Aus dem Institut für Physiologische Chemie Geschäftsführender Direktor: Prof. Dr. Gerhard Schratt des Fachbereichs Medizin der Philipps-Universität Marburg Identification of Nova1 as a novel modulator of microRNA function in neurons INAUGURAL-DISSERTATION zur Erlangung des Doktorgrades der Naturwissenschaften dem Fachbereich Medizin der Philipps-Universität Marburg vorgelegt von Juliane Thümmler aus Halle (Saale) Marburg, 2016 Angenommen vom Fachbereich Medizin der Philipps-Universität Marburg am: Gedruckt mit Genehmigung des Fachbereichs. Dekan: Herr Prof. Dr. H. Schäfer Referent: Herr Prof. Dr. G. Schratt 1. Korreferent: Herr Prof. Dr. B. Schmeck 1 Table of Contents Abstract ............................................................................................................................ 5 Zusammenfassung ........................................................................................................... 7 1 Introduction .............................................................................................................. 9 1.1 Definition of microRNAs .................................................................................. 9 1.2 Mechanisms of miRNA-mediated gene silencing ............................................. 9 1.2.1 MiRNA biogenesis pathway.......................................................................... 9 1.2.2 MicroRNA-induced silencing complex (miRISC) ...................................... 11 1.3 MiRNAs in the nervous system ....................................................................... 13 1.3.1 Role of miRNAs in neuronal development ................................................. 13 1.3.2 Role of miRNAs in mature neurons ............................................................ 14 1.3.3 Role of miRNAs in neuronal diseases ......................................................... 17 1.4 Regulation of neuronal miRNAs ..................................................................... 18 1.4.1 Activity-dependent regulation of neuronal miRNAs................................... 18 1.4.2 MiRNA regulation at the level of miRISC .................................................. 20 1.5 Aim of the study .............................................................................................. 22 2 Materials and Methods .......................................................................................... 23 2.1 Materials .......................................................................................................... 23 2.1.1 Chemicals and reagents ............................................................................... 23 2.1.2 Enzymes and inhibitors ............................................................................... 24 2.1.3 Lab equipment ............................................................................................. 25 2.1.4 Kits............................................................................................................... 26 2.1.5 Buffers and solutions ................................................................................... 26 2.1.6 Cells and culture media ............................................................................... 28 2.1.7 Oligonucleotides .......................................................................................... 28 2.1.8 RNA molecules............................................................................................ 29 2.1.9 Plasmids ....................................................................................................... 30 2.1.10 Antibodies.................................................................................................... 33 2 2.2 Cell culture methods ........................................................................................ 33 2.2.1 Primary neuronal cell culture ...................................................................... 33 2.2.2 Human embryonic kidney 293 cell culture.................................................. 34 2.2.3 Transfection of neurons with Lipofectamine 2000...................................... 34 2.2.4 Transfection of HEK293 cells with calcium-phosphate .............................. 34 2.2.5 Production of recombinant adeno-associated virus (rAAV) ....................... 35 2.2.6 rAAV infection of neurons .......................................................................... 35 2.3 Molecular methods .......................................................................................... 36 2.3.1 Annealing of miRNA duplex....................................................................... 36 2.3.2 Polymerase chain reaction (PCR) ................................................................ 36 2.3.3 Agarose gel electrophoresis ......................................................................... 36 2.3.4 Purification of plasmid DNA and PCR products ......................................... 37 2.3.5 Restriction digest, dephosphorylation, Klenow reaction and ligation of DNA ..................................................................................................................... 37 2.3.6 Transformation of competent bacteria......................................................... 37 2.3.7 Plasmid midi preparation ............................................................................. 38 2.3.8 Site-directed mutagenesis ............................................................................ 38 2.3.9 cDNA synthesis and RT-qPCR ................................................................... 38 2.4 Biochemical methods ...................................................................................... 39 2.4.1 Subcellular fractionation.............................................................................. 39 2.4.2 Protein extraction......................................................................................... 39 2.4.3 Western blot analysis ................................................................................... 40 2.4.4 Luciferase reporter assay ............................................................................. 40 2.4.5 Tethering assay ............................................................................................ 41 2.4.6 Co-immunoprecipitation (co-IP) ................................................................. 41 2.4.7 RNA-immunoprecipitation (RNA-IP) ......................................................... 42 2.5 Immunocytochemistry (ICC)........................................................................... 42 2.6 Microscopic analysis ....................................................................................... 43 2.7 Spine assay ...................................................................................................... 43 2.8 Sholl analysis ................................................................................................... 43 2.9 Statistical analysis............................................................................................ 44 3 3 Results ..................................................................................................................... 45 3.1 Background – RNAi-based screen .................................................................. 45 3.2 Validation of RNAi screen results ................................................................... 47 3.3 Expression analysis of Nova1 in primary neuronal cultures ........................... 50 3.4 Nova1 is required for the repressive function of miRNAs .............................. 51 3.4.1 Nova1 knockdown impairs miR-138 repressive activity in young cortical neurons ..................................................................................................................... 51 3.4.2 Nova1 is required for the repressive activity of endogenous miR-134 and miR-138.................................................................................................................... 53 3.4.3 Nova1 knockdown interferes with miR-134-, miR-124- and let-7-dependent repression of luciferase reporter genes ..................................................................... 55 3.4.4 Nova1-dependent regulation of miRNA activity is independent of the 3’UTR context ..................................................................................................................... 56 3.4.5 Nova1 is required for miR-134 mediated repression of endogenous Limk1 ..................................................................................................................... 58 3.5 Nova1 interaction studies ................................................................................ 59 3.5.1 Nova1 interacts with Ago in lysates obtained from the rat hippocampus ... 59 3.5.2 Validation of Nova1 / mRNA association by RNA immunoprecipitation .. 61 3.5.3 Luciferase-based tethering assay of full length Nova1 protein ................... 62 3.5.4 Luciferase-based tethering assay of Nova1 deletion constructs .................. 64 3.6 The function of Nova1 in neuronal morphogenesis and signal transduction .. 66 3.6.1 Nova1 is required for miR-134 mediated spine morphogenesis ................. 66 3.6.2 The role of Nova1 in dendrite outgrowth .................................................... 67 3.6.3 The function of Nova1 in BDNF signaling ................................................. 69 4 Discussion ................................................................................................................ 71 4.1 RNAi based screen in neurons......................................................................... 71 4.2 Nova1 is expressed in the nucleus and cytoplasm of neurons ......................... 72 4.3 Nova1 is required for miRNA function ........................................................... 73 4.4 Nova1 associates with miRISC and mRNA .................................................... 74 4 4.5 The significance of Nova1 for neuronal morphogenesis and function ............ 78 4.6 Conclusion and Outlook .................................................................................. 81 References....................................................................................................................... 84 6 Appendix ................................................................................................................. 98 6.1 List of figures and tables ................................................................................. 98 6.2 List of abbreviations ...................................................................................... 100 List of academic teachers ............................................................................................ 