November 1, 2007;
Neural induction requires continued suppression of both Smad1 and Smad2 signals during gastrulation.
Vertebrate neural induction requires inhibition of bone
morphogenetic protein (BMP) signaling in the ectoderm
. However, whether inhibition of BMP signaling is sufficient to induce neural tissues in vivo remains controversial. Here we have addressed why inhibition of BMP/Smad1
signaling does not induce neural markers efficiently in Xenopus ventral ectoderm
, and show that suppression of both Smad1
signals is sufficient to induce neural markers. Manipulations that inhibit both Smad1
pathways, including a truncated type IIB activin receptor, Smad7
, induce early neural markers and inhibit epidermal genes in ventral ectoderm
; and co-expression of BMP inhibitors with a truncated activin/nodal
-specific type IB activin receptor leads to efficient neural induction. Conversely, stimulation of Smad2
signaling in the neural plate
stages results in inhibition of neural markers, disruption of the neural tube
and reduction of head
structures, with conversion of neural to neural crest
and mesodermal fates. The ability of activated Smad2
to block neural induction declines by the end of gastrulation. Our results indicate that prospective neural cells are poised to respond to Smad2
signals to adopt mesodermal and non-neural ectodermal fates even at gastrula
stages, after the conventionally assigned end of mesodermal competence, so that continued suppression of both mesoderm
- and epidermis
-inducing Smad signals leads to efficient neural induction.
[+] show captions
Fig. 1. Neural induction in Xenopus ventral ectoderm by ectopic expression of inhibitors of TGFβ signals. RNAs encoding transmembrane (1 ng tBMPRIA and tActRIIB), cytoplasmic (0.1 ng Smad6 and Smad7) or nuclear (0.1 ng VP16-Msx1 and Ski) inhibitors of TGFβ signaling were co-injected with nβGal (0.1 ng RNA) into one ventral animal blastomere of 32- to 64-cell stage embryos. The embryos were analyzed by Red-Gal staining (red speckled stain) and in situ hybridization of the neural (Sox2 and Sox3), neural crest (Slug) and epidermal (XK70) markers at neurula stages. Inhibitors of both Smad1 and Smad2 signaling (tActRIIB, Smad7 and Ski) induced neural markers efficiently. Among the specific inhibitors of Smad1 signaling, tBMPRIA induced Sox2 and Sox3 weakly, whereas Smad6 and VP16-Msx1 were ineffective in inducing neural markers.
Fig. 2. Neural induction by inhibitors of Smad1 and Smad2 signaling occurs in the absence of the mesoderm in Xenopus. Double in situ hybridizations showed that neural marker induction in embryos injected with tActRIIB, Smad7 or Ski occurred in the absence of the mesodermal markers Chordin and MyoD. The red speckled staining is from the injected lineage tracer.
Fig. 3. Blocking both Smad1 and Smad2 signals leads to efficient neural induction in Xenopus ventral ectoderm. (A) At neurula stages, the truncated activin receptor tActRIB did not induce and tBMPRIA only weakly induced the neural markers Sox2 and Sox3. When the two truncated receptors were co-expressed, the neural markers were induced strongly to a level similar to that induced by the truncated type II receptor tActRIIB. One nanogram of each RNA was used. (B) Co-expression of tActRIB (1 ng) with Smad6 (0.1 ng) or VP16-Msx1 (0.1 ng) in ventral animal cells led to induction of the neural markers Sox2 and Sox3 in ventral ectoderm of frog neurulae.
Fig. 5. Induction of anterior, but not posterior, neural tissues in Xenopus. tActIIB, Smad7 and Ski induced the fore- and mid-brain markers Otx2 and En2, but not the hindbrain and spinal cord markers Krox20 and HoxB9.
Fig. 7. Activation of Smad2 signaling converts neural tissue to neural crest and mesodermal tissues in Xenopus. (A) In the absence of DEX, leaky GR-Smad2 activity was sufficient for neural crest induction, but not sufficient for inhibition of neural markers or induction of mesodermal genes. Activation of GR-Smad2 by DEX (2 μM) at mid-gastrula stages led to inhibition of Sox2 and Sox3 and simultaneous induction of the mesodermal markers MyoD and Chordin in the neural plate (seen more clearly in Fig. 7B and Fig. 8). GR-Smad2 RNA (0.1-0.2 ng) was used. The embryos were orientated with the head toward the left and viewed from the dorsal side. (B) Induction of mesodermal markers by activated GR-Smad2 occurred in the neural plate, as shown in transversely bisected (top) or sectioned (bottom) embryos.
Fig. 8. The ability to inhibit neural markers by activated Smad2 attenuates during gastrulation. Treatment of Xenopus embryos expressing GR-Smad2 with DEX at different stages during gastrulation showed that activated GR-Smad2 lost its neural inhibitory activity by the end of gastrulation, at which stages it also failed to induce mesoderm in the neural plate. All the embryos were viewed from the dorsal side with the anterior to the left.
Fig. 10. Activation of Smad2 at gastrula stages in the neural plate leads to defective neural development in Xenopus. (A) Activation of GR-Smad2 (0.2-0.5 ng) at mid-gastrula stages (stage 11) in the neural tissue induced neural defects in frog tadpoles. Embryos showed reduced head structures and malformed or missing eyes. Embryos without DEX treatment developed normally. (B) Histological analyses indicated that neural development at both anterior (upper panels) and posterior (lower panels) trunk levels was defective when Smad2 signaling was activated. The neural tube was disrupted and ectopic notochord (yellow arrowhead) and mesenchyme (red arrowhead) were observed in the neural derivatives. (C) In situ hybridization demonstrated that Sox2 was reduced and split from the midline and Otx2 was reduced, but the neural crest marker Twist and the muscle marker MyoD were unaffected. All embryos were viewed from the lateral side with the anterior to the left, except the second column of Sox2 in panel C, which was viewed from the dorsal direction.
Fig. S1. Neural induction occurs in the absence of the mesoderm. In situ hybridization demonstrates that the mesodermal markers Chordin and MyoD were not turned on in embryos injected with tActRIIB, Smad7 or Ski, though these molecules could induce neural marker expression in ventral ectoderm of frog neurulae.