Divergent effect of mammalian PLCζ in generating Ca²⁺ oscillations in somatic cells compared with eggs

PLCζ (phospholipase Cζ) is an isoform of phosphoinositide-specific PLC that is specifically expressed in mammalian sperm cells [1]. Microinjection of mouse PLCζ cRNA into mouse, human or bovine eggs triggers a prolonged series of Ca²⁺ oscillations similar to those observed at fertilization [1–5]. PLCζ has been identified as the protein responsible for causing Ca²⁺ oscillations in mouse eggs when either cytosolic sperm extracts or whole intact sperm are microinjected into the egg cytoplasm [1–6]. The critical level of PLCζ protein required to trigger Ca²⁺ release in mouse eggs is approximately 10–50 fg, which is similar to the estimated levels in a single mouse sperm [1]. Furthermore, reduced expression of PLCζ during spermatogenesis by transgenic RNAi (RNA interference) approaches has been shown to decrease the number of Ca²⁺ oscillations at fertilization in mouse eggs [7]. These data have led to the proposal that sperm PLCζ is the physiological activator of mammalian development that is delivered to the egg after sperm–egg membrane fusion [1–5]. PLCζ mediates its effects in eggs by virtue of its ability to hydrolyse PIP₂ (phosphatidylinositol 4,5-bisphosphate) and generate IP₃ (inositol 1,4,5-trisphosphate), since mutation of Asp²¹⁰ to an arginine residue (D210R), which is a critical residue in the PLCζ catalytic domain, abolishes its ability to cause Ca²⁺ oscillations or activate eggs [1]. Hence, PLCζ can account for all of the Ca²⁺ oscillations in eggs at fertilization since they are entirely dependent upon the stimulation of Ca²⁺ release via the type 1 IP₃ receptor [4,5,8,9].

It is not known whether PLCζ activity is regulated either in sperm or eggs. One significant feature of PLCζ that distinguishes it from many other mammalian PLCs is that its activity in vitro is retained in very low concentrations of Ca²⁺ without the requirement for other protein factors [3,10]. PLCζ is up to 50% active at resting cytoplasmic Ca²⁺ concentrations [3,10]. This implies that PLCζ would be autonomously active in generating Ca²⁺ oscillations when it is injected into the egg cytoplasm at low concentrations. In this case it is remarkable that the sperm contains a more than 1000-fold higher PLCζ concentration than the egg, yet sperm has a normal resting Ca²⁺ level and undergoes controlled Ca²⁺ oscillations in response to stimuli such as progesterone [11]. This implies that PLCζ might be inactive in sperm and that its activity is regulated differently in sperm compared with eggs. However, the nature of this differential regulation remains unclear.

Some previous studies have characterized recombinant PLCζ expressed in cell lines or somatic tissues [12,13]. It was reported that mouse PLCζ expression in COS cells is associated with the generation of Ca²⁺ oscillations [13]. Injection of boar sperm extracts (containing PLCζ) into hepatocytes has also been shown to trigger Ca²⁺ oscillations [14]. These data suggest that PLCζ can be autonomously active in somatic cells. However, another report showed that low levels of PLCζ expressed in all tissues of transgenic mice had no discernable effect upon somatic cells, despite the fact that it activated oocytes within the ovaries of females [15]. Ca²⁺ changes were not examined in that study, but it implies that low levels of PLCζ may be ineffective in somatic tissues. The only known mechanism for inactivating PLCζ is nuclear sequestration, which can account for the cessation of Ca²⁺ oscillations after fertilization in the mouse [16,17]. However, it is unclear whether PLCζ localization accounts for the different reports on the effect of PLCζ expression in somatic cells. PLCζ has been reported to be present in either the cytoplasm or the nucleus of somatic cells in a manner that may depend upon time, the cell line used or species source of PLCζ [12,18].

In the present study, we have expressed mouse PLCζ in CHO (Chinese-hamster ovary) cells, which have been widely used to study the mechanism of Ca²⁺ oscillations [19]. We find that very high levels of PLCζ expression in CHO cells has little effect upon Ca²⁺ homoeostasis, suggesting that it is inactive in these cells. Nuclear localization of PLCζ depended upon time after transfection, but is unlikely to account for the lack of Ca²⁺ oscillation in these cells. Remarkably we found that, despite the absence of an effect on the CHO cell Ca²⁺ level, the PLCζ-expressing CHO cells are able to cause Ca²⁺ oscillations when introduced into mouse eggs, and can be extracted and shown to have a high intrinsic PLC activity.

