Signal sequence recognition by SecA, a preprotein translocating molecular machine, is a major and unique cellular function in the bacterial secretory pathway. The SecA ATPase must specifically and efficiently recognize the amino-terminal signal sequences of newly synthesized polypeptides to be able to target them to the cytoplasmic membrane. To elucidate the basis of SecA mediated signal sequence recognition and the details of SecA driven preprotein translocation, biophysical and biochemical characterizations of structure-function relationship and the interactions between SecA and signal sequences have been performed and addressed in this dissertation. Our NMR study of Escherichia coli SecA demonstrates that SecA has two major assigned mobile regions: one is located within the second nucleotide-binding fold (NBF-II; residues 564–579) and the other is the extreme C-terminal segment of SecA (residues 864–901). Both mobile regions are essential for preprotein translocation activity and involved in functionally critical regulatory steps in the mechanism of SecA. Chemical cross-linking studies show that the oligomeric state of SecA in solution is altered by the presence of signal sequences and phospholipids. Addition of functional signal peptides shifts the equilibrium of the oligomeric state of SecA from the major dimeric state to the higher oligomers. In contrast, addition of phospholipids shifts the equilibrium to favor the more monomers. These results suggest that the oligomeric state of SecA may change dynamically during the translocation reaction and imply that bacteria may utilize the cycles of SecA subunit recruitment and dissociation to achieve efficient processive translocation. Lastly, the NMR study of a SecA-bound signal peptide reveals that this peptide forms an α-helical conformation in its hydrophobic h-region and extended conformations in the n- and c-regions when bound to SecA. Line-broadening effects indicate that the positively charged arginine residues of the n-region and the hydrophobic residues of the h-region appear to be more immobilized than the polar residues of the c-region during binding to SecA. TrNOEs suggest that the V15/A18/V21 side of the helical h-region is more strongly bound in the binding pocket. Therefore, the positively charged n-region and the hydrophobic helical h-region could be the selective features needed for signal sequence recognition by SecA.

Conformational and functional changes in trout hemoglobin and cod myosin were investigated at low and high pH and after subsequent refolding. Hemoglobin cooperatively unfolded at low pH to a “molten globular” state with different levels of structure depending on how low the pH was and if 500 mM NaCl was present. At low pH the heme lost its contact with the protein but was not released but the protein was fully dissociated. Alkaline pH had little effect on hemoglobin but the heme environment suggested a strong heme-distal histidine link at pH 10.5. Longer unfolding time, lower pH and higher ionic strength made it more difficult to refold hemoglobin. The more misfolded the protein was the more hydrophobic it became and the lower its solubility after refolding it to pH 5.5. The presence of salt greatly facilitated the hydrophobic aggregation. The more unfolded or misfolded the protein was the more pro-oxidative it became, maybe as a result of more exposed heme, heme and subunit dissociation and hemoglobin-membrane interactions. High pH completely suppressed pro-oxidative activity of hemoglobin, possibly due to a strong heme-distal histidine link preventing autoxidation. Helical structure of myosin rods was little affected at low and high but the head group was substantially unfolded. The protein was possibly fully dissociated at low pH while at high pH only half of the light chains dissociated. The myosin head was misfolded on refolding with increased hydrophobicity, reactive −SH groups, loss in ATPase activity and lower conformational stability. The heavy chains may have fully reassembled on refolding from acid pH but light chains failed to reassemble to the protein. Refolded myosin had almost identical solubility as native myosin at pH 7.5 while cod myofibrillar proteins had increased solubility after acid and alkali treatment compared to the untreated proteins. Myosin and myofibrillar proteins exhibited improved emulsification activity and stability at pH 7.5 after pH treatment. Myosin gelled at a lower temperature and formed stronger gels at pH 7.5 in 600 mM KCl. The pH treated myofibrillar proteins also gelled at a lower temperature than native myosin but had a different gelling mechanism on heating.

