Protein tyrosine phosphatase SHP2 suppresses podosome rosette formation in Src-transformed fibroblasts

The reciprocal regulation of protein tyrosine phosphorylation by protein tyrosine kinases and phosphatases is involved in the control of many important cellular structures and functions. The Src homolog domain-containing phosphatase 2 (SHP2), which is a cytoplasmic protein tyrosine phosphatase (PTP), has been found to play an important role in a variety of cellular functions, including cell migration (Yu et al., 1998), cell proliferation (Qu and Feng, 1998), and cell survival (Yang et al., 2006). However, the role of SHP2 in tumorigenesis remains controversial. SHP2 has been suggested to function as an oncogene by several studies (Aceto et al., 2012; Matozaki et al., 2009; Xu, 2007) and as a tumor suppressor by others (Bard-Chapeau et al., 2011). Several cellular proteins have been shown to be substrates of SHP2, including Src (Peng and Cartwright, 1995), focal adhesion kinase (FAK; Tsutsumi et al., 2006), paxillin (Ren et al., 2004), and Gab1 (Montagner et al., 2005). The dephosphorylation of Src at Y527 by SHP2 induces Src activation, which may lead to activation of the Ras/ERK signaling pathway (Zhang et al., 2004). The dephosphorylation of the focal adhesion proteins FAK and paxillin by SHP2 has been suggested to regulate focal adhesion dynamics and cell motility (Mañes et al., 1999). We have shown previously that the phosphorylation of Rho-associated kinase (ROCK) II at tyrosine 722 (Y722) by Src leads to its inhibition (Lee et al., 2010), whereas the dephosphorylation of this site by SHP2 potentiates its activation by RhoA (Lee and Chang, 2008). The reciprocal regulation of ROCK II by Src and SHP2 is important for cell contractility and motility (Lee et al., 2010).

Podosomes are F-actin-enriched structures that are important for extracellular matrix degradation and invasive cell motility (Linder, 2007). Podosomes often undergo self-assembly into large rosette-like structures in Src-transformed fibroblasts, osteoclasts, endothelial cells, and some highly invasive cancer cells (Linder and Aepfelbacher, 2003). Podosome rosettes are more potent than individual podosome dots in matrix degradation (Pan et al., 2011). Podosome rosettes are dynamic structures that have lifespans ranging from minutes to hours (Pan et al., 2011). Many scaffolding proteins, such as Tks5 (Seals et al., 2005; Stylli et al., 2009), cortactin (Oser et al., 2009), and paxillin (Badowski et al., 2008), are essential components of podosomes. The tyrosine phosphorylation of these proteins has been shown to be important for podosome formation. Several protein tyrosine kinases, such as Src and FAK, are important for the organization and function of podosome rosettes (Destaing et al., 2008; Pan et al., 2011). However, little is known regarding the identity and mechanisms of the PTPs that are involved in podosome rosette formation. The aim of the present study is to identify the PTPs that are involved in podosome rosette formation using v-Src-transformed mouse embryo fibroblasts (MEFs) as a model.

A short-hairpin RNA (shRNA) approach was used to identify the PTPs that are important for podosome rosette formation in Src-transformed MEFs. Among the examined shRNAs, those specific to SHP2 (sh-SHP2) appeared to increase the formation of podosome rosettes (supplementary material Fig. S1). We found that the shRNA-mediated knockdown of SHP2, but not LAR (a transmembrane PTP), significantly increased the formation of podosome rosettes in Src-transformed MEFs (Fig. 1). The stimulatory effect of SHP2 knockdown on the podosome rosette formation was also observed in another two independent clones of Src-transformed MEFs (supplementary material Fig. S2). This increase by SHP2 knockdown was suppressed by the re-expression of FLAG-tagged human SHP2 (FLAG-SHP2) (Fig. 2), indicating that the shRNA was specific to SHP2. In addition, the overexpression of FLAG-SHP2 apparently inhibited podosome rosette formation in Src-transformed MEFs (Fig. 3). The formation of podosome rosettes was correlated with the capability of the cells to degrade matrix proteins and to invade through Matrigel (Figs 2, 3), which was in keeping with our previous findings (Pan et al., 2011). Although SHP2 was previously reported to modulate Src activity (Zhang et al., 2004), we found that the Y416 phosphorylation of Src was not affected by SHP2 in Src-transformed MEFs (Figs 2, 3), suggesting that the inhibitory effect of SHP2 on podosome rosette formation is not due to its effect on Src. We also observed that the knockdown of SHP2 or LAR had no effect on individual podosome dots in Src-transformed MEFs (Fig. 1D,E). These findings together suggest that SHP2 negatively regulates the formation of podosome rosettes but not of podosome dots in Src-transformed fibroblasts.

