INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Cell adhesion events, coordinated both spatially and temporally, are critical to the development and maintenance of the tissue structures of an organism. Interactions with other cells and extracellular matrix components govern the growth, differentiation, migration, and ultimate location and function of a given cell. An increasing number of proteins expressed at the cell surface have been implicated in regulating cell adhesion and movement within specific developmental contexts. For example, considerable progress has been made in identifying cell surface proteins which regulate neurite outgrowth and pathfinding during nervous system development (9). These proteins include classic cell adhesion molecules such as the neural cell adhesion molecules of the immunoglobulin superfamily (58, 69), members of the cadherin superfamily (20), and integrins (47), as well as a growing number of other protein families including collapsins/semaphorins, neuropilins, netrin receptors (10, 11, 57), various peptide growth factor receptors (49), and receptor-type tyrosine phosphatases (13, 42). These cell surface proteins respond to molecular guidance cues which can arise locally or diffuse from distant sources. In one such case, control of growth cone physiology is achieved at the most intimate level of cell-to-cell contact, when neuronal cell surface proteins interact with attractive or repulsive cues present on the surface of a neighboring cell.

The Eph family of receptor tyrosine kinases and their membrane-associated activating ligands, the ephrins, are recent additions to this growing repertoire of cell surface proteins which mediate cell-to-cell interactions. They are highly expressed in the developing and adult nervous system and have been implicated in regulating cell adhesion events in both neural and nonneural tissues including neural crest cell migration (43, 62, 71), hindbrain segmentation (4, 5, 23, 75), angiogenesis (55, 72), and somitogenesis (19). At least 14 distinct Eph receptors have been identified across vertebrate and invertebrate species and are classified as type A or B depending on the ligand that they bind. Eph receptors are characterized by an intracellular catalytic domain and an extracellular cysteine-rich domain, two fibronectin type III repeats, and an amino-terminal ephrin-binding region (45, 46, 80). To date, all eight ephrins identified are membrane-bound proteins, possessing either a glycosylphosphatidylinositol anchor (ephrin A subclass, which binds to type A Eph receptors) or a transmembrane domain (ephrin B subclass, which binds to type B Eph receptors). While there is little cross-reactivity between class A and B molecules, promiscuous receptor-ligand interactions occur within each class (22).

In addition to numerous receptor-ligand-binding studies, elucidation of both partially overlapping and completely reciprocal expression patterns has indicated that various Eph receptors and ephrins regulate axon targeting, cell migration, and pattern formation by promoting repulsive and attractive intercellular interactions as well (60). Indeed, the potential dual nature of their interactions is thought to be a growing paradigm among cell surface proteins that play multiple roles during development (7, 24, 68). Ephrins activate their cognate Eph receptors most efficiently when they are membrane bound or presented in an artificially clustered state which mimics membrane attachment (22). Hence, the cell-to-cell communication they facilitate is intimately linked to cell-to-cell contact. Receptor-ligand activation studies have further indicated that in general, multimeric aggregation of an Eph receptor is optimal for its activation. Interestingly, the EphB-type receptors have a higher requirement for aggregation of their cognate ephrin B ligands (22). In fact, only a multimeric complex of ephrin B1 could induce the recruitment of a downstream phosphatase molecule to activated EphB1 receptors and concommitantly promote cell attachment (65).

The ephrin B ligands are type I transmembrane proteins which themselves are thought to transduce an intracellular signal(s) upon Eph receptor binding. The carboxy-terminal 33 amino acids of the three known ephrin B molecules are nearly identical, suggesting a signaling function. This region also contains five conserved tyrosine residues (21). Upon contact with the extracellular domain of a cognate Eph receptor, ephrin B molecules induce receptor activation and autophosphorylation on tyrosine and become tyrosine phosphorylated themselves, suggesting a bidirectional mode of signaling (33). Genetic evidence has revealed that the EphB2 receptor extracellular domain may send a reverse signal through the ephrin B molecules to which it binds (30). Deletion of the kinase domain of the EphB2 receptor did not affect its function in guiding anterior comissure axons expressing ephrin B1. EphB2 is expressed in adjacent cells which serve as a migration substrate. While ephrin B molecules become phosphorylated on tyrosine during embryogenesis (6), the mechanisms by which this modification occurs are unknown. As with activated Eph receptors, phosphorylated tyrosine residues in the ephrin cytoplasmic domain may interact with other signaling molecules. Although phosphorylation of the ephrin B1 cytoplasmic domain was detected in fibroblast cells stimulated with platelet-derived growth factor (PDGF), in cells overexpressing c-src, and in cells exposed to clustered EphB2 ectodomain (6, 33), this modification had no effects on cellular physiology. Thus, until now, the physiological significance of ephrin phosphorylation has remained unclear.

