A gradient of maternal Bicaudal-C controls vertebrate embryogenesis via translational repression of mRNAs encoding cell fate regulators

Bicaudal-C (Bicc1) proteins are evolutionarily conserved RNA-binding proteins in metazoans with important roles in a variety of biological contexts (Gamberi and Lasko, 2012). The founding member of the Bicc1 family, BicC, was identified as a maternal effect gene in Drosophila melanogaster (Bull, 1966). Mutations that disrupt maternal BicC give rise to fly embryos with anterior-posterior defects (Mahone et al., 1995; Saffman et al., 1998). Recent studies in vertebrates provide clear evidence that Bicc1 functions to control post-maternal stages of development and homeostasis of several organs. For example, homozygous Bicc1 mutant mice develop into adults but exhibit defects of their kidneys, pancreas and heart (Maisonneuve et al., 2009). Some mutations in the human BICC1 gene are associated with the kidney disease cystic renal dysplasia, whereas others are associated with defects in bone density (Kraus et al., 2012; Mesner et al., 2014). Collectively, these examples reveal the importance of zygotically expressed Bicc1 to normal vertebrate development and health, but they do not address the role of maternal Bicc1 in vertebrate embryogenesis. Thus, although it is clear that vertebrate oocytes, eggs and embryos express Bicc1 mRNA, suggesting an important role for Bicc1 in the maternal stages of embryogenesis, no study to date has directly addressed this issue by generating mutants specifically lacking maternal Bicc1 (Wessely and De Robertis, 2000; Wessely et al., 2001; Bouvrette et al., 2010).

Several lines of evidence indicate that Bicc1 proteins function as translational repressors of specific target mRNAs. However, only a few studies present data linking Bicc1-dependent translation repression of relevant target mRNAs to its biological roles (Chicoine et al., 2007; Piazzon et al., 2012; Zhang et al., 2013). In recent studies of maternally controlled stages of embryogenesis of Xenopus laevis, we have presented evidence that several maternal mRNAs encoding important cell-fate regulators can bind Bicc1 and are potential targets of Bicc1-mediated translational repression, including most notably the cripto1 (tdgf1.3 – Xenbase) mRNA for which we have defined a Bicc1 RNA-binding element in the 3′ untranslated region (3′UTR) (Zhang et al., 2013). However, although these data are consistent with a model in which the translational repression activity of Bicc1 regulates the earliest steps of vertebrate embryogenesis, such as establishing the body axes, experiments that directly address the role of maternal Bicc1 in vertebrates are needed to test this model and establish a connection between the biological and biochemical roles of Bicc1.

In this study, we directly examine the role of maternal Bicc1 in a vertebrate organism by specifically depleting both bicc1 mRNA and Bicc1 protein from Xenopus laevis eggs using two different antisense oligonucleotides in independent experiments and fertilizing these eggs using the host-transfer technique. The Bicc1-depleted embryos exhibited greatly reduced levels of Bicc1 protein, developed with an excess of dorsal-anterior structures and overexpressed several genes that define Spemann's organizer, the organizing tissue that controls cell-fate decisions after zygotic transcription begins (reviewed by Houston, 2012). In previous studies, we identified the cripto1 and coco (dand5 – Xenbase) mRNAs as potential targets of Bicc1-mediated repression (Zhang et al., 2013). These data along with previous phenotypic studies of cripto1 and coco mutants help link the biological role of Bicc1 in embryogenesis documented here to its biochemical role in translational repression (Tao et al., 2005; Bates et al., 2013). In addition, based on the phenotypes of our Bicc1-depleted embryos, we found that phenotypes of Wnt11b loss-of-function and gain-of-function (overexpression) embryos were consistent with Bicc1 regulation of Wnt11b expression (Tao et al., 2005). Therefore, we hypothesized that wnt11b mRNA is a direct Bicc1 target. Here, we provide evidence in support of this hypothesis by showing that Bicc1 could bind to wnt11b mRNA in embryos and that this mRNA contains a Bicc1-dependent translation repression element within its 3′UTR. Finally, using antibodies produced against Bicc1 we show that endogenous Bicc1 protein is present in eggs and embryos and exists in a vegetal-to-animal pole concentration gradient that mirrors Bicc1-mediated translation repression activity. In summary, this study presents the first direct examination of the role for maternal Bicc1 in vertebrates and, together with data from our previous work, begins to establish the connections between this role and the biochemical function of Bicc1 in mRNA-specific translational repression.

