Identification of β-catenin as a target of the intracellular tyrosine kinase PTK6

Protein tyrosine kinase 6 (PTK6) is an intracellular tyrosine kinase that is distantly related to Src family tyrosine kinases. Members of the PTK6 family are defined by a highly conserved exon structure that is distinct from other major intracellular tyrosine kinase families (Lee et al., 1998; Mitchell et al., 1997; Serfas and Tyner, 2003). Like Src family kinases, PTK6 is negatively regulated by phosphorylation of a C-terminal tyrosine residue (Derry et al., 2000; Qiu and Miller, 2002). Unlike Src family kinases, PTK6 is not myristoylated or specifically targeted to the membrane. As a consequence, PTK6 has been localized to different cellular compartments, including the nucleus, where it might have a distinct set of substrates and interacting proteins (Derry et al., 2003; Derry et al., 2000; Haegebarth et al., 2004).

We identified PTK6 in the mouse small intestine in a screen for factors that regulate epithelial cell differentiation, and originally named it Sik (Src-related intestinal kinase) (Siyanova et al., 1994). PTK6 was also identified in breast cancer cells where it is often referred to as BRK (breast tumor kinase) (Mitchell et al., 1994) and in cultured human melanocytes (PTK6) (Lee et al., 1993). PTK6 is predominantly expressed in epithelial cells of the skin, gastrointestinal tract (Llor et al., 1999; Vasioukhin et al., 1995; Wang et al., 2005), prostate (Derry et al., 2003) and oral epithelia (Petro et al., 2004). PTK6 is also reported to be expressed in lymphocytes (Kasprzycka et al., 2006).

PTK6 expression is developmentally regulated and detected only late in gestation in the mouse, when epithelial linings mature (Vasioukhin et al., 1995). In mature tissues, PTK6 is expressed in differentiated nondividing cells, with the highest levels in linings of the gastrointestinal tract in both humans and mice (Llor et al., 1999; Vasioukhin et al., 1995). Overexpression of PTK6 in mouse keratinocytes results in increased expression of the differentiation marker filaggrin during calcium-induced differentiation, suggesting a positive role in differentiation (Vasioukhin and Tyner, 1997). PTK6 was also shown to positively regulate expression of keratin 10 during differentiation of human cultured keratinocytes (Wang et al., 2005). Increased proliferation and impaired enterocyte differentiation in the PTK6-deficient mouse model provided further evidence that this kinase promotes epithelial cell differentiation in vivo (Haegebarth et al., 2006).

During development, Wnt signaling has an essential role in establishing the stem cell zone in the intestinal crypts (Korinek et al., 1998). In adult tissues, active Wnt signaling characterizes intestinal epithelial progenitor cells (van de Wetering et al., 2002). Activation of canonical Wnt signaling results in accumulation of nuclear β-catenin, which, in complex with T-cell factor (TCF) family members, controls proliferation versus differentiation in intestinal epithelial cells (Batlle et al., 2002; Pinto et al., 2003). Expansion of the progenitor zone in PTK6-deficient mice suggested that Wnt signaling might be affected in the intestines of these mice. Indeed, increased numbers of cells that were positive for nuclear β-catenin were observed in intestines of PTK6-deficient mice compared with their wild-type counterparts (Haegebarth et al., 2006). We explored the ability of PTK6 to regulate components of the Wnt-signaling pathway, and discovered a direct relationship between PTK6 and β-catenin. We demonstrate that PTK6 directly associates with β-catenin and that PTK6 expression leads to an inhibition of β-catenin regulated transcription in vivo and an increase in the levels of TCF4 and the co-repressor TLE/Groucho.

The increase in nuclear β-catenin observed in the intestines of PTK6-deficient mice (Haegebarth et al., 2006) led us to investigate whether PTK6 is involved in the Wnt-signaling pathway. To determine if β-catenin associates with or is a substrate of PTK6, HEK293 cells were co-transfected with wild-type PTK6, constitutively active PTK6 (PTK6 YF) or kinase-dead PTK6 (PTK6 KM) and full-length human β-catenin. In cells expressing comparable levels of PTK6, we observed an overall increase in protein tyrosine phosphorylation detected by anti-phosphotyrosine (PY) antibodies in cells expressing active forms of PTK6 (Fig. 1A). To determine whether β-catenin was one of the proteins phosphorylated in these cells, tyrosine-phosphorylated proteins were immunoprecipitated from Triton-X-100-soluble cell lysates (PY, Fig. 1B). Tyrosine-phosphorylated β-catenin was immunoprecipitated from cells expressing PTK6 or PTK6 YF and not from cells lacking PTK6 (vector) or from cells expressing PTK6 KM (β-catenin, Fig. 1B). Wild-type PTK6, and to a lesser extent PTK6 YF, were also immunoprecipitated with the anti-phosphotyrosine antibodies (PTK6, Fig. 1B).

To provide further evidence that β-catenin is tyrosine phosphorylated in cells expressing active forms of PTK6, β-catenin was immunoprecipitated from cell lysates using a β-catenin-specific antibody. Although relatively equal levels of β-catenin were immunoprecipitated from cell lysates (β-catenin, Fig. 1C), only β-catenin immunoprecipitated from cells expressing PTK6 or PTK6 YF was tyrosine phosphorylated (PY blot, asterisk, Fig. 1C). Several other tyrosine-phosphorylated proteins, including PTK6, co-immunoprecipitated with β-catenin (PTK6, Fig. 1C). Association of PTK6 with β-catenin was independent of its kinase activity, because all forms of PTK6 were found in a complex with β-catenin (PTK6, Fig. 1C).

