No description available.

During development individual neurons have to reach specific locations in the embryo to form a connected nervous system. Molecular complexes form between receptors expressed on the surface of the developing axon and molecular guidance cues in the extracellular environment, to guide the neuron to the correct location. Synapses then connect the axons with their target location. This thesis describes the interactions between the IgLON family of cell adhesion molecules and how these interactions affect their function during neuronal development. There are four members of the IgLON family in chick, namely LAMP, OBCAM, CEPU-I and Neurotractin. They are highly glycosylated proteins, predominantly anchored by a GPI anchor to the extracellular surface of the cell membrane. Initially, the comparative strengths of homophilic and heterophilic trans interactions of CEPU-I, OBCAM and LAMP were established. Generally trans heterophilic interactions within the family have a higher affinity than homophilic interactions, LAMP having the highest heterophilic affinity for CEPU and OBCAM. The lowest affinity of all the trans interactions tested was the LAMP homophilic interaction. Data is presented to suggest IgLONs also interact in cis to form heterodimeric complexes or Diglons. These putative Diglons affect the trans binding affinity for IgLON-FC recombinant proteins, possibly due to a conformational change altering the availability of the IgLON binding site. Biochemical and imaging studies provided additional physical evidence for the Diglon complex. Inhibition of the initiation of neurite outgrowth from chick forebrain neurons was found to be dependent on the presence of two IgLONs in the extracellular environment to suggest formation of putative Diglon heterodimeric complexes facilitates IgLONs to function as negative axon guidance cues. IgLONs have no cytoplasmic domain so are reliant on being part of a molecular complex on the surface of the cell membrane for signal transduction. Assays measuring neurite outgrowth from chick forebrain neurons suggested signal transduction for Diglons is linked to trimeric Go/i membrane proteins and/or to Rho GTPases.

The phosphoinositide lipids Ptdlns(3)P and Ptdlns(3,5)P2 play important roles on the endocytic pathway. Ptdlns(3)P localises to early endosomes and multivesicular bodies (MVBs) and is proposed to recruit proteins involved in fusion of early endosomes and internalisation of ubiquitinated receptors. Ptdlns(3,5)P2 is proposed to be involved in terminal maturation of Iysosomes and endosome to Golgi transport, but the precise role of this lipid in mammalian cells remains unclear. Studies in yeast have identified various proteins associated with Ptdlns(3,5)P2 metabolism, for which mammalian homologues have been found. These are the Ptdlns(3)P5- kinase Fab1/PIKfyve, Fig4 a 5-phosphatase that dephosphorylates Ptdlns(3,5)P2, Vac14 which has been shown to act as an upstream activator of PIKfyve, and WIPI-2/Svp1 a putative downstream effector of Ptdlns(3,5)P2. In this study I have further characterised three of these proteins and examined the cellular roles of all four with respect to a variety of trafficking pathways in which Ptdlns(3,5)P2 has been implicated. siRNA was used to examine the knockdown phenotypes of each of these proteins. Furthermore, I directly compared for the first time the effects of knockdown of PIKfyve, with acute pharmacological inhibition of its enzymatic activity. Loss of PIKfyve activity caused a failure in the retrieval of a variety of different cargos to the trans-Golgi network (TGN), including mannose-6-phosphate receptors, responsible for efficient delivery of lysosomal enzymes, and the TGN resident protein TGN46, leading to their accumulation in dispersed punctae. A failure in tyrosine kinase receptor downregulation was also observed following combined knockdown of PIKfyve with either Vac14 or Fig4 or following pharmacological inhibition of PIKfyve. PIKfyve knockdown alone had no effect, suggesting a low threshold of Ptdlns(3,5)P2 is necessary and sufficient for this pathway. Svp1 is the best characterised Ptdlns(3,5)P2 effector in yeast and is also an autophagy-related gene (Atg18), thereby implicating Ptdlns(3,5)P2 in this process. A family of putative mammalian Svp1 homologues have been identified, known as the WIPI family. I investigated the role of Ptdlns(3,5)P2 and WIPI-2 in mammalian autophagy. By monitoring formation of the autophagosome marker GFP-LC3 II, PIKfyve and WIPI-2 were found to have opposite effects. Furthermore, WIPI-2 redistributes to punctate structures upon induction of autophagy, which partially colocalise with autophagosome markers, in a manner dependent not on PIKfyve but on PI(3)-kinase activity. The evidence presented suggests that Ptdlns(3,5)P2 may playa role in mediating the maturation of a subset of MVBs, leading to swelling of endosomal compartments and rendering the MVBllate endosome or autophagosomes refractory to fusion with the lysosome in cells depleted of PIKfyve.

