Search

Clostridium difficile is a Gram-positive spore-forming obligate anaerobe and the most common cause of antibiotic-associated infectious diarrhoea worldwide. C. difficile infection (CDI) has increased in severity and incidence over the last decade, representing a huge economic burden to healthcare systems. Two toxins elaborated by C. difficile are widely regarded as the major causes of CDI symptoms. However, little research has focused on the assembly of the cell envelope, which is the first point of contact with the host. Vegetative C. difficile are covered in a proteinaceous paracrystalline array, known as the S-layer. Production of the S-layer comes at an enormous metabolic cost to the cell and functional analysis of the major surface layer protein (SlpA) has uncovered roles in pathogenicity and sporulation. Using a C. difficile S-layer null strain, we examined the conditional genome of C. difficile in the absence of this major surface structure. This was achieved by creating a large library of mutants via insertional mutagenesis coupled to Transposon Directed Insertion Site Sequencing (TraDIS). We have identified 55 genes as conditionally essential and 20 genes as conditionally non-essential in the absence of SlpA. For example, we have demonstrated that secA2, encoding the SlpA translocase is dispensable for growth in an S-layer mutant. Interestingly, we have also demonstrated that in the absence of SlpA, several PSII biosynthetic genes, involved in S-layer surface attachment, are conditionally non-essential for growth in vitro. As conditionally non-essential genes represent potential targets for disrupting S-layer assembly, we investigated the phenotypic effects of gene silencing via CRISPR interference of conditionally non-essential genes identified in this study. We provide evidence that knock-down of secA2 in a WT background produces the same cell morphology as an S-layer mutant. Additionally, we show depletion of Era, a poly(A) polymerase and several PSII biosynthesis genes from cells has effects on cell shape, surface boundaries, membrane permeability and localisation of septas.

Extracellular carbohydrates are an essential aspect of biology, playing central roles in cell-cell interaction, cell shape, and infectious disease. In eukaryotes, extracellular glycans and glyco-conjugates are synthesised or matured by glycosyltransferase enzymes in the Golgi apparatus. Most glycosyltransferases catalyse the formation of unique glycosidic bonds; thus, the activities of these enzymes are responsible for controlling the diversity of glycan structures. The study of glycosyltransferases can therefore grant insight into the biology of the Golgi apparatus and the ætiology of disease, as well as providing a means to engineer carbohydrate structures for therapeutic or material applications. GT43 members IRX9 and IRX14 are involved in the synthesis of xylan in Arabidopsis, and are thought to form a multimeric complex with GT47 member IRX10. In this work, I expressed transmembrane fragments of these proteins in E. coli and used a reporter assay to demonstrate the sequence-dependent oligomerisation of the IRX9 transmembrane domain. Furthermore, I showed that an IRX9 mutant containing a transmembrane G28I point mutation was unable to complement the phenotype of the irx9 mutant and appeared to be mislocalised when transiently expressed in tobacco leaves. In animals, exostosin glycosyltransferases synthesise the backbone of heparan sulphate, and contain both a GT47 and a GT64 domain. Using cryo-EM, I solved the structure of EXTL3, the largest exostosin, in complex with UDP. The structure revealed that the EXTL3 GT47 domain adopts a GT-B fold that has been inactivated by structural changes to the active site. The structure also demonstrated that the GT47 and GT64 active sites are likely separated by a surprising distance in these bi-domain enzymes. Members of GT47 clade A have particularly diverse activities in plants. I used the structure of EXTL3 to rationalise differences in substrate specificity between these enzymes. Although I was unable to explain the difference in activity between a recently characterised xylan arabinopyranosyltransferase and a homologous xylan galactosyltransferase, I was nevertheless successful in identifying a novel xyloglucan-specific enzyme from Coffea canephora on the basis of its protein sequence. As a whole, the results in this thesis provide insight into the relationship between amino acid sequence, atomic structure, and function in glycosyltransferases. The findings hint at a potential nano-scale organisation of the Golgi apparatus that will be exciting to investigate further.

The aim of the work is to investigate how asymmetric cell division and cell lineage isolation are linked to pollen cell fate determination. This was analysed by monitoring the activation of vegetative and male germline-specific markers in mutants, which disturb microspore polarity and division asymmetry (gemini pollen 1), or which disturb cytokinesis and cell lineage isolation (two-in-one). In twin-celled equally divided pollen present in late bicellular and mature anthers, the activation of VCK promoter was expressed in both daughter cells, while the activation of DUO1 and HTR10 promoters were not expressed at any stages. The centromeric chromatin fate marker HTR12-GFP is present at discrete nuclear foci in generative cell and sperm cell nuclei in wild type pollen, while in gem1 equal daughter cells the marker was initially detected in foci, but later was dispersed. Further, transposable elements activation represented by GUS reporter activity was present in wild type, twin-celled and binucleate pollen. These results support the hypothesis that asymmetric division of the microspore is required for the specification of generative cell fate and differential centromere identity between vegetative and generative cell nuclei. The analysis of cell fate in tio mutants showed that isolation of the germ cell lineage is not required for the initial activation of generative cell fate, but is essential for the maintenance of differential centromere identity. In binucleate tio pollen, generative cell fate markers were expressed despite the failure to partition daughter nuclei into separate cells. Three different cell fate categories were observed among binucleate tio pollen, including pollen with vegetative cell fate only, generative cell fate only and pollen with mixed cell fate. Moreover, the centromeric fate marker was initially present at discrete nuclear foci in both nuclei of binucleate tio pollen and later was dispersed. In a study of MICROTUBULE ORGANIZATION1 (MOR1) fused to YFP, the protein was localised as a perinuclear signal in microspores and enriched in the male germline of wild type pollen, while MOR1-YFP was not expressed in mutant gem1 and tio pollen.

The establishment of cell architecture, whether in a migrating cell, a polarized epithelial cell, or an asymmetrically dividing stem cell, requires proper intracellular patterning. One mechanism by which cells are able to locally concentrate molecules in space is through directed transport, typically through the action of cytoskeletal-motor networks. Using polarization of the C. elegans embryo as a model system, I sought to understand how actomyosin cortical flows drive efficient segregation of polarity proteins to one side of the cell. Prior data established that anterior PAR proteins are segregated into the anterior by cortical actomyosin flows, yet the mechanisms underlying this transport are unclear. More recent work suggested that oligomerization of PAR-3 and its ability to recruit other aPAR proteins is specifically required for efficient segregation. This data raised additional questions: Do all membrane-associated molecules sense flows? Do particular features of molecules such as clustering enable their segregation by advection? How is this regulated to enable correct spatiotemporal control of protein targeting? We combined single molecule tracking methods with perturbation of PAR-3 cluster dynamics to directly assess the ability of polarity molecules to be advected by flows and how this may be affected by cluster regulation. My results suggest that a variety of polarity molecules are advected by cortical flows, allowing flow to shape molecular distributions. At the same time, not all molecules are advected, or at least not advected efficiently, indicating that specific molecular features of protein complexes facilitate advection. Surprisingly, despite being required for efficient segregation, clustering of PAR-3 is not required to sense flows. Moreover, although clustering alters diffusivity, the observed changes are minimal and would not be expected to substantially alter their ability to be segregated. Rather, clustering is most likely required to shape the pattern of membrane association, potentially through positive feedback, though this remains to be definitively explored.

Phosphatidylinositol 5-phosphate 4-kinase (PIP4K) catalyses the formation of phosphatidylinositol 4,5 bisphosphate (PtdIns(4,5)P2) which is a messenger involved in important cellular signalling pathways including the PI3K/Akt/mTOR pathway and the phospholipase C pathway. Three isoforms of PIP4K have been identified, namely PIP4K2α, PIP4K2β, and PIP4K2, which have been shown to play roles in cellular activities such as growth, stress response, immune system regulation, and tumour formation. However, the exact mechanisms engaged by the kinases are still obscure. This study focused on investigating the role of PIP4K in modulating the immune system through regulatory T cells (Tregs), and in growth of cancer cells with inactivated p53. Findings from this study showed that PIP4K2β and 2 isoforms regulate PI3K/Akt/mTOR signalling in Tregs to control cell proliferation, survival, and immunosuppressive activity. This study also suggested that PIP4K controls FOXP3 expression, the master transcriptional regulator of Tregs through a novel PI3K/UHRF1/FOXP3 pathway. Previously PIP4K had been shown to play roles in cancer cell growth, and a discovery of possible synthetic lethal interaction with p53 in cancer cells has made this kinase a promising target for cancer therapy. Findings from this study indicate that the synthetic lethality of PIP4K and p53 is not present in all types of cancer. Rather, it is suggested to depend on the specific mutation of p53 and alteration in expression of PIP4K in the cells. Inhibition of PIP4K also regulates the PI3K/Akt pathway differently in different cancer cells. Together, findings from this study suggested that PIP4K modulates both the immune system and cancer cell growth, and its different role in regulating PI3K/Akt/mTOR pathway could be the key to understand how PIP4K modulates the systemic immune response, and how different cancers can benefit from the expression of this family of lipid kinases.

