Phosphorylation of Tyrosine Residues in the Kinase Domain and ...

MOLECULAR AND CELLULAR BIOLOGY, July 2000, p. 4791–4805 0270-7306/00/$04.00⫹0 Copyright © 2000, American Society for Microbiology. All Rights Reserved.

Vol. 20, No. 13

Phosphorylation of Tyrosine Residues in the Kinase Domain and Juxtamembrane Region Regulates the Biological and Catalytic Activities of Eph Receptors KATHLEEN L. BINNS,1,2 PAUL P. TAYLOR,1 FRANK SICHERI,1,2 TONY PAWSON,1,2* 1 AND SACHA J. HOLLAND † Samuel Lunenfeld Research Institute, Mount Sinai Hospital, Toronto, Ontario M5G 1X5,1 and Department of Molecular and Medical Genetics, University of Toronto, Toronto, Ontario M5G 1A8,2 Canada Received 15 November 1999/Returned for modification 5 January 2000/Accepted 28 February 2000

Members of the Eph family of receptor tyrosine kinases exhibit a striking degree of amino acid homology, particularly notable in the kinase and membrane-proximal regions. A mutagenesis approach was taken to address the functions of specific conserved tyrosine residues within these catalytic and juxtamembrane domains. Ligand stimulation of wild-type EphB2 in neuronal NG108-15 cells resulted in an upregulation of catalytic activity and an increase in cellular tyrosine phosphorylation, accompanied by a retraction of neuritic processes. Tyrosine-to-phenylalanine substitutions within the conserved juxtamembrane motif abolished these responses. The mechanistic basis for these observations was examined using the highly related EphA4 receptor in a continuous coupled kinase assay. Tandem mass spectrometry experiments confirmed autophosphorylation of the two juxtamembrane tyrosine residues and also identified a tyrosine within the kinase domain activation segment as a phosphorylation site. Kinetic analysis revealed a decreased affinity for peptide substrate upon substitution of activation segment or juxtamembrane tyrosines. Together, our data suggest that the catalytic and therefore biological activities of Eph receptors are controlled by a two-component inhibitory mechanism, which is released by phosphorylation of the juxtamembrane and activation segment tyrosine residues. abnormal migration of axon tracts in the brain (37, 39). Ephrin-induced retraction of exploratory actin filopodia has also been described in vivo in migrating Eph receptor-expressing neural crest cells (29). Eph receptors and ephrins thus appear to mediate contact-dependent repulsive guidance of migrating cells and axons in culture and in vivo. The complementary expression of Eph receptors and their cognate ligands in adjacent domains in the developing embryo suggests that these proteins may also be involved in cell sorting and boundary formation (17). In support of this hypothesis, Eph receptorephrin signaling is able to modulate both cell-cell and cellsubstrate attachment (4, 48). Furthermore, bidirectional Eph receptor-ephrin signaling (5, 24) appears important for the formation of boundaries between rhombomeres of the hind brain (34). Such cellular responses to Eph receptor stimulation indicate that these proteins may regulate signaling events which control cytoskeletal architecture and cell adhesion functions. The N-terminal extracellular region of all Eph family members contains a domain necessary for ligand binding and specificity, followed by a cysteine-rich domain and two fibronectin type III repeats (21, 25, 30). The cytoplasmic region has a centrally located tyrosine kinase domain (15, 54). C-terminal to the catalytic region is a sterile alpha motif (SAM) domain, which forms dimers or oligomers in solution and may contribute to regulation of receptor clustering (46, 49). Localization or clustering of Eph proteins may also be influenced by PDZ domain effectors which potentially interact with specific C-terminal receptor motifs (22, 50). N-terminal to the kinase domain, in the juxtamembrane region, lie two invariant tyrosine residues (tyrosines 596 and 602 of EphA4; tyrosines 604 and 610 of EphB2) which are embedded in a characteristic and highly conserved ⬃10-aminoacid sequence motif. These tyrosine residues are major sites for autophosphorylation (9, 15, 27, 54) and have been shown to

The largest group of receptor protein-tyrosine kinases (RTKs), the Eph family, currently comprises 14 highly related vertebrate members and includes receptors in Caenorhabditis elegans and Drosophila (18, 42, 53). Eph RTKs are activated by a second family of cell surface-anchored proteins, the ephrins, which are attached to the plasma membrane either via a glycosylphosphatidylinositol linkage (A class) or a transmembrane sequence (B class) (12, 25). Receptors are also divided into A and B classes corresponding to their ligand binding specificities and phylogenetic relationships (16, 17). In general, A class receptors bind A class ephrins, whereas B class ephrins stimulate B class receptors. One exception to this rule is EphA4, a receptor which can bind and respond to B as well as A class ephrins (17). The observations that both native receptors and ligands are associated with the cell surface and that membrane attachment of ephrins is necessary for efficient receptor stimulation suggest that these proteins control contactdependent signaling processes between adjacent cells (25). Oligomerization rather than dimerization is apparently required for receptor stimulation, as soluble ligand ectodomains do not efficiently activate receptors unless multiply clustered by crosslinking (12). Many Eph family members are prominently expressed in the developing nervous system (3, 7, 20, 31, 40), and ephrin stimulation of growing primary axons in vitro results in axonal retraction or repulsion, characterized by a collapse of actinrich growth cone structures at the leading edge of the cell (14, 32, 33). Consistent with these data, mice bearing homozygous null mutations in EphA8 or in both EphB2 and EphB3 exhibit * Corresponding author. Mailing address: Samuel Lunenfeld Research Institute, Mt. Sinai Hospital, 600 University Ave., Toronto, Ontario, Canada M5G 1X5. Phone: (416) 586-8262. Fax: (416) 5868869. E-mail: [email protected]. † Present address: Rigel Inc., South San Francisco, CA 94080. 4791

4792

BINNS ET AL.

interact with a number of SH2 domain-containing cytoplasmic proteins including Ras GTPase-activating protein (RasGAP), the p85 subunit of phosphatidylinositol 3⬘ kinase, Src family kinases, the adapter protein Nck, and SHEP-1, which binds the R-Ras and Rap1A GTPases (13, 15, 23, 25, 38, 47, 54). It seems likely that signaling mediated by such SH2 domain binding partners may contribute to the physiological effects of Eph receptor stimulation on cell adhesion and cytoskeletal structures. As the primary effectors in ligand-induced signaling cascades, the enzymatic activity of RTKs is tightly controlled. Activation of RTKs generally follows a common scheme beginning with ligand-initiated oligomerization or reorientation of receptor chains, allowing intermolecular autophosphorylation of specific tyrosine residues. Tyrosine phosphorylation within the activation segment of the kinase domain can directly stimulate catalytic activity, while phosphorylation of adjacent noncatalytic elements typically creates docking sites for downstream signaling effectors (26, 41). To study the regulation of Eph receptor-mediated signaling, we have generated a cellular assay system in which ephrin stimulation of EphB2 in neuronal cells leads to neurite retraction and loss of actin-rich structures and cellular adhesion. This response is dependent on EphB2 catalytic activity and integrity of the juxtamembrane tyrosine residues. Substitution of these conserved tyrosine residues in full-length EphB2 leads to a reduction in ligand-induced kinase activity and ephrinstimulated tyrosine phosphorylation, suggesting that juxtamembrane tyrosines may serve a regulatory function in addition to acting as docking sites for downstream targets. To examine Eph receptor activation more carefully, we employed a continuous assay using the bacterially expressed cytoplasmic domain of EphA4. Our results indicate that the juxtamembrane region directly regulates catalytic activity, as does a tyrosine residue in the putative activation segment, thus suggesting a complex mechanism for complete Eph receptor kinase activation. MATERIALS AND METHODS Constructs and reagents. Full-length murine EphB2 cDNA was cloned into the mammalian expression vector pcDNA3 (Invitrogen). Site-directed mutagenesis of EphB2 was performed as previously described (23). The cDNA sequence corresponding to amino acids 586 to 986 (cytoplasmic region) of murine EphA4 (EphA4CYTO) was cloned into pGEX4T2 (Pharmacia). Mutagenesis of EphA4 was performed using the QuikChange system (Stratagene). The synthetic peptide (S-1) used for enzyme kinetics has the sequence GEEIYGEFD (amide at the carboxy terminus). Polyclonal EphB2 antiserum was raised against a glutathione S-transferase (GST) fusion protein expressing the C-terminal 94 amino acids of EphB2. For immunocytochemistry, the antiserum was GST extracted and affinity purified on an antigen column. Baculovirus-produced Fc-tagged ephrin-B1 extracellular domain was purified on a protein A-Sepharose (Pharmacia) column, eluted with low pH, and neutralized with 1 M Tris (pH 8.0). Concentrated protein was dialyzed against Tris-buffered saline. To achieve efficient receptor activation, Fc-ephrin-B1 was clustered using anti-human Fc antibodies (Jackson Immunoresearch) (12). Cell culture and retraction assay. NG108-15 cells were routinely cultured in Dulbecco’s modified Eagle’s medium (DMEM) containing 10% fetal calf serum and 1⫻ hypoxanthine-aminopterine-thymidine (HAT; Gibco). To produce stable cell lines, cells were transfected with Lipofectin (Gibco) and selected as described elsewhere (23). For the cellular assay, glass coverslips were prepared by coating with poly-L-lysine and fibronectin (both from Sigma and both at 100 ␮g/ml) for 3 h at room temperature followed by three washes in phosphatebuffered saline. Cells were seeded at 3 ⫻ 104 cells per coverslip in six-well dishes in DMEM containing 5% fetal calf serum, 0.5⫻ HAT, and 1 mM N6,O2-dibutyryladenosine 3⬘:5⬘-cyclic monophosphate (dibutyryl-cAMP; Sigma) (19) and cultured for 24 h to encourage neurite outgrowth. After stimulation with soluble clustered ephrin-B1 (2 ␮g/ml), cells were fixed for 15 min in 4% paraformaldehyde (pH 7.0), permeabilized in 0.1% Triton for 5 min, and washed extensively before blocking overnight in phosphate-buffered saline containing 5% bovine serum albumin. To localize EphB2 and filamentous actin, fixed cells were double labeled with affinity-purified polyclonal anti-EphB2 antibodies followed by fluorescein isothiocyanate-conjugated goat anti-rabbit secondary antibody (Sigma) and rhodamine-conjugated phalloidin (Sigma). Specificity of the EphB2 antibody

