This thesis presents an analysis of the relationship between single nucleotide polymorphism (SNPs) and protein-protein interactions. The aim of the thesis is to investigate the distribution of non-synonymous single nucleotide polymorphism (nsSNPs) in terms of their locations in the protein core, at the protein-protein interface sites and on the other areas on the protein surface. The analysis used experimentally verified human protein-protein interactions and nsSNPs from the UniProt humsavar database. A further investigation was performed on a larger SNP dataset from the 1000 Genomes Project (1KGP). Both investigations identified a significant preference for disease-causing SNPs to occur at the protein interface compared to other areas on the protein surface. The three-dimensional structures of protein-protein interfaces were examined in order to propose stereo-chemical explanations for the disease-causing effect of nsSNPs in the humsavar dataset. In addition, three methodologies (i.e., usage of SNP server, structural analysis and usage of GMAF) that could help identify pathogenic variants were presented. Structural analysis was also performed on non-diseasecausing SNPs in order to investigate their possible effects on protein-protein interactions. The result showed that some of the previously classified non-diseasecausing SNPs could potentially be disease-causing SNPs. The myVARIANT program was developed. The program obtains SNPs from 1KGP, maps them to structures, evaluates their distribution on structures and performs a structural analysis. In conclusion, the thesis demonstrates the important role that protein-protein interactions play in disease pathogenesis.

Sensing and responding to the environment is a crucial function in all living organisms. The molecular mechanisms that facilitate these essential processes are however subject to a range of random effects and stochastic processes, which jointly affect the reliability of information transmission between receptors and e.g. the physiological downstream response. We employ an information theoretic framework to capture and characterise how extrinsic and intrinsic noise affect the transmission of signals along simple motifs of molecular interaction networks, and cell fate decision making pathways. We observe that the biomolecular information processing efficiency is profoundly but differently affected by the various sources of noise. In particular extrinsic variability is apt to generate 'apparent' information that can in extreme cases mask the actual information that would flow between the different molecular components. We show how this artificial inflation in information arises, and how the effects of different types of noise alone and in combination can be understood. By applying the same approach to the ERK signalling pathway, we determine the best mechanistic model for this system in question and shed light on the noise filtering capabilities of the system. We also consider the pathway under conditions of abnormal regulation, similar to those associated with tumerogenesis. Next we investigate the Akt pathway, using a highly simplified model under different EGF input signal conditions and in the presence of difference sources of variability, in order to determine weather, under these conditions, an information theoretic approach can still be used. We find that such an approach cannot replace the use of mechanistic models to understand the molecular dynamics. Overall, these results are a stepping stone towards understanding the complexities of cellular information processing.

Idiopathic Pulmonary Fibrosis (IPF) involves excess extracellular matrix (ECM) deposition within the lung interstitium, caused by non-resolving chronic inflammation and dysregulated repair. Alveolar macrophage (AMφ) may contribute to IPF through releasing various mediators by different subsets, investigated here in vitro, and ex vivo using a mouse model of bleomycin (BLM)-induced pulmonary fibrosis. The role of foamy AMφ in the reported increased susceptibility of Hermansky Pudlak Syndrome (HPS) 1 mice to BLM-induced pulmonary fibrosis was also assessed. Novel characterisation studies revealed that terminally differentiated AMφ are inducible into M1-like [nitric oxide synthase 2 (NOS2)hi interleukin (IL)-1βhi IL-12 p40 (total)hi major histocompatibility protein (MHC)-IIhi mannose receptor C, type 1 (MRC1)-] or M2-like [Arginase 1 (Arg1)hi Fibronectinhi IGF-1hi MHC-IIlo MRC1-/+] phenotypes following IFN-γ or IL-13 priming respectively. Lipopolysccharide (LPS) altered these AMφ subset phenotypes. AMφ heterogeneity in a novel multiple oropharyngeal dose, BLM-induced pulmonary fibrosis was evaluated from days 7 to 21. Accumulation of M1-like AMφ at day 7, and M1/M2-hybrid AMφ [Arg1hi IL-12 p40 (total)hi fibronectinhi MHC-IIlo MRC1-/+] from days 7 to 21, may promote inflammation and fibrosis respectively. Toll-like Receptor (TLR) 9 messenger ribonucleic acid (mRNA) and TLR2 surface protein, and both TLRs2 and 9 ex vivo activities were increased in BLM-challenged mice from days 7 to 21, suggesting their roles in inflammation and fibrosis. Foamy AMφ accumulated in BLM-induced pulmonary fibrosis, and their potential role in the reported increased susceptibility to BLM-induced pulmonary fibrosis of HPS1 mice was evaluated. BLM-challenged HPS1 mice (days 7-21) had increased weight loss indicating reduced BLM tolerance from days 7 to 11, but little/no difference in collagen accumulation, suggesting that reduced BLM tolerance is not correlated with increased pulmonary fibrosis. In conclusion AMφ alter their phenotype in response to their environment that contributes to different stages of BLM-induced pulmonary fibrosis. Reduced BLM tolerance in HPS1 mice is not correlated with increased pulmonary fibrosis.

