| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
HORMONES AND SIGNALING
and GSK-3
modulate -catenin stability in hyperproliferating colonic epithelia
1Division of Gastroenterology and Hepatology, Department of Internal Medicine, University of Texas Medical Branch, Galveston; and 2Department of Integrative Biology, University of Texas Health Science Center, Houston, Texas
Submitted 27 July 2006 ; accepted in final form 17 October 2006
 |
ABSTRACT |
|---|
 TOP ABSTRACT EXPERIMENTAL PROCEDURESRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
and GSK-3
phosphorylate -catenin at Ser45 (-cat45) and Thr41/Ser37,33 (-cat33,37,41) residues, thereby facilitating its ubiquitination and proteasomal degradation. We used a Citrobacter rodentium-induced transmissible murine colonic hyperplasia (TMCH) model to determine Ser/Thr phosphorylation and biological function of -catenin during crypt hyperproliferation. TMCH was associated with 3-fold and 3.3-fold increases in CKI- cellular abundance and 2-fold and 1.8-fold increase in its activity at 6 and 12 days after infection, respectively. -Catenin coimmunoprecipitated with both cellular and nuclear CKI- and cellular axin at these time points. Cellular -catenin was constitutively phosphorylated at Ser45 and underwent subcellular redistribution to cytoskeletal and nuclear fractions at days 6 and 12 of TMCH, respectively. -cat33,37,41, however, exhibited only subtle changes in either phosphorylation status or subcellular distribution even after blocking proteasomal degradation in vivo. Interestingly, GSK-3 underwent increased phosphorylation at Ser9, leading to 40% and 70% decreases in its activity at these time points, respectively. Coimmunoprecipitation studies exhibited strong association of GSK-3 with PKC-
at either time point. Cellular -cat45 stabilized and, along with unphosphorylated -catenin, underwent nuclear translocation, associated with nuclear accumulated Tcf-4 and cAMP response element binding protein binding protein, and was significantly acetylated, leading to increases in DNA binding. Priming of -catenin at Ser45 exists in vivo. However, -cat45 does not necessarily enter the degradation pathway. Impairment in linking -cat45 to subsequent GSK-3-mediated phosphorylation and degradation may account for increased steady-state levels of both unphosphorylated as well as Ser45-phosphorylated -catenin, which may be causally linked to increases in cell census during TMCH.
colon; hyperproliferation
-catenin provides a necessary link between cadherin molecules and the 
-catenin-based linkage to the actin cytoskeleton, contributing structural integrity to the complex (11, 27).
In the absence of Wnt signaling, the level of -catenin is kept low through degradation of cytoplasmic -catenin in excess of that bound to cadherins at the plasma membrane. Cytoplasmic -catenin associated with axin and adenomatous polyposis coli protein (APC) undergoes sequential phosphorylation, first at Ser45 (-cat45) by casein kinase I (CKI) and then at Ser33,37/Thr41 by glycogen synthase kinase (GSK)-3, leading to targeted ubiquitination through E3 ubiquitin ligase containing the F-box protein -transducing repeat-containing protein (TrCP) and proteasomal degradation (1, 3, 12, 16, 29). Activation of Wnt signaling leads to inhibition of GSK-3 activity, resulting in accumulation of cytoplasmic -catenin, which becomes available to bind the T cell factor (Tcf)/lymphoid enhancer factor (Tcf/LEF) family of transcription factors and to induce target gene expression (3). Appropriate defects in APC, axin, or -catenin mimic Wnt signaling and interfere with the targeted destruction of -catenin by the proteasome, thereby prolonging its signaling capacity.
Recent studies have also implicated CKI as a positive regulator of -catenin signaling (9). CKI functions downstream of dishevelled (Dvl) and upstream of GSK-3, interacts with Dvl, and coimmunoprecipitates with axin, GSK-3, and Dvl-3 (28, 35). CKI-
and CKI- are closely related, because their kinase domains are 98% identical and each contains a 53% identical COOH-terminal tail that inhibits CKI-/- when autophosphorylated. CKI- lacking the COOH terminal has been shown to 1) block induction of a secondary axis in Xenopus embryos and 2) lack coimmunoprecipitation with axin, suggesting that this domain is important for Wnt signaling (32). CKI- phosphorylates several components of the -catenin degradation complex in vitro. In addition, CKI- phosphorylates axin and -catenin on major in vivo phosphorylation sites (9). Indeed, phosphorylation of -catenin at Ser45 by CKI- has been shown to increase -catenin's affinity for the Tcf/LEF family of transcription factors, suggesting that CKI- may stabilize -catenin and propagate the Wnt signal (9). However, the mode of action of CKI- is not understood. Even less is known about CKI-'s physiological role in vivo.
