Oncogene
SEARCH     advanced search my account e-alerts subscribe register
Journal home
Advance online publication
Current issue
Archive
Press releases
For authors
For referees
Contact editorial office
About the journal
For librarians
Subscribe
Advertising
naturereprints
Contact NPG
Customer services
Site features
NPG Subject areas
Access material from all our publications in your subject area:
Biotechnology Biotechnology
Cancer Cancer
Chemistry Chemistry
Dentistry Dentistry
Development Development
Drug Discovery Drug Discovery
Earth Sciences Earth Sciences
Evolution & Ecology Evolution & Ecology
Genetics Genetics
Immunology Immunology
Materials Materials Science
Medical Research Medical Research
Microbiology Microbiology
Molecular Cell Biology Molecular Cell Biology
Neuroscience Neuroscience
Pharmacology Pharmacology
Physics Physics
Browse all publications
 
31 October 2002, Volume 21, Number 50, Pages 7690-7699
Table of contents    Previous  Article  Next   [PDF]
Original Paper
Activation of the EphA2 tyrosine kinase stimulates the MAP/ERK kinase signaling cascade
Rebecca L Pratt1 and Michael S Kinch1,2

1Department of Basic Medical Sciences, Purdue University Cancer Center, West Lafayette, Indiana, IN 47907-1246, USA

2MedImmune, Inc., Gaithersburg, Maryland, MD 20878, USA

Correspondence to: M S Kinch, MedImmune, Inc., 35 West Watkins Mill Road, Gaithersburg, Maryland, MD 20878, USA; E-mail: kinchm@medimmune.com

Abstract

Intracellular signaling by receptor tyrosine kinases regulates many different aspects of cell behavior. Recent studies in our laboratory and others have demonstrated that the EphA2 receptor tyrosine kinase critically regulates tumor cell growth, migration and invasiveness. Although the cellular consequences of EphA2 signaling have been the focus of recent attention, the biochemical changes that are triggered by ligand-mediated activation of EphA2 remain largely unknown. Herein, we demonstrate that ligand stimulation of EphA2 promotes the nucleus translocation and phosphorylation of ERK kinases, followed by an increase in nuclear induction of the Elk-1 transcription factor. Ligand-mediated activation allows EphA2 to form a molecular complex with the SHC and GRB2 adaptor proteins. Specifically, we demonstrate that tyrosine phosphorylated EphA2 interacts with the PTB and SH2 domains of SHC. We also show that the interaction of EphA2 with GRB2 is indirect and mediated by SHC and that this complex is necessary for EphA2-mediated activation of ERK kinases. These studies provide a novel mechanism to demonstrate how EphA2 can convey information from the cell exterior to the nucleus.

Oncogene (2002) 21, 7690-7699. doi:10.1038/sj.onc.1205758

Keywords

Eph kinase; Eck; signal transduction

Abbreviations

C, Control; CL, cell lysate; EA1-Fc, EphrinA1-Fc; ECM, extracellular matrix; GST, glutathione S-transferase; P-ERK, phosphorylated (activated) ERK kinase; PTB, phosphotyrosine binding domain; SH2, Src homology type 2 domain

Introduction

In the search for signals that regulate cellular behavior, much emphasis has been placed on receptor tyrosine kinases (Hunter, 1998; Reese and Slamon, 1997; Scheid and Woodgett, 2000). This interest stems in part from the knowledge that receptor tyrosine kinases rapidly convey and amplify information from the extracellular environment to the nucleus (Blumer and Johnson, 1994; Davis, 1995). In the classic situation, receptor tyrosine kinases are activated in response to the binding of a specific ligand. Ligand binding stimulates receptor autophosphorylation, which, in turn, creates binding sites for proteins that contain SH2 or PTB domains (Borg and Margolis, 1998; Pawson and Scott, 1998; van der Geer and Pawson, 1995). These interacting adaptor molecules often have direct enzymatic activity or serve as adaptors for enzymes located downstream in a signaling pathway. Receptor tyrosine kinases are often overexpressed and functionally altered in malignant cells, which is consistent with evidence that tyrosine phosphorylation is necessary for many aspects of tumorigenesis (Hunter, 1998; Reese and Slamon, 1997; Yarden and Sliwkowski, 2001).

The Eph transmembrane tyrosine kinases constitute the largest known family of receptor-like kinases, with many members displaying specific patterns of expression in the developing and adult nervous system (Carpenter et al., 1995; Davis et al., 1994). The 14 Eph receptors are divided into two groups according to ligand specificity. EphA receptors predominantly interact with ephrinA ligands and EphB receptors interact with ephrins in the B subclass, both membrane bound as well. In neuronal growth and development, specific EphA proteins are sequestered to specific regions and coordinate specific functions. Eph receptors (including EphA3 and EphB4) direct pathfinding of neurons within migratory fields of cells expressing gradients of their membrane bound ligands. EphB1 is also involved in the regulation of neural processes (Brennan et al., 1997; Nakamoto et al., 1996; Wang and Anderson, 1997). EphB1 and EphA2 direct vascular network assembly, affecting capillary morphogenesis and angiogenesis (Mccormick and Zetter, 1992; Stein et al., 1998). The task of identifying specific roles and signals of Eph receptors continues to be a challenge since many Eph receptors of the same subclass exist in overlapping regions and are controlled by overlapping ligands. Therefore it is essential to isolate these kinases so that newly identified biochemical pathways are linked to the correct Eph receptor.

