Search

The formation of amyloid fibrils is a hallmark of a wide range of prevalent and devastating human disorders, including Alzheimer's disease and Parkinson's disease. With effective treatments generally lacking, these conditions not only cause tremendous personal suffering but also impose a substantial and growing socioeconomic burden worldwide. Understanding the mechanism of amyloid formation and how exactly this process is linked to pathology is crucial to inform drug discovery efforts and to ultimately tackle these diseases. In this work, I have demonstrated how mathematical modelling based on chemical kinetics can be used to overcome current challenges in deducing the microscopic mechanism of amyloid formation, in particular in the context of how external perturbations may affect this process. Furthermore, I extended this state-of-the-art approach to account for the stochastic nature of protein aggregation. Stochastic effects affect amyloid formation in two main ways, namely through (i) variability at the molecular level and (ii) the macroscopically apparent variation in the rates of repeated experiments. Bayesian statistics provides a principled, versatile, and robust theoretical framework to account for such variability. In addition, Bayesian inference allows for the incorporation of information gained in prior experiments and enables the estimation of uncertainty in the results obtained and conclusions drawn. Using Bayesian inference in concert with automated image analysis, I addressed point (i) and investigated how heterogeneity in the fibril population affects their ability to self-replicate. To address point (ii), I developed a novel Bayesian hierarchical model to appropriately treat and quantify the systematic variability between repeated experiments observed in most kinetic data sets in the field. Having such a conceptual framework is of key relevance in the context of disease, where data are generally sparser and more heterogeneous than in vitro and where even small variations relative to the aggregation time scale can have a significant effect on patient outcome and quality of life. Thus, my work lays the foundation for the effective and transparent mechanistic analysis of in vitro, in vivo, and ultimately patient protein aggregation data.

The accumulation of misfolded proteins is a universal mechanism in neurodegenerative diseases such as Parkinson's disease (PD), Alzheimer's disease (AD) and prion disease. In mammalian cells, the majority of misfolded proteins are degraded by the ubiquitin-proteasome system (UPS), in which selective substrates are unfolded and cleaved into small peptides when passing through proteasome. The underlying pathogenesis of neurodegenerative diseases is at least partially contributed by the dysfunction of UPS. In this thesis, the kinetics of ubiquitination conjugation, which is the first step to induce selective degradation in the UPS was studied using fluorescence-based methods. From the change of FRET efficiency, the rates of specific combinations of ubiquitination enzymes E1-E2-E3 can be determined. This method does not require labelling of enzymes to study the ubiquitination activity. Next, this ubiquitination activity was investigated upon incubation with α-synuclein and tau proteins. Biochemical and biophysical analysis revealed the ubiquitination sites on the amyloid proteins and its effects on the formation of protein aggregates in vitro. To obtain further insights into how UPS maintains the quality control of the cellular proteome in response to a proteotoxic stress induced by protein aggregates, two cell models were generated to visualise the proteasome and ubiquitin (Ub) in live cells. Proteasomes were observed to form transient foci with distinct characteristics under the proteolytic stress, some foci co-localised with aggregates. Single molecule tracking experiments indicated these proteasome-containing bodies act as active proteolytic centres to provide cells with an extra defence function against stress. In addition, a photoconvertible Ub construct enables the direct observation of cellular degradation rates by UPS under various stresses. Initial results showed that proteotoxic stress inhibits the UPS function. Overall, my studies on UPS and neurodegenerative proteins provide a new assay to study ubiquitination kinetics and characterise how UPS responses to proteolytic stress. These data may also have implications for the mechanism of the formation of pathogenic protein deposits and its links with downregulated UPS function due to aging and degradation-resistant aggregates.

The aberrant aggregation of intrinsically disordered α-Synuclein (αS) is associated with Parkinson's disease (PD) and other neurodegenerative diseases termed synucleopathies, whereby αS aggregates are a major constituent of Lewy bodies, a hallmark of these diseases. Despite the apparent link between αS and neurodegeneration, the exact physiological function as well as the events initiating pathology are still elusive. Extensive literature evidence suggests that αS is involved in regulating neurotransmission, a function attributed to its metamorphic character enabling the interaction with a number of presynaptic membranes, where the protein localises. Under physiological conditions, αS is present in a dynamic equilibrium between membrane-bound and cytosolic states. Understanding the subtle variations in the membrane binding modes in different biological contexts is crucial to elucidate the physiological behaviour of αS. Therefore, the first half of this PhD aims to characterise molecular interactions and structural conformations of αS upon association with various presynaptic membranes. The results indicate that cholesterol modulates the overall affinity of αS to synaptic-like vesicles (SL-SUV) by reducing the local membrane affinity in the region 65-97. This in return enhances vesicle-vesicle interactions via a previously reported double-anchor mechanism, which has important implication on synaptic vesicle clustering (chapter III). In addition, direct evidence highlights that the docking of SL-SUVs to the surface of the inner leaflet of the plasma membrane (IPM) via the double-anchor mechanism is driven by preferential binding modes of the protein to the different membranes. Further data reveals that changes in the PM lipid composition(s), commonly associated with neurodegenerative contexts, influence the binding affinity specifically of the non-amyloid component (NAC, residues 61-95) region (chapter IV). Overall, these results conclude that lipid compositions moderate the interactions of αS with the amyloidogenic NAC region playing a critical regulatory role in the physiological properties of the protein, ultimately dictating the intricate balance between functional and pathological behaviours of αS. The second part of this PhD investigates the aggregation of αS under quiescent conditions. At low pH, as found in the lysosomal pathway, the cholesterol in SL-SUVs significantly accelerates the aggregation of αS and changes the morphology of the end-product fibrils, whilst preserving the same secondary conformation (Chapter V). The aggregation of αS is also probed at neutral pH in presence of a relevant potassium chloride background illustrating a phosphate dependent fibril formation and stability. These fibrils display characteristics of dissociation in a phosphate-depleted environment. The kinetic mechanism driving the aggregation is proposed to be secondary nucleation which was previously believed to only occur in presence of mildly acidic pH (Chapter VI). The findings reveal new crucial insights into spontaneous αS aggregation and the conversion from its soluble monomer into toxic aggregates linked to neurodegeneration. Taken together, this PhD work significantly improves our understanding of how membrane binding influences properties of αS under physiological and pathological context as well as how disruptions in the fine-tuned equilibrium of αS result in protein dysfunction and aggregation. Comprehending these properties and events will assist in elucidating the role of αS in not only the function but also the aetiology and pathogenicity of PD and other synucleinopathies.

This thesis is a compilation of multiple techniques developed to characterise solutions of proteins, or proteins themselves. Proteins are the building blocks of life, and thus they are interesting to study for a deeper understanding of the mechanisms of life but also for developing diagnostic tools as well as medical therapies. The first chapter is a general overview of the thesis with an explanation of the different points developed further in the thesis such as proteins, protein aggregation, microfluidics, and optics. The second chapter describes a microfluidic platform for performing a 2D characterisation, separating the particles using capillary electrophoresis and diffusional sizing. This allowed to characterise the components of heterogenous solutions in function of their size and the mobility of the proteins. The third chapter is looking at existing well-known microfluidic techniques, like free-flow electrophoresis and diffusional sizing, combined with confocal microscopy. The single molecule detection allows to distinguish particles in function of their brightness, breaking the homogenous limitation of these microfluidic techniques under classical epifluorescence microscopy. The fourth chapter shows an elegantly simple way to characterise an exponential decrease of the residence time distribution within a nanocavity trap. The explanation is based only on a random walk assumption without requesting potential interactions. The fifth chapter describes a new method to size particles in solution based on interferometry scattering microscopy by looking at the correlation of the intensity of the interferometric term. The last chapter before the overall conclusion is based on a study of SARS-COV2 seroprevalence in the swiss canton of Zürich between December 2019 and January 2021. This study was conducted on two cohorts, the University Hospital patients, and the blood donation services of Zürich.

Protein and peptide drugs have gained increasing importance in the portfolio of pharmaceutical companies reflected by the growing number of approved drug products in the US and Europe. One of the key challenges represents physical instability of the drug substance, in particular processes of self-association. Several excipients are available to modulate the underlying protein-protein interactions responsible for self-association processes. Here, Cucurbit[7]uril, a macrocycle able to bind hydrophobic amino acid residues is examined for its potential to improve physical stability of protein and peptide therapeutics. Studies presented here, compare solution behaviour of proteins and peptides with mutated Cucurbit[7]uril binding sites examining whether addition of Cucurbit[7]uril is capable of mitigating or fully suppressing self-association processes. One study determines the effect of Cucurbit[7]uril on the aggregation propensity of a monoclonal antibody. Results provide evidence of specific interaction between Cucurbit[7]uril and aromatic amino acid residues located on an aggregation-prone region resulting in improved solution behaviour. Another study, determines the effect of Cucurbit[7]uril on the kinetic profile and morphology of Enfuvirtide fibrils, a commercially available peptide drug. Modulation of fibrillation onset via Cucurbit[7]uril is monitored by Thioflavin T assays and circular dichroism spectroscopy. In contrast to previous reports, findings presented here, indicate Cucurbit[7]uril's potential to delay rather than fully suppress fibrillation onset through binding of a C-terminal aromatic amino acid residue. Furthermore, this work contains a study evaluating a high-throughput PEG solubility assay for its correlation with the diffusion interaction parameter. Conventional methods used in solubility assessment of proteins have high material consumption and are not amenable to high-throughput screening. A set of 8 mAbs formulated at 15 different solution conditions is used to investigate correlation with observations made from DLS measurements. Results highlight correlation of trends observed in PEG-induced phase separation with averaged protein-proteins interactions as described by the diffusion interaction parameter.

