Regulation of Mitogen-activated Protein Kinase Activation by the Cytoplasmic Domain of the α6 Integrin Subunit*

We examined the possibility that the α6A and α6B cytoplasmic domain variants of the α6β1 integrin differentially activate p42 and p44 mitogen-activated protein (MAP) kinases. P388D1 macrophages that express equivalent surface levels of either the α6Aβ1 or α6Bβ1 integrin were used to examine this issue. Adhesion to laminin-1 mediated by the α6Aβ1 integrin triggered activation of a substantial fraction of total p42 and p44 MAP kinases as assessed using a mobility shift assay, immunoblot analysis with a phosphospecific MAP kinase antibody, and an immune complex kinase assay. In contrast, ligation of the α6Bβ1 integrin did not trigger significant MAP kinase activation. These data were confirmed by antibody clustering of the α6β1 integrins. Both the α6Aβ1 and α6Bβ1 integrins were capable of activating the p70 ribosomal S6 kinase and this activation, unlike MAP kinase activation, is dependent on phosphoinositide 3-OH kinase. Activation of MAP kinase by α6β1 requires both Ras and protein kinase C activity. A functional correlate for differential activation of MAP kinase was provided by the findings that the α6Aβ1 transfectants migrated significantly better on laminin than the α6Bβ1transfectants and this migration was dependent on MAP kinase activity based on the use of the MAP kinase kinase (MEK1) inhibitor PD98059. Our findings demonstrate that the α6β1 integrin can activate MAP kinase, that this activation is regulated by the cytoplasmic domain of the α6 subunit, and that it relates to α6β1-mediated migration.

The mechanisms by which integrins modulate the activity of the ERK1 1 and ERK2 members of the MAP kinase family are of interest because these kinases regulate cell growth (1)(2)(3)(4) and have been implicated in other important functions such as cell migration (5). Initial studies demonstrated that MAP kinase could be activated by integrin-mediated attachment to the extracellular matrix (6 -8). More recently, the possible involvement of other signaling molecules including Ras and focal adhesion kinase (FAK) in MAP kinase activation by integrins has been explored, an area that remains controversial (9 -13). These studies, however, did not address the possibility that specificity exists among integrins for their ability to modulate MAP kinase activity. Recently, however, this possibility was substantiated by the finding that only a subset of integrins (␣ 1 ␤ 1 , ␣ 5 ␤ 1 , and ␣ v ␤ 3 ) can activate MAP kinase (14). The ability of these integrins to activate MAP kinase was found to be mediated by the recruitment and phosphorylation of Shc and to be specified by the membrane-proximal portion of the extracellular domain of the integrin ␣ subunit, its transmembrane segment, or both. Other integrins (␣ 2 ␤ 1 , ␣ 3 ␤ 1 , and ␣ 6 ␤ 1 ) were unable to induce MAP kinase activation in these assays (14).
An issue that has not been resolved with respect to integrinmediated activation of MAP kinase is the possible role of the cytoplasmic domain of the integrin ␣ subunit in regulating this activation. This possibility is supported by studies that have established a specific role for this cytoplasmic domain in regulating integrin-mediated functions (see, e.g., Ref. 15). In this direction, we are interested in understanding how the ␣ 6A and ␣ 6B isoforms of the ␣ 6 integrin subunit regulate the signaling properties of the ␣ 6 ␤ 1 integrin. Previously, we reported that ligation of the ␣ 6A ␤ 1 integrin in P388D1 macrophages induces a much more marked induction of protein tyrosine phosphorylation than does ligation of the ␣ 6B ␤ 1 integrin in the same cells (16). An important issue that arises from these data is whether the cytoplasmic domain sequence of the ␣ 6 subunit can influence MAP kinase activation. This issue is also timely because of the report discussed above concluding that the ␣ 6 ␤ 1 integrin cannot activate MAP kinase (14). In this study, we demonstrate that the ␣ 6 cytoplasmic domain sequence regulates MAP kinase activation and that the differential activation of MAP kinase by the ␣ 6A ␤ 1 and ␣ 6B ␤ 1 integrins is linked to the ability of these cells to migrate on laminin-1.

EXPERIMENTAL PROCEDURES
Cells and Cell Culture-Details on the generation and characterization of the P388D1/␣ 6 transfectants have been described previously (17). Cells were maintained in RPMI containing 25 mM HEPES, 15% certified fetal bovine serum, and 300 g/ml G418. The subclones used in this study had equivalent surface expression of either the ␣ 6A ␤ 1 or ␣ 6B ␤ 1 integrins as assessed by fluorescence-activated cell sorting using the ␣ 6 specific monoclonal antibody, 2B7 (17).
