An Eph receptor regulates integrin activity through R-Ras

Abstract

The ability of integrins to mediate cell attachment to extracellular matrices and to blood proteins is regulated from inside the cell. Increased ligand-binding activity of integrins is critical for platelet aggregation upon blood clotting and for leukocyte extravasation to inflamed tissues. Decreased adhesion is thought to promote tumor cell invasion. R-Ras, a small intracellular GTPase, regulates the binding of integrins to their ligands outside the cell. Here we show that the Eph receptor tyrosine kinase, EphB2, can control integrin activity through R-Ras. Cells in which EphB2 is activated become poorly adherent to substrates coated with integrin ligands, and a tyrosine residue in the R-Ras effector domain is phosphorylated. The R-Ras phosphorylation and loss of cell adhesion are causally related, because forced expression of an R-Ras variant resistant to phosphorylation at the critical site made cells unresponsive to the anti-adhesive effect of EphB2. This is an unusual regulatory pathway among the small GTPases. Reduced adhesiveness induced through the Eph/R-Ras pathway may explain the repulsive effect of the Eph receptors in axonal pathfinding and may facilitate tumor cell invasion and angiogenesis.

Integrins mediate cell adhesion to extracellular matrices and, in some cases, to other cells. They are heterodimeric transmembrane proteins that connect to the cytoskeleton and various cytoplasmic signaling molecules within cells. Through these interactions, integrins control cytoskeletal organization and transmit chemical signals into cells (1, 2). Cells control the ability of their integrins to bind to extracellular ligands. Thus, certain cells can convert from nonadherent to highly adherent cells through integrin activation. This happens in the blood, when platelets are activated and aggregate to initiate blood clotting, and when leukocytes adhere to the vascular endothelium at inflammatory sites (1–3). The degree of activation of integrins in adherent cells can also vary (4), and general lowering of adhesiveness is thought to be important in allowing tumor cells to dislodge from their original site and invade (5).

We describe here a signaling pathway controlling integrin activity. We show that an Eph receptor can regulate integrins by modulating the activity of R-Ras. Eph receptors are transmembrane receptors that consists of a ligand-binding extracellular domain and a cytoplasmic tyrosine kinase domain (6, 7). When they bind to their ligands, the ephrins, which are membrane-associated proteins, the kinase domain is activated and phosphorylates a number of cytoplasmic substrate proteins. Our results show that one of these proteins is R-Ras, a small GTPase that has previously been shown to be necessary for integrin activity (8, 9). The phosphorylated R-Ras no longer supports integrin activity, causing a loss of cell–extracellular matrix adhesion of the cells in which the Eph receptor is activated.

Materials and Methods

Cell Lines and Plasmids.

An NIH 3T3 cell line stably transfected with chicken EphB2 cDNA was generously provided by Rüdiger Klein. The EphB2 expression plasmid, pcDNA3-EphB2, has been described (10). The plasmid pcDNA3-Myc-R-Ras was obtained by inserting the Myc epitope sequence at the 5′ end or the 3′ end of the R-Ras sequence. Both Myc-tagged forms of R-Ras produced similar results. Site-directed mutagenesis of EphB2 and Myc-R-Ras was performed by overlapping PCR with appropriate mutated primers. The plasmids for expression of GST-R-Ras, GST-H-Ras, and GST-R-RasY66E (GST, glutathione S-transferase) in Escherichia coli were constructed by cloning PCR products encoding full-length R-Ras or H-Ras in the vector pGEX-4T1 (Pharmacia). The pGEX plasmid encoding GST-RBD contains the cDNA encoding amino acids 51–131 of Raf-1.

Cell Transfections.