When developing a new monoclonal antibody, the formulation is key in improving the stability of the concentrated therapeutic. The formulation design ensures the therapeutic protein is conformationally and colloidally stable and maintains a long shelf life. Excipients are added to improve the stability of proteins and are usually added at high concentrations to exert excipient effects. The aim of this thesis is to explore the effects of low concentration additives to improve conformational and colloidal stability of model proteins. Several parameters were used to measure conformational and colloidal stability. The melting temperature (T_m), midpoint of denaturation (C_m) and change in free energy between the native and unfolded state (DeltaG) measurements were conducted to measure conformational stability. Measuring colloidal stability involved static light scattering (SLS) measurements to obtain second virial coefficient (B₂₂) measurements and onset temperature of aggregation (T_{agg}) from SLS and dynamic light scattering (DLS) thermal ramps. DLVO model was applied to obtain the Fuchs stability ratio, which describes the probability of two aggregates sticking together following collision, giving another parameter to assess colloidal stability of different solution conditions. Chapter 3 highlights the effects of low concentration of tripolyphsophate (TPP) and adenosine 5'-triphosphate (ATP) on aggregation propensity. The binding of 5 mM TPP and ATP to an acidic protein, recombinant human serum albumin (rHSA), increased repulsive protein-protein interactions (PPI). This in turn significantly reduced aggregation propensity, which we show was due to an increase in the repulsive barrier upon the approach of two rHSA molecules, as quantified by the Fuchs stability ratio. Chapter 4 explores the effects of low concentrations of arginine and glutamate containing dipeptides on the aggregation behaviour of three model proteins. The effects exhibited by the dipeptides could be explained by electrostatics, the addition of the dipeptides to rHSA and ovalbumin, the most repulsive proteins, increased aggregation propensity. On the other hand, addition of the dipeptides to alpha-chymotrypsinogen, which exhibited attractive PPI, improved colloidal stability. The final chapter uses chemical denaturants to observe excipient effect on conformational and colloidal stability of partially and fully unfolded protein. Colloidal stability of unfolded protein is also defined by the size of aggregates formed following dilution refolding. These experiments provide a new method to predict aggregation propensity. Low concentration excipients are developed in this thesis to provide a novel way to reduce protein aggregation. Excipients are also used to determine effects on colloidal stability of partially unfolded protein.

The aberrant aggregation of α-synuclein in Parkinson's disease and related synucleinopathies is a histological hallmark of these conditions, as well as a cytotoxic process with no known disease-modifying interventions. While many studies have sought to develop small molecules or other biologics that may abrogate fibril formation, it is becoming increasingly clear that the toxicity associated with α-synuclein aggregation may be linked to soluble oligomeric species that serve as precursors to mature fibrils. These intermediate species are transient, amorphous and highly heterogenous in solution, which has precluded the study of their population dynamics in the context of drug discovery and screening. This thesis will aim to leverage a combination of theoretical and experimental approaches to model the flux towards oligomeric species formation within the fibril amplification process of α-synuclein, and subsequently to employ this flux as a selection parameter in drug discovery. By using this approach, possible strategies will be investigated to reduce oligomeric species formation and show that a particularly effective approach is to design small molecules that bind to the autocatalytic fibril surface of α-synuclein in a structure-based approach. Overall, the thesis aims to demonstrate the viability of the prediction and targeting of oligomeric α-synuclein species during secondary nucleation as a drug discovery pipeline that yields more robust links to the in vitro profile of an α-synuclein aggregation inhibitor and its resulting effect on α-synuclein-mediated toxicity in Parkinson's disease pathology.

Parkinson's disease (PD) is a currently incurable neurodegenerative disease with motor and non-motor symptoms that impacts the patients' everyday function and severely affects their quality of life. Current treatments are symptomatic and do not target the cause and molecular mechanism of the disease, which remains elusive. Evidence points towards the aggregation of a small presynaptic protein, alpha-synuclein (aSyn), from its functional disordered form into amyloid fibrils with characteristic β-sheet structure, as being at the centre of PD. However, many unanswered questions remain with regards to the aSyn aggregation mechanism in neurons and, in particular, the earliest steps of this process, when the soluble, functional form of monomeric aSyn misfolds. The motivation of the current thesis is the understanding of the triggers for misfolding, and the structural transitions from the monomer to the fibrillar level in different cellular environmental conditions, to aid the better design of therapeutics in the future. To this aim, different environmental conditions were interrogated with regards to aSyn aggregation propensity, monomer conformation, and amyloid fibril structure. This was approached by studying the aggregation of aSyn in vitro in three levels, with a combination of orthogonal biophysical assays: the bulk aggregation kinetics were monitored by a series of fluorescence-based kinetic assays, the monomeric conformation of aSyn was analysed via Hydrogen-Deuterium Mass Spectrometry (HDX-MS) and the fibril products formed at the end of the kinetic assays were analysed structurally by Atomic Force Microscopy (AFM). The environmental parameters interrogated were high calcium concentrations, (which are particularly relevant in the context of PD pathology), the salts NaCl and KCl (which are the principal ion components of the extracellular and intracellular space, respectively), reduced pH 4 (which roughly corresponds to the lysosomal compartment), building up to a combination of the above to mimic three cellular compartments: extracellular, intracellular, and lysosomal. The study was performed for wild type (WT) aSyn and six familial mutants (FM) (A30P, A53T, E46K, A53E, H50Q, G51D), which are implicated in hereditary PD. In parallel, the effects of aSyn-lipid association on the lipid membranes, and in particular mitochondria were studied to determine whether aSyn brings upon toxicity in additional ways to aggregation, such as mitochondrial dysfunction. This study was performed in vitro with the use of correlative super-resolution optical microscopy (structured illumination microscopy -SIM) and AFM in isolated mitochondria. The first step of this process was the optimisation of the aSyn purification protocol and the production of all the aSyn variants to sufficient amounts and purity. Following that, the combination of cutting edge HDX-MS with established biophysical assays (kinetics, AFM) highlighted a correlation between a shift in the monomeric conformation of aSyn, the protein's aggregation propensity, and the resulting fibrillar polymorphism. It was established that exposure of the structure to the solvent at the monomer level, particularly at the N-terminus and NAC region, correlates with an increase in aggregation kinetics rates and the formation of more twisted and toxic fibril as the aggregation product. With regards to the environmental conditions, the Ca2+ ion and low pH values were highlighted as an important factor in WT aSyn misfolding, while the monovalent Na+ and K+ ions had a moderate effect on aggregation and the formation of toxic fibril polymorphs. Distinct aggregation kinetics and fibril polymorphs were also observed for the FM of aSyn, pointing towards different monomer conformation distributions and possibly distinct pathways to pathology. With regards to aSyn and lipids, monomeric aSyn in its calcium-bound state was found to reduce the stiffness of mitochondria, possibly further implicating calcium in PD pathology. In this thesis, overall, the identification of aggregation-inducing conditions and the structural characteristics of the monomer and fibrils in said environments is expected to aid in the rational design of therapeutic molecules to prevent aSyn aggregation and PD pathology.

