Biochemistry and Cell Biology - Doctoral Theses
Permanent URI for this collection
Browse
Recent Submissions
Item LNX1 and LNX2 in the central nervous system: linking behavioural, cellular and molecular perspectives(University College Cork, 2025-12-31) Cioccarelli, Laura; Young, Paul; Science Foundation IrelandLigand of Numb protein-X (LNX) proteins are a conserved family of E3 ubiquitin ligases with emerging roles in neuronal development and function. In the central nervous system, LNX1p70/p62 isoforms and LNX2 show dynamic expression during development, with nearly undetectable protein abundance in the adult brain, making them notoriously difficult to study. Despite their low abundance, both proteins interact with multiple synaptic components, suggesting potential roles in shaping neuronal circuits. This dissertation investigates the behavioural, cellular and molecular consequences of Lnx1p70/p62 and/or Lnx2 deletion in mice, as explored across the following chapters. Chapter 2 revealed, through an extensive behavioural analysis, that loss of both proteins contributed to decreased anxiety-related behaviour, with LNX2 having a stronger influence on impulsivity and risk-taking phenotypes. These changes are consistent with altered stress-axis regulation. In contrast, Lnx1p70/p62 deletion is involved in altered pup ultrasonic vocalisation following maternal separation and reduced body weight. To further examine the potential role of Lnx1p70/p62 in growth and early development, Chapter 3 revealed a disruption of pruning and maturation of hippocampal CA3-mossy fiber synapses in Lnx1p70 deficient mice, associated with defects in social memory establishment. Furthermore, Lnx1p70/p62 deficient mice showed a decrease in circulating IGF-1 and leptin levels, possibly explaining the reduction in body weight. Novel LNX1p70 interactions with growth-related proteins, such as GHRHR, GIPR, TRIP6 and IGF1R were also characterised. In parallel, a possible role of LNX2 in stress response was examined in Chapter 4. In cultured cortical neurons, LNX2, but not LNX1p70, was transcriptionally induced following neuronal activation, suggesting a distinct, activity-dependent function. Proteomic analyses further revealed that Lnx2 deletion resulted in region-specific alterations in cerebellar and olfactory bulb protein networks. Finally, Chapter 5 described a complementary project that identified and characterised a functional nuclear localisation signal (NLS) within both LNX1 p80 and p70 isoforms, which is absent in LNX2. This finding partially explains why LNX1 localises to the nucleus upon transfection, whereas LNX2 remains largely cytoplasmic. Bioinformatic analysis further revealed that LNX1 NLS interacts with importin alpha family members, supporting its role in regulating LNX1 nuclear translocation. Altogether, these findings revealed distinct, non-overlapping neuronal functions of LNX1p70/p62 and LNX2, linking them to behavioural phenotypes relevant to anxiety and neurodevelopmental disorders and highlighting their potential roles in neuronal activity, growth regulation, synaptic maturation and compartment-specific signalling organisation.Item Analysis of SMAUG1 protein motifs and interacting partners - linking regulation of phase separation to human disease(University College Cork, 2025-12-31) Carey, Olivia; Dean, Kellie; Research IrelandRNA-binding proteins (RBPs) are central regulators of gene expression, influencing RNA processing, localisation, translation, and decay. A growing body of evidence shows that many RBPs exert their functions through the formation of dynamic, membrane-less organelles (MLOs) via liquid–liquid phase separation (LLPS). Dysregulation of these MLOs or “biomolecular condensates” is increasingly linked to human disease, including cancer and neurological disorders. Human SMAUG1 (SAMD4A) is a conserved translational repressor and condensate-forming protein implicated in neuronal function, metabolism, antiviral defence, and cancer. However, little is known regarding its condensate formation, localisation, and whole cell interactome. The aim of this PhD project was to define the molecular determinants of SMAUG1 phase separation, identify regulatory mechanisms controlling its condensates, and establish how condensate formation influences SMAUG1 localisation, protein interactome and RNA regulatory functions in the context of human health and disease. In Chapter 2, we defined the protein regions and molecular interactions that control SMAUG1 condensate formation in cells. Using protein domain and short linear motif (SLiMs) deletion mutants, we demonstrated that SMAUG1 phase separation is driven by its structured ‘Smaug similarity region 1’ (SSR1) dimerization domain and a novel intrinsically disordered prion-like domain. In contrast, the RNA-binding ‘sterile alpha motif’ (SAM) domain and other disordered regions were dispensable for condensation. We further identified members of the 14-3-3 protein family as potent negative regulators of SMAUG1 condensation. Dimeric 14-3-3 proteins that bind multiple phosphorylated motifs across SMAUG1 caused condensate dissolution. This work establishes a signalling dependent mechanism for dynamic control of SMAUG1 phase behaviour and with findings published as a first-author manuscript. In