Building the Vertebrate Neuromuscular Synapse

Building the Vertebrate Neuromuscular Synapse Steven J. Burden Molecular Neurobiology Program, Skirball Institute, NYU Medical School, 540 First Avenue, New York City, New York 10016 Received 10 June 2002; accepted 12 June 2002 2002 Wiley Periodicals, Inc. J Neurobiol 53: 501 511, 2002 Keywords: MuSK; Agrin; Rapsyn; neuromuscular junction; acetylcholine receptor INTRODUCTION Neuromuscular synapses form following a series of complex interactions between motor neurons, muscle fibers, and Schwann cells that culminate in the formation of a highly specialized postsynaptic membrane and a highly differentiated nerve terminal (Son and Thompson, 1995; Burden, 1998; Sanes and Lichtman, 2001; Schaeffer et al., 2001). In adult muscle, motor axon terminals are situated in shallow depressions of the muscle cell membrane, which is invaginated further into deep and regular folds, termed postjunctional folds. Acetylcholine receptors (AChRs) and additional proteins are concentrated at the crests of these postjunctional folds, while other proteins, including sodium channels, are enriched in the troughs of the postjunctional folds. The genes encoding many of these muscle-derived synaptic proteins are transcribed preferentially in the myofiber nuclei that are situated near the synaptic site, resulting in an accumulation of the encoded mrnas near synaptic nuclei. The nerve terminal is likewise spatially organized, and its substructural organization reflects that of the postsynaptic membrane. Synaptic vesicles are clustered adjacent to poorly characterized specializations of the presynaptic membrane, termed active zones, which are the sites of synaptic vesicle fusion. Active zones are organized at Correspondence to: S. J. Burden (burden@saturn.med.nyu.edu). Contract grant sponsor: NIH. 2002 Wiley Periodicals, Inc. Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/neu.10137 regular intervals and are aligned precisely with the mouths of the postjunctional folds. This precise registration of active zones and postjunctional folds insures that acetylcholine encounters a high concentration of AChRs within microseconds after release, thereby facilitating synaptic transmission. This precise organization of molecules in presynaptic and postsynaptic membranes belies the concept that the neuromuscular synapse is a simple synapse. Rather, the substructure of presynaptic and postsynaptic membranes suggests that complex mechanisms, requiring spatially restricted signaling between presynaptic and postsynaptic cells, are required to assemble the synapse and to coordinate presynaptic and postsynaptic differentiation. Myogenesis and the First Contacts between Motor Axons and Muscle Multinucleated skeletal muscle fibers form during development (beginning at E11 in the mouse) by fusion of precursor myoblasts. Shortly after myotubes begin to form, motor axons grow into the developing muscle (at E12 E13 in the mouse) and form a main intramuscular nerve, which in mammals extends through the central region of the muscle, perpendicular to the long axis of the myotubes. Individual motor axons branch and terminate adjacent to the main intramuscular nerve, resulting in a narrow, distinct endplate zone, marked by presynaptic nerve terminals, clusters of AChRs, and elevated levels of AChR gene expression, in the middle of the muscle. This spatial patterning of skeletal muscle cells has 501

502 Burden been thought to arise from the focal release of motor neuron-derived signals, including Agrin, that activate receptor tyrosine kinases in the myotube. Agrin Agrin is a 200-kDa protein that is synthesized by motor neurons, transported in motor axons, and released at synaptic sites, where it organizes postsynaptic differentiation (McMahan, 1990; Ruegg and Bixby, 1998; Sanes and Lichtman, 2001). Agrin is necessary for synapse formation, as mice lacking Agrin fail to form neuromuscular synapses (Gautam et al., 1996). Moreover, Agrin is sufficient to induce certain aspects of postsynaptic-like differentiation in cultured muscle cells and at ectopic sites in adult muscle (Wallace, 1989; Meier et al., 1997, 1998; Rimer et al., 1998). Agrin stimulates postsynaptic differentiation by activating MuSK, a receptor tyrosine kinase that is expressed selectively in skeletal muscle and concentrated at synaptic sites (Jennings et al., 1993; Valenzuela et al., 1995; Glass and Yancopoulos, 1997). Agrin contains a laminin-binding domain, nine follistatin-like repeats, one laminin type III domain, three laminin-g domains and four epidermal growth factor (EGF)-like signaling domains (Ruegg and Bixby, 1998) [Fig. 1(A)]. The four EGF-like domains and the three laminin G domains are contained in the carboxyl-terminal region, which is sufficient for activating MuSK and inducing the clustering of MuSK, AChRs, Rapsyn, and Acetylcholinesterase (AChE) in cultured myotubes. The laminin-binding site at the amino-terminus is thought to be responsible for the association of Agrin with the extracellular matrix. The agrin gene is expressed in a variety of cell types. Alternative splicing generates multiple Agrin isoforms that differ in their AChR clustering efficiency. The isoform that is most effective in clustering synaptic proteins is expressed in neurons, including motor neurons, whereas other Agrin isoforms are expressed in additional cell types, including skeletal muscle cells. The active, neuronal-specific isoforms of Agrin contain 8, 11, or 19 amino acids at a splice site, referred to as the Z site in rat Agrin and the B site in chick Agrin (Ruegg and Bixby, 1998; Sanes and Lichtman, 2001). MuSK MuSK is a critical component of an Agrin receptor complex. MuSK is expressed in Torpedo electric organ and in skeletal muscle, where it is concentrated in the postsynaptic membrane. The extracellular region Figure 1 (A) The domain organization of Agrin, MuSK and Rapsyn. Motor-neuron derived Agrin contains a laminin-binding domain (LN), nine follistatin repeats (F), a laminin-type III domain (LN-III), four EGF-like domains, three laminin G domains (LN-G) and an 8, 11, or 19 amino acid insert (Z-isoforms). MuSK contains four immunoglobulin-like domains (Ig), a cysteine-rich domain (CRD), a juxtamembrane region (JM) and a kinase domain. Rapsyn contains a myristolation consensus sequence (N-Myr), seven tetratrichopeptide repeats (TPR), a coiled-coil domain and a RING-H2 domain. (B) Model for Agrin/MuSK signaling. Agrin stimulation of MuSK leads to the clustering of critical muscle-derived proteins, including MuSK, AChRs, Rapsyn, muscle-derived Neuregulin (Nrg-1), and ErbBs, tyrosine phosphorylation of AChRs, activation of synapsespecific gene expression and the induction of a retrograde signal for presynaptic differentiation. of mammalian MuSK contains four immunoglobulinlike domains and a cysteine-rich domain; in avians, fish, and amphibians, the extracellular region also contains a kringle domain (Jennings et al., 1993; Valenzuela et al., 1995; Fu et al., 1999; Ip et al., 2000) [Fig. 1(A)]. The intracellular region of MuSK contains a 50 amino acid juxtamembrane domain, a kinase domain and a short, eight amino acid carboxyterminal tail. Mice deficient in MuSK (DeChiara et al., 1996), like agrin mutant mice, lack normal neuromus-

