JSAP1/JIP3 Cooperates with Focal Adhesion Kinase to Regulate c-Jun N-terminal Kinase and Cell Migration*

c-Jun N-terminal kinase (JNK)/stress-activated protein kinase-associated protein 1 (JSAP1) (also termed JNK-interacting protein 3; JIP3) is a member of a family of scaffold factors for the mitogen-activated protein kinase (MAPK) cascades, and it also forms a complex with focal adhesion kinase (FAK). Here we demonstrate that JSAP1 serves as a cooperative scaffold for activation of JNK and regulation of cell migration in response to fibronectin (FN) stimulation. JSAP1 mediated an association between FAK and JNK, which was induced by either co-expression of Src or attachment of cells to FN. Complex formation of FAK with JSAP1 and p130 Crk-associated substrate (p130Cas) resulted in augmentation of FAK activity and phosphorylation of both JSAP1 and p130Cas, which required p130Cas hyperphosphorylation and was abolished by inhibition of Src. JNK activation by FN was enhanced by JSAP1, which was suppressed by disrupting the FAK/p130Cas pathway by expression of a dominant-negative form of p130Cas or by inhibiting Src. We also documented the co-localization of JSAP1 with JNK and phosphorylated FAK at the leading edge and stimulation of cell migration by JSAP1 expression, which depended on its JNK binding domain and was suppressed by inhibition of JNK. The level of JSAP1 mRNA correlated with advanced malignancy in brain tumors, unlike other JIPs. We propose that the JSAP1·FAK complex functions cooperatively as a scaffold for the JNK signaling pathway and regulator of cell migration on FN, and we suggest that JSAP1 is also associated with malignancy in brain tumors.

Cell adhesion to the extracellular matrix regulates many cellular functions, including cell differentiation, proliferation, apoptosis, and migration (1). Integrin stimulation by extracellular matrix proteins such as FN 2 leads to activation of MAPKs, including JNK and extracellular signal-regulated kinase (ERK) in a variety of cell types. Among the types of MAPK signaling transmitted by integrins, JNK activation is believed to correlate particularly with increased cell migration and invasion (2,3). FN stimulation of cells is known to activate JNK through an FAK/ p130 Cas /CrkII/DOCK180/ELMO/Rac1 pathway (2, 4 -9). This signaling is highly conserved across species and seems to be important for promoting cell migration. Mutations in CrkII, DOCK180, ELMO, Rac1, and JNK homologues in Drosophila all result in phenotypes with defects in cell migration (3,4,10). Embryonic fibroblasts from JNK-deficient and MAPK kinase kinase1 (MEKK1)-deficient mice are impaired in serum-induced JNK activation and cell migration (3,(11)(12)(13). Recent studies have demonstrated that JNK regulates cell migration by phosphorylating paxillin and regulating microtubule assembly by phosphorylating microtubule-associated protein 2 (14 -16). However, the mechanism by which JNK is recruited and activated in association with focal adhesions is largely unknown.
JIP family members (JIP1, -2, and -3) were originally identified as JNK-binding proteins and are thought to serve as scaffold factors for MAPK signaling cascades (17)(18)(19). JSAP1 binds not only to MAPK pathway constituents, including all JNK isoforms, MAPK kinases (MKK) 1, 4, and 7, MEKK1, mixed-lineage protein kinase 3, and c-Raf-1, but also to the motor protein kinesin (17)(18)(19). We found previously that JSAP1 interacts with FAK (20). This association was enhanced by c-Src, and it promoted cell spreading on FN. FAK is known to be essential for cell migration and for transmitting integrin-stimulated signals to MAPK (1,6). FAK activated by cell adhesion to extracellular matrix undergoes autophosphorylation at Tyr-397 and thereby associates with Src family kinases, leading to enhancement of its tyrosine phosphorylation and kinase activity. Binding of FAK to c-Src also induces the formation of a multimolecular signaling complex in which FAK functions as a scaffold (21). Notably, JNK activation and FAK autophosphorylation at Tyr-397 are both attenuated in JSAP1 knock-out mice (22), suggesting that JSAP1 might be involved in FAK-mediated JNK activation.
