Certain cranial neural crest cells are uniquely endowed with the ability to make skeletal cell types otherwise only derived from mesoderm. As these cells migrate into the pharyngeal arches, they downregulate neural crest specifier genes and upregulate so-called ectomesenchyme genes that are characteristic of skeletal progenitors. Although both external and intrinsic factors have been proposed as triggers of this transition, the details remain obscure. Here, we report the Nr2f nuclear receptors as intrinsic activators of the ectomesenchyme program: zebrafish nr2f5 single and nr2f2;nr2f5 double mutants show marked delays in upregulation of ectomesenchyme genes, such as dlx2a, prrx1a, prrx1b, sox9a, twist1a and fli1a, and in downregulation of sox10, which is normally restricted to early neural crest and non-ectomesenchyme lineages. Mutation of sox10 fully rescued skeletal development in nr2f5 single but not nr2f2;nr2f5 double mutants, but the initial ectomesenchyme delay persisted in both. Sox10 perdurance thus antagonizes the recovery but does not explain the impaired ectomesenchyme transition. Unraveling the mechanisms of Nr2f function will help solve the enduring puzzle of how cranial neural crest cells transition to the skeletal progenitor state.

Neural crest cells (NCCs) arise from the neural plate border at all axial levels, but only a subset of those born in the cranial region give rise to skeletal derivatives such as cartilage and bone (reviewed by Fabian and Crump, 2022). Although quail cranial NCCs (CNCCs) transplanted into the trunk form ectopic cartilage nodules (Le Douarin and Teillet, 1974), trunk NCCs cannot do the same in the cranial region (Nakamura and Ayer-le Lievre, 1982), implying intrinsic potency of CNCCs. Work in chick has identified three cranially expressed transcription factors (Sox8, Tfap2b and Ets1) that, when ectopically expressed in trunk NCCs, allow them to make cartilage when transplanted into the head (Simoes-Costa and Bronner, 2016). In mouse, premigratory NCCs in the head but not the trunk also transiently reactivate the pluripotency factor Oct4 (Zalc et al., 2021). Finally, CNCCs fated to make skeleton appear to arise more laterally on the neural plate border, migrate earlier (Lee et al., 2013; Weston and Thiery, 2015), and may be transcriptionally biased to a mesenchymal fate (Soldatov et al., 2019) compared with NCCs destined for non-skeletal lineages.

A defining feature of the skeletal CNCC subpopulation is that they transition from a Sox10+ migratory neural crest (NC) state to a Sox10 skeletal progenitor or ‘ectomesenchyme’ transcriptional program as they arrive in the pharyngeal arches and frontonasal region (Blentic et al., 2008). Other CNCCs that instead retain Sox10 expression differentiate into non-ectomesenchymal lineages, including neurons, melanocytes and glia (Kelsh and Eisen, 2000). Several factors have been proposed to trigger the transition to the ectomesenchyme state: Twist1, which is expressed in migrating CNCCs, is thought to repress Sox10 expression and push cells towards the ectomesenchyme fate; however, mutants still form some head skeleton (Bildsoe et al., 2009; Das and Crump, 2012; Teng et al., 2018; Soldatov et al., 2019). Extrinsic signals also contribute: gradual loss of Bmp activity and gain of Fgf activity in migrating CNCCs are essential for activation of ectomesenchyme genes such as Dlx2 in chick and zebrafish (Blentic et al., 2008; Das and Crump, 2012). However, the precise mechanics of how these and/or other factors orchestrate the transcriptional switch are unknown.

The nuclear receptors NR2F1 and NR2F2 were first identified as NC regulators in an in vitro model of NC differentiation from human embryonic stem cells (Rada-Iglesias et al., 2012). Thousands of putative NC enhancers were co-occupied by NR2F1 and NR2F2 as cells differentiated into NC. Knockdown experiments indicated that NR2F1 drives NC gene expression in vitro and regulates craniofacial development in frogs. Patients carrying heterozygous variants in NR2F1 or NR2F2 show only mild facial dysmorphism but severe visual and cognitive (NR2F1; Bosch et al., 2014; Chen et al., 2016) or heart, diaphragm and gonad (NR2F2; Poot et al., 2007; High et al., 2016; Upadia et al., 2018; Bashamboo et al., 2018) phenotypes, suggesting that NR2F1 and NR2F2 may be functionally redundant specifically in NC. The third mammalian Nr2f gene, NR2F6, is not associated with a genetic disorder, and no craniofacial abnormalities have been reported in mouse Nr2f6 mutants (Warnecke et al., 2005).

We have previously generated stable mutant lines for the zebrafish Nr2f genes to better investigate their functions. Zebrafish have six homologs [nr2f1a, nr2f1b, nr2f2, nr2f5, nr2f6a and nr2f6b (Fig. S1E)], due to the teleost genome duplication and to retention of the fourth ancestral family member, nr2f5, which was lost in most amniote lineages (Coppola and Waxman, 2021). Although no single mutants or even nr2f1a;nr2f1b;nr2f2 triple mutants have overt NC defects, nr2f2;nr2f5 double mutants have a surprising craniofacial phenotype wherein the larval upper jaw is transformed into a mirror image of the lower jaw (Barske et al., 2018). Molecular analyses revealed misexpression of mandibular genes in maxillary precursors in mutants, indicating an essential role for Nr2f2/5 in preserving upper jaw identity. Whether mammalian NR2F1 is the functional equivalent of fish Nr2f5 has yet to be determined.

Here, we provide the first in vivo evidence that zebrafish Nr2fs are also crucial regulators of the ectomesenchyme transition. We find that the jaw patterning phenotype is preceded by a widespread delay in activation of ectomesenchyme genes in post-migratory CNCCs. Some mutant genotypes fully recover from this early defect, testifying to the complementarity of the mechanisms that drive differentiation of CNCCs into skeletal progenitors.

Delayed activation of ectomesenchyme genes in Nr2f mutants

Although our previous study focused on Nr2f-mediated regulation of jaw patterning (Barske et al., 2018), work in the human in vitro model indicates that Nr2fs may also function at earlier stages of NC development (Rada-Iglesias et al., 2012). We noted no differences in the streams of Sox10+ CNCCs arriving at the arches at 18 hpf between control and nr2f2el506/el506;nr2f5el611/el611 mutants (abbreviated to ‘el’ hereafter) (Fig. S1A). However, in 26 hpf embryos carrying the general NC reporter sox10:DsRed and the ectomesenchyme/vasculature reporter fli1a:EGFP, we noted that approximately one-quarter of each clutch had reduced EGFP in the arches but not in the forming blood vessels, while DsRed levels were unaffected (Fig. 1A). Affected individuals all genotyped as mutant for nr2f5, but were variably wild type, heterozygous or mutant for nr2f2. As endogenous fli1a (fli1) expression is activated in post-migratory CNCCs committing to ectomesenchyme fate (Brown et al., 2000), we reasoned that this transition might be impaired in nr2f5 mutants. However, the effect is transient: recovery to wild-type EGFP levels was observed at 36 hpf (Fig. S1B), and nr2f5el/el mutants that are heterozygous or wild type for nr2f2 ultimately develop normal facial skeletons and are viable.

