Involvement of Nectin in the Localization of IQGAP1 at the Cell–Cell Adhesion Sites Through the Actin Cytoskeleton in Madin–Darby Canine Kidney Cells

Involvement of nectin in the localization of IQGAP1 at the cell–cell adhesion sites through the actin cytoskeleton in Madin–Darby canine kidney cells

Tatsuo Katata, Kenji Irie, Atsunori Fukuhara, Tomomi Kawakatsu, Akio Yamada, Kazuya Shimizu and Yoshimi Takai*

Department of Molecular Biology and Biochemistry, Osaka University Graduate School of Medicine/Faculty of Medicine, Suita 565-0871, Osaka, Japan

IQGAP1, a putative downstream target of the Rho family Introduction small G proteins, Cdc42 and Rac, localizes at adherens junctions (AJs) in epithelial cells. It has been suggested Cells recognize their neighboring cells and form cell–cell that IQGAP1 localizes at AJs through its binding to junctions. This phenomenon is critical for tissue b-catenin, and negatively regulates the E-cadherin- patterning and morphogenesis as well as for the mediated cell–cell adhesion. Nectin is a Ca2+-independent, maintenance of normal tissues. Cadherin is a cell–cell immunoglobulin-like cell–cell adhesion molecule that adhesion molecule serving as a key determinant of these localizes at AJs. Nectin is associated with E-cadherin events (Gumbiner et al., 1988; Takeichi, 1988, 1991; through their respective cytoplasmic tail-binding proteins, Geiger and Ginsberg, 1991). Cadherin comprises a afadin and catenins, and involved in the formation of AJs superfamily consisting of more than 80 members, each cooperatively with E-cadherin. Here we investigated a role of which is expressed in a wide variety of cells including of nectin in the localization of IQGAP1 at AJs. Ca2+ epithelial and non-epithelial cells (Takeichi, 1995; Yagi chelation from the medium causes disruption of the E- and Takeichi, 2000). E-Cadherin, like other classical cadherin-mediated cell–cell adhesion, but not the nectin- cadherins, is a single-pass transmembrane protein whose based cell–cell adhesion, in Madin–Darby canine kidney extracellular domain mediates homophilic recognition (MDCK) cells. IQGAP1 remained at the residual nectin- and adhesive binding in a Ca2+-dependent manner based cell–cell adhesion sites where the E-cadherin (Takeichi, 1988, 1991, 1995). E-Cadherin localizes at immunofluorescence signal disappeared. Restoration of adherens junctions (AJs) and associates with the actin Ca2+ in the medium causes re-accumulation of E-cadherin cytoskeleton through peripheral membrane proteins, to the residual nectin-based cell–cell adhesion sites to re- including a-, b-, and g-catenins, a-actinin, and vinculin form AJs. Nectin inhibitors inhibit this re-accumulation of (Ozawa et al., 1989; Nagafuchi et al., 1991; Watabe- E-cadherin to re-form AJs by impairing the nectin-based Uchida et al., 1998; Weiss et al., 1998). b-Catenin cell–cell adhesion. The nectin inhibitors also reduced the directly interacts with the cytoplasmic tail of E-cadherin localization of IQGAP1 at the cell–cell adhesion sites. and connects E-cadherin to a-catenin that directly binds When MDCK cells were incubated with microbeads to F-actin (Rimm et al., 1995). a-Actinin and vinculin coated with the extracellular fragment of nectin that are also F-actin-binding proteins that directly bind to a- interacts with cellular nectin, IQGAP1 also accumulated catenin (Knudsen et al., 1995; Watabe-Uchida et al., at the bead–MDCK cell contact sites. The accumulation 1998; Weiss et al., 1998). The association of E-cadherin of IQGAP1 at the cell–cell adhesion sites was inhibited by with the actin cytoskeleton through these peripheral actin filament-disrupting agents, latrunculin A and membrane proteins strengthens the cell–cell adhesion cytochalasin D. These results indicate that nectin is activity of E-cadherin (Takeichi, 1991; Imamura et al., involved in the localization of IQGAP1 at AJs through 1999). the actin cytoskeleton. Nectin and afadin constitute another cell–cell adhe- Oncogene (2003) 22, 2097–2109. doi:10.1038/sj.onc.1206255 sion system at cell–cell AJs (Mandai et al., 1997; Takahashi et al., 1999; Takai and Nakanishi, 2003). Keywords: nectin; IQGAP1; cadherin; adherens junc- Nectin is a Ca2+-independent immunoglobulin-like cell– tions; actin cell adhesion molecule (Aoki et al., 1994; Lopez et al., 1998; Takahashi et al., 1999; Miyahara et al., 2000; Satoh-Horikawa et al., 2000; Reymond et al., 2001). Nectin comprises a family consisting of four members, nectin-1, nectin-2, nectin-3, and nectin-4, each of which *Correspondence: Y Takai; E-mail: [email protected] has two or three splicing variants (Morrison and Received 24 September 2002; revised 19November 2002; accepted 21 Racaniello, 1992; Aoki et al., 1994; Eberle´ et al., 1995; November 2002 Lopez et al., 1995; Cocchi et al., 1998, 2000; Warner A role of nectin in the localization of IQGAPI T Katata et al 2098 et al., 1998; Satoh-Horikawa et al., 2000; Reymond Rac (Hart et al., 1996; Kuroda et al., 1996; McCallum et al., 2001). Nectin-1 was originally identified as one of et al., 1996). IQGAP1 includes a calponin homology the poliovirus-receptor-related proteins, PRR-1 (Lopez domain, internal repeats, a WW domain, a calmodulin- et al., 1995). Nectin-2 was originally identified as the binding domain, and a Ras GTPase-activating protein murine homolog of human poliovirus-receptor protein homology domain (Weissbach et al., 1994; Machesky, (Morrison and Racaniello, 1992), but turned out to be 1998). It has been suggested that IQGAP1 