Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

An Intact Brachyury Function Is Necessary to Prevent Spurious Axial Development in Xenopus laevis

  • Cecilia E. Aguirre,

    Affiliation Laboratorio de Embriología Molecular, Instituto de Biología Celular y Neurociencia ‘‘Prof. E. De Robertis’’ (UBA-CONICET), Facultad de Medicina, Universidad de Buenos Aires, Ciudad Autónoma de Buenos Aires, Argentina

  • Sabrina Murgan,

    Affiliation Laboratorio de Embriología Molecular, Instituto de Biología Celular y Neurociencia ‘‘Prof. E. De Robertis’’ (UBA-CONICET), Facultad de Medicina, Universidad de Buenos Aires, Ciudad Autónoma de Buenos Aires, Argentina

  • Andrés E. Carrasco,

    Affiliation Laboratorio de Embriología Molecular, Instituto de Biología Celular y Neurociencia ‘‘Prof. E. De Robertis’’ (UBA-CONICET), Facultad de Medicina, Universidad de Buenos Aires, Ciudad Autónoma de Buenos Aires, Argentina

  • Silvia L. López

    slopez@fmed.uba.ar

    Affiliation Laboratorio de Embriología Molecular, Instituto de Biología Celular y Neurociencia ‘‘Prof. E. De Robertis’’ (UBA-CONICET), Facultad de Medicina, Universidad de Buenos Aires, Ciudad Autónoma de Buenos Aires, Argentina

An Intact Brachyury Function Is Necessary to Prevent Spurious Axial Development in Xenopus laevis

  • Cecilia E. Aguirre, 
  • Sabrina Murgan, 
  • Andrés E. Carrasco, 
  • Silvia L. López
PLOS
x

Abstract

We have previously shown that the member of the HES family hairy2 induces the ectopic expression of dorsal markers when it is overexpressed in the ventral side of Xenopus embryos. Intriguingly, hairy2 represses the mesoderm transcription factor brachyury (bra) throughout its domain in the marginal zone. Here we show that in early gastrula, bra and hairy2 are expressed in complementary domains. Overexpression of bra repressed hairy2. Interference of bra function with a dominant-negative construct expanded the hairy2 domain and, like hairy2 overexpression, promoted ectopic expression of dorsal axial markers in the ventral side and induced secondary axes without head and notochord. Hairy2 depletion rescued the ectopic dorsal development induced by interference of bra function. We concluded that an intact bra function is necessary to exclude hairy2 expression from the non-organiser field, to impede the ectopic specification of dorsal axial fates and the appearance of incomplete secondary axes. This evidence supports a previously unrecognised role for bra in maintaining the dorsal fates inhibited in the ventral marginal zone, preventing the appearance of trunk duplications.

Introduction

The Brachyury or T mutation was first described in 1927 and the mouse gene was isolated in 1990. Notochord differentiation was severely impaired in homozygous mutant embryos, which lacked mesoderm posterior to the 7th somite [1][4]. Brachyury (bra) is highly conserved in terms of sequences and expression patterns between vertebrates [5][6]. Unlike the single gene in higher vertebrates, teleosts and amphibians contain multiple copies of bra. The allotetraploid genome in Xenopus laevis resulted from the hybridisation of two similar but nonidentical ones. Thus, this species contains the homeologs (also known as allo- or pseudoalleles) Xlbra-a and Xlbra-b (which were originally named bra and bra2), and in addition, the separate gene bra3 [5] [7][10]. Most of the analysis reported in the literature was performed with Xlbra-a (bra).

The T-box DNA-binding protein encoded by bra was characterised as a transcriptional activator of mesoderm specific genes [11][15]. Bra is one of the first genes known to directly respond to mesodermal inducers and is only activated at intermediate Activin concentrations, which induce posterolateral and chordal mesoderm, while the high ones induce dorsoanterior organiser mesoderm but are unable to activate bra [5] [16][19].

In Xenopus, bra expression begins at mid-blastula transition. Although it is initially expressed in a widespread fashion, when gastrulation begins the transcripts are detected only in the marginal zone (MZ), in the prospective mesoderm. By the end of gastrulation, they only persist in the notochord and in the circumblastoporal mesoderm [5] [20][22]. In ectodermal explants, bra is sufficient to specify posterior mesoderm, including somites, but it requires at least the transcription factor foxA4a or the BMP antagonist noggin to promote notochord development [5] [23][26].

Loss-of-function experiments in Xenopus with an interfering bra construct (braEnR), in which its transcription activation domain was replaced with the repressor domain of Drosophila engrailed (EnR), resulted in loss of posterior structures, indicating that bra is essential for posterior mesoderm development [13]. Bra promotes the convergence-extension movements that elongate the notochord and is necessary but not sufficient for its differentiation [13] [27][28].

Xenopus laevis has two homeologs of hairy2, a member of the HES family of transcription factors [10] [29][30]. When overexpressed, both homeologs repress bra at gastrula stage [31][32], but at the same time, they induce the ectopic expression of organiser-specific genes on the ventral side [31] [33]. Although the effects of hairy2a on later development were not previously studied, hairy2b overexpression was able to induce an incomplete secondary axis [33]. The knock-down of hairy2a increased bra expression in the gastrula organiser, and an anterior expansion of the notochordal bra domain was observed in neurulae depleted from hairy2b [31] [33], indicating that hairy2 regulates bra expression on the dorsal midline (DML) structures or their precursors in the organiser.

Hairy2 belongs to the family of bHLH-Orange transcriptional repressors [29] [34]. It is expressed in the prospective ectoderm in blastulae, in the deep layer of the Spemann’s organiser (SO) and the notoplate during gastrulation, and later, in the prechordal mesoderm (PM), the floor plate (FP) and the neural crest cells (NCC) [30][31],[35][36]. Its dynamic expression in the borders of several territories is consistent with the role of hairy2 in controlling tissue demarcation, for example, favouring the FP fate at the expense of the notochord and maintaining the identity of the anterior PM by repressing specific genes of neighbouring tissues [31] [33],[36][38].

The basic body plan of the vertebrates is generated according to early events triggered by the Wnt/β-catenin pathway. After fertilization, the stabilization of the maternal β-catenin protein at the site opposite to the sperm entry point determines the future dorsal side of the embryo, where the Spemann-Mangold organizer will arise [39][42]. Dorsal development must also be restricted, and in order to prevent axial duplications, the accumulation of nuclear β-catenin in the ventral side of the embryo is prevented by maternal control mechanisms [41] [43][45]. Besides, zygotic mechanisms involving Wnt8 and BMPs are also necessary to maintain the inhibition of dorsal development in ventral locations [46]. In this paper we delve into this issue by showing that repression of bra target genes produce strikingly similar phenotypes to those of hairy2a overexpression, including the induction of a secondary trunk. The ectopic dorsal development induced after interfering with bra function was rescued by hairy2 depletion. Thus, we conclude that normal development requires an intact bra function to exclude the hairy2 domain from the non-organiser field in order to maintain the inhibition of dorsal axial fates in the ventral marginal zone, thus impeding the formation of spurious incomplete secondary axis.

Materials and Methods

Embryological Manipulations, RNA Synthesis, Morpholinos, and Injections

All animal studies in this report followed the rules and protocols approved by the Institutional Review Board for the Care and Use of Laboratory Animals (CICUAL) in the School of Medicine, University of Buenos Aires, Argentina. Albino and wild-type Xenopus laevis embryos were obtained using standard methods, staged and fixed with MEMFA [47]; [48]. Synthesis of capped mRNAs and the templates for bra and braEnR mRNAs were described elsewhere [13] [23],[49]. The full-length hairy2a cDNA construct fused to 6 myc-tag epitopes in pCS2+ (MT-hairy2a), kindly provided by Dave Turner, was digested with Not I and transcribed with SP6 RNA polymerase. Overexpression experiments for hairy2a were performed with MT-hairy2a mRNA. The hairy2a and hairy2b morpholino antisense oligonucleotides (hairy2a MO, hairy2b MO) used in this study were previously shown to specifically knock-down the corresponding homeolog [31] [33]. As control morpholino (Control MO), we used the standard control oligonucleotide or the random control oligonucleotide 25-N (Gene Tools, LLC, OR, USA). Except for the latter, all MOs were modified with 3′-fluorescein. 1 ng of nuc-lacZ mRNA or 10 ng of Dextran Oregon Green 488, MW 10000, anionic lysine fixable (DOG, Molecular Probes, Invitrogen) were co-injected as tracers.

In situ Hybridisation, Immunodetections, and Histology

The preparation of digoxigenin-labelled antisense RNA probes and the whole-mount in situ hybridisation (ISH) procedure were performed as previously described [50], except that the proteinase K step was omitted. For double ISH for hairy2 and bra, fluorescein-labeled antisense RNA probes were prepared with fluorescein-12-UTP (Amersham). We used either a mix of a hairy2-digoxigenin+bra-fluorescein probes, or a mix of a hairy2-fluorescein+bra-digoxigenin probes. After hybridisation, embryos were washed and blocked according to the standard protocol and incubated first with one of the antibodies (1/2000 of anti-digoxigenin-AP or 1/5000 of anti-fluorescein-AP, Fab fragments, Roche). After revealing the first probe with either BCIP or Magenta Phos (Sigma), the corresponding antibody was removed with 0.1M glycine, pH 2.0 with 1% Tween-20 (40 minutes), followed by six 10 minutes washes in MAB buffer and 15 minutes incubation in blocking reagent before adding the second AP-conjugated antibody. In some double ISH we used a chordin probe labeled with biotin-16-UTP (Roche), as previously described [51]. The c-myc epitopes were revealed by horseradish peroxidase (HRP) immunohistochemistry or by immunofluorescence as previously described [31] [51]. For histology, 20 µm sections were taken in a microtome and mounted onto gelatine coated slides.

Ventralisations, Dorsalisations and Cell Death Analysis

Embryos were exposed for 1 minute to UV light (254 nm), or were incubated for 1 hour with 0.1 M LiCl [52]. Whole-mount TUNEL staining was based on a previous protocol [53], with some modifications. After removing the vitelline membrane, embryos were fixed 1 hr in MEMFA, extensively washed in PBS, and stored in ethanol at −20°C. After rehydration in PBS, they were permeabilised in 0.25% Triton-X100/PBS, extensively washed in deionised water, incubated for 1 hour in terminal deoxynucleotidyl transferase (TdT) buffer, and transferred to TdT buffer containing 300 U/ml TdT (Invitrogen) and 1 mM digoxigenin-dUTP (Roche). Incubation was carried out overnight at 23°C. The reaction was stopped in 1 mM EDTA/PBS for 2 hours at 65°C. Embryos were washed 4 times, 30 minutes each in PBS, and twice in 100 mM maleic acid, 150 mM NaCl, pH 7.5, 5 minutes each, at room temperature. Incubation in blocking reagent and staining with NBT/BCIP were performed as for ISH [50]. LysoTracker Red (Lys) staining was based on a previous protocol [54], with some modifications. Lys (Molecular Probes, Invitrogen) was diluted in dimethyl sulfoxide and was added to 0.1× Modified Barth’s Saline (MBS) to yield a final concentration of 75 µM. Embryos were incubated in this solution for 30 minutes at room temperature and then, they were extensively washed with 0.1× MBS and fixed.

