Cold-inducible RNA binding protein (CIRP), a novel XTcf-3 specific target gene regulates neural development in Xenopus
- 4.7k Downloads
As nuclear mediators of wnt/β-catenin signaling, Lef/Tcf transcription factors play important roles in development and disease. Although it is well established, that the four vertebrate Lef/Tcfs have unique functional properties, most studies unite Lef-1, Tcf-1, Tcf-3 and Tcf-4 and reduce their function to uniformly transduce wnt/β-catenin signaling for activating wnt target genes. In order to discriminate target genes regulated by XTcf-3 from those regulated by XTcf-4 or Lef/Tcfs in general, we performed a subtractive screen, using neuralized Xenopus animal cap explants.
We identified cold-inducible RNA binding protein (CIRP) as novel XTcf-3 specific target gene. Furthermore, we show that knockdown of XTcf-3 by injection of an antisense morpholino oligonucleotide results in a general broadening of the anterior neural tissue. Depletion of XCIRP by antisense morpholino oligonucleotide injection leads to a reduced stability of mRNA and an enlargement of the anterior neural plate similar to the depletion of XTcf-3.
Distinct steps in neural development are differentially regulated by individual Lef/Tcfs. For proper development of the anterior brain XTcf-3 and the Tcf-subtype specific target XCIRP appear indispensable. Thus, regulation of anterior neural development, at least in part, depends on mRNA stabilization by the novel XTcf-3 target gene XCIRP.
KeywordsNeural Plate Morpholino Oligonucleotide Neurula Stage Antisense Morpholino Oligonucleotide Anterior Neural Plate
The most prominent nuclear transducer of wnt/β-catenin signaling in the nucleus belong to the Lef/Tcf family of sequence specific HMG-box transcription factors. Higher vertebrates express four distinct members of this family, Tcf-1 (or Tcf-7), Tcf-3, Tcf-4 (or Tcf-7 like 2) and Lef-1 [1, 2]. In Xenopus, XTcf-1, XTcf-3 and XTcf-4 are maternally provided, while the expression of XLef-1 starts only after mid-blastula transition . The expression of XTcf-1, XTcf-3 and XLef-1 is widespread and overlapping in many tissues, with highest levels in the head region including brain, head mesoderm and branchial arches [3, 4]. Zygotic expression of XTcf-4 is restricted in neurula and tailbud stages to the anterior midbrain  and in tadpoles to the entire midbrain .
A partial functional redundancy of Lef/Tcf family members has been shown in mice, where the Tcf-1/Lef-1 and Tcf-1/Tcf-4 double knock-out mice revealed more than additive effects compared to the corresponding single knock-out mice [7, 8]. In Xenopus, however, the individual Lef/Tcfs appear to have more non-redundant roles. Although XTcf-1, XTcf-3 and XTcf-4 are maternally provided, it has been shown by rescue experiments of Lef/Tcf depleted embryos that XTcf-1 mediated target gene activation in partial redundancy with XTcf-4, and XTcf-3 mediated target gene repression are required for establishing of the dorsal embryonic axis  and mesoderm patterning . Zygotic XTcf-4 has been shown to regulate midbrain development . XLef-1 is necessary for mesoderm patterning . These non-redundant roles of Lef/Tcfs during early Xenopus development implicate that the individual transcription factors regulate a different subset of target genes.
It seems most likely that Lef/Tcfs are general regulators of brain development because their expression overlaps in the developing brain and posteriorization of the developing CNS is regulated by wnt/β-catenin signaling . Thereby, they seem not to play a role as inducing factors, driving the cell fate towards anterior neural development, because they can not induce the expression of neural marker genes in Xenopus animal cap explants. Instead, they rather appear to be involved in the stabilizing and transforming step in the model of Stern . However, dorsal injection of a dominant negative Tcf-3 construct resulted in a repression of Bmp4 and subsequently in a downregulation of the pan-neural marker gene nrp1 . This is discussed as β-catenin independent target gene activation by a physical interaction between XTcf-3 and Frodo/Dapper .
