Growth Factor Stimulation of Phospholipase C-71 Activity COMPARATIVE PROPERTIES OF CONTROL AND ACTIVATED ENZYMES*

We demonstrated previously tyrosine phosphoryla- tion-dependent modulation of phospholipase C-y 1 (PLC-yl) catalytic activity (Nishibe, S., Wahl, M. I., Hernandez-Sotomayor, S. M. T., Tonks, N. K., Rhee, S. G., and Carpenter, G. (1990) Science 250, 1253-1256). The increase in PLC-yl catalytic activity in A- 431 cells occurs rapidly, with maximal activation 5 min after epidermal growth factor (EGF) stimulation. Certain other growth factors (fibroblast growth factor, platelet-derived growth factor) also stimulate PLC-yl catalytic activity, whereas insulin does not. A similar increase in PLC-yl specific activity (2-%fold) was observed in both soluble (cytosol) and particulate (membrane) preparations from EGF-treated cells. Ty- rosine-phosphorylated PLC-yl was detected in both cytosol and membrane fractions in lysates from EGF- treated A-431 cells, but the proportion of tyrosine-phosphorylated PLC-yl was higher in the cytosol (-50%) than in the membrane (-20%). Because a mi- cellar concentration of the non-ionic detergent Triton X-100 allows detection of the tyrosine phosphoryla- tion-dependent increase

reaction velocity of the control enzyme was 4-fold lower than the activated enzyme. However, at a high substrate mol fraction (e.g. 0.33), the estimated maximal reaction velocities (Vmm) for both forms of PLCy l were equivalent. PLC-yl activity from both control and EGF-treated cells was stimulated by increasing nanomolar Ca2+ concentrations. Although the catalytic activity of PLC-yl from EGF-treated cells was greater than control PLC-y1 at every Ca" concentration tested, the relative stimulation of activity was markedly greater at Ca2+ concentrations above -300 nM.
EGF' is a potent modulator of cell growth and other physiological functions (for review see Ref. 1). The EGF receptor is a cell surface glycoprotein with intrinsic, ligand-dependent tyrosine kinase activity (2). The capacity of this receptor to mediate cellular responses to EGF requires functional tyrosine kinase activity (for review see Refs. [1][2][3], suggesting that tyrosine phosphorylation of the receptor and/or exogenous substrates is crucial for signal transduction. A few cellular substrates for the tyrosine kinase activity of the EGF receptor have been identified and characterized in the attempt to define biochemical pathways that transduce the mitogenic signal. These substrate proteins include GTPase-activating protein, a regulator of ras GTPase activity (4,5); phospholipase C-yl (PLC-yl), a 145-kDa phospholipase C isozyme (for review see Ref. 6); and phosphatidylinositol 3-kinase (7).
PLC is the rate-limiting enzyme for phosphatidylinositol 4,5-bisphosphate (PtdIns 4,5-P2) hydrolysis which generates the second messenger molecules inositol 1,4,5-trisphosphate (Ins 1,4,5-P3) and diacylglycerol. A family of PLC isoenzymes (for review see Ref. 8) mediates the generation of these second messenger molecules in response to hormones; however, the biochemical mechanisms of activation of specific PLC isoenzymes are not understood. Tyrosine phosphorylation of PLCy l is believed to be involved in the mechanism by which tyrosine kinase-dependent growth factors stimulate PtdIns 4,5-P2 hydrolysis. Several groups have shown that the addition of EGF or platelet-derived growth factor (PDGF) to cells leads to rapid increases in the amount of phosphotyrosine and phosphoserine on PLC"y1 (9)(10)(11)(12)(13). Transfection of PLC-yl cDNA into NIH/3T3 cells results in overexpression of PLC-Growth Factor Stimulation of Phospholipase C-71 Activity y l protein and enhancement of PDGF-induced Ins 1,4,5-P3 generation (14). Multiple sites of tyrosine phosphorylation on PLC-y1 have been identified for both in vitro (15) and in vivo (16) phosphorylations. The sites of growth factor-stimulated serine phosphorylation have not yet been reported. Other PLC isoenzymes, 8-1 and 6, are not substrates for the EGF receptor tyrosine kinase (11,17). Although many experiments suggested that cellular PLC activity increases after growth factor treatment of cells, only recently has evidence been produced to demonstrate that tyrosine phosphorylationper se affects PLC-yl catalytic activity (18,32,33).
We reported increased catalytic activity of PLC-yl in immunoprecipitates recovered from EGF-treated A-431 cells and demonstrated that in vitro tyrosine phosphorylation of PLCy l by purified EGF receptor activates and dephosphorylation by phosphotyrosine-specific protein phosphatase deactivates the enzyme (18). We now demonstrate stimulation of PLCy l catalytic activity by EGF, PDGF, and acidic fibroblast growth factor (aFGF); characterize the subcellular localization and tyrosine phosphorylation of activated PLC-yl in A-431 cell lysates; and compare, in a PtdIns 4,5-PgTriton X-100 mixed micelle assay, the kinetic parameters of control and activated enzymes.

