Receptor-type Protein Tyrosine Phosphatase β (RPTP-β) Directly Dephosphorylates and Regulates Hepatocyte Growth Factor Receptor (HGFR/Met) Function*

Protein tyrosine phosphorylation is a ubiquitous, fundamental biochemical mechanism that regulates essential eukaryotic cellular functions. The level of tyrosine phosphorylation of specific proteins is finely tuned by the dynamic balance between protein tyrosine kinase and protein tyrosine phosphatase activities. Hepatocyte growth factor receptor (also known as Met), a receptor protein tyrosine kinase, is a major regulator of proliferation, migration, and survival for many epithelial cell types. We report here that receptor-type protein tyrosine phosphatase β (RPTP-β) specifically dephosphorylates Met and thereby regulates its function. Expression of RPTP-β, but not other RPTP family members or catalytically inactive forms of RPTP-β, reduces hepatocyte growth factor (HGF)-stimulated Met tyrosine phosphorylation in HEK293 cells. Expression of RPTP-β in primary human keratinocytes reduces both basal and HGF-induced Met phosphorylation at tyrosine 1356 and inhibits downstream MEK1/2 and Erk activation. Furthermore, shRNA-mediated knockdown of endogenous RPTP-β increases basal and HGF-stimulated Met phosphorylation at tyrosine 1356 in primary human keratinocytes. Purified RPTP-β intracellular domain preferentially dephosphorylates purified Met at tyrosine 1356 in vitro. In addition, the substrate-trapping mutant of RPTP-β specifically interacts with Met in intact cells. Expression of RPTP-β in human primary keratinocytes reduces HGF induction of VEGF expression, proliferation, and motility. Taken together, the above data indicate that RPTP-β is a key regulator of Met function.

invasion of cells through the extracellular matrix in vivo (7,8). HGF is mitogenic in many normal cell types, including epithelial cells, vascular endothelial cells, and melanocytes. HGF is also a morphogen that induces transition of epithelial cells into a mesenchymal morphology and formation of branched tubelike structures (9,10). In keratinocytes, HGF has been shown to promote motility and proliferation (11,12). Each of these biological effects exerted by HGF is triggered by stimulation of its cell surface receptor, Met (also known as HGF receptor), with concomitant activation of downstream effecter molecules (8,13,14).
Upon ligand binding, Met autophosphorylates several tyrosine residues in its carboxyl-terminal domain. Met function is controlled by its state of tyrosine phosphorylation (15)(16)(17). Among the tyrosines within the carboxyl terminus, phosphorylation of tyrosines 1234 and 1235, located within the catalytic domain, is required for tyrosine kinase activity (15)(16)(17). Phosphorylation of tyrosines 1349 and 1356, which are highly conserved among other members of the Met family, such as Sea and Ron, are necessary for most biological activities of Met (18 -20). Tyrosines 1349 and 1356, when phosphorylated, serve as multifunctional binding sites for Gab1, Grab2, PI3K, phospholipase C-␥, SHP2, and Cbl proto-oncogene. Phosphorylation of tyrosine 1356 is essential for transducing signals for cell motility and morphogenesis (19 -21).
In addition to its roles in many normal physiological processes, Met also plays important roles in human malignancy. Met was originally identified as the TPR-Met oncogene, which possesses ligand-independent tyrosine kinase activity. TPR-Met arises from chromosomal fusion of the translocated promoter region (TPR) in chromosome 1 with the Met carboxyl-terminal sequence on chromosome 7 (22,23). This rearrangement has been observed in patients with gastric carcinoma (24). A large body of evidence demonstrates that misregulation of the Met signaling pathway is involved in many types of human cancers. Inappropriate expression of HGF/Met autocrine signaling confers increased tumorigenesis and metastatic activity in vivo, and Met expression often correlates with a poor prognosis (8,25,26). The Met pathway also plays key roles in epithelial-mesenchymal transition, which is involved in tumor invasion (6).
