The Soluble Sema Domain of the RON Receptor Inhibits Macrophage ...

Download PDF
46 downloads 72991 Views 279KB Size Report
Comment
Aug 22, 2003 - sema domain of RON might contain the ligand-binding region. .... 2 nM), or free α-chain (highest concentration, 100 nM) in culture medium.
THE JOURNAL

OF

BIOLOGICAL CHEMISTRY

Vol. 279, No. 5, Issue of January 30, pp. 3726 –3732, 2004 Printed in U.S.A.

The Soluble Sema Domain of the RON Receptor Inhibits Macrophage-stimulating Protein-induced Receptor Activation* Received for publication, August 22, 2003, and in revised form, October 29, 2003 Published, JBC Papers in Press, November 3, 2003, DOI 10.1074/jbc.M309342200

Debora Angeloni‡§¶, Alla Danilkovitch-Miagkova‡¶储, Alexei Miagkov**, Edward J. Leonard‡, and Michael I. Lerman‡ ‡‡ From the ‡Laboratory of Immunobiology, National Cancer Institute, National Institutes of Health, Frederick, Maryland 21702-1201

RON is a receptor tyrosine kinase of the MET family that is involved in cell proliferation, cell survival, and cell motility in both normal and disease states. Macrophage-stimulating protein (MSP) is the RON ligand whose binding to RON causes receptor activation. RON is a trans-membrane heterodimer comprised of one ␣and one ␤-chain originating from a single-chain precursor and held together by several disulfide bonds. The intracellular part of RON contains the kinase domain and regulatory elements. The extracellular region is characterized by the presence of a sema domain (a stretch of ⬃500 amino acids with several highly conserved cysteine residues), a PSI (plexin, semaphorins, integrins) domain, and four immunoglobulin-like folds. Here we show that a soluble, secreted molecule representing the sema domain of RON (referred to as ronsema) has a dominant negative effect on the ligandinduced receptor activation and is capable of inhibiting RON-dependent signaling pathways and cellular responses. Results suggest that the sema domain of RON participates in ligand binding by the full-length receptor. The ability of ron-sema to suppress growth of MSPresponsive cells in culture, including cancer cells, points to a potential therapeutic use of this molecule, and forced expression of it could potentially be used as a gene therapy tool for treating MSP-dependent types of cancer.

The human receptor tyrosine kinases MET and RON (MST1R) (and their respective orthologs in other species) form a unique, two-member gene family that encodes proteins with identical modular structure that may perform similar functions. Oncogenic activating mutations of MET were discovered in hereditary renal carcinoma of the papillary type (1, 2) and a variety of common cancers (3). Oncogenic amplification and overexpression of MET were also reported previously (3). The involvement of RON in carcinogenesis is much less understood.

* This work was supported by NCI, National Institutes of Health Contract NO1-CO-56000. 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. § Present address: Scuola Superiore Sant’Anna and Istituto di Fisiologia Clinica-Consiglio Nazionale delle Richerche, 56100 Pisa, Italy. ¶ Both authors contributed equally to the work. 储 Present address: Osiris Therapeutics, Inc., 2001 Aliceanna St., Baltimore, MD 21231. ** Neuromuscular Research Laboratory, Johns Hopkins School of Medicine, Baltimore, MD 21287-7519. ‡‡ To whom correspondence should be addressed: Laboratory of Immunobiology, NCI, National Institutes of Health, Bldg. 560, Rm. 12-68, Frederick, MD 21702-1201. Tel.: 301-846-7323; Fax: 301-846-6145; E-mail: [email protected].

We have recently discovered that RON is negatively controlled by the tumor suppressor protein HYAL2 and becomes activated in a ligand-independent mechanism upon HYAL2 removal, leading to cell transformation (4). An oncogenic function of RON was also demonstrated in a transgenic model (5). RON is almost ubiquitously expressed (6 – 8) in a variety of normal cell types (6, 9 –14) and also in a number of tumor cell lines (15) in which RON kinase activity is deregulated. The extracellular part of RON is comprised of a sema domain, a PSI1 domain (plexin semaphorins integrins (16), also known as MRS domain (MET-related sequence) (17)), and four IPT domains (immunoglobulin-like fold shared by plexins and transcription factors (16)). The functional role of these domains remains unknown. The sema domain (whose presence in RON and MET makes them a unique family among the approximate 20 receptor tyrosine kinase families identified by the Human Genome Project) is represented by a block of ⬃500 amino acids with 15 conserved cysteine residues intermingled with stretches of conserved amino acids and a conserved potential glycosylation site (18). This domain is also present in semaphorins and plexins (a class of semaphorin receptors). Secreted semaphorins and their receptors were initially discovered by their ability to induce axon steering and collapse of the growth cone in vitro, but it is now evident their biological function is not confined to the neural system and involves at least the immune (19, 20), cardiovascular, and skeletal systems (21) in vertebrates. Semaphorins are also overexpressed in metastasizing cancer cells (22, 23), possibly facilitating cell dissociation and survival. Because the sema domain of semaphorins mediates receptor specificity (24), we reasoned that the sema domain of RON might contain the ligand-binding region. In this study, we show that two soluble molecules designed over the sema and sema ⫹ PSI domains of RON (referred to as “ron-sema” and “ron-PSI,” respectively) undergo correct posttranslational processing and are secreted when expressed in mammalian cells. Both molecules inhibit ligand binding to the RON receptor (both ␣- and ␤-MSP-binding sites), indicating that the RON sema domain may contain the ligand-binding sites and mediate ligand-binding specificity. In addition, we found that these secreted soluble molecules have a highly specific dominant-negative effect on RON kinase activation. Such activity extends to inhibiting the proliferation of MSP-sensitive cells in culture, including cancer cell lines. This highlights the potential therapeutic use of the secreted soluble molecules ron-sema and ron-PSI to inhibit MSP-dependent tumor growth. 1 The abbreviations used are: PSI, plexin semaphorins integrins; IPT, immunoglobulin-like fold shared by plexins and transcription factors; MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide; ELISA, phenylmethylsulfonyl fluoride; MSP, macrophage-stimulating protein.