Although many studies have suggested critical roles for JNK signaling in tumor cell migration and invasion, the role of JNK scaffold proteins in tumor malignancy is not well examined. Here, we provide evidence that the JSAP1⅐FAK complex functions as an effective scaffold for the JNK pathway and particularly for stimulating cell migration, and that JSAP1 mRNA is elevated in brain tumors. These results suggest that JSAP1 may also participate in the acquisition of malignancy in brain tumors.
Measurement of FAK Activity-As described previously (20), immunoprecipitated FAK was washed twice in kinase buffer (20 mM HEPES (pH 7.4), 150 mM NaCl, 10 mM MgCl 2 ) and was incubated with 25 Ci of [␥-32 P]ATP for 30 min at 37°C. The reaction mixtures were subjected to SDS-PAGE, and the 32 P-labeled materials were detected by a Typhoon 9200 Image Analyzer (Amersham Biosciences).
Immunofluorescence Staining-Glass coverslips were incubated with 10 g/ml FN in phosphate-buffered saline overnight at 4°C and then blocked with 10 mg/ml bovine serum albumin. Cells were allowed to spread onto the FN-coated coverslips for the indicated periods, fixed with 4% paraformaldehyde in phosphate-buffered saline for 15 min, and then permeabilized with 4% paraformaldehyde/0.5% Triton X-100 for 5 min. The cells were then stained with the indicated antibodies and examined by confocal laser microscopy (Carl Zeiss).
Measurement of Cell Motility-U87MG cells were co-transfected with 0.5 g of pHA262pur, pRK-GFP, FLAG-JNK1, and 1 g of His-S-JSAP1-WT, -⌬JBD, or control plasmids. Transfectants were selected with 1.25 g/ml puromycin for 48 h. This selection for transient transfectants routinely resulted in 90% positive cells expressing GFP. Puromycin-selected cells were replated on 35-mm glass-bottom dishes coated with 10 g/ml FN; the cells were cultured overnight in complete medium. Cell movements were monitored using Olympus inverted microscopes. Video images were collected at 10-min intervals for 4 h. Image stacks were converted to QuickTime movies, and the positions of nuclei were tracked to quantify cell motility using Move-tr/2D software (Library, Tokyo, Japan).
Measurement of JNK Activity-JNK activity was measured using the SAPK/JNK assay kit from Cell Signaling. Briefly, the cell lysates were incubated with GST-c-Jun (amino acid residues 1-89), and precipitates were incubated in the presence of ATP for 30 min at 30°C. Phosphorylation of c-Jun on Ser-63 by JNK bound to GST-c-Jun was detected by immunoblotting using anti-phospho-c-Jun (Ser 63) antibody.
Embryonic Stem Cell Culture-The murine ES cell line E14K (gift of Dr. Hiroshi Nishina, Tokyo Medical and Dental University) was maintained as described previously (26) with the minor modification that KnockOut Serum Replacement (Invitrogen) was used instead of fetal calf serum. The generation of Jsap1(Ϫ/Ϫ) ES cell lines will be described elsewhere.
Clinical Samples and Histology-Under an institutional review board-approved protocol, fresh human brain tumor tissues were obtained from 26 patients with astrocytic tumors who underwent therapeutic removal of brain tumors. Normal brain tissues were obtained from three patients undergoing temporal lobectomy for epilepsy. Histological diagnosis was made by standard light-microscopic evaluation. The classification of human brain tumors used in this study was based on the revised WHO criteria for tumors of the central nervous system (27). The 26 astrocytic tumors consisted of 7 low-grade astrocytomas, 8 anaplastic astrocytomas, and 11 glioblastomas. All of the tumor tissues were obtained at primary resection, and none of the patients had been subjected to chemotherapy or radiation therapy before resection.