Neuromuscular Synapse 503 cular synapses. The similar phenotype of agrin and MuSK mutant mice, at least at birth, is consistent with the idea that MuSK is a component of an Agrin receptor complex. Both agrin and MuSK mutant mice are immobile, cannot breathe and die at birth. Muscle differentiation is normal in agrin and MuSK mutant mice, but muscle fibers in MuSK mutant mice lack all known features of postsynaptic differentiation at all stages of development. Muscle-derived proteins, including AChRs, AChE, ErbB3, and ErbB4, receptors for Neuregulin-1 (Nrg-1), which are concentrated at synapses in normal mice, are uniformly distributed in MuSK mutant myofibers. In addition, AChR genes, which are normally transcribed selectively in synaptic nuclei of normal muscle fibers, are transcribed at similar rates in synaptic and nonsynaptic nuclei of muscle fibers from MuSK mutant mice. Synaptic differentiation is similarly defective in agrin and MuSK mutant mice, at least at birth. Earlier in development (E14), however, AChR clusters are evident and concentrated in the central region of muscle from agrin mutant mice (Lin et al., 2001), whereas AChR clusters are absent from MuSK mutant mice at all stages of development. These results indicate that Agrin is neither required to cluster AChRs in vivo nor to position these AChR clusters in the central region of the muscle. Therefore, Agrin appears to be necessary to maintain, rather than to induce AChR clusters in vivo (see below) (Lin et al., 2001; Yang et al., 2001). In contrast, in cultured myotubes, Agrin induces the formation of AChR clusters. These differing roles for Agrin may be a consequence of the neuralindependent mechanisms that can pattern AChR expression in developing muscle in vivo (see below) and the disruption of these muscle-autonomous patterning mechanisms in cultured myotubes, which are derived from myoblasts that are dissociated from embryonic muscle. Presynaptic differentiation is also aberrant in agrin and MuSK mutant mice, as motor axons fail to stop or differentiate, and instead wander throughout the muscle. Taken together with experiments showing that MuSK is activated by Agrin, these results indicate that Agrin stimulation of MuSK leads to the clustering of critical muscle-derived proteins, including MuSK, AChRs, and ErbBs, activation of synapse-specific gene expression, and the induction of a retrograde signal for presynaptic differentiation (DeChiara et al., 1996; Gautam et al., 1996). Although clustering of postsynaptic proteins is thought to be a direct consequence of MuSK signaling [Fig. 1(B)], defects in presynaptic differentiation and synapse-specific transcription could be attributed to a direct or indirect role for Agrin/MuSK in these aspects of synaptic differentiation. For example, although clustered MuSK could itself function as a stop/differentiation signal for presynaptic differentiation, MuSK signaling may regulate presynaptic differentiation by controlling the synthesis or organization of other proteins that serve as retrograde signals. Likewise, although synapsespecific transcription may be regulated directly by a transcriptional pathway stimulated by MuSK, activated MuSK may indirectly control synapse-specific transcription by clustering other ligands/receptors, notably Nrg-1 and ErbBs, that stimulate synaptic gene expression [Fig. 1(B), see below]. How does MuSK activation lead to postsynaptic differentiation? Agrin stimulates the rapid phosphorylation of MuSK, and the kinase activity of MuSK is essential for Agrin to stimulate clustering and tyrosine phosphorylation of AChRs. Well-characterized signaling pathways (e.g., MAP kinase, PI3-kinase, PLCgamma), however, are neither activated by Agrin/ MuSK signaling nor are required for Agrin to stimulate AChR clustering (Wallace, 1988; Herbst and Burden, 2000). Signaling downstream from MuSK depends upon phosphorylation of a tyrosine residue (Y553) in the juxtamembrane region of MuSK (Zhou et al., 1999; Herbst and Burden, 2000). Phosphorylation of this tyrosine is required to fully activate MuSK kinase activity and to recruit a downstream signaling component(s) required for all aspects of MuSK signaling, including tyrosine phosphorylation and clustering of AChRs, synapse-specific transcription and presynaptic differentiation (Herbst and Burden, 2000; Herbst et al., in press). At present, however, little is known about the protein(s) that is recruited to the juxtamembrane tyrosine in MuSK, which presumably recruits additonal proteins that mediate the signaling pathways leading to presynaptic and postsynaptic differentiation. There is evidence, however, that Rac and Cdc 42, small GTP-binding proteins that regulate actin organization, are required for Agrin to stimulate AChR clustering (Weston et al., 2000). Moreover, at least one kinase acts downstream from MuSK and is recruited and/or activated by Agrin-activated MuSK, because staurosporine, a protein kinase inhibitor, inhibits Agrin-induced AChR clustering without blocking tyrosine phosphorylation of MuSK (Wallace, 1994; Ferns et al., 1996). Agrin stimulates the rapid tyrosine phosphorylation of MuSK expressed in myotubes (Glass et al., 1996). Agrin, however, does not stimulate tyrosine phosphorylation of MuSK that is force-expressed in fibroblasts or myoblasts (Glass et al., 1996). These experiments suggest that Agrin does not bind directly to MuSK and that other components expressed selectively by skeletal muscle, termed MASC, are required

