RASSF6 Is a Novel Member of the RASSF Family of Tumor Suppressors

Oncogene (2007) 26, 6203–6211 & 2007 Nature Publishing Group All rights reserved 0950-9232/07 $30.00 www.nature.com/onc ORIGINAL ARTICLE RASSF6 is a novel member of the RASSF family of tumor suppressors

NPC Allen1, H Donninger2, MD Vos1, K Eckfeld1, L Hesson3, L Gordon2, MJ Birrer1, F Latif3 and GJ Clark1,2

1Department of Cell and Cancer Biology, National Cancer Institute, Rockville, MD, USA; 2Molecular Targets Group, Department of Medicine, JG Brown Cancer Center, University of Louisville, Louisville, KY, USA and 3Section of Medical and Molecular Genetics, University of Birmingham, Birmingham, UK

RASSF family proteins are tumor suppressors that are activated forms of Ras can also exhibit growth frequently downregulated during the development of antagonistic properties. These include the induction of human cancer. The best-characterizedmember of the senescence, cell cycle arrest and apoptosis (Mayo et al., family is RASSF1A, which is downregulated by promoter 1997; Serrano et al., 1997; Nicke et al., 2005). We can methylation in 40–90% of primary human tumors. We reconcile these apparently contradictory properties if we now identify and characterize a novel member of the consider the fact that Ras proteins can interact with a RASSF family, RASSF6. Like the other family members, wide variety of downstream effector proteins allowing a RASSF6 possesses a Ras Association domain and binds broad range of effector outputs (Malumbres and activatedRas. Exogenous expression of RASSF6 pro- Pellicer, 1998). Many Ras effector proteins contain motedapoptosis, synergizedwith activatedK-Ras to conserved structural regions that are responsible for induce cell death and inhibited the survival of specific mediating the interaction with Ras. These regions have tumor cell lines. Suppression of RASSF6 enhancedthe been designated the Ras Association (RA) domain tumorigenic phenotype of a human lung tumor cell line. (Ponting and Benjamin, 1996). Using a bioinformatics- Furthermore, RASSF6 is often downregulated in primary based approach to screen for RA domain containing human tumors. RASSF6 shares some similar overall proteins, we have previously identified several members properties as other RASSF proteins. However, there are of the RASSF family as potential Ras effectors that significant differences in biological activity between mediate growth inhibitory effects. RASSF6 and other family members including a discrete Four RASSF proteins have been characterized biolo- tissue expression profile, cell killing specificity andimpact gically thus far: RASSF1, RASSF2, RASSF4 and Nore1 on signaling pathways. Moreover, RASSF6 may play a (RASSF5) (Dammann et al., 2000; Vos et al., 2000, role in dictating the degree of inflammatory response to 2003a, b; Burbee et al., 2001; Khokhlatchev et al., 2002; the respiratory syncytial virus. Thus, RASSF6 is a novel Eckfeld et al., 2004; Agathanggelou et al., 2005). All RASSF family member that demonstrates the properties demonstrate biological properties compatible with a of a Ras effector andtumor suppressor but exhibits tumor suppressor function. These include inhibiting biological properties that are unique anddistinctfrom growth, promoting cell cycle arrest and apoptosis. They those of other family members. are all frequently downregulated during tumorigenesis by Oncogene (2007) 26, 6203–6211; doi:10.1038/sj.onc.1210440; promoter methylation (Agathanggelou et al., 2005; Akino published online 2 April 2007 et al., 2005). Moreover, RASSF1A knockout mice demonstrate enhanced tumor susceptibility (Tommasi et al., Keywords: Ras; RASSF6; tumor suppressor 2005) and loss of Nore1 (RASSF5) function is linked to a familial form of kidney cancer (Chen et al., 2003). Here, we identify and characterize a novel member of this family, RASSF6. RASSF6 interacts directly with K-Ras in a guanosine triphosphate (GTP)-dependent Introduction manner via its effector domain with an affinity compar- able to that of other known Ras effectors. RASSF6 Mutant Ras proteins play a critical role in the induces apoptosis and the ability of RASSF6 to kill development of over 30% of human cancers (Shields cells is enhanced by activated Ras. Overexpression et al., 2000; Malumbres and Barbacid, 2003). However, of RASSF6 inhibits the survival of specific tumor cell although promoting many aspects of transformation, lines and knockdown of RASSF6 by siRNA enhances the ability of tumor cell lines to grow in soft agar. In matched pair primary tumor samples, the levels Correspondence: Dr GJ Clark, JG Brown Cancer Center, Department of RASSF6 mRNA are often downregulated in the Medicine, Molecular Targets Group, University of Louisville, 580 S primary tumors. However, although RASSF6 has some Preston Street, Louisville, KY 40202, USA. E-mail: [email protected] similar overall properties as RASSF1A, the proteins Received 4 January 2007; revised 22 February 2007; accepted 22 demonstrate discrete expression profiles, and cell killing February 2007; published online 2 April 2007 specificity. RASSF6 is a potential Ras effector/tumor suppressor NPC Allen et al 6204 Recently, the RASSF6 locus has been implicated in RASSF6 and K-Ras, we generated a recombinant determining susceptibility to infection by the respiratory glutathione S-transferase (GST)-fusion of the RA domain syncytial virus (RSV) (Hull et al., 2004). RSV infection of RASSF6, designated GST-F6RA. GST-F6RA was activates the eukaryotic nuclear factor kB (NFkB) used as an affinity reagent in quantitative binding assays pathway and this activation may play a vital role in with purified, farnesylated K-Ras (Figure 3). Ras was the inflammatory response to infection. We have found loaded with GTP and incubated with 100 ng of GST or that RASSF6 expression inhibits the basal levels of GST-F6RA at decreasing concentrations for 4 h at 41Cin NFkB activity in a lung epithelial cell line, suggesting phosphate-buffered saline, 0.025% Tween-20 and 1 mM that defects in RASSF6 may facilitate viral NFkB MgCl2. GST-F6 beads were then washed and subjected to activation. Thus, we identify a further member of the western analysis using an anti-K-Ras antibody (F234 RASSF family as a novel potential Ras effector/tumor Santa Cruz Biotechnology, CA, USA). The apparent Kd suppressor with distinct biological characteristics. was taken as half the maximal binding. Binding was saturated by 300 nM,givinganapparentKd of less than or equal to 150 nM. Results

