Eur. J. Biochem. 204,857-864 (1992) 0FERS 1992

Characterization of hepatocyte-growth-factor receptors on Meth A cells Masayuki KOMADA', Keiji MIYAZAWA', Takehisa ISHI12 and Naomi KITAMURA' I Institute for Liver Research, Kansai Medical University, Osaka, Japan Research Center, Mitsubishi Kasei Corporation, Yokohama, Japan (Receivcd September 30/November 22, 1991)

-

EJB 91 1305

Hepatocyte growth factor (HGF) is a heparin-binding polypeptide mitogen for a variety of cell types including hepatocytes. H G F also has cytotoxic activity on some tumor cell lines as well as scattering activity on epithelial cells. In this study, recombinant human HGF was used to identify HGF-binding cell surface receptors on Meth A cells, whose growth is inhibited by HGF. Scatchard analysis of binding data indicated that there were two classes of binding sites with high affinity (Kd = 37 pM) and low affinity (Kd = 6.7 nM) and the average numbers were 6600 and 2600000 per cell, respectively. Affinity cross-linking of "'1-HGF to Meth A cells resulted in a major and a minor specifically labeled complex. Competition analysis followed by cross-linking indicated that the HGFbinding proteins were involved in the formation of the high-affinity binding. The existence of the two HGF-binding surface proteins was confirmed by HGF-dependent immunoprecipitation of the binding proteins with an anti-HGF polyclonal antibody. The molecular masses of the major and the minor surface proteins were 160 kDa and 130 kDa, respectively. The 160-kDa protein was autophosphorylated in vitro on tyrosine residue and was immunoprecipitated with an antiserum against the cmet proto-oncogene product. These results indicate that the 160-kDa HGF-binding surface protein on Meth A cells is the c-met protein. Furthermore, tyrosine phosphorylation of the c-met protein was stimulated by H G F treatment of Meth A cells, suggesting that it may be involved in the signal transduction of the growth inhibition of Meth A cells by HGF.

the H G F sequence is similar to that of plasminogen [13- 161. The heavy chain contains four kringle structures and the light chain is similar to the catalytic domain of serine proteases. Analysis of the genomic DNA encoding human H G F showed that the exon-intron arrangement is similar to that of human plasminogen, suggesting that the genes for human H G F and plasminogen have arisen through gene duplication events from an ancestral gene 137, 181. Like other polypeptide growth factors, HGF likely initiates its biological effects by interacting with specific receptors on the surface of its target cells. Accordingly, the different biological effects of H G F may be mediated by different intracellular signal cascades. However, the possibility cannot be ruled out that those effects are mediated by different cell surface receptors. It is therefore important to identify specific H G F receptors on the surface of its different target cells. Zarnegar et al. reported specific binding sites for HGF on rat hepatocytes [19]. They identified a 160-kDa receptor on rat hepatocytes by analyzing the affinity cross-linked lz5I-HGFreceptor complex. Higuchi and Nakamura also reported a similar molecule of 220 kDa on rat liver plasma membranes [20]. However, these molecules have not been further charac____ terized. Recently, Bottaro et al. identified an HGF receptor Correspondence to N. Kitamura, Institute for Liver Research, of the BSj589 human mammary epithelial cell line as the cKansai Medical University, Moriguchi, Osaka, Japan 570 Abbreviations. HGF, hepatocyte growth factor; MTT, 3-(4,s- met protein [2I]. They, however, used a smaller form of HGF, dimethyIthiazol-2-yl)-2,S-diphenyltetrazoliumbromide; PhMeS02F, designated HGFp28, as the cross-linking ligand. More rephcnylmethylsulfonyl fluoride; BS3, bis(sulfosuccinimidy1) suberale. cently, Naldini et al. demonstrated direct interaction of full-

Hepatocyte growth factor (HGF) was initially isolated from human and rabbit plasma and rat platelets as a potent mitogen for adult rat hepatocytes in primary culture [l -31; it is thought to be a regulator of liver regeneration following hepatic injury [4]. Recently, the samc factor was purified from the culture medium of human embryonic lung fibroblasts as a mitogen for melanocytes, endothelial and epithelial cells [ 5 ] . In addition, purified, as well as recombinant H G F has mitogenic activity on melanocytes, keratinocytes, and a variety of epithelial cells [6 - lo]. Moreover, identity of H G F with a tumor cytotoxic factor as well as scatter factor has been reported. The tumor cytotoxic factor was purified from the culture medium of human embryonic lung fibroblasts and has cytotoxic activity on some tumor cell lines [ll].Scatter factor, purified from a variety of fibroblasts, dissociates and increases the motility of epithelial cells [12]. Thus, H G F has a broad spectrum of activities on its different target cells. HGF is synthesized as a single 90-kDa precursor, which is glycosylated and cleaved at a specific proteolytic site to yield a 65-kDa heavy chain and a 35-kDa light chain [I]. The two chains are linked together by a disulfide bond. Molecular cloning and sequence analysis of HGF cDNA revealed that

858 size HGF with the c-met protein of the A549 human lung carcinoma cell line and the GTL16 gastric carcinoma cell line

P21.

In the present study, we used full-size HGF as a ligand to identify cell surface binding proteins on the Meth A mouse sarcoma cell line. We show that HGF binds Meth A cells specifically with high and low affinity. We further demonstrated that cell-bound H G F is associated with membrane proteins and one of them is the c-mel protein.

