As an embryo develops, its cells divide many times until they specialize to become distinct cell types that make up the tissues and organs. To do so, the cells need to communicate, and some send signals by making and releasing proteins that travel to more distant cells. One such signaling pathway is called Hedgehog signaling. This pathway is necessary so that the tissue and organs develop properly. If faulty, it can stop the embryo from developing properly and even lead to diseases such as cancer.

Hedgehog signaling is initiated by the Hedgehog protein, which needs to be released from the cells that produce the message to transport the signal to the target cells. A protein called Dispatched helps Hedgehog to get free and travel to its destination. Without Dispatched, Hedgehog cannot be released and the embryos will not develop. Now, Stewart, Marada et al. wanted to find out if and how Dispatched itself is controlled by studying embryo cells of mice.

The results showed that a protein called Furin activates Dispatched by splitting it at a specific point. When the break-point on Dispatch was genetically modified, Furin could no longer cleave the protein. As a consequence, Dispatched did not reach the correct location within cells to help Hedgehog move away from signal-releasing cells.

This suggests that Furin is an essential protein of the Hedgehog pathway. A next step will be to see if this is also the case in humans. Some cancer cells can produce large amounts of Hedgehog protein, which makes tumors grow faster. A better understanding of Hedgehog signaling may help to find new cancer therapies that can block this pathway in cancer cells.

Hedgehog (Hh) ligands are produced as precursor proteins that undergo autocatalytic processing whereby a carboxyl-terminal intein-like domain cleaves itself, in a cholesterol-dependent manner, from an amino-terminal signaling domain (Guy, 2000; Lee et al., 1994). The resulting ~20 kDa signaling protein is covalently modified by cholesterol on its new carboxyl-terminal cysteine and by a long chain fatty acid on its amino-terminus (Porter et al., 1996a; Chamoun et al., 2001; Pepinsky et al., 1998; Long et al., 2015). These lipid modifications contribute to physiological Sonic Hh (Shh) activity by governing ligand distribution across developing tissues, and influencing ligand potency toward target cells (Long et al., 2015; Taylor et al., 2001; Li et al., 2006; Porter et al., 1996b; Burke et al., 1999). Distribution is controlled by Shh lipid modifications conferring high membrane affinity to the mature ligand, thereby anchoring it to the producing cell surface. In order to reach long-range target cells, Shh must be deployed from producing cell membranes through a process that is dependent upon Dispatched 1 (Disp), a predicted twelve-pass transmembrane protein that shares homology with the bacterial Resistance, Nodulation and Division (RND) Transporter superfamily (Burke et al., 1999; Caspary et al., 2002; Ma et al., 2002; Kawakami et al., 2002; Amanai and Jiang, 2001). Disp1 knockout mice phenocopy animals lacking the essential Shh signal transducing component Smoothened (Smo), underscoring the importance of Disp for pathway activity during early development (Caspary et al., 2002; Ma et al., 2002; Kawakami et al., 2002).

In vertebrates, Disp functions with the secreted glycoprotein Scube2 to facilitate Shh membrane extraction (Ma et al., 2002; Creanga et al., 2012; Tukachinsky et al., 2012). The precise mechanism by which Disp and Scube2 mobilize Shh from the producing cell membrane is not yet clear. However, Disp contains a sterol sensing domain (SSD) that is thought to interact with the Shh cholesterol modification to position the ligand for transfer to Scube2 (Creanga et al., 2012; Tukachinsky et al., 2012). Despite this advance in understanding the Disp-Scube2 functional relationship, little is known about how Disp activity is regulated. Biochemical and cell biological analyses have shown Disp must organize into trimers and localize to the basolateral cell surface to release Shh (Etheridge et al., 2010). Genetic studies in Drosophila suggest a crucial role for Disp-mediated endosomal recycling during Hh deployment, demonstrating that apically localized Hh must be internalized in a Disp-dependent manner, and then retargeted to the cell surface to exit ligand-producing cells (D'Angelo et al., 2015; Callejo et al., 2011). Loss of Disp function triggers apical accumulation of Hh and disruption of long-range signaling (D'Angelo et al., 2015; Callejo et al., 2011), suggesting the ability of Disp to appropriately traffic with Hh is imperative for ligand release. The regulatory processes influencing Disp membrane targeting and recycling have not yet been established.

Herein, we demonstrate that Disp membrane targeting and recycling is dependent upon convertase-mediated cleavage. Cleavage occurs at an evolutionarily conserved site in the predicted first extracellular loop of Disp (EC1) by the proprotein convertase Furin. Mutation of the EC1 cleavage site prevents Disp processing and disrupts Shh deployment, consistent with convertase cleavage being an essential step in Disp functional maturation. Results suggest that Disp is clipped at the cell surface and that the resulting amino-terminal fragment and processed carboxyl domain are differentially trafficked post-processing. Disruption of processing by cleavage site mutation results in altered membrane distribution of Disp, leading to compromised pathway activity in vivo. Combined, these results establish cleavage as an essential step for Disp functionality, and provide novel mechanistic insight into control of Disp function in ligand-producing cells.

To begin biochemical and cell biological analysis of Disp regulation, we generated a carboxyl-terminally HA epitope-tagged murine Disp (DispHA) expression vector. All commercial and custom anti-Disp antibodies tested failed to detect the murine Disp protein, necessitating use of the epitope-tagged expression vector. Western blot of cell lysates from NIH3T3 cells transfected with plasmid encoding DispHA revealed two distinct protein bands detected by anti-HA antibody, one running near the predicted molecular weight of 175 kDa, hereafter referred to as Disp175, and a second with an apparent molecular weight of ~145 kDa, Disp145 (Figure 1A). Because membrane and secreted proteins are commonly modified by addition of N-linked glycans, we tested whether the size difference of the two species resulted from differential N-glycan modification. Lysates from cells expressing DispHA were treated with Endo H or PNGase F enzymes, and their migration on SDS-PAGE gels was assessed. Treatment with Endo H, which removes simple N-glycans added in the endoplasmic reticulum (ER), resolved a Disp protein species from Disp175, indicating a fraction of the upper band was ER-localized (Figure 1B lane 2, arrowhead). The lower band was resistant to Endo H. However, PNGase F, which strips both simple and complex post-ER glycans, significantly altered migration of Disp145, indicating post-ER localization of the smaller protein species (lane 3, arrow). PNGase F treatment collapsed Disp175 to a size similar to its Endo H-sensitive fraction, consistent with the larger protein species containing both ER and post-ER fractions (lane 3, arrowhead).

