Mr. Taibi, Mr. Mandavawala, and Ms. Noel contributed equally to this work.
Zebrafish churchill regulates developmental gene expression and cell migration
Article first published online: 29 MAR 2013
Copyright © 2013 Wiley Periodicals, Inc.
Volume 242, Issue 6, pages 614–621, June 2013
How to Cite
Taibi, A., Mandavawala, K. P., Noel, J., Okoye, E. V., Milano, C. R., Martin, B. L. and Sirotkin, H. I. (2013), Zebrafish churchill regulates developmental gene expression and cell migration. Dev. Dyn., 242: 614–621. doi: 10.1002/dvdy.23958
- Issue published online: 20 MAY 2013
- Article first published online: 29 MAR 2013
- Accepted manuscript online: 27 FEB 2013 12:00AM EST
- Manuscript Accepted: 18 FEB 2013
- Manuscript Revised: 4 FEB 2013
- Manuscript Received: 3 JAN 2013
- NIH. Grant Numbers: HD066000, GM008655
- zinc finger nuclease;
- cell migration;
Background: Regulation of developmental signaling pathways is essential for embryogenesis. The small putative zinc finger protein, Churchill (ChCh) has been implicated in modulation of both TGF-β and FGF signaling. Results: We used zinc finger nuclease (ZFN) mediated gene targeting to disrupt the zebrafish chch locus and generate the first chch mutations. Three induced lesions produce frameshift mutations that truncate the protein in the third of five β-strands that comprise the protein. Surprisingly, zygotic and maternal zygotic chch mutants are viable. Mutants have elevated expression of mesodermal markers, but progress normally through early development. chch mutants are sensitive to exogenous Nodal. However, neither misregulation of FGF targets nor sensitivity to exogenous FGF was detected. Finally, chch mutant cells were found to undergo inappropriate migration in cell transplant assays. Conclusions: Together, these results suggest that chch is not essential for survival, but functions to modulate early mesendodermal gene expression and limit cell migration. Developmental Dynamics 242:614–621, 2013. © 2013 Wiley Periodicals, Inc.†
The establishment of the vertebrate body plan depends on a carefully orchestrated series of inductive interactions and cell movements. These events determine the positions and proportions of cells that will populate each of the three germ layers. Although many molecular pathways involved in these processes have been identified, the mechanisms that integrate these signals and promote differential responses to the same signals are poorly understood. The small zinc finger protein Churchill (ChCh) has been implicated in both germ layer specification and modulation of differential responses to fibroblast growth factor (FGF) and transforming growth factor-beta (TGF-β) signals.
ChCh was identified in a differential screen to isolate genes expressed following exposure to organizer tissue in the chick and was proposed to function in germ layer specification (Sheng et al., 2003). Overexpression of ChCh in Xenopus embryos results in suppression of the mesodermal marker Xbra in embryos and animal cap assays (Sheng et al., 2003). Mopholino knockdown of ChCh in zebrafish results in enhanced expression of the zebrafish no-tail (brachyury orthologue) and somite defects (Londin et al., 2007b; Kok et al., 2010). In Xenopus, chch was found to be downstream of the OCT3/4 homologue POU91 and proposed to be a key regulator of neural competence (Snir et al., 2006). In addition, chch transcripts are induced by FGF in both chick (Sheng et al., 2003) and zebrafish (Londin et al., 2007a). Morpholino knockdown of ChCh in zebrafish also suggested that ChCh represses FGF signaling (Kok et al., 2010).
Regulation of cell migration by ChCh is thought to be central to its role in cell fate determination. Knockdown of chch in chick epiblast cells results in inappropriate migration of those cells through the primitive streak (Sheng et al., 2003). Ultimately, ChCh-compromised epiblast cells gave rise to paraxial mesoderm. Likewise, we observed that cell autonomous inhibition of ChCh in zebrafish presumptive ectoderm promotes abnormal migration causing these cells to reside in the mesoderm or superficial layers of the trunk, depending on the time of transplant (Londin et al., 2007b).