Our laboratory has been studying tyrosine kinases that regulate the growth and invasiveness of malignant epithelial cells. Much of our recent work has focused on EphA2, which is overexpressed and functionally altered in many different human carcinomas (Dodelet and Pasquale, 2000; Easty et al., 1995; Walker-Daniels et al., 1999; Zelinski et al., 2001). Interestingly, EphA2 acts as a powerful oncoprotein in cancer cells, yet in normal epithelial cells EphA2 appears to negatively regulate cell growth and migration (Zantek et al., 1999; Zelinski et al., 2001). These disparate outcomes relate to differences in ligand binding. Normal cells have low levels of EphA2 that is stably bound to its ligand, ephrinA1 (Zantek et al., 1999) which is anchored to the surface of neighboring cells (Bartley et al., 1994; Shao et al., 1995). In contrast, malignant cells have high levels of EphA2 protein that do not bind ligand (Zantek et al., 1999; Zelinski et al., 2001). The biological effects of ligand binding include decreased extracellular matrix (ECM) attachments, decreased cell migration and inhibition of malignant growth (Miao et al., 2000, 2001; Zantek et al., 1999; Zelinski et al., 2001). Unfortunately very little is known of the biochemical consequences of ligand stimulation of EphA2 or the signals that govern the biological outcomes. In our present study, we demonstrate that in response to ligand, EphA2 transmits signals to the nucleus via MAP kinases. Specifically, we show that EphA2 interacts with the SHC adaptor protein in an activation-dependent manner, and that SHC mediates a linkage between EphA2 and the GRB2 adaptor protein to facilitate ERK activation and thereby negatively regulates ECM attachments.

Results

Ligand-mediated stimulation of ERK kinases

To analyse intracellular signaling pathways that are triggered by ligand stimulation of EphA2, our initial studies utilized the ephrinA1-Fc fusion protein (Pandey et al., 1995b; Zelinski et al., 2001). EphrinA1-Fc (EA1-Fc) is a soluble, engineered form of ligand that mimics ligand binding and thus overcomes the defects in cancer cells that prevent EphA2 from binding to its endogenous ligands (Zelinski et al., 2001). Monolayers of breast cancer MDA-MB-231 cells, which overexpress EphA2 that is not tyrosine phosphorylated, were incubated with EA1-Fc at 37°C for 0-60 min and MAP kinase activation was assessed by Western blot analysis of cell lysates using antibodies that recognize activated ERK (P-ERK). ERK activation increased within 2 min after ligand binding and optimal ERK activation was observed after 5 min (Figure 1a). The stimulation of ERK was transient and began to decrease after 45 min. Results consistently demonstrated a greater than fourfold in ERK activation at 5 min, thus all subsequent experiments with EA1-Fc were performed using these conditions. Ligand-mediated activation of ERK kinases was confirmed by performing identical studies with multiple and different cell models, including cells with high levels of endogenous EphA2 (MDA-MB-231, PC-3, MDA-MB-435, BT549, MDA-MB-436) as well as cells that have been engineered to overexpress EphA2 (MCF-10EphA2, LNCaPEphA2, MCF-7EphA2) (Figures 1b, 2b and data not shown). Thus, EA1-Fc-mediated induction of MAP kinases was not unique to any particular cell model. We will focus hereafter on studies using the MDA-MB-231 cell model, which unless noted otherwise, was found to be representative of the other cell systems.

Multiple measures of MAP kinase activation all confirmed our finding that EA1-Fc activates ERK kinases. For example, immunofluorescence microscopy analyses of ERK subcellular localization revealed that ERK kinases relocated from the cytoplasm to the nucleus in response to ligand binding, which is consistent with the well-characterized behavior of MAP kinases (Gonzalez et al., 1993; Brunet et al., 1995; Lenormand et al., 1998) (Figure 1c). The downstream consequences of ERK signaling were also assessed by measuring the stimulation of Elk-1 transcriptional activity (Figure 1d). An Elk-1 luciferase reporter assay was utilized to confirm that ligand-mediated activation of EphA2 increased ERK activity by at least fourfold (P<0.0004). In contrast, parallel studies with negative control reporters Jun2 and Gal-Luc did not demonstrate significant increases. Consistent results with phospho-specific antibodies, nuclear translocation assays and Elk-1 activation all confirmed that ligand-mediated stimulation of EphA2 selectively activates ERK kinases.

EphrinA1-Fc has the potential to activate multiple EphA kinases (Pasquale, 1997). Using multiple methods, we asked whether EphA2 stimulation was in fact responsible for ERK activation. One way in which we were able to selectively activate EphA2 was through the use of well-characterized and specific monoclonal antibodies. Monolayers of MDA-MB-231 cells were incubated with B2D6, which recognizes an epitope on the extracellular domain of EphA2. EphA2 was activated by the aggregation of B2D6 at the cell surface with rabbit anti-mouse IgG. EphA2 activation was sufficient to trigger ERK activation and with a time course that was comparable to our results with EphrinA1-Fc (Figure 2a). In contrast, parallel samples that had been treated with either primary or secondary antibody alone did not demonstrate increased ERK activity. As an additional control for EphA2 specificity, we reconstituted EphA2-deficient cell lines with EphA2 by transient transfection. LNCaP cells lack endogenous EphA2 and ligand stimulation of LNCaP did not alter ERK activation (Figure 2b, left panel). In contrast, ERK activation was readily detected in EphA2-transfected LNCaP cells and mimicked the timing of ERK activation that we detected using MDA-MB-231 cells (Figure 2b, right panel). Thus, we were able to conclude that stimulation of EphA2 was both necessary and sufficient to activate ERK kinases.

Activation-dependent interaction of EphA2 with SHC

We then sought to determine the mechanism by which EphA2 stimulated ERK activity. Ligand or antibody-mediated activation of EphA2 induces rapid receptor autophosphorylation (Figure 3a) and we considered that this could create binding sites for molecules with SH2 or PTB domains. We focused on the SHC adaptor protein, since SHC has been linked to ERK activation in response to other stimuli and because SHC contains an SH2 and a PTB domain (Cattaneo and Pelicci, 1998; Luzi et al., 2000; van der Geer and Pawson, 1995). We first asked whether EphA2 could interact with endogenous SHC in vivo as measured using co-immunoprecipitation. Upon incubation with EA1-Fc, SHC co-immunoprecipitated with EphA2 and in an activation-dependent manner (Figure 3b). To prevent potential concerns about EphA2 that might have precipitated with cell-attached EA1-Fc, a parallel control was performed in which cells were treated with EA1-Fc at 0°C (to minimize activation). This confirmed that ligand stimulation increases the interaction of EphA2 with SHC. We then sought to determine the sites on SHC that bind EphA2. To accomplish this, we utilized in vitro binding assays with purified GST-fusion proteins (Figure 3c). These fusion proteins encompassed either the SH2 or PTB domain of SHC. Purified GST was utilized as a negative control. The GST fusion proteins were immobilized onto glutathione beads and incubated with cell lysates that had been harvested from either unstimulated or EA1-Fc stimulated MDA-MB-231 cells. EphA2 was bound to both the Sh2 and PTB domains of SHC in an activation-dependent manner. Neither purified GST protein nor the SH2 domain from the NCK adaptor protein bound to EphA2 even when EphA2 was stimulated with EA1-Fc, confirming the specificity of the SHC interaction. Confirmatory results with in vivo and in vitro systems indicate that ligand-mediated activation of EphA2 increases interactions with SHC via the SH2 and PTB domains.