Wnt-dependent and -independent signaling can lead to the stabilization of the cytoplasmic pool of β-catenin (Lu and Hunter, 2004). This stabilization results in nuclear accumulation of β-catenin and increased β-catenin/TCF-regulated transcription. To determine whether PTK6 influenced nuclear localization of β-catenin, HEK293 cells expressing PTK6, PTK6 YF or PTK6 KM and β-catenin were fractionated into nuclear and cytoplasmic fractions. All forms of PTK6 were detected in the nuclear and cytoplasmic fractions (PTK6, Fig. 2A) and a robust increase in tyrosine phosphorylation was only observed in fractions from cells expressing PTK6 or PTK6 YF (PY, Fig. 2A). β-catenin also localized to both nuclear and cytoplasmic fractions, and the levels of β-catenin in these fractions were not significantly changed in cells co-expressing PTK6 (β-catenin, Fig. 2A).

Immunoprecipitation of β-catenin from both nuclear and cytoplasmic fractions revealed increased tyrosine phosphorylation of β-catenin in cells expressing PTK6 or PTK6 YF (PY, Fig. 2B). Probing the β-catenin immunoprecipitates with the Myc epitope tag antibody, which detects both ectopic β-catenin and ectopic PTK6, showed that PTK6 associates with β-catenin in both nuclear and cytoplasmic fractions (Myc, Fig. 2B). All forms of PTK6 were found in a complex with nuclear and cytoplasmic β-catenin (PTK6, Fig. 2B).

The ability of PTK6 to complex with nuclear β-catenin led us to investigate whether PTK6 could regulate β-catenin transcriptional activity. Using the Super8XTOPFlash reporter construct we examined transcriptional activity of β-catenin in HEK293 cells co-expressing PTK6, PTK6 YF or PTK6 KM and β-catenin. Although all forms of PTK6 significantly inhibited β-catenin transcriptional activity (Fig. 2C), PTK6 YF consistently inhibited β-catenin transcriptional activity to the greatest extent, suggesting that there is a kinase-dependent component of transcriptional inhibition. The ability of the kinase-dead form of PTK6 to also inhibit β-catenin transcriptional activity suggests that a kinase-independent component of transcriptional inhibition also exists.

In cells co-expressing active forms of PTK6 and β-catenin, there is an increase in tyrosine phosphorylation of β-catenin. To address whether PTK6 directly phosphorylates β-catenin, an in vitro kinase assay was performed using active recombinant PTK6 and recombinant β-catenin. Activated PTK6 was detected in all samples by phosphotyrosine-specific antibodies (PY, arrowheads, Fig. 3A,D). In the presence of ATP, PTK6 directly phosphorylated wild-type β-catenin (WT) in vitro (PY, asterisk, Fig. 3A).

Co-immunoprecipitation experiments showed that PTK6 and β-catenin can be found in a complex (Fig. 1C, Fig. 2B). To determine whether this association is direct or indirect, recombinant β-catenin was incubated with GST-fusion proteins of full-length PTK6, PTK6 SH2, SH3 and SH2-SH3 domains. To explore the impact of tyrosine phosphorylation, β-catenin was incubated with active recombinant PTK6 in the presence and absence of ATP. Unphosphorylated β-catenin (—ATP) directly associated with full-length (FL) PTK6 (Fig. 3B, PTK6 FL, —ATP). Phosphorylation of β-catenin by PTK6 increased the interaction between β-catenin and PTK6 GST-fusion proteins containing the PTK6 SH2 domain, which binds phosphorylated tyrosine residues (Fig. 3B, PTK6 SH2, SH2/SH3 and FL, +ATP). These data suggest that PTK6 and β-catenin directly associate and that this interaction is increased after phosphorylation of β-catenin by PTK6.

Analysis of phosphorylated β-catenin (Fig. 3A) using mass spectrometry, identified tyrosine phosphorylation events at tyrosine residues Y64, Y142 and Y331 or Y333 (Y331,333) (Fig. 3C). Phosphorylation on Y489, Y654 or Y670, residues that are phosphorylated by Abl (Y489), EGFR and Src (Y654) (reviewed by Lilien and Balsamo, 2005) and c-Met (Y654, Y670) (Zeng et al., 2006), was not detected.

Using site-directed mutagenesis, we generated GST-β-catenin fusion proteins with single (Y64F, Y142F), double (Y64F/Y142F, Y142F/Y331,333F) and triple (Y64F/Y142F/Y331,333F) point mutations to validate the tyrosine phosphorylation sites identified by mass spectrometry. β-catenin point mutants with a mutation at Y64 were no longer phosphorylated by PTK6 (Fig. 3D), suggesting that Y64 is the key residue phosphorylated by PTK6 in vitro. Single (Y64F, Y142F, Y331F), double (Y64F/Y142F) and triple (Y64F/Y142F/Y331,333F) point mutations were also introduced into a β-catenin mammalian expression construct to assess phosphorylation of these mutants in HEK293 cells. Immunoblot analysis of immunoprecipitated β-catenin revealed that β-catenin point mutants with a mutation at Y64 were no longer phosphorylated by PTK6 YF in these cells (PY, Fig. 4A). Taken together, these data suggest that β-catenin Y64 is the major site phosphorylated by PTK6, not only in vitro, but also in vivo, and that phosphorylation of Y64 might be necessary for PTK6 to phosphorylate secondary sites in β-catenin.