Analysis of the chch primary sequence suggested that the protein is comprised of two C4 type zinc finger domains (Sheng et al., 2003). Initially, these zinc fingers were thought to confer DNA binding specificity, with Sip1 identified as a direct target (Sheng et al., 2003). However, subsequent NMR analysis suggested that ChCh does not contain canonical C4 zinc fingers but rather contains three bound zinc ions in an atypical configuration (Lee et al., 2007). The bulk of the protein is comprised of five β-strands that are configured into an unusual solvent exposed single layer β-sheet. This study also concluded that ChCh does not directly bind DNA, but suggested that the zinc-binding motif may instead be involved in protein–protein interactions (Lee et al., 2007).
To study the consequences of ChCh loss of function on embryonic development, we used zinc finger nuclease mediated gene targeting to disrupt the zebrafish chch locus. This is the first reported chch mutation in any species. To our surprise, chch mutants are viable and fertile. However, analysis of mutants revealed that ChCh is required to repress Nodal signaling and expression of the zebrafish brachyury orthologue no-tail. Surprisingly, expression of FGF targets was not altered in chch mutants. Lastly, we observed that chch mutant cells undergo inappropriate migration in a cell transplant assay revealing that ChCh is required to limit cell migration.
Genetic Disruption of the chch Locus
To study the function of chch during zebrafish development we used zinc-finger nuclease mediated gene targeting (Kim et al., 1996; Doyon et al., 2008; Meng et al., 2008). The ZiFiT algorithm (http://www.zincfingers. org/software-tools.htm) was used to predict three finger zinc finger arrays that would bind the chch locus. Because computational prediction of zinc finger array affinity to DNA is imprecise (Ramirez et al., 2008), we designed two “left” and two “right” arrays to generate four potential nucleases (which require binding of arrays on both the left and right of the target sequence). The arrays were assembled using single zinc finger clones from the Addgene ZF consortium modular assembly kit (addgene.com), that were fused to heterodimeric FokI variants in a CS2 backbone using conventional cloning approaches (Table 1). Each of the four potential combinations of mRNAs were microinjected into one-cell embryos. At 1 days postfertilization (dpf), injected embryos were screened for somatic mutations by polymerase chain reaction (PCR) using primers that flank the target sequence. Size polymorphisms reflective of small insertions and deletions were detected in 10 to 29% of the injected embryos (Table 2).
|ZFN Name||Half Target||ZF Array Sequence|
|L1 (127–140-123)||GTA GGT GAA||QSSTLTRWPSNLTRQKSNLIR|
|L2 (126–140-116)||GTA GGT GAA||QSSSLIRWPSNLTRQKGNLLR|
|R1 (107–112-138)||GAC GAT GTG||CPSNLRRISSNLQRVSSSLRR|
|R2 (2–21-49)||GAC GAT GTG||DRSNLTRTSGNLVRRSDALTR|
|ZFN pair||Somatic mutations||Germline mutations||% transmission|
|1L-2R||7/24 (29%)||1/42 (2.3%)||7/44 (15.9%)|
|2L-2R||5/32 (16%)||2/7 (28.5%)||23/98 (23.4%)|
Embryos microinjected with the two array combinations that showed the highest somatic mutation rates were grown to adulthood and screened for germline transmission of chch lesions by PCR. Three independent mutant chch alleles were recovered (Fig. 1). Sequence analysis of these mutations revealed a 7 bp insertion (SBU48), a 5 bp deletion (SBU49) and a 10 bp deletion with a 2 bp insertion (SBU50). All three mutations produce frameshifts that are predicted to prematurely truncate the protein (Fig. 1).
Wild-type chch contains a secondary structure composed of five solvent-exposed β-strands that form a single layer β-sheet (Lee et al., 2007). In addition, three zinc ions are bound in an unusual configuration (Lee et al., 2007). The lesion sites are in the 3rd β-strand, N-terminal to one of the zinc coordination sites. All of the mutations are predicted to truncate the protein in the third β-strand (Fig. 1). The complete elimination of the last two β-strands is likely to severely impact ChCh function.
chch Mutants Are Viable
Morphological analysis of all three chch alleles (SBU48, SBU49, SBU50) did not reveal any overt morphological defects during embryogenesis or larval development. Homozygous chch mutants survive to adulthood and are fertile. Because chch is supplied as a maternal transcript (Londin et al., 2007a), maternally supplied chch mRNA might provide sufficient chch activity during early development and mask effects of zygotic chch deficiency. Crossing homozygous male and female mutants produces embryos that lack both maternally supplied and zygotic chch mRNA. We found that these maternal zygotic (MZchch) mutants were morphologically indistinguishable from wild-type controls (Fig. 2) and are viable.
