No description available.

Abstract Protein folding of outer membrane and secreted proteins, including receptors, cytokines and antibodies is often linked to disulphide bond formation. Native disulphide bond formation is complex and is usually the rate limiting step in the folding of such proteins. The enzymes which catalyse the slow steps in disulphide bond formation belong to the protein disulphide isomerase (PDI) family. PDI catalyses formation, reduction and isomerization of newly synthesized disulphide bonds. The mechanisms of action of the PDIs are currently poorly understood and this not only inhibits our understanding of the biogenesis of a range of medically important proteins, and hence associated disease states, but also prevents the effective manipulation of the cellular environment by the biotechnology industry for the production of high value therapeutic proteins. Hence, understanding the mechanism of action of these enzymes is vital for a wide range of medically important processes and therapies. In this study the role of a conserved arginine residue in the catalytic activity of PDI was shown. The movement of this residue into and out of the active site locale of PDI was shown to modulate the pKa of the C-terminal active site cysteine of PDI and by that way to allow the enzyme to act efficiently as catalyst both of oxidation and isomerization reactions. The possible role of hydrogen peroxide produced by sulphydryl oxidases during disulphide bond formation was studied in an oxidative protein refolding assay. Analysis showed that hydrogen peroxide can be used productively to make native disulphide bonds in folding proteins with minimal side reactions. In addition, the kinetics of oxidation and reduction of the <b>a</b> domains of PDI and Pdi1p by glutathione was studied in this thesis. The kinetics obtained with stopped-flow and quenched-flow experiments showed the reactions to be more rapid and complex than previously thought. Significant differences exist between the kinetics of PDI and Pdi1p. This implies that the use of yeast systems to predict physiological roles for mammalian PDI family members should be treated cautiously.

Currently the mammalian endoplasmic reticulum (ER) is known to contain at least 20 different protein disulphide isomerase (PDI) family members. The oxidoreductases in the PDI family are thought to catalyse the formation and rearrangement of disulphide bonds in newly synthesised proteins. The focus of this work was to characterise two of the PDI family members: PDI and ERp46. In vitro translation reactions of major histocompatibility complex (MHC), β1-integrin (β1-I), haemagglutinin (HA), procollagen α1(III) and preprolaction (pPL) were carried out in untreated or PDI-depleted cells. The depletion of PDI decreased the rate of folding of MHC and β1-I and also prevented the oligomerisation of HA, suggesting a role for PDI in folding these putative substrates. However, when PDI was depleted neither the folding of pPL or HA was affected, implying that they may not be substrates for PDI. To determine the role of ERp46 in the cell, a substrate-trapping approach was used. Here substrates interacting with ERp46 were trapped as mixed disulphides isolated by immunoprecipitation, separated by 2D SDS-PAGE and identified by mass spectrometry. It was demonstrated that ERp46 forms mixed disulphides with at least 23 proteins, including heavily secreted proteins such as laminins, integrins and collagens. In particular, interactions with Ero1, Prx IV, EDEM3 and ERAP2 were found and confirmed by immunoprecipitation of radiolabelled in vitro translated protein. Notably nine of these clients of ERp46 have previously been identified as substrates of ERp57 (Jessop, Watkins et al. 2009). This would support the hypothesis that several different oxidoreductases, working in concert, are required to fold certain substrate proteins. Also, it was confirmed that Prx IV and Ero1 each form a mixed disulphide with PDI. These results highlight the importance of PDI family members in recruiting co-factors to substrates. Additionally, the over-expression of ERp46 led to increased cell survival following DTT treatment, yet after depletion of ERp46, cells were less able to grow, perhaps suggesting a role for ERp46 in maintaining ER redox homeostasis and cell survival. This suggestion was supported by the finding that ERp46 is able to catalyse the reduction of Prx IV in the presence of glutathione. These results suggest that Prx IV provides a novel mechanism for the transfer of disulphide bonds to nascent proteins in the ER via PDI family members such as ERp46 and PDI.