EXPERIMENTAL PROCEDURES
Materials-EGF was isolated from mouse submaxillary glands as described previously (19). PDGF (recombinant human PDGF-BB homodimer) and aFGF were purchased from Upstate Biotechnology Inc. (Lake Placid, NY). PtdIns 4,5-Pz ammonium salt was obtained from Boehringer Mannheim. Ptd[3H]Ins 4,5-P2 was obtained from Du Pont-New England Nuclear. n-Octyl P-D-glucopyranoside (octyl glucoside) and Pansorbin/Staphylococcus aureus cells were obtained from Calbiochem. Nitrocellulose was obtained from Schleicher & Schuell. lZ6I goat anti-mouse IgG was obtained from ICN Radiochemicals (Irvine, CA). Fetal calf serum and calf serum were products of GIBCO. Plastic culture dishes (35, 60, and 100 mm) and the cell scrapers were purchased from Costar (Cambridge, MA). Plastic culture plates (245 X 245 mm) were obtained from USA Scientific Plastics (Ocala, FL). Teflon homogenizer was obtained from Wheaton Scientific (Millville, NJ). BCA protein assay reagent was produced by Pierce Chemical Co.
Cell Culture-A-431 cells, human foreskin fibroblasts, and NIH/ 3T3 fibroblasts were maintained in Dulbecco's modified Eagle's medium (DMEM) plus 10% (v/v) fetal calf serum and 50 pg/ml gentamycin. A-431 cells were subcultured for experiments by plating lo4 cells/cm2 in 100-mm or 245 X 245-mm culture dishes containing DMEM plus 10% (v/v) calf serum and grown for 3-4 days until 80% confluent. Human foreskin fibroblasts or NIH/3T3 fibroblasts were subcultured by plating lo' cells/cm2 in culture dishes containing DMEM plus 10% calf serum and grown until 100% confluent. Monolayers of fibroblasts were then switched to DMEM containing 1% calf serum and cultured for 48 h before treatment with growth factors.
Preparation of Cell Extracts-A-431 cells cultured in 245 X 245mm dishes were washed twice with DMEM, incubated in 20 ml of DMEM for 20 min, then treated with or without EGF (300-500 ng/ ml) for 5 min at 37 "C. After three rapid washes with ice-cold Ca2+, M e -f r e e phosphate-buffered saline, cells were scraped in the homogenization buffer (0.75 ml/plate; 20 mM Hepes (pH 7.4), 100 mM NaCl, 10% glycerol, 50 mM &glycerophosphate, 0.2 mM N&V04, 10 pg/ml aprotinin, 10 pg/ml leupeptin, and 1 mM EGTA) and lysed with a rapidly spinning Teflon homogenizer (30 strokes, 20 strokes/ min). The lysates were centrifuged at 400,000 X g for 20 min at 4 "C. The supernatants (1.5-2.0 ml/plate, 3-6 mg of protein/ml) were recovered as the cytosolic fraction. Homogenization buffer containing 50 mM Triton X-100 (500 pl) was added to the particulate fraction to solubilize the membrane proteins. After a 20-min incubation at 22 "C, samples were centrifuged at 400,000 X g for 20 min at 4 "C, and the supernatants were recovered as the solubilized membrane fraction (500 pl, 6-9 mg of protein/ml). The cytosolic fraction produced by these procedures contained -65-75% of the total cellular PLC-yl protein.
In some experiments, cells were solubilized with homogenization buffer containing 16 mM Triton X-100 (500 pl/dish). The extracts were centrifuged at 200,000 X g for 30 min at 4 "C. The supernatants were recovered as solubilized cell extracts. Protein concentrations of samples were measured by BCA protein assay reagent using bovine serum albumin as standard.
Immunoprecipitation of PLC-yl and Zmmunoblot Analysis-Cell lysate fractions were incubated with mixed monoclonal antibodies (20) to PLC-yl (1 pg of mixed antibodies/500 pg of cell extract protein) and 60-80 pl (per 1 pg of antibodies) of 10% (v/v) heatinactivated S. aureus suspension precoated with rabbit anti-mouse IgG (2 mg of IgG/ml of S. aureus cells). After incubation (4-16 h) at 4 "C, S. aureus cells were washed twice with homogenization buffer and then resuspended in homogenization buffer with 240 p~ Triton X-100. The relative PLC-yl content in the cell lysate fractions and PLC-yl immunoprecipitates was quantified by immunoblot analysis. The proteins present in each fraction (200 pg, or immunoprecipitated from 200 pg) were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and then transferred to nitrocellulose. After blocking the nitrocellulose with 5% milk solution, the nitrocellulose was incubated with buffer containing excess monoclonal PLC-yl antibodies (20), washed, and then incubated with '261-labeled goat anti-mouse IgG. The PLC-y1 was visualized by exposing the washed and dried nitrocellulose to film. The areas of nitrocellulose containing the PLC-yl were excised and the radioactivity quantitated with a ydetector.