Phosphorylation of β-catenin on tyrosine residues is well recognized to be important for regulating β-catenin activity and intracellular associations (reviewed by Lilien and Balsamo, 2005). PTK6 is the first tyrosine kinase identified to phosphorylate Y64 of β-catenin. We therefore wanted to investigate the impact of phosphorylation of Y64 by PTK6 on β-catenin transcriptional activity. Using the Super8XTOPFlash reporter construct we determined the transcriptional activity of the β-catenin Y64F point mutant in HEK293 cells. Similarly to wild-type β-catenin, all forms of PTK6 inhibited the transcriptional activity of the β-catenin Y64F (Fig. 4B). PTK6 also inhibited the transcriptional activity of the β-catenin Y142F, Y64F/Y142F and Y64F/Y142F/Y331,333F point mutants (Fig. 4C). In all cases, PTK6 YF inhibited β-catenin/TCF-mediated transcription to the greatest extent, suggesting that although the direct phosphorylation of β-catenin is not required for the inhibition of β-catenin transcriptional activity, a kinase-dependent component of transcriptional inhibition still exists and that PTK6 might have additional, as yet unidentified, substrates that participate in the regulation of β-catenin/TCF transcription.

Although PTK6 localizes to different cellular compartments, it lacks nuclear localization and myristoylation or palmitoylation signals. In HEK293 cells, ectopically expressed PTK6 localized to the nucleus and cytoplasm and was able to inhibit ectopic β-catenin transcriptional activity. To determine whether the subcellular localization of PTK6 influenced its ability to inhibit β-catenin/TCF-mediated transcription, we assessed the transcriptional activity of endogenous β-catenin in the colorectal adenocarcinoma cell line SW620 expressing nuclear-targeted (NLS) PTK6, PTK6 YF or PTK6 KM. SW620 cells expressed low levels of endogenous PTK6 (Fig. 5B, Vector lanes) whereas ectopic PTK6 was expressed at high levels (Fig. 5B, D, PTK6). NLS.PTK6, and to a greater extent NLS.PTK6 YF, inhibited endogenous β-catenin transcriptional activity in this system (Fig. 5A). Similarly to what was observed in HEK293 cells, expression of nuclear-targeted PTK6 did not alter the levels or subcellular distribution of endogenous β-catenin in these cells (Fig. 5B). We next wanted to test the effects of tethering PTK6 to the membrane. Using SW620 cells, we determined the transcriptional activity of endogenous β-catenin in cells expressing membrane targeted (Palm) PTK6, PTK6 YF or PTK6 KM. Contrary to what we observed with NLS.PTK6, Palm.PTK6, and to a greater extent Palm.PTK6 YF, increased endogenous β-catenin transcriptional activity in this system (Fig. 5C). Expression of membrane-targeted PTK6 also did not alter the levels or subcellular distribution of endogenous β-catenin in these cells (Fig. 5D).

Although the levels of nuclear β-catenin were not affected by ectopic expression of PTK6, we wanted to examine the effects of PTK6 expression on other components of the β-catenin/TCF4 transcriptional complex. Expression of nuclear-targeted PTK6 in SW620 cells resulted, on average, in a twofold increase in nuclear TCF4 levels (Fig. 5E). More importantly, we also observed a twofold increase in the nuclear levels of the transcriptional co-repressor TLE/Groucho (Fig. 5E). Expression of all forms of NLS.PTK6 resulted in an increase in the levels of TCF4 and TLE/Groucho, suggesting that the observed increase in these proteins does not depend on PTK6 kinase activity.

Expression of nuclear-targeted and membrane-targeted PTK6 in SW620 cells had opposing effects on β-catenin/TCF-mediated transcription. To assess the physiological role of PTK6 in regulating β-catenin transcriptional activity, we knocked down endogenous PTK6 in SW620 cells using two different shRNAs (shRNA49, shRNA52) and established stable SW620 cells lines (Fig. 6B). When the transcriptional activity of β-catenin was assessed in these cell lines, an increase in endogenous β-catenin transcriptional activity was observed (Fig. 6A). Next, we determined whether the observed increase in transcriptional activity correlated with increased expression of β-catenin/TCF target genes Myc (He et al., 1998) and Survivin (Zhang et al., 2001). Although the levels of β-catenin were not increased in the PTK6-knockdown cell lines, there was a threefold increase in the expression of Myc and a twofold increase in the expression of Survivin (Fig. 6B).

To further explore the impact of PTK6 on β-catenin/TCF-regulated transcription in vivo, Ptk6-null mice were crossed with BAT-GAL transgenic mice. BAT-GAL mice contain the LacZ gene that encodes β-galactosidase under the control of β-catenin/TCF response elements (Maretto et al., 2003). In the distal colons of Ptk6^−/− BAT-GAL mice, increased numbers of crypts expressed the LacZ gene (Fig. 7A). Positive crypts were readily detected in whole-mount-stained colons from Ptk6^−/− mice, but not in wild-type mice (Fig. 7Aa,b). Comparison of cross sections of distal colons from Ptk6^+/+ and Ptk6^−/− BAT-GAL mice, revealed that entire crypts were populated with LacZ-expressing cells in Ptk6^−/− BAT-GAL animals (Fig. 7Ad,f). This suggests that PTK6 has a role in suppressing β-catenin/TCF-regulated transcription in stem or progenitor cells in the colon.

