The tumor suppressor ING3 is degraded by SCFSkp2-mediated ...

Oncogene (2010) 29, 1498–1508

& 2010 Macmillan Publishers Limited All rights reserved 0950-9232/10 $32.00 www.nature.com/onc

ORIGINAL ARTICLE

The tumor suppressor ING3 is degraded by SCFSkp2-mediated ubiquitin–proteasome system G Chen1,3, Y Wang1,3, M Garate1, J Zhou2 and G Li1 1

Department of Dermatology and Skin Science, Jack Bell Research Centre, Vancouver Coastal Health Research Institute, University of British Columbia, Vancouver, British Columbia, Canada and 2Department of Molecular Cell Biology and Toxicology, School of Public Health, Nanjing Medical University, Nanjing, PR China

The inhibitor of growth family member 3 (ING3) has been shown to modulate transcription, cell cycle control and apoptosis. We previously reported that nuclear ING3 expression was remarkably reduced in melanomas, which correlated with a poorer patient survival, suggesting that decreased ING3 expression may be associated with melanoma progression. However, the mechanism of diminished ING3 expression in melanoma is not clear. Here we show that ING3 level was decreased in metastatic melanoma cells because of a rapid degradation. Furthermore, we showed that ING3 undergoes degradation through the ubiquitin– proteasome pathway. ING3 physically interacts with subunits of E3 ligase Skp1-Cullin-F-box protein complex (SCF complex). Knockdown of F-box protein S-phase kinaseassociated protein 2 (Skp2) reduces the ubiquitination of ING3 and significantly stabilizes ING3 in melanoma cells. In addition, lysine 96 residue is essential for ING3 ubiquitination as its mutation to arginine dramatically abrogated ING3 degradation. Disruption of ING3 degradation stimulated ING3-induced G1 cell-cycle arrest and enhanced ultravioletinduced apoptosis. Taken together, our data show that ING3 is degraded by the ubiquitin–proteasome pathway through the SCF Skp2 complex and interruption of ING3 degradation enhances the tumor-suppressive function of ING3, which provides a potential cancer therapeutic approach by interfering ING3 degradation. Oncogene (2010) 29, 1498–1508; doi:10.1038/onc.2009.424; published online 23 November 2009 Keywords: ING3; protein degradation; Skp2; cell cycle; apoptosis

Introduction Fundamental cellular activities are largely maintained by coordinated protein synthesis and degradation. Breakdown of this balance usually results in aberrant Correspondence: Dr G Li, Department of Dermatology and Skin Science, Jack Bell Research Centre, Vancouver Coastal Health Research Institute, University of British Columbia, Vancouver, British Columbia, Canada V6H 3Z6. E-mail: [email protected] 3 These authors contributed equally to this work. Received 2 June 2009; revised 7 September 2009; accepted 20 October 2009; published online 23 November 2009

protein levels and their associated cellular functions, which eventually contribute to the development of many diseases including cancer and central nervous system diseases (Mani and Gelmann, 2005; Zetter and Mangold, 2005; Shah and Di Napoli, 2007). In particular, the ubiquitin-dependent proteasome pathway of protein degradation controls the turnover of many tumor suppressors and oncoproteins, thus having a crucial role in cell transformation and cancer progression. Therefore, selective inhibition of the activities of disease-specific components of the ubiquitin–proteasome pathway has arisen as a promising approach for innovative anticancer therapies (Burger and Seth, 2004; Nalepa et al., 2006; Goldberg, 2007; Newton and Vucic, 2007). Ubiquitination has an essential role in the ubiquitin– proteasome protein degradation pathway and involves three major groups of enzymes: E1-activating enzyme, E2-conjugating enzyme and E3 ubiquitin ligase. On the activation by E1 enzyme and the conjugation to E2 enzyme, ubiquitin moieties is transferred to the E3 ligase-recognized substrates, leading to polyubiquitination of the substrate protein, which will be destined for destruction by the 26S proteasome (Nandi et al., 2006; von Mikecz, 2006). The selectivity of the ubiquitin– proteasome pathway for a particular substrate protein relies on E3 ligases. Hundreds of E3 ligases have been identified in the past few decades. The SCF (Skp1-Cul1F-box) E3 ligase complex is the best-characterized RING-domain type ligase, which contains four major components: Skp1, Cul1, Roc1 and F-box protein (Schwechheimer and Calderon Villalobos, 2004; Ang and Wade Harper, 2005). Cul1 acts as a scaffold protein in this multi-subunit complex with its carboxyl-terminus interacting with Roc1 to recruit ubiquitin-conjugated E2 enzyme and its amino-terminus associating with Skp1 that binds to the F-box protein. The specificity of SCF E3 ligase complex mainly depends on the F-box protein, which directly interacts with the phosphor-degron of substrates. Approximately 70 F-box proteins have been identified so far, although only a few of them have known substrates. ING3, a member of ING tumor suppressor family proteins, is an important subunit of human NuA4 histone acetyltransferase complex (Doyon et al., 2004). It is ubiquitously expressed in normal human tissues and regulates gene transcription, cell cycle control and apoptosis (Nagashima et al., 2003). Distorted ING3

ING3 degradation G Chen et al

1499

expression has been found in human head and neck squamous cell carcinomas (HNSCCs) and several lymphoma-derived cell lines (Gunduz et al., 2002, 2008; Fadlelmola et al., 2008). In melanoma cells, we have shown that ING3 promotes ultraviolet (UV)induced apoptosis through the Fas/caspase-8-dependent pathway (Wang and Li, 2006). We have also shown in a tissue microarray study that nuclear ING3 expression is remarkably reduced in malignant melanomas compared with dysplastic nevi, and the reduced nuclear ING3 level significantly correlates with a poorer diseasespecific 5-year survival of patients with primary melanoma (Wang et al., 2007). However, the mechanism resulting in aberrant ING3 expression in melanomas is not clear. Although the data from tissue mircroarray study showed that the reduced nuclear ING3 level correlates with increased ING3 level in the cytoplasm, 48% (24/50) of melanoma metastases showed a negative-to-moderate overall ING3 expression compared with 29% (17/58) of either dysplastic nevi or primary melanomas (33/114) (Po0.05, Supplementary Figure S1), implying that mechanisms other than the nuclearto-cytoplasm shift might also contribute to the deregulation of ING3 expression in advanced melanoma. In this study, we examined the protein and mRNA levels of ING3 in normal human melanocytes and melanoma cell lines. By characterizing the ubiquitin– proteasome degradation pathway of ING3 protein, we provided the evidence that rapid ING3 protein degradation may be involved in the development of melanomas.