We previously identified and characterized a Xenopus laevis homologue of ephrin B1, XLerk (37). The 95% homology of its cytoplasmic domain to that of mammalian family members further suggested an important evolutionarily conserved function for this domain. While also highly expressed in developing neural structures including the hindbrain and retina, x-ephrin B1 (XLerk) expression was elevated early in development during gastrulation. This indicated a possible role in regulating cell adhesion during the rearrangement of embryonic blastomere cells into developmentally specific germ layers. Ectopic overexpression of x-ephrin B1 in Xenopus embryos caused a marked dissociation of embryonic cells just prior to gastrulation (in the earlier blastula stage), confirming that x-ephrin B1 could indeed modify cell adhesion (38). x-Ephrin B1 not only is an activating ligand for an Eph receptor but also exhibits the ability to signal and regulate cell adhesion.

Xenopus embryos provide a useful expression system to examine x-ephrin B1 function both biochemically and morphologically. Previously, we have found that treatment with fibroblast growth factor (FGF) reversed the cell-dissociative effect of x-ephrin B1 in Xenopus embryonic tissue explants (38). From this work, we hypothesized that the FGF receptor may modulate the function of x-ephrin B1 in regulating cell adhesion. Here, we demonstrate that FGF receptor activation inhibits x-ephrin B1-induced cell dissociation in Xenopus embryos. We show that the FGF receptor associates with and induces the phosphorylation of x-ephrin B1. We also define critical residues within a conserved region of the intracellular domain necessary for this interaction. In neural tissue, treatment of retina with FGF induces tyrosine phosphorylation of endogenous ephrin B. Our results provide the first example of a functional and physical interaction between the FGF tyrosine kinase family of receptors and the ephrin B family of cell surface ligands and implicate their cross talk in regulating adhesion.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Plasmids and mutagenesis of cDNA. cDNA encoding wild-type x-ephrin B1 was subcloned into three plasmids as follows: HindIII and BamHI restriction sites of plasmid pSP64A as previously described (38); EcoRI and NotI restriction sites of plasmid pSP64Aten (a gift from Malcolm Whitman and D. Melton, Harvard University); and EcoRI and NotI restriction sites of pSP64Aten(HindIII), a derivative of pSP64A in which the HindIII site has been abolished by restriction enzyme digestion, Klenow fill-in, and religation reactions. Carboxy-terminal truncation mutants of wild-type x-ephrin B1 were generated by PCR with plasmid pSP64A x-ephrin B1 as the template and the sense primer 5'-GGACGGCTTCTTCAACTCC-3', in conjunction with antisense primers that generated a stop codon at the appropriate positions. PCR products were purified with the agarose gel DNA extraction kit as specified by the manufacturer (Boehringer Mannheim), digested with PacI and EcoRV, repurified, and ligated into the PacI and EcoRV sites of pSP64Aten(HindIII) x-ephrin B1. Single-point mutants were also generated by the PCR method. Mutants WTY305F and WTY310F were made with pSP64Aten x-ephrin B1 as a template. The double mutant WTY305FY310F was made with WTY305F as a template to generate a second mutation at position 310. PCR products were purified, digested with PacI and SacI, repurified, and subcloned into the PacI and SacI sites of pSP64A x-ephrinB1. DNA sequencing was performed to confirm the identity of each mutant.

Preparation of synthetic RNA, microinjection of Xenopus embryos, and animal cap explant assay. For all embryo injections and manipulations, Xenopus embryos were obtained by in vitro fertilization and capped mRNA was synthesized and injected into embryos at the two-cell stage as previously described (38). All mRNA was made by using either the SP6 mMessage mMachine or T7 Megascript kit as specified by the manufacturer (Ambion). pSP64A plasmids containing x-ephrin B1 constructs were linearized with EcoRI, and pSP64Aten(HindIII) plasmids were linearized with XbaI. FGF receptor constructs in pCS2+ were linearized with NotI, and pSP64T3 containing FGFR1C289R/K420A was linearized with BamHI. pSP64AtenMekACT and pGEMtorsoACT were linearized with EcoRI, and pGEMHE2PDGFRa was linearized with NheI.