Fig. 1.

Suppression of the ectomesenchyme program in nr2f2el/el;nr2f5el/el mutants. (A) nr2f5 single and nr2f2;nr2f5 double mutants show reduced fli1a:EGFP in the arches at 26 hpf. sox10:DsRed marks NC-derived cells; dashed lines indicate arch boundaries. (B) Breeding strategy for FACS and RNAseq of sox10:DsRed+ cells from control and nr2f5 mutant embryos. (C) DESeq2-generated log2-transformed fold-change values of genes enriched in distinct NC lineages between mutants and controls. Bars indicate the mean. (D) In situ hybridizations for ectomesenchyme genes in controls and double mutants at 20 hpf (dorsal views, anterior towards the left). Arrowheads indicate diminished expression. The three bilateral CNCC streams are marked in the top left image. Bottom right numbers are log2-transformed fold-change values from the RNAseq experiment. *The fli1a fold-change value is considered unreliable due to the fli1a:EGFP transgene being present in the mutant but not in the control sample. Scale bars: 20 µm in A; 100 µm in D.

Fig. 1.

Suppression of the ectomesenchyme program in nr2f2el/el;nr2f5el/el mutants. (A) nr2f5 single and nr2f2;nr2f5 double mutants show reduced fli1a:EGFP in the arches at 26 hpf. sox10:DsRed marks NC-derived cells; dashed lines indicate arch boundaries. (B) Breeding strategy for FACS and RNAseq of sox10:DsRed+ cells from control and nr2f5 mutant embryos. (C) DESeq2-generated log2-transformed fold-change values of genes enriched in distinct NC lineages between mutants and controls. Bars indicate the mean. (D) In situ hybridizations for ectomesenchyme genes in controls and double mutants at 20 hpf (dorsal views, anterior towards the left). Arrowheads indicate diminished expression. The three bilateral CNCC streams are marked in the top left image. Bottom right numbers are log2-transformed fold-change values from the RNAseq experiment. *The fli1a fold-change value is considered unreliable due to the fli1a:EGFP transgene being present in the mutant but not in the control sample. Scale bars: 20 µm in A; 100 µm in D.

To further explore this phenotype, we separately in-crossed adult nr2f2el/+;nr2f5el/el;sox10:DsRed fish and sox10:DsRed control fish, FACS-purified DsRed+ cells from pools of 20 hpf embryos, and performed bulk RNA sequencing (Fig. 1B). DESeq2 analysis revealed 441 genes significantly upregulated and 862 significantly downregulated in mutants by ≥1.5-fold (FDR-adjusted P<0.05) (Fig. S1C, Tables S1-S3). nr2f5 was the most significantly affected gene (downregulated >10-fold; P=1.26×10−127).

To assess effects on the ectomesenchyme program, we evaluated marker genes reported in a recent scRNAseq study (Tatarakis et al., 2021) to be enriched in CNCCs differentiating into skeletal lineages at ‘early’ or ‘later’ stages. Nine out of 19 and 12/40 genes in the ‘early’ and ‘later’ gene sets were significantly downregulated >1.5-fold in mutants, respectively, whereas only 1/19 and 2/40 were significantly upregulated (Fig. 1C, Table S4). We then performed in situ hybridizations on 20 hpf nr2f5el/el single and nr2f2el/el;nr2f5el/el double mutants for eight of the skeletal genes (dlx2a, dlx3b, dlx5a, twist1a, prrx1a, prrx1b, grem2b and fli1a), plus sox9a and the endothelin receptor genes ednraa and ednrab, which are also enriched in early ectomesenchyme (Nair et al., 2007; Paudel et al., 2022). All but ednrab appeared reduced or absent in the arches of double mutants and to a lesser degree in single mutants (Fig. 1D, Fig. S2). The program eventually recovers: expression of each gene is detectable by 36 hpf (Barske et al., 2018).

To test whether alternative fates are activated in mutants instead of the ectomesenchyme program, we looked at gene sets for ‘neural/glial’ and ‘pigment’ progenitors (Tatarakis et al., 2021). Five out of 14 neural/glial markers and one out of 16 pigment genes were significantly upregulated in nr2f2el/el;nr2f5el/el mutants (Fig. 1C). We noted, however, that two of the five ‘neural/glial’ genes (sox10 and foxd3) are also markers of the early NC. We also performed in situ hybridizations for additional neural (nr4a2b) and pigment (gch2) genes that were upregulated upon knockdown of the key ectomesenchyme genes twist1a and twist1b (Das and Crump, 2012). Neither was elevated in nr2f2el/el;nr2f5el/el mutants at 20 hpf (Fig. 2A).

Fig. 2.

Behavior of non-ectomesenchyme and early NC genes in nr2f2el/el;nr2f5el/el mutants. (A) In situ hybridizations analysis showing that markers of alternate lineages are not overtly upregulated at 20 hpf. (B) In situ hybridizations for early NC genes in control and double mutants at 20 hpf. Arrowheads indicate elevated sox10 expression. In A,B, top right numbers are the average TPM values in the RNAseq data, with the log2-transformed fold-change DESeq2 values underneath. (C) Immunostaining showing ectopic Sox10 protein in the arches (outlined) at 25 hpf. sox10:kikGR marks NC-derived cells. Scale bars: 100 µm.

Fig. 2.

Behavior of non-ectomesenchyme and early NC genes in nr2f2el/el;nr2f5el/el mutants. (A) In situ hybridizations analysis showing that markers of alternate lineages are not overtly upregulated at 20 hpf. (B) In situ hybridizations for early NC genes in control and double mutants at 20 hpf. Arrowheads indicate elevated sox10 expression. In A,B, top right numbers are the average TPM values in the RNAseq data, with the log2-transformed fold-change DESeq2 values underneath. (C) Immunostaining showing ectopic Sox10 protein in the arches (outlined) at 25 hpf. sox10:kikGR marks NC-derived cells. Scale bars: 100 µm.

We next tested whether the upregulation of sox10 and foxd3 might instead reflect a failure to deactivate early NC genes. However, within the ‘early NC’ gene set (Tatarakis et al., 2021), only four (sox10, foxd3, gbx1 and her9) out of 21 genes were significantly upregulated >1.5-fold. We performed in situ hybridizations for sox10 and foxd3 as well as for other early NC genes (sox9b, msx1a, tfap2a and zic2a) at 20 hpf (Fig. 2B). Only sox10 was visually upregulated, consistent with its higher baseline expression. Sox10 protein was ectopically present in mutant CNCCs as late as 25 hpf (Fig. 2C), although the mRNA was no longer detectable (Fig. S1D). The delay in ectomesenchyme gene activation is thus accompanied by a failure to turn off some, but not all, early NC genes, such as sox10.

