The Drosophila homolog of vertebrate Glutamate Receptor-Interacting Proteins (DGrip; FlyBase name -- Grip) acts as a key component of proper muscle guidance. Mutations in Grip impair patterning of ventral longitudinal muscles (VLMs), whereas lateral transverse muscles (LTMs) that attach to intrasegmental attachment sites develop normally. Myoblast fusion, stabilization of muscle contacts, and general muscle function are not impaired in the absence of Grip. Instead, the proper formation of cellular extensions during guidance fails in Grip mutant VLMs. Grip protein concentrates at the ends of VLMs while these muscles guide toward segment border attachment sites. Conversely, LTMs overexpressing Grip form ectopic cellular extensions that can cause attachment of these muscles to other muscles at segment borders. The data suggest that Grip participates in the reception of an attractive signal that emanates from the epidermal attachment sites to direct the motility of developing muscles. This Grip phenotype should provide a valuable means to study mechanistic principles of Grip function (Swan, 2004).

The establishment of specialized cell-cell junctions plays a determining role in the formation of mature patterned organs in all multicellular organisms. The most prominent examples are synaptic connections, which are formed either between neurons or between neurons and other cells, for example, muscles. Cells form extensions such as growth cones, lamellipodia, or filopodia, which they use to sense specific guidance cues and to finally anchor at the relevant target cells. During Drosophila embryogenesis, developing muscles grow growth-cone-like projections to navigate toward specific epidermal attachment sites. Drosophila muscles are grouped into two categories. One muscle type, which includes the lateral transverse muscles (LTMs), is characterized by single muscle fibers attaching to a single epidermal tendon cell. The other muscle type, indirectly attaching muscles, including the ventral longitudinal muscles (VLMs), converges with several muscle fibers on single tendon cells, recruiting extracellular matrix, to which they adhere. Recent evidence has shown that these tendon cells, also called apodemes, are the source of secreted Slit protein. Slit is sensed as a positive guidance cue by Robo receptors expressed in the nascent VLMs. Furthermore, experimentally induced overexpression of Robo receptors causes LTMs to extend toward Slit expressing tendon sites. In addition to the Robo/Slit-system controlling VLM guidance, the Derailed receptor tyrosine kinase controls LTM guidance. Interestingly, both systems also have firmly established roles in axonal guidance processes, suggesting a common mechanistic basis for cellular motility of muscles and neurons (Swan, 2004 and references therein)

Guidance processes are controlled by a diverse array of signaling proteins, with spatiotemporal activity that is subject to subtle regulation. How the cellular metabolism of such supramolecular signaling complexes is organized is the subject of intense investigation. Proteins containing PDZ domains, a protein-protein interaction domain of ~90 amino acids, recruit components of a signaling network into larger molecular complexes in order to allow rapid and specific intracellular signaling. GRIP family proteins (GRIP1 and ABP/GRIP2) contain six or seven PDZ domains in tandem. They were first identified via an interaction of their fifth PDZ domain with the C-terminal sequence (ESVKI) of the GluR2 AMPA receptor subunit (Dong, 1997; Srivastava, 1998; Wyszynski, 1999) and are suggested to participate in the synaptic localization of AMPA receptors. Interfering with the interaction between GRIPs and GluR2/3 prevents AMPA receptor recruitment to the synapse in vitro (Dong, 1997; Osten, 2000; Xia, 2000). GRIP has also been identified as a binding partner of both ephrin receptors and ligands (Torres, 1998; Brückner, 1999; Lin, 1999; Contractor, 2002), ARF-GAP GIT1 (Ko, 2003), the kinesin motor protein KIF1A, and liprin-alpha (Ko, 2003; Wyszynski, 2002). Despite this information, the cell biological basis of GRIP function is only poorly understood. Biochemically, GRIP1 is slightly enriched in synaptic preparations but also is strongly expressed in intracellular compartments, including putative transport vesicles for glutamate receptors (Wyszynski, 1998, 2002; Dong, 1999a). GRIPs have been suggested to mediate (1) the transport of glutamate receptors directly (Dong, 1997; Wyszynski, 2002), (2) the stabilization of receptors within postsynaptic densities (Osten, 2000), and/or (3) the stabilization of intracellular stores and/or participation in sorting decisions for the destruction or recycling of internalized receptors (Shi, 2001; Hirbec, 2003). Genetic analysis in mice has shown that GRIP1 function is already required early during development (Bladt, 2002); a GRIP1 knockout was embryonic lethal at day 12 and the embryos suffered from defects in junction formation between dermis and epidermis (Swan, 2004).

