Regulator of G-protein signaling 2 (RGS2) suppresses premature calcium release in mouse eggs

Fully grown mammalian oocytes are arrested in meiotic prophase until a mid-cycle release of luteinizing hormone from the pituitary stimulates resumption of meiosis to the metaphase II (MII) stage, referred to herein as eggs. Fusion of the egg and sperm at fertilization introduces into the egg cytoplasm a sperm-specific phospholipase C (PLCζ) (Kashir et al., 2014), which initiates inositol 1,4,5-trisphosphate (IP₃)-mediated Ca²⁺ release from intracellular stores. An initial, prolonged rise in cytoplasmic Ca²⁺ is followed by several hours of repetitive, low-frequency Ca²⁺ oscillations (Runft et al., 2002). These Ca²⁺ oscillations drive the conversion of the egg to an early embryo by causing a series of downstream responses, including resumption of meiosis, cortical granule exocytosis, which prevents polyspermy, and recruitment of maternal mRNAs needed for successful embryonic development (Ducibella and Fissore, 2008).

During meiotic maturation, oocytes undergo several cytoplasmic changes that dramatically increase the ability of mature eggs to release Ca²⁺ in response to sperm or exogenous signals (Fujiwara et al., 1993; Jones et al., 1995; Mehlmann and Kline, 1994). These changes include an approximately fourfold increase in Ca²⁺ stores (Tombes et al., 1992), an increase in IP₃ receptor protein levels by 1.5- to twofold (Fissore et al., 1999; Mehlmann et al., 1996) and reorganization of the endoplasmic reticulum (ER) towards the plasma membrane (FitzHarris et al., 2007; Mehlmann et al., 1995), which places the Ca²⁺ stores proximal to exogenous signals. Priming of the egg for Ca²⁺ release, although needed for proper Ca²⁺ signaling after fertilization, is also associated with the risk of parthenogenetic activation should Ca²⁺ signals occur prior to sperm-egg fusion. Indeed, activation of overexpressed G_q-protein-coupled muscarinic receptors in mouse eggs by exogenous ligands in the absence of sperm causes IP₃-mediated Ca²⁺ release, Ca²⁺ oscillations and parthenogenetic egg activation (Moore et al., 1993; Williams et al., 1998,, 1992). Ca²⁺ signaling prevents subsequent fertilization by inducing cortical granule release, which modifies the zona pellucida (ZP), a proteoglycan-rich extracellular matrix that surrounds the egg, to cause the ‘block to polyspermy’.

Because the mature egg is exceedingly sensitive to stimuli that can activate IP₃-mediated Ca²⁺ release via G-protein-coupled receptors, we reasoned that a mechanism was in place to suppress this response prior to sperm-egg fusion. Regulator of G-protein signaling (RGS) proteins are a ubiquitous family of proteins that downregulate G-protein-coupled receptor signaling by inhibiting the activity of G-proteins (Heximer and Blumer, 2007). In general, RGS proteins stimulate the hydrolysis of GTP bound to activated Gα subunits, leading to signal termination (Willars, 2006). Here, we tested the hypothesis that RGS2, which inhibits both G_q and G_s signaling (Ingi et al., 1998; Roy et al., 2006; Wang et al., 2004), suppresses Ca²⁺ release in mature mouse eggs. We report that RGS2 translation during meiotic maturation serves as a safety mechanism to prevent parthenogenetic egg activation prior to fertilization.

To determine which RGS proteins are expressed in mouse oocytes, we searched an oocyte gene expression database (Evsikov et al., 2006). Among RGS isoforms, RGS2 had the highest expression levels, ∼500 transcripts per million (Blake et al., 2014). Rgs2 mRNA was highly expressed in germinal vesicle (GV)-stage oocytes, decreased during oocyte maturation and was greatly reduced at the 2-cell stage (Fig. 1A). By contrast, RGS2 protein was minimally detected in GV oocytes but increased ∼20-fold during maturation to the MII stage (Fig. 1B,C). These findings indicate that RGS2 is developmentally regulated and suggest that it functions during oocyte maturation or beyond.

