ABSTRACT
The Src-associated substrate during mitosis with a molecular mass of 68 kDa (Sam68) is predominantly nuclear and is known to associate with proteins containing the Src homology 3 (SH3) and SH2 domains. Although Sam68 is a Src substrate, little is known about the signaling pathway that link them. Src is known to be activated transiently after cell spreading, where it modulates the activity of small Rho GTPases. Herein we report that Sam68-deficient cells exhibit loss of cell polarity and cell migration. Interestingly, Sam68-deficient cells exhibited sustained Src activity after cell attachment, resulting in the constitutive tyrosine phosphorylation and activation of p190RhoGAP and its association with p120rasGAP. Consistently, we observed that Sam68-deficient cells exhibited deregulated RhoA and Rac1 activity. By using total internal reflection fluorescence microscopy, we observed Sam68 near the plasma membrane after cell attachment coinciding with phosphorylation of its C-terminal tyrosines and association with Csk. These findings show that Sam68 localizes near the plasma membrane during cell attachment and serves as an adaptor protein to modulate Src activity for proper signaling to small Rho GTPases.
The Src-associated substrate during mitosis with a molecular mass of 68 kDa (Sam68) is a known substrate of Src family kinases during mitosis (44). Sam68 harbors numerous proline- and tyrosine-rich regions that interact with Src family kinases, phospholipase Cγ1, Grb2, Nck, and others in a manner dependent on the Src homology 3 (SH3) and SH2 domains (22, 45, 62, 70, 76, 78, 81). Thus, Sam68 was proposed to function as an adaptor protein for Src family kinases (62). However, this putative function has remained elusive and was thought to occur during mitosis, since Sam68 is predominantly nuclear during the rest of the cell cycle (12).
Besides putatively functioning as a signaling protein, Sam68 harbors a KH-type RNA binding domain (42). Since the tyrosine phosphorylation of Sam68 was shown to negatively regulate its RNA binding activity, it is termed a STAR (signal transduction activator of RNA) protein (44). Although KH-type RNA binding domains are known to mediate specific protein-RNA interactions (40, 42, 79), few Sam68 mRNA targets have been identified (32, 56). Sam68 is also known to regulate the human immunodeficiency virus Rev nuclear export pathway (41, 59) and was proposed to mark nuclear mRNAs for their cytoplasmic fate (17). Sam68 has been shown to regulate the alternative splicing of CD44 and Bcl-x (13, 47, 55) and a subset of genes during neurogenesis (11). Sam68 may be involved in flagging variable exons for the polymerase II machinery (3). Moreover, Sam68 has been shown to regulate transcription (2, 14, 27). Sam68 knockout mice have revealed a role in bone physiology, as Sam68−/− mice display continuous bone remodeling with age and therefore are protected against age-induced osteoporosis (60).
Src family kinases are membrane-bound tyrosine kinases that have been shown to be required for key cellular roles, including cell signaling, differentiation, development, polarization, and migration (8, 30, 35, 46, 70). Integrin engagement induced when fibroblasts are plated on extracellular matrices, such as fibronectin, has been shown to promote the activation, membrane association, and redistribution of Src to focal adhesions (36). Integrin stimulation also leads to focal adhesion kinase (FAK) autophosphorylation at tyrosine 397, which in turn stimulates the association with the Src SH2 domain (68), and the subsequent phosphorylation of FAK tyrosines 576 and 577 ensues, activating its kinase function and altering cell morphology and migration (9).
Cell polarity and directional cell migration are regulated by cytoskeletal reorganization through the small family of Rho GTPases activated downstream of integrin and growth factor signaling (51, 65). RhoA activity is known to regulate the formation of stress fibers, whereas Rac1 regulates membrane ruffling (63, 64). The slow intrinsic GTPase activity of the Rho GTPases is enhanced by GTPase-activating proteins (GAPs) (77), and Src kinases are able to downregulate RhoA by inducing tyrosine phosphorylation and activation of p190RhoGAP (1, 6, 49). Moreover, Src kinases have been observed to activate Rac1 by stimulating the phosphorylation of guanine exchange factors Vav and Tiam1 (38, 49, 71).
Src kinase activity is negatively regulated by phosphorylation at the C-terminal regulatory tyrosine (Y529 in human c-Src) by another cytoplasmic tyrosine kinase, the carboxyl-terminal Src kinase (Csk) (52, 53). The phosphorylation of the Y529 adopts a catalytically inactive conformation by intramolecular interactions (82). In response to extracellular stimuli, Src kinases become active by the dephosphorylation of the regulatory site. Csk null mice are embryonic lethal, and dysregulated Src activity is observed in early embryos and cells derived from the embryos (52). A conditional allele using the keratin-5 promoter/Cre-loxP in squamous epithelial cells was generated, and the mice exhibit an epithelial cell conversion defect (83). The cells have activated Src kinase activity that leads to elevated Rac1 and diminished RhoA activities leading to cell migration defects (83). The mice developed epidermal hyperplasia, but the activation of Src kinases by Csk depletion is not sufficient for cancer initiation (83). In Csk-deficient cells, Src is activated and localized in focal adhesions, and the cells have migration defects (28), whereas in normal cells, Csk is recruited in a phosphotyrosine-dependent manner by adaptor proteins that are generally Src substrates (66).
Herein we present evidence that during cell spreading Sam68 localizes near the plasma membrane and modulates Src and p190RhoGAP activities. We observed that Sam68−/− mouse embryo fibroblasts (MEFs) or Sam68 knockdown MEFs and HeLa cells have defects in cell polarization and cell migration. The Sam68−/− MEFs contained constitutively low RhoA and elevated Rac1 GTPase activities. Src activity and p190RhoGAP tyrosine phosphorylation were sustained in Sam68−/− MEFs after cell attachment. Sam68 was localized near the plasma membrane, rapidly tyrosine phosphorylated, and associated with Csk after cell attachment. These data define a new role for the Sam68 “nuclear” protein near the plasma membrane and link it to Src during integrin signaling.
