Dissecting the differentiation process of the preplacodal ectoderm in zebrafish


  • Di Yao,

    1. The Education Ministry Key Laboratory of Cell Proliferation and Differentiation and the State Key Laboratory of Bio-membrane and Membrane Bio-engineering, School of Life Sciences, Peking University, Beijing, China
    Search for more papers by this author
    • Drs. Yao, Zhao, and Wu contributed equally to this work.

  • Feng Zhao,

    1. The Education Ministry Key Laboratory of Cell Proliferation and Differentiation and the State Key Laboratory of Bio-membrane and Membrane Bio-engineering, School of Life Sciences, Peking University, Beijing, China
    Search for more papers by this author
    • Drs. Yao, Zhao, and Wu contributed equally to this work.

  • Ying Wu,

    1. The Education Ministry Key Laboratory of Cell Proliferation and Differentiation and the State Key Laboratory of Bio-membrane and Membrane Bio-engineering, School of Life Sciences, Peking University, Beijing, China
    Search for more papers by this author
    • Drs. Yao, Zhao, and Wu contributed equally to this work.

  • Jialiang Wang,

    1. The Education Ministry Key Laboratory of Cell Proliferation and Differentiation and the State Key Laboratory of Bio-membrane and Membrane Bio-engineering, School of Life Sciences, Peking University, Beijing, China
    Search for more papers by this author
  • Wei Dong,

    1. The Education Ministry Key Laboratory of Cell Proliferation and Differentiation and the State Key Laboratory of Bio-membrane and Membrane Bio-engineering, School of Life Sciences, Peking University, Beijing, China
    Search for more papers by this author
  • Jue Zhao,

    1. The Education Ministry Key Laboratory of Cell Proliferation and Differentiation and the State Key Laboratory of Bio-membrane and Membrane Bio-engineering, School of Life Sciences, Peking University, Beijing, China
    Search for more papers by this author
  • Zuoyan Zhu,

    1. The Education Ministry Key Laboratory of Cell Proliferation and Differentiation and the State Key Laboratory of Bio-membrane and Membrane Bio-engineering, School of Life Sciences, Peking University, Beijing, China
    Search for more papers by this author
  • Dong Liu

    Corresponding author
    1. The Education Ministry Key Laboratory of Cell Proliferation and Differentiation and the State Key Laboratory of Bio-membrane and Membrane Bio-engineering, School of Life Sciences, Peking University, Beijing, China
    • Correspondence to: Dong Liu, The Education Ministry Key Laboratory of Cell Proliferation and Differentiation and the State Key Laboratory of Bio-membrane and Membrane Bio-engineering, School of Life Sciences, Peking University, Beijing 100871, China. E-mail: doliu@pku.edu.cn

    Search for more papers by this author


Background: The preplacodal region (PPR) is a region of specialized ectoderm at the border of neural and nonneural ectoderm (NNE). Coordinated Bmp, Fgf, and Wnt signals are known to drive PPR development; however, the underlying mechanism is unknown. Results: We identified key components involved in PPR differentiation. The mesoderm/marginal Wnts at the early gastrula stage trigger differentiation by allowing the adjacent NNE border cells to start adopting caudal PPR fates; otherwise, the development of caudal PPR identity is hindered due to the persistent presence of gata3 mRNA. The caudal PPR fate dominates when foxi1 expression is enhanced at the late gastrula stage, and depleting Foxi1 after 6 hours postfertilization (hpf) reduces the otic-epibranchial placodal domain. When the Gata3 level is manipulated at the fertilized egg stage or near 6 hpf, the lens is always affected. In establishing PPR polarity, both Gata3 and Foxi1 inhibit Bmp signaling, whereas Foxi1 inhibits, but Gata3 enhances, Fgf sensitivity of the PPR cells. Conclusions: Our study reveals that in zebrafish, (1) the PPR at the shield stage may enter a developmental state when the PPR cells preferentially adopt a particular placodal fate and (2) a network of genetically linked factors, including Wnt/beta-catenin, Fgfr, Bmp, Gata3, and Foxi1, direct the process of PPR differentiation. Developmental Dynamics 243:1338–1351, 2014. © 2014 Wiley Periodicals, Inc.


The vertebrate ectoderm is composed of neural ectoderm and nonneural ectoderm. The neural and nonneural ectoderm (NNE) border gives rise to the preplacodal region (PPR) and neural crest (NC) cell populations (Groves and Bronner-Fraser, 2000; Stern, 2005; Bailey and Streit, 2006; Bhattacharyya and Bronner-Fraser, 2008; Kwon et al., 2010). At the end of gastrulation, NC and PPR markers start to highlight the border, which is thought to form after neural specification (Basch et al., 2000; Kobayashi et al., 2000; Ahrens and Schlosser, 2005; Raible and Ragland, 2005). The first PPR fate map was derived from a zebrafish study and revealed a distinct area of ectoderm in which most cells adopt PPR and NC fates. This area lies along the lateral edge of the neural ectoderm at the shield stage (Kozlowski et al., 1997). Although largely intermingled, these cells are arranged rostra-caudally along the future axis, suggesting that PPR cells may preferentially adopt a particular placodal fate while the zebrafish organizer is established.

Chick fate-mapping experiments reveal that at the neural plate (one-somite) stage, the anterior and posterior placodal fates are approximately divided into rostral and caudal portions along the developing neural plate, although the placodal fates and some NC fates are not completely segregated (Streit, 2002), suggesting that the PPR may form earlier. Salamander studies show that at the late neural plate stage, if a donor PPR strip is excised, rotated 180° and placed in a host embryo of the same age whose own PPR strip has been removed, the otic placodes often develop rostrally and the olfactory placodes caudally. The frequency of such misalignments is lower when the procedure is conducted at the early neural plate stage (Jacobson, 1963). Both chick and frog transplantation studies have found that the otic fate is first determined at the one- to four-somite stage (the later neural plate stage), supporting the idea that the PPR forms much earlier, possibly before or near the early neural plate stage (Gallagher et al., 1996; Streit et al., 2000; Delaune et al., 2005). The specification of the PPR appears to be a co-developmental event or an immediate consequence of the neural specification, which is primarily determined by the organizer or during neural specification. The mechanism by which PPR differentiation is accomplished remains unknown

Combinatory signals define the NNE border (Koshida et al., 2002; Kudoh et al., 2004; Litsiou et al., 2005; Varga et al., 2011). Recently, transcription factors expressed in the nonneural ectoderm have been shown to participate in Bmp signaling regulation, although they are more involved in ectoderm patterning (Brugmann et al., 2004; Kwon et al., 2010). For example, zebrafish gata3 and foxi1 are expressed throughout the nonneural ectoderm before gastrulation begins, and together with tfap2a and tfap2c, they are activated by Bmp signals and are required to maintain each other's expression in the epidermis (Kwon et al., 2010). Because the early blockage of Bmp signaling (5 hours postfertilization [hpf]) leads to an impaired PPR (Koshida et al., 2002; Kwon et al., 2010), it would be useful to know how these transcription factors are involved in signaling.

