Characterization of TALE genes expression during the first lineage segregation in mammalian embryos

Authors

  • Wendy Sonnet,

    1. Embryologie Moléculaire et Cellulaire Animale, Institut des Sciences de la Vie, Université catholique de Louvain, Louvain-la-Neuve, Belgium
    Search for more papers by this author
  • Rene Rezsöhazy,

    1. Embryologie Moléculaire et Cellulaire Animale, Institut des Sciences de la Vie, Université catholique de Louvain, Louvain-la-Neuve, Belgium
    Search for more papers by this author
    • Rene Rezsöhazy and Isabelle Donnay contributed equally to this study.

  • Isabelle Donnay

    Corresponding author
    1. Embryologie Moléculaire et Cellulaire Animale, Institut des Sciences de la Vie, Université catholique de Louvain, Louvain-la-Neuve, Belgium
    • Université Catholique de Louvain, Institut des Sciences de la Vie, Embryologie Moléculaire et Cellulaire Animale, Place Croix du Sud 4-5 Bte L7.07.10, B-1348 Louvain-la-Neuve, Belgium
    Search for more papers by this author
    • Rene Rezsöhazy and Isabelle Donnay contributed equally to this study.


Abstract

Background: Three amino acid loop extension (TALE) homeodomain-containing transcription factors are generally recognized for their role in organogenesis and differentiation during embryogenesis. However, very little is known about the expression and function of Meis, Pbx, and Prep genes during early development. Results: In order to determine whether TALE proteins could contribute to the early cell fate decisions in mammalian development, this study aimed to characterize in a systematic manner the pattern of expression of all Meis, Pbx, and Prep genes from the precompaction to blastocyst stage corresponding to the first step of cell differentiation in mammals. To reveal to what extent TALE genes expression at these early stages is a conserved feature among mammals, this study was performed in parallel in the bovine and mouse models. We demonstrated the transcription and translation of TALE genes, before gastrulation in the two species. At least one member of Meis, Pbx, and Prep subfamilies was found expressed at the RNA and protein levels but different patterns of expression were observed between genes and between species, suggesting specific gene regulations. Conclusions: Taken together, these results suggest a previously unexpected involvement of these factors during the early development in mammals. Developmental Dynamics 241:1827–1835, 2012. © 2012 Wiley Periodicals, Inc.

INTRODUCTION

Meis, Pbx, and Prep proteins belong to the evolutionary conserved Three Amino acid Loop Extension (TALE) class of homeodomain-containing transcription factors, which are known to play crucial roles in several developmental processes. TALE proteins, characterized by an atypical homeodomain that contains a three amino acid loop extension between helix 1 and helix 2, include the PBC (Pbx1-4) and the MEINOX (Meis1-3 and Prep1-2) subfamilies in mammals (Burglin, 1997).

As transcription factors, TALE proteins establish a well-known partnership with Hox proteins. By this interaction, they increase the DNA-binding affinity and specificity of Hox proteins to enhance or repress transcription of target genes. Numerous studies revealed that TALE proteins are required for many Hox-dependent functions, such as the patterning of the body axis in the developing embryo or in oncogenesis (for a review see Moens and Selleri, 2006). TALE proteins act also at the post-transcriptional level by interacting with translation initiation factors. For example, in the mouse oocyte, Prep1 interacts with 4EHP (eukaryotic translation initiation factor 4E homolog protein) in the cytoplasm to regulate translation of Hox mRNA (Villaescusa et al., 2009).

The study of mice bearing deletions for Meis and Pbx genes demonstrated that they are essential, after gastrulation, in organogenesis, differentiation, and hematopoiesis (Selleri et al., 2001, Hisa et al., 2004, Di Giacomo et al., 2006). In that context, several members of the PBC subfamily show overlapping expression patterns and a considerable functional redundancy. Indeed, Pbx1/Pbx2-deficient mice present a more severe limb and skeletal phenotype than Pbx1-null mice, while mice lacking Pbx2 are viable (Selleri et al., 2001, 2004; Capellini et al., 2006, 2008). Unlike other TALE proteins, it is known that Prep1 plays an important role before gastrulation in the mouse, during early post-implantation stages, in the protection of epiblast cells from DNA damage (Fernandez-Diaz et al., 2010). To date, no loss of function for Meis2, Meis3, and Prep2 has been described.

Though the functions of Meis, Pbx, and Prep proteins have been mainly studied from the implantation on, two studies have provided evidence for TALE genes transcription from the embryonic genome during preimplantation development. In the mouse, mRNA corresponding to Meis1 have been found in blastocysts (Kageyama et al., 2007), and Prep1 transcripts have been detected from the 4-cell to the blastocyst stage (Fernandez-Diaz et al., 2010). In the bovine, MEIS1 transcripts have been also detected in morula and blastocyst stages (Khan et al., 2012). Likewise, early expression of Hox transcripts have been observed after the embryonic genome activation in different mammalian species, like cattle, human, mouse, and pig (Ponsuksili et al., 2001; Huntriss et al., 2006; Kageyama et al., 2007; Gao et al., 2010; Paul et al., 2011). However, to our knowledge, a global view on TALE genes activity at early stages of development is lacking, since their expression has never been studied in a systematic manner and the role they can play during pre-implantation development has never been investigated.

The first major event that takes place after the embryonic genome activation during mammalian development is the segregation of the two first cell lineages during the morula to blastocyst transition, with the establishment of the inner cell mass (ICM) and the trophectoderm (TE). The TE contributes to the embryonic portion of the placenta, while the ICM generates the embryo proper and other extra-embryonic structures, such as the yolk sac. The correct determination of these lineages is essential for implantation and further development of the embryo. The first cell fate decision begins with morula compaction. In the mouse, compaction occurs at the 8-cell stage and is characterized by an increase of intercellular adhesion mediated by E-cadherin and the formation of adherent junctions. Compaction causes the transformation of an embryo with clearly demarcated blastomeres into a tightly organized mass with indistinguishable cell outlines. Concomitant with compaction, intracellular polarity appears in all blastomeres and is manifested by cytoplasm reorganization and by distinct protein localization along the apical-basolateral axis. Up to the 8-cell stage, all blastomeres composed a uniform and totipotent population of cells. Subsequent cell divisions generate two distinct types of cells, inner apolar cells that give rise to the ICM, surrounded by outer polar cells that give rise to the TE. Around the 32-cell stage, the outer cells complete TE differentiation associated with the formation of the blastocoel cavity: the embryo is now a blastocyst, composed of the ICM at one side of the cavity, surrounded by the TE (for a review, see Cockburn and Rossant, 2010).

