伪PIX nucleotide exchange factor is activated by interaction with phosphatidylinositol 3-kinase Abstractp21-activated kinase (PAK) is a common effector protein of the small GTPases Cdc42 and Rac, leading to the activation of downstream mitogen activated protein kinases. PAK also mediates polarized cytoskeletal changes induced by these GTPases. The recently identified PAK-interacting exchange factor (PIX) acts as a guanine nucleotide exchange factor on Rac, and colocalizes with PAK in a focal complex, but little is known about the associated signaling cascades, including upstream activators of PIX. In this study, we show that one of the isoforms of PIX, 伪PIX, is activated by signaling cascades from the platelet-derived growth factor (PDGF) receptor and EphB2 receptor, and from integrin-induced signaling through phosphatidylinositol 3-kinase (PI3-kinase). 伪PIX is activated by forming a complex with these receptors either via association with PAK and Nck, or direct association with the p85 regulatory subunit of PI3-kinase. Synthetic phosphoinositide and membrane targeted PI3-kinase augmented the 伪PIX activity in vivo. In Xenopus, aggregates of mesodermal cells derived from embryos microinjected with 伪PIX significantly increased the peripheral spreading on fibronectin substrate in response to PDGF through PI3-kinase. These results indicate that 伪PIX is activated by PI3-kinase, and is involved in the receptor mediated signaling leading to the activation of the kinase activity of PAK, and the migration of mesodermal cells on extracellular matrix. IntroductionThe p21-activated kinase (PAK) family of serine/threonine kinases is known to be activated by binding to the GTP-bound form of Cdc42 or Rac1, small GTPases of the Rho family, and its activation leads to the activation of mitogen activated protein kinases (MAPK) (Brown et al., 1996; Zhang et al., 1995). PAK also acts as a modulator of polarized cytoskeletal changes. For example, overexpression of 尾PAK induces a Rac phenotype including membrane ruffling and lamellipodia formation, and membrane targeted PAK1 induces neurite outgrowth in PC12 cells (Obermeier et al., 1998; Daniels et al., 1998). The correct conformation of the amino terminus and a proline-rich region of PAK1 is important in the regulation of polarity of the cytoskeleton. A PAK1 construct with a modified amino terminus, which was more accessible to SH3 proteins such as Nck, had dramatic effects on the formation of polarized membrane ruffles and focal complexes, while the same construct with an additional mutation in the proline-rich region to abolish binding ability to SH3 protein showed no effect (Sells et al., 1997). For wild-type PAK1, equivalent conformational changes presumably occur in vivo upon binding to activated Cdc42 and Rac1. PAK likely binds to an SH3 domain-containing protein which interacts with F-actin, and eventually contributes to the regulation of cell morphology in Saccharomyces cerevisiae (Leeuw et al., 1995).Recently, the PAK-interacting exchange factor (PIX) family, which are Rho-p21 guanine nucleotide-exchange factors (GEF), have been identified (Manser et al., 1998). The human PIX family, which consists of two isoforms (伪- and 尾PIX), bind tightly through an N-terminal SH3 domain to a conserved proline-rich PAK sequence located at the C-terminus of p21-binding domain (PBD), and are colocalized with PAK to form activated Cdc42- and Rac1-driven focal complexes (Manser et al., 1998; Oh et al., 1997). Recently, it was demonstrated that the association of the PAK amino terminus with the PIX SH3 domain resulted in activation of PIX GEF function toward Rac, which leads to Rac-type morphological effects (Obermeier et al., 1998).However, little is known about the signaling cascades in which PIX is involved, and how PIX is activated. It has been shown that the adapter protein Nck links receptor tyrosine kinases to downstream signaling pathways in association with tyrosine-phosphorylated receptors via its Src homology 2 (SH2) domain, and proline-rich sites on target proteins including PAK via its SH3 domains (Lu et al., 1997; Galisteo et al., 1996). Therefore, PAK1 may associate with both Nck and PIX, as Nck and PIX binds to different sites among the three proline-rich regions of PAK (Manser et al., 1998). These findings suggest that PIX might be involved in signaling cascades from receptor tyrosine kinases. Together with Nck, PAK may function as a scaffolding protein linking PIX to receptors locating at the cell membrane, which results in activation of PIX. PDGF-stimulation of cells is known to induce active cytoskeletal rearrangement, and endogenous PAK1 redistributes into membrane ruffles and edges of lamellipodia. The findings that such changes induced by PDGF are prevented by the phosphatidylinositol 3-kinase inhibitor wortmannin, may indicate that PI3-kinase has a key role in the regulation of cytoskeletal changes mediated by GTP-ases (Hawkins et al., 1995).We have isolated cDNAs encoding human 伪- and 尾PIX which were later revealed to be identical to KIAA0006 and KIAA0142 in GenBank, respectively, from a gastric cancer cell cDNA library. In this report, we show that 伪PIX is a more effective GEF than 尾PIX toward Rac1 and also Cdc42, and that it subsequently activates PAK kinase under PDGF stimulation. 伪PIX mediates signals initiated from the receptor tyrosine kinases of PDGFR and EphB2. It also transduces the signal from cell adhesion to fibronectin substratum, leading to PAK kinase activation. These effects of 伪PIX were inhibited by coexpression of a mutated p85 regulatory subunit of PI3-kinase, which was expected to abolish its binding to the catalytic subunit of PI3-kinase. Moreover, 伪PIX-induced PAK kinase activation was at least partially inhibited by Nck with an SH2 domain mutation.In Xenopus, PDGF stimulates aggregates of embryonic mesoderm cells to spread on fibronectin substrate, and this is dependent on PI3-kinase activity (Symes and Mercola, 1996). Microinjection of RNA encoding 伪PIX into Xenopus embryos strongly activated PDGF signaling through PI3-kinase and induced spreading of mesodermal cell aggregates peripherally on fibronectin. These results suggest that 伪PIX is activated by recruitment to phosphorylated receptor tyrosine kinases or a focal complex either by the formation of a complex with PAK and Nck, or direct association with p85. We conclude that 伪PIX is a target protein of PI3-kinase associated with these receptors, which can activate PAK kinase activity and