MATERIALS AND METHODS
DNA constructs.The GFP-Sam68 expression vectors encode an N-terminal green fluorescent protein (GFP) and were described previously (12). The mouse Sam68sh vector was described previously (60). The plasmids encoding the full-length Csk and the SH2 domain of Csk were a kind gift of André Veillette (Institut de Recherches Cliniques de Montréal [IRCM], Montréal, Québec, Canada) (15). The human Sam68sh vector was constructed by ligation of the following oligonucleotides: 5′-GAT CCC Cga tgg agc cag aga aca agT TCA AGA Gac ttg ttc tct ggc tcc atc TTT TTG GAA A-3′ and 5′-GCT TTT CCA AAA Aga tgg agc cag aga aca agT CTC TTG Aac ttg ttc tct ggc tcc atc GGG-3′ (the nucleotides of the targeting sequence are shown as lowercase underlined letters). The resulting oligonucleotides were then subcloned into the BglII and HindIII sites of pSuper-Retro (OligoEngine, Seattle, WA).
TIRF microscopy.Images were acquired with an objective-based Olympus (Tokyo, Japan) total internal reflection fluorescence (TIRF) illumination arm attached to an IX81 inverted microscope using a 1.45-numerical-aperture 60× oil immersion lens (7). For GFP imaging, excitation was from the 488-nm line of a 200-mW Ar ion laser attenuated to ∼1% power with both neutral-density filters and an acousto-optic tunable filter (Prairie Technologies, Middleton, WI). Specialized Ar-coated TIRF filters, z488rdc dichroic, hq525/50m emission filter, were from Chroma Technology Corp. (Rockingham, VT). Images were acquired with MetaMorph software (Molecular Devices, Sunnyvale, CA) using a Retiga Exi charge-coupled-device camera (QImaging, Surrey, British Columbia, Canada). For wide-field images, the laser was tuned to be focused straight through the sample. Live cells were maintained at 37°C with 5% CO2 using incubation chambers from Solent Scientific Ltd. (Segensworth, United Kingdom).
Sam68-deficient cells.Sam68+/+ and Sam68−/− MEFs were isolated from 14.5-day-old embryos (60) and immortalized on a 3T3 passage protocol. Sam68sh or pSuper MEFs were obtained by stable transfection of wild-type immortalized MEFs with a pSUPERretro vector that encodes a mouse-specific short hairpin RNA (shRNA) as previously reported (60). HeLa cells Sam68sh and pSuper cells were obtained by stable transfection with a human-specific shRNA as described above.
Protein analysis.MEFs, HEK293, and HeLa cells were transfected using Lipofectamine and Plus reagent (MEFs) and Lipofectamine 2000 (HEK293 and HeLa cells) as recommended by the manufacturer (Invitrogen). For in vitro kinase assays, the cells were lysed using Triton X-100 buffer as previously described (12), and Src was immunoprecipitated using anti-Src antibodies and protein A-Sepharose. The beads were washed twice in lysis buffer and once in cold kinase buffer (25 mM HEPES [pH 7.4], 3 mM MnCl2, 5 mM MgCl2, and 100 μM sodium vanadate). The beads were divided into two portions; half of the beads were used for immunoblotting, and half were used in the in vitro kinase assay. For the kinase assay, the beads were incubated with enolase and 5 μCi [γ-32P]ATP per reaction mixture for 20 min at room temperature, and the proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Immunoblotting was performed using the anti-Sam68 AD1 antibodies (12), anti-Sam68 pY440 (43), anti-protein arginine N-methyltransferase 1 (16), antiactin (catalog no. A3853; Sigma), anti-phosphotyrosine 4G10 (catalog no. 05-1050; Millipore), anti-Csk (catalog no. 4965; Cell Signaling), anti-FAK (catalog no. 610087; BD Biosciences), anti-FAK pY576 (catalog no. 44-652ZG; Biosource), anti-FAK pY397 antibodies (catalog no. 44-624ZG; BD Biosciences), anti-p120rasGAP (catalog no. 05-178; Millipore), anti-p190RhoGAP (catalog no. 610150; BD Biosciences), and antipaxillin (catalog no. AB3794; Millipore) and anti-paxillin pY118 antibodies (catalog no. 44-722G; Biosource). Immunoreactive proteins were visualized using either goat anti-mouse (catalog no. A4416; Sigma) or goat anti-rabbit (catalog no. 55690; Cedarlane) antibodies conjugated to horseradish peroxidase and the chemiluminescence (ECL) detection kit (DuPont).
Cell migration assays.Transwell chamber filters (8-μm pore size; Costar) were coated with 15 μg of collagen/ml on both sides. A total of 104 cells resuspended in Dulbecco modified Eagle medium (DMEM) with 1% bovine serum albumin (BSA) were added to the upper chamber. DMEM supplemented with 10% bovine serum was used as the chemoattractant in the lower chamber. Cell migration was surveyed after 3, 6, 12, and 24 h. Migratory cells on the lower membrane surface were then stained with crystal blue and counted manually. For the wound-healing assay, fibronectin-coated glass coverslips were seeded with either Sam68+/+ or Sam68−/− MEFs and maintained until the cells reached confluence. Wounds were introduced using standard plastic pipette tips on the coverslip surface. Wound healing was surveyed after 16 h, the cells were fixed with paraformaldehyde and permeabilized, and the DNA was stained with 4′,6′-diamidino-2-phenylindole (DAPI) and visualized by fluorescence microscopy.
