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.

Human embryonal kinase 2 (HEK2) is a protein-tyrosine kinase that is a member of the EPH family of receptors. Transcripts for HEK2 have a wide tissue distribution. Recently, a still growing family of ligands, which we have named LERKs, for igands of the ph-elated inases, has been isolated. In order to analyze functional effects between the LERKs and the HEK2 receptor, we expressed HEK2 cDNA in an interleukin-3-dependent progenitor cell line 32D that grows as single cells in culture. Within the group of LERKs, LERK-2 and −5 were shown to bind to HEK2. Membrane-bound and soluble forms of LERK-2 were demonstrated to signal through HEK2 as judged by receptor phosphorylation. Coincubation of HEK2 and LERK-2 expressing cells induced cell-cell adhesion and formation of cell aggregates. This interaction could be inhibited by preincubation of HEK2 expressing cells with soluble LERK-2. Coexpression of HEK2 and LERK-2 in 32D cells showed reduced kinase activity and autophosphorylation of HEK2 compared with the juxtacrine stimulation, which seems to be due to a reduced sensitivity of the receptor.

The first ligand to be identified was B61 (12,13). B61, a tumor necrosis factor ␣-induced gene product from cultured human umbilical vein endothelial cells was found to bind the EPH receptor family member ECK. We have also found that B61 is a ligand for HEK and elk and have named it LERK-1 (14). Subsequently, six additional ligands for the EPH family of RTKs were identified which we have named LERK-2 through LERK-7 (14 -18). These ligands have an amino acid sequence identity ranging between 27 and 59% and are membranebound. The LERKs can be subdivided into two groups based on their mechanism of membrane attachment. LERK-1, LERK-3, LERK-4, LERK-6, and LERK-7 are anchored by glycosylphosphatidylinositol linkage while LERK-2 and LERK-5 are type 1 transmembrane proteins.
Here we describe that within the group of LERK proteins, LERK-2 and LERK-5 bind to HEK2 expressed in 32D-myeloid progenitor cell line, whereas LERK-1, LERK-3, LERK-4, and LERK-7 do not. In addition, soluble forms of LERK-2 stimulate autophosphorylation of HEK2 so that membrane attachment does not seem to be required for activation of the HEK2 receptor kinase. Furthermore, we provide evidence for stable cell-cell contacts mediated by the interaction of LERK-2 and HEK2 expressing cells and subsequent juxtacrine stimulation of receptor autophosphorylation.

MATERIALS AND METHODS
Reagents-Rabbit polyclonal antisera to HEK2 were generated by immunizing animals with a peptide corresponding to amino acid residues 897-998 (7). Monoclonal anti-phosphotyrosine antibodies PY20 were purchased from Transduction Laboratories. Sepharose-protein A beads, horseradish peroxidase-conjugated goat anti-mouse, and goat anti-rabbit antibodies were from Sigma. The mammalian expression vectors pRc/CMV and pCEP4 were supplied by Invitrogen. Gentamicin sulfate (G418) was obtained from Life Technologies, Inc., hygromycin B was from Boehringer Mannheim.
Generation of a Kinase-negative Mutant of HEK2 and Construction of Expression Vectors with HEK2 Receptors or LERK-2-The HEK2 cDNA, containing the complete open reading frame (nucleotide position 4 -3086) (7), was inserted into the HindIII/XbaI site of the mammalian expression vector pRc/CMV. A K665R kinase-deficient HEK2 receptor was generated by polymerase chain reaction-mediated site-directed mutagenesis using the following mutagenic oligonucleotide 5Ј-GTGCC-GTGGTCGACTGAAACAGCCTGGCCGCCGAGAGGTGTTTGTGGCC-ATCCGGACG-3Ј, corresponding to nucleotides 1968 -2025 of the published cDNA sequence (7). The mutated full-length receptor cDNA was cloned into the pCEP4 vector using HindIII and BamHI restriction * The Georg-Speyer-Haus is supported by the Bundesgesundheitsministerium and the Hessisches Ministerium fü r Wissenschaft und Kunst. This work was further supported by grants from the Georg und Franziska Speyer'sche-Hochschulstiftung, the Hessischer Verein zur Förderung der Jugendgesundheitspflege e.V., and the Deutsche Forschungsgemeinschaft Grants RU 242/11 u. 12-1, STR 336/6 -1. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
sites. The LERK-2 cDNA, nucleotide position 308-1348 (14), was cloned into the HindIII and BamHI cloning sites of the mammalian expression vector pCEP4. All constructs were verified by DNA sequencing.