CHO cells were cultured in complete Ham's F12 medium (Invitrogen) in 5% CO₂ at 37 °C and seeded in six-well plates or in poly-L-lysine-coated glass-bottomed culture dishes (World Precision Instruments) 24 h prior to transfection at ~80–90% confluency. For transient expression, the nuclei of cells maintained in Leibovitz L-15 media were microinjected with plasmid DNA (~10 ng/μl) in KCl Hepes buffer [20] with 2.5 mg/ml Alexa Fluor® 594-dextran (10000 Da molecular mass) or 0.5 mM OBGD {Oregon Green BAPTA [1,2-bis-(o-aminophenoxy)ethane-N,N,N′,N′-tetra-acetic acid] dextran (Invitrogen)} using a Femtojet and InjectMan NI2 (Eppendorf) with an injection pressure of 90 hPa. The injection system was mounted on an Axiovert 200 fluorescence microscope (Zeiss). After injecting 50–100 cells, the medium was exchanged back to Ham's F12 and cells were returned to the incubator prior to imaging. For stable expression, CHO cells were transfected with DNA Lipofectamine™ 2000 (Invitrogen) and then incubated for 16–18 h prior to testing for heterologous PLCζ expression. PLCζ-expressing cells were selected on the basis of resistance to G418 by using an increasing concentration of G418 (0–1 mg/ml in 100 μg increments), and cells exhibiting appreciable YFP (yellow fluorescent protein) fluorescence were enriched using FACS (DakoCytomation, MoFlo). The PLCζ–luc (PLCζ–luciferase) fusion protein construct was made in a pCR3 vector (Invitrogen) as described previously [10]. The YFP–PLCζ construct was made in pcDNA3.1 with YFP placed at the N-terminus of PLCζ. The eYFP (enhanced YFP)–^D210RPLCζ construct was derived from YFP–PLCζ by site-directed mutagenesis of Asp²¹⁰ to arginine using the Stratagene QuikChange® XL site-directed mutagenesis kit.

Metaphase II mouse eggs were collected and microinjected with ~10 pl of extracts (approximately 5% of the egg volume) as described previously [10,20]. To make cell extracts, CHO cells (~2×10⁷) stably transfected with YFP-fusion constructs were harvested by trypsinization and then pelleted and resuspended several times in KCl Hepes buffer [10]. Cells were lysed with three cycles of freeze–thaw using liquid nitrogen, before being subjected to ultracentrifugation at 55000 rev./min (rotor Beckman TLA 100.4) for 1 h at 4 °C. The resulting supernatant was concentrated using centrifugal filters and stored at −80 °C. Microinjection of cells was performed using a Piezo micro-manipulator (PrimeTech) that mimics the process of somatic cell nuclear transfer [21]. Single CHO cells were drawn into the end of a pipette narrow enough to rupture their cell membrane then immediately injected into a mouse egg by application of low pressure. Before injection, eggs were loaded with either fura red or fura PE3 by incubation for 30 min in 2–4 μM of the AM (acetoxymethyl ester) forms for the dyes. Small groups of eggs were microinjected with cell or sperm extracts, or cells, and then rapidly transferred to the imaging microscope. Ca²⁺ was measured in eggs by monitoring fluorescence with a Nikon TiU epifluorescence microscope equipped with a Coolsnap HQ₂ CCD (charge-coupled-device) camera (Photometrics), with excitation light of 430 nm/490 nm (emission measured with a 615 nm long-pass filter) for fura red, and excitation of 350 nm/380 nm (emission measured at 535nm) for fura PE3, or excitation of 550 nm and an emission of 600 nm for Rhod dextran.

Boar sperm extracts (~5 mg/ml) were prepared as the cytosolic protein fraction in a similar manner to that of lysed pig sperm, as described previously [15]. Sperm extracts were mixed with an equal volume of 10000 Da molecular mass dextran conjugated with Alexa Fluor® 647 (Invitrogen) and microinjected into confluent fluo3-loaded CHO cells [22] using injection pressures of 120–140 hPa. The successful transduction of exogenous material into cells was confirmed by visualizing Alexa Fluor® 647 fluorescence using a 633 nm He/Ne laser fitted to an SP2 confocal laser-scanning microscope (Leica Microsystems). Following microinjection, Ca²⁺-dependent fluo3 signals were acquired at 1 Hz for 20 min as described previously [14]. Post-Ca²⁺ imaging, data was only analysed from those cells that retained >75% of their initial Alexa Fluor® 647 fluorescence.