Pseudomonas aeruginosa is a versatile and prevalent opportunistic pathogen. It encodes a large arsenal of pathogenicity factors, and secrets a plethora of proteins using specialised protein secretion systems. The type VI secretion system (T6SS) delivers proteins directly into neighbouring bacteria or eukaryotic cells using a mechanism homologous to the T4 bacteriophage tail spike. Three T6SS are encoded on the P. aeruginosa genome. The study of the H1-T6SS has been facilitated by the fact it can be activated by the manipulation of the RetS/Gac/Rsm regulatory cascade by deletion of retS. However, the precise signals required for activation of this cascade, resulting in H1-T6SS activation, are unknown. This work investigates the role of subinhibitory concentrations of antibiotics in activating the system, and shows that kanamycin is able to induce production of core H1-T6SS components. This activation requires a functional Gac/Rsm cascade, but it is not known if this is due to direct signalling via the cascade, or due to a dominant effect of RsmA repression. The H2-T6SS is characterized in this work. We highlight key differences between the H2-T6SS cluster in PAO1 and PA14, including the presence of additional core T6SS components and putative secreted effectors. A strain is generated in which expression of the PA14 H2-T6SS cluster can be activated and tightly controlled by arabinose inducible promoters. The activity of the promoters is confirmed by the H2-T6SS dependent secretion of Hcp2 specifically upon arabinose induction. We further consider two putative H2-T6SS secreted substrates, VgrG14 and Rhs14. Production of these proteins is observed following arabinose induction, but their secretion is not detected. The Rhs14 protein is characterised, and its possible role as a H2-T6SS dependent effector is discussed. Finally, the H2-T6SS system in PA14 is shown to inhibit the internalisation of P. aeruginosa PA14, in contrast to the previously published observations of the H2-T6SS promoting internalisation of PAO1.

Conservation is difficult for moving targets, such as migratory species or landscapes subject to environmental change. Biodiversity offsetting is a novel approach that involves active compensation for biodiversity lost through development, with an objective of no net loss of biodiversity overall. In this thesis, I explore the use of biodiversity offsets for moving targets. My case study is the conservation of the migratory saiga antelope Saiga tatarica alongside industry in the Ustyurt plateau, Uzbekistan. Key challenges for offsetting include: specification of an appropriate frame of reference for evaluating no net loss, determination of requisite ecological gains, and the degree of flexibility permitted in biodiversity trades. I use bespoke simulation models to predict whether no net loss of biodiversity can be achieved within various hypothetical frames of reference, i.e. against different socio-ecological baselines and counterfactual scenarios. The reference frame determines the feasibility and effort required in achieving conservation objectives, and I shed light upon those ecosystem dynamics for which offsets may be appropriate. I develop a socio-ecological counterfactual for saigas and their Ustyurt habitat, relying upon satellite imagery and secondary data sets. Even with limited data, it proves possible to develop an instructive counterfactual for intervention. To calculate offset requirements, I first quantify impacts of industrial activity on the Ustyurt. Vegetation impacts are measured, mapped and projected to the landscape scale, and the influence on mobile species such as saigas is considered. Via quantitative comparison, I show that the application of different available offset calculation methodologies to these data - which all purport to achieve no net loss of biodiversity - would result in divergent offset requirements. This implies that offset methodologies should be tailored to specific moving target problems, rather than generalised. Finally, I use conservation planning software to compare the performance of flexible and non-flexible offsets. Zonation is used to model the effect of permitting flexibility in the biological, spatial and temporal constraints placed upon offsetting, and RobOff to assess the optimum return on investment under uncertainty. I find that a mixture of flexible and non-flexible offsets is desirable for conserving moving targets in the Ustyurt. We must give deeper consideration to the dynamic nature of ecosystems when designing conservation interventions. Biodiversity offsets have potential in this regard. To realise the potential, we should specify appropriate frames of reference, tailor metrics, and consider allowing flexible biodiversity trades.

