If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
* This work was supported by the European Union Framework V program for finance. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
slug gene expression is associated with the specification and migration of neural crest cells in the African clawed frog Xenopus laevis. We provide evidence that the protein Ying-Yang 1 (YY1) regulates the slug gene expression both indirectly and directly, via a YY1 cis-element in the slug promoter, during Xenopus development. The ability of the YY1 to bind this YY1 cis-element was confirmed by electromobility shift assays and reporter assays. YY1 was detected in the nuclei of ectodermal cells contemporaneously with the process of neural crest specification. The injection of anti-YY1 morpholino, which targeted both YY1α and YY1β gene products, depleted YY1 expression below 20% and was lethal at gastrulation. Sublethal depletion of YY1 reduced the length of the anterior-posterior axis and severely inhibited the expression of the neural marker Nrp1 and of the slug gene. Overexpression of YY1 or mutation of the YY1 cis-element reduced the restricted spatial expression of the slug reporter gene in the neural ectoderm border and provoked its expression in the nonneural ectoderm. Chromatin immunoprecipitation indicated that endogenous YY1 interacts directly with the YY1 cis-element of the endogenous slug gene and with the slug gene reporter sequence injected into embryos. The results suggest that YY1 is essential for Xenopus development; is necessary for neural ectoderm differentiation, a prerequisite for neural crest specification; and restricts which cells can form neural crest mesenchyme through directly blocking slug gene activity.
The restricted expression of slug mRNA in the border between the prospective neural and the non neural ectoderm corresponds to the area from which neural crest (NC)
). The injection of slug mRNA into Xenopus embryos increases the population of mesenchymal cells at the neural fold and the number of melanocytes (a NC cell derivative) at a later stage. Inhibiting slug expression in Xenopus and chicken can decrease NC cell numbers and deter their motility (
). At intermediate levels of BMP activity, ectodermal cells express markers for the anterior positioned cement gland. In the second stage, ectodermal cells experience signals that push differentiation in a more posterior direction. These signals, which may include Wnts, fibroblast growth factors, and retinoic acid, provoke those cells that experience intermediate levels of BMP activity into expressing slug.
How these signals directly affect the slug gene remains unclear. By analyzing the Xenopus slug gene promoters (
), one candidate we have identified is the DNA binding factor Ying-Yang 1 (YY1). YY1 has been described as having three related functions. It can function as a transcription factor (either activator or repressor, hence its name) and transcription initiator (
). Its function is modulated by its capacity to bind many other proteins. And interestingly, YY1 has been found to interact with Smad4, -1, and -2, such that it was able to inhibit the TGFβ-induced EMT in NMuMG cells (
YY1 has been found to participate in embryonic development. The homozygous YY1 gene knockout prevents murine embryos surviving beyond the blastocyst stage. Modulating YY1 expression in Xenopus and in mice by heterozygotic gene knockdown leads to effects on neural patterning (
). Here we present evidence that YY1 is essential for Xenopus laevis development and that its functional repertoire includes regulating the formation of the NC, in part through a direct interaction with the slug gene.
Oligonucleotides, Morpholino-modified Oligonucleotides (MOs), Antibodies, and cDNA Plasmids—Oligonucleotides were purchased from Eurogentec, France. The sequences for the oligonucleotide pairs used (design aided by Primer 3; available on the World Wide Web at frodo.wi.mit.edu/cgi-bin/primer3/primer3_www.cgi) were as follows: 5′-2.8kb sense, 5′-TGGAA ATGAA CGTGA ACTGG-3′; 5′-2.8kb antisense, 5′-CTTTC ACAGG CTGAG GCATT-3′; YY1 site sense, 5′-AGGACTCGGGTCTTCCTCTC-3′; YY1 site antisense, 5′-CAAGC AGCCT CACAG ACTGA-3′; pSB-200-GFP plasmid sense (used with YY1 site antisense), 5′-CTCAC TATAG GGCGT CGACA CT-3′; slug sense, 5′-ATTTC CATAT GGCAG CAAGG-3′; slug antisense, 5′-TGCCT GATTT AAACG CAACA-3′; snail sense, 5′-TCACA AAGGC AGTGC TTCAC-3′; snail antisense, 5′-TTGTT CTCTG TGCCA ACTGC-3′; Nrp1 sense, 5′-GACAT TGTGG ACGGT CAGTG-3′; Nrp1 antisense, 5′-GCTAC CTTTG CCGAC AGTTC-3′; twist sense, 5′-CTCAA CGAAG CCTTC TCGTC-3′; twist antisense, 5′-CGATG TATCT GGAGG CCAGT-3′; AP2 sense, 5′-TCCCA ACAGC CATAC AGACA-3′; AP2 antisense, 5′-AGTTG GTGGC TGCAG AAAGT-3′; ODC sense, 5′-GTCAA TGATG GGTGT ATGGA TC-3′; ODC antisense, 5′-TCCAT TCCGC TCTCC TGAGC AC-3′. MOs were purchased and designed together with Genetools USA: anti-YY1, (MOyy1), 5-GCGTG TCGCC CGATG CCATG TTCAT-3′; and the anti-YY1 control oligonucleotide sequence containing five mismatches (MOcon), 5′-GCGTC TCCCC CAATG CCATC TTGAT-3′. Rabbit anti-YY1 antibodies sc1703 (batch K029) and sc1703x (batch C301) and rabbit anti-HA antibody sc805x (batch Y11) (used as control) were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Rat anti-HA 3F10 (used in the immunofluorescence) was purchased from Roche Applied Science. The CS2 plasmid expressing ΔNYY1 (amino residues 18–373) was a kind gift from Dr. Arie Otte (University of Amsterdam).
