Oncogene (2015) 34, 342–350 & 2015 Macmillan Publishers Limited All rights reserved 0950-9232/15 www.nature.com/onc

ORIGINAL ARTICLE

Regulation of Mdm2 protein stability and the p53 response by NEDD4-1 E3 ligase C Xu, CD Fan and X Wang Mdm2 is a critical negative regulator of the tumor suppressor protein p53. Mdm2 is an E3 ligase whose overexpression leads to functional inactivation of p53. Mdm2 protein stability is regulated by several mechanisms including RING (Really Interesting New Gene) domain-mediated autoubiquitination. Here we report biochemical identification of NEDD4-1 as an E3 ligase for Mdm2 that contributes to the regulation of Mdm2 protein stability in cells. NEDD4-1 was identified from Jurkat cytosolic fractions using an enzyme-dead Mdm2 mutant protein as a substrate for in vitro E3 ligase assays. We show that lysates from Nedd4-1 knockout (KO) mouse embryonic fibroblasts (MEFs) have significantly diminished E3 ligase activity toward Mdm2 compared with lysates from wildtype (WT) MEFs. Recombinant NEDD4-1 promotes Mdm2 ubiquitination in vitro in a concentration- and time-dependent manner. In cells, NEDD4-1 physically interacts with Mdm2 via the RING domain of Mdm2. Overexpression of NEDD4-1, but not an enzyme-dead NEDD4-1CS mutant, increases ubiquitination of Mdm2. NEDD4-1 catalyzes the formation of K63-type polyubiquitin chains on Mdm2 that are distinct from K48-type polyubiquitination chains mediated by the Mdm2/MdmX complex. Importantly, K63-type polyubiquitination by NEDD4-1 competes with K48-type polyubiquitination on Mdm2 in cells. As a result, NEDD4-1-mediated ubiquitination stabilizes Mdm2. NEDD4-1 knockdown reduces the t1/2 (half-life) of endogenous Mdm2 from 20 to 12 min in U2OS cells. Nedd4-1 KO MEFs manifest increased p53 levels and activity, a more robust DNA damage response and increased G1 arrest compared with WT MEFs. Similarly, NEDD4-1 knockdown in WT-p53-bearing cells increases basal p53 levels and activity in an Mdm2-dependent manner, causes stronger p53 responses to DNA damage and results in p53-dependent growth inhibition compared with corresponding NEDD4-1-proficient control cells. This study identifies NEDD4-1 as a novel component of the p53/Mdm2 regulatory feedback loop that controls p53 activity during stress responses. Oncogene (2015) 34, 342–350; doi:10.1038/onc.2013.557; published online 13 January 2014 Keywords: Mdm2; p53; NEDD4-1; E3 ligases; degradation; ubiquitination

INTRODUCTION TP53 gene is the most frequently mutated tumor suppressor gene in human cancer.1 In addition to direct mutational inactivation of TP53 gene, p53 protein activity is frequently inhibited by alternative mechanisms. Among p53-negative regulators identified so far, Mdm2 has a central role in inhibition of p53 activity in vivo.2 Mdm2 inhibits p53 by promoting ubiquitindependent proteasomal degradation and by inhibition of transcriptional activity of p53. Mdm2 also has non-p53 substrates. E-cadherin ubiquitination by Mdm2 and its subsequent degradation promotes cell migration and invasiveness,3 and Mdm2 expression together with Ki-67 correlate with distant metastasis in prostate cancer.4 Mdm2 is frequently overexpressed in childhood leukemia by posttranslational mechanisms.5,6 Strikingly, 50% of pediatric acute lymphoblastic leukemia patients tested have Mdm2 protein overexpression but without Mdm2 gene amplification or p53 mutation, suggesting that Mdm2 protein overexpression is likely due to post-transcriptional mechanisms and that Mdm2 overexpression is sufficient to inactivate the p53 pathway.5 Therefore, understanding of post-transcriptional regulation of Mdm2 protein is critical for understanding of Mdm2 deregulation in human cancer. Mdm2 is a RING (Really Interesting New Gene) domain E3 ligase, whose E3 ligase activity is carried in the C-terminal RING domain

of the protein.7 Biochemical evidence indicates that Mdm2 on its own only mediates monoubiquitination of p53 at multiple sites.8,9 By interacting with MdmX through its RING domain, we previously reported that Mdm2/MdmX catalyzes polyubiquitination of p53 that targets its degradation by 26S proteasomes.9 In contrast to the intensive studies on p53 degradation by Mdm2, mechanisms underlying Mdm2 stability regulation are not well understood. Using nitric oxide, a physiological DNA damage inducer and a protein-nitrosylating agent, we previously found that downregulation of Mdm2 protein by post-translational mechanisms is involved in p53 activation by nitric oxide.10 Importantly, destabilization of Mdm2 protein also occurs in cells treated with other types of genotoxic agents.11 Mdm2 stability has been proposed to be regulated by its own RING domain through autoubiquitination.11 Surprisingly, evidence from the knock-in mice of enzyme-dead RING domain mutant Mdm2 indicates that Mdm2 RING domain is not required for Mdm2 stability regulation in vivo.12 Although RING domain mutant Mdm2 expression levels are higher than wild-type (WT) Mdm2 in co-transfection experiments, endogenous Mdm2 RING mutant in knock-in mice is expressed at levels comparable to WT Mdm2. Moreover, the kinetics of DNA damage-triggered Mdm2 degradation is identical between the enzyme-dead mutant Mdm2 and WT Mdm2.12 These findings indicate that Mdm2 stability is regulated by other E3 ligase activities rather than by its own RING domain. In line with

Department of Pharmacology and Therapeutics, Roswell Park Cancer Institute, Buffalo, NY, USA. Correspondence: Dr X Wang, Department of Pharmacology and Therapeutics, Roswell Park Cancer Institute, Elm and Carlton Street, Buffalo, NY 14263, USA. E-mail: [email protected] Received 25 June 2013; revised 30 October 2013; accepted 22 November 2013; published online 13 January 2014

NEDD4-1 positively regulates Mdm2 stability and function C Xu et al

343 autoubiquitination-independent regulation of Mdm2 degradation, it has been reported that PCAF (p300/CBP-associated factor) and Skp-cullin-F-box (b-TrCP) E3 complex SCF(b-TrCP) can act as E3 ligases for Mdm2 to promote Mdm2 degradation.13,14 In this report, we describe biochemical purification of Mdm2-associated E3 ligase activity 1 (MELA1). We provide evidence that NEDD4-1 E3 ligase is one of the MELA activities capable of ubiquitinating Mdm2, stabilizing Mdm2 protein and thus contributing to regulation of the p53 pathway.