The onset of gata3 and foxi1 expression occurs before the shield stage, presumably contributing to the nonneural ectoderm patterning. Zebrafish foxi1 mutants have obvious ear defects; in severe cases, multiple tiny ears are visible adjacent to the hindbrain (Lee et al., 2003; Nissen et al., 2003; Solomon et al., 2003). In addition, Foxi1 regulates dlx3b in the otic-epibranchial placode domain (OEPD), an area that gives rise to both otic and epibranchial fates (in fishes and frogs, the OEPD may include the lateral line fate). Foxi1 has recently been demonstrated to provide a neuronal state of the caudal PPR, supporting the idea that Foxi1 is more involved in the processes preceding otic specification (Hans et al., 2007, 2013; Padanad and Riley, 2011). Although a recent study of the zebrafish gata3 mutant indicated that a loss of otic hair cells is the most obvious defect found in mutant ears (Sheehan-Rooney et al., 2013), the role of gata3 in PPR formation or differentiation remains unknown.

We investigated PPR differentiation in zebrafish using bead implantation, small-molecule chemical treatment, conditional knocking down (photo-cleavable anti-sense morpholino oligos) and overexpression (transgenic) techniques. We have discovered that near the shield stage, Wnt signals may initially set the rostra-caudal polarity of the PPR by inhibiting nonneural ectoderm gata3 expression and that Gata3 is involved in retaining rostral PPR identity. Enhanced Wnt signaling or excessive Foxi1 levels in some way override the rostral PPR identity at approximately 6 hpf, whereas blocking Wnt signaling or depleting Foxi1 function enhances rostral PPR characteristics. PPR differentiation requires the downregulation of Bmp signals and the differential adjustment of Fgf signals, which are partially or fully mediated by Gata3 and Foxi1. Our study provides new insight into NNE patterning during gastrulation.


At the Shield Stage, PPR Cells Differentially Respond to Ectopic Fgf8 Sources

To test whether Fgf8, an important otic induction signal (Phillips et al., 2001; Leger and Brand, 2002; Liu et al., 2003), can induce the first placodal fate of the PPR, we implanted mouse Fgf8b-protein-coated beads into the mapped PPR region (Fig. 1A, red arrow). In this way, the global side effects caused by RNA injection at the one- to two-cell stage and heat treatment-induced Fgf mis-expression in the entire embryo could be avoided. One ectopic ear typically formed around each Fgf8 bead (Fig. 1C–E). When endogenous Fgf3 and Fgf8 levels were low, the ectopic ear still formed, whereas the endogenous ear failed to form (Leger and Brand, 2002; Liu et al., 2003) (Fig. 1D and inset). These results indicated that the Fgf8-loaded beads were sufficient to induce the ear fate in the PPR of fgf8/;fgf3MO embryos (Fig. 1D and inset). This ectopic ear induction did not occur when mouse Fgf3, zebrafish Bmp2a, or bovine serum albumin (BSA) -soaked beads were used (data not shown).

Figure 1.

At the shield stage, the PPR cells have distinct preference to adopt the otic fate along the A-P axis. A: A diagram of animal pole view of a shield stage embryo (6 hpf) (adopted from Streit, 2007; Kozlowski et al., 1997). It illustrates the neural ectoderm (grey shaded area), the preplacodal region (the partially overlapped color-areas between two black lines), which could be presumptively called the PPR and pointed by an arrow, and the belly ectoderm (pink shaded area). The red arrow indicates where Fgf8 soaked beads (circled asterisks) are implanted. Cells of the colored areas give rise to distinct placodal fates (indicated in A). B: An Fgf8 bead is implanted in the presumptive PPR. Arrowhead indicates the shield. The scale bar does not apply. C: An Fgf8 bead implanted into the presumptive PPR (intermediate) induces ectopic ear (blue arrow) as indicated by cldna expression. A black arrow indicates the endogenous ear. D: The endogenous ear fails to form in an fgf8/ mutant injected with fgf3MO (black arrow), but implantation of an Fgf8 bead induces an ectopic ear (blue arrow in the inlet) in the intermediate PPR. E: Ectopic ear (blue arrow) is induced caudally to the endogenous ear (black arrow) by an Fgf8 bead implanted in the posterior PPR. F: Percentage of Fgf8-beads that induce ectopic ears in the anterior (beads anterior to endogenous eyes, aPPR), intermediate (beads between endogenous eyes and ears, iPPR) and posterior (beads caudal to endogenous ears, pPPR) implants. G: An Fgf8 bead implanted just outside the presumptive PPR induces a pair of ectopic ears (blue arrows). One of the ectopic ears is seen fused to the endogenous ear (black arrow). H: An Fgf8 bead implanted in the anterior neural ectoderm fails to induce any cldna expression. The black arrow points to the endogenous otic placode. Unless indicated at the lower left corner of each panel, all are the wild-type embryos. A,B: Animal pole views. C–E and G,H: Side views with anterior to the left. A,B: 6 hpf; C: 36 hpf; D–G: 24 hpf; H: 14.5 hpf. Scale bar = 200 µm in D,E, G,H; 50 µm in C.

After scoring the ectopic ears, we found that Fgf8 induced ectopic ears most efficiently when the beads were implanted in the posterior PPR close to the margin (Fig. 1F) and that the Fgf responsiveness of presumptive PPR cells ranged from high to low (93%, 53%, and 18%) from the margin to the animal pole. We also implanted Fgf8 beads into other ectoderm areas, i.e., in neural and nonneural (belly) ectoderm. When an Fgf8 bead was implanted outside of the presumptive PPR, it induced a pair of ectopic ears (Fig. 1G) in most cases (81%; n = 72), whereas beads implanted in the neural ectoderm failed to induce ectopic ears (Fig. 1H; n = 21).

The induction of ectopic ears demonstrates that the ectoderm at the shield stage exhibits multiple Fgf8 responses, although the belly ectoderm clearly has a greater potential of adopting the otic fate than the rest of ectoderm. In addition, the strength of this response to exogenous Fgf8 varies along the rostral (low)-caudal (high) axis in the mapped PPR, suggesting that the PPR cells may differentially accommodate endogenous signaling.

