Galectin-3, a Novel Binding Partner of β-Catenin

Introduction

Wnt proteins constitute a large family of cysteine-rich molecules that control development in organisms ranging from nematode worms to mammals (1) . The intracellular signaling pathway of Wnt is evolutionally conserved and regulates cellular proliferation, morphology, motility, axis formation, and organ development (2 , 3) . β-Catenin is a downstream component of the Wnt signaling pathway. It is phosphorylated by GSK3β and CKIα in a complex that contains APC and axin, targeting β-catenin for ubiquitination and degradation (4, 5, 6) . In the current model, dephosphorylation of β-catenin leads to its accumulation and translocation into the nucleus, where it binds to the transcriptional factor Tcf/Lef and serves as a coactivator of Tcf/Lef to stimulate transcription of the Wnt target genes including c-myc and cyclin D₁, among others (7 , 8) . However, the molecular mechanism that targets β-catenin to the nucleus is still unclear. Galectins are a family of carbohydrate-binding proteins characterized by conserved amino acid sequences defined by structural similarities in their carbohydrate-binding domains and affinity for β-galactoside–containing glycoconjugates (9) . An important member of this family, e.g., gal-3, has been implicated in diverse biological functions including cell growth, differentiation, apoptosis, adhesion, malignant transformation, and RNA processing (10, 11, 12, 13) . Similarly to β-catenin, endogenous gal-3 is found in the cytoplasm, on the cell surface, and in the nucleus, where it undergoes post-translational modification by the addition of a phosphate group catalyzed by CK1. Its phosphorylation serves as a signaling switch for nuclear export. In vivo, the nuclear expression of gal-3 is associated with tumor invasion and metastatic potential in various human cancers such as colon, prostate, and tongue squamous carcinoma cells (11 , 14, 15, 16) .

We have previously shown that gal-3 overexpression regulates changes in the expression levels of cell cycle regulators, including cyclin D₁ (13) , and because cyclin D₁ has been reported to be targeted within the Wnt pathway (7) , we questioned the possible association between gal-3 and β-catenin for molecular shuttling to the nucleus of cancer cells. We propose that gal-3 complex with β-catenin activate Tcf-reporter activity and stimulate cyclin D₁ and c-myc.

Materials and Methods

Cells, Antibodies, and Culture Conditions.

The human breast cancer BT549 gal-3 null cell line, its transfectants, and culture conditions are as described were described previously (17) . The colon cancer cell line HT29 contains mutated APC and wild-type β-catenin, whereas HCT116 contains mutated β-catenin and wild-type APC; they were purchased from American Type Culture Collection (Manassas, VA). Cells were grown in Dulbecco’s modified Eagle’s medium supplemented with 10% heat-inactivated fetal bovine serum, 2 mmol/L glutamine, nonessential amino acids, and antibiotics (Invitrogen, Carlsbad, CA). Suspension cultures were performed as described previously (13) . Monoclonal rat anti-gal-3 antibody was purchased (TIB166; American Type Culture Collection). Polyclonal rabbit anti-gal-3 was described elsewhere (10) . Anti-β-catenin antibody for Western blot and immunofluorescence (BD Biosciences, San Jose, CA), anti-β-catenin antibody for immunoprecipitation (Upstate, Lake Placid, NY), anti-cyclin D₁ antibody (Sigma, St. Louis, MO), anti-c–myc antibody (Oncogene, San Diego, CA), and anti-Tcf–4 antibodies (Exalpha Biologicals, Watertown, MA).

Plasmid Construction.