Macroautophagy ('autophagy' hereafter) is a lysosome-dependent degradation system for intracellular protein quality control. Autophagy can selectively remove cargo materials, which is dependent on cargo receptors. SQSTM1/p62 ('p62' hereafter) is one of the best characterised receptors mediating the autophagic degradation of misfolded proteins, and its body formation is essential for its cargo recognition and clearance. Mounting evidence has revealed that membraneless compartments, participating in various biological processes, are assembled via liquid-liquid phase separation (LLPS). These phase-separated condensates ensure that cellular activities occur in a spatiotemporally controlled manner and provide a method for concentrating and segregating cellular components. Liquid-like non-membrane condensates present dynamic properties and can exchange macromolecules with surroundings. The formation of non-membrane-bound p62 bodies is mediated by liquid-liquid phase separation. It has been yet poorly characterised the regulation of p62 liquid-liquid phase separation, and the molecular details of the assembly and functions of p62 bodies remain elusive. This study aims to identify a mechanism of it. Via an unbiased yeast two-hybrid screening, DAXX is found a p62 binding protein. Further experiments confirmed that cytoplasmic DAXX promotes p62 puncta formation and drives p62 liquid phase condensation. In vitro experiments also confirmed that DAXX promotes p62 phase condensation. Evidence has suggested the critical role of the p62-Keap1-Nrf2 pathway in combating the oxidative stress and maintaining the cellular redox homeostasis. To examine whether DAXX has an effect on redox homeostasis, Reactive Oxygen Species (ROS) assays and Dual-luciferase assays were applied, and results showed that the DAXX-induced p62 phase condensation promotes p62 recruitment of Keap1 and subsequent Nrf2-mediated stress responses. Collectively, this study suggests that cytoplasmic DAXX drives p62 phase condensation and regulates cellular redox homeostasis, providing a mechanistic insight into p62 phase separation and the prosurvival function of DAXX. Further investigation would reveal if DAXX-induced p62 phase condensation is implicated in ROS-relevant human diseases.

Two-pore channels (TPCs) are endolysosomal ion channels of physiological and pathophysiological significance. However, fundamental properties concerning ion permeability and activation mechanisms are ambiguous. Also, as the likely targets for Ca2+-mobilizing messenger NAADP, the role of TPCs in cell-wide Ca2+ signalling is ill-defined. Importantly, their pharmacology is limited to cell-impermeable activators and a few non-selective inhibitors, which brings challenges for characterizing TPCs. In this thesis, I address the above issues. I began by examining the mechanism of action of the lysosomotropic agent, glycyl-L- phenylalanine 2-naphthylamide (GPN), which has long been appreciated for mediating Ca2+ signals from lysosomes and for probing TPC function. Its action on lysosomes has recently been questioned. However, using fibroblasts, here I show that GPN mobilises Ca2+ from acidic organelles. I move on to characterise two cell-permeable and selective TPC2 activators (A1 and H07). Additionally, I confirm that approved drugs targeting estrogen and dopamine receptors are selective TPC2 inhibitors. I go on to show that A1 and H07 activate TPC2 differentially. A1 induced larger and quicker Ca2+ signals than H07 but similar Na+ signals. A1 and H07 targeted distinct sites on TPC2. Besides, H07 but not A1-induced Ca2+ signals were regulated by external (luminal) pH. The implication is that TPC2 may be regulated in an agonist- specific manner. Finally, by using GPN and inhibiting TPC activity with novel inhibitors or siRNA knockdown, I show that TPCs are required for histamine- but not bradykinin-induced Ca2+ signals. More specifically, histamine-mediated Ca2+ signals were reduced upon TPC2 but not TPC1 knockdown. Thus, TPCs are implicated in global Ca2+ signalling evoked by physiological stimuli likely in an isoform-dependent manner. Collectively, this research has provided novel TPC modulators with which to further characterize fundamental properties and physiological roles of TPCs.

The transport of macromolecules between the nucleus and cytoplasm is controlled by nuclear pore complexes (NPC) that permeate the nuclear envelope. The NPC allows small (∼ 5 nm diameter) molecules to passively diffuse in and out of the nucleus but hinders the transport of larger material through the presence of sticky intrinsically disordered proteins (FG Nups) that occlude the inner pore channel. Through complexation with nuclear transport receptors (NTRs), which have an affinity to the FG motifs in the Nups, large cargoes are able to overcome the NPC permeability barrier. Our physical understanding of how FG Nups and NTRs give rise to selective transport remains incomplete. In this thesis, the behaviour of FG Nups and NTRs is probed using coarse-grained physical modelling approaches that treat the FG Nups and NTRs in a rather minimal fashion, and that are implemented in computer simulations and analytical models. By treating the FG Nups as homopolymers and comparing simulations with various experimental data it is here found that they behave very similarly to idealized polymer chains. Secondly, when this model is extended to recently developed biomimetic nanopores it can account for the morphology as seen in those experiments. In this nanopore system, the resealing dynamics of the FG Nups were three orders of magnitude faster than typical transport event times (~ 1 ms). The homopolymer model is further expanded to include various NTRs in an FG Nup polymer film, and simulations reveal the emergence of phase separation of different NTRs at physiologically relevant densities. Finally, the roles of sequence and surface heterogeneity of the FG Nups and NTRs on binding, diffusion in an FG Nup melt, and uptake in a nanopore are explored using modelling approaches that account for such heterogeneity at minimal complexity. Overall, this thesis makes progress towards a minimal model of the NPC.

Collective cell migration is essential in many fundamental biological processes. During morphogenesis, cells display highly coordinated shape changes, rearrangements, and motion. During wound healing, cells must coordinate their migration to close gaps in epithelial tissues to prevent infection. The efficiency of these processes is determined by the fluidity of the tissue --- the ability for cells to rearrange and remodel, which in turn is governed by the mechanics of both the cells and their environment. In this thesis, I use computational methods to investigate the origin of tissue fluidity, and its role on collective cell motion. In Part I, I investigate how tissue fluidity governs collective cell motion during wound healing. We find that the ability for cells to rearrange, preventing jamming, is essential for closure. Despite contractile tension around the gap driving closure, reducing tension in the cells increases tissue fluidity and healing rates. In Part II, I study how cells optimise their active behaviour, either contraction or crawling, to regulate tissue fluidity for rapid motion. In wound healing, we find that a balance of the two modes is most efficient over a wide range of cell, substrate, and tissue properties. In Part III, I study how tissue fluidity regulates force transmission in cell colonies. For stiff cells, the tissue acts like a giant single-cell, with traction forces distributed around the colony periphery, but as tissue fluidity increases, forces localise to the interior of the colony. Tissue fluidity requires that cells are able to rearrange and change shape, which in turn is regulated by adaptive viscoelastic properties of the junctions. Thus, in Part IV, I develop a new theory for how cell-cell junctions remodel under stress; strain above a threshold triggers tension remodelling and irreversible length changes. This enables tissue shape change and homeostasis that is robust to fluctuations in stress. Overall, this work demonstrates the important role of single cell mechanics and tissue fluidity on regulating collective cell behaviours.

In 1950 the Leeds-based biologist Irene Manton, FRS, had an international reputation as an expert in the cytology and evolution of ferns. But with the arrival of the electron microscope her research took a new and fascinating turn, which soon led to her making major discoveries on the fine structure of flagella. Drawing on previously unexamined archival materials as well as interviews with former colleagues, this thesis uses Manton's case in order to explore the impact of the electron microscope on post-WWII biological research. Particular attention is paid to (i) the transition from research based in light microscopy and its techniques; (ii) the challenges that Manton faced in gaining access to the latest electron microscopes (and how being a woman exacerbated those challenges); and (iii) her disciplinary position as someone whose integration of evolutionary and cytological research represents a path not taken for the "cell biology" that emerged in the 1950s and 1960s.