Wnt Signaling is Involved in PPR Differentiation by Regulating gata3 and foxi1

As wnt3a and wnt8 are already expressed in the margin of shield-stage embryos (zfin.org), we investigated whether the Wnt signal contributes to the observed variation in Fgf responsiveness of the PPR cells (Fig. 1F). XAV939, an Axin protein stabilizer, was used to treat embryos between 5 and 7 hpf to temporarily attenuate the sensitivity of Wnt-responsive cells, including PPR cells (Tian et al., 2013). The 2-hr treatment, spanning the shield stage, induced size changes in PPR-derived organs, i.e., yielding larger olfactory cxcr4b and lens pitx3 domains and a smaller otic stm domain; the epibranchial phox2a domains were unchanged (Fig. 2A,B,D,E,G,H,J,K). The internal blockage of the Wnt signaling cascade in Wnt-responsive cells, including PPR cells, led to a stronger expression of six4.1, a PPR marker, at 11 hpf (Fig. 3A,B). However, the reduced foxi1 domain at 10.5 hpf (Fig. 3D,E), the enhanced gata3 expression at 10.5 hpf (Fig. 3G,H) and the reduced OEPD pax2a expression at 12 hpf (Fig. 3J,K) were unexpected. Considering that the lens and olfactory placodes belong to rostral PPR-derived organs and that the otic placode and part of the epibranchial structures are caudal PPR-derived organs, it appears that inhibiting the Wnt response of the PPR cells enhanced the rostral PPR identity at the expense of suppression of the caudal PPR identity.

Figure 2.

Modulating Wnt signaling around the shield stage alters the development of PPR-derived placodes. A–F: Marker gene expression changes in the rostral PPR-derived placodes of treated embryos. The XAV939 treatment (5–7 hpf) enlarges the olfactory cxcr4b expression (B) and lens pitx3 expression (E), compared with those in control embryos (A,D). The 3F8 treatment (5–7 hpf) leads to no olfactory cxcr4b (C) and lens pitx3 expression (F). G–L: Changes in the caudal PPR-derived placodal marker gene expression in treated embryos. The XAV939 treatment (5–7 hpf) reduces the otic stm (H) or does not change the epibranchial phox2a expression (K), compared with those of controls (G,J). The 3F8 treatment (5–7 hpf) results in larger otic stm (I) and stronger epibranchial phox2a expression (L). Arrows indicate the corresponding organ or organ primordium of each embryo. A–F, J–L: Side views with anterior to the left; G–I: Dorsal views with anterior to the left. A–I: 24 hpf; J–L: 48 hpf. Scale bar = 100 µm.

Figure 3.

Modulation of Wnt signaling affects the PPR polarity and OEPD development also regulates gata3 and foxi1 expression. A–C: XAV939 treated (5–7 hpf) embryos show enhanced yet loosely packed PPR six4.1 expression (B), compared with the control (A). 3F8 (5–7 hpf) treatment enhances six4.1 expression in caudal PPR and a reduction at the anterior most region (C). D–F: foxi1 expression domains are reduced after XAV939 treatment (E), compared with the control (D). 3F8 (5–7 hpf) treatment anteriorly expands foxi1 expression and the bilateral expression domains are fused rostrally (F). G–I: gata3 expressions are enhanced by XAV939 treatment (H) and reduced after 3F8 treatment (I), compared with control (G). J–L: The OEPD but mid-hindbrain boundary (MHB) expression of pax2a is reduced in embryos treated by XAV939 (K), compared with the control (J). 3F8 (5–7 hpf) treatment expands both MHB and OEPD pax2a expression rostra-caudally (L). 12 hpf. Black arrows indicate the OEPD while blue triangles point to the MHB pax2a. M: Real-time RT-PCR reveals that modulating Wnt signaling adjusts foxi1 and gata3 expression levels at 10.5 hpf. Treating embryos with 3F8 at 5–7 hpf increases foxi1 but reduces gata3 expression. On the other hand, the XAV939 treatment of 5–7 hpf enhances gata3 but cuts down foxi1 expression. 10.5 hpf. *P < 0.05; **P < 0.01 ***P < 0.001 (ANOVA test). N: Different from 10.5 hpf, real-time RT-PCR reveals that modulating Wnt signaling adjusts only gata3 expression levels at 8 hpf. Treating embryos with 3F8 at 5–7 hpf reduces gata3 expression. On the other hand, the XAV939 treatment during 5–7 hpf enhances gata3 expression. 8 hpf. **P < 0.01; ***P < 0.001 (ANOVA test). A–L: Dorsal views with anterior to the left. A–C: 11 hpf; D–I: 10.5 hpf; J–L: 12 hpf. Scale bar = 200 µm.

Real-time reverse transcriptase-polymerase chain reaction (RT-PCR) and in situ hybridization assays all showed that in XAV939-treated (5–7 hpf) embryos, gata3 transcript levels were significantly increased at both the 8 and 10.5 hpf stages (Fig. 3G,H,M,N; data not shown), suggesting that the rostral PPR identity may be regulated by Gata3. Indeed, in embryos treated with 3F8, a GSK3 inhibitor (Zhong et al., 2009), for a period of 5–7 hpf, gata3 transcript levels remained low between 8 and 10.5 hpf (Fig. 3G,I,M,N; data not shown). Consequently, the olfactory placode, lens and retina were absent (Fig. 2A,C,D,F), and both an enlarged ear stm expression domain and enhanced phox2a expression in the epibranchial structures were evident (Fig. 2G,I,J,L), although PPR six4.1 expression appeared unchanged (Fig. 3A,C). These results indicate that Gata3 is likely responsible for rostral PPR identity.

In contrast, in the XAV939-treated (5–7 hpf) embryos, foxi1 expression was reduced at 10.5 hpf, whereas in the 3F8-treated embryos, foxi1 expression was increased at 10.5 hpf (Fig. 3M). This pattern suggests that Foxi1 has a different role than Gata3 in PPR differentiation. A real-time RT-PCR assay using total RNA from the entire embryos did not reveal any significant changes in foxi1 expression at 8 hpf (Fig. 3N), but the in situ hybridization assay did show that the expression of foxi1 in the area abutting the neural ectoderm changed in the opposite direction to that of gata3 (data not shown). In 3F8-treated embryos, the largely unchanged PPR six4.1 (Fig. 3A,C), the larger ears and epibranchial structures, and the lack of rostral PPR-derived organs (Fig. 2A,C,D,F,G,I,J,L) suggest that Foxi1 may mediate caudal PPR-derived placodes. This view is supported by the observations that the bilateral foxi1 expression domains were extended and fused at the rostral end (Fig. 3D,F), whereas OEPD pax2a expression was elongated along the anterior–posterior axis but fusion was absent upon 3F8 treatment (Fig. 3J,L). Nevertheless, we speculate that Foxi1 may be more involved in the gain of caudal PPR identity at the end of the gastrula.

Manipulating gata3 Impacts Rostral PPR Development

We then investigated whether manipulating Gata3 levels could alter PPR differentiation. The lens (marked by foxe3) and nose epithelia (cxcr4b) are derived from the rostral PPR. In embryos injected with gata3MO, the lens and olfactory epithelium of all gata3 morphants were smaller than those of control animals (Fig. 4A–D), similar to what we observed in a gata3 mutant (data not shown). Although the gata3 mutant (b1075) is not a null allele, the efficiency and specification of the gata3 morpholino anti-sense oligos has been verified in previous studies (Yang et al., 2010; Sheehan-Rooney et al., 2013).

