Sra-1 interacts with Kette and Wasp and is required for neuronal and bristle development in Drosophila

The cytoskeleton plays a central role in cell morphology and motility. In particular the F-actin cytoskeleton of most eukaryotic cells is dynamically reorganized, providing force for shape changes and movements(Borisy and Svitkina, 2000; Pantaloni et al., 2001). This is particularly evident during nervous system development, when growth cones must advance and synaptic plasticity has to be regulated. Rearrangement of actin is evoked rapidly by extracellular stimuli, and sets of actin-associated proteins are thought to act cooperatively in the polymerization, crosslinking and anchorage of actin filaments (Pollard and Borisy, 2003). Actin dynamics is controlled by a small set of proteins. The Arp2/3 complex has been shown to nucleate de novo actin polymerization (Machesky et al.,1994). Arp2/3 is normally repressed, but can be activated by the members of the Wiskott family Wasp and Wave.

The Wasp proteins are auto-inhibited, whereas the Wave proteins are trans-inhibited. Both usually require small G proteins of the Rho family for activation (Miki and Takenawa,2003). In the case of Wasp, activated, GTP-bound Cdc42 binds to the CRIB (Cdc42/Rac Interactive Binding) domain of Wasp, releasing the auto-inhibition and thereby leading to the activation of the Arp2/3 complex(Higgs and Pollard, 2001; Rohatgi et al., 1999). However, a structure-function analysis of the Drosophila Wasp has demonstrated that the Cdc42-binding domain is not strictly necessary for function, suggesting that alternative pathways, such as phosphorylation can activate WASP (Tal et al.,2002). Indeed, some tyrosine kinases have been shown to activate Wasp by phosphorylation (Cory et al.,2002; Scott et al.,2002; Suetsugu et al.,2002).

In contrast to Wasp, Wave is not auto-inhibited. It is kept in an inactive state through association with a protein complex comprising Kette/Nap1, Sra1(specifically Rac associated 1, Sra-1; also called CYFIP/p140Sra-1, from here on called Sra-1 according to FlyBase) and the Abelson-interactor protein (Abi)(Eden et al., 2002). Upon dissociation or conformational changes of this complex, Wave is assumed to be active (Eden et al., 2002; Steffen et al., 2004). Thus,Kette or Sra-1 should antagonize Wave function. This is supported by genetic studies in Dictyostelium and Drosophila(Blagg et al., 2003; Bogdan and Klämbt, 2003). Cell culture experiments show that Wave is degraded in a Ubiquitin-dependent manner following disruption of the Sra-1/Kette complex(Kunda et al., 2003; Rogers et al., 2003). These latter findings are likely to reflect the fast inactivation of Wave once activated. In vitro, activation of Wave can be mediated by Rac1 or SH3 domains, which presumably bind to Sra-1(Eden et al., 2002; Kobayashi et al., 1998; Miki et al., 2000).

The Sra-1/Kette protein complex is not only required to negatively regulate the activity of Wave but is also able to activate Wasp function at the membrane (Bogdan and Klämbt,2003). The interaction of Kette and Wasp is not direct but is likely to be mediated by the Abi, which can bind to both Kette and Wasp(Bogdan and Klämbt, 2003). Interestingly, the Nck adapter protein is also able to bind to Wasp via its third SH3 domain (Rivero-Lezcano et al.,1995). Thus, Sra-1, which can bind to the first SH3 domain of Nck is a good candidate to locate the Sra-1/Kette complex to the membrane close to Wasp (Kitamura et al., 1996; Kitamura et al., 1997; Kobayashi et al., 1998).

To test whether the mutant phenotypes of sra-1 and ketteare alike as predicted, and whether Sra-1 indeed acts through Kette to regulate actin dynamics, we conducted a functional characterization of Sra-1 during Drosophila development. Sra-1 and Kette are both required for axonal growth and perform common functions during formation and maturation of neuromuscular junctions (NMJ). Analysis of temporal and spatial distribution of the Sra-1 protein shows a prominent co-expression with Kette. Both proteins are maternally expressed and later in development become concentrated in the developing nervous system (CNS). Sra-1 is highly expressed in growth cones and neuromuscular synapses. Direct interaction of Sra-1 and Kette depends on a short C-terminal domain of the Sra-1 protein. Expression of a Sra-1 variant lacking the C-terminal domain leads to a dominant-negative phenotype that can be suppressed by expression of an activated Kette protein. In tissue culture cells as well as in vivo Sra-1 function is required for F-actin organization. Further genetic analyses demonstrate that Sra-1 function at the membrane depends on the presence of Wasp.

