γ-Catenin expression and Tcf/Lef transcription activation

Our previous work established that β-catenin proteins with mutations in presumptive amino-terminal phosphorylation sites promoted neoplastic transformation of RK3E, an adenovirus E1A-transformed rat kidney epithelial line, whereas wild-type β-catenin failed to function as an oncogene in RK3E (Kolligs et al. 1999). The finding that APC regulated γ-catenin motivated us to pursue studies of γ-catenin's transforming activity in RK3E. We first generated expression constructs encoding wild-type γ-catenin and deletion mutants (Fig. ​(Fig.2A).2A). The rationale for creating the ΔN38γ, ΔN122γ, ΔArmγ, and ΔC687γ deletion mutants was based on the conserved structures of γ- and β-catenin and prior data on the oncogenic activity of various β-catenin deletion mutants (Kolligs et al. 1999). In addition, because the only γ-catenin mutation reported in cancer was a missense mutation of a potential phosphorylation site at the amino terminus (codon 28, serine-to-leucine mutation) (Caca et al. 1999), we also prepared an expression construct encoding the S28L mutant.

Open in a separate window

Open in a separate window

Open in a separate window

Figure 2

Expression of wild-type and mutated γ- and β-catenin proteins and their effects on Tcf transcription in 293 cells. (A) Schematic illustration of the structure of 745-amino-acid γ-catenin and 781-amino-acid β-catenin proteins; the percentage identity of sequences in different regions of the proteins; and the γ- and β-catenin proteins encoded by the expression constructs used in this work. The 12 armadillo repeats in the highly conserved central regions of γ- and β-catenin are indicated by open boxes. In addition to wild-type γ- and β-catenin, several mutant proteins were expressed. Asterisks (*) indicate the position of point mutations (i.e., S33Yβ and S28Lγ), and the thin line indicates the region of γ-catenin affected by an in-frame deletion in the ΔArmγ construct. All γ- and β-catenin proteins were modified to contain carboxy-terminal Flag epitope tags. (B) Expression of wild-type and mutated γ- and β-catenin proteins. ECL–Western blot analysis with an anti-Flag antibody was carried out on whole-cell lysates prepared 48 hr after transfection of 293 cells with pcDNA3 constructs encoding the γ- and β-catenin proteins. To confirm equal loading and transfer of proteins, blots were stripped and ECL–Western blotting with an anti-actin antibody was performed. (C) Activation of Tcf transcription by wild-type and mutated forms of γ- and β-catenin in 293 cells following transfection of pcDNA3 constructs. The ratios of luciferase activities from the wild-type Tcf reporter (pTOPFLASH) and a mutant Tcf reporter gene construct (pFOPFLASH) were determined 48 hr after transfection. Luciferase activities were normalized for transfection efficiency by cotransfection with a β-galactosidase-expressing vector. Tcf activation relative to that obtained with empty pcDNA3 vector is shown. The mean values and standard deviations from three experiments are shown.

The expression of wild-type and mutated forms of γ-catenin was compared with that of wild type β-catenin and a mutated form carrying a cancer-derived missense substitution (S33Y) in β-catenin's presumptive GSK3β phosphorylation sequences, using Western blotting to detect the carboxy-terminal Flag epitope tag present on each γ- and β-catenin protein. Whereas a single missense mutation or in-frame deletions of the presumptive amino-terminal phosphorylation sites consistently stabilized the β-catenin protein (Fig. ​(Fig.2B;2B; Kolligs et al. 1999), the S28L mutation did not increase γ-catenin expression over the levels of the wild-type protein (Fig. ​(Fig.2B).2B). However, larger in-frame deletions of the amino terminus led to increased γ-catenin expression (Fig. ​(Fig.2B,2B, ΔN38 and ΔN122 mutants), implicating γ-catenin's amino terminus in the regulation of its degradation. Deletion of armadillo repeats 3–8 also led to increased γ-catenin expression (Fig. ​(Fig.2B,2B, ΔArm), consistent with the notion that the APC and conductin/axin proteins bind sequence in this region of γ-catenin and regulate its degradation.

