Search tips
Search criteria 


Logo of wtpaEurope PMCEurope PMC Funders GroupSubmit a Manuscript
Dev Biol. Author manuscript; available in PMC 2010 December 21.
Published in final edited form as:
PMCID: PMC3005709

Sequential actions of Pax3 and Pax7 drive xanthophore development in zebrafish neural crest


The Pax3/7 gene family has a fundamental and conserved role during neural crest formation. In people, PAX3 mutation causes Waardenburg syndrome, and murine Pax3 is essential for pigment formation. However, it is unclear exactly how Pax3 functions within the neural crest. Here we show that pax3 is expressed before other pax3/7 members, including duplicated pax3b, pax7 and pax7b genes, early in zebrafish neural crest development. Knockdown of Pax3 protein by antisense morpholino oligonucleotides results in defective fate specification of xanthophores, with complete ablation in the trunk. Other pigment lineages are specified and differentiate. As a consequence of xanthophore loss, expression of pax7, a marker of the xanthophore lineage, is reduced in neural crest. Morpholino knockdown of Pax7 protein shows that Pax7 itself is dispensable for xanthophore fate specification, although yellow pigmentation is reduced. Loss of xanthophores after reduction of Pax3 correlates with a delay in melanoblast differentiation followed by significant increase in melanophores, suggestive of a Pax3-driven fate switch within a chromatophore precursor or stem cell. Analysis of other neural crest derivatives reveals that, in the absence of Pax3, the enteric nervous system is ablated from its inception. Therefore, Pax3 in zebrafish is required for specification of two specific lineages of neural crest, xanthophores and enteric neurons.

Keywords: Pax3, Pax7, Pax3b, Pax7b, sox10, foxd3, dct, gch, xdh, csf1r, mitfa, zebrafish, neural crest, pigmentation, xanthophores, melanophore, enteric nervous system


The neural crest (NC) is a transient population of cells that generates a remarkably diverse range of cell types during development of the vertebrate embryo (reviewed in Le Douarin et al., 2004). NC progenitors are contained within the dorsal neural tube before undergoing an epithelial to mesenchymal transition to become pre-migratory NC. These cells can migrate to various sites within the embryo and differentiate into neurons and glia of the sensory, autonomic and enteric branches of the peripheral nervous system, craniofacial skeleton, connective tissue and pigment cells (reviewed in LaBonne and Bronner-Fraser, 1999). It is thought that NC cells initially constitute a multipotent population that is specified to distinct lineages through a process of fate restriction, characterised by distinct gene expression (reviewed in Harris and Erickson, 2007). Some gene-regulatory factors required for NC induction and specification of particular lineages are conserved across vertebrate species (reviewed in Meulemans and Bronner-Fraser, 2004). However, adaptations of NC between species provide an excellent model in which to study the evolution of development.

The Pax3/7 sub-family of genes are expressed during gastrulation, in presumptive NC domains, and have essential roles in NC development. The Pax3 mutant mouse, Splotch, has dramatically reduced NC, leading to severe defects in all NC derivatives (Tremblay and Gruss, 1994). In Splotch, fewer NC cells emigrate from the neural tube (Serbedzija and McMahon, 1997). However, it is unclear exactly how Pax3 functions in the NC (Li et al., 1999; Mansouri et al., 2001). Data from Xenopus suggest that Pax3 is necessary for NC fate decisions in the neural plate border and loss of Pax3 results in reduced NC markers during initial NC determination (Hong and Saint-Jeannet, 2007; Monsoro-Burq et al., 2005; Sato et al., 2005). Recent data from a basal vertebrate, lamprey, has also demonstrated a critical role for a Pax3 homolog in NC formation (Sauka-Spengler et al., 2007). Thus, Pax3 appears to have a conserved and fundamental role during NC development.

Pax7 also has important roles during NC development. In chick, Pax7 is required for specification of all NC during gastrulation (Basch et al., 2006). In mice, however, Pax7 is not expressed early and mice mutant for Pax7 only have defects in cranial NC derivatives (Mansouri et al., 1996). An ancestral role of Pax3 in NC is also suggested by data from zebrafish; Pax3 is the earliest known marker of NC (Lewis et al., 2004; Seo et al., 1998). However, the function of Pax3 during formation of zebrafish NC is unknown. Thus, it is likely that functions within the Pax3/7 sub-family are distributed differently between species (Meulemans and Bronner-Fraser, 2004).

The role of Pax3 during pigment development has been well studied. In humans, the congenital pigmentation disorder, Waardenburg syndrome type I, is caused by a mutation in PAX3 (Tassabehji et al., 1992). Splotch mice have reduced pigmentation (Tremblay and Gruss, 1994). However, melanocyte precursors (melanoblasts) do form and migrate, although their number is greatly diminished, suggesting a role for Pax3 in expansion of melanoblasts (Hornyak et al., 2001). Genetically, Pax3 directly regulates the melanogenic lineage genes Mitf and TRP-1 (Galibert et al., 1999; Watanabe et al., 1998), and the Pax3-Mitf network even functions in ascidian pigmentation suggesting a highly conserved role for Pax3 during pigment formation (Yajima et al., 2003). Interestingly, SOX10 enhances PAX3 regulation of MITF (Bondurand et al., 2000; Potterf et al., 2000), and transcriptional regulation of mitfa by sox10 is essential for melanocyte development in zebrafish (Elworthy et al., 2003), suggesting that Pax3-Sox10-Mitf may also cooperate in zebrafish pigment development.

Zebrafish generate three types of pigment cell (chromatophore): xanthophores and iridophores in addition to melanocytes (called melanophores in fish). Xanthophores confer a yellow colour due to the presence of pteridine pigments (Odenthal et al., 1996), and iridophores contain light-reflecting platelets (Kelsh et al., 1996). In zebrafish, the mechanism by which the different classes of chromatophores are specified is complex, and has been suggested to involve a common chromatoblast precursor (Bagnara et al., 1979; Kelsh et al., 1996). Zebrafish mutants exist that ablate all pigment cell types. However, other mutants affect only one class of chromatophore, indicating distinct genetic regulation of chromatophore types (Kelsh et al., 1996; Odenthal et al., 1996). Melanophores are formed early and migrate ventrally from the dorsal neural tube region along two pathways: medial and lateral to the somite. In contrast, xanthophores differentiate later from cells that migrate on the lateral pathway (Odenthal et al., 1996; Parichy et al., 2000).

Here we examine the function of pax3 during formation of zebrafish NC. We show that of the four pax3/7 genes expressed in NC cells, pax3 mRNA accumulates first in NC progenitors and is down-regulated in migrating NC, suggestive of an early role in NC development. Knocking down Pax3 protein led to defects in specific NC cell populations: xanthophores and enteric neurons were ablated from the trunk, whereas other NC derivatives differentiated relatively normally. Loss of pax3 function led to reduction in pax7 expression in pre-migratory and migrating xanthoblasts, consistent with loss of the xanthophore lineage. Knockdown of Pax7 alone did not ablate xanthophore lineage cells, but diminished xanthophore pigmentation, demonstrating sequential roles for Pax3 and then Pax7 during xanthophore development. The expression of a second pax7 gene, pax7b, in xanthophores raises the possibility of functional redundancy between pax7 and pax7b. Knockdown of Pax3, and loss of xanthophores, resulted in increased melanophores, providing further evidence for the long-sought common chromatoblast precursor.

Materials and Methods

Zebrafish lines and maintenance

Transgenic lines Tg(−4.9sox10:eGFP)ba2 (Carney et al., 2006), Tg(elavl3:eGFP)zf8 (Park et al., 2000) and Tg(−8.4neurog1:nRFP)sb3 (Blader et al., 2003) were crossed onto a King’s wild type background and staging and husbandry were as described previously (Westerfield, 1995).

Embryo manipulations

Morpholino (MO; Gene Tools, USA; 1 nl with 0.1% phenol red as tracer) or water control was injected into 1-2 cell stage embryos as previously described (Westerfield, 1995). MOs used were pax3 MO: 4 ng 5′-CTAATGCGGTCATATCTCCTCTGCA-3′; pax3 MO2: 4 ng 5′-AAAGGATGCACGAAGCACTTGATAG-3′ and pax7 MO: 4 ng 5′-TTCCTGGTAAAGTAGCCATTCCAGC-3′. Each MO had greater than 8 mismatches to all other members of the zebrafish pax3/7 family (Table 1). pax3-BAC DNA injection was identical with 1 nl (70 pg) of CH211-20F20 (BACPAC Resources, USA). The pax3-BAC contains the complete pax3 gene and also the ACSL3 gene (long-chain-fatty-acid-CoA-ligase-C) (Acc. No. NC_007113.2). No overt morphological defects were observed in BAC-injected embryos that could be attributed to gene dosage effects of either gene.