MAP Kinase Assays-Tissue culture dishes (60 mm, Costar) were coated at 4°C overnight with 30 g/ml laminin purified from the Englebreth-Holm-Swarm sarcoma as described (18). The dishes were rinsed twice with phosphate-buffered saline and once with Puck's Saline A prior to use. Confluent P388D1 cells were removed from tissue culture dishes by scraping and maintained in suspension for 30 min at room temperature in Puck's Saline A solution containing 0.5 mM MnCl 2 . Subsequently, the cells were either kept in suspension or plated on matrix-coated dishes and incubated at 37°C for the time periods noted in the Fig. legends. Following this incubation, the cells were washed once with ice-cold phosphate-buffered saline and then extracted in 0.2 ml of a buffer containing 20 mM HEPES, pH 7.4, 150 mM NaCl, 0.5% Triton X-100, 0.1% sodium deoxycholate, 1 mM Na 3 VO 4 , 1 mM phenymethylsulfonyl fluoride, 5 g/ml aprotinin, 5 g/ml pepstatin, and 5 g/ml leupeptin. Protein concentration was determined using the Bradford dye binding assay (Bio-Rad). Total cell extracts were electropho-resed on SDS-polyacrylamide gels, transferred to nitrocellulose, and blotted with one of the the following antibodies as indicated in the figure legends: ERK1 mAbs (Santa Cruz Biotechnology, Inc. and Transduction Laboratories, Inc.), a polyclonal ERK antibody (from John Blenis, Harvard Medical School), or a phosphospecific ERK polyclonal antibody (New England BioLabs, Inc.). MAP kinase activity was examined in the P388D1 extracts using an immune complex kinase assay following the manufacturer's protocol (New England BioLabs). In some experiments, MAP kinase activation was assessed after mAb clustering of the ␣ 6 ␤ 1 integrins on the surface of the P388D1 transfectants following the method of Kornberg et al. (19), except that 20 g/ml mouse Fc fragment was incubated with the cells prior to the addition of the integrin antibody. To examine the effects of calphostin C (Sigma), wortmannin (Calbiochem, La Jolla, CA), and PD98059 (Calbiochem) on MAP kinase activation, the cells were treated with these pharmacological reagents for 30 min in suspension prior to plating on laminin-1. Samples containing calphostin C were activated by fluorescent light.
p70 Ribosomal S6 Kinase (S6K) Assays-The same extracts of the P388D1 transfectants described above for the MAP kinase assays were also used to assay the activation of S6K by a mobility shift assay. Briefly, 20 g of total cell protein obtained from cells either maintained in suspension or adherent to laminin were electrophoresed on an 8% SDS-polyacrylamide gel. Both the phosphorylated and non-phosphorylated forms of S6K were detected by immunoblotting using an S6Kspecific polyclonal antibody (provided by John Blenis).
Dominant Negative Ras Transfections-Transient transfection experiments were performed with LipofectAMINE (Life Technologies, Inc.) according to the manufacturer's instructions. Each 100-mm dish was co-transfected with 2 g of pCDNA3-HA-ERK1 (John Blenis) and 6 g of either pMT3-RasN17 (Larry Feig, Tufts School of Medicine) or empty vector. After 40 h, the cells were removed from the dishes and either maintained in suspension or plated on laminin-1 for 30 min as described above. Cell extracts (200 g of protein) were immunprecipitated with an HA antibody (12CA5; Boehringer Mannheim). After washing three times with lysis buffer, the immune complexes were resuspended in 50 l of kinase buffer (25 mM HEPES, pH 7.4, 10 mM MgCl 2 , 1 mM dithiothreitol) containing 50 M ATP, 4 g of myelin basic protein (MBP), and 5 Ci of [␥-32 P]ATP. The samples were incubated at 30°C for 30 min and the reactions were terminated by the addition of sample buffer and boiling for 5 min. Subsequently, the samples were resolved by SDS-PAGE and the phosphorylated MBP was detected by autoradiography.
Migration Assays-Cell migration assays were performed using 6.5-mm Transwell chambers (8-mm pore size) (CoStar, Cambridge, MA) as described previously (20). Briefly, RPMI-H containing 15 g/ml laminin was added to the bottom well, and the filters were coated for 1 h at 37°C. Cells were resuspended at 10 6 cells/ml in RPMI-H, and 10 5 cells were added to the top well of the Transwell chambers. After a 24-h incubation in the presence of PD98059 (25 M) in Me 2 SO or Me 2 SO alone, the cells that had migrated to the lower surface of the filters were fixed, stained, and counted as described (20).