Figure 4.

Manipulation of Gata3 mainly changes the development of rostral PPR. A–H: In gata3 morphants, reduced olfactory cxcr4b (B) and lens foxe3 (D) are observed compared with controls (A,C). Slightly reduced otic primordial pax2a (F) and epibranchial primordial sox3 (H) expressions are observed compared with their corresponding controls (E,G). I–L: Knocking down gata3 function with an upcMO at 6 hpf, the lens foxe3 (J) is slightly smaller than that of control (I). The OEPD pax2a is essentially unchanged (K,L). Embryos are injected with double-stranded MO at one-cell stage and cleaved by UV light at 6 hpf. M–P: Mis-expression of gata3 by heat shocking a transgenic line Tg(gfp:hse:gata3) at 6 hpf expands the lens foxe3 (N), but not the OEPD pax2a (P), compared with their controls (M,O). Q,R: Expression of six4.1 is reduced when Gata3 function is knocked down using upcMO. Embryos are injected at one-cell stage and treated by UV light at 6 hpf. S,T: Over expression of gata3 by heat shocking a transgenic line Tg(gfp:hse:gata3) at 6 hpf enhances six4.1 expression. U,V: Knocking down Gata3 function reduces the PPR six4.1 expression. W,X: Over expression of gata3 by mRNA injection enhances the PPR six4.1. A–D, I–J, and M–N: Side views with anterior to the left; E–H, K–L, and O–X: Dorsal views with anterior to the left. A–D, I–J, and M–N: 24 hpf; E–F and K–J: 12 hpf; G–H, Q–X: 11 hpf; O–P 12.5 hpf. Scale bar = 200 µm.

Because gata3MO embryos showed slightly reduced otic pax2a and epibranchial sox3 expression (Fig. 4E–H) and the early onset of gata3 in the entire nonneural ectoderm at 4 hpf, it is likely that Gata3 plays a role in PPR development before its function in establishing rostral PPR identity. Using an un-caged photo-cleavable morpholino (upcMO) complementary to gata3MO, both of which were injected as double-stranded MO (gata3MO::gata3upcMO) and ultraviolet (UV) -cleaved to block gata3 mRNA translation at 6 hpf, we found weaker PPR six4.1 expression, similar to that of gata3MO embryos (Fig. 4Q,R,U,V). In addition, knocking down Gata3 function at the 1-cell (gata3MO) and 6 hpf (gata3upcMO) stages both led to a smaller lens (foxe3), although the OEPD pax2a was more affected in the gata3MO embryos (Fig. 4C–F,I–L). Therefore, Gata3 is more involved in rostral PPR differentiation after 6 hpf, whereas before the shield stage, it has a broader function in PPR development.

As enlarged rostral placodes and reduced otic placodes were observed in XAV939-treated embryos at 5–7 hpf, and as the phenotypes were somewhat related to the elevated gata3 (Figs. 2, 3M,N), we expected to observe similar phenotypes when gata3 mRNA was loaded in the embryos. The overexpression of gata3 was achieved by mRNA injection at the 1-cell stage or in hs::gata3(egfp) transgenic embryos whose transgene was activated upon heat shock at 6 hpf. Overall PPR six4.1 expression was enhanced when gata3 mRNA was injected at the 1-cell stage or when gata3 was elevated at 6 hpf (Fig. 4S,T,W,X). However, in embryos with heat-induced transgenic gata3 expression, excess Gata3 slightly increased lens size but not the OEPD (Fig. 4M–P). Therefore, Gata3 may define a ground state of PPR before 6 hpf, and it is continuously required for maintaining the rostral placodal fates. Indeed, the enlarged olfactory cxcr4b and lens pitx3 expression domains due to XAV939 (5–7 hpf) treatment returned to their typical sizes when the double-stranded MO (gata3MO::gata3upcMO) was UV-activated at 6 hpf (Fig. 5A–F). Thus, we conclude that Gata3 works downstream of Wnt signaling to promote the rostral PPR fates after 6 hpf.

Figure 5.

Foxi1 and Gata3 are involved in PPR differentiation regulation after 6 hpf. A–F: Knocking down Gata3 function can restore the enlarged cxcr4b (B,C) and pitx3 (E,F) expression domains caused by the XAV939 (5–7 hpf) treatment, compared with those of control embryos (A,D). Double-stranded MO (gata3MO::gata3upcMO) is injected at the one-cell stage and cleaved by UV light at 6 hpf. G–L: Knocking down Foxi1 function can restore the enlarged pax2a (H,I) and sox3 (K,L) expression domains caused by the 3F8 (5–7 hpf) treatment, compared with those of control embryos (G,J). Double-stranded MO (foxi1MO::foxi1upcMO) is injected at the one-cell stage and cleaved by UV light at 6 hpf. Arrows indicate a corresponding organ or organ primordium of each embryo. A–F: Side views with anterior to the left; G–L: Dorsal views with anterior to the left. A–F: 24 hpf; G–I: 12 hpf; J–L: 11 hpf. Scale bar = 120 µm in A–F; 200 µm in G–L.

Manipulating foxi1 Expression Impacts Caudal PPR Development

Using the same approach as with Gata3, we tested Foxi1 function in PPR differentiation. We found that the caudal PPR-derived ear and epibranchial domains were all reduced in foxi1 morphants (Fig. 6E–H), whereas the lens and olfactory placodes were unaffected (Fig. 6A–D), consistent with the severe caudal reduction in six4.1 in foxi1 morphants (Fig. 6M,N). In addition, in foxi1 morphants, both the otic vesicle and the lateral line primordium were reduced (Fig. 6I,J). We also used a upcMO to knock down foxi1 function at 6 hpf and found that all morphant phenotypes caused by regular MO injection were reproduced (Fig. 6M,N,Q,R). The OEPD dlx3b expression was increased by heat-shock-induced foxi1 at 6 hpf in hs::foxi1(egfp) embryos, similar to that of foxi1-mRNA-injected embryos (Fig. 6O,P,S,T). When foxi1 mRNA was injected at the 1-cell stage, both the otic vesicle and lateral line primordium were enlarged (Fig. 6K,L). These results indicate that Foxi1 likely affects caudal PPR development after 6 hpf. As further evidence, when Foxi1 function was knocked down at 6 hpf in 3F8-treated embryos (5–7 hpf), the enlarged otic pax2a and epibranchial sox3 expression domains had largely returned to their normal sizes (Fig. 5G–L). Thus, Foxi1 is mainly involved in caudal PPR formation after 6 hpf.

Figure 6.

