The Maternal Maverick/GDF15-like TGF-β Ligand Panda Directs Dorsal-Ventral Axis Formation by Restricting Nodal Expression in the Sea Urchin Embryo

Abstract

Specification of the dorsal-ventral axis in the highly regulative sea urchin embryo critically relies on the zygotic expression of nodal, but whether maternal factors provide the initial spatial cue to orient this axis is not known. Although redox gradients have been proposed to entrain the dorsal-ventral axis by acting upstream of nodal, manipulating the activity of redox gradients only has modest consequences, suggesting that other factors are responsible for orienting nodal expression and defining the dorsal-ventral axis. Here we uncover the function of Panda, a maternally provided transforming growth factor beta (TGF-β) ligand that requires the activin receptor-like kinases (Alk) Alk3/6 and Alk1/2 receptors to break the radial symmetry of the embryo and orient the dorsal-ventral axis by restricting nodal expression. We found that the double inhibition of the bone morphogenetic protein (BMP) type I receptors Alk3/6 and Alk1/2 causes a phenotype dramatically more severe than the BMP2/4 loss-of-function phenotype, leading to extreme ventralization of the embryo through massive ectopic expression of nodal, suggesting that an unidentified signal acting through BMP type I receptors cooperates with BMP2/4 to restrict nodal expression. We identified this ligand as the product of maternal Panda mRNA. Double inactivation of panda and bmp2/4 led to extreme ventralization, mimicking the phenotype caused by inactivation of the two BMP receptors. Inhibition of maternal panda mRNA translation disrupted the early spatial restriction of nodal, leading to persistent massive ectopic expression of nodal on the dorsal side despite the presence of Lefty. Phylogenetic analysis indicates that Panda is not a prototypical BMP ligand but a member of a subfamily of TGF-β distantly related to Inhibins, Lefty, and TGF-β that includes Maverick from Drosophila and GDF15 from vertebrates. Indeed, overexpression of Panda does not appear to directly or strongly activate phosphoSmad1/5/8 signaling, suggesting that although this TGF-β may require Alk1/2 and/or Alk3/6 to antagonize nodal expression, it may do so by sequestering a factor essential for Nodal signaling, by activating a non-Smad pathway downstream of the type I receptors, or by activating extremely low levels of pSmad1/5/8. We provide evidence that, although panda mRNA is broadly distributed in the early embryo, local expression of panda mRNA efficiently orients the dorsal-ventral axis and that Panda activity is required locally in the early embryo to specify this axis. Taken together, these findings demonstrate that maternal panda mRNA is both necessary and sufficient to orient the dorsal-ventral axis. These results therefore provide evidence that in the highly regulative sea urchin embryo, the activity of spatially restricted maternal factors regulates patterning along the dorsal-ventral axis.

Diversification of spatiotemporal expression and copy number variation of the echinoid hbox12/pmar1/micro1 multigene family

“…In a recent paper [40], the partial or complete protein sequences of five hbox12/pmar1/micro1 family members were inferred from one contig of P . lividus genomic sequence.…”

“…Analogously, pairwise alignments showed that the proteins predicted for the hbox12-02 , - 04 and - 27 genes displayed the highest similarity to the three remaining sequences, although various differences were detected (S2 Fig). Although these discrepancies could be at least in part due to allelic polymorphisms, in the absence of genomic sequence for the contig described in [40], we cannot rule out the possibility that we were looking at further distinct copies of the same gene family, which probably escaped from the sequencing and/or assembly of the P . lividus genomic reads.…”

“…Similarly, and in accordance with unpublished evidence obtained during the optimization of our WMISH protocol, WMISH in P . lividus embryos at the early blastula stage clearly highlighted the apparent accumulation of hbox12/pmar1/micro1 transcripts in broad spatial embryonic sectors, likely derived from both micromeres and animal blastomeres of earlier developmental stages [40]. In light of the findings described in the current study, we assume that cross-hybridization of the probe with transcripts from both the hbox12-a-like and pmar1/micro1-like subfamilies occurred during WMISH experiments in both the sea urchin species mentioned.…”

“…lividus embryos following injection of a synthetic mRNA corresponding to a not well-defined member of the hbox12/pmar1/micro1 family, at doses above or equal to 5 μg/ml [40]. By contrast, the same transcript did not produce detectable effects on dorsal-ventral polarization when injected in the developing zygote at concentration below 5 μg/ml [40].…”

Causes and evolutionary consequences of primordial germ-cell specification mode in metazoans

“…However, prioritizing these features can also yield to choices of model organisms that display some developmental similarities to each other, but are not representative of the larger taxa to which they belong. Furthermore, some cnidarians and echinoderms use induction for PGC specification (1) and yet certain data have suggested that maternal products deposited in the egg (products of the Frizzled protein family that activate Wnt signaling in cnidarians and Panda gene products in echinoderms) may partly contribute toward directing embryonic body plans in these groups, raising the possibility that the mechanism of PGC formation and axial patterning are uncoupled in some organisms (64,65). Together, further investigations of mechanisms of embryonic axis formation and PGCspecification modes across a wider range of metazoans will be needed to decipher whether a shift toward use of maternal determinants for embryonic patterning typically cooccurs with germ plasm, consistent with a spandrel effect.…”