Membrane Targeting of C2 Domains of Phospholipase C- (cid:1) Isoforms*

The C2 domain is a Ca 2 (cid:2) -dependent membrane-target-ing module found in many cellular proteins involved in signal transduction or membrane trafficking. To understand the mechanisms by which the C2 domain mediates the membrane targeting of PLC- (cid:1) isoforms, we measured the in vitro membrane binding of the C2 domains of PLC- (cid:1) 1, - (cid:1) 3, and - (cid:1) 4 by surface plasmon resonance and monolayer techniques and their subcellular localization by time-lapse confocal microscopy. The membrane binding of the PLC- (cid:1) 1-C2 is driven by nonspecific electrostatic interactions between the Ca 2 (cid:2) -induced cationic surface of protein and the anionic membrane and specific interactions involving Ca 2 (cid:2) , Asn 647 , and phosphati-dylserine (PS). The PS selectivity of PLC- (cid:1) 1-C2 governs its specific Ca 2 (cid:2) -dependent subcellular targeting to the plasma membrane. The membrane binding of the PLC- (cid:1) 3-C2 also involves Ca 2 (cid:2) -induced nonspecific electrostatic interactions and PS coordination, and the latter leads to specific subcellular targeting to the plasma membrane. In contrast to PLC- (cid:1) 1-C2 (Invitrogen) once FBS-supplemented DMEM, overlaid FBS-supple- mented DMEM g/ml ponasterone A to induce protein production. Confocal Microscopy— Images were obtained using a four-channel Zeiss 510 laser scanning confocal microscope. EGFP was excited using the 488-nm line of an argon/krypton laser. 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. An LP 505 filter was used on channel 1 for all experiments. A (cid:3) 63 magnification, 1.2 numerical aperture water immersion objective was used for all experiments. Cells for imaging were selected based on their initial intensity, which needed to fall in the upper third of the photomultiplier tube’s range. The 510 imaging soft- ware provides an option for time series imaging and was used to control the time intervals for imaging. Ca 2 (cid:1) -dependent translocation of C2 domains was monitored as follows: Thirty minutes before imaging, the cells were treated with 2 (cid:5) l of Fura Red AM (Molecular Probes). Im- mediately before imaging, induction media were removed and the cells were washed with 150 (cid:5) l of 2 m M EGTA and then overlaid with 150 which we

Mammalian phosphatidylinositol-specific phospholipases C (PLC) 1 are responsible for converting phosphatidylinositol 4,5-bisphosphate (Ins(4,5)P 2 ) into diacylglycerol and inositol 1,4,5trisphosphate (IP 3 ), which promote the activation of protein kinases C (PKC) and the release of Ca 2ϩ from intracellular stores, respectively (1,2). The PLC family comprises eleven isoforms that can be subdivided into four types (␤, ␥, ␦, and ⑀) based on their structural differences. All PLC isoforms except newly discovered PLC-⑀ possess three regulatory domains: PH, EF-hand, and C2 (2). Among PLC isoforms, Ca 2ϩ -sensitive PLC-␦1 has been the subject of extensive structure-function studies due to the availability of tertiary structural information. Crystallographic structures of PLC-␦1 lacking the aminoterminal PH domain (3) and of its isolated PH domain (4) revealed a four-module organization of the enzyme comprising the amino-terminal PH domain, the EF-hand domain, catalytic domain, and the carboxyl-terminal C2 domain. Based on these structures, it was proposed that PH and C2 domains, both of which are well-characterized membrane-targeting domains, are involved in the membrane targeting of PLC-␦1 (3). The PH domain is a ␤-barrel-like structure that is present in many membrane-binding proteins (5)(6)(7). The essential role of the PH domain in the membrane targeting of PLC-␦1 has been experimentally demonstrated both in vitro and in vivo (8). The PH domain of PLC-␦1 is capable of anchoring the protein to the membrane by specifically binding to Ins(4,5)P 2 in the membrane, and the competitive binding of the PH domain to soluble IP 3 can induce the membrane dissociation of PLC-␦1. However, the role of the C2 domain in PLC-␦1 catalysis remains unclear.
The C2 domain has been identified in many cellular proteins involved in signal transduction or membrane trafficking (9 -11). Many C2 domains bind Ca 2ϩ and mediate Ca 2ϩ -dependent membrane targeting of proteins. Structural analyses of multiple Ca 2ϩ -dependent membrane-binding C2 domains have shown that they share a common fold consisting of an eightstranded antiparallel ␤-sandwich connected by variable loops, which form the binding sites for multiple Ca 2ϩ ions at one end of the domain (3,(12)(13)(14)(15)(16)(17). The crystal structure of the PLC-␦1 C2 domain reveals three metal binding sites in the loop region (18,19), which led to the proposal that the C2 domain is involved in the Ca 2ϩ -dependent membrane targeting of the protein. However, a mutant of PLC-␦1 lacking the C2 domain Ca 2ϩ binding sites showed the same activity toward Ins(4,5)P 2 in phosphatidylcholine (PC)-containing vesicles and micelles as the wild type, suggesting the Ca 2ϩ requirement of PLC-␦1 largely reflects the binding of Ca 2ϩ to the active site (20). More recently, it was shown that, in the presence of phosphatidylserine (PS) in the assay mixture, the C2 domain plays a key role in the activation of PLC-␦1 through the formation of a C2⅐Ca 2ϩ ⅐PS ternary complex (21). To better understand the role * This work was supported by National Institutes of Health Grants GM52598 and GM53987 (to W. C.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Construction of Expression Vectors and Mutagenesis-To subclone the cDNA of PLC-␦1 into the pGEX-4T-1 vector (Amersham Biosciences, Inc., Piscataway, NJ) that contains the amino-terminal glutathione S-transferase (GST) sequence, the start codon was removed and new restriction sites (SmaI and XhoI) were constructed in the gene by overlap extension PCR mutagenesis using Pfu polymerase (Stratagene, La Jolla, CA). Using NdeI and HindIII sites, the cDNAs of PLC-␦3 and PLC-␦4 were subcloned into the pET28a vector (Novagen, Madison, WI) that encodes the amino-terminal His 6 tag and thrombin cleavage site (MGSSHHHHHHSSGLVPRGSH). The isolated C2 domains of all three isoforms were subcloned similarly into the pET28a vector. Mutants were generated by overlap extension PCR mutagenesis.
