In the Drosophila embryo, transient cell adhesion during myoblast fusion is known to lead to the formation of fusion-restricted myogenic-adhesive structures (FuRMASs). Here, we report that within these FuRMASs, a Drosophila homologue of human and mouse swiprosins (EF-hand-domain-containing proteins) is expressed, which we named Drosophila Swiprosin-1 (Drosophila Swip-1). Drosophila Swip-1 is highly conserved and is closely related to the calcium-binding proteins swiprosin-1 and swiprosin-2 that have a role in the immune system in humans and mice. Our study shows that Drosophila Swip-1 is also expressed in corresponding cells of the Drosophila immune system. During myoblast fusion, Drosophila Swip-1 accumulates transiently in the foci of fusion-competent myoblasts (FCMs). Both the EF-hand and the coiled-coil domain of Drosophila Swip-1 are required to localise the protein to these foci. The formation of Drosophila Swip-1 foci requires successful cell adhesion between FCMs and founder cells (FCs) or growing myotubes. Moreover, Drosophila Swip-1 foci were found to increase in number in sing22 mutants, which arrest myoblast fusion after prefusion complex formation. By contrast, Drosophila Swip-1 foci are not significantly enriched in blow2 and ketteJ4-48 mutants, which stop myogenesis beyond the prefusion complex stage but before plasma membrane merging. Therefore, we hypothesise that Drosophila Swip-1 participates in the breakdown of the prefusion complex during the progression of myoblast fusion.
Membrane fusion is a fundamental feature found in many processes such as vesicle fusion with plasma membranes, fusion of intracellular vesicles, viral fusion and cell–cell fusion, e.g. epithelial fusion and sperm–egg fusion (Martens and McMahon, 2008; Sapir et al., 2008). In this study, we focus on cell–cell fusion during myogenesis in Drosophila.
In higher organisms, myotubes are multinuclear and arise by myoblast fusion during development, as well as during postnatal growth and repair. In Drosophila, founder cells (FCs) determine the final identity of larval muscles, whereas fusion-competent myoblasts (FCMs) fuse to them until the final characteristic number of nuclei is reached (Abmayr et al., 2008; Chen and Olson, 2004; Maqbool and Jagla, 2007). Fusion in the somatic mesoderm progresses in two temporal phases within approximately 5.5 hours, with the individual fusion event taking only a matter of minutes (Richardson et al., 2007). These fusion events are characterised by transient F-actin plugs or foci at the contact sites of myoblasts (Richardson et al., 2008b). Studies in Drosophila have provided key insights into the underlying mechanisms and key components involved in cell adhesion, signalling, local F-actin accumulation and branching by activation of the Arp2/3 complex (Haralalka and Abmayr, 2010; Önel and Renkawitz-Pohl, 2009; Rochlin et al., 2010).
Heterologous cell adhesion is mediated by immunoglobulin (Ig) transmembrane cellular receptors. Dumbfounded/Kin of Irre (Duf/Kirre) is known to enable FCs and growing myotubes to adhere to FMCs and acts in functional redundancy to Roughest/Irregular chiasma C (Rst/IrreC) (Ruiz-Gómez et al., 2000; Strünkelnberg et al., 2001). In addition, the FCM-specific protein Sticks-and-stones (Sns) also enables cell adherence between FCMs with FCs or growing myotubes and can act in partial redundancy with Hibris (Shelton et al., 2009). Duf/Kirre, Rst/IrreC and Sns were shown to form a ring-shaped structure at the contact sites of FCs or growing myotubes and FCMs, and to establish the fusion-restricted myogenic-adhesive structure (FuRMAS) (Kesper et al., 2007). After successful cell adhesion, F-actin and its regulators accumulate as plugs or foci in the centre of this structure (Kesper et al., 2007; Richardson et al., 2007). Furthermore, prefusion complex formation, exocytosis of its vesicles and membrane breakdown have all been proposed to occur at the centre of the FuRMASs (Önel and Renkawitz-Pohl, 2009). Interestingly, FuRMASs share many features with the immunological synapse (IS), podosomes and invadopodia, such as transient adhesive rings, as well as F-actin accumulation and branching (Kesper et al., 2007; Önel and Renkawitz-Pohl, 2009). Finally, recent ultrastructural analyses have revealed that podosome-like structures are present in FCMs that are fused with FCs or growing myotube (Sens et al., 2010).
Membrane-associated, ultrastructural features during myoblast fusion
Importantly, myoblast fusion has been studied intensively at the ultrastructural level in wild-type (Doberstein et al., 1997) and mutant embryos (Berger et al., 2008; Gildor et al., 2009; Kim et al., 2007; Massarwa et al., 2007; Schröter et al., 2004; Schäfer et al., 2007). Of particular interest are the membrane-associated structures, because fusion is a membrane-merging process. First, prefusion complexes are formed by accumulation of electron-dense vesicles at opposing membranes between FCs or growing myotubes and FCMs. At the individual contact site, these pairs of vesicles are distributed over an area of 1 μm2. Occasionally, electron-dense, desmosome-like plaques are observed in the opposing membranes; however, their nature and function is currently unclear. The opposing membranes vesiculate into many membrane remnants (Doberstein et al., 1997), thereby generating cytoplasmic continuities over 4 μm wide. Subsequently, the fusing FCMs are integrated into the growing myotube.
The progression from prefusion complex to membrane breakdown
The electron-dense vesicles of the prefusion complex are proposed to exocytose fusion-relevant molecules (Doberstein et al., 1997). To date, only one protein has been characterised that might be needed to enable progression from prefusion complex to exocytosis of these vesicles. This protein was named Singles bar (Sing), a multipass transmembrane protein with a MARVEL domain, which is expressed in both FCs and FCMs (Estrada et al., 2007). However, the subcellular localisation of Sing is still unknown.
Point mutation in homozygous sing22 mutant embryos is known to lead to a missense mutation (A46V) in translation and to an accumulation of prefusion complexes (Estrada et al., 2007). In blown fuse (blow2) loss-of-function mutant embryos, the prefusion complexes are assembled and dissolved, but the opposing plasma membranes do not vesiculate (Doberstein et al., 1997; Estrada et al., 2007). Blow is solely expressed in FCMs (Schröter et al., 2006) and accumulates with F-actin plugs or foci after successful cell adhesion (Kesper et al., 2007). Genetically, blow interacts with kette, which suggests a function in F-actin regulation (Gildor et al., 2009; Schröter et al., 2004). Moreover, in ketteJ4-48 loss-of-function mutant embryos, plasma membranes remain intact (Doberstein et al., 1997; Schröter et al., 2004). Other mutants for proteins that are predicted to be relevant for actin regulation stop development at later stages during fusion. To date, there are contradictory data for the gene encoding the WASP-interacting protein (Wip), also known as Verprolin. Studies on the solitaryS1946 allele of wip revealed that Sltr might be involved in targeting and coating vesicles that are required for myoblast fusion; however, the membranes of FCMs and FCs or growing myotubes, which were brought in close proximity for fusion, were found to remain intact (Kim et al., 2007). By contrast, the loss-of-function allele wip30D leads to fusion arrest with vesiculating membranes (Massarwa et al., 2007) and Arp2/3 activation is necessary to obtain complete cytoplasmic continuity (Berger et al., 2008). However, Arp2/3 seems not to be relevant for exocytosis of the electron-dense vesicles or for membrane merging (Önel and Renkawitz-Pohl, 2009; Önel, 2009).
