Phosphorylation of serine 985 negatively regulates the hepatocyte growth factor receptor kinase.

The receptor for hepatocyte growth factor/scatter factor (HGF/SF) is an alpha beta tyrosine kinase of 190 kDa which mediates growth and motility in several cell types. We have previously shown that tyrosine autophosphorylation enhances the receptor kinase activity, while serine phosphorylation by protein kinase C or other Ca(2+)-dependent kinase(s) is inhibitory. We now identify Ser985 as the major phosphorylation site for the protein kinases responsible for such inhibition. Both phorbol esters or Ca2+ ionophore treatment induces phosphorylation of the same tryptic phosphopeptide corresponding to the sequence Leu983-Arg987 located in the juxta-membrane domain of the receptor beta chain. Purified protein kinase C phosphorylates in vitro a synthetic peptide (V14S) including Ser985. Trypsin digestion of the phosphorylated V14S generates a single phosphopeptide comigrating in reverse-phase high performance liquid chromatography with the tryptic peptide phosphorylated in vivo. Phorbol ester treatment of cultured cells inhibits the ligand-induced tyrosine autophosphorylation of the receptor. In vitro, Ser985 phosphorylation inhibits the receptor tyrosine kinase activity on exogenous substrates. Substitution of Ser985 by site-directed mutagenesis results in increased tyrosine phosphorylation of the receptor and abolishes down-modulation by protein kinase C. These data show that phosphorylation of Ser985 is a key mechanism for the negative regulation of HGF/SF receptor.

The receptor for hepatocyte growth factorlscatter factor (HGFBF) is an ap tyrosine kinase of 190 kDa which mediates growth and motility in several cell types. W e have previously shown that tyrosine autophosphorylation enhances the receptor kinase activity, while serine phosphorylation by protein kinase C or other Ca2+-dependent kinase(s) is inhibitory. W e now identify Seres6 as the major phosphorylation site for the protein kinases responsible for such inhibition. Both phorbol esters or Ca2+ ionophore treatment induces phosphorylation of the same tryptic phosphopeptide corresponding to the sequence L~U "~-A~~~' located in the juxtamembrane domain of the receptor p chain. Purified protein kinase C phosphorylates in vitro a synthetic peptide (V14S) including Seresa. Trypsin digestion of the phosphorylated V14S generates a single phosphopeptide comigrating in reverse-phase high performance liquid chromatography with the tryptic peptide phosphorylated in vivo. Phorbol ester treatment of cultured cells inhibits the ligand-induced tyrosine autophosphorylation of the receptor. In vitro, Seras6 phosphorylation inhibits the receptor tyrosine kinase activity on exogenous substrates. Substitution of Seres6 by site-directed mutagenesis results in increased tyrosine phosphorylation of the receptor and abolishes down-modulation by protein kinase C. These data show that phosphorylation of Sereea is a key mechanism for the negative regulation of HGF/SF receptor.
The receptor for hepatocyte growth factor (HGF),' a powerful mitogen for several epithelial cells, is a transmembrane tyrosine kinase encoded by the MET proto-oncogene (Bottaro et al., Naldini et al., 1991a;Giordano et al., 1993). The tyrosine kinase is also known as the receptor for "scatter factor" (SF), a molecule identical to HGF, that triggers cell motility and promotes invasiveness (Weidner et al., Naldini et al., 1991b). HGF/SF is secreted as a n inactive single chain precursor and processed to the active two chain form by extracellular proteolytic cleavage mediated by urokinase (Naldini et aE., 1992). The receptor protein is an heterodimer of 190 kDa (pl9OMET) composed of an The abbreviations used are: HGF, hepatocyte growth factor; SF, scatter factor; TPA, 12-0-tetradecanolyphorbol-13-acetate; MBP, myelin basic protein; PAGE, polyacrylamide gel electrophoresis; HPLC, high performance liquid chromatography. ship.
Fax: +39-11-650-9105, a-chain of 50 kDa, exposed at the cell surface, linked by disulfide bonds to a p-chain of 145 kDa (p145 p; Giordano et al., 1989aGiordano et al., , 1989b. The latter spans the plasma membrane and is endowed with tyrosine kinase activity (Tempest et al., 1986;Park et al., 1987;Giordano et al., 1989a). Two C-terminal truncated HGF/SF receptor forms have been identified: a 140-kDa heterodimer bound to the cell surface and a 130-kDa heterodimer released in the extracellular environment. (Prat et al., 1991b).