Protein Expression and Purification-Escherichia coli strain BL21 (for pGEX-4T-1 vector) and BL21(DE3) (for pET28a vector) were used as hosts for protein expression. One liter of Luria broth supplemented with 100 g/ml ampicillin for PLC-␦1 and 50 g/ml kanamycin for PLC-␦3 and PLC-␦4 was inoculated with 1 ml of overnight culture grown at 37°C. Cells were grown at 37°C until the absorbance at 600 nm reached ϳ0.6, then the protein expression was induced with 0.1 mM isopropyl-1-thio-␤-D-galactopyranoside (Research Products, Mount Prospect, IL). After overnight incubation at room temperature, cells were harvested by centrifugation at 5000 ϫ g for 10 min at 4°C. Cells were resuspended in 50 ml of 50 mM Tris-HCl buffer, pH 8.0, containing 50 mM NaCl, 2 mM EDTA, 0.4% (v/v) Triton X-100, 0.4% (w/v) sodium deoxycholate, and 1 mM phenylmethylsulfonyl fluoride, and sonicated. The supernatant was collected by centrifugation at 50,000 ϫ g for 20 min at 4°C. His-tagged proteins (PLC-␦3 and PLC-␦4) were purified using a Ni-NTA-agarose column according to the manufacturer's instructions. The GST fusion protein (PLC-␦1) was purified using a glutathione-Sepharose affinity system as follows. The cell lysate was incubated with GST beads for 1 h at 4°C. The unbound protein impurities were washed with 50 mM Tris-HCl buffer, pH 8.0, containing 150 mM NaCl. The GST-fused protein still bound to the beads was incubated with thrombin for 6 h at 25°C. Then, the cleaved protein was eluted with the same buffer. Isolated C2 domains were prepared as follows. One liter of Luria broth supplemented with 50 g/ml kanamycin was inoculated with 1 ml of overnight culture. Cells were grown until absorbance at 600 nm reached ϳ0.6, then the protein expression was induced with 0.5 mM isopropyl-1-thio-␤-D-galactopyranoside. After 4-h incubation at 37°C, cells were harvested by centrifugation at 5000 ϫ g for 10 min at 4°C. Cells were resuspended in 50 ml of 50 mM Tris-HCl, pH 8.0, containing 50 mM NaCl, 2 mM EDTA, 0.4% (v/v) Triton X-100, and 0.4% (w/v) sodium deoxycholate. After the suspension was soni-cated, the inclusion body pellet was obtained by centrifugation at 50,000 ϫ g for 15 min at 4°C. The pellet was resuspended in the same buffer and re-centrifuged, and the pellet was resuspended in 50 ml of 50 mM Tris-HCl, pH 8.0, containing 5 mM EDTA and 5 M urea. The pellet was stirred for 20 min at room temperature and then centrifuged at 100,000 ϫ g for 10 min at 4°C. The washed inclusion body was resuspended in 10 ml of 50 mM Tris-HCl, pH 8.0, containing 8 M guanidinium chloride. Insoluble matter was removed by centrifugation at 100,000 ϫ g for 10 min at 4°C, and the supernatant was loaded onto a Sephadex G-25 column (2.5 ϫ 25 cm) equilibrated with 50 mM Tris-HCl, pH 8.0, containing 5 M urea and 5 mM EDTA. The first major peak was collected (35 ml) and dialyzed against 50 mM Tris-HCl, pH 8.0, containing 1.5 M urea and then against 50 mM Tris-HCl, pH 8.0. The refolded C2 domain was purified using a Ni-NTA column (Qiagen) according to the manufacturer's instructions. Purity of all protein samples was higher than 90% electrophoretically. Aliquots of purified protein were stored at Ϫ20°C.
