A: RT-PCR (30 cycles) of embryos at stages 3 (4 cell), 6, 9, 14, and 17 revealed similar levels of twist, snail2, and snail1 RNAs; orthinine decarboxylase (ODC) was used in this and all other such experiments as a control. B: Equatorial dorsal marginal zone (DMZ), lateral marginal zone (LMZ), and ventral marginal zone (VMZ) regions of early gastrula (stage 10/11) embryos were dissected and analyzed by RT-PCR (28 cycles)(C); twist, snail2, and snail1 RNAs were present at highest levels in the DMZ and LMZ regions, but detectable in the VMZ. D: Fertilized eggs were injected in the animal hemisphere with RNA encoding GR-RelA (500 pg/embryo); at stage 8/9 the protein was activated with dexamethasone (+Dex) in the presence or absence of the protein synthesis inhibitor emetine (+Eme); control embryos were treated with dimethylsulfoxide alone (−Dex). Embryos were analyzed at stage 11 (3 to 4 hours after protein activation) and showed an emetine-insensitive increase in twist RNA levels. E: Fertilized eggs were injected in the animal pole region with RNAs encoding either VegT (100 pg/embryo) or Xnr1 (nodal-related protein 1)(50 pg/embryo) either alone or together with RNA encoding the Nodal-specific signaling inhibitor CerS (300 pg/embryo); ectodermal explants (animal caps) were prepared at stage 8/9 and analyzed at stage 11. Explants from uninjected embryos showed very low levels of twist, snail2, and snail1 RNAs, while both VegT and Xnr1 RNAs induced an accumulation of all three RNAs what was blocked by CerS, suggesting the Nodal dependent regulation of twist, snail2, and snail1 RNA levels. F: The effects of VegT were not blocked by co-injection of RNAs encoding either the Wnt antagonist Dickkopf (dkk) or the BMP inhibitor Noggin (both injected at 500 pg/embryo). G, H: Fertilized eggs were injected with GR-Smad2 RNA (600 pg/embryo); animal caps were prepared and treated with dexamethasone and/or emetine and analyzed by standard (G) or quantitative RT-PCR (H) at stage 11. twist, snail2, snail1, and relA RNA levels were increased in an emetine-insensitive manner following dexamethasone activation of Smad2.

Compared to uninjected embryos (A, C, E, G), injection of the Twist morpholino (7 ng/embryo)(B, D, F, H) into one cell of two cell embryos led to the loss of xbra (A, C), antipodean/vegT (B, D), and chordin (E, F) expression (arrows), and the expansion (arrow) of the expression domain of the endodermal marker endodermin (G, H). “yp” marks the yolk plug of gastrula stage embryos, “BP” marks the blastopore - all embryos are oriented similarly, with the dorsal lip oriented to the lower right.

Compared to uninjected embryos (A and top embryo in part H), the unilateral injection of either Snail2 (B) or Twist (C) morpholinos (7 ng/embryo) in two cell embryos led to the anterior loss of myoD RNA in situ hybridization staining in stage 22/24 embryos. The loss of myoD staining in Twist (D) or Snail2 (E) morphant embryos was rescued by Twist RNA injection (500 pg/embryo); similarly the loss of myoD staining in Snail2 (F) or Twist (G) morphant embryos was rescued by the injection of Snail2-GFP RNA (500 pg/embryo). All embryos are oriented with anterior to the left and dorsal to the top. H: In larvae derived from unilaterally injected 2-cell embryos there was a pronounced curvature toward the injected side. In all cases, it is worth noting that these embryos survived gastrulation and are therefore likely to represent partial, rather than complete, loss of function phenotypes.

At stage 16 there were few TUNEL positive cells in control (uninjected or control morpholino injected) embryos (data not shown). Unilateral injection of Snail2 (A) or Twist (C) morpholinos led to a dramatic increase in TUNEL staining. Co-injection of Twist-GFP RNA suppressed the increase in TUNEL positive cells in Snail2 morphant embryos (B), while Snail2-GFP RNA suppressed TUNEL staining in Twist morphant embryos (D). Arrows indicate injected sides of the embryos, which are oriented anterior up. E: Bilateral injection of the Twist morpholino leads to an increase in caspase-9 and a decrease in bclxl and xbra RNA levels. F: In animal caps prepared from fertilized eggs injected with RNA encoding GR-Twist-GFP; when activated, there was an emetine-insensitive decrease in the levels of caspase-9 RNA. This effect is similar to that produced by Snail2 (Zhang et al 2006).

