Hedgehog Signaling Promotes the Degradation of Tumor Suppressor Sufu Through the Ubiquitin–Proteasome Pathway

Oncogene (2009) 28, 492–499 & 2009 Macmillan Publishers Limited All rights reserved 0950-9232/09 $32.00 www.nature.com/onc ORIGINAL ARTICLE Hedgehog signaling promotes the degradation of tumor suppressor Sufu through the ubiquitin–proteasome pathway

S Yue1,2, Y Chen1 and SY Cheng1,2

1Laboratory of Reproductive Medicine, Nanjing Medical University, Nanjing, Jiangsu, PR China and 2The Cancer Center, Nanjing Medical University, Nanjing, Jiangsu, PR China

Sustained Sonic hedgehog (Shh) pathway activity is that binding of the ligand to itsreceptor, Patched (Ptch), associated with tumorigenesis in a wide variety of tissues. alleviatesan inhibition from Ptch on a downstream Mutational inactivation of Shh receptor Patched (Ptch) membrane protein, Smoothened (Smo), which ulti- and a downstream gene Suppressor of fused (Sufu), both mately activatestarget genesthrough a family of three of which are negative regulators of the pathway, increases Gli transcription factors (Jacob and Lum, 2007; Ruiz i susceptibility to cerebellum cancer in humans and mice. Altaba et al., 2007; Wang et al., 2007). The Gli proteins Sufu is a binding partner of Shh pathway transcription recognize a common cis-element via a zinc-finger DNA factor Gli. Recent data indicate that inactivation of Sufu, binding domain, and can either activate or repress through either gene targeting in mice or RNAi-mediated transcription depending on whether Gli2 and Gli3 are silencing in cultured fibroblasts, is sufficient to turn on Shh processed by a proteolytic cleavage that converts them target gene expression. Here, we report that Sufu is into repressors. Gli1, a target of Shh signaling itself, degraded rapidly in certain cancer cells and we show that doesnot undergo thiscleavage event, and therefore Shh signaling promotes ubiquitination of Sufu, which always acts as an activator. Expression of Gli1 is leads to its destruction in the proteasomes. We identified elevated in a variety of tumorsasthe consequence of an ubiquitin attachment site on K257 of Sufu, and showed Shh pathway activation, and ectopic expression of Gli1 that Sufu-K257R mutant is more potent as a transcription and Gli2 results in skin cancers in mice (Ruiz i Altaba repressor and cell growth inhibitor because of increased et al., 2007). stability. These results indicate that Shh signaling Suppressor of Fused (Sufu), identified as a genetic regulates Sufu activity by inducing its turnover via the suppressor of fused mutantsin Drosophila (Preat et al., ubiquitin–proteasome system. 1993), isan important regulator of Gli. Genetic and Oncogene (2009) 28, 492–499; doi:10.1038/onc.2008.403; biochemical studies indicated that Drosophila Sufu published online 10 November 2008 binds, stabilizes and retains Ci (Cubitus interruptus) in the cytoplasm (Monnier et al., 1998; Ohlmeyer and Keywords: hedgehog; Sufu; ubiquitin; lung cancer Kalderon, 1998; Ingham and McMahon, 2001). These biochemical activitiesof Sufu are conservedin mammals (Kogerman et al., 1999; Jacob and Lum, 2007), and mammalian Sufu was also found to possess a nuclear activity in repressing Gli-mediated transcription via its Introduction interaction with SAP18 (Cheng and Bishop, 2002), a component of the Sin3-HDAC corepressor complex. Sonic hedgehog (Shh) isone of the primary morpho- Sufu mutations are present in 9% of medulloblastoma genic signals that specify patterns of cell growth and patients(Taylor et al., 2002), and mice that carry one differentiation during vertebrate development (Ingham target-inactivated Sufu allele exhibited heightened risk and McMahon, 2001). Mutationsin variouscompo- of developing medulloblastoma in p53 null background nentsof itssignalingpathway frequently occur in (Lee et al., 2007). Although Drosophila Sufu is not tumors originated from the skin, cerebellum and skeletal essential for embryonic development, mouse embryos muscle (Ruiz i Altaba et al., 2002), and abnormal devoid of Sufu die early because of multiple develop- pathway activity is associated with a subset of lung, mental anomaliesincluding failure to closethe neural digestive tract, pancreatic and prostate cancers (Beachy tube (Cooper et al., 2005; Svard et al., 2006), a et al., 2004). Shh signaling follows an unusual logic, in phenotype shared by inactivation of Ptch1. RNAi- mediated gene silencing experiment in NIH3T3 cells showed that attenuation of Sufu activity is sufficient to Correspondence: Dr SY Cheng, Department of Developmental activate Gli-mediated transcription (Varjosalo et al., Genetics, Nanjing Medical University, 140 Hanzhong Road, Xianzhi 2006). Lou 1908, Nanjing, Jiangsu, PR China. E-mail: [email protected] Given the important rolesof Sufu in regulating Gli Received 4 June 2008; revised 24 September 2008; accepted 1 October activities, we set out to investigate how Sufu itself is 2008; published online 10 November 2008 regulated in response to Shh signaling. Because Sufu is a Ubiquitination-mediated degradation of Sufu SYueet al 493 unique protein with few recognizable featuresexcept an To determine whether the low level of Sufu in NCI- N-terminal globular domain, we decided to examine H322M cellswasindeed becauseof reduced protein whether itsprotein turnover issubjectto control. Our stability, we treated the cells with MG132, a specific results show that Shh promotes Sufu degradation inhibitor of proteasomes, and observed accumulation through the ubiquitin–proteasome pathway, and this of Sufu starting at 6 h and continued up to 12 h Shh-induced degradation may account for reduced (Figure 1d). Blocking proteasomes also led to accumu- stability of Sufu in certain tumor cell lines. lation of Sufu in two prostate (PC3 and DU145) and a breast (MD-MBA-231) cancer cell lines (Figure 1d). Because the above cancer cell lines all have been reported to have an elevated Shh pathway (Beachy et al., Results 2004; Yuan et al., 2007), we decided to test whether Sufu turnover in NCI-H322M cells is subject to Shh signaling We began our analyses by assessing Sufu levels in a control. To that end, we measured the rate of Sufu decay panel of lung cancer cell lines, and found that the after blocking protein synthesis with cycloheximide. steady-state level of Sufu protein was drastically reduced Under thiscondition, the level of Sufu rapidly decreased in NCI-H322M cellswhereassignificant levelsof Sufu to below half of itsoriginal amount in 6 h, but failed to were expressed in the rest of lung cancer cell lines change significantly even after 24 h when the cells were examined and in human HEK293 and mouse NIH3T3 also treated with cyclopamine, a specific inhibitor of cells(Figure 1a). The reduction of Sufu in NCI-H322M Smo (Figure 2a). To determine whether Sufu turnover is cells is likely because of post-translational regulation normally controlled by Shh signaling, we repeated the because the level of Sufu transcript in these cells is comparable with that of HEK293 and NIH3T3 cellsas determined by standard and real-time reverse transcription PCR (RT–PCR) experiments(Figures1b and c).