103 5 Abstract The proper development, differentiation and plasticity of the nervous system require an accurate regulation at multiple levels of gene expression. One important class of post- transcriptional regulators are microRNAs (miRNAs), tiny RNA molecules that inhibit protein synthesis of target mRNAs at the level of mRNA translation and stability. MiRNAs have well documented roles in the control of neuronal development and function. Specific miRNAs, such as miR-134, regulate the local dynamic translation of mRNAs at the synapse thereby controlling activity-dependent changes in synaptic strength. MiRNAs regulate mRNA translation within a large RNA-protein complex, the microRNA-induced silencing complex (miRISC). Whereas the core components of miRISC, e.g. Argonaute (Ago) and GW182 proteins, are highly conserved, cell-type specific auxiliary proteins play important roles in the modulation of miRISC activity. The molecular mechanisms by which miRISC activity is specifically regulated in the neuronal system by such auxiliary proteins however are poorly described. This project presents the validation and evaluation of the first large scale screening study that was performed in order to find novel RNA-binding proteins (RBPs) that regulate miRISC activity in primary neurons. The RNAi-based screen identified the RBPs Nova1, Ncoa3 and Ewsr1 as new modulators of miRISC in neurons and further confirmed the function of two previously reported miRISC interacting proteins (Ddx6, Tnrc6c). Subsequently, Nova1 was chosen for follow-up experiments directed at the elucidation of the underlying molecular mechanisms. Using luciferase reporter assays for targets of multiple neuronal miRNAs in addition to miR-134, I obtained evidence that Nova1 regulates miRNA activity in neurons in a general manner. In addition, I found that Nova1 can act as a regulator of miRNA activity irrespective of the 3‘UTR context. Further data demonstrated that Nova1 interacts with Ago and the miR-134 target mRNA Limk1. Investigations on the functional relevance of this mechanism revealed that Nova1 is required for miR-134 mediated spine size reduction. Furthermore, I could show that Nova1 is necessary for the upregulation of Limk1 translation upon treatment with brain- derived neurotrophic factor (BDNF), suggesting that Nova1 could be involved in the stimulus-dependent control of neuronal mRNAs. In summary, Nova1 was identified as a new modulator of miRISC activity in neurons and the underlying mechanism was 6 characterized in detail. The data obtained during the thesis project suggests that Nova1 is involved in the dynamic activity-dependent regulation of miRNA function in neurons. 7 Zusammenfassung Die korrekte Entwicklung, Differenzierung und Plastizität des Nervensystems erfordern eine genaue Regulation auf multiplen Ebenen der Genexpression. Eine wichtige Klasse post-transkriptioneller Regulatoren sind mikroRNAs (miRNAs), winzige RNA-Mole- küle, welche die Proteinsynthese von Ziel-mRNAs auf der Ebene der mRNA-Translation und -Stabilität inhibieren. MiRNAs haben wichtige Aufgaben in der Kontrolle der neuronalen Entwicklung und Funktion. Spezifische miRNAs, wie z. Bsp. miR-134, regulieren die lokale dynamische Translation von mRNAs an der Synapse, wodurch sie aktivitätsabhängige Veränderungen der Synapsenstärke kontrollieren. MiRNAs regulieren die mRNA Translation innerhalb eines großen RNA-Protein-Komplexes, des mikroRNA-induzierten Silencing Komplexes (miRISC). Während die Kernkomponenten des miRISC, wie z Bsp. Argonaute (Ago) und GW182 Proteine, hoch konserviert sind, kommen Zelltyp-spezifischen Proteinen wichtige Aufgaben in der Modulation der miRISC-Aktivität zu. Die molekularen Mechanismen, durch welche die miRISC- Aktivität speziell im neuronalen System durch solche zusätzlichen Proteine reguliert wird, sind nur unzulänglich beschrieben. Dieses Projekt stellt die Validierung und Evaluierung der ersten umfangreichen Screening-Studie dar, die durchgeführt wurde, um neue RNA-bindende Proteine (RBP) zu identifizieren, welche die miRISC-Aktivität in primären Neuronen regulieren. Der RNAi-basierte Screen identifizierte die RBPs Nova1, Ncoa3 und Ewsr1 als neue Modulatoren von miRISC in Neuronen und bestätigte außerdem zwei zuvor beschriebene miRISC-interagierende Proteine (Ddx6, Tnrc6c). Anschließend wurde Nova1 für nachfolgende Experimente ausgewählt, welche zur Aufklärung der zugrundeliegenden Mechanismen beitrugen. Durch die Anwendung von Luciferase-Reporter Assays für Ziel- mRNAs von multiplen neuronalen miRNAs neben miR-134 erbrachte ich Nachweise, dass Nova1 die miRNA-Aktivität in Neuronen auf generelle Weise reguliert. Zusätzlich fand ich heraus, dass Nova1 als Regulator der miRNA-Aktivität in Unabhängigkeit des 3’UTR Kontextes agieren kann. Weitere Daten zeigen eine direkte Interaktion von Nova1 mit der RISC Kernkomponente Ago und eine Assoziation mit der miR-134 Ziel-mRNA Limk1. Untersuchungen über die funktionale Bedeutung dieses Mechanismus zeigten, dass Nova1 für die miR-134-vermittelte Reduzierung der Größe von dendritischen 8 Dornfortsätzen notwendig ist. Außerdem wurde gezeigt, dass Nova1 für die hochregulierte Translation der Limk1 mRNA nach BDNF-Behandlung notwendig ist. Dies lässt auf eine Mitwirkung von Nova1 in der Stimulus-abhängigen Kontrolle von neuronalen mRNAs schließen. Zusammenfassend wurde Nova1 als neuer Modulator der miRISC-Aktivität in Neuronen beschrieben und die zugrundeliegenden Mechanismen genauer charakterisiert. Die während dieser Doktorarbeit erhaltenen Daten deuten zusammengenommen auf eine Beteiligung von Nova1 in der dynamischen aktivitätsabhängigen Regulation der miRNA-Funktion in Neuronen hin. INTRODUCTION 9 1 Introduction 1.1 Definition of microRNAs MicroRNAs (miRNAs) are highly conserved short RNA molecules (~22 nucleotides) that constitute an important class of small non-coding RNAs. As essential regulators of post- transcriptional gene expression miRNAs are involved in a wide range of complex cellular processes including those that occur during development, differentiation and the adaptation to environmental changes. Since their first discovery in the nematode Caenorhabditis elegans, more than 3000 distinct miRNAs have been described in a variety of organisms ranging from simple eukaryotes to mammals. Nevertheless, the physiological role of many miRNAs is still unknown (Lee et al. 2001; Griffiths-Jones et al. 2008; Friedländer et al. 2014). 1.2 Mechanisms of miRNA-mediated gene silencing MiRNAs are encoded within the genome as independent genes or in intronic regions of protein coding genes. The generation of most mammalian miRNAs is executed by the canonical miRNA pathway (Figure 1) that is described in the following chapter (reviewed by O'Carroll and Schaefer. 2013; Ha and Kim. 2014). The miRNA-mediated silencing of target mRNAs is enabled by a protein effector complex called the miRNA-induced silencing complex (miRISC) that is composed of core components and several associated factors. 1.2.1 MiRNA biogenesis pathway MiRNA coding genes are first transcribed in the nucleus by RNA polymerase II into primary miRNA transcripts (pri-miRNAs) which may be several kilobases (kb) long (Lee et al. 2004). These transcripts fold into imperfectly paired, double stranded stem loop structures that are recognized in the nucleus by a specific processing complex. This so called microprocessor complex is composed of two core enzymes, the type III ribonuclease Drosha and the RNA binding protein Dgcr8 (DiGeorge syndrome critical region 8) and additional auxiliary subunits (Gregory et al. 2004). Dgcr8 acts as a INTRODUCTION 10 molecular anchor that binds to the pri-miRNA hairpin structure and recruits Drosha by a direct interaction (Han et al. 2006). Drosha itself cleaves the hairpin of the pri-miRNA and thereby generates a shorter molecule (~ 65 nucleotides) that is called the precursor miRNA (pre-miR). Subsequently, the pre-miR is exported to the cytoplasm by Exportin- 5 in a GTP-dependent manner (Yi et al. 2003). In the cytoplasm, the loop region of the pre-miR is removed by the type III ribonuclease Dicer and its cofactor TRBP (trans- activation response RNA-binding protein) thereby releasing an about 22 nucleotide long double-stranded RNA (Chendrimada et al. 2005). The small RNA duplex is then loaded onto an Ago protein to form an effector ribonucleoprotein, known as the miRNA-induced silencing complex (miRISC). However, only one strand of the duplex (mature miRNA) serves as a guide for the recognition of the target mRNA. Figure 1: Overview of the canonical miRNA biogenesis pathway (Pfaff and Meister. 2013) MiRNA genes are transcribed into primary miRNA transcripts. In the nucleus, the primary transcripts are processed by the Drosha/DGCR8 microprocessor complex to miRNA precursors. Upon export to the cytoplasm, the miRNA precursors are further processed by Dicer/TRBP to short double-stranded RNA. One strand is incorporated with Ago in the context of the miRNA-induced silencing complex (miRISC). MiRISC either directly cleaves target mRNAs or enables translational repression and decay of target mRNAs. N, nucleus; C, cytoplasm; Pol, RNA polymerase; DGCR8, DiGeorge syndrome critical region 8; TRBP, trans-activation response RNA-binding protein INTRODUCTION 11 The detailed mechanisms of strand selection and miRISC loading in mammals are not fully understood. Recent studies revealed that first the double-stranded miRNA is loaded onto an Ago protein that is associated with Dicer and TRBP to form the minimal RISC - loading complex (RLC) (MacRae et al. 2008; Noland et al. 2011). Upon unwinding of the duplex RNA the passenger strand (miRNA* sequence) is released and degraded. The mature miRNA is now incorporated in the mature miRISC that enables the silencing of target mRNAs. 1.2.2 MicroRNA-induced silencing complex (miRISC) MiRNA-mediated silencing of target mRNAs is facilitated by the miRISC that is composed of two core proteins and additional factors that influence miRISC activity. The two core factors of miRISC which are indispensable for miRNA-dependent silencing are (i) one member of the Ago protein family and (ii) the direct Ago interaction partner GW182 (glycine-tryptophan protein of 182 kDa) (reviewed by Fabian and Sonenberg, 2012; Pfaff and Meister, 2013). The mammalian genome encodes four different Ago family members (Ago1-Ago4) that are all capable of binding miRNAs. In contrast to Ago proteins from other species, human Ago proteins generally appear to lack preference for specific miRNA sequences. (Czech et al. 2011; Burroughs et al. 2011; Dueck et al. 2012). MiRNAs loaded in Ago guide the miRISC to specific target mRNAs in order to initiate translational repression and mRNA destabilization. Generally, miRNAs bind partially complementary to target mRNAs with mismatches at the 3’ region but complementarity at the seed region (position 2 to 8) of the miRNA binding site (Bartel. 2009). In the rare case of a complete complementary base-pairing, exclusively Ago2 is able to catalyse endonucleolytic cleavage of the respective target mRNAs (Meister et al. 2004). The Ago protein structure is composed of the N-terminal domain, the PAZ domain, the middle (MID) domain and the PIWI domain. The N-terminal domain is required for miRNA loading and supports the unwinding of the RNA duplex (Kwak et al. 2012). The PAZ domain is binding to the 3’ end of the miRNA whereas the MID domain confers binding to the 5’end of the miRNA (Jinek et al. 2009). The PIWI domain structure is similar to that of the endonuclease RNase H and enables the cleavage of target mRNAs INTRODUCTION 12 at a specific nucleotide position (Song et al. 2004). Moreover, the PIWI domain functions as a binding platform for GW182 protein (Till et al. 2007). GW182 itself serves as a binding platform for additional effector proteins which in turn induce translational repression, deadenylation and/or degradation of target mRNAs. Mammalian genomes encode three GW182 paralogues which are termed TNRC6 (trinucleotide repeat-containing protein 6) A, B and C (Eulalio et al. 2009). The N- terminal part of the human GW182 contains a multitude of GW repeats that enable the direct interaction with Ago. The middle part of GW182 protein is composed of a putative ubiquitin-associated (UBA) domain and a glutamine (Q)-rich domain that is responsible for localizing GW182 to processing (P)-bodies in the cytoplasm. A silencing domain within the C-terminal part that is comprised of the motifs PAM2 (poly(A)-binding protein interacting motif 2) and RRM (RNA recognition motif) facilitates interaction of GW182 with the deadenylation machinery. One of the direct interaction partners of GW182 is the poly(A)-binding protein (PABP) that is bound through the PAM2 domain. Thereby GW182 might interfere with translational initiation of target mRNAs by preventing the binding of PABP to the eukaryotic initiation factor 4G (eIF4G). Furthermore, GW182 recruits the two cytoplasmic deadenylase complexes CCR4-NOT and PAN2-PAN3, promoting deadenylation of target mRNAs (Fabian et al. 2012; Huntzinger et al. 2013). The deadenylation of mRNAs subsequently leads to the removal of the cap structure at the 5’-end by decapping that gives rise to exonucleolytic degradation by the 5’-3’- exonuclease XRN1 (Rehwinkel. 