Table 1
Fraction of mismatched bases between pax3/7 MOs and genes

In situ mRNA hybridization, immunohistochemistry and Western blots

Embryos were treated with 0.003% 1-phenyl-2-thiourea (PTU, Sigma) to reduce melanin formation. In situ mRNA hybridization was performed as previously described (Coutelle et al., 2001; Hammond et al., 2007). Digoxigenin-labeled probes used for in situ mRNA hybridization were pax3 and pax7 (Hammond et al., 2007; Seo et al., 1998), sox10 (Dutton et al., 2001), xdh, csf1r (formerly known as fms) and gch (Parichy et al., 2000), dct (Kelsh et al., 2000b), foxd3 (Kelsh et al., 2000a) and mitfa (Lister et al., 1999). A pax3b plasmid was the kind gift of Dave Raible (University of Washington, USA), a 724 bp pax3b probe was made via a PCR product containing the consensus T7 promoter sequence allowing synthesis of a riboprobe for in situ mRNA hybridization (sense: 5′-GGAGCCATCGGCGGGACTAAAT. Antisense: 5′- TAATACGACTCACTATAGGGAGAGCGCTCGAAAGCTCTTTCCA). The resulting pax3b probe had 72% identity to zebrafish pax3. A similar strategy was used to make a pax7b probe. An EST clone (EB775328) containing the full length pax7b cDNA was obtained and a 569 bp probe with 80% identity to zebrafish pax7 was made via PCR (sense: 5′- TGCCCACTCTACCAACATACCA. Antisense: 5′- TAATACGACTCACTATAGGGCTCTCCCAGCTTCATCCTTCTCT). Immunohistochemistry was performed as previously described (Blagden et al., 1997; Hammond et al., 2007). Primary antibodies used were Pax3/7 (DP312, Davis et al., 2005; Hammond et al., 2007, 1:50), monoclonal anti-chicken Pax7 (Kawakami et al., 1997, 1:5), rabbit anti-eGFP (Molecular Probes; 1:500), anti-HuC (Molecular Probes; 1:500) and mouse zn-12 (ZIRC, 1:250). Secondary antibodies used were either HRP-conjugated (Vector) or Alexa-conjugated (Molecular Probes, 1:1000) and specimens were mounted in Citifluor mountant (Agar) for imaging. Western blots were performed as described previously (Hammond et al., 2007). Molecular weights of Pax3/7 proteins were predicted using the Compute pI/Mw tool (


Confocal images were taken on a Zeiss LSM510 confocal microscope and images were processed using Zeiss LSM 5 Image Browser and Adobe Photoshop software. For live imaging, embryos were anaesthetised in MS222 for 5 min prior to imaging on a Zeiss Axioplan2 compound microscrope. In order to visualise xanthophores, embryos were reared in 0.001% methylene blue from the first day of development as described in (Le Guyader and Jesuthasan, 2002).

Neural crest cell counts

Melanised melanophores in the dorsal stripe, head and lateral stripe along the entire length of the embryo were counted at 4 dpf and 6 dpf. To contract melanophores, larvae were incubated in 10 mg/ml adrenalin (Sigma) for 10 minutes prior to fixing in 4% paraformaldehyde, as previously described (Mellgren and Johnson, 2004). Pax7+ and dct+ NC cells were counted after in situ hybridisation or immunodetection. Individual Pax7+ NC cells were clearly visible and categorised as cephalic or trunk based on a gap in Pax7+ NC in the hindbrain region. Dct+ cells were counted on one side of the trunk, including lateral stripe. Melanised melanophores in the ventral stripe and dct+ cells in dorsal or ventral stripes were not counted because individual cells could not be confidently distinguished. Iridophores in dorsal and ventral stripes were counted in anaesthetised live embryos at 6 dpf using darkfield microscopy. Quantifications are mean ± SEM. Student’s t-test analysis was performed with Graphpad Prism statistical software.

Sequence alignments and phylogenetic reconstructions

Amino acid sequences were obtained from BLAST searches against GenBank (protein database) or the current genome assemblies at Ensembl ( of human (NCBI 360), mouse (NCBI m36), chicken (WASHUC2), frog (JGI 4.1), stickleback (BROAD S1), medaka (HdrR), fugu (FUGU 4), Tetraodon (TETRAODON 7) and zebrafish (Zv7). Usually the zebrafish gene was used as query sequence. The following sequences were used: D. melanogaster (DM) gooseberry (Gsb): NP_523863; DM gooseberry-neuro (Gsb-n): NP_523862; H. sapiens (HS) PAX3 (isoform e): NP_852124; HS PAX7 (isoform b): DQ322591; M. musculus (MM) Pax3: NP_032807; MM Pax7: NP_035169; G. gallus (GG) Pax3: NP_989600; GG Pax7: NP_990396; X. tropicalis (XT) Pax3: CAJ82363; XT Pax7: NP_001088995; G. aculeatus (GA) Pax3: ENSGACG00000014017; GA Pax7: ENSGACG00000012890; GA Pax7 (2/2): ENSGACG00000001703; O. latipes (OL) Pax3: ENSORLG00000015932; OL Pax3b: ENSORLG00000009031; OL Pax7: ENSORLG00000004269; T. rubripes (TR) Pax3: SINFRUG00000132501; TR Pax3b: SINFRUG00000161042; TR Pax7: SINFRUG00000129602; T. nigroviridis (TN) Pax3: GSTENG00028437001; TN Pax3b: GSTENG00026741001; TN Pax7: GSTENG00027574001; D. rerio (DR) Pax3: AAC41253; DR Pax3b: ENSDARG00000028348; DR Pax7: AAC41255; DR Pax7b: ENSDARG00000070818. Amino acids sequences were aligned using ClustalX (v1.81) (Chenna et al., 2003) and analysed using GeneDoc (Nicholas et al., 1997). Intron/exon boundaries were predicted from ESTs and Ensembl predictions. Only sequences with equivalent exon structure were aligned. Bayesian phylogenies were inferred using the Monte Carlo Markov chain method (four chains for 10,000 generations) as implemented by the program MrBayes (v3.1.2) (Ronquist and Huelsenbeck, 2003). Neighbour-joining (Swofford, 2000) and maximum-likelihood (Schmidt et al., 2002) trees gave identical topology. Treeview (v1.6.6) ( was used to visualise trees.


Pax3 is expressed in neural crest precursors

Pax3 mRNA is first detected in presumptive NC at the end of gastrulation and is the earliest known marker of zebrafish NC (Hammond et al., 2007; Lewis et al., 2004). Pax3 expression is maintained during early somitogenesis (Fig. 1A) and at the 7 somite stage (7s), pax3 mRNA is expressed in cranial NC, dorsolateral neural plate and in bilateral stripes within the developing spinal cord (Fig. 1B; Hammond et al., 2007; Lewis et al., 2004; Seo et al., 1998). By 7s, the pax3b gene has also initiated expression in similar cranial regions, but is not detected as early as pax3 or in neural precursors in the posterior trunk (Fig. 1C and data not shown). Similarly, pax7 mRNA is only detected in the midbrain (Fig. 1D). An antibody against the Pax3/7 protein family reacts with nuclei in all regions where pax3/7 mRNAs are detected, including the posterior trunk NC precursors, which only express pax3 mRNA (Fig. 1E). Pax3/7 immunoreactivity in trunk NC precursors reflects Pax3 protein, because immunoreactivity is eliminated in embryos injected with either of two non-overlapping antisense morpholino oligonucleotides (MOs) specific to pax3 (Fig. 1E,F; pax3 MO, 88%, n = 34; pax3 MO2, 87%, n = 30). In parallel, Western analysis reveals a reduction in a 55 kd band at the right size for Pax3 protein (Fig. 1G). Tissue immunoreactivity is reduced but still detected in cranial NC and brain regions expressing pax3b and pax7 in addition to pax3, showing specificity of the pax3 MOs (Fig. 1F). We next confirmed that injection of a BAC containing the pax3 gene results in strong mosaic accumulation of both pax3 mRNA and Pax3/7 immunoreactivity in regions that normally express pax3, including the future NC (Fig. 1H-K). That Pax3 protein is in NC precursors is confirmed by co-localisation with foxd3 mRNA (Fig. 1L-Q; Lewis et al., 2004). Thus, Pax3 is present in the nuclei of NC precursors in cranial regions and within the neural plate.

Figure 1
Pax3/7 expression in neural crest

Throughout subsequent NC development, pax3 RNA is most abundantly expressed in the dorsal portion of the CNS along the entire rostrocaudal axis, although it is also weakly detectable in the somites (Fig. 1U,V,X; Hammond et al., 2007; Seo et al., 1998). Pax3b mRNA is weakly detected in a more restricted region of dorsal neural tube and also on the lateral somite surface, possibly in dermomyotome (Fig. 1Z-E’). Although pax3 and pax3b are expressed in the location of NC precursors and pre-migratory NC all along the dorsal neural tube, neither mRNA shows the punctuate pattern characteristic of NC markers on NC migration pathways, suggesting pax3 and pax3b are down-regulated as trunk NC migrates (Fig. 1U,V,X,D’,E’). Pax3 mRNA does, however, persist in clusters of cells at the mid-hindbrain boundary, possibly corresponding to placode-derived cells of the ophthalmic lobe of the trigeminal ganglion (Fig. 1R-T; Stark et al., 1997). Thus, the data suggest that pax3 in NC functions at an early stage and is down-regulated as migration commences.