Foxi1 is specifically involved in the OEPD development. A–H: In foxi1 morphants, the olfactory (B) and lens (D) expression are unchanged, compared with their corresponding controls (A,C). The otic primordial pax2a (F) and epibranchial primordial sox3 (H) are severely reduced, compared with their corresponding controls (E, G). I,J: Smaller lateral line placode (blue arrowhead) and otic vesicle (black arrow) develop in foxi1MO embryos (J), compared with the control (I). K,L: The otic vesicles (black arrow) and lateral line placode (blue arrowhead) are enlarged when excess foxi1 mRNA (L) is present, compared with the control (K). M,N: foxi1MO injection results in the compromised caudal PPR six4.1. O,P: Over expression of foxi1 by mRNA injection enhances the OEPD dlx3b. Q,R: Expression of dlx3b is reduced caudally (in OEPD, black arrow) when Foxi1 function is knocked down by an upcMO activated by UV light at 6 hpf. S,T: Over expression of foxi1 by heat shocking a transgenic line Tg(gfp:hse:foxi1) at 6 hpf expands the caudal dlx3b expression (black arrow). A–D: Side views with anterior to the left; E–T: Dorsal views with anterior to the left. A–D and I–J: 24 hpf; E–F:12 hpf; G–H and M–T: 11 hpf; K,L: 26 hpf. Scale bar = 200 µm.

Fgf and Bmp Signaling are Differentially Regulated by Gata3 and Foxi1 in the PPR

To understand how Gata3 and Foxi1 participate in PPR differentiation, particularly with respect to Fgf and Bmp signaling, we analyzed the levels of phosphorylated Erk (p-Erk) and p-Smad1/5/8 in morphants in which either Gata3 or Foxi1 function had been knocked down. We found that Gata3 positively regulated and Foxi1 weakly downregulated Fgf signaling at the 10 hpf stage and that Gata3 negatively regulated Bmp signaling, whereas Foxi1 had an insignificant negative effect (Fig. 7A,B).

Figure 7.

Gata3 and Foxi1 differentially regulate Fgf and Bmp signaling. A: Western blot analysis of the pSmad1/5/8 level of whole embryo (10 hpf). Both foxi1MO and gata3MO embryos show an increased p-Smad level (middle and right lanes), compared with that of control (left). The changes (mean value±error) are normalized against β-Actin level of each lane. B: Western blot analysis of p-Erk level of morphants (10 hpf). The foxi1MO embryos show an increase (middle lane) but gata3MO a decrease of p-Erk (right lane), compared with that of control (left lane). The changes (mean value±error) are normalized against β-Actin level of each lane. C: Real time-PCR tests show that in gata3 morphants, all fgfr gene expressions are down regulated at 10.5 hpf. *P < 0.05; **P < 0.01 (ANOVA test). D: Over-expressing gata3 leads to the up regulated fgfr1 and fgfr2 only. 10.5 hpf. **P < 0.01 (ANOVA test). E: Real time-PCR test shows that in gata3 morphants, only bmp2b is up regulated. 10.5 hpf. ***P < 0.001 (ANOVA test). F: Real time-PCR tests show that in foxi1 morphant, all fgfr expressions are not significantly changed. 10.5 hpf. G: Over-expressing foxi1 slightly up regulates fgfr1. **P < 0.01 (ANOVA test). 10.5 hpf. H: Real time-PCR test shows that Foxi1 regulates Bmp through cv2 and/or dlx3b. 10.5 hpf. **P < 0.01; ***P < 0.001 (ANOVA test).

We also investigated the expression changes of some known Fgf signaling and Bmp signaling pathway components in embryos in which Gata3 or Foxi1 were either elevated or reduced. In gata3 morphants at 10.5 hpf, the expression of fgfr1, fgfr2, fgfr3, and fgfr4 was significantly reduced in both real-time RT-PCR and in situ hybridization assays (Fig. 7C; data not shown). At the end of the gastrula stage, bmp2b expression is normally diminished in the nonneural ectoderm, where bmp4 is not yet detectable at the early segmentation stage. Therefore, the significant increase in bmp2b in gata3 morphants was unexpected (Fig. 7E), as together with other nonneural ectoderm factors, Gata3 is known to be involved in regulating Bmp signaling. In embryos injected with gata3 mRNA, fgfr1 and fgfr2 were significantly up-regulated, whereas bmp2b and bmp4 were unchanged (Fig. 7D,E; data not shown). As the mis-expression of gata3 did not lead to lower bmp2b expression, it remains unclear whether Gata3 directly controls bmp2b expression. In contrast, real-time RT-PCR analysis indicated that Gata3 potentially adjusted the Fgf sensitivity of the PPR cells after 6 hpf, as we observed no change in Fgfrs at the mRNA level until 8 hpf (data not shown). It appears that through the control of Fgf signaling, Gata3 may direct PPR differentiation after 6 hpf.

The foxi1MO embryos did not show any obvious changes in fgfr expression (Fig. 7F; data not shown), but a mis-expression of foxi1 resulted in elevated fgfr1 at 10.5 hpf (Fig. 7G; data not shown). In foxi1 morphants at 12 hpf, fgfr1 and fgfr2 in the OEPD were reduced (data not shown). Although we did not detect any change in bmp2b and bmp4 expression in foxi1 morphants, we found a significantly downregulated expression of cv2, a Bmp antagonist gene (Esterberg and Fritz, 2009; Reichert et al., 2013), in the PPR at 10.5 hpf (Fig. 7H; data not shown) and a subsequently reduced OEPD dlx3b at 12 hpf (Fig. 7H, data not shown). Thus, Foxi1 also regulates PPR differentiation by adjusting Bmp and Fgf signaling.


The present study provides insight into how the anterior NNE border, the PPR, from which individual placodal fate occurs, is formed during gastrulation. First, we showed that at the shield stage, the presumptive PPR is likely in a developmental state, responding to exogenous Fgf8 (Fig. 1A–F) according to the gradient of Wnt signals secreted from the margin. When the internal Wnt signaling cascade of PPR cells is blocked or enhanced near the shield stage (5–7 hpf), the homeostatic state of the PPR is changed rostra-caudally (Fig. 2). In particular, with GSK-inhibitor treatment at 5–7 hpf and low Gata3 levels, the caudal fate dominates in the PPR (Figs. 2, 3). Second, through the conditional manipulation of Gata3 and Foxi1 levels, we found that the expression patterns of PPR markers such as six4.1 and dlx3b do not change significantly in shape or position (Figs. 3A–C, 4Q–X, 6M–T), whereas the PPR-derived placodes or placodal primordium are altered rostra-caudally (Figs. 2, 3J–L, 4A–P, 6A–L). We discovered that Gata3 is more involved in rostral placode development, and we postulate a role of Gata3 in establishing the pan-PPR state. In the differentiating PPR, Gata3 maintains the rostral PPR identity, whereas Foxi1 determines the caudal PPR identity. Finally, we revealed that Gata3 and Foxi1 adjust Bmp and Fgf signaling, which play critical roles in patterning the NNE. In our proposed model (Fig. 8), the differentiation of PPR, triggered by Wnt signals from the margin, begins at the caudal end and undergoes developmental steps that span the entire gastrula stage, establishing the rostra-caudal polarity.