Figure 1

Download asset Open asset

Figure 1—source data 1

Data for Figure 1D.

https://doi.org/10.7554/eLife.31678.004

Download elife-31678-fig1-data1-v2.xlsx

The observation that the Disp145 fraction was highly enriched for EndoH-resistant glycosylation, suggested Disp175 might be cleaved to generate a truncated protein after ER exit. To test this, a V5 epitope tag was inserted in the amino-terminal region of the predicted EC1, as determined using TMPred and HMMTOP 2.0 secondary structure prediction tools (Figures 1C and 2B). The V5 insertion site was chosen based upon the apparent molecular weight difference of the two DispHA protein species. Double tagged V5DispHA was expressed in NIH3T3 cells, and cell lysates were assessed by western blot. Disp175 was detected by both HA and V5 antisera (Figure 1C). Conversely, Disp145 was detected only by HA, suggesting loss of the V5 epitope from the Disp145 species. Accordingly, a ~ 30 kDa fragment was detected by V5 antisera, confirming that Disp protein is clipped to produce Disp145.

Figure 2

Download asset Open asset

To determine whether Disp175 was processed before or after reaching the plasma membrane, cell surface biotinylation experiments were performed (Figure 1D). V5DispHA-expressing NIH3T3 cells were incubated in biotin-containing culture medium for 30 min at 4°C prior to lysis, and biotinylated proteins were captured from cell lysates on streptavidin-coated beads. Bound (surface) and unbound (intracellular) fractions were examined by western blot against the V5 tag to detect the unprocessed protein and the ~30 kDa V5 cleavage fragment. Combined densitometry analysis of four independent experiments revealed that unprocessed Disp175 enriched in the non-biotinylated intracellular fraction. Approximately 60% of total Disp175 signal was detected in the unlabeled intracellular fraction with ~40% present on the cell surface (Figure 1D. lane 4 vs. 6, arrowhead and densitometry summary, purple). Although Disp175 represented the lesser pool of surface-labeled V5DispHA, its presence on the surface argued against Disp145 conversion occurring prior to it reaching the plasma membrane. Consistent with this hypothesis, the processed 30 kDa V5Disp fragment was significantly enriched on the cell surface, accounting for ~76% of the total V5Disp30 signal (Figure 4D, arrow and densitometry summary, green). Combined with the above deglycosylation analysis, these results suggest Disp145 is likely generated from Disp175 at the cell surface. Its generation in the ER or Golgi is unlikely given the low percentage (~23%) of Disp30 in the intracellular, non-biotinylated fraction.

We next sought to identify the exact cleavage site in EC1. To do so Disp-Flag was expressed in HEK293T cells, and Disp175 and 145 proteins were purified on Flag beads. Disp145 was excised and subjected to Edman degradation to identify its amino-terminal residues (Figure 2A, arrow). EVDWNF, which maps to amino acids 280–285 in EC1 of the murine protein, was identified as the amino-terminal sequence (Figure 2B, red line). This is directly adjacent to a dibasic amino acid proprotein convertase (PC) cleavage motif that is conserved in Disp proteins from Drosophila, zebrafish, chick, mouse and human (Figure 2B, yellow box) (Seidah et al., 2013). Consistent with this being a functional cleavage site, disruption of the murine Disp convertase motif by R279A, E280A mutation (DispCS) appreciably reduced Disp175 to Disp145 conversion, evidenced by nearly undetectable Disp145 or Disp 30 signal in DispCS cell lysates (Figure 2C, lane 3 and 2D, lane 3). The 175 kDa fraction of DispCS was predominantly Endo H resistant, indicating that failure to cleave was not likely due to the mutant protein being retained in the ER (Figure 2C, lane 8). Notably, CS mutation resulted in pronounced accumulation of a large molecular weight band (~250 kDa) that was also evident for the wild-type protein, albeit at reduced intensity (Figure 2C lane 3 compared to 2, bracket). The 250 kDa fraction of DispCS was largely resistant to deglycosylating enzymes, suggesting the molecular weight shift was not due to alteration of N-glycosylation status (Figure 2C, lanes 7–9). Moreover, Disp250 accumulated for both wild type and CS DispHA proteins at equal intensities following proteasome inhibition by MG132 treatment (Figure 2E). As such, Disp250 may represent a Disp175 species marked for proteosomal degradation by post-translational modifications such as ubiquitination, neddylation or sumoylation.

To confirm a proprotein convertase was responsible for Disp processing, DispHA-expressing NIH3T3 cells were treated with cell-permeable Furin Inhibitor I, which blocks activity of convertase family members Furin, PCSK1, PCSK2, PACE4, PCSK5 and PCSK7. Treatment with increasing concentrations of drug dose-dependently reduced Disp145 levels (Figure 3A, and green in bottom panel). Despite this, steady state levels of the Disp175 precursor species did not increase (magenta). Instead, as was observed following mutation of the cleavage site (Figure 2C), chemical inhibition of Furin family proteases triggered accumulation of the ~250 kDa fraction (Figure 3A, bracket, and bottom panel, black). Conversion of Disp175 to this larger species likely accounts for the lack of Disp175 accumulation following cleavage inhibition.

Figure 3 with 1 supplement see all

Download asset Open asset

Figure 3—source data 1

Data for Figure 3A.

https://doi.org/10.7554/eLife.31678.008

Download elife-31678-fig3-data1-v2.xlsx

Furin inhibitor I sensitivity, combined with results suggesting cleavage occurs after Disp reaches the cell surface (Figure 1), indicated a dibasic amino acid-specific convertase such as Furin, PCSK5, PACE4 or PCSK7 (Seidah et al., 2013; Seidah and Prat, 2012). To identify the specific PC facilitating Disp cleavage, CRISPR/Cas9 technology was used to generate knockout MEF lines for each of these genes. V5DispHA was expressed in two independent clonal lines knocked out for each of these genes, and examined for cleavage by western blot for the 30 kDa V5 fragment (Figure 3B and Figure 3—figure supplement 1). Knockout of Furin, which targets substrates in the trans-Golgi, at the cell surface and in recycling endosomes (Seidah et al., 2013), blocked formation of V5Disp30. Knockout of PCSK5, PACE4 and PCSK7 did not, identifying Furin as the candidate convertase responsible for Disp cleavage. Accordingly, over-expression of epitope-tagged Furin-Myc with V5DispHA enhanced cleavage of the wild-type protein, and induced low-level cleavage of the CS mutant, suggesting that by increasing Furin protein levels compensatory and/or off-site cleavage can occur (Figure 3C–D). To test for Furin-Disp association, V5DispHA was co-expressed with Furin-Myc in HEK293T cells and anti-Myc immunoprecipitation experiments were performed (Figure 3D). Both wild-type and CS mutant V5DispHA proteins were detected in anti-Myc immunoprecipitates, consistent with an interaction occurring between Furin and Disp (right panel, lanes 5 and 7). V5DispHA was not collected by anti-Myc in the absence of Furin-Myc expression, confirming specificity of the immunoprecipitation (right panel, lanes 1–4).