EXPERIMENTAL PROCEDURES

incubated for 15 s at room temperature. The reaction was stopped by adding 100 pl of 2.5 mg/ml sodium metabisulfite in 50 mM sodium phosphate pH 7.4 containing 0.5 M NaCI, and the labeled protein was separated from free iodine by gel filtration on a Sephadex G-25 PDlO column equilibrated with Eagle's minimum essential medium, 25 mM Hepes/NaOH pH 7.4, 0.5% bovine serum albumin (binding medium) containing 5 mM KI. The '251-HGF fraction was sterilized by filtration and stored at 4°C. The specific activity of the recombinant human '251-HGF was about lo8 cpm/pg protein. The purity of '251-HGF was examined by SDS/PAGE as described by Laemmli [26].

Materials Reagents were obtained as follows: Na['251] from Amersham (IMS 30); [y-32P]ATP(6000 Ci/mmol) from NEN; bis(sulfosuccinimidy1) suberate (BS3) and Iodobeads from Pierce Chemical Co.; Sephadex G-25 PDlO column and protein-A - Sepharose from Pharmacia LKB Biotechnology Inc. ; 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT), phenylmethylsulfonyl fluoride (PhMeS02F), aprotinin, pepstatin A, soybean trypsin inhibitor, phosphoserine, phosphothreonine, and phosphotyrosine from Sigma; leupcptin from Peptide Institute Inc. ; anti-phosphotyrosine monoclonal antibody, PY20, from ICN ImmunoBiologicals; Trans-Blot polyvinylidene difluoride membrane from BioRad ; peroxidase-conjugated rabbit anti-(mouse Ig) and porcine anti-(rabbit Ig) monoclonal antibody from Dako; and TM Blue chromogenic substrate for peroxidase from Transgenic Sciences, Inc. Recombinant human HGF was prepared as described previously [23]. An anti-HGF polyclonal antibody was prepared by immunizing a rabbit with recombinant human HGF and purified by protein A affinity chromatography. An antiserum against the c-met protein (anti-met antiserum) was prepared as follows. The peptide corresponding to the C-terminal 21 amino acids of mouse c-met protein was synthesized using a model 430A peptide synthesizer (Applied Biosystems) and conjugated to keyhole limpet hemocyanin (Sigma) using succinimidyl 4-(pmaleimidophenyl) butyrate (Pierce Chemical Co.). The conjugate was injected into a rabbit and an antiserum was collected. Meth A mouse sarcoma cells were provided from National Cancer Center and were cultured in RPMI 1640 supplemented with 10% fetal calf serum.

Cell growth assay Cells ( 5 x lo2)were cultured for 5 days with various concentrations of HGF in I00 pl RPMI 1640 supplemented with 10% fetal calf serum in 96-well flat bottomed plates [9]. Half the volume of the medium was changed every 24 h to supplement HGF. The colorimetric assay for viable cell number was processed as described by Mosmann (241. Briefly, 10 p1 of 5 mg/ml MTT was added to the cell culture and incubated for 1.5 h at 37 "C, then mixed thoroughly with 110 p140 mM HC1 in isopropanol and the A570was measured.

Radioiodination of recombinant human HGF

'

251-labeledrecombinant human HGF was prepared by chloramine-T-mediated iodination [25]. Recombinant human HGF (20 pg) was mixed with 1 mCi Na['251] in 100 p10.5 M sodium phosphate pH 7.4 containing 0.5 M NaCl, then 100 pl of 1 mg/ml chloramine T i n 50 mM sodium phosphate pH 7.4 containing 0.5 M NaCl was added to the mixture, which was

lz5I-HGF binding assay Cells ( 5 x lo4) were incubated in binding medium for 30 min at 37"C, then mixed with '251-HGF in the presence or absence of a 200-fold excess of unlabeled HGF in a final volume of 200 p1 binding medium, and incubated on ice for 2 h. After the incubation, the reaction mixture was layered onto 180 p1 phthalate oil [dibutylphthalate/bis(2-ethylhexyl)phthalate (1.5: l)] and centrifuged at 10000x g for 5 min. The sample was frozen and cut just above the cell pellet, and the cell-associated radioactivity was measured in a y-counter. The specific binding was calculated after subtracting the radioactivity of the samples containing a 200-fold excess of unlabeled H G F (nonspecific bound radioactivity).

'z51-HGF affinity cross-linking Cells ( 5 x lo6) were incubated with 50 pM 12'1-HGF on ice for 2 h, then washed twice with ice-cold 10 mM sodium phosphate pH 7.4 containing 140 mM NaCl (NaCI/P,) and treated with 100 pg/ml BS3 in 0.5 ml 10 mM sodium phosphate pH 8.3 containing 140 mM NaCl and 1 mM MgClz for 20 rnin on ice. The reaction was terminated by adding 0.5 ml 25 mM Tris/HCI pH 7.4, 140 mM NaCl and 1 mM EDTA. The cells were recovered by centrifugation at 2000 x g for 3 rnin at 4°C and lysed in 50 p1 lysis buffer (50 mM Tris/HCl pH 7.4 containing 140 mM NaCI, 0.5% Nonidet P-40, 1 mM EDTA, 2 mM PhMeS02F, 1 pg/ml leupeptin, 10 pg/ml aprotinin, 10 pg/ml pepstatin A and 100 pg/ml soybean trypsin inhibitor) on ice for 30 min. The cell lysate was centrifuged at 12000 x g for 30 rnin at 4°C and the supernatant was mixed with an equal volume of 2 x S D S sample buffer (1 xSDS sample buffer = 62.5 mM Tris/HCl pH 6.8, 5% SDS, 10Y0 glycerol and 0.0025% bromophenol blue), boiled for 5 min, and separated by SDSjPAGE (6% polyacrylamide gel). The 251-HGF-binding-protein complexes were detected by autoradiography .

