PR55α, a Regulatory Subunit of PP2A, Specifically Regulates PP2A-mediated β-Catenin Dephosphorylation

A central question in Wnt signaling is the regulation of β-catenin phosphorylation and degradation. Multiple kinases, including CKIα and GSK3, are involved in β-catenin phosphorylation. Protein phosphatases such as PP2A and PP1 have been implicated in the regulation of β-catenin. However, which phosphatase dephosphorylates β-catenin in vivo and how the specificity of β-catenin dephosphorylation is regulated are not clear. In this study, we show that PP2A regulates β-catenin phosphorylation and degradation in vivo. We demonstrate that PP2A is required for Wnt/β-catenin signaling in Drosophila. Moreover, we have identified PR55α as the regulatory subunit of PP2A that controls β-catenin phosphorylation and degradation. PR55α, but not the catalytic subunit, PP2Ac, directly interacts with β-catenin. RNA interference knockdown of PR55α elevates β-catenin phosphorylation and decreases Wnt signaling, whereas overexpressing PR55α enhances Wnt signaling. Taken together, our results suggest that PR55α specifically regulates PP2A-mediated β-catenin dephosphorylation and plays an essential role in Wnt signaling.

Cell Lines-HEK293T and SW480 cells were maintained in Dulbecco's modified Eagle's medium (Hyclone) containing 10% fetal bovine serum (Hyclone), 1% penicillin/streptomycin, and 1% glutamine at 37°C in 5% CO 2 . HEK293 cells were cotransfected with shRNA plasmid, psPAX2 packaging plasmid, and pMD2.G envelope plasmid. Virus-containing media was harvested 48 h after transfection, followed by infection of target HEK293T and SW480 cells. Stabled cells were selected with 2 g/ml puromycin 48 h after infection. The stabled cells from a mixed population were used in the experiments.
Reporter Assay-Super8xTOPFlash reporter was kindly provided by Dr. Randall Moon, University of Washington. Super8xTOPFlash reporter and Renilla reporter were cotransfected into HEK293T cells using the calcium transfection method. Luciferase and Renilla activities were analyzed with the Dual-Luciferase Reporter Assay Kit from Promega.
In Vitro Kinase and in Vitro Phosphatase Assays-Purified GST-␤-catenin was phosphorylated by purified GSK3 or CKI␣ in 1ϫ kinase buffer plus 0.2 mM ATP at 30°C for 30 min. Phosphorylated ␤-catenin can also be immunoprecipitated from HCT116 or SW480 cell extracts. Phosphorylated ␤-catenin was incubated with phosphatase buffer only, or purified PP2A, PP1, or PP5 for 30 min at 37°C. ␤-Catenin phosphorylation was analyzed by Western blot with phospho-specific Abs. Purified CKI␣ and GSK3 have been described (4). PP1 and PP2A were from New England Biolabs (P0754S) and Millipore , respectively. Recombinant human GST-PP5 was purified from bacteria. PP2A holoenzyme containing PR55␣ was purified from HEK293T cells as described below.
PP2A holoenzyme containing PR55␣ was purified according to the method described by Adams and Wadzinski (24). FLAG-tagged PR55␣ was transcribed into HEK293T cells. PP2A holoenzyme was immunoprecipitated with FLAG-conjugated beads and eluted by FLAG peptides.

RESULTS
PP2A Regulates ␤-Catenin Dephosphorylation-Our previous study (4) showed that the N terminus of ␤-catenin is phosphorylated by CKI␣ and GSK3. We further showed that Wnt inhibits GSK3 activity, resulting in a dramatic decrease of ␤-catenin phosphorylation and increase of stabilization (4). Because protein kinases and phosphatases often have opposing actions, to investigate whether phosphatases are involved in ␤-catenin dephosphorylation, and if so which phosphatases are responsible, we treated HEK293T cells with 1 nM tautomycin and 10 nM okadaic acid (OA), which preferentially inhibits PP1 and PP2A, respectively. We found that the abundance of phosphorylated ␤-catenin is elevated by OA but not tautomycin, suggesting that PP2A may regulate ␤-catenin phosphorylation (Fig. 1A, top panel).
nin is rapidly degraded in HEK293T cells. Thus, it is difficult to analyze ␤-catenin phosphorylation in these cells. Recently, we noted that ␤-catenin degradation but not phosphorylation is inhibited in the colon cancer cell line SW480 (22). When SW480 cells were treated with LiCl, which inhibits GSK3 activity, the levels of phosphorylated ␤-catenin were significantly decreased (Fig. 1B, top panel). Because ␤-catenin degradation is blocked in SW480 cells, we reasoned that the decrease in ␤-catenin phosphorylation is caused by phosphatase activity but not protein turnover. SW480 cells provide a good model to analyze ␤-catenin phosphorylation without the interference of ␤-catenin degradation.
