![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Previous Article | Next Article 
Molecular and Cellular Biology, November 2007, p. 7802-7815, Vol. 27, No. 22
0270-7306/07/$08.00+0 doi:10.1128/MCB.02179-06
Copyright © 2007, American Society for Microbiology. All Rights Reserved.
Jennifer J. Kordich,
Jason R. Spence,
Robert Opoka,
Scott Rankin,
Suh-Chin J. Lin,
Diva Jonatan,
Aaron M. Zorn,* and
James M. Wells*
Division of Developmental Biology, Cincinnati Children's Hospital Research Foundation, 3333 Burnet Avenue, Cincinnati, Ohio 45229-3039
Received 21 November 2006/ Returned for modification 22 December 2006/ Accepted 6 September 2007
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In the absence of a Wnt signal, levels of cytosolic ß-catenin are kept low via the interaction of ß-catenin with a protein complex including glycogen synthase kinase 3ß (GSK3ß), adenomatous polyposis coli (APC), and Axin. The phosphorylation of ß-catenin by the kinase GSK3ß allows ß-catenin to be ubiquitinated and targeted for degradation by the proteasome (1). The binding of a canonical Wnt ligand to the frizzled-lipoprotein receptor-related protein 5/6 receptor complex results in the repression of GSK3ß and the stabilization of ß-catenin. Stabilized ß-catenin accumulates in the nucleus, where it acts as a cofactor with the HMG box family of TCF/LEF transcription factors to regulate the expression of Wnt target genes, such as cyclin D1 and Cdx-1 (17, 22). Although the formation of a TCF-ß-catenin complex is required for the activation of all Wnt target genes (36), Wnt signaling is involved in a wide array of biological processes, including cell proliferation, cellular transformation (14), and embryonic development (24), demonstrating that the output of this pathway is highly influenced by the cellular context.
Given that aberrant activation of the canonical Wnt pathway can lead to unrestricted cell division and tumor formation (12, 26, 28, 31, 40), it is not surprising that this pathway is antagonized by several different mechanisms. For example, several extracellular antagonists that inhibit ligand-receptor interactions have been described previously, including Dickkopf (Dkk), Cerberus, and the secreted frizzled-related proteins (10, 21, 34, 35). In many instances, Wnt signaling is kept in check by a negative-feedback loop in which ß-catenin/TCF activity induces the transcription of its own negative regulators, Axin and Dkk1 (4, 20, 39). Finally, in the absence of activated ß-catenin, TCF/LEF transcription factors keep Wnt target genes off via their interaction with members of the Grouch family of transcriptional repressors (4, 20, 39).
Structurally related to TCF/LEFs, several members of the Sox family of HMG box transcription factors, including Sox17, Sox3, Sox7, and Sox9, have also been implicated in repressing ß-catenin activity by a mechanism that is not well understood (2, 48, 54, 55). In addition to acting as an antagonist, Sox17 cooperates with ß-catenin to activate the transcription of its endoderm target genes in Xenopus laevis (44). These findings suggest that, dependent on the context, Sox proteins can utilize ß-catenin as a cofactor or can antagonize ß-catenin/TCF function. While the mechanism by which Sox proteins antagonize Wnt signaling is unknown, one possibility is that they compete with TCFs for binding to ß-catenin (55).
Here, we report that Sox proteins expressed in normal and neoplastic gut epithelia can modulate canonical Wnt signaling and the proliferation of gastrointestinal tumor cells. While several Sox factors, including Sox17, Sox2, and Sox9, are antagonists of canonical Wnt signaling, others, such as Sox4 and Sox5, promote Wnt signaling activity. Gain- and loss-of-function analyses demonstrate that the Wnt antagonist Sox17 represses colon carcinoma cell proliferation while the agonist Sox4 promotes proliferation. In contrast to a proposed model in which Sox17 protein antagonizes Wnt signaling by competing with TCFs for ß-catenin binding, we found that Sox17 interacts with both TCF/LEF and ß-catenin and that Sox17 and TCF/LEF proteins interact via their respective HMG domains. Binding experiments suggest that Sox17, TCF, and ß-catenin cooperatively interact to form a complex. In contrast, Sox4 can bind to either TCF/LEF or ß-catenin alone but does not appear to cooperatively bind both proteins. Structure-function analyses indicate that Sox17 must bind directly to both ß-catenin and TCF in order to antagonize Wnt signaling and that Sox17 DNA binding activity is not required. Lastly, functional studies show that Sox17 promotes the degradation of TCF/LEF and ß-catenin proteins via a GSK3ß-independent mechanism that can be blocked by proteasome inhibitors. In contrast, Sox4 may function to stabilize ß-catenin protein. Together, these findings suggest that Sox transcription factors act in a novel pathway to modulate the stability of ß-catenin/TCF proteins and regulate the proliferation of colon carcinoma cells. These results have important implications for how Sox proteins regulate the transcriptional output of Wnt signaling in many developmental and pathological processes.