s294 Is a Hypomorphic Allele of smo with a C125Y Substitution in the Extracellular Domain

(A–C) Comparison of wild-type (WT) (A), smo^s294 mutant (B), and smo^hi1640 mutant (C) embryos at 24 hr postfertilization (hpf) shows that s294 mutants have phenotypes similar to, although weaker than, the smo null allele hi1640.

(D–F) Immunostaining for Prox1 (green) and slow myosin heavy chain (sMHC) (red) assesses medium- to low-level Hedgehog (Hh) signaling in WT (D), smo^s294 mutant (E), and smo^hi1640 mutant (F) embryos at 24 hpf.

(G–I) Immunostaining for Engrailed (red) assesses high-level Hh signaling in WT (G), smo^s294 mutant (H), and smo^hi1640 mutant (I) embryos at 24 hpf.

(J) Graphic representation of Smoothened (Smo). Each amino acid is represented by a circle; conserved cysteine residues in the extracellular domain (ECD) are shown in green. The s294 mutation is shown in red; similar Drosophila cysteine substitution alleles are shown in purple.

The Extracellular Domain of Smo Is Required for High-Level Signaling

(A–L) Purmorphamine treatment of WT and smo^s294 mutant embryos. DMSO-treated WT (A, E, and I), DMSO-treated smo^s294 mutant (C, G, and K), purmorphamine-treated WT (B, F, and J), and purmorphamine-treated smo^s294 mutant (D, H, and L) embryos at 24 hpf are shown. Purmorphamine treatment did not significantly affect morphology (B) and caused only a modest increase in Prox1 (green) and sMHC staining (red) (F) but strongly induced ectopic Engrailed (Eng)-positive muscle pioneer cells (MPs) in WT embryos (J). Purmorphamine treatment of smo^s294 mutants resulted in significant rescue of the morphological phenotype (D) and restored Prox1 (green) and sMHC (red) expression (H), but not Eng expression (L).

(M–T) Injection of SmoΔCRD mRNA in WT and smo^hi1640 mutant embryos. Prox1 (green) and sMHC (red) staining (M–P) and Eng staining (Q–T) are shown. Injection of 250 pg SmoΔCRD mRNA caused a complete rescue of Prox1 and sMHC expression in smo^hi1640 mutants (P) compared to uninjected mutants (O) and a slight increase in Prox1 and sMHC expression in WT (N) compared to control (M). In contrast, injection of SmoΔCRD mRNA had no effect on Eng expression in WT (R) compared to control (Q) and did not rescue Eng expression in smo^hi1640 mutants (T). For quantification, see Tables S1 and S2.

SmoC151Y Does Not Localize to the Cilium

Expression of the ciliary marker acetylated tubulin (green) and either Myc-tagged WT Smo or Myc-tagged SmoC151Y (red) in NIH 3T3 cells (A–D) and zebrafish embryos at 10 hpf (E–J). Nuclei of NIH 3T3 cells were visualized with DAPI (blue).

(A and B) In NIH 3T3 cells, WT Smo localized to the cilium in response to Hh (A), whereas SmoC151Y did not (B).

(C and D) Treatment with the Smo agonist purmorphamine induced ciliary localization of WT Smo (C) but did not induce detectable ciliary localization of SmoC151Y (D).

(E–I) Myc-tagged Smo mRNA was injected into Tg(-1.8gsc:GFP)^ml1 zebrafish embryos, which express GFP in the dorsal midline. WT Smo localized to the cilia of cells surrounding the dorsal midline in embryos treated with DMSO (E), whereas SmoC151Y was not detected on the cilium (F). Purmorphamine treatment increased the ciliary localization of Smo (G) but did not induce the ciliary localization of SmoC151Y (H). In contrast, SmoΔCRD localized to the cilia in untreated embryos (I). Scale bar in (E) represents 10 μm.

(J) The ciliary localization of Smo was quantified in locations away from the dorsal midline in sections (upper right), and the mean percentages of cilia exhibiting colocalization with the Myc-tagged Smo constructs are shown. Error bars indicate SEM. Total numbers of sections counted are shown below.

Inhibition of Ciliogenesis Causes Loss of Hh Target Gene Expression in Zebrafish Embryos

(A–E) Expression of the ciliary marker acetylated tubulin (green) and the basal body marker γ-tubulin (red) in Kupffer′s vesicles of 16 hpf zebrafish embryos. Injection of dnKif3b mRNA (B), qilin morpholino 1 (MO1) (C), or qilin MO2 (D) caused a reduction of KV cilia, which were restored in embryos coinjected with qilin MO2 and qilin mRNA (E). Scale bar in (A) represents 10 μm.

(F–O) In situ hybridization of 10 hpf control and injected embryos.

(F–M) In WT embryos, ptc1 (F) and myod (K) are expressed in two stripes flanking the dorsal midline. Injection of dnKif3b (G), qilin MO1 (H), or qilin MO2 (I and L) caused a reduction or loss of ptc1 and myod expression, which could be restored by coinjection of qilin MO2 with qilin mRNA (J and M).

(N and O) sonic hedgehog (shh) expression in qilin MO2-injected embryos (O) was comparable to controls (N), with a broadened expression reflecting convergence-extension defects. For quantification of these experiments, see Figure S10.