Similarly, the Wnt-induced mechanism preventing GSK-3-mediated phosphorylation of -catenin remains elusive. GSK-3 appears to be insulated from regulators of GSK-3 that lie outside of the Wnt pathway. Insulin signaling, for example, leads to inhibition of GSK-3 via Ser9/Ser21 phosphorylation but does not cause accumulation of -catenin. Conversely, Wnt signaling does not affect insulin signaling (7, 42). Thus it appears that the GSK-3 involved in -catenin phosphorylation is regulated in a unique manner.
We have used an animal model that involves the earliest molecular and functional changes that place the colon at risk for subsequent development of carcinoma. Transmissible murine colonic hyperplasia (TMCH) caused by Citrobacter rodentium is a naturally occurring disease of laboratory mice characterized by epithelial cell hyperproliferation in the distal colon (2). We showed previously (33, 37) that TMCH exhibits both functional and molecular changes typical of the earliest stages of neoplastic transformation. Before the hyperplasia, cellular -catenin increases in abundance and, subsequently, undergoes nuclear translocation. Downstream targets of -catenin, including cyclin D1 and c-myc, increase in abundance (33). These changes in -catenin signaling are critical in maintaining an elevated cell census during TMCH, since a block in -catenin's cellular abundance leads to abrogation of hyperproliferation/hyperplasia (36).
Using the TMCH model, we investigated CKI-'s expression and activity in vivo and determined its intracellular interaction with -catenin. We also determined kinetics and modulation of Ser/Thr phosphorylation and analyzed biological function of -catenin in vivo during crypt hyperproliferation. Our findings clearly demonstrate constitutive phosphorylation of -catenin at Ser45 in vivo. However, -cat33,37,41 did not increase in abundance. Activity assays confirmed GSK-3 inhibition in hyperproliferating epithelia. Inhibition of GSK-3 activity during TMCH therefore appears to mimic Wnt signaling. However, accumulation of both unphosphorylated -catenin as well as -cat45, which coimmunoprecipitates with Tcf-4 and cAMP response element binding protein binding protein (CBP), is novel and may be critical in regulating proliferatory activity of the colonic epithelium during TMCH.
|
EXPERIMENTAL PROCEDURES |
|---|
|
TOPABSTRACTEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
Treatment of animals with bortezomib (Velcade). Velcade was purchased from Millennium Pharmaceuticals (Cambridge, MA). Both normal and infected mice received Velcade intraperitoneally (1 mg/kg body wt). The treatment groups were C. rodentium-infected mice receiving vehicle alone (0.9% sodium chloride); mice receiving Velcade injection twice a week for 2 wk before and during C. rodentium infection; and normal mice receiving Velcade injection. Animals in each group tolerated the drug well, and weight gain was similar in each group (data not shown). Animals in all groups were euthanized 4 h after the last injection. Their colons were used to isolate crypts for biochemical studies as described elsewhere (37).
Isolation of crypts. Distal colons were attached to a paddle and immersed in Ca2+-free standard Krebs-buffered saline (in mM: 107 NaCl, 4.5 KCl, 0.2 NaH2PO4, 1.8 Na2HPO4, 10 glucose, and 10 EDTA) at 37°C for 10–20 min, gassed with 5% CO2-95% O2. Individual crypt units were then separated from the submucosa/musculature by intermittent (30 s) vibration into ice-cold potassium gluconate-HEPES saline (in mM: 100 potassium gluconate, 20 NaCl, 1.25 CaCl2, 1 MgCl2, 10 HEPES, 10 glucose, and 5 sodium pyruvate) and 0.1% BSA. Crypts were then concentrated by centrifugation and processed for biochemical analyses.
Subcellular fractionation and protein estimation. Crude cellular homogenates were prepared from either isolated crypts or whole distal colons of normal and C. rodentium-infected mice by homogenization in buffer [in mM: 50 Tris·HCl, 250 sucrose, 2 EDTA, 1 EGTA (pH 7.5), and 10 2-mercaptoethanol with 0.5% Triton X-100, plus protease and phosphatase inhibitors and 10 mM N-ethylmaleimide for ubiquitination studies]. After a low-speed spin (15,000 g for 15 min), the clear supernatant was saved as total solubilized protein cell extract. Nuclear extracts were prepared from freshly isolated crypts essentially as described by us (33, 39) and elsewhere (43). Protein concentrations were determined, and extracts were frozen in liquid nitrogen and stored at –70°C.
