Full-sibling heterozygous mef2ca pairs were intercrossed and their homozygous mutant offspring were scored for penetrance of ectopic bone near the opercle. Mutant progeny penetrance scores were assigned to each breeding pair. Parents are color coded for offspring penetrance on this pedigree. Heterozygous offspring from selected pairs were raised for the next round of selection. A single low-penetrance individual from generation four was crossed to a double-transgenic mef2ca wild-type unselected individual (black circle). Indicated individuals from recent generations in the low-penetrance line are viable as mef2ca homozygous mutants. Indicated pairs from recent generations in the high-penetrance strain produce heterozygous offspring with mef2ca-associated phenotypes. Uncolored circles represent homozygous mutant adults that were recovered but were not tested for mutant offspring ectopic bone phenotype penetrance.

(A) Wild-type 6 days post fertilization (dpf) zebrafish larvae were stained with Alcian Blue and Alizarin Red to label cartilage and bone then flat mounted and imaged. The following craniofacial skeletal elements are indicated in the micrograph and graphic: opercle bone (op), branchiostegal ray (br), Meckel’s (mc), ceratohyal (ch), symplectic (sy) and hyomandibular (hm) cartilages, interhyal (ih) and jaw (jj) joints. Scale bar: 50 μm. (B, C) Penetrance of most phenotypes associated with mef2ca are reduced in homozygous mutants from the low-penetrance strain compared to homozygous mutants from the high-penetrance strain including: ectopic bone (arrowheads), interhyal- and jaw-joint fusions (^), dysmorphic ch (arrows), reduced mc (double arrowhead). A shortened sy cartilage (red arrows) remains fully penetrant in both strains. (D) Penetrance of mef2ca-associated phenotypes from full-sibling low- and high-penetrance families. Asterisks indicate statistically significant penetrance differences between homozygous mutants in low-penetrance compared with high-penetrance strains (Fisher’s exact test at alpha = 0.05).

6 dpf larvae were stained with Alizarin Red (magenta) and fluorescently labelled phalloidin (green) to mark bones and muscles, respectively. (A) The following craniofacial muscles are indicated in the micrograph and graphic: intermandibularis ant. (ima), intermandibularis post. (imp), interhyal (ih), hyohyal (hh), sternohyoideus (sh). Scale bar: 50 μm. (B) In the low-penetrance strain, mef2ca mutant muscles and bones resemble wild types. (C) In the high-penetrance strain, mef2ca mutants manifest muscle phenotypes (asterisks) in conjunction with ectopic bone phenotypes (arrowheads).

(A) Low-penetrance mef2ca heterozygous adults were intercrossed and offspring were grown to adulthood. Fin amputation and genotyping identified mef2ca homozygous adults, which were imaged alongside wild-type siblings. Scale bar: 1mm. (B) 6 dpf full-sibling individuals from a single high-penetrance family were stained with Alizarin Red (bone) and Alcian Blue (cartilage), then genotyped and mef2ca heterozygotes and homozygous wild types were scored for presence of all mef2ca-associated phenotypes. (C) 6 dpf Alizarin Red and Alcian Blue stained skeletons were dissected and flat mounted, lateral view. In these individuals, a shortened symplectic (sy) cartilage phenotype is evident in mef2ca heterozygotes from the high-penetrance strain and mef2ca homozygous mutants from the low-penetrance strain. Scale bar: 50 μm.

(A) Head expression of all mef2 paralogs was quantified by RT-qPCR in full-sibling mef2ca mutant and wild-type embryos from low- and high-penetrance strains at 28 hpf. Expression of each paralog in mef2ca mutants was normalized to wild-type sibling expression levels (black line). Asterisks mark paralogs with expression levels significantly different in mef2ca mutants compared with mef2ca wild types (T-test at alpha = 0.05). Error bars are standard deviation. (B) Animals heterozygous for mef2ca from the low-penetrance strain were intercrossed and offspring treated with nonsense-mediated decay inhibitor (NMDi14) or vehicle control (DMSO) from 18–42 hpf. Treated animals were fixed and stained with Alcian Blue (cartilage) and Alizarin Red (bone) at 6 dpf. Genotyped skeletal preparations were scored for penetrance of mef2ca-associated phenotypes. (C) Intercrosses from low-penetrance heterozygotes were immunostained with anti-MEF2C antibody. Genotyped animals were imaged in whole mount. 9/9 homozygous wild types had strong signal in the pharyngeal arches, 0/9 homozygous mutants exhibited any signal in the pharyngeal arches. Scale bar: 50 μm. (D) Exonic structure of mef2ca with protein domains and the location of the mef2ca^b1086 mutation indicated. Primer pairs either span the mutant exon (F1/R1) or amplify across exon-exon boundary (F2/R1). (E) cDNA from wild-type and mutant siblings from low- and high-penetrance strains was used as a template for amplification across the mutant exon. These primers are predicted to amplify bands of indicated different sizes if the mutant exon is included versus excluded from mature transcripts. (F) cDNA from wild-type animals was collected before and after zygotic genome activation and used as a template for PCR amplification across an exon-exon junction. Reverse transcriptase (RT) negative control ensured that only cDNA and not genomic DNA was amplified under these conditions. (G) A homozygous mutant female was crossed with a heterozygous male sibling from the low-penetrance strain and offspring were fixed and stained with Alcian Blue (cartilage) Alizarin Red (bone) at 6 dpf. Genotyped maternal-zygotic mutants were scored for ectopic bone which develops between the opercle (op) and the branchiostegal ray (br) in mutants from the high-penetrance strain. Scale bar: 100 μm.