SHC-dependent interaction of EphA2 and GRB2

We next investigated the biochemical outcomes of SHC binding to EphA2. With other receptor tyrosine kinases, such as PDGF-receptor, SHC stimulates MAP kinases by linking the receptor to adaptor proteins, most notably GRB2 (Pawson, 1995). Our hypothesis was that GRB2 could interact with EphA2. We tested this hypothesis using the same types of in vitro and in vivo analyses that we had used to study SHC binding. First, untreated or EA1-Fc treated MDA-MB-231 cells were extracted and GRB2-interacting proteins were isolated using purified GRB2-GST fusion proteins. These experiments demonstrated that EphA2 interacted with GRB2 in an activation-dependent manner (Figure 4a). Furthermore, endogenous GRB2 co-immunoprecipitated with EphA2, and in an activation-dependent manner (Figure 4b).

After showing that EphA2 could form a complex with GRB2 and SHC upon stimulation with EA1-Fc, we then sought to determine whether the interaction of GRB2 with EphA2 would require SHC. Initial support for this hypothesis was obtained in studies of the interaction of GRB2 with SHC. Increasing amounts of GRB2 were found in association with SHC in response to EphA2 stimulation (Figure 5a). We then sought to determine whether SHC binding was necessary for the interaction of EphA2 with GRB2. To ask this, we utilized the GST fusion proteins described above as dominant-negative inhibitors of SHC. The SH2 or PTB fusion proteins were eluted from the GST beads and used as dominant-negative inhibitors of endogenous SHC. After stimulating EphA2 on MDA-MB-231 cells with EA1-Fc, cell lysates were harvested and incubated with the GST fusion proteins prior to assessing EphA2-SHC binding via immunoprecipitation. Dominant-negative inhibitors of either the PTB or SH2 domain of SHC were able to competitively displace the binding of EphA2 to endogenous SHC (Figure 5b). This result suggests that the interaction between EphA2 and SHC encompasses both sites on SHC. An identical experiment was then performed to ask if GRB2 binding to EphA2 requires SHC (Figure 5c). Indeed, inhibitors of either the PTB or Sh2 domains of SHC prevented EphA2 from interacting with GRB2. These findings suggest that the interaction of EphA2 with GRB2 is indirect and mediated by SHC.

SHC binding is necessary for ERK activation

Based on the biochemical analyses of the EphA2/SHC/GRB2 complex, we asked whether the interaction of EphA2 with SHC would be necessary for MAP kinases activation. To address this question, the SHC-GST fusion proteins were introduced to MDA-MB-231 cells after elution with glutathione and prior to EphA2 stimulation. The purified proteins were inserted into cells using the Chariot transfection method (Graham and van der Eb, 1973; Morris et al., 2000a,b) which successfully loads 60-80% of the cells with negligible toxicity. EphA2 was stimulated for 5 min using EA1-Fc and nuclear localization of ERK was employed as a marker of ERK activation as detailed above. We also performed a control in which samples were transfected with glutathione alone and determined that the experimental procedures did not significantly alter the level of ERK activation. This method revealed that both PTB and SH2 dominant-negative inhibitors of SHC decreased ERK activity by 61% (P<0.03) and 70% (P<0.01) respectively (Figure 6). Similar results were calculated using bulk loading (Mcneil, 1989) with GST-fusion PTB and SH2 binding domains of SHC. A control in which the samples were subjected to the bulk loading method only demonstrated that the experimental procedure itself did not alter the level of ERK activation. This method revealed that both PTB and SH2 dominant-negative inhibitors of SHC decreased ERK activity (43% (P<0.007) and 57% (P<0.002) of ERK activation, respectively, data not shown). Thus, these results indicate that SHC binding is necessary for EphA2-mediated activation of MAP kinases.

EphA2 stimulation of ERK decreases ECM attachments

Recent reports have demonstrated that ligand-mediated stimulation of EphA2 negatively regulates cell attachments to underlying ECM protein (Zantek et al., 1999; Miao et al., 2000). Since EphA2 stimulation negatively regulates MDA-MB-231 cell-ECM adhesions we employed this cell model to examine the question of whether EphA2-mediated induction of ERK activity was necessary for this biological outcome. Indeed, EA1-Fc treatment for 5 min decreased the substratum binding of MDA-MB-231 cells by at least 40% (P<0.004) (Figure 7a). Importantly, the inclusion of the MEK1 inhibitor, PD98059 at a concentration of 50 muM, prevented the ligand-mediated decrease in ECM attachments (Figure 7b). These results indicate that ligand-mediated induction of ERK activity is necessary for the ability of EphA2 to negatively regulate ECM adhesions.

Discussion

This study demonstrates that ligand-mediated stimulation of the EphA2 receptor tyrosine kinase transmits signals from the cell membrane, through MAP kinase. These signals are communicated into the nucleus via induction of Elk-1 transcription and back to the cell membrane via destabilization of cell-ECM attachments. We also show that the biochemical mechanism of EphA2 signaling involves an activation-dependent interaction of tyrosine phosphorylated EphA2 with the SHC adaptor protein. SHC, in turn, bridges EphA2 to GRB2, which facilitates the activation and nuclear translocation of ERK kinases (Figure 8).

