Roles of Catalytic Domain Residues in Interfacial Binding and Activation of Group IV Cytosolic Phospholipase A2 *

Group IV cytosolic phospholipase A2 (cPLA2) has been shown to play a critical role in eicosanoid biosynthesis. cPLA2 is composed of the C2 domain that mediates the Ca2+-dependent interfacial binding of protein and the catalytic domain. To elucidate the mechanism of interfacial activation of cPLA2, we measured the effects of mutations of selected ionic and hydrophobic residues in the catalytic domain on the enzyme activity and the membrane binding of cPLA2. Mutations of anionic residues located on (Glu419 and Glu420) or near (Asp436, Asp438, Asp439, and Asp440) the active site lid enhanced the affinity for cPLA2 for anionic membranes, implying that the electrostatic repulsion between these residues and the anionic membrane surface might trigger the opening of the active site. This notion is further supported by a biphasic dependence of cPLA2 activity on the anionic lipid composition of the vesicles. Mutations of a cluster of cationic residues (Lys541, Lys543, Lys544, and Arg488), while significantly enhancing the activity of enzyme, abrogated the specific activation effect by phosphatidylinositol 4,5-bisphosphate (PtdIns(4,5)P2). These data, in conjunction with cell activity of cPLA2 and mutants transfected into HEK293 cells, suggest that the cationic residues form a specific binding site for PtdIns(4,5)P2 and that the specific PtdIns(4,5)P2 binding is involved in cellular activation of cPLA2. Also, three hydrophobic residues at the rim of the active site (Ile399, Leu400, and Leu552) were shown to partially penetrate the membrane, thereby promoting membrane binding and activation of cPLA2. Based on these results, we propose an interfacial activation mechanism for cPLA2 which involves the removal of the active site lid by nonspecific electrostatic repulsion, the interdomain hinge movement induced by specific PtdIns(4,5)P2 binding, and the partial membrane penetration by catalytic domain hydrophobic residues.

into potent inflammatory lipid mediators, platelet-activating factor and eicosanoids that include prostaglandins, thromboxanes, leukotrienes, and lipoxins. Among multiple forms of PLA 2 s found in mammalian tissues, calcium-dependent group IV PLA 2 (cPLA 2 ) has been shown to play a key role in the biosynthesis of inflammatory lipid mediators (1)(2)(3). cPLA 2 is composed of the amino-terminal C2 domain and the carboxylterminal catalytic domain. Previous structural (4) and functional (5,6) studies have demonstrated that the C2 domain is involved in the membrane binding and subcellular targeting of the cPLA 2 molecule. However, the role of the catalytic domain in the membrane binding and activation of cPLA 2 is still poorly understood. A recent crystal structure of cPLA 2 revealed some unique structural features of the protein that provide a clue to the potential role of the catalytic domain in the interfacial activation of cPLA 2 (7). First, the C2 and the catalytic domains are connected by a narrow and flexible linker region (see Fig.  1), implying that a hinge motion of this linker might be involved in cPLA 2 activation. Second, the active site entry of the enzyme is partially blocked by a largely amphiphilic lid (residues 413-457) that is flanked by highly flexible lid hinges (see Figs. 1 and 2). Thus, the interfacial activation of cPLA 2 should involve the removal of the lid from the active site. Third, several anionic and cationic patches are present on the putative membrane-binding surface of the catalytic domain that surrounds the active site cavity (see Fig. 1). In particular, two clusters of anionic residues are present in the active site lid and in the lid hinges, respectively, whereas two prominent cationic patches are found in the molecular periphery (see Figs. 1 and 2). Finally, several hydrophobic residues are located at the rim of the active site (Fig. 2).
This study was undertaken to investigate the roles of the catalytic domain in membrane binding and activation of cPLA 2 .
In vitro mutational analyses of ionic and hydrophobic residues in the catalytic domain using surface plasmon resonance, monolayer penetration, and activity assays, in conjunction with the studies of HEK293 cells transfected with cPLA 2 and selected mutants tagged with enhanced green fluorescence protein (EGFP), provide new insights into the mechanism of cPLA 2 activation and the roles of catalytic domain residues in this process.
Mutagenesis, Expression, and Purification of cPLA 2 -Baculovirus transfer vectors encoding the cDNAs of cPLA 2 with appropriate catalytic domain mutations were generated by the overlap extension PCR using the plasmid pVL1392-cPLA 2 (10) as a template. Transfection of Sf9 cells with mutant pVL1392-cPLA 2 constructs was performed using the BaculoGold TM transfection kit from BD PharMingen. Plasmid DNA for transfection was prepared using an EndoFree Plasmid Maxi kit (Qiagen, Valencia, CA) to avoid possible endotoxin contamination. For protein expression, Sf9 cells were grown to 2 ϫ 10 6 cells/ml in 350-ml suspension cultures and infected with high titer recombinant baculovirus at a multiplicity of infection of 10. The cells were then incubated for 3 days at 27°C. For harvesting, cells were centrifuged at 1000 ϫ g for 10 min and resuspended in 35 ml of extraction buffer (20 mM Tris-HCl, pH 7.5, containing 0.1 M KCl, 1 mM EDTA, 20% (v/v) glycerol, 10 g/ml leupeptin, and 1 mM phenylmethanesulfonyl fluoride). The suspension was homogenized in a hand-held homogenizer chilled on ice. The extract was centrifuged at 100,000 ϫ g for 1 h at 4°C. To the supernatant, 1 ml of nickel-nitrilotriacetic acid-agarose (Qiagen, Valencia, CA) was added, and the mixture was kept shaking at 100 rpm for 1 h. The protein was purified according to the manufacturer's protocol. cPLA 2 fractions were concentrated and desalted in an Ultrafree-15 centrifugal filter device (Millipore, Bedford, MA). Protein concentration was determined by the bicinchoninic acid method (Pierce).