Proteasomal activity assay. The chymotrypsin-like activity of the proteasome was measured in crypt cellular homogenate prepared from the distal colons of normal, day 6, day 6+Velcade, day 12, and day 12+Velcade-treated mice. Reaction mixtures contained 10 µg of protein, 50 mM Tris·HCl (pH 8), 1 mM DTT, and 40 µM succinyl-leucine-leucine-valine-tyrosine-amino-4-methyl-coumarin (AMC). The mixture was incubated for 30 min at 37°C and then stopped by adding 100 µM monochloroacetate and 30 mM sodium acetate (pH 4.3). Fluorescence was determined by measuring the release of AMC (excitation 370 nm, emission 430 nm) with an FLx800 Multi-Detection Microplate spectrofluorometer from Bio-Tek Instruments (Winooski, VT). The concentration of liberated products was calculated with a standard curve for AMC.
Immunoprecipitation and kinase assays. For immunoprecipitation (IP) studies, crypt cytosolic or nuclear extracts were normalized for protein concentration and precleared for 1 h at 4°C with 30 µl of protein A-coated Sepharose beads. IP was carried out at 4°C by incubating the fractions for 2 h with appropriate antibodies and then for 2 h with 50 µl of protein A/G-Sepharose beads. Control experiments were performed by carrying out the IPs in the presence of the immunizing peptides, or with control IgG antisera. The immunoprecipitated proteins were recovered by boiling the Sepharose beads in 2x SDS sample buffer.
Activity assays for both CKI- and GSK-3 were carried out after IP of isolated crypt extracts from the distal colon of normal mice and mice 6 and 12 days after infection. IPs were performed at 4°C by incubating the fractions for 2 h with polyclonal antibodies against CKI- and GSK-3, respectively, and then for 1 h with 50 µl of protein A/G-Sepharose beads. Immunoprecipitates were washed twice with lysis buffer, twice with wash buffer [10 mM Tris·HCl (pH 7.4), 100 mM NaCl, 1 mM EDTA, 0.2 mM sodium vanadate, and 1 µM microcystin LR], and twice with kinase reaction buffer [in mM: 20 HEPES (pH 7.4), 10 MgCl2, 1 DTT, and 0.2 EGTA]. Kinase reactions were performed in a total volume of 40 µl of kinase buffer containing 3.75 µg of appropriate substrates: CKI, RRKDLHDDEEDEAMSITA (amino acid substitutions are underlined; this was done to make the peptide refractory to CKII); GSK-3, YRRAAVPPSPSLSRHSSP-HQ (pS) EDEEE; 15 µM of cold ATP; and 10 µCi of [32P]ATP. After 20 min of incubation at 30°C, 20 µl of SDS lysis buffer was added and incubated at 70°C to stop the reaction. Reaction mixtures were centrifuged, and 15 µl of the supernatant was spotted onto Whatman P81 phosphocellulose paper. Filters were washed in three changes of 0.75% phosphoric acid, rinsed in acetone, and dried, and 32P incorporation was measured in a liquid scintillation counter. During GSK-3 assay, unphosphorylated glycogen synthase peptide was used as a negative control. Nonspecific 32P incorporation was subtracted from values obtained with the phospho-glycogen synthase peptide.
Western blotting.
Total crypt cellular extracts, subcellular fractions (30–100 µg protein/lane), or immunoprecipitated proteins were subjected to SDS-PAGE and electrotransferred to nitrocellulose membrane. The efficiency of electrotransfer was checked by back staining gels with Coomassie blue and/or by reversible staining of the electrotransferred protein directly on the nitrocellulose membrane with Ponceau S solution. No variability in transfer was noted. Destained membranes were blocked with 5% nonfat dried milk in Tris-buffered saline [TBS; 20 mM Tris·HCl and 137 mM NaCl (pH 7.5)] for 1 h at room temperature (21°C) and then overnight at 4°C. Immunoantigenicity was detected by incubating the membranes for 1–2 h or overnight with the appropriate primary antibodies (0.5–1.0 µg/ml in TBS containing 0.1% Tween 20; Sigma). These antibodies were polyclonal anti--catenin, p-catenin-Thr41/Ser45, p-catenin-Ser45, p-catenin-Ser33,37/Thr41, pGSK-3-Ser9, acetylated lysine, and ubiquitin (Cell Signaling Technology, Beverly, MA); polyclonal anti -catenin, CBP, and lamin B (Santa Cruz Biotechnology, Santa Cruz, CA); monoclonal anti -catenin (dephospho-8E4 clone, EMD Biosciences, San Diego, CA); polyclonal anti-axin, Tcf-4, and PP2A (Upstate Biotechnology, Charlottesville, VA); and monoclonal anti-actin (Chemicon International, Temecula, CA). After washing, membranes were incubated with horseradish peroxidase-conjugated anti-mouse or anti-rabbit secondary antibodies and developed with the ECL detection system (Amersham, Arlington Heights, IL) according to the manufacturer's instructions.
