IntroductionThe hydrolysis of membrane phospholipids is one of theearliest cellular responses to many extracellular stimuli.Phospholipases C (PLCs), a crucial group of enzymes in signaltransduction, catalyses the hydrolysis of phosphatidylinositol-4,5-bisphosphate to inositol-1,4,5-trisphosphate anddiacylglycerol, two molecules that regulate the mobilization ofintracellular Ca2+ and protein-kinase-C activation, respectively(Berridge and Irvine, 1984). Phospholipase Cγ1 (PLC-γ1) is amember of the phosphoinositide-specific PLC family and isrecognized for its role in growth-factor-dependent signaltransduction (Carpenter and Ji, 1999).

All PLC isozymes contain X and Y domains that foldtogether to form the catalytic site. The unique feature of thePLC-γ subfamily, which includes two isozymes, is the presenceof a large linker between the X and Y domains that containstwo SH2 domains, one SH3 domain and two sites of tyrosinephosphorylation (Carpenter and Ji, 1999). The SH2 domainsmediate the association of PLC-γ with tyrosine-phosphorylatedproteins, especially activated growth-factor-receptor tyrosinekinases. Nearly all growth factors induce the tyrosinephosphorylation and activation of PLC-γ1 and therebystimulate phosphatidylinositol-4,5-bisphosphate turnover.SH2-domain-mediated binding of PLC-γ1 to activated

receptors leads to the tyrosine phosphorylation of PLC-γ1at three tyrosine residues: Tyr771, Tyr783 and Tyr1254.Biochemical studies have revealed that tyrosinephosphorylation activates PLC-γ1 enzyme activity (Nishibe etal., 1990) and site-directed mutagenesis (Kim et al., 1991)implicates Tyr783 as essential for PLC-γ1 activation by growthfactors.

Integrins are cell-surface receptors that mediate interactionsbetween cells and the extracellular matrix (ECM) and havea crucial role in cellular functions such as cell adhesion,migration, proliferation and apoptosis. Integrins areheterodimeric molecules comprising an α and a β subunit,which combine in a restricted manner to form various dimers,each of which exhibits different ligand-binding properties(Hynes, 2002; Plow et al., 2000). The extracellular domainsof integrin subunits are large and constitute the ligand-bindingdomain(s) of these receptors, whereas the short cytoplasmictails play a crucial role in the promotion of cell anchorage.The cytoplasmic domains interact with cytoskeletal proteinsto facilitate the connection of integrins to the cytoskeleton(van der Flier and Sonnenberg, 2001). In addition, thecytoplasmic domains also recruit cytoplasmic tyrosine kinasessuch as focal-adhesion kinase (FAK), integrin-linked kinaseand Src-family kinases, to propagate signals from the engaged

601

Although integrin engagement initiates signaling eventssuch as focal-adhesion kinase (FAK) and Src kinaseactivation, the role of phosphoinositide turnover in celladhesion is less clear. To assess PLC-γ1 function in thisprocess, Plcg1–/– fibroblasts (Null) were compared with thesame fibroblasts in which PLC-γ1 was re-expressed (Null+).Following plating on fibronectin, Null cells displayed asignificantly impaired rate of adhesion compared withNull+ cells. This defect was detected at low concentrationsof fibronectin; at high fibronectin concentrations, theNull and Null+ cells displayed equivalent adhesioncharacteristics. The differences were not due to PLC-γ1-dependent changes in integrin subunit expression, nor wasintegrin receptor clustering impaired with the absence ofPLC-γ1. Experiments with site-specific antibodies andPLC-γ1 mutants showed that fibronectin selectively

increased phosphorylation of Tyr783 and that mutagenesisof this residue, but not Tyr771 or Tyr1253, abrogatedfibronectin-dependent adhesion. The SH2 domains of PLC-γ1 were also required for maximal adhesion on fibronectin.Adhesion to fibronectin induced PLC-γ1 tyrosinephosphorylation that was inhibited by a Src-kinaseinhibitor, but not an epidermal-growth-factor-receptorkinase inhibitor. Moreover, in cells null for Src familymembers, but not in cells null for FAK family members,integrin-dependent PLC-γ1 tyrosine phosphorylation wasgreatly reduced. Finally, the data demonstrated that PLC-γ1 co-immunoprecipitated with Src following fibronectin-induced integrin activation, and this association did notdepend on FAK expression.

Key words: PLC-γ1, Fibronectin, Adhesion, Migration

Summary

Integrin-dependent PLC-γ1 phosphorylation mediatesfibronectin-dependent adhesionDenis Tvorogov1,*,‡, Xue-Jie Wang1,*,§, Roy Zent2,3 and Graham Carpenter1,¶

1Department of Biochemistry, Vanderbilt University School of Medicine, Nashville, TN 37232, USA2Departments of Medicine and Cancer Biology, Vanderbilt University Medical Center, Nashville, TN 37232, USA3Veterans Affairs Hospital, Nashville, TN 37232, USA*These authors contributed equally to this work‡Present address: MediCity Research Laboratory Tykistokatu 6A, Turku, Finland§Present address: Department of Technology Development, Biology, Discover Research, GlaxoSmithKline, 5 Moore Drive, Research Triangle Park, NC 27709, USA¶Author for correspondence (e-mail: graham.carpenter@vanderbilt.edu)

Accepted 11 November 2004Journal of Cell Science 118, 601-610 Published by The Company of Biologists 2005doi:10.1242/jcs.01643

Research Article

Jour

nal o

f Cel

l Sci

ence

602

integrins to the inside of the cell (van der Flier andSonnenberg, 2001).