504 Burden for Agrin to activate MuSK. Despite substantial effort and encouraging reports (Glass and Yancopoulos, 1997), MASC has yet to be identified. It remains possible that a coligand or posttranslational modification of Agrin or MuSK, rather than a coreceptor, is required for Agrin to stimulate MuSK. Activated MuSK clusters MuSK (Jones et al., 1999; Moore et al., 2001), and this positive-feedback mechanism may be important to achieve an adequate level of MuSK expression at the synapse, sufficient to cluster more than 10 million AChR molecules per synapse. Nonetheless, an unrestrained positive-feedback mechanism could lead to MuSK activation and postsynaptic differentiation beyond the synaptic site. Thus, it may be important to impose restraints on MuSK activation at synaptic sites. There may be instances when an absence of such restraints, leading to Agrin-independent MuSK activity, however, is exploited during development. For example, clustering of AChRs in the central region of muscle early in development, which is independent of Agrin but dependent upon MuSK (Lin et al., 2001; Yang et al., 2001), may be initiated by adventitious MuSK activation and facilitated by the absence of an inhibitory protein that otherwise restrains MuSK activity. Rapsyn, a 43-kDa peripheral membrane protein that is associated with AChRs at 1:1 stoichiometry, is a required intermediate on the pathway that couples MuSK activation to AChR clustering (Gautam et al., 1995) [Fig. 1(A)]. The tetratricopeptide repeats in Rapsyn mediate Rapsyn self-association, likely to be important for clustering Rapsyn at synapses, the coiled-coil domain in Rapsyn directs association between Rapsyn and the main intracellular loop of AChR subunits, and the RING-H2 domain in Rapsyn interacts with beta-dystroglycan (Ramarao and Cohen, 1998; Bartoli et al., 2001; Ramarao et al., 2001). The mechanisms by which MuSK activation leads to clustering of Rapsyn are poorly understood, but clustering of Rapsyn is essential for clustering AChRs. Although Rapsyn is central to building a synapse, certain aspects of synaptic differentiation do not require Rapsyn. In the absence of Rapsyn, MuSK remains concentrated at synaptic sites and AChR genes are selectively transcribed by synaptic nuclei, leading to an enrichment of AChR protein in the central region of the muscle (Gautam et al., 1995; Apel et al., 1997). The mechanisms that regulate the clustering of MuSK, independent of Rapsyn, at synaptic sites are not understood. The carboxy-terminus of MuSK can bind PDZ domain-containing proteins (Strochlic et al., 2001), but this binding site is dispensable for MuSK function in cultured muscle cells and in vivo (Zhou et al., 1999; Herbst et al., in press). It is possible that engagement by Agrin captures MuSK or that Agrin activation of MuSK stimulates MuSK clustering at synapses independent of Rapsyn and PDZ domain-containing proteins. Differing results have been reported for the role of the ectodomain of MuSK in the formation of AChR clusters. Like Agrin-induced AChR clusters, the formation of spontaneous AChR clusters in cultured myotubes depends upon Rapsyn and MuSK (Gautam et al., 1995; Zhou et al., 1999; Herbst and Burden, 2000). Because the formation of spontaneous AChR clusters is not dependent upon the ectodomain of MuSK (Zhou et al., 1999), sequences in the MuSK cytoplasmic domain are sufficient to mediate the Rapsyn-dependent formation of spontaneous AChR clusters. Consistent with this idea, MuSK mutants, lacking the ectodomain can stimulate the formation of AChR clusters at ectopic sites in adult muscle (Sander et al., 2001). These findings would suggest that the cytoplasmic domain of MuSK is sufficient to activate a pathway leading to AChR clustering. In contrast to this idea, NT3-stimulated tyrosine phosphorylation of a TrkC/MuSK chimera, expressed in cultured myotubes fails to stimulate AChR clustering (Glass et al., 1997). Moreover, Rapsyn-induced clustering of MuSK in QT6 cells requires the MuSK ectodomain, indicating that the ectodomain of MuSK interacts with Rapsyn indirectly, via a putative transmembrane protein termed RATL, to mediate Rapsyn/MuSK coclustering in QT6 cells (Apel et al., 1997). AChR Tyrosine Phosphorylation Agrin stimulates tyrosine phosphorylation of the AChR and subunits (Wallace et al., 1991; Qu and Huganir, 1994; Mittaud et al., 2001). The function of AChR tyrosine phosphorylation is not understood, but it is insufficient to cluster AChRs, as AChR tyrosine phosphorylation but not AChR clustering is stimulated in muscle cells that are transfected with a TrkC/ MuSK chimera and treated with NT3 (Glass et al., 1997). In myotubes expressing both wild-type and mutant AChR pentamers, containing an epitopetagged, mutant subunit (Y390F), mutant pentamers are clustered by Agrin, although less (twofold) efficiently than wild-type pentamers (Meyer and Wallace, 1998; Borges and Ferns, 2001). Although these results indicate that tyrosine phosphorylation of the AChR subunit contributes to AChR clustering induced by Agrin, it remains possible that AChR tyrosine phosphorylation is essential for AChR clustering. First, because the transfected myotubes express both mutant and wild-type AChR pentamers, wild-

Neuromuscular Synapse 505 type pentamers may form a scaffold that assists in the clustering of the mutant pentamers. Second, because the mutant pentamers retain a wild-type subunit, tyrosine phosphorylation of either the or subunit may be sufficient to mediate clustering of AChRs by Agrin. Tyrosine phosphorylation of AChRs could regulate AChR clustering by initiating a link between AChRs and the cytoskeleton (Wallace, 1992). Consistent with this idea, AChR pentamers containing a mutant subunit (Y390F) are extracted more readily by nonionic detergent from cultured myotubes treated with Agrin (Borges and Ferns, 2001). In addition, a phosphotyrosine (Y393)-containing sequence in the subunit serves as a docking site for Grb2 (Colledge and Froehner, 1997), raising the possibility that Grb2 serves as an adaptor to link tyrosine phosphorylated subunits to additional proteins, possibly a cytoskeletal complex. Staurosporine blocks Agrin-stimulated tyrosine phosphorylation of AChRs without inhibiting tyrosine phosphorylation of MuSK, indicating that tyrosine kinases other than MuSK catalyze tyrosine phosphorylation of AChRs (Ferns et al., 1996; Fuhrer et al., 1997). Src-family kinases were considered good candidates to catalyze AChR tyrosine phosphorylation (Swope and Huganir, 1993; Ferns et al., 1996; Fuhrer and Hall, 1996; Fuhrer et al., 1999), but Agrin stimulates tyrosine phosphorylation and clustering of AChRs in mutant muscle cells lacking either Src and Fyn or Src and Yes and in muscle cells treated with inhibitors of Src-family kinases (Smith et al., 2001; but see Mohamed et al., 2001). Thus, kinases other than Src-family kinases appear to catalyze AChR tyrosine phosphorylation and clustering. Src-like kinases, however, regulate the stability of AChR clusters (Smith et al., 2001). Following withdrawal of Agrin, AChR clusters disassemble more rapidly in myotubes lacking Src and Fyn than in