Electrophoretic mobility shift assay.
Crypt nuclear extracts were prepared from normal or Citrobacter-infected mouse distal colon essentially as described previously (33, 38). Ten micrograms of nuclear extract in 10 µl of buffer was mixed with 2 µg of poly(dI-dC) and 1 µg of BSA to a final volume of 19 µl. After 15-min incubation on ice, 1 µl of [
-32P]ATP end-labeled double-stranded Tcf-4 consensus oligonucleotide (5'-AGCTGGTAAGATCAAAGGG-3') was added to each reaction and incubated at room temperature for an additional 15 min. The reaction products were separated on a 4% native polyacrylamide-0.5% Tris-borate-EDTA gel and analyzed by autoradiography. Supershift antibodies (1 µl) were included in the binding reaction as indicated.
|
RESULTS |
|---|
|
TOPABSTRACTEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
expression and activity increase during TMCH.
Since CKI- has been implicated in priming -catenin for subsequent phosphorylation by GSK-3, we sought to determine whether the expression levels and activation status of CKI- changed during TMCH.
Using isoform specific antibody for CKI-, we measured the expression level of this kinase in Triton X-100-solubilized crypt cellular extracts prepared either at day 6 or at day 12 of TMCH. CKI- exhibited 3-fold and 3.3-fold increase in abundance at these time points (Fig. 1A). We next determined whether this could be attributed to its relocalization from either cytoskeletal or nuclear compartments. CKI- was detected in both cytoskeletal and nuclear fractions(Fig. 1B, i and ii, respectively). Neither fraction, however, exhibited any decrease in CKI- levels, suggesting that increases in cellular CKI- levels may represent increased synthesis of this enzyme during TMCH.
|
kinase assay was performed in protein A/G-agarose-recovered immune complexes. In conjunction with elevated cellular levels, CKI- activity also increased 2-fold and 1.8-fold at day 6 and day 12 of TMCH, respectively (Fig. 1C). Thus crypt hyperproliferation was associated with both increased cellular and nuclear CKI- protein level as well as cellular activity.
-Catenin interacts with CKI- in vivo and undergoes selective Ser/Thr phosphorylation during TMCH.
Coimmunoprecipitation (co-IP) studies were performed to determine the relationship between CKI- and -catenin in native epithelia. IPs were carried out with CKI- in Triton X-100-solubilized total cellular as well as nuclear extracts prepared from normal, day 6, and day 12 TMCH crypts. Immunoprecipitated proteins were then subjected to Western blotting with anti -catenin antibody. -Catenin associated with cellular CKI- at both days 6 and 12 with equal measure (Fig. 2Ai). To confirm the presence of CKI- in the immune complex, the membrane was stripped and reprobed with anti-CKI-. CKI- was successfully immunoprecipitated with anti CKI- at both day 6 and day 12 of TMCH (Fig. 2Aii). Significant nuclear translocation of CKI- along with -catenin was also observed (Fig. 2B, i and ii) as described in Fig. 1Bii. -Catenin associated with nuclear CKI- at both time points with equal measure (Fig. 2Biii).
|
has been shown to phosphorylate -catenin in the axin complex in cell lines, we next investigated CKI--axin interaction in vivo. Cellular levels of axin itself did not change during TMCH (Fig. 2Ca). Co-IP studies revealed extremely low interaction between CKI- and axin (data not shown), suggesting that CKI--axin interaction in vivo may be transient and undetectable by co-IP. At the same time, however, co-IP studies revealed association of -catenin (Fig. 2Cb) with axin both in normal crypts as well as during TMCH. We previously reported (40) association of -catenin with APC at both time points. These studies suggest that the scaffolding provided by axin and APC is not impaired in vivo.
We have observed changes in -catenin cellular expression and signaling without alterations in -catenin mRNA levels during TMCH (Ref. 33 and unpublished observations). In this study, we examined the kinetics and modulation of Ser/Thr phosphorylation of -catenin in vivo to delineate the molecular basis of accumulation of cellular/nuclear -catenin during TMCH (33). As a first step, we investigated the effect of these CKI---catenin interactions on phosphorylation of cellular -catenin.