Figure 2 Shh signaling promotes turnover of Sufu protein in vivo. (a) Western analyses of Sufu in NCI-H322M cells following cycloheximide (CHX) treatment in the absence or presence of 1 mM KAAD-cyclopamine (CPA). (b) Western analysis of Sufu turnover in primary mouse embryonic fibroblasts (MEFs) in the presence of various Shh agonists/antagonist and CHX. CHX, 20 mM; purmor- Figure 1 Expression of Sufu is regulated by the proteasome- phamine, 10 mM; KAAD-CPA, 1 mM; Shh-N-conditioned medium, mediated degradation in several cancer cell lines. (a) Western blot 50%. In each of the above conditions, MG132 (20 mM) wasalso analysis of Sufu protein level in total lysates of human lung cancer added to one cell sample for 6 h. (c) Time course of Sufu protein cellsA549, NCI-H522, NCI-H322M, HOP62, EKVX, HOP92, turnover. Line graphsrepresenttwo independent experimentsas NCI-H226, NCI-H460, mouse fibroblast NIH3T3 and human described in (b). The intensity of Sufu bands relative to that of embryonic kidney fibroblast HEK293 with antibodies against Sufu. GAPDH from each time point wasquantified by NIH ImageJ1.34s The level of GAPDH (glyceraldehyde 3-phosphate dehydrogenase) software package and plotted against the incubation time. Values wasusedasa loading control. ( b) Standard and (c) real-time RT– are the mean±s.d. (d) Western analysis of Sufu protein in torso PCR analyses of Sufu mRNA levels in NCI-H322M, A549 and tissues of e9.5 embryos showing reduced Sufu protein stability in HEK293 cellswith b-actin asthe loading control. ( d) Effect of Ptch1À/À homozygousembryoscompared with heterozygous proteasome inhibitor MG132 (20 mM, Calbiochem) on the level of littermate controls. Genotyping was carried out by PCR using Sufu protein in NCI-H322M, prostate cell lines PC3 and DU145, primers specific for neomycin resistance (knockout allele) and exon and breast cancer cell line MD-MBA-231. 2 (wild-type allele).