wild-type myotubes. Nonethelesss, the role of Src and Fyn in stabilizing AChR clusters does not require their kinase activities, because inhibitors of Src-family kinase activity do not decrease the stability of AChR clusters. These results suggest that the adapter activities, rather than the kinase activities of Src/Fyn are essential to stabilize and anchor AChRs and that Src/Fyn recruit additional kinases that modulate interactions between the postsynaptic membrane and the cytoskeleton. These ideas are supported by two observations. First, upon removal of Agrin, herbimycin and staurosporine, unlike inhibitors of Src-family kinases, disperse pre-formed AChR clusters (Ferns et al., 1996). As MuSK phosphorylation is not inhibited by staurosporine (Fuhrer et al., 1997), these findings suggest that the activity of a kinase, other than MuSK or a Src-family member, is necessary to maintain AChR clusters. Second, the stability of Agrin-induced AChR clusters is also reduced in myotubes lacking -Dystrobrevin (Grady et al., 2000). Dystrobrevin is a substrate for tyrosine kinases and is strongly tyrosine phosphorylated in Torpedo electric organ (Wagner et al., 1993). Although Dystrobrevin is not tyrosine-phosphorylated by Agrin stimulation nor required for early steps in Agrin-induced AChR clustering (Nawrotzki et al., 1998), its phosphorylation may play a role in stabilization of AChR clusters. Taken together, these studies suggest that tyrosine phosphorylation, catalyzed by kinases other than Src family members, may have a role in anchoring AChRs to cytoskeletal components, including Dystrobrevin and its associated Utrophin glycoprotein complex. Certain Genes Are Expressed Selectively in Synaptic Nuclei of Myofibers The mrnas encoding the five AChR subunits are concentrated at synaptic sites (Burden, 1998; Sanes and Lichtman, 2001; Schaeffer et al., 2001). Studies with transgenic mice carrying gene fusions between regulatory regions of AChR subunit genes and reporter genes have demonstrated that AChR genes are transcribed selectively in myofiber nuclei near the synaptic site. Thus, localized transcription of AChR genes in synaptic nuclei is responsible, at least in part, for the accumulation of AChR mrna at synaptic sites. This pathway is important for ensuring that AChRs are expressed at the required density in the postsynaptic membrane, as defects in synapse-specific gene expression of the AChR subunit gene is the cause of a congenital myaesthenia (Nichols et al., 1999; Ohno et al., 1999). Like AChR subunit genes, the utrophin gene is transcribed selectively in synaptic nuclei, resulting in accumulation of utrophin mrna and protein at synaptic sites (Gramolini et al., 1999). mrnas encoding Rapsyn, N-CAM, MuSK, sodium channels, and the catalytic subunit of AChE are also concentrated in the synaptic region of skeletal myofibers, consistent with the idea that these genes are likewise transcribed preferentially in synaptic nuclei. Thus, synapse-specific transcription is likely to be a common and important mechanism for concentrating gene products at the neuromuscular synapse.

506 Burden Figure 2 (A) Neuregulin-1 is a favored candidate for a signal that regulates synapse-specific transcription. (B) Immunostaining of frozen sections from skeletal muscles shows that Nrg-1 and its receptors, ErbB3 and ErbB4, as well as ErbB2, are concentrated at synaptic sites. Neuregulin-1 Is a Candidate for the Signal That Activates Gene Expression in Synaptic Nuclei A signal for synapse-specific transcription is contained in the synaptic basal lamina. Nonetheless, because soluble forms of Agrin do not increase AChR expression in cultured muscle, Agrin has not been considered a favored candidate for the transcriptional signal in the synaptic basal lamina. Instead, the favored candidate is Nrg-1 (Falls et al., 1993; Carraway and Burden, 1995; Rosen et al., 1996; Fischbach and Rosen, 1997; Schaeffer et al., 2001) [Fig. 2(A)]. The nrg-1 gene encodes more than a dozen alternatively spliced products that have multiple activities. Although originally purified as a ligand that stimulates tyrosine phosphorylation of the neu oncogene, Nrg-1 was independently purified an activity in the CNS, termed AChR inducing activity (ARIA), that induces AChR synthesis in cultured myotubes, and as an activity in the pituitary, termed glial growth factor (GGF), that stimulates proliferation of Schwann cells. Nrg-1 is concentrated at neuromuscular synapses, and can activate AChR gene expression in muscle cells grown in cell culture (Chu et al., 1995; Jo et al., 1995) [Fig. 2(B)]. Nrg-1 contains a single EGF-like domain, which is necessary and sufficient for cell signaling. Motor neurons synthesize Nrg-1, and Nrg-1 protein is detectable in motor axons, indicating that some of the Nrg-1 protein at synaptic sites is synthesized by motor neurons (Sandrock et al., 1995). Skeletal myofibers, however, also synthesize Nrg-1 (Moscoso et al., 1995; Rimer et al., 1998), and some of the Nrg-1 at synaptic sites is synthesized by myofibers (Rimer et al., 1998). These findings raise the possibility that Nrg-1 could act as an autocrine and/or paracrine signal at neuromuscular synapses (Fig. 3). ErbB3 and ErbB4, members of the EGF receptor family, are receptors for Nrg-1, and each receptor is concentrated in the postsynaptic membrane at neuromuscular synapses (Altiok et al., 1995; Moscoso et al., 1995; Zhu et al., 1995) [Fig. 2(B)]. The coclustering of Nrg-1, ErbB3, and ErbB4 at synapses supports the idea that Nrg-1 is a signal that regulates synaptic differentiation. Nonetheless, because mice mutant for nrg-1, erbb2, or erbb4 die, owing to a failure of heart development, at E10.5, several days prior to neuromuscular synapse formation (Gassmann et al., 1995; Lee et al., 1995; Meyer and Birchmeier, 1995), these mutant mice are not suitable for analyzing whether Nrg-1 mediated signaling is required for synapse-specific transcription. Adult mice that are heterozygous for the Ig-isoform of Nrg-1 express fewer AChRs at neuromuscular synapses, supporting the idea that Nrg-1, supplied by motor neurons and/or muscle, has a role in regulating AChR expression at adult synapses (Sandrock et al., 1997). Neuronal Nrg-1, however, is not required for synapse-specific transcription at developing synapses, Figure 3 (A) Model for synapse-specific transcription stimulated either by motor neuron-derived or musclederived Nrg-1. (B) Forced expression of neural Agrin in ectopic regions of skeletal muscle causes the clustering of AChRs, ErbBs and muscle-derived Nrg-1 at the ectopic sites, leading to the ErbB2-dependent induction of AChR genes.