The ability of the various γ-catenin proteins to activate Tcf/Lef transcription was then assessed, using a model reporter gene construct containing three Tcf-binding sites upstream of a minimal c-Fos-derived promoter element and luciferase coding sequences (i.e., TOPFLASH). As in prior studies (Kolligs et al. 1999), the S33Y mutated form of β-catenin had about 8–10× the activity of wild-type β-catenin (Fig. ​(Fig.2C).2C). In contrast, only roughly 30%–50% increases in Tcf/Lef activation were seen for the ΔN38 and S28L mutated forms of γ-catenin, respectively, when compared with wild-type γ-catenin (Fig. ​(Fig.2C).2C). These three γ-catenin proteins all had less activity than wild-type β-catenin in the assay, though the activity of the S28L mutant approached that of wild-type β-catenin. Larger deletions of the amino terminus of γ-catenin (e.g., ΔN122) as well as deletions of armadillo repeats 3–8 or the carboxyl terminus substantially inhibited or completely abrogated γ-catenin's ability to activate Tcf/Lef transcription (Fig. ​(Fig.2C).2C). Taken together, the data suggest that, as previously observed for β-catenin (Hsu et al. 1998; Kolligs et al. 1999), the abilities of γ-catenin to bind Tcf/Lef factors via sequences in armadillo repeats 3–8 and to activate transcription through carboxy-terminal and perhaps amino-terminal sequences are all crticial for γ-catenin's function in stimulating Tcf/Lef transcription.

γ-Catenin promotes neoplastictransformation

We next assessed the oncogenic activity of wild-type and mutated forms of γ-catenin in focus formation assays with the RK3E cell line. Consistent with prior results (Kolligs et al. 1999), wild-type β-catenin failed to induce transformation, whereas the S33Y mutated form of β-catenin produced dense foci of morphologically transformed cells in three weeks (Fig. ​(Fig.3A).3A). In contrast to results with β-catenin, wild-type γ-catenin had essentially the same transforming activity as two amino-terminal mutated forms of γ-catenin (S28Lγ and ΔN38γ). The ability of γ-catenin proteins to induce RK3E transformation was correlated with their ability to activate Tcf/Lef transcription, as γ-catenin mutants defective in Tcf/Lef transcriptional activation, such as ΔN122γ, ΔArmγ, and ΔC687γ, failed to induce foci (Figs. ​(Figs.2C2C and ​and3A).3A). γ-Catenin was about one-third to one-half as potent as the S33Y mutated form of β-catenin in the focus formation assay, though other transforming variants of β-catenin (e.g., in-frame amino-terminal deletions) have been previously found to possess similarly reduced focus forming activity compared with the S33Y mutant (Kolligs et al. 1999). The γ-catenin foci were generally larger in size than the S33Yβ-induced foci at equivalent time points (Fig. ​(Fig.3A).3A). Cell lines established from γ-catenin foci had a distinct morphological appearance when compared with parental RK3E cells or β-catenin-transformed RK3E lines (Fig. ​(Fig.3B;3B; data not shown), with the γ-catenin-transformed cells generally smaller, more spindle-shaped, and growing in tight clusters.

Open in a separate window

Open in a separate window

Figure 3

Neoplastic transformation of RK3E cells by γ-catenin. (A) Focus formation assay in RK3E. RK3E cells were infected with retroviruses encoding wild-type or mutated forms of γ- and β-catenin. Indicated in each panel is the protein encoded by the retrovirus used for infection of the cells. Four weeks after infection, the plates were stained and foci photographed. Representative plates from four or more independent experiments performed on each γ- and β-catenin construct are shown. (B) Morphology of parental RK3E cells, one β-catenin, and two γ-catenin transformed cell lines. Magnification for all panels, 200×.

A total of 18 clonal cell lines were established from independent γ-catenin-induced foci, with 6 lines each from RK3E cells transformed by the wild-type, S28L, or ΔN38 γ-catenin proteins. All γ-catenin-transformed lines displayed neoplastic growth properties, including the ability to grow in reduced serum media (Fig. ​(Fig.4A;4A; data not shown) and to form colonies in soft agar (Fig. ​(Fig.4B;4B; data not shown). Four γ-catenin transformed lines along with a β-catenin-transformed RK3E line were tested for tumorigenicity in nude mice. Each line formed >1.5 cm tumors at 10 of 10 injected sites within 3 weeks (data not shown). All γ-catenin-transformed lines expressed the transduced wild-type or mutant γ-catenin protein (Fig. ​(Fig.4C).4C). No consistent increases in the cytosolic or nuclear levels of β-catenin were seen in the γ-catenin-transformed RK3E lines (Fig. ​(Fig.4C;4C; data not shown). Hence, γ-catenin does not appear to transform RK3E merely by increasing β-catenin levels, a point also well supported by our data showing that γ-catenin mutants that fail to bind to Tcf and activate transcription do not promote neoplastic transformation (e.g., ΔN122, ΔArm, and ΔC687; Figs. ​Figs.2C2C and ​and3A).3A).