Integrin-dependent cell adhesion to ECM results in integrinclustering and the formation of adhesion complexes, whichprovide an intersection where mechanical forces, cytoskeletalorganization, biochemical signals and adhesion meet (Juliano,2002). Several groups have previously demonstrated that PLC-γ1 is present in adhesion complexes (Plopper et al., 1995;Miyamoto et al., 1995; Cybulsky et al., 1996). Integrin ligationby ligand (Clark and Brugge, 1995) or antibody (Kanner et al.,1993) activates phospholipase C leading to IP3- dependent Ca2+

mobilization, suggesting that PLC isozymes have a functionalrole in integrin-dependent function. Recently, it has beendemonstrated that PLC-γ1 associates with α1β1-integrincytoplasmic domains in an adhesion- and time-dependentmanner, and tyrosine phosphorylation of PLC-γ1 is notrequired for this association to occur (Vossmeyer et al., 2002).Furthermore, Vossmeyer et al. demonstrate that inhibition ofPLC activation by a chemical inhibitor leads to reducedintegrin α1β1-integrin-dependent cell adhesion to collagen.Antibody cross-linking of β2-integrins has been shown to elicittyrosine phosphorylation of PLC-γ1 (Kanner et al., 1993).

In this report, PLC-γ1-deficient cells are shown to have anadhesion, spreading and migration defect. In addition, data arepresented to show that integrin engagement by fibronectininduces tyrosine phosphorylation of PLC-γ1 at Tyr783 and thatthis signaling event requires Src activity. Mutagenesis ofTyr783 abrogates the capacity of PLC-γ1 to facilitate adhesion.

Materials and MethodsMaterialsFetal calf serum (FCS), Dulbecco’s modified Eagle’s medium(DMEM) containing L-glutamine and high levels of glucose werepurchased from Life Technologies. Trypsin inhibitor solution waspurchased from Sigma. Plcg1–/– fibroblasts (Null cells) and the samefibroblasts in which PLC-γ1 was re-expressed (Null+ cells) have beendescribed previously (Ji et al., 1998; Tvorogov and Carpenter, 2002).Cells lacking Src, Yes and Fyn expression (SYF) were obtained fromATCC and are described elsewhere (Klinghoffer et al., 1999). TFW-46 cells, representing mouse fibroblasts with tetracycline-inducibleFAK expression in a FAK nullizygous background, have beendescribed earlier (Owen et al., 1999). These cells were grown in thepresence of tetracycline (Tet+) to keep them in the uninduced state.FAK expression was induced by replacing the growth medium withfresh medium lacking tetracycline. The cells were used for analysis 2days after tetracycline withdrawal.

A monoclonal antibody against phosphotyrosine (PY20) and amonoclonal antibody to the N-terminus of PLC-γ1 were purchasedfrom Transduction Laboratories. Monoclonal antibodies to Src kinaseand polyclonal antibodies to FAK were obtained from Oncogene,polyclonal antibodies to phosphorylated PLC-γ1 (Tyr783) from SantaCruz Biotechnology. Antibodies againts phosphorylated Y771, Y783and Y1253 in the PLC-γ1 sequence were a gift from S. G. Rhee(National Institutes of Health, Bethesda, MD) (Sekiya et al., 2004).Rabbit antiserum against PLC-γ1 has been described previously(Arteaga et al., 1991). Antibodies against integrins α1, α2, α5, α6,αV, β1 and β4 were from BD Biosciences. Rat antibody againstintegrin β1 (used for immunofluorescence) and the antibody againstα3 integrin were purchased from Chemicon. Horseradish peroxidase(HRP)-conjugated protein A was from Zymed Laboratories andsecondary goat anti-mouse IgG antibody was from TransductionLaboratories. Vaccinia virus containing wild-type and site-specificmutants (Y771F, Y783F, Y1253F) of PLC-γ1 (Kim et al., 1991) was

obtained from S. G. Rhee. Null cells expressing PLC-γ1 N- and C-terminal SH2 domain loss-of-function mutations have been describedelsewhere (Ji et al., 1999).

Immunoprecipitation and immunoblottingFor ligand coating, cell-culture dishes were incubated overnight with10 µg ml–1 fibronectin, unless otherwise indicated, in PBS at 4°C.Nonspecific binding sites were blocked with heat-denatured 1%bovine serum albumin (BSA) in PBS for 1 hour at 37°C. Subconfluentcells were incubated for 12 hours in DMEM containing 0.5% FCS.Subsequently, the cells were trypsinized, washed with DMEMcontaining 0.5 mg ml–1 trypsin inhibitor and then washed withDMEM. The cells were then kept in suspension for 30 minutes orallowed to adhere to fibronectin-coated dishes for the indicatedperiods of time. Subsequently the cells were treated with lysis buffer(1% Nonidet P-40, 20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 5 mMEDTA, 2 mM Na3VO4, 100 µM phenylmethylsulfonyl fluoride and 10µg ml–1 each aprotinin and leupeptin) for 2 minutes, scraped and theinsoluble material was removed by centrifugation (15,000 g for 2minutes) at 4°C.

For immunoprecipitation, cells lysates were mixed with antibodies(generally 1 µg) for 2hours followed by the addition of 40 µl50% protein-A/Sepharose or protein-G/Sepharose beads (ZymedLaboratories) for an additional 2 hours at 4°C. Theimmunoprecipitates were then washed three times with lysis bufferand boiled in 40 µl 2× Laemmli’s buffer. Subsequently, the sampleswere subjected to 7.5% SDS-PAGE and electrotransferred topolyvinylidene difluoride membranes (Millipore).

The membranes were probed with the indicated primary antibodies,washed in Tris buffer saline (TTBS) (20 mM Tris-HCl pH 7.4, 150mM NaCl, 0.1% Tween 20) followed by HRP-conjugated secondaryantibodies or HRP-conjugated protein A. Membranes were washedand visualized by an enhanced chemiluminescence detection system(Amersham Pharmacia Biotech). When necessary, membranes werestripped by incubation in stripping buffer (62.5 mM Tris-HCl, pH 5.7,100 mM 2-mercaptoethanol, 2% SDS) for 1 hour at 70°C withconstant agitation, washed and then reprobed with other antibodies asindicated.

