Localization of the RNA-binding proteins Staufen1 and Staufen2 at the mammalian neuromuscular junction

Authors

  • Guy Bélanger,

    1. Department of Cellular and Molecular Medicine, and Center for Neuromuscular Disease, Faculty of Medicine, University of Ottawa, Ottawa, Ontario, Canada
    Search for more papers by this author
  • Mark A. Stocksley,

    1. Department of Cellular and Molecular Medicine, and Center for Neuromuscular Disease, Faculty of Medicine, University of Ottawa, Ottawa, Ontario, Canada
    Search for more papers by this author
  • Marie Vandromme,

    1. Laboratoire de Biologie Moléculaire et Cellulaire, Ecole Normale Supérieure (ENS) Lyon, Unité Médicale de Recherche (UMR) 5663, Lyon, France
    Search for more papers by this author
  • Laurent Schaeffer,

    1. Laboratoire de Biologie Moléculaire et Cellulaire, Ecole Normale Supérieure (ENS) Lyon, Unité Médicale de Recherche (UMR) 5663, Lyon, France
    Search for more papers by this author
  • Luc Furic,

    1. Département de Biochimie and Centre de Recherche en Sciences Neurologiques, Faculté de Médecine, Université de Montréal, Montréal, Québec, Canada
    Search for more papers by this author
  • Luc DesGroseillers,

    1. Département de Biochimie and Centre de Recherche en Sciences Neurologiques, Faculté de Médecine, Université de Montréal, Montréal, Québec, Canada
    Search for more papers by this author
  • Bernard J. Jasmin

    1. Department of Cellular and Molecular Medicine, and Center for Neuromuscular Disease, Faculty of Medicine, University of Ottawa, Ottawa, Ontario, Canada
    2. Ottawa Health Research Institute, Program in Molecular Medicine, Ottawa Hospital, Ottawa, Ontario, Canada
    Search for more papers by this author

Address correspondence and reprint requests to Dr Bernard J. Jasmin, Department of Cellular and Molecular Medicine, University of Ottawa, 451 Smyth Road, Ottawa, Ontario, Canada K1H 8M5. E-mail: jasmin@uottawa.ca

Abstract

Staufen is an RNA-binding protein, first identified for its role in oogenesis and CNS development in Drosophila. Two mammalian homologs of Staufen have been identified and shown to bind double-stranded RNA and tubulin, and to function in the somatodendritic transport of mRNA in neurons. Here, we examined whether Staufen proteins are expressed in skeletal muscle in relation to the neuromuscular junction. Immunofluorescence experiments revealed that Staufen1 (Stau1) and Staufen2 (Stau2) accumulate preferentially within the postsynaptic sarcoplasm of muscle fibers as well as at newly formed ectopic synapses. Western blot analyses showed that the levels of Stau1 and Stau2 are greater in slow muscles than in fast-twitch muscles. Muscle denervation induced a significant increase in the expression of Stau1 and Stau2 in the extrasynaptic compartment of both fast and slow muscles. Consistent with these observations, we also demonstrated that expression of Stau1 and Stau2 is increased during myogenic differentiation and that treatment of myotubes with agrin and neuregulin induces a further increase in the expression of both Staufen proteins. We propose that Stau1 and Stau2 are key components of the postsynaptic apparatus in muscle, and that they contribute to the maturation and plasticity of the neuromuscular junction.