3726

This paper is available on line at http://www.jbc.org

Dominant-Negative Activity of the RON Sema Domain The forced expression of these secreted soluble molecules could represent a gene therapy tool to treat MSP-dependent types of cancer. EXPERIMENTAL PROCEDURES

Materials—Human recombinant ␣-␤-heterodimeric MSP, free MSP ␣- and ␤-chains, and recombinant mouse MSP receptor RON/Fc chimera were from R&D Systems (Minneapolis, MN). Human EGF was from Invitrogen. Human recombinant HGF was a gift from Dr. P. Godowski (Genentech, San Francisco, CA). Rabbit and mouse anti-MSP antibodies were described previously (25). Rabbit anti-human RON (C-20), goat anti-human EGFR (1005-G), rabbit anti-human MET (C28), and goat anti-mouse MET (SP260) were from Santa Cruz Biotechnology (Santa Cruz, CA). Monoclonal anti-phosphotyrosine (Clone 4G10) and anti-FLAGM2 antibodies were from Upstate Biotechnology (Lake Placid, NY). Monoclonal anti-phospho-MAPK antibody (E10) was from Cell Signaling Technology (Beverly, MA). Monoclonal anti-MAPK (ERK1/2) antibody was from Transduction Laboratories (Lexington, KY). Other chemicals were from Sigma or Invitrogen. Constructs and Cloning Procedures—ron-sema was amplified by PCR from RON cDNA with the primers 17 forward (5⬘-CAG CTC GCC TCG ATG GAG-3⬘) and 1583 reverse (5⬘-GGA AAA CCT GGT CCC CAG AGG-3⬘) in 1.5 mM MgCl2 under the following PCR conditions: 5 min at 95 °C; 35 times (1 min at 95 °C, 30 s at 64 °C, and 2 min at 72 °C); and 7 min at 72 °C. The PCR product (1549 bp, 519 amino acids) was cloned into pCR2.1 (Invitrogen), sequenced, and cloned into the HindIII/EcoRV restriction sites of the pFLAG-CMV-2 expression vector (Sigma). ron-PSI was PCR amplified with the primers (ABI Applied Biosystems, Frederick, MD) 17 forward (5⬘-CAG CTC GCC TCG ATG GAG-3⬘) and 1762 reverse (5⬘-ACT GTG GGG GTG GAA CTC AGT AAG C-3⬘) in 1.5 mM MgCl2 under the same cycling conditions. The 1747-bp PCR product (578 amino acid) was first cloned into pCR2.1 (Invitrogen), sequenced, and cloned into the HindIII sites of pFLAG-CMV-2 (Sigma). The primer location is given according to GenBankTM accession number X70040. Sequencing was performed with ABI 373 stretch automated DNA sequencer (Applied Biosystems). Human MET cDNA was kindly provided by Dr. L. Schmidt (Laboratory of Immunobiology, SAIC-NCI, Frederick, MD). Cell Lines—HEK 293, HEK293T, MDCK, and colon cancer HCT116 cell lines were from ATCC (Manassas, VA). RE7 cell line (MDCK cells stably expressing human RON) was described previously (26). Cells were cultured in media containing 10% fetal calf serum according to ATCC recommendation. Transfections—HEK293 and HEK293T cells were grown in Dulbecco’s modified Eagle’s medium with 10% fetal calf serum. For transient transfection, cells were grown to 80 –90% confluence on 10-cm dishes and transfected with 30 ␮g of cDNAs of the empty vector or the ronsema or ron-PSI construct using LipofectAMINE 2000 (Invitrogen) according to the manufacturer’s protocol. 6 h after transfection, the medium was replaced by 6 ml/dish serum-free Dulbecco’s modified Eagle’s medium and cells were incubated for 3 days. The 3-day-old conditioned media were collected and analyzed for the presence of ron-sema or ron-PSI proteins by immunoprecipitation and Western blotting. HEK 293 cells were similarly transiently transfected with human MET cDNA. Cell Stimulation, Immunoprecipitation, and Western blotting—Cells were starved overnight in serum-free medium and then stimulated for 15 min with 5 nM MSP or 50 ng/ml EGF or 100 ng/ml HGF diluted in preconditioned culture medium containing secreted ron-sema or ronPSI. Preconditioned medium of non-transfected cells was used as negative control. After stimulation, cells were lysed in lysis buffer (50 mM HEPES, pH 7.4, 150 mM NaCl, 10% glycerol, 1 mM EDTA, 1 mM sodium orthovanadate, 10 mM sodium pyrophosphate, 100 mM NaF, 1% Triton X-100, 10 ␮g/ml leupeptin, 10 units/ml aprotinin, 1 mM phenylmethylsulfonyl fluoride). Insoluble material was removed by low-speed centrifugation. Supernatants were further analyzed. Molecules of interest were immunoprecipitated from cleared lysates by appropriate antibodies added with protein G-agarose to the lysates. After overnight incubation at 4 °C, immunoprecipitates were washed three times with HNTG buffer (50 mM HEPES, 150 mM NaCl, 0.1% Triton X-100, 10% glycerol). 15 ␮l of 2⫻ SDS-PAGE sample buffer were added to each tube. After 5 min of boiling, precipitated proteins were separated by SDS-PAGE and transferred onto a nitrocellulose membrane. Membrane-bound protein bands were visualized with appropriate antibodies according to standard Western blotting procedures. Cleared total cell lysates were used in SDS-PAGE and Western blotting for detection of MAPK activation.

