The Role of EF-hand Domains and C2 Domain in Regulation of Enzymatic Activity of Phospholipase Cζ*

Sperm-specific phospholipase C-ζ (PLCζ) induces Ca2+ oscillations and egg activation when injected into mouse eggs. PLCζ has such a high Ca2+ sensitivity of PLC activity that the enzyme can be active in resting cells at ∼100 nm Ca2+, suitable for a putative sperm factor to be introduced into the egg at fertilization (Kouchi, Z., Fukami, K., Shikano, T., Oda, S., Nakamura, Y., Takenawa, T., and Miyazaki, S. (2004) J. Biol. Chem. 279, 10408–10412). In the present structure-function analysis, deletion of EF1 and EF2 of the N-terminal four EF-hand domains caused marked reduction of phosphatidylinositol 4,5-bisphosphate (PI(4,5)P2)-hydrolyzing activity in vitro and loss of Ca2+ oscillation-inducing activity in mouse eggs after injection of RNA encoding the mutant. However, deletion of EF1 and EF2 or mutation of EF1 or EF2 at the x and z positions of the putative Ca2+-binding loop little affected the Ca2+ sensitivity of the PLC activity, whereas deletion of EF1 to EF3 caused 12-fold elevation of the EC50 of Ca2+ concentration. Thus, EF1 and EF2 are important for the PLCζ activity, and EF3 is responsible for its high Ca2+ sensitivity. Deletion of four EF-hand domains or the C-terminal C2 domain caused complete loss of PLC activity, indicating that both regions are prerequisites for PLCζ activity. Screening of interactions between the C2 domain and phosphoinositides revealed that C2 has substantial affinity to PI(3)P and, to the lesser extent, to PI(5)P but not to PI(4,5)P2 or acidic phospholipids. PI(3)P and PI(5)P reduced PLCζ activity in vitro, suggesting that the interaction could play a role for negative regulation of PLCζ.

PLC 1 is a novel type of PLC (the enzyme that hydrolyzes membrane PI(4,5)P 2 into IP 3 and diacylglycerol). PLC is strongly noted in developmental biology because it is specifically expressed in mammalian sperm (1) and is capable of inducing repetitive increase in [Ca 2ϩ ] i and subsequent early embryonic development when expressed in mouse eggs by injection of RNA encoding PLC (1)(2)(3). At fertilization of mammals, evidence indicates that a cytosolic sperm factor is introduced into the ooplasm upon sperm-egg fusion and induces Ca 2ϩ oscillations (4,5), which are due to Ca 2ϩ release from the endoplasmic reticulum mainly through the IP 3 receptor (6) and are a pivotal signal for egg activation (5). Ca 2ϩ oscillations similar to those at fertilization are produced by PLC expressed in a mouse egg at an estimated level comparable with the content in a single mouse sperm (1,2). Injection of recombinant PLC into mouse eggs induces Ca 2ϩ oscillations as well (7). The Ca 2ϩ oscillation-inducing activity of sperm extract (4,8) is lost when the extract is pretreated with an antibody against PLC (1). Thus, PLC is a strong candidate for the sperm-derived Ca 2ϩ oscillation-inducing protein.
PLC possesses biochemically interesting characteristics. It is composed of four EF-hand domains in the N terminus, catalytic X and Y domains, and a C2 domain in the C terminus (1) (Fig. 1A) common to other types of PLC, but PLC lacks the N-terminal PH domain found in PLC␤, -␥, -␦, and -⑀. PH domain of PLC␤, -␥, or -␦ binds to PI(3)P, PI(3,4,5)P 3 , or PI(4,5)P 2 , respectively (9 -11). The PH and C2 domains of PLC␤ also interact with the G protein subunits G␤␥ and G␣ q , respectively, for activation of PLC␤ (12,13). PLC␥ is phosphorylated at the tyrosine residues 771, 783, and 1254 in response to growth factor stimulation leading to activation of PLC␥ (9 -11). As for PLC, which lacks the PH domain, there is no prediction about the domain responsible for binding to phosphoinositides and for enzymatic activation.