The nuclear pore complex (NPC) forms a nanochannel for selective transport into and out of the cells nucleus. Notably, the transport selectivity of the NPC critically depends on an assembly of intrinsically disordered proteins with multiple phenylalanine-glycine repeats (referred to as FG nups), grafted to the inner walls of the NPC. The focus of this work is to probe the biophysical nature of this indispensable nanochannel by using mimetic systems that emulate key properties of the NPC. Using a DNA origami scaffold, it has been possible to study the configurations adopted by purifi ed FG-nups confi ned to a nanopore geometry that mimics the NPC central channel. In this thesis, atomic force microscopy (AFM) has been used extensively to compare different flavours of FG-nups and to monitor their stochastic behaviour as they form transient condensates that occlude the pore's central lumen. Rearrangements of these entanglements were observed on the timescale of seconds, suggesting that reordering occurs at little energetic cost. This supports the idea that the FG nups are sufficiently 'sticky' or 'cohesive' to form dense condensates that seal the NPC's transport barrier yet are sufficiently dynamic at the molecular scale to facilitate the transport of larger molecules. Furthermore, the dynamics of two different FG-nup (the more cohesive Nup100 and the less cohesive Nsp1) was quantifed using autocorrelation analysis and compared to analogous data acquired on native NPCs. As predicted Nup100 condensates were found to be longer lived than those formed by its less cohesive counterpart Nsp1. By contrast, no fluctuations above the background noise were observed in native NPCs. A possible explanation being due to the presence of soluble transport receptor proteins and molecules caught in transport that trap the FG-nups in given morphologies. To further probe this hypothesis additional components of the transport system have been added to the mimetic nuclear pore complexes and their binding has been visualised with single molecule resolution, both by AFM and by total internal reflection fluorescence microscopy (TIRFM). Importin was seen to bind to both Nup100 and Nsp1 whereby a systematic increase in the number of nucleoporins led to greater binding.

Intermediate filaments, along with microtubules, microfilaments and septins, comprise the cytoskeleton that provides cells with structure and shape. The cytoskeleton is known to participate in the organelle localization and signal transduction. Autophagy is an intracellular degradation mechanism that allows cells to recycle damaged proteins and organelles. When cells are exposed to stress, autophagy is activated to promote cell survival. Autophagy is highly regulated, and it is reported that the cytoskeleton participates in autophagosome formation and transport. Vimentin filaments are reported to act as autophagy inhibitors, and also as a scaffold for autophagy activation during infection. However, little is known about other roles in autophagy. Autophagy is reported to be involved in autoimmune disease like Inflammatory Bowel Disease (IBD).Crohn's disease (CD) and Ulcerative colitis (UC), are the major forms of IBD. IBDs are multifactorial, and the contributing factors include environmental triggers, microbial dysbiosis and genetic predisposition. For CD, the genes with the strongest association include the Pattern Recognition Receptors (PRR) Nucleotide Binding Oligomerization Domain (NOD)2 and Autophagy Related Protein (ATG)16L1, that participate in autophagy induction during infection. It was hypothesized that vimentin filaments participate in autophagy steps. Two objectives were investigated, i) the role of vimentin in NOD2 induced autophagy during infection, ii) the role of vimentin in autophagy. The methodology used included live/fixed cell imaging, Western blot and flow cytometry. The live cell imaging approached used for the first objective, did not provide further understanding of the underlying mechanism. For the second objective, it was shown that pharmacological vimentin inhibition through Withaferin A (WFA), results in increased autophagy as monitored through Western blot, flow cytometry and, immunostaining. Further investigation of the mechanism showed that vimentin inhibition results in re-distribution of autophagosomes and lysosomes and this results in preventing the formation of autolysosomes, as shown by the traffic light experiment, that is used to monitor the maturation of autophagosomes into autolysosomes. The main autophagy marker, Microtubule-associated protein 1-light chain 3 (LC3), is conjugated to Red Fluorescent Protein (RFP) and Green Fluorescent Protein (GFP) both expressed on the surface of autophagosome, that appear yellow. Once the autophagosome fuses with the lysosomes LC3 and GFP is degraded and autolysosomes appear red. These findings indicate that vimentin acts as a scaffold for the fusion between autophagosomes and lysosomes.

The rhomboid-like superfamily consists of the active rhomboid proteases and the inactive rhomboid pseudoproteases. The latter group includes the iRhoms, a family of polytopic transmembrane proteins composed of a long N-terminal cytoplasmic domain followed by seven transmembrane helices (TMH), with a highly conserved loop between TMH1 and TMH2 called the iRhom Homology Domain (IRHD). There are two mammalian iRhoms; iRhom1, which is broadly expressed, and iRhom2, which is predominantly expressed in immune cells and skin. A key function of iRhoms is regulating A disintegrin and metalloproteinase 17 (ADAM17), a metalloprotease responsible for the cleavage of tethered ligands on the plasma membrane, including the primary inflammatory cytokine tumour necrosis factor α. iRhoms are responsible for the trafficking of the catalytically inactive and immature form of ADAM17 from the endoplasmic reticulum (ER) to the Golgi apparatus where the inhibitory pro-domain of ADAM17 is cleaved and removed. The iRhom/ADAM17 complex traffics to the cell surface where ADAM17 is now mature. Mature ADAM17 is activated by the phosphorylation of iRhom and therefore able to cleave its substrates. iRhoms have an indirect but important role in ADAM17-dependent cleavage of substrates and are increasingly viewed as regulatory subunits of ADAM17. The complex is stabilised at the cell surface as a result of the prevention of its degradation by FRMD8. Before this thesis, little to no structural studies have been performed on the iRhom/ADAM17 complex. Understanding the complex at a structural level will enable further insight to the mechanisms by which these two proteins interact. Therefore, the aim of this project was to express and purify iRhoms alone and in complex with ADAM17 for structure determination. The project initially started by characterising the IRHD, of human iRhom1 (IRHD1), with the aim of crystallisation because it is the only major ER luminal domain of iRhom that may interact with the extracellular domain of ADAM17. However, after insect cell expression and bacterial refolding of IRHD1 it became apparent that the IRHD may be more stable as part of the full length protein. Human cell lines were generated to stably overexpress mammalian and non-mammalian iRhoms with their corresponding ADAM17 and a chimeric form of human iRhom2 fused to FRMD8 (HsFRMD8iRhom2). Expression and sub-cellular localisation of the generated cell lines was verified and optimal solubilising conditions were determined for maximal protein extraction out of the membrane and for minimal complex dissociation. HsFRMD8iRhom2 chimera was successfully purified and preliminary cryo electron microscopy data was generated. Therefore, the results reported in this thesis provide a foundation towards solving the structure of iRhom and iRhom/ADAM17 complex.

Ion channels constitute crucial components of biological membranes and represent a major class of pharmacological targets. In response to specific stimuli, conformational changes in these proteins may lead to expansion or constriction of the channel pore in its gate region(s) to permit or preclude transmembrane ion transport. With an increasing number of molecular structures being resolved for families of ion channels in alternative conformations, this thesis concerns the assessment of relative hydrophobicity of the pore lining to complement measurements of pore dimensions for predicting whether ion conduction is likely to be possible in each captured conformational state. Molecular dynamics simulations were used to examine the effect of hydrophobic gating, whereby the passage of water and ions through an apolar region may be energetically disfavoured without steric occlusion of the pore. A hydrophobic gate was identified in the bestrophin-1 channel, with simulations upon in silico mutagenesis of its distinct pore-lining motif revealing a clear dependence of pore hydration on the polarity of the substituting amino acid. The combined influence of local hydrophobicity and pore radius was quantified based on simulations and machine learning-assisted analysis of 190 ion channel structures. A simple scoring model thus derived is able to rapidly detect the presence of hydrophobic barriers to permeation among channel structures. Evaluated against levels of ionic flux supported by channel structures, simulated via computational electrophysiology, the model performed reasonably reliably as a first indicator of ion channel conductive state. On the basis of the investigation, a three-tiered pipeline has been suggested to facilitate efficient annotation and classification of ion channel structures. This offers a starting point for more detailed functional and mechanistic studies. Furthermore, the approach is anticipated to be capable of aiding the design and engineering of peptide-based nanopores that contain switchable hydrophobic gates.

Accurate membrane targeting is essential for the organisation of membrane traffic. This step is particularly crucial at the Golgi apparatus, a central sorting hub of membrane traffic in the cell. Golgins, a class of highly conserved coiled-coil vesicle tethers, have been shown to mediate specificity of membrane traffic to the Golgi via the mitochondrial relocation assay in mammalian tissue culture cells (Wong and Munro, 2014). However, little is known about the role of golgins in polarised cells and the exact subclasses of vesicles they tether in different cell types. Hence, a tissue-specific mitochondrial relocation assay was devised in Drosophila to investigate the function of seven different golgins (Golgin-97, Golgin-245, GCC88, GCC185, TMF, GMAP, and Golgin-84) in different polarised cell types and identify cargoes, Rabs, and adaptor proteins that these golgins relocate. By doing so, I have shown that the broad specificities of the golgins in tissue culture cells perdure in the physiological context: trans-Golgi localized Golgin-97 and Golgin-245 capture endosome-to-Golgi vesicles, while the intra-Golgi golgins TMF, Golgin-84, and GMAP tether Golgi-derived retrograde transport carriers. Furthermore, ultrastructural data evidence the function of these golgins as bona fide vesicle tethers in Drosophila tissues. More importantly, adding a tissue-specific dimension to the relocation assay elicited novel insights into golgin function that were not yet discovered in cultured cells. Golgins acquired or lost specific vesicle capture activities depending on the cell type, and GCC185, an inert golgin in COS and HeLa cells, unambiguously captured vesicles in Drosophila larval salivary glands (Wong and Munro, 2014). Furthermore, Golgin-97 and Golgin-245 exhibited functional differences in Drosophila tissues despite reported functional redundancy in mammalian tissue culture cells (Shin et al., 2017). This finding led to the generation and study of Drosophila knock-out and knock-in lines of Golgin-97/245 molecular receptor, TBC1D23. Finally, the combined cargo, adaptor protein, and Rab capture data evidenced partial functional redundancies between golgins, perhaps suggestive of collective vesicle tethering by the golgins. This hypothesis was partly confirmed by evidence of functional compensation between TMF and GMAP in experiments relocating ectopic golgins in golgin mutant backgrounds. Taken together, this work reveals physiologically relevant insights into the function of golgin vesicle tethers and illuminates the utility of Drosophila for understanding the complex pathways of traffic in polarised cells.