RESULTS Autoubiquitination-independent regulation of Mdm2 protein stability To understand the role of the Mdm2 RING domain in regulation of Mdm2 protein stability, p53  /  /Mdm2  /  double knockout (DKO) mouse embryonic fibroblasts (MEFs) were transfected with DNA constructs expressing either WT human Mdm2 (Hdm2) or RING mutant (Hdm2CA), and the half-life (t1/2) of Hdm2 protein was measured in cycloheximide chase experiments. As shown in Figure 1a, Hdm2CA expression levels are relatively higher than WT Hdm2, indicating that RING domain contributes to Mdm2 protein expression levels in transient transfections, as reported in the literature.7 However, the t1/2 of Hdm2 is comparable with that of Hdm2CA (t1/2, 11 min for Hdm2 vs 7 min for Hdm2CA, respectively), indicating that Hdm2CA protein is also a labile protein. To examine whether Hdm2CA undergoes proteasomal degradation, DKO cells after transfection with Hdm2 or Hdm2CA were treated with proteasome inhibitor MG132 for 4 h, and Hdm2 expression levels were analyzed by western blotting. As expected, MG132 treatment caused a significant increase in Hdm2CA levels, similar to the MG132 effect on Hdm2 (Figure 1b). Results from in vivo ubiquitination assay indicate that Hdm2CA can be ubiquitinated in cells, although to a lesser extent compared with Hdm2. MG132 treatment significantly increased Hdm2CA ubiquitination in cells (Figure 1c). These results demonstrate that Hdm2 can be degraded in cells independent of its own RING domain function via ubiquitin-dependent proteasomal pathway. Ubiquitination of Hdm2CA indicates that other E3 activities exist for Mdm2 ubiquitination in cells. We designate these cellular E3 activities as MELA and Cytosolic MELA as MELA1. We found that MELA can be easily detected by in vitro ubiquitination assay using cytosolic or nuclear fractions from Jurkat or HeLa cells and with either GST-Hdm2CA- or double-HA-tagged Hdm2CA as substrates (Figures 1d and e and data not shown). Notably, MELA is much stronger in the Jurkat cytosolic fraction (JS100) than in the nuclear fraction (JNE) (Figure 1e). Purification and identification of NEDD4-1 as MELA1 that mediates Mdm2 ubiquitination We sought to identify MELA1 by biochemical purification with our in vitro assay. We chose to use Jurkat S100 as starting material to avoid nuclear MELA2 and/or PCAF previously described as an E3 ligase for Mdm2.13 The purification procedure is outlined in Figure 2a. The peak activity of MELA1 from Mono Q column (Amersham, GE Healthcare Bio-Sciences, Pittsburgh, PA, USA) (fraction 8, MQ8) was further fractionated by sucrose gradient centrifugation. The resulting fractions were subjected to activity assay for MELA1 and silver staining of the proteins. The results indicate that the MELA1 peak activity resides in sucrose gradient fraction 8 (Figure 2b, upper panel). However, the purification strategy failed to purify MELA1 to homogeneity as shown by several protein bands in SG8 with a distribution profile in sodium dodecyl sulfate–polyacrylamide gel electrophoresis indistinct from that of the neighboring fractions, precluding excision of any candidate protein bands that potentially match the MELA1 activity profile (Figure 2b). We decided to obtain the list of identities of all & 2015 Macmillan Publishers Limited

Figure 1. RING domain-independent degradation of Mdm2 by ubiquitin proteasomal pathway. (a) Western blot analysis (WB) of transfected WT Hdm2 and its RING domain mutant (CA) decay after addition of cycloheximide (CHX) for indicated time (min) in p53/ Mdm2 DKO MEFs. nTf, non-transfected control. (b) Effects of proteasome blockade by MG132 on WT and CA Hdm2 protein levels by WB. VC, empty vector control for transfection. Tubulin serves as loading control. (c) In vivo analysis of ubiquitination of transfected WT or CA Hdm2 together with His-ubiquitin (His-Ub) in DKO MEFs. After His-Ub-pulldown, the ubiquitinated Hdm2 (Ub-Hdm2) was detected by WB analysis (immunoblotting (IB)) of Hdm2. (d) In vitro analysis of ubiquitination of GST-Hdm2CA by HeLa cytosolic fraction (S100) and nuclear fraction (NE). Human E1 and UbcH5c were used. The E3 source was either HeLa S100 or NE. GST-Hdm2CA substrate was isolated by GST-pulldown of a substrate after in vitro reaction. The ubiquitinated Hdm2CA was detected by IB for ubiquitin. GST serves as a nonspecific substrate control. (e) In vitro analysis of ubiquitination of dHA-Hdm2CA by Jurkat cytosolic fraction (JS100) and nuclear fraction (JNE). The experiment was performed similar to that in (d), except that the ubiquitinated Hdm2CA was detected by direct IB for HA without GST-pulldown.

proteins in sucrose gradient fraction 8 by mass spectrometric (MS) analysis. The MS results indicate that NEDD4-1 is the only E3 ligase among the identified proteins with a Mascot score above 90 (95% confidence interval for positive protein identification) (Figure 2c). Characterization of NEDD4-1-mediated ubiquitination of Mdm2 in vitro and in vivo To confirm whether NEDD4-1 has MELA1 activity, we conducted experiments characterizing Mdm2 ubiquitination by NEDD4-1 in vitro and in vivo. First, in in vitro ubiquitination assays using RING domain intact but enzyme-dead GST-Hdm2L468A mutant (G-L468A) as a substrate,9 we found that the MELA1 activity is significantly reduced in both cytosolic (S100) and nuclear fractions (NE) prepared from Nedd4-1 KO MEFs compared with the corresponding fractions from WT MEFs (Figure 3a). Recombinant NEDD4-1 protein efficiently ubiquitinated G-L468A in a concentration-dependent manner, showing evident activity at 30 nM (Figure 3b, upper panel). Moreover, NEDD4-1 catalyzed a timedependent G-L468A ubiquitination in vitro at fixed concentration Oncogene (2015) 342 – 350

NEDD4-1 positively regulates Mdm2 stability and function C Xu et al

344 of 300 nM (Figure 3, lower panel). In an in vivo Hdm2 ubiquitination assay using cell lysates from p53-null PC3 cells co-transfected with Hdm2 or L468A with WT NEDD4-1 or the enzyme-dead NEDD4-

1CS for 24 h and followed by 4 h treatment with proteasome inhibitor MG132, NEDD4-1 but not NEDD4-1CS mutant promoted ubiquitination of both Hdm2 and L468A (Figure 3c), indicating direct ubiquitination of Hdm2 by NEDD4-1 E3 ligase activity in cells. To demonstrate physical interaction between NEDD4-1 and Hdm2, we performed co-immunoprecipitation experiments using PC3 cells co-transfected with different forms of Hdm2 together with HA-tagged-NEDD4-1. Our results indicate that NEDD4-1 binds to Hdm2 and the splice variant Hdm2-B (deleted of amino acids 28–299 with intact RING domain) but not to the RING-deleted mutant. These results suggest that NEDD4-1 interacts with Hdm2 involving the RING domain of Hdm2 (Figure 3e). These data allow us to conclude that NEDD4-1 is a bona fide cellular E3 ligase for Mdm2. NEDD4-1-mediated ubiquitination of Mdm2 contributes to Mdm2 stabilization Protein ubiquitination has diverse effects on proteins including proteasomal degradation, signaling complex formation and subcellular translocation of proteins, depending on the types of the ubiquitin chain.15–17 As Mdm2 is a labile protein, and it is polyubiquitinated by NEDD4-1 (Figure 3b), we expected that NEDD4-1 would promote Mdm2 proteasomal degradation. Surprisingly, we noted that overexpression of NEDD4-1 is accompanied with increased levels of both WT Hdm2 and enzyme-dead L468A expression in the in vivo ubiquitination assay in the absence of proteasomal inhibitor MG132 (Supplementary Figure 1). To understand the effects of NEDD4-1 on Mdm2 stability, we characterized the types of polyubiquitin chain of Mdm2 catalyzed by NEDD4-1. As shown in Figure 3d, in in vitro ubiquitination assay, HA-Hdm2 ubiquitination by NEDD4-1 generated a similar pattern of Hdm2 polyubiquitination with WT ubiquitin and K63-only ubiquitin (K63O). In contrast, the patterns of Hdm2 ubiquitination by NEDD4-1 are similar with K48-only ubiquitin and lysine-less ubiquitin (K0 ubiquitin) (compare the Hdm2 ubiquitination patterns between wt and K63O, or between K48O and K0 in N4-1/Hdm2 of Figure 3d). These results indicate that NEDD4-1 catalyzes K63-type polyubiquitin chain formation on Hdm2. As a control, we show that HdmX-stimulated Hdm2 polyubiquitination is K48-type chain because Hdm2 polyubiquitination by Hdm2/HdmX is generated only with WT ubiquitin and K48-only ubiquitin but not with K63-only ubiquitin and K0 ubiquitin (compare the Hdm2 ubiquitination patterns between WT and K48O, or between K63O and K0 in Hdm2/HdmX of Figure 3d). Of note, WT ubiquitin gives much higher efficiency for Hdm2 polyubiquitination by either NEDD4-1 or stimulated by HdmX. Of note, a significant amount of polyubiquitinated Hdm2 catalyzed by Hdm2/HdmX was found in stacking gels (Figure 3d, Hdm2/HdmX, st.gel). To further understand the effect of NEDD4-1 on Mdm2 protein stability, we manipulated NEDD4-1 protein expression on Mdm2 steady-state levels in p53-null PC3 cells. Overexpression of WT NEDD4-1 but not the enzyme-dead NEDD41CS mutant dose-dependently increased the protein levels of both Hdm2 and L468A (Figure 4a), indicating that the Hdm2-stabilizing effect of NEDD4-1 is dependent on NEDD4-1 E3 ligase activity. To further confirm this unexpected effect, we performed experiments