PP2A Is Essential for Wingless Signaling by Regulating Armadillo Levels in Drosophila-Our results and the results from a phosphatase reconstitution study (18) suggest that PP2A could dephosphorylate and stabilize ␤-catenin, thus playing positive roles in Wnt signaling. However, the in vivo functions of PP2A in Wnt signaling remain elusive likely due to multiple roles of PP2A in the Wnt signaling cascade. To define PP2A functions in vivo, we turned to the Drosophila wing development where Wingless (Wg), a member of the Wnt gene family, exerts a highly conserved biological influence through the canonical Wnt/␤-catenin pathway. We examined the levels of ␤-catenin homolog Armadillo (Arm) and the expression of Wg target genes in Drosophila. We found that expressing dominant-negative PP2Ac, DN-Mts in Drosophila, by the wing-specific MS1096 gal4, caused significant reduction of Arm levels ( Fig.  2B, compared with wild-type Arm staining in Fig. 2A) and severe suppression of Wg target genes such as distalless (dll) and senseless (sen) (compare Fig. 2, BЈ, BЉ with AЈ, AЉ). Consistently, we found that mts RNAi blocked Arm accumulation (Fig. 2C), thus the attenuated Dll expression (Fig. 2DЈ) in Drosophila wing discs. We also performed a gain-of-function study by using the ptc-gal4 that drives gene expression in a stripe anteriorally to the anterior/posterior boundary of the wing disc (26). We found that expressing UAS-mts by ptc-gal4 elevated Arm levels (Fig. 2E).
Identification of PR55␣ as the Regulatory B Subunit That Controls ␤-Catenin Phosphorylation-PP2A holoenzyme comprises a catalytic C subunit, a scaffolding A subunit, and a regulatory B subunit, which controls substrate specificity (8 -10). There are more than 20 regulatory B subunits in the human genome (9). We hypothesize that PP2A plays different roles and regulates multiple steps in the Wnt pathway through distinct regulatory B subunits. Our in vitro and in vivo studies suggest that PP2A regulates ␤-catenin phosphorylation. The regulatory B subunit that controls ␤-catenin dephosphorylation remains to be identified.
PR55␣ comprises a seven-bladed ␤-propeller, or WD40 repeats (27). Each blade comprises four antiparallel ␤-strands. These domains are involved in protein-protein interactions. To analyze the binding between PR55␣ and PP2Ac, Myc-tagged PR55 and FLAG-tagged PP2Ac were co-transfected into HEK293T cells. As expected, PR55␣ forms a complex with PP2Ac (Fig. 3B). To confirm the siRNA results and further analyze the PR55␣ function, we generated stable SW480 cells that express PR55␣ shRNA ( Fig. 3C and supplemental Fig. S3). PR55␣ was efficiently knocked down in these cells compared with the SW480 cells with control shRNA (Fig. 3C, top panel). Consistent with the siRNA result in Fig. 3A, the levels of both CKI␣ and GSK3-phosphorylated ␤-catenin were increased in SW480 cells expressing PR55␣ shRNA (Fig. 3C, second and third panels), with the total amount of ␤-catenin unchanged (Fig. 3C, bottom panel).
Although CKI␣ and GSK3 phosphorylate the N terminus of ␤-catenin and control ␤-catenin degradation, other kinases can also phosphorylate ␤-catenin. It has been shown that AKT and PKA phosphorylated ␤-catenin at Ser-552 and Ser-675, which regulate the activation and nuclear translocation of ␤-catenin (28 -30). When SW480 cells were treated with 10 nM OA, the levels of Ser-552 and Ser-675-phosphorylated ␤-catenin were also increased (Fig. 3D, second and third panels), suggesting that these sites are also regulated by protein phosphatase. Interestingly, we found that depletion of PR55␣ in SW480 cells increased the abundance of ␤-catenin with Ser-552 and Ser-675 phosphorylation (Fig. 3C, fourth and fifth panels), suggesting that PP2A/PR55␣ may affect the phosphorylation of Ser-552 and Ser-675 thus the activation and nuclear translocation of ␤-catenin.