Compensation of the nr2f5 mutant allele allows full recovery of skeletal development

How the ectomesenchyme program recoups sufficiently to allow nr2f5el/el and nr2f2el/el;nr2f5el/el mutant CNCCs to form skeleton is unclear. Redundancy with the four other Nr2f family members may provide one mechanism of recovery. None is expressed as highly as nr2f5 at 20 hpf, but all are later upregulated in arch NCCs (Barske et al., 2018). The six genes contain highly conserved DNA-binding domains (DBDs, 84-100% identical) and ligand-binding domains (68-97% identical). In line with general redundancy, we observe that almost no facial skeleton forms when the nr2f5 mutation is combined with triple 1a/1b/2 or 2/6a/6b mutations (see Barske et al., 2018; Fig. S1F).

Another mechanism that may contribute to recovery is transcriptional adaptation of related genes, triggered by nonsense-mediated decay (NMD) of mutant transcripts (Rossi et al., 2015; El-Brolosy et al., 2019). nr2f5el611 is a 10-bp deletion that results in a premature stop codon (Barske et al., 2018), and mRNA levels were reduced more than tenfold in homozygotes, suggesting that the transcript is subject to NMD (Table S2). Reasoning that compensating genes would likely be nuclear receptors expressed above threshold levels in mutants by 20 hpf, we filtered the 73 known zebrafish nuclear receptors (Schaaf, 2017) for expression ≥10 TPM in mutant NCCs. This excluded all but the five other Nr2fs plus rarab, rarga, rargb, ppardb, rxraa and nr2c1. Only nr2f1b was significantly upregulated, by a robust log2-transformed 1.99-fold (P=1.72×10−24, Fig. S1C,G), indicating potential compensation by this paralog.

To circumvent this, we generated a second, large deletion allele for nr2f5 (nr2f5ci3000) that lacks the 5′UTR and most of the first exon (Fig. 3A). Although this deletion did not prevent transcription, nr2f5ci3000 transcripts should not contain premature stop codons and do not appear to be degraded by NMD (Fig. 3B). Translation of nr2f5ci3000 transcripts from the first methionine after the deletion would produce a truncated protein lacking the DBD. nr2f5ci/ci homozygotes show severe craniofacial defects (Fig. 3E; compare with seemingly normal nr2f5el/el skeletons in Fig. 3D) and are not viable after larval stages. The second arch-derived hyomandibula is grossly reduced (n=18/18), the ceratohyal cartilage is not angled anteriorly (n=13/13), the first ceratobranchial is reduced and ceratobranchials 2-4 from arches 4-6 do not form (n=10/10). The hyomandibula of nr2f5ci/ci mutants resembles that of nr2f2el/el;nr2f5el/el mutants heterozygous or mutant for nr2f1b (compare Fig. 3E′ with G′), supporting the observation that nr2f1b and nr2f2 are functionally compensating for the original nr2f5el611 allele.

Fig. 3.

Worsened skeletal phenotypes in new nr2f5 mutant. (A) Schematic of nr2f5, with the deletions, RT-PCR primers (rtF and rtR) and functional domains marked. (B) RT-PCR quantification showing similar normalized nr2f5 transcript levels in wild types and nr2f5ci3000 mutants at 52 hpf. (C-H′) Skeletal preparations of nr2f5 single and combinatorial mutants, visualized as ventral (C-H) and lateral (C′-H′) mounts. Black arrowheads indicate the dysmorphic hyomandibula; brackets highlight missing ceratobranchials; pink arrowheads indicate the transformed upper jaw cartilage. (I) Disrupted arch morphology in double mutants at 30 hpf revealed with the sox10:kikGR transgene. Arches are numbered, and the white dotted line surrounds the first pharyngeal pouch. (J,K) Expression of the mandibular markers dlx5a and dlx6a is expanded into the maxillary prominence (green arrowhead) in nr2f2ci/ci;nr2f5ci/ci mutants. Scale bars: 100 µm in C-I; 50 µm in J,K.

Fig. 3.

Worsened skeletal phenotypes in new nr2f5 mutant. (A) Schematic of nr2f5, with the deletions, RT-PCR primers (rtF and rtR) and functional domains marked. (B) RT-PCR quantification showing similar normalized nr2f5 transcript levels in wild types and nr2f5ci3000 mutants at 52 hpf. (C-H′) Skeletal preparations of nr2f5 single and combinatorial mutants, visualized as ventral (C-H) and lateral (C′-H′) mounts. Black arrowheads indicate the dysmorphic hyomandibula; brackets highlight missing ceratobranchials; pink arrowheads indicate the transformed upper jaw cartilage. (I) Disrupted arch morphology in double mutants at 30 hpf revealed with the sox10:kikGR transgene. Arches are numbered, and the white dotted line surrounds the first pharyngeal pouch. (J,K) Expression of the mandibular markers dlx5a and dlx6a is expanded into the maxillary prominence (green arrowhead) in nr2f2ci/ci;nr2f5ci/ci mutants. Scale bars: 100 µm in C-I; 50 µm in J,K.

We also generated a new nr2f2 allele that lacks the 5′UTR and first exon (nr2f2ci3005; Fig. S3A,B). Any translation of this allele would yield a truncated protein lacking the DBD. nr2f2ci/ci homozygotes were not more severe than nr2f2el/el mutants: both develop a normal facial skeleton but show impaired melanophore constriction and perish before 2 weeks of age (Fig. S3C,D). nr2f2ci/ci;nr2f5ci/ci double mutants present the same homeotic transformation of the upper jaw as the original nr2f2el/el;nr2f5el/el mutants (Fig. 3F-H), with ectopic upregulation of the lower jaw specifiers dlx5a and dlx6a in the maxillary prominence at 36 hpf (Fig. 3J,K). However, imaging with a sox10:kikGR transgene that labels all NCCs revealed that the arches of nr2f2ci/ci;nr2f5ci/ci mutants are more radically dysmorphic at 30 hpf (Fig. 3I), and their facial skeletons are more severely reduced (Fig. 3H).

At earlier stages (20 hpf), nr2f5ci/ci and nr2f2ci/ci;nr2f5ci/ci mutants show more severe loss of ectomesenchyme gene expression (dlx2a, twist1a and sox9a) compared with nr2f5el/el and nr2f2el/el;nr2f5el/el mutants (compare Fig. 4B with 1D, Fig. S2). However, because the skeletal phenotypes are still not as severe as that of the quadruple mutants (Fig. S1F), belated redundancy from other family members is likely still facilitating their partial recovery.