Open in a separate window

Open in a separate window

Open in a separate window

Open in a separate window

Figure 4

Analysis of γ-catenin transformed RK3E lines. (A) Growth of γ-catenin-transformed RK3E lines in medium with 0.5% FBS. Cells (2×10⁴) were initially seeded in 35-mm dishes in the presence of medium containing 10% FBS, then changed to medium with 0.5% FBS 24 hr later. Cells were trypsinized and counted at the indicated time points after the switch to 0.5% FBS. One set of assays representative of three independent experiments is shown. (B) γ-Catenin-transformed RK3E lines exhibit anchorage independent growth. Colony formation in soft agar was assessed for all transformed lines by plating 1×10⁴ cells of each line in 0.3% agar medium over 0.6% agar medium underlayers. After 3 weeks, the dishes were stained with methylene blue and photographed. Shown are representative assays for parental RK3E, one β-catenin transformed line (RK3E/S33Yβ-B), and four independent γ-catenin lines (RK3E/WTγ-A, RK3E/WTγ-B, RK3E/S28Lγ-B, and RK3E/ΔN38γ-B). (C) Expression of exogenous Flag epitope-tagged γ- and β-catenin proteins and endogenous β-catenin in transformed RK3E lines. Cytosolic lysates were separated by SDS-PAGE and analyzed by ECL–Western blotting. Expression of the transduced Flag epitope-tagged γ- and β-catein proteins was assessed with an anti-Flag antibody. β-Catenin levels were determined with an anti-β-catenin antibody. Equal loading of the lanes was confirmed by stripping the blots and reprobing with an anti-actin antibody. (D) Tcf/Lef transcription is markedly elevated in γ-catenin-transformed lines compared with parental RK3E cells. The ratio of luciferase activities from a Tcf-responsive reporter (pTOPFLASH) and a control luciferase reporter construct (pFOPFLASH) was determined 24 hr after transfection of the indicated γ- and β-catenin-transformed RK3E lines. Transfection efficiency was assessed with a β-galactosidase expression vector. Mean values and standard deviations from three independent experiments are shown in logarithmic scale because of the wide range of Tcf activities observed in the lines.

Tcf/Lef deregulation is required for γ-catenin transformation

Tcf/Lef transcription activity was measured in all γ-catenin-transformed lines. Though there was variation in the fold activation of Tcf/Lef transcription in the γ-catenin-transformed cell lines, all lines displayed elevated transcriptional activity (Fig. ​(Fig.4D).4D). Our prior studies not only demonstrated that Tcf/Lef transcription was constitutively activated in β-catenin-transformed RK3E lines, but also that transformation by β-catenin was strongly inhibited in an RK3E cell line expressing a dominant negative Tcf-4 mutant protein (RK3E/Tcf-4ΔN31; Kolligs et al. 1999). This inhibition of β-catenin transformation was confirmed, and we also found that γ-catenin transforming activity was essentially completely inhibited in the RK3E/Tcf-4ΔN31 cell line (Fig. ​(Fig.5).5). Therefore, Tcf/Lef factors play a critical role in both β- and γ-catenin transformation.

Open in a separate window

Figure 5

Tcf/Lef factors and c-Myc are required for γ-catenin transformation. Focus formation assays were carried out in parallel following infection of the RK3E/Neo, RK3E/Tcf-4ΔN31, and RK3E/MycΔ106–143 cell lines with retroviruses encoding a control LacZ protein, wild-type γ-catenin, S28L-γ-catenin, or S33Y-β-catenin. Four weeks after infection, the plates were stained and photographed. Representative portions of plates from three independent experiments are shown.