Results

Hairy2 and bra Progressively Establish Complementary Domains

We compared the expression of hairy2 and bra during the blastula-gastrula transition. Embryos were bisected in the animal-vegetal plane, and each hemisection was incubated with either a bra probe or with a hairy2 probe that recognises the transcripts of both homeologs [30], and were revealed with NBT+BCIP, which render a strong blue-purplish precipitate. We also performed double ISHs, which were revealed with BCIP and Magenta Phos. Although the reactions with these substrates are less sensitive than those carried out with NBT+BCIP, they were useful to corroborate the results obtained with single ISH. At stage 9.5, hairy2 transcripts are present in an animal-vegetal gradient, with maximum levels in the animal hemisphere (Fig. 1A). In the MZ, its expression is lower and overlaps with the bra territory (Fig. 1A–C). Double ISH in hemisections of pigmented embryos, which allowed to recognise the D–V orientation, showed that at late blastula, hairy2 expression is higher dorsally, while bra expression is higher ventrally and more exclusively expressed in the most vegetal cells of the MZ, while it overlaps with hairy2 in cells located more animally in the MZ (Fig. 1C).

thumbnail
Figure 1. Bra and hairy2 establish complementary expression domains.

Distribution of hairy2 (A,D,G, blue purplish; C,F,J,K, magenta; I,L, turquoise) and bra transcripts (B,E,H, blue-purplish; C,F,J,K, turquoise; I,L, magenta). For double ISHs, the name of each probe is written with the colour corresponding to the substrate with which it was revealed. Stage 9.5 (A–C), 10.25 (D–F), and 10.5 (G–L). (A–F,J,K) ISH in bisected embryos. The animal side is oriented to the top. d: dorsal side, v: ventral side. In C–F and J, the embryos were bisected in the sagittal plane, as represented by the broken black line in L. In K, the embryo was cut in a para-sagittal plane, as represented by the broken white line in L. Two halves of the same embryo are shown in the pairs A,B and D,E. The photographs in A,D were flipped 180° to facilitate comparison with the corresponding contralateral hemisection shown in B,E, respectively. Red arrows in A,B,D,E point to the overlapping of bra and hairy2 expression in the MZ. Yellow arrows in D,E mark the incipient blastopore dorsal lip. White asterisks in F,J indicate the overlapping of hairy2 and bra expression on the dorsal side. In J, the white, black and yellow arrowheads point to the prechordal mesoderm, the dorsal NIMZ, and the posterior, pre-involuted axial mesoderm, respectively. (G–I) Whole embryos in vegetal views. Dorsal is oriented to the top. The red broken line demarcates the blastopore. Yellow arrowheads in G point to the NIMZ. The arc between green arrows in G–I,L corresponds to the organiser region, with higher levels of hairy2 (G,I,L) and lower levels of bra expression (H,I,L). (L) Dorsal view of the same embryo shown in I.

https://doi.org/10.1371/journal.pone.0054777.g001

The domains of these genes are better distinguished in early gastrulae (Fig. 1D–F), when bra expression increases considerably in the MZ (Fig. 1E,F). While in the ventral region, their domains are already complementary, in the dorsal one, an overlapping zone still persists (white asterisk, Fig. 1F). Around mid gastrula, hairy2 is strongly expressed throughout the non-involuting MZ (NIMZ), with highest levels on the dorsal NIMZ (Fig. 1G,I,L) [31], while bra demarcates a ring of cells in the involuting MZ (IMZ) in a more vegetal location in relation to hairy2 (Fig. 1H,I). This ring shows an arc of lower bra expression in the organiser region, coincident with the maximum expression of hairy2 in the dorsal NIMZ (Fig. 1G–I,L). The spatial relationship between hairy2 and bra expression corresponding to this arc can be better appreciated in Fig. 1J, which shows a section of an embryo cut through the sagittal plane, revealing that the strong expression of hairy2 corresponds to the prechordal mesoderm (white arrowhead) and to the dorsal NIMZ (black arrowhead). On the other hand, the arc of lower bra expression corresponds to DML populations which are still segregating and show co-expression of both genes (white asterisk), and to cells of the posterior pre-involuted axial mesoderm, which mainly show bra transcripts (yellow arrowhead). Fig. 1K corresponds to a section of an embryo cut through a para-sagittal plane, where the pre-involuted non-axial mesoderm expresses high levels of bra in a complementary fashion to the strong hairy2 expression in the surrounding cells. This complementarity is also clearly observed in the ventral region, where the bra domain is larger than in the dorsal side (Fig. 1J,K). In conclusion, the domains of bra and hairy2 become mutually excluded in the MZ, suggesting a possible antagonistic regulation between them.

Hairy2 and bra are Involved in an Antagonistic Regulation

We have previously shown that the injection of hairy2a mRNA represses bra throughout its domain, while depleting hairy2a increases the expression of bra within the organiser domain [31]. Now, we extended the analysis to the effects on bra expression in the whole MZ by blocking each hairy2 homeolog with specific MOs against them. At gastrula stage, hairy2a MO expanded the bra domain toward the animal pole, both in dorsal and ventral injections (Fig. 2B,D,E,F) and also increased bra expression in the organiser, hence the arc normally expressing lower levels of transcripts was undistinguishable from the remaining domain in the MZ (Fig. 2C,D,F). Hairy2b MO also expanded the bra domain (not shown), but hairy2a MO was more effective if the percentages of embryos with bra expansions were compared at the same doses (Fig. 2E; Table 1).

thumbnail
Figure 2. Hairy2 restricts the bra domain.

(A–D) Distribution of bra transcripts (blue-purplish) at stage 10.5. Turquoise staining corresponds to chordin (chd) transcripts in the organiser. Embryos were injected into a ventral (A,B,E,G) or dorsal (C,D,E–G) cell at the 4-cell stage with 1.25 ng (C,E), 2.5 ng (A,E–G), or 5 ng (E,G) of control MO; 1.25 ng (D,E), 2.5 ng (B,E–G), or 5 ng (E,G) of hairy2a MO; 1.25 ng (E), 2.5 ng (E,G), or 5 ng (E,G) of hairy2b MO, or with a mix containing 2.5 ng of hairy2a MO +2.5 ng of hairy2b MO (G). DOG fluorescence (insets) indicates the injected side. (A,B) Whole embryos in vegetal views. Dorsal is oriented to the top. (C,D) Dorsal views of gastrulae showing bra and chd expression in the MZ. The green arrows in B,D point to up-regulations or domain expansions of bra. White bars in B,D indicate the width of the bra domain. (E) Percentage of embryos with bra expansion in the MZ. Comparison between different doses of control MO (blue line), hairy2b MO (pink line) and hairy2a MO (orange line). See Table 1. (F) Percentage of embryos displaying the arc of low bra expression that is normally observed in the organiser region (blue) or with an increase of bra expression in the organiser (light orange). (G) Percentage of embryos with bra expansion in the MZ. Comparison between the simultaneous knock-down of hairy2a and hairy2b and different doses of control MO, hairy2a MO, and hairy2b MO. n indicates the total number of injected embryos.

https://doi.org/10.1371/journal.pone.0054777.g002

Since we obtained the best performance of hairy2b MO alone with 2.5 ng (Fig. 2E), we analysed the simultaneous knock-down of both homeologs after injection of a mix consisting of 2.5 ng of each MO. This resulted in an expansion of the bra domain in 56% of the injected embryos (Fig. 2G). Therefore, the simultaneous knock-down did not increase the effect on bra in comparison with 2.5 ng of each MO alone (57% for hairy2a MO, 44% for hairy2b MO; Fig. 2G) as it would be expected if both homeologs equally contribute to modulate bra expression. Moreover, 5 ng of hairy2a MO alone resulted in a higher proportion of embryos with bra expansions (76%) than that obtained with 5 ng of the mix containing 2.5 ng of each MO in the simultaneous knock-down (56%) (Fig. 2G), indicating that the hairy2b MO could not efficiently supplement hairy2a MO. These results together support the idea that the effects of hairy2a MO on bra expression are consistently more potent than those of the hairy2b MO, suggesting that normally, hairy2a is more relevant than hairy2b in restricting the bra territory.

Our results indicate that the expression of hairy2 throughout the NIMZ confines bra to the IMZ, preventing its animal expansion, and also restricts bra expression in the organiser.

Next, we wondered whether bra could affect hairy2 expression. Since the three Xenopus laevis bra genes have nearly identical expression patterns [5] [7][8], loss-of-function analysis in this species is more complex. Therefore, to interfere with Bra activator function, we injected braEnR mRNA, a dominant-negative bra construct classically used to directly repress bra target genes, in which the transcription activator domain of Bra was replaced with EnR [13] [55]. The specificity of this kind of constructs to interfere with the function of T-box transcription factors in vivo was carefully demonstrated in previous works in Xenopus laevis. When the EnR domain was fused to a particular T-box transcription factor, only the specific wild type protein, but not other members of the family, was able to rescue the phenotypic effect of the antimorphic construct. Moreover, embryos injected with the EnR domain alone were indistinguishable from uninjected controls [13] [56][57].

We injected 1 ng of braEnR mRNA in 1 cell at the 4-cell stage and analysed hairy2 expression at gastrula and neural plate stages. In gastrulae, braEnR increased hairy2 transcripts and produced a vegetalward expansion of its domain, so that hairy2+ cells invaded the ring otherwise corresponding to bra in the MZ (78%, n = 35/45), either in ventro-lateral locations after ventral injections (79%, n = 11/14; Fig. 3B) or in dorso-lateral locations after dorsal injections (77%, n = 24/31, Fig. 3C,D). At neural plate stage, when hairy2 expression is clearly resolved into distinct subdomains, including the neural plate borders and the DML (Fig. 3E), the up-regulation of hairy2 could be very clearly distinguished in all of the braEnR-injected embryos (100%, n = 65). At this stage, ventral injections of braEnR resulted in strong ectopic expression of hairy2 in ventro-lateral locations, including the circumblastoporal collar (100%, n = 36; Fig. 3F), while dorsal injections expanded the DML hairy2 subdomain (100%, n = 29; right inset in Fig. 3F). Both in injected gastrulae and neurulae the blastopore was more opened in comparison with sibling controls (Fig. 3A–F).

thumbnail
Figure 3. Bra restricts hairy2 expression.

Distribution of hairy2 at gastrula (A–D,G,H) and neural plate stage (E,F). Embryos were injected into a ventral (B,F,H) or dorsal cell (C,D) at the 4-cell stage with 1 ng of braEnR mRNA (B–D,F) or with 1 ng of bra mRNA (H). DOG fluorescence (left insets) indicates the injected side. The upper right inset in F shows an embryo injected into a dorsal cell at the 4-cell stage with 1 ng of braEnR mRNA with its corresponding DOG image at the left. Embryos in B–D, F, and H were fixed when sibling controls as shown in A,E, and G reached stage 11.5, 14, and 10.5, respectively. Embryos are shown in vegetal (A,B,D,G,H), dorsal (C), and posterior-dorsal views (E,F). Yellow asterisks mark the normal expression domain of hairy2 in the organiser region (A–D,G,H) or in the DML (E). Black arrowheads point to the hairy2 subdomains demarcating the neural plate border. The black broken line demarcates the blastopore. The green arrows point to ectopic expression of hairy2. Comparison of the distance from the blastopore to the vegetal border of the hairy2 domain between the injected and the non-injected sides (yellow bars) show the vegetalward expansion of hairy2 expression. Red arrows point to down-regulation of hairy2.

https://doi.org/10.1371/journal.pone.0054777.g003

We also performed gain-of-function experiments by injecting bra mRNA. In contrast to the interference with bra function, bra overexpression drastically repressed hairy2 (80% n = 25/31; Fig. 3H).

