The Role of X/Y Linker Region and N-terminal EF-hand Domain in Nuclear Translocation and Ca2+ Oscillation-inducing Activities of Phospholipase Cζ, a Mammalian Egg-activating Factor*

Sperm-specific phospholipase C-zeta (PLCζ) causes intracellular Ca2+ oscillations and thereby egg activation and is accumulated into the formed pronucleus (PN) when expressed in mouse eggs by injection of cRNA encoding PLCζ, which consists of four EF-hand domains (EF1-EF4) in the N terminus, X and Y catalytic domains, and C-terminal C2 domain. Those activities were analyzed by expressing PLCζ mutants tagged with fluorescent protein Venus by injection of cRNA into unfertilized eggs or 1-cell embryos after fertilization. Nuclear localization signal (NLS) existed at 374–381 in the X/Y linker region. Nuclear translocation was lost by replacement of Arg376, Lys377, Arg378, Lys379, or Lys381 with glutamate, whereas Ca2+ oscillations were conserved. Nuclear targeting was also absent for point mutation of Lys299 and/or Lys301 in the C terminus of X domain, or Trp13, Phe14, or Val18 in the N terminus of EF1. Ca2+ oscillation-inducing activity was lost by the former mutation and was remarkably inhibited by the latter. A short sequence 374–383 fused with Venus showed active translocation into the nucleus of COS-7 cells, but 296–309 or 1–19 did not. Despite the presence of these special regions, both activities were deprived by deletion of not only EF1 but also EF2–4 or C2 domain. Thus, PLCζ is driven into the nucleus primarily by the aid of NLS and putative regulatory sites, but coordinated three-dimensional structure, possibly formed by a folding in the X/Y linker and close EF/C2 contact as in PLCδ1, seems to be required not only for enzymatic activity but also for nuclear translocation ability.

PLC 2 is a novel isozyme of PLC (the enzyme that hydrolyzes membrane PIP 2 into IP 3 and diacylglycerol) and a strong can-didate of the mammalian sperm-derived egg-activating factor (1). PLC is specifically expressed in the sperm (2) and induces repetitive increase in [Ca 2ϩ ] i called Ca 2ϩ oscillations and subsequent early embryonic development when expressed in mouse eggs by injection of RNA encoding PLC (2,3). In mammalian fertilization, accumulated evidence indicates that a cytosolic sperm factor is driven into the ooplasm upon sperm egg fusion and induces Ca 2ϩ oscillations (4,5), which are caused by Ca 2ϩ release from the endoplasmic reticulum mainly through type 1 IP 3 receptor (6) and are a pivotal signal for egg activation characterized by resumption of the second meiosis and formation of PN (5). PLC is a strong candidate of the sperm factor, because 1) fertilization-like Ca 2ϩ oscillations are produced by PLC expressed in a mouse egg at an estimated level comparable to the content in single mouse sperm (2,3). 2) Injection of recombinant PLC into mouse eggs induces Ca 2ϩ oscillations as well (7). 3) Ca 2ϩ oscillation-inducing ability of sperm extract injected into eggs (4,8) is lost when pretreated with an antibody against PLC (2). 4) PLC content in the mouse sperm and the number of Ca 2ϩ spikes at fertilization are reduced by transgenic RNA interference of PLC (9). 5) PLC has such a high Ca 2ϩ sensitivity of PIP 2 -hydrolyzing activity that the enzyme can be active in the resting cells at ϳ100 nM Ca 2ϩ (7,10), suitable for the sperm factor as the first stimulus in the egg cytoplasm at fertilization.
Another important property of PLC is nuclear translocation ability. PLC expressed by RNA injection is accumulated into the formed PN (3,11,12). This is consistent with earlier observation that sperm-derived Ca 2ϩ oscillation-inducing activity is concentrated into PN formed several hours after fertilization, as examined by transfer of the ooplasm or PN into unfertilized eggs (13). Ca 2ϩ oscillations cease at about the time of PN formation (14), but continue without stopping when PN formation was prevented by injection of a lectin, WGA (15). Therefore, it is thought that translocation of the sperm factor or PLC into PN plays a key role in cessation of Ca 2ϩ oscillations at the interphase of a cell cycle (11,13,15).