Fusion proteins. To generate the soluble fusion proteins x-EphB1-Fc and x-ephrin B1-Fc, cDNAs encoding the extracellular domains of Xenopus EphB1 and x-ephrin B1 were subcloned into the mammalian expression vector pCMV-Ig (a gift from Pantelis Tsoulfas, University of Miami), which creates a chimeric protein with the Fc domain of human IgG. Each plasmid was transfected into 293T cell monolayers with Lipofectamine as specified by the manufacturer (Life Technologies, Gaithersburg, Md.). At 1 day posttransfection, cells were grown in Optimem medium containing 2% serum with a low Ig content (GibcoBRL). The medium was harvested 5 to 7 days later and incubated with protein A-Sepharose (GibcoBRL) at 4°C for 4 to 16 h. Fusion proteins were eluted with 100 mM glycine (pH 3.0) and then neutralized to pH 7.5 with 1 M Tris (pH 11). Fusion proteins were assessed for purity and concentration by SDS-PAGE and Coomassie blue staining.

Antibodies. A rabbit polyclonal antibody, anti-x-ephrin B1, was generated to the extracellular domain of x-ephrin B1 by using purified x-ephrin B1-Fc as the immunogen. A rabbit polyclonal antibody was also generated to the carboxy terminus of x-ephrin B1 (anti-Lerk2) by using the C-terminal 14-amino-acid peptide conjugated to keyhole limpet hemocyanin. Anti-Lerk2 was also purchased from Santa Cruz Inc. Antibodies to torso and PDGF receptors were gifts from D. Morrison and M. Mercola, respectively. A monoclonal antibody to the FGF receptor has been described previously (53). Antibody specific for phosphomapkinase was purchased from New England Biolabs. Antiphosphotyrosine antibody 4G10 was purchased from UBI. Anti-glutathione S-transferase antibody was purchased from Santa Cruz Inc. Goat anti-human IgG was purchased from Boehringer Manheim.

Immunoprecipitation and Western blot analysis. Embryonic lysates were prepared with cold lysis buffer (20 mM Tris, 150 mM NaCl, 1% Nonidet P-40, 2 mM EDTA, 5 mM sodium orthovanadate, 1 mM aprotinin, 1 mM phenylmethylsulfonyl fluoride, 0.02 mM leupeptin) as previously described, followed by extraction with Freon (Sigma) at a 1:1 (vol/vol) ratio. All immunoprecipitations were conducted on 20 to 30 embryo equivalents with the indicated antibodies at 4°C for 4 to 16 h. All immunoprecipitates were washed in RIPA buffer (lysis buffer adjusted to 1% sodium dodecyl sulfate [SDS] and 0.5% deoxycholate) except for those described in Fig. 3b and c, which were washed with lysis buffer. Anti-x-ephrin B1 antibody was used for immunoprecipitations, and anti-Lerk2 antibody was used for Western blot analysis. Two embryos equivalents were examined for direct lysate analysis. For Mek experiments, buffer A was lysis buffer and buffer B was mitogen-activated protein MAP kinase buffer (44). SDS-polyacrylamide gel electrophoresis (PAGE) (8.5% polyacrylamide) was used in all cases except for the Mek experiment, which required SDS-PAGE analysis with 15% polyacrylamide. Western blots with antiphosphotyrosine antibody were conducted with buffer (10 mM Tris [pH 8], 150 mM NaCl, 0.05% Tween 20) containing 2% goat serum, 1% fish gelatin, and 1% bovine serum albumin. All other blots used 5% milk. Proteins were visualized by using an appropriate secondary antibody coupled to horseradish peroxidase, followed by application of enhanced chemiluminescence reagents as specified by the manufacturer (Amersham). The blots were stripped and washed if probing with a second antibody was necessary.