Fig. 4.

Loss of sox10 rescues the skeletal but not ectomesenchyme program in nr2f5ci/ci mutants. (A) Possible genetic relationships linking Nr2f2 and/or Nr2f5, sox10 and the ectomesenchyme program. (B) Ectomesenchyme genes were more dramatically reduced in nr2f5ci/ci and double nr2f2ci/ci;nr2f5ci/ci mutants (compare with Fig. 1D, Fig. S2). Expression is not rescued by reducing sox10 dose. (C,D) Loss of one or two copies of sox10 rescues the hyomandibular and ceratobranchial phenotypes of nr2f5ci/ci single mutants (C) but not the more severe defects of nr2f2ci/ci;nr2f5ci/ci mutants (D). In C, 10/11, 4/15 and 0/13 nr2f5ci/ci mutants that were, respectively, wild type, heterozygous or mutant for sox10 showed hyomandibular and ceratobranchial phenotypes. Scale bars: 100 µm.

Fig. 4.

Loss of sox10 rescues the skeletal but not ectomesenchyme program in nr2f5ci/ci mutants. (A) Possible genetic relationships linking Nr2f2 and/or Nr2f5, sox10 and the ectomesenchyme program. (B) Ectomesenchyme genes were more dramatically reduced in nr2f5ci/ci and double nr2f2ci/ci;nr2f5ci/ci mutants (compare with Fig. 1D, Fig. S2). Expression is not rescued by reducing sox10 dose. (C,D) Loss of one or two copies of sox10 rescues the hyomandibular and ceratobranchial phenotypes of nr2f5ci/ci single mutants (C) but not the more severe defects of nr2f2ci/ci;nr2f5ci/ci mutants (D). In C, 10/11, 4/15 and 0/13 nr2f5ci/ci mutants that were, respectively, wild type, heterozygous or mutant for sox10 showed hyomandibular and ceratobranchial phenotypes. Scale bars: 100 µm.

Reduction of sox10 levels mitigates the nr2f5 skeletal phenotype

We hypothesized that dysregulation of the ectomesenchyme program could be an indirect consequence of Sox10 perdurance (Fig. 4A, Model 1). Sox10 is required for formation of non-ectomesenchymal NC lineages (Kelsh and Eisen, 2000) and is normally suppressed in ectomesenchyme-fated cells once they reach the arches (Das and Crump, 2012). Intriguingly, persistent Sox10 expression has also been observed in other cases of disrupted ectomesenchyme activation, including mouse Twist1 mutants (Bildsoe et al., 2009), fish twist1a/twist1b morphants and mutants (Das and Crump, 2012; Teng et al., 2018), fish disc1 morphants (Drerup et al., 2009), and chick and zebrafish CNCCs expressing dominant-negative FGFR1 (Blentic et al., 2008; Das and Crump, 2012). However, the pertinence of this ectopic Sox10 expression is not clear. We therefore created a new sox10 mutant allele (sox10ci3020) and crossed it onto the nr2f2ci;nr2f5ci background. This deletion removes part of the 5′UTR and the first coding exon and leads to reduced transcription (Fig. S3E,F). Homozygotes show the classic ‘colorless’ phenotype (Kelsh and Eisen, 2000) and reduction of the hyomandibular foramen traversed by the facial nerve (Fig. S3G,H). Strikingly, loss of one or both sox10 alleles rescued the hyoid and ceratobranchial defects of single nr2f5ci/ci mutants (n=24/28; Fig. 4C), yet had no impact on the nr2f2ci/ci;nr2f5ci/ci double mutant (n=0/12; Fig. 4D). No consistent rescue of dlx2a, twist1a or sox9a expression was noted in single or double Nr2f mutants on the sox10 mutant background (Fig. 4B). Rather than instigating the ectomesenchyme defect, Sox10 perdurance may instead antagonize the onset or extent of the recovery.

Alternative mechanisms of action

An alternative to Model 1 is that the Nr2fs directly activate ectomesenchyme gene expression (Fig. 4A, Model 2). To begin to test this, we analyzed ChIPseq data for NR2F1 and NR2F2 in human NC-like cells (Rada-Iglesias et al., 2012). Out of 49 human homologs of the zebrafish ‘early’ and ‘later’ skeletal genes, 27 were associated with one to eight NR2F1 or NR2F2 peaks (Table S5). Thirty-eight of the 62 total peaks showed substantial conservation according to the UCSC ‘100 vertebrates’ track; 22/62 overlapped with a predicted human NC enhancer annotated by Rada-Iglesias et al. (2012); and 47/62 overlapped with an ENCODE candidate cis-regulatory element. Candidate regulatory sequences for homologs of many ectomesenchyme genes are thus directly bound by Nr2f proteins. Whether these elements drive NCC expression in vivo, and the Nr2f dependence of their activity, will be tested in future studies. Importantly, however, because nr2f5 is expressed in NCCs before ectomesenchyme genes are activated, additional signals must trigger the transition.

Another possible mechanism is that Nr2fs could impair Fgf receptivity in CNCCs. Fgf signaling from the arch environment may extrinsically drive ectomesenchyme activation (Blentic et al., 2008). However, no Fgfr genes were significantly downregulated in mutant NCCs (Table S2), nor were homologs of genes proposed to confer skeletal potency to CNCCs in mouse (Oct4) (Zalc et al., 2021) or avians (SOX8, TFAP2B, ETS1, BRN3, DMBX1 and LHX5) (Simoes-Costa and Bronner, 2016; Martik et al., 2019). Nr2fs may therefore not drive ectomesenchyme activation via these pathways or, alternatively, may function downstream.

We conclude that Nr2f5 is an intrinsic driver of the ectomesenchymal transition in zebrafish. Recovery of craniofacial development in mutants requires redundantly acting Nr2f family members and is improved by suppression of the non-ectomesenchyme driver sox10. It remains to be determined whether the upper jaw homeosis of nr2f2;nr2f5 mutants is attributable to a later patterning function for these genes in maxillary precursors or to aberrant differentiation caused by imperfect recovery from the ectomesenchyme defect.

The Institutional Animal Care and Use Committees of Cincinnati Children's Hospital Medical Center (2018-0076 and 2021-0048) and the University of Southern California (10885 and 20540) approved all animal procedures carried out in this study.

Zebrafish lines and genotyping

Zebrafish (Danio rerio) embryos were cultured at 28.5°C in embryo medium following standard methods (Westerfield, 2007) and staged according to Kimmel et al. (1995). Mutant and transgenic lines were made and propagated on the Tübingen background (Haffter et al., 1996) and housed in groups of less than 15 fish per tank. Previously reported mutant and transgenic lines include nr2f2el506, nr2f5el611 (Barske et al., 2018), Tg(fli1a:EGFP)y1 (Lawson and Weinstein, 2002), Tg(sox10:DsRed-Express)el10 (Das and Crump, 2012) and Tg(4.9sox10:kikGR)el2 (Balczerski et al., 2012).