induce significant morphological changes.Results and discussion伪PIX mediates signals from receptor tyrosine kinases leading to PAK1 activationWe examined whether PIX mediates signals initiated from receptor tyrosine kinases and integrin signals from cell adhesion to fibronectin substrate. In vivo and in vitro experiments indicate that Nck can mediate the relocalization and subsequent activation of PAK1 kinase (Lu et al., 1997), and can also mediate signals from PDGFR and EphB2, which is a member of Eph family receptor tyrosine kinases whose signal cascade is poorly understood, to MAPK (Stein et al., 1998; Holland et al., 1997; Meisenhelder and Hunter, 1992; Nishimura et al., 1993). Therefore, in our experiment, the kinase activity of PAK was measured under the activation of these receptors in cells coexpressing PAK and Nck with wild-type or various truncated mutants of human 伪PIX (Figure 1a). Without ligand stimulation, PAK1 was not adequately activated by transfection of any form of 伪PIX with or without Nck (data not shown). We performed the experiments under the condition that PAK1 was weakly activated by stimulation of ligands in the absence of PIX (Figure 1b and c, lanes 1 and 2). Wild-type as well as N-terminal deleted 伪PIX (螖CH) enhanced kinase activity of PAK1 in PDGF-stimulated cells (Figure 1b, lanes 3 and 4). In contrast, SH3 domain- or PH domain-deleted 伪PIX showed very little effect to activate PAK1 (Figure 1b, lanes 5 and 6). Coexpression of Nck carrying a mutation in the SH2 domain (R308K), which lacks the ability to bind phosphorylated receptors, inhibited the PAK activation by 伪PIX in PDGFR-stimulated cells (Figure 1b, lane 8). The PAK1 activation by 伪PIX was also disrupted by the coexpression of the p85伪 subunit of phosphatidylinositol 3-kinase (PI3-kinase) with a mutation that blocks binding to the p110 catalytic subunit of PI3-kinase (螖p85) (Figure 1b, lane 9), or treatment of the cells with wortmannin, which inhibits PI3-kinase activity at low concentrations (Yano et al., 1993; Arcaro and Wymann, 1993; Kanai et al., 1993) (Figure 1b, lane 11). These results indicate that 伪PIX function in the signaling pathway from PDGFR to PAK1 activation requires interaction with Nck and PI3-kinase.Figure 1伪PIX facilitates signaling from PDGF receptor and EphB2. (a) Schematic representation of wild-type and truncated PIX cDNA constructs. Proteins are depicted to scale; CH, calponin homology; SH3, Src homology 3; DH, dbl homology; PH, pleckstrin homology; Pro, proline-rich region. (b and c) 293T cells were transfected with plasmids encoding PDGFR尾 (b) or EphB2 (c); Nck, myc-tagged PAK1, wild-type or mutated forms of 伪PIX with or without Nck R308K, 螖p85 as indicated above the lanes. Forty-eight hours later, cells were lysed for immunoprecipitation and an in vitro kinase assay for PAK1 was performed. Before making cell lysates, transfected cells (b) were treated (+), or untreated (鈭? with PDGF BB. Lane 11; transfected cells were incubated for 25鈥塵in with 100鈥塶M wortmannin before being treated with PDGF. The number in the ordinate (b; lanes 1鈥夆€撯€?, c; lanes 1鈥夆€撯€?) indicates the relative PAK1 kinase activity as a fold of the activity present in cells lacking PIX (lane 2). Similar results were obtained in three independent experiments. phospho-MBP, phosphorylated-myeline basic proteinFull size imageTransiently expressed EphB2 in Cos cells is phosphorylated on tyrosine residues, including a residue crucial to the interaction with the SH2 domain of Src (Zisch et al., 1998). In 293T cells, exogenously expressed EphB2 is also highly phosphorylated on tyrosine residues, and activates MAP kinase family proteins, including extracellular signal-stimulated kinase (erk1) and c-Jun NH2-terminal kinase (JNK) (data not shown). Coexpression of 伪PIX also enhanced PAK1 activation in EphB2 receptor-expressing 293T cells (Figure 1c, lanes 3 and 4), which was inhibited by 螖p85, and partially inhibited by Nck R308K (Figure 1c, lanes 7 and 8). These results suggest that 伪PIX is involved in the signaling pathway from EphB2 to activation of PAK1 kinase activity. While this manuscript was in preparation, Daniels et al. (1999) reported truncated form 伪PIX155鈥夆€撯€?45, but not wild-type 伪PIX activated PAK kinase in Cdc42-overexpressed cells. We also detected much higher PAK1 kinase activity by 螖CH 伪PIX than by wild-type 伪PIX in cells cotransfected with PAK and Cdc42 (data not shown). However in PDGF or EphB2-stimulated cells, we could detect activation of PAK also by wild-type 伪PIX. Some conformational changes which lead to the full activation of 伪PIX may explain this observation.Next, we tested the interaction of 伪PIX with Nck and these receptors in vivo by immunoprecipitation. As shown in Figure 2a (lane 2), 伪PIX formed a complex with the PDGF receptor, Nck and PAK1 in PDGF-treated cells. Likewise, 伪PIX associated with activated EphB2 tyrosine kinase receptor under the coexpression of Nck and PAK1 (Figure 2a, lane 5). The association of 伪PIX with these receptors at least partially depends on the expression of Nck and PAK1, because coexpression of PAK carrying a mutation in the first proline-rich region which impaired its binding to the Nck SH3 domain (P13A-PAK1) inhibited the association of 伪PIX with these receptors (Figure 2a, lanes 3 and 6). These results also indicate that 伪PIX does not directly associate with Nck, but forms a complex with Nck and the receptors via PAK. Moreover, as shown in Figure 2b, wild-type p85伪 subunit of PI3-kinase, but not p85伪330鈥夆€撯€?24 (p85(III)) associates with PIX directly in vivo (lanes 2, 3 and 4) and in vitro (lanes 5 and 6). This implies that PIX binds to the N-terminal region upstream of the first SH2 domain of the p85 subunit, while SH3 domain of PIX is not involved in the association with p85. This is not the case with the Rac1-specific GEF, Vav. Tyrosine-phosphorylated Vav forms a complex with the p85 subunit of PI3-kinase via its SH2 domain (Shigematsu et al., 1997). We could also detect the association of p85 subunit with PAK1 in vivo, although its interaction was weak (Figure 2b, lane 1). Although little is known about the proteins that interact with EphB2, the SH2 domain of p85 is reported to associate with autophosphorylated EphB2 in vitro (Holland et al., 1997). We also detected the association of p85 with phosphorylated EphB2 in vivo (Figure 2b, lane 7). However, we cannot currently determine whether the phosphorylated tyrosine residues of EphB2 in this experiment are identical to