The Tetraspanin CD151 Functions as a Negative Regulator in the Adhesion-dependent Activation of Ras*

Transmembrane proteins of the tetraspanin superfamily are associated with integrins and are thought to regulate adhesion-dependent signaling. The molecular mechanisms of this regulation remain unknown. We used rat fibroblasts to analyze the contribution of the tetraspanin CD151 in the adhesion-dependent signaling. Expression of CD151 specifically attenuated adhesion-dependent activation of Ras. Furthermore, activation of PKB/c-Akt and ERK1/2, downstream targets in the Ras signaling pathway, was also diminished in cells expressing CD151. In contrast, adhesion-dependent activation of FAK and c-Src were not affected by CD151. The attenuation of Ras signaling did not correlate with phosphorylation of Tyr 925 -FAK, tyrosine phosphorylation of Shc, or with assembly of the p120RasGAP-p62Dok complex. Using mutants of CD151 we established that the cytoplasmic C-terminal portion is critical for activity of CD151 toward Ras. Taken together these results identify CD151 as a negative regulator of Ras and suggest a novel mechanism of adhesion-dependent regulation of Ras activity. The four transmembrane domain proteins of the tetraspanin superfamily are involved in regulation of various biological Although numerous reports have established that tetraspanins are involved in cell migration their pGEX-2TK- Raf1(RBD) plasmids were kindly provided by Dr. M. Schwartz, Dr. J. Collard and Dr. R. Marais, respectively. Proteins solubilized into 1% Triton X-100/PBS supplemented with 2 m M phenylmethylsulfonyl fluoride, 10 (cid:1) g/ml aprotinin, 10 (cid:1) g/ml leupeptin for 2–16 h at 4 °C. The insoluble was pelleted at 12,000 and lysates were precleared by incubation fo r at 4°C with agarose-conjugated goat anti-mouse on agarose prebound mAbs, four successive immunoprecipitation sample Analysis of Activation of analyzed

The four transmembrane domain proteins of the tetraspanin superfamily are involved in regulation of various biological phenomena (1)(2)(3). Although numerous reports have established that tetraspanins are involved in cell migration (1)(2)(3)(4)(5), their role in this process remains unclear. Despite the fact that tetraspanins are associated with various integrins, in most cells they are excluded from the focal adhesions (6) and do not affect adhesion of cells to the extracellular matrix (2,5). Therefore, it has been proposed that tetraspanins may specifically regulate post-ligand binding integrin-dependent signaling (2,5). Indeed, earlier studies showed that tetraspanins could in-fluence adhesion-dependent downstream signaling events, including tyrosine phosphorylation of cellular proteins (7) and activation of the serine-threonine kinase PKB/c-Akt (8). However, it is not known how tetraspanins target these signaling pathways or whether these events are connected with cell migration.
A number of integrin-dependent signaling pathways have been linked to cell motility. These include: association with, and activation of, receptor tyrosine kinases (9); activation of Src-family non-receptor tyrosine kinases (10); and activation of focal adhesion kinase (FAK) 1 (11). All of these signaling events, in turn, trigger a network of parallel and intersecting downstream signaling reactions (12,13). Proteins of the Ras family of small GTPases take one of the central positions in this signaling map, by regulating the activities of Erk1/2 and Rac1 (13). It has been demonstrated that several pathways could be responsible for the adhesion-dependent activation of Ras proteins. Two of them (i.e. the assembly of the phospho-FAK-(Tyr 925 )-Grb2-Sos and phospho-Shc-Grb2-Sos complexes) require the upstream activation of Src kinases (14,15). In addition, it has been reported that cell-ECM adhesion triggers the assembly of the p120RasGAP-p62Dok complex, an event that suppresses the activity of RasGAP, which, consequently, leads to the increase in active, GTP-bound, Ras in cells (16).
Recently, it has been shown that the tetraspanin CD151 is involved in regulation of migration of neutrophils (17), endothelial cells (18), and various tumor cell lines (19,20). Furthermore, it appeared that CD151 could manifest its activity only on the FAK-positive cellular background, suggesting a role for FAK in CD151-dependent migration (19). However, no biochemical signaling events have been ascribed to the activity of CD151.
