Report

Phosphorylation of Dishevelled by Protein Kinase RIPK4 Regulates Wnt Signaling

See allHide authors and affiliations

Science  22 Mar 2013:
Vol. 339, Issue 6126, pp. 1441-1445
DOI: 10.1126/science.1232253

Three Tales of Wnt Signaling

The Wnt signaling pathway has important roles in regulating many biological processes during development and is also implicated in the behavior of some cancer cells (see the Perspective by Berndt and Moon). Cruciat et al. (p. 1436, published online 14 February) describe the mechanism of action of a protein found in a screen for proteins that influence Wnt signaling. DDX3, a DEAD-box RNA helicase, is required for proper Wnt signaling in Xenopus and Caenorhabditis elegans. It appears to act not through its action as an RNA helicase or through adenosine triphosphate binding, but rather by interacting with the protein kinase, casein kinase 1, and promoting its activation. Huang et al. (p. 1441, published online 31 January) investigated the function of receptor-interacting protein kinase 4 (RIPK4), the product a gene whose mutation causes severe developmental defects in mice and humans. Over-expression of the protein in cultured human cells activated transcription of genes regulated by the Wnt signaling pathway, and loss of RIPK4 function inhibited Wnt signaling in Xenopus embryos. At the molecular level, RIPK4 interacted with the Wnt co-receptor LRP6 and the Wnt signaling adaptor protein DVL2 and promoted phosphorylation of DVL2. Habib et al. (p. 1445) used Wnt-immobilized beads to understand how external cues direct asymmetrical stem cell divisions. Spatially restricted Wnt signals oriented the plane of mitotic division and lead to pluripotency gene expression in the Wnt-proximal daughter cell while the more distal daughter cell acquired hallmarks of differentiation. Thus, asymmetric gene expression patterns can arise as a consequence of orientation by a short-range signal.

Abstract

Receptor-interacting protein kinase 4 (RIPK4) is required for epidermal differentiation and is mutated in Bartsocas-Papas syndrome. RIPK4 binds to protein kinase C, but its signaling mechanisms are largely unknown. Ectopic RIPK4, but not catalytically inactive or Bartsocas-Papas RIPK4 mutants, induced accumulation of cytosolic β-catenin and a transcriptional program similar to that caused by Wnt3a. In Xenopus embryos, Ripk4 synergized with coexpressed Xwnt8, whereas Ripk4 morpholinos or catalytic inactive Ripk4 antagonized Wnt signaling. RIPK4 interacted constitutively with the adaptor protein DVL2 and, after Wnt3a stimulation, with the co-receptor LRP6. Phosphorylation of DVL2 by RIPK4 favored canonical Wnt signaling. Wnt-dependent growth of xenografted human tumor cells was suppressed by RIPK4 knockdown, suggesting that RIPK4 overexpression may contribute to the growth of certain tumor types.

Mice lacking receptor-interacting protein kinase 4 (RIPK4) die at birth with fused external orifices that result from defective epidermal differentiation (1). In humans, RIPK4 mutations cause autosomal recessive Bartsocas-Papas syndrome, which is characterized by severe defects in face, skin, and limb development (2, 3). To investigate RIPK4 signaling, we used a PathwayFinder (QIAGEN, Hilden, Germany) polymerase chain reaction (PCR) array to determine gene expression changes in human embryonic kidney 293T cells after transfection with RIPK4. Up-regulation of Wnt target genes such as CCND1, LEF1, JUN, Myc, and TCF7 (fig. S1) prompted us to compare RIPK4 transfection to Wnt3a treatment in human PA-1 teratocarcinoma cells. RIPK4, but not the related kinases RIPK1, RIPK2, and RIPK3, caused transcriptional changes similar to those caused by Wnt3a (Fig. 1A), up-regulating AXIN2, APCDD1, GAD1, and NKX1-2 (fig. S2). RIPK4 also activated a Wnt-dependent TOPbrite (4) luciferase reporter in 293 cells (Fig. 1B).

