PLCζ: a sperm-specific trigger of Ca²⁺ oscillations in eggs and embryo development

Activation of egg development in all animals and plants is produced by the fertilising spermatozoa triggering an acute rise in cytosolic free Ca²⁺ concentration (Stricker, 1999). In mammals, the unification of sperm and egg leads to a distinctive series of cytosolic Ca²⁺ oscillations that are a prerequisite for normal embryo development (Miyazaki et al., 1993; Stricker, 1999). This striking Ca²⁺ signalling phenomenon arises from increases in inositol 1,4,5-trisphosphate (IP₃) levels, which activate IP₃ receptor-mediated Ca²⁺ release from intracellular stores in the egg (Miyazaki et al., 1993; Brind et al., 2000; Jellerette et al., 2000). However, the basic mechanism that results in stimulation of phosphoinositide (PI) metabolism following sperm-egg interaction has not been determined in any species.

The sperm factor hypothesis of signalling at fertilisation proposes that spermatozoa contain a soluble Ca²⁺ releasing factor that enters the egg after the gamete membranes fuse together and generates Ca²⁺ oscillations (Swann, 1990; Stricker, 1999). This is consistent with the finding that cytoplasmic fusion of sperm and egg is a prelude to Ca²⁺ release (Lawrence et al., 1997; Jones et al., 1998a). Direct support for this hypothesis comes from experiments where microinjection into eggs of either single spermatozoa, or soluble sperm extracts, triggers Ca²⁺ oscillations similar to those at fertilisation in mammalian and some non-mammalian eggs (Swann, 1990; Wu et al., 1997; Wu et al., 1998; Stricker, 1997; Nakano et al., 1997; Kyozuka et al., 1998; Tang et al., 2000). The mammalian sperm factor that generates Ca²⁺ oscillations is protein based (Swann, 1990), acts across species (Wu et al., 1997), and can cause Ca²⁺ release in somatic cells (Berrie et al., 1996) and in cell-free systems such as sea urchin egg homogenates (Jones et al., 1998b). Sperm specifically express a Ca²⁺ oscillation-inducing protein, because microinjecting mRNA isolated from spermatogenic cells, but not mRNA from other tissues, elicits fertilisation-like Ca²⁺ oscillations in mouse eggs (Parrington et al., 2000). Despite intensive biochemical investigation, the molecular identity of the putative sperm factor has remained elusive (Stricker, 1999). Different proteins, including a 33 kDa protein (Parrington et al., 1996) and a truncated form of the Kit receptor (Sette et al., 1997), have previously been sperm factor candidates. However, neither these two, nor any other sperm proteins, have been shown to generate Ca²⁺ oscillations in eggs (Wu et al., 1998; Wolosker et al., 1998), the single-most distinctive feature of mammalian fertilisation (Stricker, 1999).

In intact eggs and egg homogenates, mammalian sperm extracts trigger Ca²⁺ release by stimulating IP₃ production (Jones et al., 1998b; Rice et al., 2000; Jones et al., 2000; Wu et al., 2001), indicating involvement of a PI-specific phospholipase C (PLC) in the signal transduction mechanism. The high level of PLC enzyme activity measured biochemically in sperm extracts suggests that the sperm factor may itself be a PLC (Jones et al., 1998b; Rice et al., 2000). However, the PLCβ, γ and δ isoforms that exist in sperm are absent from chromatographic fractions of sperm extract that specifically cause Ca²⁺ oscillations (Wu et al., 2001; Parrington et al., 2002). In addition, when the purified, recombinant PLCβ1, γ1, γ2 or δ1 proteins are added to egg homogenates or microinjected into eggs, they fail to cause Ca²⁺ release (Jones et al., 2000). A PLCδ4 splice variant found in sperm functions in the acrosome reaction, rather than in Ca²⁺ release in eggs at fertilisation (Fukami et al., 2001). These observations led us to investigate the possible existence of a distinct, uncharacterised sperm PLC isoform. Our studies reveal that a new PLC isoform (PLCζ), specifically expressed in mammalian sperm, uniquely possesses all the essential properties of the sperm factor. These results are consistent with sperm PLCζ as the physiological trigger of egg activation, and thus an essential protein for mammalian fertilisation and embryo development.

Rabbit antisera was raised to a 19-mer sequence, GYRRVPLFSKSGANLEPSS, within the mouse testis ESTs identified as homologous to PLC (see below). Mammalian tissues and sperm cytosolic proteins (Parrington et al., 1996), were separated by 10% SDS-PAGE, transferred to PVDF membrane and probed with anti-peptide antisera. Heparin affinity and gel filtration column chromatography of soluble sperm proteins (10 mg) used an AKTA FPLC system (Amersham Pharmacia) (Parrington et al., 1996) in 50 mM sodium phosphate, 0.15 M NaCl, pH 7. Fluorometric Ca²⁺ release assays using fluo-3 in sea urchin homogenates (Jones et al., 1998b) was monitored using an LS50B (Perkin-Elmer).