Structure-function analysis of PLC is implicated, because PLC is a biologically important factor and can be practically utilized for artificial egg activation. PLC is composed of four EF-hand domains in the N terminus, X and Y catalytic domains, and C2 domain in the C terminus (2), common to other isozymes of PLC (16), but lacks N-terminal PH domain found in PLC␤, ␥, ␦, and ⑀ (2). A short form of PLC, s-PLC, is thought to be expressed in the mouse sperm, because mRNA encoding a protein, which lacks three EF-hand domains but is identical to PLC in other region, has been found to exist in the mouse testis (AK006672 in EMBL). Both 74-and 65-kDa protein bands are detected by Western blot-ting of mouse sperm extract using anti-PLC antibody (17). We have found that s-PLC expressed in mouse eggs has much less Ca 2ϩ oscillation-inducing activity and is hardly accumulated in PN (3), suggesting that EF-hand domains are responsible for these important properties. Actually, deletion of N-terminal EF-hand domains results in the loss of Ca 2ϩ oscillation-inducing activity (10,18).
In the present study, we addressed the molecular structure responsible for the nuclear translocation ability in a quantitative manner and in parallel with precise assay of Ca 2ϩ oscillation-inducing activity, focusing on the putative NLS region and EF-hand domain region. PLC or its mutants fused with a fluorescent protein Venus (19) were expressed by injection of respective cRNA into mouse eggs. Nuclear translocation was investigated in PN of 1-cell embryos after artificial activation by PLC or fertilization by the sperm. Translocation was also examined in the nucleus of cultured somatic cells after transfection of cDNA.

EXPERIMENTAL PROCEDURES
Preparation of Gametes and Insemination-Mature eggs at MII were obtained from superovulated B6D2F1 mice (see Ref. 20 for details), and freed from cumulus cells by 0.05% hyaluronidase (Sigma). M2 medium was used during egg preparation, RNA injection, [Ca 2ϩ ] i measurement, and observation of eggs or embryos. Twenty to thirty eggs were transferred to a 400-l drop of M2 medium covered with paraffin oil in a glass-bottomed plastic dish, which was placed on the stage of an inverted fluorescence microscope (TMD, Nikon) and heated at 31-33°C. Eggs were injected with cRNA (see below). In another experiments, cRNA was injected into 1-cell embryos after IVF. M16 medium was used for IVF and incubation of fertilized eggs. Spermatozoa were collected from the cauda epididymides and incubated at 37°C (5% CO 2 in air) for 1-1.5 h for capacitation (20). A small amount of sperm suspension was added to a 200-l drop of M16 medium containing M II eggs attached with cumulus cells. The eggs and spermatozoa were incubated for ϳ5 h until the male and female PN were recognized. Onecell embryos were transferred to a 400-l drop of M2 medium,

Nuclear Translocation of PLC
treated with 0.05% hyaluronidase, and after washing, injected with cRNA.
Construction of Plasmids-cDNA encoding full-length PLC (647 amino acid residues; see Fig. 1A) (GenBank TM accession number AF435950) or s-PLC lacking 110 amino acid residues from the N terminus (AK006672) was prepared using PCR techniques, fused with Venus (19) in the C terminus, and subcloned into pBluescript II SK(ϩ). The methods were the same as described previously (4).
Point replacement of an amino acid or partial deletion of amino acid sequence in cDNA of PLC was constructed by GeneTailor TM site-directed mutagenesis system (Invitrogen), using PLC-Venus-pBluescript II-SK (ϩ) as the template (see Fig. 1A for domain features and Tables for designation). To circumvent unwanted mutations, a region surrounding the targeted amino acid(s) and presenting unique restriction sites was subcloned in the parental vector and verified by DNA sequencing using Applied Biosystems ABI PRISM 310 DNA sequencer. Truncation was performed from the N terminus to a given number of amino acids between 4th to 39th residues or to the end of EF1 (D2-39), EF2 (D2-77), EF3 (D2-110), and EF4 (D2-167), leaving Met 1 . Truncated fragments were amplified by PCR using PLC-Venus-pBluescript II-KS (ϩ) as the template. Amplified fragments were digested with KpnI and SpeI, and ligated to the KpnI and SpeI sites of the parental vector. All constructs were checked by sequence analysis.
