Biochemistry and Cell Biology - Doctoral Theses
Permanent URI for this collection
Browse
Recent Submissions
Item An investigation into the molecular genetics of Major Depressive Disorder(University College Cork, 2025) Bogue, Roisin; McCarthy, Tommie V.; School of Biochemistry and Cell BiologyMajor Depressive Disorder (MDD) is a debilitating psychiatric condition with a complex aetiology involving both genetic and environmental factors. Despite its high prevalence and significant impact on global health, the precise biological underpinnings of MDD remain poorly understood, hindering the development of effective diagnostic and therapeutic strategies. This thesis addresses this critical gap by focusing on two main objectives: the identification of causative genes through linkage analysis and rare variant analysis in a human pedigree and the identification and validation of biomarkers for MDD using a rat maternal separation model. The first part of the thesis focused on a unique pedigree dataset characterised by a high incidence of mood disorders, particularly MDD, with unusually severe symptoms in a high number of affected individuals. The dataset, collected over a >30-year period, includes data from multiple sources, such as SNP arrays, exon arrays, whole exome sequencing (WES), and whole genome sequencing (WGS). Due to the length of the collection period, the data reflects a range of technological advancements in sequencing approaches. For linkage analysis these diverse datasets were combined, and the samples were stratified based on diagnostic severity criteria. Linkage analysis identified two key regions, a locus on chromosome 12q24, associated with a severe phenotype, and on chromosome 4q15, associated with a moderate phenotype, both of which co-segregate with a large number of affected individuals within the pedigree. Additionally, analysis of rare variants using WGS data, which was available for a subset of the pedigree, identified a rare variant in the STIM2 gene with an allele frequency of 0.00001002 and present in 15 affected members of the pedigree. This variation replaces an arginine (R) with a tryptophan (W) at amino acid position 25 in the STIM2 protein. Prediction software used to assess this R25W variant suggests that it is likely to have an impact on the function of the STIM2 protein. Additionally, Eukaryotic Linear Motif (ELM) analysis revealed that the mutation potentially disrupts a putative endoplasmic reticulum (ER) retention/retrieval motif. Conservation analysis of the associated sequence indicated the regions surrounding the SNP is conserved in higher-order mammals, further supporting the potential significance of this variant. The second part of this thesis employed the well-established rat maternal separation model, for inducing depressive-like behaviours to investigate the feasibility of a novel method for biomarker discovery in complex disease. The method utilises plasma from different disease states to identify compositional differences which can be used to stratify samples. This method can theoretically be applied to blood derived samples from a number of conditions. Using this method maternally separated rats and non-separated controls clustered into three distinct groups which correlated highly with forced swim test (FST) immobility counts, an established measure of depressive-like behaviour, allowing the creation of a biomarker panel. This feasibility study suggests that the method described could have broad application across many complex diseases where biomarkers are currently lacking. Together, these two approaches, combining genetic analysis of a unique pedigree with a high incidence of MDD and biomarker discovery using a rat maternal separation model, provides a comprehensive exploration of the genetic and biological factors associated with MDD. The identification of two key genomic regions, the discovery of a rare variant in the STIM2 gene, and the development of a biomarker panel offer insights into the potential mechanisms of depression. While these findings deepen our understanding of MDD, they also underscore the complexity of the disorder, and prompt further research to validate these results and better understand their potential implications for diagnosis, disease progression and treatment.Item Applications of yeast display in synthetic biology with a focus on biological synthesis and recycling of plastics(University College Cork, 2025) Jurić, Vanja; Young, Paul; Holmes, Justin; Science Foundation Ireland; Advanced Materials and Bioengineering ResearchYeast surface display has emerged as a versatile tool in synthetic biology, particularly in protein engineering. This work sought to improve the flexibility and ease-of-use of yeast display for synthetic biologists, as well as applying the yeast display system in several contexts – with a particular focus on biological synthesis and recycling of plastics. The yeast modular cloning toolkit (MoClo-YTK) utilizes type IIS restriction enzymes for the hierarchical assembly of expression vectors from standardised parts into Saccharomyces cerevisiae expression cassettes. In Chapter 3, to facilitate the use of this toolkit for yeast surface display experiments, parts that allow for the incorporation of five surface display anchor proteins with various epitope tags into MoClo-YTK-compatible expression constructs were created. Building on prior work in the Young group allows these surface anchor proteins to be combined with a library of signal peptides (SPs) and translational fusion partners (TFPs) to optimise surface display of a protein of interest – a significant challenge for many recombinant proteins. This expansion of the MoClo-YTK yeast was validated using several proteins of interest and constitutes a yeast secretion/display toolkit (MoClo-YSD) that has been made available to the research community. The MoClo-YSD toolkit allowed the efficient use of yeast display for a range of different projects. It was