Retina assays. Retinas were harvested from day 10 chicken embryos (Truslow Farms Inc., Chestertown, Mass.) as described elsewhere (32) for incubation with recombinant human basic FGF (Prepro Tech) at 250 ng/ml. The retinas were then solubilized with lysis buffer and centrifuged at 14,000 × g for 10 min at 4°C. The protein concentration of lysate samples determined by using the BCA protein kit and samples of equivalent total protein were incubated with protein A-Sepharose for 1 h at 4°C for preclearing. After removal of Sepharose by centrifugation, lysates were incubated with 5 µg of x-EphB1-Fc fusion protein that had been preincubated for 1 h at 4°C with protein A-Sepharose. Lysates were rotated in the cold for 16 h. After centrifugation to remove lysate, Sepharose beads were washed four times with lysis buffer. Proteins contained in the final pellet were analyzed by SDS-PAGE and Western blotting.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

FGF receptor activation causes x-ephrin B1 phosphorylation on tyrosine and inhibits cell dissociation. In Xenopus, the developmental effects of FGF on early embryonic tissue can be mimicked by treating ectodermal animal cap tissue explants from blastula stage embryos with exogenous FGF (39, 61). Thus, in this explant system, FGF receptor activation can be effectively initiated. Since we had observed that FGF treatment reversed the ability of x-ephrin B1 to dissociate embryonic cells in such a tissue explant (38), we inferred that FGF receptor activation was a likely corresponding event. To confirm that FGF receptor activation correlated with the inhibitory effects of FGF on x-ephrin B1 function, a constitutively activated form of Xenopus FGF receptor type I, FGFRK562E (KE) (53), was coexpressed with x-ephrin B1 in Xenopus embryos. Synthetic RNA (2.5 ng) encoding x-ephrin B1 was injected into cleaving embryos at the two-cell stage in the absence or presence of synthetic RNA (2.5 ng) encoding KE. Dissociation of embryonic cells was induced by x-ephrin B1 expression alone at the blastula stage of development (stage 8) (Fig. 1B). In contrast, coexpression of x-ephrin B1 with KE inhibited cell dissociation and generated embryos that were similar in appearance to uninjected control embryos (Fig. 1B). Expression of KE alone had no effect on whole embryos at this stage (not shown). When embryonic lysates were examined by Western blot analysis (Fig. 1A), the x-ephrin B1 (Lerk-2) expression level was the same whether expressed alone or coexpressed with KE, indicating that a reduction in x-ephrin B1 expression was not the cause of reduced cell dissociation. When lysates were examined either directly or by immunoprecipitation with an antiphosphotyrosine-specific antibody, x-ephrin B1 was not detectably phosphorylated when expressed alone but became highly phosphorylated in the presence of the activated FGF receptor (Fig. 1A). Neither cell dissociation nor the non-tyrosine-phosphorylated state of x-ephrin B1 was altered when x-ephrin B1 was coexpressed with the wild-type form of the FGF receptor or a truncated, kinase-deficient form of the FGF receptor lacking the cytoplasmic domain (Fig. 1D) (53). However, x-ephrin B1 was phosphorylated when coexpressed with the wild-type FGF receptor that was activated in response to exogenously added FGF (Fig. 2D). Taken together, x-ephrin B1 was tyrosine phosphorylated only when the FGF receptor was present in an activated state. Furthermore, this modification correlated with a block in the ability of x-ephrin B1 to induce cell dissociation, indicating that phosphorylation regulates this specific function of an ephrin B-type molecule.