Fig. 1

Stimulation of the canonical Wnt pathway by RIPK4. (A) Heat maps show gene expression patterns determined by RNA sequencing in PA-1 cells transfected with empty vector, RIPK1, RIPK2, RIPK3, or RIPK4 for 48 hours or treated with 200 ng/ml Wnt3a for 16 hours. (Left) Hierarchical clustering of variance-stabilized expression values for genes with substantial changes in the treatment groups compared with the vector group. (Right) Pearson’s correlation coefficients between samples. (B) Expression of a TOPbrite luciferase reporter in 293 cells transfected with the constructs indicated and then cultured in the absence or presence of Wnt3a for 5 hours. Error bars represent the SEM of triplicate measurements. Results are representative of three independent experiments. (C) Cytosolic β-catenin protein abundance (top) and CTNNB1 gene expression (bottom) in 293T cells transfected with RIPK4. Error bars represent the SEM of triplicate measurements. (D) Cytosolic β-catenin in 293T cells transfected with Bartsocas-Papas syndrome RIPK4 mutants. (E) Cytosolic β-catenin in A2780 or COV434 cells transfected with control (Ctrl) or RIPK4 siRNAs for 60 hours and then treated with vehicle or Wnt3a for 2 hours.

Wnt signaling stabilizes β-catenin in the cytosol, thereby facilitating its interaction with TCF transcription factors to drive Wnt-dependent gene expression (4). 293T cells overexpressing RIPK4 contained more cytosolic β-catenin than did cells transfected with empty vector, RIPK1, RIPK2, or RIPK3, but they did not contain more CTNNB1 mRNA (Fig. 1C). Therefore, RIPK4 may inhibit β-catenin protein degradation in a similar manner to Wnt signaling. The kinase activity of RIPK4 appeared to be critical for its Wnt-like effects because catalytically inactive mutant RIPK4-K51R (K51R: Lys51→Arg51 ) neither activated the TOPbrite reporter nor altered cytosolic β-catenin levels (fig. S3). RIPK4 interaction with protein kinase C δ (PKC-δ) or PKC-β was neither sufficient nor required for increased cytosolic β-catenin because the K51R mutation did not prevent PKC binding (fig. S4A), and knockdown of both PKC isoforms had no effect on β-catenin accumulation by wild-type (WT) RIPK4 (fig. S4B). Bartsocas-Papas syndrome RIPK4 point mutants (I81N and I121N; I, Ile; N, Asn) and a truncation mutant (S376X; S, Ser; X, stop codon.) did not alter the amount of cytosolic β-catenin (Fig. 1D), implying that RIPK4 signaling to β-catenin may be relevant to mammalian development.

We investigated whether RIPK4 was required for Wnt signaling in various cell lines. RIPK4 knockdown in 293T, ovarian A2780, and ovarian COV434 cells reduced Wnt3a-induced accumulation of β-catenin, TOPbrite luciferase activation, and transcription of AXIN2 and APCDD1 (Fig. 1E and fig. S5), whereas Wnt3a signaling in pancreatic PANC-1, kidney 786-O, and breast Hs578T cells was not compromised by RIPK4 knockdown (fig. S6). Therefore, the contribution of RIPK4 to Wnt signaling appears to be context-dependent.

