Activation of Phospholipase C δ1 through C2 Domain by a Ca2+-Enzyme-Phosphatidylserine Ternary Complex*

The concentration of free Ca2+and the composition of nonsubstrate phospholipids profoundly affect the activity of phospholipase C δ1 (PLCδ1). The rate of PLCδ1 hydrolysis of phosphatidylinositol 4,5-bisphosphate was stimulated 20-fold by phosphatidylserine (PS), 4-fold by phosphatidic acid (PA), and not at all by phosphatidylethanolamine or phosphatidylcholine (PC). PS reduced the Ca2+ concentration required for half-maximal activation of PLCδ1 from 5.4 to 0.5 μm. In the presence of Ca2+, PLCδ1 specifically bound to PS/PC but not to PA/PC vesicles in a dose-dependent and saturable manner. Ca2+ also bound to PLCδ1 and required the presence of PS/PC vesicles but not PA/PC vesicles. The free Ca2+concentration required for half-maximal Ca2+ binding was estimated to be 8 μm. Surface dilution kinetic analysis revealed that the K m was reduced 20-fold by the presence of 25 mol % PS, whereas V max andK d were unaffected. Deletion of amino acid residues 646–654 from the C2 domain of PLCδ1 impaired Ca2+binding and reduced its stimulation and binding by PS. Taken together, the results suggest that the formation of an enzyme-Ca2+-PS ternary complex through the C2 domain increases the affinity for substrate and consequently leads to enzyme activation.

Approximately 12 distinct isoforms of phospholipase C catalyze the Ca 2ϩ -dependent hydrolysis of phosphatidylinositol 4,5bisphosphate (PIP 2 ) 1 to yield the second messengers inositol trisphosphate (IP 3 ) and diacylglycerol (1,2). This constitutes one of the major pathways for receptor-coupled signaling at the plasma membrane of most eucaryotic cells. Three families of PLC isoforms have been described in mammals: PLC␤, PLC␦, and PLC␥ (2). The members of each family are highly homologous to one another at the amino acid sequence level, but little identity exists between members of different families (2). Three exceptions to this divergence in structure are the catalytic domain, the C2 domain, and the N-terminal pleckstrin homology (PH) domain (1,3).
Among the initial steps in activation of PLC is a transloca-tion to the plasma membrane. The enzyme binds to the lipidwater interface via the noncatalytic lipid binding PH domain, which is located near the N terminus of PLC␦1 and binds multiple phosphoinositides such as PIP 2 and IP 3 (4 -9). This domain allows the enzyme to catalyze the hydrolysis of many substrate molecules without falling off the interface, a process referred to as processive catalysis (5,9). Because PH domain binds IP 3 tightly, it also could function as a feedback regulator of catalysis.
Although much is known about the function of the PH domain in PLC␦1, very little is known about another lipid binding motif, the C2 domain. The C2 domain comprises approximately 130 amino acid residues and has been found in nearly 100 signaling molecules (10). The C2 domain was first identified in protein kinase C, and its function was implicated in Ca 2ϩ -dependent phospholipid interactions (11). Structural studies estimate three to four divalent metal binding sites in the C2 domain of PLC␦1 (12). The C2 domain in the C terminus of PLC␦1 is essential for catalysis, because partial deletion of the C2 domain results in an inactive enzyme (13). However, the molecular mechanism by which C2 domain functions in PLC␦1 catalysis still awaits further investigation.
PLC is a prototype for enzymes that operate at an interface. As for other membrane-associated enzymes, the ability of PLC to catalyze the hydrolysis of PI or PIP 2 is influenced remarkably by the presence of nonsubstrate phospholipids in the membrane (14 -19). The molecular mechanisms by which nonsubstrate phospholipid affects PLC activity is largely unknown. Nonsubstrate phospholipids could alter the structure of the PLC-membrane interface or the net charge of the interface or could promote a specific interaction between enzyme and interface.
This report examines the effect of nonsubstrate phospholipids on the activity of PLC␦1 and examines the role of Ca 2ϩ in this response. These experiments revealed that PLC␦1 was specifically stimulated by PS. PS stimulates the substrate affinity of the enzyme by virtue of its ability to bind the PLC␦1 via C2 domain in a Ca 2ϩ -dependent manner. We propose that Ca 2ϩ regulates PLC␦1 activity by promoting the formation of an enzyme-PS-Ca 2ϩ ternary complex, which leads to activation via a 20-fold reduction in the K m for substrate.