Adhesion assays24-well plates were coated with fibronectin overnight at 4°C.Nonspecific binding sites were blocked with heat-denatured 1% BSAin PBS for 1 hour at 37°C. The cells were trypsinized, washed withDMEM containing 0.5 mg ml–1 trypsin inhibitor and then washed withDMEM. Then, the cells were counted (using a Coulter Counter) and105 cells were added to each well and allowed to attach for theindicated times at 37°C. The cells were subsequently washed withPBS and the attached cells were fixed with 3.7% formaldehyde in PBSfor 1 hour and stained with crystal violet. Crystal violet was elutedwith 10% acetic acid. Cell adhesion was calculated by measuring theabsorbance of eluted dye at 595 nm with an enzyme-linkedimmunosorbent assay (ELISA) reader. The result was calculatedas an optical density (OD) percentage ratio [(ODtest – ODblank)/(ODpositive control – ODblank)]×100, where ODtest was the measured ODof cells that adhered to ligand, ODblank was the OD of cells that wereallowed to adhere to BSA-blocked wells and ODpositive control was theoptical density of cells that were allowed to adhere in the presence of10% FCS.

Migration assaysPolycarbonate transwells (6.5 mm diameter, 8 µm pore diameter)were coated on the underside with 10 µg ml–1 fibronectin overnightat 4°C. Nonspecific binding sites were blocked with heat-denatured1% BSA in PBS for 1 hour at 37°C. Cells were then trypsinized,

Journal of Cell Science 118 (3)

Jour

nal o

f Cel

l Sci

ence

603PLC-γ1 and integrin-dependent cell adhesion

washed with DMEM containing 0.5 mg ml–1 trypsininhibitor and washed with DMEM. The cells werecounted and 5×104 cells were added to eachtranswell and allowed to attach and migrate for 3hours at 37°C. Afterwards, the top of each chamberwas cleaned with a cotton swab to remove all cells.The remaining cells were fixed and stained withcrystal violet as described above and nine randomlychosen fields from triplicate wells were counted at200× magnification.

Fluorescence imagingTo visualize actin-filament distribution, trypsinizedcells were plated on fibronectin-coated glasscoverslips for different times. After the indicatedtime, adherent cells were briefly washed in PBS andfixed with 3.7% formaldehyde in PBS for 15minutes room temperature. Actin distribution wasvisualized by staining the cells with tetramethylrhodamine isothiocyanate (TRITC)-phalloidin for30 minutes at room temperature followed bywashing with PBS. For integrin-β1 staining, cellswere fixed, permeabilized in 0.1% Triton X-100 inPBS and blocked with 10% goat preimmune serum.After that, cells were incubated with primary anti-integrin-β1 antibody, washed and incubated withgoat anti-rat antibody conjugated to TRITC. Glasscoverslips were analysed on a Zeiss Axiovert 135 confocalmicroscope.

Flow cytometryCells were disassociated from tissue-culture dishes and exposed todiluted monoclonal antibodies of the appropriate integrin, followed byappropriate secondary antibodies (fluorescein-isothiocyanate-coupledrabbit anti-rat or anti-hamster immunoglobulin). Flow cytometry wasperformed with a FACScan instrument (Becton Dickinson).

ResultsNull and Null+ cells display different adhesion andmigration characteristicsTo determine the role of PLC-γ1 in cell adhesion andmigration, the previously characterized Plcg1 nullizygousmouse embryonic fibroblasts (Null cells), and Null cells inwhich PLC-γ1 has been re-expressed (Null+ cells) were used.The integrin expression profiles of the cells were determinedto identify the substrates to which these cells would adhere aswell as to detect whether there were any changes in integrinexpression caused by the restoration of PLC-γ1 (Fig. 1).Because there is no α3 subunit antibody suitable forfluorescence-activated cell sorting, integrin α3β1 expressionwas assessed by blotting with an integrin α3 antibody (data notshown). Both Null and Null+ cells express the α3β1 integrinsand there are no differences in expression levels between thecell lines. Therefore, it can be concluded that the predominantintegrins expressed by Null and Null+ cells are α5β1 andα3β1, and there are no significant difference in the levels ofexpression between the two cell lines.

Based on these integrin expression profiles, fibronectin (awell characterized ligand for integrin α5β1) was used toexamine the differences in cell adhesion between the Null andNull+ cell lines. The two cell types were plated on increasing

concentrations of fibronectin and cell adhesion was measuredafter 30 minutes. As shown in Fig. 2A, at low concentrationsof fibronectin (1-5 µg ml–1), adhesion of the Null cells issignificantly diminished relative to Null+ cells; however, athigher concentrations (5-20 µg ml–1), the adhesion of both celllines to fibronectin is equivalent. Similar results were obtainedwhen fibrinogen was used as the extracellular matrix (data notshown), although the cell adhesion for both cell lines wassignificantly less than on fibronectin. The data in Fig. 2Bdemonstrate that the observed differences in adhesion betweenNull and Null+ cells at low fibronectin concentrations isreflected in a reduced rate of adhesion. At high fibronectinconcentrations, the rate of adhesion is equivalent between thetwo cell types.

Because the Null cells adhered less than did the Null+ cellsat low concentrations of fibronectin, we investigated whetherthere were differences in cell spreading and alterations in thecell morphology at low fibronectin concentrations. The cellswere plated onto fibronection-coated coverslips for 60 minutesor 120 minutes and than stained with TRITC-phalloidin. At60 minutes, the Null cells had a more rounded morphologycompared with the elongated shape of the spread Null+ cells.By 120 minutes, the Null+ cells had developed lamellipodiaat their leading edges. At this time point, the Null cells werespread and had a polygonal shape but there was no evidenceof lamellipodia (Fig. 3A), suggesting that PLC-γ1 mightinfluence cell polarity and/or cytoskeleton dynamics.

Based on the decreased cell adhesion and spreadingobserved in the Null cells relative to Null+ cells, an assessmentwas made of differences in cell migration using transwellsprecoated with increasing concentrations of fibronectin. Theresults (Fig. 3B) demonstrate that Null cells migratesignificantly less than Null+ cells at all fibronectinconcentrations, including those concentrations at whichdifferences in adhesion were not observed.

Fig. 1. Expression ofdifferent integrinsubunits in Null andNull+ cell lines. Nulland Null+ cells weresubjected to FACSanalysis for differentintegrin subunits.

