Pseudomonas aeruginosa is an opportunistic pathogen responsible for a number of different human infections, and is the leading cause of mortality in cystic fibrosis (CF) patients. P. aeruginosa infections are difficult to treat due to a number of antibiotic resistance mechanisms and its propensity to form multicellular biofilms. It also uses a complex quorum sensing (QS) signalling mechanism to regulate virulence, but mutants in a key QS regulator (LasR) are often prevalent in CF clinical strains. Different strains of P. aeruginosa compete to dominate infections and one way they achieve this is to produce chromosomally encoded bacteriocins, called pyocins. The major classes of pyocins are the soluble (S-types) and the tailocins (R- and F-types). This study investigated the distribution of six S-type and three major groups of R-type pyocins in CF clinical isolates and their roles during strain interactions. Competition assays between strain pairs in both planktonic and biofilm modes of growth were performed between clinical strains and corresponding R-pyocin deletion mutants. Each clinical strain produced one R-pyocin but the distribution of S-pyocins was random. R-pyocins were central to strain dominance as evidenced by the reversal of competitive advantage in null-R-pyocin mutants both in planktonic and biofilm states. R-pyocins also demonstrated novel anti-biofilm activities. Genomic analysis of the most competitive strain (A026) showed that it is a lasR mutant, phylogenetically related to Liverpool Epidemic Strains (LES). Promoter fusion assays to study R-pyocin gene expression showed an increase in the expression of R-pyocin genes in a lasR mutant of PAO1 compared to the wild type. This LasR-linked pyocin expression was RecA-independent. Overall these findings establish the crucial role of R pyocins in P. aeruginosa strain competition and a link between QS and R-pyocins.

It is now well established that exercise of a strenuous and/or prolonged nature can lead to transient perturbations in immune function. The clinical significance of participating in such acute bouts of exercise in combination with other life stressors (e.g. inadequate nutrition) may be an increased incidence of upper respiratory illness (URI) (e.g. sore throat, runny nose). Many proposed nutritional countermeasures to exercise-induced immune dysfunction have been shown to be ineffective. The aims of this thesis were to determine the effects of bovine colostrum (COL) on in vitro and in vivo measures of immunity taken at rest and/or following prolonged (≥ 2 h) exercise and the incidence of URI during regular training in active males. Study 1 (Chapter 3) demonstrated that acute COL supplementation improved the recovery of bacterial stimulated neutrophil degranulation and enhanced salivary lysozyme concentration following 2.5 h of cycling. There was also greater fMLP-stimulated oxidative burst throughout the COL trial compared to PLA. These effects are suggested to be partly due to components and/or metabolites of COL that become bioavailable following digestion of the supplement which may explain why Study 2 (Chapter 4) demonstrated a greater effect on fMLP-stimulated oxidative burst with 4 weeks of COL supplementation. Study 3 (Chapter 5) found a lower proportion of URI days and lower number of URI episodes with 12 weeks of COL supplementation. Although there was no effect on selected measures of in vitro immune function taken at rest (fMLP oxidative burst, salivary secretory IgA and antimicrobial peptides), COL did blunt increases in salivary bacterial load over the winter period, which may provide a novel marker of in vivo immunity. Despite no effect of prior infection with Epstein-Barr virus (EBV) on URI incidence in Study 4 (Chapter 6), those who were seropositive and undergoing COL supplementation had a lower number of URI days vii than seronegative counterparts. Study 5 (Chapter 7) observed a lack of effect of COL supplementation on the overall magnitude of an in vivo measure of T-cell-mediated immunity to a novel antigen following prolonged exercise but there was evidence that COL may increase the sensitivity of responses to such antigenic challenge.

