Essential Role of the EF-hand Domain in Targeting Sperm Phospholipase Cζ to Membrane Phosphatidylinositol 4,5-Bisphosphate (PIP2)*

Background: The mechanism underlying sperm PLCζ interaction with its target membrane is unresolved. Results: EF-hand mutations introduced into PLCζ reduce in vivo Ca2+ oscillation inducing activity and in vitro interaction with PIP2. Conclusion: EF-hand domain is essential for targeting PLCζ to PIP2-containing membranes. Significance: We propose a novel mechanism by which sperm PLCζ is anchored to its physiological membrane substrate.

During fertilization, the spermatozoon initiates activation of egg development by triggering an acute rise in cytosolic free Ca 2ϩ concentration (1). In mammals, this manifests as a series of distinctive cytosolic Ca 2ϩ oscillations, beginning soon after sperm-egg fusion and persisting for several hours (2). The weight of evidence now suggests that Ca 2ϩ oscillations appear to be caused by a sperm-specific protein, phospholipase C-(PLC), 3 which is introduced into the egg upon sperm-egg fusion and leads to cycles of inositol 1,4,5-trisphosphate (IP 3 ) production following PIP 2 hydrolysis, thus activating IP 3 receptor-mediated Ca 2ϩ release from intracellular stores in the egg (3)(4)(5)(6)(7)(8)(9)(10)(11)(12). The closest PLC homologue of sperm PLC is PLC␦1 (47% similarity, 33% identity), which is only able to cause Ca 2ϩ oscillations in mouse eggs at non-physiological concentrations, because it has a Ͼ50-fold lower potency (2,3,12). The superior fertilization potency of the sperm PLC over somatic PLCs has not yet been fully explained.
PLC is the smallest PLC with the simplest domain organization among all the mammalian isoforms. PLC consists of four tandem EF-hand domains, the characteristic X and Y catalytic domains in the center of the molecule, and a C-terminal C2 domain. All these domains are common to the other PLC isoforms (␤, ␥, ␦, ⑀, and ), but they appear to individually have an essential role in the unique mode of regulation of this distinctive PLC isozyme (2). A notable structural difference between PLC and the other somatic PLC isoforms is that PLC lacks a pleckstrin homology (PH) domain at the N terminus (2,3,13). The membrane binding of somatic PLCs appears to be mediated by the PH domain, a well defined structural module of ϳ120-amino acid residues identified in numerous proteins (14). The PH domain of PLC␦1 is essential for interaction with its phospholipid substrate PIP 2 in the plasma membrane (15). The absence of a PH domain from PLC sequence raises questions about how PLC can bind to membranes.
We have previously proposed that the PLC XY-linker, a segment between the X and Y catalytic domains that is notably different from the corresponding XY-linker region of somatic PLCs, is involved in the targeting of PLC to its membranebound substrate PIP 2 (16,17). The XY-linker region of PLC is extended in length and consists of more basic residues relative to its PLC␦1 counterpart. The affinity of the XY-linker for PIP 2 appears to involve a polybasic charged region that is found in a number of other membrane-associated proteins (16,18). These positively charged amino acids in the XY-linker appear to assist the anchoring of PLC to membranes by enhancing the local PIP 2 concentration adjacent to the XY catalytic domain via electrostatic interactions with the negatively charged PIP 2 (16,17). However, the XY-linker might not be the only domain that mediates the binding of PLC to PIP 2 -containing membranes. We have demonstrated that the absence of the XY-linker from PLC significantly diminishes, but does not completely abolish, the in vivo Ca 2ϩ oscillation inducing activity (19). This suggests that other domain(s) may also be involved in anchoring PLC to its target membrane.
A recent study reported that the N-terminal lobe of the EFhand domain of PLC␦1 binds anionic phospholipids, and this binding is due to interactions with cationic and hydrophobic residues in the first EF-hand sequence of PLC␦1 (20). The authors propose a general mechanism that may apply to other PLC isoforms by suggesting that EF-hand domain interactions with anionic phospholipids in the target membrane provides a tether that facilitates proper substrate access and binding in the active site (20). Importantly, the cationic residues in the first EF-hand domain of PLC␦1 that contribute to anionic lipid vesicle binding are all conserved in PLC.