To our knowledge, we provide the first evidence that a tyrosine phosphorylated EphA-family kinase can interact with SHC or complex with GRB2. Similarly, a recent report demonstrated that a phosphorylated cytoplasmic domain of EphB-family kinases can bind GRB2 (Stein et al., 1998). We also find that the interaction between EphA2 and SHC involves both the PTB and SH2 domains. This finding is consistent with the facts that ligand binding promotes EphA2 autophosphorylation and with evidence that PTB and SH2 domains recognize phosphotyrosine motifs (van der Geer and Pawson, 1995). Although recent evidence has demonstrated that PTB domains can sometimes bind unphosphorylated tyrosine residues (Charest et al., 1996), the activation-dependent interaction of EphA2 with SHC suggests that the PTB domain recognizes phosphotyrosine residues in EphA2. Our studies with dominant-negative inhibitors suggest that SHC links GRB2 to EphA2 and that EphA2 stimulation increased the association of GRB2 with SHC. However, we cannot formally exclude that both Grb2 and Shc could bind the same phosphotyrosine residues.

Another novel aspect of the present study is a demonstration that EphA2 can transmit signals to promote Elk-1 induction in the nucleus. Studies of EphA2-mediated intracellular signaling have primarily focused on membrane-proximal signaling by PI3-kinase, SLAP, SHP2 and FAK (Miao et al., 2000; Pandey et al., 1994, 1995a), whereas our present study extends signaling from EphA2 at the cell membrane to the nucleus. The induction of Elk-1 is also interesting since ERK stimulation of Elk-1 has been linked with both the postive and negative regulation of cell proliferation and differentiation (Brunet et al., 1995; Clarkson et al., 1999; Davis, 1995; Lin et al., 1997; Macleod et al., 1992; Townsend et al., 1999; Treisman, 1994; Vanhoutte et al., 2001). Although the biological consequences of Elk-1 induction by EphA2 are presently unknown, ligand-mediated activation of EphA2 has been linked with the negative regulation of numerous biological outcomes, including the regulation of cell proliferation, survival, migration, invasion, differentiation and angiogenesis (Miao et al., 2000, 2001; Pandey et al., 1995a,b; Rosenberg et al., 1997; Zantek et al., 1999; Zelinski et al., 2001). We suggest that this finding warrants further investigation of whether and how EphA2-mediated regulation of ERK and Elk-1 might regulate cellular behaviors.

A recent manuscript indicated that stimulation of EphA family kinases negatively regulates MAP kinase signaling in PC-3, a human prostate cancer cell model (Miao et al., 2001). These findings contradict our present demonstration that EphA2 activates ERK signaling in the cell systems examined herein. There are a number of potential explanations that could account for these disparate outcomes. First, our results herein have largely utilized breast cancer cell models and it is formally possible that the EphA2 in breast and prostate cancer cells respond differently to ligand. However, we do not favor this explanation because we find that EphA2-transfected LNCaP prostate cancer cells respond identically to breast cancer cells. Moreover, we have investigated EphA2 signaling in PC-3 cells and have verified our findings in this model system. EphA2 induction of ERK activity was observed under many different experimental conditions and our replication of the methods reported previously did not result in decreased ERK activity. In the course of these studies, we obtained five separate isolated of PC-3 cells that yielded isolate-specific outcomes. In four out of five isolates stimulation of EphA2 increased ERK activation in PC-3 cells. It remains unclear why a single variant, of the 11 different EphA2-expressing cell models tested to date, is the only cell system in which ERK activity is reduced by EphA2 stimulation.

EA1-Fc has the potential to stimulate any of eight different EphA family kinases (Pasquale, 1997). To overcome potential complexities in data interpretation, we confirmed the ERK results using both EA1-Fc and EphA2-specific monoclonal antibodies. The study mentioned above stating that EA1-Fc negatively regulates ERK signaling only utilized EA1 ligand; which suggests that this data could have resulted from the stimulation of an EphA-kinase other than EphA2. Consistent with this, we have found that PC-3 cells express multiple EphA-family receptors (Kinch MS, not shown) and a recent study demonstrated that certain EphB family members may negatively regulate ERK signaling (Elowe et al., 2001). Another potential explanation could relate to ERK desensitization. ERK phosphorylation is negatively regulated by various phosphatases (Lu et al., 2002; Verlhac et al., 2000) and indeed, our present studies indicate that EphA2 stimulation of ERK kinases is quite transient. Moreover, we have found that EphA2 stimulation of ERK activity in PC-3 cells occurs more rapidly and more transiently than in any of the other cell systems that we have investigated (data not shown). Thus, future studies could examine whether the previous evidence of EphA-mediated inhibition of ERK kinases might actually represent this desensitization phase of EphA2-mediated induction of ERK activity.

Recent studies by our laboratory and others have demonstrated that ligand-mediated activation of EphA2 negatively regulates tumor cell growth and invasiveness (Miao et al., 2000, 2001; Zantek et al., 1999; Zelinski et al., 2001). Another novel aspect of our present findings is that EphA2-mediated induction of the ERK signaling pathway is necessary for the negative regulation of ECM attachments. A similar inhibitory action might be expected in normal epithelial cells since EphA2 stimulation of ERK kinases would predominate as cells reach confluence. Thus, ECM attachments and ECM-associated signals that promote cell growth, would be constrained by contact inhibitory signals. The activation of ERK by EphA2 is transient, returning to basal levels within 45-60 min. This timing of ERK activation is interesting in light of evidence that the duration of ERK signaling can profoundly impact the biological outcomes of receptor signaling (Aliaga et al., 1999; Brunet et al., 1995; Cowley et al., 1994; Roovers et al., 1999; Roovers and Assoian, 2000; Schramek et al., 1997). Based on these findings, we suggest that transient ERK activation could similarly inhibit tumor cell growth in response to EphA2 stimulation. This inhibitory affect may also be coupled with the induction of a MAP kinase phosphatase. Alternatively, we find that one consequence of EphA2 activation is EphA2 degradation (Kinch, manuscript submitted). Thus, ligand stimulation could decrease ERK activity over time by decreasing EphA2 protein levels.