Family alignment and tissue distribution of RASSF6 RASSF6 and RASSF1A exhibit different cell-specific TblastN searches of the EST database using the RA growth inhibitory properties domain of RASSF1A led to the identification of Tumor cell lines were transfected with pBabe expression RASSF6. During the preparation of this manuscript, constructs and selected in puromycin. Colony formation the gene was described in: http://www.genecards.org/ was scored after 2 weeks (Figure 4). Whereas neither cgi-bin/carddisp.pl?gene ¼ RASSF6. Sequences were RASSF1A nor RASSF6 inhibited the growth of the aligned using ClustalW (Figure 1a). RASSF6 shows H1299 human lung tumor cell line, both inhibited the most identity to RASSF2 (57%) and RASSF4 (55%). It growth of MCF-7 human breast tumor cells. However, lacks the N-terminal extension of Nore1 (RASSF5) and RASSF6 was significantly less effective at inhibiting the the microtubule association domain of RASSF1A (Vos growth and survival of A549 human lung tumor cells. et al., 2004; Agathanggelou et al., 2005) but shares the SARAH motif of other family members. Northern analysis of RASSF6 expression (Figure 1b) showed a RASSF6 and K-Ras synergize to induce cell death distinct pattern of expression that differs from other To determine the effects of activated Ras on the family members (Vos et al., 2000, 2003a, b). biological activity of RASSF6, we performed cotrans- fections in 293-T cells. After 48 h, the cells were stained RASSF6 interacts with Ras GTP in cells via its with trypan blue to detect cell death. A dramatic effector domain synergistic activation of cell death was observed when 293-T cells were cotransfected with FLAG-tagged the cells were transfected with activated K-Ras and RASSF6 and HA-tagged activated K-Ras (G12V) or RASSF6 together (Figure 5aand b). wild-type K-Ras. The cells were lysed after 48 h and immunoprecipitated with anti-Ras antibody for western RASSF6 induces apoptotic cell death analysis using anti-FLAG antibodies. RASSF6 pre- To determine if RASSF6 was proapoptotic, we used the ferentially associated with the active form of K-Ras as pCaspase3-Sensor system (Clontech, Mountain View, no association was seen with the inactive (wild-type) CA, USA) and fluorescent microscopy to measure protein (Figure 2a). caspase-3 activation in individual cells as described Ras associates with its effectors via its effector previously (Vos et al., 2003b). Quantification showed domain (Marshall 1993) therefore, effector mutants that addition of RASSF6 induced a fivefold increase of (White et al., 1995; Vos et al., 2003a) of activated active caspases over empty vector (Figure 6). K-Ras (G12V/35S) and (G12V/40C) were subjected to similar coprecipitation experiments with RASSF6. Figure 2b shows that mutations in the effector region of RASSF6 binds the proapoptotic protein MOAP-1 K-Ras strongly impaired its ability to interact with One of the proapoptotic pathways that has been RASSF6. identified for RASSF1A involves the direct interaction Farnesylation is important for Ras to bind to its with the protein modulator of apoptosis 1 (MOAP-1) effectors with full affinity (Williams et al., 2000). By which then binds and activates Bax (Vos et al., 2006; using a point mutant of activated K-Ras that lacks a Baksh et al., 2005). To determine if RASSF6 might also functional CAAX motif (Solski et al., 1995) and cannot use this pathway to mediate apoptosis, we performed be farnesylated, we determined that the farnesylation of coimmunoprecipitation assays with differentially tagged Ras is very important for binding to RASSF6 expression constructs. RASSF6 could coimmunopreci- (Figure 2c). pitate with MOAP-1 (Figure 6b). Moreover, the degree of binding of RASSF6 and MOAP-1 appeared to be RASSF6 binds K-Ras directly enhanced by the presence of activated K-Ras, suggesting To examine the physiological relevance of the interac- a potential mechanism by which Ras may activate the tion and to confirm a direct interaction between proapoptotic effects of RASSF6.

