Unique and shared dact1 and dact2 gene expression domains during zebrafish development. (A-C) Whole-mount in situ hybridization showing dact1, dact2, and wnt11f2 gene expression patterns. Scale bar = 100 μm. A) At 8hpf, dact2 and wntllf2 are highly expressed in the dorsal margin and presumptive Nieuwkoop center of the gastrulating embryo, with dact1 being weakly detected (arrowhead). In contrast to wnt11f2, dact1 and dact2 are expressed in the presumptive dorsal mesoderm (asterisk). B) Lateral (anterior to the left of page) and anterior (dorsal side toward top of page) views of bud-stage embryos. dact2 and wnt11f2 transcripts are detected in the anterior neural plate (arrowhead) and tailbud (asterisk) while dact2 is additionally expressed in the axial mesoderm (arrow). dact1 gene expression is concentrated to the paraxial mesoderm (open arrowhead). C) Lateral and flat-mount views of 4 ss embryos. dact2 is expressed in the anterior neural plate and polster (P), notochord (N), paraxial and presomitic mesoderm (PM) and tailbud (TB). wnt11f2 is also expressed in these cells (Thisse 2001). In contrast, dact1 is expressed in the midbrain (MP) and the paraxial and presomitic mesoderm. D) RNAscope in situ hybridization analysis of dact1 (white) and dact2 (yellow) and irf6 (green) expression in transverse section of 72 hpf embryos. dact1 is expressed in the orofacial cartilage (pq), while dact2 is expressed in the oral epithelium (oe). The epithelial marker irf6 is expressed in the oe and surface epithelium (se). Dapi (blue). Scale bar = 100 μm. E) Daniocell single-cell RNAseq analysis with a display of dact1, dact2, wnt11f2 and gpc4 clusters from 3-120 hpf of development (Farrell, Wang et al. 2018).

Impaired convergent extension in dact1 and dact2 compound mutants. A) Inter-cross of compound heterozygotes yield embryos with different degrees of axis extension that correspond to the dact1 and dact2 genotypes. Representative lateral images of embryos at 12 hpf. The yellow line indicates body axis angle measured from the anterior point of the head, the center of the yolk, to the end of the tail. B) Quantification of body axis angle. Numbers represent the difference in angle relative to the average wildtype embryo. Asterisk indicates genotypes with angles significantly different from wildtype. ANOVA p<0.5 n= 3-21 embryos. C) Representative bud stage wildtype and dact/2-/- mutant embryos stained for gsc (prechordal plate), pax2a (midbrain/hindbrain boundary) and krox20 (rhombomere 3). Asterisk indicates lack of krox20 expression in dact1/2 mutant. Scale bar = 200 μm D) Representative flat mounts of 1-2 ss wildtype and dact1/2 mutant embryos stained for zic1 (telencephalon), pax2a and tbx6 (ventrolateral mesoderm). E) Representative flat mounts of 10 ss wildtype and dact1/2-/- mutant embryos stained for ctsl1b (hatching gland), zic1, pax2a, krox20, and myo1d (somites).

Midface development requires dact1 and dact2. A) Representative brightfield images of wildtype and dact1/2-/- compound mutants at 4 dpf. Lateral and ventral views show dact1/2-/- compound mutants have a hypoplastic midface, medially displaced eyes, and a displaced lower jaw. B) Representative flat-mount images of Alcian blue stained ANC and VC elements from 4 dpf wildtype and dact1/2-/- compound mutants. Dact1/2-/- mutants have a rod-shaped ANC with no distinct lateral and medial elements. No obvious differences were found in dact1/2 mutant VC. C) Representative images of Alcian blue stained dact1/2-/-, wntllf2-/-, and wnt11f2-/-,dact1/2-/- compound mutants. Lateral and ventral views show similar craniofacial phenotypes in each mutant. D) Representative flat-mount images of Alcian blue stained ANC and VC elements show a similar phenotype between dact1/2-/-, wntllf2-/-, and wnt11f2-/-,dact1/2-/- compound mutants. Scale bar: 200 μm.

Anterior neural crest cells of the dact1/t2-/- mutant migrate to the midline and populate the ANC. Lineage tracing of wildtype and dact1/2-/- double mutant zebrafish embryos using Tg(sox10:kaede) line. sox10:kaede fluorescence is shown in green and photo-converted kaede is shown in magenta and highlighted with an arrow. Asterisks indicate that the cell population is absent. 19 hpf embryo sagittal views showing photoconversion of anterior-most neural crest population. At 36 hpf frontal images show the migration of photoconverted neural crest cells to the frontonasal prominences in wildtype and dact1/2-/- double mutants. At 55 hpf, frontal images show photoconverted neural crest cells populating the region of the developing ANC in wildtype and dact1/2-/- mutants. At 4.5 dpf ventral images show photoconverted neural crest cells populating the medial ANC in wildtype. Similarly, neural crest cells in dact1/dact2-/- mutants populate the rod-shaped ANC. Scale bar:100 μm.

