A rise in intracellular Ca2+ concentration is key to controlling both short term and long term Ca2+ dependent processes which include secretion, metabolism and gene expression, cell growth and proliferation. Store operated Ca2+ channels (SOCs), which are activated by the depletion of Ca2+ from internal Ca2+ stores, the main store being the endoplasmic reticulum (ER), are the major route for Ca2+ influx in non-excitable cell types. Stromal interacting molecule 1 (STIM1) is a Ca2+ sensing protein located in the endoplasmic reticulum (ER). Depletion of ER calcium stores triggers oligomerisation and subsequent translocation of STIM1 from its reticular location to specialized endoplasmic reticulum-plasma membrane (ER-PM) junctions where it forms STIM1 puncta and interacts with the SOC channel, Orai1. This induces the clustering of Orai1 into a functional tetrameric pore which is permeable to Ca2+ ions, enabling Ca2+ entry into the cell. The precise mechanism by which STIM1 is recruited to the plasma membrane to activate SOCs and the plasma membrane components involved in targeting STIM1 to the plasma membrane are largely unknown. In this study the mechanisms underlying movement of STIM1 to the plasma membrane and its accumulation at ER-plasma membrane junctions was explored in HeLa cells. In the initial part of this study I investigated whether the movement of STIM1 to the plasma membrane is an ATP-dependent process. I found that depletion of cytosolic ATP can stimulate STIM1 puncta formation in HeLa cells and that the formation of STIM1-Orai1 complexes at the plasma membrane is unaffected in these conditions. Inhibition of ATP synthesis also initiated the loss of phosphatidylinositol 4,5-bisphosphate (PtdIns(4,5)P2) from the plasma membrane. ATP depletion did not affect the structure of the microtubule cytoskeleton. These results suggest that the translocation of STIM1 and the formation of STIM1-Orai1 complexes is an ATP independent process which is not due to the disruption of microtubules and support a diffusional model for STIM1 puncta formation. It has been suggested that an additional interaction of the C-terminal polybasic domain of STIM1 with plasma membrane phosphoinositides could contribute to STIM1 puncta formation prior to binding to Orai1. I investigated the role of phosphoinositides in the formation of STIM1 puncta and SOCE in response to store depletion. Treatment of HeLa cells with inhibitors of the phosphatidylinositol 3-kinase (PI3K) and phosphatidylinositol 4-kinase (wortmannin and LY294002) partially inhibited formation of STIM1 puncta. Additional rapid depletion of PtdIns(4,5)P2 resulted in more substantial inhibition of the translocation of STIM1-EYFP into puncta. The inhibition was extensive at a concentration of LY294002 (50 μM) that should primarily inhibit PI3K consistent with a major role for PtdIns(4,5)P2 and PtdIns(3,4,5)P3 in puncta formation. Depletion of phosphoinositides also partially inhibited SOCE. Overexpression of Orai1 resulted in a recovery of translocation of STMI1 into puncta following phosphoinositide depletion and under these conditions SOCE was increased to above control levels. These observations support the idea that phosphoinositides are not essential but contribute to STIM1 accumulation at ER-PM junctions with a second translocation mechanism involving direct STIM1/Orai1 interactions. It was recently reported that STIM1 and Orai1 may function within a macromolecular complex involving other unidentified proteins. In this study I have identified that Golli-BG21, a member of the myelin basic protein (MBP) family, can directly interact with STIM1. Golli interacts with the C-terminal domain of STIM1 in both in vitro and in vivo binding assays and this interaction may be modulated by intracellular Ca2+ concentration. Golli also colocalises with full length STIM1 and Orai1 complexes in HeLa cells following store depletion. Overexpression of Golli reduces SOCE in HeLa cells but this inhibition is overcome by overexpressing STIM1. We therefore suggest that Golli binds to STIM1-Orai1 complexes to negatively regulate the activity of SOCs.