2005; Nishihara et al. 2013) (Figure 2). The timing and relative importance of both translational repression and mRNA destabilization is still under debate. Recent literature points towards a direct connection between both mechanisms. It is furthermore reported that translational repression occurs first and is then followed by deadenylation and decay of mRNAs (Béthune et al. 2012; Meijer et al. 2013; Huntzinger et al. 2013). On the contrary, another model suggests mRNA destabilization as the dominant effect of miRNAs (Eichhorn et al. 2014). INTRODUCTION 13 Figure 2: Mechanisms of miRNA-mediated gene silencing (Pfaff and Meister. 2013) By binding to Ago proteins miRNAs guide the miRISC to target mRNAs. Ago recruits GW182 proteins to initiate translational repression and mRNA decay. GW182 forms a binding platform for PABPC1 and the deadenylase complexes PAN2/3 and CCR4/NOT thereby inducing deadenylation of target mRNAs. Subsequently, the mRNA is decapped by DCP1/2 enzyme activity and degraded by the 5’-3’ exonuclease XRN1. 1.3 MiRNAs in the nervous system The ability of miRNAs to control cellular processes by regulating gene expression is exceedingly important for the molecular networks that regulate the complexity of the nervous system. Numerous miRNAs are exclusively expressed or enriched in the brain and were found to be abundant in certain brain areas like the hippocampus (Bak et al. 2008). Neuronal miRNAs exert different roles in neurogenesis, differentiation, maturation, physiological brain function and dysfunction (reviewed by Saba et al. 2010; Olde Loohuis et al. 2012). 1.3.1 Role of miRNAs in neuronal development Numerous studies in different model organism revealed an important contribution of miRNAs to multiple stages of neuronal development by employing loss-of-function mutants of miRNA biogenesis pathway components resulting in a global loss of mature miRNAs (reviewed by Meza-Sosa et al. 2014). Early studies in zebrafish demonstrated that the disruption of Dicer leads to abnormal cell morphogenesis in the brain INTRODUCTION 14 development. This phenotype could be rescued by introducing miR-430 suggesting an essential role of this miRNA in neuronal development (Giraldez et al. 2005). Another important study revealed that selective ablation of Dicer in the neocortex during mouse embryonic development results in decreased growth of the postnatal cortex and an impairment in neuronal layering (Pietri Tonelli et al. 2008). Furthermore, knockout of Dgcr8, caused morphological abnormalities in the central nervous system and impairments in the dendritic arborization in mice (Stark et al. 2008). More recent work studied the function of individual miRNAs in the developing brain. Thereby, miR-124 and miR-9 emerged as highly conserved brain-specific miRNAs that display important functions in neuronal development and differentiation of different organisms (Makeyev et al. 2007; Cheng et al. 2009; Coolen et al. 2013). In the developing brain miRNAs were also implicated in pathways that are required for processes involved in neuronal network formation, such as axon guidance (reviewed by Iyer et al. 2014). 1.3.2 Role of miRNAs in mature neurons The main features of mature neurons are their complex structure, the connectivity between each other through synaptic contacts and the plasticity of these connections, by which neurons can adjust to changes in neuronal activity. A main research focus was on the function of miRNAs in synaptic plasticity, a cellular correlate of memory storage. Long-lasting forms of synaptic plasticity, such as long-term potentiation (LTP) or long- term depression (LTD), involve the formation/elimination of synapses and the strengthening/weakening of pre-existing synapses. Such structural modifications often happen at the level of dendritic spines, membranous protrusions from dendrites where most excitatory synapses are formed. One mechanism that is important for activity- dependent spine modifications is the local translation of specific dendritically localized mRNAs. Recent work established a critical function for dendritically localized miRNAs in the regulation of local mRNA translation and spine plasticity. Here, I will focus on three extensively studied miRNAs, miR-132, -134 and -138 (Bicker et al. 2014) (Figure 3). miR-134 MiR-134 is the first miRNA that was shown to localize to dendrites. MiR-134 is required for multiple processes that involve dendrite remodeling, including the regulation of dendritic spine morphogenesis and activity-dependent dendritogenesis. In addition to INTRODUCTION 15 mature miR-134, pre-miR-134 was shown to localize to dendrites, suggesting a regulatory function of localized pre-miR processing. Transport of pre-miR-134 to dendrites is dependent on the RNA helicase DHX36, which interacts with the pre-miR-134 loop structure (Bicker et al. 2013). MiR-134 negatively regulates the size of dendritic spines by repressing the translation of Limk1 mRNA. It was shown in hippocampal neurons that overexpression of miR-134 leads to a spine size reduction, whereas miR-134 inhibition increased spine size (Schratt et al. 2006). Furthermore, miR-134 regulates dendritogenesis by targeting the mRNA of the RNA-binding protein Pumilio2 (PUM2) (Fiore et al. 2009). Thus, both overexpression and inhibition of miR-134 block activity- dependent dendritogenesis, suggesting that regulation of miR-134 levels within a narrow window is critical for proper neural development. Recently, the mRNA of Ubiquitin- protein ligase E3A 1 (Ube3a1) was found as a target of several miRNAs encoded by the miR-379/410 cluster, including miR-134 (Valluy et al. 2015). miR-132 MiR-132 belongs to the miR-212/132 family formed by four highly related miRNAs (miR-132 5p/3p, miR-212 5p/3p) which are highly conserved in the vertebrate lineage. MiR-132 promotes activity-dependent dendritogenesis and spine growth by inhibiting the expression of the Rho GTPase activating protein p250GAP, an upstream inhibitor of the Rac-PAK signaling pathway that regulates remodeling of the actin cytoskeleton (Wayman et al. 2008; Siegel et al. 2009; Impey et al. 2010). Dendritic miR-132 function might rely on an interaction with the RNA-binding protein FMRP, which also associates with miRISC (Edbauer et al. 2010). Recently, miR-132-mediated translational repression of the matrix metalloproteinase-9 (MMP-9) mRNA was shown to be involved in activity- dependent regulation of dendritic spine morphology (Jasińska et al. 2015). Another important target mRNA of miR-132 is methyl CpG-binding protein 2 (MeCP2) mRNA. MeCP2 is a transcriptional repressor that inhibits among other genes the expression of the neurotrophin BDNF. Since BDNF in turn activates miR-132, this constitutes a feedback control mechanism that might be involved in neuronal homeostasis (Klein et al. 2007; Su et al. 2015). miR-138 MiR-138 is another dendritic miRNA that was identified as a regulator of spine morphogenesis. It inhibits spine growth by repressing Lypla1 mRNA translation. The INTRODUCTION 16 LYPLA1 gene encodes Apt1 (acyl-protein thioesterase 1), a protein with depalmitoylase enzymatic activity (Siegel et al. 2009). The miR-138-mediated downregulation of APT1 increases palmitoylation and therefore membrane association of the RhoA signaling activator Gα12/13. The resulting activation of RhoA modifies the actin cytoskeleton in dendritic spines which leads to spine shrinkage. Another study demonstrated that miR- 138 is associated with Lypla1 mRNA at synapses where it locally mediates its translational repression. Moreover, activity-regulated proteasomal degradation of MOV10, a component of miRISC, enables the release of Lypla1 mRNA from miR-138- mediated translational inhibition (Banerjee et al. 2009). In addition, miR-138 expression was shown to correlate with short-term recognition memory in mice (Tatro et al. 2013) and later with memory performance in humans (Schröder et al. 2014). Intriguingly, mature miR-138 derives from two individual precursor forms that are encoded by two different genetic loci. The precursors pre-miR-138-1 and -2 exhibit differences in size, nucleotide sequence of the stem loop and transcription. The longer and highly expressed pre-miR-138-2 is thought to be the main source of mature miR-138 in the nervous system (Obernosterer et al. 2006). Figure 3: Control of dendritic spine size by miRNA regulatory pathways (Schratt. 2009) Dendritic spine modifications arise from dynamic changes of the actin cytoskeleton which is controlled by two signaling pathways, the Rho/ROCK cascade and the Rac-Limk1 cascade. Several dendritic miRNAs display distinct roles in the regulation of these pathways. MiR-138 mediates the repression of the depalmitoylation enzyme acyl-protein thioesterase 1 (APT1) thereby increasing the membrane association of the RhoA signaling activator Gα12/13 leading to actomyosin contraction and subsequently reduction of spine size. MiR-134 also induces spine shrinkage by repressing the expression of Limk1 which leads to a blockage of actomyosin polymerization. In contrast, miR-132 promotes spine growth by repressing the Rac-inactivating protein p250RhoGAP that results in actomyosin polymerization. INTRODUCTION 17 1.3.3 Role of miRNAs in neuronal diseases The increasing knowledge about the function of neuronal miRNAs also inspired research into possible roles of miRNAs in brain dysfunction. Cognitive impairments and learning disabilities are common characteristics of diseases of the nervous system such as Fragile X syndrome, Alzheimer disease, Parkinson’s disease, Huntington disease, schizophrenia, epilepsy and autism spectrum disorders. Several studies suggest that miRNA-mediated impairments in posttranscriptional regulation are causally involved in the development of these diseases (Saba et al. 2010; Nowak et al. 2013; Alsharafi et al. 2015). Altered miRNA expression in post mortem brains of patients as well as in animal models of neuronal diseases further support this notion (Abu-Elneel et al. 2008; Moreau et al. 2011). The X-chromosomally inherited neurodevelopmental disorder fragile X syndrome (FXS) was first associated with miRNAs (Jin et al. 2004a). Mutations in the FMR1 gene result in the loss of the RNA-binding protein FMRP (fragile X mental retardation protein) that is involved in mRNA transport and acts as a translational repressor through association with the miRISC. Loss of FMRP leads to increased translation of mRNAs at the synapse and as a consequence might cause impairments of synaptic plasticity. It was demonstrated that FMR1-knockout mice display an increase in dendritic spine length and density (Nimchinsky et al. 2001). Moreover, a FMRP knockdown study in mouse brain implicated FMRP in the regulation of dendritic spine morphology through an association with miR-125b and miR-132 (Edbauer et al. 2010). The dysregulation of individual miRNAs has also strongly been implicated in neuropsychiatric disorders such as schizophrenia and epilepsy. A high risk for developing schizophrenia-like symptoms is associated with microdeletions in the chromosomal region 22q11.2 that contains the gene encoding for the miRNA processing factor DGCR8. Deletions of this region are accompanied by alterations in brain miRNA biogenesis, thus providing a link between impaired miRNA expression and the etiology of schizophrenia (Stark et al. 2008). Moreover, the expression of numerous miRNAs was altered in post mortem brain samples of schizophrenia patients (reviewed by Beveridge et al. 2012). From these, miR-137 emerged as one of the most significant schizophrenia-associated miRNAs (Yin et al. 2014). In addition, multiple studies have linked alterations of miRNA expression to epilepsy (reviewed by Alsharafi et al. 2015). For example, expression levels of miR-132, miR-34a and miR-124 were found to be changed in experimental models of INTRODUCTION 18 induced epileptic seizures and in human epilepsy (Nudelman et al. 2010; Sano et al. 2012; Peng et al. 2013; Santarelli et al. 2011). Of note, miR-134 was found to be functionally relevant in the development of epilepsy because silencing of miR-134 prevents recurrent seizure occurrence in mice (Jimenez-Mateos et al. 2012) 1.4 Regulation of neuronal miRNAs Since miRNAs are key regulators of activity-dependent processes in the nervous system, their turnover and function have to be precisely regulated, particularly in response to developmental or environmental cues. A particular quality of neuronal miRNAs is the activity-dependent regulation that enables cellular adaptation to changes in neuronal activity, for example during synaptic plasticity. 