Burns represent one of the most devastating forms of injury with infection complications representing the highest risk of mortality. The primary objective of the Bacteriosafe project was the development of a smart wound dressing that would respond to the presence of bacteria in burn wounds. The basis of this sensing system employed the use of phospholipid vesicles, containing a self-quenchable fluorescent dye. These vesicles mimic the eukaryotic cell membrane and as such are susceptible to bacterial cytolytic factors which lyse the vesicles, generating an observable and measurable fluorescent response. My primary role in this project was to identify the vesicle lysing agents secreted from the two most frequent burn wound colonisers, Staphylococcus aureus and Pseudomonas aeruginosa. We identified the small amphipathic alpha helical peptide toxins from S. aureus and glycolipid molecules derived from P. aeruginosa as the agents responsible for vesicle lysis. The identification of these molecules led to the development of two novel phenotypic assays designed to measure these important virulence factors, as discussed in chapter 3 and 4. In chapter 5 we examined the role of toxic shock syndrome toxin-1 (TSST-1) in repressing global exoprotein expression. Our results demonstrate that TSST-1 does not repress toxin secretion and strains expressing TSST-1 retain their ability to lyse vesicles. In chapter 6 we explored the use of subinhibitory oxacillin in inducing the alternative penicillin binding protein 2a (PBP2a) in community-acquired methicillin resistant S. aureus (CA-MRSA) strains to down-regulate toxicity. Previous work in the Massey lab demonstrated that the expression of the mecA gene, which encodes PBP2a, resulted in reduced toxicity in hospital-acquired (HA) -MRSA. CA-MRSA strains are considered highly toxic and have a considerably lower level of PBP2a expression. Treatment of CA-MRSA strains with subinhibitory oxacillin did result in a down-regulation of some toxins but also the up-regulation of others, highlighting the pleiotropic effect oxacillin had on virulence regulation. In chapter 7 we developed an approach that uses the genome sequences of a set of related clinical S. aureus strains to identify novel virulence loci by associating genetic polymorphisms with specific virulence phenotypes using a genome wide association study (GWAS). This analysis resulted in the identification of four novel loci which when mutated lead to a reduction in toxicity. We demonstrate that the GWAS approach is an effective method in identifying candidate SNPs which may be important in altering virulence but do highlight limitations of this approach, primarily the generation of false positives.

Dengue is a vector-borne disease. Around 2.5 billion people are thought to be at risk of infection. It is spread primarily through the Aedes aegypti mosquito, and is endemic in tropical and subtropical regions. There are four distinct serotypes which co-circulate. Whilst infection from one serotype provides homologous immunity it does not provide heterologous immunity. In this thesis we use a range of modelling techniques to examine how the epidemiological dynamics of dengue are affected by immunological interaction between serotypes and age–dependent variation in the extent to which people are exposed to the mosquito population. We initially consider transmission dynamics for multi–serotype dengue infections and present a new framework for how secondary infections are modelled. We move on to consider age–structure and introduce a method to quantify differences between seroprevalence profiles when age–independent and age–dependent transmission rates are implemented. We combine these ideas and find that parameters associated with transmission of secondary infections can interact with age–structure and affect how easy it is to detect age–dependence in seroprevalence profiles. Finally we consider how age–dependent variation in the exposure people have to mosquitoes affects the probability of an epidemic and the optimal prevention strategy that should be implemented to ensure that the introduction of isolated infections does not lead to large epidemics. Our results show it is necessary to understand the underlying dynamics of dengue and implement the correct model, as dynamics can differ substantially. They also show the importance of public health strategies to ensure that all age–groups exposure to mosquitoes is as minimal as possible to decrease the risk of an epidemic. Therefore we have found relevant results that help to further understand the dynamics of dengue.