Oncogene Ubiquitination-mediated degradation of Sufu SYueet al 494 stability measurement in freshly isolated mouse em- examined whether activation of Shh signaling promotes bryonic fibroblasts (MEFs), which have low Shh path- ubiquitination of the endogenousSufu in primary way activity (Cooper et al., 2005; Svard et al., 2006). MEFs. In this experiment, we transfected MEFs with Our results indicated that Sufu was quite stable in the HA-tagged ubiquitin to enhance ubiquitin signal, MEFs, with an estimated half-life of 24 h (Figures 2b and treated the cellswith MG132 to let poly-ubiquiti- and c). Upon treating MEFswith Shh-conditioned nated protein accumulate. After immunoprecipitation medium or Smo agonist purmorphamine, the half-life of with anti-Sufu followed by western analysis with anti- Sufu decreased to around 4 h (Figures 2b and c). The HA antibodies, we indeed detected poly-ubiquitinated reduction induced by Shh-conditioned medium was endogenousSufu (Figure 3d). The accumulation of reversible with cyclopamine treatment, indicating it poly-ubiquitinated Sufu in the MEFsthat were treated waspathway specific.Again, the rapid turnover of Sufu with Shh-conditioned medium or purmorphamine was observed above was blocked by MG132 (Figure 2b), much higher than in those that were not treated with suggesting that activation of Shh signaling causes Sufu Shh agonists (Figure 3d). These data indicate that to be degraded in the proteasomes. To further investi- activation of Shh pathway promotesubiquitination of gate the physiological relevance of the observed regula- Sufu, which leadsto itsdegradation in the proteasome. tion of Sufu protein turnover, we isolated torsos of E9.5 Having established that Shh signaling controls Sufu embryoswith Ptch1 À/ þ or Ptch1À/À genotypes(Goodrich activity by the ubiquitin-proteasome degradation, we set et al., 1997), and quantified the levelsof Sufu by western out to identify the ubiquitin attachment site(s) on analysis. Because Ptch1 is a potent inhibitor of Smo, Sufu. We co-expressed the myc-tagged Sufu with Sufu protein isthusexpected to be unstable in the Ptch1 the HA-ubiquitin in HEK293 cells, and resolved the null embryosin which the Shh pathway isconstitutively immunopurified ubiquitin-Sufu conjugate by SDS– active (Goodrich et al., 1997). Indeed, our results PAGE (polyacrylamide gel electrophoresis). After stain- showed a dramatic reduction of Sufu protein level in ing with Coomassie blue, a band corresponding to the homozygousPtch1 À/À embryoscompared with that mono-ubiquitinated Sufu was excised and sent for liquid in the heterozygouslittermates(Figure 2d). Once again, chromatography tandem mass spectrometry analysis RT–PCR experiment showed no change in the levels of (Figure 4a). Ubiquitin isattached from itscarboxyl Sufu transcript between these two genotypes (data not terminus to substrate via an isopeptidyl bond, which is shown). Taken together, the above results indicate that resistant to trypsin cleavage. Thus, complete trypsin activation of Shh signaling either through pharmacolo- digestion would transfer the two terminal glycines from gical or genetic meansacceleratesSufu protein turnover. ubiquitin to substrate, adding 114 Da (molecular weight To determine whether Sufu is a substrate of the of the glycinyl-glycinyl dipeptide) to the mass of the ubiquitin–proteasome system, we co-expressed a myc- linked tryptic peptide. We identified a tryptic peptide tagged Sufu with a hemagglutinin (HA)-tagged ubiqui- covering amino acid sequence 255 to 272 of Sufu with tin in HEK293 cells. A protein band corresponding to the 114 Da shift after K257, indicating that K257 is the Sufu attached to a single ubiquitin moiety was clearly site of ubiquitin attachment (Figure 4b). detected after immunoprecipitation with anti-myc anti- To functionally evaluate the above mass spectrometry body (Figure 3a, Sufu-Ub), and poly-ubiquitinated result, we changed K257 to an arginine, and examined protein bandsalsoappeared when the transfectedcells the stability of Sufu-K257R mutant along with wild- were treated with MG132 (Figure 3a, Sufu-(Ub)n). To type control in NCI-H322M cells, in which Sufu verify whether the ubiquitinated proteinswere indeed degradesrapidly (Figure 1). Our resultsindicated that Sufu, we used the anti-Sufu antibody in the immuno- K257R mutant ismore stable than wild-type Sufu precipitation followed by western analysis with the same (Figure 5a). Nevertheless, K257R still underwent antibody. Once again, we were able to detect protein proteasomal degradation, suggesting that other lysine bandscorresponding to both mono- and poly-ubiquiti- residues can substitute K257 as ubiquitin attachment nated Sufu (Figure 3b). In poly-ubiquitinated proteins, sites to direct its degradation. We also assayed the the ubiquitin chainscan be linked via lysine at either ability of K257R to repress transcription in Sufu- position 48 or 63; the former being a specific signal that deficient MEFs, which have an active Shh pathway marksprotein for degradation in the proteasome, (Svard et al., 2006). When co-transfected with 8xGliBS- whereasthe latter often signalsprotein binding (Chen, luc, Sufu wascapable of attenuating the Gli-mediated 2005). We thusexamined the linkage in the poly- transcription, but K257R mutant was more potent than ubiquitinated Sufu by introducing in HEK293 cellsthe wild-type Sufu in doing so (Figure 5b), consistent with a myc-tagged Sufu along with a wild-type or mutant more stable protein. Finally, we introduced wild-type ubiquitin that lacks all seven lysine residues (KO), or has Sufu, Sufu-K257R, and the cloning vector back into only one lysine residue at position 48 (K48), or at 63 the lung cancer cell line, NCI-H322M, and examined the (K63). Our results showed that only the K48-ubiquitin effect of exogenousSufu on cell proliferation, which is supported the formation of poly-ubiquitin chains on sensitive to inhibition by cyclopamine (Yuan et al., Sufu (Figure 3c), indicating that it is a substrate of 2007). MTT assays indicated that the growth rate of ubiquitin and proteasome-mediated degradation. To NCI-H322M cellswasreduced to about 30 and 50% by determine whether ubiquitination isa physiological wild-type and K257R Sufu, respectively (Figures 5c and d), process that regulates Sufu activity or simply a again, consistent with K257R being a stronger Shh consequence of overexpression in transfected cells, we pathway repressor than its wild-type counterpart.

