Oncogene (2006) 25, 3661–3669 & 2006 Nature Publishing Group All rights reserved 0950-9232/06 $30.00 www.nature.com/onc ORIGINAL ARTICLE The translocation SYT-SSX2 recruits b-catenin to the nucleus andassociates with it in an active complex

D Pretto, R Barco, J Rivera, N Neel, MD Gustavson and JE Eid

Department of Cancer Biology, Vanderbilt University Medical Center, Nashville, TN, USA

Localization of b-catenin in the cell is a key determinant cell adhesion at the surface and its oncogenic nuclear in its decision to function as a critical mediator of cell activation that could lead to tumor formation in somatic adhesion at the surface or a transcription activator in the cells (van Es et al., 2003). nucleus. SYT-SSX2 is the fusion product of the chromo- SYT-SSX is the product of a t(X;18)(p11;q11) somal translocation, t(X;18)(p11.2;q11.2), which occurs in translocation that occurs in synovial sarcoma, an synovial sarcoma, a soft tissue tumor. SYT-SSX2 is aggressive soft tissue tumor. It is the result of a fusion known to associate with chromatin remodeling complexes between the SYT (SYnovial sarcoma Translocated) andis proposedto be involvedin controlling gene on 18 and an SSX gene on chromosome X. expression. We report that SYT-SSX2 plays a direct role The chromosomal rearrangement is detected in almost in b-catenin regulation. When expressedin mammalian all cases of synovial sarcoma. Thus SYT-SSX is cells, SYT-SSX2-induced b-catenin recruitment to the considered to play a central role in the development of nucleus. Interestingly, known target of canonical this cancer. However, the mechanism of tumor initiation Wnt were not activatedas a result of SYT-SSX2 by SYT-SSX is still unknown. SYT and SYT-SSX are expression, nor was the nuclear localization of b-catenin nuclear that associate with chromatin remodel- due to one of the signaling pathways normally implicated ing complexes (SWI/SNF) and are thought to be in this event. b-Catenin accumulation in the nucleus ledto involved in the control of gene expression (dos Santos the formation of a transcriptionally active nuclear et al., 2001; Ladanyi, 2001). complex that containedSYT-SSX2 and b-catenin. More The SSX gene family is composed of nine contiguous importantly, depletion of SYT-SSX2 in primary synovial members (SSX1–9). Although the chromosomal ar- sarcoma cells resultedin loss of nuclear b-catenin signal rangement of the SSX genes could in principle allow for anda significant decrease in its signaling activity. These heterogeneity at the breakpoint, the translocation in results unravel a novel pathway in the control of b-catenin synovial sarcoma has so far been detected either in the cellular transport andstrongly suggest that SYT-SSX2 SSX1 or SSX2, and, rarely in the SSX4 gene (Ladanyi contributes to tumor development, in part through et al., 2002). The identity of the rearranged SSX gene b-catenin signaling. has significant clinical impact. Although SYT-SSX2 is Oncogene (2006) 25, 3661–3669. doi:10.1038/sj.onc.1209413; linked to a monophasic histologic subtype, SYT-SSX1 is published online 6 February 2006 often detected in biphasic tumors and is correlated with a shorter metastasis-free period and patient survival Keywords: synovial sarcoma; SYT-SSX2; b-catenin (Ladanyi et al., 2002). SYT-SSX1 was recently shown to possess proliferative and transforming activity (Nagai et al., 2001), as well as to increase the stability of cyclin D1 (Xie et al., 2002). Introduction In the majority of synovial sarcomas, a striking cytoplasmic and nuclear accumulation of b-catenin is Nuclear translocation and retention is at the center of observed. Remarkably, b-catenin accumulation occurs b-catenin function. Its ability to integrate the multiple in the absence of genetic mutations in the tumor Wnt pathways that involve progenitor stem cell pro- suppressor APC or the b-catenin gene. This led to the liferation, differentiation and axis specification relies on belief that nuclear accumulation of b-catenin in synovial its presence in the nucleus where it associates with TCF/ sarcomas is the consequence of a distinct pathway LEF1 and induces gene expression. Moreover, b-catenin proper to the tumor (Hasegawa et al., 2001). shuttling between the nucleus and the cytoplasm is key In this report, we provide evidence that the fusion for maintaining the balance between its role in cell-to- protein SYT-SSX2, through a dominant gain of func- tion, recruits b-catenin to the nucleus. There, both Correspondence: Dr JE Eid, Department of Cancer Biology, proteins coreside in distinct foci and b-catenin signaling Vanderbilt University Medical Center, 23d Avenue South at Pierce, is activated. This regulation of b-catenin function is PRB 740, Nashville, TN 37232, USA. E-mail: [email protected] apparently independent of canonical Wnt activation and Received 3 August 2005; revised 21 December 2005; accepted 21 reveals one mechanism by which SYT-SSX likely December 2005; published online 6 February 2006 initiates tumorigenesis. SYT-SSX2 regulates b-catenin location and function D Pretto et al 3662 Results SSX2). By contrast, nuclear b-catenin was undetected in cells infected with the retroviral vector alone, SYT or Expression of the synovial sarcoma fusion protein SYTdl8 (Figure 1a, panels V, SYT and SYTdl8, SYT-SSX2 causes b-catenin nuclear accumulation respectively). All three exogenous proteins were ex- Synovial sarcoma is thought to originate from mesench- pressed at equivalent levels (Figure 1b) and the infection ymal stem cells. Therefore, we decided to conduct our efficiency was 80–90% with all four vectors. studies using NIH3T3 as target cells, in search for SYT- These results indicate that SYT-SSX2 expression SSX function. To this end, NIH3T3 cells were transi- induces nuclear accumulation of endogenous b-catenin. ently infected with POZ-retroviral vectors (Ikura et al., The change in localization of cellular b-catenin exerted 2000) expressing full-length (FL-) SYT, SYT-SSX2, or by SYT-SSX2 occurs through a gain of function caused SYTdl8 (corresponding to the full-length SYT lacking by SSX addition rather than deletion of the SYT the last carboxy-terminal eight amino acids, which are COOH-terminal eight amino acids. missing in the SYT-SSX fusion product). Remarkably, 48 h postinfection, an intense endogenous nuclear SYT-SSX2 recruits b-catenin to its nuclear foci where b-catenin signal appeared in 20–30% of cells expressing both proteins form an in vivo complex SYT-SSX2. There was no obvious cytoplasmic accumu- SYT and SYT-SSX are known to reside in nuclear foci lation of b-catenin in these cells, and no noticeable as a result of their association with chromatin modifying increase in its total cellular level (Figure 1a, panel SYT- complexes. Their nuclear structures, however, are not