To test functional roles of RGS2, we used both overexpression and knockdown approaches. Because G_s activity is crucial for the maintenance of prophase arrest prior to oocyte maturation (Mehlmann et al., 2002,, 2004), and RGS2 can inhibit the activity of G_s (Roy et al., 2006,, 2003; Sinnarajah et al., 2001), we first tested whether altering RGS2 levels affected maturation success. Overexpressing RGS2 in GV oocytes did not stimulate meiotic resumption, and depleting RGS2 in oocytes using RNA interference did not affect the progression of meiosis or meiotic spindle formation (Fig. 2A,B; see supplementary material Table S1 and Fig. S1). These results suggest that RGS2 is not required during oocyte maturation and that these approaches could be used to examine its function in MII eggs.

RGS2 potently suppresses G_q signaling (Ingi et al., 1998; Wang et al., 2004), and, in mouse eggs, activation of G_q in the absence of sperm leads to Ca²⁺-mediated resumption of meiosis and complete egg activation (Moore et al., 1993; Williams et al., 1998). Sperm do not appear to utilize this pathway (Williams et al., 1998) but instead stimulate IP₃ generation directly by introducing PLCζ (Saunders et al., 2002; Knott et al. 2005). However, G_q activation triggers the same downstream IP₃ receptor-mediated Ca²⁺ release that is essential for fertilization. To determine whether RGS2 activity could impact Ca²⁺ signals at fertilization, we tested the effect of RGS2 depletion on Ca²⁺ oscillatory patterns during in vitro fertilization (IVF). We found that RGS2-depleted eggs exhibited normal Ca²⁺ oscillations, with the exception that the duration of the first Ca²⁺ transient was slightly, but consistently, shorter (see supplementary material Fig. S2A-D). This finding suggests that the Ca²⁺ stores were depleted prior to fertilization, which could be explained by either impaired Ca²⁺ store accumulation during maturation or by premature Ca²⁺ release. To distinguish between these possibilities, we analyzed Ca²⁺ stores in control and RGS2-depleted MII eggs by measuring thapsigargin-mediated ER Ca²⁺ release. There was no difference in Ca²⁺ stores between control and RGS2-depleted MII eggs when the ZPs were intact (see supplementary material Fig. S2E-G).

As is typical for IVF experiments, the ZP was removed prior to insemination to promote rapid synchronous fertilization during Ca²⁺ imaging. We noticed that many ZP-free, RGS2-depleted eggs began emitting second polar bodies before sperm addition (Fig. 2C), indicating spontaneous activation. Spontaneous activation was only observed in ZP-free RGS2-depleted eggs, not in ZP-intact eggs (see supplementary material Fig. S3A). Our standard protocol for ZP removal is a brief treatment with acid Tyrode's medium. We therefore tested whether absence of the ZP or the acid exposure was causing the spontaneous activation by using pronase treatment as an alternative method for ZP removal. Pronase-treated eggs did not activate, whereas eggs exposed to acid had high rates of second polar body emission, and most of the activated eggs went on to form pronuclei or to cleave (Fig. 2D; see supplementary material Fig. S3A-D). These findings are consistent with previous observations of acid induction of parthenogenetic activation in mouse and human eggs (Johnson et al., 1990). In addition, spontaneous Ca²⁺ changes were observed in 25% (4/16) of the siRNA-injected cells prior to addition of sperm, but never (0/15) in controls (see supplementary material Fig. S3E). These Ca²⁺ changes could explain the shortened first transients of eggs lacking RGS2, because premature Ca²⁺ release would result in Ca²⁺ store depletion prior to fertilization. To test this idea, we analyzed Ca²⁺ stores in ZP-free control and RGS2-depleted MII eggs. Ca²⁺ stores were significantly lower in RGS2-depleted eggs when the ZPs were removed using acid Tyrode's medium, but not different when the ZPs were removed by manual microdissection (see supplementary material Fig. S3F-H). Taken together, these findings indicate that lack of RGS2 during oocyte maturation does not affect Ca²⁺ accumulation into ER stores, and suggest that acid-induced premature Ca²⁺ release in RGS2-depleted eggs causes a reduction in Ca²⁺ stores and, as a consequence, shortened first Ca²⁺ transients following fertilization.

