Research ArticleCell Biology

Hippo Pathway–Dependent and –Independent Roles of RASSF6

See allHide authors and affiliations

Science Signaling  29 Sep 2009:
Vol. 2, Issue 90, pp. ra59
DOI: 10.1126/scisignal.2000300

Abstract

The Hippo pathway restricts cell growth and proliferation and promotes apoptosis to control organ size. The Drosophila melanogaster isoform of RASSF (Ras association domain family; dRASSF) antagonizes proapoptotic Hippo signaling by inhibiting the binding of the adaptor protein Salvador to the kinase Hippo. Paradoxically, however, dRASSF also functions as a tumor suppressor. In mammals, RASSF1A induces apoptosis by stimulating the mammalian Ste20–like kinases (MSTs) 1 and 2, which are Hippo homologs. Here, we characterize the interaction between MST2 and another mammalian RASSF isoform, RASSF6. When bound to MST2, RASSF6 inhibited MST2 activity to antagonize Hippo signaling. However, RASSF6 caused apoptosis when released from activated MST2 in a manner dependent on WW45, the mammalian Salvador homolog. Thus, RASSF6 antagonizes Hippo signaling and mediates apoptosis through a pathway that is parallel to the canonical Hippo pathway. Our findings suggest that activation of MST2 causes apoptosis through the Hippo pathway, as well as through a RASSF6-mediated pathway.

Introduction

Of the 10 mammalian RASSF (Ras association domain family) isoforms (1), RASSF1 to RASSF6 are considered to be tumor suppressors. Expression of RASSF1, RASSF2, RASSF4, RASSF5 (which is also known as Nore1), or RASSF6 causes apoptosis (17). Furthermore, RASSF1A knockout mice are susceptible to spontaneously forming and chemically induced tumors (8). Moreover, CpG islands are found in genes that encode RASSF1 to RASSF5, and it has been reported that RASSF1, RASSF2, RASSF4, and RASSF5 are silenced in cancer by promoter hypermethylation (1, 911). RASSF proteins contain Ras-association domains and C-terminal Salvador-RASSF-Hippo (SARAH) domains (27).

The Drosophila Hippo pathway regulates cell proliferation and organ size (1214). The pathway is composed of the FERM (4.1 protein-ezrin-radixin-moesin) domain proteins Expanded and Merlin, the serine-threonine kinases Warts and Hippo, the WW domain–containing adaptor protein Salvador, the Warts-activating protein Mats (Mob as tumor suppressor), and the transcriptional enhancer Yorkie. Hippo interacts with Salvador through its SARAH domain, and these proteins cooperate to activate Warts, which belongs to the family of nuclear Dbf2-related (NDR) kinases (1520). Warts then phosphorylates the transcriptional activator Yorkie and inhibits its activity (21). The Drosophila homolog of RASSF, dRASSF, competes with Salvador for binding to the SARAH domain of Hippo (22) and thus antagonizes Hippo signaling and plays an antiapoptotic role. However, dRASSF also antagonizes Ras1 signaling and exhibits tumor-suppressive function. Intriguingly, dRASSF loss of function augments cell growth in Drosophila expressing the Hippo mutant lacking the SARAH domain, but not in those expressing the kinase-negative Hippo mutant. These findings suggest that dRASSF suppresses tumorigenesis in a context-dependent manner.

Hippo pathway components are conserved in mammals. Loss of function of Merlin, the Salvador homolog WW45, or the Mats homolog MOB1 (Mps-one binder 1) or the gain of function of the Yorkie homolog YAP (Yes-associated protein) are implicated in human cancers (1214). RASSF1 and RASSF5 bind the Hippo homolog MST1 (mammalian Ste20–like kinase 1) and inhibit its autophosphorylation (2325). However, the binding of Ras to RASSF5 enhances MST1 activity and promotes apoptosis. RASSF1A also releases MST2 from the inhibitory effect of Raf-1 (26). Ultimately, RASSF1A and RASSF5 activate NDR1, NDR2, and LATS1 (large tumor suppressor, homolog 1) to induce apoptosis (2629). These findings suggest that RASSF1A and RASSF5 stimulate Hippo signaling and play a distinct role from dRASSF.

RASSF6 also induces apoptosis in various cell types (6, 7). We determined that RASSF6, like RASSF1A, binds MST2 and hypothesized that RASSF6 induces apoptosis through the activation of MST2. Unexpectedly, however, MST2 inhibited RASSF6-induced apoptosis. Conversely, RASSF6 inhibited MST2 and antagonized Hippo signaling in a manner similar to that of dRASSF. After dissociating from MST2, RASSF6 induced apoptosis independently of Hippo signaling. Thus, RASSF6 plays a dual role as an inhibitor of Hippo signaling and as a mediator of apoptosis; the latter role is triggered upon the activation of the Hippo pathway but is independent of canonical Hippo signaling.

Results

RASSF6 interacts directly with MST2 through the SARAH domain

To dissect the molecular mechanism underlying RASSF6-induced apoptosis, we identified RASSF6-interacting proteins by performing a yeast two-hybrid screen of human lung and kidney complementary DNA (cDNA) libraries. Nine independent clones encoding human MST2 were obtained from the screening of 1 × 106 clones. This interaction was confirmed by pull-down and coimmunoprecipitation experiments. Maltose binding protein (MBP)–tagged RASSF6 trapped Myc-tagged MST2 (fig. S1A), and green fluorescent protein (GFP)–tagged RASSF6 coimmunoprecipitated with Myc-MST2 (fig. S1B). MBP-RASSF6 bound to glutathione S-transferase (GST)–tagged MST2, as well as a kinase-negative MST2 mutant (GST-MST2-KN), indicating that RASSF6 directly binds to MST2 and that the kinase activity of MST2 is dispensable for this interaction (fig. S1C). Endogenous RASSF6 and MST2 coimmunoprecipitated from primary cultured rat hepatocytes (fig. S1D). Homo-oligomerization of MST2 was blocked by RASSF6, indicating that MST2 interacts with RASSF6 more strongly than it does with itself (fig. S1E). Next, we determined the sequences involved in the interaction using various MST2 and RASSF6 constructs (fig. S2A). MBP-RASSF6 only trapped fragments of MST2 that encompassed the SARAH domain, such as Myc-MST2(326–491) and FLAG-MST2(435–491) (fig. S2B). Of the three MBP-tagged RASSF6 fragments tested, only the fragment containing the SARAH domain [MBP-RASSF6(301–369)] captured GST-MST2 (fig. S2C). Thus, the SARAH domain of RASSF6, like that of the human Salvador homolog WW45, interacts with the SARAH domain of MST2 (30). On closer inspection, MBP-RASSF6 bound to fragments that contained the C-terminal two-thirds of the MST2 SARAH domain [GFP-MST2(435–491) and GFP-MST2(450–491)] (fig. S2B). In contrast, GST-WW45(327–383) interacted with fragments that encompassed the N-terminal third of the MST2 SARAH domain [GFP-MST2(435–491) and GFP-MST2(435–469), but not GFP-MST2(450–491)], indicating that WW45 and RASSF6 bind to distinct regions on MST2 (fig. S2D).

