Phospholipase Cϵ Suppresses Integrin Activation*

Phospholipase Cϵ (PLCϵ) is a newly described effector of the small GTP-binding protein H-Ras. Utilizing H-Ras effector mutants, we show that mutants H-Ras(G12V/E37G) and H-Ras(G12V/D38N) suppressed integrin activation in an ERK-independent manner. H-Ras(G12V/D38N) specifically activated the PLCϵ effector pathway and suppressed integrin activation. Inhibition of PLCϵ activation with a kinase-dead PLCϵ mutant prevented H-Ras(G12V/D38N) from suppressing integrin activation, and low level expression of H-Ras(G12V/D38N) could synergize with wild-type PLCϵ to suppress integrins. In addition, knockdown of endogenous PLCϵ with small interfering RNA blocked H-Ras(G12V/D38N)-mediated integrin suppression. Suppressing integrin function with the H-Ras(G12V/D38N) mutant reduced cell adhesion to von Willebrand factor and fibronectin; this reduction in cell adhesion was blocked by coexpression of the kinase-dead PLCϵ mutant. These results show that H-Ras suppresses integrin affinity via independent Raf and PLCϵ signaling pathways and demonstrate a new physiological function for PLCϵ in the regulation of integrin activation.

Integrins are heterodimeric glycoproteins that control cellcell and cell-substratum adhesion and that regulate cell survival, proliferation, and migration (1). An essential feature of integrins is their ability to regulate the strength of ligand binding, a process termed affinity modulation (2). Various intracellular signals can induce a conformational change in the integrin heterodimer, activating or suppressing ligand binding. Members of the Ras family of small GTP-binding proteins have been shown to modulate integrin affinity (2).
Expression of a constitutively active mutant of H-Ras, H-Ras(G12V), in Chinese hamster ovary (CHO) 5 cells sup-presses integrin activation (3). In addition, activation of the H-Ras downstream effector Raf also suppresses integrin activation in CHO cells (3,4). In contrast, R-Ras, a closely related member of the Ras superfamily, activates integrins and reverses H-Ras-mediated suppression of integrin affinity (4,5). Activation of Raf by H-Ras leads to the phosphorylation and activation of ERK1/2. H-Ras-mediated suppression of integrin affinity can be reversed by expression of MAPK phosphatase-1 (MKP-1), which can dephosphorylate and inactivate ERK1/2 (3). However, studies using H-Ras/R-Ras chimeras have revealed that integrin affinity modulation does not precisely correlate with ERK1/2 activation (6,7). Remarkably, although targeting of ERK1 to the plasma membrane has been shown to be sufficient to suppress integrins (8), inhibition of ERK1/2 activation with either MKP-3 or U0126 fails to affect integrin suppression by H-Ras (7). This discrepancy in the current data might therefore be explained by the demonstration of an alternative pathway for integrin suppression that does not rely on ERK1/2 activation.
Phosphoinositide-specific phospholipase C (PI-PLC) catalyzes the hydrolysis of phosphatidylinositol 4,5-bisphosphate into the second messengers diacylglycerol and inositol 1,4,5trisphosphate (IP 3 ). Diacylglycerol stimulates protein kinase C activation, and IP 3 mobilizes intracellular Ca 2ϩ (18). Three major classes of PI-PLC have previously been identified: PLC␤, PLC␥, and PLC␦ (19). They contain an N-terminal pleckstrin homology (PH) domain, an EF-hand domain, catalytic X and Y domains, and the regulatory C2 domain. PLC␥ contains another PH domain, which is split by two SH2 domains and one SH3 domain. These PI-PLC classes are activated by distinct signaling mechanisms (20). PLC␤ is activated by the ␣ subunit (G␣) or ␤␥ subunits (G␤␥) of heterotrimeric G proteins. PLC␥ is activated by tyrosine phosphorylation following binding to tyrosine kinases of receptor or non-receptor types through its SH2 domain. PLC␦ is activated by the high molecular weight G protein G h and/or by an increase in the concentration of intracellular Ca 2ϩ .
