Search

Processing bodies (p-bodies) are cytoplasmic messenger ribonucleoprotein granules containing components of the mRNA degradation machinery, that form during stress conditions in the budding yeast Saccharomyces cerevisiae. P-bodies are conserved in eukaryotes and are related to other mRNP granules such as stress granules and neuronal granules. mRNP granules and their components self-assemble through a process of liquid-liquid phase separation, facilitated by protein-protein interaction by low complexity regions. Errors in this assembly process can cause a build-up of aggregated protein and have been implicated in the pathology of human neurodegenerative diseases such as amyotrophic lateral sclerosis. The exact role of p-bodies in normal cellular function is not known, but p-bodies are induced in conditions of stress, and can store mRNA that later re-enters translation, leading to the theory that they are sites of mRNA storage. A contrasting theory suggests them as sites of mRNA decay, due to the presence of deadenylation, decapping and exonuclease complexes within p-bodies. In this work, time-lapse fluorescence microscopy, using optimised high brightness, low photobleaching fluorescent protein fusions, was used to allow long term imaging of p-body localisation throughout the cell cycle of S. cerevisiae. Imaging was combined with microfluidic dissection of mother and daughter cells, to track single mother cells over their entire replicative lifespan (RLS) while continuing to monitor p body localisation. P-bodies were found to be localised in proximity to multiple different organelles with changes to localisation occurring through the cell cycle. Advanced replicative aged altered the profile of p-body localisation causing p-bodies to localise heavily to the mitochondria. P-bodies were also found to be inherited by daughter cells in a process dependent on the mRNA transport machinery and the p-body protein Dcp1. This inheritance was multigenerational with a single p-body moving between multiple daughter cells and was not lost in cells of advanced replicative age.

In many eukaryotes cytokinesis is facilitated by the contraction of an actomyosin ring. The exact mechanisms that lead to this contractility are unknown, although a number of models posit that actin turnover in the ring is essential. The effect of reduced actin dynamics during ring formation has been well studied in Schizosaccharomyces pombe; however the corresponding effects on ring contraction are not well understood. By using mutants of the fission yeast actin severing protein Adf1, we observed that contracting actomyosin rings display a 'peeling' phenotype. In these cells bundles of actin and myosin peel off from one side of the ring, and are then pulled across to the opposite side, which typically occurs 3 times during cytokinesis. We also found that this phenotype is dependant on the activity of the three myosin species present in the ring. We hypothesise that reduced actin turnover leads to a non-­‐ uniform distribution of tension around the ring, which causes regions of higher tension to peel off. This model predicts that adf1 mutant cells might be sensitive to additional mutations that compromise membrane anchoring of the ring, and that the tension imbalance could lead to non-­‐ uniform septum ingression during cytokinesis. Subsequent experiments confirmed these predictions, which further supported our model. We also attempted to recreate the phenotype in silico, by adapting a previously used mathematical model of the actomyosin ring. However, in doing so we encountered a number of problems, as we realised that certain parts of the model were unrealistic, and attempting to fix these aspects introduced new problems. As a result of this, we were unable to achieve this goal in the available timeframe. Finally, we also used mathematical modelling to analyse FRAP experiments, to more accurately extract information on the mobile fractions of ring proteins from FRAP recovery curves. Application of this method enabled the identification of interesting behaviour from the membrane anchoring protein Cdc15, potentially linking its mobility to its function in the ring.

Linear dichroism (LD) spectroscopy is a technique which uses polarised light to make inferences on the relative orientations of chemical groups within a molecule. A requirement for experiments using this technique is the alignment of molecules in the analyte. This thesis considers the use of fluid flows to align molecules. The use of bead-spring and bead-rod models to simulate the behaviour of polymers and particles in fluid flows has been extensively discussed in the literature. In this work, the FENE-Fraenkel spring force law is implemented in bead-spring chain simulations, to assess the alignment behaviour of molecules in flow. Analytical properties of FENE-Fraenkel springs are derived and used to inform the coding of the simulation programme and to assess its outputs. The results of these simulations were comparable to both analytical results in zero-shear conditions and to data previously published by Hsieh et al. [48] on bead-spring chains in extensional flow, albeit with a reduced signal-to-noise ratio. Simulations of the alignment of FENE-Fraenkel spring dumbbells in shear flows show increased alignment with high shear rates, similar to modelling data published by McLachlan et al. [89]. This work also considers the use of microfluidic devices manufactured from polydimethylsiloxane to perform linear dichroism studies on DNA and vesicles in solution. A wide channel and cross-slot device are developed to provide conditions for pressure-driven and extensional flow to align molecules. Both devices are shown to align DNA molecules to measure an LD signal, with the cross-slot performing better than the wide channel device. With solutions of vesicles containing 1,6-Diphenylhexa-1,3,5-triene, a membrane probe, the wide channel device is shown to provide weak vesicle alignment to measure an LD signal. However, both microfluidic devices are unable to align molecules as well as Couette flow, the method normally used to align molecules for LD experiments.

Arsenic exhibits diverse chemical reactivity depending upon its oxidation state. This distinctive reactivity has been largely overlooked in the field of polymer and biomaterials science, owing to concerns about the toxicity of arsenic. However, a recent clinical renaissance in the use of arsenicals suggests the possibility of broader acceptance and application. The aim of this work is to stimulate interest in and highlight the potential of polymeric arsenicals as a novel platform for functional and responsive biomaterials (Literature review discussed in Chapter 1). Cross-linking of polymers through pendent organic arsenic functional groups is demonstrated using three different chemistries: 1) reductive coupling forming arsine oligomers, As(I)n (Chapter 2) 2) forming arsenic-thiolate bonds with poly-thiol cross-linkers (Chapter 3) 3) polymerizing As(I) via the addition of acetylenes to form vinylene-acetylene bonds (Chapter 4). All three methods of cross-linking were able to cross-link thermally self-assembled NIPAm-PEG diblock copolymers with an arsenical acrylamide (AsAm) monomer incorporated in the NIPAm core. The first two were found to be responsive towards GSH and H2O2 under model physiological conditions. The stability of the particles can be finely tuned through varying the amounts of arsenic or varying the nature of the thiol-functional external cross-linker. The last form of cross-linking was found to be non-responsive towards the given stimuli however it enabled incorporation of further functionality to the nanoparticle, such as Rhodamine B, which helped determine the co-localization of the nanoparticle. Notably, in all three chemistries, the nanoparticles produced were not found to be toxic at 2 mg/ml by cell viability assays. Finally using the As(I)n cross-linking strategy, the formation of hydrogels (with DMA-AsAm copolymers made by FRP), was demonstrated which were responsive towards oxidation with H2O2. Synthesised hydrogels were responsive towards oxidation. 3D cell culturing of these gels were carried out by rehydration of dialyzed/lyophilized gels with trypsinized solution of cells. (Chapter 5). In addition to the investigation of polymeric arsenicals, synthesis of microscale multiblock copolymers are described (Appendix A). This enabled generation and characterisation of polymers at 2 μL scale, and procedure to synthesise Multi-block copolymers at 10 μL scale (DPn = 25, 5 blocks, 2 μL per block).

A working internal circadian clock allows a healthy organism to keep time in order to anticipate transitions between night and day, allowing the temporal optimisation and control of internal processes. The internal circadian clock is regulated by a set of core genes that form a tightly coupled oscillator system. These oscillators are autonomous and robust to noise, but can be slowly reset by external signals that are processed by the master clock in the brain. In this thesis we explore the robustness of a tightly coupled oscillator model of the circadian clock, and show that its deterministic and stochastic forms are both significantly robust to noise. Using a simple linear algebra approach to rhythmicity detection, we show that a small set of circadian clock genes are rhythmic and synchronised in mouse tissues, and rhythmic and synchronised in a group of human individuals. These sets of tightly regulated, robust oscillators, are genes that we use to de ne the expected behaviour of a healthy circadian clock. We use these “time fingerprints" to design a model, dubbed “Time-Teller", that can be used to tell the time from single time point samples of mouse or human transcriptome. The dysfunction of the molecular circadian clock is implicated in several major diseases and there is significant evidence that disrupted circadian rhythm is a hallmark of many cancers. Convincing results showing the dysfunction of the circadian clock in solid tumours is lacking due to the difficulties of studying circadian rhythms in tumours within living mammals. Instead of developing biological assays to study this, we take advantage of the design of Time-Teller, using its underlying features to build a metric, Θ, that indicates dysfunction of the circadian clock. We use Time-Teller to explore the clock function of samples from existing, publicly available tumour transcriptome data. Although multiple algorithms have been published with the aims of “time-telling" using transcriptome data, none of them have been reported to be able to tell the times of single samples, or provide metrics of clock dysfunction in single samples. Time-Teller is presented in this thesis as an algorithm that both tells the time of a single time-point sample, and provides a measure of clock function for that sample. In a case study, we use the clock function metric, , as a retrospective prognostic marker for breast cancer using data from a completed clinical trial. Θ is shown to correlate with many prognostic markers of breast cancer, and we show how could also be a predictive marker for treatment efficacy and patient survival.

Cellular blebs are fast, pressure-driven protrusions of the cell membrane that are initially devoid of F-actin. Although blebs have often been overlooked as a functional part of cell motility, blebbing has been shown to play an important role in migration in 3D mechanically resistive environments, such as movement through densely packed tissues. The location of bleb nucleation sites is often assumed to be entirely stochastic, however, cells migrating using blebbing motility have been repeatedly observed to perform persistent, directional movement. Given the compelling evidence on the role of blebbing in directional cell migration, relatively little is known on the mechanisms of bleb site selection; how bleb sites are determined and directed to the cell front remains an open question. Previously, Tyson et al. found that chemotaxing Dictyostelium cells preferentially bleb from concave regions, where membrane tension could facilitate membrane-cortex detachment. Based on this, a biophysical model for curvature dependent bleb nucleation was proposed, hinting at the possibility that polarised blebbing was due to physical forces alone. We develop a novel image based modeling approach, using real cell contours from image data to initialise the model. This enables us to link quantitative experimental data and predictive modeling on the spatial distribution of blebs for the first time. We proceed to show that the extent to which cell geometry is a good predictor of bleb site selection is highly dependent on the degree of mechanical resistance the cells experience. For cells in highly resistive environments, where we observe the front and rear of the cell to be geometrically distinct, our novel modeling approach demonstrates that physical forces are sufficient to polarise blebbing activity in chemotaxing cells. For cells in low resistive environments however, where the front and rear of the cell are not geometrically distinct, we show that an additional mechanism is required to restrict blebbing to the cell front. We propose this additional mechanism to be a front-to-rear gradient in the linker protein TalinA. Creation of a TalinA-mNeon construct allows us to experimentally confirm the existence of an asymmetric linker distribution. We incorporate the observed exponential linker gradient within the model. Inclusion of this mechanism increases the predictive power of the model, with regard to the spatial distribution of experimentally observed bleb sites, through efficiently directing blebbing activity to the cell front. We have created a method which links quantitative experimental data and predictive modeling, this tool allows us to disentangle the role of physical forces and biological mechanisms in the prediction of bleb nucleation sites.