The aim of this study is to investigate the potential importance of a conserved cluster of cationic residues at the N-terminal lobe of the EF-hand domain of PLC in association with anionic lipids and its substrate PIP 2 . A series of full-length mouse PLC mutants were prepared that sequentially neutralized two positively charged lysine and one arginine residues within the first EF-hand domain. The Ca 2ϩ oscillation-inducing properties of these mutants were experimentally tested relative to wild-type PLC by microinjection of cRNA into unfertilized mouse eggs. The various PLC mutants' enzymatic properties were analyzed using an in vitro PIP 2 hydrolysis assay. A protein-lipid overlay and a liposome binding/enzyme assay were employed to assess the binding properties of wild-type PLC to phosphatidylserine (PS), phosphatidic acid (PA), and PIP 2 . Furthermore, the binding properties of mutant EFhand PLC proteins to PIP 2 were examined. Our results suggest that PLC possesses significant affinity only for PIP 2 but not for PA or PS. We also find that sequential reduction of the net positive charge within the first EF-hand domain significantly reduces both in vivo Ca 2ϩ oscillation inducing activity and the in vitro interaction of PLC with PIP 2 . Moreover, we show that a PLC mutant where three cationic residues within the first EF-hand domain and three cationic residues within the XYlinker region of PLC were substituted by alanine is unable to trigger Ca 2ϩ oscillations in mouse eggs. In vitro biochemical characterization suggests that this PLC mutant displays dramatically reduced binding to PIP 2 -containing liposomes compared with the wild-type PLC. Thus, we propose a novel mechanism for the sperm PLC interaction with PIP 2 -containing membranes mediated by electrostatic interactions between the anionic PIP 2 with both the first EF-hand domain and the XYlinker region of PLC, which are rich in cationic residues.
All the above PLC mutants were amplified from their corresponding pCR3 plasmid with the appropriate primers to incorporate a 5-SalI site and a 3-NotI site, and the products were cloned into the pETMM60 vector to enable bacterial protein expression. The primers used for the amplifications were as follows: 5Ј-GAACGTCGACATGGAAAGCCAACTTCATG-AGCTCGC-3Ј (forward) and 5Ј-GGAAGCGGCCGCTCACT-CTCTGAAGTACCAAAC-3Ј (reverse). Successful mutagenesis and cloning of the above expression vector constructs were confirmed by dideoxynucleotide sequencing (Applied Biosystems Big-Dye Version 3.1 chemistry and model 3730 automated capillary DNA sequencer by DNA Sequencing & Services TM ).
cRNA Synthesis-Following linearization of wild-type and mutated PLC plasmids, cRNA was synthesized using the mMessage Machine T7 kit (Ambion) and then was polyadenylated using the poly(A) tailing kit (Ambion), as per the manufacturer's instructions.
Preparation and Handling of Gametes-Female mice were super-ovulated and mature MII eggs were collected from excised oviducts 13.5-14.5 h after injection of human chorionic gonadotrophin and maintained in droplets of M2 media (Sigma) under mineral oil at 37°C. Experimental recordings of Ca 2ϩ release or luciferase expression were carried out with mouse eggs in Hepes-buffered media (H-KSOM), as described previously (22). All compounds were from Sigma unless stated otherwise. All procedures using animals were performed in accordance with the United Kingdom Home Office Animals Procedures Act and were approved by the Cardiff University Animals Ethics Committee.
Microinjection and Measurement of Intracellular Ca 2ϩ and Luciferase Expression-Mouse eggs were washed in M2 and microinjected with cRNA diluted in injection buffer (120 mM KCl, 20 mM Hepes, pH 7.4). The volume injected was estimated from the diameter of cytoplasmic displacement caused by the bolus injection. All injections were 3-5% of the oocyte volume. Eggs were microinjected with the appropriate cRNA in the injection buffer, mixed with an equal volume of 1 mM Oregon Green 1,2-bis(2-aminophenoxy)ethane-N,N,NЈ,NЈ-tetraacetic acid-dextran (Life Technologies, Inc.). Eggs were then maintained in H-KSOM containing 100 M luciferin and imaged on a Nikon TE2000 microscope equipped with a cooled intensified CCD camera (Photek Ltd., UK). The luminescence (luciferase expression) and fluorescence (for Ca 2ϩ measurements) from eggs were collected by switching back and forth between the two modes on a 10-s cycle (23,24). These two signals were then displayed as two separate signals over the same time period for each egg. The fluorescent light used to measure Ca 2ϩ is shown in relative units. Luminescence was recorded as photon counts/s and plotted as a running average over 5 min. All live imaging experiments on eggs were made during a 1-month period.
Protein Expression and Purification-For NusA-His 6 -fusion protein expression, Escherichia coli (BL21-CodonPlus(DE3)-RILP; Stratagene) cells were transformed with the appropriate pETMM60 plasmid and cultured at 37°C until the A 600 reached 0.6, and protein expression was induced for 18 h at 16°C with 0.1 mM isopropyl 1-thio-␤-D-galactopyranoside (ForMedium). Cells were harvested (6000 ϫ g for 10 min), resuspended in PBS containing a protease inhibitor mixture (EDTA-free; Roche Applied Science), and sonicated four times for 15 s on ice. Soluble NusA-His 6 -tagged fusion protein was purified on nickelnitrilotriacetic acid resin following standard procedures (Qiagen) and eluted with 250 mM imidazole. Eluted proteins were dialyzed overnight (10,000 molecular weight cutoff; Pierce) at 4°C against 4 liters of PBS and concentrated with centrifugal concentrators (Sartorius; 10,000 molecular weight cutoff).