FLAG-SHP2 and its phosphatase-defective mutant (C459S mutant) were both found to be localized in podosome rosettes (Fig. 4C), indicating that the phosphatase activity of SHP2 is not required for its localization in podosome rosettes. However, the C459S mutant was deficient in the ability to suppress podosome rosette formation (Fig. 4E), indicating that the catalytic activity of SHP2 is important for the downregulation of podosome rosette formation in Src-transformed fibroblasts. A SHP2 mutant with deletion of both its Src homolog domain 2 (SH2) domains was found to be diffusively distributed in the cytoplasm and not localized in podosome rosettes (Fig. 4C). This finding suggests that the interaction of the SH2 domains with certain tyrosine phosphorylated proteins in podosome rosettes is important for the targeting of SHP2 to rosettes. Although the PTP domain per se did not specifically target podosome rosettes (Fig. 4C), it inhibited rosette formation (Fig. 4E). Because the PTP domain per se displayed strong catalytic activity (Fig. 4B), it is possible that the dephosphorylation of certain podosomal scaffolding proteins by the SHP2 PTP domain may prevent the initial assembly of podosome rosettes.

The tyrosine phosphorylation of Tks5 has been shown to be important for the formation of podosomes and invadopodia (Seals et al., 2005; Stylli et al., 2009). We found that SHP2 selectively inhibited the tyrosine phosphorylation of Tks5 in Src-transformed MEFs (Fig. 5A). In contrast, the tyrosine phosphorylation of several other scaffolding proteins, including FAK, paxillin, cortactin, and p130Cas, was not affected by SHP2 (Fig. 5A; supplementary material Fig. S3A). Tks5 was apparently co-precipitated with the substrate-trapping mutant (D425A/C459S) of SHP2 (Fig. 5B) and directly dephosphorylated by SHP2 in vitro (Fig. 5C).

Rho and ROCK have been shown to be involved in the regulation of podosome rosette formation (Moreau et al., 2003; Pan et al., 2011; Varon et al., 2006). We reported recently that Src phosphorylates ROCK II at Y722, leading to the inhibition of ROCK II activity (Lee et al., 2010). Conversely, the dephosphorylation of ROCK II at Y722 by SHP2 potentiates its activation by RhoA (Lee and Chang, 2008). In the present study, we found that although the activity and expression of Rho, Rac, and Cdc42 were not affected by SHP2 in Src-transformed fibroblasts (supplementary material Fig. S3B), the Y722 phosphorylation of ROCK II was inversely correlated with SHP2 expression (Fig. 6A,B). The reduction in Y722 phosphorylation of ROCK II by SHP2 was correlated with an increase in ROCK II activity (Fig. 6C,D).

The ROCK inhibitor Y27632 was found to reverse the inhibitory effect of SHP2 on podosome rosette formation (Fig. 7A), supporting the proposed role of ROCK in the SHP2-mediated suppression of podosome rosette formation. To test the idea that ROCK II is a negative regulator of podosome rosette formation, Myc-tagged ROCK II and its mutants were transiently expressed in Src-transformed 3T3 cells and their effects on podosome rosette formation were measured (Fig. 7B). The Y722F mutant (which is refractory to phosphorylation by Src) displayed a stronger suppressing effect on podosome rosette formation than did the wild-type ROCK II. The constitutively active form of ROCK II completely inhibited the podosome rosette formation (Fig. 7B). In contrast, knockdown of endogenous ROCK II significantly increased the formation of podosome rosettes in Src-transformed MEFs (Fig. 7C).

We demonstrated previously that the Rho-ROCK axis promotes the polymerization of vimentin intermediate filaments (VIFs) and thereby inhibits the formation of podosome rosettes (Pan et al., 2011). We found in the present study that the formation of VIFs was suppressed by the knockdown of SHP2 (Fig. 8A) but was promoted by the overexpression of SHP2 (Fig. 8B). A higher content of polymerized VIFs was correlated with a lower content of podosome rosettes (Figs 2, 3). The enhancement of VIF formation by SHP2 overexpression was suppressed by the ROCK inhibitor Y27632 (Fig. 8B), with a concomitant increase of podosome rosettes (Fig. 7A). These findings, taken together, suggest that the SHP2-mediated suppression of podosome rosette formation occurs, at least in part, through ROCK activation and VIF polymerization.

