Cytoskeletal modification of Rho guanine nucleotide ... - Nature

8 downloads 0 Views 669KB Size Report
Nov 22, 2004 - RhoGEF function(s) maybe in harmony with Sept9b- containing ..... Chanduloy S, Petty EM, Kalikin, LM, Church SW, McIlroy. S, Harkin DP ...
Oncogene (2005) 24, 65–76

& 2005 Nature Publishing Group All rights reserved 0950-9232/05 $30.00 www.nature.com/onc

Cytoskeletal modification of Rho guanine nucleotide exchange factor activity: identification of a Rho guanine nucleotide exchange factor as a binding partner for Sept9b, a mammalian septin Koh-Ichi Nagata*,1,2 and Masaki Inagaki*,1 1

Division of Biochemistry, Aichi Cancer Center Research Institute, 1-1 Kanokoden, Chikusa-ku, Nagoya 464-8681, Japan

Small GTPase Rho and septin family proteins are thought to be related to tumorigenesis. We have identified a Rhoguanine nucleotide exchange factor (GEF) as a binding partner for a mammalian septin Sept9b using yeast twohybrid screening. We termed this molecule septin-associated RhoGEF (SA-RhoGEF). Molecular dissection analyses indicated that the C-terminal area of SARhoGEF exhibited binding to the N-terminal variable region of Sept9b. SA-RhoGEF was found by immunoprecipitation analysis to associate with septin complexes in REF52 fibroblast cells, maybe through direct interaction with Sept9b. Immunofluorescence analyses revealed the colocalization of SA-RhoGEF and Sept9b along with actin stress fibers in REF52 cells, and their colocalization along stress fibers was most likely to depend on their mutual interaction. In transient expression analyses, Sept9b inhibited SA-RhoGEF-dependent Rho activation in COS7 and HeLa cells. SA-RhoGEF and its fragments expressed in REF52 cells altered endogenous septin filament structures. To our knowledge, SA-RhoGEF is the first molecule providing a link between septins and Rho signaling. Oncogene (2005) 24, 65–76. doi:10.1038/sj.onc.1208101 Published online 22 November 2004 Keywords: septin; Sept9; RhoGEF; actin

Introduction The small GTPase Rho is known to play important roles in various cellular functions such as actin cytoskeletal reorganization, transcriptional activation, tumor cell invasion, cell morphology, cell motility and cytokinesis (for a review see Etienne-Manneville and Hall, 2002; Jaffe and Hall, 2002). There has been much progress in elucidating molecular mechanisms of Rho-dependent *Correspondence: Koh-ichi Nagata or Masaki Inagaki; E-mails: [email protected] or [email protected] 2 Current address: Department of Molecular Neurobiology, Institute for Developmental Research, Aichi Human Service Center, 713-8 Kamiya-cho, Kasugai 480-0392, Japan Received 6 May 2004; revised 28 July 2004; accepted 3 August 2004; published online 22 November 2004

cellular processes such as actin reorganization (Ridley, 2001; Fukata et al., 2003). Septins are involved in cytokinesis in lower eukaryotes such as yeast and fruit fly, although the precise mechanism is largely unknown (for a review see Longtine et al., 1996; Gladfelter et al., 2001). In addition, accumulating genetic and cell biological observations on yeast septins indicate that they are also required for localized chitin deposition, bud site selection, cell cycle control, plasma membrane compartmentalization and for regulating some kinases (for a review see Field and Kellogg, 1999; Gladfelter et al., 2001; Faty et al., 2002). As for mammalian cells, 12 septin genes (Sept1–12) have been identified and some septin transcripts undergo complex splicing, showing the presence of numerous members of mammalian septin family proteins (Macara et al., 2002; Kinoshita, 2003). Sept9 was first identified as a fusion partner gene of mixed lineage leukemia in a case of therapy-related acute myeloid leukemia with a t(11;17)(q23;q25) (Osaka et al., 1999; Taki et al., 1999). Altered expression of Sept9 or deletion of the gene has been reported in some cases of breast and/or ovarian cancers, implying that Sept9 is a candidate for tumor suppressor gene (Kalikin et al., 2000; Russell et al., 2000; Burrows et al., 2003). The mutations observed may be associated with allelic loss of the 17q25 region (Saito et al., 1993; Theile et al., 1995; Kalikin et al., 1997; Phelan et al., 1998). Sept9 was recently demonstrated to interact with microtubules and to play an essential role during cytokinesis (Surka et al., 2002; Nagata et al., 2003). Although these findings provided insights into a possible role for Sept9 in leukemogenesis and oncogenesis, not only the role(s) of Sept9 in tumorigenesis but also the function(s) of Sept9 in normal cells have remained to be elucidated. It is therefore likely to be essential to identify Sept9-interacting proteins and analyse their modes of action at the molecular level for better understanding the functions of Sept9. In the present study, we identified a Rho-specific guanine nucleotide exchange factor (RhoGEF), termed SA-RhoGEF (septin-associated RhoGEF), to be a septin-binding partner. SA-RhoGEF, which binds with Sept9b in vitro and in vivo, colocalizes along with actin stress fibers and interacts with Sept9b in REF52 cells. In the COS7 cell expression system, GTP-loading activity of SA-RhoGEF on Rho was inhibited by Sept9b. SA-

A RhoGEF as a septin-binding protein K-I Nagata and M Inagaki

66

RhoGEF-mediated stress fiber formation was also blocked by coexpression of Sept9b in HeLa cells. These observations provide a clue linking septins, actin filaments and Rho signaling, and may represent initiation of Sept9 function in tumorigenesis.

