- Email: [email protected]
What are the roles of microRNAs at the mammalian synapse?
Neuroscience Letters 466 (2009) 63–68
Contents lists available at ScienceDirect
Neuroscience Letters journal homepage: www.elsevier.com/locate/neulet
Mini-review
What are the roles of microRNAs at the mammalian synapse? Anetta Konecna, Jacki E. Heraud, Lucia Schoderboeck, Alexandre A.S.F. Raposo, Michael A. Kiebler ∗ Center for Brain Research, Medical University of Vienna, Spitalgasse 4, 1090 Vienna, Austria
a r t i c l e
i n f o
Article history: Received 6 March 2009 Received in revised form 12 June 2009 Accepted 17 June 2009 Keywords: microRNA Dendritic RNA transport Neuronal RNA granules Processing bodies Translational control at the synapse RNA-binding proteins
a b s t r a c t The modification of neuronal connections in response to stimuli is believed to be the basis of long-term memory formation. It is currently accepted that local protein synthesis critically contributes to siterestricted modulation of individual synapses. Here, we summarize recent evidence implicating miRNAs in this process, leading to altered dendrite morphogenesis and synaptic plasticity. Second, we discuss findings in non-neuronal systems about how RNA-binding proteins can modulate miRNA–mRNA interactions, and how these mechanisms might apply to neurons. Finally, we review recent findings that P-bodies may be important sites for miRNA action at the synapse. © 2009 Elsevier Ireland Ltd. All rights reserved.
The reaction of specific neural circuits to stimuli and their modifications by experience at the cellular and molecular level during memory formation is the underpinning of current research in molecular neuroscience. There are two distinct forms of memory depending on the duration: short-term memory (STM) and long-term memory (LTM) [12]. Stable modifications in neuronal connections and in the strength of synaptic activity are believed to underlie LTM. Long-term potentiation (LTP) and long-term depression (LTD) serve as molecular models to study memory, and are a measure of increased or decreased synaptic strength, respectively. In analogy to memory, LTP has two temporal phases, an early phase and a late phase [12]. Early phase LTP is manifested by quick posttranslational modifications of pre-existing proteins at the synapse, whereas the late phase requires new gene expression and protein synthesis. In addition to global protein synthesis in the cell body, new proteins are also produced in dendrites [39]. This observation suggests a very attractive model where site-restricted modulation of the protein composition at a single synapse in response to various repetitive stimulations would lead to synaptic plasticity [39,40]. Obviously, a prerequisite for local translation is the presence of mRNA [10] and translation machinery in the vicinity of synapses [39]. Several hundred transcripts have been identified in synaptoneurosomes or isolated dendrites ([35] and references therein), however only a few have been reliably shown to localize to dendrites by in situ hybridization (ISH) in cultured neurons or in brain slices. Dendritically localized mRNAs are found in ribonucleoprotein particles (RNPs) or neuronal RNA granules that are
∗ Corresponding author. E-mail address: [email protected] (M.A. Kiebler). 0304-3940/$ – see front matter © 2009 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.neulet.2009.06.050
transported to distal dendrites by molecular motors, presumably in a translationally silent state [6,12]. Thus, RNA-binding proteins (RBPs) might have dual roles, one as localization determinants and another as translational regulators. A fascinating and new addition to the mechanisms of translational regulation at the synapse are small non-coding RNAs, in particular microRNAs (miRNAs). In this review, we summarize recent data that demonstrate the importance of miRNAs in regulating protein synthesis in the vicinity of the synapses that undergo plastic changes during formation of memory. We also discuss possible mechanisms for this miRNA regulation and the putative sites of action in dendrites. miRNAs are small non-coding RNA effectors of posttranscriptional regulation that are 19–22nt in length. They are transcribed in the nucleus as long primary transcripts, which are further processed in the nucleus and the cytoplasm to yield mature miRNA duplexes [16]. One strand of the duplex is then loaded into the RNA-induced silencing complex (RISC), of which the Argonaute (Ago) proteins are a major component [34]. In mammals, the vast majority of miRNAs are partially complementary to target mRNAs, in some cases only in the 6–8nt long 5 -seed region of the miRNA, and repress their targets by either blocking translation or causing their destabilization. To date, there are only a few examples of miRNA-mediated cleavage of target mRNA through perfect base pairing, as is more commonly seen in plants, in mammals [3]. Since the requirement for target complementarity is only partial, this means that one miRNA can potentially have hundreds of targets [3]. The discovery of miRNAs provoked a frenzied attempt to inventory the expression pattern of miRNAs in different organisms and organ systems. The brain expresses a disproportionately large number of tissue-specific or -enriched miRNAs, suggesting important roles in brain-specific functions [24,26]. Currently, the miRNA database,
64
A. Konecna et al. / Neuroscience Letters 466 (2009) 63–68
miR-base, reports 706 miRNAs in humans, 547 in mouse and 286 in rat (http://microrna.sanger.ac.uk/sequences), with new miRNAs constantly being identified. Recent deep sequencing of miRNAs cloned from different mouse brain regions found a total of 329 miRNAs expressed [26], therefore, more than half of all presently known mouse miRNAs are expressed in the brain. Many miRNAs are temporally regulated, with the expression of several peaking in the adult animal, suggesting roles for those miRNAs in the mature brain [24]. In addition, the presence of a large number of primate-specific miRNAs expressed in the brain has raised the interesting possibility that miRNAs have played a role in the evolution of brain function [5]. One way to analyze brain-expressed miRNAs has been to determine their subcellular localization in mature neurons. The Kosik laboratory used laser capture techniques to isolate the somatic and dendritic compartments of mature cultured hippocampal neurons, followed by multiplex real-time PCR to identify both the mRNA and miRNA components of these compartments. Using stringent criteria, only 4 miRNAs (miR-292-5p, miR-26a, miR-26b, miR-25) were significantly enriched in dendrites when comparing the ratio of abundance to the somatic compartment [25]. The dendritic localization of miR-26a was additionally confirmed by ISH and was reported to repress translation of the dendritically localized mRNA, MAP2. In an alternative approach, Lugli and colleagues [29] biochemically purified synaptoneurosomes and used a microarray to identify the miRNAs present (Table 1). The novelty here came from the identification of not only mature miRNAs, but also the premiRNAs. Coupling this with the presence of Dicer and Ago2 in this biochemical fraction ([29] and references therein), raises the intriguing possibility that mature miRNAs may be processed from their precursor in very specific subcellular domains, such as individual synapses. Some of the best indications in vivo for the role of miRNAs in differentiated neurons, however, have come from region-specific knockouts of Dicer, an essential component for miRNA biogenesis. Loss of Dicer in the forebrain results in ataxia, microcephaly and decreased lifespan. In Purkinje cells, inactivation of Dicer leads to cerebellar degeneration and development of ataxia ([12] and references therein). In addition, mice heterozygous for DGCR8, another component required for miRNA biogenesis, exhibit reduced dendritic complexity, altered dendritic spine morphology and cognitive deficits [38]. These phenotypes all point to an essential role for the miRNA pathway in mammalian neuronal survival and function. In Drosophila, pasha (DGCR8 homolog) and Dicer-1 mutants were identified in a forward genetic screen for genes involved in the targeting of projection neuron dendrites in the antennal lobe to higher brain centers, indicating a role for this pathway in the formation of correct neuronal connections [4]. It remains to be seen whether this role is conserved in mammals, but it seems plausible given the severity of the knockout phenotypes in mammals. In addition to the roles
characterized in vivo, miRNA mis-regulation has also been linked to a number of psychiatric disorders and neurodegenerative diseases, including Tourette’s syndrome, Rett’s syndrome, Parkinson’s disease and Alzheimer’s disease[8]. The current challenge is to elucidate the biological functions of individual miRNAs in neurons. That miRNAs are essential for neuron differentiation and development is well established [24], however, their activity in mature neurons and, in particular, at the synapse is still not well characterized. Recently, several important studies have pioneered functional analysis of individual miRNAs in mature post-mitotic neurons, uncovering roles in dendrite morphogenesis, synaptogenesis [17,43] and synaptic plasticity [36,37]. During development, the specific pattern of arborization of dendrites determines the connectivity of a neuron. This process is influenced by activity and is critical to neuronal function [45]. It is in part regulated by activation of gene expression through a number of transcription factors. Among their myriad of target genes, activity-induced transcription factors also up-regulate a number of miRNAs [17,43]. Expression of miR-132 is induced by the transcription factor cAMP response element-binding protein (CREB) in response to neuronal stimulation in developing neurons and consequently effects a change in neuronal morphology [43]. Induction of miR-132 is dependent on NMDA receptor activity and leads to increased levels of the mature miRNA, both in the soma and dendrites. The consequent down-regulation of the miR-132 target p250GAP, a GTPase activating protein, leads to enhanced neurite outgrowth [43]. In a similar fashion, miRNAs from the miR379-410 cluster are up-regulated by the activity-dependent transcription factor myocyte enhancing factor 2 (MEF2) in response to neuronal stimulation, some of which then effect dendritogenesis [17]. In particular, miR-134, in addition to its role in regulating spine morphogenesis in mature neurons (discussed below), regulates dendritic elaboration through down-regulation of the translational repressor Pumilio2 in response to neuronal activity [17]. This demonstrates an interesting interplay between translation control factors, miRNAs and transcription factors. Additionally, both studies provide a link between the transcriptional programs initiated in response to activity, and the post-transcriptional events controlling translation of pre-existing mRNAs that are affected by miRNAs and translational regulators. Behavioral responses, including learning and LTM formation, require regulated synthesis of new proteins at synapses in Drosophila as in mammals [1,40]. Recent identification of miRNAs and mRNAs in distal dendrites near synapses lend weight to tempting theories pertaining to a local miRNA-mediated repression of targets involved in synaptic plasticity. Indeed, in a groundbreaking study, Ashraf et al. reported that synaptic protein synthesis occurs in Drosophila antennal lobes at the sites of the formation of stable memories in response to conditioned training of odor stimuli paired
Table 1 miRNAs localizing in dendrites. Dendritic miRNA
Method
Target in neurons
Ref.
