Fundamental to cell adhesion and migration, integrins are large heterodimeric membrane proteins that link the extracellular matrix to the actin cytoskeleton. Uniquely, these adhesion receptors mediate inside-out signal transduction, whereby extracellular adhesion is activated from within the cell by talin, a large cytoskeletal protein that binds to the cytoplasmic tail of the β integrin subunit via its PTB-like F3 domain. Features of the interface between talin1 and small β3 fragments only have been described previously. Through NMR studies of full-length integrin β tails, we have found that β tails differ widely in their interactions with different talin isoforms. The muscle-specific β1D/talin2 complex exhibited particularly high affinity, leading to the X-ray crystal structure of the β1D tail/talin2 F2-F3 complex. Further NMR and biological experiments demonstrated that integrin activation is induced by a concerted series of interactions between the talin F3 domain and the β tail and between the talin F2 domain and the cell membrane. Additional studies revealed the structural determinants of tight talin2/β1D binding and the basis of more general differences between β1 and β3 talin binding. NMR studies were also performed on tyrosine-phosphorylated integrin tails binding to the PTB domains of talin1 and Dok1, an inhibitor of integrin activation; these revealed that phosphorylation can inhibit integrin activation by increasing the affinity of the β tail for talin competitors. Key residues governing this switch were identified, and proteins were engineered with reversed affinities, offering potentially useful biological tools. Taken together, these results reveal the remarkable complexity of structural features that enable talin and its competitors to mediate this important form of transmembrane signalling.

Epigenetic regulation is a complex process involving the interplay of multiple different cellular factors. Work described in this thesis concerned the characterisation of proteins involved in the binding to, and demethylation of, histone 3 (H3) tails modified by N-methylation. Initial work focussed on the biophysical characterisation of the tandem plant homeodomains (PHD) of the chromatin remodeller CHD4. NMR spectroscopy was used to investigate the solution structure of the tandem PHDs. Studies on a more native-like construct including the C terminal tandem chromodomains are also presented. Binding studies of the PHDs with H3 peptides reveal that the individual PHD fingers can independently bind a histone peptide. The remainder of the work involved characterisation of JmjC histone demethylases (KDMs), enzymes that catalyse removal of Nε-methyl groups from histone lysyl-residues. Initially, two members of the KDM7 subfamily, PHF8 and KIAA1718, were studied; a high throughput screening assay for them was developed, which enabled identification of a selective inhibitor of the KDM2/7 subfamilies of KDMs, the plant growth regulator Daminozide. A disease relevant mutation in PHF8 was studied and shown to cause mis-localisation of the enzyme to the cytoplasm, providing a potential explanation for the clinically observed phenotype. Subsequent chapters describe unprecedented activities for the JmjC KDMs. 2OG oxygenases catalyse a wide range of oxidative reactions, predominantly mediated by initial substrate hydroxylation. The activity of PHF8 with lysine analogous was tested; the results demonstrated that PHF8, and other KDMs, can oxidatively remove Nε-alkyl groups other than methyl groups, such as ethyl and isopropyl groups. The substrate scope of the JmjC KDMs thus has the potential to be wider than previously thought. Observation of β-hydroxylation of the Nε-isopropyl group of a histone peptide including Nε methylisopropyllysine by JMJD2A/E supports the presumed mechanism of histone lysine demethylation as proceeding via initial hydroxylation. This work led to the discovery that JmjC KDMs can catalyse arginine demethylation. This novel arginine demethylase activity by JmjC KDMs was characterised and the work extended to encompass potential arginine demethylase activity in cells. Biochemical characterisation of UTY, a homologue of the H3 K27 demethylases JMJD3 and UTX, which is reported to be inactive, was carried out; UTY was shown to catalyse demethylation at H3 trimethylated at K27 on peptidic substrates, albeit it at substantially lower rates than the other family members. To investigate the reason for this reduced activity, two variants were made, S1142G and P1214I; the latter variant was shown to be considerably more active than wildtype UTY, likely due to an increased peptide-binding interaction. Preliminary experiments in cells did not conclusively demonstrate histone demethylation, but a luciferase assay suggested that UTY may have catalytic activity in cells. Overall the findings in the thesis suggest that the process of cellular epigenetic regulation is likely even more complex than previously thought, with the potential that JmjC KDMs carry out multiple, context dependent functions.