Figure 1 Nuclear accumulation of b-catenin in SYT-SSX2-expressing cells. NIH3T3 cells were infected with POZ (V), POZ-SYT, POZ-SYT-SSX2 and POZ-SYTdl8 carrying both FLAG and HA tags. (a) Visualization by immunofluorescence of the expressed proteins with HA-tag antibody (left panel) and of endogenous b-catenin with the C19220 Ab (middle panel). The right panel corresponds to DAPI staining. The b-catenin signal and DAPI-stained nuclei were merged. ‘ þ nuclei’ lane shows the percentage of infected cells exhibiting nuclear b-catenin. The numbers represent mean7s.d. (n ¼ 6). In every experiment a total of 1000 cells were counted for each construct. The nuclear localization of SYT, SYT-SSX2 and SYTdl8 is due to the presence of the N-terminal nuclear localizing region (amino acids 51–90; dos Santos et al., 2000) present in all three polypeptides. A faint cytoplasmic HA signal, however, was noticed in the NIH3T3 cells infected with POZ-SYT and POZ-SYTdl8. This is likely due to minimal leakage of human proteins expressed in murine cells. Such leakage was not detected in SYT-SSX2 infected cells where the HA staining was exclusively nuclear. This could be explained by the additional presence of the nuclear localizing SSXRD region (the 34 C-terminal amino acids) of SSX2. SSXRD likely provides a more efficient binding of SYT-SSX2 on chromatin and better nuclear retention (Soulez et al., 1999). Note the absence of cytoplasmic HA staining when SYT was expressed in U2OS, a human osteosarcoma cell line (Figure 2, compare panels a and b). SYTdl8 expression gave results similar to SYT (data not shown). Images were taken at  20 magnification with the Zeiss Axioplan 2 fluorescent microscope and merging of DAPI and b-catenin signals was performed using the same contrast. (b) Levels of expressed POZ, POZ-SYT, POZ-SYT-SSX2, and POZ-SYTdl8 in infected NIH3T3 cells, visualized by the FLAG M2 Ab. Note that a FLAG/HA-positive cytoplasmic low molecular weight vector-derived polypeptide is generated from the POZ plasmid. Its synthesis was interrupted when the SYT-derived cDNAs were inserted in the POZ backbone plasmid.

