Ectodomain Shedding of SHPS-1 and Its Role in Regulation of Cell Migration*

SHPS-1 is a transmembrane protein whose cytoplasmic region undergoes tyrosine phosphorylation and then binds the protein-tyrosine phosphatase SHP-2. Formation of the SHPS-1-SHP-2 complex is implicated in regulation of cell migration. In addition, SHPS-1 and its ligand CD47 constitute an intercellular recognition system that contributes to inhibition of cell migration by cell-cell contact. The ectodomain of SHPS-1 has now been shown to be shed from cells in a reaction likely mediated by a metalloproteinase. This process was promoted by activation of protein kinase C or of Ras, and the released ectodomain exhibited minimal CD47-binding activity. Metalloproteinases catalyzed the cleavage of a recombinant SHPS-1-Fc fusion protein in vitro, and the primary cleavage site was localized to the juxtamembrane region of SHPS-1. Forced expression of an SHPS-1 mutant resistant to ectodomain shedding impaired cell migration, cell spreading, and reorganization of the actin cytoskeleton. It also increased the tyrosine phosphorylation of paxillin and FAK triggered by cell adhesion. These results suggest that shedding of the ectodomain of SHPS-1 plays an important role in regulation of cell migration and spreading by this protein.

The extracellular region of various transmembrane proteins is cleaved by proteases, such as metalloproteinases, and released as a soluble protein fragment (1,2). This process, known as "ectodomain shedding," often influences the mode of action or biological activity of the affected protein. For instance, ectodomain shedding of heparin-binding epidermal growth factor (EGF) 1 -like growth factor (HB-EGF), a membrane-anchored EGF-related protein (3), results in its activation, in that the released fragment binds to EGF receptors and thereby stimulates cell proliferation (3,4). In contrast, the membraneanchored forms of c-kit ligand (5) and of ephrins, the latter of which are ligands of Eph receptor tyrosine kinases (6), are fully functional, whereas their soluble forms generated by ectodomain shedding exhibit little or no biological activity. Ectodomain shedding is thus an important regulator of the function of certain membrane proteins that contribute to cellcell communication.
SHPS-1 (SHP substrate-1) (7), also known as SIRP␣ (8), BIT (9), and p84 neural adhesion molecule (10), is a receptor-like transmembrane protein that is particularly abundant in neurons and macrophages (10,11), although other cell types, such as fibroblasts, also express this protein (7). The putative extracellular region of SHPS-1 comprises three immunoglobulin (Ig)-like domains with multiple N-linked glycosylation sites, whereas the cytoplasmic region of the protein contains four YXX(L/V/I) motifs, which are putative tyrosine phosphorylation sites and binding sites for the Src homology 2 (SH2) domains of the protein-tyrosine phosphatases SHP-2 and SHP-1 (7,8). Tyrosine phosphorylation of SHPS-1 is regulated by various growth factors, including insulin and EGF, as well as by cell adhesion to extracellular matrix proteins (12). SHP-2 participates in the positive regulation of cell migration (13)(14)(15)(16), and SHPS-1 recruits and activates SHP-2 at the cell membrane in response to growth factors or to integrin-mediated cell adhesion (9,17). Characterization of immortalized fibroblasts from mice that lack most of the cytoplasmic region of SHPS-1 revealed a marked impairment of cell migration associated with an increased formation of actin stress fibers and focal adhesions (18). These observations suggest that the tyrosine phosphorylation of SHPS-1 and the consequent association of SHPS-1 with SHP-2 promote cell migration through regulation of cytoskeletal reorganization.
CD47, also named IAP, is implicated as a ligand for SHPS-1 (11,19). This protein, which was originally identified in association with ␣ v ␤ 3 integrin (20), is also a member of the Ig superfamily, possessing an Ig-V-like extracellular domain, five putative membrane-spanning segments, and a short cytoplasmic tail (21). CD47 and SHPS-1 appear to constitute a cell-cell communication system (the CD47-SHPS-1 system) that plays an important role in a variety of cell functions. We have recently shown that the CD47-SHPS-1 system and SHP-2 contribute to the inhibition of cell migration by cell-cell contact (22). Neutrophil migration to sites of inflammation is markedly impaired in CD47 knockout mice (23) and monoclonal antibodies (mAbs) to CD47 inhibit neutrophil transmigration (24), suggesting that the CD47-SHPS-1 system might mediate bidirectional inhibitory regulation of cell migration. The binding of CD47 on red blood cells to SHPS-1 on macrophages also inhibits phagocytosis of the red blood cells by the macrophages (25). * This work was supported by a Grant-in-Aid for Scientific Research on Priority Areas Cancer, a grant-in-aid for Scientific Research (B), and a 21st Century COE Program grant from the Ministry of Education, Culture, Sports, Science, and Technology of Japan; a grant from the Naito Foundation; a grant from the Mochida Foundation; a grant from the Nakatomi Foundation; and a grant from the Life Science Foundation. 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.