Pax3 is required for xanthophore formation

To analyse Pax3 function, we examined development of embryos injected with either pax3 MO or MO2 (hereafter referred to as pax3 morphants). At 5 days, pax3 morphants have a marked reduction in the trunk of yellow pigment normally produced in xanthophores (Fig. 2A-D; pax3 MO, 92%, n = 50; pax3 MO2; 90%, n = 42). Some xanthophore pigmentation remains in the head, but is also reduced (Fig. 2B). In wild type embryos, methylene blue specifically labels the pterinosome pigment organelles of terminally differentiated xanthophores, revealing their stellate shape (Fig. 2E; Le Guyader and Jesuthasan, 2002). In pax3 morphants, staining of xanthophores by methylene blue reveals a dose-dependent severe reduction in the trunk and also depletion in the head (Fig. 2F,H,K; pax3 MO, 91%, n = 79; pax3 MO2, 87%, n = 184). In contrast to xanthophores, black pigment-containing melanophores are present and correctly positioned (Fig. 2A-J) and iridophore formation appears grossly normal (Fig. 2I,J; controls 11.3 ± 5.8, n = 15; pax3 MO-injected 9.7 ± 5.3, n = 7). Therefore, Pax3 is required for xanthophore terminal differentiation to produce pterinosome organelles and yellow pigment.

Fig. 2
Pax3 knockdown blocks terminal differentiation of trunk xanthophores whilst other pigment cell types differentiate

Pax3 morphants have a defect in xanthophore fate specification

To analyse at which stage of xanthophore development Pax3 is required, expression of xanthophore markers was examined in pax3 morphants. Loss of four markers shows that xanthoblasts are absent (Fig. 3). Xanthine dehydrogenase (xdh) catalyses the synthesis of the pteridine xanthopterin and is expressed during differentiation of xanthophores (Parichy et al., 2000). At 24 hpf, xdh is expressed in individual NC cells in cranial regions and both pre-migratory and laterally migrating NC cells in the trunk (Fig. 3A,C). By 48 hpf, xdh expression covers the trunk of wild type zebrafish larvae (Parichy et al., 2000 and data not shown). In pax3 morphants, expression of xdh is essentially absent from the trunk at both 24 and 48 hpf and shows a reduction in the head that parallels the location and extent of loss of xanthophores (Fig. 3A-D and data not shown; pax3 MO, 94%, n = 70; pax3 MO2, 87%, n = 45). The zebrafish csf1r receptor tyrosine kinase is expressed from 18s in pre-migratory and migrating xanthoblasts and is required for xanthoblast migration (Parichy et al., 2000). In pax3 morphants, csf1r expression is missing even from pre-migratory cells, suggesting a failure of xanthophore specification rather than simply migration failure (Fig. 3F,H; pax3 MO, 90%, n = 32; pax3 MO2, 88%, n = 40). The transcription factor gene microphthalmia (mitfa) is expressed in both melanoblasts and xanthoblasts, is absolutely required for melanoblast formation and its loss causes a reduction in xanthophores (Lister et al., 1999). Expression of mitfa in pax3 morphants is reduced but, consistent with the presence of melanocytes, some mitfa+ NC cells are present at 24 hpf (Fig. 3I,J; pax3 MO, 71%, n = 14; pax3 MO2, 88%, n = 68). Notably, mitfa is missing from migratory cells at this stage, suggesting that there is a delay in melanophore migration or differentiation in pax3 morphants. GTP cyclohydrolase I (gch), is also expressed in chromatophore precursors (Pelletier et al., 2001). In pax3 morphants, expression of gch is ablated in trunk regions and greatly reduced, but not completely absent, from the head at 24 hpf (Fig. 3K-N; pax3 MO, 93%, n = 29). We conclude that knockdown of Pax3 blocks differentiation of xanthophore precursors at an early stage, prior to their migration, suggesting a role for Pax3 in xanthophore fate specification.

Fig. 3
Pax3 morphants have a defect at early stages of xanthophore development

At 25 hpf, when NC cells in anterior trunk regions are beginning their migration on the lateral pathway, strong nuclear Pax3/7 immunoreactivity is present in pre-migratory and migrating NC, whereas weaker Pax3/7 staining is present in dorsal neural tube and dermomyotome nuclei (Fig. 3O; Hammond et al., 2007). Pax3 morphants show a severe reduction of strong Pax3/7 immunoreactivity from pre-migratory and migrating NC, with residual expression in the somites and neural tube, presumably attributable to other Pax3/7 proteins (Fig. 3P; pax3 MO, 87%, n = 75; pax3 MO2, 93%, n = 42). In un-manipulated embryos, abundant pax7 mRNA and protein are expressed in the location of pre-migratory NC cells and strongly Pax7-labelled NC cells spread gradually over the lateral somite surface in parallel with NC migration on the lateral pathway (Fig. 3Q,S,U,W; Hammond et al., 2007). Furthermore, Pax7 is located in all strongly Pax3/7-reactive NC cells, suggesting that Pax3/7 immunoreactivity in migrating NC arises from Pax7 (data not shown). Pax3 morphants essentially lack NC cells containing pax7 mRNA (pax3 MO, 78%, n = 41) or Pax7 protein (pax3 MO, 87%, n = 54; pax3 MO2, 89%, n = 66) in the trunk at 25-27 hpf, whereas Pax7 immunoreactivity is still present in a reduced number of strongly-labelled NC cells in the head, and weakly in the dermomyotome (Fig. 3P,R,T,V,X). Counts of strongly labelled Pax7+ NC cells confirm their stronger depletion in trunk than in head, even in embryos in which the MO was only partially effective (Fig. 3Y; n = 18 controls, n = 19 pax3 MO injected embryos). The reduction of Pax3/7 protein revealed by Western analysis is consistent with a requirement of Pax3 for the majority of Pax7 expression (Fig. 1G). Therefore, loss of Pax7 parallels failure of xanthophore fate specification.

Pax7 is expressed in xanthophore precursors

The correlation of loss of Pax7 from NC cells with the absence of xanthophores and xanthophore markers in pax3 morphants suggests that, within NC, Pax7 is specifically expressed in the xanthophore lineage. Until 15s, pax7 is not expressed in dorsal neural regions of the trunk (Hammond et al., 2007). Thereafter, pax7 is expressed in both delaminated pre-migratory and migrating NC of the trunk and head (Figs 3Q,S and and4).4). Pax7 mRNA and protein first appear in NC at the mid-hindbrain boundary around 18s (Fig. 4A-D). By 25 hpf, individual cells outside the neural tube contain pax7 mRNA, and expression within the neural tube is absent from the most dorsal regions (Fig. 4E-G). The NC identity of migrating Pax7+ cells is confirmed by co-localisation of Pax7 with EGFP in the sox10:egfp transgenic line, which labels both pre-migratory and migrating NC (Fig. 4M-R; Carney et al., 2006). In trunk at 25 hpf, pre-migratory NC cells and isolated cells on the lateral pathway contain abundant pax7 mRNA, but this is not detected in NC cells following the medial migration pathway, which have already reached the notochord at this stage (Fig. 4I-L). The restricted expression of pax7 to lateral pathway NC cells, suggests these cells are distinct from melanophores, which also migrate medially.

Fig. 4
Pax7 is a marker of the xanthophore lineage

Only pigment cell precursors are thought to follow the lateral pathway in the trunk (Raible et al., 1992). We investigated, therefore, whether markers of distinct NC pigment lineages are co-expressed with pax7 (Fig. 4S-Y). Xdh is specifically expressed in xanthophore precursors (Parichy et al., 2000). Pax7 mRNA is co-localised with xdh in NC, showing that pax7+ cells are of the xanthophore lineage (Fig. 4S-U’). Furthermore, xdh+ and pax7+ NC cells are present in similar numbers as they migrate on the lateral pathway in the trunk, suggesting that all pax7+ NC are xanthoblasts (compare Fig. 3C,S and data not shown). In contrast, the melanophore lineage marker dopachrome tautomerase (dct; Kelsh et al., 2000b) does not co-localise with pax7, demonstrating that melanoblasts are not pax7+ (Fig. 4V-Y). In addition, among 112 melanised melanophores scored at 28 hpf none were found to contain Pax7 protein (data not shown). Taken together, these data show that pax7 marks xanthoblasts migrating on the lateral pathway.

Xanthophores are specified before Pax7 protein accumulates

To determine the sequence of events during xanthophore development, temporal expression of xanthophore markers was examined (Fig. 5). NC formation progresses from anterior to posterior, such that the most posterior NC are developmentally the ‘youngest’ (Eisen and Weston, 1993). At 20-25s, Pax7+ NC are located in the head and in pre-migratory positions in the trunk as far posteriorly as somites 4-9 (Fig. 5A,B). At the same stage, xdh is expressed as far as somite 18, where Pax7 is not detected (Fig. 5A,C). Thus, the developmentally youngest xanthoblasts express xdh without Pax7 (Fig. 5C). Where Pax7+ NC is detected, in more mature xanthoblasts, it is co-expressed with xdh (Fig. 5C). Expression of mitfa is evident posteriorly as far as somite 12, indicating that mitfa mRNA also accumulates before Pax7 (Fig. 5A,D). Spatiotemporal expression of csf1r and Pax7 appear identical (Fig. 5E). Pax7+ NC cells at all anteroposterior positions co-express csf1r mRNA, which extends as far posterior as somite 9 (Fig. 5A,E). In conclusion, xdh is expressed in the developmentally youngest identifiable cells of the xanthophore lineage, whereas Pax7 and csf1r mRNAs accumulate as xanthophore differentiation advances.

