Viswanathan PK et al. (2024), Maternal TGF-β ligand Panda breaks the radial s...

ECB-ART-53166

PLoS Biol 2024 Jun 01;226:e3002701. doi: 10.1371/journal.pbio.3002701.

Show Gene links Show Anatomy links

Maternal TGF-β ligand Panda breaks the radial symmetry of the sea urchin embryo by antagonizing the Nodal type II receptor ACVRII.

Viswanathan PK , Chessel A , Molina MD , Haillot E , Lepage T .

???displayArticle.abstract???
In the highly regulative embryo of the sea urchin Paracentrotus lividus, establishment of the dorsal-ventral (D/V) axis critically depends on the zygotic expression of the TGF-β nodal in the ventral ectoderm. nodal expression is first induced ubiquitously in the 32-cell embryo and becomes progressively restricted to the presumptive ventral ectoderm by the early blastula stage. This early spatial restriction of nodal expression is independent of Lefty, and instead relies on the activity of Panda, a maternally expressed TGF-β ligand related to Lefty and Inhibins, which is required maternally for D/V axis specification. However, the mechanism by which Panda restricts the early nodal expression has remained enigmatic and it is not known if Panda works like a BMP ligand by opposing Nodal and antagonizing Smad2/3 signaling, or if it works like Lefty by sequestering an essential component of the Nodal signaling pathway. In this study, we report that Panda functions as an antagonist of the TGF-β type II receptor ACVRII (Activin receptor type II), which is the only type II receptor for Nodal signaling in the sea urchin and is also a type II receptor for BMP ligands. Inhibiting translation of acvrII mRNA disrupted D/V patterning across all 3 germ layers and caused acvrII morphants to develop with a typical Nodal loss-of-function phenotype. In contrast, embryos overexpressing acvrII displayed strong ectopic Smad1/5/8 signaling at blastula stages and developed as dorsalized larvae, a phenotype very similar to that caused by over activation of BMP signaling. Remarkably, embryos co-injected with acvrII mRNA and panda mRNA did not show ectopic Smad1/5/8 signaling and developed with a largely normal dorsal-ventral polarity. Furthermore, using an axis induction assay, we found that Panda blocks the ability of ACVRII to orient the D/V axis when overexpressed locally. Using co-immunoprecipitation, we showed that Panda physically interacts with ACVRII, as well as with the Nodal co-receptor Cripto, and with TBR3 (Betaglycan), which is a non-signaling receptor for Inhibins in mammals. At the molecular level, we have traced back the antagonistic activity of Panda to the presence of a single proline residue, conserved with all the Lefty factors, in the ACVRII binding motif of Panda, instead of a serine as in most of TGF-β ligands. Conversion of this proline to a serine converted Panda from an antagonist that opposed Nodal signaling and promoted dorsalization to an agonist that promoted Nodal signaling and triggered ventralization when overexpressed. Finally, using phylogenomics, we analyzed the emergence of the agonist and antagonist form of Panda in the course of evolution. Our data are consistent with the idea that the presence of a serine at that position, like in most TGF-β, was the ancestral condition and that the initial function of Panda was possibly in promoting and not in antagonizing Nodal signaling. These results highlight the existence of key functional and structural elements conserved between Panda and Lefty, allow to draw an intriguing parallel between sea urchin Panda and mammalian Inhibin α and raise the unexpected possibility that the original function of Panda may have been in activation of the Nodal pathway rather than in its inhibition.

???displayArticle.pubmedLink??? 38913712
???displayArticle.pmcLink??? PMC11239237
???displayArticle.link??? PLoS Biol

???attribute.lit??? ???displayArticles.show???

