Neuronal RNA Granules: Movers and Makers
2006; Cell Press; Volume: 51; Issue: 6 Linguagem: Inglês
10.1016/j.neuron.2006.08.021
ISSN1097-4199
AutoresMichael Kiebler, Gary J. Bassell,
Tópico(s)RNA modifications and cancer
ResumoRNA localization contributes to cell polarity and synaptic plasticity. Evidence will be discussed that RNA transport and local translation in neurons may be more intimately linked than originally thought. Second, neuronal RNA granules, originally defined as intermediates involved in mRNA transport, are much more diverse in their composition and functions than previously anticipated. We focus on three classes of RNA granules that include transport RNPs, stress granules, and P bodies and discuss their potential functions in RNA localization, microRNA-mediated translational regulation, and mRNA degradation. RNA localization contributes to cell polarity and synaptic plasticity. Evidence will be discussed that RNA transport and local translation in neurons may be more intimately linked than originally thought. Second, neuronal RNA granules, originally defined as intermediates involved in mRNA transport, are much more diverse in their composition and functions than previously anticipated. We focus on three classes of RNA granules that include transport RNPs, stress granules, and P bodies and discuss their potential functions in RNA localization, microRNA-mediated translational regulation, and mRNA degradation. Translational control of localized mRNAs is a common mechanism for regulating protein expression in specific subdomains of a cell. It plays an important role in a number of processes, such as the formation of the body axes, asymmetric cell division, and cell motility (St Johnston, 2005St Johnston D. Nat. Rev. Mol. Cell Biol. 2005; 6: 363-375Crossref PubMed Scopus (420) Google Scholar). In neurons, localization of mRNAs at the synapse has been proposed as a mechanism for synaptic plasticity and thus learning and memory (Klann and Dever, 2004Klann E. Dever T.E. Nat. Rev. Neurosci. 2004; 5: 931-942Crossref PubMed Scopus (324) Google Scholar). Recently, it has also been shown that mRNA targeting and local protein synthesis can influence axon guidance and nerve regeneration (Willis et al., 2005Willis D. Li K.W. Zheng J.Q. Chang J.H. Smit A. Kelly T. Merianda T.T. Sylvester J. van Minnen J. Twiss J.L. J. Neurosci. 2005; 25: 778-791Crossref PubMed Scopus (326) Google Scholar). It is now generally accepted that localized mRNAs are often transported in large ribonucleoprotein particles (RNPs), which have been referred to as RNA granules (Ainger et al., 1993Ainger K. Avossa D. Morgan F. Hill S.J. Barry C. Barbarese E. Carson J.H. J. Cell Biol. 1993; 123: 431-441Crossref PubMed Scopus (404) Google Scholar, Knowles et al., 1996Knowles R.B. Sabry J.H. Martone M.E. Deerinck T.F. Ellisman M.H. Bassell G.J. Kosik K.S. J. Neurosci. 1996; 16: 7812-7820Crossref PubMed Google Scholar, Köhrmann et al., 1999Köhrmann M. Luo M. Kaether C. DesGroseillers L. Dotti C.G. Kiebler M.A. Mol. Biol. Cell. 1999; 10: 2945-2953Crossref PubMed Scopus (238) Google Scholar, Kiebler and DesGroseillers, 2000Kiebler M.A. DesGroseillers L. Neuron. 2000; 25: 19-28Abstract Full Text Full Text PDF PubMed Scopus (175) Google Scholar). Interestingly, there may be alternative models for RNA localization. In the Drosophila egg, hsp83 mRNA can be localized by degrading all transcripts that are not correctly localized, and transcripts can also become localized by passively diffusing through the cytoplasm until they are locally anchored (St Johnston, 2005St Johnston D. Nat. Rev. Mol. Cell Biol. 2005; 6: 363-375Crossref PubMed Scopus (420) Google Scholar, and references therein). In yeast, ER tubules are involved in the localization of