Plasmid Construction and RNA Synthesis—The expression vector pCS2-YY1 was constructed from pCS2-ΔNYY1 (start methionine 18) with the insertion of a PCR product (derived from Xenopus embryo cDNA). The YY1 constructs were subcloned into a pCS2 vector (pCS2-HA) engineered to express the hemagglutinin antigen (HA) tag (additional sequence subcloned between BamHI and XhoI sites: GGATC GATGG CATAC CCATA CGACG TCCCA GACTA CGCTG GATCC GCCAT GGACT AGTCT CGAG) on the N terminus of the protein. pCS2-YY1rescue was constructed by subcloning the PCR product spanning the coding region of YY1 back into an empty pCS2 vector, which included using sense primer containing five silent point mutations: 5′-AGGGA TCCAT GAATA TGGCC TCAGG AGATA CGCTC TACAT CGCC-3′. The reporter plasmid pSB200-GFP (green fluorescent protein) (
) was subject to point mutation alterations using the materials and methods supplied by Stratagene's QuikChange™ kit. The sequences used to design the oligonucleotides were as follows: YY1 cis-element (200Y*), 5′-CTGAA ACCCG AGACA AAATC TTACA CGATT GGACG-3′; and the GAGA cis-element, 5′-GAGGA TCTGA AACCC GACTC ACCAT CTTAC ACGA-3′ (200G*). The in situ hybridization probes were generated using the complete coding region templates of slug and Nrp1 cloned into pCS2, respectively. Plasmid integrity was verified by DNA sequencing, performed by the Service Sequençage (Institut Curie, France). Antisense RNA probes were generated by run-off in vitro translation using digoxygenin RNA labeling mix (Roche Applied Science) and T7 or Sp6 RNA polymerase (Promega). Capped RNA for injections was generated and purified from linearized plasmids using the mMessage Machine™ (Ambion) and RNeasy™ columns (Qiagen) and the accompanying protocols, respectively.
Embryos and Injections—Xenopus egg preparation and injection procedures were performed in accordance with those described previously (
) stage 1, 30–90 min after fertilization (maintaining the embryos at 14 °C). The order of plasmid injection was swapped between experiments. The embryos were analyzed by fluorescence microscopy at NF stage 17–19. For promoter activity reporter assays, embryos were injected in the dorsal blastomere at NF stage 3 with a 10-nl solution containing pCS2 expression vector with or without (control) inserted cDNA and 130 pg of reporter vector. Promoter activity was quantified by calculating the percentage of fluorescent embryos from 20 or more injected embryos that survived to NF stage 17–19. Means and S.D. values were calculated from the percentages derived from at least three manipulations performed on different embryo batches, resulting in more than 100 embryos being screened in a given experiment. RNA, MO, and other DNA injections were performed at stages 1 and 3 dependent on the experiment in volumes no greater than 20 nl and in amounts stated under “Results.”