Generation of new zebrafish mutant lines

Five new targeted mutant alleles were generated for this study. sgRNAs (100 ng/ml) were injected into one-cell stage embryos along with 100 ng/ml Cas9 mRNA prepared with the Thermo Fisher mMessage mMachine kit. Injected individuals were raised to adulthood and outcrossed, and their F1 progeny were screened for deletions by PCR and sent for Sanger sequencing verification. The nr2f2ci3005 allele (NC_007129.7:g.23874496_23875584del) is a 1089 bp deletion created using two sgRNAs (targeting 5′-GCACAGGCTGCAACTGCGCG-3′ and 5′-GTATTTTATCCCCGACTCCT-3′) that flank the 5′UTR and first coding exon of nr2f2. Genotyping was performed with nr2f2_F3 (5′-GCAGCAGTCGTGTCAAAGTT-3′), nr2f2_F4 (5′-ACTTAATCGCGCATGCTTCT-3′) and nr2f2_R3 (5′-ATACTCCCTCGCCATAACCG-3′) primers, which produce a 285 bp product for the wild-type allele and a 526-bp product for the mutant. The nr2f5ci3000 allele is a 6 bp deletion followed by inversion of the next 72 bp and then deletion of 882 bp (NC_007127.7:g.46294195_46295154delins[NC_007127.7:g.46294201_46294272inv]), created with sgRNAs targeting sequences upstream of the 5′UTR and within the first coding exon (5′-CGCCCTGCTCTGAGGCGTC-3′ and 5′-GTGGACTGCATGGTGTGCG-3′). Genotyping was performed with nr2f5_F4 (5′-GCTGCGCCTTTGTCCAATT-3′), nr2f5_R4 (5′-AGTATTGGCACTGGTTCCGA-3′) and nr2f5_R5 (5′-TGTGTTTACTAGAGGAGATGCAGT-3′) primers, which produce a 218 bp product for the wild-type allele and a 347 bp product for the mutant. The nr2f6ael648 allele contains a SNP and a 2 bp deletion in exon 1 (NM_205557.1:c.67_71delinsGAC), created with a sgRNA targeting 5′-TGACAAGGGTTACCTGAGGG-3′ and causing premature truncation after four frameshifted amino acids (Fig. S3I). Genotyping was performed with nr2f6a_F (5′-ACAAGTGACTCAGAACTCGTG-3′) and nr2f6a_R (5′-CACACATCCCTTATCATCCTCAC-3′) primers followed by digestion with Bsu36I, producing 263 and 80 bp bands in wild types and a 343 bp band in mutants. The nr2f6bel685 allele is a 11 bp deletion with a 7 bp insertion in exon 1 (NM_213239.1:c140_150delinsGTGGACT), which was created with a sgRNA targeting 5′-GGAGGACAAGGCCTGCGTGG-3′ and produces a frameshift and premature stop after 62 incorrect amino acids (Fig. S3J). Genotyping was performed with Nr2f6b_F (5′-TTCTCGACGCAGGAAGGTTT-3′) and Nr2f6b_ R (5′-AACTCTTGCACCCTTCACAG-3′) primers, and digestion with StuI, yielding 192 bp and 84 bp products in wild types and a single 276 bp band in mutants. The sox10ci3020 allele is a 1495 bp deletion that removes the second non-coding exon and first coding exon of the sox10 gene (NC_007114.7:g.1494257_1495751del), constructed using sgRNAs targeting 5′-GTTCGGGTGAACAGCAAAG-3′ and 5′-GGGTCGTGGTGACCTAGTTT-3′. Genotyping was performed with sox10_F2 (5′-CGCCAAACCTTCAACAGTAG-3′), sox10_R1 (5′-AGGCTGTTTCTAGAGTGGCT-3′) and sox10_R2 (5′-CACAACCCCTTTGCTGTTC-3′) primers, which produce a 218 bp product for the wild-type allele and a 347 bp product for the mutant.

In situ hybridization, immunostaining and skeletal staining

Colorimetric and fluorescent in situ hybridizations were performed as described previously (Barske et al., 2018). Published probes include dlx2a and dlx3b (Akimenko et al., 1994), dlx5a (Walker et al., 2006), ednraa (Zuniga et al., 2010), ednrab (Nair et al., 2007), grem2b (Zuniga et al., 2011), msx1a (Akimenko et al., 1995), nr2f5 (Barske et al., 2018), foxd3 (Cox et al., 2012), prrx1a and prrx1b (Barske et al., 2016), sox9a and sox9b (Yan et al., 2005), sox10 (Dutton et al., 2001), and nr4a2b and gch2 (Das and Crump, 2012). cDNAs for fli1a, tfap2a, twist1a and zic2a were amplified by Herculase II Fusion DNA Polymerase (Agilent) and inserted into the pCR-Blunt II-TOPO vector (Thermo Fisher). After linearization by restriction digest, antisense probes were synthesized from each plasmid using Sp6 or T7 polymerase and digoxigenin (DIG)-tagged nucleotides (Roche) (Table S6). NBT-BCIP-treated (Sigma-Aldrich) embryos were imaged on a Leica S8APO or a Zeiss StereoDiscoveryV8 equipped with an Axiocam305 camera. Fluorescent in situs and transgenic lines were imaged on a Nikon C2 or Zeiss LSM800 confocal.

For Sox10 immunostaining, embryos from a cross between nr2f2el506/+;nr2f5el611/+;sox10:kikGR+ and nr2f2el506/+;nr2f5el611/+ adults were fixed in 4% PFA for 3 h at room temperature, washed in PBS+0.1% Tween-20, permeabilized in acetone for 7 min at −20°C, rehydrated, blocked with 2% normal goat serum, then incubated with anti-Sox10 (1:200, Genetex, GTX128374) overnight at 4°C. Alexa Fluor 568-conjugated donkey anti-rabbit secondary antibody (Thermo Fisher, A-11011) was applied at 1:300 after washing the following day, and the samples were imaged on a Nikon C2 confocal.

Alcian Blue and Alizarin Red staining of cartilage and bone in larvae was performed as described previously (Walker and Kimmel, 2007), and mounted specimens were imaged on a Zeiss AxioImager.Z1 compound microscope with a 5× or 10× objective. Sample sizes for each genotype/stain combination are provided in Table S8.