Oncogene RASSF6 is a potential Ras effector/tumor suppressor NPC Allen et al 6205 a mRassF6 MTAMDHQFP-SWIVVNESTSISREQLNYLLETYNVFYENQKNLHILYGQTEDGQLIVEGM hRassF6 MTMMAHQYP-SWIFINEKTFITREQLNSLLKTYNIFYENQKNLHILYGETEDGKLIVEGM hRassF2 ---MDYSHQTSLVPCGQDKYISKNELLLHLKTYNLYYEGQ-NLQLRHRE-EEDEFIVEGL hRassF4 --MKEDCLPSSHVPISDSKSIQKSELLGLLKTYNCYHEGK-SFQLRHRE-EEGTLIIEGL * : .:.. * :.:* *:*** ::*.: .::: : : *:. :*:**:

mRassF6 LDIFWGVKRPIQLKIQDEKQISSFDLLNTPETFSSKGRMTRWGEFDD----LYRISELDR hRassF6 LDIFWGVKRPIQLKIQDEKPFSSFTSMKSSDVFSSKG-MTRWGEFDD----LYRISELDR hRassF2 LNISWGLRRPIRLQMQDDN-----ERIRPPPSSSS------WHSGCNLGAQGTTLKPLTV hRassF4 LNIAWGLRRPIRLQMQDDR-----EQVHLP--STS------WMP------RRPSCPLKE *:* **::***:*::**:. :. . :* * *

mRassF6 THVLASEARHSPEDYLS---YHS-TLTPYADEEPESPLLYRTMSEAESPLLYRTMSEAAL hRassF6 TQIPMSEKRNSQEDYLS---YHSNTLKPHAKDEPDS------PVLYRTMSEAAL hRassF2 PKVQISEVDAPPEGDQ----MPSSTDSRGLKPLQED------TPQLMRTRSDVGV hRassF4 PSPQNGNITAQGPSIQPVHKAESSTDSSGPLEEAEE------APQLMRTKSDASC . .: . * * . :. * * ** *.:..

mRassF6 V---RKRMRAPEMYRKDRMGVLSNHRASINGHVYDHETSIFTPTFGSETKVRANSIMRTE hRassF6 V---RKRMKPLMMDRKERQ----KNRASINGHFYNHETSIFIPAFESETKVRVNSNMRTE hRassF2 R--RRGNVRTPSDQRRIRR-----HRFSINGHFYNHKTSVFTPAYGSVTNVRINSTMTTP hRassF4 MSQRRPKCRAPGEAQRIRR-----HRFSINGHFYNHKTSVFTPAYGSVTNVRVNSTMTTL * ..:. :: * :* *****.*:*:**:* *:: * *:** ** * *

mRassF6 EVIKQLLQKFKIENSPRDFALYIIFGTGEQRKLKKTDVPLLQRLLQGPSKSNARIFLMDK hRassF6 EVIKQLLQKFKIENSPQDFALHIIFATGEQRRLKKTDIPLLQRLLQGPSEKNARIFLMDK hRassF2 QVLKLLLNKFKIENSAEEFALYVVHTSGEKQKLKATDYPLIARILQGPCEQISKVFLMEK hRassF4 QVLTLLLNKFRVEDGPSEFALYIVHESGERTKLKDCEYPLISRILHGPCEKIARIFLMEA :*:. **:**::*:.. :***:::. :**: :** : **: *:*:**.:. :::***:

mRassF6 D-AEEISRDVAPYINFHFSFLESILQRLDEEEKMEIERIMAKFNTERAFILKCLQSKQAA hRassF6 D-AEEISSDVAQYINFHFSLLESILQRLNEEEKREIQRIVTKFNKEKAIILKCLQNKLVI hRassF2 DQVEEVTYDVAQYIKFEMPVLKSFIQKLQEEEDREVKKLMRKYTVLRLMIRQRLE--EIA hRassF4 DLGVEVPHEVAQYIKFEMPVLDSFVEKLKEEEEREIIKLTMKFQALRLTMLQRLE--QLV * *:. :** **:*.:..*.*::::*.***. *: :: *: : : : *:

mRassF6 KTETTV hRassF6 KTETTV hRassF2 ETPATI hRassF4 EAK--- ::

b brain heart skeletal muscle colon thymus spleen kidney liver small intestine placenta lung per. blood leukocytes

hRassF6

β-actin

RASSF1 ++ ++ ++ - - + ++ ++ - ++ ++ ++ RASSF5 + ++ ++ - +++ +++ + ++ + + +++ +++ RASSF2 +++ - - - + + - + + ++ ++ nd

Figure 1 (a) Alignment of mRASSF6 with hRASSF6, RASSF2 and RASSF4 using the ClustalW program. Shaded bars, identical residues; ‘:’, conserved substitutions; ‘.’, semi-conserved substitutions. The RA domains have been shaded in gray. mRASSF6 shares 82% with hRASSF6 and RASSF6 shows greatest homology with RASSF2 (57%) followed by RASSF4 (55%). (b) Northern analysis of RASSF6 mRNA expression in normal tissue. Actin serves as the loading control. The table below the panel summarizes the relative expression of other family members in the same blot. þ , low expression; þþ, moderate expression; þþþ, high expression; À,no expression; nd, not done.

