SEARCH

SEARCH BY CITATION

Keywords:

  • Churchill;
  • Sip1;
  • FGF;
  • somitogenesis;
  • zebrafish

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Cell-type specific regulation of a small number of growth factor signal transduction pathways generates diverse developmental outcomes. The zinc finger protein Churchill (ChCh) is a key effector of fibroblast growth factor (FGF) signaling during gastrulation. ChCh is largely thought to act by inducing expression of the multifunctional Sip1 (Smad Interacting Protein 1). We investigated the function of ChCh and Sip1a during zebrafish somitogenesis. Knockdown of ChCh or Sip1a results in misshapen somites that are short and narrow. As in wild-type embryos, cycling gene expression occurs in the developing somites in ChCh and Sip1a compromised embryos, but expression of her1 and her7 is maintained in formed somites. In addition, tail bud fgf8 expression is expanded anteriorly in these embryos. Finally, we found that blocking FGF8 restores somite morphology in ChCh and Sip1a compromised embryos. These results demonstrate a novel role for ChCh and Sip1a in repression of FGF activity. Developmental Dynamics 239:548–558, 2010. © 2009 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Fibroblast growth factors (FGFs) play essential roles in cell growth and differentiation in many developmental contexts (Amaya et al., 1991; Sutherland et al., 1996; Borland et al., 2001; Coumoul and Deng, 2003; Furthauer et al., 2004; Bottcher and Niehrs, 2005; Thisse and Thisse, 2005; Huang et al., 2007; Krens et al., 2008). During vertebrate embryogenesis, FGFs are required for induction of both mesoderm and neural ectoderm, patterning of the midbrain, posteriorization of the neural plate and segmentation of the mesoderm. In vertebrates, the FGF family contains over 20 ligands that interact with four receptors. The mechanisms that account for the diverse responses evoked by FGFs have yet to be fully elucidated, but depend on cell-type specific modulators of signaling.

One such FGF effector is the zinc finger protein Churchill (ChCh), which regulates FGF activity during gastrulation. ChCh is slowly induced in response to FGF and acts as a switch between mesoderm and neural inducing activities of FGF in chick, Xenopus and zebrafish (Sheng et al., 2003; Snir et al., 2006; Londin et al., 2007a). ChCh inhibits expression of mesodermal markers including brachyury, Tbx6L, and spt and blocks mesendoderm induction, which requires FGF signaling in cooperation with Activin/Nodal activity (Kimelman and Maas, 1992; LaBonne and Whitman, 1994). ChCh also regulates cell movements during gastrulation. In the chick, Chch blocks ingression of presumptive ectoderm through the primitive streak at the end of gastrulation. Therefore, these cells adopt neural fates instead of ingressing through the streak and becoming paraxial mesoderm (Sheng et al., 2003). Similarly, when zebrafish blastomeres with compromised ChCh activity are transplanted to the animal pole of wild-type hosts, they leave the epiblast, migrate to the germ ring, and acquire mesodermal fates (Londin et al., 2007b).

ChCh was initially thought to regulate transcription of target genes by means of a direct DNA interaction. However, subsequent biophysical characterization of ChCh suggested that it may not be a DNA binding protein (Lee et al., 2007). Nonetheless, by direct or indirect mechanisms, ChCh induces sip1 (Smad Interacting Protein 1) transcription (Sheng et al., 2003; Snir et al., 2006; Londin et al., 2007a), which is key to the activity of ChCh. Sip1 is a multifunctional molecule that modulates transforming growth factor-beta (TGF-β) signaling by converting activated forms of both Smad1/5 and Smad2/3 to transcriptional repressors (Remacle et al., 1999; Verschueren et al., 1999; Postigo, 2003; Postigo et al., 2003). In addition, Sip1 regulates cell movements by directly repressing E-cadherin transcription (Remacle et al., 1999; Comijn et al., 2001) and mesoderm induction by directly impeding Xbra transcription (Remacle et al., 1999; Verschueren et al., 1999).

In zebrafish, chch is expressed after gastrulation (Londin et al., 2007a), but later roles for ChCh have not been studied. We now present data that ChCh and Sip1 are required for somitogenesis. Segmentation is an essential step in formation of the vertebrate body axis. Mesodermal segmentation is established by sequentially dividing the unsegmented presomitic mesoderm into the bilaterally segmented structures called somites. The “clock and wavefront” model describes the mechanisms of regulation of somitogenesis (Cooke, 1978; Palmeirim et al., 1997; Forsberg et al., 1998; McGrew et al., 1998; Jiang et al., 2000; Bessho et al., 2001; Saga and Takeda, 2001). In zebrafish, the “clock and wavefront” model proposes that function of a molecular oscillator in the presomitic mesoderm (PSM) results in controlled expression of a set of genes associated with the Notch pathway (Holley et al., 2000; Jiang et al., 2000; Pourquie, 2001, 2003). Simultaneously, the wavefront modulates the ability of the PSM to respond to the morphogenic signals and produce segments (Cooke and Zeeman, 1976; Dale and Pourquie, 2000; Giudicelli and Lewis, 2004). The pulses generated by the molecular clock are translated into very highly regulated spatial periodicity. Despite the similarities in somitogenesis between species, there are differences between amniotes and mammals in the segmentation program (Giudicelli and Lewis, 2004; Cinquin, 2007). For example, lunatic fringe and delta genes appear to function differently in mouse, chick, and zebrafish (Prince et al., 2001; Sato et al., 2002; Dale et al., 2003; Serth et al., 2003; Qiu et al., 2004). In addition, the Wnt pathway is important for regulation of segmentation clock in mice (Takada et al., 1994; Aulehla et al., 2003), but not zebrafish.

