A putative ortholog of the mammalian HSPC300 subunit of the WAVE/SCAR complex is annotated in FlyBase as CG30173. Using the available sequence information, we cloned Drosophila HSPC300 by RT-PCR and found it to be identical to the predicted gene sequence. In order to assess the expression profile of HSPC300, we performed quantitative real time RT-PCR on total RNA extracted from different developmental stages and determined that HSPC300 transcripts are present throughout the fly life cycle. Normalization against a housekeeping gene coding for ribosomal protein rp49 indicates that the levels of HSPC300 transcripts progressively increase during development (Figure 1a) and are particularly high in adult males and females. Transcript accumulation in ovaries is in line with high levels of HSPC300 RNA detected by in situ hybridization in embryos prior to onset of zygotic gene expression (data not shown).

A polyclonal antibody raised against the HSPC300 protein reveals a single band at the predicted molecular weight (8 kDa) in a western blot on Drosophila S2 cell extracts. Signal specificity was demonstrated first, upon transient HSPC300 overexpression, which induces significant enhancement of the 8 kDa band intensity (Figure 1b), and second, by loss of immunoreactivity in HSPC300 mutant extracts (see below). In embryos, HSPC300 primarily accumulates in the central nervous system (CNS), most prominent labeling localizing in axons along longitudinal tracts and commissures (Figure 1c). This expression pattern perfectly matches those previously reported for other components of the WAVE/SCAR complex 78. Immunolabeling specificity was further confirmed by loss of immunoreactivity in HSPC300 mutants (see below).

To address the role of HSPC300, we generated loss of function mutants. The EP(2R)0506 line harbors a P element on the right arm of the second chromosome at position 60B4 and is 100% homozygous viable. The P element lies 75 bases upstream of the HSPC300 start codon (ATG), and in the right orientation to drive HSPC300 gene expression (Figure 2a). HSPC300 is flanked by two genes for which no mutant strain is reported, CG3163, located 263 bases upstream, and PebIII, located 497 bases downstream (Figure 2a).

We generated HSPC300 alleles by imprecise excision of the EP(2R)0506 P element and recovered 116 independent excision events. Among them, four lines are homozygous lethal at late pupal stage and fail to complement each other. These excision alleles were characterized by PCR and sequencing in order to identify the breakpoints. In three of them, the entire HSPC300 coding sequences are deleted, as well as part of the neighboring PebIII gene (Figure 2b and data not shown). The fourth lethal excision line, HSPC300^Δ54.3, shows unaffected PebIII and CG3163 genes as well as the presence of an intact P element. In this line, the P element is unable to drive expression of HSPC300 as is the case with the original EP(2R)0506 line. To figure out the molecular lesions associated with this allele, we performed 3' and 5' inverse PCR. This confirmed the presence of intact junctions between the P element and surrounding genomic regions and revealed that the 3' end of the P element is flanked by a stretch of 208 bases of HSPC300 sequences followed by an insertion of unrelated sequences into the HSPC300 intron (see Figure 2b for schematic representation and Additional file 1 for molecular characterization of HSPC300^Δ54.3). We characterized this allele further by RT-PCR and detected PebIII and CG3163, but no HSPC300 transcripts (Figure 2c), indicating that the modified HSPC300 transcript is either not transcribed or very rapidly degraded. Thus, HSPC300^Δ54.3 represents a null allele that does not affect neighboring genes. This line was further used to characterize the mutant phenotypes and is referred to as HSPC300 unless otherwise indicated.

To provide formal evidence that excision line lethality is specifically due to loss of HSPC300, we generated UAS-HSPC300 transgenic animals and performed rescue experiments. Lethality observed in HSPC300^Δ54.3 and HSPC300^Δ96.1 lines is indeed rescued by HSPC300 re-expression using either actin-Gal4 or elav-Gal4 drivers (ubiquitous and pan-neuronal expression, respectively, data not shown). Altogether, these analyses demonstrate that HSPC300 is an essential gene required in the nervous system, its zygotic depletion inducing lethality prior to eclosion.