RESULTS
JSAP1 Forms a Scaffold for JNK with FAK-JSAP1 has been identified as a scaffold factor for JNK and we previously reported that JSAP1 also associates with FAK (17,20). To examine whether formation of the JSAP1 and FAK complex affects JNK signaling, e.g. as a scaffold factor, JSAP1-WT or a mutant lacking the JNK-binding domain of JSAP1 (⌬JBD) was co-expressed with FAK, c-Src, and JNK1 in 293T cells, and their interaction was analyzed by immunoprecipitation. Co-expression of FAK/c-Src significantly stimulated complex formation between JSAP1 and JNK1, and the JNK1 in the complex was phosphorylated (Fig.  1B). Although both JSAP1-WT and JSAP1-⌬JBD were tyrosine-phosphorylated and co-precipitated with FAK, only JSAP1-WT but not JSAP1-⌬JBD was co-precipitated with JNK1. To confirm that both FAK and JNK1 reside on the same JSAP1 scaffold, immunoprecipitation with an antibody against FAK was performed. JNK1 was co-precipitated with FAK only in cells co-expressing JSAP1-WT but not JSAP1-⌬JBD, whereas both JSAP1-WT and JSAP1-⌬JBD were co-precipitated with FAK ( Fig. 1C). FAK was co-precipitated with JNK1 only in the cells co-expressing JSAP1-WT but not JSAP1-⌬JBD. The tyrosine phosphorylation of JSAP1-WT and JSAP1-⌬JBD was equally augmented by adhesion of cells to FN in place of FAK⅐Src expression, and only JSAP1-WT but not JSAP1-⌬JBD associated with JNK1 (Fig. 1D). Moreover, immunoprecipitation analysis using lysates from mouse brain tis-sue demonstrated the existence of endogenous JSAP1⅐FAK⅐JNK complex (Fig. 1E). These results suggest that JSAP1 forms a scaffold for the JNK pathway with FAK following formation of a FAK⅐Src complex, and thus the association of JSAP1 with JNK is facilitated under conditions that induce FAK⅐Src complex formation such as FN stimulation.
JSAP1 Elevates FAK Activity and p130 Cas Phosphorylation-FAK is a critical regulator of FN-induced JNK activation and cell migration. FAK activity is negatively regulated by its N-terminal FERM domain (29,30), which binds to JSAP1. We next investigated whether JSAP1 affects FAK activity. In vitro kinase assays showed that expression of JSAP1-WT or JSAP1-⌬JBD increased FAK activity by 3-fold ( Fig. 2A). In contrast, JSAP1-⌬2, which fails to bind to FAK, had no effect. FAK activity was elevated 10-fold by co-expression of c-Src under the same conditions, indicating that activation by JSAP1 is substantial but not maximal.
p130 Cas is phosphorylated by a FAK⅐Src complex and mediates FAKinduced JNK activation by forming a complex with CrkII (5). The effects of JSAP1 expression on p130 Cas phosphorylation and FAK activity were studied using deletion mutants of p130 Cas (Fig. 2, B and C). p130 Cas expressed by transfection of the gene was not tyrosine-phosphorylated, and co-expression of FAK slightly induced its phosphorylation. Although expression of JSAP1 alone did not affect p130 Cas phosphorylation, co-expression of JSAP1 with FAK strongly stimulated it. This JSAP1⅐FAK-enhanced phosphorylation of p130 Cas was not observed with the mutants p130 Cas -⌬SH3 or -⌬SB, but p130 Cas -⌬SD was slightly phosphorylated under the same conditions. Because FAK is known to interact with the SH3 domain of p130 Cas through its C-terminal proline-rich region and to phosphorylate the Src-binding domain of p130 Cas , these results suggest that JSAP1 stimulates phosphorylation of p130 Cas by Src at the substrate domain followed by phosphorylation of its Src-binding domain by FAK. In vitro kinase assays showed that FAK activity was slightly elevated by p130 Cas expression, which was stimulated by co-expression of JSAP1. In contrast, p130 Cas mutants had no effect. JSAP1 phosphorylation was also induced by co-expression of FAK with p130 Cas . This is supported by the results presented in Fig. 2D showing that phosphorylation of both p130 Cas and JSAP1 induced by co-expression with FAK was reduced by Src inhibitor PP2 treatment. These results suggest that complex formation of FAK with JSAP1 and p130 Cas promotes p130 Cas phosphorylation and FAK activation through Src.