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Cell culture, treatments, cell transfection with plasmids and siRNAs, and luciferase assays. COS, 293T, and SW480 cells were obtained from the American Type Culture Collection and were cultured in Dulbecco's modified Eagle's medium (GIBCO) supplemented with 10% fetal bovine serum (HyClone). Cells were plated into 12-well plates 24 h prior to transfection and were transfected with the TOP-flash plasmid (100 to 150 ng), a ß-galactosidase or renilla plasmid (50 ng) as a transfection control, the gene encoding ß-catenin(S37A) (100 ng), and the Sox plasmids listed above (100 to 800 ng) by using Fugene 6 (Roche) or Lipofectamine 2000 (Invitrogen) according to the manufacturer's protocol. The total amount of DNA was kept constant by adding empty-vector DNA. Cell lysates were prepared 24 to 48 h posttransfection, and luciferase activity was measured according to the luciferase assay system (Promega). Transfection efficiency was normalized using either ß-galactosidase activity or renilla luciferase activity from the control plasmid. Experiments were done at least three times, and data from a representative example are shown. For small interfering RNA (siRNA) experiments, double-stranded Sox4 and Sox17 siRNAs and siRNA negative controls (Stealth RNA interference negative control med GC, catalog no. 12935-300, and BLOCK-iT Alexa Fluor red fluorescent control siRNA, catalog no. 147500-00) were obtained from Invitrogen. SW480 cells were transfected using Lipofectamine 2000 (Invitrogen) in serum and antibiotic-free OptiMEM medium with a 1:1:1 mixture of Sox4 or Sox17 siRNAs at doses of 20, 50, and 100 nM. For all experiments, the siRNA control was used at the equivalent concentrations. The cells were harvested at 48 and 72 h after transfection. The transfection efficiency as measured using the BLOCK-iT Alexa Fluor red fluorescent control siRNA was estimated at 80 to 90% after 12 h (data not shown). A combination of three double-stranded Sox17 siRNA sequences was used (only the sense strand sequences are shown): no. 10620312/116668 B07, GCACGGAAUUUGAACAGUAUCUGCA; no. 10620312/116668 B09, UGCACUUCGUGUGCAAGCCUGAGAU; and no. 10620312/116668 B11, CGCGGUAUAUUACUGCAACUAUCCU. A combination of three double-stranded Sox4 siRNA sequences was used (only the sense strand sequences are shown): no. 10620312/116668 C01, GCAAACGCUGGAAGCUGCUCAAAGA; no. 10620312/116668 C03, UCAAAGACAGCGACAAGAUCCCUUU; and no. 10620312/116668 C05, GCGACAAGAUCCCUUUCAUUCGAGA. For protein stability and degradation experiments, leupeptin, lithium chloride, and chloroquine were purchased from Sigma. Epoxomicin was purchased from Calbiochem, dissolved in dimethyl sulfoxide (1,000x stock), and diluted to 1x in culture medium. The final concentrations were as follows: epoxomicin, 0.1 µM; leupeptin, 100 µM; cholorquine, 25 µM; and LiCl, 40 mM.
For SW480 cell proliferation assays, we transfected cells by using a Lipofectamine-based Nucleofector technology kit (Amaxa, Gaithersburg, MD). For colony assays, SW480 cells were transfected with pmax-GFP, pBABE-PURO, and either Sox17-pcDNA6 or Sox4-pcDNA6. Cells were plated at three different dilutions and allowed to recover overnight. Cells were then cultured in three different concentrations of puromycin (0.7, 1.5, and 3 µg/ml), and selection was continued for 5 days, at which time puromycin-resistant, green fluorescent protein (GFP)-positive colonies were counted. Only colonies containing 
10 GFP-positive cells were counted. To monitor the total cell number, cells were trypsinized and GFP-positive cells were counted using a hemocytometer. Three wells of cells were counted for each time point and condition tested.
Immunohistochemistry and microscopy. Paraffin sections (7 µm thick) were collected onto glass slides (Superfrost Plus; Fisher), and the slides were heated overnight at 60°C, dewaxed, and equilibrated in phosphate-buffered saline (PBS). We performed antigen retrieval by heating slides in a steam chamber in 10 mM citrate buffer, pH 6, for 20 min. Anti-Sox4 (Sigma; 1:500) and anti-Sox17 (1:500) (33) immunohistochemistry analysis was performed using tissue sections from wild-type CD1 mice or APCmin/+ mice, which carry a mutation in the APC gene that predisposes them to intestinal adenomas. After deparaffinization, rehydration, and antigen retrieval, tissues were blocked for 1 h using 5% normal goat serum in PBS-0.5% Triton X. Sections were incubated overnight at 4°C with antibodies diluted in blocking reagent. On the second day, slides were washed and incubated with goat anti-rabbit biotinylated secondary antibody (Jackson ImmunoResearch) diluted in blocking reagent (1:500) for 1 h at room temperature. Slides were washed and incubated in ABC reagent (Vector Labs) for 30 min at room temperature. Instead of detecting the antigen with the traditional 3,3'-diaminobenzidine chromogen included in the ABC kit, fluorescence detection of antigens was carried out using the tyramide substrate amplification kit as described by the manufacturer (Molecular Probes). Tyramide-488 was used to produce fluorescent green immunostaining. Sections were counterstained with the red nuclear dye PoPro3 (Molecular Probes). A Leica DM IRB inverted microscope was used to capture digital images of GFP-expressing cells of immunofluorescently labeled sections.