1.4.1 Activity-dependent regulation of neuronal miRNAs Neuronal miRNAs are subject to distinct activity-dependent control mechanisms that operate at multiple levels, including transcription, pri-miR-processing, miRISC remodeling, miRNA turnover and function (Krol et al. 2010b; Treiber et al. 2012; Aksoy- Aksel et al. 2014). The various regulatory mechanisms of the miRNA pathway are either accomplished by post-translational modifications of protein core components of the miRNA pathway or RNA-binding proteins. The latter can either have enhancing or decreasing modulatory effects on miRNA function (reviewed by Loffreda et al. 2015). Activity-dependent regulatory processes were described for several individual miRNAs at the synapse. For example, miR-132 was shown to be increased by different treatments that enhance neuronal activity, such as BDNF-, KCL- or bicuculline-stimulation of neurons in vitro (Vo et al. 2005; Wayman et al. 2008) or pilocarpine-induced seizures in rodents in vivo (Nudelman et al. 2010). The BDNF-mediated induction of the miR-212/132 locus occurs through activation of ERK1/2-signaling at the transcriptional level via two cAMP response elements (CRE) within the promoter region which are bound by the transcription factor CREB (CRE-binding protein) (Remenyi et al. 2010). Recently, miR-132 expression was found to be induced more physiologically in response to learning and memory formation in mice (Hansen et al. 2013). INTRODUCTION 19 The miR-134 containing cluster miR-379-410 is another example of an activity- dependent miRNA gene locus. The expression of miR-379-410 is induced upon neuronal stimulation with BDNF or KCl via activation of the transcription factor Mef2 (myocyte enhancer factor 2) that is associated with an upstream located regulatory element (Schratt et al. 2006; Fiore et al. 2009). Interestingly, the repressive activity of miR-134 within dendritic spines can be reversed by the activation of local BDNF signalling, possibly through activation of the mTOR pathway (Figure 4). This provided the first example of a local mechanism that controls miRNA function at the level of dendritic spines (Schratt et al. 2006). However, the underlying mechanisms are yet unknown. Figure 4: BDNF-induced relief of miR-134-mediated Limk1 repression (adapted from Fiore et al. 2008) Within dendrites, the local Limk1 mRNA translation is repressed by active miR-134 which leads to an inhibition of spine growth. Upon extracellular stimulation with the neurotrophin BDNF, the miR-134- mediated repression is relieved and Limk1 translation is induced resulting in spine growth. BDNF release possibly induces the activation of the mTOR pathway that could lead to miRISC inactivation by protein phosphorylation. Furthermore, there are indications that activity-dependent regulation also occurs at the level of pre-miR-processing. A recent study showed that neuronal stimulation with N- Methyl-D-aspartate (NMDA) which activates NDMA receptors changes the miRNA expression profile in cultured hippocampal neurons and also expression levels of the microprocessor components DGCR8 and Drosha (Kye et al. 2011). Another study linked INTRODUCTION 20 pre-miRNA processing to changes in neuronal activity. It was found that in response to neuronal activity inactive Dicer is partially cleaved in a calpain-dependent manner. This in turn results in an increased RNAse III activity of Dicer at postsynaptic densities (Lugli et al. 2005). In addition, recent work demonstrated that BDNF stimulation promotes Dicer stabilization via phosphorylation of the cofactor TRBP (Huang et al. 2012). A study in retina neurons revealed that miRNA levels were dynamically increased or decreased in response to light and dark conditions, respectively (Krol et al. 2010a). A fast turnover of miRNAs was also observed in non-retinal cultured neurons, but not in non- neuronal cells. It was further shown that inhibition of neuronal action potentials by tetrodotoxin (TTX) blocks the rapid turnover of miRNAs and that glutamate receptor stimulation induces miRNA decay. This suggests that the high turnover rate of neuronal miRNAs is determined by neuronal activity (Krol et al. 2010a). The molecular mechanisms underlying the regulation of activity-dependent miRNA turnover are still not known. 1.4.2 MiRNA regulation at the level of miRISC The function of miRISC is regulated via modification of the core components, Ago and GW182, as well as via additional proteins that either positively or negatively modulate miRISC activity. Many miRISC-interacting proteins that enhance miRNA function belong to the family of DExD/H helicases, as for example Ddx6/RCK/p54 and MOV10/Armitage. RNA helicases support miRISC activity by RNA unwinding that can facilitate the incorporation of miRNAs into the miRISC as well as the interaction of miRNAs with their target mRNAs. Especially, the DEAD-box RNA helicase DDX6 is an essential miRISC component that was shown to directly interact with Ago2 thereby promoting miRNA-mediated translational repression (Chu et al. 2006). Moreover, Ddx6 was reported to affect miRNA-mediated silencing of target mRNAs by direct interaction with the CCR4-NOT1 deadenlyase complex (Rouya et al. 2014). MOV10/Armitage represents an important example of activity-dependent modification of miRISC components. The drosophila protein Armitage as well as the human homolog MOV10 (Moloney leukemia virus 10) are localized to the synapse and degraded in an activity-dependent manner by NMDA receptor stimulation. Loss of Armitage/ MOV10 INTRODUCTION 21 stimulates the translation of synaptic mRNAs, such as Limk1 and Lypla1 (Ashraf et al. 2006; Banerjee et al. 2009). Other RNA-binding proteins that facilitate miRNA-mediated repression are for example FMRP and PUF proteins. FMRP was first shown to associate with miRISC in Drosophila (Caudy et al. 2002). A later study demonstrated that FMRP cooperates with distinct miRNAs in the silencing of synaptic mRNAs in the mouse brain (Edbauer et al. 2010). Moreover, FMRP is involved in a mechanism that regulates mRNA translation at synapses in response to receptor activation. FMRP phosphorylation promotes the formation of an AGO2-miR-125a inhibitory complex on PSD-95 mRNA. Upon mGluR signaling the repressive activity of miR-125a can be relieved by dephosphorylation of FMRP. This leads to a release of AGO2 from the PSD-95 mRNA and a concomitant increase in PSD-95 translation (Muddashetty et al. 2011). The FMRP phosphorylation status also regulates the repressive activity of miR-196a, however in this case it does not affect the association between FMRP and the miR-196a target mRNA. Furthermore, FMRP directly interacts with AGO2 via a specific binding pocket within the Ago2 MID domain (Li et al. 2014). A recent study suggested a cooperative association between FMRP and MOV10. It was shown that FMRP directly interacts with MOV10 in an RNA- dependent manner and facilitates the association of MOV10 with RNAs in mouse brain (Kenny et al. 2014). Similar to FMRP the PUF proteins (e.g. Pum1/2 in mammals) enhance the miRNA repressive function. PUF binding motifs were found to be enriched in target mRNAs around miRNA binding sites suggesting an association between PUF and the miRNA regulatory system (Galgano et al. 2008). The ubiquitously expressed human Pumilio homolog 1 (PUM1) protein was shown to be phosphorylated after growth factor stimulation and to subsequently modulate the secondary structure of miRNA target mRNAs. Thereby, PUM1 promotes the accessibility of miRNA seed sequences in the 3’UTR of target mRNAs (Kedde et al. 2010). A later study with PUF proteins from nematode and humans revealed a function in the repression of mRNA translation in cooperation with Ago. Thereby, PUF forms a complex with Ago proteins and a core translation elongation factor, eEF1A, which leads to attenuation of translational elongation of target mRNAs (Friend et al. 2012). INTRODUCTION 22 RBPs that negatively regulate miRISC function have also been reported, for example Dnd1 and HuR. Dnd1 relieves miRNA repression in human germline cells by blocking miRNA interaction with the 3’UTR of target mRNAs (Kedde et al. 2007; Kedde et al. 2010). HuR associates with the 3’UTR of mRNAs close to the miRNA binding site and promotes the dissociation of miRISC from the target mRNA (Kim et al. 2009; Kundu et al. 2012). In addition, several other proteins (e.g. Rbm4, Smd1) were shown to be involved in the modulation of miRISC activity besides there originally described function in other cellular processes that are not related to miRNAs (Lin et al. 2009; Xiong et al. 2015). 1.5 Aim of the study RNA-binding proteins (RBPs) that are not core components of miRISC have been shown to modify miRNA activity in non-neuronal cells (Fabian et al. 2012). However, miRNA modulating functions of RBPs in neurons are poorly described. This study aimed to identify new modulators of miRNA activity in neurons. Previously, our laboratory investigated the role of RBPs on miRNA activity in neurons using a large-scale RNAi- based screen in primary neuron cultures (Gabriele Siegel, PhD thesis, 2011). In my PhD thesis it was the goal to validate a subset of the positive candidates obtained from the initial screen. Next, I aimed to choose one positive candidate of the secondary screen for a detailed characterization of its potential regulatory role in miRNA function. In this regard, I intended to analyze a putative involvement in the miRNA-mediated inhibition of target mRNAs. In addition, I planned to examine the molecular mechanisms underlying the function of this candidate with respect to a potential association with miRISC. Finally, I aimed to investigate whether this RBP is involved in the activity-dependent regulation of miRNA function in neurons. MATERIALS AND METHODS 23 2 Materials and Methods 2.1 Materials 2.1.1 Chemicals and reagents Agarose Biozym Ampicillin Sigma-Aldrich (2R)-amino-5-phosphonovaleric acid (AP5) Tocris Aqua-Poly/Mount Polysciences LB-Agar Sigma-Aldrich LB-Broth Carl Roth β-Mercaptoethanol Sigma-Aldrich N,N-Bis-(2-Hydroxyethyl)-2-Aminoethane Sulfonic Acid (BES) FA Applichem Bovine serum albumin (BSA) New England Biolabs Brain derived neurotrophic factor (BDNF) Peprotech Bromophenol blue Sigma-Aldrich Calcium chloride (CaCl2) Carl Roth 1,4-Dithiothreitol (DTT) Sigma-Aldrich DNA Gel Loading Dye Thermo Scientific Dulbecco's Phosphate-Buffered Saline (DPBS) Life Technologies Ethylenediaminetetraacetic acid (EDTA) Sigma-Aldrich Ethidium bromide (EtBr) Carl Roth 5-fluorodeoxyuridine (FUDR) Sigma-Aldrich Glucose Sigma-Aldrich Hydrochloric acid (HCL) Carl Roth 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) Sigma-Aldrich Hoechst Invitrogen Kanamycine Sigma-Aldrich Laminin BD Biosciences LB Broth