PR55␣ Interacts with ␤-Catenin in Mammalian Cells-We analyzed the localization of PR55␣ in SW480 cells. PR55␣ was localized in both the cytoplasm and nucleus (Fig. 4A). As a control, ␤-catenin localized on the membrane and in the cytoplasm and significantly accumulated in the nucleus (Fig. 4A). To test if PR55␣ interacts with ␤-catenin, endogenous PR55␣ protein was immunoprecipitated from SW480 cells, and the presence of ␤-catenin was analyzed by Western blot. We found that ␤-catenin indeed binds PR55␣ (Fig. 4B). Although several phosphatases and their subunits are involved in Wnt signaling, PR55␣ is the first phosphatase subunit found to bind ␤-catenin.
PR55␣ Directly Binds Both ␤-Catenin and Axin-It has been reported that PP2Ac interacts with Axin (14). To test if PR55␣ binds Axin as well, Myc-tagged PR55␣ and FLAG-tagged Axin or ␤-catenin were cotransfected into HEK293T cells. Using an immunoprecipitation assay, we found that PR55␣ not only interacts with ␤-catenin, but also strongly interacts with Axin (Fig. 5A, lane 2, top panel). Our data indicate that PR55␣ falls into the Axin-␤-catenin protein complex and further suggests that PR55␣ acts as the PP2A regulatory subunit in regulating ␤-catenin phosphorylation.
Because ␤-catenin directly binds Axin, PR55␣ could bind ␤-catenin indirectly through Axin. To test this possibility, we purified histidine-tagged PP2Ac and PR55␣ from Escherichia coli BL21(DE3). These proteins were incubated with purified GST, GST-␤-catenin, GST-Axin(CatϩS45K), and GST-Axin(G3ϩCat) (supplemental Fig. S4). GST-Axin(CatϩS45K) interacted with CKI␣ and GST-Axin(G3ϩCat) interacted with GSK3 (Fig. 5B, bottom two panels), as we reported previously (4). PP2Ac bound GST-Axin(CatϩS45K) (Fig. 5B, top panel), as previously reported (14). In this study, we found that PR55␣ binds not only GST-Axin(CatϩS45K) but also GST-Axin(G3ϩCat) and GST-␤-catenin (Fig. 5B, second panel, lane  4). Neither PP2Ac nor PR55␣ binds GST protein. These results suggest that PP2A directly binds Axin, whereas PR55␣ directly binds both Axin and ␤-catenin. PP2A and PR55␣ may bind distinct domains of Axin. The binding between Axin(Cat) and FIGURE 3. Identification of PR55␣ as the regulatory B subunit that controls ␤-catenin phosphorylation. A, SW480 cells were transfected with siRNAs targeting several different regulatory subunits of PP2A. Negative control siRNA and PP2Ac siRNA were used as controls. Total ␤-catenin and phosphorylated ␤-catenin were analyzed by Western blot with anti-Ser(P)-45 and anti-␤-catenin Abs. B, interaction between PR55␣ and PP2Ac. Myc-tagged PR55␣ was cotransfected with CS2 control plasmid or FLAG-tagged PP2Ac into HEK293T cells. PR55␣ protein was immunoprecipitated from cell lysates with an anti-Myc Ab. The presence of PP2Ac in the immunoprecipitated (IP) samples was analyzed by Western blot with an anti-FLAG Ab. The levels of PP2Ac and PR55␣ in the cell lysates were analyzed as control. C, SW480 cells were infected with lentiviruses that express control shRNA or PR55␣ shRNA. Stable cells were selected with puromycin. PR55␣ protein levels in these cells were analyzed by Western blot with an anti-PR55␣ Ab. Total ␤-catenin and phosphorylated ␤-catenin were analyzed with an anti-␤-catenin Ab and phospho-specific Abs that recognize different phosphorylation sites of ␤-catenin. D, SW480 cells were treated with dimethyl sulfoxide or 10 nM OA. Total ␤-catenin and phosphorylated ␤-catenin were analyzed with an anti-␤catenin Ab and phospho-specific Abs against ␤-catenin.