Hsp70 molecular chaperones have key functions in the cell for folding, repair and degradation of proteins. Hsp70s are a conserved family of proteins that consist of an N-terminal ATPase domain (NBD) and a C-terminal substrate-binding domain. ATP binding leads to reduced substrate-binding affinity, and reciprocally, substrate binding activates their ATPase rate. This interdomain communication is essential for Hsp70 chaperone function. Previous work suggests a key role for a stretch of conserved hydrophobic residues, 389VLLL 392, in the interdomain linker of E. coli Hsp70, DnaK, in allosteric coupling. Strikingly, we observed that the presence of these residues on the ATPase domain of DnaK (NBD392) led to a markedly enhanced rate of hydrolysis relative to either the intact DnaK protein or a construct lacking this segment (NBD388). Thermal denaturation experiments showed differences in the stability of the two ATPase domain constructs and indicated altered rigid-body movements of subdomains upon linker interactions with the ATPase domain at pH 7.0. We have further illustrated by charge state distributions with ESI-MS (electrospray ionization mass spectrometry) that the linker docking causes conformational rearrangements to a more closed conformation at this pH. We also found that this conformational switch can be reversed by varying the pH. H/D exchange mass spectroscopy of the two ATPase domain constructs revealed different exchange kinetics for the constructs in the nucleotide free-state at pD 7.6. NBD388 displayed a single slowly-exchanging population, whereas NBD392 showed two populations: one which exchanged at a rate similar to NBD388, and one which exchanged rapidly, suggestive of a global unfolding event. The differences in exchange kinetics were reversed by the addition of ADP. These data confirmed our previous findings that NBD392 exists in dynamic equilibrium between an "open", more globally dynamic state and a "closed", less dynamic state, whereas NBD388 samples only the more "open" state. Taking all of our results together, we hypothesize a possible mechanism of allostery: Ligand-induced shift in an open/closed state equilibrium along with dynamic changes in the ATPase domain may be communicated to the substrate-binding domain via the linker residues, implicating the conserved linker as an allosteric switch.

Plasmodium falciparum, the most pestilential of the malaria parasite species, is responsible for ~450,000 direct deaths annually. Clinical disease is a consequence of the blood stage of the parasite’s lifecycle involving a plethora of host-parasite interactions. Key to these interactions are the P. falciparum reticulocyte binding-like homologue (PfRh) proteins responsible for binding erythrocyte receptors and gaining entry to host cells. For example, PfRh4 binds to human complement receptor-1 (CR1) on erythrocytes for sialic-acid-independent invasion. Another protein important for invasion is the PfRh5-interacting protein (PfRipr), an essential member of the PfRh5-associated invasion complex (PAIN-complex) along with CyRPA, the cysteine-rich protective antigen. Loss of function of PfRipr in P. falciparum parasites prevents erythrocyte entry and ablates Ca2+-influx into the erythrocyte; essential events during invasion. This study aimed to biochemically and structurally investigate truncated recombinant versions of PfRh4 and PfRipr. Homology modelling suggested that PfRh4 is rich in alpha-helical secondary structure. The sequence of PfRipr suggested the presence of ten epidermal growth factor-like (EGF) modules, two towards the N-terminus and eight in the C-terminal domain. In this project, monoclonal antibodies made against recombinant PfRh4 were shown, via indirect immunofluorescent assays, to localize to the apical tip of merozoites. Monoclonal antibody 5H12, raised against PfRh4, reduces parasite invasion of erythrocytes by ~75% in growth-inhibition assays with neuraminidase pre-treated erythrocytes. Attempts to produce a stable truncated recombinant PfRh4 protein for structural studies were unsuccessful. An ELISA-based assay using ten alanine-scan mutants suggested the CR1-binding site lies outside of amino acids 283 – 341 of PfRh4. PfRipr truncations, defined by the boundaries of EGF-like repeats predicted based on sequence homology, were produced recombinantly in Escherichia coli and Pichia pastoris. These proteins had a circular dichroism signature suggestive of β-strand-containing proteins with disordered regions. EGF-containing PfRipr truncations did not bind recombinant PfRh5 according to ELISA and size-exclusion chromatography assays. EGFs 1-2, 5-7 and 7-10 of PfRipr did not bind CyRPA via size-exclusion chromatography or NMR. Crystallisation trials performed on EGF modules failed to yield crystals suitable for data collection. A 15N isotopically-labelled sample of EGF5-7 gave good quality HSQC NMR spectra. A suite of three-dimensional NMR spectra collected on a 13C,15N-EGF5-7 sample, at three different temperatures, allowed for >86% of backbone assignments. T1/T2 relaxation analysis and heteronuclear NOE data were suggestive of an elongated, rigid protein undergoing intermolecular self-association. Further evidence for EGF5-7 being an elongated protein was provided via SAXS analysis. Chemical shifts facilitated prediction of secondary structure in EGF 5-7 consistent with an EGF-like fold. Melting studies performed on EGF5-7 showed no evidence of denaturation over the temperature range 20 °C - 95 °C indicating a thermally-stable protein. The addition of Ca2+ to the 15N-EGF5-7 sample caused chemical shift perturbations consistent with high-affinity binding. The discovery of inhibitory monoclonal antibodies recognising a conformational epitope on EGF7 provided evidence of the functional importance of this region within PfRipr. The work described in this thesis provides methods for the industrially-scalable production and biophysical investigations of P. pastoris or E. coli-produced disulfide-rich P. falciparum antigens of interest to vaccinologists.

