BMP antagonists and FGF signaling contribute to different domains of the neural plate in Xenopus

Abstract

In ectodermal explants from Xenopus embryos, inhibition of BMP signaling is sufficient for neural induction, leading to the idea that neural fate is the default state in the ectoderm. Many of these experiments assayed the action of BMP antagonists on animal caps, which are relatively naïve explants of prospective ectoderm, and different results have led to debate regarding both the mechanism of neural induction and the appropriateness of animal caps as an assay system. Here we address whether BMP antagonists are only able to induce neural fates in pre-patterned explants, and the extent to which neural induction requires FGF signaling. We suggest that some discrepancies in conclusion depend on the interpretations of sox gene expression, which we show not only marks definitive neural tissue, but also tissue that is not yet committed to neural fates. Part of the early sox2 domain requires FGF signaling, but in the absence of organizer signaling, this domain reverts to epidermal fates. We also reinforce the evidence that ectodermal explants are naïve, and that explants that lack any dorsal prepattern are readily neuralized by BMP antagonists, even when FGF signaling is inhibited.

Figure 1. Ventralized embryos express residual sox2 and sox3

X. laevis embryos were injected in both blastomeres at the two cell stage with β-catenin MO, then raised to the stage indicated, fixed in MEMFA, dehydrated and stained by in situ hybridization for expression of sox2 (A) or sox3 (B). Uninjected control sibling embryos are shown at right for each stage. For β-catenin morphants, a ring of sox2 expression is evident around the blastopore between stages 14 and 16, which fades and has disappeared by stage 19. The same pattern is seen for sox3. Embryos are shown in dorsal views with anterior to the top.

Figure 2. Ventralized embryos do not express markers of differentiated neural tissue

X. laevis embryos were injected in both blastomeres at the two cell stage with β-catenin MO, then raised to the stages indicated, fixed in MEMFA and analyzed by in situ hybridization for expression of differentiated neural markers, the organizer markers chordin and gsc, or fgf8. Uninjected control sibling embryos are shown to the left. A) ncam, nrp1 and n-tubulin are not expressed in β-catenin morphants. Among stage 14 β-catenin morphants, expression was observed in 0/10 embryos for ncam, 0/10 embryos for nrp, and 0/9 embryos for n-tubulin. At stage 16, expression was observed in 1/12 embryos for ncam, 0/10 embryos for nrp, and 0/10 embryos for n-tubulin. At stage 18, expression was observed in 0/32 embryos for ncam, 0/31 embryos for nrp, and 1/34 embryos for n-tubulin. The two embryos expressing detectable ncam or n-tubulin had morphologically normal neural plates, suggesting that they received a low dose of β-catenin MO. B) fgf8 is expressed in a ring around the blastopore in β-catenin morphants at stage 14, in a pattern similar to that seen for sox2. This expression fades by stage 18. C) Gastrula-stage β-catenin morphants do not express the organizer markers chordin or gsc (For chordin, 0/8 morphants showed expression, and 0/9 morphants expressed gsc) Embryos in A and B are shown in dorsal views with anterior to the top, while embryos in C are shown in vegetal views with dorsal to the top.

Figure 3. The sox2 expressed in β-catenin morphants is FGF-dependent

X. laevis embryos were injected with β-catenin MO and cultured in SU5402 to the stage indicated, then fixed and assayed for expression of xbra (A) or sox2 (B). A) Control uninjected DMSO-treated embryos express xbra in the notochord and posterior mesoderm at stage 14, and in the posterior mesoderm only at stage 18. This expression is lost in embryos treated with the Fgf receptor inhibitor SU5402. β-catenin morphants treated with DMSO express xbra in a ring around the blastopore at stages 14 and 18, and this expression is also lost with SU5402 treatment. B) Expression of sox2 in uninjected embryos is reduced by SU5402 treatment, although small amounts of sox2 expression are retained anteriorly for all doses of SU5402 at stage 14, and for all but the highest doses of SU5402 at stage 18. In β-catenin embryos, sox2 is expressed in a ring around the blastopore, which is very sensitive to SU5402 and is lost even at low SU5402 doses both at stage 14 and 18.

Figure 4. X. laevis FCN morphants do not express sox2

X. laevis embryos were injected in each blastomere at the 2 cell stage with 20ng each of follistatin, chordin, and noggin MOS. Embryos were raised to the stage indicated, fixed in MEMFA, and stained for expression of sox2 or epidermal cytokeratin. A) Some embryos expressed low levels of sox2, but this expression was not seen in a ring. Most FCN morphants do not express sox2. The fraction of embryos falling into each class of sox2 expression are shown in the lower right corner. B) Cytokeratin is expressed strongly in FCN morphants at all stages assayed, but is excluded from the blastopore and adjacent tissue (22/24 embryos with strong circumferential expression). C) Xbra expression is expanded in neurula stage FCN morphants. Embryos are shown in dorsal views with anterior to the top.