Mitochondria are double membrane-bound organelles found in the cytosol of eukaryotic cells. They generate energy for the cell by oxidising the breakdown products of metabolism and relaying the liberated electrons through a series of membrane-embedded oxidoreductase proteins known as the electron transport chain (ETC). The first enzyme in the ETC is complex I (NADH:ubiquinone oxidoreductase). Complex I oxidises NADH in the matrix, reduces ubiquinone in the inner membrane, and couples the energy released to the translocation of four protons across the inner membrane, generating the proton motive force that powers vital cellular processes. The major question of how the energy liberated by NADH:ubiquinone oxidoreduction is captured and efficiently exploited to drive proton translocation is still unanswered. Addressing this question is important for understanding clinically relevant complex I dysfunctions and for pharmacological manipulation of the enzyme. The projects presented here aimed to investigate the mechanism of complex I by using single-particle electron cryomicroscopy (cryo-EM) to generate three-dimensional reconstructions of its structure.

Proper neuronal function relies on efficient transport within axons and dendrites, the long processes which extend from neuronal cell bodies. Within neuronal processes, the microtubule cytoskeleton provides tracks for motors to carry cargo over long distances. In axons, cytoplasmic dynein-1 (dynein) is the microtubule motor which carries cargo back towards the cell body. Dynein cargo include endolysosomal vesicles, autophagosomes and mitochondria. Dynein is activated by binding to a co-factor called dynactin. How dynactin binding activates dynein is unknown. How multiple dynein/dynactin complexes are arranged on cargo for long range transport in cells is also not clear. To understand this, I used cryo-electron microscopy (EM) combined with biochemical analysis and fluorescence microscopy to study dynein in vitro and in cells. In Chapter 2, I present the cryo-EM structure of purified full-length human cytoplasmic-1 in an inhibited state. By making structure-guided point mutations, I show that dynactin binding is the key step to promote long-range movement of dynein on microtubules and propose a model for how dynactin binding activates dynein. I then focus on visualising dynein/dynactin complexes in neurons using cryo-electron tomography. My initial aim was to perform cryo-correlative light and electron microscopy (cryo-CLEM) of dynein cargo in axons to determine if it is possible to visualise dynein/dynactin in cells and show their distribution. In Chapter 3, I show different strategies for labelling dynein cargo in primary neurons grown on EM grids for cryo-fluorescence microscopy. I describe the challenge to obtain samples with both sufficient fluorescent signal and suitable thickness for cryo-CLEM imaging. This work led me to alter my strategy and image the thin parts of mouse dorsal root ganglion (DRG) axons without guidance from cryo-FM signal. In Chapter 4, I present a survey of the intracellular architecture of thin parts of DRG axons using cryo-electron tomography. I describe the organisation of microtubules and other cytoskeletal filaments in axons. I show the morphology of microtubule plus and minus ends are similar and that microtubule inner proteins are distributed within the lumen of microtubules in DRG axons. I also describe the morphology of the membrane-bound compartments and virus-like particles present in axons. I found that short, regular tethers connect regions of the ER to microtubules while longer, more heterogeneous connections link other membrane-bound organelles to microtubules. Some of the long connections show similarity to dynein/dynactin complexes.

Cells and tissues are highly dependent on the secretion of extracellular matrix proteins like collagen to fulfil their roles in multicellular organisms and enable complex processes such as organ and bone formation during early development. Due to its unique structure, extensive post-translational modifications, requirement for fast abundant protein secretion during wound-healing processes and size of about 300 nm in length, type I procollagen (the precursor molecule of type I collagen) imposes high demand on the secretory pathway. Defects in collagen transport and secretion can lead to osteogenesis imperfecta, Ehlers-Danlos Syndrome, and fibrosis. Transport of proteins from the endoplasmic reticulum to the Golgi is facilitated via the COPII vesicle coat complex. These are typically considered to be spherical 60 – 80 nm diameter structures that bud from the endoplasmic reticulum membrane, encapsulate cargo proteins, and subsequently merge with the Golgi apparatus or, the endoplasmic reticulum-Golgi intermediate compartment. Despite having identified many factors involved in COPII organisation and regulation a major question in the field remains – how are large secretory cargo proteins like procollagen transported in a COPII-dependent manner through the early secretory pathway? Hypotheses include the use of large COPII megacarriers and direct tunnel-like connections. Here, I have used a new engineered GFP-tagged procollagen reporter, combined with experiments in primary fibroblasts, to characterise procollagen transport. In addition, fibroblasts from patients were investigated to help classify the pathogenicity of novel/rare mutations in genes involved in collagen processing and transport. GFP-procollagen was observed to be transported to the Golgi in absence of large carriers. Super resolution microscopy revealed small procollagen puncta overlapping, but not surrounded by, COPII. These data are consistent with models where large cargo proteins, including fibrillar collagens, are transported in a progressive manner in close proximity to Golgi membranes. This does not require large carriers, nor discrete cargo encapsulation by COPII.

Microtubules are highly dynamic structures with a diverse range of functions including cell division and intracellular transport. This has made them excellent targets for pharmaceuticals and agrochemicals. To understand their dynamic behaviour, microtubules need to be imaged live. However, current imaging probes, such as GFP-tubulin, suffer from high background and both EMTB-3xGFP and SiR-Tubulin are known to stabilise microtubules. Additionally, many of the current microtubule imaging methods only work in one or a few different experimental systems. Therefore, our aim was to produce a new imaging probe that overcame these deficiencies. We present three new live imaging microtubule probes, MnIV, Dynein 2219, and Kif5c that produce microtubule images on a par with or better than existing microtubule imaging probes. All three are protein-based and contain one or more microtubule-binding domains attached to an interchangeable fluorescent tag. Through the use of automated particle tracking and cell scoring we demonstrate that all of the probes analysed, including commonly used probes, modify normal microtubule function to some degree. These effects should be considered when choosing which microtubule-imaging probe to use in any imaging experiment. Increases in microtubule stabilisation were observed, with EMTB-3xGFP showing the largest increase. Additionally, tracking of microtubule plus-tips showed that all the probes we tested, including currently used probes, slowed microtubule dynamics by a small amount. Organelle movement was reduced, to a varying degree, but the overall distribution of lysosomes within the cell was not significantly different from transfected controls. This suggests that the effects on movement are not large enough to cause gross changes to organelle distribution. Of the probes we tested, Kif5c provided the best balance between clear microtubule images and its effects on normal microtubule behaviour. Further testing also showed that Kif5c is a useful tool for imaging microtubules in both Xenopus and Drosophila and is non-toxic to Drosophila development, providing the possibility of one probe that can be used to provide clear microtubule images in a wide range of organisms. This would be a useful tool in the development of new microtubule targeting compounds, as it could be used to directly observe a compounds effect on microtubules in a range of organisms, providing insight into its mode of action and specificity.

Cyclophilin D (CypD) is an approximately 18KDa protein, and it is the only reported cyclophilin family of the peptidyl prolyl cis trans isomerase that is present in the mitochondrial matrix. It functions mainly as a regulator of the open and closed state of the mitochondrial permeability transition pore (MPTP); however, the mechanism by which this occur is still largely undetermined. Being a peptidyl prolyl isomerase, CypD is believed to also function alongside other mitochondrial quality control proteins to execute the important task of folding and refolding of nascent and unfolded matrix proteins respectively. Indeed, many studies have investigated the MPTP, using CypD as ideal modulator; unfortunately, most of these studies were carried out under pathological and/or in vitro artificial experimental conditions. It was hypothesized that exploring the basal physiological role of CypD in the mitochondria, in addition to signalling changes that accompany its modulation using both knock down and knock out cell-based systems would enhance the understanding of CypD and this may prove useful in the delineation and characterisation of the MPTP. The work herein quantified the amount of CypD present in the mitochondria of both mouse tissues and HAP1 cells. It was observed that CypD was more than 5-fold higher in cancer cells, depending on the cell type than normal, nontransformed cells/tissues. Furthermore, the Cys203 on CypD directly regulates the amount of mitochondrial NAD(P)H by forming stable sulfenic acid which may constantly require the mitochondrial NADPH-dependent thioredoxin or glutaredoxin systems for its reduction. Moreover, CypD elevated basal mitochondrial ROS, membrane potential (Δψm) /proton motive force and mitochondrial ATP while decreasing the level of forward flux TCA cycle enzymes such as inhibiting PDH activity while decreasing the level of OGDH and IDH3 when compared to CypD deficient cells. Furthermore, extracellular signalling pathways such as AKT, mTORC1, ERK, AMPK and HIF-1a were differentially altered between CypD WT and deficient cells. More importantly, it was observed that CypD deficient cells were susceptible to all forms of cellular stress most likely via inhibition of autophagy. This result emphasized the physiological importance of CypD in the regulation of basal cellular ROS signalling and ATP synthesis/energy homeostasis, and this effect may be used by cancer cells to enhance the amount of circulating ROS. The findings herein suggest that inhibition of the MPTP during CypD deficiency may be associated with decreased inner membrane proton gradient and subsequent acidification of the matrix. However, CypD ablation may not be an overall useful technique against cell death, as it confers stress-like phenotype on the cells.