In humans, the HGF/SF receptor is widely expressed in cells of epithelial origin and in a variety of epithelial tumors (Prat et al., 1991a;Di Renzo et al., 1991. Recently, the receptor has been found in the microglia (Di Renzo et al., 1993) and on endothelial cells (Bussolino et al., 1992).
Both the mitogenic and the motogenic HGF/SF signals are mediated by the receptor tyrosine kinase activity which results in autophosphorylation at specific tyrosine sites recruiting intracellular signal transducers Bardelli et al., 1992). HGF/SF stimulation enhances the activity of PI 3-kinase (Ponzetto et al., 1993) of a RAS guanine nucleotide exchanger (Graziani et al., 1993) and of a tyrosine phosphatase (Villa-Moruzzi et al., 1993).
We have also shown that tyrosine autophosphorylation of the receptor enhances its own intrinsic kinase activity (Naldini et al., 1991c;Ferracini et al., 1991). Our previous work has demonstrated as well that the HGF/SF receptor can be phosphorylated on serine by two independent mechanisms involving activation of protein kinase C or activation of a protein kinase sensitive to the concentration of intracellular Ca2+ (Gandino et al., 1990(Gandino et al., , 1991.
In this work we have identified Ser985 as the major phosphate acceptor site for the phosphorylations triggered by both protein kinase C and the Ca2+-dependent kinase. We show that phosphorylation of this residue acts as a negative modulator of the tyrosine kinase activity of the receptor in vitro and in transfected cells. We suggest that phosphorylation of SergE5 is a mechanism for the negative regulation of HGF/SF receptor.
[Y-~~PIATP (specific activity 7000 Ci/ mmol), [32Plorthophosphate (10 mCi/ml), and lZ5I-protein A were obtained from Amersham. Nitrocellulose filters for Western blots were from Bio-Rad. The peptide V14S corresponding to amino acid residues 975-988 (VHTPHLDRLVSARS) of the MET sequence derived from cDNA cloning of the major transcript (Ponzetto et al., 1991: EMBY GenBank accession number X54559), was obtained from Neo System. Purified protein kinase C was kindly provided by Dr. Peter J. Parker (Imperial Cancer Research Fund, London), calciudcalmodulin-dependent protein kinase-I1 by Angus C. Nairn (Rockefeller University, New York). Anti-phosphotyrosine antibodies (anti-Qr(P)) were raised in rabbits as described (Comoglio et al., 1984). Rabbit and mouse monoclonal anti-Met antibodies, directed against the P-chain of pl9OMET receptor, were produced as described previously (Prat et al., 1991b). Human HGF/SF was purified to homogeneity from tissue culture medium con-1815 Serine Phosphorylation of HGFISF Receptor ditioned by the human fibroblast line MRC5 as described previously (Bussolino et al., 1992). GTL-16 cells are a clonal cell line derived from a poorly differentiated gastric carcinoma featuring a constitutively tyrosine-phosphorylated HGF/SF receptor due to overexpression of an amplified, structurally normal MET oncogene (Giordano et al., 1989a;Ponzetto et al., 1991). COS-7 cells were obtained from the American Type Culture Collection. The cells were cultured at 37 "C, under a humidified atmosphere of 95% air, 5% COz, in RPMI-1640 medium supplemented with 10% fetal calf serum (Seromed). P2P10rthophosphate Labeling and Immunoprecipitation-Subconfluent cultures of GTL-16 cells were incubated for 4 h at 37 "C with 1 mCi/ml of [J2P]orthophosphate in phosphate-free RPMI medium in the absence of serum. Cells were stimulated with 160 nm TPA, or dimethyl sulfoxide alone (Me2SO), for 30 min at 37 "C. The cultures were placed on ice, washed twice with cold phosphate buffer saline and lysed in RIPA buffer (20 rn Tris-HCI, pH 7.4, 150 rn NaCl, 2 m~ EDTA, 0.1% SDS, 1% Triton X-100, 1% deoxycholate, 100 p~ Na3V04, 50 rn NaF, 10 nm okadaic acid) and a mixture of protease inhibitors.