miR-134 miR-26a miR-292-5p, miR-26b, miR-25 let7, miR-128 miR-132 miR-200c, miR-339, miR-322, miR-466, miR-425, miR-182, miR-350, miR-183, miR-351, miR-297 miR-138 miR-218, miR-9 miR-29a, miR-7, miR-341, miR-137, miR-335, miR-376b, miR-98 miR-370, miR-9, miR-124
ISH ISH Laser capture and multiplex RT-PCR ISH ISH Microarray (synaptoneurosomes) ISH and microarray (synaptosomes) ISH and microarray (synaptosomes) Microarray (synaptosomes) ISHb
LIMK1 MAP2
[36] [25] [25] [13] [43] [29]a [37] [37] [37] [33]
a b
Top 10 enriched miRNAs in synaptoneurosomes only. In situ hybridization (ISH) performed on cultured neurons except where indicated, on brain sections.
p250GAP APT1
A. Konecna et al. / Neuroscience Letters 466 (2009) 63–68
65
Fig. 1. Models for miRNA-mediated regulation of protein translation at the mammalian and Drosophila synapse. (A) miR-134 represses Limk1 mRNA under resting conditions, thus restricting Limk1 levels at the synapse in hippocampal neurons (left). Upon synaptic stimulation with BDNF, miR-134 repression is relieved, which results in production of Limk1, enhanced actin polymerization and the spine growth (right). The inactivation of miR-134 is triggered by TrkB/mTOR pathway by a yet unknown mechanism. miR-138 represses APT1 mRNA resulting in a low depalmitoylation activity at the synapse and increased membrane anchoring of G␣13 protein, an activator of RhoA GTPase (left). Upon miR-138 inhibition with an LNA inhibitor, membrane-bound G␣13 levels are reduced by newly synthesized APT1 depalmitoylase leading to RhoA inactivation and spine growth (right). (B) External stimulation of olfactory neurons in Drosophila antennal lobe glomeruli (odor, electric shock) promotes proteasome-mediated degradation of the RISC component Armitage (armi). In the absence of Armitage, the RISC may not form allowing some miRNA-repressed mRNAs involved in synaptic plasticity, such as CaMKII, to be translated resulting in LTM formation.
with electric shock [1]. To visualize synaptic protein synthesis, a fluorescent reporter containing the CaMKII 3 -UTR was expressed in transgenic lines. Both olfactory training and acetylcholine receptor stimulation of explanted brains produced increased dendritic transport and localization of this reporter together with CaMKII protein synthesis at particular antennal glomeruli or synapses. This process is likely to be regulated by RISC, as mutations in proteins participating in RISC assembly or function such as, Dicer-2, Aubergine and Armitage, lead to a pronounced increase in CaMKII expression. The study also shows that the levels of the Armitage helicase drop upon synaptic stimulation and CaMKII expression increases due to proteasome activity. Therefore, a mechanism triggering proteasome-mediated degradation of RISC factors by an integrated sensory trigger was proposed to activate synaptic protein synthesis and release localized mRNAs from miRNA suppression (Fig. 1). In another recent report, the fragile X mental retardation protein (dFMRP) was shown to be involved in LTM formation via genetic interactions with the miRNA pathway in Drosophila [9]. Homozygous dFMRP mutants have impaired LTM formation due to an excess of basal protein synthesis. Interestingly, heterozygous double mutants for dFMRP and dAgo1 showed impaired LTM formation, while those of heterozygous single mutants were not affected. The authors therefore propose that FMRP mutants might lead to LTM dysfunction via disruption of miRNA-dependent regulation of protein synthesis. In mammalian neurons, only two localized mRNAs have been shown so far to be regulated in a miRNA-dependent manner [36,37]. First, Schratt et al. [36] discovered that miR-134 is distributed in a punctate pattern within dendrites and near synapses. Overexpression of miR-134 caused a small but significant decrease in spine volumes whereas its inhibition resulted in an opposite effect. A search for putative miR-134 targets identified Lim-domaincontaining kinase 1 (Limk1) mRNA, a direct regulator of actin filament dynamics that is important for spine remodeling. Additionally, Limk1 KO mice exhibit dendritic spine shrinking similar to that induced by the excess of miR-134, a strong indication that miR-134 regulates local expression of LimK1 in dendrites ([36] and references therein). Of particular interest is their finding that the repression of Limk1 mRNA translation can be released upon synaptic activation by brain-derived neurotrophic factor (BDNF), possibly
via the activation of the mTOR (mammalian target of rapamycin) signaling pathway (Fig. 1). The second miRNA, miR-138, was identified as a brain-specific, activity-regulated miRNA that is enriched in the synaptoneurosomal fraction and localizes to distal dendrites in hippocampal neurons [37]. Alteration of miR-138 levels in hippocampal neurons showed similar spine phenotypes as observed for miR-134. Bioinformatic analysis suggested several candidate mRNAs from which, only one, Acyl protein thioesterase 1 (APT1), was validated in this study as a miR-138 target in neurons. APT1 is an enzyme that catalyzes the removal of palmitate modifications, and therefore its activity could affect membrane anchoring of many post-synaptic proteins. The authors indirectly demonstrate that repression of APT1 mRNA translation via miR-138 might prevent palmitoylation of the APT1 substrate, G protein ␣ 13 (G␣13 ). Membrane-bound G␣13 acts as a RhoA activator, which in turn negatively regulates dendritic spine growth (Fig. 1) ([37] and references therein). It will be interesting to see whether the action of miR-138 or some of the other dendritically localized miRNAs are also controlled by synaptic activity, and what further functions will be found for such miRNAs. While miRNAs are now widely accepted as regulators of gene expression at the post-transcriptional level, their mode of action is still controversial [16,27]. Current models have been built on conflicting observations that miRNAs may act either at the step of translation initiation or elongation. Furthermore, there exist compelling data showing that miRNAs can induce target mRNA degradation [16,27]. The gaps in the experimental data leading to discrepancies in proposed concepts have been recently discussed [16]. A completely unexplored territory in the miRNA field has emerged only recently with the findings that RBPs might be important for miRNA functioning [7,22,42]. These proteins were not identified as core RISC/miRNP components, but their association appears to be critical for modulating the RISC/miRNP activities. It was observed in mammalian cell lines that miRNA-associated targets may shuttle between polyribosomes and certain “storage granules”, like processing bodies (PBs) or stress granules (SGs), depending on cellular conditions and the cell type [27]. Hence, a miRNA target might undergo a cycle of (reversible) repression and activation, suggesting the need for a regulatory mechanism.
66
A. Konecna et al. / Neuroscience Letters 466 (2009) 63–68
This is particularly relevant in neurons, given the observations of activity-dependent de-repression of targets discussed previously. The first example of a player that reverses miRNA repression came from work by the Filipowicz group [7]. A member of the embryonic lethal abnormal vision (ELAV) protein family, ELAV1 (also known as HuR), is required to release cationic amino acid transporter 1 (CAT1) mRNA from miR-122 repression in response to amino acid starvation in hepatocarcinoma cells. Under favorable conditions, CAT1 mRNA localizes to PBs in a miR-122-dependent manner. Upon stress, however, CAT1 mRNA is released from PBs into the cytoplasm to become translated in a process that requires ELAV1 binding to the AU-rich region in the 3 -UTR. Their great achievement is that they succeeded in visualizing an endogenous mRNA in PBs. Currently, we lack such “model” mRNAs to study their association with PBs in primary cells, probably due to a high mRNA turnover in PBs. To overcome this problem, many studies have been using overexpression of reporters containing miRNA-targeted 3 -UTRs. The second RBP found to counteract miRNA-mediated silencing, is Dead end 1 (Dnd1) [22]. Dnd1 was initially described as a protein required for germ cell viability, since its mutation in mice causes germ cell loss accompanied by testicular germ cell tumors ([22] and references therein). In a reporter-based screen, Dnd1 was found to elevate the expression of two tumor suppressor genes (LATS2 and p27) bearing functional miRNA-binding sites for miR-372 and miR-221, respectively [22]. In contrast to somatic cells in zebrafish, where miRNAs reduce nanos and tdrd7 mRNA levels, Dnd1 is able to protect these mRNAs from degradation in germ cells. The effect of Dnd1 is dependent on binding to a U-rich region close to the miRNA-binding sites where it may sterically block their access. The latest and unexpected addition to the modes of miRNA regulation is provided by a recent report demonstrating the stimulatory effects of a miRNA on protein translation [42]. The basis for this interesting discovery was the initial observation that both Ago2 and FXR1 binding to AU-rich elements (AREs) in tumor necrosis factor ˛ (TNF˛) mRNA are required to increase its translation upon cell cycle arrest ([42] and references therein). The TNF˛ 3 -UTR contains three tandem AUUUA motifs, two of which are flanked by sequences complementary to the seeds of miR-369-3. The authors showed that miR-369-3 represses TNF␣ translation in proliferating cells, whereas it enhances translation in arrested cells. The FXR protein thereby acts as the