Thus, SHPS-1 appears to play multiple roles in cellular activities in a manner dependent on or independent of its interaction with CD47. It has not been known, however, whether SHPS-1 undergoes ectodomain shedding, a process that might potentially regulate its function.
We now show that SHPS-1 indeed undergoes ectodomain shedding. Furthermore, we have characterized the molecular mechanism of this process as well as examined its physiological role in the regulation of cell migration.

EXPERIMENTAL PROCEDURES
Primary Antibodies and Reagents-A rat mAb (␣p84) to mouse SHPS-1 was purified from culture supernatants of hybridoma cells kindly provided by C. F. Lagenaur (University of Pittsburgh). Rabbit polyclonal antibodies (pAbs) to the cytoplasmic region of SHPS-1 were obtained from ProSci. A mouse mAb to SHP-2 was from Transduction Laboratories. A mouse mAb to paxillin and rabbit pAbs to focal adhesion kinase (FAK) (for immunoprecipitation) were from Upstate Biotechnology. Rabbit pAbs to FAK (for immunoblot analysis) and peroxidase-conjugated mAb PY20 to phosphotyrosine were from Santa Cruz Biotechnology. A mouse mAb to vinculin as well as 12-O-tetradecanoylphorbol 13-acetate (TPA), lysophosphatidic acid (LPA), and 1,10-phenanthroline (OPT) were from Sigma. IC-3 was from Biomol Research Laboratories. KB-R7785 was kindly provided by S. Higashiyama (Ehime University).
Determination of Ectodomain Shedding of SHPS-1 by Concanavalin A (ConA)-mediated Precipitation, Immunoprecipitation, and Immunoblot Analysis-The supernatants of cell cultures (ϳ4 ϫ 10 6 cells in a 60-mm dish) were collected and centrifuged at 21,000 ϫ g for 15 min at 4°C, and the resulting supernatants were then incubated for 2 h at 4°C with ConA-coupled agarose beads (Amersham Biosciences). Alternatively, the cell-free conditioned medium was incubated for 4 h at 4°C with the mAb ␣p84 bound to protein G-Sepharose beads (Amersham Biosciences). Both types of beads were then washed twice with 1 ml of WG buffer (50 mM Hepes-NaOH (pH 7.6), 150 mM NaCl, 0.1% Triton X-100), resuspended in SDS sample buffer, and subjected to SDS-PAGE followed by immunoblot analysis either with the ␣p84 mAb or with pAbs to SHPS-1 as described previously (22). Immune complexes were detected with an ECL detection kit (Amersham Biosciences). Lysates of cultured cells were also prepared by incubation on ice with 1 ml of lysis buffer (20 mM Tris-HCl (pH 7.6), 140 mM NaCl, 1 mM EDTA, 1% Nonidet P-40) containing 1 mM phenylmethylsulfonyl fluoride, aprotinin (10 g/ml), and 1 mM sodium vanadate. The lysates were centrifuged at 21,000 ϫ g for 15 min at 4°C, and the resulting supernatants were subjected to immunoblot analysis. Immunoprecipitation and immunoblot analysis of paxillin and FAK were also performed by similar procedures. For preparation of membrane and cytosolic fractions, CHO-Ras-SHPS-1-WT cells were lysed on ice in 1 ml of hypotonic buffer (20 mM Tris-HCl (pH 7.6), 1 mM EDTA) containing 1 mM phenylmethylsulfonyl fluoride, aprotinin (10 g/ml), and 1 mM sodium vanadate; the lysate was subjected to centrifugation at 436,000 ϫ g for 60 min at 4°C; and the resulting supernatant and pellet were designated the cytosolic and membrane fractions, respectively.
Preparation of SHPS-1-Fc and CD47-Fc Fusion Proteins-An SHPS-1-Fc fusion protein, which contained the extracellular region of mouse SHPS-1 (amino acids 1-371) fused to the Fc portion of human IgG, and a CD47-Fc fusion protein, which contained the extracellular region of mouse CD47 (amino acids 1-161) fused to the Fc portion of human IgG, were produced and purified as described previously (27).