Assay of PLC Activity-PIP 2 hydrolytic activity of recombinant PLC proteins was assayed as described previously (17,21). The final concentration of PIP 2 in the reaction mixture was 220 M, containing 0.05 Ci of [ 3 H]PIP 2 . The assay conditions were optimized for linearity, requiring a 1-min incubation of 200 pmol of PLC protein sample at 25°C. In assays to determine dependence on PIP 2 concentration, 0.05 Ci of [ 3 H]PIP 2 was mixed with cold PIP 2 to give the appropriate final concentration. In assays examining Ca 2ϩ sensitivity, Ca 2ϩ buffers were prepared by EGTA/CaCl 2 admixture, as described previously (17,21).
Protein Lipid Overlay Assay-PIP array membranes (Echelon Biosciences) were blocked for 2 h with binding buffer (TBS-T: 20 mM Tris, 137 mM NaCl, 0.1% Tween 20, pH 7.4) containing 3% bovine serum albumin (lipid-free) and incubated with 25 pmol of each NusA-PLC fusion protein for 1 h at room temperature. After washing three times in TBS-T, NusA-PLC fusion protein interaction with the inositol phosphate lipids was detected by first incubating the PIP array membranes with penta-His monoclonal antibody (Qiagen, 1:5000 dilution in 5 ml of binding buffer) overnight at 4°C, followed by three 15-min washes. This was followed by incubation with horseradish peroxidase-conjugated anti-mouse antibody in the same binding buffer for 1 h at room temperature, followed by three 15-min washes with TBS-T. Detection of horseradish peroxidase-coupled secondary antibody was achieved using enhanced chemiluminescence detection (ECL; Amersham Biosciences).
Mathematical Modeling of Oocyte Ca 2ϩ Dynamics-Theoretical predictions of the oscillatory Ca 2ϩ activity associated with the various PLC constructs were provided by a mathematical model of oocyte IP 3 /Ca 2ϩ dynamics. The mathematical model, which has previously been presented in detail (25), employs three inter-dependent variables, namely free cytosolic Ca 2ϩ , Ca 2ϩ sequestered in the endoplasmic reticulum, and intracellular concentrations of IP 3 . To account for the specific binding activity of each PLC variant, the effective activity of a PLC concentration is defined as: V e PLC ϭ V PLC ⅐b, where V PLC is the nominal PLC concentration, and b is the binding activity estimated experimentally for each construct. Coefficient b assumes a value between 0 and 1, whereas V PLC can assume values beyond the physiological range when the protein is overexpressed. The mathematical model was coded and numerically integrated on both Cϩϩ and a MATLAB platform (MathWorks).

Effect of EF-hand Mutations on PLC-mediated Ca 2ϩ Oscillations in Mouse
Eggs-To investigate the potential importance of a cluster of cationic residues within the first EF-hand unit of the first pair of PLC EF-hand domains ( Fig. 1), we performed site-directed mutagenesis to produce a panel of cumulative mutations within this positively charged region of the fulllength mouse PLC. Thus, the residues Lys-49, Lys-53, and Arg-57 were sequentially substituted by the neutral amino acid, alanine, to create three single (PLC K49A , PLC K53A , and PLC R57A ) mutants, as well as one double (PLC K53A,K57A ) and one triple (PLC K49A,K53A,R57A ) PLC mutant. To test the Ca 2ϩ oscillation inducing activity of PLC K49A , PLC K53A , PLC R57A , PLC K49A,R57A , and PLC K49A,K53A,R57A mutants and to verify that these constructs were faithfully expressed as proteins in cRNA-microinjected mouse eggs, we generated C-terminal luciferase-tagged versions of these constructs to enable quantitation of relative protein expression by luminescence detection of the expressed PLC-luciferase fusion protein, as described previously (17,21). Prominent Ca 2ϩ oscillations were observed in PLC WT -luciferase cRNA-injected mouse eggs (9.7 spikes in the 1st h of oscillations) following successful protein expression to a level indicated by a luminescence reading of 0.47 counts/s ( Fig. 2 and Table 1), in accord with previous reports (17,21). Microinjection of cRNA encoding the three single PLC mutants (PLC K49A , PLC K53A , and PLC R57A ) also triggered Ca 2ϩ oscillations (Fig. 2), but these exhibited a lower frequency relative to PLC WT (3.6, 4.4, and 4.3 spikes in the 1st h, respectively), although the proteins were expressed at comparable expression levels (Table 1). Similarly, egg microinjection with cRNA encoding either the double PLC K49A,R57A or the triple PLC K49A,K53A,R57A mutant resulted in a significant reduction in the frequency of Ca 2ϩ oscillations compared with PLC WT , causing 3.7 and 2.8 spikes/1 h, respectively, again when protein was expressed at comparable levels ( Fig. 2 and Table 1). These data indicate that the substitution of even one Lys or Arg residue for a neutral Ala within the positively charged cluster of the PLC EF-hand domain can significantly alter their Ca 2ϩ oscillation inducing activity in mouse eggs by reducing the frequency of Ca 2ϩ spikes.