View larger version (23632633K): [in this window] [in a new window]   FIG. 1.   Tyrosine phosphorylation of x-ephrin B1 by the FGF receptor and rescue of cell dissociation. Xenopus embryos were left uninjected (UN) or injected with RNA encoding x-ephrin B1, an activated form of the FGF receptor type 1 termed FGFR1-K562E (KE), or both. (A) Embryonic lysates were analyzed directly (top and middle panels) or immunoprecipitated with antibody to phosphotyrosine (IP Ab: anti-PY) (bottom panel) and separated by SDS-PAGE (8.5% polyacrylamide). Western blot analysis was conducted with antiphosphotyrosine antibody (Blot Ab: anti-PY), and the blot was stripped for reprobing with antibody to the carboxyl-terminus of x-ephrin B1 (Blot Ab: anti-Lerk2). Molecular size marker positions (kilodaltons) are indicated. Activated FGF receptor (95 kDa) and phosphorylated x-ephrin B1 (38 kDa) are indicated by arrows. (B) Embryonic cell adhesion was assessed at stage 9 of uninjected embryos and embryos expressing x-ephrin B1 alone or with activated FGFR1 (KE). Cell adhesion was disrupted only when x-ephrin B1 was expressed alone. (C) Embryos were injected with RNA encoding x-ephrin B1 alone or coinjected with RNA encoding either KE (FGFR1Act), PDGF receptor, or the TorsoAct receptor. Embryos expressing the PDGF receptor were incubated with PDGF (100 ng/ml). Receptors were immunoprecipitated from lysates with specific antibodies (IP Ab: anti-FGFR, anti-Torso, or anti-PDGFR), and these immune complexes were analyzed by SDS-PAGE and Western blotting with antiphosphotyrosine antibodies (Blot Ab: anti-PY) (top panel). Embryonic lysates were analyzed directly by SDS-PAGE and Western analysis with antiphosphotyrosine antibodies (Blot Ab: anti-PY) (middle panel) and anti-ephrinB1 antibodies (Blot Ab: anti-Lerk2) (bottom panel). Arrows indicate bands representing the different receptors and x-ephrin B1. (D) Embryos were injected with RNA encoding x-ephrin B1 alone or coinjected with RNA encoding either KE, the wild-type FGFR1 (WT), or a truncated, kinase-deficient FGFR1 (DN). Lysates were analyzed by SDS-PAGE and Western blotting by using the indicated antibodies (Blot Ab:) (Top panel, anti-phosphotyrosine antibodies; middle panel, anti-ephrin B1 antibodies; bottom panel, anti-FGF receptor antibodies). Arrows indicate the corresponding proteins. Note that tyrosine phosphorylation of x-ephrin B1 occurs with an activated FGFR.

The FGF receptor interacts with x-ephrin B1. Upon activation, the FGF receptor initiates the Ras-Raf-Mek-mitogen-activated protein (MAP) kinase signaling cascade (44). To test whether MAP kinase activation was required for x-ephrin B1 tyrosine phosphorylation, a constitutively activated form of Mek (27), which activates MAP kinase, was co-expressed with x-ephrin B1 in Xenopus embryos. Embryonic extracts were prepared with two different buffers, one that efficiently solubilizes ephrin (buffer A) (Fig. 2A, middle) and another that maximizes extraction of MAP kinase (buffer B) (Fig. 2A, bottom). Activation of MAP kinase did not alter the phosphorylation state or rescue the cell adhesion effects of x-ephrin B1 (Fig. 2A, top, and, data not shown), implying that an upstream event or alternate FGF receptor-mediated pathway is necessary for ephrin phosphorylation.

View larger version (40K): [in this window] [in a new window]   FIG. 2.   Lack of MAP kinase-induced tyrosine phosphorylation of x-ephrin B1, and existence of stable interaction between tyrosine-phosphorylated x-ephrin B1 and the FGF receptor. (A) Embryos were left uninjected (UN) or injected with RNA encoding x-ephrin B1, activated Mek (MekACT), or both. Embryos were lysed with the indicated buffers (buffer A, a lysis buffer that solublizes maximal amounts of ephrin B; buffer B, a buffer that maximally solubilizes activated MAP kinase) for Western blot analysis with antiphosphotyrosine antibody (Blot Ab: anti-PY) (top panel) and anti-Lerk2 antibodies (Blot Ab: anti-Lerk2) (middle panel) and antibody specific for phospho-MAP kinase (Blot Ab; anti-phosphomapk) (bottom panel). The arrows designate the appropriate protein. Note that x-ephrin B1 is not detectably tyrosine phosphorylated in the presence of activated MAP kinase. (B) Embryos were left uninjected (UN), injected with RNA encoding x-ephrin B1, FGFR1-K562E (KE), or a kinase-dead receptor FGFR1-C289R/K420A (CR/KA) or coinjected with RNA encoding x-ephrin B1 and KE or CR/KA. Lysates were mixed where indicated and immunoprecipitated with anti-FGF receptor antibody (indicated above panels as IP Ab: anti-FGFR). SDS-PAGE and Western blot analysis of immune complexes were conducted with the ephrin B1 antibodies (anti-Lerk2 [top panel]) or FGF receptor antibodies (anti-FGFR [bottom panel]). Arrows indicate the corresponding proteins. Note that only kinase active FGFR coprecipitates with x-ephrin B1. (C) Embryos were left uninjected (UN), injected with RNA encoding either x-ephrin B1, KE, or an activated form of FGF receptor 2, FGFR2-C332Y (CY), or coinjected with RNA encoding x-ephrin B1 and KE or CY. Lysates were mixed where indicated and immunoprecipitated with anti-FGF receptor antibody (IP Ab: anti-FGFR). SDS-PAGE and Western blot analysis of immunecomplexes was conducted with either antiphosphotyrosine antibodies (top panel) or anti-FGF receptor antibodies (bottom panel). Arrows indicate the appropriate protein. (D) Blastomere cells from injected embryos expressing x-ephrin B1 alone or in combination with the wild-type (WT) and/or kinase-deficient (DN) forms of the FGF receptor were exposed to either FGF or x-EphB2-Fc for 45 min prior to lysis and immunoprecipitation with a polyclonal antibody to the extracellular domain of x-ephrin B1 (IP Ab: anti-x-ephrinB1 [left panel]). Western blots were conducted on immunoprecipitations with antiphosphotyrosine antibodies (Blot Ab: anti-PY [left panel]). The control panels on the right show Western blot analysis of lysates before immunoprecipitations were performed with either FGF receptor-specific antibodies (top right panel) or ephrin B1-specific antibodies (bottom right panel).