Oncogene Ubiquitination-mediated degradation of Sufu SYueet al 495

Figure 3 Shh signaling promotes ubiquitin modification of Sufu. (a) Sufu is a substrate of both mono-and poly-ubiquitination as determined in HEK293 cells that were transfected with pRK5-Sufu-Myc and HA-Ub. Transfected cell were treated with MG132 to enrich ubiquitinated proteins. Ubiquitinated and total Sufu were detected after IP with anti-Myc antibody on an anti-HA western blot (top panel) or anti-Sufu western blot (bottom panel), respectively. (b) Western analysis of ubiquitinated Sufu immunopurified with anti-Sufu antibody from HEK293 cellsasin ( a). Longer exposure of the film reveals an accumulation of poly-ubiquitinated Sufu (top panel). (c) Formation of poly-ubiquitin chain on Sufu is via the K48 but not K63 linkage. The ubiquitination assay of Sufu was carried out asin ( a) except for the mutant ubiquitin that has only one lysine at 48 (K48) or 63 (K63) position, or lacks all seven lysines (K0) in samples as indicated. MG132 treatment was for 12 h. (d) Activation of Shh signaling promotes ubiquitination of Sufu. Freshly isolated primary MEFswere transfectedwith pRK5-HA-Ub and treated with DMSO, purmorphamine or Shh-N-conditioned medium along with MG132 for 12 h. Ubiquitination of endogenousSufu wasdetected on an anti-HA westernblot after anti-Sufu immunoprecipitation. IgG heavy chain ismarked by *.

Discussion response to Shh signaling in freshly isolated mouse embryonic ﬁbroblasts and in embryonic tissues. The Shh pathway playsa crucial role in tumorigenesis Although we cannot rule out the existence of other by promoting survival and proliferation of pre-tumor mechanisms because Sufu is rapidly degraded only in a cellswith differentiation potential. Here, we have shown subset of cancer cells, our data show that the Shh that Sufu failsto accumulate in certain cancer control of Sufu stability is physiological, and suggest cells because of active Shh signaling, and also that that Shh may overcome the negative inhibition of Sufu Sufu undergoesubiquitin-proteasomal degradation in on Gli-mediated target gene expression at least in part

Oncogene Ubiquitination-mediated degradation of Sufu SYueet al 496

Figure 4 Identiﬁcation of ubiquitin attachment site on Sufu. (a) Preparation of ubiquitinated Sufu in transiently transfected HEK293 cells. The #2 protein band on the 7.5% SDS gel stained with Coomassie blue corresponds to mono-ubiquitinated Sufu and was excised for mass spectrometry. (b) liquid chromatography tandem mass spectrometry spectrum of the ubiquitinated Sufu peptide. The cutting positions of y and b ions were marked on both the peptide sequence and the spectrum. All marked b ion peaks, which represent peptide fragment produced from the amino terminus direction, contain the 114 Da mass shift. The two glycines from the ubiquitin attached to K257 were shown in brackets. y axis, relative intensity; x axis, mass to charge ratio.

