Sphingolipid depletion impairs endocytic traffic and inhibits Wingless signaling

: Sphingolipids are an important part of the plasma membrane and implicated in a multitude of cellular processes. However, little is known about the role of sphingolipids in an epithelial context and their potential influence on the activity of signaling pathways. To shed light on these aspects we analyzed the consequences of changing ceramide levels in vivo in the Drosophila wing disc: an epithelial tissue in which the most fundamental signaling pathways, including the Wnt/Wg signaling pathway, are well characterized. We found that downregulation of Drosophila’s only ceramide synthase gene schlank led to defects in the endosomal trafficking of proteins. One of the affected proteins is the Wnt ligand Wingless (Wg) that accumulated. Unexpectedly, although Wg protein levels were raised, signaling activity of the Wg pathway was impaired. Recent work has spotlighted the central role of the endocytic trafficking in the transduction of the Wnt signal. Our results underscore this and support the view that sphingolipid levels are crucial in orchestrating epithelial endocytic trafficking in vivo. They further demonstrate that ceramide/sphingolipid levels can affect Wnt signaling. ABSTRACT Sphingolipids are an important part of the plasma membrane and implicated in a multitude of cellular processes. However, little is known about the role of sphingolipids in an epithelial context and their potential influence on the activity of signaling pathways. To shed light on these aspects we analyzed the consequences of changing ceramide levels in vivo in the Drosophila wing disc: an epithelial tissue in which the most fundamental signaling pathways, including the Wnt/Wg signaling pathway, are well characterized. We found that downregulation of Drosophila ’s only ceramide synthase gene schlank led to defects in the endosomal trafficking of proteins. One of the affected proteins is the Wnt ligand Wingless (Wg) that accumulated. Unexpectedly, although Wg protein levels were raised, signaling activity of the Wg pathway was impaired. Recent work has spotlighted the central role of the endocytic trafficking in the transduction of the Wnt signal. Our results underscore this and support the view that sphingolipid levels are crucial in orchestrating epithelial endocytic trafficking in vivo . They further demonstrate that ceramide/sphingolipid levels can affect Wnt signaling. compartment. It rather leads to disruption of recycling and degradative endocytic trafficking routes and consequently the accumulation of several secreted and transmembrane proteins in the cells, amongst them Wg. In spite of the higher Wg levels, the activity of the Wg signaling pathway is reduced. This reduction of Wg signaling activity is possibly connected to an early-to-late endosomal trafficking defect and highlights the importance of endocytic trafficking for Wg signal transduction. Importantly, the results reveal a hitherto unknown function for sphingolipids in Wnt signaling.


