Deltex modulates Dpp morphogen gradient formation and affects Dpp signaling in Drosophila

Decapentaplegic (Dpp) is a member of the TGF-β superfamily of signaling molecules, similar to vertebrate bone morphogenetic proteins 2 and 4 (BMP2 and BMP4, respectively), and is involved in many developmental processes, ranging from cell division to cell fate determination (Affolter and Basler, 2007; Kicheva and González-Gaitán, 2008; Perrimon et al., 2012). Any alterations in TGF-β superfamily signaling pathways often result in human disease, including developmental disorders, vascular diseases and cancer (Blobe et al., 2000; Massagué et al., 2000). Proper regulation of Dpp signaling is, therefore, indispensable for its function, and hence multiple genes potentially regulate the Dpp signaling cascade.

Dpp is expressed in a narrow stripe of anterior cells at the anterior–posterior (A–P) boundary of the wing disc (Affolter and Basler, 2007; Basler and Struhl, 1994). The secreted Dpp dimer binds and activates the receptor complex, comprising Thickveins (Tkv) and Punt (Put), which in turn phosphorylates Mothers against Dpp (Mad), the downstream target of Tkv–Put kinase, along the Dpp domain in the wing disc (Affolter and Basler, 2007). Once phosphorylated, Mad forms a complex with a co-mediator, Medea, and translocates to the nucleus, thereby activating the transcription of downstream target genes, such as spalt (sal, also known as salm) and optomotor-blind (omb, also known as bifid), at different extracellular concentrations of Dpp (Lecuit et al., 1996; Shi and Massagué, 2003).

Since different concentrations of Dpp activate expression of different target genes, it is necessary to understand the mechanism of Dpp gradient formation. A promising model of Dpp gradient establishment has been proposed by Entchev et al. (2000), wherein the Dpp gradient is formed via intracellular trafficking initiated by receptor-mediated endocytosis of the ligand in the receiving cells, and the gradient is maintained by endocytic sorting of Dpp towards recycling versus degradation. A similar mechanism of Notch signaling regulation has also been reported, in which Deltex (Dx) is required for transportation of Notch from the cell membrane to the endosomal vesicle, where the Notch protein is either recycled or degraded depending upon its ubiquitylation state (Hori et al., 2004; Yamada et al., 2011).

Dx is a cytoplasmic interactor of Notch and regulates Notch signaling in a context-dependent manner (Diederich et al., 1994; Dutta et al., 2017; Matsuno et al., 1995, 1998; Mishra et al., 2014). The E3 ubiquitin ligase activity of Dx plays an important role in determining Notch stability by mono- or poly-ubiquitylating the protein depending on the cellular context (Baron, 2012; Hori et al., 2011; Mukherjee et al., 2005; Wilkin et al., 2008). In a more recent study, Dx has been reported as an interactor of Eiger, a Drosophila homolog of tumor necrosis factor (TNF), and is involved in Eiger-mediated activation of JNK signaling (Dutta et al., 2018). Moreover, we have also reported previously that Dx synergizes with Traf6, an adaptor molecule in the JNK cascade, and activates JNK signaling at a level downstream of the ligand–receptor interaction (Sharma et al., 2020). A novel function of Dx in Toll pathway activation has also been reported recently, where Dx has been shown to activate the Toll pathway in a JNK-independent manner (Sharma et al., 2021).

In this study, we report that dx genetically interacts with different alleles of Dpp pathway components during wing development. Moreover, the Dpp mutant wing phenotype was seen to be modulated in a Dx overexpression background. Immunocytochemical studies further revealed that Dpp and its receptor Tkv localize with Dx in the same cytoplasmic vesicle. The expression of Dpp, and that of its downstream targets Mad and Spalt, was also observed to be modulated upon Dx overexpression, and we attribute this modulation to changes in Dx-mediated Dpp trafficking.