localizes at another poliovirus-receptor-related protein, PRR-2 AJs through its binding to b-catenin, and negatively (Eberle´ et al., 1995; Lopez et al., 1995). Neither PRR- regulates the E-cadherin-mediated cell–cell adhesion by 1 nor PRR-2 has thus far been shown to serve as a interacting with b-catenin and causing the dissociation poliovirus receptor. PRR-1 and PRR-2 were later shown of a-catenin from cadherin–b-catenin–a-catenin com- to serve as receptors for a-herpes viruses, facilitating plex (Kuroda et al., 1998; Fukata et al., 1999, 2001). their entry and intercellular spreading, and renamed Another group has also suggested that E-cadherin is HveC and HveB, respectively (Cocchi et al., 1998, 2000; required for the accumulation of IQGAP1 at AJs (Li Geraghty et al., 1998; Warner et al., 1998; Lopez et al., et al., 1999). In addition, it has been suggested that 2000; Sakisaka et al., 2001). It remains unknown activated Rac1 and Cdc42 positively regulate the E- whether nectin-3 and nectin-4 serve as receptors for cadherin-mediated cell–cell adhesion by inhibiting the viruses. Each member of the nectin family forms homo- interaction of IQGAP1 with b-catenin (Fukata et al., cis-dimers, followed by the formation of homo-trans- 1999, 2001). dimers, causing cell–cell adhesion (Lopez et al., 1998; Evidence is accumulating that the nectin–afadin Miyahara et al., 2000; Satoh-Horikawa et al., 2000; system is involved in the formation of the E-cadherin- Sakisaka et al., 2001; Momose et al., 2002). Nectin-3 based cell–cell AJs (Mandai et al., 1997; Asakura et al., furthermore forms hetero-trans-dimers with either nec- 1999; Ikeda et al., 1999; Takahashi et al., 1999; tin-1 or nectin-2 and the adhesion activity of each Tachibana et al., 2000; Honda et al., 2003; Takai and hetero-trans-dimers is stronger than that of each homo- Nakanishi, 2003). Thus, the nectin–afadin system might trans-dimers (Satoh-Horikawa et al., 2000; Honda et al., have a possible interaction with IQGAP1. In this study, 2003). Nectin-4 also forms hetero-trans-dimers with we investigated a role of nectin in the localization of nectin-1 (Reymond et al., 2001). Most of the nectin IQGAP1 at AJs. Our results indicate that nectin is family members have a C-terminal conserved motif of involved in the localization of IQGAP1 at AJs through four amino acids (aa) residues which interacts with the the actin cytoskeleton. PDZ domain of afadin (Takahashi et al., 1999; Satoh- Horikawa et al., 2000; Reymond et al., 2001). Afadin Results has at least two splice variants, l-afadin and s-afadin (Mandai et al., 1997). l-Afadin, the larger splice variant, Remaining of IQGAP1 at the nectin-1a-based cell–cell binds to nectin and F-actin. l-Afadin binds to the side of adhesion sites where the E-cadherin–catenin system was F-actin, but does not crosslink it to form bundles absent (Mandai et al., 1997). l-Afadin has two Ras-associated domains, a forkhead-associated domain, a dilute do- We examined the localization of IQGAP1 in compar- main, a PDZ domain, and three proline-rich domains. ison with that of nectin-1a or E-cadherin during the F-Actin binds to the region containing the third proline- disruption of cell–cell AJs in nectin-1a-Madin–Darby rich domain (Mandai et al., 1997). s-Afadin, the smaller canine kidney (MDCK) cells (MDCK cells stably splice variant, has two Ras-associated domains, a expressing FLAG-nectin-1a). Since wild-type MDCK forkhead-associated domain, a dilute domain, a PDZ cells expressed nectin-1 and nectin-2 as estimated by domain, and two proline-rich domains, but lacks the F- Western blotting and none of them was stained by any actin-binding domain. Human s-afadin is identical to currently available antibodies (Abs) (data not shown), the product of the AF-6 gene that has been identified as we utilized nectin-1a-MDCK cells in this study. When an ALL-1 fusion partner involved in acute myeloid the Ca2+ concentration in the culture medium was leukemia (Prasad et al., 1993). In this study, l-afadin is switched from 2 mm to 2 mm, the cells gradually detached simply referred to as afadin. The nectin–afadin system is from each other as described in wild-type MDCK cells ubiquitously expressed not only in epithelial cells but (Kartenbeck et al., 1991). The staining patterns for also in nonepithelial cells such as fibroblasts, neurons, afadin, IQGAP1, and E-cadherin in wild-type MDCK and Sertoli cells (Eberle´ et al., 1995; Lopez et al., 1995; cells cultured at both 2 mm Ca2+ and 2 mm Ca2+ were Mandai et al., 1997; Nishioka et al., 2000; Satoh- similar to those in nectin-1a-MDCK cells (Figure 1a and Horikawa et al., 2000; Mizoguchi et al., 2002; Ozaki- c, control). The staining patterns for a-catenin and Kuroda et al., 2002). b-catenin in wild-type MDCK cells cultured at both The Rho family small G proteins, including Rho, 2mm Ca2+ and 2 mm Ca2+ were also similar to those in Rac, and Cdc42, are involved in the regulation of nectin-1a-MDCK cells (data not shown). Thus, the cadherin-mediated cell–cell adhesion (Braga et al., 1997; exogenous expression of FLAG-nectin-1a did not Takaishi et al., 1997). IQGAP1, which was originally disturb the behaviors of these proteins during the Ca2+ identified as a putative Ras GTPase-activating protein switch. The immunofluorescence signals for nectin-1a, on the basis of aa sequence similarity (Weissbach et al., E-cadherin, and IQGAP1 were highly concentrated at 1994), is one of the downstream targets of Cdc42 and the cell–cell adhesion sites of nectin-1a-MDCK cells

