As understanding of the ways in which the complex cellular environment affects the in vivo folding of proteins improves, improved methods for their study are required. It is possible to produce limited quantities of ribosome-nascent chain complexes (RNCs) and techniques for gathering data about them are improving, but no single technique provides all the information required to understand folding of nascent proteins on the ribosome and there are still significant data that cannot be obtained experimentally. In particular, while NMR chemical shift and residual dipolar couplings may be recorded, the samples are of too low concentration and stability to conduct the most informative NOESY experiments that are traditionally used for revealing atomic-resolution structure. Recently, the ability to use chemical shifts to reveal structural details and dynamic properties of small proteins has been developed. By simulating multiple molecules and predicting the average chemical shift of the ensemble, the simulation may be restrained to conform to the experimentally measured data, making testable predictions about the atomic-resolution dynamic properties of the molecule. By adapting these methods to the macromolecular RNC structures it is theorized that the limited chemical shift data available may be used to provide structural details of the protein as it emerges from a ribosome. This, however, is faced by many challenges, including the ability to simulate such large number of atoms in a suitable timescale and applying the restraints to the nascent chain alone. The thesis presented describes the development of computational techniques to characterize RNCs, including the concepts and challenges faced, the chemical-shift restrained simulation of nascent chains alone, the development of techniques to perform chemical-shift restrained molecular dynamics simulations of the RNCs and the application of these techniques to a model system.

Atypical Protein Kinase C (aPKC) isoforms are essential regulators of polarised cell behaviour and in migrating NRK cells translocate to the leading edge in a complex with the exocyst and KIBRA. Engineered delivery of upstream signals to the plasma membrane places leading edge ERK activation downstream of aPKC and demonstrates partial sufficiency in regulating cell migration and adhesion. This model system provides the opportunity to probe the leading edge to better understand events downstream of aPKC. Multiple screening approaches have identified cytoskeletal and translation processes as putative targets of this pathway. Based on in silico candidate screening it is shown that multi-site phosphorylation of Parvin alpha is important for focal adhesion maturation. These phosphorylation events are triggered following acute focal adhesion turnover, which can be blocked by aPKC and MEK inhibition. Based upon proteomic approaches, a novel role for the putative aPKC/ERK substrate Cdc42 effector protein 1 (Cdc42ep1) has been identified. siRNA knockdown of Cdc42ep1 phenocopies aPKC loss; focal adhesions enlarge and turnover less efficiently. This impacts on polarized cell motility as knockdown prevents cell orientation and efficient wound closure. Finally, a novel role for aPKC is reported in relation to leading edge translation. Active translation at the leading edge is reduced following aPKC and MEK inhibition and compartmentalised distribution of translation factors is modulated following pathway intervention. This includes the eukaryotic translation initiation factor 3A (eIF3A), one hit identified by proteomic screening. eIF3A interacts with the exocyst and localises to the leading edge in an aPKC-dependent fashion. In addition, eIF3A is shown to regulate polarised migration and adhesion maturation. The data presented in this thesis illustrate combined screening and validation to delineate compartmentalised signalling events. Localised aPKC/exocyst/ERK activity is necessary for cytoskeletal controls and the polarized delivery and activation of translation machinery at the leading edge.

Pharmaceuticals can interact with sunlight to cause skin photosensitization and increase skin cancer risk. Interaction of drug molecules with solar UVA or visible radiation results in electronically excited states that damage biomolecules directly or indirectly via the formation of reactive species (RS). RS cause damage to DNA and its precursors, as well as to proteins and lipids. I have devised methods to examine the induction of oxidative protein damage in cultured human cells and used these to investigate the effects of UVA-activated photosensitizing drugs on the formation of protein carbonyls and the oxidation of protein thiol groups. I examined the effects of 6-thioguanine (6-TG) (a surrogate for azathioprine, an immunosuppressant), fluoroquinolone antibiotics, and the malignant melanoma therapeutic vemurafenib, each of which is associated with clinical skin photosensitivity and increased skin cancer risk in patients. All of these drugs are shown to be synergistically cytotoxic with UVA in cultured human cells and toxicity is concurrent with the generation of RS. I identify singlet oxygen as a major component of these photochemically-generated RS and show that widespread protein oxidation is caused. The Ku DNA repair heterodimer is identified as one of several targets for oxidation damage and I show using biochemical assays that damage to Ku compromises its function in the repair of DNA strand breaks. UVA irradiation of cells treated with the photosensitisers significantly compromises the removal of potentially mutagenic DNA lesions by the nucleotide excision repair pathway. Since this DNA repair pathway removes sunlight-induced DNA lesions and is the major protection against skin cancer, my findings have implications for the increased skin cancer risk associated with azathioprine. The ability of structurally dissimilar drugs to recapitulate the effects of 6-TG suggests that the observations may share a common mechanism.