Caveolae are important microdomains found in the plasma membrane, which act as signalling hubs in endothelial cells (ECs) and smooth muscle cells (SMCs). Thus, disruption of caveolae by membrane cholesterol depleting agents such as methyl β-cyclodextrin (M-β-CD) has various functional effects on arteries, including impairment of endothelium-dependent relaxation and augmentation of smooth muscle cell contraction independently of the endothelium. The aim of this study was to test the hypotheses that caveolae modulate contractility in rat femoral artery, and more specifically the response to BKCa channel activity. Different methods were used in the study including; transmission electron microscopy (TEM), immunohistochemistry, immunocytochemistry and Western blot, and myography. M-β-CD was used in this work to disrupt caveolae from the plasma membranes of SMCs and ECs of rat femoral artery. TEM examination showed that caveolae were present throughout the plasma membrane of the SMCs and ECs in femoral artery. TEM examination also showed that treatment of the artery with M-β-CD, causes removal of caveolae from most of plasma membrane of the SMCs and ECs. Immunohistochemistry and immunocytochemistry results showed that cav-1, cav-3 and BKCa channels are co-expressed in SMCs, and cav-1 and BKCa channels are present in ECs. The presence of these proteins in femoral artery tissue lysate was confirmed by Western blot. Femoral artery contraction studies using myography showed that caveolar disruption by M-β-CD caused a significant increase in the contraction in endothelium-intact artery rings in response to 20 mM K+ and 100 nM BayK-8644 (20 K/ Bay K). In endothelium-denuded artery rings and/or after incubation of endothelium-intact artery rings with L-NAME (an eNOS inhibitor), there was a non-significant increase in the contraction in response to 20K/ BayK. Incubation of endothelium-intact artery rings with TEA+ or IBTX, both BKCa channel inhibitors, caused a significant increase in the contraction to 20 K/ Bay K. Incubation of endothelium-intact artery rings with L-NAME, reduced the contraction to TEA+ and IBTX. These results provide evidence that contraction of rat femoral artery is inhibited by basal nitric oxide release from the endothelium, and BKCa channels in the SMCs might have a role in this vasorelaxation. Further femoral artery relaxation studies showed that the integrity of caveolae was important for the vasorelaxation of arteries (pre-contracted with 20 K/ Bay K) when BKCa channels were directly activated by NS-1619. In contrast, vasorelaxation by isoproterenol and forskolin in endothelium-denuded artery was not significantly altered after caveolae were disrupted. These data together suggest that caveolae are an important factor in maintaining Ca2+ homeostasis in SMCs in rat artery femoral due to regulation of contractile activation.

Calcium (Ca2+) signalling regulates many neuronal functions including neurotransmission, axonal growth and development. Neuronal calcium sensor-1 (NCS-1) has been shown to be involved in many of these processes. On Ca2+ binding, NCS-1 changes conformation and exposes a hydrophobic binding pocket. In yeast, NCS-1 binds to a PI4-kinase orthologue required for survival. In mammalian cells, NCS-1 is localised to the Golgi and plasma membranes and has been linked to multiple target proteins that have roles in neuronal signalling. NCS-1 has been shown to regulate the P/Q Ca2+ channel subunit Cav2.1; although no direct binding interaction has been identified between the proteins. The Cav2.1 C-terminal tail contains two Ca2+-sensor binding regions, the IQ domain and the calmodulin (CaM) binding domain (CBD). The first part of this study investigated NCS-1 or CaM and Cav2.1 interactions using biochemical and biophysical interactions. Pull-down analysis found that NCS-1 binds to a Cav2.1 C-terminal peptide in a Ca2+-dependent manner. Use of nuclear magnetic resonance spectroscopy also showed that the IQ domain of Cav2.1 bound to NCS-1 in the presence of Ca2+, though the NCS-1 region involved in this interaction could not be identified. The second part of this study investigated NCS-1 in the model organism C. elegans. In the worm, NCS-1 is expressed predominantly in sensory neurons. An ncs-1 null mutant worm strain (XA406) was previously shown to be defective in isothermal tracking and this was linked to a requirement for NCS-1 in memory and learning. To ensure this behaviour was not caused by a locomotion or neurotransmission phenotype, these behaviours were quantified and compared to the wild-type strain. No effect of the ncs-1 null mutation was found in a quantitative body-bend assay or in an assay of aldicarb resistance. The temperature-linked behaviour was further characterised using an acute assay for temperature-dependent locomotion (TDL). In this assay, the rate of locomotion of wild-type worms decreased when the temperature was elevated from 20oC to 28oC. In contrast, the rate of locomotion of the ncs-1 null worm was significantly increased at the higher temperature. This distinct phenotype was exploited to quantify the rescue of the null strain by expression of wild-type NCS-1 and to identify potential mechanisms involved in NCS-1 function. It was established that NCS-1 regulated TDL when expressed in AIY neurons. Using information from previous studies, key structural elements of NCS-1 were investigated by expressing NCS-1 with specific point mutations or deletions. N-terminal myristoylation of NCS-1 was not functionally required. In contrast, the N- and C-terminal clefts of the hydrophobic pocket of NCS-1 were shown to be physiologically important while the C-terminal tail was not essential for function in the TDL assay. These findings allowed discrimination between two potential modes of interaction of NCS-1 with its target proteins in a physiological context.