Fig. 5
Xanthophore specification occurs before accumulation of Pax7

Pax7 morphants have reduced xanthophore pigmentation

The appearance of Pax7 after other xanthophore markers raised the possibility that Pax7 functions at a distinct, and later, stage of xanthophore development than Pax3. To test this hypothesis we designed a morpholino against pax7. Pax7 morphants had substantial knockdown of Pax7 protein (Fig. 6A,B; pax7 MO, 4 ng dose, 76%, n = 67). However, migrating NC containing reduced levels of Pax7 were apparent (Fig. 6B). At 6 dpf, pax7 morphants had a dose-dependent reduction of yellow pigmentation in both head and trunk (Fig. 6C-F,G; pax7 MO, 4 ng dose, 66%, n = 115). Although overall yellow pigmentation was reduced, low levels were visible at dorsal locations suggesting a defect, but not an ablation, of xanthophores (Fig. 6E,F). Indeed, staining with methylene blue revealed xanthophore cells were present in equivalent number in pax7 morphants and capable of migration (Fig. 6H,I; pax7 MO, 4 ng dose, 100%, n = 89). Interestingly, pax7 morphant xanthophores had a contracted pterinosome disposition compared to the spread morphology of control xanthophores (Fig. 6H,I). In pax7 morphants expression of the xanthoblast markers xdh (Fig. 6N,O; pax7 MO, 4 ng dose, 93%, n = 97) and csf1r (data not shown; pax7 MO, 4 ng dose, 91%, n = 82) was readily detectable. Melanophore and iridophore differentiation appeared unaffected in pax7 morphants (Fig. 6J-M). In conclusion, pax7 morphants have reduced yellow pigmentation; xanthophore specification proceeds normally, however, suggesting a later role for Pax7.

Fig. 6
Xanthophore pigmentation is reduced after knockdown of Pax7

The duplicate pax7b gene is expressed in xanthophores

Even with high doses of pax7 MO (up to 20 ng per embryo), it was not possible to completely ablate Pax7 immunoreactivity in NC cells (Fig. 6B and data not shown). One hypothesis for residual Pax7 immunoreactivity is the presence of a second pax7 gene in the zebrafish genome. BLAST searching identified a duplicate pax7 gene, pax7b, on chromosome 23 in Zv7. A corresponding full length EST has 510 amino acids with 94% identity to zebrafish Pax7 and exhibits highly conserved exon boundaries (Suppl. Fig. 1B). Phylogenetic analysis demonstrates that Pax7b groups with Pax7 rather than Pax3 genes (Suppl. Fig, 1A) and synteny exists with the mammalian Pax7 loci, despite some re-arrangements (data not shown). Duplicate pax7 genes are also apparent in a variety of fish genomes, suggesting they arose early in actinopterygian evolution (Suppl. Fig. 1A). Therefore, phylogenetic sequence analysis indicates that pax7b is a duplicated pax7.

To ascertain if a role in xanthophore development was possible, the expression pattern of pax7b was examined. Prior to 15s, pax7b is expressed in the brain and somites, but not in trunk NC precursors (Suppl. Fig. 1C-F). At 25 hpf, however, pax7b mRNA is detected in both pre-migratory and migrating cranial and trunk NC (Suppl. Fig. 1G-K). Whereas pre-migratory NC cells retain pax7 expression as they mature in anterior trunk, pax7b mRNA is absent (Fig. 4H-J; Suppl. Fig. 1I-K). The dynamics and number of pax7b+ NC cells suggest that pax7b is also expressed in xanthoblasts. Co-localisation of xdh and pax7 mRNA demonstrates pax7b is expressed in the xanthophore lineage (Suppl. Fig. 1L-Q). Furthermore, pax7b continues to be expressed in NC after injection of pax7 MO (Fig. 6P,Q). In conclusion, pax7b is expressed in the xanthophore lineage, which may account for the residual Pax7 immunoreactivity and mild yellow pigmentation defect in pax7 morphants.

Increase in melanophores correlates with reduction in xanthophores after Pax3 knockdown

Interactions between different chromatophore lineages influence pigment pattern formation (Maderspacher and Nusslein-Volhard, 2003; Parichy and Turner, 2003). As pax3 morphants have a reduction in xanthophores, we examined the consequences for other chromatophore lineages. Melanoblasts migrate both medially and laterally to become distributed in four longitudinal stripes situated at the dorsal and ventral extremes of the myotome, the horizontal myoseptum and on the yolk by 48 hpf (Fig. 6E; Kelsh et al., 2000b). At 25 hpf, pax3 morphants have fewer dct+ melanoblasts than controls (Fig. 7A,B; pax3 MO, 86% of embryos affected, n = 56; in 14 control and 14 affected morphant embryos, dct+ cells were reduced by 49%, p = 0.0002). Mitfa mRNA is also reduced (Fig. 3I,J). Interestingly, despite the overall reduction in mitfa and dct mRNAs, more dct+ melanoblasts are located in a pre-migratory position in pax3 morphants, suggestive of a delay in melanophore development, although small numbers migrate successfully (Fig. 7A,B). By 35 hpf, dct (Fig. 7C,D; pax3 MO2, 92%, n = 142) and mitfa (data not shown; pax3 MO, 89%, n = 141) were relatively more abundant dorsally in pax3 morphants. However, both dct and mitfa mRNAs are still reduced in more ventral trunk regions (Fig. 7C,D and data not shown), suggesting a delay in melanophore migration. By 48 hpf, pax3 morphant dct expression appears more widespread than in controls, with many migrating dct+ cells still present between dorsal and ventral stripes (Fig. 7E,F insets; 16.7 ± 1.4 in controls, n = 10; and 29.1 ± 1.3 in pax3 MO injected embryos, n = 10; p<0.0001). Congruently, in 48 hpf pax3 morphants, melanin-expressing cells form aggregations in the dorsal stripe and are positioned in the trunk between the lateral and ventral stripes at the centre of each somite, consistent with dct expression (Fig. 7G,H; pax3 MO, 83%, n = 145). By 5 dpf, melanophore pattern appears to have resolved in pax3 morphants (Fig. 7I,J; pax3 MO, 69% with wild-type patterns, n = 102). In conclusion, pax3 morphants show an early reduction in melanophore markers and a delay in migration, which subsequently resolves.

Fig. 7
Delayed melanophore origin then increase correlate with xanthophore loss in pax3 morphants

To determine if the altered melanophore migration reflects a change in melanophore number, melanophores were analysed after migration is complete (Fig. 7K-O). The number of melanophores in the trunk dorsal stripe was significantly increased in pax3 morphants at 4 dpf and 6 dpf (Fig. 7K,L,O and data not shown). Also, dorsal stripe melanophores in pax3 morphants appeared smaller than in controls (Fig. 7K,L). Interestingly, no melanophore increase was detected in the head of pax3 morphants (Fig. 7M-O). To investigate the correlation between xanthophore loss and melanophore increase further, pax3 morphants were categorized according to severity of the xanthophore phenotype and melanophores quantified (Fig. 7P). Pax3 morphants with a more severe xanthophore phenotype had a larger increase in melanophore number (Fig. 7P). Furthermore, pax7 morphants, that retain xanthophores, have no change in melanophore number (Fig. 7O). Overall, a reduction in xanthophores correlates with an increase in melanophores after knockdown of Pax3.

Abnormal neural crest maintenance after Pax3 knockdown

As multiple chromatophore lineages are disrupted in pax3 morphants, we investigated whether foxd3 or sox10, more general markers of NC lineages, were perturbed (Dutton et al., 2001; Lewis et al., 2004; Odenthal and Nusslein-Volhard, 1998). Injection of pax3 MO leads to loss of Pax3/7 immunoreactivity in the bilateral stripes of the lateral neural plate and cranial NC, but fails to diminish either foxd3 (pax3 MO, 100%, n = 57) or sox10 (pax3 MO, 100%, n = 56) expression prior to 18s (Fig. 8A-F; Suppl. Fig. 2A-D). Thus, normal levels of Pax3 are not required for initiation of early marker expression within the NC.

Fig. 8
Pax3 is required for enteric nervous system development

At later stages of NC development, specific defects appear in NC of pax3 morphants. At 25 hpf, sox10 expression in pre-migratory and migrating trunk NC is severely disrupted in pax3 morphants. Delaminated cells expressing sox10 are detected adjacent to dorsal neural tube and migrating on medial and lateral pathways in controls, but few sox10-expressing cells are observed migrating on the medial pathway in pax3 morphants (Fig. 8I,J; pax3 MO, 81%, n = 44). Transient foxd3 expression in nascent NC remains unaltered (Fig. 8G,H; pax3 MO, 89%, n = 27). In summary, Pax3 is required for maintenance of sox10 expression during later stages of NC development.

