FBXO31 Suppresses Gastric Cancer EMT by Targeting Snail1 for Proteasomal Degradation Shuiyan Zou1, Cunying Ma1, Fenghua Yang1, Xia Xu1, Jihui Jia2, and Zhifang Liu1

Published OnlineFirst November 8, 2017; DOI: 10.1158/1541-7786.MCR-17-0432

Oncogenes and Tumor Suppressors Molecular Cancer Research FBXO31 Suppresses Gastric Cancer EMT by Targeting Snail1 for Proteasomal Degradation Shuiyan Zou1, Cunying Ma1, Fenghua Yang1, Xia Xu1, Jihui Jia2, and Zhifang Liu1

Abstract

The F-box protein FBXO31, a component of the Skp1/Cul1/ are necessary for the molecular interaction between FBXO31 and F-box (SCF) E3 ubiquitin ligase complex, plays an important Snail1. Mouse modeling experiments reveal that FBXO31 over- regulatory role in neuronal development, stress response, and expression inhibits in vivo colonization of gastric cancer cells. tumorigenesis. Our recent report indicates that FBXO31 functions Furthermore, a highly significant negative correlation between as a tumor suppressor in gastric cancer, and the loss of FBXO31 FBXO31 and Snail1 is validated in human gastric cancer clinical protein is associated with a higher malignant phenotype and specimens. Taken together, these findings identify Snail1 as a new poorer prognosis. However, little is known about the underlying target protein of FBXO31 in gastric cancer and substantiate a novel mechanism. In this study, FBXO31 inhibits gastric cancer pro- regulatory role of FBXO31 on gastric cancer progression and gression by suppressing the epithelial–mesenchymal transition metastasis. (EMT). FBXO31 overexpression decreases, whereas its inhibition increases, the protein level of the EMT transcription factor Snail1 Implication: These findings demonstrate that FBXO31 exerts the (SNAI1), respectively. Further evidence demonstrates that tumor-inhibitory role in gastric cancer by ubiquitin-mediated FBXO31 interacts with Snail1 and mediates the ubiquitin- and degradation of Snail1, which represents a viable strategy of proteasome-dependent degradation of Snail1 in gastric cancer. FBXO31 activators in the prevention and therapy of gastric cancer. The F-box domain of FBXO31 and the phosphorylation of Snail1 Mol Cancer Res; 16(2); 286–95. 2017 AACR.

Introduction cancer metastasis (11, 12). The transcription repression of E- cadherin by Snail1 has been considered the dogma of transcrip- Gastric cancer is one of the most common malignant tumors in tional regulation of EMT (13). the world and ranks third in terms of cancer-related deaths in Recent studies suggest that Snail1 protein is highly unstable and recent years (1). Invasion and metastasis of the tumor contribute extremely sensitive to proteasome inhibitors (14). Several F-box to one of the most frequent cause of death (2, 3). Cancer cells gain proteins, such as FBXW1 (b-Trcp), FBXL14 (Ppa), FBXO11, and motility and invasiveness in a dynamic and reversible process FBXL5, have been shown to promote Snail1 ubiquitination and known as epithelial-to-mesenchymal transition (EMT), which is a degradation (13–19). Zhou and colleagues show that the ubiqui- cellular conversion that takes place not only during embryonic tination and proteasomal degradation of Snail1 mediated by development but also in tumor progression (4–6). It allows b-TrCP1/FBXW1 depends on the Snail1 phosphorylation by epithelial cells to lose their cell polarity and cell–cell adhesion glycogen synthase kinase-3b (GSK-3b; ref. 15). Zheng and col- while gaining migratory and invasive properties to become mes- leagues demonstrate that FBXO11-mediated degradation of enchymal stem cells (7). The molecular hallmark of EMT is the Snail1 is dependent on Ser-11 phosphorylation of Snail by downregulation of epithelial marker E-cadherin due to the tran- protein kinase D1 (PKD1), and FBXO11 blocks Snail1-induced scriptional inhibition (8, 9). Several transcription factors have EMT, tumor initiation, and metastasis in multiple breast cancer been described to bind and repress E-cadherin (CDH1) promoter, models (16). FBXL14 modulates not only Snail1 but also Snail2, such as the Snail family (Snail1 and Slug), the Zeb family (Zeb1 Twist1, and Sip1 and acts as a master regulator of the EMT process and Zeb2), and Twist (10). As core drivers of EMT, Snail family (17). Different from FBXW1 and FBXO11, FBXL14-mediated member Snai1 (Snail) is the zinc-ﬁnger transcriptional factor ubiquitination of Snail1 does not require the phosphorylation that directly binds to the promoter of E-cadherin and inhibits of Snail1 by GSK-3b and PKD1 (14, 17). These ligases above act E-cadherin transcription and plays an essential role in EMT during exclusively in the cytosol. FBXL5 interacts with the C-terminus of Snail1 and ubiquitinates it in the nucleus in colon cancer and breast cancer cells (18). Our recent report shows that F-box 1Department of Biochemistry and Molecular Biology, School of Basic Medicine, protein FBXO31 is downregulated in gastric cancer, and the loss Shandong University, Jinan, P. R. China. 2Department of Microbiology, Key of FBXO31 is associated with higher malignance and poorer Laboratory for Experimental Teratology of Chinese Ministry of Education, prognosis (20). However, the underlying molecular mechanism School of Basic Medicine, Shandong University, Jinan, P. R. China. remains largely elusive. Corresponding Author: Zhifang Liu, School of Basic Medicine, Shandong In the present study, we demonstrate that FBXO31 inhibits University, Wenhua Xi Road 44#, Jinan 250012, P. R. China. Phone: 86-531- gastric cancer progression by suppressing the EMT and metastasis 88382092; Fax: 86-531-88382502; E-mail: [email protected] of gastric cancer. FBXO31 interacts with Snail1 and mediates the doi: 10.1158/1541-7786.MCR-17-0432 ubiquitination and degradation of Snail1. FBXO31 overexpres- 2017 American Association for Cancer Research. sion inhibits the tumorigenesis and colonization of gastric cancer