Recently, a fourth class of PI-PLC was identified, PLC⑀ (15)(16)(17)21). PLC⑀ shares the typical X, Y, and C2 domains with the other PI-PLC classes. PLC⑀ also contains putative PH and EFhand domains and is activated by G␤␥ subunits (22). Furthermore, PLC⑀ is unique in that it possesses two types of functional domains not seen in other classes. At its N terminus, PLC⑀ possesses a CDC25 homology domain (a GEF domain for the Ras family of small G proteins), which exhibits GEF activity toward Rap1 and H-Ras (17,23,24). At its C terminus, PLC⑀ possesses two H-Ras/Rap1-associating (RA) domains, RA1 and RA2. H-Ras binds to PLC⑀ in a GTP-dependent manner through its RA2 domain to stimulate the hydrolysis of phosphatidylinositol 4,5-bisphosphate into the secondary messengers IP 3 and diacylglycerol, suggesting that PLC⑀ may be a downstream effector of H-Ras and Rap1 (15)(16)(17).
Despite the characterization of these domains within PLC⑀, its physiological function remains unknown. However, given the signaling attributes and wide tissue distribution of PLC⑀, it is likely that this protein has a critical role in mammalian physiology. We therefore examined whether H-Ras utilizes PLC⑀ to suppress integrins in cells and thus modulate cell adhesion to the extracellular matrix.
Transient transfection of cells with plasmid DNA was performed with Lipofectamine TM Plus reagent (Invitrogen) following the manufacturer's instructions. Twenty-four hours after transfection, the medium containing DNA-Lipofectamine complexes was removed and replaced with fresh complete medium. For experiments when protein activity was to be assessed, the transfection medium was replaced with quiescent medium. Forty-eight hours after transfection, cells were either lysed for SDS-PAGE analyses or used for integrin affinity determination.
Small Interfering RNA (siRNA) of PLC⑀-Kelley et al. (25) have previously described rat PLC⑀-specific siRNA oligonucleotides that effectively knock down PLC⑀ in rat cells and scrambled non-targeting controls. To assess the whether these siRNA oligonucleotides could be used in CHO cells, we amplified the corresponding regions of siRNAPLC⑀#3 and siR-NAPLC⑀#5 from CHO cDNA by reverse transcription-PCR using primers based on consensus rat/mouse PLC⑀ sequence. The respective 616-and 380-bp products were gel-purified and sequenced. We confirmed that the region targeted by siR-NAPLC⑀#3 and siRNAPLC⑀#5 was 100% conserved in CHO cells. Thus, we used these siRNA oligonucleotides to knock down PLC⑀ in CHO cells. Transient transfection of cells with siRNA was performed with Oligofectamine (Invitrogen) following the manufacturer's instructions. Twenty-four hours after transfection, the medium containing siRNA-Oligofectamine complexes was removed and replaced with fresh complete medium. Cells were then transfected with plasmid constructs using Lipofectamine as indicated below. Seventytwo hours after siRNA transfection, cells were either lysed for RNA extraction or protein analysis or used for integrin affinity determination.
Fibronectin Type III Repeat 9 -11 Fragment-The soluble type III repeats 9 -11 of fibronectin (referred to as FN9 -11), which contain the RGD domain responsible for integrin binding, were expressed as a glutathione S-transferase (GST) fusion protein in Escherichia coli BL21 cells from the pGEX-4T2 vector and biotinylated as described previously (7,26).
Flow Cytometry-Integrin affinity in transfected cells was analyzed by three-color flow cytometry. Cells were transfected with test DNA (as stated) together with 0.75 g of Tac-␣ 5 transfection reporter construct. Single cell suspensions of trypsinized cells were resuspended in a total volume of 50 l containing either PAC-1 (5 g/ml) or FN9 -11 for 30 min at room temperature in 20 mM HEPES, 140 mM NaCl, 1.8 mM CaCl 2 , 1 mM MgCl 2 , and 2 mg/ml glucose (pH 7.4). Internal controls containing either 5 mM EDTA or 100 M MnCl 2 were performed for each sample. Cells were washed with cold phosphate-buffered saline (PBS) and incubated on ice with 50 l of DMEM containing 1:25 fluorescein isothiocyanate-conjugated anti-mouse IgM (for PAC-1) or fluorescein isothiocyanateconjugated streptavidin (for FN9 -11) for 30 min in the dark. Cells were washed again and incubated on ice for an additional 30 min with 50 l of DMEM containing 1:50 (v/v) R-phycoerythrin-conjugated anti-Tac antibody (ACT-1). Cells were finally washed and resuspended in cold PBS. Immediately prior to analysis on a FACSCalibur (BD Biosciences, Erembodegem, Belgium), TO-PRO-3 (Molecular Probes, Leiden, The Netherlands) at a final concentration of 1 M (in PBS) was added to each sample.