Alzheimer's, Parkinson's and Huntington's disease all belong to the group of amyloid pathologies also called protein misfolding diseases. Since the first discovery of amyloid fibrils of the aggregated protein tau in inclusion of Alzheimer brains samples, research has focussed on how amyloids form and their biological relevance in neurodegenerative diseases. Microfluidics are a well-established tool for the study of protein aggregation in a controllable environment, whereas super-resolution microscopy techniques have been developed and allow imaging protein misfolding in biological models in-vitro e.g. primary neural cell cultures. It was found that during the process of aggregation, misfolded proteins develop variations in toxicity which can be related to their size. Since their aggregation kinetics for different subspecies are again partially unstable depending on buffer conditions, we seek for methods for better and faster analysis of misfolding proteins on a single molecule level. Nanofluidics are capable of sampling single molecules from a concentrated solution but are not accessible to a broad community which would greatly benefit from their potential. To allow a broader community access to nanofluidic fabrication, first, a custom-built open-source two-photon lithography was implemented to demonstrate the usage of 2-photon direct laser writing for the fabrication of master molds of nanofiltration chips. The system was characterized and its nanolithography capabilities, using electron microscopy and atomic force microscopy, evaluated. The fabrication process is outlined and important details about the integration of nanofluidic functionalities into microfluidic masters elucidated. The produced chips were then used for a variety of applications ranging from single molecule diffusional measurements of Rhodamine 6G, GFP and human tau protein in solution, for entropic trapping of biological specimen and electrophoretic concentration of fluorescently labelled DNA. The flexible fabrication scheme allowed the enhancement of nanofluidic channels with adjascent nanotraps. The nanotraps allow to observe a variety of biologically relevant species such as 100 nm colloids, 40 nm colloids, exosomes, alpha-synuclein oligomers and 40 bp fluorescently labelled DNA with up to five times increased residence times in a confocal detection volume. Fluorescence burst microscopy is used to detect the residence times of the specimen with high precision. The characteristic Kramer escape time is extracted from the particles residence time probability distribution and allows an estimate of the particle size in the confined volume. The increased residence times are beneficial for photon expensive techniques to be conducted, which are hard to achieve without immobilizing them as with conventional approaches. To study biological samples related to nanofiltration and protein misfolding diseases in vitro, a custom-built pulsed-STED system was equipped with a second excitation beam path to allow for two-colour STED imaging using the same IR depletion wavelength. The system is then used on three different biological imaging applications. Single-color STED imaging is used to study the nanofiltration barrier of the kidney in minimal change disease (MCD) and compared to other super-resolution techniques such as STORM, SIM and expansion microscopy. Results show an effective alternative imaging pathway for the study of MCD without electron microscopy techniques. Secondly, STED imaging is combined with microfluidics which allows the observation and study of misfolding proteins in neural cell cultures in microfluidic devices. A microfluidic chip design consisting of two reservoirs which are connected via microchannels, allows the selective study of axonal synaptic transport processes by fluidically isolating two neural cell cultures from each other but keeping their axonal projections intact. Live neurons were exposed to human tau monomer and results indicate the inclusion of tau aggregates in lysosomes, which are then transported along the cellular microtubule network. Thirdly, STED is used to image the morphology of the endoplasmic reticulum (ER) in primary neurons. Calnexin is a chaperone-like membrane protein on the ER surface which is due to its close connection to glucoglycans of great interest to protein misfolding tauopathies. STED and expansion microcopy are demonstrated as methods to study Calnexin on the ER membrane and microtubular network with high resolution. Lastly, correlative 2-colour STED/AFM is presented on α-synuclein fibrils and synaptic vesicles, which combines fluorescence microscopy with sub-diffraction resolution and label-free mechanical mapping at the nanoscale.

Alzheimer's disease (AD) is a neurodegenerative disorder, the leading cause of dementia worldwide, and is one of the most pressing challenges in ageing societies. The disease is characterised by extensive brain atrophy, synaptic loss, and the accumulation of two types of protein aggregates: amyloid β (Aβ) plaques and tau neurofibrillary tangles. The deposition of Aβ can be causally linked to familial and early onset AD, however Aβ-targeting treatments against the more common sporadic form of the disease have so far shown limited to no efficacy. If early onset AD can be considered an autosomal dominant disease, sporadic AD is multi-factorial, and it might not have a unique cause. Indeed, the two largest risk factors for late-onset AD are not linked directly with Aβ processing: the first one being age and the second being the isoform APOE4 of the lipoprotein APOE-which is involved in lipid and cholesterol metabolism. In light of the high-dimensionality of this problem, in my thesis I will follow a bottom-up (biophysics to systems biology) approach in dealing with the influence of systemic complexity in protein aggregation across different components of cellular homoeostasis, in particular lipid membranes and the proteostasis network. In the first chapter of the thesis, I will present a selection of the scientific literature on AD, with a particular focus on systemic aspects, such as cellular networks of quality control, protein co-aggregation, and ageing. The second chapter will address the relationship between protein aggregation and the complexity of the lipid environment in which it occurs. This will be done by studying the aggregation of recombinant Aβ (rAβ) peptide in vitro in presence of model lipid membranes. After describing the theoretical foundations of the study of protein aggregation via means of chemical kinetics, I will show how different lipid species affect the aggregation kinetics of rAβ42 and highlight on how lipid-peptide interactions affect the ability to apply standard models of protein aggregation at the molecular scale. Then, I will present results demonstrating the acceleration of rAβ aggregation by cholesterol via a mechanism of heterogeneous primary nucleation. Finally, I will show the effect that mixtures of lipids of increasing complexity have on rAβ42 and introduce the concept of resilience in complexity-where complex lipid mixtures show an overall neutral kinetic contribution to amyloid aggregation and induce resistance to the onset of cooperative phenomena such as heterogeneous primary nucleation. The third chapter will analyse the role of a second critical component in protein aggregation: the proteostasis network. Combining mRNA and protein data from the turquoise killifish, an animal model for molecular ageing, with predictors of biophysical properties such as chaperon requirement for folding and intrinsic disorder, I will demonstrate strong links between the proteostasis network, protein stability, and protein aggregation upon ageing. In the final chapter, I will summarise and discuss some potential lines of research to expand and further generalise the work described in this thesis. AD is a complex disease, so adopting a systemic approach to the study of protein aggregation processes underlying its aetiology is of crucial importance from a conceptual point of view. Moreover, this framework has the potential to suggest network-wide therapeutic solutions for this and other neurodegenerative disorders.

The self-assembly of soluble proteins into amyloid aggregates is a mechanism which can be observed in processes that are functional or disease-associated. The resulting fibrillar structures can be formed from a variety of different proteins, but all share a β-sheet rich architecture and a variety of remarkable optical and mechanical properties. Their disease association makes amyloids relevant for medical research, but the characteristic structure and associated properties have also led to an interest in harnessing them as building blocks for functional materials. This thesis presents the development of methods for the study and manipulation of amyloid self-assembly. Currently, techniques for patterning and controlled deposition of the aggregates are lacking, which limits a wider use of the material. The first part of this thesis therefore introduces a method to pattern amyloids into arbitrary structures with macroscale order, while maintaining their characteristic properties. β-lactoglobulin proteins are electro spun into free-standing and surface deposited fibres with micron diameter and centimetre length. Structural, mechanical, and optical characterisation methods are used to demonstrate the fibres' high mechanical strength, as well as the fluorescent and light-guiding properties of the material. This new, optically active, and strong biomaterial has potential for applications in interconnected photonic and bioelectronic devices. The second part of this thesis presents the design and implementation of a structured illumination microscope (SIM) system, which can be used for the study of pathological amyloids and their effects on fast processes in living cells. By utilising a single galvo scan mirror for interferometric pattern formation, as well as a multichannel detection system, the instrument becomes faster, achromatic, and more flexible for adaptation to different imaging requirements than previous implementations of SIM. The resolution enhancement and the multi-channel capabilities of the instrument are demonstrated in up to three colours simultaneously on a variety of test samples. Due to its simple and cost-efficient design, the instrument is envisaged to make SIM available to a wider range of studies on pathological amyloids.

Alzheimer's disease (AD) is an untreatable neurodegenerative disorder that affects over 50 million people worldwide. This disease is associated with the aberrant misfolding and aggregation of the Aβ peptide into amyloid plaques in the brains of affected individuals. Despite substantial progress in the understanding of the mechanism of aggregation of Aβ, the variety of ways in which imbalances in brain metabolism can influence this process are still poorly understood. In this thesis, we address this problem by using a range of different biophysical techniques applied to ex vivo and in vivo experiments. Our three main goals are: i) to investigate the interactions of brain metabolites with Aβ, and ii) to develop an approach to systematically track the in vitro and in vivo aggregation of Aβ and iii) identify a range of natural compounds to inhibit the aggregation of Ab. We begin with an introduction of AD, with a specific emphasis on its association with Aβ aggregation. We then describe in detail the characterisation of neurotoxic oligomeric species of Aβ stabilised by an endogenous brain metabolite. In the second part, using highly reproducible chemical kinetics and imaging quantification, we describe a novel approach of using the amyloid dye X-34 to track both in vitro and in vivo the aggregation process of Aβ and further explored these findings in the context of a-synuclein, a protein closely associated with Parkinson's disease. Taken together, these results illustrate the opportunities offered by combining in vitro aggregation assays with in vivo aggregation kinetics. Finally, we explored several natural compounds in their ability to inhibit the aggregation of Ab42, characterising the molecular mechanisms of inhibition and validating our strategy in an animal model of AD. We conclude by discussing the potential impact and expected challenges of our systematic in vitro and in vivo approach, as we aim to achieve a more complete understanding of the role of brain metabolism in the pathogenesis of AD and other protein misfolding diseases.