key molecular switch from miRNA-mediated repression to activation by binding to the ARE of TNF␣ mRNA only upon cell cycle arrest. This same mechanism was shown in the same paper to also apply for two other miRNAs. Thus, the cycling of activation-repression rounds could be a more general mechanism for miRNA function, which is under the control of RBPs. All data concerning the mechanisms of miRNA-mediated regulation have so far come mostly from systems other than neuronal cells. Although the reversal of miRNA repression of localized mRNAs at stimulated synapses (CaMKII and LimK1), has been documented [1,36], the exact molecular machinery behind this process is not understood. In light of the latest findings where RBPs regulate miRNA functions, it is very likely that such regulatory events occur at mammalian synapses as well. The first candidates are the ELAV proteins (HuB, HuC ad HuD), some of whose expression is restricted to neurons [32]. ELAV4/HuD protein was previously implicated in memory formation as its levels undergo a sustained up-regulation in the hippocampus of trained mice in spatial discrimination and learning tasks [32]. Consistent with this, ELAV4/HuD is recruited to dendritic spines upon glutamate receptor activation in cultured hippocampal neurons and associates with neuronal mRNAs encoding proteins implicated in synaptic plasticity [41]. Dnd1 protein is highly expressed only in the developing brain, and mice with genetically disrupted Dnd1 are viable, with no obvious neurological deficiencies ([22] and references therein). Therefore, Dnd1 may not be the major determinant in miRNA-regulated cognitive functions,
but it may still be involved in some miRNA-regulated processes during neuronal development. And lastly, FXR1 is a known interacting partner of FMRP, whose participation in miRNA functioning is still an unresolved issue [20]. A common trait of miRNA-dependent regulation is the involvement of ARE or U-rich sequences in miRNA-targeted transcripts. ARE-containing mRNAs (TNF˛) are specifically transferred from polysomes to PBs upon translation inhibition by the ARE-binding proteins of the tristetraproline family [18]. Conversely, ARE-binding RBPs relieve miRNA repression to re-enter translation (see above). Thus, the ARE-binding proteins, which have long been thought to determine mRNA stability or degradation, might execute this function at least in part by regulating the mRNA subcellular localization possibly in concert with bound miRNAs [21]. It is now widely accepted that cytoplasmic mRNAs can be stored and silenced in RNA granules, via translation repressor proteins and miRNAs [31]. However, where and how miRNAs operate on their target mRNAs is still an open question. The first hints came from studies on miRNAs and RISC in non-neuronal cells showing that Ago proteins localize to distinct cytoplasmic foci together with markers for PBs [14,16]. PBs were originally identified as cellular sites of mRNA decay that harbor exonucleolytic (Xrn1) and decapping enzymes (Dcp1, Dcp2), decapping activators (RNA helicase Rck/p54), factors of non-sense mediated decay (Upf1, Upf2, Upf3) and other auxiliary proteins (GW182/TNRC6A, CCR4NOT deadenylase) [14,31]. Genetic studies also revealed a much closer relationship between miRNAs and mRNA degradation and repression in PBs than expected, as mutations in the Drosophila orthologues gawky (GW182), Me31B (Rck/p54), dDcp1 and dDcp2 (Dcp1 and 2), and the ccr4/twin-not3/5 complex (CCR4-NOT), disrupt miRNA-mediated translational repression and miRNAmediated mRNA degradation in flies [14]. In mammalian cells, the GW182 protein family consists of three proteins, TNRC6A (GW182), TNRC6B and TNRC6C, all of which localize to P-bodies and are all required for miRNA function [15,19,28,30]. TNRC6A also directly interacts with Ago2 [15]. Furthermore, the localization of miRNAs and their targets in PBs [27] is another strong indication that PBs are potential sites of miRNA activity. In the Drosophila nervous system, a range of established PB components (dDcp1, Upf1, Me31B/Rck) and Ago2 were observed not only in the cell body of cultured Drosophila sensory neurons, but also in dendrites [2]. Surprisingly, PBs and Ago2 extensively colocalized with components of distinct RNA granules, namely dFMRP and Staufen. In contrast to flies, however, immunostaining and dual color videomicroscopy of rat hippocampal neurons showed no significant overlap of Dcp1 and Rck with markers for transport RNPs (Stau1, Stau2, and Barentsz), suggesting a different interaction between transport RNPs and PBs in rodent neurons [46]. In another report, Cougot et al. demonstrated that dendritic PB-like structures labeled by GW182 contain Ago2 and endogenous miRNAs (miR-128, let7), as well as a let7-target reporter in either hipppocampal or hypothalamic neurons [13]. In addition, the authors detected a high degree of co-localization between GW182, the zip-code binding protein 1 (ZBP1) and FMRP in dendrites, which is more in agreement with the initial report from Drosophila neurons. Surprisingly, almost two thirds of dendritic PBs do not stain for the Xrn1 exonuclease, in contrast to PBs in the soma or in HeLa cells [16,31]. Since some of the dendritic PBs are also positive for a ribosomal component, Y10B, the authors speculate that neurons might possess two distinct populations of PB-like structures, with a dendritic population primarily being used for mRNA storage, and a cell body population where typical mRNA degradation would occur. It is important to note, however, that an absence of Xrn1 in the dendritic compartment does not exclude a possible 5 - to 3 -RNA degradation by other, unknown, exoribonucleases. To support the authors’ hypothesis, a biochemical study measuring transcript levels from
A. Konecna et al. / Neuroscience Letters 466 (2009) 63–68
different fractions would be required. Nonetheless, dendritic PB stability does seem to differ from that of HeLa PBs [14,31]. Photobleaching experiments showed that Dcp1 marked granules have a low rate of recovery, which contrasts with the rapid exchange detected previously in HeLa cells. In neurons, the efficiency of Ago2 recovery is higher than for Dcp1, although it is still below HeLa cell values (e.g., difference of ∼20% for dendrites). Remarkably, the Dcp1 exchange rate in neurons is greatly improved upon stimulation with NMDA [27]. In addition, BDNF treatment of neurons induced the translocation of PBs into dendrites and the release of most of the Ago2 from the RNPs [13]. These results show that synaptic activation can change the dynamics and composition of PBs and suggests that stimulation of neuronal activity may contribute to localized derepression of translation. In another study, Zeitelhofer and colleagues observed that both PB and Ago2 particles disassemble following neuronal stimulation with NMDA, glutamate or BDNF, supporting a model where synaptic activity induces the derepression of Ago2 and PB-based translational silencing [46]. In contrast, only a small percentage of the cellular Ago2 in non-neuronal cells is confined to PBs and Ago2 undergoes rapid exchange between PBs and the cytoplasm, where it is likely to associate with target mRNAs in polyribosomes [27]. Also in contrast to neurons, disruption of fluorescence-detectable PBs in mammalian cell lines does not interfere with miRNA functions, further suggesting that miRNAs may find distinct ways to regulate translation in different tissues [11]. In summary, these results point to substantial functional differences between neurons and the nonneuronal cell lines. The studies described above suggest that PBs in neurons might play a more prominent role in miRNA-mediated repression than in other cells since their dynamics, composition and integrity change rapidly in response to synaptic activation. However, many open questions remain. Since miRNAs can be found in PBs they might be also present in other RNA granules such as transport RNPs or SGs [23]. Whereas PBs temporarily disassemble upon synaptic activity, transport RNPs appear to be more stable and SGs only form upon cellular stress. The current working hypothesis is that mRNAs cycle through silencing and derepression and may require assembly/disassembly of RNA granules or trafficking of the transcript between different RNPs. Recent evidence favors the latter, since neuronal PBs were shown to frequently come into stable contact (termed ‘docking’) with Stau2- or ZBP1-containing RNPs in vivo [46]. Increasingly, miRNAs are being considered as one of the key players in mechanisms of translational control in neurons. We are only beginning to understand their full engagement in molecular processes underlining higher cognitive functions of the brain. Until now, there is only one piece of evidence in vivo that a particular behavior triggers de-repression of the synaptic translation of a specific miRNA-target leading to the establishment of a stable memory in Drosophila [1]. In mammals, we are currently limited to observations of morphological changes in neurons with artificially altered levels of miRNA expression [36] during dendritic outgrowth or spine remodeling. These two cellular phenotypes imply the participation of numerous gene products, which might be controlled via miRNAs. The current challenge is to discover the interaction network of miRNAs with their targets, and in particular in dendrites. In the near future, we can expect the first insights into how miRNAs operate at the synapse from identification of miRNA targets among localized transcripts in dendrites. Activation of individual synapses could release the repression of localized mRNAs by miRNAs, thereby leading to translation, and modification of individual synaptic protein composition. This hypothesis is especially attractive since miRNA-mediated regulation would allow spatial and temporal restriction of translation in a reversible manner. In addition, RBPs are likely to regulate miRNA functions depending on the cellular environment. It is possible that RBPs might attenuate