Figure 2. Biochemical purification of MELA1 from Jurkat S100. (a) Flowchart of purification procedure. (b, upper panel) Western blot analysis of Mdm2CA ubiquitination by MELA1 activity in fractions (Fr) of last step purification by sucrose gradient centrifugation. (Lower panel) The silver staining result of the same fractions. (c) Protein lists identified by MS analysis that are enriched in MELA1 peak fraction (sucrose gradient fraction 8 (SGF8)). Proteins with Mascot score above 90 (95% confidence interval for positive protein identification) are listed and NEDD4-1 (in frame) is the only E3 ligase among them. Oncogene (2015) 342 – 350

& 2015 Macmillan Publishers Limited

NEDD4-1 positively regulates Mdm2 stability and function C Xu et al

345 null MEFs and increased Hdm2 levels in U2OS cells after NEDD4-1 overexpression have effect on p53 accumulation: the basal p53 level is higher in Nedd4-1-null MEFs than WT MEFs, and p53 levels decreased by NEDD4-1 overexpression in U2OS cells (Figures 4c and d, p53). As expected, NEDD4-1 overexpression promoted Mdm2-dependent ubiquitination of endogenous p53 in transfected U2OS cells (Supplementary Figure 2). All these results together allow us to conclude that NEDD4-1-mediated ubiquitination stabilizes Mdm2, which in turn impacts the outputs of the p53-regulatory network. Consistent with these effects, in vivo assay of degradable K48-linked ubiquitin chains indicated that WT NEDD4-1 inhibited K48 ubiquitination of Hdm2; in contrast, enzyme-dead NEDD4-1 increased K48 ubiquitination of Hdm2 in PC3 cells (Figure 4e, K48-ub-Hdm2). To confirm that NEDD4-1 indeed stabilizes Mdm2 protein, we generated NEDD4-1-inducible knockdown U2OS cell line (U2OS-iN4i) and performed cycloheximide chase experiments in the presence or absence of tetracycline. Our results indicate that the t1/2 of Hdm2 protein decreased from 20 min in the presence of NEDD4-1 (U2OSiN4i, Tet  ) to 12 min in the absence of NEDD4-1 in (U2OSiN4i, Tet þ ) in these cells (Figure 4f), consistent with the expected effects.

Figure 3. Characterization of NEDD4-1 as an E3 ligase for Mdm2 in vitro and in vivo. (a) Western blot (WB) analysis of the effects of Nedd4-1 genetic status on MELA1-mediated ubiquitination of Mdm2. In vitro ubiquitination of enzyme-dead GST-Hdm2L468A (G-L468A) was performed with cytosolic (Cyt) and nuclear fraction (NE) prepared from MEFs (WT) or Nedd4-1 KO (N4  /  ) MEFs. The ubiquitinated (Ub) G-L468A was detected by immunoblotting (IB) for ubiquitin after GST-pulldown (GST-pdn). Tubulin and poly(ADPribose) polymerase (PARP) serve as purity and loading of the Cyt and NE fractions. (b) In vitro ubiquitination of G-L468A by recombinant NEDD4-1 (rN4). The experiment was performed as in (a). (Upper panel) Ubiquitination of G-L468A by increasing concentrations of rN4. (Lower panel) Ubiquitination of G-L468A by increasing incubation time with 300 nM rN4. (c) In vivo analysis of Hdm2 ubiquitination in PC3 cells transfected with Hdm2 or L468A mutant together with His-ubiquitin (His-Ub) and NEDD4-1(WT, N4-1) or enzyme-dead NEDD4-1 (CS). (Upper panel) Twenty-four hours after transfection followed by 4 h treatment with 25 mM MG132, the cell lysates were used for analysis of in vivo Hdm2 ubiquitination by Histag-pulldown (His-pdn) followed by IB for Hdm2 (mixture of 2A9 and 4B11 antibodies). (Lower panel) Direct WB analysis of indicated proteins in the cell lysates used for ubiquitination assay. (d) In vitro analysis of polyubiquitin chain types of Hdm2 catalyzed by rN4-1. In vitro Hdm2 ubiquitination was performed with HA-Hdm2 in the presence of N4-1 (N4-1/Hdm2) or HdmX (Hdm2/HdmX) with WT ubiquitin (wt), K63-only-ubiquitin (K63O), K48-only-ubiquitin (K48O) or lysineless-ubiquitin (K0). Ubiquitinated HA-Hdm2 was detected by direct IB for HA. pUb-Hdm2, polyubiquitinated Hdm2; mmUbHdm2, multi-monoubiquitinated Hdm2; St. gel, stacking gel. (e) Coimmunoprecipitation (IP) of Hdm2 and NEDD4-1, PC3 cells were transfected with pcDNA3.1-dHA-NEDD4-1 and Hdm2, (pCHDM1B), or pcDNA3.1-Hdm2 deltaRING (DelR) or Hdm2B (pCHDM1B-Hdm2B). Cell lysates were immunoprecipitated with anti-HA antibody and protein-G DYNA beads and the immunoprecipitates were immunoblotted for Hdm2 and HA. The equal inputs were shown.

using multiple cell lines and different approaches to NEDD4-1 manipulation. First, we transiently knocked down NEDD4-1 expression in p53-null PC3 cells and HCT116-p53-null cells. In both cell types, knockdown of NEDD4-1 resulted in decreased levels of endogenous Hdm2 without effects on HdmX (Figure 4b). Second, we compared the WT MEFs with Nedd4-1 KO MEFs for their levels of endogenous Mdm2. As expected, in Nedd4-1-null MEFs, Mdm2 expression level is lower compared with WT MEFs (Figure 4c). Third, we overexpressed NEDD4-1 in U2OS cells and found that NEDD4-1 overexpression increased endogenous Hdm2 levels (Figure 4d). Importantly, the decreased Mdm2 in Nedd4-1& 2015 Macmillan Publishers Limited