We observed significant increases in Ser45 phosphorylation of -catenin (3.8 ± 0.8-fold and 3.0 ± 0.8-fold, respectively) in the Triton X-100-solubilized cellular fraction at days 6 and 12 compared with uninfected control (Fig. 3Ai). When the increases in Ser45 phosphorylation were normalized to total -catenin levels, only subtle enhancement in Ser45 phosphorylation was observed, suggesting that increased phosphorylation of -catenin at Ser45 observed at days 6 and 12 may be due to a proportional increase in overall -catenin abundance. Nonetheless, priming of -catenin at Ser45 occurred during TMCH. The primary GSK-3 phosphorylation site in -catenin (Ser33,37/Thr41 residues), however, failed to show changes in phosphorylation status (Fig. 3Aii). The lack of significant change in phosphorylation on these residues was not due to mutation in exon 3, which codes for these Ser/Thr residues in -catenin (data not shown).
|
-catenin also exists in Triton X-100-insoluble (cytoskeletal) and nuclear pools, we next determined distribution of both phosphorylated and total -catenin in these subcellular compartments. Similar to the soluble fraction, significant accumulation of -cat45 was observed in both cytoskeletal (Fig. 3Bai) and nuclear (Fig. 3Bbi) fractions. Changes in -cat33,37/41 were only subtle, however (Fig. 3B, aii and bii). These studies suggest that increases in cellular -catenin/-cat45 may not be attributable to their relocalization from either cytoskeletal or nuclear compartments. Indeed, corroborating our earlier findings (33), levels of E-cadherin did not change at either time point (Fig. 3C).
After phosphorylation at Ser33,37 residues, -catenin is ubiquitinated and rapidly degraded by the ubiquitin-proteasome machinery. To rule out the possibility that the failure to detect increased levels of -cat33,37 may be due to its rapid degradation through the ubiquitin proteasomal pathway, the proteasomal activity was blocked in vivo with Velcade (see EXPERIMENTAL PROCEDURES). Cellular extracts were used for measuring proteasomal activity and for Western blotting.
Citrobacter-infected mice injected intraperitoneally with Velcade exhibited significant decrease in chymotrypsin-like activity both at day 6 and more so at day 12, confirming proteasomal inhibition in vivo (Fig. 4A). The proteasomal inhibition was also assessed by measuring the level of polyubiquitinated protein after Velcade treatment. Figure 4B shows that Velcade provoked a marked increase in accumulation of polyubiquitinated proteins. When crypt cellular extracts were probed with phospho-specific antibodies, proteasomal blockade in vivo failed to facilitate significant accumulation of -cat33,37/41, while -cat45 accumulation was unaffected (Fig. 4Cii,i). These results are consistent with a defect in the upstream signaling cascade affecting Ser33,37/Thr41 phosphorylation of -cat45 rather than rapid degradation of -cat33,37/41.
|
undergoes phosphorylation at Ser9 that leads to its inactivation during TMCH.
To explore the possibility that reduced levels of Ser33,37 phosphorylation during TMCH may be due to lack of GSK-3 action on -cat45, we next determined the phosphorylation status and activity of GSK-3 in purified crypt cellular extracts. With antibody specific for Ser9 of GSK-3, phosphorylation of GSK-3 increased significantly without changes in total GSK-3 at both day 6 and day 12 of TMCH (Fig. 5Ai). When the same crypt extracts were evaluated for the presence or absence of phosphorylated -catenin, -cat45 accumulated (Fig. 5Aii) as before, while no detectable -cat33,37/41 was observed (Fig. 5Aii), again suggesting a lack of GSK-3 action on -cat45. Because phosphorylation of GSK-3 has been associated with its inactivation (6, 8), we next determined the specific activity of GSK-3 during TMCH. Consistent with the Western blotting data, the immunoprecipitated GSK-3 exhibited 40% and 70% decreases in its activity at days 6 and 12 of TMCH, respectively (Fig. 5B).
|
along with changes in activity during TMCH (39). Since PKC- has been shown to phosphorylate and inactivate GSK-3 in cell lines, we next determined potential interactions between these two proteins in vivo. Co-IP with anti-PKC- and immunoblotting with anti GSK-3 revealed significant increases in association of GSK-3 with PKC- at both day 6 and day 12 of TMCH, compared with control (Fig. 5Ci). The kinetics of GSK-3-PKC- interaction correlated with increases in GSK-3 phosphorylation as shown in Fig. 5A. To confirm the presence of PKC- in the immune complex, the membrane was stripped and reprobed with anti-PKC-. PKC- was successfully immunoprecipitated with anti-PKC- at both day 6 and day 12 of TMCH (Fig. 5Cii). While it remains to be established whether GSK-3 is a direct target of PKC- action in vivo, our results suggest that PKC- may potentially be involved in modulating GSK-3 phosphorylation during TMCH. Thus GSK-3 inhibition in vivo may lead to a unique imbalance of phosphorylated -catenin species by impairing stepwise phosphorylation from Ser45 to Ser33,37/Thr41 positions, leading to stabilization of cytosolic -cat45.
