Skip to main content
. 2019 Nov 20;8:e49921. doi: 10.7554/eLife.49921

Figure 4. Cardiopharyngeal lineage-specific accessibility profiles and decoupling between enhancer accessibility and activity for de novo expressed genes.

(A) Differentially expressed (DE) genes vs. differentially accessible (DA) peaks in response to FGF-MAPK perturbation in the fate-restricted cells. ρ is the Spearman correlation of expression and accessibility for DA peaks associated with DE genes. (B) Relationship between accessibility and expression of de novo pan-cardiac genes as in Figure 2C. DE genes in either condition are shown on the left. (C) Time-dependent ATAC-seq peaks associated with de novo expressed pan-cardiac genes. The accessibility of these peaks is shown for 6, 10 and 18 hpf vs. the average accessibility in the controls (LacZ) and upon FGF-MAPK perturbations at either 10 or 18 hpf. Peaks were classified as ‘Open in ASM’ (less accessible in FgfrDN vs. M-RasCA or LacZ at 18 hpf), ‘Open in Heart’ (less accessible in M-RasCAvs. FgfrDNor LacZ at 18 hpf), ‘Closed in FoxfCRISPR’ (less accessible in FoxfCRISPR vs. ControlCRISPR), or ‘Open in TVC’ (less accessible in FgfrDN vs. MekS216D,S220E or LacZ at 10 hpf). Only regions changing accessibility between 6 and 10 hpf, or 10 and 18 hpf are shown. (D) A 6 kb region on chromosome four displaying expression profiles of RNA-seq and chromatin accessibility profiles of ATAC-seq normalized tag count. Peak ID refers to elements tested for reporter assay in vivo. The newly identified enhancer in Lrp4/8 locus is in the boxed region. (E) Enhancer-driven in vivo reporter expression (green) of tested ‘KhC4.137’ peak. TVCs marked with Mesp>H2B::mCherry (red). Numbers indicate observed/total of half-embryo scored. Zoom on cardiopharyngeal cell lineage (panel on the right). (F) Endogenous expression of Lrp4/8 visualized by in situ (green) in TyrosinaseCRISPR and upon CRISPR/Cas9-induced deletion of ATAC-seq peaks. Nuclei of B7.5 lineage cells are labelled by Mesp>nls::LacZ and revealed with an anti beta-galactosidase antibody (red). Mesp-driven hCD4::mCherry accumulates at the cell membrane as revealed by anti mCherry antibody (Blue). Experiment performed in biological replicates. Scale bar = 10 μm. (G) Fisher exact test; n is the total number of individual embryo halves scored per condition. (H) Summary model: patterns of chromatin accessibility dynamics and gene expression during early cardiopharyngeal fate specification.

Figure 4.

Figure 4—figure supplement 1. General characterization of FGF-MAPK perturbation.

Figure 4—figure supplement 1.

(A–D) Intersections of differentially expressed genes from bulk RNA-seq and scRNA-seq (Wang et al., 2019) comparisons in cardiopharyngeal restricted cells using Venn Diagram (A–B) and UpSet plots (C–D). FGF-MAPK activated genes at 18 hpf are defined as the intersection of genes downregulated in FgfrDN vs. LacZ at 18 hpf and genes downregulated in FgfrDNvs. M-RasCA at 18 hpf (A). FGF-MAPK inhibited genes at 18 hpf are defined as the intersection of genes downregulated in M-RasCA vs. LacZ at 18 hpf and genes downregulated in M-RasCA vs. FgfrDN at 18 hpf (FDR < 0.05 and |log2(FC)| > 1). (E–F) GSEA of significantly enriched gene sets shows opposite trends in FgfrDN vs. control and M-RasCAvs. control (LacZ). Dark gray bars indicate significant enrichment. Light gray bars are not significant (FDR < 0.05).
Figure 4—figure supplement 2. Differential accessibility in response to FGF/MAPK perturbation at 18 hpf.

Figure 4—figure supplement 2.