Patients suffering from the peripheral neuropathy Charcot-Marie-Tooth type 4B (CMT4B) present mutations in three proteins which belong to the same enzymatic complex acting to dephosphorylate the phosphoinositide lipids phosphatidylinositol 3-phosphate (PI3P) and phosphatidylinositol 3,5-bisphosphate (PI(3,5)P2). Either of the catalytically inactive myotubularin-related proteins (MTMR) 5 and 13 forms a complex with the catalytically active MTMR2 to enhance its phosphatase activity toward PI3P and PI(3,5)P2; additionally, MTMR5 and 13 act as activators of small GTPases (involved in protein recruitment to intracellular membranes) through their N-terminal DENN domain. Previous studies isolated the small GTPase Rab35 as a strong interactor with Sbf, the Drosophila melanogaster homologue for MTMR13, but did not pursue this interaction. As it indicates a potential role for the mammalian Rab35 in MTMR13-mediated processes, and thus in CMT4B, I aimed to investigate the interaction of MTMR13 with Rab35 in mammalian cells in the context of CMT4B. We first looked at a CMT4 disease mutant of MTMR13 lacking the last 80 amino acids of the DENN domain (MTMR13Δ) and found that while Rab35 interacts with MTMR13, it does not with MTMR13Δ. We continued to study the interaction by truncating MTMR13; pull-downs with Rab35 show that the DENN domain or the linker between the DENN domain and the subsequent GRAM domain independently do not bind Rab35, indicating that they are necessary but not sufficient for the interaction. However, when we truncated the protein after the first 869 amino acids, thus isolating the DENN domain and the DENN-GRAM linker (a construct we refer to as 869*), this construct strongly interacts with Rab35. Pull-downs with constitutively active or dominant negative Rab35 show that 869* binds both forms, possibly indicating that 869* traps Rab35 in an intermediate stage. 869* localises to dynamic intracellular structures resembling circles, dots, and tubules, and recruits Rab35 to these specific locations. 869* frequently encircles (i.e. appears to coat) markers of lysosomes and occasionally colocalises with them (i.e. appears to be on the same membrane). Under starvation it colocalises or is contiguous with LC3-II-positive structures, and we observed it triggers a delay in the formation of LC3-II punctae in cell expressing 869*. In order to distinguish between endosomal and lysosomal compartments, we treated cells expressing 869* with dextran. After washout dextran-positive vesicles are coated in 869*. Dextran does not exit these compartments and there is a decrease in dextran intensity, indicating it is degraded. Therefore 869* possibly coats terminal stations. LAMP1-positive structures that are either coated by a circular 869* structure or co-localise with 869*-positive rods appear enlarged and are dynamic: a tube coming off an 869* structure surrounding LAMP1 or co-localising with LAMP1 appears to transport LAMP1. We also observe no homotypic fusion of LAMP1 structures when one of them colocalises with, or encircles, 869*, and less dynamic LAMP1-positive structures when coated with 869* than when colocalising with 869* tubules. Lastly, LAMP2 is redistributed to the cell periphery. We further studied the structure of 869*-positive organelles using correlative light and electron microscopy. We found that 869* structures are coated in a filamentous array resembling a helical polymer, similar to complexes containing BAR-domain proteins, such as the retromer. Taken together our data indicate that 869* is likely involved in autophagosome maturation and fusion with lysosomes, and thus may be an intermediate in the formation of autolysosomes.

Wnts are thought to act as morphogens, spreading from producing cells and acting at a long-range to regulate various developmental processes. As such, Wnt gradients have been suggested to regulate growth, patterning and planar cell polarity (PCP) in the Drosophila wing. However, flies relying solely on membrane-tethered Wingless have normal wings with no apparent PCP phenotype, suggesting that Wingless diffusion is dispensable. This raises the following questions. (1) Could the gradients of other Wnts restore long-range signalling in membrane-tethered Wingless flies? (2) Do Wnt gradients contribute to PCP establishment in the wing? Indeed, it has been suggested that Wingless and DWnt4 act redundantly to establish PCP in the fly wing (Wu et al., 2013). To find out whether other Wnts are expressed during wing development, a set of reporters were generated for all seven DWnts. They showed that, in addition to Wingless, DWnt2, DWnt4, DWnt6 and DWnt10 are also expressed in the wing disc. Simultaneous knockouts of these Wnts were created to remove their possible contribution, leaving only membrane-tethered Wingless as the sole source of Wnt. This resulted in wings with no PCP defects, suggesting that Wnt diffusion is not required for PCP establishment in the wing. Subsequently, this study addressed the issue of whether Wnts are required for PCP. All DWnts were inhibited via a Nanobody trapping approach, which sequestered the Wnt chaperone Evi, and hence all DWnts, in the ER. The removal of all DWnts did not impact the PCP of the resulting wing, suggesting Wnts are not required for setting up PCP in the Drosophila wing. Using the same tools generated for addressing the PCP issues, this study also investigated the redundancy of Wnts in wing growth.

Cell division is essential for the propagation of all living organisms. In the vast majority of bacteria, cell division is carried out by the divisome, a multiprotein machine spanning the cell envelope. The divisome performs membrane invagination, peptidoglycan synthesis and remodelling, and eventually cell separation. Filaments of the tubulin homologue FtsZ form the scaffold for divisome assembly in the cytoplasm, the Z ring. Actin like FtsA tethers FtsZ to the inner membrane and facilitates recruitment of downstream divisome components. The late divisome component FtsN is a bitopic membrane protein that activates peptidoglycan synthesis. Recent genetic studies have suggested that the interaction of FtsA and FtsN is crucial for divisome integrity and function. So far biochemical and structural evidence for the FtsA-FtsN interaction have remained scarce or absent, and we have no mechanistic understanding of its putative signalling function. In this study, I show that Escherichia coli FtsA, upon binding the short, cytoplasmic part of FtsN, forms antiparallel double filaments on lipid monolayers. My complementary X ray crystallography studies provide a near atomic resolution structure of the FtsA double filament and a first insight into the putative FtsA FtsN binding site. FtsA filaments resemble the antiparallel double filaments formed by the actin homologue MreB involved in cell elongation. MreB filaments sense curvature and serve as a rudder in the cell elongation complex, ensuring oriented insertion of peptidoglycan around the cell circumference. Following from that I propose that curvature sensing is conserved in FtsA double filaments and, together with treadmilling FtsZ filaments, provides a guiding mechanism for membrane constriction and septal peptidoglycan synthesis during cell division.