Since bra overexpression and braEnR gave opposite results on hairy2 expression, we conclude that bra is necessary to maintain hairy2 repressed in the bra territory, preventing the vegetal expansion of hairy2 throughout the entire MZ and the ulterior expansion of the hairy2 domains in ventro-lateral and DML locations. In conclusion, our results indicate that hairy2 and bra reciprocally confine their normal domains through an antagonistic regulation.

Inhibition of bra Function, Like hairy2a Overexpression, Induces Ectopic Dorsal Markers and Impairs Morphogenetic Movements

Since braEnR expanded hairy2 in the ventrolateral marginal zone (VLMZ), and we previously showed that hairy2a overexpression induces ventral expression of DML markers [31], we wondered whether the latter effects could be reproduced by braEnR. Remarkably, ventrolateral injections of braEnR mRNA induced ectopic chd, foxA4a and not I in the VLMZ in similar proportions than those obtained with hairy2a mRNA (Fig. 4).

thumbnail
Figure 4. Ectopic induction of DML markers after interfering with bra function or overexpressing hairy2a.

(A,C,E) Uninjected sibling controls. Embryos were ventrally injected into one cell at the 4-cell stage with 1 ng of braEnR (B,B’,D,D’,F,F’,G) or of hairy2a mRNAs (G). Expression of chd (A,B,G), foxA4a (C,D,G) and not I (E–G) at stage 10.5. The injected side was detected by DOG fluorescence (B’,D’,F’). Green arrows point to the ectopic expression of the analysed markers. At this stage, the markers analysed are normally expressed only in the organiser [63] [112][113]. (G) Percentage of gastrulae ventrally injected with braEnR (blue bars) or hairy2a mRNAs (light orange bars) expressing ectopic chd, foxA4a, and not I (chd: 100%, n = 26; foxA4a: 100%, n = 14; not I: 79%, n = 11/14, for braEnR; chd: 92%, n = 23/25; foxA4a: 80%, n = 8/10; not I: 60%, n = 3/5, for hairy2a).

https://doi.org/10.1371/journal.pone.0054777.g004

At neurula stage, the severity of the alterations with braEnR depended on the distribution of the injected mRNA, as revealed by the tracer. According to the anterior-posterior extension of the dorsal axis, embryos were classified into two phenotypes. In the less severe one (phenotype A; Fig. 5), cells expressing ectopic DML markers remained arrested around the blastopore, which was unable to complete its closure. Since bra function was barely impaired in the dorsal region in these embryos, they tended to develop a normal dorsal axis. The phenotype B (Fig. 5) is more severe, since braEnR mRNA was widely distributed, thus affecting also the dorsal side. Both the primary axial cells and those expressing ectopic DML markers remained close to the blastopore. Both phenotypes were remarkably similar to those obtained at neurula stages after overexpression of hairy2a [31].

thumbnail
Figure 5. Effects of braEnR on morphogenesis and on DML markers at neurula stage.

Ventral injections in one cell at the 4-cell stage with 1 ng of braEnR mRNA +10 ng of DOG (B–C’,E–F’,H–I’,K–L’). Expression of foxA4a (A–C), shh (D–F), not I (G–I), and chd (J–L), when control embryos reached stage 14 (A,D,G,J). At this stage, these markers are normally expressed in the DML [63] [112][114]. Arrows point to the dorsal axis. Yellow arrowheads point to cells expressing ectopic DML markers. Injected embryos are shown with the corresponding fluorescence image at the right (DOG) to identify the descendants of the injected cells. Embryos were classified into phenotypes A and B according to the grade of anterior-posterior extension of the dorsal axis. Embryos are shown in dorsal (A,D,G,J), posterior-dorsal (B,E,H) or posterior (C,F,I,K,L) views. The percentages of embryos displaying each phenotype are indicated in B,C,E,F,H,I,K,L. For each marker, % A+B indicates the percentage of affected embryos (A plus B phenotypes) and N, the total number of embryos analysed.

https://doi.org/10.1371/journal.pone.0054777.g005

In conclusion, the repression of bra-target genes or gain of function of hairy2a equally promoted the ectopic expression of DML genes in the VLMZ and altered the morphogenetic movements during gastrulation, thus affecting the involution of cells with dorsal axial characteristics.

Impairing of bra Function, as well as hairy2a Overexpression, Produces Opposite Effects on chd, Depending on the Dorsal-ventral Context

It was interesting to notice that in the ventral region of embryos injected with braEnR, the patch of cells in which bra activity was blocked (green fluorescent cells in Fig. 5K,K’; brown+ cells in Fig. 6A,E) did not express chd, indicating that these cells were unable to ectopically activate chd. This marker was mainly activated in the vicinity of those cells, resembling the effects of hairy2a (not shown) or hairy2b overexpression [34]. These results suggest that ventral cells in which bra target genes were repressed with braEnR or in which hairy2 was activated secrete a signal that, in turn, induces chd in neighbouring cells. While in the ventral side, both braEnR and hairy2 overexpression were able to induce ectopic chd expression, on the dorsal side, chd was repressed, and this occurred in the same cells where bra was blocked or in which hairy2 was overexpresed (100%, n = 8 for braEnR, Fig. 6I; 83%, n = 12 for hairy2a, not shown), in agreement with previous results of hairy2a or hairy2b overexpression [31] [34].

thumbnail
Figure 6. Opposite regulation of chd by hairy2a or braEnR depending on the D

V context. Ventral (A,E) or dorsal (I) injections of 1 ng of braEnR mRNA into one cell at the 4-cell stage. E corresponds to a high-power view of the embryo shown in A. Embryos were fixed when sibling controls reached stage 14 (A) or 11 (I). Red arrows in A,I indicate the descendants of the injected cells. In A, we made use of the myc-tag epitopes fused to the braEnR construct to reveal the distribution of the encoded recombinant protein (brown dots). The injections in I included 10 ng of DOG (inset). Embryos are shown in posterior (A), vegetal (B–D,F–I) or vegetal and dorsal (J) views. (B–D,F–H,J–L) Analysis of chd expression when sibling controls (J) reached stage 11 in ventralised (B–D) or dorsalised (F–H) embryos, which were left uninjected (B,F) or were injected with 1 ng of hairy2a (C,G,K,L) or of braEnR mRNAs (D,H,K,L) before the first cleavage. Yellow asterisks mark the gaps of chd-negative cells. (K) Percentage of embryos expressing (light orange bars) or not expressing (blue bars) chd. UV-irradiated, uninjected embryos (UV) showed complete loss (67%) or very weak expression (37%) of chd (n = 88). Chd was expressed in 80% or 87% of UV-irradiated embryos injected with 1 ng of braEnR (braEnR/UV, n = 55) or of hairy2a mRNAs (hairy2a/UV, n = 96), respectively. (L) Percentage of embryos with asymmetric (blue bars) or radially symmetric chd expression (light orange bars) around the blastopore. The maximum effect of LiCl alone was found in 50% of embryos, which showed complete radial chd expression (LiCl, n = 53). Chd expression decreased in 88% or 100% of dorsalised embryos that were injected with 1 ng of braEnR (braEnR/LiCl; n = 41) or of hairy2a mRNAs (hairy2a/LiCl, n = 23), respectively.

https://doi.org/10.1371/journal.pone.0054777.g006

We made use of UV-irradiation or LiCl-treatment to confirm whether impairing of bra function or overexpression of hairy2a activate the dorsal marker chd in a ventral context but repress it in a dorsal one. UV-irradiation of fertilised eggs destabilises maternal β-catenin, producing embryos with ventral character. LiCl stabilises β-catenin, resulting in completely dorsalised embryos [39].

Chd was drastically down-regulated in UV-treated, uninjected embryos (Fig. 6B,K), but when they were injected with hairy2a or braEnR mRNAs at the 1-cell stage, chd transcripts appeared around the entire MZ (Fig. 6C,D,K). In spite of the increase of chd, these embryos showed a blastopore more open in relation to sibling controls (Fig. 6J). Therefore, chd was induced in ventralised embryos in which bra-target genes were repressed or in which hairy2a was overexpressed, but those chd+ cells could not involute like normal dorsal cells.

In LiCl-dorsalised embryos, chd expression was totally radial and symmetric throughout the MZ (Fig. 6F,L). Chd expression evidently decreased in dorsalised embryos that were injected with hairy2a or braEnR mRNAs at the 1-cell stage. Radial expression was lost in these embryos, which showed gaps of chd-negative cells and a blastopore noticeably more open in relation to dorsalised, uninjected embryos (Fig. 6G,H,L). These results confirm that in a dorsal context, repressing bra-target genes as well as overexpressing hairy2a inhibited chd and disturbed the morphogenetic movements of dorsal cells.

An Intact bra Function Prevents the Induction of Spurious Incomplete Axes

Most of the embryos injected into one cell at the 4-cell stage with either braEnR or hairy2 mRNA failed to complete gastrulation and showed generalised disturbances at tailbud stages (Fig. 7A,B). Thus, we performed more localised injections at the 16-cell stage into one ventral blastomere normally contributing to the posterior region of the embryo, but not to the notochord (Fig. 7C) [58]. We obtained dose-dependent phenotypes ranging from spina bifida to protuberances, notably including evident secondary axes without heads. Occasionally, tail-like structures were found (Fig. 7, Fig. 8, and Table 2). The most ventral blastomeres contributed in a greater proportion to the secondary axes (Fig. 7I, Table 2), which contained nervous system, since they expressed the neural markers sox2 (Fig. 8A,B) and nrp1 [59] (hairy2a 100%, n = 5; not shown). The most anterior marker otx2 was absent, indicating the lack of forebrain (Fig. 8C,D). Some of them contained midbrain cells (Fig. 8F,G), while most developed hindbrain structures (Fig. 8H,I). We observed a relatively well organised secondary trunk with myoD+ cells arranged in somites (Fig. 8C,E) and some superficial, scattered chd+ cells, apparently not organised in a notochordal rod (Fig. 8J,K). Sonic hedgehog (shh) was detected in the secondary axes (Fig. 8L,M). Therefore, the ectopic axes induced by interfering with bra function closely resemble those obtained by hairy2a overexpression.

thumbnail
Figure 7. Effects of ventral injections of braEnR and hairy2a mRNAs.

(A,B) Ventral injections into one ventral cell at the 4-cell stage produced spina bifida. (C) Mesodermal fate map of the blastomeres of the 16-cell-stage embryo (modified from [58]). Notochord (pink), somites (light blue), ventral blood islands (VBIs, light violet), Spemann’s organiser (SO). Green dots indicate the injection sites. (D–G) Ventral injections as in C occasionally induced small bumps (arrows, D,E) and tail-like structures (arrows, F,G). Embryos shown in D–G were hybridised with a sox2 probe. (H) Percentage of embryos with secondary axes+protuberances. Comparison between different doses of braEnR mRNA injected into the AB3 (blue), AB4 (pink), CD3 (yellow) or CD4 (light blue) blastomere. (I) Percentage of embryos with secondary axes classified according to the injected blastomere. The bars correspond to braEnR (blue) or hairy2a (light orange) mRNAs injections. Injections included nuc-lacZ mRNA (A,B) or DOG (D–G) as lineage tracers. Embryos were fixed when sibling controls reached stage 28 and are shown in dorsal (A,B,D,G), dorso-lateral (E) or lateral (F) views. The amounts of mRNA injected are indicated in each figure.

https://doi.org/10.1371/journal.pone.0054777.g007

thumbnail
Figure 8. Characterisation of the secondary axes induced by braEnR or hairy2a.