Chch Regulates Developmental Gene Expression
To determine whether chch mutants have subtle changes in specification or patterning of the mesoderm, we assayed chchsbu50/sbu50 mutants by RNA whole-mount in situ hybridization (Fig. 3). Expression of markers of mesoderm (no-tail, ntl) and dorsal mesoderm (chd and flh) were comparable in MZchchsbu50/sbu50 mutants and heterozygous sibling controls (Fig. 3A–F). At later stages, no changes were observed in myoD or foxa2 expression in the mutants (Fig. 3G–J). To quantitatively measure alterations in marker expression in the mutants, we prepared cDNA from MZchchsbu50/sbu50 mutants and staged-matched wild-type controls for quantitative reverse transcriptase-PCR (qRT-PCR) analysis. At 6 hours postfertilization (hpf) (shield stage), we observed an ∼2.5-fold increase in ntl expression (Fig. 4a) and a slightly larger increase at 9 hpf (90% epiboly) (Fig. 4b). This finding is consistent with our previous observations that microinjection of dominant-negative chch mRNA enhanced expression of mesodermal markers at shield stage (Londin et al., 2007b). In addition, we detected increased expression of the endodermal marker mixer and the dorsal mesodermal marker goosecoid (gsc) in chch mutants at 6 hpf (Fig. 4A). Expression levels of two other dorsal mesodermal markers, flh and chd were unaltered at this stage. These results demonstrate that ChCh acts to repress expression of mesodermal and endodermal markers during gastrulation.
Although it appears that ChCh is not a DNA binding protein, it modulates expression of specific genes, notably sip1 (Sheng et al., 2003; Londin et al., 2007b). Zebrafish have two sip1 paralogues, sip1a and sip1b (Delalande et al., 2008). Examination of these transcripts in MZchchsbu50/sbu50 mutants by qRT-PCR revealed that early expression of sip1a requires ChCh, while early sip1b expression is independent of ChCh (Fig. 4a). Based on these observations, we conclude that ChCh is required to limit expression of mesendodermal markers and to initiate or maintain early expression of sip1a. Whereas embryonic development does not require ChCh, these results suggest that it plays an important role in modulating the expression of mesendodermal transcripts.
Modulation of Nodal Signaling by ChCh
The ChCh responsive gene Sip1 binds to activated Smad2/3 to repress the response to TGF-β signals (Postigo, 2003; Postigo et al., 2003). We previously demonstrated that microinjection of dominant-negative chch increases the response to exogenous Nodal (Londin et al., 2007b) presumably by means of effects on Sip1. To determine whether chch mutants, which have reduced sip1a levels (Fig. 3a) are also sensitive to Nodal, we microinjected a low dose of sqt mRNA into MZ chchsbu50/sbu50 mutants and assayed ntl transcript levels by qRT-PCR (Fig. 5). At this dose of sqt, there was little effect on ntl levels in wild-type embryos, but ntl levels were enhanced in chch mutants. This indicates that ChCh suppresses the response to Nodal and is in accordance with our previous findings that inhibition of ChCh activity using a dominant-negative produced a similar effect on ntl levels (Londin et al., 2007b). We conclude that ChCh is required to limit the transcriptional response to Nodal.