Blast searches of the mouse EST database using mammalian PLC sequences (www.ncbi.nlm.nih.gov/BLAST) identified 12, novel PLC-related sequences (Accession Numbers, AV282878, AV278700, AV278207, AV272100, AV271735, AV270614, AV270212, AV263382, AV263095, AV258739, AV258594 and AV045146). These mouse ESTs were 232-294 basepairs with identical 3′ sequences and all derived from testis. The full-length sequence encoding this novel PLC, named PLCζ, was obtained by two-step RACE PCR amplification with pfu polymerase from a mouse spermatid cDNA library (35 ng) in lambdaZAPII. The single amplified DNA of 2.2 kb was cloned into pCR-XL-TOPO (Invitrogen), ten independent colonies were sequenced on both strands, and analysed for open reading frame by MacVector 6.5 (Oxford Molecular), for PLC homology and phylogeny by ClustalW sequence alignment (www.clustalw.genome.ad.jp) and domain structure by RPS-Blast (www.ncbi.nlm.nih.gov/structure/cdd). The GenBank Accession Number for PLCζ is AF435950.

A 1.2 kb probe from the 5′ end of mouse PLCζ, prepared by PCR as above, was cloned into pCR-BluntII-TOPO (Invitrogen) and sequenced. Antisense digoxigenin-labelled RNA synthesised from this plasmid (DIG Nucleic acid labelling system, Roche Molecular Biochemicals) was used to probe a male mouse tissue polyA⁺-RNA blot with equal loading of 2 μg polyA⁺-RNA/lane (MessageMap Northern, Stratagene). Hybridised probe was detected using the DIG Luminescence Detection Kit (Roche Molecular Biochemicals) and displayed using QuantityOne software (BioRad). Polymerase chain reaction amplification using 30 cycles was performed with oligonucleotide primers that define a 0.9 kb region within PLCζ, using cDNA prepared from mouse spermatids or mouse testis devoid of spermatids in the lambda ZAPII vector (10 ng). Negative and positive controls comprised reactions without DNA template and with PLCζ plasmid DNA (1 ng), respectively.

The 1941 bp open reading frame of mouse PLCζ was cloned into pCR-Blunt II-TOPO, sequenced and subcloned (pTarget, Promega) to generate pTarget-mPLCζ. Complementary RNA (cRNA) was synthesised from linearised pTarget-mPLCζ (Ribomax RNA synthesis, Promega) in the presence of 3 mM m⁷G(5′)ppp(5′)G, isopropanol precipitated and resuspended in DEPC-treated water containing 4 U/μl RNasin (Promega). Mutagenesis of ²¹⁰Asp to ²¹⁰Arg in PLCζ to produce ^D210RPLCζ was achieved using the QuikChange Site-Directed Mutagenesis Kit (Stratagene). Constructs and cRNAs for rat PLCδ1 and ^ΔPHPLCδ1, which encoded the full-length (756 amino acids) and PH domain-deleted PLCδ1 (Δ1-132), respectively, and ^D210RPLCζ were produced in pTarget as above. cRNA (2 μg) was expressed in vitro (Reticulocyte lysate system, Promega) in the presence of [³⁵S]methionine (Amersham Pharmacia). Radiolabelled protein, analysed by SDS-PAGE and autoradiography, was displayed using QuantityOne software (BioRad).

The 1941 bp open reading frame of mouse PLCζ was subcloned into pGBK-T7 (Clontech) with an in-frame Myc epitope tag at the 5′-end. The Myc-PLCζ was further subcloned into pcDNA3.1 and sequence-verified before cRNA synthesis from the T7 site (Ribomax) for egg microinjection, as described above. For bacterial expression, Myc-PLCζ was subcloned into pBAD (Invitrogen) with an in-frame hexahistidine tag at the 3′ end. The Myc-PLCζ-Histag protein was produced in 0.2% w/v arabinose-induced, BL21(DE3)pLysS E. coli, after extraction of the pelleted bacteria by five freeze-thaw and ultrasonication cycles, then purified by nickel affinity chromatography (ProBond, Invitrogen). Protein quantitation was performed using the BCA protein assay (Pierce).

Densitometric analysis of the Myc-PLCζ band expressed in eggs microinjected with different cRNA concentrations (Fig. 6C), Myc-PLCζ-Histag protein purified from E. coli, and calibrated sperm extract PLCζ derived from 10⁴-10⁶ mouse sperm, employed a Myc monoclonal antibody (1:2000, Santa Cruz Biotechnology) and rabbit anti-PLCζ antiserum (1:1000), respectively, using QuantityOne software (BioRad). A calibration standard plot, from analysis by immunoblot densitometry (Malek et al., 1997) using the Myc antibody, was constructed using defined amounts of Myc-PLCζ-Histag protein, purified from E. coli, to enable the calculation of the relative Myc-PLCζ content in batches of 100 microinjected eggs. For the quantitation analysis, expression of the Myc-PLCζ protein was assumed to be linear with time after cRNA microinjection, as was shown for microinjected EGFP cRNA expressed in mouse eggs (Aida et al., 2001). This assumption was necessary because the c-Myc-PLCζ protein was below the detection limit within 3 hours of cRNA microinjection (data not shown). Hence, for a single mouse egg, the calculated 440-750 fg of Myc-PLCζ protein expressed 5 hours after microinjection with 0.02 mg/ml cRNA, was equivalent to 44-75 fg expressed at 0.5 hours, the time when the first Ca²⁺ transient is normally observed (Fig. 5B). A separate calibration plot using the anti-PLCζ antibody was constructed with different Myc-PLCζ-Histag protein concentrations to enable estimation of the relative PLCζ content in defined numbers of mouse sperm (Fig. 6D).