A–M: Ventral injections into one cell at the 16-cell stage with 1 ng of braEnR (A,C,F,H,J,L) or of hairy2a mRNAs (B,D,E,G,I,K,M). Expression of sox2 (A,B), otx2 (C,D), myoD (C,E), en2 (F,G), krox20 (H,I), chd (J,K), and shh (L,M) at stage 28. For shh, embryos were overstained to preserve color during histological procedures. The injected side was detected by DOG fluorescence (insets). Red arrowheads point to the expression of different markers in the secondary axes. The red asterisk in C,D indicates the absence of otx2 expression in the secondary axis. The frequency for each marker is shown as percentage of the total number (n) of secondary axes analysed with the corresponding probe. Sox2 is a general neural marker [115]. Otx2 is normally expressed in the eyes and forebrain [116]. MyoD marks the somitic mesoderm [117][118]; en2, the midbrain-hindbrain boundary [119]; krox20, the third and fifth rhombomeres [120]; and chd and shh, the DML (see text for details). Embryos are shown with the primary axis in dorsal (A,D,F,G,K), dorso-lateral (L,M) or lateral (B,C,E,H,I,J) views.

https://doi.org/10.1371/journal.pone.0054777.g008

thumbnail
Table 2. Effects of ventral injections of braEnR or hairy2a mRNAs at tailbud stages.

https://doi.org/10.1371/journal.pone.0054777.t002

Histological analysis of braEnR-injected embryos revealed that sox2 transcripts were located in the germinal neuroepithelium of the primary neural tube, but throughout all layers of the secondary one (Fig. 9A–A’’), which also showed shh expression throughout its ventral half, indicating the presence of an expanded FP (Fig. 9B–B’’). Strikingly, no shh+ cells were detected ventrally to the FP, consistent with the observation that chd+ cells failed to organise a notochord and that the dorsal mesodermal midline was instead occupied by myoD+ somitic structures (Fig. 9C–D’’).

thumbnail
Figure 9. Histological analysis of secondary axes induced by braEnR.

Expression of sox2 (A,A’), shh (B,B’), chd (C,C’) or myoD (D,D’). A’,A’’;B’,B’’;C’,C’’; D’,D’’ are transverse sections of the embryos shown in A,B,C,D, respectively. The yellow broken lines show the plane of the sections. (A’’,B’’,C’’,D’’) Nuclear Hoescht staining. The insets in A’ and B’ correspond to high-power views of the secondary neural tubes. Red arrows point to the secondary neural tube (A’,A’’,B’,B’’) or to the somites (D’,D’’) in the secondary axes. nc: notochord; nt: neural tube; s: somites.

https://doi.org/10.1371/journal.pone.0054777.g009

At late gastrula, the PM cells express the transcriptional repressor goosecoid (gsc), when they have already segregated from the chordal population [60][62]. Ventral injections of braEnR or hairy2a mRNAs induced ectopic gsc in the VLMZ at stage 11 (Fig. 10B,C,G). Since chd is also normally expressed in the PM [63], the scattered chd+ cells observed in some secondary axes might be of PM characteristics, since the secondary axes failed to develop an organised notochordal rod (see Discussion).

thumbnail
Figure 10. BraEnR or hairy2a ventrally induce gsc and cer, but cannot consistently activate ectopic BCNE markers.

Ventral injections into one cell at the 4-cell stage with 1 ng of braEnR (B,E,I,L) or of hairy2a mRNAs (C,F,J,M). (A,D,H,K) Sibling controls. Expression of gsc (A–C) and cer (D–F) at stage 11. Expression of chd (H–J) and Xnr3 (K–M) at stage 9. Embryos are shown in vegetal (A–F), dorsal (H,K) or ventral views (I,J,L,M). Left insets in I,J,L,M are dorsal views of the embryos shown in the corresponding photography. The injected side was detected by DOG fluorescence (right insets). The green arrows point to the ectopic expression of the markers analysed. (G) Percentage of gastrulae ventrally injected with braEnR (blue bars) or hairy2a mRNAs (light orange bars) expressing ectopic gsc, and cer (gsc: 68%, n = 40/59; cer: 80%, n = 25, for braEnR; gsc: 100%, n = 25; cer: 83%, n = 30, for hairy2a). (N) Percentage of blastulae ventrally injected with braEnR mRNA (blue bars) or hairy2a mRNA (light orange bars), expressing ectopic chd or Xnr3 (chd: 75%, n = 23; Xnr3∶11%, n = 26, for braEnR; chd: 60%, n = 23; Xnr3∶10%, n = 51, for hairy2a).

https://doi.org/10.1371/journal.pone.0054777.g010

A transcriptional cascade triggered by maternal β-catenin in the dorsal side is necessary to establish the complete embryonic axis [40] [64]. An early output of this pathway is the blastula Chordin- and Noggin-expressing center (BCNE), which is important for the formation of the head and appears when zygotic transcription begins [65]. During gastrulation, secreted Cerberus (Cer) from the anterior mesendoderm is necessary to fulfil head development. When ventrally overexpressed, Cer induces secondary heads without trunk-tail structures [61] [66].

Ventral injections of braEnR or hairy2a mRNAs into one cell at the 4-cell stage induced cer expression in the gastrula VLMZ (Fig. 10E,F,G), indicating that the absence of the anterior brain is unlikely attributable to the lack of Cer signalling. We wondered whether the injections were able to induce an ectopic BCNE at late blastula. This center contains the precursors of the forebrain and the SO and expresses the neural inducers chd and nodal-related 3 (Xnr3) [65] [67][69]. While chd was detected in the ventral side, albeit in much fewer cells than in the endogenous center, ectopic Xnr3 was rarely and faintly detected (Fig 10I,J,L–N). In conclusion, although repressing bra target genes or an excess of hairy2a activity induced ventral expression of the head inducer cer, this was not preceded by a consistent activation of BCNE markers.

Hairy2a Overexpression, Like Ventral Inhibition of bra Function, Induces Cell Death in the Posterior Region

It was reported that ventral injections of braEnR promoted apoptosis in the posterior region since neurula stage [27]. When we injected the same construct at the 16-cell stage, the maximum frequency of secondary axes was around 20%, but they were very rarely observed in embryos ventrally injected with braEnR or hairy2a mRNAs into one cell at the 4-cell stage, which showed severe disturbances. Thus, we wondered whether cell death could explain the variability of the frequencies. To answer this question, we compared cell death between sibling embryos injected with braEnR mRNA or hairy2a mRNA into one ventral cell at the 4-cell stage, or into the most ventral-vegetal blastomere (CD4) at the 16-cell stage.

The TUNEL assay detects the 3′-OH ends generated by the DNA fragmentation occurring during apoptosis [70]. We performed the TUNEL assay at stage 14 and classified the embryos into three categories, according to the appearance of TUNEL+ cells in the posterior region at the neural plate stage, as follows: TUNEL+/− (none or low), TUNEL++ (moderate), and TUNEL+++ (high) (Fig. 11A,C,D–I, Table 3). Apoptotic cells in controls were found in the anterior neural plate, but the posterior region showed none or a few TUNEL+ cells (Fig. 11A and light blue bars in Fic. 11C), as previously reported [71]. Only embryos injected with either mRNA into one ventral cell at the 4-cell stage and one embryo injected with the highest amount (2 ng) of hairy2a mRNA into the CD4 blastomere at the 16-cell stage (10%, n = 10) showed high levels of apoptosis (TUNEL+++) in the posterior region (red bars in Fig. 11C; Fig. 11D,G; Table 3), and the dying cells were found around the unclosed blastopore. In comparison with the embryos ventrally injected at the 4-cell stage, apoptosis notably decreased in embryos injected with either mRNA into the CD4 blastomere at the 16-cell stage. Some of these embryos were indistinguishable from controls (light blue bars in Fig. 11C; Fig. 11E,H; Table 3) in terms of TUNEL staining, others presented moderate levels of apoptosis in the posterior region (orange bars in Fig. 11C; Fig. 11F,I; Table 3), and except for the aforementioned embryo injected with 2 ng of hairy2a mRNA, they never fell within the TUNEL+++ category (Fig. 11C, Table 3).

thumbnail
Figure 11. Apoptosis in embryos injected with braEnR or hairy2a mRNAs occurs in the posterior region.

Embryos were injected with 1 or 2 ng of braEnR (C,J–K’’ and C,D–F, respectively) or with 1 or 2 ng of hairy2a mRNAs (C,L,M’’ and C,G–I, respectively) into one ventral cell at the 4-cell stage (1V/4 cells, C,D,G,J,J’,L,L’) or into the CD4 blastomere at the 16-cell stage (C,E,F,H,I,K–K’’,M–M’’). (A,B) uninjected sibling controls. Left inset in A corresponds to the dorsal view of the same embryo. TUNEL assays (A,C,D–I) or Lys staining (B,J–K’,L–M’) were performed at stage 14 or 28, respectively. (K’’,M’’) show ISH of sox2 in the same embryos shown in K,M, respectively, to reveal the secondary axis. Turquoise, orange and red arrowheads point to low, moderate or high levels of apoptosis, respectively. Left insets in D–I,J,K,L,M show the injected side revealed by DOG fluorescence. Arrows in K,K’’,M,M’’ point to secondary axes; white arrows in J’,L’ point to spina bifida, and in M’, to the unclosed blastopore. (J’,K’,L’,M’) merge of images shown in J,K,L,M, respectively. Embryos are shown in posterior (A,D–I) or lateral (B,J–M’’) views.

https://doi.org/10.1371/journal.pone.0054777.g011

thumbnail
Table 3. TUNEL analysis in embryos injected with braEnR or hairy2a mRNAs.

https://doi.org/10.1371/journal.pone.0054777.t003

The fluorescent dye Lys accumulates in lysosomes and was previously used as indicator of cell death [54] [72]. At stage 28, faint Lys staining was scattered in control embryos, with low staining in the posterior region (Fig. 11B). Although embryos injected with braEnR or hairy2a mRNA into the CD4 blastomere developed a secondary axis, they were indistinguishable from uninjected controls in terms of Lys staining (93%, n = 15 for braEnR; 100%, n = 14 for hairy2a; Fig. 11K–K’’, M–M’’). In contrast, very strong staining was detected in the most posterior region of embryos injected with either mRNA in one ventral cell at the 4-cell stage, and they exhibited evident morphogenetic distortions (88%, n = 8 for braEnR; 100%, n = 9 for hairy2a; Fig. 11J,J’,L,L’). Moreover, Lys staining partially co-located with DOG fluorescence (Fig. 11J’,L’), indicating that not only the cells in which bra target genes were repressed or hairy2a was overexpressed were dying, but they could also induce cell death in neighbouring tissues. Notably, the secondary axes induced in CD4-injected embryos were observed in a region without appreciable cell death (arrows, Fig. 11K,K’’,M,M’’).

