Casein Kinase II Phosphorylation of E-cadherin Increases E-cadherin/β-Catenin Interaction and Strengthens Cell-Cell Adhesion
2000; Elsevier BV; Volume: 275; Issue: 7 Linguagem: Inglês
10.1074/jbc.275.7.5090
ISSN1083-351X
AutoresHeiko Lickert, Andreas Bauer, Rolf Kemler, Jörg Stappert,
Tópico(s)Kruppel-like factors research
Resumoβ-Catenin, a member of the Armadillo repeat protein family, binds directly to the cytoplasmic domain of E-cadherin, linking it via α-catenin to the actin cytoskeleton. A 30-amino acid region within the cytoplasmic domain of E-cadherin, conserved among all classical cadherins, has been shown to be essential for β-catenin binding. This region harbors several putative casein kinase II (CKII) and glycogen synthase kinase-3β (GSK-3β) phosphorylation sites and is highly phosphorylated. Here we report that in vitro this region is indeed phosphorylated by CKII and GSK-3β, which results in an increased binding of β-catenin to E-cadherin. Additionally, in mouse NIH3T3 fibroblasts expression of E-cadherin with mutations in putative CKII sites resulted in reduced cell-cell contacts. Thus, phosphorylation of the E-cadherin cytoplasmic domain by CKII and GSK-3β appears to modulate the affinity between β-catenin and E-cadherin, ultimately modifying the strength of cell-cell adhesion. β-Catenin, a member of the Armadillo repeat protein family, binds directly to the cytoplasmic domain of E-cadherin, linking it via α-catenin to the actin cytoskeleton. A 30-amino acid region within the cytoplasmic domain of E-cadherin, conserved among all classical cadherins, has been shown to be essential for β-catenin binding. This region harbors several putative casein kinase II (CKII) and glycogen synthase kinase-3β (GSK-3β) phosphorylation sites and is highly phosphorylated. Here we report that in vitro this region is indeed phosphorylated by CKII and GSK-3β, which results in an increased binding of β-catenin to E-cadherin. Additionally, in mouse NIH3T3 fibroblasts expression of E-cadherin with mutations in putative CKII sites resulted in reduced cell-cell contacts. Thus, phosphorylation of the E-cadherin cytoplasmic domain by CKII and GSK-3β appears to modulate the affinity between β-catenin and E-cadherin, ultimately modifying the strength of cell-cell adhesion. amino acid(s) casein kinase II adenomatous polyposis coli glutathione S-transferase polymerase chain reaction phosphate-buffered saline 1,4-piperazinediethanesulfonic acid polyacrylamide gel electrophoresis wild type glycogen synthase kinase-3β E-cadherin belongs to the group of classical cadherins, Ca2+-dependent adhesion molecules, that mediate cell-cell adhesion in many different tissues and various species (1.Stappert J. Kemler R. Adv. Mol. Cell Biol. 1999; 28: 27-64Crossref Scopus (8) Google Scholar). E-cadherin is a type I transmembrane protein, making mostly homophilic cell-cell interactions via its extracellular domain, whereas the cytoplasmic domain is anchored to actin microfilaments via three cytoplasmic proteins α-, β-, and γ-catenin (plakoglobin) (2.Nagafuchi A. Takeichi M. EMBO J. 1988; 7: 3679-3684Crossref PubMed Scopus (664) Google Scholar, 3.Ozawa M. Baribault H. Kemler R. EMBO J. 1989; 8: 1711-1717Crossref PubMed Scopus (1152) Google Scholar). Another catenin p120 ctn is less tightly associated to E-cadherin, and its function is presently less well understood (4.Reynolds A.B. Daniel J.M. Mo Y.Y. Wu J. Zhang Z. Exp. Cell Res. 1996; 225: 328-337Crossref PubMed Scopus (128) Google Scholar). Biochemical studies on cultured cells as well as the use of recombinant proteins have provided a detailed picture of how the different E-cadherin-associated components interact (2.Nagafuchi A. Takeichi M. EMBO J. 1988; 7: 3679-3684Crossref PubMed Scopus (664) Google Scholar, 3.Ozawa M. Baribault H. Kemler R. EMBO J. 1989; 8: 1711-1717Crossref PubMed Scopus (1152) Google Scholar, 5.Ozawa M. Kemler R. J. Cell Biol. 1992; 116: 989-996Crossref PubMed Scopus (329) Google Scholar, 6.Hinck L. Nathke I.S. Papkoff J. Nelson W.J. J. Cell