Previous studies revealed that FGF signaling at the PSM regulates the wavefront position during somitogenesis. FGF modulates somite size by impeding maturation of PSM (Dubrulle et al., 2001; Sawada et al., 2001; Delfini et al., 2005; Wahl et al., 2007). Blocking FGF signaling increases somite length, while activating signaling has the opposite effect (Sawada et al., 2001). In addition, activation of FGF signaling blocks convergence movements and extends somite width (Furthauer et al., 1997; Krens et al., 2008). It is important to define regulators of FGF signaling in the PSM to determine how precise positional cues within the PSM are generated.

Here, we demonstrate that ChCh and Sip1a modulate FGF signaling in the PSM during somitogenesis. Surprisingly, we find that ChCh and Sip1 repress FGF expression. During somitogenesis, knockdown of chch or sip1a results in somites that are less extended through the anterior–posterior (A/P) axis while they are overextended through the mediolateral axis. We also found that inhibition of ChCh and Sip1a perturbs oscillating gene expression in the forming somites. In ChCh or Sip1 compromised embryos, fgf8 expression in the paraxial mesoderm is expanded rostrally leading to altered somite size. Manipulations that reduced FGF8 signaling in ChCh and Sip1a compromised embryos restored somite size. This demonstrated that ChCh and Sip1a regulate somite morphogenesis by limiting FGF signaling. Together, these findings establish a novel role for ChCh as a repressor of FGF signaling.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Role of ChCh and Sip1 in somitogenesis

In a previous study, we observed that ChCh is required for proper somite formation (Londin et al., 2007b). However, the mechanism of action of ChCh in somitogenesis is unknown. To investigate the function of ChCh in zebrafish somitogenesis, we inhibited ChCh activity using two translation blocking morpholinos (chch ATG MO and MO2) (Londin et al., 2007b). Microinjection of chch ATG MO presents a similar, but stronger somite phenotype than that produced by chch ATG MO2 microinjection. chch morpholino injected embryos are morphologically indistinguishable from control morpholino injected siblings until 75–95% epiboly stage (Fig. 1B,C). Dorsal views of the 12-somite stage chch morphants reveals that somites are less extended through the anterior–posterior axis while they are overextended through the mediolateral axis (Fig. 1K–M; 52%; n = 128). Moreover, these embryos have a shorter and wider body axis (Fig. 1G,L; H,M). At 24 hr post fertilization (hpf), somites in ChCh compromised embryos are enlarged and lack their characteristic chevron shape (Fig. 1Q). Furthermore, although roughly 30 pairs of somites are formed in wild-type siblings, chch morphants only form 22–26 somites (Fig. 1Q and data not shown). Taken together, these results demonstrate that ChCh is required during somitogenesis.

thumbnail image

Figure 1. ChCh and Sip1a are required for somitogenesis. A–S: Living wild-type, and ChCh and Sip1a compromised embryos. A–E: Early epiboly movements appear normal in chch or sip1a morphants. F–J: By 12 somites, misshapen somites are apparent in both chch and sip1 morpholino treated embryos. K–O: In these embryos, somites are less extended through the anterior–posterior axis and overextended through the mediolateral axis. Horizontal and vertical red dotted lines span the width and length of the first four wild-type somites for comparison to the first four somites of morphant embryos, which are marked with a black line. P–S: At 24 hours post fertilization (hpf), somites are enlarged and misshapen in both ChCh and Sip1a compromised embryos. Arrowheads denote notochord and black arrows denote somites. A–J,P–S are lateral views; K–O, dorsal views.