NEDD4-1 contributes to p53 response to DNA damage and p53dependent regulation of cell growth To further understand the role of NEDD4-1 in regulation of the p53 pathway via its effects on Mdm2 stability, we assessed the effects of NEDD4-1 on p53-dependent transcription using p21-luciferase assay. As shown in Figure 5a, p53-dependent activation of p21 promoter is inhibited by co-transfection of Hdm2 in Nedd4-1 KO MEFs. The co-transfection of NEDD4-1 dose-dependently enhanced Hdm2-mediated inhibition of p53-dependent transcription, while co-transfection of NEDD4-1 with p53 did not affect p53-dependent transcription. We further assessed the effect of Nedd4-1 status on p53 response to DNA damage in vivo, we compared the p53 accumulation and p21 induction between WT and Nedd4-1 KO MEFs. As shown in Figure 5b, radiationmimicking agent neocarzinostatin (NCS) induced an earlier and stronger p53 accumulation and p21 induction in Nedd4-1-null MEFs than in WT MEFs, indicating that Nedd4-1 contributes to negative regulation of the p53 pathway. Using doxorubicin as a different type of DNA-damaging agent, we confirmed the stronger p53 accumulation in Nedd4-1 KO MEFs than WT MEFs (Figure 5c). This NEDD4-1 effect on the p53 pathway is also preserved in human cells. When NEDD4-1 was knocked down in two WT-p53bearing U2OS and HCT116 cells, p53 accumulation and p21 induction were observed. The effect of NEDD4-1 knockdown on the increased p53 level and activity is via Hdm2 because knockdown of Hdm2 abrogated this effect (Figure 5d, U2OS). However, NEDD4-1 knockdown caused an increase of endogenous Hdm2 levels in both U2OS and HCT116 cells resulting from the p53/Hdm2 feedback loop effects (Figure 5d, Hdm2). To further confirm the effect of NEDD4-1 on the p53 pathway, we used U2OS-iN4i cell line for NCS treatment. The results indicated that in the absence of NEDD4-1 after tetracycline administration NCS caused an earlier and stronger p53 response in U2OS-iN4i cells as evidenced by a higher p53 accumulation and a stronger induction of p53 target genes such as Hdm2 and p21 (Figure 5e). As p53 is a stress-responsive protein,18 as a control of tetracycline effect on the p53 pathway, we treated tet-repressor-expressing parental U2OS-TetR cells with tetracycline followed by NCS treatment similar to that in U2OS-iN4i cells, and found that tetracycline treatment on its own did not alter the p53 response to NCS (Supplementary Figure 3), suggesting that the effects of NEDD4-1 knockdown on the p53 response to NCS is NEDD4-1- but not tetracycline-dependent. DNA damage signaling stimulates Mdm2 degradation at the early time points of the p53 response.10,11 Using U2OS-iN4i cells and samples of earlier Oncogene (2015) 342 – 350

NEDD4-1 positively regulates Mdm2 stability and function C Xu et al

346 time points after NCS treatment, we found that the NEDD4-1 did not affect this stimulated Mdm2 degradation because Hdm2 downregulation still occurred in the absence of NEDD4-1 (Figure 5f). As DNA damage induces p53-dependent transient cell cycle arrest, we assessed the NEDD4-1 effect on cell cycle regulation after DNA damage. Flow cytometry analysis indicates that NCS treatment induced a much stronger G1 arrest in Nedd4-1 KO MEFs than WT MEFs: G1/S ratio change from 1.1 to 8.4 for Nedd4-1 KO MEFs compared with 1.3 to 2.5 for WT MEFs (Figures 6a and b). These results suggest that NEDD4-1 contributes to inhibition of the p53 pathway probably by its stabilizing effect on Mdm2. To further demonstrate the NEDD4-1 effects on p53dependent growth control, we used U2OS cells to knock down p53 by lentiviral short hairpin RNA vector to create p53knockdown U2OS cells (U2OS-p53-null). Then these cells were transfected with pSuperior-LacZi (LacZi) or pSuperior-N4-1i (N4i). Complete knockdown of p53 and partial knockdown of NEDD4-1 in these cells were confirmed by western blot analysis (Supplementary Figure 4), and their growth rates were monitored for 5 days. The results indicated that knockdown of NEDD4-1 decreased the growth rates of both U2OS and U2OS-p53-null cells compared with LacZi control, indicating that NEDD4-1 affects cell growth by p53-dependent and p53-independent mechanisms (Figure 6c). However, from day 3 on, the extents of growth inhibition induced by NEDD4-1 knockdown in U2OS cells (reduced growth % against LacZi control were 52%, 54% and 50% for days 3, 4 and 5, respectively) were significantly larger than that in U2OS-p53-null cells (reduced growth % against LacZi control were 35%, 30% and 18% for days 3, 4 and 5, respectively) (Po0.0001), indicating a significant contribution of p53-dependent mechanism

Oncogene (2015) 342 – 350

in growth inhibition in U20S cells upon NEDD4-1 knockdown. These results suggest that NEDD4-1-mediated Mdm2 regulation may have a significant role in p53-dependent regulation of cell proliferation.

DISCUSSION Upon discovery of Mdm2 as a member of a large RING domain E3 ligase family, it was shown that Mdm2 RING domain is critical for downregulating p53 as well as for the expression levels of Mdm2 protein itself.7 It has been widely accepted that autoubiquitination regulates Mdm2 abundance in cells. In this regard, the results from Mdm2 RING mutant–knock-in MEFs shed light on Mdm2 downregulation by DNA damage signaling: Mdm2 undergoes regulation by other E3 ligases in addition to autoubiquitinationdependent mechanisms, as Mdm2 RING mutant expressed from endogenous knock-in alleles has similar degradation kinetics as WT Mdm2 protein after DNA damage.12 Two E3 ligases have been reported that ubiquitinate Mdm2 and promote its proteasomal degradation. One of them is PCAF, a well-established acetyltransferase but was shown to also possess intrinsic E3 ligase activity toward Mdm2.13 However, whether PCAF is involved in regulated Mdm2 degradation needs to be established. Recently, SCF(b-TrCP) was discovered to be a second E3 activity. SCF(b-TrCP) promotes Mdm2 degradation in a manner dependent of Mdm2 phosphorylation by casein kinase I.14 Importantly, it was further

Figure 4. Ubiquitination by NEDD4-1 stabilizes Hdm2 and Hdm2L468A. (a) Western blot (WB) analysis of the effects of overexpression of NEDD4-1(N4-1, WT) or enzyme-dead (CS) NEDD4-1 on protein levels of transfected Hdm2 or L468A in PC3 cells. Cell lysates were analyzed by WB analysis for Hdm2, N4-1 and green fluorescent protein (GFP) after transfection of PC3 cells with indicated DNA (100 or 250 ng of N4-1 and 10 ng of Hdm2). Cotransfected GFP (50 ng) serves as a control for transfection efficiency. The relative Hdm2 band intensities are quantitated by IMAGEJ software and presented by the numbers at the bottom after they were normalized on the control Hdm2 level in the sample transfected with Hdm2 alone. (b) WB analysis of the effects of knockdown of NEDD4-1 (N4i) on the steady-state levels of endogenous Hdm2 and HdmX in p53-null PC3 and HCT116 (p53  /  ) cell lines. The indicated cells were transfected with pSuperior-LacZi (Zi) or pSuperior-N4-1i (N4i). Twenty-four hours after transfection, cell lysates were subjected to WB analysis for N4-1, Hdm2 and HdmX. The relative Hdm2 band intensities were presented at the bottom. (c) WB analysis of the effects of Nedd4-1 genetic status on the steady-state levels of endogenous Mdm2 and p53 proteins. Cell lysates from WT ( þ / þ ) or Nedd4-1  /  (  /  ) MEFs were used for the analysis. Mdm2 was detected with 2A10 and p53 with PAb421 antibodies, respectively. (d) WB analysis of the effects of NEDD4-1 overexpression on the levels of endogenous Mdm2 and p53 in U2OS cells. Increasing amounts of NEDD4-1 (0.1, 0.2 and 0.5 mg DNA) were transfected to U2OS cells and levels of endogenous NEDD4-1, Hdm2 and p53 were monitored by WB analysis with GFP and tubulin as loading control. (e) In vivo analysis of NEDD4-1 effect on K48 ubiquitination (K48-ub) of Hdm2 in PC3 cells. PC3 cells were transfected with 200 ng of Hdm2 (pCHDM1B) and 500 ng of HA-K48-only ubiquitin with 1 mg of NEDD4-1 (N4-1, WT) or enzyme-dead NEDD4-1 (N4-1, CS)) in 6-cm plates. Twentyfour hours later, the cells were harvested and the cell lysates were used for immunoprecipitation (IP) with an anti-HA antibody, followed by WB analysis for Hdm2 with a mixture of 2A9 and 4B11 antibodies. The smearing patterns of Hdm2 bands are K48ubiquitinated Hdm2. (f ) Cycloheximide (CHX) chase experiment was performed in inducible NEDD4-1-knockdown U2OS cells (U2OS-iN4i) treated with tetracycline (5 mg/ml) for 3 days (Tet þ ) or without tetracycline (Tet  ). Representative WB analysis results are shown. Hdm2 half-life (t1/2) in the presence or absence of NEDD4-1 knockdown obtained from three independent experiments was shown at the bottom. IB, immunoblotting. & 2015 Macmillan Publishers Limited