The prominent axonal localization of HSPC300 and phenotypes previously reported for mutations in other components of the WAVE/SCAR complex 6789 suggested a role for HSPC300 in axonogenesis. We therefore analyzed the integrity of the embryonic axonal network by immunolabeling with anti-FasII, which specifically recognizes six longitudinal fascicles of the CNS and motor axons, as well as with anti-BP102, which reveals connectives and commissures 3233 (Figure 3). In contrast to the results obtained with CYFIP, SCAR and Kette mutants, complete loss of zygotic HSPC300 (HSPC300^Δ54.3 and HSPC300^Δ96.1) does not induce any prominent axon defect (Figure 3c,d). We speculated that maternal HSPC300 contribution compensates for loss of zygotic gene function, masking an essential requirement during embryogenesis. To test this hypothesis, we generated homozygous HSPC300^Δ54.3 mutant clones within the germ line of heterozygous females. Deletion of both maternal and zygotic HSPC300 components leads to severe, though variable, nervous system defects ranging from broken and disorganized longitudinal connectives to remnants of axons or depletion of all CNS axon bundles (Figure 3i,j). Moreover, upon depletion of both maternal and zygotic HSPC300 components, lethality is shifted to embryonic stages and embryos appear overall disorganized. Identical observations were made using other HSPC300 alleles, such as HSPC300^Δ96.1 (data not shown).

The observed strong and pleiotropic defects did not allow us to score for specific neuronal requirement. We therefore aimed at reducing HSPC300 maternal dose without completely deleting it, using insertion line EP(2R)0506, referred to as HSPC300^EP0506 from now onwards. This line represents a hypomorphic allele since HSPC300^EP0506/HSPC300^Δ54.3 embryos show 30% viability (versus 100% viability for HSPC300^EP0506, and 0% viability for HSPC300^Δ54.3). HSPC300^EP0506/HSPC300^Δ54.3 females were mated to heterozygous HSPC300^Δ54.3 males. Amongst the progeny, we indeed noticed embryos with low maternal HSPC300 dose, reflected by reduced axonal HSPC300 immunolabeling in comparison with that observed in wild-type or zygotic null HSPC300 embryos under the same confocal microscope settings (Figure 3k–m). The resulting reduction in maternal HSPC300 dose produces embryos that overall show normal morphology (in contrast to those that are completely devoid of maternal HSPC300), allowing us to examine late CNS development. The embryonic phenotypes observed in this mutant condition, designated as maternal hypomorph, are similar, but of lower penetrance compared to those shown by other members of the WAVE/SCAR complex (midline crossing in 10% of embryos (n = 100) compared to 79% in CYFIP embryos (n = 150) 7) (Figure 3e,g). Additional defects shared by mutants affecting the other subunits of the WAVE/SCAR complex 8 are disorganized fascicles, improper separation of commissures and broken longitudinal bundles (Figure 3e–h). Taken together, these data clearly implicate HSPC300 in axonogenesis, presumably through regulation of actin cytoskeleton remodeling by SCAR- and Arp2/3-dependent mechanisms. They moreover indicate that the HSPC300 maternal component is sufficient to sustain development in embryos that are devoid of zygotic HSPC300.

In addition to the embryonic axonal defects, mutations in other subunits of the WAVE/SCAR complex lead to synaptic defects that are characterized by size reduction of the NMJ and by supernumerary buds on pre-existing boutons 7819. The fact that HSPC300 as well as CYFIP, SCAR and Kette mutants show similar embryonic axonal defects prompted us to examine the structures of synaptic terminals in HSPC300 larvae.

NMJs were labeled using anti-Disc Large antibody, a preferred marker to reveal NMJ morphology 7834. NMJs of HSPC300 larvae are severely reduced in length and display supernumerary buds (Figure 4d–h) compared to those of wild-type animals, those of rescued animals or those observed in a precise excision line (Figure 4a–c,g,h). The latter represents a revertant as it displays the same genetic background as that of imprecise excision lines (Figure 4g,h).

To quantitatively assess synaptic HSPC300 phenotypes, we measured total length of synaptic terminals using a computer-assisted program 7 and counted buds per synaptic terminal. We focused our analysis on muscle 4 in abdominal segments A2-A4, innervated by the inter-segmental nerve (ISN). Since NMJ growth varies in proportion to muscle size 35, we normalized the total length of each synapse with its corresponding muscle surface area. Normalized synapse length of HSPC300 larvae corresponds to 66% of the wild-type NMJ length, a statistically highly significant difference (1.36 × 10^-3 ± 0.11 μm^-1 versus 2.06 × 10^-3 ± 0.137 μm^-1 (mean ± standard error of the mean (SEM)); p < 0.001). Likewise, there is 88% increase in the bud number compared to wild type (6.42 ± 0.60 versus 3.51 ± 0.39 (mean ± SEM); p < 0.001) (Figure 4g,h). Strong labeling with active zone-specific antibody NC82, which recognizes the Bruchpilot protein 3536, indicates that HSPC300 synapses are functional (Figure 4i–n). Interestingly, compared to wild-type NMJs, in which NC82 labeling preferentially localizes at boutons, mutant NMJs show uniform labeling. This is in line with the finding that the characteristic bead string structure of NMJs is not preserved in HSPC300 larvae, in which boutons cannot be easily identified. This phenotype is shared by mutations affecting the other subunits of the WAVE/SCAR complex 7837 and may be related to the supernumerary budding phenotype.

