Shared molecular mechanisms regulate multiple catenin proteins: canonical Wnt signals and components modulate p120-catenin isoform-1 and additional p120 subfamily members

Abstract

Wnt signaling pathways have fundamental roles in animal development and tumor progression. Here, employing Xenopus embryos and mammalian cell lines, we report that the degradation machinery of the canonical Wnt pathway modulates p120-catenin protein stability through mechanisms shared with those regulating β-catenin. For example, in common with β-catenin, exogenous expression of destruction complex components, such as GSK3β and axin, promotes degradation of p120-catenin. Again in parallel with β-catenin, reduction of canonical Wnt signals upon depletion of LRP5 and LRP6 results in p120-catenin degradation. At the primary sequence level, we resolved conserved GSK3β phosphorylation sites in the amino-terminal region of p120-catenin present exclusively in isoform-1. Point-mutagenesis of these residues inhibited the association of destruction complex components, such as those involved in ubiquitylation, resulting in stabilization of p120-catenin. Functionally, in line with predictions, p120 stabilization increased its signaling activity in the context of the p120-Kaiso pathway. Importantly, we found that two additional p120-catenin family members, ARVCF-catenin and δ-catenin, associate with axin and are degraded in its presence. Thus, as supported using gain- and loss-of-function approaches in embryo and cell line systems, canonical Wnt signals appear poised to have an impact upon a breadth of catenin biology in vertebrate development and, possibly, human cancers.

Fig. 1.

GSK3β modulates the levels of p120-catenin protein. (A) Myc–p120-catenin mRNA (0.5 ng) was microinjected into each blastomere of two-cell-stage embryos with the indicated levels (in vitro transcribed mRNA) of GSK3β kinase, GSK3β kinase-dead mutant (KD), CK1α or CK1ε. Embryos were harvested at stages 10–11 for immunoblotting (IB) with antibody against Myc, with actin serving as an internal loading control. The right panel quantitates Myc–p120 protein levels normalized to actin, using data from four independent experiments. Shown are means + s.e.m. (B) Failure of gastrulation (blastopore closure) following expression of exogenous p120-catenin (0.5 ng mRNA injection), versus rescue upon coexpression with GSK3β mRNA (5 pg) or more-partial rescue using coexpressed axin (5 pg). Shown are means + s.e.m. (C) HeLa cells were incubated in media containing LiCl (25 μM, 4 hours), and cytoplasmic versus nuclear fractionation was determined (Pierce). Endogenous p120-catenin was detected using a monoclonal antibody directed against p120 (pp120). (D) Either Venus–sh-random or Venus–sh-GSK3 and Myc–xp120-catenin were co-transfected into HeLa cells, as indicated. Forty eight hours after transfection, cells were assayed for Myc–p120 using immunofluorescence or immunoblotting (left versus right panels). For immunofluorescence, cells were fixed with 4% PFA and probed for Myc–p120 followed by Texas-red-conjugated anti-mouse visualization. Arrowheads indicate p120-catenin. For immunoblotting, tubulin served as an internal loading control. (E) HA–p120-catenin was expressed in 293T cells. Following a 1 hour incubation with 1.48×10⁶ Bq of ³⁵S-Met and -Cys, cells were treated at the times indicated with 0.5% NP-40 lysis buffer. HA–p120 was immunoprecipitated using antibodies against HA and resolved by SDS-PAGE followed by autoradiography (band densities quantitated using ImageJ). These data were collected from two independent experiments. Shown are means ± s.e.m.

Fig. 2.

Proteasome-mediated degradation appears to modulate the levels of p120-catenin. (A) MDA-231, MDA-435 and 293T cells were treated with varying doses of the proteasome inhibitor MG132 (0, 2, 5, 10, 25 μM). After 6 hours, cells were harvested for immunoblotting (IB) with an antibody against p120 (6H11 and pp120). (B) Myc–p120-catenin was co-transfected into HeLa cells with HA–ubiquitin. After 24 hours, cells were treated with LiCl (25 μM) for 4 hours to block GSK3β function. Cells were harvested and HA–ubiquitylated proteins immunoprecipitated, followed by Myc–p120 immunoblotting. Endogenous β-catenin was used as a positive ubiquitylation control. (C) MDA-435 and 293T cells were treated with varying doses of the CK1 inhibitor D4476 (0, 2, 5, 10, 25 μM). p120-catenin was visualized using an antibody directed against this protein (pp120), and endogenous β-catenin served as an endogenous positive control. (D) Single cells of two-cell embryos were injected with Myc–p120-catenin and Myc–β-catenin mRNA (0.1 ng mRNA each) and treated for 24 hours with varying concentrations of the CK1 inhibitor D4476 (10, 50, 200 μM) in the presence of Fugene6. An antibody against Myc was used to detect both p120-catenin and β-catenin. (E) HeLa cells were grown on coverslips, transiently transfected with Myc–kaiso and treated with MG132 (10 μM) and LiCl (50 μM). Cells were fixed with 4% PFA for 10 minutes, blocked with 5% goat serum in PBS and immunostained with antibody against Myc.

