XB-ART-49657BMC Biol October 3, 2014; 12 81.
Global identification of Smad2 and Eomesodermin targets in zebrafish identifies a conserved transcriptional network in mesendoderm and a novel role for Eomesodermin in repression of ectodermal gene expression.
BACKGROUND: Nodal signalling is an absolute requirement for normal mesoderm and endoderm formation in vertebrate embryos, yet the transcriptional networks acting directly downstream of Nodal and the extent to which they are conserved is largely unexplored, particularly in vivo. Eomesodermin also plays a role in patterning mesoderm and endoderm in vertebrates, but its mechanisms of action, and how it interacts with the Nodal signalling pathway are still unclear. RESULTS: Using a combination of ChIP-seq and expression analysis we identify direct targets of Smad2, the effector of Nodal signalling in blastula stage zebrafish embryos, including many novel target genes. Through comparison of these data with published ChIP-seq data in human, mouse and Xenopus we show that the transcriptional network driven by Smad2 in mesoderm and endoderm is conserved in these vertebrate species. We also show that Smad2 and zebrafish Eomesodermin a (Eomesa) bind common genomic regions proximal to genes involved in mesoderm and endoderm formation, suggesting Eomesa forms a general component of the Smad2 signalling complex in zebrafish. Combinatorial perturbation of Eomesa and Smad2-interacting factor Foxh1 results in loss of both mesoderm and endoderm markers, confirming the role of Eomesa in endoderm formation and its functional interaction with Foxh1 for correct Nodal signalling. Finally, we uncover a novel, role for Eomesa in repressing ectodermal genes in the early blastula. CONCLUSION: Our data demonstrate that evolutionarily conserved developmental functions of Nodal signalling occur through maintenance of the transcriptional network directed by Smad2. This network is modulated by Eomesa in zebrafish which acts to promote mesoderm and endoderm formation in combination with Nodal signalling, whilst Eomesa also opposes ectoderm gene expression. Eomesa therefore regulates the formation of all three germ layers in the early zebrafish embryo.
PubMed ID: 25277163
PMC ID: PMC4206766
Article link: BMC Biol
Species referenced: Xenopus laevis
Genes referenced: eomes foxh1.2 gsc klf3 ndrg1 nodal nodal1 not smad2 smarcd3 tfap2a vgll4l zic3
Article Images: [+] show captions
|Figure 1. ndr1 overexpression in zebrafish blastulas identifies known and novel Nodal target genes. (A) Heatmap of a representative selection of genes induced on ndr1 overexpression across the full range of P values and fold changes. Genes previously identified as Nodal target genes in zebrafish are in bold. (B) In situ hybridisation of wild type and ndr1 mRNA-injected embryos for foxa, klf3, nhsl1b, notum1a and smarcd3 at 50% epiboly showing upregulation in response to ndr1 and absence in MZoep mutant embryos that have no Nodal signaling. Animal views; dorsal to the right. Numbers on each panel indicate the number of embryos showing the phenotype depicted over the total number of embryos analysed. For foxa expression in MZoep mutants the remaining 13/36 embryos showed absent expression except for a patch on one side of the embryo. For nhsl1b and notum1a the remaining 8/26 and 5/37 embryos, respectively, showed reduced expression around the margin. (C) Comparison of all genes represented on the microarray, or ndr1-responsive genes (either up- or down-regulated) to proximal Smad2 binding; comparison was performed for both all and novel ndr1-responsive genes. Compared to all genes on the microarray or those that are down-regulated in response to ndr1, up-regulated ndr1-responsive genes (both all and novel) are significantly associated with Smad2 binding. *P =7 × 10−7; **P =1 × 10−40. (D) Examples of Smad2 binding upstream of known (tbx16 and flh) and novel targets; scale in reads per million reads.|