FuRMAS, a transient adhesion-signalling centre, is comparable to the immunological synapse
In a topological model, we proposed that the prefusion complex is surrounded by a ring of cell-adhesion molecules: Duf/Kirre on the site of FCs or growing muscles and Sns on the site of the FCMs. Thereby, a transient FuRMAS signalling centre is established; a structure that is proposed to trigger myoblast fusion (Kesper et al., 2007; Önel and Renkawitz-Pohl, 2009). The prefusion complex resembles the accumulation of synaptic vesicles on the site of the axon at the neuromuscular synapse. For the neuromuscular synapse, exocytosis is known to be triggered in the presence of calcium (Schweizer and Ryan, 2006). In mouse cell culture (C2C12 cells), intracellular calcium increases before fusion and can induce myoblast fusion (Horsley and Pavlath, 2004).
Interestingly, the FuRMASs and the IS share many common features (Önel and Renkawitz-Pohl, 2009). For example, they both contain adhesion molecules and actin regulators are required to produce F-actin plugs or foci at the centre of the respective structures. Moreover, IS formation leads to increased cytoplasmic calcium concentration in T-cells (Kummerow et al., 2009). Therefore, there might be a comparable process during myoblast fusion in Drosophila. To add weight to this notion, calcium is known to be bound by EF-hand-domain-containing proteins (Gifford et al., 2007) including swiprosins (Vega et al., 2008; Dütting et al., 2011).
Mouse swiprosin-1/EFHD2 is an EF-hand and coiled-coil containing protein, which was identified in lipid rafts of mouse B-cells (Avramidou et al., 2007; Mielenz et al., 2005). Furthermore, Swiprosin-1/EFHD2 is thought to control B-cell-receptor signalling in mouse (Kroczek et al., 2010) and cytokine expression in human mast cells (Ramesh et al., 2009; Thylur et al., 2009), and it is also known to be present in human cytotoxic lymphocytes (Vuadens et al., 2004).
Here, we present our analysis of Drosophila Swip-1 distribution during myoblast fusion in wild-type embryos and show selective accumulation in FuRMASs on the site of FCMs. Our results show that Drosophila Swip-1 does not accumulate in sticks-and-stones (sns) mutants. Therefore, we predict that Drosophila Swip-1 accumulation depends on Sns-mediated cell adhesion and also probably signalling. Importantly, Drosophila Swip-1 foci were found to accumulate in singles bar (sing22) mutants; a phenomenon not observed in blown fuse (blow2) and kette (ketteJ4-48) mutants. We therefore hypothesise that Drosophila Swip-1 is one of the molecules that either directly or indirectly regulates calcium-dependent exocytosis of the electron-dense vesicles of the prefusion complex.
CG10641 encodes a protein with two EF-hand domains and one coiled-coil domain and is related to mammalian swiprosin proteins
As a result of the many shared features between FuRMASs and ISs, we asked whether calcium-binding proteins might be involved at a distinct time point during myoblast fusion. Therefore, we took a data-mining approach and searched databases (e.g. FlyBase) for putative calcium-binding proteins that are expressed preferentially in the mesoderm. This search revealed the Drosophila gene CG10641 (Tweedie et al., 2009). CG10641 appears to have a role in myogenesis, as indicated first by a screen for putative Twist targets (Furlong et al., 2001) and then in a screen for transcripts enriched in FCMs or FCs (Estrada et al., 2006).
CG10641 is predicted to encode a 217 amino acid cytosolic protein and similarly to the human and mouse swiprosins, it also has two EF-hand domains and one coiled-coil domain at the C-terminus (Fig. 1). Furthermore, CG10641 also contains one SH3 domain recognition motif at the N-terminus – three of which characterise the N-terminal region of mammalian swiprosins. Interestingly, mammalian Swip-1/EFhd2 is known to bind calcium – at least in vitro (Vega et al., 2008). Thus, in many aspects, the Drosophila protein is highly conserved with mammalian swiprosins (Fig. 1), and therefore, we named the CG10641 protein Drosophila Swip-1 in accordance with Kroczek and colleagues (Kroczek et al., 2010).
Expression in the immune system is a conserved feature between Drosophila and mammalian swiprosin
Estrada and co-workers (Estrada et al., 2006) showed that CC10641/Drosophila Swip-1 is transcribed at the extended germ-band stage shortly after the determination of FCs and FCMs in the visceral and somatic mesoderm, as well as in hemocytes of the head mesoderm at this stage. In this study, we verified these transcription data and analysed the detailed transcript distribution of Drosophila Swip-1 during myoblast fusion (supplementary material Fig. S1). We found that Drosophila Swip-1 transcripts are not maternally contributed to any significant extent, and that transcripts are first detectable at the beginning of gastrulation in the mesoderm and then persist in the somatic and visceral mesoderm until the completion of myogenesis (supplementary material Fig. S1).
To analyse the subcellular distribution of Drosophila Swip-1 during embryogenesis, we raised an antibody in rabbits against 18 amino acids from the C-terminal region of Drosophila Swip-1 (see Fig. 1 and the Materials and Methods). At the extended germ band stage, Drosophila Swip-1 was visible in the somatic and visceral mesoderm (Fig. 2A). During myoblast fusion, Drosophila Swip-1 was visible in many foci within the somatic mesoderm (Fig. 2B,C; Fig. 3A–C). Towards the end of embryogenesis, only a few Drosophila Swip-1 foci were visible in the somatic myoblasts and Drosophila Swip-1-positive stripes were detected at the muscle ends (Fig. 2D). In the pharynx, Drosophila Swip-1 accumulated at the contact sites of the individual muscles (Fig. 2F). A ventral view visualised Drosophila Swip-1 expression in the nervous system (Fig. 2E); this expression was observed from stage 13 onwards. At higher magnification, it was evident that at stage 16, Drosophila Swip-1 accumulates at the ends of the muscles towards their epidermal attachment sites (Fig. 2G, arrows). Drosophila Swip-1 localised closer to the epidermal attachment sites, which were marked by βPS integrin (for review see Brown, 1993), compared with the majority of microtubules, which were arranged in parallel to the length of the muscles (Fig. 2H,I).