MOL. CELL. BIOL. was demonstrated by including 10-times-excess purified antigen in the primary antibody incubation. Cells were photographed at 100⫻ magnification on a Leitz DM RXE microscope. Immunoprecipitation and Western blotting. Cells were serum starved overnight in DMEM, lysed in PLC lysis buffer as previously described (20), and quantitated using a bicinchoninic acid protein assay kit (Pierce); protein concentrations between cell lines were equalized. Either lysates were immunoprecipitated with crude anti-EphB2 serum as previously described (20) or, for cytoplasmic lysates, extracts were mixed with an equal volume of 2⫻ sodium dodecyl sulfate (SDS) loading buffer, boiled, and loaded directly onto the gel. Proteins were subjected to polyacrylamide gel electrophoresis (PAGE), transferred to a polyvinylidene difluoride membrane (Millipore), blotted with antiEphB2 or affinity purified polyclonal antiphosphotyrosine antibodies, and visualized using a linear enhanced chemiluminescent (ECL) substrate (ECL Plus; Amersham) according to the manufacturer’s instructions. Blots were exposed and quantitated on a Storm PhosphorImager. In vitro kinase (IVK) reactions. EphB2 was immunoprecipitated, washed twice in HNTG and twice in KRB (25 mM HEPES [pH 7.5]; 2.5 mM MgCl2, 4 mM MnCl2), and then incubated with 5 ␮g of acid-denatured enolase in the presence of 5 ␮Ci of [␥32P]ATP at 37°C for 30 min. Reaction products were electrophoresed on SDS–12% polyacrylamide gels, which were fixed, dried, and exposed to a PhosphorImager screen. Incorporation into enolase was quantitated using a Storm Phosphorimager. Bacterial expression and purification of GST fusion proteins. GST-EphA4CYTO constructs were transformed into Escherichia coli BL21, and protein expression was induced overnight at room temperature. Cell pellets were lysed in 20 mM Tris-HCl (pH 7.5)–0.5 M NaCl–1 mM EDTA–1 mM dithiothreitol (DTT) with protease inhibitors in a EmulsiFlex-C5 homogenizer (Avestin), and the fusion protein was isolated on glutathione-Sepharose (Pharmacia) according to standard protocols. Dephosphorylation (75 U of alkaline phosphatase [AP; Boehringer Mannheim]/ml of glutathione beads) and thrombin cleavage (10 U/mg of fusion protein) were carried out concurrently and to completion for 12 h at 4°C. To remove thrombin and AP, the supernatant containing cleaved EphA4CYTO was diluted 1,000-fold in 10 mM HEPES (pH 7.5)–10 mM MgCl2 and passed over an ATP-Sepharose column (Pharmacia). After extensive washing with 10 mM HEPES (pH 7.5), EphA4CYTO was eluted with a linear gradient of buffered KCl. Peak fractions were pooled, concentrated, and passed over a Sephadex 200 gel filtration column (Pharmacia) equilibrated in 10 mM HEPES (pH 7.5)–500 mM KCl–1 mM DTT. Peak fractions were again pooled, concentrated to ⬃10 mg/ml, and stored at ⫺20°C. Preliminary experiments showed that two cycles of freeze-thawing had no effect on stability of catalytic activity. All experiments were performed with thrombin-cleaved EphA4CYTO. Spectrophotometric coupling assay. Kinetic analysis of bacterial EphA4CYTO was performed using a coupled IVK assay where production of ADP is coupled to the oxidation of NADH through pyruvate kinase and lactate dehydrogenase (1). The 100-␮l reaction volume contained 1 U of lactate dehydrogenase, 1 U of pyruvate kinase, 1 mM phosphoenolpyruvate, 0.2 mM NADH, and 2 mM ATP unless otherwise noted (in 20 mM MgCl2–0.2 mM DTT–60 mM HEPES [pH 7.5]–20 ␮g of bovine serum albumin/ml). Preliminary experiments ensured that concentration of kinase was the rate-limiting component. Concurrent reactions were followed at 340 nm on a Hewlett-Packard 845-UV-Visible 7-Cell system. Where indicated, kinase was preincubated in 2 mM ATP–10 mM MgCl2 at room temperature for 1 h unless specified otherwise. For accuracy, protein concentration was determined by UV spectrometry at 280 nm using molar coefficients. Enzyme kinetics. Kinetic constants were determined under pseudo-singlesubstrate conditions using the general rate equation of Alberty for multisubstrate systems (43): V ⫽ Vmax[AX][BX]/KmB[AX] ⫹ Kmax[B] ⫹ [AX][B] ⫹ (KsAX) (KmB), where Vmax is the maximal velocity when AX and B are at saturating concentration, Km is the concentration that gives rise to 1/2 Vmax when the second substrate is saturating, and KsAX is the dissociation constant for E ⫹ AX3 EAX. At saturating concentrations of B, the general equation simplifies to a pseudo-single-substrate system as follows: V ⫽ Vmax/(1 ⫹ KmAX/[B]) ⫽ VMAX [AX]/[AX] ⫹ Kmax (the Michaelis-Menten equation). Kinetic constants were determined from a nonlinear least squares best fit of the data. Hanes plots of 1/[S] plotted against [S]/[V] were used to illustrate results. Nano-ESI-MS/MS-based phosphopeptide mapping. Purified EphA4CYTO was autophosphorylated in the presence of 10 mM MgCl2–1 mM ATP and trypsin digested to completion at 37°C for 3 h using a 50:1 ratio of protein to protease. Phosphopeptides were desalted using ZipTip desalting columns (Millipore) equilibrated in 5% formic acid, washed with equilibration buffer, and eluted with 5% formic acid–60% methanol. Tandem mass spectrometry (MS/MS) analysis was carried out on an orthogonal QqTOF mass spectrometer (PE-Sciex) with a nano-electrospray ion (ESI) source (Protana A/S). Following identification of tryptic ions of interest, product ion spectra were generated by collisionally induced dissociation. For product ion scans, collision energy was determined experimentally. Sequencing was performed using PeptideScan (EMBL).