Cell lines and tissueculture

Phoenix, RK3E, and 293 cells were grown in DMEM supplemented with 10% fetal bovine serum (FBS). HT29–APC and HT29–Gal cells (Morin et al. 1995) were provided by K. Kinzler (John Hopkins University) and B. Vogelstein, and the cells were cultured in McCoy's medium supplemented with 10% FBS and 0.6 μg/ml Hygromycin B (Roche Molecular Biochemicals). The polyclonal RK3E-derived cell lines RK3E/Neo, RK3E/Tcf-4ΔN31, RK3E/MycΔ106–143, RK3E/WTβ-1 have been described (Kolligs et al. 1999), and these lines were grown in DMEM/10% FBS with 250 μg/ml G418 (GIBCO BRL/Life Technologies, Inc.). The propagation of recombinant retroviruses and the RK3E focus formation assay have been described in detail (Kolligs et al. 1999). Briefly, RK3E cells and derived lines were infected with the indicated retroviruses at 70%–80% confluence, the medium was changed twice weekly over a period of four weeks, then the plates were stained with Hank's balanced salt solution containing 1.5% glutaraldehyde (Sigma) and 0.06 gram methylene blue (Sigma) per 100 ml, and photographed. The clonal γ-catenin-transformed RK3E lines were isolated from γ-catenin-induced foci. Altogether, 6 independent lines transformed by each of the 3 γ-catenin constructs (wild type, S28L, and ΔN38 γ-catenin cDNAs) were established, for a total of 18 independent γ-catenin-transformed RK3E lines. To assess growth of transformed cells in reduced serum, 2×10⁴ cells were seeded in triplicate in 6-well plates in DMEM with 10% FBS. After 24 hr, the medium was replaced with fresh DMEM containing 0.5% FBS. After 1, 3, 5, and 7 days of culture in 0.5% FBS, the plates were trypsinized and viable cells were counted after trypan blue staining. For colony formation in agar assays, 1×10⁴ cells were plated in 35-mm dishes in Noble agar (Difco), in a final concentration of 0.3% agar, 1× DMEM, and 20% FBS over a 0.6% agar medium layer. The agar surface was kept moist with growth medium during the three-week assay period, until the dishes were stained with methylene blue and photographed.

Immunofluorescence studies of β- and γ-catenin inHT29–APC

HT29–APC and HT29–Gal cells were grown in chamber slides (Nalge Nunc International) with normal growth media for 24 hr; they were then incubated with fresh media supplemented with 150 μm zinc chloride (Sigma). After a 6-hr exposure to zinc, no obvious apoptosis was present and APC expression was strongly induced in Western blot studies (data not shown). Hence, this was the time point used for immunofluorescence studies of APC's effects on β- and γ-catenin. After washing with phosphate buffered saline (PBS), cells were fixed with 3% paraformaldehyde (Sigma) in PBS for 15 min, then permeabilized for 15 min in 0.5% Triton X-100 (Sigma) and 1% goat serum (GIBCO) in PBS. Prior to incubation with the primary antibodies, the slides were blocked by a 30-min treatment with 20% goat serum/0.2% Triton X-100 in PBS. After washing in 2% goat serum/0.2% Triton X-100 in PBS, the slides were incubated for 2 hr with monoclonal antibodies against β- or γ-catenin (Transduction Laboratories) in 2% goat serum/0.2% Triton X-100 in PBS. After further washes, the slides were incubated with a secondary goat anti-mouse IgG antibody conjugated with Texas Red (Jackson Laboratories) for 30 min. The slides were washed with PBS, cover-slipped using VECTASHIELD mounting medium (Vector Laboratories), examined with fluorescence optics on an Olympus AX70 microscope, and photographed at a magnification of 400×.

Immunohistochemistry for β- andγ-catenin

Sections (5 μm) from formalin-fixed and paraffin-embedded carcinoma specimens were dewaxed in xylene, then rehydrated through graded alcohols. Antigen retrieval was performed by boiling in a microwave oven for 10 min in citrate buffer (pH 6.0, Biogenex). Following blocking of endogenous peroxidase activity with 0.5% hydrogen peroxide in methanol, slides were incubated overnight at 4°C with monoclonal antibodies for β- or γ-catenin (Transduction Laboratories) at concentrations of 1.0 μg/ml and 5.0 μg/ml, respectively. Detection was carried out by the avidin–biotin complex method (Vectastain Elite, Vector Laboratories) with diaminobenzidine as the chromogenic substrate. Slides were lightly counterstained with hematoxylin, then evaluated by light microscopy.