All crosses were performed at 25°C unless otherwise indicated. The following strains were used: wasp¹/TM6 wasp³/TM; kette^J4-48/TM6, kette^Δ2-6/TM6 and kette^J1-70/TM6 (Hummel et al., 2000). To determine the gain- and loss-of-function phenotype of Kette^Myr, Sra^Myr, SraΔC^Myrand Sra-RNAi, we used the GAL4 system(Brand and Perrimon, 1993). Transgenic flies carrying the following constructs were generated following standard procedures: UAS-kette^Myc;UAS-sra^Myc;UAS-sra-ΔC^Myc and UAS-sraRNAi. The Kette^Myr, Sra^Myr and SraΔC^Myrconstructs were made by fusing DNA encoding the first 88 amino acids from Drosophila Src1 (Simon et al.,1985) to the first codon of Kette, Sra-full-length and Sra^1151aa. To generate the Sra-RNAi construct, the sra DNA(coding sequence +1 to +300, containing no repetitive sequence motif) was cloned as an inverted repeat into pWIZ(Lee and Carthew, 2003). Several independent transgenic lines were analyzed in each case.

The rabbit anti-Sra antibody (Sra900) was generated against the first N-terminal 300 amino acids of Sra fused with a His₆-tag (Qiagen). The His₆-Sra fusion protein was expressed in E. coli and purified with Ni-NTA resin (Qiagen) under native conditions. Rabbits were immunized with purified proteins by Davids Biotechnologie (Germany).

In situ hybridization and immunohistochemistry was performed as described(Hummel et al., 1997; Tautz and Pfeifle, 1989). Sra-specific riboprobes were generated from EST clones LD47929 and LD19991. Larval dissections and antibody staining were carried out as previously described (Schuster et al.,1996). Antibodies were used as follows: affinity-purified anti-Sra polyclonal antibody, 1:100; mAB nc46(Reichmuth et al., 1995),1:10; mAB BP102, 1:100 (Developmental Studies Hybridoma Bank). For protein extracts, staged embryos were mashed with a pestle in lysis buffer(1×PBS, 1% NP40, 1 mM DTT and 1 mM PMSF). Lysates were centrifuged 10 minutes at 10,000 g to yield the cytoplasmic supernatant. The amount of total protein was determined by a Bradford assay (BioRad). Equivalent amounts of protein extract were separated by SDS-PAGE, and analyzed by western blot as described previously(Bogdan et al., 2001).

To test the interaction between Sra-1 and Kette, we inserted the kette ORF in frame with the GAL4 DNA-binding domain into the GBK-T7 vector (Clontech) and the Sra-1 full-length and C-terminal deletion of Sra-1 in frame with the GAL4 activation domain into the GAD-T7 vector(Clontech).

Drosophila S2R+ (Yanagawa et al., 1998) cells were propagated in 1×Schneider's Drosophila media (Gibco) supplemented with 10% FBS, 50 units/ml penicillin and 50 μg/ml streptomycin in 75 cm² T-flasks(Sarstedt) at 25°C. For immunofluorescence microscopy,2×10⁵ cells were plated on glass cover slips (pretreated with or without fibronectin) in 24-well plates, cultured for 24 hours and fixed with 4% paraformaldehyde, permeabilized with 0.5% Triton X-100, stained with Drosophila anti-Kette antibody 97/82 (1:2000 dilution) or purified anti-Sra antibody (Sra900) followed by Alexa-fluorophore-conjugated goat anti-rabbit IgG antibody (Molecular Probes, 1:1000 dilution). For filamentous actin and nucleic acid staining, cells were prepared as described above and incubated in 1-2 units Alexa-fluorophore-conjugated phalloidin (Molecular Probes) and in 300 nM DAPI (Molecular Probes), respectively. The samples were mounted in a 25% (w/v) Mowiol (Sigma) solution containing DABCO (Sigma) and visualized on a Leica LSM.