Bacterial cell integrity is maintained by peptidoglycan, a rigid polymer of alternating sugar residues (N-acetlyglucosamine and N-acetylmuramic acid) cross-linked by short peptide stems. Peptidoglycan biosynthesis is a complex three-stage process, under tight spatial and temporal control, with multiple steps exploited as antibiotic targets. The final stage occurs in the periplasm and involves the polymerisation of lipidlI (the final monomeric peptidoglycan precursor) by penicillin-binding proteins, and the attachment of cell surface proteins, containing a C-terminal LPXTG motif, by the sortase family of enzymes. The biochemical characterisation of penicillin-binding proteins and sortases has been mainly limited to investigating interactions with inhibitors and peptide analogues, due to the unavailability of the natural peptidoglycan precursors. Work presented in this thesis describes the development of protocols for the synthesis of both cytoplasmic and lipid-linked peptidoglycan intermediates. It was possible to produce approximately 8S mg of the final cytoplasmic precursor (UDP-MurNAcpentpeptide) in a single-pot incubation. UDP-MurNAc-pentapeptide was then converted to lipidII using a preparation of M jlavlls membranes. Due to the presence of cross~ linking in the peptidoglycan of many Gram-positives, a chemo-enzymatic procedure was used to attach branching amino acids. L-Ala and L-Ala-L-Ala amino acid branches, were attached to the e-amino group of L-Lys in position three of the UDP-MurNAcpentapeptide. These branched derivatives were also converted to lipidII. These substrates were then used to study the enzymology of a variety of pencillin-binding proteins. The transpeptidase activity of two high-molecular weight penicillin-binding proteins (S. aureus PBP2' and E.faecalis PBPS), with intrinsic'low affinity to ~-lactams, was investigated. However, no transpeptidase activity was detected with any monomer substrate. The requirement for the prepolymerisation of the glycan backbone prior to transpeptidation was investigated using two types of polymeric substrate (secreted uncross-linked peptidoglycan polymers and lipidII treated with a monofunctional transglycosylase to polymerise the glycan backbone of lipidII), however, no transpeptidase activity was detected using either of these polymeric substrates. The DDcarboxypeptidase and endopeptidase activity of a low-molecular weight penicillinbinding protein (E. coli PBP4) was investigated with a variety of peptidoglycan fragments. The importance of a small pocket at the base of the E. coli PBP4 transpeptidase active site for substrate recognition was demonstrated by site-direct mutagenesis. The differences in activity of these related classes of enzyme highlights the gap in the understanding of substrate recognition by the transpeptidase domain of pencillin-binding proteins. Sortase enzymes covalently attach proteins to peptidoglycan. Proteins are covalently linked to branched lipidlI, which is subsequently polymerised in to the peptidoglycan polymer. S. pnellmoniae SortaseA was crytsallised, and a fluorescence resonance energy transfer assay was used to detect transpeptidation. This work provides the basis for full structural and biochemical characterisation of Sortase mediated transpeptidation.

The insulin-secreting p-cells of the pancreas play a crucial role in blood glucose homeostasis. An elevation in blood glucose triggers the release of insulin into the systemic circulation, restoring euglycaemia. Diabetes mellitus arises when there is no longer an adequate supply of insulin available to regulate the concentration of blood glucose effectively. Beta-cells reside within cellular aggregates called the islets of Langerhans, the three-dimensional anatomy of which is crucial to their secretory function. It is well established that increased cell proximity augments glucose-evoked insulin secretion and several studies have highlighted the importance of homologous p-cell contacts within the islet configuration. It has been suggested that the improved secretory function is the result of communication via gap junctions and/or paracrine mediators between adjacent p-cells. However, there may be other molecules that are capable of modulating p-cell function in response to cell contact and recent work has provided evidence for a role for the Ca2+-dependent cell adhesion protein, E-cadherin. Using MIN6 pseudoislets as a model system, this study demonstrates that E-cadherin is a key regulator of p-cell function and is more than just an adhesion molecule, holding the three-dimensional islet together. It is not only important for the homotypic coupling of p-cells but also affects their secretory activity. Insulin secretion experiments demonstrate that the treatment of pre-formed pseudoislets with an immuno-neutralising E-cadherin antibody markedly reduces glucose-evoked insulin release. Furthermore, Fura-2 microfluorimtery shows that the synchronous glucoseinduced [Ca21i oscillations characteristic of pseudoislets are lost in the absence of Ecadherin- based cell contacts. Attenuation of fluorescent dye transfer between cells in the absence of cadherin ligation indicates a lack of functional gap junctions under such conditions providing evidence for an interaction between gap junctions and Ecadherin in p-cells. Overall these findings imply that E-cadherin has an important role in the regulation of intercellular communication between p-cells within the islet, an effect which appears to be gap junction dependent. The results of this study have potential implications for the in vitro generation of islets for future transplantation therapy.