Jour

nal o

f Cel

l Sci

ence

604

That Null cells are deficient in adhesion in a fibronectin-concentration-dependent manner might suggest that PLC-γ1influences integrin clustering. To test this, Null and Null+ cellswere allowed to adhere to low and high concentrations offibronectin for various times and then stained with an integrin-β1 antibody. As seen in Fig. 4A, although there are somedifferences in cell morphology, the pattern of integrin stainingwas similar in Null and Null+ cells plated on a high level (10µg ml–1) of fibronectin. This pattern was also present at low

levels (2.5 µg ml–1) of fibronectin,which produces clear differences inadhesion between the two cell lines.In addition, Null and Null+ cellswere plated on fibronectin for 0minutes, 30 minutes or 60 minutesand blotted with antibodiesagainst phosphorylated FAK orphosphorylated extracellular-signal-regulated kinase (ERK) to determinewhether there were differences inintegrin-dependent FAK and ERKactivation following cell adhesion.No difference was observed in theactivation of either of these kinases(Fig. 4B,C).

Cell adhesion to fibronectininduces PLC-γ1 tyrosinephosphorylationPLC-γ1 activation caused by tyrosinephosphorylation is thought tomediate cellular signaling inducedby growth factors (Carpenter and

Ji, 1999). To determine whether PLC-γ1 tyrosinephosphorylation occurred following integrin clustering, Null+cells were allowed to adhere to fibronectin for 15 minutes and30 minutes, and then the cells were lysed and PLC-γ1 wasimmunoprecipitated and blotted with site-specific antibodiesagainst phosphorylated PLC-γ1. As shown in Fig. 5A,fibronectin induces tyrosine phosphorylation of PLC-γ1 in theNull+ cells, particularly at Tyr783 and to a lesser extent atTyr1254. At Tyr771, there was a low level of phosphorylation

Journal of Cell Science 118 (3)

Fig. 2. PLC-γ1 is required formaximal adhesion to fibronectin.(A) Null (white) and Null+ (gray)cells were allowed to adhere for30 minutes on plates coated withthe indicated concentrations offibronectin. (B) Null (�) andNull+ (�) cells were incubatedfor the indicated times with platescoated with the indicatedconcentrations of fibronectin.

Fig. 3. Influence of PLC-γ1 on spreading andmigration behavior. (A) Null and Null+ cells wereallowed to adhere to cover glasses coated with theindicated concentration of fibronectin for theindicated times. Actin distribution was thenvisualized by staining with phalloidin-TRITC.(B) Null (white) and Null+ (gray) cells wereapplied to fibronectin-coated transwell chambersand allowed to migrate for 3 hours. Subsequently,the chambers were processed to remove all cellsfrom the top and the remaining cells were fixed andstained with crystal violet and counted as migratingcells.

Jour

nal o

f Cel

l Sci

ence

605PLC-γ1 and integrin-dependent cell adhesion

Fig. 4. Influence of PLC-γ1 on integrin β1 clustering, FAK phosphorylation and ERK1,2 phosphorylation following adhesion to fibronectin.(A) Equal numbers of Null and Null+ cells were allowed to adhere for the indicated times to cover glasses coated with 10 µg ml–1 or 2.5 µgml–1 fibronectin. Integrin β1 distribution was visualized by staining with anti-integrin-β1 antibody. (B) Equal numbers of Null and Null+ cellswere allowed to adhere to 10 µg ml–1 fibronectin-coated dish for the indicated times. The cells were then lysed and each lysate was blotted withantibodies against phosphorylated or unphosphorylated FAK. (C) Equal numbers of Null and Null+ cells were allowed to adhere to 10 µg ml–1

fibronectin-coated dish for the indicated times. Then cells were lysed and each lysate was blotted with antibodies phosphorylated orunphosphorylated Erk1,2. Cells maintained in suspension (Susp) were used as a control.

Fig. 5. Fibronectin-stimulated tyrosinephosphorylation of PLC-γ1. (A) Null+cells were allowed to adhere tofibronectin-coated dishes (10 µg ml–1)in serum-free medium for the indicatedtimes. The cells were then lysed andeach lysate was precipitated with anti-PLC-γ1 antibody followed by blottingwith site-specific antibodies againsttyrosine-phosphorylated-PLC-γ1 orPLC-γ1 as indicated. (B) Null+ cellsexpressing wild-type PLC-γ1 or anN+C-SH2-domain loss-of-functionPLC-γ1 mutant (N+C SH2–) (Ji et al.,1999) were allowed to adhere to

fibronectin (10 µg ml–1) for the indicated times. The cells were than lysed and the lysate wereprecipitated with anti-PLC-γ1 antibody and then blotted with antibody against tyrosine-783-phosphorylated PLC-γ1 or PLC-γ1 as indicated. Cells maintained in suspension (Susp) wereused as a control for time 0.

Jour

nal o

f Cel

l Sci

ence

606

that was not altered by the presence offibronectin.

The fibronectin-dependent tyrosinephosphorylation of PLC-γ1 predicts that loss ofthe SH2 domains of PLC-γ1, which facilitaterequisite delivery to tyrosine kinases, shouldabrogate tyrosine phosphorylation under theseconditions. This was tested and it was found (Fig.5B) that fibronectin fails to induce PLC-γ1tyrosine phosphorylation in a PLC-γ1 mutant inwhich both SH2 domains are disabled by pointmutation of a crucial lysine residue in each SH2domain (Ji et al., 1999).

The data in Fig. 5 suggest that phosphorylation of PLC-γ1at Tyr783 and Tyr1253 might be significant for adhesion tofibronectin. This was tested by expressing tyrosine mutants(Y771F, Y783F, Y1253F) or wild-type PLC-γ1 in Null cellsand measuring adhesion. The results in Fig. 6A show that anequivalent level of each PLC-γ1 construct was expressed anddemonstrate that only Tyr783 is crucial for adhesion. Althoughthe Y771F and Y1254F mutants were equivalent to wild-typePLC-γ1 in this adhesion assay, the Y783F mutant showed thesame level of deficiency as the complete absence of PLC-γ1 inthe Null cells. Similarly, the loss of PLC-γ1 SH2 domainsabrogates the capacity of PLC-γ1 to mediate fibronectin-dependent adhesion (Fig. 6B).