Kinetic Measurements-PLA 2 -catalyzed hydrolysis of mixed liposomes was carried out as described previously (9). Briefly, 0.1 M PyArPC (1 mol %) inserted in 9.9 M BLPG was incubated with enzymes at 37°C in 2 ml of 10 mM HEPES buffer, pH 7.4, containing 2 M BSA, 0.16 M NaCl, and 10 mM CaCl 2 . The progress of hydrolysis was monitored as an increase in fluorescence emission at 378 nm using a Hitachi F4500 fluorescence spectrometer with the excitation wavelength set at 345 nm. Spectral bandwidth was set at 5 nm for both excitation and emission. Specific activity was determined from the initial slope of the reaction as described previously. Activity of cPLA 2 on zwitterionic [ 14 C]SAPC vesicles was assayed by measuring the initial rate of [ 14 C]AA release as described previously (11).
SPR Measurements-The preparation of vesicle-coated Pioneer L1 sensor chip (Biacore) was described in detail elsewhere (12). The sensor surface was coated with DHPC or DHPG vesicles. In control experiments, the fluorescence intensity of the flow buffer after rinsing the sensor chip coated with vesicles incorporating 10 mM 5-carboxyfluorescein (Molecular Probes) was monitored to confirm that the vesicles remained intact on the chip. All experiments were performed with a control cell in which a second sensor surface was coated with BSA. All kinetic measurements were performed at 24°C and at a flow rate of 60 l/min. For data acquisition, five or more different concentrations (typically within a 10-fold range above or below the K d ) of each protein were used. After each set of measurements, the entire immobilized vesicles were removed by injection of 25 l of 40 mM CHAPS, followed by 25 l of octyl glucoside at 5 l/min, and the sensor chip was re-coated with a fresh vesicle solution for the next set of measurements. By using BIAevaluation 3.0 software (Biacore), the association and dissociation phases of all sensorgrams were globally fit to a 1:1 Langmuir binding model: protein ϩ (protein-binding site on the vesicle) 7 (complex). The association phase was analyzed using the equation, where RI is the refractive index change; R max is the theoretical binding capacity; C is the analyte con-centration; k a is the association rate constant, and t 0 is the initial time. The dissociation phase was analyzed using the equation, R ϭ R 0 e Ϫkd(tϪt0) , where k d is the dissociation rate constant, and R 0 is the initial response. The dissociation constant (K d ) was then calculated from the equation, K d ϭ k d /k a . It should be noted that in our SPR analysis K d is defined in terms of not the molarity of phospholipids but the molarity of protein-binding sites on the vesicle (see above). Thus, if each protein-binding site on the vesicle is composed of n lipids, nK d is the dissociation constant in terms of molarity of lipid monomer (13). Due to difficulty involved in accurate determination of the concentration of lipids coated on the sensor chip, however, only K d was determined in our SPR analysis, and the relative affinity was calculated as a ratio of K d values assuming that n values are essentially the same for wild type and mutants.
Monolayer Measurements-Surface pressure () of solution in a circular Teflon trough (4 ϫ 1 cm) was measured using a Wilhelmy plate attached to a computer-controlled Cahn electrobalance (model C-32) as described previously (13,14). Five to ten microliters of phospholipid solution in ethanol/hexane (1:9 (v/v)) or chloroform was spread onto 10 ml of subphase (20 mM Tris-HCl, pH 7.5, containing 0.1 M KCl and 0.1 mM Ca 2ϩ ) to form a monolayer with a given initial surface pressure ( 0 ). The subphase was continuously stirred at 60 rpm with a magnetic stir bar. Once the surface pressure reading of monolayer had been stabilized (after ϳ5 min), the protein solution (typically 30 l) was injected into the subphase and the change in surface pressure (⌬) was measured as a function of time. Typically, the ⌬ value reached a maximum after 30 min. The maximal ⌬ value depended on the protein concentration and reached a saturation value (i.e. 1 g/ml). Protein concentrations in the subphase were therefore maintained above such values to ensure that the observed ⌬ represented a maximal value. The ⌬ versus 0 plots were constructed from these measurements.
Cell Culture, Transfection, and Protein Production-EGFP was cloned from the vector pEGFP (CLONTECH) by PCR with its stop codon removed. It was inserted between the HindIII and the NotI site of a modified pIND vector (Invitrogen) to create the plasmid pIND/EGFP. For subcloning of cPLA 2 and its mutants into the pIND/EGFP vector, NotI and BglII sites were introduced to the 5Ј and the 3Ј ends, respectively, of corresponding pVL1392-cPLA 2 vectors by PCR. The PCR product was digested and ligated into pIND/EGFP plasmid to create an amino-terminal EGFP fusion protein with a spacer sequence, MRPH. Plasmid DNA for transfection was prepared using an EndoFree Plasmid Maxi kit (Qiagen, Valencia, CA) to avoid possible endotoxin contamination.