In order to identify genes specifically regulated by XTcf-3, we subtracted the transcriptome of neuralized animal caps of Tcf-4 morpholino injected embryos from those of Tcf-3 morpholino injected ones and identified among others Xenopus cold inducible RNA binding protein (XCIRP) as novel XTcf-3 specific target gene. XCIRP was originally identified in a screen for target genes activated after retinoic acid treatment of activin treated explants . Later it was found in a screen for genes, which are upregulated in animal caps neuralized by smad7 . In Xenopus three XCIRP isoforms, XCIRP, XCIRP1 and XCIRP2 have been reported. Because XCIRP and XCIRP1 mRNA is to 98% identical resulting in proteins of 99.4% identity they are referred in here further on as XCIRP. XCIRP protein is to 94% identical to XCIRP2, implicating that these isoforms derive from different alleles of the pseudotetraploid Xenopus laevis.
XCIRP is highly expressed in developing oocytes and forms together with FRGY2 and mRNP3 the most abundant group or RNA binding proteins in oocytes [18, 19]. It interacts with and is methylated by the protein-arginine methyltransferase 1 (XPRMT1). This methylation is supposed to play an essential role in regulating the subcellular localization of XCIRP . XCIRP interacts with ELAV-like RNA-binding protein (XELrA) in vitro and in vivo and inhibits deadenylation of synthetic AU rich elements . Immunoprecipitation followed by RT-PCR revealed that many different mRNAs including cyclin B1, CAF1 p150 (Chromatin assembly factor 1 p150 subunit) Nek2B, Cytochrome C oxidase peptide VIb and bystin bind to XCIRP protein in Xenopus oocytes . In a similar assay, Peng et al.  identified α- and β-catenin as well as C- and E-cadherin mRNA bound to XCIRP protein. Binding of XCIRP to the above mentioned mRNAs appears to increase their stability . Whether binding and stabilization of mRNA holds true only for a small subset of mRNA or if mRNA in general is stabilized by XCIRP remains elusive.
Although almost ubiquitously expressed at high levels during gastrulation , first effects of XCIRP knockdown by antisense morpholino injection were not seen before stage 22, where more than half of the embryos died . More specific effects of the XCIRP morpholino were observed only after targeting the injection to the C3 blastomere in 32 cell stage embryos. This resulted in malformation of the pronephros .
Here we show by antisense morpholino oligonucleotide injection followed by in situ hybridization that the expression of XCIRP depends on XTcf-3 but not on XLef-1 or XTcf-4. The expression fields of XTcf-3 and XCIRP broadly overlap in the neural tissue and both, the XTcf-3 specific and the XCIRP specific anitsense morpholino oligonucleotide lead to a similar phenotype, a broadening of the neuroectoderm. We provide evidence, that a general stabilization of mRNA by the XTcf-3 target gene XCIRP is an essential step in early neural development.
Among others (additional file 1), we found as putative XTcf-3 specific target genes most abundant, high mobility group nucleosomal binding protein 1 and 2 (HMGN1 and 2), HMGB3, Xenopus anterior gradient (XAG) and cold inducible RNA binding protein (XCIRP and XCIRP2).
We confirmed the specificity of the morpholinos towards their target-mRNA. Sequence alignment revealed that the subtype specific morpholino oligonucleotides although perfectly matching their target mRNA contain at least 7 mismatches to the mRNA of the other family members (Fig. 1C). In immunoblots we showed, that all three morpholino oligonucleotides suppress translation of their target mRNAs (Fig. 1D, E, F) and that the XTcf-3 specific morpholino has only minor effects on XTcf-4 translation (Fig. 1E) and vice versa (Fig. 1F). Consistently, also the RNA-levels of XTcf-3 and XTcf-4 in neuralized animal caps remained unchanged following morpholino injection (Fig. 2A) and the main XTcf-4 isoform in neuralized animal caps is XTcf-4B, regardless whether a morpholino was injected or not (Fig. 2C).