RNA and Polyadenylation-The constructed plasmids were digested with NotI, and resulting fragments were used as templates for in vitro transcription. RNA was synthesized by T3 or T7 polymerase using mMessage mMachine Kit (Ambion). To facilitate RNA translation in the egg, RNA was added with more than 200 poly(A) in the 3Ј-tail (see Ref. 21 for details). Dried RNA was resolved in 150 mM KCl solution (final concentration, ϳ1.5 g/l). RNA was diluted to the range between 10 and 1,000 ng/l and injected into MII eggs or 1-cell embryos using a glass micropipette (injected amount, ϳ4 pl per egg or embryo of which volume is 200 pl). To make the expression level of various PLC mutants comparable, the concentration of RNA for injection was adjusted in such way that fluorescence intensity (F) of Venus in the egg was in the range between 55 and 90 (arbitrary unit) at 3 h after RNA injection. A standard concentration of PLC-Venus RNA was 50 ng/l. RNA concentration used was raised up to 1,000 ng/l, when extreme overexpression was necessary.
Measurement of Venus Fluorescence-Of 30 -40 MII eggs injected with cRNA, 4 -9 eggs were left in the same dish and subjected to continuous measurement of F. Fluorescent images of eggs were acquired every 3 min at 31-33°C, using an EB- CCD camera (C7190 -23; Hamamatsu Photonics) and an image processor (Argus 50; Hamamatsu Photonics). Excitation light was passed through a 470 -490-nm bandpass filter and a 20ϫ objective lens. Emitted light was passed through the objective lens, a 510-nm dichroic mirror (DM510; Nikon), and a 520 -560-nm bandpass filter. Autofluorescence of the egg, probably derived from oxidized flavins (22), was subtracted from total fluorescence to obtain F. Other eggs were kept in another dish and subjected to precise observation at 3, 5, and 8 h after RNA injection using a confocal laser scanning microscope (LSM310, Carl Zeiss) with excitation light of 510 nm. Differential interference contrast images were recorded simultaneously by another sensor for transmitted laser light. All these procedures were also taken in the experiment in which cRNA was injected into 1-cell embryos about 5.5 h after insemination. Judgment of Nuclear Accumulation-The ratio of F in the PN to that in the cytoplasm (F PN /F C ) at 6 h after RNA injection was taken as a parameter for nuclear accumulation. The evaluation criteria were tentatively defined, as indicated in the legend of Table 1. Values at 8 h were also presented to see the progression of nuclear accumulation.
[Ca 2ϩ ] i Measurement-Ca 2ϩ oscillations were recorded in another optical system by conventional Ca 2ϩ imaging method using an image processor. Four to five MII eggs were injected with 50 M solution of the Ca 2ϩ -sensitive fluorescent dye fura dextran (Molecular Probes Inc.) together with a cRNA and were subjected to [Ca 2ϩ ] i measurement for 9 h after RNA injection. F of fura was measured without interference with that of Venus, by applying 340-and 380-nm UV lights alternatively and by leading emission light through a 400-nm dichroic mirror (DCLP; Omega) and a 500 -520-nm bandpass filter. Fluorescence was detected by an EB-CCD camera (C7190 -23; Hamamatsu Photonics). Ca 2ϩ images were acquired at intervals of 20 s and processed to calculate F 340 /F 380 later using NIH Image (a public domain image processing software for the Macintosh computer). Formation of the PN and nuclear translocation of a PLC mutant were examined 5 and 9 h after RNA injection, respectively.

RESULTS
Ca 2ϩ Oscillation-inducing Activity and Nuclear Translocation Ability of Wild-type PLC-The domain feature associated with amino acid number of PLC is illustrated in Fig. 1A. Under the present experimental conditions, expression of PLC in MII eggs was detected by Venus-derived F from 30 min after injection of 50 ng/l RNA, increased up to 3-4 h, and attained a steady level (Fig. 2A). The magnitude of expression of PLC-Venus was compared in F at 3 h after RNA injection (Tables 1  and 2). The first Ca 2ϩ transient was generated 30 -40 min after injection of RNA of wild-type PLC (Fig. 1B). The delay time was a parameter that reflects Ca 2ϩ oscillation-inducing activity of expressed PLC mutants (Tables 1 and 2); that is, the higher activity shortened the delay time. The second and third Ca 2ϩ spikes occurred at an interval of ϳ20 min. The interval was shortened up to 10 min for succeeding Ca 2ϩ spikes (Fig. 1B). These Ca 2ϩ oscillations, which are probably caused by continuously produced IP 3 (24), lasted for 3-4 h and suddenly ceased prior to the formation of (female) PN at about 5 h after RNA injection. The higher PLC activity resulted in earlier termination of Ca 2ϩ oscillations, possibly because of a negative feedback via production of diacylglycerol and subsequent activation of protein kinase C (25), and/or down-regulation of IP 3 receptor type 1 which develops as a result of Ca 2ϩ oscillations, notably ϳ4 h after fertilization or parthenogenetic activation (26,27). It should be noted that expressed PLC was continuously accumulated into the formed PN ( Fig. 2A) as described previously (4). F in the PN (F PN ) became more than twice of F in the cytoplasm (F C ) 6 h after RNA injection ( Table 1). PLC that entered PN appeared to avoid the large nucleolus, which was identified as a round structure with a clear circumference in the bright field image ( Fig. 2A, paired photographs at the right).