The physical properties of viral capsids are major determinants of vaccine efficacy for several picornaviruses which impact on human and animal health. Current picornavirus vaccines are frequently produced from inactivated virus. Inactivation often reduces the stability of the virus capsid, causing a problem for Foot and Mouth Disease Virus (FMDV) where certain serotypes fall apart into pentameric assemblies below pH 6.5 or at temperatures slightly above 37°C, destroying their effectiveness in eliciting a protective immune response. As a result, vaccines require a cold chain for storage and animals need to be frequently immunised. FMDV is a member of the Aphthovirus genus of the Picornaviridae. Globally there are seven FMDV serotypes: O, A, Asia1, C and SAT-1, -2 and -3, contributing to a dynamic pool of antigenic variation. As part of collaboration between the Division of Structural Biology, Oxford University, The Pirbright Institute, Reading University and ARC, Ondespoort, South Africa we sought to rationally engineer thermo-stable FMDV capsids either as infectious copy virus or recombinant empty capsids with improved thermo-stability for improved vaccines. In this project, in silico molecular dynamics (MD) simulations, molecular modelling, free energy calculations, X-ray crystallography, electron microscopy and various biochemical/biophysical techniques were used to design and help characterise the capsids. For the most unstable FMDV serotypes (O and SAT2), panels of stabilising mutants were characterised by MD. Promising candidates were then engineered and shown to confer increased thermo- and pH-stability. Thus, in silico predictions translate into marked stabilisation of both infectious and recombinant empty viral capsids. A novel in situ method was used to determine crystal structures for quality assessment and to verify that no unanticipated structural changes have occurred as a consequence of the modifications made. The structures of the wildtype and two of the stabilised mutants were solved and the antigenic surfaces shown to be unchanged. Animal trials showed stabilised particles can generate a similar or improved neutralising antibody response compared to the traditional vaccines and may therefore lead to a new generation of stable and safe vaccines.

Fibrillins are modular, disulphide-rich glycoproteins that assemble into microfibrils in the extracellular matrix (ECM). These microfibrils are critical structural elements of many non-elastic and elastic connective tissues. They also regulate the availability of transforming growth factor-β signalling molecules in the ECM. Defects in microfibrils are associated with acquired and inherited connective tissue disorders. In particular, mutations in the human FBN1 gene, which encodes fibrillin-1, are associated with a spectrum of diseases, including Marfan syndrome (MFS). One of the proposed initial steps in microfibril assembly is the interaction between the N- and C-terminal regions of fibrillin monomers. The minimal regions of human fibrillin-1 required for an interaction in vitro were previously identified: the four N-terminal domains, from the fibrillin unique N-terminal (FUN) domain to the third epidermal growth factor-like (EGF) domain (FUN-EGF3), and the three C-terminal calcium-binding EGF-like (cbEGF) domains (cbEGF41-43). Here, fragments corresponding to these regions were produced and shown to interact in pull-down and surface plasmon resonance assays. In addition, the structure of the FUN-EGF3 fragment was determined using nuclear magnetic resonance spectroscopy. This showed the novel structure of the FUN domain and the interdomain interfaces in this region of fibrillin. Combining structural and sequence conservation data may help to identify regions of FUN-EGF3 important for binding to cbEGF41-43. Here, the interaction was probed by site-directed mutagenesis. However, substituting individual residues in FUN-EGF3 with alanine did not abrogate binding to cbEGF41-43. Three MFS-associated residue substitutions were also introduced into the FUN-EGF3 fragment. While they did not abolish the interaction with cbEGF41-43, they did cause misfolding. Two of these substitutions, N57D and W71R, also resulted in the defective secretion of a larger N-terminal fragment by fibroblast cells, suggesting a potential mechanism of disease pathogenesis. Although specific residues involved in the N-C interaction were not identified here, the FUN-EGF3 structure will be vital for understanding the molecular surfaces involved in microfibril assembly and growth factor binding.