Oncogene RASSF6 is a potential Ras effector/tumor suppressor NPC Allen et al 6206 a IP: Ras beads IB: Flag GST-RASSF6 RA GST RASSF6

RASSF6 ++ + - - 150 nM 300 nM 600 nM 150 nM 300 nM 600 nM KRasWT - + - - + Kras(G12V) + --+ - IB: Flag RASSF6 Figure 3 RASSF6 interacts directly with K-Ras through its RA domain. The RA domain of RASSF6 was cloned into a GST vector (GST-F6RA) and used as an afﬁnity reagent to precipitate puriﬁed, IB:HA

Lysate: farnesylated K-Ras. GST served as a negative control. A total of KRas 100 ng of GST-fusion protein was incubated with decreasing concentrations of K-Ras. K-Ras was detected using a K-Ras 1 2 3 4 5 monoclonal antibody (Santa Cruz). The apparent Kd was taken as half the maximal binding concentration of 300 nM. b IP: Ras beads IB: Flag RASSF6

RASSF6 ++ ++ pBabeRASSF1A RASSF6 KRasWT + --- KRas(G12V) - + - - KRas(G12V/T35S) - - + - A549 K-Ras(G12V/Y40C) --- + IB: Flag Lysate: RASSF6

IB:HA KRas H1299

1 2 3 4

c IP: Anti-HA IB: Flag RASSF6

RASSF6 + + + MCF7 KRas(G12V) --+

KRas(G12V/C186S) --+ IB: Flag Figure 4 RASSF6 differentially inhibits growth in human tumor RASSF6 cell lines. H1299 and A549 human lung tumor cells along with Lysate IB: HA MCF-7 breast tumor cells were transfected with RASSF6 in the selectable vector pBabe. Cells were selected for 2 weeks in KRas puromycin and surviving colonies stained with crystal violet. 1 2 3

Figure 2 RASSF6 binds K-Ras in a GTP-dependant manner via its effector domain. HA-tagged K-Ras proteins (WT, activated and effector mutants) and FLAG-tagged RASSF6 were cotransfected ability to grow in soft agar compared to the control cells into 293-T cells. After 48 h, the cells were lysed using RIPA buffer transfected with a scrambled siRNA. and immunoprecipitated (IP) with sc-259 anti-Ras beads (Santa Cruz Biotechnology, Santa Cruz, CA, USA) before immunoblot- ting (IB) with FLAG antibodies. (a) RASSF6 preferentially interacts with the GTP-bound (activated) form of K-Ras compared RASSF6 expression is frequently lost in primary tumors to the wild type (compare lanes 1 and 2). The bottom two panels Using a cancer proﬁling array containing paired cDNAs show the relative amount of each protein in the lysate as examined generated from matched pair normal (N) and tumor (T)- with anti-FLAG and anti-HA antibodies. (b) In similar experi- derived tissues we show that the expression of RASSF6 ments, mutations in the effector domain of K-Ras disrupt the binding to RASSF6. (c) Lack of farnesylation (K-Ras12v/186S) is downregulated in 30–60% of tumor derived tissues of also impairs the ability of Ras and RASSF6 to interact with each the breast, colon, kidney, liver, rectum, pancreas, other. stomach and the thyroid gland compared to normal tissue (Figure 8). Intriguingly, several tumor samples showed elevated levels of RASSF6 expression. The RASSF6 siRNA enhances tumorigenicity differences seen were not owing to differential loading, H1792 cells transfected with RASSF6 siRNA were as the control incubation with the ubiquitin cDNA plated in soft agar (Figure 7a and b). The cells trans- probe yielded similar levels of expression in N vs T (data fected with RASSF6 siRNA demonstrated enhanced not shown). The reasons for the elevated levels of

Oncogene RASSF6 is a potential Ras effector/tumor suppressor NPC Allen et al 6207 a vector RASSF6 KRasWT KRas37G

KRasG12V RASSF6+KRasWT RASSF6+KRas37G RASSF6+KRasG12V

b vector RASSF6 KRasWT

DNA KRas37G KRas12V RASSF6 + KRasWT RASSF6 + KRas37G RASSF6 + KRas12V

0 50 100 150 Mean # dead cells

Figure 5 K-Ras and RASSF6 synergize to promote cell death. 293-T cells were transfected with K-Ras G12Vand RASSF6 constructs alone and in combination. (a) After 48 h, cells were stained with Trypan blue to identify dead cells. A typical experiment is illustrated and quantification of three independent experiments is shown in Figure 5b. expression in some tumors are not known, but could growth, loss of contact inhibition, reduced requirement reflect the presence of amutantRASSF6. for growth factors, enhanced motility and invasion and resistance to apoptosis (Malumbres and Barbacid, 2003). However, like other powerful oncoproteins, RASSF6 suppresses the NFkB pathway activated Ras can also induce various aspects of growth RSV induces severe bronchiolitis in asubpopulationof inhibition and death (Mayo et al., 1997; Serrano et al., infected individuals. Recent studies have genetically 1997; Hueber and Evan, 1998; Frame and Balmain, linked the RASSF6 locus to susceptibility for the RSV- 2000; Cox and Der, 2003; Diep et al., 2004; Nicke et al., induced bronchiolitis (Smyth and Openshaw, 2006). One 2005). These observations suggest that Ras proteins may of the major effects of RSV infection is to stimulate the control a balance between life and death in normal cells NFkB pathway (Hull et al., 2004). Activation of the (Feig and Buchsbaum, 2002). NFkB pathway appears to play roles in both modulat- RASSF family proteins have now been identified as ing the degree of inflammation induced by RSV and in supporting viral replication, perhaps by suppressing potential mediators of some of the growth inhibitory effects of Ras (Vos et al., 2000, 2003a, b; Khokhlatchev apoptosis (Bitko et al., 2004). As RASSF6 can induce et al., 2002; Eckfeld et al., 2004). Intriguingly, RASSF apoptosis by an unknown mechanism, we examined the family proteins are often downregulated during tumori- possibility that it might function, at least in part, via genesis (Pfeifer et al., 2002; Vos et al., 2003a, b; Eckfeld NFkB. Dual Luciferase assays were performed in A549 et al., 2004; Agathanggelou et al., 2005). This suggests lung tumor cells using an NFkB luciferase reporter and that loss of function of RASSF family proteins may be aCMV-Renillainternalcontrol. RASSF6 suppressed an important component of the development of Ras- the serum-induced basal levels of NFkB reporter dependant tumors. approximately fivefold (Figure 9). A much weaker effect We report here the first characterization of the last was observed with the most closely related family member RASSF2. member of the family, RASSF6. RASSF6 demonstrates approximately 60% identity to RASSF2 and RASSF4, with approximately 30% identity to RASSF1A. RASSF6 associates preferentially with activated Ras Discussion and this interaction is dependant on the effector domain of Ras. The interaction is also dependant on the correct Inappropriate activation of Ras proteins promotes a posttranslational modification of Ras by farnesyl. variety of protumorigenic phenotypes including enhanced Posttranslational modification of Ras has previously