To explore a role for RIPK4 during development, we overexpressed Xenopus laevis Ripk4 in the ventral-vegetal cells of Xenopus embryos, where ectopic expression of Wnt pathway agonists causes axis duplication (5). Ripk4 alone did not produce a secondary axis, but it synergized with subthreshold amounts of Xenopus Wnt8 (Xwnt8), such that a full secondary axis developed in ~25% of embryos (Fig. 2A). Ripk4 overexpression in dorsal-marginal cells, like expression of Wnt agonists (6), broadened the region expressing the dorsal organizer gene Chordin (Fig. 2B). Catalytically inactive Ripk4-K52R had no effect (Fig. 2B). To test whether Ripk4 was necessary for Wnt signaling in vivo, we suppressed Ripk4 expression in Xenopus embryos with translation-blocking morpholinos and measured expression of Wnt target gene Xnr3 in animal cap cells (7). Xnr3 induction by Xwnt8 was reduced (Fig. 2C and fig. S7), suggesting that RIPK4 is necessary for optimal Wnt pathway activation in this region. Absence of Xbra expression in the animal cap samples excluded the possibility that Xnr3 derived from contaminating mesodermal tissue. Ripk4 morpholinos also anteriorized the neural tube and caused a posterior shift in the hindbrain marker gene Engrailed2 (Fig. 2D), a phenotype linked to reduced Wnt signaling (8). Two different Ripk4 morpholinos caused this phenotype, which was reversed by injection of human RIPK4 (Fig. 2D, column 3), ruling out nonspecific, off-target effects of the morpholinos. Ripk4-K52R overexpression in dorsal-anterior cells elicited similar changes to those caused by Ripk4 morpholinos (Fig. 2D, column 4), consistent with the kinase inactive mutant interfering with signaling by endogenous Ripk4. Finally, Ripk4 or Xwnt8 overexpression had the opposite effect of Ripk4 depletion, expanding and shifting Engrailed2 expression anteriorly (Fig. 2D, columns 5 and 6). Collectively, these experiments indicate that Ripk4 can regulate Wnt signaling in vivo.

Fig. 2

Modulation of Wnt signaling in Xenopus by Ripk4. (A) Secondary axis formation in Xenopus embryos injected ventral vegetally with Ripk4 (1 ng), Xwnt8 (1 pg), or both. Two hundred sixty-nine embryos were examined in three independent experiments. Representative embryos are shown. (B) Length of Chordin expression along the blastopore lip of embryos injected at the four-cell stage with the RNAs indicated. Error bars represent the mean ± SEM of 108 embryos. Vegetal views of representative embryos are shown. au, arbitrary units. (C) Reverse transcriptase (RT) PCR analyses show the effect of Ripk4 morpholinos on Xwnt8-induced expression of Xnr3 in Xenopus animal cap cells. Control reactions contained RNA from stage 10.5 Xenopus embryos with or without RT. Amplification of epidermal keratin and ornithine decarboxylase (ODC) confirmed RNA integrity. (D) Engrailed2 expression (blue) after blastomere injection at the two-cell stage (Ripk4-MO1, Ripk4-MO2, sorted by fluorescence of a coinjected fluorescein tracer) or the four-cell stage (all others). Right, injected side; top, anterior. Red staining is due to a nuclear β-galactosidase tracer.

To investigate how RIPK4 regulates Wnt signaling, we monitored RIPK4-induced accumulation of cytosolic β-catenin and activation of a TOPbrite reporter gene after small interfering RNA (siRNA)–mediated depletion of Wnt pathway components. Loss of either the Wnt co-receptor LRP6 or the Dishevelled adaptor proteins (DVL1 to 3) blocked the increase in cytoplasmic β-catenin (fig. S8A) and decreased TOPbrite luciferase activity (fig. S8B). Endogenous RIPK4 coimmunoprecipitated with endogenous DVL2 from 293T cells, irrespective of Wnt3a treatment (Fig. 3A). A direct interaction seems likely, as in vitro translated DVL2 and RIPK4 interacted as well (fig. S8C). Wnt3a treatment induced interaction of LRP6 with RIPK4, albeit not until 15 min after treatment (Fig. 3B). These findings are consistent with RIPK4 acting at the level of the Wnt receptor complex.