Though preimplantation development progresses through the same morphological stages in mammals, the early development differs between species in several aspects. First, the kinetics of development varies among species: mouse embryos take 4.5 days from fertilization to reach the expanded blastocyst stage, while human and bovine embryos take, respectively, 6 and 7 days. Second, the major embryonic genome activation occurs at the 2-cell stage in mice, at the 4-cell stage in human, and at the 5–8-cell stage in bovine. Likewise, in human and cattle, compaction begins at the 16- and 32-cell stages, respectively (Hyttel, 2010). Evolutionary divergence in the regulation of factors determining the two first cell lineage segregation seems to have emerged to allow early TE differentiation in mice, unlike in human and cattle (Berg et al., 2011). Finally, the most marked difference is the timing of implantation. Unlike mouse and human blastocysts, which implant soon after hatching, bovine embryo implantation starts 2 weeks after blastocyst formation, after a long period of elongation during which gastrulation is already initiated (van Eijk et al., 1999). Moreover, implantation is invasive for the mouse and human embryo and non-invasive for the bovine (Hyttel, 2010).

In order to investigate the possible involvement of TALE homeoproteins in early development, this study aimed to characterize in a systematic manner the pattern of expression of all Meis, Pbx, and Prep genes. We have focused on the expression of these transcription factors during the transition from the precompaction to the blastocyst stage, corresponding to the first steps of embryonic differentiation. Two mammalian models were studied in parallel that show important differences in their kinetics of early development and implantation: in vitro–produced bovine embryos and in vitro–cultured mouse embryos. Comparison of TALE expression patterns between bovine and mouse early embryos will allow revealing conserved features but also interspecies diversity among mammals. A first complete catalogue of TALE genes expression was performed by a RT-PCR approach on pools of embryos (precompaction, compact morula, and blastocyst) in the two species. Additionally, quantitative RT-PCR was performed to draw the kinetics of expression from the precompaction to the blastocyst stage. Finally, the presence and the localization of TALE proteins were determined by immunofluorescence.

RESULTS

TALE Homeobox Genes Are Expressed During Early Mammalian Lineage Segregation

The expression of Pbx and Prep genes during early lineage segregation was evaluated by a systematic screening performed on pools of precompaction embryos, compact morulae, and blastocysts by RT-PCR using primers targeting conserved gene regions followed by cDNA cloning and sequencing. Seventy independent clones were analyzed and allowed to identify transcription of genes from each Pbx and Prep family (Table 1). Pbx3 and Prep1 gene expression were observed in both mammalian models, whereas PBX1 transcripts were only detected in the bovine and Pbx2 transcripts only in the mouse. Expression of Meis genes was also evaluated by RT-PCR on pools of precompaction embryos, compact morulae, and blastocysts, but with gene-specific primers. In mouse and bovine early embryos, Meis1 and Meis3 genes were expressed, whereas Meis2 has not been detected to be expressed.

Table 1. Repertoire of TALE Genes Expressed During Early Mammalian Lineage Segregationa
 Pbx1Pbx2Pbx3Pbx4Meis1Meis2Meis3Prep1Prep2
  • a

    Number of sequenced clones corresponding to Pbx and Prep transcripts using primers targeting conserved domains. (+) and (−) indicate, respectively, presence or absence of Pbx and Meis transcripts using specific primers.

Bovine5+180++260
Mouse+29+++60

In order to evaluate the sensitivity of the RT-PCR screening applied on pools of embryos at three stages of development (precompaction, compact morula, and blastocyst), we performed RT-PCR with specific primers at each of the three stages for those genes that were not detected to be expressed: Pbx1 in the mouse, PBX2 in the bovine, as well as Meis2, Pbx4, and Prep2 in both models. We confirmed the absence of detection of Prep2 and Meis2 transcripts in both models and of PBX4 transcripts in bovine. However, contrary to what came out of our first screening, Pbx1 (at the blastocyst stage only) and Pbx4 (at the three stages) in the mouse and PBX2 (at the blastocyst stage only) in the bovine were found to be expressed during the early lineage segregation (Fig. 1). The initial RT-PCR screening used was, therefore, not exhaustive. However among 10 TALE genes only one gene expressed in the bovine and two in the mouse had been missed.

Figure 1.

Expression of TALE genes by RT-PCR at precompaction, compact morula, and blastocyst stages in the bovine and the mouse. A: RT-PCR analysis of MEIS2, PBX2, PBX4, and PREP2 expression in bovine embryos compared to glucose-6-phosphate dehydrogenase gene (G6PD). B: RT-PCR analysis of Meis2, Pbx1, Pbx4, and Prep2 expression in mouse embryos compared to RNA polymerase II gene (RNApolII). PC, precompaction; CM, compact morula; B, blastocyst; NT, no template.

A summary of all genes found expressed in each model is presented in Table 1. Taken together, these data demonstrated the transcription of at least one gene of each Meis, Pbx, and Prep subfamilies before gastrulation during the early lineage segregation in our two models.

Distinct Profiles of TALE mRNA Expression Are Observed During Early Mammalian Lineage Segregation

Quantitative RT-PCR were run with specific primers at three developmental stages (precompaction, compact morula, and blastocyst) in both species to refine the dynamic of expression of TALE homeobox genes during the first differentiation events and to compare their expression between mouse and bovine embryos. Quantitative results were normalized with an internal normalization factor, corresponding to the geometric mean of qPCR data obtained for two reference genes. Profiles of relative expression were established for the TALE genes found reproducibly expressed at the three stages (Fig. 2). For the other genes, expression was evaluated non-quantitatively at each stage.

Figure 2.

Dynamics of TALE mRNA expression by RT-qPCR at the precompaction, compact morula, and blastocyst stages in the bovine and the mouse. A–C: Bovine: relative quantification of PBX3 (A), PREP1 (B), and PBX1 (C) mRNA determined by RT-qPCR. Results were normalized with a NF, corresponding to the geometric mean of GAPDH and YWHAZ. D–H: Mouse: relative quantification of Meis1 (D), Prep1 (E), Pbx2 (F), Pbx3 (G), and Pbx4 (H) expression assessed by RT-qPCR. Quantitative data were standardized with a NF (mean of H2a and Hprt). For each gene, normalized expression was reported to the precompaction stage (PC=1). 3<n<7. Box plot with different letters is significantly different (Kruskal-Wallis, P < 0.05).