In an attempt to identify the tyrosine kinase that inducesPLC-γ1 tyrosine phosphorylation under these conditions, themost likely candidates were screened. It is has been shown thatthe engagement of integrins can lead to activation of theepidermal growth factor (EGF) receptor (Moro et al., 2002) andwe have observed that A431 cell adhesion to fibronectin caninduce EGF-receptor tyrosine phosphorylation (data notshown). To assess whether the kinase activity of the EGFreceptor was required for PLC-γ1 tyrosine phosphorylation,Null+ cells were pretreated with the EGF-receptor tyrosinekinase inhibitor AG1478 before adhesion to fibronectin. Asshown in Fig. 7A, this inhibitor had no effect on fibronectin-induced PLC-γ1 tyrosine phosphorylation but completelyblocked EGF-induced tyrosine phosphorylation of PLC-γ1.

Because the activation of Src kinase in response to integrinengagement is also well established, we used the selective Src

kinase inhibitor PP2 to test the role of Src kinases in PLC-γ1phosphorylation. As shown in Fig. 7B, the presence of PP2significantly reduces PLC-γ1 tyrosine phosphorylationfollowing fibronectin-induced integrin activation. These resultsindicate that one or more of the Src kinase family membersmediate PLC-γ1 phosphorylation.

To investigate further whether Src kinase family membersare required for the integrin-dependent phosphorylation ofPLC-γ1, a triple-knockout fibroblast cell line lacking Src, Yesand Fyn expression (SYF–/–) was tested. When these cellsadhered to fibronectin, PLC-γ1 phosphorylation is significantlyreduced compared with Null+ cells (Fig. 8A), although PLC-γ1 tyrosine phosphorylation in both these cell types isequivalent following stimulation by EGF (data not shown). Wehave also assessed whether FAK is required for PLC-γ1phosphorylation during fibronectin-dependent adhesion. Thedata in Fig. 8B show that the absence of FAK (FAK–/– cells)does not inhibit PLC-γ1 phosphorylation.

PLC-γ1 associates with Src kinase in the absence ofFAKTo determine the mechanism whereby Src modulates PLC-γ1phosphorylation, we assessed whether these two proteinsassociate with each other. For these experiments, Null+ cellswere allowed to adhere to fibronectin, following which Src wasimmunoprecipitated and subsequently blotted with PLC-γ1antibodies. Because it is known that FAK is a substrate for Src,the Src immunoprecipitates were also tested for the presence

Journal of Cell Science 118 (3)

Fig. 6. Influence of PLC-γ1 mutants on adhesion tofibronectin. (A) Null cells were infected for 8 hourswith wild-type vaccinia virus or virus containing pointmutations Y771F, Y783F or Y1253F, or wild-typePLC-γ1, as described elsewhere (Sekiya et al., 2004).Null+ cells were used as control. Equal numbers ofcells from each cell line were allowed to adhere onfibronectin-coated plates for 30 minutes at 37°C. �,Null cells with control virus; �, Null cells with wild-type PLC-γ1 virus; �, Null cells with Y771F virus;�, Null cells with Y783F virus; �, Null cells withY1253F virus. (B) Null cells or Null cells expressingeither wild-type PLC-γ1 or a PLC-γ1 mutant thatcontains loss of function mutations in the N- and C-terminal SH2 domains [N+C SH2– (Ji et al., 1999)]were allowed to adhere to fibronectin-coated platesfor 30 minutes at 37°C. �, Null cells; �, Null+ cellsexpressing wild-type PLC-γ1; �, Null+ cellsexpressing N+C SH2– PLC-γ1.

Jour

nal o

f Cel

l Sci

ence

607PLC-γ1 and integrin-dependent cell adhesion

of FAK. As shown in Fig. 9A, both PLC-γ1 and FAK aredetected in Src immunoprecipitates from cells that adhere tofibronectin for 30 minutes. It seems likely that the capacity ofSrc to associate with and phosphorylate PLC-γ1 will dependon presence of functional SH2 domains in PLC-γ1. The datain Fig. 9B show that, when a PLC-γ1 mutant with loss-of-function mutation in each SH2 domain was tested, there wasno association of PLC-γ1 with Src following fibronectin-mediated adhesion. The SH2 domains of PLC-γ1 are probablyrequired to facilitate association with a Src phosphotyrosineresidue as a prelude to tyrosine phosphorylation of PLC-γ1.

The association of PLC-γ1 with Src raised the question asto whether this interaction is mediated by FAK. It has beenreported that the Src-FAK interaction is one of the first and

most crucial steps in integrin signaling (Thomas et al., 1998;Hanks et al., 2003) and PLC-γ1 associates with FAK afterintegrin engagement (Zhang et al., 1999). The data in Fig. 9Cshow that, when Null+ cells were plated on fibronectin,increasing amounts of PLC-γ1 immunoprecipitated with FAKover time. To test whether the PLC-γ1 interaction with Srcinvolves FAK, a FAK–/– mouse embryonic fibroblast cell linewith a tetracycline-inducible FAK construct was used (Owenet al., 1999). These cells, designated TFW-46 cells, were grownin the presence of tetracycline (Tet+) to keep them in theuninduced state. FAK expression was then induced byreplacing the growth medium with fresh medium lackingtetracycline (Fig. 10A). As shown in Fig. 10B, PLC-γ1immunoprecipitated with Src kinase in the presence or absence

Fig. 7. Src-kinase activity is necessary for fibronectin-induced PLC-γ1 phosphorylation. Null+ cells were allowed to adhere to fibronectin-coated dishes (10 µg ml–1) in serum-free medium containing 50 µM sodium pervanadate and in the presence of (A) EGF-receptor tyrosine-kinase inhibitor AG1478 or (B) Src-kinase inhibitor PP2. Then, the cells were lysed and each lysate was precipitated with anti-PLC-γ1antibodies followed by blotting with antibodies against tyrosine-783-phosphorylated PLC-γ1 or PLC-γ1 as indicated. In each experiment, 25 ngml–1 EGF was added for comparative purposes.

