Fig 1.
Medaka sox10a and sox10b are expressed in neural crest and their loss of function affects pigment cell development. (A-E) sox10a. (F-J) sox10b. (E, J) Transverse sections. (A, B, C, D, F, G, H, I) Lateral views. (A’, F’) Dorsal views. (K-O) 9 dpf. Lateral views. (A, F) At 6-somite stage (6s, 34 hpf), both sox10a and sox10b are expressed in premigratory neural crest cells (black arrows). (B-E, G-J) At 24-somite stage (24s, 58 hpf), sox10a- and sox10b- expressing cells are found in glial precursors on the posterior lateral line nerve (black arrowheads in C, E, H and J) and migrating neural crest cells between neural tube and somite (medial pathway, white arrows in D, E, I and J) and between epidermis and somite (lateral pathway, white arrowheads in C, E, H and J). (B, G) In head, both sox10a and sox10b are expressed in otic vesicle (ov) and cranial ganglia (g). (C, H) Close-up images of lateral trunk surface. (D, I) Close-up images of migrating neural crest through medial pathway. (K, L) WT and sox10b-/- hatchlings show a normal pigment pattern, composed of dorsal stripe (DS), lateral stripe (LS), ventral stripe (VS) and yolk sac cluster (YSC). (M) In sox10a-/- mutants, melanocytes in LS (*) and in posterior side of VS (from caudal end of yolk sac to tail, **) are severely reduced. (N) sox10a-/-;sox10b+/- mutants retain melanocytes only in the head and the anterior part of the body. (O) sox10a-/-;sox10b-/- mutants lack all chromatophores, except that leucophores on the head are retained (red arrows, see also Fig 4D). Scale Bars: (A) 200 μm, also applied to A’, F and F’; (B) 200 μm, also applied to G; (E) 20 μm, also applied to J; (K) 250μm, also applied to L, M, N and O.
Fig 2.
Reductions in melanoblast and iridoblast specification in sox10a mutants were partially rescued by loss of sox5. (A-D) mitfa. (E-H) dct. (I-L) ltk. (A, E, I) WT. (B, F, J) sox5-/-. (C, G, K) sox10a-/-. (D, H, L) sox10a-/-;sox5-/-. (M-O) Quantitation of number of cells expressing each fate marker in the whole embryonic body (M, N) and in the lateral patches (O). (A-H) Lateral views. (I-L) Dorsal views. (A-H) 30-somite stage (30s, 64 hpf). (I-L) 34-somite stage (34s, 74 hpf). The box in (A) indicates the enlarged region (anterior trunk) in insets. Melanoblasts, defined by expression of mitfa and dct, are unaltered in the sox5-/- mutant (B, F) as compared with WT (A, E). Expression of mitfa and dct are both dramatically reduced in sox10a-/- mutants (C, G), but substantially recovered in sox10a-/-;sox5-/- mutants (D, H). The number of mitfa (M, mitfa+)- or dct (N, dct+)- expressing cells is not different between WT (mitfa, n = 9; dct, n = 14) and sox5-/- mutants (mitfa, n = 15; dct, n = 15). p-values are p = 0.78 (mitfa) and p = 0.52 (dct). sox10a-/- mutants have significantly fewer of those cells (mitfa, n = 13; dct, n = 12) than WT (p<0.05 (mitfa) and p<0.05 (dct)). In the sox10a-/-;sox5-/- double mutants (mitfa, n = 10; dct, n = 13), cell counts are significantly increased as compared with sox10a-/- single mutant (*, p<0.05). p-values were calculated by Mann-Whitney test. Iridoblasts, as evidenced by ltk expression, in WT (I), sox5-/- mutant (J), sox10a-/- mutant (K) and sox10a-/-;sox5-/- mutant (L) are shown for the lateral patches, a region of concentrated iridophores dorsolateral to the developing gut in the anterior trunk. The number of ltk-expressing cells (O, ltk+) is indistinguishable between WT (I, n = 20) and the sox5-/- mutant (J, n = 10) (p = 0.98). In the sox10a-/- mutant (K, n = 18), the number is significantly decreased as compared with WT (I). Again, iridoblasts cell counts between the sox10a-/- mutant and the sox10a-/-;sox5-/- mutant (L, n = 23) is statistically different (*, p<0.05), showing partial rescue in the double homozygote. p-values were calculated by Mann-Whitney test. (M-O) Bars show mean and error bar (s.d.). Scale bars: (A) 200 μm, (I) 40 μm.