ChCh Does Not Modulate FGF Signaling
chch is induced by FGF signaling (Sheng et al., 2003; Londin et al., 2007a) and previous findings using a chch morpholino suggested that ChCh also acts as a feedback inhibitor of the pathway (Kok et al., 2010). To dynamically assay the FGF response in chch mutants, we crossed a dusp6:GFP transgene (Molina et al., 2007), which serves as a readout of FGF signaling, into the chch mutant background. GFP expression from the transgene was observed in the dorsal retina, midbrain–hindbrain boundary, and trigeminal ganglion at 24 hpf in both chch mutants and sibling controls (Fig. 6A–D). No differences in expression were apparent between the chch mutants and their siblings. At 48 hpf, dusp6:GFP transgene expression was observed in spinal cord neurons as well as in midbrain–hindbrain boundary and trigeminal ganglion in similar patterns in both chch mutants and their sibling controls (Fig. 6E–H). These results suggest that chchsbu50/sbu50 mutants do not have appreciably altered domains of FGF signaling.
Although the transgenic approach provides spatial resolution to assess domains of FGF signaling, it is likely not sensitive to subtle alterations in gene expression levels. To quantitatively assay the FGF signaling levels and the response to exogenous FGF in chch mutants, we micro-injected one- to four-cell wild-type and chchsbu50/sbu50 mutant embryos with FGF8 mRNA and analyzed expression of two FGF target genes, sef and sprty2 by qRT-PCR at 6 hpf (shield stage; Fig. 7). Although both genes were induced by FGF8, we found no significant change in sef or sprty2 expression resulting from loss of chch in either control or FGF8-injected chch mutants (Fig. 7). Taken together with the transgenic FGF reporter analysis, these data provide no evidence for enhanced FGF responses in chchsbu50/sbu50 mutants.
Chch Suppresses Cell Migration
Previous studies in the chick and zebrafish suggested that Chch acts to limit migration of ectodermal cells (Sheng et al., 2003; Londin et al., 2007b). To directly compare the migratory properties of Chch deficient and wild-type cells we performed a transplant assay. Wild-type donor embryos were labeled with rhodamine dextran and MZchchsbu50/sbu50 mutant embryos with fluorescein dextran. Between 4 and 4.3 hpf (sphere to dome stage), animal pole cells were removed from each of the donors and placed at the animal pole of wild-type hosts (Fig. 8). At 24 hpf, the position of both wild-type donor and MZchchsbu50/sbu50 mutant donor cells was assessed. Whereas the majority of wild-type donor cells remained in rostral neural tissue or very close to the head, the chch mutant cells behaved much differently. In 15/21 double transplants, the ChCh deficient cells were more mobile than wild-type cells and migrated to posterior regions of the embryo. These results provide the first genetic evidence that ChCh is required to suppress migration of presumptive ectodermal cells.
We used zinc finger nuclease mediated gene targeting to disrupt the zebrafish chch locus and generate the first chch mutants that have been reported in any species. Three independent chch alleles were generated. The allele used for the majority of these studies (SBU50) is predicted to truncate the protein at the third of five β-stands. The truncated region of the protein also contains Leu89 and Lys93, which connect the N-terminal loop of the protein to the β-sheet (Lee et al., 2007). The relationship of the structure of ChCh to molecular function is not known, although it is thought that ChCh indirectly influences target gene transcription (Lee et al., 2007). We strongly suspect the SBU50 mutation is a complete loss of function allele due to the structural importance of the C-terminus and the β-strands.
Because all three chch alleles are viable, we conclude that full-length chch is not essential. Consistent with our previous studies that used dominant-negative ChCh and morpholinos to inhibit ChCh function, we observed that ChCh is required to limit Nodal signaling and suppress expression of ntl and other early markers of mesoderm and endoderm (Figs. 4, 5). Early sip1a expression is reduced in chch mutants, but not expression of sip1b. Unlike many vertebrates, zebrafish have two sip1 paralogues (sip1a and sip1b). The difference in ChCh regulation may account for some of the functional divergence in these genes in zebrafish (Delalande et al., 2008).