Cell surface labeling and HGF-dependent immunoprecipitation of HGF-binding proteins by an anti-HGF polyclonal antibody Cells (2.5 x lo7) were suspended in 0.5 ml NaCl/P,. Na['251] (0.5 mCi) was incubated with Iodobeads in 100 p1 50 mM Tris/HCl pH 7.4 containing 150 mM NaCl for 5 rnin at room temperature, then mixed with the cell suspension according to the manufacture's procedure. The mixture was incubated for 30 rnin at room temperature with gentle agitation, then the Iodobeads were removed and the cells were washed three times with NaCI/Pi containing 10 mM KI. The iodinated cells were lysed in 0.5 ml lysis buffer. The cell lysate was centrifuged at 100000x g for 1 h at 4°C and the supernatant was incubated for 4 h at 4°C in the presence or absence

859

of 500 ng HGF. After the incubation, the lysate was incubated with 20 pg anti-HGF polyclonal antibody coupled to 20 p1 protein-A-Sepharose for 16 h at 4°C with gentle agitation, then washed three times with 10 mM Tris/HCl pH 7.4, 140 mM NaCl and 0.1% Nonidet P-40 (washing buffer). The immunoprecipitates were separated by SDSlPAGE on 6% (nonreduced) or 10% (reduced) polyacrylamide gel. Cellsurface HGF-binding proteins were detected by autoradiography. Otherwise, the sample was separated by nonreducingjreducing two-dimensional SDSjPAGE as described below. Nonreducing/reducing two-dimensional SDS/PAGE

Nonreducing/reducing two-dimensional SDSjPAGE was performed by the method of Sommer and Traut 1271. Briefly, ' 251-labeledHGF-binding proteins, which were HGF-dependently immunoprecipitated by an anti-HGF polyclonal antibody as described above, were first electrophoresed on an SDS/polyacrylamide disk gel (6% polyacrylamide) under nonreducing conditions. After the electrophoresis, the disk gel was equilibrated with 62.5 mM Tris/HCl pH 6.8, 2.3% SDS, 5% 2-mercaptoethanol, 10% glycerol and 0.01% bromophenol blue for 30 min at room temperature, fixed on top of the SDS/polyacrylamide slab gel (10 YO polyacrylamide) for the second dimension using melted 1Yo agarose containing 125 mM Tris/HC1 pH 6.8 and 0.1% SDS, and electrophoresed. 1251-labeledHGF-binding proteins were detected by autoradiography.

In vitro phosphorylation assay Cells (2 x lo6) were incubated with 1 nM HGF on ice for 2 h. The cells were washed and treated with or without crosslinking reagent, BS3, then lysed in 0.5 ml lysis buffer as described above. The 100000 x g supernatant of the lysate was immunoprecipitated with an anti-HGF polyclonal antibody then washed three times with washing buffer; 40 p1 20 mM Hepes/NaOH pH 7.4 containing 140 mM NaCl, 5 mM MgC12, 5 mM MnC12, 0.1% Nonidet P-40, 10 yM ATP and 10 pCi [Y-~'P]ATPwas added to the washed protein-ASepharose and the mixture was incubated for 20 min at 30°C. The protein-A- Sepharose was washed once with washing buffer and the immunoprecipitates were separated by SDS/PAGE. Phosphoproteins were detected by autoradioWPhY. Phosphorylated amino acid analysis

The band corresponding to the 140-kDa subunit of the 160-kDa HGF-binding protein, which was phosphorylated in vitro by [Y-~'P]ATP,was excised from the gel and homogenized in 200 pl water. 200 p1 0.1 M NH4HC03, 0.2% SDS, 100 pg/ ml bovine serum albumin and 6% 2-mercaptoethanol was added and the gel fragments were boiled for 10 min, shaken gently overnight at 37°C. The eluted protein was precipitated by trichloroacetic acid (20%), washed with ice-cold acetone three times, and dried under vacuum. The dried pellet was suspended in 100 pl 6 M HCI and hydrolyzed at 105°C for 2 h. The acid hydrolysates were lyophilized and analyzed by electrophoresis on Whatman 3MM paper in pyridine/acetic acidjwater (1:10: 189, by vol.) at 800 V for 2 h. [32P]Phosphorylated amino acids were detected by autoradiography and corresponding standards by staining with ninhydrin.

100, h

-f

0

100

200

300

400

500

HGF Added (pM)

Fig. 1. Effect of HGF on the growth of Meth A cells. Cells ( 5 x 10') were incubated with various concentrations of HGF for 5 days and cell growth was assayed using MTT. Growth inhibition (%) was calculated as: 100 x [ ( A s 7 0of sample in the absence of HGF)-(A5," of sample treated with HGF)]/(A570 of sample in the absence of HGF). Each point represents the mean +_ SD (n = 3).