Fig 3.
Cooperative function of Sox10 and Sox5 is required for xanthophore formation (A-H) 9 dpf hatchlings.
Lateral views. UV images. (I-M) 5 dpf embryos. In WT (A), xanthophores (white arrows) are widely scattered under the skin, concentrated in the dorsal body. Xanthophores are reduced in sox10a-/- mutants (B), and further reduced in sox10a-/-;sox10b+/- mutants while a few residual cells remain in the anterior body (C). sox10a-/-;sox10b-/- mutants completely lack xanthophores (D). Loss of sox5 function results in complete absence of xanthophores, regardless of sox10 mutant status (E-H). Xanthoblasts immunostained with anti-Sox5 antibody (I-L) were quantified in the body. (M) Specific signals cannot be detected in sox5-/- embryos. sox5+/- embryos (J) have significantly fewer Sox5-positive xanthoblasts than WT (I) (N, WT, n = 12; sox5+/-, n = 8, *p<0.05). Further reduction is observed in sox10a-/- mutants (K) (N, n = 26). The number in sox10a-/-;sox5+/- mutants (L) (N, n = 19) is comparable to sox10a-/- embryos (p = 0.288). Comparison between the genotypes was performed by Kruskal-Wallis test with Steel-Dwass-Critchlow-Fligner (SDCF) post hoc test. The compound mutant of sox10a+/-;sox10b-/- results in reduced numbers of Sox5-positive xanthoblasts compared to WT (O, WT, n = 16; sox10a+/-;sox10b-/-, n = 16). In sox10a+/-;sox10b-/-;sox5+/- mutants (n = 13), the number of Sox5-positive cells is reduced compared to both sox5+/- (n = 15) and sox10a+/-;sox10b-/- mutants (O). Comparison between the genotypes was performed by Kruskal-Wallis test with SDCF post hoc test. **, p<0.05. (N, O) Bars show mean and error bar (s.d.). Scale bars: (A, I) 200 μm.
Fig 4.
Leucophore formation does not require Sox10 function but is repressed by Sox10 and Sox5 (A-H) 9 dpf.
Dorsal views. Leucophores (orange) are distributed along the dorsal surface throughout the anterioposterior axis of WT, as scattered individual cells in the head, and along the midline in the body (A). As the number of functional sox10 alleles decreases, leucophore numbers on the head are progressively increased (A-D, I, p<0.05 by Kruskal-Wallis test), whereas in the body they are progressively decreased (J, p<0.05 by Kruskal-Wallis test). (I, J) Comparison between genotypes was performed by Kruskal-Wallis test with SDCF post hoc test. Statistical significance (p<0.05) was detected between the groups shown by letters (a, b or c) above each scatter plots. sox10a-/- mutants have some ectopic leucophores on dorsal trunk (B, arrowheads). In sox10a-/-;sox10b+/- mutants, trunk leucophores are restricted to the anterior dorsal region (C). In sox10a-/-;sox10b-/- mutants, leucophores are largely restricted to the head (D). Intriguingly, the total number of leucophores in the whole body is not statistically different among these genotypes (K, p = 0.51 by Kruskal-Wallis test), although numbers do become much more variable after loss of sox10a. (I-K) WT, n = 19; sox10b-/-, n = 16; sox10a-/-, n = 27; sox10a-/-;sox10b+/-, n = 22; sox10a-/-;sox10b-/-, n = 18. In sox5-/- mutants, leucophores are formed in excess (L) and scatter more laterally (E). In sox10a-/-;sox5-/- mutants (F), there is a further increase in leucophores, which are scattered laterally like in sox5-/- mutants (E). Abundant leucophores (greater than in sox5-/- mutants; L) are observed in sox10a-/-;sox10b+/-;sox5-/- mutants, but these are more scarce in the body from posterior trunk backwards (G). In sox10a-/-;sox10b-/-;sox5-/- mutants, most leucophores are located on the head, but with some in the anterior trunk (H). (L) WT, n = 12; sox5-/-, n = 20; sox10a-/-;sox5-/-, n = 32; sox10a-/-;sox10b+/-;sox5-/-, n = 23; sox10a-/-;sox10b-/-;sox5-/-, n = 13. Comparison between the genotypes was performed by Kruskal-Wallis test with SDCF post hoc test. Statistical difference (p<0.05) was detected between the groups shown by letters (a, b or c) above each scatter plots. (I-L) Bars show mean and error bar (s.d.). Scale bar: (A) 200 μm.