For co-immunoprecipitation of Myc-tagged Kette protein, DrosophilaSchneider S2 cells (10⁷) were co-transfected with UAS-kette-7xmyc (2 μg) and Act5C-GAL4 DNA (2 μg) by Fugene (Roche) as described previously(Bogdan et al., 2001). Transfected cells were harvested and lysed in ice-cold lysis buffer(1×PBS, 0.5% NP40, 1 mM DTT and 1 mM PMSF). Lysates were centrifuged 10 minutes at 10,000 g to yield the cytoplasmatic supernatant. Total cell lysate was subjected to immunoprecipitation with monoclonal anti-Myc (9E10) immobilized by protein G beads. Immunoprecipitates were separated by SDS-PAGE, and proteins were analyzed by western blot as described previously (Bogdan et al.,2001) using polyclonal antiserum to Sra (Sra900).

sra-1 ORF (300 bp) and kette ORF (570 bp) were inserted into the vector pLITMUS-28i (BioLabs) flanked by T7 promoters. After linearization, dsRNAs were generated using a HiScribe RNAi Transcription Kit(BioLabs) as described previously (Bogdan et al., 2001). dsRNA-treated cells were incubated for additional 2-3 days and analyzed by immunofluorescence and western blot analysis.

kette and sra-1 are initially broadly expressed during embryogenesis (Hummel et al.,2000; Schenck et al.,2003) (see Fig. S1). At the end of embryogenesis, RNA expression of both genes becomes restricted to the CNS (Fig. 1A). To further analyze the expression of Sra-1 we generated anti-Sra-1 antibodies(see Materials and methods for details). Western blot analyses show expression of Kette and Sra-1 throughout development(Fig. 1B). Antisera against the two proteins each recognize a single band, which is present in cell extracts from first cleavage stage embryos to hatching(Fig. 1B).

As both proteins are co-expressed during development and their vertebrate homologues are found in the same protein complex, we asked whether the two proteins could physically interact. First, we transfected DrosophilaS2 tissue culture cells with a vector encoding a Myc-tagged Kette protein and cell lysates were subsequently subjected to immunoprecipitation. Sra-1 protein at the expected size of 140 kDa was only detected in these precipitates, when we precipitated Kette using anti-Myc antibodies but not when we used anti-Slit antibodies as a control (Fig. 1C). This co-immunoprecipitation indicates a direct binding of Kette and Sra-1. To further verify this assumption and to map the domain on the Sra-1 protein relevant for this interaction, we used the yeast two-hybrid system. We found that full-length Sra-1 protein efficiently binds to full-length Kette protein (Fig. 1D). A small deletion of the last 100 C-terminal amino acids of the Sra-1 protein was sufficient to abrogate the interaction between Sra-1 and Kette, suggesting that the C-terminus of the Sra-1 protein is essential for the interaction with Kette. In addition to Kette, the 140 kDa protein Sra-1 has been reported to bind to Nck, Rac1 and FMRP, indicating that Sra-1 acts as a scaffolding protein linking a larger number of proteins(Fig. 1E)(Kobayashi et al., 1998; Schenck et al., 2003).

Purified anti-Sra-1 antisera allow detection of the endogenous Sra-1 protein. The specificity of these antisera was confirmed by Sra-1 RNA interference (RNAi), which removed Sra-1 staining (see Figs 3, 5). In agreement with high levels of RNA expression in the CNS, we noted highest level of Sra-1 protein in the embryonic CNS neuropile, which consists of axons, dendrites and synapses (Fig. 2A). Similarly,Sra-1 is expressed in the optic lobe neuropile of third instar larvae(Fig. 2B). Here, high levels of Sra-1 protein expression can be noted in a zone between the developing lamina and medulla (Fig. 2B,asterisk). EM-analyses have previously showed that this region consists of growth cones of the developing photoreceptor neurons which also express the 24B10 epitope (Garrity et al.,1999; Garrity et al.,1996; Van Vactor et al.,1988) (Fig. 2B). In addition, we noted some Sra-1 protein in eye imaginal discs cells(Fig. 2C).

The expression of Sra-1 in growth cones suggests that it may be required for axonal pathfinding or targeting. To test this assumption, we generated transgenes expressing a double stranded sra-1 RNA under the control of the UAS sequences to perform in vivo RNAi experiments(Lee and Carthew, 2003). In wild-type eye discs, photoreceptor neurons express Sra-1 and the 24B10 marker as they navigate through the optic stalk to the lamina and medulla in the brain (Fig. 3A). After expression of the sra-1 RNAi construct in the developing eye disc using the eyeless-GAL4 driver, Sra-1 protein expression is removed in the eye disc but not in the brain, demonstrating the specificity of the antisera (Fig. 3B). Upon reduction of Sra-1 expression, photoreceptor neurons exhibit pronounced axonal defects. The majority of the axons appear to be able to enter the brain, but then axons fail to grow towards their correct targets(Fig. 3B). In addition, fewer photoreceptor cells develop and the compound eyes appear slightly smaller and have a rough appearance compared to the wild type(Fig. 3C,D). These phenotypes are highly penetrant and can be observed in every fly expressing sra-1 RNAi (n>50).