Results ING3 is rapidly degraded in melanoma cells To explore the mechanism of reduced ING3 expression in advanced melanomas, we compared ING3 protein level in nine metastatic melanoma cell lines with that in normal human melanocytes (Figure 1a). Results showed that five melanoma cell lines (MMLH, MMRU, Sk-mel-3, KZ-13 and Sk-mel-93) had significantly lower ING3 expression in comparison with normal human melanocytes, whereas the other four melanoma cell lines (Sk-mel-110, MMAN, MEWO and Sk-mel-5) showed ING3 level similar to melanocytes. To determine the mechanism resulting in low ING3 level in MMRU, MMLH, Sk-mel-3, KZ-13 and Sk-mel-93 cells, we first checked the mRNA level of ING3 by reverse transcriptase–PCR. No dramatic difference was observed for the ING3 mRNA level between normal human melanocytes and four melanoma cell lines (MMRU, MMLH, Skmel-3 and Sk-mel-93), but KZ-13 cells had a significantly lower ING3 mRNA level compared with melanocytes (Figure 1b), implying that compromised mRNA expression is not the primary reason leading to the downregulation of ING3 in melanoma cells. Next, we examined the ING3 protein turnover rate in these cell lines by treating them with the protein synthesis inhibitor cycloheximide to inhibit de novo protein synthesis. The half-life of ING3 protein in either normal

human melanocytes or KZ-13 cells was longer than 16 h, whereas it was shortened to about 8.3–9.6 h in MEWO and MMAN cells, and to 5.0–6.6 h in MMRU, MMLH, Sk-mel-3 and Sk-mel-93 cells (Figures 1c and d). Therefore, the rapid ING3 degradation may have an important role in loss of ING3 expression in melanomas. ING3 is degraded through the ubiquitin–proteasome pathway As majority of the cellular proteins are degraded by the ubiquitin–proteasome pathway, we examined whether ING3 would be degraded by the proteasome complex. In MMRU cells, the expression levels of both p53, the well-known target of the ubiquitin–proteasome pathway, and ING3 were significantly increased upon the treatment with the proteasome inhibitor MG132 or lactacystin (Figure 2a). The level of ING3 was also significantly elevated in other melanoma cell lines including MMLH, Sk-mel-3 and Sk-mel-93 (Figure 2b), suggesting that the degradation of ING3 in these cell lines may be through the same pathway. p53, which is degraded by ubiquitin–proteasome, was also dramatically stabilized in the presence of proteasome inhibitor MG132 or lactacystin (Figure 2c). We also observed the accumulation of both endogenous and ectopic ING3 after MG132 treatment in HEK293T cells (Supplementary Figure S2). To confirm that the increased expression of ING3 is a result of improved protein stability following the inhibition of proteasome activity, we examined the half-life of ING3 by treating MMRU cells with the protein synthesis inhibitor cycloheximide. The half-life of ING3 in MMRU cells was increased from 5.1 h without any proteasome inhibitor treatment to 13.0 and 9.6 h in the presence of MG132 and lactacystin, respectively (Figures 2c and d). As we used ectopic FLAG-tagged ING3 in a number of experiments to determine the degradation pathway for ING3, we also examined the half-life of ectopic FLAGtagged ING3 (FLAG-ING3) in MMRU cells. The turnover rate of FLAG-ING3 was estimated to be 4.2 h, which was increased to 15.1 h in the presence of MG132 (Figures 2e and f). The fact that ectopic FLAGING3 (4.2 h) has a slightly shorter half-life than the endogenous ING3 (5.2 h) may be explained by the introduction of lysine residues inside FLAG motif when we constructed the FLAG-ING3 fusion protein. To distinguish whether ING3 is degraded by 20S or 26S proteasome, nuclear extracts from MMRU cells were precipitated with ammonium sulfate. Detection of 26S and 20S proteasome markers PSMC4 (subunit of 19S) and a4 (subunit of 20S) assured the enrichment of the crude preparations of 20S and 26S proteasomes (Figure 3a). However, ING3 was predominantly associated with the 26S proteasome-rich fraction and was significantly increased in the presence of MG132 (Figure 3a). To confirm that 26S proteasome is able to degrade ING3, we performed an in vitro degradation assay using isolated 26S proteasome. We found that ING3 was rapidly degraded by the 26S proteasome, similar to the degradation of p53 (Figure 3b). Oncogene


1500

Figure 1 Rapid degradation of ING3 in melanoma cells. (a) Whole cell extracts were obtained from normal human melanocytes and melanoma cell lines for western blot analysis. The dashed line indicated the average level of ING3 in melanoma cells. (b) Reverse transcriptase–PCR analysis of ING3 mRNA level in melanocytes and melanoma cell lines. (c) Melanocytes and melanoma cell lines were treated with cycloheximide (CHX, 20 mg/ml) to block de novo protein synthesis. Cells were harvested at indicated time and whole cell extracts were subjected to western blot analysis. (d) Half-life of ING3 was determined by densitometric analysis of ING3 bands in panel c from three independent experiments using the formula t1/2 ¼ Ln2/S, where S represents the slope from each linear regression. *Po0.05 and **Po0.01.