SW480 cells were fixed for 10 min in 4% paraformaldehyde and washed extensively in 1x PBS. They were then blocked using 5% normal goat serum in PBS-0.5% Triton X. Cells were incubated with primary antibodies overnight at 4°C. On the second day, cells were washed and incubated with the appropriate secondary antibodies directly conjugated to fluorophores.
Western blot analysis. Cells were transfected with Sox17 (50 to 400 ng), ß-catenin gene (100 ng), TCF4 (100 ng), and/or Lef1 (100 ng) constructs. Cell lysates were prepared 24 to 48 h posttransfection as described previously (55) and analyzed by standard Western immunoblotting procedures with the following antibodies: anti-V5-horseradish peroxidase (anti-V5-HRP; 1:4,000 [Invitrogen]), rabbit anti-ß-catenin (1:1,000; Santa Cruz Biotechnology), mouse anti-activated ß-catenin (1:1,000; Upstate Biotechnology), mouse antihemagglutinin (anti-HA) antibody (1:1,000; Roche), mouse anti-TCF4 (1:500; Upstate Biotechnology), rabbit anti-Sox4 (1:500; Sigma), mouse antitubulin (1:1,000; Sigma), guinea pig anti-Sox17 (1:1,000), affinity-purified rabbit anti-mouse/human Sox17 (1:5,000), goat anti-rabbit HRP (1:10,000; Jackson ImmunoResearch), goat anti-mouse HRP (1:10,000; Jackson ImmunoResearch), and rabbit anti-XTCF3 (1:3,000) (54).
In vitro protein binding assays. GST-mouse Sox17, GST-TCF4, and GST-ß-catenin were expressed in bacteria and purified using glutathione-coupled agarose resin. His-tagged Xenopus Sox17ß and His-tagged Xenopus Sox17ß consisting of amino acids 1 to 150 (see Fig. 4C) were expressed in bacteria and purified on Ni-agarose resin. ß-catenin, TCF4, and LEF1 genes (in pcDNA6) and the Xenopus TCF3 gene (in pcDNA1) were transcribed with T7 polymerase, and the proteins were produced in TNT translation reactions (Promega). In vitro-translated proteins were incubated with GST or His fusion proteins coupled to agarose resin as described previously (55), and bound proteins were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blotting using anti-V5-HRP or anti-HA-HRP antibody or by 35S labeling autoradiography.
|
For coimmunoprecipitation, equal amounts of nuclear extract were precleared with 1 µg of rabbit preimmune serum and protein A-agarose for 1 h at 4°C. The extracts were then incubated for 4 h at 4°C with protein A-agarose and either 1 µg of a rabbit anti-Sox17 peptide antibody (44) or 1 µg of anti-HA antibody as a negative control. To demonstrate specific binding with anti-Sox17 antibody, 2 µg of a competing Sox17 peptide was included in the incubations. Immunoprecipitates were washed four times in nuclear lysis buffer, resolved by SDS-PAGE, and subjected to anti-ß-catenin immunoblotting.
Xenopus embryo culture and manipulations.
Two-cell-stage Xenopus embryos were microinjected with Sox17
1, Sox172, and Sox17ß antisense morpholino oligonucleotides (MO; 20 ng of each) as previously described (6), 20 ng of a ß-catenin MO (13), or a synthetic mRNA encoding Xenopus Sox17ß (10 pg). At stage 11 (according to a normal table of X. laevis development [29]), embryos were harvested and pooled in groups of three for each condition.