PR55␣ was tested (supplemental Fig. S5). We found that the Axin(Cat) fragment is not sufficient for PR55␣ binding.
To test which WD40 repeats of PR55␣ bind Axin, we analyzed the binding between Axin and PR55␣ mutants (Fig.  4C). Using the immunoprecipitation assay, we found that unlike ␤-catenin, Axin only binds the PR55␣ fragments containing the last 3-4 WD40 repeats (Fig. 5C). To understand the function of PR55␣ in Wnt signaling, we performed a reporter assay in HEK293T cells using Super8xTOPFlash. PR55␣ alone has no significant effect on the reporter activity (not shown). However, when PR55␣ was cotransfected with Wnt3A, it synergized with Wnt3A and significantly increased the reporter activity (Fig. 5D). PR55␣ mutants PR55␣⌬2 and PR55␣⌬3 have less effect, whereas PR55␣⌬1 has no effect on the reporter activity; this is consistent with their binding affinities to ␤-catenin and Axin. These results further suggest that PR55␣ plays a positive role in Wnt signaling. and Myc-tagged ␤-catenin or its mutants. PR55␣ was immunoprecipitated with an anti-PR55␣ Ab. ␤-Catenin proteins were analyzed by Western blot with an anti-Myc Ab. The expression of these proteins was analyzed by Western blot with an anti-Myc Ab. E, schematic diagram of PR55␣ deletion constructs. The hexagons are WD40 repeats. F, ␤-catenin interacts with multiple domains of PR55␣. HEK293T cells were cotransfected with FLAG-tagged ␤-catenin and Myc-tagged PR55␣ or its mutants. PR55␣ proteins were immunoprecipitated with 9E10-conjugated beads and ␤-catenin was analyzed by Western blot with an anti-FLAG Ab. The levels of PR55␣ and ␤-catenin proteins in the cell lysates were analyzed as control.

PP2A/PR55␣ Regulates ␤-Catenin Degradation by Directly
Dephosphorylating ␤-Catenin-To determine the function of endogenous PR55␣ on ␤-catenin degradation, we generated a stable HEK293T cell line that expresses PR55␣ shRNA. Cytoplasmic fractions were isolated from this cell line (HEK293i) and the control cell line (HEK293c). PR55␣ protein levels were significantly reduced by shRNA (Fig. 5E, left, top panel). PR55␣ depletion resulted in ␤-catenin degradation (Fig. 5E, left, middle panel). As a loading control, ␤-tubulin levels were not affected by PR55␣ shRNA (Fig. 5E, left, bottom panel). It is worth noting that PP2Ac depletion has similar results (14), sug-gesting that PR55␣ and PP2Ac work together in ␤-catenin regulation. These cells were treated with either control medium or Wnt3A-conditioned medium. Besides the above observations, we found that Wnt3A treatment increased the cytoplasmic levels of ␤-catenin more efficiently in HEK293c cells than that of the HEK293i cells (Fig. 5E, right,  middle panel, lane 2, compared with  lane 4), suggesting that PR55␣ enhances Wnt-regulated ␤-catenin stabilization by regulating ␤-catenin dephosphorylation.
As described above (Fig. 5, A and  B), both PR55␣ and PP2Ac bind Axin. Because Axin regulates ␤catenin phosphorylation, it is also possible that PR55␣ indirectly regulates ␤-catenin phosphorylation by regulating Axin. We have shown that ␤-catenin can be dephosphorylated by a commercial PP2A holoenzyme (Fig. 1D), but it is not clear which regulatory B subunit is presented. To test whether PR55␣ directly regulates ␤-catenin dephosphorylation, Myc-tagged PR55␣ and FLAG-tagged PP2Ac were cotransfected into HEK293T cells. The PR55␣-PP2A complex was immunoprecipitated with FLAG-conjugated beads and incubated with phosphorylated ␤-catenin (Fig. 5F, left panel). Both CKI␣ and GSK3 phosphorylations of ␤-catenin were decreased upon incubation with PR55␣/PP2A, whereas total ␤-catenin levels remain unchanged (Fig.  5F, left panel), suggesting that the PR55␣-PP2Ac complex can dephosphorylate ␤-catenin. To test if PR55␣ can recruit endogenous PP2A complex for ␤-catenin dephosphorylation, we purified PP2A holoenzyme containing PR55␣ with FLAG-conjugated beads, and the holoenzyme was eluted from the beads with FLAG peptides. We demonstrated that the purified PP2A holoenzyme contains both PR55␣ and endogenous PP2Ac, and can directly dephosphorylate ␤-catenin in vitro (Fig. 5F, right  panel).