In addition, the degradation of both p53 and ING3 was abrogated in the presence of proteasome inhibitor MG132. As ubiquitination has an essential role in the protein degradation by the 26 proteasome, we examined whether ING3 is polyubiquitinated using a two-step immunoprecipitation (IP) assay to purify the ubiquitinated ING3. MMRU cells were co-transfected with HA-tagged ubiquitin (HA-Ub) and FLAG-ING3 plasmids and processed for IP assay. We first pulled down FLAG-ING3 containing immunocomplex using a mouse anti-FLAG antibody, denatured the immunoprecipitates and then pulled down ubiquitin-bound ING3 with mouse anti-HA antibody. The two-step purification assay gave a strong signal of ubiquitinated FLAG-ING3 (Figure 3c), confirming that ING3 itself is ubiquitinated. To examine the polyubiquitination of ING3, MMRU cells were transfected with HA-Ub plasmid and treated with the proteasome inhibitor Oncogene

MG132 for 6 h before being harvested. Pulldown of ubiquitinated proteins by mouse anti-HA antibody indicated that both the endogenous ING3 and ectopic FLAG-ING3 proteins were polyubiquitinated in MMRU cells and the ubiquitination level of ING3 was increased in the presence of MG132 (Figure 3d). The parallel experiment to immunoprecipitate FLAGING3 using mouse anti-FLAG antibody also showed that FLAG-ING3 is polyubiquitinated (Supplementary Figure S3). Taken together, all these data show that degradation of ING3 is mediated by an ubiquitindependent proteasome pathway. K96 residue is essential for ING3 ubiquitination and degradation In the ubiquitin–proteasome system, an initial ubiquitin moiety is conjugated through its carboxyl-terminal


1501

Figure 2 ING3 is stabilized by proteasome inhibitors. (a) MMRU cells were treated with the proteasome inhibitor MG132 or lactacystin at indicated doses for 6 h, and whole cell extracts were obtained for western blot analysis. (b) Melanoma cells were treated with MG132 (2 mM) for 6 h and harvested for western blot analysis. (c) MMRU cells were treated with cycloheximide (CHX, 20 mg/ml) in the presence of MG132 (2 mM) or lactacystin (2.5 mM) for indicated time and whole cell extracts were obtained for western blot analysis. (d) Half-life of ING3 was determined from three separate experiments in panel c. (e) MMRU cells were transfected with FLAG-ING3, treated with CHX (20 mg/ml) and subjected to western blot analysis using anti-FLAG antibody. (f) ING3 half-life was determined from three independent experiments in panel e.

glycine 76 residue to an a-NH2 group of an internal lysine residue of the substrate, which usually locates at the substrate protein’s amino-terminus (Ciechanover and Ben-Saadon, 2004). To identify the lysine residues mediating ING3 protein ubiquitination and degradation, a series of FLAG-tagged plasmids expressing amino-terminal-truncated ING3 proteins and one expressing carboxyl-terminal-truncated ING3 protein were constructed and transfected into HEK293T cells (Figure 4a). Deletion of 1–95 amino acids did not affect MG132-mediated stabilization of ING3 protein, whereas deletion of 1–111 amino acids significantly abolished MG132-triggered ING3 protein stabilization (Figures 4a and b), suggesting that amino acids 96–111 of ING3 has a crucial role in ING3 degradation. Then, we transfected plasmids for the FLAG-ING3(96–418) or FLAG-ING3(112–418) mutant into HEK293T cells followed by treatment with cycloheximide, and found that protein of ING3(112–418) mutant was more stable than that of ING3(96–418) mutant (Figure 4c). As protein ubiquitination usually occurs at lysine residues, we constructed the lysine-to-arginine point mutants for lysine residues K96, K103 and K105, which are located within the amino acids 96–111 of ING3. The K96R mutant of ING3 did not affect the subcellular location of ING3 (Supplementary Figure S4), but almost completely inhibited the ubiquitination of ING3, whereas the K103/105R mutant of ING3 did

not affect the ubiquitination of ING3 (Figure 4d). Accordingly, K96R mutation of ING3 elongates the half-life of ING3 from 4.0 to 11.3 h (Figures 4e and f), implying that K96 residue has a crucial role in the ubiquitination and degradation of ING3. SCF E3 ligase mediates the ubiquitination and degradation of ING3 As the SCF E3 complex comprises a large family of ubiquitin E3 ligases controlling the ubiquitination of many substrates including cell-cycle regulatory proteins, we examined whether ING3 is a target of the SCF complex. Co-IP assay showed that FLAG-ING3 interacted with two core subunits of SCF complex, HA-Cul1 and HA-Roc1, in HEK293T cells (Figure 5a). Although the ING3 antibody cannot be used for IP assay, pulldown of FLAG-ING3 can precipitate endogenous Cul1 and pulldown of endogenous Cul1 can precipitate endogenous ING3 in MMRU cells (Supplementary Figure S5a and b), which confirmed the ING3–Cul1 interaction at the physiological condition. Knockdown of Cul1 expression in MMRU cells significantly decreased the ubiquitination of ING3 and stabilized ING3 (Supplementary Figure S5c to f). Chk1 (checkpoint kinase 1), a well-characterized target of SCF complex, is also stabilized after repressing Cul1 (Supplementary Figure S5e and f). Next, we attempted to Oncogene


1502

Figure 3 ING3 is degraded by the ubiquitin–proteasome system. (a) MMRU cells were treated with MG132 (2 mM) for 6 h and nuclear extracts were obtained. Proteasome-enriched fractions were isolated by amnion sulfate precipitation for western blot. PSMC4 and a4 were used as specific markers for 26S and 20S proteasomes, respectively. (b) Nuclear extracts from MMRU cells were incubated at 37 1C for indicated time in the presence or absence of isolated active 26S proteasome followed by western blot analysis. (c) MMRU cells were transfected as indicated and whole cell extracts were subjected to immunoprecipitation (IP) by anti-FLAG antibody (IP1). Captured proteins were denatured and applied to IP by anti-HA antibody (IP2) for the detection of ubiquitinated FLAG-ING3 by western blotting. (d) MMRU cells were transfected as indicated for 24 h and treated with or without MG132 (2 mM) for 6 h. Whole cell extracts were immunoprecitpitated by an anti-HA antibody for the detection of polyubiquitinated ING3 by western blotting.