Figure 5. sox2 has slightly different endogenous expression in X. tropicalis and X. laevis

X. tropicalis or X. laevis embryos were raised to the stage indicated, fixed in MEMFA and analyzed by in situ hybridization for expression of sox2. In X. tropicalis, sox2 expression closes posteriorly around the blastopore, and is tightly adjacent to the blastopore margin. In X. laevis, posterior circumblastoporal expression of sox2 is weak. All embryos are shown in dorsoposterior views with anterior to the top; the embryos have been rotated slightly to clearly show the blastopore.

Figure 6. Anterior neural markers are more sensitive to the loss of Bmp antagonists than posterior neural markers

X. laevis embryos were injected at the two cell stage with increasing doses of Follisatin, Chordin, and Noggin MOs. Embryos were injected in both blastomeres at the two-cell stage; doses indicated refer to the total dose of MO injected into the embryo. Embryos were raised to stage 22, fixed and analyzed by in situ hybridization for expression of the genes indicated. nkx2.1 marks the ventral forebrain and is lost in embryos injected with low doses of FCN MOs. The forebrain/midbrain marker otx2, midbrain/hindbrain boundary marker en2 and rhombomeres 3/5 marker krox20 have reduced expression in embryos injected with low or intermediate doses of MO, and are lost in embryos injected with a high dose of MOs. The spinal cord marker hoxB9 is expressed strongly in embryos injected with low or intermediate doses of MO and is lost in embryos injected with a high dose of MO. Embryos are shown in lateral views with anterior to the left.

Figure 7. Xenopus ectodermal explants and ventralized embryos do not express neural border markers

Uninjected control X. laevis embryos, β-catenin morphants, and animal caps cut from blastula stage uninjected or noggin-injected embryos were cultured to stage 20 and assayed for expression of the neural border markers pax3, hairy2a, and zic2, and the neural crest marker slug. Expression of these genes was not detected in animal caps or β-catenin morphants (of β-catenin morphants, 0% expressed hairy2a (n=14), pax3 (n=15), and zic2 (n=14), while 6% expressed slug weakly (n=16). All four genes had some expression in animal caps cut from embryos injected with 1pg of noggin mRNA (63% for hairy2a, n=8; 50% for pax3, n=10; 79% for slug, n=14; and 67% for zic2, n=9). The epidermal marker cytokeratin is strongly expressed in β-catenin morphants but excluded from the blastopore. Cytokeratin is strongly expressed in both injected and uninjected animal caps, but was somewhat weaker in noggin-injected animal caps. Embryos are shown in dorsal views with anterior to the left.

Figure 8. Ectodermal explants derived from ventralized embryos are neuralized by Noggin

A) X. laevis embryos were injected in both blastomeres at the two cell stage with β-catenin MO, then re-injected in the animal pole at the 4 cell stage with 1 or 10pg of noggin mRNA. At stage 9, animal caps were cut, and cultured to stage 24. Animal caps from β-catenin morphants and control embryos injected with noggin mRNA were analyzed by in situ hybridization for expression of sox2 and nrp1. Animal caps cut from β-catenin noggin injected morphants were equally likely to express sox2 and nrp1 as caps cut from embryos injected with noggin mRNA alone. B, C) Quantification of the results shown in (A). A repeat of the experiment gave similar results. D) Animal caps cut from β-catenin morphants injected with noggin mRNA also strongly express both sox2 and definitive neural markers as assayed by non-quantitative RT-PCR.

Figure 9. Ectodermal explants do not require FGF signaling to be neuralized by Noggin

A) X. laevis embryos were injected in both blastomeres at the two cell stage with β-catenin MO, then re-injected in the animal pole at the 4 cell stage with 1 or 10pg of noggin mRNA. Embryos were cultured in SU5402 or 0.32% DMSO from the 32 cell stage until stage 8. At stage 8, animal caps were cut, returned to SU5402 or DMSO, and cultured to stage 24. Whole embryos treated with increasing levels of SU5402 from stage 9 to late neurula stages were stained for apoptotic nuclei using TUNEL (B). 93% of embryos treated with 100υM SU5402 had widespread TUNEL staining (n=15). Animal caps from β-catenin morphants and control embryos injected with noggin mRNA were analyzed by in situ hybridization for expression of nrp (A), or by RT-PCR for expression of nrp, and sox2 (D). Ornithine decarboxylase (ODC) was used as a loading control. The efficiency of this SU5402 treatment protocol was confirmed by assaying expression of xbra at stage 11 (C). 0% of Su5402 treated embryos expressed xbra (n=36). Expression of nrp and sox2 was also quantified using quantitative RT-PCR (F, G) and normalized to ODC expression. Animal caps cut from SU5402 treated, noggin-injected embryos still express nrp and sox2 at comparable levels to animal caps from noggin injected embrys treated only with DMSO This is also true of animal caps cut from noggin-injected, β-catenin injected, SU5402 treated embryos.,