Pax3 is required for enteric nervous system formation

The disruption of sox10 expression in pax3 morphants led us to investigate other NC derivatives, as sox10 mutant fish have disruption of numerous NC lineages (Dutton et al., 2001; Kelsh and Eisen, 2000). Among the defects reported in sox10 mutant zebrafish is mis-positioning of NC-derived dorsal root ganglion (DRG) sensory neurons (Kelsh and Eisen, 2000). Although, DRG neurons are present in equivalent numbers in pax3 morphants, they are often mis-positioned in dorsal regions (Suppl. Fig. 2E-K). No defects in several other cell populations arising from pax3-expressing dorsal neural tissue were observed: trigeminal ganglia and mechanosensory Rohon-Beard neurons were unaffected in pax3 morphants (Suppl. Fig. 2L-Z’; pax3 MO, 100%, n = 54). However, analysis of pax3 morphants at later stages revealed that sox10 mRNA is reduced at 36 hpf in enteric neuronal precursors in the vagal region of the developing gut, whereas otic placode and glial sox10 expression remains (Fig. 8K,L; pax3 MO, 86%, n = 35). Subsequently, terminally differentiated enteric neurons do not develop in the gut of pax3 morphants (Fig. 8M-P; pax3 MO, 74%, n = 27). Examination of other NC-derived cell populations failed to reveal defects. Sympathetic neurons are readily detected with HuC, above the ventral melanophore stripe and tyrosine hydroxylase (data not shown) and dopamine ß hydroxylase sympathetic neuronal markers were unaltered in pax3 morphants. (Suppl. Fig. 3B,C). Continued foxd3 (Fig. 8G,H; pax3 MO, 100%, n = 57) and sox10 expression in the lateral line of pax3 morphants suggested that glial formation occurs (Fig. 8K,L; pax3 MO, 100%, n = 57) and the cranial ectomesenchymal NC marker dlx2a is also unaffected in pax3 morphants (Suppl. Fig. 3A; pax3 MO, 100%, n = 95). Thus, Pax3 is required for sox10 expression in enteric neuronal precursors in the developing gut and subsequent enteric nervous system development.


In zebrafish, Pax3 is required for differentiation of a subset of NC cells: xanthophores and enteric neurons. Pax3 is expressed in NC precursors, but is not detected in migrating NC cells. Pax7 and Pax7b are expressed after other xanthoblast markers in maturing/migrating xanthophore lineage cells. Expression of all xanthoblast markers, including Pax7, is dependent on Pax3 function. We hypothesise that Pax3 acts at a precursor stage in NC development essential for proper specification of xanthophore and enteric neural precursor cells. Loss of xanthophores after Pax3 knockdown leads to an increase in melanophores, suggesting Pax3 may function within a multipotent chromatoblast precursor.

Pax3 is a xanthophore specification factor

Pax3 is required for fate specification of xanthophores, as compared with other pigment cells. This is evident from the loss of all known xanthophore markers, including Pax7 (see below), reflecting the absence of the xanthophore cell population in pax3 morphants. Although many mutants with reduced xanthophore pigmentation exist, little is known about the genetic hierarchies that control xanthophore specification. Zebrafish sox10 mutants have reduced non-ectomesenchymal NC, including xanthophores (Dutton et al., 2001; Kelsh and Eisen, 2000). However, xanthophores initially differentiate, but then die in sox10 null mutants, indicating that xanthophores are partially specified (Dutton et al., 2001). PAX3 and SOX10 can cooperate to drive downstream NC gene expression in some situations (Bondurand et al., 2000; Potterf et al., 2000). It seems likely that Pax3 and Sox10 also cooperate in zebrafish chromatophore development.

Severe csf1r mutant alleles lack most trunk xanthophores, but a subset of head xanthophores form, just as in pax3 morphants (Maderspacher and Nusslein-Volhard, 2003; Odenthal et al., 1996). Detailed analysis of other csf1r mutant alleles demonstrated that xanthophores are specified, but fail to migrate, leading to reduced yellow pigmentation (Parichy et al., 2000). Our data demonstrates that Pax3 functions upstream of Csf1r, because it is not expressed in pax3 morphants. Therefore, this study establishes Pax3 as the first identified xanthophore specification factor.

Residual xanthoblasts are present, contain Pax7 protein and lead to xanthophore pigmentation in head regions of pax3 morphants. We cannot exclude the possibility that incomplete knockdown of Pax3 protein or the expression of pax3b accounts for xanthophore formation in this location. However, csf1r appears to be more completely down-regulated than pax7 or xdh in head NC cells of pax3 morphants. As two distinct populations of xanthophores have been described in adult zebrafish (Mellgren and Johnson, 2006), our data raise the possibility that the residual xanthophores of the head constitute a distinct population, ones that require little or no pax3 and csf1r activity.

Pax7 as a xanthophore lineage marker

Our data show that pax7 is expressed in xanthoblasts, marked by xdh, and not in melanoblasts, which accumulate dct mRNA and melanin. However, it is not clear which, if any, NC cell types in addition to xanthophores express pax7. A recent report has suggested that Pax7 protein is present in fully differentiated melanophores and, based on Pax7+ nuclei located on the yolk extension and in close proximity to the eye, that Pax7 is also present in iridophore precursors (Lacosta et al., 2007). We have not observed melanised cells containing Pax7 and we see xanthophores in the proposed iridophore yolk and eye locations by methylene blue staining (data not shown). Furthermore, we have not noted NC phenotypes in cells other than xanthophores in pax7 morphants. Our study unequivocally shows that Pax7 is in the xanthophore lineage. However, in our view, whether Pax7 accumulates or functions in other NC cell types or at other times in development requires further critical study.

We find that gch mRNA is primarily expressed in xanthophores in 24 hpf trunk, having a similar distribution to xdh and pax7 mRNAs, markers of the xanthophore lineage. At later stages, gch expression is observed in melanophores, as well as xanthoblasts (Parichy et al., 2000; Pelletier et al., 2001; Ziegler, 2003). However, the melanophore markers mitfa and dct are expressed more ventrally than gch and Pax7 in migrating trunk NC cells at around 24 hpf. Moreover, pax3 morphants essentially lack gch, xdh and pax7 expression in the trunk, whereas dct and mitfa are still expressed dorsally. Although we can not eliminate the possibility that delayed maturation of dct-expressing melanoblasts (see below) contributes to gch mRNA absence, our findings suggest that gch primarily marks the xanthophore lineage early, before accumulating in melanoblasts.

Pax7 accumulates in pre-migratory xanthophores at 20s, shortly after gch, xdh and mitfa mRNAs and in parallel with csf1r mRNA. Mitfa expression is also detected ventrolateral to the first Pax7/gch/xdh/csf1r-expressing cells in anterior trunk NC at this stage, suggesting that melanoblasts also express mitfa and commence migration slightly before xanthoblasts.

Pax7 and xanthophore development

We show that xdh, and possibly gch and mitfa, are expressed before Pax7 in the xanthophore lineage, thereby demonstrating that xanthophores are specified before Pax7 is present. Thus, Pax7 has a fundamentally different role to Pax3 within the xanthophore lineage, and acts later. Indeed, MO-driven reduction in Pax7 correlates with late differentiation defects in xanthophores; involving reduced pigmentation and altered pterinosome disposition. The failure to disperse pterinosomes correctly is likely to contribute to the reduced yellow colouration, as previously described for other xanthophore mutants (Odenthal et al., 1996). Xanthoblasts are present and able to migrate after reduction of Pax7, consistent with retention of xdh and csf1r expression in pax7 morphants.

Our pax7 morphants retain low levels of Pax7 immunoreactivity in migrating NC cells. It is probable that the duplicate zebrafish pax7b gene, which is also expressed in laterally migrating NC and is unaffected in pax7 morphants, functions in the xanthophore lineage. As Pax7 is 94% identical to Pax7b, the Pax7 antibody may detect both Pax7 and Pax7b (Minchin and Hughes, unpublished observation; Kawakami et al., 1997). Indeed, Pax7 immunoreactivity is retained in NC cells even after injection of high doses of pax7 MO that ablate all detectable signal from somite cells, which only express pax7. Support for this view derives from the observations that pax7 mRNA declines in wild type xanthoblasts at 48 hpf, yet Pax7 immunoreactivity persists (Lacosta et al., 2007 and our unpublished observations). All teleost fish genomes analysed had evidence of two pax7 genes, suggesting that pax7 was duplicated along with the teleost genome and the duplicate genes retained. The existence of additional partial duplicate sequences in O. latipes, T. nigroviridis and T. rubripes confirm that two clades of teleost pax7 genes exist (unpublished observation). The presence of Pax7b may account for the mild xanthophore defect in pax7 morphants. Until null mutations provide a definitive analysis, our tentative conclusion is that Pax7 is dispensable for xanthophore fate specification, but necessary for normal xanthophore terminal differentiation.