determine the potential F-box proteins recognizing ING3. Skp2 and bTRCP are two important F-box proteins that behave as oncoproteins and direct many cell-cycle regulatory proteins for degradation. We co-transfected FLAG-ING3 with either Myc-Skp2 or HA-bTRCP plasmid into HEK293T cells. Results from co-IP assay showed that FLAG-ING3 was bound to Myc-Skp2 but not to HA-bTRCP (Figure 5b). Pulldown of endogenous Skp2 was also able to precipitate endogenous ING3 (Figure 5c) in MMRU cells. Co-IP analysis also showed that the carboxyl-terminal of ING3 has a key role in the interaction between ING3 and Skp2 (Supplementary Figure S6). Treatment of MMRU cells with Skp2 siRNA reduced the ubiquitination of ING3 (Figure 5d), stabilized ING3 protein (Figure 5e and Supplementary Figure S7) and elongated the half-life of ING3 to 8.7 h compared with that in MMRU cells transfected with control siRNA (5.3 h) (Figures 5f and g). Similarly, knockdown of Skp2 in MMLH, Sk-mel-3 and Sk-mel-93 cells also significantly induced ING3 accumulation (Figure 5e). Co-transfection of Skp2 with ING3 promoted ING3 degradation while the proteasome inhibitor MG132 treatment blocked this effect (Supplementary Figure S8). In addition, we also found an inverse correlation between Skp2 and ING3 expression in melanoma cell lines (Supplementary Figure S9A). As both Skp2 and ING3 are important for cell cycle progression, we examined the Skp2 and ING3 protein levels in MMRU cells during cell cycle progression and found an inverse correlation between Skp2 and ING3 expression (Supplementary Figure S9B). Therefore, our data indicated that degradation of ING3 protein is under the tight Oncogene

control of SCFSkp2 E3 complex in melanoma cells and inhibition of the activity of SCFSkp2 complex will be able to restore its expression. Blockage of ING3 degradation regulates G1 cell-cycle arrest and promotes UV-induced apoptosis Next, we examined whether blockage of ING3 degradation affects its tumor-suppressive roles in melanoma cells. ING3 has been shown to inhibit colony formation efficiency and modulate cell cycle control in RKO cells (Nagashima et al., 2003). Strategies by either silencing Skp2 (Figures 6a–c) or overexpressing ING3 (Figures 6d–f) significantly arrested MMRU cells at G1 phase, while knockdown of ING3 decreased the distribution of MMRU cells at G1-phase from 47.2 to 40.2% (Figures 6a and b), indicating that Skp2 and ING3 have opposite roles in the G1/S transition. In addition, disruption of ING3 inhibited Skp2 knockdown-induced G1 arrest, indicating that ING3 is responsible for Skp2-mediated cell cycle control. Next, we examined whether blockage of ING3 degradation affects cell cycle control in melanoma cells. Our studies showed that the G1 population was significantly increased in MMRU cells expressing the K96R ING3 mutant (58.2%) compared with those expressing wild-type ING3 (51.5%) (Figures 6d and e), confirming that the blockage of SCFSkp2mediated ING3 degradation affects G1 cell-cycle profile. These data implied that interruption of ING3 degradation can significantly enhance its role in cell cycle control in melanoma cells. Previously, we have shown that overexpression of ING3 significantly promotes UVB-induced apoptosis


1503

38.2%, respectively, which were further elevated to 29.3 and 46.1% in cells overexpressing K96R ING3 mutant, respectively (Figure 6g). In addition, we found that UVB irradiation stabilized ectopic expression of wildtype ING3 and K96R ING3 mutant (Supplementary Figure S10). These data imply that interruption of ING3 degradation can significantly enhance its tumor-suppressive roles in melanoma cells. Discussion

Figure 4 K96 is essential for ING3 ubiquitination and degradation. (a) Illustration of wild-type (WT) and truncated ING3 proteins. (b) Plasmids expressing FLAG-tagged WT or truncated ING3 were transfected into HEK293T cells for 24 h and then treated with MG132 (2 mM) for western blot analysis. (c) Plasmids expressing amino acids 96–418 and 112–418 of ING3 were transfected into MMRU cells for 24 h. Cells were treated with cycloheximide (CHX, 20 mg/ml) and harvested at indicated time for western blot analysis. (d) HEK293T cells were transfected as indicated for 24 h followed by treatment with MG132 (2 mM) for 6 h. Whole cell extracts were then processed for IP analysis of ubiquitinated ING3. (e) MMRU cells were transfected with WT or K96R ING3 for 24 h and treated with CHX (20 mg/ml) for indicated time. Whole cell extracts were then subjected to western blot analysis. (f) ING3 half-life was determined from three independent experiments in panel e.

in MMRU melanoma cells (Wang and Li, 2006). Thus, we examined whether stabilization of ING3 by K96R mutation would affect the sensitivity to UV irradiation. The MMRU cells were transfected with FLAG-ING3 and FLAG-ING3 (K96R) plasmids to overexpress wildtype ING3 protein and the ING3 mutant with K96R mutation (Figure 6g). UVB at a dose of either 200 or 600 J/m2 induced apoptosis at a rate of 14.6 and 27.5%, respectively (Figure 6g). In the presence of wild-type ING3, the apoptosis rates were increased to 22.6 and