those stimulated by the physiological interaction with ligand-expressing cells, because the association of Nck and EphB2 was weak in this experiment, comparing to that when cells which stably express EphB2 receptor were cocultured with cells stably expressing EphrinB1, a ligand for EphB2 (data not shown). This observation may explain why Nck R308K only partially inhibited PAK activation by 伪PIX in EphB2-expressing cells (Figure 1c, lane 8).Figure 2伪PIX forms a complex with PDGFR and EphB2 by association with PAK, Nck and the p85 subunit of PI3-kinase. (a) 293T cells were transfected with plasmids encoding PDGFR尾 (lanes 1, 2, 3) or EphB2 (lanes 4, 5 and 6), myc-tagged 伪PIX, wild-type or mutant (P13A) PAK1 and Nck as indicated above the lanes. Forty-eight hours later, cells (lanes 2 and 3) were treated with PDGF BB. Cells were lysed and immunoprecipitated (IP) with anti-myc antibody. Bound proteins were immunoblotted (IB) with antibodies as indicated. A comparison of expression of PDGFR, EphB2 and Nck in individual cell lysates (total lysate) is shown on the side. (b) (Lanes 1, 2, 3 and 4); in vivo association of p85 and 伪PIX. 293T cells were transfected with plasmids encoding HA-tagged wild-type p85 (WT, lanes 1, 2 and 3) or N-terminal truncated p85 (SH3(III), lane 4), myc-tagged PAK1 (lane 1) or myc-tagged wild-type 伪PIX (WT) or SH3 domain-deleted 伪PIX (螖SH3). Cells were lysed and immunoprecipitated with anti-myc antibody. Bound proteins were Western blotted and probed with anti-HA antibody. (Lanes 5 and 6); in vitro association of p85 and 伪PIX was examined as described in Materials and methods. A comparison of expression of wild-type and SH3(III) p85 in total lysates is shown at the bottom. AP, affinity precipitation. (Lanes 7 and 8); EphB2 receptor associates with p85 subunit in vivo. Plasmids encoding EphB2 with or without HA-tagged wild-type p85 as indicated were transfected into 293T cells. Cell lysate was immunoprecipitated with anti-EphB2 antibody, and subsequently immunoblotted with anti-HA antibody. (c) 伪PIX is recruited to the receptors directly by p85 subunit. 293T cells were transfected with plasmids as indicated above lanes. Cell lysates were immunoprecipitated with anti myc epitope tag antibody, and immunoblotted with antibodies for PDGF type B receptor or mouse EphB2Full size imageWe propose that there are two ways by which 伪PIX may be recruited to the activated receptors where PI3-kinase is located. Recruitment of the 伪PIX-PAK-Nck complex to the activated receptors may result in colocalization of this complex with PI3-kinase. Alternatively, 伪PIX may also be recruited to the tyrosine-phosphorylated receptors by the direct association with the p85 subunit of PI3-kinase. Consistent with this model, coexpression of p85 resulted in an increase in those receptors that coprecipitated with 伪PIX even under when the complex formation of 伪PIX with PAK and Nck was inhibited (Figure 2c, lanes 2 and 4). We also examined Rac1 or Cdc42 activation by 伪PIX in vivo by affinity precipitation of GTP-bound Rac or Cdc42 with GST fusion protein of the p21 binding domain of PAK1 (GST-PBD) as described (Bagrodia et al., 1998). Under PDGF-treatment, there was an increase in the amount of activated Rac1 and Cdc42 that precipitated with GST-PBD in 伪PIX-transfected 293T cells compared with that in vector-transfected control cells (Figure 3a). Furthermore, the mutants of Nck (R308K) and p85 (螖p85) inhibited the 伪PIX-induced activation of Cdc42, which was determined by the rapid precipitation of PAK associated with GTP-bound Cdc42 in cell lysates containing high concentrations of Mg2+ as described in the legend (Figure 3b, lanes 4 and 5). The ability of 伪PIX to promote guanine nucleotide exchange toward p21 in vivo was also reproducibly higher than that of 尾PIX (Figure 3c). Further experiments including the comparison of the degree of activation of PAK by 伪-, or 尾PIX and the other GEF, such as Dbl and Vav, might be necessary to evaluate the significance of PIX in this signaling pathway. We have also detected the stimulation of [3H]GDP dissociation from Cdc42 in vitro by 伪PIX in contrast to the decreased ability of 尾PIX to catalyze the GDP dissociation (Figure 3d). 伪PIX, but not 尾PIX also stimulated the GDP dissociation from Rac1, although less efficiently than 伪PIX toward Cdc42 (data not shown). This modest effect of 尾PIX (Cool-1) was also reported elsewhere recently (Bagrodia et al., 1998).Figure 3伪PIX activates Rac1 and Cdc42 by PDGF stimulation. (a) 293T cells were cotransfected with plasmids encoding PDGFR尾, wild-type p85, 螖CH 伪PIX indicated above the lanes and wild-type Cdc42 (lanes 1, 2, 3 and 4) or Rac1 (lanes 5, 6, 7 and 8). 伪PIX(鈭? indicate mock vector pCS2+ DNA was transfected to adjust the total amount of transfected DNA per dish. Cells were treated(+) or untreated(鈭? with PDGF BB, and lysed for affinity-precipitation (AP) with immobilized GST-PBD as described in Materials and methods. Precipitated GTP-bound p21 was detected by immunoblotting with anti-Cdc42 (lanes 1, 2, 3 and 4) or Rac1 (lanes 5, 6, 7 and 8) antibody. (b) 293T cells were transfected with plasmids encoding PDGFR尾, 螖CH, 伪PIX, myc-tagged PAK1, pEBG-Cdc42 and wild-type (WT) or mutated (R308K) Nck indicated above the lanes and treated with PDGF BB (lanes 2鈥夆€撯€?). Cells in lane 5 were also transfected with 螖p85. Cells were lysed in the same buffer as in (a) and rapidly precipitated by incubation with glutathione-sepharose for 30鈥塵in at 4掳C. After washing three times with same lysis buffer, bound proteins were Western blotted and probed with anti-myc antibody. GTP-bound Cdc42 in cell lysates was evaluated by detection of coprecipitated PAK1 protein. Precipitated GST-Cdc42 and the expression level of PAK1 and PDGFR in each total cell lystate were shown at the bottom. (c) 293T cells were transfected with plasmids encoding PDGFR尾, Nck, myc-tagged PAK1, pEBG-Rac1 (lanes 1, 2 and 3) or pEBG-Cdc42 (lanes 4, 5 and 6), and mock vector, wild-type 伪- or 尾PIX. Cells were treated with PDGF BB before lysis and then precipitated with glutathione-sepharose as in (b). Coprecipitated PAK1 protein was Western blotted and detected by anti-myc antibody. (d) Effects of 伪PIX and 尾BIX on [3H]GDP binding to the human Cdc42 protein in vitro. The time course for dissociation of [3H]GDP from GST-Cdc42 was measured in the presence of GST (鈥?, GST-伪PIX153鈥夆€撯€?45 (鈻? or GST-尾PIX11鈥夆€撯€?08 (鈻?