Oncogene A role of nectin in the localization of IQGAPI T Katata et al 2099

Figure 1 Remaining of the immunoﬂuorescence signal for IQGAP1 at the cell–cell adhesion sites during the culture at low Ca2+.(a) Nectin-1a-MDCK cells were cultured at 2 mm Ca2+ for 120 min (low Ca2+). The cells were triple stained with the anti-E- cadherin mAb, the anti-IQGAP1 pAb, and the anti-FLAG mAb. (b) Nectin-1a-MDCK cells were cultured at 2 mm Ca2+ for 120 min (low Ca2+). The cells were double stained with the anti-b-catenin mAb and the anti-IQGAP1 pAb. (c) Wild-type MDCK cells were cultured at 2 mm Ca2+ for 120 min (low Ca2+). The cells were triple stained with the anti-E-cadherin mAb, the anti-IQGAP1 pAb, and the anti-afadin mAb. Arrows, nectin-1a-based cell–cell adhesion sites remained at 2 mm Ca2+, and bars, 10 mm. The results are representative of three independent experiments

Oncogene A role of nectin in the localization of IQGAPI T Katata et al 2100 cultured at 2 mm Ca2+ (Figure 1a, control). The signal for b-catenin was also highly concentrated at the cell– cell adhesion sites (Figure 1b, control), and the signal for a-catenin was similar to that for b-catenin (data not shown). When these cells were cultured at 2 mm Ca2+ for 120 min, the signals for E-cadherin and b-catenin disappeared and partly appeared on intracellular vesicles (Figure 1a and b, low Ca2+, arrow). The signal for a-catenin also disappeared and partly appeared on intracellular vesicles (data not shown). The signal for nectin-1a mostly remained on the free surface of the plasma membrane and formed a ring-like structure as described (Fukuhara et al., 2002a). In addition, nectin- 1a still formed the cell–cell adhesion even in the absence of the E-cadherin-based cell–cell adhesion (Figure 1a, low Ca2+, arrow) as described (Fukuhara et al., 2002a). The signal for IQGAP1 remained on the free surface of the plasma membrane and at the nectin-1a-based cell– cell adhesion sites where the signals for E-cadherin and b-catenin were undetectable (Figure 1a and b, low Ca2+, arrow). The signal for a-catenin was also undetectable on the free surface of the plasma membrane and at the nectin-1a-based cell–cell adhesion sites (data not shown). Essentially the same results were obtained with a rabbit anti-IQGAP1 polyclonal Ab (pAb) and a mouse anti-IQGAP1 monoclonal Ab (mAb) that Figure 2 Speciﬁcity of anti-IQGAP1 Abs. The cell lysates of recognized a single band of 190 kDa in the crude extract nectin-1a-MDCK cells (20 mg of protein) were subjected to of nectin-1a-MDCK cells by Western blotting (Figure 2). SDS–PAGE (8% polyacrylamide gel), followed by Western These results indicate that IQGAP1 still remains on the blotting with the rabbit anti-IQGAP1 pAb (lane 1) and the mouse free surface of the plasma membrane and at the nectin- anti-IQGAP1 mAb (lane 2). The results are representative of three independent experiments 1a-based cell–cell adhesion sites even in the absence of the cell–cell adhesion based on the E-cadherin–catenin system. When wild-type MDCK cells, which did not 60 min, the signals for nectin-1a and E-cadherin express exogenous nectin-1a, were cultured at 2 mm Ca2+ concentrated at the cell–cell adhesion sites and showed for 120 min, the signal for IQGAP1 still remained on the belt-like staining (Figure 3, normal Ca2+ 60 min). free surface of the plasma membrane and at the residual During the formation of AJs, the signal for IQGAP1 cell–cell adhesion sites where E-cadherin was undetect- behaved similar to that of nectin-1a (Figure 3, normal able (Figure 1c, low Ca2+, arrow). Thus, the remaining Ca2+ 15 and 30 min, arrow and 60 min). The staining of IQGAP1 on the free surface of the plasma membrane patterns of a-catenin and b-catenin were essentially the in nectin-1a-MDCK cells cultured at 2 mm Ca2+ was not same as that of E-cadherin (data not shown). simply due to exogenous expression of nectin-1a. Inhibition of the accumulation of IQGAP1 at the cell–cell adhesion sites by nectin inhibitors Similar behavior of IQGAP1 and nectin-1a during the formation of AJs by increase of Ca2+ from a low to a high To obtain the evidence that the nectin–afadin system has concentration a role for the localization of IQGAP1, we utilized nectin inhibitors, the chimeric proteins of a fragment of Nectin-1a-MDCK cells precultured at 2 mm Ca2+ for glycoprotein D (1–285 aa) and an extracellular fragment 120 min were recultured at 2 mm Ca2+ for 15, 30, and of nectin-3 (56–400 aa) fused to IgG Fc, gD, and Nef-3, 60 min, and the time courses of the accumulation of respectively. gD is an envelope protein of herpes simplex IQGAP1 at the cell–cell adhesion sites were examined in virus type 1, one of a-herpes viruses (Campadelli-Fiume comparison with those of E-cadherin in detail. Before et al., 2000; Spear et al., 2000). gD binds to nectin-1a the incubation at 2 mm Ca2+, the immunoﬂuorescence and inhibits not only the formation of homo-trans- signal for E-cadherin was not observed at the nectin-1a- dimers of nectin-1a but also the formation of hetero- based cell–cell adhesion sites where the signals for trans-dimers between nectin-1a and nectin-3a (Sakisaka nectin-1a and IQGAP1 were observed (Figure 3, low et al., 2001; Mizoguchi et al., 2002). We have recently Ca2+, arrow). After the incubation at 2 mm Ca2+, the found that Nef-3 also has a similar activity (Honda et al., signal for E-cadherin gradually concentrated at the 2003). Since the mixture of gD and Nef-3 has a more nectin-1a-based cell–cell adhesion sites (Figure 3, nor- inhibitory activity than gD or Nef-3 alone in nectin-1a- mal Ca2+ 30 min, arrowhead) as described (Fukuhara MDCK cells (Fukuhara et al., 2002a, b), we utilized the et al., 2002a). After the incubation at 2 mm Ca2+ for mixture of gD and Nef-3 as nectin inhibitors in this

Oncogene oeo etni h oaiaino IQGAPI of localization the Katata in T nectin of role A tal et

Figure 3 Behavior of nectin-1a, E-cadherin, and IQGAP1 during the increase of Ca2+ from a low to a high concentration. Nectin-1a-MDCK cells were cultured at 2 mm Ca2+ for 120 min and then incubated with 2 mm Ca2+ for the indicated periods of time. The cells were triple stained with the anti-E-cadherin mAb, the anti-IQGAP1 pAb, and the anti-FLAG mAb. Arrows, the signal for IQGAP1 concentrated at the nectin-1a-based cell–cell adhesion sites where the signal for E-cadherin is undetectable; arrowheads, the signal for E-cadherin concentrated at the nectin-1a-based cell–cell adhesion sites; and bars, 10 mm. The results are representative of three independent experiments Oncogene 2101 A role of nectin in the localization of IQGAPI T Katata et al 2102 study. Nectin-1a-MDCK cells were first cultured at 2 mm indicate that the mixture of gD and Nef-3 inhibits the Ca2+ for 120 min and then cultured at 2 mm Ca2+ in the formation of the trans-dimers of nectin-1a, resulting in presence or absence of the mixture of gD and Nef-3. In the reduction of the colocalization of IQGAP1 with the absence of gD or Nef-3, the immunofluorescence nectin-1a. Thus, IQGAP1 localizes at AJs in a nectin- signals for nectin-1a, E-cadherin, and IQGAP1 concen- dependent manner. trated at the cell–cell adhesion sites as described above 2+ (Figure 4, control and see Figure 3, normal Ca Accumulation of IQGAP1 at the Nef-3-coated bead–cell 60 min). However, in the presence of the mixture of contact sites gD and Nef-3, the signal for nectin-1a concentrated at the cell–cell adhesion sites to a much lesser extent To obtain another line of evidence for a role of nectin in (Figure 4, nectin inhibitors). The signals for E-cadherin the accumulation of IQGAP1 at the cell–cell adhesion and IQGAP1 at the cell–cell adhesion sites were sites, we utilized the microbeads coated with Nef-3. We similarly reduced in the presence of the mixture of gD have recently shown that the microbeads coated with and Nef-3 (Figure 4, nectin inhibitors). These results Nef-3 recruit first the nectin–afadin complex and then the cadherin–catenin complex to the bead–cell contact sites using nectin-1a-MDCK cells and L fibroblasts stably expressing both E-cadherin and nectin-1a (Fukuhara et al., 2002a; Honda et al., 2003). Nectin- 1a-MDCK cells were cultured at 2 mm Ca2+ for 120 min in the presence of the Nef-3-coated beads, and further incubated at 2 mm Ca2+ for 180 min. The cells were then fixed, followed by immunostaining for cellular nectin-1a and IQGAP1. The immunofluorescence signals for cellular nectin-1a and IQGAP1 concentrated at the bead–cell contact sites (Figure 5, Nef-3). The signal for nectin-1a or IQGAP1 at the bead–cell contact sites was not observed when the control microbeads coated with human IgG were used (Figure 5, control). Analysis in the time courses of the concentration of the IQGAP1 and E-cadherin staining indicated that IQGAP1 accumulated at the bead–cell adhesion sites prior to E- cadherin. The IQGAP1 staining at the bead–cell contact sites was concentrated at 30 min after the incubation at 2mm Ca2+, whereas the E-cadherin staining was not observed at 30 min, but started to be concentrated at 1 h (data not shown). The time course of the concentration of the b-catenin staining was similar to that of E- cadherin. These results suggest that Nef-3 on the beads first forms trans-dimers with cellular nectin-1a, which then recruits IQGAP1 at the bead–cell contact sites. Taken together, these results indicate that IQGAP1 accumulates at the nectin-1a-mediated cell–cell or cell– bead adhesion sites, and that the trans-interaction of nectin-1a is necessary and sufficient for the accumulation of IQGAP1.