Therefore, the increase in cell death and the appearance of severe disturbances is a feasible explanation for the variable incidence of secondary axes in our experiments. In conclusion, bra should be fully functional and hairy2 should be maintained in low activity in the ventral half of the embryo for the survival of the posterior cells.

Bra Prevents Ectopic Dorsal Induction by Preventing hairy2 Derepression

Since braEnR and hairy2a overexpression gave identical results, we investigated whether the phenotypes observed when we blocked bra function were specifically produced by the derepression of hairy2. When braEnR mRNA was co-injected with either hairy2 MO separately into one ventral cell at the 4-cell stage, there was not a significant change in the proportion of gastrulae with ectopic chd in comparison with braEnR mRNA+control MO co-injections (Fig. 12A,C). Notably, when braEnR mRNA was co-injected with hairy2a+b MOs into one ventral cell at the 4-cell stage, most of the embryos did not reveal ectopic chd (Fig. 12D,C). In neurulae, after co-injection of braEnR mRNA with hairy2a+b MOs, the proportion of unaffected embryos increased nearly two-fold in comparison with braEnR mRNA+control MO co-injections (Fig. 12F), while the most severe phenotype (phenotype B, Fig. 12B) completely disappeared (Fig. 12F). Embryos with rescued phenotypes did not show ectopic chd, achieved a normal elongation of the notochord and completed the blastopore closure (Fig. 12E,F). In conclusion, an intact bra function is necessary to maintain both homeologs of hairy2 repressed in the VLMZ in order to prevent the ectopic expression of DML genes in this region and also contributes to keep normal morphogenetic movements during gastrulation.

thumbnail
Figure 12. The effects of braEnR are mediated by hairy2.

A–F: Ventral injections in one cell at the 4-cell stage with braEnR mRNA plus the following MOs: control MO (A–C,F), hairy2a MO (C), hairy2b MO (C), or hairy2a+b MOs (C–F). Expression of chd at stage 10.5 (A,C,D) or stage 14 (B,E,F). The doses of injection are indicated in C,F. The injected side was determined by DOG fluorescence (insets). Red arrows point to the ectopic chd+ cells. Blue arrows indicate the location of the injection in rescued embryos. Embryos are shown in vegetal (A,D), posterior-dorsal (B) or dorsal (E) views. (C) Percentage of injected gastrulae expressing ectopic chd in the VLMZ. n: total number of injected embryos. (F) Percentages of injected neurulae showing normal morphology and chd expression (Normal phenotype), or Phenotype A or B (as in Fig. 4). Blue bars: embryos injected with braEnR mRNA plus control MO (n = 36). Light orange bars: embryos injected with braEnR mRNA plus hairy2a+b MOs (n = 31).

https://doi.org/10.1371/journal.pone.0054777.g012

Discussion

Based on the dynamics of their expression and on the results of the functional experiments with localised injections, we present evidence for a previously unrecognised role for bra in preventing the appearance of trunk duplications by maintaining the dorsal fates repressed in the ventrolateral IMZ of the gastrula, through the exclusion of hairy2.

Bra is Necessary to Prevent the Derepression of hairy2 and the Occurrence of its Inductive Functions in the Ventral Side

The following line of evidence supports that hairy2 mediates the effects of bra interference: a) hairy2a and braEnR mRNAs injections result in striking similar phenotypes; b) braEnR potently induces ectopic hairy2 expression; c) the knock-down of both hairy2 homeologs significantly rescues the ectopic expression of the dorsal marker chd and contributes to restore the normal development of embryos in which bra function was interfered.

Studies on the role of bra have been mainly focused on posterior mesoderm development, but there is also evidence for the involvement of bra in delimiting contiguous territories during Xenopus embryogenesis. For example, among the two subpopulations of the axial mesoderm, bra inhibits the active cell migration behaviour of the PM and promotes the convergent-extension movements of the notochordal cells [27][28]. Within the presumptive non-axial mesoderm, bra is expressed in the animal half of the IMZ and restricts the expression of the ventral mesoderm inducer Xnr2 (Xenopus nodal-related 2) to the vegetal half [58]. In zebrafish, bra (ntl) disfavours medial FP development but promotes the notochordal fate [73]. Now, we provide evidence that bra is also necessary to prevent the specification of DML fates within the non-organiser mesoderm field by precluding the vegetal expansion of the hairy2 domain into the ventrolateral IMZ, thus impeding the appearance of spurious, incomplete axes (Fig. 13).

thumbnail
Figure 13. Bra maintains the embryonic dorsal-ventral pattern by repressing hairy2 in the MZ.

Diagrams of gastrulae in vegetal views. The black broken line represents the blastopore lip. (A,C,E) In control embryos, hairy2 is expressed in the NIMZ with maximum levels in the dorsal side (blue gradient), while bra is expressed in a more vegetal location, in the IMZ, with highest expression in the ventral region (pink gradient), while a gap of lower expression is observed in the organiser region (light pink). Through mutual antagonism, both genes establish complementary patterns and participate in a regulatory mechanism that ensures the normal development of the embryo (A). Thus, organiser specific genes (OG, green) are expressed only in the dorsal region (C) and the embryogenesis proceeds resulting in a normal larva (E). (B,D,F) However, if this equilibrium is perturbed and bra is inhibited in the VLMZ, hairy2 is derepressed, resulting in a vegetalward displacement of its domain (B), leading to the ectopic activation of OG in the VLMZ (D). This finally results in the establishment of a headless secondary axis without notochord (F). Bra should be acting indirectly, through the induction of a repressor of hairy2 (X).

https://doi.org/10.1371/journal.pone.0054777.g013

Bra overexpression inhibited hairy2 in the MZ, while braEnR gave the opposite result, indicating that the effect of the antimorphic construct was specific. These results support the hypothesis that during normal development, bra is necessary to maintain hairy2 repressed in the bra territory, preventing the vegetal expansion of hairy2 throughout the entire MZ, since both dorsal and ventral injections of braEnR resulted in up-regulation of hairy2. Since bra is a transcriptional activator [13], the most likely explanation is that bra induces a transcriptional repressor that inhibits hairy2 in the ventrolateral IMZ (X, Fig. 13). A similar model was proposed to explain the repression of Xnr2 by bra on the MZ [74]. Studies in other model organisms like mouse and zebrafish described that Bra (Ntl) binds regulatory regions of several genes, including positive and negative effectors of posterior development [75][77]. It will be necessary to elucidate how bra is related to hairy2 in this complex gene network.

Hairy2 behaves as a cell context factor which predisposes the cells towards a particular decision in response to extrinsic signals during muscle and neural differentiation [78]. Notably, hairy2 can function as a transcriptional repressor or as a non-transcription factor with inductive properties, depending on the regional context. The latter function, mediated by the C-terminal WRPW tetrapeptide motif, is responsible for the ventral activation of chd by hairy2b in neighbouring cells [34]. Our results demonstrate that repressing bra target genes mimics the effects of hairy2 overexpression both in dorsal and ventral contexts, and this is explained because braEnR was able to induce hairy2 expression both in dorsal and ventral injections. Therefore, the inductive, non-cell autonomous function of hairy2a or braEnR on chd expression only arises in a ventral context, while in a dorsal one, both manifest a repressive, cell-autonomous behaviour.

All together, these results indicate that bra is necessary throughout the non-organiser IMZ to prevent the derepression of hairy2 and the occurrence of its inductive functions. On the other hand, dorsal injections of either braEnR or hairy2a result in chd repression and similarly impair morphogenetic movements, while hairy2 MOs are able to rescue these effects of braEnR. These results suggest that in the dorsal region, bra is necessary to prevent an excess of hairy2 transcriptional repressor activity, contributing to ensure the progress of morphogenetic movements with a normal balance in the specification of the DML precursors (notochord, FP, PM).

It is worthwhile noting that in mouse chimeric embryos in which embryonic stem cells mutant for the T-gene were injected into wild type blastocyst hosts, the posterior neuropore was unable to close, the mutant cells remained in the surface of the distant tail, and small lumps of tissue on the posterior region, tails with a branch or divided tails, or even a duplicated tail containing somites were observed [79][80]. This situation is more comparable than that of homozygous T-mutants to our manipulations in Xenopus embryos, where a group of ventral-posterior cells, in which bra target genes were repressed, were able to interact with a context with an intact bra function. The frog embryo offers the advantage to perform a more spatially controlled manipulation, and we could only obtain stronger and distinguishable secondary axes when we performed more localised injections of the antimorphic Bra in ventral-posterior cells.

Hairy2 Contributes to Confine bra Expression

There is some evidence that the definitive spatial pattern of bra is achieved through the release from repression in those regions where it is required rather than through specific activation [81]. The transcriptional repressors SIP1 (a member of the δEF-1 family) and Gsc (a homeodomain protein) directly bind the bra promoter and restrict bra expression in two aspects: a) SIP1 prevents its expansion to the other germ layers (ectoderm and endoderm) [22] [82][83]; b) Gsc prevents its expansion to the PM, restricting bra to the chordamesoderm within the population of DML cells segregating during gastrulation [7] [62],[84][86].

However, other factors might contribute to confine bra, since proteins that recognise the SIP1 binding sites seem to be functionally relevant only at the beginning of gastrulation, and bra was down-regulated in animal cap cells treated with high Activin concentrations even when gsc was inhibited [22] [87]. Our previous findings after depleting hairy2a suggested that this gene is necessary to restrict bra expression within DML precursors [31]. In this work, we extended the analysis to the whole MZ by depleting each hairy2 homeolog, concluding that hairy2 participates in both aspects of bra confinement. Within the MZ, hairy2 restricts bra to the prospective mesoderm layer in the IMZ, by preventing its expansion toward the prospective ectoderm in the NIMZ.

In the MZ, hairy2a seems to be more relevant than hairy2b, as the comparison of the morphants and the simultaneous knock-down suggest (Fig. 2). Previously reported knock-down analysis indicates that both homeologs participate in the development of the DML, which expresses similar levels of hairy2a and hairy2b, but the NCC preferentially require hairy2b, which is more abundantly expressed there [30]. Although their relative expression levels in the MZ were not assessed, gain-of-function indicates that both homeologs are able to repress bra and to affect NCC in the same way [31][32]. Therefore, the local relative abundance of the transcripts might account for the differential MOs effects, perhaps indicating a divergence between homeologs in their spatial regulation rather than in the functional ability of their encoded proteins.

It is interesting to notice that while hairy2a appears to be more relevant than hairy2b to restrict bra expression in the MZ, on the other side of this regulatory network, bra prevents the derepression of both hairy2 homeologs in the VLMZ. This is supported by the observation that the single knock-down of each homeolog is insufficient, but the simultaneous knock-down of hairy2a and hairy2b is needed to significantly rescue the ectopic expression of DML genes and the perturbation of the gastrulation movements produced by braEnR mRNA injections (Fig. 12).

During DML development, the appearance of the bra expression gap correlates with the notable increase of hairy2 in the organiser region. Since dorsal depletion of hairy2 obliterated that gap, the transient down-regulation of bra might be related with the role of hairy2 in inhibiting the chordamesoderm program in a subset of DML precursors, when the dorsal cells begin to involute, allowing the specification of FP and PM [31] [33]. Indeed, the complementary patterns of both genes persist in the DML structures, since bra is only expressed in the chordamesoderm, while hairy2 transcripts are distributed in the PM and the FP but not in the notochord [5] [31],[36].