In this report we investigated in detail how expression of CD151 affects integrin-dependent signaling in non-transformed fibroblasts. We show that CD151 specifically attenuates adhesion-dependent activation of Ras and the corresponding downstream activation of ERK1/2 and PKB/c-Akt. Furthermore, we establish that this attenuation arises from the marked effect of CD151 on the inactivation of Ras proteins upon cell detachment. Finally, detailed analysis of upstream signaling events indicates that CD151 affects the adhesion-dependent Ras activation via a novel mechanism.

MATERIALS AND METHODS
Cells, Antibodies, and Reagents-Rat-1 cells and various transfectants of Rat-1 cells were routinely maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum. The Rat-1/ CD151wt cells were described previously (21). Rat-1/CD151⌬C and Rat-1/CD151⌬N cell lines were generated by transfecting pZeoSV/ CD151⌬C and pZeoSV/CD151⌬N plasmids, respectively, into Rat-1 cells. Zeocin-resistant colonies in each transfection experiment were pooled together (20 -30 individual colonies) and further selected in two cycles of cell sorting using a mixture of anti-CD151 mAbs, 5C11 (22) and 11G1B4 (kindly provided by Dr. L. Ashman). Mouse mAbs against FAK, Ras, Rac1, and p120RasGAP were purchased from BD Biosciences. Mouse mAb to RhoA and rabbit polyclonal Ab to Cdc42 were from Santa Cruz Biotechnology. Rabbit anti-phospho specific Abs to FAK and rabbit anti-Src polyclonal serum were from BioSource International. Rabbit anti-FAK Ab was purchased from Autogen Bioclear. All antibodies to PKB, anti-pTyr mAb, and anti-pSrc 416 were purchased from Cell Sig-naling Technology. Mouse anti-Src mAb, GD11, was purchased from Upstate Biotechnology. All other reagents were purchased from Sigma.
Construction of CD151 Mutants-The CD151⌬N and CD151⌬C mutants were engineered by a standard PCR protocol on the pZeoCD151 template. The N-terminal Gly 2 -Leu 13 and the C-terminal Ser 248 -Tyr 253 regions of human CD151 were substituted for the hemagglutinin tag sequence (YPYDVPDYA) to generate CD151⌬N and CD151⌬C fragments, respectively. The PCR fragments were subcloned (HindIII-EcoRI) into the pZeoSV plasmid.
Immunoprecipitation-Proteins were solubilized into 1% Triton X-100/PBS supplemented with 2 mM phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, 10 g/ml leupeptin for 2-16 h at 4°C. The insoluble material was pelleted at 12,000 rpm for 10 min, and protein lysates were precleared by incubation for 4 h at 4°C with agarose-conjugated goat anti-mouse antibodies. Immune complexes were collected on the agarose beads prebound with mAbs, followed by four successive washes with the immunoprecipitation buffer. Immune complexes were eluted from the beads with Laemmli sample buffer, and proteins were resolved in 8 -10% SDS-PAGE.
Analysis of Activation of c-Src-Serum-starved cells plated on the laminin-5 matrix were scraped into Triton X-100-based lysis buffer, and the c-Src complexes were immunoprecipitated using GD11 mAb. Equal aliquots of the immunoprecipitated proteins were resolved in 8% SDS-PAGE, transferred to the nitrocellulose membrane, and probed either with the anti-Tyr 416 -Src Ab or with the polyclonal Ab recognizing total proteins.