Association of IQGAP1 with nectin-1a and afadin The results, that IQGAP1 accumulated at the nectin-1a- mediated cell–cell or cell–bead adhesion sites, suggested the association of IQGAP1 with the nectin–afadin system. Therefore, we next examined whether IQGAP1 is associated with the nectin–afadin system. For this purpose, we performed immunoprecipitation analysis of Figure 4 Inhibition of the accumulation of IQGAP1 at the cell– nectin-1a-MDCK cells. When the cell extract of nectin- cell adhesion sites by nectin inhibitors. Nectin-1a-MDCK cells 1a-MDCK cells was subjected to immunoprecipitation were cultured at 2 mm Ca2+ for 120 min and then incubated with by the use of the anti-FLAG mAb, afadin was 2mm Ca2+ for 60 min in the absence (control) or presence (nectin coimmunoprecipitated with nectin-1a (data not shown) inhibitors) of the mixture of 60 mg/ml gD and 60 mg/ml Nef-3. The cells were triple stained with the anti-E-cadherin mAb, the anti- However, IQGAP1 was not coimmunoprecipitated with IQGAP1 pAb, and the anti-FLAG mAb. bars, 10 mm. The results nectin-1a (data not shown). When the cell extract of are representative of three independent experiments nectin-1a-MDCK cells was subjected to immunopreci-

Oncogene A role of nectin in the localization of IQGAPI T Katata et al 2103 Inhibition of the localization of IQGAP at the cell–cell adhesion sites by F-actin-disrupting agents

IQGAP1 is an F-actin-binding protein (Bashour et al., 1997). Therefore, we next examined by the use of F- actin-disrupting agents, latrunculin A and cytochalasin D, whether actin cytoskeletal structures are involved in the localization of IQGAP1 at the cell–cell adhesion sites. In nectin-1a-MDCK cells, F-actin was associated with the cell–cell adhesion sites where nectin-1a and IQGAP1 colocalized (Figure 7a, control). When these cells were incubated with latrunculin A for 30 min, the cells became flat and most of the actin cytoskeletal structures were disrupted (Figure 7a, LatA). Under these conditions, the immunofluorescence signal for IQGAP1 changed from belt-like structures at cell–cell junctions to randomly distributed clusters on the plasma membrane (Figure 7a, LatA). In contrast, the belt-like staining of nectin at the cell–cell adhesion sites was not markedly changed (Figure 7a, LatA). The essentially similar results were obtained with cytochalasin D (data not shown). These results are consistent with the previous observations that the nectin-based cell–cell adhesion does not require the cytochalasin D-sensitive actin cytoskeleton (Yokoyama et al., 2001), and that IQGAP1 colocalizes with cytochalasin D-sensitive F- actin at lamellipodia and membrane ruffles (Bashour et al., 1997). The coimmunoprecipitation analysis also revealed that the incubation of nectin-1a-MDCK cells with latrunculin A decreased the amounts of afadin and nectin-1a co-immunoprecipitated with IQGAP1 (data not shown). Next, nectin-1a-MDCK cells precultured at 2 mm Ca2+ for 120 min were incubated with or without latrunculin A. Without latrunculin A, the immunofluorescence signals for nectin-1a and IQGAP1 mostly remained on the free surface of the plasma membrane and formed a ring-like structure as described above (Figure 1a, low Ca2+ and Figure 7b, low Ca2+). With latrunculin A, the Figure 5 Accumulation of IQGAP1 at the Nef-3-coated bead–cell signal for IQGAP1 mostly disappeared from the free contact sites. The Nef-3-coated or control beads were added to nectin-1a-MDCK cells, followed by incubation at 2 mm Ca2+ at surface of the plasma membrane and the cell–cell 371C for 120 min. After the incubation, the cells were washed with adhesion sites, while the signal for nectin-1a mostly DMEM and further incubated at 2 mm Ca2+ for 180 min. Then, the remained at the cell–cell adhesion sites (Figure 7b, low cells were fixed, followed by immunostaining for IQGAP1 and Ca2+ LatA). nectin-1a using the anti-IQGAP1 and anti-nectin-1a Abs, respec- Nectin-1a-MDCK cells precultured at 2 mm Ca2+ for tively. The positions of the beads are marked with asterisks. DIC, a differential interference contrast image; arrows, bead–cell contact 120 min were incubated with or without latruculin A. sites; and bars, 10 mm. The results are representative of three The cells were then recultured at 2 mm Ca2+ for 60 min, independent experiments. and the accumulation of IQGAP1 at the cell–cell adhesion sites was examined. Without latrunculin A, the immunofluorescence signal for IQGAP1 accumu- pitation by the use of the anti-IQGAP1 mAb, afadin and lated at the cell–cell adhesion sites where the signal for nectin-1a were not coimmunoprecipitated with IQGAP1 nectin-1a colocalized (Figure 3, normal Ca2+ 60 min). (Figure 6, cross linkerÀ). Then, we exposed nectin-1a- With latruculin A, the signal for IQGAP1 did not MDCK cells with a cross linker, dithiobis succinimidyl accumulate at the cell–cell adhesion sites, while the propionate (DSP), before the cell lysis. Under these signal for nectin-1a mostly accumulated at the cell–cell conditions, afadin and nectin-1a were coimmunopreci- adhesion sites (Figure 7b, normal Ca2+ LatA). The pitated with IQGAP1 (Figure 6, cross linker+). Neither essentially similar results were obtained with cytochala- E-cadherin nor b-catenin was coimmunoprecipitated sin D (data not shown). These results suggest that the with IQGAP1 (data not shown). These results suggest latrunculin A- or cytochalasin D-sensitive actin cytos- that IQGAP1 is associated with the nectin–afadin keletal structures are essential for the localization of system. IQGAP1 at the cell–cell adhesion sites.

Oncogene A role of nectin in the localization of IQGAPI T Katata et al 2104

Figure 6 Association of IQGAP1 with nectin-1a and afadin. The cell extract of nectin-1a-MDCK cells (each 500 mg of protein) was separately subjected to immunoprecipitation with the anti-IQGAP1 pAb or control IgG in the absence (crosslinkerÀ) or presence (crosslinker +) of a crosslinker, DSP. The immunoprecipitates were then subjected to SDS–PAGE (8% polyacrylamide gel), followed by Western blotting with the anti-IQGAP1 mAb, the anti-afadin mAb, and the anti-FLAG mAb. Input; total extract, a-IQGAP1 Ab IP, immunoprecipitates of the anti-IQGAP1 pAb; and control IgG IP; immunoprecipitates of control IgG. The results shown are representative of three independent experiments