PAC-1/FN9 -11 binding was determined by gating for live and highly transfected cells (TO-PRO-3-negative and high Tac binding, respectively). To obtain numerical estimates of integrin activation, an integrin activation index (AI) was calculated, where AI ϭ ((F N Ϫ F I )/(F A Ϫ F I )) ϫ 100 and percent inhibition ϭ ((AI 0 Ϫ AI)/AI 0 ) ϫ 100. F N is the geometric mean fluorescence intensity of PAC-1/FN9 -11 binding of the native integrin, F I is the mean fluorescence intensity of PAC-1/ FN9 -11 binding in the presence of 5 mM EDTA, and F A is the mean fluorescence intensity of PAC-1/FN9 -11 binding in the presence of 100 M Mn 2ϩ , AI 0 is the activation index with the control vector, and AI is the activation index with the DNA under testing.
Gel Electrophoresis and Western Blotting-Cell lysates were resuspended in Laemmli sample buffer, separated on 10 -12% SDS-polyacrylamide gels, and transferred onto Hybond-C nitrocellulose (Amersham Biosciences, Buckinghamshire, UK). Immunoblotting was performed with appropriate antibodies diluted in 5% nonfat dried milk powder and detected by chemiluminescence (ECL, Amersham Biosciences) following the manufacturer's instructions. Relative protein expression was quantified using ImageJ software.
RNA Extraction and Reverse Transcription-PCR-Total RNA was extracted from 1 ϫ 10 6 cells using an RNeasy kit (Qiagen Inc.) according to the manufacturer's instructions. Contaminating DNA was removed by treatment with RQ1 DNase (Promega Corp.), and the RNA was quantified using a spectrophotometer and then stored at Ϫ80°C. cDNAs were generated by reverse transcription of 400 ng of RNA using TaqMan reverse transcription reagents (Applied Biosystems) according to the manufacturer's instructions and stored at Ϫ20°C. All reagents for PCR were obtained from Promega Corp. The reaction contained 2.5 l of 10ϫ buffer, 0.25 l of Taq polymerase, 1 l of dNTP (10 mM), 0.5 M forward and reverse primers, 17.25 l of nuclease-free water, and 2 l of cDNA. The PCR program was as follows: 95°C for 2 min; 35 cycles at 95°C for 45 s, 56°C for 45 s, and 72°C for 45 s; and a final extension at 72°C for 5 min. The primers used were PLC⑀for (GGCTACGTAGGCAG-GATTGTCTTA), PLC⑀rev (TTTCCCTGCACCCTTCCACT-TGC), ␤-actin-for (CCACCAACTGGGACGACATG), and ␤-actin-rev (GTCTCAAACATGATCTGGGTCATC). PCR products were resolved on a 2% agarose gel, purified using a gel extraction kit (Qiagen Inc.), and sequenced on both strands (MWG Biotech).
Ral Activity Assay-Cells were transfected with Myc-tagged Ral constructs as described above and quiesced 24 h prior to lysis. Cells were lysed in 450 l of Ral lysis buffer (50 mM Tris-HCl (pH 7.4), 200 mM NaCl, 2.5 mM MgCl 2 , 1% (v/v) Nonidet P-40, 15% (v/v) glycerol, and one Complete TM protease inhibitor tablet). Lysates were clarified by centrifugation at 13,000 rpm for 10 min at 4°C, and the protein concentration estimated. Protein concentrations were normalized, and equal volumes were incubated with 50 l of the Ral-binding domain of RLIP76 coupled to GST-agarose beads (28) for 60 min at 4°C. Samples of whole cell lysates were taken for loading controls of Ral proteins. Agarose beads were washed four times with Ral lysis buffer and finally resuspended in Laemmli sample buffer. Whole cell lysates and Ral-binding domain-bound samples (20 l) were analyzed by SDS-PAGE on a 15% gel, and Western blots were probed with anti-Ral antibody.