A stable HEK293 cell line expressing the ecdysone receptor (Invitrogen) was used for all cell studies. Briefly, cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS) at 37°C in 5% CO 2 and 98% humidity until 90% confluent. Cells between the 5th and 20th passage were passaged into 8 wells of a Lab-Tech TM chambered coverglass. For transfection, these cells were exposed to 150 l of unsupplemented DMEM containing 0.5 g of endotoxin-free DNA and 1 l of LipofectAMINE TM reagent (Invitrogen) for 7-8 h at 37°C. After exposure, the transfection medium was removed; the cells were washed once with FBS-supplemented DMEM and overlaid with FBS-supplemented DMEM containing Zeocin TM (Invitrogen) and 140 g/ml Ponasterone A to induce protein production.
Confocal Microscopy-Images were obtained using a 4-channel Zeiss LSM 510 laser scanning confocal microscope. EGFP was excited using the 488-nm line of an argon/krypton laser. A 505-nm line pass filter and a ϫ63, 1.2 numerical aperture water immersion objective were used for all experiments. All experiments were carried out at the same laser power, which was found to induce minimal photobleaching over 30 scans, and at the same gain and offset settings on the photomultiplier tube. Cells for imaging were selected based on their initial intensity. The LSM 510 imaging software was used to control the time intervals for imaging calcium-dependent translocation of cPLA 2 . Immediately before imaging, induction media were removed, and the cells were washed twice with 150 l of 2 mM EGTA and then overlaid with 150 l of HEK buffer (1 mM HEPES, pH 7.4, containing 2.5 mM MgCl 2 , 1 mM NaCl, 0.6 mM KCl, 0.67 mM D-glucose, 6.4 mM sucrose) containing 2 mM EGTA and 10 M ionomycin. After initial imaging of the cells, 150 l of HEK buffer containing 6 mM CaCl 2 was added to transfected cells to induce the calcium-dependent translocation of cPLA 2 and its mutants. Control experiments were done with dimethyl sulfoxide in place of ionomycin.
Cellular AA Release-Although the expression of cPLA 2 and mutants in HEK293 cells was induced as described above, cells were radiolabeled with 0.1 Ci/ml [ 3 H]AA (200 Ci/mmol). After 20 h at 37°C, medium was removed, and cells were washed three times with the DMEM, 10% FBS medium. Cells were stimulated with 10 M ionomycin in 0.5 ml of the medium for 30 min at 37°C. The culture medium was collected and briefly centrifuged, and 0.25-ml aliquots were subjected to scintillation counting. Cells were then removed from the plate by adding 200 l of trypsin/EDTA solution and collected by brief centrifugation. The cell pellet was resuspended in 150 l of the medium and was subjected to scintillation counting to obtain total radioactivity of the cells. The [ 3 H]AA release is expressed in terms of the percentage of total radioactivity incorporated into cells.

Mutations of Anionic Residues in the Catalytic Domain-The
x-ray crystal structure of cPLA 2 (7) indicates the presence of two clusters of anionic residues in the lid blocking the active site and the hinge region that connects the lid to the rest of the molecule, respectively (see Figs. 1 and 2). Glu 418 , Glu 419 , and Glu 420 on the lid are surface-exposed in the structure, and the Asp 436 , Asp 438 , Asp 439 , and Glu 440 in the hinge region are also expected to be exposed although their coordinates have not been defined due to high flexibility. Because all intracellular membranes, including perinuclear membranes, contain considerable amounts of anionic phospholipids (15), the docking of cPLA 2 to these membranes by the C2 domain would create energetically unfavorable contacts between the membrane surface and the catalytic domain. This suggests that the interfacial activation of cPLA 2 , i.e. the removal of the lid from the active site, might be triggered by the electrostatic repulsion. To test this notion, we mutated these surface anionic residues to lysines, i.e. D436K/D437K/D438K/E440K. For Glu 419 and Glu 420 , E419A/E420A mutation was selected over E419K/ E420K due to low stability of the latter. We then measured the membrane binding affinities of these mutants for zwitterionic (DHPC) and anionic vesicles (DHPG) by SPR analysis. DHPG was selected over its phosphatidylserine (PS) derivative due to difficulty involved in the synthesis of latter. We also measured the enzyme activities of wild type and mutants toward zwitterionic [ 14 C]SAPC vesicles and anionic mixed vesicles of PyArPC (1 mol %)/BLPG. We previously showed that PyArPC is selectively hydrolyzed in PyArPC/BLPG mixed vesicles because of extremely low activity of cPLA 2 on BLPG (9, 10). As summarized in Table I, cPLA 2 wild type has ϳ4-fold higher affinity for PC vesicles than for PG vesicles. As expected, E419A/E420A and D436K/D438K/D439K/E440K had wild type-like affinities for zwitterionic PC vesicles. Similarly, the mutations yielded activities toward [ 14 C]SAPC vesicles within a factor of 2 for wild type. Modestly decreased activities are likely to reflect minor local conformational changes that interfere with the activation process. When the affinity for anionic PG vesicles was measured, the quadruple mutant D436K/D438K/D439K/ E440K showed 5-fold higher affinity than wild type; as a result, it binds PG vesicles slightly better than PC vesicles. The large increase in affinity for anionic vesicles by the charge reversal indicates that the hinge region containing the Asp 436 /Asp 438 / Asp 439 /Glu 440 cluster would make energetically unfavorable contact with the anionic membrane surface as cPLA 2 approaches the membrane. Interestingly, D436K/D438K/D439K/ E440K was only 60% more active than wild type toward Py-ArPC/BLPG mixed vesicles. This is much lower than expected from its enhanced affinity for PG vesicles, even after taking into account a partial loss of enzyme activity caused by the mutation (see above). The lower-than-expected activity suggests that electrostatic attraction between cationic residues introduced to the lid hinge region of cPLA 2 and the anionic membrane surface interferes with the enzyme activation process. This in turn supports the notion that interfacial activation of cPLA 2 involves the removal of the active site lid by the electrostatic repulsion. On the other hand, E419A/E420A had a modest but definite 30% gain in affinity for PG vesicles and a slight 10% decrease in activity toward PyArPC/BLPG vesicles. These changes are much smaller than those seen with D436K/ D438K/D439K/E440K, largely because E419A/E420A has a two-site charge removal in lieu of a four-site charge reversal in the latter mutant. The data still support the notion, however, that the two lid anionic residues unfavorably interact with the anionic membrane surface.