RESULTS AND DISCUSSION
We examined whether CD151 modifies adhesion-dependent signaling using Rat-1 fibroblasts. This model system provided two main advantages for our study. First, these cells express low levels of the endogenous CD151. 2 Second, adhesion-dependent signaling in Rat-1 cells is well characterized (23). Rat-1 cells were transfected with human CD151 cDNA and further selected by bulk sorting to obtain a population of cells expressing the tetraspanin at a level similar to that found in some cancer cell lines (e.g. HT1080, MDA-MB-231). Initially, adhesion-dependent signaling was analyzed in cells plated on laminin-5-containing matrix (LN5M), an ECM ligand for ␣ 3 ␤ 1 , which is a principal integrin partner for CD151 in Rat-1/CD151 cells. The ectopic expression of CD151 in Rat-1 cells did not affect adhesion to, and spreading on, LN5M (Fig. 1, upper  panels). Furthermore, the number and distribution of vinculincontaining adhesion complexes were similar in Rat-1 and Rat-1/CD151 cells (Fig. 1, lower panels). As illustrated in Fig. 2, the expression of CD151 had a negative effect on the kinetics and amplitude of adhesion-dependent activation of ERK1/2 and PKB/c-Akt (Fig. 2, A and B, lower panels, compare lanes 2-4  with 6 -8). Specifically, we found that, in control Rat-1 cells, the level of ERK phosphorylation reached the plateau at 10 min (the earliest time point analyzed). By contrast it required 40 min for the pERK1/2 to reach their maximum in Rat-1/CD151 cells ( Fig. 2A, lower panel, compare lanes 2-5 with 7-10). On the other hand, adhesion of Rat-1 and Rat-1/CD151 cells to LN5M induced comparable increases in tyrosine phosphorylation of Tyr 397 -FAK and Tyr 416 -c-Src (Fig. 2, A and B, top panels). The differences in the adhesion-dependent activation of ERK1/2 were observed when the cells were plated on other ECM ligands (e.g. fibronectin, laminin-10/11) (Fig. 2C, compare  FIG. 1. Ectopic expression of CD151 in Rat-1 cells does not affect cell adhesion and spreading. Upper panels, morphology of serum-starved cells plated on LN5M. Photographs were taken 1 h after plating. Lower panels, assembly of focal adhesions in cells plated on LN5M. Serum-starved cells were plated on the LN5M for 1 h. Cells were fixed with paraformaldehyde and permeabilized with 0.1% Triton X-100. Indirect immunofluorescence staining was carried out using mAb to vinculin, hVIN-1. Staining was visualized using fluorescein isothiocyanate-conjugated goat anti-mouse IgG. FIG. 2. The role of CD151 in adhesion-dependent signaling. A, the role of CD151 in the adhesion-dependent activation of FAK and ERK1/2. Cells were prepared as described in the legend to A. Cellular lysates were resolved in 8% SDS-PAGE. Proteins were transferred to a nitrocellulose membrane and probed with indicated polyclonal Abs. B, the role of CD151 in the adhesion-dependent activation of c-Src and PKB/c-Akt. Serum-starved cells were kept in suspension for 60 min before plating on LN5M. The PKB and c-Src complexes were immunoprecipitated using specific mAbs (5G3-PKB; GD11-cSrc). Cellular lysates and the immunoprecipitated complexes were resolved in 8% SDS-PAGE. Proteins were transferred to a nitrocellulose membrane and probed with indicated polyclonal Abs. C, activation of ERK1/2 in cells plated on FN and LN10/11 was carried out as described in the legend to B. D, serum-starved cells were stimulated with EGF (1 ng/ml) and activation of ERK1/2 was analyzed as above.

Tetraspanin CD151 as a Negative Regulator in Activation of Ras 26324
lanes 2-5 with 7-10). In contrast, the kinetics and degree of ERK1/2 activation were comparable when serum-starved Rat-1 and Rat-1/CD151 cells were stimulated with EGF (Fig. 2D). These results indicated that CD151 specifically attenuated adhesion-dependent activation of ERK1/2 and PKB/c-Akt but had no effect on the EGF-induced signaling.
Small GTP-binding proteins of the Ras family are the best characterized activators of ERK1/2 and PI3K in integrin-dependent signaling (24,25). Hence, we analyzed the role of CD151 in adhesion-dependent activation of Ras. As illustrated in Fig. 3A cell attachment increases the levels of GTP-loaded Ras in both Rat-1 and Rat-1/CD151 cells. However, the level of GTP-Ras in Rat-1 cells was higher than that in Rat-1/CD151 cells (up to ϳ1.8-fold at 20 min after plating to LN5M). On the other hand, LN5M-dependent modulation of Rho-proteins was similar for Rat-1 and Rat-1/CD151 cells (results are not shown). It has been previously reported that integrin-dependent activation of Ras may involve two signaling pathways: one requires Fyn-dependent phosphorylation of Shc and subsequent recruitment of the Grb2-Sos1 complex (15), and another relies on Src-dependent phosphorylation of FAK on tyrosine 925 followed by assembly of the FAK-Grb2-Sos1 signaling complex (14). Although phosphorylation of Tyr 925 -FAK was induced upon cell adhesion, no apparent differences were observed between Rat-1 and Rat-1/CD151 cells (Fig. 3B, lower  panel). Furthermore, we were unable to detect adhesion-induced phosphorylation of Shc above the existing baseline in both cell lines (Fig. 3B, upper panel). Therefore, it is highly unlikely that CD151-induced differences in the adhesion-dependent activation of Ras linked to signaling through Fak or Shc. Rather, we propose that CD151 affects Ras activation through a novel pathway.