In the small intestine, β-catenin/TCF-regulated transcription was extinguished in most cells in both Ptk6^+/+ and Ptk6^−/− BAT-GAL mice. However single cells expressing the LacZ gene could be detected in numerous crypts of Ptk6^−/− BAT-GAL mice, but these were rarely observed in the Ptk6^+/+ BAT-GAL small intestine (Fig. 7B). The β-galactosidase-positive cells in the BAT-GAL small intestine contained granules characteristic of Paneth cells, differentiated epithelial granulocytes that are localized at the base of the crypts (Fig. 7Bc,d).

In normal adult epithelial linings, PTK6 is most highly expressed in nondividing differentiated epithelial cells (Derry et al., 2003; Haegebarth et al., 2006; Llor et al., 1999; Petro et al., 2004; Vasioukhin et al., 1995; Wang et al., 2005). Characterization of the Ptk6-null mouse revealed that PTK6 has a role in promoting differentiation of enterocytes in the small intestine. Increased intestinal epithelial cell turnover, growth and impaired enterocyte differentiation correlated with increased levels of active Akt in Ptk6-null mice (Haegebarth et al., 2006). In addition to increased activation of Akt, nuclear β-catenin was more readily detected in the Ptk6-null mice, suggesting that PTK6 functions to negatively regulate β-catenin transcriptional activity. Here, we demonstrate that expression of PTK6 leads to an inhibition of β-catenin/TCF-regulated transcription.

The importance of Wnt signaling and β-catenin-regulated transcription in intestinal epithelial cell renewal has been well established (reviewed by Gregorieff and Clevers, 2005). Inhibition of β-catenin signaling might be a prerequisite for differentiation and normal tissue homeostasis. Although TCF-4 has an essential role in establishing the intestinal stem cell zone in the developing crypts (Korinek et al., 1998), high levels of TCF-4 are expressed in differentiated nondividing epithelial cells, where it might function to repress activation of genes positively regulated by β-catenin (Gregorieff et al., 2005). PTK6 expression is also high in differentiated nondividing epithelial cells of the gastrointestinal tract (Vasioukhin et al., 1995). In Ptk6-null animals, enterocyte differentiation was delayed (Haegebarth et al., 2006). A recently established Ptk6-null epithelial cell line derived from colonic mucosa displays characteristics of progenitor cells (Whitehead et al., 2008), whereas an almost twofold increase in PTK6 message was observed in intestinal stem cells that were induced to differentiate by the deletion of β-catenin (Fevr et al., 2007).

We demonstrate that PTK6 associates with and phosphorylates signaling pools of β-catenin and can negatively regulate β-catenin/TCF transcriptional activity. Although direct phosphorylation of β-catenin is not essential for PTK6-mediated negative regulation of β-catenin/TCF transcription, the ability of PTK6 to associate with nuclear β-catenin might still have a role in the inhibition of β-catenin transcriptional activity. The adenomatous polyposis coli (APC) tumor suppressor protein inhibits β-catenin transcriptional activity by sequestering it away from promoter targets in the nucleus (Sierra et al., 2006). Similarly, PTK6 might compete for binding of other positive factors, leading to decreased transcription. Alternatively, PTK6 might prevent β-catenin from effectively competing away transcriptional co-repressors such as TLE/Groucho from TCF/Lef proteins.

Earlier, we proposed that intracellular localization of PTK6 might have a crucial role in determining the outcomes of PTK6-regulated signaling (Derry et al., 2003; Haegebarth et al., 2004). Although it lacks a myristoylation or palmitoylation signal, and a nuclear localization signal, PTK6 has been shown to have both membrane-associated and nuclear substrates. Recently Kim and Lee showed that targeting PTK6 to the plasma membrane enhanced oncogenic signaling, whereas targeting PTK6 to the nucleus inhibited its oncogenic functions in HEK293 cells (Kim and Lee, 2009). Here, we show that targeting PTK6 to the membrane or nucleus has opposing effects on endogenous β-catenin/TCF-regulated transcription in SW620 cells. Although nuclear-targeted PTK6 negatively regulates endogenous β-catenin/TCF transcriptional activity, membrane-targeted PTK6 results in activation of β-catenin/TCF-regulated transcription. Interestingly, expression of nuclear-targeted PTK6 leads to increased levels of TCF4 and TLE/Groucho (Fig. 5E), which form a repressor complex that inhibits target gene expression.

The importance of proper β-catenin/TCF signaling is underscored by the variety of mechanisms that have evolved to regulate this signaling pathway. PTK6 and the Kruppel-like factor 4 transcription factor (KLF4) share similar patterns of expression in the normal intestine, with expression being localized to differentiated epithelial cells (Haegebarth et al., 2006; Zhang et al., 2006). KLF4 was found to interact with the transcriptional activation domain of β-catenin, leading to inhibition of β-catenin-regulated transcription (Zhang et al., 2006). Here, we demonstrate that similarly to KLF4, PTK6 can associate with and inhibit β-catenin-regulated transcription.