RESULTS Kinase activity and juxtamembrane tyrosine residues are required for EphB2-mediated neurite retraction. To develop

VOL. 20, 2000

REGULATION OF Eph RECEPTORS

4793

FIG. 1. Ligand activation of EphB2 in NG108 cells results in neurite retraction. Parental (A⬘ and B⬘) and WT or mutant EphB2-transfected (C to M; D⬘ to M⬘) NG108 cells were differentiated with dibutyryl-cAMP and left untreated (columns 1 and 3) or challenged with 2 ␮g of soluble clustered Fc-ephrin-B1 per ml for 10 min (columns 2 and 4). Fixed cells were double stained for EphB2 (green; C to M) and filamentous actin (red [rhodamine-phalloidin]; A⬘ to M⬘). (A⬘ and B⬘) NG108 cells; (C to M and D⬘ to M⬘) NG108 clones stably expressing WT (C to E, D⬘, and E⬘), KDIIM (F, F⬘, G, and G⬘), YJX1⫹2F (H, H⬘, I, and I⬘), YJX1F (J, J⬘, K, and K⬘), or YJX2F (L, L⬘, M, and M⬘) forms of EphB2. (C) Specific EphB2 staining was competed with an excess of immunizing peptide. Arrows represent bundled actin filopodia.

an assay for EphB2 function, we have established NG108-15 cell lines stably expressing wild-type (WT) EphB2 (NGEphB2WT cells [23]). NG108-15 cells display characteristics of motor neurons (19, 36), a cell type in which EphB2 is expressed

in the developing embryo (20). They do not, however, endogenously express EphB2 or respond biochemically to stimulation with B class ephrins and therefore provide a negative background for testing EphB2 function (23). In differentiated

4794

BINNS ET AL.

MOL. CELL. BIOL.

FIG. 2. EphB2 and EphA4 mutant constructs. (A) Schematic of EphA4 and EphB2 structures depicting juxtamembrane tyrosine residues JX1 (Y596 of EphA4/Y604 of EphB2), JX2 (Y602 of EphA4/Y610 of EphB2), activation segment tyrosine YACT (Y779 of EphA4), and the invariant lysine residue in kinase subdomain II (KDII; K661 of EphB2). Tyrosine residues were replaced with phenylalanine and KDII was replaced with methionine to produce mutant proteins as shown. The portion of the receptor used in EphA4CYTO is marked with a bracket. (B) Alignment of EphA4 and EphB2 amino acid sequences in the juxtamembrane and kinase subdomain regions.

neurite-bearing NG-EphB2WT cells, EphB2 protein is detected throughout the cell body and projections but is particularly concentrated distally in actin-rich filopodial and growth cone structures (Fig. 1D), as evidenced by double staining with rhodamine-phalloidin (Fig. 1D⬘). The specificity of the EphB2 antibody was established by competing EphB2 staining with the immunizing antigen (Fig. 1C). After stimulation of NGEphB2WT cells with clustered Fc-ephrin-B1, polymerized actin structures were disassembled and neurites retracted, accompanied by cell rounding and loss of substrate attachment (Fig. 1E and E⬘), reminiscent of Eph receptor responses previously observed in primary neurons (14, 32, 33). In contrast, parental NG108 cells were morphologically unaffected by ephrin-B1 stimulation (Fig. 1A⬘ and B⬘). WT EphB2 protein became relocalized to punctate structures upon ephrin-B1 stimulation, possibly indicative of receptor clustering and internalization (Fig. 1E). Using the NG108 cell assay, we evaluated the importance of specific EphB2 residues for the morphological response to ephrin stimulation. Mutant EphB2 receptors with amino acid substitutions in the juxtamembrane or kinase domains (Fig. 2A) were expressed in parental NG108 cells. These variant proteins were localized indistinguishably from WT EphB2 in unstimulated cells (Fig. 1F, H, J, and L). In contrast to NGEphB2WT cells, cells expressing a kinase-inactive form of

EphB2 containing a substitution in the conserved lysine in kinase subdomain II, which is critical for phosphotransfer activity (EphB2KDIIM), retained fully elaborated neurites and prominent microspikes after ephrin-B1 stimulation (Fig. 1G and G⬘). A similar lack of cytoskeletal remodeling was observed in cells expressing an EphB2 variant (EphB2YJX1⫹2F) in which both conserved juxtamembrane tyrosines, Y604 (JX1) and Y610 (JX2), were replaced with phenylalanine (Fig. 1I and I⬘). For both of these mutants, EphB2 protein appeared to relocalize proximally, becoming concentrated in the base rather than the tip of filopodia (Fig. 1G and data not shown). Cells expressing EphB2 with single substitutions in one or other of the juxtamembrane tyrosine residues exhibited a partial phenotype. Ephrin stimulation of cells expressing receptors mutated at the first juxtamembrane tyrosine (EphB2YJX1F) resulted in loss of short bundled actin filopodia, which were replaced by long trailing retraction fibers, but only an incomplete retraction of neurite length (Fig. 1K and K⬘). Little change in morphology was noted upon stimulation of EphB2YJX2F cells; neurites remained extended, and most cells retained well-formed filopodia (Fig. 1M and M⬘). These results suggest that EphB2 catalytic activity and juxtamembrane tyrosine residues, particularly JX2, are required for the morphological response of a neuronal cell to ephrin-B1 stimulation,

VOL. 20, 2000

and we performed biochemical experiments to further investigate the molecular basis for these observations. EphB2YJX2F substitution reduces EphB2 and p62dok tyrosine phosphorylation. We have previously found that the scaffolding protein p62dok becomes prominently tyrosine phosphorylated in NG-EphB2WT cells upon EphB2 activation (23). Ephrin-B1-induced tyrosine phosphorylation of both EphB2 and p62dok was abrogated in cells expressing the double juxtamembrane mutant or EphB2KDIIM (23) (data not shown). In contrast, mutation of the first juxtamembrane tyrosine, JX1, had only a minor effect on ephrin-B1-induced tyrosine phosphorylation of EphB2, whereas ligand-stimulated tyrosine phosphorylation of EphB2YJX2F was more notably reduced at both time points studied (30 and 60 min) (Fig. 3A to C). Similarly, while ligand-induced tyrosine phosphorylation of p62dok was almost normal in cells expressing EphB2YJX1F, mutation of JX2 virtually abolished ephrin-B1-dependent phosphorylation of the scaffolding protein (Fig. 3B). These differences might be due to failure of the mutant receptors to physically associate with downstream targets. Alternatively, EphB2 catalytic activity might be directly regulated by juxtamembrane tyrosine residues (2, 54). Juxtamembrane mutations regulate ligand-induced EphB2 kinase activity. To distinguish between these two possibilities, IVK assays were performed using WT and mutant EphB2 proteins. First, immunoprecipitated EphB2 receptors from unstimulated cells were assessed for the ability both to autophosphorylate and to phosphorylate an exogenous substrate, enolase. Mutations in the juxtamembrane region did not lead to a significant reduction in incorporation into enolase in comparison to EphB2KDIIM, which was essentially inactive for enolase phosphorylation (Fig. 4A and reference 23). In contrast, autophosphorylation of the EphB2 receptor was markedly reduced in EphB2YJX1⫹2F, YJX1F, and YJX2F, potentially due in part to loss of tyrosine phosphorylation sites. Ephrin-B1 stimulation of NG-EphB2WT cells led to a rapid increase in EphB2 IVK activity, evidenced both by receptor autophosphorylation and exogenous substrate phosphorylation (Fig. 4B). In contrast to basal levels, mutations in the juxtamembrane region differentially affected ligand-induced catalytic function. The EphB2YJX1⫹2F receptor, although retaining basal catalytic activity (Fig. 4A; also see reference 23), was impaired in ligand-induced kinase activation (Fig. 4B and C). This effect was primarily due to substitution of the JX2 site (Y610) since the single EphB2YJX2F mutant was also deficient in ephrin-induced kinase activity, whereas the EphB2YJX1F receptor exhibited only a small reduction in catalytic stimulation compared to WT at early time points. These results suggest that the juxtamembrane tyrosine residues of EphB2 may play a role in modulating ephrin-induced catalytic activation. To address the mechanistic basis for such regulation, and whether A as well as B class receptors may be subject to similar control, we turned to an in vitro system employing the highly related receptor EphA4 (Fig. 2). Bacterially expressed EphA4 is tyrosine phosphorylated in the juxtamembrane motif and kinase activation segment. The cytoplasmic region of EphA4, from the juxtamembrane region to the C terminus, was expressed in bacteria as a GST fusion protein (GST-EphA4CYTO) and became autophosphorylated in E. coli. The isolated GST fusion protein was cleaved with thrombin to release the EphA4CYTO polypeptide, which was purified to homogeneity (see Materials and Methods). We used nano-ESI-MS/MS to map phosphorylation sites within the thrombin-cleaved EphA4CYTO protein (Fig. 5 and 6). Initial spectra (MS1) of the peptide mixture following trypsin digestion of EphA4CYTO indicated ⬃70% representation of