Nuclear translocation of PLC was observed as well, when RNA was injected into the 1-cell embryo in which male and female PN were recognized 5 h after insemination. In the 1-cell embryo, Ca 2ϩ oscillations induced by IVF had already ceased (14,28), and another series of Ca 2ϩ spikes were induced by expressed PLC after a long delay of ϳ80 min and at long intervals of 40 -60 min (Fig. 1C). Phosphoinositide signaling pathway and/or IP 3 receptor-mediated Ca 2ϩ release seems to be suppressed in the 1-cell embryo, at the interphase of cell cycle (11). As shown in Fig. 2B (line 2), F PN was lower than F C at the early stage after RNA injection, but exceeded the latter at 3.5 h. Subsequently, F PN continuously increased and became twice of F C at about 8 h. Venus alone was more expressed than PLC-Venus (Fig. 2B, line 1) because of the smaller molecule. For Venus alone, F PN /F C was close to 1.0 (photographs of the inset in Fig. 2B; Table 1), indicating free diffusion through nuclear pores.
The nuclear translocation of PLC expressed in the 1-cell embryo after IVF served as a control for mutants that had quite low or no Ca 2ϩ oscillation-inducing activity and were incapable of activating the egg. For example, the mutant in which Asp 210 in the X catalytic domain (see Fig. 1A) was replaced with arginine (D210R) was defective in Ca 2ϩ oscillation-inducing activity even when overexpressed ( Fig. 1D and Table 1), as shown previously (2). Nuclear accumulation of D210R took place (Fig.  2C), but it was substantially slower, compared with that of wildtype PLC (Table 1). F in the nucleolus was comparable to that in the nucleoplasm 12 h after RNA injection (Fig. 2C, paired photographs at the right). Some fraction of PLC may to be accumulated to the nucleolus after a long delay.
Nuclear Localization Signal in PLC-According to NLS sequence searched from data base, PLC possesses two regions containing a cluster of basic amino acid residues (lysine and arginine), which is found in many nuclear proteins (29). One is in the C terminus of the X domain from 299 -308 (KFKIL-VKNRK) and another is in the X/Y linker region from 374 to 381 (KKRKRKMK). Paired mutation in Lys 432 and Lys 434 of PLC␦1 has been shown to prevent nuclear import (30). These residues correspond to Lys 299 and Lys 301 of PLC. The Ca 2ϩ oscillation-inducing activity was lost by replacement of Lys 299 or Lys 301 with acidic amino acid, glutamate, or both Lys 299 and Lys 301 with neutral amino acid, alanine (Table 1, K299A &  K301A). Nuclear translocation ability was also prevented by these mutations (Table 1). Thus, these lysine residues are important for nuclear import of PLC as in PLC␦1.
It has been shown (11) that K377E lacks nuclear translocation ability and that Ca 2ϩ oscillations induced by RNA encod-ing K377E continue without stopping at the stage of PN formation. Each amino acid from Val 373 to Ile 382 was replaced with glutamate. The mutants induced Ca 2ϩ oscillations that were similar to those induced by wild-type PLC (Table 1) and terminated a little prior to PN formation (Fig. 3A for K379E).
Expression level of these mutants might be higher than that in the previous work (11), and the feedback inhibition on Ca 2ϩ oscillations described above might predominantly operate in our experiments. Nuclear accumulation was negative for R376E, K377E, and K379E (Table 1; Fig. 4A for K379E). Replacement of Lys 377 with alanine showed prominent accumulation in PN, similar to wild-type PLC ( Table 1). Accumulation was faint for R378E and K381E (see Legend of Table 1 for the evaluation criteria), positive but delayed for K374E, and positive for V373E, K375E, M380E, and I382E (Table 1). Thus, Arg 376 , Lys 377 , Arg 378 , Lys 379 , and Lys 381 are essential for the nuclear translocation ability. To examine whether the NLS sequence is autonomously functional, a short fragment of Lys 374 -Ala 383 fused with Venus was co-expressed with Venus-free PLC. F PN /F C was close to 1.0, comparable to that of Venus alone (Table 1); that is, positive accumulation was not detected in PN.