In summary, the major finding of our present study is that ligand stimulation of the EphA2 receptor tyrosine kinase transmits signals to the nucleus and back to the cell membrane via induction of the MAP kinase pathway. We also show that tyrosine phosphorylated EphA2 interacts with SHC and that SHC provides a linkage between EphA2 and GRB2. Finally, the interaction of EphA2 with the SHC-GRB2 complex is necessary for the transmission of downstream signaling to MAP kinase. The novel interactions described in this paper may help to link the oncoprotein, EphA2, to recent biological functions including the regulation of ECM attachments, cell growth and invasiveness.

Materials and methods

Cell lines and antibodies

Human breast (MDA-MB-231, MDA-MB-435, MCF10EphA2, BT549, MDA-MB-436, MCF7EphA2) and prostate (PC-3, LNCaP) carcinoma cells were cultured as described previously (Walker-Daniels et al., 1999; Zelinski et al., 2001). Anbodies specific for P-ERK and ERK1 were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA) and BD Transduction Laboratories (Lexington, KY, USA), respectively. SHC polyclonal antibodies, which recognizes the 66, 52 (preferred) and 46 kDa protein forms, and GRB2 monoclonal antibodies were purchased from Transduction Laboratories (Lexington, KY, USA). Monoclonal antibodies specific for EphA2 (clones D7 and B2D6) were produced in the laboratory or purchased from Upstate Biotechnology Inc (Lake Placid, NY, USA). Antibodies specific for P-Tyr (4G10) were purchased from Upstate Biologicals, Inc. Ephrin-A1-Fc (EA1-Fc) was a generous gift from Dr B Wang (Case Western Reserve University).

Western blot analysis

Western blot analyses were performed on normalized cell lysates as described previously (Zelinski et al., 2001) and antibody binding was detected by enhanced chemiluminescence (Pierce, Rockford, IL, USA) and autoradiography (Kodak X-OMAT; Kodak, Rochester, NY, USA).

EphA2 stimulation

Aggregation of EphA2 receptors was performed as previously described (Zelinski et al., 2001). Unless noted otherwise, all stimulations of EphA2 were with 1 mug/ml of the soluble ligand EA1-Fc for 0 (-) or 5 (+) min incubation at 37°C.

Immunofluorescence staining

Staining of cell monolayers with P-ERK antibodies was performed as previously described (Zelinski et al., 2001). Briefly, cells were grown on glass coverslips and observed at densities of 70% confluence. Samples were fixed in 3.7% formaldehyde solution, extracted in 0.5% Triton X-100, and stained. Immunostaining was visualized using rhodamine-conjugated donkey anti-mouse antibodies (Chemicon; Temecula, CA, USA). Nuclear localization was determined using Bisbenzimide (Hoechst) stain (1 : 1000) in Universal buffer added in tandem with the conjugated secondary antibody. Coverslips were viewed using a BX60 Olympus (Lake Success, NY, USA).

Reporter constructs

Induction of Elk-1 was determined by cotransfection with 5xGAL4-Luc, which contains the luciferase gene driven by a minimal promoter containing tandem GAL4 DNA-binding sites (Graham et al., 1996) and GAL4-Elk-1, a fusion protein containing the GAL4 DNA-binding domain together with the transactivation domain of Elk-1 (Zohn et al., 1998). The Jun2-Luc reporter plasmid contains three tandem copies of the Jun/ATF-2 DNA binding motif present in the c-jun promoter (Zohn et al., 1998). All of the reporter constructs, Gal4-Elk-1, Jun2-Luc and 5xGAL4-Luc, were generously provided by Dr C Der (University of North Carolina, Chapel Hill, NC, USA). Briefly, MDA-MD-231 cells were transiently transfected with indicated plasmid cDNAs by LipofectAMINE Plus reagents (Life Technologies Inc., Grand Island, NY, USA). Approximately 30-40 h after transfection and appropriate incubation time with EA1-Fc, cells lysates were prepared with Luciferase reporter lysis buffer (Promega, Madison, WI, USA). After 15 min on ice, cell lysate was quick-frozen in liquid nitrogen prior to use, thawed and then prepared with 100 mul of the Luciferase Assay System substrate (Promega), and analysed using a Monolight 2010 Luminometer.

Transient transfection of EphA2

Monolayers of LNCaP cells were transfected with 2 mug of wild type pNeoMSV-EphA2 (generously provided by Dr T Hunter, Scripps Institute, La Jolla, CA, USA) following the protocol provided with LipofectAMINE Plus reagents (Life Technologies, Inc., Grand Island, NY, USA). Cells were then provided with serum rich media and incubated at 37°C for the remaining 30-40 h. LNCaP cells were lysed on ice for 15 min. As a control, a parallel transfection was performed using the pNeoMSV vector. Parental cells were also used for comparative controls in the LNCaP experiment. Cell lysates were normalized and resolved by 10% SDS-PAGE as described (Zelinski et al., 2001).

Immunoprecipitation

Immunoprecipitation experiments were performed as previously described (Zelinski et al., 2001) for 1.5 h at 4°C with the appropriate antibody and conjugated protein A-Sepharose (Sigma). To confirm equal loading, blots were stripped and reprobed with antibodies specific for ERK1, SHC or EphA2 as were applicable.

GST fusion protein pull down assay

GST-fusion proteins, constructed in pGEX vectors, were generously provided by Drs J O'Bryan (NIEHS) (pGEX-GST-SHC-SH2 and pGEX-GST-SHC-PTB), L Quilliam (Indiana University) (p-GEX-GST-GRB2 and p-GEX-GST-NCK-SH2), and R Geahlen (Purdue University) (p-GEX-GST). The vectors listed above were induced by the addition of 0.5 to 2.0 mM IPTG for appropriate times. Bacteria were lysed in B-Per reagent (Promega) and purified for use with lysates. MDA-MB-231 cells grown to confluence, were left unstimulated or stimulated with 1 mug/ml EA1-Fc for 5 min at 37°C. Cells were then placed on ice and washed twice with chilled PBS and lysed with Triton X-100 buffer for 15 min. Five hundred mug of clarified lysate was incubated with purified GST-containing constructs from target proteins at 4°C for 1 h. GST pellets were spun down, washed with lysis buffer three times, and resuspended in SDS sample buffer, then resolved by 10% SDS-PAGE as previously described (Zelinski et al., 2001).

