Clostridium difficile is the leading cause of nosocomial diarrhoea in the world, and a considerable burden to healthcare services. For colonisation of C. difficile to occur in the gut of an individual, the resident gut flora must first be quantitatively or qualitatively altered, normally through antibiotic treatment for an unrelated infection. At present, broad-spectrum antibiotics such as vancomycin and metronidazole, the frontline choices for the treatment of C. difficile infection; disturb the normal gut flora, suffer from poor recurrence rates, and have received reports of sporadic emergence of resistance. Development of novel narrow-spectrum antimicrobials would circumvent these problems but depend on the identification of novel essential genes. Molecular techniques available to identify and study essential genes in other organisms have not yet been applied to C. difficile. In this study, we identified 208 candidate essential genes via in silico analysis based upon similarity to known Bacillus subtilis essential genes. In order to provide experimental evidence of essentiality, we developed a novel method utilizing partial diploids and functionally characterised three C. difficile genes as essential; CD0274, metK, and trpS. This method not only identified CD0274, a candidate narrow-spectrum drug target and the first essential genes in C. difficile, but also provides a reliable method to identify further essential genes for novel antimicrobial targeting. In addition, we developed a lac repressor system, a rationally designed theophylline-responsive riboswitch, and most importantly, a Tet inducible promoter system able to conditionally control expression of a catP reporter through a wide dynamic range in both C. difficile and C. sporogenes, to maximum induction factors of 192.89 and 1,275.63, respectively. The combinations of these characteristics make this Tet system not only a powerful tool for identifying essential genes, but bestows a great potential for further analysing gene function far beyond the scope of this project.

Staphylococcus aureus is an important human pathogen. The emergence of multiple antibiotic resistant bacteria and lack of new antibiotics has highlighted the need for better understanding of staphylococcal physiology, molecular biology and virulence. In S. aureus the agr quorum sensing (QS) system is a global regulator of virulence. In the agr system an autoinducing peptide (AIP) activates the histidine protein kinase (HPK), AgrC, leading to a switch from the production of colonization factors to exotoxins. The S. aureus agr system has diverged such that there are four different agr groups, each with a distinct AIP capable of activating its cognate AgrC but inhibiting the AgrC of the other groups. To investigate the molecular basis for the recognition of AIPs by AgrC, transmembrane topology modelling together with site-specific mutagenesis were used. The transmembrane topology of AgrC was predicted to consist of six transmembrane helices (TMHs) and three extracellular loops with both the N- and C-terminals on the cytoplasmic side. Since AIP-1 and AIP-4 differ by a single amino acid residue, the S. aureus AgrC1 and AgrC4 proteins were compared to identify extracellular amino acids likely to be involved in AIP recognition. Site-specific mutagenesis was used to exchange the key AgrC1 and AgrC4 amino acid residues in extracellular loops 1 and 2. The response of these AgrC mutants was evaluated using a novel bioluminescent AIP reporter. The data obtained showed that differential recognition of AIP-1 and AIP-4 depend primarily on three amino acid residues in loop 2, but that loop 1 plays an essential for activation but not for inhibition of AgrC. The data obtained also revealed that a single mutation in the AgrC1 loop2 results in conversion of (ala5)AIP-1 from a potent antagonist to an activator, essentially forcing the evolution of a fifth agr group. Attempts to identify AgrC in the cytoplasmic membrane using Western blotting failed, but data obtained using an N-terminal gfp tag showed that AgrC is evenly distributed through out the membrane. Since the processing of AgrD by AgrB to generate an AIP requires at least 3 steps - two endopeptidase steps and the formation of a thiolactone bond to form the macrocycle, it is likely that other proteins are involved in the processing of AgrD and export of the AIP. To identify potential AgrB partners, yeast two hybrid assay was employed which revealed a potential role for the putative ABC transporter Rlp in the processing and/or secretion of AIP. In summary, the data presented define the key amino acid residues involved in AIP/AgrC interactions and imply a role for proteins such as Rlp in AIP synthesis and export.