RT-PCR analysis. Total RNA was extracted from embryonic tissue or cells by using the TRIzol reagent (Invitrogen) and treated with RNase-free DNase (Roche). cDNA was synthesized with oligo(dT) primers. Real-time reverse transcription-PCR (RT-PCR) and semiquantitative analysis were performed using the Opticon machine (MJ Research-Bio-Rad). The primers used in this study were as follows: human Sox17 forward, 5'-GAC GAC CAG AGC CAG ACC-3', and human Sox17 reverse, 5'-CGC CTC GCC CTT CAC C-3' (5); human Sox4 forward, 5'-TTC AGC AAC CAG CAT TC-3', and human Sox4 reverse, 5'-TCC CTC TCT CTC GCT CTC TC-3' (37); human cyclin D1 forward, 5'-AAT GAC CCC GCA CGA TT-3', and human cyclin D1 reverse, 5'-GCA CAA GAC GCA ACG AAG G-3' (49); human c-myc forward, 5'-GGT GAA GTA CAT GAA CTC AGG GC-3', and human c-myc reverse, 5'-GGC TTT GAA TCT GCT GGA TTG-3' (43); human ß-tubulin gene forward, 5'-GAT ACC TCA CCG TGG CTG CT-3', and human ß-tubulin gene reverse, 5'-AGA GGA AAG GGG CAG TTG AGT-3'; Xnr3 forward, 5'-AAA TCC ATG TGA GCA CCG TTC-3', and Xnr3 reverse, 5'-GCA TTC TCT GTC TCA TTC TGT G-3'; Hnf1ß forward, 5'-GCA TAT GGC ACA GCA GCC ATT-3', and Hnf1ß reverse, 5'-CAC CAT GCT TGC AAA GGA CAC-3'; Edd forward, 5'-AGA ACG GAG ACT CAG ATC AAT-3', and Edd reverse, 5'-TTA CTC ATT GCT CCC ACC ATC-3'; Plakoglobin forward, 5'-GCT CGC TGT ACA ACC AGC ATT C-3', and Plakoglobin reverse, 5'-GTA GTT CCT CAT GAT CTG AAC C-3'; and ornithine decarboxylase (ODC) gene forward, 5'-GCC ATT GTG AAG ACT CTC TCC AAT C-3', and ODC gene reverse, 5'-TTC GGG TGA TTC CTT GCC AC-3'. SYBR green dye (QIAGEN) was included in the PCR mix. For each experiment and primer pair, a dilution series of cDNA was included in order to generate a standard curve from which the amount of product in each sample was estimated at the log-linear amplification phase. These values were normalized to the level of expression of the ODC gene (53) or the ß-tubulin gene, and the data are presented as a ratio of the expression of each gene to ODC or ß-tubulin gene expression. In each experiment, samples with water alone or controls without reverse transcriptase were included and failed to give specific products. Each experiment was repeated at least three times, and results from a representative example are shown.
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
Sox17 and Sox4 have opposite effects on Wnt activity and the proliferation of colon carcinoma cells. The fact that Sox17 represses Wnt signaling and is down regulated in Wnt-dependent tumors while Sox4 enhances Wnt signaling and is expressed at high levels in APCmin/+ tumors suggests that these transcription factors may play opposite functional roles in gut tumorigenesis. As a rapid way to investigate this idea, we examined the impact of Sox17 or Sox4 overexpression and loss of function on ß-catenin/TCF activity and on the proliferation of SW480 cells. SW480 cells are transformed colon carcinoma cells that contain a mutant APC gene that results in high endogenous ß-catenin/TCF activity as measured by a TOP-flash reporter assay (28, 30). Moreover, immunohistochemical analysis revealed endogenous Sox17 and Sox4 protein in SW480 cells (Fig. 2A). Sox17 protein was restricted to the nucleus, whereas Sox4 was present in both the nucleus and cytoplasm. For gain-of-function analysis, we transfected SW480 cells with increasing levels of Sox17 or Sox4 plasmids. Sox17 caused a dose-dependent decrease in endogenous ß-catenin/TCF activity, whereas Sox4 caused a dose-dependent increase in ß-catenin/TCF activity (Fig. 2B and C). We investigated whether the Sox-mediated regulation of Wnt/ß-catenin/TCF activity correlated with the proliferation of SW480 cells by using a colony formation assay (51). Consistent with its ability to repress Wnt activity, Sox17 expression caused a significant reduction in the number of SW480 cell colonies (Fig. 2D) but had no impact on apoptosis (data not shown). In addition, we observed a significant decrease in the absolute number of transfected SW480 cells that expressed Sox17 over the course of 4 days (data not shown). In contrast, Sox4 had the opposite effect, causing an increase both in colony numbers and in the average number of cells per colony (Fig. 2E). Interestingly, the cells transfected with Sox17 tended to be flatter and more adherent than cells transfected with Sox4, which were often rounder, smaller, and less adherent (Fig. 2D and E insets). These phenotypes are consistent with the possibility that Sox17 inhibits cellular transformation whereas Sox4 promotes it.
|
70%. These data suggest that Sox4 protein may be more stable than Sox17. Sox4 knockdown resulted in a significant reduction in SW480 cell proliferation and an accompanying reduction in the levels of Wnt target genes c-myc and cyclin D1 (Fig. 2K, M, and N), whereas apoptosis was not significantly affected (Fig. 2L). The efficacy and specificity of Sox4 siRNA knockdown were confirmed by the analysis of two Sox4 target genes, Puma and Tle-1 (22a), both of which had significantly reduced levels of expression in cells transfected with Sox4 siRNA (data not shown). Sox17 knockdown caused a modest increase in proliferation and cyclin D1 mRNA levels. The fact that Sox17 does not have more of an effect in colon carcinoma cells is likely due to the presence of other Sox factors, such as Sox9, that are functionally redundant in relation to Sox17. For example, Sox9 is also a Wnt antagonist (Fig. 1) that is expressed in the gut and represses gut cell proliferation (2, 27). Sox17 antagonizes endogenous TCF/ß-catenin activity in embryos. Sox17 was previously shown to antagonize canonical Wnt signaling through an unidentified mechanism when overexpressed in Xenopus embryos (44, 55). To investigate whether endogenous Sox17 may perform a similar function in vivo, we depleted Xenopus embryos of either Sox17 or ß-catenin by the injection of antisense MO and examined the expression of direct TCF/ß-catenin target genes (Fig. 2F and G) (6, 13). As predicted, the depletion of Xenopus embryos of Sox17 resulted in the up regulation of the Wnt target gene Xnr3, whereas the overexpression of Sox17 repressed Xnr3 expression (Fig. 2G) (55). These data further indicate that one normal function of Sox17 is to restrict the activity of ß-catenin/TCF. As a control, we demonstrated that the knockdown of ß-catenin resulted in the loss of expression of both ß-catenin/TCF and ß-catenin/Sox17 target genes (Xnr3 and Hnf1ß, respectively) (Fig. 2F and G) (13, 44). Taken together, these findings suggest that Sox proteins normally function to either positively or negatively modulate Wnt/ß-catenin/TCF activity in a variety of cellular contexts.