EXPERIMENTAL PROCEDURES
Materials-The expression vector pRSETA was from Invitrogen. To express PLC␦1 under the control of the T7 promoter, the coding sequence for PLC␦1 was cloned into pRSETA. The resulting expression construct (pRSETAplc) was transformed into the Escherichia coli strain BL21(DE3)pLys (Novagen), and the protein was isolated and purified as described previously (20). Phosphatidylethanolamine, PA, PC, and PS were obtained from Avanti Polar Lipids Inc. PIP 2 and dodecyl maltoside was obtained from Calbiochem. The double point mutant PLC␦1 (E341G,E390G) in which the calcium-binding residues Glu-341 and Glu-390 in the cleavage center were both changed to Gly. In the C2 loop deletion mutant (⌬646 -654), residues 646 -654 implicated in the divalent metal binding were deleted. These mutant forms of PLC␦1 were constructed, expressed, and purified as described previously (20,21).
Phospholipid Binding Assay-Phospholipid vesicles composed of PS/PC or PA/PC mixtures were prepared as described by Mueller et al. (22) with slight modifications (20,21). A dry phospholipid film was formed by slowly blowing 0.25 ml of chloroform/methanol (2:1 v/v) containing mixed lipids (300 nmol or the indicated concentration of each of the indicated phospholipids) under a stream of nitrogen followed by freeze-drying under vacuum for 4 h. The phospholipid film was hydrated under nitrogen with 0.5 ml of nitrogen-aerated 0.18 M sucrose for 18 h at 4°C followed by mixing with an equal volume of distilled H 2 O. Vesicles were isolated from the pellet by centrifuging the hydrated phospholipids at 1200 ϫ g for 20 min. The phospholipid vesicles were washed once with 1 ml of 50 mM HEPES, pH 7.0, 100 mM KCl, 2 mM EGTA (binding buffer) and resuspended in 0.5 ml of the same buffer.
Centrifugation Binding Assay-The binding of PLC␦1 to phospholipid vesicles was estimated by a centrifugation assay (20,23). The free Ca 2ϩ concentration was calculated according to Fabiato and Fabiato (24). To perform the assay, 1 g of enzyme was incubated with 200 l of 50 mM HEPES, pH 7.0, 100 mM KCl, 2 mM EGTA, 150 M phospholipid vesicle, and various concentrations of CaCl 2 to yield the indicated concentration of free calcium. The reaction was carried out at 30°C for 15 min. The free and bound PLC␦1 were separated by sedimentation at 50,000 ϫ g for 30 min. An equal proportion of the supernatant and pellet fractions were resolved by 12% SDS-polyacrylamide gel electrophoresis, and the amount of PLC␦1 in each fraction was estimated by Western blotting analysis.
Calcium Binding Measurements-Calcium binding was determined by a nitrocellulose membrane binding assay similar to that described by Nakamura (25) and Kawasaki et al. (26). 150 M PS/PC or PA/PC phospholipid vesicles containing the indicated mole fractions of lipids were incubated with PLC␦1 (final concentration, 0.15 M) in 100 l of 50 mM HEPES, pH 7.0, 100 mM KCl, 10 M EGTA containing 5-120 M of 45 CaCl 2 (total cpm, 1.5 ϫ 10 7 ) to yield the indicated concentration of free Ca 2ϩ . Millipore polyvinylidene difluoride Immobilon-P transfer membrane was immersed in methanol and washed three times with 10 ml of 50 mM HEPES, pH 7.0, 100 mM KCl, and 10 mM EGTA. The membrane was mounted on a 96-well Bio-Dot filtration apparatus according to the manufacturer's instructions (Bio-Rad). After incubating at 30°C for 30 min, the reaction was filtered through the membrane at a constant flow rate of 0.6 -1 ml/min. Each well was washed six times with 100 l of ice-cold 50 mM HEPES, pH 7.0, 100 mM KCl, 10 mM EGTA. The membrane was cut into sections corresponding to each sample, and the retention of 45 Ca 2ϩ on the membrane was determined by liquid scintillation counting. The amount of 45 Ca 2ϩ bound to the enzyme was determined in the presence or absence of vesicles containing the indicated composition of phospholipids. Calcium binding to phospholipid vesicles in the absence of enzyme was determined under the same experimental conditions. The amount of 45 Ca 2ϩ binding directly to phospholipid vesicles was relatively