Biol. 1994; 125: 1327-1340Crossref PubMed Scopus (557) Google Scholar, 7.Nathke I.S. Hinck L. Swedlow J.R. Papkoff J. Nelson W.J. J. Cell Biol. 1994; 125: 1341-1352Crossref PubMed Scopus (274) Google Scholar). Whereas β-catenin or plakoglobin bind directly to the cytoplasmic domain of E-cadherin, the linkage of the whole complex to the cytoskeleton is mediated by α-catenin, which binds to β-catenin or plakoglobin on one side and to actin on the other side (8.Knudsen K.A. Soler A.P. Johnson K.R. Wheelock M.J. J. Cell Biol. 1995; 130: 67-77Crossref PubMed Scopus (564) Google Scholar, 9.Rimm D.L. Koslov E.R. Kebriaei P. Cianci C.D. Morrow J.S. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 8813-8817Crossref PubMed Scopus (637) Google Scholar, 10.Obama H. Ozawa M. J. Biol. Chem. 1997; 272: 11017-11020Abstract Full Text Full Text PDF PubMed Scopus (46) Google Scholar). Plakoglobin, originally found as a major component of desmosomal junctions, and β-catenin form mutually exclusive complexes with E-cadherin, but the function of these two different complexes is not yet known (7.Nathke I.S. Hinck L. Swedlow J.R. Papkoff J. Nelson W.J. J. Cell Biol. 1994; 125: 1341-1352Crossref PubMed Scopus (274) Google Scholar, 11.Butz S. Kemler R. FEBS Lett. 1994; 355: 195-200Crossref PubMed Scopus (114) Google Scholar). Both proteins belong to the so-called Arm-repeat protein family, a class of proteins characterized by a repeating motif originally identified in the product of the Drosophilasegment polarity gene armadillo (12.Peifer M. Berg S. Reynolds A.B. Cell. 1994; 76: 789-791Abstract Full Text PDF PubMed Scopus (550) Google Scholar). Armadillo, as well as β-catenin, have been shown to be part of the Wingless cascade inDrosophila (Wnt in vertebrates) (for review, see Ref.13.Willert K. Nusse R. Curr. Opin. Genet. & Dev. 1998; 8: 95-102Crossref PubMed Scopus (666) Google Scholar). Cell-cell adhesion is a rather dynamic process for which interactions between cell adhesion molecules and the cytoskeleton must be continually modified. Thus E-cadherin and especially catenins are major target sites for posttranslational modifications, primarily phosphorylation and dephosphorylation events. Tyrosine phosphorylation of β-catenin and plakoglobin decreases cell-cell adhesion (14.Matsuyoshi N. Hamaguchi M. Taniguchi S. Nagafuchi A. Tsukita S. Takeichi M. J. Cell Biol. 1992; 118: 703-714Crossref PubMed Scopus (450) Google Scholar, 15.Behrens J. Vakaet L. Friis R. Winterhager E. Vanroy F. Mareel M.M. Birchmeier W. J. Cell Biol. 1993; 120: 757-766Crossref PubMed Scopus (842) Google Scholar, 16.Hoschuetzky H. Aberle H. Kemler R. J. Cell Biol. 1994; 127: 1375-1380Crossref PubMed Scopus (673) Google Scholar, 17.Ozawa M. Kemler R. J. Biol. Chem. 1998; 273: 6166-6170Abstract Full Text Full Text PDF PubMed Scopus (177) Google Scholar), whereas ectopic expression of tyrosine phosphatases strengthens it (18.Müller T. Choidas A. Reichmann E. Ullrich A. J. Biol. Chem. 1999; 274: 10173-10183Abstract Full Text Full Text PDF PubMed Scopus (268) Google Scholar), but the exact underlying mechanisms are poorly understood. Even less is known about the serine/threonine kinases involved in phosphorylating catenins or E-cadherin. Treatment of human epidermal cells with okadaic acid disrupts cell-cell contacts and results in hyperphosphorylation of β-catenin (19.Serres M. Grangeasse C. Haftek M. Durocher Y. Duclos B. Schmitt D. Exp. Cell Res. 1997; 231: 163-172Crossref PubMed Scopus (51) Google Scholar). Serine/threonine phosphorylation of E-cadherin during pre- implantation development of the mouse embryo was also found to increase in the early eight-cell stage, and it was hypothesized that phosphorylation of E-cadherin is necessary for its redistribution during compaction (20.Sefton M. Johnson M.H. Clayton L. McConnell J.M. Mol. Reprod. Dev. 1996; 44: 77-87Crossref PubMed Scopus (23) Google Scholar). In a previous study we have shown that phosphorylation of E-cadherin is concentrated to a short stretch of 30 aa1 in the cytoplasmic domain (21.Stappert J. Kemler R. Cell Adhes. Commun. 