Results and discussions Identification of SA-RhoGEF as a Sept9b-binding partner by yeast two-hybrid screening To clarify the physiological significance of Sept9, we attempted to identify binding partners for Sept9b. A yeast two-hybrid cDNA library (human brain) was screened using full-length Sept9b as a bait. Partial cDNA sequence analysis revealed that one clone showing a strong positive interaction with Sept9b

corresponded to the C-terminal region (starting at aa 646) of KIAA0521(Genebank Accession no. AB011093), which was first isolated from a human brain cDNA library in a random cloning project (Nagase et al., 1998) (Figure 1a). As depicted in Figure 1a and b, this molecule contains a tandem of DH and PH domains, being the DH domain closely related to those of Lbc (Toksoz and Williams, 1994) (55% identity, 72% similarity), PDZ-RhoGEF/KIAA0380 (Togashi et al., 2000) (29% identity, 46% similarity), p115-RhoGEF (Hart et al., 1996) (31% identity, 45% similarity) and Lsc (Whitehead et al., 1996) (30% identity, 45% similarity). As for the PH domain, the sequence revealed 55% identity and 72% homology to Lbc, 31% identity and 49% homology to PDZRhoGEF. The DH and PH domains were more distantly related to the DH domain of Dbl (15% identity, 31% similarity) and PH domain of pleckstrin

Figure 1 Characterization of SA-RhoGEF, a Sept9-binding protein. (a) Structure of SA-RhoGEF. Structural domains of the proteins are abbreviated as follows: DH, Dbl-homology; PH, pleckstrin-homology; Pro, proline-rich motif. Position of the original fragment identified in the screening is underlined. Numbers refer to amino-acid position. (b) Sequence comparison of SA-RhoGEF with proteins possessing DH and PH domains. Accession numbers: SA-RhoGEF, AB011093; Lbc, AB055890; PDZ-RhoGEF, AB002378; p115RhoGEF, U64105; Lsc, U58203; proto-Dbl, P10911; pleckstrin, P08567. Numbers refer to amino-acid position. (c) Northern blot analysis of SA-RhoGEF. Molecular size markers are at left Oncogene

A RhoGEF as a septin-binding protein K-I Nagata and M Inagaki

67

(14% identity, 20% similarity), respectively. This molecule also contains a Pro-rich motif, which may be involved in protein–protein interactions. We named this putative RhoGEF as Septin-Associated (SA)-RhoGEF. The transcript with approximately 7.5 kb detected in all tested tissues, except the liver, skeletal muscle and testis (Figure 1c), suggesting that SA-RhoGEF may function in various tissues.

SA-RhoGEF-FL (Figure 2D, a–c) or -DH/PH (Figure 2D, g–i) showed a flat phenotype with enhanced actin stress fiber and cortical F-actin formation. It is notable that Myc-SA-RhoGEF-FL is partially colocalized with stress fibers at cell peripheral areas in the cells (Figure 2D, a–c). This may suggest masking of SA-RhoGEFbinding sites on stress fibers in the central region of the

SA-RhoGEF activates Rho in vitro and in vivo Since the DH domain is responsible for the nucleotide exchange activity of GEFs, the identified molecule might represent an exchange factor for Rho family proteins. We thus asked if SA-RhoGEF directly activates Rho family proteins both in vitro and in vivo. We engineered expression plasmids for full-length SA-RhoGEF (FL), N-terminal fragment containing the DH/PH domains (DH/PH), a mutant termed D283–298, in which aa 283– 298 of the DH/PH fragment was deleted, and Cterminal fragment (C, aa 646–1015) containing the Pro-rich motif (Figure 2a). As for in vitro experiments, recombinant Rho, Rac and Cdc42 were first complexed with nonradioactive GDP, and then incubated with [35S]GTPgS in the presence or absence of GST-SARhoGEF-DH/PH or -D283–298. As shown in Figure 2b, the DH/PH accelerated the binding of GTPgS to Rho, but not to Rac or to Cdc42. In contrast, D283–298 showed no GEF activity on Rho. These results indicate that SA-RhoGEF is a Rho-specific GEF in vitro and the DH domain is essential for Rho activation as is the case of other RhoGEFs (Figure 2b). On the other hand, some Dbl family proteins show different substrate specificities in vitro and in vivo (Olson et al., 1996; Crespo et al., 1997). We therefore determined if SARhoGEF would activate Rho specifically in vivo, using metabolic labeling with [32P]orthophosphate. FLAGRho, -Rac or -Cdc42 was expressed in COS7 cells with or without Myc-SA-RhoGEF-FL, -DH/PH or -D283– 298. We then examined the effects of these constructs regarding activation of the GTPases as described (Togashi et al., 2000). As shown in Figure 2c, coexpression of FL increased the GTP-bound form of Rho, but not Rac or Cdc42. Although SA-RhoGEF-DH/PH activated Rho in the assay, D283–298 did not alter the GTP-binding level of Rho under the same conditions. These results are consistent with in vitro observations and strongly suggest that SA-RhoGEF is a physiologically Rho-specific GEF and the DH domain is essential for Rho activation. RhoGEFs as well as Rho-GTPase activating proteins (GAPs) are important regulators of Rho-dependent actin polymerization. To confirm the involvement of SA-RhoGEF in actin reorganization, we expressed Myc-tagged SA-RhoGEF-FL, -DH/PH, -C or -D283– 298 into HeLa cells. Cells expressing the constructs were identified by staining the Myc-epitope and F-actin organization was examined. A population (B20%) of cells expressing SA-RhoGEF-FL or -DH/PH exhibited a compacted structure with saturated F-actin (data not shown). However, majority of the cells expressing

Figure 2 Specific activation of Rho by SA-RhoGEF in vitro and in vivo. (A) Schematic representation of full-length SA-RhoGEF and its mutants. Positions of DH and PH domains, Pro-rich motif and the deleted region (aa 283–298) are shown. Numbers refer to amino-acid position. (B) Specific activation of Rho by SARhoGEF in vitro. Rho, Rac, Cdc42, SA-RhoGEF-DH/PH and D283–298 were expressed as GST-fusion proteins in E. coli and affinity-purified using glutathione–Sepharose beads. Radioactivities of [35S]GTPgS binding to Rho, Rac or Cdc42 were measured as described in ‘Materials and methods’. (C) Specific activation of Rho by SA-RhoGEF in vivo. FLAG-tagged Rho, Rac or Cdc42 was transiently expressed with or without Myc-SA-RhoGEF-FL, DH/PH or D283–298 in COS7 cells. After metabolic labeling with [32P]orthophosphate, cells were lysed and the FLAG-tagged GTPases were immunoprecipitated. Radioactive nucleotides bound to the GTPases were eluted and resolved by thin-layer chromatography. The positions of GDP and GTP standards are indicated. (D) Actin reorganization in HeLa cells expressing various truncated mutants of SA-RhoGEF. Cells expressing Myc-SARhoGEF-FL (a–f), -DH/PH (g–i), -C (j–l) and -D283–298 (m–o) were fixed after 15 h of transfection. Myc-epitope (a, d, g, j and m) and F-actin (b, e, h, k and n) were double-stained with 9E10 and anti-actin antibodies, respectively. The merged images are also shown (c, f, i, l and o). Scale bar, 20 mm Oncogene