Mutations of Cationic Residues in the Catalytic Domain-Although the full-length cPLA 2 has ϳ4-fold higher affinity for PC vesicles than for PG vesicles, this preference is significantly less than that of the isolated C2 domain that strongly prefers PC vesicles (16,17). Furthermore, it has been reported that cPLA 2 can tightly bind anionic phosphatidylmethanol vesicles (18) or PtdIns(4,5)P 2 -containing PC vesicles (19) even in the absence of calcium. These data suggest that the catalytic domain favorably interacts with anionic membrane surface, independently of its calcium-dependent, PC-favoring C2 domain. Consistent with this notion, the catalytic domain of cPLA 2 contains a number of surface-exposed cationic residues, form- ing two prominent patches (i.e. Arg 488 /Lys 541 /Lys 543 /Lys 544 and Lys 271 /Lys 273 /Lys 274 /Arg 467 ) on its putative membrane-binding surface (see Fig. 1). To assess the roles of these cationic residues in the membrane binding of cPLA 2 , we selectively mutated them to either Glu or Ala, depending on the relative stability of mutants. Specifically, K271A/K273A/K274A, R467A, K541A/K543A/K544A, and R488E were generated, and their membrane affinity and lipolytic activity were measured. Mutations of one cationic patch, i.e. K271A/K273A/K274A and R467A, led to 2-3-fold reduction in both affinity for PG vesicles and activity on PyArPC/BLPG vesicles while exhibiting the wild type-like affinity for and activity on PC vesicles. Thus, these clustered residues seem to play some role in binding to anionic membranes. On the other hand, the mutations in the other cationic patch, K541A/K543A/K544A and R488E, did not significantly change the affinity for either PC or PG vesicles, indicating that they do not make close contact with the membrane. Unexpectedly, however, K541A/K543A/K544A had Ͼ2fold higher activity than wild type toward both SAPC and PyArPC/BLPG vesicles, whereas R488E was modestly more active than wild type toward both substrates. These data indicate that the mutations of cationic patch composed of Arg 488 / Lys 541 /Lys 543 /Lys 544 result in a conformational change of protein that leads to enzyme activation. This in turn suggests that the cationic cluster might be involved in keeping the enzyme in an inactive conformation in the resting state.
Requirements of Anionic Phospholipids for cPLA 2 Activation-It has been long known that some anionic phospholipids can increase cPLA 2 activities (20). In particular, PtdIns(4,5)P 2 was shown to specifically activate cPLA 2 (19). However, the origin of this activation is not fully elucidated. To investigate this aspect systematically, we measured the enzyme activity and membrane affinity of cPLA 2 and selected mutants in the presence of different anionic phospholipids in mixed vesicles with PC. First, we measured the effect of PG on the vesicle affinity and the enzyme activity of cPLA 2 . As expected from the preference of cPLA 2 for PC vesicles (see Table I), the affinity of cPLA 2 for DHPC/DHPG-mixed vesicles determined by the SPR analysis decreased monotonously with the increase in PG composition in the mixed vesicles, reaching a minimum at 50 mol % of PG (Fig. 3). However, the enzyme activity of cPLA 2 showed a sharp biphasic dependence on PG composition. As the DHPG composition in [ 14 C]SAPC/DHPG-mixed vesicles increased, the enzyme activity measured in terms of [ 14 C]AA release steeply rose up to 90% at 7 mol % PG then decreased with the further increase in PG composition. This unique PG dependence suggests that PG enhances the activity of cPLA 2 not by increasing its membrane affinity but by facilitating its activation process at the membrane surface, presumably by triggering the removal of anionic lid from the active site by electrostatic repulsion. The sharp bell-shape dependence is likely to reflect the combination of the PG-facilitated activation of cPLA 2 that reaches the plateau at ϳ7 mol % of PG and the monotonous PG-dependent decrease of membrane affinity. To preclude the possibility that the observed effect is a PG-specific artifact, we performed a similar measurement with PS. As is the case with PG, cPLA 2 has been shown to have lower affinity for PS than for PC (19). Due to the difficulty involved in the synthesis of a non-hydrolyzable PS derivative in large quantity, we used DPPS toward which cPLA 2 shows low activity (i.e. Ͻ1% activity toward SAPC) 2 and measured the enzyme activity as a function of the PS composition in SAPC/DPPS-mixed vesicles. As illustrated in Fig. 3, the PS dependence of cPLA 2 activity was essentially the same as the PG dependence. These data thus indicate that the biphasic dependence of cPLA 2 activity on anionic lipids is common to those anionic lipids for which cPLA 2 has lower affinity than for PC.