Fig 5.
Sox10-mediated pigment cell formation is modulated by Sox5 in zebrafish. (A, B, C, E) Swim bladder inflation stage (10 dpf). Lateral views. (C, E) UV images. (D, F) 24 dpf. Lateral views. (G, I-O) 4 dpf. (G) Dorsolateral view. (I) Lateral view. (J-O) Fluorescing xanthophores. Dorsal views anterior to the left. The sox5 mutant is indistinguishable from WT, exhibiting four stripes of melanocytes (A, B) and having fluorescing xanthophores (C, E) and gch-expressing xanthoblasts (D, F). Melanocytes are almost completely absent from sox10baz1/baz1 mutant (S8B Fig), but are partially recovered in sox10baz1/baz1 mutants that have also lost sox5 WT allele(s) (G, H). The ratio of the embryos without melanocytes or with more than one melanocytes was compared among genotypes by Chi-squared test (*p<0.05). A few xanthophores develop on surface of the head in sox10baz1/baz1 mutant (S8E, S8I and S8J Fig), and reduction of sox5 is likely to elevate xanthophore formation (K, L). Whereas sox10t3/t3 mutant almost completely lacks xanthophores (M), a few xanthophores are rescued as the sox5 WT allele(s) are reduced (N, O). The counts are shown for xanthophores on the t3 background (P). Comparison between the genotypes was performed by Kruskal-Wallis test with SDCF post hoc test. **p<0.05. A substantial number of iridophores are formed in sox10baz1/baz1 mutants (S8G and S8H Fig), and the counts are not significantly altered with reduction of the sox5 WT allele(s) (Q, p = 0.775 by Kruskal-Wallis test). Iridophores are almost completely lost in sox10t3/t3 mutant (S8I Fig), but are partially recovered in sox10t3/t3;sox5+/- and sox10t3/t3;sox5-/- mutants (R **p<0.05 by Kruskal-Wallis test with SDCF post hoc test). (H) sox10baz1/baz1;sox5+/+, n = 22; sox10baz1/baz1;sox5+/-, n = 34; sox10baz1/baz1;sox5-/-, n = 33. (Q) sox10baz1/baz1;sox5+/+, n = 46; sox10baz1/baz1;sox5+/-, n = 73; sox10baz1/baz1;sox5-/-, n = 47. (P) sox10t3/t3;sox5+/+, n = 7; sox10t3/t3;sox5+/-, n = 19; sox10t3/t3;sox5-/-, n = 16. (R) sox10t3/t3;sox5+/+, n = 24; sox10t3/t3;sox5+/-, n = 48; sox10t3/t3;sox5-/-, n = 40. (H, P-R) Bars show mean and error bar (s.d.). Arrowheads point to weakly melanised cells (G) and at the corresponding position of the head (I-O). Scale bars: (A, D) 200 μm.
Fig 6.