Analysis for the Binding of Fc Fusion Proteins to CHO-Ras Cells-The binding of SHPS-1-Fc to CHO-Ras cells stably expressing CD47 was determined essentially as described previously (27). Briefly, confluent CHO-Ras cells stably expressing CD47 in 96-well plates were incubated for 30 min at 37°C with various concentrations of SHPS-1-Fc in the culture medium described above, after which the cells were washed with ice-cold phosphate-buffered saline (PBS) and incubated for 30 min at 4°C with horseradish peroxidase-conjugated goat pAbs to the Fc fragment of human IgG (Jackson ImmunoResearch). The cells were again washed with PBS, and Fc fusion protein binding was determined by measurement of peroxidase activity with o-phenylenediamine dihydrochloride (Sigma) as substrate. The absorbance of each well at 492 nm was monitored with a microplate reader. The binding of CD47-Fc to CHO-Ras cells stably expressing wild-type or a mutant SHPS-1 was also performed as described above.
Identification of the SHPS-1 Peptide Bond Cleaved by Matrix Metalloproteinases-Incubation of SHPS-1-Fc with MMP-1 or MMP-9 was performed according to the method described by Franzke et al. (28), with minor modifications. The purified fusion protein (15 g) was incubated for 24 h at 37°C with 5 g of MMP-9 or MMP-1 (Calbiochem) in 120 l of a solution containing 50 mM Tris-HCl (pH 7.4), 0.2 M NaCl, 5 mM CaCl 2 , 0.1 mM ZnCl 2 , and 2 mM p-aminophenyl mercuric acetate, after which the reaction mixture was evaporated. The samples were then subjected to SDS-PAGE, the separated proteins were transferred to a polyvinylidene difluoride membrane, and the membrane was stained with Coomassie Brilliant Blue R-250. Stained bands corresponding to proteins of 35 and 30 kDa were excised, and their NH 2terminal amino acid sequences were determined with a gas-phase amino acid sequence analyzer (Procise 492; Applied Biosystems) as described previously (29).
Generation of CHO-Ras and CHO Cells Expressing a Mutant SHPS-1-To generate the plasmid for expression of the cleavage-resistant SHPS-1 mutant (SHPS-1-FLAG-JM), we performed the polymerase chain reaction with an expression plasmid for wild-type mouse SHPS-1 (full-length cDNA for mouse SHPS-1 (30) subcloned into the EcoRI and NotI sites of pTracer-CMV (Invitrogen)) as the template and the primers 5Ј-CCGATATCGACTACAAGGACGACGATGACAAGACC-CACAACTGGAATGTCTTC-3Ј (sense) and 5Ј-CCGATATCCCCTTGAT-CACTCGAGTG-3Ј (antisense). The sense primer encodes the artificial amino acid sequence DIDYKDDDDK, the last eight residues of which correspond to the FLAG epitope tag. The amplification product thus encoded a mutant SHPS-1 protein (SHPS-1-FLAG-JM) with this 10amino acid sequence instead of the juxtamembrane sequence SMQTF-PGNNA 368 (Fig. 5A); it was digested with EcoRV and then self-ligated, and its entire nucleotide sequence was verified by DNA sequence analysis with an ABI PRISM310 Genetic Analyzer (Applied Biosystems). CHO-Ras cells were then transfected with the plasmid containing the mutant SHPS-1 cDNA with the use of LipofectAMINE 2000 (Invitrogen). The cells were cultured in ␣MEM supplemented with 2 mM Lglutamine, 10 mM Hepes-NaOH (pH 7.6), 10% FBS, and Zeocin (200 g/ml) (Invitrogen), and colonies were isolated 14 -21 days after transfection. CHO cells were also transfected with the plasmid containing the mutant SHPS-1 cDNA and selected as described above. The cells were cultured in F-12 medium (Sigma) supplemented with 10% FBS and Zeocin (200 g/ml). Colonies were isolated 14 -21 days after transfection, and several cell lines expressing the mutant SHPS-1 protein were identified by immunoblot analysis of cell lysates with the mAb ␣p84.
Cell Migration Assay-Cell migration was assayed with a Transwell apparatus (Corning) as described previously (22). In brief, CHO, CHO-Ras, and derived cell lines were detached from culture dishes and resuspended in ␣MEM containing 10% FBS. A portion of the cell suspension (1 ϫ 10 5 cells in 100 l) was then transferred to a polycarbonate filter (pore size, 8 m; Corning) in the upper compartment of a Transwell apparatus, and 600 l of fresh culture medium was placed in the lower compartment. The number of cells that had migrated into the lower compartment after incubation for 16 h at 37°C was then counted in triplicate with a hemocytometer and was expressed as a percentage of the total number of cells added to the upper compartment.