RT-PCR

Semi-quantitative reverse-transcriptase PCR (RT-PCR) was performed to estimate the transcript levels associated with different mutant alleles. Single heterozygous nr2f5ci3000, nr2f2el506, nr2f2ci3005 and sox10ci3020 fish were separately incrossed, and anesthetized offspring were fin-clipped at 24 hpf and genotyped. For each line, 6-10 mutants and 6-10 sibling wild-type controls were pooled at 48-52 hpf and frozen at −80°C. RNA was extracted using the RNAqueous-4PCR Total RNA Isolation Kit (Invitrogen), and equivalent amounts were used to synthesize cDNA with the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems). RT-PCR was run with a minimum of two biological replicates per genotype under the following protocol: 95°C for 2 min, then 28-30 cycles of 95°C for 15 s, 56°C for 20 s and 72°C for 20 s, followed by 72°C for 5 min. eef1g expression was used for normalization. Band intensity was quantified with Image Lab (BioRad) and analyzed in Prism 9 (GraphPad). RT-PCR primers are listed in Table S7.

FACS and RNA sequencing

Embryos were generated from in-crosses of control sox10:DsRed or mutant nr2f2el/+;nr2f5el/el;sox10:DsRed fish, raised to 20 hpf at 28°C, and sorted for DsRed expression. Mutant samples were therefore all nr2f5el/el but could be wild type, heterozygous or mutant for nr2f2. The parents of the mutant but not control embryos also carried fli1a:EGFP. In order to increase our mutant sample size, we did not exclude GFP+ embryos from the analysis. Embryo dissociation, FACS, RNA extraction, cDNA library preparation and RNA sequencing were performed as previously described (Barske et al., 2016). Two biological replicates were performed on separate days for each genotype, using pools of 120 and 110 (control) or 88 and 35 (mutant) embryos.

Data analysis

Raw Fastq files were trimmed, pseudoaligned and quantified with Kallisto (Bray et al., 2016) to determine counts and transcript per kilobase million (TPM) values. Counts were submitted to DESeq2 (Love et al., 2014) analysis to determine differential expression values and the associated significance. Published ChIP data (Rada-Iglesias et al., 2012) were transferred from hg18 to hg38 coordinates using the UCSC LiftOver program, then subjected to GREAT (McLean et al., 2010) to assign each peak to the nearest gene(s).

We thank Danielle Fritsch and Eric Alley of CCHMC Veterinary Services and Megan Matsutani at USC for fish care, Krishna Roskin of the CCHMC Bioinformatics Collaborative Core, Matthew Kofron of the CCHMC Confocal Imaging Core, Jeffrey Boyd at the USC Flow Cytometry Core Facility, and the USC Molecular Genomics Core for core services. We also thank Peter Fabian for critical reading of the manuscript.

Author contributions

Conceptualization: C.O., L.B.; Methodology: C.O., S.P., L.B.; Validation: C.O., L.B.; Formal analysis: L.B.; Investigation: C.O., D.P., A.R., C.P., L.B.; Resources: C.S.T.; Writing - original draft: C.O., L.B.; Visualization: L.B.; Supervision: L.B.; Project administration: L.B.; Funding acquisition: L.B.

Funding

This work was supported by the National Institutes of Health (R00 DE026239 to L.B.) and lab startup funds from the Cincinnati Children's Research Foundation. Deposited in PMC for release after 12 months.

Data availability

Raw RNAseq files have been deposited in GEO under accession number GSE206903.