PLC-yl Tyrosine Phosphorylation-The percentage of PLC-y1 molecules in each lysate fraction containing phosphotyrosine was estimated by incubating aliquots of each fraction (500 pg of protein) with 1G2-anti-phosphotyrosine (21) Sepharose matrix (150 pl) at 4 "C overnight. For both cytosol and membrane fractions, this ratio of protein to antibody matrix was nonsaturating (Le. there was excess antibody). The nonadsorbed proteins were recovered, the matrix rapidly washed once with homogenization buffer (1,000 pl), and specifically adsorbed proteins eluted with homogenization buffer (200 pi) containing phenyl phosphate (20 mM). Proteins present in equivalent portions of the nonadsorbed and eluate fractions were separated by gel electrophoresis, and the relative PLC-yl content of each fraction was quantified by immunoblot analysis as described above.
PLC Assay-The hydrolysis of Ptd[3H]Ins 4,5-P2 was measured in a reaction mixture (50 pl) that contained 35 mM NaHZPO4 (pH 6.8), 70 mM KCl, 0.8 mM EGTA, 0.8 mM CaCl2, Ptd[3H]Ins 4,5-P2, and Triton X-100 at the concentrations indicated, and PLC-yl immunoprecipitated from 50-100 pg of cytosol or solubilized membrane fraction. Some experiments included 5 mM octyl glycoside (as described previously (18)) or 2.4 mM deoxycholate in the reaction mixture. An aliquot of PLC-yl immunoprecipitate suspension (5-10 pl) was added to the substrate solution (40-45 pl) and the reaction mixture incubated at 35 'C or for the indicated periods of time, then stopped by transfer to an ice bath with the addition of 100 p1 of 1% (w/v) bovine serum albumin and 250 pl of 10% (w/v) trichloroacetic acid. Precipitates were removed by centrifugation (13,600 X g for 4 min) and the supernatants collected for quantitation of the release of [3H]Ins 1,4,5-P3 by liquid scintillation counting. The radioactivity present in a buffer-only control assay (approximately 1-2% of total) was subtracted from the experimental values. Ptd[3H]Ins 4,tj-P~ hydrolysis was calculated from the release of 13H]Ins 1,4,5-P3 based on the specific radioactivity of Ptd[3H]Ins 4,5-P2 (1,500-2,000 cpm/ nmol). Data presented in the figures represent the average of duplicate assays. All enzyme activity values are based upon the total protein in the aliquot from which the assayed PLC-yl was immunoprecipitated rather than the actual amount of PLC-yl present. All experiments were repeated at least three times using cell lysates prepared on separate occasions, with similar results. Each figure contains data from a single, representative experiment.

RESULTS
Formation of Activated PLC-yl in Intact Cells-Treatment of A-431 cells with EGF yields a detectable increase in the formation of inositol phosphates and the tyrosine phosphorylation of PLC-yl within 1 min of stimulation (22,23). T o determine whether the catalytic activity of PLC-y1 is similarly increased, we evaluated the time course of PLC-yl activation in EGF-treated and control A-431 cells, using PLCy l immunoprecipitates of solubilized extracts and the in vitro PLC assay (Fig. 1). When cells were incubated with EGF at 37 "C, increased PLC-yl catalytic activity was detected within 1.0 min and reached a maximum of 5-fold at 5 min after the addition of EGF. PLC-yl activity recovered from EGFtreated cells then decreased significantly (-25%) between 5 and 10 min after stimulation but remained elevated (3.1-fold) for a t least 60 min. Activation of PLC-yl by the addition of EGF to cells maintained at 4 "C was slower than at 37 "C but continued to increase for 60 min. Both PDGF and aFGF stimulate tyrosine phosphorylation of PLC-y1 in intact cells, whereas insulin does not affect PLC-yl tyrosine phosphorylation (24). To determine whether our observation of EGF-stimulated PLC-yl catalytic activation in A-431 cells can be extended to other growth factors and cell lines, we evaluated the capacity of several growth factors to influence the in vitro catalytic activity of PLC-yl (Table I). NIH/3T3 and human foreskin fibroblasts were stimulated with PDGF, aFGF, EGF, or insulin. After solubilization of cell proteins, PLC-yl was immunoprecipitated, and PLC-yl activity was assayed. PDGF (BB homodimer) increased PLC-yl catalytic activity approximately 3-fold in both cell lines. aFGF activated PLC-y1 about 2-fold in NIH/ 3T3 cells but did not detectably activate PLC-yl in human fibroblasts. EGF activated PLC-yl about 20% in human fibroblasts but did not activate PLC-yl in NIH/3T3 cells which have very low numbers of EGF receptors (25). Insulin did not affect PLC-y1 activity in either cell line. These data support the hypothesis that there is a general, but variable, capacity for mitogenic polypeptides to generate an intracellular signal by phosphorylation and activation of PLC-y1 in different cell types. The different capacities of PDGF, aFGF, and EGF to elicit a detectable response may reflect differences in phosphorylation of PLC-yl, perhaps caused by differences in receptor concentration on the cell surface. Insulin, a nonmitogenic polypeptide, is incapable of signaling by this pathway even when the receptor is present at high concentration on the cell surface (24).