Immunoprecipitationof '"I-HGF cross-linked complex with an anti-met antiserum

Cells ( 5 x lo6) were cross-linked with lZ5I-HGFas described above and lysed with 0.5ml lysis buffer. The 100000 x g supernatant of the lysate was immunoprecipitated with 5 p1 anti-met antiserum coupled to 20 pl protein-ASepharose and separated by SDSjPAGE on 6% polyacrylamide gel under nonreducing conditions. lz5I-HGF crosslinked complex was detected by autoradiography. Immunoblotting of the c-met protein with an anti-phosphotyrosine antibody and an anti-rnet antiserum

Cells (4 x 10') were incubated with or without 5 nM HGF in binding medium at 37 'C for 15 min, washed, and lysed with 1 ml lysis buffer containing 30 mM sodium pyrophosphate, 50mM NaF, and 0.1 mM sodium orthovanadate. The 100000 x g supernatant of the lysate was immunoprecipitated with 50 p1 of an anti-met antiserum coupled to 25 p1 proteinA-Sepharose and separated by SDSjPAGE on 6% polyacrylamide gel. After SDS/PAGE, proteins were transferred to a Trans-Blot transfer membrane, according to the procedure of Towbin et al. [28] with a slight modification. The membrane was blocked with 5% bovine serum albumin in 10 mM Tris/HCl pH 7.4 containing 150 mM NaCI, incubated with 1 pg/ml anti-phosphotyrosine antibody or 1 :100 dilution of the anti-met antiserum in bovine serum albumin solution, followed by incubation with 1 : 1000 dilution of rabbit anti(mouse Ig) or porcine anti-(rabbit Ig) antibody-peroxidase conjugate, respectively. The color was developed by TM Blue. RESULTS Growth inhibition of Meth A cells by recombinant human HGF

Among several cell lines examined, we found that Meth A cells had relatively large number of HGF binding sites. Shima et al. reported that purified fibroblast-derived tumor cytotoxic factor, which is identical to HGF, inhibited the growth of Meth A cells [9]. We thus examined whether the recombinant human HGF used in this study had the same activity on Meth A cells. The viable cell numbers increased about 150-fold in 5 days of culture in the absence of HGF. As shown in Fig. 1, this cell growth was inhibited dose-dependently by incubation

860

b

Fig. 2. Autoradiography of lZ5I-HGF analyzed by SDS/PAGE. Radioiodinated HGF was analyzed by SDSjPAGE (10% polyacrylamide gel) under nonreducing (lane 1) and reducing (lane 2) conditions. Molecular mass standards were phosphorylase h (94 kDa), bovine serum albumin (67 kDa), ovalbumin (43 kDa) and carbonic anhydrase (30 kDa).

with HGF. Maximal inhibition (about 80% inhibition) was observed at about 100 pM HGF. Thus, recombinant human HGF had an inhibitory effect on growth of Meth A cells. Purity of radiolabeled HGF H G F was radioiodinated by the chloramine T method. Fig. 2 shows the SDS/PAGE analysis of '251-labeled HGF. The preparation migrated as a broad band with a molecular mass of 80 - 90 kDa under nonreducing conditions (lane 1). This molecular mass corresponds to that of unlabeled recombinant human H G F [ 2 3 ] .Under reducing conditions, the band separated into three bands of molecular mass 65 kDa, corresponding to the heavy chain, and 35 kDa and 32 kDa, corresponding to the light chain (lane 2). The apparent heterogeneity in the light chain, which may be due to differences in glycosylation, was also observed in unlabeled recombinant human H G F [23]. The iodinated H G F retained more than 50% of the biological activity of the native protein as assessed by growth inhibition of Meth A cells.

125I-HGF

0

HGF bound to receptors on Meth A cells was cross-linked with BS3. I2'1-HGF (50 pM) was incubated with Meth A cells for 2 h at O"C, washed with NaCI/P, and cross-linked to its receptors. The resulting complex was analyzed by SDS/PAGE

3

4

Bound (pmolel1O'cells)

Fig. 3. Binding of 1251-HGFto Meth A cells. (A) Equilibrium binding analysis. Cells ( 5 x lo4)wereincubated with '251-HGFfor 2 h a t 0 ,C. ( 0 )Specific binding (total binding minus nonspecific binding) and (0)nonspecific binding (means of duplicate determinations) at (a) low and (b) high concentrations of '251-HGF.(8) Scatchard representations of specific binding replotted from A. Data was analyzed by a computer program. ( 0 )Experimental values; (-) a computerfitted curve; the latter was resolved by computer analysis into two separate components (- - - -).

Origin212-

1 2 3 4 5 6 7 7

Q*

250kDa 230kDa

170-

Characterization of binding of *2s1-HGFto Meth A cells

Aftinity cross-linking of lZ5I-HGFto Meth A cells

2

1

116-

I2'1-HGF binding to Meth A cells at 0°C reached equilibrium after 2 h (data not shown). Thus, further binding experiments were performed at 0 'C for 2 h. Incubation of increasing concentrations of lz51-HGF with Meth A cells showed that the specific binding consisted of a saturable component (Fig. 3 A). The level of nonspecific binding was linearly dependent on the concentration of free I2'I-HGF and was about 15% of the total binding. Scatchard plot analysis of the binding data revealed that there were two classes of binding sites on Meth A cells (Fig. 3B). The Kd values of the high-affinity and low-affinity sites were 17 pM and 6.7 nM, respectively, and the average numbers were 6600 and 2600000 per cell, respectively.