The carboxy terminus of x-ephrin B1 interacts with the FGF receptor. Ephrin B molecules have a high degree of sequence homology in their cytoplasmic domains, particularly within their last 33 amino acids (21). To determine if this region interacted with the FGF receptor, a series of deletion mutations were made within the last 33 residues of x-ephrin B1 (Fig. 3a). Since this region also contained five tyrosine residues that are conserved among ephrin B molecules, the series of deletions generated sequential removal of these residues as well. Each mutant x-ephrin B1 molecule was coexpressed with the activated FGF receptor, and immunoprecipitations were performed with either x-ephrin-specific antibodies (Fig. 3b, top panels) or FGF receptor specific antibodies (bottom panels). When coexpressed with KE, 33 was not phosphorylated, and it did not coimmunoprecipitate with KE. While 26 showed weakly detectable phosphorylation, it did not associate with KE (data not shown), implying that a region(s) downstream of residue 301 was required for stable interaction. 19, 15, and 6 did coimmunoprecipitate with KE (Fig. 3B, third panel from top) and were phosphorylated (Fig. 3B, first and third panels from top), indicating that a region within residues 302 to 308 was important for interacting with the FGF receptor. Full-length ephrin proteins containing a mutation at either position 305 or 310 (tyrosine changed to phenylalanine) also coimmunoprecipitated with KE, and cell dissociation was rescued (data not shown) (62% and 65% respectively [Fig. 3d]). However, when both residues 305 and 310 were mutated simultaneously (WTY305FY310F), association with KE was markedly reduced (Fig. 3c, top left panel) and cell dissociation was not effectively rescued (15%; Fig. 3d). Moreover, mutation of position 305 in the context of the 19 truncation mutant (19Y305F), which lacks residue 310, also resulted in a reduced interaction with the FGF receptor (Fig. 3b, third panel from top). These results demonstrate that both tyrosine residues contained in the region from residues 302 to 310 are critical for stable FGF receptor interaction. Although 26, 19Y305F, and WTY305FY310F did not coimmunoprecipitate with the FGF receptor, all three mutants were phosphorylated. These results suggest that when overexpressed in cells, these mutants may transiently or weakly associate with overexpressed KE protein but cannot remain stably associated with the receptor without the presence of the region from residues 302 to 310 or without the presence of both tyrosine residues 305 and 310. 