Does nuclear accumulated -cat45 have a biological role?
Paralleling our earlier studies, significant nuclear accumulation of -cat45 was observed during TMCH, while -cat33,37 accumulation was minimal (Fig. 6A). The same membrane, when stripped and reprobed with anti -catenin, showed accumulation of unphosphorylated -catenin in the nuclei (Fig. 6A). Similar results were obtained when Velcade-treated nuclear extracts were probed with these phospho-antibodies (data not shown). Interestingly, the same crypt nuclear extracts, when probed for Tcf-4, also exhibited significantly elevated levels at both time points (Fig. 6B).
|
-cat45 along with unphosphorylated -catenin has a functional role, we investigated interaction of both species of -catenin with Tcf-4 as well as interaction of -cat45 with the transcriptional coactivator CBP. Co-IP studies with both anti-dephospho--catenin (8E4 clone) and anti -cat45 followed by blotting with anti-Tcf-4 revealed strong association of Tcf-4 with both species of -catenin both at day 6 and more so at day 12 of TMCH (Fig. 6C, 1i and 2i). This may be related to apparent increase in Tcf-4 levels at day 12 (Fig. 6B). Figure 6C1ii and 2ii represents successful IP of these species at both time points in the nuclear extracts. Co-IP also revealed significant association of -cat45 with CBP at both time points (Fig. 6Di), suggesting that phosphorylation of -catenin at Ser45 may not be antagonistic to its transcriptional coactivator function. When the same membrane was stripped and reprobed with anti-acetylated lysine, significant acetylation of -cat45 was observed, particularly at day 12 TMCH (Fig. 6Dii). Thus CBP-mediated alteration in -cat45 may lead to presence of a transcriptionally competent protein in hyperproliferating colonic epithelia. These results suggest that, along with unphosphorylated -catenin, nuclear -cat45 may be equally important for its pro-proliferating role during TMCH.
In melanoma cells, -cat33,37 translocates to the nucleus and interacts with LEF-1 but fails to form a tertiary complex with DNA (30). To investigate whether -cat45-Tcf-4 complex binds DNA, gel shift assay was performed with a radiolabeled probe corresponding to the consensus DNA sequence of the Tcf-4 binding site in normal and day 12 TMCH crypt nuclear extracts. As shown in Fig. 7, Tcf-4 binding and Tcf-4--catenin binding with DNA increased significantly at day 12. Supershift with antibodies against Tcf-4 and -catenin (8E4 clone) confirmed these bindings. We also observed supershift with anti--cat45, and the extent of supershift was comparable with that observed with unphosphorylated -catenin antibody (Fig. 7). These findings suggest that -cat45, like unphosphorylated -catenin, accumulates in the nucleus, forms complexes with CBP and Tcf-4, and may therefore modulate transcriptional coactivation of target genes in the native epithelia.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
-catenin abundance with nuclear translocation and upregulation of downstream targets including c-myc and cyclin D1. These increases in -catenin abundance are not due to mutational inactivation in -catenin or dysfunction related to APC (40). We hypothesized that the increase in -catenin may be more likely related to an epigenetic alteration in the degradation pathway. The present study analyzed the kinetics and modulation of Ser/Thr phosphorylation and the biological function of -catenin in vivo during crypt hyperproliferation.
Using the TMCH model, we show that hyperproliferation was associated with 1) increases in CKI- abundance (3-fold and 3.3-fold) and nuclear translocation and activity (2-fold and 1.8-fold) at days 6 and 12 of TMCH; 2) significant -catenin-CKI- interaction in the cellular and nuclear compartments as well as cellular -catenin-axin interaction at both time points; 3) constitutive phosphorylation of cellular -catenin at Ser45 and compartmentalization into cytoskeletal and nuclear fractions, respectively; 4) unaltered phosphorylation status of Ser33,37/Thr41 residues even with inhibitor of proteasomal degradation; 5) significant increase in phosphorylation of GSK-3 at Ser9, leading to 40% and 70% decreases in its activity at these time points; and 6) increased interaction of nuclear accumulated -catenin with Tcf-4 and CBP, leading to significant increase in DNA binding.