In the bovine, expression profiles revealed two main trends. On the one hand, PBX3 and PREP1 expression showed a decrease in relative expression at the blastocyst stage compared to the precompaction stage (Fig. 2A,B). The relative amount of PBX3 and PREP1 transcripts displayed a significant stepwise decrease from the precompaction to the blastocyst stage: respectively, 30 and 43% reduction in compact morulae, and further 9 and 12% in blastocysts compared with the precompaction stage (Fig. 2A,B). On the other hand, the pattern of expression of PBX1 (Fig. 2C) was similar except that the relative transcription level was stable between the precompaction and the compact morula stage, followed by a significant decrease at the blastocyst stage (13% of reduction compared with the morula stage). MEIS1 expression appeared to be weak as transcripts were scarcely detected with RT-qPCR. Nested PCR was necessary to detect its expression along the transition from the precompaction to the blastocyst stages (Fig. 3C). MEIS3 was reproducibly detected only at the blastocyst stage and, therefore, the relative mRNA expression profile could not be drawn. When performing PCR instead of qPCR, we found that MEIS3 was expressed from the compact morula stage (Fig. 3A). Finally, PBX2 was detected only at the blastocyst stage (Fig. 1A).

Figure 3.

RT-PCR of TALE genes at precompaction, compact morula, and blastocyst stages in the bovine and the mouse. A,B: RT-PCR analysis of Meis3 expression in bovine (A) and mouse (B) embryos compared to glucose-6-phosphate dehydrogenase gene (G6PD) and RNA polymerase II gene (RNApolII). C: Nested RT-PCR analysis of MEIS1 expression in bovine embryos compared to G6PD. PC, precompaction; CM, compact morula; B, blastocyst; NT, no template.

In the mouse, four patterns of expression were observed. The first one was an increase of relative expression from the precompaction to the blastocyst stage for Meis1, Meis3, and Prep1. Meis1 displayed a significant stepwise increase from the precompaction to the blastocyst stage (Fig. 2D). Most significantly, a 19.6-fold higher Meis1 expression was observed in blastocysts compared to the precompaction stage, which contrasts with its very weak expression in the bovine. Unlike the pattern observed in the bovine, Prep1 expression increased significantly between the precompaction and the blastocyst stage (1.8-fold compared with the precompaction stage; Fig. 2E). The second pattern was the dynamic of expression of Pbx2 (Fig. 2F). It decreases between the precompaction and the compact morula stages, before rising up at the blastocyst stage (2.8-fold compared with precompaction expression level). The third pattern was a decrease in expression for Pbx3; its relative mRNA expression decreases along the early lineage specification: 64% reduction in compact morulae, and a further 36% decrease in blastocysts compared with the precompaction stage (Fig. 2G). Fourthly, no significant change in expression from precompaction to the blastocyst stage was observed for Pbx4 (Fig. 2H). As for the bovine, Meis3 mRNA was detected reproducibly only at the blastocyst stage using RT-qPCR. RT-PCR were then performed at each of the three stages, and indicated that Meis3 is transcribed from compaction, as in the bovine (Fig. 3B). Finally, Pbx1 was reproducibly detected by RT-PCR only at the blastocyst stage (Fig. 1B).

Taken together, these data show that in both models Meis3 and Pbx3 present a quite similar pattern, whereas bovine and mouse Pbx1, Pbx2, and Prep1 show different kinetics of expression. Intriguingly, mouse Meis1 showed an important increase of expression at the blastocyst stage, contrary to its very low expression in the bovine. Likewise, Pbx4 showed a stable expression in the mouse while its expression remained undetectable in the bovine.

TALE Proteins Are Translated During Early Mammalian Lineage Segregation

The next step was to check if the TALE proteins corresponding to the expressed genes are translated. When specific antibodies were obtained that were suitable for immunofluorescence, whole-mount immunofluorescence detection was performed. This was possible for MEIS1, PBX1, PBX3, and PREP1 in the bovine and for Meis1, Pbx1, Pbx2, Pbx3, and Prep1 in the mouse. Cellular distribution of the proteins was then revealed by confocal microscopy analysis. Results are illustrated in Figure 4. The specificity of each antibody was first controlled by Western blot for each species (see Experimental Procedures section; see also Fig. 6). Immunoblots showed that TALE antibodies detected the corresponding protein at the expected molecular weight. However, a few additional bands were also stained for Meis1 antibody in bovine extracts.

Figure 4.

Distribution of TALE proteins during bovine and mouse early lineage segregation. Whole-mount confocal immunofluorescence analysis of TALE proteins distribution at precompaction (PC), compact morula (CM), and blastocyst (B) stages. A–D: Distribution in bovine embryos of PREP1 (A), PBX3 (B), PBX1 (C), and MEIS1 (D). E–I: Distribution in mouse embryos of Prep1 (E), Pbx3 (F), Pbx2 (G), Meis1 (H), Pbx1 (I). For each protein, immunofluorescence was performed in triplicate and similar results were obtained for each replicate. Top shows specific TALE protein staining, and bottom represents merge of TALE and DAPI nuclear staining. Representative confocal Z-section. N=3. Bar=30 μm.

In the bovine, PBX1, PBX3, and PREP1 proteins were detected at the three stages (Fig. 4A–C). The proteins were observed in each cell and in both cell lineages (ICM and TE) at the blastocyst stage. However, the intracellular localization differed between proteins. While PBX3 and PREP1 appeared to be predominantly located in the nucleus, PBX1 was mainly observed in the cytoplasm. In correlation with its low level of expression, we failed to detect MEIS1 protein at those stages of development (Fig. 4D). These results demonstrated that in cattle three TALE proteins are present during the first differentiation steps and show different intracellular patterns.

In the mouse, Meis1, Pbx1, Pbx2, Pbx3, and Prep1 proteins were all found to be translated before implantation (Fig. 4E–I). Two patterns were observed. Pbx2, Pbx3, and Prep1 proteins were detected at the three stages of interest. As in the bovine, they were detected in all cells at each stage, but interestingly the intracellular localization of Pbx2, Pbx3, and Prep1 proteins differed from each other. Prep1 seemed to be present at similar levels in the nucleus and the cytoplasm at the precompaction and compact morula stages, whereas its localization became more strictly nuclear at the blastocyst stage (Fig. 4E). Pbx2 shared a similar distribution as Prep1, although nuclear staining was more important than cytoplasmic staining at the precompaction and compact morula stages. At the blastocyst stage, Pbx2 was likewise more strictly located inside the nucleus (Fig. 4G). The Pbx3 protein was detected in both the cytoplasm and the nucleus at each stage, but its pattern of expression in the nucleus appeared spotty (Fig. 4F). Contrary to the other TALE proteins, Meis1 and Pbx1 proteins were detected only at the blastocyst stage, whereas no immunoreactivity was detected in precompaction embryos and compact morulae. In the blastocyst, Meis1 was observed only in the nucleus, while Pbx1 protein was present in the cytoplasm (Fig. 4H,I). For both proteins, no obvious difference was observed between ICM and TE cells. These results are in agreement with the fact that Meis1 mRNA displayed a remarkable increase of 19.6-fold in blastocysts compared to the precompaction stage, and with the fact that Pbx1 mRNAs were only detected at the blastocyst stage.