3727

In Vitro Cell Proliferation Assay—RE7 cells (MDCK cells stably expressing the human RON receptor) (26) were plated (2 ⫻ 103 cells/ well) in triplicate in 96-well plates and incubated with 5 nM MSP diluted in serum-free Dulbecco’s modified Eagle’s medium plus preconditioned culture medium containing secreted ron-sema or ron-PSI (final ratio 1:1). Conditioned medium of non-transfected cells was used as negative control. HCT116 colon cancer cells (2 ⫻ 103 cells/well) were distributed in duplicates in 24-well plates and incubated with or without 5 nM MSP in McCoy serum-free medium plus preconditioned culture medium containing ron-sema or ron-PSI (final ratio 1:1). Cell number was determined after 72 h by adding 3-(4,5-dimethylthiazol-2-yl)-2,5diphenyltetrazolium bromide (MTT) into culture wells and measuring A570 2 h later. Cell number was determined from an MTT calibration curve specific for each cell line. ELISA—All of the experiments were performed as described previously (27). ELISA microtiter plates were coated overnight at 4 °C with recombinant mouse MSP-receptor RON/Fc (mrRON/Fc) chimera (R&D Systems, Inc.) (100 ␮l/well at 0.5 ␮g/ml in carbonate-bicarbonate buffer, pH 9.6). Unspecific binding sites were blocked with 1% bovine serum albumin (1 h at 37 °C). Subsequently, RON-coated ELISA plates were incubated for 2 h with a series of 2-fold dilutions of human recombinant MSP (highest concentration, 2 nM), free ␤-chain (highest concentration, 2 nM), or free ␣-chain (highest concentration, 100 nM) in culture medium from HEK293T cells transiently transfected with empty vector (negative control) or ron-sema cDNA construct. ELISA plates were then washed three times with Tris-buffered saline-Tween 20. Plates were then incubated for 2 h with rabbit polyclonal anti-MSP antibodies. Bound rabbit anti-MSP was detected by equilibration with alkaline phosphatase-conjugated anti-rabbit IgG. Substrate 104 (Sigma) was used for quantifying alkaline phosphatase activity. A405 was measured with an ELISA reader. At least three independent experiments were performed. The effect caused by ron-sema on MSP binding, ␤-chain binding, and ␣-chain binding to mrRON/Fc was calculated in percent inhibition versus the effects of the control culture medium at each experimental point. The binding of MSP ␣- or ␤-chains in the presence of the blank medium (preconditioned medium of non-transfected cells) was counted as 100%. RESULTS

The Expression Constructs Designed over the sema and sema ⫹ PSI Domains of RON Are Correctly Processed and Secreted after Translation—The sema domain is characterized by the presence of ⬃15 highly conserved cysteine residues interspersed in ⬃500 amino acids (18). The PSI domain contains eight conserved cysteines distributed in ⬃50 residues. Initially, it was part of the definition of sema domain but because not all of the semaphorines contain a PSI domain, this modular structure was redefined (16). The minimal consensus sequence for the PSI domain was described previously (16, 17) as C-X-(5– 6)-C-X-(2)-C-X-(6 – 8)-C-X-(2)-C-X-(3–5)-C. To obtain a soluble molecule that could mimic the activity of ron-sema and sema ⫹ PSI domains, we amplified by PCR these regions from RON cDNA and cloned them into expression vectors. On the basis of a consensus sequence for the sema domain (16), we designed PCR primers to amplify a cDNA stretch corresponding to the N-terminal part of the protein up to amino acid Phe-518 (this molecule is referred to as ron-sema in Fig. 1B). A region comprising up to amino acid Ser-578 was amplified for the longer sema ⫹ PSI domain (referred to as ron-PSI). All of the residues in the N-terminal sequence were amplified, starting from the first ATG to include the first 24 amino acids of the predicted signal peptide (9). An expression vector was chosen that provides an N-terminal FLAG but no artificial signal peptide. Western blot in reducing conditions of proteins immunoprecipitated with anti-FLAG antibodies from the conditioned culture medium of HEK293T cells transfected with each construct, respectively, shows two bands of approximately 70 and 35 kDa. The upper bands (Fig. 1B, left panel) correspond to the uncleaved soluble molecules ron-sema and ron-PSI. The lower band (Fig. 1B, left panel) corresponds to the ␣-chain alone (derived from the predicted cleavage at the KRRR site), identifiable through the fusion with the FLAG epitope. The frag-

3728

Dominant-Negative Activity of the RON Sema Domain

FIG. 1. A, schematic representation of the RON receptor and experimental constructs. Red, the sema domain; yellow, PSI domain; orange, IPT domains; black, transmembrane region; blue, kinase domain; gray, cytoplasmic tail. Pink bar, representation of the cell membrane. An unknown number of disulfide bridges connect ␣- and ␤-chain. ron-sema construct includes amino acids 1–518. ron-PSI construct includes amino acids 1–578. PCR primers were designed on the plexin domain described by Pfam (the Protein Families Database at the Sanger Center: www.sanger.ac.uk/Software/Pfam/) (amino acids 526 –568). This domain approximately corresponds with the definition of PSI (16) and MRS (17). Green, FLAG epitope fused by cloning into pFLAG-CMV-2 (Sigma) (see ‘‘Experimental Procedures’’). Orange, 10 amino acids of the first IPT domain are inserted in the construct. B, analysis of the ronsema and ron-PSI proteins produced by transfected HEK293T cells. HEK293T cells were transiently transfected with ron-sema or ron-PSI DNA constructs (see ‘‘Experimental Procedures’’). Both proteins were immunoprecipitated with monoclonal anti-FLAG antibody from 1.5 ml of HEK293T-conditioned tissue culture medium, separated with 8 –16% gradient SDS-PAGE under reducing (left panel) and non-reducing conditions (right panel), and transferred onto a nitrocellulose membrane. Protein bands were visualized by Western blotting with anti-FLAG monoclonal antibody. Control, conditioned tissue culture medium from HEK293T cells transfected with an empty vector. Proteins detected under reducing conditions (left) show the presence of both the unprocessed single chain (upper band, ⬃70 kDa) and processed disulfidelinked heterodimeric ron-sema and ron-PSI (lower band, ␣-chain containing the N-terminal FLAG tag, molecular mass below 40 kDa). IP, immunoprecipitation; WB, Western blotting.