CKI- and the closely related CKI- are positive regulators of the Wnt signaling pathway, functioning downstream of Wnt and upstream of the APC-axin-GSK-3-PP2A complex. The data presented here show that both abundance (cellular and nuclear) and activity of CKI- increase significantly during TMCH. These increases in CKI- were apparently not related to its redistribution from either cytoskeletal or nuclear compartments (see Fig. 1). These results suggest that other relevant posttranscriptional/translational mechanisms may be involved in modulating CKI- levels in vivo.
CKI- has also been shown to bind Dvl-1, albeit less strongly than CKI-, suggesting that the differential interaction of CKI- and - with Dvl-1 may be due to their divergent COOH termini (31, 26). We did not measure CKI- expression or activity and thus cannot rule out the possibility that CKI- may have contributed toward increases in CKI- activity during TMCH. Nonetheless, the observed changes in CKI- abundance and activity are, to our knowledge, the first report of such changes in a native, nonmalignant colonic epithelium.
Recent studies in vitro have shown that overexpression of CKI- mimics Wnt signaling by stabilizing -catenin, thereby increasing expression of -catenin-dependent genes (26). In fact, inhibition of endogenous CKI- either by the kinase-defective form of CKI- or antisense oligonucleotide attenuated gene transcription stimulated by Wnt (26). CKI- also responds to Wnt signaling directly by undergoing transient dephosphorylation of critical inhibitory sites present in its COOH-terminal domain (19, 20). Indeed, mutation of CKI- inhibitory autophosphorylation sites creates a kinase that is significantly more active than CKI- and is not activated further on Wnt stimulation. Given that the aberrant mucosal cytokinetics and inhibition of GSK-3 observed during TMCH mimic Wnt signaling, it is tempting to speculate that the cytoplasmic component of the Wnt pathway may be modulating CKI- expression and activity in vivo. Since alterations in CKI- observed in the present study could be the molecular basis for TMCH, efforts are under way to block Wnt signaling in vivo to see whether changes in -catenin phosphorylation/stability, mediated probably by CKI-, are abrogated. Taken together, our results are consistent with studies in cell lines and during Xenopus embryogenesis (4, 35, 28), suggesting that CKI- activation in vivo may be a biologically relevant mechanism to modulate -catenin stability in hyperproliferating epithelia.
We have observed significant increases in association of CKI- with -catenin during TMCH. We were, however, unable to demonstrate a direct interaction between CKI- and axin during TMCH (data not shown), which may be attributed to transient interaction between these proteins leading to weak signal. In contrast, we did observe significant interaction of axin with -catenin at both time points (see Fig. 2). We have also reported recently (40) that -catenin-APC interactions exist in vivo during TMCH. These studies suggest that scaffolding provided by axin and APC proteins is not compromised during epithelial hyperproliferation.
Since scaffolding proteins facilitate phosphorylation of -catenin by both CKI- and GSK-3, we next investigated phosphorylation status of -catenin during TMCH. Our results clearly show that Ser45 phosphorylation of -catenin occurs in vivo. This does not necessarily mean that it is mediated by CKI-, since a causal relationship between increases in CKI- and concomitant increase in -cat45 remains to be established. Nevertheless, accumulation of Ser45-phosphorylated -catenin in hyperproliferating colonic epithelium clearly suggests that priming of -catenin, possibly by CKI-, exists in vivo for subsequent phosphorylation by GSK-3 and degradation via proteasomal pathway. Yet, despite the priming, GSK-3-mediated phosphorylation at Ser33,37/Thr41 was minimal. The lack of any change in phosphorylation at these residues was not due to mutational defect. Rapid degradation of -cat33,37/41 was considered unlikely because of the lack of a demonstrable effect of Velcade. These studies suggested that other epigenetic signaling events in addition to CKI- may be altered during TMCH. Indeed, we observed dramatic increase in phosphorylation of GSK-3 at Ser9 concomitant with significant decreases in its activity (see Fig. 6). Thus, despite the existence of Ser45-phosphorylated -catenin, -cat45 was only minimally phosphorylated at Ser33,37/Thr41 due to GSK-3 inactivation. Moreover, we have shown recently (40) that cellular abundance and interaction of wild-type APC with -catenin increase significantly at day 6 of TMCH. -Catenin, however, does not degrade as a result of this association, again suggesting the existence of a deranged pathway for degradation of -cat45 probably due to GSK-3 inactivation.
Given that GSK-3 activity, particularly at day 6, drops only to 50% of the normal level, it is indeed intriguing why accumulation of -cat33,37/41 is not observed at this time point. It is, however, not surprising since the same level of inhibition for GSK-3/ is sufficient to allow glycogen synthesis to proceed in response to insulin signaling (5, 34). Interestingly, however, while both insulin and Wnt signaling cause inhibition of GSK-3, only Wnt signaling stabilizes -catenin (7). But unlike insulin signaling, which involves phosphorylation of both GSK-3 and -, it is unclear whether GSK-3 is regulated by phosphorylation during Wnt signaling.