Soluble extracts (Parrington et al., 1999) prepared from hamster sperm were incubated for 1 hour at 4°C with control IgG or anti-PLCζ. antibody that had been covalently attached to Protein G beads (1 mg/ml, Seize X Kit, Pierce). The PLCζ content of the supernatant and precipitated beads was determined by immunoblot analysis with anti-PLCζ antibody. Antibody-treated sperm supernatants were also analysed for Ca²⁺ release activity by fluo-3 fluorometry with sea urchin egg homogenates, as described above, and for ability to generate Ca²⁺ oscillations by microinjection into mouse eggs, as described below. Maximal immunodepletion of the sperm PLCζ protein was achieved by using an optimised ratio of antibody beads to sperm extract for each experiment (n=4). The optimal ratio was empirically determined for each sperm extract preparation as the minimum concentration of sperm extract (0.3-0.8 mg/ml) that still retains Ca²⁺ release activity after treatment with the control IgG beads.

Mouse egg procedures were carried out either in Hepes-buffered KSOM or amino acid supplemented KSOM (Summers et al., 2000). Female MF1 mice were superovulated by injection with 5 IU of PMSG followed 48 hours later by HCG (Intervet). Eggs were collected 13.5-14.5 hours after HCG, maintained in 100 μl droplets of H-KSOM under mineral oil at 37°C and cRNA microinjection performed within 1 hour. Expression of Myc-PLCζ in eggs was examined 5 hours after cRNA microinjection, by adding SDS sample buffer to pelleted eggs and incubating at 95°C for 5 minutes prior to SDS-PAGE, immunoblot then densitometric analysis with the Myc monoclonal antibody, as described above. Calibrated mouse sperm pellets were resuspended in 10 mM Tris-HCl pH 7.5, 15 mM dithiothreitol (Perry et al., 1999), then subjected to five freeze-thaw cycles in liquid N₂ and centrifuged at 20,000 g at 4°C for 10 minutes, before densitometric analysis of the soluble extract with PLCζ antibody, as described above. For in vitro fertilisation studies, sperm were capacitated for 2-3 hours before being added to eggs. Egg activation and development studies were in H-KSOM containing 2 μM cytochalasin D for 4 hours. Further development to two-cell stage, morula and blastocyst stage was carried out in 50 μl droplets of KSOM under mineral oil at 37°C in a 5% CO₂ incubator.

Eggs loaded with 4 μM Fura red-AM (Molecular Probes) for 10 minutes were washed in H-KSOM and placed on a Nikon Diaphot stage. Loading media included sulfinpyrazone to prevent dye compartmentalisation and extrusion (Lawrence et al., 1997). cRNA solutions in 120 mM KCl, 20 mM Hepes, pH 7.4, were microinjected to 3-5% of egg volume as previously described (Swann, 1990). Protein synthesis was inhibited in control experiments (Lawrence et al., 1998; Jones et al., 1995) where eggs were preincubated in solution containing 10 μM cycloheximide for 30 minutes before microinjection with PLCζ cRNA (0.02 mg/ml; n=9). Injection volume was estimated from the displacement caused by bolus injection. Ca²⁺ measurements were performed on a CCD-based imaging system as previously described (Lawrence et al., 1997), or a Zeiss Axiovert 100 with illumination from a monochromator (Photonics) controlled by MetaFluor v4.0 (Universal Imaging Corp).

Analysis of the mouse EST database for PLC-related sequences reveals twelve testis-derived expressed sequence tags (ESTs) with identical 3′ ends, apparently from a single mouse testis gene, encoding the C terminus of a putative novel PLC (see Materials and Methods). The putative testis PLC sequence is not found in ESTs from any other tissue. An antiserum raised to an unique peptide antigen deduced from the mouse testis ESTs recognises a single protein band of ∼70 kDa) in immunoblots of mouse, boar and hamster sperm (Fig. 1A). Soluble protein extracts from several other tissues are devoid of this immunoreactivity, suggesting that the ∼70 kDa protein is specifically enriched in sperm (Fig. 1A). Gel filtration chromatography of sperm extracts shows that the immunoreactive sperm protein elutes between the 150 kDa and 29 kDa markers, consistent with a ∼70 kDa monomer in solution (Fig. 1B). Importantly, the ∼70 kDa protein specifically co-migrates with Ca²⁺ release activity in fluorometric assays using egg homogenate (Fig. 1B). This is in contrast to previous chromatographic studies where antibodies to the PLCβ, γ and δ isoforms showed that they did not co-migrate with Ca²⁺ releasing activity (Wu et al., 2001; Parrington et al., 2002). The elution profile further indicates that the ∼70 kDa sperm protein is unrelated to the recently discovered PLCε, which has a molecular mass of ∼250 kDa (Lopez et al., 2001; Song et al., 2001), and therefore it could be a new PLC.