/span>)Full size image伪PIX mediates the signal leading to PAK activation by cell adhesion to fibronectin substratumIntegrin-mediated cell adhesion to fibronectin leads to tyrosine phosphorylation of several proteins associated with the focal complex, such as focal adhesion kinase (p125FAK) and p130cas and activates MAP kinase by interacting with c-Src, Grb2 and Nck (Schlaepfer et al., 1997). We examined whether PIX is also involved in the fibronectin-induced signaling cascade of PAK activation through interactions with Nck and/or PI3-kinase. Expression of 伪PIX enhanced PAK1 kinase activity in cells plated on fibronectin-coated dishes (Figure 4a, lanes 3 and 4). Moreover, fibronectin-dependent PAK1 activation by 伪PIX was remarkably reduced by the coexpression of 螖p85 or Nck carrying a mutation in its SH2 domain as well as dominant negative Rac1 or Cdc42 (Figure 4a, lanes 8鈥夆€撯€?1). To confirm that the association with 伪PIX is required for the activation of PAK1, PAK1 with mutation which abolished its binding ability with PIX (PAK1 P191G,R192A) was used (Manser et al., 1998). Compared to the strong activation of wild-type PAK, PAK1 (P191G,R192A) was not activated in the presence of 伪PIX when transfected cells were plated on fibronectin (Figure 4a, lanes 15 and 16). These results indicate an essential role for 伪PIX in PAK1 activation by fibronectin. Tyrosine-phosphorylated p125FAK directly binds to the p85 subunit of PI3-kinase, which subsequently activates PI3-kinase (Chen et al., 1996). Under stimulation of fibronectin, 伪PIX formed a complex with p125FAK and p130cas in cells cotransfected with wild-type Nck and PAK, or the p85 subunit (Figure 4b). However, the weaker association of 伪PIX with p125FAK or p130cas under the coexpression of Nck R308K and PAK1 mutant (P13A) may indicate that 伪PIX does not directly bind to FAK or p130cas (Figure 4b, lanes 1 and 4). As in the case of stimulation by PDGFR or EphB2, association with PAK and Nck, as well as with p85 is considered to be important for this pathway of activation of 伪PIX.Figure 4Involvement of 伪PIX in PAK1 activation induced by cell attachment to fibronectin substratum. (a) Myc-tagged PAK1 (lanes 1鈥夆€撯€?4) or myc-tagged PAK1(P191G, R192A) (lanes 15 and 16; m), Nck together with wild-type or mutated forms of 伪PIX with or without Nck R308K, 螖p85, T17N Rac1, T17N Cdc42 or T17N RhoA were transiently expressed in 293T cells as indicated above the lanes. Forty-eight hours later, cells (lanes 2鈥夆€撯€?6) were plated on fibronectin-coated dishes for 30鈥塵in. Then cells were lysed for immuno-precipitation and in vitro kinase assay of PAK1 as described in Materials and methods. Expression of PAK1 and PIX in individual cell lysates was shown at the bottom. The number in the ordinate (lanes 1鈥夆€撯€?) indicates the relative PAK1 kinase activity as a fold of the activity present in the cells lacking PIX (lane 2). Similar results were obtained in three independent experiments. (b) 293T cells were transfected with plasmids encoding p130cas (lanes 1鈥夆€撯€?) or p125FAK (lanes 4鈥夆€撯€?), myc-tagged 伪PIX, HA-tagged p85, wild-type (WT) or mutated (R308K) Nck and wild-type (WT) or P13A PAK1 as indicated above lanes. Forty-eight hours later, cells were plated on fibronectin-coated dishes, then lysed and immunoprecipitated with anti-myc antibody as described above. Bound proteins were Western blotted and probed with anti-p130cas or anti-p125FAK antibody. Expression of p130cas or p125FAK in individual cell lysates (total lysate) is shown at the bottom. (c) Plasmids encoding myc-tagged PAK1 and wild-type or mutated forms of 伪PIX as indicated were transfected into 293T cells. Cells were treated with cisplatinum (40鈥壩糶/ml) for 3鈥塰, then cell lysates were immunoprecipitated for in vitro kinase assay of PAK1. A comparison of expression of PAK1 in individual cell lysates is shown at the bottom. Similar results were obtained in three independent experimentsFull size imageIn contrast, when the effect of 伪PIX was examined on PAK1 activation by exogenous stress including DNA damage induced by cisplatinum, PIX had no effect (Figure 4c). 伪PIX did not enhance the level of PAK1 kinase activity induced by cisplatinum. This result might be consistent with the finding that we could not detect the effective inhibition of PAK1 kinase activity by coexpression with T17N Rac1 or T17N Cdc42 in cells treated with cisplatinum (data not shown).These results indicate 伪PIX involvement in PAK activation is limited to signaling cascades from receptor tyrosine kinases and cell adhesion to fibronectin. In conclusion, two ways for relocalization of 伪PIX to tyrosine-phosphorylated sites of receptors and proteins associated with the focal complex were identified; complex-formation with Pak and Nck, or direct association with p85.伪PIX is activated by the lipid products of PI3-kinaseThe results that PAK1 activation by 伪PIX was blocked by coexpression of 螖p85 described above demonstrate that the activity of 伪PIX may depend on the interaction with PI3-kinase which is localized at the activated receptors. We next examined whether 伪PIX is activated by the membrane-targeted p110 catalytic subunit of PI3-kinase. 伪PIX was cotransfected into 293T cells with the p110 subunit of PI3-kinase with the addition of a myristoylation signal at amino terminus (myr p110) together with wild-type p85. The membrane-localized p110 (myr p110) is known to be sufficient to trigger PI3-kinase dependent signals including the activation of p70 S6 kinase and Akt (Klippel et al., 1996). Myristoylated p110 activated wild-type and 螖CH 伪PIX, which was detected by the increased amount of GTP-bound Rac1 in vivo (Figure 5a, lanes 4 and 5). In contrast, as expected 伪PIX lacking the PH domain (螖PH) was not activated by the coexpression with myristoylated p110 (Figure 5a, lane 6). Moreover, 尾PIX showed less activity than wild-type 伪PIX to activate Rac1 in response to coexpression with myristoylated p110 (lane 9), although 尾PIX also associated with p85 and was thought to be translocated to the cell membrane by myr p110 in vivo (data not shown).Figure 5伪PIX is activated by PI3-kinase. (a) Membrane localized PI3-kinase activated 伪PIX. 293T cells were transfected with plasmids encoding wild-type Rac1, p85, N-terminal myristoylated p110 subunit of PI3-kinase (myr p110), wild-type or mutants of 伪PIX, 尾PIX, or mock vector as indicated above the lanes. Cells were lysed and affinity-precipitated (AP) with immobilized GST-PBD. Precipitated GTP-bound Rac1 was detected by immunoblot with anti-Rac1 antibody, lanes 1 and 2 (鈭?; control cells without transfection of p110. (b) Lipid products of PI3-kinase activated 伪PIX in vivo. Plasmids encoding wild-type Rac1, p85, myr p110 and 螖CH or 螖PH 伪PIX were transiently transfected into 293T cells. 