Immunofluorescence Analysis-CHO-Ras and derived cell lines were plated on cover glasses 12-24 h before analysis. They were fixed for 20 min either with 4% paraformaldehyde in PBS for vinculin staining or with 4% paraformaldehyde and 0.1% glutaraldehyde in PBS for actin and paxillin staining. The cells were then permeabilized for 1 h at room temperature in buffer G (PBS containing 5% goat serum and 0.1% Triton X-100) before incubation for 1 h at room temperature with a mAb to vinculin (1/400 dilution), a mAb to paxillin (5 g/ml), or rhodamine-conjugated phalloidin (2 units/ml) (Molecular Probes) diluted in buffer G. They were then washed three times with PBS. For staining of vinculin or paxillin, the cells were incubated for 30 min with Alexa488conjugated sheep pAbs to mouse IgG (Molecular Probes) and then washed another three times with PBS. All cells were examined with a confocal laser-scanning microscope (LSM 5 Pascal, Carl Zeiss).
Cell Spreading and Adhesion Assay-Detached cells were suspended in fresh culture medium and transferred at a density of 1 ϫ 10 5 cells/ml to 35-mm dishes coated with fibronectin (Asahi Technoglass). After incubation for 1 or 4 h, the cells were examined with a light microscope equipped with phase-contrast optics (Leica DM IRBE), and digital images of random fields were captured with a charge-coupled device camera (Penguin 600CL, Pixera).

RESULTS
Ectodomain of SHPS-1 in Conditioned Medium of SHPS-1expressing Cells-We first examined whether the ectodomain of SHPS-1 was shed from CHO-Ras-SHPS-1-WT cells, CHO cells that were transformed as a result of expression of an active form of H-Ras and that also express mouse SHPS-1. Given that SHPS-1 is extensively glycosylated and therefore binds to ConA (7), we incubated the conditioned medium obtained by overnight culture of these cells with ConA-coupled agarose beads and then examined bead-bound proteins by immunoblot analysis with a mAb (␣p84) that specifically recognizes the extracellular region of mouse SHPS-1 (31). An immunoreactive protein of ϳ100 kDa was detected in the conditioned medium from CHO-Ras-SHPS-1-WT cells (Fig. 1A) but not in that from parental CHO-Ras cells (data not shown). The same protein was also observed in immunoprecipitates prepared with the ␣p84 mAb from the medium conditioned by CHO-Ras-SHPS-1-WT cells (Fig. 1A). In contrast, immunoblot analysis with rabbit pAbs that react with the cytoplasmic region of SHPS-1 failed to detect the ϳ100-kDa protein in precipitates prepared from this conditioned medium with either ConA or the ␣p84 mAb (Fig. 1A). Immunoblot analysis with these pAbs to SHPS-1 did detect an ϳ20-kDa protein in lysates of CHO-Ras-SHPS-1-WT cells but not in those of parental CHO-Ras cells (Fig. 1B). This ϳ20-kDa protein appeared to be localized predominantly in a membrane fraction, rather than a cytosolic fraction, of CHO-Ras-SHPS-1-WT cells (Fig. 1B). The molecular size of intact SHPS-1 in lysates of CHO-Ras-SHPS-1-WT cells was ϳ120 kDa (Fig. 1A), whereas that of the transmembrane plus cytoplasmic regions of SHPS-1 was estimated to be ϳ20 kDa on the basis of the predicted amino acid sequence of SHPS-1 (30). Together, these results thus suggested that the ϳ100-kDa protein detected in the conditioned medium of CHO-Ras-SHPS-1-WT cells was the extracellular domain of SHPS-1, whereas the ϳ20-kDa protein present in the membrane fraction of these cells was the transmembrane-cytoplasmic region of SHPS-1 remaining after cleavage (see also Fig. 5B).
Endogenous SHPS-1 was detected in lysates prepared from primary cultured mouse hippocampal neurons, B16F10 mouse melanoma cells, or RAW264.7 mouse macrophages (Fig. 1C). We also detected ϳ80to 100-kDa proteins that were reactive with the ␣p84 mAb (Fig. 1C), but not with the pAbs to SHPS-1 (data not shown), in the conditioned media obtained from these cells. The molecular sizes of the SHPS-1 fragments in these conditioned media were also ϳ20 kDa smaller than that of the intact SHPS-1 protein in the corresponding cell lysates. These results indicated that the ectodomain of SHPS-1 is shed into the medium of cultured cells that express endogenous SHPS-1 as well as into that of CHO-Ras-SHPS-1-WT cells. We also observed that the ectodomain of human SHPS-1 was present in the culture medium of CHO-Ras cells expressing the intact human protein (data not shown).
Regulation of Ectodomain Shedding of SHPS-1-We next examined whether the shedding of the ectodomain of SHPS-1 occurs in response to extracellular or intracellular stimulation.