PLC has domain features similar to those of PLC␦ except for the PH domain (Fig. 1A). PLC␦ is suggested to be regulated by Ca 2ϩ because PLC␦1 has Ca 2ϩ -binding sites in not only the EF-hand domains but also in the X and Y domains and the C2 domain (11). We have recently shown (7) that the PI(4,5)P 2hydrolyzing activity of recombinant PLC is ϳ100-fold more sensitive to Ca 2ϩ than that of PLC␦1 and is 70% maximal at 100 nM [Ca 2ϩ ], which is usually the basal [Ca 2ϩ ] i of cells. Recombinant PLC has ϳ20-fold higher Ca 2ϩ oscillation-inducing activity than PLC␦1 (7). Thus, PLC possesses remarkable properties that are favorable for the sperm factor to initiate (7) and maintain (3) Ca 2ϩ oscillations via continuously produced IP 3 . Using mutational analysis in the present study, we addressed which domains of PLC are responsible for the Ca 2ϩsensitive PLC activity in vitro, Ca 2ϩ oscillation-inducing activity in mouse eggs, and phosphoinositide binding activity in vitro and the effect of phosphoinositide binding activity on the PLC activity. We focused on four N-terminal EF-hand domains because they are likely to be Ca 2ϩ -binding sites and could be critical for the high Ca 2ϩ sensitivity. We also focused on the C2 domain, as three Ca 2ϩ -binding sites are predicted in the C2 domain of PLC␦1 (9,14). The C2 domain of PLC could play another role such as interaction with phosphoinositides. Our study revealed that both EF-hand domains and the C2 domain are critical for Ca 2ϩ oscillation-inducing activity as well as the PLC activity of PLC. Furthermore, we found that the C2 domain of PLC binds to PI(3)P or PI(5)P, and this interaction may negatively regulate the PLC activity.
The FYVE domain of mouse Hrs (hepatocyte growth factor-regulated tyrosine kinase substrate) (17) was cloned from a mouse testis cDNA library and subcloned to the pGEX vector. A construct containing FYVE in duplicate (2xFYVE) was also prepared as described previously (18). Protein Expression and Purification-Recombinant PLC mutant proteins were prepared by the baculovirus/Sf9 cell expression system (7). As expression of GST-tagged PLC protein in Sf9 cells was low, His-tagged PLC was expressed. Similarly, ⌬EF1 and ⌬EF1-2, and EF1AA and EF2AA were His-tagged. Other proteins including ⌬EF1-4 were GST-tagged. Both His-tagged PLC⌬EF1-3 and GST-tagged PLC⌬EF1-3 were expressed for comparison (see Table I).
Infected Sf9 cells (0.5-1.5 liters) were lysed with a French pressure cell in a lysis buffer and centrifuged at 15,000 ϫ g for 20 min. After filtration through a 0.22-m filter, supernatant was applied to a Sepharose column (His-tagged protein to nickel-chelating Sepharose and GST-tagged protein to glutathione-Sepharose (Amersham Biosciences)). The former protein was eluted with 300 mM imidazole, and the latter protein was eluted with 10 mM glutathione in elution buffer. Purified proteins were dialyzed against an intracellular buffer (120 mM KCl and 20 mM HEPES/KOH, pH 7.5).
Measurement of PLC Activity-The PLC activity was assayed by hydrolysis of PI(4,5)P 2 (19) in a 50-l reaction mixture containing 25,000 dpm of [ 3 H]PI(4,5)P 2 (PerkinElmer Life Sciences), 40 M PI(4,5)P 2 (Sigma), and 50 M PE (Sigma) as phospholipid micelles (see Ref. 7 for details). In some experiments, one of other phospholipids was added to the micelles to examine its interference in the PI(4,5)P 2hydrolyzing activity of PLC. After incubation at 37°C for 5 min, the reaction was stopped by adding 2 l of chloroform:methanol (2:1, v/v). Radioactive IP 3 was extracted with 0.5 ml of 1 N HCl (19), and radioactivity in the upper aqueous phase was measured for 1 min in a liquid scintillation counter. [Ca 2ϩ ] of the reaction mixture was adjusted at various values between 10 Ϫ9 and 3 ϫ 10 Ϫ5 M using EGTA/CaCl 2 (20).
