(A) Confocal image of facial cartilages expressing col2a1a:GFP and sox10:Dsred at 72 hpf. Lateral view with anterior to left. Scale bar = 100 μm. (B) Venn diagram indicating distal elements with open chromatin accessibility in col2a1a:GFP+; sox10:Dsred+ versus col2a1a:GFP−; sox10:Dsred− cells. (C) Peak intensity plots (heatmap) of μATACseq show differentially enriched open chromatin regions in double-positive versus double-negative cells. (D) The top five transcription factor (TF) motifs recovered from the top 2000 μATACseq peaks enriched in chondrocytes (after removing redundant motifs). (E) GO analysis of nearest neighbor genes of μATACseq peaks enriched in chondrocytes.

(A) col2a1a locus showing μATACseq reads from the indicated experiments. Several regions become accessible in 48 hpf CNCCs and then remain open in chondrocytes at 72 hpf. Boxed region shows the published R2 cartilage enhancer. (B) Comparison of consensus motif sequences recovered from zebrafish μATACseq peaks enriched in facial chondrocytes (left) and mouse Sox9 ChIP-Seq peaks in rib chondrocytes (right).

(A) Confocal images of CNCCs expressing fli1a:GFP and sox10:Dsred at 36 and 48 hpf. Lateral view with anterior to left. Scale bar = 100 μm. (B) Peak intensity plots of cartilage-accessible distal elements shown for chondrocytes at 72 hpf and CNCCs at 36 and 48 hpf. Chondrocyte accessible elements are pooled into three categories based on dynamics of chromatin accessibility across stages. (C, E, G) De novo motif enrichment recovered by Homer analysis among the three categories. Top six motifs are shown with associated p values after removing redundant motifs. (D, F, H) GO term analysis among the three categories.

(A) Peak intensity plots of μATACseq show differentially enriched open chromatin regions in 36 hpf and 48 hpf fli1a:GFP+; sox10:Dsred+ CNCCs. (B, C) De novo motif enrichment recovered by Homer analysis for peaks with decreasing (B) and increasing (C) accessibility from 36 to 48 hpf. Top six motifs are shown with associated p values after removing redundant motifs. (D, E) GO term analysis and associated p values for peaks with decreasing (D) and increasing (E) accessibility from 36 to 48 hpf.

(A,B) Confocal images show dorsal CNCCs of the first two arches labeled by sox10:Dsred and pou3f3b:Gal4; UAS:nlsGFP in control and foxc1a^–/–; foxc1b^–/– mutant embryos at 48 hpf. Scale bar = 100 μm. (C,D) Confocal images show loss of dorsal cartilages in foxc1a^–/–; foxc1b^–/– mutant embryos at 6 dpf. sox10:Dsred+ cartilages are seen in single channels in C’ and D’, with dashed lines highlighting boundaries of dorsal arch and otic cartilage. (E) Peak intensity plots of Group I and Group II elements in control and foxc1a^–/–; foxc1b^–/– mutant embryos. Peaks above the dashed lines are reduced in mutants. (F,G) De novo motif enrichment of Foxc1-dependent and Foxc1-independent Group I and Group II elements. Top six motifs are shown with associated p-values after removing redundant motifs.

(A,B) In situ hybridization of foxc1a or foxc1b with pou3f3b:Gal4; UAS:nlsGFP (pou3f3b>GFP, detected by anti-GFP antibody) at 36 hpf. Dashed lines indicate regions of foxc1a or foxc1b expression, carried over to the pou3f3>GFP-only and merged images for reference. Scale bar = 100 μm. (C) Quantification of pou3f3b>GFP+; sox10:DsRed+ cartilage size in control and foxc1a^–/–; foxc1b^–/– mutant embryos at 6 dpf. Units were normalized to one based on the average size of the control cartilage. n indicates number of embryos analyzed. Shown is mean ± sem, and p value was calculated using Students’ T test (tails = 2, type = 3).

Venn diagrams and heatmaps show distal accessible elements in μATACseq data from fli1a:GFP+; sox10:Dsred+ (pan-CNCCs) and pou3f3b>GFP+; sox10:DsRed+ (dorsal CNCCs) at 36 hpf (A) and 48 hpf (B). Numbers list elements unique for each population or shared (center of circles).