Added (nM)

76-

Dye-

m

+125~

- - --

HGF

Fig. 4. 1251-HGFcross-linking analysis in Meth A cells. Cells ( 5 x lo6) were incubated with 50 pM "*I-HGF, then chemically cross-linked and analyzed by SDS/PAGE on 6% polyacrylamide gel under nonreducing conditions (lane 2). Cells were not treated with BS3 (lanc 1). For competition analysis, 25 pM (lane 3), 50 pM (lane 4), 100 pM (lane S ) , 200 pM (lane 6), and 400 pM (lane 7) unlabeled HGF was added together with IZ51-HGF. IZ51-HGF binding complexes were detected by autoradiography. Molecular mass standards were myosin (212 kDa), a2-macroglobulin (170 kDa), 8galactosidase (1 16 kDd) and transferrin (76 kDa).

under nonreducing conditions. As shown in Fig. 4, a major band with a molecular mass of 250 kDa and a minor band of 230 kDa were detected, in addition to that of free HGF (lane 2). In the absence of cross-linking reagent, the only cellassociated radiolabeled band detected was that of free HGF

86 1

Fig. 5. lmmunoprecipitationof cell-surface HGF-binding proteins using HGF and an anti-HGF polyclonal antibody. The surface of 2.5 x 10' cells was '251-labeled and the cell lysate was incubated with (A, lanes 1, 3) or without (A, lanes 2,4) HGF, then immunoprecipitated with an antiHGF polyclonal antibody. lmmunoprecipitates were analyzed by SUSjPAGE under nonreducing (A, lanes 1, 2) or reducing (A, lanes 3. 4) conditions on 6% or 10% polyacrylamide gel, respectively. (B) The sample from lane 1 ofA was separated by nonreducing (6% polyacrylamide)/ reducing (10% polyacrylamide) two-dimensional SDS/PAGE. HGF-binding proteins were detected by autoradiography.

mass and heterodimeric subunit structure. The sequence of the c-met protein has features characteristic of the tyrosine kinase family of growth factor receptors [30] and the protein product is autophosphorylated in vitro on tyrosine residue [31]. We therefore examined whether the 160-kDa surface protein is autophosphorylated in vitro. HGF was incubated with Meth A cells and the H G F and surface binding protein complex was recovered from the cell membrane by solubilization and immunoprecipitated with an anti-HGF polyclonal Detection of HGF-binding surface proteins antibody. The immunoprecipitates were incubated with To detect the directly labeled HGF-binding surface [y-32P]ATP and analyzed by SDS/PAGE. As shown in proteins, cells were surface-labeled with Na[lZ5I],then lysed. Fig. 6A, a band with a molecular mass of 160 kDa was detectThe resulting detergent lysatcs were incubated with HGF. ed under nonreducing conditions (lane 1). Under reducing The HGF and cell-surface binding protein complexes were conditions, the band shifted to the position of 140 kDa (lane immunoprecipitated with an anti-HGF polyclonal antibody. 3). This molecular mass was identical to that of the large The specifically bound proteins were eluted and analyzed by subunit of the 160-kDa protein detected by cell surface SDSIPAGE. As shown in Fig. 5A, a major and a minor band labeling. When the 160-kDa protein was cross-linked with with molecular masses of 160 kDa and 130 kDa, respectively, HGF, the band of the 160-kDa protein had almost disapwere detected under nonreducing conditions (lane 1). These peared and a band with a larger molecular mass, 250 kDa, bands represent the specifically bound cell-surface proteins was detected (Fig. 6B, lane 1). This result indicates that the because they were not detected from immunoprecipitates with- 360-kDa protein was directly associated with H G F to form out HGF (lane 2). The molecular masses of these specifically the 250-kDa complex. The apparent molecular mass of the bound proteins isolated from the surface-iodinated cells corre- band exactly corresponded to that of the major product crosssponded to the net molecular masses of those identified in linked with '251-labeledHGF. We further analyzed the amino cross-linking studies (Fig. 4). Under reducing conditions, the acid which was phosphorylated in the 160-kDa protein. The two bands separated into three with molecular masses of band corresponding to the 140-kDa subunit of the 160-kDa 140 kDa, 100 kDa and 45 kDa (lane 3). To relate the three protein in lane 3 of Fig. 6A was excised from the gel, acidbands to those detected under nonreducing conditions, the hydrolyzed, and analyzed by high-voltage paper electrophoproteins were analyzed by nonreducing/reducing two-dimen- resis. As shown in Fig. 6C, the predominant radioactivity sional separation. The 160-kDa band separated into 140-kDa corresponded to phosphotyrosine, not to phosphoserine or and 45-kDa bands, while the 130-kDa band separated into phosphothreonine. These results further suggested that the 100-kDa and 45-kDa bands (Fig. 5B). These results suggest 160-kDa protein was the c-met protein. that the 160-kDa and 130-kDa proteins are structurally related. lmmunoprecipitation of 1251-HGFcross-linked complex