Minimal ectodomain shedding of SHPS-1 was apparent after incubation for 4 h in serum-free medium of CHO cells that stably express mouse SHPS-1 (CHO-SHPS-1-WT cells) ( Fig.  2A). In contrast, incubation of the cells for 4 h in the presence of either 10% FBS or 1 M LPA resulted in a marked increase in the amount of the SHPS-1 ectodomain in the medium ( Fig.  2A). LPA induces activation of protein kinase C (PKC) (32), and PKC activation in turn triggers ectodomain shedding of various membrane-associated proteins (33,34). TPA, a potent activator of PKC, also induced marked release of the SHPS-1 ectodomain into the culture medium of serum-deprived CHO-SHPS-1-WT cells in a time-dependent manner (Fig. 2B), indicating that activation of PKC elicits ectodomain shedding of SHPS-1. We also examined the TPA-induced ectodomain-shedding of endo- geneous SHPS-1 in RAW264.7 cells. Incubation of the RAW264.7 cells for 4 h in the presence of TPA induced a marked release of the SHPS-1 ectodomain into the culture medium (Fig. 2B). The abundance of SHPS-1 in CHO-SHPS-1-WT cells was similar to that in CHO-Ras-SHPS-1-WT cells (Fig. 2C). However, the amount of the SHPS-1 ectodomain present in culture medium after incubation of the two cell types in the absence of serum for 4 h was markedly greater for CHO-Ras-SHPS-1-WT cells than for CHO-SHPS-1-WT cells (Fig. 2C). Furthermore, in contrast to its effect in CHO-SHPS-1-WT cells, TPA did not induce additional shedding of the SHPS-1 ectodomain in CHO-Ras-SHPS-1-WT cells (Fig. 2C). These results thus suggest that the activation of Ras also promotes ectodomain shedding of SHPS-1.
Minimal CD47-binding Activity of the Shed SHPS-1 Ectodomain-To determine whether the cleaved ectodomain of SHPS-1 binds to CD47, we examined the binding of recombinant SHPS-1-Fc (the extracellular region of mouse SHPS-1 fused to the Fc portion of human IgG) to mouse CD47 expressed in CHO-Ras cells (CHO-Ras-CD47 cells). The binding of SHPS-1-Fc to CHO-Ras-CD47 cells increased in a concentration-dependent manner (Fig. 3A). We prepared conditioned medium containing the shed SHPS-1 ectodomain and concentrated it by a factor of 10; the concentration of the SHPS-1 ectodomain in the concentrated medium was substantially greater than 5 g/ml, as estimated by immunoblot analysis with the ␣p84 mAb to SHPS-1 (Fig. 3B). The binding of recombinant SHPS-1-Fc to CHO-Ras-CD47 cells was not inhibited, however, by the presence of concentrated conditioned medium containing the shed SHPS-1 ectodomain (Fig. 3A). The shed ectodomain of SHPS-1 thus appears to possess minimal CD47-binding activity. Given that the shed ectodomain of SHPS-1 was ϳ100 kDa, the ϳ100-kDa portion of SHPS-1-Fc produced by MMPs (band 1) was predicted to be an NH 2 -terminal fragment of the fusion protein and the 35-kDa (band 2) and 30-kDa (band 3) pieces to be COOH-terminal fragments (Fig. 4B). The 35-kDa fragment (band 2) was subjected to direct sequencing and its NH 2 -terminal amino acid sequence was found to be MQTFPGNNA (Fig. 5A). In contrast, the NH 2 -terminal amino acid sequence of the 30-kDa fragment (band 3) revealed it to be derived from the Fc portion of the fusion protein. These data thus suggested that a metalloproteinase cleaves the Ser 359 -Met 360 peptide bond in the juxtamembrane region of SHPS-1 in vitro (Fig. 5A). To confirm that ectodomain shedding of SHPS-1 results from cleavage in the juxtamembrane region of the protein, we generated an SHPS-1 mutant (SHPS-1-FLAG-JM) in which residues 359 to 368 were replaced with a 10-amino acid sequence, including the FLAG epitope (Fig. 5A). We replaced these 10 residues because point mutations in the juxtamembrane region of ephrin-A2 had no effect on ectodomain shedding of this protein (40). We then isolated CHO-Ras cell lines that stably express SHPS-1-FLAG-JM.