(A) Head expression of dlx5a was quantified by RT-qPCR in full-sibling mef2ca mutant and wild-type embryos from low- and high-penetrance strains at 28 and 48 hpf. Expression levels for all conditions were normalized to mef2ca wild types at 28 hpf. Asterisks at 28 hpf mark significantly different expression levels between mef2ca mutants compared with mef2ca wild types within each strain. Asterisks at 48 hpf mark significantly different expression levels between low- and high-penetrance strains within each mef2ca genotype (T-test at alpha = 0.05), error bars are standard deviation. (B) Heterozygous dlx5a^j1073Gt adults from an unselected background were intercrossed, 6 dpf offspring were fixed and stained with Alcian Blue (cartilage) and Alizarin Red (bone). Stained skeletons were genotyped, and flat mounts were imaged. Arrows mark ectopic cartilage nubbins, caret indicates jaw-joint fusion. Scale bar: 50 μm. (C) Penetrance scores for phenotypes associated with dlx5a heterozygotes and mutants in an unselected background. Asterisk indicates significance at alpha = 0.05 Fisher’s exact test compared with dlx5a+/+. (D) Head expression of dlx5a was quantified by RT-qPCR in full-sibling dlx5a wild-type and mutant embryos at 28 hpf. Expression in dlx5a mutants was normalized to wild-type sibling expression levels (black line). Asterisk marks expression level significantly different in dlx5a mutants compared with dlx5a wild types (T-test at alpha = 0.05). Error bars are standard deviation.

(A) Animals double heterozygous for mef2ca and dlx5a were intercrossed, raised to 6 dpf and fixed and stained with Alcian Blue (cartilage) and Alizarin Red (bone). Genotyped animals were flat mounted and imaged. In this family, dlx5a homozygous mutants appear wild type. But, removing functional copies of dlx5a increases mef2ca mutant penetrance. This includes phenotypes that we interpret to be ventral-to-dorsal transformations: br-to-op transformation (arrowheads), and ch-to-hm transformation (red arrow). Scale bar: 50 μm. (B) Genotyped mef2ca;dlx5a skeletal preparations were scored for penetrance of mef2ca-associated phenotypes. Asterisks indicate significance at alpha = 0.05 Fisher’s exact test compared with dlx5a+/+. (C, D) Genotyped mef2ca;dlx5a skeletal preparations were scored for penetrance of dlx5a-associated phenotypes.

(A) Animals heterozygous for mef2ca from the high-penetrance strain were intercrossed and offspring treated with Notch inhibitor (DBZ) or vehicle control (DMSO) from 18–48 hpf. Treated animals were fixed and stained with Alcian Blue (cartilage) and Alizarin Red (bone) at 6 dpf. Genotyped skeletal preparations were scored for penetrance of mef2ca-associated phenotypes. Asterisks indicate significance at alpha = 0.05 Fisher’s exact test compared with DMSO. (B) Imaged flat mounts of dissected skeletons from stained, genotyped animals treated with DMSO or DBZ were imaged. Arrows indicate dysmorphic ch and arrowheads mark ectopic bone. Scale bar: 50 μm. (C) Animals heterozygous for mef2ca from an unselected strain were intercrossed and offspring treated with either DMSO from 18–48 hpf, or Notch inhibitor (DBZ) from 18–30, or DBZ from 30–48 hpf. Treated animals were fixed and stained with Alcian Blue (cartilage) and Alizarin Red (bone) at 6 dpf. Genotyped skeletal preparations were scored for penetrance of mef2ca-associated phenotypes. Asterisks indicate significance at alpha = 0.05 Fisher’s exact test compared with DMSO.

(A) Animals double heterozygous for mef2ca and jag1b were intercrossed, raised to 6 dpf and fixed and stained with Alcian Blue (cartilage) and Alizarin Red (bone). Genotyped animals were flat mounted and imaged. Removing functional copies of jag1b decreased mef2ca mutant penetrance. Scale bar: 50 μm. (B) Genotyped mef2ca;jag1b skeletal preparations were scored for penetrance of mef2ca-associated phenotypes. Asterisks indicate significance at alpha = 0.05 Fisher’s exact test compared with jag1b+/+. (C) Genotyped mef2ca;jag1b skeletal preparations were scored for penetrance of jag1b-associated phenotypes. Asterisks indicate significance at alpha = 0.05 Fisher’s exact test compared with mef2ca+/+.