HES proteins directly bind to N- or E-box DNA sequences of target promoters [88]. Noteworthy, a fragment of the Xenopus bra promoter, sufficient to drive reporter gene expression to the MZ, contains two E-box motifs which are conserved in mouse [22]. It will be interesting to elucidate whether hairy2 can directly repress bra through these sites.

Bra is Involved in a Zygotic Checkpoint that Prevents the Induction of an Incomplete Secondary Axis

The secondary axes that we obtained as a consequence of localised bra interference or hairy2a overexpression were headless and lacked notochord. They had trunk characteristics and were remarkably similar to those obtained with hairy2b overexpression [33].

We presume that the absence of notochord in the secondary axes might be explained because the ectopic dorsal cells adopt FP or PM fates. At stage 11, the patches of ectopic chd obtained after injections with braEnR or hairy2a mRNA in the CD4 blastomere (in the same conditions than those resulting in secondary axes) were similar (80%, n = 10 for braEnR; 85%, n = 13 for hairy2a; not shown) to those obtained after injections in one ventral cell at the 4-cell stage. Notably, when analysed at stage 28, chd+ cells were observed in the secondary axes in a lower proportion of the injected embryos and were detected in lower levels in comparison to earlier stages (43%, n = 7 for braEnR; 38%, n = 8 for hairy2a; Fig. 8J,K), but never formed a notochord, while the majority of the secondary axes were positive for other markers normally expressed in the trunk region, like shh, sox2, or myoD (Fig. 8). Indeed, in embryos injected with hairy2a or with braEnR mRNAs into one ventral cell at the 4-cell stage, ectopic chd+ cells were unable to involute, since they remained in the border of the still opened blastopore in neurulae (this work and [31]).

It was previously shown that, among the precursors of the embryonic DML in the Spemann’s organiser, hairy2 impedes the involution of notochordal cells by promoting FP in Xenopus [31] and also favours PM fates at the expense of the notochord [33]. Indeed, injections of either mRNA into one ventral cell at the 4-cell stage (Fig. 10B,C,G) or into the CD4 blastomere (85%, n = 13 for braEnR; 64%, n = 14 for hairy2a; not shown) resulted in ectopic expression of the PM marker gsc at stage 11. Although at this stage we probably still observe a co-ectopic expression of gsc and chd in the VLMZ corresponding to DML precursors that have not yet completely segregated, they might later be unable to adopt a notochordal fate but preferentially choose the other alternative fates of the DML populations, i.e., the PM or the FP, accounting for the lower levels of chd+ cells and the lower proportion of secondary axis with chd+ cells in the injected embryos. This hypothesis is supported by the following evidence: a) the secondary axes that we obtained have a widened FP (Fig. 9). This is supported by the fact that zebrafish mutants for no tail (ntl), the homolog of bra, lack notochord but they contain a wider FP, indicating that bra antagonises the development of the FP in favour of the notochord [73] [89][90]; b) chd is normally expressed in the notochord and in the PM [63], but the notochord is absent from the secondary axes in the hairy2a- or braEnR-injected embryos. Therefore, the scattered chd+ cells that they contain might have migrated haphazardly, according to the active migration behaviour of the PM, which was shown to be induced by gsc [91]. The fact that embryos injected with either mRNA at the 4-cell or the 16-cell stage activate ectopic gsc expression indicates that they were able to specify cells with PM characteristics.

In addition, in embryos injected with hairy2a or with braEnR mRNA into one ventral cell at the 4-cell stage, we observed high levels of apoptosis surrounding the unclosed blastopore, where the ectopic chd+ cells remained arrested. Although embryos injected into the CD4 blastomere showed significantly lower levels of apoptosis in the posterior region, a considerable proportion of them revealed moderate levels of TUNEL staining, when sibling controls presented low or none apoptotic cells in the same region (Fig. 11C,F,I). Therefore, cell death might in part account for the absence or the lower expression of chd in the secondary axes, indicating that some chd+ cells occasionally migrate anteriorly and survive in these conditions.

Among the genes directly activated by maternal β-catenin, only siamois and twin are able to induce secondary axes containing trunk and head. Both are the earliest genes known to be zygotically activated in the BCNE center, and in turn, they induce other BCNE or organiser factors like chd, noggin, foxA4a and gsc [61] [69],[92][95]. On the other hand, genes that are not direct targets of β-catenin but are later activated in the BCNE or in the organiser, such as chd and gsc, induce incomplete secondary axis similar to the ones we obtained [63] [96]. These genes were ectopically induced by braEnR or hairy2a mRNAs at gastrula stage, when the expression of chd is normally restricted to the organiser and becomes independent of β-catenin signalling [65]. However, at stage 9, ectopic chd was very faintly induced, and neither injection consistently induced ectopic Xnr3, which was proposed as a direct target of β-catenin [69] [97][98]. Moreover, while ventral injections of β-catenin failed to induce hairy2 expression [33], braEnR strongly activated hairy2 and besides, promoted chd expression in embryos depleted of β-catenin activity by UV treatments.

All this evidence suggests that bra is involved in a later zygotic checkpoint, down-stream of the earliest transcription event triggered by maternal β-catenin. Thus, in early embryogenesis, there are two critical times in which dorsal development must be repressed in the ventral side: one, under maternal control, which prevents the accumulation of nuclear β-catenin [41] [43][45]; the other one, under zygotic control mechanisms [46], among which we now propose an additional one involving bra, through the repression of hairy2. It is likely that the mechanism by which bra prevents ectopic axis development extends beyond gastrulation, since we have occasionally observed secondary tail-like structures. In Xenopus, the tail develops from secondary neurulation as a continuation of events triggered during gastrulation in the late blastopore, and bra expression persists in the tailbud [99][100].

The complete development of the head requires the early expression of BMP inhibitors in the BCNE center, which predispose the prospective neuroectoderm to form the brain, but it also requires vertical signals from the anterior endoderm during gastrulation, among which Cer is found. Notably, Cer induces secondary heads which do not contain any trunk-tail axial structures such as somites or notochord [61] [65]. We found out that braEnR- or hairy2-injected embryos ectopically express cer, and still they fail to develop head structures in the secondary axes. Since these injections were unable to consistently induce ectopic BMP inhibitors at stage 9, it is tempting to speculate that the ventral ectoderm in our experiments was incompetent to respond to Cer signalling.

Similarly to braEnR injections, hairy2 overexpression induced programmed cell death at neurula stage in the posterior region. Apoptosis is an evolutionary conserved physiological mechanism fundamental to developmental processes, since it sculpts or deletes structures, controls cell number, and eliminates abnormal, misplaced, or nonfunctional cells [101]. Since hairy2 promotes cell survival in NCC progenitors, where it is normally expressed [102][103], the apoptotic effect that we observed might be triggered as a protective mechanism to remove wrongly specified cells. Thus, through the inhibition of hairy2, bra prevents the activation of dorsal genes in ectopic regions, guaranteeing the survival of cells with ventral-posterior fates. However, if dorsal mis-specified cells happen to escape cell death, they are able to organise an incomplete secondary axis.

Fetus in fetus and parasitic conjoined twins are human pathologies that consist in a defective fetus attached externally, or located within a relatively normal fetus, respectively. The worldwide incidence of the former is about 1∶500,000 births, while the latter occurs with an incidence of 1∶100,000 births. DNA and chromosomal analysis indicates that the host and the defective fetus are genetically identical. The classical fusion theory, based on the pattern of the anomalies but not on embryological models of their pathogenesis, proposes that these pathologies result from the secondary union of two monozygotic embryonic discs [104][108]. In spite that the pathogenic mechanisms remain theoretical and are unclear, it is remarkable that the brain is always vestigial or missing in the defective twin fetus, which joins the normal twin at either ventral or dorsal locations [109]. This strikingly resembles the secondary axis without anterior brain that we obtained, which joined the primary axis either dorsally (Fig. 8A,B) or ventrally (Fig. 8C,E). Interestingly, a truncated BMP-4 receptor induces ectopic expression of dorsal genes and secondary axis like those obtained with braEnR or hairy2 [110][111], indicating that BMP signalling normally maintains ectopic trunk development suppressed in ventral regions. Thus, although we cannot exclude the fusion theory, failure in control mechanisms resulting in ectopic derepression of dorsal specific genes with the capability to induce a truncated secondary axis might as well contribute to these human pathologies.

Acknowledgments

We thank the following colleagues for providing us with DNA constructs: Abraham Fainsod (bra); Ariel Ruiz i Altaba (foxA4a); Bruce Blumberg (cer); Christof Niehrs (myoD); Dave Turner (hairy2; hairy2a, MT-hairy2a); David Kimelman (not I); David Wilkinson (krox20); Eddy De Robertis (chd); Eric Bellefroid (nrp1); Frank Conlon (braEnR); Ira Blitz (otx2); Ken Cho (gsc); Richard Harland (Xnr3, nuc-lacZ); Steve Ekker (shh); Thomas Hollemann (en2), and Yoshiki Sasai (sox2). We also thank members of our laboratory (Helena Acosta, Ignacio Bocles, Laura Boyadjián López, Victoria Gnazzo, Ernesto González Stivala, Alejandra Paganelli, and Diego Revinski) for helping with embryos and reagents preparation and participating in discussions.

Author Contributions

Conceived and designed the experiments: CEA AEC SLL. Performed the experiments: CEA SM SLL. Analyzed the data: CEA AEC SLL. Contributed reagents/materials/analysis tools: CEA SM SLL. Wrote the paper: CEA SLL.