A RhoGEF as a septin-binding protein K-I Nagata and M Inagaki

68

Figure 2

Continued

cells. Although the DH/PH domain also induced stress fibers, it distributed throughout the cytoplasm and obvious overlapping with stress fibers was not observed (Figure 2D, g–i). In some flat cells expressing Myc-SARhoGEF-FL, many protrusions were formed. In such cells with protrusions, Myc-SA-RhoGEF-FL was enriched in the protrusions and colocalized with polymerized actin there (Figure 2D, d–f). On the other hand, Myc-SA-RhoGEF-C (Figure 2D, j–l) and -D283–298 (Figure 2D, m–o) localized in the cytoplasm and did not induce actin reorganization in the cells, indicating the importance of the DH domain for Rho activation followed by actin reorganization. The observations with D283–298 are consistent with the results that this mutant did not activate Rho in vitro and in vivo (Figure 2b and c). Mapping of interactive regions of SA-RhoGEF and Sept9b, and in vivo interaction of these proteins Molecular dissection analyses using the yeast twohybrid assay were carried out to determine the domains in SA-RhoGEF and Sept9b essential for interaction with each other. As shown in Figure 3a, the N-terminal 147 amino acids of Sept9b is sufficient for interaction with SA-RhoGEF-C (aa 646–1015). On the other hand, Oncogene

SA-RhoGEF-C contains the binding domain to Sept9b, while the proline-rich motif is not required for the binding (Figure 3b). The data obtained by two-hybrid analyses suggest that SA-RhoGEF and Sept9b make an in vivo complex in cells. To test this hypothesis, FLAG-Sept9b, -Sept7, Sept6, -Sept2, -Sept8 and -Sept4 were expressed with Myc-SA-RhoGEF in COS7 cells. Figure 3c shows that SA-RhoGEF can be co-immunoprecipitated specifically with Sept9b, but not with Sept7, Sept6, Sept2, Sept8 and Sept4 under the conditions used. Taken together, our results imply that in cells SA-RhoGEF interacts with septin filaments through direct association with Sept9b (see below). Interaction of endogenous SA-RhoGEF with Sept9bcontaining septin complexes in REF52 cells To further explore the interaction between Sept9b and SA-RhoGEF, we developed a rabbit polyclonal antibody (anti-SA-RhoGEF-N) against the bacterially synthesized SA-RhoGEF-DH/PH. It was affinity-purified on a column to which the antigen had been conjugated. Specificity of the antibody was confirmed by Western blot analyses using lysates from COS7 cells expressing Myc-SA-RhoGEF or PDZ-RhoGEF

A RhoGEF as a septin-binding protein K-I Nagata and M Inagaki

69

Figure 3 Specific interaction of Sept9b with SA-RhoGEF. (a) Interactions of SA-RhoGEF-C (aa 646–1015) with various fragments of Sept9b. Numbers refer to amino-acid positions. Y190 cells cotransformed with pYTH9 harboring various truncated mutants of Sept9b and pACT2-SA-RhoGEF-C were analysed for growth on medium lacking histidine, tryptophan and leucine but with 3-aminotriazole. The plus sign represents the growth of the transformed yeast colonies in 3 days. Minus sign represents failure of growth of the transformed yeast colonies in 7 days. (b) Interactions of various truncated mutants of SA-RhoGEF with Sept9b. pACT2 vectors harboring various SA-RhoGEF fragments were cotransformed into yeast Y190 cells with pYTH9-Sept9b. The interactions were determined as in (b). (c) Co-immunoprecipitation of Myc-SA-RhoGEF with FLAGSept9b, -Sept7, -Sept2, -Sept8 or -Sept4. Lysates (90 mg of protein) from transfected COS7 cells were immunoprecipitated with antiMyc polyclonal antibody (Santa Cruz Inc.). Western blotting was carried out with a mixture of 9E10 and M2. The upper panel is a Western blot showing total protein levels in transfected cells and the lower panel is a Western blot of immunoprecipitated material. Molecular size markers are at left

(Fukuhara et al., 1999) (Figure 4a). We also tested if anti-SA-RhoGEF-N specifically precipitates SA-RhoGEF with Sept9b expressed in COS7 cells. As shown in Figure 4b, anti-SA-RhoGEF-N precipitated SA-RhoGEF with Sept9b, but not with Sept7 or Sept2, confirming the results in Figure 3c. The association between SA-RhoGEF and Sept9b was lost when Sept9b mutants lacking N-terminus (Sept9b-DN, Sept9b-N3