ASH1 mRNA to the bud, suggesting a myosin-dependent cotransport of tubules and localized RNPs (Schmid et al., 2006Schmid M. Jaedicke A. Gu T.G. Jansen R.P. Curr. Biol. 2006; 16: 1538-1543Abstract Full Text Full Text PDF PubMed Scopus (107) Google Scholar). In this review, we focus on progress to define RNPs involved in mRNA localization, sorting, and degradation. Significant progress has been made in identifying components of such localized RNPs. A thorough comparison revealed that some RNP components are conserved across species and used in different cellular contexts (Kiebler and DesGroseillers, 2000Kiebler M.A. DesGroseillers L. Neuron. 2000; 25: 19-28Abstract Full Text Full Text PDF PubMed Scopus (175) Google Scholar, and references therein). However, very few of these proteins are essential for RNA localization in polarized neurons. This leaves us with several important questions. First, are there different types of RNA granules for different RNAs, or do all RNPs contain an obligatory subset of trans-acting factors? Second, how are RNAs translationally repressed until they reach their final destinations? Do these neuronal RNA granules, for example, contain microRNAs that silence the transported messages? Third, how do transport RNPs relate to stress granules and processing bodies, two recently discovered RNA granules in mammalian cells? Are these different RNA granules invariantly particles with specific functions and independent assembly pathways, or are these dynamic structures that share components or even interconvert? We will focus on neuronal RNA granules and review the evidence for their role in neuronal RNA transport, translational regulation, and possibly in RNA processing. Many in situ hybridization studies (ISH) studies had previously revealed the nonrandom localization of specific mRNAs to subcellular domains of polarized cells, such as yeast, oocytes, fibroblasts, and neurons (St Johnston, 2005St Johnston D. Nat. Rev. Mol. Cell Biol. 2005; 6: 363-375Crossref PubMed Scopus (420) Google Scholar). However, these studies only provided information on the steady-state distribution of mRNAs, and alternative approaches were needed in live cells to understand the dynamic process of mRNA localization. In a seminal study, Ainger et al., 1993Ainger K. Avossa D. Morgan F. Hill S.J. Barry C. Barbarese E. Carson J.H. J. Cell Biol. 1993; 123: 431-441Crossref PubMed Scopus (404) Google Scholar fluorescently labeled and microinjected myelin basic protein (MBP) mRNA into cultured oligodendrocytes. The MBP mRNA formed granules, which were rapidly transported along microtubules (MT) into processes at a rate of 0.2 μm/s. The observed RNA granules were heterogeneous in size and displayed persistent, oscillatory, or immobile characteristics. This work represented the first characterization of mRNA movements in living cells and suggested RNA granules as intermediates in a multistep pathway to localize mRNAs (Ainger et al., 1993Ainger K. Avossa D. Morgan F. Hill S.J. Barry C. Barbarese E. Carson J.H. J. Cell Biol. 1993; 123: 431-441Crossref PubMed Scopus (404) Google Scholar). Wilhelm and Vale were the first to call them “RNA transport particles” (Wilhelm and Vale, 1993Wilhelm J.E. Vale R.D. J. Cell Biol. 1993; 123: 269-274Crossref PubMed Scopus (298) Google Scholar); we therefore refer to them as transport RNPs. In neurons, the first evidence for the existence of transport RNPs (Table 1) came from Knowles et al., 1996Knowles R.B. Sabry J.H. Martone M.E. Deerinck T.F. Ellisman M.H. Bassell G.J. Kosik K.S. J. Neurosci. 