Immunofluorescence, Western Blots—Xenopus embryos were fixed by gentle agitation for 25 min in a 0.1 m sodium phosphate buffer, pH 7.4, containing 4% paraformaldehyde followed by 60 min of gentle agitation in 0.1 m sodium phosphate buffer, pH 7.4, containing 0.15 m sucrose. The embryos were dehydrated in graded ethanol solutions. Fixed embryos were embedded into paraffin wax blocks. Eight μm sections were treated for antigen recovery by incubation in 0.1 m sodium citrate, pH 8.0, at 100 °C for 20 min followed by acclimatization to room temperature for 40 min. Sections were incubated in a blocking buffer (10 mm maleic acid, 15 mm NaCl, 20 ng/ml glycine, 500 ng/ml heparin (Sigma), 5 ng/ml levamisole (Sigma)), including 10% dialyzed fetal calf serum (Diachron), 0.2% Triton X-100 in conjunction with the Vector Biotin Blocking System™ (Vector Laboratories), followed by similar repeated incubations with the blocking buffer containing the same constituents but including 2% dialyzed fetal calf serum, 0.4% Triton X-100, and antibody. Biotin anti-rabbit secondary with a fluorescent tertiary streptavidin-linked fluorophore was used to detect the YY1 primary antibody. Immunofluorescence detection was enabled by using AlexaFluor488™ (green) or AlexaFluor594™ (red) secondary conjugated antibodies/streptavidin. Western analysis was performed using standard protocols using SDS-polyacrylamide gels for the electrophoresis separation of proteins and with blotting onto Immobilon™ membranes (Millipore Corp.). Antibody-mediated detection of the proteins was visualized by ECL™ chemiluminescence (AP Biotech).
Nuclei Preparation, Electromobility Shift Assay (EMSA), and Chromatin Immunoprecipitation (ChIP)—Embryos were subjected to mild lysis in 0.15 m sucrose in E50 buffer (20 mm HEPES-KOH, pH 7.6, 50 mm KCl, 10 mm sodium β-glycerophosphate, 5 mm MgCl2, 0.1 mm EDTA, 1 mm dithiothreitol, 0.1 mm spermine, Complete Protease Inhibitors® (used in accordance with manufacturer's instructions; Roche Applied Science). The nuclei suspension was centrifuged twice through an equal volume of a sucrose cushion (0.9 m sucrose in E50 buffer). For nuclear extracts, the nuclei were prepared by resuspension in 150 μl/50 embryos of 25% glycerol in E50 buffer, followed by lysis of the nuclei by vortexing. Debris was removed by centrifugation. In the EMSAs, double-stranded oligonucleotide probes containing 4-nucleotide 5′-overhangs (CTAG sense strand, GAGT antisense strand; see Fig. 1A) were radiolabeled with [α-32P]dCTP using Klenow DNA polymerase (AP Biotech). Protein-DNA complexes were prepared in a 10-μl volume containing 20,000 cpm of labeled oligonucleotide, 5 μl of nuclear extract, 1 μg of poly(dI-dC) (Sigma) in E75 buffer (same as E50 but with 75 mm KCl) and incubated for ∼60 min. When necessary, 5 ng of unlabeled competing oligonucleotides or 2 μg of antibody were also included. Protein-DNA complexes were size-separated at 4 °C by electrophoresis in 5% polyacrylamide gels containing 0.5× TBE and 10% glycerol in 0.5× TBE running buffer. Radiolabeled oligonucleotides were detected in the dried gels using Eastman Kodak Co. autoradiographic film.
Nuclei for ChIP analysis were resuspended in 0.5 ml/200 nuclei of 0.15 m sucrose E50 buffer and fixed by gentle agitation at room temperature for 30 min with the further addition of 0.5 ml of 0.1 m sodium phosphate buffer, pH 7.4, containing 4% paraformaldehyde. The nuclei were resuspended in 1 ml of lysis buffer (1% SDS, 50 mm Tris, pH 8.0, 10 mm EDTA, 1 mm phenylmethylsulfonyl fluoride, Complete Protease Inhibitors®) and sonicated on ice. The lysate was precleared, which included incubation with protein G-Sepharose. An aliquot of lysate was removed for the assessment of the input. The lysate was incubated with gentle agitation at 4 °C, with 5 μg of antibody overnight and then with protein G-Sepharose for 2 h. The protein G-Sepharose was then isolated by centrifugation and washed in radioimmune precipitation assay buffer (10 mm Tris, pH 8.0, 140 mm NaCl, 1% Triton X-100, 0.1% SDS, 1% deoxycholic acid); radioimmune precipitation assay buffer containing 500 mm NaCl instead of 150 mm; LiCl lysis buffer (10 mm Tris, pH 8.0, 1 mm EDTA, 250 mm LiCl, 0.5% Nonidet P-40, 1% deoxycholic acid); and Tris-buffered saline. The protein G-Sepharose was resuspended in a buffer containing (0.8% SDS, 50 mm Tris, pH 8.0, 10 mm EDTA) and incubated for 4–6 h at 65 °C followed by overnight treatment with Proteinase K (Qiagen) at 55 °C. A similar dilution and treatment was prepared with the input sample. The DNA was then purified by a typical phenol/chloroform procedure.