(A,B) Heatmaps of the 5,736 distal elements with accessibility in 72 hpf chondrocytes. Their accessibility is then plotted according to μATACseq data from fli1a:GFP+; sox10:Dsred+ (pan-CNCCs) and pou3f3b>GFP+; sox10:Dsred+ (dorsal CNCCs) at 36 hpf (A) and 48 hpf (B). At 36 hpf, 132 peaks were more accessible in pan-CNCCs and 106 peaks more accessible in dorsal CNCCs. At 48 hpf, 2,470 peaks were more accessible in pan-CNCCs and 10 peaks more accessible in dorsal CNCCs. Peaks commonly or not accessible in both data sets are also shown. (C,D) Pearson correlation graphs of the behavior of cartilage-associated distal elements in pan- versus dorsal CNCCs at 36 and 48 hpf. x and y axes represent raw read counts from BAM files, and ‘r’ is the Pearson correlation coefficient.

Comparison of the chromatin accessibility of Group I and Group II peaks between wild types and FoxC mutant CNCCs. The x-axis represents the log2 fold change value of mutant versus control peaks, and the y-axis represents the −log10 adjusted p values associated with these regions. Peaks above a −log10 adjusted p value of 1 (purple dots) are considered significant. Many more peaks display decreased accessibility at 48 hpf than at 36 hpf in Foxc1 mutants.

(A) Genomic regions (gene loci and GRCz10 coordinates listed) for enhancer testing on the left and GFP expression driven by the indicated peaks in stable transgenic zebrafish at 6 dpf on the right. Peaks (p) tested are shown, with Foxc1-dependent regions in purple and Foxc1-independent elements in green. μATACseq reads are shown in each row, with chondrocyte and non-chondrocyte peaks from 72 hpf embryos. Confocal projections show cartilages of the first two arches in lateral view with anterior to the left. Arrows indicate enriched expression at joint regions, and arrowheads denote relative lack of expression. (B) Confocal projections show selective loss of sox10_p2:EGFP, ucmab_p1:EGFP, and epyc_p1:EGFP transgene expression in the dorsal cartilage domains (dashed outlines) of foxc1a−/−; foxc1b−/− mutants at 72 hpf. (C) Confocal projections of in situ hybridization show selective loss of sox10, lect1, col9a3, and epyc in the dorsal cartilage domain (dashed outline) of foxc1a−/−; foxc1b−/− mutants at 56 hpf. Numbers indicate proportion of embryos in which the displayed patterns were observed. bh, basihyal; DNCCs, dorsal CNCCs; jj, jaw joint; hj, hyoid joint; hm-op, hyomandibular-opercular joint; hm-otic, hyomandibular-otic junction; hm-sy, hyomandibular-symplectic junction; M-M, Meckel’s–Meckel’s joint. Scale bars = 100 μm.

(A) Schematic of the reporter construct for enhancer testing. Tol2, transposase integration site. E1b, minimal promoter. GFP, green fluorescent protein. pA, polyadenylation sequence. CFP, cerulean fluorescent protein. (B) Snapshots of genomic regions (gene loci and GRCz10 coordinates listed) for enhancer testing. Peaks (p) tested are shown, with Foxc1-dependent regions in purple and Foxc1-independent elements in green. μATACseq reads are shown in each row for the experiments indicated, with chondrocyte and non-chondrocyte peaks from 72 hpf embryos. DNCCs, dorsal CNCCs. (C) GFP expression (yellow) driven by the indicated peaks in stable transgenic zebrafish at 6 dpf. Confocal projections of the cartilages of the first two arches are shown in lateral view for acan_p1, slc35d1a_p1 and sparc_p1 and ventral view for col9a1a_p1. For acan_p1, sox10:Dsred labels all chondrocytes for reference in magenta. Arrows indicate expression in cartilages. sparc_p1 also drives expression in ligamentocytes and osteoblasts, and slc35d1a_p1 in pharyngeal muscles. (D) Confocal sections of representative embryos injected with enhancer constructs are shown in lateral view at 6 dpf. Mosaic expression in chondrocytes (yellow, arrows) is shown, with DIC (white channel) providing embryo context. Numbers indicate proportion of embryos positive for the co-selectable lens:CFP marker (also on the injected plasmid) in which the displayed patterns were observed. Scale bar = 100 μm.

(A) Genomic regions (gene loci and GRCz10 coordinates listed) for enhancer testing. Peaks (p) tested are shown in green, and μATACseq reads are shown in each row. DNCCs, dorsal CNCCs. (B) GFP expression (yellow) driven by the indicated peaks at 36 hpf and 6 dpf. For emx3_p1 and prrx1a_p1, confocal projections of whole arch CNCCs of stable transgenic zebrafish are shown in lateral view at 36 hpf. For satb2-p1, confocal section of representative injected embryo is shown in lateral view at 6 dpf, with DIC providing embryo context in white. Numbers indicate proportion of embryos positive for the co-selectable lens:CFP marker (also on the injected plasmid) in which the displayed patterns were observed. Scale bar = 100 μm.