Download figure to PowerPoint

Because ChCh activity is mediated by Sip1 during gastrulation (Sheng et al., 2003) and homozygous Sip1 knockout mice have a somite phenotype similar to chch morphants (Maruhashi et al., 2005), we asked whether Sip1 inhibition produced similar somite defects in zebrafish. Zebrafish have two sip1 genes (sip1a and sip1b; Delalande et al., 2008). To investigate the function of Sip1 in zebrafish somitogenesis, we inhibited Sip1 activity using previously characterized sip1a and sip1b morpholinos (Delalande et al., 2008). The sip1a ATG MO injected embryos have a somite phenotype similar to chch MO injected embryos, but the overall phenotype is more severe (Fig. 1I,N (89%; n =112) and data not shown). As in chch morphants, the somites are shorter through the A/P axis and overextended through the mediolateral axis (Fig. 1N). Microinjection of sip1a splice morpholino also produced embryos that were short in the A/P axis and elongated in the mediolateral axis (Fig. 1J,O). Similar to chch morphants, somites are enlarged and lost their characteristic chevron shape at 24 hours postfertilization (hpf; Fig. 1R,S). Because this morpholino alters mRNA structure, we were able to monitor the effectiveness of the knockdown on eliminating wild-type mRNA (Supp. Fig. S1, which is available online). Until dome stage, only wild-type mRNA is detected, which is presumably maternal message that is unaltered by the splice morpholino. Beginning at dome stage (4.3 hpf), the smaller misspliced product is detected. By 75% epiboly (8 hpf), no wild-type mRNA is apparent.

An alternatively spliced form of sip1a that lacks a portion of exon 8 has been described (Delalande et al., 2008). The alternative splice form lacks one zinc finger that is present in the longer form (Supp. Fig. S2, blue bar, Supp. Fig. S3A, dark blue box). We identified a similar alternatively spliced form of Sip1b that contains a deletion of 66 bp of exon 8 (Supp. Fig. S2). Because both sip1a and sip1b contained forms that lack the same region, we sought to determine whether either form of sip1a was required for somite formation. Both sip1a forms were detected by reverse transcriptase-polymerase chain reaction (RT-PCR) during a series of stages spanning the first 24 hpf (data not shown). We designed morpholinos to target the exon 8/intron 8 boundary (sip1a splice MO2, Supp. Fig. S3A, pink bar) to drive production of the shorter form and a morpholino (sip1a splice MO3, Supp. Fig. S3A, orange bar) that targets the alternative splice site in exon 8 and blocks production of the shorter form (leaving only the long form). Analysis of cDNA from embryo injected with each morpholino revealed that sip1a splice MO2 efficiently altered splicing so that only the shorter form was produced, while sip1a splice MO3 eliminated the shorter form (Supp. Fig. S3B,C). Somitogenesis was not altered by microinjection of either morpholino (data not shown). Due to toxic effects of co-injecting high doses of the two morpholinos (10 ng each), we were unable to analyze the effects of blocking production of both the long and short Sip1a forms in the same embryos. However, these experiments suggest that each Sip1a form is likely sufficient for normal somite development.

In a previous report, sip1b morphants were reported to have severe defects and produce only a few somites (Delalande et al., 2008). We observed a more severe phenotype than Delalande et al. when we microinjected low doses of sip1b MO (data not shown). Therefore, we were unable to address the role of sip1b in somite formation using morpholino approaches.

ChCh and Sip1a Are Required for Pattering of Presomitic Mesoderm

During somitogenesis, a series of highly regulated morphogenetic processes produce periodic and symmetrically formed somite boundaries. To characterize the origin of the severe somite defects in ChCh and Sip1a-compromised embryos, we performed a series of RNA in situ hybridizations with 14-somite stage chch and sip1a MO injected embryos. Because somites are derived from paraxial mesoderm cells, we analyzed the state of the presomitic mesoderm cells in ChCh and Sip1a compromised embryos. papC encodes a structural transmembrane protein and regulates the mesodermal segmentation during zebrafish development. In wild-type control embryos, four papC expression stripes corresponds to the (prospective) somites at the segmentation plate (Yamamoto et al., 1998; Fig. 2A). However, in chch and sip1a morphants, the number of stripes ranges from five to eight (Fig. 2B,C). Moreover, tail bud expression domain of papC in ChCh and Sip1a compromised embryos is much broader mediolaterally than in wild-type siblings (Fig. 2A–C).

thumbnail image

Figure 2. Patterning of the presomitic and somitic mesoderm is disrupted in ChCh and Sip1a compromised embryos. A–L: Whole-mount (A–C,G–I) and flat-mount (D–F,J–L) RNA in situ hybridization of somite markers in wild-type, ChCh and Sip1a compromised embryos. All views are dorsal; anterior to the top (A–C,G–I) and anterior to the left (D–F,J–L). A–C: The expression domains of the PSM marker, papC is broader mediolaterally in ChCh and Sip1a compromised embryos than the wild-type siblings. The number of the papC expression stripes corresponding to the prospective and formed somites at the segmentation plate in chch and sip1a MO is higher than in wild-type siblings (compare asterisks number). D–F: The myogenic regulatory factor, myoD, is expressed in the posterior somite compartment in chch and sip1a MO injected embryos as in wild-type embryos. G–L: ephB2 and dld expression at the anterior half of the somites is reduced and diffuse. Asterisks denote (prospective) somites.