Scanning ion conductance microscopy (SICM) is a nanopipette-based technique that has historically been used for the topographical imaging of soft samples. This thesis demonstrates the development of SICM into a multifunctional tool, capable of providing a host of additional information about both biological and inert samples, whilst maintaining the structural mapping capability for which it is usually employed. Two approaches are taken to extend the functionality of SICM: (i) designing sophisticated potential, and positional, control functions that are then used with traditional single-channel nanopipettes; and (ii) incorporating an ion conductance channel into a multi-barrelled probe. In the single-channel setup, a pulsed-potential profile allows the extraction of surface charge density on extended substrates, and a ramped-potential profile permits spatially resolved mapping of redox reactions on an electrode substrate. When integrated into a more complex probe, SICM is used to study molecular uptake at cellular surfaces, and to print Cu microstructures on a Au substrate. While this thesis is primarily concerned with technique development, the studies herein have broad applications in cell biology, pharmaceuticals, materials science and beyond. In addition to developing imaging modes that allow the extraction of functional information at a surface, this thesis also contributes to the fundamental understanding of the SICM system. Finite element method simulations are performed alongside experimental studies, in order to fully understand the contributions of the pipette geometry, ion current rectification, and pipette-surface interactions on the measured ionic current. The theoretical treatment herein provides a foundation upon which future multifunctional SICM regimes could be designed, extending the scope of this increasingly powerful technique.

This thesis presents the development and analysis of a macro-chemical kinetics model of microbial metabolism and growth, belonging to the class of Variable-Internal-Stores (VIS) models. Such models can account for changes in the internal state of a micro-organism in response to variations in environmental conditions. In addition to VIS, the model developed in the present thesis describes the adaptive allocation of molecular building blocks among various types of catalytic machinery by means of so-called ‘regulatory rules’ which can be reconstructed from observational data, as shown in Chapter 4. In Chapter 2 I show that this novel VIS-with-reallocation model has a unique, stable equilibrium point and reduces to classical models of microbial growth and metabolism. In Chapter 3 this mathematical model is treated as a piecewise smooth system, which allows me to investigate the behaviour of the model under conditions of starvation, and, in particular, to describe this behaviour in terms of Fillipov sliding modes. In Chapter 4, I study the response of the model to time-varying environmental conditions of the ‘feast-or-famine’ type of nutrient limitation, where I address the question of maximally adaptive regulation of reserve densities and show that optimal reserve management is governed by the frequency of environmental fluctuations on time scale comparable to that of the organism’s physiology.

Personalised medicine for cancer treatment promises benefits for patient survival and effective use of medical resources. This goal requires the development of predictive models for the identification and implementation of biomarkers for the prediction of patient survival given treatment options. This thesis addresses research questions in this area. The systematic review detailed in Chapter 2 investigates the literature concerning the prediction of resistance to chemotherapy for ovarian cancer patients using statistical methods and gene expression measurements. The range of models used by studies in the systematic review highlights the popularity of traditional models, such as Cox proportional hazards, with few more complex models being utilised. In Chapters 3 and 4, new methods are presented for modelling right-censored survival data. Due to the nature of biomedical data, the methods used need to be flexible and adequately account for high dimensional, noisy data. Gaussian processes fulfil these requirements and were hence used for the development of three Gaussian process models for right-censored survival data. Chapter 3 details these models, and they are applied to synthetic and cancer data in Chapter 4. In all cases the Gaussian processes for survival were found to equal or outperform all comparison models, as measured by concordance index. Given the application to molecular cancer data, it was expected that the data would be high dimensional. Two feature selection methods are investigated in Chapter 5 for use with Gaussian processes to address this. In Chapter 6 a program is developed for the analysis of the data produced by a test for cancer mutations using qPCR. The automated program was designed to isolate the analysis from the user and produce results and reports for clinical use. It is observed that this approach of automated analysis would be suitable for application to any form of clinical test or complex predictive model without the requirement of user guidance.

Functional magnetic resonance imaging (fMRI) is a non-invasive method for studying the human brain that is now widely used to study functional connectivity. Functional connectivity concerns how brain regions interact and how these interactions change over time, between subjects and in different experimental contexts and can provide deep insights into the underlying brain function. Multiregression Dynamic Models (MDMs) are dynamic Bayesian networks that describe contemporaneous, causal relationships between time series. They may therefore be applied to fMRI data to infer functional brain networks. This work focuses on the MDM Directed Graph Model (MDM-DGM) search algorithm for network discovery. The Log Predictive Likelihood (model evidence) factors by subject and by node, allowing a fast, parallelised model search. The estimated networks are directed and may contain the bidirectional edges and cycles that may be thought of as being representative of the true, reciprocal nature of brain connectivity. In Chapter 2, we use two datasets with 15 brain regions to demonstrate that the MDM-DGM can infer networks that are physiologically-interpretable. The estimated MDM-DGM networks are similar to networks estimated with the widely-used partial correlation method but advantageously also provide directional information. As the size of the model space prohibits an exhaustive search over networks with more than 20 nodes, in Chapter 3 we propose and evaluate stepwise model selection algorithms that reduce the number of models scored while optimising the networks. We show that computation time may be dramatically reduced for only a small trade-off in accuracy. In Chapter 4, we use non-local priors to derive new, closed-form expressions for the model evidence with a penalty on weaker, potentially spurious, edges. While the application of non-local priors poses a number of challenges, we argue that it has the potential to provide a flexible Bayesian framework to improve the robustness of the MDM-DGM networks.

Cellular force is essential in maintaining the normal function of a biological cell. The primary goal of this study is to develop experimental methods to quantitatively determine forces generated from cell contraction and cell-to-cell adhesion. A novel method has been developed to measure the cell contraction forces exerted within a cell-embedded collagen matrix. The technique provides a 3D cell-matrix model which allows estimation of the cell contraction forces over a certain period of time. It was found that embedded fibroblast cells are able to cause a shrinkage of their surrounding matrix due to cell contractility. Tailored equipment which has ultimate force and displacement resolutions of 10 nN and 100 nm respectively has been constructed to accurately determinate the elasticity of cell-embedded collagen matrix. In combination with a mathematical model, the cell contraction force can be calculated based on the geometric parameters of the collagen matrix before and after the shrinkage. Reagents of agonist (histamine) or antagonist (ML-7) have been used to stimulate or block the fibroblast contraction force. They both show the effect of altering the stiffness of an extracellular matrix which is critical in the determination of cell contraction forces. More importantly, the analysis of the measured data based on a non-linear mechanical model have also confirmed that the elasticity of the extracellular matrix will influence the fibroblast contraction force. Cell-to-cell adhesion is an intricate interplay of mechanical, chemical and electrical signals between cells. A novel method based on Atomic Force Microscopy Single Cell Force Spectroscopy (AFM-SCFS) has been applied to examine cell-to-cell adhesion of human kidney proximal tubule HK2 cells. Ketamine was used to evoke early changes in expression of proteins (E-cadherin, N-cadherin and β-catenin) central to adherens junctions that lead to the loss of cell-to-cell adhesion force. The results provide strong evidence that the illicit substance Ketamine has major impacts on renal function and the loss of intracellular adherent energy. Overall, both cell contraction and cell-to-cell adhesion experiments demonstrated that the changes of biological states evoke protein interactions which would ultimately lead to the biomechanical alterations. Therefore, characterising the changes of mechanobiological properties can provide new insight into the investigation of physiological and pathological issues.

Clathrin is a well-studied molecule yet its assembly properties have been characterised only in fairly simple biochemical terms to date. When clathrin is isolated for study in vitro, certain important aspects of its in vivo functionality are conserved: • purified clathrin units can be induced to assemble, reversibly, into cage-like forms that exhibit an array of polygonal binding motifs, similar to those observed in a vesicle's coat complex; • the artificial clathrin cages so produced are found to be heterogeneous, resulting in particles of different size according to the number of units each incorporates. Laboratory methods that investigate mechanisms of cage disassembly have become integral to the supply of knowledge about clathrin's structural biology. Yet there has been little progress towards a quantitative approach for exploring the assembly mechanisms that might explain clathrin cage dispersity; indeed, for some biophysical techniques the latter is an unfortunate property that might even preclude a meaningful analysis. This thesis shows that when purified clathrin is made to assemble in vitro by an overnight dialysis method (x2.8.3), the distribution of coalescent particle sizes that results can be predicted according to the pH and composition of the buffer solution and the final concentration of the clathrin sample: • Within the confines of a certain parameter space, variation in the distribution of particle sizes is found to be consistent with the predictions of a Becker-D�oring model for reversible coalescence: the empirical evidence suggests that a dynamic population of supramolecular cage structures provides a mechanism for energy minimisation towards achieving thermodynamic equilibrium. • Addition to, or substitution for, the basic compounds that constitute the assembly buffer solution exerts detectable effects upon the clathrin particle size distribution; the nature and magnitude of these effects may be expressed in terms of the state function Gibbs energy. To demonstrate these, a methodology was developed to focus on the preparation of samples suitable for analysis by the dynamic light scattering (DLS) technique, thereby aiming to provide an extensible and far richer analysis than was a afforded previously by the somewhat erroneous, certainly tedious, task of manually counting and classifying particles by size to construct a histogram from electron microscopy images. DLS determines the frequency profile of fluctuations in the electromagnetic field issued by a bulk sample to measure the diffusion processes of its particles. To then derive a number distribution relies upon exploiting prior knowledge of the physical system or else assumptions for the same. So, the research presented here has involved both a theoretical and an experimental component: over the course of the project there has been a process of iteration between developing experimental methods at the bench and interpreting the data recovered, until the physical picture and the theoretical model that have emerged appear to be consistent. And the results described in this thesis do indeed encourage further work. A number of very promising themes have not been advanced further here due only to the limitations of time: • 3. By developing a quantitative approach that relates the physical properties of the clathrin particle in solution to thermodynamic measures, the empirical results that have been collected over the course of this project are expressed in terms that translate more readily to the language of simulation and dynamical systems modelling. • 4. Analysis of DLS makes a valuable contribution to the project that aims to resolve accessory proteins bound to clathrin cages using the SANS contrast variation technique, with auxilin as a model. • Perhaps the most important conclusion to be drawn, however, is that: 5. The additional insight made available by analysis of DLS allows greater control to be exercised over the directed assembly of clathrin in vitro, which in turn provides fresh opportunities for experimental design. Chapter 1: The scientific questions addressed by this thesis; key concepts; review of the academic literature; project aims and objectives. Chapter 2: Details of the practical methods and materials used to conduct experiments in the laboratory with clathrin; software resources. Chapter 3: Description of the theoretical framework that has been developed over the course of the project; mathematical methods. Chapter 4: The empirical results for clathrin that have been recovered over the course of the project; statistical methods and models. Chapter 5: How the main findings of Part II have been used to solve a problem for clathrin studies when using a high resolution application (SANS). Chapter 6: Discussion of the main findings with respect to the project aims and objectives; directions for further work and unresolved questions. Figure 1: The narrative of this thesis. Part IV consists of appendices: statistical diagnostics, examples of programming and supplementary images.

