February 1, 2007;
FoxI1e activates ectoderm formation and controls cell position in the Xenopus blastula.
The segregation of the vertebrate embryo
into three primary germ layers is one of the earliest developmental decisions. In Xenopus, where the process is best understood, the endoderm
is specified by a vegetally localized transcription factor, VegT
, which releases nodal
signals that specify the adjacent marginal zone
of the blastula
to become mesoderm
. However, little is known about how the ectoderm
becomes specified. In this paper, we show that the forkhead protein FoxI1e
(also known as Xema
) is required at the blastula stage
for normal formation of both the central nervous system
, the two early derivatives of the ectoderm
. In addition, FoxI1e
is required to maintain the regional identity of the animal cells of the blastula
, the cells that are precursors of ectodermal structures. In its absence, they lose contact with the animal cap
, mix with cells of other germ layers and differentiate according to their new positions. Because FoxI1e
is initially expressed in the animal region of the embryo
and is rapidly downregulated in the neural plate
, its role in neural and epidermal gene expression must precede the division of the ectoderm
into neural and epidermal. The work also shows that FoxI1e
plays a role in the embryo
in the poorly understood process of differential adhesion, which limits cell mixing as primary germ layers become specified.
[+] show captions
Fig. 2. FoxI1e is not expressed in all ectodermal cells and does not co-localize with epidermal cilia. (A) In situ hybridization for FoxI1e at stage 10 showing salt and pepper expression. (B,C) Coimmunostaining for α-tubulin (brown, arrows) and in situ hybridization for FoxI1e (purple, arrowheads) demonstrates that these two markers are expressed in different cell populations. Staining was done in whole mount (B), and stained embryos were then embedded and sectioned. A high-power picture is shown in C. (D) Twenty high-power fields were analyzed for expression of the two markers. Sixty-one FoxI1e-positive cells and 83 ciliated cells were counted. There were five instances of overlapping expression.
Fig. 3. FoxI1e expression is sufficient for ectoderm formation. RTPCR analysis of vegetal explants injected with 300-600 pg FoxI1e mRNA and cultured to stage 11. (A) Expression of ectoderm-specific markers is increased, including E-cadherin, epidermal markers epidermal cytokeratin and AP-2, the neural marker Sox-2 and the neural crest marker slug. (B) Endodermal markers Sox17α and endodermin were reduced. (C) Schematic of experimental design to determine behavior of vegetal hemispheres ectopically expressing FoxI1e. (D) In control vegetal hemispheres, pigmented mesodermal cells formed aggregates that extended from bases, whereas in FoxI1e-positive hemispheres, pigmented mesodermal cells formed a layer around the explant (E). Immunostaining for α-tubulin to mark cilia was negative in control embryos (F), but demonstrated surface ciliation on FoxI1e-positive hemispheres (G).
Fig. 6. Loss of FoxI1e causes a rescuable loss of epidermal cilia.Control embryos (A,B) injected with RLDX (red) in an animal, ventral cell at the eight-cell stage and stained for α-tubulin (green) have a normal cilia pattern in injected cells. (C,D) Co-injection of SBMO resulted in loss of cilia in cells that received the morpholino. (E,F) SBMO-resistant FoxI1e mRNA was injected after the SBMO, restoring normal cilia formation.
Fig. 8. FoxI1e controls regional position of ectodermal cells. (A,B) Stage 25 embryos injected at the 32-cell stage in the A4 blastomere, a precursor of epidermis, with RLDX (red) alone (A) or with RLDX + SBMO (B) and then stained for α-tubulin (green). With RLDX alone, injected cells co-localize with cilia (A). With SBMO, the cells are located in the interior of the embryo (B). At stage 47, embryos were sectioned and embedded. (C-E) DIC and fluorescent images of a control embryo injected with RLDX in A4. RLDX is localized to the epidermis. (F-H) Co-injection with SBMO causes cells to localize to the gut. These cells give rise to morphologically normal endodermal structures, including the intestinal epithelium.
Fig. 9. FoxI1e is required for normal ectodermal cell adhesion in the gastrula. (A,B) Brightfield and fluorescent images of a hemisected control stage 11 embryo injected with FLDX alone into A4 at the 32-cell stage. (C,D) Co-injection of SBMO causes the cells to disaggregate and fall into the blastocoel. (E-G) This effect is rescued by morpholino-resistant mRNA injected at the two-cell stage. *P=8.8×10-5, **P=0.03.
Fig. 1. FoxI1e mRNA is animally localized and upregulated in the absence of VegT. RT-PCR analysis of FoxI1e expression. (A) FoxI1e is upregulated in VegT-depleted (VegT–) equatorial (eq) and vegetal (base) explants compared with control explants. (B) Comparison of isolated animal caps, equators and bases dissected from wild-type embryos shows enrichment of FoxI1e expression in prospective ectoderm. Data are normalized to ODC. (C) Loss of Nodal signaling by injection of 1 ng CerS mRNA at the two-cell stage increases expression of FoxI1e in whole embryos.
Fig. 4. Injection of FoxI1e mRNA in D-tier cells at the 32-cell stage causes their descendents to enter other germ layers. (A,C) Wholemount and transverse section of embryos injected with GFP alone in D-tier cells. GFP signal is restricted to the endoderm. (B,D) Co-injection of 10-30 pg FoxI1e mRNA causes cells to enter other germ layers as shown in whole mount (B). The section shown in D shows GFP/ FoxI1e positive cells located in the epidermis.
Fig. 5. An SBMO prevents mRNA maturation and causes developmental abnormalities in embryos. (A) Schematic showing oligo binding site and RT-PCR demonstrating inhibition of mRNA maturation by the SBMO. Both pseudoalleles are affected. Genomic DNA template was used as a positive control for unspliced product. (B) An alternative morpholino, targeting the splice donor site, does not efficiently inhibit mRNA maturation. (C) RTPCR analysis of XBra and FGF8 expression in animal caps depleted of FoxI1e and treated with 1 g/ml Activin A. SBMO-injected animal caps showed levels of mesoderm induction similar to uninjected caps. (D,E) Embryos injected animally at the two cell stage with a dose range of SBMO. There were dose-dependent abnormalities in development, shown at the neurula stage (stage 19, D) and in tailbud stage embryos (stage 27, E). (F) Epidermal cytokeratin, slug, sox-2, FoxG1 (BF-1) and Xlhbox6 are all reduced at stage 23. E. keratin, slug and sox- 2 are partially rescued by 50 pg FoxI1e mRNA. (G) Xbra, Xwnt-8 and chordin are unaffected by SBMO injection at stage 11.5.
Fig. 7. FoxI1e actively promotes ectoderm formation. Embryos were injected at the two-cell stage with SBMO and then at the four-cell stage with cmBMP7. Animal caps were excised at stage 7 and cultured to stage 14. (A,B) Epidermal cytokeratin and E-cadherin are expressed in uninjected (epidermal) animal caps. Loss of FoxI1e downregulates expression of these markers. cmBMP7 neuralizes these caps, so the epidermal markers are reduced. (C-E) The neural markers Sox-2, NCAM and NRP-1 are induced relative to controls in cmBMP7-injected (neural) caps. This induction requires the presence of FoxI1e. Therefore, FoxI1e is necessary for expression of both epidermal and neural genes. (F) Expression of the mesodermal marker Xbra at stage 11 in caps excised at stage 7. Whole embryo expression is shown as a reference.