286 Mol Cancer Res; 16(2) February 2018

Ubiquitination and Degradation of Snail1 by FBXO31

cells in vivo. Furthermore, FBXO31 and Snail1 protein levels are sequences specific for FBXO31 mRNA: 50-CCGGCGGGAGGCAG- negatively correlated in gastric cancer patient specimens. Taken GAGGAGT-30 (forward) and 50-GCGGCGGTAGGTCAGGCA- together, we identify FBXO31 as a novel Snail1 ubiquitin ligase in GTTG-30 (reverse); for Snail1 mRNA: 50-GGAAGCCTAAC TAC- gastric cancer and verify the regulatory role of FBXO31 on gastric AGCGAGCT-30 (forward) and 50-TCCCAGATGAGCATTGGCA-30 cancer progression. (reverse); for b2-M mRNA: 50-GAATTGCTATGTGTCTGGGT-30 (forward) and 50-CATCTTCAAACCTCC ATGATG-30 (reverse). The Materials and Methods first-strand cDNA was synthesized with random primers (N6). The mRNA expression level was determined by qRT-PCR using the Cell lines and clinical samples Bio-Rad CFX96TM Real-Time PCR System (Bio-Rad) with the Human gastric cancer cell lines BGC-823 and SGC-7901 were SYBR Green Kit (TaKaRa Biotechnology Co., Ltd) according to the purchased from Cell Resource Center, Institute of Biochemistry protocol of manufacturer. Calculation of target mRNA levels was and Cell Biology at the Chinese Academy of Sciences, Shanghai, PR based on the CT method and normalization to human b2-M China. The cell bank routinely performed cell line authentication expression. All reactions were run in triplicate. by short tandem repeat profiling, and all of the cell lines were preserved in our laboratory for no more than 6 months after Western blot analysis receipt. The cells were cultured in RPMI 1640 medium (Invitrogen) Total proteins from the cells or tissues were extracted with RIPA supplemented with 10% FBS, 100 mg/mL penicillin, 100 mg/mL lysis buffer supplemented with protease inhibitors. Protein con- streptomycin, and maintained in an incubator with a humidified centration was measured with the BCA Kit. After incubation at atmosphere of 95% air and 5% carbon dioxide at 37C. 100C for 10 minutes in SDS loading buffer, the proteins were Human gastric cancer tissue specimens and surrounding nor- separated by SDS-PAGE and transferred to a PVDF membrane mal tissues were obtained from primary gastric cancer patients (Millipore). After blocking with TBST containing 5% nonfat milk who underwent gastrectomy at the Cancer Hospital of Shandong for 2 hours, the membrane was incubated with specific primary Province in 2012–2014 and the Central Hospital of Taian city in antibodies against FBXO31 (1:1,000; Abcam, ab86137), Snail1 2016. All specimens were stored at 80C. We collected the data (1:1,000; Cell Signaling Technology, C15D3), E-Cadherin of all patients, including age, sex, tumor histology, differentiation (1:1,000; Santa Cruz Biotechnology, sc-7870), c-Myc (1:1,000; status, size (diameter), invasiveness, and regional and distant TransGen Biotech, HT101e), Flag (1:1,000; Sigma-Aldrich, metastases at the time of surgery (pathologic tumor—node– F1804), HA (1:1,000; Cell Signaling Technology, C29F4), fol- metastasis classification). The study protocol was approved by lowed by anti-mouse or anti-rabbit horseradish peroxidase–con- the patients and ethics committee of School of Basic Medicine, jugated secondary antibodies (1:5,000; Bio-Rad) and visualized Shandong University. by the chemiluminescence method (ECL; Millipore). b-Actin (1:3,000; Cell Signaling Technology, 8H10D10) served as a siRNAs and plasmids loading control. The chemically modified Stealth siRNAs targeting FBXO31 and control siRNA were from RiboBio (Guangzhou RiboBio Co., Coimmunoprecipitation assay LTD). The sequences for FBXO31 siRNA1, FBXO31 siRNA2, The gastric cancer cells were seeded in 10-cm culture plates and and control siRNA were 50-GCCACATCCAGATTGTGAA-30,50- 0 0 cotransfected with two different expression vectors. After 48 GCCTGGAGATTGTGATGCT-3 , and 5 -CCUACAU CCCGAUC- m 0 hours, the cells were treated with 10 mol/L MG-132 for 6 hours GAUGAUGUUGA-3 , respectively. In order to increase the inter- and then lysed in IP lysis buffer. About 1.5 mg total proteins were fi ference ef ciency, a mixture of FBXO31 siRNA1 and siRNA2 was incubated with 5 mg of Flag (Sigma-Aldrich, F1804) or c-Myc used to cotransfect into the gastric cancer cells. pCMV-myc- (TransGen Biotech, HT101e) antibody overnight at 4Cona FBXO31 vector and the F-box domain deletion mutation vector vertical roller. The immunoglobulin G (Sigma-Aldrich) served as D (pCMV-myc-FBXO31 F) were kindly provided by Professor an isotype control in the experiments. The mixture was incubated David F. Callen (Dame Roma Mitchell Cancer Research Labora- with 30 mL Protein A/G PLUS Agarose beads (Santa Cruz tories, Department of Medicine, University of Adelaide and Han- Biotechnology) for 4 hours. The immunoprecipitated proteins son Institute). Flag-Snail1-WT (Plasmid #16218), Flag-Snail1- were subjected to SDS-PAGE and were transferred to a PVDF 6SA (Plasmid #16221), Flag-Snail1-8SA (Plasmid #16222), and membrane. The membranes were probed with c-Myc (1:1,000; HA-Ubiquitin (Plasmid #18712) were purchased from Addgene. TransGen Biotech, HT101e) or Flag (1:1,000; Sigma-Aldrich, F1804) antibody and then with horseradish peroxidase– Transient transfection conjugated secondary antibody (1:5,000; Bio-Rad) and visualized X-tremeGENE HP Transfection Reagent (Roche Applied by the chemiluminescence method (ECL; Millipore). Science) was used to transfect plasmids into BGC-823 or For in vivo ubiquitination assay, SGC-7901 cells were cotrans- SGC-7901 cells. Transient transfection of FBXO31 siRNAs was fected with HA-Ubiquitin vector in combination with pCMV or performed with Lipofectamine 2000 (Invitrogen). All transfection FBXO31 or FBXO31DF vector. Immunoprecipitation with 5 mgof procedures were performed according to the protocol of the anti-HA antibody (Cell Signaling Technology, C29F4) was per- manufacturer. formed according to the protocol above. The immunoprecipitated proteins were identified with Snail1 (1:1,000; Cell Signaling RNA extraction, reverse transcription, and qRT-PCR Technology, C15D3) antibody by Western blotting. Total RNA was extracted from gastric cancer cells or tissues using Trizol reagent (Invitrogen). The PCR primers for FBXO31, Scratch wound-healing assay Snail1, and b-microglobulin (b2-M) were synthesized from Invi- The gastric cancer cells were cultured in 6-well plates and were trogen. The PCR primers used in the study were as follows: transfected with FBXO31 expression vector or FBXO31 siRNAs.