low but consistent with published apparent K d values for Ca 2ϩ -PS interactions (27,28). The calcium bound to membranes or to phospholipid vesicles was considered nonspecific binding and was subtracted from the amount of Ca 2ϩ bound by samples containing protein plus phospholipid. PIP 2 Hydrolysis Assay-PIP 2 hydrolysis in dodecyl maltoside/PIP 2 mixed micelles was performed in a manner similar to that described by Cifuentes et al. (9) with slight modifications. In brief, the indicated amount of PIP 2 /[ 3 H]PIP 2 (4 ϫ 10 5 cpm) in chloroform/methanol (19:1) in the presence or absence of the indicated phospholipids was dried under a stream of N 2 and lyophilized for 30 min. Lipids were solubilized by probe sonication in 0.95 ml of 50 mM HEPES, pH 7.0, 100 mM NaCl, 2 mM EGTA plus the indicated concentration of dodecyl maltoside. Bovine serum albumin in the same buffer was added to 500 g/ml. PLC activity was determined as a function of the concentration of substrate by keeping the total concentration of nonsubstrate phospholipid plus dodecyl maltoside at 500 M and varying the mole fraction of PIP 2 . The reaction at 30°C was initiated by adding various concentration of CaCl 2 to yield the indicated concentration of free calcium. The reaction was continued for 1-5 min and was stopped by adding 0.34 ml of 10% ice-cold trichloroacetic acid and 0.17 ml of bovine serum albumin (10 mg/ml). After incubation on ice for 15 min, the unhydrolyzed [ 3 H]PIP 2 (pellet) was separated from [ 3 H]IP 3 (supernatant) by centrifugation at 2000 ϫ g for 10 min at 4°C.
Analysis of Kinetic Data-Surface dilution kinetics were employed to study PIP 2 hydrolysis catalyzed by PLC␦1. Case III conditions previously described for phospholipase A 2 (29) and PLC␤ (30) were employed. In Case III conditions, the total concentration of diluent deter-gent (dodecyl maltoside plus nonsubstrate phospholipid) was fixed, and the PLC activity was measured with increasing concentrations of substrate. A dual phospholipid binding model of catalysis (Equations 1 and 2) (29) was used to analyze the kinetic data.
This model takes into account the fact that the reaction catalyzed by PLC␦1 occurs at the water-lipid interface of the phosphoinositide/dodecyl maltoside mixed micelle. Initial binding of the enzyme to the waterlipid interface of the micelle is described by the micellar dissociation constant, K s ϭk Ϫ1 /k ϩ1 (molar unit). This constant is dependent on both the total enzyme concentration and the total substrate. The binding of the second lipid molecule and the subsequent catalysis by PLC␦1 is described by the interfacial Michaelis constant, K m ϭ (k Ϫ2 ϩ k 3 )/k ϩ2 (mole fraction, unitless). Initial rates of catalysis (v) as a function of total concentration of PIP 2 in the vesicle with a fixed concentration of diluent nonsubstrate phospholipids and dodecylmaltoside (T o ) were fitted using Equation 3 (29) to obtain the values of V max , K s , and where the absolute rate (V max ) occurs at an infinite substrate concentration and the saturated substrate mole fraction (S o ) is the total substrate concentration.

Effect of Nonsubstrate Phospholipids on the Ca 2ϩ
Dependence of PLC␦1 Catalysis-The concentration of free Ca 2ϩ and nonsubstrate phospholipids can greatly influence the activity of PLC␦1. As shown in Fig. 1A, anionic phospholipids such as PS and PA stimulated PLC␦1 hydrolysis of PIP 2 by a factor of 20 and 4, respectively, whereas cationic phospholipids such as phosphatidylethanolamine and PC had no effect or were inhibitory. Although PS and PA are both anionic phospholipids, PS stimulation was much greater than that of PA. As shown in Fig. 1B, the maximal stimulation of PLC␦1 activity by PS (12 mol/min/mg) was at least 5-fold higher than that by PA (2.3 mol/min/mg). The concentration (mol %) of PA for half-maximal stimulation of PLC␦1 is lower than that of PS. This analysis revealed that the affinity of PLC␦1 appears to be greater for PA than for PS, whereas the maximal stimulation of PLC␦1 activity by PS is much greater than that by PA.