The adhesion GPCRs are a poorly understood and evolutionarily ancient family of cell surface receptors, several of which have emerging functions in the development of the nervous system. aGPCRs comprise a large extracellular domain, providing binding sites for a variety of ligands, alongside a seven transmembrane domain characteristic of GPCRs. It has been proposed that aGPCRs may function as "context-recognisers", using their large ectodomains to bind different combinations of ligands depending on the molecular make-up of the environment. However there is a lack of direct evidence for this at a molecular level. The ectodomain of one subfamily of aGPCRs, the Latrophilins (Lphns) has been shown to directly interact with several ligands with roles in synaptogenesis and neural guidance. The best-validated of these interactions are those with Fibronectin Leucine-Rich Transmembrane (FLRT) proteins and the Teneurins. In addition, FLRT proteins, also interact with Uncoordinated5 (Unc5) proteins, mediating cell repulsion. Here I reveal that the FLRT-binding site of Lphn is bifunctional, mediating both cell adhesion and repulsion, and that Unc5 is capable of influencing the functional outcome of this interaction. Biophysics and structural studies show that fragments of the Lphn, FLRT and Unc5 ectodomains interact in an unusual and homologue-dependent stoichiometry. Despite the fact that Teneurin interacts with Lphn at a distinct site, Teneurin seems incapable of interacting with the Lphn-FLRT-Unc5 complex, but can form a ternary complex with Lphn and FLRT in the absence of Unc5. Alongside this I present a crystal structure of a large portion of a Teneurin ectodomain, revealing the ancient evolutionary origins of this receptor. Together these data provide strong molecular evidence for a role of Lphns as context-recognisers, by their abilities to bind diverse ligands in distinct combinations and variable stoichiometries.