1996; 16: 7812-7820Crossref PubMed Google Scholar. The fluorescent vital RNA dye SYTO14 depicted the dynamic movements of endogenous RNA granules into dendrites of cultured cortical neurons. These transport RNPs displayed rapid anterograde and retrograde trajectories that were dependent on MTs, consistent with an active transport process (Knowles et al., 1996Knowles R.B. Sabry J.H. Martone M.E. Deerinck T.F. Ellisman M.H. Bassell G.J. Kosik K.S. J. Neurosci. 1996; 16: 7812-7820Crossref PubMed Google Scholar). Furthermore, they contained many translational components, such as elongation factors and ribosomal proteins or even clusters of ribosomes, supporting the previous findings on MBP mRNA granules. These findings led to the hypothesis that transport RNPs move as discrete units together with at least some of the machinery needed to initiate translation upon reaching their final destination.Table 1Diversity of Cytoplasmic RNA Granules in NeuronsNameSynonymsDefinitionsKey ComponentsKey ReferencesTransport RNPsNeuronal RNA granules, RNA particlesMotile granules transporting mRNA.Staufen1, Staufen2, FMRP, ZBP1, hnRNPA2, CPEB, Purα, SMNKnowles et al., 1996Knowles R.B. Sabry J.H. Martone M.E. Deerinck T.F. Ellisman M.H. Bassell G.J. Kosik K.S. J. Neurosci. 1996; 16: 7812-7820Crossref PubMed Google Scholar, Köhrmann et al., 1999Köhrmann M. Luo M. Kaether C. DesGroseillers L. Dotti C.G. Kiebler M.A. Mol. Biol. Cell. 1999; 10: 2945-2953Crossref PubMed Scopus (238) Google Scholar, Zhang et al., 2001Zhang H.L. Eom T. Oleynikov Y. Shenoy S.M. Liebelt D.A. Dictenberg J.B. Singer R.H. Bassell G.J. Neuron. 2001; 31: 261-275Abstract Full Text Full Text PDF PubMed Scopus (296) Google Scholar, Zhang et al., 2006Zhang H.L. Xing L. Rossoll W.M. Wichterle H. Singer R.H. Bassell G.J. J. Neurosci. 2006; 26: 8622-8632Crossref PubMed Scopus (155) Google Scholar, Tang et al., 2001Tang S.J. Meulemans D. Vazquez L. Colaco N. Schuman E. Neuron. 2001; 32: 463-475Abstract Full Text Full Text PDF PubMed Scopus (177) Google Scholar, Huang et al., 2003Huang Y.S. Carson J.H. Barbarese E. Richter J.D. Genes Dev. 2003; 17: 638-653Crossref PubMed Scopus (196) Google Scholar, Kanai et al., 2004Kanai Y. Dohmae N. Hirokawa N. Neuron. 2004; 43: 513-525Abstract Full Text Full Text PDF PubMed Scopus (801) Google ScholarContain translational components.mRNAs in these RNA granules are translationally arrested during transport by the action of regulatory RNAs and RNA-binding proteins.Biochemical evidence for RNA granules with and without ribosomes.Stress granulesHeat stress granulesHarbor translationally arrested mRNAs that form in cells exposed to a broad range of stresses.TIA-1, TIAR, PABP, G3BP, 40S ribosomal subunitAnderson and Kedersha, 2006Anderson P. Kedersha N. J. Cell Biol. 2006; 172: 803-808Crossref PubMed Scopus (802) Google Scholar, Vessey et al., 2006Vessey J.P. Vaccani A. Xie Y. Dahm R. Karra D. Kiebler M.A. Macchi P. J. Neurosci. 2006; 26: 6496-6508Crossref PubMed Scopus (141) Google ScholarSort, remodel, and export specific RNAs for reinitiation or storage.Processing bodies (P bodies)Cytoplasmic bodies, Dcp1 bodies, GW bodiesSites of translational repression and/or mRNA degradation.Dcp1a, Lsm proteins, Rck/p54, GW182Anderson and Kedersha, 2006Anderson P. Kedersha N. J. Cell Biol. 2006; 172: 803-808Crossref PubMed Scopus (802) Google Scholar, Schratt et al., 2006Schratt G. Tuebing F. Nigh E.A. Kane C. Sabatini M.W. Kiebler M.A. Greenberg M.E. Nature. 2006; 439: 283-289Crossref PubMed Scopus (1369) Google Scholar, Vessey et al., 2006Vessey J.P. Vaccani A. Xie Y. Dahm R. Karra D. Kiebler M.A. Macchi P. J. Neurosci. 