In Situ Hybridization and TdT-mediated dUTP Digoxigenin Nick End Labeling (TUNEL) Assay—Whole mount in situ hybridizations and TUNEL assays were performed following previously published procedures (
Reverse Transcription (RT) and PCR—Total RNA was extracted using Tri-Pure reagent (Roche Applied Science). First strand cDNA was subsequently synthesized using Superscript KSII polymerase (Invitrogen), primed with an oligo(dT)15 according to the manufacturer's instructions. PCR assays were performed in the exponential phase of amplification as described elsewhere (
) using Tfl polymerase (Promega), using Vistra Green® (AP Biotech) for visualization. Real time quantitative PCR (qPCR) was performed using SYBR® Green reagent with an ABI Prism® thermal cycler (Applied Biosystems) at an annealing temperature of 58 °C. With cloning and some ChIP analyses, a conventional PCR protocol for detection was used with either Dy-NAzyme® polymerase (Finnzymes) or Pfu polymerase (Stratagene).
Three conserved cis-elements were identified in the proximal region of the slug gene, by comparing the Xenopus sequences with the human (
R. K. Wilson (Genome Sequencing Center, Washington, D. C.), Gen-Bank™ accession number AADN01074414.
slug gene sequences (Fig. 1A). These cis-elements include a putative binding site for YY1 (consensus CGCCATNTT), an E-box (consensus MYGCACCTGA potentially capable of binding Slug and Snail; data not shown), and a TATA box (
). The E-box consistently appeared within 30 bases of the putative TATA box sequence (except in the human gene, where a TATA box motif was separated from this E-box by an extra 48 bp) and provided a useful landmark for comparing the slug gene promoters. The E-box and TATA box were also identified in two fish gene promoters (Takifugu rubripes and Tetraodon nigroviridis; not shown). The distance between the more upstream YY1 cis-element and the E-box ranged approximately between 130 bp (Xenopus), 180 bp (mammalian), and 450 bp (chicken; Fig. 1B). A GAGA factor cis-element was also identified adjacent to the YY1 cis-element in the Xenopus sequences (Fig. 1A, underlined) (
The potential for YY1 to bind the Xenopus and murine cis-elements was confirmed by EMSAs using double-stranded oligonucleotides as probes, spanning 22 nucleotides of the YY1 cis-element and flanking regions (see Fig. 1A). Retarded YY1-DNA complexes from the electrophoresis were identified in samples containing the Xenopus YY1 cis-element radiolabeled probe (WT), incubated with Xenopus NF stage 18 nuclear extracts (Fig. 1C, lane 1). The specific association of YY1 with the cis-element was confirmed, because the appearance of one complex was inhibited by the prior addition of excess unlabeled probe (WT) but not by excess unlabeled double point-mutated probe (Y*; Fig. 1C, lanes 3 and 4). The mobility of this complex was lowered by the addition of an anti-YY1 antibody (αY) but not with an anti-HA antibody control (Fig. 1C, lanes 1, 2, 12, and 13), and the appearance of this lower mobility “supershifted” complex was also inhibited by the addition of excess unlabeled probe but not by excess mutated probe (Fig. 1C, lanes 5 and 6). A similar profile of probe retardation was observed when using a sequence from the murine slug gene YY1 cis-element (mWT; Fig. 1C, lanes 7–11). The appearance of the YY1-Xenopus WT probe complex was only partially inhibited with the addition of excess unlabeled probe double point-mutated in the GAGA sequence (G*; Fig. 1C, compare lanes 14 and 16). When either the radiolabeled Y* probe or the G* probe were used, YY1-DNA complexes were not identified (Fig. 1C, lanes 12–14). A YY1-cis-element-specific complex of similar mobility was observed when Xenopus YY1 was synthesized using an in vitro reticulocyte lysate method (and this YY1 protein also had the same mobility by Western analysis; these data are not shown).