Detergent Requirement for Detection of PLC-yl Activation-Of the many factors that influence the catalytic activity detected in a PLC assay, the detergent for solubilization of the substrate appears to be particularly important to the detection of tyrosine phosphorylation-dependent modulation of PLC-yl activity. Although PLC-yl activation by tyrosine phosphorylation was not detected in our initial studies, which employed substrate solubilized in either octyl glucoside (16) or deoxycholate (15), we subsequently observed phosphorylation-dependent changes in PLC-y1 activity when Triton X-100 was used (18), alone or together with octyl glycoside, to prepare the substrate. Fig. 2 demonstrates the critical influence of Triton X-100 concentration in the PLC assay, when the substrate (Ptd[3H]Ins 4,5-P2) is solubilized in the commonly used detergents.
In Fig. 2 A , Ptd[3H]Ins 4,5-P2 was solubilized in octyl glucoside, and increasing concentrations of Triton X-100 (0-2.4 mM) were included in the assay. The final concentration of octyl glycoside in the assay is 5 mM. A substantial and differential influence of Triton X-100 on the activity of PLC- NIH/3T3 cells or human fibroblasts were incubated in the absence or presence of PDGF (50 ng/ml), aFGF (50 ng/ml), insulin (100 ng/ ml), or EGF (50 ng/ml) for 5 min at 37 "C. Cells were washed and solubilized in homogenization buffer containing 16 mM Triton X-100. PLC-71 was immunoprecipitated, and the activity was measured with a reaction mixture containing 200 p M Ptd[3H]Ins 4,5-P,, 5 mM octyl glucoside, and 0.08 mM Triton X-100 (18). The data represent the PLC-v1 activitv recovered from an entire dish of cells.  y l immunoprecipitated from EGF-treated and control cells was detected. Low concentrations of Triton X-100 (up to approximately 0.24 mM) increased the PLC-y1 activity detected in the immunoprecipitates from both EGF-treated and control cells. Higher concentrations of Triton X-100 (0.48-2.4 mM) produced a selective decrease in the PLC-yl activity from control cells. Thus, a 2-5-fold increase in PLC-y1 activity from EGF-treated cells relative to PLC-y1 activity from control cells is detectable at Triton X-100 concentrations above -0.24 mM. Fig. 2B shows the results of a similar assay, performed with substrate solubilized in Triton X-100 (without octyl glycoside). Low concentrations of Triton X-100 (0.08-0.48 mM) decreased the activity of PLC-yl from both EGF-treated and control cells. However, at higher concentrations of Triton X-100 (>0.48 mM) a differential influence on enzyme activity was observed again. Thus, a 2-6-fold activation of PLC-y1 from EGF-treated cells relative to control cells was detectable at higher Triton X-100 concentrations.
In Fig. 2C, Ptd[3H]Ins 4,5-P2 was solubilized in deoxycholate, the detergent most commonly used for in uitro PLC assays, and PLC-yl activity was measured in the presence of increasing concentrations of Triton X-100 (0-2.4 mM). In the presence of 2.4 mM deoxycholate, Triton X-100, at concentrations above 0.16 mM, produced equivalent decreases of PLCy l activity from both control and EGF-treated cells (panel C). Thus, EGF activation of PLC-y1 was not detectable in the assay containing this concentration of deoxycholate, a concentration used in previous studies of PLC-yl activity (15,39).
Actiuated PLC-yl in Cytosolic and Membrane Fractions-Before growth factor stimulation, the majority of PLC-yl is present in the soluble (cytosol) fraction of cell lysates, whereas a smaller portion is detected in the particulate (membrane) fraction (26). After EGF stimulation, a redistribution of PLCy l to the membrane fraction has been documented (26). In that study, cells were gently lysed (through a needle) in a homogenization buffer containing divalent cation ( M e ) . In the present study, cell lysates were prepared by vigorous homogenization in a buffer containing a chelator (EGTA) and no divalent cation. A-431 cells were treated with or without EGF for 5 min at 37 "C, lysed, and the cytosol and membrane fractions separated by centrifugation. After solubilization of protein in the membrane fraction with Triton X-100, PLCy l was quantitatively immunoprecipitated from the cytosol and membrane fractions with specific monoclonal antibodies (20). The relative PLC-yl protein content in the immunoprecipitates was quantified after gel electrophoresis by anti-PLCy l immunoblot analysis, and the PLC-y1 catalytic activity was quantified by PLC assay. Fig. 3 demonstrates a comparison of the fractionation of PLC-yl from control and EGF-treated cells with regard to the subcellular location of protein and enzyme activity, the relative PLC-yl tyrosine phosphorylation in each subcellular fraction, and specific enzyme activity in each fraction. Under these homogenization conditions, approximately one-third of the PLC-yl protein was present in the particulate (membrane) fraction, and two-thirds was located in the soluble (cytosol) fraction from both control and EGF-treated cells (Fig. 3A). Assay of PLC activity in PLC-yl immunoprecipitates from each fraction revealed a similar distribution: approximately two-thirds of total cell lysate PLC-yl activity was in the soluble fraction, and one-third remained in the particulate fraction (Fig. 3B). Although the relative proportion of PLC-yl activity in membrane and cytosol fractions was similar for control and EGF-treated cell lysates, the absolute PLC-yl activity was 2-%fold greater in the EGFtreated preparation. When PLC-yl activities were normalized for the amount of total protein present in each fraction (soluble or particulate) to yield apparent PLC-yl specific activities for each fraction, a 2-3-fold increase in PLC-yl specific activity was observed in both the soluble and particulate fractions of lysates obtained from EGF-stimulated cells (Fig. 3C).