Fig. 8. Fibronectin-dependent PLC-γ1 phosphorylation in SYF–/– and FAK–/– cells. (A) SYF–/– and Null+ cells were allowed to adhere tofibronectin-coated dish (10 µg ml–1) in serum-free media containing 50 µM sodium pervanadate. The cells were then lysed and each lysate wasprecipitated with anti-PLC-γ1 antibodies followed by blotting with site-specific antibodies against tyrosine-783-phosphorylated PLC-γ1 orPLC-γ1 as indicated. Cells maintained in suspension (Susp) were used as a control. (B) TFW-46 cells were grown in the presence oftetracycline (Tet+) to maintain an uninduced state. To induce FAK expression, cells were incubated in tetracycline-free growth medium. Cellswere allowed to adhere on fibronectin-coated dish (10 µg ml–1) in serum-free medium containing 50 µM sodium pervanadate. Then, the cellswere lysed and each lysate was precipitated with PLC-γ1 antibodies, followed by blotting with site-specific antibodies against tyrosine-783-phosphorylated PLC-γ1 or PLC-γ1 as indicated.

Jour

nal o

f Cel

l Sci

ence

608

of FAK expression, which supports the conclusion that thePLC-γ1/Src interaction does not require FAK.

To determine whether the PLC-γ1/FAK interaction dependson Src, SYF–/– and Null+ cells were plated on fibronectin-coated dishes for different periods of time. The cells were lysedand the lysates were precipitated with anti-FAK antibody andthen blotted with PLC-γ1 antibody. The data in Fig. 10C showthat PLC-γ1 precipitates with FAK in a fibronectin-induciblemanner in both SYF–/– and Null+ cells. Taken together, thesedata indicate that PLC-γ1 interactions with FAK and Src arenot dependent on each other after integrin engagement.

DiscussionIn this study, mouse embryonic fibroblasts isolated fromPlcg1–/– mice were used to demonstrate that phosphorylationof PLC-γ1 at Tyr783 is stimulated by fibronectin and isrequired for maximal cell adhesion on fibronectin. The PLC-γ1-dependent defect in adhesion was clearly evident at low, butnot at high, concentrations of fibronectin. Similar to theadhesion results, Null cells took longer to spread fully thanNull+ cells. These results suggest that PLC-γ1 is not anabsolute requirement for either cell adhesion or spreading but

rather facilitates these cell functions. PLC-γ1 might alsoinfluence other aspects of adherent physiology, such as cellpolarity.

Using Plcg1-null cells, this report extends the observationsof many others who have shown that inhibitors of PLC activitydecrease cell adhesion (Schottelndreier et al., 1999). However,the PLC-selective basis of these pharmacological inhibitors isunclear. Published data showing that Plcg2–/– platelets have aspreading defect following adhesion to collagen, fibrinogen orvonWillebrand factor (Asselin et al., 1997; Goncalves et al.,2003; Inoue et al., 2003; Wonerow et al., 2003; Rathore et al.,2004) concurs with our data. In contrast to our data, Plcg2–/–

platelets do not show deficits in cell adhesion. However, theconcentrations of extracellular matrix ligand used in thosestudies was high and not varied. It is possible that lowerconcentrations of those extracellular matrix molecules mighthave revealed a PLC-γ2 requirement for platelets in theadhesion process. Plcg2–/– platelet adhesion to a collagensurface was decreased relative to wild-type platelets, when ashear force was applied (Suzuki-Inoue et al., 2003).

In contrast to the adhesion results, a defect in cell migrationwas present in Null cells at both low and high concentrationsof fibronectin. These data indicate that the decrease in

Journal of Cell Science 118 (3)

Fig. 9. Fibronectin-induced PLC-γ1 association with Src or FAK. (A) Null+ cells were allowed to adhere for the indicated times on fibronectin-coated (10 µg ml–1) dish in serum-free medium. The cells were then lysed and each lysate was precipitated with anti-Src-kinase antibodyfollowed by blotting with PLC-γ1, FAK or Src antibodies. Whole-cell lysate was loaded as a marker for migration pattern of each protein.(B) Null+ cells expressing wild-type (wt) PLC-γ1 or N+C SH2– PLC-γ1 mutant were allowed to adhere for the indicated times on fibronectin-coated dishes (10 µg ml–1) in serum-free medium. The cells were then lysed and each lysate was precipitated with anti-Src-kinase antibodiesfollowed by blotting with anti-PLC-γ1 or anti-Src antibodies. (C) Null+ cells were allowed to adhere to fibronectin-coated dishes (10 µg ml–1)for the indicated times. Cells were lysed and each lysate was precipitated with anti-FAK antibody followed by blotting with anti-PLC-γ1 oranti-FAK antibodies.

Jour

nal o

f Cel

l Sci

ence

609PLC-γ1 and integrin-dependent cell adhesion

migration of the Null cells is not due to just the adhesiondefects. PLC-γ1 is important in growth-factor-inducedmigration, in which it is thought to play a role in modulatingevents related to cytoskeleton organization at the leading edgeof migrating cells (Piccolo et al., 2002).

FAK is a crucial component in the transduction of signalingpathways following integrin ligation (Sieg et al., 1999).The fact that FAK phosphorylation was similar in Null andNull+ cells suggests that activation of this protein is notdependent on the presence of PLC-γ1. In addition, althoughFAK associates with PLC-γ1, FAK is not required forintegrin-induced phosphorylation of PLC-γ1. These resultssupport the report that, although FAK autophosphorylationsite Tyr397 mediates a direct interaction with PLC-γ1,FAK does not directly phosphorylate PLC-γ1 (Zhang et al.,1999).