A nonoverlapping functional role for dact1, dact2, and gpc4 and wls. A) Representative Alcian blue stained whole-mount images of wildtype, dact1/2-/- double mutant, gpc4-/- mutant, and gpc4-/-,dact1/2-/- triple mutants at 4 dpf. Low magnification lateral images of embryos showing tail truncation in dact1/2-/- mutants, shortened and kinked tail in gpc4-/- mutants, and a combinatorial effect in gpc4-/-,dact1/2-/- triple mutants. Higher magnification lateral images show a shortened midface and displaced lower jaw in dact1/2-/- mutants, a shortened midface in gpc4-/- mutant, and a combinatorial effect in gpc4-/-,dact1/2-/- triple mutants. B) Representative flat-mount images of dissected Alcian blue-stained cartilage elements. dact1/2-/- mutants have a narrow rod-shaped ANC while gpc4-/- mutants have a broad and shortened ANC. dact1/2/gpc4 triple mutants have a combinatorial effect with a short, broad rod-shaped ANC. In ventral cartilages, dact1/2-/- mutants have a relatively normal morphology while Meckel’s cartilage in gpc4-/- mutants and gpc4-/-,dact1/2-/- triple mutants is truncated. C,D) Same as above except wls-/- mutant and wls-/-,dact1/2-/- triple mutant, with similar findings. E) Combinatorial genotypes of dact1, dact2, and gpc4. dact2-/- contributed the dact/gpc4 compound phenotype while dact1-/- did not. Scale: 200 μm.

Single-cell RNAseq of 4 ss wildtype, dact1/2-/- mutant and gpc4-/- mutants. A) UMAP showing cluster identification. B) Dot plot showing the most differentially expressed genes between clusters. C) UMAP showing dact1, dact2, and gpc4 expression in wildtype embryos.

Pseudobulk differential expression analysis of single-cell RNAseq data. A) Heatmaps showing the 50 most differentially expressed genes in 3 major cell types; ectoderm (clusters 4,5,6,7), paraxial mesoderm (clusters 10,11,12), and lateral plate mesoderm (clusters 15, 16, 17,18) between dact1/2-/- mutants and wildtype and gpc4-/- mutants and wildtype. B) Venn diagrams showing unique and overlapping DEGs in dact1/2-/- and gpc4-/- mutants. C) GO analysis of dact1/2-/- mutant-specific DEGs in lateral plate mesoderm showing enrichment for proteolytic processes.

Expression of capn8 is significantly dysregulated in dact1/2-/- mutants. A) Single-cell RNAseq gene expression analysis of capn8 in wildtype and dact1/2-/- mutants. In wildtype embryos, capn8 expression is restricted predominantly to the epidermis whereas capn8 is widely expressed throughout the embryo in dact1/2-/- mutants, especially in the mesoderm. B) Whole-mount in situ hybridization of capn8 expression in wildtype and dact1/2-/- mutant embryos at 2 ss. Staining corroborates the single-cell RNAseq data, with expanded ectopic expression of capn8 throughout the embryo. C) Brightfield images and alcian blue staining of the ANC show ectopic expression of capn8 mRNA (200 pg) at the 1-cell stage in wildtype embryos recapitulates the dact1/2-/- compound mutant craniofacial phenotype at a low frequency (1/142 injected embryos). The mutant craniofacial phenotype did not manifest in gfp mRNA (200 pg) injected 1-cell stage embryos (0/192 injected embryos).

Characterization of CRISPR/Cas9 generated dact1-/- and dact2-/- mutants. A) Schematic representations of dact1 and dact2 exons, positions of guide RNA target site, introduced premature stop codon (arrow), and sequences of mutations. B) Expression levels of dact1 and dact2 mRNA by RT-qPCR in 12 hpf dact1-/- mutants, dact2-/- mutants, and dact1/2-/- compound mutants. 8 embryos were pooled for mRNA isolation per sample. C) Injection of dact1 mRNA, dact2 mRNA, or a combination of dact1 and dact2 mRNA rescues the rod-shaped ANC phenotype in dact1/2-/- compound mutants. Representative images of Alcian blue stained dact1/2-/- double mutant treated with 300 pg dact1 mRNA and 300 pg dact2 mRNA. Arrow highlights normal ANC. D) Quantification of the mutant craniofacial phenotype observed in a dact1-/-,dact2+/- breeding in-cross. Without mRNA injection, the mutant phenotype was observed at approximately the expected Mendelian ratio of 25%. Injection with dact1 mRNA, dact2 mRNA, or a combination of dact1 and dact2 mRNA decreased the frequency that the mutant craniofacial phenotype was observed.

Loss of gpc4 and loss of dact1/2 leads to distinct changes in gene expression profiles but with some overlapping functions. GO analysis of DEGs identified between gpc4-/- and wildtype embryos and dact1/2-/- and wildtype embryos identified changes in calcium ion binding and actin interaction in both mutants.