Mating-type and other ‘sex-related’ genes in the filamentous ascomcyete genera Aspergillus and Penicillium, were examined to investigate the potential sexual capacity of supposedly asexual species and also the possible evolutionary route and ancestry of mating strategy and mating-type genes. Two heterothallic and one homothallic sexual species were screened to determine the presence and genomic organisation of mating-type genes. An additional gene has previously been detected in Neosartorya fischeri, N. fumigata and Penicillium marneffei. This gene was also detected and sequenced in the heterothallic species, Emericella heterothallica and the homothallic species, Eurotium repens. The expression of this gene was investigated under conditions that cause expression of mating-type genes in these species. Mating-type and other ‘sex-related’ genes were investigated in asexual Aspergilli that have been genome sequenced. Expression of mating-type, α-factor pheromone precursor, pheromone receptor and two transcription factor encoding genes were also investigated. Gene expression varied between species, but no genes displayed mating type-dependent expression. Previous studies had developed a degenerate PCR diagnostic approach to identify putative MAT1-1-1 and MAT1-2-1 gene fragments. This degenerate PCR diagnostic was performed on Penicillium species in the subgenus Penicillium to determine the presence or absence of mating-type genes. Mating-type gene fragments or whole open reading frames were sequenced from four of these Penicillium species. RT-PCR analyses were also performed on these species, and MAT1-1-1 and MAT1-2-1 gene expression was confirmed in three of the four Penicillium species. The overall structure of the mating-type loci and idiomorphs of the Aspergillus and Penicillium species revealed certain common features. The ancestral mating strategy of the Eurotiomycetes has been suggested to be homothallism. Whilst this remains possible, alternative evolutionary scenarios are suggested from this investigation.

The integrity of all genomes is under constant threat, with DNA double strand breaks being particularly dangerous. Double strand breaks can be repaired by homologous recombination, a process catalysed by recombinase proteins of the RecA family. The archaeal recombinase, RadA, is homologous to eukaryotic and bacterial Rad51/RecA. Euryarchaea encode an additional Rad51/RecA homologue, RadB. RadB shares homology with the core domain of RadA and has been shown to bind both single and double stranded DNA, binds ATP and possesses a very weak ATPase activity. However, RadB does not catalyse strand exchange. RadB has been shown to interact with RadA, a Holliday junction resolvase (Hjc) and a DNA polymerase (PolD), suggesting a role in recombination. In this study, radB was deleted from the halophilic archaeon, Haloferax volcanii. 'delta' radB strains were slow growing, sensitive to mitomycin C and UV irradiation, and deficient for both crossover and non-crossover recombination. Deletion of radA results in similar phenotypic characteristics, and complete abrogation of recombination. Strains deleted for both radA and radB are equally defective as 'delta' radA strains, demonstrating that RadA is epistatic to RadB. A suppressor of 'delta' radB was isolated and identified as a mutation in the polymerisation domain of RadA (RadA-A196V). radA-A196V suppresses the slow growth, crossover and non-crossover recombination defects associated with 'delta' radB, as well as UV and mitomycin C sensitivity phenotypes. On account of the nature of this suppressor, the observed interaction between RadA and RadB, and the epistatic relationship between RadA and RadB, a role for RadB as a recombination mediator protein is proposed. Finally, strains were deleted for hjc. 'delta' hjc strains exhibit no growth, crossover and non-crossover recombination defects and no UV and mitomycin C sensitivity. This suggests that another, as yet unidentified, Holliday junction resolvase is encoded by Haloferax volcanii.