Spt4/5 and its bacterial homologue NusG are the only known universally conserved RNAP- associated transcription elongation factors. In the hyperthermophilic archaeon Methanocaldococcus jannaschii, Spt5 comprises an N-terminal NGN domain and a C-terminal KOW domain, and is bound at its NGN domain by Spt4. NusG and Spt5 increase the processivity of RNAP by binding to the RNAP clamp via the NGN domain. This maintains the RNAP clamp in a closed conformation, thereby enabling RNAP to remain bound to the template DNA. The NusG KOW domain interacts with ribosomes, thereby coupling transcription to translation. The functions of Spt4/5 in archaea are less well characterised. The work contained within this thesis demonstrates that in the context of M. jannaschii cell extract, Spt4/5 is found in the same fractions as ribosomes and RNAP, and therefore has the potential to couple transcription and translation. Furthermore, data obtained by microscale thermophoresis suggests that the KOW domain of Spt5 interacts with purified ribosomes. Electron paramagnetic resonance was performed on Spt4/5, demonstrating that Spt5 is conformationally flexible, and that the presence of Spt4 restricts its mobility. Limited proteolysis and thermofluor assays support the notion that Spt4 stabilises the Spt5 NGN domain. In E. coli, NusA binds to RNAP as a component of the antitermination complex, along with NusG, NusB, and NusE. This enables RNAP to enter a pause and termination-resistant state. M. jannaschii NusA consists of two KH domains. Mutational analysis identified the contribution of the two KH domains to RNA binding and identified additional residues involved in the interaction. Archaeal NusA does not coelute with RNAP, raising the possibility that archaeal NusA does not have antitermination functions. In summary this thesis argues that Spt4/5 likely couples transcription and translation in archaea and indicates that archaeal NusA binds to RNA via a novel binding site.

For successful gene expression in eukaryotes, mRNA transcripts are processed in the 5’ and 3’ untranslated regions (UTRs) and non-coding mRNA is spliced out. These processes are crucial in determining the fate of the mRNA transcript. Rna14 and Rna15 are subunits of Cleavage factor 1A (CF1A) and required for 3’ end processing in S. cerevisiae. Structural and biophysical data have determined a number of residues within the RNA recognition motif (RRM) of Rna15 that interact directly with RNA. However, although the crystal structure of the RRM revealed a mechanism for the preferential recognition of G/U nucleotides by Rna15 the sequence specificity for Rna15 is still extensively debated. This thesis applies combination of in vivo and in vitro techniques aimed to characterise Rna15-RNA binding, Rna15 interaction with Rna14 and examine consequences for RNA processing and yeast viability in vivo. An in vitro mutational/biophysical analysis is presented that reveals the residues essential for the Rna15-RNA interaction and application of NMR-quantified Scaffold Independent Analysis (SIA) demonstrates a clear GU-bias in the in vitro consensus sequence. However, given these strong effects, surprisingly only extensive mutation of the RRM produces growth defects in S. cerevisiae and qRT-PCR experiments employing a small subset of genes show only slight effects on polyA site selection. By contrast, an RNA-sequencing (RNA-Seq) global analysis of expression and transcriptional readthrough reveals that expression of over 100 S. cerevisiae genes is severely affected when the RRM of Rna15 is deleted and in addition the 3’-UTR of the mRNA of a sample set of 40 genes is significantly different to wild type. These results indicate that only severe reduction of Rna15-RNA interactions result in defects in transcriptional and 3’ end processing, hypothesized to be due in part to functional redundancy. Nevertheless, the global changes observed upon deletion of the Rna15 RRM are striking and reinforce the link between 3’-end processing, transcriptional regulation and gene expression.