ment of ␤-chain is released after the reduction of the disulfide bonds, which in the mature receptor connect the two chains together, and remains undetectable by anti-FLAG antibodies used in this experiment. The upper band migrates with a molecular weight higher than expected, most probably because of post-translational events such as N-glycosylation and Nmyristoylation predicted by sequence analysis (28). In nonreducing conditions (Fig. 1B, right panel), the soluble molecules, consisting of ␣- and ␤-chain fragment held together by the disulfide bonds, migrate with a molecular mass slightly higher than 70 kDa. These results show that the constructs are

FIG. 2. The sema domain of the RON receptor inhibits MSP binding to RON. A, RE7 cells were stimulated with 5 nM MSP for 15 min. in the presence of blank medium or preconditioned culture medium containing ron-sema-soluble molecule. Cells were then lysed, and the RON receptor immunoprecipitated from cell lysates with anti-RON antibodies. MSP presence in RON precipitates was detected by Western blotting (WB) with anti-MSP antibodies (upper panel). Lower panel, reblotting with anti-RON antibodies to estimate the amount of RON in precipitates (upper band, RON precursor; lower band, mature RON). Arrows on the right represent MSP band and the immunoglobulin G heavy chains band (IgH). Molecular mass markers is indicated on the right (in kDa). The optical density measurement of MSP bands in three independent experiments revealed approximately a 3-fold decrease of the amount of MSP co-immunoprecipitated with RON. B, the effect of medium containing ron-sema on the binding of recombinant (disulfidelinked ␣-␤-chains) MSP or free ␣- and free ␤-chain to recombinant full-length RON adsorbed to the wells of a microtiter plate. ron-sema caused 50 and 63% inhibition of 0.05 nM ␤-chain and ␣-␤-MSP binding to RON respectively, and up to 90% inhibition of 5 nM free ␣-chain binding to RON. Binding was determined by ELISA (see ‘‘Experimental Procedures’’). Similar results were obtained with conditioned medium containing ron-PSI (data not shown). The data are presented as doseresponse curves for the three ligands, each point being the amount of ligand bound in the presence of ron-sema, expressed as a percentage of the amount bound in the absence of ron-sema. Each value represents the mean ⫾ S.E. of three independent experiments.

translated and the resulting proteins are processed as expected with the final product conveyed to the cell membrane and released in the culture medium. The Soluble Molecule ron-sema Interferes with the Ligand Binding Activity of the Full-length Receptor—The extracellular part of RON is known to contain the ligand binding pocket; however, the site of RON interaction with its ligand MSP has not been mapped. We investigated effects of soluble ron-sema (and ron-PSI, data not shown) on the ligand binding activity of RON. The addition of MSP to RON-expressing cells causes the

Dominant-Negative Activity of the RON Sema Domain formation of ligand-receptor complexes. As shown in Fig. 2A, MSP can be co-immunoprecipitated with RON. The presence of ron-sema in the culture medium during cell stimulation with MSP reduces the amount of MSP co-immunoprecipitated with RON (Fig. 2A, lane 4). MSP is a heterodimer comprised of ␣- and ␤-chains linked by a disulfide bond. Two binding sites for RON have been identified in MSP. MSP ␤-chain contains the high affinity binding site (29), whereas the ␣-chain contains a second low affinity receptor-binding site (30). Neither ␣- nor ␤-chain of MSP alone induces RON tyrosine phosphorylation and activation, indicating that both binding sites of MSP are required for biological activity (29). We used a competitive ELISA to investigate the effect of soluble ron-sema on the binding of recombinant disulfidelinked ␣-␤-heterodimeric MSP and free ␣- and ␤-chains of MSP to recombinant RON receptor adsorbed to an ELISA plate (Fig. 2B). The binding of recombinant disulfide-linked ␣-␤-MSP, free ␣-chain, and free ␤-chain to recombinant ron receptor is inhibited by the presence of ron-sema in the reaction mixture (Fig. 2B). The Soluble Molecules ron-sema and ron-PSI Inhibit MSPinduced Signaling and Cell Growth—Our finding that ronsema (Fig. 2B) and ron-PSI (data not shown) inhibits MSPbinding to RON suggested that these soluble molecules might inhibit MSP-induced RON activation via competition for MSP binding. To investigate whether the soluble molecules ronsema and ron-PSI may functionally interfere with the activity of the full-length receptor, we stimulated RON-expressing RE7 cells (26) with MSP in a culture medium containing either molecule (or none as a control). The RON receptor was then immunoprecipitated from these cell lysates with anti-RON antibodies and a Western blot performed with anti-phosphotyrosine antibodies. The result shows a greatly reduced level of phosphorylation of the receptor immunoprecipitated from cells exposed to a conditioned culture medium containing ron-sema or ron-PSI molecules compared with the control (Fig. 3A). Ligand-induced RON-tyrosine phosphorylation is a key event leading to up-regulation of RON catalytic activity and consequent activation of intracellular signaling pathways that mediate the biological effects of MSP (31). A decrease in RON tyrosine phosphorylation by soluble ron-sema or ron-PSI molecules resulted in the inhibition of RON-dependent signaling pathways. In RE7 cells exposed to non-conditioned medium, MSP induces a strong phosphorylation of MAPK (Fig. 3B), a kinase known to be part of signaling pathways that control MSP-dependent cell proliferation (reviewed in Refs. 32–34). When RE7 cells are exposed to ron-sema or ron-PSI, the level of MSPinduced MAPK phosphorylation is strongly reduced (Fig. 3B). This inhibiting effect on MAPK phosphorylation translates directly into cell proliferation inhibition. A growth rate experiment was performed on RON-transfected cells stimulated with MSP. Under the experimental conditions described, MSP has a strong mitogenic effect that is greatly reduced in the presence of soluble ron-sema and ron-PSI with ron-PSI causing a 3-fold reduction of growth compared with the control (Fig. 3C). These results show that the secreted soluble molecules designed over the sema and sema ⫹ PSI domains of RON are capable of inhibiting RON phosphorylation and therefore activation induced by MSP. MSP-induced RON activation in tumor cells including breast carcinoma-derived ZR75.1 cell line, colon cancer cell line HCT116, and lung carcinomas correlates with cell proliferation, migration, and invasion (14, 35–37). Thus, ligand-induced RON stimulation can be associated with tumor progression. We