Replication protein A (RPA) is a single-stranded DNA-binding protein that is present in all three domains of life. The roles of RPA include stabilising and protecting single- stranded DNA from nuclease degradation during DNA replication and repair. To achieve this, RPA uses an oligosaccharide-binding fold (OB fold) to bind single- stranded DNA. Haloferax volcanii encodes three RPAs – RPA1, RPA2 and RPA3, of which rpa1 and rpa3 are in operons with genes encoding associated proteins (APs). The APs belong to the COG3390 group of proteins found in Euryarchaeota and feature an OB fold. Genetic analysis of deletion mutants was employed to determine if all three RPAs are essential for cell viability, and if there is an element of redundancy between RPA1 and RPA3. The hypothesis that the RPAs form a complex with their respective APs, as opposed to a heterotrimeric RPA complex, was also investigated. Furthermore, it was tested whether the RPAs and their respective APs are specific for each other, or whether they are interchangeable. The genetic analysis showed that RPA2 is essential for cell viability, but that neither RPA1 nor RPA3 are. The rpa3, rpa3ap and the rpa3 operon deletion mutants showed sensitivity to DNA damage but only a slight growth defect. By contrast, the rpa1, rpa1ap, rpe and rpa1 operon mutants did not show any DNA damage sensitivity and an even milder growth defect. The double rpa1 rpa3 operon deletion was difficult to generate but unexpectedly lacked a significant DNA damage sensitivity and growth defect. The inability to make the double rpa1 rpa3ap and rpa1ap rpa3 deletion mutants suggests that the APs are specific for their respective RPAs. Biochemical analysis involving histidine-tagged RPAs and APs was used to confirm the conclusions of the genetic analysis. The RPAs did not interact with each other, but instead co-purified with their respective APs. This finding reiterates that the RPAs do not form a heterotrimeric complex, as seen in eukaryotes, but instead form a novel complex with their respective APs.

The aims of this thesis are to establish the therapeutic value of two anti-glycan mAbs produced in-house, to develop an immunisation protocol with the aim of improving the immunogenicity tumour-associated glycolipids with the intention of producing therapeutically valuable mAbs and to determine the implication of a mAb with the ability to induce apoptosis in colorectal cancer. The anti-glycan mAbs 692/29 and 505/4 have previously been produced in-house and this study aimed to determine their fine specificity using a glycan array. 692/29 displayed binding predominantly to Lewis b as well as Lewis y-containing glycans. 505/4 was discovered to bind to sialyl Lewis a as well as sialyl di-Lewis a, with no cross-reactivity with other blood group antigens. This was compared to other anti-Lewis mAbs, with differences in specificity being observed. Characterisation of 505/4 mAb distribution showed binding to 80% of colorectal tumours and low levels of binding to normal tissues by IHC, suggesting it may be therapeutically useful. This thesis aimed to assess the ability of 505/4 and 692/29 to meditate immune mediated and non-immune mediated cell death as well as to determine whether non-immune-mediated cell death would be a desirable therapeutic property. Resistance to apoptosis is one of the hallmarks of cancer cells and mAbs stimulating apoptosis may not be very effective. Alternatively, cancer cells are driven to initiate apoptosis by genomic and other aberrations thus if pro-apoptotic pathways are stimulated these cells may be more susceptible to death than normal cells. To investigate the significance of apoptosis in cancer a large tissue microarray of colorectal tumours was assessed for apoptosis and its relationship to patient prognosis. Cleaved caspase-3 is a good marker of apoptosis as it is the executioner caspase for both the extrinsic and intrinsic pathways. Immunohistochemical analysis of colorectal tumour samples revealed that a high expression of cleaved caspase-3 in tumour was associated with good prognosis in colorectal cancer. This suggested that some tumours were still susceptible to apoptotic death but some are resistant and an alternative mechanism of cell death may be an advantage in these tumours. High expression of cleaved caspase-3 in the tumour-associated stroma was also an independent marker of good prognosis in colorectal cancer. This may be because apoptosis of the tumour-associated stroma reduces the level of pro-tumour signals originating from tumour-associated immune cells and stromal cells. As the tumour microenvironment can act in an immunosuppressive and pro-tumour manner, the ability of a mAb to induce direct cell death without the need for effector cells or complement would be an advantage. Lewis y and Lewis b are blood group antigens commonly overexpressed on the surface of a range of cancers. Characterisation of effector functions of 505/4 and 692/29 demonstrated that both mAbs have the ability to mediate apoptosis by antibody dependent cellular cytotoxicity, complement dependent cytotoxicity and cause direct cell death in an oncosis-like manner. Comparison with other anti-Lewis mAbs demonstrated that a number of anti-Lewis mAbs can induce direct cell death independently of