www.aacrjournals.org Mol Cancer Res; 16(2) February 2018 287

Zou et al.

Figure 1. FBXO31 siRNA promotes EMT-like alteration in morphology and the metastasis of gastric cancer cells. A, qRT-PCR was used to analyze FBXO31 mRNA level in BGC-823 and SGC-7901 cells transfected with the control siRNA or a mixture of FBXO31 siRNA1 and 2. Data, mean SD. B, Western blot analysis of FBXO31 protein level in the cells transfected with the control siRNA or FBXO31 siRNA1 and 2. The transfection efficiency was confirmed with protein band quantitative analysis. C, EMT-like morphologic changes occur in the cells transfected with FBXO31 siRNAs for 72 hours. D, The scratch wound-healing assays in gastric cancer cells transfected with the control or FBXO31 siRNAs. Representative images are shown. E, Quantitative analysis of the cell migration in the scratch wound-healing assay. The data are expressed as the mean SD from three experiments. F, Transwell assays in the cells transfected with the control or FBXO31 siRNAs. G, Quantitative analysis of the cell number passing through the transwell filter. The data are expressed as the mean SD from three experiments.

Scratch wound was created by using a fine end of 10-mL pipette removed with cotton swabs and cells that had migrated the tips. Images of migrated cells were captured under a phase- membrane filter were fixed with 100% methanol and stained contrast microscopy at different times. with crystal violet and photographed under a microscope. The number of migration cells was counted in three independent Transwell migration assay experiments. The gastric cancer cells BGC-823 or SGC-7901 with different transfection were harvested and resuspended in serum-free RPMI- Animal experiment 1640 medium. Cells (5 104) were seeded into the upper of Athymic BALB/c nude mice (5–6-week-old) were purchased 24-well chambers. RPMI-1640 medium containing 20% FBS was from Vital River Laboratory Animal Technology Company and added to the lower chambers as a chemoattractant. After 24 hours, bred under specific pathogen-free conditions. The FBXO31-BGC- cells on the upper surface of polycarbonate membrane were 823 cells that were transfected stably with FBXO31 expression

288 Mol Cancer Res; 16(2) February 2018 Molecular Cancer Research

Ubiquitination and Degradation of Snail1 by FBXO31

Figure 2. Overexpression of FBXO31 inhibits the metastasis of gastric cancer cells. A, qRT-PCR analysis of FBXO31 mRNA level in BGC-823 and SGC-7901 cells transfected with the control vector (pCMV), FBXO31-overexpressing vector (FBXO31), or the F-box domain deletion mutation vector (DFBXO31). Data are mean SD from experiments with three replicates. B, Western blot analysis of FBXO31 protein level in the cells transfected with the vectors above. C, The scratch wound-healing assays of SGC-7901 transfected with pCMV, FBXO31, and DFBXO31. D, Quantitative analysis of the cell migration in the scratch wound- healing assay. Data are mean SD from experiments with three replicates. E, Transwell assays in BGC- 823 and SGC-7901 cells transfected with pCMV, FBXO31, or DFBXO31. Representative images are shown. F, Quantitative analysis of the migration number in the two cell lines with different transfection. The data are expressed as the mean SD from three experiments.

vector were collected. Cells (6 105) were suspended in 100 mL the cells transfected with FBXO31 siRNAs (Fig. 1C). Because PBS and injected into the tail vein of nude mice. After 6 weeks, the the change in the cell phenotype facilitates cell metastasis (21), mice were sacrificed and lungs were weighed and fixed in form- we thus used scratch wound-healing assay and transwell assay aldehyde solution and stained with hematoxylin and eosin to investigate the effect of FBXO31 on the migration potential (H&E). The percentage of H&E staining was analyzed. The animal of the gastric cancer cells. The results show that FBXO31 experiments were approved by the local ethics committee of siRNAs promote gastric cancer cell migration significantly Shandong University. (Fig. 1D–G).

