Retinoic acid production, regulation and containment through Zic1, Pitx2c and Cyp26c1 control cranial placode specification

Abstract

All paired sensory organs arise from a common precursor domain called the pre-placodal region (PPR). In Xenopus, Zic1 non-cell autonomously regulates PPR formation by activating retinoic acid (RA) production. Here, we have identified two Zic1 targets, the RA catabolizing enzyme Cyp26c1 and the transcription factor Pitx2c, expressed in the vicinity of the PPR as being crucially required for maintaining low RA levels in a spatially restricted, PPR-adjacent domain. Morpholino- or CRISPR/Cas9-mediated Cyp26c1 knockdown abrogated PPR gene expression, yielding defective cranial placodes. Direct measurement of RA levels revealed that this is mediated by a mechanism involving excess RA accumulation. Furthermore, we show that pitx2c is activated by RA and required for Cyp26c1 expression in a domain-specific manner through induction of FGF8. We propose that Zic1 anteriorly establishes a program of RA containment and regulation through activation of Cyp26c1 and Pitx2c that cooperates to promote PPR specification in a spatially restricted domain.

Keywords: Degradation; Patterning; Placode; Retinoic acid; Xenopus.

Fig. 1.

Cyp26c1 is expressed anteriorly and is a target of Zic1. (A) In situ hybridization for cyp26c1 expression at late neurula stage (left panel). The dashed white line indicates the plane of section. Right panel shows the corresponding section (lateral view) stained with Hematoxylin and Eosin. The prospective hindbrain (magenta arrow) and the PPR-adjacent anterior (green arrow) domains are indicated. (B) In situ hybridization for cyp26c1 on control, Zic1GR- and Zic1MO-injected embryos (upper panel). Representative images are shown, with the injected side on the right. (C) Quantification of the phenotypes. The number of embryos analyzed for each condition is on the top of each bar; ****P<0.0001, χ² test. (D) Schematic representation of an animal cap explant assay. (E) RT-PCR analysis of cyp26c1, cyp26a1, six1 and ODC (ornithine decarboxylase 1) expression in Zic1GR-injected animal cap explants. (F,G,H) Developmental qRT-PCR expression profile of cyp26c1 (F), six1 (G) and pitx2c (H). NF stages are indicated on the x-axis, values are normalized to ODC. A representative experiment is shown. (I) In situ hybridization for cyp26c1 and pitx2c expression. NF stages are indicated at the top of each panel. The onset of expression of cyp26c1 and pitx2c in the PPR-adjacent domain is indicated with a red arrow. (J) Double in situ hybridization for zic1 (teal) and cyp26c1 (magenta) (upper panels); and foxi4.1 (teal) and cyp26c1 (magenta) (lower panels). Higher magnifications of stage 15 embryos are shown. (K) Schematic representation of zic1, cyp26c1 and foxi4.1 expression domains. (E) W.E, whole embryo; (J) Lat, lateral view. All images show anterior views with dorsal towards the top, unless otherwise indicated.

Fig. 2.

Measurement of at-RA and 4-oxo RA levels in the embryo. (A) Schematic representation of the metabolic pathway for synthesis and degradation of endogenous RA by RALDH2 and Cyp26c1, respectively. (B,C) Endogenous retinoids, at-RA (B) and 4-oxo-RA (C), in wild-type Xenopus embryo at stages 9, 10.5 and 13, as quantified by LC-MS/MS. NF stages are indicated on the x-axis. Data are for 130 embryos per group and are shown as mean±s.d., n=3 groups per time-point; *P<0.0311 (unpaired t-test).

Fig. 3.