The complete cDNA sequence encoding a novel sperm PLC homologue was cloned from a mouse spermatid cDNA library by PCR using oligonucleotide primers designed with the mouse ESTs identified above. Within the ∼2.2 kb sequence, untranslated regions at the 5′ and 3′ ends, of 194 basepairs (bp) and 52 bp (excluding polyA⁺-tract), respectively, are found flanking a single open reading frame (ORF) of 1941 bp. The ORF encodes a novel protein sequence of 647 amino acids, with a predicted molecular mass of 74 kDa and pI of 5.3 (Fig. 2A). The novel 74 kDa protein includes the C-terminal peptide sequence used to produce the antiserum and is consistent with the native sperm protein of ∼70 kDa detected in immunoblots (Fig. 1A). Blastp sequence analysis suggests that the sperm protein is a novel PLC isoform, smaller than all those previously identified (PLCβ, γ, δ and ε) (Katan and Williams, 1997; Rebecchi and Scarlata, 1998; Rhee, 2001), which we accordingly assign PLCζ.

Clustal alignment of PLCζ with PLCδ1 reveals that the most notable difference of this new isoform is that it lacks an N-terminal PH domain (Fig. 2A,B). A single PH domain is found at the N terminus of all the PLCβ, γ and δ isoforms, and the PH domain of PLCδ1 has been shown to be involved in membrane phospholipid interactions (Katan and Williams, 1997; Rebecchi and Scarlata, 1998; Rhee, 2001). The sequence analysis also indicates that PLCζ possesses the typical X and Y catalytic domains found in all known PLCs (residues 168-307 and 386-502 of PLCζ, respectively). The X and Y domains are between a tandem pair of N-terminal EF hand-like domains and a C-terminal C2 domain (residues 20-150 and 521-625, respectively), both of which are present in most PLCs (Fig. 2B). The X and Y domains of PLCζ contain the PLCδ1 active site residues, corresponding to ¹⁷⁸His, ²¹⁰Asp and ²²³His, that have been shown to be involved in catalysis by site-directed mutation studies of PLCδ1 (Ellis et al., 1993; Ellis et al., 1998) and are conserved across the entire PLC family (Katan, 1998; Rebecchi and Pentyala, 2000). Another distinction between PLCζ and PLCδ1 is the extended X-Y linker sequence in PLCζ (residues 308-385), which has a high proportion of charged residues (Fig. 2A). The X-Y linker region is the only part of PLCδ1 that was not determined in the 3D crystal structure (Williams, 1999). Multiple alignment of PLCζ with the other mammalian PLC isoforms shows that it has the highest degree of similarity with the PLCδ group (33% identity with PLCδ1) and the lowest with PLCε (9% identity; Fig. 2C). The classification of PLCζ as a distinct isoform is supported by phylogeny analysis of the twelve identified mammalian PLCs, which suggests that ζ is the least divergent PLC isoform from a hypothetical precursor, with the rank order ζ<δ<β<ε<γ (Fig. 2D). In accordance with this observation, the domain structure of ζ is similar to plant PLCs that also lack an N-terminal PH domain but retain normal enzymatic properties (Rebecchi and Pentyala, 2000). No plant PLCs with domain structures of the mammalian β, γ, δ or ε isoforms have been identified (Rebecchi and Pentyala, 2000).

Northern blot analysis with mouse tissue mRNAs shows that PLCζ is present as a relatively abundant 2.3 kb transcript only in the testis (Fig. 3A). The transcript abundance is consistent with the significant number of mouse testis ESTs found in the database. The transcript size of 2.3 kb for PLCζ matches the spermatid cDNA clone with a 1941 bp ORF plus ∼300 bp of untranslated sequence (Fig. 2A). The PLCζ transcript distribution also is congruent with immunoblot analysis of a panel of mouse tissues which suggests testis-specificity, as PLCζ protein expression is not detected in any sample other than sperm (Fig. 3B). Sperm cell-specificity of PLCζ expression within testis was examined by performing PCR on cDNA from mouse spermatids and mouse testis devoid of spermatids. PLCζ amplification is observed with spermatid cDNA but not with testis cDNA devoid of spermatids (Fig. 3C), suggesting that PLCζ expression within testis is sperm cell-specific. No PLC isoform has previously been found to be sperm specific, although a splice variant of PLCδ4 enriched in testis (Nagano et al., 1999) was shown to be involved in the zona pellucida-induced acrosome reaction (Fukami et al., 2001).