In addition, Drosophila Swip-1 expression was also detectable in hemocytes in the head region (Fig. 2B, arrow), as well as in macrophage-like hemocytes in the region between the epidermis and the nervous system, where it was concentrated in distinct areas (Fig. 2J, arrow). β3-tubulin is known to be expressed in the macrophage-like hemocytes (Leiss et al., 1988). Therefore, we used antibodies against Drosophila Swip-1 and β3-tubulin in double-labelling studies and found that Drosophila Swip-1 does not overlap with the majority of microtubules (Fig. 2K). Taken together, we conclude that the role of swiprosin in the immune system is a conserved feature between Drosophila and mammals and that Drosophila Swip-1 is involved in further physiological and developmental processes.
Western blot analysis revealed a band at 25 kDa in extracts of embryos, larvae, pupae and hemocyte-like SL2 cells, which appeared as a double band in embryos, larvae and pupae (Fig. 2L, arrow). In contrast to vertebrates (for review see Dütting et al., 2011), there is no evidence for splice variants (Tweedie et al., 2009) and therefore, the additional band might indicate that Drosophila Swip-1 can carry a secondary modification in Drosophila. Indeed, vertebrate Swip-1 is known to be phosphorylated (Dütting et al., 2011). Finally, the specificity of our Drosophila Swip-1 antibody was demonstrated by transfecting SL2 cells with Drosophila Swip-1 constructs including or lacking the antigen (supplementary material Fig. S2).
Drosophila Swip-1 is mainly visible in FCMs during myoblast fusion and accumulates transiently in foci at the contact sites between FCs or growing myotubes and FCMs
As mentioned above, during fusion-relevant stages, Drosophila Swip-1 was detected in numerous foci within the somatic mesoderm. The pattern of these foci varied from segment to segment and from stage to stage (Fig. 3A–C, arrowheads). Thus, not all myoblasts showed these Drosophila Swip-1 foci at the same time. Late in myogenesis, at stage 15, only a few fusion events took place, and thus, fewer Drosophila Swip-1 foci were present and these were restricted to the FCMs that were in contact with a growing myotube (Fig. 3C, arrowheads). These observations were confirmed by visualisation of Drosophila Swip-1 foci in the rP298-LacZ enhancer trap line (Nose et al., 1992), where nuclei of FCs or growing myotube were labelled (Fig. 3Da–c, optical sections following one FCM (arrow) with respect to the growing myotube). These analyses revealed that Drosophila Swip-1 is not expressed in FCs and growing muscles, but is restricted to FCMs. In FCMs without contact to a FC or growing myotube Drosophila Swip-1 was distributed throughout the cytoplasm at a low level. Because Drosophila Swip-1 was barely detectable in the growing myotube, it could be degraded after successful fusion or transported towards the attachment sites.
Loss of Drosophila Swip-1 does not interfere with myogenesis
To better determine the exact biological function of Drosophila Swip-1, we first analysed P-element insertions upstream of the translation initiation codon of the Drosophila Swip-1 gene or localised in the intron. All of these P-element line insertions yielded homozygous viable flies that expressed Drosophila Swip-1 at a level indistinguishable from that in the wild type and exhibited normal myoblast fusion (see supplementary material Table S1).
We also mobilised the P-element P(SUPor-P)CG10641KG08194 and screened 1000 jump-out lines for Drosophila Swip-1 expression using PCR. No line was found to have lost Drosophila Swip-1 expression during myoblast fusion (data not shown). Once again, no evidence was found that Drosophila Swip-1 expression was affected during myoblast fusion.
Therefore, we next tested several Drosophila Swip-1-deficient fly lines and deficient fly lines in trans (supplementary material Fig. S3 and Table S2). Note that these deficient fly lines also delete mind bomb2 (mib2) (Carrasco-Rando and Ruiz-Gómez, 2008; Nguyen et al., 2007). The absence of Drosophila Swip-1 did not enhance the mib2 phenotype and therefore, we conclude that a loss of Drosophila Swip-1 causes no obvious defects during myoblast fusion.
In addition, we performed RNAi experiments to reduce the level of Drosophila Swip-1 specifically in the mesoderm taking RNAi for blow as a control (Fig. 4). However, although RNAi nearly completely knocked down Drosophila Swip-1 expression in the mesoderm, expression was maintained in the nervous system (Fig. 4A) and the musculature develops similarly to the wild type (Fig. 4B). Visualisation of Drosophila Mef2 in all mesodermal nuclei revealed that the number and localisation of the nuclei in these muscles was similar to that in wild-type embryos. For example, the VA2 muscle contained 7 nuclei, whereas in the wild type, this was between 6 and 13 (Beckett and Baylies, 2007) (Fig. 4E,F). We therefore conclude that the absence of Drosophila Swip-1 expression does not significantly affect myoblast fusion.
Thus, we asked whether other proteins could compensate for a loss of Drosophila Swip-1. There are 70 EF-hand-domain-containing proteins predicted to be encoded by the Drosophila genome, of which 63 are proposed to be cytosolic (Chintapalli et al., 2007). We have so far analysed CG2256, Calmodulin, CG31650 and Troponin C (1–4; next paragraph), which contain a similar domain structure to Drosophila Swip-1 (supplementary material Table S3).
(1) FlyBase (Chintapalli et al., 2007) predicts a single gene (CG2256) on the X-chromosome that encodes a small protein similar to Drosophila Swip-1 with two EF-hand domains and one coiled-coil domain. We found no CG2256 transcripts in the embryo (data not shown). Thus, it is unlikely that CG2256 acts in redundancy to Drosophila Swip-1. (2) Calmodulin is the best-characterised cytosolic and highly conserved EF-hand-domain-containing and calcium-binding protein. The Drosophila genome contains a single calmodulin gene (cam) (Yamanaka et al., 1987), which we analysed. (3) CG31650 encodes a protein with five EF-hand domains but no coiled-coil domain. The P-element insertion line P(SUPor-P)CG31650KG09054 has an integration upstream of the predicted ATG of CG31650 (Chintapalli et al., 2007). (4) Troponin C is a calcium-binding protein that acts in concert with Tropomyosin in muscle contraction. TroponinC73F (TpnC73F) (Fyrberg et al., 1994) is expressed in the mesoderm at fusion-relevant stages (Herranz et al., 2004); a finding that is corroborated by our in situ hybridisation data (supplementary material Table S3). We examined double mutants and visualised the musculature by observing β3-tubulin distribution. All corresponding mutants showed only the mib2 phenotype caused by Df(2L)Exel6042 and thus are concluded not to be potential redundancy partners to Drosophila Swip-1 (supplementary material Table S4). Thus, it remains to be clarified whether there are other proteins that can replace or compensate for the function of Drosophila Swip-1.