PIP₂ hydrolytic activity of PLC present in YFP- or YFP–PLCζ-transfected CHO cell extracts was assayed as described previously [10]. The assay mixture final volume was 50 μl containing 100 mM NaCl, 0.4% sodium cholate, 2 mM CaCl₂, 4 mM EGTA, 20 μg of BSA, 5 mM 2-mercaptoethanol and 20 mM Tris/HCl (pH 6.8). The PIP₂ concentration in the reaction mixture was 220 μM, containing 0.05 μCi of [³H]PIP₂. Assay conditions were optimized for linearity, requiring a 10 min 25 °C incubation with 20 pmol of protein. Transfected CHO cell line proteins were separated by SDS/PAGE and immunoblot analysis was performed as described previously [10]. Proteins were probed with a peptide affinity-purified rabbit anti-PLCζ polyclonal antibody (1:10000 dilution) generated against the EF-hand domain (amino acid residues 16–31).

Luciferase expression was monitored from cells incubated in 100 μM luciferin using an Axiovert 100 microscope (Carl Zeiss) with a ×40, 1.4 NA (numerical aperture) oil-immersion objective lens and a cooled ICCD (intensified CCD) camera (Photek) [20]. Luminescence was collected for 20 min and then Ca²⁺ was monitored in the same cells for ~1 h using low-level fluorescence excitation light from a halogen lamp with a 490 nm excitation filter. Fluorescence from these cells was integrated every 10 s with sensitivity of the ICCD at 10% [20]. Resting Ca²⁺ concentrations were estimated using the equation

with F_min taken as zero and the K_d for OGBD in cytoplasm taken as 430 nM [23]. The F_max was obtained by adding ionomycin to saturate the fluorescence of OGBD or fluo3. Cells expressing YFP-fusion proteins were imaged with either a confocal or wide-field microscope. For confocal imaging we used a TCS SP5 and a DMI 6000 inverted microscope (Leica). Cells were imaged with a ×63, 1.23 NA oil-immersion objective lens, and a 512×512 pixel resolution. The YFP, fluo3 and fura red fluorescence was excited with the 488 nm laser line with emission at 510–540 nm for YFP or fluo3, and at 580–610 nm for fura red fluorescence. Alexa Fluor® 594 fluorescence was visualized using the 561 nm He/Ne laser and emission at 600–630 nm. For wide-field imaging we used a Nikon TiU epifluorescence microscope with a Photometrics Coolsnap HQ₂ CCD camera equipped with excitation and emission filter wheels. Excitation was with a 490 nm and 420 nm band-pass filter, and emission was monitored with a 530 nm band-pass filter (for YFP) or a 620 long-pass filter (for fura red). Where specified, cells were loaded with the Ca²⁺ indicator dyes by incubation in 10 μM of the AM form in non-supplemented Ham's F12 medium for 60–90 min before being washed and transferred to Krebs–Ringer–Hepes buffer [22,24]. For the confocal analysis of intracellular Ca²⁺ transients, cells were imaged every 100 ms for 100 frames before the addition of ATP (100 μM) and imaging for a further 500 frames (1 min total). SV (signal variability) is an index of the magnitude of amplitude and temporal characteristics of Ca²⁺ fluxes during steady-state Ca²⁺ cycling processes [22,24]. For SV analysis, data was acquired every 100 ms for 30 s. Following an experimental series, 15–20 cells were selected per area (typically four to six areas per dish), and data were analysed from ROIs (regions of interest) of ~40 μm² (>150 pixels) using Leica software. Further analysis was carried out using Microsoft Excel and GraphPad Prism. Derivation of mathematical operations used to calculate SV is given in a previous paper [23]. For wide-field imaging of YFP-transfected cells, an initial YFP image was collected, and the fura red images were collected every 10 s with a 420 nm/490 nm excitation, and data were analysed using Image-Pro Plus and Sigma Plot. Results were analysed for significance using ANOVA or a Student's t test.