Statistical analysis Overexpression of FBXO31 inhibits the metastasis of gastric The expression differences between two groups were per- cancer cells, and the F-box domain of FBXO31 is essential for formed with GraphPad Prism v5.0 software (GraphPad Soft- the inhibition ware) using a paired Student t test. Correlation analyses Next, we wondered whether FBXO31 overexpression sup- between FBXO31 and Snail1 expression in gastric cancer tumor presses the EMT program and cell metastasis. We transfected the tissues and matched nontumor tissues were analyzed with the control vector (pCMV-myc), the vectors expressing wild-type Spearman correlation analysis. Statistical analyses were per- FBXO31 (pCMV-myc-FBXO31) and the F-box domain deletion formed with the Statistical Package for the Social Sciences, mutation vector (pCMV-myc-FBXO31DF) into the gastric cancer version 17.0 (SPSS Inc.), and P < 0.05 was considered statis- cells BGC-823 and SGC-7901, respectively. The transfection effi- tically significant. ciency was validated with qRT-PCR and Western blot (Fig. 2A and B). Then scratch wound-healing assay and transwell assay were Results used to detect the migration ability in the transfected gastric cancer Knockdown of FBXO31 induces a mesenchymal phenotype and cells. The results show that the wild-type FBXO31 inhibits the promotes metastasis of gastric cancer cells migration (Fig. 2C–F) of gastric cancer cells distinctly, whereas the When we knocked down FBXO31 with the specificsiRNAsin FBXO31 with F-box domain deletion mutation has no such effect. gastric cancer cells BGC-823 and SGC-7901 (Fig. 1A and B), we Therefore, the F-box domain of FBXO31 is essential for the observed an elongated fibroblast-like phenotype alterations in inhibition.

www.aacrjournals.org Mol Cancer Res; 16(2) February 2018 289

Zou et al.

Figure 3. FBXO31 directs degradation of Snail1. A, SGC-7901 cells were treated with the proteasome inhibitor MG-132 for 6 hours. Western blot was used to detect EMT-TFs and related protein expression. B, Western blot was used to detect Snail1, Slug, Twist1, Twist2, and E-cadherin in gastric cancer cells transfected with empty vector or FBXO31 expression vector (pCMV-myc-FBXO31). C, Western blot was used to detect Snail1 protein level in SGC-7901 cells transfected with empty vector or FBXO31 expression vector for 42 hours and then treated with or without MG-132 for 6 hours. D, qRT-PCR was used to detect the mRNA level of Snail1 in gastric cancer cells transfected with empty vector or FBXO31 expression vector. E, Empty or Snail1 expression vector was cotransfected with pCMV or FBXO31 expression vector into the gastric cancer cells. Western blot analyzes the Snai1 protein abundance and the downstream E-cadherin protein level. F, Western blot was used to detect Snail1 in gastric cancer cells with different transfection. G, Western blot was used to detect Snail1 protein level in gastric cancer cells transfected with Control or FBXO31 siRNAs. H, qRT-PCR was used to detect Snail1 mRNA level in gastric cancer cells transfected with Control or FBXO31 siRNAs.

FBXO31 decreases the protein level of EMT transcription factor fected the FBXO31 expression vector (pCMV-myc-FBXO31) into Snail1 the SGC-7901 cells and used Western blot to detect the protein Because FBXO31 takes part in the formation of SCF E3 ubiqui- level of EMT-TFs and the epithelial marker E-Cadherin. The results tin ligase complex (22) and many EMT transcription factors show that FBXO31 overexpression decreased the protein level of (EMT-TF) are short half-life proteins that are susceptible to be Snail1 distinctly and increased the Snail1 downstream E-cadherin degraded by ubiquitin-dependent proteasome pathway (17), we level slightly (Fig. 3B). We added MG-132 into the transfected wondered whether FBXO31 is involved in the degradation of cells and found that FBXO31-mediated downregulation of Snail1 EMT-TFs in gastric cancer. We ﬁrst used proteasome inhibitor MG- can be completely blocked by the treatment with MG-132 (Fig. 132 to treat gastric cancer cells and found that MG-132 treatment 3C). We then investigated the effect of FBXO31 on Snail1 mRNA increased Snail1 protein level distinctly (Fig. 3A). Then we trans- level and found that FBXO31 has no effect on the mRNA level of

290 Mol Cancer Res; 16(2) February 2018 Molecular Cancer Research

Ubiquitination and Degradation of Snail1 by FBXO31

Figure 4. FBXO31 interacts with Snail1 and directs polyubiquitination of Snail1. A and B, pCMV or Myc- FBXO31 or FBXO31DF and Flag-Snail1 were cotransfected into SGC-7901 cells for 48 hours and then the cells were treated with 10 mmol/L MG-132 for 6 hours. Coimmunoprecipitation was performed with either Flag antibody to probe for FBXO31 by anti-Myc antibody (A) or Myc antibody to probe for Snail1 by anti-Flag antibody (B). C, SGC-7901 cells were cotransfected with HA-Ubiquitin in combination with pCMV or FBXO31 or FBXO31DF. Immunoprecipitation with anti-HA antibody followed by Snail1 immunoblotting was determined.

