(2008) 27, 4056–4064 & 2008 Nature Publishing Group All rights reserved 0950-9232/08 $30.00 www.nature.com/onc ORIGINAL ARTICLE VprBP targets to the Roc1-Cul4A-DDB1 E3 ligase complex for degradation

J Huang1 and J Chen1

Department of Therapeutic Radiology, Yale University School of Medicine, New Haven, CT, USA

Inactivation of the neurofibromatosis type 2 (NF2) tumor schwannomas (Gutmann, 1997). The NF2 encodes suppressor gene function has been observed not only in a 595 amino-acid Merlin (schwannomin), which familial schwannomas and other central nervous system isrelated to the --(ERM) protein tumors, but also in malignant tumors unrelated to the NF2 family and washence dubbed Merlin for Moesin,Ezrin, syndrome, indicating a broader role of NF2 in Radixin-like protein (Rouleau et al., 1993; Trofatter tumorigenesis. The NF2-encoded protein Merlin is closely et al., 1993). Merlin is putative tumor suppressor that is related to the Ezrin-Radixin-Moesin family of membrane/ anchored at where it linkstransmem- linker , and has been demonstrated brane receptorsto the cytoskeleton (Xu and to suppress tumor growth by inhibiting extracellular Gutmann, 1998; den Bakker et al., 2000). Merlin can signal-regulated (ERK) and Rac1 activation. form homo- and heterotypic interactionswhich in turn Interestingly, serum deprivation has been shown to regulate itsbinding to other proteins.Unlike Merlin in regulate Merlin at the protein level, however, exactly an open conformation, head-to-tail binding of Merlin, how such condition affects Merlin remains elusive. In this which leadsto a closedconformation, isthought to study, we provide evidence to show that Merlin is act as a tumor suppressor and regulates cell growth regulated in a Roc1-Cullin4A-DDB1-dependent manner. (Gonzalez-Agosti et al., 1999; Bretscher et al., 2000). Following serum stimulation, Merlin is recruited to the E3 of a C-terminal (S518) by - ligase complex through a direct interaction with the activated kinase (PAK)- or cAMP-dependent protein WD40-containing adaptor protein VprBP. Loading of kinase A weakens Merlin’s self-association and is Merlin to the E3 complex resulted in believed to inactivate its growth-suppressing activity. its polyubiquitination, and consequently its proteasome- On the other hand, the phosphatase MYPT-1-PP1-d mediated degradation. Consistently, VprBP depletion activates the tumor suppression activity of Merlin abolished the in vivo interaction of Merlin and Roc1- through S518 dephosphorylation, which subsequently Cullin4A-DDB1, which resulted in Merlin stabilization inhibits the Ras/ERK (extracellular signal-regulated and inhibited ERK and Rac activation. Together, our data kinase) pathway (Shaw et al., 2001; Alfthan et al., revealed a novel regulatory mechanism for the tumor 2004; Jin et al., 2006a). In addition, Merlin can also be suppressor function of Merlin. regulated at the protein level by growth arrest stimuli. Oncogene (2008) 27, 4056–4064; doi:10.1038/onc.2008.44; Under conditionsof serumdeprivation, Merlin protein published online 10 March 2008 level increases significantly (Shaw et al., 1998; Lee et al., 2006). However, the mechanism underlying the regula- Keywords: VprBP; Merlin; E3 ligase; degradation tion of Merlin stability under such condition is not well understood. The Roc1-CUL4-DDB1 ubiquitin E3 complex isa new class of -containing ubiquitin E3 ligases. The Cullin E3 ligases have been shown to target a number of Introduction proteinsfor ubiquitination-dependent proteolysisvia different adaptor proteins that afford specific substrate Inactivation of the neurofibromatosis type 2 (NF2) recognition (Higa and Zhang, 2007; Lee and Zhou, leads to the development of 2007). Recently, multiple WD40 domain-containing several types of tumors commonly found in the central proteinswere found to interact with DDB1 and serve nervous system (CNS), in particular, meningiomas and as the substrate-recognition subunits of the CUL4- DDB1 ubiquitin ligase complex (Angers et al., 2006; He Correspondence: Professor J Chen, Department of Therapeutic et al., 2006; Higa et al., 2006; Jin et al., 2006b). One Radiology, Yale University School of Medicine, Hunter Building of these WD40 domain-containing proteins, VprBP/ 213C, 333 Cedar Street, PO Box 208040, New Haven, CT 6520, USA. DCAF1, containsfour WD40 domainsand a LisH E-mail: [email protected] domain. VprBP wasoriginally identified due to its 1JH and JC designed experiments and interpreted the data; JH performed all experiments; JH and JC prepared the manuscript. interaction with the protein from the human Received 4 November 2007; revised 22 January 2008; accepted 2 immunodeficiency virus, however, the physiological February 2008; published online 10 March 2008 function and substrates of VprBP/DCAF1 are still VprBP targets Merlin for degradation J Huang and J Chen 4057 unknown (Zhang et al., 2001; Belzile et al., 2007; Table 1 List of proteins associated with SFB-Merlin identified Hrecka et al., 2007; Le