PLCζ–luc has previously been shown to trigger Ca²⁺ oscillations when expressed in mouse eggs [10]. CHO cells were microinjected with cDNA encoding either PLCζ–luc or luciferase, together with the Ca²⁺ indicator OGBD. Figure 1(a) shows a luminescence image from a typical group of PLCζ–luc expression plasmid-injected cells indicating successful transient expression of the fusion protein, as reported by the corresponding fluorescence image of OGBD specifically in injected cells (Figure 1b). Intracellular Ca²⁺ was measured in PLCζ–luc-expressing cells 3, 5 and 24 h following microinjection, and the mean luminescence in single CHO cells expressing PLCζ–luc was in the range of 1–10 cps, compared with a range of 1–15 cps from cells injected with control luciferase vector. Figure 1(c) shows typical recordings of Ca²⁺ levels in PLCζ–luc and luciferase control cells, as well as OGBD control injected cells, 3 h after microinjection. It can be seen that there were small low-frequency Ca²⁺ oscillations in all three types of cells, but there was no clear difference between PLCζ-transfected cells and other cells transfected or not (see Table 1). Since CHO cells express the endogenous purinergic P2Y1, P2Y2 and P2X7 receptors, they were tested for responsiveness to extracellular ATP [25]. All of the cells underwent a single large rise in Ca²⁺ in response to ATP (Figure 1c). Table 1 summarizes data from transfected cells analysed at 3, 5 and 24 h post-microinjection. At all three time points there were no differences in the proportion of cells showing low-frequency Ca²⁺ oscillations, or in those showing a Ca²⁺ increase in response to addition of ATP, between PLCζ–luc and luciferase control CHO cells and control OGBD-injected cells (see Table 1). The resting Ca²⁺ levels in CHO cells were estimated by the addition of a Ca²⁺ ionophore. Figure 1(d) shows that the resting Ca²⁺ levels were not significantly different between PLCζ–luc or luciferase control cells, or in comparison with OGBD-injected cells, at either 3, 5 or 24 h after microinjection. These data suggest that transient PLCζ–luc expression does not have a significant effect upon Ca²⁺ homoeostasis in CHO cells. The presence of low-frequency small amplitude Ca²⁺ oscillations in some CHO cells evidently occurs in these cells under certain conditions.

Nuclear sequestration of PLCζ appears to cause the cessation of Ca²⁺ oscillations in mouse eggs and could account for the lack of CHO cell Ca²⁺ oscillations in the above experiments [16,17]. We examined the distribution of PLCζ using a YFP-tagged PLCζ which, similar to VenusGFP–PLCζ (GFP is green fluorescent protein), is also effective in causing Ca²⁺ oscillations in mouse eggs (see Supplementary Figure S1a at http://www.BiochemJ.org/bj/438/bj4380545add.htm). Figure 2(a) shows confocal images of CHO cells at 3, 5, 24 and 48 h after being injected with the YFP–PLCζ expression plasmid. Distinct levels of recombinant YFP expression were detected as early as 3 h after DNA microinjection and persisted for at least 48 h for all of the DNA constructs examined. Expression of YFP alone was homogeneous throughout the nucleus and cytoplasm in all cells. In contrast, YFP–PLCζ showed two distinct patterns of localization. At 3 and 5 h post-injection, all cells showed a cytoplasmic distribution, whereas after 24 and 48 h the majority of cells showed nuclear translocation with enhanced staining in the nucleoli. Some cells that did not show obvious nuclear translocation at 24 and 48 h maintained a clear cytoplasmic distribution identical with that observed at 3 and 5 h. Unlike YFP–PLCζ, YFP–^D210RPLCζ did not show any obvious signs of nuclear accumulation at any time points, and the majority of cells (>90%) maintained the same cytoplasmic localization.

As well as studying transiently transfected CHO cells, we also examined the distribution of PLCζ in stable cell lines. We prepared CHO cell lines that stably expressed YFP–PLCζ or YFP alone after two rounds of cell sorting. Figure 2(b) shows that after such sorting there was a clear difference in YFP channel fluorescence between selected cells and the wild-type CHO cells. The distribution of YFP–PLCζ in CHO cells was variable and again could be seen in either the cytoplasm or the nucleus (Figure 2b). Along with the transient transfection studies, these data support the previous observations that PLCζ can localize to the nucleus of cells [12]. However, since localization does not occur in all cells or takes more than 5 h, this suggests that nuclear translocation cannot account for its apparent inactivity in CHO cells.