and Sept9b-C) and/or an SA-RhoGEF mutant lacking C-terminus (DH/PH) were used instead of the wild types (Figure 4b and data not shown). These results support the observations in the mapping analyses (Figure 3a and b) that the N-terminal region of Sept9b and the Cterminal region of SA-RhoGEF are essential for their association. In the next set of experiments, we detected endogenous SA-RhoGEF with an apparent molecular mass of 115 kDa in lysates from REF52 and COS7 cells by Western blotting (Figure 4c). Preincubation of the antibody with the antigen inhibited the immunoreactivity (Figure 4c). We then asked if SA-RhoGEF is physiologically associated with Sept9b in cells. We used REF52 cells for this experiment since the cell dominantly expresses Sept9b among Sept9 subfamily proteins (Figure 4c and data not shown), and thus we can attribute obtained data to one splicing variant, Sept9b. When we immunoprecipitated endogenous Sept9b from REF52 cell lysate, SA-RhoGEF was detected in the immunoprecipitate (Figure 4D, a and b). Since septins are a family of heteropolymeric filament-forming proteins, it is most likely that SA-RhoGEF interacts with septin complexes through direct association with Sept9b in REF52 cells. To test this possibility, we tried to detect other septin molecules in the immunoprecipitates. Consequently, endogenous Sept7, Sept8 and Sept11 were detected in the immunoprecipitates by Western blotting using respective antibodies (Figure 4D, c–e), supporting the above hypothesis. Actin was also detected in the immunoprecipitate, maybe reflecting the septin–actin interaction as described (Kinoshita et al., 1997, 2002; Trimble, 1999; Xie et al., 1999) (Figure 4D, f). On the other hand, tubulin, which has been recently reported to be involved in septin structure and functions (Surka et al., 2002; Nagata et al., 2003), was hardly detected in the precipitate (Figure 4D, g) and vimentin was not detected (Figure 4D, h). Localization of SA-RhoGEF in REF52 cells We examined the subcellular localization of SA-RhoGEF in REF52 cells, using a confocal microscope. REF52 cells were double-stained with anti-actin antibody and anti-SA-RhoGEF-N or anti-Sept9. As shown in Figure 5A, both SA-RhoGEF and Sept9b were significantly localized along with stress fibers. It is reported that Sept9 proteins are localized along microtubules completely and partially in HMEC and HeLa cells, respectively (Surka et al., 2002; Nagata et al., 2003). HMEC cells express only Sept9a among Sept9 subfamily proteins and HeLa cells express Sept9a–c, suggesting that Sept9a is specifically localized along microtubules and Sept9b is present along stress fibers. Alternatively, one septin isoform is possible to interact with different cytoskeletons in cell and/or tissue typespecific manners as suggested recently for Sept11; Sept11 is colocalized with microtubules and stress fibers in HMEC and REF52 cells, respectively (Hanai et al., 2004). Further studies are required to clarify the molecular manners of septin–cytoskeleton interactions. Oncogene

A RhoGEF as a septin-binding protein K-I Nagata and M Inagaki

70

Figure 4 In vivo association of SA-RhoGEF with Sept9b-containing septin complex. (A) Western blot analyses with anti-SARhoGEF-N. SDS-sample loading buffer containing 10 mM dithiothreitol was used for extraction. The extracts from COS7 cells (30 mg) expressing Myc-SA-RhoGEF (left lane of each panel) or Myc-PDZ-RhoGEF (right lane of each panel) were immunoblotted with antiSA-RhoGEF-N (left panel) or 9E10 (right panel). Molecular size markers are at left. (B) Immunoprecipitation analyses with anti-SARhoGEF-N. Myc-SA-RhoGEF-FL (FL) or -DH/PH (DH/PH) was transfected into COS7 cells with FLAG-Sept9b, -Sept9b-DN, Sept7 or -Sept2. For negative control, FLAG-Sept9b alone was transfected. Immunoprecipitation was carried out with anti-SARhoGEF-N as described in ‘Materials and methods’. The left panel is the Western blot of the immunoprecipitated materials. Expression of each construct was confirmed by Western blotting of cell lysates with 9E10 (right panel). (C) Detection of endogenous SA-RhoGEF in COS7 and REF52 cells. COS7 and REF52 cell lysates (30 mg) were stained with Coomassie brilliant blue (CBB) (left panel), immunoblotted with anti-SA-RhoGEF-N (middle panel) or with the antibody preabsorbed by the antigen (right panel). AntiSA-RhoGEF-N recognized a band of about 115 kDa in samples obtained from COS7 and REF52 cells. (D) Interaction of SARhoGEF with Sept9b-containing septin complexes in vivo. Immunoprecipitates prepared from REF52 cell lysates, using anti-Sept9 (right lane in each panel) and control rabbit IgG (center lane in each panel), were suspended in SDS-sample loading buffer containing 10 mM N-ethylmaleimide and subjected to Western blot analysis with anti-Sept9 (a), anti-SA-RhoGEF-N (b), anti-Sept7 (c), anti-Sept8 (d), anti-Sept11 (e), anti-actin (f), anti-tubulin (g) and anti-vimentin (h). Whole-cell lysates (Input, right lane in each panel) were also used as a control. Positions of Sept9b (a) and SA-RhoGEF (b) are indicated by arrowheads Oncogene

A RhoGEF as a septin-binding protein K-I Nagata and M Inagaki

71

Figure 5 SA-RhoGEF is colocalized with Sept9b in REF52 cells. (A) REF52 cells were double-stained using anti-SA-RhoGEF-N (a) or anti-Sept9 (d) with anti-actin (b and e). The merged images are also shown (c and f). (B) After 16 h of transfection of GFP-SARhoGEF-FL (a–c), -DH/PH (d–f), -C (g–i), GFP-Sept9b (j–l) or GFP-Sept9b-DN (m–o), REF52 cells were prepermeabilized with 80 mM Pipes (pH 7.2) containing 5 mM EGTA, 1 mM MgCl2 and 0.1% saponin for 1 min at 371C, fixed and stained for F-actin (b, e, h, k and n). The merged image is also shown (c, f, i, l and o). Bar, 20 mm

We also asked if exogenously expressed SA-RhoGEF and Sept9b are localized along stress fibers. In the experiments, we used a prepermeabilization method to wash excessively expressed proteins away. As shown in Figure 5B, GFP-SA-RhoGEF-FL (a–c) and -C (g–i) were partially but clearly colocalized along with stress fibers, whereas the DH/PH domain was not (d–f). It should be noted here that Myc-SA-RhoGEF-C did not

colocalize with actin filaments in HeLa cells (Figure 2D, j–l). We assume that the discrepancy was due to the epitope type (Myc or GFP) and/or cell types. GFP-SARhoGEF-DH/PH, lacking the Sept9b-binding domain, distributed throughout the cytoplasm and did not show colocalization along stress fibers (Figure 5B, d–f). GFPSept9b also colocalizes along with stress fibers (Figure 5B, j–l), while Sept9b-DN, lacking SA-RhoGEF-binding Oncogene

