A sperm-induced intracellular Ca2+ signal([Ca2+]i) underlies the initiation of embryo development in most species studied to date. The inositol 1,4,5 trisphosphate receptor type 1 (IP3R1) in mammals, or its homologue in other species, is thought to mediate the majority of this Ca2+ release. IP3R1-mediated Ca2+ release is regulated during oocyte maturation such that it reaches maximal effectiveness at the time of fertilization, which, in mammalian eggs, occurs at the metaphase stage of the second meiosis (MII). Consistent with this, the [Ca2+]ioscillations associated with fertilization in these species occur most prominently during the MII stage. In this study, we have examined the molecular underpinnings of IP3R1 function in eggs. Using mouse and Xenopus eggs, we show that IP3R1 is phosphorylated during both maturation and the first cell cycle at a MPM2-detectable epitope(s),which is known to be a target of kinases controlling the cell cycle. In vitro phosphorylation studies reveal that MAPK/ERK2, one of the M-phase kinases,phosphorylates IP3R1 at at least one highly conserved site, and that its mutation abrogates IP3R1 phosphorylation in this domain. Our studies also found that activation of the MAPK/ERK pathway is required for the IP3R1 MPM2 reactivity observed in mouse eggs, and that eggs deprived of the MAPK/ERK pathway during maturation fail to mount normal[Ca2+]i oscillations in response to agonists and show compromised IP3R1 function. These findings identify IP3R1 phosphorylation by M-phase kinases as a regulatory mechanism of IP3R1 function in eggs that serves to optimize[Ca2+]i release at fertilization.
In preparation for fertilization, ovulated mammalian eggs are arrested at the metaphase stage of the second meiosis (MII). Sperm entry triggers a series of increases in the intracellular free-Ca2+ concentration([Ca2+]), termed [Ca2+i]ioscillations, which enable exit from the MII arrest and induce egg activation(Miyazaki et al., 1993). Egg activation entails the orderly initiation of several events, including cortical granule exocytosis, extrusion of the second polar body (2PB),pronuclear (PN) formation and progression into interphase, the completion of which is required for the normal initiation of development(Ducibella et al., 2002; Schultz and Kopf, 1995).
Fertilization-associated [Ca2+]i responses are known to involve stimulation of the phosphoinositide pathway, whereby a phospholipase C (PLC) enzyme hydrolyzes phosphatidylinositol 4,5-bisphosphate[PIP2; PtdIns(4,5)P2] to produce inositol 1,4,5-trisphosphate [IP3; Ins(1,4,5)P3] and diacylglycerol (DAG) (Halet et al.,2004; Stith et al.,1993; Turner et al.,1984). In mammals, a recently identified sperm-specific PLC isoform, PLCζ, is thought to represent the sperm factor responsible for the sustained [Ca2+]i oscillations and production of IP3 (Kouchi et al.,2004; Saunders et al.,2002). IP3 promotes Ca2+ release by binding to the IP3 receptors (IP3Rs), which are predominantly located on the endoplasmic reticulum (ER) - the main reservoir of intracellular Ca2+ (Berridge,2002; Berridge et al.,2000). IP3Rs are thought to mediate all of the intracellular Ca2+ release required to induce egg activation(Miyazaki et al., 1992; Runft et al., 1999).
IP3Rs are well suited to mediate the highly specialized spatiotemporal patterns of [Ca2+]i responses that underlie fertilization. For example, IP3Rs function as tetramers and each monomer consists of three functional domains: an N-terminal domain that contains the IP3-binding site; a modulatory domain; and a C-terminal domain that includes the channel region of the molecule(Bezprozvanny, 2005; Bosanac et al., 2004; Patel et al., 1999). In addition, the entire sequence of IP3Rs is lined with highly conserved consensus sites for interacting partner proteins that are likely to influence the properties of the receptor, such as affinity and/or conductivity(Patterson et al., 2004b), its location and distribution (Vermassen et al., 2004b).
Mammalian oocytes and eggs almost exclusively and abundantly express the type 1 IP3R isoform (IP3R1)(He et al., 1999; Parrington et al., 1998);oocytes of other vertebrates and invertebrates also express a single IP3R isoform that is closely related to IP3R1(Iwasaki et al., 2002; Kume et al., 1993). The most conspicuous mode of regulation of IP3R1 function in these cells is associated with maturation and cell-cycle transitions. For example,Ca2+ release through IP3R1 is greatly enhanced after the initiation of oocyte maturation (Chiba et al., 1990; Fujiwara et al.,1993; Mehlmann and Kline,1994) and, in vertebrate eggs, maximal IP3R1-mediated Ca2+ release is closely timed to coincide with sperm entry, which takes place at one of the two M-phase stages of meiosis according to the species (Fujiwara et al.,1993; Kume et al.,1993; McDougall and Levasseur,1998). Likewise, exit from M-phase and progression into interphase is associated with attenuation and cessation of the[Ca2+]i responses(Jones et al., 1995; Parrington et al., 1998),which is accompanied by a pronounced loss of IP3R1 function(FitzHarris et al., 2003; Jones and Whittingham, 1996). Given that during maturation and after activation/fertilization the changes in IP3R1 concentrations and content of the Ca2+ stores are small (Brind et al., 2000; Iwasaki et al., 2002; Jellerette et al., 2000), it is likely that other mechanisms might regulate IP3R1 function in eggs.