3729

FIG. 3. Supernatants containing ron-sema and ron-PSI inhibit MSP-induced RON signaling and cell growth. A, tyrosine phosphorylation of RON induced by MSP. RE7 cells were stimulated with 5 nM MSP for 15 min in the presence of blank or conditioned culture medium containing ron-sema or ron-PSI-soluble molecules. Cells were then lysed, and the RON receptor immunoprecipitated from cell lysates with anti-RON antibodies. RON tyrosine phosphorylation was detected by Western blotting (WB) with anti-phosphotyrosine (PY) antibodies (upper panel). Lower panel, reblotting with anti-RON antibodies to estimate the amount of RON protein in precipitates (upper band, RON precursor; lower band, mature RON). Molecular mass markers are indicated on the right (in kDa). B, MAPK activation induced by MSP. RE7 cells were stimulated with MSP as described above. Activation of MAPK was determined in cell lysates with anti-phospho-MAPK antibodies (upper panel). The amount of MAPK (lower panel, p42 and p44 ERK1/ERK2) was estimated by reprobing the membrane with antiMAPK antibodies. Molecular mass markers are indicated in kDa to the right. C, cell proliferation induced by MSP. RE7 cells (2 ⫻ 103/well) were distributed in triplicate in 96-well plate and incubated with or without 5 nM MSP in serum-free Dulbecco’s modified Eagle’s medium in the presence of blank medium or conditioned medium containing either ron-sema or ron-PSI-soluble molecules. Cell number was measured after 72 h of cell culturing by adding MTT to culture wells and measuring A570 2 h later. The number of cells was determined with a standard calibration curve. Error bars, means ⫹ S.E. for three independent experiments. The inhibitory effect of ron-sema on RON and MAPK phosphorylation (Fig. 3, A and B) seems stronger that its effect on cell proliferation (Fig. 3C) and vice versa for ron-PSI. This difference could be because of the different duration of the biochemical (15 min) and proliferation studies (3 days). In these conditions, it is likely that the relative stability of soluble ron-sema and ron-PSI are different and certainly affected by the cell type-specific proteases secreted in the culture medium (compare Figs. 3C and 4B).

investigated the effect of soluble ron-sema and ron-PSI on MSP-dependent RON activation and cell proliferation of the colon cancer cell line HCT116. In this cell line, MSP induces RON-tyrosine phosphorylation (Fig. 4A) and promotes cell growth (Fig. 4B). The presence of

3730

Dominant-Negative Activity of the RON Sema Domain

FIG. 4. Supernatants containing ron-sema and ron-PSI proteins inhibit MSP-induced RON tyrosine phosphorylation and cell growth of HCT116 colon cancer cells. A, RON tyrosine phosphorylation induced by MSP. HCT116 cells were stimulated with 5 nM MSP for 15 min in the presence of blank medium or conditioned media containing either the ron-sema or ron-PSI-soluble molecules. Cells were then lysed, and RON receptor immunoprecipitated from cell lysates with anti-RON antibodies. RON tyrosine phosphorylation was detected by Western blotting (WB) with anti-phosphotyrosine (PY) antibodies (upper panel). Lower panel, reblotting with anti-RON antibodies to estimate RON amount in precipitates (upper band, RON precursor; lower band, mature RON). Molecular mass markers (in kDa) are indicated on the right. B, cell proliferation induced by MSP. HCT116 cells (2 ⫻ 103/well) were distributed in duplicates in 24-well plates and incubated with or without 5 nM MSP in serum-free McCoy medium in the presence of blank medium or conditioned media containing either ron-sema or ron-PSI. Cell number was measured after 72 h of culturing by adding MTT to culture wells and measuring A570 after 2 h. The cell number was determined from a standard calibration curve. Error bars, means ⫹ S.E. for two independent experiments. In the naturally RONexpressing HCT116 cells, the inhibitory effect on RON phosphorylation of ron-sema seems weaker than the ron-PSI effect. This is the opposite of what is shown in the transfected cells RE7. This difference shows that possibly RON activation could be inhibited not only via blocking MSP binding but also by interfering with its binding to other molecules (such as integrins, HYAL2, and so forth). Also, a cell-type specific pattern of post-translational modification of RON may affect a relative sensitivity to the presence of ron-sema and ron-PSI.

either ron-sema or ron-PSI inhibits MSP-mediated RON activation (Fig. 4A) and cell growth (Fig. 4B). These results suggest that the dominant-negative molecules ron-sema and ron-PSI may have a therapeutic application. They may prevent MSP/ RON-mediated cancer cell growth. ron-sema Effect Is Highly Specific for the RON Receptor— RON and MET have overall 33% identity and share the same modular organization, including the presence of sema and PSI domains in their extracellular part. To rule out the possibility that ron-sema might interfere non-specifically with MET (which shares the highest homology with RON) or other growth factor receptors, we incubated MET-ligand HGF in precondi-