Oncogene SYT-SSX2 regulates b-catenin location and function D Pretto et al 3663 identical and they exhibit cell-to-cell variability in size and number (dos Santos et al., 1997; Thaete et al., 1999). In our studies, SYT and SYT-SSX2 were both seen in small foci evenly distributed in the nuclear compartment (Figure 2a of NIH3T3 cells, HA). Upon translocation to the nucleus, b-catenin localized in a subset of the SYT- SSX2 foci (Figure 2a, panel SYT-SSX2 and merge). No b-catenin could be detected in the SYT nuclear speckles (Figure 2a, panel SYT and merge). Furthermore, in human osteosarcoma (U2-OS; a mesenchymal tumor) cells, SYT-SSX2 formed large distinct nuclear bodies, whereas SYT was distributed in fine nuclear dots (Figure 2b of U2-OS, HA). In these cells, a subpopula- tion of the endogenous nuclear b-catenin was found recruited to a subset of the SYT-SSX2 nuclear bodies (Figure 2b, SYT-SSX2 and merge). No b-catenin colocalized with nuclear SYT (Figure 2b, SYT and merge). SYTdl8 exhibited a behavior similar to SYT, both in NIH3T3 and U2OS cells. It formed fine nuclear foci devoid of b-catenin in both cell lines (data not shown). Together, these data reveal that SYT-SSX2 resides in unique structures within the nucleus. These structures likely represent SYT-SSX2 protein complexes of differ- ent biological activities. SYT-SSX2 recruits b-catenin to its own distinct foci in the nucleus. The association of b-catenin in a subpopulation of these foci signifies that it participates in mediating, in part, SYT-SSX function. In vivo complex formation between SYT-SSX2 and b-catenin could be confirmed at the biochemical level. A specific antibody (SV11) that recognizes SYT, SYT- SSX2 and SYTdl8, was able to co-precipitate b-catenin and SYT-SSX2, but failed to do so significantly in NIH3T3 cells expressing SYT, SYT-dl8 or POZ vector (Figure 2c). Similarly, a specific SYT-SSX2/b-catenin complex was detected when both polypeptides were coexpressed along with TCF4 (Figure 2d). Further in vitro studies, however, showed that SYT-SSX does not bind directly to b-catenin, implying that in vivo, other partners likely serve as a bridge between the two proteins to form a nuclear complex (Figure 2e). Figure 2 Recruitment of endogenous b-catenin to SYT-SSX2 nuclear foci. (a) NIH3T3 and (b) U2-OS infection with POZ-SYT b-catenin nuclear recruitment by SYT-SSX2 is and POZ-SYT-SSX2. SYT and SYT-SSX2 were visualized with independent of canonical Wnt HA Ab and b-catenin with C19220 Ab. The right image is a merge of both signals. Images were taken at 63 magnification with the In human cancers, accumulation of b-catenin in the Zeiss Axioplan 2 fluorescent microscope and merging of HA and b- nucleus is normally due to incapacitation of its catenin signals was performed using the same contrast. b-Catenin degradation machinery. This occurs when GSK-3b- signal and merge results in cells infected with POZ vector alone or dependent components of the Wnt pathway, such as expressing SYTdl8 were similar to SYT (data not shown). (c) In vivo SYT-SSX2/b-catenin complex formation in infected cells. APC, axin or b-catenin itself are mutated. As a result, NIH3T3 cells were infected with POZ, POZ-SYT, POZ-SYT-SSX2 b-catenin is stabilized, reaches the nucleus where it and POZ-SYTdl8. The SV11 Ab was used for co-IP. SV11 associates with TCF/LEF1 and induces the expression recognizes both SYT and SYT-SSX2. Levels of endogenous b- of proliferation genes such as cyclin D1 and MYC, or catenin in lysates of infected cells are shown. All the POZ proteins the matrilysin (MMP7) gene. These events are impli- were expressed at equivalent levels (data not shown). (d) In vivo complex formation of cotransfected b-catenin, SYT-SSX2 and cated in the formation of colon, pancreatic and liver TCF4 in NIH3T3 cells. NIH3T3 were transfected with 1 mg of wt- cancers (Doucas et al., 2005). b-catenin, 1 mg of TCF4 cDNAs, and 3 mg of POZ-SYT or POZ- In SYT-SSX2-expressing cells, the cellular levels and SYT-SSX2. Lanes 1 and 2 show FLAG (M2-affinity gel) co- the stability of b-catenin remained the same. Further- immunoprecipitation (co-IP) and lanes 3 and 4 are TCF4 co-IP. (e) GST in vitro binding assays. Upper panel: 35S-b-catenin binding to more, there was no noticeable increase in any of the the indicated GST-fusion polypeptides. Lower panel: levels of known Wnt targets; cyclin D1, MYC (Figure 3a), or GST-TCF4, GST-SYT and GST-SYT-SSX2 in the bacterial matrilysin (data not shown). The level of unpho- lysates, using the TCF4 and SV11 Abs, respectively.