Figure 5 Sufu-K257R mutant is a more potent pathway repressor than its wild-type counterpart. (a) Western analysis of the K257R mutant and wild-type Sufu in NCI-H322M cells. (b)8Â GliBS-Luc luciferase reporter assay showing the enhanced repressor activity of K257R. Open bar represents 0, shaded bar 500, and ﬁlled bar 1000 ng of pRK5-hSufu-myc or K257R plasmids in the assay. (c)MTT assays of NCI-H322M cells transfected with a cloning vector, wild-type, or K257R mutant Sufu. (d) Western analysis showing comparable expression of wild-type and K257R Sufu in the transfected cells as in (c).

by inducing itsdegradation. We further identiﬁed a mutant, conﬁrming the validity of K257 asa major lysine residue, K257, as an ubiquitin attachment site on ubiquitin attachment site. Sufu by liquid chromatography tandem mass spectro- Controlling Sufu activity by regulating itsprotein metry analysis. Like phosphorylation, ubiquitination stability is an attractive mechanism that connects the usually occurs on multiple lysine residues on target action of Smo to the downstream transcriptional events. proteins. Our data do not exclude the existence of other Proteasomal degradation could generate a gradual lysines that can also serve as the ubiquitin acceptors. inactivation of Sufu, an essential negative regulator, Moreover, ubiquitin ligases often select secondary sites which would establish a graded responsiveness of if the primary lysine residues were mutated (Trotman differentiating cellsto the Shh morphogenic gradient. et al., 2007). Nevertheless, we were able to show altered In tumors, loss of Sufu could render cells more sensitive stability and repressor activity of the Sufu-K257R to Shh pathway activation that promotescell survival.