We propose that the attenuation of the adhesion-dependent activation of Ras is reflective of the lower level of the GTP-Ras complexes in the detached Rat-1/CD151 cells (Fig. 3A, lanes 1  and 4). Indeed, in separate experiments we found that detach-ment of serum-starved Rat-1/CD151 cells induced a greater decrease in the amount of the precipitated GTP-Ras complexes than detachment of Rat-1 cells (Fig. 4A, lanes 2 and 4). The dissociation of the p120RasGAP/p62-Dok complex is thought to be a critical event in the detachment-induced inactivation of Ras proteins (16). However, while the p120RasGAP/p62-Dok complex dissociated upon detachment of both Rat-1 and Rat-1/ CD151 cells, no quantitative differences were observed between the cell lines (Fig. 4B, lanes 2 and 4). These results suggest that the effect of CD151 on inactivation of Ras involves an alternative mechanism(s).
To determine which region(s) of CD151 is responsible for the adhesion-dependent regulation of Ras activity, we carried out Tetraspanin CD151 as a Negative Regulator in Activation of Ras 26325 cell detachment experiments using the Rat-1 cells expressing various CD151 mutants at the levels similar to that of the wild-type protein. As illustrated in Fig. 5A the deletion of the N-terminal cytoplasmic portion of CD151 did not affect significantly the "Ras-inactivating function" of the tetraspanin. On the other hand, deletion of the C-terminal cytoplasmic portion completely abolished the detachment-dependent activity of CD151 toward Ras (Fig. 5B, lanes 2, 4, and 6). These data indicate that the C-terminal end is essential for the Ras-inactivating function of CD151. Activation of Ras proteins is one of the central events in the attachment-dependent signaling mediated by integrins (12,26). Here we demonstrate that the tetraspanin CD151 plays an important role in this process. Specifically, we identified CD151 as a new negative regulator of Ras proteins, which potentiates the decrease in the amount of the GTP-Ras complex in cells upon detachment. Consequently, adhesion-dependent accumulation of GTP-Ras and subsequent activation of ERK1/2 and PKB is attenuated in cells expressing CD151. Importantly, the effect of CD151 on the adhesion-dependent signaling is specific. Indeed, activation of FAK and c-Src and modulation of the activity of Rho GTPases were comparable in the Rat-1 and Rat-1/CD151 cells. Importantly, CD151 exhibits its activity toward Ras only when integrins are disengaged from their ligands. This function of CD151 may be particularly relevant for cell migration. In migrating cells the dynamic adhesiondeadhesion cycles are taking place at the front edge of extending lamellae. Thus, it is conceivable that CD151 facilitates the GTP⅐Ras u GDP⅐Ras cycle and specifically affects this step of the migratory process.
The molecular mechanisms that link integrin disengagement with the "Ras-inactivating" function of CD151 remain to be established. CD151 is directly associated with ␣ 3 ␤ 1 and ␣ 6 ␤ 1 integrins through its large extracellular loop (27,28). Furthermore, the conformation of LECL CD151 can be influenced by these interactions (29). In addition, CD151 is indirectly associated with other integrins within the tetraspanin-enriched microdomains (30). Thus, it is possible that the initial events in signal transduction involve lateral transmission of conformational changes from the disengaged integrins to CD151. Subsequently, the signal may be transmitted allosterically to the cell interior and further involve the C-terminal cytoplasmic part of CD151.
Our results indicate that detachment-induced disassembly of the RasGAP-Dok complex, the only known pathway that regulates decrease in the amount of the GTP-Ras complex in suspended cells, is not affected by CD151. This suggests that the tetraspanin operates through a novel mechanism. This may involve CD151-dependent membrane recruitment and/or compartmentalization of direct regulators of Ras proteins (e.g. Ras-GAP proteins of the GAP1 family, Ras guanine nucleotide exchange factors (31,32)). In this regard, palmitoylation of CD151 and its association with phosphatidylinositol 4-kinase may play an important role in this process.
Future study should establish whether function of other "pro-" or "anti-migratory" tetraspanins (e.g. CD82 and CO-029) also linked to the Ras pathway. CD151 is associated with various other tetraspanins, each of which may recruit a unique subset of regulatory proteins into integrin complexes (2,5). Hence, by changing a tetraspanin balance in the integrin/ CD151 microdomains it may be possible to amplify, negate, or even reverse the effect of CD151 in the adhesion-dependent activation of Ras.