RASSF6, WW45, and MST2 form a tripartite complex in vitro and in vivo, and the complex is disrupted by the activation of MST2

In Drosophila, dRASSF and Salvador compete for binding to Hippo (22). However, because RASSF6 and WW45 bind to different sequences in the SARAH domain of MST2, we hypothesized that these proteins might form a tripartite complex. Indeed, MBP-RASSF6 bound to GST-WW45(327–383) only in the presence of GST-MST2 in vitro (Fig. 1A), and in human embryonic kidney (HEK) 293 cells, Myc-WW45 coimmunoprecipitated with FLAG-RASSF6 only when Myc-MST2 was also expressed (Fig. 1B). Thus, RASSF6, WW45, and MST2 form a complex, which contrasts with the situation in Drosophila. Because WW45 is extensively phosphorylated by MST2 (30), we hypothesized that the activation of MST2 might disrupt the complex. Indeed, activation of MST1 and MST2 with a phosphatase inhibitor, okadaic acid (OA), which enhances the autophosphorylation of MST1 and MST2, reduced the coimmunoprecipitation of RASSF6 but not that of WW45 (Fig. 1C) (25). Furthermore, Myc-RASSF6 coimmunoprecipitated with FLAG-WW45 by way of Myc-MST2 from untreated cells, but not from OA-treated cells (Fig. 1D). However, suppression of endogenous WW45 enabled Myc-RASSF6 to interact with FLAG-MST2 in OA-treated cells (Fig. 1E), suggesting that endogenous WW45 competes with Myc-RASSF6 for the binding to FLAG-MST2 in OA-treated cells. We also examined how the interaction of endogenous RASSF6 and MST2 changes during apoptosis. Rat hepatocytes undergo apoptosis when they are treated with OA (fig. S3), an event that was suppressed by knockdown of RASSF6, suggesting a role for RASSF6 in OA-induced apoptosis. We confirmed that OA treatment reduces the interaction of endogenous RASSF6 and MST2 in rat hepatocytes (Fig. 1F).

Fig. 1

Interactions among RASSF6, MST2, and WW45. (A) One hundred picomoles of GST-WW45(327–383) or GST-MST2, or both, were incubated with control MBP or MBP-RASSF6 immobilized on amylose resin beads. GST-WW45(327–383) was captured by MBP-RASSF6 in the presence of GST-MST2 (arrow). IB, immunoblotting. (B) FLAG-RASSF6, Myc-MST2, and Myc-WW45 were expressed in HEK293 cells. Myc-MST2 coimmunoprecipitated with FLAG-RASSF6 (white arrowhead). Myc-WW45 (black arrowhead) immunoprecipitated with FLAG-RASSF6 from cells coexpressing Myc-MST2. The asterisk indicates the immunoglobulin heavy chain. (C) Myc-RASSF6 did not coimmunoprecipitate with FLAG-MST2 from the OA-treated cells (white arrowhead). The coimmunoprecipitation of Myc-WW45 was not affected by OA (black arrowhead). (D) FLAG-WW45, Myc-MST2, and Myc-RASSF6 were coexpressed in HEK293 cells. Myc-RASSF6 was not detected in FLAG-WW45 immunoprecipitates from cells that were treated with OA. The asterisk indicates the immunoglobulin heavy chain. (E) When endogenous WW45 was silenced, Myc-RASSF6 coimmunoprecipitated with FLAG-MST2 even after OA treatment (arrow). The lower panel shows the suppression of WW45 at the protein level. (F) OA treatment (10 nM for 12 hours) reduced the coimmunoprecipitation of RASSF6 and MST2 from rat primary cultured hepatocytes. The asterisk indicates the immunoglobulin heavy chain. IP, immunoprecipitation.

MST2 blocks RASSF6-induced apoptosis

To examine how MST2 influences RASSF6-induced apoptosis, RASSF6 was expressed alone or with MST2 in HeLa cells. Thirty-six and 48 hours after transfection, HeLa cells expressing FLAG-RASSF6 were reduced in number and exhibited nuclear condensation (Fig. 2A, top panel). In contrast, expression of GFP-MST2 alone did not cause nuclear condensation (Fig. 2A, middle panel). Intriguingly, HeLa cells expressing both FLAG-RASSF6 and GFP-MST2 remained viable and did not show nuclear condensation (Fig. 2A, bottom panel). Coexpression of MST2 also inhibited RASSF6-mediated induction of several other hallmarks of apoptosis, including release of cytochrome c, Bax translocation, and release of apoptosis-inducing factor (Table 1). The kinase-negative Myc-MST2-KN mutant and a construct containing the MST2 SARAH domain [Myc-MST2(326–491)], which do not cause apoptosis, also blocked RASSF6-induced apoptosis (Fig. 2B, top and middle panels, and fig. S4). In contrast, Myc-MST2(2–327) (which does not contain the SARAH domain) was detected in the nucleus, caused apoptosis, and did not block RASSF6-induced apoptosis (Fig. 2B, bottom panel, and fig. S4). Quantification of cells exhibiting nuclear condensation is shown (Fig. 2C). MST1 also inhibited RASSF6-induced apoptosis in HeLa cells (fig. S5, A to C). Fluorescence-activated cell sorting (FACS) analysis by tetramethylrhodamine methyl ester (TMRM) staining corroborated that RASSF6 enhanced mitochondrial membrane permeability, an effect that was blocked by MST2 independently of its kinase activity (Fig. 2D). Conversely, suppression of MST1 and MST2 slightly enhanced RASSF6-induced apoptosis in HeLa cells and OA-induced apoptosis in rat hepatocytes (Fig. 2E and fig. S3).