Sigma-Aldrich Lipofectamine 2000 Invitrogen Magnesium acetate (MgAc) Fisher Scientific Magnesium chloride (MgCl2) Carl Roth Methanol VWR Chemicals MATERIALS AND METHODS 24 Milk powder Carl Roth NP-40 (Igepal) Sigma-Aldrich OptiPrep™ Density Gradient Medium Sigma-Aldrich Paraformaldehyde Carl Roth Poly-L-Lysine Sigma-Aldrich Poly-L-Ornithine Sigma-Aldrich Potassium acetate (KAc) Carl Roth Potassium chloride (KCl) Carl Roth Potassium hydroxide (KOH) Carl Roth Precision Plus Protein Dual Color Standard Bio-Rad Sodium chloride (NaCl) Carl Roth Sodium deoxycholate (NaDOC) Sigma-Aldrich Sodium dodecyl sulfate (SDS) Carl Roth Sucrose Carl Roth Tris(hydroxymethyl)aminomethane (Tris) Carl Roth Triton X-100 Carl Roth Tween 20 Carl Roth yeast tRNA Sigma-Aldrich 2.1.2 Enzymes and inhibitors Restriction enzymes AscI, EcoRI, HindIII New England Biolabs BamHI , BsrGI, DpnI, MssI, NotI, SacI, XbaI Thermo Scientific Other enzymes Benzonase® Nuclease Sigma-Aldrich DNA Polymerase I, Large (Klenow) Fragment New England Biolabs PfuPlus! DNA Polymerase Roboklon RNase A Ambion Shrimp alkaline phosphatase Promega T4 DNA ligase Invitrogen Turbo DNase Ambion Inhibitors Complete Protease inhibitors Roche MATERIALS AND METHODS 25 Phosphatase inhibitors cocktail 2 and 3 Sigma-Aldrich SUPERase inhibitor Ambion 2.1.3 Lab equipment Amicon Ultra-15 centrifugal filter with Ultracel-100 membrane Merck Millipore Cell culture dishes (15 cm2) Corning Cell culture flasks (25, 75, 175 cm2) Sarstedt Cell culture plates (24-well, 6-well) Corning Cell-lifter Corning Centrifuge JC-MC Beckman Coulter Coverslips Carl Roth Curix 60 Tabletop processor Agfa Digital gel documentation system E-Box Vilber Dounce potter (15 ml) Sartorius Eppendorf reaction tubes (1.5 ml/2 ml) Eppendorf GloMax R96 Microplate Luminometer Promega Hyperfilm ECL GE Healthcare Intell Mixer RM-2L Elmi LSM 5 Pascal microscope Zeiss Magnetic Particle Concentrator DynaMagTM -2 Invitrogen Microscope slides neoLab Microtiter plate (96-well) Nunc Mini-PROTEAN® Tetra Vertical Electrophoresis Cell Bio-Rad Nano Drop 2000c Thermo Scientific Needle (26G) B.Braun Protein G beads Sigma-Aldrich Protein G Dynabeads® Life technologies PVDF membrane Millipore Rotor type 70 Ti Beckman Coulter Sonifier 250 Branson Step One Plus instrument Applied Biosystems Thermal cycler S1000TM Bio-Rad Tubes type 70 Ti Beckman Coulter Ultracentrifuge Optima XE-90 Beckman Coulter MATERIALS AND METHODS 26 2.1.4 Kits Amersham ECL Prime Western Blotting Detection Reagent GE Healthcare Dual-Luciferase® Reporter Assay System Promega iScript cDNA synthesis Kit Bio-Rad iTaq SYBR Green Supermix with ROX Bio-Rad mirVanaTM Isolation Kit Ambion NucleoBond® Xtra Midi Kit Macherey-Nagel Pierce™ BCA Protein Assay Kit Thermo Scientific QIAquick Gel Extraction Kit Qiagen QIAquick Purification Kit Qiagen 2.1.5 Buffers and solutions 2x HBS: 50 mM HEPES 10 mM KCl 12 mM Glucose 280 mM NaCl 1.5 mM Na2HPO4 pH 7.05 2x BBS: 50 mM BES 1.5 mM Na2HPO4 280 mM NaCl pH 6.95 Tris/NaCl: 50 mM Tris 150 mM NaCl pH 8.0 PBS-MK: 1 x DPBS 1 mM MgCl2 2.5 mM KCl 5x Annealing buffer: 100 mM KAc 30 mM HEPES-KOH, pH 7.4 2 mM MgAc 50x TAE buffer: 2 M Tris-acetate 50 mM EDTA KCM solution: 100 mM KCL 30 mM CaCl2 50 mM MgCl2 MATERIALS AND METHODS 27 HLB: 10 mM KCL 1 mM EDTA 1 mM DTT 10 mM Tris-HCl pH 7.5 10x RIPA: 0.5 M Tris pH 8.0 1.5 M NaCl 10% Triton X-100 5% NaDOC 0.5% SDS 20 mM EDTA lysis buffer (protein extraction): 150 mM NaCl 50 mM Tris pH 7.5 1% Triton X-100 4x Laemmli sample buffer: 250 mM Tris-HCl (pH 6.8) 8% SDS 40% glycerol 8% β-mercaptoethanol 0.04% bromophenol blue TBS-T: 50 mM Tris 150 mM NaCl 0.1% Tween 20 pH 7.5 Running buffer: 25 mM Tris 192 mM glycine 0.1% SDS Blotting buffer: 25 mM Tris 192 mM glycine 20% methanol lysis buffer (co-IP): 20 mM Tris pH 7.5 150 mM NaCl 1% NP-40 2 mM EDTA lysis buffer (RNA-IP): 10 mM HEPES, pH 7.5 200 mM NaCl 30 mM EDTA 0.5% Triton X-100 0.5 U/µl SUPERase inhibitor GDB: 20 mM Sodium phosphate buffer, pH 7.4 450 mM NaCl 0.3% Triton X-100 0.1% gelatine MATERIALS AND METHODS 28 2.1.6 Cells and culture media Prokaryotic cells Escherichia coli (E. coli) (XL1-Blue) LB medium: LB-Broth LB Agar plates: 1.5% LB-Agar in LB medium Eukaryotic cells Human embryonic kidney 293 (HEK293) cells Growth medium: Minimum essential medium (MEM) Gibco 10% fetal calf serum (FCS) Invitrogen 2 mM L-Glutamine Invitrogen 100 U/ml Penicillin/Streptomycin Invitrogen Primary neuronal cells prepared from embryonic day 18 Sprague Dawley rats (Charles River Laboratories) Growth medium (NB+): Neurobasal medium (NB) Invitrogen 2% B27 supplement Invitrogen 1 mM GlutaMax Invitrogen 100 U/ml Penicillin/Streptomycin Invitrogen 2.1.7 Oligonucleotides Oligonucleotide sequences are notated from 5’ to 3’ direction. Primers used to clone 3’UTRs to pGL4: LIN41-FW GATCTCTAGACACTTTCTTCTTGCTCTTTAC LIN41-REV GATCGCTAGCTTTATTCCAATTATGTTATCAG LIMK1-FW CTTCTAGAGATACTTGGAGGATAGACCCTCACC LIMK1-REV GCCCCGACTCTAGCTAGCGGGAGCACAGAATTGAT Primers used for site directed mutagenesis: Nova1R-FW GATTTTTATCCAGGTACCACCGAGCGGGTTTGCTTGATCCAGG Nova1R-REV CCTGGATCAAGCAAACCCGCTCGGTGGTACCTGGATAAAAATC MATERIALS AND METHODS 29 Oligonucleotides used for cloning rAAV-Nova1 shRNA: Nova1-shRNA-fw1 GATCCCCGGTACTACTGAGAGGGTTTTCAAGAGA Nova1-shRNA-fw2 AAACCCTCTCAGTAGTACCTTTTTA Nova1-shRNA-rev1 AGCTTAAAAAGGTACTACTGAGAGGGTTTTCTCTTGAA Nova1-shRNA-rev2 AAACCCTCTCAGTAGTACCGGG Primers used for cloning of NHA-fusion constructs: NHA-Nova1-Xba1-FW TCTCTAGAATGGCGGCAGCTCCCATTCAGCAGAACG NHA-Nova1-Not1-REV CCGGTGGCGGCCGCGTCAACCCACTTTCTGAGGATTGGCAG NHA-Nova1-del1-Not1-REV CTGTGGATCCTCTGCGGCCGCCTGCTAGATAAGTTCAACAG NHA-Nova1-del2-Xba1-FW TGAATCTAGAATCCAGAAGATACAAGAGGATCCACAGAGTG NHA-EGFP-FW CCACCGGTCGACACCATGGTGAGCAAGGGCG NHA-EGFP-REV ATCTAGAGTCGCGGCCGCTTTACTTGTAC Primers used for qPCR: U6 snRNA FW CTCGCTTCGGCAGCACA U6 snRNA REV AACGCTTCACGAATTTGCGT GAPDH FW GCCTTCTCTTGTGACAAAGTGGA GAPDH REV CCGTGGGTAGAGTCATACTGGAA Limk1 FW CCTCCGAGTGGTTTGTCGA Limk1 REV CAACACCTCCCCATGGATG Rgs4 FW ACAAGCCGGAACATGTTAGAG Rgs4 REV AGACTTGAGGAAACGACGGT 2.1.8 RNA molecules Table 1: MiRNA sense and antisense sequences miRNA Sequence (5’ – 3’) miR-134 sense Phospho-UGUGACUGGUUGACCAGAGGGA miR-134 antisense Phospho-CCUCUGGUCAACCAGUUAUACU miR-138 sense Phospho-AGCUGGUGUUGUGAAUCAGGCCG miR-138 antisense Phospho-GCCUGAUUCACAACACCAGAUUU MiRNA sense and antisense oligos (Table 1) (Integrated DNA Technologies) were used for miRNA duplex annealing (chapter 2.3.1). MATERIALS AND METHODS 30 Table 2: SiRNA subset of custom siRNA library (Ambion) siRNA siRNA sense sequence siRNA antisense sequence Ewsr1 siRNA-1 GGAUAUGCACAGACCACCCtt GGGUGGUCUGUGCAUAUCCtt Ncoa3 siRNA-2 GGAGACAGUGAGACAGAUAtt UAUCUGUCUCACUGUCUCCtt Nova1 siRNA-3 GGUACUACUGAGAGGGUUUtt AAACCCUCUCAGUAGUACCtg Nxf1 siRNA-1 CGAUUUCCCAAGUUAUUACtt GUAAUAACUUGGGAAAUCGtt Nxt1 siRNA-2 GCCACUUUAGUAUGGAAUGtt CAUUCCAUACUAAAGUGGCtg Rbm25 siRNA-2 GGACAUUUUCCGUAGAUUUtt AAAUCUACGGAAAAUGUCCtc Rpl5 siRNA-3 GGAGAUGUAUAAGAAAGCUtt AGCUUUCUUAUACAUCUCCtc Tnrc6c siRNA-3 GGUUCAAGCACAGCUUUUGtt CAAAAGCUGUGCUUGAACCtg U2af1 siRNA-2 CCUUUAGCCAGACCAUUGCtt GCAAUGGUCUGGCUAAAGGtt 2.1.9 Plasmids Luciferase reporter plasmids The pGL4.13 vector (Promega) encodes the luciferase reporter gene luc2 (firefly). The vector contains an SV40 early enhancer/promoter. The firefly reporter constructs were generated by PCR amplification and/or restriction digest of the corresponding 3’UTRs (Table 3) from pGL3-3’UTR reporter constructs. The 5’overhangs of the resulting inserts were filled by Klenow reaction (chapter 2.3.5). The 3’UTRs were inserted downstream of the firefly gene into the XbaI restriction site of pGL4.13. The 138-sponge-luc reporter was generated by inserting six binding sites for miR-138 carrying a bulge at position 9- 12 (Ebert et al. 2007) into pGL4.13 vector. The control (ctrl) sponge was created by inserting the reverse sequence of the 138 sponge into pGL4.13. In order to generate the renilla luciferase vector pGL4-RL, the renilla luciferase gene was excised by restriction digest from the plasmid phRL-TK and inserted into pGL4.13 by replacing the firefly coding sequence. MATERIALS AND METHODS 31 Table 3: Luciferase reporter plasmids Luciferase reporter plasmid 3’UTR / Reference Restriction sites used for cloning 3’UTR pGL4-UBE3A UBE3A rat (606 bp) (Valluy et al. 2015) BamHI, XbaI pGL4-APT1 APT1 mouse (Siegel et al. 2009) AscI, EcoRI pGL4-LIN41-WT LIN41 C. elegans (D. Bartel) NotI, MssI pGL4-LIN41-mutant LIN41 C. elegans (D. Bartel) NotI, MssI pGL4-LIMK1-WT LIMK1 rat (Schratt et al. 2006) XbaI, NheI pGL4-LIMK1-mutant LIMK1 rat (Schratt et al. 2006) XbaI, NheI pGL4-IQGAP1-WT IQGAP1 human (D. Bartel) SacI, XbaI pGL4-IQGAP1-mutant IQGAP1 human (D. Bartel) SacI, XbaI pGL4-HMGA2-WT HMGA2 mouse (D. Bartel) XbaI, NotI pGL4-HMGA2-mutant HMGA2 mouse (D. Bartel) XbaI, NotI pGL4-138-sponge 6x miR-138 binding site (S. Bicker, K. Weiß) BsrGI, HindIII pGL4-ctrl-sponge 6x miR-138 binding site reversed (S. Bicker, K. Weiß) BsrGI, HindIII ShRNA expression plasmids In order to generate shRNA expression vector pAAV-Nova1 shRNA, DNA oligonucleotides (chapter 2.1.7) were annealed followed by ligation into the BglII and HindIII restriction sites of the pAAV vector (Christensen et al. 2010). The pAAV vector contains inverted terminal repeats (ITRs) that flank a U6 promoter driven shRNA sequence upstream of a chicken β-actin (CBA) promotor driven GFP gene. Table 4: ShRNA expression plasmids shRNA expression plasmid Reference pSuper basic Oligoengine pSuper-control shRNA Störchel and Thümmler et al. 2015 pSuper-Nova1 shRNA Störchel and Thümmler et al. 2015 pAAV-control shRNA pAM/U6-sHRNA-EGFP-CBA-hrGFP (obtained from M. Schwarz) pAAV-Nova1 shRNA generated from pAAV-control shRNA MATERIALS AND METHODS 32 GFP-Nova1 overexpression vector To generate the vector expressing an shRNA-resistant Nova1 protein (Nova1R), silent mutations were introduced into the GFP-Nova1 expressing vector pEGFP-C1-Nova1 (G. Siegel) by site directed mutagenesis (chapter 2.3.8) using the primer Nova1R-FW and Nova1R-REV resulting in the vector pEGFP-C1-Nova1R. Table 5: GFP-Nova1 overexpression plasmids GFP-Nova1 overexpression plasmid Reference pEGFP-C1 Clontech pEGFP-C1-Nova1 Störchel and Thümmler et al. 2015 pEGFP-C1-Nova1R generated from pEGFP-C1-Nova1 Plasmids for tethering assay To generate pCI-neo-NHA-Nova1 and pCI-neo-NHA-GFP, the respective coding sequences were amplified by PCR from the vector pEGFP-C1-Nova1 and pEGFP-N1. Subsequently, the PCR product was inserted into the vector pCI-Neo-NHA downstream of the N-HA peptide sequence by use of the restriction sites that are indicated in the primer name (chapter 2.1.7). Table 6: Plasmids used for tethering assays Tethering assay plasmid Reference RL-5boxB obtained from R. Pillai pCI-neo-NHA obtained from R. Pillai pCI-neo-HA-Tnrc6c obtained from R. Pillai pCI-neo-NHA-Tnrc6c obtained from R. Pillai pCI-neo-NHA-Nova1 generated from pCI-Neo-NHA pCI-neo-NHA-GFP generated from pCI-Neo-NHA pCI-neo-NHA-Ncoa3 Störchel and Thümmler et al. 2015 MATERIALS AND METHODS 33 2.1.10 Antibodies Table 7: Antibodies used for Western blot (WB), immunocytochemistry (ICC), RNA-IP and co-IP Antibody Species Source Application/ Dilution Primary antibodies anti-MAP2 mouse clone HM-2, Sigma-Aldrich ICC/ 1:2,000 anti-α-Tubulin rabbit cat.# 2144, Cell Signaling WB/ 1:5,000 anti-β-Actin mouse clone AC-15, Sigma-Aldrich WB/ 1:10,000 anti-HDAC2 rabbit ab32117, Abcam WB/ 1:5,000 anti-Limk1 mouse clone 42, BD Biosciences WB/ 1:200 anti-Nova1 rabbit Merck Millipore cat.