To address the relevance of maternal Bicc1 to X. laevis development, we performed maternal mRNA depletion experiments using antisense oligonucleotides (oligos) (Heasman et al., 1991; Olson et al., 2012). Full-grown X. laevis oocytes were injected to identify antisense oligos that would cleave and deplete maternal bicc1 mRNA. Two different antisense oligos (as9460 and as9463) were effective at reducing bicc1 mRNA levels as monitored by RNA blot hybridization and qRT-PCR (Fig. 1A,B). Importantly, these oligos targeted both Bicc1 homologs present in the allotetraploid X. laevis genome that are predicted to encode the same Bicc1 protein (http://www.xenbase.org). Examination of endogenous maternal Bicc1 protein has not been reported, but assessing Bicc1 protein levels was crucial to interpretation of these experiments. Therefore, we raised an antibody to the N-terminal half of the X. laevis Bicc1 protein and used it to examine Bicc1 protein in control and oligo-treated oocytes and eggs by protein immunoblotting (Fig. 1C,D). Bicc1 protein was not detectable in oocytes but accumulated during oocyte maturation, indicating that maternal Bicc1 protein is already present in unfertilized eggs (Fig. 1C). However, oocytes injected with the antisense oligos as9460 and as9463 failed to accumulate Bicc1 protein during maturation (Fig. 1D), indicating that these oligos effectively eliminated the majority of maternal bicc1 mRNA and Bicc1 protein.

To perform maternal loss-of-function experiments, antisense oligos were synthesized in a phosphorothioate-modified form and injected into oocytes. These oocytes were then matured in vitro, transferred into ovulating host females, recovered as eggs, and fertilized to examine the effects of bicc1 depletion on development (Olson et al., 2012). Depletion with each oligo yielded similar results. Bicc1-depleted embryos developed normally up to gastrulation, but were delayed in completing blastopore closure and often formed thickened blastopore lips with deep-set yolk plugs (Fig. 2A). This phenotype is often associated with axis defects. By the tailbud stages, bicc1-depleted embryos displayed an excess of dorsal-anterior structures that included enlarged heads, often accompanied by expansion of the cement gland. Dorsoanterior index (DAI) scores ranged from 6 to 7 (Fig. 2B,C,E; Table 1) (Kao and Elinson, 1988). To verify that the phenotypes were due to loss of maternal Bicc1, rescue experiments were performed in which Bicc1-depleted embryos were injected with in vitro-synthesized bicc1 transcript (20 pg) shortly after fertilization. This relatively low dose and post-fertilization injection regimen was necessary, as excess Bicc1 was a highly potent inhibitor of dorsal development and of gastrulation. The addition of the bicc1 mRNA reduced the proportion of embryos exhibiting exaggerated dorsal-anterior structures and even induced ventralization in some samples, including a fraction of the control samples (Fig. 2E; Table 1). These findings suggest that the concentration and localization of Bicc1, both challenging to control precisely in microinjection experiments, are crucial to normal embryogenesis. Nevertheless, these data indicated that bicc1 transcript could reverse the dorso-anteriorization associated with Bicc1 depletion and thus indicated the specificity of the antisense oligos for bicc1 mRNA. Thus, we conclude that Bicc1 depletion and not off-target effects cause the mutant phenotypes.

During X. laevis development, the formation of dorsal-anterior structures relies on activities of organizer-specific genes, such as goosecoid homeobox (De Robertis, 2006; Houston, 2012). The phenotypes of Bicc1-depleted embryos indicated that organizer function was overactive. To test this directly, we analyzed organizer gene expression at gastrulation (stage 11) using qRT-PCR. Expression of the organizer-specific genes siamois homeodomain 1 (sia1), goosecoid homeobox (gsc), nodal3.1 (also known as Xnr3) and noggin (nog) were increased in Bicc1-depleted embryos relative to uninjected controls, whereas the expression of genes expressed by the ventral/posterior cells that flank the organizer, ventx1.2 and wnt8a, were decreased (Fig. 2F). Thus, depletion of Bicc1 led to over-development of the organizer.