Download figure to PowerPoint

Rostrocaudal polarity of the somites is also disrupted in ChCh and Sip1a compromised embryos. The segmental expression of myogenic regulatory factor myoD in the posterior half of the somites (Weinberg et al., 1996) is extended in the mediolateral axis (Fig. 2D–F). On the other hand, expression of ephB2 (Durbin et al., 1998; Fig. 2G–I), dld (Jiang et al., 2000; Holley et al., 2002; Fig. 2J–L), and fgf8 (Fig. 3G–I) in the anterior half of the somites is reduced and expression is no longer restricted to the anterior region of the somites in both chch and sip1a morphants. Therefore, we conclude that rostrocaudal somite polarity requires ChCh and Sip1.

thumbnail image

Figure 3. Inhibition of ChCh and Sip1a affects the components of the “clock and wavefront model.” A–I: Flat-mount RNA in situ hybridization of 10-somite stage wild-type, and ChCh and Sip1a compromised embryos. All views are dorsal; anterior to the left. A–F: In wild-type embryos, periodic activation of Notch signaling provides cycling gene expression of Notch pathway genes such as her1 and her7. The number of the her7 expressing stripes in chch and sip1a ATG MO injected embryos ranges from 4 to 5 (B,C; E,F). G–I: fgf8 expression domain at tail bud is expanded anteriorly in ChCh and Sip1a compromised embryos. A–F: Asterisks denote each her1 or her7 expressing premesoderm stripe. G–I: Arrows denote fgf8 tail bud expression domain (G–I).

Download figure to PowerPoint

Inhibition of ChCh and Sip1a Alters Periodic Gene Expression

The “clock and wavefront” model describes the timing and positioning of somite boundaries during segmentation of the mesoderm (Cooke, 1978; Palmeirim et al., 1997; Forsberg et al., 1998; McGrew et al., 1998; Jiang et al., 2000; Bessho et al., 2001). In this model, somite size is the synchronous function of frequency of “molecular clock” oscillations and of the pace of “wavefront” progression. In chch and sip1a morphants, formed somites are smaller through the A/P axis. Either a faster ticking “molecular clock” or slowed down “wavefront” progression during somitogenesis can trigger reduction in somite size.

To test whether cyclic gene expression is altered in ChCh and Sip1a compromised embryos, we assayed her1 and her7 expression by RNA in situ hybridization. her1 and her7 are both the output of the “molecular clock” and have characteristic one- to two-stripe expression domain in the PSM at the 10-somite stage (Fig. 3A,D; Oates and Ho, 2002). However, the number of stripes observed in the ChCh and Sip1a compromised embryos ranged from three to five (Fig. 3A–C; D–F, marked with asterisks) indicating that impeding ChCh or Sip1a averts the proper termination of cyclic her1 and her7 expression in the anterior PSM. However, the performance and pace of the “molecular clock” is not substantially altered because, despite the size difference, duration of somite formation in ChCh and Sip1a compromised embryos is comparable to control morpholino injected siblings (Supp. Fig S4, and data not shown).

Failure to properly terminate her1 and her7 cyclic expression in the anterior PSM can be the result of slowed wavefront progression because the “wavefront” facilitates the transition of the PSM cells from the immature state to mature state by arresting the oscillating her1 and her7 wave. If the pace of the wavefront progression is slower, the overall rate of maturation of the PSM and therefore arresting the expression of cyclic her1 and her7 expression would also be slowed.

FGF signaling at the PSM is required for the regulation of the position of the wavefront during somitogenesis (Dubrulle et al., 2001; Sawada et al., 2001; Delfini et al., 2005; Wahl et al., 2007). A threshold level of FGF signaling facilitates the transition of the PSM cells from the immature state to mature state (Dubrulle et al., 2001; Sawada et al., 2001). Ectopic activation of the FGF signaling in zebrafish by surgically inserting FGF8-soaked beads in one side of the PSM gives rise to narrower somites in the region anterior to the FGF bead, while blocking FGF signaling has the opposite effect. On the FGF8-soaked bead implanted side, her1 expression domain also extends more anteriorly than the control side (Sawada et al., 2001).

To determine whether the somite alterations in chch or sip1 morphants are associated with altered fgf8 gene expression, we assayed fgf8 expression in these embryos by RNA in situ hybridization. In both chch and sip1a ATG morpholino injected embryos, fgf8 expression domain is expanded rostrally and extends into the midline and sometimes the (putative) somitic mesoderm (Fig. 3G–I). This suggests that the progression of the wavefront is much slower in ChCh and Sip1a compromised embryos. Slowing the wavefront is predicted to result in a broader zone of immature PSM. A similar phenomenon was also observed in Sip1a knockout mice (Maruhashi et al., 2005).

