Axin-mediated CKI phosphorylation of beta-catenin at Ser 45: a molecular switch for the Wnt pathway

Abstract

The Wnt pathway controls numerous developmental processes via the beta-catenin-TCF/LEF transcription complex. Deregulation of the pathway results in the aberrant accumulation of beta-catenin in the nucleus, often leading to cancer. Normally, cytoplasmic beta-catenin associates with APC and axin and is continuously phosphorylated by GSK-3beta, marking it for proteasomal degradation. Wnt signaling is considered to prevent GSK-3beta from phosphorylating beta-catenin, thus causing its stabilization. However, the Wnt mechanism of action has not been resolved. Here we study the regulation of beta-catenin phosphorylation and degradation by the Wnt pathway. Using mass spectrometry and phosphopeptide-specific antibodies, we show that a complex of axin and casein kinase I (CKI) induces beta-catenin phosphorylation at a single site: serine 45 (S45). Immunopurified axin and recombinant CKI phosphorylate beta-catenin in vitro at S45; CKI inhibition suppresses this phosphorylation in vivo. CKI phosphorylation creates a priming site for GSK-3beta and is both necessary and sufficient to initiate the beta-catenin phosphorylation-degradation cascade. Wnt3A signaling and Dvl overexpression suppress S45 phosphorylation, thereby precluding the initiation of the cascade. Thus, a single, CKI-dependent phosphorylation event serves as a molecular switch for the Wnt pathway.

Figure 1

Axin induces β-catenin phosphorylation exclusively on S45, yet proteasomal degradation requires further GSK-mediated phosphorylation at T41, S37, and S33. (a) β-Catenin (wild-type or N-terminal-mutated β-catenin; DP) stability in response to transfection of axin and GSK-3β. A dominant-negative Flag–ΔF-box–β-TrCP fragment (WD) was included in lane 7. A GFP expression vector was included in all transfections as an expression reference. (b) β-Catenin phosphorylation analysis with anti-phosphopeptide antibodies in proteasome-inhibited cells. (c) β-Catenin phosphorylation analysis by mass spectrometry (MS). Immunopurified Flag–β-catenin bands were digested with Asp-N endoproteinase, and the resulting N-terminal peptides, including Asp17–Leu31 and Asp32–Glu55 (shown on top) were analyzed by nanoelectrospray MS. The m/z ratios shown (six top panels) correspond to triple-charged [M+3H]³⁺ ions of the Asp32–Glu55 peptide, marked by asterisks (including all its isotopic variants). Displayed are spectra of peptides containing 1 (+1P) and 4 (+4P) phosphate groups from β-catenin alone (panels 1,2), β-catenin coexpressed with axin (panels 3,4), and axin plus GSK-3β (panels 5,6). [M+3H]³⁺ ions from bar-marked peptides (gray and black) were selected for fragmentation. The spectra in the lower panels (7,8) correspond to fragments derived from MS/MS of the gray bar peptide (from panel 3) and black bar peptide (from panel 6). The Asp32–Glu55 peptide was fragmented by MS/MS into halves, a C-terminal half containing Ser 45 and an N-terminal half containing Ser 33/Ser 37/Thr 41. Whereas the gray bar C-terminal fragment is mostly phosphorylated, the N-terminal fragment is mostly unphosphorylated (panel 7). The relationship of nonphosphorylated to phosphorylated (+1P) fragments is indicated by angled arrows. The black bar N-terminal fragment (panel 8) is mostly triply phosphorylated (+3P). Its C-terminal fragment is singly phosphorylated (data not shown), similarly to that of the gray bar. MS/MS analysis of the Asp17–Leu 31 peptide (from +1P of the MS) reveals minor, yet significant Ser 29 phosphorylation (data not shown).

Figure 2

Mutations at four β-catenin N-terminal phosphorylation sites curtail the phosphorylation cascade, leading to β-catenin stabilization. (a) Wild-type and mutated β-catenin from MG-132 treated cells were detected with anti-Myc or anti-phospho-β-catenin antibodies. (b) Stability analysis of phosphorylation-site β-catenin mutants in MG-132-untreated cells.