Phosphorylation has been shown to be an important regulatory mechanism of IP3R1 function (Bezprozvanny,2005; Patterson et al.,2004a). Among the protein kinases that phosphorylate IP3R1 are: protein kinase A and protein kinase C(Ferris et al., 1991; Vermassen et al., 2004a);protein kinase G (Koga et al.,1994); Ca2+/calmodulin-dependent protein kinase II(Ferris et al., 1991; Zhu et al., 1996); the tyrosine kinases Fyn (Jayaraman et al.,1996) and Lyn (Yokoyama et al., 2002); Rho kinase(Singleton and Bourguignon,2002); and, very recently, protein kinase B(Khan et al., 2006) (V.V.,H.D.S. and J.B.P., unpublished). In most cases, IP3R1 phosphorylation by these kinases enhances Ca2+ conductivity, but none of these kinases appears to be intimately associated with cell-cycle transitions. Most importantly, abrogation of their activities by pharmacological inhibitors does not affect IP3R1 function in eggs(Carroll and Swann, 1992; Smyth et al., 2002; Swann et al., 1989). However,a recent report has shown in vitro and in vivo IP3R1 phosphorylation at several highly conserved consensus sites by Cdc2/cyclin B[also known as maturation promoting factor (MPF)], the kinase responsible for promoting resumption of meiosis and cell-cycle transitions(Malathi et al., 2003). Furthermore, Cdc2/cyclin B is, together with mitogen-activated protein kinase(MAPK), responsible for arresting vertebrate eggs at the MII stage(Masui and Markert, 1971). It is presently not known whether Cdc2/cyclin B phosphorylates IP3R1 in oocytes and eggs, but our findings demonstrating cell-cycle stage-specific IP3R1 phosphorylation in mouse eggs is in keeping with the hypothesis that receptor phosphorylation underlies, at least in part, the entrainment of the cell cycle with [Ca2+]i responses in eggs and embryos (Jellerette et al.,2004). We detected IP3R1 phosphorylation using the MPM2 antibody, which recognizes a large group of phospho-proteins active at mitosis(Davis et al., 1983; Westendorf et al., 1994). However, it is not known what kinase(s) is responsible for this phosphorylation, at which site(s) or domain(s) this modification takes place,or whether it affects the ability of mouse eggs to mount[Ca2+]i oscillations.
In this study, we have analyzed IP3R1 phosphorylation in oocytes, eggs and zygotes. We show that IP3R1 MPM2 immunoreactivity is first detected during the early stages of oocyte maturation and decreases after fertilization, immediately preceding PN formation. We report the presence of a highly conserved MAPK phosphorylation consensus site within the IP3-binding domain of IP3R1 and show in vitro phosphorylation of this site by MAPK. Finally, we reveal that the MAPK signaling pathway is required for MPM2-detectable IP3R1 phosphorylation in vivo and that abrogation of this pathway impairs the oscillatory activity of mouse eggs.
MATERIALS AND METHODS
Recovery of oocytes/eggs and culture conditions
Germinal vesicle (GV) oocytes and MII eggs were collected from the ovaries and oviducts of 6- to 8-week-old CD-1 female mice, respectively. Females were superstimulated with an injection of 5 IU PMSG (Sigma, St Louis, MO; all chemicals were from Sigma unless otherwise specified). GV oocytes were recovered 40 hours post-PMSG in HEPES-buffered Tyrode-Lactate solution(TL-HEPES) supplemented with 5% heat-treated fetal calf serum (FCS; Gibco,Grand Island, NY) and 100 μM 3-isobutyl-1-methylxanthine (IBMX). GV oocytes were matured for 12-16 hours in Chatot, Ziomek, and Bavister (CZB) medium(Chatot et al., 1989)containing 3 mg/ml bovine serum albumin (BSA) under paraffin oil at 36.5°C and in a humidified atmosphere containing 6% CO2. In vivo MII eggs were recovered 12-14 hours after injection of 5 IU hCG, which was administered 45-48 hours after PMSG stimulation, and, after the removal of cumulus cells,were transferred into 50 μl drops of KSOM (Potassium Simplex Optimized Medium; Specialty Media, Phillipsburg, NJ) and cultured as above.
Parthenogenetic egg activation
The release of MII arrest and generation of zygotes was accomplished by exposing eggs to Ca2+-free CZB medium supplemented with 10 mM SrCl2 for 2 hours, as described by our laboratory(Jellerette et al., 2000). Activated eggs were transferred to drops of KSOM, cultured as above, and monitored for signs of activation, such as 2PB extrusion and PN formation.
Microinjection of eggs and preparation of PLCζ mRNA
Eggs were microinjected as previously described(Kurokawa et al., 2004). Reagents were diluted in injection buffer [100 mM KCl and 10 mM HEPES (pH 7.0)], loaded into glass micropipettes and delivered by pneumatic pressure(PLI-100 picoinjector, Harvard Apparatus, Cambridge, MA). Each egg received 7-12 pl (∼1-3% of the total volume of the egg). pBluescript containing the full-length coding sequence of mouse PLCζ (a gift from Dr K. Fukami,Tokyo University of Pharmacy and Life Science, Japan) downstream of a T7 promoter was in vitro transcribed using the T7 mMESSAGE mMACHINE Kit (Ambion,Austin, TX), as reported by us (Kurokawa et al., 2004).
Fluorescence recordings and [Ca2+]ideterminations
Measurements of [Ca2+]i were performed using Fura2AM(Molecular Probes, Eugene, OR) as previously reported by our laboratory(Kurokawa et al., 2005). Eggs were monitored in drops of TL-HEPES under mineral oil. Up to ten eggs were monitored simultaneously using the software SimplePCI (C-Imaging System,Cranberry Township, PA), which controls a high-speed filter wheel rotating between excitation wavelengths of 340 and 380 nm. Illumination was provided by a 75 W Xenon arc lamp and the emitted light above 510 nm was collected by a cooled Photometrics SenSys CCD camera (Roper Scientific, Tucson, AZ). Fluorescence ratios of 340/380 nm were obtained every 20-30 seconds.