sra-1 function is not restricted to axonal pathfinding. Recently,defects in synaptic architecture were found in sra-1 mutants(Schenck et al., 2003). Similarly, inactivation of sra-1 by RNAi specifically in neuronal cells using elav-GAL4 led to synaptic boutons with supernumerary buds(2.0-fold increase compared with wild type, terminal boutons of 60 neuromuscular junctions on muscle 4 of 15 larvae were counted) and a pronounced bulged 3D structure (Fig. 4B). The cell type specific sra-1 knockdown demonstrates a cell-autonomous requirement for sra-1 function in the developing presynaptic neuron.

We then asked whether the interacting protein Kette is similarly required for synaptic development. Certain hypomorphic allele combinations allow development of third instar larvae (Hummel et al., 2000). kette^J1-70/kette^Δ2-6gives rise to larvae with smaller neuromuscular junctions carrying bulged synaptic boutons of irregular shape resembling the sra-1 phenotype(Fig. 4C; 2.1-fold increase of bud number of terminal boutons on muscle 4, n=60). As Sra-1 was shown to bind Kette via its C terminus (Fig. 1), we assumed that expression of a Sra-1 protein lacking this domain (Sra-1ΔC) may interfere with synaptic development. As wild-type Kette appears to function at the membrane, we generated a membrane tethered Sra-1ΔC protein variant (Sra-1ΔC^Myr).

Indeed, when we expressed such a construct in all neurons using the elav-GAL4 driver, we noted a synaptic phenotype similar to the one caused by reduction of sra-1 function by RNAi(Fig. 4D; 2.3-fold increase of bud number of terminal boutons on muscle 4, n=60).

A number of reports have recently shown that Sra-1 and Kette stabilize the Arp2/3 activator Wave (Blagg et al.,2003; Kunda et al.,2003; Rogers et al.,2003). Degradation of Wave following disruption of the regulatory complex may be an efficient way to terminate Wave activity. If all protein members of the complex are required to keep Wave inactive, one might expect that removal of Kette or Sra-1 should also lead to a downregulation of the remaining proteins of the complex. We have tested this assumption by following Kette or Sra-1 protein expression in S2 cells in which kette or sra-1 function was knocked down by RNAi. As previously shown, gene silencing of kette and sra-1 in Drosophila S2 by RNAi cell line induced identical morphological phenotypes(Blagg et al., 2003; Bogdan and Klämbt, 2003; Kunda et al., 2003; Rogers et al., 2003). In comparison with wild-type cells, dsRNA-treated cells assume a starfish-like morphology with multiple filopodia like extensions and a prominent accumulation of cytosolic F-actin (Fig. 5, arrowheads). Immunofluorescence and western blot analyses demonstrate that kette RNAi efficiently eliminates Kette protein expression (Fig. 5B,D). Interestingly, kette RNAi also reduces Sra-1 expression(Fig. 5). Conversely, when we reduced sra-1 function by RNAi, we blocked both Sra-1 and Kette protein expression (Fig. 5C,D). These data suggest that Kette and Sra-1 are closely linked and stabilize each other.

It has recently been shown that kette is required for F-actin formation (Bogdan and Klämbt,2003; Hummel et al.,2000; Kunda et al.,2003; Rogers et al.,2003). As the dynamics of the F-actin cytoskeleton is difficult to analyze in Drosophila neurons, we turned to the analysis of bristle development which crucially depends on F-actin dynamics(Jacinto and Baum, 2003). Hypomorphic kette mutants(kette^J1-70/kette^Δ2-6)are characterized by typical bent bristles(Fig. 6A, arrow; about 15% of the heads have one or more bent bristles, n=250). Given the general colocalization of Kette and Sra-1 and the mutual stabilization of the two proteins, we anticipated that sra-1 is also required for bristle development. Following expression of sra-1 dsRNA in the scabrous pattern, and formation of microchaete and macrochaete, we observed bent bristles similar to those in kette mutants(Fig. 6B, arrow; about 10% of the heads show one or more bent bristles, n=170). When we increased the number of sra-1 RNAi transgenes, the penetrance of the bent bristle phenotype increased (about 30% of the heads show one or more bent bristles, n=100). In addition, we observed a loss of some bristles in flies expressing high levels of sra-1 RNAi (in 5% of the heads, n=100). A similar increase in the phenotypic penetrance was observed when we lowered the dose of kette in flies expressing one copy of the sra-1 RNAi construct.