Akimenko
,
M. A.
,
Ekker
,
M.
,
Wegner
,
J.
,
Lin
,
W.
and
Westerfield
,
M.
(
1994
).
Combinatorial expression of three zebrafish genes related to distal-less: part of a homeobox gene code for the head
.
J. Neurosci.
14
,
3475
-
3486
.
Akimenko
,
M. A.
,
Johnson
,
S. L.
,
Westerfield
,
M.
and
Ekker
,
M.
(
1995
).
Differential induction of four msx homeobox genes during fin development and regeneration in zebrafish
.
Development
121
,
347
-
357
.
Balczerski
,
B.
,
Matsutani
,
M.
,
Castillo
,
P.
,
Osborne
,
N.
,
Stainier
,
D. Y. R.
and
Crump
,
J. G.
(
2012
).
Analysis of sphingosine-1-phosphate signaling mutants reveals endodermal requirements for the growth but not dorsoventral patterning of jaw skeletal precursors
.
Dev. Biol.
362
,
230
-
241
.
Barske
,
L.
,
Askary
,
A.
,
Zuniga
,
E.
,
Balczerski
,
B.
,
Bump
,
P.
,
Nichols
,
J. T.
and
Crump
,
J. G.
(
2016
).
Competition between Jagged-Notch and endothelin1 signaling selectively restricts cartilage formation in the zebrafish upper face
.
PLoS Genet.
12
,
e1005967
.
Barske
,
L.
,
Rataud
,
P.
,
Behizad
,
K.
,
Del Rio
,
L.
,
Cox
,
S. G.
and
Crump
,
J. G.
(
2018
).
Essential role of Nr2f nuclear receptors in patterning the vertebrate upper jaw
.
Dev. Cell
44
,
337
-
347.e5
.
Bashamboo
,
A.
,
Eozenou
,
C.
,
Jorgensen
,
A.
,
Bignon-Topalovic
,
J.
,
Siffroi
,
J.-P.
,
Hyon
,
C.
,
Tar
,
A.
,
Nagy
,
P.
,
Sólyom
,
J.
,
Halász
,
Z.
et al.
(
2018
).
Loss of function of the nuclear receptor NR2F2, encoding COUP-TF2, causes testis development and cardiac defects in 46,XX children
.
Am. J. Hum. Genet.
102
,
487
-
493
.
Bildsoe
,
H.
,
Loebel
,
D. A. F.
,
Jones
,
V. J.
,
Chen
,
Y.-T.
,
Behringer
,
R. R.
and
Tam
,
P. P. L.
(
2009
).
Requirement for Twist1 in frontonasal and skull vault development in the mouse embryo
.
Dev. Biol.
331
,
176
-
188
.
Blentic
,
A.
,
Tandon
,
P.
,
Payton
,
S.
,
Walshe
,
J.
,
Carney
,
T.
,
Kelsh
,
R. N.
,
Mason
,
I.
and
Graham
,
A.
(
2008
).
The emergence of ectomesenchyme
.
Dev. Dyn.
237
,
592
-
601
.
Bosch
,
D. G. M.
,
Boonstra
,
F. N.
,
Gonzaga-Jauregui
,
C.
,
Xu
,
M.
,
de Ligt
,
J.
,
Jhangiani
,
S.
,
Wiszniewski
,
W.
,
Muzny
,
D. M.
,
Yntema
,
H. G.
,
Pfundt
,
R.
et al.
(
2014
).
NR2F1 mutations cause optic atrophy with intellectual disability
.
Am. J. Hum. Genet.
94
,
303
-
309
.
Bray
,
N. L.
,
Pimentel
,
H.
,
Melsted
,
P.
and
Pachter
,
L.
(
2016
).
Near-optimal probabilistic RNA-seq quantification
.
Nat. Biotechnol.
34
,
525
-
527
.
Brown
,
L. A.
,
Rodaway
,
A. R. F.
,
Schilling
,
T. F.
,
Jowett
,
T.
,
Ingham
,
P. W.
,
Patient
,
R. K.
and
Sharrocks
,
A. D.
(
2000
).
Insights into early vasculogenesis revealed by expression of the ETS-domain transcription factor Fli-1 in wild-type and mutant zebrafish embryos
.
Mech. Dev.
90
,
237
-
252
.
Chen
,
C.-A.
,
Bosch
,
D. G. M.
,
Cho
,
M. T.
,
Rosenfeld
,
J. A.
,
Shinawi
,
M.
,
Lewis
,
R. A.
,
Mann
,
J.
,
Jayakar
,
P.
,
Payne
,
K.
,
Walsh
,
L.
et al.
(
2016
).
The expanding clinical phenotype of Bosch-Boonstra-Schaaf optic atrophy syndrome: 20 new cases and possible genotype-phenotype correlations
.
Genet. Med.
18
,
1143
-
1150
.
Coppola
,
U.
and
Waxman
,
J. S.
(
2021
).
Origin and evolutionary landscape of Nr2f transcription factors across Metazoa
.
PLoS ONE
16
,
e0254282
.
Cox
,
S. G.
,
Kim
,
H.
,
Garnett
,
A. T.
,
Medeiros
,
D. M.
,
An
,
W.
and
Crump
,
J. G.
(
2012
).
An essential role of variant histone H3.3 for ectomesenchyme potential of the cranial neural crest
.
PLoS Genet.
8
,
e1002938
.
Das
,
A.
and
Crump
,
J. G.
(
2012
).
Bmps and id2a act upstream of Twist1 to restrict ectomesenchyme potential of the cranial neural crest
.
PLoS Genet.
8
,
e1002710
.
Drerup
,
C. M.
,
Wiora
,
H. M.
,
Topczewski
,
J.
and
Morris
,
J. A.
(
2009
).
Disc1 regulates foxd3 and sox10 expression, affecting neural crest migration and differentiation
.
Development
136
,
2623
-
2632
.
Dutton
,
K. A.
,
Pauliny
,
A.
,
Lopes
,
S. S.
,
Elworthy
,
S.
,
Carney
,
T. J.
,
Rauch
,
J.
,
Geisler
,
R.
,
Haffter
,
P.
and
Kelsh
,
R. N.
(
2001
).
Zebrafish colourless encodes sox10 and specifies non-ectomesenchymal neural crest fates
.
Development
128
,
4113
-
4125
.
El-Brolosy
,
M. A.
,
Kontarakis
,
Z.
,
Rossi
,
A.
,
Kuenne
,
C.
,
Gunther
,
S.
,
Fukuda
,
N.
,
Kikhi
,
K.
,
Boezio
,
G. L. M.
,
Takacs
,
C. M.
,
Lai
,
S. L.
et al.
(
2019
).
Genetic compensation triggered by mutant mRNA degradation
.
Nature
568
,
193
-
197
.
Fabian
,
P.
and
Crump
,
J. G.
(
2022
).
Reassessing the embryonic origin and potential of craniofacial ectomesenchyme
.
Semin. Cell Dev. Biol.
Haffter
,
P.
,
Granato
,
M.
,
Brand
,
M.
,
Mullins
,
M. C.
,
Hammerschmidt
,
M.
,
Kane
,
D. A.
,
Odenthal
,
J.
,
Van Eeden
,
F. J.
,
Jiang
,
Y. J.
,
Heisenberg
,
C. P.
et al.
(
1996
).
The identification of genes with unique and essential functions in the development of the zebrafish, Danio rerio
.
Development
123
,
1
-
36
.
High
,
F. A.
,
Bhayani
,
P.
,
Wilson
,
J. M.
,
Bult
,
C. J.
,
Donahoe
,
P. K.
and
Longoni
,
M.
(
2016
).
De novo frameshift mutation in COUP-TFII (NR2F2) in human congenital diaphragmatic hernia
.
Am. J. Med. Genet. A
170
,
2457
-
2461
.
Kelsh
,
R. N.
and
Eisen
,
J. S.
(
2000
).
The zebrafish colourless gene regulates development of non-ectomesenchymal neural crest derivatives
.
Development
127
,
515
-
525
.
Kimmel
,
C. B.
,
Ballard
,
W. W.
,
Kimmel
,
S. R.
,
Ullmann
,
B.
and
Schilling
,
T. F.
(
1995
).
Stages of embryonic development of the zebrafish
.
Dev. Dyn.
203
,
253
-
310
.
Lawson
,
N. D.
and
Weinstein
,
B. M.
(
2002
).
In vivo imaging of embryonic vascular development using transgenic zebrafish
.
Dev. Biol.
248
,
307
-
318
.
Le Douarin
,
N. M.
and
Teillet
,
M.-A. M.
(
1974
).