Histone 1 (H1) and myelin basic protein (MBP) kinase assays and kinase inhibitors
The activities of MPF and MAPK in eggs and zygotes were assayed as described previously (Gordo et al.,2002). Lysates from five eggs were mixed in kinase buffer with a cocktail consisting of ATP, [γ-32P]ATP (Amersham, Arlington Heights, IL), H1 (as a substrate for MPF) and MBP (as a substrate for MAPK). The reaction was allowed to proceed for 30 minutes at room temperature and was terminated by the addition of an equal volume of 2×Laemmli sample buffer(SB) (Laemmli, 1970). Proteins were separated on 15% SDS-polyacrylamide gels, and H1 and MBP phosphorylations visualized by autoradiography. Autoradiographs were scanned and quantified as described for western blotting.
U0126 (Calbiochem, San Diego, CA), a MEK-specific inhibitor, was prepared in dimethyl sulfoxide (DMSO) and routinely used at 25 μM; the inactive analog U0124 was used as a negative control. Roscovitine (Ros; Calbiochem), an inhibitor of Cdk1 and Cdk2, was prepared in DMSO and used at 75 μM for all experiments.
Immunological detection of IP3R1 was carried out using the Rbt03 polyclonal antibody raised against C-terminal amino acids 2735-2749 of mouse IP3R1 (Parys et al.,1995). For detection of IP3R1 and IP3R3, the panspecific antibody Rbt475 was used, the epitope of which (amino acids 127-141 of mouse IP3R1) is conserved between isoforms and across species (Bultynck et al.,2004). The anti-cytI3b-1 antibody, which has amino acids 378-450 as its epitope (Sipma et al.,1999), was used to identify the GST-fusion protein containing domain 2 (amino acids 346-922) of IP3R1. The MPM2 monoclonal antibody (Upstate, Lake Placid, NY), which recognizes an epitope characterized by a phosphorylated serine (S)/threonine (T) followed by a proline (P)residue, and the 16B4 monoclonal antibody (Alexis Biochemicals, Lausen,Switzerland), which recognizes a phosphorylated S followed by either a P or a K residue, were used to ascertain IP3R1 phosphorylation.
Preparation of Xenopus eggs/zygotes lysates and IP3R1 immunoprecipitation
Xenopus eggs were collected from mature females and in vitro fertilized, as per standard protocols. For immunoprecipitation experiments,groups of 25 unfertilized eggs or eggs collected after insemination were frozen on dry ice and solubilized with 500 μl cold embryo solubilization buffer containing 1.0% Triton X-100(Cousin et al., 2000). Cellular debris was pelleted by centrifugation at 4°C and discarded. Supernatants were incubated overnight at 4°C with preimmune serum, Rbt03 antibody or MPM2 antibody, with head-over-head rotation. Incubation of protein A sepharose beads (Amersham) with the immunocomplexes occurred for an additional 3 hours before several washes with PBS. Samples were denatured by the addition of 2×SB and stored at -80°C until western blotting was performed.
Cell lysates from 15 to 100 mouse eggs or 0.5 to 6.0 Xenopus eggs were mixed with 15 μl of 2×SB, boiled and loaded onto NuPAGE Novex 3-8% Tris-Acetate gels (Invitrogen, Carlsbad, CA). After electrophoresis,proteins were transferred onto nitrocellulose membranes (Micron Separations,Westboro, MA). Successive MPM2 and IP3R1 western blotting were performed as described by our laboratory(Jellerette et al., 2004). Membranes were washed and incubated for 1 minute in chemiluminescence reagent(NEN Life Science Products, Boston, MA) and developed according to the manufacturer's instructions. Each nitrocellulose membrane was digitally captured and quantified using an imaging system (Kodak Imaging Station 440 CF,Rochester, NY); quantification was performed in the TIFF files before any rendering was carried out. The intensity of the MPM2 immunoreactive band (also the phosphorylated substrate bands in kinase assays) from MII eggs was arbitrarily given the value of 1 and values in other lanes were expressed relative to this band from MII eggs. Intensities were plotted using Sigma Plot(Jandel Scientific Software, San Rafael, CA). Figures were prepared from the TIFF files using ImageJ software (NIH; http://rsb.info.nih.gov/ij/)and Microsoft Powerpoint.
IP3R1 GST constructs and mutagenesis
For domain analysis we expressed GST-fusion proteins corresponding to the various IP3R1 domains that can be obtained by limited proteolysis(Yoshikawa et al., 1999). The cDNAs encoding domains 1-6 of mouse IP3R1 were amplified by PCR using the full-length mouse IP3R1 cDNA as a template (a kind gift from Dr K. Mikoshiba, Tokyo, Japan) and the primers listed in Table 1. Purified PCR products were ligated into the pGEX-6p2 vector and transformed into E. coliDH5α or Bl21 (DE3). Site-directed mutagenesis was performed using the Quick-Change pointmutation kit (Stratagene, La Jolla, CA, USA). Forward primers were designed according to the manufacturer's recommendation and reverse primers were the complementary sequence of the forward primers. Single mutations were made using pGEX6p2-IP3R1 domain 2 as a template,whereas the double mutation was made using pGEX6p2-IP3R1 domain 2 S421A as template cDNA. GST-fusion proteins were purified as previously described (Bultynck et al.,2001). All constructs were sequenced to confirm mutations and frame.
