Genomic libraries of S. lavendulae and S. subrutilus were constructed in S. lividans using the technique of shotgun cloning. S. lividans is a genetically well characterised recipient for heterologous DNA, but plasmid deletions or the entry of plasmids with small inserts occurred during transformation of the DNA into S. lividans This was not due to a straightforward restriction-modif ication system as this possibility was checkea using the KC301 phage. Several gene libraries were produced, using high and low copy number plasmids, and the resultant transformants screened. The nature of deoxynojirimycin (DNJ) and its lack of microbial activity prevented use of a bioassay. However, the inhibition of cc-glucosidase by DNJ was exploited by development of a quantitative assay system f or DNJ and nojirimycin (NOJ). The assay could not only detect DNJ and NOJ-producing clones, but could also assess the titre of DNJ and NOJ in culture broths. The assay was used to demonstrate differential production of DNJ and NOJ by selected StreRtOIFYces strains of cluster 61 (the S-lavendulae species group). The assay also confirmed the effectiveness of using microtitre plates as an effective screening procedure. The microtitre screening programme generated further data and statistical treatment of the results delimited the number of isolates for further examination. No DNJ-producing colony was detected and examination of the size of the DNA inserts showed almost all were too small to contain the DNJ gene cluster. Additionally, blocked DNJ production mutants were characterised by the feeding of NOJ, one of the mutants successfully converted NOJ to DNJ.

VanS and VanR form the two-component regulatory system that controls inducible glycopeptide resistance in Enterococci. Upon induction VanS, a sensor histidine kinase, phosphorylates itself on a conserved histidine residue. This phosphate is transferred to a conserved aspartate residue on VanR. The phosphorylated form of VanR is the transcriptional activator of the vanAHX genes which, when expressed, directly confer vancomycin resistance. VanS also possesses VanR phosphatase activity, providing a mechanism by which to repress vanAHX transcription. This thesis describes approaches used towards the crystallisation of VanS and VanR. These are based upon previous crystallisation studies resulting in full-length VanS crystals which diffracted to 8Å, as well as a cytoplasmic structure of an analogous histidine kinase from Thermatoga maritima (Marina et al., 2005). Full-length and truncated forms of VanSA and full-length VanRA were cloned, expressed in E.coli and purified for crystallisation studies. Autokinase activity was biochemically characterised using radiolabelling and spectrophotomic assays, in tandem with a novel application of mass spectrometry. Site-directed mutagenesis of VanSA, led to the observation that ATP hydrolysis may occur independently of the autokinase function of VanSA. Adenosine 5’ tetraphosphate was also discovered, as a novel product of VanSA. Based upon these data an expanded model for VanS autokinase activity has been proposed. This may be expanded to include the phosphotransfer and phosphatase mechanisms and validated through the measurement of associated product formation. Finally a new mechanism for the control of the VanRS two component system has been proposed. Future studies will validate and expand this model. This work has significantly increased our knowledge of this system providing the tools and foundations that will lead to the elucidation of the way that this two component system functions. This has the potential for the development of novel inhibitors that either complement or supersede existing therapies.

Methylocella silvestris BL2 is a moderately acidophilic facultative methanotroph isolated from forest soil in 2003. Uniquely, it has the ability to grow on a wide range of multi-carbon compounds in addition to methane. An analysis of growth conditions identified the requirements for robust and predictable growth on a wide range of substrates. A simple and effective method of targeted mutagenesis was developed, which relies on electroporation with a linear DNA fragment, and several strains with deletions of key enzymes were constructed using this method. Deletion of isocitrate lyase demonstrated that this enzyme is required for growth on both one-carbon and two-carbon compounds. The second enzyme of the glyoxylate cycle, malate synthase, was shown to be essential for growth on two-carbon compounds. However, surprisingly, deletion of glyoxylate cycle enzymes had a dramatic effect on expression of methanol dehydrogenase. Possible causes of this effect are discussed. Surprisingly, M. silvestris was able to grow on propane and the presence and expression of a gene cluster encoding a putative propane monooxygenase was confirmed. This enzyme was found to be a second soluble diiron monooxygenase (SDIMO) with homology to the propane monooxygenase from Gordonia TY5, identifying M. silvestris as the first known methanotroph to contain SDIMOs from more than one group. Deletion of these enzymes in turn was used to determine the requirement for each during growth on methane or propane. The soluble methane monooxygenase (sMMO) was found to be capable of oxidising propane, whereas the propane monooxygenase (PrMO) was unable to oxidise methane. However, although a strain lacking the PrMO was capable of growth on 2.5% (v/v) propane, it was unable to grow on this gas at 20% (v/v), and at 2.5%, assimilation into biomass was less efficient in comparison to the wild-type. Evidence is presented that products of oxidation of propane by the sMMO may be toxic to the cell or inhibitory to growth in the absence of the PrMO. Both the sMMO and the PrMO were found to be capable of oxidation of a wide range of aliphatic and aromatic compounds, including xenobiotics, suggesting a possible role in bioremediation. M. silvestris BL2 was found to oxidise propane at both terminal and sub-terminal positions, resulting in 1- propanol and 2-propanol respectively, and biochemical methods were used to assay the enzymes of terminal and sub-terminal pathways. Assimilation of 1-propanol was found to be by the methylmalonyl-CoA pathway, and the data suggested that 2- propanol was oxidised to acetone and acetol. The final gene of the PrMO genecluster, predicted to encode a flavin adenine dinucleotide (FAD)-containing enzyme with homology to characterised membrane-bound D-gluconate dehydrogenase from Gluconobacter spp., was found to be essential for growth on 2-propanol and acetone and may be involved in the oxidation of acetol during propane metabolism by the sub-terminal pathway.