Determination of PLC Activity-Activity of PLC was assayed by measuring the initial rate of Ins(4,5)P 2 hydrolysis as described by Cifuentes et al. (23) with some modifications. Small unilamellar vesicles (500 M) containing 1% Ins(4,5)P 2 , a trace of [ 3 H]Ins(4,5)P 2 (2 ϫ 10 4 dpm) and bulk phospholipids (POPC, POPG, or POPS; each 495 M) were prepared in 10 mM HEPES buffer, pH 7.0, containing 0.1 M KCl, 500 g/ml BSA, and 0.5 mM Ca 2ϩ . Free calcium concentration was adjusted using a mixture of EGTA and CaCl 2 according to the method of Bers (24). The reaction was initiated by adding the indicated amount of enzyme, continued for 5 min, and quenched by adding 0.25 ml of 10% ice-cold trichloroacetic acid and 25 l of 20% Trition X-100. Samples were kept on ice for 15 min, and the precipitate containing [ 3 H]Ins(4,5)P 2 was separated from the supernatant containing [ 3 H]IP 3 by centrifugation at 12,000 ϫ g for 2 min at 4°C. To the supernatant, 0.5 ml of CHCl 3 /MeOH (2:1), was added and the aqueous phase containing [ 3 H]IP 3 was extracted. Radioactivity of the hydrolyzed product was measured by liquid scintillation counting.
Monolayer Measurements-Surface pressure () of solution in a circular Teflon trough was measured using a Wilhelmy plate attached to a computer-controlled tensiometer (25). The trough (4 cm diameter ϫ 1 cm depth) has a 0.5-cm deep well for a magnetic stir bar and a small hole drilled at an angle through the wall to allow an addition of protein solution. Five to ten milliliters of phospholipid solution in ethanol/ hexane (1:9 (v/v)) was spread onto 10 ml of subphase (20 mM Tris-HCl, pH 7.5, containing 0.1 M KCl and 0.5 mM free Ca 2ϩ ) to form a monolayer with a given initial surface pressure ( o ). The subphase was continuously stirred at 60 rpm with a magnetic stir bar. Once the surface pressure had been stabilized (after about 5 min), the protein solution (typically 50 l) was injected into the subphase through the hole, and the change in surface pressure (⌬) was measured as a function of time. Typically, the ⌬ value reached a maximum after 20 min. The maximal ⌬ depended on the protein concentration at the low concentration range and reached saturation when the protein concentration was higher than 3 g/ml. Protein concentrations in the subphase were therefore maintained above such values to ensure the observed ⌬ represented a maximal value.
SPR Measurements-The preparation of vesicle-coated Pioneer L1 sensor chip (BIAcore) was described in detail elsewhere (26). The sensor surface was coated with POPC/POPS (7:3) or POPC/POPG (7:3) vesicles. In control experiments, the fluorescence intensity of the flow buffer after rinsing the sensor chip coated with vesicles incorporating 10 mM 5-carboxyfluorescien (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 POPC, because all C2 domains of PLC-␦ isoforms showed negligible binding to POPC-coated chip. The drift in signal for both sample and control flow cells was allowed to stabilize below 0.3 resonance unit/min before any kinetic experiments were performed. All kinetic experiments were performed at 24°C, and a flow rate of 60 l/min in 10 mM HEPES, pH 7.4, containing 0.1 M NaCl and varying concentration of Ca 2ϩ . A high flow rate was used to circumvent mass transport effects. The association was monitored for 90 s and dissociation for 4 min. The immobilized vesicle surface was then regenerated for subsequent measurements using 10 l of 50 mM NaOH. The regeneration solution was passed over the immobilized vesicle surface until the SPR signal reached the initial background value before protein injection. 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. All data were evaluated using BIAevaluation 3.0 software (BIAcore). For each trial, the signal was corrected against the control surface response to eliminate any refractive index changes due to buffer change. Furthermore, the derivative plot was used to monitor potential mass transport effects. Once these factors were checked for each set of data, the association and dissociation phases of all sensorgrams were globally fit to a 1:1 Langmuir binding model: [protein ϫ vesicle] 7 protein ϩ vesicle. The association phase was analyzed using an equation, Rϭ[k a C/(k a Cϩk d )] R max (1Ϫe Ϫ(kaCϩkd)(tϪt0 ))ϩRI where RI ϭ refractive index change, R max is the theoretical binding capacity, C is analyte concentration, k a is the association rate constant, and t 0 is the time at start of fit data. The dissociation phase was analyzed using an equation, RϭR 0 e Ϫkd (tϪt0) where k d is the dissociation rate constant and R 0 is the response at the start of fit data. The curve fitting efficiency was checked by residual plots and 2 . The dissociation constant (K d ) was then calculated from the equation, Construction of Gene Constructs of C2 Domains Fused with EGFP-All constructs were ligated into the modified pIND vector (Invitrogen). An amino-terminal EGFP fusion was found to yield higher gene expression than the carboxyl-terminal counterpart. The spacer sequence between EGFP and the gene was AAA.
Cell Culture-A stable HEK293 cell line expressing the ecdysone receptor (Invitrogen) was used for all experiments. 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 were passaged into eight wells of a Lab-Tech-chambered coverglass for later transfection and visualization. Only cells between the 5th and 20th passages were used.
Transfection and Protein Production-80 -90% confluent cells in Lab-Tech-chambered cover glass wells were exposed to 150 l of unsupplemented DMEM containing 0.5 g of endotoxin-free DNA and 1 l of LipofectAMINE reagent (Invitrogen) for 7-8 h at 37°C. After exposure, the transfection medium was removed, and the cells were washed once with FBS-supplemented DMEM, and overlaid with FBS-supplemented DMEM containing Zeocin and 140 g/ml ponasterone A to induce protein production.
