A Positive Role for the PP2A Catalytic Subunit in Wnt Signal Transduction*

Protein phosphatase-2A (PP2A) is a multisubunit serine/threonine phosphatase involved in intracellular signaling, gene regulation, and cell cycle progression. Different subunits of PP2A bind to Axin and Adenomatous Polyposis Coli, components of the Wnt signal transduction pathway. Using early Xenopusembryos, we studied how PP2A functions in Wnt signal transduction. The catalytic subunit of PP2A (PP2A-C) potentiated secondary axis induction and Siamois reporter gene activation by Dishevelled, a component of the Wnt pathway, indicating a positive regulatory role of this enzyme in Wnt signaling. In contrast, small t antigen, an antagonist of PP2A-C, inhibited Dishevelled-mediated signal transduction, as did the regulatory PP2A-B′ε subunit, consistent with the requirement of PP2A function in this pathway. Although Wnt signaling is thought to occur via regulation of β−catenin degradation, PP2A-C did not significantly affect β−catenin stability. Moreover, the pathway activated by a stabilized form of β−catenin was sensitive to PP2A-C and its inhibitors, suggesting that PP2A-C acts downstream of β-catenin. Because previous work has suggested that PP2A can act upstream of β-catenin, we propose that PP2A regulates the Wnt pathway at multiple levels.

The Wnt family of secreted glycoproteins regulates many biological processes including cell growth, cell polarity, and tissue specification (1)(2)(3). In Xenopus embryos, microinjection of RNAs encoding certain Wnts into ventral-vegetal blastomeres leads to generation of a secondary dorsal-ventral axis (4). Furthermore, inhibition of downstream Wnt signaling suppresses the formation of the primary axis (5,6). Thus, activation of the Wnt pathway plays a key role in dorsal axis development in vertebrates.
Genetic and biochemical approaches have identified many components of the Wingless/Wnt signal transduction pathway. Wnts bind to transmembrane Frizzled receptors, which leads to activation of the cytoplasmic Dsh (Dishevelled) protein (7,8).
Dsh forms a complex with proteins of the Axin family (9 -12), which also bind glycogen synthase kinase-3␤ (GSK-3␤), 1 the adenomatous polyposis coli gene product (APC), protein phosphatase-2A (PP2A), and ␤-catenin (13)(14)(15)(16). In the absence of Wnt signaling, ␤-catenin within this complex is phosphorylated by GSK-3, and this leads to its rapid degradation via the ubiquitin pathway (17). In response to Wnt signals, ␤-catenin is no longer targeted for degradation and accumulates to high levels in the cytoplasm (18). This stabilized ␤-catenin enters the nucleus and, complexed with transcription factors of the Lef-1/Tcf family, promotes the transcription of target genes (19,20) such as Siamois (21). The biochemical mechanism by which the Axin⅐GSK-3⅐APC⅐␤-catenin complex is regulated by Wnt signaling is not yet known. Regulation of phosphorylation is likely to play an important role, because Dsh, Axin, APC, GSK-3, and ␤-catenin are all phosphoproteins. The serine/threonine phosphatase PP2A has a wide range of substrates and is important in many cellular processes. It comprises a regulatory A subunit, a catalytic C subunit, and variable regulatory B subunits, which may target the location and/or action of the holoenzyme (22)(23)(24). Mouse embryos lacking the PP2A-C␣ gene die 6.5 days post coitum, indicating that PP2A has a critical function during early development (25). PP2A-C was shown to bind to Axin (15), whereas regulatory subunits of the PP2A-BЈ family can interact with APC (26). Also, PP2A can promote dephosphorylation of both APC and Axin (27,28). Overexpression of the PP2A-BЈ⑀ subunit causes a decrease in levels of ␤-catenin in mammalian tissue culture cells (26). Although these observations indicate that PP2A may regulate ␤-catenin degradation, the specific role of the catalytic subunit and the function of the holoenzyme in the Wnt pathway remain unclear.
We have used microinjection assays in Xenopus laevis to probe the action of PP2A in Wnt signaling. We show that PP2A-C is a necessary positive component of Wnt signaling in Xenopus. Unexpectedly, we find that PP2A can act downstream or parallel to ␤-catenin, and it may serve to activate Tcf transcription factors. Because PP2A has also been shown to act upstream of ␤-catenin, we propose that this enzyme regulates the Wnt pathway at multiple levels.