Role of a Metalloproteinase in Ectodomain
We chose two cell lines expressing SHPS-1-FLAG-JM (CHO-Ras-SHPS-1-FLAG-JM cells, clones 7 and 12) for experiments, because the intracellular abundance of the mutant protein was similar to that of SHPS-1 in CHO-Ras-SHPS-1-WT cells (Fig.  5B). In contrast to wild-type SHPS-1, minimal ectodomain shedding of SHPS-1-FLAG-JM was detected (Fig. 5B), consistent with our identification of the cleavage site of SHPS-1 targeted by MMPs in vitro as Ser 359 -Met 360 in the juxtamembrane region. In addition, immunoblot analysis with pAbs to SHPS-1 revealed that, in contrast to CHO-Ras-SHPS-1-WT cells, CHO-Ras-SHPS-1-FLAG-JM cells did not contain an SHPS-1 fragment of ϳ20 kDa (Fig. 5B), consistent with the notion that this fragment corresponds to the transmembrane and cytoplasmic regions of SHPS-1 that remain after ectodomain shedding. We also examined the binding of a recombinant CD47-Fc fusion protein, which comprised the extracellular region of mouse CD47 fused to the Fc region of human IgG, to CHO-Ras-SHPS-1-WT and CHO-Ras-SHPS-1-FLAG-JM cells. The binding of CD47-Fc to these two cell lines was virtually indistinguishable (Fig. 5C), suggesting that the ability of the SHPS-1-FLAG-JM mutant to bind to CD47 was similar to that of wild-type SHPS-1. Moreover, we examined tyrosine phosphorylation of SHPS-1 and its association to SHP-2 in CHO-Ras-SHPS-1-WT cells and CHO-Ras-SHPS-1-FLAG-JM cells. Tyrosine phosphorylation of SHPS-1-FLAG-JM and its consequent association of SHP-2 in either presence or absence of pervanadate was almost comparable to those of wild-type SHPS-1 (Fig. 5D).

Impairment of Cell Migration, Cell Spreading, and Cytoskeletal Reorganization by Expression of SHPS-1-FLAG-JM-
SHPS-1 positively regulates cell migration, cell adhesion, cell spreading, and cytoskeletal reorganization through complex formation with SHP-2 (18,22). To investigate whether ectodomain shedding of SHPS-1 is important in such regulation of cell migration and adhesion, we examined the effects of forced expression of SHPS-1-FLAG-JM. CHO-Ras, CHO-Ras-SHPS-1-WT, or CHO-Ras-SHPS-1-FLAG-JM cells were applied to polycarbonate filters in the upper compartments of a Transwell apparatus, and the number of cells that migrated into the lower compartments was determined. Whereas ϳ15% of the applied CHO-Ras cells or CHO-Ras-SHPS-1-WT cells had migrated across the membrane after incubation for 16 h, the corresponding values for both CHO-Ras-SHPS-1-FLAG-JM cell lines were less than 5% (Fig. 6A). We also detached CHO-Ras-SHPS-1-WT and CHO-Ras-SHPS-1-FLAG-JM cells from their culture dishes and replated them on fibronectin for examination of cell spreading by phase-contrast microscopy. After 1 h, most CHO-Ras-SHPS-1-WT cells had become phase-dark and had begun to attach to the dish and spread, and almost all of the cells exhibited a well spread morphology 4 h after plating (Fig. 6B). In contrast, most CHO-Ras-SHPS-1-FLAG-JM cells of both clone 7 (Fig. 6B) and clone 12 (data not shown) attached to the dish but still exhibited a phasebright, rounded morphology 1 h after plating and failed to spread well even 4 h after plating.
Immunofluorescence analysis with rhodamine-conjugated phalloidin also revealed that both CHO-Ras cells and CHO-Ras-SHPS-1-WT cells exhibited a well spread, polarized morphology with prominent actin stress fibers, although the latter cell line appeared to contain fewer stress fibers than did the former (Fig. 6C). In addition, patchy staining for paxillin and vinculin, both of which localize to focal adhesions (41), was observed around the cell periphery as well as at the basal level of the cell body in both CHO-Ras and CHO-Ras-SHPS-1-WT cells (Fig. 6C). In contrast, the failure of CHO-Ras-SHPS-1-FLAG-JM cells to spread was accompanied by the presence of condensed stress fibers in the cell periphery and a consequent smaller cell size compared with that of CHO-Ras or CHO-Ras-SHPS-1-WT cells (Fig. 6C). Some CHO-Ras-SHPS-1-FLAG-JM cells also manifested podosome-like structures at the periph-ery. In addition, the staining for focal adhesions was markedly reduced, especially at the basal level of the cell body, in CHO-Ras-SHPS-1-FLAG-JM cells (Fig. 6C). These observations suggest that expression of the ectodomain shedding-resistant SHPS-1 mutant impaired the reorganization of the actin cytoskeleton that is required for cell adhesion and spreading and thereby impeded cell migration.