Confocal Microscopy-Images were obtained using a four-channel Zeiss 510 laser scanning confocal microscope. EGFP was excited using the 488-nm line of an argon/krypton laser. 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. An LP 505 filter was used on channel 1 for all experiments. A ϫ63 magnification, 1.2 numerical aperture water immersion objective was used for all experiments. Cells for imaging were selected based on their initial intensity, which needed to fall in the upper third of the photomultiplier tube's range. The 510 imaging software provides an option for time series imaging and was used to control the time intervals for imaging. Ca 2ϩ -dependent translocation of C2 domains was monitored as follows: Thirty minutes before imaging, the cells were treated with 2 l of Fura Red AM (Molecular Probes). Immediately before imaging, induction media were removed and the cells were washed 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, and 6.4 mM sucrose). After initially imaging a cell, 150 l of HEK buffer containing ionomycin and various concentrations of Ca 2ϩ was added to PLC-␦-C2-transfected cells. Control experiments were done with dimethyl sulfoxide in place of ionomycin.
Homology Models and Electrostatic Potential Computation-The electrostatic properties of the structure of PLC-␦1-C2 and homology models for PLC-␦3-C2 and PLC-␦4-C2 and mutants were calculated and visualized in the program GRASP (27). In each panel of the figure, the red and blue meshes represent, respectively, the Ϫ25 and ϩ25 mV electrostatic equipotential contours in 0.1 M KCl. All homology models were built based on the alignment of the sequence being modeled to the PLC-␦1-C2 sequence and using the structure of PLC-␦1-C2 as a template. Previous studies have suggested that PLC-␦1-C2 binds three Ca 2ϩ ions (18). Therefore, in the modeling studies, we used the structure of the PLC-␦1-C2 complexed with three lanthanum ions (1djg, residues 626 -756) (18), which we replaced with Ca 2ϩ ions. Indistinguishable results were obtained with the structure of the C2 domain complexed with two calcium ions (1dji, residues 626 -756) into which we modeled a third Ca 2ϩ ion based on the structural alignment with the lanthanum-bound structure. Homology models were built for PLC-␦3-C2 and PLC-␦4-C2 as well as the S717D and K717E/R718E mutants of PLC-␦4-C2 with the program PRISM (28). Hydrogen atoms were added to the heavy atoms of the structure and homology models with the program CHARMM (29). The structures with hydrogens were subjected to conjugate gradient minimization with a harmonic restraint force of 50 kcal/mole/Å 2 applied to the heavy atoms located at the original crystallographic coordinates to minimize atomic clashes. Each model was evaluated using the program Verify 3D (30), which scores structures according to how well each residue fits into its structural environment based on criteria derived from statistical analyses of the Protein Data Bank; all models scored well relative to the PLC-␦1-C2 domain structure. Rather than build separate models for both the Ca 2ϩ -free and Ca 2ϩ -bound forms of PLC-␦3-C2 and PLC-␦4-C2 and to facilitate comparison of their electrostatic properties, we assumed that the structures in the absence of Ca 2ϩ are similar to those in its presence, and the calcium-free forms of the C2 domains were derived from the Ca 2ϩ -bound models by deleting the Ca 2ϩ ions from the models' coordinate files. This assumption is consistent with the studies of PLC-␦1-C2 (18). Confirming this, our results for Ca 2ϩ -free PLC␦1-C2 are insensitive to whether we used the Ca 2ϩ -free structure (1isd) or deleted the Ca 2ϩ ions from the Ca 2ϩ -bound form.

RESULTS
Membrane Binding of PLC-␦-C2 Domains-Four distinct PLC-␦ isoforms have been identified so far (2). Although all four isoforms are homologous in general, some sequence variations are noticed in the calcium-binding loops in the C2 domains (see Fig. 1). For instance, PLC-␦4-C2 has Ser in place of a calcium-ligating Asp residue (i.e. Asp 708 for PLC-␦1). To determine how these variations affect the membrane binding properties of the C2 domains, we measured the membrane binding of the C2 domains of PLC-␦1, -3, and -4 isoforms by SPR analysis. We have shown that the SPR analysis allows direct determination of membrane association (k a ) and dissociation (k d ) rate constants for peripheral proteins (25,26). We first measured the binding of PLC-␦1-C2 to immobilized POPC/ POPS (7:3) vesicles as a function of Ca 2ϩ concentration (see Table I). In the absence of Ca 2ϩ (i.e. 0.1 mM EGTA), no appreciable binding was detected with protein concentration up to 1 M, indicating that PLC-␦1-C2 has ϾM affinity for POPC/ POPS (7:3) under this condition. As illustrated in Fig. 2, an increase in Ca 2ϩ from 0.5 M to 0.5 mM resulted in a 50-fold increase in the affinity (K d ) of PLC-␦1-C2 for POPC/POPS (7:3), demonstrating that it is a Ca 2ϩ -dependent membrane targeting domain. Interestingly, Ca 2ϩ affected both k a (ϳ6-fold) and k d (ϳ9-fold) to a comparable degree. Our previous study indicated that