Neurotensin (NT) is a tridecapeptide neurotransmitter found in the central nervous system and gastrointestinal tract. Neurotensin receptor 1 (NTS1), a high affinity receptor for NT, is a member of the GPCR superfamily and is a putative target for the treatment of conditions such as Schizophrenia, Parkinson’s Disease and drug addiction. Overexpression and purification are typically limiting steps in the high resolution structure determination of GPCRs. In this study, through the optimisation of the E.coli strain used for overexpression of rat NTS1 (NTS1) and the inclusion of phospholipids in the purification buffers to prevent delipidation, an approximate 3-fold improvement in active receptor yield was obtained relative to existing protocols. Preliminary electron microscopy (negative stain and cryo) confirmed a monodisperse receptor population. Purified NTS1 is now being produced at a sufficient level for high resolution structural studies, including 3D crystallisation and further electron microscopy studies. The existing construct for the expression of NTS1 in E.coli, termed NTS1B, was modified to contain a fusion to the genes encoding either the eCFP or eYFP fluorescent proteins. These constructs were used for the E.coli expression of NTS1 tagged with either fluorescent protein at the C-terminus. Tagged receptor was successfully expressed at levels of up to 0.29 ± 0.03 mg per l of culture. Successful purification and proteolytic removal of the MBP and TrxA-His10 fusion partners was achieved whilst retaining both fluorescence and ligand binding capability (K_d = 0.91 ± 0.17 nM). Purified, fluorescent receptor was reconstituted into brain polar lipid (BPL) liposomes in an active conformation which was both fluorescent and able to bind NT. Experimentation with alternative lipid compositions suggested that specific lipids are required in order to maintain ligand-binding activity. FRET between the eCFP- and eYFP-tagged receptors was observed in reconstituted samples. The FRET efficiency was comparable to that observed in vivo for other GPCRs, including the yeast α-factor receptor, which is believed to be dimeric. This suggests that NTS1 could also be multimeric. In contrast, no FRET was observed in detergent samples. Therefore, a functioning in vitro system has been developed which enables the study of NTS1 multimerisation in lipid bilayers and future studies will attempt to implement single molecule fluorescence techniques. In addition, fluorescent derivatives of NT were successfully synthesised and purified. Radioligand competition assays and fluorescence correlation spectroscopy (FCS) confirmed that the fluorescent peptides bound to purified NTS1 in specific competition with unlabelled NT. Surface plasmon resonance (SPR) was used to confirm the ligand binding activity of purified NTS1. A novel approach was utilised which involved the measurement of the binding of detergent-solubilised NTS1 to immobilised, N-terminally biotinylated NT on the sensor surface. The use of a rigorous control, which consisted of immobilised ‘scrambled sequence’ NT, demonstrated a specific interaction. Analysis of the kinetics revealed a multiphasic interaction with a K_d in the nanomolar range. In summary, improvements to the expression and purification of NTS1, the generation of fluorescent constructs as useful tools in the study of receptor multimerisation and the optimisation of lipid-reconstitution protocols have opened up several preliminary lines of study which show considerable potential for future research.

Correct regulation of transcription is essential for maintaining a healthy cellular state. During transcription RNA polymerase II (Pol II) proceeds in a regulated manner through several transitions to ensure appropriate control of synthesis and enable correct processing of the pre-RNA. Shortly after initiation Pol II is caused to pause by the binding of factors, DSIF and NELF. To enable transition of Pol II into the elongation phase CDK9/cyclin T phosphorylates the C-terminal domain (CTD) of Pol II, DSIF and NELF. This phosphorylation releases the paused state and provides an alternative set of post-transcriptional modifications on the CTD to generate a binding platform for elongation, histone modifying and termination factors. CDK9/cyclin T is itself regulated within multicomponent complexes. A small activated complex, containing Brd4, recruits CDK9/cyclin T to active sites of transcription, thereby promoting the elongation of transcription. The role of CDK9/cyclin T in the regulation of transcription has resulted in its validation as a drug target against several disease states including cancer, HIV and cardiac hypertrophy. In this thesis, I present the crystallographic structures of a series of 2-amino-4-heteroaryl-pyrimidine compounds and the roscovitine derivative, (S)-CR8, bound to CDK9/cyclin T and CDK2/cyclin A. In combination with thermal denaturation data and kinetic analysis, these structures have suggested chemical modifications that might be made to increase the CDK9 specificity of these compounds. I have also validated the use of a mutated form of cyclin T for use in the development of CDK9/cyclin T inhibitors. In addition, I present both structural and kinetic analysis of the Brd4-CDK9/cyclin T interaction. I show that C-terminal fragments of Brd4 enhance the in vitro kinase activity of CDK9/cyclin T against the Pol II CTD. Furthermore, I demonstrate that this enhancement may be inhibited by Plk1-mediated phosphorylation of Brd4. Finally, I show that Brd4 binds to a site that spans CDK9 and cyclin T and I propose detailed molecular models of the Brd4-cyclin T interaction.