RESULTS AND DISCUSSION
The ␣ 6A ␤ 1 and ␣ 6B ␤ 1 Integrins Differentially Induce MAP Kinase Activation in P388D1 Cells-To determine whether the ␣ 6 ␤ 1 integrin can induce MAP kinase activation and whether the cytoplasmic domain of the ␣ subunit plays a role in this event, we used stable subclones of P388D1 macrophages that expressed either the ␣ 6A ␤ 1 or ␣ 6B ␤ 1 integrins at equivalent levels of surface expression (16,20). P388D1 cells do not express the ␣ 6 integrin subunit, and expression of either the ␣ 6A ␤ 1 or ␣ 6B ␤ 1 integrin enables them to adhere to laminin-1 (20). Activation of ERK1 and ERK2 was assessed initially by a mobility shift assay using cells that were maintained in suspension or plated on laminin-1-coated dishes for 40 min. In ␣ 6A ␤ 1 -expressing cells plated on laminin-1, the phosphorylated forms of both ERK1 and ERK2 increased markedly in comparison to cells maintained in suspension (Fig. 1A, upper panel). Under identical conditions, significantly less ERK1 and ERK2 were shifted to the slower mobility species in ␣ 6B ␤ 1 -expressing cells. Densitometric analysis of data obtained from three separate experiments revealed that the ratio of the increase in ERK2 phosphorylation induced by laminin-1 in the ␣ 6A ␤ 1expressing cells compared with the increase in the ␣ 6B ␤ 1 -expressing cells was 6.0 Ϯ 1.5.
The mobility shift data were confirmed using an antibody that recognizes the phosphorylated form of the MAP kinases (Fig. 1A, middle panel). A significant amount of phosphorylated ERK2 was detected in the ␣ 6A ␤ 1 -expressing cells adherent to laminin-1 but not in the ␣ 6B ␤ 1 -expressing cells. Adhesion to fibronectin, however, induced ERK1 and ERK2 activation in The same membrane was stripped, and phosphorylated ERK1 and ERK2 were detected using a phospho-MAP kinase specific Ab (middle panel). MAP kinase activity was assayed using an Elk1 fusion protein as the substrate in an immune complex kinase assay (lower panel). Phosphorylated Elk1 was detected by immunoblotting with a phospho-Elk1 specific Ab. The negative control (Ϫ) represents an immune complex kinase assay performed using nonspecific IgG. The positive control represents an immune complex kinase assay performed using 20 ng of purified phospho-MAP kinase. B, MAP kinase activation induced by Ab clustering of the ␣ 6 ␤ 1 integrins. The ␣ 6A ␤ 1 or ␣ 6B ␤ 1 integrins were clustered on the surface of P388D1 cells using an ␣ 6 -specific mAb, 2B7. As a control, cells were incubated in the presence of a nonspecific mouse IgG. Cell extracts (15 g of protein) were resolved by SDS-PAGE and immunoblotted using an anti-ERK mAb (Transduction Laboratories), which recognizes ERK2 predominantly. C, time course of MAP kinase activation induced by ␣ 6A ␤ 1 -mediated adhesion to laminin-1. P388D1 cells that expressed the ␣ 6A ␤ 1 integrin were maintained in suspension for 30 min and then plated on laminin-1-coated dishes for the times indicated on the x axis. At these time points, the cells were extracted and aliquots of the extracts were resolved by SDS-PAGE and immunoblotted with an ERK mAb. ERK2 bands were analyzed by densitometry, and the ratio of phosphorylated to unphosphorylated ERK2 was plotted against time.
The activity of MAP kinase was assayed using an immune complex kinase assay with Elk1 as the substrate. Densitometric analysis of the blot shown in Fig. 1A (lower panel) revealed that MAP kinase activity in the ␣ 6A ␤ 1 -expressing cells adherent to laminin was approximately 4-fold higher than the activity in the same cells maintained in suspension. In contrast, an increase in Elk1 phosphorylation induced by adhesion of the ␣ 6B ␤ 1 -expressing cells adherent to laminin was barely detectable (0.5-fold increase).