1994; 2: 319-327Crossref PubMed Scopus (200) Google Scholar). This region is necessary and sufficient for the interaction with β-catenin or plakoglobin, and it harbors a cluster of 8 Ser residues. When all Ser were substituted by alanines, phosphorylation of E-cadherin in vivo was completely abolished. The lack of phosphorylation also prevented association between E-cadherin and β-catenin or plakoglobin. In fact, phosphorylation of two other binding partners of β-catenin, the adenomatous polyposis coli (APC) tumor suppressor protein and the DF3/MUC1 breast carcinoma associated antigen, was also shown to increase their in vitro binding affinity to β-catenin (22.Rubinfeld B. Albert I. Porfiri E. Fiol C. Munemitsu S. Polakis P. Science. 1996; 272: 1023-1026Crossref PubMed Scopus (1310) Google Scholar, 23.Yamamoto M. Bharti A. Li Y.Q. Kufe D. J. Biol. Chem. 1997; 272: 12492-12494Abstract Full Text Full Text PDF PubMed Scopus (284) Google Scholar). Thus Ser/Thr phosphorylation may be a common mechanism to modulate the interaction between β-catenin and its binding partners, but little is known about the kinases involved. Therefore, we analyzed the phosphorylation of E-cadherin in more detail. Here we show that E-cadherin is phosphorylated in the Ser cluster by casein kinase II (CKII) and glycogen synthase kinase-3β (GSK-3β). Preventing this phosphorylation not only reduces the interaction between β-catenin and E-cadherin in vitro but also reduces E-cadherin-mediated cell-cell adhesion in transfected mouse NIH3T3 fibroblasts. Thus phosphorylation of E-cadherin appears to be a crucial mechanism by which cell-cell adhesion is modulated. NIH3T3 fibroblasts were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum. E-cadherin-expressing cells were generated by stable transfection of NIH3T3-lacZ and NIH3T3-Wnt-1 with the following expression plasmids: pcDNA3.1zeo-MMEC (musmusculus E- cadherin) and the substitution mutants S840A, S853A, S855A, and S853/855A in the same vector. NIH3T3-lacZ and NIH3T3-Wnt-1 cells were kindly provided by A. Kispert, MPI-IB, Freiburg, Germany. Stable clones were selected for resistance to 500 μg/ml Zeocin. The E-cadherin expression of several clones was analyzed by flow cytometry on a Becton Dickinson FACScan using a monoclonal antibody against hE-cadherin and a fluorescein isothiocyanate-conjugated second antibody (3.Ozawa M. Baribault H. Kemler R. EMBO J. 1989; 8: 1711-1717Crossref PubMed Scopus (1152) Google Scholar). For further analysis clones with an equal amount of E-cadherin on the cell surface were used. Mouse monoclonal antibodies against human E-cadherin and mouse β-catenin were from Transduction Laboratories (Lexington, KY) and the monoclonal anti-GST antibody from Sigma; 3000 mCi/ml γ-[32P]ATP and [32P]orthophosphate were from Amersham Pharmacia Biotech. Plasmids coding the E-cadherin Ser-Ala substitutions were generated by PCR-based site-directed mutagenesis using the pSK+UM (Uvomorulin, E-cadherin) plasmid as template (21.Stappert J. Kemler R. Cell Adhes. Commun. 1994; 2: 319-327Crossref PubMed Scopus (200) Google Scholar). pSK+UMS853A, pSK+UMS855A, and pSK+UMS853/855A were generated with the universal 3′-primer MECHindIII, 5′-GGTCGACGGTATCGATAAGC-3′, in combination with one of the following 5′-primers: MEC-S853A, 5′-AAGCCGCTAGCCTGAGCTCACTGAACTCCGCTGAGTCGGATCAGG-3′; MEC855A, 5′-AAGCCGCTAGCCTGAGCTCACTGAACTCCTCTGAGGCGGATCAGGACC-3′; and MEC-S853/855A, 5′-AAGCCGCTAGCCTGAGCTCACTGAACTCCGCTGAGGCGGATCAGGACC-3′.The pSK+UMS840A cloning fragment was amplified with the 5′-primer MECBstEII, 5′-AAAGACCAGGTGACCACGTTG-3′and the 3′-primer MECS840A, 5′-GCTCAGGCTAGCGGCTTCAGAACCAGCGCCCTCGTAATC-3′. For MECS853A, MECS855A and MECS853A/S855A, the PCR fragments and the pSK+UM vector were cut with NheI andHindIII and ligated. The MECS840A PCR product was cut withNheI and BstEII and ligated into pSK+UM, cleaved with the same enzymes. All constructs were confirmed by DNA sequencing. For eukaryotic expression of the various