View larger version (14432969K): [in this window] [in a new window]   FIG. 3.   Association of x-ephrin B1 mutants with the FGF receptor. (a) A series of X-ephrin B1 cDNAs were constructed to encode proteins with mutations in their cytoplasmic domains. The last 33 amino acids of wild-type (WT) x-ephrin B1 are shown. Numbers indicate the positions of the five conserved tyrosine residues (bold). (b) Embryos were coinjected with RNA encoding FGFR1-K562E (KE) in combination with each x-ephrin B1 mutant. Lysates were immunoprecipitated with either the x-ephrin-specific antibody (IP Ab: anti-x-ephrinB1 [top panels]) or the FGF receptor antibody (IP Ab: anti-FGFR [bottom panels]), transferred, and blotted with phosphotyrosine-specific antibodies (anti-PY) to visualize the phosphorylated ephrin mutants (first and third panels from top) or with the precipitating antibody (second and fourth panels from top). (c) Embryos were coinjected with RNA encoding either wild-type x-ephrin B1 (WT) or the WTY305FY310F double mutant and the FGFR1-K562E (KE). (Left panel) Lysates were immunoprecipitated with the FGF receptor-specific antibody (IP Ab: anti-FGFR), transferred, and probed with phosphotyrosine specific antibodies (Blot Ab: anti-PY) or FGF receptor antibodies (Blot Ab: anti-FGFR) as indicated. (right panel) Control lysates from coinjected embryos that were analyzed directly by Western blotting with antiphosphotyrosine (top right) or x-ephrin B1 (bottom right) antibodies. Note the reduction in association of WTY305FY310F and 19Y305F with FGF receptor in panels b and c. (d) Morphology of embryos coinjected with KE and wild-type x-ephrin B1 (WT), a tyrosine 305 mutant (WTY305F), a tyrosine 310 mutant (WTY310F), or a tyrosine 305 and 310 double mutant (WTY305FY310F). The graph indicates the percentage of embryos injected with the wild type or indicated mutants that display the dissociated or rescued phenotype. Error bars represent standard deviations, and the absence or presence of coinjected activated FGF receptor RNA is indicated as KE  or +. Note that double-mutant-induced cell dissociation is not effectively rescued by KE. Dissociation was scored by at least 15 to 20% of the animal pole region displaying the deadhesion phenotype. These are representative of three separate experiments and a total of 34 to 36 injected embryos for each sample.

FGF induces ephrin B1 phosphorylation in the chicken retina. Both the FGF and Eph receptor tyrosine kinase families and their ligands have been implicated in similar instructional developmental processes such as axon outgrowth and guidance. For example, during development of the vertebrate visual system, both receptor family members and ephrins regulate retinal axon pathfinding (49, 56, 70). In the developing vertebrate retina, inhibition of the FGF receptor impairs proper neurite outgrowth (50). In addition, Eph receptors and ephrins establish gradients along specific axes that later define the topographical map which migrating axons follow (18). This has been extensively studied in the chicken retina where ephrin B molecules act as repulsive cues to migrating axons which express EphB receptors (8, 17, 52, 63).

View larger version (45K): [in this window] [in a new window]   FIG. 4.   FGF stimulates tyrosine phosphorylation of ephrin molecules in retinal neurons. Embryonic chicken retina was incubated with basic FGF (250 ng/ml) at 37°C for increasing times. Lysates were made and incubated with x-EphB1-Fc (2 µg) for 4 h followed by protein A-Sepharose for 1 h at 4°C. The proteins were resolved by SDS-PAGE, and Western blot analyses were conducted with antiphosphotyrosine antibodies (Blot Ab: anti-PY [top panel]), antiephrin antibodies (anti-Lerk2 [middle panel]), or human IgG antibodies (Blot Ab: Human IgG [bottom panel]). Anti-human IgG was used to detect the x-EphB1-Fc protein. The level of Tyrosine-phosphorylated ephrin increased sixfold over the indicated time course, as determined by scanning densitometry of the autoradiograph.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Transmembrane (B-type) ephrins have been implicated in bidirectional signaling upon contact with cognate receptor-expressing cells or reverse signaling through phosphorylation of its cytoplasmic domain. Our results identify the FGF receptor as a molecule that can associate with and induce tyrosine phosphorylation of the ephrin B1 ligand, thus modulating its effect on cell adhesion when ectopically expressed in pluripotent ectodermal cells. Additional support for the importance of the association between FGFR and ephrin B1 comes from experiments with the Y305FY310F double mutant. When expressed in embryos, this mutant ephrin B1 causes cell dissociation but demonstrates a marked reduction in the ability to associate with the FGFR and fails to be effectively rescued by FGFR activity (Fig. 3c and d). These data also indicate that the rescue of cell adhesion is not the result of a general cellular process or a downstream signaling event that results from activation of the FGF receptor pathway. It is also worth noting that while the cytoplasmic domain residues 305 and 310 are important for the ephrin-FGF receptor association, they are not the only required residues. An x-ephrin B1 mutant lacking the entire extracellular domain (38) is also unable to interact with the FGF receptor (data not shown). Therefore, it is likely that regions in both the intracellular and extracellular domains mediate ephrin B1-FGF receptor interaction.