Transient transfections were performed with 293T cells. Cells were transfected in 10-cm plates at 60% confluency with 5–10 μg of each plasmid DNA by using Superfect Transfection Reagent (Qiagen) or, for cell attachment assays, with N-[1-(2,3-dioleoyloxy)propyl]-N,N,N-trimethylammonium methylsulfate (DOTAP) transfection reagents (Boehringer Mannheim). The total amount of transfected DNA was kept constant by addition of pcDNA3. To examine cell morphology, the transfected cells were marked by cotransfection with 1 μg of the enhanced green fluorescent protein vector pEGFP (CLONTECH). For ligand stimulation of EphB2, ephrin-B1 Fc (11) was preclustered with a 1:10 molar ratio of anti-Fc antibodies and added to cells at 2 μg/ml for 15 min.

Immunoprecipitation and Immunoblot Analysis.

Cell extracts were prepared in lysis buffer [50 mM Hepes, pH 7.6/1% Triton X-100/0.1% SDS/1% sodium deoxycholate/2 mM EDTA/2 mM EGTA/10% (vol/vol) glycerol/150 mM NaCl/50 mM NaF]. For immunoprecipitations, cell lysates were first incubated with antibodies at 4°C for 1.5 hr, then incubated with GammaBind Sepharose beads for another 1.5 hr at 4°C. For GST pull-down assays, GST fusion proteins immobilized on glutathione-agarose beads were incubated with cell lysates for 1 hr at 4°C. Immunoprecipitates and proteins associated with GST fusion proteins were boiled in SDS-containing sample buffer, separated by SDS/PAGE, and transferred to Immobilon membranes (Millipore). An anti-phosphotyrosine antibody conjugated to horseradish peroxidase (PY-20H; Transduction Laboratories) was used for immunoblotting at a dilution of 1:2,000. An anti-Myc antibody (9E10; Santa Cruz Biotechnology) was used at a dilution of 1:1,000, followed by a secondary goat anti-mouse IgG antibody conjugated to horseradish peroxidase. Polyclonal anti-EphB2 antibodies (12) were used at a concentration of 3 μg/ml, followed by staphylococcal protein A conjugated to horseradish peroxidase (Sigma). Immunoblots were developed by enhanced chemiluminescence (Amersham).

In Vitro Kinase Assays.

EphB2 immunoprecipitates were split into aliquots and washed with a kinase buffer consisting of 50 mM Hepes, pH 7.6, 20 mM MnCl₂, 0.25 mM Na₃VO₄, 10 μg/ml leupeptin, 5 μg/ml aprotinin, and 2.5 mM PMSF. Kinase reactions were carried out for 15 min at room temperature in kinase buffer containing 1 μCi (1 μCi = 1 kBq) of [γ-³²P]ATP. Reactions were stopped by adding SDS sample buffer.

Mass Spectrometry.

293 cells transiently cotransfected with EphB2 and Myc-tagged R-Ras (or a vector containing only the Myc tag as a control) were lysed in RIPA buffer (0.15 mM NaCl/0.05 mM Tris⋅HCl, pH 7.2/1% Triton X-100/1% sodium deoxycholate/0.1% SDS) containing protease inhibitors, 1 mM Na₃VO₄, 10 mM NaF, and 5 mM EDTA. Myc tag antibody immunoprecipitates were washed in 0.1% 1-O-octyl β-d-glucopyranoside (98%; Aldrich)/100 mM NH₄HCO₃, pH 7.5/1 mM Na₃VO₄ and digested for 24 hr at 37°C with sequence grade endoproteinase Asp-N (20 ng/μl; Sigma) alone or endoproteinase Asp-N followed by overnight digestion with endoproteinase Glu-C (50 ng/μl; Promega) in 100 μl of 100 mM NH₄HCO₃, pH 8. Endoproteinases were inactivated by boiling for 5 min before inhibiting further with PMSF (100 μg/ml) and leupeptin (2 μg/ml). Peptides phosphorylated on tyrosine were isolated as described before (13) on anti-phosphotyrosine antibodies conjugated to agarose (clone PT-66; Sigma). Mass spectra were collected on a Voyager matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) mass spectrometer (PerSeptive Biosystems) as described previously (13), and the bound phosphopeptide mixtures were analyzed in the delayed extraction, reflector mode. External mass calibrations were performed with angiotensin I (MH⁺ = 1296.6853) and adrenocorticotropic hormone fragment (amino acids 7–38; MH⁺ = 3657.93). Phosphopeptides were identified from a list of predicted molecular masses generated from theoretical cleavage with specific endoproteinases by using the MS-Digest program (http://prospector.ucsf.edu). The observed masses of the phosphopeptides in the mass spectra were matched to the calculated masses from the theoretical digest with an accuracy of ≈0.1% or better.