E-cadherin mutants, the HindIII/BglII fragments of the described pSK+UM constructs were inserted into the respective sites of the expression vector pcDNA3 (Invitrogen; Carlsbad, CA). The expression plasmid pGEXUC1, coding for the cytoplasmic domain of murine E-cadherin (ECT, amino acids 580–728) fused to the C terminus of GST from pGEX2T (Amersham Pharmacia Biotech), has been described (24.Marrs J.A. Napolitano E.W. Murphy-Erdosh C. Mays R.W. Reichardt L.F. Nelson W.J. J. Cell Biol. 1993; 123: 149-164Crossref PubMed Scopus (138) Google Scholar). pGEXUC1 was cut withBamHI/EcoRI, and the fragment coding for ECT was cloned into the respective sites of pGEX4T1. The region coding for the cytoplasmic part of the mutant E-cadherin was isolated from the pSK+UM vector, cleaved withClaI/XhoI, and ligated into theClaI/SalI-cut pGEX4T1-ECT to generate the following plasmids: pGEX4T1-ECTS840A, pGEX4T1-ECTS853A, pGEX4T1-ECTS855A, and pGEX4T1-ECTS853/855A. All cDNAs were expressed as ECT-tagged proteins in Escherichia coli. The recombinant glutathioneS-transferase (GST)-tagged cytoplasmic domain (ECT) of mouse E-cadherin and the E-cadherin substitution mutants were expressed inE. coli and affinity purified on glutathione S-transferase-agarose beads as described (25.Aberle H. Butz S. Stappert J. Weissig H. Kemler R. Hoschuetzky H. J. Cell Sci. 1994; 107: 3655-3663Crossref PubMed Google Scholar). 2 μg of GST-ECT or the GST-ECT substitution mutants were immobilized on 30 μl of glutathione-agarose beads (50% w/v) and washed in CKII kinase buffer (0.1 m HEPES, pH 7.5; 0.1 m MgCl2; 1 m KCl; 50 mm dithiothreitol) or GSK-3β kinase buffer (0.2 m Tris-HCl, pH 7.5; 0.1 mMgCl2; 50 mm dithiothreitol). The proteins were phosphorylated in a 50-μl reaction for 15–60 min at 30 °C with either 1 unit of casein kinase II (New England Biolabs, Beverly, MA) or 0.1 unit of glycogen synthase kinase-3β (New England Biolabs), in the presence of 10 μCi of γ-[32P]ATP. To remove unincorporated radioactivity, the beads were washed in association buffer (0.05% (v/v) Triton X-100 in PBS). Then the GST-ECT fusion proteins were eluted and separated on a 7.5% SDS-polyacrylamide gel. Radioactive gels were Coomassie Blue-stained and quantified with a BAS 1000 Bioimaging Analyzer (Fuji). Coomassie Blue staining was quantified using the NIH Imager program version 1.59. 2 μg of GST-ECT was immobilized as described above and washed in association buffer (0.05% (v/v) Triton X-100 in PBS). For reconstitution of the protein complex 10 or 100 nm recombinant His-tagged full-length mouse β-catenin (26.Aberle H. Bauer A. Stappert J. Kispert A. Kemler R. EMBO J. 1997; 16: 3797-3804Crossref PubMed Scopus (2172) Google Scholar) was added and incubated for 1 h at 37 °C. The beads bearing the complexes were washed and eluted for separation by 7.5% SDS-PAGE and immunoblotting. For the phosphorylation-dependent association analysis, the GST-ECT fusion proteins were phosphorylated as described above before performing the binding reactions. NIH3T3-Wnt-1 cells stably transfected with either MMEC, MMECS840A, MMECS853A, MMECS855A or MMEC8531855A were lysed in CSK buffer (150 mm NaCl; 300 mm sucrose; 10 mm PIPES, pH 6.8; 3 mm MgCl2; 0.5% (v/v) Triton X-100) and separated into detergent-soluble and -insoluble fractions by centrifugation at 12,000 × g for 10 min. For immunoprecipitation, the supernatant was precleared with 10% protein A-Sepharose beads (Amersham Pharmacia Biotech) for 1 h at 4 °C. Approximately 2 μg of E-cadherin antibody was coupled to protein A-Sepharose beads. The antigen-antibody complex was washed with CSK buffer, boiled in SDS sample buffer, and analyzed by SDS-PAGE as described (3.Ozawa M. Baribault H. Kemler R. EMBO J. 1989; 8: 1711-1717Crossref PubMed Scopus (1152) Google Scholar). HEK 293 cells (6·106) were transiently transfected with 20 μg of either MMEC or MMECS853/855A. After 24 h the cells were grown for 2 h in phosphate-free medium followed by an incubation period of 16 h in the presence of 200 μCi/ml [32P]orthophosphate. Immunoprecipitation