伪PIX(鈭? indicate mock vector pCS2+ DNA was transfected to adjust the total amount of transfected DNA per dish. Cells were treated(+) or untreated(鈭? with DiC16PI-3,4,5-P3 at a concentration of 10鈥壩?span >M for 20鈥塵in before making cell lysates for affinity-precipitation with immobilized GST-PBD. Precipitated GTP-bound Rac1 was detected by immunoblot with anti-Rac1 antibody. Expression of PIX and Rac1 in each cell lysate were shown at the bottom. DiC16PI-3,4,5-P3 dipalmitoyl phosphatidylinositol-3,4,5-trisphosphate. (c) Measurement of the dissociation of [3H]GDP from GST-Cdc42 in the presence of GST or GST-伪PIX153鈥夆€撯€?45 incubated with or without DiC16PI-3,4,5-P3Full size imageTo confirm that 伪PIX is a target protein of PI3-kinase, we examined whether 伪PIX is also activated by the lipid products of PI3-kinase. As an example, Vav is fully activated by binding of phosphatidylinositol-3,4,5-triphosphate (PI-3,4,5-P3) or another lipid product of PI3-kinase to the PH domain of Vav (Han et al., 1998; Gringhuis et al., 1998). Although 伪PIX may be translocated to the cell membrane and interact with phospholipid products via its PH domain, such direct relocalization of 伪PIX by the PH domain itself may not be the major mechanism of receptor mediated activation of 伪PIX, because coexpression with mutated forms of Nck (Nck R308K) or p85 (螖p85) inhibited 伪PIX activation despite the presence of an intact PH domain. Therefore, activation of 伪PIX by treatment of the cells with phospholipids was examined under conditions in which 伪PIX is recruited to the cell membrane by the coexpression with myr p110 and p85. PI-3,4,5-P3 did not stimulate Rac1 in the absence of PIX (Figure 5b, lanes 1 and 2). As shown in Figure 5b, 螖CH 伪PIX was synergistically activated by the expression of myristoylated p110 and the treatment of cells with PI-3,4,5-P3, resulting in an increase in the GTP-bound form of Rac1. Without expression of myr p110, activation of 伪PIX by the treatment of cells with PI-3,4,5-P3 was weak, and we also detected the synergistic effect of myristoylated p110 and PI-3,4,5-P3 on wild-type 伪PIX (data not shown). In contrast, the activity of 伪PIX lacking the PH domain was not affected by the phospholipid (Figure 5b, lanes 5 and 6). These results indicate 伪PIX is activated by PI3-kinase through binding of the lipid product of PI3-kinase (PI-3,4,5-P3) to the PH domain of 伪PIX. The activation of exchange activity of 伪PIX153鈥夆€撯€?45 by PI-3,4,5-P3 in vitro was also detected by the examination of [3H]GDP dissociation from Cdc42. However, the effect of PI-3,4,5-P3 itself was not so great (Figure 5c). It is possible that other regions of 伪PIX may be required for the full activation of 伪PIX by PI-3,4,5-P3, or other molecules locating at cell membrane may modify the conformation of PIX and activate it. Alternately PH-domain-deleted construct has other negative effects and suppresses background exchange activity.伪PIX activates PDGF-induced spreading of Xenopus mesoderm aggregates on fibronectin substratumWe further analysed whether 伪PIX is involved in PDGF-stimulated signaling to modify cell movement on extracellular matrix by using a system of involuting mesodermal cells in gastrula-stage Xenopus embryos. Migration of mesoderm cells over the fibronectin-rich inferior surface of the blastocoel is known to be important for gastrulation and subsequent anterior-posterior neural patterning of the embryo (Keller and Jansa, 1992). PDGF-A is a candidate signaling molecule involved in gastrulation (Jones et al., 1993) During Xenopus gastrulation, PDGF receptor-伪 is expressed in involuting marginal zone cells which migrate over ectodermal cells expressing PDGF-A (Ataliotis et al., 1995). To examine the ability of 伪PIX to modify PDGF-induced mesodermal cell movement, we used an in vitro assay in which mesoderm cell aggregates spread on fibronectin when exposed to PDGF-A, but form dense clumps in the absence of PDGF-A (Symes and Mercola, 1996).We microinjected mRNA encoding 伪PIX together with mRNA encoding human PAK1 into marginal zones of the two dorsal blastomeres in four-cell stage embryos. Dorsal marginal zone (DMZ) aggregates, which also contain some endodermal cells, from embryos coinjected with 伪PIX and PAK1 formed filopodia and spread peripherally on fibronectin substratum, while control embryos microinjected with wild-type PAK1 alone showed less spreading of DMZ aggregates (Figure 6a and b). Microinjection of 伪PIX mRNA alone also stimulated the peripheral spreading compared to the control, but was always less effective than coinjection with PAK1 mRNA (data not shown). Without PDGF-stimulation, aggregates expressing both 伪PIX and PAK1 did not show any peripheral spreading. The ability of 伪PIX to stimulate PDGF-dependent peripheral spreading of mesodermal aggregates on fibronectin was independent from kinase activation of PAK, as coinjection of kinase-activated PAK1 plus 伪PIX also induced aggregate spreading to almost the same degree as with coinjection of wild-type PAK1 (Figure 6c). This observation is consistent with the previous report that a PAK1 mutant which does not have an active catalytic domain also showed dramatic effects on the polarity of the cytoskeleton in fibroblast cells (Sells et al., 1997; Obermeier et al., 1998). Furthermore, microinjection of PAK1 with mutation in its p21-binding domain (PAK1 H83L,H86L), which lacks the binding ability to GTP-p21, also showed the similar effect with wild-type PAK on the spreading of mesoderm aggregates (data not shown). From this result, we can rule out the possibility that association of PAK with GTP-p21 protected Rac1/Cdc42 from GAPs, thus induced mesoderm aggregates spreading. In contrast, coinjection of dominant-negative mutants of Rac1, or pretreatment of dorsal marginal zone cells with wortmannin inhibited the spread of aggregates on fibronectin under PDGF treatment (Figure 6d and e). These results suggest that 伪PIX activation by PDGF-treatment also occurs in the mesodermal cells of DMZ presumably via PI3-kinase. In previous reports, PI3-kinase may also act as an effector protein of activated GTPase to spread mesodermal aggregates on fibronectin, as GTP-Rac and GTP-Cdc42 bind to p85 and are proposed to activate PI3-kinase directly (Zheng et al., 1994; Tolias et al., 1995), and Cdc42 and Rac1 induce integrin-mediated