To directly measure the effect of ROCK on VIFs, ROCK II was suppressed or overexpressed in Src-transformed MEFs. The knockdown of endogenous ROCK II apparently decreased the content of polymerized VIFs in the cells (Fig. 8C). The transient expression of the constitutively active form of ROCK II enhanced the polymerization of VIFs (Fig. 8D). In contrast, the kinase-deficient mutant (K121G) of ROCK II had no such effect (Fig. 8D). These results suggest that the catalytic activity of ROCK II is required for promoting VIFs.

In this study, we report that SHP2 functions as a negative regulator of podosome rosette formation in Src-transformed fibroblasts (Figs 1–f02,3). Based on our findings, we propose that SHP2 suppresses podosome rosette formation through at least two routes: (i) the dephosphorylation of certain critical podosomal scaffolding proteins (e.g. Tks5); (ii) ROCK-mediated VIF polymerization (Fig. 9). The phosphatase activity of SHP2 is not required for the targeting of SHP2 to podosome rosettes but is essential for the suppression of podosome rosette formation by SHP2 (Fig. 4). Consistent with this finding, the tyrosine phosphorylation of Tks5 was selectively decreased by SHP2 (Fig. 5). Tks5 is a Src substrate and its tyrosine phosphorylation is important for podosome formation (Seals et al., 2005; Stylli et al., 2009). Tks5 has never been reported to be a substrate of SHP2 to the best of our knowledge. Although FAK is a known SHP2 substrate (Tsutsumi et al., 2006), its tyrosine phosphorylation was not affected by SHP2 in Src-transformed fibroblasts (Fig. 5).

We demonstrated previously that depletion of FAK leads to activation of Rho and ROCK and polymerization of VIFs and that these processes inhibit the assembly of podosome rosettes (Pan et al., 2011). In the present study, the SHP2 depletion reduced ROCK II activity, whereas SHP2 overexpression activated ROCK II (Fig. 6). The increased activation of ROCK by FAK depletion (Pan et al., 2011) or by SHP2 overexpression (present study) leads to enhanced VIF polymerization (Fig. 8). We provide direct evidence to support that ROCK II facilitates VIF polymerization. Our results indicate that a higher content of polymerized VIFs was correlated with a lower content of podosome rosettes, in keeping with our previous findings (Pan et al., 2011). The mechanism for the inhibitory effect of VIFs on podosome rosettes is currently under investigation.

To assess the possible role of SHP2 in the podosomal organization of other types of cells such as osteoclasts, we depleted SHP2 in mouse RAW264.7 cells, which can differentiate into osteoclast-like cells by receptor activator of NFκB ligand (RANKL; Boyle et al., 2003). However, SHP2 depletion did not appear to affect the podosomal organization of such osteoclast-like cells (supplementary material Fig. S4), indicating that SHP2 is not involved in the regulation of podosomal organization in osteoclasts. PTPε was shown to play a positive role in the podosomal organization and bone absorption function of osteoclasts (Chiusaroli et al., 2004). Taken together, these findings suggest that SHP2 may affect podosome rosette formation in certain types of cells but not in others. Experiments are in progress to examine the effect of SHP2 on podosome rosette formation in endothelial cells and carcinoma cells.

Podosomes and invadopodia are similar structures and are both important in extracellular matrix degradation and invasive cell motility (Linder, 2007). Very little is known regarding the role of PTPs in these structures. Only two PTPs have been reported to participate in the regulation of podosomes or invadopodia: PTPε in osteoclasts (Chiusaroli et al., 2004) and PTP1B in breast cancer cells (Cortesio et al., 2008). The stimulatory effects of PTPε and PTP1B on podosomes/invadopodia are through activating Src (Chiusaroli et al., 2004; Cortesio et al., 2008). In fact, many prior reports support a critical role of Src in podosomes/invadopodia in various types of cells. Src appears to be generally involved in podosome formation. We propose that when Src is highly activated, it phosphorylates and inhibits ROCK, which facilitates the formation of podosome rosettes. Under such circumstances, SHP2 becomes to counteract the action of Src on ROCK in order to facilitate the dynamics of podosomal structures. However, it has been reported that SHP2 can also activate Src by dephosphorylating Src Y527 (Zhang et al., 2004). From this point of view, SHP2 may play a positive role in podosome formation by activating Src. Therefore, the role of SHP2 in podosomes may be cell-type and/or stimulus-dependent. We have demonstrated here that SHP2 plays an inhibitory role in podosome rosette formation in Src-transformed fibroblasts. Additional studies are needed to clarify the mechanisms whereby PTPs affect the formation of podosomes/invadopodia and the assembly of podosomal super structures in various types of cells.