Overexpression of PLC K49A,K53A,R57A in Mouse Eggs Rescues Its Defective Ca 2ϩ Oscillation-inducing Phenotype-Judging by the number of Ca 2ϩ spikes observed within the 1st h of oscillations per unit of recombinant fusion protein expression (cps), PLC WT can be seen to be about ϳ3.5 times more effective at causing Ca 2ϩ oscillations than the PLC K49A,K53A,R57A triple mutant. To investigate whether we could rescue the low frequency of Ca 2ϩ oscillations induced by PLC K49A,K53A,R57A , we overexpressed this PLC mutant in mouse eggs. As shown in Fig. 3 and Table 1, the overexpression of PLC K49A,K53A,R57A (7.65 cps) indeed led to 8.6 spikes in the 1st h of oscillations, comparable with that for PLC WT , suggesting that loading the egg with large amounts of this PLC mutant can rescue its defective Ca 2ϩ oscillation-inducing phenotype.
Expression and Enzymatic Characterization of PLC EFhand Mutants-Each of the PLC K49A , PLC K53A , PLC R57A , PLC K49A,R57A , and PLC K49A,K53A,R57A mutants was subcloned into the pETMM60 vector and purified as NusA-His 6 fusion proteins by affinity chromatography. We have recently demonstrated that NusA is an effective fusion protein partner for PLC, significantly increasing soluble expression of PLC protein in E. coli, as well as enhancing the enzymatic stability of the purified protein over time (11). Following expression of NusA-PLC fusion proteins in E. coli and purification by nickel-nitrilotriacetic acid affinity chromatography, samples of each protein were analyzed by SDS-PAGE followed by Coomassie Brilliant Blue staining and immunoblotting using an anti-NusA monoclonal antibody. Fig. 4A shows that the major protein band following affinity isolation, with mobility corresponding to the predicted molecular mass of ϳ134 kDa for each construct, was present for all fusion proteins analyzed (left panel), and these major bands were also recognized in the corresponding anti-NusA immunoblot (right panel), confirming the appropriate expression of all PLC mutants. Some intermediate molecular mass bands detected by the anti-NusA antibody are the probable result of some degradation occurring through the various protein expression and purification procedures. Similarity of protein expression profile, including degradation products, for the various PLC constructs being examined suggests that experimental comparison of relative enzymatic data may be appropriate. Hence, the specific PIP 2 hydrolytic enzyme activity for PLC WT and each recombinant mutant protein was determined by the standard micellar [ 3 H]PIP 2 hydrolysis assay. The histogram of Fig. 4B and Table 2 summarize the enzyme specific activity values obtained for each recombinant protein. The enzymatic activities of all recombinant proteins was very similar, suggesting that mutating the basic residues of the first pair of EF-hands to a neutral residue has no effect on the ability of PLC to hydrolyze PIP 2 in vitro. Moreover, to investigate the impact of the EF-hand mutations on Ca 2ϩ sensitivity of PLC enzyme activity, we assessed the ability of these PLC recombinant proteins to hydrolyze [ 3 H]PIP 2 at different Ca 2ϩ concentrations ranging from 0.1 nM to 0.1 mM. These experiments indicated that there was no significant difference in the Ca 2ϩ sensitivity of PIP 2 hydrolysis for the wild type, and the five EF-hand mutants (Fig. 4C) with a very similar EC 50 value (67-85 nM) displayed by all recombinant PLC proteins (Table 2). To compare the enzyme kinetics of wild-type and mutant PLCs, the Michaelis-Menten constant, K m , was calculated for each construct ( Table 2). The K m values obtained were similar for human PLC WT (84 M), PLC K49A (121 M), and PLC R57A (115 M), whereas the K m value for PLC K53A (169 M) and PLC K49A,R57A (219 M) mutants was ϳ2and ϳ2.6-fold higher compared with that of PLC WT . Interestingly, the K m value for PLC K49A,K53A,R57A (432 M) was ϳ5.1-fold higher compared with PLC WT (84 M), suggesting that replacement of these three positively charged residues within the first EF-hand domain affects the in vitro affinity of PLC for PIP 2 without affecting the Ca 2ϩ sensitivity of this enzyme.