Evidence is presented that CG14447, the single GRIP homolog in Drosophila and therefore named Grip, participates in muscle development during embryogenesis. Loss of Grip function causes severe defects in VLM but not LTM patterning. Grip is required for the guidance of developing VLMs toward the apodemes. Other processes such as myoblast fusion, stabilization of muscle attachments, and muscle function per se are not affected in Grip mutant embryos and larvae. Mesodermal expression of Grip using transgenes rescues the Grip mutant phenotype. Consistent with its specific function in VLM guidance, Grip protein progressively concentrates at the ends of these muscles as they establish contact to their target position. Furthermore, when Grip was overexpressed within embryonic mesoderm, LTMs were guided toward ectopic attachment sites at segment borders. The Grip protein therefore appears to be used by a subset of muscles to direct their motility, likely by transporting and/or localizing signaling components of a novel pathway (Swan, 2004).

Elimination of GRIP1 in mice results in embryonic lethality (Bladt, 2002) associated with defective dermoepidermal junctions. These results were interpreted to indicate that the architecture of this contact requires PDZ domain interactions mediated through GRIP1, in order to maintain proper cell adhesion. The contact between epidermis and specific muscles is not properly formed in Drosophila embryos mutant for Grip. This observation on first sight might hint toward a defect in the stabilization of cell adhesion in Drosophila Grip mutants as well. However, escaping adult Drosophila from Grip null alleles showed no signs of adhesion loss. Moreover, defects in Grip mutants are limited to one muscle group in a way that argues against cell adhesion defects and favors a role of Grip in muscle guidance, in which the pathways identified to date are found to act in a muscle subgroup-specific manner. Consistently, the mechanical attachment of muscles in Grip mutants is unaffected, because the mutant VLMs form integrin-expressing attachment sites. Furthermore, the attachment is stable upon contraction, since the Grip mutant larvae are able to locomote robustly (Swan, 2004).

The findings exclude the possibility that the phenotypes of Grip mutants are due to an effect on cell adhesion properties in the process of stabilizing the muscle attachment sites versus upcoming muscular contraction force. Instead, direct evidence is provided that the motility of VLMs is specifically affected in the absence of Grip, by visualizing the morphological development of VLMs during guidance. Wild-type VLMs form growth-cone-like extensions invariantly projecting in the anterior direction. However, in Grip mutant muscles, the direction of cellular extensions appears randomized from the beginning, and often extensions appear collapsed. Furthermore, upon overexpression of Grip, ectopic cellular extensions form specifically from LTMs, which normally do not express the protein. These aberrant extensions frequently contacted and anchored at the segment borders, where obviously they became stabilized, since they still are detected in late larval muscles (Swan, 2004).

The data imply that Grip mediates a motility response within developing muscles toward an attractive signal expressed at the segment border. It has been reported that Robo receptors are required to extend toward Slit-expressing muscle attachment sites at segment borders. Loss of Robo-Slit function eliminates segment border attachment in VLMs, whereas overexpression of Robos leads to segment border attachment in LTMs. Moreover, Robo receptors are expressed at the edges of developing muscles in a spatiotemporal pattern very similar to the expression profile of Grip. Because of these obvious parallels between Robo/Slit and Grip, a potential interaction of these factors was extensively addressed by genetic and biochemical means. No evidence for a functional or physical interaction could be obtained. It therefore appears most likely that Grip organizes the response to a novel signal working in parallel to the Robo/Slit-system. The finding that Grip overexpression provokes changes in LTMs, whereas Robos are reported to be absent from these muscles, also argues in this direction. In principle, Grip could be involved in the execution of a signaling event, or alternatively, it might be important for the stabilization of first interactions pioneered, for example, by Robo/Slit signaling. Because Grip mutant muscles show defective extensions early during muscle guidance and, secondly, overexpression of Grip directly causes the formation of cellular extensions, the first alternative is favored (Swan, 2004).