successfully applied to develop a new approach to co-immunoprecipitation, where instead of immobilising an antibody on agarose beads, the antibody is displayed on the yeast surface, effectively using yeast as beads. As a proof of concept, an anti-GFP nanobody was used to capture GFP from bacterial lysates and GFP-tagged LNX1 protein from mammalian cell lysates. Co-precipitation of the LNX1 binding partners liprin-α1 and Numb, showed that yeast display can be used for studying protein-protein interactions. Furthermore, in Chapter 4, the MoClo YSD toolkit was adapted to co-display enzymes capable of degrading polyethylene terephthalate (PET), a traditional plastic that presents an environmental challenge due to its non-biodegradable nature and the resultant microplastic pollution. Enzymes such as LC-cutinase and MHETase, as well as hydrophobin, a protein that facilitates yeast attachment to hydrophobic surfaces, were successfully co-displayed, achieving notable degradation of low-crystallinity PET. Additionally, in Chapter 5, a method for developing biosensors for small molecules utilising yeast display was tested in this work. One application of such biosensors would be the detection of metabolites of interest during strain development and optimization for the production of biobased plastics. The biosensor design strategy employed yeast display with the goal of identifying protein-protein interactions that can be disrupted by the target small molecule, and thereafter adapted into a working biosensor. While a nanobody library was successfully screened to select for nanobodies that bind the target proteins (small molecule binders Dig10.2 and FDCA decarboxylase) the disruption of the nanobody-protein interactions by the small molecule analytes has not yet been demonstrated. However, nanobodies that specifically bind Dig10.2 were expressed in bacteria, purified and characterised in a plate-based binding assay and may be useful in other lines of research. Overall, this work highlights the diverse applications of yeast display in synthetic biology, from studying protein-protein interactions, to enhancing the binding of an enzyme to its substrate, environmental solutions and biosensing technologies.Item Functions of PDLIM2 protein in maintaining intestinal homeostasis(University College Cork, 2024) Ward, Stephanie; O'Connor, RosemaryDisruption to epithelial integrity is a common feature across disease pathogenesis and in tumour biology. Regulation of important cellular signalling pathways often becomes dysregulated resulting in aberrant activity, which contributes to the disease progression. In the context of cellular adhesion, loss of cell-cell and cell-extracellular matrix adhesions weakens the integrity of the epithelium which sensitises tissues to injury, exposure to microbial content, or permits cancer cells to undergo epithelial-mesenchymal transition (EMT) and metastasise. PDLIM2 is a cytoskeletal and nuclear associated protein which regulates the stability of many transcription factors associated with adhesion, inflammation, and EMT. Its repression during disease and cancer alters the stability of adhesion molecules, disrupts epithelial polarisation, and sustains activation of pro-inflammatory pathways which further aggravates symptoms. Nevertheless, high PDLIM2 expression in the tumour epithelia also contributes to an aggressive cancer phenotype. Thus, the balance in PDLIM2 expression is essential for maintaining epithelial homeostasis. This thesis, therefore, aims to investigate how expression of PDLIM2 may contribute to disease pathophysiology and the development of cancer. PDLIM2’s function was evaluated in the colorectal cancer cell line, Caco-2, as well as in murine models. In Caco-2 cells, PDLIM2 expression was either suppressed or knocked out to characterise the phenotypic changes which occur in the absence of PDLIM2, and to identify potential mechanistic factors. PDLIM2 wildtypes (PDLIM2+/+) and knockout (PDLIM2-/-) mice were assessed in relation to ulcerative colitis and colitis-associated colorectal cancer development to analyse PDLIM2’s function in disease and this data was compared to existing datasets of human colitis. Suppression and knockout of PDLIM2 in the Caco-2 cell line altered key adhesion pathways within the cell. These pathways were strongly associated with integrin signalling, which changed the cell’s capacity to adhere and migrate. Aberrant activation of adhesion pathways upon PDLIM2 suppression was sufficient to alter the metabolic dependencies of Caco-2 cells and increase their oxidative stress, which further contributed to the disruption of their adhesion signalling. In murine models, the absence of PDLIM2 increased epithelial injury during chronic colitis. PDLIM2-/- mice also retained higher numbers of pro-inflammatory immune populations in their colon tissue, with corresponding elevation in the expression of pro-inflammatory genes. The expression of PDLIM2 was also decreased in PDLIM2+/+ mice during acute colitis and, despite the recovery of their epithelium following chronic colitis, PDLIM2 levels were not restored. The absence of PDLIM2 also altered the biodiversity of the microbial populations in the colon, including the loss of commensal colonies. In human patients, the expression of PDLIM2 is reduced in patients with ulcerative colitis even in the absence of inflammation. This loss correlated with increased expression of adhesion-associated genes and decreased those associated with mitochondrial function. Overall, the data from this thesis highlight PDLIM2 as a key regulator of adhesion signals. The absence of its expression in colorectal epithelial cells dysregulates their adhesion signalling, with downstream impairment of antioxidant activity and cellular metabolism as a consequence. In the context of colitis, this loss of PDLIM2 likely contributed to the severity of the phenotype observed in our murine model, which showed reduced epithelial integrity and an altered