Fig. 2

MST2 blocks RASSF6-induced apoptosis. (A) HeLa cells expressing GFP-MST2 and FLAG-RASSF6 were fixed and immunostained with antibody to FLAG at the indicated time periods after transfection. Nuclei were visualized with Hoechst 33342. Cells expressing FLAG-RASSF6 exhibited nuclear condensation (arrowheads), whereas the cells expressing both FLAG-RASSF6 and GFP-MST2 did not (arrows). Scale bar, 50 μm. (B) HeLa cells expressing GFP-RASSF6 and Myc-MST2(2–327) exhibited nuclear condensation (arrowheads), whereas cells expressing GFP-RASSF6 with Myc-MST2-KN or Myc-MST2(326–491) did not (arrows). Scale bar, 50 μm. (C) Quantification of cells with nuclear condensation. Fifty to 100 cells were analyzed for each treatment. Error bars indicate SD of three independent experiments. ***P < 0.001. n.s., not significant. (D) HeLa cells were transfected with either pClneoGFP or pLL3.7 EF-2 FH-RASSF6-ires-GFP to express GFP or RASSF6 together with GFP (RASSF6 + GFP), respectively. Myc-MST2 or Myc-MST2-KN was expressed as indicated. The cells were loaded with 200 nM TMRM. Mitochondrial membrane permeability was evaluated in GFP-positive cells by FACS. *P < 0.05. (E) HeLa cells were transfected with control (white bars) or MST1- and MST2-specific siRNA (black bars), then GFP or FLAG-RASSF6 together with GFP (RASSF6 + GFP). Nuclear condensation was evaluated in FLAG-RASSF6–positive cells (left). Mitochondrial membrane permeability was evaluated in GFP-positive cells (middle). The suppression of MST1 and MST2 was confirmed by immunoblotting (right). **P < 0.01; ***P < 0.001.

Table 1

Quantification of Figs. 2 and 3. One hundred HeLa cells expressing various proteins were observed at 24 and 36 hours after transfection for each experiment to evaluate the ratio of cells showing various changes characteristic of cell death. The data are shown as the means ± SD of three independent experiments. AIF, apoptosis-inducing factor; EndoG, endouclease G.

View this table:

WW45 enables RASSF6 to induce apoptosis in the presence of MST2

Because WW45 and MOB1 interact with MST2 (3032), we tested the effect of these proteins on the ability of MST2 to counter RASSF6-induced apoptosis. We transfected a dual promoter vector encoding GFP-MST2 and FLAG-RASSF6 alone or with Myc-WW45 or Myc-MOB1. Forty-eight hours after transfection, almost no HeLa cells expressing GFP-MST2, FLAG-RASSF6, and Myc-WW45 were viable, in contrast to cells expressing GFP-MST2, FLAG-RASSF6, and Myc-MOB1 (Fig. 3A). Among the deletion mutants tested, a construct that contained the MST2 binding site [Myc-WW45(327–383)], but not one lacking this region [Myc-WW45(2–334)], enabled RASSF6 to induce apoptosis (Fig. 3, B and C). Quantification of the number of cells exhibiting nuclear condensation indicated that WW45 enabled RASSF6 to induce apoptosis even in the presence of MST2 (Fig. 3C). We confirmed that none of the WW45 proteins or MOB1 caused apoptosis (fig. S4).

Fig. 3

WW45 enables RASSF6 to induce apoptosis in the presence of MST2. (A) HeLa cells expressing FLAG-RASSF6, GFP-MST2, and Myc-WW45 showed nuclear condensation (arrowheads), whereas cells expressing FLAG-RASSF6, GFP-MST2, and Myc-MOB1 did not (arrows). Scale bar, 50 μm. (B) HeLa cells expressing GFP-MST2 and FLAG-RASSF6 with Myc-WW45(2–334) or Myc-WW45(327–383). Cells expressing Myc-WW45(2–334) showed flat morphology without nuclear condensation (arrows) in contrast to those expressing Myc-WW45(327–383), which showed nuclear condensation (arrowheads). Scale bar, 50 μm. (C) Quantification of cells with nuclear condensation. Fifty to 100 cells were analyzed per treatment. Error bars indicate SD of three independent experiments. ***P < 0.001.

RASSF6 inhibits MST2 activity

dRASSF inhibits Hippo, whereas in mammals, RASSF1A activates the Hippo homolog MST2 (22, 26, 27, 29). These reports are apparently inconsistent. We therefore examined the effect of RASSF6 on the activity of MST2. FLAG and hexahistidine (FH)–tagged MST2 was immunoprecipitated from HEK293 cells and incubated with adenosine 5′-triphosphate (ATP) and MOB1 in vitro to measure the incorporation of 32P into MOB1. RASSF6 significantly inhibited the phosphorylation of MOB1, whereas WW45 increased it and abrogated the inhibitory effect of RASSF6 (Fig. 4A). Immunoblotting with antibody against phospho-PAK1 (Thr423), which cross-reacts with phospho-MST2 (Thr180), indicated that autophosphorylation of MST2, which is essential for activation, is reduced in cells coexpressing RASSF6 (Fig. 4B). MST2 activity was inhibited by a RASSF6 deletion mutant containing the SARAH domain (ΔN), but not one lacking the SARAH domain (ΔC) (Fig. 4C). We further confirmed that RASSF1A enhanced the phosphorylation of MOB1 by MST2 under the same conditions in which RASSF2, RASSF4, and RASSF6 were inhibitory (Fig. 4D). In addition, MST2 activity was inhibited by a chimera containing the N terminus of RASSF6 and the C terminus of RASSF1A (6/1A) but stimulated by a chimera containing the N terminus of RASSF1A and the C terminus of RASSF6 (1A/6), indicating that the N-terminal region of RASSF1A is required for the activation of MST2.

Fig. 4

RASSF6 inhibits the phosphorylation of MOB1 by MST2. (A) FH-MST2, Myc-RASSF6, and Myc-WW45 were expressed in HEK293 cells in various combinations. FH-MST2 was immunoprecipitated, eluted with FLAG peptide, quantified by Sypro Orange staining (left), and used in vitro kinase assays with GST-MOB1 as a substrate. On the right, the upper panel shows the autoradiography of GST-MOB1. The lower graph shows the amount of 32P incorporated into GST-MOB1. FH-MST2 alone (open circles); with Myc-RASSF6 (filled circles); with Myc-WW45 (open triangles); and with Myc-RASSF6 and Myc-WW45 (filled triangles). (B) Cell lysates were immunoblotted with an antibody to phospho-PAK1 that recognizes the autophosphorylation of MST2. RASSF6 reduced the signal (lane 3), whereas WW45 slightly increased it (lane 4) and negated the reduction caused by RASSF6 (lane 5). (C) The effect of various RASSF6 proteins on the activity of MST2. RASSF6 lacking the SARAH domain (ΔC) had no effect. (D) MST2 activity was inhibited by RASSF2, RASSF4, and RASSF6, but enhanced by RASSF1A and a chimera containing the N-terminal RASSF1A and the C-terminal RASSF6. *P < 0.05; **P < 0.01; ***P < 0.001. n.s., not significant.