In common with other hollow organs, the gastric mucosa consists of exocrine and endocrine epithelial cells and sub-epithelial stromal cells, all of which secrete mediators into the tissue microenvironment that define mucosal architecture and maintain organ function. Gastric myofibroblasts are an important stromal cell population involved in maintaining mucosal integrity, and are implicated in pathophysiological processes including chronic inflammation, fibrosis and cancer. Urokinase plasminogen activator (uPA) and its inhibitor plasminogen activator inhibitor (PAI)-1 are secreted by gastric epithelial and stromal cells and, together with the uPA receptor (uPAR), are believed to be involved in epithelial-mesenchymal signalling. However, the role of the uPA system in regulating gastric mucosal morphological homeostasis and responses to acute and chronic challenges is not yet fully understood. In this thesis, it was shown that transgenic PAI-1-H/Kβ mice, which have increased expression of PAI-1 in gastric parietal cells, have elevated concentrations of circulating PAI-1 and develop age-dependent increases in corpus mucosal thickness, independent of changes in parietal cell and myofibroblast abundance. PAI-1 did not have a direct trophic effect on gastric epithelial cells, indicating that PAI-1 might act via an indirect mechanism to modulate gastric epithelial cell turnover. In order to determine whether PAI-1 influences the gastric tissue microenvironment via myofibroblast gene expression, global transcript expression profiles of gastric myofibroblasts from wild-type and PAI-1 null (PAI-1-/-) mice were compared. Whole genome microarrays and subsequent validation by immunofluorescence indicated that mouse antral myofibroblasts are highly heterogeneous, including both desmin positive and desmin negative cells with distinct global transcript expression profiles. Furthermore, there was evidence that antral myofibroblasts displayed phenotypic plasticity in vitro, developing a neuroendocrine-like phenotype marked by expression of secretogranin-2. Plasticity may be functionally significant in vivo, supporting the role of myofibroblasts in wound healing. The role of PAI-1 in gastric mucosal responses to chronic Helicobacter infection was investigated using the H. felis model in wild-type, PAI-1-/- and PAI-1-H/Kβ mice. Both absence of PAI-1 and increased expression of gastric PAI-1 were protective against Helicobacter-induced preneoplastic gastric histopathology. Responses to acute gastric mucosal injury were investigated using intragastric indomethacin administration. Gastric PAI-1 protected against the development of lesions, whilst exogenous PAI-1 exacerbated the development of indomethacin induced gastric mucosal injury. Gastric uPA expression did not effect the development of lesions, whilst absence of uPAR tended to exacerbate lesion development. Taken together, the data presented in this thesis suggest a broadly protective role of PAI-1 in the gastric mucosa. A common therapeutic strategy for the prevention of NSAID-induced gastric injury, ulcer complications and progression of Helicobacter-induced gastric preneoplasia might be emerging, aimed at specifically increasing PAI-1 bioavailability in the gastric mucosa.