Polyclonal anti-Cdc42, anti-cortactin (H-191), anti-Tks5 (M300), and anti-ROCK II (H-85 for immunoblotting) antibodies and monoclonal anti-SHP2 (B-1), anti-Myc (9E10), anti-RhoA (26C4), and anti-β tubulin (D-10) antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Polyclonal anti-Myc antibody was purchased from Cell Signaling Technology (Beverly, MA). Monoclonal anti-p130Cas, anti-LAR, anti-FAK, anti-phosphotyrosine (4G10), anti-paxillin, and anti-Rac1 antibodies, and Matrigel were from BD Transduction Laboratories (San Jose, CA). Monoclonal anti-FLAG, and anti-vimentin (clone VIM 13.2 for immunofluorescence staining) antibodies, myelin basic protein (MBP), gelatin, ρ-NPP (para-nitrophenylphosphate), and protein A-Sepharose beads were from Sigma-Aldrich (St Louis, MO). Polyclonal anti-FAK pY577, anti-FAK pY397, anti-paxillin pY118, and anti-Src pY416 antibodies, Dulbecco's modified Eagle's medium (DMEM), Zeocin, and Lipofectamine were from Life Technologies-Invitrogen (Carlsbad, CA). Polyclonal anti-ROCK II (for immunoprecipitation) antibody, fibronectin, puromycin, and Y27632 were from Millipore (Billerica, MA). Polyclonal anti-ROCK II pY722 antibody was prepared as described previously (Lee and Chang, 2008). Mouse ascites containing monoclonal anti-Src (peptide 2-17) antibody produced by hybridoma (CRL-2651) was prepared in our laboratory.

The plasmids pFLAG-CMV2-human SHP2 WT and C459S were kindly provided by D.-L. Wang (Tzu Chi University, Hualien, Taiwan). The plasmids pLKO-AS2.zeo-FLAG-SHP2 WT, C459S, and PTP [amino acids (aa) 268–593] were constructed in our laboratory. The plasmids pEF-myc-ROCKII WT and Y722F were described previously (Lee and Chang, 2008). pEF-Bos-myc-ROCK C.A. (constitutively active; aa 85–355) was described previously (Chen et al., 2002). The plasmids pEF-myc-ROCK II KD (aa 6–553; K121G) and pFLAG-CMV2-human SHP2 D425A/C459S (substrate-trapping mutant) were kindly provided by H.-H. Lee (National Yang-Ming University, Taipei, Taiwan). The plasmid pCMV-Tag-2B-Tks5 was kindly provided by T. Oikawa (Keio University, Tokyo, Japan) and described previously (Oikawa et al., 2012).

v-Src-transformed MEFs and SrcY527F-transformed NIH3T3 cells were described previously (Pan et al., 2011). For transient expression, cells were transfected with plasmids encoding FLAG-SHP2, FLAG-Tks5, or Myc-ROCK II using Lipofectamine.

The lentiviral expression system was provided by the National RNAi Core Facility, Academia Sinica, Taiwan. For FLAG-SHP2 expression, human FLAG-SHP2 cDNA was amplified by polymerase chain reaction and subcloned in frame to the NheI and EcoRI site of the pLKO-AS2-zeo vector. The plasmids pLKO-AS1-puro, which encodes shRNAs, were obtained from the National RNAi Core Facility, Academia Sinica. The target sequences for mouse SHP2 are 5′-CCTGATGAGTATGCGCTCAAA-3′ (#1) and 5′-GACATCCTTATTGACATCATT-3′ (#2). The target sequences for mouse LAR are 5′- GCCGTATGTGAAATGGATGAT-3′ (#1) and 5′- CCACCAGTGTTACTCTGACAT-3′ (#2). The target sequences for mouse ROCKII are 5′-GCATCTCTTGAAGAAACAAAT-3′ (#1) and 5′-CCCATGGATCAGAGATAATTA-3′ (#2). To produce lentiviruses, HEK293T cells were co-transfected with pCMV-ΔR8.91 (2.25 µg), pMD.G (0.25 µg), and pLKO-AS1-puro-shRNA (or pLKO-AS2-zeo-FLAG-SHP2; 2.5 µg) using Lipofectamine. After 3 days, medium containing lentivirus particles was collected and stored at −80°C. Cells were infected with lentiviruses encoding shRNAs or FLAG-SHP2 for 24 hours and subsequently selected in the growth medium containing 1.8–2.5 µg/ml puromycin.