Effects of Truncation at EF-hand Domain Region-Four EF-hand domains were defined by referring to those of PLC␦1, as indicated in Fig. 1A (18). s-PLC, which is expressed in the mouse testis and lacks EF1-3 from the N terminus, could induce Ca 2ϩ oscillations only when it was extremely overexpressed (⌬2-110 in Table 2). However, the first Ca 2ϩ transient appeared after a long delay of 3 h from the instance of RNA injection (Fig. 3B). The Ca 2ϩ oscillation-inducing activity of s-PLC was at least two orders of magnitude lower than that of full-length PLC, estimated from F at the time of the first Ca 2ϩ transient. Repetitive Ca 2ϩ spikes were generated at long intervals of 30 -40 min, and stopped at about the time of PN formation 5-6 h after the occurrence of the first Ca 2ϩ transient (Fig. 3B). No nuclear accumulation of expressed s-PLC was observed (Fig. 4B), even when RNA at a high concentration was injected into the 1-cell embryo ( Table 2).

Nuclear Translocation of PLC
extremely overexpressed (Table 2). A mutant of shorter truncation up to EF2 (⌬2-77; ⌬EF1-2-tr) exhibited delayed Ca 2ϩ oscillations, comparable to ⌬EF1-3-tr. Even shorter truncation up to EF1 (⌬EF1-tr) was incapable of inducing any Ca 2ϩ spike (Table 2). Thus, EF1 plays an important role in the Ca 2ϩ oscillation-inducing activity. The reason for discrepancy between ⌬EF1-tr and ⌬EF1-2-tr or ⌬EF1-3-tr is unclear, but EF2 apparently serves as an inhibitory factor in the absence of EF1. None of these truncation mutants underwent nuclear accumulation ( Table 2), indicating that the EF-hand region is also important for nuclear translocation.
Effects of Mutation at EF1-As aforementioned results lead us to predict the significant role of EF1, effects of modification of the N terminus and EF1 on nuclear translocation ability were precisely analyzed. ⌬2-4 had no significant effect ( Table 2). For ⌬2-9, however, nuclear accumulation was delayed and recognized 8 h after RNA injection ( Table 2). Ca 2ϩ oscillations were also affected in the prolongation of the delay time to ϳ50 min.
Surprisingly, both Ca 2ϩ oscillation-inducing ability and nuclear accumulation ability were lost for ⌬2-14 or ⌬2-19 (Table 2; Fig. 3C and Fig. 4C). Deletion of Glu 10 -Gln 19 associated with intact N terminus (⌬10 -19) deprived both abilities. Replacement of Glu 10 and Arg 12 with alanine had no significant effect, but substitution of three residues Arg 12 , Trp 13 , and Phe 14 caused loss of both abilities ( Table 2). With precise analysis of the two residues having an aromatic side chain, Trp 13 and Phe 14 , W13A or F14A could induce Ca 2ϩ oscillations, but the delay time was prolonged; that is, the activity is substantially lowered. W13A or F14A could not translocate to PN, whereas R12E had normal activities (Table 2). When tryptophan was replaced with phenylalanine (W13F) or the two residues were exchanged (W13F & F14W), the onset of Ca 2ϩ oscillations was substantially delayed, and active nuclear import was lost ( Table  2). Both abilities were preserved upon replacement of phenylalanine with tryptophan. Thus, Trp 13 is essentially necessary, and Phe 14 is replaceable with tryptophan. S16A, K17A, or K17E was ineffective. In contrast, V18A showed no nuclear translocation ability, even when extremely overexpressed in 1-cell embryos. V18A induced Ca 2ϩ oscillations, but the delay time was significantly prolonged (Table 2). Thus, Trp 13 , Phe 14 , and Val 18 are critical for nuclear translocation ability and are necessary to keep normal Ca 2ϩ oscillationinducing activity as well. The region between Glu 10 and Gln 19 functionally looks like an NLS. However, the sequence Met 1 -Gln 19 fused with Venus showed no positive accumulation into PN ( Table 2).