nonspecific electrostatic interactions primarily accelerate the association of protein to anionic membrane surfaces, whereas hydrophobic interactions and specific interactions, whether electrostatic interactions or hydrogen bonding, mainly slow the membrane dissociation (26). Thus, it appears that Ca 2ϩ ions are involved in both nonspecific electrostatic interactions and specific and/or hydrophobic interactions. As was the case with PLC-␦1-C2, PLC-␦3-C2 (up to 1 M) showed no detectable affinity for immobilized POPC/POPS (7:3) vesicles in the absence of Ca 2ϩ . Again, the membrane affinity of PLC-␦3-C2 increased as a function of Ca 2ϩ . For this C2 domain, increasing the Ca 2ϩ concentration from 0.5 M to 0.5 mM led to a more pronounced 280-fold increase in affinity. Unlike the case with PLC-␦1-C2, Ca 2ϩ affected k a (ϳ90-fold) much more significantly than k d (ϳ3-fold) (see Fig. 3), suggesting that for PLC-␦3-C2 the primary role of Ca 2ϩ is to enhance nonspecific electrostatic interactions with the anionic membrane surface. Lastly, we measured the Ca 2ϩ dependence of membrane binding of PLC-␦4-C2. In contrast to PLC-␦1-C2 and PLC-␦3-C2 that lack the Ca 2ϩ -independent membrane affinity, PLC-␦4-C2 showed relatively high affinity for immobilized POPC/POPS (7:3) vesicles in the absence of Ca 2ϩ (see Table I). Also, raising the Ca 2ϩ level to 0.5 mM resulted in an 18-fold increase in k a and an 1.3-fold decrease in k d ; hence, resulting in a 23-fold decrease in K d . These results indicate that, although PLC-␦4-C2 has high Ca 2ϩ -independent membrane affinity, Ca 2ϩ further promotes its membrane binding by enhancing the nonspecific electrostatic interactions with the anionic membrane surface.
We then measured the PS selectivity of the three C2 domains in the presence of 0.5 mM Ca 2ϩ . As shown in Table I, PLC-␦1-C2 clearly has PS selectivity: It binds POPC/POPS (7:3) membranes 10-fold more strongly than POPC/POPG (7:3) membranes. The difference in K d originates from a 2-fold larger k a and a 5-fold smaller k d for PS-containing membranes. PLC-␦3-C2 also prefers PS to PG, albeit to a smaller degree (i.e. 5-fold smaller K d for PS), which was exclusively due to smaller k d . The fact that the PS preferences of PLC-␦1-C2 and PLC-␦3-C2 derive primarily from k d effects indicates that PS is involved in specific interactions with these C2 domains, which enhances their membrane affinity primarily by slowing the membrane dissociation step. Unlike these two C2 domains, PLC-␦4-C2 did not show PS selectivity. Given that PLC-␦4-C2  Table  I for absolute values of parameters.   Table  I for absolute values of parameters.
has the lowest degree of Ca 2ϩ dependence in PS binding, these data suggest that PLC-␦1-C2 and PLC-␦3-C2 specifically coordinate a PS molecule via Ca 2ϩ .
To further characterize the membrane binding properties of the three C2 domains, we measured their interactions with POPC/POPS (7:3) monolayers in the presence and absence of Ca 2ϩ . The monolayer technique has been shown to be a sensitive tool for assessing the relative membrane penetrating ability of peripheral proteins (25,31). As shown in Fig. 4, all PLC-␦-C2 domains showed significantly lower monolayer penetration than the C2 domain of cytosolic phospholipase A 2 (cPLA 2 ) that has been shown to penetrate into the membrane. Ca 2ϩ and PS did not have much effect on the monolayer penetration of PLC-␦-C2 domains (data not shown). Thus, the membrane binding of PLC-␦-C2 domains does not seem to involve significant membrane penetration and hydrophobic interactions. The predominantly electrostatic nature of membrane binding of PLC-␦-C2 domains is further supported by the SPR binding data showing the lack of appreciable binding of the C2 domains to immobilized POPC/POPS (7:3) membranes in the presence of 0.5 M NaCl (with 0.5 mM Ca 2ϩ ) (data not shown).
PS-dependent Activities of PLC-␦ Isoforms-It was shown that PS enhances the enzymatic activity of PLC-␦1 via the formation of a C2⅐Ca 2ϩ ⅐PS complex (21). Differential PS selectivity of the PLC-␦-C2 domains seen in our membrane binding measurements predicts that the enzyme activities of PLC-␦1, -␦3, and -␦4 should also display different PS dependence. To test this notion, we measured the PS-dependent enzymatic activities of PLC-␦1, -␦3, and -␦4. As shown in Fig. 5, PLC-␦1 showed the most pronounced PS selectivity. The enzyme activity in the presence of POPS vesicles was ϳ10 times higher than that in the presence of POPC or POPG vesicles. PLC-␦3 was activated ϳ7-fold by POPS, whereas PLC-␦4 showed no activation by POPS. Thus, the increases in the enzyme activity of PLC-␦ isoforms by POPS correlate with the increases in the membrane affinity of their C2 domains by POPS. In general, PLC-␦3 and PLC-␦4 were much less active than PLC-␦1. Even in the presence of POPS vesicles, PLC-␦3 was ϳ10 times less active than PLC-␦1. At present, little is known about the activities and specificities of PLC-␦ isoforms for different phosphoinositides. It is thus possible that PLC-␦3 and PLC-␦4 are more active toward other phosphoinositides.