Bulk loading of dominant negative inhibitors

Purified fusion proteins that encompass the SH2 or PTB domains of SHC were introduced into viable cell monolayers using a modification of a published protocol (Mcneil, 1989). Briefly, monolayers of MDA-MB-231 cells, grown overnight on glass coverslips, were incubated in a KCl solution containing 6% Pluronic F-68 (Sigma, St Louis, MO, USA), 2% 2-mercaptoethanol and purified fusion protein for 30 s. The coverslips were then covered with 425 mum glass spheres for 30 s before restoring normal culture conditions. The samples were then stimulated using EA1-Fc and ERK activation was determined microscopically as detailed above.

Chariot protein transfection

Purified GST-fusion proteins of both SH2 and PTB domains of SHC were eluded in glutathione and transfected into viable cell monlayers following the protocol provided with Chariot Transfection reagents (Active Motif Inc., Carlsbad, CA, USA). One mug of protein was transfected into cells in monolayer and incubated in serum-free media at 37°C, 5% CO2 for 1 h. One millilitre of complete growth medium was added directly to the cells and returned in the incubator for 2 h prior to staining (as detailed above).

Statistical analyses

To determine statistical significance, all analyses were assessed in at least three independent experiments and were analysed using a Student's t-test, defining P<0.05 as significant.

Extracellular matrix (ECM) attachment assays

For relative attachment assays 100 000 MDA-MD-231 cells were plated in a 24 well dish and allowed to sit for 12 h prior to stimulation with EA1-Fc for 5 min. The plate was then tapped on all four sides; media was aspirated and remaining cells were suspended with Trypsin and counted on a hemacytometer. MEK 1 inhibitor PD98059 purchased from Cell Signaling Technology Inc. (Beverly, MA, USA) was used at a final concentration of 50 muM and added to media at 37°C for 1 h prior to EA1-Fc treatment and ECM assessment.

Acknowledgements

We thank Drs C Der, T Hunter, J O'Bryan, L Quilliam and R Geahlen for reagents and Drs J Fernandez, D Zelinski and the Kinch Laboratory for critical advice and support.

References

Aliaga JC, Deschenes C, Beaulieu JF, Calvo EL, Rivard N. (1999). Am. J. Physiol., 277: G631-G641.

Bartley TD, Hunt RW, Welcher AA, Boyle WJ, Parker VP, Lindberg RA, Lu HS, Colombero AM, Elliott RL, Guthrie BA. (1994). Nature, 368: 558-560. MEDLINE

Blumer KJ, Johnson GL. (1994). Trends Biochem. Sci., 19: 236-240. MEDLINE

Borg JP, Margolis B. (1998). Curr. Top. Microbiol. Immunol., 228: 23-38. MEDLINE

Brennan C, Monschau B, Lindberg R, Guthrie B, Drescher U, Bonhoeffer F, Holder N. (1997). Development, 124: 655-664.

Brunet A, Brondello JM, L'Allemain G, Lenormand P, McKenzie F, Pages G, Pouyssegur J. (1995). Comptes Rendus des Seances de la Societe de Biologie et de Ses Filiales, 189: 43-57.

Carpenter MK, Shilling H, Van den Bos T, Beckmann MP, Cerretti DP, Kott JN, Westrum LE, Davison BL, Fletcher FA. (1995). J. Neurosci. Res., 42: 199-206.

Cattaneo E, Pelicci PG. (1998). Trends Neurosci., 21: 476-481. Article MEDLINE

Charest A, Wagner J, Jacobs S, McGlade CJ, Tremblay ML. (1996). J. Biol. Chem., 271: 8424-8429. Article MEDLINE

Clarkson RW, Shang CA, Levitt LK, Howard T, Waters MJ. (1999). Mol. Endocrinol., 13: 619-631.

Cowley S, Paterson H, Kemp P, Marshall CJ. (1994). Cell, 77: 841-852. MEDLINE

Davis RJ. (1995). Mol. Repro.Dev., 42: 459-467.

Davis S, Gale NW, Aldrich TH, Maisonpierre PC, Lhotak V, Pawson T, Goldfarb M, Yancopoulos GD. (1994). Science, 266: 816-819. MEDLINE

Dodelet VC, Pasquale EB. (2000). Oncogene, 19: 5614-5619.

Easty DJ, Herlyn M, Bennett DC. (1995). Int. J. Cancer, 60: 129-136. MEDLINE

Elowe S, Holland SJ, Kulkarni S, Pawson T. (2001). Mol. Cell. Biol., 21: 7429-7441. Article MEDLINE

Gonzalez FA, Seth A, Raden DL, Bowman DS, Fay FS, Davis RJ. (1993). J. Cell Biol., 122: 1089-1101.

Graham FL, van der Eb AJ. (1973). Virology, 52: 456-467. MEDLINE

Graham SM, Vojtek AB, Huff SY, Cox AD, Clark GJ, Cooper JA, Der CJ. (1996). Mol. Cell. Biol., 16: 6132-6140. MEDLINE

Hunter T. (1998). Philos. Trans. R. Soc. Lond. B. Biol. Sci., 353: 583-605.

Lenormand P, Brondello JM, Brunet A, Pouyssegur J. (1998). J. Cell Biol., 142: 625-633. MEDLINE

Lin TH, Aplin AESY, Chen Q, Schaller M, Romer LH, Aukhil I, Juliano RL. (1997). J. Cell Biol., 136: 1385-1396. Article MEDLINE

Lu Q, Dunn RL, Angeles R, Smith GD. (2002). Biol. Repro., 66: 29-37.

Luzi L, Confalonieri S, Di Fiore PP, Pelicci PG. (2000). Curr. Opin. Genet. Dev., 10: 668-674. Article MEDLINE

Macleod K, Leprince D, Stehelin D. (1992). Trends Biochem. Sci., 17: 251-256. MEDLINE

McCormick BA, Zetter BR. (1992). Pharmacol. Ther., 53: 239-260.

McNeil PL. (1989). Fluorescence Microscopy of Living Cells in Culture Wang YL and Taylor DL (eds) San Diego: Academic Press, Inc, pp 153-173.