a Absolute specific activity values for wild type were 100 Ϯ 15 and 20 Ϯ 4 nmol/mg min for SAPC and PyArPC/BLPG vesicles, respectively. concentrations (19). We also found that with 10 M Ca 2ϩ , PtdIns(4,5)P 2 specifically activates cPLA 2 , up to 3.5-fold at 5 mol % of PtdIns(4,5)P 2 in SAPC vesicles, whereas PtdIns(3)P and PtdIns(3,4)P 2 up to 5 mol % had negligible effect (see Fig.  4A). We did not measure the effect of these phosphoinositides beyond 5 mol %, as higher concentrations would be physiologically irrelevant. When measured by SPR analysis, 5 mol % PtdIns(4,5)P 2 in DHPC vesicles did not significantly enhance the affinity of cPLA 2 (data not shown), indicating that the activity increase is not due to enhanced affinity. To account for this specific effect of PtdIns(4,5)P 2 , it was proposed that cPLA 2 has a PH domain-like structure in the region surrounding the Lys 271 /Lys 273 /Lys 274 anionic cluster (19), which was later disputed on the basis of the x-ray structure of cPLA 2 (7). To explore the possibility that any of the two prominent cationic clusters are involved in PtdIns(4,5)P 2 binding, we measured the enzyme activity of cationic site mutants, K271A/K273A/ K274A, K541A/K543A/K544A, and R488E, in the presence of PtdIns(4,5)P 2 . As shown in Fig. 4B, K271A/K273A/K274A was as active as the wild type toward SAPC vesicles (also see Table  I), and its the activity increased with the increase of PtdIns(4,5)P 2 concentration (up to 300%), confirming that this cationic cluster is not involved in PtdIns(4,5)P 2 binding. In contrast, neither K541A/K543A/K544A nor R488E was activated by PtdIns(4,5)P 2 . In fact, PtdIns(4,5)P 2 modestly reduced the activities of K541A/K543A/K544A and R488E. This suggests that these residues, Arg 488 , Lys 541 , Lys 543 , and Lys 544 , might form a specific binding site for PtdIns(4,5)P 2 .
Cellular Membrane Translocation and Activation-To assess the physiological relevance of the effect of PtdIns(4,5)P 2 on cPLA 2 activation, we transiently transfected HEK293 cells with cPLA 2 wild type and two mutants, K271A/K273A/K274A and K541A/K543A/K544A, tagged with EGFP at their carboxyl termini. We then determined their cellular activities by monitoring the [ 3 H]AA release from radiolabeled cells and their subcellular localization by time-lapse confocal imaging. It has been shown that calcium induces the perinuclear translocation of cPLA 2 (21,22). As shown in Fig 5, cPLA 2 was evenly dispersed in the cytoplasm when cells were incubated with a Ca 2ϩ -depleted medium and moved to the perinuclear region upon cell activation with external Ca 2ϩ and ionomycin, showing a clear annular pattern. Under this condition, HEK293 cells overexpressing wild type cPLA 2 had ϳ6-fold higher AA releasing activity than parent cells when compared at 30 min after stimulation (see Fig. 6). K271A/K273A/K274A and K541A/K543A/K544A showed the wild type-like subcellular location patterns (see Fig. 5). When compared with the wild type, K271A/K273A/K274A had only 40% of net activity (i.e. corrected for the control background), which compares well with its 30% wild type activity toward PyArPC/BLPG vesicles. This suggests that perinuclear membranes of HEK293 cells, including nuclear envelope, contain a considerable amount of anionic lipids. Interestingly, K541A/K543A/K544A, which was ϳ2.5-fold more active than wild type in in vitro assays with both SAPC and PyArPC/BLPG, was only as active as the wild type in the cellular AA assay. In this regard, it is noteworthy that the wild type cPLA 2 and K541A/K543A/K544A have comparable in vitro activity in the presence of 5 mol % PtdIns(4,5)P 2 in SAPC vesicles (see Fig. 4B). Thus, it is possible that perinuclear membranes contain a high enough concentration of PtdIns(4,5)P 2 to fully activate cPLA 2 when it is targeted to these membranes in response to the rise in intracellular calcium concentration. To test this possibility, we measured the effect of the PH domain of phospholipase C-␦1, which was shown to have high specificity for PtdIns(4,5)P 2 (23), on the cellular cPLA 2 activity. When the PH domain of phospholipase C-␦1 tagged with EGFP in its carboxyl terminus was overexpressed, it was localized