Oncogene Ubiquitination-mediated degradation of Sufu SYueet al 497 Thus, elucidating the mechanism by which Shh signaling Materials and methods counteractsthe repressivefunction of Sufu haspro- found implication in both normal development and Plasmids and cell lines cancer. Human Sufu expression construct with a myc tag at the carboxyl Current understanding of Shh signaling mechanism terminusin pRK5 vector wasdescribedearlier (Cheng and Bishop, 2002). The Sufu-K257R mutant was created using the hasembraced the crucial role of primary cilium, a non- Quickchange site-directed mutagenesis kit (Stratagene, La Jolla, motile flagella-like protrusion present in almost all CA, USA). The HA-tagged ubiquitin, K48, K63 and KO interphase eukaryotic cells (Huangfu and Anderson, mutantswere giftsfrom Dr Ying E Zhang (Maryland, USA). 2006). Several key componentsof the Shh pathway such Human HEK293 and mouse NIH3T3 cells were purchased from asSmo, Gli2 and Gli3, and Sufu have all been found to ATCC and lung cancer cell linesA549, NCI-H522, NCI-H322M, be enriched in thisunique organelle, and recent data HOP62, EKVX, HOP92, NCI-H226, NCI-H460, prostate cancer indicated that localization of Smo in primary cilia is linesPC3 and DU145, and breastcancer line MD-MBA-231 dependent on itsformation of a protein complex with b- were obtained from the DTP/NCI and cultured in a humidified arrestin and Kif3A (Kovacs et al., 2008), a kinesin-like incubator with 5% CO2 in RPMI 1640 medium with 10% fetal molecule. Thisrequirement isreminiscentof the hedge- bovine serum, supplemented with penicillin and streptomycin. SufuÀ/À MEFswere giftsfrom Dr Rune Toftgard (Svard et al., hog signaling complex of Costal-2, Fused and Ci in 2006). Shh-conditioned medium wasproduced from HEK293 Drosophila (Stegman et al., 2000). AsDrosophila Sufu cells transiently transfected with pRK5-Shh-N (a gift from Dr is also a member of the hedgehog signaling complex, David J Robbins). and there isevidence indicating that it functionsto keep full-length Ci transcription activator in a stable but RNA isolation, RT and real-time PCR inactivated state, the mammalian Sufu may play an Total RNA was isolated from culture cells by using RNAiso analogousrole in primary cilia. However, in Drosophila reagent (TaKaRa, Madison, WI, USA), and reverse transcrip- salivary gland and in transfected mammalian cells, the tion wascarried out usingPrimeScript RT reagent Kit most readily assayable activity of Sufu is to restrict Ci/ (TaKaRa). Standard RT–PCR primersfor human Sufu Gli1 from entering the nucleus(Kogerman et al., 1999; (50-CCAATCAACCCTCAGCGGCAGAATG-30 and 50-GTA Sisson et al., 2006). It isnot clear whether thisactivity GGTGAGAAAGAGGGCTGTC-30) and a b-actin control (50-CTACAATGAGCTGCGTGTGGC-30 and 50-CAGGTC requiresthe functional integrity of primary cilia. In 0 addition, several studies also suggested a nuclear CAGACGCAGGATGGC-3 ) were synthesized by Invitrogen (Shanghai, China). Real-time PCR wascarried out in a function of Sufu (Cheng and Bishop, 2002; Paces-Fessy EvaGreen qPCR Master Mix (OPE Technology Development et al., 2004), in which Sufu may function asa component Co, Hongkong, China) on a 7500 Real-time PCR System of a transcription repressor complex. Regardless of the (Applied Biosystems, Foster City, CA, USA) with primer sets mode of action, attenuation of Sufu activity through for human Sufu (50-GTTGGAGGATTTAGAAGATTTG-30 degradation would be an effective way to alleviate the and 50-CGTCAGGCAGGATGGAGA-30) and b-actin control negative inhibition of Sufu on Shh signaling. (50-CTACAATGAGCTGCGTGTGGC-30 and 50-CAGGTC Prior to thisstudy,little isknown about how the CAGACGCAGGATGGC-30). Cycling conditionswere 50 1C activity of Sufu itself is regulated by Shh signaling, for 2 min, 95 1C for 10 min followed by 40 cyclesof despite recent reports showing phosphorylation of Sufu amplification at 95 1C for 30 s, 55 1C for 30 sand 72 1C for in Drosophila (Lum et al., 2003; Ho et al., 2005). Crystal 32 s. Experiments were repeated at least three times, and samples were analysed in triplicates. structure showed that both human and mouse Sufu contain four consensus protein kinase A (PKA) Isolation of MEFs and the preparation of protein extract from recognition sites that are likely present on the protein À surface (Merchant et al., 2004). When tested in vitro, Ptch1 embryos Primary mouse embryonic fibroblasts were isolated from E13.5 these sites can be phosphorylated by recombinant PKA, embryos. Briefly, mouse embryos were decapitated and and co-expressing Sufu with the PKA catalytic subunit degutted under phosphate-buffered saline. The remaining in HEK293 cellsled to stabilization of Sufu (our torsos were minced and digested in 0.25% trypsin-EDTA at unpublished data), suggesting that PKA phosphoryla- 37 1C for 30 min to disperse the cells, which were then plated tion may increase the stability of Sufu. Because PKA is a out in 10 cm plates. Ptch1 knockout mice were obtained from well-known negative regulator of Shh signaling, it is the Jackson Laboratory (Bar Harbor, ME, USA). The total tempting to speculate that Shh may induce Sufu protein extractsof Ptch1 À embryoswere prepared at E9.5 from degradation through Smo by counteracting the function torso tissues without limbs and tails in RIPA buffer (50 mM of PKA in a mechanism akin to the control of Tris-HCl, pH 7.5, 150 mM NaCl, 1% NP-40, 0.5% sodium proteolytic processing of Gli2 and Gli3. However, deoxycholate and 0.1% SDS). The protein concentration was determined by BCA (bicinchoninic acid) assay and equal mutant Sufu lacking all four PKA consensus sites still amount wasloaded in each lane. Genotyping of embryonic turned over at the same rate as their wild-type counter- tissue was done by PCR with primer sets given on the Jackson part in HEK 293 cells(our unpublished data). Thus,it is Lab website. possible that PKA may normally inactivate the E3 ubiquitin ligase specific for Sufu by phosphorylation, Measuring protein turnover and Shh signaling may reverse this inactivation through Approximately 0.4 Â 105 cellswere seededper well of a 12-well the action of Smo. Final resolution of this issue awaits plate 3–4 days(or until cellsbecame confluent) before identification of the E3 ligase responsible for catalysing cycloheximide treatment to block de novo protein synthesis. the ubiquitination of Sufu. Shh pathway agonists/antagonists treatment was started at the