To examine whether the microtuble structures are cell–cell adhesion sites where the signals for E-cadherin, involved in the localization of IQGAP1 at the cell–cell a-catenin, and b-catenin were undetectable. Second, the adhesion sites, nectin-1a-MDCK cells were treated with signal for IQGAP1 behaved similarly to that of nectin- a microtubule-disrupting agent, nocodazole. When these 1a during the formation of AJs by increase of Ca2+ from cells were incubated with nocodazole for 45 min, the a low to a high concentration. Third, the localization of immunoﬂuorescence signal for a Golgi marker, wheat IQGAP1 at the cell–cell adhesion sites was inhibited by germ agglutinin, disappeared (Figure 8, nocodazole). nectin inhibitors, which inhibited the trans-interaction Under these conditions, the belt-like staining of nectin- of nectin. Fourth, when nectin-1a-MDCK cells were 1a or IQGAP1 was not markedly changed (Figure 8, incubated with the microbeads coated with the extra- nocodazole). These results suggest that the microtubule cellular fragment of nectin, which interacts with cellular structures are not required for the localization of nectin, IQGAP1 accumulated at the bead–cell contact IQGAP1 at the cell–cell adhesion sites. sites. Finally, immunoprecipitation analysis of IQGAP1 revealed that it was associated with nectin and afadin. Discussion Taken together, our results suggest that IQGAP1 localizes at AJs in a manner dependent on nectin in IQGAP1, a putative downstream target of the Rho epithelial cells that does not absolutely require the E- family small G proteins, Cdc42 and Rac, localizes at AJs cadherin–catenin system, although we cannot comple- in epithelial cells (Kuroda et al., 1998; Fukata et al., tely neglect the possibility that the cadherin–catenin 1999). It has been suggested that IQGAP1 localizes at system has also a role in the localization of IQGAP1 AJs through its binding to b-catenin (Kuroda et al., at AJs. 1998). It has also been suggested that E-cadherin is It has been suggested that IQGAP1 negatively required for the accumulation of IQGAP1 at AJs (Li regulates the E-cadherin-mediated cell–cell adhesion et al., 1999). In this paper, we have investigated a role of through dissociation of a-catenin from the E-cadherin– nectin in the localization of IQGAP1 at AJs in epithelial b-catenin–a-catenin complex (Kuroda et al., 1998; cells. We have found several lines of evidence that Fukata et al., 2001). However, the E-cadherin-mediated IQGAP1 localizes at AJs in a nectin-dependent manner cell–cell adhesion sites, where a-catenin was present, in epithelial cells. First, during disruption of the E- were clearly formed during the formation of AJs in cadherin-mediated cell–cell adhesion by Ca2+ chelation nectin-1a-MDCK cells, although IQGAP1 also accu- in nectin-1a-MDCK cells, the immunoﬂuorescence mulated at the E-cadherin-based cell–cell adhesion sites. signal for IQGAP1 remained on the free surface of the If IQGAP1, the negative regulator of E-cadherin- plasma membrane and at the residual nectin-1a-based mediated cell–cell adhesion, accumulated at the cell–cell

Oncogene oeo etni h oaiaino IQGAPI of localization the Katata in T nectin of role A tal et

Figure 7 Inhibition of the localization of IQGAP1 at the cell–cell adhesion sites by an F-actin-disrupting agent, latrunculin A. (a) Nectin-1a-MDCK cells cultured at 2 mm Ca2+ with or without latrunculin A. The cells were cultured at 2 mm Ca2+ in the absence (control) or presence (LatA) of 20 nm latrunculin A for 30 min. The cells were triple stained with rhodamine– phalloidin, the anti-IQGAP1 pAb, and the anti-FLAG mAb.bars, 10 mm. (b) Nectin-1a-MDCK cells cultured at 2 mm Ca2+ with or without latrunculin A. The cells were cultured at 2 mm Ca2+ for 120 min (low Ca2+ ); the cells were cultured at 2 mm Ca2+ in the presence of 20 nm latrunculin A for 30 min (low Ca2+ LatA); and the cells precultured at 2 mm Ca2+ were then cultured at 2 mm Ca2+ in the presence of 20 nm latrunculin A for 120 min (normal Ca2+ LatA). The cells were triple stained with rhodamine–phalloidin, the anti-IQGAP1 pAb, and the anti-FLAG mAb. bars, 10 mm. The results are representative of three independent experiments Oncogene 2105 A role of nectin in the localization of IQGAPI T Katata et al 2106 a-catenin complex is present, IQGAP1 may have a positive role on cell–cell adhesion. A previous study indicates that the immunofluorescence signal for IQGAP1 disappears from the free surface of the plasma membrane and the cell–cell adhesion sites with that of E-cadherin during the disruption of the E-cadherin-mediated cell–cell adhesion by Ca2+ chelation in MDCK cells (Nakagawa et al., 2001). This is inconsistent with our present observation that the signal for IQGAP1 remained on the free surface of the plasma membrane and at the residual nectin- 1a-based cell–cell adhesion sites where the signals for E- cadherin, a-catenin, and b-catenin were undetectable during the disruption of the E-cadherin-mediated cell– cell adhesion by Ca2+ chelation. The signal for IQGAP1 still remained on the free surface of the plasma membrane and at the residual cell–cell adhesion sites where E-cadherin was undetectable in wild-type MDCK cells, which did not express exogenous nectin-1a, when these cells were cultured at 2 mm Ca2+ for 120 min. Thus, the difference is not simply because of the exogenous expression of nectin-1a in nectin-1a-MDCK cells. Since our results were confirmed by the stainings with two different Abs, it is evident that IQGAP1 remains on the free surface of the plasma membrane and at the residual nectin-based cell–cell adhesion sites where E-cadherin is undetectable. We have shown here that the accumulation of IQGAP1 at the cell–cell adhesion sites was inhibited by F-actin-disrupting agents, latrunculin A and cytochalasin D, but not by a microtubule-disrupting agent, nocodazole. Thus, IQGAP1 localizes at AJs in a manner dependent on nectin and F-actin. How does IQGAP1 localize at the cell–cell adhesion sites in a manner dependent on nectin and F-actin in epithelial cells? Recently, we have shown that the heterotypic trans- Figure 8 No effect of a microtubule-disrupting agent, nocodazole, interaction between nectin-2 in Sertoli cells and nectin-3 on the localization of IQGAP1 at the cell–cell adhesion sites. Nectin-1a-MDCK cells were cultured in the absence (control) or in spermatids is formed at Sertoli–spermatid junctions, presence (nocodazol) of 33 mm nocodazole. The cells were triple heterotypic AJs in the testis, and that each nectin-based stained with the wheat germ agglutinin conjugates (WGA)’ the adhesive membrane domain exhibits one-to-one coloca- anti-IQGAP1 pAb, and the anti-FLAG mAb. bars, 10 mm. The lization with each actin bundle underlying Sertoli– results are representative of three independent experiments spermatid junctions (Ozaki-Kuroda et al., 2002). Afadin also colocalizes with nectin at Sertoli–spermatid junctions. Thus, IQGAP1 may recognize the specific actin adhesion sites solely by binding to b-catenin, b-catenin structures organized by the nectin–afadin system and would be occupied by IQGAP1 and the resulting E- localize at the cell–cell adhesion sites where nectin and cadherin-mediated cell–cell adhesion would lack a- afadin are present. The accumulation of IQGAP1 at the catenin and form weak adhesion. In addition, it has Nef-3-bead cells contact sites in nectin-1a-MDCK cells been suggested that the formation of E-cadherin- suggests that the trans-interaction of nectin-1a is mediated cell–cell adhesion activates Rac and the necessary and sufficient for the accumulation of activated Rac inhibits the binding of IQGAP1 to b- IQGAP1 in nectin-1a-MDCK cells. We have recently catenin (Fukata et al., 1999; Nakagawa et al., 2001). If found that the trans-interaction of nectin-1a induces the IQGAP1 localized at AJs solely by binding to b-catenin, formation of filopodia and lamellipodia through the IQGAP1 would not localize at the E-cadherin-based respective activation of Cdc42 and Rac (Kawakatsu cell–cell adhesion sites where Rac is activated. Taken et al., 2002). Thus, IQGAP1 may recognize activated together, it is less likely that IQGAP1 localizes at AJs Cdc42 and Rac at the nectin-based cell–cell adhesion solely by binding to b-catenin and that IQGAP1 sites and localize there. It is of crucial importance to negatively regulates the E-cadherin-mediated cell–cell clarify the actin structure organized by the nectin–afadin adhesion through dissociation of a-catenin from the E- system to understand how IQGAP1 localizes. In cadherin–b-catenin–a-catenin complex. Since IQGAP1 addition, more detailed analyses of the ordered assembly localizes at AJs where the E-cadherin–b-catenin– of nectin, afadin, F-actin, Cdc42, Rac, and IQGAP1