In vitro phosphorylation of IP3R1 by ERK/MAPK
GST-IP3R1 fragments (0.5 μg) were diluted in phosphorylation buffer supplemented with 30 μM ATP and 20 μCi[γ-32P]ATP (Vermassen et al., 2004a). Full-length mouse IP3R1 and full-length rat IP3R3 (1 μg) were expressed in Sf9 insect cells, purified and diluted in phosphorylation buffer supplemented with 0.18% CHAPS and 0.072%L-α-phosphatidylcholine, as previously described(Vermassen et al., 2004a). Free Ca2+ concentrations in in vitro phosphorylation reactions (40μM Ca2+) were calculated using the CaBuf software (Dr G. Droogmans, K.U. Leuven, Belgium; available at ftp://ftp.cc.kuleuven.ac.be/pub/droogmans/). Reactions were performed at 30°C for 1 hour and initiated by the addition of 250 ng (GST-IP3R1 fragments) or 400 ng (IP3R1 and IP3R3) of the activated ERK2 kinase (Calbiochem). Reactions were stopped by heating the samples for 10 minutes at 70°C in 2×SB. Incorporation of 32P was determined using the Storm 840 PhosphorImager (Molecular Dynamics, Sunnyvale, CA), as previously reported(Vermassen et al., 2004a).
Values from three or more experiments, performed on different batches of eggs or zygotes, were used for the evaluation of statistical significance. Statistical comparisons of the intensity of IP3R1 bands, kinase assays and [Ca2+]i parameters were performed using the Student's t-test or one-way ANOVA and, if differences were observed between groups, comparisons between treatments were carried out by applying the Tukey/Kramer test using the JMP-IN software (SAS Institute, Cary, NC). Differences were considered to be significant when P<0.05. Significance among groups/treatments is denoted in bar graphs by different superscripts (western blots) or by the presence of one or two asterisks(kinase assays).
IP3R1 is differentially phosphorylated in mouse oocytes,eggs and zygotes
Our previous study demonstrated that IP3R1 is phosphorylated in a M-phase stage-specific manner in mouse eggs and interphase zygotes(Jellerette et al., 2004). In the present investigation, we sought to extend those findings by examining IP3R1 MPM2 reactivity in mouse oocytes during maturation and in zygotes throughout the first cell cycle. Western blotting was performed on lysates of oocytes at the GV stage (0 hours of maturation), at the onset of germinal vesicle breakdown (GVBD; 2 hours), at meiosis I (MI; 8 hours) and at MII (IVM/MII; 12 hours). MPM2 immunoreactivity of a protein of ∼270 kDa molecular mass corresponding to the IP3R1(Jellerette et al., 2004) was nearly undetectable at the GV stage, but showed a dramatic increase at GVBD and reached near maximal reactivity at MI. This reactivity remained unchanged until MII (Fig. 1A, upper panel and bar graph). The changes in MPM2 reactivity were not due to changes in IP3R1 mass during the same period of time, as stripping of the same blots followed by re-probing with an anti-IP3R1 antibody revealed no such changes in the IP3R1 signal(Fig. 1A, lower panel). Examination of MPF and MAPK activity using an in vitro kinase assay revealed that, consistent with the notion that MPM2 reactivity is associated with M-phase kinase activity, both MPF and MAPK were at basal levels at the GV stage but their activity had increased by the time of GVBD and was at near peak levels at both MI and MII stages (Fig. 1B).
We next examined IP3R1 MPM2 immunoreactivity during the first cell cycle. Exit from MII arrest was induced by exposure to SrCl2(Kline and Kline, 1992a). Samples were collected at the time of 2PB release (2 hours after activation),at PN formation (6 hours post-activation, the time at which zygotes had reached interphase) and at first mitosis (Mit I), the beginning of which was noted by the breakdown of the PN envelope (PNBD; ∼17 hours post-activation). Surprisingly, despite exit from MII as evidenced by the release of the 2PB, IP3R1 MPM2 immunoreactivity was nearly unchanged 2 hours after activation (Fig. 1C). Conversely, a dramatic reduction of MPM2 reactivity took place as zygotes reached interphase and remained low until the time of Mit I(Fig. 1C, upper panel and bar graph); in three out of five replicates, MPM2 reactivity became nearly undetectable at Mit I. The overlapping IP3R1 reactivity remained unchanged in SrCl2-activated zygotes(Fig. 1C, lower panel). Concomitant evaluation of MPF and MAPK activities in these zygotes revealed the expected rapid inactivation of MPF and protracted loss of MAPK activity during the transition into interphase (Fig. 1D), and the increase in MPF activity, which occurred without a concurrent increase in MAPK activity, as zygotes gained entry into Mit I(Fig. 1D). Altogether, these results show that although IP3R1 acquisition of MPM2 reactivity coincides with the increase of both MPF and MAPK activities during oocyte maturation, the presence and loss of this reactivity during the zygotic cell cycle are more closely associated with MAPK than with MPF.
IP3R1 is differentially phosphorylated in Xenopuseggs and zygotes
To extend our findings to other species, and taking into account that cycling Xenopus egg extracts also show cell cycle-restricted[Ca2+]i responses(Tokmakov et al., 2001), we examined whether IP3R1 phosphorylation in Xenopus eggs exhibited the same association with the cell cycle. Xenopus egg extracts were prepared from unfertilized eggs and from fertilized eggs ∼60 minutes after fertilization, which represented the MII and interphase stages,respectively. The results show that in Xenopus eggs, IP3R1 also undergoes cell cycle-associated phosphorylation, as MPM2 reactivity was observed only in MII extracts (Fig. 2A, upper panel). Once again, IP3R1 immunoreactivity was unchanged (Fig. 2A, lower panel).