cell motility and invasiveness through PI3-kinase (Keely et al., 1997). Therefore, it is likely that 伪PIX and PI3-kinase constitute a positive feedback. When interpreting our observations, we cannot exclude the possibility that 伪PIX may have changed the dorso-ventral fate of mesodermal cells resulting in alteration of the expression level of PDGF receptor or integrin. Although spreading of mesoderm aggregates on fibronectin by PDGF-treatment was reported to depend on the activity of PI3-kinase, its significance in the signaling cascade is not known. Further experiments will be necessary including analysis of the morphological changes (i.e. membrane ruffling) of dispersed-single cells of 伪PIX-injected marginal zone mesoderm to determine whether the effect of 伪PIX on the spread of mesoderm aggregates described above is completely relevant to the actin-organization effect of Rac1.Figure 6伪PIX stimulated peripheral spreading of mesoderm aggregates on fibronectin. Aggregates of dorsal marginal zone cells from control embryos microinjected with mRNA of wild-type PAK1 (0.5鈥塶g per embryo) (a), mRNA of 伪PIX and PAK1 (0.5鈥塶g per embryo for each) (b and e), 伪PIX and kinase-inactive mutant (K299R) PAK1 (c), 伪PIX, PAK1 and T17N Rac1 (0.5鈥塶g per embryo for each) (d), were plated on fibronectin-coated dishes in 0.6脳MMR containing PDGFF AA (a, b, c and d) or PDGFF AA with wortmannin (e) as described in Materials and methods. Photographs were taken when sibling embryos reached stage 12. (f) High magnification of (b), taken by phase-contrast microscope. Bar represents 100鈥壩糾. (g) Expression of injected PAK and PIX were indicated by immunoblot. Four explants for each incubated in fibronectin-uncoated plates were corrected and lysed at stage 12 for Western blotting with anti-myc antibodyFull size imageBased on our experiments, 伪PIX is likely to be a crucial regulator of the actin cytoskeleton involved in cell motility and invasiveness induced by the tyrosine kinase receptors such as PDGFR. We have demonstrated that 伪PIX is also a target protein of PI3-kinase associated with the receptors. Therefore, there are at least two ways for recruitment of 伪PIX to the receptors; direct association with p85, or association with PAK and Nck (Figure 7). In a recent report, a PAK-PIX interaction led to PIX activation, and PAK was suggested to act upstream of Rac-mediated morphological changes through interaction with PIX (Obermeier et al., 1998). Our finding that association of PIX with PAK and Nck resulted in relocation of PIX to receptors at the cell membrane might support this model. Activated p21 associates with PAK, and may change N-terminal conformation of PAK to a more accessible shape for 伪PIX binding, which would form a positive feedback for these molecules. Daniels et al. (1999) reported recently that PAK1 mutant that does not bind Rac or Cdc42 is activated in the absence of GTPase by truncated form of 伪PIX and proposed an exchange factor-independent mechanism of PAK activation by PIX. There is also a possibility that binding of PAK to PIX directly changes the conformation, and results in activation of PIX and PAK each other.Figure 7Integrated models for the PDGFR, EphB2 and integrin-induced signaling cascades, which are mediated by 伪PIX. (a) Upon engagement of the receptors by ligands, the tyrosine residues in the cytoplasmic domain of the receptors are autophosphorylated. The SH2 domains of the p85 subunit of PI3-kinase bind to phosphotyrosine residues of the receptors. 伪PIX associates with p85, which serves to recruit 伪PIX to the complex. 伪PIX is also recruited to the complex by the subsequent association with PAK and Nck. Upon ligation of the receptors, the kinase activity of the p110 subunit of PI3-kinase is activated, which generates phosphoinositides including PI-3,4,5-P3. (b) Upon binding of phosphoinositide (PI) to the PH domain of 伪PIX, 伪PIX is fully activated, possibly by a conformation change, and activates Rac1/Cdc42 leading to the activation of PAK, or cytoskeletal rearrangement. To clarify this, the same scheme is presented for both receptors, although the phosphorylation site of EphB2 which serves as a binding site for p85 is not known, and EphB2 has only one catalytic domain in the cytoplasmic region. TK, tyrosine kinase domain. (c) Integrin receptor engaged with fibronectin stimulates p125FAK autophosphorylation, which creates the c-Src SH2 domain binding motif. Src family proteins associated with p125FAK can potentiate tyrosine phosphorylation of p130cas, which provides SH2 binding sites for the Nck. On the other hand, the SH2 domains of the p85 subunit bind to other phosphotyrosine residues of the p125FAK. Thus, 伪PIX is recruited to the complex by either subsequent association with PAK and Nck, or association with the p85 subunit, following the activation of Rac1/Cdc42 as described aboveFull size imageIt is also tempting to speculate that activated 伪PIX may lead to oncogenic transformation as well as invasiveness and metastasis of tumors due to its effects on cytoskeletal organization and cell motility on extracellular substratum. The determination of the role of 伪PIX itself in early embryonal development may require another series of experiments.Materials and methodsPlasmids and antibodiesPlasmids encoding full-length cDNAs of human 伪PIX and 尾PIX were gifts from T Nagase (KIAA0006 and KIAA0142, respectively). Deletion mutants of PIX tagged with a myc epitope at the C-terminus were generated using a PCR-based technique (Tanaka et al., 1997). All constructs were subcloned into the pCS2+ vector for transfection into 293T cells, into pGEX2T (Pharmacia) to make bacterial GST fusion constructs, and into pGHXP (a derivative of pGEM-HE, a gift from L Zon) to make mRNA for microinjection into Xenopus embryos. pGEX-PBD was generated by cloning of a PCR-amplified fragment of putative Cdc42 and the Rac binding domain of human PAK1 (amino acids 70鈥夆€撯€?33) into pGEX2T for affinity precipitation. Plasmid encoding full-length cDNA of mouse EphB2 was a gift from T Pawson. Amino-terminal myc-tagged wild-type and mutant (P13A, K299R and H83L H86L) human PAK1 clones, wild-type and mutant (R308K) Nck clones, and wild-type and mutant (T17N) clones of Cdc42, Rac1 and RhoA were kindly provided by Bruce J Mayer (Lu et al., 1997; Tanaka et al., 1995). PAK1 (P191G,R192A) tagged with myc epitope at the C-terminus was generated and cloned into the pCS2+ vector. Wild-type and mutated forms (螖p85, p85(III)) of bovine p85伪, and p110 subunit