pGEX-2T/β-catenin, pGEX-2T/β-catenin (1–131), pGEX-2T/β-catenin (132–423), pGEX-KG/β-catenin (423–781), pGEX-2T/β-catenin (1–423), and pGEX-2T/β-catenin (175–423) were described elsewhere (6) . Production of recombinant glutathione S-transferase (GST)-fusion protein was conducted according to the manufacturer’s recommendation (Amersham Biosciences Corp., Piscataway, NJ). PGEX-6P-2/gal-3 was also described elsewhere. To establish the deletion mutant of gal-3, a Quick Change Site-directed Mutagenesis kit (Stratagene, La Jolla, CA) was used, using pGEX-6P-2/gal-3 as a template. To construct Gal-3 deletion mutant 1–62, 1–93, and 1–142, tyrosine⁶³, glycin⁹⁴, and asparagine¹⁴³ were substituted to the stop codon, TAG, using specific primers as follows: sense for 1–62, 5′-CAGGCACCTCCAGGCGCCTAGCATGGAGCACCTGGAGCTTATCCC-3′; antisense, 5′-GGGATAAGCTCCAGGTGCTCCATGCTAGGCGCCTGGAGGTGCCTG-3′; sense for 1–93, 5′-GGGGCCTACCCATCTTCTGGATAGCCAAGTGCCCCCGGAGCCTAC-3′; antisense, 5′-GTAGGCTCCGGGGGCACTTGGCTATCCAGAAGATGGGTAGGCCCC-3′; sense for 1–142, 5′-GAGTCATTGTTTGCAATACATAGCTGGATAATAACTGGGGAAGGG-3′; and antisense, 5′-CCCTTCCCCAGTTATTATCCAGCTATGTATTGCAAACAATGACTC-3′. To check accuracy of site-directed mutation, produced plasmids were sequenced at Wayne State University DNA Sequencing Core (Detroit, MI). pGEX-6P-2/gal-3 (62–250) was constructed by an insertion of an EcoRI-EcoRI fragment coding cleaved human gal-3 sequence into the EcoRI site in the pGEX-6p-2. To generate cleaved gal-3 fragment coding the COOH-terminal domain (NH₂-terminal 62 amino acids deleted) of human gal-3, a sense primer (5′-CAGAATTCTGTTATGTACCATGGAGCACCTG-3′) and an antisense primer (5′-CCTGAATTCAGATTATATCATGGTATATGAAGC-3′) containing a flanking EcoRI cleavage site were synthesized using plasmid pBK-CMV-Gal-3 as a template. A PCR product was purified and inserted into EcoRI-digested vector pGEX-6p-2 [designated as pGEX-6P-2/Gal-3 (63–250)].

Immunoprecipitation.

To determine whether gal-3 forms a complex with β-catenin, immunoprecipitation was performed. Cell supernatants were immunoprecipitated with anti-gal-3, anti-β-catenin, or anti-Tcf–4 antibody for 60 minutes at 4°C. Then 30 μL of 1:2 slurry of protein G-Sepharose were added for 60 minutes at 4°C. After centrifugation, the precipitate was mixed with 20 μL of SDS-PAGE sample buffer and boiled. In all immunoprecipitation, we confirmed that no coimmunoprecipitants were detected when precipitated with control normal serum.

Luciferase Assay.

To examine whether galectin overexpression augments transcriptional activity of the Tcf-reporter plasmid, BT549 parental cells were transiently transfected with pcDNA3.1+/Zeo/gal-3 (1 μg), cotransfected with 0.5 μg of pTOPFLASH or pFOPFLASH (Upstate). Forty-eight hours after transfection, the cells were collected and cultivated on poly-HEMA–coated dishes for the indicated time. After lysis with reporter lysis buffer, luciferase activity was measured using a Luciferase Assay System (Promega, Madison, WI) and luminophotometer TD-20/20 (Turner Designs, Sunnyvale, CA). To observe whether transiently transfected gal-3 also augments Tcf-reporter activity in HT29, pcDNA3.1+/Zeo/gal-3 was transiently transfected into HT29 cells and cotransfected with pTOPFLASH or pFOPFLASH (0.5 μg). After 48 hours, the cells were lysed, and activity was measured using a Luciferase Assay System and luminophotometer TD-20/20. The results were calculated as pTOPFLASH to pFOPFLASH ratio.

Mapping the Binding Region between β-Catenin and Galectin-3.

For determination of the region of β-catenin that binds to gal-3, various deletion mutants of GST–β-catenin (each at 250 nmol/L) were incubated with 250 nmol/L gal-3 (full length) for 1 hour at 4°C in 50 μL of reaction mixture [20 mmol/L Tris-HCl (pH 7.5) and 1 mmol/L dithiothreitol). GST–β-catenin deletion mutants were precipitated with glutathione-Sepharose 4B and probed with anti-gal-3 antibody. Reciprocally, deletion mutants of gal-3 protein (250 nmol/L) were incubated with 250 nmol/L β-catenin for 1 hour at 4°C in 50 μL of reaction mixture. The mixture was precipitated with anti-β-catenin antibody and probed with anti-gal-3 antibodies.