To further test for a specific requirement for Furin in facilitating Disp cleavage, V5DispHA was expressed in Furin-deficient colorectal adenocarcinoma-derived LoVo cells, and generation of the V5Disp30 cleavage fragment was assessed (Takahashi et al., 1993). When expressed in control HCT-15 colorectal cells, both 175 kDa and 30 kDa protein species were evident (Figure 3E, lane 1). Conversely, murine V5DispHA protein expressed in LoVo cells failed to produce V5Disp30, and instead migrated in a manner similar to the CS mutant (lanes 2–3). Cleavage disruption was specific to Furin loss because its re-expression in LoVo cells rescued V5Disp30 production (lane 4). Combined with the above, these results support that Disp is cleaved by Furin.

Proprotein convertases such as Furin typically act on inactive proproteins to remove inhibitory or regulatory domains as a requisite step in functional maturation of the substrate. However, a small number of substrates are inactivated by convertase cleavage (Seidah et al., 2013). To test how Furin cleavage affected Disp, Shh transcriptional reporter assays were performed by co-culturing Shh-responsive LightII reporter cells with Disp-/- mouse embryonic fibroblasts (MEFs) engineered to stably express Shh (Ma et al., 2002; Taipale et al., 2000). Shh-expressing Disp-/- MEFs were transiently transfected with vectors encoding GFP control, wild-type DispHA, DispHA-CS or a published nonfunctional Disp mutant, DispHA-TM. This mutant harbors mutations of conserved residues in the predicted transporter motifs in TM domains 4 and 10 (Figure 2B [Ma et al., 2002]). Ligand-producing cells were co-cultured with LightII reporter cells for ~48 hr, and reporter induction was measured (Figure 4A). Co-culture of LightII cells with Disp-/- MEFs expressing GFP +Shh failed to induce a significant change in reporter gene activity over that of the GFP control. Conversely, co-culture of reporter cells with Shh-expressing MEFs transiently expressing wild-type DispHA induced a statistically significant reporter response, consistent with Shh deployment being rescued by re-expression of wild-type DispHA protein. Expression of DispCSHA in Shh-expressing Disp-/- MEFs affected LightII cell reporter activity similarly to the GFP control. This reduced signaling level was also similar to what was observed in LightII cells co-cultured with Shh-stable Disp-/- MEFs expressing nonfunctional DispTM-HA (Ma et al., 2002). As such, attenuation of Disp cleavage likely compromises Shh deployment to target cells.

Figure 4

Download asset Open asset

Figure 4—source data 1

Data for Figure 4A.

https://doi.org/10.7554/eLife.31678.010

Download elife-31678-fig4-data1-v2.xlsx

Figure 4—source data 2

Data for Figure 4B.

https://doi.org/10.7554/eLife.31678.011

Download elife-31678-fig4-data2-v2.xlsx

Figure 4—source data 3

Data for Figure 4C.

https://doi.org/10.7554/eLife.31678.012

Download elife-31678-fig4-data3-v2.xlsx

To directly test the ability of DispCS to deploy Shh, ligand release into culture media of Shh-expressing Disp-/- MEFs was examined. In vertebrates, Disp functions with the secreted glycoprotein Scube2 to promote ligand release (Creanga et al., 2012; Tukachinsky et al., 2012). Therefore, wild type or DispCSHA proteins were co-expressed with increasing Scube2-Flag in Shh-stable Disp-/- cells, and ligand accumulation in culture media was monitored by western blot and densitometry analysis of protein-normalized media samples (Figure 4B). In the absence of Scube2, Shh was not detected in culture media of GFP and Shh-expressing Disp-/- cells (Figure 4B, lane 2, media and light gray in densitometry analysis). Co-expression of Scube2-Flag was unable to bolster Shh release from GFP-transfected cells, despite efficient Scube2-Flag secretion from the Shh-stable Disp-/- cells (lanes 3–5, media and light gray). Low-level re-expression of wild-type DispHA in MEFs modestly increased Shh release into culture media over that of the GFP control (lane 7 compared to 2, media and dark gray in densitometry analysis). Consistent with Scube2 partnering with Disp to facilitate ligand extraction from the membrane, co-expression of increasing levels of Scube2-Flag with DispHA prompted a dose-dependent increase in Shh protein detectable in conditioned culture media (lanes 8–10 and dark gray). Conversely, cleavage-deficient DispCS failed to effectively promote Shh release into conditioned media when expressed alone or in combination with Scube2-Flag (Figure 4B, lanes 12–15 and black). Shh release was similarly affected by genetic elimination of Furin (Figure 4C). Whereas control MEFs released Shh into culture media, CRISPR/Cas9 generated Furin-/- MEF clones, which failed to effectively cleave V5DispHA, were compromised in their ability to release ligand (lanes 1–2 vs 3–4). Combined, these results support that Disp cleavage is necessary for Shh deployment.

To assess whether cleavage site disruption would compromise Disp activity in vivo, we turned to the Drosophila system, which is a robust and genetically tractable model for Hh signal transduction (Jiang and Hui, 2008; Lee et al., 2016). We first confirmed processing of endogenous Drosophila Disp (dDisp) in cultured fly cells using a polyclonal antibody raised against predicted EC4 of dDisp (Figure 5A–A’). Wing imaginal disc-derived Clone 8 (Cl8) cells were treated with control or 5’dispUTR dsRNA to assess endogenous dDisp, or transfected with a dDispHA expression vector to assess over-expressed protein. The dDisp antibody detected two distinct bands, one migrating at the predicted molecular weight of ~150 kDa (dDisp150) and a second species with an approximate molecular weight of ~110 kDa (dDisp110). The intensity of both bands decreased following disp dsRNA treatment and increased with over-expression of epitope-tagged dDispHA (Figure 5A). However, the ratio of the fractions shifted from being approximately equal for the endogenous protein to the upper band being predominant when over-expressed (Figure 5A, lane 1 vs. 3). Similar to what was observed for mouse Disp, a ~ 30 kDa V5 fragment was released from double-tagged V5dDispHA expressed in Cl8 cells (Figure 5B). To test for processing of endogenous dDisp protein in vivo, dDisp was immunoprecipitated from wing imaginal disc lysate prepared from third instar larvae using the dDisp antisera (Figure 5A’). Both bands were evident at equal levels in anti-Disp immunoprecipitates, but not in IgG control immunoprecipitates, confirming that endogenous dDisp protein is processed in cultured fly cells and in vivo in wing imaginal discs.

Figure 5 with 1 supplement see all

Download asset Open asset

In addition to the dibasic convertase cleavage motif at 237/238 of the fly protein that aligns with the mouse cleavage site (Figure 2B), we identified additional consensus motifs at amino acids 209 and 218. Mutation of each of the three sites on their own did not block cleavage (not shown). We therefore engineered an in frame deletion to remove sequence encompassing all three putative cleavage sites (Δ206–238), and tested for processing of the ΔCS mutant in Cl8 cells. Deletion of the three putative cleavage sites ablated generation of dDisp110 (Figure 5C lane 4 compared to 1). Endo H and PNGase F sensitivity analysis revealed that like the mouse protein, dDisp110 of the wild type protein harbored complex N-linked glycans, indicative of post-ER localization (Figure 5C, lane 3). The dDispΔCS mutant showed both ER and post-ER fractions, indicating loss of cleavage did not result from ER retention (Figure 5C lanes 4–6).