Sox proteins interact with ß-catenin and TCF/LEF proteins. Previous studies demonstrated that Xenopus Sox17 binds to ß-catenin in vitro (55), suggesting that Sox17-mediated repression in colon carcinoma cells may involve direct interaction with ß-catenin. To test this possibility, we performed a coimmunoprecipitation assay using a polyclonal antibody that recognizes mouse and human Sox17 and found that ß-catenin can be coimmunoprecipitated with Sox17 from SW480 nuclear extracts (Fig. 3A). This interaction was blocked with a Sox17 peptide used to generate the antibody but not with a control peptide generated from a nonconserved region of Xenopus Sox17ß. This finding confirms that the interaction between Sox17 and ß-catenin is conserved in Xenopus and humans and suggests that this interaction may be involved in the Sox17-mediated repression of Wnt signaling.
|
Sox17 but not Sox4 forms a stabilized protein complex containing both TCF and ß-catenin. To further investigate whether Sox and TCF proteins interact, we performed a series of in vitro binding assays with recombinant Sox17, Sox4, ß-catenin, and TCF/LEF proteins (Fig. 4 and 5). Using Sox17-GST beads, we determined that Sox17 binds directly to TCF4, TCF3, and LEF1 (Fig. 4A), which indicates that TCF/LEF transcription factors physically interact with Sox proteins.
|
To further test the hypothesis that the interaction of all three proteins serves to stabilize a complex, we repeated the in vitro binding experiments with higher salt concentrations at which ß-catenin would not efficiently bind to GST-Sox17 but at which TCF could (Fig. 4D and E). Under these conditions, we found that the ability of Sox17-GST to pull down ß-catenin protein was restored in the presence of TCF4 or TCF3 (Fig. 4D and E). Moreover, we demonstrated that the ß-catenin interaction domain of TCF3 was necessary for the ability of Sox17 to pull down ß-catenin (Fig. 4E).
In summary, rather than Sox17's and TCF's competing for ß-catenin binding, as had been previously proposed, we found that all three proteins interact in a complex. The interaction between TCF/LEF and ß-catenin proteins is stabilized in the presence of Sox17, and this stabilization depends on the ability of Sox17 to interact with ß-catenin. Similarly, Sox17-ß-catenin interactions are stabilized in the presence of TCF/LEF and depend on the ability of Sox17 to interact with TCF/LEF.
The fact that Sox4 has an impact on ß-catenin/TCF activity opposite that of Sox17 prompted us to investigate whether this difference is manifested at the level of ß-catenin/TCF protein-protein interaction. We found that Sox4, like Sox17, could bind directly to TCF4 or ß-catenin (Fig. 5A). Interestingly, in three separate experiments, Sox4 was unable to enhance the binding between TCF4 and ß-catenin, as Sox17 could (Fig. 5B and C). In fact, our data suggested that Sox4, TCF/LEF, and ß-catenin may interfere with binding to one another (Fig. 5C and data not shown). One implication of these data is that the differences in Sox17 and Sox4 binding to ß-catenin and TCF may account for the opposite effects of these proteins on Wnt signaling activity.
Sox17 has distinct TCF and ß-catenin interaction domains.
To further investigate the nature of the Sox17-ß-catenin-TCF protein complex, we performed reciprocal structure-function binding assays to identify the domains that mediate this interaction. We generated a series of deletions and several point mutations in Sox17 and assayed the ability of mutant proteins to bind to TCF4-GST or ß-catenin-GST beads (Fig. 6A). An N-terminal deletion of Sox17 (leaving amino acids 129 to 419 intact) that removed 80% of the HMG domain greatly reduced the ability of the protein to bind to TCF4, whereas the deletion of the C-terminal ß-catenin interaction domain (55) had little effect (Fig. 6A). A point mutation (G to R at amino acid 103) was generated in the HMG domain of Sox17, which based on the crystal structure of the HMG domain of LEF1 (23) was predicted to disrupt the -helical structure of the HMG domain. This mutation abolished the ability of Sox17 to interact with TCF4 but did not reduce ß-catenin binding (Fig. 6A). We generated a second point mutation in the HMG domain (M to A at amino acid 76), which was not predicted to affect HMG domain structure (50) but rather impair DNA binding. As predicted, the Sox17 M-to-A mutant form retained the ability to interact with TCF4 and ß-catenin proteins but was incapable of activating the transcription of a Sox17 reporter gene (pH4X8-Luci) (Fig. 6A and D). These data suggest that the HMG domain of Sox17 mediates the interaction with TCF4 and is distinct from the ß-catenin interaction domain.