Figure 8.

A model states how Wnt signaling, Gata3 and Foxi1 are differentially involved in the PPR differentiation. The left panel: the Wnt signals, from the margin of the zebrafish shield-staged embryo, promote the differentiation of the presumptive PPR by enhancing foxi1 and the foxi1 positive cells are gradually restricted to the caudal PPR during the gastrulation (6–10 hpf). The Wnt signals slightly inhibit gata3 expression at the beginning of PPR differentiation. The right panel: Gata3 positively regulates Fgf signaling through adjusting fgfr genes and inhibits bmp2b expression. Foxi1 slightly inhibits both Fgf and Bmp signaling until the OEPD is being defined at early segmentation stage (10–12 hpf), when Cv2 inhibits Bmp signaling, while the local OEPD fgfr genes are positively controlled by Foxi1. D, dorsal; V, ventral; NE, the neural ectoderm; EE, the nonneural (epidermis) ectoderm.

All ectodermal (neurogenic) placodes are specified in the PPR (Groves and Bronner-Fraser, 2000; Phillips et al., 2001; Leger and Brand, 2002; Liu et al., 2003; Schlosser and Ahrens, 2004; Bhattacharyya et al., 2004; Mackereth et al., 2005; Martin and Groves, 2006; Bailey and Streit, 2006; Hans et al., 2007; Padanad and Riley, 2011; Chen and Streit, 2013), and the PPR is guided by the Bmp and Fgf signals (gradients), whereby an intermediate level of Bmp favors NC and PPR development (Koshida et al., 2002; Kudoh et al., 2004; Furthauer et al., 2004; Rentzsch et al., 2004; Raible and Ragland, 2005; Delaune et al., 2005; Litsiou et al., 2005; Kwon et al., 2010; Pieper et al., 2012). In Xenopus, neural plate grafts induce certain PPR characters nearby at the belly epidermis (Delaune et al., 2005), highlighting the necessity of neural specification before PPR formation. Because the neural ectoderm, ventral epidermis and PPR respond to implanted Fgf8 differentially in a dorsal (low)-ventral (high) manner (Fig. 1), the PPR likely forms before the shield stage, perhaps coinciding with Fgf-mediated neural induction (Streit et al., 2000; Koshida et al., 2002; Stern, 2005; Delaune et al., 2005) at 4–5.5 hpf. Blocking and stimulating the internal Wnt signaling cascade at 5–7 hpf led to a larger retina (in addition to a larger lens) and the absence of the entire eye field (Fig. 2), respectively, indicating that the abnormal neural patterning/specification may be responsible for the PPR defects. However, PPR six4.1 expression showed only minor changes upon Wnt signaling manipulation (Fig. 3A–C), suggesting that PPR differentiation is more directly linked to Wnt signaling and is not affected by dispensable from the Wnt-caused neural development changes near the shield stage. One study of chicks concluded that the attenuation of Wnt signals is necessary for the neural and nonneural border cells to obtain the PPR fate (Litsiou et al., 2005), although both Wnt and Fgf signals have long been recognized as critical to neural specification. When endogenous Fgf3 and Fgf8 are largely diminished or absent, caudal neural development is severely impaired and the endogenous ear fails to form. However, the ectopic ear is still induced by Fgf8 in fgf8a/;fgf3MO PPR (Fig. 1D) (Maves et al., 2002; Liu et al., 2003), suggesting that the neural defects and PPR development are dispensable, although the integrity of such a PPR requires verification.

In zebrafish, high levels of Bmp/Bmp-regulated genes are present in the nonneural ectoderm during the late blastula and early gastrula stages, presumably to initiate the expression of ventral/epidermis competence factors tfap2a/2c, gata3 and foxi1 (Kwon et al., 2010; Bhat et al., 2013). From the onset of gastrulation onward, the ventral ectoderm is competent to form the PPR (Ahrens and Schlosser, 2005; Delaune et al., 2005; Kwon et al., 2010; Leung et al., 2013). In this study, we focused on Gata3 and Foxi1 to evaluate whether they can provide such competence. Because both regular and upc morpholino oligos lead to similar PPR defects (Fig. 6M,N,Q,R) and as enhancing foxi1 expression at the 1-cell stage and 6 hpf expands the caudal PPR (OEPD) (Fig. 6O,P,S,T), it is likely that Foxi1 functions primarily to define the territory and support the differentiation of the caudal PPR. Molecularly, Foxi1 functions to inhibit Fgf signaling through one or more unknown pathways and suppresses Bmp signaling by regulating cv-2 (Fig. 7). The homeostatic changes in rostral and caudal PPR preference appear to depend on Wnt signaling through Foxi1, whereas the elevation of foxi1 expression is delayed until the end of gastrulation (Figs. 2, 3), suggesting the participation of additional factors that are directly regulated by Wnt signaling and/or other Wnt-activated regulators. The mechanism by which foxi1 is upregulated by the end of gastrula remains unknown.

Knocking down Gata3 function at the 1-cell stage and 6 hpf leads to a smaller lens and/or olfactory placode; however, 3F8-treated embryos lack both the lens and olfactory placode, with an apparent reduction in gata3 expression during gastrulation (Figs. 2A,C,D,F, 3G,I,M,N, 4A–D,I,J). In contrast, XAV939 treatment results in a larger lens and/or olfactory placode, which is due, at least in part, to an increase in gata3 expression during gastrulation, and mis-expression of gata3 at the one-cell stage (data not shown) and 6 hpf slightly increases lens size (Figs. 2A,B,D,E, 3G,H,M,N, 4M,N). These results suggest a role of Gata3 in rostral PPR development. However, knocking down Gata3 at the one-cell stage (Fig. 4E,F) leads to a large reduction in OEPD pax2a expression, whereas doing so at 6 hpf produces no apparent changes (Fig. 4K,L), suggesting a Gata3 function other than maintaining or defining rostral PPR identity before 6 hpf. Considering these results along with the reduced gata3 and enhanced OEPD pax2a expression in 3F8-treated embryos (Fig. 3J,L–N), it appears likely that the enlarged otic vesicles and enhanced epibranchial phox2a expression (Fig. 2G,I,J,L) are due to elevated Foxi1 levels (Fig. 3D,F,M) in the caudal PPR by the end of gastrulation. We can exclude the possibility that Gata3 regulates foxi1 during gastrulation, and we find an overall reduction of PPR gene expression when both Gata3 and Foxi1 functions are knocked down (unpublished data), thus indicating that both transcription factors are required for normal PPR development. Spatially, although both genes are expressed throughout the entire nonneural ectoderm, a late attenuation of gata3 is evident when foxi1 expression is restricted to two stripes adjacent to the developing hindbrain to mark the OEPD (zfin.org). Therefore, either the rostral PPR retains a relatively undifferentiated state by the end of gastrula (when gata3 expression persists throughout the entire nonneural ectoderm) or rostral PPR determination occurs later than caudal PPR determination (the olfactory, lens and pituitary placodes become morphologically visible at much later stages than the otic and epibranchial placodes). Alternatively, Gata3 may function predominantly in shaping the PPR before the shield stage, retaining the rostral PPR identity after the shield stage, or it may exert both functions throughout gastrulation. The stronger role of Gata3 in promoting Fgf signaling and inhibiting Bmp signaling than that of Foxi1 (Fig. 7) and the large changes in placodal development that result when gata3 expression is up- or downregulated (Figs. 2, 3G–I) support the ideas that Gata3 may serve to provide an initial state of PPR and that this state is primarily of rostral identity. In Gata3 knockout mouse embryos, lens development (and possibly other development) is impaired (Maeda et al., 2009), and in chick embryos, PPR cells have been shown to initially share a ground state with the lens (Bailey et al., 2006). However, the mechanism by which gata3 is regulated during zebrafish gastrula remains unknown.

