Most cells, including bone cells, release ATP into the extracellular environment. A considerable body of previous work has shown that ATP, acting through the P2 receptors, inhibits bone formation by osteoblasts and increases bone resorption by osteoclasts. This work focuses on the action of two key breakdown products of ATP, pyrophosphate and adenosine on bone cell function. Pyrophosphate, a ubiquitous physicochemical inhibitor of mineralisation, is formed from extracellular ATP by the action of ecto-nucleotide pyrophosphatase phosphodiesterases (NPPs); in bone these enzymes act in opposition to alkaline phosphatase. Adenosine, which can be generated in a number of ways from ATP, has been previously reported to stimulate both osteoblast and osteoclast function. However, using in vitro cultures, I found that it had little or no effect on the differentiation and bone forming capacity of rat osteoblasts, nor on the formation and resorptive function of mouse osteoclasts. I investigated the possibility that osteocytes, which form an interconnected cellular network within bone, might regulate mineralisation via NPPs. I found that cultured, primary osteocyte-like cells derived from mouse bone expressed Enpp1 mRNA. Osteocyte lacunae in the femora of Enpp1-/- mice imaged by scanning electron microscopy were found to be reduced in area by about 35%; indirect estimates of lacunar size using microCT imaging were in agreement. These results are consistent with the notion that ATP-derived pyrophosphate is important for maintenance of osteocyte lacunae size. Enpp1-/- mouse bones (humerus) were found to have reduced cortical bone diameter, reduced cortical porosity and an increased endosteal diameter compared to wild types, suggesting that the knockout phenotype also involves increased bone resorption and decreased bone formation. Histology and microCT of Enpp1-/- mice confirmed inappropriate joint mineralisation and showed that cartilage in the trachea and ear pinna was also mineralised, as were whisker sheaths. Osteoblasts, osteoclasts and osteocytes cultured in vitro from Enpp1-/- mice were found to release less ATP compared to cells from Enpp1+/+ mice in static conditions and after fluid flow stimulation. Enpp1-/- osteoblasts and osteoclasts also contained higher levels of intracellular ATP. Enpp1-/- osteoblasts showed increased bone production in vitro compared to Enpp1+/+; no effects of Enpp1 knockout on the formation or resorptive activity of osteoclasts were noted. Sclerostin, an osteocyte-derived inhibitor of WNT signalling and bone formation, was found to increase Enpp1 mRNA expression and NPP activity of osteoblasts, without affecting ALP in vitro. These results emphasise the importance of ATP and its breakdown product pyrophosphate in regulating mineralisation.

The majority of CATH domain structure superfamilies have small populations and are conserved in sequence and function. However, previous studies have shown that 4% are highly populated and functionally diverse. Previous analyses of some of these showed that relatives with different functions tend to exploit different functional sites to perform their function. In this work, functional site diversity was explored with a much larger dataset of superfamilies, by examining residues involved in protein interfaces and catalytic sites. This was done using a novel protocol to map sites across each superfamily. Functional site locations were shown to be least diverse for catalytic sites and most diverse for protein- protein binding sites. However, although protein interaction sites can vary considerably, in 79% of superfamilies analysed there is a common protein interface site, used by at least 80% of the functionally diverse relatives. By contrast with protein interactions, enzyme superfamilies tend to use the same active site in functionally diverse relatives. However, sometimes the nature and location of catalytic residues vary. We examined changes in catalytic machinery over one hundred enzyme superfamilies by considering physicochemical properties and sequence/structure positions. Reaction mechanisms were also compared to explore how enzyme chemistry has evolved between functionally diverse relatives and how changes in chemistry relate to changes in catalytic residues. A complex relationship was found and several examples are discussed to illustrate the different trends identified. In the final chapter, we assigned metagenome sequences to functional families in CATH and used KEGG pathway annotations to identify differences in the functional abilities of two metagenome environments, the human tongue and gut. Bacteroidetes, Firmicutes, and Proteobacteria phyla dominate both microbiomes. Enriched functional terms in the tongue and gut environments suggested an enrichment of bacterial cell wall building proteins in the mouth and an enrichment of denitrifying enzymes in the gut.