Clathrin mediated endocytosis (CME) is an integral process in eukaryotic cells and governs a wide range of processes in higher organisms including neurotransmitter release and cell signalling, to development and cell polarity. Due to its wide ranging roles this mechanism has been implicated in various disease states such as Huntington’s disease and various cancers as well as being a method of entry into cells for many viruses and bacteria. The process of CME involves the formation of a clathrin coat, primarily consisting of the protein clathrin, that drives uptake of cargo at the plasma membrane. These interactions are facilitated through a large family of proteins known as adaptor proteins, which drive the process of CME through specific interactions with clathrin, cargo, the plasma membrane and the cytoskeleton. Once the cargo has been endocytosed the process must be reversed through the actions of the disassembly proteins auxilin/GAK and Hsc70. A number of questions remain as to how adaptors promote assembly and how auxilin and Hsc70 drive disassembly through interaction with clathrin and potentially through interactions with the adaptor proteins. By purifying adaptor proteins and clathrin I have used various biochemical and biophysical techniques to investigate these interactions in vitro. Using a novel assembly assay based on dynamic light scattering I have shown that it is possible to measure the effect of adaptors on clathrin cage size distribution during assembly. In disassembly I have shown how mutations in the disassembly protein auxilin affect its ability to catalyse the disassembly of clathrin cages and how the presence of various adaptor proteins alters the ability of auxilin and Hsc70 to disassemble these structures. Finally, I demonstrate an inhibitory effect on disassembly by the adaptor protein epsin and propose a mechanism of interaction with clathrin that can be disrupted through mutations to epsin clathrin-binding motifs and discuss the implications of this effect for the role of adaptors in vivo.

Clathrin-mediated endocytosis is the major route of internalisation of a wide range of cargo at the plasma membrane. Despite the high frequency of endocytic events, it is difficult to predict when and where a clathrin-coated pit will form, complicating any analysis of initiation mechanisms. This thesis describes the development of an inducible model of endocytosis that provides spatiotemporal control over initiation. This is achieved through chemical- or light-induced dimerisation of a clathrin-binding protein (hook) to a plasma membrane cargo (anchor), bypassing any preparation steps prior to clathrin recruitment. It was found that this rapidly induced clathrin-mediated endocytosis that could function independently of AP‑2 and without detriment to endogenous CME. This synthetic endocytosis was found to be a useful measure of functional interaction between clathrin and its adaptors, for example β3 adaptin hinge + appendage was non-functional for CME despite previous reports of in vitro clathrin binding. Also, these clathrin hooks can be easily mutated to pinpoint interaction sites. Mutagenesis of a β2 hinge + appendage hook revealed that a low level of endocytic activity could be maintained when either the hinge or appendage site was disrupted, but mutation of both sites or removal of the entire appendage inhibited synthetic endocytosis. Additionally, the ability of clathrin alone to promote clathrin-coated vesicle formation was shown by the success of a GTSE1 hook, a protein not known to interact with endocytic accessory proteins. This technique offers great potential for further examination of molecular level clathrin/adaptor interactions as well as CME mechanics on a cell-wide scale. In particular, the localised initiation of endocytosis demonstrated here using optogenetics is ideally suited to manipulate cells displaying polarised endocytosis.

During the course of this work the development and optimisation of the dichroism field of spectroscopic techniques regarding their application towards living biological systems has been examined. To this end we have studied the theory of light scattering, as one of the principle flaws that can prevent accurate analysis of spectral data. The fields of Rayleigh-Gans-Debye and Mie theory were investigated and implemented into a program for the calculation of scattering for various different systems. The developed program was then optimised using previously collected data regarding the scattering produced by liposomes under flow, and further data regarding the deformation of liposomes validated. A secondary feature to be addressed was the inherent complexity of cellular spectra that comprise a variety of different components. Experimental studies were therefore conducted regarding the deconvolution of several biological systems in terms of their individual chromophores, in increasing degrees of complexity. An approach in which each component was converted into a series of Gaussian curves and the contribution of sets of these Gaussians (from each component) to the overall system spectrum was developed. Since biological systems contain various components that interact with each other, such as the hypochromicity of DNA, the influence on this spectral decomposition approach was also outlined and studied. Unusual features were observed in the most complex cellular samples of red blood cells and fission yeast, regarding the linear dichroism (LD) spectra produced in the absence of flow. Further studies were conducted regarding the simpler of these systems, red blood cells, to determine the source of these spectral features. The principle component was determined to be oxy-hemoglobin, and that both cellular and protein orientation was required to produce the observed LD spectrum. Sedimentary forces were then discovered to be the signal source within the sample vessel, and the spectrum comprised of both absorbing and scattering elements. Of the latter, both typical RGD/Mie type scattering was observed in addition to resonance scattering. In addition, it was observed that alterations to the cellular geometry, due to various environmental influences such as pH and tonicity, were directly related to spectral behaviour.

Dynamic microtubules control cell shape, cell locomotion and the proper segregation of chromosomes. End Binding (EB) proteins are the key components of the microtubule (MT) plus tip (+TIP) protein network. EBs bind to the MT plus end and regulate microtubule dynamics. EBs localise to the microtubule tip by recognising the nucleotide state of tubulin. Mammalian cells express three members of the EB family (EB1, EB2 and EB3) that localise to spatial distinct sites on the microtubule in cells. Perturbation experiments in cells and in vitro reconstitution experiments have shown that EB1 and EB3 accelerate MT assembly and increase catastrophe frequency. This is a paradoxical effect, as an increase in growth speed should increase the size of the GTP cap thus decreasing the probability of catastrophe. To study this paradoxical effect an image analysis routine was developed to gain insight into any structural re-arrangement at the microtubule tip. An algorithm was developed to extract fluorescence intensity data along the length of a microtubule from time-lapse images. Curve fitting to these data allowed determination of the MT end position with sub-pixel resolution, the measurement of taper (i.e. the length difference of protofilaments at the microtubule end) and the quantitative analysis of the comet-shaped distributions of EB proteins at microtubule ends. The method was verified using synthetic images of MTs and then applied to time lapse movies of dynamic MTs from in vitro experiments where either the tubulin concentration or the EB3 concentration was varied. It was discovered that EB3 may increase microtubule taper, thereby de-stabilising the microtubule tip structure. Binding of the three EB proteins to spatial distinct sites at the MT tip was carefully re-investigated in-vitro by pair-wise comparison, and in relation to the MT tip. All three EB proteins were found to localise to distinct sites with EB3 found to bind closest to the MT tip and EB2 being the furthest from the MT tip. Based on structural data that became available during the course of the project, and additional evidence of different nucleotide preferences between EB1/EB3 and EB2, a dual nucleotide recognition model was conceived to explain these spatially distinct locations. The model assumes that an EB protein is sensitive to the nucleotide state at both E-sites close to its binding site at the interface of 4 tubulin dimers. All three EB proteins showed evidence of dual nucleotide recognition in mixed nucleotide lattice experiments designed to directly test the model. EB proteins recognise spatial distinct sites by recognising the pairwise nucleotide state of tubulin. As EB3 binds closest to the MT tip, it is best placed to affect microtubule dynamics by increasing taper, promoting a quicker growth phase and destabilising the microtubule. Within cells this is a useful concept as it can be up regulated to increase the dynamicity of MTs ensuring more efficient re-organisation of the cytoskeleton during cell differentiation or neuronal elongation.

We examine mathematical models for small deformations of membranes. First we review physically well-established models, posed in the Monge gauge, from a mathematical perspective. We produce a variational framework in which well posedness can be studied and finite element methods applied. The methods are used to investigate the effects of point forces, point displacement constraints and point curvature constraints. Such models are suitable for the study of deformations induced by filaments contained in the cell cytoskeleton and by embedded protein inclusions. In particular we study the membrane mediated interactions between filaments and also between inclusions. We then introduce a new linearised model which describes small deformations of closed surfaces that are minimisers of Helfrich-type energies. The deformed surface is described as a graph over the Helfrich minimising undeformed surface. This is the natural generalisation of the Monge gauge to initially curved surfaces. We focus on a Willmore energy which gives rise to spheres and a family of tori as undeformed surfaces and also introduce surface tension on a sphere. Again we study deformations induced by filaments. A variational formulation is produced which is similar to the Monge gauge case and we formulate a numerical method to study membrane mediated interactions. Finally we introduce an abstract splitting method which allows a high order PDE to be solved by an equivalent system of lower order equations. We give conditions which ensure well posedness of the system and produce a finite element method whose solution converges to the solution of the full system. The theory is applied to show convergence for the numerical methods used for the surface deformations model. We provide examples which show the theoretical error estimates are achieved.

