XB-ART-40165PLoS One August 6, 2009; 4 (8): e6543.
Mad is required for wingless signaling in wing development and segment patterning in Drosophila.
A key question in developmental biology is how growth factor signals are integrated to generate pattern. In this study we investigated the integration of the Drosophila BMP and Wingless/GSK3 signaling pathways via phosphorylations of the transcription factor Mad. Wingless was found to regulate the phosphorylation of Mad by GSK3 in vivo. In epistatic experiments, the effects of Wingless on wing disc molecular markers (senseless, distalless and vestigial) were suppressed by depletion of Mad with RNAi. Wingless overexpression phenotypes, such as formation of ectopic wing margins, were induced by Mad GSK3 phosphorylation-resistant mutant protein. Unexpectedly, we found that Mad phosphorylation by GSK3 and MAPK occurred in segmental patterns. Mad depletion or overexpression produced Wingless-like embryonic segmentation phenotypes. In Xenopus embryos, segmental border formation was disrupted by Smad8 depletion. The results show that Mad is required for Wingless signaling and for the integration of gradients of positional information.
PubMed ID: 19657393
PMC ID: PMC2717371
Article link: PLoS One
Genes referenced: cnga1 dspp egfr egr2 gsk3b lgals4.2 mad1l1 mapk1 mlc1 myod1 nkd1 otx2 ptch1 rax smad1 smad7 smad9 sox2 szl tbx2 wnt10b wnt3a
Morpholinos: smad9 MO1
Article Images: [+] show captions
|Figure 7. Smad5/8 Is Required for Segment Border Formation in Xenopus Embryos. (A) Illustration of C3 Xenopus blastomere injection at the 32 cell stage. The fate of C3 cells in the somites is indicated in a stage 28 tadpole. (B and C) In Xenopus, microinjection of Smad8-MO at the 16 or 32 cell stage (in C2 or C3 blastomeres) erased segmental somite borders on the injected side. Somites are composed mostly of segmental muscles, which were stained for myosin light chain (a-MLC). (D-F0) Smad8-MO effects on segment borders were cell autonomous (co-injection at 32-cell stage with rhodamine dextran amine lineage tracer), and were rescued by human Smad1 mRNA (n = 55, 44 with somite fusions, n = 19, with 17 completely rescued, n = 24, all normal, respectively, 3 independent experiments).|
|The Mad12 mutant, which lacks the C-terminal phosphorylation sites, retains posteriorizing activity in Xenopus, in particular when the GSK3 sites are also mutated. (A) Whole mount in situ hybridization of tail bud stage Xenopus embryos (n = 16). Embryos are stained for Rx2a (eye), Krox20 (Hindbrain) and Sizzled (Ventral/Belly). (B) Microinjection of Mad12 mRNA reduced the anterior head region of the embryo, indicated by decreased Rx2a expression (n = 15). (C) Elimination of the anterior head structures in MGM12 microinjected mRNAs (mutation of the GSK3 phosphorylation sites mimics Mad receiving a maximal Wg signal) resulted in a severely posteriorized embryo with almost complete loss-of Rx2a expression (n = 13). (D) Similar posteriorized phenotypes are generated when Wnt10b DNA is microinjected into Xenopus embryos. In addition to this, there is an increase in the expression of the BMP responsive gene sizzled, presumably because it affects the stability of endogenous Smad1/5/8 (n = 7).|
|Knockdown of xSmad8 Dorsalizes Xenopus Embryos. (A) Whole-mount in situ hybridization for the pan-neural marker Sox2 in an uninjected control embryo, stage 22, dorsal view. (B) Injection of xSmad8 morpholino (0.5 mM, 4 nl injected four times radially) leads to expansion of the neural plate. (C) Control embryo stained for Otx2 (forebrain and midbrain marker) and Krox20 (hindbrain, rhombomeres 3 and 5), lateral view. (D) Smad8-depleted embryos are dorsalized (anti-BMP phenotype) and show expansion of head structures. It should be mentioned here that the original depletion of Xenopus laevis Smad8 by Miyanaga et al  yielded a very different result, namely apoptosis via activation of caspases. However, it should be noted that their methods for depletion were different. They used DNA oligonucleotides to deplete Smad8 transcripts in oocytes that were then subjected to maternal transfer and fertilization. We used morpholino oligos (of a different sequence) injected at the 4-cell stage, and therefore the depletion of maternal transcripts must have been less extensive. This explains why we did not observe apoptosis, but instead dorsalization (anti-BMP phenotype) of the embryo. The morpholino described here provides a useful reagent for knockdown of the maternal Xenopus laevis BMP-Smad. Smad8 probably corresponds to the homolog of zebrafish Smad5  and is therefore referred to below as Smad5/8.|