We also examined ectodomain shedding of SHPS-1-FLAG-JM in CHO cells, which were not transformed with H-Ras. To this end, we generated CHO cells stably expressing SHPS-1-FLAG-JM (CHO-SHPS-1-FLAG-JM). In contrast to wild-type SHPS-1, minimal ectodomain shedding of SHPS-1-FLAG-JM was detected in CHO cells even in the presence of serum or TPA (Fig. 7A), consistent with the results obtained in CHO-Ras cells. CHO-SHPS-1-FLAG-JM cells exhibited slightly impaired migratory response, compared with CHO cells or CHO-SHPS-1-WT cells, although we could not find statistical difference between them (data not shown). The spreading of CHO-SHPS-1-FLAG-JM cells on fibronectin was also impaired either 1 or 4 h after replating, compared with that apparent in CHO-SHPS-1-WT cells (Fig. 7B). The effect of SHPS-1-FLAG-JM expression on spreading of CHO cells was not marked as that observed in CHO-Ras-SHPS-1-FLAG-JM cells. These weak effects of SHPS-1-FLAG-JM on cell migration and spreading in CHO cells might be partly due to lower expression of SHPS-1-FLAG-JM in CHO cells, compared with that in CHO-Ras cells (data not shown).
Finally, we examined the effects of SHPS-1-FLAG-JM expression on the tyrosine phosphorylation of paxillin and FAK, given the importance of such modification of these focal adhesion-associated proteins in cell spreading and migration (41). The time-dependent increase in the tyrosine phosphorylation of paxillin and FAK apparent in CHO-Ras-SHPS-1-WT cells as they attached to fibronectin was markedly enhanced in CHO-Ras-SHPS-1-FLAG-JM cells (Fig. 8). DISCUSSION We have shown that shedding of the ectodomain of SHPS-1 occurs in cultured CHO cells engineered to express both SHPS-1 and an active form of Ras. In addition, ectodomain shedding of endogenous SHPS-1 was apparent in primary cultured neurons as well as in melanoma and macrophage cell lines. We also demonstrated that serum and LPA each induced shedding of the SHPS-1 ectodomain, suggesting that this process is regulated by extracellular stimuli such as growth factors and lipid mediators. Furthermore, in addition to the effect of forced expression of an activated form of Ras, TPA promoted ectodomain shedding of SHPS-1. Given that LPA induces the activation of both PKC and Ras (32), it is possible that LPA promotes ectodomain shedding of SHPS-1 through activation of these two signaling molecules. Both PKC and Ras mediate ectodomain shedding of various other transmembrane proteins (38,42). SHPS-1 thus appears to be a new member of the group of membrane proteins that undergo ectodomain shedding in response to activation of PKC and Ras. Ephrin-A2, a ligand of the receptor tyrosine kinase Eph, undergoes ectodomain shedding in response to its binding to this receptor (40). However, engagement of SHPS-1 by CD47-Fc failed to promote ectodo-main shedding of SHPS-1, 2 suggesting that, unlike the Ephrin-Eph system, ectodomain shedding of SHPS-1 is not triggered by the binding of its ligand.
Ectodomain shedding often changes the mode of action or biological activity of transmembrane proteins. The SHPS-1 ectodomain present in conditioned medium of CHO-Ras-SHPS-1-WT cells did not inhibit the binding of SHPS-1-Fc to CD47, suggesting that the shed fragment of SHPS-1 has no or minimal ability to bind to CD47. Similarly, in contrast to the membrane-anchored forms of c-kit ligand (5) and of ephrins (6), the soluble forms of these proteins generated by ectodomain shedding exhibit little or no biological activity. We previously proposed a model for the contribution of the CD47-SHPS-1 system to the inhibition of cell migration by cell-cell contact. According to this model, when a migrating SHPS-1-expressing cell makes contact with a nearby CD47-expressing cell, engagement of SHPS-1 by CD47 results in inhibition of cell migration (22). It is therefore possible that ectodomain shedding of SHPS-1 disrupts the interaction of SHPS-1 with CD47 and thereby reverses the inhibition by the CD47-SHPS-1 system of cell migration. In addition, activation of Ras or PKC contributes to the development of various types of cancer. Thus, ectodomain shedding of SHPS-1 in response to the activation of Ras or PKC in cancer cells might result in inactivation of the CD47-SHPS-1 system and thereby promote cancer cell migration and invasion. their culture dishes, replated on dishes coated with fibronectin, and incubated for the indicated times at 37°C. Cell lysates were then prepared and subjected to immunoprecipitation (IP) with mAbs to paxillin (␣Paxillin) or pAbs to FAK (␣FAK), and the resulting precipitates were subjected to immunoblot analysis with horseradish peroxidaseconjugated mAb PY20 to phosphotyrosine (␣PY). Duplicate immunoprecipitates were probed with mAbs to paxillin or pAbs to FAK, as indicated. Data are representative of three separate experiments.