Rostral expansion of fgf8 expression domain at tail bud can be explained by local changes in cell number in the PSM. We tested whether rate of cell proliferation is also altered in ChCh and Sip1a compromised embryos using anti-phosphorlyated histone H3 antibody as a cell proliferation marker. There is no significant difference between ChCh or Sip1a compromised embryos and their wild-type siblings (n = 29, 31, and 19, respectively, data not shown). This indicates that rostral expansion of fgf8 expression domain is not due to altered cell proliferation.

Effects of ChCh Knockdown on Somite Morphology Are Due to Expanded FGF8 Activity

In ChCh or Sip1a compromised embryos, fgf8 expression in the paraxial mesoderm is expanded rostrally and somites are narrower at the A/P axis and wider in the mediolateral axis. We hypothesized that the expansion in FGF8 expression in these embryos caused alterations in somite morphology. To test that hypothesis, we determined the effects of chch and sip1a knockdown on embryos with compromised FGF signaling. If rostral expansion of fgf8 is the cause of the somite phenotype observed in ChCh or Sip1a compromised embryos, then reduction of FGF8 activity will restore wild-type somite morphology. We used two means to attenuate FGF activity, microinjection sprouty4 (spry4) mRNA and acerebellar (fgf8/ace) mutant embryos.

spry4 is a feedback induced antagonist of FGF signaling. FGF8 induces spry4 expression which in turn inhibits FGF activity (Furthauer et al., 2001). Microinjection of spry4 sense RNA weakly impedes FGF8 signaling in zebrafish (Furthauer et al., 2001). Injection of spry4 sense RNA into chch ATG morphants reduced the number of embryos with narrowed somites from 58% (ChCh ATG MO+lacZ mRNA) to 45% (chch ATG MO+spry4 RNA) (P = 0.0063; Supp. Fig. S5). Similarly, injection of spry4 sense RNA into sip1a ATG morphants reduced the number of embryos with narrowed somites from 93% (sip1a ATG MO+lacZ mRNA) to 66% (sip1a ATG MO+spry4 RNA) (P = 0.0044) (Supp. Fig. 5). The ability of spry4 to restore somite morphology in only a subset of embryos likely reflects the poor stability of the injected RNA (Furthauer et al., 2001).

To circumvent this difficulty, we assayed the effects of blocking ChCh and Sip1a in ace/FGF8 mutants (Reifers et al., 1998; Draper et al., 2001). ace mutants show a loss of the isthmus and cerebellum, but do not have overt somite defects due to redundant FGF activity (Reifers et al., 1998; Sawada et al., 2001). Injection of chch ATG MO into ace homozygous embryos (somite length = 47.4 μm ± 2.4, n = 15; P = 0.003) did not produce the alterations in somite morphology observed in control morpholino injected embryos (somite length = 48.1 μm ± 1.4; n = 20). In the same experiments, chch morpholino injection decreased somite length in wild-type (somite length = 41.5 μm ± 4.6; n = 32) and ace heterozygous siblings (somite length = 40.4 μm ± 2.4; n = 29; Fig. 4A).

thumbnail image

Figure 4. Somite malformation in ChCh and Sip1a compromised embryos can be rescued by reduction of fibroblast growth factor-8 (FGF8). A,B: Dorsal views of 10-somite stage (A) and 12-somite stage (B) living embryos, anterior to the top. A: Somites are narrower at the anterior–posterior (A/P) axis, broader at the mediolateral axis in wild-type and ace heterozygous chch morphants, but not ace homozygous chch morphants. ace homozygous sip1a ATG morphants do not have somite phenotype, but wild-type and ace heterozygous sip1a morphants still have the somite defects. B: Horizontal and vertical red dotted line spans the first five somites of the control injected embryo width and length for comparison to the first five somite of morphant embryos which is indicated with a black line. Embryos were genotyped for the ace allele after imaging.

Download figure to PowerPoint

Similarly, injection of sip1a ATG MO into ace homozygous embryos did not result in significant narrowing of somites in the A/P axis (somite length = 47.1 μm ± 2.4; n = 9; P < 0.0001) compared with control-injected embryos (somite length = 49.0 μm ± 2.1; n = 22). In the same experiments, somite length was decreased in sip1a ATG MO injected wild-type (somite length = 40.1 μm ± 2.7; n = 17) and ace heterozygous siblings (somite length = 41.3 μm ± 2.3; n = 16).

These findings suggest that the consequences of chch knockdown on somite size are largely due to expansion of FGF signaling. However, in Sip1a compromised embryos, the overall phenotype is more severe than Chch compromised embryos and reduction of FGF signaling in these embryos is not sufficient to fully rescue the defects. Because Sip1 has wide range of functions that are distinct from known roles for ChCh, including regulation of TGF-β and bone morphogenetic protein (BMP) pathways (Remacle et al., 1999; Verschueren et al., 1999; Postigo, 2003; Postigo et al., 2003; Nitta et al., 2004; van Grunsven et al., 2007) and cell fate determination (Remacle et al., 1999; Verschueren et al., 1999), some of the phenotypes observed in sip1a morphants may be due to FGF-independent activities of Sip1a.