In summary, after the shield stage, the PPR enters the differentiation process, the onset of which is presumably triggered by Wnt signaling originating from the margin and Gata3 (Fig. 8). To fully differentiate, the PPR cells obtain the caudal characteristics upon caudally enhanced foxi1 expression; otherwise, Gata3 demarcates the rostral PPR. Mechanistically, the PPR cells express gata3 and foxi1 to inhibit Bmp signaling, and Gata3 promotes, whereas Foxi1 modestly inhibits, Fgf signaling, allowing the individual placode to progressively develop (Fig. 8).

Experimental Procedures

Fish Care and Treatments

Wild-type Tubingen strain and all transgenic fish lines were raised under standard conditions. All animal experiments were approved by the Institutional Animal Care and Use Committee (IACUC) of Peking University. The reference from IACUC of Peking University is LSC-LiuD-01. Embryos were produced using standard procedures and staged by hours post fertilization (hpf) or by morphological criteria (Kimmel et al., 1995). The wild-type TU line was used. The mutant and transgenic lines have been described previously: acerebellarti282a, a strong hypomorphic allele of fgf8a mutant; foxi1em1 (from A. Fritz) and foxi1hi3747tg (from N. Hopkins) are two presumptive null alleles of foxi1; and created Tg(gata3::CMV-8xhse-CMV::egfp) and Tg(foxi1::CMV-8xhse-CMV::egfp), two transgenic lines in which the gene of interest (gata3 and foxi1) and egfp are driven by two identical CMV promoters that share eight heat response elements (8xhse). We refer to homozygous mutants as foxi1 and transgenic lines as hs::gata3(eGFP) and hs::foxi1(eGFP), respectively.

Tg(gata3:CMV-8×hse-CMV:egfp) or Tg(foxi1:CMV-8×hse-CMV:egfp)

To obtain the constructs, the KOD-PLUS PCR was performed using primer pairs for gata3 (5′-TATGAATTCAACGGACGGACTTTGTAAAAAC-3′, the underlined is an EcoRI site; 5′-GACTAGTTTCTTGGCGTGGCATTCTTGTAT-3′, the underlined is a SpeI site) or foxi1 (5′-TATGAATTCGAAACTCCATGTTTCTGGA-3′, the underlined is an EcoRI site; 5′-GACTAGTATGGACCAGCTGTAGATCA-3′, the underlined is a SpeI site) in a volume of 50 μl (95°C 10 min; 95°C, 30 s/56°C, 30 s/72°C, 90 s for 30 cycles; 72°C, 10 min; 16°C, 5 min). The right PCR product was chosen after sequencing. After sequencing confirmation, the PCR fragment and vector pSGH2, which contains a gfp coding sequence on one side of the dual CMV promoters (Aghaallaei et al., 2007), were restricted by EcoRI and SpeI, and ligated.

A total of 50 ng plasmid DNA was digested in a volume of 8 μl (plasmid 2 μl, water 4.8 μl, 10 × buffer 0.4 μl, I-SecI (NEB) 0.8 μl) at room temperature for 20 min and dialyzed against 0.5 × TE pH7.5 for 10 min. The dialyzed restriction solution was then used in microinjection. A total of 12.5 pg of digested plasmid (with I-SecI) was injected into one-cell stage embryos. The raised fish (F0) was out-crossed with wild-type fish to collect F1 embryos subjected to heat shock treatment starting at 14 hpf, in 38–39°C water bath for 30 min, with gentle inversion or shaking 2–3 times. If green fluorescent protein (GFP) -positive embryos were observed 6–8 hr after the heat shock, their parent(s) was kept for future experiments and the passage of the transgene to the next generation.

Small Chemical Treatment

The GSK3-β inhibitor 3F8 (TOCRIS) and tankyrase (TNKS) inhibitor XAV939 (TOCRIS) were initially dissolved in DMSO, and the working solution were diluted from the stock by embryo medium E2. A total of 20 μM 3F8 and 35 μM XAV939 were used to treat embryos, respectively.

Probes and mRNAs

Probe synthesis and single or double color in situ hybridization was performed essentially as previously described (Jowett and Yan, 1996; Phillips et al., 2001). The probe information was described in Table1. The cDNA of foxi1 was restricted by EcoRI and SpeI and gata3 was restricted by EcoRI and KpnI, and sub-cloned into pXT7 vector to synthesize stable foxi1 and gata3 mRNA. In vitro mRNA synthesis was performed using an RNA synthesis kit (Ambion). We purified synthesized mRNA and in situ probes with RNeasy mini columns (Qiagen), and inject mRNAs, foxi1 (10–100 ng/μl), and gata3 (150 ng/μl) into blastomeres of embryos at the 1 or 2-cell stage.

Table 1. Primers Used for In Situ Hybridization Probes
GeneForward primerReverse primer


All MOs were purchased from Gene Tools, LLC (Philomath, OR), and their sequences are listed in Table 2. The use of these morpholino oligos, such as fgf3MO, foxi1MO and gata3MO and their specificity and efficiency were described elsewhere (Solomon et al., 2003; Yang et al., 2010; Sheehan-Rooney et al., 2013). Microinjection was performed at the one- to four-cell stage. In our trial experiments, we also co-injected with the zebrafish p53MO oligo, and did not find any difference from using fgf3MO, foxi1MO and gata3MO alone. We always used the un-injected embryos as our negative controls. In our hands, the foxi1MO injected embryos and larvae showed the otic phenotypes identical to that of foxi1 mutants such as foo(hi3747) and hsy (Nissen et al., 2003; Solomon et al., 2003).