PLC Activity Assay-CHO-K1 cells were maintained in inositol-free DMEM (ICN Biochemicals, Basingstoke) supplemented with dialyzed 10% (v/v) fetal bovine serum (Labtech International, East Sussex, UK). Cells were transfected as described above and quiesced 24 h prior to lysis in serum-free DMEM containing 5 Ci of myo-[ 3 H]inositol (Amersham Biosciences). Cells were then incubated with 20 mM lithium chloride for 60 min, washed, and lysed with 500 l of 0.5 M trichloroacetic acid. Lysates were clarified at 13,000 ϫ g for 10 min, and the supernatant (400 l) was neutralized by addition to a 50:50 mixture of 1,1,2-trichlorotrifluoroethane/tri-n-octylamine (750 l). The mixture was vortexed and separated by centrifugation at 13,000 ϫ g for 5 min. The aqueous phase was collected (300 l), diluted by the addition of 10 ml of ice-cold distilled H 2 O, and loaded onto an AG 1-X8 200 -400 mesh formate form column (Bio-Rad, Hertfordshire, UK). Following a column wash with 10 ml of 60 mM ammonium formate and 5 mM sodium tetraborate, [ 3 H]IP 1-3 was eluted with 5 ml of 1.2 M ammonium formate and 0.1 M formic acid and measured by liquid scintillation counting.
Cell Adhesion Assay-Cells were transfected with test DNA, and CHO(␣␤-py) cells were also transfected with 2 g of pEGFP-C3. Cells were harvested after 48 h and resuspended at 1 ϫ 10 6 cells/ml in DMEM. 96-Well cell culture cluster plates were coated with 5 g/ml von Willebrand factor (vWF) (CHO(␣␤-py) cells) or 10 g/ml fibronectin (CHO-K1 cells) in PBS for 60 min at 37°C and blocked with 2% (w/v) bovine serum albumin in PBS for 60 min at room temperature. The cell suspension (200 l) was added to each well and allowed to adhere for 15 min at 37°C. Unattached cells were removed by gentle shaking for 3 ϫ 10 s. Total cell adhesion was assessed by centrifugation of the plate at 1000 rpm for 5 min. For CHO(␣␤py) cells, adhesion was quantified on a fluorescence plate reader. For CHO-K1 cells, adhesion was quantified by staining adherent cells with methylene blue (0.4%) for 5 min, three washes with PBS, and elution of stain with 0.1 M HCl. The absorbance of each sample was read at 640 nm on an optical plate reader. Adhesion was expressed as a percentage compared with total cell adhesion with background cell adhesion to plastic subtracted against all values. The data represented are shown as the percent change in cell adhesion with values normalized against control cell adhesion to the extracellular matrix.
Statistical Analysis-Data were analyzed by one-way analysis of variance, and the appropriate post-test analyses were applied. p values Ͻ0.05 were considered to be significant.

Integrin Suppression by H-Ras(G12V) Effector Mutants T35S
and E37G-It has previously been reported that a constitutively active mutant of H-Ras (H-Ras(G12V)) suppresses integrin function (3). However, the signaling pathways downstream of H-Ras are still poorly understood. To examine the role of downstream effectors of H-Ras in integrin suppression, we utilized the CHO(␣␤-py) cell line, a CHO cell line stably expressing a chimeric integrin that contains the extracellular and transmembrane domains of ␣ IIb ␤ 3 fused to the cytoplasmic domains of ␣ 6A ␤ 1 (3,4). H-Ras(G12V) effector mutants were transfected into CHO(␣␤-py) cells, and their effect upon integrin affinity modulation was assessed. The integrin AI was quantified from changes in the binding levels of the ␣ IIb ␤ 3 ligand-mimetic monoclonal antibody PAC-1, detected by flow cytometry as described under "Experimental Procedures." Fig. 1A shows that the H-Ras(G12V) effector mutants E37G (25.3 Ϯ 14.6%) and T35S (27.9 Ϯ 6.6%) both induced a significant reduction in the AI compared with control vector-transfected cells (70.6 Ϯ 8.5%, p Ͻ 0.01). This was comparable with H-Ras(G12V)-induced reduction in the integrin AI (23.1 Ϯ 4.3%).