As immunofluorescence could not be set up for Meis3 (both species), Western blot analysis was instead performed, but only at the blastocyst stage considering the important number of embryos required for this technique. Meis3 protein was detected by Western blot in the two species. Moreover, we localized the expression of the gene by whole-mount in situ hybridization. Meis3 mRNA localized both in the ICM and TE in the two species (Fig. 5). Finally, we could not evaluate the presence of Pbx4 proteins in mouse embryos and of PBX2 proteins in bovine embryos as no specific signal was obtained with the tested antibodies in control tissues.

Figure 5.

Expression patterns of Meis3 in bovine and mouse blastocysts. A,B: Western blot analysis of Meis3 proteins in bovine (A) and mouse (B) blastocyst extracts compared to β-actin (41 kDa). Mouse embryos aged 11.5 days (E11.5) were used as positive control and two specific bands are visualized corresponding to two Meis3 isoforms (41 and 39 kDa). A long exposure of the autoradiography was necessary to detect Meis3 in mouse blastocyst extracts due to the low cell number of the blastocysts. C,D: Distribution of Meis3 transcripts by whole-mount in situ hybridization in bovine (C) and mouse (D) blastocysts. Negative controls with sense probe are also shown. B, blastocyst. Bar = 30 μm.

Taken together, the results obtained from immunofluorescence were consistent with the mRNA expression patterns assessed by RT-PCR. For the first time, we have demonstrated the presence of Meis, Pbx, and Prep proteins during early lineage segregation in mammals: Meis1, Meis3, Pbx1, Pbx2, Pbx3, and Prep1 in the mouse; MEIS3, PBX1, PBX3, and PREP1 in the bovine. Among the TALE proteins analyzed at the three stages, Meis1 and Pbx1 were only detected at the blastocyst stage, in the mouse.

DISCUSSION

This study aimed to characterize the expression of all TALE homeobox genes after the embryonic genome activation, throughout the first cell lineage segregation, and to compare their expression profile between the mouse and bovine embryo, two species showing important differences during pre-implantation development. Using for the first time a systematic approach, we demonstrated the transcription and translation of TALE homeobox genes, before gastrulation in the two species. All Meis, Pbx, and Prep subfamilies were found expressed at the RNA and protein levels but different patterns of expression were observed between genes and between species, suggesting specific gene regulations. Taken together, these results suggest a previously unexpected involvement of these proteins during preimplantation development in mammals.

We first investigated the expression of Pbx and Prep genes by a RT-PCR screening using primers targeting conserved gene regions. We observed that Pbx3 and Prep1 were expressed in both mammalian models, whereas PBX1 and Pbx2 expression were only detected in bovine or in mouse embryos, respectively. This approach had the advantage to check for the expression of several genes in a small amount of biological material. However, amplification or cloning efficiency may differ according to gene sequences, for example, and the sensitivity of this approach on pools of embryos from different stages might also be questioned. We have, therefore, examined in a second step the accuracy of the first strategy by performing RT-PCR with specific primers at each embryonic stage for the genes not detected as expressed in the screening. This confirmed that PBX4 expression could not be detected in the bovine, as well as Meis2 and Prep2 expression in neither species. Conversely, it resulted in the detection of Pbx1 and Pbx4 expression in the mouse and PBX2 expression in the bovine. Among 10 TALE genes, only one expressed in the bovine and two in the mouse have, therefore, been ignored by the first screening approach. This approach is thus not exhaustive but is nevertheless appropriate to provide a good indication of expressed transcripts among large families of genes in a small amount of scarcely available biological material.

Our data are in accordance with those obtained in previous studies showing the expression of Meis1 (Kageyama et al., 2007) and Prep1 (Fernandez-Diaz et al., 2010) during early development in mice, as well as the presence of MEIS1 transcripts in bovine early embryos (Khan et al., 2012). The expression of Meis, Pbx, and Prep families in both species suggests some conservation among mammals. Looking at the dynamics of TALE mRNA expression during early lineage specification, different profiles were observed between genes and species suggesting specific regulation of gene expression among TALE class members, and different functions for the corresponding proteins.

When comparing orthologous genes in the two species, only Meis3 and Pbx3 presented a similar pattern of expression. The observed interspecies discrepancies could be related to the well-known differences between mouse and bovine embryos, for example, in the timing of activation of the embryonic genome, in the kinetics of development, in the type and the timing of implantation. Specificities in the regulation of early lineage specification were previously demonstrated for the pluripotency transcription factor Oct4 (Berg et al., 2011). Indeed, OCT4 protein is not restricted to the ICM in bovine blastocysts, while it is in the mouse (Nichols et al., 1998; Kirchhof et al., 2000; Cauffman et al., 2005; Niwa et al., 2005). This is related to divergence in a regulatory element controlling Oct4 expression (Berg et al., 2011). In our study, Meis1 expression in the mouse showed an important increase at the blastocyst stage, while in the bovine its expression remains low and the protein was not detected. This coincides with the commitment of ICM and TE in mouse blastocysts to allow rapid implantation (Berg et al., 2011), while in the bovine implantation starts 2 weeks after blastocyst formation, after a long period of elongation during which gastrulation is already initiated (van Eijk et al., 1999). It will be interesting to evaluate mRNA expression profiles of MEIS1 in the bovine at later stages, to analyse whether its level also increases before implantation.

In most studies about gene expression in early embryos, particularly in bovine embryos, only the mRNA was considered, as no available and/or specific antibodies exist. However, it is well known that mRNA expression does not always reflect the presence of the corresponding protein. The presence and the localization of TALE proteins were thus assessed in preimplantation embryos. Whole-mount immunofluorescence was the technique of choice, but Western blot was also performed at the blastocyst stage when immunostaining could not be set up with the available antibodies. However, the high number of embryos necessary for immunoblotting prevented us from performing this technique at each stage. For the first time, we showed the presence of Meis, Pbx, and Prep proteins, before gastrulation, during the first differentiation steps of mammalian embryos: Meis1, Meis3, Pbx1, Pbx2, Pbx3, and Prep1 in the mouse; MEIS3, PBX1, PBX3, and PREP1 in the bovine. Overall, immunofluorescence analysis of the proteins showed a good correlation with mRNA expression in both species. In particular, Meis1 expression showed an important increase at the blastocyst stage in the mouse, which corresponds to the detection of the corresponding protein only at this stage, while in the bovine its expression remains low and the protein was not detected.