NEDD4-1 positively regulates Mdm2 stability and function C Xu et al

347

Figure 5. NEDD4-1 contributes to regulation of the p53 functionality. (a) P21-luciferase assay for the effects of NEDD4-1 overexpression on Mdm2-mediated p53 inhibition. Nedd4-1 KO MEFs were transfected with 20 ng of p21-luciferase with 100 ng of p53 with or without 200 ng of Hdm2 and increasing amounts of NEDD4-1 (N4-1, 200 or 400 ng). Twenty-four hours later, cell lysates were used for luciferase assay. Numbers on top of each column are fold change on p21-luciferase basal activity. (b) Western blot (WB) analysis of the effects of Nedd4-1 status on DNA damage-induced p53 accumulation and activation of Mdm2 and p21 expression. The MEFs from WT or Nedd4-1  /  were treated with 200 ng/ml NCS for indicated hours (h) and cell lysates were subjected to WB analysis for the indicated proteins. Fold changes of p21 normalized on the control level of WT MEFs are presented at the bottom. (c) WB analysis of the effects of Nedd4-1 status on the p53 response to doxorubicin (Dox), as performed in (b). Fold changes of p53 normalized on the detectable p53 band of WT MEFs 1 h after Dox are presented at the bottom. (d) WB analysis of the effects of knockdown of endogenous NEDD4-1 in U2OS or HCT116 cells on levels of p53, Hdm2 and p21. Knockdown of the target genes was performed using NEDD4-1 siRNA or control siRNA with or without siHdm2. (e) WB analysis of the effects of inducible knockdown of endogenous NEDD4-1 in U2OS-iN4i cells on NCS-induced p53 accumulation and expressions of Hdm2 and p21. NEDD4-1-inducible knockdown U2OS cells (U2OSiN4i) pretreated with 5 mg/ml of tetracycline for 72 h (Tet þ ) or without tetracycline (Tet  ) were treated with 200 ng/ml of NCS for the indicated hour (h) and the cell lysates were subjected to WB analysisfor the indicated proteins. Fold changes of p21 normalized on the control level of WT MEFs are presented at the bottom. (f ) WB analysis of the effects of inducible knockdown of endogenous NEDD4-1 in U2OS-iN4i cells on Hdm2 and p53, as performed in (e) but with earlier time points. Hdm2(lo), the same result of Hdm2 of short exposure to reveal downregulation of Hdm2 levels at 30 min and 1 h time points.

shown that ATM-dependent phosphorylation of CKId drives its nuclear accumulation, and phosphorylation of Mdm2 and its subsequent degradation is regulated by SCF(b-TrCP)-mediated ubiquitination.19 These findings strongly support the idea that SCF(b-TrCP) might be the E3 ligase responsible for DNA damage-induced Mdm2 degradation independent of Mdm2 autoubiquitination. In this report, we identified NEDD4-1 as a novel E3 ligase for Mdm2 through biochemical purification of activities that ubiquitinate RING domain mutant Mdm2. Our approach is unbiased, independent of stable protein interaction and based on enzymatic activity assay for E3 identification in chromatographic fractions. This approach has been successfully applied in identification of hPRC1L (human polycomb repressive complex 1like) E3 complex for H2A ubiquitination, MULE as an E3 ligase for Mcl-1 and NEDD4-1 as a PTEN E3 ligase.20–22 In this study, we used cytosolic fraction as our starting material for purification; WT ubiquitin in our assay catches the collection of all E3 ligase activities in the cytoplasm but not the nucleus, and thus excluding the reported PCAF activity as PCAF is a nuclear protein. Depletion of Cullin-1 from our peak fractions to remove Cullin1-based SCF & 2015 Macmillan Publishers Limited

(including SCF(b-TrcP)) activities did not decrease MELA activity (data not shown). Because the absence of NEDD4-1 in cell lysates caused marked decrease in MELA activity toward Mdm2 (Figure 3a), we conclude that NEDD4-1 is the major cytosolic E3 activity that mediates K63 ubiquitination of Mdm2 in non-stressed condition. NEDD4-1 as a new E3 ligase for Mdm2 was established in this study through several lines of evidence: NEDD4-1 can ubiquitinate both WT and enzyme-dead Mdm2 in concentrationand time-dependent manners, and NEDD4-1 physically interacts with Hdm2 and its overexpression enhances Mdm2 ubiquitination in cells, whereas the cell lysates from Nedd4-1 KO MEFs have significantly reduced E3 ligase activity toward enzyme-dead Mdm2 in vitro (Figures 3 and 4). Different from PCAF and SCF(b-TrCP), the effects of NEDD4-1mediated ubiquitination exerts stabilizing effects on Mdm2 by increasing Mdm2 protein stability (Figure 4f), apparently antagonizing the degradable ubiquitin chain formation on Mdm2 (Figure 4e) by Mdm2/MdmX and other E3 ligases. This serves as another layer of regulation to maintain the stable levels of cellular Mdm2 protein. We found that Mdm2 RING domain is involved in physical interaction with NEDD4-1 (Figure 3e). However, whether Oncogene (2015) 342 – 350

NEDD4-1 positively regulates Mdm2 stability and function C Xu et al

348 this interaction is affected by DNA damage signaling needs further investigation. The NEDD4-1-mediated regulation of Mdm2 abundance under normal growth conditions affects the p53 levels and the p53 DNA-damage response. Loss of this regulatory mechanism

has evident effects on basal Mdm2 and p53 abundance and the p53 response to DNA damage in terms of p53 accumulation and its downstream gene induction, as demonstrated in Nedd4-1 KO MEFs and U2OS cells after inducible knockdown of NEDD4-1. Moreover, NEDD4-1 overexpression decreases p53 levels (Figure 4d) and inhibits p53-dependent transcription in luciferase reporter assay (Figure 5a), and loss of NEDD4-1 increases G1 arrest after DNA damage (Figure 6a) and reduces cell growth rate in a p53-dependent manner (Figure 6c). These data support a role of NEDD4-1, via its effect on Mdm2 stabilization, as a novel regulatory component of the p53 pathway. Recent studies indicate that NEDD4-1 is overexpressed in non-small-cell-lung carcinoma and colorectal cancer.23,24 Interestingly, NEDD4-1 overexpression correlates with downregulation of the PTEN pathway in nonsmall-cell-lung carcinoma but not in colorectal cancer, suggesting a cell-type-specific effect of NEDD4-1 on cancer pathways. Therefore, NEDD4-1 overexpression in colon cancer might contribute to inhibition of the p53 pathway via its effect on Mdm2 stabilization. Of note, Mdm2 promotes cancer development through p53dependent and -independent mechanisms.25 Therefore, the NEDD4-1 effect on Mdm2 stabilization may have broader effects beyond the p53 pathway in promoting cancer progression. Our studies have identified multiple substrates for NEDD4-1 E3 ligase that are critically involved in cancer. NEDD4-1 negatively regulates PTEN tumor suppressor function, while it positively regulates Mdm2 abundance and phospho-AKT nuclear trafficking in IGF1 signaling.22,26 The collective effects of NEDD4-1 in the PTEN/AKT and p53 pathways may explain why NEDD4-1 is oncogenic in certain cell types such as immortalized human astrocytes, in which upregulation of NEDD4-1 by FoxM1B leads to cellular transformation and full malignant phenotype.27 In this context, NEDD4-1 may be a promising drug target for development of new cancer therapies.