The interaction between endogenous YY1 and the endogenous slug gene was confirmed by ChIP assays, at a developmental stage when slug expression is first observed in the neural plate border. Nuclear preparations from NF stage 12 embryos were fixed and sonicated to fragment the DNA to roughly below 1000 bp (Fig. 2A). Antibodies against YY1 or HA (as a negative control) were used to pull down protein-DNA complexes. This DNA was purified, and PCR was used to either detect a 120-bp fragment flanking the slugα gene YY1 cis-element or a 100-bp fragment (in a conserved region in Xenopus genes) 2.8 kb upstream and assumed not to bind YY1 (due to the absence of YY1 consensus motifs). DNA fragments of the expected sizes were detected in the samples subjected to 35 cycles of PCR (Fig. 2B). qPCR quantification of the same samples gave similar results (Fig. 2C). The ratio of detectable YY1 target site DNA to the control region DNA (normalizing for their respective input amounts) was equal to 3.8 in the sample subjected to anti-YY1 pull-down, whereas this ratio was equal to 0.8 in the sample subjected to the control antibody (anti-HA pull-down). In further repetitions of samples subjected to αYY1 pull-down, the ratio was calculated as 7.5, 14, and 24, giving an average of 12 over the four experiments.
The ability of YY1 to directly regulate the slug gene promoter was demonstrated by reporter gene assays in Xenopus embryos (Fig. 2D). Plasmid DNA was injected into NF stage 3 embryos. The reporter plasmids were based upon 200 bp of the slug gene promoter upstream of the GFP reporter gene (pSB-200-GFP; 200WT) and included versions mutated to disrupt YY1 binding (200Y* or 200G*). No significant difference in the number of fluorescent embryos was observed between those embryos injected with different reporter plasmids, suggesting that endogenous YY1 levels are not able/sufficient to regulate the transcriptional activity of the ectopic plasmid reporter. However, the co-injection of a YY1-expressing plasmid caused a significant reduction in 200WT reporter activity compared with 200Y* activity. This reduction was also significant and greater when an N-terminal truncated mutant of YY1 (ΔNYY1) was co-injected with 200WT versus 200Y* or with 200WT versus 200G* (Fig. 2D) and was concentration-dependent (Fig. 2E). The N-terminal domain has been shown to have an activating role in cell culture models, so its removal concurred with this enhanced repressive activity (
The spatial and temporal expression profile of YY1 was compatible with a function of regulating the slug gene. YY1 was detected with an estimated size of 65 kDa (greater than the 41.4 kDa predicted from the amino acid sequence) by Western analysis of nuclear extracts from NF stage 8 to NF stage 22 embryos (Fig. 3A, upper panel). YY1 was barely detectable at NF stage 8–9 but was clearly detectable by NF stage 12. In the same extracts, PCNA could also be detected at similar levels to YY1 at each stage (Fig. 3A, lower panel). Quantification of the Western analysis suggested that there was a significant elevation of YY1 relative to PCNA at stages 15 and 22 compared with NF stage 8–9 (p < 0.05; Fig. 3A, graph).
YY1 was detected primarily in the nuclei of Xenopus embryonic cells in the developmental stages examined from NF stage 8 to 22 (Fig. 3, B–M, and data not shown) by immunofluorescence (and by horseradish peroxidase immunohistochemistry, also not shown) on 8 μm paraffin-embedded embryo sections. The nuclear localization of YY1 was confirmed by examining embryos that were injected with plasmid DNA coding for HA-tagged YY1. Elevated nuclear YY1 detected by anti-YY1 antibody co-localized with the protein detected by the anti-HA antibody in NF stage 12 lateral ectoderm (Fig. 3, B–E). Confocal microscopic detection of YY1 was similar to the detection of DAPI, although it appeared slightly more punctate than DAPI (Fig. 3, F–H). YY1 was detected in the endoderm, mesoderm, and ectoderm at NF stage 10 and NF stage 12. At NF stage 12, YY1 was detected at high levels (relative to DAPI fluorescence) in the nuclei of endodermal cells compared with the other cell types. YY1 levels were higher in the nuclei of lateral and ventral nonneural ectoderm relative to the prospective neural ectoderm and mesoderm (Fig. 3I). Also, in the region of the dorsal blastopore lip, YY1 was detected at higher levels in the nuclei of cells in the surface epidermis relative to the underlying mesoderm (Fig. 3, J and K), with a similar but not so obvious difference at stage 10 (Fig. 3, L and M).
The effect of YY1 upon the endogenous slug gene expression was examined by depleting YY1 levels using MO injections. The MOyy1 was designed to target the two mRNA isoforms revealed by expressed sequence tag sequence alignment (Fig. 4A). Two expressed sequence tags contained sequences identical to the 5′-untranslated region (5′-UTR) of the known YY1 sequence (
), whereas two expressed sequence tags were found to differ from this. Accordingly, the known gene was termed YY1α, and it was assumed that the nonhomologous sequences were derived from the YY1β gene (because of the pseudotetraploidal nature of X. laevis). The YY1β gene also differed from YY1α, such that a PCR fragment generated from nucleotides 160–260 in the coding region sequence could be cut twice by the restriction enzyme Sau3A instead of only once (Fig. 4B). Both gene products were detected in Xenopus cDNA samples (Fig. 4C). Quantitative analysis suggested a ratio of expression levels of 16 ± 3%:84 ± 3% for the α and β genes, respectively, from NF stages 8–24.