TALE proteins were found in all embryonic cells, and localized in both ICM and TE at the blastocyst stage. Moreover, no difference in the intracellular distribution of these factors has been observed between these lineages. The segregation of the ICM and TE in mice is accompanied by the differential expression of transcription factors essential for the determination of cell fate, such as Oct4 and Cdx2 for ICM and TE, respectively (Nichols et al., 1998; Niwa et al., 2005; Strumpf et al., 2005). The complementary expression patterns and reciprocal repression of these transcription factors define the determination of the two first cell lineages (Niwa et al., 2005). As TALE proteins are not segregated between the ICM and TE, it could be deduced that they are not directly involved in the differentiation process. However, several studies demonstrated that upstream factors involved in the segregation of Cdx2 expression are expressed in both ICM and TE lineages, like the transcription factor Tead4 whose partnership with Yap in the nucleus of the outer cells activates Cdx2 expression (Nishioka et al., 2008, 2009). So an involvement of TALE genes in the regulation of factors determining cell fate is not excluded, especially as Chan and collaborators (2009) demonstrated the direct involvement of PBX1 in the regulation of NANOG, a well-known factor of embryonic cell pluripotency, in human embryonic stem cells (hESCs). Indeed, they demonstrated that PBX1 directly binds and regulates NANOG promoter. Moreover, PBX1, PBX2, and PBX3 show redundant functions on the NANOG promoter (Chan et al., 2009). Such redundancy is related to their extensive sequence similarities combined to their overlapping patterns of expression as shown in the present study in the mouse. However, it should be mentioned that the intracellular distribution seems to differ between the proteins, Pbx1 being essentially nuclear, while Pbx2 and 3 show principally a cytoplasmic localization. This could reflect different functions instead of redundant ones.

While the pattern of gene expression showed important variations between the two species, Prep1 proteins are present ubiquitously during early lineage segregation, in both bovine and mouse embryos. An important role of Prep1 has been shown before gastrulation, during early post-implantation stages in the mouse, in the protection of epiblast cells from DNA damage (Fernandez-Diaz et al., 2010). Embryonic lethality is observed at this stage for Prep1-null mice and Prep1−/− embryos show a high level of p53-dependent apoptosis of pluripotent epiblast cells, probably due to an increase in DNA damage (Fernandez-Diaz et al., 2010). Indeed, a role of Prep1 in the maintenance of genome stability has been recently demonstrated (Iotti et al., 2011). Similar roles of Prep1 at earlier stages could be hypothesized. While the conserved presence of the Prep1 protein in bovine and mouse embryos suggests that the functions of this protein at those stages might be conserved through evolution, the different patterns of gene expression between species might be related to the differences in kinetics of development and implantation, as mentioned before.

Meis, Pbx, and Prep transcription factors are known to be essential for Hox protein activity (Mann et al., 2009). Other studies demonstrated the expression of several Hox genes at the same stages of development. In particular, HOXA3, HOXB7, HOXB9, HOXC9, and HOXD4 transcripts have been detected in the bovine (Ponsuksili et al., 2001; Paul et al., 2011), as well as Hoxa3, Hoxa5, Hoxb7, Hoxb9, and Hoxc9 in the mouse (Kageyama et al., 2007; Paul et al., 2011). Though the presence and localization of the corresponding proteins have not been addressed during preimplantation development, transcription of these genes by the embryonic genome in different mammalian models, as well as our results showing the presence of Meis, Pbx, and Prep proteins, suggest the translation of Hox proteins before gastrulation. Therefore, distinct partnerships between Hox and TALE transcription factors could take place during early lineage specification.

In conclusion, using a systematic approach, we demonstrated, for the first time, the transcription and translation of TALE homeobox genes before gastrulation in the bovine and mouse mammalian models. We showed that at least one member of each Meis, Pbx, and Prep family was found expressed at the RNA and protein levels. These data open the way to studies aiming to understand the functions of TALE factors during early development.

EXPERIMENTAL PROCEDURES

Unless otherwise indicated, all products were purchased at Sigma Aldrich (Saint-Louis, MO).

In Vitro Production of Bovine Embryos

Cumulus-oocyte complexes (COCs) were recovered from abattoir ovaries of slaughtered cows by aspirating follicles from 3 to 8 mm of diameter. Only COCs showing a compact complete cumulus and a homogenous dark oocyte cytoplasm were selected, and then washed 4 times in Hepes-buffered TCM 199. Groups of 60–80 COCs were transferred in 4-well plates containing 500 μl of maturation medium, consisting of a serum-free enriched 199 medium (Donnay et al., 2004). Maturation was performed for 24 hr at 39°C in a humidified air containing 5% CO2.

Motile frozen-thawed bull spermatozoa (kindly provided by the Association wallonne de l'Elevage, Ciney, Belgium) were purified on a Percoll density gradient (45–90%) and diluted to a final concentration of 2×106 ml-1. They were co-incubated with matured COCs for 18 hr in 500 μl of fertilization medium consisting of Tyrode Albumin Lactate Pyruvate medium containing 6 mg/ml fatty acid-free fraction V BSA, 4 mg/ml Na-lactate, 0.11 mg/ml Na-pyruvate, and 1.7 IU/ml heparin, at 39°C in a humidified air containing 5% CO2. The same batch of bull sperm was used throughout the experiments.

After removal of cumulus cells (vortex), presumptive zygotes were cultured in groups of 25–30 embryos in 30-μl droplets of culture medium covered with paraffin oil (Fertipro, Beernem, Belgium) in an atmosphere of 5% O2, 5% CO2, and 90% N2 at 39°C. The culture medium was modified Synthetic Oviduct Fluid (Holm et al., 1999) containing 0.7 mM Na-pyruvate, 4.2 mM Na-lactate, 2.8 mM myo-inositol, 0.2 mM glutamine (Gibco, Paisley, Scotland), 0.3 mM citrate, 30 ml/l essential amino acids, 10 ml/l non-essential amino acids, 50 μg/ml gentamycin, 5 μg/ml insulin, 5 μg/ml transferrin, 5 ng/ml selenium, and 4 mg/ml fatty acid-free fraction V BSA. Embryos were collected at the following stages: precompaction (17–32 cells, 101 hr post-insemination, hpi), compact morula (33–64 cells, 120 hpi), and blastocyst (65–120 cells, 168 hpi), corresponding to the mean time of appearance of each stage in our culture conditions, previously determined by time-lapse cinematography (see Supp. Table S1, which is available online). Only embryos showing good morphological characteristics and from efficient in vitro embryo production (at least 25% blastocyst rates) were selected for the following experiments.