References

  1. 1. Dobrovolskaïa-Zavadskaïa N (1927) Sur la mortification spontanée de la queue chez la souris nouveau-née et sur l’existence d’un caractère heriditaire “nonviable”. C R Soc Biol 97: 114–116.
  2. 2. Chesley P (1935) Development of the short-tailed mutant in the house mouse. J Exp Zool 70: 429–459.
  3. 3. Gluecksohn-Schoenheimer S (1938) The Development of Two Tailless Mutants in the House Mouse. Genetics 23: 573–584.
  4. 4. Herrmann BG, Labeit S, Poustka A, King TR, Lehrach H (1990) Cloning of the T gene required in mesoderm formation in the mouse. Nature 343: 617–622.
  5. 5. Smith JC, Price BM, Green JB, Weigel D, Herrmann BG (1991) Expression of a Xenopus homolog of Brachyury (T) is an immediate-early response to mesoderm induction. Cell 67: 79–87.
  6. 6. Holland PW, Koschorz B, Holland LZ, Herrmann BG (1995) Conservation of Brachyury (T) genes in amphioxus and vertebrates: developmental and evolutionary implications. Development 121: 4283–4291.
  7. 7. Latinkić B V, Umbhauer M, Neal KA, Lerchner W, Smith JC, et al. (1997) The Xenopus Brachyury promoter is activated by FGF and low concentrations of activin and suppressed by high concentrations of activin and by paired-type homeodomain proteins. Genes Dev 11: 3265–3276.
  8. 8. Hayata T, Eisaki A, Kuroda H, Asashima M (1999) Expression of Brachyury-like T-box transcription factor, Xbra3 in Xenopus embryo. Dev Genes Evol 209: 560–563.
  9. 9. Martin BL, Kimelman D (2008) Regulation of canonical Wnt signaling by Brachyury is essential for posterior mesoderm formation. Dev Cell 15: 121–133.
  10. 10. Gilchrist MJ (2012) From expression cloning to gene modeling: The development of Xenopus gene sequence resources. Genesis 50: 143–154.
  11. 11. Kispert A, Herrmann BG (1993) The Brachyury gene encodes a novel DNA binding protein. EMBO J 12: 3211–3220.
  12. 12. Kispert A, Koschorz B, Herrmann BG (1995) The T protein encoded by Brachyury is a tissue-specific transcription factor. EMBO J 14: 4763–4772.
  13. 13. Conlon FL, Sedgwick SG, Weston KM, Smith JC (1996) Inhibition of Xbra transcription activation causes defects in mesodermal patterning and reveals autoregulation of Xbra in dorsal mesoderm. Development 122: 2427–2435.
  14. 14. Smith JC, Armes NA, Conlon FL, Tada M, Umbhauer M, et al. (1997) Upstream and downstream from Brachyury, a gene required for vertebrate mesoderm formation. Cold Spring Harb Symp Quant Biol 62: 337–346.
  15. 15. Smith J (1999) T-box genes: what they do and how they do it. Trends Genet 15: 154–158.
  16. 16. Green JB, New H V, Smith JC (1992) Responses of embryonic Xenopus cells to activin and FGF are separated by multiple dose thresholds and correspond to distinct axes of the mesoderm. Cell 71: 731–739.
  17. 17. Green JB, Smith JC, Gerhart JC (1994) Slow emergence of a multithreshold response to activin requires cell-contact-dependent sharpening but not prepattern. Development 120: 2271–2278.
  18. 18. Gurdon JB, Harger P, Mitchell A, Lemaire P (1994) Activin signalling and response to a morphogen gradient. Nature 371: 487–492.
  19. 19. Gurdon JB, Mitchell A, Mahony D (1995) Direct and continuous assessment by cells of their position in a morphogen gradient. Nature 376: 520–521.
  20. 20. Wilkinson DG, Bhatt S, Herrmann BG (1990) Expression pattern of the mouse T gene and its role in mesoderm formation. Nature 343: 657–659.
  21. 21. Panitz F, Krain B, Hollemann T, Nordheim A, Pieler T (1998) The Spemann organizer-expressed zinc finger gene Xegr-1 responds to the MAP kinase/Ets-SRF signal transduction pathway. EMBO J 17: 4414–4425.
  22. 22. Lerchner W, Latinkic B V, Remacle JE, Huylebroeck D, Smith JC (2000) Region-specific activation of the Xenopus brachyury promoter involves active repression in ectoderm and endoderm: a study using transgenic frog embryos. Development 127: 2729–2739.
  23. 23. Cunliffe V, Smith JC (1992) Ectopic mesoderm formation in Xenopus embryos caused by widespread expression of a Brachyury homologue. Nature 358: 427–430.
  24. 24. Cunliffe V, Smith JC (1994) Specification of mesodermal pattern in Xenopus laevis by interactions between Brachyury, noggin and Xwnt-8. EMBO J 13: 349–359.
  25. 25. O’Reilly MA, Smith JC, Cunliffe V (1995) Patterning of the mesoderm in Xenopus: dose-dependent and synergistic effects of Brachyury and Pintallavis. Development 121: 1351–1359.
  26. 26. Tada M, O’Reilly MA, Smith JC (1997) Analysis of competence and of Brachyury autoinduction by use of hormone-inducible Xbra. Development 124: 2225–2234.
  27. 27. Conlon FL, Smith JC (1999) Interference with brachyury function inhibits convergent extension, causes apoptosis, and reveals separate requirements in the FGF and activin signalling pathways. Dev Biol 213: 85–100.
  28. 28. Kwan KM, Kirschner MW (2003) Xbra functions as a switch between cell migration and convergent extension in the Xenopus gastrula. Development 130: 1961–1972.
  29. 29. Davis RL, Turner DL (2001) Vertebrate hairy and Enhancer of split related proteins: transcriptional repressors regulating cellular differentiation and embryonic patterning. Oncogene 20: 8342–8357.
  30. 30. Murato Y, Nagatomo K, Yamaguti M, Hashimoto C (2007) Two alloalleles of Xenopus laevis hairy2 gene–evolution of duplicated gene function from a developmental perspective. Dev Genes Evol 217: 665–673.
  31. 31. López SL, Rosato-Siri M V, Franco PG, Paganelli AR, Carrasco AE (2005) The Notch-target gene hairy2a impedes the involution of notochordal cells by promoting floor plate fates in Xenopus embryos. Development 132: 1035–1046.
  32. 32. Nichane M, De Crozé N, Ren X, Souopgui J, Monsoro-Burq AH, et al. (2008) Hairy2-Id3 interactions play an essential role in Xenopus neural crest progenitor specification. Dev Biol 322: 355–367.
  33. 33. Yamaguti M, Cho KWY, Hashimoto C (2005) Xenopus hairy2b specifies anterior prechordal mesoderm identity within Spemann’s organizer. Dev Dyn 234: 102–113.
  34. 34. Murato Y, Yamaguti M, Katamura M, Cho KWY, Hashimoto C (2006) Two modes of action by which Xenopus hairy2b establishes tissue demarcation in the Spemann-Mangold organizer. Int J Dev Biol 50: 463–471.
  35. 35. Turner DL, Weintraub H (1994) Expression of achaete-scute homolog 3 in Xenopus embryos converts ectodermal cells to a neural fate. Genes Dev 8: 1434–1447.
  36. 36. Tsuji S, Cho KWY, Hashimoto C (2003) Expression pattern of a basic helix-loop-helix transcription factor Xhairy2b during Xenopus laevis development. Dev Genes Evol 213: 407–411.
  37. 37. Davis RL, Turner DL, Evans LM, Kirschner MW (2001) Molecular targets of vertebrate segmentation: two mechanisms control segmental expression of Xenopus hairy2 during somite formation. Dev Cell 1: 553–565.
  38. 38. Iso T, Kedes L, Hamamori Y (2003) HES and HERP families: multiple effectors of the Notch signaling pathway. J Cell Physiol 194: 237–255.
  39. 39. Heasman J (1997) Patterning the Xenopus blastula. Development 124: 4179–4191.
  40. 40. Moon RT, Kimelman D (1998) From cortical rotation to organizer gene expression: toward a molecular explanation of axis specification in Xenopus. Bioessays 20: 536–545.
  41. 41. Weaver C, Kimelman D (2004) Move it or lose it: axis specification in Xenopus. Development 131: 3491–3499.
  42. 42. Marikawa Y (2006) Wnt/beta-catenin signaling and body plan formation in mouse embryos. Semin Cell Dev Biol 17: 175–184.
  43. 43. Saneyoshi T, Kume S, Amasaki Y, Mikoshiba K (2002) The Wnt/calcium pathway activates NF-AT and promotes ventral cell fate in Xenopus embryos. Nature 417: 295–299.
  44. 44. Liao G, Tao Q, Kofron M, Chen J-S, Schloemer A, et al. (2006) Jun NH2-terminal kinase (JNK) prevents nuclear beta-catenin accumulation and regulates axis formation in Xenopus embryos. Proc Natl Acad Sci U S A 103: 16313–16318.
  45. 45. Hayward P, Kalmar T, Arias AM (2008) Wnt/Notch signalling and information processing during development. Development 135: 411–424.
  46. 46. Glinka A, Wu W, Onichtchouk D, Blumenstock C, Niehrs C (1997) Head induction by simultaneous repression of Bmp and Wnt signalling in Xenopus. Nature 389: 517–519.
  47. 47. Nieuwkoop PD, Faber J (1994) Normal Table of Xenopus laevis (Daudin). New York and London: Garland Publishing.
  48. 48. Harland RM (1991) In situ hybridization: an improved whole-mount method for Xenopus embryos. Methods Cell Biol 36: 685–695.
  49. 49. Franco PG, Paganelli AR, López SL, Carrasco AE (1999) Functional association of retinoic acid and hedgehog signaling in Xenopus primary neurogenesis. Development 126: 4257–4265.
  50. 50. Pizard A, Haramis A, Carrasco AE, Franco P, López S, et al.. (2004) Whole-mount in situ hybridization and detection of RNAs in vertebrate embryos and isolated organs. Curr Protoc Mol Biol Chapter 14: Unit 9.
  51. 51. Revinski DR, Paganelli AR, Carrasco AE, López SL (2010) Delta-Notch signaling is involved in the segregation of the three germ layers in Xenopus laevis. Dev Biol 339: 477–492.
  52. 52. Kao KR, Elinson RP (1988) The entire mesodermal mantle behaves as Spemann’s organizer in dorsoanterior enhanced Xenopus laevis embryos. Dev Biol 127: 64–77.
  53. 53. Hensey C, Gautier J (1997) A developmental timer that regulates apoptosis at the onset of gastrulation. Mech Dev 69: 183–195.
  54. 54. Zucker RM, Hunter ES, Rogers JM (1999) Apoptosis and morphology in mouse embryos by confocal laser scanning microscopy. Methods 18: 473–480.
  55. 55. Yamada A, Martindale MQ, Fukui A, Tochinai S (2010) Highly conserved functions of the Brachyury gene on morphogenetic movements: insight from the early-diverging phylum Ctenophora. Dev Biol 339: 212–222.
  56. 56. Ryan K, Garrett N, Mitchell A, Gurdon JB (1996) Eomesodermin, a Key Early Gene in Xenopus Mesoderm Differentiation. Cell 87: 989–1000.
  57. 57. Horb ME, Thomsen GH (1997) A vegetally localized T-box transcription factor in Xenopus eggs specifies mesoderm and endoderm and is essential for embryonic mesoderm formation. Development 124: 1689–1698.
  58. 58. Kumano G, Ezal C, Smith WC (2001) Boundaries and functional domains in the animal/vegetal axis of Xenopus gastrula mesoderm. Dev Biol 236: 465–477.
  59. 59. Knecht AK, Good PJ, Dawid IB, Harland RM (1995) Dorsal-ventral patterning and differentiation of noggin-induced neural tissue in the absence of mesoderm. Development 121: 1927–1935.
  60. 60. Cho KWY, Blumberg B, Steinbeisser H, De Robertis EM (1991) Molecular nature of Spemann’s organizer: the role of the Xenopus homeobox gene goosecoid. Cell 67: 1111–1120.