Oncogene SYT-SSX2 regulates b-catenin location and function D Pretto et al 3664 sphorylated b-catenin (Figure 3a), considered the active location in SYT-SSX expressing cells. In addition, levels form of the protein in the nucleus (Staal et al., 2002) was of TCF4 remained constant (Figure 3a) and there was also unchanged. Moreover, in vitro assays showed no detectable LEF (data not shown). The latter is normal GSK-3b kinase activity and constant levels of known to escort b-catenin and retain it in the nucleus as its target b-catenin species phosphorylated on residues a result of integrin-linked kinase (ILK)-mediated S33, S37, T41 and S45 (data not shown). induction (D’Amico et al., 2000). In summary, SYT-SSX2 recruitment of nuclear b- To complete our studies, we treated SYT-SSX2- catenin did not appear to be a consequence of canonical expressing cells with specific pathway inhibitors and Wnt pathway activation. monitored their effect on b-catenin nuclear accumula- In sarcomas, aberrant activation of growth factor tion. Addition of LY29004 or wortmannin (PI3Kinase/ and receptor tyrosine kinase (RTK) signaling pathways AKT inhibitors), SB203580 (the p38 MAPK (mitogen- and their downstream effectors is frequently observed activated protein kinase) inhibitor), and UO126 or (Allander et al., 2002). It is also known that tyrosine PD98059 (inhibitors of ERK (extracellular signal phosphorylation regulates the behavior of b-catenin. A regulated kinase)), failed to abolish nuclear b-catenin number of RTKs such as those of epidermal growth (Figure 3b and data not shown). factor (EGF; Hoschuetzky et al., 1994), insulin-like Together, these data indicate that the nuclear growth factor (IGF; Playford et al., 2000), and translocation protein SYT-SSX2 controls b-catenin hepatocyte growth factor-like ligands (RON and localization through a novel mechanism independent MET; Danilkovitch-Miagkova et al., 2001) were shown of canonical Wnt or other pathways known to regulate to phosphorylate b-catenin and may lead to its nuclear this process. translocation. Tyrosine residue 142 (Tyr 142; Piedra et al., 2003) phosphorylation is a common target. Recently, the oncogene homolog BCL9-2 was also SYT-SSX2 activates b-catenin signaling reported to phosphorylate b-catenin on Tyr 142, and in a p300-dependent manner to assist in its transport to and retention in the nucleus Nuclear translocation of b-catenin is normally required (Brembeck et al., 2004). but not sufficient for its transcriptional activation In our studies, however, we were unable to detect a function (Prieve and Waterman, 1999). In order to change in the tyrosine phosphorylation status of cellular determine whether SYT-SSX nuclear recruitment of b-catenin (Figure 3a), excluding the possibility that one b-catenin activates its signaling, we conducted transcrip- of the aforementioned pathways is altering b-catenin tion studies where we compared the effect of SYT, SYT- SSX2 and SYTdl8 on b-catenin/TCF4 activity, using the TOP-FLASH reporter construct. Although SYT and SYTdl8 exhibited a modest stimulation over back- ground effect, we observed a steep surge in b-catenin activation by SYT-SSX2 (Figure 4a). All three proteins were expressed at adequate levels (Figure 4a, inset). SYT and SYT-SSX2 are not DNA-binding transcrip- tion factors. It is therefore thought that they exert their gene regulation function by associating with coactivator protein partners (Clark et al., 1994). We have previously identified SYT as one of the transcriptional coactivator p300-interacting proteins. SYT binds p300 at its NH2- terminus, a region also present in SYT-SSX2 (Eid et al., 2000). Recently, we found that, similar to SYT, SYT- SSX2 makes direct contact with p300 in its C/H1 and C/H3 (unpublished data; Barco and Eid) domains. In order to acquire insight into the regulation of the nuclear b-catenin/SYT-SSX2 complex, we therefore asked whether p300 contributed to its activation. For this, we inhibited p300 by coexpressing in the TOP- Figure 3 SYT-SSX2 induces b-catenin nuclear localization FLASH activation assay described above, wt 12S-E1A, through a novel pathway. (a) Levels of the designated proteins in the transforming adenovirus polypeptide known to bind NIH3T3 cells infected with POZ, POZ-SYT, POZ-SYT-SSX2 and p300 and inhibit its function. This resulted in complete POZ-SYTdl8. a-Tubulin shows equal protein loading. All are suppression of SYT-SSX2 effect on b-catenin signaling. derived from boiled total cellular lysates with the exception of tyrosine-phosphorylated b-catenin, which was visualized by an IP Significantly, an E1A mutant (E1A D2–36) unable to with the b-catenin H-102 Ab and blotting with the HRPO-p-Tyr bind p300 rescued SYT-SSX2 stimulatory effect by Ab. (b) Visualization of endogenous b-catenin in SYT-SSX2- more than 50% (Figure 4b). expressing cells after an overnight treatment with 25 mM of We were next interested in investigating whether the LY29004, SB203580 and UO126 before fixation. DMSO served as control. SYT-SSX2 was visualized with HA Ab. Images were transcription coactivating effect of SYT-SSX2 is con- taken at  20 magnification in a Zeiss Axioplan 2 fluorescent text-specific, or a general feature of the fusion protein microscope. conferred on all gene promoters. To answer this

Oncogene SYT-SSX2 regulates b-catenin location and function D Pretto et al 3665 question, we compared the three SYT-derived proteins stimulatory effect is context-dependent, with b-catenin for their ability to stimulate p300 activity in a GAL- as one of its specific targets. Both proteins may form a promoter reporter assay. Neither SYT, SYT-SSX2, nor transcriptionally active complex in the nucleus. SYTdl8 exerted any significant change on GAL-p300 function (Figure 4c). SYT-SSX2 regulates b-catenin function in primary Altogether, these results suggest that the translocation human synovial sarcoma cells protein, through its SSX region, acquires a positive gain We next endeavored to explore the biological signifi- of function that stimulates b-catenin signaling. Intact cance of our findings as they applied to the cancer itself p300 function seems to be required for such effect. and whether the regulation of b-catenin localization Failure of SYT-SSX2 to enhance transcription driven and signaling by exogenous SYT-SSX2 seen in NIH3T3 by the GAL promoter indicates that the SYT-SSX2 is a reflection of actual events in synovial sarcoma. For this, we assayed primary human synovial sarcoma cells (Syn-1 with an SYT-SSX2 translocation). As reported previously, accumulation of b-catenin is frequently observed in synovial sarcomas (Hasegawa et al., 2001). Therefore, we were not surprised to discover b-catenin in approximately 90% of untreated Syn-1 nuclei. Specific depletion of the SYT-SSX2 protein by transient RNA interference (RNAi) resulted in the disappearance of more than half the nuclear b-catenin from 98% of Syn-1 nuclei (Figure 5a, panel b-catenin and merge, and panel 5b). The 38% nuclear enrichment of cellular b-catenin in the control cells (INV) dropped to 17% in SiRNA cells (Si-SSX2). Meanwhile, the steady-state levels of total cellular b-catenin remained constant (Figure 5b and c, insets; b-catenin), suggesting that SYT-SSX2 likely influences b-catenin function primarily by controlling its cellular compartmentalization rather than its expression. We then used the TOP-FLASH reporter assay to test whether the nuclear exit of b-catenin from Syn-1 RNAi cells resulted in a con- comitant decrease in its signaling. We observed that Syn-1 cells contained a constitutively active b-catenin/ TCF4 complex. Upon RNAi of SYT-SSX2, the en- dogenous b-catenin-specific TOP-FLASH activity was reduced to background FOP-FLASH levels (Figure 5c). To summarize, based on our SYT-SSX2 RNAi data in Syn-1 cells, we conclude that in synovial sarcoma b- catenin accumulates in the nucleus and is constitutively active. Reversal of these two processes upon depletion of