Hh patterns the Drosophila wing by controlling gene expression in the wing imaginal disc (De Celis, 2003). Alteration of Hh signaling during wing development triggers phenotypes in the adult wing, providing a robust system for monitoring changes in pathway activity in vivo. Wild type or ΔCS UAS-dispHA transgenes were expressed in the dorsal compartment of wing imaginal discs using the apterous-GAL4 driver, and adult wings were screened for phenotypes. Consistent with the established positive role of dDisp in Hh release (Burke et al., 1999), over-expression of wild-type dDispHA triggered obvious blistering of the adult wing (Figure 5E compared to D). This phenotype is similar to what is observed in response to dorsal wing disc over-expression of activating mutants of the Hh signal transducing component Smo (Marada et al., 2013). Blistering is indicative of overgrowth of the dorsal face of the wing blade, likely resulting from over-proliferation of dorsal compartment cells in response to enhanced Hh release by over-expressed dDispHA. By comparison, over-expression of dDispΔCS triggered modest wing curling, but did not induce pronounced blistering (Figure 5F). These results suggest compromised in vivo activity by the cleavage-deficient mutant.

To directly test for the effect of dDisp cleavage disruption on Hh export, a transgene encoding a Hh protein with an internal GFP that is retained post ligand processing (HhGFP [Torroja et al., 2004; Hartman et al., 2013]) was expressed alone or in combination with wild type or ΔCS V5dDispHA proteins in Drosophila salivary glands (Figure 5G–I). Salivary glands were chosen because they are large and do not express endogenous disp (modENCODE). These characteristics allowed for clear visualization of dDispHA effects on HhGFP without compensation by the endogenous protein. UAS-hhGFP and UAS-V5dispHA transgenes were recombined onto the same chromosome, and then expressed under control of SGS-GAL4. In control non-dDisp expressing salivary glands, Hh accumulated in large puncta on the basal surface of salivary gland cells (Figure 5G). Expression of wild-type V5dDispHA in salivary gland cells resulted in a mostly uniform membrane localization of HhGFP with distinct puncta evident throughout basolateral optical sections (Figure 5H and Figure 5—figure supplement 1). HhGFP was largely depleted at the basal surface but was evident in a small number of distinct puncta. Conversely, cells expressing V5dDispHAΔCS showed an overt increase in HhGFP puncta throughout basolateral optical sections with accumulation of large puncta at the basal surface that resembled puncta observed in the absence of V5dDispHA expression (Figure 5I and Figure 5—figure supplement 1). This punctate organization is similar to that observed for Hh protein expressed in embryonic and larval tissues (Callejo et al., 2011; Gallet et al., 2003), suggesting the puncta could represent ‘packaged’ HhGFP that has been primed for release. Accumulation of these puncta at the basal surface of DispΔCS-expressing salivary glands, where the majority of DispΔCSHA signal was detected, suggests that although HhGFP may appropriately package for release, it cannot effectively deploy when Disp cleavage is compromised. Thus, dDisp cleavage is required for ligand deployment in vivo.

Having established an evolutionarily conserved requirement for Disp processing for effective ligand release, we next wanted to examine the mechanism by which processing impacted Disp functionality. Disp is predicted to assemble into functional trimers (Etheridge et al., 2010), raising the possibility that cleavage might control oligomerization. To determine whether cleavage disruption attenuated Disp trimer assembly, wild type and CS murine DispHA proteins were expressed in NIH3T3 cells, and lysates were examined by native gel electrophoresis and western blot (Figure 6A). A ~ 480 kDa fraction consistent with the predicted molecular weight of the Disp trimer was evident for both wild type and cleavage deficient DispHA proteins, indicating that blocking cleavage did not block trimer formation (Figure 6A, top). Moreover, larger molecular weight fractions (~750 kDa) were present at equal intensities for both wild type and CS proteins, suggesting cleavage disruption did not prevent murine Disp from forming higher order assemblies.

Figure 6

Download asset Open asset

Figure 6—source data 1

Data for Figure 6E.

https://doi.org/10.7554/eLife.31678.016

Download elife-31678-fig6-data1-v2.xls

Having confirmed DispCS was not deficient in trimer formation, we next assessed its ability to bind Shh. Disp is thought to bind Shh through a sterol sensing domain (SSD)-mediated association with the carboxyl-terminal Shh cholesterol modification, which is subsequently transferred to Scube2 (Tukachinsky et al., 2012). Disp EC1, which contains the processing site, is situated directly adjacent to the SSD (Figure 2B). As such, Disp cleavage could potentially influence Shh association by governing SSD access. To test this, wild type and CS mutant V5DispHA proteins were co-expressed with Shh in Disp-/- cells, and the ability of Shh to co-immunoprecipitate with DispHA from cellular lysates was examined. Similar amounts of Shh co-immunoprecipitated on anti-HA beads with both WT and CS mutant DispHA proteins (Figure 6B, lanes 3 and 4). Shh-DispHA binding was specific because Shh failed to bind HA beads in the absence of DispHA (lane 2). These results suggest that Disp cleavage does not regulate ligand binding.

Disp predominately localizes to basolateral membranes in polarized epithelial cells, but a minor sub-apical, vesicular pool has been reported (Etheridge et al., 2010; Callejo et al., 2011). Cholesterol-modified Hh ligand enriches apically, placing the majority of Disp protein and its target ligand in non-overlapping membrane domains (D'Angelo et al., 2015; Callejo et al., 2011). In Drosophila, dDisp achieves ligand release by capturing apical Hh in recycling endosomes, which subsequently retarget to the plasma membrane for ligand deployment (D'Angelo et al., 2015; Callejo et al., 2011). Because convertase-mediated cleavage can impact protein function by affecting intracellular trafficking (Constam, 2014), we hypothesized Disp cleavage might affect its membrane targeting. To assess its subcellular localization in polarized cells in vivo, we tested V5dDispHA subcellular localization in ovarian follicle cells, which are large and polarized, making them ideal for monitoring protein trafficking and subcellular localization. Wild type and ΔCS V5dDispHA proteins were expressed using the follicle cell driver C204-GAL4, and localization of V5 (amino) and HA (carboxyl) epitope tags was examined in stage 10 ovaries (Figure 6C–D). Colocalization between amino-V5 (magenta) and carboxyl-HA (green), indicative of the unprocessed protein, was evident for wild type V5dDispHA in basal vesicles and along basolateral membranes (Figure 6C, white; F-Actin marks apical, blue). Consistent with processing removing the amino-terminal fragment from dDisp110, a clear separation of the two signals was observed. The released amino-terminal V5 fragment, evidenced by V5 signal not colocalized with HA, localized to apical and basolateral membrane and in vesicles throughout the cell (Figure 6C, magenta). Notably, dDisp110HA depleted apically, enriching on basolateral membrane and in basally-localized vesicles (Figure 6C, green), suggesting differential trafficking of the two dDisp domains post-cleavage. Although we cannot rule out a functional role for the amino-terminal fragment post-cleavage, we do not think it contributes to signaling because its over-expression in wing disc-derived Cl8 cells did not alter induction of Hh-dependent luciferase reporter gene activity (Figure 6E). We were unable to directly confirm activity of DispΔN lacking the amino-terminal prodomain due to it being retained in the ER (not shown). However, the functional pool of Disp protein is thought to enrich basolaterally in both murine and Drosophila systems, which is consistent with what we observed for dDisp110HA in follicle cells (Etheridge et al., 2010; Callejo et al., 2011).