Fig. 3

Phosphorylation of DVL by RIPK4 promotes canonical Wnt signaling. (A) In 293T cells, endogenous RIPK4 coimmunoprecipitated with endogenous DVL2 in the presence or absence of 200 ng/ml Wnt3a for 30 min. Control (Ctrl) and RIPK4 siRNAs confirmed RIPK4 antibody specificity. WCL, whole-cell lysates; IP, immunoprecipitation. (B) Endogenous RIPK4 coimmunoprecipitated with endogenous LRP6 in 293T cells treated with Wnt3a. (C) In vitro kinase assays using the RIPK4 kinase domain and the DVL2 DEP (or PDZ) domain as a substrate. M.W., molecular weight. (D) RIPK4-dependent phoshorylation of endogenous DVL2 in 293T cells treated with Wnt3a for 10 min. Asterisks indicate nonspecific bands. (E) Effect of RIPK4-GFP on DVL2-FLAG cellular distribution in HeLa cells. Cells containing DVL2 puncta were enumerated by counting 250 cells per condition. Scale bars, 10 μm.

We explored whether RIPK4 phosphorylated LRP6, DVL, or their associated proteins. Finding no evidence of LRP6 phosphorylation, we focused on the DVL proteins as potential RIPK4 substrates. When DVL2 was coexpressed in 293T cells with RIPK4 or K51R catalytic inactive mutant RIPK4, two phosphopeptides, GDGGIYIGS298IMK and KYAS480GLLK (G, Gly; D, Asp; Y, Tyr; M, Met; A, Ala; L, Leu), derived from the PDZ and DEP domains of DVL2, respectively, were enriched in cells expressing WT RIPK4 (fig. S9A). Both sites are conserved, being found in Xenopus, zebrafish, mouse, and human DVL isoforms (fig. S9B).

To determine if RIPK4 phosphorylated DVL2 directly, we purified the kinase domain of RIPK4 (amino acids 1 to 300) from Sf9 insect cells and tested its ability to phosphorylate the PDZ or DEP domains of DVL2 in vitro. The WT RIPK4 kinase domain, but not the K51R mutant, phosphorylated both domains, and mutation of DVL2 Ser298 and Ser480 prevented this phosphorylation (Fig. 3C). Antibodies recognizing each DVL2 phosphorylation site detected WT DVL2 overexpressed in 293T cells, but not DVL2 with the relevant serine residue mutated to alanine (fig. S9C). Further confirming the specificity of the antibodies, the WT DVL2 bands were not detected when the lysates were treated with calf intestinal alkaline phosphatase (fig. S9D). In keeping with DVL2 being a RIPK4 substrate, phosphorylation at Ser298 and Ser480 increased dramatically when DVL2 was cotransfected with RIPK4, but not RIPK4-K51R (fig. S9D).

When 293T cells were stimulated with Wnt3a, phosphorylation of DVL2 at Ser298 and Ser480 increased transiently after 10 min (fig. S9E); this was attenuated by RIPK4 depletion (Fig. 3D and fig. S9E). To determine if DVL2 phosphorylation was necessary for Wnt3a signaling, we depleted Dvl3 from Dvl1−/−Dvl2−/− mouse embryo fibroblasts to obtain DVL-null cells that were reconstituted with either WT or phospho-site mutant S298A/S480A DVL2. Cells expressing DVL2 S298A/S480A contained less cytosolic β-catenin after Wnt3a treatment than did cells expressing WT DVL2 (fig. S9F), suggesting that phosphorylation at one or both sites was necessary for maximal Wnt3a signaling.

Wnt3a treatment redistributes cytoplasmic DVL proteins into large signaling complexes detected as punctate structures by immunofluorescence microscopy (9). Approximately 20% of HeLa cells expressed DVL2-FLAG in puncta, whereas the rest contained cytoplasmic DVL2-FLAG (Fig. 3E). Cotransfection of RIPK4–green fluorescent protein (GFP) increased the percentage of cells with DVL2 puncta to more than 75% (Fig. 3E and fig. S10), suggesting that RIPK4 facilitates assembly of the DVL2 signalosome. DVL2 S298A/S480A did not form more puncta when coexpressed with RIPK4, whereas DVL2 containing a single Ser-to-Ala mutation behaved like WT DVL2 (Fig. 3E). These data indicate that RIPK4 phosphorylation at Ser298 and Ser480 promotes DVL2 signalosome assembly.