Embryos depleted of Bicc1 developed with expanded dorsal-anterior structures (Fig. 2). This phenotype showed similarity to that described for embryos that overexpress the Wnt11b ligand (Tao et al., 2005), raising the possibility that wnt11b mRNA was a direct or indirect target of Bicc1 translational repression. Although wnt11b mRNA was not identified as a Bicc1 target in our previous studies using an ectopically expressed HA-Bicc1 immunoprecipitation strategy that identified several mRNAs relevant to cell-fate decisions, there were several possible technical reasons that could cause relevant direct targets to be missed (Zhang et al., 2013). Therefore, we tested whether wnt11b mRNA encoded elements sufficient to allow Bicc1-mediated repression using a luciferase reporter mRNA assay described in our previous work (Zhang et al., 2013). Specifically, an mRNA containing the luciferase coding region fused to the X. laevis wnt11b mRNA 3′UTR was injected into animal cells of 4- to 8-cell embryos, with and without a synthetic mRNA encoding HA-tagged Bicc1 (HA-Bicc1) (Fig. 3A,B). When embryos reached stage 8-9, extracts were prepared and assayed for luciferase activity. The extent of repression by Bicc1 was measured by determining the ratio of luciferase activity for each reporter mRNA in the presence or absence of Bicc1. The reporter mRNA containing the wnt11b 3′UTR was repressed by co-expression of HA-Bicc1 to the same extent as the reporter containing the cripto1 3′UTR that we previously identified and characterized as a Bicc1-regulated mRNA (Fig. 3C) (Zhang et al., 2009,, 2013). The effect was specific, as a reporter mRNA containing the 3′UTR from the cyclin B1 (ccnb1) mRNA, an mRNA not regulated by Bicc1, was not repressed. Additional controls indicated that the effects on luciferase expression were not caused by differences in the amount of HA-Bicc1 expressed or the stability of the mRNA reporters (Fig. 3D,E).

If Bicc1 regulation of wnt11b mRNA were direct, then Bicc1 should be capable of physically associating with this mRNA. To test this possibility, an mRNA encoding HA-Bicc1 was injected into animal cells of 8-cell embryos. When the embryos developed to the blastula stage (stage 7) the HA-Bicc1 protein was immunoprecipitated and the bound RNAs were analyzed by qRT-PCR using specific primers (Fig. 4A). wnt11b mRNA was present in the HA-Bicc1 immunoprecipitate, as was cripto1 mRNA, a previously established Bicc1 target (Fig. 4B) (Zhang et al., 2013). The Bicc1 association was specific as the abundant cyclin B1, odc1 or eif1 mRNAs were not immunoprecipitated. We note that that HA-Bicc1 was less efficient at immunoprecipitating the endogenous wnt11b mRNA compared with cripto1 mRNA. Perhaps this observation explains why wnt11b mRNA was not identified in our previous screen for Bicc1 target mRNAs (Zhang et al., 2013). Regardless, these experiments provided evidence that Bicc1 could physically associate with wnt11b mRNA in embryos.

Our previous studies of cripto1 mRNA indicate that its regulation by Bicc1 is mediated via the cripto1 mRNA 3′UTR. Therefore, we investigated whether the 3′UTR of the wnt11b mRNA was sufficient for Bicc1 association by performing HA-Bicc1 immunoprecipitation experiments in embryos co-injected with reporter luciferase reporter RNAs containing the indicated 3′UTRs (Fig. 3B). The HA-Bicc1 immunoprecipitates were analyzed by qRT-PCR using primers to the luciferase-coding region. HA-Bicc1 associated with the reporter mRNAs containing either the 3′UTR of wnt11b mRNA or of the positive control cripto1 mRNA. By contrast, HA-Bicc1 did not associate with the negative control reporter mRNA containing the 3′UTR of cyclin B1 mRNA (Fig. 4C). Together, these results indicate that Bicc1 can associate directly with the wnt11b mRNA in embryos and that this interaction is mediated by sequences within the 3′UTR of wnt11b mRNA.