Following microinjection of CHO cells with YFP–PLCζ, we measured cytoplasmic Ca²⁺ changes using confocal microscopy. We studied the response of transduced cells to the addition of ATP, as described above with PLCζ–luc-expressing cells, but the faster rate of image acquisition (10 Hz) enabled by confocal microscopy permits a more detailed kinetic analysis of basal Ca²⁺ handling and ATP-evoked Ca²⁺ release in these cells. Cells were loaded with fluo3 and, after exposure to ATP, the amplitude and relative change in fluorescence was taken as a measure of response for YFP-, YFP–PLCζ- and YFP–^D210RPLCζ-transfected cells, and for control-injected cells. The initial increase in Ca²⁺ was unaffected by the removal of extracellular Ca²⁺, indicating that P2Y and IP₃-dependent intracellular Ca²⁺ store release was involved (S. Phillips and K. Swann, unpublished work). Figure 3(a) shows examples of the Ca²⁺ response to a maximal concentration of ATP (at the arrow) for all four cell types, which consisted of a large and rapid response followed by a gradual decline, as shown in Figure 1(c). The responses in Figure 3(a) are at 5 and 24 h, but similar results were seen at 3 and 48 h after DNA injection. Figure 3(b) summarizes the data from for all four types of cells, at all time points. There was no difference in the proportion of cells responding to ATP, the relative amplitude of the Ca²⁺ rise or in the time to peak of the Ca²⁺ rise evoked by ATP at any of the different times after microinjection of DNA. These data again suggest that P2Y-receptor-driven IP₃ production and Ca²⁺ store release is unaltered by PLCζ expression.

The use of YFP–PLCζ and confocal microscopy allowed us to examine more subtle changes in Ca²⁺ homoeostasis than may be revealed by the experiments described above. Previous studies of cardiac RyR2 (ryanodine receptor 2)-transfected CHO cells have shown that there is an increased variability in resting cytoplasmic Ca²⁺ fluxes associated with heterologous expression of RyR2 [22,24]. Such SV is an index that robustly describes the signal noise in the basal Ca²⁺-dependent fluorescence from Ca²⁺ dyes in the cell cytoplasm, and has previously been used to characterize subtle perturbations of steady-state Ca²⁺ cycling and their functional consequences (e.g. induction of apoptosis) [22,24]. To examine whether recombinant expression of YFP–PLCζ resulted in subtle changes to CHO cell Ca²⁺ handling, we monitored Ca²⁺ with fura red since the emission peaks of YFP and fura red are separated by >100 nm. Figure 3(b) (bottom right-hand panel) shows the analysis of fura red fluorescence noise as measured by SV for YFP, YFP–PLCζ and YFP–^D210RPLCζ-transfected cells, and for control-injected cells at 5 h and 24 h after microinjection. There was no difference observed in the SV from all four groups of cells. These experiments suggest that there are no subtle alterations in Ca²⁺ homoeostasis in CHO cells expressing YFP–PLCζ, supporting the conclusion that the YFP–PLCζ expressed is functionally inactive in CHO cells.

As well as examining Ca²⁺ dynamics in CHO cells transiently expressing recombinant protein, we analysed Ca²⁺ in the YFP–PLCζ stable cell lines and studied the Ca²⁺ response to a full range of ATP concentrations. Figure 4(a) shows the Ca²⁺ transients generated by different concentrations of ATP in control CHO cells and in YFP–PLCζ-transfected cells. We categorized the cell populations into those that expressed high or low levels of recombinant protein, based on the FACS YFP fluorescence signals, and analysed the populations on this basis. As with the transient transfections, there were no differences in the Ca²⁺ response to high concentrations of ATP. Nevertheless, the data shown in Figure 4 indicate that, for low concentrations of ATP, a smaller proportion of both types of YFP–PLCζ-expressing cells responded with a Ca²⁺ transient. The response in YFP–PLCζ cells was also significantly delayed compared with control CHO cells (Figure 4b). However, this small difference in response was independent of the expression level of YFP–PLCζ since it was equivalent in those cells expressing either high or low levels of YFP–PLCζ recombinant protein.