Abbreviations used
AChR

acetylcholine receptor

EDL

extensor digitorum longus

MuSK

muscle-specific kinase

PMSF

phenylmethylsulfonyl fluoride

SDS

sodium dodecyl sulfate

TA

tibialis anterior

Stau1

Staufen1

Stau2

Staufen2

The postsynaptic sarcoplasm of the neuromuscular junction represents a specialized domain within skeletal muscle fibers. In addition to the accumulation of morphologically distinct myonuclei and several cytoskeletal and membrane proteins, which highlight the structural compartmentalization, the postsynaptic sarcoplasm also shows evidence of functional specialization (Cartaud and Changeux 1993). For example, several laboratories have shown that transcripts encoding several synaptic proteins display a clear asymmetric distribution as they are preferentially expressed within the postsynaptic sarcoplasm. Transgenic mouse studies and promoter analyses have revealed that in the case of the acetylcholine receptor (AChR) subunits (Burden 1998; Sanes and Lichtman 2001; Schaeffer et al. 2001), acetylcholinesterase (Chan et al. 1999) and utrophin (Gramolini et al. 1998) the synaptic accumulation of these transcripts results, at least partially, from a local transcriptional activation of their respective genes. More recent studies have further highlighted the contribution of a DNA motif termed an N-box (Koike et al. 1995; Duclert et al. 1996) and of the ets-related transcription factor GA-binding protein (GABP) in these transcriptional regulatory events (Schaeffer et al. 1998; Fromm and Burden 1998; Chan et al. 1999; Gramolini et al. 1999; Khurana et al. 1999).

Although transcriptional regulation is a key mechanism in maintaining high levels of specific proteins at the postsynaptic membrane, converging lines of evidence indicate that additional regulatory steps are involved (for review see Chakkalakal and Jasmin 2003). For example, translational regulatory mechanisms have recently been shown to contribute directly to the development and maintenance of Drosophila neuromuscular junctions (Sigrist et al. 2000). Furthermore, a recent study has shown that the synaptic localization of α-dystrobrevin 1 transcripts in mammalian muscle fibers relies on post-transcriptional events (Newey et al. 2001). Finally, it has also been shown that the overall levels in muscle cells of several synaptic transcripts, such as those encoding utrophin (Gramolini et al. 2001a,b), acetylcholinesterase (Boudreau-Larivière et al. 2000; Angus et al. 2001; Fuentes and Taylor 1993), and various subunits of protein kinase A (Hoover et al. 2002), are regulated at least partially via post-transcriptional mechanisms.

Previous work using Drosophila oocytes has shown that the polarized localization of distinct mRNAs such as oskar and bicoid relies on the presence of RNA-binding proteins that can direct transcripts to precise subcellular compartments during early development (for review see Kiebler and DesGroseillers 2000; Roegiers and Jan 2000; Jansen 2001; Palacios and Johnston 2001; Kloc et al. 2002). In this context, recent studies have highlighted the contribution of Staufen in the compartmentalization of specific mRNAs in developing oocytes. Staufen is a double-stranded RNA-binding protein involved in the intracellular transport of mRNAs via microtubules. Mammalian homologs of Staufen have recently been cloned and two genes, termed Staufen1 (Stau1) and Staufen2 (Stau2), have been described (Buchner et al. 1999; Marion et al. 1999; Wickham et al. 1999; Monshausen et al. 2001; Tang et al. 2001; Duchaîne et al. 2002). In the present study, we have therefore examined the expression and distribution of Stau1 and Stau2 in skeletal muscle in relation to the presence of neuromuscular junctions, and determined whether the levels of these RNA-binding proteins are modulated according to the state of differentiation and innervation of muscle fibers.

Materials and methods

Animal care and surgery

Control C57BL/6 mice were obtained from Charles River Laboratories (St Constant, PQ, Canada) and housed in the University of Ottawa Animal Care Facility. All surgical procedures were performed in accordance with the strict guidelines established by the Canadian Council on Animal Care. For these studies, the tibialis anterior (TA), extensor digitorum longus (EDL) and soleus muscles were used. In some cases, the lateral and medial gastrocnemius muscles were also excised. Because we were also interested in determining the levels of Staufen proteins in the fast and slow regions of a given muscle, these latter muscles were further dissected into their red (slow) and white (fast) compartments. Following excision, muscles were either flash frozen in liquid nitrogen or embedded in OCT compound (Thermo Shandon, Pittsburgh, PA, USA) and frozen in melting isopentane precooled with liquid nitrogen.