and immunoblotting were done as described. Western blots (25.Aberle H. Butz S. Stappert J. Weissig H. Kemler R. Hoschuetzky H. J. Cell Sci. 1994; 107: 3655-3663Crossref PubMed Google Scholar) were quantified by Lumi-Imager (Roche Molecular Biochemicals). NIH3T3 fibroblasts or stably transfected NIH3T3 cells were seeded at a very low density (0.5 × 106/10-cm dish) and cultivated for 12–16 h at 37 °C. The cells were washed in PBS and trypsinized (0.01% (v/v) trypsin in HEPES-buffered saline) in the presence of 2 mmCaCl2 for 10 min at 37 °C. The cells were resuspended in aggregation buffer (Dulbecco's modified Eagle's medium/PBS = 1:1 plus 5% heat inactivated fetal calf serum), washed, sedimented at 1000 × g for 3 min, and resuspended in the same buffer for counting. For adhesion assays on 1% agarose-covered 6-well dishes, 3–6 × 105 cells were incubated in 3 ml of aggregation buffer containing 5 μg/ml DNase I at a constant rotation of 70 rpm at 37 °C. At 0, 15, 30, and 45 min aggregates were randomly photographed at × 10 magnification with a computer-controlled digital C4880 camera (Hamamatsu, Hamamatsu City, Japan) on an Axioskop microscope (Zeiss, Jena). Camera and microscope were controlled by the computer program Openlab (Improvision, Conventry, UK). Cell aggregation was quantified using the index (N 0 −N t)/N 0 (2.Nagafuchi A. Takeichi M. EMBO J. 1988; 7: 3679-3684Crossref PubMed Scopus (664) Google Scholar).N 0 is the number of aggregates at time 0, andN t is the number of aggregates after timet. The average and standard deviation for three experiments are depicted in the graphical analysis. Previously we had shown that a 30-aa region within the mouse E-cadherin cytoplasmic domain (aa positions 833–862) is necessary and sufficient for the interaction with β-catenin (21.Stappert J. Kemler R. Cell Adhes. Commun. 1994; 2: 319-327Crossref PubMed Scopus (200) Google Scholar). This region contains a cluster of eight Ser residues, which were highly phosphorylated in vivo. Consensus sequence analysis for various serine/threonine kinases indicated that Ser-840, -853, and -855 fit the classical consensus motif SXX(E/D) for casein kinase II (CKII), whereas Ser-849 basically fit the consensus motif SXXXS(P) for glycogen synthase kinase-3β (GSK-3β) (Fig.1 A). The ability of CKII and GSK-3β to phosphorylate E-cadherin in vitro was tested on recombinant fusion protein containing the cytoplasmic domain of E-cadherin (ECT) fused N-terminal to GST. Both kinases were able to phosphorylate the recombinant GST-ECT fusion protein (Fig. 1B, right panel, autoradiogram). Whereas CKII was able to phosphorylate GST-ECT directly, the phosphorylation of GST-ECT by GSK-3β was dependent on pre-phosphorylation by CKII. This finding is in agreement with previous results demonstrating that some target proteins of GSK-3β require pre-phosphorylation four residues C-terminal to the GSK-3β site by a distinct protein kinase (e.g. CKII or protein kinase A). GST alone was not phosphorylated by CKII or GSK-3β (Fig. 1 B, left panel). Upon phosphorylation, no apparent change in the mobility of the recombinant GST-ECT fusion protein was detectable by SDS-PAGE. Coomassie staining confirmed that equal amounts of GST-ECT protein were used in all assays (Fig. 1 B, Coomassie blue). To test whether E-cadherin can be phosphorylated by protein kinases from whole cell lysates of mouse NIH3T3 fibroblasts, recombinant GST-ECT was incubated with the lysate in the presence of [γ-32P]ATP, and phosphorylation was analyzed by autoradiography. GST-ECT, but not GST, was strongly phosphorylated under these conditions (Fig. 1 C). To test whether this phosphorylation might be mediated by CKII or GSK-3β activity, the kinase assays were done in the presence of heparin, a specific inhibitor of CKII, or LiCl, which blocks GSK-3β activity (Fig.1 C). For both inhibitors we observed a strong reduction of the E-cadherin phosphorylation, indicating that CKII and GSK-3β are likely the major kinases involved in the phosphorylation of GST-ECT. However, a