The lack of any dramatic change in Ca²⁺ homoeostasis in YFP–PLCζ-expressing cells (Figure 3b) did not appear to be due to any degradation of the PLC portion of the YFP–PLCζ fusion protein, nor in any loss of its intrinsic PLC activity. We made cytosolic extracts from YFP–PLCζ cells and tested then for PLCζ activity and expression. We specifically assayed PLC activity of cells in a mixed micelle assay with a free Ca²⁺ concentration at 200 nM which is near maximal for stimulating PLCζ, but below the level required for stimulating most somatic PLC isoforms [3,10]. Figure 5(a) shows that cytosolic extracts made from YFP–PLCζ-expressing cells had a much higher PLC activity in vitro compared with control YFP-transfected cells. This is consistent with the robust expression of the full-length YFP–PLCζ fusion protein (103 kDa), as illustrated by the clear immunoreactive protein of ~100 kDa recognized by an affinity-purified anti-PLCζ antibody on the Western blot (Figure 5b).

One reason why PLCζ may not be active in somatic cells could be a requirement for a specific factor(s) present only within the egg. To test this idea we sought to introduce recombinant PLCζ synthesized in CHO cells into mouse eggs. Cell extracts were prepared from untransfected CHO cells (negative control) or from cells exhibiting stable expression of YFP (~73.4±0.6% positively labelled), or else YFP–PLCζ (97.6±0.1% positively labelled). These extracts were prepared in the same way as described previously for making cytosolic sperm extracts [1,26]. Figure 6(a) shows that injecting a ~5% volume of YFP–PLCζ CHO cell extract into mouse eggs triggered a series of Ca²⁺ oscillations in all eggs injected (n=14). The frequency of the Ca²⁺ oscillations observed was 11.4±1.4 spikes/h, which persisted for the duration of the recording (typically 1–1.5 h). These Ca²⁺ oscillations are similar in form to those previously described after injection of mouse eggs with cytosolic sperm extracts [26]. Injection of extracts from untransfected CHO cells or YFP-transfected CHO cells did not cause any changes in Ca²⁺ in mouse eggs (Figure 6b).

We also tested the activity of PLCζ expressed in CHO cells by directly injecting these cells into the egg cytoplasm in the manner used in the somatic cell transfer procedure [21]. Intracellular Ca²⁺ was monitored immediately following microinjection of either untransfected CHO cells (negative control) or cells expressing YFP–PLCζ (from an enriched stable cell population). Single CHO cells were injected into single mouse eggs and then transferred to the fluorescence imaging system within 1–2 min. Figure 6(c) shows that the injection of cells expressing YFP–PLCζ induced a series of repetitive Ca²⁺ oscillations in the majority (15/23) of eggs. As was observed with the CHO cell extract injections, the egg Ca²⁺ response to single CHO cell injection displayed rapid spikes, with the initial transient taking place within minutes of the cell injection. The frequency of the Ca²⁺ oscillations observed was 4.3±0.8 per h, which persisted for the duration of the recording and is similar to that seen during mouse egg fertilization [2,4]. In contrast, Figure 6(d) shows that no change in intracellular Ca²⁺ occurred following microinjection of negative control cells. These data suggest that the YFP–PLCζ-transfected CHO cells are expressing YFP–PLCζ at levels capable of inducing Ca²⁺ oscillations in mouse eggs. Furthermore, the rapid onset of Ca²⁺ oscillations following injection into egg cytoplasm suggests that the PLCζ expressed in CHO cells exhibits full functional activity.

Since PLCζ appears to be inactive in CHO cells we considered whether this might be due to an intrinsic inability of these cells to respond to PLCζ. Previously, sperm extracts were used as a source of an endogenously active Ca²⁺-releasing factor and they contain high levels of PLCζ [1]. We made cytosolic boar sperm extracts and microinjected them into CHO cells. Figures 7(a)–7(d) show experiments in which boar sperm extracts were injected into cells along with a fluorescent tracer dye. In all cases, the injection of such sperm extracts failed to cause any changes in resting Ca²⁺, and their response to sperm extracts was indistinguishable from microinjection of fluorescent tracer dye alone (Figures 7c and 7d). The same extracts caused high-frequency Ca²⁺ oscillations when injected into mouse eggs (Figure 7e); so these data suggest that CHO cells are not able to generate a response to native sperm PLCζ, even when it is introduced directly by microinjection into their cytoplasm.