Rouzic et al., 2007; Tan et al., by mass spectrometry analysis 2007; Wen et al., 2007). Protein name No. of peptides In this study, using a tandem affinity-purification obtained approach, we have identified VprBP and DDB1 astwo Merlin (Moesin-Ezrin-Radixin-like Protein) 98 major binding partnersof Merlin. Our data show that VPRBP (KIAA0800) 48 the Roc1-Cul4A-DDB1 E3 ligase complex controls the DNA damage-binding protein 1 41 stability of Merlin, which in turn regulates ERK and Heat shock 70 kDa protein 1 11 Rac activities. Merlin is recruited to the E3 ligase Heat shock 70 kDa protein 1l 8 through VprBP, where it ispolyubiquitinated and Homo sapiens trifunctional subunit-a, mi- 6 tochondrial precursor (tp-a) targeted for proteasome-dependent degradation. We Heat shock cognate 71 kDa protein 6 propose that the ubiquitination-dependent proteolysis 78 kDa glucose-regulated protein precursor (grp 78) 4 of Merlin providesa fine-tune mechanismfor itstumor Hypothetical protein DDA1 (DDA1) 4 suppressor activity. Zinc-finger ran-binding domain-containing protein 2 3 zinc-finger protein 265 Stress-70 protein, mitochondrial precursor (75 kDa 3 glucose-regulated protein) Dihydrolipoamide branched chain transacylase 2 Results Hypothetical protein LOC92345 2 Tubulin a-2 chain (a-tubulin 2) 2 Identification of VprBP and DDB1 as Merlin-associated Heat shock-related 70 kDa protein 2 (heat shock 2 proteins 70 kDa protein 2) Nucleolar phosphoprotein p130 (nucleolar 130 kDa 2 To identify Merlin-associated proteins, we purified protein) Merlin-containing complex from human embryonic Propanoyl-coa:carbon dioxide ligase subunit-b 1 kidney 293T cells stably expressing full-length Merlin Methylcrotonoyl-coa carboxylase-b chain 1 with an N-terminal triple tag that consists an S tag, a Tubulin a-6 chain 1 Flag epitope and a streptavidin-binding peptide (S-Flag- mRNA encoding b-tubulin 1 KIAA1604 protein 1 SBP (SFB)-Merlin). Two major protein bands, sizes of Tubulin, b-polypeptide 1 approximately 130 and 170 kDa, were co-purified with Serine/threonine-protein phosphatase 2A 55 kDa 1 SFB-Merlin. Mass spectrometry analysis revealed these regulatory subunit B a-isoform proteinsasDDB1 and VprBP, respectively (Figure 1a). Heat shock 70 kDa protein 6 1 RNA-binding protein with serine-rich domain 1 1 VprBP and DDB1 have been shown to assemble with RNA-binding protein 1 Cul4A and Roc1 to form an E3 ubiquitin ligase complex (Angers et al., 2006; He et al., 2006; Higa et al., 2006; Jin Proteins of interest are shown in bold. et al., 2006b). In addition, several peptides derived from DDA1, an additional component of the Cul4A ligase complex, were also identified from the purification but not with DDB1 (Figure 1c), indicating that the (Table 1). Since VprBP has been speculated to serve as direct interaction of Merlin and VprBP isrequired for a substrate-recognition subunit of the Cul4A-containing the association of Merlin with the E3 ubiquitin ligase ligase complex, we reasoned that VprBP might target complex in vivo. Coimmunoprecipitation experiments Merlin to the Roc1-Cul4A-DDB1 E3 ubiquitin ligase confirmed the interaction between endogenousMerlin complex. and VprBP (Figure 1d). Consistent with the proposed role of VprBP in targeting Merlin to the Roc1-Cul4A-DDB1 complex, Adaptor protein VprBP links Merlin to VprBP depletion abolished the interaction between Roc1-Cul4A-DDB1 E3ligase complex Merlin and Roc1-Cul4A-DDB1 complex (Figure 1e). To verify the associations between Merlin and Phosphorylation of Merlin at Ser518 was previously VprBP, DDB1 and other componentsof the Cul4A reported to play an important role in regulating ligase complex, we coexpressed Merlin with various Merlin tumor suppressor activity (Shaw et al., 2001; componentsof the Cul4A complex. Asshownin Alfthan et al., 2004; Jin et al., 2006a). Interestingly, the Figure 1b, Merlin interacted with VprBP and DDB1. interaction between Merlin and VprBP isindependent of Merlin also coimmunoprecipitated with Roc1 and Ser518 phosphorylation of Merlin (data not Cul4A, but not with an unrelated protein Rap80 shown), suggesting that the Merlin-VprBP interaction (Figure 1b). These data support the notion that Merlin might represent a new means for Merlin regulation interactswith the Roc1-Cul4A-DDB1-VprBP E3 ubi- in vivo. quitin ligase complex. Since WD40 domain-containing proteinshave been shownto interact directly with DDB1 and serve as the substrate-recognition subunits of The very C terminus of VprBP is required for its the Roc1-CUL4-DDB1 ubiquitin ligase, we suspected interaction with Merlin that VprBP might mediate the interaction between To determine the domain of VprBP that mediatesits Merlin and the Roc1-Cul4A-DDB1 E3 ligase complex. interaction with Merlin, we used a series of VprBP Indeed, bacterially expressed glutathione S-transferase deletion mutants(Figure 2a). While neither the LisHnor (GST)-tagged Merlin interacted robustly with VprBP the WD40 domain wasrequired for thisinteraction, the