Metallothioneins are a superfamily of small, cysteine-rich, metal-binding proteins, which have been the subject of intense scientific interest since their discovery in the late 1950s. Over the last decade, the advancement of genome sequencing, microarray and high-throughput protein identification techniques has resulted in an exponential increase in the number of plant MT sequences in protein and translated nucleotide databases. These studies show that plant metallothioneins display significant variation in sequence and are discretely expressed during various developmental stages and in different organs. Therefore, it is likely that plant metallothioneins carry out compartmentalised functions for which they possess specific properties. However, there is a relative paucity of academic literature on the structure and biochemical properties of plant MTs and this is a void that needs to be filled before we can establish structure/property and structure/function relationships for these proteins. This research has focussed on a comprehensive characterisation of the solution structure and metal binding dynamics of the wheat Eel/II metallothionein using a range of analytical and biophysical techniques. Mass spectrometry and nuclear magnetic resonance studies have identified six zinc ions in two distinct domains with hitherto unprecedented stoichiometries for MTs: Zn2Cys6 and Zn4CysllHis2. Structure calculations revealed that in addition to these unprecedented stoichiometries, the individual Ee domains possess unique structural features not previously reported in MT literature. In Domain 1, the binding of two zinc ions in a binuclear cluster was observed. When compared to Domain 2, the binuclear cluster displayed remarkable stability towards Wand EDTA attack. In Domain 2 an isolated binding site with stoichiometry ZnCys2His2 was observed, which is proposed to confer a well-defined structure for this domain. However, disruption of this site through mutation or cadmium binding abolishes ordered structure along with the defined binding of six zinc ions. These novel structural features confer distinct backbone and metal dynamic properties in each domain, and are most likely to have a functional significance. A separate avenue of investigation identified a potential role for Ee in zinc donation. A re-examination of previously published literature on the expression levels of Ee mRNA and protein demonstrated that it is conspicuously accumulated during desiccation, which suggests that it is perfectly poised to satisfy the microquota of zinc-dependent sites during rehydration processes. In conclusion, this research has made significant advances in the understanding of the structure/function relationship of the wheat Eel/II metallothionein.

Ferulic acid is a secondary metabolite usually found as esters in plants and dihydrocaffeic acid results from the microbial metabolism of flavonoids and of caffeic acid, both widely distributed in food. Even though ferulic acid and flavonoids have been proposed to exert several beneficial effects on health, their in vivo activities could partly result from their microbial metabolites and strongly depend on their bioavailability. The absorption and metabolism of phenolic acids were studied in vitro with a model for the colonic epithelium composed of absorptive and mucus secreting cells, ex vivo with everted colonic sacs and liver slices, and in vivo with rats. The photoprotective effect of phenolic acids was tested in vitro on the keratinocytes HaCaT.The ferulic acid permeation was mainly by transcellular diffusion and also by a facilitated transport (S-MCT and MCTI). Intestinal cells conjugated ferulic acid with sulphate or glucuronide and reduced its unsaturated side chain. In rats, intestinal cells were more potent for glucuronidation of dihydrocaffeic acid, whereas the liver favoured sulphation, the methylation being regio-selective.

In Paracoccus pantotrophus, periplasmic nitrate reduction is performed by NapAB. This dissimilatory system is implicated in maintaining membrane redox homeostasis by linking ubiquinol oxidation with the reduction of nitmte, during aerobic metabolism of highly reduced carbon sources. NapA contains a bis-molybdopterin guanine dinucleotide cofactor and an iron-sulphur cluster. The associated di-haem subunit, NapB mediates catalytic electron flow from the quinol-oxidase electron donor, Nape. Protein film voltammetry (PFV) revealed steady-state electrocatalysis from NapAB in the presence of nitrate. At pH 7, nitrate reduction is onset <0.15 V and increases across the potential mnge where ubiquinollevels predominate in vivo. Though NapAB cannot opemte upstream of cytochrome be], its opemting potential is entirely consistent with the proposed physiological role of this enzyme in productive redox balancing. The magnitude of the electrocatalytic current-potential profile was greatest at acidic pH (~7) and responded readily to micromolar concentmtions of nitrate. At pH 6 and 7, KM values of ca. 23 and 45 J.lM were defined at the activity peak potential and 40 V, respectively. These values are in good agreement with the KM value of <15 J.lM defined for NapAB in intact cells of P. pantotrophus M-6 and thus bring the KM value for this enzyme into line with those of related proteins. Electrocatalytic activity was further modulated below ca. -0.15 V, where attenuation was observed. Such 'tunnel-diode' behaviour was maintained at [S]»KM and the high potential wave flank showed greatest pH dependence. At pH 6, potentiometric titration confirmed that activity attenuation occurred in the vicinity of the [4Fe-4Sf+/I + redox couple. A competitive inhibitor, thiocyanate had little impact on wave shape. However, NapAB purified from bacteria grown in tungsten-supplemented media showed a similar waveform to the molybdoenzyme, but where catalytic onset was displaced to lower potentials, while the position of the low potential wave flank was unchanged. With activity attenuation attributed to a redox couple distinct from the active site, modelling indicates the M06 + 15 + and M05 +/ 4+ potentials likely lie in close proximity in the high potential wave flank, thus a semi-cooperative redox process may describe the redox chemistry of the molybdenum site, consistent with the presence of sub-stoichiometric M05+ populations in EPR experiments.