Immunoblotting and immunoprecipitation were performed as described previously (Chen and Chen, 2006). Chemiluminescent signals were detected and quantified using a luminescence image system (LAS-3000, Fujifilm).

FLAG-SHP2 WT and its mutants were transiently expressed in HEK293 cells and FLAG-SHP2 proteins were immunoprecipitated by anti-FLAG antibody. The immune complexes were washed three times with 1% NP-40 lysis buffer (1% Nonidet P-40, 20 mM Tris-HCl, pH 8.0, 137 mM NaCl, 10% glycerol) and twice with washing buffer (50 mM Tris pH 7.4, 137 mM NaCl, 2 mM EDTA) and subjected to a phosphatase activity assay using ρNPP as the substrate. This assay was performed in phosphatase buffer (50 mM Tris pH 7.4, 137 mM NaCl, 2 mM EDTA, 5 mM dithiothreitol, 20 mM ρNPP) at 37°C for 1.5 hours and the reaction was terminated by the addition of 0.2 N NaOH. Upon release of the phosphate group from ρNPP, a ρ-nitrophenolate anion was produced and determined by absorbance at 405 nm. To ensure equal amounts of FLAG-SHP2 proteins, immune complexes were analyzed by immunoblotting with anti-SHP2.

To assay whether Tks5 is a substrate of SHP2, FLAG-Tks5 proteins were transiently co-expressed with v-Src in 293 cells and purified by anti-FLAG affinity chromatography. The purified Tks5 proteins (100 ng) were incubated with immobilized FLAG-SHP2 or its phosphatase mutants in 50 µl of the buffer (25 mM HEPES pH 7.4, 150 mM NaCl, 5 mM EDTA, and 2 mM dithiothreitol) at 37°C for 60 minutes. The reaction was terminated by SDS sample buffer and the proteins were fractionated by SDS-PAGE. The tyrosine phosphorylation of Tks5 was analyzed by immunoblotting with anti-phosphotyrosine.

Cell lysates were incubated with 2 µg polyclonal anti-ROCK II antibody for 1.5 hours at 4°C. Immunocomplexes were collected on protein A-Sepharose beads and washed three times with 1% Nonidet P-40 lysis buffer and twice with 20 mM Tris-HCl, pH 7.4. Kinase reactions were performed in 40 µl of kinase buffer (20 mM Tris pH 7.4, 10 mM MgCl₂, 3 mM NaCl, 1 mM dithiothreitol, 1 mM EDTA) containing 10 µCi of [γ-³²P]ATP and 2.5 µg of MBP for 30 minutes at 25°C. The reaction was terminated by the addition of SDS sample buffer and the proteins were fractionated by SDS-PAGE.

This assay was performed as described previously (Pan et al., 2011). v-Src-transformed MEFs were seeded on coverslips coated with Alexa-Fluor-488-conjugated fibronectin for 48 hours. The matrix degraded areas were determined using the Image-Pro Plus® software version 5.1. Ten random fields equivalent to 2 mm² were measured.

This assay was performed in a 24-well transwell chamber (pore size 8 µm, Costar) coated with 100 µl Matrigel. Cells (5×10³) were added to the upper compartment of the chamber. After 48 hours, the cells that had migrated through the Matrigel were fixed by methanol, stained by Giemsa Stain, and counted.

Immunofluorescence staining was performed as described previously (Pan et al., 2011). The primary antibodies used were monoclonal anti-Myc (1:100), anti-FLAG (1:500), and anti-vimentin (1:100) and polyclonal anti-cortactin (1:200). Rhodamine-conjugated phalloidin and Alexa-Fluor-488-conjugated phalloidin (Molecular Probes, Invitrogen) were used to stain actin filaments. Coverslips were mounted in anti-Fade Dapi-Fluoromount-G™ (Southern Biotechnology Associates). Images were detected using a Zeiss LSM510 laser-scanning confocal microscopic imaging system (LSM 510; Carl Zeiss) with a Zeiss 63 Plan-Apochromat (NA 1.2 W Korr).

The data were analyzed by Student's t-test. Differences were considered to be statistically significant at P<0.05.

We are grateful to D.-L. Wang (Tzu Chi University, Hualien, Taiwan) for SHP2 cDNA and T. Oikawa (Keio University, Tokyo, Japan) for Tks5 cDNA. All authors declare no potential conflict of interest.