Mutations that enhance Wnt signaling are found in various human cancers (10), so we examined human tumors for overexpression of RIPK4. Microarray data revealed increased RIPK4 mRNA in many ovarian, skin, and colorectal tumors (Fig. 4A). We detected RIPK4 protein and cytosolic β-catenin in several human ovarian adenocarcinomas; by contrast, RIPK4 was less abundant, and we did not detect cytosolic β-catenin in most noncancerous ovarian tissue samples (Fig. 4B and fig. S11A). To determine if RIPK4 contributes to tumor growth by enhancing Wnt signaling, we used a RIPK4 short hairpin RNA (shRNA) to deplete RIPK4 from the Wnt-dependent human teratoma–derived NTERA-2 xenograft tumor model (11). We used colon HCT116 cells with an activating β-catenin mutation (12) as a control. Although RIPK4 depletion from the NTERA-2 and HCT116 cells was incomplete (fig. S11B), growth of the NTERA-2 cells in athymic nude mice was suppressed (Fig. 4, C and D). By contrast, RIPK4 depletion had no effect on HCT116 tumor growth. RIPK4 depletion from NTERA-2 tumors, but not HCT116 tumors, decreased expression of two Wnt-responsive genes, GAD1 and AXIN2 (fig. S11C), implying a critical role for RIPK4 in Wnt signaling upstream of β-catenin stabilization.

Fig. 4

Delayed growth of Wnt-dependent xenograft tumors depleted of RIPK4. (A) Box-and-whisker plots show RIPK4 mRNA abundance measured by microarray in human colorectal, ovarian, and melanoma samples. N, noncancerous tissues; C, cancer. P values were derived from Student’s t tests. (B) Western blots show increased RIPK4 and cytosolic β-catenin in human ovarian adenocarcinomas compared to noncancerous ovarian tissue samples. (C and D) NTERA-2 and HCT116 cells were transduced with lentiviral particles encoding RIPK4 or control (Ctrl) shRNAs and injected subcutaneously into athymic nude mice. (C) Representative tumor-bearing mice and their dissected tumors after 47 days are shown. (D) Graphs show tumor volumes; error bars represent the SEM (n = 10 mice).

We have identified a RIPK4 signaling mechanism that might explain why its mutation in mammals causes severe developmental defects. RIPK4 appears to be recruited to the LRP6 co-receptor and phosphorylates DVL proteins after Wnt stimulation, leading to maximal stabilization of β-catenin and transcription of Wnt-responsive genes (fig. S12). Mutant forms of RIPK4 associated with human Bartsocas-Papas syndrome are compromised in this signaling ability. RIPK4 expression is restricted to vertebrates, indicating that phosphorylation of DVL proteins in lower organisms might require a different kinase. Finally, the finding that RIPK4-mediated Wnt signaling may promote human ovarian cancer suggests that small-molecule inhibitors of RIPK4 might hold therapeutic potential.

Supplementary Materials

www.sciencemag.org/cgi/content/full/science.1232253/DC1

Materials and Methods

Figs. S1 to S12

Table S1

References (1417)

References and Notes

  1. Acknowledgments: We thank K. O’Rourke, J. Huang, and the next-generation sequencing and baculovirus groups for technical assistance. The transcriptome sequencing data for PA1 cells have been submitted to the National Center for Biotechnology Information, NIH, Gene Expression Omnibus (www.ncbi.nlm.nih.gov/geo/) under accession no. GSE43362.
View Abstract

Stay Connected to Science

Navigate This Article