Oncogene VprBP targets Merlin for degradation J Huang and J Chen 4058 VprBP C terminus encompassing residues 1311–1507 ing residues 1–225 and a small region in the middle was both necessary and sufficient for the binding of (residues 300–333) were capable of binding to VprBP, VprBP to Merlin (Figure 2b). indicating that there may be multiple binding sites Similarly, we generated a series of Merlin deletion on Merlin that are involved in itsinteraction with mutantsand examined which domainsof Merlin might VprBP (Figure 2d). The existence of multiple VprBP- be required for itsinteraction with VprBP (Figure 2c). binding regionson Merlin isreminiscent to We found that both the N terminusof Merlin-contain- the interaction between Merlin and PAK1 (Kissil

0 A 1234

SBP-VprBP SBP-DDB1 SBP-Cul4 SBP-Roc1 SBP-RAP8 - - - - - VprBP HA-merlin + + + + + IP S-beads IB α-HA DDB1 merlin

VprBP merlin DDB1 RAP80 Cul4A Total lysate Total Roc1

merlin

Flag-VprBP Flag-DDB1 IP α-merlin Con SiRNA

VprBP INPUT GST GST-merlin INPUT GST GST-merlin

IB α-Flag DDB1

VprBP

DDB1 Cul4A

Roc1 IP

merlin

Con-IgG Anti-VprBP Lysate IB α-merlin Con SiRNA

VprBP VprBP

merlin merlin Total lysate Total

Oncogene VprBP targets Merlin for degradation J Huang and J Chen 4059 et al., 2003), suggesting that Merlin may have evolved to were used to deplete VprBP and both of them led to interact with itspartnersthrough multiple significant downregulation of endogenous VprBP contact sites. expression (Figure 4a). Merlin protein levels were clearly increased in cells transfected with VprBP siRNAs when compared to those transfected with control siRNA The Roc1-Cul 4A-DDB1 ubiquitin ligase catalyses Merlin (Figure 4a). In agreement with a role of the Roc1- ubiquitination and degradation Cul4A-DDB1 complex in Merlin degradation, depletion Since VprBP targetsMerlin to the Roc1-Cul 4A-DDB1 of endogenous DDB1 also led to an increase in Merlin ubiquitin ligase, we next tested whether VprBP pro- protein level (Figure 4a). motesMerlin ubiquitination and degradation. Cells It waspreviouslyreported that Merlin protein level is were cotransfected with expression plasmids encoding upregulated following serum starvation (Shaw et al., Merlin, or an unrelated control protein Amot-p80, 1998). Stimulation of serum-starved cells resulted in a together with VprBP. Overexpression of VprBP resulted rapid disappearance of Merlin protein within 5–15 min in the downregulation of Merlin, but did not signifi- (Shaw et al., 1998). We first examined whether Merlin cantly affect the levelsof the unrelated protein downregulation following serum stimulation is mediated Amot-p80 (Figure 3a, left). Preincubation with the by the proteasome-dependent degradation pathway. As proteasome inhibitor MG132 alleviated VprBP-induced shown in Figure 4b, the addition of proteasome Merlin degradation (Figure 3a, right), suggesting that inhibitor MG132 efficiently blocked Merlin degradation VprBP promotesproteasome-dependent degradation following serum stimulation. Notably, Merlin protein of Merlin. To ascertain that the binding of VprBP levelsdid not change following serumaddition in to Merlin isrequired for VprBP-mediated degradation, VprBP-depleted cells (Figure 4c), suggesting that serum we examined the effect of a VprBP deletion mutant, stimulation-induced Merlin degradation depends on VprBP-N (residues 1–1311) that lacks the C-terminal VprBP. Merlin-binding region. Asshownin Figure 3b, over- Rac and ERK activitiesare tightly controlled in the expression of this VprBP mutant did not markedly alter cell. Merlin has been suggested to restrain tumor growth Merlin expression, establishing that the interaction through inhibiting Rac and ERK activities(Shaw et al., between VprBP and Merlin isimportant for Merlin 2001; Kissil et al., 2003; Okada et al., 2005; Morrison degradation. et al., 2007). To further substantiate a role of VprBP in To directly demonstrate that VprBP promotes Merlin Merlin function regulation, we tested whether Rac and ubiquitination, we transfected cells with plasmids ERK activitiesmight be altered in VprBP-depleted cells. encoding Flag-tagged Merlin and -tagged ubiquitin, As shown in Figure 4d, VprBP depletion, which resulted in combination with empty control plasmid or a plasmid in Merlin upregulation, was associated with a decrease encoding hemagglutinin (HA)-VprBP. Asshown in of Rac1 and ERK activities. Together, these data Figure 3c, VprBP significantly increased Merlin ubiqui- support that VprBP targets Merlin for degradation tination. Moreover, the addition of MG132 further following serum stimulation and thus regulates Rac and enhanced Merlin ubiquitination (Figure 3c). Taken ERK activitiesand cell proliferation. together, these observations suggest that VprBP negatively regulatesMerlin by targeting Merlin to the Roc1-Cul4A-DDB1 complex for ubiquitination and degradation. Discussion