Recent advances in biotechnology have enabled the development and large scale production of protein therapeutics, thereby providing new treatments for numerous diseases. However, oral delivery of biologics remains a challenge as macromolecules are poorly absorbed and rapidly degraded in the gastrointestinal tract. Several strategies have been studied to improve bioavailability, but to date, no safe, efficient and clinically relevant delivery system has been discovered. Here, the reprogramming of bacterial toxins for therapeutic purposes was investigated. Following cellular uptake, Pseudomonas aeruginosa exotoxin A (PE) and Vibrio cholerae cholix toxin (Cho) escape lysosomal degradation and are capable of both trafficking to the cytosol of non-polarised (NP) cells and undergoing transcytosis across polarised cells. This project aimed to investigate the trafficking routes exploited by these toxins to determine how they avoid degradation, and define their potential as vehicles for oral protein delivery. The nature of the conformational change experienced by PE in acidic conditions was first investigated. It was shown that this transition consists of a subtle alteration in the protein’s structure which does not affect its overall size, but results in the exposure of additional trypsin cleavage sites and hydrophobic residues. It was concluded that the transition is involved in the protein’s escape from lysosomal degradation and was most likely caused by the protonation of two histidine residues when the pH was lowered, resulting in the formation of additional salt bridges and thus different structural constraints. The interaction between PE and monosialoganglioside GM1 was also examined to determine whether it could act as a secondary receptor for PE. High-affinity binding was established in both acidic and physiological conditions, and interaction with GM1 was shown to be required for efficient protein internalisation by NP cells. These findings were concluded to be in agreement with GM1 acting as a secondary receptor for PE, leading the toxin away from lysosomal degradation following conformational change. The potential of PE as a vehicle for delivery of biologics was studied using several versions of the toxin connected to a cargo protein. Results showed that a truncated version of PE is capable of carrying a cargo protein (green fluorescent protein, human growth hormone) inside NP cells in vitro and across polarised epithelium in vivo. These data strongly support the idea that PE has the potential to be used as a vehicle for oral delivery of biologics. Similarly, these studies were also carried out on cholix to assess its ability to act as a drug carrier. Cholix was shown to undergo a conformational change similar to that experienced by PE. It was also demonstrated that an interaction with GM1 occurred, although in this case, increased binding appeared to result in decreased internalisation by NP cells. Finally, full-length and truncated cholix could transport a payload inside NP cells in vitro and across polarised epithelial cells in vivo. These results confirm that cholix could also represent a powerful tool for oral administration of macromolecular drugs.

The global transcriptional regulator mexT, is a mutational hotspot; the sequence variants commonly seen to co-exist within the P. aeruginosa population are: drug susceptible (e.g. PAO1) and chloramphenicol and norfloxacin non-susceptible (nfxC mutant). The nfxC phenotype, selected for on chloramphenicol agar is characterised by reduced virulence. The conversion between PAO1 and nfxC phenotypes is associated with an 8-bp repeat sequence in mexT. To investigate the effects of the 8-bp repeat on the adaptive mode of survival of P. aeruginosa, isogenic mutants were generated: PA (8-bp, two copies) and PAdel (8-bp, one copy). The mutants were characterised using phenotypic microarrays (PM), motility, antibiotic susceptibility, Galleria virulence models and RNA-seq in defined media. PM revealed differences in central metabolism indicating that PAdel/PAnfxC were associated with a biological metabolic cost. Strains with the single copy of the 8-bp sequence showed reduced motility and virulence. Transcriptome analysis revealed that mexT, in PA, consists of two regulatory elements defined by an intact helix-turn-helix motif (across the repeat region) which is capable of regulating the downstream LysR region via repressor and autoregulative mechanisms. Whole genome sequencing identified regions of compensatory mutations that were associated with differences in phenotype between PAdel (genetically modified) and PAnfxC (selected). To link phenotype and genotype and to understand the metabolic effects of this mutation, a genome wide metabolic reconstruction was performed. This revealed differences in key metabolic pathways such as glycolysis, gluconeogenesis and oxidative phosphorylation. This study has shown that an 8-bp repeat in mexT is a driver of genetic diversity. Regulatory elements linked to the effect of the 8-bp sequence on antibiotic resistance, central metabolism, chemotaxis, motility and virulence have also been identified. These methods can be used to define phenotype in any pair of isogenic mutants, at the genome level, and to investigate the clinical risk of strains.

The Type VII/Ess protein secretion system (T7SS) is found in many Gram positive bacteria, including the opportunistic human and animal pathogen Staphylococcus aureus. Previous work has shown that the S. aureus T7SS machinery comprises six core components of which four, EsaA, EssA, EssB and EssC are integral membrane proteins. EssC, the largest T7 component, is a member of the FtsK/SpoIIIE family of membrane-bound ATPases that harness the energy of ATP hydrolysis with the movement of macromolecules. The aim of this thesis was to probe the organisation of the membrane-embedded components of the secretion system and to determine whether EssC plays a role in substrate recognition. A range of detergents was tested for their ability to solubilise EsaA, EssB and EssC from the S. aureus inner membrane. The zwitterionic detergent Foscholine-12 was subsequently used for further study as it was the only detergent identified that was able to extract reasonable levels of each protein from the membrane. Blue-native PAGE analysis identified homomeric complexes of Foscholine-solubilised EsaA and EssB proteins. Crosslinking analysis in native membranes provided further evidence for EsaA and EssB homo-dimerisation, and also revealed the presence of high molecular weight multimers of EssC. Surprisingly, there was no evidence from crosslinking that any of the components interacted with each other, or with the EssA protein. Furthermore, EssC multimers detected in whole cells were not dependent upon the presence of any other Ess component or substrate protein for their assembly. The EssC protein from S. aureus strains can be grouped into four variants (EssC1-EssC4) that have sequence variability in their C-terminal domains. These variants are associated with unique clusters of candidate substrate-encoding genes. Here it was shown that each EssC variant can interact with the remaining Ess components from strain RN6390 to facilitate secretion of EsxA but not the substrate protein EsxC. Thus, EssC appears to be a specificity determinant for T7 substrate secretion in S. aureus.