weak phosphorylation was still detectable even in the presence of both inhibitors, suggesting that additional kinases are also involved. The inhibition of CKII had a stronger effect on phosphorylation than the inhibition of GSK-3β, suggesting that phosphorylation of E-cadherin by GSK-3β may require pre-phosphorylation by CKII. The putative CKII and GSK-3β sites are located within the binding site of β-catenin in E-cadherin. As we have shown earlier, deletion of the entire 8 Ser cluster located here completely abolished the E-cadherin/β-catenin interaction, suggesting that this interaction may be modulated by phosphorylation. To test this directly, in vitro binding assays were performed with recombinant β-catenin and GST-ECT which was nonphosphorylated or pre-phosphorylated either with CKII alone or with CKII and GSK-3β together (Fig.2 A). For the binding analysis β-catenin was used at a concentration of 10 or 100 nm, both still in the non-saturating range (data not shown). GST-ECT-bound β-catenin was separated by SDS-PAGE and detected by Western blotting. As shown in Fig. 2 A, the relative amount of precipitated β-catenin was much higher when GST-ECT was phosphorylated. The amount of β-catenin bound to GST-ECT was increased by a factor of 4 if GST-ECT was pre-phosphorylated by CKII alone, and 6 times higher after phosphorylation by both CKII and GSK-3β (Fig. 2 B). To verify that in all assays equal amounts of GST-ECT had been used, the blots were also incubated with anti-GST antibodies (Fig. 2 A, lower panel). These experiments clearly demonstrated that the interaction of β-catenin and ECT is much stronger if the cytoplasmic domain of E-cadherin is phosphorylated and suggest that this event may also be an important regulatory process in vivo. To demonstrate unequivocally that the Ser residues at aa positions 840, 853, and 855 of mature E-cadherin are in fact target sites for CKII, these three sites were substituted by Ala residues, either individually (S840A, S853A, and S855A) or pairwise (S853A/S855A). Since these mutations should also influence the phosphorylation of GST-ECT by GSK-3β, the putative GSK-3β site at Ser-849 was not mutated separately. Mutant proteins were subjected toin vitro kinase assays with CKII and analyzed as above (Fig.3, A and B). All mutants showed a clear reduction in phosphorylation. Phosphorylation was completely abolished in the single mutant S855A and the double mutant S853/855A, and phosphorylation in S840A and S853A mutants was reduced to 43 and 28% that of wt GST-ECT. The results demonstrate that all putative CKII sites located within the β-catenin-binding site of E-cadherin are used by CKII in vitro, and the observed differences in the residual phosphorylation of the three mutants may reflect differences in the preferences of CKII. Mutants were also subjected to in vitro kinase assays with GSK-3β, and phosphorylation was found to be drastically reduced (data not shown), in good agreement with the model that pre-phosphorylation by CKII at Ser-855 and/or Ser-853 of E-cadherin is required before GSK-3β can phosphorylate at Ser-849. To confirm the results obtained in vitro, human 293 cells were transiently transfected with wt E-cadherin or with the double mutant E-cadherin form S853A/S855A. The cells were radioactively labeled with orthophosphate and subjected to an E-cadherin immunoprecipitation (Fig. 3 C). Under these conditions, phosphorylation of the E-cadherin double mutant S853A/S855A was reduced by 25% as compared with wt E-cadherin. This is in the range of what can be expected since E-cadherin is also a putative target for other serine/threonine kinases, e.g. protein kinase C or cAMP-dependent kinases, which were not blocked in this assay. The results described so far indicate that the Ser residues at aa positions 853/855 of E-cadherin are the major CKII target sites and that CKII phosphorylation directly influences β-catenin/E-cadherin interaction, at least in vitro. To extend this analysis, E-cadherin-negative mouse NIH3T3 fibroblasts