Neuromuscular Synapse 507 Figure 4 Mutually reinforcing, crossregulatory pathways for synaptic differentiation. Agrin/MuSK signaling (pink arrow) regulates Nrg-1/ErbB signaling by clustering (black arrow) muscle-derived Nrg-1 and ErbBs at synaptic sites, and Nrg-1/ErbB signaling (green and blue arrows) regulates Agrin/MuSK by stimulating MuSK expression (red arrow). as the pattern of AChR transcription is normal in newborn mice lacking Nrg-1 specifically in motor and sensory neurons (Yang et al., 2001). These experiments leave open the possibility that muscle-derived Nrg-1 may be required for synapse-specific transcription or that neuronal- and muscle-derived Nrg-1 have redundant roles in regulating synapse-specific transcription. Neural Agrin, or constitutively active MuSK, can cluster muscle-derived Nrg-1 and ErbBs, and stimulate transcription of AChR genes (Meier et al., 1997, 1998; Rimer et al., 1998) (Fig. 3). Ectoptic induction of AChR transcription is inhibited by a dominant negative form of ErbB2, demonstrating that Agrin/ MuSK stimulate AChR transcription in an ErbB-dependent manner, at least at ectopic sites (Meier et al., 1998; Moore et al., 2001). Moreover, a dominant negative form of GABP, a transcription factor implicated in Nrg-1 mediated induction of AChR transcription in cultured myotubes, likewise inhibits induction of AChR transcription at ectopic Agrin sites (Briguet and Ruegg, 2000). Taken together, these experiments suggest that neural Agrin regulates synapse-specific transcription by defining the limits of Nrg-1 and ErbB expression, thus restricting an autocrine Nrg-1 signaling pathway to synaptic sites in muscle [Figs. 1(B) and 3]. Moreover, because Nrg-1 can stimulate MuSK expression (Ip et al., 2000), Agrin/MuSK and Nrg-1/ ErbB signaling pathways mutually reinforce one-another (Fig. 4), thereby sharpening the boundaries of transcription and protein localization at synaptic sites. Nrg-1 can activate AChR gene expression in muscle cells grown in cell culture, and the Nrg-1 response element in the AChR subunit gene is a binding site for Ets-domain transcription factors (Burden, 1998; Schaeffer et al., 2001). In the AChR subunit gene, both an Ets-site as well as a binding site for Sp1 are thought to mediate Nrg-1 responsiveness (Alroy et al., 1999). Mutations in the Ets-binding site in the human AChR subunit gene result in reduced AChR expression and a myopathy, termed congenital myasthenic syndrome (Nichols et al., 1999; Ohno et al., 1999); these findings are currently the best evidence that synapse-specific transcription has a critical role at neuromuscular synapses. Nrg-1 stimulation of ErbBs leads to activation of MAP kinase signaling pathways, and both Erk and Jnk kinases are required for Nrg-1 to induce AChR genes in cultured myotubes (Si et al., 1998; Schaeffer et al., 2001). The mechanisms that couple MAP kinase signaling to activation of GABP are poorly understood. Muscle-Autonmous Patterning of AChR Expression Analysis of agrin mutant mice supported the idea that neural signals are required to pattern AChR gene Figure 5 Model for patterning AChR expression independent of innervation and for refinement of this pattering by innervation. In the absence of innervation, and possibly prior to innervation, weak MuSK activation in the central region of the muscle leads to clustering (black arrows) of AChRs, muscle-derived Nrg-1 and ErbBs in this central region (light blue zone). Muscle-derived Nrg-1 activates ErbBs (blue arrow), leading to an increase in AChR and MuSK transcription (red arrow). Motor innervation modifes the prepattern to the more refined pattern of AChR transcription and AChR clustering characteristic of mature synapses. This conversion appears to depend upon two distinguishable nerve-dependent programs: one program utilizes motor neuron-derived Agrin to maintain AChR expression at nascent synaptic sites and a second program, possibly triggered by electrical activity, extinguishes AChR expression throughout the muscle. Together, the two programs ensure the stable expression of AChR clusters selectively at nascent synapses (blue zone).

508 Burden expression and AChR clusters in skeletal muscle cells (Gautam et al., 1996). Small AChR clusters are present in muscle from agrin mutant newborn mice, but they were found to be scattered throughout the muscle (Gautam et al., 1996). These findings indicated that Agrin is required to restrict AChR gene expression and AChR clusters to the central region of the muscle, consistent with the notion that neuronal signals are required to pattern muscle. Studies of mice lacking Topoisomerase II (Top II ), however, led to a reconsideration of the possibility that muscle may be patterned independent of motor innervation (Yang et al., 2000). Although motor axons in top 2 mutant embryos fail to grow into or branch within diaphragm and limb muscles, AChRs are clustered in the central region of muscle, in a zone that is only twofold wider than in normal mice (Yang et al., 2000). These findings indicated that the pattern of AChR clusters in skeletal muscle might be determined, at least in part, by mechanisms that are independent of motor innervation. Subsequent analysis of additional mouse mutants that affect motor neuron generation, motor axon projections, and motor neuron-derived signals confirmed that AChR gene transcription and AChR clustering are patterned in developing skeletal muscle in the absence of motor neurons and motor axons (Lin et al., 2001; Yang et al., 2001). Moreover, these studies indicated that neural signals have a role in maintaining and refining, rather than initiating this pattern of AChR expression (Yang et al., 2000, 2001; Lin et al., 2001). These findings suggest that an initial, spatially restricted pattern of AChR expression is generated in muscle-independent of neurally derived Agrin (Fig. 5). Arrival of the nerve, and its attendant signals, modifies the AChR prepattern into the more refined pattern of AChR transcription and AChR clustering characteristic of mature synapses. This alteration appears to depend upon two distinguishable nerve-dependent programs: one program utilizes neurally derived Agrin to maintain AChR expression at nascent synaptic sites and a second program, possibly triggered by electrical activity (Goldman et al., 1988), extinguishes AChR expression throughout the muscle. Together, the two programs ensure the stable expression of AChR clusters selectively at nascent synapses (Fig. 5). Prior studies of developing mammalian muscle provided hints of such nerve-independent, musclepatterning mechanisms (Braithwaite and Harris, 1979; Harris, 1981). In the earlier studies, however, it remained unclear whether developing muscle was indeed deprived of innervation or whether some motor axons had established transient contacts with muscle prior to their ablation. The lack of a well-defined synaptic zone in chick muscles may explain the discrepant reports on AChR expression in developing chick muscles deprived of innervation (Sohal, 1988; Dahm and Landmesser, 1991). Muscle patterning does not appear to be required for synapse formation, as motor neurons can form synapses with muscle cells in cell culture, which do not display an AChR prepattern (Anderson and Cohen, 1977; Frank and Fischbach, 1979). Likewise, regenerating motor axons can form synapses at nonsynaptic regions of denervated adult muscle (Frank et al., 1974). Nevertheless, muscle patterning may provide a preferred region for motor innervation during development in vivo. Physiologic reasons may dictate why the central region of a muscle is commonly the preferred site for innervation, as the contraction mechanism is optimally synchronized by initiating an action potential from a synapse near the middle of a muscle fiber. Although the muscle AChR prepattern does not require Agrin, this prepattern is not observed in mice lacking MuSK (Lin et al., 2001; Yang et al., 2001). These results indicate that MuSK can be activated independent of Agrin, and suggest that the mechanisms that control MuSK activation have an important role in establishing the AChR prepattern. Because activated MuSK recruits and clusters MuSK (Jones et al., 1999), an early and modest bias in MuSK activity in the central region of the muscle may be converted, via this positive feedback loop, into a significant increase in MuSK activity in the central region of the muscle, leading to patterned expression of additional muscle genes. Although such mechanisms could explain how MuSK regulates muscle patterning, this model does not address how MuSK might initially be expressed preferentially in the prospective synaptic region of developing muscle. Developing myotubes grow, in large part, by the fusion of myoblasts symmetrically to either side of the neuromuscular synapse. Because a large number of genes expressed in skeletal muscle, including MuSK, are activated concomitant with fusion, gene activation will proceed temporally from the central to the peripheral regions of the muscle. Because activated MuSK clusters and recruits MuSK, an early bias in MuSK expression would function to stabilize and enhance this disposition. Thus, a low level of ligandindependent MuSK activation, initiated when MuSK is first expressed, may be sufficient to establish a muscle pre-pattern in the absence of innervation.