Somite Defects in ace Mutants Are Unaltered by ChCh Knockdown

Previous studies have demonstrated the ChCh is an effector of FGF signaling. However, our results reveal that ChCh also represses fgf8 expression. FGF signaling in the somites is required to induce myoD expression and terminal differentiation of subset of fast muscle cells. ace mutants have reduced somatic myoD expression (Groves et al., 2005). If FGF8 functions downstream of ChCh in somitogenesis, somite defects present in ace mutants will not be altered by chch knockdown. Therefore, we performed RNA in situ hybridization with myoD probe on ChCh compromised ace mutant embryos. As expected, we observed that lateral myoD expression in the somites is lost in ace mutant embryos, but adaxial cell expression was unaffected (Groves et al., 2005; Fig. 5C). We observed similar reduction of myoD expression in ChCh compromised ace−/− embryos as in control MO injected ace−/− embryos (Fig. 5D). This observation is consistent with the model that FGF8 acts downstream of ChCh during somite formation.

thumbnail image

Figure 5. Somite defects in ace mutants are unaltered by ChCh knockdown. A–D: Flat-mount RNA in situ hybridization of myoD in 17-somite stage ChCh compromised wild-type, ace+/−, and ace−/− embryos. All views are dorsal; anterior to the left. C: Lateral myoD expression in the somites is lost in ace mutant embryos. D: myoD expression pattern in ChCh compromised ace−/− embryos is similar to control MO injected ace−/− siblings.

Download figure to PowerPoint

To determine whether ChCh and Sip1a modulate FGF signaling in other tissues, we examined the consequence of knockdown of ChCh and Sip1a in the isthmus, which is a well-characterized site of FGF activity. We also observed that the expression of the FGF target gene pax2a is expanded anteriorly in the isthmus in chch morphants, but not in sip1a morphants (Fig. 6A–C). While the hindbrain is wider in ChCh and Sip1a compromised embryos (probably due to convergence defects), a similar alteration was not observed in krox20 expression (Fig. 6D–F). These results suggest that negative regulation of FGF8 by ChCh is not limited to the mesoderm, but that the ChCh has a broader function in modulating FGF signaling.

thumbnail image

Figure 6. Repression of fibroblast growth factor-8 (FGF8) by ChCh is not limited to the mesoderm. A–F: Whole-mount RNA in situ hybridization of pax2a and krox20 in 10-somite stage ChCh and Sip1a compromised embryos. All views are dorsal; anterior to the top. A–C: FGF target gene pax2a is expanded anteriorly in the isthmus in chch morphants, but not in sip1a morphants. D–F: The anterior extent of krox20 expression is not altered by chch or sip1a knockdown. Arrows denote isthmus.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

ChCh and Sip1 modulate FGF-dependent processes and act as a switch between mesodermal and neural inducing activities of FGF in chick, Xenopus, and zebrafish (Sheng et al., 2003; Londin et al., 2007b). Although their regulatory properties and function during gastrulation have been studied extensively, very little is known about the requirements for ChCh and Sip1a at other stages. In zebrafish, both genes are expressed in the developing mesoderm (Londin et al., 2007a; Delalande et al., 2008). Previous studies using SIP1 knockout mice and sip1 morphant zebrafish embryos did not clearly establish the function of Sip1 in somitogenesis (Maruhashi et al., 2005; Delalande et al., 2008).

In the present study, we characterized the function of ChCh and Sip1a in zebrafish somitogenesis. Our data revealed that chch and sip1a knockdown resulted in embryos with somites that are less extended through the A/P axis while overextended in the mediolateral axis. In addition, cyclic expression of her1 and her7 is maintained in formed somites in these embryos. We observed that these defects correlated with an anterior expansion of FGF8 expression in the tail bud. In ChCh morphants, the defects in somite morphology could almost entirely suppressed by blocking FGF8, while the same treatment only partially restored the overall defects in sip1a morphants. Taken together, these findings demonstrate that ChCh and Sip1a regulate somitogenesis by mediating the position of the FGF8 mediated wavefront in the zebrafish PSM.