(A) Head expression of jag1b was quantified by RT-qPCR in full-sibling mef2ca mutant and wild-type embryos from low- and high-penetrance strains at 28 and 48 hpf. Expression of jag1b in mef2ca mutants was normalized to wild-type sibling expression levels (black line). Asterisks mark expression levels significantly different in mef2ca mutants compared with mef2ca wild types (T-test at alpha = 0.05), error bars are standard deviation. (B) Head expression of the Jag/N transcriptional targets hey1 and her6 was quantified by RT-qPCR in full-sibling jag1b mutant and wild-type embryos at 28 and 48 hpf. Expression of each target in jag1b mutants was normalized to wild-type sibling expression levels (black line). Asterisks mark genes with expression levels significantly different in jag1b mutants compared with jag1b wild types (T-test at alpha = 0.05). Error bars are standard deviation. (C) Head expression of the Jag/N transcriptional targets hey1 and her6 was quantified by RT-qPCR in full-sibling mef2ca mutant and wild-type embryos from low- and high-penetrance strains at 28 and 48 hpf. Expression of each gene in mef2ca mutants was normalized to wild-type sibling expression levels (black line). Asterisks mark genes with expression levels significantly different in mef2ca mutants compared with mef2ca wild types (T-test at alpha = 0.05). Error bars are standard deviation. (D) Spatial expression of her6 was determined by in situ hybridization. her6 is expressed in three discrete patches in pharyngeal arch 2 (arrows) in low-penetrance wild types, low-penetrance mutants and high-penetrance wild types. In only high-penetrance mef2ca mutants her6 expands ventrally (arrowhead). dlx2a expression was used to delineate pharyngeal arches one (a1) and two (a2) (dashed outlines). In the high penetrance strain, her6 was expanded in 0/7 homozygous wild types and 7/8 homozygous mutants. In the low-penetrance strain, her6 was expanded in 0/3 homozygous wild types and 0/5 homozygous mutants. Scale bar: 30 μm.

(A) Animals triple heterozygous for mef2ca, dlx5a and jag1b were intercrossed, offspring were raised to 6 dpf and fixed and stained with Alcian Blue (cartilage) and Alizarin Red (bone). Genotyped skeletal preparations were scored for penetrance of mef2ca-associated phenotypes. Removing functional copies of jag1b from mef2ca;dlx5a double mutants decreases penetrance of mef2ca-associated phenotypes. Asterisks indicate significance at alpha = 0.05 Fisher’s exact test compared with mef2ca-/-;dlx5a+/+;jag1b+/+. (B) Genotyped mef2ca;dlx5a;jag1b skeletal preparations were flat mounted and imaged. mef2ca mutant associated phenotypes are indicated including ectopic bone (arrowheads), dysmorphic ch (arrows) and ch-to-hm transformation (red arrow). Stick-like opercle bone is indicated with a white arrow. Scale bar: 50 μm. (C) Head expression of Notch target genes was quantified by RT-qPCR in full-sibling dlx5a mutant and wild-type embryos at 48 hpf. Expression in dlx5a mutants was normalized to wild-type sibling expression levels (black line). Asterisks mark genes with expression levels significantly different in dlx5a mutants compared with dlx5a wild types (T-test at alpha = 0.05). Error bars are standard deviation. (D) Spatial her6 expression was monitored by in situ hybridization in wild types and dlx5a mutants. In wild-type embryos her6 is expressed in three discrete patches in arch 2 (arrows). In dlx5a homozygous mutants, her6 expands ventrally resulting in a larger, more continuous expression domain (arrowhead). dlx2a expression was used to delineate pharyngeal arches one (a1) and two (a2) (dashed outlines). her6 was expanded in 0/4 homozygous wild types and 3/3 homozygous mutants Scale bar: 30 μm.

(A) The liability-threshold model illustrates the inheritance of ectopic bone penetrance in mef2ca mutants following selective breeding. The difference between the mean of the population and the threshold is defined as x. x_F0-op = 0.468, x_high-op = -1.080 and x_low-op = 1.918. (B) The multiple threshold model illustrates how the same modifiers can shape liability and penetrance of different mef2ca mutant phenotypes to different extents. x_{high-jaw joint} = -4, x_{low-jaw joint} = -.806 x_high-op = -1.08 x_low-op = 1.476. (C) Model for how tuning down of Jag/N-driven transcription of her6 can result in mef2ca low penetrance. This genetic circuitry downstream of mef2ca is consistent with mutant, genetic epistasis, and gene expression analyses in this study.