As a substructure of cell cytoskeleton, the crosslinked actin filament networks (CAFNs) play a major role in different cell functions, however, the mechanical properties and the deformation mechanisms of CAFNs still remain to be understood. In this research, numerical simulations have been performed on a three-dimensional (3D) finite element (FE) model to mimic the mechanical properties of actin filament (F-actin) networks crosslinked by filamin A (FLNA). The simulation results indicate that although the Young’s moduli of CAFNs varies in different directions for each random model, the statistical mean value is in-plane isotropic. The crosslinking density and the actin filament volume fraction are found to strongly affect the in-plane shear modulus of CAFNs. In addition, a cantilever beam model is developed for dimensional analysis on the shear stiffness of CAFNs, which indicates that the in-plane shear modulus of CAFNs is mainly dominated by FLNA (i.e., crosslinkers). The dimensional analysis results agree well with the simulation results. In addition, the deformation mechanism of the CAFNs is also investigated by performing dimensional analysis and conducting numerical simulations. In physiological conditions, cells can undergo large deformation to respond to external stimulations and to support different cell functions. In large deformation situations, the CAFNs always show strong nonlinear elasticity to maintain the cell shape and integrity, which is known as strain stiffening. In this research, the nonlinear elastic properties of CAFNs are also studied by conducting numerical simulations. According to the results of numerical simulations, it can be demonstrated that the nonlinear elastic properties of crosslinked actin filament networks can be greatly influenced by the actin filament volume fraction, crosslinking density and components’ properties. The stress strain relationship of crosslinked actin filament networks is obtained and compared with experimental measurements reported in literature. In addition, the negative normal stresses of CAFNs are obtained from simulations, and the effects of the components’ properties on the negative normal stresses are discussed. In addition, to study the deformation mechanism of CAFNs in large strain regime, the effects of the components’ bending, torsional and tensile stiffnesses on the stress-strain curves of CAFNs are investigated. III Most of the biopolymer networks are proved to be viscoelastic and their mechanical properties are highly dependent on time and frequency. In physiological conditions, the viscoelastic behaviour of CAFNs could greatly affect the mechanical responses of cells, and different cell functions. Thus, it is important to study the viscoelasticity of CAFNs. The viscoelastic properties of actin filament networks crosslinked by filamin A are also investigated by conducting numerical simulations in this research. Both the creep and relaxation simulations are conducted in finite element method (FEM) software. In FEM simulations, different viscoelastic properties are applied to actin filament and FLNA respectively to probe their effects on the creep and relaxation behaviour of CAFNs. It is found that the FLNA affects the viscoelastic behaviour of CAFNs greater than actin filament. In addition, the effects of applied stress on the creep strain of CAFNs as well as the effects of applied strain on the relaxation stress of CAFNs are studied respectively. Simulation results also show that CAFNs with larger actin filament volume fraction creep less, however, larger actin filament volume fraction results in more relaxation in CAFNs. The creep strain and relaxation stress of CAFNs are found to reduce and increase with the actin filament volume fraction respectively. The crosslinking density is proved to have similar influences on the creep and relaxation behaviour of CAFNs. In addition, the dependence of dynamic shear modulus of CAFNs on the frequency and amplitude of applied loads are investigated. Results indicate that the storage shear modulus of CAFNs develops a plateau at low-frequency conditions, however, the loss shear modulus of CAFNs shows an increasing trend. At high-frequency conditions, both the storage and loss moduli of CAFNs scale with the frequency of applied loads. Moreover, the storage and loss shear moduli of CAFNs remain almost constant when the amplitude of applied load is small. However, they increase with the amplitude of applied load when the applied load reaches a critical value. The simulation results obtained in this research show good agreement with the experimental measurements and theoretical predictions reported in literature.

The inner mitochondria membrane (IMM) is densely packed with proteins necessary for mitochondrial activity, including oxidative phosphorylation (OXPHOS) complexes and transporters. Mitochondrial respiratory chain (MRC) complexes, which generate ATP by OXPHOS, associate in higher-order assemblies, known as supercomplexes (SC), that are structurally interdependent. Numerous patients carrying mutations in a single complex, indeed, present with combined enzyme deficiencies and, in particular, the destabilisation of complex I (CI) has been often described in the absence of complex III (CIII). To clarify the structural and functional relationships between complexes, we have analysed a MTCYB-deficient human cell line, unable to assemble CIII. Our results showed that in this line, CI biogenesis was blocked by preventing the incorporation of the NADH module, the last step of CI assembly, rather than decreasing its stability. Moreover, complex IV (CIV) biogenesis was impaired as well, and CIV subunits appeared sequestered within CIII subassemblies. Therefore, we propose that CIII is central not only for the formation of SC but also for the maturation of the other electron transport chain complexes. These results challenge the previous SC model that described the formation of fully assembled individual complexes before the association in SC. In contrast, they support a cooperative-assembly model in which the main role of CIII in SC is to provide a structural and functional platform for the completion of overall MRC biogenesis. Next, we identified and characterised the interaction between CIV and the mitochondrial calcium uniporter complex (MCUC), responsible for mitochondrial calcium uptake and homeostasis. Our data, indeed, showed a specific physical interaction between this ETC enzyme and various subunits of MCUC in first and second dimension blue native PAGE, SILAC labelling/pulldown, and BioID proteomic-based methods. We then investigated the effects of this association with CIV, measuring enzyme activity and mitochondrial respiration, but also on MCUC and CIV distribution in the IMM, by N-structured illuminated super-resolution microscopy (N-SIM). Our results showed a specific reduction in CIV activity in the absence of MICU1, the main regulator of the uniporter, but surprisingly no effects were observed after MCU downregulation or pharmacological inhibition of mitochondrial Ca2+ entry. Instead, the lack of CIV seems to have an impact on MCU-containing complexes formation and induces a re-localisation of MCU from the cristae membrane to the inner boundary membrane. Further experiments will be performed to shed light on the physiological relevance of this interaction.

Protein secretion in eukaryotic cells begins with COPII-mediated transport from the endoplasmic reticulum (ER). The COPII coat assembles on the cytoplasmic side of the ER membrane in consecutive stages. The inner coat comprises the Sar1-Sec23/Sec24 heterotrimeric complex and is recruited through the activation of the Sar1 GTPase. Sar1 regulates the assembly/disassembly cycle of the coat, Sec23 acts as its GTPase activating protein (GAP), whereas Sec24 is involved in selecting cargo. The cage-forming Sec13/Sec31 outer coat is recruited by Sec23 to deform the membrane and allow budding of the cargo-loaded vesicle. My work provides evidence that the essential connection between the two coat layers is established by multiple different interfaces between Sec23 and the Sec31 disordered proline-rich domain. The Sec31 unstructured domain comprises known Sec23 binding sites, namely the ‘active fragment’ and short PPP motifs. My data demonstrate that although these interfaces are important for the inner/outer layer interaction, they are individually dispensable. This highlighted the existence of another previously undefined interface that uses an electrostatic interaction between net positively charged clusters of the Sec31 disordered domain and a negatively charged patch on the membrane-distal surface of Sec23. My work further reveals that PPP motifs and positively charged clusters within a disordered region could maintain the inner/outer coat interface irrespective of the underlying sequence. Moreover, my experiments suggest that in the regulatory protein Sec16, which also contains disordered domains, similar features have been adapted to prime COPII components for vesicle formation and to regulate their assembly. Lastly, I demonstrate that the Golgi-localised tether Grh1 can compete with Sec31 using similar disordered domain features to potentially stimulate uncoating and thus permit membrane fusion. An interface generated through multiple transient interactions highlights the capacity for their utilisation by sequence-divergent proteins to compete with the coat components during coat regulation, or to achieve its efficient disassembly. The proposed type of combinatorial interface explains how a stable structure can form robustly enough to generate vesicular carriers from a protein rich membrane, but also remain permissive to modifications and able to be easily disassembled in the appropriate context.

Ubiquitination and deubiquitination are the important types of post-translational modifications that are involved in the regulation of most complex cellular processes. USP17, a member of the ubiquitin specific protease (USP) family of deubiquitinases, has been shown to be required for cell cycle progression, cell motility, endocytosis and our previous studies have preliminarily shown it is required for lysosome positioning. The research presented in this thesis further confirm the role of USP17 in lysosome trafficking and explore the impact upon lysosomal function as well as its potential mechanism. The data presented in Chapter 3 of this thesis shows the impact of USP17 upon lysosomal trafficking and lysosome-related processes. USP17 depletion affects lysosome peripheral trafficking and the degradation of extracellular matrix. Additionally, USP17 is essential for EGF triggered lysosome periphery trafficking and proper plasma membrane repair (PMR). The results in this chapter also suggest that USP17 may impact upon autophagy and TFEB nuclear translocation, however more research is still required to confirm these impacts of USP17. The data presented in Chapter 4 studies the potential mechanisms by which USP17 regulates the lysosome. Firstly, RCE1 iso2, a previously identified substrate of USP17, is shown not to be involved in lysosome trafficking. YIPF proteins are found to be regulated by USP17 and YIPF1, as well as YIPF2, may be implicated in lysosome trafficking. However, a clear mechanism via which USP17 directly regulates YIPF proteins is lacking. Moreover, the impact of USP17 upon Sec31A could indicate an impact upon ER-to-Golgi transport, but again, a clear mechanism is not apparent. Finally, USP17 is shown to be involved in the RNF26-based network which can regulate lysosome positioning via the ubiquitination of p62 to tether vesicles to the ER, and in particular USP17 is shown to deubiquitinate p62. Overall, the data illustrates that USP17 is important for lysosome trafficking and lysosomal function. These findings provide us with more understanding about the function of USP17 as well as providing additional insight into the impact of any inhibitors that target USP17 for future drug research and development.