The efficacy of MOyy1 treatment was confirmed by observing the depletion of YY1 protein, the lack of phenotypic change with injections of similar (or greater) levels of the control MO (MOcon; see Fig. 4A), and the rescue of viability by co-injection of YY1 mRNA (YY1rescue; engineered to contain five silent point mutations in the target zone of MOyy1; see Fig. 4A). The injection of a range of MOyy1 amounts from 3.5 to 16.4 ng/embryo did not affect PCNA levels but greatly depleted YY1 protein levels (normalized relative to PCNA) at NF stage 11–12 (Fig. 4D). The injection of similar amounts of MOcon slightly depleted YY1 protein levels compared with the controls. Further repetitions with the injection of 9.8 ng of MOyy1/embryo consistently reduced protein levels to below 20% of the controls. The injection of 9.8 ng of MOyy1/embryo was lethal by NF stage 13 (Table I). The injection of 8.2 ng also reduced viability. Co-injection of YY1rescue mRNA with the lethal 9.8-ng dose of MOyy1 restored viability in a proportion of the embryos (Table I). TUNEL analysis of embryos lethally affected by MOyy1 at NF stage 11–12 extensively labeled the ectoderm cells but not the endoderm cells (9/9; visible at the blastopore), suggesting that the ectodermal cells were preferentially affected by apoptosis (Fig. 4G). Few cells were labeled in the embryos treated with the same concentration of MOcon (5/5; Fig. 4E) or in the noninjected embryos (5/5; Fig. 4F).
Table IThe effects upon Xenopus embryo viability by MO and YY1rescue RNA injections
Morphological changes in the embryos were observed with MOyy1 injections, especially a decrease in the length of the anterior posterior axis (Fig. 4, H–L). Those embryos injected with 8.2 ng of MOyy1 displayed abnormalities in common with a Dorsoanterior index score of 4 that included having a diminished forehead and smaller eyes (Fig. 4, K and L) (
). A third of these embryos displayed the more severe phenotype shown in Fig. 4L. Those embryos injected with 9.8 ng of MOyy1 failed to pass through gastrulation (Fig. 4M). However, the majority of those embryos co-injected with 9.8 ng of MOyy1 and 200 pg of YY1rescue RNA passed through gastrulation and displayed a morphological phenotype range similar to those embryos injected with 8.2 ng of MOyy1 (Fig. 4, N and O). The surviving embryos co-injected with 9.8 ng of MOyy1 and 400 pg of YY1rescue RNA displayed a phenotype similar to the more severely affected of the 9.8-ng MOyy1 plus 200-pg YY1rescue RNA batch (Fig. 4P). Those embryos injected with 16.4 ng of MOcon displayed no obvious morphological defects (not shown).