Post-translational modification of proteins via ubiquitination is mediated by three enzyme families; E1 activating enzymes, E2 conjugating enzymes and E3 ligases, all of which work in a hierarchical manner to facilitate different forms of protein ubiquitin ranging from mono-ubiquitination to the formation of different forms of ubiquitin chains (Ciechanover et al., 2000). Deubiquitinating enzymes (DUBs) act to remove ubiquitin from modified substrates. Apart from the classic interactions within the E1-E2-E3 enzymatic cascade, an unusual non-hierarchical interaction has been observed between some E2 enzymes and a DUB called OTUB1 (Markson et al., 2009). This observation raises interesting questions concerning the molecular mechanisms and specificity of this unusual E2:DUB partnership. In this study, systematic yeast two-hybrid (Y2H) screens were performed between all human E2 and DUB proteins to analyse the extent of E2:DUB interactions. Putative partnerships between OTUB1 and UBE2D1, UBE2D2, UBE2D3, UBE2D4, UBE2E1, UBE2E2, UBE2E3 and UBE2N were identified. These data correlate well with data from other independent studies, including HTP Y2H screens (Markson et al., 2009) and mass spectrometry (Sowa et al., 2009). An N-terminal truncated form of OTUB1 (ΔNOTUB1) was generated by removing a predicted 39aa N-terminal disordered region (Edelmann et al., 2009). Using this construct in combination with wild type (WT) OTUB1, complementary biophysical studies were performed to investigate the formation of complexes with UBE2D2 and UBE2E1 as these represented the strongest interactions detected in preliminary Y2H studies. Gel filtration chromatography showed convincing complex formation for both ΔNOTUB1:UBE2D2 and ΔNOTUB1:UBE2E1 in 1:1 stoichiometry. The thermodynamic profile of each complex was measured by ITC suggested a stronger affinity between ΔNOTUB1:UBE2D2 (Kd 3.89 µM) than observed for the ΔNOTUB1:UBE2E1 complex (Kd 16.55 µM). The n values for both complexes are 1.16±0.06 sites and 0.92±0.03 sites respectively, confirming that both complexes adopt a 1:1 stoichiometry. Observing the UBE2D2 (1H15N)-HSQC NMR spectral changes that occurred upon addition of unlabelled ΔNOTUB1 allowed the identification of potential residues of contact between the two proteins. From this study, we were able to predict that the 1st α-helix, the L1 loop of the 3rd and 4th β-sheet, the L2 loop connecting the 4th β-strand and the H2 α-helix within UBE2D2 were likely to be the binding surfaces for OTUB1. Point mutants corresponding to predicted contact residues in UBE2D2 were generated and tested in Y2H studies to determine their role in facilitating the formation of both E2:OTUB1 and E2:E3-RING complexes. This data suggests that in some, but not all cases, OTUB1 and E3-RINGs bind competitively to the same interface on E2 proteins. Preliminary immunofluorescence studies show that partner proteins predominantly co-localise in the cytoplasm, except UBE2E1 which is predominantly nuclear. Data from this study allowed us to propose a model of how OTUB1:UBE2D2 complex may forms and functions. Significantly, many of these predictions have now been verified by independent structural studies and subsequent live cell microscopy studies in our lab.

Prostaglandin F2α (PGF2α) is a myometrial stimulant, both PGF2α and its receptor are reported to increase towards parturition. With high PGF2α levels correlating with pre-term birth, especially those caused by infection. The mechanism by which PGF2α exerts its affects on uterine excitation-contraction coupling is unknown. The aim of this work was to describe the relationship between force and [Ca2+]i in pregnant rat myometrium and the effects of PGF2α on this relationship. The mechanism by which PGF2α exerts its affects was also investigated, focused on the role of Ca2+ entry mechanisms. PGF2α was examined on both longitudinal strips and isolated myocytes from pregnant Wistar rats. Strips were loaded with the Ca2+ sensitive indicator Indo-1AM and simultaneous recording of [Ca2+]i and force were made using a photometric system combined with force measurements, at a sampling rate of 1KHz to get good temporal resolution. Cells were isolated using Liberase Blendzyme 3, loaded with Fluo-4AM, and [Ca2+]i recorded using a Nipkow Disk based confocal imaging system. Pregnant myometrial tissue gave three contractility patterns in control conditions; two spontaneously active, giving either irregular contractions, or smooth phasic contractions, while the third was not spontaneously active but responsive to high-K+ stimulation. All contractions were preceded by a rise in [Ca2+]i which was dependent upon Ca2+ entry through VOCC, while synchronisation was dependent upon gap junctions. PGF2α increased myometrial contractility. In spontaneous tissue there were two responses; firstly an increase in amplitude, duration and frequency of phasic contraction while the second resulted in a tonic-like contraction which lasted for the duration of agonist application. On quiescent tissue, PGF2α resulted in the imitation of spontaneous activity, which ceased upon removal of the agonist. [Ca2+]i mirrored force in respect to frequency and duration, but PGF2α did not increase the amplitude of Ca2+ transient above that seen under normal spontaneous activity. The increase in amplitude of force induced by PGF2α is caused by an increase in the frequency of Ca2+ spikes within the Ca2+ spike burst. PGF2α resulted in the oscillatory release of Ca2+ from the SR in the form of propagating Ca2+ waves, initiated at one end of the cell. Re-admission of external Ca2+ resulted in activation of a nifedipine-resistance Ca2+ influx sensitive to La3+, which suggests that the stimulant action of PGF2α is associated with activation of a Ca2+-release Ca2+-entry coupling mechanism leading to opening of a SOCE pathway. In addition to this PGF2α increase both force and [Ca2+]i when applied in the presence of CPA and nifedipine, suggesting that PGF2α also works in part through a receptor operated and / or non-selective cation channels.