Binding of PLC to PS, PA, and PIP 2 -To examine the ability of PLC to bind the membrane lipids, PS, PA, and PIP 2 , we employed three different approaches. First, we used a proteinlipid overlay assay to assess the binding of PLC to membranespotted arrays of inositol phospholipids containing PS, PA, or PIP 2 . As shown in Fig. 5A, no binding to PS or PA was evident, although PLC was able to bind to membrane arrays containing PIP 2 . This result is consistent with our liposome binding assays (Fig. 5B). For these binding assays, we made unilamellar liposomes composed of phosphatidylcholine/CHOL/phos- phatidylethanolamine (4:2:1) with incorporation of either 5% PS, 1 or 5% PA, and 1% PIP 2 . To diminish any nonspecific protein binding to highly charged lipids, the liposome binding assays were performed in the presence of a near-physiological concentration of MgCl 2 (0.5 mM). PLC displayed robust binding only to liposomes containing 1% PIP 2 , whereas the protein was only detected in the supernatant of liposomes containing 5% PS and 1 or 5% PA (Fig. 5B). Finally, we incubated 1 g of PLC recombinant protein with the liposomes composed of the different phospholipids, and after centrifugation, the supernatants were separated and assayed for their ability to hydrolyze PIP 2 in vitro, using the standard [ 3 H]PIP 2 hydrolysis assay. As shown in Fig. 5C, only the supernatant obtained after the interaction of recombinant PLC protein with the liposomes containing 1% PIP 2 showed a dramatic ϳ94% reduction in its PIP 2 hydrolytic activity. All these data suggest that the PLC binds specifically to PIP 2 , not generically to any anionic phospholipid.

Properties of PLC-luciferase EF-hand mutants expressed in unfertilized mouse eggs
Ca 2ϩ oscillation inducing activity (number of Ca 2ϩ spikes in the 1st h of oscillations) and luciferase luminescence levels (counts/s of luminescence in 1st h of oscillating) are summarized for mouse eggs microinjected with each of the PLC-luciferase mutants as follows: PLC K49A , PLC K53A , PLC R57A , PLC K53A,R57A , PLC K49A,K53A,R57A , PLC DMM , and wild type PLC-luciferase (see Figs. 2, 3, and 8B). The data are expressed to two significant figures, with means Ϯ S.E. Ratios of total luminescence (counts) per spike in the 1st h of oscillating are also shown expressed to two significant figures with means Ϯ S.E. This is not shown where the number of spikes is zero or where the expression of PLC is sufficiently high that the relationship between number of spikes and expression is no longer linear. The results of Mann-Whitney tests for significant differences between the number of spikes for each of the wild-type and mutant PLC-luciferase constructs are indicated with p values. The results of this same test for a significant difference between the PLC mutant ratios (counts/spike) and the wild-type PLC-luciferase ratio are also denoted in this way. NA, not applicable. Binding of PLC EF-hand Mutants to PIP 2 -containing Liposomes-To investigate the effect of cumulative EF-hand mutations on the PIP 2 -binding properties of wild-type PLC, we employed the liposome/activity binding assay as described above (see Fig. 5C). Thus, 1 g of recombinant protein corresponding to PLC WT and the five EF-hand mutants were each incubated with liposomes containing 1% PIP 2 . After centrifugation, the supernatants were separated, and the PIP 2 hydro-lytic activity was assayed using the standard [ 3 H]PIP 2 hydrolysis assay. Based on the percentage of the PIP 2 hydrolytic activity pre-and post-liposome binding, we estimated the relative binding of each PLC protein to the PIP 2 -containing liposomes. As shown in Fig. 6, although 94% of PLC WT bound to the liposomes, the three single EF-hand mutants (PLC K49A , PLC K53A , and PLC R57A ) showed ϳ71-75% liposome binding. The effect of the double and the triple mutation was even more notable, as PLC K53A,K57A displayed ϳ59% and PLC K49A,K53A,R57A ϳ49% relative liposome binding. These data indicate that sequential neutralization of the basic residues within the EF-hand region substantially reduces the PIP 2 -binding ability of PLC.