FIG. 5. The ron-sema-soluble molecule does not interfere with HGF-induced activation of MET (A) and with EGF-induced activation of EGFR (B). A, HGF-induced MET tyrosine phosphorylation. MDCK cells expressing endogenous MET were starved overnight in serum-free medium and then stimulated with 100 ng/ml HGF for 15 min. in the presence of blank medium or conditioned medium containing the ron-sema-soluble molecule. Cells were then lysed, and the MET receptor was immunoprecipitated from cell lysates with anti-MET antibodies. MET tyrosine phosphorylation was detected by Western blotting (WB) with anti-phosphotyrosine (PY) antibodies (upper panel). Lower panel, re-blotting with anti-MET antibodies to estimate the amount of MET in precipitates (upper band, MET precursor; lower band, mature MET). Molecular mass markers (in kDa) are indicated to the right. B, EGF-induced EGFR tyrosine phosphorylation. MDCK cells were starved in serum-free medium and then stimulated by 50 ng/ml EGF for 15 min in the presence of blank medium or conditioned medium containing ron-sema. After cell lysis, the EGFR receptor was immunoprecipitated from cell lysates by anti-EGFR antibodies. EGFR tyrosine phosphorylation was detected by WB with PY antibodies (upper panel). Lower panel, reblotting with anti-EGFR antibodies to estimate the amount of EGFR in precipitates. Molecular mass markers (in kDa) are indicated to the right. Similar results were obtained with ron-PSIsoluble molecule (data not shown).

tioned culture medium with or without ron-sema (or ron-PSI, data not shown). In this experiment, MET was immunoprecipitated with antiMET antibodies from cell lysates of MDCK (cells naturally expressing MET) previously treated with HGF. Western blotting performed with anti-phosphotyrosines antibodies (Fig. 5A) shows that HGF increases MET tyrosine phosphorylation and ron-sema does not affect MET phosphorylation. Similar results were obtained with HEK 293 cells transiently transfected with human MET cDNA (data not shown). As a further control, the effect of ron-sema on EGF-induced EGFR phosphorylation was investigated. Cells expressing EGFR were stimulated with EGF in the presence or absence of ron-sema. As shown (Fig. 5B), the presence of ron-sema has no effect on the receptor phosphorylation triggered by EGF.

Dominant-Negative Activity of the RON Sema Domain These results show that the soluble molecule ron-sema does not affect ligand-stimulated phosphorylation of other growth factor receptors such as EGFR and, more importantly, MET, which has a highly homologous sema domain in the extracellular region. DISCUSSION

This paper reports experimental evidence that secreted soluble proteins representing the sema or sema ⫹ PSI domains of the tyrosine kinase receptor RON have highly specific dominant-negative effects on ligand-induced RON activation and downstream signaling events. The sema domain is represented by a block of ⬃500 amino acids with 15 conserved cysteine residues spaced by stretches of conserved amino acids and a conserved potential glycosylation site (18). A similar domain is also present in semaphorins and plexins (a class of semaphorin receptors). At least in plexins, it was shown previously (38) that the cysteine-rich domain present in the extracellular region of the receptor mediates cell adhesion via plexin-plexin interaction, i.e. through homophilic binding. It would be interesting to investigate whether the PSI domain present on integrin ␤-chain exerts a direct interaction with the sema or PSI domains of RON in the context of the MSP-mediated integrin-dependent mechanism of cell adhesion and motility in which RON is involved (39). The observation (Fig. 2A) that the presence of soluble ronsema proteins in the medium inhibits RON activation suggests that the ron-sema molecule interferes with MSP binding, either by sequestering MSP or by competitive binding to the sema domain of the receptor itself. Further experiments are needed to clarify this point. The observation that RON can be expressed on the cell surface as a non-covalently linked dimer (40) raises the possibility that RON-RON association might be mediated by homophilic interactions between sema or PSI domains or both of two RON molecules. However, using cross-linking reagents, we were unable to detect a complex between full-length transmembrane RON and ron-sema or ron-PSI molecules (data not shown). These results argue that other parts of the receptor might be necessary for homophilic dimerization. Possibly, the sema and/or PSI domain could also mediate interaction with other sema domain-containing proteins. Recently, a complex between Plexin B1 (a semaphorin receptor) and the MET receptor, co-expressed on the surface of the same cell, has been detected (41). The binding of SEMA4D to Plexin B1 up-regulates MET kinase activity and induces tyrosine phosphorylation of both MET and Plexin B1 receptors. The receptor activation results in triggering cell dissociation and invasive growth. Cells lacking MET expression did not respond to SEMA4D (41). Plexin B1 and MET association is specific, non-covalent, and requires the extracellular domains of both receptors (41). These data suggest that the sema domain of MET might be responsible for association with Plexin B1. At the present time, there are no sema-containing proteins identified as RON partners. However, the observations that (i) RON knock-out is lethal but MSP (RON ligand) knock-out is not (42, 43) and (ii) Met-interacting semaphorins have been identified (41) suggest that in addition to MSP, RON might interact with other ligands or may have other sema-containing partners during normal development. We hypothesized and investigated the possibility that RON sema-domain determines the ligand-binding specificity of the full-length receptor. MSP is a protein that belongs to the plasminogen-related kringle protein family (44). It is most closely related to hepatocyte growth factor. Biologically active MSP is a disulfide-linked ␣-␤-heterodimer (31) that carries two bind-