INTRODUCTION
Sphingolipids influence a variety of developmental processes including cell growth, differentiation and apoptosis. They are synthesized in the ER and Golgi apparatus and are subsequently transported to the plasma membrane, where they reside in the external membrane leaflet (Futerman and Riezman, 2005). Sphingolipids were proposed to form the controversial detergent resistant microdomains, the so-called "lipid rafts", together with cholesterol. These microdomains within the plasma membrane form, as sphingolipids tend to cluster due to lateral interactions; additionally cholesterol intercalates between their mostly saturated fatty acids, giving rise to a more rigid or "liquid-ordered" patch of membrane. Due to the differential membrane composition some lipid-modified and transmembrane proteins might preferentially associate with these lipid microdomains, whereas others would avoid them (Lingwood and Simons, 2010;Simons and Ikonen, 1997). This led to the notion that sphingolipids are involved in protein sorting and trafficking (Le Roy and Wrana, 2005). It also suggested a potential role in regulating signaling activity by providing platforms for signaling complexes (Simons and Toomre, 2000). However, the effect of changing sphingolipid levels in an epithelial context, for example the consequences for signaling pathway activities, is not well understood. Ceramide, the precursor of all complex sphingolipids, is known to be an important second messenger, as well as a modulator of membrane properties. Ceramide production is induced during various stress situations and is thought to be involved not only in apoptotic signaling (Kolesnick and Krönke, 1998), but also in cell growth and differentiation. Due to its physicochemical properties ceramide tends to lateral clustering which, combined with its inverted conical shape, results in negative membrane curvature and inwards budding of membrane (Holopainen et al., 2000). It is still debated if ceramide functions directly as a second messenger or indirectly via its ability to induce changes in the membrane structure (Kolesnick et al., 2000;Van Blitterswijk et al., 2003). This influence on membrane structure further suggests a role of ceramide in endocytosis and intracellular trafficking: when fibroblasts and macrophages are treated with an exogenous sphingomyelinase in the absence of ATP the formation of endocytic vesicles was observed (Zha et al., 1998).
There is growing evidence that endocytosis and endocytic trafficking of the ligandreceptor-complexes are essential for Wnt/Wg signal transduction both in mammalian cells and in Drosophila (Blitzer and Nusse, 2006;Seto and Bellen, 2006). This highly conserved signaling pathway plays a crucial role in a variety of developmental processes and deregulated Wnt signaling is implicated in various diseases, including colon cancer and degenerative diseases (Clevers and Nusse, 2012). In the canonical branch of Wnt/Wg signaling Wnt forms a ternary complex with its receptors Frizzled (Fz) and LRP5/6 (Arrow in Drosophila), which are activated by a series of phosphorylation events. This leads to relocalization of Dishevelled (Dvl) to the plasma membrane and the formation of receptor complexes on oligomerized Dvl clusters, giving rise to signaling platforms called "signalosomes" (Bilic et al., 2007).
The binding of the scaffold protein Axin leads to the recruitment of APC and GSK3ß, which are also components of the ß-catenin destruction complex. As a consequence ßcatenin can no longer be degraded, accumulates and translocates to the nucleus, where it initiates the transcriptional activation of target genes. Several observations highlight the importance of endocytic trafficking in Wnt signaling: first, the Fz receptor stimulates its own endocytosis via Gαo and the early-endosomal GTPase Rab5, and localization in the respective endosomal compartment is reported to influence the balance between canonical and non-canonical signaling branch (Purvanov et al., 2010). Second, Niehrs and colleagues showed that acidification of an intracellular compartment is required for LRP6 receptor phosphorylation and thus Wnt signaling induction Cruciat et al., 2010). Third, it was shown in an elegant study that GSK3ß is sequestered into multivesicular bodies (MVBs) upon Wnt stimulation, so that the enzyme is secluded from its cytosolic substrate ß-catenin and by this the Wnt signal can be transduced (Taelman et al., 2010).
Given the dearth of knowledge about the role of sphingolipids in an epithelial context and their potential influence on the activity of signaling pathways we set out to analyze the consequences of changed ceramide levels, using the Drosophila wing disc as a suitable model for an epithelium. Since most fundamental signaling pathways, including the Wnt/Wg signaling pathway, are well characterized in the wing imaginal disc we hoped to gain insights into the effects of ceramide depletion on these pathways. To this end, we used RNAi-mediated gene knockdown against components of the de novo ceramide synthesis pathway (Acharya and Acharya, 2005). Our study focuses on the effects of depleting ceramide levels by downregulation of schlank, the only ceramide synthase gene in Drosophila. We found that downregulation of schlank does not obviously influence endocytosis and formation of the early-endosomal compartment. It rather leads to disruption of recycling and degradative endocytic trafficking routes and consequently the accumulation of several secreted and transmembrane proteins in the cells, amongst them Wg. In spite of the higher Wg levels, the activity of the Wg signaling pathway is reduced. This reduction of Wg signaling activity is possibly connected to an early-to-late endosomal trafficking defect and highlights the importance of endocytic trafficking for Wg signal transduction. Importantly, the results reveal a hitherto unknown function for sphingolipids in Wnt signaling.