4795

FIG. 3. Juxtamembrane mutations reduce tyrosine phosphorylation of EphB2 and p62dok in response to ephrin-B1 stimulation. NG-EphB2 clones expressing WT or mutant receptors (as indicated) were stimulated with 2 ␮g of soluble clustered Fc-ephrin-B1 per ml. Anti-EphB2 immunoprecipitates (A) or cytoplasmic lysates (B) were electrophoresed and blotted with antibodies to phosphotyrosine (anti-pTyr) (A and B, top panels), stripped and reprobed with anti-EphB2 (lower panels), and developed using a linear ECL system followed by PhosphorImager scanning. Ephrin-B1-induced phosphorylation of EphB2 (A and B) and p62dok (B) is reduced in EphB2YJX2F cells. An unidentified tyrosine phosphorylated protein of ⬃70 kDa is absent from immunoprecititates of EphB2YJX1F and EphB2YJX2F receptors (arrow in panel A). (C) Graphical representation of EphB2 tyrosine phosphorylation in panel B, measured with a PhosphorImager.

the predicted tryptic peptides. Ion peaks corresponding to the juxtamembrane region peptide in the unphosphorylated and the singly and doubly phosphorylated states were then identified (Fig. 5A). Fragmentation of these parent ions produced secondary spectra (MS2) allowing peptide sequencing, thus confirming peptide identity (Fig. 5B) and specific phosphorylation of tyrosines 596 (JX1 [Fig. 5C]) and 602 (JX2 [Fig. 5D]) of EphA4. Identical results were obtained using protein phosphorylated in bacteria or in an IVK reaction. These results are consistent with previous phosphopeptide mapping experiments using EphA4 (15). A similar approach was used to determine whether a conserved tyrosine in the putative activation segment, a 5- to 20-amino-acid region that in other tyrosine kinases interacts structurally and functionally with the catalytic cleft, may be phosphorylated. Ion peaks corresponding to the tryptic peptide containing the activation segment tyrosine Y779 (YACT) in the phosphorylated and unphosphorylated forms were identified (Fig. 6A). MS2 sequencing confirmed phosphorylation of Y779 (Fig. 6B and C). In addition to the sites above, phosphorylation

4796

BINNS ET AL.

MOL. CELL. BIOL.

FIG. 4. Juxtamembrane mutations regulate ligand-induced EphB2 kinase activity. Lysates of parental NG108 (NG) cells or clones expressing WT or mutant EphB2 proteins (as indicated) were immunoprecipitated (IP) with anti-EphB2 or preimmune (PI) serum. The catalytic activity of the immunoprecipitated proteins was assessed using an IVK assay with enolase as an exogenous substrate. (A) Uninduced cells. (B) WT and mutant clones were stimulated with soluble clustered Fc-ephrin-B1 (2 ␮g/ml) for 0, 5, 15, 30, and 60 min (represented by open triangles) prior to lysis, immunoprecipitation, and IVK reaction. EphB2 expression in the clones was determined by blotting untreated samples for EphB2 (bottom panel) and developing using a linear ECL system. (C) Graphical representation of the mean fold increase in 32P incorporation into enolase measured from three representative experiments performed as for panel B. Incorporation was quantitated with a PhosphorImager.

of Y798 and Y841 was observed (data not shown), consistent with phosphopeptide mapping of EphB2 and EphB5 (27). EphA4CYTO requires an induction time before reaching maximum activity. To investigate the influence of the catalytic domain and juxtamembrane tyrosine residues on EphA4 activation, we performed a coupled IVK assay using a peptide substrate (coupling assay). By coupling the conversion of ATP to ADP with the oxidation of NADH, this assay allows continuous monitoring of a kinase reaction spectrophotometrically (1). Due to the high homology between the catalytic domains of Eph receptors and Src family kinases, a peptide (S-1; see Materials and Methods) based on Src family kinase specificity (45) was selected. Initial IVK reactions and MS analysis confirmed that S-1 was a suitable substrate for EphA4. Using the coupling assay, an intrinsic EphA4CYTO ATPase activity was

observed at ATP concentrations above 5 mM, which was insignificant compared to consumption of ATP by peptide phosphorylation (data not shown). Starting the kinase assay with purified, dephosphorylated protein (Fig. 7A, i and ii), we observed a nonlinear progress curve of peptide phosphorylation with two somewhat distinct phases; a low initial rate of NADH oxidation of 0.56 (⌬A340/s)/ ␮mol of kinase followed by a higher steady-state rate of 2.12 (⌬A340/s)/␮mol of kinase (Fig. 7A, iii). The time required to reach a steady-state activity (the lag or induction time) was obtained from the graph by estimation of the bisection of the asymptotes of the two distinct phases. In several other tyrosine kinases, including Src and the fibroblast growth factor receptor, it has been shown that this lag period corresponds to the time required for autophosphorylation of the regulatory tyro-

FIG. 5. EphA4CYTO is autophosphorylated at both juxtamembrane tyrosines JX1 and JX2. (A) A nano-ESI-MS (MS1) of in vitro-phosphorylated EphA4CYTO depicting the doubly charged ion peaks 958.4, 998.4, and 1,038.4 corresponding to the tryptic peptide TY596VDPFTY602EDPNQAVR (monoisotopic weight ⫽ 1,916.8 atomic mass units [amu] [M]) in the unphosphorylated (M), singly phosphorylated (M ⫹ P), and doubly phosphorylated (M ⫹ 2P) states, respectively (P ⫽ phosphate [amu]; q ⫽ charge). (B) nano-ESI-MS (MS2) product ion spectra generated by collisionally induced dissociation of the unphosphorylated 958.4 ion. Singly charged peptide ion peaks, differing in mass by one amino acid, are marked by arrows (y series, C-terminal portions of peptide fragments, boldface arrows; b series, N-terminal portions of peptide fragments, lightface arrows). (C and D) MS2 product ion spectra of the 998.4 ion peak with single phosphorylation on Y596 (C) or Y602 (D). An increase in mass difference of 80 amu (corresponding to one phosphate) is observed between the singly charged ions b1 and b2 (C) and between y8 and y9 (D) compared to the corresponding ion peaks in the unphosphorylated spectrum (B). Ion peaks representing the fragments of the tryptic peptide that do not contain the phosphorylated tyrosine (pTyr) (b1 in C and y8 in D) are identical between the spectra.

VOL. 20, 2000


4797

4798

BINNS ET AL.

MOL. CELL. BIOL.