Snail1 (Fig. 3D). In order to further verify FBXO31 functions as an cancer cells. Either empty vector or Myc-FBXO31 was cotransfected E3 ubiquitin ligase of Snail1 in gastric cancer, we investigated the with Flag-Snail1 expression vectors into SGC-7901 cells, and regulatory role of FBXO31 on the exogenous Snail1 protein level. coimmunoprecipitation (co-IP) assay was performed to detect the Myc-FBXO31 and Flag-Snail1 expression vectors were cotrans- interaction between FBXO31 and Snail1. The result showed that fected into SGC-7901 cells. Snail1 and E-cadherin protein level ectopically expressed myc-FBXO31 protein was coimmunopreci- were investigated. The results show that exogenous Snail1 protein pitated by Flag-Snail1 (Fig. 4A). A similar result was obtained in a level was downregulated by coexpression of FBXO31, and Snail1- reciprocal co-IP experiment using myc antibody (Fig. 4B). We mediated downregulation of E-cadherin can be abrogated by further investigated whether the interaction depends on the F-box FBXO31 (Fig. 3E). Then, we investigated whether F-box domain domain of FBXO31. The F-box domain deletion mutation vector of FBXO31 is essential for the degradation of Snail1. As shown myc-FBXO31DF and Flag-Snail1 were cotransfected into SGC-7901 in Fig. 3F, DFBXO31 has no effect on both endogenous and cells. The co-IP assay results showed that the myc-FBXO31DFdoes exogenous Snail1 protein level, which indicates that the F-box not coimmunoprecipitate with Flag-Snail1 (Fig. 4A and B). These domain of FBXO31 is essential for the degradation of Snail1. data suggest that FBXO31 interacts with Snail1 in gastric cancer We further investigated the effect of FBXO31 siRNAs on the cells, and the interaction depends on the F-box domain of FBXO31. protein level of Snail1 and found that Snail1 protein level is We then performed ubiquitination assay to investigate whether increased in the cells transfected with FBXO31 siRNAs, whereas FBXO31 is involved in the ubiquitination of Snail1 in vivo. Either the mRNA level of Snail1 has no signiﬁcant change in the empty vector or myc-FBXO31 or myc-FBXO31DF was cotrans- transfected cells (Fig. 3G and H). fected with HA-tagged ubiquitin into SGC-7901 cells. Co-IP assay was performed with HA antibody to immunoprecipitate the FBXO31 interacts with Snail1 and mediates ubiquitin- Snail1. The results showed that Snail1 ubiquitination was remark- dependent proteasome degradation of Snail1 protein ably increased in the cells transfected with FBXO31 expression Because substrate binding plays a critical role in E3 ligase– vector, whereas FBXO31DF does not increase the ubiquitination mediated ubiquitination and subsequent proteasome degradation, level of Snail1 (Fig. 4C). These results suggest that FBXO31 we examined whether FBXO31 can interact with Snail1 in gastric facilitates Snail1 degradation through ubiquitination-dependent

www.aacrjournals.org Mol Cancer Res; 16(2) February 2018 291

Zou et al.

Figure 5. Phosphorylation of Ser is essential for Snail1 to be recognized by FBXO31. A, SGC-7901 cells were treated with or without 10 mmol/L MG132 or 40 mmol/L LiCl, or both, for 5 hours and detected by Western blot with antibody against Snail1. The result showed that alone MG132 or LiCl treatment increased the protein level of Snail1, whereas both MG132 and LiCl treatment elevated the concentration of Snail1 markedly. B, GSK-3b phosphorylation motif of Snail1 and the mutation site of the plasmid were shown. C and D, Flag- Snail1 proteins (wild-type and variants) and Myc- FBXO31 protein were coexpressed in SGC-7901 cells. Proteins were immunoprecipitated with either anti-Flag antibody to probe for FBXO31 by anti-Myc antibody (C) or Myc antibody to probe for Snail1 by anti-Flag antibody (D).

pathway, and the F-box domain of FBXO31 is essential for the constructed the cell line that stably expressed FBXO31 in BGC- ubiquitination and sequential degradation. 823 (called BGC-FBXO31) and injected the cells and the control cells into nude mice through the tail vein. The mice were sacrificed Phosphorylation of Ser is essential for Snail1 to be recognized after 6 weeks, and the tumor colonization in the lungs was by FBXO31 examined. We observed that both the mice injected with the Because it has been reported that F-box proteins generally control cells and the BGC-FBXO31 cells formed metastatic recognize and interact with the phosphorylated protein substrate nodules at the surface of the lungs, but the lungs from mice (23), we wondered whether Snail1 phosphorylation is required for injected with BGC-FBXO31 cells were significantly smaller and initiating FBXO31-mediated ubiquitination and degradation. lighter than those from the control mice (Fig. 6A and B). Also, a Therefore, we treated SGC-7901 cells with the inhibitors of several bigger size of metastases were found in the lungs of mice injected phosphorylation pathways and found that treatment with GSK-3b with control cells, compared with mice injected with BGC- inhibitor LiCl led to a marked increase of the Snail1 protein level FBXO31 cells (Fig. 6C and D). Tumors derived from BGC-FBXO31 (Fig. 5A). Furthermore, treatment with proteasome inhibitor MG- cells expressed much higher levels of FBXO31 mRNA compared 132 and GSK-3b inhibitor LiCl increased Snail1 protein level with those from the control cells (Fig. 6E). synergistically (Fig. 5A). It is known that Snail1 contains two identical GSK-3b phosphorylation motifs. GSK-3b binds to and FBXO31 is negatively correlated with Snail1 in human gastric phosphorylates the two motifs to double-regulate the function of cancer tissues Snail1. The phosphorylation of the first motif regulates ubiquitina- Our recent report demonstrated that FBXO31 is downregulated tion, and the phosphorylation of the second motif regulates its in gastric cancer tissues, and the low expression of FBXO31 is subcellular localization (15). We obtained two Snail1 mutant associated with high malignancy and poor prognosis in gastric plasmid (Flag-Snail1-6SA and Flag-Snail1-8SA; Fig. 5B) and exam- cancer patients (20). We then explored whether FBXO31 was ined whether the phosphorylation of the motifs was required for associated with Snail1 in the gastric cancer tissues. We analyzed Snail1 to be recognized by FBXO31. Flag-Snail1 proteins (wild-type the expression of FBXO31 and Snail1 in the paired 77 gastric cancer and variants) and Myc-FBXO31 protein were coexpressed in SGC- and corresponding noncancerous tissues by Western blot. The 7901 cells. Co-IP assay results showed that the mutation of the amount of FBXO31 and Snail1 protein was further evaluated by phosphorylation site of Snail1 in the GSK-3b phosphorylation densitometry. The results showed that the gastric cancer tissues had motifs abrogates the interaction between FBXO31 and Snail1 (Fig. a lower expression of FBXO31 but a higher level of Snail1 com- 5C and D). This suggested that the phosphorylation of the GSK-3b pared with adjacent normal tissues (Fig. 7A and B). Further analysis phosphorylation motif is critical for FBXO31 to recognize Snail1. indicated that low-grade tumor (G1 and G2) had a higher FBXO31 expression compared with high-grade tumor (G3). In contrast, FBXO31 suppressed the tumorigenesis and colonization of low-grade tumor had a lower Snail1 expression compared with gastric cancer cells in vivo high-grade tumor (Fig.7C and D). The statistical analysis indicated We then investigated whether FBXO31 overexpression sup- that FBXO31 was negatively correlated with Snail1 expression (Fig. pressed the metastasis of gastric cancer cells in vivo. We thus 7E). Some representative results were shown in Fig. 7F. These data