Depletion of schlank leads to growth defects, accumulation of Wg protein and reduced Wg signaling activity
To determine the effects of sphingolipid depletion in an epithelial context, we examined the effect of reducing schlank function (CG3576, Bauer et al 2009). This choice was based on the facts that its product is Drosophila's only ceramide synthase and that its downregulation results in reduced ceramide levels, and on the availability of suitable alleles. Schlank was previously identified in a genome-wide RNAi screen conducted in our lab to uncover growth regulators (G. Reim and K. Basler, unpublished). Consistent with this, schlank depletion by two independent RNAi lines These data indicate that besides the growth phenotypes, reduction of ceramide levels also leads to accumulation of Wg protein in vesicle-like structures. Unexpectedly although there is more Wg ligand, high-level Wg signaling activity seems to be impaired upon schlank depletion.

Wg accumulation is due to impaired ceramide synthase activity and independent of apoptosis
To confirm that the observed phenotypes are a consequence of downregulation of schlank and to exclude potential RNAi mediated off-target effects, we tested the hypomorphic P-element-based allele schlank P365 , in which the P-element is inserted downstream of the first schlank exon. We observed the same Wg accumulation phenotype as in schlank RNAi expressing wing discs (data not shown). Additionally, we generated a presumptive null allele, schlank 331 , via imprecise excision of this P-element, deleting the downstream half of the first exon including a predicted transmembrane domain (Fig. 1A). Again, we observed the cell-autonomous accumulation of Wg protein in schlank 331 clones ( Fig. 2A). Due to the impaired growth upon schlank depletion, we mostly worked in a Minute background to equip the schlank mutant cells with a growth advantage. These experiments show that the observed defects are caused by the depletion of schlank and are not an RNAi artifact.
We next performed rescue experiments using an actin-Gal4 driven UAS-schlank transgene (either native or HA-tagged) and a genomic construct comprising the entire schlank locus. All of these constructs were able to fully rescue either schlank P365 or schlank 331 mutant flies. In a wild-type background they did not cause any phenotype.
Importantly, we also tested the allele H215D, in which the enzymatic activity of Schlank is abrogated (Bauer et al., 2009), and found it unable to rescue our schlank alleles (data not shown). RNAi induced Wg accumulation was diminished by expression of Schlank but not the Schlank mutant H215D (Fig. S2A). We further induced schlank P365 mutant clones and expressed the HA-tagged Schlank protein in the posterior compartment: again the schlank P365 induced Wg accumulation was reverted (Fig. 2B). The rescue capability of Schlank and the inability of the H215D mutant strongly suggest that the Wg accumulation is due to reduced ceramide synthase activity.
To test if the Wg accumulation is merely a consequence of elevated levels of apoptosis, we provoked apoptosis by expressing the pro-apoptotic protein Hid in the posterior compartment. In this case, the apoptosis induction ( Taken together these experiments indicate that the Wg accumulation phenotype is caused by reduced ceramide synthase activity and is independent of apoptosis.

Several secreted and transmembrane proteins accumulate, but their respective signaling pathway activities are not changed
To test if the reduction of ceramide levels affects wing imaginal disc cell morphology, we assessed the overall integrity of the wing disc epithelium by electron microscope analysis of wing discs expressing schlank RNAi1 in the posterior compartment. For visualization of these cells we additionally expressed a membrane-bound Horseradish Peroxidase that could be detected via DAB staining. We were able to identify schlank mutant tissue using this method and found that the apical-basal cell morphology and the tissue integrity of schlank depleted cells was not disrupted (Fig. S2D).
Next we asked if Wg was the only protein accumulating as a consequence of schlank loss/reduction of function. We tested several other secreted and transmembrane proteins for accumulation upon schlank depletion. Antibody stainings against the Wg receptor Fz2 and its co-receptor Arrow, as well as Hedgehog (Hh), its receptor In summary, we found that schlank depletion does not grossly affect the overall integrity of the wing disc epithelium. However it does cause the accumulation of a subset of secreted and transmembrane proteins. The extent of the accumulation varied, but with the exception of Wg there appeared to be no detectable changes in the activity of the signaling pathways. Therefore we decided to concentrate our attention on the Wg signaling pathway.