Role of Sox5 in adult xanthophore development in medaka remains opposite to that in zebrafish. (A, B) Medaka. 6 mpf. (D, E) Zebrafish. 2 mpf. (D’, E’) Schematics of stripe pattern. (D”, E”) Enlarged images of X0 interstripe. (F, G) Zebrafish. 1.5 years old. In WT medaka (A), melanocytes, xanthophores (yellow arrowhead) and leucophores (white arrowhead) are scattered on the body surface. sox5-/- mutant medaka (B) have fewer xanthophores and more leucophores than WT (C, Xan and Leu, *p<0.05 by Mann-Whitney test; WT, n = 5; sox5-/-, n = 5; Bars show mean and error bar (s.d.)). Pigment cells were counted from a 1 mm2 area on the dorsal body surface. The melanocyte numbers were not significantly different between WT and sox5-/- (C, Mel, p = 0.1 by Mann-Whitney test). The adult pigment pattern of WT zebrafish is composed of 5 melanocyte stripes (2D, 1D, 1V, 2V, 3V) and xanthophore interstripes (X1D, X0, X1V, X2V) (D, D’). Zebrafish sox5-/- mutants lack two ventral interstripes (2V and 3V) (E, E’), and thus have fewer stripes than WT. This is the case regardless of sex (F, female; G, male) after the mutant fish get larger and older than 1.5 years. The sox5-/- zebrafish mutants have wider X1D and X0 (two-way arrow in D” and E”) interstripes and larger numbers of xanthophores in these interstripes than WT. (H) Scatter plot of stripe width or pigment cell numbers in each stripe, comparing sox5-/- and their WT siblings. X axis shows the standard body length of zebrafish examined. Analysis of covariance was performed to examine the differences in width or cell numbers between WT and sox5-/- mutants, by using standard length as a covariate. The p values are as follows; width X1D (p<0.05), 1D (p<0.05), X0 (p<0.05), cell number X1D (p<0.05), 1D (p = 0.985), X0 (p<0.05). The width and number of xanthophore in the xanthophore stripes (X0, X1D) in sox5-/- mutants (black boxes) show significant increase compared with WT siblings (white boxes). The number of melanocyte in the 1D stripe is comparable between WT (white boxes) and sox5-/- mutant (black boxes), but the width is slightly but significantly different (p<0.05). Scale bar: (A) 200 μm, (D) 3 mm, (D”) 200 μm, (F) 3 mm.
Fig 7.
Misexpression of Sox5 can reverse the xanthophore vs leucophore fate choice in medaka. (A) Transgenic construct. (B-C”) 3 dpf. (B, C) Bright field. (B’, C’) Dsred. (B”, C”) Venus. (D-G) 7 dpf. Dorsal views. (D’, E’) lateral views. UV images. (F’, G’) Magnified images of the head in (F, G). (B, D) sox5-/-;Tg(pax7a:loxp-dsred-loxp-sox5WT). (C, E) sox5-/-;Tg(pax7a:loxp-dsred-loxp-sox5WT) injected with cre mRNA. (F) WT(sox5+/+);Tg(pax7a:loxp-dsred-loxp-sox5WT);Tg(hsp70:cre) without heat shock. (G) WT;Tg(pax7a:loxp-dsred-loxp-sox5WT);Tg(hsp70:cre) with heat shock. A transgenic fish in which the pax7a promoter drives sox5-2A-venus expression once the upstream dsred region is excised at the flanking loxp sites by Cre recombinase TgBAC(pax7a:loxp-dsred-loxp-sox5WT) was used for Sox5 misexpression in a shared xanthophore/leucophore progenitor. In the absence of cre mRNA, sox5-/-;Tg embryos (sox5-/-;TgBAC(pax7a:loxp-dsred-loxp-sox5WT)) were strongly expressing Dsred (B-B”, n = 8). When cre mRNA was injected in the 1-cell sox5-/-;Tg embryos, they became positive for Venus (C-C”, n = 7). Venus-expressing animals showed rescued xanthophore and leucophore phenotypes in sox5-/- mutants (D, E). When sox5+/+ WT embryos obtained from crossing the transgenic fish Nagoya (sox5+/+);TgBAC(pax7a:loxp-dsred-loxp-sox5WT) with Tg(hsp70:cre) were heat shocked, Dsred became barely detectable. Leucophores (Leu) were significantly reduced in the hatchlings treated with heat shock (+) compared with those without heat shock (-) whereas melanocyte (Mel) formation was not altered by heat shock (F-H). (H) Bars show mean and error bar (s.d.). Comparison between with and without heat shock was performed by paired t-test (*p<0.0001).
Fig 8.
Based on a new working model for pigment cell development in medaka and zebrafish based on our previous work in zebrafish [46], we propose that a chromatoblast, a multipotent common progenitor of chromatophores, gives rise to three pigment cell types, melanocytes (M), iridophores (I) and xanthophores (X) in zebrafish, development of which is all dependent on SoxE with antagonistic modulation by Sox5. Likewise in medaka, Sox5 counteracts SoxE in the melanocyte and iridophore lineages and, by extension, presumably in formation of xanthophore/leucophore progenitor (XL). In an additional step, an evolutionary novelty of medaka, Sox5 likely functions in collaboration with Sox10 to promote xanthophore specification, but to repress leucophore (L) formation, from a novel shared progenitor.