of PI3-kinase with myristoylation signal (myr p110) were provided by W Ogawa (Yonezawa et al., 1992; Hara et al., 1994). Plasmid encoding p130cas was provided from H Hirai (Nakamoto et al., 1996), and chicken p125FAK from JT Parsons. The monoclonal antibody for Nck was from Transduction Laboratories. The monoclonal antibodies for the myc epitope tag and HA epitope tag were from Santa Cruz and Berkeley Antibody Co., respectively. Polyclonal antibodies for Cdc42 and Rac1 were from Santa Cruz, and PDGF type B receptor was from Upstate Biotechnology. Rabbit polyclonal serum for mouse EphB2 was a gift from EB Pasquale.Cell culture, transfections and in vitro kinase assay293T human embryonal kidney cells (expressing simian virus 40T antigen) were transfected with a maximum of 20鈥壩糶 of plasmid DNA per 10鈥塩m diameter dish by a calcium phosphate coprecipitation method with concurrent treatment with 25鈥壩?span >M chloroquine essentially as described previously (Tanaka et al., 1995). At 18鈥夆€撯€?4鈥塰 after transfection, cells were placed in medium with 0.5% FBS and incubated for a further 12鈥塰. Cells were either untreated or treated with human, recombinant PDGF-BB (Sigma) at final concentration 40鈥塶g/ml for 15鈥塵in prior to lysis. For fibronectin stimulation, transfected cells were plated onto culture dishes coated with human plasma fibronectin (Gibco BRL) for 30鈥塵in. For treatment of transfected cells with synthetic dipalmitoyl phosphoinositides, phosphatidylinositol 3-phosphate (DiC16PI3-P), phosphatidylinositol-3,4-bisphosphate (DiC16PI-3,4-P2) or phosphatidylinositol-3,4,5-trisphosphate (DiC16PI-3,4,5-P3) (Matreya, Inc.) was sonicated and added to the medium at a concentration of 10鈥壩?span >M for 20鈥塵in as described (Derman et al., 1997). Cells were harvested for immunoprecipitation and kinase assays as previously described (Tanaka et al., 1995). To purify myc-tagged PAK protein, monoclonal antibody against myc epitope (1鈥壩糶) was incubated with 500鈥壩糽 cell lysate for 2鈥塰 at 4掳C and precipitated with protein G-agarose (Boehringer Mannheim). Immunoprecipitates were washed extensively before immunoblotting or in vitro kinase assay with MBP as substrate. The results were visualized by Bio Imaging Analyzer (BAS1000 Fuji), and the image was quantitated with NIH Image software. The relative activity of PAK was calculated by adjusting the densitometric value and standardization.In vitro association of HA-p85 with GST-PIX293T cell lysates transiently expressing p85 tagged with HA epitope at the amino-terminus were prepared as described above. Immobilized GST-尾PIX was incubated with cell lysates for 4鈥塰 at 4掳C. Precipitates were washed three times with lysis buffer, and bound proteins were eluted in sample buffer subjected to SDS鈥夆€撯€塒AGE, Western blotted, and probed with anti-HA mAb.Affinity precipitationAffinity precipitation with GST-PBD was as described previously (Taylor and Shalloway, 1996; Bagrodia et al., 1998) except that 293T cells were lysed in lysis buffer [20鈥塵M Tris-HCl (pH 7.5), 150鈥塵M NaCl, 20鈥塵M MgCl2, 1鈥塵M Na3VO4, 0.5% Triton X-100, 5鈥壩糶/ml aprotinin, 1鈥塵M PMSF], and precipitates were washed three times in the same buffer.GDP dissociation assaysThe [3H]GDP dissociation assay was performed as described (Zheng et al., 1995). Briefly, the [3H]GDP-bound form of Cdc42 or Rac1 was first made by incubating E. coli-derived GST-tagged p21 (2鈥塸mol) with 1鈥壩?span >M [3H]GDP for 10鈥塵in at 30掳C. After the first incubation, 0.2鈥?span >M MgCl2 was added to give a final concentration of 20鈥塵M and the mixture was immediately cooled on ice. The dissociation of [3H]GDP from GST-Cdc42 was started in a mixture containing 6.75鈥塵M MgCl2, 67.5鈥壩?span >M GTP and 50鈥塶M of purified GST-伪PIX153鈥夆€撯€?45, GST-尾PIX11鈥夆€撯€?08 (corresponding to the region containing SH3, DH and PH domain) or GST followed by incubation with or without 50鈥壩?span >M of DiC16PI-3,4,5-P3 at 30掳C. The reaction was stopped by the addition of 2鈥塵l of ice-cold 20鈥塵M Tris (pH 7.5) containing 25鈥塵M MgCl2 and 100鈥塵M NaCl. The diluted mixture was filtered through nitrocellulose filters and the filters were washed with the same solution four times. Time courses for the dissociation of [3H]GDP from GST-p21 were measured.Embryos, microinjection and mesoderm aggregate spreading assayCapped synthetic mRNAs were generated as described (Krieg et al., 1989) by in vitro transcription of linearized pGHXP vectors with the inserts. Fertilized embryos were prepared as described (Newport and Kirschner, 1982). Staging was according to Nieuwkoop and Faber (1967). RNA was microinjected into two dorsal blastomeres at the four-cell stage following published procedures (Tanaka et al., 1998). Mesoderm aggregate spreading assay was performed as described with modification to use dorsal marginal zone cells (Symes and Mercola, 1996). Briefly, dorsal marginal zones (DMZ) of gastrulae at stage 10 were microdissected and dispersed in Ca2+- and Mg2+-free medium (CMFM), and the epidermal cell layer was removed. The dispersed DMZ cells, which contain mesoderm cells and some endoderm cells, were reaggregated in 0.6脳Marc\'s Modified Ringer\'s (1脳MMR; 100鈥塵M NaCl, 2鈥塵M KCl, 1鈥塵M MgCl2, 2鈥塵M CaCl2, 5鈥塵M HEPES) containing 0.5% bovine serum albumin with PDGF AA (Gibco BRL) at a concentration of 50鈥塶g/ml, or PDGF AA plus wortmannin at 100鈥塶M. Aggregates were then plated on dishes coated with human fibronectin (Gibco BRL) in 0.6脳MMR containing PDGF AA, or PDGF AA plus wortmannin at the same concentration as described above. dorsal marginal zone ReferencesArcaro A and Wymann MP. . 1993 Biochem. J. 296: 297鈥?01.Ataliotis P, Symes K, Chou MM, Ho L and Mercola M. . 1995 Development 121: 3099鈥?110.Bagrodia S, Taylor K, Jordon KA, Van Aelst L and Cerione RA. . 1998 J. Biol. Chem. 273: 23633鈥?3636.Brown JL, Stowers L, Baer M, Trejo J, Coughlin S and Chant J. . 1996 Curr. Biol. 6: 598鈥?05.Chen HC, Appeddu PA, Isoda H and Guan JL. . 1996 J. Biol. Chem. 271: 26329鈥?6334.Daniels RH, Hall PS and Bokoch GM. . 1998 EMBO J. 17: 754鈥?64.Daniels RH, Zenke FT and Bokoch GM. . 1999 J. Biol. Chem. 274: 6047鈥?050.Derman MP, Toker A, Hartwig JH, Spokes K, Falck JR, Chen CS, Cantley LC and Cantley LG. . 1997 J. Biol. Chem. 272: 6465鈥?470.Galisteo ML, Chernoff J, Su YC, Skolnik EY and Schlessinger J. . 1996 J. Biol. Chem. 271: 20997鈥?1000.Gringhuis SI, de Leij LF, Coffer PJ and Vellenga E. . 1998 Mol. Cell. Biol. 18: 1725鈥?735.Han J, Luby-Phelps K, Das B, Shu X, Xia Y, Mosteller RD, Krishna UM, Falck JR, White MA and Broek D. . 