Cell Line, Culture Conditions, and Transfection-32D cells are an immature murine myeloid cell line that is absolutely dependent on exogenously supplied interleukin-3 for maintenance of its growth and survival in culture. These cells were maintained in RPMI 1640 supplemented with 10% fetal calf serum, 2 mM L-glutamine, antibiotics (streptomycin/penicillin), and 5% conditioned medium derived from the interleukin-3 producing cell line WEHI-3B. Cells were subjected to electroporation (250 V, 960 microfarad) in RPMI 1640 containing 25 g of linearized plasmid DNA. Following 2 weeks of selection in 600 g/ml G418 (pRc/CMV) or 2 mg/ml hygromycin B (pCEP4) stable transfectants were tested by immunoblotting or Northern blotting.
LERK-2 expressing cells were kept in culture for 2 days and subsequently spun down, and the supernatant was used as conditioned medium for further experiments.
Immunoblotting-Cells were washed twice with ice-cold phosphatebuffered saline (PBS). The cell pellet was resuspended in 500 l of lysis buffer (50 mM HEPES (pH 7.5), 150 mM NaCl, 10% glycerol, 1% Triton X-100, 1.5 mM MgCl 2 , 1 mM EGTA) freshly supplemented with 10 g of leupeptin per ml, 10 g of aprotinin per ml, 1 mM phenylmethylsulfonyl fluoride, and 1 mM Na 3 VO 4 . Cell lysates were incubated on ice for 15 min with occasional vortexing and then clarified by centrifugation for 10 min at 12,000 ϫ g. The protein concentration was measured using the Pierce BCA protein assay reagents. 25 g of protein were fractionated per lane on a SDS-7.5% polyacrylamide gel by electrophoresis. After transfer of the proteins onto Immobilon P (Millipore), the filters were preincubated 2 h with 1% bovine serum albumin, 1% gelatin in PBS, washed with PBST (0.05% Tween 20 in PBS), incubated 1 h at 22°C with primary antibody in reaction buffer (1% bovine serum albumin, 10% fetal calf serum, 0.1% Triton X-100 in PBS), washed three times, and incubated for 1 h with horseradish peroxidase-conjugated goat anti-mouse or goat anti-rabbit antibody. Immunoblots were developed using the ECL system (DuPont NEN) or the diaminobenzidine staining method. In order to reprobe proteins with a second antibody, filters were incubated for 30 min in strip buffer (62.5 mM Tris (pH 6.8), 100 mM ␤-mercaptoethanol, 2% SDS) at 50°C.
Immunoprecipitation-Prior to lysis, cells were stimulated with LERK-2 for 30 min at 37°C. 200 g of proteins were incubated for 1-2 h at 4°C with HEK2-specific antibodies coupled to Sepharose-protein A beads. Immunoprecipitates were washed three times with lysis buffer, mixed with SDS-gel sample buffer, heated at 95°C for 5 min, and subjected to gel electrophoresis.
In Vitro Kinase Assay-Washed immunoprecipitates were suspended in 20 l of 50 mM Pipes (pH 7.0), 10 mM MnCl 2 , 5 Ci of [-32 P]ATP (3000 Ci/mmol, DuPont NEN) and incubated at 30°C for 10 min. Reaction mixtures were mixed with an equal volume of 2 ϫ sample buffer, boiled for 2 min, and fractionated on a 7.5% SDS-polyacrylamide gel. The dried gel was exposed to an x-ray film for 20 min.