Liposome Binding Assay-The binding of C2D, ␦PHD, or 2xFYVE to PI, PI(3)P, PI(5)P, PI(4,5)P 2, PS, or phosphatidic acid (Cell Signals, Inc.) was examined in a micelle containing PC, PE (Sigma), and one of the above phosphoinositides or other acidic lipids at various percentages. The lipid mixture (100 g) was dried to form a film in a microtube and suspended in an intracellular buffer. Liposomes formed by sonication were incubated with C2D, ␦PHD, or 2xFYVE for 15 min at room temperature and centrifuged at 15,000 ϫ g at 4°C. Both supernatant and pellet were used for analysis by SDS-PAGE and Coomassie Blue staining.
Expression of Venus-PLC Mutants in Mouse Eggs-Venus PLC mutants integrated into pBluescript II (SKϩ) were digested with NotI, and the resulting fragments were used as templates for in vitro transcription. cDNA of Venus was a kind gift from Dr. A. Miyawaki (Brain Science Institute, RIKEN). RNA was synthesized by T7 polymerase using the T7 MessageMachine kit (Ambion), treated with phenol-chloroform, and then precipitated with ethanol. To facilitate expression in mouse eggs, RNA was added with 200 -250 bases of poly(A) (21) by incubation for 30 min at 37°C in the presence of 200 M ATP and yeast poly(A) polymerase (Amersham Biosciences). Dried RNA was resolved in 150 mM KCl solution and adjusted to 50 ng/l prior to injection into mouse eggs. To avoid folding of RNA, the diluted RNA was heated at 85°C for 3 min and cooled on ice. RNA was injected into ϳ10 eggs through a glass micropipette (injected amount, ϳ5 pl/egg). The estimated concentration of RNA in the egg was ϳ1.2 ng/l, assuming the egg volume was ϳ200 pl.
Preparation of Eggs and [Ca 2ϩ ] i Measurement-Mature mouse eggs were obtained from the oviducts of B6D2F1 female mice superovulated by intraperitoneal injection of gonadotropins (8). They were loaded with the Ca 2ϩ -sensitive dye fura-2 (5 M; dextran, Molecular Probes Inc.) in M2 medium for 8 min at 37°C and then transferred to a plastic dish mounted on an inverted fluorescence microscope and heated at 30 -32°C. Fluorescence intensity (F) of fura-2 to 340 and 380 nm excitation lights was measured by a conventional Ca 2ϩ imaging method using an Measurement of Venus Fluorescence-Venus, which was ligated to PLC and expressed in the egg, was excited by light passed through a 470 -490-nm bandpass filter and a ϫ40 objective lens. Emitted fluorescence was passed through the objective lens, a 510-nm dichroic mirror (DM510, Nikon), and a 520 -560-nm bandpass filter (see Ref. 2 for details). The relation between F and the injected amount of recombinant Venus alone in the egg was obtained for calibration of expressed PLC. Table I presents the specific PI(4,5)P 2hydrolyzing activity of recombinant PLC protein and its mutants measured in vitro at [Ca 2ϩ ] between 1 and 30 M, at which a maximal activity was obtained. PLC⌬EF1-4, which lacks all of the four EF-hand domains, had no PLC activity at 1 and 10 M [Ca 2ϩ ] (Table I and Fig. 2A). s-PLC (PLC⌬EF1-3), a PLC variant that is expressed in the mouse testis and lacks the EF1, EF2, and EF3 domains, had substantial activity (20 nmol/min/mg at 10 M [Ca 2ϩ ]). Deletion of the C2 domain (PLC⌬C2) or the 37 amino acid residues in the C-terminal region next to the C2 domain (PLC⌬611-647) resulted in complete loss of PLC activity (less than 0.1 nmol/min/mg) at 1 and 10 M [Ca 2ϩ ] (Table I). GST-tagged PLC⌬EF1-3 serves as a positive control for these inactive mutants that were tagged with GST. These results indicate that whole N-terminal EFhand domains are critical for PLC activity and that the C2 domain including the C-terminal region is also prerequisite for PLC activity.

EF-hand Domains and C2 Domain Are Necessary for Catalytic Activity of PLC-
To examine the role of the residual region except for the EF-hand domains, a chimera was formed by replacing the region from the X domain to the C terminus of PLC with that of PLC␦1 (PLCEF/␦1, Fig. 1A). This chimera showed no detectable PLC activity at 1 and 10 M [Ca 2ϩ ] (Table I) even if four EF-hand domains are normally prepared. Therefore, the enzymatic activity of PLC seems to require a specific matching in the PLC molecule between the four N-terminal EF-hand domains and the other region containing the catalytic X and Y domains and C-terminal C2 domain.