Cell spreading and immunofluorescence.Cells were fixed with 4% paraformaldehyde in 1× phosphate-buffered saline (PBS) for 5 min at room temperature and permeabilized with 1% Triton X-100 in PBS for 5 min at room temperature as previously described (12). To assess cellular spreading, cells were detached using trypsin and subsequently inactivated with serum-free DMEM containing 0.5 mg/ml soybean trypsin inhibitor (Sigma, St. Louis, MO) and 0.5% BSA. The cells were washed twice with DMEM, then resuspended in DMEM with 0.1% BSA, and tumbled for 1 h at 37°C. The cells were then plated on coverslips previously coated overnight at 4°C with 1 μg/ml fibronectin (catalog no. F4759; Sigma). Before the cells were plated, the coverslips were washed once with 1× PBS and incubated for 1 h at 37°C, and the cells were fixed with 4% paraformaldehyde in PBS. The cells were permeabilized with 0.5% Triton in PBS for 5 min. The reagents used for fluorescence microscopy were antivinculin antibodies (catalog no. V9131; Sigma), anti-myc antibodies (clone 9E10), and Alexa Fluor 546 phalloidin (catalog no. A12380; Invitrogen). Images were processed in Adobe Photoshop CS3 for publication.
Pulldown assays.Rho GTPase activity was assessed using the Rho activation assay kit (catalog no. 14-383; Millipore) and the Rac1/Cdc42 activation assay kit (catalog no. 17-441; Millipore). GTPase protein levels were detected using antibodies provided with the respective activation kit. Cyanogen bromide-activated Sepharose 4B beads (catalog no. C9142; Sigma) were coupled to Sam68 C-terminal peptides (unphosphorylated, pY435, pY440, and pY443) (43) and used in pulldown assays with HEK293 cells transfected with Csk (15).
RESULTS
Sam68−/− MEFs display an accelerated spreading phenotype.In the majority of cell types, Sam68 is known to reside predominantly within the nucleus; thus, the link with membrane-bound proteins, such as Src, is thought to occur mainly during mitosis (44). However, Sam68 has been observed in the cytoplasm of specialized cells, such as neurons and spermatocytes (25, 56), but its role in the cytoplasm and its link to Src kinases remain undefined.
To assess the role of Sam68 in the early events of cell adhesion, we plated primary and immortalized mouse embryo fibroblasts from wild-type (Sam68+/+) and Sam68-deficient (Sam68−/−) 14.5-day-old embryos (60) on fibronectin-coated coverslips and monitored cell spreading by staining actin filaments with Alexa Fluor 546 phalloidin (Fig. 1A). The Sam68−/− MEFs attached and spread extensively on fibronectin and as early as 15 min, whereas wild-type Sam68+/+ MEFs required >60 min to attach and spread (Fig. 1A). The area that the Sam68−/− MEFs occupied was approximately 2.5 times larger than that of Sam68+/+ MEFs as assessed by MetaMorph software (data available upon request). In addition, cell shape was measured as the width divided by the length of the cell by MetaMorph, and the Sam68+/+ MEFs had values of ∼0.05 to 0.15, consistent with an elongated fibroblast, whereas the Sam68−/− MEFs had values of ∼0.05 to 0.35, consistent with a more oval shape (data available upon request). We also noted that the Sam68−/− MEFs were more difficult to detach from the tissue culture dishes with trypsin than the Sam68+/+ MEFs were (data not shown). These data demonstrate that Sam68−/− MEFs display an “accelerated spreading” phenotype. Since the primary Sam68−/− MEFs also displayed this phenotype (Fig. 1A), this indicated that the defect was intrinsic to the MEFs and not due to the immortalization.
Sam68−/− MEFs spread rapidly on fibronectin. (A) Comparative spreading of primary and immortalized Sam68−/− and Sam68+/+ MEFs. The cells were plated for 15 and 60 min on fibronectin-coated coverslips and stained with Alexa Fluor 546 phalloidin. Bar, 100 μm. (B) Sam68 regulation of cell spreading requires the C-terminal nine amino acids, including Y435 and Y440. Expression vectors encoding GFP-Sam68 and mutants were transfected in Sam68−/− MEFs, and their expression was verified by immunoblotting with anti-GFP (α-GFP) antibodies, and antiactin (α-Act) antibodies were used as a loading control. The positions (in kilodaltons) of molecular mass markers are shown to the left of the gel. The spreading efficiency of the transfected cells was recorded 60 min after seeding on fibronectin using Alexa Fluor 546 phalloidin. A cell was considered “spread” if the area of the phalloidin signal was greater than twice the DAPI signal. Over 400 cells from triplicate transfections were analyzed for each GFP fusion protein. The bars depict the means plus standard deviations (error bars) from three experiments. Values that were significantly different (P < 0.001) from the value for GFP-transfected cells are indicated by an asterisk. WT, wild type.