The cyanobacterium Acaryochloris marina was isolated in 1996 and solved a 50 year old mystery as to the origin of the pigment chlorophyll d, which was thought to be a pigment of red algae or a breakdown product of the universal chlorophyll, chlorophyll a. Over the next decade, new Acaryochloris spp. were isolated from all over the world as the genus received international interest from the scientific community, with the majority of research directed towards understanding the mechanisms of photosynthesis of this uniquely pigmented cyanobacterium, using A. marina as the model organism of the genus. During this project, characterisation of different aspects of photosynthesis in Acaryochloris spp. was performed including an investigation of pigment adaptation and composition and the growth and characterisation of A. marina biofilms. However, the main focus of the thesis concerns the isolation and characterisation of cyanophages A-HIS1 and AHIS2, which infect A. marina as a basis to investigate and understand the impact of phage on host physiology in this new model system. A-HIS1 and A-HIS2 were characterised by their morphology, growth behaviour and genomes. Experiments were designed and implemented to investigate interactions between the phages and host. Interestingly, an analysis of novel genes in these phages revealed a surprising evolutionary history of phages A-HIS1 and A-HIS2 providing new insights into the origin of DNA polymerase, which is found only in the mitochondria of eukaryotes.

Siderophores are small, high-affinity ferric iron chelators released by many microorganisms and some plants to solubilize iron. They are of great interest due to their clinical use to treat iron overload in humans, and also in relation to the development of novel antibiotics that target the biosynthetic and uptake pathways for iron in pathogens. Pathogens such as Bacillus anthracis excrete more than one type of siderophore. This is linked to increased pathogenicity. The Gram-positive soil bacterium Streptomyces coelicolor A3(2) excretes three siderophores: desferrioxamine B, desferrioxamine E and coelichelin. These displace iron from insoluble ferric hydroxides, and the resulting ferric complexes are transported into the cell via siderophore-binding proteins (lipoprotein receptors) associated with ATP-binding cassette (ABC) transporters. Previous studies showed that some of the genes in the biosynthetic clusters of the desferrioxamines (des) and coelichelin (cch) were required for efficient uptake of ferrioxamine E and ferri-coelichelin respectively and a third ABC transporter gene cluster (cdt), not associated with siderophore biosynthesis genes, was implicated in the import of ferrioxamine B. In this study, the lipoprotein receptors encoded within the des, cch and cdt clusters - DesE, CchF and CdtB – were recombinantly overproduced in E. coli and purified by immobilized metal affinity chromatography. Also, ferri-coelichelin was purified from cultures of S. coelicolor. The binding of the ferric complexes of the three cognate siderophores, as well as the xenosiderophores ferrichrome and ferrialbomycin, to the lipoprotein receptors was monitored by intrinsic fluorescence quenching. Dissociation constants of receptor-siderophore complexes were found to be in the nanomolar range, and a revised model of cognate siderophore transport in S. coelicolor was proposed. In collaboration with researchers at St. Andrews University, an X-ray crystal structure was solved for apo-DesE and DesE bound to ferrioxamine B, which demonstrated the similarity of DesE to other bacterial siderophore-binding proteins and the negligible conformational change on substrate binding. Ferrioxamine B also exhibited an unusual configuration not observed before in X-ray crystals of this ferri-siderophore. Also, a forcefield was constructed to model the structure and distortions ferric-tris-hydroxamate complexes, which could be used in the future to investigate the molecular basis of the tight and specific binding of ferri-siderophores to siderophore-binding proteins.