Cell migration is a key cellular function involved in normal conditions such as embryonic development and immune cell travelling and in diseases such as in cancer metastasis. Cell migration is a tightly regulated process that requires the coordinated activity of different cellular structures including cytoskeletal networks and focal adhesions. Focal adhesions connect cells to the extracellular environments and are critical in cell migration by transmitting the force generated by cytoskeletal networks to the extracellular matrix allowing cells to move through these matrices. During cell migration, existing focal adhesions are disassembled, and new focal adhesions are formed at new sites allowing cells to detach from the matrix and continue migration. SUMOylation has been recently suggested to be involved in the regulation of these adhesion structures. However, previous studies lack the evidence of a direct role of SUMOylation at focal adhesion sites. The aims of this project are to investigate the role of SUMOylation in the regulation of focal adhesions and cancer cell migration. Treating MDA-MB-231 breast cancer cells with SUMOylation inhibitors significantly reduced their migration. These inhibitors also increased the size and turnover time of focal adhesions suggesting that SUMOylation is likely to facilitate cancer cell migration by enhancing focal adhesion turnover. In addition, bioinformatic analysis predicted the presence of SUMOylation consensus and SUMOylation interacting motifs in various focal adhesion proteins indicating a possible role of this posttranslation modification in the regulation of focal adhesions. SUMO specific proteases (SENPs) are critical in cancer metastasis by deSUMOylating associated proteins. Here, SENP2 was found to interact with major FA proteins including talin, vinculin, FAK and paxillin. This interaction could indicate a possible role of this protease in the regulation of SUMOylation levels of focal adhesion proteins. Knocking down SENP2 significantly reduced MDA-MB-231 cancer cell migration. It also increased the size and turnover time of focal adhesions suggesting that the reduction in cell migration of these cells was caused by impairment in FA dynamics. In order to investigate the direct role of SUMOylation in the regulation of focal adhesion dynamics, one of the main aims of this project is to investigate the effects of preventing SUMOylation of vinculin on Focal dynamics and cancer cell migration. This was achieved by replacing lysine residues with arginine in highly scored SUMOylation motifs in vinculin using mutagenesis. This way of investigation provides the opportunity to investigate the consequences of preventing SUMOylation of specific targets without influencing global SUMOylation that may affect many important cellular mechanisms. Replacing lysine at position 80 in the amino acid sequence in vinculin reduced its interaction with SUMO2 indicating that this site is a major SUMOylation acceptor site in vinculin. Expressing the K80R mutated vinculin in different cancer cells caused a significant increase in focal adhesion size and turnover time and significantly reduced their migration speed. As one of the main functions of SUMOylation is mediating protein-protein interaction, the last aim of this project is to investigate the consequences of K80R mutation on talin-vinculin interaction. Talin is one of the early recruited focal adhesion proteins and its interaction with vinculin enhances the stability of focal adhesions. Photobleaching FRET assay revealed that the K80R mutation enhances talin-vinculin interaction. This finding indicates that SUMO2 attachment to vinculin at K80 disassociates talin-vinculin interaction and mutating this site to prevent SUMO2 attachment allowed their interaction to last longer. Taken together, the finding in this project provides novel evidence that indicate a direct role of SUMOylation in the regulation of focal adhesion dynamics and consequently cell migration. It clearly suggests that SUMOylation regulates the disassembly of focal adhesions by SUMO2 conjugation to vinculin at K80 to disassociate its interaction with talin. The disassociation of talin-vinculin interaction induces the disassembly of existing focal adhesions, a critical requirement during cell migration. The prediction of several potential SUMO substrates in focal adhesions indicates a wider role of SUMOylation in the regulation of different functions of focal adhesions. Small molecule inhibitors against specific SUMO substrates could be promising therapeutic strategy for the development of anticancer drugs for cancer intervention.

Apoptotic cell death is almost universally considered a beneficial process that acts to both prevent and treat cancer. Exploiting this, drugs called BH3-mimetics that inhibit anti-apoptotic Bcl-2 proteins and thereby induce apoptosis, have been designed to sensitise cancer cells to apoptosis. There are a wide variety of BH3-mimetics targeting Bcl-2 proteins, and they display potent effects in some cancer types. However, the role of apoptosis in cancer may be more complex than initially thought, as oncogenic properties of apoptosis have also been described. Here, I have investigated the effects of BH3-mimetics (apoptotic-inducers) in different cancer types. Amongst other non-canonical effects such as inflammation, I have also found metabolic effects induced by BH3-mimetics. Most of these metabolic effects were cell-type dependent, however, the specific BCL-2 inhibitor ABT-199 (also called venetoclax) demonstrated the most consistent results. ABT-199 altered cellular metabolism by inhibiting mitochondrial respiration and inducing reductive carboxylation. Surprisingly, these metabolic alterations were independent of its well-known target BCL-2, although they were dependent on the transcription factor ATF4. ATF4 is involved in the integrated stress response (ISR) and it regulates the transcription of several metabolic enzymes. Importantly, both metabolic reprogramming and ATF4 upregulation have been described as mechanisms of chemoresistance. Thus, by demonstrating that ABT-199 rewires metabolism through ATF4 activation, it may indicate a mechanism to promote tumour resistance upon ABT-199 treatment, which may have further implications for therapy. Overall, I have found that induction of mitochondrial stress rather than apoptosis per-se induced by BH3-mimetics may be oncogenic. Thus, I have uncovered non-canonical roles of these drugs that are implicated in inflammation and metabolism of cancer cells.

Membrane type-1 matrix metalloproteinase (MT1-MMP) is a type I transmembrane proteinase that has been shown to promote the progression of different diseases, including rheumatoid arthritis (RA) and cancer, by enhancing cellular invasion. MT1-MMP promotes cellular invasion by degrading pericellular extracellular matrix (ECM) at the leading edge of migrating cells. MT1-MMP has been shown to localise to various motility-associated structures, including lamellipodia, filopodia, invadopodia, and podosomes. However, the molecular mechanisms of MT1-MMP localisation to these structures are not well- understood. Dr. Itoh’s research group and others have found that MT1-MMP cell surface exposure is achieved by intracellular trafficking of MT1-MMP-containing vesicles along microtubules and that kinesin superfamily motor proteins (KIFs) play a role. We have employed a siRNA library to screen for the responsible KIFs and found that the knockdown (KD) of four KIFs (KIF13A, KIF3A, KIF9, KIF1C) notably affected MT1-MMP-mediated cellular functions at the cell-matrix interface of human fibrosarcoma cells (HT1080). Interestingly, the KD of these KIFs did not affect the overall level of MT1-MMP at the cell surface, but it significantly influenced MT1-MMP localisation at the substrate-attached side of the plasma membrane. Live cell imaging experiments confirmed that KIF13A, KIF3A and KIF9 associate with MT1-MMP-positive vesicles. These experiments also showed that when cells are cultured atop a substratum, KIF13A and KIF3A collaborate to transport MT1-MMP within the central cytoplasmic areas. However, only KIF13A seemed to reach the cell periphery to deliver the proteinase to the tips of motility-associated structures. Bioinformatic analysis revealed that the expression profile of KIF13A and KIF9 is altered in several cancer types and that these two motor proteins might have cancer-specific roles, partially related to their involvement in MT1-MMP intracellular trafficking. Taken together these findings shed light on KIF-dependent MT1-MMP cell surface exposure and suggest KIF-driven MT1-MMP intracellular trafficking mechanisms as potential targets to control the un-wanted cell invasion that sustains the progression of diseases like RA and cancer.

Understanding the interactions between vesicles within a vesicle dispersion is key to understanding the stability of the dispersion. If the interactions between the vesicles are sufficiently repulsive the system will be stable. Whilst this behaviour is relatively easy to predict for simple colloid and vesicle dispersions. In complex systems, containing various additives such as salts, polymers, and polyelectrolytes, the interactions effecting the stability of the dispersion are more complex. In this study methodologies for directly measuring the interactions between two supported cationic surfactant bilayers using the colloidal probe atomic force microscopy technique were successfully developed. The surfactant vesicle system of interest in this work was the double tailed, cationic surfactant DIPEDMAMS, often used in commercial fabric softeners. The system was observed to form two co-existing lamellar phases which melt to forma a single fluid lamellar phase at around SI 40°C. The addition of different alkyl alcohols was used to modify the transition temperature to create supported bilayers for the colloidal probe technique. Electrolytes were observed to decrease the repulsive interactions between the bilayers with the identity of the salt determining the strength of the effect. Polyelectrolytes whilst having a similar impact as to the salts the depletion attraction was also observed and shown to be dependent on the molecular weight of the polymer. The interactions with uncharged polymers were complex due to the absorption of the polymer to the bilayers as observed by AFM imaging. The addition of fatty acids to simulate the breakdown of the surfactant molecules within the bilayers exhibited complex behaviour dependent on the nature of the fatty acid in question. All these measurements were supported by AFM imaging and x-ray scattering to decouple changes to the supported bilayers and the interactions between them. Finally, direct force measurements were compared to the behaviour of the bulk vesicle dispersions.