Consistent with previous expression studies using Venus- or YFP-tagged PLCζ in COS cells, we found that either YFP– or luciferase–PLCζ could be expressed in CHO cells [12,13]. The previous reports of PLCζ expression in non-gametes gave apparently contradictory results in that a study using cell lines showed that PLCζ expression triggered Ca²⁺ oscillations, whereas transgenic expression of PLCζ throughout the mouse soma did not lead to any clear phenotypic change [15]. The PLCζ transgenic whole soma expression levels were probably much lower than our studies with mammalian cell lines since we could readily detect expression, and yet expression in the transgenic mice was not detectable using a Venus–PLCζ fusion protein [15]. Notably, our results indicate that readily detectable expression of two different PLCζ-fusion proteins can be achieved in a cell line with no apparent perturbation of resting Ca²⁺ levels, no major change in Ca²⁺ homoeostasis and no altered response to agonists that increases cytoplasmic Ca²⁺. The absence of any effect of expressed PLCζ on cytoplasmic Ca²⁺ is surprising since previous work implied that PLCζ is autonomously activated inside cells. Moreover, our results also demonstrate that the PLCζ expressed within CHO cells is either already active, or can subsequently become activated when it is assayed in vitro or introduced into eggs, thus implying that PLCζ may be held in an inactive state in CHO cells. This is also consistent with our findings that cytosolic sperm extracts, which contain an endogenous active PLCζ [1], were also incapable of causing any Ca²⁺ release when directly microinjected into normal CHO cells.

Previous studies showing expression of mouse PLCζ in somatic cells lines have reported either nuclear or cytoplasmic localization [12]. These observations could provide an explanation for some of the differences in the reported functional effects of PLCζ expression. The results of the present study show that the specific localization of PLCζ within cells may be dependent upon the elapsed time after transfection. We observed this time-dependence because in our studies we microinjected the plasmid DNA into cells, hence the timing of cDNA transduction was more precisely determined than with other methods. The localization of YFP–PLCζ is primarily retained within the cytoplasm for the first 3–5 h after injection, and is subsequently found in the nucleus after 24 h. This time period should be sufficient for some cells to complete a cell cycle and hence may imply that differential localization occurs after a round of mitosis. It was of interest that YFP–^D210RPLCζ did not undergo any nuclear localization in CHO cells, which is similar to what we have seen in mouse eggs (Y. Yu and K. Swann, unpublished work). Since this mutant form of PLCζ lacks the ability to bind its substrate, this may suggest that PIP₂ binding plays some role in nuclear translocation. Previous studies with Venus–PLCζ shows that a rapid (<1 h) localization occurred in the nuclei of mouse zygotes and embryos [16,17], suggesting that the mouse PLCζ expressed in CHO cells is either very slowly transported into the nucleus, or alternatively it may require a mitotic transit to become localized. In either case, the time-dependent localization we observe may explain the variation in previous reports of localization of PLCζ that may be due to the uncertain time at which transcription starts using liposome-based methods. Nevertheless, the observed nuclear localization is unlikely to account for the lack of PLCζ activity in CHO cells since as it occurs many hours after transfection and with high levels of YFP–PLCζ protein expression, some cytoplasmic localization will invariably overlap with nuclear localization. In stably transfected YFP–PLCζ cells there was also some variability in nuclear localization, but the homogeneity of Ca²⁺ handling in all cells confirmed that nuclear sequestration cannot readily account for an apparent lack of activity of PLCζ in cells.

Semi-quantitative estimates suggest that we have expressed relatively high levels of recombinant PLCζ protein in CHO cells. From our previous calibration of luciferase expressed in mouse eggs [10], we can determine that the luminescence from single CHO cells expressing PLCζ–luc (~10 cps) corresponds to >100 fg of the fusion protein. We can also calculate the amount of YFP–PLCζ protein present in our experiments. The amount of YFP–PLCζ in the selected high-expression cells is at least 10-fold the detection threshold level, which is ~1 μM [27]. So our levels should correspond to approximately 10 μM. Given a CHO cell volume of ~2 pl, our transfected and selected cells should contain ~2 pg of YFP–PLCζ protein (relative molecular mass ~100 kDa). The estimated ~2 pg/CHO cell is significantly greater than the 20–50 fg of PLCζ that was previously shown to trigger a physiological pattern of Ca²⁺ oscillations in mouse eggs [1]. Given that CHO cells are only ~1% of the volume of a mouse egg, it appears that CHO cells can tolerate expression levels of ~200-fold PLCζ–luc and ~2000-fold YFP–PLCζ higher than is required to cause IP₃-dependent Ca²⁺ oscillations in mouse eggs.

