Casein kinase I phosphorylates and destabilizes the beta-catenin degradation complex

Abstract

Wnt signaling plays a key role in cell proliferation and development. Recently, casein kinase I (CKI) and protein phosphatase 2A (PP2A) have emerged as positive and negative regulators of the Wnt pathway, respectively. However, it is not clear how these two enzymes with opposing functions regulate Wnt signaling. Here we show that both CKI delta and CKI epsilon interacted directly with Dvl-1, and that CKI phosphorylated multiple components of the Wnt-regulated beta-catenin degradation complex in vitro, including Dvl-1, adenomatous polyposis coli (APC), axin, and beta-catenin. Comparison of peptide maps from in vivo and in vitro phosphorylated beta-catenin and axin suggests that CKI phosphorylates these proteins in vivo as well. CKI abrogated beta-catenin degradation in Xenopus egg extracts. Notably, CKI decreased, whereas inhibition of CKI increased, the association of PP2A with the beta-catenin degradation complex in vitro. Additionally, inhibition of CKI in vivo stabilized the beta-catenin degradation complex, suggesting that CKI actively destabilizes the complex in vivo. The ability of CKI to induce secondary body axes in Xenopus embryos was reduced by the B56 regulatory subunit of PP2A, and kinase-dead CKI epsilon acted synergistically with B56 in inhibiting Wnt signaling. The data suggest that CKI phosphorylates and destabilizes the beta-catenin degradation complex, likely through the dissociation of PP2A, providing a mechanism by which CKI stabilizes beta-catenin and propagates the Wnt signal.

Figure 1

CKIδ/ɛ interacts with Dvl-1. (A) CKI coimmunoprecipitates Dvl-1. In vitro translated proteins were loading directly (Input), or incubated either alone (-), or with CKIɛ or K38R. The samples were immunoprecipitated with anti-HA and coimmunoprecipitating proteins were detected by SDS/PAGE and PhosphorImager analysis. (B) Dvl-1 coimmunoprecipitates CKIδ/ɛ, but not CKIα. HA: Dvl-1 was immunoprecipitated from 293 cells with anti-HA and the lysate and pellet were probed with anti-HA, anti-CKIδ/ɛ, and anti-CKIα antibodies. The experiments were repeated twice with similar results.

Figure 2

CKI phosphorylates multiple components of the Wnt pathway. The indicated epitope-tagged Wnt pathway components were expressed in 293 cells and then immunoprecipitated and incubated with [γ-³²P]ATP with or without CKIδΔ319. Phosphorylation was assessed by SDS/PAGE and PhosphorImager analysis, and tagged proteins in the pellet were assessed by immunoblotting. The experiments were repeated twice with similar results.

Figure 3

CKI phosphorylates β-catenin and axin on major in vivo sites. Tryptic phosphopeptide maps of β-catenin (A–C) and axin (E–G). The major phosphopeptides are indicated with letters; uppercase letters denote shared peptides, whereas lowercase letters denote unique peptides. (D) The immunoblot confirms the effects of Wnt3a on endogenous β-catenin accumulation in 293 cells. α-Tubulin serves as a loading control. (H) The diagram represents the directions of electrophoresis and chromatography. The experiments were repeated twice with similar results.

Figure 4

CKI abrogates β-catenin degradation in Xenopus egg extracts. Reticulocyte lysate-expressed [³⁵S]β-catenin and luciferase were incubated in Xenopus egg extracts with or without 1.15 μM CKIɛΔ319. Protein degradation was assessed by SDS/PAGE and autoradiography. Data from three separate experiments are plotted as the ratio of β-catenin to luciferase ± SEM.

Figure 5

CKIɛ destabilizes the β-catenin degradation complex in vitro and in vivo. Lysates from 293 cells expressing (A) Myc:axin or (B) Myc:β-catenin were incubated alone (-), with 180 nM CKIɛ (CKIɛ), or with 480 nM K38R and 250 μM CKI inhibitor CKI-7 (K38R) in the presence of anti-Myc-Sepharose. The pellet and 10% of the supernatant were analyzed by SDS/PAGE and immunoblotting. (C) 293 cells expressing Myc:β-catenin were incubated with DMSO or 40 μM IC261 before harvest and anti-Myc-Sepharose immunoprecipitation. The immunopellets were analyzed by SDS/PAGE and immunoblotting. Each experiment was repeated twice with similar results.

Figure 6

B56 acts downstream of CKI. Xenopus embryos were injected with CKIδ and (A) β-gal or (B) B56α; or CKIɛ and (D) β-gal or (E) B56α RNA. (C and F) Diagrams depicting the degree of axis formation in CKIδ and CKIɛ injected embryos, respectively. Vent., ventralized; WT, wild type; Vest., vestigial secondary axis or slightly dorsalized; Incom., incomplete secondary axis; Com., complete secondary axis. The data shown are from a representative experiment which was performed five times for an n of 236 (CKIδ/β-gal) or 231 (CKIδ/B56α), or six times for an n of 290 (CKIɛ/β-gal) or 288 (CKIɛ/B56α). The statistical analysis was carried out on the cumulative experiments, and the difference between the phenotypes resulting from β-gal and B56α RNA injections in both the CKIδ and CKIɛ experiments is highly significant (P value < 10⁻⁴).

Figure 7

B56 and kinase-dead CKI synergistically inhibit Wnt signaling. Xenopus embryos were injected with Xwnt-8 and (A) β-gal, (B) B56α, (C) K38A, or (D) B56α/K38A RNA. Equal amounts of each RNA were injected per embryo, except for β-gal, which was used to normalize total RNA injected per embryo. (E) The graph is a plot of the fitted distribution of responses across the phenotypes from a representative experiment. The injections were performed eight times for an n of 365 (Xwnt-8), 365 (Xwnt-8/B56α), 369 (Xwnt-8/K38A), or 368 (Xwnt-8/K38A/B56α). The statistical analysis shows that there is a synergistic interaction between K38A and B56α, because the phenotypic shift is greater than that predicted for an additive interaction (P value = 0.0066).