Expanded FGF8 expression in PSM would be predicted to alter the FGF gradient that regulates somite length (Sawada et al., 2001). However, our finding that somite width was also altered by the enhanced FGF8 expression was surprising. Because expression of dominant-negative ChCh blocks FGF mediated mesoderm induction in animal cap assays (Sheng et al., 2003), ChCh is generally thought of as a positive effector of FGF signaling. ChCh is induced in response to FGF in chick, Xenopus, and zebrafish (Sheng et al., 2003; Londin et al., 2007a), and our results demonstrate that ChCh represses expression of FGF8, suggesting that it functions in a negative feedback loop to repress FGF signaling. In a broad context, this relationship is consistent with functions of ChCh in cell movement. Previous studies have shown that ChCh impedes cell movement (Sheng et al., 2003; Londin et al., 2007b). While FGFs have complex roles in cell migration, in many tissues, FGF promotes cell migration (Sun et al., 1999; Ciruna and Rossant, 2001; Yang et al., 2002). Therefore it is conceivable that the effects of ChCh on cell movement are due to blocking, rather then promoting FGF signaling. Further studies are needed to elucidate the precise mechanisms by which ChCh modulates FGF signal transduction.

The somite phenotype observed in sip1a ATG morphants is stronger than chch ATG morphants, and although somite phenotype can be rescued by blocking FGF activity, the overall defects cannot be fully restored. Thus far, Sip1 is the only described transcriptional target of ChCh. In several situations, overexpression of sip1 is sufficient to compensate for ChCh deficits (Sheng et al., 2003; Snir et al., 2006). However, Sip1 has a wide range of activities that are likely to be ChCh independent. These include regulation of TGF-β pathways (Remacle et al., 1999; Verschueren et al., 1999; Postigo, 2003; Postigo et al., 2003), expression of BMP4 (Nitta et al., 2004; van Grunsven et al., 2007), mesodermal gene expression (Remacle et al., 1999; Verschueren et al., 1999), and E-cadherin transcription (Remacle et al., 1999; Comijn et al., 2001).

The functional differences between the alternatively spliced forms of zebrafish Sip1a remains unclear. The previously described alternatively spliced form of sip1a lacks a zinc finger (Delalande et al., 2008). Because we identified a structurally similar form of Sip1b (Supp. Fig. 2), we reasoned that the two forms have unique functions. To test for distinct activities during somitogenesis, we used splice morpholinos to block production of each form while leaving the other intact. Our approach demonstrated the effectiveness of morpholinos to specifically eliminate alternative spliced message. However, we were unable to detect somite defects or any other patterning defects in the respective Sip1a form-specific morphants. This suggests that each form can compensate for loss of the other during somitogenesis or that protein derived from maternal mRNA is sufficient to compensate for the loss of wild-type zygotic mRNA.

In conclusion, we studied the functions of ChCh and Sip1a during zebrafish somitogenesis and found that ChCh and Sip1a modulate somite morphogenesis by repressing FGF8 expression at the PSM. Significantly, we determined that Fgf8 is downstream of ChCh, suggesting a negative feedback loop between Chch/Sip1a and Fgf8. Our data also demonstrate that regulation of FGF signaling by ChCh is not limited to the PSM. FGF signaling has diverse functions in many biological processes, and investigation of the vertebrate EST databases reveals that ChCh transcripts are detected in low levels in a wide array of tissues (data not shown). It will therefore be important for future studies to determine the importance of modulation of FGF signaling by ChCh in these contexts.

EXPERIMENTAL PROCEDURES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Adult Fish and Embryo Maintenance

Adult zebrafish strains and embryos obtained from natural crossings were maintained at 28.5°C. Developmental stages of the embryos were determined according to Kimmel et al. (1995).

Expression Constructs, mRNA Synthesis, and Morpholinos

spry4 ORF was amplified from 10-somite stage wild-type total first-strand cDNA using GTTCTAGAGGCTCGAGGAAGGTCCTGCAAACCAT/TCTTTTTGCAGGATCCTGAGGAACACGACCTACA primer pair. Amplified fragment is then cloned to pCS2+ at BamHI and XhoI sites. Capped sense mRNA was synthesized using mMESSAGE mMACHINE Kit (Ambion).

The sequences of the morpholinos used are chch ATG MO 5′-CAGTATAGTCCAGATCAGAAGACGC-3′ and chch ATG MO2 5′-GCTTCTGGACACAACCGGTACACAT-3′(Londin et al., 2007b). sip1a and sip1b ATG and splice morpholinos are kindly provided by Iain T. Shepherd (Delalande et al., 2008); sip1a splice variant targeting MO2 5′-GTCTAAATGTGATATAC CTGTGC-3′ and sip1a splice variant targeting MO3 5′-CGCGTACATACCACTTTCAGTCTTC-3′. Primers used to monitor the efficiency of the splice variant targeting MO are MO2: ATGTACGCGTGTGACTTGTG/CATTTGT CGCACTGGTAAGG, MO3: TTAAGAAGACTGAAAGTGGAAAGC/CAT TTGTCGCACTGGTAAG.

Standard control oligo (Gene Tools) is used as control. mRNA and morpholino solutions were diluted to desired concentration with 0.2 M KCl supplemented with phenol red. Typically, 500 pg of spry4 mRNA, 10–15 ng of ChCh ATG morpholino1 and ChCh ATG morpholino2, 2–4 ng of sip1a ATG, 10 ng of sip1a splice morpholino, 5 ng of sip1b ATG, 1ng of sip1b splice morpholino is injected to one- to two-cell stage embryos.