VLM-type muscles by far show the strongest defects within Grip mutants, affecting ~100% of VLMs 6 and 7. However, other indirectly attaching muscles did show defects as well. Although the defects were weaker in these cells than in VLMs, they clearly were significant in comparison to control animals. Consistently, although Grip expression seems strongest at VLM attachment sites, the contacts of more dorsal muscles, which also attach indirectly, also express the protein. A similar situation, characterized by VLMs being most affected and expressing the most Grip between the indirectly attaching muscles, is reported for the Robo/Slit muscle guidance pathway. It might be that spatiotemporal specificities in the development of the VLMs make this particular muscle group especially dependent on robust guidance signaling between the indirectly attaching muscles. Drosophila muscle guidance has not so far been subject to saturating genetic analysis and besides few seminal studies, understanding of the process is still rather poor. In several other models of cellular motility, for example, growth cone migration, distinct pathways partially working in parallel have also been identified (Swan, 2004).

Even in the complete absence of myoblast fusion, muscle founder cells still form properly attached mini-muscles. Hereby, the initial polarization of these specific muscle precursors seemingly does not depend on tendon cells. However, the tendon cells provide essential guidance cues that direct muscle extension. It is essentially unknown, how cellular polarity is organized throughout the time course of guidance and subsequent muscle attachment. Most likely the polarized transport of relevant proteins toward the 'active muscle ends' is important already early within muscle guidance. In fact, developing muscles display a polarized microtubule network with the + ends facing the attachment sites (Swan, 2004).

Grip appears concentrated at ends of muscle cells before any proper attachment between the muscle and its prospective attachment site is established. As an intracellular adaptor molecule, Grip might organize signaling processes, for example, by clustering transmembrane receptors, or it might act downstream of the actual signaling processes, for example, by executing transporting events that are essential for directed muscle cell motility. In fact, the correct targeting of the EGF receptor ligand Vein to the site of muscle tendon attachment has been shown to be an essential step in organizing proper muscle pattern. The data suggest that after supplying Grip to muscles that normally do not express the protein, they start to sense a distant guidance cue, which in turn causes the formation of cellular extensions. Grip thus might switch on dormant receptors in muscles, for example, by mediating their transport to relevant cellular locations. Interestingly, Grip has been suggested to control the transport of transmembrane receptors and signaling molecules, such as glutamate receptors and ephrins, from intracellular compartments to the cell surface (Torres, 1998; Wyszynski, 1998; Brückner, 1999; Dong, 1999b; Braithwaite, 2002; Hirbec, 2003). In Drosophila muscles Grip is found localizing to discrete punctae similar to punctae formed by Grip in culture cells. Colocalization experiments in culture cells showed that Grip punctae often colocalized with endosomal markers, whereas no colocalizations with ER, Golgi, plasma membrane, lysosomal, or mitochondrial markers were observed. Thus the hypothesis is favored that Grip mediates signaling throughout muscle motility by regulating the endosomal trafficking of receptor complexes. Specialized proteins regulating signaling by endosomal trafficking have recently emerged as key players in animal development. Regulation of membrane protein composition by GRIPs might be subtle; different receptor populations such as AMPA/Kainate receptors have been suggested to be regulated by GRIP in opposing manners (Hirbec, 2003). Palmitoylation close to the N-terminal end has been described for the Grip family members GRIP1b and pABP-L, and is suggested to control their intracellular distribution (DeSouza, 2002; Yamazaki, 2001). Indeed, the absolute N terminus of Grip contains a conserved cysteine residue at position 13 and is similar to the N-terminal sequences demonstrated to mediate palmitoylation of GRIP1b and pABP-L. Initial experimental data suggest post-translational modification of Grip with palmitate (Swan, 2004).

The highly penetrant embryonic phenotype of Grip presented in this study should thus be especially well suited to further study mechanisms of GRIP function in the genetically well-tractable Drosophila model (Swan, 2004).

GENE STRUCTURE

cDNA clone length - 3780 bp

Bases in 5' UTR - 286

Bases in 3' UTR - 317

PROTEIN STRUCTURE

Amino Acids - 1058

Structural Domains

The Drosophila genome encodes a single GRIP homolog (CG14447). CG14447 is represented by several embryonic cDNA isolates, which all predict the same protein sequence. Comparison of this sequence with mouse GRIP1 in respect to both position and sequence of PDZ domains clearly identifies CG14447 as a GRIP family member. Comparison between mouse GRIP1 and Drosophila protein CG14447 (Grip) reveals that both encode seven individually conserved PDZ domains (Swan, 2004).

date revised: 30 May 2004

Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.