Cell lysates were clarified at 30,000 revolutiondmin for 30 min a t 4 "C, and immunoprecipitated for 2 h a t 4 "C with anti-Met antibodies coupled to Sepharose-protein A. Bound proteins were washed several times and eluted in boiling Laemmli buffer. Eluted proteins were subjected to 8% SDS-PAGE followed by autoradiography at -70 "C using intensifying screens.
HPLC Phosphopeptide Mapping and Phosphwmino Acid Analysis-For phosphopeptide mapping the 32P-labeled bands were excised from SDS-PAGE and further processed as described elsewhere . The phosphopeptides were analyzed on a reversephase Cz-CIR Superpack Pep-S column (Pharmacia LKB), resolved on a gradient of acetonitrile in buffer A of 0.46%/min (from 0 to 32% acetonitrile in 82 min) with a flow rate of l mumin. The eluted radioactivity was monitored by a Radiomatic A-I00 radioactive flow detector (Packard). The fractions corresponding to the largest radioactive peaks were pooled, lyophilized, and subjected to phosphoamino acid analysis. Phosphoamino acid analysis was performed as described by Cooper et (11. (1983).
In Vitro Phosphorylation of the Synthetic Peptide VI4S-The peptide V14S was phosphorylated in vitro by protein kinase C as follows: 50 pg of peptide were incubated a t 30 "C for 10 min in 40 pl of a buffer containing 50 mM HEPES, pH 7.5,0.5 rn EGTA, 12.5 IIIM MgClZ, 0.75 rn CaCIz, 50 pg of phosphatidylserine, 50 ng of "PA, together with 30 pCi of [Y-~*P]ATP and 3 units of purified kinase. Phosphorylation by calciudcalmodulin-dependent kinase-I1 was performed in a buffer containing 120 rn NaCI, 16 rn Tris-HC1 pH 8.00, 10 rn MgCIZ, 0.5 m~ CaClZ, 0.07 rn EGTA, 0.1 rn p-mercaptoethanol, 1 p~ calmodulin, 50 p~ [y-32PlATP, and 300 ng of purified kinase. The reaction was terminated by adding trichloroacetic acid a t a final concentration of 10%. After 1 hr on ice the reaction mixture was centrifuged at 15,000 revolutiondmin a t 4 "C for 15 min. The pellet was dissolved in Laemmli buffer and analyzed in an 8% SDS-PAGE followed by autoradiography. The supernatant was subjected to SeppakTM chromatography and analyzed in reverse-phase HPLC as described above. The peak was collected, lyophilized, resuspended in 50 rn NH,C03, pH 7.8, and incubated with 50 pg of ~-l-to~yl-amido-2-phenylethyl chloromethyl ketonetreated trypsin for 2 h a t 37 "C. The desiccated sample was then resuspended in buffer A and analyzed on HPLC. The phosphorylated peak was recovered from HPLC, hydrolyzed, and analyzed in thin layer chromatography using phosphorylated amino acids as standards.
Myelin Basic Protein Phosphorylation Assay-GTL-16 cells treated or not with TPA were washed twice with ice-cold phosphate-buffered saline and lysed in RIPA buffer in the presence of serine phosphatase inhibitors (see above) to prevent serine dephosphorylation. 'Qrosine phosphatase inhibitors were deliberately omitted to allow in vitro dephosphorylation of tyrosyl residues. Immunoprecipitation was performed as described above with anti-Met antibodies. Immunocomplexes were washed in KB buffer (25 rn Tris-HC1, pH 8.0, 10 m~ MgC12, and 0.1 IIIM dithiothreitol) in the presence of different concentrations of MBP or [y-"PIATP. After 10 min at 37 "C, the reaction was stopped by addition of equal volume of 2-fold concentrated Laemmli buffer (Laemmli, 1970). Samples were separated in 15% SDS-PAGE and autoradiographed. The relative amount of phosphate incorporated in MBP was estimated by measuring the optical density of the corresponding autoradiographic band with a laser densitometer (LKB 2202 Ultroscan). Alternatively, phosphorylation of MBP was directly quantified by liquid scintillation counting of the 32P-labeled protein bound to phosphocellulose filters as described previously . washed and incubated in the presence of HGF/SF (50 ng/ml) for 10 min a t 37 "C. Monolayers were washed with ice-cold phosphate-bufferd saline and lysed in RIPA buffer. Immunoprecipitation was performed as described above with anti-Met antibodies. Eluted proteins were subjected to 8% SDS-PAGE, transferred to nitrocellulose sheets, and probed with 10 pg/ml of purified anti-'Ilr(P) antibodies followed by 1251-labeled protein A as described elsewhere (Di . To quantify the amount of HGF/SF receptor, anti-Met antibodies were used to reprobe the blots. Filters were subjected to autoradiography as above.