(lane 3 ) . The binding to form these complexes was saturated at about 50 pM Iz5I-HGF (data not shown) and was dosedependently competed in the presence of unlabeled H G F (lanes 3-7). About 50% competition was observed with 25 pM unlabeled HGF. These indicate that the binding was specific for HGF and the binding proteins were involved in the high-affinity binding.

with an anti-met antiserum

In vitro phosphorylation of HGF-binding surface protein During the course of this experiment, Bottaro et al. reported that the c-met protein is an HGF receptor [21]. The c-met protein is a 190-kDa disulfide-linked heterodimeric protein composed of a 50-kDa tl chain and a 145-kDa p chain [29]. The 360-kDa surface protein which we identified on Meth A cells could be the c-met protein because of its similar molecular

To confirm the identity of the 160-kDa HGF-binding protein as the c-met protein, we immunoprecipitated 1251-HGF cross-linked complex with a polyclonal antiserum specific to the C-terminal 21 amino acids of the mouse c-mef protein. As shown in Fig. 7, the cross-linked 250-kDa complex was immunoprecipitated by the antibody. Since the 250-kDa complex was not immunoprecipitated by a pre-immune serum

862

Fig. 6. In vitro phosphorylation and phosphorylated amino acid analysis of HGF-binding protein. Cells (1 x 10') were incubated with (A, lanes 1 , 3; B, lane 1) or without (A, lanes 2,4; B, lane 2) 1 nM H G F and treated with (B) or without (A) the cross-linker, BS3. The cells were lysed

and immunoprecipitatcd with an anti-HGF polyclonal antibody, then incubated with [y-3zP]ATP.The samples were separated by SDSjPAGE on 6% polyacrylamide gels under nonreducing (A, lanes 1, 2; B, lanes 1, 2) or reducing (A, lanes 3,4) conditions. The 140-kDa band in lane 3 of A was excised and eluted from the gel, acid-hydrolyzed, and electrophoresed on a Whatman 3MM paper (C). Phosphorylated amino acid standards are phosphoserine (p-Ser), phosphothreonine (p-Thr), and phosphotyrosine (p-Tyr). 32P-labeled phosphoprotein and amino acids were detected by autoradiography.

Fig. 7. Immunoprecipitationof '"I-HGF cross-linked complex with an anti-met antiserum. Cells ( 5 x lo6) were incubated with 200 pM lZ5IHGF. chemically cross-linked, and immunoprecipitated with an antimet antiserum. Immunoprecipitates were analysed by SDSjPAGE under nonreducing conditions on 6% polyacrylamide gel. Total lysate (lane 1) was immunoprccipitated with an anti-met antiserum (lane 2) or a pre-immune serum (lane 3). The immunizing peptide (1 mM) was added to the immunoprecipitation (lanc 4).

Fig. 8. Immunoblotting of the c-met protein with an anti-phosphotyrosine antibody and an anti-met antiserum. Cells (4 x 10') were treated (A, lanes 2, 4; B, lane 2) or untreated (A, lanes 1, 3; B, lane 1) with HGF, lysed, and immunoprecipitated with an anti-met antiserum. Immunoprecipitates were separated by SDS/PAGE (6% polyacrylamide gel) under nonreducing (A, lanes I , 2) or reducing (A, lanes 3,4; B, lanes 1,2) conditions and immunoblotted with an antiphosphotyrosine antibody (A) or an anti-met antiserum (B).

(lane 3) or in the presence of excess (1 mM) immunizing peptide (lane 4), it was specific to the peptide of the c-met protein. From the results, it can be concluded that the 160-kDa HGF-binding cell surface protein of Meth A cells is the c-met protein.

protein on Meth A cells was markedly stimulated by HGF treatment (Fig. 8A), whereas the amount of the c-met protein was unchanged by H G F treatment (Fig. 8B).

DISCUSSION HGF-stimulated tyrosine phosphorylation of the c-met protein

Bottaro et al. [21] have shown that H G F stimulates the tyrosine phosphorylation of the c-met protein on a mammary epithelial cell line, whose growth is promoted by HGF. Therefore we examined whether HGF stimulates the tyrosine phosphorylation of the c-met protein on Meth A cells, whose growth is inhibited by HGF. Meth A cells were incubated with or without HGF, then lysed and immunoprecipitated with an anti-met antiserum. The immunoprecipitate was immunoblotted with an anti-phosphotyrosine antibody or an anti-nzet antiserum. Tyrosine phosphorylation of the c-met

The results presented here demonstrate that human HGF can be covalently cross-linked to two surface proteins on Meth A cells with net molecular masses of about 160 kDa and 130 kDa. Independent confirmation of the existence of the two distinct HGF-binding surface proteins was obtained using a strategy that did not require chemical cross-linking. The HGF-binding surface proteins associated with HGF was immunoprecipitated with an anti-HGF polyclonal antibody. When the 160-kDa major binding protein was reduced, it separated into two subunits with molecular masses of 140 kDa and 45 kDa. This characteristic structure, together with the molecular masses, indicated that the 160-kDa protein was

