Wnt5a-Ror-Dishevelled signaling constitutes a core developmental pathway that controls tissue morphogenesis

Abstract

Wnts make up a large family of extracellular signaling molecules that play crucial roles in development and disease. A subset of noncanonical Wnts signal independently of the transcription factor β-catenin by a mechanism that regulates key morphogenetic movements during embryogenesis. The best characterized noncanonical Wnt, Wnt5a, has been suggested to signal via a variety of different receptors, including the Ror family of receptor tyrosine kinases, the Ryk receptor tyrosine kinase, and the Frizzled seven-transmembrane receptors. Whether one or several of these receptors mediates the effects of Wnt5a in vivo is not known. Through loss-of-function experiments in mice, we provide conclusive evidence that Ror receptors mediate Wnt5a-dependent processes in vivo and identify Dishevelled phosphorylation as a physiological target of Wnt5a-Ror signaling. The absence of Ror signaling leads to defects that mirror phenotypes observed in Wnt5a null mutant mice, including decreased branching of sympathetic neuron axons and major defects in aspects of embryonic development that are dependent upon morphogenetic movements, such as severe truncation of the caudal axis, the limbs, and facial structures. These findings suggest that Wnt5a-Ror-Dishevelled signaling constitutes a core noncanonical Wnt pathway that is conserved through evolution and is crucial during embryonic development.

Fig. 1.

Generation and characterization of conditional Ror1 and Ror2 mutant mice. (A) Schematic of the Ror1 gene targeting strategy. The Ror1 conditional allele (Ror1^f) was generated by flanking exons 3 and 4 of the Ror1 genomic locus with loxP sequences. The Ror1-null allele (Ror1⁻) was generated by crossing the Ror1^f/f mice to the EIIA-Cre deleter line. (B) Schematic of the Ror2 genomic locus targeting strategy. The Ror2 conditional allele was generated by flanking exons 3 and 4 of the Ror2 gene with lox2272 sequences. Lox2272, which can undergo Cre-mediated recombination with itself, but not with loxP sequences, was used to avoid potential interchromosomal recombination with the targeted Ror1 locus. The Ror2 null allele (Ror2⁻) was generated by crossing the Ror2^f/f mice to the EIIA-Cre deleter line. (C) Immunoblot of Ror1 protein in E12.5 embryo lysates from WT (Ror1^+/+), Ror1^f/f, and Ror1^−/− mice. Protein bands marked by asterisks are proteins unrelated to Ror1 that cross-react with the anti-Ror1-C antibody (Fig. S1). (D) Immunoblot of Ror2 protein in E12.5 embryo lysates from WT (Ror2^+/+), Ror2^f/f, and Ror2^−/− mice.

Fig. 2.

Ror1 and Ror2 double KO embryos exhibit morphogenesis defects. (A–D) Representative images of unfixed E12.5 Ror1^+/−;Ror2^+/− embryo (A), Ror1^−/−;Ror2^−/− embryo (B), Wnt5a^+/− mice (C), and Wnt5a^−/− mice (D). Embryos shown in A and B are littermates. Embryos shown in C and D are littermates. (E and F) Images of a Bouin's fixed E12.5 Ror1^+/−;Ror2^+/− embryo (E, littermate control of F) and Ror DKO embryo with exencephaly (F). Arrows indicate truncated and asymmetric hindlimbs; arrowhead indicates truncation of the posterior body axis. f, facial malformation; ed, edema; ex, exencephaly; fl, forelimb.

Fig. 3.

Ror1 and Ror2 double mutant embryos exhibit sympathetic axon branching defects. In situ RNA hybridization of Ror1 (A) and Ror2 (B) in the SCG of P0.5 mice. (C) Coimmunostaining of the SCG of a P0.5 Ror2^LacZ/+ embryo with anti-TH and anti–β-gal antibodies. Asterisks in A–C denote the carotid arteries used as landmarks during tissue sectioning. (D–I) TH immunostaining in E17.5 control Ror1^f/+;Ror2^f/+ spleen (D), kidney (E), and bladder (F) and littermate E17.5 Ror1^f/f;Ror2^f/f;Wnt1-cre spleen (G), kidney (H), and bladder (I). Arrows denote axonal branches that are normally seen in control target organs but are compromised in mutant target organs. (J and K) Sympathetic chain ganglia of the Ror1^f/f;Ror2^f/f;Wnt1-cre embryos appear grossly intact and show normal coalescence as shown by whole-mount TH staining.