Hindlimb musculature was denervated by cutting and removing a short segment of the sciatic nerve in the mid-thigh region while the animals were anesthetized with halothane. One to 10 days later, animal were re-anesthetized, and soleus and EDL muscles were excised and rapidly frozen in liquid nitrogen. Formation in vivo of ectopic synapses was induced by directly injecting TA muscles with 30 μL of a cDNA solution containing 2 μg rat agrin, 2 μg mouse muscle-specific kinase (MuSK) and 1 μg nlsLacZ expression vector dissolved in physiological saline. To increase the uptake of plasmids, injected muscles were subsequently electroporated by applying 10 pulses of 20 ms at 200 V/cm at a frequency of 2 Hz via needle electrodes. Mice were killed 2 weeks after injection, and TA muscles were removed and fixed for 30 min in 3.7% formalin before LacZ staining. LacZ-positive fibers were then microdissected and processed for fluorescence experiments.

Cell culture

The mouse C2C12 cell line was used in these studies; cells were obtained from American Type Culture Collection (Manassas,VA, USA). Myoblasts were seeded on 60-mm culture dishes coated with Matrigel (Collaborative Biomedical Products, Bedfort, MA, USA) and grown in Dulbecco's modified Eagle's medium (Life Sciences/Gibco; Burlington, ON, Canada) supplemented with 50 U/mL penicillin, 50 μg/mL streptomycin and 10% (vv) fetal bovine serum in a water-saturated chamber at 37°C containing 5% CO2. Confluent myoblasts were induced to differentiate and fuse into myotubes by replacing the growth medium with differentiation medium that contained a low concentration of serum (5% horse serum). The culture medium was changed every 48 h. Myoblasts were harvested when they were approximately 50% confluent whereas myotubes were used after being in differentiation medium for 4 days. Agrin and neuregulin treatment of myotubes was performed as described previously (Gramolini et al. 1998, 1999). Muscle cells were treated simultaneously with these two factors because it has been shown that agrin and neuregulin activate transcription of synaptic genes through a common signaling cascade (Meier et al. 1998; Jones et al. 1999).

Immunoblotting

Cultured cells were washed in phosphate-buffered saline and solubilized in BX homogenization buffer containing 0.3 m sucrose, 60 mm NaCl, 15 mm Tris-HCl, pH 8.0, 10 mm EDTA, 0.1 mmβ-mercaptoethanol, 0.01 mm phenylmethylsulfonyl fluoride (PMSF), 0.01 mm benzamidine, 1 μg leupeptin, 10 μg peptatin A and 1 μg aprotinin (Bag and Wu 1996). Muscle tissue was homogenized with a Polytron in Tris-HCl, pH 8.0, containing 1% sodium deoxycholate, 10% sodium dodecyl sulfate (SDS), 0.5% Triton X-100, 1 mm PMSF, 5 mm iodoacetamide and 1 μg/mL aprotinin (Gramolini et al. 2001b). After centrifugation, the supernatant was recovered, aliquoted and stored at − 80°C. The concentration of proteins in each sample was determined using the Bradford assay (Bio-Rad, Hercules, CA, USA).

Immunoblotting experiments were performed as described by Gramolini et al. (1998). Briefly, extracts containing 100 μg protein were denaturated in SDS loading buffer and subjected to SDS polyacrylamide gel electrophoresis using an 11% gel. The proteins were then transferred on to a polyvinylidene difluoride membrane (Sigma, St Louis, MO, USA) and equal loading was confirmed by Ponceau staining of the membrane. Following transfer, the membranes were incubated with rabbit polyclonal antibodies directed against Stau1 (Wickham et al. 1999) or Stau2 (Duchaîne et al. 2002) and revealed using a commercially available chemifluorescence kit (NEN Life Sciences, Boston, MA, USA). As shown before (Wickham et al. 1999; Duchaîne et al. 2002), these polyclonal Stau1 and Stau2 antibodies are specific and do not cross-react with each other. The identity of the Staufen isoforms expressed in skeletal muscle cells was ascertained by comparing the molecular mass of Stau1 and Stau2 between muscle tissues and Neuro2A and HeLa cells. To further verify equal loading, the membranes were also incubated with a monoclonal antibody against α-tubulin (Sigma).