This paper presents the results of a large scale empirical study of coherent dependence clusters. All statements in a coherent dependence cluster depend upon the same set of statements and affect the same set of statements; a coherent cluster's statements have ‘coherent’ shared backward and forward dependence. We introduce an approximation to efficiently locate coherent clusters and show that it has a minimum precision of 97.76%. Our empirical study also finds that, despite their tight coherence constraints, coherent dependence clusters are in abundance: 23 of the 30 programs studied have coherent clusters that contain at least 10% of the whole program. Studying patterns of clustering in these programs reveals that most programs contain multiple substantial coherent clusters. A series of subsequent case studies uncover that all clusters of significant size map to a logical functionality and correspond to a program structure. For example, we show that for the program acct, the top five coherent clusters all map to specific, yet otherwise non-obvious, functionality. Cluster visualization also brings out subtle deficiencies in program structure and identifies potential refactoring candidates. A study of inter-cluster dependence is used to highlight how coherent clusters are connected to each other, revealing higher-level structures, which can be used in reverse engineering. Finally, studies are presented to illustrate how clusters are not correlated with program faults as they remain stable during most system evolution.

Collective cell migration is a fundamental process, occurring during embryogenesis and cancer metastasis. Neural crest cells exhibit such coordinated migration, where aberrant motion can lead to fatality or dysfunction of the embryo. Migration involves at least two complementary mechanisms: contact inhibition of locomotion (a repulsive interaction corresponding to a directional change of migration upon contact with a reciprocating cell), and co-attraction (a mutual chemoattraction mechanism). The thesis explores three different types of models to understand collective cell migration of the neural crest cells. Chapter 2 provides a review of the current literature. Chapter 3 presents the experimental data that are used to develop the models and Chapters 4, 5 and 6 explore three different modelling approaches. In Chapter 6, a parameterized discrete element model of neural crest cells is employed to investigate how the mechanisms of contact inhibition of locomotion and co-attraction contribute to long-range directional migration during development. Motion is characterized using a coherence parameter and the time taken to reach, collectively, a target location. The simulated cell group is shown to switch from a diffusive to a persistent state as the response-rate to co-attraction is increased. Furthermore, the model predicts that when co-attraction is inhibited, neural crest cells can migrate into restrictive regions. This suggests that the interplay between the complementary mechanisms may contribute to guidance of the neural crest. Directional migration is a system property and does not require action of external chemoattractants. The results of the thesis suggest that the cranial neural crest cells could acheive collective cell migration by self-organisation that is caused by aggregation.

The neuronal ceroid lipofuscinoses (NCLs) are a group of severe neurodegenerative lysosomal storage disorders characterised by accumulation of autofluorescent ceroid lipopigments in most cells. NCLs are caused by mutations in at least fourteen recessively inherited human genes. The NCL genes encode both soluble and transmembrane proteins localised to the endoplasmic reticulum, Golgi apparatus or endosomal/lysosomal organelles. Mutations in the CLN3 gene result in juvenile neuronal ceroid lipofuscinoses (JNCL, Batten disease). JNCL represents the worldwide most common form of NCL. Currently more than 40 mutations have been characterised in the CLN3 gene. However, the most common mutation causes a 1-kb deletion. CLN3 encode a multi-pass type III transmembrane protein, which is conserved in single-celled eukaryotes such as the fission yeast Schizosaccharomyces pombe, suggesting a fundamental role for this protein in eukaryotic cells. CLN3 has been functionally linked to many different cellular processes, including lysosomal homeostasis, autophagy, lipid synthesis or modification, cytoskeleton organisation and trafficking. Despite these endeavours, the function of CLN3 remains unclear. The main goal of this project was to investigate the role of the Golgi apparatus in the pathogenesis of juvenile CLN3 disease. The role of CLN3 at the Golgi apparatus was studied in mammalian cells and in fission yeast model. The morphology of the Golgi complex was studied in fibroblast cell lines from patients and in HeLa cells depleted for CLN3 using RNAi. The observed changes in morphology were accompanied by manganese dyshomeostasis within the Golgi complex, ER stress and apoptosis. The morphology of the Golgi complex was studied in S. pombe using electron microscopy in order to confirm the changes observed in mammalian cells. Finally, drugs shown to ameliorate aspects of the yeast model of CLN3 disease were tested for their efficacy in mammalian cells as an early step in therapeutic development. In this study I have shown that both morphology and size of the Golgi apparatus result to be affected by the loss/depletion of CLN3. Moreover, the changes in Golgi complex morphology and size are accompanied by manganese dyshomeostasis within the Golgi complex with activation of ER stress and activation of the proapoptotic protein caspase 2. Together, these data suggest that the loss/depletion of CLN3 activates secretory stress pathways and cell death. A dysfunctional Golgi apparatus may be the key to uncover the role of CLN3 and find new targets for therapeutic development.