VprBP is required for Merlin degradation following serum The NF2 tumor suppressor gene product, Merlin, is stimulation structurally related to the ERM family of proteins that To further explore the physiological role of the Roc1- anchor the actin cytoskeleton to specific membrane Cul4A-DDB1 ubiquitin ligase in Merlin degradation, we proteins. Inactivation of the NF2 tumor suppressor gene examined the steady-state levels of endogenous Merlin was first observed in familial schwannomas and other with or without VprBP depletion. Two different siRNAs CNS tumors. Notably, mice with heterozygous NF2

Figure 1 Identification of VprBP asthe mediator between Merlin and Roc1-Cul4A-DDB1 E3 ubiquitin complex. ( a) Coomassie blue staining of affinity-purified Merlin-containing protein complexes. Cell extracts prepared from 293T cells stably expressing S-Flag-SBP (SFB)-Merlin were subjected to two rounds of affinity purification. Crude cell lysate (lane 1), first eluent (lane 2), proteins bound to streptavidin-conjugated beads after elution (lane 3) and final eluent (lane 4) were analysed by SDS-polyacrylamide gel electrophoresis (PAGE) and visualized by Coomassie blue staining. Proteins were identified by matrix-assisted laser desorption/ionization time-of- flight (MALDI-TOF) mass spectrometry analysis and summarized in Table 1. Lines indicate protein bands corresponding to VprBP and DDB1. (b) The interaction between ectopically expressed Merlin and Roc1-Cul4A-DDB1-VprBP complex. Immunoprecipitation (IP) reactions were performed using S beads and were subjected to western blot analyses using anti-hemagglutinin (HA) or anti-Flag antibodies. (c)Direct interaction between Merlin and VprBP. Glutathione S-transferase (GST) or GST-Merlin immobilized on sepharose beads were incubated with cell lysates containing exogenously expressed Flag-tagged VprBP or DDB1. Bound VprBP or DDB1 was analysed by anti-Flag immunoblotting. (d) Interaction between endogenousMerlin and VprBP. IP reactionswere performed usingcontrol immunoglobulin G (IgG, Con-IgG) or anti-VprBP antibody. Western blotting was conducted with indicated antibodies. (e) VprBP mediatesthe interaction between Merlin and Roc1-Cul4A-DDB1 complex. HeLa cellswere treated with indicated siRNAsto specificallydownregulate VprBP. Cell lysateswere immunoprecipitated with anti-Merlin antibody and immunoblotted with the indicated antibodies.

Oncogene VprBP targets Merlin for degradation J Huang and J Chen 4060 VprBP LisH WD40 Merlin binding FL 1507 + D1 1000 - D2 1311 - D3 801 1311 - D4 801 + D5 901 1311 - D6 901 + D7 1000 + D8 1311 +