Finally, NMJ phenotypes are rescued by elav-Gal4-driven HSPC300 expression but not by postsynaptic HSPC300 expression induced by the muscle-specific mhc-Gal4 driver (Figure 4g,h, and data not shown). Taken together, these results reveal that HSPC300 is required by motoneurons for synaptic morphogenesis.

The phenotypic similarity of axonal and synaptic architecture of HSPC300, CYFIP, Kette and SCAR mutants strongly suggests synergy amongst the wild-type molecules and indicates that HSPC300 represents a functional subunit of the Drosophila WAVE/SCAR complex in vivo. To provide evidence for the physical association of HSPC300 and the other WAVE/SCAR complex subunits, we performed co-immunoprecipitation experiments in Drosophila S2 cells. Both CYFIP and SCAR proteins were found to co-immunoprecipitate with HSPC300 (Figure 5a), as revealed by using anti-CYFIP (band at 145 kDa) and anti-SCAR (68 kDa) antibodies 79. Control experiments performed in parallel with unrelated rabbit IgG did not co-immunoprecipitate either of these proteins, thus confirming the specificity of the observed interactions. Thus, Drosophila HSPC300 appears to represent, like its mammalian counterpart, a subunit of the WAVE/SCAR complex.

We and others have previously demonstrated that CYFIP, Kette and SCAR are essential for the integrity and the stability of the WAVE/SCAR complex in vivo. If any of them is missing, the other subunits fail to be detected 82122232425, most likely due to proteasome-mediated degradation 23. In line with the observation that zygotic HSPC300 null embryos show neither significant reduction of HSPC300 labeling (Figure 3l) nor axonal defects (Figure 3c,d), other WAVE/SCAR complex proteins (SCAR and CYFIP) accumulate at wild-type levels in these embryos (Additional file 2). In order to determine the stabilizing potential of HSPC300, we therefore assessed the levels of CYFIP and SCAR later in development, in third instar HSPC300 mutant larvae, a developmental stage in which maternal contribution has faded. Indeed, we noticed a considerable decrease in CYFIP, Kette, SCAR and Abi protein levels compared to those observed in wild-type larvae. Qualitatively similar results were obtained in zygotic null larvae and in larvae that completely lack the zygotic component and also are maternal hypomorphs, even though the latter genotype shows a stronger defect (Figure 5b). Interestingly, there appears to be yet stronger reduction of these proteins in the extracts prepared from CYFIP zygotic null animals (Figure 5b).

To provide formal evidence that HSPC300-induced phenotypes depend on its functional relation with the WAVE/SCAR complex, we performed genetic interaction experiments at the NMJ. We took advantage of our previous observation that single mutations in the subunits of the WAVE/SCAR complex show reduced synapse length in a dose-dependent manner. Heterozygous mutants indeed already show significantly reduced synaptic length 8 (Figure 4g). Interestingly, while double heterozygous HSPC300, CYFIP NMJs are similar to those from larvae that are heterozygous for either gene, imbalance between HSPC and CYFIP content using a sensitized background results in a marked further decrease of synaptic terminal length: the strong CYFIP transheterozygous allelic combination CYFIP^Δ85.1/CYFIP^EP3267 (that is, null over hypomorph), in an otherwise heterozygous HSPC300 background, leads to a significant reduction in the synapse length (Table 1). This result indicates that HSPC300 and CYFIP synergistically affect synapse growth and provides a first assessment of the molecular pathway requiring HSPC300.

The above data clearly demonstrate that HSPC300 animals share nervous system defects with SCAR, CYFIP and Kette mutants. Flies mutant for subunits of the WAVE/SCAR complex are also characterized by misshaped macrochaetae 51920, a phenotype due to defects in F-actin. The phenotype of bent bristles is also observed in HSPC300 pharate adults (Figure 6, arrows). However, specification of sensory organ precursors (SOPs) is not affected and the number of bristles on the head and thorax are normal (Figure 6), as in mutant SCAR and CYFIP conditions 9 (our unpublished data). This result is particularly interesting in view of the observation that HSPC300 also directly interacts with Abi 29, which in turn interacts with and activates WASP 530. The Abi-WASP biochemical interaction is further sustained by the common phenotype of WASP and knockdown Abi mutant flies, which lack macrochaete on the thorax, due to SOP cell fate defects 520. On the basis of the phenotypical similarity, we conclude that loss of HSPC300 affects signal transduction through SCAR, not WASP.