Motivated by an intricate mechanism to transport folded proteins across biological membranes, known as the Twin-arginine translocation (Tat) pathway, we construct lattice protein models in an attempt to study the aggregation of the membrane protein TatA, which plays an integral role during active Tat translocation. We develop force field that characterizes intra- and inter-residue interactions, as well as how each residue interacts with its environment. Although written with the Tat process in mind, this thesis is mainly devoted to developing efficient Monte Carlo schemes for biomolecular simulations, which are often challenged and impeded by complex energy landscapes. To tackle the local trap problem that is typical in Metropolis sampling, the idea of dynamic weighting is incorporated into the parallel tempering (PT) algorithm. Our results show that, when applied to the lattice-protein model, the modified PT algorithm is capable of locating the low energy state much more quickly, but does not produce reliable estimates for equilibrium expectations. A modern method for free energy calculation, called the multistate Bennett acceptance ratio (MBAR) estimator, is reviewed from a statistical perspective, reminiscent of the underlying statistical theory which the method is based upon. Instead of adopting the common practice of using MBAR as a post-simulation analysis tool, we propose a new approach that integrates MBAR into simulation, allowing the simulation to benefit from the statistical optimality of the MBAR estimator. We show that the MBAR-enhanced Monte Carlo improves simulation efficiency of the lattice-protein aggregation model and, since it can also be applied to continuous models, provides a promising alternative to the study of more realistic systems. The new method is then applied to our model of TatA, where the protein features both a transmembrane and an amphipathic helix. The effect of individual helices on dimerization was studied and problem with the move set was identified. In this thesis, we used pull move and translation move as our Monte Carlo trial moves. Implementation details of pull moves, which are often omitted by many researchers who use them for sampling configuration space, are given in Chapter 1. We show that, for our double-helix TatA model, pull moves are no longer efficient moves and therefore, for future study of more realistic systems, we point to several methods which all attempt to design efficient trial moves. Aggregation of more than two polymer chains was also considered in this thesis.

Directed cell motility plays a key role in many areas of biology, with cells able to reorient quickly in response to changes of an extracellular stimuli. A complex signalling network directs this response, which motivates the use of conceptual mathematical models that replicate aspects of this behaviour and can be more readily analysed. Comparisons between such models have more focused on the qualitative differences between them. We wished to construct a framework for the rigorous comparison between models, using cell repolarisation in response to shear ow change. We fitted three reaction-diffusion models of cell polarity to experimental data of dictyostelium amoeba repolarising in response to mechanical shear flow. Experiments performed under different conditions were fitted simultaneously, to provide models with a range of cellular dynamics, with the models being fit to spatio-temporal data of cortex fluorescence of an F-actin reporter. All models were able to give a satisfactory t, with parameter identifiability determined using the pro le likelihood. The Meinhardt and Levchenko models were able to obtain better fits than the Otsuji model. Analysis of the model behaviour and parameter identifiability prompted alterations of the models, which resulted in a fully identi able two-variable Meinhardt model. Simulations of the Meinhardt and Levchenko models were used to test their behaviour over time frames past which the models had been tfitted. This motivated changes to the model parameters to obtain the desired long-term behaviour. Further simulations were run to elicit the model response to a changing external signal beyond that seen in the fitting, with the models being able to adapt to a moving signal, and respond to multiple simultaneous signals. Further fitting of the Meinhardt and Levchenko models was conducted using single cell data. The models were able to t well to data taken from both repolarising and unstimulated cells, showing that these models are able to replicate both mean and single cell spatio-temporal imaging data.

Single particle tracking (SPT) trajectories are fundamentally stochastic, which makes the extraction of robust biological conclusions difficult. This is especially the case when trying to detect heterogeneous movement of molecules in the plasma membrane. This heterogeneity could be due to a number of biophysical processes such as: receptor clustering, traversing lipid microdomains or cytoskeletal barriers. Working in a Bayesian framework, we developed multiple hidden Markov models for heterogeneity, such as confinement in a harmonic potential well, switching between diffusion coefficients, and diffusion in a fenced environment (or "hop" diffusion). We implement these models using a Markov chain Monte Carlo (MCMC) methodology, developing algorithms that infer model parameters and hidden states from single trajectories. We also calculate model selection statistics, to determine the most likely model given the trajectory. For LFA-1 receptors diffusing on T cells we show that 12-26% of trajectories display clear switching between diffusive states, depending on treatment. We also demonstrated that allowing for measurement noise is essential, as otherwise false detection of heterogeneity may be observed. Analysis of the motion of GM1 lipids bound to the cholera toxin B subunit (CTxB) in model membranes confirmed transient confinement. On this dataset we also demonstrated a clear signature in the confinement shape for individual tagging molecules, and showed that confinement events are not exponentially distributed. Finally, we developed an algorithm which detects hopping diffusion, validating on simulated data. Rather than methods which rely on generic properties of Brownian motions, our approach allows us to test which biophysical model best fits a trajectory. Using a model-based approach we can also extract biophysical parameters, segment trajectories into different motion states, and hence analyse particle motion in high detail. With the continuing improvement in spatial and temporal resolution of trajectories, these methods will be important for biological interpretation of SPT experiments.

Approximately one third of genes in the human genome (1) encode transmembrane (TM) proteins and form more than half of all drug targets (2). However, our understanding of how these proteins fold into their functional form, as well as how they may misfold into a disease-associated form, remains a difficult area of study. By observing the effects of single point mutations in the context of a native sequence, in addition to adding and mutating interhelical interaction motifs on a low complexity sequence background, we aim to elicit ‘rules’ of TM protein domain association. For the single point mutation in the context of a native sequence, the TM domain of the sequence Neu, along with its oncogenic substitution V664E form Neu*, were selected. Using molecular dynamics (MD) a united atom model of each dimer in a model bilayer system was subjected to umbrella sampling along an interhelical reaction coordinate to yield a free energy profile of self-association. The lipid order, bilayer thickness, and peptide tilt angle were calculated from trajectories taken from three points along the reaction coordinate. Helical composition, solvent accessible surface area, and hydrogen bond analysis (for the V664E substitution) were performed at the free energy minimum. Low complexity sequences of polyleucine and polyleucine-alanine heptad repeat sequences, with and without interaction motifs similar to those present in the Neu model, were ligated into PBLM100 plasmids. Transformed E. coli cells were subjected to semi-quantitative homo-interaction analysis using the GALLEX assay. The same TM sequences were modelled using a coarse grained (CG) forcefield. Umbrella sampling along an interhelical reaction coordinate was performed to yield a free energy profile of self-association. Single-linkage cluster analysis of peptides was performed at the global free energy minimum. A representative structure from each set was compared to an averaged structure from the clusters of an atomistic conformational search. The results presented in this study, could contribute to what in theory would be a large database of motif-driven rules for TM helical domain oligomerisation. This may encourage further investigation into TM protein design for novel application.

The identification and treatment of bacterial infections remains a major healthcare challenge, especially to ensure appropriate application of a limited spectrum of antibiotics. Therefore the development of alternatives to antibiotics and new analytical tools to probe pathogenic infection processes and as point-of-care biosensors is crucial to combat the spread of infectious diseases. Glycopolymers offer many opportunities for interfacing synthetic materials with biological systems. However, the nature of the interactions between glycopolymers and their biological targets, lectins, and the structural features necessary to obtain high-affinity materials are not fully understood. The application of synthetic glycopolymers to anti-adhesive therapies has so far been limited by their lack of lectin specificity. Herein a number of tandem post-polymerisation modification methods are utilised to probe the multivalent inhibition of a bacterial toxin as a function of linker length, carbohydrate density, and glycopolymer chain length. Guided by structural-biology information, the binding-pocket depth of the toxin was probed and used as a means to specifically improve inhibition of the toxin by the glycopolymer. Glycosylated gold nanoparticles that change colour due to lectin-mediated aggregation may find use as biosensors to aid in the detection of infectious diseases and biological warfare agents such as ricin. Here, carbohydrate-functionalised, gold nanoparticles have been used to discriminate between lectins and bacterial phenotypes. Optimisation of the particle coating is required to ensure stability in complex media, but still allow for rapid detection readouts.

This thesis advances the understanding of mitochondrial dynamics in adipocytes. Chapter 1 outlines mitochondrial dynamics, health, function and quality control in the context of obesity and T2DM. Chapters 2 and 3 are concerned with characterising mitochondrial dynamics in human abdominal subcutaneous adipose tissue and how the balance of this process may be augmented by adiposity and bariatric surgery. As an extension of this, Chapter 4 explores the potential regulatory role of p38 within adipogenesis in the 3T3-L1 cell line model together with the interplay between p38 and mitochondrial dynamics throughout differentiation. As part of this study, the impact of modulating p38 by chemical means or via the introduction of an overexpression vector was used to investigate the functional consequences of this gene on mitochondrial bioenergetics. The subtle balance of mitochondrial fusion and fusion markers is critical, not only for controlling the abundance and bioenergetic health of mitochondria but also, for overall cellular health. Finally, Chapter 5 extends analysis of the regulatory role of p38 on mitochondrial dynamics through computational modelling of the p38β. From the entire AstraZeneca compound collection, a series of novel and putative p38β inhibitors were selected for further functional assays to assess the effect of these compounds on mitochondrial bioenergetics in 3T3-L1.

With every corner of science, engineering and business generating vast amounts of data, it is becoming increasingly important to be able to understand what these data mean, and make sensible decisions based on the findings. One tool that can assist with this aim is the type of program called a self-organising map (SOM). SOMs are unsupervised Artificial Neural Networks (ANNs) that are used for pattern recognition, dimensionality-reduction of datasets, and can give a visual representation of the data using topology. For this project, SOMs were used to do pattern recognition on circular dichroism (CD) and myoelectric signal (MES) data, among other applications. To the first of these SOMs, we gave the name SSNN for Secondary Structure Neural Network, as it analyses CD spectra to find structures of proteins. CD is a polarised UV light spectroscopy, it is a useful for estimating structures (conformations) of chiral molecules in solution. In this work we report on its use with proteins and lipoproteins. The problem with using CD spectra is that they can be difficult to interpret, especially if quantitative results are required. We have improved the structure estimations compared with similar methodologies. The overall error across all structures for SELCON3 was 0.2, for CDSSTR: 0.3, for K2d: 0.2, but for our methodology, SSNN, it was 0.1. Another difficult problem the world faces is that thousands more people every year have limb amputations or are born with non-fully-functioning limbs. Robotic limbs can help people with these afflictions, and while many are available, none give much dexterity or natural movements, or are easy to use. To help rectify the situation we adapted the SOM tool we developed, SSNN, to work as part of a software platform that is used to control robotic prostheses, calling it HASSANN, Hand Activation Signals, SOM Artificial Neural Network. The system works by performing pattern recognition on myoelectric signals, which are electrical signals from muscles. The software platform is called BioPatRec, and was developed by Max Ortiz-Catalan and his other collaborators. The SOM HASSANN was written by the author, who also tested how well the software works at predicting which robotic limb movements are needed.