The observation of single biomolecules via optical microscopy eliminates all the implicit averaging of ensemble techniques and thereby provides access to the heterogeneity of molecular systems that will be the key to at least some biological functions. The implementation of photothermal microscopy at the University of Liverpool to achieve the detection of single gold nanoparticles over long times at high signal-to-noise-ratio is presented here, along with the development of Photothermal Raster Image Correlation Spectroscopy, PhRICS. PhRICS was shown to be equally effective as Photothermal Absorption Correlation Spectroscopy, PhACS, in the determination of the hydrodynamic diameter of colloidal gold nanoparticles in solution. The use of gold nanoparticles as labels for biomolecules has been of great interest due to their favorable optical properties and surface chemistry. The development of a new strategy for the covalent biofunctionalisation of gold nanoparticles with a single maleimide group is described. Nanoparticles functionalised this way were used to label FGF-2 protein and heparin-derived oligosaccharides. Both the PhRICS and the new nanoparticles developed in this thesis are combined to investigate the heterogeneity of FGF binding to heparin-derived oligosaccharides and to HS in the pericellular matrix of Rama 27 fibroblasts. The cooperativity of the interaction of FGF-2 with a dodecasaccharide is investigated. Although oligomerization of FGF-2 on the dodecasaccharide is observed, it is not cooperative. The first photothermal imaging of FGF-1 in the pericellular matrix of Rama 27 fibroblasts reveals that its diffusion is quite different from FGF-2. Imaging of FGF-2 on live cells is also revisited and probed with PhRICS. In comparison to photothermal tracking, PhRICS indicates that FGF-2 diffuses faster than first thought, and that the pericellular matrix is remodeling at timescales much shorter than previously observed.

Snake venom is a critical evolutionary innovation enabling venomous snakes to become successful limbless predators; it is therefore vital that snakes possess a highly efficient venom production system to maintain their predatory arsenal. The dynamics of venom synthesis and the regulatory mechanisms by which the expression of venom protein-encoding genes is controlled are little understood. The overarching aim of the work described in this thesis was to investigate the dynamics of venom synthesis in terms of the production of venom in juvenile snakes from birth and in the immediate replenishment of depleted venom stores using the African Puff Adder (Bitis arietans) as a model viperid species. We also aimed to investigate the underlying control mechanisms which regulate venom production Initial studies revealed a remarkable degree of intra-species variation in the protein profile, immunoreactivity and enzyme activity of venom between B. arietans specimens originating from different geographical origins across sub-Saharan Africa and Arabia, and within the same geographical origin. Variation was most evident in the snake venom metalloproteinases (SVMPs); toxins with a primary role in the haemorrhagic and tissue-necrotic pathologies suffered by envenomed victims. Our findings are of therapeutic importance as observations could translate into variations in the clinical manifestation of B. arietans envenoming and affect the patient response to antivenom treatment. To monitor the synthesis of venom proteins, we exploited the unusual stability of messenger RNA in lyophilised snake venoms as an alternative source of transcriptionally active mRNA to venom gland tissues, thus avoiding the requirement to sacrifice specimens for transcriptome analysis. Our optimised approach was used to quantitatively track changes in expression of venom protein-encoding genes. Our results showed that the gene expression, protein composition and functional activity of juvenile B. arietans venom did not appear to significantly change over time from birth to four years indicating that some aspects of venom are genetically hard-coded. We also showed that venom resynthesis triggered by venom expulsion peaked between days 3-7 following depletion of venom, with different protein families expressed in parallel. It appeared that venom production in both adult and juvenile specimens occurs very rapidly, presumably to ensure that venomous snakes retain their ability to efficiently predate and remain defended from predators. Our findings suggest that highly regulated mechanisms may be in place to ensure the rapid synthesis of venom. As it appeared that different venom protein families shared similar expression levels during venom replenishment, we investigated whether venom genes also showed similarities in their genomic location, organisation and structure, and regulatory elements responsible for controlling expression levels. We have taken the first steps to begin to investigate the genomic structure and organisation of genes encoding venom protein families expressed in B. arietans venom.