A custom-built bioinformatics pipeline is used to costruct a species-level phylogeny for beetles (Coleoptera) using publicly available sequences from Genbank, including 8441 terminals 5 gene loci. 152 of 189 described beetle families are included, representing 2.17% of described species. The overall structure of publicly available data and its fit with Linnaean classifications are discussed. The dataset is further expanded by the inclusion of additional gene loci and relaxation of concatenation conditions, bringing total species to ~12,000. To overcome incomplete or incorrect identifications, a multi-partite matching algorithm is applied, for concatenation of partially conflicting taxon labels between gene loci, using species-level sequence clusters. The method is modified through the addition of country/specimen weighting between loci, and the incorporation of the the GMYC method of sequence-based species delimitation into the bioinformatics pipeline. GMYC and BlastClust approaches are compared, in terms of accuracy of species delimitation, supermatrix structure and topology of resulting trees. GMYC clusters are used as a framework for broad-scale comparisons of intraspecificvariation across the Coleoptera. The Coleoptera tree is used to illustrate a novel method for estimating total extant diversity by extrapolating from higher-taxon diversification rates, generating an estimate of 3.1 million beetle species globally. The sensitivity of the method to phylogenetic uncertainty within the data, and undersampling of families and subfamilies, is examined. Partial and complete mitochondrial genomes are used to generate the largest and most comprehensive phylogeny ever produced fromthis type of data. This tree is used as the basis for a molecular dating analysis, and the quantification of compositional heterogeneity among genes, taxa and sites within protein-coding genes. Non-homogenous substitution models are applied to help resolve problematic regions of the phylogeny, and the effects on topology and phylogenetic diversity of adding a previously unsampled regional fauna from Borneo are assessed.

Phytophthora ramorum is the cause of Sudden Oak Death in the USA and also infects many ornamental shrub species in both North America and Europe. In addition, it is now causing widespread disease in UK commercial larch plantations. The epidemiology of P. ramorum including sporulation potential, pathogenicity, ability to persist and the changing threat posed by the recently discovered EU2 lineage of the pathogen were investigated. Results indicated that all three species of commercially grown larch in the UK are not only able to support prolific sporulation of P. ramorum (approximately 960 sporangia per cm2) but are similarly vulnerable to bark colonisation, resulting in extensive dieback and mortality. Sporulation of P. ramorum on larch foliage exceeded that of all known sporulating hosts, including Californian bay laurel (approximately 77 sporangia per cm2), which drives epidemics in North America. Sporulation potential on foliage and larch bark susceptibility varied significantly over the growing season and with genotype. Sporulation on larch foliage was highest in summer but lower in spring and autumn. In contrast, susceptibility of larch bark was highest in spring and decreased in late summer. Asymptomatic infection of foliage occurred early in the year with symptom development mainly seen on old foliage and late in the year. Phytophthora ramorum also persisted in the litter layer but not the soil for up to two years after removal of infected Japanese larch in south-east England. In comparison with the widespread EU1 lineage, the EU2 had a faster grow rate over 2.5-29°C and was significantly more pathogenic to Japanese larch, European larch and English oak bark. However, isolates of the EU2 lineage produced significantly lower inoculum loads than the EU1 suggesting a trade-off between pathogenicity and sporulation. The implications of the aggressive nature of P. ramorum infection on commercially grown larch, its persistence on infested sites and the new established adaptively different EU2 lineage are discussed in relation to the future of disease management in the UK.