Immunofluorescence

Longitudinal serial cryostat sections (12 μm) of TA, EDL and soleus muscles were placed on Superfrost microscope slides (VWR Canlab, Montreal, PQ, Canada). Staufen immunoreactivity was detected using anti-Stau1 (rabbit polyclonal; Wickham et al. 1999) or anti-Stau2 (rabbit polyclonal; Duchaîne et al. 2002) antibodies and appropriate Cy3-conjugated secondary antibody (Jackson Laboratories, Bio/Can Scientific, Mississauga, ON, Canada). For double immunofluorescence experiments, the anti-Stau2 monoclonal antibody IC6 (Duchaîne et al. 2002) was also used. FITC-conjugated α-bungarotoxin (Molecular Probes, Eugene OR, USA) was used to label AChRs at the neuromuscular junctions.

Formalin-fixed, LacZ-positive fibers were incubated for 10 min in Tris-buffered saline containing 0.5% Triton X-100. After two washes in Tris-buffered saline, TA muscle fibers were permeabilized with a pre-incubation solution containing 0.1% Triton X-100. The fibers were then incubated overnight at 4°C with anti-Staufen antibodies followed by 1-h incubations first with biotin-conjugated anti-rabbit antibodies and then streptavidin–Texas red. These fibers were also stained with FITC-conjugated α-bungarotoxin and with Hoechst stain (Sigma).

Results

In a first set of experiments, we examined the distribution of Stau1 and Stau2 along skeletal muscle fibers of mouse TA muscle by immunofluorescence. Although a low level of staining was apparent in extrasynaptic regions of muscle fibers, expression of Stau1 (Figs 1a and b) and of Stau2 (not shown) was clearly not homogeneous; there was a pronounced accumulation at the neuromuscular junction as judged by their co-localization with AChR. As expected, however, the staining pattern did not coincide perfectly with the presence of the AChR because Staufen proteins are not integral membrane proteins. Immunofluorescence experiments performed on single microdissected fibers further confirmed the presence of both Stau1 (not shown) and Stau2 (Figs 1c–f) at the neuromuscular junction; the pattern was very intricate and staining appeared in between myonuclei identified by Hoechst stain. It is important to note that throughout these experiments both Stau1 and Stau2 displayed a similar pattern of distribution and, in fact, their localization was nearly identical in muscle (Figs 1g–i).

Figure 1.

Localization of Stau1 and Stau2 at mammalian neuromuscular junctions. (a, b) Example of a muscle section stained for the presence of neuromuscular junctions by α-bungarotoxin labeling of AChR and Stau1 respectively. Note the accumulation of Stau1 at the neuromuscular junction and the faint labeling elsewhere in the sarcoplasm. (c–f) Single TA dissected muscle fibers stained for AChR (c) and Stau2 (d). Myonuclei labeled with Hoechst stain (e). (f) Superimposed view obtained with the three different signals. (g–I) Muscle sections stained for Stau1 (g) and Stau2 (h) in double immunofluorescence experiments, and (i) the merged images; the distribution of Stau1 and Stau2 appears nearly identical in muscle. Bar in (b) 100 μm, (f) 10 μm and (i) 100 μm.

In separate experiments, we directly injected cDNAs encoding agrin and MuSK into TA muscles to induce the formation of ectopic pseudo-synapses at sites distant from the original postsynaptic membrane domains. At such ectopic synapses, a functional postsynaptic apparatus develops, and synapse-specific genes are expressed locally (Cohen et al. 1997; Meier et al. 1998; Jones et al. 1999). Accumulation of both Stau1 (not shown) and Stau2 (Fig. 2) at ectopic synapses, identified by the accumulation of AChR (Fig. 2a), were observed 2 weeks after the injection procedure. The degree of co-localization was somewhat less than that observed at mature neuromuscular junctions (see Fig. 1) most probably because these ectopic synapses are relatively immature. In any case, the results of these experiments taken together suggest that the RNA-binding proteins Stau1 and Stau2 are important components of the postsynaptic apparatus in skeletal muscle.