A new method for deconvolution of electrospray ionisation mass spectrometry (ESI-MS) spectra was produced, allowing for the masses of overlapping charge state series to be correctly indentified. The algorithm also determines the abundance of individual molecular species with a much higher accuracy for congested spectra. Several new methods for representing TWIM-MS data were developed. The combination of the deconvolution algorithm with travelling wave ion mobility data creates plots with collision cross section (CCS) axes which can be directly compared with X-ray crystallography structures and computational models. Difference plots have allowed multidimensional analysis of changes in condition, and spectral averaging can produce a single representative spectrum from multiple replicates. Gas-phase unfolding experiments using TWIM-MS are a popular method for probing protein stability in response to conditions such as ligand binding. The algorithms for processing these data are however in their infancy. This thesis describes the first deconvolution algorithm for gas-phase unfolding data, allowing for the accurate interpretation of conformation cross sections and abundances during the unfolding procedure. The methodologies developed were then applied to 1-antitrypsin, a metastable, aggregation prone protein. The protein was bound to a ligand, Ac- TTAI-NH2, which has been shown to block aggregation as a titration and the MS deconvolution method was used to quantify the abundances of each bound state in each mass spectrum. The first use of IM-MS to analyse ex vivo aggregates are shown, and the ion mobility methods created were used to determine the CCS values of the monomeric and dimeric species. The interaction between 1-antitrypsin and Ac-TTAI-NH2 was probed using gas-phase unfolding experiments, determining that the ligand stabilises the protein, with a specific pattern of gas-phase unfolding observed for each state.

The adult mammalian brain is a highly plastic structure capable of cellular and molecular remodelling in response to its interactions with the outside world. The addition of new neurons to the hippocampus throughout life is one of the most striking manifestations of this plasticity. New neurons here are generated from a population of stem cells that, although existing primarily in a dormant or quiescent state, they can become activated upon the reception of neurogenic signals. How stem cells integrate these signals from the environment to ultimately control neuronal production is currently under investigation. During embryonic development, transcription factors of the basic helix-loop-helix family promote progenitor proliferation and differentiation to ensure the production of neurons in correct numbers and at the correct positions. We found Ascl1, a proneural factor in this family, to be expressed by stem cells of the adult hippocampus when in an active state. Here we used pharmacological and genetic approaches to show that Ascl1 expression is rapidly induced in response to neurogenic stimuli, and that deletion of this factor with a conditional inactivation approach results in an inability of stem cells to respond to signals and exit their quiescent state. Moreover, by examining the genes deregulated in Ascl1-deleted stem cells, we show that Ascl1 promotes the proliferation of hippocampal stem cells by directly regulating cell cycle regulatory genes, among which the cyclin D genes are of great importance. The data presented here supports a model whereby Ascl1 acts as a central factor in adult hippocampal stem cells to integrate both stimulatory and inhibitory signals and translate them into a transcriptional programme that controls stem cell activity. With this work we also highlight that understanding how Ascl1 is regulated will contribute, in the future, to the development of stem cell therapies for the treatment of neurological disorders.

Transcriptional enhancers and their associated transcription factors (TFs) have a profound impact in gene regulation during development. The TFs Ascl1 and the Sox factors play important roles in many aspects of neurogenesis in vertebrates. Ascl1 induces neuronal differentiation in neural progenitor cells. Sox2 is an essential TF in the maintenance of neural stem cell characteristics, but also functions in the correct specification of neuronal subtypes. SoxC factors have a fundamental role in the establishment of neuronal traits in differentiating neurons. Data previously generated in our lab have identified genomic regions bound by both Ascl1 and Sox2 in close proximity through a ChIP-seq approach, raising questions about possible interactions between these TFs as transcriptional regulators. Here, I have characterised these regions as neural enhancers regulated by Ascl1, Sox2, and SoxC factors using NS5 cells as an in vitro model of neural stem cells through luciferase assays. Enhancers activated by overexpression of Ascl1 were classified as neuronal enhancers, since overexpression of this TF induces neuronal differentiation in NS5 cells. Cotransfection experiments and luciferase assays demonstrated that Ascl1 and Sox2 counteract each other in the regulation of the enhancers, whereas Ascl1 and SoxC factors synergistically activate the neuronal enhancers. Mutations of the binding motifs of these TFs and comparisons between the transcriptional activity of the wt and mutant enhancers suggest direct binding of Ascl1 on all the enhancers activated by this TF, but different mechanisms of regulation exist for the Sox factors on the different enhancers analysed. I have shown direct binding of Ascl1 on one of these enhancers, MSB4, by EMSA. In conclusion, I have identified a novel transcriptional regulatory network in the regulation of neuronal differentiation where Ascl1 and SoxC factors synergistically activate neuronal enhancers, and Sox2 counteracts Ascl1 in this activation.