In Drosophila SL2 cells, the coiled-coil domain of Drosophila Swip-1 is required for localisation to the plasma membrane
We next investigated the role of the Drosophila Swip-1 domains for subcellular localisation. For this purpose, we first analysed the distribution of Drosophila Swip-1 in Drosophila Schneider cells (Schneider, 1972). We used the subclone SL2 because SL2 cells have many characteristics of hemocytes (Armknecht et al., 2005) and Drosophila Swip-1 is known to be expressed in hemocytes, macrophage-like, cells in the embryo. We found that Drosophila Swip-1 was endogenously expressed in SL2 cells close to the plasma membrane and in dots close to the nucleus. These Swip-1-positve dots did not overlap with the Golgi marker syntaxin (Fig. 5C).
We next deleted the EF-hand domain and the proline-rich region (DSwip-1ΔEFPX10xMyc), the coiled-coil domain (DSwip-1ΔCC10xMyc) or the N-terminal proline-rich region only (DSwip-1ΔPX10xMyc) (for constructs see Fig. 5A and Table 1) and then tested protein extracts of SL2 cells transfected with these Drosophila Swip-1 variants under the control of the UAS/GAL4 system (Brand and Perrimon, 1993) on western blots with anti-Myc antibody. Constructs were expressed at comparable levels and were of the expected size (Fig. 5B). Additionally, we probed these western blots with the antibody against Drosophila Swip-1. As expected, we detected the full-length protein and DSwip-1ΔEFPX10xMyc, but not DSwip-1ΔCC10xMyc, which lacks the antigen site used for immunisation of rabbits (supplementary material Fig. S2). Thus, these findings confirm the specificity of our anti-Drosophila Swip-1 antibody.
To gain more insight into the role of the individual Drosophila Swip-1 protein domains in regard to the subcellular localisation of DSwip-1, we analysed the subcellular distribution of the protein variants in the transfected SL2 cells (Table 1). Interestingly, our immunofluorescence studies revealed that although the full-length DSwip-1-fl10xMyc protein localised to the plasma membrane (Fig. 5D), deletion of the coiled-coil domain (DSwip-1ΔCC10xMyc) led to localisation solely in the cytoplasm (Fig. 5E). In double-labelling studies with the Rab5 antibody (marker for early endosomes), we excluded the possibility that Drosophila Swip-1 colocalises with these endosomes. Despite the loss of the PX and EF-hand domains, the truncated DSwip-1ΔEFPX10xMyc protein localised to the plasma membrane (Fig. 5F). In addition, deletion of only the proline-rich (PX) region did not change the subcellular localisation (data not shown). We therefore conclude that the coiled-coil domain is responsible for localisation to the plasma membrane, at least in SL2 cells.
During myoblast fusion the EF-hand and the coiled-coil domains are required to recruit Drosophila Swip-1 to foci in FCMs
We established transgenic Drosophila lines with DSwip-1-fl10xMyc, DSwip-1ΔCC10xMyc, DSwip-1ΔEFPX10xMyc and DSwip-1ΔPX10xMyc with a modified Gateway-phiC31 system using a landing site at the third chromosome (Bischof et al., 2007). Driving expression with wg-GAL4 in the wingless domains of the epidermis showed that the full-length, as well as the truncated proteins, were synthesised in the embryo (Table 1).
We next asked whether the full-length protein and the truncated versions localised in FCMs similarly to the endogenous protein. To answer this question, we drove expression ofDSwip-1-fl10xMyc, DSwip-1ΔCC10xMyc, DSwip-1ΔEFPX10xMyc and DSwip-1ΔPX10xMyc selectively in FCMs with the sns4,5-GAL4 driver line (Stute et al., 2006) and analysed these embryos with respect to subcellular localisation during myoblast fusion. The FCM-specific expression through the sns4,5-GAL4 driver line was analysed in parallel with standardised conditions. The DSwip-1-fl10xMyc protein localised to characteristic foci on the side of the FCMs when they attached to the FC/growing myotubes (Fig. 6A, arrowheads), whereas DSwip-1-fl10xMyc was not detectable in FCMs without contact to a growing myotube. This reflects the subcellular protein expression pattern known for the endogenous Drosophila Swip-1 protein (Fig. 3A–C) and is evidence that the 10xMyc tag does not disturb subcellular localisation. However, the truncated proteins lacking the coiled-coil (DSwip-1ΔCC10xMyc; Fig. 6B) or proline-rich region and EF-hand domain (DSwip-1ΔEFPX10xMyc; Fig. 6C) were expressed in the cytoplasm, but do not concentrate into foci, whereas DSwip-1ΔPX10xMyc was localised in the characteristic foci (supplementary material Fig. S4).
We next addressed the question whether the expression of the full-length or truncated versions of Drosophila Swip-1 disturbs myoblast fusion. To answer this question, we drove the Drosophila Swip-1 variants with DMef-GAL4 and twist driver lines. The distribution of β3-tubulin revealed that DSwip-1-fl10xMyc, DSwip-1ΔCC10xMyc, DSwip-1ΔEFPX10xMyc and DSwip-1ΔPX10xMyc do not interfere with muscle development (Table 1). We therefore conclude that the truncated versions of Drosophila Swip-1 variants do not act as a dominant negative. In the cases of DSwip-1ΔEFPX10xMyc and DSwip-1ΔCC10xMyc, this is presumably due to failure to localise to the plasma membrane.
Drosophila Swip-1 foci are in close proximity to transient F-actin foci
The Drosophila Swip-1 foci strongly resemble the known F-actin foci in the FuRMASs (Kesper et al., 2007). Live imaging has shown that these F-actin foci are a transient feature during myoblast fusion (Richardson et al., 2008a; Richardson et al., 2007). We addressed the question whether Drosophila Swip-1 foci overlap with F-actin foci. Therefore, we combined the visualisation of F-actin by the Phalloidin-TRITC marker with Drosophila Swip-1 distribution and found that Drosophila Swip-1 foci and F-actin foci were at different positions from segment to segment (supplementary material Fig. S5). Moreover, we observed that the Drosophila Swip-1 foci sometimes overlapped with the F-actin foci and sometimes did not (Fig. 7A–C). We therefore conclude that both molecules are localised in close proximity in the tip of the filopodia in FCMs as far as the resolution of fluorescence microscopy allows (Fig. 7), and that there might be a functional connection between Drosophila Swip-1 and F-actin (see the Discussion).