Morpholino and dominant-negative inhibition of ChCh function produced defects in somitogenesis that attributed to elevated levels of FGF signaling (Kok et al., 2010). We found no evidence of either somite defects or elevated FGF signaling in chch mutants (Figs. 6, 7). Three possibilities could explain the disparity. First, we cannot definitively conclude that the SBU50 allele is a null mutation; therefore, the intact N-terminal portion of ChCh may function to limit FGF signaling. A more detailed understanding of the biophysical function of ChCh, antibody reagents or additional alleles will be needed to investigate autonomous functions of the N-terminus. Alternatively, the presumed low levels of ChCh that are present in morpholino treated embryos may create a different milieu for FGF signaling than is created in the mutant and produce an different phenotypic outcome. Finally, although mRNA rescue controls were done to establish specificity of the chch morpholinos and dominant-negative constructs, it is still possible that off target effects of the morpholinos produce more robust effects on FGF signaling than are observed in the mutant. It is worth noting that wild-type chch mRNA did not restore all the effects of the morpholinos and dominant-negative mRNA (Londin et al., 2007b). Only ntl expression levels and the cell migration defects were rescued with wild-type chch mRNA. This is precisely the set of phenotypes that are observed in chch mutants.
Our previous studies, the work of others, as well as the current effort reveal a requirement for ChCh in limiting cell migration. Influences of ChCh on cell movement likely occur by means of modulation of Sip1, which directly regulates E-Cadherin expression (Comijn et al., 2001). Although the effects of ChCh inhibition on cell migration in transplant assays are robust (Fig. 8), we did not observe overt developmental abnormalities resulting from inappropriate migration. Such defects may be subtle or restricted to specific cell types or developmental stages. The transplant assays created situations where relatively small numbers of chch mutant cells were placed in a sea of wild-type cells. The differential adhesive properties of wild-type and mutant cells may enhance the migratory properties of the chch mutant cells.
In conclusion, we used zinc finger nuclease mediated gene targeting to disrupt the zebrafish chch locus. The resulting mutations truncate ChCh and eliminate nearly half of the residues in the β-sheet that comprise the bulk of the protein and eliminate one zinc coordination site. Nonetheless, we observed that both zygotic and maternal chch mutants are viable, suggesting that ChCh is not essential for survival. However, we found that ChCh is required to limit expression of early mesendodermal markers and to suppress cell migration. Future studies will be needed to elucidate the migratory events during development that are influenced by ChCh.
Wild-type and mutant lines were obtained through natural matings and maintained at 28.5°C. Developmental stages were determined according to the Kimmel staging series (Kimmel et al., 1995).
Generation of Zinc Finger Nucleases
To construct the zinc finger nucleases, plasmids containing individual fingers were isolated from the Zinc Finger Consortium Modular Assembly Library Kit (Addgene.com). Arrays of three zinc fingers were generated by successively ligating the F2 and F3 fingers into the Age1/BamH1 sites of the F1 vector. Each zinc finger array was cloned into the KpnI/BamHI sites of the heterodimeric FokI-RR/FokI-DD vectors (Meng et al., 2008). mRNA corresponding to each construct was synthesized using mMESSAGE mMACHINE Kit (Ambion).
Microinjection of Zinc Finger Nucleases
A total of 40–50 pg of total zinc finger nuclease (zfn) mRNA was injected into the blastomere of a one-cell stage wild-type embryo. Each clutch of injected embryos were screened for somatic mutations in chch by PCR to detect small indels using the following primers that flank the target site: Chch genomic F 5′-CTGTGAAGCGAGGGCTCA-3′ and R 5′-CTTTTCCTCAGACATGTGCAA-3′. PCR products were analyzed on a conventional 2.5% agarose gel. Zinc finger array combinations yielding the highest rate of somatic mutation were again micro-injected into wild-type embryos which were grow to adulthood. To screen for germline transmission of chch lesions mature injected fish were crossed and progeny were screened for lesions by PCR.
Whole-Mount RNA In Situ Hybridization
In situ hybridizations were performed according to Thisse et al. (1993). Digoxigenin labeled antisense RNA probes were synthesized using T3, T7, SP6 RNA polymerase (Roche). Embryos were stained with nitro-blue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate (NBT/BCIP) for detection of anti-digoxigenin conjugated alkaline phosphatase. Stained embryos were fixed in 4% paraformaldehyde (Sigma) in phosphate buffered saline then dehydrated and stored in methanol at −20°C. For photographs, embryos were cleared in 2:1 benzyl benzoate to benzyl alcohol and mounted in 2.5% methyl salicylate in Canada balsam (Sigma). Embryos were then photographed on a Zeiss Axioplan 2 microscope with Axiocam mounted camera and AxioVision Software.