(A) Differentially expressed (DE) genes vs. differentially accessible (DA) peaks using bulk RNA-seq and ATAC-seq in Foxf>M RasCA vs. LacZ at 18 hpf. (B) Relationship between expression and accessibility of DE genes associated with DA peaks. The top axis shows the scale for log2(FC) of bulk RNA-seq (black dots). The bottom axis shows the log2(FC) scale for ATAC-seq (colored diamonds). Top 50 genes inhibited and activated by FGF-MAPK (based on log2(FC)) along with STVC genes. Heatmaps show gene expression over time. (C) A 15 kb region on chromosome five displaying expression profiles from RNA-seq and chromatin accessibility profiles from ATAC-seq (normalized by total sequencing depth). Transcript model is indicated as black bars, percentage of conservation score between C. savignyi and C. robusta obtained obtained from WASHU browser (Brozovic et al., 2018) (yellow peaks), accessome (light grey bars), TVC- (green bar) and ATM-specific peaks (dark gray bars). Peak accessible specifically in the heart (KhC5.1641) of Mmp21 locus is in the boxed region. (D–E) Enhancer-driven in vivo reporter expression (green) of a ~ 3 kb region upstream the coding ATG and including the DA peak (KhC5.1641). (D) dorsal view, (E) lateral view; GFP is detected in the mesenchyme (asterisk) surrounding the cardiopharyngeal progenitors (D) and in the epidermis (orange arrow) (E). B7.5 cells are marked with Mesp>H2B::mCherry (red). Dotted line: ventral midline. Numbers indicate observed/total of half-embryos scored. Scale bar = 25 µm.
Figure 4—figure supplement 3. Accessibility of elements annotated to de novo ASM genes.

Figure 4—figure supplement 3.

(A) Relationship between accessibility and expression of de novo ASM genes in either condition indicated. (B) Time-dependent ATAC-seq peaks associated with de novo expressed ASM genes. The accessibility of these peaks is shown for 6, 10 and 18 hpf vs. the average (avg.) accessibility in the controls and upon FGF-MAPK perturbation either at 10 and 18 hpf. Peaks are clustered based on their accessibility as ‘Open in ASM’ less accessible in Hand-r>FgfrDN vs. Foxf>M RasCA or Hand-r>LacZ at 18 hpf, ‘Open in Heart’ less accessible in Foxf>M RasCA vs. Hand-r>FgfrDN or Hand-r>LacZ at 18 hpf, ‘Closed in FofxCRISPR’ less accessible in FoxfCRISPR vs. ControlCRISPR, ‘Open in TVC’ less accessible in Mesp>FgfrDN vs. Mesp>MekS216D,S220E or Mesp>LacZ at 10 hpf. Only peaks changing accessibility between 6 hpf and 10 hpf or 10 hpf and 18 hpf are shown.
Figure 4—figure supplement 4. Accessibility of binding motifs over time for elements annotated to de novo-expressed genes.

Figure 4—figure supplement 4.

(A) TF binding motifs enriched in peaks associated with de novo cardiac and pharyngeal expressed genes parsed based on primed and de novo accessibility (see Material and methods). log2(OR) (see Materials and methods) are shown for motifs significantly enriched (one-tailed hypergeometric test, p < 0.05) in the indicated peak classes. Motif accessibility from chromVAR (Schep et al., 2017) is shown for all peaks associated with a de novo-expressed cardiac or ASM gene. Only the motif with the highest log2(OR) for each TF is shown. (B) Conservation of TVC-accessible peaks closed in FoxfCRISPR in Lrp4/8 locus between C. savignyi and C. robusta. The gene body is shown in black. Conservation score was obtained from WASHU browser (Brozovic et al., 2018). (C) Alignment of conserved region of ‘KhC8.137’ peak between C. robusta and C. savignyi. (D) C. robusta Hand locus, exon IV sequence highlighted in bold and black. Only the core nucleotides are shown for candidate Pitx binding sites.