in both the inner plasma membrane and the perinuclear region (see Fig. 5). Thus, it is evident that PtdIns(4,5)P 2 is present in both the plasma membrane and the perinuclear region of HEK293 cells. We then doubly transfected HEK293 cells with both cPLA 2 (or K541A/K543A/ K544A) and the PH domain. As shown in Fig. 6, the AA releasing activity of cPLA 2 was greatly reduced by the co-transfection with the PH domain. The AA release from the doubly transfected cells was only twice higher than that from parent cells (i.e. 6.5-fold drop in net activity). The reduced activity could be due to either specific PtdIns(4,5)P 2 depletion by the PH domain or nonspecific competition for perinuclear membrane-binding sites between the PH domain and cPLA 2 (or a combination of both). The activity of K541A/K543A/K544A was also reduced, albeit by only 2-fold; as a result, the mutant was three times more active than the wild type, which is comparable with the difference in their in vitro activities toward vesicles (see Table  I). Because the activity of K541A/K543A/K544A is independent of PtdIns(4,5)P 2 , this ϳ2-fold reduction in activity should reflect the nonspecific inhibition by the PH domain. This in turns indicates that the depletion of PtdIns(4,5)P 2 by the PH domain is responsible in large part for the 6.5-fold decrease in cPLA 2 inhibition by the PH domain. Most importantly, the selective inhibitory effect of PH domain on the wild type cPLA 2 corroborates the notion that Lys 541 , Lys 543 , and Lys 544 are involved in PtdIns(4,5)P 2 binding and also indicates that PtdIns(4,5)P 2 plays a significant role in cellular activation of cPLA 2 under the experimental conditions used herein.
Mutations of Hydrophobic Residues in the Catalytic Domain-The rim of the active site of cPLA 2 contains hydrophobic residues, including Ile 399 , Leu 400 , and Leu 552 . To understand the role of these residues, we mutated them to Ala and measured the effects of mutations on enzyme activity and membrane affinity. Because Ile 399 and Leu 400 are contiguous, we made a double-site mutant for these residues. As shown in Table II, L552A showed 2-3-fold decreases in affinity for and activity on PC and PG vesicles. I399A/L400A also had ϳ2.5-fold lower affinity for PC and PG vesicles, but it was less active than wild type by an order of magnitude, suggesting that these mutations might unfavorably alter the active site geometry to some degree. For these hydrophobic site mutants, their membrane binding was rigorously analyzed using SPR and monolayer analyses. Association (k a ) and dissociation (k d ) rate constants for binding to PC vesicles determined by SPR analysis show that reduced PC affinities of mutants are mainly due to increases in k d . We showed previously that in membrane-protein binding nonspecific long range electrostatic interactions mainly affect k a , whereas specific short range interactions as well as hydrophobic interactions resulting from the membrane penetration of protein largely influence k d (12). Because these hydrophobic side chains are unlikely to form specific interactions with the membrane, our data suggest that they penetrate the membrane during the interfacial catalysis of cPLA 2 . To test this notion, we measured the penetration of wild type and mutants into the mixed monolayer of DHPC/DHPG (9:1). 10 mol % of PG was included in the monolayer to allow the removal of the active site lid and the exposure of hydrophobic residues. As illustrated in Fig. 7, L552A and I399A/L400A had significantly lower monolayer penetration than wild type and could not penetrate the monolayer when 0 Ն 23 dynes/cm, supporting that the hydrophobic residues are involved in partial membrane penetration during interfacial catalysis. To prove the specific nature of the hydrophobic site mutations on monolayer penetration, we measured the monolayer penetration of R467A as a control, and it behaved essentially the same as wild type. Finally, the catalytic domain (residues 148 -749) of cPLA 2 exhibited much lower but definite monolayer penetration, again showing that hydrophobic residues in the catalytic domain make some contribution to membrane penetration and hydrophobic interactions.

