Distinct Activities of the α-Catenin Family, α-Catulin and α-Catenin, on β-Catenin-Mediated Signaling

Abstract

α-Catenin, an integral part of cadherin-catenin adhesion complexes, is a major binding partner of β-catenin, a key component of the Wnt pathway, which activates T-cell factor (TCF)/lymphoid enhancer factor (LEF) transcription and is often upregulated in cancers. Recently, we identified an α-catenin-related protein, α-catulin, whose function is poorly understood, as part of a Rho GTPase signaling complex. Here, based on evidence suggesting that α-catulin may associate with a β-catenin fraction, we investigated the role of α-catenin family members in β-catenin-mediated signals. Expression of the full length or a 103-residue region of α-catenin strongly inhibits the induction of the TCF/LEF-responsive TOPFLASH reporter in HEK293T cells expressing activated β-catenin or in cancer cells with constitutively upregulated Wnt signaling, whereas α-catulin expression had no effect. Interestingly, α-catulin expression attenuates the activation of the cyclin D1 promoter, a target of Wnt pathway signals. α-Catulin appears to inhibit Ras-mediated signals to the cyclin D1 promoter, rather than β-catenin signals, and the synergy between Ras and β-catenin required to fully activate this promoter. Data suggesting the involvement of Rho in this response are presented and discussed. These results suggest a novel function for α-catulin and imply that α-catenin and α-catulin have distinct activities that downregulate, respectively, β-catenin and Ras signals converging on the cyclin D1 promoter.

The Wnt signaling pathway is a key pathway in the mediation of cell growth and development (4). Stimulation of this pathway via Wnt ligand binding to the cell surface seven-transmembrane receptor Frizzled leads to the activation of a series of complex intracellular signals which are still not completely understood (20). This potentially oncogenic pathway is activated by mutations in several of its signaling components, notably the adenomatous polyposis coli (APC) tumor suppressor and β-catenin, which result in the stabilization of cytoplasmic β-catenin (21, 27). The accumulation of cytoplasmic β-catenin allows β-catenin to complex with, and translocate to the nucleus with, members of the T-cell factor (TCF)/lymphoid enhancer factor (LEF) family of transcription factors (26). This process subsequently results in the transcription of multiple oncogenes, including c-myc (12) and cyclin D1 (39, 43). Activating mutations in components of the pathway, such as β-catenin, are thought to lead to constitutive Wnt signaling and subsequent tumorigenesis and are frequently associated with many types of cancer, including colon cancer, desmoid and gastric cancers, hepatocarcinoma, medulloblastoma, melanoma, ovarian cancer, pancreatic cancer, and prostate cancer (21).

Cyclin D1, the regulatory subunit for cdk4 and cdk6, plays a key role in the progression of cells through the cell cycle and past the G₁ checkpoint (14, 38). The importance of cyclin D1 in G₁-phase progression is supported by the finding that the induction of cyclin D1 in rat fibroblasts can shorten G₁ duration (35, 36). Importantly, cyclin D1 is upregulated in many types of human cancer, including B-cell lymphoma and breast, gastric, esophageal, and colon carcinomas (14, 42).

Prior to its discovery as a critical component of the Wnt signaling pathway, β-catenin was originally identified as a key component of the cadherin-catenin complex, which mediates cell-cell adhesion via the cadherin transmembrane receptor. Within this complex, β-catenin links the cytoplasmic portion of cadherin to α-catenin, a distinct protein (6, 18, 19). α-Catenin links cadherin complexes to the actin cytoskeleton and is necessary for functional cell-cell adhesion by these complexes, which requires this connection with the internal actin cytoskeleton (16). The region of α-catenin necessary for binding to β-catenin has been mapped to within the N-terminal 163 amino acids of the protein (3, 13, 30, 31). Loss-of-function mutations in α-catenin can lead to tumorigenesis, and the expression of wild-type protein in cells containing these mutations leads to tumor suppression and a restoration of growth control (3). Tumor progression associated with loss of α-catenin has been reported in human cancers (44). The phenotypic effects in these tumors are generally thought to be due to loss of adhesion, but it is also conceivable that aberrations of intracellular growth control pathways may contribute to the effect. In support of this idea, evidence increasingly indicates that α-catenin plays a role in cell signaling. An early study showed that α-catenin strongly inhibits Wnt pathway-mediated developmental processes in Xenopus embryos (37). Moreover, α-catenin expression inhibits β-catenin-dependent activation of TCF-mediated transcription in various cell types (7, 8, 40). The N-terminal 210 amino acids of the protein, which contain the β-catenin binding site, are required for these functions (37, 40). Recently, it has been proposed that the mechanism by which α-catenin downregulates β-catenin-dependent transcriptional activation is via α-catenin interference with the interaction of the β-catenin-TCF/LEF complex with its target DNA in the nucleus (7).

We recently described an α-catenin-related protein, α-catulin, as part of a Rho GTPase signaling complex. Our findings indicated that α-catulin interacts with the Rho GTP/GDP exchange factor (GEF), Lbc, and may be a putative scaffold for Lbc (32). Little else is known about the function(s) of the recently described α-catulin, which shares sequence similarity with α-catenin and vinculin, members of the vinculin superfamily. α-Catenin is a well-characterized binding partner of β-catenin (6, 18, 19), and vinculin has also been suggested to bind β-catenin (11). On the basis of these findings, here we analyzed the potential role of α-catulin in β-catenin-mediated signaling pathways.

MATERIALS AND METHOD

Plasmids

Flag epitope-tagged pcDNA:β-catenin Δ45 (28) was a gift from H. Clevers (University Medical Center, Utrecht, The Netherlands). Hemagglutinin (HA)-tagged pCGN:α-catenin and mutant constructs (3) were gifts from L. Bullions (Wyeth). The pGFP:vinculin construct was a gift from B. Geiger (Weizmann Institute). The pTOPFLASH reporter construct (28) was a gift from B. Vogelstein (Johns Hopkins School of Medicine). Cyclin D1 reporter constructs were gifts from F. McCormick (University of California, San Francisco). pEXV:RasV12 and pEXV:RhoL63 constructs were gifts from C. Marshall and A. Hall (Institute of Cancer Research, University College London). pcDNA and pEGFPN3 backbone vectors were obtained from Clonetech Laboratories, and pRenilla was obtained from Promega.

The Myc epitope-tagged α-catenin 46-149 expression plasmid (pcDNA:α-catenin 46-149), containing an N-terminal KpnI site and a C-terminal SacII site on the insert, was created as follows. The α-catenin 46-149 insert was generated by PCR with Platinum Pfx DNA polymerase (Invitrogen) and full-length α-catenin in the pCGN vector (3) as a template with the primers 5′-CGGGGTACCATGGGGCCCTCTAATAAGAAG (forward) and 5′-TCCCCGCGGTTTGTAGACATCTGCCAT (reverse). The PCR product was then subcloned into the pcDNA4/Myc-His B vector (Invitrogen) between the KpnI and SacII restriction sites with T4 DNA ligase (Invitrogen) for 16 h at 14°C.