It has been demonstrated that integrins associate with EGFreceptors in a macromolecular complex during the early phasesof cell adhesion, which results in phosphorylation of the EGFreceptor (Moro et al., 2002). Data showing that an EGF-receptor kinase inhibitor does not inhibit cell-adhesion-dependent PLC-γ1 tyrosine phosphorylation suggests thatintegrin-dependent PLC-γ1 activation does not require EGF-receptor tyrosine kinase activity. By contrast, addition of theSrc-kinase inhibitor PP2 resulted in decreased PLC-γ1 tyrosinephosphorylation, implying that this kinase is required forfibronectin-dependent PLC-γ1 phosphorylation. Results using

the SYF–/– cells, which are deficient in Src-family kinases,confirm the need for Src for PLC-γ1 phosphorylation. Src hasbeen demonstrated to activate PLC-γ1, resulting in Ca2+ releasein fertilization-induced egg activation (Sato et al., 2003). Thisresult contrasts with the growth-factor literature, in which nodifferences are reported in PLC-γ1 activation followingplatelet-derived-growth-factor stimulation of SYF–/– cells(Klinghoffer et al., 1999).

The interaction of Src-family members with FAK is wellestablished (Hanks et al., 2003; Zhang et al., 1999). Inaddition, it has been demonstrated that Src can phosphorylatePLC-γ1 in vitro in the absence of FAK (Zhang et al., 1999)and at the Y783 site (Sekiya et al., 2004). The data presentedin Fig. 9 demonstrate that Src associates with PLC-γ1following integrin-dependent cell adhesion. FAK-null cellswere used to show that the capacity of Src to phosphorylatePLC-γ1 is independent of Src association with FAK. In somemodels, such as CAS phosphorylation, it has been proposedthat FAK is required to mediate a productive ternary complexof FAK-CAS-Src (Hanks et al., 2003). Based on our data, thismodel would not be applicable to PLC-γ1 phosphorylation bySrc.

This research was supported by NIH grant CA 75195 and coreservices support from CA 68485. RZ is supported by a VA AdvancedCareer Development and Merit Award, a Grant in Aid from AHA anda Clinician Scientist Award from NKF.

Fig. 10. PLC-γ1 association with Src and FAK in TFW-46 andSYF–/– cells. (A) TFW-46 cells were grown in the presence oftetracycline (Tet+) to keep them in the uninduced state. To induceFAK expression, the cells were incubated in tetracycline-free growthmedium for the indicated periods of time, then lysed and blotted withanti-FAK and anti-PLC-γ1 antibody. (B) TFW-46 cells withuninduced (Tet+) and induced (Tet–) FAK expression were allowedto adhere for the indicated times on fibronectin(10 µg ml–1)-coateddishes in serum-free medium. Then, cells were lysed and each lysatewas precipitated with anti-Src-kinase antibodies followed by blottingwith anti-PLC-γ1 or anti-Src antibodies. PDL, cells that wereallowed to adhere on a poly-D-lysine-coated dish. (C) SYF–/– andNull+ cells were allowed to adhere for the indicated times on

fibronectin-coated dishes in serum-free medium. Then, cells were lysed and each lysate was precipitated with anti-FAK antibodies followed byblotting with anti-PLC-γ1 or anti-FAK antibodies. PDL, cells that were allowed to adhere on a poly-D-lysine-coated dish.

Jour

nal o

f Cel

l Sci

ence

610

ReferencesArteaga, C. L., Johnson, M. D., Todderud, G., Coffey, R. J., Carpenter, G.

and Page, D. L. (1991). Elevated content of the tyrosine kinase substratephospholipase C-gamma 1 in primary human breast carcinomas. Proc. Natl.Acad. Sci. USA 88, 10435-10439.

Asselin, J., Gibbins, J. M., Achison, M., Lee, Y. H., Morton, L. F.,Farndale, R. W., Barnes, M. J. and Watson, S. P. (1997). A collagen-likepeptide stimulates tyrosine phosphorylation of SYK and phospholipase Cgamma2 in platelets independent of the integrin alpha2beta1. Blood 15,1235-1242.

Berridge, M. J. and Irvine, R. F. (1984). Inositol trisphosphate, a novelsecond messenger in cellular signal transduction. Nature 312, 315-321.

Carpenter, G. and Ji, Q. (1999). Phospholipase C-gamma as a signal-transducing element. Exp. Cell Res. 253, 15-24.

Clark, E. A. and Brugge, J. S. (1995). Integrins and signal transductionpathways: the road taken. Science 14, 233-239.

Cybulsky, A. V., McTavish, A. J. and Papillon, J. (1996). Extracellularmatrix stimulates production and breakdown of inositol phospholipids. Am.J. Physiol. 271, 579-587.

Goncalves, I., Hughan, S. C., Schoenwaelder, S. M., Yap, C. L., Yuan, Y.and Jackson, S. P. (2003). Integrin alpha IIb beta 3-dependent calciumsignals regulate platelet-fibrinogen interactions under flow. Involvement ofphospholipase C gamma 2. J. Biol. Chem. 278, 34812-34822.

Hanks, S. K., Ryzhova, L., Shin, N. Y. and Brabek, J. (2003). Focaladhesion kinase signaling activities and their implications in the control ofcell survival and motility. Front. Biosci. 8, 982-996.

Hynes, R. O. (2002). Integrins: bidirectional, allosteric signaling machines.Cell 110, 673-687.

Inoue, O., Suzuki-Inoue, K., Dean, W. L., Frampton, J. and Watson, S. P.(2003). Integrin alpha2beta1 mediates outside-in regulation of plateletspreading on collagen through activation of Src kinases and PLCgamma2.J. Cell Biol. 160, 769-780.

Ji, Q. S., Ermini, S., Baulida, J., Sun, F. L. and Carpenter, G. (1998).Epidermal growth factor signaling and mitogenesis in Plcg1 null mouseembryonic fibroblasts. Mol. Biol. Cell 9, 749-757.

Ji, Q. S., Chattopadhyay, A., Vecchi, M. and Carpenter, G. (1999).Physiological requirement for both SH2 domains for phospholipase C-gamma1 function and interaction with platelet-derived growth factorreceptors. Mol. Cell. Biol. 19, 4961-4970.

Juliano, R. L. (2002). Signal transduction by cell adhesion receptors and thecytoskeleton: functions of integrins, cadherins, selectins, andimmunoglobulin-superfamily members. Annu. Rev. Pharmacol. Toxicol. 42,283-323.