To ascertain whether the band of ∼270 kDa recognized by the MPM2 antibody was IP3R1, we performed immunoprecipitation studies using either anti-IP3R1 or MPM2 antibodies followed by blotting with the reciprocal antibody. These experiments were performed with Xenopusegg extracts, as they provide an abundant source of material(Parys et al., 1992). Probing of the material precipitated by either of the antibodies with the alternate antibody recognized a strong immunoreactive band, consistent with the notion that the ∼270 kDa MPM2 immunoreactive band is IP3R1(Fig. 2B, upper panel,IP3R1 blotting; lower panel, MPM2 blotting). Immunoprecipitation with pre-immune serum or beads alone produced negative results. Together,these results demonstrate that IP3R1 in eggs is phosphorylated at a MPM2 consensus site(s) in a M-phase-specific manner.
M-phase kinase phosphorylation motifs in IP3R1
The MPM2 antibody recognizes more than 50 proteins phosphorylated during the M-phase of the cell cycle (Westendorf et al., 1994). Remarkably, although the identity of the majority of proteins phosphorylated at the MPM2 epitope is known, that of the kinases responsible for most of these phosphorylations is not(Che et al., 1997). Likewise,the kinase(s) responsible for MPM2 IP3R1 phosphorylation also remains to be identified. The preferred MPM2 epitope sequence reportedly consists of a hydrophobic amino acid-phosphoserine or phosphothreonine-proline-hydrophobic amino acid or an uncharged amino acid(Che et al., 1997; Ding et al., 1997). Given that MPM2 kinases are likely to be proline-targeted kinases and that the S/T-P motif coincides with the minimal phosphorylation motif for Cdc2/cyclin B and MAPK, we examined the presence of phosphorylation motifs for these kinases in IP3R1 using an `insilico' approach. In a recent manuscript, the presence of two highly conserved Cdc2/cyclin B consensus sites centered on amino acids S421 and T799, which lie in the consensus motif S/T-P-X-K/R (Nigg,1991), of IP3R1 has been reported(Malathi et al., 2003). Whereas S421 was restricted to IP3R1, T799was also present in IP3R3(Malathi et al., 2003). Here,we report that IP3R1 also contains two conserved consensus sequences for MAPK/ERK phosphorylation: P-X-S/T-P(Che et al., 1997), which are centered on S436 and T945, respectively. S436is conserved in IP3R1 from mouse to Drosophila, whereas T945 is conserved only among IP3R1 in vertebrates(Fig. 3). A third site,centered on S1765, is present, but only in the IP3R1 of some vertebrates, such as mouse, rat and bovine, but not in human or in Xenopus. None of these sites is conserved in IP3R2 or IP3R3 (Fig. 3).
In vitro phosphorylation of IP3R1 by MAPK
Whereas cellular MPM2 epitopes have been shown to be phosphorylated by both Cdc2/cyclin B and MAPK/ERK, in vitro phosphorylation studies using Xenopus egg extracts and a 19-residue peptide containing two representative MPM2 epitopes as a substrate, one of which closely matched the IP3R1 S421PLK motif, showed that the MPM2 phosphorylating activity of these extracts could be attributed almost exclusively to MAPK (Che et al.,1997; Kuang and Ashorn,1993). Che et al. also showed that MPM2 epitope recognition by the antibody was enhanced by the presence of a proline residue near the N-terminal end of the MPM2 epitope (Che et al.,1997). Given that PVS436P and PMT945P are highly conserved in IP3R1 and that they both have a proline residue at the -2 position, which renders them stronger MAPK consensus sites, we focused our in vitro IP3R1 phosphorylation studies on this kinase. We first performed in vitro phosphorylation of the full-length IP3R1 and -3 proteins that were expressed and purified from Sf9 insect cells in the presence of [γ-32P] ATP. In vitro phosphorylation of IP3Rs by activated ERK2 showed that, whereas IP3R1 is an adequate substrate for the kinase, IP3R3 is not (Fig. 4A). This finding is consistent with the absence of MAPK motifs in IP3R3. Similar results were obtained with both of the monoclonal antibodies 16B4 and MPM2(not shown). Equal loading of the proteins was confirmed by western blotting of the same membrane using the Rbt475 pan-IP3R antibody(Fig. 4B).
To gain insight into the possible ERK phosphorylation sites on IP3R1, in vitro assays were performed as above, but using GST-fusion proteins of large domains of the IP3R1 that corresponded to the IP3R1 fragments that can be generated in vitro by limited trypsinolysis (Yoshikawa et al.,1999). The cDNA constructs were generated by PCR and together encompassed the complete IP3R1 coding sequence (except for the transmembrane regions; see Table 2). The results show that domains 2 and 4, both of which contain optimal consensus sequences for MAPK phosphorylation, were most strongly phosphorylated by ERK2 (Fig. 4C). Although a higher relative phosphorylation level was observed in domain 4, it is important to realize that the only potential MAPK-phosphorylation site in this domain, S1765, is not conserved between mammals. For that reason, domain 4 was not investigated further.
Instead, we focused on GST-IP3R1 domain 2, as it contained the most highly conserved site for MAPK phosphorylation (PVS436P). Moreover, it is known that MPM2 reactivity is enhanced by the presence of consensus sites in close succession and that S436 is preceded by a phosphorylation consensus site for Cdc2/cyclin B (S421PLK)(Westendorf et al., 1994). We therefore chose to introduce inactivating mutations by site-directed mutagenesis of S421 (S→A) and S436 (S→A) to examine whether or not any of these sites was phosphorylated in vitro by MAPK. The results show that although the S421→A mutation decreased overall phosphorylation of the domain, the S436→A mutation completely abrogated phosphorylation of domain 2 by ERK2(Fig. 4D). Together, these results show that at least one of the predicted MAPK/ERK sites within IP3R1 is phosphorylated by the kinase during in vitro phosphorylation.