Because PLC-yl present in both particulate and soluble fractions from cell lysates is activated after EGF treatment, we measured the proportion of PLC-yl protein that is phosphorylated on tyrosine in each fraction (Fig. 30). Cytosol and (solubilized) membrane fractions were adsorbed to excess anti-phosphotyrosine matrix. After recovery of nonadsorbed and specifically adsorbed proteins from the matrix, the relative PLC-yl protein content in each sample was quantified, after gel electrophoresis, by anti-PLC-yl immunoblot analysis. In unstimulated A-431 cells, a small fraction (2-3%) of PLC-yl present in both the soluble and the particulate fractions specifically adsorbed to the anti-phosphotyrosine matrix. In EGF-treated cells, approximately 20% of the particulate PLC-yl, and 54% of the soluble PLC-yl bound to antiphosphotyrosine. In the soluble fraction, this represents the EGF (500 ng/ml) for 5 min at 37 "C. Cell lysates were centrifuged to yield soluble (cytosol) and particulate (membrane) fractions, and then PLC-yl protein and activity in each fraction were measured, as described under "Experimental Procedures." Panel A demonstrates the relative PLC-yl protein content (percent of total lysate PLC-yl) of the cytosol and membrane fractions, measured by immunoblot analysis. Panel B demonstrates the relative PLC-yl activity (percent of total lysate PLC-yl activity) of the same cytosol and membrane fractions as in panel A. The PLC assay was performed at 35 "C with Ptd[3H]Ins 4,5-P2 (200 p~) and Triton X-100 (0.80 mM above the critical micelle concentration), 1.0 p M free Ca2+, and PLC-yl immunoprecipitated from 50 pg of protein from each fraction. The total PLC-yl activity present in each fraction, using this assay, was calculated to be: (-)EGF cytosol, 6.77 nmol/min; (+)EGF cytosol, 18.8 nmol/min; (-)EGF membrane, 3.90 nmol/min; (+)EGF membrane, 8.60 nmol/min. Panel C demonstrates the PLC-y1 specific activity in each fraction, calculated by dividing the total PLC-yl activity in each fraction (panel B ) by the total protein content of that fraction. The total protein content in each fraction was calculated to be: (+/-)EGF cytosol, 18.0 mg; (+/-)EGF membrane, 11.5 mg. Panel D demonstrates the relative tyrosine phosphorylation of PLC-yl in each fraction, i.e. the percentage of PLC-yl in each fraction which is specifically adsorbed to anti-phosphotyrosine matrix. The relative amount of PLC-yl from each fraction which was adsorbed or nonadsorbed to the matrix was measured by subsequent immunoblot analysis. fraction of PLC-yl molecules containing phosphotyrosine. PLC-yl in the particulate fraction may adsorb to the antiphosphotyrosine matrix directly or indirectly as a result of association with the EGF receptor (10). However, we have found very little PLC-yl co-precipitating with the EGF receptor in A-431 cells under a variety of isolation conditions (data not shown). If 20% of the PLC-yl molecules in the particulate fraction contain phosphotyrosine, it is surprising that the degree of activation is similar to that observed in the cytosol fraction, in which 50% of the PLC-yl molecules are tyrosinephosphorylated. Several possible explanations for this difference exist, including: 1) unrecognized activating factors in the membrane fraction or 2) differences in the phosphorylation of individual tyrosine residues (32).
Gel Chromatographic Analysis of Substrate Micelles-Our previous experiments (Fig. 2) indicate that the mode of presentation of the substrate phospholipid to the enzyme has a significant influence on the detection on the tyrosine phosphorylation-dependent modulation of PLC-y 1 activity in uitro. Because the differential influence of Triton X-100 on the activity of control and phosphorylated PLC-yl is observed only with Triton X-100 concentrations greater than its nominal critical micelle concentration (0.24 mM), detection of PLC-yl activation in vitro may require presentation of PtdIns 4,5-P2 as a mixed micelle with Triton X-100. We began the evaluation of the mechanistic basis of the apparent PLC-y1 activation in this assay with gel chromatographic analysis of PtdIns 4,5-P2:Triton X-100 micelles followed by kinetic analysis of hydrolysis of the substrate by control and activated PLC-yl. The coelution of PtdIns 4,ij-P~ and Triton X-100 from the gel filtration column suggests that the detergent and phospholipid substrate combine to form a mixed micelle. Alternatively, they may individually form micelles that coincidentally have similar sizes. Because the elution volume does not vary detectably with the molar ratio of substrate to detergent, the size of the presumed mixed micelles appears to be invariant.
Kinetic Analysis of PLC-yl in a Mixed Micelle Assay-Neither the mechanism of PLC-y1 hydrolysis of PtdIns 4,5-P2 nor the role of tyrosine phosphate in PLC-yl activation in vivo or in vitro is well understood. Because the presence of Triton X-100 allows a clearly measurable difference between the in vitro activities of control and tyrosine-phosphorylated PLC-71, we evaluated and compared the kinetic parameters of control and activated (by treatment of intact cells with EGF) PLC-yl using a Ptd[3H]Ins 4,5-P2:Triton X-100 mixed micelle assay.
