutations into segment 5 had much less effect on virus growth, with most mutant viruses replicating normally and only two showing replication defects. Nevertheless, the most defective virus showed evidence of poor genome packaging. Serial passage of defective segment 7 mutants produced pseudorevertants that regained wild-type growth properties despite retaining their original mutations; preliminary work to characterise the genetic interactions in these viruses is presented. In summary, this thesis characterises a series of influenza A mutations which provide information on various <i>cis-</i>acting RNA functions, particularly genome packaging.

I present a three-dimensional structural analysis of (1) the simian immunodeficiency virus (SIV) gp120-gp41 envelope glycoprotein complex, (2) D1D2-Igαtp, a CD4-antibody construct with extreme HIV-neutralizing capability, and (3) the complex of D1D2-Igαtp with native SIV gp120, using cryo electron tomography and atomic force microscopy, in order to better understand structural determinants of HIV neutralization using this model system. Cryo electron tomographic density maps and atomic force micrographs of D1D2-Igαtp show that the 3-9 antibody-like units of this high-molecular-weight polymer appear to be more flexible than the 5-6 antibody arms of IgM, a naturally-occurring polymeric antibody that was analyzed in parallel. When imaged by cryo electron tomography in complex with the gp120-gp41 spikes of frozen-hydrated SIV virions, D1D2-Igαtp molecules were observed to bridge spikes within and between virions. Notably, these complexes were associated with an increased proportion of ruptured virions compared to control experiments. Further, inter-virus spike crosslinking was associated with close apposition of the membranes of the crosslinked virions, effectively neutralizing all spikes on the closely apposed surfaces by preventing access to the target cell surface. These effects suggest that the ratio of neutralized to CD4-liganded gp120 spikes can be much greater than unity, and imply that the potency of D1D2-Igαtp may be due to its flexibility more than to its polyvalence or steric bulk <i>per se. </i>These observations suggest polyvalence presented on a flexible scaffold as a key design criterion for small molecule and low-molecular-weight HIV entry inhibitor drugs.

<i>Erwinia carotovora </i>subsp. <i>atroseptica </i>(Eca) mutants resistant to a variety of phages were isolated and characterized, identifying lipopolysaccharide (LPS) as a frequently used phage receptor. Various strains deficient in steps in LPS assembly were isolated. One mutation mapped to horizontally acquired island (HAI) 5. Although lateral transmission of this locus to Eca has enhanced virulence in potato, it has also rendered Eca sensitive to phage infection. The genome sequence of Eca identified 17 HAIs, including 2 prophages. Genomic deletion of the prophages resulted in decreased motility, the ability of these strains to cause soft rot of potato was diminished. One phage newly isolated on Eca, φTE, was sequenced. Only 20% of the ORFs could be assigned a function. Characterization of this phage demonstrated that infection depended on the presence of a flagellum. This phage was able to mediate generalized transduction, indicating a potential use in functional genomic analysis. The generalized transducing phage that infects <i>Serratia </i>sp. ATCC39006 (φ0T8) was also shown to be flagellatropic following mutational analysis of the host. Further characterization of this phage identified an alternative and unexpected host, <i>Pantoea agglomerans. </i>It was possible to move a plasmid-encoded antibiotic resistance determinant between the two bacterial genera with high efficiency by φ0T8-mediated transduction. This φ0T8-sensitive strain of <i>P. agglomerans </i>has antifungal capacity, and is therefore of potential application as a biocontrol agent. The antifungal activity was investigated. <i>P. agglomerans </i>was found to produce an <i>N</i>-acyl homoserine lactone signalling molecule that is typically used in quorum sensing. The potential impact of quorum sensing on antifungal production was investigated via mutagenesis.