In contrast, effector mutant Y40C (59.0 Ϯ 15%) did not significantly reduce the AI in our system. Furthermore, we confirmed previous results (3) that inhibition of phosphatidylinositol 3-kinase activity with LY294002 (10 M) does not affect H-Ras(G12V)-mediated integrin suppression in CHO(␣␤-py) cells (data not shown), in agreement with the inability of H-Ras(G12V/Y40C) to suppress integrins. Each mutant was expressed at similar levels in the transfected CHO(␣␤-py) cells (Fig. 1B). H-Ras(G12V) and the effector mutant T35S both stimulated ERK1/2 phosphorylation as detected by anti-phospho-ERK1/2 antibody (clone MAPK-YT; 4.6 Ϯ 0.6-and 3.2 Ϯ 0.5-fold increase (mean Ϯ S.E.) compared with vector, respectively; n ϭ three independent experiments) (Fig. 1B). As expected, the effector mutants E37G and Y40C did not stimulate ERK1/2 phosphorylation above that in control transfected cells (0.9 Ϯ 0.2-and 1.1 Ϯ 0.2-fold change, respectively). Transfection of the H-Ras mutants did not alter the level of ERK2 expression in these transfected cells. Fig. 1C shows the effect of H-Ras(G12V/E37G) expression in the CHO(␣␤-py) cell assay. Expression of H-Ras(G12V/E37G) caused a marked reduction in PAC-1 binding, resulting in a leftward shift in the highly transfected cell population (Fig. 1C, upper quadrants). The untransfected/poorly transfected cell population (Fig. 1C, lower quadrants) did not display any significant change in PAC-1 binding. PAC-1 binding to the chimeric integrin in CHO(␣␤-py) cells was inhibited by EDTA (Fig. 1C, left panel). In contrast, H-Ras(G12V/E37G)-transfected cells in the presence of Mn 2ϩ displayed a slight rightward shift in the whole cell population as a result of increased PAC-1 binding (Fig. 1C, right panel). The ability of Mn 2ϩ to override H-Ras(G12V/E37G) suppression of PAC-1 binding by activating integrins indicates that the effect of H-Ras is not due to changes in integrin expression in this system, but rather to suppression of integrin activity. These results suggest that H-Ras(G12V) mediates integrin suppression by two separate effector pathways: a Raf/ERK-dependent signaling pathway utilized by T35S and a Raf/ERK-independent pathway utilized by E37G.
Integrin Suppression by H-Ras(G12V/E37G) Is Not Mediated by RalA-The small GTP-binding protein Ral is a downstream effector of H-Ras(G12V) that is activated by the effector mutant E37G (29). We therefore examined whether RalA plays a role in the suppression of integrin affinity. CHO(␣␤-py) cells were transfected with Myc-tagged wild-type RalA, constitutively active RalA(G23V), or dominant-negative RalA(S28N) in the presence of either the control vector or H-Ras(G12V/E37G). Integrin suppression was measured by the percent change in the integrin AI from the control as described under "Experimental Procedures." Cotransfection with either the RalA(G23V) (41.6 Ϯ 3.9%) or RalA(S28N) (47.3 Ϯ 3.3%) mutant had no significant effect on H-Ras(G12V/E37G)-mediated suppression (57.1 Ϯ 0.5%) ( Fig. 2A). Furthermore, cells cotransfected with either RalA(G23V) or RalA(S28N) and the control vector also had no significant effect on integrin affinity compared with control cells.
Ral activity was assessed in duplicate samples using the Ralbinding domain of RLIP76 fused to GST to pull down active Ral (28). Cells cultured in serum-free medium for 12 h prior to assay showed a low basal level of RalA activity, which was H-Ras Mutant D38N Mediates Integrin Suppression-PLC⑀ is a newly described Ras effector that is activated by the H-Ras-(G12V/E37G) mutant (15)(16)(17). In addition to H-Ras(G12V/ E37G), PLC⑀ activity is also stimulated by the H-Ras(G12V/ D38N) mutant (15). The D38N mutant does not stimulate Ral GEFs or bind to Raf, and so the effect of PLC⑀ can be examined in the absence of Ral or Raf signaling (30). Expression of H-Ras(G12V/D38N) in CHO(␣␤-py) cells caused marked integrin suppression as demonstrated by inhibition of PAC-1 bind-ing (Fig. 3A). Neither the D38N nor E37G mutant significantly affected ERK1/2 phosphorylation compared with control cells (0.93 Ϯ 0.19-and 0.98 Ϯ 0.09-fold change (mean Ϯ S.E.) compared with vector-transfected cells, n ϭ three independent experiments) (Fig. 3B). Equal expression of the HA-tagged H-Ras mutants is shown by Western blot analysis of whole cell lysates with anti-HA antibody (Fig. 3B). The level of integrin suppression induced by H-Ras(G12V/D38N) (53.5 Ϯ 3.75%) was comparable with that induced by H-Ras(G12V/E37G) (51.1 Ϯ 3.5%). This level of integrin suppression demonstrated by the effector mutants E37G and D38N, neither of which interacts with Raf, indicates that H-Ras may utilize PLC⑀ as a Raf-independent alternative pathway to suppress integrin activation.