Miao H, Burnett E, Kinch MS, Simon E, Wang B. (2000). Nat. Cell Biol., 2: 62-69. Article MEDLINE

Miao H, Wei B, Li Q, Terry A, Schelling JR, Rhim JSSJR, Burnett E, Wang B. (2001). Nat. Cell Biol., 3: 527-530. Article MEDLINE

Morris MC, Chaloin L, Heitz F, Divita G. (2000a). Curr. Opin. Biotechnol., 11: 461-466. Article MEDLINE

Morris MC, Heitz A, Mery J, Heitz F, Divita G. (2000b). J. Biol. Chem., 275: 28849-28857. Article MEDLINE

Nakamoto M, Cheng HJ, Friedman GC, McLaughlin T, Hansen MJ, Yoon CH, O'Leary DD, Flanagan JG. (1996). Cell, 86: 755-766. MEDLINE

Pandey A, Duan H, Dixit VM. (1995a). J. Biol. Chem., 270: 19201-19204. Article MEDLINE

Pandey A, Shao H, Marks RM, Polverini PJ, Dixit VM. (1995b). Science, 268: 567-569. MEDLINE

Pasquale EB. (1997). Curr. Opin. Cell Biol., 9: 608-615. Article MEDLINE

Pawson T. (1995). Nature, 373: 573-580. MEDLINE

Pawson T, Scott JD. (1998). Science, 278: 2075-2080.

Reese DM, Slamon DJ. (1997). Stem Cells, 15: 1-8. MEDLINE

Roovers K, Assoian RK. (2000). Bioessays, 22: 818-826. Article MEDLINE

Roovers K, Davey G, Zhu X, Bottazzi ME, Assoian RK. (1999). Mol. Biol. Cell., 10: 3197-3204. MEDLINE

Rosenberg IM, Goke M, Kanai M, Reinecker HC, Podolsky DK. (1997). Am. J. Physiol., 273: G824-G832.

Scheid MP, Woodgett JR. (2000). Curr. Biol., 10: R191-R194. MEDLINE

Schramek H, Feifel E, Healy E, Pollack V. (1997). J. Biol. Chem., 272: 11426-11433. MEDLINE

Shao H, Pandey A, O'Shea KS, Seldin M, Dixit VM. (1995). J. Biol. Chem., 270: 5636-5641. MEDLINE

Stein E, Lane AA, Cerretti DP, Schoecklmann HO, Schroff AD, Van ERL, Daniel TO. (1998). Genes Dev., 12: 667-678. MEDLINE

Townsend KJ, Zhou P, Qian L, Bieszczad CK, Lowrey CH, Yen A, Craig RW. (1999). J. Biol. Chem., 274: 1801-1813. MEDLINE

Treisman R. (1994). Curr. Opin. Genet. Dev., 4: 96-101. MEDLINE

van der Geer P, Pawson T. (1995). Trends Biochem. Sci., 20: 277-280. MEDLINE

Vanhoutte P, Nissen JL, Brugg B, Gaspera BD, Besson MJ, Hipskind RA, Caboche J. (2001). J. Biol. Chem., 276: 5189-5196.

Verlhac MH, Lefebvre C, Kubiak JZ, Umbhauer M, Rassinier P, Colledge W, Maro B. (2000). EMBO J., 19: 6065-6074.

Walker-Daniels J, Coffman K, Azimi M, Rhim JS, Bostwick DG, Snyder P, Kerns BJ, Waters DJ, Kinch MS. (1999). Prostate, 41: 275-280. Article MEDLINE

Wang HU, Anderson DJ. (1997). Neuron, 18: 383-396. Article MEDLINE

Yarden Y, Sliwkowski MX. (2001). Nat. Rev. Mol. Cell Biol., 2: 127-137. Article MEDLINE

Zantek ND, Azimi M, Fedor-Chaiken M, Wang B, Brackenbury R, Kinch MS. (1999). Cell Growth Differ., 10: 629-638. MEDLINE

Zelinski DP, Zantek ND, Stewart J, Irizarry A, Kinch MS. (2001). Cancer Res., 61: 2301-2306. MEDLINE

Zohn IE, Symons M, Chrzanowska-Wodnicka M, Westwick JK, Der CJ. (1998). Mol. Cell. Biol., 18: 1225-1235. MEDLINE

Figures

Figure 1 Ligand-mediated stimulation of EphA2 activates ERK kinases. (a) MDA-MB-231 breast cancer cells were incubated in the presence of 1 mug/ml Ephrin A1-Fc (EA1-Fc) at 37°C for 0-60 min. Equal amounts of cell lyates were then resolved by SDS-PAGE and subjected to Western blot analysis using P-ERK specific antibodies (top). Duplicate cell lysates were resolved on the same gel and probed with ERK 1 antibody as a loading control (bottom). (b) EphA2-induction of ERK applies to multiple cell systems. MDA-MB-435, MCFEphA2, PC-3, and MDA-MB-436 cell models were incubated in the presence or absence of 1 mug EA1-Fc for 5 min at 37°C. Equal amounts of cell lysates were resolved by SDS-PAGE and subjected to Western blot analyses using P-ERK specific antibodies (top), with ERK 1 antibodies serving as a loading control (bottom). (c) The nuclear translocation of ERK was assessed following EA1-Fc treatment of MDA-MB-231 cells using immunofluorescence microscopy. Note that in both assays, the highest levels of ERK activation were detected at early time points (5 and 20 min; P<0.002 and P<0.003 relative to untreated controls, respectively). (d) Elk-1 is induced upon ligand-mediated EphA2 stimulation. MDA-MB-231 cells were transfected with a luciferase reporter of the Elk-1 promoter, which revealed and activation-dependent induction of Elk-1 in response to EphA2 stimulation (P<0.0004). The expression levels of a Jun2-Luc reporter construct and a negative control reporter (C; Gal4) were not significantly altered (P<0.55 and P<0.19, respectively)