RASSF6 inhibits NDR1 and LATS2 activity

Because MST2 and MOB1 cooperate to activate NDR kinases, we tested the effect of RASSF6 on these kinases (3133). In vitro assays showed that RASSF6 inhibits NDR1 and LATS2 activity (Fig. 5A). Thr444 of NDR1 and Thr1041 of LATS2 are phosphorylated by MST2, and antibodies that recognize the phosphorylated forms of these sites were prepared. Phosphorylation was detected when MOB1 and MST2 were coexpressed, but was reduced when RASSF6 was also transfected, supporting the idea that RASSF6 inhibits the phosphorylation of NDR1 and LATS2 by MST2 (Fig. 5B).

Fig. 5

RASSF6 inhibits NDR1 and LATS2. (A) The effect of RASSF6 on the in vitro kinase activities of NDR1 and LATS2. FLAG-NDR1 and LATS2-FLAG were immunoprecipitated from transfected HEK293 cells and used for in vitro assays. The activation of NDR1 and LATS2 by MST2 and MOB1 was abolished by RASSF6. NDR1 assays were performed with synthetic peptide as the substrate for 90 min at 30°C (left). LATS2 assays were performed with His-YAP1 as the substrate for 60 min at 30°C (right). LATS2-FLAG and GST-MOB1 (open triangle); LATS2-FLAG, GST-MOB1, and FH-MST2 (open circles); and LATS2-FLAG, GST-MOB1, FH-MST2, and Myc-RASSF6 (filled circles). **P < 0.01. (B) FLAG-NDR1 and LATS2-FLAG were expressed in HEK293 cells in various combinations with Myc-MOB1, Myc-MST2, and Myc-RASSF6. Cell lysates were immunoblotted with the indicated antibodies. Antibodies to pNDR1 and pLATS2 recognize the phosphorylation of NDR1 and LATS2 by MST2, respectively.

RASSF6-induced apoptosis is independent of MST2, NDR1 and NDR2, and LATS1 and LATS2

MST2 activates NDR and LATS kinases, thereby promoting apoptosis (2629, 33). However, because MST2 is inhibited by interaction with RASSF6, RASSF6-induced apoptosis is unlikely to be mediated by MST2. A RASSF6 mutant lacking the SARAH domain consistently induced apoptosis, whereas a mutant lacking the N-terminal region or the Ras-association domain did not (fig. S6A), thus indicating that MST2 is not involved in RASSF6-induced apoptosis. Knockdown of NDR or LATS kinases did not affect RASSF6-induced apoptosis (fig. S6B). We also expressed a YAP1 mutant, in which Ser127 and Tyr407 (corresponding to Tyr357 of YAP2) are mutated to alanine and phenylalanine (YAP1 SAYF), respectively. Phosphorylation of Ser127 by LATS1 or LATS2 causes YAP1 to translocate from the nucleus to the cytosol (34, 35), whereas phosphorylation of Tyr357 by c-Abl triggers selective activation of p73-mediated transcription of genes encoding proapoptotic factors (36); thus, the YAP1 SAYF mutant is localized in the nucleus and does not promote proapoptotic gene transcription. RASSF6 caused apoptosis in cells coexpressing this mutant (fig. S6C). Together, these findings indicate that RASSF6-induced apoptosis is independent of MST2, NDR kinases, and YAP1.

MOAP1 is implicated in RASSF6-induced apoptosis and MST2 blocks the interaction of RASSF6 and MOAP1

The interaction of RASSF1A with the Bax-activating protein MOAP1 (modulator of apoptosis 1) has been implicated in inducing apoptosis (37, 38). RASSF6 also interacts with MOAP1, although the molecular determinants of this interaction have not been elucidated (7). The acidic sequence Glu-Glu-Glu-Glu (EEEE) in the SARAH domain of RASSF1A is necessary for MOAP1 binding (37). Although the SARAH domain of RASSF6 contains a similar sequence [Glu-Glu-Glu-Lys (EEEK)], MOAP1 bound to RASSF6 lacking the N-terminal region, the Ras-association domain, or the SARAH domain, suggesting that MOAP1 binds to more than one site in RASSF6 (fig. S7A). Knockdown of MOAP1 partially suppressed RASSF6-induced apoptosis (Fig. 6A). GFP-MOAP1 coimmunoprecipitated with Myc-RASSF6 from HEK293 cells, but coexpression of FH-MST2 reduced the coimmunoprecipitation, the effect of which was abrogated by the additional expression of WW45 (Fig. 6B). Suppression of MOAP1 also attenuated OA-induced apoptosis in rat hepatocytes (fig. S3).

Fig. 6

MOAP1 is involved in RASSF6-induced apoptosis and MST2 inhibits the interaction between MOAP1 and RASSF6. (A) HeLa cells were transfected with either control or MOAP1-specific siRNA, then with FLAG-RASSF6. Cells expressing FLAG-RASSF6 and control siRNA showed nuclear condensation. Arrows and arrowheads indicate normal and condensed nuclei, respectively. Knockdown of MOAP1 decreased the number of cells exhibiting nuclear condensation. Knockdown was confirmed by quantitative RT-PCR. ***P < 0.001. (B) GFP-MOAP1, FLAG-RASSF6, Myc-MST2, and Myc-WW45 were expressed in HEK293 cells in various combinations. Myc-MST2 blocked the coimmunoprecipitation of GFP-MOAP1 with FLAG-RASSF6, but the further expression of Myc-WW45 recovered this interaction.

dRASSF inhibits MST2 activity and dRASSF-induced apoptosis is inhibited by the kinase-negative Hippo

dRASSF inhibits Hippo (22), so we tested whether dRASSF inhibits MST2. dRASSF interacted with MST2 and inhibited it (fig. S8, A and B). Immunoblotting with antibody to phospho-PAK1 showed that dRASSF attenuates the autophosphorylation of MST2 (fig. S8C). We next tested the effect of Hippo on dRASSF-induced apoptosis. dRASSF induced apoptosis in HeLa cells, but not S2 cells (figs. S8, D to F, and S9). dRASSF consistently bound to MOAP1, implying that MOAP1 is involved in dRASSF-induced apoptosis (fig. S7B). Unlike MST2, transfection of Hippo alone caused apoptosis in HeLa cells, but the kinase-negative Hippo mutant, which does not induce apoptosis, inhibited dRASSF-induced apoptosis and abrogated the interaction of dRASSF and MOAP1 (figs. S7B and S8, D and F).