Because Ca 2ϩ participates directly in the PLC␦1 catalyzed hydrolysis of PIP 2 , we also examined the effect of PS concentration on the Ca 2ϩ dependence of PLC␦1 activity. As illustrated in Fig. 2, the concentration of Ca 2ϩ required for halfmaximal stimulation in 25 mol % PS mixed micelles was 0.45 M, whereas it was greater than 5 M in PS-free micelles. This result demonstrated that PS significantly increases the potency for Ca 2ϩ activation of PLC␦1.
Ca 2ϩ Regulates Phospholipid Binding to PLC␦1-PS could activate PLC␦1 by interacting with the enzyme directly or through nonspecific mechanisms. To distinguish between these possibilities, centrifugation binding experiments were performed with vesicles to examine the interaction between PLC␦1 and PS. Fig. 3 shows that PLC␦1 accumulates in the pellet fraction, a consequence of direct binding of PLC␦1 to the sucrose-loaded PS/PC vesicles. Furthermore, this binding was dependent on Ca 2ϩ . In the absence of free Ca 2ϩ , very little PLC␦1 bound to the PS/PC vesicles. PLC␦1 binding to PS/PC vesicles increased as the concentration of free Ca 2ϩ increased, reaching saturation (100% of PLC␦1 in the pellet) at 50 M free Ca 2ϩ .
It is the PS that is essential for PLC␦1 to bind to the PS/PC phospholipid vesicles. No binding of PLC␦1 was detected in vesicles devoid of PS. As the concentration of PS increased (Fig.  4a), more PLC␦1 bound to the vesicle; saturation of PLC␦1 binding occurred at a concentration of 35 mol % PS. The concentration of PS required in the mixed micelles for half-maximal binding was estimated to be 15 mol %. The binding of phospholipid vesicles by PLC␦1 was remarkably specific for PS; in contrast, only a small amount of PLC␦1 bound to PA/PC vesicles (Fig. 4b).
Phophatidylserine Regulates Ca 2ϩ Binding to PLC␦1-A nitrocellulose filter Ca 2ϩ protein binding assay was used to investigate the role of Ca 2ϩ in regulating the binding of phospholipid to PLC␦1. In the absence of PLC␦1, 45 Ca 2ϩ was poorly retained on the nitrocellulose membrane (Fig. 5A). Even the binding between 0.15 M PLC␦1 and 40 M free 45 Ca 2ϩ (1.5 ϫ 10 7 cpm) was barely detectable (7,000 -10,000 cpm). However, the binding of 45 Ca 2ϩ to PLC␦1 was significantly increased by including PS/PC but not PA/PC vesicles in the binding mixture. As shown in Fig. 5A, as much as 1% (150,000 cpm) of the total 45 Ca 2ϩ was bound to PLC␦1 when 150 M of PS/PC (mole ratio, 1:1) vesicles were co-incubated with 0.15 M PLC␦1 and 40 M of free Ca 2ϩ . Binding of Ca 2ϩ to PLC␦1 specifically required PS, because the binding was minimal with PA/PC phospholipid vesicles.
PS stimulated PLC␦1-Ca 2ϩ binding in a dose-dependent and saturable manner. As shown in Fig. 5B, Ca 2ϩ binding by PLC␦1 was significantly stimulated by as low as 5 mol % PS, and the binding plateaued at 40 mol % PS. The concentration of PS required for half-maximal Ca 2ϩ binding under these conditions was estimated to be 10 mol %. In contrast, stimulation of Ca 2ϩ -PLC␦1 binding by PA/PC vesicles was minimal. The maximal PA-dependent Ca 2ϩ binding by PLC␦1 was lower than that of PS and occurred at 10 mol % PA; no further stimulation of binding was observed even if the concentration of PA was increased to 75 mol %. This result demonstrates the specificity for the head group of PS required for PLC␦ to bind Ca 2ϩ .