PLC
Modeling of Ca 2ϩ Oscillations Induced by PLC EF-hand Mutants-The Ca 2ϩ oscillatory activity associated with each of the PLC mutants constructed was simulated by using the parameters calculated in Fig. 6 and Tables 1 and 2. The most marked differentiation between constructs is the binding activity of each protein (Fig. 6), which is in agreement with a progressive destabilization of the EF-hand binding regime. By contrast, the Ca 2ϩ dependence of IP 3 production (plotted in Fig. 4 and quantified in Table 2 as Ca 2ϩ -dependent EC 50 value) is very similar for each of the PLC constructs. Ca 2ϩ oscillations simulated with this set of parametric values (Fig. 7, top panel) closely match those observed experimentally for each construct (Fig. 2) in terms of frequency. The theoretical relationship between Ca 2ϩ oscillatory frequency and binding activity was produced by the mathematical model for EC 50 ϭ 65, 75, and 85 nM (Fig. 7, bottom panel,, lines left to right). The experimentally computed operating points of PLC wild-type and its various constructs (Fig. 7, bottom panel, circles) are located very close to the theoretical curves, confirming that the variability in Ca 2ϩ oscillatory frequency can be accounted for almost exclusively by the gradual reduction in binding activity. When PLC K49A,K53A,R57A was highly overexpressed, the oscillatory activity was largely restored, as indicated by the operating point of this scenario (Fig. 7, bottom panel, solid circle at the right of the panel). The fact that the circle lies below the theoretical frequency curve (Fig. 7, bottom panel, dashed line) may be due to the sub-optimal binding of the protein to PIP 2 at non-physiologically elevated concentrations.
Ca 2ϩ Oscillation Inducing Activity of PLC Double Motif Mutant Expressed in Mouse Eggs-To investigate whether there is synergy between the cationic residues of the first EF-hand FIGURE 5. In vitro binding of wild-type PLC to PS, PA and PIP 2 . A, PLC protein-lipid overlay assays. Recombinant protein binding to spotted phospholipids on the PIP arrays was detected using the monoclonal penta-His antibody. B, liposome "pulldown" assay of PLC. Unilamellar liposomes containing either PS (5%), or PA (1 or 5%), or PIP 2 (1%) were incubated with PLC recombinant protein. Following liposome centrifugation, both the supernatant (s) and liposome pellet (p) were subjected either to SDS-PAGE and Coomassie Brilliant Blue staining. C, supernatants were assayed for their ability to hydrolyze PIP 2 in vitro, using the standard [ 3 H]PIP 2 hydrolysis assay, n ϭ 4 Ϯ S.E., using two different preparations of recombinant protein. Significant statistical differences (asterisks) were calculated by an unpaired Student's t test; ***, p Ͻ 0.0005 (GraphPad, Prism 5).  domain and the XY-linker region of PLC and whether these residues are necessary and sufficient to anchor this sperm protein to its PIP 2 -containing membranes, we generated a PLC mutant, in which charge-neutralization mutations were introduced within these two PLC motifs. Thus, the residues Lys-49, Lys-53, and Arg-57 within the first EF-hand domain and the residues Lys-374, Lys-375, and Lys-377 within the XY-linker of PLC were substituted by the neutral Ala residue giving rise to a PLC double motif mutant (PLC K49A,K53A,R57A,K374A,K375A,K377A ; PLC DMM ) containing six neutralization mutations (Fig. 8A). Interestingly, microinjection of cRNA encoding a luciferase-tagged version of PLC DMM failed to cause any Ca 2ϩ release, even after relatively high levels of protein expression in unfertilized mouse eggs ( Fig. 8B and Table 1).
To investigate whether the luciferase-tagged PLC WT and PLC DMM fusion constructs were expressed as structurally intact proteins in mouse eggs, we performed immunoblot anal-  Fig. 6 and Tables 1 and 2. The theoretical relationship between PIP 2 binding activity of the PLC constructs and the Ca 2ϩ oscillatory frequency is plotted in the bottom panel as a solid line for EC 50 ϭ 75 nM. The solid curve is framed by two dotted lines corresponding to EC 50 ϭ 65 and 85 nM (left and right panels, respectively) to account for the small variability in EC 50 estimated for the various constructs (indicated by circles). The curve is plotted against a normalized range of 0 to 1 to account for the binding activity estimated as a percentile in Fig. 6. Oscillatory activity associated with overexpressed PLC K49A,K53A,R57A is indicated by the solid circle (top right). This point lies below the theoretical binding activity versus frequency curve (dashed line).
ysis of two groups of mouse eggs microinjected with 0.5 g/l cRNA encoding either PLC WT -LUC or the PLC DMM -LUC mutant. Expression was followed for ϳ3 h and then the two groups of eggs were analyzed by SDS-PAGE and immunoblot detection using an anti-luciferase antibody. A single protein band was observed with mobility corresponding to the predicted molecular mass (ϳ129 kDa) for both PLC WT -LUC and PLC DMM -LUC fusion proteins (Fig. 9), suggesting that each of the two cRNAs was faithfully expressed as full-length PLC-lu-ciferase proteins and at similar expression levels in the cRNAinjected mouse eggs.