To characterize the acid sensitivity of RGS2-depleted eggs, we examined the effect of gradually lowering pH on Ca²⁺ release. The majority of control eggs treated with acid did not exhibit increases in Ca²⁺ even at a pH as low as 5 (Fig. 2E,F). However, most RGS2-depleted eggs showed marked increases in Ca²⁺ starting between pH 6.2 and 6.9, suggesting that RGS2 inhibits acid-induced Ca²⁺ release. Similar results were obtained using an Rgs2-targeted morpholino oligonucleotide (Fig. 2G), indicating that this response was not due to a non-specific effect of siRNA.

We also examined the Ca²⁺ response to increased acid in GV-stage oocytes, which have low levels of RGS2 protein, and throughout oocyte maturation. We found that GV- and GVBD-stage oocytes (maturing oocytes that have undergone nuclear envelope breakdown) were very sensitive to lower pH, with virtually all oocytes displaying Ca²⁺ release starting at pH 6.4-6.6 (Fig. 3A,B). Acid sensitivity decreased during maturation such that fewer metaphase I oocytes and MII-stage eggs responded to lower pH with Ca²⁺ release (Fig. 3A,B). This decrease in sensitivity during maturation correlated with increased translation of Rgs2. To examine directly the role of RGS2 in inhibiting acid-induced Ca²⁺ release, we overexpressed RGS2 protein in GV-stage oocytes and tested their response to acid treatment. Control oocytes released Ca²⁺ in response to lowering pH beginning at pH 6.4-6.6. By contrast, RGS2-overexpressing oocytes had a greatly reduced Ca²⁺ response, with fewer cells responding, and the total amount of Ca²⁺ released being much lower when Ca²⁺ release occurred at all (Fig. 3C,D).

In somatic cells, RGS2 inhibits acid-induced responses, such as those induced by alterations in pH that result from inflammatory conditions including asthma (Liu et al., 2013). Studies in airway epithelial cells suggest that lower pH activates the G_q-coupled proton sensor GPR68 (also called OGR1) to release Ca²⁺, which stimulates the production of MUC5AC (Liu et al., 2013; Ludwig et al., 2003; Saxena et al., 2012). RGS2 overexpression prevents acid-induced secretion of MUC5AC, and RGS2 depletion increases this response (Liu et al., 2013). These results indicate that RGS2 acts as an inhibitory regulator of acid-induced cellular responses by binding to G_q. GPR68 or a similar receptor could act as a proton sensor linked to Ca²⁺ signaling by activating G_q in oocytes. Indeed, GV-stage oocytes have significant levels of Gpr68 transcripts compared with those in MII eggs and early embryos (Fig. 3E). It is unclear whether mouse eggs are exposed to low pH under physiological conditions within the ovarian follicle or oviduct, as direct measurements have not been reported. In larger species, follicular fluid and oviduct pH is generally in the range of 7-8 [reviewed by Edwards (1974); Stone and Hamner (1975)], but pH has been measured as low as 6.8 in bovine follicles and pig oviducts (Smiljaković et al., 2008; Zachariae and Jensen, 1958). In addition, follicle or oviductal pH could drop during an inflammatory process as has been observed in inflamed airways (Liu et al., 2013).

In addition to an acidic environment, eggs could be exposed to other stimuli that activate G_q in vivo. One such stimulus is acetylcholine (ACh). Choline acetyltransferase, which synthesizes ACh, is expressed in granulosa cells from large antral follicles of human, monkey and rat ovaries (Fritz et al., 1999,, 2001; Mayerhofer et al., 2006). Moreover, ACh has been measured in cultured human and rat granulosa cells (Fritz et al., 2001) and in adult rat ovaries (Mayerhofer et al., 2006). Interestingly, the amount of ACh in rat granulosa cells is significantly increased by follicle stimulating hormone (Mayerhofer et al., 2006), suggesting that maturing oocytes are exposed to ACh in vivo. ACh treatment stimulates Ca²⁺ release via G_q-coupled muscarinic receptors, which are expressed in oocytes of several species including mouse (Caratsch et al., 1984; Eusebi et al., 1979,, 1984). In addition, stimulation of the muscarinic receptor induces Ca²⁺ release in immature growing mouse oocytes (Carroll et al., 1994) but not in MII eggs (Williams et al., 1998).