Fig. 3.

p120-catenin associates with components of the destruction complex. (A) HA–p120-catenin (1 ng) was co-injected with either Myc–GSK3β or Myc–GSK3β KD (0.5 ng) into both blastomeres of two-cell embryos. Myc–GSK3β immunoprecipitates (IP) were immunoblotted (IB) with antibody against HA to detect p120. (B) HA–axin (0.5 ng) was microinjected with Myc–p120-catenin (0.5 ng) or Myc–Kaiso (0.5 ng) into both blastomeres of two-cell embryos, subsequently harvested at gastrulation. Anti-HA antibody (axin) immunoprecipitates were blotted with antibody against Myc to detect p120-catenin versus Kaiso (negative control). WCL, whole-cell lysate. (C) HA–CK1α was injected into both blastomeres of two-cell embryos along with Myc–p120, Myc–C-cadherin or Myc–Kaiso, and embryos harvested at early–mid gastrulation (stage 10–11). This was followed by anti-HA immunoprecipitation (CK1α) and then anti-Myc (p120, or negative controls C-cadherin or kaiso) or anti-HA immunoblotting (CK1α). (D) The association of Myc–β-TrCP (0.5 ng) and HA–p120-catenin (1 ng) was resolved using methods analogous to those used in C. (E) 293T and MDA-435 cells were grown in 10 cm dishes, and lysates immunoprecipitated for endogenous GSK3β (BD Transduction). The association of endogenous p120-catenin (6H11) with GSK3β was resolved by immunoblotting (β-catenin positive control).

Fig. 4.

Mapping of the association of p120 with GSK3β, CK1α and ubiquitin. (A) Myc-tagged p120-catenin deletion constructs (a–f). The table summarizes the relative p120-construct:GSK3β association, as shown in B. (B) HA–GSK3β (0.5 ng mRNA) was coexpressed with varying p120–catenin deletion constructs (a–f) (0.5 ng), followed by HA–GSK3β immunoprecipitation and Myc-construct immunoblotting (IB). (C) Either Myc–p120-catenin (fl) or Myc–ΔN-p120-catenin (construct d) was co-transfected with HA-tagged ubiquitin into HeLa cells. Following HA–ubiquitin immunoprecipitation, co-associated (versus not associated) Myc–p120 constructs were visualized by immunoblotting. (D) Myc–p120 constructs (fl, c and d) (0.5 ng mRNA) were co-injected with HA–CK1α (0.5 ng) in both blastomeres of two-cell embryos. Gastrula embryo extracts were immunoprecipitated for HA–CK1α and blotted for co-associated (versus not associated) constructs of Myc–p120. Right panel, lysate indicates Myc–p120 construct expression.

Fig. 5.

Phosphorylation-dependent p120-catenin ubiquitylation and proteasomal degradation. (A) Cross-species sequence alignment of p120-catenin regions harboring conserved predicted GSK3β phosphorylation and ubiquitylation sites. Highlights shown in each region include conserved serine residues, and a DSEXXS motif for β-TrCP recognition. (B) HA–p120 (0.3 ng mRNA), versus the HA–p120^4SA point mutant (4S→A, see Fig. 4A, 0.3 ng), was co-injected with HA–GSK3β (0.1 ng) into both cells of two-cell embryos. Gastrula embryo lysates (stages 11–12) were immunoblotted (IB) for the HA–p120 constructs, revealing a differing response to GSK3β. (C) HA–p120 versus HA–p120^4SA (0.5 ng) were co-injected with Myc–β-TrCP (0.5 ng) into one blastomere of two-cell embryos. Gastrula embryo lysates were immunoprecipitated for Myc–β-TrCP and assayed for co-associated HA–p120 or HA–p120^4SA. (D) Myc–p120 or Myc–p120^4SA was co-transfected with HA–ubiquitin into HeLa cells. HA–ubiquitin was immunoprecipitated and then immunoblotting used to detect ubiquitylated Myc–p120 versus Myc–p120^4SA. (E) The half-lives of HA–p120 versus HA–p120^4SA were monitored by pulse–chase analysis. Following a 1 hour pulse–chase with [³⁵S] Met and Cys, HeLa cells transfected with HA–p120 and HA–p120^4SA were harvested at the times indicated, and anti-HA immunoprecipitates were resolved by SDS-PAGE and visualized by autoradiography (quantitation of band intensities employed Image J and was normalized to the zero time-point). These data are representative of two independent experiments. (F) Myc–GSK3β (5 pg mRNA) or Myc–kaiso (0.25 ng) was microinjected with HA–p120 versus HA–p120^4SA (0.25 ng) into one blastomere of two-cell embryos. Total mRNA injection loads were equalized using β-galactosidase mRNA. Gastrula embryo cDNA (stages 10–12) was subject to RT-PCR to assay endogenous xWnt-11 transcript levels. Band intensities are indicated relative to the β-gal control (set at 1), following normalization to the internal loading control (encoding histone H4). (G) Myc–kaiso or β-gal (negative control) was injected alone (0.25 ng), or Myc–kaiso was co-injected with HA–p120 (WT) versus HA–p120^4SA (0.25 ng), into one blastomere of two-cell embryos. Gastrula embryo cDNA was assayed by real-time RT-PCR for xWnt-11 transcript levels. ODC, ornithine decarboxylase.