Effects of Deletion of EF-hand Domains-It is interesting to examine whether nuclear translocation takes place for a mutant in which EF1 is connected to the catalytic domain by deleting EF2-4 (⌬45-163; ⌬EF2-4). This mutant turned out to have no nuclear translocation ability and Ca 2ϩ oscillation-inducing ability (Table 2). Similarly, a mutant in which EF1 was connected to EF4 by deleting Asp 45 -Met 110 (⌬45-110; s-PLC preceded by EF1) lacked both abilities. Thus, EF1 is incapable of causing nuclear translocation without EF2-4; that is, the structure of EF-hand domain region as a whole is necessary.
Effects of Deletion or Modification of C2 Domain-Deletion of the C2 domain has been shown to cause the loss of PIP 2 -hydrolyzing activity in vitro (18) and Ca 2ϩ oscillation-inducing ability (10,18). In the present experiment, deletion mutant of the C2 domain (⌬522-625; ⌬C2) had no nuclear translocation ability ( Table 2). Point mutation was constructed at Asp 542 , because it corresponds to one of the putative Ca 2ϩ -ligating aspartates in the C2 domain of all four PLC␦ subtypes (31). Replacement of Asp 542 with alanine or arginine did not affect nuclear translocation ability as well as Ca 2ϩ oscillation-inducing activity ( Table 2).
Ca 2ϩ Oscillations and Nuclear Translocation in Cultured Somatic Cells-Ca 2ϩ oscillation-inducing activity and nuclear accumulation ability of PLC and its mutants tagged with Venus were investigated in COS-7 cells 24 -72 h after transfection with respective cDNA. COS-7 cells showed no spontaneous Ca 2ϩ spike (Fig. 5A). PLC expressed in COS-7 cells was capable of inducing Ca 2ϩ oscillations. Repetitive Ca 2ϩ spikes at intervals of ϳ3 min were recorded in 12/17 cells at 24 h after transfection (Fig. 5B), while no Ca 2ϩ spike was induced by D210R (Fig. 5C) in all 18 cells examined.
As to nuclear accumulation examined at 48 h, expression of Venus alone (Fig. 6A) showed that F in the nucleoplasm (F N ) was comparable to that in the cytoplasm (F C ); that is, Venus passively diffuses into the nucleus of somatic cells. Venus was not accumulated in the nucleoli, as indicated by black spots in the nucleus (Fig. 6A). Wild-type PLC translocated into the nucleus. Particularly, F in the nucleoli was strikingly enhanced, whereas F N was rather lower than F C at 48 h (Fig. 6B). It appears that the nuclear import of PLC is relatively slow, while PLC that entered the nucleoplasm is concentrated to nucleoli. At 72 h, PLC was accumulated in the nucleoplasm as well as nucleoli (Fig. 6C).
For D210R, F N was clearly higher than F C (Fig. 6D). Accumulation into nucleoli was observed to the lesser extent, compared with wild-type PLC. K377E completely lacked nuclear translocation ability (Fig. 6E). ⌬EF1-tr was hardly accumulated in the nucleoplasm (Fig. 6F), indicating that truncation of the N terminus causes remarkable suppression of nuclear translocation. The putative NSL sequence Lys 374 -Ala 383 fused with Venus was clearly accumulated into both nucleoplasm and nucleoli (Fig. 6G), when compared with Venus alone (Fig. 6A). In contrast, the sequence Met 1 -Gln 19 or Glu 296 -Val 309 fused with Venus showed comparable F N and F C (Fig. 6, H and I), indicating no positive accumulation into the nucleus.

DISCUSSION
NLS Sequence-The present study demonstrated the site or region of the PLC molecule responsible for the nuclear translocation ability by quantitative assay in mouse embryos and confocal microscopy of cultured somatic cells. NLS sequence was identified as the residues 374 -381 in the X/Y linker region. Addition of the sequence Lys 374 -Ala 383 to Venus (27-kDa protein), which diffuses through nuclear pores (F PN /F C ϭ 1.0) caused active nuclear import of Venus in COS-7 cells. Nuclear translocation of PLC-Venus (74 ϩ 27 kDa) relies primarily on the NLS in which basic amino acids, Arg 376 , Lys 377 , Arg 378 , Lys 379 , and Lys 381 , were essential. NLS is thought to be a binding site of the nuclear transport receptor (NTR). The residues 371-381 are well conserved in mouse, rat, human, monkey, pig, and cow (NCBI data bank).