2006; 26: 6496-6508Crossref PubMed Scopus (141) Google ScholarContain RISC machinery including microRNAs.Do not contain ribosomal subunits.Since the first description of motile transport RNPs in oligodendrocytes and neurons, at least two additional classes of neuronal RNA granules exist: stress granules and P bodies that are sites for RNA storage and degradation, respectively. With regard to possible shared components between these neuronal RNA granules, there is evidence that several mRNA binding proteins, i.e. staufen1, staufen2, FMRP, SMN, CPEB, and pumilio2, can associate with SGs upon overexpression (Anderson and Kedersha, 2006Anderson P. Kedersha N. J. Cell Biol. 2006; 172: 803-808Crossref PubMed Scopus (802) Google Scholar). The presence of components in more than one type of RNA granule suggests that these RNPs are not homogeneous, but represent dynamic structures that a cell uses to sort mRNAs and regulate their translation and degradation. Open table in a new tab Since the first description of motile transport RNPs in oligodendrocytes and neurons, at least two additional classes of neuronal RNA granules exist: stress granules and P bodies that are sites for RNA storage and degradation, respectively. With regard to possible shared components between these neuronal RNA granules, there is evidence that several mRNA binding proteins, i.e. staufen1, staufen2, FMRP, SMN, CPEB, and pumilio2, can associate with SGs upon overexpression (Anderson and Kedersha, 2006Anderson P. Kedersha N. J. Cell Biol. 2006; 172: 803-808Crossref PubMed Scopus (802) Google Scholar). The presence of components in more than one type of RNA granule suggests that these RNPs are not homogeneous, but represent dynamic structures that a cell uses to sort mRNAs and regulate their translation and degradation. This leads us to an important question: are RNA granules homogeneous in composition and function, or do various types exist? Evidence for the latter hypothesis came from several laboratories that have identified mRNA binding proteins involved in mRNA localization in polarized cells. One key player is Staufen, a double-stranded RNA binding protein that plays important roles in localization of bicoid and oskar RNA to the anterior and posterior pole, respectively, in Drosophila oocytes (St Johnston, 2005St Johnston D. Nat. Rev. Mol. Cell Biol. 2005; 6: 363-375Crossref PubMed Scopus (420) Google Scholar). Köhrmann and colleagues were the first to express the mammalian Staufen homolog tagged with GFP in hippocampal neurons and to observe the MT-dependent recruitment of Staufen into RNA granules, which exhibited rapid bidirectional movements in dendrites (0.1–0.4 μm/s). This was the first demonstration of dynamic movement of an RNA binding protein in transport RNPs in living neurons (Köhrmann et al., 1999Köhrmann M. Luo M. Kaether C. DesGroseillers L. Dotti C.G. Kiebler M.A. Mol. Biol. Cell. 1999; 10: 2945-2953Crossref PubMed Scopus (238) Google Scholar). This approach was subsequently used by Zhang and colleagues to analyze the movement of granules (over 1 μm/s) containing the β-actin mRNA binding protein zipcode binding protein 1 (ZBP1) into growth cones of developing axons (Zhang et al., 2001Zhang H.L. Eom T. Oleynikov Y. Shenoy S.M. Liebelt D.A. Dictenberg J.B. Singer R.H. Bassell G.J. Neuron. 2001; 31: 261-275Abstract Full Text Full Text PDF PubMed Scopus (296) Google Scholar). Furthermore, anterograde trafficking of RNA granules containing ZBP1 and β-actin mRNA was stimulated by the neurotrophin NT-3, suggesting that granules can be regulated by physiological signals. Rook and colleagues took a different approach to visualize RNPs in neurons by employing the MS2-GFP tagging system to show that movement of granules containing the 3′-UTR for CaMKIIα mRNA in dendrites is dependent on synaptic activity (Rook et al., 2000Rook M.S. Lu M. Kosik K.S. J. Neurosci. 