Expression, Enzymatic Characterization, and in Vitro Binding of PLC Double Motif Mutant to PIP 2 -containing Liposomes-PLC DMM was then subcloned into the pETMM60 vector and bacterially expressed and purified as a NusA-His 6tagged fusion protein. Fig. 10A shows NusA-His 6 -PLC DMM recombinant protein analyzed by SDS-PAGE (left panel) and immunoblot detection with the anti-NusA monoclonal anti- body (right panel). The corresponding protein with the appropriate molecular mass (ϳ134 kDa) was observed as the top band in both Coomassie Brilliant Blue staining and on the immunoblot (Fig. 10A). Some low molecular weight bands were also detected by the anti-NusA antibody, and these are probably the result of protein degradation occurring through the bacterial expression and purification processes. Enzymatic analysis using the [ 3 H]PIP 2 hydrolysis assay showed that PLC DMM retained ϳ80% of the enzymatic activity of PLC WT (434 Ϯ 28 versus 544 Ϯ 23 nmol/min/mg) (Fig. 10B) and that there was no significant difference in the Ca 2ϩ sensitivity of PIP 2 hydrolysis for PLC WT and PLC DMM , with a very similar EC 50 value (72 versus 108 nM) ( Fig. 10C and Table 2). However, the K m value for PLC DMM (4975 M) was ϳ59-fold higher compared with PLC WT (84 M). More interestingly, when we performed the liposome/activity binding assay for PLC DMM , we found that this mutant displayed only ϳ15% relative liposome binding compared with PLC WT (Fig. 10D). These data indicate that neutralization of the positively charged residues within the first EF-hand and the XY-linker region dramatically reduces the binding of PLC to PIP 2 , leading to complete loss of its in vivo Ca 2ϩ oscillation inducing activity.

Discussion
A significant body of scientific and clinical evidence suggests that the sperm-specific PLC protein is the physiological molecule that, following sperm-egg fusion, stimulates cytoplasmic Ca 2ϩ oscillations, egg activation, and early embryonic development to effect mammalian fertilization (3,5,7,8,11,21,27). The most compelling observation is that solely introducing PLC mimics all of the signaling processes initiated by the sperm, triggering the same pattern of Ca 2ϩ release as seen at normal fertilization and leading to the successful development of a blastocyst embryo. Thus, the current model of egg activa-tion at fertilization is that the PLC of a fertilizing spermatozoon is introduced into the egg cytoplasm where it catalyzes PIP 2 hydrolysis, stimulating the IP 3 signaling pathway, and leading to Ca 2ϩ oscillations (5,13).
The sperm PLC is the smallest, with the most elementary domain organization, of all the mammalian PLC isoforms (3). Hence, the intrinsic ability of sperm PLC to cause robust Ca 2ϩ oscillations in eggs is significant because all the other PI-specific PLCs are unable trigger Ca 2ϩ oscillations in eggs at physiological protein expression levels. It therefore appears most plausible that PLC employs a novel mechanism to potently induce Ca 2ϩ release in eggs and each of its individual domains appears to play an important role in the distinct molecular and biochemical characteristics, as well as in the unique regulatory mechanism of this sperm-derived PLC isozyme (2,12). PLC shares the greatest homology with PLC␦1, but one major structural difference that distinguishes PLC from PLC␦1 is the lack of an N-terminal PH domain (2,13). This is mechanistically interesting because the PH domain of PLC␦1 in particular is known to specifically bind PIP 2 in the plasma membrane (15,28). In contrast, we have recently shown that PLC does not localize to the plasma membrane-bound PIP 2 , but instead it targets distinct vesicular structures inside the egg cortex (29). Interestingly, the chimeric addition of a PH domain at the N terminus of the PLC sequence does not alter the ability of PLC to trigger Ca 2ϩ oscillations in mouse eggs, and the PH-PLC chimera is unable to target PLC to the plasma membrane PIP 2 (25). The precise mechanism employed by PLC to enable interaction with the PIP 2 -containing vesicular membranes inside the egg cytosol is not understood.