GST-Merlin pulldown Flag-VprBP FL D1 D2 D3 D4 D5 D6 D7 D8

Pulldown IB α-Flag

INPUT IB α-Flag

Merlin FERM α-HELIX VprBP binding FL 595 + D1 225 + D2 333 + D3 480 +

D4 226 + D5 300 +

D6 333 - D7 475 -

GST-VprBP-(1311-1507) pulldown

Flag-Merlin FL D1 D2 D3 D4 D5 D6 D7

Pulldown IB α-Flag

INPUT IB α-Flag

Figure 2 Determine the domainsrequired for the interaction between Merlin and VprBP. ( a, c) Schematic presentation of human VprBP, Merlin and their deletion mutants used in this study. (b, d) Mapping of regionsrequired for the Merlin/VprBP interaction. Beadscoated with glutathione S-transferase (GST)-Merlin (b) or GST-VprBP-C (residues 1311–1507; d) were incubated with cell lysates containing exogenously expressed S-Flag-SBP (SFB)-tagged wild-type (WT) VprBP or deletion mutants (b) or Merlin or deletion mutants( d). After extensive washing, bound Merlin or VprBP was analysed by western blotting using anti-Flag antibody.

developed a wide variety of malignant tumors, believed to play a critical role in regulating Merlin suggesting that NF2 has a much broader role in cancer activity. Moreover, studies have shown that Merlin is development and progression (McClatchey et al., 1998). also regulated at the protein level. Accordingly, under The phosphorylation status of Merlin at Ser518 is conditions of serum deprivation, Merlin expression

Oncogene VprBP targets Merlin for degradation J Huang and J Chen 4061 Myc-Merlin+HA-Amot-p80 (control) Myc-Merlin+HA-Amot-p80 (control)

MG132 - + Flag-VprBP 0 0.2 0.4 0.6 0.8 1.0 0 0.2 0.4 0.6 0.8 1.0

IB α-Flag VprBP

IB α-Myc Merlin

IB α-HA Amot-p80 (control)

IB α-actin

Flag-Merlin Myc-Merlin+HA-Amot-p80 (control) Myc-Ubi- + + + HA-VprBP --+ + Flag-VprBP-N 0 0.2 0.4 0.6 0.8 1.0 MG132 - --+ IB α-Flag VprBP-N

IB α-Myc IP α-Flag (Ub)n Merlin (Merlin)

α IB -HA IB α-Myc Amot-p80 (ubi) (control) IgG-H

α IB -actin IB α-Flag Merlin IB α-HA VprBP

Figure 3 VprBP targetsMerlin for proteasome-dependent degradation. ( a) Overexpression of VprBP leads to Merlin downregulation. 293T cells were transfected with the indicated expression constructs. Forty-eight hours after transfection, cells were treated with MG132 (10 mM) or left untreated for 6 h before they were collected. Cell lysates were analysed by western blotting using indicated antibodies. (b) VprBP mutant lacking the Merlin-binding region cannot promote Merlin degradation. Experimentswere performed similar to that described in (a). (c) VprBP targetsMerlin for ubiquitination in vivo. 293T cellswere transfectedwith indicated expression constructs. Forty-eight hours after transfection, cell lysates were immunoprecipitated (IP) with anti-Flag antibody and immunoblotted with anti-myc antibody (top) to detect ubiquitinated Merlin. Expression of Merlin and VprBP was detected by immunoblotting with anti-Flag (middle) and anti-HA (bottom) antibodies, respectively. All of the experiments were repeated at least three times and similar results were obtained. increases significantly, whereas it is downregulated in Merlin inhibitstumor formation through itsability to response to serum stimulation. However, exactly how interfere with the activation of Rac and ERK following Merlin protein level iscontrolled remainslargely serum stimulation. In agreement with this model, we unknown. showed that VprBP downregulation leads to an increase With the goal to understand the normal cellular in Merlin expression and inhibited Rac and ERK function of Merlin and how it isregulated in the cell, we activation, indicating that the Roc1-Cul4A-DDB1- utilized a tandem affinity-purification approach to VprBP complex may play a role in the control of cell isolate potential Merlin-interacting proteins. Here, we growth and transformation. In summary, our study describe the identification of a novel Merlin interaction revealed a previously unknown mechanism for the protein, the WD40-containing adaptor protein VprBP, control of Merlin expression in response to physiologi- which targetsMerlin to the Roc1-Cul4A-DDB1 E3 cal stimuli, which represents an additional layer of ligase complex for ubiquitination and subsequent Merlin regulation and the control of its tumor suppres- proteasome-dependent degradation. sion function in vivo.

Oncogene VprBP targets Merlin for degradation J Huang and J Chen 4062 VprBP DDB1 Control Control SiRNA SiRNA1 SiRNA2 SiRNA SiRNA1 SiRNA2

VprBP DDB1 VprBP Control merlin merlin SiRNA SiRNA1 SiRNA2 p-ERK1 p-ERK2 actin actin (Long exposure) p-ERK1 p-ERK2 - + MG132 (Short exposure) Serum 0 5 10 0 5 10 mins