The defining character of the mammalian sperm factor is the ability to elicit Ca²⁺ oscillations that mimic the fertilisation-associated transients displayed by mammalian eggs (Swann, 1990; Fissore et al., 1998). To examine whether sperm PLCζ could trigger such Ca²⁺ oscillations, we introduced PLCζ complementary RNA (cRNA) by microinjection into MII-arrested mouse eggs, as described previously for spermatogenic cell mRNA (Parrington et al., 2000). Eggs microinjected with a pipette concentration of 2 mg/ml PLCζ cRNA, corresponding to <0.1 mg/ml in the egg after a 3-5% injection volume, underwent a prolonged series of Ca²⁺ oscillations that commence within 15-20 minutes (Fig. 4A, top trace). The high oscillation frequency is similar to that observed upon microinjection of concentrated sperm extracts into mouse eggs (Tang et al., 2000). Ca²⁺ oscillations of similar amplitude, but lower frequency, are obtained with a 1000-fold dilution to 0.002 mg/ml PLCζ cRNA (Fig. 4A, middle trace; 0.0001 mg/ml in egg). None of the eggs treated with cycloheximide to block protein synthesis showed any Ca²⁺ transients after PLCζ cRNA-microinjection (0.02 mg/ml, n=9; Fig. 4A, bottom trace). Robust Ca²⁺ oscillations were observed in 100% of the eggs microinjected with the four different PLCζ cRNA concentrations tested, ranging from 0.002-2 mg/ml (Fig. 4B). Importantly, the frequency, but not the amplitude, of Ca²⁺ oscillations varies with PLCζ cRNA concentration, directly matching the same phenomenon observed with different concentrations of sperm extract (Swann, 1990). The highest pipette concentration used, 2 mg/ml, produces Ca²⁺ oscillations with a mean interspike interval of 7.3±3.2 minutes (Fig. 4B). The lowest pipette concentration of PLCζ cRNA that gives oscillations within 2 hours of injection (0.002 mg/ml), displayed a mean interspike interval of 20.1±5.4 minutes (Fig. 4B). Both of these values are significantly different to the mean interspike interval produced with in vitro fertilisation (IVF) of mouse eggs (12.1±5.8 minutes). However, the interspike intervals for 0.2 and 0.02 mg/ml PLCζ cRNA (13.6±3.2 and 12.7±6.0 minutes, respectively) are not significantly different from IVF (Fig. 4B).

The Ca²⁺ oscillations at fertilisation (Cuthbertson and Cobbold, 1985) display some unique features. The first Ca²⁺ transient invariably lasts longer than subsequent oscillations (Fig. 5A), and exhibits a set of intriguing, smaller sinusoidal increases on top of the main peak (Fig. 5A-I). Microinjection of a pipette concentration of PLCζ cRNA that produces an interspike interval matching IVF (i.e. 0.02 mg/ml; Fig. 4B) results not only in the same, longer initial Ca²⁺ transient, but also displays a similar pattern of smaller sinusoidal increases (Fig. 5B and Fig. 5B-I). The first Ca²⁺ increase after 0.02 mg/ml PLCζ cRNA microinjection matches the first IVF transient in both average duration (PLCζ 2.8±0.6 minutes, n=39 versus IVF 3.0±0.7 minutes, n=16), and also in reproducibly producing the cluster of smaller Ca²⁺ increases superimposed on the first transient (Fig. 5B-I). A concentration of 0.02 mg/ml PLCζ cRNA was used for subsequent microinjection experiments, unless stated otherwise, to provide the precise Ca²⁺ signalling conditions that are stereotypical of fertilisation.

The ability of sperm PLCζ to initiate Ca²⁺ oscillations in eggs is specific to this novel PLC isoform because microinjecting a PLCδ cRNA (2 mg/ml), structurally the most similar mammalian isoform to PLCζ (Fig. 2B), does not trigger a Ca²⁺ increase in any of the 14 eggs tested (Fig. 5C). The lack of effect of PLCδ1 cRNA is consistent with the inability of microinjected PLCδ1 protein to cause any Ca²⁺ changes in mouse eggs (Jones et al., 2000). As the lack of an N-terminal PH domain is the most distinctive difference between PLCζ and PLCδ1 (Fig. 2B), the function of PLCζ could possibly be mimicked by a truncated PLCδ1 without the PH domain. Therefore, a deletion construct of PLCδ1 minus the N-terminal 132 residue PH domain (^ΔPHPLCδ1) was prepared, resembling the domain structure of sperm PLCζ. Microinjection of ^ΔPHPLCδ1 cRNA into eggs does not result in any detectable Ca²⁺ changes (Fig. 5D, 2 mg/ml, n=12 eggs), suggesting that additional factors unique to sperm PLCζ are crucial for Ca²⁺ mobilisation in mammalian eggs. To determine whether catalytically active sperm PLCζ is required for Ca²⁺ mobilisation in the egg, ζ cRNA with a mutation at ²¹⁰Asp, a putative active site residue critical for PLCζ enzyme function (Katan, 1998; Williams, 1999; Rebecchi and Pentyala, 2000), was microinjected. Mutation of the corresponding residue in PLCδ1, ³⁴³Asp to ³⁴³Arg, was shown to be the most severe of numerous site-directed alterations, causing a 180,000-fold reduction in PIP₂-mediated hydrolysis of PLCδ1 (Ellis et al., 1998). The microinjection of ^D210RPLCζ cRNA into eggs, even at high concentration (2 mg/ml), does not produce any Ca²⁺ increase (Fig. 5E, n=22). This suggests that ²¹⁰Asp is crucial for PLCζ enzyme activity, and that IP₃ production is necessary for Ca²⁺ release to occur in eggs (Miyazaki et al., 1983; Stricker, 1999; Brind et al., 2000; Jellerette et al., 2000). The four different cRNAs used in this study, PLCζ, ^D210RPLCζ, ^ΔPHPLCδ1 and PLCδ1, were also expressed in vitro in rabbit reticulocyte lysates, illustrating that they are correctly synthesised and yield the predicted protein sizes of 74, 74, 70 and 85 kDa, respectively (Fig. 5F).