Cell culture

HEK293T cells (a human kidney epithelial cell line), Caco2A colon carcinoma cells, and A427 lung carcinoma cells were obtained from the American Type Culture Collection. HEK293T and Caco2A cells were maintained in Dulbecco's modified Eagle medium supplemented with 10% heat-inactivated fetal bovine serum. A427 cells were maintained in minimal essential medium (Life Technologies) supplemented with nonessential amino acids and 10% heat-inactivated fetal bovine serum. Cells were maintained in a humidified 5% CO₂ incubator at 37°C. Where indicated, cells were treated with 15 mM lithium chloride (Sigma) for the indicated times immediately preceding lysis.

Cell transfection

Cells at 80 to 90% confluence were transfected by using the Lipofectamine reagent (Invitrogen) as recommended by the manufacturer. Six hours posttransfection, medium was replaced with fresh medium containing serum, and cells were harvested 24 to 36 h posttransfection. In all cases, backbone vector was added to the transfection mixture to ensure that all groups were transfected with equal amounts of total plasmids, and experiments were performed in triplicate.

Cell selection

Selection of transfected cells was performed by using the MACS K^k.II system (Miltenyi Biotec) according to the manufacturer's instructions 24 h posttransfection. Cells were cotransfected with 2 μg of the pMACSK^k.II vector to select for transfected cells. Where indicated, cells were treated for 3 h with 15 mM LiCl (Sigma) immediately prior to the cell selection process. Selected cells were then immediately lysed in cell nucleus lysis buffer (50 mM Tris-Cl [pH 8.0], 250 mM NaCl, 2 mM EDTA, 50 mM NaF, 10 mM Na₃VO₄, 1% NP-40, 0.1% sodium dodecyl sulfate [SDS], 1× Protease Arrest [Calbiochem]).

Transcriptional reporter assays

Cells were transfected in 6-well dishes at 90% confluence with either the TOPFLASH, the FOPFLASH, or the cyclin D1 luciferase reporter constructs and the indicated expression plasmids. Each well was also cotransfected with the pRenilla luciferase expression plasmid to provide an internal control for the normalization of all data for transfection efficiency, cell viability, and cell number. Backbone vector was used to ensure the addition of equal amounts of total DNA to all groups during transfections. Thirty-six hours posttransfection, cells were washed with phosphate-buffered saline and 500 μl of Promega reporter lysis buffer was added to each well. Lysates were then collected and assayed for luciferase activity by using the dual luciferase reporter assay system (Promega) according to the manufacturer's instructions. Each experiment was performed in triplicate and repeated at least two times, for a total of six or more experiments.

Immunoprecipitation

Cells were lysed in a solution of 1% Triton X-100, 1% Nonidet P-40, 50 mM Tris-Cl [pH 7.5], and 150 mM NaCl containing 1 mM phenylmethylsulfonyl fluoride and 10 μg of aprotinin/ml. Lysates were then cleared of insoluble material by centrifugation at 10,000 × g for 10 min and precleared with 1 μg of mouse immunoglobulin G (IgG) (Santa Cruz Biotechnology) and 20 μl of protein A-agarose conjugate (Santa Cruz Biotechnology) for 1 h at 4°C. For immunoprecipitations, 2 μg of either mouse anti-c-Myc (9E10; Santa Cruz Biotechnology), goat anti-p120 (M-19; Santa Cruz Biotechnology), mouse anti-β-catenin (E-5; Santa Cruz Biotechnology), rabbit anti-α-catenin (H-297; Santa Cruz Biotechnology), or normal mouse IgG (Santa Cruz Biotechnology) antibody was used as indicated. Immunoprecipitation was carried out as described previously (32).

Nucleus isolation

Nuclei were isolated essentially as described previously (10). Briefly, about 10⁷ to 10⁸ cells were resuspended in 5 ml of buffer A (60 mM KCl, 15 mM NaCl, 15 mM HEPES [pH 7.8], 1 mM phenylmethylsulfonyl fluoride, 10 μg of aprotinin/ml)-0.3 M sucrose-0.5% NP-40 (Calbiochem). The cell suspension was then disrupted with 50 strokes in a Dounce homogenizer and layered onto 2.5 ml of buffer A plus 0.9 M sucrose. Samples were centrifuged at 3,100 rpm (2,000 × g) in a Beckman J-6B centrifuge for 10 min, and the supernatant was removed. The pellet was resuspended in 5 ml of buffer A plus 0.3 M sucrose plus 0.2% NP-40 and centrifuged over 2.5 ml of buffer A plus 0.9 M sucrose. The supernatant was then removed, and the pellet was resuspended in the cell nucleus lysis buffer described above.

Immunoblotting

Immunoblotting was carried out as described previously (32) and was visualized with the Perkin-Elmer enhanced chemiluminescence detection system.

Antibodies

The following primary antibodies were used: anti-c-Myc (9E10; Santa Cruz Biotechnology), anti-Flag (M2; Sigma), anti-HA tag (HA-7; Sigma), anti-β-catenin (E-5; Santa Cruz Biotechnology), anti-α-catenin (H-297; Santa Cruz Biotechnology), anti-cyclin D1 (HD11; Santa Cruz Biotechnology), anti-α-tubulin (B-7; Santa Cruz Biotechnology), and rabbit polyclonal anti-α-catulin (32). Secondary antibodies were horseradish peroxidase-conjugated goat anti-mouse IgG (goat; NEN) and horseradish peroxidase-conjugated goat anti-rabbit IgG (Promega).

Statistical analysis

Data were analyzed with Student's t test; a P value of <0.05 was considered to indicate significance.