292 Mol Cancer Res; 16(2) February 2018 Molecular Cancer Research

Ubiquitination and Degradation of Snail1 by FBXO31

Figure 6. FBXO31 overexpression inhibits in vivo colonization of BGC-823 cells. A, FBXO31-overexpressing stable cell lines (FBXO31-BGC-823) and its control cell lines (GFP-BGC-823) were injected directly into the tail vein of nude mice. Six weeks later, the nude mice were dissected and the lungs of the mice were taken photograph. B, The lung weight of the control and FBXO31-BGC-823 mice groups was weighed. Data, mean SD. C, H&E staining of lung tissues was used to detect the metastasis nodules. D, The quantity of percentage of H&E staining. E, The mRNA level of FBXO31 with qRT-PCR in lung tissue of nude mice injected with FBXO31-BGC-823 cells and GFP-BGC-823 cells, respectively. Data, mean SD. suggested that Snail1 and FBXO31 protein levels are negatively oncogene and promotes malignancy in lung cancer and esoph- correlated in human gastric cancer specimens. ageal squamous cell carcinoma (32, 33). Therefore, FBXO31 acts in a context-dependent manner in cancer, as has been reported for other F-box proteins (28, 34). Our previous report shows that Discussion FBXO31 is downregulated in gastric cancer, and the loss of FBXO31, encoded by the fbxo31 gene located in chromosome FBXO31 is associated with higher malignance and poorer prog- 16q24 (22), belongs to F-box protein family with no recognizable nosis (20). However, how FBXO31 is involved in gastric cancer substrate binding domains and is the substrate recognition com- progression remains poorly understood. In the present study, we ponent of SCF E3 ubiquitin ligase complex which mediated the provide evidence that FBXO31 plays an important role in EMT ubiquitination and degradation of the targeting substrates (24, regulation of gastric cancer cells by targeting EMT transcription factor 25). It has been reported that FBXO31 acts as a putative tumor Snail1, thereby contributing to inhibit gastric cancer progression. suppressor in breast cancer, melanoma, hepatocellular carcino- When we transfected FBXO31 siRNA into the gastric cancer ma, and gastric cancer and is frequently inactivated due to LOH or cells, we noticed that FBXO31 siRNA induced an elongated noncoding miRNAs-mediated posttranscriptional regulation (20, fibroblast-like phenotype alterations in the transfected gastric 22). FBXO31 inhibits cell growth and cell-cycle progression by cancer cells. Because EMT plays a critical role in tumor metastasis, targeting the cell-cycle protein CyclinD1 for ubiquitination and we further investigated the role of FBXO31 on gastric cancer degradation (20, 22, 26, 27). In addition to CyclinD1, FBXO31 metastasis and found that FBXO31 knockdown promoted, where- exerts the tumor-suppressor role by mediating the ubiquitination as its overexpression inhibited, the metastasis of gastric cancer and degradation of FOXM1 during the G2–M transition (28) and cells. Furthermore, the nude mice experiment showed that targeting DNA replication factor Cdt1 for proteolysis in the G2 FBXO31 overexpression suppressed the metastasis of gastric can- phase of cell cycle to prevent re-replication (29). Besides, FBXO31 cer cells in vivo. These results suggest that FBXO31 functions as a also plays an important role in DNA damage repair by degrading tumor suppressor to inhibit EMT and metastasis in the progres- MDM2 (30), a negative regulator of p53, and MKK6 (31), an sion of gastric cancer. activator of the p38 MAPK pathway. However, FBXO31 performs It has been reported that EMT is regulated by a cohort of conflicting proapoptotic and antiapoptotic roles by regulating transcription factors, including members of the Snail1 and Twist MDM2 and MKK6 after DNA damage (30, 31). Similarly, in families (10). Thus, we investigated the regulatory role of FBXO31 contrast to the tumor-suppressor role, FBXO31 is regarded as an on these EMT transcription factors and verified that FBXO31

www.aacrjournals.org Mol Cancer Res; 16(2) February 2018 293

Zou et al.

Figure 7. FBXO31 is negatively correlated with Snail1 in human gastric cancer tissues. A and B, Statistical analysis of FBXO31 and Snail1 protein expression in 77 pairs gastric cancer and corresponding noncancerous normal tissues with paired t tests. C and D, Statistical analysis of FBXO31 and Snail1 protein expression in gastric cancer tissues with low tumor grade (G1, G2) and high tumor grade (G3). E, Regression analysis of correlation of FBXO31 and Snail1 expression in gastric cancer tissues. F, Western blot analysis of the expression of FBXO31 and Snail1 in gastric cancer and corresponding noncancerous normal tissues. Eight pairs of representative tissues were shown.

