crasher mutants have fewer and lighter melanophores. (a, a') Lateral (a) and dorsal (a’) view of crasher siblings with typical the wild-type (WT) phenotype at 3 days postfertilization (3 dpf). (b, b′) Lateral (b) and dorsal (b′) view of homozygous recessive (mutant) phenotype. Arrowheads indicate punctate melanophores. (c) Melanophore count boxplot at 3 dpf. Mutants average 25% fewer melanophores (μ = 188, SD = 20.4) than WT siblings (μ = 255, SD = 20.9). (Student's t-test, t[28] = −8.87, p = 1.27 × 10⁻⁹***). (d) Iridophore count boxplot at 3 dpf. Mutants vary more (median = 25, interquartile range [IQR] = 23 to 29.5), but have similar iridophore numbers as WT siblings (median = 28, IQR = 22–29) (Mann–Whitney U test, W = 11.05 × 10¹, p = 0.95). Scale bar is 1 mm.

crasher mutant phenotype recapitulation using CRISPR knockout. (a) The crasher mutation maps to a ~ 375 kb region on chromosome 22. R = recombination event. (b–g) CRISPR ap3d1 knockout in *AB lines recapitulates the crasher phenotype. (b) Uninjected *AB control. (c) tyr gRNA injected *AB fish have mosaic tyrosinase activity. (d) crasher mutant with slight lightening of the retinal pigmented epithelium. (e) ap3d1 gRNA injected *AB fish display light, punctate crasher-like melanophores and a lighter retinal pigmented epithelium than crasher mutant. (f) Melanophore count boxplot in uninjected and injected tyr and ap3d1 fish. Melanophore number is reduced in both populations injected with gRNA (two-way ANOVA, guide F(1, 56) = 5.50, p = <0.02; injection status F(1, 56) = 247.50, p = <0.22 × 10⁻¹⁵; interaction F(1, 56) = 23.68, p = <9.68 × 10⁻⁶, Tukey HSD performed to determine group differences, p = <0.001***). (g) Boxplot shows iridophore number is unchanged between uninjected and injected for both guides (two-way ANOVA, injection status F(1, 56) = 0.02, p = 0.88; guide F(1, 56) = 46.05, p = 7.80 × 10⁻⁹; interaction F(1, 56) = 0.89, p = 0.35, Tukey HSD performed to determine group differences, p = <0.001***). Scale bar is 500 μm. 78 eggs injected for tyr; 58 eggs injected for ap3d1.

ap3d1 gene expression is reduced overall in crasher mutants, but exon 14 of ap3d1 is overexpressed. (a) RNA-Seq data bar graph shows no reduction in any candidate gene in crasher mutants as compared to *AB to a log₂ fold difference of 1. Log₂ fold difference is the difference in mutant and *AB expression. However, ap3d1 expression is reduced the most. Negative values indicated reduced expression in crasher mutants as compared to *AB. A difference of 1 is considered of import and equivalent to a twofold difference. Significance testing was not performed. (b) ap3d1 expression is reduced in crasher mutants according to qRT-PCR (Student's t-test, t[4] = 2.94, p = 0.04*, n = 3). Error bars are standard deviation. (c) Exon 14 overexpression in ap3d1 line graph. The normalized reads ratio is defined as the normalized (transcripts per million) number of crasher reads at that exon divided by the normalized number of *AB reads at that exon.

Specification and early differentiation markers are unchanged in crasher (ap3d1 mutants). (a and b) Melanoblast/iridoblast bipotent precursor marker, foxd3, expression is unchanged in progeny of crasher heterozygotes. (c–h) mitfa expression is not altered in crasher heterozygote progeny. Melanoblast marker, mitfa, expression at 24 h postfertilization (hpf) in *AB control (c) and crasher heterozygote progeny (d) is not different. (e–h) Representative early differentiated melanophore marker, mitfa, expression at 30 hpf in *AB control (e) and crasher heterozygote progeny (f) from 2 replicates is not different. (g) mitfa + cell counts are not significantly different for early differentiating melanophores at 30 hpf between progeny of *AB parents and crasher heterozygote parents (Student's t-test, t(49) = 0.36, p = 0.72). (h) crasher heterozygote progeny have similar mitfa + signal, but less mitfa + signal on the ventral side. Ratio is mitfa + dorsal cell number divided by mitfa + ventral cell number (Student's t-test, t(45) = −0.94, p = 0.35). Plots are violin box plots. Scale bars are 500 μm. Images labeled ^‡crasher are representative fish from the clutch.

Melanogenesis gene expression is reduced in crasher (ap3d1) mutants. (a, b and e) dct expression at 24 hpf is reduced in 1/4 of crasher heterozygote progeny (b) as assessed by a chi-square goodness-of-fit test to a 3 wild-type phenotype:1 mutant phenotype ratio (χ² = 1.08, χ = 3.84, p = 0.30). dct + cell number is significantly reduced in mutant phenotype fish at 24 hpf (Student's t-test, t[6] = 2.83, p = 0.03*). (c, d and f) dct expression is reduced in the crasher clutch overall by 30 cells on average at 30 hpf (Student's t-test, t[6] = 2.39, p = 0.02*). (f) crasher clutch cell counts at 30 hpf segregate into a 3:1 ratio (χ² = 1.61, χ = 3.84, p = 0.20). (g–l) tyr is not significantly reduced at 24 hpf (Student's t-test, t[57] = 0.90, p = 0.38) (g, h and k), but is slightly reduced at 30 hpf in the crasher clutch (Student's t-test, t[48] = 2.14, p = 0.04*) (i, j and l). (m-r) tyrp1b + cells are reduced at 24 hpf in the crasher clutch (Student's t-test, t[57] = 3.51, p = 8.73 × 10⁻⁴***), but no obvious mutant phenotype is observed (m, n and q). At 30 hpf, tyrp1b + cells are significantly reduced (Student's t-test, t[52] = 2.66 p = 0.01**), but no obvious mutant phenotype is observed (o,p, and r). (s) Expression heatmap of melanogenesis genes at 3 dpf using RNA-Seq. tyrp1b is reduced by >twofold (log₂ fold difference of 1), while dct is also reduced. tyr expression is slightly increased. mitfa expression is not altered. (a, b, g, h, m and n) Scale bars are 500 μm. (c, d, i, j, o and p) Scale bars are 1 mm. tyrp1b 24 hpf and tyr 30 hpf experiment repeated twice. dct 30 hpf experiment repeated once. Plots are violin box plots. Images labeled ^‡crasher are representative fish from the clutch.

Autophagy inhibition with bafilomycin A1 negatively impacts melanophore survival in crasher (ap3d1) mutants. (a–d) Melanophore morphology of mutant fish treated with 50 nM bafilomycin A1 appears more punctate. (e) Melanophore count boxplot of crasher mutants and wild-type siblings treated with bafilomycin A1 or DMSO. Bafilomycin A1-treated mutants have 39.51% fewer melanophores (μ = 39.2, SD = 4.75) than DMSO-treated mutants (μ = 64.8, SD = 6.76). Bafilomycin A1-treated wild-type siblings have 15.74% fewer melanophores (μ = 68.0, SD = 8.70) than DMSO-treated wild-type siblings (μ = 80.7, SD = 3.39). (two-way ANOVA, phenotype F(1, 20) = 77.03, p = 2.71 x 10⁻⁸, treatment F(1, 20) = 56.74, p = 2.92 × 10⁻⁷, interaction F(1, 20) = 6.53, p = 0.02, Tukey HSD performed to determine group differences, p = <0.001 ***, p = <0.01 **, p = <0.05*). Scale bar is 500 μm.