Wg accumulation is not due to more production and secretion
Two possible scenarios could account for the observed Wg accumulation: first, the accumulating Wg could be due to more production and secretion of the protein.
Second, there could be less degradation of Wg protein. The cell autonomy of the observed phenotype already hinted to less degradation as a secretion defect would be expected to be non-autonomous. To test this in more detail, we monitored wg transcription using a lacZ-reporter and performed extracellular Wg stainings, which constitute a very sensitive readout for Wg secretion. Neither the transcription of wg ( Fig. 4A) nor the extracellular level of the Wg protein was changed upon schlank depletion ( Fig. 4B,C), demonstrating that the observed accumulation is not caused by more production and secretion of the Wg protein. Furthermore we did not detect changes in the uptake of fluorescently labeled Dextran in schlank depleted cells, suggesting that the initial step of endocytosis is not disturbed (data not shown). Thus the Wg accumulation is probably due to impaired degradation after the protein was taken up.
Consistent with the hypothesis that upon schlank depletion there is a defect in endosomal trafficking we observed that the typical punctate localization of Wntless in the Golgi apparatus in Wg-producing cells was disrupted (Fig. 4D). Wg secretion is dependent on the Wntless/Evi (Wls) protein, which promotes Wg release and is afterwards recycled to the Golgi apparatus in a retromer-dependent manner (Port et al., 2008). Wls was apparently not as efficiently recycled back to the Golgi apparatus upon schlank depletion, whereas Wg secretion seemed unchanged. Therefore, we speculate that the amount of Wls in the Golgi apparatus is still sufficient to ensure proper Wg secretion, but cannot rule out secondary effects as, e.g., a change in the potency of the secreted ligand to induce Wg signaling (Franch-Marro et al., 2008).

Schlank depletion leads to enlarged Rab4/7/11 positive endosomal structures
To see where the endosomal trafficking is defective, we made use of various Rab GTPases as endosomal markers. Rab5 GTPase marks early endosomal structures, whereas Rab4 and Rab11 GTPases label the fast and slow recycling endosomes, respectively. Rab7 is used to label the late endosomal trafficking route leading to lysosomal degradation. Constructs consisting of the tubulin promoter driven Rab GTPase fused with an YFP-tag were employed and schlank clones were induced in this background. Whilst Rab5 positive endosomal compartments were not detectably altered by schlank depletion (Fig. 5A), Rab4/7/11 positive endosomes were massively enlarged ( Fig. 5B,C,D). To check if other intracellular compartments were affected, we analyzed the status of Golgi apparatus and Endoplasmic Reticulum (ER) by testing the Golgi component ß-Cop and ER-resident KDEL-carrying proteins. We could not detect any differences in the antibody stainings against these proteins (Fig.   S4A,B).
These experiments show that the processes of endocytosis or early endosome formation are not so sensitive to alteration in sphingolipid levels, in contrast to the more downstream intracellular trafficking steps.