67
or augment repressive effects of miRNAs and have an important impact on target recognition and miRNA loading [44]. Finally, the most critical point will be to establish the link between miRNAmediated silencing and mechanisms of synaptic plasticity in in vivo behavioral tests in mammals. In this respect, the availability of knock-out mice for a number of miRNAs to the neuroscientists will be of crucial importance in the future. Acknowledgements We apologize for omitted references due to space restraints. We thank Natasha Bushati, Gunter Meister and Stefan Hüttelmaier for critical comments. MAK is supported by the FWF, the RNA quality programme (Eurocores, ESF) and the HFSP. References [1] S.I. Ashraf, A.L. McLoon, S.M. Sclarsic, S. Kunes, Synaptic protein synthesis associated with memory is regulated by the RISC pathway in Drosophila, Cell 124 (2006) 191–205. [2] S.A. Barbee, P.S. Estes, A.M. Cziko, J. Hillebrand, R.A. Luedeman, J.M. Coller, N. Johnson, I.C. Howlett, C. Geng, R. Ueda, A.H. Brand, S.F. Newbury, J.E. Wilhelm, R.B. Levine, A. Nakamura, R. Parker, M. Ramaswami, Staufen- and FMRPcontaining neuronal RNPs are structurally and functionally related to somatic P bodies, Neuron 52 (2006) 997–1009. [3] D. Bartel, MicroRNAs: target recognition and regulatory functions, Cell 136 (2009) 215–233. [4] D. Berdnik, A. Fan, C. Potter, L. Luo, MicroRNA processing pathway regulates olfactory neuron morphogenesis, Curr. Biol. 18 (2008) 1754–1759. [5] E. Berezikov, F. Thuemmler, L. Van Laake, I. Kondova, R. Bontrop, E. Cuppen, R. Plasterk, Diversity of microRNAs in human and chimpanzee brain, Nat. Genet. 38 (2006) 1375–1377. [6] F. Besse, A. Ephrussi, Translational control of localized mRNAs: restricting protein synthesis in space and time, Nat. Rev. Mol. Cell Biol. 9 (2008) 971–980. [7] S.N. Bhattacharyya, R. Habermacher, U. Martine, E.I. Closs, W. Filipowicz, Relief of microRNA-mediated translational repression in human cells subjected to stress, Cell 125 (2006) 1111–1124. [8] S. Bicker, G. Schratt, microRNAs: tiny regulators of synapse function in development and disease, J. Cell. Mol. Med. 12 (2008) 1466–1476. [9] F.V. Bolduc, K. Bell, H. Cox, K.S. Broadie, T. Tully, Excess protein synthesis in Drosophila fragile X mutants impairs long-term memory, Nat. Neurosci. 11 (2008) 1143–1145. [10] C.R. Bramham, D.G. Wells, Dendritic mRNA: transport, translation and function, Nat. Rev. Neurosci. 8 (2007) 776–789. [11] C.Y. Chu, T.M. Rana, Translation repression in human cells by microRNA-induced gene silencing requires RCK/p54, PLoS Biol. 4 (2006) e210. [12] M. Costa-Mattioli, W.S. Sossin, E. Klann, N. Sonenberg, Translational control of long-lasting synaptic plasticity and memory, Neuron 61 (2009) 10–26. [13] N. Cougot, S.N. Bhattacharyya, L. Tapia-Arancibia, R. Bordonne, W. Filipowicz, E. Bertrand, F. Rage, Dendrites of mammalian neurons contain specialized Pbody-like structures that respond to neuronal activation, J. Neurosci. 28 (2008) 13793–13804. [14] A. Eulalio, I. Behm-Ansmant, E. Izaurralde, P bodies: at the crossroads of posttranscriptional pathways, Nat. Rev. Mol. Cell Biol. 8 (2007) 9–22. [15] A. Eulalio, E. Huntzinger, E. Izaurralde, GW182 interaction with Argonaute is essential for miRNA-mediated translational repression and mRNA decay, Nat. Struct. Mol. Biol. 15 (2008) 346–353. [16] W. Filipowicz, S.N. Bhattacharyya, N. Sonenberg, Mechanisms of posttranscriptional regulation by microRNAs: are the answers in sight? Nat. Rev. Genet. 9 (2008) 102–114. [17] R. Fiore, S. Khudayberdiev, M. Christensen, G. Siegel, S. Flavell, T. Kim, M. Greenberg, G. Schratt, Mef2-mediated transcription of the miR379-410 cluster regulates activity-dependent dendritogenesis by fine-tuning Pumilio2 protein levels, EMBO J. 28 (2009) 697–710. [18] T.M. Franks, J. Lykke-Andersen, TTP and BRF proteins nucleate processing body formation to silence mRNAs with AU-rich elements, Genes Dev. 21 (2007) 719–735. [19] A. Jakymiw, K.M. Pauley, S. Li, K. Ikeda, S. Lian, T. Eystathioy, M. Satoh, M.J. Fritzler, E.K. Chan, The role of GW/P-bodies in RNA processing and silencing, J. Cell Sci. 120 (2007) 1317–1323. [20] P. Jin, R.S. Alisch, S.T. Warren, RNA and microRNAs in fragile X mental retardation, Nat. Cell Biol. 6 (2004) 1048–1053. [21] Q. Jing, S. Huang, S. Guth, T. Zarubin, A. Motoyama, J. Chen, F. Di Padova, S.C. Lin, H. Gram, J. Han, Involvement of microRNA in AU-rich element-mediated mRNA instability, Cell 120 (2005) 623–634. [22] M. Kedde, M.J. Strasser, B. Boldajipour, J.A. Vrielink, K. Slanchev, C. le Sage, R. Nagel, P.M. Voorhoeve, J. van Duijse, U.A. Orom, A.H. Lund, A. Perrakis, E. Raz, R. Agami, RNA-binding protein Dnd1 inhibits microRNA access to target mRNA, Cell 131 (2007) 1273–1286. [23] M.A. Kiebler, G.J. Bassell, Neuronal RNA granules: movers and makers, Neuron 51 (2006) 685–690.
68
A. Konecna et al. / Neuroscience Letters 466 (2009) 63–68
[24] K. Kosik, The neuronal microRNA system, Nat. Rev. Neurosci. 7 (2006) 911–920. [25] M. Kye, T. Liu, S. Levy, N. Xu, B. Groves, R. Bonneau, K. Lao, K. Kosik, Somatodendritic microRNAs identified by laser capture and multiplex RT-PCR, RNA 13 (2007) 1224–1234. [26] P. Landgraf, M. Rusu, R. Sheridan, A. Sewer, N. Iovino, A. Aravin, S. Pfeffer, A. Rice, A. Kamphorst, M. Landthaler, A mammalian microRNA expression atlas based on small RNA library sequencing, Cell 129 (2007) 1401–1414. [27] A.K. Leung, P.A. Sharp, Function and localization of microRNAs in mammalian cells, Cold Spring Harb. Symp. Quant. Biol. 71 (2006) 29–38. [28] J. Liu, F.V. Rivas, J. Wohlschlegel, J.R. Yates III, R. Parker, G.J. Hannon, A role for the P-body component GW182 in microRNA function, Nat. Cell Biol. 7 (2005) 1261–1266. [29] G. Lugli, V. Torvik, J. Larson, N. Smalheiser, Expression of microRNAs and their precursors in synaptic fractions of adult mouse forebrain, J. Neurochem. 106 (2008) 650–661. [30] G. Meister, M. Landthaler, L. Peters, P.Y. Chen, H. Urlaub, R. Luhrmann, T. Tuschl, Identification of novel argonaute-associated proteins, Curr. Biol. 15 (2005) 2149–2155. [31] R. Parker, U. Sheth, P bodies and the control of mRNA translation and degradation, Mol. Cell 25 (2007) 635–646. [32] A. Pascale, M. Amadio, A. Quattrone, Defining a neuron: neuronal ELAV proteins, Cell. Mol. Life Sci. 65 (2008) 128–140. [33] J. Pena, C. Sohn-Lee, S. Rouhanifard, J. Ludwig, M. Hafner, A. Mihailovic, C. Lim, D. Holoch, P. Berninger, M. Zavolan, T. Tuschl, miRNA in situ hybridization in formaldehyde and EDC–fixed tissues, Nat. Meth. 6 (2009) 139–141. [34] L. Peters, G. Meister, Argonaute proteins: mediators of RNA silencing, Mol. Cell 26 (2007) 611–623. [35] M.M. Poon, S.H. Choi, C.A. Jamieson, D.H. Geschwind, K.C. Martin, Identification of process-localized mRNAs from cultured rodent hippocampal neurons, J. Neurosci. 26 (2006) 13390–13399. [36] G.M. Schratt, F. Tuebing, E.A. Nigh, C.G. Kane, M.E. Sabatini, M. Kiebler, M.E. Greenberg, A brain-specific microRNA regulates dendritic spine development, Nature 439 (2006) 283–289.
[37] G. Siegel, G. Obernosterer, R. Fiore, M. Oehmen, S. Bicker, M. Christensen, S. Khudayberdiev, P.F. Leuschner, C.J. Busch, C. Kane, K. Hubel, F. Dekker, C. Hedberg, B. Rengarajan, C. Drepper, H. Waldmann, S. Kauppinen, M.E. Greenberg, A. Draguhn, M. Rehmsmeier, J. Martinez, G.M. Schratt, A functional screen implicates microRNA-138-dependent regulation of the depalmitoylation enzyme APT1 in dendritic spine morphogenesis, Nat. Cell Biol. 11 (2009) 705–710. [38] K. Stark, B. Xu, A. Bagchi, W. Lai, H. Liu, R. Hsu, X. Wan, P. Pavlidis, A. Mills, M. Karayiorgou, J. Gogos, Altered brain microRNA biogenesis contributes to phenotypic deficits in a 22q11-deletion mouse model, Nat. Genet. 40 (2008) 751–760. [39] O. Steward, E.M. Schuman, Protein synthesis at synaptic sites on dendrites, Annu. Rev. Neurosci. 24 (2001) 299–325. [40] M.A. Sutton, E.M. Schuman, Dendritic protein synthesis, synaptic plasticity, and memory, Cell 127 (2006) 49–58. [41] D.M. Tiruchinapalli, M.D. Ehlers, J.D. Keene, Activity-dependent expression of RNA binding protein HuD and its association with mRNAs in neurons, RNA Biol. 5 (2008) 157–168. [42] S. Vasudevan, Y. Tong, J. Steitz, Switching from repression to activation: MicroRNAs can up-regulate translation, Science 318 (2007) 1931–1934. [43] G.A. Wayman, M. Davare, H. Ando, D. Fortin, O. Varlamova, H.Y. Cheng, D. Marks, K. Obrietan, T.R. Soderling, R.H. Goodman, S. Impey, An activity-regulated microRNA controls dendritic plasticity by down-regulating p250GAP, Proc. Natl. Acad. Sci. U.S.A. 105 (2008) 9093–9098. [44] L. Weinmann, J. Hock, T. Ivacevic, T. Ohrt, J. Mutze, P. Schwille, E. Kremmer, V. Benes, H. Urlaub, G. Meister, Importin 8 is a gene silencing factor that targets argonaute proteins to distinct mRNAs, Cell 136 (2009) 496–507. [45] R. Wong, A. Ghosh, Activity-dependent regulation of dendritic growth and patterning, Nat. Rev. Neurosci. 3 (2002) 803–812. [46] M. Zeitelhofer, D. Karra, P. Macchi, M. Tolino, S. Thomas, M. Schwarz, M. Kiebler, R. Dahm, Dynamic interaction between P-bodies and transport ribonucleoprotein particles in dendrites of mature hippocampal neurons, J. Neurosci. 28 (2008) 7555–7562.