Cell Attachment Assays.

Non-tissue-culture 96-well plates (Linbro/Titertek) were precoated with 10 μg/ml fibronectin (Chemicon), 5 μg/ml collagen I (Collaborative Biomedical Products), or BSA overnight at 4°C. The wells were then blocked with 1% BSA. Forty-eight hours after transfection, the cells were detached with 2.5 mM EDTA, washed with serum-free medium, and added to the wells (5 × 10⁴ cells per well). After incubation at 37°C for 1 hr, the wells were washed with serum-free medium and attached cells were fixed and stained with 0.5% crystal violet in methanol/water. The crystal violet incorporated into cells was collected in 1% SDS and quantified by measuring A₅₉₀. For ligand stimulation experiments, the EphB2-stably transfected NIH 3T3 cells were treated for 5 min with 2 μg/ml preclustered ephrin-B1 Fc, before seeding into the wells (10⁵ cells per well).

Results and Discussion

Inhibition of Cell Adhesion by EphB2.

Forced expression of the EphB2 receptor in 293T human epithelial kidney cells at high levels results in activation of the transfected receptor (Fig. 1A). The cells expressing the activated receptor partially detached from the substrate and rounded up (Fig. 1B). In contrast, cells transfected with a kinase-inactive mutant, EphB2K662R, or empty vector (Fig. 1A) remained well spread and spindle-shaped (Fig. 1B). Cells in which the activation of the EphB2 receptor required treatment with an ephrin ligand (Fig. 1C) showed reduced cell substrate adhesion upon treatment with chimeric ephrin-B1 Fc (Fig. 1D). Activation of an Eph receptor (EphB1) by ligand presented as part of the substrate can increase substrate adhesion (14). However, in our hands, EphB2 (and EphB3; H. Miao and B.W., unpublished results), whether activated upon overexpression or by ligand, consistently reduces cell adhesion.

Figure 1.

Activation of the EphB2 receptor and its effects on cell adhesion. (A) Activation of EphB2 in transiently transfected 293 cells. EphB2 was immunoprecipitated from cells transfected with EphB2, kinase-inactive EphB2K669, or pcDNA3. The immunoprecipitates (IP) were probed by immunoblotting as indicated. (B) Morphological changes caused by expression of activated EphB2 in 293T cells. An enhanced green fluorescent protein (EGFP) vector was cotransfected to identify the transfected cells. The transfected cells were shown to be alive by trypan blue exclusion. EphB2K662R and pcDNA3 vector were used as controls. (Upper) Fluorescence microscopy. (Lower) Phase-contrast microscopy of same fields. (×200.) (C) Activation of EphB2 by ephrin-B1 ligand. EphB2 was immunoprecipitated from NIH 3T3 cells that had been stably transfected with EphB2 and treated with soluble ephrin-B1 Fc ligand or left untreated. The immunoprecipitates were probed by immunoblotting as indicated. (D) Ligand stimulation of EphB2 decreases adhesion to collagen I in NIH 3T3 cells stably transfected with EphB2.

R-Ras Phosphorylation.

When the activated EphB2 receptor was coexpressed with R-Ras, R-Ras became phosphorylated on tyrosine, whereas the kinase-inactive EphB2K662R did not cause R-Ras tyrosine-phosphorylation (Fig. 2 A and B). Our earlier results have shown that the activity of endogenous R-Ras is required for cells to maintain their ability to attach to extracellular matrix substrates through integrins (8). We have, therefore, explored the possibility that EphB2 would directly phosphorylate R-Ras, and that this would interfere with the ability of R-Ras to support integrin-mediated cell adhesion. In in vitro kinase reactions, immunoprecipitated EphB2 phosphorylated GST-R-Ras but not GST-H-Ras or GST (Fig. 2C). Furthermore, the tyrosine-phosphorylated form of R-Ras did not bind to an immobilized GST fusion protein containing the Ras-binding domain of Raf-1 (GST-RBD) (Fig. 2D). Because Raf-1 is a downstream effector of Ras proteins (15–17) to which R-Ras also binds (18), this result suggests that tyrosine phosphorylation by EphB2 affects the functional properties of R-Ras.