Drosophila Swip-1 recruitment to the FuRMASs on the site of the FCMs depends on successful cell adhesion
We analysed whether successful cell adhesion due to the chemoattracting function of Duf/Kirre and Rst/IrreC and their interaction with Sns is essential for Drosophila Swip-1 foci formation (Galetta et al., 2004; Ruiz-Gomez et al., 2000; Strünkelnberg et al., 2001). For this purpose, we analysed embryos homozygous for Df(1)w67k30/z(1)v(1), which do not contain both Duf/Kirre and Rst/IrreC. Here, Drosophila Swip-1 was expressed only in the cytoplasm of FCMs; however, in the absence of cell adhesion, no characteristic Drosophila Swip-1 foci were formed (data not shown).
Because Sns is the most relevant cell adhesion molecule in FCMs, we analysed the localisation of Drosophila Swip-1 in sns20-2 (formerly rost20-2) (Paululat et al., 1995) loss-of-function mutants. Without Sns, Drosophila Swip-1 was expressed (visible as diffuse weak staining), but it rarely accumulated in the characteristic foci (Fig. 8B) typical of the wild type (Fig. 8A). Hibris is known to be able to replace Sns, and both proteins act in functional redundancy to some degree (Shelton et al., 2009). Thus, we suspect that this functional redundancy is responsible for the appearance of the few observed Drosophila Swip-1 foci. This would be in agreement with the rare fusion events observed in sns mutants. Taken together, we conclude that Drosophila Swip-1 accumulation in foci depends on successful cell adhesion.
We next performed anti-Drosophila Swip-1 staining on sns20-5 mutant embryos (Paululat et al., 1995). sns20-5 encodes a truncated protein of 1163 amino acids that contains the extracellular domain and the transmembrane domain but lacks the majority of the intracellular domain (Bour et al., 2000). In sns20-5 mutants, Drosophila Swip-1 did not accumulate into characteristic foci (Fig. 8C). Thus, we conclude that the accumulation of Drosophila Swip-1 in foci depends on Sns-mediated cell adhesion and probably signalling.
Drosophila Swip-1 foci accumulate in singles bar mutants, but not in blow and kette mutant embryos
To obtain further insight into the role of Drosophila Swip-1 during myoblast fusion, we analysed Drosophila Swip-1 distribution in fusion mutants. We cannot present quantification of the Drosophila Swip-1 foci accumulation in wild-type and mutants because this would require knowing the situation in the wild type at a given time and in an individual segment. Moreover, Drosophila Swip-1 (Fig. 9A,C,F) and F-actin (supplementary material Fig. S5) foci vary highly in number and position during fusion. This is in contrast to muscle identity conferring factors such as nuclear markers that allow fusion efficiency to be monitored (Richardson et al., 2007). We therefore analysed the distribution of Drosophila Swip-1 in a number of fusion mutants with the aim of correlating its expression pattern to the known status of myoblast fusion at the ultrastructural level (for mutant fly strains used see Table 2). In this respect, it is of prime interest to analyse mutants in which cell adhesion was successful, but where the opposing membranes remain intact. This is the case for singles bar mutants (sing22) where prefusion complexes accumulate as well as for blown fuse (blow2) and kette (ketteJ4-48) mutants where the prefusion complex is dissolved but plasma membranes remain intact (Doberstein et al., 1997; Estrada et al., 2007). Of further interest are mutants that stop myoblast fusion when the membranes between FCMs and the FC or growing myotube vesiculate. One example of this is wip30D, which lacks the WASP-interacting protein (Massarwa et al., 2007). In our study, we analysed arp3 (arp3schwächling) (Berger et al., 2008) mutant embryos, which stop myogenesis after membrane vesiculation, removal of the membrane remnants and the formation of cytoplasmic continuity (Önel and Renkawitz-Pohl, 2009; Önel et al., 2011; Önel and Renkawitz-Pohl, 2009). Furthermore, the expression of Drosophila Swip-1 in both sing22 and blow2 mutant embryos appears to be different compared with that in the wild type, whereas in the remaining mutants tested, Drosophila Swip-1 expression resembled that in the wild type (Fig. 9). For example, in contrast to wild-type embryos (Fig. 9A,C,F), in sing22 mutants, many FCMs contacted an individual FC or growing myotube at the same time (Fig. 9B,D,G), and all these FCMs contained Drosophila Swip-1 foci (Fig. 9D,G). In blow2 mutant embryos, fewer Drosophila Swip-1 foci accumulated compared with sing22 mutant embryos, and they appeared smaller and more diffuse (Fig. 9E,H). In ketteJ4-48 mutant embryos, Drosophila Swip-1 distribution resembled that in the wild type (Fig. 9I). Moreover, wip30D and arp3schwächling both lacked a protein that regulates F-actin branching, but also had a Drosophila Swip-1 distribution that resembled that in the wild type (Fig. 9J,K). Taken together, these data clearly show that Drosophila Swip-1 only accumulated significantly in sing22 mutants, and therefore, we hypothesise that Drosophila Swip-1 participates in regulating vesicle exocytosis from the prefusion complex.
Drosophila Swip-1 is the Drosophila homologue of human and murine swiprosins
The CG10641-encoded protein Drosophila Swip-1 is homologous to mammalian swiprosins, EFHD1 and EFHD2. Drosophila Swip-1 and mammalian EFHD1 and EFHD2 share two EF-hand domains and one coiled-coil domain. The mouse proteins have been described to localise to lipid rafts in B-cells and to be involved in different processes such as swiprosin-dependent calcium influx and efflux in B-cells of the adaptive immune system (for a review, see Dütting et al., 2011). Additionally, cytokine secretion in human mast cells is thought to be guided by swip-1/EFHD2 (Ramesh et al., 2009). Hemocytes, macrophage-like cells of the invertebrate innate immune system, are known to transcribe CG10641 (Estrada et al., 2006; Kroczek et al., 2010) and our study found a polar distribution of Drosophila Swip-1 in the macrophage-like hemocytes of the embryo. Taken together, these findings suggest a conserved function of swiprosin in the innate immune system of invertebrates and vertebrates.
In Drosophila SL2 cells, the coiled-coil domain of Drosophila Swip-1 is required for localisation to the plasma membrane whereas in embryos, the EF-hand domains are also necessary
Drosophila Swip-1 is transcribed in SL2 cells (Henikoff et al., 2009) – cells that have a hemocyte-like character (Armknecht et al., 2005). In this study, we show that Drosophila Swip-1 is associated with the plasma membrane in SL2 cells and that this localisation depends on the coiled-coil domain. We conclude that the behaviour of Drosophila Swip-1 in SL2 cells, where only loss of the coiled-coil domain leads to failure to localise to the plasma membrane, is different from the situation in the embryo, where the EF-hand domains are also essential for Drosophila Swip-1 localisation in FCMs. These findings strongly indicate that Drosophila Swip-1 is under a different control mechanism for membrane localisation in hemocytes compared with FCMs.