1998 Science 279: 558鈥?60.Hara K, Yonezawa K, Sakaue H, Ando A, Kotani K, Kitamura T, Kitamura Y, Ueda H, Stephen L and Jackson TR. . 1994 Proc. Natl. Acad. Sci. USA 91: 7415鈥?419.Holland SJ, Gale NW, Gish GD, Roth RA, Songyang Z, Cantley LC, Henkemeyer M, Yancopoulos GD and Pawson T. . 1997 EMBO J. 16: 3877鈥?888.Hawkins PT, Eguinoa A, Qiu RG, Stokoe D, Cooke FT, Walters R, Wennstrom S, Claesson-Welsh L, Evans T, Symons M, Stephens L. . 1995 Curr. Biol. 5: 393鈥?03.Jones SD, Ho L, Smith JC, Yordan C, Stiles CD and Mercola M. . 1993 Dev. Genet. 14: 185鈥?93.Kanai F, Ito K, Todaka M, Hayashi H, Kamohara S, Ishii K, Okada T, Hazeki O, Ui M and Ebina Y. . 1993 Biochem. Biophys. Res. Commun. 195: 762鈥?68.Keely PJ, Westwick JK, Whitehead IP, Der CJ and Parise LV. . 1997 Nature 390: 632鈥?36.Keller R and Jansa S. . 1992 Dev. Dyn. 195: 162鈥?76.Klippel A, Reinhard C, Kavanaugh WM, Apell G, Escobedo MA and Williams LT. . 1996 Mol. Cell. Biol. 16: 4117鈥?127.Krieg PA, Sakaguchi DS and Kintner CR. . 1989 Nucleic Acids Res. 17: 10321鈥?0335.Leeuw T, Fourest-Lieuvin A, Wu C, Chenevert J, Clark K, Whiteway M, Thomas DY and Leberer E. . 1995 Science 270: 1210鈥?213.Lu W, Katz S, Gupta R and Mayer BJ. . 1997 Curr. Biol. 7: 85鈥?4.Manser E, Loo TH, Koh CG, Zhao ZS, Chen XQ, Tan L, Tan I, Leung T and Lim L. . 1998 Mol. Cell 1: 183鈥?92.Meisenhelder J and Hunter T. . 1992 Mol. Cell. Biol. 12: 5843鈥?856.Nakamoto T, Sakai R, Ozawa K, Yazaki Y and Hirai H. . 1996 J. Biol. Chem. 271: 8959鈥?965.Newport J and Kirschner M. . 1982 Cell 30: 687鈥?96.Nieuwkoop PD and Faber J. . 1967 Normal Table of Xenopus laevis (Daudin). North Holland: Amsterdam. Google Scholar聽 Nishimura R, Li W, Kashishian A, Mondino A, Zhou M, Cooper J and Schlessinger J. . 1993 Mol. Cell. Biol. 13: 6889鈥?896.Obermeier A, Ahmed S, Manser E, Yen SC, Hall C and Lim L. . 1998 EMBO J. 17: 4328鈥?339.Oh WK, Yoo JC, Song YH, Kim MG and Park D. . 1997 Biochem. Biophys. Res. Commun. 235: 794鈥?98.Schlaepfer DD, Broome MA and Hunter T. . 1997 Mol. Cell. Biol. 17: 1702鈥?713.Sells MA, Knaus UG, Bagrodia S, Ambrose DM, Bokoch GM and Chernoff J. . 1997 Curr. Biol. 7: 202鈥?10.Shigematsu H, Iwasaki H, Otsuka T, Ohno Y, Arima F and Niho Y. . 1997 J. Biol. Chem. 272: 14334鈥?4340.Stein E, Huynh-Do U, Lane AA, Cerretti DP and Daniel TO. . 1998 J. Biol. Chem. 273: 1303鈥?308.Symes K and Mercola M. . 1996 Proc. Natl. Acad. Sci. USA 93: 9641鈥?644.Tanaka M, Gupta R and Mayer BJ. . 1995 Mol. Cell. Biol. 15: 6829鈥?837.Tanaka M, Lu W, Gupta R and Mayer BJ. . 1997 Proc. Natl. Acad. Sci. USA 94: 4493鈥?498.Tanaka M, Wang DY, Kamo T, Igarashi H, Wang Y, Xiang YY, Tanioka F, Naito Y and Sugimura H. . 1998 Oncogene 17: 1509鈥?516.Taylor SJ and Shalloway D. . 1996 Curr. Biol. 6: 1621鈥?627.Tolias KF, Cantley LC and Carpenter CL. . 1995 J. Biol. Chem. 270: 17656鈥?7659.Yano H, Nakanashi S, Kimura K, Hanai N, Saitoh Y, Fukui Y, Nonomura Y and Matsuda Y. . 1993 J. Biol. Chem. 268: 25846鈥?5856.Yonezawa K, Ueda H, Hara K, Nishida K, Ando A, Chavanieu A, Matsuba H, Shii K, Yokono K, Fukui Y, Calas B, Grigorescu F, Dhand R, Gout I, Otsu M, Waterfield MD and Kasuga M. . 1992 J. Biol. Chem. 267: 25958鈥?5965.Zhang S, Han J, Sells MA, Chernoff J, Knaus UG, Ulevitch RJ and Bokoch GM. . 1995 J. Biol. Chem. 270: 23934鈥?3936.Zheng Y, Bagrodia S and Cerione RA. . 1994 J. Biol. Chem. 269: 18727鈥?8730.Zheng Y, Hart M and Cerione RA. . 1995 Methods in Enzymology 256: 77鈥?4.Zisch AH, Kalo MS, Chong LD and Pasquale EB. . 1998 Oncogene 16: 2657鈥?670.Download referencesAcknowledgementsWe thank Drs Bruce J Mayer, Tony Pawson, Wataru Ogawa, Hisamaru Hirai, JT Parsons, Elena B Pasquale and Takahiro Nagase of Kazusa DNA Research Institute for providing us with plasmids and the antibody used in these experiments. This work was supported by the Uehara Memorial Foundation, Smoking Research Foundation, Grants in Aid for Cancer Research from the Ministry of Health and Welfare of Japan (9-3), Grant in Aid for Scientific Research (B), on Priority area, and Encouragement of Young Scientists, the Ministry of Education, Culture, Sports and Science of Japan, and Foundation for advancement of international science (FAIS). Yue Zhu is a recipient of an AIEJ short-term student exchange program, peace and friendship scholarship.Author informationAffiliationsThe First Department of Pathology, Hamamatsu University School of Medicine, 3600 Handa-cho, Hamamatsu, 431-3192, JapanShigeto Yoshii,聽Masamitsu Tanaka,聽Yoshirou Otsuki,聽Dong-Yu Wang,聽Rong-Jun Guo,聽Yue Zhu,聽Ririko Takeda聽 聽Haruhiko SugimuraFirst Department of Medicine, Hamamatsu University School of Medicine, 3600 Handa-cho, Hamamatsu, 431-3192, JapanShigeto Yoshii,聽Hiroyuki Hanai聽 聽Eizo KanekoAuthorsShigeto YoshiiView author publicationsYou can also search for this author in PubMed聽Google ScholarMasamitsu TanakaView author publicationsYou can also search for this author in PubMed聽Google ScholarYoshirou OtsukiView author publicationsYou can also search for this author in PubMed聽Google ScholarDong-Yu WangView author publicationsYou can also search for this author in PubMed聽Google ScholarRong-Jun GuoView author publicationsYou can also search for this author in PubMed聽Google ScholarYue ZhuView author publicationsYou can also search for this author in PubMed聽Google ScholarRiriko TakedaView author publicationsYou can also search for this author in PubMed聽Google ScholarHiroyuki HanaiView author publicationsYou can also search for this author in PubMed聽Google ScholarEizo KanekoView author publicationsYou can also search for this author in PubMed聽Google ScholarHaruhiko SugimuraView author publicationsYou can also search for this author in PubMed聽Google ScholarCorresponding authorCorrespondence to Haruhiko Sugimura.Rights and permissionsReprints and PermissionsAbout this articleCite this articleYoshii, S., Tanaka, M., Otsuki, Y. et al. 伪PIX nucleotide exchange factor is activated by interaction with phosphatidylinositol 3-kinase. Oncogene 18, 5680鈥?690 (1999). https://doi.org/10.1038/sj.onc.1202936Download citationReceived: 14 December 1998Revised: 26 April 1999Accepted: 26 April 1999Published: 08 October 1999Issue Date: 07 October 1999DOI: https://doi.org/10.1038/sj.onc.1202936KeywordsPAKPDGFEphB2Nckphosphoinositide Young Seo Kim, Min Young Noh, Ji Young Kim, Hyun-Jeung Yu, Kyung Suk Kim, Seung Hyun Kim Seong-Ho Koh Molecular Neurobiology (2013) Takashi Yokota, Jun Kouno, Koji Adachi, Hiroshi Takahashi, Akira Teramoto, Koshi Matsumoto, Yuichi Sugisaki, Masamitsu Onda Tatsuhiko Tsunoda Acta Neuropathologica (2006)