Oncogene RASSF6 is a potential Ras effector/tumor suppressor NPC Allen et al 6208 a 80 a Control siRNA RASSF6 siRNA

Number of cells with active caspase-3 0 pCaspase-3 sensor pCaspase-3 sensor + vector + RASSF6

b Myc-MOAP-1 + + + 1.5 i) HA-K-Ras + - + b FL-RassF6 -++

IP MYC RassF6 IB:FL 1

IB:FL RassF6 0.5 Figure 6 RASSF6 promotes apoptosis in MCF-7 cells and binds MOAP-1. (a) The pCaspase3-Sensor plasmid7RFP-RASSF6 was transfected into MCF-7 cells and localization examined after 48 h. Fold change in RASSF6 levels 0 (a) Quantification of cells exhibiting caspase activation in RASSF6 expressing cells. Data presented are derived from three separate Control RASSF6 experiments, showing standard error. (b) A total of 2 mg FLAG- siRNA siRNA tagged RASSF1A was cotransfected in 293-T cells with 2 mg MYC- tagged MOAP-1 in the presence or absence of activated K-Ras. Figure 7 Knockdown of RASSF6 enhances the tumorigenic Upper panel shows the immunoprecipitate probed with FLAG, the phenotype. (a) H1792 human lung tumor cells were transfected lower panel shows the input lysate probed with FLAG. with a RASSF6 siRNA. The cells were plated in soft agar and scored for growth after 10 days. Two representative experiments are shown. (b) qRT–PCR analysis of RASSF6 expression in the A549 cells. The siRNA RASSF6 construct reduced the expression of RASSF in the cell line tested. Data represent two separate been shown to play an important role in the binding to experiments in duplicate. Error bars are standard errors. effectors. In the case of RASSF6, this seems to be particularly important. Thus, studies using unprocessed Ras to examine the binding of Ras to RASSF proteins RASSF6. This implies that RASSF1A and RASSF6 do may underestimate the binding affinity. The interaction not mediate identical functions. between farnesylated Ras and RASSF6 is likely to be To date, all RASSF family members with the physiological as the apparent Kd of the interaction was exception of RASSF3 (Tommasi et al., 2002) have determined to be less than or equal to 150 nM compared shown frequent loss of expression during tumor deve- to 50 nM previously determined for Ras and its lopment primarily due to promoter methylation physiological effector Raf-1 (Okada et al., 1996). (Agathanggelou et al., 2005). Here, we show that RASSF family proteins have been shown to inhibit RASSF6 shows reduced levels of RASSF6 mRNA in growth and participate in proapoptotic programs between 30 and 60% of primary tumor tissues of the (Agathanggelou et al., 2005). RASSF6 also promotes breast, colon, kidney, liver, pancreas, stomach and cell death in a Ras-dependant manner and can induce thyroid gland. Curiously, some tumor samples appeared apoptotic cell death in human tumor cell lines. to exhibit elevated levels of RASSF6 compared to the RASSF1A mediates some of its apoptotic functions by normal tissue. The reason for this is unknown but might directly binding the proapoptotic Bax activator MOAP- represent a mutant form of RASSF6 that has lost tumor 1 (Baksh et al., 2005; Vos et al., 2006). RASSF6 can also suppressor function. So far, we have not identified any interact with MOAP-1 and so shares this pathway with tumor-associated point mutants of RASSF6. RASSF1A. However, not all tumor lines are sensitive to To determine amechanism for the reduced expression killing by RASSF family overexpression. Neither of RASSF6 in tumors, we examined the promoter of RASSF1A nor RASSF6 was very effective at killing RASSF6 and found that approximately only 1/7 of the the lung tumor cell line H1299. The killing also shows tumor cell lines examined demonstrated partial promo- some specificity between family members as RASSF1A ter methylation (data not shown). This level was lower was much more effective at killing A549 cells than than expected, based on the results in the primary tumor