Kinases and phosphatases have the potential of switching a gene on and off and, in turn, regulate associated signaling cascades. To check whether kinases have some role in Dx-associated phenotypes, we performed an RNAi-based modifier screen of kinases and phosphatases that regulate a broad array of biological and cellular processes. We identified tkv as a potent interactor of Dx (Fig. S1). When tkv was downregulated in the dorsal–ventral boundary, a crumpled wing phenotype was observed. This was rescued appreciably when dx was overexpressed in the same background (Fig. S1C).

Tkv is the receptor of the ligand Dpp, and hence it was intriguing to study the role of Dx in the Dpp pathway. Dpp signaling plays a prominent role in wing development, and a reduction in Dpp level affects the differentiation of veins and cross-veins (Segal and Gelbart, 1985; Spencer et al., 1982). Loss-of-function mutation in either the ligand or the receptor, or at the transcription factor level of the Dpp pathway disturbs wing vein formation and wing growth (Affolter and Basler, 2007). Conversely, ectopic activation of Dpp signaling results in the formation of extra veins (Affolter and Basler, 2007; Eivers et al., 2009; Sotillos and De Celis, 2005). The dx gene, on the other hand, shows haplosufficiency, with heterozygous females not exhibiting any visible phenotype. However, hemizygous flies were observed to show wing vein thickening (Fig. 1A).

To address the functional relationship between dx and dpp, we further investigated whether mutations in dx and dpp display genetic interactions. It is interesting to note that both the dpp mutant alleles tested (dpp⁵ and dpp⁶) showed lethality in homozygous conditions (data not shown); however, in heterozygous conditions, they did not exhibit any phenotype (Fig. 1B,C). When a loss-of-function allele of dpp was brought in trans-heterozygous combination with dx (dx/+; dpp/+), the resulting flies had wings that were indistinguishable from those of wild-type flies (data not shown). However, reducing the dose of dpp in a dx-null background (dx¹⁵²) resulted in enhanced notching and wing vein thickening when compared with the wings of dx hemizygotes in an otherwise wild-type background (Fig. 1B,C). Given that genetic interaction between dx and dpp was observed in the absence of all dx functions, it is explicit that a complete absence of dx creates a sensitized genetic background that makes the development of the wing margin sensitive to a decrease in the dosage of dpp.

We also checked whether dx interacts with other Dpp pathway components, namely the receptor tkv, and the downstream target Mad. The loss-of-function alleles of tkv (tkv⁸ and tkv¹²) and Mad (Mad^k00237) resulted in wing nicking and vein thickening phenotypes in the trans-heterozygous condition with dx hemizygotes (Fig. 1D–F). Moreover, we observed a spread out wing phenotype with dx hemizygotes in the dpp heterozygous condition (Fig. 1G). The graph (Fig. 1H) highlights the frequency of wing notching phenotypes observed in the indicated genetic combinations. Taken together, these results suggest that dx and dpp genetically interact, which may have functional implications for wing development.

To investigate further, we checked the subcellular localization of Dpp and Dx. Dpp being a morphogen means that it shows diffuse expression in endogenous conditions (Teleman and Cohen, 2000); however, when co-expressed with Dx, a punctate pattern of Dpp was observed that strongly colocalized with Dx-positive vesicles in the same subcellular compartment (Fig. 2A). Moreover, we also observed a sparse localization of smaller endogenous Dpp puncta with FLAG-tagged UAS-Dx (Fig. S2A). Interestingly, with endogenous Dx, a less punctate Dpp expression pattern was observed (Fig. S2B). We also tried to check the localization status of Tkv along with Dx, and we found a strong colocalization when Dx and Tkv were overexpressed with dpp-GAL4 in the same subcellular compartment (Fig. 2B).