The self-assembly of proteins and peptides into complex supramolecular structures provides an important avenue in the development of materials for biomedical applications. In particular, the design and fabrication of protein and peptide-based hydrogels has emerged as an attractive route toward novel materials with tunable three-dimensional chemical and physical structure, biodegradability, biocompatibility, as well as drug loading and release properties. These unique characteristics offer great potential for the utilisation of self-assembled hydrogels as drug delivery vehicles, biomedical materials, and tissue engineering scaffolds. A class of materials highly suitable for such applications due to their biodegradability and lack of cellular toxicity is silk-derived proteins. Regenerated silk fibroin (RSF) in particular, retains most of the properties of native silk but is readily available through scalable processes. This natural block copolymer has the propensity to self-assemble into a fibrillar network that is β-sheet rich, and its biocompatibility coupled with its remarkable mechanical properties make this protein an excellent candidate as a material for biomedical applications. A facile and reproducible route towards the fabrication of novel materials, featuring organised and multicompartmental structure with high particle monodispersity can be achieved through microfluidics, and in particular, droplet-based approaches. In this thesis, microfluidic techniques are employed to generate microcapsules comprised of a protein-fibrillar network. Capsules exhibiting complex internal structure were fabricated, while droplets ranging from hundreds of nanometers up to hundreds of microns, were generated. Furthermore, different approaches as to how such microcapsules/microgels can be used for biomedical applications are presented through the generation of hierarchical emulsions, but also through the formation of Janus-like microgels, exhibiting enhanced release kinetic profiles. Furthermore, by integrating silver nanoparticles with silk-based microcapsules, I was able to systematically form organic/inorganic microgels, and investigate their properties both in vitro and in vivo. Not only did these hybrid microgels display potent antimicrobial properties, but in contrast to conventional treatments involving silver, the organic/inorganic microgels showed minimal cytotoxicity towards mammalian cells, making them ideal for wound healing purposes. To this effect, the antibacterial and wound healing properties of the hybrid microgels were investigated through the use of a murine model. It was determined that the efficacy of this system is comparable to results obtained when using a conventional antibiotic such as ampicillin, which clearly demonstrates the potential of the hybrid microgels for treatment of surgical site infections. The successful delivery of cargo molecules for cell related applications mostly depends on particle size and its distribution. Although traditional methods of generating nanoparticles have resulted in significant advances, some of which are currently used for pharmaceutical treatments, systematic control over size and monodispersity remains challenging. In order to address this issue, I developed a microfluidic/nanofluidic device capable of generating nanosized water-in-oil emulsions, with sizes ranging from 2500 down to 50 nm. By adding monomeric protein to the aqueous phase, these nanoemulsions acted as templates to form nanogels, which had the ability to permeate through mammalian cancer cell membranes and deliver intracellular cargo. Finally, in order to gain a better understanding of the processes involved in protein self-assembly, novel label-free approaches which utilise the fluorescence of the intrinsic amino acid, tryptophan, were investigated. It was found that during protein self-assembly, or more generally during protein phase transitions which involve hydrophobic burial, an increase in the tryptophan fluorescence signal was observed. This allowed for a systematic study of protein phase transitions, such as fibril, spherulite and crystal formation, without resorting to extrinsic fluorophores, paving the way for a label-free method to monitoring self-assembly events. In order to fully explore the potential of observing such events in a massively parallel way, I developed a microfluidic device capable of trapping thousands of individual droplets under zero-flow conditions. This device was used to spatially confine and temporally monitor biomolecular interactions within thousands of droplets simultaneously, and the statistical character of self-assembly could thus be highlighted.

A wide range of human disorders, including Parkinson's disease (PD), are associated with protein misfolding and aggregation. Misfolded a-synuclein is a major constituent of intracellular inclusions called Lewy bodies, which are a key hallmark of PD. A large body of work shows that a-synuclein can self-assemble into amyloid fibrils via a nucleation dependent pathway. However, the relationship between a-synuclein self-assembly and inclusion formation remains unclear. Recently, it has become apparent that many proteins can phase separate into dense droplets with liquid properties. Aberrant phase separation facilitates the formation of amyloid and intracellular inclusions. As such, we proposed that phase separation by a-synuclein into a liquid droplet state is a critical step in the formation of PD relevant inclusions. In this work, we show that a-synuclein undergoes liquid-liquid phase separation by forming a liquid droplet state, both in vitro and in a Caenorhabditis elegans model of PD, which converts into an amyloid-rich hydrogel. Importantly, this maturation process towards the amyloid state is modulated in the presence of model synaptic vesicles in vitro. As the maturation of a-synuclein droplets can be at least in some cases pathological, we investigated whether a-synuclein phase separation could be pharmacologically modulated. By establishing a robust screening platform, we identified modulators for the maturation of a-synuclein droplets into amyloid. To establish the generality of our observations for a-synuclein droplet formation, we explored phase separation as a proteome-wide concept and developed the FuzDrop method to predict droplet-promoting regions. This method is based on the concept that the droplet state is stabilised by the large conformational entropy associated with non-specific side-chain interactions. Our results indicate that the droplet state could be, at least transiently, accessible to most proteins under conditions found in the cellular environment.

The phenomenon of protein misfolding and aggregation has been associated with over 50 human diseases, including Parkinson's disease (PD). While the hallmark deposits in aggregation-associated diseases are primarily composed of fibrillar species, intermediate oligomeric species that form during the aggregation process are believed to be a major cause of toxicity. Such species are relatively poorly characterised due to several challenges that render them inaccessible to most conventional techniques; oligomers are only present at extremely low concentrations in the aggregation reaction, and are additionally highly heterogeneous and often transient in nature. In this thesis, I present complementary approaches to address these difficulties. Firstly, an ensemble of stable, kinetically trapped oligomers with varying biophysical characteristics was established, enabling detailed structure-toxicity relationships to be investigated. Secondly, a method for the simultaneous single-molecule level characterisation and fractionation of oligomeric species under native conditions was established and applied to studying intermediate species in Parkinson's disease-associated protein aggregation. Finally, I present a general approach to optimising experimental design, which maximises both the efficiency and information gain of experiments, and demonstrate its validation and application to several experimental systems.

Neurodegenerative diseases, including Alzheimer's and Parkinson's diseases, are characterised by the progressive, irreversible loss of neuronal tissue accompanied by protein aggregation in the brain. Although identified as a leading pathology, the role of protein aggregates in the pathogenesis remains elusive. They are a highly heterogeneous population spanning from the fibrous forms visible in histology to the nanoscopic oligomers, both characterised by various structural conformations. Many of these distinct proteoforms have been created in vitro allowing detailed biochemical studies. However, there are gaps between our in vitro understandings of aggregates and their in vivo behaviours. The slow progression of the diseases, low concentration of proteins and involvement of cellular machinery are in contrast with the idealised test tube conditions. In this work, super-resolution fluorescence imaging is firstly applied for the study of alpha-synuclein, the protein associated with Parkinson's disease. A new method is described combining single-molecule localisation microscopy and an alpha-synuclein aggregate specific aptamer to characterise surface immobilised aggregates. This method is validated on an in vitro alpha-synuclein aggregation time course, before being applied to the imaging of supernatants from a familial Parkinson's disease induced pluripotent stem cell model. The method determines the number density and size distribution of alpha-synuclein oligomers that are below the diffraction limit. Single-molecule localisation microscopy is also applied to characterise alpha-synuclein molecules in mouse brain tissue sections. The last chapter of this work studies the distribution of histological tau tangles in post mortem brain sections which are associated with Alzheimer's disease and several tauopathies. By using a digital pathology approach, cellular segmentation and pathological quantification are conducted with efficiency and reproducibility, allowing single-cell analysis into the distribution pattern of tau aggregates in the brain. Both the single-molecule and single-cell approaches aim to address the heterogeneity associated with aggregated proteins and provide tools for identifying the key molecular or cellular event in the pathogenesis.

A wide range of human disorders, including Alzheimer's and Parkinson's diseases, are associated with the process of misfolding and aggregation of proteins. Therefore, modulating the aggregation pathway of disease-related proteins, in particular the Amyloid-β peptide (A) in Alzheimer's disease and α-Synuclein (αSyn) in Parkinson's disease, is key to establishing therapeutics for these disorders. One promising method is the blockade of aggregation-prone regions of proteins using antibodies, but a major challenge in establishing antibodies against pre-selected epitopes is the effectiveness of the target regions to behave as antigens. To address this problem, a method of rational design was recently established in our group to construct complementary peptides in the complementarity determining regions (CDRs) of single-domain antibodies (sdAbs) towards the target regions. This was shown to bind disordered epitopes specifically and with good affinity. The binding was also shown to inhibit the aggregation of αSyn. This approach was also utilized to create multiple sdAbs binding different epitopes across the length of Aβ. It was found that systematically 'scanning' the sequence of Aβ in this way led to markedly different effects upon the Aβ aggregation pathway and the specificity of the sdAbs towards different aggregated species. To build on these advances and improve the designed sdAbs, in my work I combined this rational design method with established methods of protein evolution to improve binding affinities and specificity. The expectation was that this strategy would enable the establishment of a repertoire of high-affinity antibodies towards epitopes, which until recently, were deemed to be 'un-targetable'. Using a rationally designed αSyn antibody as a template for evolution, it was found that protein engineering via directed evolution and phage display is indeed capable of producing stable variants of the designed antibodies with improved affinity and crucially, retention of the pre-selected and poorlyimmunogenic epitope that it was originally designed to bind. Next, we aimed to alter the specificity of two of the rationally designed Aβ antibodies towards maximising the inhibition of secondary protein nucleation, in order to tune the specificity of the antibodies away from binding the monomer, and towards the toxic oligomeric species. Indeed, using directed evolution and phage display with competitive selection, it was found that it is possible to further alter the inhibition of the microscopic nucleation events that govern protein aggregation, albeit with compromised sdAb thermal stability. This was likely due to both the library design strategy and the heterogeneous selection employed. Nevertheless, these results show that combining rational design and directed evolution can offer novel opportunities in antibody discovery against targets not readily accessible through existing methods.