Visualization of cleavage-deficient V5dDispΔCSHA revealed a strikingly altered localization from that of the wild-type protein (Figure 6D). The cleavage site mutant showed pronounced accumulation on both apical and basolateral membranes, along with uniform vesicular distribution throughout the cell, suggestive of altered membrane trafficking upon cleavage loss. Notably, signal intensity of cleavage-deficient dDisp was increased compared to wild type, potentially indicating that dDisp protein turnover might be affected by altered membrane recycling.

To directly test whether dDisp membrane recycling was altered by cleavage disruption, we expressed wild type and ΔCS V5dDispHA proteins in S2 cells, and tested for dDisp colocalization with the early endosomal marker Rab5 by immunofluorescence confocal microscopy (Figure 6F, white). Imaris image analysis software was used to perform colocalization analysis of ~50 V5dDispHA-expressing cells per condition across three independent experiments. In cells expressing wild-type V5dDispHA, approximately 50% of the HA signal was colocalized with endogenous Rab5 signal. Cleavage site deletion lessened V5dDispHA-Rab5 colocalization, reducing the percent of HA signal colocalized with Rab5 signal to ~35% (Figure 6G). These results are consistent with compromised membrane recycling, and taken together with in vivo experiments, suggest that Disp cleavage is necessary for proper membrane trafficking.

Disp was first identified as a crucial regulator of Hh ligand deployment in 1999 through a genetic screen conducted in Drosophila (Burke et al., 1999). A number of vertebrate genetic studies subsequently established the importance of Disp in Shh morphogen gradient formation and activity during tissue development (Caspary et al., 2002; Ma et al., 2002; Kawakami et al., 2002; Nakano et al., 2004). Owing to the comparatively small number of cell biological and biochemical interrogations of Disp activity (Creanga et al., 2012; Tukachinsky et al., 2012; Etheridge et al., 2010), mechanistic insight into its regulation and functionality has remained limited. The study presented here improves understanding of Disp regulation by revealing an evolutionarily conserved cleavage event that influences the ability of Disp to deploy Hh family ligands from ligand-producing cells. We report that Disp is cleaved at a conserved processing site in its predicted first extracellular loop by the proprotein convertase Furin. Cleavage site mutation compromises Disp-mediated ligand deployment in vitro and in vivo, leading to reduced pathway activation in target cells. As such, this study is the first to provide mechanistic insight into a process promoting functional maturation of Disp for its role in ligand deployment.

Proprotein convertase-mediated cleavage of substrate proteins typically promotes their maturation by removing inhibitory prodomains, revealing active domains, releasing bioactive fragments, priming substrates for cleavage by additional proteases, or by influencing substrate subcellular localization (Seidah et al., 2013; Seidah and Prat, 2012). How prodomains affect substrate trafficking is not yet fully understood, but a logical hypothesis is that cleavage regulates association with trafficking molecules and/or tethering proteins along secretory or endosomal recycling routes. Such a model has been proposed for the convertase substrate Nodal, which accumulates on the cell surface following processing inhibition (Constam, 2014; Blanchet et al., 2008a, 2008b). Because cleavage disruption altered full-length dDisp membrane localization in polarized epithelial cells to mimic what was observed for the processed 30 kDa V5 fragment, we hypothesize that like Nodal, Disp membrane trafficking is regulated in cleavage-dependent manner. The observed colocalization of amino- and carboxyl-terminal dDispWT epitope tags on basolateral membranes of Drosophila follicle cells suggests cleavage occurs after basolateral targeting of unprocessed Disp. Consistent with this hypothesis, Furin has been demonstrated to traffic basolaterally in polarized epithelial cells (Simmen et al., 1999).

Intriguingly, whereas Disp is predominantly observed to localize to basolateral membrane, Hh enriches on apical membrane, from which it must be endocytosed in a Disp-dependent manner to facilitate its release upon plasma membrane recycling (D'Angelo et al., 2015; Callejo et al., 2011; Gallet et al., 2003). Our observations that cleavage-deficient Disp (1) accumulated uniformly along apical and basolateral membranes of Drosophila follicle cells and (2) showed reduced colocalization with the Rab5 endosomal marker when expressed in S2 cells suggest that its endosomal trafficking is likely compromised by cleavage disruption. As such, we suggest a testable model in which basolateral Disp cleavage activates the protein for endosomal recycling, allowing it to capture, recycle and deploy apically localized Hh. We do not believe Disp cleavage is required to interact with ligand because both wild type and cleavage-deficient murine Disp proteins co-immunoprecipitated with Shh. Cleavage-deficient Disp was also capable of forming multimers, diminishing the likelihood that EC1 clipping regulates self-association.

Results obtained using both murine and Drosophila experimental systems demonstrate that Disp processing is evolutionarily conserved. However, whereas mutation of a single consensus cleavage motif in EC1 of mouse Disp was sufficient to disrupt cleavage, multiple predicted sites had to be targeted in Drosophila Disp EC1. That three predicted motifs had to be deleted to block Drosophila Disp cleavage suggests cleavage site redundancy in the fly protein. Multiple redundant motifs may indicate increased reliance upon Disp cleavage for function in the Drosophila system. Notably, Drosophila lack a Scube2-like protein that partners with dDisp to extract Hh from ligand-producing cells (Creanga et al., 2012; Tukachinsky et al., 2012). Multiple redundant sites might serve as fail-safes to assure dDisp cleavage and efficient Hh membrane release in the absence of Scube2-mediated assistance. Another possible explanation is that Drosophila Disp is cleaved by additional or alternative proteases with different cleavage site preferences or efficiencies. It has been reported that although Drosophila convertases can share substrate specificity with their vertebrate counterparts, cleavage efficiency will often vary between the two systems (De Bie et al., 1995).