MAPK activity is required for MPM2 IP3R1 phosphorylation of mouse eggs
Next, we examined whether MAPK/ERK2 was required for the MPM2-detectable phosphorylation of IP3R1 observed in mouse eggs. Accordingly, to prevent ERK activity in oocytes, we used U0126, a specific pharmacological inhibitor of MEK1 and MEK2, the upstream kinases required for phosphorylation and activation of ERK (Favata et al.,1998). Exposure of the oocytes and eggs of several species to U0126 has already been shown to specifically block activation of the ERK pathway (Philipova et al.,2005; Phillips et al.,2002). In mouse eggs, this exposure produced cellular phenotypes consistent with the abrogation of this pathway by genetic methods(Colledge et al., 1994; Hashimoto et al., 1994). In our studies, oocytes were exposed to 25 μM U0126 or 25 μM U0124 (its inactive analog) prior to and during maturation. Because it is well established that oocytes matured in the absence of MAPK activity, as is the case here, either do not reach MII or progress into a MIII arrest(Araki et al., 1996),maturation was performed in the presence of colcemid, a microtubule disruptor that prevents exit from metaphase (Winston et al., 1995). After maturation, oocytes were examined for MPM2 and IP3R1 immunoreactivity, and for in vitro kinase activity. Our results show that, whereas IP3R1 in oocytes matured under control conditions acquired MPM2 reactivity, IP3R1 reactivity at this epitope was precluded in oocytes matured in the presence of U0126(Fig. 5A). Our data also show that U0126 induced the expected inactivation of MAPK activity, without diminishing MPF activity (Fig. 5B), which is consistent with our findings that U1026 affected neither the rate nor the timing of oocyte maturation (not shown).
Given the above results, we next determined whether untimely activation of MAPK could induce IP3R1 MPM2 reactivity. As the low levels of IP3R1 MPM2 reactivity at the time of PN formation coincide with basal levels of MPF and MAPK activities, we sought to increase MAPK activity at this stage by exposing zygotes to okadaic acid (OA). OA has been shown to induce precocious activation of MAPK and PNBD in mouse zygotes by inhibiting protein phosphatases PP1 and PP2A (Gordo et al., 2002; Moos et al.,1995). Treatment of PN zygotes with 10 μM OA for 60 minutes resulted in a sharp increase of IP3R1 MPM2 immunoreactivity(Fig. 5C). This increase was largely precluded by U0126, but not by roscovitine(Meijer et al., 1997)(Fig. 5C). As expected, OA treatment strongly increased the level of MAPK activity, and the increase was inhibited by U0126 (Fig. 5D). Collectively, these results demonstrate that the acquisition of MPM2 reactivity by IP3R1 at MII and in PN zygotes requires elevated activity of MAPK, but not necessarily of MPF.
MPM2 IP3R1 phosphorylation influences IP3R1 function
We next investigated whether or not abrogation of MAPK activity and loss of MPM2 reactivity influenced the ability of eggs to initiate and maintain[Ca2+]i oscillations. Accordingly, we matured oocytes as above in the presence of U0126 and colcemid and induced[Ca2+]i oscillations 16 hours after initiation of maturation. First, we investigated the effects of U0126 on SrCl2-induced oscillations. We reasoned that, as SrCl2promotes Ca2+ release via IP3R1 without inducing IP3 production (Brind et al.,2000; Jellerette et al.,2000), it would test IP3R1 function in an independent way. In the second approach, we initiated physiological oscillations by injecting PLCζ mRNA, which induces oscillations by triggering IP3 production (Saunders et al., 2002). Our results show that oocyte maturation in the presence of U0126 greatly reduced the ability of both activators to induce[Ca2+]i oscillations. For example, although control oocytes mounted normal [Ca2+]i oscillations in response to SrCl2 stimulation (Fig. 6A, left panel), oscillations in U0126-treated eggs either terminated prematurely or were absent altogether(Fig. 6A, right panel). In addition, in the latter group, the observed rises were of lower amplitude and shorter duration, which is clearly evidenced in the side-by-side comparison of the parameters for the first SrCl2-induced[Ca2+]i rise between the control and U0126-treated eggs. For instance, the amplitude of the first rise, as assessed by the magnitude of the change in the fluorescence ratio of 340/380 nm (F340/F380), was of 1.5±0.06 for control eggs, whereas it was of 0.6±0.03 for U0126-treated eggs. Similarly, the duration of this rise was almost double for control eggs than for U0126-treated eggs (8.1±1.39 min versus 4.3±0.30 min, respectively). Likewise, oscillations initiated by injection of PLCζ mRNA terminated prematurely in U0126-treated oocytes and each of the [Ca2+]i transients observed showed reduced amplitude and duration (Fig. 6B). For example, control eggs injected with 0.1 μg/μl PLCζ mRNA showed an average of 9.1±1.80[Ca2+]i rises in the first 2 hours of monitoring,whereas U0126-treated eggs only showed 1.3±0.45 rises in the same period of time. These effects cannot be attributed to inhibition of mRNA translation, as IP3R1 degradation in these eggs was unaffected by U1026 (not shown).