Results and Discussion

β-Catenin–Galectin-3 Interaction.

Previously, it was reported that gal-3 induces cyclin D1 promoter activity and up-regulates cyclin D₁ expression in suspension culture. It was suggested that this regulation is through enhancement/stabilization of cAMP-responsive element-associated complex (13) . These findings prompted us to examine a possible link between Wnt pathway and gal-3, because both cyclin D₁ and c-myc are targets of this pathway. Thus, we have activated gal-3 by anoikis and found an induction of transcriptional activity of β-catenin–dependent genes, e.g., cyclin D₁ and c-myc only in cells expressing gal-3 (Fig. 1, II) ⇓ . To explore whether gal-3 forms a complex with β-catenin, the lysates of BT549-Gal cells from monolayer clusters were immunoprecipitated with either anti-gal-3 or anti-β-catenin antibodies and reciprocally blotted with anti-β-catenin or anti-gal-3 antibodies. When the lysates of the cells were pulled down with antibody directed against gal-3, separated by electrophoresis, and Western analyzed with anti-β-catenin antibody, β-catenin associated with gal-3 was detected (Fig. 2A) ⇓ . In reciprocal experiments, whereby the lysates were immunoprecipitated with anti-β-catenin antibody and Western blot analysis was done with anti-gal-3 antibody, gal-3–associated β-catenin was detected (Fig. 2B) ⇓ . Next, we questioned whether the gal-3–β-catenin interaction is carbohydrate dependent. To this end, we tested lactose, a specific gal-3 inhibitor. Formation of the complex between β-catenin and gal-3 was inhibited by low (50 mmol/L) lactose concentration present in the lysis buffer; specificity of the inhibition was confirmed by using 50 mmol/L sucrose as a control (Fig. 2C) ⇓ . It should be noted that β-catenin is not a glycoprotein, suggesting that specific sugar inhibition could have resulted from a conformational change of gal-3 due to occupancy of the sugar-binding site. These findings are supported by the observation that a naturally occurring complex carbohydrate rich in galactose, i.e., pectin, inhibits β-catenin expression in the colon (18) . This might explain the inhibitory effect of modified pectin on the growth and metastasis of human breast carcinoma in nude mice (19) , possibly due to inhibition of gal-3–β-catenin complex formation.

We also confirmed that β-catenin–gal-3 colocalized in vivo by immunofluorescence (Fig. 2D) ⇓ . Confocal microscopic analysis revealed that in BT549-vCTR cells (gal-3 null; Fig. 2D, b ⇓ ), β-catenin showed no nuclear localization and is localized in the cytoplasm and at the membrane (Fig. 2D, a and c) ⇓ , whereas in cells expressing gal-3, both β-catenin and gal-3 proteins were distributed throughout the cytoplasm and the nucleus (Fig. 2D, a and b) ⇓ and were colocalized in the nucleus (Fig. 2D, f) ⇓ . Of note, loss of nuclear gal-3 expression is associated with tumor progression of colon carcinoma (14) ; in prostate carcinomas, gal-3 is excluded from cell nuclei (16) , and in tongue squamous carcinoma cells, the levels of nuclear gal-3 markedly decreased during the progression from normal to cancerous states, whereas the cytoplasmic expression level increased (15) . The nuclear presence of gal-3 is consistent with its role in the regulation of transcription, pre-mRNA splicing, and transport processes (11, 12, 13) . Because of its small size, it is plausible that gal-3 shuttles freely between the cytosol and the nucleus. Gal-3 could be involved in the nuclear retention of β-catenin, leading to the activation of the Wnt-targeted genes. However, because both proteins lack an NLS sequence (19) , it is still unclear whether they independently import/export to/from the nucleus.

Tcf-4 and Galectin-3 Interaction.