Figure 2.

Expression of Stau2 at ectopic synapses. TA muscles were injected with cDNAs encoding agrin, MuSK and LacZ. Two weeks later, β-galactosidase-positive fibers were dissected and stained for AChR (a) and Stau2 (b). Myonuclei were also labeled with Hoechst stain (c). (d) Superimposed view obtained with the three different signals. Note that Stau2 accumulates at ectopic synapses induced to form at sites distant from the original neuromuscular junctions. Bar 10 μm.

The preferential localization of Stau1 and Stau2 at the level of the neuromuscular junction was not restricted to TA muscles as an enrichment of both Staufen proteins was also observed in synaptic regions of soleus and EDL muscles (data not shown). As these two muscles are of similar size but with different contractile properties, they offer an opportunity to examine the relative abundance of Staufen proteins in a typical fast (EDL) and slow (soleus) muscle. Western blot analyses revealed that slow soleus muscles contained significantly more Stau1 and Stau2 than their fast-contracting counterparts (Fig. 3a). This difference could not be ascribed to differences in the functional role of these two muscles, i.e. ankle flexor versus ankle extensor, because a similar trend was observed when extracts from the red (slow) and white (fast) regions of both medial and lateral gastrocnemius muscles were examined (Fig. 3b). Because both Stau 1 and Stau 2 are known to exist in different isoforms (Wickham et al. 1999; Duchaîne et al. 2002), these experiments also suggested that skeletal muscles primarily express the Stau155 and Stau259 isoforms. This was confirmed by comparing the pattern of Staufen expression in muscle to that seen in HeLa and Neuro2A cells, which have been shown previously to contain specific isoforms (data not shown). Occasionally, distinct isoforms were also observed, albeit at very low levels (see Figs 4 and 5).

Figure 3.

Stau1 and Stau2 are differentially expressed in slow and fast muscles. (a) Western blot analysis of mouse fast (EDL) and slow (soleus; SOL) muscles showing that both Stau1 and Stau2 are more abundant in slow muscle whereas levels of α-tubulin, used as an loading control, are similar. (b) This difference is also evident between the red (R; slow) and white (W; fast) portions of the medial and lateral gastrocnemius muscles. Molecular mass markers are shown on the left.

Figure 4.

Expression of Stau1 and Stau2 is increased in denervated muscle. (a) Western blot analysis showing that denervation rapidly increases the levels of Stau1 and Stau2 in EDL muscle; α-tubulin was used as a loading control for these experiments. Similar results were obtained with soleus muscle. (b–e) Double fluorescence experiments using cryostat sections from denervated EDL muscle. Neuromuscular junctions were identified by α-bungarotoxin labeling of AChR (b and c). Note that in comparison to the staining pattern seen in control muscle (Fig. 1), the expression of both Stau1 (d) and Stau2 (e) increases markedly in extrasynaptic regions of denervated muscle but persists at the postsynaptic membrane. Bar 100 μm.

Figure 5.

Expression of Stau1 and Stau2 is regulated during myogenic differentiation and by nerve-derived factors. C2C12 cells were grown and harvested when they reached 50% confluence (myoblasts; MB) and after 4 days in differentiation medium (myotubes; MT). Note the increase in the levels of Stau1 (a) and Stau2 (b) in differentiated myotubes; (c) shows the levels of α-tubulin in these cells. Treatment of myotubes with neuregulin and agrin (+) for 4 days induced a further increase in the expression of both Stau1 and Stau2. Molecular mass markers are shown on the left.