Dysfunction of the myometrium is a major contributor to preterm labour. The current limitations to effective management of preterm labour is in large part due to our, as yet, incomplete insight into the mechanisms that underlie uterine excitability. The uterus goes through a period of activation towards the end of pregnancy during which the myometrium develops the ability to deliver the baby. Gap junctions are among the most important molecular entities involved in this activation process. In this thesis, a mathematical model based on FitzHugh-Nagumo dynamics is used to study the role that gap junctions play in controlling the excitability of the myometrium. Spatial heterogeneity is introduced into the network through stochastic assignment of coupling strengths based on physiological statistics. It is demonstrated that heterogeneity amplifies the ability of a locally applied stimulus to generate global activity and that the ability of the stimulus to excite the network is strongly dependent on the local spatial correlation structure of the couplings. In networks driven by a pacemaker cell, large coupling strengths preclude activity by reducing the frequency of the stimulating oscillations. The model is extended to incorporate voltage-dependent gap junctions. It is established in the literature that gap-junctional conductance is dependent on the transjunctional voltage of neighbouring myocytes. Two conductance relationships have been observed corresponding to gap junctions composed of connexin-43 or connexin-45 proteins. It is demonstrated here that networks with only connexin-45 proteins are unable to exhibit global excitability whereas networks with connexin-43 proteins always display full activity. The mathematical models are supported by analysis of human and rat RNA expression data which shows that connexin-45 is down-regulated at term. It is hypothesised that connexin-45 blocks activity in the uterus throughout gestation, and is down-regulated at term to allow the uterus to deliver the powerful contractions associated with labour.

The determination of protein structure and the exploration of protein folding landscapes are two of the key problems in computational biology. In order to address these challenges, both a protein model that accurately captures the physics of interest and an efficient sampling algorithm are required. The first part of this thesis documents the continued development of CRANKITE, a coarse-grained protein model, and its energy landscape exploration using nested sampling, a Bayesian sampling algorithm. We extend CRANKITE and optimize its parameters using a maximum likelihood approach. The efficiency of our procedure, using the contrastive divergence approximation, allows a large training set to be used, producing a model which is transferable to proteins not included in the training set. We develop an empirical Bayes model for the prediction of protein β-contacts, which are required inputs for CRANKITE. Our approach couples the constraints and prior knowledge associated with β-contacts to a maximum entropy-based statistic which predicts evolutionarily-related contacts. Nested sampling (NS) is a Bayesian algorithm shown to be efficient at sampling systems which exhibit a first-order phase transition. In this work we parallelize the algorithm and, for the first time, apply it to a biophysical system: small globular proteins modelled using CRANKITE. We generate energy landscape charts, which give a large-scale visualization of the protein folding landscape, and we compare the efficiency of NS to an alternative sampling technique, parallel tempering, when calculating the heat capacity of a short peptide. In the final part of the thesis we adapt the NS algorithm for use within a molecular dynamics framework and demonstrate the application of the algorithm by calculating the thermodynamics of allatom models of a small peptide, comparing results to the standard replica exchange approach. This adaptation will allow NS to be used with more realistic force fields in the future.

Reduced neuron models are essential tools in computational neuroscience to aid understanding from the single cell to network level. In this thesis I use these models to address two key challenges: introducing experimentally verifi�ed heterogeneity into neocortical network models, and furthering understanding of post-spike refractory mechanisms. Neocortical network models are increasingly including cell class diversity. However, within these classes significant heterogeneity is displayed, an aspect often neglected in modelling studies due to the lack of empirical constraints on the variance and covariance of neuronal parameters. To address this I quantified the response of pyramidal cells in neocortical layers 2/3-5 to square-pulse and naturalistic current stimuli. I used standard and dynamic I-V protocols to measure electrophysiological parameters, a byproduct of which is the straightforward extraction of reduced neuron models. I examined the between- and within-class heterogeneity, culminating in an algorithm to generate populations of exponential integrate-and-�re (EIF) neurons adhering to the empirical marginal distributions and covariance structure. This provides a novel tool for investigating heterogeneity in neocortical network models. Spike threshold is dynamic and, on spike initiation, displays a jump and subsequent exponential decay back to baseline. I examine extensions to the EIF model that include these dynamics, fi�nding that a simple renewal process model well captures the cell's response. It has been previously noted that a two-variable EIF model describing the voltage and threshold dynamics can be reduced to a single-variable system when the membrane and threshold time constants are similar. I examine the response properties of networks of these models by taking a perturbative approach to solving the corresponding Fokker-Planck equation, �finding the results in agreement with simulations over the physiological range of the membrane to threshold time constant ratio. Finally, I found that the observed threshold dynamics are not fully described by the inclusion of slow sodium-channel inactivation.

This thesis is concerned with the study of biophysicochemical processes using electrochemistry and related techniques. The first part of the thesis discusses the electrochemical detection of biological species, and characterisation of the electrode materials employed. A comparison of two novel forms of carbon electrode, namely carbon nanotubes and polycrystalline boron doped diamond (pBDD), with more conventional carbon electrode materials reveals their enhanced characteristics for bioelectrochemistry, with improved sensitivity and resistance to fouling. These materials are further characterised using novel high-resolution electrochemical imaging methods, to determine heterogeneous electron transfer rates for a number of different redox species. The kinetic rate constants are determined from measured electrochemical currents using finite element method (FEM) modelling, which proves to be a powerful technique for the quantitative analysis of intrinsic system parameters that cannot be studied directly. The electrochemical response of isolated regions of pristine SWNTs is investigated using scanning electrochemical cell microscopy, demonstrating high electrochemical activity at the nanotube sidewalls. A similar analysis of the different facets of pBDD is performed using intermittent contact scanning electrochemical microscopy coupled with FEM simulations, revealing that the electroactivity is strongly in uenced by the local density of states of the material. New techniques are also presented for the investigation of transport processes at membrane interfaces. A new method of bilayer formation is developed, which overcomes many of the limitations of current techniques, and is used to investigate the permeation rates of a series of aliphatic carboxylic acids. Using confocal laser scanning microscopy (CLSM) with a pH-sensitive uorophore, the pH change as a weak acid permeates across the bilayer can be visualised, and the permeation coefficient determined by comparison with FEM simulations. This reveals a trend of increasing permeability with lipophilicity. Finally, CLSM is used to study the lateral diffusion of protons at lipid bilayers and other surfaces. Protons are generated galvanostatically by a UME positioned close to the substrate, altering the local pH which can be visualised by means of a pH-sensitive uorophore. The uorescence profile is again compared to FEM simulations, allowing the lateral diffusion coefficient to be determined.

There is a continuing need for improvements in the cryopreservation of clinically relevant cells, tissues and organs as advances in transplantation science and regenerative medicine rise alongside an aging populace that intensifies demand. Antifreeze (glyco)proteins (AF(G)Ps) and antifreeze proteins (AFPs) are classes of proteins found in cold acclimatized species. Ice recrystallization is a highly damaging process that occurs upon the thawing of frozen specimens with AF(G)Ps and AFPs limiting this effect in a process termed ice recrystallization inhibition (IRI). However AF(G)Ps and AFPs largely fail to improve in vitro and ex vivo cryopreservation due to their secondary property of dynamic ice shaping. The biocompatible and synthetically accessible polymer poly(vinyl alcohol) (PVA) has been shown to process a strong IRI activity. The IRI property of PVA along with numerous other polymers and polyols is investigated to highlight the uniqueness of PVA (Chapter 2). PVA is then explored as a cryoprotectant with red blood cells (Chapter 3), immortalized mammalian cell lines (Chapter 4) and primary cells (Chapter 5) with a significant advantageous effect observed with each cell type in terms of the number of cells recovered post thaw. However, this is despite the use of proportionately low concentrations of PVA compared to traditional membrane permeable cryoprotectants. The application of PVA as a cryoadjuvant could therefore improve the cryopreservation of cells, tissues and organs resulting in widespread clinical benefits.

The interplay between protein structure, flexibility, mobility and function is a central concern in structural biology. Here, we have studied the interactions of two different proteins, HIV-1 protease and cyclophilin A (CypA), with a variety of ligands. HIV-1 protease is a key modern drug target due to its role in the lifecycle of the virus HIV-1. CypA is a multifunctional protein which acts principally as an enzyme during protein folding, but also as the primary binding partner for the immunosuppressant drug cyclosporin A (CsA). Using a computational approach we have studied the manner in which different drugs affect the flexibility of HIV-1 protease. We have used experimental and computational approaches to investigate the binding effect of CsA on the structure, flexibility and mobility of CypA. The wealth of structural data available, particularly from X ray crystallography studies, creates an opportunity for computational scientists to provide data on flexibility and mobility in order to augment the structural data and inform future experiments. We use rapid computational tools to model the flexibility (first) and mobility (froda) with a variety of structural data as input. We characterised, with first, the effect of ligand binding on the rigidity of more than 200 X-ray crystal structures of HIV-1 protease. A similar approach has been applied to the structural dataset available for CypA. In addition, we have used froda to simulate the mobility of the protein. A new procedure, of tracking surface exposure during large-scale simulations of protein motion and combining the information with rigidity analysis data, was used to try to predict results of hydrogen-deuterium exchange NMR experiments (HDX). Experimentally, we have characterised CypA and its binding with CsA and subsequently performed NMR spectroscopy including HDX on both the unliganded CypA and the CypA-CsA complex. Small changes in chemical shift and the HDX folding core were observed upon ligand binding. This is the first study of the CypA-CsA complex using HDX. Our method for predicting the results of HDX for CypA is an improvement on the first-only approach, but we find that it is not yet sufficiently sensitive to predict the effects of ligand binding on CypA using these experiments. It is hoped that this work will contribute to the general understanding of ligand interactions, the particular interactions involving HIV-1 protease and CypA, and the applications of computational simulation using rigidity analysis and large-scale coarse-grained motion.