Pax3 and melanophore development

Pax3 has been shown to have an evolutionarily conserved role in melanocyte development from ascidians to mammals (Lang et al., 2005; Tremblay and Gruss, 1994; Yajima et al., 2003). Our data indicate that, in zebrafish, Pax3 also has a role in melanophore development. Initially, pax3 morphants exhibit a reduced number of migrating melanoblasts, that express dct, during early melanogenesis. Although this early reduction in melanoblasts could reflect loss of a sub-population of embryonic melanophores, a more likely possibility is that it reflects a delay in melanophore development. More dct and mitfa mRNA is detected in un-migrated NC cells in pax3 morphants. By 48 hpf, melanophore migration is recovering in pax3 morphant zebrafish and melanophores appear more abundant. Subsequently, melanophore numbers are increased, particularly in the dorsal stripe.

The early reduction in melanoblast maturation and migration in zebrafish lacking Pax3 is reminiscent of the Splotch (Sp; Pax3 mutant) mouse, in which reduced numbers of melanoblasts form and migrate, despite deficient melanoblast expansion (Hornyak et al., 2001). In mice, it is unclear if melanocytes would be permanently reduced, as the Sp/Sp null phenotype is lethal at E14 (Tremblay and Gruss, 1994). However, heterozygous Sp/+ mutants have melanocyte migration defects and humans with Waardenburg syndrome type I, a dominant PAX3 loss of function mutation, also have pigmentation alterations. In neither case is it clear if melanocyte number is reduced or if melanocyte distribution is simply altered (Tassabehji et al., 1992; reviewed in Chi and Epstein, 2002). These parallels suggest that Pax3 is required for timely melanoblast development in mouse and zebrafish.

The migration failure of melanophores prior to 48 hpf could reflect either a cell autonomous requirement for Pax3 in early NC cells, or an environmental effect on NC migration. Pax3 is expressed in the developing somitic dermomyotome in both zebrafish and mice (Goulding et al., 1991; Hammond et al., 2007). The dermomyotome serves a substratum over which laterally-migrating trunk NC moves, and a recent study has highlighted the importance of this somitic region in melanophore patterning mediated by sdf1a (Svetic et al., 2007). There remains considerable debate on the cell autonomy in Pax3 NC phenotypes (Chan et al., 2004; Li et al., 1999; Mansouri et al., 2001). The ease of cell transplantation in zebrafish should permit the precise role of Pax3 in NC lineages to be determined.

Pax3 and a chromatoblast stem cell

At least two models could account for the correlated loss of xanthophores and increase in melanophores in pax3 morphants: xanthophore-melanophore interaction or a fate switch in a common chromatophore precursor. Although xanthophores interact with melanophores in adult patterning, we do not favour this model because, in contrast to the pax3 morphant phenotype, xanthophore loss causes a reduction of melanophores (Maderspacher and Nusslein-Volhard, 2003; Parichy et al., 2000; Parichy and Turner, 2003). As NC lineages are specified through progressive fate restriction (Le Douarin et al., 2004), we prefer a cell autonomous model in which Pax3 controls xanthophore fate choice within a common chromatoblast precursor.

A common chromatophore precursor or stem cell has been postulated (Bagnara et al., 1979). A dorsal location for such stem cells is suggested by better melanophore regeneration in dorsal regions (Yang and Johnson, 2006). The co-expression of melanoblast and xanthoblast markers such as mitfa with csf1r, xdh and, as we show, Pax7 in a sub-population of dorsal chromatophore precursors is consistent with this idea (Parichy et al., 2000; Parichy and Turner, 2003). As pax3 is expressed in NC precursors, we hypothesise that Pax3 drives xanthophore choice in a common chromatoblast stem cell. Reduction of Pax3, and loss of the xanthophore lineage, results in defective fate choice within this precursor leading to more abundant melanophores.

Zebrafish have three chromatophore lineages (reviewed in Lister, 2002) and these seem to form a developmental hierarchy: xanthophore, melanophore and iridophore. Although melanophores undergo terminal differentiation before xanthophores, pax3, a xanthophore specification gene, is expressed prior to mitfa, the earliest known melanophore specification factor. Iridophores differentiate much later and have no molecularly identified specification genes. If Pax3 is functioning within a multipotent chromatoblast precursor to specific xanthophores, why do pax3 morphants show an increase in melanophores but not iridophores? Mitfa mutants fail to specify melanophores and have an increase in iridophores together with a partial loss of xanthophores (Lister et al., 1999). As some xanthoblasts express mitfa, we speculate that Mitfa inhibits iridophore fate, whereas Pax3 inhibits melanophore fate, even in the presence of Mitfa. Interestingly, a similar hierarchy is apparent in adult chromatophore patterning, in which xanthophores regulate melanophore pattern and melanophores are required for proper iridophore disposition (Johnson et al., 1995; Parichy et al., 2000; Parichy and Turner, 2003).

A mechanism by which Pax3 could function to repress melanophore fate in a chromatoblast stem cell is suggested by studies showing that murine Pax3 represses melanocyte terminal differentiation by competition with Mitf on a Dct enhancer (Lang et al., 2005). A similar action of zebrafish Pax3 might contribute to melanophore increase in pax3 morphants by allowing Mitfa to drive dct (and possibly other target gene) expression in more cells than normal, thereby forcing a common precursor away from a xanthoblast fate and towards a melanoblast. It should be noted that such a view would not exclude Pax3 having other functions within NC, for example promoting rapid melanophore differentiation.

Pax3 is required for enteric nervous system development in zebrafish

Pax3 morphants lack enteric neurons, similar to the situation in Splotch mice (Lang et al., 2000). In the mouse, Pax3 acts with Sox10 to enhance c-Ret expression (Lang and Epstein, 2003). Either SOX10 or c-RET mutation can cause Hirschsprung disease, but no PAX3 involvement has yet been described (OMIM #142623). Interestingly, although sox10 expression in pre-migratory and early migrating NC cells is normal in pax3 morphants, sox10 mRNA is diminished in later migrating cells. Lack of Sox10 would account for failure of enteric neurogenesis (Elworthy et al., 2005). Alternatively, lack of specification of enteric neuronal precursors prior to ventral migration could account for the reduction of sox10 mRNA at later stages. Indeed, DRG and sympathetic neuronal precursors migrate from the dorsal region early (Raible and Eisen, 1994; Raible and Eisen, 1996), and these cell types appear normal in pax3 morphants. Perhaps Pax3 acts to select enteric neuronal precursors from a sox10+ population in an analogous way to its selection of xanthoblasts. In any case, our findings promote zebrafish as a useful model for investigating the potential role of Pax3 in Hirschsprung disease.

Evolutionary change in roles of Pax3/7 gene family

The single pax3/7 genes in Amphioxus and tunicates are expressed at the dorsal neurectodermal border, suggesting an ancient role in NC patterning (Holland et al., 1999; Sauka-Spengler et al., 2007; Wada et al., 1997). Our data indicate that although it is expressed early, Pax3 is less important in NC specification in zebrafish than in other vertebrates (Monsoro-Burq et al., 2005; Sato et al., 2005; Sauka-Spengler et al., 2007; Tremblay and Gruss, 1994). Meulemans and Bronner-Fraser (2004) have suggested that Pax3/7 genes have a conserved function high in the hierarchy of NC specification. In chicken, evolutionary change appears to have allowed Pax7 to fulfil the major NC specification role (Basch et al., 2006). As zebrafish pax7 and pax7b expression in NC appears late and in xanthoblasts, and Pax7+ NC is greatly reduced in pax3 morphants, pax7 genes are unlikely to rescue other NC lineages when Pax3 is absent. Another possible reason for the presence of most NC lineages in pax3 morphants is the duplication in teleosts of the pax3 gene (Lewis, 2005). Phylogenetic analysis of Pax3 proteins showed that a broad range of teleost fish have two pax3 genes, forming the pax3 and pax3b sub-groups. Pax3b is expressed in similar regions to pax3, but expression initiates at later stages and lower levels in trunk neurectoderm. Our use of an antibody that recognises Pax3/7 proteins across the metazoa (Davis et al., 2005), shows that pax3 morphants lack Pax3/7 immunoreactivity in NC at early stages. This suggests that other Pax3/7 family members are unlikely to support expression of early NC markers such as sox10 and foxd3 in zebrafish. Nevertheless, the early expression of pax3 in mice, Xenopus and teleost fish and their requirement for at least some NC lineages suggests that pax3 was an ancestral NC cell fate specification gene following the whole genome duplications that accompanied vertebrate evolution.

What is the significance of the apparent differences in the role of Pax3 in NC between vertebrate species? Mice lacking Pax3 have drastic reductions in melanocytes prior to their death, whereas pax3 morphant fish, although they have reductions in early melanoblast markers, eventually have more melanophores. It seems Pax3 is more essential for driving melanophore differentiation in mammals, perhaps reflecting a more prominent role in promoting early melanoblast differentiation. But an alternative possibility relates to the loss of xanthophores in amniotes. Xanthophores are found in both fish and amphibians; therefore parsimony suggests that amniote ancestors possessed xanthophores (reviewed in Kelsh, 2004; Silver et al., 2006). It is possible that Pax3-requiring xanthoblasts did not disappear in amniote evolution, but evolved into a ‘new’ kind of melanoblast; one that required Pax3 function. If this were the case, mice lacking Pax3 would lose ‘new’ but retain ‘ancestral’ melanocytes, leading to the observed phenotype of melanocyte reduction (Hornyak et al., 2001).