However, analysis of XLef-1 and XTcf-1 in XTcf-3- and 4 depleted neuralized animal caps (Fig. 2A) and in caps overexpressing Lef/Tcfs (Fig. 2B) revealed a complex cross-regulation of these transcription factors. Expression of XTcf-1 and XLef-1 depended on the presence of XTcf-3 and XTcf-4 and overexpression of XTcf-3 induced both, XTcf-1 and XLef-1 (Fig. 2B). XTcf-4 overexpression, however, induced XLef-1 but repressed XTcf-1 (Fig. 2B). The complexity of this cross-regulation is further sustained by the finding that XTcf-3 expression is specifically upregulated following XLef-1 and XTcf-1, but not XTcf-4 mRNA injection, regardless whether an activating (XTcf-4C) or repressing (XTcf-4A) isoform was used. Regulation of the direct wnt target gene XTcf-4 by Lef/Tcfs is seen by the induction of XTcf-4 following XTcf-3, XTcf-1 and XLef-1 overexpression (Fig. 2B).
Thus, a complex cross-regulation of Lef/Tcfs seems to complicate the analysis.
The molecular mechanism of XCIRP, regulating proper development of neural tissue is most likely due to its function as RNA-binding protein. XCIRP is known to stabilize a large number of different mRNAs, including N-cadherin, E-cadherin, β-catenin, cyclin B1, Nek2B, Cytochrome C oxidase peptide VIb and bystin [19, 21]. In order to determine, whether a small subset of mRNA or mRNA in general is stabilized by XCIRP, we compared the mRNA/RNA relation of control embryos with XCIRP depleted ones. Indeed, following XCIRP morpholino injection, the mRNA pool was reduced by 10–15% (Fig. 6B, C) indicating that XCIRP as one of the most abundant RNA binding proteins in early Xenopus development is involved in stabilizing many different mRNA populations.
Thus, we identified in here the RNA binding protein XCIRP as novel Tcf-subtype specific target gene, which by stabilization of a large pool of mRNAs allows proper anterior neural development.
In a subtractive screen specific for Tcf-subtype specific target genes one of the most frequently candidate genes, cold inducible RNA binding protein XCIRP, was found to be regulated by XTcf-3, but not by other Lef/Tcfs. Cement gland specific marker genes, although not co-localized with XTcf-3 in the embryo, have been shown to be repressed by injection of dominant negative XTcf-3 . Consistently, among our XTcf-3 target genes we identified the cement gland specific marker genes XAG and XAG2. Other candidate genes enriched in the subtractive screen are supposed to be expressed rather in ectodermal than in neuroectodermal tissue (α-tubulin, cytokeratine type II, cytokeratine 81 [25, 26] or their expression did not strictly depend on the presence of XTcf-3 (HMGN1, HMGN2, HMGX). Interestingly, we did not enrich any "classical" wnt target gene in our screen, which might be explained by their partial redundant roles in transducing wnt/β-catenin signaling or by cross-regulation of their expression. Based on our results, one could suspect, that general wnt/β-catenin target genes are not enriched in the subtractive screen, because both, the XTcf-3 depleted and the XTcf-4 depleted pool also contained less XTcf-1 and XLef-1. Thus, only subtype-specific target genes will be detected, while general wnt/β-catenin target genes are underrepresented.
Although differential activation of wnt/β-catenin target gene promoters by individual Lef/Tcf family members have been reported over the last years [9, 24, 27, 28, 29], most studies dealing with wnt/β-catenin signaling still unite Lef-1, Tcf-1, Tcf-3 and Tcf-4 as Lef/Tcfs and implicate that they are exchangeable. Thus far, non-sequence specific DNA binding and recruitment of p300 by the Tcfs E-tail [30, 31] and epigenetic modifications of wnt target gene promoters  provide the best explanations for subtype specific target gene regulation. Recently, it has been shown in Drosophila, that pangolin/armadillo also promotes target gene repression by binding to DNA sequences different to the "classical" Lef/Tcf binding side . Whether a similar mechanism also regulates Tcf-subtype specificity in vertebrates remains elusive.
With XCIRP, we describe for the first time a novel Lef/Tcf target gene, which is endogenously regulated in a Tcf-subtype specific manner. XTcf-3 but not XTcf-4 and XLef-1 is required for the XCIRP expression in the neural plate.