Contents lists available at ScienceDirect
Neuroscience Letters journal homepage: www.elsevier.com/locate/neulet
Mini-review
What are the roles of microRNAs at the mammalian synapse? Anetta Konecna, Jacki E. Heraud, Lucia Schoderboeck, Alexandre A.S.F. Raposo, Michael A. Kiebler ∗ Center for Brain Research, Medical University of Vienna, Spitalgasse 4, 1090 Vienna, Austria
a r t i c l e
i n f o
Article history: Received 6 March 2009 Received in revised form 12 June 2009 Accepted 17 June 2009 Keywords: microRNA Dendritic RNA transport Neuronal RNA granules Processing bodies Translational control at the synapse RNA-binding proteins
a b s t r a c t The modification of neuronal connections in response to stimuli is believed to be the basis of long-term memory formation. It is currently accepted that local protein synthesis critically contributes to siterestricted modulation of individual synapses. Here, we summarize recent evidence implicating miRNAs in this process, leading to altered dendrite morphogenesis and synaptic plasticity. Second, we discuss findings in non-neuronal systems about how RNA-binding proteins can modulate miRNA–mRNA interactions, and how these mechanisms might apply to neurons. Finally, we review recent findings that P-bodies may be important sites for miRNA action at the synapse. © 2009 Elsevier Ireland Ltd. All rights reserved.
The reaction of specific neural circuits to stimuli and their modifications by experience at the cellular and molecular level during memory formation is the underpinning of current research in molecular neuroscience. There are two distinct forms of memory depending on the duration: short-term memory (STM) and long-term memory (LTM) [12]. Stable modifications in neuronal connections and in the strength of synaptic activity are believed to underlie LTM. Long-term potentiation (LTP) and long-term depression (LTD) serve as molecular models to study memory, and are a measure of increased or decreased synaptic strength, respectively. In analogy to memory, LTP has two temporal phases, an early phase and a late phase [12]. Early phase LTP is manifested by quick posttranslational modifications of pre-existing proteins at the synapse, whereas the late phase requires new gene expression and protein synthesis. In addition to global protein synthesis in the cell body, new proteins are also produced in dendrites [39]. This observation suggests a very attractive model where site-restricted modulation of the protein composition at a single synapse in response to various repetitive stimulations would lead to synaptic plasticity [39,40]. Obviously, a prerequisite for local translation is the presence of mRNA [10] and translation machinery in the vicinity of synapses [39]. Several hundred transcripts have been identified in synaptoneurosomes or isolated dendrites ([35] and references therein), however only a few have been reliably shown to localize to dendrites by in situ hybridization (ISH) in cultured neurons or in brain slices. Dendritically localized mRNAs are found in ribonucleoprotein particles (RNPs) or neuronal RNA granules that are
∗ Corresponding author. E-mail address: [email protected] (M.A. Kiebler). 0304-3940/$ – see front matter © 2009 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.neulet.2009.06.050
transported to distal dendrites by molecular motors, presumably in a translationally silent state [6,12]. Thus, RNA-binding proteins (RBPs) might have dual roles, one as localization determinants and another as translational regulators. A fascinating and new addition to the mechanisms of translational regulation at the synapse are small non-coding RNAs, in particular microRNAs (miRNAs). In this review, we summarize recent data that demonstrate the importance of miRNAs in regulating protein synthesis in the vicinity of the synapses that undergo plastic changes during formation of memory. We also discuss possible mechanisms for this miRNA regulation and the putative sites of action in dendrites. miRNAs are small non-coding RNA effectors of posttranscriptional regulation that are 19–22nt in length. They are transcribed in the nucleus as long primary transcripts, which are further processed in the nucleus and the cytoplasm to yield mature miRNA duplexes [16]. One strand of the duplex is then loaded into the RNA-induced silencing complex (RISC), of which the Argonaute (Ago) proteins are a major component [34]. In mammals, the vast majority of miRNAs are partially complementary to target mRNAs, in some cases only in the 6–8nt long 5 -seed region of the miRNA, and repress their targets by either blocking translation or causing their destabilization. To date, there are only a few examples of miRNA-mediated cleavage of target mRNA through perfect base pairing, as is more commonly seen in plants, in mammals [3]. Since the requirement for target complementarity is only partial, this means that one miRNA can potentially have hundreds of targets [3]. The discovery of miRNAs provoked a frenzied attempt to inventory the expression pattern of miRNAs in different organisms and organ systems. The brain expresses a disproportionately large number of tissue-specific or -enriched miRNAs, suggesting important roles in brain-specific functions [24,26]. Currently, the miRNA database,
64
A. Konecna et al. / Neuroscience Letters 466 (2009) 63–68
miR-base, reports 706 miRNAs in humans, 547 in mouse and 286 in rat (http://microrna.sanger.ac.uk/sequences), with new miRNAs constantly being identified. Recent deep sequencing of miRNAs cloned from different mouse brain regions found a total of 329 miRNAs expressed [26], therefore, more than half of all presently known mouse miRNAs are expressed in the brain. Many miRNAs are temporally regulated, with the expression of several peaking in the adult animal, suggesting roles for those miRNAs in the mature brain [24]. In addition, the presence of a large number of primate-specific miRNAs expressed in the brain has raised the interesting possibility that miRNAs have played a role in the evolution of brain function [5]. One way to analyze brain-expressed miRNAs has been to determine their subcellular localization in mature neurons. The Kosik laboratory used laser capture techniques to isolate the somatic and dendritic compartments of mature cultured hippocampal neurons, followed by multiplex real-time PCR to identify both the mRNA and miRNA components of these compartments. Using stringent criteria, only 4 miRNAs (miR-292-5p, miR-26a, miR-26b, miR-25) were significantly enriched in dendrites when comparing the ratio of abundance to the somatic compartment [25]. The dendritic localization of miR-26a was additionally confirmed by ISH and was reported to repress translation of the dendritically localized mRNA, MAP2. In an alternative approach, Lugli and colleagues [29] biochemically purified synaptoneurosomes and used a microarray to identify the miRNAs present (Table 1). The novelty here came from the identification of not only mature miRNAs, but also the premiRNAs. Coupling this with the presence of Dicer and Ago2 in this biochemical fraction ([29] and references therein), raises the intriguing possibility that mature miRNAs may be processed from their precursor in very specific subcellular domains, such as individual synapses. Some of the best indications in vivo for the role of miRNAs in differentiated neurons, however, have come from region-specific knockouts of Dicer, an essential component for miRNA biogenesis. Loss of Dicer in the forebrain results in ataxia, microcephaly and decreased lifespan. In Purkinje cells, inactivation of Dicer leads to cerebellar degeneration and development of ataxia ([12] and references therein). In addition, mice heterozygous for DGCR8, another component required for miRNA biogenesis, exhibit reduced dendritic complexity, altered dendritic spine morphology and cognitive deficits [38]. These phenotypes all point to an essential role for the miRNA pathway in mammalian neuronal survival and function. In Drosophila, pasha (DGCR8 homolog) and Dicer-1 mutants were identified in a forward genetic screen for genes involved in the targeting of projection neuron dendrites in the antennal lobe to higher brain centers, indicating a role for this pathway in the formation of correct neuronal connections [4]. It remains to be seen whether this role is conserved in mammals, but it seems plausible given the severity of the knockout phenotypes in mammals. In addition to the roles
characterized in vivo, miRNA mis-regulation has also been linked to a number of psychiatric disorders and neurodegenerative diseases, including Tourette’s syndrome, Rett’s syndrome, Parkinson’s disease and Alzheimer’s disease[8]. The current challenge is to elucidate the biological functions of individual miRNAs in neurons. That miRNAs are essential for neuron differentiation and development is well established [24], however, their activity in mature neurons and, in particular, at the synapse is still not well characterized. Recently, several important studies have pioneered functional analysis of individual miRNAs in mature post-mitotic neurons, uncovering roles in dendrite morphogenesis, synaptogenesis [17,43] and synaptic plasticity [36,37]. During development, the specific pattern of arborization of dendrites determines the connectivity of a neuron. This process is influenced by activity and is critical to neuronal function [45]. It is in part regulated by activation of gene expression through a number of transcription factors. Among their myriad of target genes, activity-induced transcription factors also up-regulate a number of miRNAs [17,43]. Expression of miR-132 is induced by the transcription factor cAMP response element-binding protein (CREB) in response to neuronal stimulation in developing neurons and consequently effects a change in neuronal morphology [43]. Induction of miR-132 is dependent on NMDA receptor activity and leads to increased levels of the mature miRNA, both in the soma and dendrites. The consequent down-regulation of the miR-132 target p250GAP, a GTPase activating protein, leads to enhanced neurite outgrowth [43]. In a similar fashion, miRNAs from the miR379-410 cluster are up-regulated by the activity-dependent transcription factor myocyte enhancing factor 2 (MEF2) in response to neuronal stimulation, some of which then effect dendritogenesis [17]. In particular, miR-134, in addition to its role in regulating spine morphogenesis in mature neurons (discussed below), regulates dendritic elaboration through down-regulation of the translational repressor Pumilio2 in response to neuronal activity [17]. This demonstrates an interesting interplay between translation control factors, miRNAs and transcription factors. Additionally, both studies provide a link between the transcriptional programs initiated in response to activity, and the post-transcriptional events controlling translation of pre-existing mRNAs that are affected by miRNAs and translational regulators. Behavioral responses, including learning and LTM formation, require regulated synthesis of new proteins at synapses in Drosophila as in mammals [1,40]. Recent identification of miRNAs and mRNAs in distal dendrites near synapses lend weight to tempting theories pertaining to a local miRNA-mediated repression of targets involved in synaptic plasticity. Indeed, in a groundbreaking study, Ashraf et al. reported that synaptic protein synthesis occurs in Drosophila antennal lobes at the sites of the formation of stable memories in response to conditioned training of odor stimuli paired
Table 1 miRNAs localizing in dendrites. Dendritic miRNA
Method
Target in neurons
Ref.