To understand how PS facilitates Ca 2ϩ binding to PLC␦1, the effect of PS on the dose dependence of Ca 2ϩ binding to PLC␦1 was examined. 45 Ca 2ϩ bound to PLC␦1 in a dose-dependent and saturable manner when the binding was carried out in the The free Ca 2ϩ concentration in an EGTA-CaCl 2 buffer was calculated according to Fabiato and Fabiato (24). The reaction was carried at 30°C for 1-5 min, stopped, and quantitated as described under "Experimental Procedures." presence of phospholipid vesicles containing PS. As shown in Fig. 6, in the presence of PS/PC vesicles containing 35 mol % PS the maximal binding of Ca 2ϩ to 0.15 M PLC␦1 was 35 pmol, which corresponds to approximately 2.3 pmol Ca 2ϩ /pmol protein. The concentration of Ca 2ϩ required for half-maximal binding was estimated to be 7 M. In the presence of vesicles containing 10% PS, the total binding at 1 mM Ca 2ϩ was reduced to 23 pmol, corresponding to 1.4 pmol Ca 2ϩ /pmol protein. With vesicles containing 2.5 mol % PS, saturation of Ca 2ϩ binding was not reached, even at a free Ca 2ϩ concentration of 1 mM. These results demonstrate that PS specifically increases the affinity of Ca 2ϩ for PLC␦1.

PS-dependent Ca 2ϩ Binding to the Cleavage Center Mutant (E341G,E390G) and C2 Domain Deletion Mutant (⌬646 -654)-
Two structure features in PLC␦1 have been shown to be involved in divalent metal ion binding, the cleavage center and the C2 domain near the C terminus of the enzyme. To determine whether PS-dependent Ca 2ϩ binding is mediated by these structural determinants, we examined the Ca 2ϩ binding activity of a cleavage center double point mutant (E341G,E390G) and a C2 domain loop deletion mutant (⌬646 -654). The double mutant enzyme E341G,E390G is completely defective in catalysis (Table I), presumably because of the loss of Ca 2ϩ binding at the cleavage center. In contrast, C2 loop deletion ⌬646 -654 mutant catalyzes the hydrolysis of PI and PIP 2 in a manner comparable with that of the native enzyme ( Table I).
The E341G,E390G mutant PLC␦1, although defective in cleavage activity, was able to bind Ca 2ϩ in a PS-dependent manner indistinguishable from the native enzyme (Fig. 7A), and the binding also displayed Ca 2ϩ dependence similar to that of the native enzyme (Fig. 7B). Although the ⌬646 -654 deletion mutant is as active as the native enzyme in catalyzing the hydrolysis of PI or PIP 2 , this mutant was severely defective in PS-dependent Ca 2ϩ binding. The ⌬646 -654 deletion mutant enzyme bound Ca 2ϩ poorly, even at saturating concentrations of either PS or free Ca 2ϩ (Fig. 7, A and B). These results identify the C2 domain as the structural motif responsible for mediating PS-dependent Ca 2ϩ binding.
The ⌬646 -654 Deletion Mutant Is Defective in PS Stimulation and PS Binding-To study whether the ⌬646 -654 mutant is also defective in PS mediated stimulation, we examined the effect of PS on the Ca 2ϩ -dependent catalysis of substrate. In the absence of PS, the ⌬646 -654 mutant catalyzed the hydrolysis of PIP 2 with a similar Ca 2ϩ dependence as the native enzyme (Fig. 8). In contrast, in the presence of 25 mol % PS, the activity of the ⌬646 -654 mutant is at least lower than that of the native enzyme by a factor of 3 (Fig. 8). This result illustrates that residues 646 -654 are not essential for the basal activity of PLC␦1 but are required for PS-dependent enzyme activation.
The ⌬646 -654 mutant is also impaired in its ability to bind PS. Although E341G,E390G was able to bind a PS vesicle in a Ca 2ϩ -dependent manner (Fig. 9a), the ⌬646 -654 mutant was severely defective in Ca 2ϩ -dependent PS binding (Fig. 9b). This finding revealed that residues 646 -654 in the C2 domain of PLC␦1 are also involved in PS binding.