A RhoGEF as a septin-binding protein K-I Nagata and M Inagaki

72

domain, was not colocalized along stress fibers but distributed in the cytoplasm (Figure 5B, m–o). These results suggest that the interaction between SA-RhoGEF and Sept9b may be essential for their distribution along stress fibers, although the precise molecular mechanism should be elucidated. These observations also raise the possibility that the N-terminal of Sept9b and/or the C-terminal of SA-RhoGEF are important for interaction with yet unidentified protein(s) linking these molecules (and septin filaments) to stress fibers. In this context, as an adaptor protein linking septin filaments and actin ones, anillin was recently identified (Kinoshita et al., 2002). Anillin seems to be important as a physical and structural linker in actin–septin interaction. It is not clear if some physiological relationship is present between anillin and SA-RhoGEF. Alteration of the septin filament structure in REF52 cells by expression of SA-RhoGEF and its fragments Since SA-RhoGEF interacts with Sept9b, we asked if SA-RhoGEF has effects on the septin structure as well as actin filament one in REF52 cells. As shown in Figure 6, expression of SA-RhoGEF or the fragments induced characteristic alteration of the Sept9b-containing septin filament structure. SA-RhoGEF or its DH/ PH fragment disturbed the stress fiber-like septin structure, although some filaments were still observed (Figure 6, a–d). These results suggest that yet unidentified effector molecule(s) activated by the SA-RhoGEF/ Rho signaling pathway is involved in the septin filament reorganization in REF52 cells. On the other hand, septin filaments disappeared when SA-RhoGEF-C was

Figure 6 Expression of SA-RhoGEF and its fragments induces morphological change of the septin filament structure in REF52 cells. After 16 h of transfection of Myc-SA-RhoGEF-FL (a and b), -DH/PH (c and d), -C (e and f), REF52 cells were fixed and stained for Sept9b (a, c and d) and Myc-tag (b, d and f). Bar, 20 mm Oncogene

expressed (Figure 6, e and f). The observed disruption and disappearance of the septin filaments may be caused by the interference of septin–septin interactions by the C-terminal fragment, since the N-terminal SA-RhoGEF-binding region of Sept9b is also essential for interaction with other septin molecules (Nagata and Inagaki, submitted). Inhibition of SA-RhoGEF-dependent Rho activation by Sept9b In the next set of experiments, we examined the effects of Sept9b on SA-RhoGEF-dependent Rho activation using the immobilized GST-RBD as a bait for selective binding of Rho-GTP (Ren et al., 1999). In COS7 cells, FLAG-Rho was coexpressed with Myc-SA-RhoGEF, Myc-PDZ-RhoGEF, FLAG-Sept7, FLAG-Sept8 and FLAG-Sept9b in various combinations. The lysates of the transfectants were incubated with GST-RBD bound on glutathione beads, and the precipitates were immunoblotted with M2 antibody. As shown in Figure 7a, SA-RhoGEF-induced Rho activation was inhibited by the presence of Sept9b, while PDZ-RhoGEF-induced activation was hardly affected by coexpression of Sept9b. On the other hand, when effects of Sept7 and Sept8 on SA-RhoGEF-induced Rho activation were analysed, they were seen to have little effects on the SARhoGEF-induced Rho activation (Figure 7b). These results suggest that Sept9b negatively regulates SARhoGEF activity in vivo. The effects of Sept9b on SARhoGEF-dependent signaling were further investigated by monitoring stress fiber formation in cells. HeLa cells were used in the experiments, since the cell has low level of stress fibers compared to REF52 cells and thus it is easy to monitor changes of stress fibers. As shown in Figure 7C, stress fiber formation by SA-RhoGEF was inhibited by coexpressing Sept9b (a–f) but not Sept7 (g– h). These observations again support the hypothesis that Sept9b negatively regulates SA-RhoGEF activity in vivo. Since artificial protein expression may be toxic to the cells and disrupt actin structure, we transfected FLAGSept9bDN, a Sept9b mutant lacking the SA-RhoGEFbinding region (aa 1–147), with GFP-SA-RhoGEF. As shown in Figure 7C, j–l, Sept9bDN did not inhibit SARhoGEF-mediated stress fiber formation. As a next set of experiments, we tried to test if Sept9b-N2, SARhoGEF-binding domain of Sept9b, is sufficient to inhibit the RhoGEF activity in vivo. We expressed FLAG-Rho and Myc-SA-RhoGEF with various amount of FLAG-Sept9b-N2, and analysed GTPbinding status of FLAG-Rho. Consequently, Sept9bN2-induced inhibition of SA-RhoGEF activity was not observed under the conditions used. Next, we expressed various amount of Sept9b-N2 with SA-RhoGEF in HeLa cells, and analysed Rho activation by monitoring stress fiber formation. Inhibition of SA-RhoGEFdependent Rho activation was not observed in the assay. One possible explanation of these results is that full-length Sept9b is required to function as scaffolds for SA-RhoGEF and to regulate its activity. However, it cannot be ruled out that the expression level is