Saccharomyces cerevisiae is an established biotechnological workhorse for the production of various compounds including food products, biopharmaceuticals and biofuels. Previously, the Ashe lab has demonstrated that the biofuel n-butanol can be produced by overexpressing a clostridial butanol synthetic pathway in the yeast cytoplasm. However, the titre of n-butanol produced is below what would be viewed as industrially significant. Hence, the first objective of this project was to compartmentalise the butanol synthetic pathway into yeast peroxisomes in order to concentrate the enzymes to investigate whether this improved butanol titres. A yeast strain was constructed bearing a yeast endogenous gene ERG10 and four other Clostridium beijerinckii genes carrying a peroxisome targeting signal (PTS1). Although peroxisome morphology was visualised, there was little evidence that the butanol synthetic enzymes were localised to peroxisomes and little evidence for improved butanol production. However, the localisation studies suggested that the butanol synthetic enzymes may be present in protein aggregates. In order to test this, the solubility of each butanol synthetic enzyme was biochemically quantified to reveal that all of the enzymes involved in the butanol synthesis pathway were present in insoluble aggregates, regardless of the presence of PTS1. As a result, for one of the butanol production enzymes, AdhE2, strategies such as chaperone overexpression, altered growth temperature and analysis of different protein domains were undertaken with the aim of reducing protein aggregation. A small reduction in the level of insoluble protein was observed for AdhE2 at lower growth temperatures, yet these lower temperatures did not improve butanol levels from the butanol production strains. Interestingly, in some of these butanol production strains, 2,3-butanediol (2,3-BD) accumulates to high levels. 2,3-BD is an important high value bulk chemical with numerous industrial applications. Indeed, the presence of the clostridial bifunctional aldehyde-alcohol dehydrogenases, AdhE2, leads to significantly higher levels of 2,3-BD. In particular, the alcohol dehydrogenase domain plays a key role in 2,3-BD accumulation. On the whole, the project has applied several metabolic engineering strategies to improve bioalcohols production in yeast. However, further optimisation would be required to explore and understand the dynamic system of cell metabolism in yeast especially in the context of highly expressed exogenous proteins and their effect on cell homeostasis.

The behaviour and function of biomolecules represent a fundamental aspect in modulating the activity of micro and macro-scale complexes evolved in cells and tissues. The network of interactions of such biomolecules allow for the formation and regulation of the basic machinery of life, yet is commonly studied under non-physiological conditions. In order to characterise the behaviour and function of such biomolecules, they have to be analysed under relevant conditions, ideally in biofluids, cells or artificial systems which mainly imitate these properties. Recent microfluidic applications present an orthogonal approach for determining the interactions between a wide range of biomolecules, thus allow the study of molecular binding in the condensed phase with no need for extensive dilution, sample immobilisation or changes to the molecular environment from the liquid to the gas phase. As part of my PhD, I capitalised on microfluidic diffusion approaches, developed in the Knowles lab to systematically study the binding and thermodynamics of small heat shock proteins, such as clusterin, αB-crystallin and the Brichos chaperone domain, to aggregated forms of amyloid-beta and α-synuclein, protein aggregates that are associated with a wide range of neurodegenerative diseases. The three chaperones are crucial components of the cellular proteostasis network and characteristically overexpressed during cell stress. Each chaperone type shows distinct binding behaviour to protein aggregates, which can be related to its inhibitory function. While αB-crystallin binding to α-synuclein is entropically driven by conformational rearrangement, the binding of Brichos to amyloid-beta fibrils is shown to be enthalpically driven as it inhibits specifically secondary nucleation processes. In contrast to the specific secondary nucleation inhibition by Brichos, clusterin inhibits specifically fibril elongation of amyloid-beta. I could show that these two specific aggregation processes are affected by the two chaperones, Brichos and clusterin, in a non-cooperative manner. These molecular details are particularly relevant in the context of the rational design of drug molecules that could, potentially in combination, target multiple specific aggregation steps in a selective manner. Therefore, I further screened the binding of a wide range of monoclonal antibodies to either amyloid-beta monomers or fibrils, which are currently at different stages of clinical phase trials for Alzheimer's therapy. I thus show that the obtained stoichiometry and affinity information of the drug correlates with the distinct inhibition mechanisms and consequently provides mechanistic and structural information. In contrast to studying disease related model systems in vitro under homogeneous conditions, measurements in complex body fluids are key in medical applied science, e.g. cancer treatment or immunological characterisation. In my research, I have undertaken the challenge of extending the platform developed above to characterise the binding of a wide range of molecules under complex solution conditions. Preliminary data obtained during my PhD underlines the extraordinary capability of the diffusion-based microfluidics to being applicable for investigating the binding parameters of molecules involved in alloimmunisation in human serum. Along with my main focus on measuring protein interactions with diffusion-based microfluidics, I further developed a technique using selective separation properties, such as particle charge, hydrophobicity, size or immunoaffinity and coupled it with a series of microfluidic devices for an instantaneous and full biophysical characterisation of heterogeneous solutions. This new technique can be used to explore the formation of protein oligomers or protein complexation, characterisation and identification of complex mixtures in the context of amyloid formation and protein homeostasis.

The self-assembly of soluble proteins is a common phenomenon observed in nature. It ranges from the formation of functional biomaterials, such as actin filaments and tubulin microtubules that form the cytoeskeleton, to the abnormal deposition of aberrant aggregates that lead to disease. There are nearly 50 diseases currently associated with protein misfolding and amyloid formation, among them several neurodegenerative disorders such as Alzheimer's, Parkinson's and prion diseases. Recent evidence indicates that the most toxic species are the low molecular weight oligomers formed during the onset of the aggregation process. These oligomeric species are commonly difficult to analyse due to their low abundance, transient nature and heterogeneity in their structure. In fact, prions, known to be associated with transmissible spongiform encephalopathies, have not yet been fully characterised as parts of the prion composition seem to lose infectivity through various in vitro purification and isolation procedures. In nature, non-homogeneous solution conditions and molecular crowding are key factors as there might be multiple species associated with the disease. However, traditional biophysical techniques allow the study of biomolecules under ideal conditions, such as high concentrations and monodisperse size distribution. Furthermore, among the limitations of the traditional methods are the sensitivity as well as the sample preparation, which in some cases may require to chemically modify the molecule. In this thesis, I describe the development of a sensitive label-free method to enable the study of protein complexes and proteins under biologically relevant environment, such as brain homogenates, by combining immunochemistry with microfluidic strategies. In this thesis I describe the development of Immuno Diffusional Sizing (IDS), a method that combines diffusional sizing with the sensitivity and specificity of immunoassay detection. In Chapter 3, I explain the optimisation of the devices developed to size a range of particle sizes in heterogeneous mixtures, from small protein monomers to large aggregates. IDS is a label-free method capable of sizing proteins down to the nano- and picomolar concentrations, in complex solutions. The applications of this method to monomeric and aggregated forms of disease-related proteins are described in Chapter 4. IDS has been applied to study the Prion Protein (PrP) in its native and infectious forms. Furthermore, I studied of the proteinase resistant form (PrPres) extracted from the brains of infected mice specimens. IDS was applied for studying disease-relevant species, such as intrinsically disordered proteins, e.g. α-Synuclein (aSyn) and Aβ40, and their respective oligomers. Another advantage of this method is its applicability for studies of brain aSyn from transgenic overexpressing mice (OVX). Furthermore, I explore the potential of IDS for the study of post-translational modifications (PTMs) of specific proteins, such as the ubiquitination of aSyn in vitro. The efforts performed during this work to achieve TR-FRET detection on-chip are described in Chapter 5. A comparison of two commercial immunoassays was performed in order to determine the suitability for its adaptation from a platereader to a microfluidic device. The design and adaptation of an optics platform for FRET detection on-chip is also discussed. My research on the mechanisms of the molecular chaperone αB-Crystallin (aBC) is described in Chapter 6. I investigated the aggregation process of insulin in the presence of aBC. I have used a combination of bulk techniques to study the kinetics of the systems, as well as microfluidics to generate microdroplets that allow the direct observation of rare nucleation events leading to protein aggregation. Overall, in this work I describe the development of a novel microfluidic technique, its optimisation and application to multiple systems related to protein misfolding diseases, including prions and intrinsically disordered proteins. Purification of the sample is not necessary to study the target protein because the detection of the target protein after diffusional separation is made with a sandwich immunoassay, which allows to study disease-related proteins under native conditions in a label-free manner, which is usually difficult with traditional sizing techniques. The applications of this technology could be expanded to characterise other antigens in body fluids, like cerebrospinal fluid (CSF) of blood serum.