Oncogene RASSF6 is a potential Ras effector/tumor suppressor NPC Allen et al 6209 1 2 3 4 5 6 7 8 9 10 11 NT NT NT NT NT NT NT NT NT NT NT

TUMOR DOWNREGULATED UPREGULATED 1. Breast 3/10 0/10 2. Ovary 0/10 2/10 3. Colon 6/10 1/10 4. Stomach 5/10 1/10 5. Kidney 3/10 0/10 6. Cervix 0/10 2/10 7. Rectum 3/10 2/10 8. Thyroid 5/10 0/10 9. S. Intestine 1/10 1/10 10. Pancreas 4/10 1/10 11. Liver 3/10 0/10

Figure 8 Expression of RASSF6 is reduced in primary tumor cells. A cancer proﬁling array (Clontech) containing cDNAs generated from total RNA of tumor samples and corresponding normal tissue was subjected to northern analysis using the human RASSF6 as a cDNA probe. An ubiquitin cDNA probe was used and conﬁrmed approximately equal loading in all samples (data not shown). A summary of the frequency of alterations of expression is shown below.

150 infection is a leading cause of infant hospitalization. Certain individuals appear to be particularly susceptible to RSV-induced bronchiolitis, which is characterized by 100 severe inﬂammation (Hull et al., 2004) and can lead to serious complications or death. The reason why only some individuals develop bronchiolitis after infection is 50 not known. However, agenetic locus for the susceptibility to RSV-induced bronchiolitis has now been

Relative luciferase units mapped to RASSF6 (Hull et al., 2004). This implies 0 that defects in RASSF6 might alter the cellular response Vector RASSF6 RASSF2 to RSV infection. One of the major effects of infection Transfected DNA by RSV is activation of the NFkB pathway and this Figure 9 RASSF6 suppresses the activation of NFkB in A549 appears to play a critical role in both promoting human lung tumor cells. A549 cells were transfected with RASSF6, inflammation and supporting viral replication by anNFkB-luciferase reporter construct and a Renilla luciferase internal control. Results are normalized to the Renilla internal suppressing apoptosis (Bitko et al., 2004). As RASSF6 control. Data presented represent three experiments performed in is proapoptotic, we wondered if RASSF6 might be able duplicate. Error bars are standard errors. to modulate the NFkB pathway. Indeed, we found that in a lung tumor cell line, RASSF6 was highly effective at suppressing the NFkB pathway (Figure 9). This appears samples. This suggests that either our examination of to be a specific effect as the closest family member, the promoter did not extend sufficiently upstream to RASSF2, exhibited only amodest effect on NF kB detect important sites of methylation, or that non- activation (Figure 9). This is the first example of a methylation-based mechanisms are involved in the RASSF family member being linked to NFkB function. downregulation of RASSF6. RASSF6 is located at Thus, impaired function of RASSF6 may facilitate the 4q21.21 in the genome. This region has been reported to activation of NFkB by RSV, enhancing the infection. It suffer deletions during tumor development (Diep et al., may be interesting to test the activity of RSV in cells 2004). Therefore, the loss of expression in the primary specifically knocked down for RASSF6. It may also be tumors could also reflect gene deletions as well as interesting to examine the role of RASSF6 in inflamma- epigenetic mechanisms of silencing. tion in cancer. Recently, genetic mapping studies have implicated Thus, although RASSF6 demonstrates some simila- RASSF6 as playing an unexpected role in the response rities to other previously characterized RASSF proteins, to infection by RSV (Smyth and Openshaw, 2006). RSV it also demonstrates some unique properties. These