A RhoGEF as a septin-binding protein K-I Nagata and M Inagaki

73

Figure 7 Sept9b inhibits SA-RhoGEF-dependent Rho activation. (A) COS7 cells were transfected with FLAG-RhoA, Myc-SARhoGEF, Myc-PDZ-RhoGEF and FLAG-Sept9b in various combinations. Whole-cell lysates were prepared, and a portion of each was subjected to SDS–PAGE and Western blotting with M2 and 9E10 to measure expressed proteins (lower panel). The remaining portion of the lysates was affinity-precipitated with the GST-fused RBD of Rhotekin and immunoblotted with M2. Only activated Rho binds to the GST-RBD (top panel). The amounts of precipitated GST-RBD in each sample were detected using an anti-GST antibody (Santa Cruz Inc.) (middle panel). (B) COS7 cells were transfected with FLAG-Rho, Myc-SA-RhoGEF, FLAG-Sept9b, FLAG-Sept7 and FLAG-Sept8 in various combinations. The cell extracts were analysed as in (A). Activated Rho bound to the RBD was shown in the top panel. GST-RBD in the precipitates and expressed proteins were visualized as in the middle and lower panels, respectively. (C) After 24 h of transfection of pEGFP-SA-RhoGEF (0.2 mg) with 0.6 mg of pRK5-FLAG-vector (a–c), -FLAG-Sept9b (d–f), -FLAGSept7 (g–i), or -FLAG-Sept9bDN, HeLa cells were fixed and stained for FLAG-tag (b, e, h and k) and F-actin (c, f, i and l). Bar, 20 mm

insufficient to inhibit SA-RhoGEF activity, since more than 2.0 and 0.8 mg of Sept9b-N2 cDNA in six- and 24well culture vessels, respectively, led to cytotoxicity (data not shown). The obtained results might suggest that septin filaments function as scaffolds for SA-RhoGEF through direct interaction with Sept9b, and keep SA-RhoGEF in

an inactive state. If this is the case, Sept9b may play an important role in SA-RhoGEF regulation since it specifically inhibited SA-RhoGEF-mediated Rho activation. However, we assume that other septins forming physiological complexes with Sept9b are also important since Sept9b is most likely to function as a component of heteroseptin filaments. It is therefore an open question Oncogene

A RhoGEF as a septin-binding protein K-I Nagata and M Inagaki

74

as to how Sept9b in tightly regulated septin filament networks controls SA-RhoGEF function(s). In this context, it is tempting to speculate that regulation of SA-RhoGEF activity by Sept9b is modified by interaction of Sept9b with other septin molecule(s) forming physiological complexes with Sept9b. Alternatively, proteins other than septins are possible to control SARhoGEF function(s) maybe in harmony with Sept9bcontaining septin filaments. It is thus essential to identify SA-RhoGEF-binding proteins and to analyse their functions. In addition, the identity of the signal which releases SA-RhoGEF from Sept9b-containing septin filaments and induces its subsequent activation remains to be elucidated. The presence of up to 18 possible splice variants of Sept9 has been demonstrated (McIlhatton et al., 2001). Among these variants, the protein coding regions of Sept9-beta and -zeta are identical and lacking the Nterminal SA-RhoGEF-binding domains, although the two mRNAs have different 50 UTRs. It is reported that expression of Sept9-zeta is detectable in the majority of ovarian tumors and cell lines, but not in the range of nonmalignant adult and fetal tissues (Burrows et al., 2003). It is notable that Sept9-zeta expression is accompanied by loss of the ubiquitous Sept9-beta transcript (Burrows et al., 2003). It is thus suggested that loss of expression of one Sept9 isoform can be compensated for by alteration in the expression of other isoforms. It is not known if altered Sept9 gene expression affects the septin filament conformation and/or activation status of SA-RhoGEF in tumor cells. There is growing evidence for crosstalk between the septin filaments and actin cytoskeleton, which is known to be regulated by Rho family proteins, although molecular mechanisms governing the crosstalk have so far been unknown. In the present study, we identified SA-RhoGEF for the first time as a possible component interconnecting septin filaments and Rho-signaling system. SA-RhoGEF is possible to utilize septin filament as a scaffold rather than an anillin-like structural linker. Biochemical and biological interactions between Sept9b and SA-RhoGEF, in addition to their colocalization, may provide a clue to elucidate the functional relationship between septins and Rho signaling, and may represent initiation of Sept9 function in tumorigenesis. Further extensive studies are essential for understanding the physiological significance of septindependent SA-RhoGEF regulation in cells.

Materials and methods Plasmid construction Human Sept9b, Sept7, Sept6, Sept2, Sept8 and Sept4 were produced by PCR with Marathon-Ready cDNA (human brain) (Clontech Inc.) and then subcloned into the mammalian expression vectors pRK5-FLAG and/or pEGFP-C2 (Clontech Inc.). The cDNA fragments of Sept9b (FL, amino acid (aa) 1– 568), Sept9bDN (aa 148–568), Sept9b-N1 (aa 1–258), Sept9bN2 (aa 1–147), Sept9b-N3 (aa 148–258), Sept9b-Cent (aa 259– Oncogene