|Figure 2. The transcriptional network directed by Nodal signalling is substantially conserved in vetebrates. (A) Venn diagram indicating the number and overlap of genes with Smad2 binding ±10 kb of TSSs in ChIP-seq datasets for each of four vertebrates. (B) Examples of genes with proximal Smad2 binding in all four species; scale in reads per million reads. Colour coded as in A. (C) Relationship between ndr1-responsive genes and proximal Smad2 binding in 1 (left), 2 (middle) or 3+ species (right). Colour coded as in A. †P ≤1 × 10−12; †††P ≤1 × 10−42; ††††P ≤1 × 10−48. (D) Fold enrichment for genes expressed at sites of Nodal activity amongst genes with proximal Smad2 binding in zebrafish, or zebrafish and other species. (E) Fold enrichment for genes involved in known Nodal-mediated processes amongst genes with proximal Smad2 binding in zebrafish, or zebrafish and other species. Key defined as gene subsets as in A. *P ≤5 × 10−2; **P ≤3 × 10−4; ***P ≤5 × 10−6; ****P ≤1 × 10−20. ChIP-seq, chromatin immunoprecipitation sequencing; TSSs, transcription start sites.|
|Figure 3. Smad2 and Eomesa bind common regulatory elements proximal to ndr1 -responsive genes and regulate the developmental functions of Nodal signalling. (A) De novo motif analysis identifies the known Smad binding element (SBE) and consensus T-box binding site within Smad2 ChIP-seq peaks and Eomesa ChIP-seq peaks. (B) Venn diagram indicating the overlap between Smad2 and Eomesa ChIP-seq peaks. (C) Venn diagram of the overlap between genes with proximal Smad2 and Eomesa binding (within ±10 kb of their TSS). (D) Examples of genes with common proximal peaks of Smad2 and Eomesa binding; scale in reads per million reads. (E-G) Functional and anatomical analysis of genes showing Smad2 and/or Eomesa binding. Colour coded as in C. (H) Comparison of ndr1-responsive genes with subsets of genes with proximal Smad2 and/or Eomesa binding as defined in C. ** P ≤2 × 10−3; **** P ≤1 × 10−105. ChIP-seq, chromatin immunoprecipitation sequencing; TSS, transcription start site.|
|Figure 4. Eomesa and Foxh1 combinatorially regulate early mesendoderm marker expression and are required for endoderm formation. In situ hybridisation of ntla (A-E), gsc (F-L) and sox32 (M-S) at 30% epiboly in wild-type embryos, foxh1 morphants, MZeomesa mutants, MZeomesa;foxh1 morphants, MZoep mutants, and in wild type embryos and MZeomesa mutants injected with foxh1 mRNA. Rescued expression of sox32 expression in the ventral margin of MZeomesa mutants injected with foxh1 mRNA is indicated by an asterisk. In situ hybridisation of ttna (T-W) at 24 hpf in wild-type embryos, foxh1 morphants, MZeomesa mutants and MZeomesa;foxh1 morphants showing expression in the heart (arrow heads) and somites. The open arrow heads indicate cardia bifida. Animal views; dorsal to the right. Numbers on each panel indicate the number of embryos showing the phenotype depicted over the total number of embryos analysed. For sox32 expression in MZoep mutants faint staining in the YSL was detected in 11/15 embryos as previously reported . hpf, hours post fertilization.|
|Figure 5. Eomesa positively regulates mesendoderm and negatively regulates ectoderm markers and chromatin assembly genes. (A) Heatmap of expression differences between wild-type and MZeomesa mutants at sphere stage; π – endoderm; ∑ - mesoderm; ∂ - ectoderm. (B) Comparison of all genes, or those up- or down-regulated in MZeomesa embryos compared to wild type with genes that have Smad2 only, common, uncommon or Eomesa only binding within ±10 kb of their TSSs (colour coded as in Figure 3C). Compared to all genes, those that are up- or down-regulated in MZeomesa embryos are significantly associated with lone Eomesa binding. Genes that have common and uncommon binding are associated with down-regulated genes, whilst Smad2 only and common binding is associated with up-regulated genes in MZeomesa embryos. † P ≤2 × 10−5; †† P ≤1 × 10−7; ††† P ≤1 × 10−18. (C) Functional annotation analysis of genes upregulated in MZeomesa mutants. (D) Anatomical analysis of genes downregulated in MZeomesa mutants. (E) Anatomical analysis of genes upregulated in MZeomesa mutants. TSSs, transcription start sites.|
|Figure 6. Eomesa negatively regulates neural marker gene expression. (A) ChIP-seq data showing binding of Eomesa at sphere stage proximal to neural marker genes; scale in reads per million reads. (B) In situ hybridisation of wild type and MZeomesa mutant embryos for stm, tfap2a, vgll4l and zic3 showing upregulation of these genes (lateral view, animal to the top). (C) In situ hybridisation of control injected and eomesa injected embryos for stm, tfap2a, vgll4l and zic3 showing downregulation of these genes (lateral view, animal to the top). ChIP, chromatin immunoprecipitation sequencing.|
References [+] :