To the colonising bacterium, the human body represents a number of ecological niches, some of which it could be argued are as hostile to the coloniser as even the most extreme of ex vivo environments; by the activities of the immune system, these living host niches are actively dedicated to the prevention of their colonisation. Paradoxically, the normal human microbiota of each human extends in estimated total number to approximately 1014 per human. This thesis is a compilation of published work that focuses on two anaerobic bacteria of the normal human microbiota: Bacteroides fragilis predominantly found in the large intestine; and Propionibacterium acnes predominantly found in the skin microbiota. When given the opportunity these bacteria can cross the divide between commensal and pathogen and cause infection. The papers included in this thesis address aspects of the characteristics of these bacteria that may relate to virulence and the association of these bacteria with clinical infection.

Clostridium difficile was identified as an infective agent in the late 1970s and early 1980s and causes a spectrum of disease ranging from asymptomatic carriage, mild colitis, pseudomembranous colitis (PMC) to fulminant colitis and even death. Since its recognition as an infective pathogen, C. difficile has become the principal cause of nosocomial diarrhoea in adults. The main aims of this four-part thesis were to determine the extent of Clostridium difficile infection (CDI) within the local in-patient population and to establish the epidemiology of CDI within the specialty of colorectal surgery. The first study focused on the burden of CDI to the diagnostic laboratory and the relative incidence of disease in different clinical specialties over an 8-year period (2000 to 2007) in a region that had not been affected by the hypervirulent 027 strain. A 27-fold increase in the number of faecal samples analysed by the enteric laboratory occurred from 2000 to 2006 and the total number of potential CDI cases increased over the same period, with a decline finally seen in 2007. One-fifth of all toxinpositive samples were from age groups under 60 years of age providing further evidence that CDI was not just a disease of the elderly. Although Medicine of the Elderly provided the greatest faecal analysis workload; Renal Medicine/Transplant Surgery, Intensive Care, Infectious Disease and Gastrointestinal Medicine all had higher incidences of CDI than Medicine of the Elderly. Similarly the low risk group of Paediatrics was also starting to show a small but notable increase in potential incidence. Potential excess costs for CDI in this region rose from £3.5 million to £29 million over the study period. The second study aimed to assess the potential impact of CDI within all surgical services. In the absence of 027, a further aim of this study was to assess if the more severe and extreme forms of C. difficile disease were occurring from 2000 to 2006. Colorectal surgery had the greatest number of CDI episodes followed by Upper Gastrointestinal Surgery and Urology. Despite the total number of C. difficile toxinpositive in-patients increasing each year, a similar increase was not demonstrated in the number of patients diagnosed with more severe forms of CDI or in the number of CDI patients treated with surgical intervention. In the cases requiring surgical intervention up to 40% of patients did not present with diarrhoea and up to 50% of patients did not have a C. difficile toxin-positive faecal sample prior to surgery. Demonstrating the importance of clinical recognition of the entire spectrum of C. difficile related disease. The post-operative mortality rate for fulminant CDI was 26%. High mortality figures for fulminant CDI treated surgically have not changed significantly over the last two decades and may relate to surgical referral for CDI often occurring late when the patient is in extremis. The third and fourth studies examined the specific burden of C. difficile in the colorectal surgical patient population and the environmental surface contamination within colorectal wards. An asymptomatic carrier rate of 6.1% was identified in the out-patient colorectal surgical population. Asymptomatic carriers admitted from the community play an important role in sustaining the transmission of disease within the hospital environment with 42.8% of C. difficile strains only identified in the in-patients faecal samples but not on the surface environment of the wards. Standard enteric hospital laboratory CDI diagnosis using enzyme immuno-assay for toxin A+B detection was 52% less sensitive then toxigenic culture with a false positive rate of 2.5%. Toxigenic culture identified a further 58 colorectal surgical in-patients with CDI. Of all the C. difficile isolates identified from in-patients and the surface environment ribotype 001 was the commonest strain, consistent with other local studies where ribotype 001 has emerged as the dominant strain. A large proportion of the in-patient ribotype 001 isolates showed resistance to ceftriaxone and ciprofloxacin. The ribotype 001 isolates from the surface environment showed decreased resistance to ceftriaxone compared with the in-patient strains. Similarly 4.6% of all in-patient isolates showed intermediate resistance to vancomycin but no vancomycin resistance was demonstrated in the environmental surface isolates and may represent increased development of C. difficile resistance mechanisms in the host. The patient bed frames were the commonest contaminated environmental surface with C. difficile, followed by the patient's bedside lockers and tables. Therefore the risk of a patient ingesting a C. difficile spore from the surface environment is high. Following the introduction of a new cleaning protocol during the environmental sampling study a statistically significant reduction in environmental C. difficile surface contamination and in the number of CDI colorectal in-patients was demonstrated. Acquisition of CDI from the surface environment in hospitals is not to be under-estimated and judicious application of infection control measures remains an important factor in preventing CDI transmission.