HGFISF Stimulation and Detection
Site-directed Mutagemsis and Expression in COS-7 Cells-The cloning of the MET c-DNA has been reported previously (Ponzetto et al., 1991;Giordano et al., 1992). The nucleotide tract comprised between the EcoRI restriction site and the 3' end of the MET cDNA sequence was subcloned in the p S E L E C F -1 vector (Promega). Mutagenesis was performed using an in vitro oligonucleotide site-directed mutagenesis system (Promega), as specified by the manufacturer. Oligonucleotides were synthesized with an Applied Biosystem 391 apparatus. Bacterial colonies containing the mutated MET cDNA were identified by sequencing (T7 sequencing kit from Pharmacia). Full-size MET cDNA carrying the SePR5-AlagR5 mutation was reconstructed in the PMTB vector. The PMTB vector, containing the major late adenovirus promoter, was transiently transfected in COS-7 cells by the lipofection procedure (GIBCO-BRL). Protein experiments were performed 48 h after transfection.

RESULTS
Activation of Protein Kinase C with TPAZnduces Serine Phosphorylation in a Single Zkyptic Peptide of the HGFISF Receptor P-Chain-Due to amplification and overexpression of the MET oncogene in GTG16 cells the HGFISF receptor is constitutively phosphorylated on tyrosine even in the absence of the ligand (Giordano et al., 1989a;Ponzetto et al., 1991). After metabolic labeling with [32P]orthophosphate, the 145-kDa P-chain immunoprecipitated with anti-Met antibodies was heavily phosphorylated both in tyrosine and in serine. After TPA treatment, the overall amount of phosphate detected in the P-chain was unaffected (Fig. IA). Phosphoamino acid analysis showed that TPA treatment decreased the phosphotyrosine and increased the phosphoserine content (Fig. lB ).
32P-Labeled 145-kDa P-chains, immunoprecipitated from control or TPA-treated cells, were digested with trypsin and analyzed by reverse-phase HPLC (Fig. 2). In the 0-chain from control cells, the peptide eluting a t 35 min contained the phosphorylated Qr1235, which has been previously identified as the major autophosphorylation site of the HGFISF receptor (Fer- racini et al., 1991). The amount of 32P incorporated in this peptide was reduced in the P-chain immunoprecipitated from TPA-treated cells. On the other hand, a significant amount of radioactivity was incorporated in a peak, barely detectable in control cells, eluting at 20 min. Phosphoamino acid analysis of this peak showed that phosphoserine was the only labeled amino acid.

Activation of a Ca2+-dependent Kinase or Protein Kinase C Induces Serine Phosphorylation of the HGFISF Receptor P-Chains in the Same %ptic Peptidt+32P]orthophosphate-
labeled GTL-16 cells were treated for 5 min with the calcium ionophore A23187 (10 p~) .
As described previously, under these conditions a Ca2+-dependent protein kinase phosphorylates the HGF/SF receptor in a single major phosphopeptide (Gandino et al., 1991). After A23187 or TPA treatment, the receptors were immunoprecipitated with anti-Met antibodies, the p-chains were trypsin digested, and compared in HPLC. In both conditions an identical major tryptic phosphopeptide was eluted at approximately 20 min (Fig. 3). In both cases, serine was the only phosphorylated amino acid (not shown).

The Major Serine Phosphorylation Site of the HGFISF Receptor p-Chain Maps at
Residue 985-A number of peptides derived from the sequence of the cytoplasmic domain of the HGF/SF receptor p-chain were synthesized and incubated in vitro with purified protein kinase C in the presence of [y-32P]ATP. These included the sequence between Valg7'%3ePe9 (V14S). The peptide V14S was phosphorylated in vitro by protein kinase C (Fig. 4A) and, at lower efficiency, by calcium/ calmodulin-dependent kinase-I1 (data not shown). The phosphorylated peptide V14S was subjected to trypsin digestion and compared with the tryptic phosphopeptide generated from HGF/SF receptor P-chains phosphorylated in vivo after TPA treatment. Identity was proven by the same retention time in HPLC and by co-elution when mixed (Fig. 4, B and C ) . The only phosphorylated amino acid was serine, showing that protein kinase C and the Ca2+-activated kinase trigger the phosphorylation of serine 985.