To rescue the accelerated spreading phenotype of the immortalized Sam68−/− MEFs, we transfected Sam68−/− MEFs with GFP expression plasmids encoding fusion proteins with wild-type Sam68 (GFP-Sam68), an RNA binding-defective Sam68 mutant lacking the KH domain (GFP-Sam68ΔKH), and a C-terminal deletion mutant lacking the last nine C-terminal amino acids, including tyrosines 435, 440, and 443 (GFP-Sam68Δ435-443) (12, 43). We expressed the spreading of Sam68−/− MEFs as the percent spread after 60 min on fibronectin. A cell was identified as spread if its area of phalloidin staining was more than two times the size of the nucleus visualized by DAPI staining. This value was 71% for untransfected Sam68−/− MEFs and 62% for GFP-transfected Sam68−/− MEFs (Fig. 1B) (n > 400). Approximately 15 to 19% of the Sam68−/− MEFs transfected with GFP-Sam68 or GFP-Sam68ΔKH were spread at 60 min (Fig. 1B), demonstrating that GFP-Sam68 and GFP-Sam68ΔKH “reversed” or rescued the accelerated spreading defect of the Sam68−/− MEFs. These data showed that the KH-type RNA binding domain did not contribute to the defect observed in Sam68−/− MEFs. In contrast, Sam68−/− MEFs transfected with GFP-Sam68Δ435-443 (58%) were not statistically different from untransfected and GFP-transfected Sam68−/− MEFs (Fig. 1B). These findings show that the last nine amino acids of Sam68 are necessary to contribute to the accelerated spreading phenotype defect observed in Sam68−/− MEFs. We next transfected Sam68−/− MEFs with GFP-Sam68 expression vectors harboring phenylalanine substitutions at individual tyrosines at positions 435, 440, and 443. Interestingly, Sam68−/− MEFs with GFP-Sam68Y435F (62%) and GFP-Sam68Y440F (61%) were not rescued, while Sam68−/− MEFs with Sam68Y443F protein (34%) were significantly rescued. These findings demonstrate that both Sam68 Y435 and Y440 are necessary to contribute to the accelerated spreading phenotype defect of Sam68−/− MEFs, whereas Y443 may weakly contribute. The facts that both GFP-Sam68ΔKH and GFP-Sam68Δ435-443 are known to localize in the cytoplasm (12, 43) and that GFP-Sam68Δ435-443 rescued and GFP-Sam68ΔKH did not rescue indicate that it is not the aberrant localization of the GFP-Sam68 mutant proteins that rescues the defect of Sam68−/− MEFs. The GFP-Sam68 proteins were expressed at equivalent levels by immunoblotting (Fig. 1B). The accelerated attachment and spreading phenotype on fibronectin was also observed in HeLa cells harboring a Sam68 shRNA, albeit with lower kinetics, further supporting the idea that Sam68 is involved in cell spreading and attachment. These findings show that the defects observed are not cell or species specific (data available upon request).
Sam68−/− MEFs have actin architecture, focal adhesion, and cell migration defects.We examined the actin cytoskeleton and the formation of focal adhesions in Sam68−/− MEFs plated on fibronectin by visualizing the actin filaments with Alexa Fluor 546 phalloidin and the focal adhesions with antivinculin antibodies. The Sam68−/− MEFs displayed a nonpolarized, rounded shape phenotype, unlike the wild-type MEFs, which have a polarized morphology with an elongated and polygonal shape (Fig. 2A, left panels). The actin architecture was disorganized in the Sam68−/− MEFs, with highly enriched actin filaments localized at the cell periphery, creating cortical actin ring structures and membrane ruffles. In contrast, the wild-type MEFs showed actin bundles organized in stress fibers along the axis of the cell. Sam68−/− MEFs contained many focal adhesions compared to the wild-type MEFs (Fig. 2A, right panels). Moreover, adhesions were distributed over the entire ventral surface, whereas in wild-type MEFs, the focal adhesions were mostly distributed at the cell periphery.
Sam68−/− MEFs have stress fiber architecture and focal adhesion defects. (A) Sam68+/+ and Sam68−/− MEFs were plated on fibronectin coverslips, and stress fibers and focal adhesions were visualized by phalloidin staining and the use of antivinculin antibodies followed by fluorescence microscopy. Bar, 50 μm. (B) Chemotactic migration assays performed on MEFs expressing endogenous levels of Sam68 (Sam68+/+ MEFs), low levels of Sam68 (Sam68 shRNA), and no Sam68 (Sam68−/− MEFs). The proportion of migrating cells was normalized to the number of migrating Sam68+/+ MEFs (counted in four random fields) that migrated to the lower chamber (n = 1,429). The levels of Sam68 and the protein arginine N-methyltransferase 1 (PRMT1) (loading control) expressed in the different cell lines used for the migration assays as visualized by SDS-PAGE and immunoblotting are shown in the gel insert (α-Sam68, anti-Sam68 antibody). (C) Wound-healing assays performed on Sam68+/+ and Sam68−/− MEFs plated on fibronectin-coated tissue culture dishes. The directional migration of the cells was visualized by DAPI staining and represented as percentage gap closed ± standard deviation (shown as a percentage also).
The accelerated spreading phenotype and the lack of cellular polarity observed suggested that Sam68−/− MEFs may have cell migration defects. Chemotactic migration was measured with a Boyden transwell assay using serum as the chemoattractant. As anticipated, therefore, the Sam68−/− MEFs were severely impaired in their ability to migrate compared to Sam68+/+ MEFs (Fig. 2B). Wild-type MEFs harboring a Sam68 shRNA (Sam68sh MEFs) displayed cell migration defects, but to a lesser extent than Sam68−/− MEFs as expected from a knockdown. Reduced chemotactic cell migration was also observed in HeLa cells expressing a Sam68 shRNA (data not shown). Sam68 was reintroduced in HeLa Sam68sh cells, and migration was rescued by the wild-type (GFP-Sam68) and GFP-Sam68ΔKH, but not by Sam68Δ435-443 (data available upon request), confirming the need for an intact Sam68 C terminus for chemotactic cellular migration.
A wound-healing assay was performed to further examine the migratory capabilities of the Sam68−/− MEFs. While the Sam68+/+ MEFs closed the wound in 16 h, migration of the Sam68−/− MEFs was significantly impaired (Fig. 2C), further supporting the idea that the directional migration of MEFs requires Sam68.