interacted with Snail1 and mediated the ubiquitination and degra- the cytoplasm (15). Different from the b-Trcp/FBXW1, F-box E3 dation of Snail1 protein in gastric cancer cells. The F-box domain of ubiquitin ligase FBXl14 interacts with Snail1 and promotes its FBXO31 is essential for the interaction and degradation. Moreover, ubiquitylation and proteasome degradation independently of we provided evidence that Snail1 and FBXO31 protein levels were phosphorylation by GSK-3b (14). FBXO11-mediated Snail1 pro- negatively correlated in human gastric cancer specimens. Therefore, tein ubiquitylation and degradation depends on Ser-11 phos- our study identiﬁed Snail1 as a novel ubiquitination and degrada- phorylation on the SNAG domain of Snail1 by protein kinase D1 tion substrate of FBXO31, and FBXO31-mediated degradation of (PKD1; ref. 16). So, different F-box proteins ubiquitinate and Snail1 inhibited tumor migration in gastric cancer cells. degrade Snail1 protein with different mechanisms. It has been It has been reported that Snail1 protein stability can be regu- reported that FBXO31 recognizes both the phosphorylated target b lated by other several E3 ubiquitin ligases, including SCF -TrCP1, substrates, such as CyclinD1 (27), FOXM1 (28), MDM2 (30), and b SCFPpa/FBXL14, and SCFFBXO11 (14–17). SCF -TrCP1-mediated ubi- MKK6 (31), and unphosphorylated target substrate, such as Cdt1 quitination and degradation of Snail1 depend on the phosphor- (29). Therefore, we investigated whether FBXO31-mediated ubi- ylation of this protein by GSK-3. GSK-3b binds to and phosphor- quitination and degradation of Snail1 need the phosphorylation ylates Snail1 at two consensus motifs to dually regulate the of this protein. We used GSK-3b inhibitor LiCl to treat the cells function of this protein. b-Trcp–mediated ubiquitination and and found that the treatment led to a marked increase of the degradation need the phosphorylation of the ﬁrst motif, whereas Snail1 protein level. Furthermore, the Snail1 variant that abol- phosphorylation of the second motif causes Snail1 to localize in ished GSK-3b-phosphorylation deprived the interaction with

294 Mol Cancer Res; 16(2) February 2018 Molecular Cancer Research

Ubiquitination and Degradation of Snail1 by FBXO31

FBXO31, indicating the interaction between Snail1 and FBXO31 Acquisition of data (provided animals, acquired and managed patients, depended on the phosphorylation of Snail1 mediated by GSK-3b. provided facilities, etc.): S. Zou, F. Yang In summary, we identified that FBXO31 is a novel E3 ubiquitin Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): S. Zou, C. Ma, X. Xu, Z. Liu ligase of Snail1, and its low expression contributes to an aberrant Writing, review, and/or revision of the manuscript: S. Zou, J. Jia, Z. Liu accumulation of Snail1 in gastric cancer. Furthermore, F-box Administrative, technical, or material support (i.e., reporting or organizing domain of FBXO31 and the phosphorylation of Snail1 are essen- data, constructing databases): C. Ma, F. Yang, J. Jia, Z. Liu tial for the interaction and degradation. These findings reveal the Study supervision: J. Jia, Z. Liu important function of FBXO31 on the regulation of Snail1 protein Acknowledgments level in gastric cancer, which might represent a viable strategy for This study was supported by the National Natural Science Foundation of effective treatment of gastric cancer. China (81671981) and the Natural Science Foundation of Shandong Province (ZR2014HM024 and ZR2015HM027). Disclosure of Potential Conflicts of Interest No potential conflicts of interest were disclosed. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Authors' Contributions Conception and design: S. Zou, Z. Liu Received August 28, 2017; revised September 23, 2017; accepted October 27, Development of methodology: S. Zou, F. Yang, X. Xu, Z. Liu 2017; published OnlineFirst November 8, 2017.