Background: The heart undergoes major changes at birth: shifting preference from glycolysis to fatty acid oxidation, rapid development of the electrical conduction system and closure of the ductus arteriosus. The ultimate physiological trigger(s) causing the transitions are not known and I hypothesised that the postpartum increase in oxygen availability plays an important role. A changing oxygen environment also occurs ex utero, and I investigated the role of oxygen in the successful genetic adaptation of yaks and cattle to high altitude, and the genetic adaptations that correlate to athletic success at high altitude in humans. Results: I documented the changes in gene expression, cardiac electrical conduction, mitochondrial maturation and ductus arteriosus closure around birth in a mouse model. In all instances, maturation from fetal to adult phenotypes became evident within hours after birth. Changing oxygen conditions in the neonatal mice had no effect on ductus arteriosus closure but affected metabolic gene expression, mitochondrial morphology and maturation of the cardiac conduction system. I found that amongst elite long-distance runners the ACE ‘insertion’ allele and UCP3 T polymorphism were associated with a faster time in the Leadville high altitude ultra-marathon. My study in highland-dwelling cattle in Wyoming uncovered the first polymorphism (rs29016420 myosin heavy chain 15, T allele) associated with a reduction in hypoxia-induced pulmonary hypertension in cattle. I found positive selection of this polymorphism in yaks. Conclusions: Oxygen plays an important and variable role in postnatal maturation of the cardiovascular system. I discuss my findings as they may relate to the aetiology of Sudden Infant Death Syndrome and heart failure. I have also shown the important role of oxygen in genetic adaptation to a ‘fetal-like’ environment such as high altitude. Genetic studies in high altitude runners revealed significant associations with alleles involved in metabolism and the renin-angiotensin system and has prompted further study of this population for novel polymorphisms associated with high altitude adaptation. I also evaluated the first gene correlated with success at high altitude in cattle and the closely related yak.

Excitatory synapses preferentially form onto dendritic spines. Importantly, ab-normal spine morphology and density are associated with several neurological dis-eases. Thus, the appropriate formation of spines is critical for the proper function of neuronal circuits. However, the mechanisms that regulate spinogenesis remain poorly characterized. Recently, our lab demonstrated that Wnt7a stimulates the formation and matura-tion of excitatory synapses by regulating spine morphogenesis. The aim of my thesis was to dissect the mechanisms by which Wnts promote spinogenesis. I found that Wnt7a-mediated spine morphogenesis requires both translation and transcription. Moreover, my studies revealed that Wnt7a activates, through Ca2+/calmodulin-dependent (CaM) kinases, the transcription factor cAMP response element-binding protein (CREB) to induce spinogenesis. Interestingly, Wnt7a in-creases BDNF transcripts suggesting that Wnt7a is upstream of BDNF during neu-ronal circuit formation. Together, my studies demonstrate that Wnt7a induces spino-genesis in a transcription-dependent manner that requires CREB. Neuronal activity is known to modulate spine morphogenesis. However, little is known about molecules regulating activity-mediated spinogenesis. I observed that long-term potentiation (LTP) upregulates transcripts of several Wnt signalling com-ponents. This finding together with previous results indicates that activity-dependent changes in spine morphogenesis might be modulated by Wnt signalling. I also examined the role of microspherule protein 1 (MCRS1), an mRNA-binding protein, in Wnt-mediated spinogenesis to further elucidate the mechanism by which Wnt7a induces spine morphogenesis through translation. I showed that MCRS1 localizes to spines and interacts with Dishevelled-1, a key protein in Wnt signalling. My results suggest that MCRS1 might mediate some aspects of Wnt-mediated spine maturation. In conclusion, my studies demonstrate new roles for Wnt signalling in spinogen-esis. Neuronal activity, which results in LTP, increases mRNA levels of Wnt signal-ling components. Moreover, Wnt7a mediates spine morphogenesis through a CaM kinase-CREB pathway in a similar manner than activity. Collectively, this suggests that LTP-associated spinogenesis is regulated by Wnt signalling.