2000; 20: 6385-6393Crossref PubMed Google Scholar). Neuronal depolarization increased the fraction of the CaMKIIα reporter mRNA that was in the anterograde motile pool, suggesting that oscillatory granules were capable of sampling multiple nearby synapses. These finding indicate a possible role of synaptic input in the regulation of RNA granule motility in dendrites and localization at synapses. Taken together, these studies underscore that not only are mRNAs packaged into transport RNPs for directed movement but that there are signaling mechanisms likely regulating distinct populations of these transport RNPs and their dynamic interrelationships. Live cell imaging studies are consistent with the notion that transport RNPs may be propelled by molecular motors. First, MT-depolymerizing drugs were shown to decrease the levels of RNAs or mRNA binding proteins in neuronal processes (Knowles et al., 1996Knowles R.B. Sabry J.H. Martone M.E. Deerinck T.F. Ellisman M.H. Bassell G.J. Kosik K.S. J. Neurosci. 1996; 16: 7812-7820Crossref PubMed Google Scholar, Köhrmann et al., 1999Köhrmann M. Luo M. Kaether C. DesGroseillers L. Dotti C.G. Kiebler M.A. Mol. Biol. Cell. 1999; 10: 2945-2953Crossref PubMed Scopus (238) Google Scholar, Rook et al., 2000Rook M.S. Lu M. Kosik K.S. J. Neurosci. 2000; 20: 6385-6393Crossref PubMed Google Scholar). Second, the observed velocities of RNA granules (0.2–1.5 μm/s) and their anterograde or retrograde trajectories over long distances further suggest that they depend on kinesin and dynein motors. Indeed, previous work from the Carson group showed that antisense knockdown of the conventional kinesin heavy chain (KIF5b) impaired the ability of microinjected MBP mRNA to translocate into oligodendrocyte processes. Kanai and colleagues found that KIF5b associates with large RNA granules that contain 42 proteins and two well-studied dendritically localized mRNAs, CaMKIIα and Arc (Kanai et al., 2004Kanai Y. Dohmae N. Hirokawa N. Neuron. 2004; 43: 513-525Abstract Full Text Full Text PDF PubMed Scopus (801) Google Scholar). Overexpression of KIF5b increased mRNA localization into distal dendrites, whereas mRNA localization was reduced following its knockdown. Further work is needed to define the specific molecular interactions between transport RNP components and kinesin subunits to enable directed RNA movement. Nonetheless, this study provides compelling evidence that the anterograde transport of at least some transport RNPs is mediated by conventional kinesin in hippocampal neurons. It will also be important to study how transport RNPs that are capable of bidirectionally trafficking may become selectively captured within stimulated dendritic spines. Based on the landmark findings of Carson and others, a compelling model has been put forth of how mRNA localization could be achieved in an ordered, multistep pathway (Wilhelm and Vale, 1993Wilhelm J.E. Vale R.D. J. Cell Biol. 1993; 123: 269-274Crossref PubMed Scopus (298) Google Scholar). It predicts (1) the formation of transport RNPs as a functional complex, (2) their motor-dependent translocation to their destinations, (3) their anchoring to the local cytoskeleton, and (4) the translational derepression of the localized mRNAs. Since the formulation of this model, the challenge for the last 15 years has been to identify the key players involved in these processes and to understand these two processes at the molecular level. First attempts have been made to purify neuronal RNA granules in order to utilize proteomics to characterize the molecular composition of neuronal RNA granules in a more systematic manner and to begin validating their functional significance in RNA transport. First, Krichevsky and Kosik, 2001Krichevsky A.M. Kosik K.S. Neuron. 2001; 32: 683-696Abstract Full Text Full Text PDF PubMed Scopus (410) Google Scholar identified an unusually heavy sucrose gradient fraction beyond polysomes that contained ribosomes and Staufen1 and are believed to represent a fraction enriched for components of RNA granules. Kanai et al., 2004Kanai Y. Dohmae N. Hirokawa N. Neuron. 