Figure 2.

R-Ras phosphorylation on tyrosine by EphB2 impairs binding to Raf-1. (A) Phosphorylation of R-Ras in cells expressing activated EphB2. Extracts of 293T cells cotransfected with EphB2, or EphB2K662R, and Myc-tagged R-Ras were subjected to immunoprecipitation (IP) with anti-Myc antibody. The immunoprecipitates, and cell lysates, were probed by immunoblotting as indicated. Anti-PY, anti-phosphotyrosine. (B) The labeling of R-Ras with anti-phosphotyrosine antibody is specific. R-Ras immunoprecipitated with anti-Myc antibody was probed with anti-phosphotyrosine antibody, or this antibody together with 1 mM phosphotyrosine, phosphoserine, or phosphothreonine. Only phosphotyrosine blocked antibody labeling of R-Ras. (C) Phosphorylation of R-Ras by EphB2 in vitro. Equal amounts of purified GST-R-Ras, GST-H-Ras, or GST alone were incubated with equal amounts of immunoprecipitated EphB2 in the presence of [γ-³²P]ATP. Phosphorylation of R-Ras was analyzed by SDS/PAGE followed by autoradiography. The lower molecular weight phosphorylated band in the R-Ras lane is presumably a degradation product of GST-R-Ras. Coomassie blue protein staining shows the amount of each GST protein used in the assay. The arrows point to the Ras proteins. (D) Tyrosine-phosphorylated R-Ras lacks the ability to bind to Raf-1. Extracts of cells transfected with Myc-tagged R-Ras and EphB2 were incubated with the immobilized GST-Ras binding domain of Raf-1 (GST-RBD). Cell extracts and the supernatant of the GST-RBD incubation were immunoprecipitated with anti-Myc antibody. Immunoprecipitates (lanes 1 and 2) and proteins bound to GST-RBD (lane 3) were analyzed by immunoblotting as indicated. The large amount of GST-RBD in lane 3 interfered with the migration of the R-Ras band, making it appear of lower molecular weight than in the other lanes.

Because the interaction with Raf-1 is mediated by the effector domain of Ras proteins (15–17), the loss of Raf-1 binding activity suggested that the tyrosine residue phosphorylated by EphB2 is located in the effector domain of R-Ras. Among the tyrosines in R-Ras, the tyrosine residue at position 66 (Y66) is in the effector domain and in a sequence context (EDSYT) that is favorable for phosphorylation by certain tyrosine kinases (19). We engineered an R-Ras mutant in which Y66 was converted to glutamic acid (R-RasY66E). This introduces a nonconservative substitution that may additionally mimic the effects of the negative charge in phosphotyrosine (20, 21). To confirm that Y66 is a phosphorylation site of EphB2 on R-Ras, wild-type and R-RasY66E were expressed as GST fusion proteins and tested in in vitro kinase assays. Immunoprecipitated EphB2 phosphorylated GST-R-Ras, but not GST-R-RasY66E (Fig. 3A). The phosphorylation of R-Ras by EphB2 was further characterized in living cells. Myc-tagged R-Ras cotransfected with EphB2 became phosphorylated on tyrosine, whereas phosphorylation of the Y66E mutant and of a mutant in which Y66 was converted to phenylalanine (R-RasY66F) was markedly diminished (Fig. 3B). This result suggests that Y66 of R-Ras is a major, but not the only, EphB2 phosphorylation site in cells.

Figure 3.