Although the ␣ 6 ␤ 1 integrin is the only laminin-1 receptor used by P388D1 macrophages (20), we wished to confirm that the activation of MAP kinase observed in response to laminin adhesion was actually the result of ␣ 6 ␤ 1 ligation. For this purpose, the ␣ 6A ␤ 1 and ␣ 6B ␤ 1 integrins were clustered on the surface of the P388D1 transfectants using the 2B7 mAb and a secondary Ab that recognizes 2B7. It is important to note that 2B7 recognizes the extracellular domain of the ␣ 6 subunit that is identical in the ␣ 6A ␤ 1 and ␣ 6B ␤ 1 integrins (21). Ligation of the ␣ 6A ␤ 1 integrin induced a more marked increase in the appearance of the phosphorylated form of ERK2 in comparison to ligation of ␣ 6B ␤ 1 (Fig. 1B). Thus, these data substantiate the laminin adhesion data presented above.
To determine the kinetics of MAP kinase activation induced by the ␣ 6A ␤ 1 integrin, we examined the mobility shift of ERK1 and ERK2 at different times after plating the cells on laminin-1 (Fig. 1C). Significant activation of both ERK1 and ERK2 was detected at 10 min. Activation peaked at 40 min and was sustained for an additional 30 min. These results are consistent with a sustained activation of MAP kinase that occurs in response to fibronectin adhesion, and they are in contrast to the more rapid and transient activation induced by growth factors (8).
The above data obtained using P388D1 macrophages demonstrate that the interaction of the ␣ 6 ␤ 1 integrin with laminin can trigger the activation of the ERK1 and ERK2 family MAP kinases. More importantly, these data implicate a key role for the cytoplasmic domain of the integrin ␣ subunit in regulating MAP kinase activation because a substantial fraction of total ERK1 and ERK2 was activated by ligation of the ␣ 6A ␤ 1 integrin. Significant activation of MAP kinase by ligation of ␣ 6B ␤ 1 was not observed in any of our assays. Clearly, the results we obtained differ from the recent report that neither ligation of the ␣ 6 ␤ 1 integrin nor cell adhesion to laminin can induce MAP kinase activation (14). A likely explanation for this discrepancy is that integrin-mediated regulation of MAP kinase activity is cell-type specific and that it is not activated by ␣ 6 ␤ 1 ligation in the NIH 3T3 cells used by Wary et al. (14). However, it should be noted that MAP kinase activation in response to NIH 3T3 cell adhesion to laminin has been reported previously (6).
Both the ␣ 6A ␤ 1 and ␣ 6B ␤ 1 Integrins Induce p70 S6 Kinase Activation in P388D1 Macrophages-The issue of whether the ␣ 6B ␤ 1 integrin is capable of activating other kinases in P388D1 cells was addressed by examining the phosphorylation of the p70 S6K, a Ser/Thr kinase that phosphorylates the 40 S ribosomal subunit protein S6 (22,23). We chose S6K because its activation has been shown to be independent of the MAP kinase pathway (24) and because it can be activated by integrin signaling (25). The ability of the ␣ 6A ␤ 1 -and ␣ 6B ␤ 1 -expressing cells to activate S6K in response to laminin adhesion was examined by a mobility shift assay. Extracts from cells that were maintained in suspension or adherent to laminin-1 were immunoblotted with a polyclonal antibody that recognizes both the phosphorylated and non-phosphorylated forms of S6K. Adhesion to laminin-1 induced a dramatic shift to the more slowly migrating, phosphorylated forms of S6K in both the ␣ 6A ␤ 1 -and ␣ 6B ␤ 1 -expressing cells ( Fig. 2A), indicating that both the ␣ 6A ␤ 1 and ␣ 6B ␤ 1 integrins can activate S6K in these cells. Interestingly, S6K activation was completely inhibited by 100 nM wortmannin ( Fig. 2A), but MAP kinase activation was affected minimally by this inhibitor (Fig. 2B). These findings demonstrate that the ␣ 6B ␤ 1 integrin is capable of signaling.