Expression of the Sra-1ΔC^Myr protein lacking the Kette interaction domain, resulted in a similar bristle phenotype. When one copy of the transgene was expressed, all heads lacked bristles. The few remaining bristles are frequently bended (Fig. 6C). When we increased the dose of the sra-1ΔC^Myr transgene, almost no bristles develop (data not shown).

To more directly test a possible role of sra-1 on F-actin formation, we performed phalloidin staining of bristles during pupal development. In wild-type animals, bristles are prefigured by pointed F-actin bundles (Fig. 7A). Cortical F-actin is found in a sharp zone close to the cell membrane. When we expressed sra-1 dsRNA, pointed F-actin bundles still form, but they show bent morphology (as do the later bristles) (Fig. 7B). When we expressed the Sra-1ΔC^Myr construct,no F-actin bundles were detected at the site where bristles are expected to develop, corresponding to the loss of bristles observed in adult stages.

We have recently shown that a membrane-tethered form of Kette behaves like an activated protein variant that efficiently reorganizes the formation of F-actin via modulating the activity of Wasp. When expressed in the developing notum of a fly, this results in the formation of branched or brushed bristles(Bogdan and Klämbt, 2003). If both Kette and Sra-1 act in the same complex, expression of a membrane-bound Sra-1 is expected to act like the membrane-bound Kette protein. Indeed, expression of a myristylated full-length Sra-1 protein(Sra-1^Myr) in the scabrous pattern induced the formation of brushed F-actin bundles (Fig. 7D, arrow). Furthermore the cortical F-actin appeared fuzzier and many irregular F-actin fibers can be recognized(Fig. 7D, arrowhead).

The F-actin phenotype seen following expression of the Sra-1^Myrprotein resulted in the formation of split bristles on the notum of the adult fly (Fig. 8B, about 10% of the microchaete, see Fig. 8I). Expression of a membrane-tethered Kette protein resulted in a similar,Wasp-dependent, split bristle phenotype(Fig. 8C,I)(Bogdan and Klämbt, 2003). Animals expressing both membrane-tethered proteins display a synergistic effect (Fig. 8C,I). To test whether the induction of the Sra-1^Myr induced phenotype also depends on the function of both kette and wasp we expressed Sra-1^Myr in heterozygous kette and wasp flies. Whereas the reduction in the gene dose of kette did not significantly modify the Sra-1^Myr induced phenotype, heterozygous loss of wasp function clearly suppressed the branched bristle phenotype evoked by Sra-1^Myr expression(Fig. 8E,F,I). This indicates that either Sra-1 does not require Kette to activate Wasp or the reduction in Kette expression level is not sufficient to detect phenotypic changes.

Expression of the dominant-negative Sra-1ΔC^Myr protein variant did not induce a branched bristle phenotype but led to a loss of bristles (Fig. 6C, Fig. 8G,I). This bristle phenotype could be suppressed by the expression of activated Kette^Myr (Fig. 8H,I). Furthermore, flies expressing both Sra-1ΔC^Myr and Kette^Myr show similarly split bristles as do flies expressing only the Kette^Myr protein. Thus,activated Kette acts downstream or independent of Sra-1 to regulate the activity of the F-actin regulator Wasp.

Axonal pathfinding and synaptic growth both required well regulated F-Actin dynamics. We have analyzed the role of Sra-1, which (together with Kette)constitutes an important regulator of Wasp mediated F-actin dynamics.

In addition to the CNS phenotypes, mutations in sra-1 and kette both lead to synaptic defects that are characterized by an overall reduction in size of the neuromuscular junction, as well as the induction of supernumerary buds in sra-1 and kette mutant synaptic boutons. It is known that new boutons often arise from existing ones by asymmetrical budding or symmetrical division(Zito et al., 1999), which in turn requires an intact regulation of the actin cytoskeleton(Luo, 2002; Notarangelo and Ochs, 2003). The increased number of branches as well as the bulged appearance of the synaptic boutons after depletion of kette or sra-1 function may reflect their function in regulating wasp(Bogdan and Klämbt, 2003). Indeed wasp mutants display synaptic phenotypes similar to those of sra-1 and kette (Coyle et al., 2004). Recently, it has been found that the adaptor protein Nervous wreck (Nwk) binds Wasp and is also required for normal synapse morphology (Coyle et al.,2004). Thus, Nwk might act as a scaffolding protein in the synapse assembling a Wasp activation complex comprising Sra-1, Kette and Abi.