Oncogene A role of nectin in the localization of IQGAPI T Katata et al 2107 should be also required. A function of the nectin- agglutinin, and rhodamine-phalloidin. The samples were then mediated localization of IQGAP1 at cell–cell adhesion washed with PBS for 5 min three times and mounted in sites also remains to be elucidated. As recent studies Immuno-fluore mounting medium (ICN Biomedicals, Inc.). have suggested that IQGAP1 functions as a molecular The samples were analysed by Radiance 2000 confocal laser link between Cdc42 and Rac and the actin cytoskeleton scanning microscope (Bio-Rad Laboratories). (Bashour et al., 1997; Erickson et al., 1997; Swart- 2+ Mataraza et al., 2002), one possibility is that IQGAP1 Ca switch assay functions in the formation of the specific actin structures Ca2+ switch experiments using nectin-1a-MDCK cells were together with the nectin–afadin system and Cdc42 by done as described (Kartenbeck et al., 1991). Briefly, the cells organizing F-actin to form the specified morphology at (1 Â 105) were seeded on an 18-mm glass coverslip in 12-well the cell–cell adhesion sites in epithelial cells. It should be culture dishes. After 48 h, the cells were washed with PBS and 2+ elucidated how the nectin–afadin system, Cdc42, Rac cultured at 2 mm Ca in DMEM without serum for 60 min. m 2+ m and IQGAP1 function in the formation and disruption The cells were then cultured at 2 m Ca (DMEM with 5 m EGTA) in the presence or absence of the mixture of 60 mg/ml of the junctional complex and the actin structures at the gD and 60 mg/ml Nef-3 for 120 min. After the culture, the cells cell–cell adhesion sites. were washed with PBS and cultured at 2 mm Ca2+ in DMEM without serum in the presence or absence of the mixture of 60 mg/ml gD and 60 mg/ml Nef-3 for the indicated periods of Materials and methods time.

Materials and chemicals Preparation of Nef-3-coated beads A rabbit anti-nectin-1a pAb was prepared as described Nef-3-coated beads were prepared as described (Honda et al., (Takahashi et al., 1999). A mouse anti-afadin mAb was 2003). Latex-sulfate microspheres (3 Â 108, 5.2-mm diameter; prepared as described (Sakisaka et al., 1999). A rat anti-E- Interfacial Dynamics Corporation, Portland, OR, USA) were cadherin mAb (ECCD-2) was supplied by Dr M Takeichi washed, resuspended in 0.2 ml of 0.1 m borate buffer, pH 8.0, (Center for Developmental Biology, RIKEN, Kobe, Japan). A and incubated with 100 mg of a goat anti-human IgG (Fc- mouse anti-FLAG mAb and a rabbit anti-b-catenin pAb were speciﬁc) Ab (Sigma) with gentle mixing at room temperature from Sigma. A rabbit anti-IQGAP1 pAb was from Santa Cruz for 18 h. The beads were then centrifuged at 16 000 g at 41C for Biotechnology, Inc. A mouse anti-IQGAP1 mAb was from 10 min and washed with 1 ml PBS three times. The beads were Transduction Laboratories. Texas Red-X-conjugated wheat then incubated with 0.2 ml PBS containing 5 mg/ml BSA germ agglutinin and rhodamine phalloidin were from Mole- (BSA/PBS) at room temperature for 180 min. Aliquots of cular Probes. Secondary Abs for immunoﬂuorescence micro- 0.2 ml of the bead suspension (6 Â 107 microspheres) were then scopy were obtained from Chemicon International, Inc. gD, incubated with 30 mg of Nef-3 or human IgG at room the chimeric protein of a fragment of glycoprotein D (1– temperature for 180 min. After the incubation, the beads were 285 aa) fused to IgG Fc, and Nef-3, the chimeric protein of an washed with 1 ml of BSA/PBS three times and resuspended in extracellular fragment of nectin-3 (56–400 aa) fused to IgG Fc, 0.2 ml of BSA/PBS. were prepared as described (Tachibana et al., 2000; Honda et al., 2003). Bead–cell adhesion assay

Cell culture and DNA transfection Bead-cell adhesion assay was done as described (Honda et al., 2003). Nectin-1a-MDCK cells (3 Â 104) were seeded on a 14- MDCK cells were kindly supplied by Dr W Birchmeier (Max- mm glass coverslip in 24-well culture dishes. After 48 h, the Delbruck-Center for Molecular Medicine, Berlin, Germany). cells were washed with PBS and cultured at 2 mm Ca2+ in An MDCK cell line stably expressing FLAG-nectin-1a (nectin- DMEM without serum for 60 min. The cells were then cultured 1a-MDCK cells) was prepared as described (Takahashi et al., at 2 mm Ca2+ (DMEM with 5 mm EGTA) with latex-sulfate 1999). Briefly, MDCK cells were transfected with pCAGI- microspheres (3 Â 104) coated with Nef-3 or human IgG for puro-FLAG-nectin-1a using the LipofectAMINE reagent 120 min. After the culture, the cells were washed with DMEM (Gibco BRL). The cells were then cultured for 24 h, replated, and cultured at 2 mm Ca2+ in DMEM with 10% FCS at 371C and selected by being cultured in the presence of 5 mg/ml for 180 min. puromycin (Sigma). Crosslinking experiment Immunofluorescence microscopy A crosslinking experiment was done as described (Kuroda Immunofluorescence microscopy was done as described et al., 1998). Nectin-1a-MDCK cells were suspended in PBS in (Mandai et al., 1997; Takaishi et al., 1997). Briefly, the cells the absence or presence of 0.75 mm DSP, and incubated at were fixed in the mixture of 50% acetone and 50% methanol at room temperature for 20 min. The DSP activity was quenched À201C for 1 min. The fixed cells were then washed three times by the addition of 50 mm glycine in PBS. The cells were then with phosphate-buffered saline (PBS). After being blocked in sonicated in an ice-cold lysis buffer (25 mm Tris/HCl, pH 7.5, PBS containing 1% bovine serum albumin (BSA) for 60 min, 1mm CaCl2,1mm MgCl2, 100 mm NaCl, 1% Triton X-100, the cells were incubated in the same buffer with various 2 mg/ml aprotinin, 10 mg/ml leupeptin, and 100 mg/ml PMSF), combinations of the anti-E-cadherin, anti-IQGAP1, anti- followed by centrifugation at 47 000 g for 15 min. The super- FLAG, anti-a-catenin, and anti-b-catenin mAbs, and the natant (4.5 mg of protein) was incubated with the anti- anti-IQGAP1 and anti-nectin-1a pAbs for 60 min. The samples IQGAP1 pAb bound to protein G-Sepharose beads (20 mgof were washed with PBS for 5 min three times and incubated for Ab/20 ml of beads) (Amersham Pharmacia Biotech) at 41C 30 min in PBS containing 1% BSA with various combinations overnight. After the beads were extensively washed with the of the secondary pAbs, Texas Red-X-conjugated wheat germ lysis buffer, the bound proteins were eluted by boiling the