miR-134 miR-26a miR-292-5p, miR-26b, miR-25 let7, miR-128 miR-132 miR-200c, miR-339, miR-322, miR-466, miR-425, miR-182, miR-350, miR-183, miR-351, miR-297 miR-138 miR-218, miR-9 miR-29a, miR-7, miR-341, miR-137, miR-335, miR-376b, miR-98 miR-370, miR-9, miR-124
ISH ISH Laser capture and multiplex RT-PCR ISH ISH Microarray (synaptoneurosomes) ISH and microarray (synaptosomes) ISH and microarray (synaptosomes) Microarray (synaptosomes) ISHb
LIMK1 MAP2
[36] [25] [25] [13] [43] [29]a [37] [37] [37] [33]
a b
Top 10 enriched miRNAs in synaptoneurosomes only. In situ hybridization (ISH) performed on cultured neurons except where indicated, on brain sections.
p250GAP APT1
A. Konecna et al. / Neuroscience Letters 466 (2009) 63–68
65
Fig. 1. Models for miRNA-mediated regulation of protein translation at the mammalian and Drosophila synapse. (A) miR-134 represses Limk1 mRNA under resting conditions, thus restricting Limk1 levels at the synapse in hippocampal neurons (left). Upon synaptic stimulation with BDNF, miR-134 repression is relieved, which results in production of Limk1, enhanced actin polymerization and the spine growth (right). The inactivation of miR-134 is triggered by TrkB/mTOR pathway by a yet unknown mechanism. miR-138 represses APT1 mRNA resulting in a low depalmitoylation activity at the synapse and increased membrane anchoring of G␣13 protein, an activator of RhoA GTPase (left). Upon miR-138 inhibition with an LNA inhibitor, membrane-bound G␣13 levels are reduced by newly synthesized APT1 depalmitoylase leading to RhoA inactivation and spine growth (right). (B) External stimulation of olfactory neurons in Drosophila antennal lobe glomeruli (odor, electric shock) promotes proteasome-mediated degradation of the RISC component Armitage (armi). In the absence of Armitage, the RISC may not form allowing some miRNA-repressed mRNAs involved in synaptic plasticity, such as CaMKII, to be translated resulting in LTM formation.
with electric shock [1]. To visualize synaptic protein synthesis, a fluorescent reporter containing the CaMKII 3 -UTR was expressed in transgenic lines. Both olfactory training and acetylcholine receptor stimulation of explanted brains produced increased dendritic transport and localization of this reporter together with CaMKII protein synthesis at particular antennal glomeruli or synapses. This process is likely to be regulated by RISC, as mutations in proteins participating in RISC assembly or function such as, Dicer-2, Aubergine and Armitage, lead to a pronounced increase in CaMKII expression. The study also shows that the levels of the Armitage helicase drop upon synaptic stimulation and CaMKII expression increases due to proteasome activity. Therefore, a mechanism triggering proteasome-mediated degradation of RISC factors by an integrated sensory trigger was proposed to activate synaptic protein synthesis and release localized mRNAs from miRNA suppression (Fig. 1). In another recent report, the fragile X mental retardation protein (dFMRP) was shown to be involved in LTM formation via genetic interactions with the miRNA pathway in Drosophila [9]. Homozygous dFMRP mutants have impaired LTM formation due to an excess of basal protein synthesis. Interestingly, heterozygous double mutants for dFMRP and dAgo1 showed impaired LTM formation, while those of heterozygous single mutants were not affected. The authors therefore propose that FMRP mutants might lead to LTM dysfunction via disruption of miRNA-dependent regulation of protein synthesis. In mammalian neurons, only two localized mRNAs have been shown so far to be regulated in a miRNA-dependent manner [36,37]. First, Schratt et al. [36] discovered that miR-134 is distributed in a punctate pattern within dendrites and near synapses. Overexpression of miR-134 caused a small but significant decrease in spine volumes whereas its inhibition resulted in an opposite effect. A search for putative miR-134 targets identified Lim-domaincontaining kinase 1 (Limk1) mRNA, a direct regulator of actin filament dynamics that is important for spine remodeling. Additionally, Limk1 KO mice exhibit dendritic spine shrinking similar to that induced by the excess of miR-134, a strong indication that miR-134 regulates local expression of LimK1 in dendrites ([36] and references therein). Of particular interest is their finding that the repression of Limk1 mRNA translation can be released upon synaptic activation by brain-derived neurotrophic factor (BDNF), possibly
via the activation of the mTOR (mammalian target of rapamycin) signaling pathway (Fig. 1). The second miRNA, miR-138, was identified as a brain-specific, activity-regulated miRNA that is enriched in the synaptoneurosomal fraction and localizes to distal dendrites in hippocampal neurons [37]. Alteration of miR-138 levels in hippocampal neurons showed similar spine phenotypes as observed for miR-134. Bioinformatic analysis suggested several candidate mRNAs from which, only one, Acyl protein thioesterase 1 (APT1), was validated in this study as a miR-138 target in neurons. APT1 is an enzyme that catalyzes the removal of palmitate modifications, and therefore its activity could affect membrane anchoring of many post-synaptic proteins. The authors indirectly demonstrate that repression of APT1 mRNA translation via miR-138 might prevent palmitoylation of the APT1 substrate, G protein ␣ 13 (G␣13 ). Membrane-bound G␣13 acts as a RhoA activator, which in turn negatively regulates dendritic spine growth (Fig. 1) ([37] and references therein). It will be interesting to see whether the action of miR-138 or some of the other dendritically localized miRNAs are also controlled by synaptic activity, and what further functions will be found for such miRNAs. While miRNAs are now widely accepted as regulators of gene expression at the post-transcriptional level, their mode of action is still controversial [16,27]. Current models have been built on conflicting observations that miRNAs may act either at the step of translation initiation or elongation. Furthermore, there exist compelling data showing that miRNAs can induce target mRNA degradation [16,27]. The gaps in the experimental data leading to discrepancies in proposed concepts have been recently discussed [16]. A completely unexplored territory in the miRNA field has emerged only recently with the findings that RBPs might be important for miRNA functioning [7,22,42]. These proteins were not identified as core RISC/miRNP components, but their association appears to be critical for modulating the RISC/miRNP activities. It was observed in mammalian cell lines that miRNA-associated targets may shuttle between polyribosomes and certain “storage granules”, like processing bodies (PBs) or stress granules (SGs), depending on cellular conditions and the cell type [27]. Hence, a miRNA target might undergo a cycle of (reversible) repression and activation, suggesting the need for a regulatory mechanism.
66
A. Konecna et al. / Neuroscience Letters 466 (2009) 63–68
This is particularly relevant in neurons, given the observations of activity-dependent de-repression of targets discussed previously. The first example of a player that reverses miRNA repression came from work by the Filipowicz group [7]. A member of the embryonic lethal abnormal vision (ELAV) protein family, ELAV1 (also known as HuR), is required to release cationic amino acid transporter 1 (CAT1) mRNA from miR-122 repression in response to amino acid starvation in hepatocarcinoma cells. Under favorable conditions, CAT1 mRNA localizes to PBs in a miR-122-dependent manner. Upon stress, however, CAT1 mRNA is released from PBs into the cytoplasm to become translated in a process that requires ELAV1 binding to the AU-rich region in the 3 -UTR. Their great achievement is that they succeeded in visualizing an endogenous mRNA in PBs. Currently, we lack such “model” mRNAs to study their association with PBs in primary cells, probably due to a high mRNA turnover in PBs. To overcome this problem, many studies have been using overexpression of reporters containing miRNA-targeted 3 -UTRs. The second RBP found to counteract miRNA-mediated silencing, is Dead end 1 (Dnd1) [22]. Dnd1 was initially described as a protein required for germ cell viability, since its mutation in mice causes germ cell loss accompanied by testicular germ cell tumors ([22] and references therein). In a reporter-based screen, Dnd1 was found to elevate the expression of two tumor suppressor genes (LATS2 and p27) bearing functional miRNA-binding sites for miR-372 and miR-221, respectively [22]. In contrast to somatic cells in zebrafish, where miRNAs reduce nanos and tdrd7 mRNA levels, Dnd1 is able to protect these mRNAs from degradation in germ cells. The effect of Dnd1 is dependent on binding to a U-rich region close to the miRNA-binding sites where it may sterically block their access. The latest and unexpected addition to the modes of miRNA regulation is provided by a recent report demonstrating the stimulatory effects of a miRNA on protein translation [42]. The basis for this interesting discovery was the initial observation that both Ago2 and FXR1 binding to AU-rich elements (AREs) in tumor necrosis factor ˛ (TNF˛) mRNA are required to increase its translation upon cell cycle arrest ([42] and references therein). The TNF˛ 3 -UTR contains three tandem AUUUA motifs, two of which are flanked by sequences complementary to the seeds of miR-369-3. The authors showed that miR-369-3 represses TNF␣ translation in proliferating cells, whereas it enhances translation in arrested cells. The FXR protein thereby acts as the