Furthermore, XCIRP expression can be restored by XTcf-3 but not by XTcf-1 and XTcf-4, indicating that Lef/Tcfs are not exchangeable. However, the fact that XLef-1 can restore XCIRP expression in a similar extent as XTcf-3 revealed some redundant functions of XTcf-3 and XLef-1. This redundant function is somehow unexpected, because we recently could show that XTcf-3 in general behaves as a repressor and XLef-1 as an activator [27, 28]. Keeping in mind the redundant functions of murine Lef/Tcfs revealed by loss of function studies and double knockouts in mice [7, 8] it is surprising, that we could detect such specific regulation of XCIRP expression by individual Lef/Tcfs, even more, since it has been shown in different contexts that the expression of Lef/Tcf transcription factors is regulated by Lef/Tcfs (this study, 33,34,35). However, this complex cross-regulation provides an explanation for the unexpected reconstitution of XCIRP expression following XLef-1 injection in XTcf-3 depleted embryos: Because overexpression of XLef-1 upregulates XTcf-3, the recovery of XCIRP might be a secondary effect and still due to a subtype-specific regulation by XTcf-3. ChIP with XTcf-subtype specific antibodies will help to define whether endogenous XCIRP, indeed, is directly regulated by XTcf-3 and not by XLef-1.
Lef/Tcf transcription factors as transducer of the wnt/β-catenin signaling pathway have been reported to regulate anterior posterior patterning of the CNS. While activation of wnt/β-catenin signaling in general posteriorizes the CNS , wnt-2b/XTcf-4 regulates the development of the midbrain . Here we show that XTcf-3 is required for proper development of the neural plate. Depletion of XTcf-3 results in a lateral expansion of the pan-neural marker gene sox2, the midbrain specific marker gene Meis3 and to a shift of the placodal marker gene eya1. A similar broadening of the anterior neural plate was observed following injection of dominant negative XTcf-3 . Thus, inhibition of wnt/β-catenin signaling is required for neural induction and injection of dn XTcf-3 is thought to mimick the effect of endogenous wnt inhibitors. In contrast to the dominant negative XTcf-3 in Heeg-Truesdells study , the XTcf-3 morpholino did not result in neural induction. This might be due to the fact that the N-terminal deletion mutant competes with all Lef/Tcfs, thus is not strictly specific for XTcf-3.
A similar broadening of the neuroectoderm was also observed following depletion of the transcription factors zic-1  and six1 . In contrast to the depletion of zic-1 and six1, the broadening of the anterior neural tissue following depletion of XTcf-3 did not go ahead with a reduction of placodal marker genes, thus, did not affect the organization of the neural plate border. Although wnt/β-catenin signaling is involved in the induction of cranial placodes , the expression of the placodal marker gene eya1 seems to be unaffected by depletion of XTcf-3. This opens up the question, how XTcf-3 can lead to a broadening of the neural plate without affecting the integrity of the adjacent placodal region. We provide evidence, that stabilization of a large pool of mRNAs by the XTcf-3 target XCIRP might be an explanation.
Depletion of the XTcf-3 target gene XCIRP resulted in a similar lateral expansion of neural tissue, indicating that XCIRP is an important downstream component of XTcf-3 for neural development. One obvious mechanism for this broadening would be a failure in convergent extension movements, which might be caused by a destabilization of the mRNA of cell adhesion molecules including C- and E-Cadherin, β-Catenin, and α-Catenin . According to our dorsal marginal zone explants, we can exclude an effect on convergent extension movements, because injection of the CIRP morpholino did not result in an inhibition of elongation or constriction of the explants. The molecular mechanism of XCIRP action during early embryonic development is most likely due to a stabilization of mRNA. CIRP has been shown to bind to and stabilize many different mRNAs, including those encoding for cell adhesion molecules and cell proliferation regulators [19, 21]. Thus, CIRP is supposed to act as a general mRNA binding protein, which regulates the stability of a large pool of different mRNAs. Consistently, we can show that the total mRNA pool is decreased following XCIRP morpholino injection, indicating that apart the thus far identified XCIRP target mRNAs a big pool of additional mRNAs is regulated by XCIRP. Thus, we conclude that the XTcf-3 target gene XCIRP regulates the correct positioning of the neural border by regulating the stability of many different mRNAs. We think that deciphering the identity of the XCIRP regulated mRNA pool will give enormous progress in understanding the mechanisms underlying early neural development and the positioning of the neuroectoderm/ectoderm border. Furthermore, the specific effect of XCIRP depletion on neural development together with the distinct expression domains of different RNA binding proteins including XCIRP, SEB4 , nrp-1  and elav-like proteins  might reflect a general mechanism in development: stabilization of a distinct cell fate by regulating the stability (or translation) of a broad variety of mRNAs in a tissue specific manner.