JSAP1 but Not JIP1 Enhances JNK Activation-Although JSAP1 is a member of the JIP family, the primary structure of JSAP1 is different from the others (18). To compare JSAP1 with JIP1 in terms of association with FAK, JSAP1 or JIP1 tagged with FLAG epitope was co-expressed with FAK and c-Src, and immunoprecipitation analysis was carried out. As shown in Fig. 3A, only JSAP1 but not JIP1 co-precipitated FAK. JNK activation was faintly induced by cultivation of cells on FN or by JSAP1 expression when cultured on PLL (Fig. 3B). In contrast, consistent with its binding to FAK, the expression of JSAP1 but not JIP1 significantly enhanced the JNK activating signal from FN (Fig. 3B). To confirm the role of JSAP1 in JNK activation, U87MG cells co-transfected with either JSAP1-WT or JSAP1-⌬JBD and JNK1 were cultured on FN, and JNK phosphorylation was examined (Fig. 3C). JSAP1-WT expression augmented JNK phosphorylation in a dose-dependent manner, whereas JSAP1-⌬JBD expression did not affect it. As shown in Fig.  3D, the JNK phosphorylation enhanced by JSAP1 was suppressed by the Src inhibitor and a dominant negative form of p130 Cas (⌬SD). These results suggest that JSAP1 expression promotes FN-induced JNK activation by facilitating the FAK/p130 Cas pathway.
Because FAK, p130 Cas , CrkII, and Rac1 are known to accumulate at the leading edge in migrating cells (31,32), we next examined the localization of JSAP1 in U87MG cells cultured on FN. Immunofluorescence staining demonstrated the extension of tubulin-containing microtubules toward the cell periphery (Fig. 3E) and the well organized distribution of paxillin (Fig. 3F) and autophosphorylated FAK (Fig. 3G) at the  NOVEMBER 11, 2005 • VOLUME 280 • NUMBER 45 leading front of migrating U87MG cells. It has been reported that JSAP1 is transported to the tips of neurites through its interaction with the tetratricopeptide repeat domain of kinesin light chains (24,33). We observed that JSAP1 is distributed diffusely in the cytoplasm but is also concentrated at the cell periphery where paxillin is well organized and microtubules have elongated. These results indicate that JSAP1 functions together with FAK scaffold at the leading edge.

JSAP1 Modulates JNK Activation and Cell Migration
JSAP1-deficient ES Cells Are Impaired in Lamellipodial Protrusion-ES cells provide a highly informative in vitro cell system to study the phenotypic effects of gene disruption at the cellular level. Jsap1 Ϫ/Ϫ murine ES cell lines have been described elsewhere (Fig. 4A). Attachment of wild-type Jsap1 ϩ/ϩ ES cells to FN induced lamellipodial protrusion and membrane ruffling (Fig. 4, B and C). In contrast, Jsap1-null (Jsap1 Ϫ/Ϫ ) ES cells displayed a substantial decrease of the rate of lamellipodial protrusion and membrane ruffling (Fig. 4, B and D). Rac1 and p130 Cas /Crk has been shown to induce lamellipodial protrusion and membrane ruffle formation associated with cell migration (5,38). These observations suggest that JSAP1 is required for FN-induced lamellipodial protrusion and membrane ruffling through the p130 Cas /Crk/Rac1 pathway, which led us to test the hypothesis that JSAP1 contributes to cell motility.
JSAP1 Expression Promotes Cell Migration-To explore whether JSAP1 is involved in cell migration, the distribution of JSAP1 and JNK1 in wound-edge cells was monitored. As shown in Fig. 5A, JNK1 was barely localized at the leading edges of cells at the edge of in vitro wounds in cell monolayers expressing JNK1 alone. In contrast, co-expression of JSAP1 with JNK1 induced accumulation of JNK1 at the leading edge, suggesting the recruitment of JNK to the leading edge where JSAP1 and FAK also accumulate. Next, we examined the effect of JSAP1 expression on cell migration induced by FN and serum. Expres- sion of JSAP1-WT but not JSAP1-⌬JBD enhanced U87MG migration by 18% compared with control cells (p Ͻ 0.01, Fig. 5B). The cell migration of mock and JSAP-WT-transfected cells was reduced to 44 and 42% of control cells by treatment with the JNK inhibitor SP600125. In parallel with suppression of cell migration, SP600125 treatment resulted in the loss of motile cell morphology with an increase in the number and size of focal adhesions and less-oriented actin stress fiber formation (Fig. 5C) and the suppression of JNK activity (Fig. 5D). These results indicate that activation of JNK by JSAP1 can contribute to cell migration.