Tight regulation of gene expression is crucial for the survival of an organism, and allows certain genes to be switched on or off depending on the growth conditions and the state of differentiation. Transcription initiation is the most highly regulated step of gene expression, preventing wastage and being subject to the action of sophisticated signalling pathways. The RNA polymerases are further regulated through multiple different activators (stimulating transcription) and repressors (inhibiting transcription). Notably, promoter recognition specificity in bacteria is regulated by several dissociable sigma factors which can bind to the RNA polymerase core enzyme. Sigma54 (σ54) is the major alternative sigma factor in E. coli and historically known for its role in nitrogen metabolism. One unique property of σ54-dependent transcription is that it recognises -12 and -24 promoter elements rather than the traditional -10 and -35 sequences. Another distinguishing feature of σ54-dependent transcription initiation is that it absolutely requires the activity of a cognate activator and ATP hydrolysis, potentially giving tight control over gene expression and a wider dynamic range than with σ70-dependent transcription. Interestingly, this activator requirement has been shown to be bypassed in vitro with σ54 mutants where the interaction with the -12 promoter DNA element is disrupted. However, σ54-dependent transcription is not readily observed in vivo for the same mutants, suggesting further barriers exist in vivo inhibiting activator independence from occurring. The -12 promoter element has been shown to be somewhat expendable for σ54 promoter binding, and activator-independent σ54 transcription would obviate the need for additional proteins and ATP. This raises the question as to why σ54-dependent transcription has not evolved towards activator independence over time, by bypassing the interaction with the -12 promoter DNA. In this study, I show that σ54-dependent inhibition of transcription was detected for several genes with σ54 binding sites in their promoter region (ytfJ, chaC, patA, argT, mdfA ybhK, and acrD). Additionally, local transcriptional repression in the RNASeq data directly correlated with proximal σ54 binding sites, suggesting a novel role for σ54 as a transcriptional repressor. Strikingly, the -12 consensus motifs were more conserved in promoters linked to σ54-dependent transcriptional inhibition, underlining the importance of this promoter element in this newly discovered putative repressive function. However, extensive screens for genes inhibiting bypass transcription in vivo failed to identify any major repressive genes that keep activator independence in check in vivo. The mechanisms maintaining activator bypass transcription at low levels in vivo still remain to be characterised.

The basic leucine zipper transcription factor E4bp4 is essential for various immunological processes, most notably the development of natural killer (NK) cells. E4bp4 is required at the point of lineage commitment when NK cell progenitors develop from common lymphoid progenitors. E4bp4 promotes NK cell development by directly regulating the expression of other transcription factors, including Eomes and Id2. Despite its critical role, little is known about how the functions of the E4bp4 protein are regulated within the cellular environment. This study demonstrates that E4bp4 is post translationally modified by phosphorylation and SUMOylation and that these modifications directly regulate the protein's function. In the absence of either modification, E4bp4 is both a more potent transcriptional activator and transcriptional repressor. In NK cells, the absence of post translational modifications on E4bp4 promotes the expression of target genes Notch1 and E2A. Most strikingly, when expressed in hematopoietic progenitor cells, versions of E4bp4 lacking phosphorylation or SUMOylation sites, promote the development of up to three times more NK cells than the wild type form of E4bp4. This work has implications for the production of NK cells for use in immunotherapy and provides the first evidence that SUMOylation of an individual transcription factor protein can regulate a highly complex cellular process.

The opportunistic human pathogen Pseudomonas aeruginosa is a threat for immunocompromised individuals, a major cause of nosocomial infections, and is prevalent in patients with cystic fibrosis. The bacterium can form biofilms that help it evade the immune response. It adheres to host cells using molecular adhesins, such as pili assembled by chaperone-usher pathways (Cup). Understanding the adhesion could, therefore, help develop treatments that prevent the establishment of infections. This thesis considers the CupB system, consisting of an usher (CupB3), two chaperones (CupB2 and CupB4) and two pilin subunits (CupB1 and CupB6). The chaperones target the pilin subunits to the usher assembling a CupB1-containing pilus with a putative CupB6 adhesin at its tip. The cupB operon also encodes the TpsA-like protein (two partner secretion) CupB5, previously suggested to be secreted by CupB3. The aim of this work was to understand the CupB1-containing pilus assembly and CupB5 secretion mechanism. Genetic and biochemical techniques were used, such as deletion or point mutation, qRT-PCR, pull-down assays, shearing assays, and protein structure prediction or resolution. They led to the following results. First, each chaperone likely has a cognate substrate: CupB1 interacts with CupB2 and CupB6 with CupB4. Second, the crystal structure solved for CupB6 showed that it has a pilin and a putative adhesin domain, connected by a poly-proline linker. Third, CupB5 secretion was observed to be CupB3-independent and TpsB4-dependent. tpsB4 is encoded with its substrate tpsA4. The expression of the cupB and tpsB4/tpsA4 operons was shown to be controlled by the same regulatory pathway, Roc1, and deletion of the tpsB4 transporter gene abolished CupB5 secretion. Fourth, a structural analysis indicated that TpsB4 has two POTRA domains, and POTRA-1 interacts with the highly homologue TPS motifs of CupB5 and TpsA4. Based on these results, the thesis presents a model of CupB pilus assembly and CupB5 secretion.