Novel tumor suppressor ING3 has been shown to modulate transcription, cell cycle control and apoptosis. Recent studies by our group and Gunduz et al. suggest that ING3 is deregulated and functions as a tumor suppressor in both melanomas and HNSCCs (Gunduz et al., 2002, 2008; Wang and Li, 2006; Wang et al., 2007). However, the mechanisms causing distorted ING3 expression in melanomas are different from that in HNSCCs. In HNSCCs, the ING3 gene is rarely mutated, but 50% of primary HNSCCs has decreased or no expression of ING3 mRNA because of loss of heterozygosity (Gunduz et al., 2002, 2008). In this study, we found that rapid ING3 degradation has an important role in the loss of ING3 expression in advanced melanomas, thereby providing another explanation of aberrant ING3 expression in melanomas in addition to our previous finding that subcellular relocalization of ING3 causes its deregulation in both primary and metastatic melanomas (Wang et al., 2007). Meanwhile, ING1b has been shown to be shifted from the nucleus to the cytoplasm in melanocytic lesions (Nouman et al., 2002), which may be mediated by 14-3-3 proteins (Gong et al., 2006). Both nuclear ING2 and overall ING4 expressions are also downregulated in melanomas, although only the reduced ING4 expression was significantly associated with melanoma progression (Lu et al., 2006; Li et al., 2008). The reduced ING4 expression also correlated with tumor thickness, ulceration and a poorer 5year patient survival with primary melanomas (Li et al., 2008). These studies indicate that ING proteins are deregulated in melanomas and may have a crucial role in the development of melanoma. However, the reasons causing the aberrant expression of ING proteins, particularly ING2 and ING4, are largely unknown. As ING genes are located on different chromosomes (He et al., 2005), the aberrant expression of ING proteins are unlikely under a common genetic control. However, the structural similarity of ING proteins may provide an alternative explanation that deregulation of ING proteins in melanoma may be controlled by certain common mechanisms in their rapid protein turnover and/or subcellular relocalization. Future studies are needed to achieve a complete understanding of aberrant ING proteins in melanomas. It is worthy to notice that ING4 is also subjected to N-terminal ubiquitination and degradation through the ubiquitin– proteasome pathway (Tsai et al., 2008). Thus, it will be necessary to investigate whether downregulation of ING4 in melanoma is affected by enhanced degradation through the ubiquitin–proteasome system. Oncogene


1504

Figure 5 SCF E3 complex targets ING3 for degradation. (a, b) HEK293T cells were co-transfected as indicated for 24 h followed by treatment with MG132 (2 mM) for 6 h. Whole cell extracts were subjected to immunoprecipitation (IP) analysis by anti-FLAG or HA antibodies. (c) MMRU cells were treated with MG132 (2 mM) for 6 h and whole cell extracts were obtained for IP assay with an antiSkp2 antibody. (d) MMRU cells were co-transfected as indicated for 48 h followed by western blot analysis of ING3. Whole cell extracts were applied for IP assay with anti-HA antibody followed by western blot analysis of ubiquitinated ING3. (e) Melanoma cells were transfected with control (Ctrl) or Skp2 siRNA for 48 h and processed for western blot analysis of ING3. Numbers below ING3 bands indicate the fold of ING3 accumulation after Skp2 knockdown. (f) MMRU cells were transfected with Ctrl or Skp2 siRNA for 48 h. Cells were then treated with CHX (20 mg/ml), harvested at indicated time and whole cell extracts were subjected to western blot analysis. (g) ING3 half-life was estimated from three separate experiments in panel f.

It appears that ING3 expression is drastically reduced when primary melanomas become metastatic, as overall ING3 expression is not reduced in primary melanomas compared with that in dysplastic nevi (Supplementary Figure S1), suggesting a crucial role of ING3 in melanoma progression and metastasis. It is most likely that the reduced ING3 expression is caused by the rapid degradation of ING3 in advanced melanomas. Possible explanations for the rapid ING3 degradation may include increased E3 ligase level in melanoma metastases and/or any mutation or modification of ING3 favoring its recognition by E3 ligase. On the other hand, it is reported that mdm2-mediated p53 ubiquitination not only contributes to p53 degradation but controls the nucleus-to-cytoplasm shift of p53, especially in the presence of a low level of mdm2 (Inoue et al., 2001). Considering that the expression of Skp2 is increased in melanomas and correlated with melanoma progression (Li et al., 2004; Woenckhaus et al., 2005), it is possible that the increased expression of Skp2 may mainly target ING3 to the cytoplasm in primary melanomas and subsequently enhances both subcellular translocation Oncogene

and degradation of ING3 in melanoma metastases. Future studies will be required to test this hypothesis. Furthermore, our study also shows that the loss of ING3 expression in the metastatic melanoma cell line KZ-13 is a result of decreased ING3 mRNA expression (Figure 1b), implying that there are various mechanisms leading to aberrant ING3 expression in melanoma. Our current data strongly support the crucial role of the ubiquitin–proteasome system in the proteolysis of ING3. The SCF E3 ligase complex has an important role in controlling the level of proteins governing cell cycle progression (Ang and Wade Harper, 2005). Skp2 is well known as a critical S-phase-promoting molecule by targeting the cdk inhibitor p27 for proteolysis (Frescas and Pagano, 2008). It also targets many other cellular proteins for degradation, including cyclin-dependent kinase inhibitors p21, p27 and p57, TOB1 (transducer of Erbb2), RASSF1 (Ras association domain family 1) and FOXO1 (forkhead box-containing transcription factor 1) (Frescas and Pagano, 2008). Upregulation of Skp2 has been shown in several types of cancers and may be an attractive target for the development of novel cancer-


1505

Figure 6 Blockage of ING3 degradation regulates G1 cell-cycle arrest and promotes ultraviolet (UV)-induced apoptosis of MMRU cells. (a–c) MMRU cells transfected with control siRNA (siCtrl), Skp2 siRNA (siSkp2), ING3 siRNA (siING3), and both Skp2 and ING3 siRNA (siSkp2 þ siING3) for 48 h. Cells were collected for FACS analysis of (a) cell cycle distribution, (b) quantification of cells in G1 phase and (c) WB analysis of ING3 and Skp2. (d–f) MMRU cells were transfected with vector, WT or K96R ING3 for 24 h. (d) Cells were collected for FACS analysis of cell cycle distribution, (e) quantification of cells in G1 phase and (f) WB analysis for expression of ING3. (g) MMRU cells transfected with vector, WT or K96R ING3 for 24 h were irradiated with UVB at indicated doses. After incubation for another 24 h, cells were collected for analysis of cell cycle distribution. Cells at sub-G1 phase were considered apoptotic. *Po0.05; **Po0.01; ***Po0.001.