Membrane Binding of PLC-␦-C2 Domain Mutants-To better
understand the mechanisms of Ca 2ϩ -dependent membrane binding of PLC-␦-C2 domains, we prepared selected mutants of PLC-␦1-C2 and PLC-␦4-C2 and measured their membrane interactions by SPR analysis. The x-ray crystal structure of the C2 domain of PKC-␣ showed that a Ca 2ϩ ion and an Asn residue (Asn 189 ) in the Ca 2ϩ -binding loop are involved in specific PS coordination (15). Our recent study confirmed that Asn 189 is the determinant of its PS specificity. 2 The sequence alignment shows that PLC-␦1-C2 also contains an Asn residue (Asn 647 ) in the corresponding position (see Fig. 1). PLC-␦3-C2 and PLC-␦4-C2 have Glu and Thr, respectively, in the position. We thus mutated Asn 647 of PLC-␦1-C2 to Ala (N647A) and measured the mutational effect on PS selectivity. As shown in Table II, N647A showed no PS selectivity; it actually binds PG slightly better. This demonstrates that Asn 647 is directly involved in PS coordination and PS selectivity of PLC-␦1-C2. We then prepared two mutants of PLC-␦4-C2 to elucidate the origin of its unique membrane binding properties. In the x-ray structure of the PLC-␦1-C2, four Asp residues, Asp 653 , Asp 706 , Asp 708 , and Asp 714 , directly participate in metal coordination (18,19). PLC-␦3-C2 has the corresponding four Asp residues. However, PLC-␦4-C2 contains Ser (Ser 717 ) in place of Asp 708 and has a more conservative Asp to Asn substitution in position 714 (of PLC-␦1). Based on this structural information, we first generated S717D to see if the mutation can convert the PLC-␦4-C2 into a PLC-␦1-C2-like molecule. S717D showed no detectable affinity (in terms of K d ) for POPC/POPS membranes in the absence of Ca 2ϩ but at 0.5 mM Ca 2ϩ , S717D had 2-fold higher affinity for POPC/POPS membranes than wild type; as a result, it displayed much larger Ca 2ϩ dependence than wild type (see Table II). Also, S717D had definite PS selectivity. These results thus indicate that the unique membrane binding properties of PLC-␦4-C2 derive in large part from the Asp to Ser substitution at the position 717. We then mutated two cationic residues (Lys 714 and Arg 718 ) in the Ca 2ϩ -binding loop of PLC-␦4-C2 to Glu to assess their contribution to membrane binding. These residues are predicted to be surface-exposed in a homology model structure (see Fig. 6). In the absence of Ca 2ϩ , K714E/R718E had no appreciable affinity for POPC/POPS (7:3) membranes (see Table II), indicating that these cationic residues are the main contributors to Ca 2ϩ -independent mem- brane affinity of PLC-␦4-C2. Even at 0.5 mM Ca 2ϩ , the mutant had 7-fold lower affinity than the wild type, due largely to lower k a , again showing that the two residues are involved in nonspecific electrostatic interactions with the anionic membrane surface, whether the binding is Ca 2ϩ -dependent or not.
Calculation of the Electrostatic Potential-Our SPR and monolayer measurements described above indicated that the membrane binding of the three C2 domains is largely driven by electrostatic interactions. To account for the differential membrane binding properties of the C2 domains and mutants, we therefore compared the electrostatic equipotential profiles of these molecules in the absence and presence of Ca 2ϩ ions. Results are illustrated in Fig. 6. In the absence of Ca 2ϩ , PLC-␦1-C2 and PLC-␦3-C2 had neutral to negative electrostatic potential in the Ca 2ϩ -binding loops, respectively, consistent with their lack of Ca 2ϩ -independent binding to anionic membranes. In contrast, PLC-␦4-C2 that has unique Ca 2ϩ -independent membrane affinity possesses a significant degree of positive electrostatic potential in the Ca 2ϩ -binding loops, particularly around Lys 714 and Arg 718 , even in the absence of Ca 2ϩ . Under the same conditions, however, K714E/R718E has a highly negative electrostatic potential in the same region. Also, the S711D mutation generates a large negative electrostatic potential that overshadows the positive electrostatic potential around Lys 714 and Arg 718 . This accounts for the extremely low Ca 2ϩ -independent affinity of the two mutants for anionic membranes. The addition of three Ca 2ϩ ions (see the "Experimental Procedures") to the C2 domains leads to the development of highly positive electrostatic potentials for all wild type and mutant C2 domains, albeit to different extents. The largest change in electrostatic potential is seen with PLC-␦3-C2, primarily due to the highly negative electrostatic potential in the absence of Ca 2ϩ , which agrees with its most pronounced Ca 2ϩ dependence in binding to anionic membranes. Even with three Ca 2ϩ ions, PLC-␦4-C2-K714E/R718E has a smaller cationic lobe in the Ca 2ϩ -binding loop region than the wild type, again consistent with its lower affinity for anionic membranes than wild type. Taken together, our electrostatic potential calculation is in excellent agreement with the observed membrane binding properties of the PLC-␦-C2 domains and mutants.