RESULT

α-Catulin, a member of the vinculin superfamily, associates with a β-catenin fraction in vivo

Previously (32), the α-catulin protein, which has 50% overall similarity to, and 29% identity to, α-catenin, was identified. α-Catulin is a member of the vinculin superfamily of proteins, which includes both α-catenin and vinculin, and previous analysis of the overall amino acid sequences and genomic structures of these proteins indicated that α-catulin is more closely related to α-catenin than to vinculin (17). Previously, multiple studies (3, 13, 30, 31) mapped the β-catenin binding site on α-catenin to an N-terminal region of the protein. Based on these reports, the smallest consensus region of α-catenin suggested to bind β-catenin comprises residues 46 to 149, shown in Fig. 1A, and BLAST analysis (National Center for Biotechnology Information) indicates that this region is more similar to the corresponding region of α-catulin (31% identity, 52% similarity) than to that of vinculin (28% identity, 45% similarity). On this basis, we tested whether α-catulin, like α-catenin, associates with β-catenin in vivo by using coimmunoprecipitation. As a control, we initially determined complex formation between endogenous α-catenin and β-catenin by immunoprecipitating β-catenin from HEK293T cells with an anti-β-catenin antibody. Following separation by SDS-polyacrylamide gel electrophoresis, immunoblotting of the immunoprecipitated material with anti-α-catenin antibody showed coimmunoprecipitation of substantial amounts of endogenous α-catenin, as represented by the ∼95-kDa band of the expected size in Fig. 1B (top panel, third lane), in keeping with previous reports (1, 18). The reciprocal experiment was also done, and following immunoprecipitation of endogenous α-catenin and immunoblotting for β-catenin, substantial amounts of β-catenin were observed to coimmunoprecipitate with α-catenin (middle panel, second lane). Next, complex formation between endogenous α-catulin and β-catenin was evaluated. Upon immunoprecipitation of endogenous β-catenin with anti-β-catenin antibody, immunoblotting of immunoprecipitated material with anti-α-catulin antiserum (32) showed coimmunoprecipitation of endogenous α-catulin (bottom panel, third lane), albeit at a level lower than that of α-catenin coimmunoprecipitation (top panel, third lane). As a control for specificity, no α-catulin coimmunoprecipitated with p120 catenin (bottom panel, fourth lane) or with mouse IgG (bottom panel, second lane). Since the previously described anti-α-catulin antiserum (32) may not efficiently immunoprecipitate endogenous α-catulin, to test the reciprocal coimmunoprecipitation of β-catenin with α-catulin, Myc-tagged α-catulin expression plasmid (pcDNA:Myc-α-catulin) was transiently transfected into HEK293T cells and Myc-α-catulin was immunoprecipitated with anti-Myc antibody. Following immunoblotting with anti-β-catenin antibody, we observed coimmunoprecipitation of the ∼90-kDa β-catenin with Myc-α-catulin precipitate (middle panel, fourth lane). In contrast, this band was not observed upon immunoprecipitation with the mouse IgG control (middle panel, third lane), and α-catenin did not coimmunoprecipitate with Myc-α-catulin (top panel, fourth lane), indicating specificity of the coimmunoprecipitate. In keeping with this finding, a larger amount of β-catenin coimmunoprecipitated with endogenous α-catenin (middle panel, second lane) than with transfected Myc-α-catulin (middle panel, fourth lane), although it should be noted that these two coimmunoprecipitates cannot be directly compared, since exogenously expressed α-catulin is presumably expressed in only a fraction of the cells. Together, these results suggest that endogenous β-catenin associates with α-catulin, although apparently not as extensively as it associates with α-catenin.

FIG. 1.

α-Catulin, a vinculin superfamily member, associates with a β-catenin fraction. (A) Alignment of the conserved β-catenin binding region of α-catenin (residues 46 to 149) to corresponding α-catulin and vinculin sequences by BLASTp (National Center for Biotechnology Information) analysis, showing 31% identity and 52% similarity between α-catenin and α-catulin and 28% identity and 45% similarity between α-catenin and vinculin. (B) In the top panel, HEK293T whole-cell lysate from untransfected cells was immunoprecipitated with mouse IgG control antibody (second lane) or anti-β-catenin antibody (third lane), and whole-cell lysate from HEK293T cells transfected with 1 μg of pcDNA:Myc-α-catulin was immunoprecipitated with anti-Myc antibody (fourth lane). Whole-cell lysate from transfected cells (first lane) and immunoprecipitates (IP) (second, third, and fourth lanes) were then immunoblotted (IB) with an anti-α-catenin antibody to detect coimmunoprecipitation of α-catenin. In the middle panel, whole-cell lysate from HEK293T cells was immunoprecipitated with anti-α-catenin antibody (second lane), and lysate from HEK293T cells transfected with 1 μg of pcDNA:Myc-α-catulin was immunoprecipitated with mouse IgG control antibody (third lane) or anti-Myc antibody (fourth lane), and whole-cell lysate (first lane) and immunoprecipitates (second, third, and fourth lanes) were immunoblotted with an anti-β-catenin antibody to detect coimmunoprecipitation of β-catenin. In the bottom panel, HEK293T whole-cell lysate from untransfected cells was immunoprecipitated with mouse IgG control antibody (second lane), anti-β-catenin antibody (third lane), or anti-p120 catenin antibody (fourth lane), and whole-cell lysate (first lane) and immunoprecipitates (second, third, and fourth lanes) were immunoblotted with an anti-α-catulin antibody to detect coimmunoprecipitation of α-catulin.

α-Catenin, but not α-catulin or vinculin, inhibits TCF/LEF-mediated transcription

Next, we investigated the potential activities of α-catenin, α-catulin, and vinculin in β-catenin-dependent signaling by using a transcriptional reporter assay with the synthetic luciferase reporter construct pTOPFLASH, which, as shown in Fig. 2A, contains three copies of the TCF/LEF-binding site (28). A mutant reporter, pFOPFLASH, which is unable to bind the TCF/LEF transcription factor (28), was also used as a negative control to measure background transcription, and the levels of TCF/LEF-mediated transcription were determined by calculating the ratios of pTOPFLASH activity to pFOPFLASH activity for each group. To monitor reporter activation, luciferase levels were assayed by using the dual luciferase reporter system, which provides an internal firefly luciferase control (Renilla) for normalization. In the absence of any specific stimuli, the TOPFLASH reporter had negligible activity when cotransfected into HEK293T cells with α-catenin plasmid (Fig. 2B, lane 1), thus providing an excellent background-to-signal ratio. In this system, pTOPFLASH activity was readily induced by various stimuli, including expression of the activated mutant β-catenin Δ45 (28), which, at the dose used in the following experiments, induced approximately 60-fold higher activation of the TOPFLASH reporter than the background (FOPFLASH) reporter, and LiCl treatment (data not shown), which leads to endogenous β-catenin accumulation (23). As shown in Fig. 2B, the expression of β-catenin Δ45 (lane 2) led to robust activation of the TOPFLASH reporter, and coexpression of increasing amounts of α-catenin potently inhibited this β-catenin-mediated pTOPFLASH activation (lanes 3 to 6). In contrast, coexpression of increasing amounts of α-catulin (lanes 7 to 10) did not inhibit β-catenin-mediated pTOPFLASH activation. α-Catulin and α-catenin had effects similar to those described above on the activation of pTOPFLASH induced by other stimuli, such as 15 mM LiCl treatment (data not shown). Finally, based on the similarity of vinculin to α-catenin (shown in Fig. 1A), we also tested whether vinculin expression from pGFP:vinculin could inhibit β-catenin-mediated activation of pTOPFLASH. Like α-catulin, vinculin did not inhibit pTOPFLASH activation induced by β-catenin Δ45 (Fig. 2B, lanes 11 to 13), and it was not included in subsequent studies.