All endosomal compartments exhibit increased Wg concentration
Given that we see accumulation of ligand (Wg) and receptor (Arr, Fz2) we were surprised to find that Wg signaling was reduced. Recent work on the Wg signaling transduction mechanism offers two possible explanations for this observation in the context of endocytic trafficking: first, it was shown that Wg has to enter the Rab5 positive early endosome for full strength signaling (Purvanov et al., 2010). It is possible that, due to defects in endocytic sorting and traveling, there is in absolute terms less Wg in Rab5 positive compartments, which is masked by Wg accumulation in other endosomal compartments. Alternatively, the underlying defect could be occurring in the more downstream steps of the degradative route such as impaired lysosomal acidification, which is essential for Wg signal transduction Cruciat et al., 2010), or due to impaired formation of multivesicular bodies, which are reportedly necessary for GSK3ß sequestration in vertebrate systems (Taelman et al., 2010).
To test if less Wg in Rab5 early endosomes could account for the Wg signaling defect, we quantified the Wg protein in Rab5 early endosomes, as well as in Rab4 recycling and in Rab7 late endosomes. As Wg endocytosis is believed to be receptormediated (Dubois et al., 2001) and by staining for Frizzled receptor we would not be able to distinguish between ligand-bound endocytosed receptor and "empty" recycling receptor, we used Wg staining to detect endocytosed ligand-receptor complex. As markers of the endosomal compartments, we again used the YFP-Rab constructs, now in combination with schlank RNAi driven in the posterior compartment. In this setup, we first defined the endosome volumes in control and schlank depleted compartments in the same disc to test how much the different Rab-positive compartments are enlarged upon schlank depletion. Next, we quantified the amount of Wg in the mutant and the neighboring control compartment. This yielded quantitative information about how much Wg is accumulating overall in the schlank depleted situation. Third, we measured the amount of Wg that could be found inside a particular type of endosome.
To do so, we analyzed the ratio of Wg signal in mutant compared to control endosomes and normalized for the endosomal volume.
Confirming our microscopic analysis of the respective Rab-positive endosomal compartments, we did not see an enlargement of the Rab5-marked endosomal compartments. In contrast, we observed a twofold enlargement of Rab4-positive endosomal structures and an even more pronounced enlargement of Rab7 late endosomal structures (Fig. 6A). Overall we found a twofold accumulation of Wg protein upon schlank depletion. Rab4-and Rab5-positive compartments showed a twofold increase in the concentration of Wg. In Rab7-positive endosomes, more than fourfold more Wg was measured upon schlank reduction (Fig. 6B).
These results indicate that Wg protein levels in Rab5-positive endosomes are increased, rather than decreased, upon schlank depletion (also when measured in absolute amounts and not corrected for the slightly decreased Rab5 endosomal volume), ruling out the notion that less Wg protein in Rab5 early endosomes is the cause for the reduced signaling activity.
Based on current models of Wnt/Wg signaling, a defect more downstream in the early-to-late endosomal pathway could result in reduced Wg signal transduction. The ligand-receptor complexes are thought to be sorted in a Hrs-dependent manner from Rab5-positive early endosome into multivesicular endosomes (Taelman et al., 2010).
Therefore we analyzed the localization of Hrs and Vps16, both markers of multivesicular bodies, in our schlank reduced background. The localization of neither Hrs nor Vps16 was changed upon schlank depletion (Fig. 6C, S4C). In contrast, Lamp1-GFP and LysoTracker, markers of the lysosomal compartment were enlarged in the schlank depleted situation (Fig. 6D, S4D).
Taken together, the observed protein accumulation and the defects in the formation of Rab7 endosomes and lysosomes suggest that ceramide depletion impedes Wg signaling by disrupting the early-to-late endosomal route.