543) and Sept9b-C (aa 544–568) were produced by PCR and subcloned into the yeast GAL4 DNA-binding domain vectors pYTH9, pRK5-FLAG and/or pEGFP-C2. The cDNA of SARhoGEF/KIAA0521 was kindly provided from Dr T Nagase (Kazusa DNA Institute, Japan) (Nagase et al., 1998) and the cDNA fragments, FL (aa 1–1015), Dbl homology (DH) pleckstrin homology (PH) (aa 1–645), C (aa 646–1015), Pro (aa 916–1015) and D283–298, in which aa 283–298 of the DH/ PH was deleted, were produced by PCR and constructed into the yeast GAL4 DNA-activation domain vectors pACT2, pRK5-Myc and/or pEGFP-C2. The DH/PH and D283–298 were also constructed into pGEX4T-3 (Clontech Inc.). Plasmids harboring RhoA, Rac, Cdc42 and RBD (aa 7–89 of Rhotekin) were kind gifts from Dr A Hall (University College London, England). All constructs were verified by DNA sequencing. Yeast two-hybrid analysis pYTH9-Sept9b was used in the two-hybrid screen of human brain cDNA library fused to pACT2 vector (Clontech Inc.), following the Matchmaker Two-hybrid System Protocol. Subsequent two-hybrid interaction analyses were carried out as described (Nagata et al., 1998). Northern blot analysis A human multiple tissue Northern blot (Clontech Inc.) was hybridized with a [a-32P]dCTP-labeled cDNA probe encompassing nt 2045–3795 of SA-RhoGEF sequence. Hybridization was performed in accordance with the manufacturer’s instructions. Expression and purification of recombinant proteins Rho family GTPases, SA-RhoGEF-DH/PH, D283–298 and RBD, were expressed in Escherichia coli strain DH5 as glutathione S-transferase (GST)-tagged proteins and purified according to the manufacturer’s protocol. Purity of the protein preparations was confirmed on Coomassie Blue-stained SDS– polyacrylamide gels. Preparation of antibodies GST-fused SA-RhoGEF-DH/PH served as the antigen. A rabbit polyclonal antibody specific for SA-RhoGEF (anti-SARhoGEF-N) was produced and affinity-purified. Anti-Sept9 was prepared as described (Nagata et al., 2003). Rabbit polyclonal antibodies specific for Sept7, Sept8 and Sept11 were also produced (characterization of anti-Sept7 and anti-Sept8 will be described elsewhere). Cell culture, transfection and immunofluorescence COS7, HeLa and REF52 cells were cultured as described (Nagata et al., 2003). Transient transfection was carried out in six- or 24-well cluster culture vessels using the Lipofectamine method (Gibco-BRL Inc.). For immunofluorescence analyses, cells grown on 13-mm coverslips were fixed in methanol for 15 min at 201C. Anti-SA-RhoGEF, anti-Sept9, monoclonal anti-actin (Santa Cruz Inc.), monoclonal anti-Myc 9E10 and monoclonal anti-FLAG M2 (Sigma Inc.) were used as the primary antibodies. Alexa488-, Alexa350- or Cy3-linked antibody (Molecular Probes Inc.) was used as a secondary antibody. When analysing the cells, we used an Olympus LSM-GB200 confocal microscope.

A RhoGEF as a septin-binding protein K-I Nagata and M Inagaki

75 In vitro and in vivo analyses of GDP/GTP exchange activity of SA-RhoGEF In vitro GDP/GTP exchange activity of recombinant Rho, Rac or Cdc42 (10 pmol) was measured as described (Nagata et al., 1992) in the presence or absence of recombinant GST-SARhoGEF-DH/PH (20 pmol). In vivo analysis of guanine nucleotides bound to the small GTPases was made as described (Togashi et al., 2000). Immunoprecipitation Myc-SA-RhoGEF-FL or -DH/PH was expressed in COS7 cells with various FLAG-tagged septins. Cells were lysed on ice for 20 min in lysis buffer consisting of 0.5% TritonX-100, 20 mM Tris/HCl (pH 7.5), 50 mM NaCl, 1 mM EDTA, 10 mM PMSF, 10 mg/ml leupeptin, 1 mM Na3VO4 and 10 mM NaF. Lysates were clarified by centrifugation at 15 k r.p.m. for 30 min. Expressed proteins were immunoprecipitated from cell lysates (B90 mg of protein) with anti-SA-RhoGEF-N, antiMyc polyclonal antibody (Santa Cruz Inc.) or control rabbit IgG, and protein G–sepharose beads, and washed four times with the lysis buffer. The resultant precipitates were suspended in SDS-sample loading buffer (60 mM Tris/HCl (pH 6.8) containing 2% SDS, 10% glycerol, 0.025% (wt/vol) bromophenol blue) containing 10 mM dithiothreitol, subjected to SDS–PAGE (10%), and Western blotting was carried out with 9E10 and M2 to detect proteins. Endogenous Sept9b was immunoprecipitated with the anti-Sept9 from REF52 cell lysate prepared as above. In this experiment, the resultant precipitates were suspended in SDS-sample loading buffer containing 10 mM N-ethylmaleimide to help prevent the reduction of the IgG molecule. The precipitates were analysed by Western blotting with anti-SA-RhoGEF-N, anti-Sept9, anti-Sept7, anti-Sept8, anti-Sept11, anti-actin (Santa Cruz Inc.), anti-tubulin (Sigma Inc.) and anti-vimentin 1B8 antibodies. Immunoreactive bands were visualized by making use of a horseradish-peroxidase-conjugated anti-rabbit antibody

and the ECL Western blotting detection system (Amersham Inc.). Assessment of GTP-bound Rho The RBD pull-down assay to determine the activation status of Rho was carried out as described (Ren et al., 1999). Briefly, COS7 cells were cotransfected with pRK5-FLAG-Rho (0.5 mg) and vectors expressing Myc-SA-RhoGEF, Myc-PDZ-RhoGEF, FLAG-Sept7, FLAG-Sept8 and FLAG-Sept9b (0.5 mg of each) in various combinations in 35 mm dishes. After 48 h post-transfection, the cell lysates were incubated with GSTRBD bound on glutathione beads. The expression of each protein was confirmed on Western blots using 9E10 or M2. The amount of Rho-GTP pulled down by the bead-associated RBD was detected by immunoblotting using M2.

Abbreviations GEF, guanine nucleotide exchange factor; DH, Dbl homology; PH, pleckstrin homology; GST, glutathione S-transferase; aa, amino acid; nt, nucleotide; RBD, active Rho-binding domain of Rhotekin. Acknowledgements We thank N Saitoh, C Yuhara and I Iwamoto for technical assistance. We are grateful to M Ohara for language assistance and to Dr T Nagase (Kazusa DNA Inst., Japan) for providing KIAA0521 clone. KN thanks Drs T Asano (Inst. Dev. Res., Kasugai) and Y Nozawa (Gifu International Inst. of Biotechnol., Gifu) for encouragement. This work was supported in part by grants-in-aid for scientific research and for cancer research from Ministry of Education, Science, Technology, Sports and Culture of Japan, and by the grant from Yamanouchi Foundation for Research on Metabolic Disorders.