The bacterial flagellar motor is one of the few rotary motors found in nature, and an excellent example of a complex molecular machine. Flagellar motors in the model organism Escherichia coli are products of the coordinated expression of ∼50 different genes. The E. coli flagellar motor is powered by the proton-motive force (pmf), an electrochemical gradient across the cell membrane. Motor torque is gen- erated by proton flow through membrane-embedded stator units which bind to the basal body of the motor. This thesis aimed to investigate the relationship between the pmf and the flag- ellar motor. A novel pmf control system was developed, based on the light-driven proton pump proteorhodopsin (pR). This system enabled pmf -dependent changes in motor behaviour to be precisely monitored in vivo. Expression of pR in E. coli was shown to be sufficient to drive the flagellar motor at wild-type speeds. Using the pR-based pmf control system, the motor was shown to respond to changes in pmf on a timescale of milliseconds. Surprisingly, motor speed increase was observed when pmf was increased above the physiological norm. Reduction of pmf to low levels enabled individual steps in motor rotation to be observed. Motor response to loss of pmf was investigated. Motors were shown to exhibit a two-stage speed decrease after disruption of pmf , with motor speed falling to ∼20 % of its initial value within milliseconds, reaching a complete stop after 1 s. Extended periods of pmf loss was shown to lead to disengagement of stators from the motor, with motor speed increasing in a stepwise fashion after pmf restoration. The integrity of the motor at different pmf levels was investigated by using TIRF microscopy to directly image positioning of fluorescently tagged motor components. The stator protein MotB was shown to physically leave and rejoin the motor after pmf disruption and restoration, with MotB dispersal following motor stop.

Signal transduction is critical during the lifetime of a neuron as it navigates to reach its targets, forms functional synaptic connections and adjusts the molecular architecture of these connections in an activity-dependent manner. Understanding the molecular organisation of components required for neuronal signalling will provide novel biological insight and can contribute to the design of therapeutics for neurodevelopmental and neurodegenerative disorders. The focus of the thesis is on determining mechanistic molecular details of a number of distinct cell surface systems implicated in neuronal signalling. Crystallographic studies on the cell surface complex between Eph receptor A4 and ephrinA5 contributed to understanding how the modes of higher order arrangements of receptors involved in guidance affect signal transduction across the membrane. A set of structural and biophysical studies addressed the proteoglycan regulation of RPTPσ-TrkCtrans-synaptic interaction and contributed to deciphering the principles of the switch from axonal growth to synapse establishment and formation. A crystallographic and biochemical analysis of the neuronal C1q-like family, enabled mapping their interactions with potential synaptic partners, and guided functional studies aimed at elucidating their roles in the maintenance of synaptic integrity. Preliminary work on the neuronal Sigma-1 receptor chaperone laid the foundations for the structural determination of this receptor.

The molecular mechanism of transmembrane signaling is unknown. Investigations have been hampered by the limits of current biophysical methods. Recently developed solid-state nuclear magnetic resonance (NMR) techniques provide a new approach for selective distance measurements probing structure and function of membrane proteins including ligand interactions. The environmental signal transduced by the serine receptor of bacterial chemotaxis is initiated by the attractant molecule serine. We initially demonstrated the feasibility of direct internuclear distance measurements between the 15N-amino labeled serine ligand and phenylalanine 13C-carbonyl labeled receptor. The two 4.0 ± 0.2 Å distances measured from the serine receptor to the ligand match the distances observed in the crystal structure of the ligand-binding domain fragment of the homologous aspartate receptor. Demonstration of the structural similarity between the aspartate receptor and serine receptor instigated further investigations into the mechanism of ligand specificity. To probe transmembrane signaling we developed a new constant time version of the rotational resonance solid-state NMR technique with improved reliability and efficiency. Combined with a site-directed strategy, this is a valuable and general tool for probing structure in large membrane proteins. A CO( i) to Cβ(i + 3) distance measurement along the periplasmic edge of the second membrane-spanning helix, provides a structural constraint for an unmapped region of the serine receptor. The 5.4 Å internuclear distance measured in the presence and in the absence of serine shows that any signaling mechanism must conserve the helical pitch of the second transmembrane domain. To test abundant and conflicting models of transmembrane signaling we measured an inter-helical distance across the dimer interface in the transmembrane region of the receptor. Computer modeling of this distance predicted sensitivity to proposed long-range ligand-induced conformational changes including piston, scissors, and rotational motions. Ile measured 5.0 Å distance provides a valuable structural constraint of tertiary structure. Both ligand-free and -bound signaling states of the receptor show the same inter-helical distance, suggesting that conformational changes are not propagated into the transmembrane domain. This approach provides a means for testing proposed mechanisms and mapping conformational changes involved in transmembrane signaling.