The control of the flagellar motor in Rhodobacter sphaeroides was investigated. Unlike most flagellar motors which are controlled by reversing the direction of rotation, the R. sphaeroides motor is controlled via a stop-start mechanism. Advanced optical microscopy was employed alongside genetic, biochemical, and behavioural techniques. High-resolution measurements of rotating beads on flagellar stubs revealed that the R. sphaeroides motor is similar to its E. coli counterpart, rotating counterclockwise at comparable torques/speeds (1,300 pNnm/rad at stall torque), and exhibiting transient step changes in speed. The mean stop duration, mean stop frequency (number of stops per s), and run bias (fraction of time spent rotating) of wild-type at steady-state were 0.66 ± 1.01 s, 0.31 ± 0.19 s-1, and 0.80 ± 0.20, respectively. Manipulating signal inputs to the motor genetically, or by exposing cells to chemotactic stimuli revealed that (i) without chemotactic stimulation the motor rotates continuously, (ii) phosphorylated CheYs are required to stop the motor, and (iii) the chemotaxis system cannot control the speed of rotation of the motor (termed chemokinesis) as previously reported. Complementation studies revealed that CheY3, CheY4, and CheY5 are functionally equivalent. The copy numbers per cell of important CheYs were found to vary greatly under the conditions tested (<1,000, ~3,000, ~60,000 for CheY3, CheY4, and CheY6 respectively). To determine how CheY-P binding causes the motor to stop, external force (viscous flow or optical tweezers) was applied to chemotactically stopped motors. CheY-P binding might either cause the torque-generating units to disengage from the rotor, analogous to a clutch, or trigger the rotor to jam, analogous to a brake. The rotor resisted re-orientation during a chemotactic stop implying that the motor was held in a locked state. The value of torque resisting forward motion (keeping it locked) was estimated to be 2-3 x stall torque (2,500-4,000 pNnm/rad). Furthermore beads attached to flagellar stubs stop at fixed angles for several seconds, showing no large-scale Brownian motion. Step analysis revealed that these stop events occur at 27-28 discrete angles around the motor, which most likely reflect the periodicity of the rotor (i.e. copies of FliG). This represents the first experimental resolution of steps in the rotation of a wild-type bacterial flagellar motor with a full complement of torque-generating units.

The ubiquitin proteasome system targets proteins to the proteasome where they are degraded. Substrate recognition and processing prior to degradation take place at the 19S regulatory particle of the proteasome. A polyubiquitin chain, linked through isopeptide bonds formed between the C-terminal G76 and K48, is the signal responsible for delivery to the proteasome. Because chains linked via any of the seven lysine residues of ubiquitin exist in vivo and encode signals unrelated to protein degradation it is crucial for cells to avoid crosstalk between these different pathways. Several ubiquitin receptors related to proteasomal degradation have been identified but the selectivity between the different ubiquitin chains has not been assessed quantitatively while avoiding artefacts attributed to GST-dimerisation. By employing isothermal titration calorimetry, analytical ultracentrifugation and nuclear magnetic resonance, discrimination between K48- and K63-linked diubiquitin was established for the S. pombe proteasomal receptor Rpn10 and the shuttle protein Rhp23. The same methods allowed us to propose a discriminatory model for Rpn10. The crystal structures of the 19S regulatory particle subunits Rpn101-193 and Rpn121-224 have been determined and possible protein-protein interaction sites were identified by surface conservation and electrostatics analysis. Rpn12 surface residues were identified that had a negative effect on Rpn10-binding. This interaction was studied by surface plasmon resonance, fluorescence anisotropy and nuclear magnetic resonance. These experiments revealed a binding site on Rpn10 that is exclusively occupied by either ubiquitin or Rpn12 and for the first time demonstrated the interaction of a ubiquitin interacting motif with a protein other than ubiquitin.