Figure 4 p300-dependent stimulation of b-catenin signaling by SYT-SSX2. (a) Transcription activation studies in transfected NIH3T3 cells using the TOP-FLASH reporter (a luciferase reporter gene with an optimized TCF binding site) assay. FOP- FLASH (luciferase reporter gene with a mutated and inactive TCF site) was included for specificity. The indicated amounts of transfected plasmids were included in a total of 5 mg of transfected DNA of which 0.5 mgofCMV-b-GAL was used for internal control. The difference in amount of transfected DNA was compensated for with POZ backbone vector. Luciferase activities were measured 48 h post-transfection. All activities were measured relative to background levels obtained with POZ vector alone. Inset shows the protein levels of the transfected POZ-derived cDNAs by immunoblotting (IB) with FLAG M2 Ab. Similar activation results were obtained in at least five independent experiments. (b) Transfections and activation assays were conducted as described in (a), with the excepted addition of the indicated amounts of E1A and E1A D2–36. (c) Activation studies in NIH3T3 cells were conducted as described above and as indicated in the figure. In total, 0.5 mg of GAL-luciferase reporter vector and 5 mg of total transfected DNA were used. All activities were measured relative to GAL-p300.

Oncogene SYT-SSX2 regulates b-catenin location and function D Pretto et al 3666

Figure 5 SYT-SSX2 regulates b-catenin signaling in synovial sarcoma. RNAi of SYT-SSX2 in primary human synovial sarcoma cells (Syn-1). Si-SSX2 is the specific SSX targeting oligo and INV is control oligo. (a) Immunofluorescence in untreated (UNRx), INV and Si-SSX2 Syn-1 cells. Left panel: endogenous SYT-SSX2 visualized with the B56 Ab. Middle panel: b-catenin visualized with the C19220 Ab. In the right panel, nuclei are stained with DAPI. b-Catenin and DAPI signals were merged. ‘ þ nuclei’ shows the percentage of b-catenin-positive nuclei. The numbers represent mean7s.d. (n ¼ 3). Images were taken at  20 magnification in a Zeiss Axioplan 2 fluorescent microscope and the merge panels were acquired using the same contrast. Inset panel reveals SYT-SSX2 depletion in Si-SSX2 cells shown with both B56 and SV11 Abs. a-Tubulin reveals equal protein loading (b) fractionation of INV and Si-SSX2 RNAi Syn-1 cells. The top panel shows the relative subcellular distribution of b-catenin in both Syn-1 populations. Relative intensities were measured using the FluorChem program, then standardized with the cytoplasmic (a-tubulin; lower panel) and the nuclear (lamin A/C; middle panel) controls. Inset panel reveals SYT-SSX2 depletion in Si-SSX2 with the SV11 Ab. a-Tubulin reveals equal protein loading. (c) TOP and FOP-FLASH activities post-RNAi in INV and Si-SSX2 cells. A total of 5 mg of transfected DNA included 2 mg either of TOP- or FOP-FLASH, 1 mg of CMV-b-GAL for internal control, and 2 mg of pCDNA3 as carrier. Luciferase activities were measured 16 h after transfection. Transcriptional activities were measured relative to the control (INV) cells transfected with TOP-FLASH. Inset panel shows SYT-SSX2 depletion in Si-SSX2 cells with the SV11 Ab. a-Tubulin reveals equal protein loading. All data in a–c were obtained at 48 h post-RNAi transfection.

SYT-SSX2 in the primary cancer cells implies that the Exogenous expression of SYT-SSX2 in mesenchymal translocation protein is directly responsible for both, cells induces accumulation of b-catenin in the nucleus and this could represent part of its function in tumor where both proteins form an active complex. Depletion initiation. of SYT-SSX2 in primary synovial sarcoma cells reverses this process and inactivates b-catenin signaling. The effect of the fusion protein on b-catenin localization and Discussion function appeared to be independent of canonical Wnt and the signaling pathways known to exert similar events. In this report, we describe a new regulation of b-catenin b-catenin transduces the Wnt pathway by translocat- function by a cancer translocation protein, SYT-SSX2. ing to the nucleus and associating with TCF factors. In