Neuromuscular Synapse 509 REFERENCES Alroy I, Soussan L, Seger R, Yarden Y. 1999. Neu differentiation factor stimulates phosphorylation and activation of the Sp1 transcription factor. Mol Cell Biol 19:1961 1972. Altiok N, Bessereau JL, Changeux JP. 1995. ErbB3 and ErbB2/neu mediate the effect of heregulin on acetylcholine receptor gene expression in muscle: differential expression at the endplate. EMBO J 14:4258 4266. Anderson MJ, Cohen MW. 1977. Nerve-induced and spontaneous redistribution of acetylcholine receptors on cultured muscle cells. J Physiol (Lond) 268:757 773. Apel ED, Glass DJ, Moscoso LM, Yancopoulos GD, Sanes JR. 1997. Rapsyn is required for MuSK signaling and recruits synaptic components to a MuSK-containing scaffold. Neuron 18:623 635. Bartoli M, Ramarao MK, Cohen JB. 2001. Interactions of the rapsyn RING-H2 domain with dystroglycan. J Biol Chem 276:24911 24917. Borges LS, Ferns M. 2001. Agrin-induced phosphorylation of the acetylcholine receptor regulates cytoskeletal anchoring and clustering. J Cell Biol 153:1 12. Braithwaite AW, Harris AJ. 1979. Neural influence on acetylcholine receptor clusters in embryonic development of skeletal muscles. Nature 279:549 551. Briguet A, Ruegg MA. 2000. The Ets transcription factor GABP is required for postsynaptic differentiation in vivo. J Neurosci 20:5989 5996. Burden SJ. 1998. The formation of neuromuscular synapses. Genes Dev 12:133 148. Carraway KL III, Burden SJ. 1995. Neuregulins and their receptors. Curr Opin Neurobiol 5:606 612. Chu GC, Moscoso LM, Sliwkowski MX, Merlie JP. 1995. Regulation of the acetylcholine receptor epsilon subunit gene by recombinant ARIA: an in vitro model for transynaptic gene regulation. Neuron 14:329 339. Colledge M, Froehner SC. 1997. Tyrosine phosphorylation of nicotinic acetylcholine receptor mediates Grb2 binding. J Neurosci 17:5038 5045. Dahm LM, Landmesser LT. 1991. The regulation of synaptogenesis during normal development and following activity blockade. J Neurosci 11:238 255. DeChiara TM, Bowen DC, Valenzuela DM, Simmons MV, Poueymirou WT, Thomas S, Kinetz E, Compton DL, Rojas E, Park JS, Smith C, DiStefano PS, Glass DJ, Burden SJ, Yancopoulos GD. 1996. The receptor tyrosine kinase MuSK is required for neuromuscular junction formation in vivo. Cell 85:501 512. Falls DL, Rosen KM, Corfas G, Lane WS, Fischbach GD. 1993. ARIA, a protein that stimulates acetylcholine receptor synthesis, is a member of the neu ligand family. Cell 72:801 815. Ferns M, Deiner M, Hall Z. 1996. Agrin-induced acetylcholine receptor clustering in mammalian muscle requires tyrosine phosphorylation. J Cell Biol 132:937 944. Fischbach GD, Rosen KM. 1997. ARIA: a neuromuscular junction neuregulin. Annu Rev Neurosci 20:429 458. Frank E, Fischbach GD. 1979. Early events in neuromuscular junction formation in vitro: induction of acetylcholine receptor clusters in the postsynaptic membrane and morphology of newly formed synapses. J Cell Biol 83: 143 158. Frank E, Jansen JKS, Lømo T, Westgaard R. 1974. Maintained function of foreign synapses on hyperinnervated skeletal muscle fibres of the rat. Nature 247:375 376. Fu AK, Smith FD, Zhou H, Chu AH, Tsim KW, Peng BH, Ip NY. 1999. Xenopus muscle-specific kinase: molecular cloning and prominent expression in neural tissues during early embryonic development. Eur J Neurosci 11:373 382. Fuhrer C, Hall ZW. 1996. Functional interaction of Src family kinases with the acetylcholine receptor in C2 myotubes. J Biol Chem 271:32474 32481. Fuhrer C, Gautam M, Sugiyama JE, Hall ZW. 1999. Roles of rapsyn and agrin in interaction of postsynaptic proteins with acetylcholine receptors. J Neurosci 19:6405 6416. Fuhrer C, Sugiyama JE, Taylor RG, Hall ZW. 1997. Association of muscle-specific kinase MuSK with the acetylcholine receptor in mammalian muscle. EMBO J 16: 4951 4960. Gassmann M, Casagranda F, Orioli D, Simon H, Lai C, Klein R, Lemke G. 1995. Aberrant neural and cardiac development in mice lacking the ErbB4 neuregulin receptor [see comments]. Nature 378:390 394. Gautam M, Noakes PG, Moscoso L, Rupp F, Scheller RH, Merlie JP, Sanes JR. 1996. Defective neuromuscular synaptogenesis in agrin-deficient mutant mice. Cell 85:525 535. Gautam M, Noakes PG, Mudd J, Nichol M, Chu GC, Sanes JR, Merlie JP. 1995. Failure of postsynaptic specialization to develop at neuromuscular junctions of rapsyndeficient mice [see comments]. Nature 377:232 236. Glass DJ, Yancopoulos GD. 1997. Sequential roles of agrin, MuSK and rapsyn during neuromuscular junction formation. Curr Opin Neurobiol 7:379 384. Glass DJ, Apel ED, Shah S, Bowen DC, DeChiara TM, Stitt TN, Sanes JR, Yancopoulos GD. 1997. Kinase domain of the muscle-specific receptor tyrosine kinase (MuSK) is sufficient for phosphorylation but not clustering of acetylcholine receptors: required role for the MuSK ectodomain? Proc Natl Acad Sci USA 94:8848 8853. Glass DJ, Bowen DC, Stitt TN, Radziejewski C, Bruno J, Ryan TE, Gies DR, Shah S, Mattsson K, Burden SJ, DiStefano PS, Valenzuela DM, DeChiara TM, Yancopoulos GD. 1996. Agrin acts via a MuSK receptor complex. Cell 85:513 523. Goldman D, Brenner HR, Heinemann S. 1988. Acetylcholine receptor -, -, -, and -subunit mrna levels are regulated by muscle activity. Neuron 1:329 333. Grady RM, Zhou H, Cunningham JM, Henry MD, Campbell KP, Sanes JR. 2000. Maturation and maintenance of the neuromuscular synapse: genetic evidence for roles of the dystrophin glycoprotein complex. Neuron 25:279 293. Gramolini AO, Angus LM, Schaeffer L, Burton EA, Tinsley