Although the inhibition of [Ca2+]i oscillations in those eggs by U0126 could strictly be due to the effects of abrogating MAPK activity on IP3R1 function, the possibility cannot be excluded that, unrelated to IP3R1 function, more extensive effects caused by the drug treatment could, at least in part, account for the cessation of oscillations. For example, the Ca2+ content of the stores or the Ca2+ influx required to refill these stores (also known as capacitative Ca2+ entry) could be compromised by lack of MAPK activity, thereby reducing the persistence of the oscillations. We first assessed whether the content of the intracellular Ca2+ stores was affected by maturation in the presence of U0126. To do this, eggs were treated with thapsigargin, a specific inhibitor of the sarcoplasmic/ER Ca2+ATPase pumps (Thastrup et al.,1990) that has been widely used to estimate the Ca2+content of IP3-sensitive stores(Shuttleworth and Thompson,1992; Kline and Kline,1992b). In vitro oocyte maturation was performed as above in the presence of U0126 and colcemid. After 14-16 hours, eggs were placed in medium devoid of external Ca2+ for 30 minutes, after which they were treated with 10 μM thapsigargin. The presence of U0126 did not affect the[Ca2+]i responses elicited by thapsigargin(Fig. 6C), as the mean change in the F340/F380 ratio was of 0.7±0.16 for control eggs and of 0.7±0.14 for U0126-treated eggs. To investigate the effect of reduced MAPK activity during maturation on capacitative Ca2+ entry, we examined whether the increase in [Ca2+]i elicited by the addition of CaCl2 to the extracellular medium after thapsigargin treatment was affected. We found that Ca2+ influx into these eggs was not influenced by maturation in the presence of U0126(Fig. 6C), as evidenced by the similarity in the mean change in the F340/F380 ratio between the control and U0126-treated eggs (0.30±0.13 versus 0.30±0.19, respectively). Collectively, our results are consistent with the model that the MAPK pathway is involved in the regulation of IP3R1 function in eggs.
We have examined whether IP3R1 phosphorylation at the MPM2 epitope, an epitope commonly phosphorylated by M-phase kinases, may underpin the enhanced functional activity of IP3R1 in MII eggs. Our results show that MPM2 IP3R1 phosphorylation is associated with the presence of [Ca2+]i oscillations in mouse eggs and zygotes. Using in vitro assays, we found that MAPK/ERK2 phosphorylates IP3R1 at a conserved consensus site and that MAPK activity is required for in vivo MPM2-detectable IP3R1 phosphorylation. We also observed that abrogation of MPM2 IP3R1 phosphorylation during maturation coincides with the failure of eggs to mount[Ca2+]i oscillations. These results establish an unmistakable molecular link between the cell cycle and the Ca2+releasing machinery.
The role of Ca2+ at fertilization represents, perhaps, the clearest manifestation of cell-cycle regulation by a second messenger, as a sperm-induced Ca2+ response is required to induce cell-cycle progression in all species studied to date(Stricker, 1999; Whitaker and Patel, 1990). An important feature of Ca2+ release during fertilization is that it almost universally unfolds during M-phase stages of the cell cycle(Stricker, 1999; Whitaker, 2006) and that, in those species in which the sperm initiates oscillations, attenuation of[Ca2+]i oscillations coincides with transition into the interphase stages (Kono et al.,1996; McDougall and Levasseur,1998; Stricker and Smythe,2003). Appropriately, both IP3R1 function, as examined by IP3-induced Ca2+ release, and IP3R1 cellular distribution, especially its reorganization into cortical clusters,coincide with these highly oscillatory M-phase stages(FitzHarris et al., 2003; Goud et al., 1999; Jellerette et al., 2004; Kline et al., 1999; Parrington et al., 1998). However, the molecular mechanisms that bring about this enhanced function and organization of IP3R1 are unknown. In the present study, we considerably extend our previous findings(Jellerette et al., 2004) and demonstrate that during maturation and fertilization, IP3R1 undergoes phosphorylation at a MPM2 epitope(s) in concert with the wax and wane of M-phase kinase activities in both mouse and Xenopus eggs and zygotes.
In mammals, where sperm-induced oscillations are long-lasting, other mechanisms besides the role of IP3R1 have been proposed to account for the cell-cycle dependence of [Ca2+]i oscillations(Carroll et al., 2004; Nixon et al., 2000; Nomikos et al., 2005),including the demonstration that PLCζ is sequestered away in the PN,presumably leading to reduced IP3 production(Larman et al., 2004). Nonetheless, evidence in the literature shows that changes in the maternal Ca2+-releasing machinery downstream of IP3 production are also crucial for the association of Ca2+ release with M-phase stages in mammals and other species. For example, it has been shown that uninterrupted administration of IP3 into PN-stage mouse zygotes does not restore oscillations (Jellerette et al., 2004; Jones and Whittingham, 1996). Likewise, in oscillating fertilized zygotes bisected after initiation of oscillations such that one half contains all nuclear structures, [Ca2+]i oscillations cease at approximately the same time in both halves(Day et al., 2000). Lastly,cycling Xenopus egg extracts also show M-phase restricted IP3R1-mediated Ca2+ release, regardless of whether cell-cycle resumption of the extracts was induced by the sperm or by a parthenogenetic agent (Tokmakov et al.,2001). Collectively, evidence suggests that the Ca2+-releasing machinery of eggs undergoes a functional optimization during M-phase stages of the cell cycle; we propose that MPM2 IP3R1 phosphorylation is one of the mechanisms that underlie this optimization.
In this study, we have identified IP3R1 as a novel target for MPM2-detectable phosphorylation during the M-phase stages of the cell cycle. The kinase(s) responsible for IP3R1 phosphorylation in eggs is not yet known, although it is logical to envisage the participation of MPF and MAPK. Cdc2/cyclin B has already been shown to phosphorylate IP3R1 in vitro and in vivo (Li et al.,2005; Malathi et al.,2003). In addition, substrate-binding motifs for this kinase were recently reported in IP3Rs and, consistent with this,immunoprecipitation studies in breast cancer cells have shown that cyclins and IP3R3 can interact (Soghoian et al., 2005). Despite this evidence, attempts to in vitro phosphorylate IP3Rs using starfish oocyte extracts and recombinant Cdc2 kinase have produced negative results(Lim et al., 2003; Santella et al., 2003), which is consistent with our own unpublished results. The presumed role of MPF in MPM2 IP3R1 reactivity is further undermined by our finding that the decline in MPM2 IP3R1 reactivity and MPF activity during egg activation are not synchronous. Moreover, the preferred MPM2 epitope differs greatly from the preferred MPF phosphorylation motif(Holmes and Solomon, 1996). Nonetheless, it is still possible that MPF could actively phosphorylate IP3R1 at a site undetectable by the MPM2 antibody. Future investigations should be pursued with more site-specific antibodies, such as those used by others (Malathi et al.,2003; Soghoian et al.,2005).