Ricin is a heterodimeric, toxic plant protein. It is able to deliver its catalytic A chain (RTA) into the cytosol of target cells. RTA crosses the endoplasmic reticulum (ER) membrane into the cytosol, masquerading as a substrate of ER-associated degradation (ERAD) to do so. Therein, RTA inactivates ribosomes. This thesis shows that RTA is prone to lose solubility in vitro near the physiological temperature and pH of target cells. This instability is hypothesised to cause RTA to misfold in the ER lumen, promoting chaperone and membrane interactions therein. Fittingly, this thesis shows that Grp94, a lumenal chaperone, promotes the toxicity of RTA, and that liposomes constructed of negatively-charged phospholipid interact with RTA in vitro. This instability resembles that of other toxic A chains that exploit ERAD. Cholera toxin A chain and pertussis toxin A chain, for instance, are also relatively unstable (Pande et al., 2007 & 2008). The Hrd1 complex now seems the strongest candidate for retrotranslocating RTA from the ER lumen (Li et al., 2010). In the cytosol, the proteasomal cap has been shown to be involved in downstream processing of RTA – enabling toxicity (Li et al., 2010). This thesis reports that, in mammals, the balance of cytosolic chaperones and their co-factors helps to dictate the success of retrotranslocated RTA, putatively by determining its escape from terminal degradation in the proteasomal core. The effect of these chaperones occurs at a stage beyond access of the toxin subunit to the ER, and can result in both activation and inactivation of cytosolic RTA. It has been shown that, on one hand, Hsc70 is responsible for activating RTA. Hsc70 may aid RTA in attaining an active conformation in the cytosol after retrotranslocation. Alternatively, it might supplant the effectors of its degradation. On the other hand, Hsc70 also enters RTA into a sequential triage with Hsp90. Unlike Hsc70, Hsp90 deactivates RTA. This effect is dependent upon the lysines of this toxin subunit, suggesting Hsp90 may participate in the lysine-ubiquitination of RTA. Supporting this conclusion, Hsc70 and Hsp90 can both ubiquitinate RTA in vitro. This ubiquitination can be promoted if RTA is first incubated with liposomes. This implies that RTA may be particularly vulnerable to ubiquitination during retrotranslocation, where it might also be partially solvated by phospholipid. Contrasting to RTA, Hsp90 actually aids the toxicity and dislocation of ER-retrotranslocating cholera toxin A chain (Taylor et al., 2010). It seems that RTA may have fortuitously evolved to exploit Hsc70 rather than Hsp90 to promote its cytosolic activation. Provocatively, the hydrophobic C-terminal tail of RTA demarcates it from a homologous toxin, saporin, which does not exploit ERAD to achieve toxicity. Indeed, this region may be an adaptation RTA has acquired to promote interaction of the toxin subunit with Hsc70, which even seems to occur in RTA’s native, folded state. This interaction may be another reason why the region is apparently significant to the cytotoxicity of the protein (Simpson et al., 1995) Finally, because the reactivation of RTA after retrotranslocation involves proteins with broad specificity (Hsc70, Hsp90), this thesis hypothesises that this pathway may operate for other ERAD substrates. Prior investigators have shown isolated examples of this phenomenon. For example the degradation-independent retrotranslocation of extracellularly-applied luciferase (Giodini & Cresswell, 2008) and of endogenous calreticulin (Afshar et al., 2006). This thesis hypothesises that the success of a protein in being reactivated post-dislocation will be determined by stringency of the cell’s chaperone network, the propensity of the substrate to be degraded, and its propensity to refold. As a protein which is toxic to the cell when refolded in the cytosol, RTA will be a useful tool to investigate this putative, broadly relevant, post-dislocation activation pathway.