The gene underlying the classic neurodegenerative lysosomal storage disorder (LSD) juvenile neuronal ceroid lipofuscinosis (JNCL) in humans, CLN3, encodes a polytopic membrane spanning protein of unknown function. Several studies using simpler models have been performed in order to further understand this protein and its pathological mechanism. Schizosaccharomyces pombe provides an ideal model organism for the study of CLN3 function, due to its simplicity, genetic tractability and the presence of a single orthologue of CLN3 (Btn1p), which exhibits a functional profile comparable to its human counterpart. In this study, this model was used to explore the effect of different mutations in btn1 as well as phenotypes arising from complete deletion of the gene. Different btn1 mutations have different effects on the protein function, underlining different phenotypes and affecting the levels of expression of Btn1p. So far, there is no cure for JNCL and therefore it is of great importance to identify novel lead compounds that can be developed for disease therapy. To identify these compounds, a drug screen with btn1Δ cells based on their sensitivity to cyclosporine A, was developed. Positive hits from the screen were validated and tested for their ability to rescue other specific phenotypes also associated with the loss of btn1. The same hits were also tested in JNCL patient fibroblasts and in a zebrafish model of the disease. Promising results were obtained for three compounds: alloxazine, prochlorperazine dimaleate and E-64, with the latest being the one with the most potential for developing therapeutic tools. Yeast models for other LSDs (Chédiak Higashi Syndrome, Niemann-Pick disease type C2 and congenital CLN10 disease) were also characterised in terms of cellular phenotypes and the compounds described above were also tested in these models. Overlapping phenotypes were observed on all the yeast models, suggesting at least one common pathway between these LSDs.

Contamination of groundwater with the toxic soluble arsenic species, arsenite (AsIII) and arsenate (AsV) has led to an epidemic of arsenic poisoning effecting over 100 million people worldwide. The World Health Organisation (WHO) recommended maximum contaminant level (MCL) for arsenic in water is 0.13 μM (10 μg L−1). Accurate quantification of arsenic below the MCL usually requires highly sensitive laboratory based techniques, the practical uses of which are limited within the effected populations, principally due to cost. Biosensors are a potentially powerful technology for overcoming this problem. Amperometric biosensors couple the analytical sensitivity of electrochemistry with the selectivity of enzyme substrate interactions. The bioenergetic metalloenzyme AsIII oxidase (Aio) catalyses the oxidation of AsIII to AsV in a number of physiologically diverse microorganisms including the Rhizobium sp. str. NT-26. To develop a biosensor for AsIII it was first necessary to optimise the expression and purification of the biological recognition element, a recombinant NT-26 Aio in Escherichia coli str. DH5α. with final a yield of ca. 1.1 mg per L of culture. The recombinant NT-26 Aio was characterised using biophysical techniques to confirm the correct insertion of the enzyme cofactors during heterologous expression in E. coli. The reduction midpoint potentials of the 3Fe-4S (270 mV) and the Rieske 2Fe-2S (225 mV) clusters were confirmed by redox titration. The thermostability of the recombinant Aio was ≤ 64.5 °C. The oxidised structure of the Mo at the active site was confirmed to have a di-oxo (Mo = O2) coordination. The kinetics and pH dependence of AsIII oxidation were investigated using various artificial and physiological electron acceptors. Electrochemical studies were performed to develop a system for AsIII concentration determination, using the biological recognition element Aio. The electron transfer mediator ferrocene methanol was found to produce the greatest currents during catalytic voltammetry experiments at pH 8.0. A chronoamperometric detection system incorporating the electron transfer mediators ferrocene methanol and potassium ferricyanide was able to resolve AsIII concentrations of 0.07 – 0.53 μM (5 – 40 μg L−1), below the WHO MCL for arsenic, suggesting such a system would be capable of determining the safe levels of arsenic in drinking water.