Discussion

The Hippo pathway restricts cell growth and proliferation and promotes apoptosis to control organ size in Drosophila (1214). dRASSF binds the SARAH domain of Hippo and blocks its ability to interact with Salvador, thereby antagonizing Hippo signaling, and thus has been identified as an additional component of the Hippo pathway in Drosophila (22). However, suppression of dRASSF enhances cell growth in the Hippo mutant lacking the SARAH domain and the Ras1 mutant, suggesting a paradoxical function for dRASSF in tumor suppression. Another contradictory observation is that dRASSF loss of function does not affect cell growth in the kinase-negative Hippo mutant. These findings suggest that dRASSF plays a tumor-suppressive role in a context-dependent manner. The human genome contains six dRASSF-related genes that harbor C-terminal SARAH domains (1), two of which (RASSF1 and RASSF5) are well studied tumor suppressors (2). RASSF2, RASSF4, and RASSF6 have also been characterized as proapoptotic and tumor-suppressive proteins (47). Because the amino acid sequences of mammalian RASSFs are homologous to that of dRASSF, RASSF proteins are thought to play a role in Hippo signaling in mammalian cells. Accordingly, RASSF1A interacts with MST kinases and activates NDR1, NDR2, and LATS1 (2427, 29). In addition, RASSF1A releases MST2 from inhibition by Raf-1 and subsequently activates MST2 and LATS1 to cause apoptosis (26), thereby showing that RASSF1A stimulates Hippo signaling. It remains to be determined whether the mammalian RASSF proteins play a different role in mammalian Hippo signaling compared to that of dRASSF in the Drosophila pathway.

Interactions mediated by the SARAH domains

The SARAH domain sequences of Salvador and WW45 are 42% conserved, those of Hippo and MST2 are 56% conserved, and those of dRASSF and RASSF6 are 25% conserved. The SARAH domain of MST2 binds the SARAH domains of WW45 and RASSF6, whereas the SARAH domains of WW45 and RASSF6 do not interact. Our interaction studies reveal the possibility that the SARAH domain of MST2 has two distinct subdomains that individually mediate the interactions with WW45 and RASSF6. However, nuclear magnetic resonance experiments on the SARAH domains of MST1, RASSF5, and WW45 suggest that these proteins do not form a stable complex and thus do not support the existence of two subdomains (39). Because the sequences of the SARAH domains of RASSF5 and RASSF6 are only 23% conserved, it will be necessary to directly analyze the SARAH domain of RASSF6 to determine whether it has different properties.

Competition of WW45 and RASSF6 for binding to MST2

A previous study using Drosophila cells showed that the interactions of Hippo with Salvador and dRASSF are mutually exclusive (22). Our biochemical data indicate that WW45, MST2, and RASSF6 can form a tripartite complex in vitro and in cells. However, when MST2 is activated, RASSF6 is released from the complex in a manner dependent on WW45. Based on the extensive phosphorylation of WW45 by MST2, we speculate that phosphorylation may change the conformation or the surface charge of WW45 so as to block RASSF6 from binding to MST2.

Inhibition of RASSF6-induced apoptosis by MST2

In HeLa cells, RASSF6 expression leads to caspase activation and the release of proapoptotic factors from the mitochondria (6). MST kinases are involved in RASSF1A- and RASSF5-induced apoptosis (2429). We therefore speculated that MST2 might be involved in RASSF6-induced apoptosis, but unexpectedly found that MST2 suppresses RASSF6-induced apoptosis. The SARAH domain of MST2 is essential for this inhibition, although its kinase activity is dispensable. This observation is consistent with the fact that dRASSF functions as a tumor suppressor in Drosophila with a Hippo mutant lacking the SARAH domain but not with a kinase-negative Hippo mutant (22). Furthermore, we observed that WW45 recovers RASSF6-induced apoptosis. Because MST1 also blocks RASSF6-induced apoptosis, binding of the MST kinases to the SARAH domain of RASSF6 is likely important for the inhibition. Together with the in vitro interaction experiments, we speculate that RASSF6 is released from MST2 in a WW45-dependent manner and then induces apoptosis. Indeed, the interaction between endogenous RASSF6 and MST2 is abolished in rat hepatocytes treated with OA, which would be expected to activate MST2. Although MST2 is cleaved by caspase-3 (40), the amount of intact MST2 does not change in OA-treated rat hepatocytes, thus suggesting that the disruption of the MST2-RASSF6 interaction is not likely due to the cleavage of MST2. No antibody is available to detect endogenous rat Salvador, so it remains to be determined how rat Salvador is involved in the release of RASSF6 in rat hepatocytes. Nonetheless, it is clear that the interaction between RASSF6 and MST2 is lost during apoptosis. The finding that a RASSF6 mutant lacking the SARAH domain also induces apoptosis indicates that MST2 does not mediate RASSF6-induced apoptosis.

Inhibition of Hippo signaling by RASSF6

In vitro kinase assays and immunoblotting to detect the activation of MST2 in situ indicate that RASSF6 inhibits MST2. We speculate that RASSF6 blocks the homo-oligomerization of MST2 and the phosphorylation and subsequent activation. Thus, RASSF6, like dRASSF, antagonizes Hippo signaling. This finding stands in contrast to the reported functions of RASSF1A. Under the same conditions in in vitro assays, RASSF1A activates MST2, whereas RASSF6 inhibits it. RASSF2 and RASSF4 also inhibit MST2. Moreover, dRASSF can inhibit MST2. The C-terminal region of RASSF6 suppresses MST2 activation, but the addition of the N-terminal region of RASSF1A activates MST2. These findings imply that the N-terminal region of RASSF1A has stimulatory properties and that RASSF1A is different from other RASSF proteins. Another study has shown that WW45 is necessary for the activation of MST1 and LATS2 (41). Because MST1 is also inhibited by RASSF6, we speculate that WW45 is necessary to release MST1 from the inhibitory effects of RASSF proteins.