Sam68−/− MEFs are defective in activating RhoA.Defects in actin architecture and stress fiber formation in Sam68−/− fibroblasts suggested that alterations in the activity of the Rho GTPases had occurred (34). To verify this hypothesis, we performed Rhotekin-BD (binding domain) or PAK1-BD pulldown assays to determine the level of RhoA or Rac1 in the GTP-bound (active) conformation. Sam68−/− MEFs showed an approximately fivefold reduction in basal RhoA activity and an approximately fourfold increase in Rac1 activity (Fig. 3A). In addition, we performed a time course study of RhoA activity and observed a transient activation of RhoA in Sam68+/+ MEFs but not in Sam68−/− MEFs (data available upon request). To substantiate these findings, we reasoned that constitutively active RhoL63 would rescue the disorganized actin structure phenotype observed in Sam68−/− MEFs. An expression plasmid encoding myc-tagged RhoL63 was transfected in Sam68+/+ and Sam68−/− MEFs. The cells were prepared for immunofluorescence by the addition of anti-myc antibodies to visualize the transfected cells and the addition of Alexa Fluor 546 phalloidin to visualize the stress fibers. RhoL63 rescued the phenotype in Sam68−/− MEFs, restoring organized actin stress fiber bundles (Fig. 3B). Moreover, the addition of the Rac inhibitor (NSC23766 [23]) at 100 μM (data available upon request) or a dominant-negative form of Rac1 (Rac1 N17; data not shown) decreased spreading and lamellipodium formation of the Sam68−/− MEFs after 30 min, suggesting that the Sam68−/− MEF phenotype is mainly caused by elevated Rac1 activity.
Sam68−/− MEFs have altered activity of the Rho GTPases. (A) Cellular extracts prepared from Sam68+/+ and Sam68−/− MEFs were prepared, and Rac1 and RhoA pulldown assays were performed. The bound Rac1 and RhoA were visualized by SDS-PAGE, followed by immunoblotting with anti-Rac1 and RhoA antibodies. The amount bound to GTP was normalized to the total levels and normalized to 1.00. (B) Sam68−/− MEFs were transfected with dominant active RhoA, the transfected cells were visualized with anti-myc antibody (α-myc), and the stress fibers were visualized using Alexa Fluor 546 phalloidin. Bar, 50 μm. (C) The formation of LPA-induced stress fibers is defective in Sam68−/− MEFs. Sam68+/+ and Sam68−/− MEFs were treated for 10 min with 20 μM of LPA with or without 10 μM Y27632. The cells were stained with Alexa Fluor 546 phalloidin. Bar, 50 μm.
To investigate the role of Sam68 in de novo stress fiber formation, the cells were treated with lysophosphatidic acid (LPA), a known inducer of stress fiber formation through RhoA activation (63). Treatment of wild-type MEFs with LPA led to an increase in stress fibers which was prevented in the presence of the Rho-associated kinase (ROCK) inhibitor Y27632 (Fig. 3C). In contrast, LPA did not induce stress fiber formation in Sam68−/− MEFs, presumably because the cells are defective in RhoA activation. In Sam68−/− cells, stress fiber dynamics was, nevertheless, sensitive to treatment withY27632, suggesting that ROCK is still functional in these cells (Fig. 3C). Taken together, these findings show that Sam68 is required for stimulating RhoA activity and actin stress fiber organization and that Sam68 likely functions upstream of RhoA.
Sam68−/− MEFs harbor constitutive activation of p190RhoGAP and Src kinases.It is known that Src kinases can inhibit RhoA through the activation of p190RhoGAP (6). In Sam68+/+ MEFs, p190RhoGAP was tyrosine phosphorylated and associated with p120rasGAP at time zero and 5 and 15 min after plating cells on fibronectin as detected by immunoblotting with antiphosphotyrosine and anti-p120rasGAP antibodies (Fig. 4A). The p190RhoGAP tyrosine phosphorylation and association with p120rasGAP decreased after 30 to 60 min (Fig. 4A), suggesting that spreading and adhesion lead to Src kinase deactivation and activation of RhoA. In contrast, Sam68−/− MEFs constitutively maintained the tyrosine phosphorylation of p190RhoGAP and its association with p120rasGAP.
Sustained Src and p190RhoGAP activities in Sam68−/− MEFs. (A) Cells were plated for the indicated time on fibronectin-coated dishes, and cell extracts were prepared. p190RhoGAP was immunoprecipitated using anti-p190RhoGAP antibodies, and the bound proteins were separated by SDS-PAGE and visualized by immunoblotting with anti-p190RhoGAP(α-p190RhoGAP), antiphosphotyrosine (α-p-Tyr), and anti-p120rasGAP (α-p120rasGAP) antibodies as indicated. The ratio of associated p120 to total p190 was determined by densitometric scanning. The bars show means plus standard deviations (error bars). IP, immunoprecipitation. (B) Cells were plated for the indicated times on fibronectin, and the cells were lysed in lysis buffer. Src immunoprecipitations were performed, and the activity of Src was assessed by an in vitro kinase assay using enolase as an exogenous substrate and [γ-32P]ATP. The activity of Src was visualized by SDS-PAGE followed by autoradiography. To ensure equivalent levels of Src, immunoprecipitations were visualized by immunoblotting with anti-c-Src antibodies (α-Src). The activity of Src is expressed as the intensity of the 32P-labeled enolase/Src protein level by densitometric scanning.
Src was transiently activated in Sam68+/+ MEFs plated on fibronectin, and its activity increased slightly after 15 and 30 min (Fig. 4B), as assessed by in vitro kinase activity using [γ-32P]ATP and enolase as an exogenous substrate. However, Src activity was approximately threefold higher in Sam68−/− MEFs than in Sam68+/+ MEFs, and its activity was sustained after plating on fibronectin (Fig. 4B). These findings are consistent with our previous findings that mammary tumors induced with polyomavirus middle T antigen in Sam68+/− mice have activated Src compared to Sam68+/+ mice as visualized by using anti-Src Y416 antibodies (61). Treatment of Sam68−/− MEFs with 100 nM of the Src kinase inhibitor PP2 {4-amino-5-(4-chlorophenyl)-7-(t-butyl)pyrazolo[3,4-d]pyramidine} decreased the accelerated spreading phenotype, and the cells were more polarized (data available upon request). We also noted the sustained tyrosine phosphorylation of FAK and paxillin in Sam68sh-treated HeLa cells consistent with sustained Src activation (data available upon request). These findings suggest that Sam68 is required to regulate Src kinase activity in order to regulate RhoA activation and actin stress fiber organization.