There is a need for new and innovative biosensing technologies. Biosensors are extremely useful in: relieving the diagnostics burden of healthcare professionals, environmental safety screening, drug and food testing, etc. Here, we have utilized the inherent advantages of continuous wave electron spin resonance (CW-ESR) over other biosensing platforms to develop ESR biosensing metrologies. Initially, we extrapolate the work of the ESR biosensing company Syva, established in the 1970’s. They previously performed various small molecule competition immunoassays, through spin labeling of small antigens e.g. morphine, but not large biomolecules. We attempted to strategically develop an assay for large biomolecule sensing - using similar assay principles to Syva enabled by epitope tagging. However, competition immunoassays suffer from a range of pitfalls, so a CWESR direct immunoassay was developed for the first time. Consequently, a proof-of-concept ESR direct immunoassay for a cystine stabilized single chain variable fragment (sscFv) for the carcinoembryonic antigen (CEA) was developed using disulfide bond labeling. The direct immunoassay was performed successfully in human plasma and whole human blood, both without the need for any sample pretreatment. Furthermore, using the direct immunoassay method, a thermal stability shift assay (TSSA) was developed. To truly evaluate the potential of the direct immunoassay, the disulfide bridging nitroxide spin labeling methodology was utilized to modify full-length antibodies. An anti-FLAG antibody, two human IgG1 antibodies varying in light chains only and anti-C-reactive protein (CRP) antibodies were modified and show interesting binding effects. Finally, a competition glucose sensing assay utilizing the displacement of concanavalin-A bound nitroxide spin labeled mannan complexes by glucose and a direct sensing assay utilizing chromium ions were developed. Overall, the biosensing capabilities of ESR are explored, and the respective advantages and disadvantages are discussed. Furthermore, the commercialization potential of the work is considered and critically evaluated.

Cyanobacteria hold significant potential as platforms for the production of a wide variety of high-value products and biofuel molecules such as biohydrogen, isoprenoids and alkanes. Currently, the genome sequences of over 120 cyanobacterial species are publicly available, and techniques for the genetic manipulation of a few species are well established. However, more advanced metabolic engineering technologies are required for high-throughput production/evaluation of modified strains with improved biofuel characteristics. The most popular species for genetic studies is the freshwater species, Synechocystis sp. PCC 6803 – not least because it is naturally transformable and is a facultative phototroph. We demonstrated in chapter 3, that a deletion of an exonuclease gene in Synechocystis greatly increases transformation rates, probably through the promotion of non-HR events. However the mutant generated did not display random insertional mutagenesis but rather foreign DNA was being targeted to specific locations in the genome. On the other hand, targeted gene knock-outs and knock-ins are easily achieved in Synechocystis via homologous recombination (HR), but the creation of each transforming plasmid is rather laborious and involves several PCR and cloning steps. Chapter 4 of this thesis describes detailed study conducted to determine the minimum length of homologous DNA sequence flanking the foreign DNA that is necessary for efficient integration into the cyanobacterial genome. Our findings suggest that as little as 50 bases is sufficient for HR and open up the possibility of a quick, single step PCR strategy using long-tail primers to achieve any desirable knock-out or knock-in. However, the efficiency of the PCR needs to be optimised in order to make the technique more efficient. Chapter 5 describes genetic modification of the isoprenoid pathway for production of the novel high-value product geraniol (C10 monoterpene) and the fuel molecule farnesene (C15 sesquiterpene). Finally, chapter 6 exploits the possibility of expressing a biological plug-in for facilitated subtract delivery and product removal of hydrocarbons in Synechocystis.