PLC-yl was quantitatively immunoprecipitated from the cytosol of untreated and EGF-treated A-431 cells. Western blot analysis confirmed that a similar amount of PLC-yl was recovered in the immunoprecipitates regardless of EGF treatment (data not shown). Control experiments verified that the hydrolysis of Ptd subsequent kinetic analysis of the influence of EGF treatment on the velocity of PLC-71-catalyzed substrate hydrolysis. The maximal substrate concentration evaluated was limited by the method of substrate preparation. The minimal substrate concentration evaluated (100 p~) was limited by the assumed Triton x-100 non-micellar concentration (0.24 mM).
Ptd[3H]Ins 4,5-P2 hydrolysis by control and activated PLCy l was analyzed with respect to both the bulk concentration of the substrate and relative (ie. mol fraction)' concentration of the substrate. This analysis follows the precedents established for kinetic analysis of phospholipase AP activity by Dennis and colleagues (27)(28)(29)(30)(31). Double reciprocal plot analyses of the relationship between substrate bulk concentration and reaction velocity for control and activated PLC-71 are shown, respectively, in Fig. 4, A and B. (Note that the y axis scales differ between A and B.) Each set of data points represents the reaction velocities observed with increasing substrate bulk concentrations at different molar ratios of Ptd[3H]Ins 4,5-P2 to Triton X-100. As the molar concentration of PtdIns 4,5-P2 decreased, the relative velocity of the reaction decreased; this decrease is a result of surface dilution of the substrate (28,29). It is possible to determine the kinetic parameters K,,,, V,,,,,, and K, for the system by plotting the intercept and the slope of each line in Fig. 4, A and B, versus the mol fraction. The results of the kinetic analysis are summarized in Table 11. This secondary analysis of the data presented in Fig. 4B revealed that the plot of intercept or slope versus mol fraction for the EGF-activated enzyme is linear (Fig. 4, C and D). Extrapolation of the line in Fig. 4C to the y intercept allowed the evaluation of V,,,, maximal velocity at infinite mol fraction of PtdIns 4,5-P2, which was 5.5 nmol/min. mg. K, was evaluated from the relationship of slope = K,,,/V,., of the line in Fig. 4C. K,,, is

TABLE I1
Kinetic parameters of PLC-71 isolated from control and EGFstimulnted cells Kinetic constants were calculated from Fig. 4, A-D, as explained under "Results."
The plot of intercept or slope versus mol fraction for the control enzyme deviated from linearity (Fig. 4, C and D). This deviation from linearity suggested that the kinetics for the control enzyme are more complex than the activated enzyme. Evaluation of kinetic parameters (Table 11) for control PLC-71 is difficult from the secondary analyses plots because the relationship of slope or intercept versus mol fraction was not linear. At a substrate mol fraction of 0.07, there is a 4-fold difference in reaction velocities of control and activated enzymes. As the mol fraction increases, the lines converge so that at 0.33 substrate mol fraction the reaction velocities are equivalent (Fig. 4C). A Hill plot of the data for the control enzyme allowed the estimation of the substrate concentration where half-maximal activity is observed, or So.a (0.95 mol fraction PtdIns 4,5-P2; data not shown). K, for the control enzyme was estimated from the double reciprocal plots in Fig.   4 A , with Ks being equivalent to the -l/x value of the point at which all lines converge. K, was determined to be approximately 1.5 mM.
A double reciprocal plot of the reaction velocity versus mol fraction also deviated from linearity for the control enzyme (Fig. 5A). This behavior suggests that the control enzyme displays apparent allosteric kinetics, whereas the EGF-acti- vated enzyme does not (Fig. 5B). A Hill plot of the data in Fig. 5A (inset) revealed that the cooperativity index did not change as the substrate concentration increased, therefore allosterism cannot be overcome by increasing the substrate concentration. Kinetic parameters calculated from Fig. 5B (EGF-activated enzyme) were similar to those calculated from Fig. 4B (data not shown).
Comparison of PLC-yl Activity in Membrane and Cytosol Fractions-The data in Fig. 4 were obtained using PLC-yl from the cytosol fraction of control and EGF-treated cells. Because a portion (-33%) of PLC-yl is present in the membrane fraction and is activated in EGF-treated cells (Fig. 3), we directly compared the activity of PLC-y1 from membrane and cytosol fractions, using increasing concentrations of Ptd[3H]Ins 4,5-P2 at a 0.33 mol fraction. At this relative substrate concentration, PLC-y1 from cytosol of control and activated PLC-71 had similar relative velocities (Fig. 4). A double reciprocal plot of the reaction velocities versus the substrate bulk concentration is shown in Fig. 6. Extrapolation of the data revealed similar relative velocities (-5.00 nmol/ mine mg) values for each enzyme preparation in this assay. These data suggest that in this assay PLC-yl activity recovered from membrane fraction performs similarly to PLCy l activity recovered from the cytosol fraction. Extrapolation also revealed that membrane and cytosolic enzyme from control or EGF-treated cells intersected the x axis at a common point.