MATERIALS AND METHODS Cell culture, chemicals and treatment Nedd4-1 þ / þ and Nedd4-1  /  primary MEFs (from Dr Baoli Yang at Carver College of Medicine, University of Iowa, Iowa City, IA, USA) and p53/ Mdm2 DKO MEFs28 (from Gigi Lozano, MD Anderson Cancer Center, Houston, TX, USA) were maintained in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal calf serum (Atlanta Biologicals Inc., Norcross, GA, USA) and antibiotics. Transfection was carried out with Lipofectamine 2000 (Invitrogen, Grand Island, NY, USA). Proteasome inhibitor MG-132 was purchased from Calbiochem, La Jolla, CA, USA (catalog number (cat. no.) 474790). MG132 treatment was carried out at 25 mM for 4 h. Jurkat cells used in biochemical purification of MELA were obtained from National Cell Culture Center (Minneapolis, MN, USA). HCT116 and HCT116-p53  /  cells (from Bert Vogelstein, Johns Hopkins University, Baltimore, MD, USA) were maintained Dulbecco’s modified Eagle’s medium with 10% fetal calf serum and the –antibiotics.

Figure 6. NEDD4-1 contributes to DNA damage-induced cell cycle arrest and p53-dependent cell growth inhibition. (a) Cell cycle analysis of the effects of Nedd4-1 status on NCS-induced G1 arrest. WT or Nedd4-1 KO (N4  /  ) MEFs were treated with 200 ng/ml NCS for 24 h, followed by flow cytometry analysis of cell cycle distribution. The cell cycle profiles are shown. (b) Bar graph of fold changes in G1/S ratio (the numbers on top of each column) before and 24 h after NCS treatment. (c) Effects of NEDD4-1 knockdown on p53-dependent and -independent cell growth. U2OS cells or U2OS cells with p53 knockdown by pLKO.1shp53 (U2OS-p53-null) were transfected with pSuperior-LacZi (LacZi) or pSuperior-NEDD4-1i (N4i). Then cell growth rate was monitored daily for 5 days. Fold change of cell number was presented. Reduced growth rate (Rg in %) by NEDD4-1 knockdown was normalized on LacZi control of the same time point. The changes in Rg between U2OS and U2OS-p53null cells on days 3, 4 and 5 are significant (Po0.00001). PI, propidium iodide. Oncogene (2015) 342 – 350

& 2015 Macmillan Publishers Limited

NEDD4-1 positively regulates Mdm2 stability and function C Xu et al

349 Plasmids and oligos NEDD4-1 expression plasmids including pcDNA3.1-dHA-NEDD4-1 and pcDNA3.1-NEDD4-1C867S were described previously.26 The E3-dead pGEX-6P-1Hdm2CA (C461A,C464A), E2-binding mutant pGEX-6P-1Hdm2L468A and pcDNA3.1-Hdm2 deltaRING (DelR, del439-491) were generated by site-directed mutagenesis using pGEX-6P-1-Hdm2 as a template (a gift from Dr Yosef Shiloh, Tel Aviv University, Tel Aviv, Israel). The mammalian expression plasmid for Hdm2, pCHDM1B, was a kind gift from Dr Jiandong Chen (H Lee Moffitt Cancer Center and Research Institute, Tampa, FL, USA), and pCHDM1B-L468A and pCHDM1B-Hdm2B (del28-299) were generated using site-directed mutagenesis. All the mutations were confirmed by DNA sequencing. His-ubiquitin plasmid (pMT107) was a gift from Dr Dirk P Bohmann (University of Rochester Medical Center, Rochester, NY, USA). Hdm2 small interfering RNA (siRNA) (HSS142909; Invitrogen), control siRNA (Dharmacon; ON_TARGET SMART pool siControl; cat. no. L-001816-10) or siRNA for NEDD4-1 (Dharmacon; ON_TARGET SMART pool NEDD4-1; cat. no. L-007178-00) were prepared and used as per the manufacturer’s instructions.

Antibodies Mouse monoclonal antibody for Mdm2 (2A10, 2A9 and 4B11) and for p53 (PAb421) were kind gifts from Dr Moshe Oren (Weizmann Institute of Science, Rehovot, Israel). Rabbit monoclonal poly(ADP-ribose) polymerase (cat. no. 9532) was purchased from Cell Signaling Technology (Danvers, MA, USA). Rabbit polyclonal NEDD4-1 antibody (cat. no. 07-049) was purchased from Millipore (Billerica, MA, USA). Lamin B1 (cat. no. ZL-5) and integrin-a (cat. no. sc-271034) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). g-Tubulin (cat. no. T6557), a-tubulin (cat. no. T9026) and a-actin (cat. no. A2066) were purchased from Sigma (St Louis, MO, USA). Monoclonal anti-HA antibody was purchased from Covance (Princeton, NJ, USA) (HA.11; Clone 16B12). Mouse GFP antibody was obtained from Roche (Nutley, NJ, USA) (cat. no. 11814460001; a mixture of 7.1 and 13.1). Monoclonal anti-ubiquitin antibody was purchased from BD (Franklin Lakes, NJ, USA) (cat. no. 550944).

Cytoplasmic and nuclear fractionation with NP40 detergentcontaining buffer Cytosolic and nuclear fractionation was carried out as described previously.26

Cycloheximide chase experiment The DKO MEFs (50–80% confluence in 60 mm plates) were transfected with Hdm2 or Hdm2CA expression plasmids with Lipofectamine 2000 (Invitrogen). After 24 h, the cells were cultured in 5 ml of complete Dulbecco’s modified Eagle’s medium containing 50 mg/ml of cycloheximide at 37 1C for 15, 30, 60 and 90 min before harvest. The cells were washed with cold phosphate-buffered saline and lysed with lysis buffer (50 mM Tris (pH 8.0), 0.5% Triton X-100 supplemented with protease inhibitors). Then the cell lysates were analyzed by western blotting for Hdm2 using a mixture of 2A9 and 4B11 antibodies. The band intensities of Hdm2 were quantified by IMAGEJ software (NIH, Bethesda, MD, USA) and the Hdm2 t1/2 values were obtained from the best-fit median-effect plots using CompuSyn software (ComboSyn Inc, Paramus, NJ, USA).29 Student’s t-tests were used to define significance of differences in t1/2 obtained from three independent experiments and presented as mean±s.e.m.

In vivo Mdm2 ubiquitination In vivo Mdm2 ubiquitination assay was carried out with Mdm2/p53 DKO MEFs as described previously,26 with minor modification for cell lysate preparation. Briefly, the cellular protein samples were prepared by directly lysing cells in plates with 200 ml of urea buffer A (9 M urea, 0.1 M phosphate, 0.01 M Tris (pH 8.0), 15 mM imidazole, 10 mM NaF, 2 mM Na3VO4, 0.2% Triton X-100), scraped and collected to microtubes (B300 ml), followed by sonication on the Bioruptor for 10 min. After centrifugation at 22 000 g for 10 min, 50 ml of the supernatant was saved for direct western blot analysis and the rest was used for His-tag-pulldown (Dynabeads, Invitrogen), followed by western blot analysis with a mixture of monoclonal Hdm2 antibodies 2A9 and 4B11.