Mouse Embryo Culture

All the experiments were carried out on F1 hybrid mice (DBA/2N X C57BL/6j) (Charles River Laboratories, Brussels, Belgium). Experimental procedures in animals were performed in accordance to the guidelines of the animal ethics committee of the Université catholique de Louvain. According to light cycle (5a.m.–7p.m.), 8-week-old females were superovulated with 10 UI PMSG injected intraperitoneally between 1p.m. and 2p.m. (Folligon, Intervet International, Boxmeer, The Netherlands), followed 47 hr later by 5 UI hCG (Chorulon, Intervet International). After superovulation, females were mated overnight with adult males. Presumptive zygotes were collected from the oviducts 21 hr post-hCG, and cumulus cells were removed using a treatment with hyaluronidase (300 μg/ml). Embryos were cultured in groups of 30–50 embryos in 50-μl droplets of M16 culture medium covered with paraffin oil in an atmosphere of 5% O2, 5% CO2, and 90% N2 at 37°C. Embryos were collected at the following stages: precompaction (8 cells, 68 hr post-hCG), compact morula (16 cells, 80 hr post-hCG), and blastocyst (33–64 cells, 100 hr post-hCG). Only embryos showing good morphological characteristics and from efficient in vitro embryo production (at least 85% blastocyst rates) were selected for the following experiments.

RNA Extraction and cDNA Synthesis

Embryos were collected, washed in nuclease-free PBS, frozen in liquid nitrogen, and stored at −80°C until use. Total RNA was isolated using TRIZOL Reagent (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions and 20 μg of GlycoBlue (Ambion, Austin, TX) was added as a nucleotide carrier. After precipitation with isopropanol, the pellet was washed with 70% ethanol, dried, and resuspended in 8 μl of nuclease-free water. DNase digestion was then carried out to remove possible contamination from genomic DNA by treating the total RNA with 2 U of RQ1 RNase-free DNase (Promega, Leiden, The Netherlands). The RNA quantity and quality could not be evaluated because of the very small cell numbers used for RNA extraction.

The RNA was reverse transcribed using 200 ng of poly-dT (Roche, Penzberg, Germany). All samples were denatured at 65°C for 10 min and then reverse transcribed at 42°C for 1 hr. Reverse transcription reaction mix contained 40UI of Expand Reverse Transcriptase (Roche, Penzberg, Germany), 10 mM DTT, 1 mM dNTP, and 32 UI of RNAse OUT (Roche).

PCR Screening

Meis, Pbx, and Prep gene sequences were obtained from an NCBI database for the two species. Intron spanning primers targeting conserved domains (Pbx and Prep genes) and specific primers (Meis genes) were designed to amplify TALE genes in bovine and mouse embryos (Table 2). All primers were designed to amplify cDNA in both animal models.2

Table 2. Primers Used in RT-PCR
GeneSpecieSequenceAnnealing temperature
Meis1Bovine and mouse5′-AACTGACCAGCCCTCTTGGA-3′60°C
5′-TCGGTTGGACTGGTCTATCA-3′
Meis2Bovine and mouse5′-TTTGAGAAGTGCGAGCT-3′58°C
5′-AGGATGGTGAGTCCCGTGTC-3′
Meis3Bovine and mouse5′-TCCCAGAGTGGGGACAACTC-3′58°C
5′--TGCCACTCTCCTTCTAAGTT-3′
Pbx1, Pbx2, Pbx3, Pbx4Bovine and mouse5′-CATTCAGATGCAGCT-3′43°C
5′-AGCTCCTCTTTGGCTTC-3′
Prep1, Prep2Bovine and mouse5′-TGTTTGAGAAATGTGAACAG-3′47°C
5′-TTGATGAACCAGTTGTT-3′
GAPDHBovine5′-GAGGCTGTCACCAAAAACAT-3′48°C
5′-GGTAGTGGTCAATGCGGTAGA-3′
RNApolIIMouse5′-TTTGCAGAGATTGTCACACCC-3′59°C
5′-AGCCAGTCTGCAGATGAAGG-3′
Table 3. Primers Used in RT-PCR
GeneSpeciesSequenceAnnealing temperature
Pbx1Mouse5′-GAAGCGCAGGCCAGAAAACATGC-3′60°C
5′-AGGATCATGACGGCCTCGCAC-3′
PBX2Bovine5′-GAGGCCGAGGGAAGCAAGAC-3′60°C
5′-CAGGCTTCATGCGGTGGCAGTT-3′
PBX4Bovine5′-GGCATCACGGTCTCCCAGGTCT-3′60°C
5′-TAAGTGCGTCTCCGGGGCCA-3′
Pbx4Mouse5′-GTTCCTGGATGCCAGGCGCAA-3′60°C
5′-CTCTGACCCCTGGGACGCCG-3′
PREP2Bovine5′-CAGCACCGGCCCACCCAAAG-3′60°C
5′-TCCAAGCCCAGGTCGCTGACA-3′
Prep2Mouse5′-AGCTGGAAGTGAGAAGAGGGGGC-3′60°C
5′-GCACAGGTGTGCTGGCGGTA-3′

PCR reactions were carried out in a total volume of 25 μl consisting of 0.5UI of TaKaRa ExTaq polymerase (TAKARA Bio INC., Shiga, Japan), 1× Ex Taq Buffer, 0.25 mM of each dNTP, 0.25 μM of each primer, and 5 μl of cDNA. The cycling program consisted of an initial denaturation step at 95°C for 5 min, followed by 34 cycles of denaturation at 95°C for 30 sec, annealing at primer specific temperature (Table 2) for 30 sec and elongation at 72°C for 1 min. PCR was completed by a final elongation step at 72°C for 10 min.

Generation of appropriate PCR products was analyzed and confirmed on 1% agarose gel stained with ethidium bromide. Mixed PCR products were then inserted into pCR® 2.1-TOPO® vector and transformed into TOP10F′ One Shot Chemically Competent cells using TOPO TA Cloning® Kit (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions. Following blue/white selection, clones were amplified by PCR, cleaned on column with NucleoSpin® Extract II (Macherey-Nagel, Düren, Germany) according to the manufacturer's instructions, and then sequenced. Each sequence was analysed using the GenBank Nucleotide BLAST search tool to assign PCR product to a specific gene.