In a screen for Tcf-subtype specific target genes we identified cold inducible RNA binding protein XCIRP as novel XTcf-3 specific target gene. Both, XTcf-3 and XCIRP are required for proper anterior neural development. We provide evidence, that RNA stabilization by the XTcf-3 specific target gene XCIRP is essential for anterior neural development.
Plasmids, constructs, in vitrotranscription
Capped mRNAs were transcribed from linearized DNA templates using mMESSAGE mMACHINE (Ambion). Digoxigenin labeled antisense probes for in situ hybridization were synthesized with DIG RNA labeling kit (Roche). Probes for in situ hybridization are described elsewhere: sox2 , Tcf-4 , HMGN1 and HMGN2 , HMGX , Meis3 and eya1 . The open reading frame of XCIRP was amplified using the following primers: XCIRPstart: 5'-tcagaattcaatgtctgacgaaggaaaact-3' and XCIRPstop 5'-cttctcgagttactcgtgtgtagcatagctg-3' and sub-cloned into pGEMT for creating labeled antisense RNA and into pCS2 for synthesizing myc tagged mRNA. cDNA corresponding to aa 63–328 of XTcf-3  was inserted into pGEMT for creating labeled antisense RNA. The following antisense morpholino oligonucleotides (Gene Tools, LaJolla) were used: XTcf-3 (Tcf3Mo, MO#-Tcf7l1): 5'-cgctgttgagctgaggcatgatgag-3' (directed against BC077764); XTcf-4 (Tcf4Mo, MO#-tcf4): 5'-cgccattcaactgcggcatctctgc-3' (directed against BC106476), XLef-1 LefMo1, MO#-LEF1): 5'-tacggagtcgagagacctcgtccc-3' and (LefMo2, MO#-LEF1): 5'-caagacctcggtacggagtcgagag-3' (directed against NM_001088655 und AC151234) and XCIRP (CIRP-Mo, MO#-CIRBP): 5'-agtacagactgcttccttttgaga-3' (directed against XCIRP NM_00108600, XCIRP1 AF278702 and XCIRP2 NM_001086325). The concentrations injected are given in the figure legends and were tested to be non-toxic.
Selective subtractive screen
Eggs from HCG treated females were fertilized by standard methods and staged according to . For animal cap explants eight picomoles of XTcf-3 and XTcf-4 morpholino where co-injected with 100 pg mRNA encoding for a truncated version of the BMP receptor (tBr) into both blastomeres of 2-cell stage embryos. At late blastula, animal caps where explanted and cultivated until sibling embryos reached stage 16. The transcriptome of Tcf-3 depleted neuralized animal caps was subtracted from the transcriptome of Tcf-4 depleted ones using the "PCR-Select cDNA Subtraction Kit" (BD Biosciences, Heidelberg, Germany) according to manufacturers description.
Embryo treatment and in situhybridization
For marker gene analysis, the indicated morpholinos where co-injected into one blastomere of Xenopus 2-cell stage embryos with 4 pg dextrane or 100 pg mRNA encoding for myc-tagged GFP as lineage tracer. At neurula stages, embryos were fixed in MEMFA (0.1 M MOPS pH 7.2, 2 mM EGTA, 1 mM MgSO4, 3.7% formaldehyde). For vibratome sections, fixed embryos were embedded in 3% agarose and sectioned using a Leica VT 1000S vibratome. Whole-mount in situ hybridization was performed according to previously described procedures . Localization of mRNA was visualized using anti-digoxigenin antibodies conjugated to alkaline phosphatase, followed by incubation with nitro blue tetrazolium (NBT) and 5-bromo 4-chloro 3-indolyl phosphate (BCIP). Images were captured on a Leica MZFLIII microscope using a digital camera (Qimaging) and Improvision software (Openlab).
For analyzing convergent extension movements, dorsal marginal zone explants were cut at blastula stages, cultivated until siblings reached stage 16 and analyzed according to .