As in Drosophila, mutations in several of the vertebrate orthologs of the above mentioned genes are associated with learning deficits,demonstrating the pivotal importance of F-actin dynamics for precise neuronal function (Eden et al., 2002; Endris et al., 2002; Soderling et al., 2003; Suzuki et al., 2000; Yamamoto et al., 2001).

Rapid remodeling of the F-actin cytoskeleton is mostly brought about by the Arp2/3 complex, which in turn is activated by members of the Wasp and Wave protein families (Machesky et al.,1994; Miki and Takenawa,2003; Nakagawa et al.,2001; Takenawa and Miki,2001). Wasp as well as Wave are potent F-actin nucleation factors. Obviously within the cell their activity must be tightly regulated. Whereas Wasp is auto-inhibited, Wave is trans-inhibited and requires the inhibiting Sra-1 Kette protein complex (Eden et al.,2002). Upon dissociation of this complex or conformational changes within the complex, Wave is active and presumably remains active until it is degraded via ubiquitination (Blagg et al.,2003; Eden et al.,2002; Kunda et al.,2003; Rogers et al.,2003; Steffen et al.,2004). This latter mechanism, which is frequently used in regulating the effective concentration of active proteins(Di Fiore et al., 2003; Moon et al., 2002; Ou et al., 2003), ensures that Wave activity lasts for only a short time period.

As we have shown here, Wave is not the only protein of the complex that is degraded upon disruption of the protein complex. Depletion of Kette not only leads to a loss of Wave but also of Sra-1. Vice versa, depletion of Sra-1 leads to a loss of Kette. Thus, ultimately the stability of all proteins of the inhibitory Sra-1 complex appears to be interdependent.

We have previously shown that Kette can activate Wasp-mediated F-actin formation (Bogdan and Klämbt,2003). We have shown that Sra-1 function also depends on Wasp. In both cases, the membrane localization of Kette or Sra-1 is essential,indicating that in vivo regulation of membrane recruitment of Sra-1 and Kette is important for function. The data presented in this work also suggest that membrane-bound Sra-1 or Kette proteins are both able to activate Wasp independently of each other.

Vertebrate homologues of Sra-1 and Kette were first identified in a complex with the SH2 SH3 adapter Nck (Kitamura et al., 1996; Kitamura et al.,1997). The N-terminal SH3 domain of Nck is thought to bind to Sra-1 (Kitamura et al., 1996),evoking a model where Nck recruits Sra-1 and the associated Kette protein to the membrane. We have previously shown that kette and dock,which encodes the Drosophila Nck homologue interact during axonal pathfinding (Garrity et al.,1996; Hummel et al.,2000). Cell signaling and cell adhesion leads to the activation of a number of receptor systems which in turn mediate anchorage-dependent recruitment of adapter proteins such as Nck(Howe, 2001; Howe et al., 2002; Li et al., 2001). Nck in turn is able to connect cell-surface receptors via different signal cascades to the F-actin cytoskeleton (Buday et al.,2002).

However, in vivo Nck cannot mediate all aspects of Sra-1/Kette function, as the complete loss of maternal and zygotic Nck results in similar but not identical phenotypes when compared with kette or sra-1loss-of-function phenotypes (Fan et al.,2003; Garrity et al.,1996; Hummel et al.,2000; Schenck et al.,2003). In the developing synapse, the function of Nck in recruiting Sra-1 and Kette to the membrane may be fulfilled by Nwk, which as Nck also binds Wasp (Coyle et al.,2004; Zeng et al.,2003). Interestingly, both adaptor proteins are involved in Slit Robo signaling, where they may mediate different biological effects(Fan et al., 2003; Wong et al., 2001). This suggests that combinatorial and tissue specific factors are assembled in response to specific cues to activate Wasp in different cell types or compartments.

We thank E. Buchner for the generous supply of the nc46 antibody. We are grateful to S. Yanagawa for providing the S2R+ cells, A. Giangrande and A. Schenk for sharing results prior publication, and T. Hummel for critically reading the manuscript. This work was supported by a grant of the DFG to S.B. and C.K.