Galectin-3 (gal-3), a member of the β-galactoside–binding proteins family, was identified as a binding partner of β-catenin. Analysis of the human gal-3 sequence reveled a structural similarity to β-catenin as it also contains the consensus sequence (S92XXXS96) for glycogen synthase kinase-3β (GSK-3β) phosphorylation and can serve as its substrate. In addition, Axin, a regulator protein of Wnt that complexes with β-catenin, also binds gal-3 using the same sequence motif identified here by a deletion mutant analysis. The data presented here give credence to the suggestion that gal-3 is a key regulator in the Wnt/β-catenin signaling pathway and highlight the functional similarities between gal-3 and β-catenin.
- glycogen synthase kinase-3β
- casein kinase I
Galectins are a family of carbohydrate-binding proteins characterized by conserved amino acid sequences of their carbohydrate-binding domains and affinity for β-galactoside–containing glycoconjugates ( 1). Galectin-3 (gal-3) exhibits pleiotropic biological functions and has been implicated in cell growth, differentiation, apoptosis, adhesion, malignant transformation, and RNA processing ( 1– 4).
Previously, we have reported that gal-3 overexpression regulates the expression levels of cell cycle targets of Wnt pathway, like cyclin D1 and c-myc ( 5– 7), and found that gal-3 is a novel binding partner of β-catenin and is phosphorylated, like β-catenin, by casein kinase I (CKI; ref. 7– 9). β-catenin is phosphorylated by a dual kinase system of CKIα and glycogen synthase kinase-3β (GSK-3β) in a complex containing adenomatous polyposis coli and axin, targeting β-catenin for ubiquitination and degradation ( 10– 14). Gene mutations in APC, axin, or β-catenin augment phosphorylation, which, in turn, leads to its accumulation in the nucleus, resulting in activation of transcription of Wnt-target genes ( 11, 14). Because the nuclear import-export of gal-3, like that of β-catenin, is phosphorylation dependent ( 9), we questioned whether gal-3 may also be phosphorylated by a similar dual kinase system. A search of the human gal-3 protein amino acid sequence revealed that in addition to a CKI phosphorylation site (Ser6; ref. 9), it also contains the GSK-3β phosphorylation consensus sequence (S92XXXS96). This prompted us to question whether gal-3 is phosphorylated by GSK-3β and whether gal-3 will bind Axin in a phosphorylation-dependent manner.
Materials and Methods
Cells and reagents. The origin and the culture conditions of the human breast cancer cell line BT549, the Gal-3–transfected cell clones (BT549-Gal), and the control transfectants (BT549-vCTR) were as described ( 15, 16). TIB166, monoclonal rat anti–gal-3, was purchased from American Type Culture Collection (Manassas, VA). hL31, anti–gal-3 antibody, was obtained as described ( 7). Anti-HA antibody was purchased from Cell Signaling Technology, Inc. (Beverly, MA) and anti-phospho-serine antibody from Sigma (St. Louis, MO). Recombinant GSK3β was purchased from New England BioLabs, Inc. (Beverly, MA).
Plasmid construction. pGEX-2T/rat Axin (rAxin; 298-832), pGEX-2T/rAxin (298-506), pGEX-2T/rAxin (1-529), pGEX-2T/rAxin (508-732), and pGEX-2T/rAxin (713-832) were described elsewhere ( 7). pcDNA3.1+/Zeo-HA-Axin was kindly provided by Dr. Shuichi Kusano (St. Marianna University School of Medicine, Kawasaki, Japan). pGEX-6P-2/gal-3 and gal-3 deletion mutants were as described ( 7). Production of recombinant glutathione S-transferase (GST) fusion proteins were produced and purified according to the manufacturer's instruction (Amersham Biosciences, Piscataway, NJ).
Immunoprecipitation. To determine whether gal-3 forms a complex with axin, BT549-Gal cells were transiently transfected in a 100-mm-diameter dish with pcDNA 3.1+/Zeo-HA-Axin using LipofectAMINE 2000 (Invitrogen, Carlsbad, CA). Twenty-four hours after transfection, cells were lysed in 800 μL ice-cold lysis buffer (procedural and technical detail as in ref. 7). The supernatant (200 μg protein) were immunoprecipitated with anti–gal-3 or anti-HA antibodies for 60 minutes at 4°C. Protein separations and identification were as described ( 7).
Mapping the Axin–galectin-3 binding region. Various deletion mutants of GST-β-rAxin (each at 250 nmol/L) were incubated with 250 nmol/L gal-3 (full length) for 1 hour at 4°C in 50 μL reaction mixture [20 mmol/L Tris-HCl (pH 7.5) and 1 mmol/L DTT]. GST-Axin deletion mutants were precipitated with glutathione-Sepharose 4B, and then the precipitates were probed with anti–gal-3 antibody. To examine the region of gal-3 that binds to rAxin, various deletion mutants of gal-3 (250 nmol/L each) were incubated with 250 nmol/L of GST-rAxin (298-832) for 1 hour at 4°C in 50 μL reaction mixture. GST-rAxin deletion mutants were precipitated with glutathione-Sepharose 4B and probed with the anti–gal-3 antibody.