We also show that in neural tissue (chicken retina), FGF treatment results in the phosphorylation of the resident ephrin B protein (Fig. 4), providing supportive evidence for an interaction between the endogenous proteins. In Xenopus, the association between ectopically expressed FGF receptor and ephrin B1 is stable even under stringent conditions (Fig. 2B), but it is not clear whether the association between these two molecules is direct or requires another protein. A complex between the active FGF receptor and ephrin B1 is observed in extracts from embryos coexpressing the two proteins but is not detected in a mixture of two extracts, each expressing only one of the molecules (Fig. 2B), suggesting that another protein may be required for complex formation. It is also possible that while the interaction between the activated FGF receptor correlates with x-ephrin B phosphorylation, tyrosine phosphorylation occurs through an indirect mechanism requiring another kinase.

Both Eph and FGF RTK families and their activating ligands coordinate cell adhesion and migration during development (14, 16, 54, 78). While FGF signaling has been implicated in regulating the cell fate and differentiation of neuronal precursors, a role in controlling the migratory pathway of mesodermal and neuronal precursors has also emerged (16, 54, 76). FGF has also been implicated in modulating cell-adhesive interactions between neuroepithelial cells and the extracellular matrix during neural-tube development (40). The role of FGF in migration events is not limited to vertebrates. In Drosophila, FGF signaling controls morphogenetic movements occurring during gastrulation, neurogenesis, and trachea formation (26, 41). Similarly, ephrins and their receptors also play a role in morphogenetic events during development. Ephrins control the migration of neural crest cells in the branchial arches and somites (29, 43, 62, 71). EphA4 and ephrin B2 functions are necessary for the proper formation of somites (19) and the restriction of cell intermingling (51). Mutations in the Caenorhabditis elegans Vab-1 Eph receptor disrupt epidermal morphogenesis (25). Losses of both cell adhesion and blastocoel formation result from the expression of the activated Eph receptor tyrosine kinase, Pag (x-EphA4) in Xenopus embryos (74). Our findings raise the possibility that cross talk between the FGF receptor and ephrins allows further regulation of cell adhesion events.

Recent work has demonstrated an additional role for ephrins in vasculogenesis (55), where clustered ephrin B1 promoted cell adhesion and capillary-like assembly in cognate receptor-bearing endothelial cells in vitro (65). In another study, ephrin B2-deficient mice had defects in angiogenesis (72). Ephrin B1, like angiopoietins, can also induce sprouting in an in vitro capillary assay, while a GST-Tie2 receptor fusion can phosphorylate the cytoplasmic domain of ephrin B1 in vitro. Thus, it has been suggested that cross talk may exist between signaling pathways triggered by angiogenic factors and ephrin B proteins (1). In this regard, it is interesting that FGFs exhibit strong angiogenic activities in vivo and that FGF signaling is thought to play an important role in the normal development of the vascular system (35, 73). While we observed little detectable phosphorylation of ephrin B1 in response to PDGF in our system (Fig. 1C), the strong phosphorylation of ephrin B1 (Lerk2) in NIH 3T3 cells treated with PDGF (6) is also suggestive of a cross talk scheme. Future analysis of a potential ephrin B1-PDGF receptor interaction will provide further insight into the complex nature of ephrin signaling within different cellular contexts.

It is intriguing that ephrin Bs may play a role in cell adhesion and movement in several cell contexts (endothelial, neural crest, axon guidance, etc.), and that these proteins may be regulated by a number of different receptors in the developing embryo. Although the mechanism of how x-ephrin B1 influences cell adhesion is unclear, its association with and phosphorylation induced by the FGF receptor family affect this event. Phosphorylation may facilitate the interaction with SH2 domain-containing proteins and/or affect ephrin binding to cognate Eph receptors. The carboxyl-terminal PDZ domain of ephrin B1 has recently been shown to bind syntenin, a protein proposed to function as an adapter that couples syndecans to cytoskeletal proteins or cytosolic downstream signal effectors (28, 48). However, this binding may be reduced when the PDZ domain is tyrosine phosphorylated (48). Elucidation of other associated proteins and how they may be organized and interact within a signaling complex is under way. Interestingly, both FGF and Eph receptors functionally interact with L1, another protein involved in cell adhesion (15, 59, 79). Extensive studies have explicitly demonstrated that the FGF receptor is necessary for neurite outgrowth stimulated by the cell adhesion molecules L1, N-cadherin, and NCAM (15, 59). Our data suggest the possibility of an even broader scheme for FGF receptor-related regulation of cell adhesion.