Neisseria meningitidis is an obligate human pathogen causing over 1000 cases of meningococcal disease within the U.K., 10 % of which result in long-term disability or fatality. 10-70 % of the population carry N. meningitidis in their nasopharynx, the natural reservoir of this bacterium, as a commensal. The host-pathogen interactions of this species are complex and a greater understanding of the molecular mechanisms involved in pathogenesis and immune evasion is required. Three aspects of N. meningitidis pathogenesis were explored in this study. One mechanism of immune evasion which promotes serum resistance of N. meningitidis is recuitment of complement factor H through domains 6 and 7 (fH₆₇) by factor H binding protein (fHbp). In this study, mouse fH₆₇ was recombinantly expressed and purified. fHbp did not bind mouse fH₆₇ at physiologically relevant protein concentrations. The structure of mouse fH₆₇ was solved, showing differences in domain orientation and surface chemistry compared to the human version of this protein, potentially accounting for the host specificity of this interaction. Type IV pili (T4P) are crucial adhesins of N. meningitidis, the fibre of which is composed of thousands of copies of PilE. A method was developed to recombinantly produce large quantities of this protein from a variety of meningococcal strains and the structure was solved of one PilE protein. Subsequent analysis was performed with the PilE proteins investigating their interaction with the putative pilus receptor CD46 and human epithelia as well as their immunogenicity. A method was also established to produce PilC, the proposed tip-assocoated adhesin of T4P. ZapE has recently been identified as an important protein in pathogen colonisation, functioning as an ATPase linked to Z-ring formation in bacterial cell fission. Both N. meningitidis and E. coli ZapE were recombinantly produced. The domain boundaries were mapped and ATPase activity was confirmed. No interaction was seen with FtsZ but DNA binding and modulation was observed by shift assays, the exact function of which remains to be elucidated in future studies.

DNA translocases are molecular motors that use energy from nucleotide triphosphate (NTP) hydrolysis to move along, pump, remodel or clear DNA. Unlike helicases, double-stranded DNA (dsDNA) translocases do not unwind DNA; their action has no net product apart from inducing supercoils as a result of groove-tracking, which has hampered their characterisation. Many dsDNA translocases appear to have biased directionality. However, the inherent symmetry of dsDNA requires that translocase activity is regulated by specific sequences or through modulation by interaction partners. FtsK is a highly conserved bacterial cell-division protein, localised to the dividing septum, that coordinates chromosome segregation with cytokinesis. It is responsible for the resolution of chromosome dimers by activating the tyrosine recombinases XerCD bound to the 28bp chromosomal site dif. The C-terminal domain of FtsK (FtsKC) is a dsDNA translocase (speed ~5 kb/s, stall force ~60 pN) most closely related to superfamily 4 helicases and is active as a hexameric ring. A winged-helix subdomain at the C-terminus of FtsKC, FtsKgamma, binds to specific 8 bp sequences, KOPS, that are polarised in the bacterial chromosome from the origin to towards dif. FtsKgamma also interacts with XerD, activating it for catalysis. Studies of FtsK translocation have differed over whether KOPS act as a loading or a reversal sequence for FtsK. In Chapter 2, I use a continuous ensemble assay for dsDNA translocation to show that FtsK initiates rapidly at KOPS, with loading dependent on FtsKgamma. Translocation requires moderately cooperative ATP binding, while ATP hydrolysis has a more relaxed cooperativity. I have determined the ATP coupling efficiency of translocation to be ~1.6 bp/ATP, in line with theoretical estimates. Though FtsK probably strips most proteins from DNA, I show in Chapter 3 that FtsK stops translocating when it encounters XerCD bound to dif. The interaction is most likely a specific down-regulation, but surprisingly does not depend on FtsKgamma or on the catalytic or synaptic activity of XerCD. In Chapter 4, I show some preliminary structural data of FtsKC bound to dsDNA, with the aim of determining the first high resolution structure of a ring dsDNA translocase bound to nucleic acid.