Oncogene SYT-SSX2 regulates b-catenin location and function D Pretto et al 3667 development it controls cell fate determination, axis SYT-SSX2 or b-catenin and identify the common formation and stem cell differentiation. In somatic cells, targets and oncogenic signaling pathways activated by constitutive activation of b-atenin is oncogenic (Moon the two oncogenes, SYT-SSX2 and b-catenin. et al., 2002). For these events b-catenin makes the Finally, synovial sarcoma is an aggressive tumor critical switch between adhesion and nuclear signaling. resistant to conventional therapy. As in other neoplasms The mechanism of b-catenin nuclear shuttling is still characterized by a simple and distinct cytogenetic defect largely unknown and an active area of research is such as leukemias, translocation proteins provide an currently to identify the cellular factors involved in ideal target for therapy. Reversal of b-catenin functions escorting b-catenin to the nucleus and those that help by RNAi of SYT-SSX2 in Syn-1 cells suggests an active anchor it on the chromatin for its transcriptional role of the fusion protein in b-catenin deregulation activation. Two such factors have recently been isolated commonly detected in cases of synovial sarcoma. In this in Drosophila: Legless (the human oncogene BCL9-2 regard, b-catenin signaling is an intriguing candidate for homolog) and Pypogus. The former is required therapeutic intervention in this cancer. for targeting b-catenin to the nucleus and the latter for anchoring it and providing chromatin access to b-catenin or TCF. There, all three proteins are thought to form a quaternary active complex with TCF (Bienz, Materials andmethods 2004). Cells, lysis, immunoprecipitation and reagents SYT-SSX2 is primarily an obligate nuclear chromatin NIH3T3 cells were grown in DMEM with 10% FCS. Syn-1 binding protein. Thus, its ability to direct b-catenin to a cells are primary human synovial sarcoma cells (a kind gift subset of its nuclear foci implies that either it escorts b- from Jonathan Fletcher; Brigham and Women’s Hospital, catenin to the nucleus in a fashion similar to Legless/ Boston, USA). They were cultured in RPMI 1640 supplemen- BCL9-2 or it activates an inappropriate pathway that ted with 20% FCS. To assess protein expression levels, cells induces a yet unknown cellular protein to accompany b- were lysed in RIPA buffer (50 mM Tris 7.5, 150 mM NaCl, 1% catenin to the nucleus, where, like Pypogus, SYT-SSX2 DOC, 1% Triton X-100, 0.1% SDS, 1 mM DTT) for 30 min at anchors it on chromatin. Conversely, SYT-SSX2 could 41C. Lysates were then boiled in 2 Â sample buffer and equal be retaining b-catenin in the nucleus by overriding the protein amounts loaded on SDS–PAGE. For immunoprecipi- function of its tumor suppressor exporters such as APC tation (IP), cells were lysed in 150 NTN buffer (20 mM Tris pH 8.0, 150 mM NaCl, 0.5% NP40), then pelleted to remove or Axin. We are currently exploring these eventual insoluble material. The resulting supernatant was incubated scenarios. with primary antibody (Ab) for 4 h at 41C, and, where needed, When observed in various human tumors or as a protein A-Sepharose beads (Amersham Biosciences, Uppsla, result of oncogenic signaling pathways, nuclear accu- Sweden) were added to purify the immunocomplex. After three mulation of b-catenin is usually accompanied by an washes with 150 NTN, the beads were boiled in sample buffer. inappropriate activation of its signaling (D’Amico et al., SV11 is a SYT-specific monoclonal antibody raised against 2000; Persad et al., 2001; Patturajan et al., 2002; the COOH-terminal region (residues: 331–387; Jim de Caprio Thompson et al., 2002; Tissier et al., 2005). The outcome lab; Dana Farber Cancer Institute, Boston, MA, USA). B56 is of such activation varies depending on the cellular an SSX-specific polyclonal antibody raised against the peptide context. Our data describing b-catenin nuclear exit HRLRERKQLVI (Open Biosystems, Huntsville, AL, USA). Rabbit b-catenin (C2206) and HA (H 6908), as well as the consequent to SYT-SSX2 depletion in synovial sarcoma mouse a-tubulin (T 5168) and FLAG (M2 and M2-affinity gel) cells and the formation of a transcriptionally active in Abs were all purchased from Sigma (Saint Louis, MO, USA). vivo complex that contains b-catenin and SYT-SSX2, Mouse anti-b-catenin (C19220) and HRPO-phosphotyrosine imply that the sarcoma chimeric protein contributes to (P11625-050) Abs were obtained from BD Biosciences (San cancer formation by activating one of b-catenin target Diego, CA, USA). Rabbit anti-b-catenin (H-102) and mouse programs. Identifying the genes activated by the SYT- cyclin D1 (sc-450) were acquired from Santa Cruz, Santa Cruz, SSX2/b-catenin complex in a microarray analysis will be CA, USA). Anti-MYC tag (9E10) was kindly provided by key in determining such programs. These studies are David Livingston (Dana Farber cancer Institute, Boston, underway. The transcription coactivator p300 and USA). Rabbit anti-MYC (06-340) and mouse anti-TCF4 components of the SWI/SNF chromatin remodeling (clone 6H5-3) were purchased from Upstate (Lake Placid, NY, USA). Mouse anti-dephospho-b-catenin Ab (8E4) was complex such as BRG1, are likely candidates for acting obtained from A.G. Scientific, San Diego, CA, USA. UO126, as protein bridges in the active complex. Both SYT- SB203580 and LY294002 were purchased from Calbiochem, SSX2 and b-catenin are able to associate with them La Jolla, CA, USA. directly. Moreover, both BRG1 (Barker et al., 2001) and p300 (Miyagishi et al., 2000) have been shown to regulate b-catenin signaling. Cell fractionation 6 As illustrated in our colocalization studies, b-catenin After 2 washes with PBS, 5 Â 10 Syn-1 cells were pelleted then is not the only partner of SYT-SSX2. It is present in a suspended and swollen in 2 volumes of hypotonic buffer (10 mM Hepes, pH 7.9, 1.5 mM MgCl2, 10 mM KCl and 0.5 mM subset of SYT-SSX2-containing nuclear bodies. There- DTT) for 10 min. Nuclei and cytoplasmic fractions were fore, in order to determine the b-catenin proper obtained by brief lysing with 5% NP-40 (2 ml of 5% NP-40 for biological function affected by SYT-SSX2 in synovial 50 ml of lysate), then separated by slow centrifugation sarcoma, an informative approach will be to generate (2500 rpm for 5 min). All procedures were performed on ice. extensive gene expression profiles of Syn-1 cells lacking After an additional wash of the nuclei with hypotonic buffer,