3731

ing sites for RON, a low affinity site and a high affinity site located on the ␣- and the ␤-chains, respectively (27, 29). Although free ␣- and ␤-chains can bind to RON, only mature heterodimeric MSP induces RON kinase activation and cellular responses (29). According to the proposed model, a single molecule of MSP induces RON dimerization (30). In this work, we found that possibly both low and high affinity MSP binding sites are localized in the RON sema domain (Fig. 2). This finding, together with the proposed MSP/RON interaction model (31), suggests that soluble ron-sema may inhibit MSP/ RON-mediated cellular responses via competition for MSP binding. We found that indeed incubation of RON-expressing cells with soluble ron-sema or ron-PSI causes a decrease of MSPinduced RON phosphorylation (Figs. 3A and 4A) and RON-dependent MAPK activation (Fig. 3B) along with the inhibition of MSP-mediated cell growth (Fig. 4B). The activity of soluble ron-sema and ron-PSI is specific, because there is no effect on MET or EGF receptor ligand-induced tyrosine phosphorylation (Fig. 5). Thus, the soluble molecule corresponding to the sema domain of RON, expressed separately from the receptor protein scaffold, can act as a specific dominant-negative regulator of RON activity. Truncated dominant-negative receptor forms on the cell surface, tissue culture supernatants, and biological fluids have been detected for a number of growth factor receptors including Met (45, 46). C-terminally truncated receptors (lacking the kinase domain) can interfere with ligand-induced stimulation via the formation of inactive heterodimers when they are expressed together with the full-length receptors or by competing for ligand binding when they are present in soluble forms. Two C-terminal truncated forms of MET were identified (45). They carry an intact ligand-binding domain but lack the kinase domain because of truncation of the ␤-chain and arise from alternative post-transcriptional processing of the mature form. One truncated form is soluble and released from the cells (45). These isoforms may have a physiological role in down-regulating MET functional activity. Truncated naturally occurring forms of RON lacking the kinase domain have not yet been identified. The ability of ron-sema and ron-PSI to inhibit ligand-dependent RON activation and subsequent MSP/RON-mediated biological responses in vitro suggests a potential therapeutic use of such soluble molecules. RON expression has been detected in a number of cancer cell lines and tumors (reviewed in Ref. 15), whereas it is barely detectable in the corresponding normal cells. Autocrine and paracrine MSP-dependent mechanisms of RON activation could contribute to tumor growth and metastasis. Simultaneous RON and MSP expression has been found in a subset of non-small lung carcinomas (14). MSP supports cell growth and migration of breast and colon cancer cell lines (35, 36). Thus, interfering with MSP/RON autocrine/paracrine loops might be an approach for therapy of MSP-sensitive tumors. In the present work, we investigated the effect of the soluble molecule ron-sema on growth of the colon cancer cell line HCT116. It was shown that MSP stimulation induces RON tyrosine-phosphorylation and invasive growth of HCT116 cells (36). Our results are in agreement with previously published data (36). Moreover, the incubation in conditioned media containing soluble molecules ron-sema or ron-PSI inhibited MSPdependent RON tyrosine phosphorylation (Fig. 4A) and cell growth (Fig. 4B) of HCT116 cancer cells. In conclusion, we demonstrated that the sema domain of RON (which presumably contains the MSP binding sites) might serve as a dominant negative form of the RON receptor via competition for MSP binding to the full-length trans-mem-

3732

Dominant-Negative Activity of the RON Sema Domain

brane RON, thus possibly preventing RON-RON active dimer formation. These data suggest a potential therapeutic use of the soluble RON sema domain for preventing proliferation and migration of MSP-sensitive cancer cells and therefore inhibiting growth of MSP-sensitive types of cancer. Acknowledgment—We thank Massimiliano Andreazzoli for critical reading of this paper. REFERENCES 1. Schmidt, L., Duh, F. M., Chen, F., Kishida, T., Glenn, G., Choyke, P., Scherer, S. W., Zhuang, Z., Lubensky, I., Dean, M., Allikmets, R., Chidambaram, A., Bergerheim, U. R., Feltis, J. T., Casadevall, C., Zamarron, A., Bernues, M., Richard, S., Lips, C. J., Walther, M. M., Tsui, L. C., Geil, L., Orcutt, M. L., Stackhouse, T., Zbar, B., et al. (1997) Nat. Genet. 16, 68 –73 2. Zbar, B. (2000) Semin. Cancer Biol. 10, 313–318 3. Danilkovitch-Miagkova, A., and Zbar, B. (2002) J. Clin. Invest. 109, 863– 867 4. Danilkovitch-Miagkova, A., Duh, F. M., Kuzmin, I., Angeloni, D., Liu, S. L., Miller, A. D., and Lerman, M. I. (2003a) Proc. Natl. Acad. Sci. U. S. A. 100, 4580 – 4585 5. Wang, M. H., Wang, D., and Chen, Y. Q. (2003) Carcinogenesis 24, 1291–1300 6. Gaudino, G., Avantaggiato, V., Follenzi, A., Acampora, D., Simeone, A., and Comoglio, P. M. (1995) Oncogene 11, 2627–2637 7. Angeloni, D., Danilkovitch-Miagkova, A., Ivanov, S. V., Breathnach, R., Johnson, B. E., Leonard, E. J., and Lerman, M. I. (2000) Genes Chromosomes Cancer 29, 147–156 8. Okino, T., Egami, H., Ohmachi, H., Takai, E., Tamori, Y., Nakagawa, K., Nakano, S., Akagi, J., Sakamoto, O., Suda, T., and Ogawa, M. (1999) Int. J. Oncol. 15, 709 –714 9. Ronsin, C., Muscatelli, F., Mattei, M. G., and Breathnach, R. (1993) Oncogene 8, 1195–1202 10. Iwama, A., Wang, M. H., Yamaguchi, N., Ohno, N., Okano, K., Sudo, T., Takeya, M., Gervais, F., Morissette, C., Leonard, E. J., et al. (1995) Blood 86, 3394 –3403 11. Kurihara, N., Iwama, A., Tatsumi, J., Ikeda, K., and Suda, T. (1996) Blood 87, 3704 –3710 12. Banu, N., Price, D. J., London, R., Deng, B., Mark, M., Godowski, P. J., and Avraham, H. (1996) J. Immunol. 156, 2933–2940 13. Wang, M. H., Dlugosz, A. A., Sun, Y., Suda, T., Skeel, A., and Leonard, E. J. (1996) Exp. Cell Res. 226, 39 – 46 14. Willett, C. G., Wang, M. H., Emanuel, R. L., Graham, S. A., Smith, D. I., Shridhar, V., Sugarbaker, D. J., and Sunday, M. E. (1998) Am. J. Respir. Cell Mol. Biol. 18, 489 – 496 15. Danilkovitch-Miagkova, A. (2003b) Curr. Cancer Drug Targets 3, 31– 40 16. Bork, P., Doerks, T., Springer, T. A., and Snel, B. (1999) Trends Biochem. Sci. 24, 261–263 17. Artigiani, S., Comoglio, P. M., and Tamagnone, L. (1999) IUBMB Life 48, 477– 482 18. Kolodkin, A. L., Matthes, D. J., and Goodman, C. S. (1993) Cell 75, 1389 –1399 19. Hall, K. T., Boumsell, L., Schultze, J. L., Boussiotis, V. A., Dorfman, D. M., Cardoso, A. A., Bensussan, A., Nadler, L. M., and Freeman, G. J. (1996) Proc. Natl. Acad. Sci. U. S. A. 93, 11780 –11785 20. Delaire, S., Elhabazi, A., Bensussan, A., and Boumsell, L. (1998) Cell Mol. Life