This schematic summarizes the key observations and interactions presented here. The maternal factors Sox3 (in the animal hemisphere) and VegT (in the vegetal hemisphere) regulate Nodal expression (Zhang and Klymkowsky, 2007). Nodal signaling via Smad2 positively regulates twist, snail2, snail1, and relA RNA levels in an emetine-independent (“possibly direct”) manner. relA, the major zygotically expressed NF-κB gene in Xenopus (Zhang et al., 2006), positively regulates twist, snail2, and snail1 RNA levels in an emetine-independent (“possibly direct”) manner. Twist and Snail2 (and perhaps Snail1 as well), negatively regulate cerberus RNA, which encodes a Nodal, BMP and Wnt antagonist, in an emetine-independent manner. The net effect is that Twist and Snail2 will increase Nodal activity by inhibiting cerberus expression, which will, in turn increase twist, snail2, snail1, and relA expression, forming a positive feedback loop, at least in the absence of other regulatory inputs.

A, B: Fertilized eggs were injected with RNA encoding UTR-Twist-GFP RNA (500 pg/embryo) either alone (“UN”) or together with the Twist morpholino (“Twist MO”) or a control (“Con MO”), both at 7 ng/embryo; injected embryos were analyzed at stage 11. A: The Twist morpholino inhibited the accumulation of Twist-GFP, as monitored by immunoblot analysis using an anti-GFP antibody; the control morpholino did not. Blotting with an antiTcf3 antibody served as a loading control. B: Injection of the Twist morpholino inhibited Twist-GFP fluorescence while the control morpholino did not. C, D: Embryos were injected bilaterally with either Twist or Snail2 morpholinos and analyzed at stage 11; injection of the Twist morpholino (C) inhibited the accumulation of snail2 and snail1 RNAs, while injection of the Snail2 morpholino (D) inhibited the accumulation of twist RNA; injection of a control morpholino had no apparent effect on twist RNA levels. E, F: Fertilized eggs were injected with either GR-Twist-GFP or GR-Snail2-GFP RNAs (500 pg/embryo); embryos were treated with dexamethasone and/or emetine beginning at stage 8, RNA was isolated at stage 10/11, and levels of snail2, snail1 (E), and twist (F) RNAs were determined by quantitative RT-PCR; the GR-Snail2 experiment was performed only once. snail2 and snail1 RNA levels were indirectly regulated by Twist, while twist RNA levels were indirectly regulated by Snail2.

The effect of Twist MO on xbra expression (A, C) was rescued by the injection of either Twist-GFP RNA (B) or Snail2-GFP RNA (D), both injected at 500 pgs/embryo. Similarly, the loss of xbra expression in Snail2 morpholino injected embryos (E) was rescued by the injection of Twist-GFP RNA (F)(500 pgs/embryo). Arrows indicate the β-galactosidase (red) lineage marker. “yp” indicates the yolk plug and “BP” indicates the blastopore - all embryos are oriented similarly, with the dorsal lip oriented to the lower right.

Unilateral injection of the Twist morpholino inhibited sox9 expression (A) and the loss of craniofacial cartilages in stage 45 tadpoles (B, C), as visualized by Alcian blue staining. The injected side of the embryos is indicated by an arrow, and all embryos are oriented anterior up. A wild type embryo is presented in part D for comparison. sox9 expression was rescued by the co-injection of Twist-GFP (E); this led to partial rescue of the craniofacial phenotype - an array of representative embryos is presented, ranging from almost complete (F) to partial examples of rescue (G, H). I: Injection of Snail2-GFP RNA rescued sox9 expression in Twist morphant embryos and also rescued the effects on craniofacial cartilage to various extents (J-L). In parts E and I it is possible to see the β-galactosidase (red) lineage marker.

A: Animal caps were prepared from fertilized eggs injected with RNA encoding GR-Twist; when activated, there was an emetine-insensitive decrease in the levels of cerberus RNAs. Quantitative RT-PCR analysis using GR-Twist-GFP (B) or GR-Snail2-GFP (C) indicates that both Twist and Snail2 inhibited the accumulation of cerberus RNA in a protein synthesis independent manner. D: Injection of cerberus RNA reduced the level of the relA RNA as determined by RT-PCR. E: In animal caps, both Cerberus and CerS (300 pg RNA/embryo) inhibited the activity of the 3×κB luciferase reporter, normalized to TK-Renilla luciferase activity; injection of a control GFP RNA (500 pg/embryo) had no effect on reporter activity. F: Injection of xnr1 RNA activated the NF-κB-regulated 3×κB luciferase reporter, while co-injection of either cerberus or CerS RNAs blocked this effect. Injection of Smad2, Smad4b (G) or Smad4a (H) RNAs (500 pg/embryo) led to an increase in 3×κB-luciferase reporter activity; co-injection of CerS RNA did not block Smad4b activation of the 3×κB reporter. The effect of Smad4a was reproducibly lower than that of either Smad2 or Smad4b.