The gene expression of one panneural and four NC markers was examined in embryos injected with sublethal doses of MOyy1 with or without the co-injection of Wnt8-expressing plasmid (Fig. 5). The injection of Wnt8-expressing plasmid has been shown to augment and extend the expression domain of slug more anteriorly (
). RT-PCR was used to analyze NF stage 13 embryos after being injected at NF stage 1 with 8.2 ng of MOyy1 and/or 20 pg of Wnt8-expressing plasmid (Fig. 5A). PCR products of the expected size were detected for each of the primer sets used detecting the neural marker Nrp1; the NC markers slug, snail, AP2, and twist; and the loading control ODC. The levels of Nrp1, slug, and snail mRNA were diminished severely with the injection of MOyy1, whereas the levels of twist and AP2 remained unchanged. The levels of Nrp1, slug, snail, and twist mRNA increased with the injection of Wnt8-expressing plasmid, whereas the levels of slug, snail, and twist mRNA increased with the co-injection of Wnt8 plus MOyy1. The level of AP2 mRNA remained unchanged with the injection of Wnt8-expressing plasmid. Whole mount in situ hybridization was used to analyze the spatial expression of slug and Nrp1 (Fig. 5, B–I). Embryos were subject to injections of 20 pg of Wnt8 plasmid at NF stage 1 and injections of 4.4 ng of MOyy1 in the left blastomere at NF stage 2 and analyzed at stage NF stage 13. As expected, slug mRNA was detected in two similarly intense discrete stripes lateral to the dorsal midline in the more anterior portion of the embryo (10 of 12 embryos examined; Fig. 5B). slug mRNA was not detected on the left side of embryos subjected to injection of MOyy1 but was detected on the right side, in common with the noninjected controls (8 of 9; Fig. 5C). slug mRNA was detected more anterior (15 of 19) sometimes appearing as a single semicircular loop linking across both hemispheres of the embryos (6 of 19) when subject to Wnt8 plasmid injections (Fig. 5D). slug mRNA was not detected on the left side of six embryos and was only detected faintly in the remainder (4 of 10) of the embryos subjected to co-injection of Wnt8 plasmid and MOyy1 in the region that corresponded to the more anterior portion of the slug-expressing domain on the right side (Fig. 5E). Again, slug expression on the right side was similar to that found in the embryos only injected with Wnt8 plasmid. As expected, Nrp1 mRNA expression was detected in a diffuse domain on the dorsal side, extending from the midline to what would seem to be the slug mRNA-expressing zones (8 of 8; Fig. 5F). The Nrp1 mRNA expression was diminished in the MOyy1-injected left side of the embryos, compared with the right side of the same embryos and with the controls (16 of 25; Fig. 5G). Nrp1 expression appeared unaffected by Wnt8 plasmid injections (17 of 17 and 7 of 12; Fig. 5, H and I, respectively).
The direct regulation of the spatial expression of the slug gene by endogenous YY1 was suggested by using a GFP reporter assay in combination with the ChIP assay. The injection of slug 200WT reporter plasmid into embryos between 30 and 90 min after fertilization at NF stage 1 spatially restricted the expression of GFP at NF stage 16 (Fig. 6A). In the majority of embryos that displayed fluorescence, the expression of GFP was found in the neural ectoderm and neural/nonneural ectoderm border (Table II, top). However, when the slug 200Y* plasmid was injected, the expression of GFP in the majority of fluorescent embryos was detected more laterally and ventrally in the nonneural ectoderm (Fig. 6B). The proportion of embryos displaying fluorescence did not differ markedly between those injected with 200WT and 200Y* (Table II). 50 embryos from each treatment were subject to anti-YY1 ChIP. The level of plasmid detected by qPCR in the anti-YY1 ChIP (and normalized for the respective inputs) was 12-fold higher in the samples injected with 200WT compared with the samples injected with 200Y*. Co-injection of 400 pg of YY1 RNA with 200WT reporter did not reduce the total number of fluorescent embryos but reduced the proportion of the embryos with spatially restricted expression of GFP to a similar level found with those embryos injected with 200Y* alone. Co-injection of 400 pg of YY1 RNA with 200Y* reporter slightly increased the total number of fluorescent embryos and only slightly reduced the minority proportion of embryos with spatially restricted expression of GFP (Table II, bottom).
Table IIThe scoring of the GFP spatial expression patterns identified in the embryos injected with either the 200WT or the 200Y* reporter (top) and in combination with YY1 RNA (bottom)
Injection conditions, reporter + RNA
Restricted to neural ectoderm and border
Nonrestricted in ectoderm
Proportion of fluorescent embryos (total embryos examined)
The interaction of YY1 with the X. laevis slug gene was demonstrated by sequence analysis, EMSAs, ChIP, and reporter assays. YY1 binding to the Xenopus slug gene promoter was sensitive to sequences peripheral to the core consensus as observed with the G* mutation. Little evidence was found for other proteins specifically binding to the GAGA motif as inferred from studies of PHO binding in Drosophila (
). The ChIP assays also demonstrated that endogenous YY1 directly interacts with the YY1 cis-element of the Xenopus endogenous slug gene during the period of NC specification. The capacity to bind YY1 seems to be maintained by the mammalian slug genes, but YY1 may have a different role in mammals than in Xenopus. slug expression is detected later during the migration phase of NC and other mesenchymal cells (
The results here indicate that YY1 is primarily located in the nucleus in the developmental stages examined and that encompassed NC specification. If it is assumed that PCNA levels reflect nuclei number from NF stage 8 to 22 (
), then the nuclear concentration of YY1 increased by stage 15. YY1 was detected throughout the nucleus and not obviously restricted to a particular part. In common with this study, a punctate localization of YY1 in the nuclei was also detected in murine blastocysts (
). YY1 may not have been detected immunohistochemically in that study, because the antibodies and protocols used were different than here.