	Fig 1. ACVRII is essential for D/V patterning and overexpression of acvrII causes dorsalization.(A) Morphological phenotypes resulting from injection into the egg of antisense morpholino oligonucleotide targeting the translation start site of the acvrII transcript. Knockdown of acvrII resulted in the loss of D/V polarity and caused a radialized phenotype, mimicking knockdown of nodal. The magnification bar is 10 μm. (B) Expression of nodal in wild-type embryos and acvrII morphants. acvrII morphants exhibited loss of nodal expression in the ventral ectoderm. (C) Morphological defects at different developmental stages caused by acvrII overexpression. acvrII overexpressing embryos appeared radialized at 24 hpf. At 48 hpf and 72 hpf, acvrII overexpressing embryos appeared strongly dorsalized, with ectopic pigmentation and presence of elongated spicules supporting a thin and irregular ectoderm. (D) acvrII overexpressing embryos adopt a morphology typical of that of embryos dorsalized by treatment with recombinant BMP4 or of embryos overexpressing bmp2/4.
	Fig 2. Co-injection of panda mRNA with acvrII mRNA suppressed the dorsalization caused by acvrII and partially rescued the D/V polarity.(A) Overexpression of panda mRNA (1,500 μg/ml) reduces nodal expression at blastula stage. (B) While overexpression of acvrII mRNA alone (600 μg/ml) dorsalizes the embryos, co-injection of panda mRNA partially rescues normal development and allows the D/V axis to be established. (C) pSmad1/5/8 immunostaining at the very early blastula stage (VEB) in wild-type embryos and embryos overexpressing panda, acvrII, and panda+acvrII. At the VEB stage, control embryos lacked endogenous pSmad1/5/8. Overexpression of panda (800 μg/ml) did not activate pSmad1/5/8 signaling. Embryos injected with acvrII (600 μg/ml) showed activation and nuclearization of pSmad1/5/8 in all cells at the VEB stage. Co-injection of panda mRNA (400 μg/ml) with acvrII mRNA (600 μg/ml) blocked the ectopic activation of pSmad1/5/8 caused by acvrII overexpression. (D) Overexpression of acvrII up-regulates nodal expression starting at early blastula stage but it does not cause ectopic expression of bmp2/4.
	Fig 3. Panda blocks the ability of ACVRII to orient the D/V axis and Panda acts upstream of the Nodal receptors.(A) Scheme of the axis specification assay. This assay is based on the fact that In Paracentrotus, there is no correlation between the D/V axis and the first cleavage plane. mRNA encoding the factor to be tested is injected locally into one blastomere at the two-cell stage and the positions of the injection clones is scored at prism/pluteus stages. (B) Effect of local overexpression of acvrII mRNA in the presence or absence of panda mRNA on the orientation of the D/V axis. Injection of acvrII locally into one blastomere at the two-cell stage imposes a ventral identity to the progeny of the injected blastomere in nearly 100% of the injected embryos. Local co-injection of panda with acvrII blocks the ability of acvrII to impose a ventral identity to the clone of injected cells resulting in a mixture of ventral and dorsal clones. (C) Histogram representing the percentage of embryos displaying dorsal, ventral, or lateral injection clones following local overexpression of acvrII, or of both acvrII and panda. The number of embryos scored in each category is indicated. (D) Effect of local overexpression of alk4/5/7 QD mRNA in the presence or absence of co-injected panda mRNA on the orientation of the D/V axis. Panda does not block the ability of Alk4/5/7 QD to orient ventrally, suggesting that Panda activity is needed upstream of the Nodal receptors. (E) Histogram representing the percentage of embryos displaying dorsal, ventral, or lateral injection clones following local overexpression of panda, alk4/5/7 QD, or both panda and alk4/5/7 QD. The number of embryos scored in each category is indicated in Table 1.
	Fig 4. Structural and biochemical analyses of Panda and Panda protein interaction studies show that Panda can interact with the ACVRII in vitro.(A) Immunoblot of full-length Panda, Pandamat, and PandaIgKSP tagged with 3× HA tags under reducing and non-reducing conditions showing the presence of these proteins in the culture supernatant of HEK293T cells transfected with these constructs. Full-length Panda was secreted predominantly as an 80 kDa precursor form (pre), and to a significantly lesser extent, as a processed 27 kDa mature form (mat). Among the pro-domain deleted versions of Panda, Pandamat showed a significantly higher secretion into the culture supernatant than PandaIgKSP. The schemes on the right represent monomeric and dimeric forms of wild-type and mutant Panda molecules. N-linked glycosylation sites are represented by ψ (B) Co-immunoprecipitation of Pandamat with ACVRII or the Nodal co-receptor Cripto. Cell lysates immunoprecipitated (IP) with anti-Myc (ACVRII or Cripto) in the presence of co-transfected Pandamat and probed with anti-HA show a 27 kDa band likely corresponding to the glycosylated Panda monomer. (C) Co-immunoprecipitation of Pandamat with the soluble form of ACVRII in the culture supernatant. Pandamat binds ACVRIIECD secreted in the culture medium, suggesting direct binding. (D) Co-immunoprecipitation of Pandamat with the Inhibin-alpha co-receptor TBR3 (Betaglycan). Cell lysate immunoprecipitated (IP) with anti-Myc (TBR3) in the presence of co-transfected Pandamat and probed with anti-HA shows a band at the size of the Panda monomer. The lines in C and D indicate that the 2 lanes shown originate from the same blot but were juxtaposed after cutting an irrelevant lane. (E) Co-immunoprecipitation of PandaP461Smat with ACVRII or Cripto, showing that Panda P461Smat can form complexes with ACVRII or Cripto.