Purmorphamine stabilizes, but cannot induce full activity of SmoC151Y.
(A-B) Western blots showing that purmorphamine treatment increased SmoC151Y levels from less than that of wt Smo to levels similar to, or above, wt Smo in untreated cells. (A) Zebrafish embryos from a smo^hi1640+/- heterozygote incross were injected with mRNA encoding Myc-tagged wt Smo or SmoC151Y, and treated with DMSO or purmorphamine. (B) Smo^-/- MEFs were transfected with wt Smo or SmoC151Y and treated with DMSO or purmorphamine as indicated. (C) Whereas wt Smo caused a robust activation of the Gli-luciferase reporter in response to either Shh or purmorphamine treatment, SmoC151Y did not display full activity in Smo^-/- MEFs.

SmoC151Y fails to localize to the cilium.
Expression of the ciliary marker acetylated Tubulin (green) and Myc-tagged (red) wild-type Smo (A), Myc-tagged SmoC151Y (B) or CLDSmo (C-D) in NIH 3T3 cells. Nuclei were visualized with DAPI (blue). Low levels of wild-type Smo expression did not cause its ciliary localization in the absence of Shh or purmorphamine; SmoC151Y or CLDSmo was not detected on the cilia even in the presence of Shh or purmorphamine (see also Suppl. Fig. 5).

CLDSmo and CLDSmoΔCRD do not localize to cilia in zebrafish embryos.
Expression of the ciliary marker acetylated Tubulin (green) and Myc-tagged (red) CLDSmo (A) or CLDSmoΔCRD (B) in the dorsal midline in cross-sections of 10 hpf zebrafish embryos. Nuclei were visualized with Hoechst (blue). Like in NIH-3T3 cells, CLDSmo was never detected on the cilium in zebrafish embryos (A). Similarly, the CLD mutations abolished the ciliary localization of SmoΔCRD in zebrafish embryos (B).

Areas of scanning.
A cross-section of a 10hpf zebrafish Tg(-1.8gsc:GFP)^ml1 embryo, labeled with anti-acetylated Tubulin and anti-GFP antibodies (both green). The scale bar is 100 μm. To assess ciliary localization of Smo at different dorso-ventral positions, we scanned three 90 μm areas of each cross-section, as shown. We also examined Smo localization on the ventral side, 180° from the dorsal midline.

The ciliary localization of SmoΔCRD is independent of Hh pathway activation.
Expression of acetylated Tubulin (green) and Myc-tagged SmoΔCRD (red) in NIH-3T3 cells (A-C). Nuclei were visualized with DAPI (blue). SmoΔCRD localized to the cilium in the absence of pathway activation (A), as well as in the presence of Shh (B) or purmorphamine (C). (D) Quantification of ciliary localization of wild-type Smo, SmoΔCRD and SmoC151Y in NIH-3T3 cells. For quantification, all transfected cells were assessed, independent of expression level. High level expression of wild-type Smo was observed to be sufficient to promote ciliary localization.

Morphological phenotypes in cilia knockdown embryos.
Injection of either dnKif3b mRNA (B-D), qilin MO1 (E-G) or qilin MO2 (H-J) gave similar morphological phenotypes in 24 hpf embryos. (A) shows an uninjected control embryo. 3 broad classes of phenotypes were found: class I embryos (B, E, H) resemble uninjected control embryos, class II embryos (C,F,I) have shortened body axes, and partial or full cyclopia, class III embryos (D, G, J) have severe gastrulation defects. The majority of embryos (>80% in all injections) in this latter class did not survive long after gastrulation (K) shows the percentage of embryos in each class for each injection experiment. Co-injection of qilin MO2 with qilin mRNA gave a significant rescue of the severity of phenotypes. The total number of embryos injected and counted for each injection is given in parenthesis, and is the result of a minimum of 3 independent experiments.

Convergence-extension and dorsal specification defects in cilia knock-down embryos
In situ hybridization for convergence-extension (A-D) and dorsal specification (E-H) marker genes at gastrula stages. Injection of either dnKif3b mRNA, qilin MO1 or qilin MO2 caused convergence-extension defects, suggesting perturbation of Wnt/Planar Cell Polarity signaling, and expansion of chd but not gsc expression, suggesting an upregulation of canonical Wnt signaling. Shown are controls and embryos injected with qilin MO1. All injections gave similar phenotypes. (A-D) in situ hybridization for ntl (A-B) and papc (C-D) at 10 hpf, showing convergence-extension defects. (E-H) in situ hybridization for chd (E-F) and gsc (G-H) at 6hpf. chd expression was expanded in qilin MO1 injected embryos (F) compared to controls (E), whereas gsc expression (H) was comparable to controls (G).

Laterality defects in cilia knock-down embryos.
In situ hybridization for myocardial marker cmlc2 at 28hpf. In uninjected control embryos, 100% of hearts jogged to the left (A). In 33 – 65% of embryos injected with dnKif3b mRNA, qilin MO1 or qilin MO2, the heart jogged to the right (B) or remained in the midline (C). The table shows the percentage and total number of embryos counted showing left-sided hearts and midline or right-sided hearts for uninjected control embryos and embryos injected with either 200 pg dnKif3b mRNA, 5 ng qilin MO1, or 15ng qilin MO2.

Variable loss of Hh target gene expression in later stage qilin morphant embryos.
Knock-down of cilia by injection of qilin MO1 caused a loss of Hh target gene expression at 18hpf. (A-B) Floorplate expression of GFP under the control of the nkx2.2a promoter in Tg(nkx2.2a:GFP)^vu16 embryos is dependent on high level Hh signaling. In controls, 30/30 embryos showed GFP expression (A), whereas in 4/15 embryos injected with qilin MO1 (B), GFP expression was absent. Severe or complete knock-down of cilia affected gastrulation movements, and such embryos showed a high rate of early lethality (prior to 18 hpf), providing a possible explanation for the low percentage of embryos that showed a loss of GFP expression in these experiments.