Eukaryotic cells have evolved molecular mechanisms that control organelle size, number and position. Molecular tethers are required for organelle positioning, multiplication and establishment of interorganellar contact sites. The balance between organelle tethering and motility determines the intracellular distribution of organelles and their segregation during cell division. In Saccharomyces cerevisiae, correct peroxisome distribution is achieved by the opposing processes of cortical anchoring in the mother cell and Myosin-dependent transport towards the bud. The Inp1-Pex3 tethering complex is required for peroxisome retention during cell division and for peroxisome positioning along the mother cortex. As has been postulated for other organelles, yeast peroxisomes interact with many cellular structures including the plasma membrane, ER, vacuole, mitochondria and lipid bodies. Components of some interorganellar peroxisomal contact sites have recently been identified whereas others are still completely uncharacterised including the plasma membrane- peroxisome (PM-PER) contact site. The work presented in this thesis identifies Inp1 as the first known plasma membrane-peroxisome (PM-PER) tether by demonstrating that Inp1 meets the predefined criteria which a contact site tether protein must adhere to. This thesis first describes a conserved Pex3 binding motif in the C-terminal region of Inp1. This motif bears a striking resemblance to the Pex3 binding site present on Pex19 and both in vitro and in vivo evidence is presented which illustrates that Pex19 and Inp1 can compete for binding to Pex3. In addition, the N-terminal 100 amino acids of Inp1 are shown to localise to the plasma membrane, bind to phosphatidylinositol 4,5-bisphosphate (PI(4,5)P2) and, when artificially attached to the peroxisomal membrane, restore retention by relocating peroxisomes to the cell periphery in inp1Δ cells. In this study, Inp1 is shown to be present in the correct sub-cellular location to interact with both the plasma membrane and peroxisomal membrane and the data illustrates the structural and functional capacity of Inp1 to be a PM-PER tether. Through detailed analysis of the molecular function of Inp1, the work in this thesis identifies a novel role for Inp1 as a PM- PER tether and concludes that tethering of peroxisomes to the plasma membrane is required for peroxisome retention. This is the first molecular characterisation of the PM-PER tether and has allowed for the proposal of a new model for peroxisome retention.

Ezrin, a member of the ERM (Ezrin/Radixin/Moesin) family of proteins, serves as a crosslinker between plasma membrane and actin cytoskeleton. It provides structural links to connect cell cortex and plasma membrane, acting as a signal transducer in multiple pathways during migration, proliferation, and development. Ezrin is also considered crucial during cancer metastasis, due to its altered expression levels and subcellular localization. Nevertheless, the mechanism how Ezrin promotes cancer progression remains unclear. In this thesis, the primary role of Ezrin phosphorylation in the relationship between cell motility, cell mechanical properties and cytoskeleton organization was investigated as follows: Firstly, long-term live-cell imaging was used to evaluate the effect of Ezrin phosphorylation on cell migration. The key results showed that cells expressing active Ezrin T567D migrated faster, with more protein accumulating at the cell rear. Secondly, the changes of cell mechanical properties caused by Ezrin were explored by atomic force microscopy (AFM). It revealed that transfection of active Ezrin T567D decreased cortical stiffness but increased cytoskeleton stiffness. Thirdly, image quantification approaches were carried out to study the effect of Ezrin on three cytoskeleton (actin filaments, microtubules and vimentin) as well as nuclear mechanical properties. The results showed that Ezrin T567D transfection led to thicker actin stress fibers and longer vimentin filaments. In the end, a sandwich model was developed to study bleb based migration in which Ezrin is also involved. Bleb based migration was observed by confining cells between two pieces of polyacrylamide gels. This model showed potential for the future investigation of Ezrin phosphorylation in bleb based migration. Altogether, we have fundamentally revealed that dynamic regulation of phosphorylated Ezrin played a pivotal role in cell motility and cell mechanical properties by altering the cytoskeleton’s microstructure. These findings are useful for a better understanding of how Ezrin is biophysically involved in cancer progression.

The plasma membrane separates the cell interior from the external environment, protecting the cell from adverse conditions and pathogens. However, cells need to continually interact with their surroundings to survive and perform their metabolic functions. These tasks are performed by a great variety of macromolecules that crowd the cellular membrane, from cell-surface receptors to transport channels. The physical and dynamic properties of many of these receptors, as well as their interaction with their target ligand, are poorly understood. This is mainly due to the technological challenges posed by studying single-molecule properties and interactions on living cells, in particular the ability to reach the required spatial and temporal resolution at which these interactions take place. This thesis focuses on the design and development from scratch of an innovative, fully custom-made microscopy platform, which combines light-sheet fluorescence microscopy and optical tweezers for the study of the physical properties of receptors in the plasma membrane of living cells. The technological developments regarding fluorescence microscopy include the demonstration of single-molecule resolution using light-sheet microscopy on single cells. With this method, the tracking of individual receptors in the plasma membrane is achieved with a temporal resolution of tens of milliseconds. Only the top surface of adherent cells is imaged. This prevents artefacts that typically arise in the cell membrane in contact with hard glass substrates, and enables measurements of cell-surface receptor mobility with increased physiological relevance. Furthermore, we have developed an optical tweezers setup optimised for the application of vertical forces on living cells using a counter-propagating beam configuration. Single-molecule force spectroscopy on living cells is demonstrated by quantification of the strength of antigen-antibody bonds. The response of cells to weak and localised mechanical stimuli is likewise measured.

The paired lobes of the male Drosophila accessory gland (AG) contain two epithelial cell types, main cells (MCs) and secondary cells (SCs). SCs are a useful in vivo genetic model system to study fundamental mechanisms regulating cell growth and secretion, and they share several properties with the human prostatic gland. The SCs in adult virgin males grow with age in a BMP-dependent manner, and this growth can be further enhanced upon mating. During my DPhil, I have developed the SC as a model to study steroid receptor-dependent growth regulation. The fly nuclear steroid receptor, the ecdysone receptor (EcR), and the human androgen receptor (AR) share similar structural and functional domains. I show that a complex interplay between the BMP and EcR signalling pathways is involved in the regulation of SC growth in both virgin and mated males, an interaction which may allow mated males to rapidly replenish the expelled contents of their AG. The growth that occurs after mating is partly accounted for by genome endoreplication which is BMP/EcR-dependent, as is the secretory activity of SCs. I show that SCs can synthesise the hormone ecdysone, much like the prostate gland which produces the most active form of androgen. Interestingly, SC growth that occurs in virgin males requires ecdysone, but in mated males, endoreplication-dependent growth occurs independently of this hormone. The switch from hormone-dependent steroid receptor signalling to hormone-independent signalling mirrors the transition from androgen-dependent prostate cancer to the more aggressive and incurable form of prostate cancer, androgen-independent ‘castration-resistant prostate cancer’ (CRPC). I demonstrate that the cell cycle regulators, Cyclin D (CycD), Cyclin E (CycE), Retinoblastoma (Rbf) and E2F1, which are all implicated in prostate cancer, are key regulators of SC growth and endoreplication both in virgin and mated males. Furthermore, CycD/Rbf/E2F1/CycE interacts with the BMP/EcR signalling axis to regulate the transition from hormone-dependent to hormone-independent steroid receptor signalling in SCs, mirroring some of the mechanisms associated with the transition to CRPC. Overall, my work suggests that there are parallels between the physiological roles of the EcR’s switch to hormone-independence in SCs and the pathological changes seen in CRPC, and indicates that further genetic analyses in SCs may unravel the complex signalling mechanisms that mediate the transition to this lethal form of cancer.

The shift in lactate metabolism of industrial mammalian cell lines is a widely studied phenomenon that has yet to be fully understood. Many mammalian cell lines, including Chinese Hamster Ovary (CHO) cells, which are mainly used to manufacture therapeutic glycoproteins, exhibit an unusual shift in lactate metabolism when grown in batch or fed-batch suspension cultures. Understanding the underlying mechanism of this metabolic switch and whether it is beneficial in terms of productivity, might open up large areas of currently infeasible process space for exploitation. Herein, a combined experimental and computational approach utilizing Flux Balance Analysis (FBA) coupled with Multivariate Data Analysis (MVDA) was developed. The latest available CHO genome-scale metabolic model (iCHO1766) was studied in detail, in order to gain further insights into the metabolic effects that cause this shift in lactate metabolism and its impact on culture performance. Experimental data pre- and post-lactate shift were collected from batch and fed-batch cultures across multiple cell lines and were used to constrain the model. A novel carbon-based constraining algorithm was developed and employed in order to rationally and systematically refine the accuracy of the model’s predictions. Using principal components analysis (PCA) and partial least squares discriminant analysis (PLS-DA) well defined and clearly separated clusters in the solution space corresponding to the lactate producing (LP) and lactate consuming (LC) metabolic states were identified. Furthermore, through analysis of the PC loadings and variable importance in projection (VIP) scores key reaction groups whose activities differed significantly between the two metabolic states were identified. These key metabolic differences were used to formulate hypotheses regarding potential mechanisms and actuators of the switch in lactate metabolism. Based on the computational analysis, experiments exploring enzymatic inhibitors, additional media components, and the introduction of a non-native enzyme into the CHO genome were performed in an effort to test the initial hypotheses.