2004; 43: 513-525Abstract Full Text Full Text PDF PubMed Scopus (801) Google Scholar isolated large CaMKIIα and Arc RNA-containing granules from adult mouse brain that associate with KIF5. Using the RNA interference (RNAi) approach, they showed that four of the proteins identified were important for the localization of a CaMKIIα reporter: Purα, hnRNP U, polypyrimidine tract binding protein-associated splicing factor (PSF), and Staufen1. SYNCRIP (hnRNP-Q1) was another mRNA binding protein identified, although its knockdown in neurons had no effect on CaMKIIα mRNA localization. Also identified in this study was the fragile X mental retardation protein (FMRP), an mRNA binding protein that traffics in RNA granules in dendrites (Antar et al., 2004Antar L.N. Afroz R. Dictenberg J.B. Carroll R.C. Bassell G.J. J. Neurosci. 2004; 24: 2648-2655Crossref PubMed Scopus (303) Google Scholar). Taken together, the study by Kanai et al. was groundbreaking in three ways. First, it showed that conventional kinesin is involved in the transport of RNA granules to dendrites, although it is still unclear which RNP component binds directly to which kinesin subunit(s). Second, it identified with Purα, hnRNP U, and PSF three additional essential trans-acting factors for dendritic RNA transport. Third, it strongly suggested that granules are composed of many mRNA binding proteins and a number of these may not be essential for mRNA transport, but instead regulate aspects of RNP assembly, translation, or stability. Elvira et al., 2006Elvira G. Wasiak S. Blandford V. Tong X.K. Serrano A. Fan X. del Rayo Sanchez-Carbente M. Servant F. Bell A.W. Boismenu D. et al.Mol. Cell. Proteomics. 2006; 5: 635-651Crossref PubMed Scopus (202) Google Scholar biochemically isolated RNA granules in developing rat brain that were enriched for β-actin but not for CaMKIIα mRNA and contained ribosomes, a large set of RNA binding proteins, MT-associated proteins, and several novel proteins. One protein identified was ZBP1, which was known to be required for localization of β-actin mRNA granules (Zhang et al., 2001Zhang H.L. Eom T. Oleynikov Y. Shenoy S.M. Liebelt D.A. Dictenberg J.B. Singer R.H. Bassell G.J. Neuron. 2001; 31: 261-275Abstract Full Text Full Text PDF PubMed Scopus (296) Google Scholar). In addition, Staufen and hnRNP-A2 were identified as also being required factors for mRNA localization. Interestingly, both proteomic studies identified a number of RNA binding proteins, such as Staufen and SYNCRIP, which was shown by Bannai and colleagues to colocalize with GFP-tagged Staufen1 and inositol 1,4,5-trisphosphate receptor type 1 mRNA in dendritic granules (Elvira et al., 2006Elvira G. Wasiak S. Blandford V. Tong X.K. Serrano A. Fan X. del Rayo Sanchez-Carbente M. Servant F. Bell A.W. Boismenu D. et al.Mol. Cell. Proteomics. 2006; 5: 635-651Crossref PubMed Scopus (202) Google Scholar, and reference therein). Another noteworthy protein family detected in both studies is the family of DEAD box helicases, which has previously been implicated in RNP assembly. Elvira et al., 2006Elvira G. Wasiak S. Blandford V. Tong X.K. Serrano A. Fan X. del Rayo Sanchez-Carbente M. Servant F. Bell A.W. Boismenu D. et al.Mol. Cell. Proteomics. 2006; 5: 635-651Crossref PubMed Scopus (202) Google Scholar observed that DEAD Box 3 colocalized with RNA granule markers and also exhibited BDNF-regulated movements in live hippocampal neurons. There were also some interesting distinctions between these two studies. Elvira et al., 2006Elvira G. Wasiak S. Blandford V. Tong X.K. Serrano A. Fan X. del Rayo Sanchez-Carbente M. Servant F. Bell A.W. Boismenu D. et al.Mol. Cell. Proteomics. 