Quantitative Real-Time PCR
RNA was extracted from pools of five embryos, which were either chchsbu50/sbu50 mutants or wild-type. cDNA was then prepared using SmartScribe Reverse Transcriptase (Clontech) for use in quantitative real-time PCR. Amplification was performed using PerFectA Sybr Green (Quanta Biosciences) and analyzed using Roche Lightcycler 480. All samples were run in duplicate and normalized to β-actin. All quantifications are shown as relative to wild-type, error bars represent standard error of the mean of three groups of five embryos. Primers were used as previously described in Londin, et al. 2005, and Londin et al., 2007 except: sef F- 5′-GAATTGTGCCGTTTCCAACT-3′ R- 5′-GAATGGTTTGCAGACGAGGT-3′ and sprouty2 F- 5′-CATGCAGCCAAACATTCAAC-3′ R- 5′-CAAAATGCGCCGAGTTTTAT-3′.
We thank the many members of the community who provided reagents. In particular we thank Michael Tsang for the Dusp6:GFP fish line. We thank Bernadette Holdener, for comments on the manuscript Erika Wunderlich for fish care and Otto Mullings for technical assistance. H.S. was funded by the NIH J.N. received a NIH MARC grant, K.M. received a Stony Brook URECA fellowship and E.O. received a NIH Bridges to the Baccalaureate grant.
- 2001. The two-handed E box binding zinc finger protein SIP1 downregulates E-cadherin and induces invasion. Mol Cell 7:1267–1278. , , , , , , , , .
- 2008. Zebrafish sip1a and sip1b are essential for normal axial and neural patterning. Dev Dyn 237:1060–1069. , , , .
- 2008. Heritable targeted gene disruption in zebrafish using designed zinc-finger nucleases. Nat Biotechnol 26:702–708. , , , , , , , , , , , , .
- 1996. Hybrid restriction enzymes: zinc finger fusions to Fok I cleavage domain. Proc Natl Acad Sci U S A 93:1156–1160. , , .
- 1995. Stages of embryonic development of the zebrafish. Dev Dyn 203:253–310. , , , , .
- 2010. Churchill and Sip1a repress fibroblast growth factor signaling during zebrafish somitogenesis. Dev Dyn 239:548–558. , , .
- 2007. Embryonic neural inducing factor churchill is not a DNA-binding zinc finger protein: solution structure reveals a solvent-exposed beta-sheet and zinc binuclear cluster. J Mol Biol 371:1274–1289. , , , , .
- 2007a. Expression and regulation of the zinc finger transcription factor Churchill during zebrafish development. Gene Expr Patterns 7:645–650. , , , .
- 2007b. Churchill regulates cell movement and mesoderm specification by repressing Nodal signaling. BMC Dev Biol 7:120. , , .
- 2008. Targeted gene inactivation in zebrafish using engineered zinc-finger nucleases. Nat Biotechnol 26:695–701. , , , , .
- 2007. Generation of FGF reporter transgenic zebrafish and their utility in chemical screens. BMC Dev Biol 7:62. , , .
- 2003. Opposing functions of ZEB proteins in the regulation of the TGFbeta/BMP signaling pathway. EMBO J 22:2443–2452. .
- 2003. Regulation of Smad signaling through a differential recruitment of coactivators and corepressors by ZEB proteins. EMBO J 22:2453–2462. , , , .
- 2008. Unexpected failure rates for modular assembly of engineered zinc fingers. Nat Methods 5:374–375. , , , , , , , , , , , , .
- 2003. Churchill, a zinc finger transcriptional activator, regulates the transition between gastrulation and neurulation. Cell 115:603–613. , , .
- 2006. Xenopus laevis POU91 protein, an Oct3/4 homologue, regulates competence transitions from mesoderm to neural cell fates. EMBO J 25:3664–3674. , , , .
- 1993. Structure of the zebrafish snail1 gene and its expression in wild-type, spadetail and no tail mutant embryos. Development 119:1203–1215. , , , .