DISCUSSION
JNK is activated in response to a variety of extracellular and intracellular stimuli, and it plays crucial roles in cellular processes that include cell proliferation, differentiation, apoptosis, and migration. However, the mechanisms determining specificity and efficiency of JNK activation remain to be clarified. In the present study, we have tested for functional cooperation between JSAP1 and FAK. Our novel findings are as follows: 1) the function of JSAP1 as a JNK scaffold is augmented by its binding to FAK; 2) JSAP1 and p130 Cas function cooperatively to transmit signaling to JNK and to regulate FAK activity; 3) Jsap1-null ES cells are impaired in FN-induced lamellipodial protrusion formation and membrane ruffling; 4) JSAP1 expression stimulates FN-induced cell migration, and this stimulation depends on the ability of JSAP1 to promote JNK activation; and 5) JSAP1 mRNA expression correlates with the malignant phenotype of brain tumors. These studies define a signaling pathway that stimulates FN-induced JNK activation and cell migra- tion that depends on mutual regulation and functional cooperation between JSAP1 and FAK.
FAK autophosphorylation at Tyr-397 following adhesion depends on its association with Src family kinases, leading to the formation of multimolecular signaling complexes in which FAK serves as a scaffold (1,6,34). FAK scaffolding is thought to function in integrin-mediated signal transduction (1,6). Among various signaling pathways, the FAK/ p130 Cas pathway is believed to be essential for FN-induced JNK activation and cell migration (2,4,5,(7)(8)(9)15), even though the precise mechanism by which JNK pathway constituents associate with the FAK scaffold is largely unknown. We have previously shown that the association of JSAP1 with FAK is enhanced by binding of c-Src to FAK, and adhesion of cells to FN induces FAK⅐c-Src⅐JSAP1 complex formation (20). Biochemical and gene disruption studies indicate that JSAP1 functions as a scaffold for JNK signaling (17)(18)(19)35). In the present study, we showed that the association of JSAP1 with JNK and subsequent JNK phosphorylation were enhanced by either FAK⅐Src expression or attachment of cells to FN. In contrast, JIP1, which associates with JNKpathway components but not with FAK, and JSAP1⌬JBD, which associates with FAK but not with JNK, both failed to augment FN-induced JNK activation. These results indicate that JSAP1 but not JIP1 serves as a specific scaffold for FN-induced JNK activation, supporting the importance of the association with FAK for this JIP protein to function as a scaffold.
Phosphorylation of p130 Cas is initiated by FAK, followed by interaction of the SH3 domain of p130 Cas with the C-terminal proline-rich region of FAK and hyperphosphorylation by Src. Hyperphosphorylated p130 Cas is critical for FAK-mediated JNK activation by binding to CrkII (25). We find that complex formation between JSAP1 and FAK induces activation of FAK and phosphorylation of both JSAP1 and p130 Cas ; this process requires p130 Cas hyperphosphorylation and is abolished by inhibition of Src (Fig. 2). FAK activity is known to be negatively regulated by its N-terminal FERM domain (29,30), which binds to JSAP1, and thus binding of JSAP1 to the N terminus of FAK might abrogate the autoinhibitory interaction of FAK. Our demonstration that FAK activity is significantly elevated by co-expression of JSAP1 with p130 Cas and is reduced by Src inhibition suggests that release from autoinhibition by JSAP1 cooperatively enhances FAK activity with the p130 Cas ⅐CrkII complex. This interpretation is supported by the fact that FAK autophosphorylation and JNK activation are both attenuated in JSAP1-deficient mice (22) and that complex formation of p130 Cas with CrkII followed by integrin engagement promotes FAK autophosphorylation, probably through Src family kinases (36). The stimulation of FN-induced JNK activation by JSAP1 was suppressed by expression of p130 Cas -⌬SD and by inhibiting Src, indicating essential roles for p130 Cas and Src (Fig. 3). Thus, the cooperation between JSAP1 and p130 Cas not only functions to mediate FAK-mediated JNK activation, but it may also affect the activation of FAK as an upstream regulator. We speculate that JSAP1 plays an important role in the spatial regulation of JNK pathway players together with the FAK scaffold in response to FN stimulation.