To confirm the data obtained with PLC⑀(H1433L), we sought to knock down endogenous PLC⑀ in CHO(␣␤-py) cells and to investigate the effect on H-Ras(G12V/D38N)-mediated integrin suppression. siRNA oligonucleotides targeting conserved sequences from rat, human, and mouse PLC⑀ have previously been described and shown to reduce PLC⑀ expression by up to 97% (25). Our analysis of the surrounding regions of siRNAPLC⑀#3 and siRNAPLC⑀#5 showed that CHO PLC⑀ is nearly identical between rat and hamster and that the two siRNA target sites of siRNAPLC⑀#3 and siRNAPLC⑀#5 are both completely conserved. Therefore, we transfected CHO-(␣␤-py) cells with either the PLC⑀-targeting siRNA or a scrambled non-targeting oligonucleotide and assessed the effect on integrin suppression mediated by transient expression of H-Ras(G12V/D38N).
PLC⑀ Contributes to H-Ras-mediated Reduction in Cell Adhesion-Modulation of integrin affinity is a central process in the control of cell adhesion to the extracellular matrix (2). To investigate whether PLC⑀ participates in H-Ras-mediated reduction in cell adhesion, we assessed adhesion of CHO(␣␤-py) cells to vWF, a ligand for the chimeric ␣ IIb ␣ 6A ␤ 3 ␤ 1 integrin present in this stable cell line. Transient transfection of H-Ras(G12V/D38N), which mediates integrin suppression, produced a significant reduction in cell adhesion (46 Ϯ 10%; p Ͻ 0.01) (Fig. 8A). Coexpression of H-Ras(G12V/D38N) with PLC⑀(H1433L) significantly blocked this reduction in adhesion to vWF back to control cell levels of adhesion (120 Ϯ 20%; p Ͻ 0.05). This effect on adhesion by PLC⑀ was not due to a reduction in expression of H-Ras(G12V/D38N) (Fig. 8B). Reversing H-Ras(G12V/D38N)-mediated integrin suppression by coex-  pression of PLC⑀(H1433L) therefore restores the cells' ability to adhere to vWF.
We went on to investigate the role of PLC⑀ in the suppression of endogenous integrins. It has previously been shown that a soluble fragment of fibronectin composed of type III repeats 9 -11 (FN9 -11) can bind to endogenous ␣ 5 ␤ 1 integrins in CHO-K1 cells and that FN9 -11 binding can be suppressed by activated Raf-1 signaling (6). We therefore investigated the effect of transfection of H-Ras(G12V/D38N) with or without PLC⑀(H1433L) on FN9 -11 binding in CHO-K1 cells as described under "Experimental Procedures." We found that transfection of H-Ras(G12V/D38N) significantly suppressed FN9 -11 binding and that this was blocked by cotransfection of dominant-negative PLC⑀(H1433L) (46.2 Ϯ 1.9 and 13.5 Ϯ 2.1%, respectively) (Fig. 8, B  and C). Notably, Mn 2ϩ activated and EDTA suppressed FN9 -11 binding similarly in CHO-K1 cells regardless of the DNA transfected, indicating that changes in FN9 -11 binding are not due to changes in integrin expression. In parallel, we allowed CHO-K1 cells transfected with H-Ras(G12V/ D38N) in the absence or presence of PLC⑀(H1433L) to adhere to fibronectin-coated plastic. In support of the FN9 -11 binding data, H-Ras(G12V/D38N) reduced cell adhesion, which was blocked by PLC⑀(H1433L) (Fig. 8D). Thus, PLC⑀ modulates H-Ras suppression of endogenous ␣ 5 ␤ 1 integrin binding to fibronectin.

DISCUSSION
Modulation of integrin affinity plays a central role in the regulation of integrin function (1,2). We have shown for the first time that PLC⑀, a novel H-Ras effector, suppresses integrin affinity. The constitutively active mutant of H-Ras suppresses integrins via two independent H-Ras effector pathways: a Raf-dependent T35S pathway and a Rafindependent E37G/D38N pathway. This alternative mechanism of inte-  OCTOBER 6, 2006 • VOLUME 281 • NUMBER 40 grin suppression utilizes PLC⑀, as shown in Fig. 9. The suppression of integrins via PLC⑀ also reduces cell adhesion to a vWFcoated surface and to the extracellular matrix component fibronectin. Our results thus identify a new physiological function for PLC⑀.