Tyrosine 66 of R-Ras is phosphorylated by EphB2 and is important for binding to Raf-1 but not for association with EphB2. (A) EphB2 phosphorylates R-Ras but not R-RasY66E in vitro. In vitro kinase reactions were carried out as described in the legend of Fig. 2. The faint phosphorylated band in the GST-R-RasY66E (YE) and GST lanes presumably represents phosphorylated IgG heavy chain. The amount of GST-R-Ras and GST-R-RasY66E proteins used for the assay is shown by Coomassie blue staining (Lower). (B) R-Ras phosphorylation in EphB2-transfected cells is Y66 dependent. EphB2 was cotransfected with Myc-tagged R-Ras or R-Ras variants into 293T cells. Equal amounts of cell lysates were immunoprecipitated with anti-Myc antibody. and the immunoprecipitates were probed by immunoblotting as indicated. R-RasY66F protein (YF), and to a lesser extent R-RasY66E (YE), migrated faster than the wild-type R-Ras (Wt), consistent with a lower level of phosphorylation.

The phosphorylation of Y66 in the EphB2-transfected cells was confirmed by chemical analysis of phosphopeptides from R-Ras immunoprecipitates. After digestion with the protease Asp-N, mass spectrometry revealed a peak corresponding to a monophosphorylated peptide containing Y66 (Fig. 4A). This peak was not present in samples that were additionally digested with a second protease, and new peaks with masses that were compatible with peptides containing the appropriate cleavages and phosphorylated Y66 appeared (Fig. 4B).

Figure 4.

Matrix-assisted laser desorption ionization (MALDI) mass spectra of R-Ras phosphopeptides. Tyrosine-phosphorylated peptides resulting from digestion with endoproteinase Asp-N alone (A) or endoproteinase Asp-N followed by endoproteinase Glu-C (B) were isolated from the peptide mixtures by using anti-phosphotyrosine antibodies conjugated to agarose and analyzed directly by MALDI mass spectrometry. Peaks representing peptides phosphorylated at Y66 are indicated by their respective masses and sequences. Peaks marked B are from background ions detected in control samples and unmarked peaks are unidentified. The peptide at m/z 1836.16 (Ox) has a mass consistent with its being an oxidized form of that at m/z 1819.21.

EphB2 and R-Ras Associate with One Another in Cells.

Wild-type EphB2 was detected in R-Ras immunoprecipitates, but no interaction was seen between kinase-inactive EphB2 and R-Ras (Fig. 5A). Similar results were obtained by probing EphB2 immunoprecipitates for the presence of R-Ras (not shown). Thus, the interaction of EphB2 with R-Ras in cells is specific and depends on the kinase activity or phosphorylation status of the receptor. This suggests that a protein with an SH2 (Src homology 2) or a PTB (phosphotyrosine-binding) domain mediates the interaction between EphB2 and R-Ras. Mutation of Y66 did not prevent the interaction of R-Ras with EphB2 (Fig. 5A). The Y66E mutation, however, prevented binding to Raf-1 (Fig. 5B), similar to what was seen upon phosphorylation of R-Ras. This is consistent with the observation that mutations in the effector domain of Ras proteins can alter their functions (15–17, 22). It should be noted that the R-RasY66E mutant bound GTP to a similar extent as wild-type R-Ras (not shown). It is therefore conceivable that a modification of Y66, such as phosphorylation, may selectively inactivate signaling pathways downstream of GTP-bound (activated) R-Ras that depend on Y66. The Y66F mutation did not prevent binding of R-Ras to Raf-1 (Fig. 5B), possibly because phenylalanine is a conservative substitution for nonphosphorylated tyrosine.

Figure 5.

Association of EphB2 with R-Ras in cells. (A) EphB2 or EphB2K662R was cotransfected with wild-type or mutated Myc-tagged R-Ras, or pcDNA3. Equal amounts of cell lysates were immunoprecipitated with anti-Myc antibody. The immunoprecipitates and cell lysates were probed by immunoblotting with the antibodies indicated. (B) Raf-1 binding by R-RasY66 mutants. Extracts of cells transfected with wild-type R-Ras, R-RasY66F, and R-RasY66E were incubated with the immobilized GST fusion protein of the Ras-binding domain of Raf-1 (GST-RBD) or GST as a control. Bound proteins and cell extracts were analyzed by immunoblotting with anti-Myc antibody. The amounts of GST-RBD and GST were examined by Coomassie blue staining.