510 Burden JM, Davies KE, Changeux JP, Jasmin BJ. 1999. Induction of utrophin gene expression by heregulin in skeletal muscle cells: role of the N-box motif and GA binding protein. Proc Natl Acad Sci USA 96:3223 3227. Harris AJ. 1981. Embryonic growth and innervation of rat skeletal muscles. III. Neural regulation of junctional and extra-junctional acetylcholine receptor clusters. Philos Trans R Soc Lond B Biol Sci 293:287 314. Herbst R, Burden SJ. 2000. The juxtamembrane region of MuSK has a critical role in agrin-mediated signaling [published erratum appears in EMBO J 2000 Mar 1;19(5):1167]. EMBO J 19:67 77. Herbst R, Avetisova E, Burden SJ. 2002. Restoration of synapse formation in Musk mutant mice expressing a Musk/Trk chimeric receptor. Development, in press. Ip FC, Glass DG, Gies DR, Cheung J, Lai KO, Fu AK, Yancopoulos GD, Ip NY. 2000. Cloning and characterization of muscle-specific kinase in chicken. Mol Cell Neurosci 16:661 673. Jennings CG, Dyer SM, Burden SJ. 1993. Muscle-specific trk-related receptor with a kringle domain defines a distinct class of receptor tyrosine kinases. Proc Natl Acad Sci USA 90:2895 2899. Jo SA, Zhu X, Marchionni MA, Burden SJ. 1995. Neuregulins are concentrated at nerve-muscle synapses and activate ACh- receptor gene expression. Nature 373:158 161. Jones G, Moore C, Hashemolhosseini S, Brenner HR. 1999. Constitutively active MuSK is clustered in the absence of agrin and induces ectopic postsynaptic-like membranes in skeletal muscle fibers. J Neurosci 19:3376 3383. Lee KF, Simon H, Chen H, Bates B, Hung MC, Hauser C. 1995. Requirement for neuregulin receptor erbb2 in neural and cardiac development [see comments]. Nature 378: 394 398. Lin W, Burgess RW, Dominguez B, Pfaff SL, Sanes JR, Lee KF. 2001. Distinct roles of nerve and muscle in postsynaptic differentiation of the neuromuscular synapse. Nature 410:1057 1064. McMahan UJ. 1990. The agrin hypothesis. Cold Spring Harb Symp Quant Biol 55:407 418. Meier T, Hauser DM, Chiquet M, Landmann L, Ruegg MA, Brenner HR. 1997. Neural agrin induces ectopic postsynaptic specializations in innervated muscle fibers. J Neurosci 17:6534 6544. Meier T, Masciulli F, Moore C, Schoumacher F, Eppenberger U, Denzer AJ, Jones G, Brenner HR. 1998. Agrin can mediate acetylcholine receptor gene expression in muscle by aggregation of muscle-derived neuregulins. J Cell Biol 141:715 726. Meyer D, Birchmeier C. 1995. Multiple essential functions of neuregulin in development [see comments] [published erratum appears in Nature 1995 Dec 14;378(6558):753]. Nature 378:386 390. Meyer G, Wallace BG. 1998. Recruitment of a nicotinic acetylcholine receptor mutant lacking cytoplasmic tyrosine residues in its beta subunit into agrin-induced aggregates. Mol Cell Neurosci 11:324 333. Mittaud P, Marangi PA, Erb-Vogtli S, Fuhrer C. 2001. Agrininduced activation of acetylcholine receptor-bound Src family kinases requires Rapsyn and correlates with acetylcholine receptor clustering. J Biol Chem 276:14505 14513. Mohamed AS, Rivas-Plata KA, Kraas JR, Saleh SM, Swope SL. 2001. Src-class kinases act within the agrin/musk pathway to regulate acetylcholine receptor phosphorylation, cytoskeletal anchoring, and clustering. J Neurosci 21:3806 3818. Moore C, Leu M, Muller U, Brenner HR. 2001. Induction of multiple signaling loops by MuSK during neuromuscular synapse formation. Proc Natl Acad Sci USA 98:14655 14660. Moscoso LM, Chu GC, Gautam M, Noakes PG, Merlie JP, Sanes JR. 1995. Synapse-associated expression of an acetylcholine receptor-inducing protein, ARIA/heregulin, and its putative receptors, ErbB2 and ErbB3, in developing mammalian muscle. Dev Biol 172:158 169. Nawrotzki R, Loh NY, Ruegg MA, Davies KE, Blake DJ. 1998. Characterisation of alpha-dystrobrevin in muscle. J Cell Sci 111:2595 2605. Nichols P, Croxen R, Vincent A, Rutter R, Hutchinson M, Newsom-Davis J, Beeson D. 1999. Mutation of the acetylcholine receptor epsilon-subunit promoter in congenital myasthenic syndrome. Ann Neurol 45:439 443. Ohno K, Anlar B, Engel AG. 1999. Congenital myasthenic syndrome caused by a mutation in the Ets-binding site of the promoter region of the acetylcholine receptor epsilon subunit gene [see comments]. Neuromuscul Disord 9:131 135. Qu Z, Huganir RL. 