ERK merlin

actin Rac1-GTP

VprBP SiRNA Control SiRNA Rac1 Serum 0 5 10 0 5 10 mins

VprBP merlin

actin

Figure 4 VprBP is required for Merlin degradation following serum stimulation. (a) siRNA-mediated downregulation of VprBP or DDB1 stabilizes endogenous Merlin protein. HeLa cells were treated with indicated siRNAs two times at 24-h intervals. Forty-eight hours after transfection, cells were harvested and western blot analysis was performed using antibodies as indicated. (b) Serum stimulation leads to Merlin ubiquitination and proteasome-dependent degradation. Serum-starved HeLa cells were treated with MG132 (10 mM) or left untreated for 6 h. Thereafter, 10% serum was added to the media for indicated time period before cells were collected. Cell lysates were prepared and analysed by western blotting using indicated antibodies. (c) VprBP isrequired for Merlin degradation following serum stimulation. HeLa cells were treated with VprBP siRNA or control siRNA twice at 24-h intervals. After the second transfection, cells were serum starved for 48 h and then 10% serum was added to the media for indicated time before cells were collected and analysed by western blotting. (d) Knockdown of VprBP inhibits Rac and extracellular signal-regulated kinase (ERK) activation following serum stimulation. Experiments were performed similar to that described in (c).

Materials and methods RPMI 1640 medium supplemented with 10% fetal bovine serum at 37 1Cin5%CO2 (v/v). For serum starvation Plasmids experiments, subconfluent cells were placed in 0.5% fetal Plasmids encoding human wild-type S518A and S518D mutant bovine serum for 48 h and then restimulated with normal of Merlin and wild-type VprBP were kind giftsfrom Dr KQ medium for the indicated time. Ye (Emory University School of Medicine) and Dr JW Harper (Harvard Medical School), respectively. The full-length or deletion mutantsof Merlin, VprBP, Roc1, Cul4A and DDB1 siRNA were generated by PCR and subcloned into the pDONR201 All siRNA duplexes used in this study were purchased from vector using Gateway Technology (Invitrogen, San Diego, CA, Dharmacon Research (Lafayette, CO, USA). The sequences of DDB1 siRNA1 and siRNA2, VprBP siRNA1 and siRNA2 are USA). For transient expression of Merlin, VprBP and their 0 0 0 mutants, their corresponding coding sequences in the entry respectively 5 -ACACTTTGGTGCTCTCTTTUU-3 ,5-AGA TCGCGATAATAAAGAAUU-30,50-GAUGGCGGAUGCU vector were transferred into a Gateway-compatible destination 0 0 vector, which harborsan N-terminal triple-epitope tag (S- UUGAUAUU-3 and 5 -UCACAGAGUAUCUUAGAGAU U-30. The sequence of control siRNA is 50-UUCAAUAAAUUC protein tag, Flag epitope tag and streptavidin-binding peptide 0 tag) or HA or Myc epitope tag. DNA fragmentsencoding UUGAGGUUU-3 . siRNAs were introduced into cells VprBP and Merlin were also generated by PCR and subcloned using oligofectamine according to the manufacturer’s protocol into pGEX-4T-1 vector (Pharmacia, Uppsala, Sweden) to (Invitrogen). generate expression constructs of GST-VprBP-C (residues 1311–1507) and GST-Merlin, respectively. Antibodies, transfection and immunoprecipitation procedures Anti-Merlin polyclonal antibodies sc331 and sc332 and anti- Cell culture, serum starvation and release ERK1 polyclonal antibody sc-94 were purchased from Santa HeLa and 293T cellswere purchasedfrom American Type Cruz Biotechnology (Santa Cruz, CA, USA). Anti-phospho- Culture Collection (Manassas, VA, USA) and maintained in ERK1/2 polyclonal antibody 9101S waspurchasedfrom Cell