PS Reduces the K m for PIP 2 -PS stimulation of PIP 2 hydrolysis in the mixed micelles could be because of an increase in the maximal turnover of the enzyme, an increase in affinity for the substrate, or an increase in affinity of the enzyme for the membrane. To distinguish between these possibilities or a combination thereof, the effects of PS concentration in the mixed micelles on the substrate dependence of PLC␦1 catalysis was examined. The rate of hydrolysis of PIP 2 in mixed micelles containing either 0, 2.5, or 35 mol % PS was compared as a function of the total PIP 2 concentration (from 0.5 to 50 M, corresponding to a molar fraction from 0.1 to 9 mol %). When the concentration of PIP 2 was increased from 0.1 to 9%, the hydrolysis of PIP 2 in mixed micelles containing 25 mol % PS sharply increased from 3 to 18 mol/min/mg and was saturated as the PIP 2 concentration approached 4 mol % (Fig. 10). PIP 2 hydrolysis in PS-free mixed micelles increased slowly and did not reach a maximum even at a 9 mol % PIP 2 (Fig. 10). The most dramatic stimulatory effect of PS on the hydrolysis of PIP 2 was found at a low substrate concentration. As shown in Fig. 10, when PIP 2 concentration was less than 2 mol %, PIP 2 hydrolysis in mixed micelles containing 25 mol % PS was at least 10-fold higher than in PS-free mixed micelles. The stimulatory effect of PS diminished as the concentration of PIP 2 increased (Fig. 10).
The kinetic parameters of PLC␦1 (Table II) demonstrate that the stimulation of PIP 2 hydrolysis by PS was primarily because of a reduction in the interfacial Michaelis constant (K m ), which governs the affinity between the catalytic site and the substrate. The K m was reduced by a factor of 20, from 0.065 to 0.003, as PS in the mixed micelles was increased from 0 to 25 mol %. In contrast, PS had little effect on the affinity of the enzyme for the membrane (K s ) and its maximal rate of catalysis (V max ).   (20), expressed as release of mol of IP 3 or IP/min/mg of enzyme.

DISCUSSION
In this report we examine in detail the effects of nonsubstrate phospholipids on PLC␦1 activity and the molecular mechanism of their interaction. Although both PA and PS stimulate PLC␦1, the mechanism of stimulation may be different for PA and PS. They differ in affinity and maximal stimulation of PLC␦1. A direct interaction between PLC␦1 and PS but not PA was demonstrated by the binding assay. Furthermore, Ca 2ϩ binding to PLC␦1 requires PS not PA. These observations suggested that there is a specific interaction between PS and PLC␦1. The similar effect of PS on enzyme stimulation has also been demonstrated for protein kinase C (31,32).
Although both C2 domain and cleavage center could be the Ca 2ϩ binding site, the present data show that PS-dependent Ca 2ϩ binding occurs at the C2 domain of PLC␦1. In the presence of saturation level of PS but no substrate, the maximal binding of Ca 2ϩ to PLC␦1 was measured to be 2.1 pmol Ca 2ϩ / pmol protein. The concentration of Ca 2ϩ required for halfmaximal binding was estimated to be 7 M. This stoichiometry is in good agreement with other studies indicating that a total of three to five metal ions bind to a single molecule of PLC␦1 (3). One site exists in the catalytic domain where a single calcium ion binds and is required for catalysis (3). Substrate facilitates calcium binding at this site. A second site exists in the C2 domain, where four putative metal ion binding sites are present. Two of these sites have been shown to bind calcium ions in structural studies, and three can be occupied by lanthanum (12). The actual number of calciums binding to the C2 domain of PLC␦1 is unknown. Because calcium analogs such as lanthanum may actually have greater affinity for true calcium binding sites, three calcium binding sites could well be an overestimate. The present studies indicate a minimum of two calciums binding to the C2 domain.