␣ 6A ␤ 1 Integrin-mediated MAP Kinase Activation Depends on the Activity of Ras and PKC-To investigate the role of the small G protein Ras in the ␣ 6 ␤ 1 -mediated MAP kinase activation, ␣ 6A ␤ 1 -expressing P388D1 cells were co-transfected with cDNAs that encode a hemagglutinin epitope-tagged ERK1 (HA-ERK1) and a dominant negative form of Ras (RasN17) that blocks guanine nucleotide exchange factors involved in the activation of endogenous Ras (26). Cells that expressed these constructs were either maintained in suspension or plated on laminin-1 for 30 min and HA-immune complexes obtained from extracts of these cells were assayed for their ability to phosphorylate MBP. As shown in Fig. 3A, MBP phosphorylation in response to laminin-1 attachment was significantly inhibited in cells that expressed RasN17. A RasN17-dependent reduction in the basal level of MBP phosphorylation of cells in suspension was also observed (Fig. 3A). Although these findings suggest that ␣ 6 ␤ 1 -stimulated MAP kinase activation is through the canonical Ras pathway in these cells, it should be noted that the involvement of Ras, as well as FAK, in integrin-mediated MAP kinase activation remains controversial in other cells (9 -13). At the very least, we can conclude that FAK is not important for ␣ 6 ␤ 1 -mediated activation of MAP kinase in P388D1 cells because these cells, similar to other macrophages, do not express this kinase. 2 We also examined the possible role of PKC in the ␣ 6 ␤ 1mediated MAP kinase activation because of its involvement in MAP kinase activation by growth factor receptors (28). The involvement of PKC in MAP kinase activation has been implicated largely from the activating effect of phorbol 12-myristate 13-acetate on these kinases (28). More recently, it was shown that overexpression of certain PKC isoforms can activate MEK1 as well as ERK directly, suggesting that these PKC isoforms could be authentic upstream activators of the MEK-ERK pathway (29). To examine the involvement of PKC, cells were pretreated with the PKC inhibitor calphostin C and either maintained in suspension or plated on laminin-1 in the presence of 0.5 mM MnCl 2 . This divalent cation was included to maintain cell attachment to laminin-1 (20). Mobility shift assay revealed that the ␣ 6A ␤ 1 -induced phosphorylation of ERK1 and ERK2 was specifically inhibited by calphostin C with halfmaximal inhibition observed at approximately 25 nM (Fig. 4). Thus, PKC is a likely intermediate in the ␣ 6 ␤ 1 -mediated activation of MAP kinase in P388D1 cells.
MAP Kinase Activation Is Required for P388D1 Migration on Laminin-We reported previously that ␣ 6A ␤ 1 -expressing P388D1 cells migrate markedly better on laminin-1 than the ␣ 6B ␤ 1 cells (20). Indeed, this finding has been substantiated by other studies that have linked the ␣ 6A ␤ 1 integrin to cell migration (30,31). For this reason, we examined the possibility that the differential activation of MAP kinase by the ␣ 6A ␤ 1 and ␣ 6B ␤ 1 integrins is linked to the ability of these integrins to mediate cell migration on laminin-1. For this purpose, migration assays were performed in the presence of the MEK1 in-hibitor PD98059. This inhibitor blocked the ␣ 6A ␤ 1 -induced phosphorylation of both ERK1 and ERK2 ( Fig. 4A) but had no effect on S6K activation (data not shown). The ␣ 6A ␤ 1 -expressing cells were approximately 2.5-fold more migratory on laminin-1 than the ␣ 6B ␤ 1 cells (Fig. 4B) consistent with our previous results (20). PD98059 inhibited the migration of ␣ 6A ␤ 1 cells by 60% at a concentration range of 12.5-25 M. The slower migration of the ␣ 6B ␤ 1 -expressing cells was also inhibited slightly by PD98059, suggesting that the haptotactic migration of these cells is dependent on MAP kinase and that the differential activation of MAP kinase by the ␣ 6A ␤ 1 and ␣ 6B ␤ 1 integrins regulates this migration. The enhanced ability of the ␣ 6A ␤ 1 -expressing cells to migrate on laminin-1 was not observed on other matrix proteins (20) indicating that a specific, ligand-dependent activation of MAP kinase is required. These results are interesting in view of the recent report that MAP kinase can regulate cell migration by its ability to phosphorylate myosin light chain kinase (5). For this reason, it will be informative to determine if the ␣ 6A ␤ 1 and ␣ 6B ␤ 1 integrins differ in their ability to induce the phosphorylation of this kinase.
In summary, our data demonstrate that the cytoplasmic domain sequence of the ␣ 6 integrin subunit regulates the ability of the ␣ 6 ␤ 1 integrin to activate MAP kinase. This finding highlights the importance of integrin ␣ subunits in regulating downstream signaling events. Moreover, our finding that differential MAP kinase activation by the ␣ 6 ␤ 1 integrin isoforms is linked to the ability of P388D1 cells to migrate on laminin-1 could have widespread functional implications because the preferential expression of the ␣ 6A ␤ 1 over the ␣ 6B ␤ 1 integrin has been associated with cell migration in several systems (20,30,31).