expressing Wnt-1 were stably transfected with either wt E-cadherin or with each of the mutant forms of E-cadherin. Fibroblasts expressing Wnt-1 were used mainly for two reasons. First, it has been reported that in theDrosophila clone 8 cell line, CKII activity is up-regulated upon Wg signaling (27.Willert K. Brink M. Wodarz A. Varmus H. Nusse R. EMBO J. 1997; 16: 3089-3096Crossref PubMed Scopus (203) Google Scholar). Second, expression of Wnt-1 was shown to increase cadherin-mediated cell-cell adhesion in AtT20 cells and PC12 pheochromocytoma cells (28.Hinck L. Nelson W.J. Papkoff J. J. Cell Biol. 1994; 124: 729-741Crossref PubMed Scopus (381) Google Scholar, 29.Bradley R.S. Cowin P. Brown A.M. J. Cell Biol. 1993; 123: 1857-1865Crossref PubMed Scopus (232) Google Scholar). We therefore anticipated that the effects of CKII on E-cadherin/β-catenin interaction and cell-cell adhesion would be much more pronounced in NIH3T3 fibroblasts expressing Wnt-1. Immunoprecipitations were done on Wnt-1-positive NIH3T3 cells, stably transfected with wt E-cadherin or mutant E-cadherin. When precipitated complexes were separated by SDS-PAGE and analyzed by Western blotting with antibodies against β-catenin and E-cadherin (Fig. 4), β-catenin was found to be strongly associated with E-cadherin. More importantly, the association between β-catenin and each of the four mutant E-cadherins was greatly reduced, with the double mutant form S853A/S855A showing the most striking effect. Hence, substitution of the two CKII sites Ser-853 and Ser-855 leads to a drop in the β-catenin/E-cadherin interaction not only in vitro but also in live cells. As a direct consequence of this reduced interaction between E-cadherin and β-catenin in vivo, the E-cadherin-mediated cell-cell adhesion should be diminished as well. The effect on cell-cell adhesion was tested by examining wt E-cadherin or S853A/S855A double mutant E-cadherin-expressing Wnt-1 NIH3T3 fibroblasts in a cell-cell adhesion assay (Fig. 5). Expression of wt E-cadherin led to an increase in cell-cell adhesion (squares), whereas the adhesion of mutant E-cadherin-expressing E-cadherin was significantly reduced (triangles). The relatively strong adhesion already observed in E-cadherin-negative Wnt-1-expressing NIH3T3 fibroblasts is presumably due to the presence of N-cadherin in those cells, since basal cell adhesion of NIH3T3 Wnt-1-negative cells was much lower. Nonetheless, cell adhesion was increased when these cells were transfected with wt E-cadherin and reduced if they were transfected with mutated E-cadherin (data not shown). In summary, the expression of the E-cadherin double mutant S853A/S555A in NIH3T3 cells not only reduces the interaction with β-catenin but prevents cell-cell adhesion as well.Figure 5Expression of the E-cadherin double mutant S853A/S855A in NIH3T3 cells expressing Wnt-1 reduces cell-cell adhesion. Cell adhesion assay of Wnt-1 NIH3T3 cells either untransfected or stably expressing wt or mutant E-cadherin. 3–6 × 105 cells were incubated for the given time at 37 °C. Expression of the double mutant S853A/S855A in NIH3T3 cells leads to a clear reduction of E-cadherin-mediated cell adhesion.View Large Image Figure ViewerDownload Hi-res image Download (PPT) Taken together, from the data presented here and from interaction studies of other binding partners for β-catenin, it seems evident that Ser/Thr phosphorylation is a common mechanism by which the binding to β-catenin is modulated. The clustering of serines within the binding region for β-catenin seems to be a characteristic feature for at least some interaction partners of β-catenin. For example, a phosphorylation-dependent binding of β-catenin was also found for two other binding partners, the tumor suppressor protein APC and the DF3/MUC1 breast carcinoma-associated antigen (22.Rubinfeld B. Albert I. Porfiri E. Fiol C. Munemitsu S. Polakis P. Science. 