In Gram-negative bacteria such as Escherichia coli, signal sequence-bearing secretory preproteins are targeted post-translationally from the cytosol to their final destinations. This mechanism is mainly performed by the ubiquitous Sec machinery, a multiprotein complex containing the molecular motor ATPase SecA, the secretion-dedicated chaperone SecB and a heterotrimeric protein-conducting channel consisting of the SecY, SecE and SecG subunits. Crystal structures have been obtained for rnonomeric, detergent-solubilised SecYEG in its 'closed' and SecA-bound states, revealing that the channel lies at the centre of a single protomer. However, many aspects of preprotein translocation remain uncertain, including the functional significance of the observation that in membranes SecYEG is predominantly dimeric. To address these uncertainties, total internal reflection fluorescence microscopy (TIRFM) was exploited to investigate the functional oligomeric state and monitor dynamics of the translocon at a single molecule level. This approach revealed that while monomers are sufficient for the SecA- and ATPdependent association of SecYEG with preproteins, active transport requires SecYEG dimers associated through the SecE subunit. In collaboration with the Collinson group (University of Bristol, UK), a molecular model of the functional translocon was proposed, rationalising the need for both SecYEG copies. The SecY channel is closed at the periplasmic side of the membrane by a small helical region termed the 'plug'. Relocation of the latter towards SecE during polypeptide translocation was investigated kinetically and spatially by ensemble and single molecule Forster resonance energy transfer (FRET), respectively. Intra-molecular conformational changes within SecA were also probed using FRET, the results suggesting that monomerisation of the SecA dimer occurs as a pre-activation step upon binding to SecYEG. Overall, the results presented in this Thesis describe the first use of single molecule imaging to study the bacterial Sec-translocon, and represent an integral part of the emerging applications of single molecule techniques in the membrane protein field.

Legionella pneumophila is the causative agent of Legionnaires’ disease, severe pneumonia acquired from inhalation of contaminated water droplets. In the lung, L. pneumophila infects alveolar macrophages and creates a compartment named the Legionella containing vacuole (LCV), which avoids degradation and recruits components of the secretory pathway. LCV creation depends on the action of the Dot/Icm system that translocates over 275 effectors into the host cell. To study the function of these effectors, models that approximate human disease are required. Here, I characterise the larvae of Galleria mellonella as an infection model for L. pneumophila. Infection resulted in larval mortality and bacterial replication in a strain- and Dot/Icm-dependent manner. Flagella expression was dispensable for bacterial virulence, however secreted phospholipases and the Dot/Icm effector SdhA were shown to be important in virulence. Deletion of SdhA resulted in disruption of the LCV membrane and destruction of haemocytes. The importance of SdhA expression was confirmed in a mammalian model, validating the utility of G. mellonella. In the second part of this study, the novel protein LtpD was characterised. LtpD was translocated via the Dot/Icm secretion system and localised to the LCV. A series of truncation mutants defined a C-terminal 153 amino acid domain as required for LCV localisation. This region was shown to bind directly to the lipid phosphoinositide 3-phosphate. Further analysis revealed that LtpD also interacted with the enzyme inositol monophosphatase 1, however did not change the enzyme’s activity in vitro. Deletion of LtpD resulted in a subtle growth defect in mammalian macrophages at late time points during infection. This growth defect was also seen the G. mellonella and mouse lungs, confirming that LtpD is a virulence factor of L. pneumophila. In summary, here I present an infection model to investigate L. pneumophila virulence and further characterisation of the Dot/Icm effector LtpD.