In vitro ubiquitination of Mdm2 For in vitro assay of MELA activity, the reaction was carried out at 30 1C for 1 h in a volume of 20 ml containing 40 mM Tris-HCl (pH 7.5), 2 mM & 2015 Macmillan Publishers Limited

dithiothretiol, 5 mM MgCl2, 70 mM of ubiquitin, 50 nM E1, 400 nM UbCH5c, 5 mM ATP (Sigma; cat. no. A-7699). Different enzyme-dead Mdm2 substrate proteins were used in different experiments during the assay setup stage: 0.5 mg of recombinant GST-Hdm2CA (C461A, C464A) produced in Escherichia coli and affinity purified with glutathione sepharose (GE; cat. no. 17075601) or 40 ng of double-HA-tagged Hdm2 CA (C461A, C464A) (dHA-Hdm2CA) purified from baculovirus-infected sf9 insect cells as a substrate. In the case of GST-Hdm2CA as a substrate, the ubiquitinated Mdm2 species were detected by the smearing patterns of ubiquitinated GST-Hdm2CA on western blot analysis using an anti-ubiquitin antibody (BD) after GST-pulldown following our previously described protocol.22 While in the case of dHA-Hdm2CA as a substrate, Mdm2 ubiquitination was detected by smearing pattern obtained by western blot analysis for HA-tag. For characterization of recombinant NEDD4-1 as an MELA in vitro, 2.75 mg of enzyme-dead GST-Hdm2L468A in total volume of 40 ml was used as a substrate. The use of GST-Hdm2L468A mutant in later characterization experiments is because GST-Hdm2L468A serves equally as an E3-dead Hdm2 but adopts intact RING domain because the point mutation results in E2-binding deficiency without affecting RING structure.9,30 Recombinant NEDD4-1 was purified from baculovirus-infected sf9 insect cells as described previously.22

Purification and identification of NEDD4-1 as MELA1 from Jurkat S-100 All purification steps were carried out at 4 1C, and chromatography was performed with an Amersham FPLC system, AKTA purifier 100/10 (Amersham, GE Healthcare Bio-Sciences, Pittsburgh, PA, USA). Jurkat cytosolic fraction (JS100) was obtained by homogenizing Jurkat cells in Buffer A (20 mM Tris-HCl (pH 7.5), 50 mM NaCl, 1 mM dithiothretiol) supplemented with protease inhibitors, followed by centrifugation at 100 000 g for 2 h at 4 1C. The MELA activity assay was monitored by in vitro Mdm2 ubiquitination using dHA-Hdm2CA as a substrate. For the purification, 40 ml of JS100 (402 mg total protein) diluted with Buffer A into 100 ml was applied to a Q-Sepharose column (5-ml bed volume) equilibrated with Buffer A. After washing the column with Buffer A containing 0.15 M NaCl, the eMad activity was eluted into 40 ml of Buffer A containing 0.30 M NaCl (Q30, 84 mg protein). The activity was then subjected to 30% (saturation) ammonium sulfate (AS) precipitation. The protein pellet, which contained the activity, was dissolved in 40 ml Buffer A (AS30P, 15.8 mg protein). After dialysis against Buffer A for 2 h, the activity was run through a 5-ml HiTrap SP-Sepharose column (Amersham, GE Healthcare Bio-Sciences) equilibrated with Buffer A. After washing the column with Buffer A containing 0.15 M NaCl, the MELA was eluted into 12 ml of Buffer A containing 0.40 M NaCl (SP40BP, 5.7 mg). The eluate was adjusted to 7.5% saturation of AS and run through a 5-ml HiTrap PhenylSepharose column equilibrated with 10% saturation of AS in Buffer A. After washing the column with 7.5% AS, the activity was eluted with a 7.5–0% gradient of AS in 36 ml Buffer A, followed by 28 ml Buffer A. Sixteen fractions of 4 ml were collected, dialyzed and assayed for activity. The fractions containing the activity peak were pooled (f9–12, 0.76 mg protein) and further fractionated with a Mono Q column (Amersham, GE Healthcare Bio-Sciences) with a 150–300 mM NaCl gradient in 20 ml Buffer A. Fractions of 1 ml were collected. After dialysis, activity assay was performed using 2 ml of each fraction, and the MELA activity resides in fraction 8 (MQ8). A 0.5 activity ml of MQ8 was subjected to sucrose gradient (15–35%) centrifugation at 37 500 g for 16.5 h at 4 1C on OPTIMA MAX-E Ultracentrifuge (Beckman Coulter Inc., Indianapolis, IN, USA) in an MLS 50 rotor. The activity was recovered into 14 fractions at 0.3 ml per fraction. The MELA activity assay were performed using 4 ml of each fraction and 25 ml /fraction was loaded on an 8% sodium dodecyl sulfate–polyacrylamide gel electrophoresis, followed by silver staining. The remaining Mono Q fractions (MQ7 and MQ8) were subjected to protein identity determination by performing mass spectrometry analysis (MALDI-TOF-MS/MS).

Co-immunoprecipitation of Hdm2 with NEDD4-1 PC3 cells were transfected with pCDNA3.1-dHA-NEDD4-1 plasmid and full-length Hdm2 or truncated Hdm2. Twenty-four hours after transfection, the cells were lysed with 400 ml lysis buffer (phosphate-buffered saline (pH 7.4) containing 20% glycerol, 10 mM dithiothreitol, 0.5% NP40 and proteinase inhibitor cocktail). The supernatants were incubated with anti-HA antibody or mouse IgG in the lysis buffer for 1 h at 4 1C. Then the lysates were transferred to bovine serum albumin-blocked protein-G DYNA beads and further incubated at 4 1C for 1 h. After three times Oncogene (2015) 342 – 350

NEDD4-1 positively regulates Mdm2 stability and function C Xu et al

350 washing with lysis buffer, the proteins bound to the beads were eluted by sodium dodecyl sulfate sample buffer and subjected to western blot analysis for Hdm2.

P21-luciferase activity assay Nedd4-1  /  MEFs in a 3.5 cm plate were transfected with pGL3-p21luciferase and pCMV-p53 and pCHDM1B-HDM2 and pCDNA3.1-dHANEDD4-1 plasmids with the aid of Lipofectamine 2000 (Invitrogen; cat. no. 11668-019). Cells were lysed 24 h later with 100 ml Passive Lysis Buffer (Promega, Madison, WI, USA; Part cat. no. E194A) at room temperature for 20 min. The cell lysates were pipetted up and down three times and followed by centrifugation at 16 000 g at 4 1C for 3 min. The supernatants were transferred to another tube and frozen in liquid nitrogen immediately and stored at  80 1C. For the luciferase activity assay, the samples were thawed quickly in water bath, 50 ml of luciferase substrate and 15 ml of samples were added into 96-well LUMITRAC 200 white immunology plate (USA Scientific Inc., Ocala, FL, USA) and mixed with pipette, and the luciferase activity was immediately measured on Turner Biosystems Veritas Microplate Luminometer (Conquer Scientific, San Diego, CA, USA).

Cell cycle analysis WT and Nedd4-1-null MEFs were treated with NCS at 200 ng/ml for 24 h. Then both cells were harvested by trypsinization at a confluence of about 50–60%, and fixed in 70% EtOH for 1 h. The cells were then pelleted by centrifugation and resuspended in 500 ml of phosphate-buffered saline containing 50 mg/ml propidium iodide and 20 mg/ml RNase A. The content of propidium iodide was analyzed by a flow cytometer (BD LSRII), and cell cycle analysis was performed using the Modfit LT software (Verity Software House, Topsham, ME, USA).

Cell proliferation assay U2OS cells were infected with lentivirus-expressing GFP short hairpin RNA (control) or p53 short hairpin RNA for three rounds (8 h each round) in the presence of 10 mg/ml polybrene. Positive cells were selected in the media containing 2 mg/ml puromycin. After examination of the stable U2OS cells with p53 knockdown by western blot analysis, the control cells and p53 KO U2OS cells were transfected with pSuperior-NEDD4-1i (target sequence: 50 TggCgATTTgTAAACCgAA-30 ) or pSuperior-LacZi plasmid. Then the transfected cells were seeded in 96-well plates for cell number test by Cell Counting Kit-8 (CCK8; Dojindo Molecular Technologies Inc., Rockville, MD, USA). The cell number on day 1 was detected 6 h after seeding, and the value was taken as onefold at first. The values of CCK8 in the following 2–5 days were normalized on the reading of first day to get the fold increase of cell growth.

CONFLICT OF INTEREST The authors declare no conflict of interest.