Oncogene SYT-SSX2 regulates b-catenin location and function D Pretto et al 3668 both fractions were boiled in protein sample buffer and et al., 2002). The virus was harvested 2 days after transfection analysed on SDS–PAGE. and experimental analyses were conducted on NIH3T3 cells 2 days after infection. Cloning of SYT and SYT-SSX expression vectors GST-TCF4 was a kind gift from Lynn Matrisian (Vanderbilt University, Nashville, TN, USA). GST-SYT was previously Immunofluorescence described (Eid et al., 2000). GST-SYT-SSX2 was generated as Cells were plated on cover slips coated with 0.2% gelatin in follows. The SSX component of the oncogene SYT-SSX2 was 1 Â PBS. They were fixed for 15 min in PBS containing 3% amplified by RT–PCR (Invitrogen, Carlsbad, CA, USA) from paraformaldehyde and 2% sucrose. After permeabilization a human synovial sarcoma tumor. The primers used were with PBS/0.2% Triton X-100 for 5 min, they were blocked with JE-41: 50-GCGCAGGCCTTATGGATATGACCAG-30 and 3% goat serum in PBS and incubated with primary Abs for 2 h JE-42: 50-GCGCGAATTCTTACTCGTCATCTTCCTC-30. at room temperature. After three washes with PBS, coverslips 1 GST-SYT-SSX2 was constructed by ligating the SSX2 PCR were incubated with Alexa-conjugated 2 Abs (Molecular product cut with Stu1 and EcoR1, the Sma1/Stu1 SYT-coding Probes, Eugene, OR, USA) for 30 min, washed and mounted fragment derived from pCDNA3/SYT-HA, and the PGEX-2T with DAPI-Vectastain (Vector, Burlingame, CA, USA). Zeiss digested with Sma1 and EcoR1. Wt b-catenin-expressing (Axioplan 2) fluorescence microscope was used for cell vector (pCI-neo-b-cateninXL) was a kind gift from Bert visualization and counting. Vogelstein (Johns Hopkins Hospital, Baltimore, MD, USA). The POZ-based expression vectors (a kind gift from Pat RNAi of SYT-SSX2 in synovial sarcoma (Syn-1) cells Nakatani; Dana Farber Cancer Institute, Boston, USA) were RNAi in SYN-1 cells was performed as described previously constructed by inserting, in frame, between the Not1 and the (Elbashir et al., 2001), using 3 ml of oligofectamine (Invitrogen) Xho1 sites of POZ, SYT, SYT-SSX and SYTdl8. SYT and and 10 ml of the indicated RNA duplex for each well. The SYTdl8 were PCR products amplified from GST-SYT, and specific SSX2 sequence targeted for SiRNA was 50-AAC SYT-SSX from GST-SYT-SSX. The primers used were, SYT- 0 0 0 CAACTACC TCTGAGAAGA-3 (Si-SSX2). The nonspecific 5 :5-GCGCCTCGAGTAGACCACCATGTCTGTGGCTT 0 0 0 0 control sequence (INV) was 5 -TCCTGGGTATCCGTATAA TCGCGGCC-3 and SYT-3 :5-GCGCGCGGCCGCCTGCT 0 0 0 0 0 GAC-3 (Qiagen, Valencia, CA, USA). INV corresponds to an GGTAATTTCCATA-3 for FL-SYT; SYT-5 and D8-3 :5- inverted sequence selected from the SYT cDNA. GCGCGCGGCCGCCTGGTCATATCCATAAGG-30 for SYT-D8; and SYT-50 with SSX-30:50-GCGCGCGGCCGCC TCGTCATCTTCCTCAGG-30 for SYT-SSX. All three pro- Activation/luciferase assays teins acquired at the COOH-terminus both an HA and a For activation studies in Figure 4, NIH3T3 cells were FLAG tags built in POZ. Cloned fragments were verified by transfected using the Superfect (Qiagen) reagent according to sequence analysis. manufacturer’s instructions. Experiments were conducted in triplicates. Luciferase activities were measured using the In vitro binding assay Luciferase reporter assay (Promega, Madison, WI, USA). 35S-labeled in vitro-translated (IVT) b-catenin was generated from wt-b-catenin plasmid following vendor’s instructions (Promega, Madison, WI, USA). For the in vitro assay, 5 mlof Acknowledgements IVT b-catenin were incubated with GST, GST-SYT, GST- SYT-SSX and GST-TCF4 immobilized on glutathione beads. We thank Kelli Gunn and the Al Reynolds group for technical The binding buffer was 25 mM Tris 7.5, 150 mM NaCl, 1 mM support, Pat Nakatani for the POZ retroviral vector, Hans DTT and 0.1% NP40. The binding reaction was carried at 41C Clevers and Bert Vogelstein for the TOP-FLASH, FOP- for 30 min, after which the complexes were washed three times FLASH, MYC-and wt-b-catenin constructs, Jonathan Fletch- in binding buffer without DTT. The beads were then boiled in er for the Syn-1 cells, Ray Mernaugh for help with peptide Laemmli buffer, resolved on SDS–PAGE and subjected to design and Open Biosystems for the SSX2 antibody. This autoradiography. Of IVT b-catenin, 2 ml, were included in the work was supported by Grants K22 CA098008-01 and input lane. R01CA106481-01, and in part by a Cancer Center Support Grant CA68485 and Center in Molecular Toxicology Support Retroviral infection Grant EB00672. Some of the work began in David Living- Transient retroviral infections were conducted using the ston’s laboratory (Dana Farber Cancer Institute) and was phoenix packaging cell system, as described previously (Ireton supported by Grant NCI CA 15751.