Acinetobacter spp, particularly A. baumannii, are implicated in a wide spectrum of nosocomial infections, including bacteraemia, secondary meningitis, and urinary tract infections, but their most important role seems to be as agents of nosocomial pneumonia. Until 1988, the new fluorinated quinolones showed very good activity against Acinetobacter strains. However, reports of a high prevalence of fluoroquinolone resistance among Acinetobacter isolates have appeared. Therefore, understanding the mechanisms by which this phenomenon occurs in this pathogen is important. The two target proteins of the quinolones are the DNA gyrase (topoisomerase II) and topoisomerase IV. Studies with clinical isolates of A. baumannii have shown that mutations in the gyrA and parC genes were likely to be the major mechanisms of resistance. However, alteration in drug permeability or drug efflux could also contribute to quinolone resistance in A. baumannii. Initially, the purpose of this thesis is to compare the activities of the original fluoroquinolones (e.g. ciprofloxacin) with those of novel fluoroquinolones (e.g. moxifloxacin) against 47 A. baumannii isolates collected from five Scottish hospitals. MICs were determined by a standard agar dilution method and had shown that most of the new fluoroquinolones had better activity compared to ciprofloxacin. Nine clinical resistant isolates that exhibited decreased sensitivities to ciprofloxacin (MIC > 4mg/L) were investigated for their mechanisms of resistance. The mutations, in the gyrA and parC genes that might be responsible for resistance to ciprofloxacin were examined by PCR amplification, restriction fragment length polymorphism and sequencing. It was found that all but one isolate, for which the MIC of ciprofloxacin was 4mg/L, showed a change at Ser 83- Leu in the GyrA subunit. The remaining strain had a change at Gly 81-Cys in the GyrA subunit. Two strains for which the ciprofloxacin MICs were 64mg/L showed a double change of Ser 83-Leu or Gly 78- Cyst in Topoisomerase IV subunit C. The outer membrane profiles were also investigated and revealed that one isolate for which the ciprofloxacin MIC was 64mg/L lacked a 33KDa band. In vitro mutation studies, with challenges by ciprofloxacin and moxifloxacin, showed that all resistant mutants had a mutation in the gyrA gene where Serine has been substituted by Leucine at codon 83. However no parC mutations were found, even within highly resistant mutants. In addition the outer membrane profiles showed a deletion of 20 KDa band, and this might suggest that the pathway of resistance to fluoroquinolones is different between mutants and clinical isolates. By employing a biotin- and PCR- assisted capture method, a 500 bp of unknown segment of gyrB gene was amplified and sequenced, to establish if this gene was a contributory factor towards resistance.