Fig. 4.

Dvl2 phosphorylation is a target of noncanonical Wnt5a-Ror signaling in vitro and in vivo. (A–D) Representative immunoblots showing Dvl2 phosphorylation in protein lysates prepared from MEFs and embryos. (A) Dvl2 protein in lysates from WT MEFs, WT MEF lysates treated with calf intestinal phosphatase (CIP), Wnt5a^−/− MEFs, WT MEFs treated with sFRP-3, WT MEFs treated with DKK-1, and Lrp6^−/− MEFs. (B) Dvl2 protein in lysates from WT MEFs, Ror1^−/− MEFs, Ror2^−/− MEFs, Ror1/2 DKO MEFs, Ror1^+/−;Ror2^+/− MEFs, and Wnt5a^−/− MEFs. (C) Ror1 protein, Ror2 protein, and Dvl2 protein in WT MEFs and Ror1^f/f;Ror2^f/f;ER-Cre MEFs with or without 72 h treatment of 4-OHT. (D) Dvl2 protein in E12.5 embryo lysates from WT mice, Wnt5a^−/− mice, Ror1^−/− mice, Ror2^−/− mice, Ror DKO mice, and Ror1^+/−;Ror2^+/− mice. α-Tubulin was used for loading controls in all experiments. Percent Dvl2 phosphorylation was calculated by dividing the upshifted band by total Dvl2 signal.

Fig. 5.

Rors function as receptors for Wnt5a. (A) Dvl2 protein in lysates from WT MEFs treated for 18 h with native control rabbit IgG at 1:500 or 1:100 dilutions, WT MEFs treated for 18 h with both native anti-Ror1 ECD antibodies and anti-Ror2 ECD antibodies at 1:1,000 or 1:200 dilutions of each antibody, and untreated WT MEFs. The stock concentration of control rabbit IgG, anti-Ror1 ECD, and anti-Ror2 ECD antibodies are all 12 mg/mL (total IgG fraction purified from whole serum). (B) Coomassie G-250 stained gel showing control rabbit IgG, anti-Ror1 ECD antibodies, and anti-Ror2 ECD antibodies with or without pretreatment with papain. (C) Dvl2 protein in lysates from WT MEFs treated for 18 h with papain-cleaved control rabbit IgG at 1:100 or 1:25 dilutions, WT MEFs treated for 18 h with papain-cleaved anti-Ror1 ECD antibodies and papain-cleaved anti-Ror2 ECD antibodies at 1:200 or 1:50 dilutions of each antibody, and untreated WT MEFs. The stock concentration of papain-cleaved control rabbit IgG, anti-Ror1 ECD, and anti-Ror2 ECD antibodies are all 7 mg/mL (cleaved total IgG fraction purified from whole serum).

Fig. 6.

Rors are not required for Wnt5a-dependent inhibition of canonical Wnt signaling or c-Jun phosphorylation, and Wnt5a and Wnt3a both induce c-Jun phosphorylation in MEFs. (A) Wnt5a-induced inhibition of Wnt3a-stimulated β-catenin–responsive luciferase reporter activity in WT MEFs and Ror DKO (Ror1^−/−;Ror2^−/−) MEFs. (B) Immunoblot showing levels of phospho-c-Jun (ser 63) and total c-Jun proteins in lysates from WT MEFs, Wnt5a^−/− MEFs, and Ror DKO (Ror1^−/−;Ror2^−/−) MEFs. Phospho-c-Jun to total c-Jun ratio was calculated by using quantitative Western blotting. (C and D) E12.5 MEFs were stimulated with Wnt3a (C) or Wnt5a (D) at the indicated concentrations. Protein samples were collected at 0, 1, and 3 h after stimulation with Wnt proteins and analyzed by Western blotting using the anti–phospho-c-Jun (S63) antibody. α-Tubulin was used for loading controls.