Subcellular Translocation of PLC-␦-C2 Domains-To determine the role of PLC-␦-C2 domains in the subcellular localization of PLC-␦ enzymes and also to assess the physiological relevance of our in vitro measurements, we transfected PLC-␦-C2 domains and mutants tagged with EGFP into HEK293 cells and measured their spatiotemporal dynamics by timelapse confocal microscopy. EGFP was linked to the amino-terminal end of each protein, because the carboxyl-terminal attachment interfered with protein overexpression. As shown in Fig. 7, PLC-␦1-C2-EGFP and PLC-␦3-C2-EGFP were evenly dispersed in the cytoplasm in the resting state. When the cells were activated by the Ca 2ϩ ionophore, ionomycin, these C2 domains rapidly translocated to the plasma membrane: The translocation was completed within 5 min. It has been shown that the inner plasma membrane of mammalian cells is rich in PS (32,33). It thus appears that the subcellular localization patterns of PLC-␦1-C2 and PLC-␦3-C2 are consistent with their PS selectivity. For PLC-␦4-C2-EGFP, partial pre-localization of protein near the plasma membrane was seen prior to Ca 2ϩ import. This is not likely due to intracellular Ca 2ϩ , because the cells were pre-treated with excess Fura Red AM (K ca ϭ 140 nM). Interestingly, the cytoplasmic population of PLC-␦4-C2-EGFP moved to various cellular membranes, including plasma membrane and the perinuclear membranes when the cells were activated by ionomycin. This data is consistent with the lack of PS selectivity of PLC-␦4-C2. On the other hand, the S717D mutant of PLC-␦4-C2, which has extremely low Ca 2ϩ -independent affinity but definite PS selectivity, showed no prelocalization and translocated to plasma membrane in response to Ca 2ϩ import. The most dramatic change in subcellular localization was seen with the N647A mutant of PLC-␦1-C2, which has lower overall affinity for anionic phospholipids and greatly reduced PS selectivity (see Table II). For this mutant, clear and preferential translocation to the nuclear envelope was observed in response to Ca 2ϩ import. This result is consistent with the finding that the nuclear envelope has lower content of anionic phospholipids, PS in particular, than plasma membrane (32,33). Together, these cell data show that PLC-␦-C2 domains are Ca 2ϩ -dependent membrane targeting domains and that their subcellular localization is governed in large part by their membrane binding properties. DISCUSSION The role of the C2 domain in the membrane binding and activation of PLC-␦ isoforms has been controversial. This work represents the first systematic in vitro and cell studies on the isolated C2 domains of PLC-␦ isoforms. Our SPR and monolayer measurements, electrostatic potential calculation, and cell translocation studies show that PLC-␦-C2 domains are Ca 2ϩ -dependent membrane targeting domains, the membrane binding of which is driven mainly by electrostatic interactions. The predominantly electrostatic nature of their membrane binding is supported by low monolayer penetration and high dependence on the ionic strength. At least three roles have been proposed for the C2-bound Ca 2ϩ ions (11,34): i.e. nega- tive-to-positive electrostatic potential switch, formation of a protein-Ca 2ϩ -anionic lipid complex, and inducing conformational changes. Previous studies have suggested that PLC-␦1-C2 binds three Ca 2ϩ ions (18). Although we did not determine the stoichiometry of Ca 2ϩ binding to other PLC-␦-C2 domains, it is reasonable to assume, based on the sequence homology, that PLC-␦3-C2 binds three Ca 2ϩ ions. Due to the substitution of two of four Ca 2ϩ -coordinating Asp residues, PLC-␦4-C2 might bind less Ca 2ϩ ions. Our electrostatic potential calculation showed that binding of two Ca 2ϩ ions to PLC-␦4-C2 results in only a slightly smaller positive electrostatic potential than that due to three ions (data not shown). Thus, we analyzed our membrane binding data of PLC-␦-C2 domains based on the assumption that all C2 domains bind three Ca 2ϩ ions, albeit with different affinities. Our recent studies on the C2 domains of PKC-␣ and cPLA 2 , both of which bind two Ca 2ϩ ions, showed that the two Ca 2ϩ ions play distinct roles (35)(36)(37). Likewise, these studies indicate that Ca 2ϩ ions play two different roles for PLC-␦1-C2; i.e. bridging to PS and inducing electrostatic switch. For the C2 domain of PKC-␣ that has been shown to directly coordinate a PS molecule via Ca 2ϩ and an Asn residue (37), Ca 2ϩ has comparable effects on k a and k d and PS has an effect primarily on k d . As shown in Table I, similar observations were made for PLC-␦1-C2 in terms of effects of Ca 2ϩ and PS on k a and k d , respectively. This, in conjunction with high PS selectivity of PLC-␦1-C2, suggests that at least one Ca 2ϩ ion of PLC-␦1-C2 is directly involved in PS coordination. As was the case with the C2 domain of PKC-␣, the putative PS coordination by PLC-␦1-C2 is also mediated by Asn 647 , the mutation of which drastically reduces the PS selectivity. In addition to the PS-coordinating role of Ca 2ϩ , the good agreement between our electrostatic potential calculations and SPR binding data points to the importance of nonspecific electrostatic interactions that are induced by Ca 2ϩ binding. Ca 2ϩ ions switch the predominantly negative electrostatic potential of the Ca 2ϩ binding