Whole-Mount In Situ Hybridization and Immunohistochemistry

In situ hybridizations were performed according to Thisse et al. (1993). Digoxigenin labeled probes for in situ hybridization was synthesized using T7, T3, or Sp6 RNA polymerase (Roche). Hybridized probes were detected using NBT/BCIP (nitroblue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate) system (Roche). Stained embryos were re-fixed in 4% paraformaldehyde and either stored in 100% methanol or cleared in Benzyl benzoate/benzyl alcohol solution (2:1) and mounted in Canada balsam/methyl salicylate (2.5% v/v) or flat-mounted in 70% glycerol. Embryos were viewed with Zeiss Axioplan microscope, digitally photographed with Zeiss Axiocam camera. Images were processed and assembled with Zeiss Axiovision and Adobe Photoshop.

Genotyping acerebellar (ace) Mutants

chch and sip1a ATG MO injected embryos obtained from ace heterozygous incross were genotyped following imaging. PCR primers to genotype ace embryos are GCCAAGCTTATAGTAGAGAC/ AAGTGATGACTTTTTCAGATA. Following PCR, products were cut with EcoRV, which digests the mutant but not wild-type alleles.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

We thank Bernadette Holdener, Jerry Thomsen, and members of their laboratories for helpful suggestions; Laura Mentzer for technical support; Heena Rana and Andrew Taibi for fish care. We also thank the many labs that provided reagents including fish stocks, constructs and probes. H.S. was funded by a NIH grant.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
DVDY_22201_sm_SuppFig1.tif256KSupp. Fig. S1. sip1a splice morpholino is effective in eliminating wild-type sip1a mRNA. Reverse transcriptase-polymerase chain reaction (RT-PCR) of mRNA from staged sip1a splice MO injected embryos with sip1a specific primers reveals that the smaller misspliced product is detected by dome stage (4.3 hours postfertilization [hpf]). By 75% epiboly (8 hpf), no wild-type mRNA is detected.
DVDY_22201_sm_SuppFig2.tif6675KSupp. Fig. S2. Comparison of zebrafish Sip1a and Sip1b proteins. Genedoc software is used to align the predicted protein sequences. Predicted domains of Sip1a and Sip1b is lined with colored bars: C2H2 type zinc fingers present in both Sip1a and Sip1b (red), C2H2 type zinc finger region that is absent from Sip1a and Sip1b short forms (blue). Identical and similar amino acids conserved among all proteins are shown in black and dark gray boxes, respectively. Lighter shades of gray or no shading represent low levels of amino acid conservation and the lack of conservation, respectively.
DVDY_22201_sm_SuppFig3.tif5846KSupp. Fig. S3. sip1a splice variant targeting morpholinos efficiently eliminate short and long forms. A: Diagram of the alternative splicing within the 3′ region of the sip1a pre-mRNA. Alternative splicing of exon 8 eliminates one zinc finger (dark blue box), which is present in the longer form. Sip1a MO3 (orange bar) targeted to alternative splice site in exon 8 (denoted by black arrow) blocks the alternating splicing event and eliminates production of the short form. Sip1a MO2 (pink bar) blocks the pre-mRNA splicing event needed to generate the long form by targeting the splice site at the 3′ end of exon 8. B,C: Reverse transcriptase-polymerase chain reaction (RT-PCR) of mRNA from staged sip1a splice MO2 and MO3 injected embryos with sip1a full length specific primers (B, green arrows in A) and sip1a short form specific primers (C, orange arrows in A ). B,C: sip1 splice MO2 efficiently altered splicing so that the full length message is eliminated and only the shorter form was produced (B) while sip1 splice MO3 blocked production of the shorter form (C).
DVDY_22201_sm_SuppFig4.tif18565KSupp. Fig. S4. Pace of the “molecular clock” is not significantly altered in ChCh-compromised embryos. Living wild-type and ChCh-compromised embryos beginning at the seven-somite stage. A–J: All views are dorsal, anterior to the top. Duration of somite formation in both control (A–E) and ChCh-compromised embryos (F–J) is approximately 45 min at 23°C. Black arrows denote already formed somite boundaries and white arrows denote newly forming segmentation furrow.
DVDY_22201_sm_SuppFig5.tif494KSupp. Fig. S5. Somite malformation in ChCh and Sip1a compromised embryos can be partially rescued by fibroblast growth factor (FGF) antagonist Spry4. Bar graph representation of spry4 rescue assay in chch and sip1a ATG morphants. Injection of spry4 sense RNA into chch and sip1a ATG morphants (blue bar) reduced the penetrance of the somite phenotype with respect to their lacZ injected siblings (gray bar).

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.