Myoblast fusion is not affected in Drosophila Swip-1 mutants, implying functional redundancy
Surprisingly, myoblast fusion takes place in Drosophila Swip-1 mutants implying functional redundancy with – most likely – another calcium-binding protein. Interestingly, functional redundancy is known for proteins involved in vesicle exocytosis in the nervous system of mice. Here, synaptophysin, a major synaptic vesicle protein, contains a MARVEL domain and is proposed to have a role in SNARE assembly into the fusion core complex, fusion pore formation and neurotransmitter release (Hübner et al., 2002). Syp is also known to be essential for Ca2+-induced neurotransmitter release in vitro. However, single-gene knockout of Syp in mice revealed no obvious phenotype (McMahon et al., 1996; McMahon and Monroe, 1996), despite the many proposed roles for this protein (Valtorta et al., 2004). This finding is a strong indicator that other proteins are compensating for the role of synaptophysin – a scenario that we also propose is the case for Drosophila Swip-1 during myoblast fusion.
Examples of functional redundancy for myogenesis in Drosophila are already known. For example, the Ig domain cell adhesion molecules Duf/Kirre and Rst/IrreC (Strünkelnberg et al., 2001), as well as the small GTPases Rac1 and Rac2 (Hakeda-Suzuki et al., 2002; Ng et al., 2002) can completely replace each other. The maternal contribution of the actin regulator WASP can completely rescue a zygotic loss-of-function mutant (Schäfer et al., 2007). Moreover, there is also partial functional redundancy between Sns and Hibris (Hakeda-Suzuki et al., 2002; Ng et al., 2002; Shelton et al., 2009). Because over 60 cytosolic calcium-binding proteins have been predicted for Drosophila (Chintapalli et al., 2007), it is a major challenge to identify the exact protein(s) that can act in functional redundancy to Drosophila Swip-1.
Drosophila Swip-1 accumulates in foci in FCMs during myoblast fusion independently of actin regulators Kette, Wip and Arp3
The aim of this study was to better determine the function of Drosophila Swip-1 during myoblast fusion. We found that Drosophila Swip-1 accumulates transiently in the FuRMASs on the site of the FCM and that this accumulation is dependent on the transmembrane proteins Sns, Duf/Kirre and Rst/IrreC. Thus, we propose that Sns-mediated cell adhesion and probably signalling triggers the accumulation of Drosophila Swip-1 into the foci in the FCM.
During myoblast fusion, F-actin accumulates transiently in foci within the FuRMASs (for a review, see Önel et al., 2011). Similarly, Drosophila Swip-1 also appears to accumulate transiently in foci; however, these foci are found exclusively in FCMs. Interestingly, the FCM-specific protein Blow, a putative regulator of F-actin and Kette (Kesper et al., 2007; Richardson et al., 2007; Schröter et al., 2006) also accumulates in a similar manner to Swip-1. In human mast cells, EFHD2/swiprosin-1 was shown to colocalise to the actin-cytoskeleton (Kroczek et al., 2010; Ramesh et al., 2009). This raises the question of whether Drosophila Swip-1 is a regulator of F-actin on the side of the FCMs. Therefore, we analysed further suitable mutants. Mutations in genes that encode actin regulators such as Kette, Wip and Arp3 do not affect Drosophila Swip-1 localisation. By contrast, F-actin foci are not dissolved in kette4-48 and arp3 mutants (Richardson et al., 2007) and in the sltr allele of wip, an abnormally high expression of actin was observed (Kim et al., 2007). Thus, we propose that Drosophila Swip-1 is not directly involved in F-actin regulation in connection with Kette, Wip and Arp3, but rather acts in a parallel process such as regulation of vesicle exocytosis (see below).
Drosophila Swip-1 might act during exocytosis of electron-dense vesicles
Regulated exocytosis is characteristic of neurotransmitter-carrying vesicles (Schweizer and Ryan, 2006; Wojcik and Brose, 2007), secretion by cells of the adaptive immune system (Benado et al., 2009) and exocytosis of acrosomal vesicles during fertilisation (Ackermann et al., 2009). Regulated exocytosis has also been proposed to take place during myoblast fusion in Drosophila (Doberstein et al., 1997; Estrada et al., 2007). Moreover, Vega and colleagues (Vega et al., 2008) showed that the highly conserved mammalian swiprosin-1/EFHD2 exhibits calcium-binding activity. Because swiprosins are known to be calcium-binding proteins, and because calcium has often been described to trigger exocytosis (Barclay et al., 2005; Burgoyne and Morgan, 2003), we asked whether Drosophila Swip-1 might be a regulator of exocytosis of the vesicles of the prefusion complex and/or during plasma membrane fusion.
To address this question, we correlated Drosophila Swip-1 distribution in fusion mutants to the known arrest of myoblast fusion at the ultrastructural level (Fig. 10, Table 2). From ultrastructural analyses, it is known that after recognition and adhesion between FCM and the growing myotube, electron-dense vesicles accumulate at opposing membranes and participate in the formation of the prefusion complex (Fig. 10). After resolving this complex, electron-dense plaques are established. In the following steps, these plaques dissolve, the opposing membranes become vesiculated and the fusion pore expands to integrate the FCM into the growing myotube to build the mature muscle (Önel and Renkawitz-Pohl, 2009; Önel et al., 2011).
In this study, we have shown that Drosophila Swip-1 is transiently expressed during myoblast fusion and becomes concentrated in foci at FCMs – a process that is dependent on successful cell adhesion as well on Sns signalling. Interestingly, we also found that Drosophila Swip-1 accumulates in sing22 mutants. sing22 mutants are known to arrest fusion after establishing the prefusion complexes, and show no resolution of these complexes (Estrada et al., 2007). Therefore, we suggest that the calcium-binding protein Drosophila Swip-1 might directly or indirectly trigger exocytosis of these electron-dense vesicles (Fig. 10). We further propose cooperation between Sing and Drosophila Swip-1 in a regulatory cascade.
Estrada and colleagues (Estrada et al., 2007) hypothesised that Sing might mediate fusion of the electron-dense vesicles with the plasma membrane because Sing contains a MARVAL domain (Estrada et al., 2007). These domains are implicated in vesicle trafficking and cholesterol-rich membrane opposition events (Sánchez-Pulido et al., 2002). Our analyses of blow2 and ketteJ4-48 mutants has added substantial weight to this hypothesis, because: (1) the prefusion complexes are dissolved in these mutants; (2) the plasma membranes are not vesiculated; and (3) Drosophila Swip-1 foci are not accumulated above wild-type level (Fig. 10).
We also asked whether Drosophila Swip-1 might be needed in further fusion-relevant steps such as vesiculation of the plasma membranes. Our analyses of blow2 and ketteJ4-48 mutants revealed that failure of plasma membrane merging does not lead to an accumulation of Drosophila Swip-1 foci. Thus, we propose that Drosophila Swip-1 is specific for prefusion complex resolution (Fig. 10) and hypothesise that calcium-binding proteins such as Drosophila Swip-1 are regulators participating in exocytosis of vesicles of the prefusion complex.