Oncogene A role of nectin in the localization of IQGAPI T Katata et al 2108 beads in an SDS sample buffer (60 mm Tris/HCl, pH 6.7, 3% Other procedures SDS, 2% 2-mercaptoethanol, and 5% glycerol), and subjected Protein concentrations were determined with BSA as a to SDS–polyacrylamide gel electrophoresis (PAGE) (8% reference protein (Bradford, 1976). SDS–PAGE was done as polyacrylamide gel), followed by Western blotting. described (Laemmli, 1970). Disassembly of the actin cytoskeleton Nectin-1a-MDCK cells (1 Â 105) were seeded on an 18-mm glass coverslip in 12-well culture dishes. After 48 h, the cells Abbreviations were washed with PBS and cultured at 2 mm Ca2+ in DMEM aa, amino acid(s); Ab, antibody; AJs, adherens junctions; without serum for 60 min. Nectin-1a-MDCK cells were then BSA, bovine serum albumin; DSP, dithiobis succinimidylpro- cultured at 2 mm Ca2+ in DMEM with 20 nm latrunculin A for pionate; gD, fragment of glycoprotein D fused to IgG Fc; 30 min. On the other hand, after being cultured at 2 mm Ca2+ mAb, monoclonal antibody; MDCK, Madin–Darby canine in DMEM without serum for 60 min, the cells were then kidney; Nef-3, extracellular fragment of nectin-3 fused to IgG cultured at 2 mm Ca2+ (DMEM with 5 mm EGTA) for 120 min. Fc; pAb, polyclonal antibody; PAGE, polyacrylamide gel After the culture, nectin-1a-MDCK cells were cultured at 2 mm electrophoresis; PBS, phosphate-buffered saline. Ca2+ in DMEM with 5 mm EGTA and 20 nm latrunculin A for 30 min. Nectin-1a-MDCK cells were then cultured at 2 mm Ca2+ in DMEM with 20 nm latrunculin A for 120 min. For the treatment with cytochalasin D, nectin-1a-MDCK cells Acknowledgements were incubated with 4 mm cytochalasin D for 60 min instead of We thank Dr A Nagafuchi (Kumamoto University, Kuma- latrunculin A. moto, Japan) for valuable discussions and critical readings of the manuscript, Dr M Takeichi (Center for Developmental Biology, RIKEN, Kobe, Japan) for providing us with the anti- Disassembly of the microtubule cytoskeleton E-cadherin mAb, and Dr W Birchmeier (Max-Delbruck- Nectin-1a-MDCK cells (1 Â 105) were seeded on an 18-mm Center for Molecular Medicine, Berlin, Germany) for provid- glass coverslip in 12-well culture dishes. After 48 h, the cells ing us with MDCK cells. This investigation was supported by were washed with PBS and cultured at 2 mm Ca2+ in DMEM grants-in-aid for Scientiﬁc Research and for Cancer Research without serum for 60 min. The cells were then cultured in from the Ministry of Education, Science, Sports, Culture, and DMEM with 33 mm nocodazole for 45 min. Technology, Japan (2001, 2002).

References

Aoki J, Koike S, Ise I, Sato-Yoshida Y and Nomoto A. (1994). Geiger B and Ginsberg D. (1991). Cell Motil. Cytoskeleton, 20, J. Biol. Chem., 269, 8431–8438. 1–6. Asakura T, Nakanishi H, Sakisaka T, Takahashi K, McCal- Geraghty RJ, Krummenacher C, Cohen GH, Eisenberg RJ lum K, Nishimura M, Sasaki T and Takai Y. (1999). Genes and Spear PG. (1998) Science, 280, 1618–1620. Cells, 4, 573–581. Gumbiner B, Stevenson B and Grimaldi A. (1988). J. Cell Bashour AM, Fullerton AT, Hartv MJ and Bloom GS. (1997). Biol., 107, 1575–1587. J. Cell Biol., 137, 1555–1566. Hart MJ, Callow MG, Souza B and Polakis P. (1996). EMBO Bradford MM. (1976). Anal. Biochem., 72, 248–254. J., 15, 2997–3005. Braga VM, Machesky LM, Hall A and Hotchin NA. (1997). J. Honda T, Shimizu K, Kawakatsu T, Yasumi M, Shingai T, Cell Biol., 137, 1421–1431. Fukuhara A, Ozaki-Kuroda K, Irie K, Nakanishi H and Campadelli-Fiume G, Cocchi F, Menotti L and Lopez M. Takai Y. (2003). Genes Cells (in press). (2000). Rev. Med. Virol., 10, 305–319. Ikeda W, Nakanishi H, Miyoshi J, Mandai K, Ishizaki H, Cocchi F, Lopez M, Menotti L, Aoubala M, Dubreuil P and Tanaka M, Togawa A, Takahashi K, Nishioka H, Yoshida Campadelli-Fiume G. (1998). Proc. Natl. Acad. Sci. USA, H, Mizoguchi A, Nishikawa S and Takai Y. (1999). J. Cell 95, 15700–15705. Biol., 146, 1117–1132. Cocchi F, Menotti L, Dubreuil P, Lopez M and Campadelli- Imamura Y, Itoh M, Maeno Y, Tsukita S and Nagafuchi A. Fiume G. (2000). J. Virol., 74, 3909–3917. (1999). J. Cell Biol., 144, 1311–1322. Eberle´ F, Dubreuil P, Mattei MG, Devilard E and Lopez M. Kartenbeck J, Schmelz M, Franke WW and Geiger B. (1991). (1995). Gene, 159, 267–272. J. Cell Biol., 113, 881–892. Erickson JW, Cerione RA, Hart MJ. (1997). J. Biol. Chem., Kawakatsu T, Shimizu K, Honda T, Fukuhara T, Hoshino T 272, 24443–24447. and Takai Y. (2002) J. Biol. Chem., 277, 50749–50755. Fukata M, Nakagawa M, Itoh N, Kawajiri A, Yamaga M, Knudsen KA, Soler AP, Johnson KR and Wheelock MJ. Kuroda S and Kaibuchi K. (2001). Mol. Cell. Biol., 21, (1995). J. Cell Biol., 130, 67–77. 2165–2183. Kuroda S, Fukata M, Kobayashi K, Nakafuku M, Nomura Fukata M, Nakagawa M, Kuroda S and Kaibuchi K. (1999). N, Iwamatsu A and Kaibuchi K. (1996). J. Biol. Chem., 271, J. Cell Sci., 112, 4491–4500. 23363–23367. Fukuhara A, Irie K, Nakanishi H, Takekuni K, Kawakatu T, Kuroda S, Fukata M, Nakagawa M, Fujii K, Nakamura T, Ikeda W, Yamada A, Katata T, Honda T, Sato T, Shimizu Ookubo T, Izawa I, Nagase T, Nomura N, Tani H, Shoji I, K, Ozaki H, Horiuchi H, Kita T and Takai Y. (2002a) Matsuura Y, Yonehara S and Kaibuchi K. (1998). Science, Oncogene, 21, 7642–7655. 281, 832–835. Fukuhara A, Irie K, Yamada A, Katata T, Honda T, Shimizu Laemmli UK. (1970). Nature, 227, 680–685. K, Nakanishi H and Takai Y. (2002b) Genes Cells, 7, 1059– Li Z, Kim SH, Higgins JM, Brenner MB and Sacks DB. 1072. (1999). J. Biol. Chem., 274, 37885–37892.