PCR

Specific intron-spanning primers were designed based on NCBI database sequences to amplify TALE genes by PCR (Table 3). PCR reactions were carried as described above. Appropriate PCR products were analyzed on 2% agarose gel stained with ethidium bromide.

Nested PCR

Specific intron-spanning primers were designed based on NCBI database sequences to amplify Meis1 cDNA first by PCR with 5′-CCCTTCCAGCGGGGGCCATA-3′ and 5′-TGGACCACCCCGGGCTACAT-3′. PCR reactions were carried out as described above. Appropriate PCR products were analyzed on 2% agarose gel stained with ethidium bromide. In order to increase our detection sensitivity, 10 times diluted PCR product was used for nested PCR with 5′-TCCAGCGGGGGCCATACGTC-3′ and 5′-AGCCTTCCCCCTTGCTTTGCG-3′ with the same protocol.

Real-Time qPCR

All specific primer pairs were designed using Primer Express Software (Applied Biosystems, Foster City, CA), based on NCBI database sequences (Tables 4 and 5). The primer specificity was tested using a BLAST analysis against the genomic NCBI database. Primer pairs for Luciferase, GAPDH, YWHAZ, H2afz, and Hprt were chosen respectively from Pennetier et al. (2006), Goossens et al. (2005), Paul et al. (2011), and Mamo et al. (2007). For all the primer pairs, PCR amplification efficiency calculated using a relative standard curve from a pooled cDNA mixture was between 80–100%.

Table 4. Primers Used in RT-qPCR for the Bovine
GeneGenbank accession numberSequenceConcentration (nM)
MEIS1NM_001083507.15′-GCCGATGGCGCAAAGGTACG-3′300
5′-GGTGGACGGGCTGCATGGAC-3′
MEIS3XM_002695108.15′-AGCCTTTGCCAAGCAGGTC-3′100
5′-CAATAGGTGGAAGCGGAGCAC-3′
Pbx1XM_002685833.15′-AGCGTGCAGTCACTCAATGG-3′100
5′-GCGAAGGGTATCCACCTGTG-3′
PBX3NM_001046453.15′-TCTTCAGTGTCCTGTGTGAGATCA-3′100
5′-ATTGTCCAGTCTCATTAGCTGGG-3′
PREP1NM_001046350.15′-GGAGAAGGAGGGAAAGCCC-3′100
5′-AATACGCAGAACCTGGATGGC-3′
GAPDHNM_001034034.15′-TTCAACGGCACAGTCAAGG-3′250
5′-ACATACTCAGCACCAGCATCAC-3′
YWHAZNM_174814.25′-GGTCATCTTGGAGGGTCGCT-3′250
5′-CAGCACCTTCCGTCTTTTGC-3′
Table 5. Primers Uused in RT-qPCR for the Mouse
GeneGenbank accession no. no.NumberSequenceConcentration (nM)
Meis1NM_010789.25′-TAGACCAGTCCAACCGAGCAG-3′200
5′-CGTCCATTACAAAACCTCCCA-5′
Meis3NM_008627.25′-CCTGTGCGACAACTTCTGTCA-3′100
5′-TCCCGATCCTCAATCACCAG-3′
Pbx2NM_017463.25′-CCCTTCCTCTGCAGGCTCTG-3′200
5′-GAAGCAGGGTAGGAATCGCC-3′
Pbx3XM_001478299.15′-CTGGAGAAATATGAACAGGCATGTAA-3′100
5′-AAATGGGACGCGTTCTACTCTG-3′
Pbx4NM_001024954.15′-GGCAACAAAAGAATCCGGTA-3′450
5′-GCAGAGTGAGAACCACGTCA-3′
Prep1NR_027493.15′-CTGACAACCTAATGGTGAAAGCAA-3′350
5′-AATGTACCGACTACAGAAATCTTTACAGAG-3′
H2afzNM_0167505′-ACAGCGCAGCCATCCTGGAGTA-3′100
5′-TTCCCGATCAGCGATTTGTGGA-3′
Hprt1NM_0135565′-GCTTGCTGGTGAAAAGGACCTCTCGAAG-3′100
5′-CCCTGAAGTACTCATTATAGTCAAGGGCAT-3′

Quantitative PCR were performed using the StepOne™ Real Time PCR system (Applied Biosystems, Foster City, CA). Each qPCR reaction mix contained 50% SYBR Green mix (ABgene, UK), 50 to 450 nM of each specific primer, and 2 μl of diluted cDNA. cDNA was synthesized as described above and diluted to obtain one bovine embryo equivalent per reaction, or 2.5 mouse embryo equivalent per reaction. The cycling program consisted of an initial denaturation and enzyme activation step at 95°C for 15 min, followed by 40 cycles of denaturation at 95°C for 20 sec, and combined annealing-elongation step at 60°C for 1 min during which fluorescence was measured. At the end of qPCR reaction, melt curve analysis was performed for all genes to assess primers specificity by the presence of a single peak. Each reaction was run in duplicate and a negative control containing no template was included.

The gene expression data (Cq) were reported to a normalization factor corresponding to the geometric mean of the two accurate reference genes: GAPDH and YWHAZ in bovine preimplantation embryos and H2afz and HprtI in mouse preimplantation embryos (Goossens et al., 2005; Mamo et al., 2007; Vandesompele et al., 2002). Normalised expression of each gene was reported to the precompaction stage (PC=1).

Kruskal-Wallis statistical analyses were performed with SAS Enterprise Guide 4 software (Cary, NC) to compare the relative level of expression between stages. Differences were considered significant at P < 0.05.

Control of TALE Antibodies Specificity

Rabbit anti-Meis1 was kindly provided by M. Torres (Centro Nacional de Investigaciones Cardiovasculares Carlos III, Spain). Rabbit anti-Pbx1 (sc-889), rabbit anti-Pbx2 (sc-890), rabbit anti-Pbx3 (sc-891), and rabbit anti-Prep1 (sc-6245) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Rabbit anti-Prep4 (ARP37406_P050) was purchased from Aviva System Biology (San Diego, CA).

We checked the specificity of each antibody by Western blot analysis on total proteins extract from both bovine and mouse genital tract, based on previous data in mouse and human (Wagner et al., 2001, Villaescusa et al., 2004, Dintilhac et al., 2005, Ota et al., 2008). We used the same antibodies in both species according to amino acid sequence similarities between mouse and bovine TALE proteins. Specificity controls for TALE antibodies by Western blot analysis are shown in Figure 6.