Acetolactate decarboxylase (ALDC) is a bacterial enzyme of the butanediol fermentation pathway that decarboxylates (S)-acetolactate into (R)-acetoin. Remarkably this enzyme also catalyses the decarboxylation of the opposite enantiomer, (R)-acetolactate, to give the same product, (R)-acetoin. It is unusual for an enzyme to convert racemic substrate into an enantiomerically pure product. This unusual stereochemical control has led to extensive study of the ALDC mechanism and the hypothesis that ALDC catalyses the rearrangement of (R)-acetolactate into (S)-acetolactate prior to decarboxylation. The research presented in this thesis sought to gain insight into the molecular mechanism of the ALDC catalysed reaction by using a combination of kinetic and structural techniques. Bacillus subtilis alsD encoding ALDC was cloned into an expression vector and a series of active site mutants were prepared. The activity of mutant AlsD were determined using a circular dichroism based assay, which identified that the two active site glutamates and a basic residue are required for catalysis. A series of chiral transition state analogues were prepared in a two-step synthesis to give enantiomerically enriched 2,3-dihydroxy-2-methylbutanoic acid in reasonable yields. Three of the compounds were identified as competitive inhibitors, co-crystallised with Bacillus brevis ALDC and structures solved to 1.1-1.6 Å. These structures, coupled with inhibition studies and site-directed mutagenesis, provide an improved understanding of the molecular processes involved in the stereoselective decarboxylation of acetolactate. A mechanism for the transformation of each enantiomer of acetolactate is proposed.

In this work, trans-diazido Pt(IV) complexes with general formula [Pt(N3)2(OH)(OCOR)(pyr)2] (where OCOR is a carboxylate axial ligand) and [Pt(N3)2(OH)2(L1)(L2)] (where L1 and L2 are aromatic N-heterocyclic ligands) have been synthesised and characterised. The chemical and photochemical properties of these complexes, as well as their photobiological behaviour, have been studied in order to check their potential as photoactivatable anticancer drugs. Four trans-diazido Pt(IV) complexes with general formula trans, trans, trans- [Pt(N3)2(OH)(OCOR)(pyr)2] (OCOR= succinate, 4-oxo-4-propoxybutanoate, Nmethylisatoate and succinate-(RGD)f peptide ligands) have been obtained by axial derivatisation of one hydroxido ligand from trans, trans, trans- [Pt(N3)2(OH)2(pyr)2]. The crystal structures of three axially-derivatised complexes have been determined by X-ray diffraction. Photoirradiation studies have shown an improved photoactivity of the carboxylate versus the dihydroxido complexes at the longer wavelengths. Release of the axial ligands was observed in the studied complexes. This fact is especially relevant in the case of the Pt(IV)-(cRGD)f complex, where the RGD was incorporated as a tumour cell targeting moiety. DFT-TDDFT calculations performed on the complex trans, trans, trans- [Pt(N3)2(OH)(Succ)(pyr)2] showed dissociative transitions at longer wavelength, which could explain the photolability observed in these carboxylate derivatives. Studies of photoactivation of the diazido Pt(IV) complexes in the presence of 5’- GMP indicate the formation of a mono-GMP Pt(II) adduct as main photoproduct, therefore DNA could be considered a potential target site for these anticancer compounds. Additionally, EPR studies showed that azidyl radical release was observed when complexes bearing the succinate and 4-oxo-4-propoxybutanoate ligands were irradiated with green light. No such result was obtained for the dihydroxo precursor showing that these complexes could be phototoxic with longer wavelength light activation. Seven trans-diazido Pt(IV) complexes, trans, trans, trans- [Pt(N3)2(OH)2(L1)(L2)] (where L1 and L2 are pyridine, 2-picoline, 3-picoline, 4- picoline, thiazole or 1-methylimidazole ligands), have been obtained by oxidation of the corresponding trans-[Pt(N3)2(L1)(L2)] precursor. The X-ray crystal structures have been determined for four Pt(IV) diazido complexes from this family of compounds. Photoirradiation studies indicate that the incorporation of a sterically demanding ligand, e.g. trans, trans, trans-[Pt(N3)2(OH)2(2-pic)(pyr)], greatly enhances the photoactivity in these complexes. DFT-TDDFT calculations are in agreement with these results, since higher intensity transitions were observed for such complex at longer wavelength. Phototoxicity studies carried out on A2780, A2780cis and OE19 cell lines with the trans, trans, trans-[Pt(N3)2(OH)2(pyridine)(n-picoline)] family concluded that steric hindrance close to the platinum centre does not favour phototoxicity. Most of the complexes were equally potent in cisplatin resistance against the ovarian cancer cell line (A2780cis), except [Pt(N3)2(OH)2(3-pic)2] and [Pt(N3)2(OH)2(4-pic)2] which exhibited some cross resistance. All of the complexes tested in both OE19 and A2780 cell lines have shown less sensitivity to OE19 than to A2780. Studies in S. pombe yeast strains (WT and ΔRad3) with trans, trans, trans-[Pt(N3)2(OH)2(pyr)2] suggest that DNA is potentially an important target for this type of compounds, although other targets are not excluded. Furthermore, live-cell confocal microscopy was performed on A2780 cells treated with the complex trans, trans, trans-[Pt(N3)2(OH)2(pyr)2] and irradiated with a low dose of blue light. The cell death, monitored by propidium iodide uptake, was captured occurring 2 h 30 min post activation.

Protein disulphide-isomerase (PDI) is the key element of the machinery which ensures correct and efficient folding of proteins that pass through the export compartments of the cell. PDI brings about both conformational and chemical change in the proteins on which it acts, but it has been impossible to probe its mechanism due to a lack of information regarding its inter-domain motion. PDI comprises a series of thioredoxin-like domains, represented as a-b-b’-x-a’-c, where a and a’ are highly similar domains containing dithiol active sites, the b and b’ –domains form a stable base, x is a flexible linker and c is an acidic tail. Complex dithiol-disulphide chemistry occurs at the active sites whilst the b’ domain is the major site for non-covalent binding of ligands; influenced by interaction with the x-linker region. The overall aim of this investigation was to characterise the flexibility of PDI, explore the possibility that it accesses a range of conformations corresponding to different functional states and interconverts between these states in response to the binding of peptides or unfolded protein. A system has been developed in which the motion of PDI can be measured in real time on the high milliseconds to seconds timescale by monitoring the inter-site distance using smFRET. PDI mutants have been produced to facilitate this (Chapter 3) and a method of specifically labelling the a and a’ active sites with pairs of fluorescent dyes has been developed to allow measurement of this distance (Chapter 4). The distribution of inter-site distances of PDI revealed by this technique confirm that PDI is highly flexible and that distinct sub-sets of conformations exist within PDI independent of active site chemistry (Chapter 5). Mutation of the b’x region of PDI has be shown to abolishe this behavior, suggesting it is related to ligand binding.

Meningococcal infection due to serogroup B disease is the leading cause of meningitis across Europe, particularly in infants. Prevention through vaccination will soon be possible utilising a new multicomponent vaccine (4CMenB or Bexsero®) principally directed against serogroup B Neisseria meningitidis. In addition to OMVs, the vaccine contains three novel recombinant proteins discovered by reverse vaccinology. Studies were performed on the biophysical characteristics of these antigens to better understand their structural properties in solution and aspects of their immunogenicity. Initially, the antibody responses to the vaccine components were assessed in a low-dose mouse model. This extreme immunogenicity model was chosen to highlight any subtle differences in the immune response to individual antigens which may have been overlooked by previously high-dose immunisation studies. The recombinant vaccine components are synthetic homologues of proteins naturally present on the bacterial cell and changes in their structure can induce loss of functional responses. By deliberately stressing the antigens to lose functional immune responses, it was possible to study the structural changes at the molecular level. Using molecular modeling and protein-specific monoclonal and polyclonal antibody binding in parallel with UV circular dichroism (CD) spectroscopy, it was possible to assess at which point changes in the structure of proteins affected their ability to induce functional antibody responses in vivo. The two fusion proteins present unique biophysical characteristics derived from their component parts. As a result, they were shown to possess distinctive profiles by CD. When combined with fluorescence unfolding studies, in which the tryptophan residues were used as reporters for integrity, CD uncovered significant structural differences among the vaccine antigens. The NHBA-FP comprises two domains having different thermodynamics while in the fHbp-fusion both halves cooperate in the unfolding process. The choice of including a β-barrel protein in the chimeric constructs proved successful as both the NHBA and fHbp exhibited excellent solvent accessibility profiles due to the numerous inter- and intra-strand hydrogen bonds stabilising the structure. This had a direct effect on their immunogenicity which could only be adversely affected by prolonged harsh treatments. Maintenance of the α-helical profile was found to be critical for the immunogenicity of NadAΔ351–405 component as was the maintenance of its trimeric organisation. Overall, results confirmed the requirement and importance of the three additional protein antigens to the bactericidal response, even when directed against the homologous OMV bacterial strain. Structural studies confirmed the quality of the vaccines antigens and supported their critical contribution to the vaccine.

This thesis is concerned with the development of new electrochemical scanning probe techniques and the application of these to biological problems. These techniques allow high resolution quantitative investigations of surface processes through measurements at a precisely placed electrode probe. A new technique, called intermittent contact scanning electrochemical microscopy, which allowed the probe-surface distance to be decisively determined through the physical interaction of the probe with the surface was developed. Separately, a new type of dual electrode probe was developed and characterised, and a new instrument (including both hardware and software) capable of a wide range of electrochemical imaging modes was developed with wide applications. The quantitative analysis of the electrochemical signal, typically measured at the probe, requires understanding the mass transport between the probe and the surface. Finite element modelling was used extensively throughout to solve the mass transport problem and therefore quantitatively analyse experimental results. Intermittent contact scanning electrochemical microscopy was used to quantify the mass transport through a porous biological membrane, dentin, that separates the pulp and enamel in teeth. Oxygen generation and consumption rates during photosynthesis were determined by measuring the local oxygen flux at an electrode placed a precise distance above a monolayer of isolated chloroplasts or thylakoid membranes. Finally, the new dual electrode probe was used to measure the reduction of an artificial electron acceptor by isolated thylakoid membranes.