Kanner, S. B., Grosmaire, L. S., Ledbetter, J. A. and Damle, N. K. (1993).Beta 2-integrin LFA-1 signaling through phospholipase C-gamma 1activation. Proc. Natl. Acad. Sci. USA 90, 7099-7103.

Kim, H. K., Kim, J. W., Zilberstein, A., Margolis, B., Kim, J. G.,Schlessinger, J. and Rhee, S. G. (1991). PDGF stimulation of inositolphospholipid hydrolysis requires PLC-gamma 1 phosphorylation ontyrosine residues 783 and 1254. Cell 65, 435-441.

Klinghoffer, R. A., Sachsenmaier, C., Cooper, J. A. and Soriano, P. (1999).Src family kinases are required for integrin but not PDGFR signaltransduction. EMBO J. 18, 2459-2471.

Miyamoto, S., Teramoto, H., Coso, O. A., Gutkind, J. S., Burbelo, P. D.,Akiyama, S. K. and Yamada, K. M. (1995). Integrin function: molecularhierarchies of cytoskeletal and signaling molecules. J. Cell Biol. 131, 791-805.

Moro, L., Dolce, L., Cabodi, S., Bergatto, E., Erba, E. B., Smeriglio, M.,Turco, E., Retta, S. F., Giuffrida, M. G., Venturino, M. et al. (2002).

Integrin-induced epidermal growth factor (EGF) receptor activation requiresc-Src and p130Cas and leads to phosphorylation of specific EGF receptortyrosines. J. Biol. Chem. 277, 9405-9414.

Nishibe, S., Wahl, M. I., Hernandez-Sotomayor, S. M., Tonks, N. K., Rhee,S. G. and Carpenter, G. (1990). Increase of the catalytic activity ofphospholipase C-gamma 1 by tyrosine phosphorylation. Science 250, 1253-1256.

Owen, J. D., Ruest, P. J., Fry, D. W. and Hanks, S. K. (1999). Induced focaladhesion kinase (FAK) expression in FAK-null cells enhances cell spreadingand migration requiring both auto- and activation loop phosphorylation sitesand inhibits adhesion-dependent tyrosine phosphorylation of Pyk2. Mol.Cell. Biol. 19, 4806-4818.

Piccolo, E., Innominato, P. F., Mariggio, M. A., Maffucci, T., Iacobelli, S.and Falasca, M. (2002). The mechanism involved in the regulation ofphospholipase Cgamma1 activity in cell migration. Oncogene 21, 6520-6529.

Plopper, G. E., McNamee, H. P., Dike, L. E., Bojanowski, K. and Ingber,D. E. (1995). Convergence of integrin and growth factor receptor signalingpathways within the focal adhesion complex. Mol. Biol. Cell 6, 1349-1365.

Plow, E. F., Haas, T. A., Zhang, L., Loftus, J. and Smith, J. W. (2000).Ligand binding to integrins. J. Biol. Chem. 275, 21785-21788.

Rathore, V., Wang, D., Newman, D. K. and Newman, P. J. (2004).Phospholipase Cgamma2 contributes to stable thrombus formation on VWF.FEBS Lett. 573, 26-30.

Sato, K., Tokmakov, A. A., He, C. L., Kurokawa, M., Iwasaki, T.,Shirouzu, M., Fissore, R. A., Yokoyama, S. and Fukami, Y. (2003).Reconstitution of Src-dependent phospholipase Cgamma phosphorylationand transient calcium release by using membrane rafts and cell-free extractsfrom Xenopus eggs. J. Biol. Chem. 278, 38413-38420.

Schottelndreier, H., Mayr, G. W. and Guse, A. H. (1999). Beta1-integrinsmediate Ca2+-signalling and T cell spreading via divergent pathways. CellSignal. 11, 611-619.

Sekiya, F., Poulin, B., Kim, Y. J. and Rhee, S. G. (2004). Mechanism oftyrosine phosphorylation and activation of phospholipase C-gamma 1.Tyrosine 783 phosphorylation is not sufficient for lipase activation. J. Biol.Chem. 279, 32181-32190.

Sieg, D. J., Hauck, C. R. and Schlaepfer, D. D. (1999). Required role of focaladhesion kinase (FAK) for integrin-stimulated cell migration. J. Cell Sci.112, 2677-2691.

Suzuki-Inoue, K., Inoue, O., Frampton, J. and Watson, S. P. (2003). MurineGPVI stimulates weak integrin activation in PLCgamma2–/– platelets:involvement of PLCgamma1 and PI3-kinase. Blood 102, 1367-1373.

Thomas, J. W., Ellis, B., Boerner, R. J., Knight, W. B., White, G. C., IIand Schaller, M. D. (1998). SH2- and SH3-mediated interactions betweenfocal adhesion kinase and Src. J. Biol. Chem. 273, 577-583.

Tvorogov, D. and Carpenter, G. (2002). EGF-dependent association ofphospholipase C-gamma1 with c-Cbl. Exp. Cell Res. 277, 86-94.

van der Flier, A. and Sonnenberg, A. (2001). Function and interactions ofintegrins. Cell Tissue Res. 305, 285-298.

Vossmeyer, D., Hofmann, W., Loster, K., Reutter, W. and Danker, K.(2002). Phospholipase Cgamma binds alpha1beta1 integrin and modulatesalpha1beta1 integrin-specific adhesion. J. Biol. Chem. 277, 4636-4643.

Wonerow, P., Pearce, A. C., Vaux, D. J. and Watson, S. P. (2003). A criticalrole for phospholipase Cgamma2 in alphaIIbbeta3-mediated plateletspreading. J. Biol. Chem. 278, 37520-37529.

Zhang, X., Chattopadhyay, A., Ji, Q. S., Owen, J. D., Ruest, P. J.,Carpenter, G. and Hanks, S. K. (1999). Focal adhesion kinase promotesphospholipase C-gamma1 activity. Proc. Natl. Acad. Sci. USA 96, 9021-9026.

Journal of Cell Science 118 (3)

Jour

nal o

f Cel

l Sci

ence