Chapter 1 is an introduction to the work detailed in this thesis and details the background of the main biological molecules that are studied. The recent discovery of novel long-non coding RNAs (lncRNAs) reveals a new class of molecules with immense biological relevance and promising candidates for therapeutic targeting. Classified as RNA molecules longer than 200 nucleotides in length, lncRNAs have been found to be dysregulated in various disease conditions, and their return to steady-state levels through transcript level manipulation has been observed to induce cell and tissue survival. Of great interest is the lncRNA MALAT1 which is known to be dysregulated in various cancer tissues and neurodegenerative diseases. Localized within the nucleus, MALAT1 is involved in a variety of important biological processes including mRNA transcriptional regulation, alternative splicing, and stabilizing nascent nuclear speckles. A detailed analysis of direct protein binders to MALAT1 reveals TDP-43 as a strong binder with common functionalities present. Both MALAT1 and TDP-43 are dysregulated in neurodegenerative diseases, and their common functionalities overlap in alternative splicing and transcript regulation. An increase in MALAT1/TDP-43 binding is also observed in neurodegenerative diseased tissues, but the reason for this interaction is yet to be explored fully in literature. Thus, it is important to study the common functions of MALAT1 and TDP-43 in the context of neurodegeneration and understand if the MALAT1/TDP-43 complex plays a role in regulating biological mechanisms for cell survivability, which can then be used for therapeutic advancements in this field.

Chapter 2 is focused on the regulation of TDP-43 binding to mRNA targets by MALAT1. The increase in MALAT1/TDP-43 interaction in neurodegenerative disease tissues indicates a potential function for this interaction. Given TDP-43’s known function of regulating gene expression level, we hypothesized that MALAT1 may play a role in regulating TDP-43 function and gene regulation. We found that altering TDP-43 expression alters MALAT1 expression, and disruption of the steady-state level of either transcript induces cell death in neuroblastoma samples. TDP-43 binds to the 3’UTR of mRNA genes, and reduced MALAT1 expression can increase the binding ability of TDP-43 to target 3’UTR genes in HEK293 and SH-SY5Y cells. This suggests that MALAT1 regulates the binding activity of TDP-43 to target genes in the nucleus. This hypothesis is extended in a Parkinson’s Disease model through the induction of MPP+ to SH-SY5Y cells, and a similar alteration in TDP-43 binding and gene expression regulation is observed in altered MALAT1 conditions. From the collected data, we can infer the existence of a coordinated network of RNA-protein binding that mediates gene expression levels and ensures homeostatic concentrations in cell. Dysregulation of the MALAT1 lncRNA results in aberrant TDP-43 function, which has severe consequences for cell viability.

Chapter 3 details the role of MALAT1 in regulating alternative splicing through sequence-specific binding interactions. Alternative splicing (AS) regulation is a common function of both TDP-43 and MALAT1, and the increase in TDP-43/MALAT1interaction may affect gene expression level and molecular pathway regulation through changes in AS. We hypothesized that MALAT1 and TDP-43 may concurrently regulate gene expression of a common splice target, the spermine/spermidine acetyltransferase 1 (SAT1) pre-mRNA. Further investigation revealed the existence of overlapping sequence-specific binding regions between MALAT1/SAT1 and MALAT1/TDP-43 that regulated SAT1 AS. These regions help efficiently recruit SAT1 pre-mRNA to TDP-43, with MALAT1 acting as a bridge to direct localization. This is a novel proposed mechanism for AS regulation by lncRNA. We then extended this hypothesis to another splicing factor and pre-mRNA target and observed a similar recruitment phenomenon for the MALAT1/CSTF2/PPFIA3 interactome. We hence present a model in which MALAT1 can recruit both pre-mRNA and splicing factor protein via an overlapping binding region with distinct sequence-specific targeting for efficient AS.

Long non-coding RNAs (lncRNAs) have been implicated as regulators of cell growth and survival in a variety of cancer types, but the specific biological functions and interaction partners of many lncRNAs are still not well understood. In this dissertation, I used multiple biochemical and quantitative techniques to study how a novel cancer-associated lncRNA, termed Growth Regulator Antisense 1 (GRAS1), mechanistically regulates cell viability in cancer cells. In Chapter 2, I discussed the development of a system to investigate 5 new cancer-associated lncRNAs from screening data from multiple published studies, including linc01977, linc00883, PDCD4-AS1, linc00909, and GRAS1. More specifically, we established a working platform of overexpression and knockdown for these lncRNAs, performed cell survival screening on these lncRNAs using five different cancer cell lines, examined transcriptomic alterations in altered expression of these lncRNAs, and validated biotinylated probes to capture these lncRNAs. Based on the data in that chapter, we selected GRAS1 to be our lncRNA of interest to focus on. In Chapter 3, I reported that GRAS1 directly and specifically interacts with STAT5 in K562 cells using RNA antisense purification with mass spectrometry (RAP-MS). In addition, we found that knockdown of GRAS1 reduces phosphorylation of STAT5 as well as the activity of the STAT5 signaling pathway in K562 cells. In Chapter 4, I characterized the function of GRAS1 in non-small cell lung cancer cells. GRAS1 is critical for A549 cell survival, and depletion of GRAS1 RNA transcript induces cell death. We found that knockdown of GRAS1 resulted in a significant increase in protein levels of p53 and the DNA damage marker γH2AX. GRAS1 knockdown leads to differential expression of many mRNA transcript levels, including upregulation of DNA damage response factors involved in double strand break repair. Further, we identified the NF-κB activating protein (NKAP) as a direct interaction partner of GRAS1 using RAP-MS. Decreased GRAS1 levels in lung cancer cells caused a significant decline in NKAP level and induction of cell death through ferroptosis. The addition of the ferroptosis inhibitors ferrostatin and α-tocopherol could partially rescue the effect of GRAS1 knockdown on lung cancer cell survival. We found NKAP overexpression could completely reverse the ferroptosis as well as rescue the cell viability promoted by GRAS1 downregulation in lung cancer cells. Besides, NKAP overexpression could significantly reverse the effect of GRAS1 knockdown on lung cancer DNA damage. Chapter 4, in full, is currently being submitted for publication. Finally in Chapter 5, I listed the methods and materials we used in this study.