Fig. 6.

Axin promotes the degradation of members of the p120-catenin subfamily. (A) The indicated Myc-tagged p120 subfamily members (1 ng mRNA each) were co-injected with Myc–axin (0.1 ng) into both blastomeres of two-cell embryos, and gastrula embryo lysates were Myc-immunoblotted (IB). HA–kazrin and HA–EWS (1 ng each) were co-injected with Myc–axin (0.5 ng) as negative controls for the effects of axin, whereas actin and GAPDH served as internal loading controls. (B) The indicated HA-tagged p120 subfamily members (1 ng each) were co-injected with Myc–axin (1 ng) into both blastomeres of two-cell embryos. Gastrula (stage 11–12) embryo lysates were immunoprecipitated for Myc–axin, followed by immunoblotting for HA-tagged p120 subfamily catenins. The bottom panel confirms axin immunoprecipitations. HA–kaiso and HA–Dyrk served as negative controls. (C) HA–axin was co-injected with either Myc-tagged β-catenin or δ-catenin into both blastomeres of two-cell embryos. HA–axin immunoprecipitates were immunoblotted with antibody against Myc to detect β-catenin or δ-catenin. (D) Either HA–p120-catenin (0.25 ng) or Myc–β-catenin mRNA (0.25 ng) was co-injected with Frodo morpholino [10 ng versus standard morpholino (STD-MO) or negative control], into both blastomeres of two-cell embryos. Embryos were harvested at stage 11–12 for immunoblotting with antibody against Myc (β-catenin) or HA (p120). Actin served as a loading or negative control. (E) Either Myc–β-catenin (0.5 ng) or Myc–p120-catenin (0.5 ng) was co-injected with HA–Frodo (0.5 ng) into both blastomere of embryos at the two-cell stage. Embryos were harvested at early-mid gastrula stages (10–11) and lysates immunoprecipitated for HA–Frodo followed by immunoblotting. WCL, whole-cell lysate.

Fig. 7.

Upstream Wnt pathway components modulate the levels of p120-catenin. (A) Co-injection of Frodo or Wnt8 counter the effectiveness of Myc–GSK3β (5 pg) or Myc–axin (0.1 ng) to reduce HA–p120 (0.5 ng) levels. Embryos expressing up to two exogenous constructs (as noted) in addition to HA–p120 were harvested at gastrulation, and the corresponding lysates immunoblotted (IB) with antibody against HA to detect p120. Actin was used as an internal loading control. Total mRNA injection loads were equalized using β-galactosidase. (B) Exogenous Wnts (0.5 ng of Wnt8, Wnt11 or Wnt5a mRNA) were co-injected with both HA–GSK3β (0.1 ng) and HA–p120 (0.25 ng) into both blastomere of two-cell embryos. At stage 12, Wnt protection (versus no protection) from GSK3β impact upon HA–p120 was assessed by immunoblotting. Actin served as an internal loading control. (C) An increasing dose of Wnt8 mRNA was co-injected with both HA–p120 (0.5 ng) and Myc–axin (0.5 ng) into two blastomeres of two-cell embryos that were later collected at gastrulation (stage 11, with 20 embryos per condition). The protein levels of p120-catenin were assessed by immunoblotting. (D) MDA-231, HeLa and 293T cells were transfected with siRNAs directed against axin-1 and axin-2 (50 pmol), as indicated, for 48 hours. Endogenous p120-catenin levels were monitored through anti-p120 immunoblotting (6H11). Each experiment was repeated at least three times. nc, negative control. (E) HeLa cells were transiently transfected with siRNA for LRP5 or LRP6 (50 pmol), along with either Myc-tagged LRP5 or LRP6, and effects assessed by means of Myc-immunoblotting. (F) MDA-435 cells were seeded in six-well plates, followed by transfection with LRP5 and LRP6 siRNAs. The effect of LRP5 and LRP6 depletion on p120-catenin was monitored using distinct antibodies directed against p120 (pp120 or 6H11). The asterisk (*) indicates a nonspecific band serving as an additional negative control. (G) The stability of p120-catenin in HeLa cells was resolved as described for F. (H) Myc–p120 full-length (isoform-1) or ΔN-p120 were transiently transfected into 293T cells for 24 hours, followed by MG132 or D4476 treatment (6 hours) at the doses indicated. The levels of p120 were monitored through Myc-immunoblotting. (I) 293T cells were transiently transfected with mouse p120 isoform-1A or isoform-3A. p120-catenin isoforms were monitored in the presence of BIO or D4476 (6 hours), employing an antibody against Myc.