Although phosphorylation of β-catenin on serine and threonine is well recognized as important for regulating intracellular signaling pools of β-catenin, its phosphorylation on tyrosine is also important for modulation of its activities and intracellular associations (reviewed by Daugherty and Gottardi, 2007; Lilien and Balsamo, 2005). Tyrosine kinases that can phosphorylate β-catenin include the intracellular kinases Abl (Rhee et al., 2002), Src (Behrens et al., 1993), Fyn and Fer (Piedra et al., 2003), as well as the epidermal growth factor receptor (EGFR) (Hoschuetzky et al., 1994) and the hepatocyte growth factor receptor Met (Monga et al., 2002). The Bruton's tyrosine kinase (BTK) has recently been identified as a negative regulator of the Wnt-signaling pathway. BTK increases the levels of the polymerase-associated factor transcriptional elongation complex member CDC73, which acts as a repressor of β-catenin/TCF transcription (James et al., 2009). Similarly to BTK, we show that nuclear PTK6 enhances expression of TCF4 and TLE/Groucho, a well-characterized corepressor of Wnt/β-catenin-regulated transcription.

Using a solid-phase peptide library, Shin and colleagues (Shin et al., 2008) identified a consensus sequence for PTK6 phosphorylation. PTK6 preferentially phosphorylated peptides with acidic amino acids at the C-terminus of the sequence [xxY(D/E)x or xxYx(D/E)]. Based on our in vitro phosphorylation data, we identified several residues in β-catenin that can be phosphorylated by PTK6 (Y64, Y142, and Y331 and/or Y333). Analysis of the sequence surrounding these tyrosine residues showed that three of the four residues are within the preferred phosphorylation sequence of PTK6 [Y64 (xxYEx), Y142 (xxYxD) and Y333 (xxYEx)].

A recent report identified Y64, along with Y86 in β-catenin, as major phosphorylation sites in F9 cells that lack α-catenin, following treatment with phosphatase inhibitors (Tominaga et al., 2008). Expression of a Y64 phosphomimetic (Y64E) resulted in decreased β-catenin transcriptional activity compared with wild-type protein. The endogenous kinase responsible for phosphorylating β-catenin at Y64, however, was not identified. Here we demonstrate that PTK6 preferentially phosphorylates β-catenin at Y64, and phosphorylation at this site might allow for phosphorylation at secondary sites.

In addition to Y64, Y142, Y331 and/or Y333 were identified as sites phosphorylated by PTK6. Similarly to PTK6, Fyn, Fer (Piedra et al., 2003) and Met (Brembeck et al., 2004) phosphorylate β-catenin on Y142. Phosphorylation of β-catenin by Fyn and Fer, which are also localized within the adherens complex, negatively regulates β-catenin binding to α-catenin, leading to decreased adhesion (Piedra et al., 2003). Met-regulated phosphorylation of β-catenin Y142 was reported to lead to its increased association with Bcl9-2, promoting transport to the nucleus and enhancing transcription (Brembeck et al., 2004). Phosphorylation of β-catenin on Y331/Y333, Y654 and Y670 has been reported in the Caco-2 colon carcinoma cell line following treatment with acetaldehyde, which induced disruption of the E-cadherin/β-catenin complex (Sheth et al., 2007). Although PTK6 can phosphorylate β-catenin on several tyrosine residues (Y142, Y331/333) that have been linked to decreased adhesion and increased signaling, the lack of reliable phosphospecific antibodies for these sites has impeded our ability to determine whether these phosphorylation events occur in a physiological setting.

The inverse correlation between PTK6 expression and nuclear β-catenin in the gastrointestinal tract makes PTK6 an attractive candidate for a negative regulator of β-catenin. Here, we show that PTK6 negatively regulates endogenous β-catenin/TCF transcriptional activity in SW620 cells. Similarly to what we previously observed in the PTK6 knockout animals (Haegebarth et al., 2006), loss of PTK6 in SW620 cells resulted in increased β-catenin/TCF transcription and increased expression of the β-catenin target genes Myc and Survivin.

We found that PTK6 is a negative regulator of β-catenin transcription in the small and large intestine in BAT-GAL reporter mice. Although β-catenin/TCF-driven reporter gene expression is evident during development and in intestinal adenomas, little activity can be detected in the mature intestinal epithelium (Maretto et al., 2003). The absence of strongly positive β-galactosidase-expressing cells in wild-type BAT-GAL mice suggests that activation of β-catenin/TCF transcription is transient or very weak in adult crypts and the reporter enzyme activity cannot be easily detected at the level of sensitivity of the system. However, disruption of the Ptk6 gene in BAT-GAL transgenic animals resulted in increased expression of the β-galactosidase reporter in epithelial cells of the mature colon and small intestine (Fig. 7). In the distal colon, entire crypts were populated by β-galactosidase-positive cells, suggesting that PTK6 is important for modulating Wnt signaling in either stem or progenitor cells in this region of the intestine. By contrast, enhanced LacZ expression in the small intestine was detected in single cells in the small intestine, the majority of which had granules that are characteristics of Paneth cells, a cell type whose differentiation is positively regulated by Wnt signaling and β-catenin/TCF activity (Andreu et al., 2005; van Es et al., 2005).