Northern Blot Analysis-Total RNA was isolated from 5 ϫ 10 7 cells by using the guanidinium isothiocyanate/CsCl cushion technique. The RNA was fractionated on a 1% formaldehyde agarose gel before being transferred onto a Hybond N ϩ membrane (Amersham Corp.). Hybridization with polymerase chain reaction-generated radiolabeled probes was performed as described before (7). pBluescript (KS ϩ ) HEK2 cDNA and pCEP4 LERK-2 cDNA were used as templates in the synthesis of probes. Blots were exposed to an x-ray film for 2 h.

RESULTS
Expression of HEK2 in the 32D Myeloid Cell Line-To study the effects of ligand binding on HEK2 activation, we inserted the complete HEK2 open reading frame into the pRc/CMV vector, placing the cDNA under the transcriptional control of the cytomegalovirus enhancer-promoter. As a control, we generated a receptor lacking intrinsic protein-tyrosine kinase activity (HEK2/KIN Ϫ ). A single amino acid mutation replaces the conserved lysine, which is proposed to provide an essential role for binding of ATP, by arginine (21). This mutated cDNA sequence was cloned into the cytomegalovirus early promotorbased expression vector pCEP4. Subsequently, 32D cells transfected with these HEK2 expression vectors were cultured for 2 weeks in the presence of a selective drug (G418 for pRc/CMV and hygromycin B for CEP4). Stable transfectants were analyzed by Western blotting of whole cell lysates. Rabbit polyclonal antibodies directed against the C terminus of HEK2 (amino acids 897-998) specifically recognized polypeptides of approximately 110 kDa in cells transfected with either HEK2 or mutagenized HEK2/KIN Ϫ cDNA (Fig. 1, lanes 1 and 2). No HEK2-related polypeptides were detected in the parental clone of 32D cells (Fig. 1, lane 3). In all subsequent experiments, cell lines expressing the highest number of receptors (clone 32DH20 for HEK2 and clone HEK2/KIN Ϫ for the kinase-negative mutant) were used.
Binding of LERK Proteins to HEK2-The binding characteristics of HEK2 with different LERKs were analyzed utilizing fusion proteins consisting of the extracellular domain of the LERKs linked to the Fc domain of human IgG1, resulting in soluble forms of LERKs, LERK/Fc. Binding of these molecules was measured using an indirect method in which 32D cells expressing HEK2 were incubated with varying concentrations of LERK/Fc followed by saturating concentrations of 125 I-labeled mouse anti-human IgG antibodies directed against the Fc portion of the molecule. Only LERK-2 and LERK-5 were found to bind to HEK2 in this assay (Table I) Immunoprecipitates of HEK2 from lysates of ligand-stimulated cells were fractionated by SDS-polyacrylamide gel electrophoresis and probed with anti-phosphotyrosine antibodies. As shown in Fig. 3 soluble LERK-2/Fc caused dose-dependent phosphorylation of the wild-type HEK2 receptor and reached a maximum at a concentration of 250 ng of LERK-2/Fc per ml in the cell culture medium. Incubation of 32D-HEK2 cells without addition of LERK-2/Fc resulted in a low but detectable level of tyrosine phosphorylation (Fig. 3, lane 1) that might be due to an intrinsic basal level of HEK2 tyrosine phosphorylation, to a stimulatory effect exerted by factors present in the cell culture medium, or to proteins other than LERK-2 expressed on the cell surface of 32D cells. In summary, our results indicate that soluble LERK-2/Fc-protein can functionally interact with the HEK2 receptor.
Membrane-bound LERK-2 Induces Phosphorylation of Wildtype HEK2-The ability of membrane-bound LERK-2 to stimulate the HEK2 autophosphorylation was examined as well. For this purpose, we produced 32D-derived cell lines expressing transmembrane forms of LERK-2 or HEK2. Furthermore, LERK-2 and HEK2 coexpressing 32D-derived cells were generated. A Northern blot analysis using specific probes demonstrated equal levels of ligand and receptor transcripts in corresponding cell lines (Fig. 4).