Malaria is a disease of global importance, causing hundreds of thousand of deaths a year. The majority or deaths are caused by Plasmodium falciparum, a parasite transmitted by the mosquito Anopheles. Its pathogenicity largely results from an ability to transform infected erythrocytes by creating knob-like structures that result in endothelial adhesion. Two major components of these knob structures have been identified as P. falciparum erythrocyte membrane protein 1 (PfEMP1) and knob-associated histidine rich protein (KAHRP). The extracellular fragment of PfEMP1 is responsible for antigenic variability and cytoadherence while its intracellular domain (ATS) connects to the cytoskeleton via interactions with other plasmodium-encoded proteins. In addition, perforin-like proteins (PLPs) with a MACPF domain have been identified in the genome of Plasmodium. PLPs are highly conserved and are expressed in various life-cycle stages of the parasite. They are believed to form pores in membranes of the host cell but their structure is yet unknown. The aim of the work in this thesis was to obtain new information about the structure and role of malaria proteins, thus giving a better understanding of the disease and its possible treatment. Studies of numerous designed constructs of the ATS family were carried out using biophysical methods including high resolution NMR and CD. These revealed that ATS domains are mainly unstructured with a relatively small folded core, consisting of a bundle of &alpha;-helices. Surprisingly, no evidence could be found for ATS binding to KAHRP in solution conditions although previous pull-down data had indicated an interaction. Bioinformatics analysis and yeast-two-hybrid data suggested, however, that there is a conserved protein interaction epitope on the central flexible part of ATS. It was shown, using fluorescence anisotropy measurements, that this part of ATS associates with a parasite protein containing a PHIST (Plasmodium helical interspersed sub-telomeric) domain. Expression constructs of the PLP protein family were designed and manufactured, with the aim of enabling structural studies of this putative pore protein.

Transient receptor potential (TRP) channels are expressed throughout the central nervous system and have a unique ability to detect a wide range of stimuli including changes in voltage, temperature, pH, lipid environment, small molecule agonists, and mechanical stress. While it is known that TRP channels contain the same six transmembrane helix (S1-S6), tetrameric architecture as voltage-gated channels, the degree to which functional and structural analogies are relevant remains poorly understood. This thesis describes a multidisciplinary approach toward understanding the structure and function of TRP channel transmembrane domains by focusing on the S1-S4 transmembrane helices of the TRPV1. This focus is inspired by the voltage-sensor domain (VSD) of the S1-S4 helices of voltage-gated channels, for which a range of studies show functional and structural independence. While some TRP channels are voltage-sensitive, their S4 helix does not contain the positive string of amino acids of canonical VSDs. However, the S1-S4 helices are functionally significant as the binding site of small molecule ligands in both TRPV1 and TRPM8 (for capsaicin and menthol, respectively). The question of TRP channel transmembrane domain modularity is addressed in this thesis by expression and purification trials as well as radioligand-binding assays. It is demonstrated that the S1-S4 and S1-S6 helices of TRPV1 can be properly inserted, overexpressed, and show signs of stability upon detergent-extraction from Saccharomyces cerevisiae membranes. However the TRPV1 S1-S4 and S1-S6 helices do not show wildtype (WT)-like binding in [³H]-RTX binding assays. These results indicate that the TRPV1 transmembrane domains are likely structural but not functional domains. The S. cerevisiae expression system remained promising for the overexpression of TRP transmembrane domains as well as the production of functional, though not stable upon detergent-extraction, WT TRPV1. This WT TRPV1 was subsequently found to functionally bind both RTX, used in ligand binding assays, as well as the double-knot toxin (DkTx), targeted to the pore domain (the S5-S6 helices). An effect of DkTx on RTX binding affinity demonstrates an allosteric interaction indicative of a possible tighter packing between the two transmembrane domains than is seen in voltage-gated channels containing the canonical VSD. Computational approaches additionally allowed for the investigation of the intramembrane capsaicin binding site in the TRPV1 S1-S4 helices, crucial to the initial motivations of this study. While the literature locates the capsaicin binding site to the TRPV1 S1-S4 helices, a `binding pocket' has yet to be defined, with regards to the orientation of bound capsaicin and its access route to the site via the bilayer. Using molecular dynamics (MD) simulations the preferred location of capsaicin within the bilayer is defined, as well as the elucidiation of capsaicin flip-flop between bilayer leaflets as a key event prior to TRPV1 binding. A transient binding was also observed between a homology model of the TRPV1 S1-S4 helices and capsaicin, possibly encouraging the idea that the S1-S4 helices still contain a partial binding site, though of too low affinity to be observed in the binding experiments performed here.