It is not clear why PLCζ is apparently inactive in CHO cells. One possibility is that our transfected CHO cells have adapted to high levels of PLCζ by down-regulating the number of IP₃ receptors. Given the high levels of expressed PLCζ, this would amount to the effective ablation of IP₃ receptors. However, this seems unlikely to be occurring in our experiments because such down-regulation only occurs after generation of considerable amounts of IP₃ that are sufficient to cause high-frequency Ca²⁺ oscillations [8,9]. Notably, we failed to detect any Ca²⁺ changes, even at early times after transfection with PLCζ. There is also insufficient time for IP₃ receptor down-regulation to explain the lack of response to the sperm extracts. Furthermore, any InsP₃ receptor down-regulation leads to a loss of agonist-induced Ca²⁺ increase [28]. We did find that there was a reduction in the Ca²⁺ response to low doses of ATP for stably transfected YFP–PLCζ cells. However this effect was not related to the expression level of YFP–PLCζ. It is also noteworthy that previous work on ATP-induced Ca²⁺ response in CHO cells showed that PKC (protein kinase C) stimulation causes a marked feedback inhibition of Ca²⁺ transients [29]. Therefore if PLCζ was causing sustained PIP₂ hydrolysis and hence DAG (diacylglycerol) production, we would anticipate a substantial block of the ATP Ca²⁺ transients. Since we consistently failed to detect any change in Ca²⁺ response at high ATP concentrations, the simplest interpretation of our data is that the recombinant PLCζ is inactivated when expressed in CHO cells.

The apparent lack of PLCζ activity in transfected CHO cells, coupled with clear Ca²⁺-releasing activity when these cells are introduced into eggs, resembles the physiological situation with regards to fertilization. Given the size difference between the sperm and egg, the mammalian sperm should contain levels of PLCζ that must be at least three orders of magnitude higher than is effective in causing Ca²⁺ release in eggs. The mammalian sperm has a resting Ca²⁺ level similar to that of other cells and can exhibit Ca²⁺ increase in response to agonists, some of which appear to work via IP₃ generation [11]. This implies that the PLCζ present in sperm is also inactive. However, PLCζ is capable of causing the first Ca²⁺ increase at fertilization within minutes of sperm–egg membrane fusion [30]. Therefore the experiments where we inject transfected CHO cells into the egg mimics the fertilization of the egg by a sperm in that an inactive PLCζ rapidly becomes active upon exposure to the egg cytoplasm. The high intrinsic PLC activity of sperm extracts also resembles the high in vitro PLC activity we find in YFP–PLCζ-transfected CHO cell cytosolic extracts. There may be two explanations for such effects. One is that PLCζ may be held in an inactive state in sperm, and that whatever the mechanism the inhibitory components are also present in CHO cells. Otherwise the activity of PLCζ observed upon entry into eggs could be because it interacts specifically with a protein(s) that is only expressed in the egg cytoplasm. This possibility would be consistent with the fact that other known mammalian phosphoinositide-specific PLCs are also stimulated by interaction with accessory proteins [31].

AM

acetoxymethyl ester

CCD

charge-coupled-device

ICCD

intensified CCD

CHO

Chinese-hamster ovary

IP₃

inositol 1,4,5-trisphosphate

OBGD

Oregon Green BAPTA [1,2-bis-(o-aminophenoxy)ethane-N,N,N',N'-tetra-acetic acid] dextran

PLC

phospholipase C

PLCζ–luc

PLCζ–luciferase

PIP₂

phosphatidylinositol 4,5-bisphosphate

RyR2

ryanodine receptor 2

SV

signal variability

YFP

yellow fluorescent protein

Sally Phillips carried out most of the experiments on CHO cells, and Yuansong Yu carried out experiments on eggs and some of the studies on CHO cells. Andreas Rossbach prepared the DNA constructs. Christopher George contributed to the analysis of confocal studies of CHO cells and microinjected extracts for Figure 7. Michail Nomikos conducted biochemical analysis of the PLC protein, and Vyronia Vassilakopoulou, Evangelia Livaniou and Bevan Cumbes prepared the rabbit anti-PLC antibody. F. Anthony Lai contributed towards the conception of the project and drafting of the paper. Karl Swann supervised the project and prepared the final version of the paper.

We thank Nazar Amso for his support of S.V.P. on the early part of this project.

This work was supported by the Biotechnology and Biological Sciences Research Council [grant number BB/C513142/1]; Medical Research Council [grant number G0801676]; and the Wellcome Trust [grant number 080701/Z/06/Z]. V.V. holds a research scholarship from NCSR Demokritos.