Cyp26c1 knockdown results in a broad loss of anterior structures. (A) Schematic representation of the working model for PPR specification based on previous work (Jaurena et al., 2015). Anterior neural plate (ANP, purple) expresses Zic1, which induces raldh2/aldh1a2 and lpgds/ptgds. Raldh2 converts retinal to RA. A putative refractory region (orange) expressing cyp26c1 lies adjacent, across which RA is transported. At the PPR (blue), RA induces expression of PPR genes six1, eya1 and foxi1c/foxi4.1. (B, top) Schematic representation of the cyp26c1 gene structure showing the target site of the splice-blocking morpholino (MO; green bar) and the target sites of sgRNA SL1 and SL2. (B, bottom) RT-PCR analysis showing intron 4 retention in embryos injected with increasing doses of Cyp26c1 morpholino (Cyp26MO). Control indicates uninjected embryos. ODC is shown as loading control and Xenopus laevis genomic DNA as a positive control for intron 4 detection. (C,D) In situ hybridization for the indicated genes in control and Cyp26MO-injected embryos at stage 15. Injected side (right) showing the lineage tracer Red-Gal. Expression domains of dmrta1 (red arrows), snai2 (red arrowheads) and the placode domain of sox2 (white arrows) are indicated. (E) Quantification of the phenotypes. The number of embryos analyzed is indicated at the top of each bar. P-values were calculated using an unpaired t-test for the major phenotype, **P≤0.01, ***P≤0.001, ****P<0.0001. (F,G) qRT-PCR analysis of six1 (F) and pitx2c (G) expression in Zic1GR and Zic1GR+Cyp26MO-injected animal cap explants cultured in the presence of dexamethasone. A representative experiment is shown, normalized to ODC. (H) In situ hybridization for six1, foxi4.1 and tbx2 expression in tailbud stage control and bilaterally Cyp26MO-injected embryos. Nasal and epibranchial placodes are visualized using six1 and foxi4.1, respectively (white arrows). The lens and otic vesicle are visualized by tbx2 (white arrows). Top panels show anterior views, dorsal towards the top. Middle and bottom panels show lateral views, dorsal towards the top. (I) Quantification of the phenotypes. The number of embryos analyzed is indicated at the top of each bar. ****P<0.0001, χ² test. (J) In situ hybridization for foxi4.1 and six1 expression in embryos injected with Cas9 alone or with Cyp26c1-sgRNA+Cas9. The injected side is indicated by an asterisk. (K) Quantification of the phenotypes. The number of embryos analyzed for each condition is indicated at the top of each bar. ***P≤0.001, unpaired t-test. (L) Schematic representation of eight-cell stage injection. D, dorsal side; V, ventral side. (M) In situ hybridization for six1 and foxi4.1 in control and Cyp26MO-injected embryos at stage 15. Injected side (right) showing the lineage tracer Red-Gal. (N) Quantification of the phenotypes. The number of embryos analyzed is indicated at the top of each bar. ****P<0.0001, χ² test. Quantification from at least two (I) or three (E,K,N) biological replicates. (F,G) Un, uninjected caps.

Fig. 4.

Measurement of RA levels in wild-type and Cyp26MO-injected embryos. Endogenous retinoids in Xenopus embryos of either wild type (WT) or bilaterally injected with Cyp26MO at stages 10.5 and 13. (A) at-RA, (B) 13-cis-RA, (C) retinol and (D) RE. Data are for 25 embryos per group and are shown as mean±s.d., n=4 groups per time-point/genotype; *P<0.0247, unpaired t-test. 4-oxo-RA was below the assay limit of detection for groups of 25 embryos.

Fig. 5.

Cyp26a1 knockdown does not affect cyp26c1 and six1 expression. (A) Schematic representation of the cyp26a1 gene structure, indicating the target site for the splice-blocking morpholino (SMO2; green bar). For representation purposes, the cyp26a1.L form is shown; however, the morpholino targets both forms. The primers used to detect intron exclusion are indicated in red (E1, E2) and in black for intron retention (Int1F, Int1R). (B) qRT-PCR analysis of total RNA from embryos bilaterally injected with increasing doses of Cyp26a1 as indicated. The fold change of intron 1 retention over exclusion (exon 1-exon 2) is shown. Values are normalized to ODC prior to calculation of fold change. (C) In situ hybridization for cyp26c1 and six1 in Cyp26a1SMO2-injected embryos (TexasRed dextran was used as a lineage tracer). The asterisks indicate the injected side. (D) Quantification of the phenotypes in C from at least three biological replicates. The number of embryos analyzed for each condition is indicated at the top of each bar. ns, not significant.

Fig. 6.