# 07- 637 WB/ 1:1,500; ICC/ 1:250 anti-Nova1 rabbit ab183723, Abcam RNA-IP anti-Nova-pan rabbit obtained from R. Darnell WB/ 1:5,000 anti-pan-Ago mouse clone 2A8, Merck Millipore co-IP/ 1:200 anti-Ago2 rat clone 6F4, obtained from G. Meister WB/ 1:10 (serum) anti-HA rabbit ab9110, Abcam ICC/ 1:1,000 rabbit IgG normal Santa Cruz Biotechnology RNA-IP mouse IgG normal Santa Cruz Biotechnology co-IP Secondary antibodies anti-mouse HRP rabbit Calbiochem WB/ 1:20,000 anti-rabbit HRP goat Calbiochem WB/ 1:20,000 anti-rat HRP goat Calbiochem WB/ 1:20,000 Alexa 488 anti-rabbit goat Life technologies ICC/ 1:1,000 Alexa 647 anti-mouse donkey Life technologies ICC/ 1:1,000 Alexa 546 anti-mouse goat Life technologies ICC/ 1:1,000 2.2 Cell culture methods 2.2.1 Primary neuronal cell culture Dissociated primary cortical and hippocampal neurons were prepared from embryonic day 18 rats as described in Schratt et al. (2004). After dissociation cortical neurons were MATERIALS AND METHODS 34 directly plated on Poly-L-Ornithine (Sigma) coated 24-well or 6-well cell culture plates. Hippocampal neurons were plated on nitric acid-treated coverslips that were coated with Poly-L-Lysine and Laminin in a 24-well format. Both types of primary neurons were cultured in Neurobasal medium (NB+) at 37 °C and 5% CO2. Primary neurons were routinely prepared by Gertraud Jarosch, Eva Becker, Renate Gondrum and Heinrich Kaiser. For the analysis of subcellular fractions, hippocampal neurons were cultured in medium supplemented with 10 µM of 5-fluorodeoxyuridine (FUDR) to prevent the growth of glial cells. For BDNF stimulation, cortical neurons were treated with BDNF at a final concentration of 100 ng/ml for 5 min to 5 h before cell lysis. 2.2.2 Human embryonic kidney 293 cell culture Human embryonic kidney 293 (HEK293) cells were cultivated in 6-well- or 12-well- culture plates at subconfluent density in minimum essential medium (MEM) supplemented with FCS, L-Glutamine and Penecillin/Streptomycin at 37 °C and 5% CO2. 2.2.3 Transfection of neurons with Lipofectamine 2000 Cultured neuronal cells were transfected with the Lipofectamine 2000 transfection reagent. Per well (24-well format), a total of 1 µg of RNA or DNA was mixed with a 1:66 dilution of Lipofectamine 2000 in 100 µl Neurobasal medium (NB, without supplements). The mixture was incubated at room temperature (RT) for 20 min, then diluted 1:5 in NB medium and gently applied to the cells. After 2 h of incubation at 37 °C and 5% CO2, the cells were washed with NB. Subsequently, the cells were incubated with the NMDA- receptor antagonist (2R)-amino-5-phosphonovaleric acid (AP5, 20 µm) in NB+ for 45 min at 37 °C. After an additional washing step with NB the cells were provided with one third of fresh NB+ medium mixed with two third of medium cells were incubated in before (conditioned medium) that was collected before the transfection procedure. 2.2.4 Transfection of HEK293 cells with calcium-phosphate HEK293 cells were transfected at a confluency of 60% using calcium-phosphate. Per well (12-well format) a total amount of 3.2 µg of DNA was mixed with 6 µl of 2 M CaCl2 in H2O (total volume of 50 µl). Then 50 µl of 2x HBS were added to the mixture while MATERIALS AND METHODS 35 slowly vortexing. The mix was incubated for 3 min at RT and gently applied drop-wise to the cells medium. After 5 h incubation at 37 °C and 5% CO2 the medium was replaced with fresh MEM. The cells were lysed after incubation for 18-24 h. 2.2.5 Production of recombinant adeno-associated virus (rAAV) Recombinant adeno-associated viruses (rAAV) expressing either Nova1 shRNA or control shRNA were generated by co-transfection of the shRNA expressing pAAV plasmid and the helper plasmids pDP1 and pDP2 (Grimm et al. 2003) into HEK293 cells using calcium-phosphate. The cells were plated onto 15 cm2 cell culture dishes and transfected at 80% confluency. The DNA mix was prepared for a total of 8 dishes (96 µg pAAV, 192 µg pDP1, 192 µg pDP2) and mixed with 2.5 ml 2 M CaCl2 and 17.5 ml H2O. After splitting the solution to 8 vessels (2.5 ml each) 2.5 ml of 2x BBS were added. The solution was gently mixed by inverting the tube and incubated at RT for 3 min. 20 ml of MEM were added and the mixture was applied to the cells after removing the complete culture medium. Following 5 h of incubation at 37 °C and 5% CO2, the cells were washed with 25 ml fresh MEM. After 3 days, the cells were harvested with a cell-lifter and collected in 4 tubes. The pelleted cells from 8 plates were united in 25 ml DPBS and resuspended in 8.5 ml Tris/NaCl. 2 µl of Benzonase (25 U/µl) were added followed by 500 µl of 10% NaDOC. After incubation for 30 min at 37 °C 584 mg NaCl were dissolved in the cell solution followed by 30 min of incubation at 56 °C. The cells were frozen overnight at -20 °C, thawed and centrifugated for 30 min at 16,000 g. After another freeze-thaw cycle the virus was purified by discontinuous iodixanol gradient centrifugation (Zolotukhin et al. 1999). The sample was loaded onto the gradient (7 ml of 15%, 4 ml of 25%, 3 ml of 40% and 3 ml of 54% iodixanol) in 70 Ti rotor tubes and spun in the ultracentrifuge for 2 h at 55,000 rpm (rotor type 70 Ti). To recover the virus, the 40% phase and interphase between 40% and 54% were aspirated with a syringe and 1:1 diluted in PBS-MK. The virus was concentrated through filter (Amicon Ultra-15 centrifugal filter with Ultracel-100 membrane) by centrifugation at 2,000 g. 2.2.6 rAAV infection of neurons In order to yield a high number of shRNA expressing cells in the neuronal culture, cells were infected with shRNA expressing recombinant adeno-associated virus (rAAV). Cortical neurons were infected at 4 days in vitro (DIV) with shRNA expressing rAAV MATERIALS AND METHODS 36 (2*106 IFU (infectious units)/ml). Virus infected cells were lysed 10 days after infection (14 DIV). 2.3 Molecular methods 2.3.1 Annealing of miRNA duplex MiRNA duplex RNA was synthesized by annealing single stranded miRNA sense and miRNA antisense strands (Table 1). 20 µl of each strand (100 µM) were mixed with 10 µl 5x Annealing buffer. The mixture was incubated in the thermal cycler for 2 min at 95 °C, for 1 h at 37 °C and cooled down to 4 °C. 2.3.2 Polymerase chain reaction (PCR) The amplification of specific DNA fragments for cloning was performed by polymerase chain reaction (PCR) using the PfuPlus! DNA Polymerase. In accordance with the manuals’ instruction, PCR reactions were set up as follows and incubated in a thermal cycler with the following program. PCR reaction (50 µl): 5 µl 10x Pfu buffer 4 µl 2.5 mM dNTP mix 2 µl 10 µM forward primer 2 µl 10 µM reverse primer 1 µl plasmid DNA (50 ng) 0.5 µl PfuPlus! DNA polymerase (2.5 U) 35.5 µl ddH2O PCR program: Initial denaturation 95 °C, 3 min Denaturation 95 °C, 50 sec Annealing 50 - 68 °C, 30 sec 35 x Extension 72 °C, 1 min/kb Terminal extension 72 °C, 7 min 4 °C, ∞ 2.3.3 Agarose gel electrophoresis The separation of DNA fragments according to their size was performed by agarose gel electrophoresis. The DNA samples were mixed with DNA Gel Loading Dye (Thermo ScientificTM) and loaded onto a 1% agarose gel in 1x TAE buffer containing ethidium bromide (1:200). The gel was run at 120 V. The bands were visualized by UV-light in a digital gel documentation system. MATERIALS AND METHODS 37 2.3.4 Purification of plasmid DNA and PCR products Restricted DNA fragments were separated by agarose gel electrophoresis and the respective bands were excised from the gel followed by purification using the QIAquick Gel extraction kit. PCR products were purified by use of the QIAquick purification kit according to the manuals’ instructions. 2.3.5 Restriction digest, dephosphorylation, Klenow reaction and ligation of DNA For cloning of recombinant DNA constructs, plasmid DNA or PCR-products were digested with restriction enzymes (chapter 2.1.2) according to the manuals’ instructions. In order to generate blunt ends of fragments, 5’-overhangs were filled up by Klenow reaction. DNA polymerase I, large (Klenow) fragment was used according to the manufactures’ protocol (1U/µg DNA). To prevent self-ligation of the linearized vector, DNA was dephosphorylated using shrimp alkaline phosphatase to prevent self-ligation. The ligation of DNA fragments into linearized vector DNA was performed using the T4 DNA ligase in accordance with the manuals’ instructions. The dephosphorylated, linearized vector DNA was mixed with restricted insert DNA at a ratio of 1:3 for sticky end ligation or 1:5 for blunt end ligation. 2.3.6 Transformation of competent bacteria Transformation of chemical competent E. coli cells was performed by heat shock. Cells were thawed on ice and 1:1 diluted in KCM solution. 100 µl of the cell solution were mixed with 10 ng plasmid DNA or 1 µl ligation mixture. Following heat shock for 45 sec at 42 °C, cells were chilled on ice for 10 min. In the case of ampicillin-resistant plasmids, the transformed cells were directly plated onto LB plates containing 100 µl/ml ampicillin. In case of kanamycin-resistant plasmids, 1 ml of pre-warmed antibiotic-free LB was added to the cell suspension and the mixture was shaken for 1 h at 37 °C. The cells were centrifuged for 3 min at 1,200 g and the pellet was resuspended in 100 µl LB. The suspension was plated on LB plates containing 50 µl/ml kanamycin. The plates were incubated over night at 37 °C. MATERIALS AND METHODS 38 2.3.7 Plasmid midi preparation Plasmids for midi preparation were isolated from transformed E. coli. A single bacteria colony was inoculated in 200 ml of antibiotic supplemented LB medium and grown over night at 37 °C. The cell suspension was pelleted by centrifugation at 4 °C. The preparation of plasmids was performed using the NucleoBond® Xtra midi kit according to the manuals’ instructions. Plasmid concentrations were measured with a Nano Drop 2000c. 2.3.8 Site-directed mutagenesis The introduction of point mutations in the DNA sequence of a plasmid was generated by PCR, based on the manufacturer’s instructions of the Quick Change® Site directed mutagenesis kit (Stratagene). The PCR reaction was set up as follows and incubated in a thermal cycler with the following program. PCR reaction (50 µl): 5 µl 10x Pfu buffer 4 µl 2.5 mM dNTP mix 1,25 µl forward primer (125 ng) 1,25 µl reverse primer (125 ng) 1 µl plasmid DNA (12.5 ng) 0.5 µl PfuPlus! DNA polymerase (2.5 U) 37 µl ddH2O PCR program: Initial denaturation 95 °C, 1 min Denaturation 95 °C, 30 sec Annealing 55 °C, 1 min 18 cycles Extension 68 °C, 1 min/kb Terminal extension 68 °C, 7 min To cleave the template plasmid, the PCR reaction was then digested with DpnI (10 U) for 1 h at 37 °C. 1 µl of digested PCR product was transformed into bacteria. 2.3.9 cDNA synthesis and RT-qPCR Complementary DNA (cDNA) was synthesized from purified DNA-free RNA samples by reverse transcription. 1 µg total RNA was reverse transcribed using the iScript cDNA synthesis Kit and additional MgCl2. The reaction mix was incubated in a thermal cycler according to the manufactures’ instructions. The cDNA was diluted 1:5 in nuclease-free water. Real time-quantitative PCR (RT-qPCR) was processed in duplicates using the iTaq SYBR Green Supermix with ROX according to the manufactures’ protocol. The PCR run was performed on a Step One Plus instrument. MATERIALS AND METHODS 39 cDNA synthesis reaction (20 µl): 4 µl 5x iScript reaction mix 1 µl 100 mM MgCl2 1 µl iScript reverse transcriptase 14 µl RNA template cDNA reaction program: 25 °C, 5 min 42 °C, 30 min 85 °C, 5 min 4 °C, ∞ RT-qPCR reaction (20 µl): 10 µl iTaq SYBR Green supermix 0.5 µl 10 µM forward primer 0.5 µl 10 µM reverse primer 4 µl nuclease-free dH2O 5 µl cDNA RT-qPCR program: 95 °C, 3 min 95 °C, 15 sec 45 cycles 60 °C, 30 sec 2.4 Biochemical methods 2.4.1 Subcellular fractionation For the subcellular fractionation of lysates from hippocampal neurons, the cells were washed with ice-cold DPBS and detached in HLB containing protease inhibitors and 0.2 %NP-40 by shaking for 10 min at 4 °C. The cell lysis was achieved through 20 strokes with a Teflon potter. One part of the sample was taken as input and diluted with 10x RIPA buffer to 1x. For nuclear and cytosolic fractions, the lysate was centrifugated at 800 g for 5 min at 4 °C. The supernatant was recovered as cytosolic fraction. The pellet that contains the nuclei was washed once with HLB containing 0.25 M sucrose and dissolved in 1x RIPA (containing protease inhibitors). The input and nuclear sample were lysed by sonication. Subsequently, all fractions were centrifugated by 16,000 g for 10 min at 4 °C and the respective supernatants were recovered as protein extracts. The subcellular fractionation was performed by Peter Störchel. 