Pancreatic acinar cells release hydrolytic enzymes into the gut where they are responsible for digestion. When this process goes awry, these enzymes can be activated within the pancreas itself, characteristic of a severe and systemic inflammatory disease: acute pancreatitis. A large body of data has highlighted that the IP3R Ca2+-release channel is a central mediator of the toxic Ca2+ signals associated with the onset of early enzyme activation and cellular necrosis. Not only is the pancreatic acinar cell a focus of investigation into the onset of this acute disease, it is also investigated as a paradigm cell type in the fields of Ca2+ signalling; stimulus secretion coupling; and cellular polarity. The mitochondria, which are the ATP-producing powerhouses and Ca2+ signal modulators of the cell, have a well documented polarised distribution in pancreatic acinar cells. This is a thesis of three main parts. Initial experiments aimed to determine the mechanisms of inhibition of IP3-mediated Ca2+ signals by xanthines. Xanthines were shown to inhibit the IP3R and did not inhibit Ca2+ influx directly. A range of methylxanthines were shown to have varied effects on Ca2+ oscillations, however another role of methylxanthines, PDE inhibition, is likely to be involved in this. No significant differences were seen between xanthines when inhibiting IP3-mediated Ca2+ signals or necrosis, but inhibition of the Ca2+ signal does protect the cell against mitochondrial dysfunction. Whilst no xanthines tested offered any superior protection against pathological Ca2+ signals, this only goes to highlight the need for better pharmacological modulators of IP3-mediated Ca2+ signals. The development of multiplex assays to record pathological criteria such as Ca2+ signals and necrosis in a rapid and more user-friendly way was also undertaken. Assays for determining ∆ΨM and necrosis were successfully developed but throughput was limited by the primary cell isolation procedure. Finally, the most fascinating part of this research involves the mitochondrial network in pancreatic tissue. We investigate the mechanisms and signals involved in how the mitochondrial distribution comes into being and what effect pathological signals have on this distribution. This mitochondrial belt is formed in a Ca2+-dependent manner upon stimulation. Redistribution is tubulin-mediated and the maintenance of subplasmalemmal mitochondria is dependent upon Ca2+-influx, highlighting a potential role for Ca2+ microdomains in inhibiting tubulin-mediated motility. Hyperstimulation induced actin-mediated redistribution of cellular contents, including mitochondria into the apical pole. Finally we determined the functional importance of mitochondrial redistribution away from the plasma membrane by determining the effects of redistribution on store-operated Ca2+ entry. This investigation strongly advocates the importance of intact tissue to bolster molecular and cellular tools in the study of organ physiology.

TAR DNA Binding protein (TDP-43) is a member of the heterogeneous nuclear ribonucleoprotein family with crucial splicing, transport and regulatory function of genetic material inside mammalian cells. Unfortunately, TDP-43 positive cytoplasmic aggregates occurring with post-translational modifications are a common hallmark in neurodegenerative diseases observed in Alzheimer’s, Parkinson’s, amyotrophic lateral sclerosis (ALS) and fronto-temporal lobar degeneration (FTLD) diseases. Mutations in the TARDBP gene responsible for encoding TDP-43, have been directly correlated with onset of ALS and FLTD. Disease models describing TDP-43 proteinopathy suggests onset may derive through either cytoplasmic mis-localisation or a loss of nuclear function but it is unclear if or how disease associated point-mutations contribute to these observations. In order to determine the effects these mutations have on the protein, a fragment containing the tandem RRM domains (residues 101-265), responsible for the proteins nucleic acid binding function was tested. Using small angle X-ray scattering, circular dichroism, isothermal titration calorimetry and thermal assay methodology it was demonstrated that initial structures of all variants are similar but mutations (D169G and K263E) confer resistance to thermal denaturation by up to 4.9 ± 0.6˚C. This stability positively correlated with an increase in half-life when tested in the full-length variant using a neuron cell model suggesting that protein turn-over is a contributing disease factor. This study was also concerned with solving an X-ray crystallographic DNA binding complex structure for E.coli DPS and mapping interactions with neighbouring DPS complexes. These mechanisms are important in DPS function to protect nucleic acids during prokaryote stress. DPS is conserved in almost all prokaryotes however not all species can interact with DNA. Using X-ray crystallography, a model of E.coli DPS was built to 2.8 Å resolution from DNA containing samples showing both DNA and N-terminal residues were absent. Stabilising polar interactions were shown to form between neighbouring dodecamer structures involving T12, R18, D20, N99, S100, S106 and K134. Polar contacts are observed in all compared crystallographic structures from different species but the residues involved are poorly conserved, despite strong similarities between sequence and structure. This suggests that these contacts may contribute to stabilising the DNA-DPS complexes but form indiscriminately between exposed polar residues available on the dodecamer surface. These interactions are likely to contribute to the thermal stability of DNA-DPS complexes to aid in the proteins protective function.