VOL. 20, 2000

ine(s) in the activation segment (1, 10, 11, 28, 35). To investigate whether the lag time seen with EphA4 also reflects a required phosphorylation event, and to ensure that this phenomenon was not due to any of the coupling assay components, we varied the concentrations of all reaction mixture constituents and examined their effects on the induction time. Increasing EphA4CYTO concentration led to a decrease in lag time, suggesting an intermolecular event (Fig. 7B, i). However, of the other assay components, only ATP concentration affected the rate of peptide phosphorylation or the induction time. Increases in ATP concentration resulted in a decrease in induction time (Fig. 7B, ii). Furthermore, an immediate steady-state reaction was obtained both upon preincubation of the dephosphorylated EphA4CYTO with MgATP (Fig. 7B, iii) and by using EphA4CYTO which had not been subjected to prior phosphatase treatment (data not shown). Together, these results provide strong evidence that the induction time reflects a phosphorylation event(s) which is required for maximal kinase activity. EphA4YACTF exhibits decreased kinase activity but maintains a lag time before complete activation. To determine if the lag period could be attributed more directly to phosphorylation of Y779, this residue was substituted with phenylalanine within the EphA4CYTO protein. If phosphorylation of this residue plays a role in regulating catalytic activity, this substitution (YACTF) should affect the overall rate of the kinase reaction. Furthermore, if the induction time seen with WT EphA4CYTO is due solely to autophosphorylation of this residue, we would not expect to observe a lag period before full EphA4YACTF activity. The overall specific activity and kcat of the YACTF mutant protein were ⬃6-fold lower than those of WT EphA4CYTO (Fig. 8A; Table 1), suggesting an inhibitory role for this tyrosine prior to receptor phosphorylation. Surprisingly, the EphA4YACTF protein still exhibited a lag phase before reaching maximal (although reduced) velocity. This suggested that multiple independent events might be required to achieve full catalytic activity. The induction time of EphA4YACTF was dependent on kinase and ATP concentrations, indicating that an intermolecular phosphorylation event in addition to that involving Y779 might be important in regulating receptor activity (Fig. 7C, i and ii). Both lack of phosphatase treatment (data not shown) and preincubation of the EphA4YACTF protein with ATP (Fig. 7C, iii) abolished the lag time, resulting in an immediate steady-state velocity, strongly supporting a phosphorylation event as the basis for this additional level of control. From the results obtained in the NG108 cellular assay with EphB2, we hypothesized that this regulatory phosphorylation could involve one or both of the conserved juxtamembrane tyrosines. Using MS we determined that phosphorylation of both the juxtamembrane region and the activation segment occurs during the lag phase in the WT protein (data not shown). Tyr-to-Phe changes in the juxtamembrane motif inhibit kinase activity. To investigate this possibility, tyrosines JX1 and JX2 of EphA4 were replaced with phenylalanine, and the mutant EphA4CYTO proteins were introduced into the coupling assay. Both EphA4YJX1F and EphA4YJX2F exhibited a decreased ability to phosphorylate the peptide substrate (Fig. 8A;


4799

Table 1), with the largest decrease resulting from substitution of JX2, consistent with the results seen for full-length EphB2 isolated from neuronal cells. The double Tyr-to-Phe mutant EphA4YJX1⫹2F displayed the most significant (10-fold) decrease in catalytic activity (Fig. 8) under the conditions tested. Surprisingly, no induction was observed, with the mutant lacking both juxtamembrane tyrosines with a steady-state velocity being maintained over 2 h. Preincubation with ATP had no effect on the progress curve (Fig. 7D), nor did lack of phosphatase treatment (data not shown). To ensure that the juxtamembrane region was not an essential adjunct of the kinase domain, directly involved in the kinase reaction, an EphA4CYTO protein lacking the juxtamembrane region (EphA4JX-) was expressed. EphA4JX- protein exhibited a catalytic activity virtually identical to that of WT EphA4CYTO (Fig. 8) but had a shorter lag time than WT kinase. Determination of kinetic parameters. To investigate the mechanism of catalytic regulation by both the juxtamembrane and putative activation loop tyrosines, kinetic analysis was performed on the kinase reactions of WT and mutated EphA4 proteins. For ease of comparison, kinetic constants for the multisubstrate reaction were determined by varying a single substrate concentration while holding the second substrate at a fixed, saturating concentration, creating a pseudo-single-substrate mechanism. Results are summarized in Table 1. Representative Hanes plots demonstrating the alterations in the Michaelis constants for ATP (KmATP) and peptide (KmPEP) are illustrated in Fig. 9A and B, respectively. Calculations of constants was performed by nonlinear least squares fitting to the velocity substrate curves directly. Substitution of the activation loop tyrosine resulted in a 200-fold increase in KmPEP, suggesting a significant decrease in affinity for the peptide prior to phosphorylation of YACT. Single juxtamembrane tyrosine mutants also showed up to a 50-fold increase in KmPEP, with the biggest increase seen for EphA4YJX2F, consistent with the differences seen in their catalytic activities. Calculated KmPEP values for the double tyrosine mutant displayed significant variability, likely due to the relatively small increases in activity with increasing peptide concentrations. Estimations, however, were as high as 25 mM. Attempts to achieve more significant changes by increasing the range of peptide concentration were hampered by solubility of the peptide. Similar analysis was performed to investigate the effects of all substitutions on the affinity for ATP. No corresponding increases in KmATP were observed. Only EphA4YACTF gave a relatively small but reproducible increase of twofold compared to WT EphA4CYTO. This suggests that binding of ATP and peptide are independent and that inhibition, or conversely, activation by phosphorylation, of both activation loop and juxtamembrane regions arises through a decrease in the binding affinity for the phosphoacceptor without affecting the affinity for ATP. DISCUSSION The activation of signaling pathways by RTKs generally follows a common framework, beginning with binding of extracellular ligands, receptor oligomerization or reorientation, and

FIG. 6. EphA4CYTO is phosphorylated at the putative activation loop tyrosine YACT. (A) A nano-ESI-MS (MS1) spectrum indentifying doubly charged ion peaks corresponding to the unphosphorylated (740.3) and phosphorylated (780.3) forms of the tryptic peptide VLEDDPEAY779TTR (1480.6 amu). (B and C) MS2 spectra of the phosphorylated and unphosphorylated ions illustrating the 80 amu (one phosphate) increase in mass of singly charged peptides (arrows) retaining YACT in the 780.3-derived spectrum (B) versus the 740.3-derived spectrum (C). Peak intensity ratios of phosphorylated and unphosphorylated ions (panel A and Fig. 5A) are not quantitative due to differing ionization potentials and desalting-column elution and retention profiles between peptides.

4800

FIG. 7. EphA4CYTO requires an induction time for maximum catalytic activity. (A) i, Sephadex 200 size exclusion column fractions were subjected to SDS-PAGE and stained with Coomassie blue to analyze protein purity. Fractions 10 and 11 were routinely concentrated and used for subsequent analysis. ii, equivalent amounts of bacterially expressed EphA4CYTO, were thrombin cleaved with (⫹) and without (⫺) the presence of AP, purified, electrophoresed, and blotted with antibodies to phosphotyrosine (anti-pTyr), illustrating the complete dephosphorylation upon AP treatment (described in Materials and Methods). Protein concentration was determined by UV spectrometry using molar coefficients and confirmed by SDS-PAGE iii, progress curve of the phosphorylation of 0.3 mM S-1 peptide in a standard reaction mixture containing 0.5 mM ATP and 0.2 ␮M dephosphorylated EphA4CYTO depicting the initial and steady-state phases of a typical reaction. Induction time was estimated from the intersection of asymptotes as illustrated. (B) The induction time is dependent on kinase and ATP concentrations. Reactions were performed with 0.15 mM S-1 peptide under standard conditions and WT EphA4CYTO concentrations from 0.36 ␮M to 4.8 ␮M in 0.5 mM ATP (i) or ATP concentrations from 0.5 to 2 mM with 0.24 ␮M WT EphA4CYTO (ii). Induction times were calculated from progress curves as illustrated in panel A, iii. iii, progress curves of the phosphorylation of 0.15 mM S-1 peptide in standard reaction conditions with (■) or without (Œ) 1 h of preincubation of the WT EphA4CYTO in 2 mM ATP–10 mM MgCl2. Final concentration of kinase was 3.6 ␮M. (C) Dependence of the induction time of EphA4YACTF catalytic activity on kinase concentration from 0.7 to 4 ␮M at 0.5 mM ATP (i) and on ATP concentration from 0.5 to 2 mM with 3.4 ␮M EphA4YACTF (ii) iii, progress curves of the phosphorylation of 1.2 mM S-1 peptide with (■) or without (Œ) 1 h of preincubation of the EphA4YACTF in 2 mM ATP–10 mM MgCl2. (D) Progress curves of the phosphorylation of 0.31 mM S-1 peptide by 0.76 ␮M EphA4YJX1⫹2F kinase with (■) or without (Œ) 2 h of preincubation of the kinase in 2 mM ATP–10 mM MgCl2.

4801

4802

BINNS ET AL.