Stimulation of PLC-yl Activity by CP-The catalytic activity of all mammalian PLC isozymes is critically dependent on the concentration of Ca2+ (8). We examined the CaZ+dependent stimulation of the catalytic activity of PLC-yl immunoprecipitated from the cytosol of control and EGFtreated A-431 cells using 200 PM Ptd[3H]Ins 4,5-P2 at a substrate:detergent molar fraction of 0.33. We observed (Fig.  7 A ) that both control and activated PLC-yl are stimulated by Ca2+ concentrations in the nanomolar range. PLC-y1 from EGF-treated cells demonstrated a maximal activity of -1.9 nmol/min. mg in this assay, whereas the control enzyme maximal activity was -0.6 nmol/min. mg. Although PLC-yl from EGF-treated cells displays higher catalytic activity at each Ca2+ concentration than the control enzyme, the degree of activation is greater with Ca2+ concentrations higher than -300 nM. The data in Fig. 7A were reevaluated with regard to the relative stimulation (percent maximal) of activity at each Ca2+ concentration. Fig. 7B demonstrates that the halfmaximal stimulation of control enzyme was achieved with a Ca2+ concentration of -56 nM, whereas activated PLC-yl required -316 nM Ca2+ for half-maximal stimulation. When the experiment was performed with a mol fraction of 0.2, a similar result was obtained (data not shown).

DISCUSSION
Using a Triton X-100-based PLC assay, we have characterized several parameters of PLC-yl activation. The rapid time course of PLC-yl activation demonstrated in Fig. 1 is indistinguishable from the time course of EGF-stimulated PLCy l tyrosine phosphorylation documented previously (9). Other growth factors known to induce PLC-yl tyrosine phosphorylation likewise rapidly increase PLC-y1 catalytic activity (Table I). Importantly, these data show that PLC-y1 catalytic activation is not restricted to EGF-stimulated A-431 cells (a carcinoma that overexpresses the EGF receptor) but occurs in nontransformed cells expressing normal complements of growth factor receptors. The capacity to detect tyrosine phosphorylation-dependent modulation of PLC-y1 in this assay requires the presence of Triton X-100 at a micellar concentration (Fig. 2) and is compatible with the presence of octyl glycoside but not the ionic detergent deoxycholate, although a variety of deoxycholate concentrations and conditions have not been tested.
Although previous studies have characterized various PLC activities in vitro (34)(35)(36)(37)(38)(39)(40)(41), little information is available, particularly in comparison with studies of phospholipase Az (27)(28)(29)(30)(31), regarding either the mechanism of PLC-mediated hydrolysis of PtdIns 4,5-Pz or the influence of any hormonally mediated activation steps upon this reaction mechanism. Our analysis attempts to follow the theory and experimental conditions outlined by Dennis and his colleagues in their investigations of the mechanism of phospholipase Az catalytic activities (27)(28)(29)(30). The primary goals in this study are to evaluate the kinetic parameters of control and activated PLCy l under our experimental conditions and to assess the influence of tyrosine phosphorylation on the catalytic activity of this enzyme. To our knowledge, this is the first example of such a study for a tyrosine kinase substrate (other than the autophosphorylation activity of receptor tyrosine kinases).
The kinetic analysis was performed with PLC-y1 immunoprecipitated from the soluble fraction of cell lysates. We have termed this PLC-yl pool "cytosolic" as an operational designation. Because a substantial portion of total cellular PLC-yl remains in the particulate ("membrane") fraction of cell lysates (Fig. 3, A and B ) , we compared its activity with the cytosolic PLC-yl. Although the tyrosine phosphorylation of PLC-y1 tightly associated with the membrane is significantly less than the tyrosine phosphorylation of soluble PLCy l (Fig. 3 0 ) , both fractions have higher activity in EGFtreated than control cells (Fig. 3C). Comparison of PLC-yl from control and EGF-treated cell lysates suggests that activated enzyme from cytosolic or membrane preparations performs similarly in the assay and likewise for the control enzyme (Fig. 6).
We performed the kinetic analysis of PLC-yl activities to evaluate the basis for the differential rate of hydrolysis of substrate by control and activated PLC-yl using a mixed micelle assay system. In this assay system V,,, is equivalent to maximal velocity at infinite mol fraction of PtdIns 4,5-P2, K,,, is the concentration that yields half-maximal velocity, and K, is equivalent to the dissociation constant for PLC-71 binding to the micelle. The reaction velocity of both activated and control PLC-y1 increased as bulk substrate concentration in the reaction mixture increased from 100 to 800 p~ (Fig. 4,  A and B ) . The substrate bulk concentration-dependent increase in catalytic activity was observed at each PtdIns 4,5-P2 ratio tested (Fig. 4, A and B ) . Our analysis of the kinetic data included not only double reciprocal analysis (Fig. 4, A  and B, and Fig. 5, A and B ) but also secondary analysis of the apparent kinetic constants (Fig. 4, C and D).  Table 11). Because the secondary analysis of the control enzyme deviated from linearity, it is difficult to determine the kinetic parameters. Examination of the secondary plot of intercept uersus mol fraction (Fig. 4C) revealed that the line for control enzyme approached that of the EGF-activated enzyme. It is possible that the V,,, of the control enzyme may be the same as the activated enzyme. Determining the exact V,,, for the control enzyme experimentally is difficult because at mol fractions greater than 0.33 the exact nature of the micelle is not known and may no longer be uniform. When the data from Fig. 4A were transformed using the Hill equation, it was possible to determine an SOs for the control enzyme. This parameter was determined to be 0.95 mol fraction of PtdIns 4,s-P~. K, was evaluated from the double reciprocal plot (Fig. 4A). It is equivalent to -l/x value of the point at which all lines intersect, this point being 1.5 mM. From our kinetic analysis, EGF stimulation of PLC-yl activity resulted in a 7-fold reduction in the K, (Table 11). Kinetic analysis also revealed that EGF stimulation of PLC-yl activity had a positive cooperative effect on enzyme association with PtdIns 4,5-P2 mixed micelles.