Both YY1 gene products were targeted efficiently by MOyy1. YY1β mRNA was over 5-fold more prevalent than the known YY1α, and its 5′-UTR differed sufficiently so as not to be good target for a previously described anti-YY1 MO (MOKwon et al. in Fig. 4A) (
) post-midblastula transition. The levels of PCNA did not appear to alter with MOyy1 treatment, suggesting that cell division was not inhibited at these developmental stages. The level of YY1 protein appeared lower in those cells that undergo the higher rates of proliferation (
) after midblastula transition into gastrulation. Presumably, therefore, inhibiting de novo YY1 synthesis post-midblastula transition adversely affected the ectoderm (compared with the endoderm) and the mesoderm in the blastopore lip because of the greater decline of maternal stocks, inducing an apoptotic-like response in the ectoderm and gastrulation arrest. Sublethal depletion of YY1 affected the morphology of the embryos concomitant with an effect on neural tissue, such as shortened axis, loss of or deformed forehead, and eye defects. These results compare favorably with YY1 loss of function in mice and PHO loss of function in Drosophila. Murine embryos with the YY1–/– genotype develop to the blastocyst stage but die at a stage of rapid proliferation and differentiation. The YY1+/– genotype causes neural defects (
YY1 depletion dramatically inhibited expression of the slug gene according to the PCR and in situ hybridization analyses. This was probably an indirect effect of YY1 on the slug gene, because the neural ectoderm was adversely affected as reflected by the loss of Nrp1 expression and by the decline of the other early NC marker Snail. However, Twist and AP2 expression were refractive to the MOyy1 treatments. This is probably because these markers are not restricted to the prospective NC at the stage analyzed and thus reflect the lack of change in the prospective epidermis and mesoderm, respectively (
). The results suggest that YY1 performs a critical role in the differentiation of the neural ectoderm, and therefore, without proper neural ectoderm differentiation, the specification of the NC territory was disrupted. This explains the low level of melanocytes in the most severely affected embryos that survived to tadpole stages.
The direct effect of endogenous YY1 on the slug gene was indicated from experiments where reporter plasmids were injected soon after fertilization (to promote an association with chromatin comparable with of the endogenous gene) (
), it was spatially restricted to the neural ectoderm and its lateral border. This restricted expression was regulated by the YY1 cis-element, which in turn was associated with YY1 protein. The direct effect of YY1 on the slug gene was also suggested by the overexpression of YY1, which reduced to a minority (with a dependence upon the YY1 cis-element), the proportion of embryos with a spatially restricted GFP expression. However, an indirect effect was also suggested by the elevation of those embryos displaying a nonrestricted GFP expression. This may be caused by YY1 promoting neural differentiation more extensively across the ectoderm (
) (concurring with YY1 depletion, which inhibits neural differentiation) and thus enhancing reporter activity.
The results support a hypothesis where YY1 directly restricts the activation of the slug gene through regulating its silenced state in the lateral and ventral ectoderm. This silencing could be polycomb-like, in that it is probably not operating through inhibiting transcriptional activity but by preventing gene activation. Although YY1 expression levels may have differed across the ectoderm, it seems likely that YY1 would require co-factor(s) to qualitatively mediate this silencing. Furthermore, depleting YY1 should make the slug gene more responsive to NC-inducing factors. An example of such a factor is Wnt8. And although there was a slight induction in a region corresponding to an area more anterior of the slug expressing domain, there was not a more general ectopic expression in the lateral and ventral ectoderm depleted of YY1. However, these ectodermal cells may have been relatively unresponsive to Wnt8 if they did not express the appropriate Frizzled receptor (
In conclusion, YY1 was found to be a functional nuclear protein and essential in Xenopus development, which may directly target the slug gene and regulate its silencing in the nonneural ectoderm. If so, this may reflect a mechanism by which the ectoderm loses its competence to express the slug gene (
). YY1 was also found to regulate the slug gene indirectly through regulating the differentiation of the neural ectoderm. How YY1 performs these functions will probably be answered through identifying the protein or RNA complexes with which it is associated spatially and temporally in the Xenopus embryo.
We thank J. Vallin for help in initiating this project and for access to reagents; P-Y. Bourillot, F. Broders, O. Eder, N. Elkhatib, M. Faraldo, L. Gibbs, R. Mudge, A. Nennot, T. Pietri, J. Teulière, and J. Veltmaat for technical advice and assistance; and G. Mainguy, Y. Bellaïche, and A-H. Monsoro-Burq for useful suggestions with drafting the manuscript.