Integrin Suppression via PLC⑀
Hughes et al. (3) have previously shown that both H-Ras-(G12V) and activated Raf can suppress the binding of the ␣ IIb ␤ 3 ligand-mimetic antibody PAC-1 to the stably expressed chimeric ␣ IIb ␣ 6A ␤ 3 ␤ 1 integrin in CHO cells. Fluorescein isothiocyanate-labeled fibronectin binding to endogenous ␣ 5 ␤ 1 in CHO cells is also suppressed by H-Ras(G12V) expression, indicating that inside-out signaling can act upon a native integrin and be detected by a physiological integrin ligand. H-Ras(G12V/T35S) is a well established activator of the Raf effector arm of H-Ras (9) and can mimic many of the putative effects of H-Ras through the activation of the ERK1/2 signaling cascade (36). Integrin suppression by H-Ras(G12V/T35S) indicates that a Raf-dependent mechanism is capable of modulating integrin affinity, consistent with previously published data (3). However, the requirement of ERK1/2 activation for integrin affinity suppression is still unclear. Integrin suppression occurs in the absence of bulk ERK1/2 activation and is not reversed by coexpression of MKP-3 (Fig. 5A). Targeting ERK1 to the plasma membrane with the CAAX box of H-Ras leads to integrin suppression (8), suggesting that localized ERK1/2 activity can allow integrin suppression in the absence of bulk ERK activation. The MEK inhibitor U0126 fails to inhibit both H-Ras(G12V)-and activated Raf-mediated integrin suppression; whether ERK1-CAAX activation at the plasma membrane is sensitive to MEK inhibition remains to be tested. The ability of H-Ras(G12V) to suppress integrins via PLC⑀ in a Raf-independent manner provides a mechanism to resolve the conflicting data relating to the need for ERK1/2 activation. This alternative H-Ras pathway can compensate for and maintain integrin suppression in the presence of either MEK inhibitors or MKP-3 coexpression.
PLC⑀ has previously been identified as an H-Ras effector (15)(16)(17). Kelley et al. (15) showed that PLC⑀ activity is stimulated by the constitutively active H-Ras(Q61L) mutant. PLC⑀ activity is also stimulated by the effector mutant H-Ras(Q61L/ E37G), implying that the H-Ras(E37G) mutant may not solely be an activator of Ral GEFs. In contrast, the H-Ras effector mutants T35S and Y40C fail to significantly increase PLC⑀ activity compared with the control. However, the H-Ras(Q61L/ D38N) mutant increases inositol production by 50 -60% (15), implicating this mutant as an activator of PLC⑀ along with E37G. The D38N mutation within the H-Ras effector domain is known to abrogate Raf binding and activation in vitro and displays minimal transformation potency in NIH/3T3 cells (37). Therefore, the ability of H-Ras(G12V/D38N) to selectively activate PLC⑀ allowed us to specifically examine this pathway in isolation of other effector pathways in our system, viz. Raf, Ral GEFs, and phosphatidylinositol 3-kinase.
We have shown that H-Ras(G12V/D38N) induces a level of integrin suppression similar to that induced by H-Ras(G12V/E37G) (Fig. 3), indicating that the PLC⑀ effector pathway can suppress integrin affinity. Unfortunately, no constitutively active mutant of PLC⑀ is available to test whether overexpression of active PLC⑀ would activate integrin affinity. Furthermore, although physiological agonists of PLC⑀ have been described (25), we (4, 6, 7) and others (5) have demonstrated previously that detection of integrin suppression requires sustained rather than transient activation of H-Ras. However, the ability of kinase-dead PLC⑀(H1433L) and endogenous PLC⑀ knockdown to block H-Ras(G12V/ D38N)-mediated integrin suppression (Fig. 4A) provides good evidence that PLC⑀ catalytic signaling modulates integrin affinity. Our demonstration that PLC⑀(H1433L) does not affect H-Ras(G12V)-stimulated ERK phosphorylation or integrin suppression (Fig. 6) even at high ratios (15:1) suggests that this dominant-negative mutant is not acting through sequestration of H-Ras. Furthermore, our demonstration in Fig. 6A that coexpression of both PLC⑀ and MKP-3 is needed to reverse integrin suppression caused by H-Ras(G12V) is important. It shows that blocking either effector pathway individually is insufficient to fully prevent integrin suppression and provides strong evidence for two independent pathways of H-Ras-mediated integrin suppression.