Ligand-induced 2)rosine Autophosphorylation of the HGFISF
Receptor in Vivo-GTG16 cells, due to overexpression of an amplified MET gene, display a basal level of constitutive tyrosine phosphorylation of the HGF/SF receptor. As previously described, the level of tyrosine phosphorylation can be further increased upon binding of HGF/SF to the receptor (Naldini et al., 1991a). After treatment of GTL-16 cells with TPA for 30 min, the level of constitutive tyrosine phosphorylation of the receptor P-chain was reduced (Fig. 5A). Successive treatment with HGF/SF, a t concentrations effective in control cells, failed to induce autophosphorylation on tyrosine of the receptor. TPA treatment did not apparently change the amount of HGF/SF receptor protein, as detected in Western blots probed with anti-Met antibodies (Fig. 5B ).

Phosphorylation of SePm Inhibits the HGFISF Receptor 8-
rosine Kinase Activity in Vitro-The effect of Sep85 phosphorylation on the HGF/SF receptor kinase activity was analyzed in vitro by studying tyrosine phosphorylation of an exogenous substrate. Identical amounts of HGF/SF receptor, immunoprecipitated with anti-Met antibodies from GTL-16 cells stimulated or not with TPA, were analyzed. Immunoprecipitation was performed in a buffer supplemented with EDTA and Sed Thr phosphatase inhibitors, but lacking Na3V04 in order to allow dephosphorylation of tyrosine residues. The amount of protein precipitated and the content in phosphotyrosine of samples solubilized from control and treated cells was checked by Western blots probed with anti-Met and anti-l)r(P) antibodies (Fig. 6A). Under these conditions the difference in basal phosphorylation on tyrosine of the HGF/SF receptor samples from control and TPA-treated cells was abolished while the difference in SePs5 phosphorylation was maintained as judged from phosphoamino acid analysis (not shown). The kinase assay was performed by incubating equal amounts of the receptor kinase with 50 p~ MBP in the presence of increasing concentrations of [y3TlATP. Assays were also performed by varying the concentration of MBP a t fixed concentrations of ATP. After 10 min at 0 "C, the reaction was stopped and the phosphorylated MBP was quantified either in 15% SDS-PAGE (Fig. 6B) or by binding to phosphocellulose paper (data not shown). Serss5 phosphorylation reduced the kinase activity of the HGF/SF receptor toward the exogenous substrate. The reduction affected mostly the V , , (-50%) leaving unaffected the KM for both substrates. The KM for MBP was in the range of 3-15 p~; the KM for ATP was -50 p~.

Mutation of SepE6 Abolishes Negative Regulation of HGFI SF Receptor Kinase Activity by Protein Kinase C"Ib fur-
ther confirm the regulatory role of SePS5 on the HGF/SF receptor tyrosine kinase activity, we used in vitro site-directed mutagenesis to replace Ser985 with alanine. TheAlass5 HGF/SF receptor cDNA construct was cloned into PMT2 vector, as described under "Experimental Procedures" and transfected into COS-7 cells. Wild-type receptor cDNA was transfected as control. In transfected cells, due to over-expression (Bardelli et al., 1992), the recombinant wild-type HGF/SF receptor is tyrosine phosphorylated in the absence of the ligand. As in GTG16 cells, the level of autophosphorylation can be further increased in the presence of HGF/SF (not shown). In both cases, tyrosine phosphorylation was reduced in response to TPA treatment (Fig. 7). Tyrosine phosphorylation of the AlasE5 mutant was significantly higher than phosphorylation of the wild-type receptor, both in the presence or in the absence of the ligand. Moreover, negative regulation by TPA treatment was abolished (Fig. 7).