Oncogene VprBP targets Merlin for degradation J Huang and J Chen 4063 Signaling Technology Inc., Danver, MA, USA, Anti-actin screened for the expression of tagged proteins by western and anti-Flag (M2) antibodieswere obtained from Sigma blotting. For affinity purification of tagged protein complexes, (St Louis, MO, USA). Anti-VprBP polyclonal antibody was 293T cells stably expressing tagged proteins were collected and a kind gift from Dr LJ Zhao (St LouisUniversitySchool lysed with NETN buffer on ice for 20 min. Crude lysates were of Medicine), and anti-Cul4A, anti-DDB1 and anti-Roc1 cleared by centrifugation at 14 000 r.p.m. at 4 1C for 10 min, polyclonal antibodieswere giftsfrom Dr Y Xiong (University and supernatants were incubated with 300 ml streptavidin- of North Carolina). Quantification of active Rac was conjugated beads (Amersham, Piscataway, NJ, USA). Beads performed using the Rac Activation Assay Kit, according to were washed three times with NETN buffer and then proteins manufacturer’s instructions (Upstate, Temecula, CA, USA). bound to beadswere eluted with 1 ml NETN buffer containing Transient transfection of expression constructs was performed 1mgmlÀ1 biotin (Sigma). Eluted supernatant was incubated using Fugene 6 transfection reagent according to manufac- with 80 ml S protein beads(Novagen, Gibbstown,NJ, USA). turer’s instructions (Roche Inc., Indianapolis, IN, USA). For S beadswere washedthree timeswith NETN buffer and immunoprecipitation, cells were washed with ice-cold phos- proteinswere eluted by boiling in samplingbuffer and phate-buffered saline and then lysed in NETN buffer (0.5% separated by SDS–PAGE. Coomassie blue staining was Nonidet P-40, 20 mM Tris(pH 8.0), 50 m M NaCl, 50 mM NaF, performed to visualize protein bands, which were excised, 100 mM Na3VO4, 1 mM DTT and 50 mgmlÀ1 PMSF) at 4 1C digested and subjected to mass spectrometry analysis. for 10 min. Crude lysates were cleared by centrifugation at 14 000 r.p.m. and 4 1C for 5 min, and supernatants were GST pulldown assay incubated with protein A-agarose-conjugated primary anti- S bodies. The immunocomplexes were washed three times with Glutathione -transferase fusion proteins were expressed in Escherichia coli NETN buffer and then subjected to SDS–polyacrylamide gel and purified as previously described (Hofer et al electrophoresis (PAGE). Western blotting was performed ., 1994). Two microgramsof GST-fusionprotein or GST using the antibodies specified in the figure legends. Total cell alone wasimmobilized on glutathione-Sepharose4B beadsand lysates were obtained by using radioimmunoprecipitation incubated with lysates prepared from cells that were transiently transfected with plasmids encoding indicated proteins for 2 h assay buffer (except for immunoprecipitation) (50 mM Tris at 4 1C. After washing with NETN buffer, the samples were (pH 8.0), 0.1% SDS, 150 mM NaCl, 1% Nonidet P-40, 0.5% sodium deoxycholate). analysed by western blotting analysis.

The establishment of stable cell lines and affinity purification Acknowledgements of SFB-tagged protein complexes To establish stably cell lines, 293T cells were transfected with We thank membersof the Chen Laboratory for helpful plasmids encoding SFB-tagged proteins. Forty-eight hours discussions and technical support. We also thank Michael after transfection, cells were split at 1:10 ratio and cultured in Huen and Jamie Wood for proofreading the paper. Thiswork the medium containing puromycin (2 mgmlÀ1) for 3 weeks. wassupportedby the DOD Era of Hope Scholar Award to JC Individual puromycin-resistant colonies were isolated and (W81XWH-05-1-0470).

References

Alfthan K, Heiska L, Gronholm M, Renkema GH, Carpen O. Higa LA, Wu M, Ye T, Kobayashi R, Sun H, Zhang H. (2006). CUL4- (2004). Cyclic AMP-dependent protein kinase phosphory- DDB1 ubiquitin ligase interacts with multiple WD40-repeat proteins lates Merlin at serine 518 independently of p21-activated kinase and regulateshistonemethylation. Nat Cell Biol 8: 1277–1283. and promotesMerlin-Ezrin heterodimerization. J Biol Chem 279: Higa LA, Zhang H. (2007). Stealing the spotlight: CUL4-DDB1 18559–18566. ubiquitin ligase docks WD40-repeat proteins to destroy. Cell Div 2: AngersS, Li T, Yi X, MacCossMJ,Moon RT, Zheng N. (2006). 5. Molecular architecture and assembly of the DDB1-CUL4A Hofer B, BackhausS, TimmisKN. (1994). The biphenyl/polychlori- ubiquitin ligase machinery. Nature 443: 590–593. nated biphenyl-degradation (bph) of Pseudomonas sp. LB400 Belzile JP, Duisit G, Rougeau N, Mercier J, Finzi A, Cohen EA. encodesfour additional metabolic . Gene 144: 9–16. (2007). HIV-1 Vpr-mediated G2 arrest involves the DDB1- Hrecka K, Gierszewska M, Srivastava S, Kozaczkiewicz L, CUL4A(VPRBP) E3 ubiquitin ligase. PLoS Pathog 3: e85. Swanson SK, Florens L et al. (2007). Lentiviral Vpr usurps Cul4- Bretscher A, Chambers D, Nguyen R, Reczek D. (2000). ERM-Merlin DDB1[VprBP] E3 ubiquitin ligase to modulate . Proc Natl and EBP50 protein familiesin plasmamembrane organization and Acad Sci U S A 104: 11778–11783. function. Annu Rev Cell Dev Biol 16: 113–143. Jin H, Sperka T, Herrlich P, Morrison H. (2006a). Tumorigenic Den Bakker MA, Riegman PH, Suurmeijer AP, Vissers CJ, Sainio M, transformation by CPI-17 through inhibition of a Merlin phospha- Carpen O et al. (2000). Evidence for a cytoskeleton attachment tase. Nature 442: 576–579. domain at the N-terminusof the NF2 protein. J Neurosci Res 62: Jin J, AriasEE, Chen J, Harper JW, Walter JC. (2006b). A family of 764–771. diverse Cul4-Ddb1-interacting proteins includes Cdt2, which is Gonzalez-Agosti C, Wiederhold T, Herndon ME, Gusella J, required for destruction of the replication factor Cdt1. Mol Ramesh V. (1999). Interdomain interaction of Merlin isoforms Cell 23: 709–721. and itsinfluence on intermolecular binding to NHE-RF. J Biol Kissil JL, Wilker EW, Johnson KC, Eckman MS, Yaffe MB, Jacks T. Chem 274: 34438–34442. (2003). Merlin, the product of the Nf2 tumor suppressor gene, is an Gutmann DH. (1997). Molecular insights into neurofibromatosis 2. inhibitor of the p21-activated kinase, Pak1. Mol Cell 12: 841–849. Neurobiol Dis 3: 247–261. Le Rouzic E, Belaidouni N, Estrabaud E, Morel M, Rain JC, Transy C He YJ, McCall CM, Hu J, Zeng Y, Xiong Y. (2006). DDB1 functions et al. (2007). HIV1 Vpr arrests the cell cycle by recruiting DCAF1/ asa linker to recruit WD40 proteinsto CUL4-ROC1 VprBP, a receptor of the Cul4-DDB1 ubiquitin ligase. Cell Cycle 6: ubiquitin ligases. Dev 20: 2949–2954. 182–188.