References Burrows JF, Chanduloy S, McIlhatton MA, Nagar H, Yeates K, Donaghy P, Price J, Godwin AK, Johnston PG and Russell SE. (2003). J. Pathol., 201, 581–588. Crespo P, Schuebel KE, Ostrom AA, Gutkind JS and Bustelo XR. (1997). Nature, 385, 169–172. Etienne-Manneville S and Hall A. (2002). Nature, 420, 629–635. Faty M, Fink M and Barral Y. (2002). Curr. Genet., 41, 123–131. Field CM and Kellogg D. (1999). Trends Cell Biol., 9, 387–394. Fukata M, Nakagawa M and Kaibuchi K. (2003). Curr. Opin. Cell Biol., 15, 590–597. Fukuhara S, Murga C, Zohar M, Igishi T and Gutkind JS. (1999). J. Biol. Chem., 274, 5868–5879. Gladfelter AS, Pringle JR and Lew DJ. (2001). Curr. Opin. Microbiol., 4, 681–689. Hanai N, Nagata K, Kawajiri A, Shiromizu T, Saitoh N, Hasegawa Y, Murakami S and Inagaki M. (2004). FEBS Lett., 568, 83–88. Hart MJ, Sharma S, elMasry N, Qiu RG, McCabe P, Polakis P and Bollag G. (1996). J. Biol. Chem., 271, 25452–25458. Jaffe AB and Hall A. (2002). Adv. Cancer Res., 84, 57–80. Kalikin LM, Frank TS, Svoboda-Newman S, Wetzel JC, Cooney KA and Petty EM. (1997). Oncogene, 14, 1991–1994.

Kalikin LM, Sims HL and Petty EM. (2000). Genomics, 63, 165–172. Kinoshita M, Kumar S, Mizoguchi A, Ide C, Kinoshita A, Haraguchi T, Hiraoka Y and Noda M. (1997). Genes Dev., 11, 1535–1547. Kinoshita M, Field CM, Coughlin ML, Straight AF and Mitchison TJ. (2002). Dev. Cell, 3, 791–802. Kinoshita M. (2003). J. Biochem., 134, 491–496. Longtine MS, DeMarini DJ, Valencik ML, Al-Awar OS, Fares H, DeVirgilio C and Pringle JR. (1996). Curr. Opin. Cell Biol., 8, 106–119. Macara IG, Baldarelli R, Field CM, Glotzer M, Hayashi Y, Hsu SC, Kennedy MB, Kinoshita M, Longtine M, Low C, Maltais LJ, McKenzie L, Mitchison TJ, Nishikawa T, Noda M, Petty EM, Peifer M, Pringle JR, Robinson PJ, Roth D, Russell SE, Stuhlmann H, Tanaka M, Tanaka T, Trimble WS, Ware J, Zeleznik-Le NJ and Zieger B. (2002). Mol. Cell. Biol., 13, 4111–4113. McIlhatton MA, Burrows JF, Donaghy PG, Chanduloy S, Johnston PG and Russell SE. (2001). Oncogene, 20, 5930–5939. Nagase T, Ishikawa K, Miyajima N, Tanaka A, Kotani H, Nomura N and Ohara O. (1998). DNA Res., 5, 31–39. Nagata K, Suzuki T, Shibagaki Y, Mizumoto K, Okano Y, Kaziro Y and Nozawa Y. (1992). J. Biol. Chem., 267, 19600–19606. Oncogene

A RhoGEF as a septin-binding protein K-I Nagata and M Inagaki

76 Nagata K, Puls A, Futter C, Aspenstrom P, Schaefer E, Nakata N, Hirokawa N and Hall A. (1998). EMBO J., 17, 149–158. Nagata K, Kawajiri A, Matsui S, Takagishi M, Shiromizu T, Saitoh N, Izawa I, Kiyono T, Itoh JT, Hotani H and Inagaki M. (2003). J. Biol. Chem., 278, 18538–18543. Olson MF, Pasteris NG, Gorski JL and Hall A. (1996). Curr. Biol., 6, 1628–1633. Osaka M, Rowley JD and Zeleznik-Le NJ. (1999). Proc. Natl. Acad. Sci. USA, 96, 6428–6433. Phelan CM, Borg A, Cuny M, Crichton DN, Baldersson T, Andersen TI, Caligo MA, Lidereau R, Lindblom A, Seitz S, Kelsell D, Hamann U, Rio P, Thorlacius S, Papp J, Olah E, Ponder B, Bignon YJ, Scherneck S, Barkardottir R, Borresen-Dale AL, Eyfjord J, Theillet C, Thompson AM, Devilee P and Larsson C. (1998). Cancer Res., 58, 1004–1012. Ren XD, Kiosses WB and Schwartz MA. (1999). EMBO J., 18, 578–585. Ridley AJ. (2001). Trends Cell Biol., 11, 471–477. Russell SEH, McIlhatton MA, Burrows JF, Donaghy PG, Chanduloy S, Petty EM, Kalikin, LM, Church SW, McIlroy S, Harkin DP, Keilty GW, Cranston AN, Weissenbach J,

Oncogene

Hickey I and Johnston PG. (2000). Cancer Res., 60, 4729–4734. Saito H, Inazawa J, Saito S, Kasumi F, Koi S, Sagae S, Kudo R, Saito J, Noda K and Nakamura Y. (1993). Cancer Res., 53, 3382–3385. Surka MC, Tsang CW and Trimble WS. (2002). Mol. Cell. Biol., 13, 3532–3545. Taki T, Ohnishi H, Shinohara K, Sako M, Bessho F, Yanagisawa M and Hayashi Y. (1999). Cancer Res, 59, 4261–4265. Theile M, Hartmann S, Scherthan H, Arnold W, Deppert W, Frege R, Glaab F, Haensch W and Scherneck S. (1995). Oncogene, 10, 439–447. Togashi H, Nagata K, Takagishi M, Saitoh N and Inagaki M. (2000). J. Biol. Chem., 275, 29570–29578. Toksoz D and Williams DA. (1994). Oncogene, 9, 621–628. Trimble WS. (1999). J. Membr. Biol., 169, 75–81. Whitehead IP, Khosravi-Far R, Kirk H, Trigo-Gonzalez G, Der CJ and Kay R. (1996). J. Biol. Chem., 271, 18643–18650. Xie H, Surka M, Howard J and Trimble WS. (1999). Cell Motil. Cytoskeleton, 43, 52–62.