In this paper we attempted to analyze PLC-yl kinetics using the ordered sequential model outlined by Dennis and co-workers for phospholipase A2 (27)(28)(29)(30)(31). Inherent in this model are the assumptions that the transition of the enzyme to the interface of the micelle occurs before the hydrolysis of the substrate and that the binding sites for the micelle and the binding site for the substrate are independent and noncooperative, For the EGF-activated enzyme, the assumptions apparently hold true, but for the control enzyme the curves of reaction velocity uersw mol fraction of PtdIns 4,5-P2 appear sigmoidal in nature. Therefore, that the binding site for the micelle and the binding site for PtdIns 4,5-P2 are independent and noncooperative is not true. The sigmoidal behavior displayed by the control enzyme could result from a greater sensitivity to Triton inhibition than the activated enzyme. It is also possible that the apparent cooperativity is caused by the binding of PtdIns 4,5-P2 to a nonsubstrate site on the enzyme or that PLC-yl is a multisubunit or multidomain enzyme. The apparent cooperativity resulting from the binding of lipid to a nonsubstrate site on the protein has been seen with many membrane-associated proteins (29,42,43), although it is not clear why the activated enzyme does not display some degree of cooperativity. The possibility of PLCy l being a multisubunit or multidomain protein is intriguing. If this is the case, then not only does tyrosine phosphorylation shift the equilibrium of PLC-y1 so that it is now more likely to associate with micelles by lowering the K, 7-fold (1.5 to 0.22 mM), but it can act as a positive allosteric modifier of PLC-yl activity. The allosteric behavior of PLC-y1 may explain why no difference in activity caused by tyrosine phosphorylation is seen when pure liposomes containing a high mol fraction of PtdIns 4,fj-P~ are used (15,33). This hypothesis may also explain why a difference in activity resulting from tyrosine phosphorylation is seen when PLCy l activity is assayed in the presence of profilin (33). Profilin binding to PtdIns 4,5-P2 effectively lowers the PtdIns 4,5-P2 concentration in the liposome (45). At this effectively lowered PtdIns 4,5-P2 concentration the control enzyme cannot bind PtdIns 4,5-P2 as readily as the activated enzyme and therefore displays an overall lower rate of PtdIns 4,5-P2 hydrolysis.
A major consequence of hormone-stimulated PLC activity is an increase in the intracellular free Ca" level. Whereas Ca2+ affects the activities of many cellular proteins, it is also an important regulator of PLC activities. We demonstrated previously that EGF stimulation of PLC activity in A-431 cells is biphasic (23). The initial phase is brief, correlates with induction of PLC-yl tyrosine phosphorylation, and occurs in the absence of extracellular Ca2+. The second phase of EGFstimulated inositol phosphate formation is prolonged and, despite high levels of PLC-yl tyrosine phosphorylation, occurs only in the presence of extracellular Ca2+. Other research has documented a biphasic rise in intracellular Ca2+ concentration in EGF-treated cells which depended largely, but not entirely, on the presence of extracellular Ca2+ (44, 46-48). The precise relationships between Ca2+ mobilization and regulation of PLC-y1 activity in EGF-stimulated A-431 cells remain to be clarified. We investigated the influence of Ca2+ concentrations on PLC-yl activities in the mixed micelle assay. The catalytic activity of PLC-yl from EGF-treated cells was greater than control PLC-y1 activity at each Ca2+ concentration tested. However, at the higher Ca2+ concentrations (above -300 nM) the relative stimulation of activated PLC-yl was markedly enhanced. It is possible that this result reflects catalytic activation of a portion of the PLC-y1 protein molecules which requires a high Ca2+ concentration. Alter-natively, at the higher Ca2+ concentrations, activated PLCyl, but not control PLC-yl, may bind with higher affinity to the substrate mixed micelles. This interpretation would be consistent with our interpretation of kinetic data in Fig. 4 that Triton X-100 does not affect activated PLC-yl association with mixed micelles. (Note that the kinetic analysis was performed with a Ca2+ concentration of -1.0 p~. ) Finally, the in vitro Ca2+ stimulation of activated PLC-yl is consistent with our previous observations that extracellular Ca2+ is required to maintain the second phase of EGF stimulation of PLC activity in A-431 cells.