microbiome. Furthermore, data from human ulcerative colitis patients support the cell line and murine data reported in this thesis, solidifying PDLIM2’s importance in protecting against disease development.Item Modernising the Ribo-Seq approach: a new gold standard(University College Cork, 2024) O'Connell, Aoife Myra; Loughran, Gary; Moore, Thomas F.; Irish Research CouncilRibosome profiling (Ribo-Seq) provides a genome-wide snapshot of protein synthesis by identifying the locations of translating ribosomes at single-nucleotide resolution via deep sequencing of ribosome protected mRNA fragments. First described in 2009, this technique has generated novel insights and a greater understanding of translation processes. However, the time-consuming, labour-intensive protocol, and lack of standardisation has impeded its uptake by the broader research community. Our aim is to modernise the Ribo-Seq approach through incorporation of cutting-edge technology, to develop a protocol at a fraction of the current cost, time, and labour. First, we investigated a number of cell lysis parameters to identify biases and to streamline the harvesting process. We find that the standard cell lysis procedure for Ribo-seq is depleted for a subset of mRNAs encoding cytoskeletal proteins. This issue can be easily resolved by omitting one step of the protocol. The second issue we tackled was the removal of contaminating rRNA. The majority of reads in Ribo-Seq libraries align to rRNA (over 80%), resulting in a reduction of useful mapping reads. We implemented a CRISPR-Cas9 approach to remove rRNA derived sequences from the final libraries. Additionally, we tested a novel approach termed Ribo-FilterOut, which involves EDTA treatment and ultrafiltration of samples to separate the small and large ribosomal subunits and release the mRNA footprints from the complex and reduce rRNA contamination. However, we failed to see an improvement in mapped coding reads or reduction in rRNA contamination. Lastly, a number of optimisations were employed to implement a streamlined Ribo-Seq approach, including adaptation of an existing low-input, one-pot protocol, with removal of critical laborious gel RNA-isolation steps, which helps facilitate high-throughput Ribo-Seq using microfluidic robotics. These modifications reduce the risk of sample loss throughout the procedure and allows for library generation in 2-3 days instead of the standard 5-7 days. As little as 1µg of total RNA input is required, allowing for expansion of ribosome profiling to a wider range of cells and tissues. Additionally, many cDNA library generation protocols exhibit consistent ligation bias at the 5’ end of reads, including our standard approach using SMARTer technology. To circumvent this, we implemented a protocol that uses ordered two-template relay that effectively resolves this heavy CG bias at the 5’ end.Item Cancer proteogenomics – identifying coding, long non-coding RNAs relevant in cancer(University College Cork, 2024) Zaheed, Oza; Dean, Kellie; Baranov, Pavel V.; Mészáros, Bálint; Science Foundation IrelandLong non-coding RNAs (lncRNAs) are RNA molecules exceeding 500 nucleotides that were traditionally considered non-functional and not translated (Mattick et al., 2023). However, recent advancements in next-generation sequencing have unveiled pivotal roles for lncRNAs in cellular processes, especially in disease contexts such as cancer. My work proposes that the non-coding transcriptome significantly impacts cancer cell biology, influencing aggressiveness and offering potential applications in patient diagnosis, disease stratification, and treatment. Cancer cells, characterised by altered metabolism and adaptations to cellular stress, may tip the balance between non-coding and coding functions. This shift could potentially result in the production of peptides from RNAs previously annotated as non-coding. Emerging evidence suggests that some of these 'non-coding' RNAs particularly lncRNAs may, in fact, encode proteins, often originating from small open reading frames (sORFs). Examples of these peptides generated from sORFs within non-coding transcripts span across species, from plants to humans, and many exhibit critical biological functions. My research focuses on identifying bifunctional RNAs that exhibit dual non-coding and coding functions in cancer. Leveraging RNA sequencing (RNA-seq) and ribosome profiling (Ribo-seq) datasets, I aim to challenge the conventional classification of 'noncoding' RNAs. My hypothesis posits that a single RNA molecule may serve both noncoding and coding roles, influenced by contextual cues and cellular conditions. To explore this hypothesis, I propose a multi-omic approach, aiming to identify potential peptides with pivotal roles in oncogenesis and cancer progression. This approach uncovered a total of 400 lncRNAs that were differentially expressed in malignant breast cancer cell lines. From this list of differentially expressed long non-coding RNAs, 64 candidate translated open reading frames were identified. By employing sequence analysis in combination with proteomic evidence and translated open reading frame curation, a scoring system was created to rank these candidates on the likelihood of producing a detectable microprotein. A total of nine parameters were assessed and fit into evidencebased categories including, translation, sequence information, and proteomics. This scoring system identified ten candidates most likely to produce a stable detectable microprotein with the three best scoring candidates arising from the lncRNAs ENSG00000253477, LINC02163, and MRPS30-DT. This integrated methodology holds promise for uncovering hidden coding potential within lncRNAs, offering novel insights into therapeutic targets and biomarkers. By examining the interplay between non-coding and coding functions in cancer, this research seeks to reshape our understanding of lncRNAs and their applications in cancer diagnosis and treatment.