DISCUSSION
␤-Catenin phosphorylation by protein kinases has been well studied (31). How phosphatase regulates ␤-catenin is less understood. PP2A has been shown to dephosphorylate ␤-catenin in vitro (18). However, PP2A has been suggested to play FIGURE 5. PR55␣ directly interacts with ␤-catenin and Axin and regulates PP2A-mediated ␤-catenin dephosphorylation. A, PR55␣ interacts with both Axin and ␤-catenin. HEK293T cells were cotransfected with Myc-tagged PR55␣ and FLAG-tagged Axin, or FLAG-tagged ␤-catenin. Axin and ␤-catenin were immunoprecipitated from the cell lysates using an anti-FLAG Ab. The presence of PR55␣ in the immunoprecipitated samples were analyzed with an anti-Myc Ab. The protein levels of Axin and ␤-catenin were analyzed with an anti-FLAG Ab. B, PR55␣ directly interacts with ␤-catenin and Axin. Purified histidine-tagged PP2Ac and PR55␣ were incubated with purified GST, GST-␤-catenin, GST-Axin(CatϩS45K), and GST-Axin(G3ϩCat), respectively. These GST-Axin fusion proteins have been described previously (4). GST fusion proteins were pulled down with glutathione-agarose beads. PP2Ac and PR55␣ were analyzed with anti-PP2Ac and anti-PR55␣ Abs. The binding between these GST proteins and CKI␣ and GSK3 were analyzed as controls. C, the C-terminal region of PR55␣ interacts with Axin. HEK293T cells were cotransfected with FLAG-tagged Axin and Myc-tagged PR55␣ or its mutants. PR55␣ proteins were immunoprecipitated with 9E10-conjugated beads and Axin was analyzed by Western blot with an anti-FLAG Ab. The levels of PR55␣ and Axin proteins in the cell lysates were analyzed as control. D, PR55␣ enhances Wnt signaling. Super8XTOPFlash was cotransfected with Wnt3A plus CS2 or Wnt3A plus PR55␣ constructs into HEK293T cells. Luciferase activity was analyzed and normalized. E, PR55␣ regulates ␤-catenin stability. Left panel, HEK293T cells were infected with lentiviruses that express control shRNA or PR55␣ shRNA. Stable cells were selected with puromycin. Cytoplasmic PR55␣ and ␤-catenin protein levels were analyzed by Western blot with anti-PR55␣ and anti-␤-catenin Abs. Right panel, the HEK293T cell line contains PR55␣ shRNA (HEK293i) and the control HEK293T cell line were treated with control conditioned medium or Wnt3A-conditioned medium for 6 h. Cytoplasmic fractions were isolated from these cells. ␤-Catenin and PR55␣ levels were analyzed by Western blot using anti-␤-catenin and anti-PR55␣ Ab. ␤-Tubulin was analyzed as a loading control. F, PP2A holoenzyme containing PR55␣ directly dephosphorylates ␤-catenin. GST-␤-catenin was phosphorylated by CKI and GSK3 in vitro. Left panel, Myc-tagged PR55␣ and FLAG-tagged PP2Ac were cotransfected into HEK293T cells. HEK293T cells transfected with empty vector were used as control. The PR55␣-PP2Ac complex was immunoprecipitated with 9E10-conjugated beads. The beads were resuspended in 1ϫ phosphatase buffer and incubated with phosphorylated GST-␤-catenin for 30 min. PR55␣, PP2Ac, total ␤-catenin, and phosphorylated ␤-catenin were analyzed by Western blot. Right panel, PP2A holoenzyme containing FLAG-tagged PR55␣ was purified from HEK293T cells and incubated with phosphorylated ␤-catenin for 0, 10, 20, and 30 min. PR55␣ and the catalytic subunit of PP2A (PP2Ac) were analyzed by Western blot with anti-PR55␣ and anti-PP2Ac Abs. ␤-Catenin was analyzed by Western blot with Abs against phosphorylated ␤-catenin or total ␤-catenin.
both positive and negative roles in Wnt signaling. Our study demonstrated that PP2A regulates ␤-catenin phosphorylation both in vitro and in vivo. The specificity of PP2A in Wnt signaling is regulated by the regulatory B subunit. We demonstrated that PR55␣ directly interacts with ␤-catenin and regulates PP2A-mediated ␤-catenin dephosphorylation.