Teleost pigment patterns have roles in sexual selection and predator avoidance, and are therefore strongly selected for in evolution; close relatives of zebrafish exhibit a spectacularly diverse array of pigment patterns (reviewed in Parichy, 2006). The xanthophore lineage has been shown to have an important role governing adult melanophore patterning within closely related Danio species (Parichy et al., 2000; Parichy and Turner, 2003). Therefore, change in regulation or action of pax3 is a strong molecular candidate for adaptation during the evolution of pigment pattern in the xanthophore lineage and beyond.

Supplementary Material

Supplementary Material


Thanks to Dr. Chrissy Hammond for the Western blot, and to Drs R. Kelsh, D. Raible, J. Lewis, D. Parichy and N. Patel for reagents. SMH is a member of MRC External Scientific Staff with Programme Grant support and JENM had an MRC PhD studentship.


  • Bagnara JT, Matsumoto J, Ferris W, Frost SK, Turner WA, Jr., Tchen TT, Taylor JD. Common origin of pigment cells. Science. 1979;203:410–415. [PubMed]
  • Basch ML, Bronner-Fraser M, Garcia-Castro MI. Specification of the neural crest occurs during gastrulation and requires pax7. Nature. 2006;441:218–222. [PubMed]
  • Blader P, Plessy C, Strahle U. Multiple regulatory elements with spatially and temporally distinct activities control neurogenin1 expression in primary neurons of the zebrafish embryo. Mech Dev. 2003;120:211–218. [PubMed]
  • Blagden CS, Currie PD, Ingham PW, Hughes SM. Notochord induction of zebrafish slow muscle mediated by sonic hedgehog. Genes Dev. 1997;11:2163–2175. [PubMed]
  • Bondurand N, Pingault V, Goerich DE, Lemort N, Sock E, Caignec CL, Wegner M, Goossens M. Interaction among sox10, pax3 and mitf, three genes altered in waardenburg syndrome. Hum Mol Genet. 2000;9:1907–1917. [PubMed]
  • Carney TJ, Dutton KA, Greenhill E, Delfino-Machin M, Dufourcq P, Blader P, Kelsh RN. A direct role for sox10 in specification of neural crest-derived sensory neurons. Development. 2006;133:4619–4630. [PubMed]
  • Chan WY, Cheung CS, Yung KM, Copp AJ. Cardiac neural crest of the mouse embryo: Axial level of origin, migratory pathway and cell autonomy of the splotch (sp2h) mutant effect. Development. 2004;131:3367–3379. [PubMed]
  • Chenna R, Sugawara H, Koike T, Lopez R, Gibson TJ, Higgins DG, Thompson JD. Multiple sequence alignment with the clustal series of programs. Nucleic Acids Res. 2003;31:3497–3500. [PMC free article] [PubMed]
  • Chi N, Epstein JA. Getting your pax straight: Pax proteins in development and disease. Trends Genet. 2002;18:41–47. [PubMed]
  • Coutelle O, Blagden CS, Hampson R, Halai C, Rigby PW, Hughes SM. Hedgehog signalling is required for maintenance of myf5 and myod expression and timely terminal differentiation in zebrafish adaxial myogenesis. Dev Biol. 2001;236:136–150. [PubMed]
  • Davis GK, D’Alessio JA, Patel NH. Pax3/7 genes reveal conservation and divergence in the arthropod segmentation hierarchy. Dev Biol. 2005;285:169–184. [PubMed]
  • Dutton KA, Pauliny A, Lopes SS, Elworthy S, Carney TJ, Rauch J, Geisler R, Haffter P, Kelsh RN. Zebrafish colourless encodes sox10 and specifies non-ectomesenchymal neural crest fates. Development. 2001;128:4113–4125. [PubMed]
  • Eisen JS, Weston JA. Development of the neural crest in the zebrafish. Dev Biol. 1993;159:50–59. [PubMed]
  • Elworthy S, Lister JA, Carney TJ, Raible DW, Kelsh RN. Transcriptional regulation of mitfa accounts for the sox10 requirement in zebrafish melanophore development. Development. 2003;130:2809–2818. [PubMed]
  • Elworthy S, Pinto JP, Pettifer A, Cancela ML, Kelsh RN. Phox2b function in the enteric nervous system is conserved in zebrafish and is sox10-dependent. Mech Dev. 2005;122:659–669. [PubMed]
  • Galibert MD, Yavuzer U, Dexter TJ, Goding CR. Pax3 and regulation of the melanocyte-specific tyrosinase-related protein-1 promoter. J Biol Chem. 1999;274:26894–26900. [PubMed]
  • Goulding MD, Chalepakis G, Deutsch U, Erselius JR, Gruss P. Pax-3, a novel murine DNA binding protein expressed during early neurogenesis. Embo J. 1991;10:1135–1147. [PubMed]
  • Hammond CL, Hinits Y, Osborn DP, Minchin JE, Tettamanti G, Hughes SM. Signals and myogenic regulatory factors restrict pax3 and pax7 expression to dermomyotome-like tissue in zebrafish. Dev Biol. 2007;302:504–521. [PMC free article] [PubMed]
  • Harris ML, Erickson CA. Lineage specification in neural crest cell pathfinding. Dev Dyn. 2007;236:1–19. [PubMed]
  • Holland LZ, Schubert M, Kozmik Z, Holland ND. Amphipax3/7, an amphioxus paired box gene: Insights into chordate myogenesis, neurogenesis, and the possible evolutionary precursor of definitive vertebrate neural crest. Evol Dev. 1999;1:153–165. [PubMed]
  • Hong CS, Saint-Jeannet JP. The activity of pax3 and zic1 regulates three distinct cell fates at the neural plate border. Mol Biol Cell. 2007;18:2192–2202. [PMC free article] [PubMed]
  • Hornyak TJ, Hayes DJ, Chiu LY, Ziff EB. Transcription factors in melanocyte development: Distinct roles for pax-3 and mitf. Mech Dev. 2001;101:47–59. [PubMed]
  • Johnson SL, Africa D, Walker C, Weston JA. Genetic control of adult pigment stripe development in zebrafish. Dev Biol. 1995;167:27–33. [PubMed]
  • Kawakami A, Kimura-Kawakami M, Nomura T, Fujisawa H. Distributions of pax6 and pax7 proteins suggest their involvement in both early and late phases of chick brain development. Mech Dev. 1997;66:119–130. [PubMed]
  • Kelsh RN. Genetics and evolution of pigment patterns in fish. Pigment Cell Res. 2004;17:326–336. [PubMed]
  • Kelsh RN, Brand M, Jiang YJ, Heisenberg CP, Lin S, Haffter P, Odenthal J, Mullins MC, van Eeden FJ, Furutani-Seiki M, Granato M, Hammerschmidt M, Kane DA, Warga RM, Beuchle D, Vogelsang L, Nüsslein-Volhard C. Zebrafish pigmentation mutations and the processes of neural crest development. Development. 1996;123:369–389. [PubMed]
  • Kelsh RN, Dutton K, Medlin J, Eisen JS. Expression of zebrafish fkd6 in neural crest-derived glia. Mech Dev. 2000a;93:161–164. [PubMed]
  • Kelsh RN, Eisen JS. The zebrafish colourless gene regulates development of non-ectomesenchymal neural crest derivatives. Development. 2000;127:515–525. [PubMed]
  • Kelsh RN, Schmid B, Eisen JS. Genetic analysis of melanophore development in zebrafish embryos. Dev Biol. 2000b;225:277–293. [PubMed]
  • LaBonne C, Bronner-Fraser M. Molecular mechanisms of neural crest formation. Annu Rev Cell Dev Biol. 1999;15:81–112. [PubMed]
  • Lacosta AM, Canudas J, Gonzalez C, Muniesa P, Sarasa M, Dominguez L. Pax7 identifies neural crest, chromatophore lineages and pigment stem cells during zebrafish development. Int J Dev Biol. 2007;51:327–331. [PubMed]
  • Lang D, Chen F, Milewski R, Li J, Lu MM, Epstein JA. Pax3 is required for enteric ganglia formation and functions with sox10 to modulate expression of c-ret. J Clin Invest. 2000;106:963–971. [PMC free article] [PubMed]
  • Lang D, Epstein JA. Sox10 and pax3 physically interact to mediate activation of a conserved c-ret enhancer. Hum Mol Genet. 2003;12:937–945. [PubMed]
  • Lang D, Lu MM, Huang L, Engleka KA, Zhang M, Chu EY, Lipner S, Skoultchi A, Millar SE, Epstein JA. Pax3 functions at a nodal point in melanocyte stem cell differentiation. Nature. 2005;433:884–887. [PubMed]
  • Le Douarin NM, Creuzet S, Couly G, Dupin E. Neural crest cell plasticity and its limits. Development. 2004;131:4637–4650. [PubMed]
  • Le Guyader S, Jesuthasan S. Analysis of xanthophore and pterinosome biogenesis in zebrafish using methylene blue and pteridine autofluorescence. Pigment Cell Res. 2002;15:27–31. [PubMed]