In vitro kinase assay. To examine whether gal-3 is phosphorylated by GSK-3β, in vitro kinase assay was done. Purified gal-3 (1 μg protein) was incubated in a kinase buffer containing 10 μCi [γ-32P]ATP with indicated units of GSK-3β in the presence or absence of Axin (indicated amount). LiCl (30 mmol/L), a specific GSK-3β inhibitor, was used to establish specificity. The reaction was stopped by boiling, followed by SDS-PAGE and autoradiography ( 9).
Results and Discussions
Axin–galectin-3 interaction. Previously, it was reported that the β-catenin–Axin association promotes β-catenin phosphorylation by GSK3β ( 12, 13), whereas we found that gal-3 is a binding partner of β-catenin ( 7). Because human gal-3 contains the consensus sequence motif of GSK3β phosphorylation (S92XXXS96), we questioned whether gal-3 could bind Axin and be phosphorylated by GSK-3β. Thus, BT549-Gal cells were transiently transfected with HA-Axin cDNA, lysed, and immunoprecipitated with either anti-HA ( Fig. 1A, right; B, left ) or anti–gal-3 antibodies ( Fig. 1A, left; B, right) and Western blotted with the reciprocal antibodies ( Fig. 1A and B, bottom). The results show that Axin coprecipitated with gal-3 and vice versa, suggesting that the two are complexed in vivo. Next, we confirmed that Axin–gal-3 are indeed colocalized in vivo by immunofluorescence ( Fig. 1C) by using confocal microscopy analysis of BT549 (gal-3 null; Fig. 1C a, b, and c) parental and BT549-Gal ( Fig. 1C d, e, and f) cells transiently transfected with HA-Axin ( Fig. 1C). In cells expressing both Axin and gal-3, colocalization of the two proteins is readily observed ( Fig. 1C, f). Subsequently, we have constructed and expressed deletion mutants of Axin ( Fig. 2A and B ) and gal-3 ( Fig. 2D and E) to assist in determining the Axin–gal-3 interacting motifs necessary for their interaction. The Axin mutant peptides were purified as GST fusion proteins and gal-3 was precipitated with GST-Axin (1-528), GST-Axin (298-508), and GST-Axin (298-832) peptides but not with GST-Axin (1-229), GST-Axin (508-832), or GST-Axin (713-832) peptides ( Fig. 2C). In the reciprocal experiments ( Fig. 2F), only gal-3 (full length) and gal-3 (63-250) peptides were recognized by GST-Axin (298-508) peptide. Thus, we have concluded that the internal domain of Axin interacts with the COOH terminus of gal-3, encompassing amino acid residues 298-508 and 143-250, respectively.
GSK-3β phosphorylates galectin-3. Because the human gal-3 contains a GSK-3β phosphorylation consensus sequence (S92XXXS96), we first questioned its substrate suitability utilizing an in vitro kinase assay. Following incubation of gal-3 with or without GSK-3β in a reaction mixture containing [γ-32P]ATP, we found that gal-3 was phosphorylated by GSK-3β and that the phosphorylation was specifically inhibited by a GSK-3β inhibitor, e.g., LiCl ( Fig. 3A ). Similar to β-catenin whereby Axin enhances its GSK-3β–dependent phosphorylation ( 10– 13), Axin promoted the GSK-3β–dependent phosphorylation of gal-3 ( Fig. 3A). Of note, the phosphorylation of gal-3 by GSK-3β was specific in a time- and dose-dependent manner ( Fig. 3B and C). The above results and the previous data ( 7, 9) prompted the proposed model ( Fig. 4 ) that revises the Wnt/β-catenin signaling pathway to include gal-3.
It was surprising to find that both gal-3 and β-catenin are substrates of CKI and GSK-3β. As for gal-3, phosphorylation of Ser6 by CKI serves as a molecular switch for the sugar binding ( 17) and regulation of nuclear export ( 9); unlike the phosphorylation of β-catenin by CK1α and GSK3β that promotes its proteosomal degradation, the consequence of GSK-3β gal-3 phosphorylation mediated by Axin is yet to be determined. In addition, we will need to resolve whether the phosphorylation of gal-3 affects the status of phosphorylation of β-catenin or vice versa and to establish whether gal-3 engages/mediates/inhibits the ubiquitination/signaling of β-catenin.
Grant support: National Cancer Institute, NIH, grant CA46120 (A. Raz).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Received January 13, 2005.
- Revision received February 14, 2005.
- Accepted March 1, 2005.
- ©2005 American Association for Cancer Research.