G protein-coupled receptor (GPCR) signalling cascades are a highly conserved and important method of regulating a wide range of physiological processes and responses. The conserved nature of GPCR signalling throughout eukaryotic cells allows the use of model systems such as the pheromone-response pathway in the fission yeast Schizosaccharomyces pombe, to investigate the intricacies of the signalling interactions without the complexities associated with higher eukaryotic cells. This thesis describes the use of Sz. pombe to investigate the signalling characteristics and regulation of G protein α-subunits and the further characterisation of events within the pheromone-response pathway of Sz. pombe. Human Gα subunits were found not to couple to the Sz. pombe pheromone-response pathway as they failed to reach the plasma membrane. Such localisation is essential for the functions of the endogenous Gα subunit, Gpa1, which is targeted to the plasma membrane via myristoylation and palmitoylation. The Gpa1 N-terminal region was also found to influence volume of cells by influencing the length of the cell cycle. The regulation of Gpa1 activity was investigated by mutational analysis of residues with predicted involvement in Gα activation and GTP hydrolysis. Mutations within the conserved G5 loop of the nucleotide-binding pocket resulted in enhanced levels of spontaneous Gpa1 activation, due to destabilisation of GDP binding. Attenuating GTP hydrolysis was found to have both positive and negative effects upon Gpa1 signalling, leading to the hypothesis that each Gpa1 protein can only activate one effector protein per bound GTP molecule, before entering a GTP-bound inactive state. Subsequently, the acceleration of GTP hydrolysis by the regulator of G protein signalling (RGS) protein Rgs1, acts to enhance signalling at high levels of pheromone stimulation by freeing Gpa1 from this GTP-bound inactive state, to propagate further signalling reactions. High-affinity binding of the effector protein to GTP-bound Gpa1 may constituent the GTP-bound inactive state of Gpa1, necessitating the hydrolysis of GTP to dissociate the complex. Investigating the interactions of pheromone-signalling components revealed the importance of Ral2 in coordinating the activities of Gpa1 and Ras1 within the pheromone response pathway and suggested that Rgs1 may be the basis of a signalling complex at the plasma membrane. Additionally, the dominant negative activity of inactive Gpa1 mutants suggested the existence of a Gβγ-like protein within the pheromone-response pathway. The results presented in this thesis should allow further development of the Sz. pombe model system for investigating GPCR signalling reactions.

Photosynthetic picoeukaroytes (PPEs), defined here as single celled organisms <3 μm in diameter, are significant contributors to primary production. Until recently, marine PPEs had received relatively little research attention in contrast to the more numerous picocyanobacteria. Molecular studies have now started to reveal the diversity of this group, using both the nuclear 18S rRNA gene and the plastidtargeted 16S rRNA gene as taxonomic markers. The latter marker has the advantage of directly targeting the PPE community, counteracting the problem of heterotrophic sequences dominating clone libraries. As well as PCR based molecular approaches, genomic studies of PPEs are starting to reveal the metabolic capabilities of these organisms. In this thesis, taxonomic information obtained on two flow-sorted PPE populations (Euk-A and Euk-B) showed that pico-prymnesiophytes, largely representing lineages with no close cultured counterpart, dominated the Euk-A and Euk-B libraries (54 and 58%, respectively) in tropical and sub-tropical waters of the Atlantic Ocean. Radiotracer work performed elsewhere had shown these PPE groups contribute up to 19% and 38% (Euk-A and Euk-B, respectively) to total CO2 fixation, demonstrating the importance of these PPE groups in marine carbon cycling. To further assess the taxonomic composition and distribution of these Euk-A and Euk-B PPE populations at the ocean-basin scale, clone libraries were constructed along an Atlantic Meridional Transect (AMT18). Major components of these flow cytometry sorted PPE populations were Prymnesiophyceae and Chrysophyceae using plastid markers, or Prasinophyceae and Dinophyceae (nuclear markers) including several lineages with no cultured counterparts. In surface waters a latitudinal diversity gradient was observed with a peak in PPE diversity found in the equatorial region. Distribution patterns of specific PPE groups and OTUs were subsequently correlated with measured environmental parameters, although most of the variation in PPE diversity was not explained by the measured variables. Attempts were undertaken to obtain into culture novel PPEs, especially those representative of oligotropic regions. However, the majority of isolates obtained were related to Prasinoderma or Chlorella which are cosmopolitan, fast-growing genera. Even so, some isolates more relevant of open ocean environments were obtained, including a clade VIIA prasinophyte and a Pelagomonas sp. Trancriptomics was used to further assess the functional potential of specific PPE populations, firstly in cultures using both an Ochromonas sp. and a prasinophyte as being representative of organisms present along AMT18. This approach revealed a C4 carbon concentrating mechanism in the clade VIIA prasinophyte and enzymes required for a functioning urea cycle in the Ochromonas sp. A pipeline was also developed to undertake a metatranscriptomic approach on a flow cytometrically sorted PPE population from the south Atlantic gyre. This approach revealed a diatom-like C4 carbon concentrating system in the metatranscriptome. Overall, this thesis has given new insights into the diversity of specific PPE groups at the ocean basin-scale, developed a new pipeline for the transcriptomic analysis of PPEs both in culture and in the environment, and in so doing has provided new information on the functional potential of these important photosynthetic organisms.