FIG. 2.

α-Catenin, but not α-catulin or vinculin, inhibits β-catenin-induced pTOPFLASH reporter activity. (A) Diagram of the composition of the TOPFLASH reporter used to measure β-catenin-induced transcriptional activation showing the encoded TCF/LEF binding sites. (B) HEK293T cells were cotransfected with increasing amounts of either pCGN:HA-α-catenin (α-catenin) (100 to 800 ng), pcDNA:Myc-α-catulin (α-catulin) (0.5 to 4 μg), or pGFP:vinculin (vinculin) (0.5 to 2 μg), as indicated, and 500 ng of pcDNA:Flag β-catenin Δ45 (β-catenin Δ45) (except in lane 1) to induce TCF-mediated transcription. (C) Caco2A and A427 cells were cotransfected with 1 to 4 μg of pcDNA:Myc-α-catulin (α-catulin), pCGN:HA-α-catenin (α-catenin), pcDNA:Myc-α-catenin 46-149 (α-catenin 46-149), or pCGN:HA-α-catenin 129-907 (α-catenin 129-907), as indicated. (D) HEK293T cells were cotransfected with a constant amount of pCGN:HA-α-catenin (α-catenin) (100 ng) and increasing amounts of pcDNA:Myc-α-catulin plasmid (α-catulin) (0.5 to 4 μg). For all panels, backbone vector was used to ensure that cells were transfected with equal amounts of plasmid, all groups contained 20 ng of pRenilla, and each group was tested again with either the TOPFLASH or the FOPFLASH reporter. Experiments were performed in triplicate and repeated at least twice. Immunoblots show expression levels of α-catenin, α-catulin, and vinculin proteins. In panels B to D, the average pTOPFLASH and pFOPFLASH luciferase activities were normalized against Renilla luciferase activities, and the ratios of the normalized pTOPFLASH:pFOPFLASH luciferase activities were calculated for each group. In panels B and D, the pTOPFLASH:pFOPFLASH ratios are plotted as percentages of the ratio obtained from cells transfected with pcDNA:Flag β-catenin Δ45 plus pcDNA (lane 2), which was set to 100%, and in panel C, these ratios are plotted as percentages of the ratio obtained from cells transfected with pcDNA alone (lane 1), which was set to 100%. Error bars represent standard deviations.

To validate these results in a more physiologically relevant system, we next used two human cancer cell lines which exhibit constitutively upregulated TCF/LEF-mediated signaling, as shown by the activation of the TOPFLASH reporter in the absence of exogenous β-catenin Δ45 expression in these cells (Fig. 2C, lane 1). These cell lines are the Caco2A colon carcinoma cell line, which contains mutations in APC and β-catenin, resulting in the accumulation of cytoplasmic β-catenin (28), and the A427 lung adenocarcinoma line, which also exhibits upregulated TCF/LEF-mediated signaling, although the exact mutation(s) in this line has not been conclusively identified (41). In Fig. 2C, the 100% activation in cells transfected with reporter by itself (lane 1) represents the activation of pTOPFLASH at a level approximately 150-fold higher than the background (pFOPFLASH) level in Caco2A cells and at a level approximately 12-fold higher than the background (pFOPFLASH) level in A427 cells. The quantitative difference between the two cell lines in the degree of pTOPFLASH activation may be due to mutations in different Wnt pathway genes, leading to different levels of cytosolic β-catenin and TCF/LEF transcriptional activation in the two cell types. In both of these cell lines, α-catenin expression led to the inhibition of pTOPFLASH activation (lanes 4 to 5), whereas α-catulin expression did not (lanes 2 to 3), similar to their effects in HEK293T cells. Based on data from previous studies (3, 13, 30, 31) which mapped the region of α-catenin necessary for β-catenin binding to a region in the N terminus as mentioned above, we attempted to further define the necessary and sufficient region for the inhibitory effect on β-catenin-induced pTOPFLASH activation. For this purpose, an α-catenin 46-149 mutant (pcDNA:Myc-α-catenin 46-149) that contains the consensus N-terminal 103-amino-acid β-catenin binding region determined in those previous studies was constructed. Expression of this truncated α-catenin mutant (lanes 6 to 7) inhibited the activation of the TOPFLASH reporter in these cancer cell lines to an extent comparable to that observed with full-length α-catenin. Similar effects were also observed when this N-terminal α-catenin domain was coexpressed with β-catenin Δ45 in HEK293T cells (data not shown). Serving as a control, expression of an α-catenin mutant with a truncation in the N terminus, α-catenin 129-907, which was previously shown to lack β-catenin binding (3), did not inhibit the activity of pTOPFLASH in either cancer cell line (lanes 8 to 9) or in HEK293T cells (data not shown).

Given the complex formation between β-catenin and α-catulin, we next tested whether α-catulin could compete with α-catenin's activity in this system. As shown in Fig. 2D, coexpression of increasing amounts of α-catulin (lanes 5 to 9) with a constant amount of α-catenin which yields a submaximal level of inhibition (lane 4) did not lessen this inhibition of pTOPFLASH activation mediated by α-catenin, suggesting that α-catulin is unable to compete with α-catenin in this inhibitory effect.

α-Catenin, but not α-catulin, cotranslocates with β-catenin into the nucleus after LiCl treatment

Based on previous reports suggesting that α-catenin can cotranslocate to the nucleus with β-catenin (7, 8), we evaluated the subcellular localization of α-catenin and α-catulin in response to LiCl-induced β-catenin accumulation and translocation to the nucleus by isolating nuclear fractions from LiCl-stimulated and unstimulated cells, and immunoblotting for the endogenous proteins. Figure 3 shows the presence of β-catenin in unstimulated (0 h) HEK293T whole-cell lysate, while no β-catenin was detected in the nuclear fraction of unstimulated cells. Similarly, no α-catenin or α-catulin was detected in the nuclear fraction of unstimulated cells. Following stimulation of the cells with 15 mM LiCl treatment, a time-dependent increase in total cellular levels of β-catenin was observed, while levels of α-catenin and α-catulin did not change. Moreover, β-catenin was detected in the nuclear fraction within 1 h of treatment, and its level was further increased by 3 h. In addition, an accompanying appearance of endogenous α-catenin in the nuclear fraction was observed, in agreement with Giannini et al.'s reports (7, 8). Interestingly, unlike α-catenin, α-catulin showed no nuclear accumulation in response to LiCl treatment. To monitor fraction contents, α-tubulin was used as a cytosolic marker and histone H1 was used as a nuclear marker. As shown by the immunoblots in Fig. 3, α-tubulin was absent from the nuclear fraction, whereas histone H1 was enriched in this fraction, validating the purity of the fractionations and the results obtained.