PP2A Regulates ␤-Catenin Dephosphorylation-The PP2A inhibitor, OA, increased the levels of phosphorylated ␤-catenin in HEK293T cells (Fig. 1A). However, the levels of ␤-catenin are elevated rather than attenuated by OA treatment (Fig. 1A, bottom panel). This is probably because OA regulates multiple phosphatases that play different roles in ␤-catenin phosphorylation and degradation. For example, OA also inhibits PP1 and PP5 when used at higher concentrations. Another possibility is that PP2A may regulate the degradation of phosphorylated ␤-catenin (32).
Because ␤-catenin degradation but not phosphorylation is inhibited in colon cancer cell line SW480, the SW480 cell line provides a better model to study ␤-catenin phosphorylation. The siRNA experiments in SW480 cells suggests that PP2A is essential for ␤-catenin dephosphorylation. The in vitro phosphatase assay further demonstrated that PP2A is both necessary and sufficient for ␤-catenin dephosphorylation. Within 30 min, only PP2A dephosphorylated ␤-catenin (Fig. 1D). However, after 2 h of incubation, PP1 can also dephosphorylate ␤-catenin, and PP5 weakly dephosphorylated ␤-catenin (supplemental Fig. S2), suggesting that PP1 and PP5 are active, but not as specific as PP2A in ␤-catenin dephosphorylation. In addition, siRNA results suggest that PP1 and PP5 are not required for ␤-catenin dephosphorylation.
PP2A Function in Wingless Signaling in Drosophila-To test PP2A function in vivo, we analyzed Arm and its target genes in Drosophila. Loss-of-function of PP2A resulted in decreased levels of Arm (Fig. 2, B and C). Overexpression of PP2A resulted in increased levels of Arm (Fig. 2E). However, under such conditions, we did not observe changes in Wg target genes expression ( Fig. 2EЈ; not shown). It could be possible that Arm dephosphorylation by PP2A was not sufficient to activate Arm, or, PP2A has additional role(s) in regulating Arm signaling activity. Our finding that PP2A plays a positive role in Wnt signaling is consistent with a previous study (11). However, PP2A also plays negative roles in other studies probably by regulating different substrates. It is important to further determine how PP2A specificity is regulated in Wnt signaling.
PP2A Regulatory Subunit, PR55␣, Regulates ␤-Catenin Dephosphorylation-We performed siRNA screening for the regulatory B subunits of PP2A. Our result demonstrates that PR55␣ is required for ␤-catenin dephosphorylation and further suggests that PP2A regulates different steps of Wnt signaling through distinct regulatory subunits (Fig. 3A). We found that PR55␣ also regulates the phosphorylation of ␤-catenin Ser-552 and Ser-675 (Fig. 3, C and D). This may be one of the reasons why the elevation of Arm caused by overexpressing Mts did not ectopically activate Wg target genes (Fig. 2EЈ, data not shown). The Ser-675 site is highly conserved among species although the function of this phosphorylation has not yet been demonstrated in Drosophila or other in vivo models.
Our mechanistic studies of ␤-catenin dephosphorylation by PP2A/PR55␣ are fully consistent with the genetic results of the PR55␣ mutant in Drosophila (11). The Drosophila PR55␣ homolog is encoded by twins (tws). In the tws Ϫ wing disc, both Arm and its target genes were down-regulated, suggesting that Twins is required for Arm stabilization (11). We also found that overexpression of Twins increased Arm in the wing disc. 5 In summary, we provide first and direct evidence that PR55␣ regulates the specificity of PP2A-mediated ␤-catenin dephosphorylation. Our work further suggests that PP2A may perform different roles in Wnt signaling by taking advantage of different regulatory B subunits. These findings are not only important for cell signaling and developmental biology, but also have important implications for cancer biology. It has been reported that PR61␥ (B56␥) but not PR55␣ suppressed human cell transformation (33), suggesting that PP2A regulatory B subunits play distinct roles in tumorigenesis. This may be partially due to their distinct roles in Wnt signaling.