The formation of macroscopic biofilaments from monomeric peptides is a process central to both normal and disease biology, and a very active area of research. An example of particular significance is the formation of amyloid plaques from normally soluble proteins in several neurodegenerative conditions including Alzheimer's disease. Traditionally, attention has focused on the proliferation of macroscopic quantities of biofilaments, and experimentalists have relied on bulk measurements to quantify this. By contrast, the initiation of the aggregation reaction has been comparatively understudied, despite the involvement of oligomeric intermediates that are considered to be the key pathogenic agents in many amyloidogenic diseases. This is due in large part to the key quantities measured in bulk experiments being insensitive to the details of the initiation step, and to the difficulties faced until recently in detecting comparatively rare oligomeric species. In this thesis I describe a number of theoretical approaches towards filling this crucial gap in our knowledge. These approaches fall into 3 main themes: modelling the early stages of bulk aggregation kinetics in greater detail; developing general kinetic theories of filament formation via initial oligomeric intermediates; and investigating heterogeneity within oligomer populations using equilibrium statistical mechanical modelling. In several instances I have already been able to apply these theories to experimental results, furthering our practical understanding of these processes in vitro. A final chapter is devoted to a highly general and insightful theoretical method for developing analytical solutions for the kinetics of a wide range of self-assembly phenomena. Although not focused on early events in amyloid aggregation, it is already finding application in this field, as in many others.

The overall aim of this PhD project was to use chemical denaturants such as urea and guanidine hydrochloride (Gdn HCl) to increase the relative population of intermediate and/or unfolded states of proteins as known precursors to protein aggregation. Measurements of protein-protein interactions (PPIs) under denaturing conditions of both the model protein lysozyme and a set of five pharmaceutically relevant monoclonal antibodies (mAbs) were made. Light scattering techniques were employed including static light scattering (SLS) to determine the second osmotic viral coefficient, B22 and dynamic light scattering (DLS) to determine the interaction parameter, kD. SLS and DLS was also used to measure aggregate growth rates under denaturant and/or temperature accelerated conditions and were ratified by measuring monomer loss kinetics using size exclusion chromatography couple with multi-angle laser light scattering (SEC-MALLS). The data generated in the first results chapter highlighted the importance of measuring the refractive index increment of the protein under constant solvent chemical potential, to ensure correct values of B22 and molecular weight at infinite protein dilution are obtained as well as excluding short delay times of an autocorrelation decay in solutions of high co-solvent concentrations to similarly calculate correct values of kD and hydrodynamic radius at infinite protein dilution. The data generated in the second and third results chapter showed that PPIs of intermediate, aggregate prone states could be quantified using SLS at high protein concentrations (20 g/L) and were related to the aggregation propensity of the mAbs under both native and denaturing conditions. Furthermore, addition of 0.5 M arginine HCl under denaturing conditions caused a reduction in aggregation rates and concordantly an increase in repulsive PPIs.

In this thesis, we investigate the mechanisms driving the self-assembly of peptides and proteins using computational and theoretical tools, always validating our results with experimental measures when possible. In the first part, Chapters 2-5, we focus on the Aβ system, a peptide whose aggregation is intimately linked with the development of Alzheimer's Disease. We begin by simulating the major alloforms of the peptide, Aβ_40 and Aβ_42, demonstrating that the two populate similar disordered ensembles and matching experimental data. Next we investigate how disordered Aβ_42 monomers interact with each other, finding that oligomerisation into amorphous aggregates is driven largely by hydrophobic, non-specific forces. We then move on to probing the aggregation of Aβ_42 into amyloid structures using a native-centric coarse-grained model, and explain the results with a novel Markov state analysis from which we are able to extract structural, kinetic and thermodynamic information on elongation reactions. Finally, we probe the interactions of Aβ_42 monomers with Aβ_42 fibrillar surfaces using a specially designed enhanced sampling scheme, which allows us to obtain enthalpy-driven binding thermodynamics consistent with experiments and to propose major polar binding modes. In the second part of the thesis, Chapters 6 and 7, we model the aggregation of two other self-assembling systems, viruses and a truncated form of the molecular chaperone Hsp70. We first develop a data analysis platform to extract information on the microscopic mechanisms of viral capsid self-assembly from experimental data, synthesising the results from several different systems to draw general evolutionary conclusions about the assembly mechanism. Finally, we model the oligomerisation of Hsp70 thermodynamically and kinetically, showing that its self-assembly is a highly cooperative reaction that is under strong structural constraints.

Upon ageing, a progressive disruption of protein homeostasis often leads to extensive protein aggregation and neurodegeneration. It is therefore important to study at the proteome level the origins and consequences of such disruption, which so far have remained elusive. Addressing this problem has recently become possible by major advances in mass spectrometry-based (MS) proteomics, which allows the identifications and quantification of thousands of proteins in a variety of biological samples. In the first part of this thesis, I analyse proteome-wide MS data for the nematode worm C. elegans upon ageing, in wild type (WT), long-lived and short-lived mutant strains. By comparing the total abundance and the soluble abundance for nearly 4000 proteins, I provide extensive evidence that proteins are expressed in adult worms at levels close to their solubility limits. With the use of sequence-based prediction tools, I then identify specific physico-chemical properties associated with this age-related protein homeostasis impairment. The results that I obtained reveal that the total intracellular protein content remains constant, in spite of the fact that the proteome undergoes wide remodeling upon ageing, resulting into severe protein homeostasis disruption and widespread protein aggregation. These results suggest a protein-dependent decrease in solubility associated with the protein homeostasis failure. In the second part of the thesis, I determine and classify potential interactions of misfolded protein oligomers with other proteins. This phenomenon is widely believed to give rise to cytotoxicity, although the mechanisms by which this happens are not fully understood. To address this question, I process and analyse MS data from structurally different oligomers (toxic type A and nontoxic type B) of the protein HypF-N, incubated in vitro with proteins extracted from murine cell cultures. I find that more than 2500 proteins are pulled down with the misfolded oligomers. These results indicate that the two types of oligomers interact with the same pool of proteins and differ only in the degree of binding. Functional annotation analysis on the groups reveals a preference of the oligomers to bind proteins in specific biological pathways and categories, including in particular mitochondrial membrane proteins, RNA-binding proteins and molecular chaperones. Overall, in this study I complement the powerful and high-throughput experimental approach of MS proteomics with bioinformatics analyses and prediction algorithms to define the physical, chemical and biological features of protein homeostasis disruption upon ageing and the interactome of misfolded oligomers.

Protein deposition in the form of amyloid fibrils is the hallmark of more than 40 human pathologies, including Alzheimer's disease (AD) and Parkinson's disease (PD). Misfolded protein oligomers formed as intermediates during the aggregation process have been strongly implicated in the onset and progression of these diseases. In this thesis, I describe our efforts to uncover molecular agents that can reduce the toxicity caused by protein aggregation via targeting the generation, the physiochemical properties or the membrane affinity of oligomeric species. We employed an integrative approach combining in vitro techniques, including chemical kinetics, atomic force microscopy, and biophysical measurements, and in vivo methods, including neuroblastoma cells and C. elegans models of AD and PD, to identify a range of small molecules and antibodies that can suppress the toxicity related to protein aggregation through a variety of mechanisms. In Chapter 3, we show that the deleterious effects of protein aggregation can be suppressed in AD and PD worms by interfering with the aggregation rates of the amyloid-β peptide (Aβ) and the α-synuclein protein (αS). In Chapter 4, we resolve the mechanism of action for a molecule that enhances the rate of Aβ42 aggregation in AD worms with the result that toxicity is reduced, and find that it potentiates the secondary nucleation microscopic step in vitro. In Chapter 5, we characterize molecules and antibodies that modify the physiochemical properties and self-association of oligomers comprised of several proteins into clusters with reduced diffusibility. In Chapter 6, we classify a family of molecules that protect the cell by displacing several types of oligomeric species from the membrane through a generic mechanism. These results demonstrate strategies by which one can target the aggregation process to alter its resulting toxicity, provide insight into modifying the properties of the most deleterious species associated with protein aggregation and suggest that the protection of the cell from the oligomer-induced cytotoxicity associated with numerous protein misfolding diseases is a promising strategy to combat protein misfolding diseases.

The misfolding and aggregation of proteins is closely associated with more than fifty human disorders, including Alzheimer's and Parkinson's diseases, all of which are currently incurable and many represent a major threat to human life. The mechanism of protein aggregation is subject to extensive studies. The damaging effects associated with protein aggregation have been attributed to amyloidogenic species that are present during the misfolding process. In particular, oligomeric species are, however, intrinsically difficult to study as a consequence of their low abundance and highly heterogeneous nature. The first chapter of my thesis gives an introduction into the field of protein folding and misfolding with a focus on the study of protein aggregation, and toxic effects relevant to human disorders. The second chapter of my thesis describes the development of a methodology that enables the study of aggregate induced lipid bilayer permeability, possibly the most general mechanism of protein aggregate toxicity. Surface-tethered lipid vesicles functioning as optochemical probes sensitive to membrane integrity are imaged using total internal reflection microscopy. It is shown that oligomeric species of the 42-residue form of the Aβ peptide (Aβ42) are responsible for the membrane disruption. The methodology can be applied to the study of other proteins such as α-synuclein and tau, and the ability of antibodies and chaperones to counteract the aggregate induced lipid bilayer permeability can be assessed. Furthermore, lipid bilayer permeability induced by aggregates formed in human induced pluripotent stem cells can be studied. The third chapter presents a new approach for the measurement of protein aggregation kinetics by following the development of the lipid bilayer permeability over the course of the aggregation process of Aβ42. The aggregation kinetics can be modulated with molecular chaperones and pre-formed seed fibrils, which allows secondary nucleation to be identified as the process that drives the formation of species responsible for the lipid bilayer permeability. The fourth chapter describes the development of a three-pronged strategy to study the mechanism of α-synuclein amyloid formation. The aggregation is studied in the presence of lipid vesicles or pre-formed fibrils at neutral or acidic pH of the solution. The influence of single-point mutations on the aggregation of α-synuclein is described. Furthermore, the strategy is applied to the characterisation of the ability of antibodies and small molecules to inhibit the aggregation, and thus has the potential for the development of therapeutical agents. The work presented in the fifth chapter characterises the amyloid fibril populations formed by α-synuclein and mutational variants associated with familial Parkinson's disease. X-ray crystallography, circular dichroism spectroscopy, Fourier transform infrared spectroscopy, transmission electron microscopy and atomic force microscopy have all been applied to the analysis of these amyloid fibrils. Finally, the sixth chapter summarises the results described in this thesis and points out future opportunities in the context of fundamental and translational studies related to the research area of protein misfolding disorders.