Prior to stimulation LERK-2 expressing cells were washed thoroughly with fresh medium, subsequently mixed at a ratio of 1:1 (6 ϫ 10 6 cells of each cell type per ml) with HEK2 expressing cells, and incubated for 30 min. Immunoprecipitation of HEK2 from those cells was followed by immunoblotting with anti-phosphotyrosine antibodies. As shown in Fig. 5A increased tyrosine phosphorylation is induced by coincubation of HEK2 and LERK-2 expressing cells (Fig. 5A, lane 7). We found that during the period of stimulation the culture medium of cocultivated cells was devoid of factors such as soluble ligands that could activate receptor phosphorylation.
Membrane-anchored LERK-2 Released to the Cell Culture Medium Activates Kinase Activity of HEK2-Conversion of membrane-anchored growth factors into soluble forms involves a proteolytic system, which acts at or near the cell surface and shows broad proteolytic activity (22). We therefore determined whether LERK-2 is released from 32D cells into the medium and whether it could activate HEK2. As shown in Fig. 5A (lanes  5 and 6), conditioned medium from LERK-2 expressing cells but not from nontransfected cells induces elevated phosphorylation of HEK2 compared with the previously described coincubation experiments, in which the membrane-bound ligand induced receptor activity (Fig. 5A, lane 7).
Activation of the HEK2 Receptor Is Reduced in HEK2/ LERK-2 Coexpressing Cells-We have further analyzed the influence of HEK2/LERK-2 coexpression in 32D cells on the kinase activity of the receptor. This coexpression caused a very low level of autophosphorylation that was at the limit of detection in this analysis (Fig. 5A, lane 8). Prolonged exposure  showed that the signal in HEK2/LERK-2 coexpressing cells was comparable with the one seen in HEK2 expressing cells. Thus, the degree of HEK2 autophosphorylation seems to depend on the mode of activation, autocrine or juxtacrine stimulation by LERK-2 (Fig. 5A, lanes 7 and 8). We found that the reduced phosphorylation of HEK2 is not due to different concentrations of ligand and receptor in both experiments. Still, steric hindrance of receptor-ligand interaction in coexpressing cells could be the reason for the diminished stimulatory efficiency of LERK-2. To study this effect in more detail, we have assayed the interaction of LERK-2 and HEK2 under conditions that exclude steric hindrance exerted by coexpression of both molecules on the same cell surface and measured the tyrosine kinase activity in vitro. Immunoprecipitates of HEK2 with polyclonal antiserum were subjected to an in vitro kinase reaction in the presence of [-32 P]ATP. Our results suggest that LERK-2 from coexpressing cells is less efficient in stimulating the kinase activity of the HEK2 receptor compared with LERK-2 expressed in the absence of HEK2 (Fig. 5B, lanes 6 -8). This observation could be due to a different nature of the LERK-2 protein in single and coexpressing 32D cells. Still, comparing both cell types we could demonstrate that LERK-2 is expressed at the same level on each surface and displays the same mobility on denaturing gels. Furthermore, the affinity constants (K a ) for LERK-2/Fc binding to HEK2/32D cells and to HEK2/LERK-2 coexpressing cells was not significantly different (data not shown).
We further analyzed whether the phosphorylation of HEK2 could be up-regulated in HEK2/LERK-2 coexpressing cells, but soluble LERK-2/Fc did not exhibit the ability to induce phosphorylation of the receptor (Fig. 5C, lane 5). Moreover, conditioned medium from HEK2/LERK-2 coexpressing cells did not stimulate HEK2 activity in 32D/HEK2 cells suggesting that no ligand or no functional ligand is released into the medium (Fig. 5C, lane 4).
Adhesion of 32D Cells Expressing HEK2 to LERK-2 Expressing Cells-HEK2 expressing 32D cells were coincubated with LERK-2 expressing cells. In contrast to the parental 32D line and to cells transfected with HEK2 (Fig. 6A) or LERK-2 alone, which grow as single cells, coincubation of LERK-2 and HEK2 expressing cells resulted in the formation of cell aggregates within 10 min after mixing of the cells (Fig. 6B). This cell aggregation was also seen in cells coexpressing both, the receptor and the ligand (Fig. 6C). To verify the specificity of this receptor-ligand interaction, we tested the ability of LERK-2/Fc to inhibit the binding of 32D-HEK2 cells to 32D-LERK-2 cells. Addition of the soluble ligand prevented the formation of cell aggregates (Fig. 6D). In summary, cell-cell adhesion is mediated by the membrane-anchored ligand LERK-2 to 32D-HEK2 cells. This interaction is inhibited by soluble LERK-2.