RA sensitivity and dose response of PPR-related genes. (A) In situ hybridization for cyp26c1, pitx2c, foxi4.1 and hnf1b expression on stage 15 embryos treated with either 1 μM DMSO with increasing doses of RA as shown. (B) In situ hybridization analysis of stage 22 embryos treated with either 1 μM DMSO or 1 μM RA at stages 10 and 11. For double in situ hybridization (middle and bottom rows), the gene being assessed is indicated with bold letters, pitx2c (middle rows) and hnf1b (bottom rows). Red arrows indicate the position of the cement gland region. The red dotted lines indicate the anterior boundary of hnf1b expression under normal conditions, which expands anteriorly under conditions of excess RA (1 μM, white dashed line). (A,B) The phenotypes are fully penetrant; n>45 embryos per conditions from three biological replicates. (A,B) All images show anterior views with dorsal towards top, except for hnf1b, where dorsal views are presented with anterior towards bottom.

Fig. 7.

Regulation of Cyp26c1 expression by the transcription factor Pitx2c. (A,B) qRT-PCR analysis of pitx2c (A) and six1 (B) expression in animal cap explants injected with Zic1GR mRNA and cultured for 8 h in the presence of dexamethasone with or without disulfiram (DSF; 100 μM). Values are normalized to ODC. A representative experiment is shown. (C) Schematic representation of the pitx2c gene structure, indicating the target site for the translation-blocking (MO1; black arrow) morpholino. (D) Western blot analysis of protein lysates from control embryos (lane 1), Pitx2c-FLAG mRNA-injected embryos (lane 2) and Pitx2c-FLAG mRNA co-injected with 20 ng of PitxMO1 (lane 3). Tubulin is shown as a loading control (bottom panel). (E,G) In situ hybridization for cyp26c1 and six1 expression in PitxMO1- (E) and PitxSMO2-injected embryos (G). Red-Gal (E) or RFP (G) were used as a lineage tracer. The asterisks indicate the injected side. (I,J) qRT-PCR analysis for cyp26c1 (I) and fgf8a (J) expression in animal cap explants injected with Pitx2cGR mRNA and cultured in the presence of dexamethasone. The values are normalized to ODC. cyp26c1 expression was assessed after 8 h in culture, while fgf8a expression was evaluated after 4 h in culture. (F,H) Quantification of the phenotypes from E and G, respectively, from at least three biological replicates. The total number of embryos is indicated at the top of each bar. P-values for the association between morpholino injection and level of gene expression were calculated using an unpaired t-test for the major phenotype, **P≤0.01, ***P≤0.001, ****P<0.0001 in F,H. (E,G) All images show anterior views with dorsal to the top.

Fig. 8.

FGF signaling is required for cyp26c1 expression and PPR formation. (A) qRT-PCR analysis of cyp26c1 expression in animal cap explants injected with Pitx2cGR mRNA and cultured for 8 h in the presence of dexamethasone with DMSO or SU5402 (25 μM). Values are normalized to ODC. A representative experiment is shown. (B) In situ hybridization for pitx2c, cyp26c1 and foxi4.1 on NF stage 15 embryos treated with DMSO or SU5402. (D) In situ hybridization for pitx2c, cyp26c1 and foxi4.1 expression in control NF stage 15 embryos and embryos injected with Fgf8aMO. Injected side with Red-Gal is on the right. (C,E) Quantification of the phenotypes from B and D, respectively, from at least three biological replicates. Total embryos analyzed are indicated at the top of each bar. P values were calculated using an unpaired t-test for the major phenotype (reduced cyp26c1 expression in the anterior domain or reduced foxi4.1 expression), ***P≤0.001, ****P<0.0001; ns, not significant (C,D). (A,B) All images show anterior views with dorsal towards the top.

Fig. 9.

Model for Zic1-regulated PPR formation. (A) Under normal conditions (optimal RA levels), at the end of gastrulation, Zic1 is expressed in the prospective anterior neural plate (orange), where it activates the expression of raldh2/adh1a2. RA synthesized by raldh2/adh1a2 is transported to the adjacent PPR-adjacent region (blue) where it activates pitx2c expression. Subsequently, Pitx2c induces the expression of cyp26c1 in the same domain via activation of Fgf8a. Cyp26c1 in turn maintains RA at low levels for sustained pitx2c expression and elicits the activation of PPR-specific genes. (B) Under conditions of excess RA, upon Cyp26c1 knockdown (left panel) or on exposure to exogenous RA (right panel), pitx2c and pitx2c/cyp26c1 are downregulated, respectively, preventing the modulation of RA to levels necessary for PPR gene activation.