|GSK3-resistant Activated Forms of Smad1 Disrupt Segmentation in Xenopus Embryos. (A and B) Immunostainings for myosin light chain showing loss of segmental borders in somites in the injected side (B) of a Xenopus embryo (stage 26), compared to the uninjected side. C2 or C3 blastomeres were injected with 100 pg of hSmad1 resistant to GSK3 phosphorylation  and 50 pg of nuclear LacZ mRNA (n = 32/42). (C) In situ hybridizations for the somite marker MyoD shows a disruption of the segmental pattern in Xenopus embryos (stage 30) injected with activated forms of Smad1. Uninjected control embryos express MyoD in the somitic segments in a typical chevron shape (n = 27). Overexpression of Smad1 wild-type mRNA (SWT) does not change this pattern (n = 21). An activated mutant of Smad1 that has phospho-mimetic amino acid substitutions on the C-terminus (the two most c-terminal serines mutated into glutamic acids, designated EVEmutant ) displays mild disruptions of the somite pattern (n = 9/12), while the same phospho-mimetic form of Smad1 with an additional mutation in the GSK3-phosphorylation site in the linker (named EVE-GM exhibits strongly impaired segmentation (n = 14/18). Nuclear LacZ marks the injected cells.|
|Figure 1. Phosphorylation-Resistant Mad Proteins are Hyperactive.(A) Model summarizing the integration of Dpp, EGFR and Wg signaling at the level of Mad phosphorylations in Drosophila. (B) Diagrams of Mad Wild Type (MWT), Mad MAPK Mutant (MMM) and Mad GSK3 Mutant (MGM) proteins. (C–F) Microinjection of MMM and MGM mRNAs into Xenopus embryos had stronger ventralizing activity than MWT, causing upregulation of sizzled (n = 17, 32, 26, and 30, two independent experiments). Brain markers otx2 and krox20 were repressed. (G–J) Driving MMM and MGM with patched-Gal4 in the anterior wing compartment caused formation of ectopic crossvein-like tissue. This tissue links longitudinal veins two and three in both proximal and distal regions, pulling the two veins closer together. (K–N) Driving phosphorylation-resistant Mads with apterous-Gal4 induced ectopic vein tissue and blistering, indicating increased Dpp signaling. (O) Polyubiquitinylation of Mad requires GSK3 and MAPK phosphorylation sites. Lane 1, 293T cells cotransfected with MWT-Flag, Drosophila Smurf and HA-ubiquitin all cloned in pCS2. The strong smear represents polyubiquitinylated Mad tagged with HA-ubiquitin. Lanes 2 and 3, polyubiquitinylation was greatly decreased in the MMM and MGM mutant proteins. The lower panel shows equal levels of immunoprecipitated Mad (α-Flag).|
|Figure 2. Mad GSK3 Mutants Mimic Wg Overexpression.(A–C) Additional bristles are induced in the wing margin by MGM driven by scalloped-Gal4. Arrows indicate chemosensory bristles and arrowheads stout mechanosensory bristles. Insets show that the number of Senseless-expressing bristle precursors in wing imaginal discs is increased by MGM, but not MWT.(D–G) Expression domains of sd-Gal4 driver, Wg protein, Armadillo stabilized by Wg (expression in the proveins is noted), and Senseless in wing discs. (H and I) Senseless protein expression was inhibited in Mad RNAi clones marked by GFP. (J–O) Quantitative RT-PCR of wing discs showing that MGM increased both a Wg target gene (senseless) and Dpp target genes (spalt and optomotor blind), while not affecting wg or hh levels. dpp was inhibited by MGM RNA. Samples were normalized for rp49, except for the inset in L in which Gal4 mRNA was used. (P and Q) Clonal overexpression of MGM does not change Wg levels. (R and S) MGM clone in the anterior margin causes ectopic expression of Senseless within the clone (inset). (T) Overexpression of MGM in clones marked by yellow induced duplications of the wing margin (n = 24). Yellow arrowheads indicate ectopic margins and a black arrowhead the wild type one. (U) Knockdown of Mad in clones partially eliminates the wing margin In this large clone the remaining margin bristles are yellow (y).|
