October 1, 2013;
Coco regulates dorsoventral specification of germ layers via inhibition of TGFβ signalling.
One of the earliest steps in embryonic development is the specification of the germ layers, the subdivision of the blastula embryo
. Maternally expressed members of the Transforming Growth Factor β (TGFβ) family influence all three germ layers; the ligands are required to induce endoderm
, whereas inhibitors are required for formation of the ectoderm
. Here, we demonstrate a vital role for maternal Coco
, a secreted antagonist of TGFβ signalling, in this process. We show that Coco
is required to prevent Activin and Nodal
signals in the dorsal marginal side of the embryo
from invading the prospective ectoderm
, thereby restricting endoderm
- and mesoderm
-inducing signals to the vegetal and marginal zones of the pre-gastrula
Xenopus laevis embryo
[+] show captions
Fig. 1. Knockdown of Coco causes anterior truncations at tadpole stages. (A-H) Xenopus embryos at stage 28. Overexpression of Coco results in embryos with an ectopic head; compare uninjected embryo (A) with Coco-injected (B, asterisk). Knockdown of Coco causes anterior truncations (C). Hoxb9 is expressed in the spinal cord (D, arrow); Coco morphant embryos have no expression of Hoxb9 (E). Host transfers were also performed with MO (F-H). Embryos have a loss of anterior structures (G, arrowheads, compared with control embryo in F). Anterior structures, such as the cement gland (H, arrowheads) could be rescued by injection of Coco RNA. (A,B) Lateral view, anterior to the left. (C-H) Anterior to the top. cg, cement gland; fb, forebrain; hb, hindbrain; mb, midbrain; sc, spinal cord.
Fig. 2. Knockdown of Coco causes germ layer defects at blastula stage. (A-M) Xenopus embryos were injected with either CocoMO or ControlMO and analysed at stage 9.5. Whole-mount in situ hybridisation was performed to identify endoderm (Sox17β), mesoderm (Xbra) and presumptive dorsal tissue (Chordin). Injection of CocoMO causes both a shift in Sox17β expression (A-C, arrowheads; using the blastocoel floor as a reference this shift is clearly seen in sections shown in A′,B′) and an upregulation of expression (D; compare RT-PCR of uninjected embryo and CocoMO-injected embryo). CocoMO-injected embryos were additionally injected with β-Gal in either a dorsal or ventral blastomere at the four-cell stage. Compared with an uninjected control (E, arrows) the shift of Sox17β expression is on the same side as dorsally injected β-Gal (F) but on the opposite side as ventrally injected β-Gal (G). Loss of Coco also caused a reduction of both Xbra (H-J′, arrows) and Chordin (K-M, arrowheads) expression, effects that are not seen following ControlMO injections. The shift of Sox17β expression following CocoMO injection (N,O, arrows) is rescued with an injection of Coco mRNA (P), demonstrating specificity. A-C,E-J,K-P are whole-mount lateral views and A′-B′;H′-J′ are sagittal sections.
Fig. 3. Germ layer defects in Coco morphant embryos are caused by an increase in Xnr5/6 and Activin signalling. (A) Western blot analysis demonstrating an increase of P-Smad2 following CocoMO and Activin overexpression, compared with uninjected and ControlMO-injected embryos. A dorsoventral bias of P-Smad2 is detected in normal development (compare dorsal half with ventral half). (B) Coco depletion increases Nodal-induced transcription of a reporter gene. Error bars represent standard errors. Assays were performed in triplicate. (C-N) Embryos were injected with CocoMO, CocoMO+TGFβMO or TGFβMO and compared with uninjected control embryos at stage 9.5. (C-F) Injection of CocoMO causes a shift in Sox17β expression (D, arrows), which is rescued by co-injection with Xnr5/6MO (E). (F) Injection of Xnr5/6MO alone does not affect the Sox17β domain, but reduces the intensity of expression. (G-J′) Injection of CocoMO causes a reduction in Xbra staining in the marginal zone (H,H′), which is clearly rescued by co-injection with ActivinMO (I,I′), an effect different to that of ActivinMO alone (J,J′; compare black arrows in G-J and red arrowheads in G′,I′). (K-N) Injection of CocoMO causes a shift in Sox17β expression (L, arrows), which is also rescued by co-injection of ActivinMO (M, arrows). (N) ActivinMO alone caused a slight reduction in Sox17β. D, dorsal half; V, ventral half.
Fig. 4. Coco controls germ layer specification via an inhibition of both Activin and Nodal signals. (A) A dorsoventral gradient of Coco activity (brown) restricts Activin and Xnr signals (green) from acting in the animal pole, ensuring correct spatial organization of the germ layers. (B) Knockdown of Coco allows dorsal marginal Activin and Nodal signals to become active in a more animal domain, disrupting mesoderm and endoderm formation resulting in a loss of anterior structures. D, dorsal; V, ventral.
Suppl. Figure 1: Efficacy of as-oligo and MO to Coco. (A) RT-PCR demonstrating the loss of endogenous Coco in oocytes after injection with an as-oligo specific for Coco. (B) Oocytes were injected with Coco RNA and a Western performed with an anti-Coco antibody. A strong band can be seen which is lost after co-injection with CocoMO.
Suppl. Figure 2: Knockdown of Coco results in a loss of head mesoderm. (A) CocoMO injected embryo at st9.5. Purple staining represents Gsc and red staining, CocoMO. Gsc expression is lost where Coco is knocked down. (B) Uninjected control embryo showing endogenous Gsc expression
Suppl. Figure 3: qRT-PCR of explants from st16 Coco morphant embryos. (A) Animal caps from Coco depleted embryos were analysed for a variety of molecular markers. A strong upregulation of Xbra was detected. (B) Analysis of equatorial region of the embryo also had an increase in mesodermal markers, as seen by the upregulation of MyoD. y-axis: expression ratio (2-log scale)