R-Ras Y66F Mutant Counteracts the EphB2 Effect.

To test whether tyrosine-phosphorylation of R-Ras by EphB2 might contribute to the decreased cell adhesion, we performed cell attachment assays with cells transfected with EphB2 and various R-Ras mutants. The adhesiveness of 293 cells expressing EphB2 alone was reduced 3- to 4-fold on fibronectin and collagen (Fig. 6A). As the transfection efficiency in the 293 cells is about 70% (not shown), the level of reduction in adhesion that we have achieved with EphB2 indicates that the transfected cells have become essentially nonadhesive. Transfecting R-RasY66E alone decreased the adhesiveness of these cells 2- to 3-fold, whereas wild-type R-Ras, or R-Ras66YF, had no effect on adhesion (Fig. 6A). The adhesion of the 293 cells to fibronectin and collagen was mediated by integrins, because function-perturbing antibodies against the β1 and αv integrin subunits inhibited it (not shown).

Figure 6.

Phosphorylation of R-Ras mediates the effects of EphB2 on cell adhesion. (A) Effects of R-RasY66 mutants on cell adhesion. The 293T cells were transfected with various R-Ras vectors, together with EphB2, and plated on surfaces coated with fibronectin, collagen I, or BSA as a control. (B) R-Ras38V Y66F suppresses the effect of EphB2 on cell adhesion. The constitutively activated R-Ras38V and its Y66 mutants were cotransfected with EphB2 in 293T cells, and cell attachment was quantitated.

The results described above suggest that EphB2 inhibits cell adhesion through phosphorylation of R-Ras, which suppresses the ability of R-Ras to support integrin activity. R-RasY66E may be a dominant negative capable of suppressing the cell attachment-supporting activity of endogenous R-Ras. If phosphorylation at Y66 inactivates this function of R-Ras, cells expressing an R-Ras mutant that lacks the phosphorylation site but is otherwise active should be resistant to the EphB2 effect. To test this possibility, we used R-Ras38V because the activating 38V mutation makes it more potent than wild-type R-Ras. R-Ras38V cotransfected with EphB2 had no effect on the ability of EphB2 to reduce cell adhesion (Fig. 6B). However, when the R-Ras38V Y66F double mutant was cotransfected with EphB2, there was no decrease in cell adhesion. As expected, cells cotransfected with EphB2 and R-Ras38V Y66E, which contains the inactivating Y66E mutation, were nonadhesive. The neutralization of the EphB2 effect on cell adhesion by R-Ras38V Y66F, which is active and cannot be phosphorylated at position 66, strongly supports the hypothesis that the ability of EphB2 to decrease cell adhesion is mediated by the phosphorylation of R-Ras.

The findings reported here have a number of important implications. First, they reveal a way of regulating R-Ras activity by tyrosine phosphorylation that has not been described previously for R-Ras or for the Ras protein superfamily. Second, our findings may explain the molecular mechanism underlying some of the activities ascribed to the Eph receptors. Activation of Eph receptors causes neurite retraction and steers axons away from incorrect targets (23, 24). Our results suggest that down-regulation of integrin-mediated adhesion could be important in this phenomenon. Growth cone migration depends on integrins, at least in vitro (25), and growth cones that lose their attachment to the extracellular matrix would be likely to retract. Third, the recently discovered role of Eph receptors in vasculogenesis (26, 27) suggests that the Eph-R-Ras connection may also be important in shaping the developing vasculature. Finally, increased expression of Eph receptors in tumors (28, 29) suggests that phosphorylation of R-Ras by Eph receptors, and perhaps by oncogenic kinases such as Src, may also be responsible for the reduced adhesiveness of tumor cells.

Abbreviation

GST

glutathione S-transferase