In platelets, PLC stimulation is essential for integrin affinity modulation of ␣ IIb ␤ 3 during platelet activation, and inhibition of PLC activity with the inhibitor U73122 suppresses adhesion of BaF3 cells to fibronectin (38,39). Integrin activation is prevented pharmacologically by inhibiting the second messengers from PLC activation, viz. intracellular calcium and protein kinase C activation (40). However, we confirmed previously published results (3) showing that PLC␤ and protein kinase C do not suppress integrin affinity in the CHO(␣␤-py) system (data not shown). Furthermore, CDC42 expression, a known activator of PLC␤ in the CHO(␣␤-py) system, fails to suppress integrins (3,31). Expression of protein kinase C isoforms or artificial stimulation of protein kinase C by the phorbol ester phorbol 12-myristate 13-acetate in CHO cells also fails to modulate integrin affinity (41), implying that the suppressing effect of PLC⑀ on integrin affinity is isotype-specific.
Talin, a large actin-binding protein and a component of focal adhesions, links the cytoskeleton with ␤ integrin tails (42). It plays a crucial role in focal adhesion assembly and in integrinmediated signaling recruiting focal adhesion kinase. Talin is a member of the FERM domain-containing protein family; this 30-kDa domain is a cysteine-rich basic charged motif and has been reported to bind phosphoinositides and to target proteins to the plasma membrane (49). These motifs, rich in basic and polar amino acids, may facilitate membrane association of these proteins by interacting with the negatively charged head groups of phosphoinositides controlling protein/protein interactions and biological activities (43). Recent data have revealed that the interaction between the globular head domain of talin and the ␤ 3 cytosolic tail occurs in the membrane-proximal region of the integrin cytoplasmic domain (44 -46). Interaction between the talin head domain and integrin cytoplasmic tails has been shown to be sufficient for integrin activation (47). Talin binding to integrins is regulated by phosphoinositides, which may stabilize the activated conformational state (48,49). The talin head domain can activate integrin affinity through its binding to the ␤ integrin cytoplasmic tail (44,46). Thus, lipid metabolism by PLC⑀ activity at the membrane may play a major role in the sequential assembly of focal adhesions and modulate integrin affinity by regulating talin binding to ␤ integrin tails and other focal adhesion proteins.
The Ras family small GTPases are the commonest mutated oncogenes in human cancer, and these mutations will be particularly relevant in integrin suppression in cancer cells. Both H-Ras and Raf play an important role in cellular transformation and have been implicated in human tumorigenesis. H-Ras has been shown to be an upstream and downstream effector of PLC⑀, depending on the cell type and stimulus examined, and binds to the RA domain, leading to PLC⑀ activation. Additionally, the CDC25 domain of PLC⑀ can act as a Ras GEF. Our results show conclusively that H-Ras is an upstream effector of PLC⑀, leading to integrin suppression. However, the precise role of the CDC25 and RA domains of PLC⑀ in integrin-mediated suppression will be determined by point mutational analysis and forms the basis for further ongoing studies.
PLC⑀ plays a role in tumorigenesis, although its precise role remains controversial. PLC⑀ expression in platelet-derived growth factor-stimulated BaF3 cells enhances cell proliferation rates and inhibits apoptosis (32). However, colorectal cancer cells show a reduction in expression of PLC⑀, and restoring PLC⑀ expression in these tumor lines inhibits tumor cell viability and proliferation (50). In a chemically induced skin tumor model, PLC⑀ Ϫ/Ϫ mice display a reduction in the number of squamous tumors, and conversion of papillomas into carcinomas fails to occur in these mice (51). PLC⑀ may therefore provide the link between integrin activation state, which regulates proliferation and survival, and the conversion of papilloma to carcinoma associated with anchorage independence and downregulation of integrin function.
The novel function of PLC⑀ to modulate integrin affinity allows the H-Ras integrin suppression pathway to occur in a Raf/ERK-independent manner. Ras activation is a focal point for growth factor signaling and is associated with cell cycle control and proliferation (52). Activation of ERK1/2 via Raf is central to the control of cyclins and cyclin-dependent kinases (27). H-Ras-mediated integrin suppression via the PLC⑀ pathway could therefore avoid inappropriate signaling to the cell cycle machinery. In summary, this study has demonstrated a novel and physiologically relevant function of PLC⑀ in suppressing integrin affinity modulation. PLC⑀, a novel H-Ras effector, modulates integrin affinity via the effector mutants E37G and D38N in a Raf-independent manner.