Effects of Deletion of EF-hand Domains on PLC Activity and Its Ca 2ϩ Sensitivity-The PLC activity of PLC was remarkably reduced when the EF-hand domains were deleted one by one from the N terminus (Table I) To analyze the Ca 2ϩ sensitivity, the PLC activity was assayed at [Ca 2ϩ ] between 10 Ϫ9 and 3 ϫ 10 Ϫ5 M and presented as the percentage relative to the maximal activity ( Fig. 2A). As reported previously (7), the PLC activity of PLC was significantly recognized at [Ca 2ϩ ] as low as 10 nM and reached a maximum at 1 M [Ca 2ϩ ]. There was ϳ80% maximal activity at 100 nM, which is the resting [Ca 2ϩ ] i level in mouse eggs (8) as well as somatic cells (22). The [Ca 2ϩ ] for giving a half-maximal PLC activity, EC 50 , was obtained by fitting a curve to the data using the Hill equation. The EC 50 of PLC was calculated as 32 nM. Deletion of EF1 and EF2 did not cause a marked decrease in the Ca 2ϩ sensitivity ( Fig. 2A). The EC 50 was 98 nM for PLC⌬EF1 and 93 nM for PLC⌬EF1-2 (Table I). In contrast, the PLC activity of PLC⌬EF1-3 was about 15% maximal at 100 nM [Ca 2ϩ ] and reached the maximum at 30 M [Ca 2ϩ ] (Fig.  2A). The EC 50 was 373 nM, 12-fold higher than that of PLC (Table I). These results indicate that the region containing EF3 contributes to the high Ca 2ϩ dependence of the PLC activity. However, the EC 50 of PLC⌬EF1-3 is still 15-fold lower than that of PLC␦1 (5.7 M) (7). The Hill constant obtained from curves in Fig. 2A was around 1.0 for EF-hand domain deletion mutants as well as for PLC. There may be a coordinating structural determinant(s) other than EF1-EF3 for the highly Ca 2ϩ -sensitive enzymatic activity.
Effects of Point Mutation in EF1 and EF2 on PLC Activity and Its Ca 2ϩ Sensitivity-According to the ProDom EF-hand pattern (23), EF1 and EF2 of PLC as well as those of PLC␦1 (24) contain the Ca 2ϩ -binding loop sequence homologous to a well conserved Ca 2ϩ -binding site of calmodulin or troponin C (25, 26) (Fig. 1C). Of the Ca 2ϩ -coordinating residues at the x, y, z, and Ϫy, Ϫx, and Ϫz positions of troponin C, the oxygencontaining side chains of amino acids at x, z, and Ϫz play a critical role in Ca 2ϩ binding (25,26). To analyze the function of EF1 or EF2, two point mutations to alanine were introduced at x and z (Fig. 1C), as they have been shown to impair Ca 2ϩ binding to canonical EF-hand proteins (25).
The specific activity of PLCEF1AA and PLCEF2AA was 650 and 590 nmol/min/mg at 1 M [Ca 2ϩ ], respectively (Table  I); the point mutation reduced the PLC activity to about half. The EC 50 was 66 nM for EF1AA and 64 nM for EF2AA. The values ranged between the EC 50 of PLC and those of PLC⌬EF1 and PLC⌬EF1-2 (Table I). The Ca 2ϩ dependence of the PLC activity of these mutants was thought to be approximately similar to that of PLC (Fig. 2B). Ca 2ϩ binding, if any, to the putative Ca 2ϩ binding loop in EF1 and EF2 does not play a significant role in the regulation of the Ca 2ϩ sensitivity of PLC, consistent with the deletion mutant study of EF1 and EF2.