Sam68 is localized near the plasma membrane during cell attachment.To fulfill a nonnuclear role, Sam68 has to relocalize or be localized in the cytoplasm. To investigate this possibility, we examined Sam68 localization by total internal reflection fluorescence microscopy. In traditional fluorescence microscopy, signals from the membrane molecules are dwarfed by the fluorescence background (26). In TIRF microscopy, only a thin slice (∼100 nm) near the glass-water interface is excited, and thus, proteins near the plasma membrane, and not all the proteins within the cell, are excited (26). We examined the locations of GFP-Sam68 and endogenous Sam68. First, Sam68−/− MEFs were transfected with GFP-Sam68 to facilitate the imaging. Twenty-four hours after transfection, the cells were detached from the tissue culture dish and plated on fibronectin, and the localization of GFP-Sam68 protein was visualized by TIRF microscopy. GFP-Sam68 was observed near the plasma membrane (Fig. 5A), since TIRF microscopy excites only fluorescent proteins within ∼100 nm of the plasma membrane. Next, endogenous Sam68 was visualized near the plasma membrane by TIRF microscopy (Fig. 5B), and predominantly in the nucleus by wide-field microscopy, as expected (Fig. 5B). The usage of a preimmune serum followed by the secondary antibody did not detect any signal, as expected (Fig. 5C). These findings confirm the presence of Sam68 near the plasma membrane during cell adhesion.
Sam68 is localized near the plasma membrane during cell attachment using TIRF illumination. (A) Sam68−/− MEFs were transfected with GFP-Sam68 for 24 h and subsequently plated on fibronectin for 1 h and visualized by differential interference contrast (DIC), wide-field (midlevel), and TIRF field microscopy. Bar, 50 μm. (B) Localization of endogenous Sam68 near the plasma membrane in wild-type MEFs. Sam68 was visualization by indirect immunofluorescence microscopy. (C) Preimmune rabbit serum was incubated with wild-type MEFs and visualization by indirect immunofluorescence. Bar, 100 μm.
Sam68 is rapidly tyrosine phosphorylated and associates with Csk after plating on fibronectin.Since Sam68 is a known substrate of Src, we examined whether Sam68 was tyrosine phosphorylated in wild-type MEFs after cell spreading on fibronectin. Tyrosine-phosphorylated Sam68 was observed after 5 and 15 min of cell attachment on fibronectin, as visualized by immunoblotting with the antiphosphotyrosine antibody and the anti-Sam68 pY440 (43) antibody (Fig. 6A).
Sam68 is tyrosine phosphorylated and associates with CSK after cellular adhesion. (A) Wild-type MEFs were plated for the indicated time on fibronectin, and cell extracts were prepared. Sam68 immunoprecipitations were performed, the bound proteins were separated by SDS-PAGE, and immunoblotting was performed with the antiphosphotyrosine antibody (α-p-Tyr), a pY440 Sam68-specific antibody (α-pY440), and total Sam68 using anti-Sam68 (α-Sam68) (AD1) antibody. (B) Sam68 peptides were coupled to cyanogen bromide Sepharose beads and utilized to perform Csk pulldown assays from HEK293 transfected with a Csk expression vector. IB, immunoblotting; α-Csk, anti-Csk antibody. (C) Wild-type MEFs were plated for 60 min on fibronectin and lysed in lysis buffer. The lysates were incubated with GST or GST-Csk-SH2 domain covalently coupled to Sepharose beads. The presence of Sam68 was visualized by immunoblotting with anti-Sam68 antibodies. (D) Wild-type MEFs were plated for the indicated time on fibronectin, and anti-Csk immunoprecipitations were performed. The coimmunoprecipitation of Sam68 was visualized by immunoblotting with anti-Sam68 (AD1) antibody. Cell extract from each time point was immunoblotted with anti-Csk antibody to demonstrate equivalent levels of Csk at each time point. For each panel, the positions (in kilodaltons) of molecular mass markers are shown to the left of the gels.
As Sam68 Y435 and Y440 contribute to the Sam68−/− MEF phenotype, we compared their surrounding sequences to SH2 domain binding consensus sequences to identify SH2 domain-containing proteins that may be docking on these phosphorylated tyrosines. We noted that the Csk and Crk SH2 domains were the only consensus sequences that weakly resembled the binding site surrounding Y440 (Y440GRY443; Y443 being the last residue of Sam68) with a small amino acid at position +1 and an arginine at position +2 (29, 74). As Csk regulates Src activity, we tested whether Sam68 peptides containing phosphorylated Y435, Y440, or Y443 associated with Csk. We performed pulldown assays with the Sam68 C-terminal unphosphorylated peptide KGAY435REHPY440GRY443 and peptides harboring phosphorylated Y435, Y440, or Y443. Sam68 peptides containing pY435 or pY440 associated with Csk, whereas the pY443 peptide weakly associated with Csk, as detected by immunoblotting (Fig. 6B). The unphosphorylated peptide did not associate with Csk (Fig. 6B), consistent with an interaction mediated by the Csk SH2 domain. Indeed, the GST-Csk SH2 domain could pull down Sam68 from cell lysates plated on fibronectin for 30 min (Fig. 6C).