|
Sox17 requires both its TCF and ß-catenin interaction domains, but not transcriptional activity, for Wnt-antagonizing activity. We used the mutant forms of Sox17 to investigate whether the interaction between Sox17, TCF/LEF, and ß-catenin was necessary for Wnt-antagonizing activity. We found that Sox17 mutants that did not interact with either TCF/LEF or ß-catenin failed to antagonize Wnt signaling in TOP-flash assays (Fig. 6C). One possibility was that Sox17 regulates Wnt signaling indirectly by activating the transcription of a gene encoding a Wnt antagonist. However, the Sox17 mutant protein with the M-to-A mutation (at amino acid 76), which was predicted to have impaired DNA binding ability and cannot stimulate the transcription of a Sox17 reporter gene (pH4X8-Luci), was still capable of repressing ß-catenin-stimulated TOP-flash activity (Fig. 6C). Importantly, this Sox17 M-to-A mutant protein retained the ability to bind to both TCF/LEF and ß-catenin (Fig. 6A). This finding suggests that Sox17 does not antagonize Wnt signaling via transcription but rather does so by direct interaction with TCF/LEF and ß-catenin.
Sox factors regulate the degradation of ß-catenin and TCF/LEF proteins. While performing transfection experiments with Sox17 and ß-catenin plasmids, we observed reduced ß-catenin protein levels in cells cotransfected with Sox17. We therefore analyzed the effect of Sox17 overexpression on levels of endogenous ß-catenin protein in SW480 cells. We found that the unphosphorylated signaling pool of ß-catenin protein was significantly reduced in the presence of elevated levels of Sox17 protein (Fig. 7A) (the activated-ß-catenin antibody recognizes unphosphorylated ß-catenin). In contrast, ß-catenin protein levels were elevated when Sox4 was overexpressed (Fig. 7B). In RNA interference experiments, we found the opposite; a reduction of Sox4 protein coincided with a significant reduction of endogenous unphosphorylated ß-catenin protein, and Sox17 siRNAs caused a modest increase in ß-catenin protein (Fig. 7C). Similarly, endogenous TCF4 protein levels were slightly elevated by Sox17 siRNAs; however Sox4 siRNAs had no impact on TCF4 levels. These data suggest that Sox factors modulate ß-catenin/TCF activity through regulating protein levels.
|
While it has been previously reported that Sox9 affects ß-catenin protein levels (2), we have demonstrated that Sox17 also interacts with TCF/LEF proteins. Therefore, it is possible that Sox17 may regulate the levels of TCF/LEF proteins. To investigate this possibility, COS cells were cotransfected with TCF4 or LEF1 expression plasmids and increasing levels of a Sox17 plasmid, and extracts were harvested 48 h later (Fig. 7F and G). Interestingly, an increase in Sox17 correlated with a dramatic reduction in TCF4 and LEF1 protein levels in a dose-dependent manner (Fig. 7F and G). To determine if Sox17 needed to interact directly with TCF4 and ß-catenin protein to modulate the stability of these proteins, we used the mutant forms of Sox17 depicted in Fig. 6A. Sox17 mutant proteins that did not interact with TCF4 also failed to promote its degradation, and mutant proteins that did not interact with ß-catenin did not promote its degradation (Fig. 7I). Interestingly, a Sox17 mutant protein (containing only amino acids 1 to 367) that failed to interact with ß-catenin but still interacted with TCF4 was unable to promote TCF4 degradation. The converse was true in that interaction with TCF4 was required for ß-catenin degradation. Together, these findings suggest that Sox17 must directly interact with both ß-catenin and TCFs to trigger the degradation of the complex and to antagonize Wnt signaling.
We next investigated whether Sox4 overexpression affected TCF4 or TCF3 protein levels. The Sox and TCF4 proteins were both V5 tagged, which allowed for the direct comparison of protein levels by Western blotting with an anti-V5 antibody. As expected, Sox17 overexpression coincided with a reduction in TCF4 and TCF3. In three separate experiments, Sox4 overexpression either caused no change or effected a slight increase in TCF4 protein levels and had little effect on TCF3 (Fig. 7J and K). Also, Sox4 overexpression had no effect on levels of coexpressed ß-catenin(S37A) protein (data not shown). These findings are the first to indicate that Sox factors regulate Wnt signaling by modulating the stability of TCF/LEF family members.