References 1. Torre LA, Bray F, Siegel RL, Ferlay J, Lortet-Tieulent J, Jemal A. Global cancer 20. Zhang X, Kong Y, Xu X, Xing H, Zhang Y, Han F, et al. F-box protein FBXO31 statistics, 2012. CA Cancer J Clin 2015;65:87–108. is down-regulated in gastric cancer and negatively regulated by miR-17 and 2. Eccles SA, Welch DR. Metastasis: recent discoveries and novel treatment miR-20a. Oncotarget 2014;5:6178–90. strategies. Lancet 2007;369:1742–57. 21. Friedl P, Alexander S. Cancer invasion and the microenvironment: plas- 3. Wan L, Pantel K, Kang Y. Tumor metastasis: moving new biological insights ticity and reciprocity. Cell 2011;147:992–1009. into the clinic. Nat Med 2013;19:1450–64. 22. Kumar R, Neilsen PM, Crawford J, McKirdy R, Lee J, Powell JA, et al. 4. Kalluri R, Weinberg RA. The basics of epithelial-mesenchymal transition. FBXO31 is the chromosome 16q24.3 senescence gene, a candidate breast J Clin Invest 2009;119:1420–8. tumor suppressor, and a component of an SCF complex. Cancer Res 5. Nieto MA. The snail superfamily of zinc-ﬁnger transcription factors. Nat 2005;65:11304–13. Rev Mol Cell Biol 2002;3:155–66. 23. Cardozo T, Pagano M. The SCF ubiquitin ligase: insights into a molecular 6. Thiery JP. Epithelial-mesenchymal transitions in tumour progression. Nat machine. Nat Rev Mol Cell Biol 2004;5:739–51. Rev Cancer 2002;2:442–54. 24. Zhou W, Wei W, Sun Y. Genetically engineered mouse models for func- 7. Brabletz T. To differentiate or not–routes towards metastasis. Nat Rev tional studies of SKP1-CUL1-F-box-protein (SCF) E3 ubiquitin ligases. Cell Cancer 2012;12:425–36. Res 2013;23:599–619. 8. Batlle E, Sancho E, Francí C, Domínguez D, Monfar M, Baulida J, et al. The 25. Zheng N, Schulman BA, Song L, Miller JJ, Jeffrey PD, Wang P, et al. transcription factor snail is a repressor of E-cadherin gene expression in Structure of the Cul1-Rbx1-Skp1-F boxSkp2 SCF ubiquitin ligase epithelial tumour cells. Nat Cell Biol 2000;2:84–9. complex. Nature 2002;416:703–9. 9. Cano A, Perez-Moreno MA, Rodrigo I, Locascio A, Blanco MJ, del Barrio MG, 26. Huang HL, Zheng WL, Zhao R, Zhang B, Ma WL. FBXO31 is down- et al. The transcription factor snail controls epithelial-mesenchymal transi- regulated and may function as a tumor suppressor in hepatocellular tions by repressing E-cadherin expression. Nat Cell Biol 2000;2:76–83. carcinoma. Oncol Rep 2010;24:715–20. 10. Peinado H, Olmeda D, Cano A. Snail, Zeb and bHLH factors in tumour 27. Santra MK, Wajapeyee N, Green MR. F-box protein FBXO31 mediates progression: an alliance against the epithelial phenotype? Nat Rev Cancer cyclin D1 degradation to induce G1 arrest after DNA damage. Nature 2007;7:415–28. 2009;459:722–5. 11. Zeisberg M, Neilson EG. Biomarkers for epithelial-mesenchymal transi- 28. Jeffery JM, Kalimutho M, Johansson P, Cardenas DG, Kumar R, Khanna KK. tions. J Clin Invest 2009;119:1429–37. FBXO31 protects against genomic instability by capping FOXM1 levels at 12. Thiery JP, Acloque H, Huang RY, Nieto MA. Epithelial-mesenchymal the G2/M transition. Oncogene 2017;36:1012–22. transitions in development and disease. Cell 2009;139:871–90 29. Johansson P, Jeffery J, Al-Ejeh F, Schulz RB, Callen DF, Kumar R, et al. SCF- 13. Jin Y, Shenoy AK, Doernberg S, Chen H, Luo H, Shen H, et al. FBXO11 FBXO31 E3 ligase targets DNA replication factor Cdt1 for proteolysis in the promotes ubiquitination of the Snail family of transcription factors in cancer G2 phase of cell cycle to prevent re-replication. J Biol Chem 2014;289: progression and epidermal development. Cancer Lett 2015;362:70–82. 18514–25. 14. Vinas-Castells~ R, Beltran M, Valls G, Gomez I, García JM, Montserrat-Sentís 30. Malonia SK, Dutta P, Santra MK, Green MR. F-box protein FBXO31 directs B, et al. The hypoxia-controlled FBXL14 ubiquitin ligase targets SNAIL1 for degradation of MDM2 to facilitate p53-mediated growth arrest following proteasome degradation. J Biol Chem 2010;285:3794–805. genotoxic stress. Proc Natl Acad Sci 2015;112:8632–7. 15. Zhou BP, Deng J, Xia W, Xu J, Li YM, Gunduz M, et al. Dual regulation of 31. Liu J, Han L, Li B, Yang J, Huen MS, Pan X, et al. F-box only protein 31 Snail by GSK-3beta-mediated phosphorylation in control of epithelial (FBXO31) negatively regulates p38 mitogen-activated protein kinase mesenchymal transition. Nat Cell Biol 2004;6:931–40. (MAPK) signaling by mediating lysine 48-linked ubiquitination and 16. Zheng H, Shen M, Zha YL, Li W, Wei Y, Blanco MA, et al. PKD1 phosphor- degradation of mitogen-activated protein kinase kinase 6 (MKK6). J Biol ylation-dependent degradation of SNAIL by SCF-FBXO11 regulates epithe- Chem 2014;289:21508–18. lial-mesenchymal transition and metastasis. Cancer Cell 2014;26:358–73. 32. Huang HL, Jiang Y, Wang YH, Chen T, He HJ, Liu T, et al. FBXO31 promotes 17. Lander R, Nordin K, LaBonne C. The F-box protein Ppa is a common regulator cell proliferation, metastasis and invasion in lung cancer. Am J Cancer Res of core EMT factors Twist, Snail, Slug, and Sip1. J Cell Biol 2011;194:17–25. 2015;5:1814–22. 18. Vinas-Castells~ R, Frías A, Robles-Lanuza E, Zhang K, Longmore GD, García 33. Kogo R, Mimori K, Tanaka F, Komune S, Mori M. FBXO31 determines De Herreros A, et al. Nuclear ubiquitination by FBXL5 modulates Snail1 poor prognosis in esophageal squamous cell carcinoma. Int J Oncol DNA binding and stability. Nucleic Acids Res 2014;42:1079–94. 2011;39:155–9. 19. Wu W, Ding H, Cao J, Zhang W. FBXL5 inhibits metastasis of gastric cancer 34. Skaar JR, Pagan JK, Pagano M. Mechanisms and function of substrate through suppressing Snail1. Cell Physiol Biochem 2015;35:1764–72. recruitment by F-box proteins. Nat Rev Mol Cell Biol 2013;14:369–81.

www.aacrjournals.org Mol Cancer Res; 16(2) February 2018 295

FBXO31 Suppresses Gastric Cancer EMT by Targeting Snail1 for Proteasomal Degradation

Shuiyan Zou, Cunying Ma, Fenghua Yang, et al.

Mol Cancer Res 2018;16:286-295. Published OnlineFirst November 8, 2017.

Updated version Access the most recent version of this article at: doi:10.1158/1541-7786.MCR-17-0432

Cited articles This article cites 34 articles, 6 of which you can access for free at: http://mcr.aacrjournals.org/content/16/2/286.full#ref-list-1

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://mcr.aacrjournals.org/content/16/2/286. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.