Although the precise identity of the intracellular PIP 2 -containing vesicles is currently unknown, we have proposed that PLC associates with vesicular PIP 2 via electrostatic interactions mediated by the positively charged XY-linker region, assisting in anchoring PLC to membranes, while enhancing local concentrations of the negatively charged PIP 2 (16,17). In PLC, the XY-linker region is more extended compared with that of PLC␦1, and the proximal part to the Y catalytic domain contains a distinctive cluster of basic amino acid residues not found in the homologous region of any of the other somatic PLC isoforms (3). It is also notable that the XY-linker of somatic PLCs confers potent inhibition of their enzymatic activity (30,31). In contrast, the XY-linker of PLC does not confer enzymatic autoinhibition but conversely appears to be required for maximal enzymatic activity (19). We have recently shown that deletion of PLC XY-linker significantly diminishes its in vivo Ca 2ϩ oscillation inducing activity but does not completely abolish it (19). This suggests that the XY-linker is essential for the association of PLC with PIP 2 -containing vesicular membranes, but it is not the sole region of PLC responsible for this association.
Another candidate region that might be involved in the sequestration of PLC to membranes containing its substrate PIP 2 is the C2 domain. The current data indicate that the C2 domain of PLC may interact, albeit with low affinity, with membrane phospholipids (17,32). Indeed, such interactions were observed in vitro with phosphatidylinositol 3-phosphate and phosphatidylinositol 5-phosphate. It is possible that the association of the C2 domain with phosphatidylinositol 3-phosphate may FIGURE 9. Confirmation of expression of PLC WT -and PLC DMM -LUC fusion proteins in mouse eggs. Two sets of mouse eggs (50 eggs each) were microinjected with 0.5 g/l cRNA corresponding to either PLC WT or PLC DMM . Expression was allowed for ϳ3 h, and then the two sets of eggs were analyzed by SDS-PAGE and Western blotting using an anti-firefly luciferase antibody (1:10,000; Pierce). play a role in PLC localization, or even perhaps regulation of enzymatic activity, as the presence of phosphatidylinositol 3-phosphate reduced PIP 2 hydrolysis by PLC in vitro (32).
A recent study demonstrated that the N-terminal lobe of the EF-hand domain of PLC␦1 binds to anionic phospholipid-containing vesicles, suggesting that the EF-hand domain aids substrate binding in the active site when the protein is membraneanchored (20). The binding of the PLC␦1 EF-hand domain to anionic phospholipid is mediated by a number of cationic residues within the first EF-hand motif of PLC␦1. Interestingly, the positively charged residues that have been shown to contribute to the binding of PLC␦1 (Arg-182, Lys-183, and Arg-186) by vesicles containing anionic lipids are specifically conserved in PLC. We have shown that PLC EF-hand domains play an important role in the high Ca 2ϩ sensitivity relative to the other PLC isoforms, especially in comparison with PLC␦1 (21). PLC appears to be 100-fold more sensitive to Ca 2ϩ than PLC␦1, which would enable the enzyme to be active at the resting nanomolar Ca 2ϩ levels within the egg cytosol (21). Deletion of one or both pairs of EF-hand domains of PLC completely abolishes its Ca 2ϩ oscillation inducing activity in mouse eggs (21). Our current data suggest that this might be the result of both altered Ca 2ϩ sensitivity and loss of ability to associate with PIP 2 -containing membranes, as these PLC EF-hand deletion constructs were unable to trigger Ca 2ϩ release even when overexpressed in mouse eggs (21). Our mutagenesis analysis indicates that the substitution of even one Lys or Arg residue to Ala within the positively charged cluster of the PLC EF-hand domain diminishes the Ca 2ϩ oscillation inducing activity of PLC (Fig. 2) without affecting its ability to hydrolyze PIP 2 in vitro or the Ca 2ϩ sensitivity of its enzymatic activity (Fig. 4). Interestingly, the K m value for the triple mutant PLC K49A,K53A,R57A (432 M) was ϳ5.1-fold higher compared with PLC WT (84 M), suggesting that replacement of these three positively charged residues within the first EF-hand domain has an effect on the in vitro binding ability of PLC to PIP 2 (Table 2). Moreover, we used a variety of approaches and demonstrated that PLC binds only to PIP 2 -containing liposomes, and sequential neutralization of these basic residues within the first EF-hand region of PLC can significantly diminish the PIP 2 -binding ability of PLC (Figs. 5 and 6). As shown in our proposed mechanism in Fig. 11, which is supported by our studies on the PLC K49A,K53A,R57A,K374A,K375A,K377A mutant (PLC DMM ), it is plausible that PLC is attracted to the anionic PIP 2 -containing  component of the intracellular vesicular membranes through electrostatic interactions with both the first EF-hand domain and the XY-linker regions, which are rich in basic residues.
Our study provides an important advance in understanding the complex regulatory mechanism of PLC and suggests that the N-terminal lobe of the EF-hand domain of PLC has an essential role in the interaction of this enzyme with its target membrane, which together with the XY-linker may combine to provide a tether that facilitates proper PIP 2 substrate access and binding in the PLC active site.