In vertebrates, genetic loss-of-function of proprotein convertases such as Furin, PCSK5 and PACE4 triggers developmental defects leading to embryonic lethality (Seidah et al., 2013; Roebroek et al., 1998; Essalmani et al., 2008; Constam and Robertson, 2000a; Constam and Robertson, 2000b). Although Furin and Disp1 knockout mice both show axial rotation and heart looping defects that lead to death at or before embryonic days ~ E9.5-10.5, their phenotypes are not indistinguishable (Caspary et al., 2002; Ma et al., 2002; Kawakami et al., 2002; Roebroek et al., 1998). Most notably, whereas Disp1 mutant embryos show clear disruption of left-right asymmetry, Furin mutants do not (Ma et al., 2002; Kawakami et al., 2002; Roebroek et al., 1998). This could be due to functional compensation by other convertases in vivo. Consistent with this notion, functional redundancy between Furin, PACE4, PCSK5 and PCSK7 has been reported (Seidah et al., 2013; Roebroek et al., 1998). It is also possible that Furin-mediated Disp cleavage occurs in temporal or tissue-specific manners to scale Shh release efficiency commensurate with increased need. In such a scenario, Disp cleavage would be predicted to occur during later developmental stages or in larger developing tissues to bolster Shh deployment for a growing population of target cells. Future in vivo studies using vertebrate model systems will be required to explore these hypotheses, and to determine how Disp cleavage disruption impacts Shh-dependent developmental patterning.

1. 1. Amanai K
   2. Jiang J

   (2001)

   Distinct roles of central missing and dispatched in sending the hedgehog signal

   Development 128:5119–5127.

   • PubMed
   • Google Scholar
2. 1. Bischof J
   2. Maeda RK
   3. Hediger M
   4. Karch F
   5. Basler K

   (2007) An optimized transgenesis system for Drosophila using germ-line-specific phiC31 integrases

   PNAS 104:3312–3317.

   https://doi.org/10.1073/pnas.0611511104

   • PubMed
   • Google Scholar
3. 1. Blanchet MH
   2. Le Good JA
   3. Mesnard D
   4. Oorschot V
   5. Baflast S
   6. Minchiotti G
   7. Klumperman J
   8. Constam DB

   (2008a) Cripto recruits furin and PACE4 and controls Nodal trafficking during proteolytic maturation

   The EMBO Journal 27:2580–2591.

   https://doi.org/10.1038/emboj.2008.174

   • PubMed
   • Google Scholar
4. 1. Blanchet MH
   2. Le Good JA
   3. Oorschot V
   4. Baflast S
   5. Minchiotti G
   6. Klumperman J
   7. Constam DB

   (2008b) Cripto localizes Nodal at the limiting membrane of early endosomes

   Science Signaling 1:ra13.

   https://doi.org/10.1126/scisignal.1165027

   • PubMed
   • Google Scholar
5. 1. Burke R
   2. Nellen D
   3. Bellotto M
   4. Hafen E
   5. Senti KA
   6. Dickson BJ
   7. Basler K

   (1999) Dispatched, a novel sterol-sensing domain protein dedicated to the release of cholesterol-modified hedgehog from signaling cells

   Cell 99:803–815.

   https://doi.org/10.1016/S0092-8674(00)81677-3

   • PubMed
   • Google Scholar
6. 1. Callejo A
   2. Bilioni A
   3. Mollica E
   4. Gorfinkiel N
   5. Andrés G
   6. Ibáñez C
   7. Torroja C
   8. Doglio L
   9. Sierra J
   10. Guerrero I

   (2011) Dispatched mediates Hedgehog basolateral release to form the long-range morphogenetic gradient in the Drosophila wing disk epithelium

   PNAS 108:12591–12598.

   https://doi.org/10.1073/pnas.1106881108

   • PubMed
   • Google Scholar
7. 1. Carroll CE
   2. Marada S
   3. Stewart DP
   4. Ouyang JX
   5. Ogden SK

   (2012) The extracellular loops of smoothened play a regulatory role in control of hedgehog pathway activation

   Development 139:612–621.

   https://doi.org/10.1242/dev.075614

   • PubMed
   • Google Scholar
8. 1. Caspary T
   2. García-García MJ
   3. Huangfu D
   4. Eggenschwiler JT
   5. Wyler MR
   6. Rakeman AS
   7. Alcorn HL
   8. Anderson KV

   (2002) Mouse Dispatched homolog1 is required for long-range, but not juxtacrine, Hh signaling

   Current Biology 12:1628–1632.

   https://doi.org/10.1016/S0960-9822(02)01147-8

   • PubMed
   • Google Scholar
9. 1. Chamoun Z
   2. Mann RK
   3. Nellen D
   4. von Kessler DP
   5. Bellotto M
   6. Beachy PA
   7. Basler K

   (2001) Skinny hedgehog, an acyltransferase required for palmitoylation and activity of the hedgehog signal

   Science 293:2080–2084.

   https://doi.org/10.1126/science.1064437

   • PubMed
   • Google Scholar
10. 1. Constam DB
    2. Robertson EJ

    (2000a)

    SPC4/PACE4 regulates a TGFbeta signaling network during axis formation

    Genes & Development 14:1146–1155.

    • PubMed
    • Google Scholar
11. 1. Constam DB
    2. Robertson EJ

    (2000b)

    Tissue-specific requirements for the proprotein convertase furin/SPC1 during embryonic turning and heart looping

    Development 127:245–254.

    • PubMed
    • Google Scholar
12. 1. Constam DB

    (2014) Regulation of TGFβ and related signals by precursor processing

    Seminars in Cell & Developmental Biology 32:85–97.

    https://doi.org/10.1016/j.semcdb.2014.01.008

    • PubMed
    • Google Scholar
13. 1. Creanga A
    2. Glenn TD
    3. Mann RK
    4. Saunders AM
    5. Talbot WS
    6. Beachy PA

    (2012) Scube/You activity mediates release of dually lipid-modified Hedgehog signal in soluble form

    Genes & Development 26:1312–1325.

    https://doi.org/10.1101/gad.191866.112

    • PubMed
    • Google Scholar
14. 1. D'Angelo G
    2. Matusek T
    3. Pizette S
    4. Thérond PP

    (2015) Endocytosis of Hedgehog through dispatched regulates long-range signaling

    Developmental Cell 32:290–303.

    https://doi.org/10.1016/j.devcel.2014.12.004

    • PubMed
    • Google Scholar
15. 1. De Bie I
    2. Savaria D
    3. Roebroek AJ
    4. Day R
    5. Lazure C
    6. Van de Ven WJ
    7. Seidah NG

    (1995) Processing specificity and biosynthesis of the Drosophila melanogaster convertases dfurin1, dfurin1-CRR, dfurin1-X, and dfurin2

    Journal of Biological Chemistry 270:1020–1028.

    https://doi.org/10.1074/jbc.270.3.1020

    • PubMed
    • Google Scholar
16. 1. De Celis JF

    (2003) Pattern formation in the Drosophila wing: The development of the veins

    BioEssays 25:443–451.

    https://doi.org/10.1002/bies.10258

    • PubMed
    • Google Scholar
17. 1. Essalmani R
    2. Zaid A
    3. Marcinkiewicz J
    4. Chamberland A
    5. Pasquato A
    6. Seidah NG
    7. Prat A

    (2008) In vivo functions of the proprotein convertase PC5/6 during mouse development: Gdf11 is a likely substrate