apoptosis. Thus, they could effectively target apoptotic sensitive and resistant colorectal cancers. Tumours aberrantly express glycolipids and these molecules may be involved in a number of cellular pathways. In addition a large proportion of anti-glycan mAbs, including 505/4 and 692/29 in this thesis, have displayed the ability to induce direct cell death. Therefore this thesis aimed to develop an immunisation protocol capable of increasing the immunogenicity of tumour-associated glycolipid for the production of anti-tumour glycolipid mAbs directed against ovarian cancer. This study suggests that the incorporation of tumour glycolipid into liposomes and their immunisation along with the iNKT cell adjuvant α-galactosylceramide, elicits an anti-tumour glycolipid immune response, which can yield IgG mAbs capable of binding a high proportion of ovarian cancers. In summary, this thesis confirmed specificity of 692/29 to Lewis y and Lewis b and 505/4 to sialyl Lewis a and sialyl di-Lewis a. Furthermore, this thesis demonstrated a promising tissue distribution of 505/4 in vitro. Characterisation of mAb effector functions suggest that both Lewis y and sialyl Lewis a directed mAbs have the ability to cause direct cell death, independently of apoptosis in antigen positive cells, as well as the ability to cause immune-mediated cell death. This may be an important factor in the immune-suppressive tumour microenvironment. Furthermore, this thesis provides the basis for the production of new anti-glycolipid antibodies that may also be able to induce direct cell death.

The scientific community has gathered an extremely detailed and sophisticated understanding of the genetic and molecular underpinnings of microbial communication. How these microbial communication systems arise and are maintained over evolutionary time-scales however has received relatively little attention. Some major questions remain unanswered such as; what is the function of small diffusible molecules? How does population structure affect the dynamics of social communication and what is the link between the ecology of communication and the virulence of a pathogenic population? Borrowing concepts from evolutionary theory can help to unravel these fundamental questions in the context of microbial communication as it has done in other taxa displaying social behaviours. In addition microbial model organisms in which molecular and genetic tools are abundant lend enormous power to empirical tests of evolutionary theory. This work combines both of these in an attempt to understand the evolution of bacterial communication using the model organism Pseudomonas aeruginosa and its well characterised Quorum Sensing systems. Specifically the focus is in three areas. Firstly this study reveals that the stability of bacterial signalling is vulnerable to perturbations in cost and benefit and genetic conflict. Secondly this study finds that spatial structure (biofilm vs planktonic) influences the outcome of social competition over signalling and reduces population viability. Thirdly this study finds that interspecific and intraspecific competition over public goods impose divergent selective pressures on communication.

Over the course of the last two decades, research into the role of quorum sensing (QS) in regulating diverse bacterial behaviours has exploded, and around twelve years ago, a QS network was identified in the enteropathogenic bacterium Yersinia pseudotuberculosis, which was shown to control motility and cellular clumping. This thesis seeks to expand this regulatory relationship and explore the causes and consequences of the link between QS and motility, which affects pleiotropic processes including the type 3 secretion system (T3SS) and biofilm formation. Indeed, the clumping phenotype first explored by Atkinson et al. (1999), is linked to QS-dependent regulation of the T3SS, since the deletion of several QS genes results in liquid culture biofilm (LCB) formation. This is concomitant with T3S protein secretion into culture supernatant, which occurs under normally non-inducing conditions, while deleting the T3SS structural component yscJ prevents secretion and LCB formation. De-repression of the T3SS and the development of LCBs also occurs following mutation of the flagella regulators flhDC and fliA, revealing that QS and the flagella system co-regulate LCBs. However, interestingly it was found that LCB formation and secretion also occurs following mutation of the flagella structural gene flhA. The ΔflhA mutant represents a flagella-minus strain, in which the underlying regulatory circuit mediated by FlhDC and FliA is intact, suggesting that an element of the flagella structure that depends on FlhA activity acts as a check-point governing expression of the T3SS. Both QS and the flagella system positively regulate biofilm formation by Y. pseudotuberculosis on the surface of the nematode worm, Caenorhabditis elegans. Surprisingly, the up-regulated T3SS was found to be responsible for mediating down-regulation of biofilm formation by Y. pseudotuberculosis QS mutants, since subsequent deletion of yscJ could restore biofilms to wild-type levels. This suggested that a component of the injectisome was capable of influencing cellular processes in addition to its role in secretion. In light of