Although PTK6 inhibits β-catenin-regulated transcription in normal intestinal cells, its role in cancer might be different. PTK6 is not expressed in the normal mammary gland, but it is frequently induced in breast cancers and has been implicated in promoting oncogenic signaling (reviewed by Harvey and Crompton, 2004). In normal prostate epithelial cells, PTK6 is largely found in the nucleus, but in prostate cancer cells it is relocalized to the cytoplasm (Derry et al., 2003). Even in the normal small intestine, PTK6 has been shown to have distinct functions under different conditions. Although PTK6 inhibits growth and promotes epithelial cell differentiation during normal intestinal tissue homeostasis, it is induced in proliferating progenitor cells of the small intestinal crypts following irradiation-induced DNA damage, where it promotes apoptosis (Haegebarth et al., 2009). It is becoming clear that functions of PTK6 might differ depending on environmental conditions, the cell type in which it is expressed, its expression level and its intracellular localization (reviewed by Brauer and Tyner, 2009).

Full-length wild-type PTK6, PTK6 YF and PTK6 KM constructs in the pRcCMV vector were described previously (Kamalati et al., 1996) and were a gift from Mark Crompton (Royal Holloway University of London, Surrey, UK). PTK6 YF has a substitution of the regulatory tyrosine, Y447, to phenylalanine, resulting in a constitutively active mutant. PTK6 KM has a substitution of a critical lysine (K219) in the ATP binding site to methionine resulting in a kinase dead mutant. Coding sequences from the pRcCMV constructs were subcloned into the pcDNA3 vector containing a Myc epitope tag (Invitrogen). PTK6 constructs tagged with an SV40 nuclear localization signal (NLS) were generated in the pcDNA4/TO vector (Invitrogen). Duplexed oligonucleotides encoding a Myc epitope tag (bold) and the SV40 NLS (underlined) (5′-AGCTCATGGAACAAAAGCTGATTAGCGAAGAGGACCTGCCTAAAAAGAAGCGTAAAGTGAACCGGTGCTAGCA-3′; 5′-AGCTTGCTAGCACCGGTTCACTTTACGCTTCTTTTTAGGCAGGTCCTCTTCGCTAATCAGCTTTTGTTCCATG-3′) were introduced into the pcDNA4/TO vector. The coding sequences for wild-type PTK6, PTK6 YF and PTK6 KM were subcloned into the vector, and constructs were verified by sequencing. PTK6 constructs tagged with the Lyn tyrosine kinase myristoylation/palmitoylation sequence (Palm) were generated in the pcDNA4/TO vector. Duplexed oligonucleotides encoding a Myc epitope tag (bold) and the Lyn Palm sequence (underlined) (5′-AGCTCATGGGCTGCATCAAGAGCAAGAGGAAGGAACAAAAGCTGATTAGCGAAGAGGACCTGAACCGGTGCTAGCA-3′; 5′-AGCTTGCTAGCACCGGTTCAGGTCCTCTTCGCTAATCAGCTTTTGTTCCTTCCTCTTGCTCTTGATGCAGCCCATG-3′) were introduced into the pcDNA4/TO vector. The coding sequences for wild-type PTK6, PTK6 YF and PTK6 KM were subcloned into the vector, and constructs were verified by sequencing. GST-tagged mouse PTK6 fusion constructs (full-length, SH2, SH3 and SH2-SH3 domains), in the bacterial expression vector pGEX KG, were described previously (Vasioukhin and Tyner, 1997). Wild-type Myc-tagged human β-catenin, in the mammalian expression vector pCAN, was provided by Paul Polakis (Genentech). Wild-type GST-tagged mouse β-catenin and the GST-tagged β-catenin Y142F point mutant, in the bacterial expression vector pGEX KG, were provided by Jack Lilien (University of Iowa, Iowa City, IA). The Super8XTOPFlash (TOPFlash) luciferase reporter construct containing eight TCF/Lef binding sites and the control Super8XFOPFlash (FOPFlash) construct containing eight mutated TCF/Lef binding sites (Veeman et al., 2003) were a gift from Randall Moon (University of Washington, Seattle, WA) and were used to assess β-catenin transcriptional activity. The packaging plasmids HIVtrans and VSVG (Feng et al., 2007) were provided by Bin He (University of Illinois at Chicago, Chicago, IL). The pRL-TK construct was purchased from Promega.

Human embryonal kidney (HEK) 293 cells (ATCC CRL-1573) and the human colorectal adenocarcinoma cell line SW620 (ATCC CRL-227) were cultured according to ATCC recommendations. Transfections of HEK293 cells were performed using the Lipofectamine Transfection Reagent in combination with PLUS Reagent (Invitrogen) as per the manufacturer's instructions. For immunoprecipitation and fractionation experiments, HEK293 cells were transfected with DNA encoding PTK6 (pcDNA3.Myc.PTK6) and β-catenin (pCAN.Myc.β-catenin). Transfections of SW620 cells were performed using Lipofectamine 2000 Transfection Reagent (Invitrogen) as per the manufacturer's instructions. For fractionation experiments, SW620 cells were transfected with NLS.PTK6 (pcDNA4/TO.Myc/NLS.PTK6) or Palm.PTK6 (pcDNA4/TO.Myc/Pal.PTK6) DNA. To examine the effects of PTK6 on wild type and mutant β-catenin transcriptional activity, HEK293 cells were transfected with Super8XTOPFlash or Super8XFOPFlash DNA, β-catenin (pCAN.Myc.β-catenin) DNA, and PTK6 (pcDNA3.Myc.PTK6) DNA and pRL-TK DNA as an internal transfection control. To examine the effects of PTK6 on endogenous β-catenin transcriptional activity, SW620 cells were transfected with Super8XTOPFlash or Super8XFOPFlash DNA and NLS.PTK6 or Pal.PTK6 DNA and pRL-TK DNA as an internal transfection control.