FIG. 3.

α-Catenin and β-catenin, but not α-catulin, translocate into the nucleus in response to LiCl treatment. HEK293T cells at 75% confluence were treated with 15 mM LiCl for the indicated times. Cells were collected by scraping, and nuclei were isolated as described in Materials and Methods. A small fraction of the collected cells were lysed in cell-nucleus lysis buffer prior to nucleus isolation to yield the whole-cell lysate fractions. Cell fractions were resolved by SDS-polyacrylamide gel electrophoresis and separately immunoblotted with antibodies against β-catenin, α-catenin, α-catulin, α-tubulin, and histone H1.

α-Catulin partially inhibits β-catenin-induced activation of the cyclin D1 promoter

Next, we extended our analysis beyond the use of the synthetic TOPFLASH reporter to include reporters containing regions of the native promoter of cyclin D1, a gene known to be responsive to β-catenin/TCF-mediated signaling (39, 43). Initially, we examined the effect of α-catenin family members on β-catenin-induced cyclin D1 transcription by using a luciferase reporter construct containing 1,748 bp of cyclin D1 5′ sequence, designated −1748 CD1, as shown in Fig. 4A (29). In HEK293T cells, coexpression of the −1748 CD1 reporter with β-catenin Δ45 (Fig. 4B, lane 2) resulted in the activation of this reporter, as indicated by a fivefold increase in luciferase activity over the background stimulation observed in cells transfected with the reporter plus the vector alone (lane 1). Similar to the results we obtained with the TOPFLASH reporter, α-catenin expression strongly inhibited β-catenin Δ45-induced activation of the −1748 CD1 reporter (Fig. 4B, lane 3). Interestingly, in contrast to the results obtained with the TOPFLASH reporter, increased α-catulin expression also inhibited β-catenin-induced activation of the −1748 CD1 reporter by up to ∼50% (Fig. 4B, lanes 4 to 6), although this inhibition by α-catulin was not as strong as that observed with α-catenin (lane 3). The inhibitory effect of α-catulin on cyclin D1 promoter activation was also confirmed in Caco2A colon carcinoma cells, which exhibit constitutive β-catenin-mediated signaling. As expected, the −1748 CD1 reporter was activated in Caco2A cells, as shown by increased luciferase activity in cells transfected with the −1748 CD1 reporter by itself (Fig. 4C, lane 2) compared to that in cells transfected with the backbone luciferase expression construct, pGL3 basic, which lacks the 5′ cyclin D1 promoter sequence (lane 1). Further analysis showed that this activation of the −1748 CD1 reporter was strongly inhibited by α-catenin coexpression (lane 3). Moreover, α-catulin expression led to a slightly weaker inhibition of −1748 CD1 reporter activity (lane 4) (∼50%), confirming the results observed for the HEK293T cells.

FIG. 4.

α-Catulin inhibits β-catenin-induced cyclin D1 reporter activation. (A) In a diagram adapted from Tetsu and McCormick (43), the different transcription factor binding sites (boxes) for the −1748 CD1, the wild-type, and two mutant −962 CD1 luciferase reporter constructs are shown. (B) HEK293T cells were cotransfected with the −1748 CD1 reporter (500 ng), pRenilla (20 ng) as an internal control, and pcDNA:Flag-β-catenin Δ45 expression plasmid (500 ng) to induce TCF-mediated cyclin D1 promoter transcription with either 500 ng of pCGN:HA-α-catenin (α-catenin) or 1 to 4 μg of pcDNA:Myc-α-catulin (α-catulin), as indicated. (C) Caco2A cells were cotransfected with 1 μg of the −1748 CD1 reporter (lanes 2 to 4) or pGL3 Basic (lane 1), pRenilla (20 ng), and either pCGN:HA-α-catenin (α-catenin) (4 μg) or pcDNA:Myc-α-catulin (α-catulin) (4 μg), as indicated. (D) HEK293T cells were cotransfected with one of the cyclin D1 reporters (500 ng) described in panel A, pRenilla (20 ng), and either pcDNA vector alone (4.5 μg) or pcDNA:Flag-β-catenin Δ45 (beta-catenin) (500 ng) along with either pcDNA vector (4 μg), pCGN:HA-α-catenin (alpha-catenin) (2 μg), or pcDNA:Myc-α-catulin (alpha-catulin) (4 μg), as indicated. Backbone vector was used to equalize the total amounts of transfected plasmid. For panels B to D, luciferase activity was measured as an indication of promoter activation. The means of the ratios of the reporter/Renilla luciferase readings from at least two independent experiments performed in triplicate were calculated and are presented as percentages of the value for cells transfected with reporter plus pcDNA:Flag-β-catenin Δ45 in panels B and D and as percentages of the value for cells transfected with the −1748 CD1 reporter alone in panel C. For panel D, background levels from cells transfected with reporter plus pcDNA alone were subtracted from all values prior to calculating percentages and plotting. Error bars show standard deviations. In panels B and D, asterisks indicate statistically significant (P < 0.005) decreases in reporter activation compared to that in cells transfected with β-catenin Δ45 plus pcDNA vector (panel B, lane 2; panel D, dark shaded bars).

We next investigated why α-catulin can inhibit β-catenin-induced activation of the −1748 CD1 reporter but not that of the TOPFLASH reporter. In contrast to the synthetic TOPFLASH reporter construct, which exclusively contains three copies of the TCF/LEF transcription factor binding site (28) (Fig. 2A), as shown in Fig. 4A, the −1748 CD1 reporter contains a large region of the native cyclin D1 promoter which encodes multiple elements that respond not only to β-catenin-mediated signals, but also to signals from other stimuli, such as the Ras oncogene (43). On this basis, Ras and β-catenin cooperate with each other to stimulate transcription of cyclin D1 to a greater extent than that resulting from either one of these oncogenes by itself (43). Hence, we tested whether the inhibitory effect of α-catulin on cyclin D1 promoter activation may be via the Ras pathway, which does not induce pTOPFLASH activation. For this purpose, we used the −1748 CD1 reporter and three previously described (43) derivatives shown in Fig. 4A: two shorter forms, −962 CD1 and −962 EtsA mt, both previously demonstrated (43) to respond to both β-catenin and Ras signals, and one shorter form, −962EtsB mt, previously shown (43) to be activated by β-catenin but not by Ras due to a deletion of the EtsB binding site, which is critical for cyclin D1 responsiveness to Ras signals. As shown in Fig. 4D, α-catulin expression partially inhibited β-catenin Δ45-induced activation of the −1748 CD1, −962 CD1, and −962EtsA mt CD1 reporters to similar magnitudes (∼50%) but had no effect on the activation of −962EtsB mt CD1. In contrast, α-catenin expression significantly inhibited β-catenin-induced activation of all four reporters. In summary of our data thus far, α-catulin partially inhibited β-catenin-induced activation of three transcriptional reporters responsive to both β-catenin and Ras (−1748 CD1, −962 CD1, and −962EtsA mt CD1) but not that of two reporters sensitive to only β-catenin (pTOPFLASH and −962EtsB mt CD1).