Our in vitro data reveal that MAPK phosphorylates IP3R1 but not IP3R3, and that this phosphorylation occurred within the domains that contain consensus sites for MAPK. Moreover, in vitro mutagenesis studies of the most-conserved site, S436, showed that its substitution abolishes the phosphorylation of this domain by MAPK. Although mutation at a nearby conserved MPF site, S421, decreased ERK-mediated phosphorylation within this GST-IP3R1 fragment, it did not eliminate it. However, these results suggest that phosphorylation by one kinase may modify the effectiveness of phosphorylation by the second kinase,which is reminiscent of the effects observed after sequential phosphorylation of IP3R1 by PKA and PKC(Vermassen et al., 2004a). Whether S436 is in vivo phosphorylated by MAPK/ERK and whether it becomes a MPM2 epitope require additional investigation. However, the recent demonstration in the peripheral Golgi protein Nir 2 that a S residue within a S382 PVE site, which is remarkably similar to the S436PAE site in IP3R1, becomes a MPM2 epitope at the onset of mitosis (Litvak et al.,2004) supports the concept that the S436 in IP3R1 may also be actively modified during the MII stage.
Besides the previous demonstration in Xenopus egg extracts that MAPK/ERK was one of the kinases responsible for generating MPM2 reactivity sites (Kuang and Ashorn,1993), our data showing that IP3R1 MPM2 reactivity is not regained in Mit I zygotes, a stage that is devoid of MAPK activity,further implicates this pathway in playing a role in IP3R1 phosphorylation. This association is further strengthened by the finding that MPM2 IP3R1 reactivity was largely prevented in oocytes matured in the presence of the MEK inhibitor U0126. Moreover, ectopic activation of MAPK/ERK in PN-stage zygotes by the addition of OA re-established MPM2 reactivity. OA is known to promote PNBD by activating MAPK/ERK in the absence of MPF/Cdc2/cyclin B (Moos et al.,1995; Moos et al.,1996), although it is likely that other kinases are also activated by OA. Nevertheless, the OA-induced IP3R1 phosphorylation was precluded by U0126, which supports the hypothesis that MAPK/ERK activity is required for IP3R1 MPM2 phosphorylation. However, it is not possible to discern from these studies how MAPK/ERK brings about MPM2 IP3R1 reactivity. For example, it could be by directly phosphorylating IP3R1, which would support our in vitro studies. However, it is also feasible that ERK may activate other downstream kinases that are ultimately responsible for the phosphorylation. Given the pivotal role of MAPK/ERK in the cytoskeletal organization of the oocyte(Lefebvre et al., 2002; Verlhac et al., 2000), it is also plausible that this kinase could control the cellular distribution of IP3R1 and the putative active kinase(s) such that they overlap at MII. Which of these possibilities, or what combination of them, underlies the role of MAPK/ERK on IP3R1 MPM2 reactivity in eggs will require additional investigation.
Abrogation of the MAPK pathway and IP3R1 MPM2 phosphorylation affected the oscillatory capacity of eggs. Our studies reveal that inhibition of IP3R1 MPM2 reactivity by U0126 coincided with eggs showing[Ca2+]i oscillations of shorter duration in response to SrCl2 exposure or to PLCζ mRNA injection. In addition, the duration and amplitude of individual [Ca2+]i rises were severely reduced in U0126-matured eggs. Importantly, the content of intracellular Ca2+ stores and the capacitative Ca2+entry of these eggs appeared to be unaffected. Collectively, these results suggest that the oscillatory activity in general, and IP3R1 function in particular, is compromised in eggs matured in the absence of the MAPK/ERK signaling pathway. A similar observation regarding the role of the MAPK pathway has been reported in sea urchin eggs, where abrogation of the MAPK signaling pathway prevented the [Ca2+]i rise associated with nuclear-envelope breakdown and cell-cycle progression after fertilization (Philipova et al.,2005). These results differ, at least in part, from a recent publication that indicated that acute exposure of MII mouse eggs to U0126 was without consequences on sperm-initiated oscillations despite the reduction of MAPK/ERK activity (Marangos et al.,2003). However, whether or not the phosphorylation status of any of the downstream targets of MAPK/ERK was altered by the U0126 treatment, as demonstrated in our study for IP3R1, was not determined in the aforementioned study.
In summary, we show that the IP3R1 of vertebrate eggs is differentially phosphorylated at a MPM2 site(s) during oocyte maturation and after egg activation. We provide evidence that a M-phase kinase that phosphorylates IP3R1 in vitro, MAPK/ERK, is required for the IP3R1 MPM2-detectable in vivo phosphorylation observed in mouse eggs, and that elimination of the MPM2 reactivity may undermine the function of IP3R1 during fertilization.
This work was supported by grants from the USDA and the NIH/NICHD to R.A.F.; and by grants G.0210.03 of the Fund for Scientific Research -Flanders, by P5/05 of the Interuniversity Attraction Poles Program of the Belgian Science Policy and by 04/07 of the Concerted Actions of the K.U.Leuven to H.D.S. and J.B.P. R.A.F. was a recipient of a visiting postdoctoral fellowship of the Fund for Scientific Research - Flanders. D.A. is supported by NIH grant DE016289. We also acknowledge the excellent technical assistance of I. Willems, L. Bauwens, S. Vangeel, T. Luyten and C. He, and comments on the manuscript by Dr Jeremy Smyth.