2006; 5: 635-651Crossref PubMed Scopus (202) Google Scholar identified ZBP1 in their proteomic analysis of developing brain, whereas Kanai et al., 2004Kanai Y. Dohmae N. Hirokawa N. Neuron. 2004; 43: 513-525Abstract Full Text Full Text PDF PubMed Scopus (801) Google Scholar did not. The authors hypothesized that there may be different types of RNA granules that are developmentally regulated and respond to distinct physiological signals. Together, these studies open the door toward a more complete understanding of the composition of neuronal RNA granules and allow us to address some exciting new questions. One shortcoming of these studies, however, is that very few components that came out of these proteomics experiments have been validated in RNA localization assays, and it is unclear up to now whether the majority of the identified components are indeed part of transport RNPs. What is the experimental evidence for an involvement of trans-acting factors in RNA localization? The most-studied example is ZBP1 and its role in the localization of β-actin mRNA into neurites and growth cones (Zhang et al., 2001Zhang H.L. Eom T. Oleynikov Y. Shenoy S.M. Liebelt D.A. Dictenberg J.B. Singer R.H. Bassell G.J. Neuron. 2001; 31: 261-275Abstract Full Text Full Text PDF PubMed Scopus (296) Google Scholar). Antisense oligonucleotides directed against the 54 nt β-actin zipcode, which disrupt ZBP1 binding to the zipcode in vitro, blocked the NT-3-induced localization of β-actin mRNA into neurites and growth cones (Zhang et al., 2001Zhang H.L. Eom T. Oleynikov Y. Shenoy S.M. Liebelt D.A. Dictenberg J.B. Singer R.H. Bassell G.J. Neuron. 2001; 31: 261-275Abstract Full Text Full Text PDF PubMed Scopus (296) Google Scholar). Furthermore, morpholino antisense oligonucleotides to knock-down ZBP1 resulted in reduced dendritic β-actin mRNA localization in cultured hippocampal neurons, whereas the localization of CaMKIIα mRNA was unaffected (Eom et al., 2003Eom T. Antar L.N. Singer R.H. Bassell G.J. J. Neurosci. 2003; 23: 10433-10444PubMed Google Scholar). Another trans-acting factor that has been implicated in dendritic RNA transport has been mammalian Staufen2. Overexpression of dominant-negative Staufen2 significantly reduced the level of ethidium bromide-stained RNA in dendrites of polarized neurons (Tang et al., 2001Tang S.J. Meulemans D. Vazquez L. Colaco N. Schuman E. Neuron. 2001; 32: 463-475Abstract Full Text Full Text PDF PubMed Scopus (177) Google Scholar). Unfortunately, there is still not a single mRNA identified in mammalian cells that is recognized by Staufen proteins and that has been shown to be transported within Staufen-containing RNA granules. This is in sharp contrast to Drosophila Staufen, which is involved in the localization of bicoid, oskar, and prospero RNAs in the Drosophila oocyte and embryo (St Johnston, 2005St Johnston D. Nat. Rev. Mol. Cell Biol. 2005; 6: 363-375Crossref PubMed Scopus (420) Google Scholar). Finally, a recent study reported an essential role of the cytoplasmic polyadenylation element binding protein 1 (CPEB1) in dendritic RNA transport. Huang et al., 2003Huang Y.S. Carson J.H. Barbarese E. Richter J.D. Genes Dev. 2003; 17: 638-653Crossref PubMed Scopus (196) Google Scholar showed that functional but not mutated CPEs within the 170 nt 3′-UTR of wild-type CaMKIIα mRNA are sufficient to target a reporter RNA into dendrites of hippocampal neurons. To demonstrate a requirement for CPEB1 in dendritic mRNA localization, cultured neurons from CPEB1 knockout mice showed reduced localization of EGFP reporters harboring CPE sequences in the 3′-UTR. The study by Huang et al., 2003Huang Y.S. Carson J.H. Barbarese E. Richter J.D. Genes Dev. 2003; 