Figure 2 Stimulation of EphA2 is sufficient to activate ERK. (a) The EphA2 on MDA-MB-231 breast cancer cells was activated by aggregating EphA2-specific antibodies (1°; B2D6) with goat anti-mouse IgG (2°). Cells treated with primary antibodies (1°) alone or secondary antibodies alone (2°) were compared with samples where EphA2 was aggregated (1°+2°) for 0-20 min at 37°C. Cell lysates were resolved by SDS-PAGE and P-ERK (top) and ERK1 levels were assessed by Western blot analysis as a loading control (bottom). Note that a similar increase in P-ERK was observed in Figure 1a. (b) ERK activation in LNCaP cells, which lack endogenous EphA2, was assessed in response to ligand (EA1-Fc) stimulation (left). These results were then compared with a parallel experiment using LNCaP cells that had been transfected with vector (pNeoMSV) only or EphA2 (right). The first lane indicates a negative control in which parental LNCaP cells were stimulated for 5 min with EA1-Fc

Figure 3 Activation-dependent interaction of EphA2 with SHC. (a) The phosphotyrosine content of immunoprecipitated EphA2 was assessed over a 0 (no stimulation) to 20 min time course after EA1-Fc treatment of MDA-MB-231 cells. (b) Complexes of the SHC adaptor protein were immunoprecipitated from MDA-MB-231 cells that had been incubated in the absence or presence of EA1-Fc. The samples were then probed for associated EphA2 (top). The membranes were then stripped and reprobed with SHC (52 kDa) antibodies as a loading control. Note that lane 1 represents samples that had been incubated with EA1-Fc on ice to control for any EphA2 that had been precipitated by the EA1-Fc. (c) Purified GST-fusion proteins that encode for SH2 or PTB binding domains of SHC were immobilized and used to identify complexes that contain EphA2. To assure equal loading, the amount of GST fusion proteins were assessed using a standard Coomassie assay (data not shown). Note that EphA2 interacted with both the SH2 and PTB domains of SHC, but not NCK, in an activation-dependent manner. A parallel sample that had been incubated with immobilized GST is included as an additional negative control (C)

Figure 4 Activation-dependent interaction of EphA2 and GRB2. (a) Immobilized GST-fusion proteins that encompass GRB2 were utilized to identify associated protein. Cell lysates were harvested from MDA-MB-231 cells that had been incubated in the presence or absence of EA1-Fc for 5 min. GRB2-associated EphA2 was then observed by Western blot analyses of isolated complexes. To assure equal loading, the amount of GST fusion proteins were assessed using a standard Coomassie assay (data not shown). A parallel sample that had been incubated with immobilized GST is included as an additional negative control (C). (b) MDA-MB-231 cells were treated in the presence or absence of EA1-Fc before immunoprecipitation of EphA2 with specific antibodies. The samples were then resolved by SDS-PAGE and Western blot analysis was performed using GRB2-specific antibodies (top). The samples were then stripped and reprobed with EphA2 antibodies as a loading control (bottom)

Figure 5 SHC binding to EphA2 is necessary for GRB2 binding. (a) MDA-MB-231 cells were treated in the presence or absence of EA1-Fc for 5 min at 37°C. The samples were then extracted and immunoprecipitated using SHC antibodies and probed for the presence of associated GRB2. (b) MDA-MB-231 cells were treated in the presence or absence of EA1-Fc for 5 min at 37°C. The samples were then extracted and immunoprecipitate using SHC antibodies in the presence or absence of purified GST fusion proteins that encode for either the SH2 or PTB domains of SHC as indicated, with purified GST serving as a negative control. The immunoprecipitated complexes were resolved by SDS-PAGE and probed with EphA2-specific antibodies (top). The membranes were then stripped and re-probed with SHC antibodies as a loading control (bottom). 'CL' denotes cell lysate, which provided a positive control for antibody reactivity for SHC. (c) Cell lysates were harvested from samples of MDA-MB-231 cells that had been stimulated with EA1-Fc for 5 min at 37°C. The extracts were then immunoprecipitated with EphA2 antibodies in the presence or absence of the indicated GST fusion proteins (GST, PTB, SH2). GST is included as a negative control. After probing for associated GRB2 (top), the samples were stripped and re-probed with EphA2 antibodies as a loading control (bottom). Note that the interaction of EphA2 with SHC and GRB2 required interactions with both the SH2 and PTB domains

Figure 6 SHC binding to EphA2 is necessary for ERK activation. Samples of MDA-MB-231 cells, grown on glass coverslips overnight, were transiently transfected with 1 mug protein in glutathione using the Chariot protocol as detailed in the Materials and methods. EphA2 was stimulated using EA1-Fc for 5 min at 37°C before assessment of ERK activation. Shown are the fraction of cells that demonstrated ERK nuclear localization. Note that the dominant-negative PTB (61%, P<0.03) and SH2 (70%, P<0.01) domains of SHC significantly inhibited EphA2-mediated activation of ERK. Cells that have been transfected with glutathione alone (C) and glutathione-GST served as negative controls. The first bar represents cells that did not receive any ligand treatment. Fraction of positive nuclear staining was assessed by comparing EA1-Fc stimulated cells bulk loaded with glutathione-GST to those receiving the SHC fusion protein

Figure 7 EphA2 induction of MEK is necessary for reduced ECM attachments. 1´105 MDA-MB-231 cells (-/+ EA1-Fc) were suspended and incubated onto ECM-coated supports in the absence (a) or presence (b) of the MEK 1 inhibitor, PD98059. After removing detached cells, the remaining attached cells were suspended with trypsin and counted microscopically using a hemacytometer. Note that EA1-Fc induced a significant decrease (43%, P<0.004) but that the addition of the MEK 1 inhibitor prevented the decrease in ECM attachments

Figure 8 Proposed model of EA1-Fc stimulated EphA2 signaling. Shown is a proposed model of the biochemical mechanism of EphA2 mediated ERK activation. Tyrosine phosphorylated EphA2 interacts with the PTB and SH2 domains of the SHC adaptor protein. SHC, in turn, bridges EphA2 to GRB2, which facilitates the activation and nuclear translocation of ERK kinases, where they induce the Elk-1 transcription factor

Received 28 April 2002; revised 6 June 2002; accepted 6 July 2002
31 October 2002, Volume 21, Number 50, Pages 7690-7699
Table of contents    Previous  Article  Next    [PDF]