The SCAR, CYFIP, Kette and Abi subunits of the WAVE/SCAR complex have been shown to impact on various processes and structures that depend on actin cytoskeletal remodeling. These include Dictyostelium motility 21, development of plant trichomes 27, as well as egg chamber structure and nuclear positioning in the blastoderm of Drosophila 9. Crucial for the versatile functions of the complex is its structural integrity. Loss or mutations in any subunit (CYFIP, Kette or SCAR) leads to proteasome-mediated degradation of the others and, as a consequence, identical mutant phenotypes 82123. Surprisingly, HSPC300 knockdown in Drosophila S2 cells has been reported to produce much milder cytoskeletal phenotypes than those produced by targeting other components of the WAVE/SCAR complex 23. In addition, in in vitro experiments, lack of HSPC300 affects neither complex assembly 29 nor the complex's ability to activate Arp2/3 in vitro 22, two features that characterize other complex subunits. These data raise questions as to the real role of HSPC300 in the context of the WAVE/SCAR complex.

Presented data demonstrate that Drosophila HSPC300 constitutes a subunit of the WAVE/SCAR complex in vitro and in vivo. Like the other subunits of the complex, HSPC300 is highly expressed in the developing fly nervous system and is crucially required for axonogenesis and neuromuscular synapse morphogenesis. Moreover, HSPC300 loss of function conditions are marked by a decrease in all members of the WAVE/SCAR complex, a result that is in keeping with a model in which each subunit, including HSPC300, significantly contributes towards the stability of the complex, notably in vivo. By using HSPC300 allele combinations that lead to different amounts of HSPC300 protein, we have revealed that a sharp threshold exists and that the maternal component is sufficient to ensure normal embryonic development and viability. The stringent genetic conditions we generated (loss of zygotic in addition to partial or complete loss of maternal HSPC300) also allowed us to reveal that strong loss of HSPC300 protein is necessary to cause dramatic consequences comparable to those observed in mutants affecting other complex components, thereby suggesting that the previously obtained mild phenotypes (RNAi-mediated knockdown in cells 23) merely result from the limitation of the utilized technique. Thus, highly similar requirements for HSPC300 and the WAVE/SCAR complex components control cell morphology in CNS neurons and at NMJs.

The genetic interaction observed upon disrupting the balance between CYFIP and HSPC300 levels shows a positive/synergistic role for HSPC300 in the Rac1-WAVE/SCAR complex pathway to control synapse length, thereby providing first genetic evidence for HSPC300 functioning in this pathway. Taken together, the phenotypes and genetic interactions in the fly nervous system, as well as the phenotypes previously described in Arabidopsis 2627, provide strong evidence for HSPC300 being an evolutionarily important integral part of the Rac1-WAVE/SCAR-Arp2/3 pathway.

Interestingly, it has been shown that the control of cotyledon cell size requires Arabidopsis Brick1 but not WAVE/SCAR-Arp2/3 27, suggestive of an additional function that does not depend on the WAVE/SCAR complex. These data are in line with the existence of a significant fraction of free and soluble vertebrate HSPC300 and with our finding that phenotype and genetic interactions can only be revealed upon strong HSPC300 depletion and imbalance, respectively. Whether HSPC300, similar to CYFIP, Kette and Abi, works on additional pathways that are independent of the WAVE/SCAR complex remains to be elucidated. Since HSPC300 is the most conserved subunit of the WAVE/SCAR complex not only in the animal kingdom and Dictyostelium, but also in plants, and since no HSPC300 paralogous gene exists in flies, we expect our data to be of predictive value for HSPC300 indispensability with respect to the function of its associated complex in other organisms. We further expect the generated mutant animals to facilitate identification of novel HSPC300-dependent pathways.

The wild-type strain used was Sevelen. Line EP(2R)0506 was provided by the Szeged Stock Center (Szeged, Hungary), and elav-Gal4 (C155) by the Bloomington Stock Center (Bloomington, Indiana, USA). HSPC300 excision mutants were obtained upon transposon mobilization in the EP(2R)0506 line after isogenization. Transgenic lines were obtained using standard protocols. UAS-HSPC300 (line 4b.1) was recombined onto the HSPC300^Δ54.3 carrying chromosome. FRT42B was recombined separately with HSPC300^Δ54.3 and HSPC300^Δ96. Germ line clones were induced in hs-FLP; FRT42B Ovo^D/FRT42B, HSPC300^Δ54.3 flies as described 45. Other utilized strains were CYFIP^Δ85.1, Kette⁰³³³⁵ and SCAR^Δ37 8.