key molecular switch from miRNA-mediated repression to activation by binding to the ARE of TNF␣ mRNA only upon cell cycle arrest. This same mechanism was shown in the same paper to also apply for two other miRNAs. Thus, the cycling of activation-repression rounds could be a more general mechanism for miRNA function, which is under the control of RBPs. All data concerning the mechanisms of miRNA-mediated regulation have so far come mostly from systems other than neuronal cells. Although the reversal of miRNA repression of localized mRNAs at stimulated synapses (CaMKII and LimK1), has been documented [1,36], the exact molecular machinery behind this process is not understood. In light of the latest findings where RBPs regulate miRNA functions, it is very likely that such regulatory events occur at mammalian synapses as well. The first candidates are the ELAV proteins (HuB, HuC ad HuD), some of whose expression is restricted to neurons [32]. ELAV4/HuD protein was previously implicated in memory formation as its levels undergo a sustained up-regulation in the hippocampus of trained mice in spatial discrimination and learning tasks [32]. Consistent with this, ELAV4/HuD is recruited to dendritic spines upon glutamate receptor activation in cultured hippocampal neurons and associates with neuronal mRNAs encoding proteins implicated in synaptic plasticity [41]. Dnd1 protein is highly expressed only in the developing brain, and mice with genetically disrupted Dnd1 are viable, with no obvious neurological deficiencies ([22] and references therein). Therefore, Dnd1 may not be the major determinant in miRNA-regulated cognitive functions,
but it may still be involved in some miRNA-regulated processes during neuronal development. And lastly, FXR1 is a known interacting partner of FMRP, whose participation in miRNA functioning is still an unresolved issue [20]. A common trait of miRNA-dependent regulation is the involvement of ARE or U-rich sequences in miRNA-targeted transcripts. ARE-containing mRNAs (TNF˛) are specifically transferred from polysomes to PBs upon translation inhibition by the ARE-binding proteins of the tristetraproline family [18]. Conversely, ARE-binding RBPs relieve miRNA repression to re-enter translation (see above). Thus, the ARE-binding proteins, which have long been thought to determine mRNA stability or degradation, might execute this function at least in part by regulating the mRNA subcellular localization possibly in concert with bound miRNAs [21]. It is now widely accepted that cytoplasmic mRNAs can be stored and silenced in RNA granules, via translation repressor proteins and miRNAs [31]. However, where and how miRNAs operate on their target mRNAs is still an open question. The first hints came from studies on miRNAs and RISC in non-neuronal cells showing that Ago proteins localize to distinct cytoplasmic foci together with markers for PBs [14,16]. PBs were originally identified as cellular sites of mRNA decay that harbor exonucleolytic (Xrn1) and decapping enzymes (Dcp1, Dcp2), decapping activators (RNA helicase Rck/p54), factors of non-sense mediated decay (Upf1, Upf2, Upf3) and other auxiliary proteins (GW182/TNRC6A, CCR4NOT deadenylase) [14,31]. Genetic studies also revealed a much closer relationship between miRNAs and mRNA degradation and repression in PBs than expected, as mutations in the Drosophila orthologues gawky (GW182), Me31B (Rck/p54), dDcp1 and dDcp2 (Dcp1 and 2), and the ccr4/twin-not3/5 complex (CCR4-NOT), disrupt miRNA-mediated translational repression and miRNAmediated mRNA degradation in flies [14]. In mammalian cells, the GW182 protein family consists of three proteins, TNRC6A (GW182), TNRC6B and TNRC6C, all of which localize to P-bodies and are all required for miRNA function [15,19,28,30]. TNRC6A also directly interacts with Ago2 [15]. Furthermore, the localization of miRNAs and their targets in PBs [27] is another strong indication that PBs are potential sites of miRNA activity. In the Drosophila nervous system, a range of established PB components (dDcp1, Upf1, Me31B/Rck) and Ago2 were observed not only in the cell body of cultured Drosophila sensory neurons, but also in dendrites [2]. Surprisingly, PBs and Ago2 extensively colocalized with components of distinct RNA granules, namely dFMRP and Staufen. In contrast to flies, however, immunostaining and dual color videomicroscopy of rat hippocampal neurons showed no significant overlap of Dcp1 and Rck with markers for transport RNPs (Stau1, Stau2, and Barentsz), suggesting a different interaction between transport RNPs and PBs in rodent neurons [46]. In another report, Cougot et al. demonstrated that dendritic PB-like structures labeled by GW182 contain Ago2 and endogenous miRNAs (miR-128, let7), as well as a let7-target reporter in either hipppocampal or hypothalamic neurons [13]. In addition, the authors detected a high degree of co-localization between GW182, the zip-code binding protein 1 (ZBP1) and FMRP in dendrites, which is more in agreement with the initial report from Drosophila neurons. Surprisingly, almost two thirds of dendritic PBs do not stain for the Xrn1 exonuclease, in contrast to PBs in the soma or in HeLa cells [16,31]. Since some of the dendritic PBs are also positive for a ribosomal component, Y10B, the authors speculate that neurons might possess two distinct populations of PB-like structures, with a dendritic population primarily being used for mRNA storage, and a cell body population where typical mRNA degradation would occur. It is important to note, however, that an absence of Xrn1 in the dendritic compartment does not exclude a possible 5 - to 3 -RNA degradation by other, unknown, exoribonucleases. To support the authors’ hypothesis, a biochemical study measuring transcript levels from
A. Konecna et al. / Neuroscience Letters 466 (2009) 63–68
different fractions would be required. Nonetheless, dendritic PB stability does seem to differ from that of HeLa PBs [14,31]. Photobleaching experiments showed that Dcp1 marked granules have a low rate of recovery, which contrasts with the rapid exchange detected previously in HeLa cells. In neurons, the efficiency of Ago2 recovery is higher than for Dcp1, although it is still below HeLa cell values (e.g., difference of ∼20% for dendrites). Remarkably, the Dcp1 exchange rate in neurons is greatly improved upon stimulation with NMDA [27]. In addition, BDNF treatment of neurons induced the translocation of PBs into dendrites and the release of most of the Ago2 from the RNPs [13]. These results show that synaptic activation can change the dynamics and composition of PBs and suggests that stimulation of neuronal activity may contribute to localized derepression of translation. In another study, Zeitelhofer and colleagues observed that both PB and Ago2 particles disassemble following neuronal stimulation with NMDA, glutamate or BDNF, supporting a model where synaptic activity induces the derepression of Ago2 and PB-based translational silencing [46]. In contrast, only a small percentage of the cellular Ago2 in non-neuronal cells is confined to PBs and Ago2 undergoes rapid exchange between PBs and the cytoplasm, where it is likely to associate with target mRNAs in polyribosomes [27]. Also in contrast to neurons, disruption of fluorescence-detectable PBs in mammalian cell lines does not interfere with miRNA functions, further suggesting that miRNAs may find distinct ways to regulate translation in different tissues [11]. In summary, these results point to substantial functional differences between neurons and the nonneuronal cell lines. The studies described above suggest that PBs in neurons might play a more prominent role in miRNA-mediated repression than in other cells since their dynamics, composition and integrity change rapidly in response to synaptic activation. However, many open questions remain. Since miRNAs can be found in PBs they might be also present in other RNA granules such as transport RNPs or SGs [23]. Whereas PBs temporarily disassemble upon synaptic activity, transport RNPs appear to be more stable and SGs only form upon cellular stress. The current working hypothesis is that mRNAs cycle through silencing and derepression and may require assembly/disassembly of RNA granules or trafficking of the transcript between different RNPs. Recent evidence favors the latter, since neuronal PBs were shown to frequently come into stable contact (termed ‘docking’) with Stau2- or ZBP1-containing RNPs in vivo [46]. Increasingly, miRNAs are being considered as one of the key players in mechanisms of translational control in neurons. We are only beginning to understand their full engagement in molecular processes underlining higher cognitive functions of the brain. Until now, there is only one piece of evidence in vivo that a particular behavior triggers de-repression of the synaptic translation of a specific miRNA-target leading to the establishment of a stable memory in Drosophila [1]. In mammals, we are currently limited to observations of morphological changes in neurons with artificially altered levels of miRNA expression [36] during dendritic outgrowth or spine remodeling. These two cellular phenotypes imply the participation of numerous gene products, which might be controlled via miRNAs. The current challenge is to discover the interaction network of miRNAs with their targets, and in particular in dendrites. In the near future, we can expect the first insights into how miRNAs operate at the synapse from identification of miRNA targets among localized transcripts in dendrites. Activation of individual synapses could release the repression of localized mRNAs by miRNAs, thereby leading to translation, and modification of individual synaptic protein composition. This hypothesis is especially attractive since miRNA-mediated regulation would allow spatial and temporal restriction of translation in a reversible manner. In addition, RBPs are likely to regulate miRNA functions depending on the cellular environment. It is possible that RBPs might attenuate