loops, due to the presence of anionic Ca 2ϩ ligands, to a strongly positive electrostatic potential and thereby drive the membrane association. Although it was reported that the Ca 2ϩ analogue Sm 3ϩ induces the local conformational changes of PLC-␦1-C2 (19), the importance of this conformational effect is not evident from our studies. For PLC-␦3-C2, the main role of Ca 2ϩ ions appears to be the electrostatic switch, because PLC-␦3-C2 shows the highest degree of Ca 2ϩ dependence in membrane affinity (K d ) that is mainly due to a k a effect. PLC-␦3-C2 has a significant degree of PS selectivity (ϳ5-fold) even though it contains a Glu in place of the PS-binding Asn 647 of PLC-␦1-C2 (see Fig. 1). It is unclear whether this residue or an alternative site is involved in binding to a PS molecule. Further studies are required to identify residues, if any, that are involved in putative PS coordination of PLC-␦3-C2. Lastly, it is clear that the primary role of Ca 2ϩ ions in the membrane binding of PLC-␦4-C2 is the electrostatic switch. Ca 2ϩ -enhanced affinity for anionic lipids is insensitive to the phospholipid headgroup and is ascribed exclusively to a k a effect. Thus, Ca 2ϩ ions shift the electrostatic potential of the calcium bind- ing loop region of PLC-␦4-C2 to a highly positive one and thereby allow it to interact more strongly but nonspecifically with anionic membranes. The lack of PS selectivity of PLC-␦4-C2 indicates that it does not directly coordinate PS via Ca 2ϩ and other protein residues. For this C2 domain, cationic residues, Lys 714 and Arg 718 , confer significant Ca 2ϩ -independent affinity for anionic membranes. The high Ca 2ϩ -independent affinity of PLC-␦4-C2, relative to PLC-␦1-C2 and PLC-␦3-C2, also derives from the lower negative electrostatic potential in its Ca 2ϩ -binding loops due to the lack of two Asp residues (Asp 708 and Asp 714 in PLC-␦1). The undetectably low Ca 2ϩindependent affinity of S717D, which introduces an anionic residue into the Ca 2ϩ -binding loops, is consistent with this model.
The mechanisms by which PLC-␦ isoforms are regulated in the cell and their coupling to membrane receptors remain unclear (2). PLC-␦ isoforms show the highest Ca 2ϩ sensitivity among all known PLC enzymes (2). Our studies indicate that differential Ca 2ϩ -dependent membrane binding properties of the three C2 domains, different PS selectivity in particular, result in their different subcellular localization behaviors. Although the exact lipid composition of different cellular membranes of HEK293 cells has not been determined yet, it is expected from the known lipid compositions of mammalian subcellular membranes (32, 33) that its inner plasma membrane is rich in PS and the perinuclear membranes, including the nuclear envelope, contains higher PC content and lower anionic lipids. In response to Ca 2ϩ import, PLC-␦1-C2 and PLC-␦3-C2, which exhibit PS selectivity, translocate to the PS-rich plasma membrane. N647A of PLC-␦1-C2, which has dramatically reduced PS selectivity, is localized to both the plasma membrane and the nuclear envelope, supporting the notion that the specific targeting of PLC-␦1-C2 to plasma membrane is due to its PS selectivity. Similarly, the plasma membrane targeting of PLC-␦3-C2 appears to be due to its PS selectivity. It has been shown that the PH domain plays a critical role in the binding of PLC-␦1 to the membrane containing Ins(4,5)P 2 . Our results indicate that the favorable Ca 2ϩ -dependent interactions between PLC-␦1-C2 (and PLC-␦3-C2) and PS molecules would augment the targeting of PLC-␦1 (and PLC-␦3) to the plasma membrane containing Ins(4,5)P 2 . Thus, PLC-␦1-C2 (and PLC-␦3-C2) in response to Ca 2ϩ rise would not only accelerate the membrane binding of PLC-␦1 but also ensure its specific targeting to the plasma membrane. Lastly, the subcellular localization pattern of PLC-␦4-C2 is consistent with its membrane binding affinity: A definite degree of pre-localization to the plasma membrane is seen in the absence of Ca 2ϩ import and the cytoplasmic population of PLC-␦4-C2 translocates nonspecifically to various subcellular membranes in response to Ca 2ϩ import. In addition, the Ca 2ϩ -independent prelocalization and the non-selective Ca 2ϩ -dependent targeting were not observed for the S717D mutant that has greatly reduced Ca 2ϩ -independent binding but definite PS selectivity. Although the physiological significance of this finding awaits further investigation that is beyond the scope of this work, the unique properties of PLC-␦4-C2 suggest the existence of a novel regulatory mechanism for PLC-␦4. For instance, PLC-␦4 might be selectively activated at lower intracellular Ca 2ϩ levels. Also, the specific targeting of PLC-␦4 might involve protein-protein interactions, given the nonspecific nature of its membrane targeting.
In summary, this work establishes the C2 domains of PLC-␦ isoforms as Ca 2ϩ -dependent membrane targeting domains that have distinct membrane binding properties and play a major role in their subcellular localization behaviors. As such, these studies lay the foundation for further investigation of the complex mechanisms of membrane targeting and activation of PLC-␦ isoforms, which involve the interplay among different domains and the interactions with calcium, phosphoinositides, and other factors.