Non-coding RNA transcripts regulate diverse functions including gene expression, splicing, and cell growth in human cells through their interactions with RNA binding proteins. The full biological relevance of these RNA/protein interactions and the role of RNA in regulating protein function is still unclear in many cases. This work investigates two specific cases of lncRNA regulating gene expression to better understand the cellular functions and molecular details of lncRNA/protein interactions. Specifically, the ability of the lncRNAs, MALAT1 and XIST, to regulate gene expression through their interactions with the RNA binding proteins TDP-43 and SHARP, respectively, is examined in this thesis. Chapter 2 is focused on the interaction of TDP-43 protein and MALAT1 non-coding RNA. TDP-43 protein is associated with neurodegenerative disease and has been shown to bind directly to MALAT1 non-coding RNA. We investigated the dependence of TDP-43 localization to the 3’ untranslated regions (3’ UTRs) of other mRNAs upon the level of MALAT1 expression in human cells. Target RNAs selected for analysis were representative of high, moderate, and low expression mRNAs bound by TDP-43 in multiple previous CLIP-seq studies. To measure the affinity of TDP-43 binding to MALAT1 long noncoding RNA and 3’ UTR mRNA targets, we used a newly designed fluorescent multiplexed version of the electrophoretic mobility shift assay (mEMSA), described in Chapter 3. Binding affinities of TDP-43 for mRNA 3’ UTRs correlated positively with the number of UG repeats in the identified CLIP-seq peak binding region. In addition, CLIP-seq peak signal height directly correlated with TDP-43 in vitro binding affinity for mRNA 3’ UTR targets. The binding of TDP-43 protein to messenger RNA 3’ UTR targets was evaluated by immunoprecipitation and quantitative real-time PCR (RT-qPCR). Decreased MALAT1 expression resulted in re-localization of TDP-43 protein to mRNA 3’ UTRs. TDP-43 binding stabilized the mRNA expression levels of the mRNA transcripts. Neuroblastoma cells showed an upregulation of MALAT1, downregulation of mRNA targets, and increased cell death in response to MPP+ treatment, a cellular condition modeling Parkinson’s disease. Knockdown of MALAT1 reversed these results, indicating that the MALAT1/TDP-43 interaction contributes directly to cell death. Our data suggest that MALAT1 lncRNA expression is required to maintain the correct balance of TDP-43 binding to messenger RNA 3’ UTR targets in mammalian cells. These studies confirm the interdependence of TDP-43 and MALAT1 functions and suggest that mechanisms of neurodegenerative disease may depend in part on the dysregulation of non-coding RNA transcripts. Chapter 3 describes an updated, multiplexed method of the electrophoretic mobility gel-shift assay which can be used for measuring the binding affinities (Kds) for several RNAs in a single gel. This method is performed by first fluorescently labeling each RNA of interest with a unique fluorophore. RNA and protein binding partners are mixed, incubated, and separated by acrylamide gel electrophoresis. The binding of RNAs labeled with different fluorophores can then be measured simultaneously using a fluorescent imaging system, allowing the calculation of a dissociation constant (Kd) for each pair of RNA-protein interactions. This chapter outlines the mEMAS method and validates the approach with specific results for three RNA binding partners of TDP-43 (positive control, TARDBP, TDP-43’s own 3’UTR which has previously been shown to bind TDP-43 in vitro, the negative control BARD1, a region from a non-interacting mRNA, and CSNK1E, a known TDP-43 interactor without a previously published Kd). We found that this multiplexed method allows more rapid analysis and direct comparison of Kd measurements for multiple RNAs with the same protein. Results of the mEMSA approach are also found in Chapters 2 and 4. Chapter 4 provides another specific example of lncRNA/RBP interactions regulating gene expression. XIST non-coding RNA promotes the initiation of X chromosome silencing by recruiting the protein SHARP (SMRT and HDAC Associated Repressor Protein) to one X chromosome in female mammals. SHARP recruits N-CoR2 and HDAC3 to initiate histone deacetylation on the silenced X chromosome, leading to the formation of repressive chromatin marks and silencing gene expression. We show that binding of SHARP to XIST RNA requires RRM4 of the protein, in contrast to the requirement of RRM3 for specific binding to SRA RNA. Evaluation of SHARP binding to full-length, dimeric, or truncated versions of the A-repeat region revealed that high affinity binding of XIST to SHARP requires only four A-repeat segments. SHARP binding to XIST A-repeat RNA results in changes in the flexibility and structure of the RNA, reflected in chemical structure probing data, and an apparent conformational change of the XIST RNA when bound to SHARP protein by electron microscopy. These findings provide new information on the molecular basis of the XIST and SHARP interaction and how specificity in RNA-protein binding is achieved, furthering our understanding of the mechanisms ensuring proper gene dosage in mammals.