Cysteine rich with EGF-like domains 2 (Creld2) is a novel endoplasmic reticulum (ER) resident molecular chaperone that has been recently implicated in the ER stress signalling response (ERSS) and the unfolded protein response (UPR). Global transcriptomic data derived from in vivo mouse models of rare chondrodysplasias; Multiple Epiphyseal Dysplasia (MED Matn3 p.V194D) and Metaphyseal chondrodysplasia type Schmid (MCDS Col10a1 p.N617K), identified a significant upregulation in Creld2 expression in mutant chondrocytes. These chondrodysplasias share a common disease signature consisting of aberrant folding of a matrix component often as a result of inappropriate alignment of intramolecular disulphide bonds. This in turn culminates in toxic protein aggregation, intracellular retention mutant polypeptides and a classical ER stress response. The aim of this study was to further analyse the function of Creld2 in cartilage development and chondrodysplasias in which endochondral bone growth is perturbed. Protein disulphide isomerases (PDIAs) were amongst the most up-regulated genes in the MED and MCDS mouse models, consistent with the prolonged exposure of normally 'buried' cysteine residues. This led to the hypothesis that Creld2 was functioning as a novel PDI-like oxidoreductase to assist in the correct folding and maturation of aggregated misfolded polypeptide chains through REDOX regulated thiol disulphide exchange. A series of Creld2-CXXA substrate trapping mutants were generated in order to determine whether Creld2 possessed inherent isomerase activity. Here potential substrates interacting with Creld2 were 'trapped' as mixed disulphide intermediates, then isolated by immunoprecipitation and identified by mass spectrometry analysis. It was demonstrated that Creld2 possessed a catalytic active CXXC motif in its N-terminus that enabled the molecular chaperone to participate in REDOX regulated thiol disulphide exchange with at least 20 potential substrates including; laminin (alpha3,β3,γ2), thrombospondin 1, integrin alpha3 and type VI collagen. There was also numerous co-chaperones and foldases thought to be part of a specialised protein-protein interactome (PPI) for folding nascent polypeptides translocating the ER lumen. Moreover, co-immunoprecipitation experiments supported a protein-protein interaction between Creld2 and mutant matrilin-3, thereby inferring a potential chondro-protective role in resolving non-native disulphide bonded aggregates in MED. An established biochemical approach was employed to test the hypothesis that all MATN3-MED disease causing mutations have a generic cellular response to the β-sheet V194D mutation, consisting of intracellular retention, protein aggregation and ER stress induction. Several missense mutations were selected for analyses which encompassed a spectrum of disease severity and included examples of both β-sheet and alpha helical mutations. It was possible to define a reliable and reproducible assay for categorising MATN3 missense mutations into pathological or benign based on these basic parameters. This study was extended further to determine whether there were common pathological mechanisms behind MED and Bethlem myopathy (BM) caused by missense mutations in von Willebrand Factor A domain (vWF-A) containing proteins (matrilin-3 and type VI collagen respectively). We chose to compare and contrast the effects of an archetypal MATN3-MED causing mutation (R121W) with the equivalent COL6A2-BM causing mutation (R876H). These mutations compromised protein folding and maturation, resulting in the familiar disease profile of intracellular retention, protein aggregation and an ER stress response in an artificial overexpression system. However, the mutant C2 domain was efficiently targeted for degradation whilst mutant matrilin-3 vWF-A domain appeared to be resistant to these molecular processes.Molecular genetics was employed to study the role of Creld2 in vivo. Creld2-/- null mice (both global and conditional) were generated to directly examine the role of Creld2 in endochondral bone growth. Global knock-out mice were viable with no overt phenotype at birth. However, female Creld2-/- null mice showed a significant reduction in body weight and tibia bone length at 3 weeks of age. A cartilage specific knock-out was generated to determine whether these skeletal abnormalities were attributed to a systemic or a direct effect on cartilage development. [Creld2Flox/Flox Col2Cre (+)] demonstrated a severe chondrodysplasia with significantly reduced body weight and long bone growth compared to control littermates. Morphological and histochemical analysis of mutant growth plates revealed gross disorganisation of the chondrocyte columns with extensive regions of hypocellularity. These pathological features were confirmed to be the result of reduced chondrocyte proliferation and increased/spatially dysregulated apoptosis throughout all zones of differentiation. Taken together, these data provide evidence that Creld2 possesses isomerase activity and exhibits distinct substrate specificity. Furthermore, Creld2 has a fundamental role in post-natal cartilage development and chondrocyte differentiation in the growth plate.

The aim of the work described in this thesis is to investigate the nature and mechanisms of action of repellent cues for growing axons. In particular I try to resolve the controversy in the literature regarding the need for protein synthesis in the growth cone in response to external guidance cues. My results resolve the conflicting data in the literature on Semaphorin-3A signalling, where differing labs had shown that inhibiting protein synthesis either blocks or has no effect upon repulsion. They demonstrate the presence of at least two independent pathways, protein synthesis-dependent mTOR activation and -independent GSK3? activation. The higher sensitivity of the synthesis-dependent pathway, and its redundancy at higher concentrations where synthesis-independent mechanisms can evoke a full collapse response alone, resolve the apparent conflict. My experiments also demonstrated that Nogo-?20, a domain of Nogo-A, requires local protein synthesis to cause collapse. Unlike Semaphorin-3A, the dependence of collapse upon protein synthesis is concentration-independent and does not involve guanylyl cyclase, but it does share a dependence upon mTOR activity and the synthesis of RhoA, sufficient to cause collapse downstream of Semaphorin-3A. The other axon-repelling domain of Nogo-A, Nogo-66, is partially dependent upon the proteasome instead. It does not share a common pathway with Nogo-?20, except that both are RhoA-dependent. I further attempted to identify the nature of a repulsive activity found in grey matter, ruling out a previously suggested candidate identity. Finally, I examined the phenomenon of nitric oxide-induced growth cone collapse. My experiments revealed that S-nitrosylated glutathione causes growth cone collapse through the activity of protein disulphide isomerase. This mechanism shows only a partial dependence upon soluble guanylyl cyclase, but I argue that it has total dependence upon an S-nitrosylated donor. Coupled with its apparent relation to S-palmitoylation, the reciprocal of S-nitrosylation, I propose that nitric oxide causes collapse by crossing the cell membrane to inhibit S-palmitoylation-determined localisation of proteins. These results reveal some of the many pathways involved in growth cone collapse, whose further characterisation may provide new targets for the treatment of injuries of the central nervous system.