Chapter 3, we tested SMAUG1 condensation in vitro and defined its intrinsic phase separation behaviour. Recombinant SMAUG1 formed liquid-like condensates at low micromolar concentrations (0.1-2μM) under physiological salt (0-200 mM NaCl) and pH conditions (pH 7.5), in the absence of RNA or molecular crowders. SMAUG1 condensates exhibited hallmark liquid properties including fusion, surface wetting and ageing. Addition of purified 14-3-3γ dissolved SMAUG1 condensates in vitro, confirming that 14-3-3 proteins regulate SMAUG1 phase behaviour through direct interactions. In Chapter 4, we investigated how SMAUG1 phase separation influences its subcellular localisation and RNA interactome. Using motif prediction, in silico protein complex modelling, cell-based assays, and in vitro phenylalanine/glycine (FG) phase separation assays, we showed that SMAUG1 undergoes nucleocytoplasmic shuttling and identified a functional C-terminal nuclear export signal that promotes rapid export to the cytoplasm via exportin-1. Crucially, investigation of SMAUG1’s RNA targets in and out of condensates revealed a potential nuclear role for SMAUG1 in splicing through binding of intronic and exonic transcript regions. These findings expand SMAUG1’s functionality beyond translational repression and implicate it as a regulator of splicing, with the ability to bind more transcripts in its disperse state. Top RNA targets revealed further roles for SMAUG1 in nervous system development and degeneration, cell structure integrity, cell motility and guanosine triphosphatase (GTPase) enzyme regulation. Collectively, this thesis establishes SMAUG1 as a dynamically regulated phase-separating RBP that integrates intrinsic sequence features and interactions with 14-3-3 proteins to modulate phase behaviour. Meanwhile, phase state dictates SMAUG1’s RNA interactome in the nucleus and cytoplasm. These findings provide a mechanistic framework linking SMAUG1 condensate biology to processes associated with neurodevelopment, neurodegeneration and cancer, and highlight SMAUG1 as a potential target for future therapeutic intervention.Item Secondary bile acids as modulators of colonic secretion and immunity: mechanistic insights and implications for irritable bowel syndrome(University College Cork, 2025) Xiao, Qiao; O'Malley, Dervla; Joyce, Susan; China Scholarship Council; University College CorkBile acids (BAs) are best known for aiding lipid digestion but are increasingly recognized as signaling molecules that regulate gut function. Through receptors such as FXR, VDR, and TGR5—widely expressed in intestinal epithelial cells and enteric neurons—BAs influence secretion, motility, gut-brain communication, and immune responses. Disruptions in these pathways have been linked to disorders of gut-brain interaction, particularly irritable bowel syndrome (IBS), where altered profiles of luminal BA profiles are commonly observed. However, key questions remain relating to their role as signalling molecules. For example, how do BAs communicate with enteric neurons, how do their receptors respond to inflammatory signals, and do interactions between BAs, immune factors, and neural pathways contribute to IBS symptoms? We propose that disrupted BA signaling may drive bowel dysfunction through their modulatory actions on interconnected neuroendocrine and immune pathways intrinsic and extrinsic to the gut. This project used mass spectrometry, Ussing chamber electrophysiology, ELISA, calcium imaging and transcriptomic analyses to investigate BA profiles in the lumen and gut tissues of healthy Sprague-Dawley (SD) rats and stress-sensitive Wistar Kyoto (WKY) rats. Impaired microbial deconjugation markedly reduced luminal BA levels, including lithocholic acid (LCA) in WKY rat tissue. This reduction was associated with a remodeled immune environment marked by concurrent upregulation of pro- and anti-inflammatory mediators, indicating a newly established immune homeostasis. Furthermore, we specifically focused on the modulatory actions of LCA, a secondary BA, on colonic function. LCA potently inhibited colonic secretory currents evoked by veratridine, carbachol, and capsaicin, an effect that was independent of classical endocrine mediators such as GLP-1 and 5-HT, but appeared to be linked to local secretion of interleukin-6 (IL-6). Indeed, LCA had the capacity to activate the JAK-STAT signaling pathway, suggesting a link between BA signaling and mucosal immune activation. These findings identify a novel immunomodulatory role for LCA in regulating colonic secretion and suggest that disrupted BA metabolism and immune adaptation may contribute to bowel dysfunction in IBS.Item Identification and characterisation of non-coding RNAs involved in ductal carcinoma in situ(University College Cork, 2025) Derlipanska, Magdalina; Dean, Kellie; Das, Sudipto; University College CorkDuctal carcinoma in situ (DCIS) is an early form of breast cancer, the most common cancer in women worldwide. DCIS can develop into potentially fatal invasive disease, but studies show that less than half of all cases progress if untreated. Currently, most patients receive surgery, as there is no reliable way to distinguish which cases will become invasive. The goal of this work is to contribute to developing a