EF-hand Domains and C2 Domain Are Necessary for Ca 2ϩ Oscillation-inducing Activity-To analyze whether the enzymatic property of PLC mutants is correlated to the Ca 2ϩ oscillation-inducing activity, PLC mutant proteins were expressed in mouse eggs by injection of 50 ng/l RNA encoding the respective proteins as described previously (1,2). Fusion of Venus to the N or C terminus of PLC did not affect the Ca 2ϩ oscillation-inducing activity (2). Changes in [Ca 2ϩ ] i were recorded for 3 h, and fluorescence intensity (F) of the Venus-PLC mutant was measured at 3 h after RNA injection to estimate the expressed amount of the protein. Venus-PLC induced fertilization-like Ca 2ϩ oscillations starting from 25-30 min after RNA injection (Fig. 3A). PLCEF1AA (Fig. 3B) and PLCEF2AA also produced repetitive Ca 2ϩ spikes in a similar pattern (Table II) and induced egg activation as indicated by the formation of the second polar body and the (female) pronucleus (not shown). In contrast, none of the PLCEF deletion mutants caused any detectable [Ca 2ϩ ] i rise for 3 h after RNA injection ( Fig. 3C and Table II). Both PLC⌬611-647 and PLC⌬C2 (Fig. 3D) failed to induce any Ca 2ϩ spike (Table II). The expression level of these five inactive mutants was higher than that of PLC at 3 h after RNA injection (Table II). However, as shown in Fig. 3E, there was a large (ϳ40-fold) difference between F of Venus-PLC at 25-30 min (the time when the first Ca 2ϩ spike appeared, arrow 1) and F of Venus-s-PLC at 3 h (up to the time Ca 2ϩ spike had never occurred, arrow 2). The results characterize an all-or-none mode of the Ca 2ϩ oscillation-inducing activity of these mutants in this series of assays. The PLC mutants showing the specific PLC activity lower than one-quarter of the activity of PLC (Table I) are incapable of inducing Ca 2ϩ oscillations (Table II).
Binding of C2 Domain to Phosphoinositides-The C2 domain of PLC␦1 (14), cytosolic phosholipase A 2 (27), or protein kinase C (28) has been shown to interact with phospholipids and induce the selective translocation and activation of these enzymes. PLC is devoid of the PH domain that is found in other PLC isozymes and is known to interact with phosphoinositides. Therefore, interaction of the C2 domain of PLC with phosphoinositides was examined by preparing the C2 domain with the neighboring C-terminal region added and with GST in the N terminus (C2D). The protein-lipid overlay assay showed that C2D bound clearly to PI(3)P and weakly to PI(5)P but not to PI (4)P, PI(3,4)P 2 , PI(3,5)P 2 , PI(4,5)P 2 , and PI(3,4,5)P 3 (Fig.  4A). 2xFYVE (18) and ␦1PHD (16,29) have been used for specific probes for PI(3)P and PI(4,5)P 2 , respectively. In this assay, 2xFYVE and PLC␦1PHD specifically bound to PI(3)P and PI(4,5)P 2 , respectively (Fig. 4A).
The binding ability of C2D to phosphoinositides and other phospholipids was also assayed using liposomes containing the respective lipid at increasing percentages (Fig. 4B). C2D-PI(3)P binding was detected in the pellet of liposomes containing 10 or 20% PI(3)P (Fig. 4B, 1), although higher levels of PI(3)P content in liposomes were required than those for 2xFYVE (Fig. 4B, 7). On the other hand, binding of C2D to PI(4,5)P 2 was not detected (Fig. 4B, 6), although it was clearly  Ϫ 62 0/6 detected for ␦1PHD even at a much lower content of PI(4,5)P (Fig. 4B, 8) but not seen for ␦1PHDK30N/K32N (replacement of both Lys 30 and Lys 32 with Asn) (Fig. 4B, 9). The liposome assay failed to show the binding of C2D to PI(5)P (Fig. 4B, 2). Binding to PI (Fig. 4B, 3), PS (Fig. 4B, 4), or phosphatidic acid (Fig. 4B, 5), negatively charged phospholipids like PI(3)P or PI(5)P, was not detected. Thus, C2D has substantial affinity to PI(3)P and slight affinity to PI(5)P but not to other phosphoinositides or acidic lipids. Effect of Phosphoinositides on PI(4,5)P 2 -hydrolyzing Activity of PLC in Vitro-The effects of PI(3)P and PI(5)P, which bind to the PLC C2 domain, on the PI(4,5)P 2 -hydrolyzing activity of PLC were examined. The activity of PLC was reduced to 0 or 13% in the co-existence of micelles containing 40 M PI(4,5)P 2 as the substrate and another micelle containing 200 M