1996; 272: 1023-1026Crossref PubMed Scopus (1310) Google Scholar, 23.Yamamoto M. Bharti A. Li Y.Q. Kufe D. J. Biol. Chem. 1997; 272: 12492-12494Abstract Full Text Full Text PDF PubMed Scopus (284) Google Scholar). APC plays a central role in the ubiquitin-mediated degradation of β-catenin and has been shown to become phosphorylated by GSK-3β upon Wnt signaling (30.Morin P.J. Sparks A.B. Korinek V. Barker N. Clevers H. Vogelstein B. Kinzler K.W. Science. 1997; 275: 1787-1790Crossref PubMed Scopus (3517) Google Scholar). The APC protein contains several binding sites for β-catenin. Those sites, necessary for the degradation of β-catenin, were found to bind β-catenin much more strongly if pre-phosphorylated by GSK-3β in vitro (22.Rubinfeld B. Albert I. Porfiri E. Fiol C. Munemitsu S. Polakis P. Science. 1996; 272: 1023-1026Crossref PubMed Scopus (1310) Google Scholar). The β-catenin-binding sites of APC and DF3/MUC1 contain the Ser motif SXXXXXSSL, which closely resembles the sequence SXXXXXSSE of E-cadherin, which harbors the CKII phosphorylation site 853. The serines localized in the β-catenin-binding site of E-cadherin and mutated here are conserved among all classical cadherins, supporting the idea that these residues play an essential role for β-catenin binding. The participation of CKII in regulating E-cadherin/β-catenin interaction may be deduced from the increased cell adhesion observed in Wnt-1-transfected ATt20 and PC12 cells (28.Hinck L. Nelson W.J. Papkoff J. J. Cell Biol. 1994; 124: 729-741Crossref PubMed Scopus (381) Google Scholar, 29.Bradley R.S. Cowin P. Brown A.M. J. Cell Biol. 1993; 123: 1857-1865Crossref PubMed Scopus (232) Google Scholar). Considering that CKII activity was found to be increased in Drosophila cells receiving a Wg signal (27.Willert K. Brink M. Wodarz A. Varmus H. Nusse R. EMBO J. 1997; 16: 3089-3096Crossref PubMed Scopus (203) Google Scholar), CKII activity may also be enhanced in Wnt-1-expressing mouse NIH3T3 fibroblasts, thereby strengthening cell adhesion. Another hint that phosphorylation may be in fact a general mechanism to modulate the interaction between β-catenin and its binding partners comes from the crystal structure analysis of the central repeat region of β-catenin (31.Huber O. Krohn M. Kemler R. J. Cell Sci. 1997; 110: 1759-1765Crossref PubMed Google Scholar). The core region of β-catenin is composed of 12 copies of the 42-amino acid Armadillo repeat motif. Through these repeats, β-catenin interacts with E-cadherin and APC (32.Hulsken J. Birchmeier W. Behrens J. J. Cell Biol. 1994; 127: 2061-2069Crossref PubMed Scopus (586) Google Scholar). The three-dimensional structure of this region was found to form a superhelix of helices that features a long, positively charged groove. From this it was hypothesized that proteins with acidic regions could interact with the basic groove of β-catenin. In fact the β-catenin-binding region is highly acidic in E-cadherin, APC, and LEF-1 (a DNA binding factor of the high mobility group box protein family). As mentioned before, phosphorylation of APC by GSK-3β was found to increase the interaction between APC and β-catenin, and it was concluded that interaction between β-catenin and its binding partner, in this case APC, would be enhanced by the phosphorylation of the β-catenin-binding region, which would make this region even more acidic. The increased binding to β-catenin upon phosphorylation of E-cadherin presented here confirms this conclusion. The biological significance of our results is also underlined by an E-cadherin mutant identified in a human ovarian tumor (33.Risinger J.I. Berchuck A. Kohler M.F. Boyd J. Nat. Genet. 1994; 7: 98-102Crossref PubMed Scopus (260) Google Scholar). Here the Ser at aa position 838 of human E-cadherin, which corresponds to Ser 840 of mouse E-cadherin, has been replaced by Ala due to an A to G transition. Thus, phosphorylation of the β-catenin-binding region of E-cadherin by CKII and/or GSK-3β appears to be an essential mechanism to modify cell-cell adhesion. We gratefully acknowledge the expert technical assistance of Andreas Rolke. We thank Dr. Randy Cassada for critical reading the manuscript.
Referência(s)