Sox17 stimulates ß-catenin and TCF4 protein degradation by the proteasome in a GSK3-independent manner. Our data suggest that the Sox17-mediated degradation of ß-catenin is GSK3ß independent because mutant forms of ß-catenin that cannot be phosphorylated by GSK3ß and thus are constitutively stable were still targeted for degradation by Sox17 (Fig. 7D and E). To confirm this finding, we inhibited GSK3ß activity by including lithium chloride in the culture medium and monitored levels of wild-type ß-catenin (myc tagged) alone or in the presence of increasing concentrations of a Sox17 plasmid. As expected, wild-type ß-catenin protein did not accumulate in COS cells (Fig. 8A, lane 1); however, including LiCl (40 mM) in the culture medium resulted in the accumulation of wild-type ß-catenin protein (Fig. 8A, lane 2). The coexpression of Sox17 in LiCl-treated cultures caused a dose-dependent loss of wild-type ß-catenin protein. This finding confirms that Sox17 promotes GSK3ß-independent degradation of wild-type ß-catenin protein.
|
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Nontranscriptional functions of Sox factors. We have demonstrated that Sox17 and Sox4 can directly interact with ß-catenin and TCF/LEF proteins. In the case of Sox17, it forms a complex with both ß-catenin and TCF. This finding is significant because one of the prevailing models is that Sox17 may antagonize Wnt activity by competing with TCFs for binding to ß-catenin, in which case Sox17 and TCF binding would be mutually exclusive. Our data show that Sox17 antagonizes the Wnt pathway by promoting the degradation of TCF/LEF and ß-catenin proteins via a GSK3ß-independent mechanism. Structure-function analyses indicate that Sox17 must be able to interact directly with both TCF and ß-catenin in order to promote their turnover. In contrast, while Sox4 can bind to TCF and ß-catenin individually, the three proteins do not appear to from a stable complex. In addition, Sox4 overexpression does not promote ß-catenin/TCF degradation like the overexpression of Sox17 does. Rather, reduced levels of Sox4 were associated with a reduction of ß-catenin protein, and the overexpression of Sox4 either had no impact on TCF4 levels or in some instances caused an increase in TCF4 protein levels. To our knowledge, this is the first demonstration that Sox proteins modulate canonical Wnt signaling by influencing levels of TCF/LEF factors.
In addition to a reduction in ß-catenin protein levels, the knockdown of Sox4 coincided with a decrease in the activation of ß-catenin and TCF target genes, such as cyclin D1, and reduced the proliferation of colon carcinoma cells, whereas reduced Sox17 levels had the opposite effect in these cells and in Xenopus embryos, resulting in the hyperactivation of Wnt target genes. Together, these data suggest that Sox proteins normally function to modulate ß-catenin/TCF activity in vivo. One possible mechanism by which Sox4 may stabilize ß-catenin was suggested by the finding of Lee et al. (19), who demonstrated that TCF3 can stabilize ß-catenin protein by preventing its association with Axin and APC.
Transcriptional functions of Sox, ß-catenin, and TCF proteins. Although we have demonstrated that the ability of Sox17 to antagonize ß-catenin and TCF is separable from its role as a transcription factor, it is possible that the transcriptional function of Sox proteins may also have an impact on Wnt signaling in some contexts. For example, Sox and TCF proteins may compete for common DNA binding sites or shared cofactors. In Xenopus, there is evidence that Sox3 and TCF bind to adjacent DNA sites to coordinately regulate the expression of the Wnt and ß-catenin target gene Xnr5 (54). In addition, it has recently been shown that the HMG domain of Sox factors can interact with other transcription factors, including homeodomain proteins, zinc finger proteins, and basic helix-loop-helix and leucine zipper proteins (3, 52), each of which may affect the ability of Sox factors to bind DNA or interact with cofactors such as ß-catenin and TCF proteins. Thus, in addition to Sox proteins' regulating ß-catenin and TCF protein stability, Sox-TCF complexes may regulate target gene transcription.
In the case of Sox17, there is evidence that it has two distinct functions, to repress or limit ß-catenin and TCF activity and to use ß-catenin as a cofactor to activate Sox17 target gene expression in Xenopus embryos. The depletion of Xenopus embryos of endogenous Sox17 protein results in the hyperactivation of Wnt target genes, and the depletion of ß-catenin results in reduced Sox17 target gene expression. Although transcriptional targets of Sox17 in the intestinal epithelium are unknown, we postulate that a similar mechanism may act there. Together, the present and previously published findings show that Sox17 can either utilize ß-catenin as a transcriptional cofactor or promote the degradation of ß-catenin and TCFs. Thus, Sox17 may radically shift the transcriptional output of the Wnt signaling pathway from TCF to Sox target genes. Ultimately, the function of Sox17 as a Wnt antagonist or a transcriptional activator may depend on the relative levels of Sox17, ß-catenin, and TCF/LEF. In a case in which the signaling pool of ß-catenin is in excess, we predict that high Sox17 levels could both repress Wnt signaling by promoting the degradation of ß-catenin-TCF complexes and activate the expression of Sox17 target genes.
Sox proteins and cancer. Oncogenic, activating mutations in ß-catenin stabilize the protein by preventing its phosphorylation by GSK3ß and subsequent degradation and are associated with the formation of colorectal carcinoma (26, 28, 46). Our data show that Sox17 promotes the degradation of oncogenic forms of ß-catenin, suggesting its possible role as a tumor suppressor. Consistent with such a function, Sox17 expression levels are down regulated in APCmin/+ tumors and in colon cancer as adenomas progress to carcinomas (32). In addition, Sox17 mRNA was almost undetectable in 66 human primary tumors derived from various tissues (16). Finally, our functional analysis of Sox17 clearly indicates that this factor represses colon carcinoma cell proliferation and Wnt signaling, possibly via targeting ß-catenin and TCF for degradation. These findings suggest that down regulating Wnt antagonists like Sox17 may be an obligatory step in tumor formation and progression.