Oncogene A role of nectin in the localization of IQGAPI T Katata et al 2109 Lopez M, Aoubala M, Jordier F, Isnardon D, Gomez S and Rimm DL, Koslov ER, Kebriaei P, Cianci CD and Dubreuil P. (1998). Blood, 92, 4602–4611. Morrow JS. (1995). Proc. Natl. Acad. Sci. USA, 92, 8813– Lopez M, Cocchi F, Menotti L, Avitabile E, Dubreuil P and 8817. Campadelli-Fiume G. (2000). J. Virol., 74, 1267–1274. Sakisaka T, Nakanishi H, Takahashi K, Mandai K., Miyahara Lopez M, Eberle F, Mattei MG, Gabert J, Birg F, Bardin F, M, Satoh A, Takaishi K and Takai Y. (1999). Oncogene, 18, Maroc C and Dubreuil P. (1995). Gene, 155, 261–265. 1609–1617. Machesky LM. (1998). Curr. Biol., 8, R202–R205. Sakisaka T, Taniguchi T, Nakanishi H, Takahashi K, Mandai K, Nakanishi H, Satoh A, Obaishi H, Wada M, Miyahara M, Ikeda W, Yokoyama S, Peng YF, Yamanishi Nishioka H, Itoh M, Mizoguchi A, Aoki T, Fujimoto T, K and Takai Y. (2001). J. Virol., 75, 4734–4743. Matsud Y, Tsukita S and Takai Y. (1997) J. Cell Biol., 139, Satoh-Horikawa K, Nakanishi H, Takahashi K, Miyahara M, 517–528. Nishimura M, Tachibana K, Mizoguchi A and Takai Y. McCallum SJ, Wu WJ and Cerione RA. (1996). J. Biol. Chem., (2000). J. Biol. Chem., 275, 10291–10299. 271, 21732–21737. Spear PG, Eisenberg RJ and Cohen GH. (2000). Virology, 275, Miyahara M, Nakanishi H, Takahashi K, Satoh-Horikawa K, 1–8. Tachibana K and Takai Y. (2000). J. Biol. Chem., 275, 613– Swart-Mataraza JM, Li Z and Sacks DB. (2002). J. Biol. 618. Chem., 277, 24753–24763. Mizoguchi A, Nakanishi H, Kimura K, Matsubara K, Ozaki- Tachibana K, Nakanishi H, Mandai K, Ozaki K, Ikeda W, Kuroda K, Katata T, Honda T, Kiyohara Y, Heo K, Yamamoto Y, Nagafuchi A, Tsukita S and Takai Y. (2000). Higashi M, Tsutsumi T, Sonoda S, Ide C and Takai Y. J. Cell Biol., 150, 1161–1176. (2002). J. Cell Biol., 156, 555–565. Takahashi K, Nakanishi H, Miyahara M, Mandai K, Momose Y, Honda T, Inagaki M, Shimizu K, Irie K, Satoh K, Satoh A, Nishioka H, Aoki J, Nomoto A, Nakanishi H and Takai Y. (2002). Biochem. Biophys. Res. Mizoguchi A and Takai Y. (1999). J. Cell Biol., 145, Commun., 293, 45–49. 539–549. Morrison ME and Racaniello VR. (1992) J. Virol., 66, 2807– Takai Y and Nakanishi H. (2003). J. Cell Sci., 116, 17–27. 2813. Takaishi K, Sasaki T, Kotani H, Nishioka H and Takai Y Nagafuchi A, Takeichi M and Tsukita S. (1991). Cell, 65, 849– (1997). J. Cell Biol., 139, 1047–1059. 857. Takeichi M. (1988). Development, 102, 639–655. Nakagawa M, Fukata M, Yamaga M, Itoh N and Kaibuchi Takeichi M. (1991). Science, 251, 1451–1455. K. (2001). J. Cell Sci., 114, 1829–1838. Takeichi M. (1995). Curr. Opin. Cell Biol., 7, 619–627. Nishioka H, Mizoguchi A, Nakanishi H, Mandai K, Warner MS, Geraghty RJ, Martinez WM, Montgomery RI, Takahashi K, Kimura K, Satoh-Moriya A and Takai Y. Whitbeck JC, Xu R, Eisenberg RJ, Cohen GH and Spear (2000). J. Comp. Neurol., 424, 297–306. PG. (1998). Virology, 246, 179–189. Ozaki-Kuroda K, Nakanishi H, Ohta H, Tanaka H, Kurihara Watabe-Uchida M, Uchida N, Imamura Y, Nagafuchi A, H, Mueller S, Irie K, Ikeda W, Sakai T, Wimmer E, Fujimoto K, Uemura T, Vermeulen S, van Roy F, Adamson Nishimune Y and Takai Y. (2002) Curr. Biol., 12, 1145– ED and Takeichi M. (1998). J. Cell Biol., 142, 847–857. 1150. Weiss EE, Kroemker M, Rudiger AH, Jockusch BM and Ozawa M, Baribault H and Kemler R. (1989) EMBO J., 8, Rudiger M. (1998). J. Cell Biol., 141, 755–764. 1711–1717. Weissbach L, Settleman J, Kalady MF, Snijders AJ, Murthy Prasad R, Gu Y, Alder H, Nakamura T, Canaani O, Saito H, AE, Yan YX and Bernards A. (1994). J. Biol. Chem., 269, Huebner K, Gale RP, Nowell PC, Kuriyama K, Miyazaki 20517–20521. Y, Croce CM and Canaani E. (1993). Cancer Res., 53, 5624– Yagi T and Takeichi M. (2000). Genes Dev., 14, 1169–1180. 5628. Yokoyama S, Tachibana K, Nakanishi H, Yamamoto Y, Irie Reymond N, Fabre S, Lecocq E, Adelaide J, Dubreuil P and K, Mandai K, Nagafuchi A, Monden M and Takai Y. Lopez M. (2001). J. Biol. Chem., 276, 43205–43215. (2001). Mol. Biol. Cell, 12, 1595–1609.

Oncogene