FIG. 8. Comparison of reaction velocities of the EphA4CYTO WT and mutant proteins. (A) Comparison of specific activities of EphA4CYTO WT and mutant proteins. Reactions were performed with 0.1 mM S-1 peptide and 10.8 ␮M preincubated kinase using WT, mutant, and juxtamembrane-deleted (JX-) EphA4CYTO protein as indicated. Velocities are the means of three separate determinations. (B) Progress curves of EphA4CYTO WT ({) and (JX-) EphA4CYTO (}) illustrating the reduced lag time observed upon truncation of the juxtamembrane region. Reactions were performed as described for panel A.

consequent trans phosphorylation by the cytoplasmic kinase domain (reviewed in reference 26). The resulting autophosphorylation sites generally fall into two classes: those involved in the regulation of kinase activity (primarily in the activation segment of the catalytic domain), and those that serve as docking sites for cytoplasmic signaling molecules containing SH2 or phosphotyrosine binding (PTB) domains (generally in noncatalytic elements). The extent to which Eph receptors conform to this scheme is not well defined, and indeed they exhibit several unusual features that may pertain to their regulation, such as slow kinetics of ligand activation in cells, interaction with monovalent membrane-bound ligands, and the presence of intrinsic oligomerizing regions. Regulation of in vitro Eph receptor catalytic activity by multiple phosphorylation events. To characterize the role of tyrosine phosphorylation in Eph receptor function, we have used two complementary approaches, a ligand-inducible system to examine the involvement of specific EphB2 residues in the cellular response to ephrin stimulation, and a coupled IVK assay to analyze direct effects on EphA4 catalytic activity. To identify EphA4 autophosphorylation sites, we carried out phosphopeptide mapping of bacterially expressed protein using nano-ESI-MS/MS. This approach revealed that a conserved tyrosine within the activation segment of the kinase domain (reviewed in reference 26) is a site for autophosphorylation and also confirmed the phosphorylation of the two conserved juxtamembrane tyrosines.

MOL. CELL. BIOL.

To investigate the involvement of the activation loop and juxtamembrane tyrosines in Eph receptor activation, we used a continuous coupled kinase assay for peptide phosphorylation by the EphA4 cytoplasmic region. Although the use of a truncated receptor dissociates the kinase domain from possible constraints imposed by the extracellular and transmembrane regions, the coupling assay facilitated a detailed kinetic analysis of the kinase reaction. Using dephosphorylated EphA4CYTO, we observed a lag time preceding maximal kinase activity and found that this reflects a requirement for autophosphorylation. The effects of SAM domain-mediated oligomerization were also investigated; although oligomerization may influence receptor activation, we have not detected changes in activation kinetics of either EphA4CYTO or full-length EphB2 bearing SAM domain truncations or mutations in the SAM dimer interface. Substitution of YACT with phenylalanine, however, resulted in a significant drop in activity, suggesting that phosphorylation of this residue might be required to activate EphA4CYTO or release an inhibition imposed on the unstimulated receptor. Kinetic analysis revealed that, similar to Src and the fibroblast growth factor receptor 1 kinases (35, 44, 51), a decrease in the binding affinity for substrate is the basis for this inhibition while ATP binding is largely unaffected by the phosphorylation state of YACT. Interestingly, EphA4YACTF still appeared to require autophosphorylation for full activation. Although previous studies using constitutively active cellular overexpression systems or in vitro experiments omitting an initial dephosphorylation step (9, 15) have not detected an effect of juxtamembrane substitutions on the catalytic activity of Eph receptors, our results using full-length, ligand-stimulated EphB2 suggested that the juxtamembrane region can indeed influence kinase activity (discussed below). To examine this further, we investigated the role of the corresponding EphA4 residues in kinase activation in vitro. The EphA4YJX1⫹2F mutant, in which the two conserved juxtamembrane tyrosine residues are replaced with phenylalanine, had a significantly impaired catalytic activation indicating that the juxtamembrane region can inhibit kinase activity, although it is not required for catalytic activity per se. Kinetic analysis showed that the mutated juxtamembrane region also inhibits kinase activity by preventing peptide substrate acquisition rather than ATP binding. These results suggest an inhibitory role for the juxtamembrane tyrosines that is relieved upon phosphorylation. This model is further supported by our observation that the EphA4JX- protein which lacks the juxtamembrane region does not appear to require a commensurate lag period for full WT activity. The proposed inhibitory role of these residues in Eph receptors is reminiscent of observations made for the insulin-like growth factor 1 and platelet-derived growth factor receptors, where

TABLE 1. kcat and Km values of ATP and S-1 peptide for WT and mutant EphA4CYTO proteinsa Protein

WT YACTF YJX1F YJX2F YJX1⫹2F

Km (mM) ATP

S-1 peptide

kcat (dA340/mm/mM)

0.15 ⫾ 0.01 0.40 ⫾ 0.07 0.17 ⫾ 0.03 0.19 ⫾ 0.01 0.14 ⫾ 0.03

0.01 ⫾ 0.002 2.00 ⫾ 0.02 0.08 ⫾ 0.03 0.51 ⫾ 0.02 NA

719 ⫾ 46.6 178.4 ⫾ 14.8 302.9 ⫾ 18.8 117.2 ⫾ 41.4 NA

a Reactions were performed as described in the legend to Fig. 9. NA indicates values that were not available due to solubility (see text). Values are means of at least four determinations. Errors reported are standard deviations.

FIG. 9. (A) Representative Hanes plots ([S] versus [S/V]) of the substrate concentrations and velocities derived from a single determination of KmATP values for preincubated EphA4CYTO WT (Œ), YACTF (■), YJX1F (}), YJX2F (F), and YJX1⫹2F (E). ATP concentration was varied from 0.1 to 2.0 mM. Peptide concentration was held fixed at 8.15 mM for YACTF and YJX1⫹2F and 2 mM for WT, YJX1F, and YJX2F proteins based on calculations of peptide Km and maximum solubility. (B) Representative Hanes plots derived from KmPEP for preincubated EphA4CYTO WT (Œ), YACTF (■), YJX1F (}), and YJX2F (F). Peptide concentrations were varied from 0.012 to 1.5 mM. ATP concentrations were held fixed at 2 mM.

4803

4804

BINNS ET AL.

analogous juxtamembrane mutations result in decreased catalytic activity via a 10-fold decrease in peptide affinity and a lack of ligand response, respectively (2, 6). An unexpected feature of the progress curves of EphA4YJX1⫹2F was the lack of activation despite the presence of the activation loop tyrosine, with the kinase appearing to remain in a basal state. One possible explanation for this lack of induction (or extended lag phase) is that phosphorylation of YACT is partially occluded by the unphosphorylated juxtamembrane region, which upon phosphorylation may undergo a conformational change, thus allowing kinase access to the active site. Alternatively, phosphorylation of Y596/604 might aid or induce receptor oligomerization, resulting in trans phosphorylation of Y779. Experiments to order the phosphorylation events and to determine whether juxtamembrane phosphorylation is a prerequisite for YACT-based activation are under way. EphB2 juxtamembrane tyrosine residues are required for kinase activation and cytoskeletal responses in an ephrin-B1stimulated neuronal cell line. Our experiments using EphB2transfected NG108 cells suggest that substitution of juxtamembrane tyrosine residues with phenylalanine preferentially influences ligand-induced as opposed to basal kinase catalytic activity. We have previously demonstrated (23) and confirm in Fig. 4A that substitution of tyrosine residues in the juxtamembrane region of the full-length EphB2 protein causes only a slight (⬃2-fold) decrease in basal catalytic activity compared to WT receptor. Similar reductions in kinase activity were observed with yeast Lex-A-EphB2 cytoplasmic domain fusion proteins bearing juxtamembrane mutations (54). In comparison, substitution of the phosphotransfer lysine in EphB2KDIIM abolished catalytic activity as predicted. In contrast to the data obtained using unstimulated EphB2, experiments with ligand-activated receptors showed very different trends for WT versus mutant EphB2. Ephrin activation of WT EphB2 led to an approximately fivefold increase in catalytic activity, whereas receptors in which tyrosine JX2 or both JX2 and JX1 were substituted with phenylalanine had markedly impaired ligand activation kinetics. We also found that substitution of juxtamembrane tyrosine residues, especially JX2, had profound effects on the biological activity of EphB2 in a neuronal cell line. Upon ephrin stimulation, WT EphB2-expressing cells showed a striking loss of actin-rich structures and started to detach from the substrate. However, cells expressing EphB2KDIIM or EphB2YJX1⫹2F essentially failed to respond morphologically to ligand stimulation, retaining elongated neurites with well-bundled polymerized actin microspikes, and maintaining substrate attachment. Cells expressing EphB2YJX2F were similarly compromised in their response to ephrin stimulation. Extrapolating from the biochemical and kinetics experiments, it appears likely that the failure of the EphB2YJX2F and EphB2YJX1⫹2F mutants to induce a biological response may be due not only to the loss of these specific SH2 domain binding sites but also to impaired catalytic activity which could influence autophosphorylation of additional tyrosines on the receptor. Decreased Eph receptor catalytic function might also lead to diminished tyrosine phosphorylation of receptor substrates. NG108 cells expressing the JX2 and JX1⫹2 mutants of EphB2 exhibit reduced tyrosine phosphorylation of p62dok, a component of the ephrin-stimulated signaling cascade (8, 23, 52). This may directly reflect the compromised catalytic activity of EphB2, although we cannot exclude the possibility that p62dok phosphorylation requires the docking of an SH2 domain-containing signaling intermediate (e.g., RasGAP) to the receptor. In NG108 cells, JX1 mutations in full-length EphB2 cause a