EXPERIMENTAL PROCEDURES
DNA Constructs-The Siamois-luciferase reporter construct (pSia-Luc) (29), pE1b-luciferase (30), dominant negative GSK-3 (DN-GSK-3) (31), and the hemagglutinin-tagged form of Xenopus Dsh (Xdsh) (12) have been described previously. For synthesis of mRNA for injection experiments, bovine PP2A-C␣ was subcloned into pCS2 (33) using BamHI and EcoRI restriction sites; SV40 small t antigen was subcloned into pXT7 (31) via KpnI and BglII sites. Xenopus PP2A-BЈ⑀ cDNA was isolated in a yeast two hybrid screen using Xdsh as a bait, as described (12). Of the 189 positive clones isolated in this screen 67 were sequenced, and 9 encoded Xenopus PP2A-BЈ⑀. The Xenopus PP2A-BЈ⑀ sequence has been deposited in GenBank™ under accession number AF298157. A cDNA comprising the entire coding sequence of PP2A-BЈ⑀ was subcloned into pXT7HA, 2 using available restriction sites. Detailed * This work was supported by National Institutes of Health grants (to S. Y. S.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM  Xenopus Embryos, Microinjections, and Luciferase Assays-These were performed as described previously (34). Luciferase activity data are presented as the average of triplicate samples, each comprising five embryos, and every experiment was reproduced on at least three separate occasions. The following amounts of RNA were injected: Xdsh, 500 or 125 pg as indicated; PP2A-BЈ⑀, 200 or 50 pg; small t antigen, 50 pg; PP2A-C␣, 1 ng; DN-GSK-3, 200 pg; stabilized ␤-catenin (XBC26), 200 pg; and XTcf-3, 250 pg.
Protein Analysis-For ␤-catenin stability assays, Myc-␤-catenin RNA (XBC40; 50 pg) was injected into the two ventral animal cells at the 4 -8-cell stage, alone or in the presence of PP2A-C␣, small t antigen or DN-GSK-3␤ RNA. Five embryos were harvested at stage 10.5 in 100 l of lysis buffer (50 mM Tris⅐HCl, pH 7.5, 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, 10 mM NaF, 1 mM Na 3 VO 4 ), and cell debris and yolk platelets were removed by centrifugation at 14,000 rpm for 5 min. The supernatant was mixed 1:1 with 2ϫ sample loading buffer (35), and samples were analyzed by standard SDS polyacrylamide gel electrophoresis and Western blotting procedures as described (12). Myc antibodies were from the 9E10 hybridoma (36), and ␤-tubulin monoclonal antibodies were from Sigma.

PP2A Is an Essential Positive Regulator of the Wnt Pathway-
The role of PP2A in Wnt signaling was studied using RNA microinjection assays in early frog embryos. Dsh is an essential component of the Wnt pathway and, when overexpressed, induces an ectopic secondary axis (37,38). mRNA encoding the catalytic subunit of bovine PP2A (PP2A-C␣) was coinjected with mRNA for Xdsh into ventral-vegetal blastomeres of four cell-stage embryos. PP2A-C␣ did not inhibit the ability of Xdsh to induce a secondary axis nor was it able to induce a secondary axis when expressed alone (data not shown). When low doses of Xdsh were used such that secondary axes were often only partial and formed at lower frequencies, PP2A-C␣ was found to increase the incidence of secondary axis formation ( Fig. 1A and B). Furthermore, a higher percentage of these axes were complete, as scored by the presence of eyes and cement gland (see Fig. 1 and Table I). This result shows that PP2A-C cooperates with Xdsh in axis induction.
We also studied transcriptional activation of a reporter gene comprising the promoter for the Wnt target Siamois linked to luciferase (pSia-Luc) (27). This construct was coinjected with Xdsh mRNA into Xenopus ventral animal blastomeres at the 4 -8-cell stage. Embryos were harvested at stage 10.5, and luciferase activity was assayed. This site of injection was chosen, because the pSia-Luc reporter alone is not significantly induced in ventral animal blastomeres (29). The reporter activity was up-regulated by Xdsh, and this activation was further increased by PP2A-C␣ ( Fig. 2A). Together, these data suggest that the catalytic subunit of PP2A plays a positive role in Wnt signal transduction.