mRNAs containing the 3′UTRs of Bicc1 target mRNAs are efficiently repressed in vegetal cells but not in animal cells of X. laevis embryos (Zhang et al., 2013). To determine whether this differential repression might be attributed to the distribution of endogenous maternal Bicc1 protein, X. laevis blastula stage embryos (stage 7) were dissected into three regions (animal, marginal zone and vegetal) and analyzed for endogenous Bicc1 protein by protein immunoblotting. Vegetal cells contained the highest levels of Bicc1 and animal cells contained the lowest levels, whereas marginal zone cells contained a level of Bicc1 protein intermediate between these two poles (Fig. 5A). By contrast, cytoskeletal actin was uniformly distributed between these three regions of the embryo. Thus, endogenous Bicc1 protein was differentially distributed along the animal-vegetal axis. To address whether the repression activity attributed to Bicc1 followed this same gradient, reporter mRNAs containing Bicc1-responsive 3′UTRs were injected into animal, marginal zone, or vegetal cells of 16-cell-stage embryos. Extracts were prepared from injected embryos after they reached stage 8-9 and assayed for luciferase activity. The luciferase activity of each reporter mRNA introduced into each injection site was normalized to animal cell injections. The repression of cripto1 and wnt11b reporter mRNAs was highest in vegetal cells and lowest in animal cells, with marginal zone cells exhibiting an intermediate level of repression (Fig. 5B). By contrast, the negative control reporter containing the 3′UTR of cyclin B1 mRNA was translated at similar levels in all three embryonic locations. The differences in luciferase activity observed for the cripto1 and wnt11b reporter mRNAs were due to translational repression and not differences in mRNA stability as all the reporter mRNAs were present in similar amounts in all three regions analyzed as assayed by qRT-PCR (Fig. 5C). Thus, our results strongly suggest a model in which the gradient of endogenous maternal Bicc1 protein establishes the animal-vegetal gradient of Bicc1 translational repression necessary for the maternally controlled stages of X. laevis embryogenesis.

In a previous study and in the experiments described above, we have shown that HA-Bicc1 protein generated in animal cells, by injection of HA-bicc1 mRNA, is sufficient to cause Bicc1-dependent translational repression of reporter mRNAs and to associate with endogenous target mRNAs in animal cells (Fig. 3) (Zhang et al., 2013). These observations provide evidence that Bicc1 protein itself is the limiting component for Bicc1-mediated translational repression in animal cells and that everything else needed for repression must be present (Zhang et al., 2013) (Fig. 3). To test the converse hypothesis, specifically that Bicc1 is present in excess in vegetal cells, the reporter containing the 3′UTR of the cripto1 mRNA was injected into animal cells or vegetal cells of 8-cell X. laevis embryos in the presence or absence of HA-bicc1 mRNA. Extracts from the injected cells were analyzed for luciferase activity and the amount of Bicc1-dependent repression was determined. As expected, animal cells repressed the cripto1 reporter mRNA only when additional Bicc1 was provided (Fig. 5D,E). However, and in contrast to these observations, the cripto1 reporter mRNA was not repressed to any further extent in vegetal cells when additional Bicc1 was provided. Thus, Bicc1 translational repression activity was in excess in vegetal cells, consistent with our observation that the endogenous Bicc1 protein was concentrated in these cells.

This study presents the first direct evidence that the maternal Bicc1 translational repressor regulates the earliest maternally controlled patterning events in vertebrates. Specifically, depletion of the maternal bicc1 mRNA and Bicc1 protein from X. laevis embryos caused them to develop with an excess of dorsal-anterior structures. This phenotype is most likely to be caused by defects in the maternal steps required for formation of Spemann's organizer because several organizer-specific genes are overexpressed in the Bicc1-depleted embryos. We examine endogenous maternal Bicc1 protein and show that it forms a concentration gradient across the embryo from high in the vegetal pole to low in the animal pole that mirrors the activity levels of Bicc1-dependent translational repression. Thus, a concentration gradient of maternal Bicc1 protein establishes the gradient of Bicc1-dependent translational repression function. We propose that this maternal Bicc1 gradient establishes additional molecular asymmetries of cell-fate regulators, including key signaling proteins that drive embryonic patterning and cell-fate decisions. Finally, the phenotype of Bicc1-depleted embryos led us to identify a new relevant maternal Bicc1 target, wnt11b mRNA. Our data provide evidence that Bicc1 attenuated translation of wntb11 mRNA directly by associating with the 3′UTR of this mRNA. The observations presented here, together with our previous study that identifies coco and cripto1 mRNAs as targets of Bicc1-mediated repression, provide a link between the biological roles of Bicc1 in the maternal stages of vertebrate embryogenesis to its biochemical role as a translational repressor (Zhang et al., 2013). Thus, we propose a model in which Bicc1-specific translational repression helps establish many of the animal-vegetal embryonic asymmetries of cell-fate regulatory proteins, which in turn direct the embryonic patterning and earliest cell-fate decisions of vertebrate development (Sheets et al., 2016).