α-Catulin inhibits Ras-induced activation of the cyclin D1 promoter and synergistic activation of this reporter by Ras and β-catenin

To provide further support for the above findings, the effect of α-catulin on cyclin D1 transcription induced by activated Ras was tested (Fig. 5A). The expression of activated RasV12 in HEK293T cells resulted in activation of the −1748 CD1 reporter (lane 2) by approximately 2.5-fold over the background stimulation observed in cells transfected with the vector alone (lane 1). Coexpression of α-catulin inhibited Ras-induced activation of the −1748 CD1 reporter (lanes 3 to 4), in contrast to α-catenin, which had no effect on the induction of this reporter by RasV12 (lane 5). In fact, α-catulin decreased cyclin D1 reporter activation to below the background activation level (lane 1) of the reporter; a possible explanation for this finding is discussed later.

FIG. 5.

α-Catulin inhibits Ras signaling to the cyclin D1 promoter and its synergy with β-catenin signaling to the same promoter. The effect of α-catulin and α-catenin on −1748 CD1 activation was observed in HEK293T cells transfected with pEXV:RasV12 (1 μg) (RasV12) (A) or with both pEXV:RasV12 (1 μg) and pcDNA:Flag-β-catenin Δ45 (500 ng) (β-catenin) (B). For each panel, cells were also transfected with the −1748 CD1 reporter (500 ng), pRenilla (20 ng) as an internal control, and pCGN:HA-α-catenin (2 μg) (α-catenin) or pcDNA:Myc-α-catulin (1 to 4 μg) (α-catulin), as indicated. Backbone vector was used to equalize total amounts of plasmid transfected. Luciferase activity was measured as an indicator of promoter activation, and the means of the ratios of the −1748 CD1/Renilla luciferase readings were calculated. Values were normalized to the reading for cells transfected with reporter plus vector alone (lane 1), which was set to a value of 1. Experiments were performed in triplicate and repeated at least twice, and data shown are those from a representative experiment. Error bars represent standard deviations.

We next examined the effect of α-catulin on the synergistic activation of −1748 CD1 by Ras and β-catenin by expression of these oncogenes at submaximal levels in HEK293T cells along with increasing amounts of α-catulin. Figure 5B shows that the expression of Ras V12 alone (lane 3) induced a twofold increase in −1748 CD1 reporter activation over the background level (lane 1) but that the expression of β-catenin Δ45 alone (lane 2) led to a fourfold increase in the activation of this reporter. Coexpression of RasV12 and β-catenin Δ45 (lane 4) led to a very robust (∼12-fold) increase in luciferase activity, indicating cooperation between the two proteins in stimulating the cyclin D1 promoter, as reported previously (43). Coexpression of α-catenin with β-catenin Δ45 and RasV12 (lane 8) strongly inhibited this cooperation and reduced reporter activation to a level similar to that observed in cells expressing RasV12 alone (lane 3). Interestingly, α-catulin expression (lanes 5 to 7), although not quite as effective as α-catenin expression, led to a dose-dependent decrease in the synergistic activation of the −1748 CD1 reporter by β-catenin Δ45 and RasV12, and ultimately, with the highest amount of α-catulin transfected (lane 7), the reporter activation was reduced to a level that approached the level of the activation induced by β-catenin Δ45 alone (lane 2). These results support the concept that α-catulin inhibits signaling to cyclin D1 by Ras and, consequently, the synergy between Ras and β-catenin.

α-Catulin inhibits cyclin D1 reporter activation induced by activated RhoA

Recently, Liberto et al. (24) showed that Rho can activate cyclin D1 transcription and that Rho function is required for epidermal growth factor (EGF)-induced activation of the cyclin D1 promoter and for both EGF- and activated Ras-induced S-phase entry. These results suggest that Rho may mediate Ras-induced proliferation and cell cycle progression by regulating cyclin D1 expression. Hence, we evaluated whether α-catulin may modulate cyclin D1 promoter activation induced by activated RhoA. As shown in Fig. 6, coexpression of α-catulin (lane 3) potently inhibited activation of the −1748 CD1 reporter induced by activated RhoL63 (lane 2), whereas the expression of α-catenin had no effect (lane 4), similar to its lack of effect on Ras-induced cyclin D1 transcription.

FIG. 6.

α-Catulin inhibits cyclin D1 reporter activation induced by activated RhoA. HEK293T cells were cotransfected with the −1748 CD1 luciferase reporter construct (500 ng), pRenilla (20 ng), and pEXV:RhoL63 (RhoL63) (1 μg), pcDNA:Myc-α-catulin (α-catulin) (4 μg), and pCGN:HA-α-catenin (α-catenin) (2 μg), as indicated. Backbone vector was used to ensure that all cells were transfected with equal amounts of plasmid. Ratios of −1748 CD1/Renilla luciferase readings were calculated, and the data shown represent the means of the values from three independent experiments performed in triplicate and are presented as percentages of the activity from cells transfected with the reporter and RhoL63 alone (lane 2). Error bars represent standard deviations. An asterisk indicates statistical significance (P < 0.005) for the comparison of group 3 (RhoL63 plus α-catulin) with group 2 (RhoL63 alone).

Cyclin D1 protein expression and Caco2A colony formation are partially inhibited by α-catulin expression

We next examined the effect of α-catulin expression on β-catenin-induced increases in cyclin D1 protein levels stimulated by LiCl treatment. For this purpose, HEK293T cells were transiently transfected with either pcDNA vector, pcDNA:α-catulin, pCGN:α-catenin, or both pcDNA:α-catulin and pCGN:α-catenin in the presence of the pMACSK^k.II expression plasmid. Following LiCl treatment, transfected cells were magnetically selected by using an antibody against the H-2K^k protein expressed by the pMACSK^k.II plasmid coupled to magnetic beads. Cell lysates were then immunoblotted using α-tubulin as a loading control, and relative cyclin D1 protein levels were determined by normalization of cyclin D1 band densities to α-tubulin band densities. LiCl treatment of pcDNA-transfected cells for 3 h induced an approximately twofold increase in cyclin D1 protein expression (Fig. 7A, lane 2) over that in mock treated cells (lane 1). Moreover, the expression of either α-catulin (lane 3) or α-catenin (lane 4) by itself under the same conditions led to significant attenuation of cyclin D1 protein expression, and as indicated by the relative band densities shown in the graph, α-catenin inhibited cyclin D1 expression more strongly than did α-catulin, a result that parallels the data from the transcriptional reporter assays. Furthermore, coexpression of both proteins (lane 5) resulted in an even stronger inhibition of cyclin D1 protein expression than that observed with the expression of either protein by itself. Levels of the tubulin control remained comparable throughout the experiment.