In contrast to those of Sox17, elevated Sox4 levels are correlated with cancer in many contexts; Sox4 was identified as one of the most highly tumor-associated genes in an insertional mutagenesis screen for oncogenes (47). Elevated Sox4 levels are associated with small-cell lung carcinomas and medulloblastomas (9, 18, 45) and APCmin/+ adenomas (Fig. 1) (8, 32, 38). Our data provide a novel mechanism by which Sox4 may act as an oncogene, namely, by stimulating Wnt signaling, possibly by regulating ß-catenin or TCF protein stability and/or activity. Together, our findings suggest a novel mechanism by which some Sox proteins may act as protooncogenes while others may act as tumor suppressors by regulating ß-catenin and TCF protein activity.
Mammals have over 20 Sox proteins that are found in many cell types (41). The fact that most Sox proteins can have an impact on canonical Wnt responses in TOP-flash assays suggests that Sox-mediated regulation of ß-catenin and TCF proteins is a broadly utilized mechanism that may regulate Wnt signaling in a variety of physiological and pathological contexts, ranging from embryonic development to cancer. Moreover, this novel GSK3ß-independent means to degrade TCF/LEF and ß-catenin proteins represents a new molecular target for drug therapies that target the canonical Wnt pathway.
|
ACKNOWLEDGMENTS |
|---|
This research was supported by a career development award and a postdoctoral fellowship from the Juvenile Diabetes Research Foundation (J.M.W. and S.-C.J.L.), an NIH training grant in developmental and perinatal endocrinology, no. T32 HD07463 (J.R.S.), and the NIH grants HD42572 (A.M.Z.) and GM072915 (J.M.W.).
|
FOOTNOTES |
|---|
Published ahead of print on 17 September 2007.
These authors contributed equally.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2. Akiyama, H., J. P. Lyons, Y. Mori-Akiyama, X. Yang, R. Zhang, Z. Zhang, J. M. Deng, M. M. Taketo, T. Nakamura, R. R. Behringer, P. D. McCrea, and B. de Crombrugghe. 2004. Interactions between Sox9 and beta-catenin control chondrocyte differentiation. Genes Dev. 18:1072-1087.
3. Ambrosetti, D. C., C. Basilico, and L. Dailey. 1997. Synergistic activation of the fibroblast growth factor 4 enhancer by Sox2 and Oct-3 depends on protein-protein interactions facilitated by a specific spatial arrangement of factor binding sites. Mol. Cell. Biol. 17:6321-6329.[Abstract]
4. Cavallo, R. A., R. T. Cox, M. M. Moline, J. Roose, G. A. Polevoy, H. Clevers, M. Peifer, and A. Bejsovec. 1998. Drosophila Tcf and Groucho interact to repress Wingless signalling activity. Nature 395:604-608.[CrossRef][Medline]
5. Chang, C. M., C. L. Kao, Y. L. Chang, M. J. Yang, Y. C. Chen, B. L. Sung, T. H. Tsai, K. C. Chao, S. H. Chiou, and H. H. Ku. 2007. Placenta-derived multipotent stem cells induced to differentiate into insulin-positive cells. Biochem. Biophys. Res. Commun. 357:414-420.[CrossRef][Medline]
6. Clements, D., I. Cameleyre, and H. R. Woodland. 2003. Redundant early and overlapping larval roles of Xsox17 subgroup genes in Xenopus endoderm development. Mech. Dev. 120:337-348.[CrossRef][Medline]
7. Clevers, H. 2006. Wnt/beta-catenin signaling in development and disease. Cell 127:469-480.[CrossRef][Medline]
8. Du, Y., S. E. Spence, N. A. Jenkins, and N. G. Copeland. 2005. Cooperating cancer-gene identification through oncogenic-retrovirus-induced insertional mutagenesis. Blood 106:2498-2505.
9. Friedman, R. S., C. S. Bangur, E. J. Zasloff, L. Fan, T. Wang, Y. Watanabe, and M. Kalos. 2004. Molecular and immunological evaluation of the transcription factor SOX-4 as a lung tumor vaccine antigen. J. Immunol. 172:3319-3327.
10. Glinka, A., W. Wu, H. Delius, A. P. Monaghan, C. Blumenstock, and C. Niehrs. 1998. Dickkopf
C Sox17). ß-catenin and its associated proteins were precipitated with an anti-myc antibody, and TCF3 protein was analyzed by SDS-PAGE and autoradiography. IP, immunoprecipitate. (D and E) Under high-stringency conditions, Sox17-GST beads would not efficiently pull down ß-catenin. However, in the presence of TCF4 (D) or TCF3 (E), Sox17 was able to pull down ß-catenin. This interaction depends on the ß-catenin interaction domain of TCF3, which is deleted in the N terminus-lacking TCF3 protein (