|Figure 3. Phospho-Specific Antibodies Reveal Wg-regulated Segmental Expression Patterns of pMadMAPK and pMadGSK3.(A and B) Western blot analysis of pMadMAPK and pMadGSK3 antibodies demonstrating that they were phospho-specific, and that GSK3 phosphorylation had an absolute requirement for MAPK priming. Drosophila S2 cells were transiently transfected with the plasmids indicated. (C) Cultured 293T cells stably transfected with Mad-Flag treated with L-cell control conditioned medium (CMed), Wnt3a medium, control DMEM (Con), or 30 mM LiCl in DMEM for 2 hours. Wnt3a and LiCl inhibited the MadGSK3 phosphorylation band and increased β-Catenin levels (indicating that the Wnt treatment was effective). (D and E) pMadMAPK and pMadGSK3 antibodies stain the entire blastoderm and a Dpp-dependent dorsal stripe (inset). (F) pMadMAPK tracks ventral EGFR-activated MAPK (inset shows diphospho-Erk staining). (G and H) Segmental staining of pMadMAPK (Stage 9) and pMadGSK3 (Stage 17). (I) In Wg null mutants segmental expression is lost. Mutant embryos were identified by lack of staining with Wg antibody. Inset shows same embryo stained with DAPI to indicate that, despite its abnormal shape, it reached late stages of development. (J–L) Wg stabilizes pMadMAPK, overlapping with Wg stripes. (M–O) Nuclear pMadGSK3 accumulates in between Wg stripes, indicating that Wg inhibits Mad phosphorylation at GSK3 sites in vivo. (P–R) Wg overexpression driven with prd-Gal4 stabilizes pMadMAPK over a broader domain compared to just MWT alone (compare brackets in P and R). This experiment shows that Wg expression stabilizes pMADMAPK.|
|Figure 4. Mad10 and Mad12 Alleles Are not Nulls; Mad RNAi Is an Effective and Specific Loss-of-Function Reagent.A) Schematic representation of Mad, showing RNAi and mutant sites. (B–E) Mad10 and Mad12 mutants are phosphorylated in the linker region in the absence of C-terminal phosphorylation. (F) UAS-Mad RNAi depleted stage 15 embryos of endogenous pMadCter when driven by daughterless-Gal4. (G) Repression of brinker-LacZ reporter by Dpp (demarcated by hatched lines) was inhibited by Mad RNAi in wing imaginal discs. (H–K) Mad RNAi driven by MS1096-Gal4 causes complete vein loss at room temperature, which was rescued by UAS-hSmad1. (L and M) Anterior margin mechanosensory bristles are lost when two copies of Mad RNAi were driven with A9-Gal4; this phenocopies Wg loss-of-function. (N–Q) Epistasis by QRT-PCR showing that Wg overexpression in the wing pouch, driven by sd-Gal4, increased transcript levels of the reporter genes optomotor blind, senseless, distalless and vestigial, and that this induction required Mad Samples were normalized for Gal4 mRNA levels Inset shows that levels of Wg transcripts were not affected by Mad RNAi.|
|Figure 5. Mad Is Epistatic to Wg Signaling During Neurogenic Induction.(A) Mad was phosphorylated at its MAPK sites in developing CNS neuroblasts. (B) pMadCter was excluded from the neurogenic ectoderm marked by SoxNeuro (stage 8). (C and D) Mad RNAi driven in the egg by pUASp knocked down pMadMAPK staining (stage 7). (E and F) Maternal Mad RNAi knocked down pMADMAPK centrosomal staining (stage 7, see Figure S5 for asymmetric centrosomal staining). (G–J) Mad RNAi increased neurogenic ectodermal nuclei marked by SoxNeuro at stage 8, Wg overexpression reduced it, and the double Wg;RNAi embryos displayed the Mad depletion phenotype. All images were taken at identical exposure conditions. This experiment shows that Mad is epistatic to Wg.|
|Figure 6. Mad Is Required for Segmentation in Drosophila.(A–B) Early depletion of Mad caused wider (ventralized) denticle belts and internalized posterior spiracles in embryonic cuticles (n = 259 cuticles, 20% ventralized and 34% ventralized with denticle belt fusions). (C and C') Denticle belt fusions showing large (row 5-like) denticles. (D and D') Wg loss-of-function caused a ventral lawn of denticles. Note that these are large denticles with a small refringent spot (row 5 denticles) resembling those seen in Mad RNAi depletion. (E) DppH46 mutant embryo showing fusion of two denticle belts. (F and F') Overexpression of UAST-MGM driven by mat-Gal4-VP16 caused patches of naked cuticle at the expense of denticle rows. (G–I) Embryos stained for Engrailed at stage 9, showing that Mad depletion disrupts abdominal segmental bands, while MGM overexpression does not.|