The molecular mechanisms underlying RASSF6-induced apoptosis

RASSF1A induces apoptosis through the activation of NDR1, NDR2, and LATS1. However, RASSF6-induced apoptosis is not mediated by NDR kinases. Because the binding of MST2 inhibits apoptosis, we hypothesize that MST2 prevents the interaction between RASSF6 and an apoptosis-exerting target molecule. One such putative target is MOAP1 (7). RASSF1A binds MOAP1, induces the interaction of MOAP1 and Bax, and results in the activation of Bax (37, 38). RASSF6 also interacts with MOAP1 (7). Although the acidic sequence in the SARAH domain is a putative binding site, the in vitro binding experiments suggest that MOAP1 binds to RASSF6 at more than one site. Together with the fact that MOAP1 binds to dRASSF, in which the acidic sequence is not conserved, it seems likely that the interaction between MOAP1 and RASSF does not solely depend on the acidic sequence in the SARAH domain. Indeed, this would be consistent with the observation that RASSF6 lacking the SARAH domain can still induce apoptosis.

MST2 attenuates the interaction between RASSF6 and MOAP1. The coexpression of WW45 enables the interaction of RASSF6 with MOAP1 in the presence of MST2. These observations suggest that MST2 prevents RASSF6 from binding to MOAP1 and that once RASSF6 is released from MST2, it binds MOAP1 and triggers apoptosis. However, the suppression of MOAP1 by RNA interference does not block RASSF6-induced apoptosis as much as the coexpression of MST2. Therefore, additional molecular mechanisms are likely to play roles in RASSF6-induced and MST2-suppressible apoptosis. Because RASSF6 does not induce apoptosis in S2 cells, the machinery underlying RASSF6-induced apoptosis is missing in these cells. It is necessary to test other Drosophila cells to determine whether such machinery is conserved in Drosophila.

Mutual inhibition of RASSF6 and MST2

Our findings illustrate the previously unknown aspects of the role of RASSF in the mammalian Hippo pathway (Fig. 7). RASSF6 and MST2 inhibit each other under static conditions. When MST2 is activated, RASSF6 is released from MST2 and induces apoptosis. Free of RASSF6, MST2 is further activated and, in a complex with WW45, activates NDR kinases. Therefore, a trigger that causes the initial activation of MST2 leads to apoptosis through two pathways, the canonical Hippo signaling and RASSF6-induced apoptosis.

Fig. 7

Hippo-dependent and -independent roles of RASSF6. RASSF6 and MST2 inhibit each other under static conditions. Because RASSF6 inhibits MST2, RASSF6 is a negative regulator of Hippo signaling. When a trigger activates MST2, RASSF6 is released from MST2. When free of RASSF6, MST2 is further activated, so that the complex of MST2 and WW45 activates NDR kinases, such as LATS2. However, RASSF6 also induces apoptosis independently of Hippo signaling. RASSF6 is therefore a downstream mediator regulated by MST2 and WW45.

Materials and Methods

Construction of expression vectors and recombinant proteins

cDNAs encoding human MST2 (BC010640), RASSF2 (BC057402), RASSF3 (BC055023), and RASSF4 (BC032593) were purchased from Open Biosystem. pcDNA human LATS2-FLAG and cDNA of human MST1 are gifts of T. Yamamoto (University of Tokyo) and Y. Gotoh (University of Tokyo), respectively. pCIneoMyc, pCIneoFLAG-His6 (FH), pCIneoFLAG-His6-FLAG (FLAG), pClneoHA, pClneoGFP, pGex4T-1 (GE Healthcare Bio-Sciences), and pMalC2 (New England Biolabs) vectors were used to generate expression constructs (6). The MOB1 and NDR1 constructs were described previously (32). MST2 constructs were based on AAH10640, RASSF6 constructs were based on NP_958834, and WW45 constructs were based on AAH20537. The kinase-negative mutant of MST2 (MST2-KN), in which Lys56 was replaced with Arg, was generated by polymerase chain reaction (PCR). pLL3.7 was modified as follows (42). First, pBudCE4.1 (Invitrogen) was digested with Nhe I and Not I to release the elongation factor promoter, which was then ligated into the Xba I and Not I sites of pLL3.7. Next, a linker (5′-ggcctctagaattcggatcccgggctagctcgaggccgcgtttaaac-3′ and 5′-aattgtttaaacgcggcctcgagctagcccgggatccgaattctaga-3′) was ligated into the Not I and Eco RI sites to generate pLL3.7 EF-2. RASSF6 expression vectors were generated using the pClneoFH vector and encode the following amino acid residues of human RASSF6b (NP_958834) and RASSF1A (NP_009113): RASSF6ΔN, 212 to 369; RASSF6ΔRA, 1 to 213 and 305 to 369; RASSF6ΔC, 1 to 308; RASSF1A/6, 1 to 191 of RASSF1A and 218 to 369 of RASSF6; and RASSF6/1A, 1 to 217 of RASSF6 and 194 to 340 of RASSF1A. pEGFPN3 (Clontech) was digested with Bam HI and Not I and the isolated GFP fragment was ligated into the Bam HI and Not I sites of pIRES (Clontech) to generate pIRES-GFP. The Nhe I and Sal I fragments from the pClneoFH expression vectors, which cover various FH-tagged RASSF-coding sequences, were ligated with the Xho I– and Not I–digested fragment from pIRES-GFP into the Nhe I and Not I sites of pLL3.7 EF-2 to express various RASSF proteins together with GFP. dRASSF (Fly base ID: FBgn0039055) and Hippo (Fly base ID: FBcl0136453) were obtained through DNAFORM (Yokohama, Japan). The kinase-negative mutant of Hippo (Hippo-KN), in which Lys71 was replaced with arginine, was generated by PCR. pAc5.1/V5-HisA (Invitrogen) was digested with Eco RI and Mlu I and a linker (5′-aattgctagcaagcttgaattcacgcgtctagactcgagcggccg-3′ and 5′-cgcgcggccgctcgagtctagacgcgtgaattcaagcttgctagc-3′) was ligated to generate pAc-2. pClneoFH-RASSF6, dRASSF, and Hippo were digested with Nhe I and Not I and ligated into the same sites of pAc-2. Rat MOAP1 was ligated into Eco RI and Sal I sites of pClneoFH, pClneoGFP, and pClneoMyc. Nhe I and Not I fragments from pClneoMyc and pClneoFH MOAP1 were ligated into the same sites of pAc-2.