Sphingolipids and endosomal trafficking
In this study, we reduced sphingolipid levels in the Drosophila wing imaginal discs by genetic means to study the effects of changed ceramide availability, with a focus on the activity of signaling pathways in an epithelial context. Not unexpected given the coordination of energy metabolism and growth with lipid metabolism, we observed reduced cell size and cell number due to apoptosis upon depletion of the ceramide synthase gene schlank. When we examined the Drosophila wing imaginal disc we found that overall tissue integrity was not affected. However, we did observe an accumulation of various secreted and transmembrane proteins.
Analysis of the endocytic compartments revealed that the slow and fast recycling endocytic routes, as well as the late endosomal route, were disrupted. In contrast, exocytosis (secretion of Wg and Dpp), the initial steps of endocytosis and the formation of Rab5-positive early endosomal compartments were not affected.
Ceramide is able to induce membrane curvature in artificial membranes (Holopainen et al., 2000). Additionally, changes in membrane morphology and fusion were reported in lipids extracted from mice defective in long chain ceramide synthesis (Silva et al., 2012). Thus, via physical changes in membrane budding/fusion sphingolipid composition could influence endocytic trafficking.
Consistent with a problem in the early-to-late endosomal route leading to lysosomal protein degradation we observed the accumulation of various proteins upon ceramide synthase depletion. These included Hedgehog (Hh), its receptor Patched and the surface receptor Notch (Fig. 3). We saw no effect on the Hh or Notch signaling pathways, whereas earlier studies in Drosophila implicated glycosphingolipids in various steps in development and different signal pathways (Kraut, 2011;Singh et al., 2011). For example, in an impressive study Hamel and colleagues show that glycosphingolipids modulate the signaling activity of Notch ligands (Hamel et al., 2010). The fact that we do not find reduced signaling activity apart from Wg signaling suggests that sufficient glycosphingolipids are still produced, either by the remaining de novo synthesized ceramide or via salvage pathways. Taking this into account, our results show that the Wnt/Wg signaling pathway is particularly sensitive to the reduction in ceramide levels.

Sphingolipids and Wingless signaling
One of the many proteins that accumulated was the morphogen Wg. Although there was more Wg ligand in the system, the pathway activity was reduced upon schlank depletion. We showed that this effect is dependent on ceramide synthase activity and is not a secondary effect of apoptosis. Furthermore, we demonstrated that the accumulation of Wg is not due to increased expression or secretion and happens in the late endosomal route and to a lesser extent in the recycling endosomes.
The Wg signaling defect that we see could result from impaired protein sorting.
Sphingolipids are, beside their implied role in membrane morphology, thought to be involved in protein sorting (Le Roy and Wrana, 2005;Lippincott-Schwartz and Phair, 2010;Simons and Ikonen, 1997). The partitioning or differential sorting of proteins into lipid microdomains is implicated in promoting the assembly of signaling complexes. In this regard, it is interesting that the Arrow homolog LRP6 is associated with membrane domains composed of cholesterol and sphingolipids (Yamamoto et al., 2006). A simple scenario is that this association is important for signaling by promoting receptor clustering. Interfering with or abolishing this clustering could be envisaged to result in reduced signaling efficiency. However, given the growing body of evidence that connects endosomal trafficking and Wnt/Wg signaling, a parsimonious explanation is that a disruption of endosomal routes is responsible for the observed effects on the Wg pathway.
The first link between Wg signaling activity and internalization was made in 2006 (Blitzer and Nusse, 2006;Seto and Bellen, 2006). Later in an elegant study by the Niehrs lab it was shown that in the mammalian system Wnts induce clustering of receptors and Dvl in endocytosed complexes called signalosomes, which facilitates pathway induction (Bilic et al., 2007). More recently a series of studies has highlighted various aspects of the relationship between Wnt signaling and endosomal trafficking: it could be shown that localization of Wg in Rab5-positive endosomal compartments is required for full-strength signal induction (Purvanov et al., 2010).
Furthermore, acidification occurring along the early-to-late endosomal route was demonstrated to be important for Wnt signaling (Cruciat et al., 2010). Finally, it was reported that the GSK3ß kinase has to be sequestered into multivesicular endosomes for proper Wnt signaling induction (Taelman et al., 2010).
Given that reduced schlank expression leads to defects in endosomal trafficking and Wnt/Wg signaling is critically linked to endosomal transport it is tempting to suggest that reduced Wg signaling we observed upon schlank depletion is caused by the impairment of endocytic trafficking. To further clarify the connection of Wnt signaling and ceramide levels, a cell-biological system has to be developed that provides the necessary resolution to dissect endocytic pathways during Wnt signaling upon ceramide depletion.
In summary, we present evidence that reducing sphingolipid levels in an epithelial context leads to the impairment of distinct endocytic routes. Furthermore, we show that the Wg signaling pathway is particularly sensitive to changes in cellular ceramide levels, and we believe this to be a consequence of defects at or below the level of early-to-late trafficking. This is the first demonstration of a connection between ceramide/sphingolipid levels and Wg signaling and highlights the importance of these lipids in regulation of signaling processes.