DISCUSSION
The EPH-related kinases constitute the largest known family of orphan receptor tyrosine kinases with several members being expressed in the developing and adult nervous system. The recent characterization of a family of ligands that we call LERKs is a first step in trying to understand what cellular processes might be regulated by the EPH-related receptor proteins. In this article we have demonstrated that within this family of ligands only the type 1 transmembrane members, LERK-2 and LERK-5, bind to HEK2. Moreover, LERK-2 was shown to stimulate the kinase activity of the HEK2 receptor. In previous indirect binding assays LERK-2 protein expressed on CV-1 transfectants was shown to display also high affinity binding with a K a of 1.08 ϫ 10 9 M Ϫ1 to elk/Fc (14). During the preparation of this manuscript, a study was published demonstrating that LERK-2 binds to CEK5 and CEK10, two EPHrelated receptors from chicken tissues (23). The affinity constants for LERK-2 binding to these receptors are K a ϭ 3.7 ϫ 10 9 M Ϫ1 for CEK10 and K a ϭ 2.3 ϫ 10 9 M Ϫ1 for CEK5. Taken together, LERK-2 shows extensive cross-binding to EPH-related RTKs, as demonstrated for other members of the LERKfamily (15), indicating functional redundancy.
HEK2/LERK-2 coexpressing cells and coincubated HEK2and LERK-2 expressing cells were compared with respect to receptor phosphorylation. Interestingly, in the case of coexpressing cells, we observed reduced autophosphorylation of HEK2. This effect might be due to steric hindrance that results in an inappropriate interaction of the ligand to its receptor. Alternatively, the ligand could be modified in a way that reduces its ability to induce receptor phosphorylation. To elucidate this phenomenon, we performed in vitro kinase assays and found that kinase activity with immunoprecipitated HEK2 receptors from coexpressing cells is reduced compared with HEK2-and LERK-2 coincubated cells. Since conformational aspects are essential for receptor activation in vivo and have been shown to be of less importance in kinase assays, we speculate that steric hindrance is not the cause for reduced receptor activation. This was further supported by identical binding affinities in single and coexpressing cells. In addition, we demonstrated that the nature of the ligand from both cells types is very similar as judged by their migration in denaturing gels. Thus, we suggest that the intracellular portion of the LERK-2 inhibits receptor activation by reducing its kinase activity. A negative regulatory role of the intracellular portion of LERK-2 has been postulated in the case of coexpression with chimeras consisting of chicken EPH-related receptors and TrkB as well (23). A functional role for the cytoplasmic domain of LERK-2 is also suggested because of the high amino acid sequence identity between the cytoplasmic domains of LERK-2 and LERK-5 (19).
Cell-cell adhesion and juxtacrine activation that is a cell-cell stimulation that is mediated by interaction of a membraneattached ligand with its receptor has been demonstrated for different cell-surface receptors (27)(28)(29)(30). We asked the question whether the expression of the HEK2 receptor and LERK-2 in 32D cells induces the ability of 32D cells to adhere. Several lines of evidence support the conclusion that adhesion of these two transfectants is mediated by binding to membrane LERK-2 via HEK2 receptors. Cell adhesion mediated by membraneattached ligands might support cell home to tissue locations that express a certain factor. Especially members of the EPH family containing cell adhesion domains in their extracellular portion are expressed in embryonic development of multicellular organisms at high levels. For example, the subcellular localization of Nuk receptor is concentrated at sites of cell-cell contact (31). High levels of the Nuk protein are found within initial axon outgrowth and associated nerve fibers. Our observations on cell adhesion using the 32D test system and on signaling through an EPH-related receptor could help to study those developmental processes in more detail.
Acknowledgment-We thank G. Raab for helpful support and for providing 32D cells.