Difference in Nuclear Distribution of PLC between Pronucleus and Nucleus-The active nuclear import of Venus fusion with K374-A383 did not occur in PN, unlike in the nucleus of COS cells. Additional import signal in PLC may be required for translocation into PN (see below). Another difference existed in the finding that PLC was hardly accumulated to the large nucleolus of PN, while Lys 374 -Ala 383 as well as PLC was localized in nucleoli of a COS cell. Lys 374 -Ala 383 involving a cluster of basic amino acids may serve as a nucleolar localization signal (32,33). In the nucleolus of PN, however, the presumptive nucleolar localization signal receptor might be not expressed.
In COS cells, expressed PLC was little accumulated in the nucleoplasm at 48 h after transfection, while strongly concentrated to the nucleolus (Fig. 6B). The rate of nucleolar accumulation may be much higher than the rate of net nuclear import. At 72 h, PLC was accumulated in the nucleoplasm (Fig. 6C) probably after saturation in nucleoli. This preferential nucleolar localization was less marked for D210R (Fig. 6D), suggesting that the rate of nucleolar targeting could be enhanced by Ca 2ϩ oscillations which are produced by PLC but not by D210R in COS cells (Fig. 5).
Import and Export Signals-PLC␦1 is a PLC isozyme similar to PLC (2) (38% identity and 49% similarity in 647 amino acid residues of PLC), although the PH domain is present in PLC␦1 but absent in PLC. PLC␦1 is not accumulated to PN (3) or the nucleus (34). PLC␦1 has a nuclear export signal at a leucine-rich sequence in EF1 (34) and an import signal at lysine-rich sequence in the C terminus of X domain and the X/Y linker of PLC␦1 (30). Import and export are balanced (30). I31-C43 in EF1 of PLC may correspond to the export signal sequence of PLC␦1. Lys 299 and Lys 301 in the C terminus of X domain of PLC were found to be responsible for nuclear import as in PLC␦1 (30). However, the region itself is not NLS, since the sequence E296-V309 fused with Venus did not show active nuclear import in COS cells. Lys 299 and Lys 301 (Lys 299 /Lys 301 ) are thought to be a supplemental component enabling nuclear translocation.  (16,35). D210R was able to undergo nuclear accumulation, although its rate was substantially lowered. Lys 299 /Lys 301 is close to one of the residues interacting with the substrate, PIP 2 , if considered in analogy with PLC␦1 (35). Hence they could affect Ca 2ϩ oscillationinducing activity. Lys 299 /Lys 301 are also close to the X/Y linker, hence they could affect nuclear translocation activity as well.
EF1 Domain-Ca 2ϩ oscillations and nuclear translocation are not prerequisite for each other, as they are dissected by point mutation in NLS and Asp 210 . The two independent abilities are affected in a parallel manner by mutational modifications in EF-hand domains and C2 domain. EF1 is an important domain for both abilities. According to the ProDom EF-hand pattern (36), Asp 20 -Ile 31 in EF1 is the Ca 2ϩbinding loop sequence. Ca 2ϩ oscillations and nuclear translocation are little affected by replacement at the x and z positions of the loop (D20A and G24A) (18). Thus, the Ca 2ϩ binding site in EF1 does not play a critical role in these abilities. ⌬2-14, ⌬2-19, or ⌬10 -19 is defective in both abilities. Point mutation in hydrophobic amino acids Trp 13 , Phe 14 , and Val 18 caused the loss of nuclear translocation and prolonged the onset of Ca 2ϩ oscillations. Trp 13 and Phe 14 are common to mouse, rat, human, monkey, pig, and cow. The residue 18 is Val 18 in mouse, pig, and cow, and Ile 18 in rat, human, and monkey. Thus, these hydrophobic residues are conserved. Although both Trp 13 and Phe 14 have an aromatic side chain, Trp 13 was not replaceable with phenylalanine while Phe 14 was replaceable with tryptophan. It seems that tryptophan at the exact positions in a hydrophobic moiety is the essential requirement. The sequence Glu 10 -Gln 19 fused with Venus did not show active translocation into the nucleus of COS-7 cells. The sequence corresponds neither to any known NLS nor to a non-classical NLS found in phospholipids scramblase 1 to interact with importin ␣ (37). Thus, the region is not NLS, but may necessary to take appropriate conformation for nuclear translocation of PLC (see below).