2.4.2 Protein extraction For the preparation of whole cell lysates from neurons, cells were washed once with DPBS and resuspended in lysis buffer containing protease inhibitors. BDNF-treated neurons were lysed in lysis buffer additionally containing Phosphatase Inhibitors Cocktail 2 and 3. MATERIALS AND METHODS 40 2.4.3 Western blot analysis The protein concentration of cell extracts was determined using the Pierce™ BCA Protein Assay Kit. Equal protein amounts or lysate volumes were mixed with Laemmli sample buffer, boiled for 5 min at 95 °C and loaded onto a 10% SDS-polyacrylamide gel and separated by electrophoresis (SDS-PAGE) in running buffer. Precision Plus Protein Dual Color Standard was loaded as size standard next to the samples. The proteins were then transferred by tank blotting at 90 V and 4 °C for 100 min in blotting buffer to an Amersham Hybond PVDF (poly-vinylidene difluoride) membrane, which was prior activated in 100% methanol. To saturate unspecific binding, the membrane was incubated on a shaker for 1 h with 5% milk powder in TBS-T. The blot was then incubated with a primary antibody (Table 7) diluted in milk containing TBS-T for 2-3 h at RT or overnight at 4 °C. After three times washing for 10 min at RT with TBS-T milk, the blot was incubated with the HRP-conjugated secondary antibody (Table 7) for 1 h at RT, followed by three wash steps with TBS-T. HRP was detected using the ECL Prime Western Blotting Detection Reagent. The band intensities of scanned films were measured with ImageJ software. 2.4.4 Luciferase reporter assay Luciferase reporter assays were performed by use of the Dual Luciferase Reporter System. Therefore, hippocampal neurons (11 DIV) or cortical neurons (5 DIV, 12 DIV) were transfected in 24 well-plates with equal amounts of firefly (pGL4-3’UTR reporter constructs) and renilla luciferase (pGL4-RL) expressing vectors (Table 3). The luciferase vectors were co-transfected in duplicates as indicated with either pSuper vector (Table 4) or siRNA duplex (Table 2) for knockdown of RBPs, and in the case of miRNA overexpression with miRNA duplex RNA (miR-134 or miR-138) (Table 1). After 2-3 days the cells were washed once in 500 µl DPBS. The DPBS was replaced by 100 µl/well of 1x lysis buffer (Promega) and lysis was carried out by shaking the plate for 20 min at RT. The solutions for the firefly and renilla luciferase measurement were prepared as indicated in the manufactures instructions. After the lysis, 20 µl of each lysate were transferred to a 96-well microtiter plate. The luciferase activity of both firefly and renilla was measured using the GloMax R96 Microplate Luminometer. Firefly activity was normalized to renilla activity and a basal condition which did not contain an shRNA expression vector or an siRNA duplex was set as 1.0 RLA (relative luciferase activity). MATERIALS AND METHODS 41 Table 8: Vector and RNA duplex quantities used for transfection 2.4.5 Tethering assay Tethering experiments were performed by co-transfecting cortical neurons at 5 DIV with 150 ng pCI-Neo-NHA or 250-400 ng NHA-fusion constructs, 50 ng of pRL-5boxB and 50 ng of pGL4.13. The cell lysis and luciferase measurement was performed 2 days after transfection as described in chapter 2.4.4. The renilla activity was normalized to firefly activity. 2.4.6 Co-immunoprecipitation (co-IP) Co-immunoprecipitation (co-IP) of Ago and Nova1 was performed in lysates that were prepared from hippocampi of adult female rat brain. The hippocampi of one brain were dissected in ice-cold DPBS and transferred to a pre-chilled 15 ml dounce potter and homogenized on ice by 14 strokes in 2.5 ml lysis buffer containing protease inhibitors. The homogenate was transferred to 2 ml tubes and centrifugated for 10 min at 16,000 g at 4 °C. 50 µl of the supernatant were kept as input, the rest was used for further steps. For RNase treatment RNase A (1:20) was added to one half of the cell lysate. For immunoprecipitation of Ago 500 µl lysate were incubated with mouse anti pan-Ago (1.25 µg). As a control, 500 µl lysate were incubated with mouse IgG (1.25 µg). Antibody incubation occurred in Eppendorf tubes rotating over night at 4 °C. Protein G beads (25 µl per condition) were equilibrated in lysis buffer by rotating for 5 min at 4 °C followed by centrifugation for 2 min at 300 g (twice). Afterwards the beads were blocked for 2 h in BSA diluted in lysis buffer (1 mg/ml). Following two times washing in 1 ml lysis buffer, the beads were resuspended in 30 µl lysis buffer and incubated with lysate and antibody for 1 h at 4 °C followed by four washes in lysis buffer. The last two wash steps contained 350 mM NaCl. After final washing the supernatant was discarded and beads were mixed with sample buffer and prepared for SDS-PAGE. Plasmid/ duplex RNA Amount per well (24-well format) pGL4-3’UTR 50 – 100 ng pGL4-RL 50 – 100 ng pSuper 10 ng siRNA duplex 7.5 pmol miRNA duplex 5 pmol MATERIALS AND METHODS 42 2.4.7 RNA-immunoprecipitation (RNA-IP) The hippocampi of an adult female rat were dissected and homogenized in a dounce potter by 20 strokes in 3.5 ml lysis buffer containing complete protease inhibitors. The homogenate was then passed five times through a 26G needle. After centrifugation for 20 min at 70,000 g at 4 °C, the supernatant was supplemented with 100 µg/ml yeast tRNA. Protein G Dynabeads were equilibrated in 1 ml lysis buffer (without inhibitors) by rotating for 5 min at 4 °C. To separate the beads from the solution, the tube was placed on a magnet and the supernatant was removed. The lysate was then precleared with the beads for 30 min at 4 °C. After preclearing, 160 µl of the supernatant were collected as input. The immunoprecipitation was performed by addition of 10 µg antibody (rabbit anti- Nova1(Abcam) or normal rabbit IgG (SCBT)) and 30 µl Protein G Dynabeads to 1 ml of precleared lysate. The solution was rotating for 2 h at 4 °C. Afterwards, the beads were washed five times with lysis buffer (0.1 U/ml SUPERase inhibitors and 10 µg/ml yeast tRNA) and once with lysis buffer without additives. The beads were resuspended in 1 ml lysis buffer and divided to 200 µl (protein) and 800 µl (RNA) fractions. After pelletizing the beads with the magnet, the supernatant was discarded. For protein detection by Western Blot the beads were diluted in sample buffer and prepared for SDS-PAGE. From the other part of the beads RNA was extracted by use of the mirVanaTM Isolation Kit. RNA was eluted in 70 µl nuclease-free water followed by digestion of DNA by treatment with Turbo DNase according to the manufacturer’s protocol. After denaturation of the DNase for 10 min at 75 °C in the presence of 15 mM EDTA, RNA samples were kept on ice. For detection of specific mRNAs, cDNA was synthesized and RT-qPCR was performed (chapter 2.3.9). 2.5 Immunocytochemistry (ICC) For microscopic analysis, cultured hippocampal neurons were fixed in 500 µl 4% paraformaldehyde/sucrose in DPBS for 15 min at RT. Subsequently, the cells were washed four times in 500 µl DPBS (2x 30 sec, 2x 5 min). For the analysis of dendritic spines and dendritic growth, the GFP expressing cells were mounted on microscope slides in Aqua-Poly/Mount. For immunocytochemistry the coverslips were transferred to a wet chamber. Primary and secondary antibodies were diluted in GDB (gelatine detergent buffer). The cells were incubated with the primary antibody for 1 h at RT in the dark. MATERIALS AND METHODS 43 After four wash steps with DPBS (1x 30 sec, 3x 5 min) the secondary antibody was applied for 1 h at RT protected from light. After three further wash steps with DPBS, Hoechst was incubated at 1:5000 in DPBS for 5 min. After all, the cells were mounted as described above (chapter 2.5). 2.6 Microscopic analysis All microscopic images were taken with a Confocal Laser Scanning microscope. The presented immunocytochemistry images are the maximum projections of 2 - 4 confocal stacks with a z-distance of 0.5 – 1 µm. The measurements of average signal intensity and x-y-distance as well as further image processing were carried out with the ImageJ software. 2.7 Spine assay For dendritic spine analysis, hippocampal neurons (13 DIV) were co-transfected with 200 ng pEGFP-Amp, 5 ng of shRNA expressing pSuper construct and miR-134 duplex RNA. After 6 days (19 DIV) the cells were prepared for confocal imaging (chapter 2.5). Confocal z-stack images were taken with a 63x objective (7 z-stacks at 0.4 µm interval) and projected to a single plane image. The individual spine size was quantified by its mean grey value in a 2.2 µm2-circle using ImageJ software. More than 100 spines per cell were measured and normalized to the cell’s mean grey values. Spines from three independent experiments were measured. The imaging and analysis were carried out blinded to the experimental conditions. The spine assay was performed by Anna Tsankova under supervision of Peter Störchel. 2.8 Sholl analysis For the determination of dendritic complexity by Sholl analysis, pyramidal neurons (12 DIV) expressing GFP were imaged by fluorescence microscopy. The images were layered with a pattern of concentric circles with increasing radii, centered to the cell body (8 circles with distances from soma between 25 – 200 µm). The dendritic outgrowth was determined by quantification of intersections between dendrites and each circle. The sum MATERIALS AND METHODS 44 of all dendrite intersections at each distance represents the total number of intersections. Per condition 10 to 12 cells were analyzed. 2.9 Statistical analysis Statistical analysis for pairwise comparison was carried out by two-tailed, unpaired Student’s t-test. For multiple conditions, one-way Anova test (http://vassarstats.net/) was applied with Tukey HSD post-test comparing single conditions pairwise. Significance was set at p < 0.05. RESULTS 45 3 Results 3.1 Background – RNAi-based screen Previously, a large-scale RNAi-based screen in primary mouse cortical neuron cultures was performed in our laboratory to identify novel neuronal RBPs that modulate miRNA activity (Gabriele Siegel, PhD thesis, 2011). The study was based on the idea that the knockdown of crucial effector proteins should relieve the miRNA-mediated repression of translation that can be monitored by a luciferase reporter assay. The luciferase reporter construct contained downstream of the luciferase coding sequence the 3‘UTR of the gene Ube3a which is responsive to the neuronal miRNA miR-134 (Valluy et al. 2015). A subset of 286 RBPs that are expressed in the postnatal mouse cortex were chosen as candidate proteins (McKee et al. 2005). In the screening experiment each of the candidates was individually knocked down by transfection of synthetic siRNAs. Specifically, three different siRNAs for each RBP were separately co-transfected with the Ube3a luciferase reporter and miR-134-duplex RNA into primary mouse cortical neurons (5 days in vitro (DIV)). After three days the luciferase activity was measured (Figure 5 A). Transfection of miR-134 duplex RNA led to an about 40% repression of the Ube3a- luciferase reporter expression. In the case that an RBP was required for the inhibitory function of miR-134, we expected that the reporter repression was relieved in the presence of the respective siRNAs. Candidates were considered as positive hits if at least two out of three siRNAs could relieve the reporter repression by at least 