Oncogene RASSF6 is a potential Ras effector/tumor suppressor NPC Allen et al 6210 include adiscrete tissue expression profile, subcellular 100 nM of RASSF6 SiRNA (ID # 130201 Ambion Inc., localization, cell killing specificity and impact on Austin, TX, USA) using oligofectamine (Invitrogen). signaling pathways. Moreover, we may have identified Cells were plated in soft agar after 24 h as described a mechanism behind which RASSF6 may play a role in previously (Vos et al., 2003b). Dual luciferase assays dictating the degree of inflammatory response of the were performed as described in Ellis et al. (2002). RSV. Consequently, we propose that RASSF6 is a Ras effector/tumor suppressor of the RASSF family but qRT–PCR studies. quantitative real-time polymerase with distinct characteristics and function compared to chain reaction (qRT–PCR) was performed on an iCycler other family members. Real-Time Detection System (Bio-Rad Laboratories Inc., Hercules, CA, USA) using the Quantitect SYBR Experimental procedures Green RT–PCR Kit (Qiagen Inc., Valencia, CA, USA). Cloning and vectors. Murine RASSF6 was isolated by The fold change for the RASSF6 gene was calculated ÀDDC polymerase chain reaction (PCR) from IMAGE clone using the 2T method and using b-actin as the ID: 3669766 using primers ggatccatgacagcaatggatcaccag reference gene. and gaattcctagactgtggtctccgttttagc and cloned in pCDNA.Flag and pBabe.HA. The RA domain of Cell death/apoptosis assay. MCF-7 cells were trans- RASSF6 (amino acid residues 184–273) was isolated fected with 1 mg each of pCaspase3-sensor reporter by PCR (oligomers gacggatccaagcctctgatgatggacaga and plasmid (15) and pHcRed-RASSF6. Location of the acacaattgctaaactgttgtctctgtttttattactagtttatt) and cloned enhanced green fluorescent protein-sensor reporter into pGEX-2T vector (Phamacia, Piscataway, NJ, protein in cotransfected cells was examined after 24– USA). 48 h using fluorescence microscopy. In situ Trypan blue assays were performed on 293-T K-Ras/RASSF6 binding assays. 293-T cells were trans- cells (ATCC, Manassas, VA, USA) as described fected with pCGN-K-ras (Fiordalisi et al., 2000) and previously (Vos et al., 2003b). pcDNAF-RASSF6. Cells were lysed in radioimmuno precipitation assay (RIPA) buffer (50 mM Tris, pH 7.5, Expression in normal tissue and cancer profiling array. 1% (v/v) IGEPAL, 150 mM NaCl) and immunoprecipi- Human RASSF6 was used as a probe for northern tated with anti-Ras antibodies (sc-259, Santa Cruz analysis of a Normal Tissue Blot (Clontech, Paolo Alto, Biotechnology) followed by western analysis. CA, USA). To examine the expression of RASSF6 in Direct binding assays were carried out using a GST- primary tumors, southern analysis of the Clontech fusion of the RASSF6 RA domain and purified, matched pair tissue blot was performed, using hRASSF6 farnesylated K-Ras (a kind gift from D Stokoe, and an ubiquitin cDNA probe as a control. University of California, San Francisco). K-Ras was loaded as described previously (Clark et al., 1996) and added to 100 ng of the RA fusion protein immobilized to Abbreviations glutathione-coated beads at a range of Ras concentrations from 600 to 150 nM. Ras binding was determined GST, glutathione S transferase; RA, Ras association; RSV, as described previously (Vos et al., 2003b). respiratory syncytial virus.

Cell cultures and transfections. Human tumor cell lines Acknowledgements were grown in Dulbecco’s modiﬁed Eagle’s medium/ 10% fetal bovine serum and transfected using lipofecta- This work was supported in part by Intramural funds of mine-2000 (Invitrogen, Carlsbad, CA, USA). SiRNA the National Cancer Institute and RR018733 (GJC), FL is studies were performed by transfecting H1792 cells with supported in part by Cancer Research UK.

References

Agathanggelou A, Cooper WN, Latif F. (2005). Role of the virus: potential relevance in ocular inﬂammation and Ras-association domain family 1 tumor suppressor gene in respiratory infection. BMC Microbiol 4: 28. human cancers. Cancer Res 65: 3497–3508. Burbee DG, Forgacs E, Zochbauer-Muller S, Shivakumar L, Akino K, Toyota M, Suzuki H, Mita H, Sasaki Y, Ohe- Fong K, Gao B et al. (2001). Epigenetic inactivation of ToyotaM et al. (2005). The Ras effector RASSF2 is a novel RASSF1A in lung and breast cancers and malignant tumor-suppressor gene in human colorectal cancer. Gastro- phenotype suppression. J Natl Cancer Inst 93: 691–699. enterology 129: 156–169. Chen J, Lui WO, Vos MD, Clark GJ, Takahashi M, Baksh S, Tommasi S, Fenton S, Yu VC, Martins LM, Schoumans J et al. (2003). The t(1;3) breakpoint-spanning Pfeifer GP et al. (2005). The tumor suppressor RASSF1A genes LSAMP and NORE1 are involved in clear cell renal and MAP-1 link death receptor signaling to Bax cell carcinomas. Cancer Cell 4: 405–413. conformational change and cell death. Mol Cell 18: Clark GJ, Drugan JK, Terrell RS, Bradham C, Der CJ, 637–650. Bell RM et al. (1996). Peptides containing a consensus Ras Bitko V, Garmon NE, Cao T, Estrada B, Oakes JE, Lausch RN binding sequence from Raf-1 and theGTPase activating et al. (2004). Activation of cytokines and NF-kappa B in protein NF1 inhibit Ras function. Proc Natl Acad Sci USA corneal epithelial cells infected by respiratory syncytial 93: 1577–1581.