the link regulatory link between flagella and T3S, this raised the possibility that the injectisome could play a role in the reciprocal regulation of motility. Since the genetic regulatory network underpinning expression of the T3SS is intact in the ΔyscJ mutant, like the ΔflhA mutant for flagella, the ΔyscJ mutant can reveal the role of the injectisome structure in modulating gene expression. By phenotypic observation, it was determined that the ΔyscJ mutant displayed aberrant flagella mediated motility, swimming vigorously under conditions in which the wild-type did not, and, similar to the over-production of Yop proteins in the ΔflhA mutant, the ΔyscJ mutant over-produces flagellin. This suggests that a component of the T3SS injectisome acts as a checkpoint to regulate motility, which appears to be at the level of transcription, since the ΔyscJ mutant displays up-regulation of the flagella regulators flhDC and fliA. Indeed, the relationship between T3S and motility appears to require a direct influence on QS, since subsequent mutation of ypsI and ytbI abolishes ΔyscJ-dependent hyper-motility, the ΔyscJ mutant displays altered expression of the QS system genes. Furthermore, for the emerging transcriptional relationship between these systems, the flagella and QS mutants which are up-regulated for the production of Yop proteins also over-express the virulence regulator virF, completing the transcriptional regulatory circuit which appears to be crucial for the regulation of lifestyle choices by Y. pseudotuberculosis.

Introduction: The vacuolating cytotoxin is an important H. pylori virulence factor and is naturally polymorphic, one of the most diverse regions being the signal region (type s1 or s2). Signal type determines vacuolating activity: type s1 strains are vacuolating and are associated with peptic ulcers and gastric adenocarcinoma; and type s2 strains are non-vacuolating. Heterogeneity in VacA levels between strains also exists, and different strains produce different amounts: this is due to differences in vacA transcription and differences in protein secretion between strains. A vacA promoter and a transcriptional start point (TSP) have been identified. Sequence analysis revealed that -35 and -10 motifs are well conserved. Mutagenesis of this region resulted in a decrease in vacA transcription in vitro. However, investigation of vacA expression between 8 different H. pylori strains using real-time PCR revealed that differences in vacA transcription observed between different strains were unrelated to the putative -35 and -10 motifs. VacA type s1 and s2 signal sequences differ in the cleavage recognition site at key positions. In type s1 the more favourable serine and proline residues are at positions -3 and -6 respectively, whereas type s2 strains have leucine present at -3 and glycine at -6. In addition to VacA type, the determinants of VacA production are of importance in H. pylori pathogenesis: It has been shown that larger amounts of toxin produce greater damage in animal models. However there is limited information on the level of vacA expression in vivo and disease. Aims: To analyse the cysS – vacA intergenic regions of a large number of strains to identify motifs and natural polymorphic differences that contribute to different vacA mRNA levels between strains in vivo and in vitro using sequencing and real-time PCR. Then to investigate these regions of interest using site directed mutagenesis. Next, to determine whether naturally occurring differences in the signal peptide cleavage recognition sites between type s1 and s2 strains affect the secreted VacA levels between the two strains using isogenic mutants and ELISA and Western blotting. Finally, to determine whether there is an association between the level of vacA mRNA in vivo and pre-cancerous changes, infection density and disease outcome in infected patients. Results and Conclusions: Analysis of the cysS – vacA intergenic region revealed that differences between strains in the -35 and -10 regions were not significantly associated with levels of vacA transcription, but that other regions were associated with differences in vacA production between strains. These regions differed in vitro and in vivo, which may imply in vivo regulation of vacA transcription. Strains with a fully palindromic inverted repeat upstream of the -35 motif expressed more vacA than other strains in vitro (P=0.01). Analysis of vacA transcription from the same strains in vivo revealed that strains with an A within a second stem loop located in the 5’ UTR of the vacA transcript expressed more vacA than those with a G at this position (P=0.006). Site directed mutagenesis confirmed that replacing the G with an A in an isogenic mutant strain resulted in an increase in vacA transcript levels. As found for other bacterial secreted proteins, these results show that positions -6 and -3 are important in the VacA signal sequence and may affect signal sequence processing efficiency. The level of vacA mRNA in vivo was significantly positively associated with the severity of chronic inflammation, neutrophilic activity, glandular atrophy and infection density in the infected patient. Differences in toxin production between H. pylori strains could help explain why only some infections result in disease.