PI(3)P or PI(5)P at 1 M [Ca 2ϩ ] (Fig. 5A). The activity was not affected by PC, PE, or PS (Fig. 5A). When assayed in the same micelles containing both PI(4,5)P 2 and PI(3)P or PI(5)P at the ratio of 1:1, the activity of PLC was reduced to 40 or 50%, respectively (Fig. 5B). The value was 20 or 30% when the ratio was 1:5. No reduction of the PLC activity occurred in the presence of 5-fold higher PC or PS. These results suggest that PLC as well as C2D has a specific affinity to PI(3)P or PI(5)P and binds preferentially to PI(3)P or PI(5)P possibly via the C2 domain, resulting in reduced accession of PLC to its substrate. The magnitude of this effect was not markedly different at 100 nM, 1 M, and 10 M [Ca 2ϩ ] (Fig. 5C), suggesting that the suppressing effect of the PI(3)P-or PI(5)P-PLC interaction on the PI(4,5)P 2 -hydrolyzing activity operates independently of the Ca 2ϩ -regulated activation process of PLC. DISCUSSION The present study demonstrated the significant role of the EF-hand and C2 domains of PLC in the highly Ca 2ϩ -sensitive PI(4,5)P 2 -hydrolyzing activity in vitro and Ca 2ϩ oscillationinducing activity in the mouse egg. Our study also demonstrated that the C2 domain of PLC interacts with PI(3)P or PI(5)P, and this association may cause an inhibitory regulation of PLC activity.
The Role of EF-hand Domains of PLC in Regulation of Its Ca 2ϩ Sensitivity-Deletion of EF1 and EF2 remarkably reduced the specific activity of PLC, indicating that the two N-terminal EF-hand domains are obligatory for enzymatically active conformation. PLC contains the putative Ca 2ϩ -binding loop in EF1 and EF2 in which an Asp residue at the x position and a Gly between the z and Ϫy positions are conserved. Point mutation of EF1 or EF2 at the x and z positions, which are critical for Ca 2ϩ binding in EF-hand proteins such as troponin C (25,26), reduced the PLC activity to half, suggesting that Ca 2ϩ binding to EF1 and EF2 might play a substantial role in the PLC activity. However, the point mutation in EF1 or EF2 did not cause a marked decrease in the Ca 2ϩ sensitivity of the PI(4,5)P 2 -hydrolyzing activity. The two N-terminal EF-hand domains, therefore, are considered to play a structural role to form the active conformation rather than a Ca 2ϩ -binding site for activation of the catalytic activity. Similarly, in PLC␦1, deletion of the N-terminal EF-hand domain markedly reduces the PLC activity, but point mutation of EF1 at the x, z, and Ϫz positions does not affect the PLC activity and the Ca 2ϩ sensitivity (24).
Deletion of the EF1, EF2, and EF3 remarkably reduced the Ca 2ϩ sensitivity of PLC, as indicated by the 12-fold higher EC 50 of [Ca 2ϩ ] (Fig. 2A). The result indicates that EF3 is an important domain necessary for the high Ca 2ϩ sensitivity. EF3 could serve as a high affinity Ca 2ϩ -binding site for activation of PLC. However, the putative Ca 2ϩ -binding loop is less conserved in EF3 than in EF1 or EF2. In addition, PLC⌬EF1-3 still has more than 10-fold higher Ca 2ϩ sensitivity than that of PLC␦1 (7). We performed direct measurement of Ca 2ϩ binding to PLC and its variants by a Ca 2ϩ overlay assay. The Ca 2ϩ binding activity of PLC⌬EF1-3 tended to be reduced compared with full-length PLC or PLC⌬EF1-2 (data not shown). However, the difference was not statistically significant because of fluctuation in the obtained values. The Ca 2ϩ binding level to the EF-hand region of PLC and mutants was much lower than that of calmodulin. Taken together, it seems that the high Ca 2ϩ sensitivity of PLC may be derived from the highly coordinated structure of the EF-hand region rather than the primary sequence in the Ca 2ϩ binding loop. It is considered that the PLC␦1 is folded at the X-Y region in such a way that the N-terminal domains and C-terminal structure are closely apposed and form the catalytic core composed of the catalytic domain, EF-hand domain, and C2 domain (30 -32). Replacement of the residual region except for the EF-hand domains with that of PLC␦1 (PLCEF/␦1) could not conserve the PLC activity. A close apposition of the N-terminal EF-hand domains and a specific C-terminal structure might determine the enzymatic activity.