DISCUSSION
The membrane binding mechanism of secretory PLA 2 s (sPLA 2 ) has been extensively studied (24). In general, an sPLA 2 has a wide open active site, the rim and the wall of which are composed of mainly hydrophobic residues and also contains ionic (primarily cationic), hydrophobic, and aromatic residues on its membrane-binding surface (25). Structure-function studies of various sPLA 2 s have shown that their membrane binding is mainly driven by long range electrostatic interactions between surface cationic residues and the anionic membrane surface (26) and complex interactions between aromatic residues and phospholipid head groups (27,28). Neither gross conformational changes nor a significant degree of membrane penetration by sPLA 2 s is involved in their membrane binding. As is the case with sPLA 2 s, the catalytic domain of cPLA 2 contains two prominent cationic patches and some hydrophobic residues on its putative membrane-binding surface. However, the catalytic domain of cPLA 2 is distinct from sPLA 2 in that its active site is partially covered by a lid containing multiple anionic residues. The presence of lid blocking the entry to the active site in the membrane-free state of enzyme strongly suggests that upon membrane binding gross conformational changes of the enzyme might take place to remove the lid from the active site and allow a phospholipid molecule to enter the active site, as is the case with the interfacial activation of lipases (29). The lid (residue 413-457) is made of a short ␣-helix and a short turn that are flanked by hinges (i.e. residues 408 -412 and 434 -456) both of which are not defined in the x-ray structure due to high mobility. The external face of the lid contains three anionic residues (Glu 418 /Glu 419 /Glu 420 ), whereas a hinge region possesses four anionic residues (Asp 436 / Asp 438 /Asp 439 /Glu 440 ), most of which are expected to be surface-exposed. The presence of these anionic clusters on the membrane-binding surface of cPLA 2 that acts on intracellular membranes containing some degree of anionic lipids implies that the removal of active site lid might be triggered by the electrostatic repulsion at the membrane surface. The notion that the interfacial activation of cPLA 2 is triggered by electrostatic repulsion is supported by several findings. First, the mutations of these anionic residues enhance the affinity for anionic vesicles, indicating that indeed these residues unfavorably interact with anionic membranes. Second, the unexpectedly low activity of D436K/D438K/D439K/E440K on PyArPC embedded in anionic BLPG vesicles suggests that the electrostatic repulsion between the anionic membrane surface and these anionic residues is involved in the removal of the lid from the active site. Third, cPLA 2 activity exhibits a unique bellshaped dependence of on the PG (and PS) composition of PC/ PG(S)-mixed vesicles, despite the monotonously negative effect of PG on the membrane affinity of cPLA 2 . It should be noted, however, that cPLA 2 shows relatively high activity on purely zwitterionic vesicles (i.e. maximal activation by PG is 90%), indicating that the active site lid could be removed even in the absence of anionic phospholipids in the membrane. This is presumably because the lid might be non-specifically removed to avoid the desolvation penalty associated with bringing highly anionic residues close to the membrane (30). Thus, it would seem that the lid removal by PG and other nonspecific anionic lipids is not absolutely required for cPLA 2 activation but that it facilitates the activation under physiological conditions.
Unlike nonspecific anionic lipids (e.g. PG and PS), PtdIns(4,5)P 2 appears to activate cPLA 2 in a specific manner. It was shown that cPLA 2 binds with high affinity and specificity to PtdIns(4,5)P 2 in a 1:1 stoichiometry (19). We found that 5 mol % PtdIns(4,5)P 2 in SAPC vesicles enhances the cPLA 2 activity by 3.5-fold, with a minimal effect on vesicle affinity. This degree of activation is much less than the reported value (Ͼ100-fold increase in activity) (19), which might be due to different experimental conditions (i.e. different calcium concentrations and vesicle versus mixed micellar systems). The crystal structure of cPLA 2 did not reveal a well defined PH domainlike structure in the catalytic domain (7). Our mutational analysis suggests, however, that the cationic cluster composed of Lys 541 , Lys 543 , Lys 544 , and Arg 488 forms at least a part of PtdIns(4,5)P 2 -binding site. Although this part of catalytic domain does not have the PH domain-like structure, it should be noted that a phosphoinositide can be specifically recognized by different structural modules, as evidenced by the specific binding of PtdIns(3)P by FYVE (31) and phox (PX) domains (32). Thus, it is possible that the PtdIns(4,5)P 2 -binding site in cPLA 2 might belong to a yet unidentified phosphoinositide-binding motif. Unexpectedly high enzyme activities of two mutants, K541A/K543A/K544A and R488E, raise an interesting possibility about the mechanism by which PtdIns(4,5)P 2 specifically activates cPLA 2 . Obviously, these cationic residues do not participate in nonspecific interactions with the anionic membrane surface, as their mutations did not influence the binding to anionic PG vesicles. Instead, it would seem that they are somehow involved in keeping the enzyme in an inactive conformation at the resting state. As shown in Fig. 1, the C2 domain of cPLA 2 also has clustered cationic residues (Arg 57 , Lys 58 , Arg 59 , and Arg 61 ) that are separated from the Lys 541 /Lys 543 /Lys 544 / Arg 488 cluster by Ն13 Å over the molecular groove. By taking into the account the flexibility of molecular linker region con-  necting the C2 domain and the catalytic domain, it is possible that the binding of PtdIns(4,5)P 2 could bring two cationic patches closer by neutralizing positive charges. This is similar to the electrostatic switch role proposed for the calcium in the C2 domain (33). It is even possible that the two cationic patches together form a PtdIns(4,5)P 2 -binding site, because they form the wall of molecular groove between the C2 domain and the catalytic domain (see Fig. 1). The membrane-binding mode of the isolated C2 domain of cPLA 2 has been extensively characterized by various biophysical techniques (17,34,35). Based on crystal structure of the whole cPLA 2 molecule, however, it is difficult to explain how the active site of the catalytic domain can be juxtaposed to the membrane while docking the C2 domain to the membrane in the orientation that was suggested by the studies of isolated C2 domain. Interestingly, the hypothetical hinge motion of the linker region induced by PtdIns(4,5)P 2 binding can bring the active site closer to the membrane surface with the C2 domain bound to the membrane in its optimal orientation (see Fig. 8).