Figure 3

Axin-mediated S45 phosphorylation is GSK-3β-independent. (a) S45 phosphorylation is not affected by LiCl. Analysis of wild-type or T41A β-catenin from MG-132-treated cells, with or without LiCl. No phosphorylation signal at S33/37 was detected in the T41A mutant (see Fig. 2a). (b) L525P-Axin does not bind GSK-3β. Anti-Flag immunoprecipitation from cells transfected with Flag-GSK-3β and wild-type or L525P Myc-axin. (c) L525P-Axin induces S45 phosphorylation and supports a phosphorylation–degradation cascade in the presence of exogenous GSK-3β. All cells were transfected with Myc-β-catenin and the indicated expression vectors.

Figure 4

Axin-independent, surrogate phosphorylation at S45 promotes the GSK-3β phosphorylation cascade, resulting in β-catenin degradation. All cells were transfected with the β-catenin mutant 45PKA and the indicated expression vectors. HA–GSK-3β (WT) or Flag–R96A–GSK-3β (R96A) was detected using anti-GSK-3β antibodies (GSK-3β position varies according to its tag).

Figure 5

Identification of the β-catenin S45 kinase. (a) Immunopurified axin and recombinant CKIδ (aa1–318) phosphorylate β-catenin in vitro at S45, creating a priming site for sequential phosphorylation at S33/37 by recombinant GSK-3β. Flag–β-catenin was immunoprecipitated from transfected 293 cells and subjected to an in vitro kinase assay using the indicated kinase or immunopurified Flag–axin from 293 transfectants, with or without ATP (30 μM). β-Catenin phosphorylation was analyzed by Western blot, using anti-phospho-β-catenin antibodies. (b) Dominant-negative CKI inhibits axin-induced S45 phosphorylation. 293 cells were transfected with Myc–β-catenin alone (lane 1), with axin (lanes 2–5) and wild-type (lane 3), or dn-CKIɛ (D-N, lane 4; K-R, lane 5). The bottom CKI arrow points to the endogenous hCKIɛ, whereas the top arrow points to the exogenous xCKIɛ proteins detected by monoclonal anti-CKIɛ. (c) CKI-7 inhibits the axin-induced S45 phosphorylation of wild-type β-catenin, but not the constitutive phosphorylation of 45PKA. (d) Effects of specific kinase inhibitors on the in vivo phosphorylation of β-catenin at T41, S45, and S33 in HeLa cells. CKI-7 was used at 100 μM, H89 at 5 μM, and LiCl at 40 mM. β-Catenin was detected with anti-β-catenin antibody or anti-phospho-β-catenin antibodies.

Figure 6

The Wnt pathway targets the axin-mediated priming phosphorylation of β-catenin at S45. (a) Wnt3A inhibits S45 phosphorylation. Mouse L cells, β-catenin transfected Jurkat T cells, SNU449 hepatoma cells (Satoh et al. 2000), and HeLa cells were incubated (5 h) with conditioned medium of wild-type or Wnt3A-secreting L cells (Shibamoto et al. 1998). MG-132 or okadaic acid (OKA, 1 μM) was added to the indicated cultures. Endogenous (HeLa, L cells, and SNU449) and exogenous (Jurkat) β-catenin was detected with anti-β-catenin antibody, or anti-phospho-β-catenin antibodies. The numbers under the different lanes are relative densitometry figures, referring to the signal intensity of untreated cultures. (b) Dvl-1 inhibits axin-mediated S45 phosphorylation. Mass spectra of the Asp32–Glu55 peptide with one phosphate group (+1P): from β-catenin coexpressed with L525P-axin (m-axin), compared with that of β-catenin coexpressed with L525P-axin and Dvl-1. (c) Dvl-1 overexpression results in inhibition of S45 phosphorylation and β-catenin stabilization. β-Catenin of 293 transfectants was detected using anti-Myc or anti-phospho-β-catenin antibodies. (d) Dvl-1 has no effect on axin-independent, S45-phosphorylation-initiated GSK-3β cascade: Analysis of 45PKA-transfected cells.

Figure 7

A model depicting the β-catenin phosphorylation–degradation cascade. (a) Axin recruits CKI to phosphorylate β-catenin at S45. (b,c) S45 phosphorylation primes β-catenin for the succeeding GSK-3β phosphorylation cascade, finally hitting the S33/37 sites. (d) Phosphorylation at S33/37 creates a docking site for β-TrCP/E3RS, promoting the ubiquitination and degradation of β-catenin. (e) Wnt signaling, possibly through Dvl and CKI, regulates S45 phosphorylation.