The protein DC-SIGNR (Dendritic-cell specific ICAM3 grabbing non-integrin related) is a C-type (calcium-dependent) lectin, which binds highly-branched mannose oligosaccharides. DC-SIGNR interacts with a range of deadly diseases via surface glycans on pathogenic glycoproteins, and the ability of DC-SIGNR to increase the rate of infection of viruses including human immunodeficiency virus (HIV) and hepatitis C virus (HCV) makes the study of DC-SIGNR/oligosaccharide interactions very attractive. The research described in this thesis sought to gain insight into the calcium and ligand binding properties of the DC-SIGNR carbohydrate recognition domain (CRD) in solution by utilising solution state nuclear magnetic resonance spectroscopy (NMR). A protocol for the production of uniformly 15N /13C labelled DC-SIGNR CRD was developed, allowing the acquisition of heteronuclear NMR experiments and the first assignment of the calcium-bound (holo) DC-SIGNR CRD to be reported. The assignment has allowed investigation of calcium and glycan binding, as well as the pH dependence of the DC-SIGNR CRD. The data presented in this thesis reveal that the DC-SIGNR CRD is highly dynamic in the calcium-free state, with the addition of calcium resulting in global conformational and dynamic changes throughout the CRD. While calcium binding hinders the protein dynamics (particularly in the calcium binding regions), a large degree of mobility remains. The evidence that ligands are released at low pH suggests that DC-SIGNR may act as an endocytic receptor. In addition to calcium binding, interactions of the DC-SIGNR CRD with a range of ligands were investigated. In particular, interactions with the oligosaccharide Man9GlcNAc (present on the HIV viral envelope) are described, representing the first direct study of the CRD interacting with a diseaseassociated ligand. The glycans employed in this study all bind to the primary calcium binding site, supporting previous crystal data. However, each glycan displays distinct patterns of chemical shift perturbations implying that they each have different, extended binding modes. Particularly striking is the difference between the disease-associated Man9GlcNAc ligand and the ligand present in a previously published crystal structure, (GlcNAc)2Man3. An investigation of the dynamics of the CRD in the holo form and bound to the ligand Man5 shows that the CRD is highly dynamic and that glycan binding further hinders, but does not abolish, the molecular motions. The dynamics data also suggests that a ligand-induced conformational change may occur and indicates potential new binding sites which are not present in any published crystal structures. The dynamic nature of the DC-SIGNR CRD may explain the wide range of ligand specificities and affinities of the C-type lectin scaffold and suggests that the study of the ligand binding properties and dynamics of proteins such as DC-SIGNR in solution is essential to further understanding of this class of proteins.

Marine picocyanobacteria are the most abundant photosynthetic bacterioplankton occupying a wide range of habitats across the world’s oceans. In order to survive in such diverse habitats, these organisms have developed various mechanisms to respond to specific environmental challenges they might encounter. One such challenge for cyanobacteria is the acquisition and homeostasis of micronutrients such as Zn and Cu, especially for those organisms occupying a variable ecosystem with an erratic nutrient supply. Metallothioneins are metal-binding proteins that potentially participate in such metal homeostasis mechanisms in these marine picocyanobacteria. Metallothioneins are small, cysteine-rich proteins capable of binding multiple metal ions, and have attracted intense scientific interest since their discovery in the 1950s. Over the last decade, they have been reported in every kingdom, from prokaryotic to eukaryotic, from bacteria to plants, from worms to mammals. Eukaryotic metallothioneins have been extensively studied. However, characterisation of bacterial metallothioneins is still rare. This research focused on the coastal cyanobacterium Synechococcus sp. CC9311, which is unusual in possessing four metallothionein genes (sync_0853, sync_1081, sync_2379 and sync_2426) while most marine picocyanobacteria contain only one, or none. Three metallothioneins were comprehensively characterised using a range of analytical and biophysical techniques. Mass spectrometry and nuclear magnetic resonance studies combined with homology modelling led to an unambiguous Zn3Cys8His cluster for BmtA0853, a highly likely Zn4Cys9His2 cluster for BmtA2426, and three possible configurations for BmtA1081. Analysis of gene expression profiles revealed that the four metallothioneins selectively participated in zinc scavenging, zinc homeostasis, cadmium detoxification, or protection from oxidising conditions. Growth of Synechococcus sp. CC9311 under various metal treatments also revealed that this coastal strain has developed a metal intensive physiology compared to the open ocean strain Synechococcus sp. WH8102.

The complexity and size of the higher animal genome and relative scarcity of DNA-binding factors with which to regulate it imply a complex and pleiotropic regulatory system. Cisregulatory modules (CRMs) are vitally important regulators of gene expression in higher animal cells, integrating external and internal information to determine an appropriate response in terms of gene expression by means of direct and indirect interactions with the transcriptional machinery. The interaction space available within systems of multiple CRMs, each containing several sites where one or more factors could be bound is huge. Current methods of investigation involve the removal of individual sites or factors and measuring the resulting effect on gene expression. The effects of investigations of this type may be masked by the functional redundancy present in some of these regulatory systems as a result of their evolutionary development. The investigation of CRM function is limited by a lack of technology to generate and analyse combinatorial mutation libraries of CRMs, where putative transcription factor binding sites are mutated in various combinations to achieve a holistic view of how the factors binding to those sites cooperate to bring about CRM function. The principle work of this thesis is the generation of such a library. This thesis presents the development of microstereolithography as a method for making microfluidic devices, both directly and indirectly. A microfluidic device was fabricated that was used to generate oligonucleotide mixtures necessary to synthesise combinatorial mutants of a CRM sequence from the muscle regulatory factor MyoD. In addition, this thesis presents the development of the optimisation algorithms and assembly processes necessary for successful sequence assembly. Furthermore, it was found that the CRM, in combination with other CRMs, is able to synergistically regulate gene expression in a position and orientation independent manner in three separate contexts. Finally, by testing a small portion of the available combinatorial mutant library it was shown that mutation of individual binding sites within of the CRM is not sufficient to show a significant change in the level of reporter gene expression.

Live Cell Imaging and High Throughput Screening are rapidly evolving techniques and have found many applications in recent years. Modern microscopy enables the visualisation of internal changes in the cell through the use of fluorescently tagged proteins which can be targeted to specific cellular components. A system is presented here which is designed to track cells at low temporal resolution within large populations, and to extract fluorescence data which allows relative expression rates of tagged proteins to be monitored. Cell detection and tracking are performed as separate steps, and several methods are evaluated for suitability using timeseries images of Hoechst-stained C2C12 mouse mesenchymal stem cells. The use of Hoechst staining ensures cell nuclei are visible throughout a time-series. Dynamic features, including a characteristic change in Hoechst fluorescence intensity during chromosome condensation, are used to identify cell divisions and resulting daughter cells. The ability to detect cell division is integrated into the tracking, aiding lineage construction. To establish the efficiency of the method, synthetic cell images have been produced and used to evaluate cell detection accuracy. A validation framework is created which allows the accuracy of the automatic segmentation and tracking systems to be measured and compared against existing state of the art software, such as CellProfiler. Basic tracking methods, including nearest-neighbour and cell-overlap, are provided as a baseline to evaluate the performance of more sophisticated methods. The software is demonstrated on a number of biological systems, starting with a study of different control elements of the Msx1 gene, which regulates differentiation of mesenchymal stem cells. Expression is followed through multiple lineages to identify asymmetric divisions which may be due to cell differentiation. The lineage construction methods are applied to Schizosaccharomyces pombe time-series image data, allowing the extraction of generation lengths for individual cells. Finally a study is presented which examines correlations between the circadian and cell cycles. This makes use of the recently developed FUCCI cell cycle markers which, when used in conjunction with a circadian indicator such as Rev-erbα-Venus, allow simultaneous measurements of both cycles.

G protein-coupled receptors (GPCRs) comprise the largest family of cell-surface receptors involved in sensing a multitude of ligands and are consequently attractive pharmacological targets. Their study is complicated by cross-talk between signalling pathways and altered receptor pharmacology due to, for instance, receptor oligomerization. Difficulties in obtaining structural information of the receptors hinder the understanding of oligomerization and therefore it is desirable to develop alternative approaches in which to study this phenomenon. The fungal pheromone GPCRs, STE2 and Mam2, from Saccharomyces cerevisiae and Schizosaccharomyces pombe respectively are both known to oligomerize and a GxxxG motif in the first transmembrane (TM) domain of STE2 has previously been shown to mediate receptor oligomerization. Previous work on polytopic proteins suggest that individual TM helices may be treated as individually stable domains, and it may therefore be possible to study oligomerization via single TM peptides as opposed to full-length receptor. This thesis describes the use of STE2 and Mam2 to explore TM helix oligomerization and the effects of mutations on receptor trafficking, localization and cellular signalling. The development of a luminescent reporter assay for Sz. pombe, which proved more sensitive than previously used assays and is capable of generating high-throughput data, is also discussed. It was found that STE2 could couple to the Sz. pombe pheromone-response pathway and mutations in the GxxxG dimerization motif affected both signalling and trafficking. Expression of the first TM GxxxG containing domain of STE2 was insufficient for oligomerization, in line with previous reports suggesting that the presence of the second domain is required for receptor oligomerization. In Mam2, a motif was identified that appeared homologous to the STE2 dimerization motif and mutations of this motif also affected trafficking and signalling. This domain could oligomerize in isolation, and mutations of the motif abolished oligomerization. In contrast the study of more polar TM domains appeared more complicated. These findings suggest that relatively hydrophobic TM domains can be studied as individually stable units, whereas more polar domains may require the presence of other TM domains.