References

Allander SV, Illei PB, Chen Y, Antonescu CR, Bittner M, Danilkovitch-Miagkova A, Miagko A, Skeel A, Nakaigawa N, Ladanyi M et al. (2002). Am J Pathol 161: 1587–1595. Zbar B, Leonard EJ. (2001). Mol Cell Biol 21: 5857–5868. Barker N, Hurlstone A, Musisi H, Miles A, Bienz M, Clevers dos Santos NR, de Bruijn DRH, van Kessel AG. (2001). H. (2001). EMBO J 20: 4935–4943. Genes, Chrom Cancer 30: 1–14. Bienz M. (2004). Curr Biol 15: R64–R67. dos Santos NR, de Bruijn DRH, Kater-Baats E, Otte AP, van Brembeck FH, Schwraz-Romond T, Bakkers J, Wilhelm S, Kessel AG. (2000). Exp Cell Res 256: 192–202. Hammerschmidt M, Birchmeier W. (2004). Genes Dev 18: dos Santos NR, de Bruin DRH, Balemans M, Janssen B, Gartner 2225–2230. F, Lopes JM et al. (1997). Hum Mol Genet 6: 1549–1558. Clark J, Rocques PJ, Crew AJ, Gill S, Shipley J, Chan AM-L. Doucas H, Garcea G, Neal CP, Manson MM, Berry DP. (1994). Nat Genet 7: 502–508. (2005). Eur J Cancer 41: 365–379. D’Amico M, Hulit J, Amanatullah DF, Zafonte BT, Albanese Eid JE, Kung AL, Scully R, Livingston DM. (2000). Cell 102: C, Bouzahzah B et al. (2000). J Biol Chem 275: 32649–32657. 839–848.

Oncogene SYT-SSX2 regulates b-catenin location and function D Pretto et al 3669 Elbashir SM, Harborth J, Lendeckel W, Yalcin A, Weber K, Persad S, Troussard AA, McPhee TR, Mulholland DJ, Tuschl T. (2001). Nature 411: 494–498. Dedhar S. (2001). J Cell Biol 153: 1161–1173. Hasegawa T, Yokoyama R, Matsuno Y, Shimoda T, Piedra J, Miravet S, Castano J, Palmer HG, Heisterkamp N, Hirohashi S. (2001). Hum Pathol 32: 257–263. Garcia de Herreros A et al. (2003). Mol Cell Biol 23: 2287–2297. Hoschuetzky H, Aberle H, Kemler R. (1994). J Cell Biol 127: Playford MP, Bicknell D, Bodmer WF, Macaulay VM. (2000). 1375–1380. Proc Natl Acad Sci 97: 12103–12108. Ikura T, Ogryzko VV, Grogoriev M, Groisman R, Wang J, Prieve MG, Waterman ML. (1999). Mol Cell Biol 19: Horikoshi M et al. (2000). Cell 102: 463–473. 4503–4513. Ireton RC, Davis MA, van Hengel J, Mariner DJ, Soulez M, Saurin AJ, Freemont PS, Knight JC. (1999). Barnes K, Thoreson MA et al. (2002). J Cell Biol 159: Oncogene 18: 2739–2746. 465–476. Staal FJT, van Noort M, Strous GJ, Clevers HC. (2002). Ladanyi M. (2001). Oncogene 20: 5755–5762. EMBO rep 3: 63–68. Ladanyi M, Antonescu CR, Leung DH, Woodruff JM, Kawai Thaete C, Brett D, Monaghan P, Whitehouse S, Rennie G, A, Healy JH et al. (2002). Cancer Res 62: 135–140. Rayner E et al. (1999). Hum Mol Genet 8: 585–591. Miyagishi M, Fujii R, Hatta M, Yoshida E, Araya N, Thompson B, Townsley F, Rosin-Arbesfeld R, Musisi H, Nagafuchi A et al. (2000). J Biol Chem 275: 35170–35175. Bienz M. (2002). Nat Cell Biol 4: 367–373. Moon RT, Bowerman B, Boutros M, Perrimon N. (2002). Tissier F, Cavard C, Groussin L, Perlemoine K, Fumey G, Science 296: 1644–1646. Hagnere A-M et al. (2005). Cancer Res 65: 7622–7627. Nagai M, Tanaka S, Tsuda M, Endo S, Kato H, Sonobe H van Es JH, Barker N, Clevers H. (2003). Curr Opin Genet Dev et al. (2001). Proc Natl Acad Sci 98: 3843–3848. 13: 28–33. Patturajan M, Nomoto S, Sommer M, Fomenkov A, Hibi K, Xie Y, Skytting B, Nilsson G, Gasbarri A, Haslam K, Zangen R et al. (2002). Cancer Cell 1: 369–379. Bartolazzi A et al. (2002). Cancer Res 62: 3861–3867.

Oncogene