Oncogene VprBP targets Merlin for degradation J Huang and J Chen 4064 Lee J, Zhou P. (2007). DCAFs, the missing link of the CUL4-DDB1 Shaw RJ, McClatchey AI, JacksT. (1998). Regulation of the ubiquitin ligase. Mol Cell 26: 775–780. neurofibromatosis type 2 tumor suppressor protein, Merlin, by Lee JY, Moon HJ, Lee WK, Chun HJ, Han CW, Jeon YW et al. adhesion and growth arrest stimuli. J Biol Chem 273: 7757–7764. (2006). Merlin facilitatesubiquitination and degradation of Shaw RJ, Paez JG, Curto M, Yaktine A, Pruitt WM, Saotome I et al. transactivation-responsive RNA-binding protein. Oncogene 25: (2001). The Nf2 tumor suppressor, Merlin, functions in Rac- 1143–1152. dependent signaling. Dev Cell 1: 63–72. McClatchey AI, Saotome I, Mercer K, Crowley D, Gusella JF, Tan L, Ehrlich E, Yu XF. (2007). DDB1 and Cul4A are required for Bronson RT et al. (1998). Mice heterozygousfor a mutation at the human immunodeficiency virustype 1 Vpr-induced G 2 arrest. Nf2 tumor suppressor locus develop a range of highly metastatic J Virol 81: 10822–10830. tumors. Genes Dev 12: 1121–1133. Trofatter JA, MacCollin MM, Rutter JL, Murrell JR, Duyao MP, Morrison H, Sperka T, Manent J, Giovannini M, Ponta H, Herrlich P. Parry DM et al. (1993). A novel moesin-, ezrin-, radixin-like gene is a (2007). Merlin/neurofibromatosis type 2 suppresses growth by candidate for the neurofibromatosis 2 tumor suppressor. Cell 75: 826. inhibiting the activation of Rasand Rac. Cancer Res 67: Wen X, DuusKM, Friedrich TD, De Noronha CM. (2007). The HIV1 520–527. protein Vpr actsto promote G 2 cell cycle arrest by engaging a DDB1 Okada T, Lopez-Lago M, Giancotti FG. (2005). Merlin/NF-2 and Cullin4A-containing ubiquitin ligase complex using VprBP/ mediates contact inhibition of growth by suppressing recruitment DCAF1 asan adaptor. J Biol Chem 282: 27046–27057. of Rac to the plasma membrane. J Cell Biol 171: 361–371. Xu HM, Gutmann DH. (1998). Merlin differentially associates with Rouleau GA, Merel P, Lutchman M, Sanson M, Zucman J, Marineau the microtubule and actin cytoskeleton. J Neurosci Res 51: 403–415. C et al. (1993). Alteration in a new gene encoding a putative Zhang S, Feng Y, Narayan O, Zhao LJ. (2001). Cytoplasmic retention membrane-organizing protein causes neuro-fibromatosis type 2. of HIV-1 regulatory protein Vpr by protein–protein interaction with Nature 363: 515–521. a novel human cytoplasmic protein VprBP. Gene 263: 131–140.

Oncogene