PLCEF1AA and PLCEF2AA were capable of inducing Ca 2ϩ oscillations in the mouse egg at about 30 min after RNA injection when expression of the protein was still at a low level (Fig. 3). In contrast, PLC⌬EF1, PLC⌬EF1-2, and PLC⌬EF1-3 were incapable of inducing any Ca 2ϩ spike within 3 h even if expression of the protein reached a much higher level than the level of wild type PLC critical for initiation of Ca 2ϩ oscillations. In our experimental condition, PLC mutants showed an all-or-none mode of Ca 2ϩ oscillation-inducing ability. Furthermore, all of the PLC mutants that induced Ca 2ϩ spikes caused formation of a second polar body and pronucleus. Thus, the N-terminal half of the EF-hand structure is critical for the physiological function of PLC. The Ca 2ϩ oscillation-inducing ability was lost by deletion of EF1 and EF2 but not by point mutation in the putative Ca 2ϩ -binding loop of EF1 and EF2. The ability may be correlated to the specific activity and overall structure of the mutant.
The Role of C2 Domain of PLC-The C2 domain, including 37 external amino acids in the C terminus of PLC, is necessary for the catalytic activity in vitro and the Ca 2ϩ oscillationinducing activity in the egg. This region could be essential for the positioning of the enzyme in an active conformation or for the membrane targeting of the enzyme. The C2 domain is known to play a significant role in the Ca 2ϩ -dependent subcellular membrane targeting of several lipid-metabolizing enzymes such as PLC␦1 or cytosolic phospholipase A 2 (27, 30 -33). In our previous study (2), Venus-PLC␦1 expressed in the mouse egg was located on the surface in association with the plasma membrane, but this was not detected for Venus-PLC. Therefore, targeting of PLC to phospholipid is unknown. Screening of C2 domain-interacting phosphoinositides revealed that the C2 domain binds to PI(3)P and, to a lesser extent, to PI(5)P. To our knowledge, this is the first example of the binding of the C2 domain to PI(3)P, although the C2 domain of synaptotagmin or JFC1 has been shown to interact mainly with PI(3,4,5)P 3 , and to a lesser extent, with other 3Ј-phosphoinositides (34,35).
The presence of PI(3)P or PI(5)P remarkably reduced the PLC-mediated hydrolysis of PI(4,5)P 2 in micelles. It is likely that the accession of PLC to its substrate is perturbed simply because PLC binds preferentially to PI(3)P or PI(5)P. Alternatively, the association of the C2 domain with PI(3)P or PI(5)P could interfere with the suitable positioning of PLC on the lipid membrane to hydrolyze the substrate. This perturbation of the PLC activity was Ca 2ϩ -independent, suggesting that the binding of PI(3)P or PI(5)P to PLC occurs irrespective of changes in [Ca 2ϩ ]. The physiological significance of the suppressing effect of PI(3)P or PI(5)P on PLC remains to be elucidated.
Mammalian PLCs are activated by agonist-induced anchoring to the membrane (11). The PH and C2 domains of PLC␤ bind to G␤␥ and G␣ q in the plasma membrane, respectively, upon stimulation of the G protein-coupled receptor (12,13). The PH domain of PLC␥ binds to PI(3,4,5)P 3 with a high affinity upon agonist stimulation (36,37). PLC⑀ is a Ras effector and localized to the plasma membrane or the perinuclear region upon growth factor stimulation when coexpressed with activated Ha-Ras or Rap1 mutant (38 -40). PLC␦1 also associates tightly with PI(4,5)P 2 at the PH domain and is recruited to the plasma membrane (41). In the case of PLC, activation of the enzyme accompanied by membrane association is not found at present. What is known is that PLC is so Ca 2ϩ -sensitve in the PLC activity that it could be active at the resting state of the cell and cause repetitive [Ca 2ϩ ] i rises continuously. This situation seems to be unfavorable for resting cells. The interaction between the C2 domain and PI(3)P or PI(5)P might play a role in the inhibitory regulation of the enzymatic activity of PLC, for example in the sperm before fertilization. Further studies on other targets of the C2 domain of PLC together with the Ca 2ϩ dependence of the binding are required to reveal the physiological significance of the C2 domain.