Previous reports have shown the potent role of Dx in vesicular trafficking of Notch from the cell membrane to the cytoplasmic vesicles (Hori et al., 2004; Wilkin et al., 2008); therefore, we hypothesized that Dx can bind with Dpp and facilitate trafficking of Dpp, like that of Notch. To address this hypothesis, it was crucial to check whether Dpp physically interacts with Dx. Co-immunoprecipitation experiments using FLAG-tagged Dx revealed that FLAG-tagged Dx pulled down Dpp–GFP when co-expressed in eye tissue (Fig. 2C). Likewise, FLAG-tagged Dx was coimmunoprecipitated with Dpp–GFP using an anti-GFP antibody when the two proteins were expressed together (Fig. 2D). Additionally, we also checked whether Dx physically interacts with the receptor Tkv and found that FLAG-tagged Dx pulled down Tkv–GFP when co-expressed in the eye tissue (Fig. 2E).

Taken together, our data suggest that there is a direct interaction of Dx with Dpp and its receptor Tkv, and we hypothesize that Tkv possibly facilitates the binding of cytoplasmic Dx to Dpp. The functional implications of this interaction are a crucial aspect for further investigation.

Dpp overexpression using C96-GAL4 as a driver resulted in an enhanced wing vein thickening and ectopic vein formation phenotype (Fig. 3A, panel 3). Interestingly, induction of Dx expression in flies overexpressing Dpp resulted in a reduction of vein thickening and ectopic vein formation (Fig. 3A, panel 4). Furthermore, lowering the dose of tkv in the wing margin induced wing serration, which was significantly rescued when Dx was overexpressed in the same background (Fig. 3B; n=100). Conversely, reducing the levels of Dx in flies overexpressing Dpp or underexpressing Tkv did not result in a change in phenotype (data not shown). We hereby hypothesize that overexpression of Dx might facilitate the proper distribution of Dpp. The results presented here indicate a potential role of Dx in modulating the phenotypes associated with Dpp signaling.

To further investigate the involvement of Dx in Dpp signaling regulation, we tried to observe whether the expression of Dpp and its targets is affected by dx activity. Dpp organizes the Drosophila wing patterning by acting as a graded morphogen. The gene is transcribed in a narrow stripe along the A–P compartment boundary of the wing imaginal disc. However, when FLAG-Dx was overexpressed in the dorsal compartment of the wing imaginal disc, a broader Dpp stripe was seen, suggesting a wider gradient of Dpp, compared to that observed in wild-type wing imaginal discs (Fig. 3C). A similar broadening of the Dpp expression domain was observed when Dx was overexpressed using A–P domain-specific dpp-GAL4 (Fig. S3).

In addition, upon expression of Dx at the A–P border cells using dpp-GAL4, we observed abnormal alteration of the distribution of phosphorylated Mad (pMad) (Fig. 3D), which is the signal transducer downstream of Dpp. pMad levels in general are high in the central region of the wing disc and decline gradually towards the anterior and posterior distal cells. As there are no other known transducers of Dpp signaling, pMad levels are taken to reflect the intensity of Dpp signal transduction activity. pMad regulates the expression of Dpp target genes (including spalt, optomotor-blind, and vestigial) in a Dpp-dependent manner in domains that straddle the Dpp stripe, with Spalt having the narrowest domain and Vestigial the widest (Hamaratoglu et al., 2014). Of the abovementioned markers, we checked the expression of Spalt in wing imaginal discs overexpressing Dx. Spalt is expressed in the wing pouch in a broad stripe of cells centered along the A–P compartment boundary (de Celis et al., 1996; Nellen et al., 1996). An ectopic Spalt expression was observed in the posterior compartment of the disc where Dx was overexpressed using en-GAL4 (Fig. 3E). In addition to the ectopic expression, the localization of Spalt was also altered upon Dx overexpression, resulting in a more elongated region of Spalt expression with a tapering end towards the ventral region of the disc.

These results indicate that Dx may facilitate Dpp signaling by enhancing the trafficking of the morphogen Dpp, thereby altering the expression of effector target genes.