1994. Comparison of innervation and agrin-induced tyrosine phosphorylation of the nicotinic acetylcholine receptor. J Neurosci 14:6834 6841. Ramarao MK, Cohen JB. 1998. Mechanism of nicotinic acetylcholine receptor cluster formation by rapsyn. Proc Natl Acad Sci USA 95:4007 4012. Ramarao MK, Bianchetta MJ, Lanken J, Cohen JB. 2001. Role of rapsyn tetratricopeptide repeat and coiled-coil domains in self-association and nicotinic acetylcholine receptor clustering. J Biol Chem 276:7475 7483. Rimer M, Cohen I, Lomo T, Burden SJ, McMahan UJ. 1998. Neuregulins and erbb receptors at neuromuscular junctions and at agrin- induced postsynaptic-like apparatus in skeletal muscle. Mol Cell Neurosci 12:1 15. Rosen KM, Sandrock AW Jr, Goodearl AD, Loeb JA, Fischbach GD. 1996. The role of neuregulin (ARIA) at the neuromuscular junction. Cold Spring Harb Symp Quant Biol 61:427 434. Ruegg MA, Bixby JL. 1998. Agrin orchestrates synaptic differentiation at the vertebrate neuromuscular junction. Trends Neurosci 21:22 27. Sander A, Hesser BA, Witzemann V. 2001. MuSK induces in vivo acetylcholine receptor clusters in a ligand-independent manner. J Cell Biol 155:1287 1296. Sandrock AW, Jr., Dryer SE, Rosen KM, Gozani SN, Kramer R, Theill LE, Fischbach GD. 1997. Maintenance of acetylcholine receptor number by neuregulins at the neuromuscular junction in vivo. Science 276:599 603. Sandrock AW, Jr., Goodearl AD, Yin QW, Chang D,

Neuromuscular Synapse 511 Fischbach GD. 1995. ARIA is concentrated in nerve terminals at neuromuscular junctions and at other synapses. J Neurosci 15:6124 6136. Sanes JR, Lichtman JW. 2001. Induction, assembly, maturation and maintenance of a postsynaptic apparatus. Nat Rev Neurosci 2:791 805. Schaeffer L, de Kerchove d Exaerde A, Changeux JP. 2001. Targeting transcription to the neuromuscular synapse. Neuron 31:15 22. Si J, Tanowitz M, Won S, Mei L. 1998. Regulation by ARIA/neuregulin of acetylcholine receptor gene transcription at the neuromuscular junction. Life Sci 62: 1497 1502. Smith CL, Mittaud P, Prescott ED, Fuhrer C, Burden SJ. 2001. Src, Fyn, and Yes are not required for neuromuscular synapse formation but are necessary for stabilization of agrin-induced clusters of acetylcholine receptors. J Neurosci 21:3151 3160. Sohal GS. 1988. Development of postsynaptic-like specializations of the neuromuscular synapse in the absence of motor nerve. Int J Dev Neurosci 6:553 565. Son YJ, Thompson WJ. 1995. Nerve sprouting in muscle is induced and guided by processes extended by Schwann cells. Neuron 14:133 141. Strochlic L, Cartaud A, Labas V, Hoch W, Rossier J, Cartaud J. 2001. MAGI-1c: a synaptic MAGUK interacting with musk at the vertebrate neuromuscular junction. J Cell Biol 153:1127 1132. Swope SL, Huganir RL. 1993. Molecular cloning of two abundant protein tyrosine kinases in Torpedo electric organ that associate with the acetylcholine receptor. J Biol Chem 268:25152 25161. Valenzuela DM, Stitt TN, DiStefano PS, Rojas E, Mattsson K, Compton DL, Nunez L, Park JS, Stark JL, Gies DR, et al. 1995. Receptor tyrosine kinase specific for the skeletal muscle lineage: expression in embryonic muscle, at the neuromuscular junction, and after injury. Neuron 15:573 584. Wagner KR, Cohen JB, Huganir RL. 1993. The 87K postsynaptic membrane protein from Torpedo is a protein-tyrosine kinase substrate homologous to dystrophin. Neuron 10:511 522. Wallace BG. 1988. Regulation of agrin-induced acetylcholine receptor aggregation by Ca and phorbol ester. J Cell Biol 107:267 278. Wallace BG. 1989. Agrin-induced specializations contain cytoplasmic, membrane, and extracellular matrix-associated components of the postsynaptic apparatus. J Neurosci 9:1294 1302. Wallace BG. 1992. Mechanism of agrin-induced acetylcholine receptor aggregation. J Neurobiol 23:592 604. Wallace BG. 1994. Staurosporine inhibits agrin-induced acetylcholine receptor phosphorylation and aggregation. J Cell Biol 125:661 668. Wallace BG, Qu Z, Huganir RL. 1991. Agrin induces phosphorylation of the nicotinic acetylcholine receptor. Neuron 6:869 878. Weston C, Yee B, Hod E, Prives J. 2000. Agrin-induced acetylcholine receptor clustering is mediated by the small guanosine triphosphatases Rac and Cdc42. J Cell Biol 150:205 212. Yang X, Arber S, William C, Li L, Tanabe Y, Jessell TM, Birchmeier C, Burden SJ. 2001. Patterning of muscle acetylcholine receptor gene expression in the absence of motor innervation. Neuron 30:399-410. Yang X, Li W, Prescott ED, Burden SJ, Wang JC. 2000. DNA topoisomerase IIbeta and neural development. Science 287:131 134. Zhou H, Glass DJ, Yancopoulos GD, Sanes JR. 1999. Distinct domains of MuSK mediate its abilities to induce and to associate with postsynaptic specializations. J Cell Biol 146:1133 1146. Zhu X, Lai C, Thomas S, Burden SJ. 1995. Neuregulin receptors, erbb3 and erbb4, are localized at neuromuscular synapses. EMBO J 14:5842 5848.