Luciferase reporter geneassays

Cells were plated in 35-mm dishes 12 hr prior to transfection. Transfections were performed with FuGENE6 (Boehringer Mannheim), according to the manufacturer's protocol, using 2 μl of FuGENE6 per 1 μg of transfected DNA. To measure Tcf reporter activity in the various RK3E lines, 0.5 μg of pTOPFLASH or pFOPFLASH was cotransfected with 0.5 μg of the control plasmid pCH110. To assess the ability of the various β- and γ-catenin proteins to activate Tcf transcription, 293 cells were transfected with 1 μg of the respective pcDNA3 expression construct, 0.5 μg of pTOPFLASH or pFOPFLASH, and 0.5 μg of pCH110. Analogous studies were carried out with the c-MYC reporter constructs Del-2, Del-3, and Del-4. The total mass of transfected DNA in the assays was kept constant by adding empty pcDNA3 plasmid, if necessary. Two days after transfection, the dishes were washed with PBS, and the cells were scraped from the plates and resuspended in reporter lysis buffer (Promega). Luciferase activities were measured in a luminometer after adding luciferase assay reagent (Promega) to the cell lysates. β-Galactosidase activities were determined according to standard methods and were used to control for transfection efficiency.

Tumorigenicity in nude mice

Groups of 5 mice (nu/nu-nuBR, Charles River Laboratories) were injected subcutaneously into each lower flank with 5×10⁶ cells of each of the following cell lines: RK3E/WTγ-A, RK3E/WTγ-B, RK3E/S28Lγ-B, RK3E/ΔN38γ-B, RK3E/S33Yβ-B, and parental RK3E cells. Prior to injection, cells were trypsinized, washed twice with DMEM, and resuspended in DMEM at a final concentration of 2.5×10⁷ cells/ml. After 3 weeks, all mice injected with γ- or β-catenin-transformed RK3E cells were sacrificed, and tumor diameters and weights were measured. Mice injected with parental RK3E cells were monitored for 8 weeks, without any evidence of tumor formation.

Western blotting

For all protein extracts, confluent cell culture dishes were washed with PBS and then scraped into PBS. After microcentrifugation, the cells were resuspended either in RIPA lysis buffer [Tris-buffered saline (TBS), 0.5% deoxycholate, 0.1% sodium dodecyl sulfate (SDS), 1% Nonidet P-40] for whole-cell extracts or in NP-40 lysis buffer (100 mm NaCl, 50 mm Tris-HCl at pH 7.5, 0.5% Nonidet P-40) for cytosolic extracts. PBS and lysis buffers were supplemented with proteinase inhibitors (complete proteinase inhibitors, Boehringer Mannheim). After a final centrifugation step, the protein concentrations of all lysates were determined by the bicinchoninic acid assay (Pierce). Equal masses of proteins were separated by electrophoresis in SDS–8% polyacrylamide gels. After semidry transfer of the proteins to Immobilon P membranes (Millipore), the blots were incubated in TBS containing 0.1% Tween 20 (Sigma) and 10% nonfat dry milk to block nonspecific antibody binding. The antibody incubation steps were performed in TBS containing 0.05% Tween 20 and 5% nonfat dry milk. The anti-Flag M2 antibody (Sigma), the anti-γ-catenin and anti-β-catenin antibodies (Transduction Laboratories), and the anti-actin antibody (Sigma) were used at 1:5000 dilutions. The secondary horseradish peroxidase-conjugated goat anti-mouse antibody (Pierce) was used at a 1:20,000 dilution. The blots were subjected to enhanced chemiluminescence (ECL; Amersham Life Science) and exposed to X-Omat film (Kodak).

We thank Drs. Kenneth Kinzler and Bert Vogelstein for generously providing the HT29–APC and HT29–Gal cells and the TOPFLASH, FOPFLASH, and c-MYC reporter vectors; Dr. Gary Nolan for the pBMN vector; and Drs. David Ginsburg, Elizabeth Petty, and Steve Weiss for their comments on the manuscript. This work was supported by NIH grants RO1CA70097 and 5 T32 CA09676-08 and Deutsche Forschungsgemeinschaft grant KO1826/1.

The publication costs of this article were defrayed in part by payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 USC section 1734 solely to indicate this fact.