HEK293 cells or SW620 cells were transfected as stated above and lysed 18-24 hours after transfection and luciferase activity was measured using the Dual-Luciferase Reporter Assay System (Promega) and the Clarity Microplate Luminometer (Bio-Tek Instruments). Reporter expression was normalized to co-transfected Renilla luciferase activity.

The MISSION TRC shRNA Target Set directed against PTK6 was purchased from Sigma. Lentivirus expressing TRCN0000021549 (shRNA 49), TRCN0000021552 (shRNA 52) and empty vector were produced in the HEK293FT packaging cell line by co-transfection with compatible packaging plasmids HIVtrans and VSVG (Feng et al., 2007). SW620 cells were infected with retrovirus (50% viral supernatant and 50% growth medium containing 5 μg/ml polybrene) and placed in selection medium containing 2 μg/ml puromycin for 2 weeks.

Human PTK6 [Brk (C-18)], phosphotyrosine [p-Tyr (PY20)], Sp1 (PEP 2), E-cadherin (H-108), Myc (N-262) and Survivin (FL-142) antibodies were purchased from Santa Cruz Biotechnology. The anti-phosphotyrosine, clone 4G10 antibody was purchased from Upstate. A β-catenin monoclonal antibody (clone 14) was purchased from BD Transduction Laboratories. The TCF4 antibody (C9B9), the TLE1/2/3/4 antibody and an antibody directed against the Myc epitope tag (9B11) were purchased from Cell Signaling Technology. A monoclonal antibody against GST (4C10) was purchased from Covance. α-tubulin and β-actin (AC-15) antibodies were purchased from Sigma. Donkey anti-rabbit or sheep anti-mouse antibodies conjugated to horseradish peroxidase were used as secondary antibodies (Amersham Biosciences) and detected by chemiluminescence with SuperSignal West Dura Extended Duration Substrate from Pierce.

Protein fractionation from transfected HEK293 cells was done using the NE-PER Nuclear and Cytoplasmic Extraction Reagents from Pierce according to the manufacturer's instructions. Protein fractionation from transfected SW620 cells was done using the ProteoExtract Subcellular Proteome Extraction Kit from Calbiochem according to manufacturer's instructions. One-tenth volume of each fraction was subjected to SDS-PAGE and transferred onto Immobilon-P membranes (Millipore) for immunoblotting. Alternatively, transfected HEK293 cells were lysed in 1% Triton X-100 lysis buffer (1% Triton X-100, 20 mM HEPES, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 10 mM sodium pyrophosphate, 100 mM NaF, 10 mM Na₃VO₄, 5 mM iodoacetic acid, 0.2 mM PMSF, protease inhibitor cocktail). All transfected cell were harvested 18-24 hours after transfection. Tyrosine-phosphorylated proteins were immunoprecipitated using a cocktail of anti-phosphotyrosine antibodies (4G10 and PY20). Changes in proteins levels were quantified using the ImageJ program.

Recombinant human PTK6 (Invitrogen) alone or in combination with recombinant human β-catenin (Upstate), was incubated in kinase buffer (10 mM HEPES, pH 7.5, 150 mM NaCl, 2.5 mM DTT, 0.01% Triton X-100, 10 mM MnCl₂) with or without 200 μM ATP for 10 minutes at 30°C. One-tenth of each reaction was subjected to SDS-PAGE and transferred onto Immobilon-P membranes (Millipore) for immunoblotting. The remainder of each reaction was resolved on a Tris-Glycine pre-cast gel (Invitrogen). Proteins were visualized with Coomassie Brilliant Blue R-250 Staining Solution (BioRad). β-catenin was excised from the gel and the gel slice was washed in 50% acetonitrile in preparation for mass spectrometry.

Phosphorylated GST-tagged human β-catenin was subjected to in-gel digestion using both trypsin and chymotrypsin separately followed by C₁₈ reversed-phase microcapillary LC/MS/MS using a LTQ 2D linear ion trap mass spectrometer (Thermo Scientific) in positive ion data-dependent acquisition mode. MS/MS spectra were searched against the reversed Swissprot protein database using Sequest (Proteomics Browser, Thermo Scientific) with differential modifications for STY phosphorylation (+79.97) and methionine oxidation (+15.99). Phosphorylation sites were identified if they initially passed the following Sequest scoring thresholds: 2+ ions, Xcorr≥2.0, Sf≥0.4, P>0; 3+ ions, Xcorr≥2.75, Sf≥0.5, P>0 against the forward database. Peptides with gas phase charges of 1+ and 4+ were not accepted owing to difficulty of interpretation. Passing MS/MS spectra were then manually inspected to be sure that all b- and y- fragment ions aligned with the assigned protein database sequence. Determination of the exact sites of phosphorylation was aided using GraphMod software (Proteomics Browser, Thermo Scientific), resulting in a false-positive identification rate of less than 2%.