  • Lewis JL. PhD thesis. University of Washington; 2005. Reiterated signal transduction pathways and development of the zebrafish neural crest.
  • Lewis JL, Bonner J, Modrell M, Ragland JW, Moon RT, Dorsky RI, Raible DW. Reiterated wnt signaling during zebrafish neural crest development. Development. 2004;131:1299–1308. [PubMed]
  • Li J, Liu KC, Jin F, Lu MM, Epstein JA. Transgenic rescue of congenital heart disease and spina bifida in splotch mice. Development. 1999;126:2495–2503. [PubMed]
  • Lister JA. Development of pigment cells in the zebrafish embryo. Microsc Res Tech. 2002;58:435–441. [PubMed]
  • Lister JA, Robertson CP, Lepage T, Johnson SL, Raible DW. Nacre encodes a zebrafish microphthalmia-related protein that regulates neural-crest-derived pigment cell fate. Development. 1999;126:3757–3767. [PubMed]
  • Maderspacher F, Nüsslein-Volhard C. Formation of the adult pigment pattern in zebrafish requires leopard and obelix dependent cell interactions. Development. 2003;130:3447–3457. [PubMed]
  • Mansouri A, Pla P, Larue L, Gruss P. Pax3 acts cell autonomously in the neural tube and somites by controlling cell surface properties. Development. 2001;128:1995–2005. [PubMed]
  • Mansouri A, Stoykova A, Torres M, Gruss P. Dysgenesis of cephalic neural crest derivatives in pax7−/− mutant mice. Development. 1996;122:831–838. [PubMed]
  • Mellgren EM, Johnson SL. A requirement for kit in embryonic zebrafish melanocyte differentiation is revealed by melanoblast delay. Dev Genes Evol. 2004;214:493–502. [PubMed]
  • Mellgren EM, Johnson SL. Pyewacket, a new zebrafish fin pigment pattern mutant. Pigment Cell Res. 2006;19:232–238. [PubMed]
  • Meulemans D, Bronner-Fraser M. Gene-regulatory interactions in neural crest evolution and development. Dev Cell. 2004;7:291–299. [PubMed]
  • Monsoro-Burq AH, Wang E, Harland R. Msx1 and pax3 cooperate to mediate fgf8 and wnt signals during xenopus neural crest induction. Dev Cell. 2005;8:167–178. [PubMed]
  • Nicholas KB, Jr., N.H.B. Deerfield DWI. Genedoc: Analysis and visualization of genetic variation. EMBNEW.NEWS. 1997;4:14.
  • Odenthal J, Nüsslein-Volhard C. Fork head domain genes in zebrafish. Dev Genes Evol. 1998;208:245–258. [PubMed]
  • Odenthal J, Rossnagel K, Haffter P, Kelsh RN, Vogelsang E, Brand M, van Eeden FJ, Furutani-Seiki M, Granato M, Hammerschmidt M, Heisenberg CP, Jiang YJ, Kane DA, Mullins MC, Nüsslein-Volhard C. Mutations affecting xanthophore pigmentation in the zebrafish, danio rerio. Development. 1996;123:391–398. [PubMed]
  • Parichy DM. Evolution of danio pigment pattern development. Heredity. 2006;97:200–210. [PubMed]
  • Parichy DM, Ransom DG, Paw B, Zon LI, Johnson SL. An orthologue of the kit-related gene fms is required for development of neural crest-derived xanthophores and a subpopulation of adult melanocytes in the zebrafish, danio rerio. Development. 2000;127:3031–3044. [PubMed]
  • Parichy DM, Turner JM. Temporal and cellular requirements for fms signaling during zebrafish adult pigment pattern development. Development. 2003;130:817–833. [PubMed]
  • Park HC, Kim CH, Bae YK, Yeo SY, Kim SH, Hong SK, Shin J, Yoo KW, Hibi M, Hirano T, Miki N, Chitnis AB, Huh TL. Analysis of upstream elements in the huc promoter leads to the establishment of transgenic zebrafish with fluorescent neurons. Dev Biol. 2000;227:279–293. [PubMed]
  • Pelletier I, Bally-Cuif L, Ziegler I. Cloning and developmental expression of zebrafish gtp cyclohydrolase i. Mech Dev. 2001;109:99–103. [PubMed]
  • Potterf SB, Furumura M, Dunn KJ, Arnheiter H, Pavan WJ. Transcription factor hierarchy in waardenburg syndrome: Regulation of mitf expression by sox10 and pax3. Hum Genet. 2000;107:1–6. [PubMed]
  • Pujic Z, Malicki J. Mutation of the zebrafish glass onion locus causes early cell-nonautonomous loss of neuroepithelial integrity followed by severe neuronal patterning defects in the retina. Dev Biol. 2001;234:454–469. [PubMed]
  • Raible DW, Eisen JS. Restriction of neural crest cell fate in the trunk of the embryonic zebrafish. Development. 1994;120:495–503. [PubMed]
  • Raible DW, Eisen JS. Regulative interactions in zebrafish neural crest. Development. 1996;122:501–507. [PubMed]
  • Raible DW, Wood A, Hodsdon W, Henion PD, Weston JA, Eisen JS. Segregation and early dispersal of neural crest cells in the embryonic zebrafish. Dev Dyn. 1992;195:29–42. [PubMed]
  • Ronquist F, Huelsenbeck JP. Mrbayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics. 2003;19:1572–1574. [PubMed]
  • Sato T, Sasai N, Sasai Y. Neural crest determination by co-activation of pax3 and zic1 genes in xenopus ectoderm. Development. 2005;132:2355–2363. [PubMed]
  • Sauka-Spengler T, Meulemans D, Jones M, Bronner-Fraser M. Ancient evolutionary origin of the neural crest gene regulatory network. Dev Cell. 2007;13:405–420. [PubMed]
  • Schmidt HA, Strimmer K, Vingron M, von Haeseler A. Tree-puzzle: Maximum likelihood phylogenetic analysis using quartets and parallel computing. Bioinformatics. 2002;18:502–504. [PubMed]
  • Seo HC, Saetre BO, Havik B, Ellingsen S, Fjose A. The zebrafish pax3 and pax7 homologues are highly conserved, encode multiple isoforms and show dynamic segment-like expression in the developing brain. Mech Dev. 1998;70:49–63. [PubMed]
  • Serbedzija GN, McMahon AP. Analysis of neural crest cell migration in splotch mice using a neural crest-specific lacz reporter. Dev Biol. 1997;185:139–147. [PubMed]
  • Silver DL, Hou L, Pavan WJ. The genetic regulation of pigment cell development. Adv Exp Med Biol. 2006;589:155–169. [PubMed]
  • Stark MR, Sechrist J, Bronner-Fraser M, Marcelle C. Neural tube-ectoderm interactions are required for trigeminal placode formation. Development. 1997;124:4287–4295. [PubMed]
  • Svetic V, Hollway GE, Elworthy S, Chipperfield TR, Davison C, Adams RJ, Eisen JS, Ingham PW, Currie PD, Kelsh RN. Sdf1a patterns zebrafish melanophores and links the somite and melanophore pattern defects in choker mutants. Development. 2007;134:1011–1022. [PubMed]
  • Swofford D. Paup*: Phylogenetic analysis using parsimony (* and other methods) Sinauer Associates; 2000.
  • Tassabehji M, Read AP, Newton VE, Harris R, Balling R, Gruss P, Strachan T. Waardenburg’s syndrome patients have mutations in the human homologue of the pax-3 paired box gene. Nature. 1992;355:635–636. [PubMed]
  • Tremblay P, Gruss P. Pax: Genes for mice and men. Pharmacol Ther. 1994;61:205–226. [PubMed]
  • Wada H, Holland PW, Sato S, Yamamoto H, Satoh N. Neural tube is partially dorsalized by overexpression of hrpax-37: The ascidian homologue of pax-3 and pax-7. Dev Biol. 1997;187:240–252. [PubMed]
  • Watanabe A, Takeda K, Ploplis B, Tachibana M. Epistatic relationship between waardenburg syndrome genes mitf and pax3. Nat Genet. 1998;18:283–286. [PubMed]
  • Westerfield M. The zebrafish book-a guide for the laboratory use of zebrafish (Danio rerio) 1995
  • Yajima I, Endo K, Sato S, Toyoda R, Wada H, Shibahara S, Numakunai T, Ikeo K, Gojobori T, Goding CR, Yamamoto H. Cloning and functional analysis of ascidian mitf in vivo: Insights into the origin of vertebrate pigment cells. Mech Dev. 2003;120:1489–1504. [PubMed]
  • Yang CT, Johnson SL. Small molecule-induced ablation and subsequent regeneration of larval zebrafish melanocytes. Development. 2006;133:3563–3573. [PubMed]
  • Ziegler I. The pteridine pathway in zebrafish: Regulation and specification during the determination of neural crest cell-fate. Pigment Cell Res. 2003;16:172–182. [PubMed]