    PNAS 105:5750–5755.

    https://doi.org/10.1073/pnas.0709428105

    • PubMed
    • Google Scholar
18. 1. Etheridge LA
    2. Crawford TQ
    3. Zhang S
    4. Roelink H

    (2010) Evidence for a role of vertebrate Disp1 in long-range Shh signaling

    Development 137:133–140.

    https://doi.org/10.1242/dev.043547

    • PubMed
    • Google Scholar
19. 1. Gallet A
    2. Rodriguez R
    3. Ruel L
    4. Therond PP

    (2003) Cholesterol modification of hedgehog is required for trafficking and movement, revealing an asymmetric cellular response to hedgehog

    Developmental Cell 4:191–204.

    https://doi.org/10.1016/S1534-5807(03)00031-5

    • PubMed
    • Google Scholar
20. 1. Goetz JA
    2. Singh S
    3. Suber LM
    4. Kull FJ
    5. Robbins DJ

    (2006) A highly conserved amino-terminal region of sonic hedgehog is required for the formation of its freely diffusible multimeric form

    Journal of Biological Chemistry 281:4087–4093.

    https://doi.org/10.1074/jbc.M511427200

    • PubMed
    • Google Scholar
21. 1. Guy RK

    (2000) Inhibition of sonic hedgehog autoprocessing in cultured mammalian cells by sterol deprivation

    PNAS 97:7307–7312.

    https://doi.org/10.1073/pnas.97.13.7307

    • PubMed
    • Google Scholar
22. 1. Hartman TR
    2. Strochlic TI
    3. Ji Y
    4. Zinshteyn D
    5. O'Reilly AM

    (2013) Diet controls Drosophila follicle stem cell proliferation via Hedgehog sequestration and release

    The Journal of Cell Biology 201:741–757.

    https://doi.org/10.1083/jcb.201212094

    • PubMed
    • Google Scholar
23. 1. Jiang J
    2. Hui CC

    (2008) Hedgehog signaling in development and cancer

    Developmental Cell 15:801–812.

    https://doi.org/10.1016/j.devcel.2008.11.010

    • PubMed
    • Google Scholar
24. 1. Kawakami T
    2. Kawcak T
    3. Li YJ
    4. Zhang W
    5. Hu Y
    6. Chuang PT

    (2002) Mouse dispatched mutants fail to distribute hedgehog proteins and are defective in hedgehog signaling

    Development 129:5753–5765.

    https://doi.org/10.1242/dev.00178

    • PubMed
    • Google Scholar
25. 1. Lee JJ
    2. Ekker SC
    3. von Kessler DP
    4. Porter JA
    5. Sun BI
    6. Beachy PA

    (1994) Autoproteolysis in hedgehog protein biogenesis

    Science 266:1528–1537.

    https://doi.org/10.1126/science.7985023

    • PubMed
    • Google Scholar
26. 1. Lee RT
    2. Zhao Z
    3. Ingham PW

    (2016) Hedgehog signalling

    Development 143:367–372.

    https://doi.org/10.1242/dev.120154

    • PubMed
    • Google Scholar
27. 1. Li Y
    2. Zhang H
    3. Litingtung Y
    4. Chiang C

    (2006) Cholesterol modification restricts the spread of Shh gradient in the limb bud

    PNAS 103:6548–6553.

    https://doi.org/10.1073/pnas.0600124103

    • PubMed
    • Google Scholar
28. 1. Long J
    2. Tokhunts R
    3. Old WM
    4. Houel S
    5. Rodgriguez-Blanco J
    6. Singh S
    7. Schilling N
    8. J Capobianco A
    9. Ahn NG
    10. Robbins DJ

    (2015) Identification of a family of fatty-acid-speciated sonic hedgehog proteins, whose members display differential biological properties

    Cell Reports 10:1280–1287.

    https://doi.org/10.1016/j.celrep.2015.01.058

    • PubMed
    • Google Scholar
29. 1. Ma Y
    2. Erkner A
    3. Gong R
    4. Yao S
    5. Taipale J
    6. Basler K
    7. Beachy PA

    (2002) Hedgehog-mediated patterning of the mammalian embryo requires transporter-like function of dispatched

    Cell 111:63–75.

    https://doi.org/10.1016/S0092-8674(02)00977-7

    • PubMed
    • Google Scholar
30. 1. Marada S
    2. Navarro G
    3. Truong A
    4. Stewart DP
    5. Arensdorf AM
    6. Nachtergaele S
    7. Angelats E
    8. Opferman JT
    9. Rohatgi R
    10. McCormick PJ
    11. Ogden SK

    (2015) Functional divergence in the role of N-linked glycosylation in smoothened signaling

    PLoS Genetics 11:e1005473.

    https://doi.org/10.1371/journal.pgen.1005473

    • PubMed
    • Google Scholar
31. 1. Marada S
    2. Stewart DP
    3. Bodeen WJ
    4. Han YG
    5. Ogden SK

    (2013) The unfolded protein response selectively targets active smoothened mutants

    Molecular and Cellular Biology 33:2375–2387.

    https://doi.org/10.1128/MCB.01445-12

    • PubMed
    • Google Scholar
32. 1. Nakano Y
    2. Kim HR
    3. Kawakami A
    4. Roy S
    5. Schier AF
    6. Ingham PW

    (2004) Inactivation of dispatched 1 by the chameleon mutation disrupts Hedgehog signalling in the zebrafish embryo

    Developmental Biology 269:381–392.

    https://doi.org/10.1016/j.ydbio.2004.01.022

    • PubMed
    • Google Scholar
33. 1. Ogden SK
    2. Ascano M
    3. Stegman MA
    4. Suber LM
    5. Hooper JE
    6. Robbins DJ

    (2003) Identification of a functional interaction between the transmembrane protein Smoothened and the kinesin-related protein Costal2

    Current Biology 13:1998–2003.

    https://doi.org/10.1016/j.cub.2003.10.004

    • PubMed
    • Google Scholar
34. 1. Pepinsky RB
    2. Zeng C
    3. Wen D
    4. Rayhorn P
    5. Baker DP
    6. Williams KP
    7. Bixler SA
    8. Ambrose CM
    9. Garber EA
    10. Miatkowski K
    11. Taylor FR
    12. Wang EA
    13. Galdes A

    (1998) Identification of a palmitic acid-modified form of human Sonic hedgehog

    Journal of Biological Chemistry 273:14037–14045.

    https://doi.org/10.1074/jbc.273.22.14037

    • PubMed
    • Google Scholar
35. 1. Porter JA
    2. Ekker SC
    3. Park WJ
    4. von Kessler DP
    5. Young KE
    6. Chen CH
    7. Ma Y
    8. Woods AS
    9. Cotter RJ
    10. Koonin EV
    11. Beachy PA

    (1996b) Hedgehog patterning activity: role of a lipophilic modification mediated by the carboxy-terminal autoprocessing domain

    Cell 86:21–34.

    https://doi.org/10.1016/S0092-8674(00)80074-4

    • PubMed
    • Google Scholar
36. 1. Porter JA
    2. Young KE
    3. Beachy PA

    (1996a) Cholesterol modification of hedgehog signaling proteins in animal development

    Science 274:255–259.

    https://doi.org/10.1126/science.274.5285.255

    • PubMed
    • Google Scholar
37. 1. Roebroek AJ
    2. Umans L
    3. Pauli IG
    4. Robertson EJ
    5. van Leuven F
    6. Van de Ven WJ
    7. Constam DB

    (1998)

    Failure of ventral closure and axial rotation in embryos lacking the proprotein convertase Furin

    Development 125:4863–4876.