Using the QuikChange XL Site-Directed Mutagenesis Kit (Stratagene) we introduced single, double, and triple point mutations (tyrosine to phenylalanine) into Myc-tagged human β-catenin in the mammalian expression vector pCAN, and GST-tagged mouse β-catenin in the bacterial expression vector pGEX KG according to manufacturer's instructions. Mutagenic primers used for human β-catenin include: Y64F, forward, 5′-TGGATACCTCCCAAGTCCTGTTTGAGTGGGAAC-3′ and reverse, 5′-GTTCCCACTCAAACAGGACTTGGGAGGTATCCA-3′; Y142F, forward, 5′-GTAAACTTGATTAACTTTCAAGATGATGCAGAA-3′ and reverse, 5′-GTTCTGCATCATCTTGAAAGTTAATCAAGTTTAC-3′; Y331F/Y333F, forward, 5′-CCAAGCTTTAGTAAATATAATGAGGACCTTTACTTTCGAAAAACTACTGTGGAC-3′ and reverse, 5′-GTCCACAGTAGTTTTTCGAAAGTAAAGGTCCTCATTATATTTACTAAAGCTTGT-3′. Mutagenic primers used for mouse β-catenin include Y64F, forward, 5′-GACACCTCCCAAGTCCTTTTTGAATGGGAGCAAGG-3′ and reverse, 5′-CCTTGCTCCCATTCAAAAAGGACTTGGGAGGTGTC-3′; Y331F/Y333F forward, 5′-AGTAAACATAATGAGGACCTTCACTTTTGAGAAGCTTCTGTGGAC-3′ and reverse, 5′-GTCCACAGAAGCTTCTC AAAAGTGAAGGTCCTCATTATGTTTACT-3′. All constructs were sequenced to verify point mutations.

GST-tagged mouse β-catenin proteins (wild type and point mutant) as well as GST-tagged mouse PTK6 proteins (full-length, SH2, SH3 and SH2/SH3 domains) were generated in bacteria. Cleared sonicated bacterial cell lysates were run over a glutathione-Sepharose 4B column (GE Healthcare) and proteins were eluted with reduced glutathione in 50 mM Tris-HCl, pH 8.0. Eluted proteins were dialyzed against 4 litre storage buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1 mM EDTA, 0.05% Triton X-100, 5 mM DTT, 25% glycerol). In vitro kinase assays were performed using recombinant human PTK6 (Invitrogen) and the mouse GST.β-catenin wild-type and point mutant proteins, as described above. For GST pull-down experiments β-catenin was cleaved from the GST tag by incubation in thrombin (Sigma) and the eluted protein was dialyzed against cold PBS and then against storage buffer as above. The cleaved β-catenin was incubated with recombinant human PTK6 (Invitrogen) and an in vitro kinase assay was first performed in the presence or absence of ATP. Mouse PTK6 GST fusion proteins (1 μg full-length, SH2, SH3 and SH2/SH3 domains) were incubated with glutathione-Sepharose 4B beads for 30 minutes and then incubated with 20 ng phosphorylated or unphosphorylated β-catenin overnight at 4°C. GST pull-downs were washed four times in wash buffer (1% Triton X-100, 20 mM HEPES, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 10 mM Na-pyrophosphate). Samples were resuspended in 30 μl of 2× Laemmli sample buffer, boiled for 5 minutes and resolved by SDS-PAGE.

Female Ptk6^−/− mice in the C57BL/6J genetic background (Haegebarth et al., 2006) were crossed with male BAT-GAL transgenic mice described previously (Maretto et al., 2003) and purchased from the Jackson Laboratory [B6.Cg-Tg(BAT-lacZ)₃Picc/J] to generate Ptk6^+/− BAT-GAL animals. Heterozygous animals were crossed to generate Ptk6^+/+ BAT-GAL and Ptk6^−/− BAT-GAL mice. Adult animals were sacrificed and small intestines and colons were isolated, washed in fixative (1% formaldehyde, 0.2% glutaraldehyde, 0.02% NP-40 in PBS), and incubated in a β-galactosidase solution [5 mM K₃FE(CN)₆, 5 mM K₄Fe(CN) ^.₆3H₂O, 2 mM MgCl₂, 0.02% NP-40, 0.1% sodium deoxycholate, 1 mg/ml 5-bromo-4-chloro-3-indolyl-β-D-galactoside] overnight. Tissues were fixed overnight in 4% paraformaldehyde and paraffin blocks were prepared using standard methods. Tissue sections (5 μm) were prepared and counterstained with either Nuclear Fast Red or hematoxylin and eosin (Vector Laboratories).

The authors would like to thank Mark R. Crompton, Paul Polakis, Jack Lilien, Randall T. Moon and Bin He for their generous gifts, and Katherine Weaver for reviewing the manuscript. This work was supported by National Institutes of Health Grants DK44525 and DK068503 (A.L.T.). J.J.G. was supported by an AGA Foundation Graduate Student Research Fellowship Award and is supported by an NRSA/NIH Institutional T32 training grant, ‘Training Program in Signal Transduction and Cellular Endocrinology’, T32 DK07739 from the NIDDK. P.M.B. was supported by a DOD Predoctoral Traineeship Award, Army W81XWH-06-1-000. Deposited in PMC for release after 12 months.