17: 638-653Crossref PubMed Scopus (196) Google Scholar showed that endogenous mRNAs that contain a CPE, e.g., CaMKIIα and MAP2 mRNAs, were reduced in synaptosome preparations isolated from cultured neurons that were infected with a dominant-negative CPEB construct. Reduced mRNA levels of MAP2 mRNA were also confirmed by FISH, further suggesting a role for CPEB in mRNA localization. As other studies have identified a role for different localization elements within the CaMKIIα mRNA 3′-UTR other than the CPE (Rook et al., 2000Rook M.S. Lu M. Kosik K.S. J. Neurosci. 2000; 20: 6385-6393Crossref PubMed Google Scholar), it is likely that there may be additional localization elements, and further work is needed to assess whether the CPE is necessary for mRNA localization. An aspect of mRNA localization that is as yet unclear is how transport RNPs are assembled. An RNP complex containing the survival of motor neuron protein (SMN) and gemin proteins facilitates the assembly of spliceosomal RNPs, which may play a comparable role in assembly of transport RNPs in neurons (Monani, 2005Monani U. Neuron. 2005; 48: 885-896Abstract Full Text Full Text PDF PubMed Scopus (241) Google Scholar). Rossoll and coworkers have shown that SMN binds SYNCRIP (Monani, 2005Monani U. Neuron. 2005; 48: 885-896Abstract Full Text Full Text PDF PubMed Scopus (241) Google Scholar, and reference therein), which was identified in the above proteomics screens of RNA granules (Kanai et al., 2004Kanai Y. Dohmae N. Hirokawa N. Neuron. 2004; 43: 513-525Abstract Full Text Full Text PDF PubMed Scopus (801) Google Scholar, Elvira et al., 2006Elvira G. Wasiak S. Blandford V. Tong X.K. Serrano A. Fan X. del Rayo Sanchez-Carbente M. Servant F. Bell A.W. Boismenu D. et al.Mol. Cell. Proteomics. 2006; 5: 635-651Crossref PubMed Scopus (202) Google Scholar). SMN-GFP granules are actively transported in association with gemin proteins in live neurons (Zhang et al., 2006Zhang H.L. Xing L. Rossoll W.M. Wichterle H. Singer R.H. Bassell G.J. J. Neurosci. 2006; 26: 8622-8632Crossref PubMed Scopus (155) Google Scholar, and references therein). The inherited loss of SMN is the cause of the neurodegenerative disease spinal muscular atrophy (SMA). Mouse neurons cultured from a transgenic mouse model of SMA have impaired localization of β-actin mRNA in axonal growth cones, suggesting possible interactions of SMN complexes with mRNA binding proteins, e.g., ZBP1 and SYNCRIP involved in mRNA localization (Monani, 2005Monani U. Neuron. 2005; 48: 885-896Abstract Full Text Full Text PDF PubMed Scopus (241) Google Scholar, and references therein). Future work is needed to understand whether impaired mRNP assembly and localization contribute to SMA. Most approaches described so far tried to understand the molecular mechanisms of how individual mRNAs are transported in the form of granules using cultured neurons as a model. The identification of various components of the translational machinery in dendrites and near synapses (Klann and Dever, 2004Klann E. Dever T.E. Nat. Rev. Neurosci. 2004; 5: 931-942Crossref PubMed Scopus (324) Google Scholar, and references therein) revealed several important questions. Is RNA transport coupled with translation, and how is this achieved? Are mRNAs generally repressed within the observed transport RNPs? What causes the activation of translation? In addition, more than one class of transport RNP might exist. In the case of mammalian staufen proteins, there is biochemical evidence that some RNA granules contain ribosomes, whereas others do not. When fractionated by size, the largest Staufen pools contained ribosomal and ER markers, whereas the smaller RNA granules were free of ribosomes and ER but cofractionated with conventiona
Referência(s)