Antibodies

Mouse monoclonal antibody 4C2 was raised against an MBP-rat RASSF6 fusion protein. Rabbit MST2, phosphospecific NDR1/2, and phosphospecific LATS1/2 antibodies were raised against GST-MST2-KN, CVFINY(phospho-)TYKRFE, and CAFYEF(phospho-)TFRRFF [single letters code amino acid residues; cysteine is added for the conjugation to Keyhole Limpet Hemocyanin (Thermo Fisher Scientific)] (43). The phosphospecific antibodies were affinity-purified on nonphosphorylated and phosphorylated peptide columns. Antibodies and reagents used in this study were obtained from the following sources: mouse monoclonal antibody to Myc 9E10 (American Type Culture Collection); mouse monoclonal and rabbit polyclonal antibodies to FLAG, mouse monoclonal antibodies to MBP, GST, HA, endonuclease G, and actin (Sigma-Aldrich); mouse antibodies to cytochrome C and MST1 (BD Pharmingen); mouse antibody to GFP and goat antibody to apoptosis-inducing factor (Santa Cruz); rabbit antibody to phospho-PAK1 (Cell Signaling); antibodies to NDR1 and WW45 (Abnova); and fluorescein isothiocyanate (FITC)-, rhodamine- and Cy5-conjugated secondary antibodies (Chemicon International). Antibody to Phospho-PAK1 (Thr423) cross-reacts with phospho-MST1 (Thr183) and phospho-MST2 (Thr180).

Cell culture and transfection

COS-7, HeLa, and HEK293FT (Invitrogen) cells were grown in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum (FBS), penicillin (100 U/ml), and streptomycin (100 μg/ml) under 5% CO2 at 37°C. COS-7 cells were transfected with the DEAE-dextran method. S2 cells were cultured in Schneider’s insect medium (Sigma-Aldrich) supplemented with 10% FBS at 23°C and transfected with Effectene (Qiagen). HeLa and HEK293FT cells were transfected with Lipofectamine 2000 reagent (Invitrogen). Primary culture of hepatocytes was performed as described (44). The animals and procedures used in this study were in accordance with the guidelines and approval of Tokyo Medical and Dental University Animal Care and Use Committee.

RNA interference

HeLa cells and rat hepatocytes were transfected with the following 21-nucleotide oligomers (Ambion) or Stealth small interfering RNA (siRNA; Invitrogen) with Lipofectamine 2000. Silencer negative control #1 siRNA (Ambion) or Stealth RNAi (RNA interference) negative control (Invitrogen) was used as a control. The validity of the knockdown was confirmed by quantitative reverse transcription (RT)-PCR or by immunoblotting. For quantitative RT-PCR, total RNAs were extracted with TRIzol Reagent (Invitrogen). First-strand cDNA was synthesized from 3 μg of total RNA in a 20-μl reaction containing 500 μM each dNTP (deoxynucleotide triphosphate), 10 μM dithiothreitol (DTT), 2.5 μM oligo(dT) primer, 10 U of SuperScript II reverse transcriptase (Invitrogen), and 8 U of RNaseOUT Recombinant Ribonuclease Inhibitor (Invitrogen). The reaction mixture was incubated at 42°C for 90 min and then at 70°C for 15 min. Quantitative RT-PCR analysis was performed with SYBR Green (Roche) and the DNA Engine Opticon system (Bio-Rad). Actin messenger RNA was measured as an internal control. Cycling conditions comprised a 3.5-min denaturing step, followed by 40 cycles of denaturing at 95°C for 1 min, annealing at 60°C for 30 s, and extension at 72°C for 30 s. The data were analyzed by Opticon Monitor software (Bio-Rad). The sequences of 21-nucleotide oligomers and the primers for RT-PCR are summarized in tables S1 and S2.

Apoptosis assays

HeLa cells expressing various RASSF proteins together with GFP were loaded with 200 nM TMRM for 15 min at 37°C and the mitochondria membrane permeability in GFP-positive cells was analyzed by FACS. The detection of sub-G1 populations was performed as described (6). Lactose dehydrogenase (LDH) activity was measured with the LDH-Cytotoxic Test Wako (Wako Pure Chemical Industries). Caspase-3 activity was measured as described with DEVD-MCA (Peptide Institute, Inc.) as a substrate, and the released MCA was measured (excitation, 355 nm; emission, 460 nm) (45).

In vitro interaction assay

Proteins with various tags were expressed in HEK293 cells. Cells from 3.5-cm plates were homogenized in 400 μl of lysis buffer A [25 mM tris-HCl (pH 7.4), 4-amidinophenylmethanesulfonyl fluoride (1 mg/liter), leupeptin (1 mg/liter), pepstatin A (1 mg/liter), and aprotinin (1 mg/liter)] containing 100 mM NaCl and 1% (w/v) Triton X-100 and centrifuged at 100,000g for 15 min at 4°C. The supernatant was incubated with antibody against FLAG on protein G Sepharose 4 fast-flow beads (GE Healthcare Bio-Sciences) or with antibody against FLAG M2 agarose gel (Sigma-Aldrich). The beads were washed three times with 25 mM tris-HCl (pH 7.4), 100 mM NaCl, and 0.33% (w/v) Triton X-100, analyzed by SDS–polyacrylamide gel electrophoresis (SDS-PAGE), and immunoblotted with antibodies against FLAG or Myc.

Immunoprecipitation

Primary cultured rat hepatocytes (4 × 107 cells) were harvested and lysed in 7.5 ml of buffer B [25 mM tris-HCl (pH 8.0), 100 mM NaCl, 1% (w/v) Triton X-100, 4-amidinophenylmethanesulfonyl fluoride (10 mg/liter), leupeptin (10 mg/liter), pepstatin A (10 mg/liter), and aprotinin (10 mg/liter)]. The lysates were centrifuged at 100,000g for 15 min at 4°C. Supernatant (3 ml) was incubated with antibody to rat RASSF6 on 10 μl of protein G Sepharose 4 fast-flow beads. The beads were washed three times with 25 mM tris-HCl (pH 8.0), 100 mM NaCl, and 0.5% (w/v) Triton X-100. The precipitates were analyzed by SDS-PAGE and immunoblotted with antibody to RASSF6 or MST2.