50%. Using this criteria, 12 RBPs that are required for the function of miR-134 were identified in this study (Figure 5 B). Among the positive hits were Tnrc6c and Ddx6, two proteins that are known to be directly associated with miRISC (Chu et al. 2006; Chen et al. 2009). This confirmed that the screen setup allowed the identification of miRISC-associated factors. Importantly, ten new candidates (Ewsr1, Fubp1, Lsm7, Ncoa3, Nova1, Nxf1, Nxt1, Rbm25, Rpl5 and U2af1) were detected that had not yet been described in the context of miRISC regulation (Figure 5 B). A miRNA-independent function in RNA metabolism was previously reported for most of these RBPs. For example, Nova1, U2af1 and Rbm25 are involved in the regulation of RNA splicing (Jensen et al. 2000; Mollet et al. 2006; Zhou et al. 2008) RESULTS 46 while others influence mRNA translation (Rpl5) (Meskauskas et al. 2001) or mRNA decay (Lsm7) (Tharun et al. 2000). The RBPs Ewsr1, Fubp1 and Ncoa3 had been implicated in transcription, whereas Nxf1 and Nxt1 were shown to regulate the nuclear export of mRNA. Figure 5: Large scale RNAi-based screen in primary mouse neurons A) Schematic overview of the RNAi screen in mouse primary neurons performed by G. Siegel. Cortical Neurons (5 days in vitro (DIV)) were co-transfected with pGL3-Ube3a 3’UTR luciferase reporter, miR- 134 duplex RNA and one siRNA targeting individual RBPs. Three different siRNAs were used for every RBP. After three days of incubation luciferase activity was determined. B) List of a representative subset of genes that were tested in the screen. The dashed line designates “known genes” that are involved in the regulation of miRNA activity. Thick line marks candidate genes that relieved reporter repression by at least 50% with at least two out of three tested siRNAs. Values represent the relief of miRNA-mediated repression after siRNA treatment, obtained from three independent experiments. RESULTS 47 3.2 Validation of RNAi screen results For the validation of the RNAi-based screen results (Siegel, 2011) I performed a secondary screen using only candidates that were classified as positive hits in the first study. To exclude false positives, the experimental setup was modified in the following points. First, I used an improved luciferase reporter system that should reduce potential non- specific siRNA effects. The Ube3a 3’UTR was cloned downstream of the firefly coding sequence of the pGL4 plasmid (UBE3A-luc). Compared to the previously used pGL3 vector, pGL4 contains less regulatory elements in the vector backbone. Thus, it minimizes the possibility of non-specific siRNA effects caused by factors that can potentially interact with regulatory elements in the vector. The UBE3A-luc reporter was co- transfected with the miR-134 duplex RNA and the siRNAs for specific RBPs individually into immature rat cortical neurons (5 DIV). Furthermore, a condition was included where the reporter was co-transfected with the siRNA in the absence of miR-134 duplex. This condition should control for possible non-specific effects on luciferase mRNA translation caused by the siRNAs. As a further control, an empty luciferase vector lacking 3’UTR sequence was co-transfected with the siRNA in the presence or absence of miR-134 duplex. In the secondary screen only nine out of the 12 positive candidates (Nova1, Ncoa3, Ewsr1, Tnrc6c, Rpl5, Nxt1, Rbm25, U2af1, and Nxf1) were tested. Since the secondary screen was performed in rat, an siRNA was selected for each candidate that targets a sequence conserved between rat and mouse genomes and showed high efficacy in the primary screen. Lsm7 and Fubp1 were excluded, as none of the respective siRNAs was conserved between rat and mouse. Furthermore, Ddx6 was not included in the secondary screen since it was already shown to be involved in the regulation of miRNA activity (Chu et al. 2006). The known miRISC protein Tnrc6c was used as positive and a scrambled siRNA nucleotide (control siRNA) as a negative control in the screen. Luciferase activity was measured three days after transfection (8 DIV) and the experiment was performed three times. The results are presented as ratio between the condition that contains the miR-134 duplex RNA (“+miR”) and the condition without the miR-134 duplex RNA (“-miR”). In the basal condition where no siRNA was co-transfected as well as in the condition with control siRNA the ratio of “+miR” to “-miR” was about 0.5 (0.53 ± 0.05 (basal); 0.51 ± RESULTS 48 0.05 (control siRNA)) (Figure 6 A). This reflects a nearly 50% downregulation of the UBE3A-luc reporter expression mediated by miR-134 overexpression. Compared to control siRNA (0.51 ± 0.05) the downregulation was significantly relieved when siRNAs against Nova1 (0.87 ± 0.20), Ncoa3 (0.84 ± 0.16), Ewsr1 (0.69 ± 0.08) and Tnrc6c (0.73 ± 0.14) were co-transfected. This suggests that Nova1, Ncoa3 and Ewsr1 are partially required for the repressive function of miR-134, in a comparable manner to Tnrc6c. Surprisingly the siRNAs against Rpl5 (0.62 ± 0.27), Nxt1 (0.52 ± 0.20), Rbm25 (0.51 ± 0.07), U2af1 (0.49 ± 0.05) and Nxf1 (0.46 ± 0.03) did not significantly change the ratio “+miR” to “-miR” in the UBE3A-luc reporter. This suggests that knockdown of these RBPs might affect reporter gene expression independent of miR-134, for example due to a general role of these RBPs in mRNA translation. None of the siRNAs had a significant effect on the empty luciferase vector demonstrating that the effects of the siRNAs were strictly dependent on the presence of 3’UTR regulatory sequences (Figure 6 B). Among all candidates, the knockdown of Nova1 had the strongest activating effect on reporter gene expression. We therefore decided to study this RBP in more detail. The focus of this thesis was to characterize the role of Nova1 in the regulation of miRNA function and to understand the molecular mechanism whereby Nova1 knockdown impairs miRNA mediated inhibition of target mRNAs. RESULTS 49 Figure 6: Luciferase reporter assay for the validation of positive hits using RNAi A) The pGL4-Ube3a-3’UTR luciferase reporter (UBE3A-luc) was co-transfected with miR-134 duplex RNA and one of the indicated siRNAs into rat cortical neurons at 5 DIV. Numbers of siRNAs refer to their original designation in the first screen. Plotted are normalized luciferase values as ratio of a condition with miR-134 duplex (“+miR”) to a condition without miR-134 duplex (“-miR”). Values are the average from three independent experiments ± standard deviation. *, p < 0.05 (unpaired Student’s t-test compared to control siRNA). B) An empty pGL4 plasmid was co-transfected together with miR-134 duplex RNA and one of the indicated siRNAs into rat cortical neurons at 5 DIV. Values are plotted as in (A). *, p < 0.05 (unpaired Students t-test compared to control siRNA) RESULTS 50 3.3 Expression analysis of Nova1 in primary neuronal cultures The endogenous expression of Nova1 protein was examined in cultured neurons using Western blot analysis and immunocytochemistry. First, protein lysates of rat hippocampal and cortical neurons were prepared at five different time points ranging from 7 DIV to 17 DIV (hippocampus) or 5 DIV to 14 DIV (cortex). As shown in figure 7 A and B, Nova1 was detected at all time points in both cell types, suggesting that it is expressed throughout the in vitro neuronal development. Expression of β-actin in the same lysates served as a loading control. In the next step, the subcellular localization of Nova1 protein was analyzed in more detail by immunocytochemistry (ICC) of hippocampal neurons (14 DIV) with an anti-Nova1 antibody. The neuronal cell morphology was visualized by staining for the microtubule - associated protein 2 (MAP2, green). The nuclei were stained with Hoechst dye (blue). Nova1 signal, shown in red, was observed in the cytoplasm and nucleus (Figure 7 C). The majority of cytoplasmic Nova1 was detected as granules surrounding the nucleus. Few Nova1 positive granules were also observed in the proximal parts of dendrites (arrows Figure 7 D). The subcellular localization of Nova1 assessed by ICC could be confirmed by a Western blot analysis of cytoplasmic and nuclear fractions prepared from rat cortical neurons. Nova1 was detected in both fractions, with a more prominent expression in the nuclear fraction (Figure 7 E). Western Blots for α -tubulin and HDAC2 (Histone Deacetylase 2) confirmed the proper separation of cytoplasmic and nuclear fractions in this experiment. In conclusion, Nova1 protein is expressed in both the nucleus and cytoplasm of hippocampal neurons during the neuronal development between 7 and 14 DIV. RESULTS 51 Figure 7: Expression of Nova1 in primary rat neurons in culture A), B) Western blot analysis of whole-cell protein lysates prepared from hippocampal (HC) (A) or cortical (CTX) (B) neuronal cultures at the indicated days in vitro (DIV). β-actin was used as a loading control. C), D) Immunocytochemistry analysis of Nova1 expression in hippocampal culture at 14 DIV. C) Signal of Nova1-specific antibody is pictured in red. The neuronal cell was visualized by a Map2 staining (green). Nuclei were stained with Hoechst (blue). D) Magnification of the insert depicted in (C) shows Nova1 signal in grayscale and nuclear and cellular outlines in yellow. Scale bar = 10 µm. E) Western blot analysis of Nova1 of whole cell lysates (input) and cytoplasmic (cyt) or nuclear (nuc) fractions obtained from FUDR- treated rat hippocampal neurons at 7 DIV. α-tubulin was used as cytoplasmic marker. HDAC2 was used as nuclear marker. The fractionation experiment was performed in contribution with Peter Störchel. 3.4 Nova1 is required for the repressive function of miRNAs 3.4.1 Nova1 knockdown impairs miR-138 repressive activity in young cortical neurons To characterize the specific role of Nova1 in the regulation of miRNAs it was of interest to investigate whether Nova1 is required for the repressive function of other miRNAs apart from miR-134. To address this point, I examined the activity of the neuronal miRNA miR-138 in the context of Nova1 knockdown using luciferase reporter gene assays. RESULTS 52 Luciferase assay was carried out using the miR-138 responsive APT1 3’UTR luciferase construct in the pGL4 vector (APT1-luc). APT1 (Acyl-Protein-Thioesterase-1) is encoded by the LYPLA1 gene which contains a binding site for miR-138 in the 3’UTR (Siegel et al. 2009). Analogous to the secondary screen (chapter 3.2) the luciferase reporter was co-transfected into young rat cortical neurons (5 DIV) together with the miR-138 duplex RNA and either control siRNA or the siRNAs targeting Nova1 or Ewsr1, respectively. The luciferase activity was measured three days after transfection. The results are presented as ratio between the condition with miR-138 duplex RNA (“+miR”) and the condition without miR-138 duplex RNA (“-miR”) (Figure 8). Overexpression of miR-138 only (basal) led to a significant repression of the APT1 luciferase reporter (0.34 ± 0.07). The knockdown of Nova1 caused a significant relief of the repression (0.76 ± 0.26) whereas the knockdown of Ewsr1 had no effect on the reporter expression (0.32 ± 0.07). This result suggests that Nova1 knockdown interferes with miR-138 mediated repression of reporter gene expression similar to miR-134 function. In conclusion, this suggests that Nova1 regulates the activity of several neuronal miRNAs, whereas Ewsr1 might be more specific for the regulation of miR-134. Figure 8: Luciferase reporter assay of APT1-luc in cortical neurons overexpressing miR-138 pGL4-APT1 3’UTR (APT1-luc) was co-transfected into rat cortical neurons at 5 days in vitro (DIV) with indicated siRNAs in the presence (“+miR”) or absence (“-miR”) of miR-138 duplex RNA. Values represent ratio of the two conditions (“+miR”/”-miR”) and are the average from three independent experiments ± standard deviation. *, p < 0.