MOL. CELL. BIOL.

partial defect in the neurite retraction response. We have shown that ligand-induced kinase activity and tyrosine phosphorylation of EphB2YJX1F approximates WT levels, and this residue has only a minor effect on catalytic activity in the coupling assay. It is therefore likely that the reduction in biological function seen in EphB2YJX1F in part represents loss of a docking site for a cytoplasmic signaling partner(s). Consistent with this notion, tyrosine JX1 has previously been shown to regulate cellular attachment responses and activation of JNK by EphB1, potentially through direct interaction with the adapter protein Nck (47). In our experiments, the loss of a ⬃70-kDa tyrosine-phosphorylated protein was observed in EphB2 immunoprecipitates from NG-EphB2YJX1F cells, which might also correspond to an associated effector. In the present study, we have used complementary cellular and in vitro approaches to investigate the role of conserved tyrosine phosphorylation sites in the regulation of catalytic and biological activity of Eph RTKs. Our data are consistent with models for other RTKs where phosphorylation of the active loop tyrosine controls kinase access to the peptide substrate. We have also demonstrated that the conserved juxtamembrane tyrosine-based motif is critical for biological responses to ephrin stimulation and appears to have two distinct functions: not only to provide docking sites for signaling effectors, but also to contribute to the intrinsic regulation of catalytic activity. ACKNOWLEDGMENTS We thank Sarang Kulkarni and Jerry Gish for helpful discussions and Renping Zhou for the EphA4 cDNA. This work was supported by a grant from the Medical Research Council of Canada (MRC) and a Howard Hughes Medical Institute International Research Scholar Award to T.P. S.J.H. was supported by a postdoctoral fellowship from the MRC, and K.L.B. was supported by a Bank of Montreal predoctoral Fellowship in Medical Research. T.P. is a Distinguished Scientist of the MRC. REFERENCES 1. Barker, S. C., D. B. Kassel, D. Weigl, X. Huang, M. A. Luther, and W. B. Knight. 1995. Characterization of pp60c-src tyrosine kinase activities using a continuous assay: autoactivation of the enzyme is an intermolecular autophosphorylation process. Biochemistry 34:14843–14851. 2. Baxter, R. M., J. P. Secrist, R. R. Vaillancourt, and A. Kazlauskas. 1998. Full activation of the platelet-derived growth factor beta-receptor kinase involves multiple events. J. Biol. Chem. 273:17050–17055. 3. Becker, N., T. Seitanidou, P. Murphy, M. G. Mattei, P. Topilko, M. A. Nieto, D. G. Wilkinson, P. Charnay, and H. P. Gilardi. 1994. Several receptor tyrosine kinase genes of the Eph family are segmentally expressed in the developing hindbrain. Mech. Dev. 47:3–17. 4. Bohme, B., T. VandenBos, D. P. Cerretti, L. S. Park, U. Holtrich, H. Rubsamen-Waigmann, and K. Strebhardt. 1996. Cell-cell adhesion mediated by binding of membrane-anchored ligand LERK-2 to the EPH-related receptor human embryonal kinase 2 promotes tyrosine kinase activity. J. Biol. Chem. 271:24747–24752. 5. Bruckner, K., E. B. Pasquale, and R. Klein. 1997. Tyrosine phosphorylation of transmembrane ligands for Eph receptors. Science 275:1640–1643. 6. Cann, A. D., S. M. Bishop, A. J. Ablooglu, and R. A. Kohanski. 1998. Partial activation of the insulin receptor kinase domain by juxtamembrane autophosphorylation. Biochemistry 37:11289–11300. 7. Carpenter, M. K., H. Shilling, T. VandenBos, M. P. Beckmann, D. P. Cerretti, J. N. Kott, L. E. Westrum, B. L. Davison, and F. A. Fletcher. 1995. Ligands for EPH-related tyrosine kinase receptors are developmentally regulated in the CNS. J. Neurosci. Res. 42:199–206. 8. Carpino, N., D. Wisniewski, A. Strife, D. Marshak, R. Kobayashi, B. Stillman, and B. Clarkson. 1997. p62 (dok): a constitutively tyrosine-phosphorylated, GAP-associated protein in chronic myelogenous leukemia progenitor cells. Cell 88:197–204. 9. Choi, S., and S. Park. 1999. Phosphorylation at Tyr-838 in the kinase domain of EphA8 modulates Fyn binding to the Tyr-615 site by enhancing tyrosine kinase activity. Oncogene 18:5413–5422. 10. Cobb, M. H., B. C. Sang, R. Gonzalez, E. Goldsmith, and L. Ellis. 1989. Autophosphorylation activates the soluble cytoplasmic domain of the insulin receptor in an intermolecular reaction. J. Biol. Chem. 264:18701–18706. 11. Cunningham, M. E., R. M. Stephens, D. R. Kaplan, and L. A. Greene. 1997.

VOL. 20, 2000

12.

13. 14.

15.

16. 17.

18. 19. 20.

21. 22.

23.

24. 25.

26. 27. 28.

29.

30. 31.