Experimental analysis of the migration and differentiation of neuroblasts of the autonomic nervous system and of neurectodermal mesenchymal derivatives, using a biological cell marking technique
.
Dev. Biol.
41
,
162
-
184
.
Lee
,
R. T.
,
Nagai
,
H.
,
Nakaya
,
Y.
,
Sheng
,
G.
,
Trainor
,
P. A.
,
Weston
,
J. A.
and
Thiery
,
J. P.
(
2013
).
Cell delamination in the mesencephalic neural fold and its implication for the origin of ectomesenchyme
.
Development
140
,
4890
-
4902
.
Love
,
M. I.
,
Huber
,
W.
and
Anders
,
S.
(
2014
).
Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2
.
Genome Biol.
15
,
550
.
Martik
,
M. L.
,
Gandhi
,
S.
,
Uy
,
B. R.
,
Gillis
,
J. A.
,
Green
,
S. A.
,
Simoes-Costa
,
M.
and
Bronner
,
M. E.
(
2019
).
Evolution of the new head by gradual acquisition of neural crest regulatory circuits
.
Nature
574
,
675
-
678
.
Mclean
,
C. Y.
,
Bristor
,
D.
,
Hiller
,
M.
,
Clarke
,
S. L.
,
Schaar
,
B. T.
,
Lowe
,
C. B.
,
Wenger
,
A. M.
and
Bejerano
,
G.
(
2010
).
GREAT improves functional interpretation of cis-regulatory regions
.
Nat. Biotechnol.
28
,
495
-
501
.
Nair
,
S.
,
Li
,
W.
,
Cornell
,
R.
and
Schilling
,
T. F.
(
2007
).
Requirements for Endothelin type-A receptors and Endothelin-1 signaling in the facial ectoderm for the patterning of skeletogenic neural crest cells in zebrafish
.
Development
134
,
335
-
345
.
Nakamura
,
H.
and
Ayer-Le Lievre
,
C. S.
(
1982
).
Mesectodermal capabilities of the trunk neural crest of birds
.
J. Embryol. Exp. Morphol.
70
,
1
-
18
.
Paudel
,
S.
,
Gjorcheska
,
S.
,
Bump
,
P.
and
Barske
,
L.
(
2022
).
Patterning of cartilaginous condensations in the developing facial skeleton
.
Dev. Biol.
486
,
44
-
55
.
Poot
,
M.
,
Eleveld
,
M. J.
,
Van ‘T Slot
,
R.
,
Van Genderen
,
M. M.
,
Verrijn Stuart
,
A. A.
,
Hochstenbach
,
R.
and
Beemer
,
F. A.
(
2007
).
Proportional growth failure and oculocutaneous albinism in a girl with a 6.87 Mb deletion of region 15q26.2→qter
.
Eur. J. Med. Genet.
50
,
432
-
440
.
Rada-Iglesias
,
A.
,
Bajpai
,
R.
,
Prescott
,
S.
,
Brugmann
,
S. A.
,
Swigut
,
T.
and
Wysocka
,
J.
(
2012
).
Epigenomic annotation of enhancers predicts transcriptional regulators of human neural crest
.
Cell Stem Cell
11
,
633
-
648
.
Rossi
,
A.
,
Kontarakis
,
Z.
,
Gerri
,
C.
,
Nolte
,
H.
,
Hölper
,
S.
,
Krüger
,
M.
and
Stainier
,
D. Y. R.
(
2015
).
Genetic compensation induced by deleterious mutations but not gene knockdowns
.
Nature
524
,
230
-
233
.
Schaaf
,
M. J. M.
(
2017
).
Nuclear receptor research in zebrafish
.
J. Mol. Endocrinol.
59
,
R65
-
R76
.
Simoes-Costa
,
M.
and
Bronner
,
M. E.
(
2016
).
Reprogramming of avian neural crest axial identity and cell fate
.
Science
352
,
1570
-
1573
.
Soldatov
,
R.
,
Kaucka
,
M.
,
Kastriti
,
M. E.
,
Petersen
,
J.
,
Chontorotzea
,
T.
,
Englmaier
,
L.
,
Akkuratova
,
N.
,
Yang
,
Y.
,
Haring
,
M.
,
Dyachuk
,
V.
et al.
(
2019
).
Spatiotemporal structure of cell fate decisions in murine neural crest
.
Science
364
,
eaas9536
.
Tatarakis
,
D.
,
Cang
,
Z.
,
Wu
,
X.
,
Sharma
,
P. P.
,
Karikomi
,
M.
,
Maclean
,
A. L.
,
Nie
,
Q.
and
Schilling
,
T. F.
(
2021
).
Single-cell transcriptomic analysis of zebrafish cranial neural crest reveals spatiotemporal regulation of lineage decisions during development
.
Cell Rep.
37
,
110140
.
Teng
,
C. S.
,
Ting
,
M.-C.
,
Farmer
,
D. T.
,
Brockop
,
M.
,
Maxson
,
R. E.
and
Crump
,
J. G.
(
2018
).
Altered bone growth dynamics prefigure craniosynostosis in a zebrafish model of Saethre-Chotzen syndrome
.
eLife
7
,
e37024
.
Upadia
,
J.
,
Gonzales
,
P. R.
and
Robin
,
N. H.
(
2018
).
Novel de novo pathogenic variant in the NR2F2 gene in a boy with congenital heart defect and dysmorphic features
.
Am. J. Med. Genet. A
176
,
1423
-
1426
.
Walker
,
M. B.
and
Kimmel
,
C. B.
(
2007
).
A two-color acid-free cartilage and bone stain for zebrafish larvae
.
Biotech. Histochem.
82
,
23
-
28
.
Walker
,
M. B.
,
Miller
,
C. T.
,
Coffin Talbot
,
J.
,
Stock
,
D. W.
and
Kimmel
,
C. B.
(
2006
).
Zebrafish furin mutants reveal intricacies in regulating Endothelin1 signaling in craniofacial patterning
.
Dev. Biol.
295
,
194
-
205
.
Warnecke
,
M.
,
Oster
,
H.
,
Revelli
,
J.-P.
,
Alvarez-Bolado
,
G.
and
Eichele
,
G.
(
2005
).
Abnormal development of the locus coeruleus in Ear2(Nr2f6)-deficient mice impairs the functionality of the forebrain clock and affects nociception
.
Genes Dev.
19
,
614
-
625
.
Westerfield
,
M.
(
2007
).
THE ZEBRAFISH BOOK: A Guide for the Laboratory use of Zebrafish (Danio rerio)
.
Eugene, OR
:
University of Oregon Press
.
Weston
,
J. A.
and
Thiery
,
J. P.
(
2015
).
Pentimento: Neural Crest and the origin of mesectoderm
.
Dev. Biol.
401
,
37
-
61
.
Yan
,
Y.-L.
,
Willoughby
,
J.
,
Liu
,
D.
,
Crump
,
J. G.
,
Wilson
,
C.
,
Miller
,
C. T.
,
Singer
,
A.
,
Kimmel
,
C.
,
Westerfield
,
M.
and
Postlethwait
,
J. H.
(
2005
).
A pair of Sox: distinct and overlapping functions of zebrafish sox9 co-orthologs in craniofacial and pectoral fin development
.
Development
132
,
1069
-
1083
.
Zalc
,
A.
,
Sinha
,
R.
,
Gulati
,
G. S.
,
Wesche
,
D. J.
,
Daszczuk
,
P.
,
Swigut
,
T.
,
Weissman
,
I. L.
and
Wysocka
,
J.
(
2021
).
Reactivation of the pluripotency program precedes formation of the cranial neural crest
.
Science
371
,
eabb4776
.
Zuniga
,
E.
,
Stellabotte
,
F.
and
Crump
,
J. G.
(
2010
).
Jagged-Notch signaling ensures dorsal skeletal identity in the vertebrate face
.
Development
137
,
1843
-
1852
.
Zuniga
,
E.
,
Rippen
,
M.
,
Alexander
,
C.
,
Schilling
,
T. F.
and
Crump
,
J. G.
(
2011
).
Gremlin 2 regulates distinct roles of BMP and Endothelin 1 signaling in dorsoventral patterning of the facial skeleton
.
Development
138
,
5147
-
5156
.

Competing interests

The authors declare no competing or financial interests.

Supplementary information