Materials and Methods
Fly stocks and genetics
Flies were grown on standard medium at 25°C. Balancers and chromosome markers are as described in FlyBase (Tweedie et al., 2009) unless otherwise specified. Stocks were grown under standard conditions. w1118 (BL6326) were taken as a wild-type strain. If/CyOhglacZ strain (kindly provided by Markus Affolter, Basel, Switzerland) was used to rebalance the stocks with a blue balancer.
The following strains were obtained from the Bloomington Drosophila Stock Center (Bloomington, IN); Df(2)Exel6042, Df(2)ED1202 and Df(2L)Exel8039. The following fusion mutants were used: sns20-2/CyOhglacZ (Paululat et al., 1995), sns20-5/CyOhglacZ (Paululat et al., 1995), blow2/CyOhglacZ (Doberstein et al., 1997), sing22/Fm7, twiGAL4-UAS-2EGFP (kindly provided by Alan Michelson, National Heart, Lung and Blood Institute, Bethesda, MD); ketteJ4-48/Sb, TDlz (Hummel et al., 2000; Schröter et al., 2004), wip30D (kindly provided by Eyal D. Schejter, Weizmann Institute, Rehovot, Israel) and arp3Schwächling/SbTDlz (Berger et al., 2008).
We used meiotic recombination in females to generate flies that carry Df(2L)Exel6042 and camn339 or Df(2L)Exel6042 and CG31650KG09054 together on the second chromosome, respectively. For generating the double mutants Df(2L)Exel6042; tpnC73FMB03946, we used the double blue balancer CyOhglacZ/Sp; Tm2/Sb, TDlz (laboratory of Renate Renkawitz-Pohl, Philipps-Universitaet Marburg, Marburg, Germany). Except for the double blue balancer strain, flies were obtained from the Bloomington Drosophila Stock Center. As mesoderm-specific driver lines, we used DMef2-GAL4 (Bloomington Drosophila Stock Center), twist-GAL4 (TG-X and SG24, kindly provided by Alan Michelson) and sns4,5-GAL4 (Stute et al., 2006). To determine cell-type specificity, we used the enhancer trap line rp298-lacZ (Nose et al., 1992).
For generating fly stains, which carried a construct to silence Drosophila Swip-1 by RNAi, we crossed different fly lines with the mesoderm-specific driver line DMef2-GAL4 (BL25756, Bloomington Drosophila Stock Center, USA), which drives GAL4 expression specifically in the somatic mesoderm and expresses Dicer additionally. For Drosophila Swip-1-specific RNAi induction we used fly lines named v31307, v31308 and BL31585, which knock down Drosophila Swip-1. As a positive control we used v10177, a fly strain, in which blow is silenced (Dietzl et al., 2007). Crossings were done on 30°C, embryos were fixated and immunohistochemistry with antibodies against β3-tubulin and Drosophila Swip-1 was done as described below.
Immunostaining and Phalloidin labelling of Drosophila embryos
For immunohistochemistry, embryos were collected from juice agar plates, rinsed with TNX (0.7% NaCl and 0.01% Triton X-100) and dechorionised with 50% Klorix. After fixation for 15 minutes in 4% F-PBS (4% formaldehyde with PBS: 0.13 mM NaCl, 7 mM Na2HPO4, 3 mM NaH2PO4), embryos were devitellinised by shaking in 1:1 methanol and heptane. After rehydration in PBT (PBS, 0.5% Tween-20), embryos were incubated overnight at 4°C with primary antibodies in PBT at the following concentrations: 1:4000 anti-DSwip-1 (raised in rabbits; Pineda Antibody Service, Berlin), 1:10,000 anti-β3-tubulin (rabbit) (Leiss et al., 1988), 1:1000 anti-Myc (mouse, Millipore, Billerica, MA), 1:25 CF.6G11 (βPS integrin, mouse, Hybridoma Bank, Iowa City, IA); 1:1000 anti-DMef2 (rabbit, kindly provided by Hanh T. Nguyen, University of Erlangen, Erlangen, Germany); and 1:5000 anti-β-galactosidase (mouse, Promega, Heidelberg). Embryos were then incubated for 1 hour in blocking solution [2% goat serum (Vector Laboratories, Burlingame, CA) in PBT] and afterwards with biotinylated secondary antibody (anti-rabbit 1:500; anti-mouse 1:500; anti-guinea pig 1:500; Vector Laboratories) respectively with Cy2- and Cy3-conjugated secondary antibodies (anti-rabbit, 1:200; anti-mouse, 1:200; anti-guinea pig, 1:200; Dianova, Hamburg) for 2 hours at room temperature (RT). After amplifying the reaction (not for fluorescent immunohistochemistry) using Vectastain Elite AMC Standard Kit (Vector Laboratories), staining was carried out with 10 μl diaminobenzidine, 5 μl 20% H2O2 and 1 μl 10% NiCl2 in 600 μl PBT. Numerous embryos were mounted in Epon and examined with a Zeiss Axiophot (Zeiss, Germany). For fluorescence immunohistochemistry, embryos were incubated with Cy2- and Cy3-conjugated secondary antibodies, washed in PBT several times and mounted in Fluoromount G™ (Southern Biotech, Birmingham, AL) and examined under Zeiss Axioplan 2 imaging with ApoTom (Zeiss, Germany).
For double labelling with Phalloidin-TRITC, the embryos were treated as before until devitellinisation. Then, we devitellinised the embryos in PBT with a sharp glass needle. Embryos were incubated with the first antibody and Phalloidin–TRITC [anti-DSwip-1, 1:2000; anti-β-galactosidase, 1:2500; Phalloidin–TRITC, 1:60 (1 μg/μl; Sigma, Steinheim)] overnight at 4°C. The primary antibodies were detected using the corresponding Cy2- and Cy3-conjugated secondary antibodies (1:200, Dianova, Hamburg). Embryos were mounted in Fluoromount G.
Whole-mount in situ hybridisation was performed as described previously (Tautz and Pfeifle, 1989). DIG-labelled RNA antisense probes were synthesised by in vitro transcription using a RNA-DIG labelling kit (Roche Diagnostics, Mannheim). We used the SD04693 cDNA clone (CG10641 in pOT2; obtained from Drosophila Genomics Resource Center, Bloomington, IN); as a template for transcribing the RNA probe using SP6 polymerase (Roche Diagnostics, Mannheim).
Antibody against Drosophila Swip-1 was generated by the Pineda Antibody Service (Berlin, Germany) against the peptide N-CEERAQPRQQFQQRAAIF-C in rabbits. The serum was affinity purified.