A number of kinases including, but not limited to, protein kinase B/AKT, mitogen-activated protein kinases, and PKC- may be involved in inhibiting GSK-3 activity (23, 10, 15, 22). Indeed, significant interaction between GSK-3 and PKC- existed at both day 6 and day 12 of TMCH, which may have facilitated GSK-3 phosphorylation and inactivation in vivo (see Fig. 6). Thus PKC- potentially may be involved in modulating temporal inhibition of GSK-3, thereby affecting cellular levels of -catenin during TMCH.
Nuclear accumulation of both phosphorylated and unphosphorylated -catenin.
-Catenin participates in transcription and is regulated through multiple phosphorylations that control its access to the cell nucleus. The rapid proteolysis of cytosolic -catenin is such that constitutive nuclear localization of -catenin is a marker for activation of -catenin signaling and can be detected in tumors containing APC, axin, and -catenin mutations (27). Our results showed that GSK-3 inhibition resulted in stabilization of cellular -catenin by failing to efficiently couple -cat45 to Ser33,37/Thr41 residues. This led to significant nuclear accumulation of -cat45 along with unphosphorylated -catenin during TMCH. -cat45 did not simply accumulate in the nuclei but, along with unphosphorylated -catenin, exhibited significant interaction with Tcf-4 both at day 6 and more so at day 12 of TMCH. In a separate experiment, the same nuclear extracts when immunoprecipitated with anti -cat45 and probed for 1) CBP and 2) acetylated lysine exhibited significant increases in -cat45-CBP interaction, leading to acetylation of -cat45 at both time points. Phosphorylation of -catenin at Ser45 is not an absolute prerequisite for -catenin's biological activity, as several human cancers harbor mutation at this site yet exhibit constitutively active Wnt signaling (27). However, our findings suggest that nuclear -cat45, along with its unphosphorylated counterpart, may be involved in transcriptional coactivation in vivo. CBP is an acetyl transferase that promotes the activity of several transcriptional factors by their acetylation. CBP has also been shown to acetylate -catenin in cell lines (13, 17, 41). Thus, while mechanistic investigations such as transcriptional coactivation assays remain to be carried out, our findings clearly suggest that nuclear accumulation of both species of -catenin may be required for full control of its transcriptional coactivator function in vivo.
The nuclear accumulation of Ser45-phosphorylated -catenin is not unprecedented. In Chinese hamster ovary cells, cotransfection of -catenin with LEF-1 led to extensive nuclear localization of -cat45 but not -cat33,37/Thr41. Similarly, -cat45 selectively accumulated in the nucleus of presenilin 1 (PS1)-deficient epidermal tumors and in medulloblastomas with activating -catenin mutations (14). Notably, -cat45 nuclear reactivity was elevated not only in the neoplastic tissue but often also in regions that were hyperplastic or histologically normal in PS1-deficient skin (14). Similarly, mouse embryonic stem cells in which both GSK-3 and - isoforms have been "knocked out" exhibit massive increases in cytoplasmic and nuclear -catenin that remains phosphorylated on Ser45 regardless of treatment with Wnt-3a and is unphosphorylated on residues 33, 37, and 41 as expected (24). We also observed nuclear translocation of -cat45 at both time points. Thus our results both support earlier findings, albeit in a different model, and, more importantly, extend the current canonical model of -catenin signaling in the colonic epithelium. Indeed, long-term kinetics clearly indicates that reduction in levels of both species of -catenin is critical to completely abolish hyperplasia (manuscript in preparation).
The temporal inhibition of GSK-3 along with CKI- overexpression and nuclear translocation highlights the significance of epigenetic modifications of Wnt/-catenin signaling affecting proliferation in vivo. The dual alterations of critical modulators of -catenin (i.e., inhibition of GSK-3 coupled with nuclear accumulation of both unphosphorylated -catenin and -cat45) not only may be vital in maintaining an elevated cell census during TMCH but may eventually promote mucosal priming for subsequent neoplasia (18, 21, 25).
|
GRANTS |
|---|
|
TOPABSTRACTEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
|
REFERENCES |
|---|
|
TOPABSTRACTEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
/
) is a convergent downstream target of the insulin-stimulated phosphatidylinositol 3-kinase and TC10 signaling pathways. J Cell Biol 64, 279–290, 2004.