Pull-down assays

COS-7 cells expressing various proteins were homogenized in 500 μl of buffer B per 10-cm plate. The lysates were centrifuged at 100,000g for 15 min at 4°C. The supernatants were incubated with 250 pmol of either GST-fusion or MBP-fusion proteins and 10 μl of glutathione-Sepharose 4B (GE Healthcare Bio-Sciences) or amylose resin beads (New England BioLabs). After the beads were washed, the precipitates were analyzed by SDS-PAGE and immunoblotted with the appropriate antibodies. To test direct interaction between proteins, 50 pmol of MBP-fusion proteins were incubated with 10 pmol of GST-fusion proteins fixed on glutathione-Sepharose 4B beads.

Kinase assays

MST2 kinase assays were performed with GST-MOB1 as a substrate (32). The phosphorylation of GST-MOB1 by MST1 was performed with the same method. To evaluate the effect of RASSF6, dRASSF, and WW45 on MST2 activity, FH-MST2 was expressed in HEK293 cells in various combinations with Myc-RASSF6, Myc-dRASSF, and Myc-WW45, immunoprecipitated with antibody to FLAG M2 agarose gel, eluted with FLAG peptide, quantified by Sypro Orange staining and an FLA-3000 Image Analyzer (FUJIFILM), and used to phosphorylate GST-MOB1. To compare the effects of various RASSF proteins on MST1 and MST2, Myc-MST1 or Myc-MST2 was immunoprecipitated with antibody to Myc or coimmunoprecipitated with FLAG-RASSF proteins, quantified, and used to phosphorylate GST-MOB1. To test the effects on NDR1 activity, FLAG-NDR1 was expressed in HEK293 cells in various combinations with Myc-MST2, Myc-RASSF6, and Myc-MOB1, immunoprecipitated with antibody to FLAG M2 agarose gel, quantified by Sypro Orange staining, and used to phosphorylate the synthetic peptide for 90 min at 30°C. For LATS2 kinase assays, HEK293 cells were transfected with pcDNA LATS2-FLAG, harvested in ice-cold phosphate-buffered saline, and homogenized in 500 μl of buffer A containing 100 mM NaCl and 1% (w/v) Triton X-100. After centrifugation at 20,400g for 15 min, the supernatant was incubated for 3 hours at 4°C with antibody to FLAG M2 agarose gel. LATS2-FLAG was immunoprecipitated and washed four times with buffer A containing 1 M NaCl and 1% (w/v) Triton X-100 and twice with buffer A containing 100 mM NaCl. Kinase abundance was quantified with Sypro Orange staining. FH-MST2 alone or with Myc-RASSF6 was expressed in HEK293 cells, isolated with antibody to FLAG M2 agarose gel, eluted with FLAG peptide, and quantified by Sypro Orange staining. LATS2-FLAG (0.5 pmol) was mixed with 0.2 pmol of MST2, 5 pmol of His-YAP1, and 5 pmol of GST-MOB1 in 30 μl of 20 mM tris-HCl (pH 7.4), 10 mM MgCl2, 1 mM DTT, 2 mM Na3VO4, 20 mM β-glycerophosphate, and 20 μM [γ-32P]ATP (3000 cpm/pmol). The incubation was performed at 30°C. The reaction was stopped by the addition of 15 μl of SDS-PAGE loading buffer. The samples were analyzed by SDS-PAGE. The incorporated 32P into His6-YAP1 was counted by a liquid scintillation counter (LSC-5100, Aloka).

Statistical analysis

Statistical analyses were performed with Student’s t test for the comparison between two samples and analysis of variance (ANOVA) with Bonferroni post hoc test for multiple comparisons by GraphPad Prism software (GraphPad Software). Data expressed as percentages were subjected to arcsine square root transformation before Student’s t test or ANOVA. P values less than 0.05 were considered statistically significant.

Other procedures

Purified recombinant proteins and various amounts of bovine serum albumin were separated by SDS-PAGE. Western blotting was performed with ECL reagent (GE Healthcare Bio-Sciences) or SuperSignal West Femto (Pierce).

Acknowledgments

This study was supported by grants-in-aid for Scientific Research and Priority Areas from the Ministry of Education, Culture, Sports, Science and Technology of Japan and Araki Memorial Medical Research Fund. M.I. and A.K. were supported by the Tokyo Medical and Dental University 21st century COE program “Brain Integration and its Disorders.” We thank T. Yamamoto (University of Tokyo), Y. Gotoh (University of Tokyo), K. Emoto (National Institute of Genetics), H. Nishina (Tokyo Medical and Dental University), J. Hirayama (Tokyo Medical and Dental University), R. Honda (Tokyo Medical and Dental University), and C. Rokukawa (Tokyo Medical and Dental University) for materials, advice, and technical assistance. We especially thank T. Yoshimoto (Tokyo Medical and Dental University) and T. Masuda (Tokyo Medical and Dental University) for advice on statistical analysis.

Author contributions: A.K. started this study and contributed to figs. S1 and S2, Figs. 1, 2, 3, and 6, and Table 1. M.I. completed this study and contributed to figs. S1 to S9 and Figs. 1, 2, 4, 5, and 6. M.N. and Y.T. contributed to figs. S1 and S5 to S8 and Figs. 2, 4, 5, and 6 under the instruction of M.I.

Supplementary Materials

www.sciencesignaling.org/cgi/content/full/2/90/ra59/DC1

Fig. S1. Interaction of RASSF6 with MST2.

Fig. S2. Molecular determinants of the interaction of MST2 with RASSF6 and WW45.

Fig. S3. OA treatment induces apoptosis in rat hepatocytes.

Fig. S4. The effect of the expression of various MST2 and WW45 proteins.

Fig. S5. MST1 inhibits RASSF6-induced apoptosis, whereas RASSF6 inhibits the phosphorylation of MOB1 by MST1.

Fig. S6. RASSF6-induced apoptosis is independent of the interaction with MST2 and NDR kinases.

Fig. S7. Interaction of MOAP1 with RASSF6 and dRASSF.

Fig. S8. dRASSF interacts with and inhibits MST2.

Fig. S9. dRASSF or RASSF6 does not induce apoptosis in S2 cells.

Table S1. Twenty-one–nucleotide oligomers used in this study.

Table S2. Primers for quantitative RT-PCR.

References and Notes

  1. Abbreviations for the amino acid residues are as follows: A, Ala; C, Cys; E, Glu; F, Phe; I, Ile; K, Lys; N, Asn; R, Arg; T, Thr; V, Val; and Y, Tyr.
View Abstract

Navigate This Article