Given the accumulation of Staufen proteins within the postsynaptic membrane domain of the neuromuscular junction, we next determined whether their expression was sensitive to the presence of the motor nerve. As shown in Fig. 4(a), expression of Stau1 and Stau2 was rapidly increased following denervation. Indeed, 1 and 2 days after cutting the sciatic nerve, levels of both Stau1 and Stau2 were increased by approximately twofold to threefold in EDL muscle. At later time points, for example 10 days after denervation, the increased expression of Staufen proteins was still apparent, particularly for Stau1. In soleus muscles, we also observed an increase in the expression of Stau1 and Stau2 following denervation (data not shown) suggesting that expression of this RNA-binding protein is regulated in a similar fashion in fast and slow muscles. Immunofluorescence experiments confirmed these findings and further demonstrated that, although staining within the postsynaptic sarcoplasm remained, a considerable increase in the amount of Staufen proteins occurred in extrasynaptic regions of muscle fibers (Figs 4b–e). These findings are consistent with the known increase in the expression of AChR subunits in extrajunctional compartments of denervated muscle fibers (Fromm and Burden 1998; Sanes and Lichtman 2001; Schaeffer et al. 2001).

Finally, we also examined the expression of Staufen proteins during myogenic differentiation using the mouse C2 cell line which has been used extensively to study the pattern of expression of synaptic proteins in differentiating muscle cells (for example Fuentes and Taylor 1993; Schaeffer et al. 1998; Angus et al. 2001). In mononucleated myoblasts, a low level of expression of both Stau1 and Stau2 was observed (Fig. 5). Expression of Staufen proteins increased markedly upon fusion of these myoblasts into multinucleated myotubes (∼ 3–4-fold). As expression of several synaptic components of the neuromuscular junction is known to be under the influence of nerve-derived trophic factors, such as agrin and neuregulin (see for review Fromm and Burden 1998; Sanes and Lichtman 2001; Schaeffer et al. 2001), we simultaneously treated differentiated myotubes with these factors because previous studies have shown that they can indeed activate synaptic gene transcription through a common signaling pathway (Jones et al. 1999). In agreement with the localization of Staufen proteins at the neuromuscular junction, treatment of myotubes with agrin and neuregulin led to a further increase in the expression of both Stau1 and Stau2 (Fig. 5).

Discussion

In recent years, it has become apparent that local transcriptional activation of genes cannot solely account for the compartmentalized expression of synaptic proteins at the postsynaptic membrane of the neuromuscular junction (Chakkalakal and Jasmin 2003). However, in comparison to the progress made in characterization of the molecular mechanisms regulating transcription of synaptic genes in muscle, there is a paucity of information concerning post-transcriptional, translational and post-translational events. Staufen is an RNA-binding protein that was first identified for its role in Drosophila oogenesis and CNS development. Since these early genetic studies, Staufen has become one of the best characterized RNA-binding proteins and two Staufen genes have been identified in mammals (Kiebler and DesGroseillers 2000; Roegiers and Jan 2000; Jansen 2001; Palacios and Johnston 2001; Kloc et al. 2002). Stau1 and Stau2 share approximately 50% sequence identity at the amino acid level (Buchner et al. 1999; Marion et al. 1999; Wickham et al. 1999; Monshausen et al. 2001; Tang et al. 2001; Duchaîne et al. 2002). Structurally, these proteins contain multiple copies of evolutionarily conserved consensus domains that are necessary for binding to RNA molecules, protein co-factors or microtubules. In contrast to Stau1, which appears ubiquitously distributed, Stau2 is mostly expressed in the brain. Here, we showed that both Stau1 and Stau2 are enriched within the postsynaptic compartment of neuromuscular junctions. In addition, we observed that expression of these RNA-binding proteins is modulated during myogenic differentiation and in response to muscle denervation. Moreover, levels of both Stau1 and Stau2 are clearly higher in slow compared with fast muscles. Although the Stau1 and Stau2 genes can each yield several protein isoforms via alternative splicing, our results further showed that skeletal muscle cells preferentially express one isoform of Stau1 and Stau2.