    • PubMed
    • Google Scholar
38. 1. Seidah NG
    2. Prat A

    (2012) The biology and therapeutic targeting of the proprotein convertases

    Nature Reviews Drug Discovery 11:367–383.

    https://doi.org/10.1038/nrd3699

    • PubMed
    • Google Scholar
39. 1. Seidah NG
    2. Sadr MS
    3. Chrétien M
    4. Mbikay M

    (2013) The multifaceted proprotein convertases: their unique, redundant, complementary, and opposite functions

    Journal of Biological Chemistry 288:21473–21481.

    https://doi.org/10.1074/jbc.R113.481549

    • PubMed
    • Google Scholar
40. 1. Simmen T
    2. Nobile M
    3. Bonifacino JS
    4. Hunziker W

    (1999) Basolateral sorting of furin in MDCK cells requires a phenylalanine-isoleucine motif together with an acidic amino acid cluster

    Molecular and Cellular Biology 19:3136–3144.

    https://doi.org/10.1128/MCB.19.4.3136

    • PubMed
    • Google Scholar
41. 1. Taipale J
    2. Chen JK
    3. Cooper MK
    4. Wang B
    5. Mann RK
    6. Milenkovic L
    7. Scott MP
    8. Beachy PA

    (2000) Effects of oncogenic mutations in Smoothened and Patched can be reversed by cyclopamine

    Nature 406:1005–1009.

    https://doi.org/10.1038/35023008

    • PubMed
    • Google Scholar
42. 1. Takahashi S
    2. Kasai K
    3. Hatsuzawa K
    4. Kitamura N
    5. Misumi Y
    6. Ikehara Y
    7. Murakami K
    8. Nakayama K

    (1993) A mutation of furin causes the lack of precursor-processing activity in human colon carcinoma LoVo cells

    Biochemical and Biophysical Research Communications 195:1019–1026.

    https://doi.org/10.1006/bbrc.1993.2146

    • PubMed
    • Google Scholar
43. 1. Taylor FR
    2. Wen D
    3. Garber EA
    4. Carmillo AN
    5. Baker DP
    6. Arduini RM
    7. Williams KP
    8. Weinreb PH
    9. Rayhorn P
    10. Hronowski X
    11. Whitty A
    12. Day ES
    13. Boriack-Sjodin A
    14. Shapiro RI
    15. Galdes A
    16. Pepinsky RB

    (2001) Enhanced potency of human Sonic hedgehog by hydrophobic modification

    Biochemistry 40:4359–4371.

    https://doi.org/10.1021/bi002487u

    • PubMed
    • Google Scholar
44. 1. Torroja C
    2. Gorfinkiel N
    3. Guerrero I

    (2004) Patched controls the Hedgehog gradient by endocytosis in a dynamin-dependent manner, but this internalization does not play a major role in signal transduction

    Development 131:2395–2408.

    https://doi.org/10.1242/dev.01102

    • PubMed
    • Google Scholar
45. 1. Tukachinsky H
    2. Kuzmickas RP
    3. Jao CY
    4. Liu J
    5. Salic A

    (2012) Dispatched and scube mediate the efficient secretion of the cholesterol-modified hedgehog ligand

    Cell Reports 2:308–320.

    https://doi.org/10.1016/j.celrep.2012.07.010

    • PubMed
    • Google Scholar

Author details

1. Daniel P Stewart

   Department of Cell and Molecular Biology, St. Jude Children’s Research Hospital, Memphis, United States

   Contribution

   Data curation, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing—original draft, Writing—review and editing

   Contributed equally with

   Suresh Marada

   Competing interests

   No competing interests declared

2. Suresh Marada

   Department of Cell and Molecular Biology, St. Jude Children’s Research Hospital, Memphis, United States

   Contribution

   Data curation, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing—original draft, Project administration, Writing—review and editing

   Contributed equally with

   Daniel P Stewart

   For correspondence

   suresh.marada@stjude.org

   Competing interests

   No competing interests declared

3. William J Bodeen

   1. Department of Cell and Molecular Biology, St. Jude Children’s Research Hospital, Memphis, United States
   2. Integrated Program in Biomedical Sciences, University of Tennessee Health Sciences Center, Memphis, United States

   Contribution

   Data curation, Validation, Investigation, Methodology, Writing—original draft

   For correspondence

   William.Bodeen@STJUDE.ORG

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-0557-4826

4. Ashley Truong

   Department of Cell and Molecular Biology, St. Jude Children’s Research Hospital, Memphis, United States

   Contribution

   Data curation, Formal analysis, Validation, Investigation, Methodology

   For correspondence

   Ashley.Truong@STJUDE.ORG

   Competing interests

   No competing interests declared

5. Sadie Miki Sakurada

   1. Department of Cell and Molecular Biology, St. Jude Children’s Research Hospital, Memphis, United States
   2. Center for Advanced Genome Engineering, St. Jude Children’s Research Hospital, Memphis, United States

   Contribution

   Resources, Formal analysis, Validation, Investigation

   For correspondence

   sadie.sakurada@stjude.org

   Competing interests

   No competing interests declared

6. Tanushree Pandit

   Department of Cell and Molecular Biology, St. Jude Children’s Research Hospital, Memphis, United States

   Contribution

   Data curation, Formal analysis, Validation, Methodology

   For correspondence

   Tanushree.pandit@STJUDE.ORG

   Competing interests

   No competing interests declared

7. Shondra M Pruett-Miller

   1. Department of Cell and Molecular Biology, St. Jude Children’s Research Hospital, Memphis, United States
   2. Center for Advanced Genome Engineering, St. Jude Children’s Research Hospital, Memphis, United States

   Contribution

   Resources, Data curation, Formal analysis, Methodology, Writing—review and editing

   For correspondence

   shondra.miller@stjude.org

   Competing interests

   No competing interests declared

8. Stacey K Ogden

   Department of Cell and Molecular Biology, St. Jude Children’s Research Hospital, Memphis, United States

   Contribution

   Conceptualization, Formal analysis, Supervision, Funding acquisition, Writing—original draft, Project administration, Writing—review and editing

   For correspondence

   stacey.ogden@stjude.org

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0001-8991-3065

• 2,658

  views

• 425

  downloads

• 25

  citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.