intervention strategies (Masuda et al., 2002; Yang et al., 2002; Yokoi et al., 2004; Sonoda et al., 2006; Chiappetta et al., 2007; Hershko, 2008). In melanoma, Skp2 expression is significantly increased, which correlates with a poorer patient survival (Li et al., 2004; Woenckhaus et al., 2005). Suppression of Skp2 by RNA interference has been shown to inhibit melanoma growth effectively (Katagiri et al., 2006; Sumimoto et al., 2006). Consistent with the role of ING3 in cell growth control (Nagashima et al., 2003), we found that degradation of ING3 protein is under the tight control of SCFSkp2 E3 complex. As the SCFSkp2 E3 ligase complex interacts with their substrates for ubiquitination through the phosphorylated consensus sequence in their target proteins (Frescas and Pagano, 2008), we found that the carboxyl-terminal of ING3 was involved in the interaction between ING3 and Skp2 (Supplementary Figure S6); hence, the ongoing work to identify the phospho-motif of ING3 will improve our knowledge on the subtle mechanism controlling ING3 protein turnover. Stabilization of ING3 protein in the metastatic MMRU melanoma cells resulted in enhanced tumor-

suppressive functions of ING3 (Figure 6), which provides possible explanation for our previous finding that reduced nuclear ING3 expression was significantly correlated with a poorer disease-specific 5-year survival of melanoma patients (Wang et al., 2007). Although ING3 has been reported to inhibit growth and promote apoptosis in RKO cells (Nagashima et al., 2003), the mechanism of inhibitory effect of ING3 in melanoma progression and metastasis is unclear. Thus, it will be of interest to investigate the role of ING3 on migration and invasion in melanoma cells, and whether disruption of ING3 protein degradation inhibits melanoma progression and metastasis in vivo. As metastatic melanoma is very resistant to conventional radiotherapy and chemotherapy, it is necessary to develop novel regimens for the treatment of melanoma. This study showed that interruption of ING3 degradation regulates G1 cellcycle arrest and promotes UV-induced apoptosis (Figure 6). Future study can be designed to examine whether interruption of ING3 degradation increases sensitivity of melanoma cells to radiotherapy and chemotherapy. In summary, we showed that the tumor suppressor ING3 is degraded by SCFSkp2 complex by the ubiquitin– Oncogene


1506

proteosome pathway, and the lysine 96 of ING3 is crucial for its ubiquitination and degradation. Our results suggest that interfering ING3 degradation pathway can affect its tumor-suppressive functions and may provide a novel therapeutic approach for malignant melanoma.

Materials and methods Cell culture and reagents The MMRU, MMLH and MMAN cell lines were gifts from Dr HR Byers (Boston University School of Medicine). The MEWO, Sk-mel-3, Sk-mel-93, and Sk-mel-110 cell lines were provided by Dr AP Albino (Memorial Sloan-Kettering Cancer Center). KZ-13, Sk-mel-5 and HEK293T cell lines were obtained from Dr H Silver (British Columbia Cancer Research Center), the Tissue Bank at the National Institutes of Health (USA) and American Type Culture Collection, respectively. These cell lines and normal human epidermal melanocytes (Clonetics, Norwalk, CT, USA) were cultured as described (Ng et al., 2004; Wang and Li, 2006). The specific proteasome inhibitors MG132 and lactacystin were purchased from Calbiochem (San Diego, CA, USA) and Sigma (Mississauga, ON, Canada), respectively. The protein synthesis inhibitor cycloheximide was purchased from Sigma. Plasmids and transfection HA-tagged pcDNA3-ubiquitin plasmid (HA-Ub) was a kind gift from Dr R Zhang (University of Alabama at Birmingham). HA-Cul1 and HA-Roc1 plasmids were provided by Dr M Pagano (New York University). Myc-Skp2 and HAbTRCP plasmids were gifts from Dr J Hsieh (Washington University) and Dr S Sun (Pennsylvania State University), respectively. ING3 cDNA was a kind gift from Dr M Gunduz (Okayama University), which was subcloned into p3 FLAG vector (Sigma) between BglII and XbaI restriction sites to generate the FLAG-ING3 plasmid. Plasmids expressing ING3-truncated proteins and K96R or K103/105R point mutants were generated with PCR from FLAG-ING3 using site-specific primers and subcloned into the p3 FLAG vector. All the constructs were sequenced across the newly created junctions to confirm no reading frame shift induced by PCR. Plasmids were transfected into cells using the Effectene Transfection Kit (Qiagen, Mississauga, ON, Canada) according to the manufacturer’s procedures. siRNA and transfection The Cul1 siRNA oligomers (50 -CGAAGAGUUCAGGUUUACC-30 ) were synthesized by Dharmacon (Lafayette, CO, USA). The validated Skp2 siRNA oligomers (Hs_SKP2_5 and Hs_SKP2_8) were purchased from Qiagen. siRNA was transfected into cultured cells using the siLentFect Reagent (Bio-Rad, Mississauga, ON, Canada) according to the manufacturer’s instruction. Reverse transcriptase–PCR Total RNA was prepared by TRIzol extraction (Invitrogen, Burlington, ON, Canada) and reverse transcribed into cDNA with the SuperScript First-Strand Synthesis System (Invitrogen) according to the manufacturer’s protocol. Hotstart PCR was performed with Taq DNA PCR (Qiagen). The sequences of human ING3 primers were 50 -CAGCCTCTTCTAACAATGC CTA-30 (sense) and 50 -CTTCATCAAACAAAAGGACCAC-30 (antisense). The primers for human glyceraldehyde-3-phosphate dehydrogenase were 50 -CTCATGACCACAGTCCATGCCA Oncogene