It has been well established that cPLA 2 translocates to and acts on perinuclear membranes in response to a rise in calcium concentration (21,22). This calcium-dependent subcellular targeting has been ascribed to the C2 domain that has high preference for PC (16,17,36) that is abundant in perinuclear membranes (37)(38)(39). Our cell data using EGFP-tagged cPLA 2 wild type and mutants indicate that perinuclear membranes of HEK293 cells contain a significant amount of anionic lipids (estimated from the low activity of K271A/K273A/K274A), particularly PtdIns(4,5)P 2 in high enough concentration to fully activate cPLA 2 upon its membrane translocation (estimated from the comparable activities of wild type and K541A/K543A/ K544A). PtdIns(4,5)P 2 has been found mainly in the plasma membrane (40), but our confocal imaging of the PH domain of phospholipase C-␦ shows its presence in the perinuclear region of HEK293 cells. A selective inhibitory effect of the PH domain on cPLA 2 , with a much lesser degree of inhibition for the PtdIns(4,5)P 2 -independent K541A/K543A/K544A mutant, strongly suggests that the cellular cPLA 2 activity can be regulated by the spatiotemporal dynamics of PtdIns(4,5)P 2 in pe-rinuclear membranes. Although it was proposed that PtdIns(4,5)P 2 might be able to activate cPLA 2 in a calciumindependent manner, calcium-independent cPLA 2 activity was extremely low even in the presence of Ͼ5 mol % PtdIns(4,5)P 2 in our in vitro and cell measurements, 2 indicating that PtdIns(4,5)P 2 would regulate cPLA 2 activity in concert with calcium rather than independently of calcium. Undoubtedly, further studies are necessary to elucidate this important aspect of cPLA 2 activation.
In contrast to sPLA 2 s, the membrane binding of which does not involve the penetration of hydrophobic residues into the membrane, the membrane binding of cPLA 2 involves the calciumdependent membrane penetration (10) that is mainly mediated by the C2 domain (17). Our monolayer penetration data show that the isolated catalytic domain also has lower but definite membrane penetrating capability, which appears to be attributed at least in part to the hydrophobic residues located in the rim of the active site. The effects of mutations of these hydrophobic residues on membrane binding and activity indicate that they might play some role in elongating the membrane residence time of cPLA 2 and perhaps properly orienting the active site at the membrane surface to allow facile movement of a substrate molecule to the active site On the basis of these results, we propose a model for the membrane binding and activation of cPLA 2 illustrated in Fig.  8. In this model, cPLA 2 in the resting state exists as an inactive conformation because of active site lid and the electrostatic repulsion between cationic patches in the C2 domain (Arg 57 / Lys 58 /Arg 59 /Arg 61 ) and in the catalytic domain (Lys 541 /Lys 543 / Lys 544 /Arg 488 ). Upon the increase of calcium concentration, the C2 domain drives the cPLA 2 molecule to the membrane surface. This initial binding involves the membrane penetration of aliphatic and aromatic residues in the calcium-binding loops of the C2 domain and the interaction of cationic residues in the catalytic domain (Lys 271 /Lys 273 /Lys 274 /Arg 467 ) with anionic phospholipids (Fig. 8A). The binding would also bring the anionic residues located in the active site lid and hinge regions close to the anionic membrane surface. The resulting electrostatic repulsion swings the lid away from the active site. How-FIG. 8. A hypothetical model for the interfacial binding of cPLA 2 . A, calcium ions (brown) drive the membrane translocation of the C2 domain (orange ribbon), which brings anionic residues (red) located in the active site lid and hinge regions in contact with the anionic membrane surface. The electrostatic repulsion removes the lid away from the active site. However, the active site and the residues at the active site rim (magenta) are not juxtaposed to the membrane (cyan, only a monolayer is shown) in this binding mode. B, a PtdIns(4,5)P 2 molecule (green) migrates into the groove and specifically interacts with the cationic residues (blue) in the catalytic domain and in the C2 domain. This binding induces the conformational change of cPLA 2 that brings its active site closer to the membrane surface. These two types of conformational changes lead to the interfacial activation of the enzyme and allow a substrate molecule to enter the active site for catalytic turnover. ever, this type of membrane binding cannot lead to the optimal docking of the active site to the membrane surface as long as cPLA 2 exists in its inactive conformation. Only after the conformational change that is induced by specific interactions between a PtdIns(4,5)P 2 molecule and a group of cationic residues in the catalytic domain (Lys 541 /Lys 543 /Lys 544 /Arg 488 ), the active site of cPLA 2 is juxtaposed to the membrane (Fig. 8B). These two types of conformational changes, i.e. the removal of active lid by nonspecific electrostatic repulsion and the interdomain hinge movement by specific PtdIns(4,5)P 2 binding, expose hydrophobic residues at the rim of the active site and allow them to make energetically favorable hydrophobic interactions with membranes by partially penetrating the hydrophobic core of the membranes. Eventually, these processes allow a substrate molecule to enter the active site for catalytic turnover. It should be noted that all our studies were performed with partially phosphorylated proteins and that the effect of phosphorylation is not taken into account in the current model. In view of the fact that the major phosphorylation site of cPLA 2 (Ser 505 ) is located in the interdomain hinge region, it is tempting to assume that the effect of its phosphorylation (i.e. introduction of negative charges) might be similar to that of the PtdIns(4,5)P 2 binding. Further studies on the interplay of phosphorylation and PtdIns(4,5)P 2 binding would provide an important clue to the understanding of the complex cellular membrane targeting and activation mechanism of cPLA 2 .