Amyloid $\beta$ (1-42) (A$\beta$42) is a seminal neuropathic agent in Alzheimer’s disease (AD), a multifaceted neurodegenerative disorder for which no preventative measures or disease modifying therapies currently exist. Aggregation of this peptide plays a key role in the synaptic dysfunction and neuronal death associated with the disease. Perturbing the aggregation process, therefore, represents a key strategy for the development of new AD therapeutics. A variety of issues with current screening methods, including lack of reproducibility, high reagent consumption and spectral interference from the test molecules, can limit efforts to identify new small molecule inhibitors. Furthermore, the lack of robust, time- and cost-efficient methods for screening compounds in cellular or in vivo models limits the throughput with which seemingly active small molecules can be validated and prioritised. Herein, this thesis describes efforts to overcome such limitations through the development of a unified in vitro to in vivo assay system, in which hits identified in the ‘nanoFLIM’ microfluidic-based assay can quickly be tested in cellular and whole organism disease models. The assay platform designed relies on the use of an amyloid aggregation fluorescence lifetime sensor. A$\beta$42 aggregation is monitored by changes in the fluorescence lifetime of an attached fluorophore, which is significantly quenched upon amyloid formation. To take advantage of the benefits associated with miniaturisation, an in vitro microfluidic platform was employed. A microfluidic chip capable of trapping 110 precisely ordered droplets was designed, allowing for increased sample size and greatly lowering reagent consumption relative to conventional assay formats. Optimisation of the lifetime sensor technique permitted real-time compound screening in SH-SY5Y neuroblastoma cells, as well as in disease model Caenorhabditis elegans (C. elegans). To demonstrate the potential of this assay, a selection of novel chemical libraries developed in the Spring research group was screened, resulting in the identification of a key library of interest. The inhibitory activity of the lead compound from this collection was validated using a variety of biophysical tests, and was also shown to suppress amyloid aggregation in the live cell fluorescence lifetime sensor assay, as well as in whole organism disease model C. elegans. Whilst assay development was underway, additional screening of structurally diverse chemical libraries was performed using a conventional Thioflavin T spectroscopic assay. Such work identified another molecular scaffold capable of exerting a strong inhibitory effect against A$\beta$42 aggregation. A selection of analogues was synthesised to improve the in vivo profile of this library, giving rise to a second lead inhibitory compound. The activity of this compound was subsequently validated in biophysical and cellular tests, and was also tested in disease model Drosophila melanogaster. The aggregation of A$\beta$42 lies at the root of Alzheimer’s disease. In light of the relatively few drug candidates in clinical trials for this disorder, the development of improved translational screening approaches and continued screening of novel chemical libraries is necessary to identify new potential therapeutics. In this study, an in vitro to in vivo fluorescence lifetime imaging assay has been established. Using this assay system and conventional screening approaches, two A$\beta$42 aggregation inhibitors have been identified and validated. These represent promising candidates for the development of new AD therapeutic agents, or for use as molecular probes to further dissect the mechanisms underlying this devastating disease.

The aggregation of alpha-synuclein (αS) protein from soluble monomer into solid amyloid fibrils in the brain is associated with a range of devastating neurodegenerative disorders such as Parkinson’s disease. Soluble oligomers formed during the aggregation process are highly neurotoxic and are thought to play a key role in the onset and spreading of disease. Despite their importance, these species are difficult to study by conventional experimental approaches owing to their transient nature, heterogeneity, low abundance and a remarkable sensitivity of the oligomerisation process to the chosen experimental conditions. In this thesis, well-established single-molecule techniques have been utilised to study the aggregation and oligomerisation of αS in solution.

Monoclonal antibodies (mAbs) represent one of the fastest growing classes of therapeutic proteins. This success is due to a number of attractive properties such as high binding affinity, specificity, low immunogenicity and high aqueous solubility. Despite this, mAbs can suffer from undesirable physical instabilities, especially reversible self-association (RSA), which can lead to aggregation and phase separation. One aspect of formulation is therefore to find solution conditions which minimise mAb aggregation propensity during storage at high concentrations. Hence, the buffer, excipient and pH must be carefully considered to obtain the optimal formulation. Currently, if a platform formulation process is non-ideal for a particular candidate mAb, then an alternative strategy is to utilise high-throughput screening to measure various physical parameters indicative of physical stability. Arginine (in the form of hydrochloride salt Arg·HCl) is often used in formulations exhibiting high RSA and a propensity for aggregation. The interaction of Arg with the protein surface is complex and dependent on both the salt form and concentration. Here the focus was on the glutamate salt of arginine (Arg·Glu), to quantify its effect on mAb conformational and colloidal stability under different pH conditions. Arg·Glu was able to decrease the propensity of the mAbs to aggregate, particularly at pH values closer to their pI.The work also included the use of in vitro cell culture models to examine cell viability in the presence of the various arginine salts over a range of osmolalities. Whilst Arg·Glu is composed of two naturally occurring amino acids and both of which are considered non-toxic individually, the effect of the increased concentrations of their combination, on cells has not been explored previously. In vitro cell lines were chosen to represent the subcutaneous tissue, the effect of Arg·Glu on cell viability was compared against NaCl, Arg·HCl and sodium glutamate (NaGlu). The work concluded there was no additional toxicity associated with the presence of Arg·Glu in the cell culture models studied, therefore Arg·Glu has the potential as an excipient as it reduces aggregation and is nontoxic. Another aspect of the work was to assess the use of solution NMR spectroscopy as an orthogonal technique in mAb formulation characterisation. 1H NMR spectroscopy was used to measure a number of experimental parameters for high concentration mAb solution. The work proposed that 1H NMR spectroscopy can serve as a valuable orthogonal method for mAb characterization and formulation.

Peroxiredoxins are ubiquitous, thiol-specific proteins that have multiple functions in stress protection, including oxidative stress. Tsa1 is the major yeast peroxiredoxin and we show that it functions as a specific antioxidant to protect against oxidative stress caused by nascent protein misfolding and aggregation. Yeast mutants lacking TSA1 are sensitive to misfolding caused by exposure to the proline analogue azetidine-2-carboxylic acid (AZC). AZC promotes protein aggregation and its toxicity to a tsa1 mutant is caused by reactive oxygen species (ROS). Generation of [rho0] cells lacking mitochondrial DNA rescues the tsa1 mutant AZC sensitivity indicating that mitochondria are the source of ROS. Inhibition of nascent protein synthesis with cycloheximide prevents AZC-induced protein aggregation and abrogates ROS generation confirming that aggregate formation causes ROS production. Protein aggregation is accompanied by mitochondrial fragmentation and we show that Tsa1 localizes to the sites of protein aggregation, which are formed adjacent to mitochondria. Further investigation reveals that AZC-induced protein aggregation leads to an inhibition of mitochondrial respiration and the depolarisation of the mitochondrial membrane. Remarkably, this was entirely dependent on the presence of Tsa1. We show that the effects of protein aggregation on mitochondrial function are mediated by the Ras/PKA pathway and that Tsa1 appears to influence the activity of this pathway through its effects on the yeast phosphodiesterase, Pde2. Together, these data indicate a new role for peroxiredoxins in the response to ROS, generated as a result of protein misfolding and aggregate formation. Finally, we analysed the characteristics of proteins found within protein aggregates, isolated from different conditions during the course of the study. Our results highlight the differences between proteins that aggregate under normal, mid-exponential growth conditions (physiological aggregates) and those which aggregate during cellular stress. We were able to establish the characteristics of an archetypical physiological aggregate, through an assessment of a range of properties, identifying factors that significantly differed from genomic expectations. Furthermore, our observations indicate that, in general, cellular stress reduces the threshold of metrics associated with protein aggregation propensity. We also found that different stresses result in the aggregation of proteins that are, largely, physicochemically indistinct from one another, regardless of the mode of toxicity. Finally we show that a significant number of proteins, identified in our protein aggregates, were also present in protein aggregates isolated from aged C. elegans. This suggests that the factors and components of protein aggregates are conserved.