Figure 6.

Specificity control of Meis1, Pbx1, Pbx2, Pbx3, and Prep1 antibodies in mouse and bovine uterus by Western blot analysis. Thirty micrograms of total protein extracts from bovine and mouse genital tract were loaded on an SDS-PAGE gel and probed with each of the five TALE antibodies. A major band for each antibody is observed at the expected size, except for bovine MEIS1 where minor additional bands are also observed. M, mouse; B, bovine. Molecular weight (kDa) is indicated.

The polyclonal rabbit anti-Meis1 antibody detected as expected a single band at ∼50 kDa in mouse uterus total protein extracts. However, in bovine extracts, additional bands are observed probably corresponding to MEIS1 isoforms or modified forms of the protein. Polyclonal rabbit anti-Pbx1 antibody recognized a band at ∼50-kDa protein in mouse ovaries (Villaescusa et al., 2004), as, respectively, in bovine and mouse uterus extracts. Likewise, polyclonal rabbit anti-Pbx2 antibody stained a band at ∼50-kDa protein in mouse uterus, as previously showed in mouse ovary extracts (Villaescusa et al., 2004). The same anti-Pbx2 antibody was also tested on bovine uterus tissues but gave several aberrant bands, which was considered as not specific in this species. The polyclonal rabbit anti-Pbx3 antibody stained a single band at ∼50 kDa in mouse uterus extracts. In bovine extracts, this antibody detected a single band as well but at a lower molecular weight according to the smaller amino acid sequence of the bovine protein compared to the mouse. The polyclonal rabbit anti-Prep1 antibody detected a single band at ∼64 kDa in both bovine and mouse extracts, as previously showed in mouse ovary (Villaescusa et al., 2004). The polyclonal rabbit anti-Pbx4 antibody was only tested in the mouse as no expression was observed in the bovine, but no specific reproducible signal could be obtained on mouse testis extracts. Therefore, the presence of Pbx4 protein was not investigated in mouse embryos.

Whole-Mount Immunofluorescence

Groups of 9 embryos per stages (precompaction, compact morula, and blastocyst) were washed in PBS and fixed with 4% paraformaldehyde in PBS overnight at 4°C. After washing three times in PBS, embryos were permeabilized with 1% Triton X-100 (Roche) in PBS for 30 min at room temperature. Non-specific binding sites were blocked with blocking buffer in PBS containing 5% goat serum for 1 hr at room temperature. Embryos were then incubated with primary antibody O/N at 4°C. After washes, embryos were incubated with secondary antibody, goat anti-rabbit Alexafluor555 conjugated (Invitrogen, Carlsbad, CA), for 1 hr at room temperature in the dark. The nuclei were stained in Vectashield mounting medium containing DAPI (Vector Laboratories H1200). A negative control without primary antibody was performed simultaneously to check for the non-specific binding of the secondary antibody. Samples were analysed with confocal microscope (LSM710, Zeiss, Jena, Germany). For each protein, three embryos of each stage (precompaction, compact morula, and blastocyst) were handled at the same time and the experiment was repeated on three different batches of embryos.

Western Blot (Meis3)

Goat anti-Meis3 and anti-goat HRP conjugated were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Pools of 150 bovine or 500 mouse blastocysts were collected, washed twice in PBS, and loaded in SDS buffer (250 mM, pH 6.8 Tris-HCl, 500 mM DTT, 10% SDS, 0.5% bromophenol blue, and 50% glycerol). Subsequently, 11.5E embryo and tissues from female genital tract were collected, washed twice in PBS, lysed in RIPA buffer (250 mM NaCl, 50 mM, pH 7.5 TrisHCl, 1% Nonidet P40, 0.5 % NaDeoxycholate, 0.1 % SDS, 1 mM EDTA, 20 mM Imidazole, and 5× protease inhibitor). Total protein extracts were run on an SDS-PAGE and blotted on nitrocellulose membranes. Non-specific binding sites were blocked with TBS containing 0.1% Tween20 and 10% milk for 1 hr at room temperature. Nitrocellulose membranes were then incubated with TBS containing primary antibody and 1% milk O/N at 4°C. After washes, membranes were incubated with TBS containing anti-goat HRP conjugated and 1% milk for 1 hr at room temperature. The anti-β-actin HRP conjugated was used after detection with TALE antibody to control the protein load.

Whole-Mount In Situ Hybridization (Meis3)

Primers used for bovine MEIS3 riboprobes were 5′-CAGGATCAATGGGGATGAGT-3′ and 5′-GAAGGGAACCTCCCCAAAT-3′; mouse Meis3 riboprobes were 5′-CCCAGCTCCGCAGGAGAGGA-3′ and 5′-GGGTGCAGGACCTGGAGGCT-3′. Standard alkaline phosphatase whole-mount in situ hybridizations were performed according to Piette et al. (2008) with some modifications, with only embryos showing normal morphological characteristics for their stage and from efficient in vitro embryo production. In brief, embryos were fixed in 4% PFA in PBS overnight at 4°C, and then stored in 100% methanol. After rehydration, embryos were permeabilized with RIPA buffer for 10 min at RT, post-fixed with 4% paraformaldehyde and 0.2% glutaraldehyde for 20 min at RT, and then prehybridized in hybridization solution (50% formamide, 1% blocking reagent, 5× SSC, 1 mg/ml yeast extract, 0.1 mg/ml heparine, 0.1% CHAPS, 5 nM EDTA) for 1 hr at 65°C. Probes were hybridized overnight at 65°C. Post-hybridization washes were performed to remove excess probe. The reactions were blocked with manufacturer's blocking solution for 3 hr at RT, and then treated with 1:500 anti-DIG antibody conjugated to alkaline phosphatase (Roche) overnight at 4°C. Colorimetric reactions were performed with BM purple (Roche).

Acknowledgements

We gratefully acknowledge Miguel Torres for the gift of anti-Meis1 antibody, Francesco Blasi and Licia Selleri for their precious advice. We thank Abdelmounaim Errachid for his technical support and advice about confocal microscopy analysis. We thank Françoise Gofflot for the fruitful discussions about whole-mount in situ hybridization and all members of Isabelle Donnay and René Rezsöhazy's research groups for suggestions and help. We also acknowledge Philippe Bombaerts, Raphaël Chiarelli, Fabienne George, Delphine Paul, and Emmanuelle Ghys for their assistance in embryo production and Marie-Anne Mauclet for her help with administrative procedures. Wendy Sonnet holds a FRIA fellowship, FNRS, from the Communauté française de Belgique.

Ancillary