In Drosophila, Staufen has been shown to play multiple roles in the localization of distinct mRNAs (St Johnston 1995). During oogenesis, Staufen anchors bicoid transcripts to the anterior pole of oocytes, and it transports and localizes oskar mRNAs to the posterior pole. Additional studies have also shown that in the latter case Staufen participates in the regulation of oskar translation (Kim-Ha et al. 1995). Within the developing CNS, Staufen has been shown to play a similar role by contributing to the asymmetric distribution of prospero transcripts in dividing neuroblasts (Li et al. 1997; Broadus et al. 1998). By contrast, the role of Staufen in mammalian cells remains unclear. In neurons, however, Stau1 was shown to be expressed in the soma and dendrites but was absent from axons (Kiebler et al. 1999; Krichevsky and Kosik 2001; Monshausen et al. 2001). Recent studies using a Stau1/green fluorescence protein (GFP)-labeled fusion protein, have shown the movement of Stau1 in dendrites of living neurons (Köhrmann et al. 1999). Interestingly, several of these moving particles co-localized with the RNA-labeling dye SYTO-14, indicating that Stau1 may be a component of a ribonucleoprotein complex. Stau1 was recently co-immunoprecipitated from pur alpha-containing messenger ribonucleoproteins (mRNPs), which also contain the fragile X mental retardation protein and myosin Va (Ohashi et al. 2002). Together, these observations strongly support the involvement of Stau1 in the somatodendritic export of cellular RNA via microtubules. Consistent with a function in regulating translation in Drosophila, Stau1 has also been shown to associate with ribosomes and polysomes in mammalian neurons (Marion et al. 1999; Krichevsky and Kosik 2001; Luo et al. 2002). Although there is less information available on the function of Stau2, its localization in dendrites (Tang et al. 2001) and its relatively high sequence identity with Stau1 indicate nonetheless that Stau2 probably participates in the transport of RNA in a manner analogous, but not necessarily identical, to that involving Stau1 (Duchaîne et al. 2002).

The preferential localization of Stau1 and Stau2 within the postsynaptic sarcoplasm raises the interesting possibility that they fulfill a similar role in RNA transport and translation. Transcripts encoding several synaptic proteins, including AChR subunits, acetylcholinesterase and utrophin, have been shown to accumulate in the vicinity of synaptic myonuclei (Burden 1998; Sanes and Lichtman 2001; Schaeffer et al. 2001). In some cases, this local expression of mRNAs was shown to result from the local transcriptional activation of their respective genes. On the basis of the findings presented here, it may therefore be envisaged that once these mature synaptic transcripts enter the postsynaptic sarcoplasm, they are transported and/or anchored to specific subcellular sites for their subsequent translation. In this context, it is important to note that previous studies have shown the presence within the postsynaptic sarcoplasm of a specialized Golgi apparatus (Jasmin et al. 1989; Jasmin et al. 1995; for review Marchand and Cartaud 2002), a distinct network of microtubules (Jasmin et al. 1990) and the association in muscle cells of utrophin transcripts with cytoskeleton-bound polysomes (Gramolini et al. 2001a). Together these findings indicate that the postsynaptic compartment of muscle fibers contains not only the transcriptional machinery to ensure the local activation of a specific subset of genes but also all of the biosynthetic machinery necessary for the post-transcriptional, translational and post-translational processing of mature transcripts and proteins.

In comparison to the situation in neurons, the mechanisms of RNA transport at the neuromuscular junction may be more restricted given the relatively shorter distances separating the synaptic myonuclei and the postsynaptic membrane and cytoskeleton. This may be especially true for genes that are locally transcribed within synaptic myonuclei. However, the relative importance of RNA transport may be significantly greater when mRNAs encoding synaptic proteins accumulate within the postsynaptic sarcoplasm despite an absence of local transcriptional control. In particular, the accumulation of α-dystrobrevin 1 transcripts at the neuromuscular junction, which relies on post-transciptional events (Newey et al. 2001), may involve the preferential targeting and transport of these mRNAs to the synaptic sites. Nevertheless, in both mechanisms, anchoring of the transcripts within the postsynaptic sarcoplasm seems a further essential step in preventing the diffusion of the transcripts before local translation. Given their characteristics documented in other systems, Staufen proteins are good candidates for regulating RNA transport, anchoring and/or local translation. Indeed, components of the translational machinery have been shown to accumulate within this specialized region of muscle fibers in both Drosophila (Sigrist et al. 2000) and mouse (J. Thompson and B. J. Jasmin, unpublished observations).