UAS-diap1
In general, we analyzed male larvae for dissection in RNAi experiments and female larvae for schlank P365/331 clonal analyses, as schlank is located on the X chromosome.

Cloning and transgene production
The Lag1 open-reading frame was based on amplification of the cDNA clone LD18904 (BDGP) and cloned into a pUASattB vector for C-terminal 3xHA tagging (Bischof et al 2012, under revision). For the ceramide synthase dead version H215D (Bauer et al., 2009)  library BACPAC Resources Center.

Generation of presumptive null allele schlank331
For the schlank 331 allele, the Bloomington schlank P365 fly stock was used to generate an imprecise excision. This deleted the last 63 base pairs of the first schlank exon (including a predicted transmembrane domain) and fused the intronic region downstream of the P-element P 365 insertion site to the remaining first exon.

Electron microscopy
For analysis, male larvae with the genotype ywf; UAS-CD2-HRP/Sp or CyO; schlank RNAi 1 /hhGal4 were dissected and immediately fixed with 2.5% glutaraldehyde in PBS for 15 minutes. After washing in PBS they were stained for 2 minutes using the DAB Peroxidase Substrate Kit (Vector Laboratories). Following 4 further washing steps the discs were postfixed for 1h with 1% OsO4 in PBS and block contrasted with 2% uranyl acetate in H2O. After dehydration in an alcohol series (50, 70, 96, 2x100% 15 min each) the discs were embedded in Epon and polymerized for 24h at 60°C.
Thin sections were cut with a Reichert Ultracut E microtome and imaged with a Gatan Orius 1000 CCD camera in a Tecnai G2 spirit transmission electron microscope (FEI, Eindhoven, Netherlands).

Immunohistochemistry
Immunostaining of Drosophila wing imaginal discs and embryos was performed according to standard protocols. Primary antibodies used in this study were: mouse anti-Wg (4D4, DSHB, 1:1000), guinea pig anti-Sens (GP55, gift  Pictures were taken with a Zeiss LSM710 confocal microscope and the Zen software.
Images were processed using ImageJ and Photoshop Elements.
The images were further processed using the IMARIS software. A cubic volume was defined in control and mutant compartments and the overall Wg fluorescence was measured. To mark the endosomal compartments the Surface tool was used with automatic creation and the following settings: local Background: 0.6, Threshold: 20, enable split objects, Quality: above 10. This allowed assessment of the ratio of endosomal volumes in mutant versus control compartment of each disc analyzed (N=7 for each genotype). To quantify the amount of Wg in the respective endosomes, the fluorescence intensity of Wg signal in the previously defined endosomal compartments was measured and again set into relation of mutant to control compartment in each disc per endosome volume.

realtime PCR
Flies carrying the schlank RNAi 1 , and as control yw flies, were crossed to the C765Gal4 driver at 29°C. Male larvae were dissected in three independent experiments and RNA was extracted using the Nucelospin RNA II kit (Machery-Nagel). Following an additional DNA digestion (DNA-free™ kit, Ambion) we used the Transcriptor High Fidelity cDNA synthesis kit (Roche Applied Science) for cDNA synthesis. Tubulin, actin and TBP expression was used for normalization of each experiment; otherwise the experiments were not normalized to show the variability between the replicates.         To test lysosomes we combined hh-Gal4 mediated schlank RNAi with a tubulin promoter driven Lamp1-GFP fusion and found this lysosomal marker enlarged in the schlank mutant posterior compartment.