FIG.7.

α-Catulin attenuates LiCl-induced cyclin D1 protein expression and colony formation in Caco2A cancer cells. (A) HEK293T cells were cotransfected with the pMACS K^k.II plasmid (2 μg) and either pcDNA vector (5 μg), pcDNA:Myc-α-catulin (2.5 μg) plus pcDNA (2.5 μg) (pcDNA:α-catulin), pCGN:HA-α-catenin (2.5 μg) plus pcDNA (2.5 μg) (pCGN:α-catenin), or pcDNA:Myc-α-catulin (2.5 μg) plus pCGN:HA-α-catenin (2.5 μg) (pcDNA:α-catulin + pCGN:α-catenin), as indicated. Twenty-four hours posttransfection, transfected cells were treated with 15 mM LiCl for 3 h where indicated and isolated by using the pMACSK^k.II magnetic cell selection system. Isolated cells were lysed in cell nucleus lysis buffer, and 100 μg of lysate from each group was immunoblotted for cyclin D1 and α-tubulin as described in Materials and Methods. Relative band densities were quantified by using an Alpha Innotech IS-2200 Digital Imaging System and software. The ratios of cyclin D1 band densities to α-tubulin band densities were calculated for each group and normalized to the ratio obtained for LiCl-treated cells transfected with pcDNA vector alone (lane 2), which was set to a value of 1. These values are presented in the graph below the immunoblots. (B) Caco2A cells were transfected with equimolar (0.25-pmol) amounts of pcDNA vector or pcDNA:Myc-α-catulin (pcDNA: α-catulin), each of which confers neomycin resistance, and grown in the presence of Geneticin (1.75 μg/ml) for 3 weeks, with the media being changed every third day. Colonies were then stained with crystal violet and counted. The average number of resistant colonies for three experiments relative to the average number for cells transfected with pcDNA vector alone are shown for each group in the graph. Error bars represent standard deviations.

Finally, we determined whether α-catulin could affect cell growth in Caco2A cells by using a colony formation assay. Following transfection of Caco2A cells with pcDNA vector or pcDNA:α-catulin, both of which express the neomycin resistance gene, equal numbers of cells were seeded into dishes with media containing G418. After 3 weeks of growth, colonies were stained with crystal violet and counted. As shown in Fig. 7B, Caco2A cells transfected with pcDNA:α-catulin formed ∼40% fewer colonies than Caco2A cells transfected with pcDNA vector alone.

DISCUSSION

Despite the fact that β-catenin is a key component of the Wnt pathway, much remains to be understood about β-catenin and its binding partners. While we consistently detected an α-catulin:β-catenin complex in vivo upon immunoprecipitation of β-catenin, the amount of α-catulin associated with β-catenin appeared to be smaller than that of α-catenin associated with β-catenin. Given that the majority of β-catenin is normally bound to α-catenin, this result was not altogether unexpected. β-Catenin has several other binding partners, including APC and axin (20), which also show much lower stoichiometric association than α-catenin, yet these partners are biologically relevant inhibitors. Hence, the stoichiometry of a physical association as determined by immunoprecipitation may not necessarily be a true indicator of activity, and this was the rationale for the subsequent experiments.

The dramatic inhibition of pTOPFLASH activation that was observed upon α-catenin expression in HEK293T and cancer cells is consistent with data from other reports (7, 8, 40) and represents an intriguing activity for a protein that is generally viewed as a cytoskeletal linker protein. The effect of full-length α-catenin on pTOPFLASH activation was significantly stronger in HEK293T cells than in the Caco2A and A427 cells, a finding we presently attribute to lower transfection efficiencies in these cancer cell lines. Since activated β-catenin does not need to be exogenously expressed in these two cell types to activate TCF/LEF-mediated transcription due to endogenously upregulated β-catenin levels, lower expression of the transfected α-catenin would result in a weaker inhibition of pTOPFLASH activation than that observed in HEK293T cells.

In addition, we were able to further define an α-catenin core bioactive region to a 103-residue stretch (residues 46 to 149), a finding which may be useful in potential therapeutic approaches to treating cancers with upregulated Wnt signaling. This region represents the smallest consensus region necessary for β-catenin binding, as mapped by four previous studies (3, 13, 30, 31), although the signaling activity of this domain was not determined. While two of those studies suggest that smaller portions of this region (residues 97 to 148 [30] and residues 117 to 143 [13]) are sufficient to bind β-catenin, we observed that residues 84 to 907 did not inhibit β-catenin-mediated activation of pTOPFLASH in HEK293T cells (data not shown), suggesting that residues 46 to 149 span the smallest region presently determined to both bind β-catenin (3) and inhibit its signaling function.

The contrasting finding that α-catulin did not inhibit β-catenin-induced pTOPFLASH activity indicates a different functional role for α-catulin, despite its sequence homology with α-catenin. Similarly, vinculin had no inhibitory effect, although this finding was not as unexpected, as cDNA and genomic analyses indicate that vinculin is not as closely related to α-catenin as α-catulin is (17, 46).

LiCl treatment led to a robust increase in total cellular levels of β-catenin, although levels of the other proteins assayed did not change. In addition, the appearance of nuclear β-catenin was observed, reflecting its required function in TCF/LEF transcription, which is activated by LiCl treatment, as indicated by activation of the TOPFLASH reporter (data not shown). Two mechanisms have been proposed for the inhibitory effect of α-catenin on β-catenin-mediated transcription. One proposed mechanism involves sequestration of β-catenin in the cytosol through its binding to α-catenin (40). In contrast, Giannini et al. (7) proposed that the association of α-catenin with β-catenin in the nucleus interferes with protein-DNA interactions required for TCF-mediated transcription. Our finding that LiCl treatment of cells correlates with nuclear translocation of endogenous α-catenin is in keeping with Giannini et al.'s proposal and their corresponding finding, by use of immunofluorescent staining, of endogenous α-catenin in the nuclei of two colon cancer cell lines with upregulated Wnt signaling (7). Since we did not observe nuclear translocation of α-catulin upon treatment with LiCl, it is unlikely that α-catulin can interfere with these protein-DNA interactions, and this improbability may, at least partially, explain the different effects of α-catenin and α-catulin on β-catenin-mediated transcription that we observed.