We found that the levels of both Stau1 and Stau2 are considerably higher in slow-twitch muscles than in fast muscles. Although the functional implication of this difference may not be evident, it is important to note that several transcripts encoding synaptic proteins, including AChR subunits (Kues et al. 1995), ColQ (Krejci et al. 1999), utrophin (Gramolini et al. 2001a) and isoforms of protein kinase A (Hoover et al. 2002), have been shown to be more abundant in slow muscles as a result of transcriptional and/or post-transcriptional events. Therefore, it is also possible that the RNA transport may be of greater importance in slow-contracting muscle fibers. In other systems, the levels of expression of Stau1 and Stau2 also correlate with the amounts of human immunodeficiency virus type 1 genomic RNA found in virus particles (Mouland et al. 2000) and cellular RNA transported in dendrites (Tang et al. 2001) respectively.

Our results also suggest that, in addition to the myogenic program, expression of Staufen proteins is regulated by a combination of neural signals. First, nerve-derived factors such as agrin and neuregulin can increase the expression of these RNA-binding proteins. This observation is consistent with the preferential expression of Stau1 and Stau2 within the postsynaptic sarcoplasm of muscle fibers, and with the known effects of these factors on the expression of other synaptic proteins (Burden 1998; Sanes and Lichtman 2001; Schaeffer et al. 2001). In addition, the elimination of electrical activity via surgical denervation markedly affected Stau1 and Stau2 expression as observed for AChR subunits. Finally, the difference in the expression of Stau1 and Stau2 in slow and fast muscles further indicates that the pattern of nerve-evoked activity, i.e. tonic or phasic activation, is a key regulator of the levels of Staufen in muscle fibers. Therefore, the combined effects of patterned electrical activity and nerve-derived factors offer the possibility of significantly modulating the abundance of Staufen proteins in muscle which, in turn, may directly impact on the mechanisms controlling synaptic plasticity via its involvement in RNA transport and translation. Our results now raise the exciting possibility that physiological modulators regulate the function of Staufen-containing granules in both nerves and muscles, allowing coordinated responses at the neuromuscular junction.

In neurons, Stau1 and Stau2 define two specific populations of RNA granules (Duchaîne et al. 2002). Whether the same organization occurs in muscles is still unclear. Identification of the synaptic mRNAs that interact with Staufen proteins might prove difficult (Palacios and Johnston 2001) but should help to resolve this issue and indeed represents one of the key unanswered questions relating to the role of Staufen proteins in mRNA transport/anchoring at the neuromuscular junction. Therefore, future experiments should examine whether Stau1 and Stau2 interact specifically with a subset of mRNAs encoding synaptic proteins, or whether these RNA-binding proteins are part of a larger ribonucleoprotein complex that participates in RNA transport and translation at the neuromuscular junction. In any event, our present findings highlight the importance of post-transcriptional mechanisms in the assembly, maintenance and plasticity of the neuromuscular junction.

Acknowledgements

We thank John Lunde for expert technical assistance. This work was supported by operating grants from the Canadian Institutes of Health Research (CIHR), the Muscular Dystrophy Association of America, the Association Française contre les Myopathies (AFM), the Ontario Neurotrauma Foundation and an Ontario Premier's Research Excellence Award. GB is supported by a postdoctoral fellowship from the AFM and LF by a studentship from Fonds Formation Chercheur et Aide Recherche (FCAR)/Fonds de la Recherche en Santé du Québec (FRSQ) Santé. BJJ is a CIHR Investigator.

Ancillary