Autophosphorylation of activation loop tyrosines regulates signaling by the TRK nerve growth factor receptor. J. Biol. Chem. 272:10957–10967. Davis, S., N. W. Gale, T. H. Aldrich, P. C. Maisonpierre, V. Lhotak, T. Pawson, M. Goldfarb, and G. D. Yancopoulos. 1994. Ligands for EPHrelated receptor tyrosine kinases that require membrane attachment or clustering for activity. Science 266:816–819. Dodelet, V. C., C. Pazzagli, A. H. Zisch, C. A. Hauser, and E. B. Pasquale. 1999. A novel signaling intermediate, SHEP1, directly couples Eph receptors to R-Ras and Rap1A. J. Biol. Chem. 274:31941–31946. Drescher, U., C. Kremoser, C. Handwerker, J. Loschinger, M. Noda, and F. Bonhoeffer. 1995. In vitro guidance of retinal ganglion cell axons by RAGS, a 25 kDa tectal protein related to ligands for Eph receptor tyrosine kinases. Cell 82:359–370. Ellis, C., F. Kasmi, P. Ganju, E. Walls, G. Panayotou, and A. D. Reith. 1996. A juxtamembrane autophosphorylation site in the Eph family receptor tyrosine kinase, Sek, mediates high affinity interaction with p59fyn. Oncogene 12:1727–1736. Eph Nomenclature Committee. 1997. Unified nomenclature for Eph family receptors and their ligands, the ephrins. Cell 90:403–404. Gale, N. W., S. J. Holland, D. M. Valenzuela, A. Flenniken, L. Pan, T. E. Ryan, M. Henkemeyer, K. Strebhardt, H. Hirai, D. G. Wilkinson, T. Pawson, S. Davis, and G. D. Yancopoulos. 1996. Eph receptors and ligands comprise two major specificity subclasses and are reciprocally compartmentalized during embryogenesis. Neuron 17:9–19. George, S. E., K. Simokat, J. Hardin, and A. D. Chisholm. 1998. The VAB-1 Eph receptor tyrosine kinase functions in neural and epithelial morphogenesis in C. elegans. Cell 92:633–643. Hamprecht, B. 1977. Structural, electrophysiological, biochemical, and pharmacological properties of neuroblastoma-glioma cell hybrids in cell culture. Int. Rev. Cytol. 49:99–170. Henkemeyer, M., L. E. Marengere, J. McGlade, J. P. Olivier, R. A. Conlon, D. P. Holmyard, K. Letwin, and T. Pawson. 1994. Immunolocalization of the Nuk receptor tyrosine kinase suggests roles in segmental patterning of the brain and axonogenesis. Oncogene 9:1001–1014. Himanen, J. P., M. Henkemeyer, and D. B. Nikolov. 1998. Crystal structure of the ligand-binding domain of the receptor tyrosine kinase EphB2. Nature 396:486–491. Hock, B., B. Bohme, T. Karn, T. Yamamoto, K. Kaibuchi, U. Holtrich, S. Holland, T. Pawson, H. Rubsamen-Waigmann, and K. Strebhardt. 1998. PDZ-domain-mediated interaction of the Eph-related receptor tyrosine kinase EphB3 and the ras-binding protein AF6 depends on the kinase activity of the receptor. Proc. Natl. Acad. Sci. USA 95:9779–9784. Holland, S. J., N. W. Gale, G. D. Gish, R. A. Roth, Z. Songyang, L. C. Cantley, M. Henkemeyer, G. D. Yancopoulos, and T. Pawson. 1997. Juxtamembrane tyrosine residues couple the Eph family receptor EphB2/Nuk to specific SH2 domain proteins in neuronal cells. EMBO J. 16:3877–3888. Holland, S. J., N. W. Gale, G. Mbamalu, G. D. Yancopoulos, M. Henkemeyer, and T. Pawson. 1996. Bidirectional signalling through the EPH-family receptor Nuk and its transmembrane ligands. Nature 383:722–725. Holland, S. J., E. Peles, T. Pawson, and J. Schlessinger. 1998. Cell-contactdependent signalling in axon growth and guidance: Eph receptor tyrosine kinases and receptor protein tyrosine phosphatase ␤. Curr. Opin. Neurobiol. 8:117–127. Hubbard, S. R. 1999. Structural analysis of receptor tyrosine kinases. Prog. Biophys. Mol. Biol. 71:343–358. Kalo, M. S., and E. B. Pasquale. 1999. Multiple in vivo tyrosine phosphorylation sites in EphB receptors. Biochemistry 38:14396–14408. Kendall, R. L., R. Z. Rutledge, X. Mao, A. J. Tebben, R. W. Hungate, and K. A. Thomas. 1999. Vascular endothelial growth factor receptor KDR tyrosine kinase activity is increased by autophosphorylation of two activation loop tyrosine residues. J. Biol. Chem. 274:6453–6460. Krull, C. E., R. Lansford, N. W. Gale, A. Collazo, C. Marcelle, G. D. Yancopoulos, S. E. Fraser, and F. M. Bronner. 1997. Interactions of Ephrelated receptors and ligands confer rostrocaudal pattern to trunk neural crest migration. Curr. Biol. 7:571–580. Labrador, J. P., R. Brambilla, and R. Klein. 1997. The N-terminal globular domain of Eph receptors is sufficient for ligand binding and receptor signaling. EMBO J. 16:3889–3897. Lhotak, V., P. Greer, K. Letwin, and T. Pawson. 1991. Characterization of


4805

elk, a brain-specific receptor tyrosine kinase. Mol. Cell. Biol. 11:2496–2502. 32. Meima, L., I. J. Kljavin, P. Moran, A. Shih, J. W. Winslow, and I. W. Caras. 1997. AL-1-induced growth cone collapse of rat cortical neurons is correlated with REK7 expression and rearrangement of the actin cytoskeleton. Eur. J. Neurosci. 9:177–188. 33. Meima, L., P. Moran, W. Matthews, and I. W. Caras. 1997. Lerk2 (ephrinB1) is a collapsing factor for a subset of cortical growth cones and acts by a mechanism different from AL-1 (ephrin-A5). Mol. Cell. Neurosci. 9:314– 328. 34. Mellitzer, G., Q. Xu, and D. G. Wilkinson. 1999. Eph receptors and ephrins restrict cell intermingling and communication. Nature 400:77–81. 35. Mohammadi, M., J. Schlessinger, and S. R. Hubbard. 1996. Structure of the FGF receptor tyrosine kinase domain reveals a novel autoinhibitory mechanism. Cell 86:577–587. 36. Nelson, P., C. Christian, and M. Nirenberg. 1976. Synapse formation between clonal neuroblastoma ⫻ glioma hybrid cells and striated muscle cells. Proc. Natl. Acad. Sci. USA 73:123–127. 37. Orioli, D., M. Henkemeyer, G. Lemke, R. Klein, and T. Pawson. 1996. Sek4 and Nuk receptors cooperate in guidance of commissural axons and in palate formation. EMBO J. 15:6035–6049. 38. Pandey, A., D. F. Lazar, A. R. Saltiel, and V. M. Dixit. 1994. Activation of the Eck receptor protein tyrosine kinase stimulates phosphatidylinositol 3-kinase activity. J. Biol. Chem. 269:30154–30157. 39. Park, S., J. Frisen, and M. Barbacid. 1997. Aberrant axonal projections in mice lacking EphA8 (Eek) tyrosine protein kinase receptors. EMBO J. 16: 3106–3114. 40. Pasquale, E. B., T. J. Deerinck, S. J. Singer, and M. H. Ellisman. 1992. Cek5, a membrane receptor-type tyrosine kinase, is in neurons of the embryonic and postnatal avian brain. J. Neurosci. 12:3956–3967. 41. Pawson, T. 1995. Protein modules and signalling networks. Nature 373: 573–580. 42. Scully, A. L., M. McKeown, and J. B. Thomas. 1999. Isolation and characterization of Dek, a Drosophila eph receptor protein tyrosine kinase. Mol. Cell. Neurosci. 13:337–347. 43. Segel, I. H. 1975. Enzyme kinetics: behaviour and analysis of rapid equilibrium and steady state enzyme systems. John Wiley & Sons, New York, N.Y. 44. Sicheri, F., I. Moarefi, and J. Kuriyan. 1997. Crystal structure of the Src family tyrosine kinase Hck. Nature 385:602–609. 45. Songyang, Z., K. L. Carraway III, M. J. Eck, S. C. Harrison, R. A. Feldman, M. Mohammadi, J. Schlessinger, S. R. Hubbard, D. P. Smith, C. Eng, M. J. Lorenzo, B. A. J. Ponder, B. J. Mayer, and L. C. Cantley. 1995. Catalytic specificity of protein-tyrosine kinases is critical for selective signalling. Nature 373:536–539. 46. Stapleton, D., I. Balan, T. Pawson, and F. Sicheri. 1999. The crystal structure of an Eph receptor SAM domain reveals a mechanism for modular dimerization. Nat. Struct. Biol. 6:44–49. 47. Stein, E., D. U. Huynh, A. A. Lane, D. P. Cerretti, and T. O. Daniel. 1998. Nck recruitment to Eph receptor, EphB1/ELK, couples ligand activation to c-Jun kinase. J. Biol. Chem. 273:1303–1308. 48. Stein, E., A. A. Lane, D. P. Cerretti, H. O. Schoecklmann, A. D. Schroff, R. L. Van Etten, and T. O. Daniel. 1998. Eph receptors discriminate specific ligand oligomers to determine alternative signaling complexes, attachment, and assembly responses. Genes Dev. 12:667–678. 49. Thanos, C. D., K. E. Goodwill, and J. U. Bowie. 1999. Oligomeric structure of the human EphB2 receptor SAM domain. Science 283:833–836. 50. Torres, R., B. L. Firestein, H. Dong, J. Staudinger, E. N. Olson, R. L. Huganir, D. S. Bredt, N. W. Gale, and G. D. Yancopoulos. 1998. PDZ proteins bind, cluster, and synaptically colocalize with Eph receptors and their ephrin ligands. Neuron 21:1453–1463. 51. Xu, W., S. C. Harrison, and M. J. Eck. 1997. Three-dimensional structure of the tyrosine kinase c-Src. Nature 385:595–602. 52. Yamanashi, Y., and D. Baltimore. 1997. Identification of the Abl- and rasGAP-associated 62 kDa protein as a docking protein, Dok. Cell 88:205–211. 53. Zhou, R. 1998. The Eph family receptors and ligands. Pharmacol. Ther. 77: 151–181. 54. Zisch, A. H., M. S. Kalo, L. D. Chong, and E. B. Pasquale. 1998. Complex formation between EphB2 and Src requires phosphorylation of tyrosine 611 in the EphB2 juxtamembrane region. Oncogene 16:2657–2670.