A challenge for examining maternal effect genes in vertebrates is that the protein might already be present in the egg from maternal stores. An important contribution of this study is that the antisense oligos used to abolish the maternal bicc1 mRNA in oocytes allowed for the depletion of both bicc1 mRNA and Bicc1 protein. Using an antibody to detect endogenous Bicc1, we show that substantial amounts of Bicc1 are already present in the egg prior to fertilization, having accumulated during oocyte maturation. Thus, to examine the maternal role for Bicc1, depleting bicc1 mRNA from oocytes and then using the host-transfer method to allow for fertilization of the depleted eggs was key. Indeed, a previous study injected morpholinos to disrupt bicc1 mRNA translation in 2- to 4-cell-stage Xenopus embryos and showed that embryos developed to later stages with kidney defects (Tran et al., 2007). These data are fully consistent with studies of homozygous Bicc1 mutants in mice and support a role for zygotic Bicc1 in normal kidney formation in vertebrates (Maisonneuve et al., 2009; Tran et al., 2010). Importantly, our work shows that by the 2- to 4-cell stage a substantial amount of Bicc1 protein is already present. Thus, the kidney defects from morpholino treatment are probably the result of disrupting zygotic Bicc1 function but leaving maternal Bicc1 function intact.

The X. laevis bicc1 maternal mRNA is localized to the vegetal cortex during oogenesis (Wessely and De Robertis, 2000). Here, we show that the bicc1 mRNA itself is translationally regulated during oocyte maturation; the protein is absent from oocytes, but accumulates during oocyte maturation. Our data also reveal that the Bicc1 protein is distributed in a gradient along the animal-vegetal axis. This protein gradient is likely to be a direct result of bicc1 mRNA localization achieved during oogenesis. The Bicc1 translational repression activity we measured tracked with this concentration gradient. Thus, mRNA localization during oogenesis sets the stage for a maternal Bicc1 protein gradient that goes on to have a profound impact on embryogenesis by establishing additional molecular gradients via translational repression of cell-fate regulatory mRNAs such as coco, cripto1 and wntb11 (Sheets et al., 2016). Essentially, a cascade of gradient amplification is achieved through Bicc1-mediated repression, with this initial gradient establishing multiple secondary gradients. Precisely how Bicc1-repression is tailored or fine-tuned for specific targets to achieve the proper combination of cell-fate regulatory gradients remains unknown.

A previous study of mouse embryos revealed that Bicc1 affected development at least in part through the Wnt signaling pathway (Maisonneuve et al., 2009). Thus, it was particularly notable that in this study we identified wnt11b mRNA as a direct target of Bicc1. Our earlier work identified several potential maternal mRNA targets for Bicc1 making it likely that the Bicc1-repression mechanism fine-tunes the expression of multiple signaling pathways (Zhang et al., 2013). However, it is interesting to note that some of the other Bicc1 mRNA targets we identified in our previous studies have also been implicated as modulators of Wnt signaling. For example, we have shown that cripto1 mRNA is a Bicc1 target, and the Cripto1 protein affects multiple signaling events, including those of the Wnt pathway (Nagaoka et al., 2012,, 2013; Klauzinska et al., 2014). We also identified ddx5 mRNA as a Bicc1 target (Zhang et al., 2013), and Ddx proteins are documented modulators of Wnt signaling (Cruciat et al., 2013). These observations raise the possibility that Bicc1 is a major regulator of Wnt signaling localization and strength via its attenuation of expression of multiple proteins that regulate this pathway. Whether the bulk of Bicc1 regulation of maternal development can ultimately be explained by Wnt pathway modulation or instead involves modulation of other independent signaling pathways remains an open question.