Residues 646 -654 constitute the C terminus of the loop that connects ␤1 and ␤2 strands in the C2 domain of PLC␦1 (3). The phenotype of the ⌬646 -654 deletion mutant revealed that this region of C2 domain is required for PLC␦1 to interact with PS and Ca 2ϩ . This finding is consistent with the role of the C2 domain in other molecules. The C2 domains of cytosolic phospholipase A2, synaptotagmin I, and protein kinase C all mediate calcium-dependent phospholipid binding (33)(34)(35). Our data indicated that abolishing Ca 2ϩ binding to the C2 domain did not affect the basal activity of PLC␦1. This is consistent with the findings of Grobler and Hurley (36), which show that deleting similar loop in PLC␦1 does not affect its catalysis of PIP 2 Hydrolysis of increasing concentrations of PIP 2 by PLC␦1 in PIP 2 /dodecyl maltoside mixed micelles containing 0 (q), 10 (Ⅺ), and 25 (f) mol % PS is shown. The mol % of PIP 2 was increased, while the concentration of PS was fixed at 0, 10, and 25 mol %, and the combined total concentration of PS plus dodecyl maltoside was maintained at 495 M (see "Experimental Procedures"). The reaction was carried out and stopped, and the [ 3 H]IP 3 was separated and quantitated as described under "Experimental Procedures." hydrolysis in PC vesicles. However, this loop is required for the enzyme to interact with and is stimulated by PS, because ⌬646 -654 mutant enzyme was much less active than the native enzyme when catalysis was performed in the presence of PS. Although the loop connecting ␤1 and ␤2 strand in the C2 domain is required for PS mediated activation and Ca 2ϩ binding of PLC␦1, other residues in the C2 domain were mapped and found essential for PS to bind and stimulate PLC␦1. 2 The relationship between calcium, PS, and PLC␦1 appears to be very complex because all three are interdependent. While calcium regulates PS binding, PS also regulates calcium binding to PLC␦1. PS increases the potency for free calcium on the activation of PLC␦1. PS facilitates the binding of calcium to PLC␦1 by increasing the affinity. In the absence of PS, very little calcium binds to PLC␦1. PA has no effect on calcium binding correlating with its weak ability to activate PLC␦1. Furthermore, deleting residues 646 -654 not only abolished Ca 2ϩ binding but also eliminated the activation and binding of PLC␦1 by PS. All these observations indicated that the simultaneous presence of Ca 2ϩ , PLC␦1, and PS stabilizes a ternary complex formed by these three components.
The predominant effect of PS on the stimulation of PLC␦1 is to reduce the interfacial Michaelis constant (K m ) (increasing the substrate affinity for the catalytic site). The K m was reduced by a factor of 20. A similar result was also found for protein kinase C; a decrease in K m results when rat brain enzyme binds PS and Ca 2ϩ (37). PS has also been found to increase the affinity between coagulation factor IXa and VIIIa and to reduce the K m for the factor IXa catalyzed proteolysis reaction (38). Both K s (which governs the affinity of the enzyme for membrane) and V max (which reflects the velocity at infinite substrate concentrations) were not affected by the presence of PS. The present kinetic analysis may not have been sensitive enough to detect the contribution of PS on membrane anchorage (K s ), because membrane association may be primarily driven by PIP 2 binding to the PH domain. The binding of PIP 2 to the PH domain is at least 10-fold tighter than that for the Ca 2ϩ -dependent binding of PS by the enzyme (5).
Although two distinct mechanisms have been proposed by isolated reports (39,40), the binding of calcium and PS to the C2 domain may also be an important mechanism for regulation of this enzyme in vivo. Our data demonstrated that in the absence of PS, PLC␦1 is relatively inactive to catalyze the hydrolysis of physiological concentrations of PIP 2 (Ͻ1 mol % of total membrane phospholipids). However, PS and calcium stimulate PLC␦1 at physiological concentration of PIP 2 by increasing the affinity for the substrate. The concentration of PS (41) and Ca 2ϩ required to stimulate PLC␦1 are within limits of their intracellular concentrations. The intracellular concentration of calcium is usually below 1 M; however, the local calcium can rise to almost mM levels after stimulation by various calcium mobilizing agonists. We suggest that the formation of a stimulatory ternary complex is a mechanism by which calcium can regulate this isoform. If we assume that the mol % of PS is relatively constant in the plasma membranes of cells, then calcium becomes the primary modulator of PLC␦1 activ-ity. Of course, this fits into a well established and broad paradigm in which fluxes of calcium are modulators of protein function and cellular effects.
In summary, the present results show that the formation of an enzyme-PS-calcium ternary complex through the C2 domain increases its affinity for substrate and consequently stimulates the enzyme. We postulate that the formation of this ternary complex plays a role in the in vivo activation and regulation of PLC␦1.