Certain Src substrates are known to recruit Csk at the plasma membrane to inhibit subsequent Src activity (66). To determine whether this function applies to Sam68 and represents the mechanism by which Src kinase activity is elevated and sustained in Sam68−/− MEFs, we examined whether Sam68 and Csk associated in wild-type MEFs after cell adhesion. Indeed, Sam68 coimmunoprecipitated with Csk, and the interaction increased with time, even though there were equivalent Csk levels expressed in all lysates (Fig. 6C). These findings suggest that Sam68 functions to recruit Csk via its C-terminal tyrosines to modulate Src kinase activity during cell attachment.
DISCUSSION
We present evidence that Sam68 is required for cell polarization, morphology, and migration. We observed that Sam68−/− MEFs have an accelerated spreading phenotype with sustained Src kinase and p190 RhoGAP activation resulting in constitutively low levels of RhoA activity and high levels of Rac1 activity. Our results show that Sam68, a predominantly nuclear protein, is localized near the plasma membrane after cell attachment. Sam68 is rapidly tyrosine phosphorylated and associated with Csk after cell attachment. These findings define a signaling role for cytoplasmic Sam68 as an adaptor protein in the modulation of Src activation near the plasma membrane after integrin engagement.
The lack of cell polarity and the spreading and migration defects observed in Sam68−/− MEFs were rescued by wild-type and RNA binding-defective Sam68 proteins, but not by Sam68Δ435-443. These data show that the C-terminal nine amino acids of Sam68 and not RNA binding activity is required for the spreading defects of Sam68−/− MEFs. Within the C-terminal nine amino acids of Sam68, amino acid substitutions of Sam68Y435F and Sam68Y440F also failed to rescue, while the amino acid substitution of Y443F significantly rescued the spreading defect of Sam68−/− MEFs (Fig. 1B). These findings suggest that both Sam68 Y435 and Y440 are critical residues, while Y443 is dispensable for maintaining cell polarity and migration. Sam68 Y443 is the most C-terminal residue and does not contain C-terminal sequences required for SH2 domain docking and in fact had a low relative affinity for Csk (Fig. 6B). The fact that the single phosphorylation of Y435 or Y440 in the context of peptides efficiently associated with Csk in vitro (Fig. 6B) suggests that GFP-Sam68Y435F and Sam68Y440F should rescue the spreading defect of Sam68−/− MEFs, as either Y435 or Y440 associates with Csk (Fig. 6B). We suspect that both Sam68 435 and 440 tyrosine residues are required in addition to one phosphate for association.
We do not know whether Sam68 near the plasma membrane is generated by new protein translation or whether it is exported from the nucleus. However, at the plasma membrane, Sam68 may be recruited by Src itself. Subsequently, Sam68 tyrosine phosphorylated by Src then recruits Csk and negatively modulates Src activity. As Sam68 harbors 16 C-terminal tyrosines, it may also interact with other SH2 domain-containing proteins near the plasma membrane. Once Sam68 has fulfilled its function, what happens to the phosphorylated Sam68? The ability of the tyrosine-phosphorylated form of Sam68 to bind RNA is impaired (80), and the tyrosine phosphorylation at Y440 of Sam68 is required for nuclear import and retention (43). Thus, we speculate that the phosphorylated Sam68 fulfills its adaptor role at or near the plasma membrane and that phosphorylation of Sam68 at tyrosines including Y440 is required for its nuclear import. It remains to be determined whether Csk plays a role in escorting Sam68 to the nucleus or whether another phosphotyrosine binding protein fulfills this function.
The presence of RNA binding proteins near the plasma membrane has been observed by quantitative mass spectrometry; however, their roles were not defined. de Hoog and coworkers (18) differentially labeled cells that were attached and unattached to a matrix with light and heavy isotopes. They immunoprecipitated vinculin, paxillin, and talin complexes and identified several associated RNA binding proteins (18). They showed that the electroporation of antibodies against the RNA binding proteins hnRNP E1, hnRNPK, and FUS/TLS prevented cell migration. The results with hnRNPK have recently been corroborated by others, and another RNA binding protein, RasGAP SH3 domain binding protein (G3BP), was also shown to regulate cell migration (31, 58, 73, 84). It is interesting to note that hnRNPK and G3BP are substrates of Src kinases (54, 58). Thus, the link with Csk may be a general property of tyrosine-phosphorylated RNA binding proteins. Recently, mRNAs were identified at sites of membrane attachment (50). Thus, specific RNAs are present near the plasma membrane during cell attachment, suggesting that RNA binding proteins are likely to also participate in RNA regulation near the plasma membrane.
The presence of Sam68 at the plasma membrane was suggested by coimmunoprecipitation assays and cell fraction studies that show its association with plasma membrane receptors (33, 67); however, as Sam68 is an abundant protein, it was not clear from these studies whether there was some cytoplasmic Sam68 that associated with the receptors before cell lysis or whether the association occurred after lysis between the nuclear Sam68 and the plasma membrane receptor. In addition, by traditional fluorescence microscopy, Sam68 appears to be located exclusively in the nucleus in most cells (12). Our findings using TIRF microscopy show that Sam68 is indeed localized near the plasma membrane after fibronectin stimulation and corroborate a role for Sam68 at the plasma membrane.
The elevated and sustained Src activity observed in the Sam68−/− MEFs likely leads to the constitutive activation of FAK. Moreover, the tyrosine phosphorylation levels of paxillin and p190RhoGAP were also increased. The phosphorylation of p190RhoGAP has a direct effect on RhoA activity and increases the hydrolysis rate of GTP-bound RhoA, resulting in a smaller pool of activated RhoA (1, 6, 49), as observed in the Sam68−/− MEFs. RhoA inactivation is not always observed with Src activation, as v-Src-transformed cells maintain elevated levels of active RhoA (4, 69). Moreover, Src has been shown to regulate Rho GTPase activity by Rho GDP dissociation inhibitor (RhoGDI) phosphorylation and by decreasing the association of RhoGDI with Rho GTPases (19). Src kinases also link with Rac guanine exchange factors and thus contribute directly to increased Rac1 activity via Vav and Tiam1 (38, 49, 71). Interestingly, Sam68 is known to associate with Vav and may also provide a direct link to Rac activity (39).