  61. 61. Bouwmeester T, Kim S, Sasai Y, Lu B, De Robertis EM (1996) Cerberus is a head-inducing secreted factor expressed in the anterior endoderm of Spemann’s organizer. Nature 382: 595–601.
  62. 62. Artinger M, Blitz I, Inoue K, Tran U, Cho KW (1997) Interaction of goosecoid and brachyury in Xenopus mesoderm patterning. Mech Dev 65: 187–196.
  63. 63. Sasai Y, Lu B, Steinbeisser H, Geissert D, Gont LK, et al. (1994) Xenopus chordin: a novel dorsalizing factor activated by organizer-specific homeobox genes. Cell 79: 779–790.
  64. 64. Heasman J (2006) Patterning the early Xenopus embryo. Development 133: 1205–1217.
  65. 65. Kuroda H, Wessely O, De Robertis EM (2004) Neural induction in Xenopus: requirement for ectodermal and endomesodermal signals via Chordin, Noggin, beta-Catenin, and Cerberus. PLoS Biol 2: E92.
  66. 66. Piccolo S, Agius E, Leyns L, Bhattacharyya S, Grunz H, et al. (1999) The head inducer Cerberus is a multifunctional antagonist of Nodal, BMP and Wnt signals. Nature 397: 707–710.
  67. 67. Wessely O, Agius E, Oelgeschläger M, Pera EM, De Robertis EM (2001) Neural induction in the absence of mesoderm: beta-catenin-dependent expression of secreted BMP antagonists at the blastula stage in Xenopus. Dev Biol 234: 161–173.
  68. 68. De Robertis EM, Kuroda H (2004) Dorsal-ventral patterning and neural induction in Xenopus embryos. Annu Rev Cell Dev Biol 20: 285–308.
  69. 69. Ishibashi H, Matsumura N, Hanafusa H, Matsumoto K, De Robertis EM, et al. (2008) Expression of Siamois and Twin in the blastula Chordin/Noggin signaling center is required for brain formation in Xenopus laevis embryos. Mech Dev 125: 58–66.
  70. 70. Gavrieli Y, Sherman Y, Ben-Sasson SA (1992) Identification of programmed cell death in situ via specific labeling of nuclear DNA fragmentation. J Cell Biol 119: 493–501.
  71. 71. Hensey C, Gautier J (1998) Programmed cell death during Xenopus development: a spatio-temporal analysis. Dev Biol 203: 36–48.
  72. 72. Barbosky L, Lawrence DK, Karunamuni G, Wikenheiser JC, Doughman Y-Q, et al. (2006) Apoptosis in the developing mouse heart. Dev Dyn 235: 2592–2602.
  73. 73. Amacher SL, Draper BW, Summers BR, Kimmel CB (2002) The zebrafish T-box genes no tail and spadetail are required for development of trunk and tail mesoderm and medial floor plate. Development 129: 3311–3323.
  74. 74. Cha YR, Takahashi S, Wright CVE (2006) Cooperative non-cell and cell autonomous regulation of Nodal gene expression and signaling by Lefty/Antivin and Brachyury in Xenopus. Dev Biol 290: 246–264.
  75. 75. Morley RH, Lachani K, Keefe D, Gilchrist MJ, Flicek P, et al. (2009) A gene regulatory network directed by zebrafish No tail accounts for its roles in mesoderm formation. Proc Natl Acad Sci U S A 106: 3829–3834.
  76. 76. Garnett AT, Han TM, Gilchrist MJ, Smith JC, Eisen MB, et al. (2009) Identification of direct T-box target genes in the developing zebrafish mesoderm. Development 136: 749–760.
  77. 77. Evans AL, Faial T, Gilchrist MJ, Down T, Vallier L, et al. (2012) Genomic targets of brachyury (T) in differentiating mouse embryonic stem cells. PloS One 7: e33346.
  78. 78. Cui Y (2005) Hairy is a cell context signal controlling Notch activity. Develop Growth Differ 47: 609–625.
  79. 79. Wilson V, Rashbass P, Beddington RS (1993) Chimeric analysis of T (Brachyury) gene function. Development 117: 1321–1331.
  80. 80. Wilson V, Manson L, Skarnes WC, Beddington RS (1995) The T gene is necessary for normal mesodermal morphogenetic cell movements during gastrulation. Development 121: 877–886.
  81. 81. Smith JC (2001) Making mesoderm–upstream and downstream of Xbra. Int J Dev Biol 45: 219–224.
  82. 82. Verschueren K, Remacle JE, Collart C, Kraft H, Baker BS, et al. (1999) SIP1, a novel zinc finger/homeodomain repressor, interacts with Smad proteins and binds to 5′-CACCT sequences in candidate target genes. J Biol Chem 274: 20489–20498.
  83. 83. Papin C, Van Grunsven LA, Verschueren K, Huylebroeck D, Smith JC (2002) Dynamic regulation of Brachyury expression in the amphibian embryo by XSIP1. Mech Dev 111: 37–46.
  84. 84. Steinbeisser H, De Robertis EM (1993) Xenopus goosecoid: a gene expressed in the prechordal plate that has dorsalizing activity. C R Acad Sci 316: 959–971.
  85. 85. Vodicka MA, Gerhart JC (1995) Blastomere derivation and domains of gene expression in the Spemann Organizer of Xenopus laevis. Development 121: 3505–3518.
  86. 86. Latinkić B V, Smith JC (1999) Goosecoid and mix.1 repress Brachyury expression and are required for head formation in Xenopus. Development 126: 1769–1779.
  87. 87. Papin C, Smith JC (2000) Gradual refinement of activin-induced thresholds requires protein synthesis. Dev Biol 217: 166–172.
  88. 88. Fischer A, Gessler M (2007) Delta-Notch–and then? Protein interactions and proposed modes of repression by Hes and Hey bHLH factors. Nucleic Acids Res 35: 4583–4596.
  89. 89. Halpern ME, Ho RK, Walker C, Kimmel CB (1993) Induction of muscle pioneers and floor plate is distinguished by the zebrafish no tail mutation. Cell 75: 99–111.
  90. 90. Halpern ME, Hatta K, Amacher SL, Talbot WS, Yan YL, et al. (1997) Genetic interactions in zebrafish midline development. Dev Biol 187: 154–170.
  91. 91. Luu O, Nagel M, Wacker S, Lemaire P, Winklbauer R (2008) Control of gastrula cell motility by the Goosecoid/Mix.1/Siamois network: basic patterns and paradoxical effects. Dev Dyn 237: 1307–1320.
  92. 92. Lemaire P, Garrett N, Gurdon JB (1995) Expression cloning of Siamois, a Xenopus homeobox gene expressed in dorsal-vegetal cells of blastulae and able to induce a complete secondary axis. Cell 81: 85–94.
  93. 93. Laurent MN, Blitz IL, Hashimoto C, Rothbächer U, Cho KW (1997) The Xenopus homeobox gene twin mediates Wnt induction of goosecoid in establishment of Spemann’s organizer. Development 124: 4905–4916.
  94. 94. Carnac G, Kodjabachian L, Gurdon JB, Lemaire P (1996) The homeobox gene Siamois is a target of the Wnt dorsalisation pathway and triggers organiser activity in the absence of mesoderm. Development 122: 3055–3065.
  95. 95. Fagotto F, Guger K, Gumbiner BM (1997) Induction of the primary dorsalizing center in Xenopus by the Wnt/GSK/beta-catenin signaling pathway, but not by Vg1, Activin or Noggin. Development 124: 453–460.
  96. 96. Steinbeisser H, De Robertis EM, Ku M, Kessler DS, Melton DA (1993) Xenopus axis formation: induction of goosecoid by injected Xwnt-8 and activin mRNAs. Development 118: 499–507.
  97. 97. McKendry R, Hsu SC, Harland RM, Grosschedl R (1997) LEF-1/TCF proteins mediate wnt-inducible transcription from the Xenopus nodal-related 3 promoter. Dev Biol 192: 420–431.
  98. 98. Nelson RW, Gumbiner BM (1999) A cell-free assay system for beta-catenin signaling that recapitulates direct inductive events in the early xenopus laevis embryo. J Cell Biol 147: 367–374.
  99. 99. Gont LK, Steinbeisser H, Blumberg B, De Robertis EM (1993) Tail formation as a continuation of gastrulation: the multiple cell populations of the Xenopus tailbud derive from the late blastopore lip. Development 119: 991–1004.
  100. 100. Beck CW, Slack JM (1999) A developmental pathway controlling outgrowth of the Xenopus tail bud. Development 126: 1611–1620.
  101. 101. Fuchs Y, Steller H (2011) Programmed Cell Death in Animal Development and Disease. Cell 147: 742–758.
  102. 102. Nagatomo K, Hashimoto C (2007) Xenopus hairy2 functions in neural crest formation by maintaining cells in a mitotic and undifferentiated state. Dev Dyn 236: 1475–1483.
  103. 103. Nichane M, Ren X, Souopgui J, Bellefroid EJ (2008) Hairy2 functions through both DNA-binding and non DNA-binding mechanisms at the neural plate border in Xenopus. Dev Biol 322: 368–380.
  104. 104. Spencer R (2000) Theoretical and analytical embryology of conjoined twins: part I: embryogenesis. Clin Anat 13: 36–53.
  105. 105. Kaufman MH (2004) The embryology of conjoined twins. Childs Nerv Syst 20: 508–525.
  106. 106. Brand A, Alves MC, Saraiva C, Loío P, Goulão J, et al. (2004) Fetus in fetu–diagnostic criteria and differential diagnosis–a case report and literature review. J Pediatr Surg 39: 616–618.
  107. 107. Escobar MA, Rossman JE, Caty MG (2008) Fetus-in-fetu: report of a case and a review of the literature. J Pediatr Surg 43: 943–946.
  108. 108. Gupta SK, Singhal P, Arya N (2010) Fetus-in-fetu: A Rare Congenital Anomaly. J Surg Tech Case Rep 2: 77–80.
  109. 109. Spencer R (2001) Parasitic conjoined twins: external, internal (fetuses in fetu and teratomas), and detached (acardiacs). Clin Anat 14: 428–444.
  110. 110. Suzuki A, Thies RS, Yamaji N, Song JJ, Wozney JM, et al. (1994) A truncated bone morphogenetic protein receptor affects dorsal-ventral patterning in the early Xenopus embryo. Proc Natl Acad Sci U S A 91: 10255–10259.
  111. 111. Hoppler S, Moon RT (1998) BMP-2/−4 and Wnt-8 cooperatively pattern the Xenopus mesoderm. Mech Dev 71: 119–129.
  112. 112. Ruiz i Altaba A, Jessell TM (1992) Pintallavis, a gene expressed in the organizer and midline cells of frog embryos: involvement in the development of the neural axis. Development 116: 81–93.
  113. 113. Von Dassow G, Schmidt JE, Kimelman D (1993) Induction of the Xenopus organizer: expression and regulation of Xnot, a novel FGF and activin-regulated homeo box gene. Genes Dev 7: 355–366.
  114. 114. Ekker SC, McGrew LL, Lai CJ, Lee JJ, Von Kessler DP, et al. (1995) Distinct expression and shared activities of members of the hedgehog gene family of Xenopus laevis. Development 121: 2337–2347.
  115. 115. Mizuseki K, Kishi M, Matsui M, Nakanishi S, Sasai Y (1998) Xenopus Zic-related-1 and Sox-2, two factors induced by chordin, have distinct activities in the initiation of neural induction. Development 125: 579–587.
  116. 116. Pannese M, Polo C, Andreazzoli M, Vignali R, Kablar B, et al. (1995) The Xenopus homologue of Otx2 is a maternal homeobox gene that demarcates and specifies anterior body regions. Development 121: 707–720.
  117. 117. Frank D, Harland RM (1991) Transient expression of XMyoD in non-somitic mesoderm of Xenopus gastrulae. Development 113: 1387–1393.
  118. 118. Hopwood ND, Pluck A, Gurdon JB (1991) Xenopus Myf-5 marks early muscle cells and can activate muscle genes ectopically in early embryos. Development 111: 551–560.
  119. 119. Hemmati-Brivanlou A, De la TorreJR, Holt C, Harland RM (1991) Cephalic expression and molecular characterization of Xenopus En-2. Development 111: 715–724.
  120. 120. Bradley LC, Snape A, Bhatt S, Wilkinson DG (1993) The structure and expression of the Xenopus Krox-20 gene: conserved and divergent patterns of expression in rhombomeres and neural crest. Mech Dev 40: 73–84.