Many high throughput sequencing protocols for RNA and DNA require that the polynucleic acid is fragmented so that the identity of a limited number of nucleic acids of one or both of the ends of the fragments can be determined by sequencing. The nucleic acid sequence allows the fragment to be located within the genome, and the fragment distribution can then be used for a variety of different purposes. In the case of DNA this includes identifying the locations where specific proteins are bound to the genome. In the case of RNA this includes quantifying the expression levels of different gene variants or transcripts. If the locations of the polynucleic acid fragments are partly determined by the underlying nucleic acid sequence this could bias any results derived from the data. Unfortunately, such sequence dependencies have already been observed in the distribution of both RNA and DNA fragments. Previous analyses of such data in order to reduce the bias have examined the role of regional characteristics such as GC bias, or the bias towards a specific sequence at the start of the fragments. This thesis introduces a new method for modelling the bias which considers the degree to which the nucleotide sequence affects the likelihood of a fragment originating at that location. This shows that there is often not a single bias characteristic, but multiple, alternative sequence biases that coexist within a single dataset. This also shows that the nucleotide sequence immediately proximal to the fragment also has a significant effect on the fragment likelihood. This new approach highlights characteristics that were previously hidden and provides a more powerful basis for correcting such bias. Multiple alternative sequence biases are observed when both RNA and DNA are fragmented, but the more detailed information provided by the new technique shows in detail how the characteristics are different for RNA and DNA and indicates that very different molecular mechanisms are responsible for the biases in the two processes. This thesis also shows how removing the effect of this bias in ChIP-seq experiments can reveal more subtle features of the distribution of the fragments. This can provide information on the nature of the binding between proteins and the DNA with per-nucleotide precision, revealed through the change in likelihood of the DNA fragmenting at each position in the binding site. It is also shown how the model fitting technique developed to analyse sequence bias can also be used to obtain additional information from the results of ChIP-chip experiments. The approach is used to find the nucleotide sequence preference of DNA binding proteins, and also the cooperative effects associated with binding at multiple binding sites in close proximity.

Neuronal interactions with electric fields depend on the biophysical properties of the neuronal membrane as well as the geometry of the cell relative to the field vector. Biophysically detailed modeling of these spatial effects is central to understanding neuron-to-neuron electrical (ephaptic) interactions as well as how externally applied electrical fields, such as radio-frequency radiation from wireless devices or therapeutic Deep Brain Stimulation (DBS), interact with neurons. Here we examine in detail the shape-dependent response properties of cells in oscillating electrical fields by solving Maxwell's equations for geometrically extended neurons. Early modeling for compact (spherical) cells in alternating fields predicts a smaller effective membrane time constant for the field-cell system compared to direct current injection via whole-cell patch clamp. This result, predicting that cells should respond strongly to field oscillations in the kHz range, was verified later in vitro for murine myeloma cells. However, recent experiments on CA3 pyramidal cells (highly elongated neurons) in the hippocampus do not exhibit this high frequency response. In this thesis we examine the implications of modeling full two-way coupling between three-dimensional cylindrical neurons and the extracellular field utilizing three different methodologies, namely: cable equation, finite-difference and finite-element. Our modeling demonstrates that the electrotonic length and orientation of the cell to the field are key determinants of the neuronal response to oscillating fields. This explains the experimentally observed absence of the high frequency response for pyramidal neurons when the applied field direction is oriented along their dendritic axis. Additionally, we developed biophysically detailed models of neuronal membranes with quasi-active electrical properties stemming from voltage-gated currents. These are known to lead to resonances at characteristic frequencies in the case of current injection via whole-cell patch clamp. Interestingly, in the field-cell system, the resonance was masked in compact, spherical neurons but recovered in elongated neurons. Utilizing our cable and finite-element models, we investigate the effect of point-source stimulation on cylindrical neurons and find a novel type of "passive resonance" not reported before in the literature. We further extend our modeling by incorporating Hodgkin Huxley channels in to the membrane and construct a fully active, spiking model of a neuron, fully coupled to the applied electric fields. We then go on to embed the neuron in to an array of cells to validate our results at the tissue-level. These findings delineate the relationship between neuron shape, orientation and susceptibility to high frequency electric fields, with implications for DBS efficacy, ephaptic coupling in networks and the filtering properties of cortical tissue.

Those key factors, supporting re-expansion of beta cell numbers after injury in various model systems, are largely unknown. Insulin-like growth factor II (IGF-II), an important member from the IGF family, plays a critical role in supporting cell division and differentiation during ontogeny, but its role in the adult is not known. In this study we investigated the effect of IGF-II in beta cell regeneration. An in vivo model of „switchable‟ c-Myc-induced beta cell ablation, pIns-c-MycERTAM (Pelengaris, Khan et al. 2002), which exhibits beta cell regeneration once Myc is deactivated, is employed in this study of the IGF-II function. Here we show for the first time that IGF-II is re-expressed in the adult pancreas following beta cell injury. Moreover, whereas a 90% beta cell ablation was induced in both pIns-cMycERTAM/IGF-II WT (MIG) and pIns-cMycERTAM/IGF-II KO (MIGKO) mice, a recovery up to 3 months was performed. By investigating the beta cell mass and numbers our results demonstrate that re-expression of IGF-II is important in supporting the beta cell regeneration in adult mice. Moreover this study supports the utility of using such ablation-recovery models for identifying other potential factors critical for underpinning successful beta cell regeneration in vivo. Both Myc and PML contribute to the regulation of apoptosis. Recent studies suggest that Myc and PML may interact at several levels in control of cell fate. Here we examined whether loss of the PML protein, which has been shown to regulate apoptosis via the p53 pathway, can prevent or affect c-Myc-induced beta cell apoptosis in pIns-c-MycERTAM transgenic mice. Together with the Applied Neuroinformatics Group at the University of Bielefeld in Germany, we have been developing and validating a machine learning based system to analyze beta and alpha cell numbers (Herold, Zhou et al. 2009). Comparative results between traditional techniques and this new method are presented here.

One of the challenges of quantitative approaches to biological sciences is the lack of understanding of the interplay between form and function. Each cell is full of complex-shaped objects, which moreover change their form over time. To address this issue, we exploit recent advances in confocal microscopy, by using data collected from a series of optical sections taken at short regular intervals along the optical axis to reconstruct the Endoplasmic Reticulum (ER) in 3D, obtain its skeleton, then associate to each of its edges key geometric and dynamic characteristics obtained from the original filled in ER specimen. These properties include the total length, surface area, and volume of the ER specimen, as well as the length surface area, and volume of each of its branches. In a view to benefit from the well established graph theory algorithms, we abstract the obtained skeleton by a mathematical entity that is a graph. We achieve this by replacing the inner points in each edge in the skeleton by the line segment connecting its end points. We then attach to this graph the ER geometric properties as weights, allowing therefore a more precise quantitative characterisation, by thinning the filled in ER to its essential features. The graph plays a major role in this study and is the final and most abstract quantification of the ER. One of its advantages is that it serves as a geometric invariant, both in static and dynamic samples. Moreover, graph theoretic features, such as the number of vertices and their degrees, and the number of edges and their lengths are robust against different kinds of small perturbations. We propose a methodology to associate parameters such as surface areas and volumes to its individual edges and monitor their variations with time. One of the main contributions of this thesis is the use of the skeleton of the ER to analyse the trajectories of moving junctions using confocal digital videos. We report that the ER could be modeled by a network of connected cylinders (0.87μm±0.36 in diameter) with a majority of 3-way junctions. The average length, surface area and volume of an ER branch are found to be 2.78±2.04μm, 7.53±5.59μm2 and 1.81±1.86μm3 respectively. Using the analysis of variance technique we found that there are no significant differences in four different locations across the cell at 0.05 significance level. The apparent movement of the junctions in the plant ER consists of different types, namely: (a) the extension and shrinkage of tubules, and (b) the closing and opening of loops. The average velocity of a junction is found to be 0.25μm/sec±0.23 and lies in the range 0 to 1.7μm/sec which matches the reported actin filament range.

A mesh-free numerical method for solving partial differential equations (PDE) on geometrically complex domains is presented. The approach is based on the general Partition of Unity Method (PUM), a framework developed by Babuska and Melenk in [8, 122]. In order to underline the universality of our method, we summarise the current state-of-the-art of the PUM in a comprehensive, self-contained way. Additionally, as an interesting research direction, fundamental and recent a posteriori error estimation techniques are discussed and transferred to the PUM. These are employed to steer an adaptive refinement process with the aim to reduce the overall error in an efficient way, Le. with the least additional computational complexity. Particular attention is paid to the targeted application area: the life sciences and, more specifically, transport processes in cell biology. Biological environments exhibit several features which make them quite different to classical fields of application. We discuss some of these properties and propose a novel approach to define and examine complex transport mechanisms based on 'elementary modules' describing typical phenomena. This systematic approach culminates in the definition of a biological transport toolbox whose building blocks can be plugged together to form more complex systems. One of the paramount challenges one has to tackle in biological simulations is the spatial discretisation of c'omplex geometries. Hence, we develop a general method to automatically handle complicated shapes which may be organised hierarchically, thus allowing for the simulation of processes spanning several nested domains. Additionally, processes may also take place on the interfaces, reminiscent of biological binding and transport phenomena on membranes. The numerical technique employed in this case is a mixed mesh-based and mesh-free discretisation. '!\vo interesting and important biological systems are discussed in order to relate the presented ideas to real-world applications. First, we summarise some properties of the nucleus of the eukaryote cell, a compartment where nearly the entire genetic information of a cell is contained. Second, the thylakoid, an organelle of the chloroplast of the plant cell, is depicted. Typical transport processes encountered in these model systems are then examined to demonstrate the developed methodology. The powerful combination of a systematic and modular modelling approach and 'in silica' experiments based on a versatile numerical method represents, we hope, an important contribution to the interdisciplinary research required to attack the multitude of challenges the study of biological systems poses.

This thesis presents a method for the induction and selection of cytoplasmic respiratory deficient mutants of Saccharomyces cerevisiae, distinct from the pleiotropic rhoˉ petites. The procedure utilizes a parental stock carrying the nuclear recessive pet 9 (op1) mutation. Mutagenesis is performed under growing conditions in the presence of manganese. As a result of this treatment twenty one mutants, putatively designated mitˉ, have been isolated. Among these are strains that are defective in the biosynthesis of either cytochrome b or cytochrome oxidase. Two mutants, D603-3B/8 and D603-3B/9, have been intensively studied following the observation that they possess spectrophot-metrically detectable levels of cytochromes c, b and a + a3. The mitˉ mutations carried by these strains have been further genetically characterized and mapped with respect to known mitochondrial markers. These lesions have been located close to, and are possibly allelic with, the pho 1 segment of the genome. In addition, a very tight linkage to the olir2 antibiotic resistance locus has been established. Attempts to determine whether the mitˉ mutations of D603-3B/8 and D603-3E/9 are allelic, have proved unsuccessful. The effect of these lesions, in constructed mitˉ diploids, on growth in glucose and glycerol liquid media has also been investigated. It is suggested that D603-3B/8 and D603-3B/9 are defective in the biosynthesis of the CS-ATPase complex. Their potential use in studies of the biogenesis of this multisubunit enzyme is also discussed.