SV40 small t antigen (Smt) is a commonly used inhibitor of the catalytic subunit of PP2A (39). Overexpression of Smt completely blocked the ability of Xdsh to activate pSia-Luc (Fig.  2B). An E1b-luciferase promoter was used as a control to confirm that this decrease in signal from pSia-Luc was not because of nonspecific inhibition of basal transcription or translation in response to Smt. In fact, there was a slight increase in signal from the E1b-luciferase reporter (Fig. 2B) and a cytomegalovirus-luciferase reporter (data not shown) in the presence of Smt, suggesting that PP2A-C has a negative influence on transcription and/or translation. These data indicate that PP2A-C function is required for Wnt signal transduction.
The PP2A-BЈ⑀ Subunit Is an Antagonist of Xdsh-In addition to Smt, regulatory B subunits of PP2A can modulate PP2A activity (40,41). They have been proposed to target PP2A-C to specific cellular locations and/or confer substrate specificity (24,42,43). Biochemical studies of purified PP2A have shown that dissociation of the B subunit from the core enzyme (AC) leads to increased phosphatase activity, suggesting that B subunits can act as negative regulators of the catalytic subunit (40,41). A cDNA encoding the PP2A-BЈ⑀ subunit was isolated in a yeast two-hybrid screen of a Xenopus gastrula cDNA library using Xenopus Dishevelled as a bait. This Xenopus PP2A-BЈ⑀ was highly conserved, revealing 96% identity with human PP2A-BЈ⑀ at the amino acid level (data not shown). Coinjection of mRNA encoding the Xenopus PP2A-BЈ⑀ subunit with Xdsh mRNA blocked the ability of Xdsh mRNA to induce a secondary axis (see Fig. 1, C and D and Table I). Furthermore, PP2A-BЈ⑀, like Smt, inhibited activation of the pSia-Luc reporter by Xdsh (Fig. 2B) while slightly increasing activation of a control E1b-Luc reporter. Thus, overexpression of the PP2A-BЈ⑀ subunit has an effect opposite that of the catalytic subunit. These data are consistent with the hypothesis that the PP2A-BЈ⑀ subunit acts as a negative regulator of the catalytic subunit and further support a requirement of PP2A in Wnt signal transduction.
PP2A-C Acts Independently of ␤-Catenin Stabilization-In mammalian cells, overexpression of PP2A-BЈ⑀ leads to ␤-catenin degradation (26), suggesting that PP2A-C may promote ␤Ϫcatenin stabilization. This result could provide an explanation for the positive role of PP2A in Wnt signaling. To determine the effect of PP2A-C on the stability of ␤-catenin, low doses (50 pg) of RNA encoding Myc-␤-catenin were injected  with PP2A-C␣ or Smt RNAs into ventral animal blastomeres of Xenopus embryos at the 4-cell stage. Embryos were harvested for protein analysis at stage 10.5. PP2A-C␣ or Smt had no effect on the levels of Myc-␤-catenin (Fig. 3A). As a positive control, dominant negative GSK-3 RNA caused a clear increase in the levels of ␤-catenin in the same experiment. The inability of PP2A-C or Smt to affect ␤-catenin levels suggested that the potentiation of Wnt signaling by PP2A-C might occur by a mechanism other than increasing ␤-catenin stability. To investigate this possibility, we used a form of ␤-catenin, XBC26, which lacks the N-terminal phosphorylation sites necessary for targeting to the ubiquitin-mediated degradation pathway (18). Injection of PP2A-C␣ RNA increased the induction of pSia-Luc by XBC26 RNA (Fig. 3B). Conversely, Smt inhibited activation of pSia-Luc by XBC26 (Fig. 3C). Activation was also blocked by high doses of PP2A-BЈ⑀ (data not shown). Thus the PP2A catalytic subunit plays a positive role in Wnt signaling that is downstream or parallel to ␤-catenin stabilization.