Although Dpp is one of the most extensively studied morphogens, the molecular mechanism by which the Dpp gradient is formed and sustained is poorly understood. Endocytosis of Dpp, however, is proposed to play a potent role in formation of the Dpp gradient (Entchev et al., 2000). Dx plays a key role in receptor-mediated endocytosis of Notch (Hori et al., 2004, 2011). Dx depletes Notch from the apical cell surface and leads to accumulation of Notch in the late endosomal vesicles. Furthermore, if transport of Notch into the late endosome is impaired, Dx- mediated Notch signal activation is hampered (Hori et al., 2004). These previous reports prompted us to study the role of Dx in Dpp trafficking.

We first tried to check whether Dx colocalizes with key Rab components along with Dpp, which would indicate that all three molecules are present in the same vesicle. We overexpressed dpp-GFP and FLAG-Dx in third-instar wing imaginal discs using dpp-GAL4 and checked for colocalization of the two proteins with the early endosome marker Rab5. Since we already observed that Dx and Dpp do physically interact (Fig. 2), we focused on whether Dx and Dpp-positive vesicles were also positive for Rab5. Our results showed that Dx colocalizes with Dpp and Rab5 (Fig. S4), and it is possible that these three proteins form a trimeric complex. We next questioned whether this putative complex has some role in Dpp trafficking. We centered our study on Rab7, since Rab7 targets endocytic cargo from early endosomes to late endosomes and lysosomes for degradation (Méresse et al., 1995; Vitelli et al., 1997), and Dx promotes accumulation of Notch in the late endosomal compartment, where Dx colocalizes with Notch and Rab7 (Hori et al., 2004). We tried to monitor the Dpp signaling range with respect to Dx. For this, we checked the status of Dpp in a Rab7 overexpression background. A reduction in the width of the Dpp gradient was observed upon ectopic expression of Rab7; however, this reduction was rescued when Dx was overexpressed in the same background (Fig. 4A–C). Moreover, we also visualized the expression of the Dpp target Spalt. When UAS-Rab7 was ectopically expressed in the posterior compartment of the wing disc, a reduction in Spalt expression was observed (Fig. 4D,E) (Entchev et al., 2000), indicating that sorting of endocytic cargo towards degradation limits the range of Dpp signaling. This reduction in Spalt expression upon Rab7 overexpression was, however, rescued when Dx was co-expressed in the same compartment of the wing disc (Fig. 4F).

In conclusion, we have uncovered an important function of the cytoplasmic protein Dx in the formation of gradients of the morphogen Dpp. Our studies show that Dx genetically and physically interacts with Dpp and Tkv. Our results further reveal that Rab7, along with Dx, expands the gradient of Dpp and its target Spalt. We attribute this expansion to Dx-mediated Dpp trafficking; however, further analysis is required to unravel the detailed mechanistic aspects of Dpp endocytosis and trafficking mediated by Dx.

All fly stocks were maintained on standard medium containing cornmeal, yeast, molasses and agar at 25°C, as per standard procedures. UAS-FLAG-Dx, dx, and dx¹⁵² flies were kindly provided by Professor Spyros Artavanis-Tsakonas (Department of Cell Biology, Harvard Medical School, USA). tkv⁸ and tkv¹² flies were kind gift from Professor Yu Cai (National University of Singapore, Singapore). The UAS-tkv-GFP line was obtained from Professor Thomas Kornberg (University of California San Francisco, USA). dpp⁵ (BL-20620), dpp⁶ (BL-2071), Mad^k00237 (BL-10474), UAS-dpp-GFP (BL-53716), UAS-Rab7-GFP (BL-42706), en-GAL4, dpp-GAL4, ap-GAL4, GMR-GAL4 and C96-GAL4 stocks were obtained from Bloomington Drosophila Stock Center. UAS-tkv-IR (VDRC-105834) was obtained from Vienna Drosophila Resource Center.

All crosses were performed at 25°C unless otherwise mentioned. To induce the expression of genes in a specific domain, the GAL4–UAS binary system was used (Brand and Perrimon, 1993). Combination stocks were made with help of appropriate genetic crosses.