RIPA lysates corresponding to one half embryo were separated by SDS-PAGE and transferred onto nitrocellulose. After blocking, the membrane was incubated with α-myc (9E10) or α-CIRP antibody (kindly provided by Ken Matsumoto) and peroxidase-coupled secondary antibody. Proteins were visualized using ECL-Plus Western Blotting Detection System (Amersham). As loading control, blots were incubated with α-Gemin antiserum kindly provided by U. Fischer, or a parallel Gel was stained with Coomassie brillant blue.
Marker gene analysis
For marker gene analysis by RT-PCR, total RNA was extracted from whole embryos or animal caps, reverse transcribed using random hexamer primers and MMLV reverse transcriptase. For amplification the following primers were used. Histone 4: 5'-cgggataacattcagggtatcact-3' and 5'-atccatggcggtaactgtcttctt-3'; XCIRP: 5'-gctgatcaggcggggccacc-3' and 5'-gcacccaggctctgtcctgc-3'; XCIRP2: 5'-ccattcaggctgatcagg-3' and 5'-ctggagagagacgaacac-3'; XAG: 5'-gagttgcttctctggca-3' and 5'-ctgactgtccgatcagac-3'; XANF: 5'-actgacctacaagagagaa-3' and 5'-agtgcatcattgttccacag-3'; ODC: 5'-tggcagcagtacagacagca-3' and 5'-gatgggctggatcgtatcct-3'; XTcf-1: 5'-ctcggctcacatgga agatg-3' and 5'-cagatcgacaggaaggtg tc-3'; XLef-1: 5'-ggagacctcgcagatatcaa-3' and 5'-ccagccc aatggtggtgtcat-3'; XTcf-3: 5'-catggcagcatgctcgactc-3' and 5'-ctggtcactagagaaagggg-3'; XTcf-4: 5'-ctcacgccgctcattacctacagcaac-3' and 5'-catgtacagcatgaacgcgtttagggg-3'; HMGN1: 5'-ggct cctgcttctgatggag-3' and 5'-ccagtatcaatctctgcctg-3'; HMGN2: 5'-ggggacacaaagtgaagc-3' and 5'-cacagtcgaagtaccgcc-3' and HMGX: 5'-cgtccccagataaagagcgag-3' and 5'-gctgcactcgagttcac attc-3'. NCAM: 5'-ggaatcaagcggtacaga-3' and 5'-cacagttccaccaaatgc-3'. The XTcf-4 isoforms were determined by digesting the amplicons with RsaI and XbaI as previously described .
mRNA/total RNA ratio
For analyzing the mRNA/total RNA ratio, 1 μg RNA of stage 18 embryos was reverse transcribed with MMLV reverse transcriptase and random hexamer primers (total RNA) or oligo-dT primers (mRNA) in the presence of 5 μCi [32P]-α-dCTP (3000 Ci/mmol). Radioactive transcripts were separated from unincorporated [32P]-α-dCTP nucleotides by denaturating polyacrylamide gel electrophoresis. Quantity of transcripts was determined via intensity of radioactive signals (Canberra Packard) and normalized to total radioactivity.
We thank Karolin Rahm for technical assistance, Isabell Albert for initial characterization of the XLef-1 antisense morpholino oligonucleotides and Bianca Kraft for helping with the dorsal marginal zone explants. The work was funded by the DFG grant GR1802.
- 17.Munoz-Sanjuan I, Bell E, Altmann CR, Vonica A, Brivanlou AH: Gene profiling during neural induction in Xenopus laevis: regulation of BMP signaling by post-transcriptional mechanisms and TAB3, a novel TAK1-binding protein. Development. 2002, 129: 5529-5540. 10.1242/dev.00097.CrossRefPubMedGoogle Scholar
- 21.Peng Y, Yang PH, Tanner JA, Huang JD, Li M, Lee HF, Xu RH, Kung HF, Lin MC: Cold-inducible RNA binding protein is required for the expression of adhesion molecules and embryonic cell movement in Xenopus laevis. Biochem Biophys Res Commun. 2006, 344: 416-424. 10.1016/j.bbrc.2006.03.086.CrossRefPubMedGoogle Scholar
- 46.Nieuwkoop PD, Faber J: Normal tables of Xenopus laevis (Daudin). 1967, Elsevier North-Holland Biomedical Press, AmsterdamGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.