Met is one member of a large family of protein tyrosine kinases (PTK). PTK-mediated tyrosine phosphorylation is balanced by the family of protein tyrosine phosphatases (PTPs) (27)(28)(29). Aberrant regulation of either PTKs or PTPs can lead to abnormal cellular behavior and diseases such as cancer and autoimmunity (30). There are 107 genes in the human genome that encode for PTPs (81 active enzymes), and 90 genes encode for PTKs (85 active enzymes). Similar levels of complexity suggest that the two families have a comparable substrate specificity (31). PTKs have been implicated in controlling the amplitude of a signaling response, whereas PTPs are thought to have important roles in controlling the rate and duration of the response (32,33).
Among the PTP superfamily, there are 38 classical, tyrosinespecific PTPs. These classical PTPs can be further subdivided into 21 receptor-type PTPs (RPTPs) and 17 non-transmembrane PTPs. RPTPs contain an intracellular region containing catalytic activity, a transmembrane region, and an extracellular region. All PTPs, regardless of their subtype, contain at least one catalytic domain with a highly conserved active site signature motif ((I/V)HCXAGXXR(S/T)G), where the cysteine residue is absolutely required for catalytic activity. The specificity of PTPs toward substrates can be achieved by their tissue-/cellspecific expression, subcellular localization/compartmentalization, posttranslational modification, and/or specific interaction between PTP active sites and target sequences (34,35).
We report here that Met tyrosine phosphorylation and function are regulated by receptor-type protein tyrosine phosphatase ␤ (RPTP-␤). RPTP-␤ directly dephosphorylates Met tyrosine 1356, the major binding site for multiple effecter molecules that drive downstream signaling pathways.
Adeno-X Expression Vector Construction and Adenovirus Production-pShuttle RPTP-␤ was used to generate the Adeno-X expression vector using the Adeno-X expression system (Clontech Laboratories, Inc.). HEK293 cells were used for adenovirus production (38).
Preparation of Membrane Extract and Membrane Met Kinase Assay-Human primary keratinocytes were washed twice with ice-cold hypotonic buffer (20 mM Tris-HCl, (pH7.6) with 10 mM NaCl) and scraped from the culture plates in hypotonic buffer supplemented with protease inhibitor mixture (10 g/ml aprotinin, 10 g/ml leupeptin, 10 g/ml pepstatin A, 1 mM phenylmethylsulfonyl fluoride). Cell suspensions were disrupted in a Dounce homogenizer. Unbroken cells were separated from membrane fraction by centrifuge at 500 ϫ g for 10 min. Supernatant was centrifuged at 20,000 ϫ g for 30 min. The membrane pellet was resuspended in hypotonic buffer with protease inhibitor mixture and homogenized by passing the suspension back and forth repeatedly through a 25-gauge needle. The membrane suspension was then supplemented with 100 M ATP, 0.2% ␤-mercaptoethanol, and 30 mM MgCl 2 . The kinase reaction was carried out by addition of phosphatase inhibitors and incubating at room temperature for 30 min. Laemmli sample buffer was added to stop the reactions. Tyrosine phosphorylation of Met was analyzed by Western blotting using phospho-Met (pY1349/1356) antibody.
Transient Transfection of HEK293 Cells-pShuttle RPTPs were transiently transfected into HEK293 cells using Superfec-tion according to the manufacturer's protocol (Qiagen, Chatsworth, CA). Cells were treated 24 h after transfection.
Lentivirus-mediated shRNAi Silencing of RPTP-␤ in Primary Human Keratinocytes-The MISSION TurboGFP shRNA control vector and shRNA constructs targeting RPTP-␤ (5Ј-CCGGCGGGTGTATCAGACTAATTATCTCGAGA-TAATTAGTCTGATACACCCGTTTTT-3Ј) were purchased from Sigma-Aldrich (St. Louis, MO). Lentivirus was produced in 293FT cells after transfection of the RPTP-␤ shRNA construct and helper plasmids using the Superfection method as described by the manufacturer (Qiagen). Two days after transfection, medium from 293FT cells was collected and used to infect human primary keratinocytes to knock down endogenous RPTP-␤.