Protein aggregation is a frequently cited problem during the development of liquid protein formulations, which is especially problematic since each protein exhibits different aggregation behaviour. Aggregation can be controlled by judicious choice of solution conditions, such as salt and buffer type and concentration, pH, and small molecule additives. However, finding conditions is still a trial and error process. In order to improve formulation development, a fundamental understanding of how excipients impact upon protein aggregation would significantly contribute to the development of stable protein therapeutics. The underlying mechanisms that control effects of excipients on protein behaviour are poorly understood. This dissertation is directed at understanding how excipients alter the conformational and colloidal stability of proteins and the link to aggregation. This knowledge can be used for finding novel ways of either predicting or preventing/inhibiting protein aggregation. Experiments using static and dynamic light scattering, intrinsic fluorescence, turbidity and electrophoretic light scattering were conducted to study the effect of solution conditions such as pH, salt type and concentration on protein aggregation behaviour for three model systems: lysozyme, insulin and a monoclonal antibody. Emphasis is placed on understanding the effects of solution additives on protein-protein interactions and the link to aggregation. This understanding has allowed the rational development of stable formulations with novel additives, such as arginine containing dipeptides and polycations.

Protein solubility plays an important role whether for biophysical and structural studies, or for production and delivery of therapeutic proteins. Poor solubility could lead to protein aggregation, which is an undesired physicochemical mechanism at any stage of recombinant proteins production. To date, more than half of all recombinant therapeutic proteins are produced in mammalian cells, mainly due to the high similarity of the final product to human protein structures. However, poor secretion can occur, due to misfolded proteins or aggregates leading to cellular stress and proteolysis. Another widely-used expression system is E. coli, which can offer a cost-efficient alternative. This system has an important limitation, since proteins tends to form insoluble protein aggregates in the cytoplasm upon heterologous overexpression. Several strategies are being implemented to improved soluble expression, ranging from culture conditions to solubility enhancing tags. However, there is no universal approach or technology that solves protein aggregation. In this thesis two recently published hypotheses from our group have been applied. One stated that soluble expression of proteins was inversely correlated with the size of the largest positively-charged patch on the protein surface. The second hypothesis (of protein solubility), arose from the finding that the relative content of lysine and arginine residues separated E. coli proteins by solubility. Both hypotheses arose from a study of an extensive dataset of experimental solubilities determined for cell-free expression of E. coli proteins. In combination with other widely used strategies, such as lowering expression temperature and inducer concentration, decreasing non-charged (hydrophobic) patches and addition of helical capping for increasing stability, a rational understanding for directed alteration of solubility in a variety of recombinant proteins has been explored. This includes three protein models to test: (i) recombinant human erythropoietin (rHuEPO) (one of the top selling therapeutics) (ii) recombinant 6-Phosphofructo-2-Kinase/fructose-2,6-bisphosphatase (rPFKFB3) (a product for which over-expression has been sought for characterisation and insight into possible cancer therapy) and (iii) a set of three selected E. coli proteins containing high ratios of lysines to arginines: thioredoxin-1 (TRX), cold shock-like protein cspB (cspB), and the histidine-containing phosphocarrier protein (HPr). It was found that single or multiple point mutations (changing amino acids from positive to negative charge or vice versa; or lysines to arginines) verified the predicted effect on rHuEPO, rPFKFB3, TRX, cspB, and HPr solubility (experimentally defined as the distribution between soluble and total fractions) for expression in E. coli. In addition, the redesigned set of rHuEPO transiently expressed in HEK 293-EBNA cells, suggesting that positively-charged patch size may also influence protein secretion. Further application of these computational and experimental approaches could provide a valuable tool in the design and engineering of proteins, with enhanced solubility, stability and secretion.

The aim of this thesis is to investigate and better understand the mechanisms of protein self-assembly. Specifically, I study three protein systems which form morphologically and structurally distinct brillar protein aggregates. The first of these studies is concerned with the self-assembly of amyloid brils formed from bovine insulin. Amyloid brils are associated with human diseases such as Alzheimers Disease and type-2 diabetes, and are also garnering interest in biomaterial applications. Fragmentation-dominated models for the self-assembly of amyloid brils have had important successes in explaining the kinetics of amyloid bril formation but predict bril length distributions that do not match experimental observations. Here I resolve this inconsistency using a combination of experimental kinetic measurements and computer simulations. I provide evidence for a structural transition demarcated by a critical bril mass concentration, or CFC, above which fragmentation of the brils is suppressed. Our simulations predict the formation of distinct bril length distributions above and below the CFC, which I confirm by electron microscopy. These results point to a new picture of amyloid bril growth in which structural transitions that occur during self-assembly have strong effects on the final population of aggregate species with small, and potentially cytotoxic, oligomers dominating for long periods of time at protein concentrations below the CFC. I further show that the CFC can be modulated by environmental conditions, pointing to possible in vivo strategies for controlling cytotoxicity. I probe the structural nature of the transition by performing small angle neutron scattering. Secondly, I study the formation of amyloid-like brils from the protein ovalbumin. I undertake kinetic experiments of self-assembly and find two key features emerge: the lack of a lag time and the existence of a slow growth regime in the long-time limit. I observe, using TEM, that these brils are worm-like in nature and form closed-loops. I find the growth kinetics are intimately connected to this particular morphology. I present a simple kinetic model which captures the features of the kinetics found in experiments by incorporating end-to-end association of brils. I comment on the ramifications this type of amyloid bril assembly may have on oligomeric toxicity. Thirdly, the DNA-mimic protein ocr is highly charged (-56e at pH 8) and forms non-amyloid brillar assemblies at very high ammonium sulphate concentrations (3.2M). The fact that ocr forms translucent brillar gels at such high salt concentrations is extremely unique. Typically under such high salt conditions, non-specific amorphous aggregates are formed. In order to better understand the mechanism of why ocr forms specific bril aggregates, I used variants of the wile-type protein in which extensive regions of surface have been removed or modified. The structural characteristics of gels formed from the variants were probed using microrheological techniques. I find that non-specific electrostatic charge screening plays an important role in ocr aggregation. However, I also locate a potentially important α-helical region which may play a part in establishing specific interactions so that ocr may form ordered brillar assemblies.

In the past forty years advancements in fluorescence-based methods including imaging (e.g. confocal and multi-photon) and quantitative spectroscopies (e.g. Fluorescence Correlation Spectroscopy) have been applied to systems ranging from solutions to in vivo models: such methods possess the ability to monitor fluorescence intensity fluctuations and offer the potential to unravel biophysical and biochemical phenomena. A major disadvantage associated with these methods is their ever-increasing cost resulting in the development of image analysis tools that offer the potential to exploit hidden information contained in confocal images.The hypothesis pertaining to this thesis is that image analysis tools developed in recent years exemplified by Raster Image Correlation Spectroscopy (RICS), Spatial Intensity Distribution Analysis (SpIDA) and Fluorescence Intensity Gaussian Mixture Model Analysis (FIGMMA) will provide a new insight into current pharmaceutical problems. The application of these methods to the quantification of protein aggregation, monomer/dimer equilibria, p-glycoprotein efflux activity and transcytosis are presented in this thesis.Protein aggregation poses a major challenge to the biotechnology industry which currently lacks analytical capabilities to profile broad particle size ranges. An in-house RICS (ManICS) software was validated against Dynamic Light Scattering and Fluorescence Correlation Spectroscopy (FCS) to determine Bovine Serum Albumin (BSA) aggregate population distributions under accelerated stability conditions. Initial stages implicated in the growth of aggregates are vital to the mechanistic assessment of protein aggregation. Hence, real-time in situ examination of monomer loss and aggregation of BSA was performed at 50 °C to enable continuous assessment with imaging and subsequent SpIDA analysis. Results obtained from this study suggested reversible fluctuation between monomers and dimers for up to four hours.To correlate membrane receptor and transporter expression with activity and enable the comparison of expression in multiple cell lines, population densities of p-glycoprotein transporters and transferrin receptors were determined using SpIDA in samples subjected to immunofluorescence labelling.The Calcein retention assay is a routine approach to determining multidrug resistance associated with p-glcoprotein efflux and the traditional plate reader approach omits microscopic aspects of p-glycoprotein Calcein-AM uptake and efflux. Confocal microscopy and data obtained from image analyses supported the subcellular and intercellular assessment of Calcein accumulation in MDR1-transfected and control cell lines as a function of time and verapamil concentration. Finally, live cell imaging of transferrin vesicular transport and Cell TraceTM Calcein red-orange AM internalisation in combination with traditional Transwell® assays were assessed to compare their transcellular transport and intracellular concentrations in multiple cell lines. Images obtained enabled visualisation of internalisation and following analysis using SpIDA, RICS and FIGMMA the number of intracellular vesicles and dynamic parameters of Cell TraceTM Calcein red-orange diffusion and intracellular concentration were determined.In conclusion, image analysis tools were applied to providing new parametric insights into a number of pharmaceutically-relevant processes and in some instances this is the first example of such studies. Despite current phenomenal advances in image acquisition capabilities, there remains a broad scope for the validation of image analysis tools and their application to a multitude of areas of interest to pharmaceutical and biomolecular research.