JSAP1 is transported to the tips of neurites along microtubules (24,33), which are known to associate with JNK, FAK, MEKK1, paxillin, and JSAP1, and to regulate cell migration (37)(38)(39). Our observations show-ing that JSAP1 co-localizes with autophosphorylated FAK and JNK at the leading edge (Fig. 5) suggests that a JSAP1⅐FAK scaffold may cooperatively enhance FAK and JNK activation at the leading edge. In fact, JSAP1 expression promoted FN-induced cell migration, which depended on its JNK binding domain and was suppressed by a JNK inhibitor accompanied by an increase in the number and size of focal adhesions and less-oriented actin stress fiber formation (Fig. 5). FAK, p130 Cas , CrkII, and Rac1 accumulate at the leading edge in migrating cells (31,32). Translocation of JSAP1 and FAK to the leading edge may place them at an appropriate site for facilitating FN-induced JNK activation and subsequent cell migration (Fig. 7).
JIP1-deficient mice are developmentally normal and viable (40,41), whereas JSAP1 null mice exhibit severe developmental defects (22,33). JSAP1-deficient mice show various developmental deficits in the brain, including axon guidance defects of the corpus callosum, suggesting importance of JSAP1 in mice development. The present study has shown that JSAP1 but not JIP1 associates with FAK and facilitates JNK activation and cell migration and that Jsap1-null ES cells are impaired in the formation of lamellipodial protrusions and membrane ruffling. JSAP1 is different from other JIP family members in its primary structure (18), consistent with a distinct role from that of JIP1. Transgenic expression of JIP1 in JSAP1-deficient mice only partially rescues the Jsap1 deficiency-induced developmental defects in parallel with partial restoration of JNK activation (22). From this study, an explanation for these findings is that even though JIP1 may partially compensate for JSAP1 by scaffolding JNK pathway components, the processes of JNK activation, FAK autophosphorylation, and cell migration induced by integrin-mediated signaling are not restored by JIP1 expression. The cooperative spatial regulation of JNK pathway components by JSAP1 and FAK may thus play critical roles in specific cellular and developmental processes.
Increased FAK expression and tyrosine phosphorylation are observed in many malignant tumors (42). We report here that the mRNA expression of JSAP1 is elevated, but expression of JIP1 and JIP2 is diminished in malignant brain tumors (Fig. 6), suggesting that JSAP1 may serve predominantly as a JNK scaffold in advanced brain tumors. Intriguingly, the expression of RhoA and its effectors, which regulate focal adhesion turnover, microtubule and actin stress fiber formation, and cell migration, as well as affecting FAK and JNK activation, are reduced in JSAP1deficient mice (33). Taken together with our result that JSAP1 facilitates JNK activation and cell migration together with FAK, it is possible that JSAP1 regulates integrin-mediated JNK activation and cell migration by controlling both the FAK and Rho pathways and that JSAP1 may thereby contribute to the acquisition of malignancy in brain tumors.
The data presented here provide the first evidence that JSAP1 serves as a scaffold for the efficient activation of JNK in response to FN stimulation, which is required to promote FN-induced cell migration and depends on mutual regulation and functional cooperation between JSAP1 and FAK. We conclude that cooperation between JSAP1 and FAK can assemble the elements of signaling modules that can include focal adhesion components, JNK pathway players, and microtubules, and that it can modulate FAK-mediated cell migration by regulating JNK activation.