863 possibly the c-met protein, which was recently identified as an HGF receptor in a human mammary epithelial cell line [21]. The 160-kDa protein was autophosphorylated in vitro on tyrosine residue by incubation with [p3'P]ATP and was immunoprecipitated with an anti-met antiserum. These results indicate that the 160-kDa HGF-binding surface protein on Meth A cells is the c-met protein. The met oncogene was originally isolated by transfection analysis in NIH 3T3 cells from a human osteogenic sarcoma cell line treated in vitro with the chemical carcinogen N-methyl-N-nitro-N-nitrosoguanidine [32]. The oncogene encoded a truncated tyrosine kinase activated by chromosomal rearrangement [33]. Sequence analysis of the cDNA for the proto-oncogene, e-met, revealed that the gene encoded a 1408-amino-acid protein with features characteristic of the tyrosine kinase family of growth factor receptors. A putative intracellular domain was highly similar to the src family of tyrosine kinases. However, a putative ligand-binding extracellular domain had no apparent sequence similarity to other growth factor receptors 1301. Thus, the ligand for this putative receptor had not been identified. Bottaro et al. and Naldini et al. showed that H G F induced the tyrosine phosphorylation of the c-met protein and HGF bound to the c-met protein [21, 22, 341. Our results, together with theirs, indicate that the c-met protein may function as a receptor for HGF or may be a component of the functional receptor for HGF. We also demonstrated in this study that H G F specifically bound Meth A cells with high and low affinity, the Kd values being 17 pM and 6.7 nM, respectively. The Kd value for high affinity is similar to that (24 pM) for rat hepatocytes reported by Higuchi and Nakamura [20], and the Kd value for low affinity is similar to that (3.5 nM) for rat hepatocytes reported by Zarnegar et al. [19]. The competition analysis followed by cross-linking revealed that the 3 60-kDa protein (c-met protein) was involved in the formation of the high-affinity binding. However, it is unknown at present whether the highaffinity HGF binding requires only the c-met protein or its association with one or more unknown molecules. We have detected the 130-kDa HGF-binding protein in addition to the 160-kDa protein. The competition analysis in cross-linking experiments suggested that the 130-kDa protein might be involved in the high-affinity binding. The 330-kDa protein is composed of 100-kDa and 45-kDa subunits. Since the 45-kDa subunit had the same molecular mass as that of small subunit of the 160-kDa protein, the 130-kDa protein is likely to be a smaller form of thc 160-kDa protein. If this is so, the large subunit of the 130-kDa protein lacks the intracellular kinase domain of the large subunit of the 260-kDa protein, because the 130-kDa protein was not autophosphorylated in vitro. Thus, the 130-kDa protein could be a product derived from the 160-kDa protein by proteolytic processing. However, we cannot exclude the possibility that proteolytic degradation may have occurred during the course of the preparation. Molccules which are involved in the low-affinity sites on Meth A cells are not characterized in this study. HGF has an affinity for heparin [l] and Naldini et al. reported that the low-affinity HGF binding to carcinoma cell lines was eluted by excess heparin [22]. Thus, the low-affinity binding sites which we detected on Meth A cells are possibly cell-surface-associated heparin-like molecules. HGF has multi-functional properties. These are growth stimulation of a variety of cell types including hepatocytes, growth inhibition of some tumor cell lines and scatter activity on epithelial cells. HGF binds the c-met protein and stimulates its tirosine phosphorylation on an epithelial cell line whose

growth is stimulated by H G F [21] and on a lung carcinoma cell line which is scattered by HGF [22]. We indicated in this study that HGF bound the c-met protein and stimulated its tyrosine phosphorylation on Meth A cells, whose growth is inhibited by HGF. We also detected a 160-kDa protein, which shifted to 140 kDa after reduction, by HGF-dependent immunoprecipitation and in vitro phosphorylation, in MDCK and A431 cells (target cells of scatter activity) as well as rat hepatocytes in primary culture (unpublished results). Thus, the c-met protein appears to be one of the common molecules which mediate the biological effects of HGF on its target cells. Further characterization of the receptor molecules as well as the molecules involved in intracellular signal cascades will be required to understand the diverse biological activity of HGF on different target cells. This work was supported in part by research grants from the Ministry of Education, Science and Culture of Japan, and Japan Private School Promotion Foundation.