Preparation of Whole Cell Lysates and Western Blot-Cells were washed twice with ice-cold PBS, scraped from the culture dishes in WCE buffer (25 mM HEPES (pH7.2), 75 mM NaCl, 2.5 mM MgCl 2 , 0.2 mM EDTA, 0.1% Triton X-100, 0.5 mM DTT, 20 mM ␤-glycerophosphate), supplemented with protease inhibitor mixture and 1 mM sodium orthovanadate, and transferred to microfuge tubes. Following 10 min of incubation at 4°C, cell homogenates were centrifuged at 14,000 ϫ g for 10 min, and supernatants were collected and used as whole cell extracts. Equal amounts of protein from whole cell extracts were resolved by 10% SDS-polyacrylamide gel (Invitrogen) and transferred to Immobilon-P filter paper (Millipore, Bedford, MA). Immunoreactive proteins were visualized by enhanced chemifluorescence according to the manufacturer's protocol (GE Healthcare). Quantification of chemifluorescence was performed using a STORM PhosphorImager (GE Healthcare).
GST-RPTP-␤-IC Fusion Protein Expression and Purification-GST-RPTP-␤-IC fusion protein was expressed in Escherichia coli strain BL21 and purified by GST affinity column as described previously (39). Purity was at least 90%, as judged by SDS-PAGE. The concentration of purified protein was determined using Bio-Rad protein assay reagent (Bio-Rad).
In Vitro Direct Dephosphorylation of Purified TPR-Met by RPTP-␤-Purified tyrosine phosphorylated TPR-Met was mixed with purified GST-RPTP-␤-IC fusion protein in PTP assay buffer (150 mM NaCl, 1% IGEPAL CA-630, 50 mM Tris (pH 8.0)), supplemented with 10 mM DTT, and incubated at 37°C for the indicated times. Dephosphorylation reactions were terminated by addition of Laemmli sample buffer, and the levels of TPR-Met tyrosine phosphorylation were analyzed by Western blotting probed with total and phospho-Met antibodies.
Real-time RT-PCR-Total cellular RNA was purified using the EZgene total RNA purification kit according to the manufacturer's protocol (Biomiga, San Diego, CA). Reverse transcription of total RNA was carried out using a Taqman reverse transcription kit (Applied Biosystems, Foster City, CA). Realtime PCR was performed on a 7300 sequence detector (Applied Biosystems) using the Taqman universal PCR master mix kit (Applied Biosystems). Primer/probe combinations were purchased from Applied Biosystems. Target gene mRNA levels were normalized to endogenous housekeeping gene 36B4 mRNA levels.

Inhibition of Membrane-associated Protein Tyrosine Phosphatase Activity Increases Met Tyrosine Phosphorylation-We
initially investigated regulation of Met by RPTP activity in the membrane fraction from primary human keratinocytes. Inhibition of membrane-associated PTP activity by addition of PTP inhibitors orthovanadate or pervanadate significantly elevated tyrosine phosphorylation of Met. Tyrosine phosphorylation of tyrosines 1349/1356 was elevated 4-to 6-fold (Fig. 1). Similar elevations of phosphorylation of tyrosines 1234/1235 were also observed (data not shown). These data suggest that keratinocytes express membrane-associated PTP activity that dephosphorylates Met.
Specific Dephosphorylation of Met by RPTP-␤ in HEK293 Cells-To identify membrane-associated RPTPs that dephosphorylate Met, we performed expression screening in HEK293 cells. On the basis of expression profiling of RPTPs in human keratinocytes (36), we chose to characterize six candidate RPTPs. These six RPTPs were separately expressed in HEK293 cells, and their ability to reduce the level of HGF-induced phosphorylation of tyrosine 1349/1356 in endogenous Met was determined. The expression of each candidate RPTP has been confirmed by Western blotting (Fig. 2A). Of the six RPTPs, only RPTP-␤ significantly decreased the level of Met phosphorylation ( Fig. 2A). Overexpression of two different catalytically inactive mutant forms of RPTP-␤ (D1870A and C1904S) or wild-type RPTPdid not decrease Met tyrosine phosphorylation, indicating that reduction of Met tyrosine 1349/1356 phosphorylation requires catalytic activity of RPTP-␤ (Fig. 2B).