We also showed that the TPA-induced shedding of the ectodomain of SHPS-1 was blocked by inhibitors of metalloproteinases. During the course of this study, Guo et al. (43), with a proteomic approach, identified SHPS-1 as a transmembrane protein that underwent ectodomain shedding in macrophages in an IC-3-sensitive manner. In our experiment, MMP-9 or MMP-1 catalyzed the cleavage of SHPS-1-Fc in vitro. A member of the MMP family might thus be responsible for the ectodomain shedding of SHPS-1 triggered by TPA or other stimuli. ADAMs (A disintegrin and metalloproteinases) are transmembrane-type metalloproteinases that participate in ectodomain shedding of certain transmembrane proteins (35,37). Given that KB-R7785 inhibits ADAM12-mediated ectodomain shedding of HB-EGF (37), it is possible that an ADAM isoform, rather than MMPs, mediates ectodomain shedding of SHPS-1.
Determination of the NH 2 -terminal sequence of a fragment of SHPS-1-Fc generated by an MMP indicated that the primary cleavage site of SHPS-1 during ectodomain shedding is located in the juxtamembrane region. Ectodomain shedding was thus not observed in CHO-Ras cells expressing a mutant (FLAG-JM) of SHPS-1 in which 10 amino acids in this region were replaced by an unrelated sequence. The cleavage sites for ectodomain shedding of other transmembrane proteins are also located in their juxtamembrane regions (44,45).
We found that forced expression of SHPS-1-FLAG-JM markedly impaired cell migration and spreading of CHO-Ras cells. This effect was also observed in non-Ras-transformed CHO cells. In addition, the cytoskeletal architecture of CHO-Ras cells expressing this mutant was abnormal, with condensed actin fibers apparent at the cell periphery and a reduced number of focal adhesions. In addition, the tyrosine phosphorylation of paxillin and FAK triggered by the cell adhesion was markedly enhanced in these cells. Consistently, treatment of CHO-Ras-SHPS-1-WT cells with the metalloproteinase inhibitor KB-R7785, which prevented ectodomain shedding of SHPS-1, also inhibited cell migration and spreading. 2 Together, these results suggest that SHPS-1 ectodomain shedding is required for cell migration, cell spreading on substrate, and the accompanying reorganization of the actin cytoskeleton.
The mechanism responsible for the observed phenotypes of CHO-Ras-SHPS-1-FLAG-JM cells remains to be determined. The ability of the SHPS-1-FLAG-JM mutant to bind to CD47 was similar to that of wild-type SHPS-1. Tyrosine phosphorylation of SHPS-1-FLAG-JM and consequent association of SHP-2 in either presence or absence of pervanadate was comparable to those of wild-type SHPS-1. However, it is possible that the shedding-resistant mutant of SHPS-1 somehow acts in a dominant negative manner. Indeed, the phenotypes of CHO-Ras-SHPS-1-FLAG-JM cells resemble those of fibroblasts from mice that lack most of the cytoplasmic region of SHPS-1, which also manifest a marked impairment of cell migration and spreading (18). In addition, formation of a complex between SHPS-1 and SHP-2 promotes cell migration through regulation of actin cytoskeletal reorganization (18,22). Cell migration or cell spreading is markedly impaired in SHP-2-deficent cells or cultured fibroblasts expressing a dominant negative mutant of SHP-2 (13,46). In addition, the tyrosine phosphorylation of paxillin or FAK is markedly increased in these cells (13,46), indicting that SHP-2 activity is required for tyrosine dephosphorylation of these proteins. The recruitment of SHP-2 to lipid rafts was recently shown to contribute to the spatial regulation of Rho and thereby to promote cell migration (16). We found that a proportion of SHPS-1 molecules was localized to lipid rafts in CHO-Ras-SHPS-1-WT cells; however, a similar proportion of SHPS-1-FLAG-JM was detected in lipid rafts in CHO-Ras-SHPS-1-FLAG-JM cells. 3 The ectodomain shedding of SHPS-1 thus might not influence the localization of this protein and subsequent recruitment of SHP-2 to lipid rafts.
The phenotypes of CHO-Ras-SHPS-1-FLAG-JM cells resemble those of cultured cells that overexpress wild-type Csk (47). Although Csk was shown to regulate the activity of Src family kinases negatively, overexpression of Csk also increased the tyrosine phosphorylation of paxillin and FAK triggered by cell adhesion (47). SHPS-1 has been shown to bind Csk (48), and thus the shedding-resistant mutant of SHPS-1 could contribute to the recruitment of Csk to focal adhesions, thereby increasing the tyrosine phosphorylation of paxillin and FAK. Further investigation is thus necessary to clarify the mechanism by which ectodomain shedding of SHPS-1 modulates the function of this protein in regulation of cell spreading and migration.