Table 2. Morpholino Oligos Used in This Study
GeneMO sequence

We used sense uncaged photo cleavable MO to complement the antisense MO for a particular gene knock down assay. To block the regular antisense MO from functioning, both strand MOs were mixed by a ratio of antisense MO: sense MO = 1:1.5 and annealed in water: denatured at 95°C for 5 min, quick chilled to −1°C for 30, and stored at 4°C. We injected one or two cell-stage embryos, and the injected embryos were exposed to UV light (365 nm, Gene Tools Inc.) for 400 s at a chosen stage. To test the efficiency, we monitored if the ears became obviously smaller than normal (foxi1) or smaller lens (gata3). The Morpholino oligos used in this study are listed in Table 2.

Bead Implantation

Bead experiments were performed similarly to previously published reports with a few modifications (Reifers et al., 2000). Briefly, 45 μm polystyrene beads (Polysciences) were rinsed in phosphate buffered saline (PBS), treated with 0.5 mg/ml heparin for 20 min at room temperature, then incubated in 100–250 μg/ml mouse FGF8b (R&D Systems), which was reconstituted in 0.5% BSA in PBS, for 2 hr at room temperature. Beads were ready to use after an additional 2 rinses in PBS. Beads were similarly coated with Bmp2a or Fgf3 proteins. Control beads were incubated in 0.5% BSA in PBS only. Shield stage embryos were dechorionated and mounted with the animal pole up in 2% methylcellulose buffered by Hepes solution (pH 7.2). With a sharp glass needle, a small cut was made near the target area, based on the 50% epiboly fate map. Using a slightly blunt (broken by touch) glass needle, a coated bead was brought close to the implantation site and pushed into the cut. Implanted embryos remained in mounting cellulose with Ringer's solution for at least 1 hour and were then transferred into standard embryo medium. All solutions were filter sterilized and contained antibiotics. To test the efficacy of FGF8 beads, embryos with implanted beads at the shield stage were labeled for erm or gbx2 expression. In all experiments (n = 31), embryos showed expression of either gene around the FGF8b beads, whereas no embryos with BSA or Bmp2a beads showed erm or gbx2 expression.

Total RNA Extraction and Reverse Transcription

Total RNA were extracted by TRIzol (Invitrogen) and transcribed into the first-strand cDNA by M-MLV reverse transcriptase (Invitrogen). For the details: embryos (20 embryos, 10 hpf) were collected and mixed with 500 μl of TRIzol (Invitrogen), until the embryos were dissolved, after occasional pipetting and standing for 5 min at room temperature. The TRIzol/embryos solution was then mixed with 100μl of chloroform, vortexed for 15 s, and settled for 3min at room temperature. The resultant solution was centrifuged at 4°C, 13,000 rpm for 15 min, and the supernatant was sucked up and transferred to a new tube. To precipitate the RNA, 250 μl isopropyl alcohol was added, followed by shortly vortex, room temperature settlement of 10 min and centrifugation at 13,000 rpm for 10 min. We carefully discarded the supernatant, added 75% alcohol to wash 4°C, and 7,500 rpm centrifuged the tube at 7,500 rpm for 5 min. Dry and dissolve RNA by standard protocols. After removing the co-precipitated genomic DNA by DNase I digestion, the RNA was purified using RNeasy (Qiagen). To set up the reaction, 1 μl of Oligo (dT18), 1 μl of dNTPs, treated RNA solution and nuclease-free water were mixed to a total volume of 12 μl, and resultant solution was heat treated at 65°C for 5 min, chilled on ice for at least 1 min. A total of 4 μl of 5 × first-strand synthesis buffer, 2 μl of 0.1M DTT and 1 μl of RNaseOUT were added, mixed gently, and incubated at 37°C for 2 min. A total of 1 μl of M-MLV reverse transcriptase was then added for additional incubation at 37°C for 50 min, the reaction was stopped by heat treatment at 70°C for 15 min and stored at −20°C.

Real-Time PCR (Invitrogen Platinum SYBR Green SuperMix-UDG)

Each gene-specific first-strand cDNA sample was appropriately diluted into a series of concentrations and PCR was performed using specific primers based on a one-fold of template dilution. For β-actin (a constitutively expressed gene to serve as a control), template dilution was 1/8. In a reaction volume of 25 μl, 12.5 μl of qPCR mix (Invitrogen Platinum SYBR Green SuperMix-UDG) was added. Each experiment was repeated three times. PCR reactions were performed in 96-well plates (Thermo Scientific, AB-1100) with optical adhesive covers (Applied Biosystems, P/N: 4360954), using an Applied Biosystems 7500 real time PCR system and following the manual. We used the 2−ΔΔCT method to calculate changes of gene expression. Primers used for real time PCR reactions are listed in Table 3.

Table 3. The primers used in real time-PCR reactions
GeneForward primerReverse primer

Protein Extraction and Western Blotting

We disconnected the yolk and embryo axis through mechanical force, and removed the yolk by low-speed centrifuging. Through gel electrophoresis (10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis), protein bands were separated and transferred overnight to polyvinylidene fluoride (PVDF) membrane(s) (Millipore) at 100mA. The PVDF membrane(s) was blocked for 1 hr in the TBST buffer (Tris buffered saline with 0.1% Tween-20) containing 5% BSA, and incubated with primary antibodies overnight at 4oC with gentle agitation. Membrane(s) was rinsed several times in TBST before adding the secondary antibodies for 2 hr at room temperature in the dark. Membrane(s) was washed in TBST intensively and TBS before imaging on a LiCor infrared scanner.

Protein band intensities were measured using Odyssey 3.0 software. The primary antibodies used in this study include: rabbit anti-p-Smad1/5(S463/465)/Smad8 (S426/428) (CST, 1:1,000), mouse IgG anti-Actin (CST, 1:2,000), mouse anti-p-Erk (CST, 1:2,000). Secondary antibodies were conjugated to IRDye680 or IRDye800CW (ODYSSEY), with a 1:5,000 dilution.


All the experiments in this study were repeated for three times. Statistical data was analyzed using one-way ANOVA and P-values were generated using GraphPad Prism version 5 (GraphPad Software).


We are grateful to the critical comments from and discussion with the past and current members of Monte Westerfield lab and members of Liu lab for the period of current study, and good suggestion from Drs. J. Peng and J. Chen. We thank J. Wegner, J. Pierce, and Y. Zhuang for their technical assistance. This study has been supported by the National Basic Research Program of the Chinese Ministry of Science and Technology through a 973 Grant and was funded by a National Science Foundation of China grant, and was partially supported by Peking-Tsinghua Center for Life Sciences (to D.L.). Y.W. is a recipient of the PKU President Graduate Scholarship.