DISCUSSION
Hepatocyte growth factor, also known as scatter factor, exerts pleiotropic cellular responses mediating regulation of cell division, motility, and differentiation (for a review, see Gherardi and Stoker, 1991). Such critical functions are mediated by binding and activation of a cell surface tyrosine kinase receptor whose structure, biosynthesis, and biochemical properties have been elucidated in a series of previous reports from this and other laboratories (for a review, see Cooper, 1992;. The response to growth factors can be modulated both positively and negatively by covalent modifications of the corresponding receptors. It is widely accepted that phosphorylation plays an important role in regulation of signal transduction (Yarden and Ullrich, 1988;Nishizuka, 1988;Ullrich and Schlessinger, 1990). Positive regulation by tyrosine phosphorylation has been shown for the v-fps oncogene product (Weinmaster et al., 1986) and the insulin receptor (Rosen et al., 1983;Czech, 1984,1986;Klein et al., 1986). Furthermore, the kinase activity of pp60"-"" (CartWright et al., 1987;Kmiecik and Shalloway, 1987;Piwnica-Worms et al., 1987), of ppl2Ogagf p s (Meckling-Hansen et al., 1987) and of the insulin receptor (Ellis et al., 1986) was shown to be greatly diminished upon substitution of a critical tyrosine residue with a phenylalanine.
The HGF/SF receptor tyrosine kinase is activated by tyrosine autophosphorylation. "he raise in phosphorylation rate is due to an increase in the V , , of the enzymecatalyzed phosphotransfer reaction . The major phosphorylation site of the HGF/SF receptor has been mapped to Tyr1235 . Tyr1235 is located within the kinase domain in a segment homologous to the major autophosphorylation sites of other receptor and non-receptor kinases (Hanks et al., 1988). We recently demonstrated that this site is directly involved in the kinase activation of the receptor.2 Negative regulation by serine or threonine phosphorylation has been shown to operate in a number of membrane receptor kinases including the receptors for insulin (Jacobs et al., 1983;Takayama, et al., 19881, insulin-like growth factor-1 (Jacobs et al., 19831, and epidermal growth factor (Davis and Czech, 1984). Epidermal growth factor receptor is phosphorylated by protein kinase C a t a specific site, Thr654 . Phosphorylation of this residue inhibits the receptor tyrosine kinase activity (Downward et al., 1985;Friedman et al., 1984;Cochet et al., 1984Countaway et al., 1990) and, as a consequence, the ability to trigger the mitogenic response (Livneh et al., 1988). Other reports indicate that the epidermal growth factor receptor may also be negatively modulated upon threoninekerine phosphorylation mediated by a yet unidentified protein kinase activated by the increased concentration of intracellular Ca2+ (Fearn and King, 1985;Friedman et al., 1989;Verheijden et al., 1990).
Previous work has shown that serine phosphorylation induced by TPA treatment reduces tyrosine autophosphorylation of the HGF/SF receptor (Gandino et al., 1990). A similar effect was also observed after serine phosphorylation induced by a Ca2+-sensitive kinase, acting through a protein kinase C-independent mechanism (Gandino, et al., 1991). We now report that in both cases a single residue, Ser986, is phosphorylated. Substitution of this residue with alanine resulted in complete loss of negative regulation.
SergS5 is located within the juxtamembrane domain of the HGF/SF receptor, a position similar to that of the regulatory Thr654 of epidermal growth factor receptor. SergS5 is surrounded by a stretch of amino acids that fits the consensus sequence for protein kinase C phosphorylation. Interestingly, SergS5 and the 3 residues upstream generate a second motif that is the canonical consensus sequence for phosphorylation by calciudcalmodulin-dependent kinase I1 ( Fig. 8; Kennelly and Krebs, 1991). The synthetic peptide V14S, corresponding to the sequence Va1975-Ser988 of the HGF/SF receptor and including SerSs5, was phosphorylated in vitro by purified protein kinase C and by calciwdcalmodulin-dependent kinase-11, suggesting that these enzymes might directly phosphorylate the HGF/SF receptor @chain in uiuo. However, attempts to phosphorylate the immunoprecipitated receptor have been so far unsuccessful (data not shown).
Phosphorylation of SergE5 causes inhibition of the HGF/SF receptor kinase as shown by reduced phosphorylation of an exogenous substrate in vitro and inhibition of autophosphorylation in vivo. Kinase activation and subsequent tyrosine phosphorylation of growth factor receptors are critical to recruit and to activate cytoplasmic signal transducers which ultimately generates the biological response (for a review, see Cantley et al., 1991). In this context phosphorylation of SerSs5 is a negative regulatory event interfering with the signal transduction pathways triggered by HGF/SF.