Oncogene Ubiquitination-mediated degradation of Sufu SYueet al 498 same time as cycloheximide for the time indicated. At the end MTT assay of each time point, cells were collected for western analysis. Approximately 2 Â 105 H322M cellswere transfectedwith 1 mg The protein concentration in cell lysates was determined by pRK5-hSufu-Myc, K257R mutant or pRK5 vector control BCA (Bio-Rad, Hercules, CA, USA) assay, and equal amount plasmids in a 6-well plate. The next day, the cells were of total proteinswasloaded in each lane. Shh-conditioned reseeded in triplicates in a 96-well plate at 2 Â 104 per well for medium wasproduced from HEK293 cellstransientlytrans- MTT assay. One day after reseeding, the serum content fected with pRK5-Shh-N (a gift from Dr David J Robbins). in the culturing medium wasreduced to 0.5%, and the cell Cycloheximide (Sigma, St Louis, MO, USA), purmorphamine growth rate was assayed 48 h thereafter. MTT reaction (Calbiochem, San Diego, CA, USA), KAAD-cyclopamine wascarried out with 500 mg/ml 3-(4,5-dimethylthiazol-2-yl)- (Calbiochem) and proteasome inhibitor MG132 (Calbiochem) 2,5-diphenyltetrazolium bromide in the 96-well plate, were purchased from commercial vendors. and read by Microplate Reader (Bio-Rad, Model 680) at 570 nm. Ubiquitination assay To assay for ubiquitination of exogenous Sufu, HA-tagged Luciferase reporter assay ubiquitin and myc-tagged Sufu were co-transfected into The Gli-responsive 8 Â GliBS-Luc reporter wasdescribed HEK293 cells. After 24 h of transfection, the cells were treated earlier (Cheng and Bishop, 2002). In all, 0.4 Â 105 SufuÀ/À with MG132 for 16 h or asindicated before they were lysedin MEFs(giftsfrom Dr Rune Toftgard, Svard et al., 2006) were RIPA buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1% NP- transfected with 250 ng of 8 Â GliBS-Luc reporter, 25 ng 40, 0.5% sodium deoxycholate and 0.1% SDS). Immunopre- pRL-TK (Promega, Madison, WI, USA) and 250 ng pRK5- cipitation wascarried out with anti-myc or anti-Sufu (Santa hSufu-myc or mutant Sufu plasmids in each well of a 12-well Cruz Biotech, Santa Cruz, CA, USA) coupled to Protein G or plate. After 2 daysof transfection,the culture medium was Protein A agarose beads (Upstate, Lake Placid, NY, USA), replaced with a low-serum (0.5% fetal bovine serum) assay and immunopuriﬁed proteinswere subjectedto 8% SDS– medium together with 20 mM purmorphamine asindicated. The PAGE, followed by western blot analysis. luciferase activities were assayed after 24 h using the dual- reporter luciferase system on a GloMax-96 luminometer Mass spectrometry analysis of ubiquitin attachment sites (Promega). Ubiquitinated Sufu wasprepared by transfecting4 mg pRK5- hSufu-Flag and 4 mg pRK5-HA-Ub in 2 Â 106 HEK293 cells with Fugen HD (Roche, Mannheim, Germany). After 48 h, the Acknowledgements cells were lysed in RIPA buffer and the transfected Sufu was immunopuriﬁed from 2 mg cell lysates with anti-Flag M2 We thank DrsYE Zhang, A Ruiz i Altaba, R Toftgard, agarose beads (Sigma) before being resolved by 7.5% PAGE. DJ Robbinsand CM Fan for valuable reagentsand After Coomassie blue staining, the band corresponding to suggestions. This work is supported by a Chinese ubiquitinated Sufu wasexcised.The liquid chromatography National Science Foundation Grant (30771079 to SYC) tandem mass spectrometry analysis was carried out at the and a 973 State Key Research Project Grant (2009CB918403 Proteomics Center of Children’s Hospital Boston (MA, USA). to SYC).