Four EF-hand Domains-Besides the presence of EF2-4, the C2 domain was necessary for both abilities. Deletion of EF2-3 or EF2-4 resulted in the loss of both abilities, even if EF1 was present. Truncation of EF-hand domains from the N terminus showed that Ca 2ϩ oscillations were barely induced after a long delay by extremely overexpressed s-PLC (⌬EF1-3-tr), but were never produced by ⌬EF1-4-tr. Thus, s-PLC is considered to be the minimal structure required for Ca 2ϩ oscillationinducing activity. EF3 is responsible for high Ca 2ϩ sensitivity of PIP 2 -hydrolyzing activity in vitro of PLC, as shown previously (18). Deletion of C2 domain caused the loss of both abilities, although the abilities were conserved upon point mutation of Asp 542 in C2 of PLC corresponding to the putative Ca 2ϩ -ligating aspartate of PLC␦ subtypes (31). A chimera formed by replacing the region from X domain to the C terminus of PLC with that of PLC␦1 has no detectable PLC activity (18). Taken together, the PLC activity of PLC seems to be derived from highly coordinated structure of EF-hand region and C2 domain.
A Model of Functional Structure-We tried to imagine a model of molecular structure to explain present results, as schematically drawn in Fig. 7. Crystal analysis of three-dimensional structure of PLC␦1 has shown that it is folded at the X/Y linker region in such a way that C2 domain in the C terminus makes extensive contact with EF-hand domains in the N terminus and the catalytic domain, forming the catalytic core (16,38). PLC is supposed to take the three-dimensional structure basically similar to that of PLC␦1. In wild-type PLC, association of EF1 with C2 is considered to be essential to take a compact form as the active conformation (Fig. 7A). The hydrophobic residues in the N terminus of EF1 may play an important role in EF1-C2 interaction. The site for substrate binding is given by close apposition of X and Y catalytic domains. The X/Y linker (309 -385) is a flexible region protruding from the catalytic domains. Lys 299 / Lys 301 in X domain is located close to the N terminus of the X/Y linker, and NLS (374 -381) is located at the C terminus of the X/Y linker, so that the two regions are likely to be neighboring each other. We postulate that K299/K301 is a component of the NTR binding site and that PLC-NTR association is accomplished by binding at both Lys 374 -Lys 381 and Lys 299 /Lys 301 regions (Fig. 7A).
⌬EF-tr or mutation in the N terminus of EF1 will cause dissociation of EF-hand domains from C2 domain. It is deduced that this conformation change may prevent the close apposition of X and Y, and thereby, perturb substrate binding (Fig.  7B). The change may also disturb NTR binding to both Lys 374 -Lys 381 and Lys 299 /Lys 301 region which became substantially distant each other. Thus, both Ca 2ϩ oscillation-inducing ability and nuclear targeting ability are lost.
From results presented in Table 2, EF2 apparently serve as an inhibitory factor in the absence of EF1. EF2 might enhance the dissociation between EF3-4 and C2 (Fig. 7B). When EF2 was deleted, the region of EF3-4 or EF4 could be substantially closer to C2 than in ⌬EF1-tr, yielding a slight Ca 2ϩ oscillationinducing activity (Fig. 7C). As ⌬EF1-4-tr is incapable of forming the catalytic core (Fig. 7D), it has no Ca 2ϩ oscillation-inducing activity. Nuclear translocation ability is lost by deletion of any EF-hand domains.
Biological Significance-A biological significance of translocation of the sperm factor or PLC into PN is postulated to be turning Ca 2ϩ oscillations off at the entrance of the interphase in the first cell cycle (11,15,39). At the transition from G 2 to M phase, Ca 2ϩ spikes resume upon nuclear envelope breakdown prior to the first cleavage and then disappear at the 2-cell stage (11,39). Artificially expressed PLC that entered the PN disperses into the cytoplasm upon nuclear envelope breakdown and is accumulated again into the nuclei of the 2-cell embryo (11,12). Thus, cytoplasm/nucleus shuttling of PLC is thought to be related to turning off and on Ca 2ϩ oscillations in a cell cycle stage-dependent manner.
A phosphoinositide signaling pathway is known to exist in the nucleus (40). For example, PLC␤1 translocates into the nucleus during G2/M transition in immature mouse oocytes and participates in germinal vesicle breakdown via diacylglycerol and protein kinase C (41,42). Nuclear accumulation of PLC might have some roles other than regulating Ca 2ϩ oscillations in early embryonic development.