We examined the ability of ACh to stimulate Ca²⁺ release in oocytes, eggs and RGS2-depleted eggs. All of the GV-stage oocytes and RGS2-depleted eggs released Ca²⁺ in response to addition of ACh, whereas the majority of control eggs that contained RGS2 showed no response (Fig. 4A,B). Treatment of GV-stage oocytes with the muscarinic receptor antagonist atropine almost completely suppressed ACh-induced Ca²⁺ release (Fig. 4C), but did not affect acid-induced Ca²⁺ release (data not shown). These findings indicate that RGS2 effectively inhibits G_q-mediated Ca²⁺ release in response to ACh, and suggest that one function of the maturation-associated accumulation of RGS2 protein is to suppress this physiological response.

Ca²⁺ release at fertilization triggers cortical granule exocytosis and release of the protease ovastacin, which cleaves the ZP protein ZP2 to ZP2_f (Burkart et al., 2012). This cleavage event is responsible for the ZP block to polyspermy. Because cortical granule release in eggs is easily triggered in response to rises in cytoplasmic Ca²⁺ (Ducibella et al., 2002), ZP2 cleavage can be used as a proxy for Ca²⁺ release over time. To test whether RGS2 regulates Ca²⁺ release during oocyte maturation, we microinjected GV-stage oocytes with Rgs2 siRNA, matured them and then measured the percentage of ZP2-to-ZP2_f conversion. RGS2-depleted eggs consistently had elevated levels of ZP2 conversion compared with controls (Fig. 4D,E). These findings demonstrate that during in vitro maturation of RGS2-depleted oocytes, Ca²⁺ increases occurred that were sufficient to cause cortical granule release and promote ZP2 conversion.

A full RGS2 knockout mouse has been developed and this mouse has at least two defects due to inappropriate Ca²⁺ regulation. RGS2 knockout pancreatic acinar cells have significantly higher IP₃-mediated Ca²⁺ release in response to muscarinic receptor activation (Wang et al., 2004). In addition, these knockout mice have high blood pressure due to increased Ca²⁺ release in response to vasoconstrictors, which act through G_q-coupled receptors. To determine whether loss of RGS2 in vivo resulted in abnormal responses to G-protein-coupled receptor agonists, we tested the effects of ACh and acidic pH on MII eggs from Rgs2^+/+ and Rgs2^−/− females. Consistent with our findings in RGS2-depleted eggs, 100% of the Rgs2^−/− eggs responded to ACh by releasing Ca²⁺ (Fig. 4F). Only 12/26 Rgs2^+/+ eggs showed some degree of response to ACh, but far less Ca²⁺ was released than in Rgs2^−/− eggs (Fig. 4G). Similarly, Rgs2^−/− eggs responded at a higher pH and released more Ca²⁺ than Rgs2^+/+ eggs (Fig. 4H,I). Of note, control eggs in these experiments, which were from C57BL/6J females, were more sensitive to both ACh and acid when compared with the CF-1 eggs used in previous experiments, as demonstrated by the greater proportion showing responses. These findings draw attention to mouse strain differences that probably underlie conflicting reports regarding responses of MII eggs to ACh (Kang et al., 2003; McGuinness et al., 1996; Williams et al., 1992). In fact, the MF-1 mouse strain, which produces eggs that have a high incidence of activation following acid-mediated ZP removal (Johnson et al., 1990), was the subject of a quantitative trait locus analysis that identified the Rgs2 gene locus as a modulator of anxiety (Yalcin et al., 2004). These findings suggest that differences in RGS2 expression or function contribute to the increased sensitivity of MF-1 eggs to ACh and acid activation.

RGS2 knockout mice are viable and fertile, but no formal breeding study has been reported. We analyzed litter sizes from ongoing production of Rgs2^+/+ and Rgs2^−/− mice over a 32-month period. The average litter size of Rgs2^−/− females was significantly lower than that of Rgs2^+/+ females (Fig. 4J). This finding was due to subfertility of the Rgs2^−/− females, because litters of Rgs2^+/+ females mated to Rgs2^−/− males were not smaller than those of wild-type pairs (data not shown). A possible explanation for the reduced litter size was premature cleavage of ZP2 similar to that observed in RGS2-depleted eggs, which could result in impaired fertilization. To test this idea, we collected MII eggs from Rgs2^+/+ and Rgs2^−/− females 16 h after hCG administration and quantified ZP2-to-ZP2_f conversion. Indeed, Rgs2^−/− eggs had increased levels of ZP2 conversion compared with those from Rgs2^+/+ females (Fig. 4K,L). These findings indicate that RGS2 suppresses Ca²⁺ release in vivo. The finding that the mice were not completely infertile suggests that, in addition to RGS2, other mechanisms are in place to help prevent premature Ca²⁺ release.