Sci. 54, 1265–1276 21. Behar, O., Golden, J. A., Mashimo, H., Schoen, F. J., and Fishman, M. C. (1996) Nature 383, 525–528 22. Yamada, T., Endo, R., Gotoh, M., and Hirohashi, S. (1997) Proc. Natl. Acad. Sci. U. S. A. 94, 14713–14718 23. Christensen, C. R., Klingelhofer, J., Tarabykina, S., Hulgaard, E. F., Kramerov, D., and Lukanidin, E. (1998) Cancer Res. 58, 1238 –1244 24. Feiner, L., Koppel, A. M., Kobayashi, H., and Raper, J. A. (1997) Neuron 19, 539 –545 25. Wang, M. H., Skeel, A., Yoshimura, T., Copeland, T. D., Sakaguchi, K., and Leonard, E. J. (1993) J. Leukocyte Biol. 54, 289 –295 26. Wang, M. H., Ronsin, C., Gesnel, M. C., Coupey, L., Skeel, A., Leonard, E. J., and Breathnach, R. (1994a) Science 266, 117–119 27. Danilkovitch, A., Miller, M., and Leonard, E. J. (1999) J. Biol. Chem. 274, 29937–29943 28. Bairoch, A., Bucher, P., and Hofmann, K. (1997) Nucleic Acids Res. 25, 217–221 29. Wang, M. H., Julian, F. M., Breathnach, R., Godowski, P. J., Takehara, T., Yoshikawa, W., Hagiya, M., and Leonard, E. J. (1997) J. Biol. Chem. 272, 16999 –17004 30. Miller, M., and Leonard, E. J. (1998) FEBS Lett. 429, 1–3 31. Wang, M. H., Yoshimura, T., Skeel, A., and Leonard, E. J. (1994b) J. Biol. Chem. 269, 3436 –3440 32. Danilkovitch-Miagkova, A., and Leonard, E. J. (2001) Apoptosis 6, 183–190 33. Santoro, M. M., Penengo, L., Minetto, M., Orecchia, S., Cilli, M., and Gaudino, G. (1998) Oncogene 17, 741–749 34. Santoro, M. M., Collesi, C., Grisendi, S., Gaudino, G., and Comoglio, P. M. (1996) Mol. Cell. Biol. 16, 7072–7083 35. Maggiora, P., Marchio, S., Stella, M. C., Giai, M., Belfiore, A., De Bortoli, M., Di Renzo, M. F., Costantino, A., Sismondi, P., and Comoglio, P. M. (1998) Oncogene 16, 2927–2933 36. Chen, Y. Q., Zhou, Y. Q., Angeloni, D., Kurtz, A. L., Qiang, X. Z., and Wang, M. H. (2000) Exp. Cell Res. 261, 229 –238 37. Willett, C. G., Smith, D. I., Shridhar, V., Wang, M. H., Emanuel, R. L., Patidar, K., Graham, S. A., Zhang, F., Hatch, V., Sugarbaker, D. J., and Sunday, M. E. (1997) J. Clin. Invest. 99, 2979 –2991 38. Ohta, K., Mizutani, A., Kawakami, A., Murakami, Y., Kasuya, Y., Takagi, S., Tanaka, H., and Fujisawa, H. (1995) Neuron 14, 1189 –1199 39. Danilkovitch-Miagkova, A., Angeloni, D., Skeel, A., Donley, S., Lerman, M., and Leonard, E. J. (2000) J. Biol. Chem. 275, 14783–14786 40. Follenzi, A., Bakovic, S., Gual, P., Stella, M. C., Longati, P., and Comoglio, P. M. (2000) Oncogene 19, 3041–3049 41. Giordano, S., Corso, S., Conrotto, P., Artigiani, S., Gilestro, G., Barberis, D., Tamagnone, L., and Comoglio, P. M. (2002) Nat. Cell Biol. 4, 720 –724 42. Waltz, S. E., Eaton, L., Toney-Earley, K., Hess, K. A., Peace, B. E., Ihlendorf, J. R., Wang, M. H., Kaestner, K. H., and Degen, S. J. (2001) J. Clin. Invest. 108, 567–576 43. Muraoka, R. S., Sun, W. Y., Colbert, M. C., Waltz, S. E., Witte, D. P., Degen, J. L., and Friezner-Degen, S. J. (1999) J. Clin. Invest. 103, 1277–1285 44. Yoshimura, T., Yuhki, N., Wang, M. H., Skeel, A., and Leonard, E. J. (1993) J. Biol. Chem. 268, 15461–15468 45. Prat, M., Crepaldi, T., Gandino, L., Giordano, S., Longati, P., and Comoglio, P. (1991) Mol. Cell. Biol. 11, 5954 –5962 46. Crepaldi, T., Prat, M., Giordano, S., Medico, E., and Comoglio, P. M. (1994) J. Biol. Chem. 269, 1750 –1755