TC-30 (sense) and 50 -CTGCTTCACCACCTTCTTGATGTC-30 (antisense) that served as an input control. Western blot analysis and IP Western blot analysis and IP were performed as described (Garate et al., 2007). The primary antibodies included rabbit anti-ING3 antibodies (Proteintech, Chicago, IL, USA), mouse monoclonal antibodies against p53, PSMC4, a4, Chk1, Cul1, Skp2 (Santa Cruz Biotechnology, Santa Cruz, CA, USA), HA, myc (Epitope Biotech, Burnaby, BC, Canada), FLAG and actin (Applied Biological Materials, Richmond, BC, Canada), and goat anti-Skp2 (Santa Cruz Biotechnology). In vivo ubiquitination assay in vivo ubiquitination was assayed as described (Kuo et al., 2004). Briefly, cells were treated with 2 mM MG132 for 6 h after transfection with HA-Ub plasmid or co-transfection together with FLAG-ING3 plasmid for 24 h. Cells were then harvested and lyzed in modified RIPA buffer (50 mM Tris (pH 7.4), 150 mM NaCl, 1% NP-40, 0.25% sodium deoxycholate, EDTA 1 mM) containing protease inhibitor cocktails and 10 mM N-ethylmaleimide (Sigma) to inhibit ubiquitin hydrolases. Mouse anti-HA monoclonal antibody was used to precipitate HA-tagged ubiquitinated proteins that were denatured and resolved on 8% SDS–PAGE (polyacrylamide gel electrophoresis) for the analysis of ING3. Isolation of proteasome The nuclear extracts obtained from MMRU cells were applied to the isolation of proteasome as previously described (Ben-Shahar et al., 1999; Garate et al., 2008). Briefly, 26S proteasome was precipitated in the presence of ammonium sulfate added to 38% saturation, whereas 20S proteasome was precipitated in the presence of ammonium sulfate added to 70% saturation. The precipitated fractions were loaded on a Superose 6 column (GE Healthcare, Scarborough, ON, Canada), previously calibrated with a Gel Filtration HMW Calibration Kit (GE Healthcare). Fractions were collected for western blot analysis of 26S proteasome marker PSMC4 and 20S proteasome marker a4. In vitro degradation assay The 26S proteasome-containing fraction was electrophoresed through a native 4.5% PAGE, where the 26S proteasome complex was identified by overlay staining using the fluorogenic substrate Suc-LLVY-AMC (BIOMOL International, Plymouth Meeting, PA, USA) as described (Asher et al., 2005). The band containing 26S proteasome was excised and eluted from the gel. For in vitro degradation assay, 50 mg of nuclear extracts were incubated at 37 1C in the presence or the absence of 50 ng of isolated 26S proteasome for different time points.

Conflict of interest The authors declare no conflict of interest.

Acknowledgements We gratefully thank Drs HR Byers, AP Albino, H Silver, R Zhang, M Pagano, S Sun and J Hsieh for kindly providing the materials. This study was supported by Cancer Research Society, the Canadian Institutes of Health Research (MOP93810) and the Canadian Dermatology Foundation (GL).


1507 GC is a recipient of Postdoctoral Trainee Award from Michael Smith Foundation for Health Research. YW is a recipient of Roman M Babicki Fellowship from the University of

British Columbia and a research fellow of the Terry Fox Foundation through an award from the National Cancer Institute of Canada.

References Ang XL, Wade Harper J. (2005). SCF-mediated protein degradation and cell cycle control. Oncogene 24: 2860–2870. Asher G, Tsvetkov P, Kahana C, Shaul Y. (2005). A mechanism of ubiquitin-independent proteasomal degradation of the tumor suppressors p53 and p73. Genes Dev 19: 316–321. Ben-Shahar S, Komlosh A, Nadav E, Shaked I, Ziv T, Admon A et al. (1999). 26 S proteasome-mediated production of an authentic major histocompatibility class I-restricted epitope from an intact protein substrate. J Biol Chem 274: 21963–21972. Burger AM, Seth AK. (2004). The ubiquitin-mediated protein degradation pathway in cancer: therapeutic implications. Eur J Cancer 40: 2217–2229. Chiappetta G, De Marco C, Quintiero A, Califano D, Gherardi S, Malanga D et al. (2007). Overexpression of the S-phase kinaseassociated protein 2 in thyroid cancer. Endocr Relat Cancer 14: 405–420. Ciechanover A, Ben-Saadon R. (2004). N-terminal ubiquitination: more protein substrates join in. Trends Cell Biol 14: 103–106. Doyon Y, Selleck W, Lane WS, Tan S, Cote J. (2004). Structural and functional conservation of the NuA4 histone acetyltransferase complex from yeast to humans. Mol Cell Biol 24: 1884–1896. Fadlelmola FM, Zhou M, de Leeuw RJ, Dosanjh NS, Harmer K, Huntsman D et al. (2008). Sub-megabase resolution tiling (SMRT) array-based comparative genomic hybridization profiling reveals novel gains and losses of chromosomal regions in Hodgkin lymphoma and anaplastic large cell lymphoma cell lines. Mol Cancer 7: 2. Frescas D, Pagano M. (2008). Deregulated proteolysis by the F-box proteins SKP2 and beta-TrCP: tipping the scales of cancer. Nat Rev Cancer 8: 438–449. Garate M, Campos EI, Bush JA, Xiao H, Li G. (2007). Phosphorylation of the tumor suppressor p33(ING1b) at Ser-126 influences its protein stability and proliferation of melanoma cells. FASEB J 21: 3705–3716. Garate M, Wong RP, Campos EI, Wang Y, Li G. (2008). NAD(P)H quinone oxidoreductase 1 inhibits the proteasomal degradation of the tumour suppressor p33(ING1b). EMBO Rep 9: 576–581. Goldberg AL. (2007). Functions of the proteasome: from protein degradation and immune surveillance to cancer therapy. Biochem Soc Trans 35: 12–17. Gong W, Russell M, Suzuki K, Riabowol K. (2006). Subcellular targeting of p33ING1b by phosphorylation-dependent 14-3-3 binding regulates p21WAF1 expression. Mol Cell Biol 26: 2947–2954. Gunduz M, Beder LB, Gunduz E, Nagatsuka H, Fukushima K, Pehlivan D et al. (2008). Downregulation of ING3 mRNA expression predicts poor prognosis in head and neck cancer. Cancer Sci 99: 531–538. Gunduz M, Ouchida M, Fukushima K, Ito S, Jitsumori Y, Nakashima T et al. (2002). Allelic loss and reduced expression of the ING3, a candidate tumor suppressor gene at 7q31, in human head and neck cancers. Oncogene 21: 4462–4470. He GH, Helbing CC, Wagner MJ, Sensen CW, Riabowol K. (2005). Phylogenetic analysis of the ING family of PHD finger proteins. Mol Biol Evol 22: 104–116. Hershko DD. (2008). Oncogenic properties and prognostic implications of the ubiquitin ligase Skp2 in cancer. Cancer 112: 1415–1424. Inoue T, Geyer RK, Howard D, Yu ZK, Maki CG. (2001). MDM2 can promote the ubiquitination, nuclear export, and degradation of p53 in the absence of direct binding. J Biol Chem 276: 45255–45260.