Based on the data presented above, we questioned (1) whether the β-catenin–gal-3 complex also contains Tcf-4 protein and (2) whether β-catenin–gal-3 complex formation is affected by either APC or β-catenin amino acid mutations. To address these, we examined two human colon cancer cell lines: (1) the HT29 cell line, which contains mutated APC and wild-type β-catenin proteins, and (2) the HCT116 cell line, which harbors β-catenin mutation and wild-type APC proteins. Pull-down immunoprecipitations followed by Western analyses revealed that both β-catenin and gal-3 complex could be detected in the cell lysates of both cell lines (Fig. 3) ⇓ . β-Catenin was pulled down by antigalectin antibodies (Fig. 3, I ⇓ , Lanes 1 and 2), and gal-3 was pulled down by anti-β-catenin antibodies (Fig. 3, I ⇓ , Lanes 3 and 4). Gal-3 pulled-down Tcf-4 (Fig. 3, I ⇓ , Lane 5) and reciprocally was pulled down by anti-Tcf-4 antibodies (Fig. 3, I ⇓ , Lane 6). Thus, we report that gal-3 forms a complex with β-catenin independent of either APC or β-catenin mutations or that in the cells tested here, gal-3 forms a ternary complex with β-catenin and Tcf-4.

To determine whether gal-3 affects β-catenin–Tcf signaling, we examined its effect on Tcf-reporter activity. BT549 parental cells (gal-3 null) were transiently transfected with pcDNA3.1+/Zeo/gal-3 (1 μg) and collected, and the activity of Tcf-reporter plasmid was assayed. The pTOPFLASH/pFOPFLASH ratio after transient transfection was enhanced up to 5-fold (Fig. 3, II) ⇓ Moreover, the ratio was further enhanced up to 13-fold when the cells were cultured in suspension (Fig. 3, II) ⇓ , a condition that up-regulates cyclin D1 and c-myc expression (Fig. 1) ⇓ . In addition, we questioned whether gal-3 overexpression may augment Tcf-transcriptional activity in cells that have activated Tcf due to APC mutation. HT29 cells were transfected with pcDNA3.1+/Zeo/gal-3 and cotransfected with pTOPFLASH or pFOPFLASH. Luciferase activities were enhanced up to 3.8-fold in a dose-dependent manner (Fig. 3, II, B) ⇓ .

β-Catenin– and Galectin-3–Binding Domains.

Next, we attempted to determine the target motifs in β-catenin and gal-3 sequences that are required for their molecular interaction. We performed in vitro pull-down assays with GST proteins and peptides. We have generated various deletion mutant proteins of β-catenin (Fig. 4A) ⇓ and gal-3 (Fig. 4B) ⇓ , spanning the entire sequence. The proteins were expressed and purified as GST fusion proteins. The GST–fusion-β-catenin mutant proteins were incubated with full-length gal-3 protein and precipitated thereafter with glutathione-Sepharose. As shown in Fig. 4C ⇓ , gal-3 was coprecipitated with GST–β-catenin (full length), GST–β-catenin (1–423) and (1–131) peptides, but not with either GST–β-catenin (132–423) and GST–β-catenin (175–423) or GST–β-catenin (423–781) peptides. Reciprocally, to determine the motif of gal-3 that binds to β-catenin, full-length β-catenin protein and deleted peptides were incubated with gal-3 and immunoprecipitated thereafter (Fig. 4D) ⇓ . β-Catenin was coprecipitated with intact recombinant gal-3 (full length) and gal-3 (63–250)-deleted peptide but not with gal-3 (1–62), gal-3 (1–92), or gal-3 (1–142) peptides. Thus, we concluded that the NH₂ terminus of β-catenin interacts with the COOH terminus of gal-3 encompassing amino acid residues 1–131 and 63–250, respectively. Of note, this structural domain contains the β-galactoside–binding motif of gal-3 and thus might explain the observed inhibitory effect of lactose on β-catenin–gal-3 complex formation (Fig. 2C) ⇓ . The β-catenin region (1–131 amino acids), which is also involved in its interaction with α-catenin and is prerequisite for oncogenicity (20) , might suggest that a competition between gal-3 and α-catenin for β-catenin binding leads to its the displacement from the membrane.

Gal-3 has been implicated in adhesion-mediated processes, cell cycle regulation, apoptosis resistance, tumorigenesis, and metastasis. However, the molecular mechanism(s) that regulates these functions are not fully deciphered. The findings presented here revealed a molecular interaction of gal-3 with β-catenin/Tcf complex, and the similarities in the pleiotrophic functions of both β-catenin and gal-3 should provide an insight to the functional regulation of both molecules, leading to a better understanding of the molecular machinery of the Wnt signaling pathway.