HGF has been shown to modulate angiogenesis through induction of VEGF (40,41). This induction is mediated by activation of the Ras/MAP kinase pathway (42,43). Given the ability of RPTP-␤ to reduce HGF-induced Erk activation, we investigated the ability of RPTP-␤ to inhibit HGF-induced VEGF gene expression in primary human keratinocytes. Consistent with the above results, expression of RPTP-␤ abolishes HGFinduced VEGF mRNA expression (Fig. 4C).
Knockdown of Endogenous RPTP-␤ Increases Met Phosphorylation-We next investigated the effect of knockdown of endogenous RPTP-␤ on basal and HGF-induced tyrosine phosphorylation of Met in primary human keratinocytes. Lentivirus-mediated expression of shRNA reduced the level of RPTP-␤ protein by approximately 70% (Fig. 5A). This knockdown of endogenous RPTP-␤ elevated both basal and HGFstimulated phosphorylation of Met tyrosine 1356 (Fig. 5B) but has no effect on HGF-induced phosphorylation of Met tyrosine 1349 (Fig. 5C). Furthermore, knockdown of RPTP-b potentiates both basal and HGF-stimulated Erk activation (Fig. 5D).
RPTP-␤ Directly Dephosphorylates Met Tyrosine 1356-RPTP-␤ may act directly or indirectly to reduce phosphorylation of Met tyrosine 1349/1356 in intact cells. To distinguish between these possibilities, we performed in vitro phosphatase assays using the purified intracellular domain of RPTP-␤ and purified phosphorylated TPR-Met as substrates. In these assays, RPTP-␤ reduced Met phosphorylation in a dose-and time-dependent manner. As shown in Fig. 6, phosphorylation of Met tyrosine 1356 was reduced by 75% within 15 min by RPTP-␤. In contrast, RPTP-␤ had no significant effect on the  level of tyrosine 1349 phosphorylation (Fig. 6). These data indicate that RPTP-␤ directly dephosphorylates Met tyrosine 1356 in vitro and is consistent with our findings in intact cells (Fig. 3,  A and B).
RPTP-␤ Substrate-trapping Mutant Specifically Binds Met in Intact Cells-PTPs and their substrates normally do not form stable complexes. Mutation of a conserved active site aspartic acid (D1870A in RPTP-␤) to alanine prevents completion of phosphate ester hydrolysis and therefore traps PTP and the substrate in a stable complex (37). Substrate-trapping mutants  have been employed to identify physiological substrates for several PTPs (36,37,44). To further substantiate the role of RPTP-␤ in Met regulation in intact cells, wild-type, substratetrapping mutant, and active site cysteine mutant RPTP-␤ were separately coexpressed with His-tagged TRP-Met in HEK293 cells. Pull-down assays were performed with Ni-NTA beads to purify His-tagged TRP-Met from HEK293 lysates, and TRP-Met associated proteins were analyzed by Western blotting. As shown in Fig. 7, substrate-trapping mutant RPTP-␤ bound to TRP-Met. In contrast, neither wild-type nor active site cysteine mutant RPTP-␤ bound to Met. These data indicate that RPTP-␤ forms a catalytically active complex with Met in intact cells and that Met is a physiological substrate for RPTP-␤.
RPTP-␤ Regulates Met-mediated Proliferation and Migration in Human Primary Keratinocytes-HGF/Met has been shown to be a key mediator of wound healing through its ability to promote keratinocyte proliferation and migration (11,12,45,46). Therefore, we investigated the role of RPTP-␤ in the regulation of keratinocyte proliferation and migration. We found that blockade of Met tyrosine kinase activity by a specific inhibitor (47,48) suppressed proliferation of primary human keratinocytes (Fig. 8A). Expression of RPTP-␤ similarly inhibited cell proliferation (Fig. 8B). Furthermore, expression of RPTP-␤ abolished HGF-dependent migration of primary human keratinocytes in a scratch assay (Fig. 8C).
Taken together, the above data demonstrate that RPTP-␤ negatively regulates several important biological functions of HGF/Met via direct dephosphorylation of Met tyrosine 1356.