Sam68-deficient cells showed an increase in actin filaments assembled into lamellipodium structures at the periphery of the cell and a decrease in stress fibers traversing the cells. Furthermore, they were unresponsive to LPA stress fiber induction (63). These phenotypes are consistent with the low levels of active RhoA and elevated levels of active Rac1. Plated on fibronectin, the Sam68−/− MEFs appear to be spreading in every direction, resulting in a loss of cell polarity and directional migration. The fact that the RNA binding domain of Sam68 was not required to rescue the cell polarity and migration defects observed in Sam68−/− MEFs suggested that its role in alternative splicing and RNA metabolism is not required. The expression of CD44 variants correlates with invasiveness of certain cancers (57), and CD44 is a known alternative splice target of Sam68 (13, 47). It could be argued that altered CD44 expression may explain the cell migration phenotype that we observed for Sam68−/− MEFs, but this is unlikely, as the role we propose for Sam68 herein is devoid of its KH domain, while the Sam68-dependent CD44 alternative splicing event requires binding to an RNA element within the V5 exon (47).
The mechanism by which Sam68 controls cell polarization and migration should shed some light on the phenotypes observed with the Sam68−/− mice. The Sam68−/− female mice have delayed mammary gland development (61) that may be caused by inadequate cell migration during the formation of mammary branching (21). Moreover, bone marrow-derived mesenchymal stem cell fate is impaired in Sam68−/− mice, leading to an increase in osteoblasts with aging (60). Interestingly, healthy human mesenchymal stem cells are modulated by endogenous RhoA activity and mesenchymal cells that adhere, flatten, and spread undergo osteogenesis, while unspread, round cells become adipocytes (48). Since Sam68−/− MEFs have an accelerated spreading phenotype, we speculate that Sam68−/− MEF-derived mesenchymal stem cells are favored to become osteoblasts because of altered cell spreading properties.
Some Src substrates are known to function in the recruitment of Csk; these include Cbp/PAG (phosphoprotein associated with glycosphingolipid-enriched microdomains) (5, 37), paxillin (66), caveolin-1 (10, 24), and p140Cap (20). Our findings that a nuclear RNA binding protein is recruited near the plasma membrane to recruit Csk and modulate Src activity raises the following questions. Why does Src require many substrates to modulate its activity? Why is a nuclear protein recruited near the plasma membrane and why an RNA binding protein? Although the answers will require further experimentation, in the case of the Sam68/Csk complex, this function seems to target Src during cellular adhesion. It was recently shown that Sam68 complexes with Cbp/PAG, Fyn, and RasGAP in stimulated T cells, suggesting that Cbp/PAG may be a recruitment factor for Sam68 in T cells to regulate Src family kinases via Csk (72). Csk increases Src activity, but this increased Src activity in mice is insufficient for cancer initiation (83). Similarly, Sam68−/− mice live a normal life span and are not prone to tumorigenesis (60). Although the Sam68−/− MEFs exhibit sustained Src activation, they did not display the cytoskeleton rearrangements normally observed in transformed cells (Fig. 1 and 2). Moreover, Sam68−/− MEFs did not display the formation of invasive adhesions or podosomes (75), demonstrating that the loss of Sam68 and immortalization (3T3 protocol) is insufficient for transformation.
Our findings suggest that the tyrosine phosphorylation of Sam68 may regulate cell mobility and hence cancer metastasis. Indeed, the Sam68-deficient cells harboring the polyomavirus middle T antigen had reduced lung metastases in nude mice (61). Thus, inhibitors that block Sam68 relocalization near the plasma membrane or disrupt Sam68 tyrosine phosphorylation may be useful to prevent cell metastasis and invasiveness. Actually blocking antibodies to hnRNPK were identified in a screen as metastasis inhibitors (31).
In conclusion, we have shown that Sam68 links with Src during cell polarization and cell migration. Although Sam68 resides predominantly in the nucleus, it localizes near the plasma membrane after cell adhesion and becomes tyrosine phosphorylated and associates with Csk to modulate Src kinase activity. These findings define a new role for Sam68 near the plasma membrane as an adaptor protein to recruit Csk to modulate Src activity.
ACKNOWLEDGMENTS
We thank Dominique Davidson and Josée Lavoie for reagents and helpful discussions. Images for Fig. 5 were collected on a TIRF microscope at the McGill Life Sciences Complex Imaging Facility.
This work was supported by MT13377 from the Canadian Institutes of Health Research to S.R. M.-E.H. is supported by fellowships from the Fonds de la Recherche en Santé du Québec (FRSQ) and the Cancer Research Society Inc. N.L.-V. is an FRSQ Chercheur-Boursier Senior, and S.R. is an FRSQ Chercheur-National.
FOOTNOTES
- Received 5 November 2008.
- Returned for modification 8 December 2008.
- Accepted 5 January 2009.
- Accepted manuscript posted online 12 January 2009.
- Copyright © 2009 American Society for Microbiology
REFERENCES
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.↵
- 7.↵
- 8.↵
- 9.↵
- 10.↵
- 11.↵
- 12.↵
- 13.↵
- 14.↵
- 15.↵
- 16.↵
- 17.↵
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.↵
- 35.↵
- 36.↵
- 37.↵
- 38.↵
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.↵
- 45.↵
- 46.↵
- 47.↵
- 48.↵
- 49.↵
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.↵
- 55.↵
- 56.↵
- 57.↵
- 58.↵
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.↵
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.↵
- 77.↵
- 78.↵
- 79.↵
- 80.↵
- 81.↵
- 82.↵
- 83.↵
- 84.↵