REFERENCES 1. Gohda, E., Tsubouchi, H., Nakayama, H., Hirono, S., Sakiyama, O., Takahashi, K., Miyazaki, H., Hashimoto, S . & Daikuhara, Y.(1988) J . Clin. Invest. 81,414-419. 2. Zarnegar, R. & Michalopoulos, G. (1989) Cancer Res. 49,33243320. 3. Nakamura, T., Nawa, K., Ichihara, A,, Kaisc, N. & Nishino, T. (1987) FEBS Lett. 224, 311 -316. 4. Michalopoulos, G. (1990) FASEB J . 4 , 176- 187. 5 . Rubin, J. S., Chan, A. M.-L., Bottaro, D. P., Burgess, W. H.. Taylor, W. G., Cech, A. C., Hirschfield, D. W., Wong, J., Miki, T., Finchi, P. W. & Aaronson, S. A. (1991) Proc. Natl Acad. Sci. USA 88,415-419. 6. Kan, M., Zhang, G., Zarnegar, R., Michalopoulos, G.. Myokcn, Y . , Mckcehan, W. L. & Stevens, J. I. (1991) Biochem. Biophys. Res. Cornrnun. 174,331 -337. 7. Igawa, T., Kanda, S., Kanetake, H., Saito, Y., Ichihara, A., Tomita, Y . & Nakamura, T. (1991) Biochem. Biophys. Res. Comrnun. 174,831 - 838. 8. Matsumoto, K., Tajima, H. & Nakamura, T. (1991) Biochem. Biophys. Rex Cornmun. 176,45 - 51. 9. Shima, N., Itagaki, Y., Nagao, M., Yasuda, H., Morinaga, T. & Higashio, K. (1991) Cell Biol. Inl. Rep. 15,397-408. 10. Matsumoto, K., Hashimoto, K., Yoshikdwa, K. & Nakamura, T. (1991) Exp. Cell Rex 196, 114-120. 11. Higashio, K., Shima, N., Goto, M., Itagaki, Y., Nagao, M., Yasuda, H . & Morinaga, T. (1990) Biochern. Biophp. Res. Cornrnun. 170, 397 -404. 12. Weidner, K. M., Arakaki, N., Hartmann, G., Vandekcrckhove, J . , Weingart, S., Rieder, H., Fonatsch, C., Tsubouchi. H., Hishida, T., Daikuhara, Y. & Birchmeier, W. (1991) Proc. Nut/ Acad. Sci. USA 88,7001 -7005. 13. Miyazawa, K., Tsubouchi, H., Naka, D., Takahashi, K., Okigaki, M., Arakaki, N., Nakayama, H., Hirono, S., Sakiyama, O., Takahashi, K., Gohda, E., Daikuhara, Y. & Kitamura, N . (1989) Biochern. Biophys. Res. Cornrnun. 163. 967-973. 14. Nakamura, T., Nishizawa, T., Hagiya, M., Seki, T., Shimonishi, M., Sugimura, A., Tashiro, K. & Shimizu, S. (1989) Nature 342,440 -443. 15. Tashiro, K., Hagiya, M., Nishizawa, T., Seki, T.. Shimonishi, M., Shimizu, S. & Nakamura, T. (1990) Proc. Nut1 Acud. Sci. U S A 87,3200- 3204. 16. Okajima, A., Miyazawa, K. & Kitamura, N. (1990) Eur. J . Riochern. 193,375-381. 17. Miyazawa, K., Kitamura, A., Naka, D. & Kitamura, N. (1991) Eur. J . Biochem. 197, 15 - 22. 18. Miyazawa, K., Kitamura, A. & Kitamura, N. (1991) Biochernisrry 30,9170-9176.

864 19. Zarnegar, R., DeFrances, M. C., Oliver, L. & Michalopoulos, G. (1990) Biochem. Biophys. Res. Commun. 173, 3 179- 1185. 20. Higuchi, 0. & Nakamura, T. (1991) Biochem. Biophys. Res. Commun. 176, 599 - 607. 21. Bottaro, D. P., Rubin, J . S . , Faletto, D. L., Chan, A. M.-L., Kmiecik, T. E., Vande Woude, G. F. & Aaronson, S. A . (1991) Science 251, 802 - 804. 22. Naldini, L., Weidner, K . M., Vigna, E., Gaudino, G., Bardelli, A,, Ponzetto, C., Narsirnhan, R. P,, Hartmann, G., Zarnegar, R.. Michalopoulos, G. K., Birchmeier, W. & Comoglio, P. M. (1991) EMBO J . 10,2867-2878, 23. Strain. A. J., Ismail, T.. Tsubouchi, H., Arakaki, N., Hishida, T., Kitamura, N., Daikuhara, Y. & McMaster, P. (1991) J . Clin. Invest. 87, 1853 - 1857. 24. Mosmann, T. (1983) J. Immunol. Methods 65, 55 -63. 25. Hunter. W. M. & Greenwood, F. C. (1962) Nature 194, 495496. 26. Laemrnli, U. K . (1970) Nature 227, 680-685.

27. Sommer, A. & Traut, R. R. (1975) J . Mol. Bid. Y7,471-481. 28. Towbin, H., Staehelin, T. & Gordon, J. (1979) Proc. Natl Acud. Sci. U S A 76,4350 - 4354. 29. Giordano, S., Ponzetto, C., Di Renzo, M. F., Cooper, C. S. & Cornoglio, P. M. (1989) Nature 339, 155- 156. 30. Park, M., Dean, M., Kaul, K., Braun, M. J., Gonda, M. A. & Vande Woude, G. (1987) Proc. Natl Acad. Sci.USA 84,6379 6383. 31. Gonzatti-Haces, M., Seth, A,, Park, M., Copeland, T., Oroszlan, S. & Vande Woude, G. F. (1988) Proc. Natl Acad. Sci. USA 85,21-25. 32. Cooper, C. S., Park, M., Blair, D. G., Tainsky, M. A,, Huebner, K., Croce, C. M. & Vande Woude, G. F. (1984) Nuture 311, 29 - 33. 33. Park, M., Dean, M., Cooper, C. S., Schmidt, M., O’Brien, S. J., Blair, D. G. & Vande Woude, G. F. (1986) CeN45, 895-904. 34. Naldini, L., Vigna, E., Narsimhan, R., Gaudino, G., Zarnegar, R., Michalopoulos, G. & Comoglio, P. M. (1991) Oncogene 6, 501 - 504.

Characterization of hepatocyte-growth-factor receptors on Meth A cells.

Hepatocyte growth factor (HGF) is a heparin-binding polypeptide mitogen for a variety of cell types including hepatocytes. HGF also has cytotoxic acti...
916KB Sizes 0 Downloads 0 Views