Constructs for cell culture experiments and generation of transgenic flies
Deletion constructs (DSwip-1ΔCC10xMyc, DSwip-1ΔEFPX10xMyc and DSwip-1ΔPX10xMyc) and full-length coding region of Drosophila Swip-1 (DSwip-1-fl10xMyc) were amplified by PCR using the following primers:
(1) DSwip-1ΔCC10xMyc constructs: ΔCCfwd: 5′CACCATGTCCGTTTCCTCGAACGCCTCATCC-3′ and ΔCCrev: 5′GAAGAACTTCTTGGCTCCGCT-3′; (2) DSwip-1DEFPX10xMyc constructs: ΔEFPXfwd: 5′-CACCATGAGCGGAGCCAAGAACTTCTTC-3′ and ΔEFPXrev: 5′ CTCGAAGATCGCTGCCCGCTG-3′; (3) DSwip-1DPX10xMyc constructs: ΔPXfwd: 5′CACCATGAGCACCACCAACACGGACA and DEFPXrev; (4) DSwip-1-full-length constructs: ΔCCfwd and ΔEFPXrev.
PCR products were cloned into the pENTR/D-TOPO vector (Invitrogen, Karlsruhe). With recombination catalysed by the LR Clonase II plus enzyme mix (Invitrogen, Karlsruhe), the constructs were cloned into modified Gateway vectors (pUAST-attB-rfa-10xMyc kindly provided by Sven Bogdan, Universität Münster, Münster, Germany). These vectors were injected into embryos of fly strain BL24749 (Bloomington Drosophila Stock Center), where the integration of the construct takes place on the third chromosome. Table 1 summarises constructs and corresponding transgenic Drosophila lines.
SL2 cell culture, transfection and antibody staining
SL2 cells were cultured in Schneider's Drosophila Medium (Invitrogen, Karlsruhe, Germany) with 1% penicillin-streptomycin (Invitrogen) and 10% fetal calf serum (Sigma) at 25°C. For each transfection, we split cells on multititer plates (3×105 cells/ml per well) and after 24 hours, transfected with 0.5 μg DNA, 100 μl FCS-free medium, 3 μl Transfectin (Bio-Rad, Munich), act-GAL4 and 400 μl medium for 2 days.
For antibody staining, we incubated glass dishes with concanavalin A (Sigma) for 40 minutes at room temperature (RT). After incubation with the transfected cells for 2 hours, we fixed them with 4% F-PBS for 15 minutes at RT. Cells were then washed twice with PBS, before permeabilising with 0.5% Triton X-100 in PBS and washing again. After this, cells were blocked with 3% BSA-PBS for 30 minutes and washed three times in PBS and H2O. Cells were next incubated with the primary antibodies [(anti-DSwip-1, 1:1000; anti-Myc, 1:1000; anti-syntaxin-16, 1:50 (mouse, Abcam, Cambridge, MA), anti-Rab5, 1:200 (rabbit, Abcam)] for 2 hours at RT and washed three times in PBS and H2O. The primary antibodies were detected using the corresponding Cy2- and Cy3-conjugated secondary antibodies (1:100, Dianova, Hamburg, Germany) and by incubating the cells for 1 hour at RT. After washing again, cells were incubated with DAPI (1 μg/μl; Sigma) to visualise nuclei, washed and finally mounted in Fluoromount G.
Preparation of protein extracts and western blot analysis
Cells were transfected as described above (3×105 cells/ml in 24-well plates). Two days after transfection, protein extracts were prepared. Cells were harvested and after centrifugation (3 minutes, 1000 g) dissolved in lysis buffer [1% Triton X-100, 1 mM PMSF, 1 mM DTT, protease inhibitor (Roche Diagnostics) in PBS]. After centrifugation, the supernatant was taken for SDS-PAGE.
For generation of protein extracts of Drosophila embryos, larvae and pupae, we homogenised each tissue and added two volumes of lysis buffer (50 mM Tris-HCl, 125 mM NaCl, 5% glycerol, 0.3% NP4O, 1.5 mM MgCl2, 1 mM Na3VO4, 1 mM EDTA, 1 mM DTT, proteinase inhibitor in PBS). After centrifugation (20 minutes, 15,000 g), supernatant was taken for SDS-PAGE. Larvae were homogenised and additionally precipitated with TCA. After centrifugation (20 minutes, 15,000 g), the pellet was washed in 20% TCA, followed by 5% TCA, 100% ethanol and 1 M Tris-HCl (pH 8). After centrifugation (20 minutes, 15,000 g), supernatant was taken for SDS-PAGE.
SDS-PAGE (15% acrylamide) was carried out as described by (Laemmli, 1970). Proteins were blotted on a nitrocellulose membrane (1 hour, 0.8 mA/cm2). After washing the membranes in TBST (10 mM Tris-HCl, 0.15 M NaCl, 0.1% Tween20) for 10 minutes, blocking in 5% milk powder in TBST for 1 hour and washing again in TBST, extracts were incubated with primary antibody overnight at 4°C (anti-Myc, 1:2000; anti-DSwip-1, 1:1000; anti-actin (rabbit, Cell Signaling Technology, Boston, MA), 1:1000). Membranes were then washed four times in TBST and incubated with peroxidase-coupled corresponding secondary antibody (Dianova, Hamburg, 1:1000) and developed with the help of an ECL Western Blotting Analysis System (GE Healthcare, Munich).
We thank Christiana Stute and Markus Kallin for the initial analysis of Drosophila Swip-1 and Nadja Ring for generating some constructs. We also thank Sven Bogdan (Universität Münster) for providing modified Gateway vector. We would also like to thank Katja Gessner for excellent secretarial assistance, SusanneÖnel for critical reading, Anja Rudolf and Jessica Kuckwa for helping with Phalloidin and anti-DMef2 staining, Ruth Willmott for excellent editing as well as Bettina Bonn for advice with cell culture transfection and Monika Hassel for sharing cell culture facilities. The Hybridoma antibody developed by Danny Brower was obtained from the Developmental Studies Hybridoma Bank developed under the auspices of the NICHD and maintained by The University of Iowa, Department of Biology, Iowa City, IA52242. We received gratefully fly strains from Vienna Drosophila RNAi Center, Austria. This work was supported by the Deutsche Forschungsgemeinschaft Re 628/14-3 and Re 628/15-2 to R.R.-P. and the EU Network of Excellence MYORES.
↵* Present address: Justus-Liebig Universität Giessen, Fachbereich Medizin, Institut für Biochemie, 35392 Giessen, Germany
Supplementary material available online at http://jcs.biologists.org/lookup/suppl/doi:10.1242/jcs.083907/-/DC1
- Accepted May 30, 2011.
- © 2011.