AMPA receptors (AMPARs) are crucial for fast excitatory synaptic transmission throughout the mammalian central nervous system (CNS). At the molecular level, these receptors are tetramers of GluA1-4 subunits, which can be homomeric or heteromeric, incorporating the Ca2+-impermeable GluA2 subunit. Although the basic tetrameric AMPAR is sufficient for expression of functional channels in a recombinant system, native AMPARs are associated with a number of auxiliary proteins, which modify their trafficking as well as their biophysical and pharmacological properties. The first AMPAR auxiliary protein discovered was stargazin (γ-2), a transmembrane protein related to the γ-1 subunit of voltage gated Ca2+- channels. The lack of γ-2 in the naturally occurring mouse mutant stargazer was found to produce a loss of AMPAR-mediated transmission at the cerebellar mossy fibre to granule cell synapse. This led to the discovery, that γ-2 is crucial for AMPAR trafficking. Subsequent studies focusing on this protein showed that it is also able to modify the functional properties of AMPARs, such as deactivation and desensitization kinetics, single-channel conductance and susceptibility of Ca2+-permeable AMPARs to block by intracellular and extracellular polyamines. In addition, a family of related proteins, known as Transmembrane AMPA Receptor Regulatory Proteins (TARPs) has been identified. As well as γ-2, this includes the subunits γ-3,-4,-5,-7 and -8. All TARPs are related to γ-2 in structure and function, although certain properties vary between the family members. Since the discovery of TARPs, more proteins modulating AMPAR function have been identified, so that a complex network of AMPAR interacting partners is apparent from the literature in the field. Surprisingly, little focus had been put on how AMPARs interact with auxiliary proteins at the molecular level. Enhancing our knowledge about these interactions could provide insights into the role of AMPARs in synaptic transmission, plasticity and neurological diseases. The aim of my thesis was thus to provide insight into the role of TARPs in modulation of AMPAR complexes and the molecular mechanisms that underlie this modulation. I used a recombinant expression system to identify the physiological relevance of different AMPAR/TARP interaction regions, as well as to investigate the role of TARP stoichiometry on AMPAR/TARP function. The initial part of my thesis focused on the molecular interactions of AMPARs with their TARP auxiliary subunits. Our experiments showed that various regions of the first extracellular loop (Ex1) of γ-2 play a role in modulation of a number of functional properties of both GluA1 and GluA2(Q) subunits. We also aimed to obtain insight into the corresponding AMPAR regions important for interaction with the TARP. We therefore focused on the possible novel role of the AMPAR N-terminal domain (NTD) in TARP-AMPAR interaction and found that, although this domain plays a role in AMPAR function, it has little effect on the modulation of AMPARs by TARPs. The first two Chapters of my thesis were aimed at answering the question - how do TARPs and AMPARs interact at the molecular level? The work presented in the final Chapter attempted to expand on this question and focused on the role of TARP stoichiometry in AMPAR function. Using tandem TARP-AMPAR constructs expressed in a recombinant system, we have studied the properties of TARP/AMPAR assemblies that contain a known number of γ-2 molecules. This allowed us to identify some of the AMPAR properties that are sensitive to TARP stoichiometry. As most neurons in the brain express more than one TARP, we also investigated the pharmacological consequences of the simultaneous presence of two TARPs (γ-2 and γ-7) within a single TARP/AMPAR assembly, a possibility that has previously received little attention.