Despite being closely related viruses with similar replication cycles, HIV-2 replicates more slowly than HIV-1 and produces fewer particles, resulting in a lower plasma viral load. Expression of the major structural gene, gag, from HIV-1 and HIV-2 proviruses was compared to investigate whether this could play a role in the difference in particle production observed between HIV-1 and HIV-2 infection. Using quantitative RT-PCR, significantly less full-length HIV-2 gag mRNA was found to be transcribed from its provirus than for HIV-1. Sub-cellular fractionation allowed us to determine HIV-1/2 gag mRNA levels in the nucleus and cytoplasm throughout a time course. RNA export of HIV-2 gag mRNA was shown to be slower than for HIV-1 gag mRNA. HIV-2 full-length gag RNA was shown to be translated much less efficiently than HIV-1 in a range of cell lines. Both HIV-1 and HIV-2 Gag have been proposed to be translated by internal ribosome entry. Shutting down capdependent translation (by poliovirus-mediated eIF4G cleavage) significantly reduced translation from both HIV-1/2 gag RNAs, with no evidence of compensatory IRES activity. This suggests that cap-dependent translation is the predominant mechanism for translation of both HIV-1 and HIV-2 RNA. Additional work explored HIV RNA-protein interactions by UV cross-linking experiments using cellular proteins. Several proteins differentially binding to HIV-1/2 5’ UTR RNAs were identified and, in particular, a 45 kDa protein binding only to the HIV-1 5’ UTR. Attempts were made to characterise the proteins binding with different affinities to HIV-1 and HIV-2 RNAs. Confocal microscopy was used to visualise HIV-1/2 Gag expression within the cell. Both HIV-1 and HIV-2 Gag expression was shown to be reduced when siRNA was used to inhibit the cellular clathrin adaptor protein AP-1. In conclusion, HIV-2 Gag gene expression was found to be less efficient than HIV-1 at the level of transcription, RNA export and translation. Future work will continue to investigate the mechanisms behind these differences.

The bacterial twin-arginine translocation (Tat) pathway is able to export pre-folded cofactor containing proteins across the cytoplasmic membrane. Tat substrates bear cleavable N-terminal signal peptides that are characterized by the presence of a critical and highly conserved twin-arginine motif which lends the Tat pathway its name. In Escherichia coli and many other Gram-negative bacteria, three integral membrane proteins: TatA, TatB and TatC are essential for Tat-dependent translocation. In contrast Bacillus subtilis possesses a simpler TatAC system which lacks the TatB component. In E. coli the TatA protein assembles into homo-oligomeric complexes that vary considerably in size. The TatA proteins found in B. subtilis do not exhibit the same degree of heterogeneity and this suggested mechanistic differences between the Tat pathways of Gram-negative and Gram-positive bacteria. How the Tat system works is still poorly understood, and the work presented in this thesis sought to gain insights into the assembly and mechanism of E. coli and B. subtilis Tat pathways. This work focused on the study of two previously uncharacterized components: the E. coli TatA paralog TatE subunit and B. subtilis TatAc subunit. In this thesis the purification and characterization of E. coli TatE complexes is reported. Using analytical gel filtration chromatography, blue-native gelelectrophoresis (BN-PAGE) and single-particle analysis of purified TatE complexes, it was found that the TatE complexes are more discrete than the highly heterogeneous TatA complexes. This finding, together with the ability of TatE to support the translocation of the 90-kDa TorA protein, suggested alternative translocation models in which single TatE complexes do not contribute the bulk of the translocation channel, similar to the B. subtilis model. In addition, co-purification and BN-PAGE experiments demonstrated for the first time that TatE interacts with TatA to form TatAE mixed complexes in the membrane, and reveals a completely novel form of Tat complex that might be functionally significant. A soluble population of TatE was also identified in E. coli cell extracts, and phase separation experiments using Triton X-114 suggested it may be mis-localized. In a separate set of studies, the ability of the B. subtilis TatAc protein to form active translocases in combination with the B. subtilis TatCd or TatCy proteins was investigated for the first time. The TatAcCd and TatAcCy mixed translocases were able to translocate several E. coli Tat substrates including, TorA, AmiA and AmiC. Finally BN-PAGE and gel filtration chromatography showed that the TatAcCd and TatAcCy complexes were significantly smaller than the previously described E. coli TatABC substrate-binding complex.