Oncogene RASSF6 is a potential Ras effector/tumor suppressor NPC Allen et al 6211 Cox AD, Der CJ. (2003). The dark side of Ras: regulation of for activation of, but not for association with, B-Raf. J Biol apoptosis. Oncogene 22: 8999–9006. Chem 271: 4671–4678. Dammann R, Li C, Yoon JH, Chin PL, Bates S, Pfeifer GP. Pfeifer GP, Yoon JH, Liu L, Tommasi S, Wilczynski SP, (2000). Epigenetic inactivation of a RAS association domain Dammann R. (2002). Methylation of the RASSF1A gene in family protein from the lung tumour suppressor locus human cancers. Biol Chem 383: 907–914. 3p21.3. Nat Genet 25: 315–319. Ponting DP, Benjamin DR. (1996). A novel family of Ras- Diep CB, TeixeiraMR, Thorstensen L, Wiig JN, EknaesM, binding domains. Trends Biochem Sci 21: 422–425. Nesland JM et al. (2004). Genome characteristics of primary Serrano M, Lin AW, McCurrach ME, Beach D, Lowe SW. carcinomas, local recurrences, carcinomatoses, and liver (1997). Oncogenic ras provokes premature cell senescence metastases from colorectal cancer patients. Mol Cancer 3:6. associated with accumulation of p53 and p16INK4a. Cell Eckfeld K, Hesson L, Vos MD, Bieche I, Latif F, Clark GJ. 88: 593–602. (2004). RASSF4/AD037 is a potential ras effector/tumor Shields JM, Pruitt K, McFall A, Shaub A, Der CJ. (2000). suppressor of the RASSF family. Cancer Res 64: 8688–8693. Understanding Ras: ‘it ain’t over ‘til it’s over’. Trends Cell EllisCA,VosMD,HowellH,VallecorsaT,FultsDW,ClarkGJ. Biol 10: 147–154. (2002). Rig is a novel Ras-related protein and potential Smyth RL, Openshaw PJ. (2006). Bronchiolitis. Lancet 368: neural tumor suppressor. Proc Natl Acad Sci USA 99: 312–322. 9876–9881. Solski PA, Quilliam LA, Coats SG, Der CJ, Buss JE. (1995). Feig LA, Buchsbaum RJ. (2002). Cell signaling: life or death Targeting proteins to membranes using signal sequences for decisions of ras proteins. Curr Biol 12: R259–R261. lipid modification. Methods Enzymol 250: 435–454. Fiordalisi JJ, Johnson RL 2nd, Ulku AS, Der CJ, Cox AD. Tommasi S, Dammann R, Jin SG, Zhang XX, Avruch J, (2001). Mammalian expression vectors for Ras family Pfeifer GP. (2002). RASSF3 and NORE1: identification and proteins: generation and use of expression constructs to cloning of two human homologues of the putative tumor analyze Ras family functions. Meth Enzymology 332: 3–36. suppressor gene RASSF1. Oncogene 21: 2713–2720. Frame S, Balmain A. (2000). Integration of positive and Tommasi S, Dammann R, Zhang Z, Wang Y, Liu L, Tsark negative growth signals during ras pathway activation WM et al. (2005). Tumor susceptibility of Rassf1a knockout in vivo. Curr Opin Genet Dev 10: 106–113. mice. Cancer Res 65: 92–98. Hueber AO, Evan GI. (1998). Traps to catch unwary Vos MD, Dallol A, Eckfeld K, Allen NP, Donninger H, oncogenes. Trends Genet 14: 364–367. Hesson L et al. (2006). The RASSF1A tumor suppressor Hull J, Rowlands K, Lockhart E, Sharland M, Moore C, activates Bax via MOAP-1. J Biol Chem 281: 4557–4563. Hanchard N et al. (2004). Haplotype mapping of the Vos MD, Ellis CA, Bell A, Birrer MJ, Clark GJ. (2000). Ras bronchiolitis susceptibility locus near IL8. Hum Genet 114: uses the novel tumor suppressor RASSF1 as an effector to 272–279. mediate apoptosis. J Biol Chem 275: 35669–35672. Khokhlatchev A, Rabizadeh S, Xavier R, Nedwidek M, Vos MD, Ellis CA, Elam C, Ulku AS, Taylor BJ, Clark GJ. Chen T, Zhang XF et al. (2002). Identification of a novel (2003a). RASSF2 is a novel K-Ras-specific effector and Ras-regulated proapoptotic pathway. Curr Biol 12: 253–265. potential tumor suppressor. J Biol Chem 278: 28045– Malumbres M, Barbacid M. (2003). RAS oncogenes: the first 28051. 30 years. Nat Rev Cancer 3: 459–465. Vos MD, Martinez A, Elam C, Dallol A, Taylor BJ, Latif F Malumbres M, Pellicer A. (1998). RAS pathways to cell cycle et al. (2004). A role for the RASSF1A tumor suppressor in control and cell transformation. Front Biosci 3: d887–d912. the regulation of tubulin polymerization and genomic Marshall MS. (1993). The effector interactions of p21ras. stability. Cancer Res 64: 4244–4250. Trends Biochem Sci 18: 250–254. Vos MD, Martinez A, Ellis CA, Vallecorsa T, Clark GJ. Mayo MW, Wang CY, Cogswell PC, Rogers-Graham KS, (2003b). The pro-apoptotic Ras effector Nore1 may serve as Lowe SW, Der CJ et al. (1997). Requirement of NF-kappaB a Ras-regulated tumor suppressor in the lung. J Biol Chem activation to suppress p53-independent apoptosis induced 278: 21938–21943. by oncogenic Ras. Science 278: 1812–1815. White MA, Nicolette C, Minden A, Polverino A, Van AL, Nicke B, Bastien J, Khanna SJ, Warne PH, Cowling V, Cook SJ Karin M et al. (1995). Multiple Ras functions can contribute et al. (2005). Involvement of MINK, a Ste20 family kinase, to mammalian cell transformation. Cell 80: 533–541. in Ras oncogene-induced growth arrest in human ovarian Williams JG, Drugan JK, Yi GS, Clark GJ, Der CJ, surface epithelial cells. Mol Cell 20: 673–685. Campbell SL. (2000). Elucidation of binding determinants Okada T, Masuda T, Shinkai M, Kariya K, Kataoka T. and functional consequences of Ras/Raf-cysteine-rich (1996). Post-translational modification of H-Ras is required domain interactions. J Biol Chem 275: 22172–22179.

Oncogene