DISCUSSION
The present study demonstrates regulation of Met tyrosine phosphorylation and function by RPTP-␤ via specific dephosphorylation of Met tyrosine 1356. Tyrosine 1356 is the major   binding site for multiple Met effecter molecules that mediate downstream pathways, including the Ras/MAP kinase pathway. Overexpression of RPTP-␤ suppresses multiple Met functions, including proliferation and migration, and inhibits HGFinduced MAP kinase activation and gene expression. Using multiple techniques, including substrate trapping and direct in vitro phosphatase assay, we demonstrate that Met is a bona fide substrate for RPTP-␤.
RPTP-␤ is encoded by the ptprb gene, and is composed of an extracellular domain with cell adhesion molecule-like motif (multiple fibronectin type III-like domains), a transmembrane domain, and a single intracellular PTP domain (49,50). RPTP-␤ is localized to chromosome 12 of the human genome, in a region associated with tumors whose cells have lost contact inhibition (51). Because of historical reasons (e.g. cloned and named by different groups), RPTP-(encoded by the gene ptprz) is sometimes referred as RPTP-␤/ or even RPTP-␤ in the literature. While there are many published studies regarding RPTP-␤/, there are only a few published studies on true RPTP-␤. Two studies describe in vitro analysis of the RPTP-␤ intracellular domain (52,53), and a very recent study indicates that RPTP-␤ regulates angiopoietin-Tie2 signaling pathway in human endothelial cells. This latter study did not elucidate mechanism of action or identify any RPTP-␤ substrates (54). Interestingly, knockout of the mouse homologue of human RPTP-␤, VE-PTP, is embryonic lethal because of defective angiogenesis (55,56). This observation is consistent with the important role that Met plays in angiogenesis (57).
Most RPTPs have two intracellular PTP domains, referred to as membrane-proximal (D1) and membrane distal (D2) domains. Usually the membrane-proximal PTP (D1) domain is catalytically active, whereas the membrane-distal PTP (D2) domain possesses little, if any, enzymatic activity (58). The D2 domain is thought to play important roles in substrate recognition or protein-protein interactions (59). RPTP-␤ has only a D1 domain; it lacks a D2 domain. We found that purified RPTP-␤ maintains its catalytic specificity toward Met tyrosine 1356 in vitro, indicating that a D2 domain is not necessary to confer substrate recognition and specificity.
Biological function of Met is determined by phosphorylation of multiple C-terminal tyrosines. In addition to our finding that RPTP-␤ dephosphorylates Met, other RPTPs have been shown to influence Met tyrosine phosphorylation (44,60,61). DEP-1 has been shown to preferentially dephosphorylate tyrosines 1349 and 1365 (44). However, DEP-1 expression did not affect MET-dependent MAP kinase activation (44). PTP1B and TCPTP have been shown to dephosphorylate tyrosine 1234/ 1235 of Met (61). Using only the antisense method, Kulas et al. (60) have shown that knockdown of RPTP-LAR increases tyrosine phosphorylation of insulin receptor, insulin receptor substrate-1, EGFR, and Met. The mechanism for the observed increased phosphorylation was not determined. Our data demonstrate that RPTP-␤ directly and specifically dephosphorylates Met tyrosine 1356. This finding raises the possibility that fine-tuning of ligand-induced Met signaling specificity can be achieved by the levels and activities of subsets of RPTPs in different cell types and tissues.
We have previously described direct and specific dephosphorylation of EGFR by RPTP- (36). Therefore, the functions of two important RPTK pathways are specifically controlled by distinct RPTPs. Activation of EGFR or Met leads to both common and distinct cellular responses. There is also cross-talk between these two RTKs. For example, activation of EGFR leads to ligand-independent constitutive activation of Met (62). These overlapping functions of EGFR and Met allow lung cancer to develop resistance toward EGFR-targeting therapy by switching to Met-dependent pathways for proliferation and survival (63). On the other hand, in Met-amplified gastric cancer cells, selective blocking of Met abolishes the cross-talk activation of EGFR. However, EGFR and downstream signaling pathways (e.g. Ras/MAPK and PI3K/Akt) can still be activated by the EGFR ligand in a Met-independent manner (64). Our studies suggest that the availability and abundance of RPTPand RPTP-␤ will play important roles in regulation of EGFR and Met functions in human physiology and cancer.