67
or augment repressive effects of miRNAs and have an important impact on target recognition and miRNA loading [44]. Finally, the most critical point will be to establish the link between miRNAmediated silencing and mechanisms of synaptic plasticity in in vivo behavioral tests in mammals. In this respect, the availability of knock-out mice for a number of miRNAs to the neuroscientists will be of crucial importance in the future. Acknowledgements We apologize for omitted references due to space restraints. We thank Natasha Bushati, Gunter Meister and Stefan Hüttelmaier for critical comments. MAK is supported by the FWF, the RNA quality programme (Eurocores, ESF) and the HFSP. References [1] S.I. Ashraf, A.L. McLoon, S.M. Sclarsic, S. Kunes, Synaptic protein synthesis associated with memory is regulated by the RISC pathway in Drosophila, Cell 124 (2006) 191–205. [2] S.A. Barbee, P.S. Estes, A.M. Cziko, J. Hillebrand, R.A. Luedeman, J.M. Coller, N. Johnson, I.C. Howlett, C. Geng, R. Ueda, A.H. Brand, S.F. Newbury, J.E. Wilhelm, R.B. Levine, A. Nakamura, R. Parker, M. Ramaswami, Staufen- and FMRPcontaining neuronal RNPs are structurally and functionally related to somatic P bodies, Neuron 52 (2006) 997–1009. [3] D. Bartel, MicroRNAs: target recognition and regulatory functions, Cell 136 (2009) 215–233. [4] D. Berdnik, A. Fan, C. Potter, L. Luo, MicroRNA processing pathway regulates olfactory neuron morphogenesis, Curr. Biol. 18 (2008) 1754–1759. [5] E. Berezikov, F. Thuemmler, L. Van Laake, I. Kondova, R. Bontrop, E. Cuppen, R. Plasterk, Diversity of microRNAs in human and chimpanzee brain, Nat. Genet. 38 (2006) 1375–1377. [6] F. Besse, A. Ephrussi, Translational control of localized mRNAs: restricting protein synthesis in space and time, Nat. Rev. Mol. Cell Biol. 9 (2008) 971–980. [7] S.N. Bhattacharyya, R. Habermacher, U. Martine, E.I. Closs, W. Filipowicz, Relief of microRNA-mediated translational repression in human cells subjected to stress, Cell 125 (2006) 1111–1124. [8] S. Bicker, G. Schratt, microRNAs: tiny regulators of synapse function in development and disease, J. Cell. Mol. Med. 12 (2008) 1466–1476. [9] F.V. Bolduc, K. Bell, H. Cox, K.S. Broadie, T. Tully, Excess protein synthesis in Drosophila fragile X mutants impairs long-term memory, Nat. Neurosci. 11 (2008) 1143–1145. [10] C.R. Bramham, D.G. Wells, Dendritic mRNA: transport, translation and function, Nat. Rev. Neurosci. 8 (2007) 776–789. [11] C.Y. Chu, T.M. Rana, Translation repression in human cells by microRNA-induced gene silencing requires RCK/p54, PLoS Biol. 4 (2006) e210. [12] M. Costa-Mattioli, W.S. Sossin, E. Klann, N. Sonenberg, Translational control of long-lasting synaptic plasticity and memory, Neuron 61 (2009) 10–26. [13] N. Cougot, S.N. Bhattacharyya, L. Tapia-Arancibia, R. Bordonne, W. Filipowicz, E. Bertrand, F. Rage, Dendrites of mammalian neurons contain specialized Pbody-like structures that respond to neuronal activation, J. Neurosci. 28 (2008) 13793–13804. [14] A. Eulalio, I. Behm-Ansmant, E. Izaurralde, P bodies: at the crossroads of posttranscriptional pathways, Nat. Rev. Mol. Cell Biol. 8 (2007) 9–22. [15] A. Eulalio, E. Huntzinger, E. Izaurralde, GW182 interaction with Argonaute is essential for miRNA-mediated translational repression and mRNA decay, Nat. Struct. Mol. Biol. 15 (2008) 346–353. [16] W. Filipowicz, S.N. Bhattacharyya, N. Sonenberg, Mechanisms of posttranscriptional regulation by microRNAs: are the answers in sight? Nat. Rev. Genet. 9 (2008) 102–114. [17] R. Fiore, S. Khudayberdiev, M. Christensen, G. Siegel, S. Flavell, T. Kim, M. Greenberg, G. Schratt, Mef2-mediated transcription of the miR379-410 cluster regulates activity-dependent dendritogenesis by fine-tuning Pumilio2 protein levels, EMBO J. 28 (2009) 697–710. [18] T.M. Franks, J. Lykke-Andersen, TTP and BRF proteins nucleate processing body formation to silence mRNAs with AU-rich elements, Genes Dev. 21 (2007) 719–735. [19] A. Jakymiw, K.M. Pauley, S. Li, K. Ikeda, S. Lian, T. Eystathioy, M. Satoh, M.J. Fritzler, E.K. Chan, The role of GW/P-bodies in RNA processing and silencing, J. Cell Sci. 120 (2007) 1317–1323. [20] P. Jin, R.S. Alisch, S.T. Warren, RNA and microRNAs in fragile X mental retardation, Nat. Cell Biol. 6 (2004) 1048–1053. [21] Q. Jing, S. Huang, S. Guth, T. Zarubin, A. Motoyama, J. Chen, F. Di Padova, S.C. Lin, H. Gram, J. Han, Involvement of microRNA in AU-rich element-mediated mRNA instability, Cell 120 (2005) 623–634. [22] M. Kedde, M.J. Strasser, B. Boldajipour, J.A. Vrielink, K. Slanchev, C. le Sage, R. Nagel, P.M. Voorhoeve, J. van Duijse, U.A. Orom, A.H. Lund, A. Perrakis, E. Raz, R. Agami, RNA-binding protein Dnd1 inhibits microRNA access to target mRNA, Cell 131 (2007) 1273–1286. [23] M.A. Kiebler, G.J. Bassell, Neuronal RNA granules: movers and makers, Neuron 51 (2006) 685–690.
68
A. Konecna et al. / Neuroscience Letters 466 (2009) 63–68
[24] K. Kosik, The neuronal microRNA system, Nat. Rev. Neurosci. 7 (2006) 911–920. [25] M. Kye, T. Liu, S. Levy, N. Xu, B. Groves, R. Bonneau, K. Lao, K. Kosik, Somatodendritic microRNAs identified by laser capture and multiplex RT-PCR, RNA 13 (2007) 1224–1234. [26] P. Landgraf, M. Rusu, R. Sheridan, A. Sewer, N. Iovino, A. Aravin, S. Pfeffer, A. Rice, A. Kamphorst, M. Landthaler, A mammalian microRNA expression atlas based on small RNA library sequencing, Cell 129 (2007) 1401–1414. [27] A.K. Leung, P.A. Sharp, Function and localization of microRNAs in mammalian cells, Cold Spring Harb. Symp. Quant. Biol. 71 (2006) 29–38. [28] J. Liu, F.V. Rivas, J. Wohlschlegel, J.R. Yates III, R. Parker, G.J. Hannon, A role for the P-body component GW182 in microRNA function, Nat. Cell Biol. 7 (2005) 1261–1266. [29] G. Lugli, V. Torvik, J. Larson, N. Smalheiser, Expression of microRNAs and their precursors in synaptic fractions of adult mouse forebrain, J. Neurochem. 106 (2008) 650–661. [30] G. Meister, M. Landthaler, L. Peters, P.Y. Chen, H. Urlaub, R. Luhrmann, T. Tuschl, Identification of novel argonaute-associated proteins, Curr. Biol. 15 (2005) 2149–2155. [31] R. Parker, U. Sheth, P bodies and the control of mRNA translation and degradation, Mol. Cell 25 (2007) 635–646. [32] A. Pascale, M. Amadio, A. Quattrone, Defining a neuron: neuronal ELAV proteins, Cell. Mol. Life Sci. 65 (2008) 128–140. [33] J. Pena, C. Sohn-Lee, S. Rouhanifard, J. Ludwig, M. Hafner, A. Mihailovic, C. Lim, D. Holoch, P. Berninger, M. Zavolan, T. Tuschl, miRNA in situ hybridization in formaldehyde and EDC–fixed tissues, Nat. Meth. 6 (2009) 139–141. [34] L. Peters, G. Meister, Argonaute proteins: mediators of RNA silencing, Mol. Cell 26 (2007) 611–623. [35] M.M. Poon, S.H. Choi, C.A. Jamieson, D.H. Geschwind, K.C. Martin, Identification of process-localized mRNAs from cultured rodent hippocampal neurons, J. Neurosci. 26 (2006) 13390–13399. [36] G.M. Schratt, F. Tuebing, E.A. Nigh, C.G. Kane, M.E. Sabatini, M. Kiebler, M.E. Greenberg, A brain-specific microRNA regulates dendritic spine development, Nature 439 (2006) 283–289.
[37] G. Siegel, G. Obernosterer, R. Fiore, M. Oehmen, S. Bicker, M. Christensen, S. Khudayberdiev, P.F. Leuschner, C.J. Busch, C. Kane, K. Hubel, F. Dekker, C. Hedberg, B. Rengarajan, C. Drepper, H. Waldmann, S. Kauppinen, M.E. Greenberg, A. Draguhn, M. Rehmsmeier, J. Martinez, G.M. Schratt, A functional screen implicates microRNA-138-dependent regulation of the depalmitoylation enzyme APT1 in dendritic spine morphogenesis, Nat. Cell Biol. 11 (2009) 705–710. [38] K. Stark, B. Xu, A. Bagchi, W. Lai, H. Liu, R. Hsu, X. Wan, P. Pavlidis, A. Mills, M. Karayiorgou, J. Gogos, Altered brain microRNA biogenesis contributes to phenotypic deficits in a 22q11-deletion mouse model, Nat. Genet. 40 (2008) 751–760. [39] O. Steward, E.M. Schuman, Protein synthesis at synaptic sites on dendrites, Annu. Rev. Neurosci. 24 (2001) 299–325. [40] M.A. Sutton, E.M. Schuman, Dendritic protein synthesis, synaptic plasticity, and memory, Cell 127 (2006) 49–58. [41] D.M. Tiruchinapalli, M.D. Ehlers, J.D. Keene, Activity-dependent expression of RNA binding protein HuD and its association with mRNAs in neurons, RNA Biol. 5 (2008) 157–168. [42] S. Vasudevan, Y. Tong, J. Steitz, Switching from repression to activation: MicroRNAs can up-regulate translation, Science 318 (2007) 1931–1934. [43] G.A. Wayman, M. Davare, H. Ando, D. Fortin, O. Varlamova, H.Y. Cheng, D. Marks, K. Obrietan, T.R. Soderling, R.H. Goodman, S. Impey, An activity-regulated microRNA controls dendritic plasticity by down-regulating p250GAP, Proc. Natl. Acad. Sci. U.S.A. 105 (2008) 9093–9098. [44] L. Weinmann, J. Hock, T. Ivacevic, T. Ohrt, J. Mutze, P. Schwille, E. Kremmer, V. Benes, H. Urlaub, G. Meister, Importin 8 is a gene silencing factor that targets argonaute proteins to distinct mRNAs, Cell 136 (2009) 496–507. [45] R. Wong, A. Ghosh, Activity-dependent regulation of dendritic growth and patterning, Nat. Rev. Neurosci. 3 (2002) 803–812. [46] M. Zeitelhofer, D. Karra, P. Macchi, M. Tolino, S. Thomas, M. Schwarz, M. Kiebler, R. Dahm, Dynamic interaction between P-bodies and transport ribonucleoprotein particles in dendrites of mature hippocampal neurons, J. Neurosci. 28 (2008) 7555–7562.