The sperm factor hypothesis predicts that a single sperm contains sufficient activating factor to initiate Ca²⁺ release upon sperm-egg fusion (Swann, 1990; Stricker, 1999). The observation of sperm PLCζ cRNA triggering fertilisation-like Ca²⁺ oscillations in eggs (Figs 4 and 5) is of physiological significance only if the PLCζ protein expressed in a single egg is similar to the native PLCζ present in a single sperm. In order to quantitate the PLCζ expressed in microinjected eggs, a Myc epitope tag was introduced at the N terminus of PLCζ (Lopez et al., 2001). Microinjected Myc-PLCζ cRNA at different concentrations is as effective at generating Ca²⁺ oscillations in eggs (Fig. 6A, 0.02 mg/ml) as the untagged PLCζ (Fig. 4B), indicating that the N-terminal attachment of the Myc tag is not deleterious to PLCζ activity, as was shown for Myc-PLCε (Lopez et al., 2001). Furthermore, the Myc-PLCζ protein expressed in eggs is readily detected in immunoblots using an anti-Myc monoclonal antibody, as a single band with the predicted mass of 78 kDa, whereas uninjected eggs exhibit no immunoreactivity (Fig. 6B). Comparison of the relative mobility of native mouse sperm PLCζ (Fig. 6c, 74 kDa) and recombinant Myc-PLCζ protein [Fig. 6C, 78 kDa (74 kDa PLCζ + 4 kDa Myc tag)] indicates that the deduced ORF of the PLCζ cDNA clone (Fig. 2A, 74 kDa) represents the complete sperm PLCζ sequence. Densitometric analysis of the immunoreactive 78 kDa Myc-PLCζ protein expressed in eggs (Fig. 6C, 100 eggs microinjected with each Myc-PLCζ cRNA concentration), compared with calibrated amounts of purified recombinant Myc-PLCζ protein produced in bacteria, enabled the determination of 44-75 fg/egg (n=4) as the amount of PLCζ protein that triggers Ca²⁺ oscillations using 0.02 mg/ml cRNA (see Materials and Methods). This cRNA concentration is the one that most closely mimics the IVF response, though tenfold lower levels (i.e. 4-8 fg PLCζ protein/egg using 0.002 mg/ml cRNA) are also able to cause Ca²⁺ oscillations (Fig. 4).

The PLCζ content of sperm was also determined by densitometry with a PLCζ polyclonal antibody using a defined number of mouse sperm and compared with calibrated amounts of recombinant PLCζ protein (Fig. 6D). Using densitometric values within the recombinant PLCζ protein calibration plot, obtained from samples comprising 10⁴-10⁶ mouse sperm, a single mouse sperm was calculated to contain 20-50 fg PLCζ protein (n=4). The level of PLCζ able to produce Ca²⁺ oscillations in a single egg similar to fertilisation (4-75 fg, i.e. with 0.002-0.02 mg/ml cRNA) is therefore in the same range as the single sperm content of PLCζ (20-50 fg). The observed quantitative correlation indicates that the PLCζ from a single sperm is sufficient to produce the Ca²⁺ oscillations observed upon sperm-egg fusion.

The features of sperm PLCζ at the functional (Figs 4 and 5) and quantitative (Fig. 6) level are fully consistent with characteristics observed for the sperm factor present in mammalian sperm extracts (Swann, 1990; Berrie et al., 1996; Wu et al., 1997; Stricker, 1997; Wu et al., 1998; Jones et al., 1998b; Kyozuka et al., 1998; Tang et al., 2000). However, it remains possible that sperm components other than PLCζ are also involved in causing Ca²⁺ release in eggs. To address whether the PLCζ in sperm is uniquely responsible for Ca²⁺ mobilisation in eggs, the PLCζ content of sperm extracts was specifically depleted using an anti-PLCζ antibody. Immunoblot analysis indicates that sperm extract supernatant retains the PLCζ protein after control antibody treatment, in contrast to PLCζ antibody-treated supernatant where the PLCζ is absent (Fig. 7A, S– and S+, respectively). Analysis of the corresponding precipitated antibody samples reveals that the sperm PLCζ is effectively removed by PLCζ antibody, but not by the control antibody (Fig. 7A, P+ and P–, respectively). Assessment of Ca²⁺ release activity in antibody-treated sperm extracts using sea urchin egg homogenate assays shows that PLCζ-depleted samples lack any Ca²⁺ mobilising activity, whereas a robust Ca²⁺ release is observed with the control antibody-treated sperm extract containing PLCζ protein (Fig. 7B, S+ and S–, respectively). Moreover, microinjection of antibody-treated sperm extracts into mouse eggs illustrates that the ability of untreated samples to generate IVF-like Ca²⁺ oscillations (Fig. 7C, top trace) is fully preserved in control antibody-treated samples (Fig. 7C, second trace, n=13), while PLCζ-depletion effectively abrogates Ca²⁺ release activity (Fig. 7C, bottom two traces, n=13). These PLCζ antibody depletion experiments (n=4) suggest that PLCζ is the sole component of sperm extracts possessing the ability to cause Ca²⁺ release in mouse eggs. Taken together with evidence that the PLCζ level in a single mouse sperm is sufficient to trigger IVF-like Ca²⁺ oscillations in a single mouse egg (Figs 4,Fig. 5-6), the immunodepletion data provides compelling evidence that PLCζ is synonymous with the previously described mammalian sperm factor (Swann, 1990; Berrie et al., 1996; Wu et al., 1997; Stricker, 1997; Wu et al., 1998; Jones et al., 1998b; Kyozuka et al., 1998; Tang et al., 2000).