References

Beachy PA, Karhadkar SS, Berman DM. (2004). Tissue repair and KovacsJJ, Whalen EJ, Liu R, Xiao K, Kim J, Chen M et al. (2008). stem cell renewal in carcinogenesis. Nature 432: 324–331. Beta-arrestin-mediated localization of smoothened to the primary Chen ZJ. (2005). Ubiquitin signalling in the NF-kappaB pathway. cilium. Science 320: 1777–1781. Nat Cell Biol 7: 758–765. Lee Y, Kawagoe R, Sasai K, Li Y, Russell HR, Curran T et al. (2007). Cheng SY, Bishop JM. (2002). Suppressor of Fused represses Loss of suppressor-of-fused function promotes tumorigenesis. Gli-mediated transcription by recruiting the SAP18-mSin3 core- Oncogene 26: 6442–6447. pressor complex. Proc Natl Acad Sci USA 99: 5442–5447. Lum L, Zhang C, Oh S, Mann RK, von Kessler DP, Taipale J et al. Cooper AF, Yu KP, Brueckner M, Brailey LL, Johnson L, McGrath (2003). Hedgehog signal transduction via Smoothened association JM et al. (2005). Cardiac and CNS defectsin a mousewith targeted with a cytoplasmic complex scaffolded by the atypical kinesin, disruption of suppressor of fused. Development 132: 4407–4417. Costal-2. Mol Cell 12: 1261–1274. Goodrich LV, Milenkovic L, HigginsKM, Scott MP. (1997). Altered Merchant M, VajdosFF, UltschM, Maun HR, Wendt U, neural cell fates and medulloblastoma in mouse patched mutants. Cannon J et al. (2004). Suppressor of fused regulates Gli activity Science 277: 1109–1113. through a dual binding mechanism. Mol Cell Biol 24: 8627–8641. Ho KS, Suyama K, Fish M, Scott MP. (2005). Differential regulation Monnier V, Dussillol F, Alves G, Lamour-Isnard C, Plessis A. (1998). of Hedgehog target gene transcription by Costal2 and Suppressor of Suppressor of fused links fused and Cubitus interruptus on the Fused. Development 132: 1401–1412. hedgehog signalling pathway. Curr Biol 8: 583–586. Huangfu D, Anderson KV. (2006). Signaling from Smo to Ci/Gli: Ohlmeyer JT, Kalderon D. (1998). Hedgehog stimulates maturation of conservation and divergence of Hedgehog pathways from Droso- Cubitusinterruptusinto a labile transcriptional activator. Nature phila to vertebrates. Development 133: 3–14. 396: 749–753. Ingham PW, McMahon AP. (2001). Hedgehog signaling in animal Paces-Fessy M, Boucher D, Petit E, Paute-Briand S, Blanchet- development: paradigmsand principles. Genes Dev 15: 3059–3087. Tournier MF. (2004). The negative regulator of Gli, Suppressor of Jacob L, Lum L. (2007). Hedgehog signaling pathway. Sci STKE 2007: fused (Sufu), interacts with SAP18, Galectin3 and other nuclear cm6. proteins. Biochem J 378: 353–362. Kogerman P, Grimm T, Kogerman L, Krause D, Unden AB, Preat T, Therond P, Limbourg-Bouchon B, Pham A, Tricoire H, Sandstedt B et al. (1999). Mammalian suppressor-of-fused Busson D et al. (1993). Segmental polarity in Drosophila melano- modulates nuclear-cytoplasmic shuttling of Gli-1. Nat Cell Biol 1: gaster: genetic dissection of fused in a Suppressor of fused background 312–319. revealsinteraction with costal-2. Genetics 135: 1047–1062.

Oncogene Ubiquitination-mediated degradation of Sufu SYueet al 499 Ruiz i Altaba A, MasC, Stecca B. (2007). The Gli code: an Taylor MD, Liu L, Raffel C, Hui CC, Mainprize TG, Zhang X et al. information nexus regulating cell fate, stemness and cancer. Trends (2002). Mutations in SUFU predispose to medulloblastoma. Nat Cell Biol 17: 438–447. Genet 31: 306–310. Ruiz i Altaba A, Sanchez P, Dahmane N. (2002). Gli and hedgehog Trotman LC, Wang X, Alimonti A, Chen Z, Teruya-Feldstein J, Yang in cancer: tumours, embryos and stem cells. Nat Rev Cancer 2: H et al. (2007). Ubiquitination regulatesPTEN nuclear import and 361–372. tumor suppression. Cell 128: 141–156. Sisson BE, Ziegenhorn SL, Holmgren RA. (2006). Regulation of Ci Varjosalo M, Li SP, Taipale J. (2006). Divergence of hedgehog signal and Su(fu) nuclear import in Drosophila. Dev Biol 294: 258–270. transduction mechanism between Drosophila and mammals. Dev Stegman MA, Vallance JE, Elangovan G, Sosinski J, Cheng Y, Cell 10: 177–186. RobbinsDJ. (2000). Identiﬁcation of a tetrameric hedgehog Wang Y, McMahon AP, Allen BL. (2007). Shifting para- signaling complex. J Biol Chem 275: 21809–21812. digmsin Hedgehog signaling. Curr Opin Cell Biol 19: Svard J, Heby-Henricson K, Persson-Lek M, Rozell B, Lauth M, 159–165. Bergstrom A et al. (2006). Genetic elimination of Suppressor of Yuan Z, Goetz JA, Singh S, Ogden SK, Petty WJ, Black CC et al. fused reveals an essential repressor function in the mammalian (2007). Frequent requirement of hedgehog signaling in non-small Hedgehog signaling pathway. Dev Cell 10: 187–197. cell lung carcinoma. Oncogene 26: 1046–1055.

Oncogene