In conclusion, a dramatic rise in intracellular Ca²⁺ is of paramount importance for successful egg activation and embryo development in all animals (Kashir et al., 2013). Mammalian oocytes do not develop the ability to efficiently release large amounts of Ca²⁺ until immediately before ovulation, thereby preventing premature Ca²⁺ release that could preclude successful fertilization. As the oocyte becomes fertilization-competent, it is exquisitely sensitive to signals that release Ca²⁺ (Fig. 4M). Our findings suggest that RGS2 functions to inhibit premature G-protein-mediated Ca²⁺ release in eggs that are poised for activation by PLCζ from the fertilizing sperm.

All mice used (see details in the supplementary Materials and Methods) were maintained under approved protocols and complied with the Institute of Laboratory Animal Research Guide for the Care and Use of Laboratory Animals. Oocytes, eggs, embryos and sperm were collected and oocytes were microinjected with 5-10 pl volume, essentially as previously described (Bernhardt et al., 2011; Jefferson et al., 2009). Pipette concentrations of microinjected reagents were as follows: Rgs2 siRNA (Ambion) and control nontargeting siRNA (Santa Cruz), 2 μM; Rgs2 and control morpholinos (GeneTools), 2 mM; and HA-RGS2 cRNA, 1 µg/µl. siRNA- and morpholino-injected oocytes were cultured in 10 µM milrinone for 5-8 h to allow protein turnover and then matured. cRNA-injected oocytes were incubated in milrinone-containing medium for 18-20 h prior to imaging to allow for protein overexpression. For some experiments, ZPs were removed by brief exposure to acidic Tyrode's solution, pH 1.6, by incubation for 4-8 min in 5 mg/ml pronase, or by piezoelectric-actuated drilling of a 50-µm slit in the ZP, followed by gentle pipetting using a 70-μm inner diameter capillary. Culture media used are detailed in the supplementary Materials and Methods.

Ca²⁺ imaging was performed as previously described (Miao et al., 2012). For fertilization experiments, ZP-free eggs were adhered to Cell-Tak-coated dishes in BSA-free medium under mineral oil. Capacitated sperm were added to a final concentration of 10⁵ sperm/ml. For acid addition and ACh experiments, ZP-intact cells were used. HCl (1 N) was diluted 1:100, and 150 μl of this solution was added to 2 ml medium several minutes apart during imaging. pH measurements were made using parallel additions of the dilute HCl to the same medium volume. ACh was prepared as a 5 mM stock in water and added to L-15 medium during imaging to achieve final concentrations of 2 µM, 200 µM and 1 mM.

For spindle staining, eggs were fixed, extracted and blocked as previously described (Bernhardt et al., 2012). Oocytes, eggs or embryos were lysed in sample buffer, separated under reducing conditions and immunoblotted as previously described (Jefferson et al., 2013). Details regarding antibodies are provided in the supplementary Materials and Methods.

Total RNA was isolated from 50 oocytes, eggs or embryos using an Arcturus Pico Pure kit (Life Technologies). EGFP cRNA was generated as previously described (Miao et al., 2012) and 10 pg was added to each sample prior to RNA isolation as an internal control. Real-time RT-PCR was performed as previously described (Jefferson et al., 2013), using cDNA from two oocytes or embryos per reaction. Primer sequences are provided in the supplementary Materials and Methods. Relative gene expression was calculated by the ΔCt method (Pfaffl, 2001) using EGFP expression for normalization.

Data were analyzed using one-way ANOVA, Student's t-test, Mann–Whitney U-test or chi-square test, as indicated in the figure legends. Statistical tests were performed using GraphPad Prism.

We thank Yaping Tu for providing Rgs2^−/− mice, Dean Betts and Nicole Edwards for obtaining Rgs2^−/− eggs, Jurrien Dean for the ZP2 antibody, and Laurinda Jaffe and Bruce White for helpful discussions.