This project is an attempt to have a better understanding of the role of carbohydrates in recognition and uptake of allergens by the innate immune system. Glycosylation analysis of different allergens like Der p 1, Fel d 1, Ara h 1, Ber e 1, Der p 2, Bla g 2, Can f 1, Bromelain and Papain was made using labelled lectins that can detect specific carbohydrate moieties on proteins to reveal their pattern of glycosylation. These experiments showed that all major allergens are glycosylated and this represented a major first step towards demonstrating a link between glycosylation and allergen recognition by the innate immune system. N- and O-glycosylation patterns were predicted in different allergens and a difference in mannosylation and fucosylation was detected between allergens and non-allergenic proteins. We found that the main dominant sugars on allergens are 1,2 1,3 and 1,6 mannose, as detected by GNA lectin. We have also showed that Der p 1 and Der p 2 possess 1,3 fucose in their natural forms, thus concluding that Der p1 and Der p 2 have part of the CCDs which are epitope structures for IgE. O-glycosylation in allergens was also studied giving a better understanding of the whole glycan structure in allergens. The role of mannosylation in allergen recognition by the immune system was investigated further. Different methods, including recombinant expression, enzymatic and chemical deglycosylation, were optimised to produce glycoforms of Der p 1. A recombinant preparation of Der p 1 produced in Pichia pastoris was used as a hypermannosylated form of the allergen. These glycoforms served as useful tools in addressing the nature of glycoallergen recognition by looking at the uptake of hyper- and hypo-glycosylated preparations by DCs, with confocal microscopy, ELISA and FACS as readouts. Results indicate that deglycosylated forms of Der p 1 exhibited minimal uptake by DCs compared to the natural and hyperglycosylated recombinant allergen. Comparative analysis of the hypermannosylated preparation of Der p 1 and its natural counterpart, possessing less mannan, showed that the recombinant form was taken up more readily by DCs at 37°C and at 4°C. We also showed that these glycoforms bind to the MR subfragment CTLD 4-7-FC, the C-type lectin carbohydrate recognition domain. This binding significantly decreased when the Der p 1 allergen was deglycosylated. These results were confirmed further using confocal microscopy imaging which also showed that recombinant Der p 1 uptake is immediate, starting at 5 mins of incubation and that a higher quantity accumulates inside the DC compared to natural allergen. Recombinant and natural Der p 1 both co-localised with MR, DC-SIGN and LAMP-2 lysosomal marker, suggesting a key role for these receptors in allergen uptake and a common fate for these preparations inside the DC. Further experiments were done to show the effect of Der p 1 and Der p 1 glycoforms on TSLP secretion by epithelial cells, which is known to induce Th2 driven immune responses. The results show that TSLP secretion decreases significantly when epithelial cells are challenged with deglycosylated preparations of the same allergen. This may indicate a change in the outcome of adaptive immune responses when a deglycosylated allergen challenges epithelial cells. In conclusion, this work has demonstrated a link between allergenicity and glycosylation patterns in allergens. It therefore appears that mannosylation is the dominant sugar moiety associated with allergen uptake and recognition by humans DCs, and this is in line with MR being the main receptor involved in allergen binding by these cells.