X. laevis oocyte and embryos were obtained and injected as described (Sive et al., 2000). Host-transfer experiments were performed using antisense oligonucleotides (oligos) against bicc1. These were synthesized as HPLC-purified phosphorothioate-phosphodiester chimeric oligos (Integrated DNA Technologies) with the following sequences (asterisks indicate phosphorothioate linkages): bicc1 9460, 5′-G*C*GTGTTTGTCTCTTC*C*A-3′ (nucleotides 180-162 where the A of the AUG start codon is nucleotide 1); bicc1 9463, 5′-T*G*TAACATTGTCTCGAG*C*T-3′ (nucleotides 374-357). Oocytes were injected in the vegetal pole and cultured for 24 h at 18°C before being matured by treatment with 2.0 μM progesterone. Matured oocytes were colored with vital dyes, transferred to egg-laying host females, recovered and fertilized essentially as described (Heasman et al., 1991; Olson et al., 2012). For rescue experiments, HA-bicc1 mRNA, which encodes Bicc1 with an N-terminal HA epitope tag, was injected into vegetal cells of bicc1-depleted embryos shortly after fertilization. The injected HA-bicc1 mRNA used for rescue was not affected by the antisense oligos because the oligos are degraded a few hours after injection.

Firefly luciferase reporter mRNAs were generated that contained the translational control element of the cripto1 mRNA 3′UTR, the 3′UTR of the cyclin B1 mRNA or the 3′UTR of the wnt11b mRNA (Sheets et al., 1994; Fritz and Sheets, 2001; Zhang et al., 2009). All plasmids were linearized with appropriate restriction enzymes and used to generate capped mRNAs as described (Sheets et al., 1994; Fritz and Sheets, 2001; Zhang et al., 2009,, 2013).

Each reporter mRNA was diluted to 2.5 nM and 5 nl (12.5 amol) was injected into embryonic cells. When injected embryos reached the appropriate stage, extracts were prepared and analyzed for luciferase activity (Sheets et al., 1994; Fritz and Sheets, 2001; Zhang et al., 2009).

Total RNA from embryos injected with reporter mRNAs was analyzed by RNA blot hybridization (Sheets et al., 1994; Zhang et al., 2009) using a radiolabeled probe to detect the bicc1 mRNA.

The analysis of proteins by immunoblotting was performed as described (Zhang et al., 2013) using mouse monoclonal anti-HA-tag antibody and anti-Actin antibody. Anti-Bicc1 polyclonal antibodies were raised in rabbits (Harlan Labs, Madison, WI, USA) immunized with a GST-Bicc1 fusion protein encoding the N-terminal 486 amino acids of X. laevis Bicc1 expressed in Escherichia coli. Antibodies were affinity purified using recombinant MBP-Bicc1 fusion protein (the N-terminal 486 amino acids; Fig. S1) and used at 1:2000 for immunoblotting.

Quantitative RT-PCR to analyze mRNAs associated with Bicc1 was performed as described (Park et al., 2011). Primers used to analyze changes in gene expression have been previously published (Kofron et al., 2004; Heasman et al., 1991; Luxardi et al., 2010).

Embryos were injected with mRNA encoding Bicc1 fused at the N-terminus with an HA epitope tag (Ha-Bicc1) (Zhang et al., 2013). When injected embryos reached the appropriate blastula stage (stage 7), injected embryos were lysed in 100 μl of TNMEN-150 buffer (Cooke et al., 2010). The lysate was centrifuged (4°C, 10 min at 5000 rpm, 2700 g) and the supernatant incubated with anti-HA antibody coupled to protein-G agarose (2 h, 4°C). The beads were collected (1 min, 3000 rpm, 960 g) and washed 4 times in 1 ml TNMEN-150 buffer. For each wash, the beads were incubated in buffer at 4°C for 5 min, spun at 3000 rpm (960 g) for 1 min and supernatant removed. RNA was isolated from the washed beads for analysis.

We thank Catherine Fox (University of Wisconsin-Madison) for useful discussions and edits on the manuscript. We also thank Mary Lou King (University of Miami) for providing a plasmid containing the Wnt11b 3′UTR and Laura Vanderploeg for preparing the figures.