The activation of mammalian eggs is caused by sperm-induced Ca²⁺ oscillations at fertilisation (Cuthbertson and Cobbold, 1985; Kline and Kline, 1992). Microinjection of sperm extract into eggs also produces activation and the consequent cellular processes leading to embryo development (Stice and Robl, 1990; Fissore et al., 1998; Sakurai et al., 1999). The isolated sperm factor molecule therefore is predicted to support embryo development after egg activation, providing a crucial test for any putative sperm factor candidate (Fissore et al., 1998). Because eggs that are microinjected with PLCζ cRNA (0.02mg/ml) display all the properties of Ca²⁺ oscillations indistinguishable from those of IVF (Fig. 4B, Fig. 5B) and is equivalent to the PLCζ content of a single sperm (Fig. 6), their ongoing development was monitored for several days after PLCζ-microinjection. PLCζ-microinjected eggs underwent activation (Fig. 8A) because normal development proceeds to the two-cell stage within 24 hours (78%, n=147), and many reach the morula or blastocyst stages by 4-5 days (62%, n=76). None of the eggs microinjected with buffer control reach the two-cell stage, indicating activation as an artefact of microinjection procedure has not occurred (data not shown). The proportion of PLCζ-induced embryos that develop to either the two-cell, or morula and blastocyst stages, is the same as for eggs that are either parthenogenetically activated (Bos-Mikich et al., 1997) by strontium ions (n=75), or when embryos are collected at the one-cell stage from female mice after in vivo fertilisation (n=101) upon mating with males (Fig. 8A). Photomicrographs taken at 24 hours and 5 days after PLCζ-microinjection into mouse eggs show the appearance of normal embryo development to the two-cell stage and blastocyst stage (left and right panel, respectively, Fig. 8B). There are no morphological differences to embryos obtained after fertilisation with sperm (data not shown). Thus, after inducing Ca²⁺ oscillations in the egg, sperm PLCζ-microinjection also triggers the entire cascade of events required for activation and embryo development, in the same manner as sperm at fertilisation.

The possibility remains that a novel action of PLCζ other than PIP₂ hydrolysis is responsible for egg activation, such as a protein-protein interaction with a distinct egg molecule. To test whether an enzymatically active PLCζ is required for egg activation and embryo development, the ^D210RPLCζ cRNA (0.02mg/ml), which has been shown to be defective in triggering Ca²⁺ oscillations (Fig. 5E) (Ellis et al., 1998; Katan, 1998; Williams, 1999; Rebecchi and Pentyala, 2000), was microinjected and egg activation assessed after 24 hours. None of the ^D210RPLCζ cRNA-microinjected eggs were found to proceed to the pronuclear or two-cell stage (Fig. 8C, n=20), suggesting that the enzymatic function of sperm PLCζ is crucial for egg activation.

Cytoplasmic oscillations in intracellular free Ca²⁺ is a remarkable signalling phenomenon observed in many cell types that can regulate a wide variety of physiological processes (Berridge et al., 2000). However, as the original observation of Ca²⁺ oscillations at mammalian fertilisation (Cuthbertson and Cobbold, 1985), the molecular mechanism has remained an enigma. A popular model of Ca²⁺ signalling at fertilisation involves a sperm surface ligand interacting with a receptor on the egg plasma membrane. The ligand-bound membrane receptor couples with an egg PLC to stimulate IP₃ production and Ca²⁺ release, analogous to the signalling pathway found ubiquitously in somatic cells (Berridge et al., 2000). However, the egg and sperm molecules required for the operation of this ‘receptor’ model have not been identified despite extensive studies (Stricker, 1999).