For immunoprecipitation of Dx and Dpp, protein lysates were prepared in 1× RIPA buffer (Millipore, #20-188) from Drosophila head tissue expressing UAS-FLAG-Dx, UAS-Dpp-GFP, or both UAS-Dpp-GFP and UAS- FLAG-Dx using a GMR-GAL4 driver. Crude lysates containing 2 mg of total protein were mixed with 5 µl of rabbit anti-FLAG antibody (Sigma, #F7425) or rabbit anti-GFP antibody (Invitrogen, #PA1-980A) and 30 µl of protein A/G beads before being incubated overnight with end-over-end rotation at 4°C. UAS-dpp-GFP or UAS-FLAG-Dx lysates were used as control samples. Beads were collected after washing thrice with 1× RIPA buffer and separated on a 12% denaturing SDS polyacrylamide gel followed by transfer onto Immuno-blot PVDF membranes (Bio-Rad). After washing for 10 min in TBST (50 mM Tris base, 150 mM NaCl, 0.1% Tween-20) and blocking (4% skimmed milk in TBST) for 30 min, blots were probed with rabbit anti-GFP antibody at 1:1500 dilution or rabbit anti-FLAG antibody at 1:1000 dilution. Again, after washing thrice in TBST and blocking for 1 h, goat anti-rabbit IgG–AP conjugate at 1:2000 dilution (Molecular Probes) in blocking solution was added for 90 min, followed by three washes in TBST. Colorimetric detection was performed using Sigma FAST BCIP/NBT. A similar approach was followed to perform immunoprecipitation of Dx and Tkv.

Drosophila third-instar larval wing discs were dissected out in ice-cold 1× PBS, and tissues were fixed in a 1:1 mixture of 3% paraformaldehyde in PBS at room temperature for 1 min, followed by a second fixation in 3% paraformaldehyde and 5% DMSO for 20 min. Immunostaining was performed as described previously (Sharma et al., 2020). The following primary antibodies were used in this study: rabbit anti-FLAG (1:00; Sigma, #F7425), rabbit anti-phospho-Smad1/5 (1:50; Cell Signaling Technology, #41D10), rabbit anti-Dpp (1:100; kindly gifted by Professor Matthew Gibson, Stowers Institute for Medical Research, Kansas City, MO, USA), guinea pig anti-Spalt (1:20,000; kindly gifted by Professor Antonia Monteiro, National University of Singapore, Singapore), guinea pig anti-Rab5 (1:1000; a generous gift from Professor Akira Nakamura, Institute of Molecular Embryology and Genetics, Kumamoto, Japan). Alexa Fluor 488-, Alexa Fluor 555- or Alexa Fluor 405-conjugated secondary antibodies (1:200, Molecular Probes) were used to detect the primary antibodies. Imaging was performed using a Carl Zeiss LSM 780 laser scanning confocal microscope with 20×, 40× and 63× objectives, and images were processed using Adobe Photoshop 7.

Intensity profiles were generated using ImageJ software (NIH, Bethesda, MD, USA). The average intensity for each image was measured using the plot function. For measuring gradient profiles in ImageJ, we used average intensity projections of single line profiles for each image in triplicates. Representative plots of intensity projections are shown in respective figures. Gradient profiles were extracted using ImageJ software. For Fig. 3 and Fig. S3, the entire pouch up to the edge of the wing disc was measured. For Fig. 4, the pouch region was measured. Measurement of the Spalt (Sal) domain was done taking into consideration the A–P boundary.

The authors extend sincere thanks to Professor Spyros Artavanis-Tsakonas, Professor Yu Cai, Professor Thomas Kornberg, Professor Matthew Gibson, Professor Antonia Monteiro, Professor Akira Nakamura, the Vienna Drosophila Resource Center and the Bloomington Drosophila Stock Center for fly stocks and antibodies. Some of the antibodies used in the work were obtained from Developmental Studies Hybridoma Bank, University of Iowa. We also acknowledge the confocal facility of DBT-BHU-ISLS, Banaras Hindu University.