The inability of excess α-catulin to compete out, or reduce, α-catenin's inhibition of β-catenin-induced pTOPFLASH activation again suggests that α-catulin may associate with a different β-catenin fraction than α-catenin associates with. Thus, it is possible that α-catulin associates with one of the multiple, distinct β-catenin pools proposed to exist other than the major pool associated with cadherin-catenin adhesion complexes (9) or with distinct forms of modified β-catenin, such as tyrosine phosphorylated β-catenin, whose protein interactions are altered (15, 22). Hence, it is conceivable that the α-catulin-associated β-catenin fraction has a function other than transcriptional activation.

The finding that α-catulin inhibits up to 50% of the β-catenin-induced activation of the −1748 cyclin D1 promoter in both HEK293T and Caco2A cells provided the first clue for an activity for α-catulin in transcriptional responses. One possible explanation for why α-catulin inhibits β-catenin-mediated activation of the cyclin D1 promoter but not that of the TOPFLASH reporter may be the difference in the number of TCF sites on the two promoters. Since pTOPFLASH contains three TCF sites, the inhibition of the β-catenin-induced activation of this reporter by α-catulin may be more difficult to detect. However, since α-catulin does not inhibit β-catenin-induced activation of the −962 EtsB mt cyclin D1 reporter but does inhibit the activation of the wild-type −962 cyclin D1 reporter, both of which have the same number of TCF sites, we think this is an unlikely possibility, although it cannot presently be ruled out.

As depicted in the model in Fig. 8, the cyclin D1 promoter is also activated by Ras, which signals to this promoter through the EtsB binding site (43). The use of cyclin D1 promoter mutants suggests that α-catulin attenuates cyclin D1 activation by endogenous Ras rather than by β-catenin signals. The inhibition of RasV12-induced −1748 CD1 activity by α-catulin in HEK293T cells and of the synergistic activation of this reporter by RasV12 and β-catenin Δ45 further supports these findings. Since Ras and β-catenin synergistically stimulate cyclin D1 transcription to a level higher than that induced by either protein alone (43), the total activation of the cyclin D1 promoter by β-catenin Δ45 likely results from synergy of this activated β-catenin with endogenous Ras signals induced by the presence of serum in the media used in the assays described here. Thus, the inhibition of such endogenous Ras signals to the cyclin D1 promoter by α-catulin would result in the partial inhibition (up to 50%) of β-catenin-induced cyclin D1 promoter activation which we observed. In keeping with this idea, α-catulin not only completely inhibited RasV12-induced −1748 CD1 activation but also further lowered the reporter activity to below background levels when expressed in the highest amounts, also likely due to its effect on serum-induced Ras signals. Whether α-catulin can attenuate Ras signals in contexts other than the cyclin D1 promoter remains to be determined.

FIG. 8.

Proposed model for the observed inhibitory effects of α-catulin and α-catenin on the activation of cyclin D1 transcription. The cyclin D1 promoter is activated by serum via a Ras-mediated pathway which signals to the EtsB binding site and by a β-catenin-mediated signaling pathway (inducible by LiCl) which signals to the TCF-1 binding site. α-Catenin likely inhibits β-catenin-mediated activation through its direct interaction with β-catenin, as proposed by others (7). We propose that α-catulin may inhibit Ras-induced cyclin D1 promoter activation via attenuation of Rho-mediated signaling downstream of Ras.

Activated Rho induces cyclin D1 promoter activation, and Rho function is required for EGF-induced activation of this promoter (24). Rho function is also necessary for Ras-induced transformation and S-phase entry (24, 33, 34). As depicted in Fig. 8, one possible mechanism by which α-catulin may down-modulate Ras may be via Rho, and our results suggest that α-catulin may target some aspect of Rho function in this context. Our previous report of an α-catulin fraction as part of a Rho signaling complex via association with Lbc Rho GEF (32) also implies a functional link between α-catulin and the Rho pathway. Interestingly, the transforming ability of oncogenic forms of Rho GEF family members, such as Lbc, is associated with increased transcription of cyclin D1 (45) rather than with other endpoints such as activation of immediate early gene transcription via the serum response factor (SRF). Moreover, the binding of the N terminus of α-catulin to wild-type Lbc (proto-Lbc) but not to the transforming form (onco-Lbc) (32) is in keeping with the downregulatory effect of α-catulin on Rho signals proposed here.

We propose an antiproliferative role for α-catulin based on its attenuation of cyclin D1 transcription. Previous studies (32) did not detect an effect of α-catulin on a different Rho-dependent transcriptional pathway mediated by the SRF, although α-catulin coexpression enhanced Lbc-induced SRF activity. While the basis for this difference is not precisely understood, a number of possibilities exist. The activation of different Rho-dependent transcriptional pathways is likely associated with distinct Rho complexes containing particular scaffolds and Rho effectors. On this basis, the presence of a putative scaffold, such as α-catulin, in a Rho complex may disrupt a particular downstream Rho effector that is required for cyclin D1 transcription but not for SRF-mediated responses. Alternatively, the presence of α-catulin may limit the availability of common Rho effectors for a Rho complex required for cyclin D1 transcription by sequestering them into a distinct Rho complex associated with SRF activation. While at this point these and other possibilities remain to be explored, the identification of other binding partners of α-catulin may yield further clues concerning its mechanisms of signal modulation.

Our finding that α-catulin expression downregulates cyclin D1 protein levels is in keeping with the results obtained by using cyclin D1 transcriptional reporter assays and, together with our finding that α-catulin expression attenuates colony formation in Caco2A cells, supports the notion that α-catulin can modulate endogenous growth signaling pathways. α-Catenin expression was also previously shown to lead to reduced colony formation in cells containing mutant α-catenin (3), suggesting a possible function for this protein family in cell growth control. While our observations are based on the exogenous expression of α-catenin and α-catulin, it is certainly conceivable that levels of the endogenous proteins are modulated. The vinculin gene contains SRF-responsive elements which respond to serum stimulation (25), and whether the α-catenin and α-catulin genes contain SRF binding elements remains to be determined. Such a possibility would potentially allow for the downregulation of Wnt-associated signaling by increased endogenous levels of these two proteins in response to serum, thus establishing regulatory loops between different growth pathways and providing tight, time-dependent control over the expression of genes involved in specific stages of inductive cell growth.

The Wnt pathway is upregulated in many human cancers, including colon cancer (2), the second most commonly occurring cancer, and activating Ras mutations are present in approximately 20% of all cancers (5). At least in the context described here, our findings indicate that α-catenin and α-catulin may down-modulate the activity of dominantly active oncogenes such as β-catenin Δ45 and RasV12, respectively, along with their endogenously activated signals found in transformed cells, and thus potentially provide a basis for future cancer therapies.