Recent evidence has shown that Wnt signaling can be regulated at the level of Tcfs (44). Although XTcf-3 is required for transmission of the Wnt signal, when injected alone it is unable to induce transcription of dorsal-specific genes and instead acts as a repressor (19,45). Consistent with these reports, we found that XTcf-3 mRNA was unable to induce pSia-Luc reporter activity when injected into animal-ventral blastomeres (Fig. 4). Furthermore, overexpression of XTcf-3 blocked activation of pSia-Luc by Xdsh. However, coinjection of XTcf-3 mRNA and PP2A-C␣ mRNA caused a significant increase in promoter activity (Fig. 4). This result suggests that PP2A-C may promote the conversion of XTcf-3 from a repressor to an activator. DISCUSSION In this report, we show that PP2A plays a positive role in Wnt signal transduction. Coinjection of the catalytic subunit potentiates signaling by Xdsh, as assayed by the ability of Xdsh to induce a secondary axis or activate a Siamois-luciferase reporter. Inhibition of endogenous PP2A-C by small t antigen blocks the ability of Xdsh to activate Siamois, demonstrating that PP2A-C is required for Wnt signal transduction. These findings are unexpected, because previously published data suggested that PP2A-C might be an inhibitor of the pathway in mammalian cells (26).
The PP2A-BЈ⑀ regulatory subunit of PP2A has been shown to bind to APC and is a negative regulator of Wnt signaling (26). We isolated a highly conserved Xenopus homologue of PP2A-BЈ⑀ using a yeast two-hybrid screen for proteins interacting with Xdsh. However, we were unable to demonstrate binding of PP2A-BЈ⑀ to Xdsh in embryo lysates, suggesting that the binding may be weak or transient. Consistent with the earlier report (26), our data show that PP2A-BЈ⑀ inhibits axis induction by Xdsh. Thus, the PP2A-BЈ⑀ subunit acts in an opposite manner to the catalytic subunit and is likely to function as an antagonist of PP2A-C. Our in vivo experiments support biochemical studies in which the presence of certain regulatory B subunits was correlated with reduced activity of the enzyme (40,41).
In mammalian tissue culture cells, Myc-␤-catenin levels are increased by okadaic acid, an inhibitor of PP2A-C (26). In our experiments we did not observe a significant change in the stability of ␤-catenin in response to PP2A-C or Smt, whereas ␤-catenin was clearly stabilized by dominant negative GSK-3 Luciferase activity was assayed at stage 10.5. Neither XTcf-3 nor PP2A-C␣ alone significantly activate pSia-Luc. XTcf-3 inhibits activation of pSia-Luc by Xdsh. When both XTcf-3 and PP2A-C␣ are injected together the reporter is activated. (Fig. 3A). The assay we used may not be sufficiently sensitive. It is also possible that PP2A may behave differently in mammalian and Xenopus systems. Such differences in the Wnt pathway have been previously documented for APC, a potential PP2A substrate that inhibits Wnt signaling in mouse cells (46) but activates the same pathway in Xenopus embryos (47).
Because PP2A had no discernable effect on ␤-catenin stability in our experiments, we tested whether this phosphatase functions downstream of ␤-catenin. We found that target gene activation by stabilized ␤-catenin is potentiated by PP2A-C and blocked by Smt. Therefore, PP2A can regulate the pathway via a mechanism independent of ␤Ϫcatenin stabilization.
We next tested the possibility that PP2A acts at the level of Tcfs, which are necessary downstream components of Wnt signal transduction. Tcfs function as transcriptional repressors (29,45,48), but they can be converted into activators of Wnt target genes when complexed with ␤-catenin (19,20,49). In Xenopus, overexpression of XTcf-3 represses endogenous Wnt signaling in the dorsal marginal zone (45). We also found that XTcf-3 suppressed signaling by Xdsh or ␤-catenin ( Fig. 4 and data not shown). However, upon coinjection of XTcf-3 and PP2A-C, a Siamois reporter construct was markedly activated (Fig. 4). This result suggests that PP2A-C can promote the conversion of XTcf-3 from a repressor to an activator. We propose that endogenous PP2A-C levels are insufficient for activation of overexpressed XTcf-3, which therefore behaves as a repressor of Wnt signaling. However, when PP2A-C is provided in excess, XTcf-3 now behaves as a transcriptional activator. XTcf-3 is known to bind corepressor molecules such as Cterminal binding protein (45) or Groucho (50), and signaling by PP2A-C might relieve this repression. Although XTcf-3 may be a direct substrate for PP2A, it is also possible that one of the many pathways involving PP2A could integrate with Wnt signaling downstream of ␤-catenin. These possibilities remain to be tested by future studies.