molecular-based method that can differentiate indolent and potentially invasive DCIS. To do this, I identified non-coding ribonucleic acids (ncRNAs) that are involved in DCIS and aimed to discover their molecular functions. Through target-capture RNA sequencing, I identified 181 ncRNAs overexpressed in high-grade DCIS cell lines (ETCC-006 and ETCC-010), compared to normal-like breast cells. Next, the specificity of the ncRNAs was validated across a panel of cancer cell lines. From that, two candidates were selected for in-depth analysis: vtRNA2-1 and MEG3. Their expression was also measured in a small cohort of DCIS patients (n=12) and the overexpression observed in cell lines was confirmed. Vault RNA 2-1 (vtRNA2-1) is an RNA that has been implicated in some cancers; however, its mechanism of action is not clear. Through bisulfite sequencing I found it was hypomethylated in DCIS cell lines, driving its high expression levels. Antisense oligonucleotide knockdown resulted in reduced migration and overexpression increased proliferation in DCIS cells, suggesting it may act as an oncogene. Using RNA-affinity chromatography and UV-crosslinking, followed by mass spectrometry, protein interactors of vtRNA2-1 were identified. These showed enrichment in cytoskeletal processes. Furthermore, I examined genes with altered expression following knockdown of vtRNA2-1, and many of these were associated with various cancer processes. Taken together, my findings suggest an oncogenic role for vtRNA2-1 in the context of DCIS. Maternally expressed gene 3 (MEG3) is a tumour-suppressive long ncRNA that is usually downregulated in various cancers; however, I found that this RNA was highly expressed in DCIS cell lines and patient samples. Like vtRNA2-1, MEG3 was epigenetically regulated in DCIS cell lines. Previous observations that its knockdown increases proliferation and suppresses overexpression were confirmed. However, MEG3 appeared to promote migration, in contrast to what has been shown in breast cancer previously. I confirmed MEG3 to be primarily nuclear and examined genes altered after its knockdown. From gene set enrichment analysis (GSEA), pathways related to the cell cycle and neuronal maturation were prominent. Overall, MEG3 appears to be tumour suppressive in DCIS and could be a marker of indolent disease. By discovering these ncRNAs found in abundance in high-grade DCIS, but not in normal-like or invasive breast cancer cell lines and showing that they have a role in the cancer characteristics of DCIS, I believe these RNAs could eventually be used as biomarkers to stratify patients.Item Advancing ovarian cancer research through subtype specific models, molecular stratification, and a patient guided approach to identify long non-coding RNA biomarkers(University College Cork, 2025) McCabe, Aideen; Dean, Kellie; McDade, Simon; Research IrelandOvarian cancer is the most fatal gynaecological malignancy, responsible for over 200,000 deaths worldwide each year. Low ovarian cancer survival rates are primarily driven by a lack of effective early detection methods, poor understanding of subtype specific heterogeneity and non-specific symptoms. Therefore, there is a striking need to both improve our understanding of the molecular differences underlying ovarian cancer subtypes and to identify more effective early-detection biomarkers. This research should also be informed by the priorities of those affected by the disease in order to ensure relevance and increase the translation of results into the clinic. This thesis aims to address these gaps in ovarian cancer research via three main aims: improving pre-clinical model selection, creating accessible molecular subtyping tools, and identifying lncRNA biomarker candidates. Chapter 2 evaluates the transcriptomic and genomic features of a panel of 56 ovarian cancer cell lines to identify those that best match the gene expression profiles of various ovarian cancer subtypes. This chapter also shows that A2780, one of the most commonly used cell lines, is not representative of any subtype of ovarian cancer. Chapter 3 describes the development of classifieRov, a user-friendly Shiny application that implements consensus molecular subtyping for high-grade serous ovarian cancer (HGSOC) tumours, the most common and aggressive ovarian cancer subtype. This tool facilitates stratification of HGSOC tumours into four clinically relevant molecular subtypes, while also allowing researchers without bioinformatics expertise to apply a number of sample annotation tools to their data, making exploring these subtypes more accessible to the broader ovarian cancer research community. Finally, Chapter 4 details a patient and public involvement (PPI) guided approach to investigate long non-coding RNAs as potential blood-based biomarkers for ovarian cancer detection. In reponse to PPI input, single-cell RNA sequencing of peripheral blood mononuclear cells and analysis of tumor-educated platelet and primary tumour RNA profiles identified highly specific lncRNAs with potential for non-invasive diagnosis. Collectively, this work advances ovarian cancer research by establishing a panel of subtype-specific cell line models, creating accessible bioinformatics tools and implementing a patient-informed biomarker discovery pipeline.