Katagiri Y, Hozumi Y, Kondo S. (2006). Knockdown of Skp2 by siRNA inhibits melanoma cell growth in vitro and in vivo. J Dermatol Sci 42: 215–224. Kuo ML, den Besten W, Bertwistle D, Roussel MF, Sherr CJ. (2004). N-terminal polyubiquitination and degradation of the Arf tumor suppressor. Genes Dev 18: 1862–1874. Li J, Martinka M, Li G. (2008). Role of ING4 in human melanoma cell migration, invasion, and patient survival. Carcinogenesis 29: 1373–1379. Li Q, Murphy M, Ross J, Sheehan C, Carlson JA. (2004). Skp2 and p27kip1 expression in melanocytic nevi and melanoma: an inverse relationship. J Cutan Pathol 31: 633–642. Lu F, Dai DL, Martinka M, Ho V, Li G. (2006). Nuclear ING2 expression is reduced in human cutaneous melanomas. Br J Cancer 95: 80–86. Mani A, Gelmann EP. (2005). The ubiquitin–proteasome pathway and its role in cancer. J Clin Oncol 23: 4776–4789. Masuda TA, Inoue H, Sonoda H, Mine S, Yoshikawa Y, Nakayama K et al. (2002). Clinical and biological significance of S-phase kinase-associated protein 2 (Skp2) gene expression in gastric carcinoma: modulation of malignant phenotype by Skp2 overexpression, possibly via p27 proteolysis. Cancer Res 62: 3819–3825. Nagashima M, Shiseki M, Pedeux RM, Okamura S, Kitahama-Shiseki M, Miura K et al. (2003). A novel PHD-finger motif protein, p47ING3, modulates p53-mediated transcription, cell cycle control, and apoptosis. Oncogene 22: 343–350. Nalepa G, Rolfe M, Harper JW. (2006). Drug discovery in the ubiquitin–proteasome system. Nat Rev Drug Discov 5: 596–613. Nandi D, Tahiliani P, Kumar A, Chandu D. (2006). The ubiquitin– proteasome system. J Biosci 31: 137–155. Newton K, Vucic D. (2007). Ubiquitin ligases in cancer: ushers for degradation. Cancer Invest 25: 502–513. Ng KC, Campos EI, Martinka M, Li G. (2004). XAF1 expression is significantly reduced in human melanoma. J Invest Dermatol 123: 1127–1134. Nouman GS, Anderson JJ, Mathers ME, Leonard N, Crosier S, Lunec J et al. (2002). Nuclear to cytoplasmic compartment shift of the p33ING1b tumour suppressor protein is associated with malignancy in melanocytic lesions. Histopathology 40: 360–366. Schwechheimer C, Calderon Villalobos LI. (2004). Cullin-containing E3 ubiquitin ligases in plant development. Curr Opin Plant Biol 7: 677–686. Shah IM, Di Napoli M. (2007). The ubiquitin–proteasome system and proteasome inhibitors in central nervous system diseases. Cardiovasc Hematol Disord Drug Targets 7: 250–273. Sonoda H, Inoue H, Ogawa K, Utsunomiya T, Masuda TA, Mori M. (2006). Significance of skp2 expression in primary breast cancer. Clin Cancer Res 12: 1215–1220. Sumimoto H, Hirata K, Yamagata S, Miyoshi H, Miyagishi M, Taira K et al. (2006). Effective inhibition of cell growth and invasion of melanoma by combined suppression of BRAF (V599E) and Skp2 with lentiviral RNAi. Int J Cancer 118: 472–476. Tsai KW, Tseng HC, Lin WC. (2008). Two wobble-splicing events affect ING4 protein subnuclear localization and degradation. Exp Cell Res 314: 3130–3141. von Mikecz A. (2006). The nuclear ubiquitin–proteasome system. J Cell Sci 119: 1977–1984. Wang Y, Dai DL, Martinka M, Li G. (2007). Prognostic significance of nuclear ING3 expression in human cutaneous melanoma. Clin Cancer Res 13: 4111–4116. Oncogene


1508 Wang Y, Li G. (2006). ING3 promotes UV-induced apoptosis via Fas/caspase-8 pathway in melanoma cells. J Biol Chem 281: 11887–11893. Woenckhaus C, Maile S, Uffmann S, Bansemir M, Dittberner T, Poetsch M et al. (2005). Expression of Skp2 and p27KIP1 in naevi and malignant melanoma of the skin and its relation to clinical outcome. Histol Histopathol 20: 501–508. Yang G, Ayala G, De Marzo A, Tian W, Frolov A, Wheeler TM et al. (2002). Elevated Skp2 protein expression in human prostate cancer:

association with loss of the cyclin-dependent kinase inhibitor p27 and PTEN and with reduced recurrence-free survival. Clin Cancer Res 8: 3419–3426. Yokoi S, Yasui K, Mori M, Iizasa T, Fujisawa T, Inazawa J. (2004). Amplification and overexpression of SKP2 are associated with metastasis of non-small-cell lung cancers to lymph nodes. Am J Pathol 165: 175–180. Zetter BR, Mangold U. (2005). Ubiquitin-independent degradation and its implication in cancer. Future Oncol 1: 567–570.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc)

Oncogene