The second major hypothesis involves a sperm cytosolic protein that enters the egg and causes Ca²⁺ release (Swann, 1996; Yamamoto et al., 2001). This ‘sperm factor’ model, though hindered by initial quandaries (Parrington et al., 1996; Sette et al., 1997), has gained increasing credence due to the numerous studies demonstrating the potency of sperm extracts in effecting Ca²⁺ release in eggs (Stice and Robl, 1990; Swann, 1990; Nakano et al., 1997; Stricker, 1997; Wu et al., 1997; Jones et al., 1998b; Kyozuka et al., 1998; Wu et al., 1998; Parrington et al., 1999; Dong et al., 2000; Jones et al., 2000; Rice et al., 2000; Tang et al., 2000; Wu et al., 2001; Yamamoto et al., 2001; Parrington et al., 2002). Moreover, the sperm factor model is congruous with the amount of activity contained in a single sperm (Nixon et al., 2000) and is further supported by the technique of intracytoplasmic sperm injection (ICSI), a clinically effective IVF procedure that has produced thousands of live births (Bonduelle et al., 1999). In the ICSI method, which bypasses the possibility of sperm-egg membrane interaction, a single spermatozoa is injected directly into a human egg to cause Ca²⁺ oscillations, activation and development to term (Bonduelle et al., 1999; Yanagida et al., 2001). Interestingly, the ICSI practice of breaking off the sperm tail to enhance the rate of egg activation (Dozortsev et al., 1995; Yanagida et al., 2001) could be explained by the facilitated release of sperm cytosolic contents, including the sperm factor.

These two major models for Ca²⁺ signalling at fertilisation have developed an overlap following recent observations indicating that a PLC activity is indeed involved in triggering Ca²⁺ oscillations, but the PLC is in the sperm, not the egg (Jones et al., 1998b; Rice et al., 2000; Wu et al., 2001; Parrington et al., 2002). However, comprehensive analysis of known PLC isoforms by different approaches have all concluded that none of them could be the sperm factor (Mehlmann et al., 1998; Jones et al., 2000; Fukami et al., 2001; Mehlmann et al., 2001; Wu et al., 2001; Parrington et al., 2002). There remains the possibility of an undiscovered sperm PLC with the requisite Ca²⁺ signalling properties, and this was directly addressed in the present study.

The revelation of abundant testis-derived ESTs with PLC homology led to our characterisation of a novel sperm PLC isoform (Fig. 1). The new isoform, PLCζ, is the smallest PLC identified to date, most closely resembling the PLCδ class, but without an N-terminal PH domain and a longer X-Y domain linker sequence (Fig. 2). The tissue transcript and protein expression profile indicates sperm-specific enrichment of the PLCζ protein, consistent with a gamete-specific role (Fig. 3). Functional analysis by expression in mammalian eggs provides exquisite evidence that PLCζ possesses the mandatory properties of the sperm factor. PLCζ exhibits the unique ability to produce Ca²⁺ oscillations with the characteristic interspike interval (Fig. 4), and the intriguing, first transient profile-specificity (Fig. 5), found in Ca²⁺ signalling at fertilisation. The inability of the PH-domain-deleted PLCδ1 to mimic fertilisation Ca²⁺ transients, suggests an exclusive functional specificity for the sperm PLCζ domains inside mammalian eggs (Fig. 5). Similarly, the functionally ineffective PLCζ with a catalytic site mutation (Fig. 5) is consistent with the previously shown vital role of a PLC and IP₃ production in mobilising Ca²⁺ in eggs (Miyazaki et al., 1993; Brind et al., 2000; Jellerette et al., 2000). Quantitative correlation of the PLCζ level that produces an IVF-like Ca²⁺ response with that found in a single sperm (Fig. 6), together with demonstration of the unique role of the PLCζ within sperm extracts in effecting Ca²⁺ release in eggs (Fig. 7), directly support the tenet that sperm PLCζ has a physiologically relevant role in egg activation. Furthermore, the normal development of PLCζ-microinjected eggs to the blastocyst stage (Fig. 8) shows that Ca²⁺ oscillations, which are triggered solely via PLCζ, are both necessary and sufficient to initiate the entire network of cellular processes that operate from egg activation through early embryo development to blastocyst. These decisive features of PLCζ argue that it is an important component of the augured mammalian sperm factor and also that there is a physiological role for PLCζ in egg activation and embryo development during mammalian fertilisation.

Discovery of PLCζ as a novel mediator of intracellular Ca²⁺ regulation will enable an increased understanding of the propagation mechanism of large amplitude, low-frequency cytosolic Ca²⁺ oscillations (Berridge et al., 2000). Identification of PLCζ as a component of the putative physiological sperm factor should help to reveal the molecular mechanisms involved in subsequent stages of embryo development after egg activation. Analysis of human sperm PLCζ may also provide a new framework for understanding some cases of male factor infertility where the sperm are ineffective in stimulating development (Rybouchkin et al., 1996; Battaglia et al., 1997). Finally, PLCζ could be applied in approaches to improve egg activation rates, for example, after somatic cell nuclear transfer into enucleated eggs, in the production of stem cells for therapy of human diseases (Aldhous, 2001).

This work was supported by a SIF grant to F. A. L. from the University of Wales College of Medicine. K. S. holds a Wellcome Trust grant and J. P. is an MRC senior fellow. The mouse spermatid and mouse testis devoid of spermatids cDNA libraries were kindly provided by P. Burgoyne and the PLCδ1 plasmid by M. Katan. We are grateful for the advice and encouragement of M. J. Berridge, L. K. Borysiewicz and D. R. Trentham.