ACKNOWLEDGEMENTS This work was supported in part by start-p funds from Roswell Park Cancer Institute (XW). The Roswell Park Core Grant CA16056 is also gratefully acknowledged. We sincerely thank Dr Baoli Yang for providing Nedd4-1  /  and parental MEFs; Dr Yosef Shiloh and Dr Jiandong Chen and Dr Dirk P Bohmann for providing DNA constructs; Dr Moshe Oren for antibodies, Bert Vogelstein for providing HCT116p53  /  cells and Dr Gigi Lozano for p53/Mdm2 DKO MEFs, respectively. We also thank Dr David Goodrich and other faculty colleagues at the Department of Pharmacology and Therapeutics, Roswell Park Cancer Institute for their critical reading of the manuscript.

REFERENCES 1 Levine AJ, Oren M. The first 30 years of p53: growing ever more complex. Nat Rev Cancer 2009; 9: 749–758. 2 Wade M, Wang YV, Wahl GM. The p53 orchestra: Mdm2 and Mdmx set the tone. Trends Cell Biol 2010; 20: 299–309. 3 Yang JY, Zong CS, Xia W, Wei Y, Ali-Seyed M, Li Z et al. MDM2 promotes cell motility and invasiveness by regulating E-cadherin degradation. Mol Cell Biol 2006; 26: 7269–7282.

4 Khor LY, Bae K, Paulus R, Al-Saleem T, Hammond ME, Grignon DJ et al. MDM2 and Ki-67 predict for distant metastasis and mortality in men treated with radiotherapy and androgen deprivation for prostate cancer: RTOG 92-02. J Clin Oncol 2009; 27: 3177–3184. 5 Zhou M, Yeager AM, Smith SD, Findley HW. Overexpression of the MDM2 gene by childhood acute lymphoblastic leukemia cells expressing the wild-type p53 gene. Blood 1995; 85: 1608–1614. 6 Marks DI, Kurz BW, Link MP, Ng E, Shuster JJ, Lauer SJ et al. High incidence of potential p53 inactivation in poor outcome childhood acute lymphoblastic leukemia at diagnosis. Blood 1996; 87: 1155–1161. 7 Fang S, Jensen JP, Ludwig RL, Vousden KH, Weissman AM. Mdm2 is a RING fingerdependent ubiquitin protein ligase for itself and p53. J Biol Chem 2000; 275: 8945–8951. 8 Lai Z, Ferry KV, Diamond MA, Wee KE, Kim YB, Ma J et al. Human mdm2 mediates multiple mono-ubiquitination of p53 by a mechanism requiring enzyme isomerization. J Biol Chem 2001; 276: 31357–31367. 9 Wang X, Wang J, Jiang X. MdmX protein is essential for Mdm2 protein-mediated p53 polyubiquitination. J Biol Chem 286: 23725–23734. 10 Wang X, Michael D, de Murcia G, Oren M. p53 Activation by nitric oxide involves down-regulation of Mdm2. J Biol Chem 2002; 277: 15697–15702. 11 Stommel JM, Wahl GM. Accelerated MDM2 auto-degradation induced by DNAdamage kinases is required for p53 activation. EMBO J 2004; 23: 1547–1556. 12 Itahana K, Mao H, Jin A, Itahana Y, Clegg HV, Lindstrom MS et al. Targeted inactivation of Mdm2 RING finger E3 ubiquitin ligase activity in the mouse reveals mechanistic insights into p53 regulation. Cancer cell 2007; 12: 355–366. 13 Linares LK, Kiernan R, Triboulet R, Chable-Bessia C, Latreille D, Cuvier O et al. Intrinsic ubiquitination activity of PCAF controls the stability of the oncoprotein Hdm2. Nat Cell Biol 2007; 9: 331–338. 14 Inuzuka H, Tseng A, Gao D, Zhai B, Zhang Q, Shaik S et al. Phosphorylation by casein kinase I promotes the turnover of the Mdm2 oncoprotein via the SCF(beta-TRCP) ubiquitin ligase. Cancer Cell 2010; 18: 147–159. 15 Ciechanover A. The ubiquitin-proteasome pathway: on protein death and cell life. EMBO J 1998; 17: 7151–7160. 16 Chen ZJ, Sun LJ. Nonproteolytic functions of ubiquitin in cell signaling. Mol Cell 2009; 33: 275–286. 17 Mosesson Y, Chetrit D, Schley L, Berghoff J, Ziv T, Carvalho S et al. Monoubiquitinylation regulates endosomal localization of Lst2, a negative regulator of EGF receptor signaling. Dev Cell 2009; 16: 687–698. 18 Oren M. Regulation of the p53 tumor suppressor protein. J Biol Chem 1999; 274: 36031–36034. 19 Wang Z, Inuzuka H, Zhong J, Fukushima H, Wan L, Liu P et al. DNA damageinduced activation of ATM promotes beta-TRCP-mediated Mdm2 ubiquitination and destruction. Oncotarget 2012; 3: 1026–1035. 20 Wang H, Wang L, Erdjument-Bromage H, Vidal M, Tempst P, Jones RS et al. Role of histone H2A ubiquitination in Polycomb silencing. Nature 2004; 431: 873–878. 21 Zhong Q, Gao W, Du F, Wang X. Mule/ARF-BP1, a BH3-only E3 ubiquitin ligase, catalyzes the polyubiquitination of Mcl-1 and regulates apoptosis. Cell 2005; 121: 1085–1095. 22 Wang X, Trotman LC, Koppie T, Alimonti A, Chen Z, Gao Z et al. NEDD4-1 is a proto-oncogenic ubiquitin ligase for PTEN. Cell 2007; 128: 129–139. 23 Amodio N, Scrima M, Palaia L, Salman AN, Quintiero A, Franco R et al. Oncogenic role of the E3 ubiquitin ligase NEDD4-1, a PTEN negative regulator, in non-smallcell lung carcinomas. Am J Pathol 2010; 177: 2622–2634. 24 Eide PW, Cekaite L, Danielsen SA, Eilertsen IA, Kjenseth A, Fykerud TA et al. NEDD4 is overexpressed in colorectal cancer and promotes colonic cell growth independently of the PI3K/PTEN/AKT pathway. Cell Signal 2013; 25: 12–18. 25 Li Q, Lozano G. Molecular pathways: targeting mdm2 and mdm4 in cancer therapy. Clin Cancer Res 2013; 19: 34–41. 26 Fan CD, Lum MA, Xu C, Black JD, Wang X. Ubiquitin-dependent regulation of phospho-AKT dynamics by the ubiquitin E3 ligase, NEDD4-1, in the insulin-like growth factor-1 response. J Biol Chem 2013; 288: 1674–1684. 27 Dai B, Pieper RO, Li D, Wei P, Liu M, Woo SY et al. FoxM1B regulates NEDD4-1 expression, leading to cellular transformation and full malignant phenotype in immortalized human astrocytes. Cancer Res 2010; 70: 2951–2961. 28 Montes de Oca Luna R, Wagner DS, Lozano G. Rescue of early embryonic lethality in mdm2-deficient mice by deletion of p53. Nature 1995; 378: 203–206. 29 Chou TC. Theoretical basis, experimental design, and computerized simulation of synergism and antagonism in drug combination studies. Pharmacol Rev 2006; 58: 621–681. 30 Linke K, Mace PD, Smith CA, Vaux DL, Silke J, Day CL. Structure of the MDM2/MDMX RING domain heterodimer reveals dimerization is required for their ubiquitylation in trans. Cell Death Differ 2008; 15: 841–848.

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

Oncogene (2015) 342 – 350

& 2015 Macmillan Publishers Limited

Regulation of Mdm2 protein stability and the p53 response by NEDD4-1 E3 ligase.

Mdm2 is a critical negative regulator of the tumor suppressor protein p53. Mdm2 is an E3 ligase whose overexpression leads to functional inactivation ...
2MB Sizes 2 Downloads 0 Views