	Fig 5. Structural basis for the antagonistic activity of Panda.Alignment of the mature domain of Panda with other members of the TGFβ family. Unlike Nodals and Activins that possess a serine in the ß 6 region within the knuckle domain or ACVRII binding region, proteins of the Panda family, like all the Leftys, possess a proline (red asterisk). Also note that while in most of the Nodals, BMPs, and Inhibins ß, the ß 6 and ß 7 strands are separated by 4–6 amino acids, in both Panda and in all the Leftys, as well as in the Inhibin α proteins, the region separating the ß 6 and ß 7 strands is about twice this size (8–12 residues).
	Fig 6. Morphological phenotypes resulting from the global and local injection of pandaP461S mRNA into the egg or two-cell zygotes.(A) pandaP461S overexpressing embryos appear radialized at the late gastrula stage and partially ventralized at the pluteus stage. (B) In situ hybridization for nodal and chordin in wild-type and pandaP461S overexpressing embryos at the late blastula stage. Note that pandaP461S injection causes a dramatic ectopic expression of both nodal and chordin. (C) qRT-PCR analysis of nodal transcript levels in wild-type and pandaP461S injected embryos at the swimming blastula stage. pandaP461S injection results in a 2.7-fold increase in nodal expression when compared to wild-type embryos. The values underlying the graph can be found in S1 Data. (D) Synergy between Panda P461S and ACVRII. Co-injection of panda P461S mRNA and acvrII mRNA result in strong ventralization of the embryos. (E) Effect of local overexpression of panda or pandaP461S mRNA on D/V axis formation. While local injection of panda into one blastomere at the two-cell stage imposes a dorsal identity to the progeny of the injected blastomere, local injection of pandaP461S predominantly assigns a ventral identity to the progeny of the injected cell. (F) The panda P461S mutant does not work as a dominant negative form of Panda. Introduction of Y426P in the type I receptor binding region of Panda P461S abrogates its ventralizing activity. (G) Histogram representing the percentage of embryos displaying dorsal, ventral, or lateral injection clones following local overexpression of panda, panda P461S, or panda P461S + Y426P.
	Fig 7. Phylogenetic analysis of metazoan TGF-β ligands: All Panda sequences from sea urchins, like all Leftys, carry a proline in position equivalent to Panda P461.Left, maximum likelihood phylogenetic tree of TGF-β sequences from deuterostomes (vertebrates, cephalochordates, hemichordates, tunicates, hemichordates, and echinoderms) and protostomes (arthropods, cnidarians, molluscs, annelids). The full-length sequences of the TGF-β were used for the analysis. The tree was calculated using the maximum likelihood method with PhyML with the substitution model WAG. A consensus tree with a 50% cut off value was derived from the support values of the alRT test. Numbers above nodes represent the approximate likelihood ratio values as percentage of the values supporting the node. The scheme on the right is a close up of the first clades of the tree on the left. The nature of the residue present at the position equivalent to proline 461 of Panda in each species is indicated by a colored dot. The list of the accession numbers of the 182 sequences is provided in the Supporting information (S1 Text).
	Fig 8. Model of symmetry breaking during D/V axis specification.Nodal is ubiquitously induced at the 32-cell stage while maternal panda mRNA is enriched in the presumptive dorsal territory. Translation of panda mRNA generates a broad gradient of Panda protein along the D/V axis. Panda antagonizes the Nodal type II receptor ACVRII, likely resulting in an asymmetry of Nodal signaling along the D/V axis. Nodal is unable to auto-stimulate its own expression in the dorsal territory and as a result, expression of nodal is restricted to the presumptive ventral territory. During later stages, this asymmetry is maintained by Nodal/Lefty reaction–diffusion mechanism.
	Fig 9. Proposed model of the mechanism by which Panda antagonizes Nodal signaling.Panda is proposed to antagonize Nodal signaling by forming asymmetrical heterodimers with Nodal. In the Panda-Nodal heterodimer, Nodal can interact with high affinity with ACVRII while Panda cannot do so due to the presence of proline 461 in its ACVRII binding motif. In contrast, Panda interacts with high affinity with Betaglycan. This complex, made of ACVRII-Nodal-Panda-Betaglycan, cannot signal due to the absence of the type I receptor in the receptor complex and therefore inhibits Nodal signaling by sequestering ACVRII. Note that Panda may also form complexes with other factors such as Cripto or Alk4/5/7 and these complexes may contribute to the antagonism of Panda of Nodal signaling. Also note that the Nodal-Panda heterodimer, which attenuates Nodal signaling, would play a role opposite to that of the Nodal-Univin heterodimer, which promotes strong Nodal signaling.