Regulation of endoplasmic reticulum Ca²⁺ oscillations in mammalian eggs

At fertilization, an increase in the intracellular Ca²⁺ concentration ([Ca²⁺]_i), which is generally induced following interaction of the gametes, is the universal signal for the completion of meiosis and initiation of embryo development (Stricker, 1999). In mammals, this signal takes the form of periodical increases in [Ca²⁺]_i, termed [Ca²⁺]_i oscillations, which last for several hours after sperm entry (Miyazaki et al., 1986). The [Ca²⁺]_i responses provide spatio-temporal information that is decoded by downstream effectors, whose actions underpin several distinct cellular events such as the release of cortical granules, progression into interphase and pronuclear formation (PN). These and other subtler events underlie the egg-to-embryo transition and are collectively known as ‘egg activation’ (Ducibella et al., 2002; Schultz and Kopf, 1995). Although [Ca²⁺]_i oscillations are a hallmark of mammalian fertilization and required for egg activation, the underlying molecular mechanism(s) that sustain them remain elusive.

Inositol 1,4,5-trisphosphate (IP₃)-mediated Ca²⁺ release from the endoplasmic reticulum ([Ca²⁺]_ER), the main Ca²⁺ reservoir of the cell (Berridge, 2002), is primarily responsible for the initial [Ca²⁺]_i wave and oscillations during fertilization (Miyazaki et al., 1992). Therefore, the general properties of Ca²⁺ release during fertilization have been described in the context of IP₃ production and regulation of its cognate receptor, IP₃R (Jellerette et al., 2000; Miyazaki, 1993). Nevertheless, the regulation of [Ca²⁺]_i oscillations in eggs or cells is likely to be far more complex, because oscillations require Ca²⁺ influx and clearing mechanisms. For instance, for [Ca²⁺]_i oscillations to continue without attenuation, eggs must replenish [Ca²⁺]_ER, because in the absence of external Ca²⁺ ([Ca²⁺]_e), sperm-initiated [Ca²⁺]_i oscillations run down and cease prematurely (Igusa and Miyazaki, 1983; Winston et al., 1995). Thus, [Ca²⁺]_e must cross the plasma membrane (PM) and access the ooplasm by a variety of channels and mechanisms (Smyth et al., 2006). Store-operated Ca²⁺ entry (SOCE) is one of these proposed mechanisms that is activated by the depletion of [Ca²⁺]_ER (Putney, 1990). However, excessive or prolonged [Ca²⁺]_i elevations abolish the oscillatory behavior, and can cause fragmentation and apoptosis (Gordo et al., 2002; Ozil et al., 2005). To prevent this, elevated [Ca²⁺]_i is rapidly removed by pumps and exchangers. The PM Ca²⁺-ATPase (PMCA) and Na⁺/Ca²⁺ exchanger extrude free cytosolic Ca²⁺ to the extracellular space and the sarco-endoplasmic reticulum Ca²⁺-ATPases (SERCA) reuptake it into the ER stores, thereby refilling them (Berridge et al., 2000; Bootman et al., 2001). The mitochondria also uptake cytosolic Ca²⁺, thereby contributing to shape the spatio-temporal patterns of [Ca²⁺]_i responses, while simultaneously promoting a number of events that sustain ATP levels in cells (Hajnóczky et al., 1995). Despite the pivotal role of these mechanisms of Ca²⁺ homeostasis, few studies have examined their underlying molecules and regulation in mammalian oocytes and eggs.

In this regard, few factors are as important to sustain fertilization-associated [Ca²⁺]_i oscillations than [Ca²⁺]_ER, because it is the source of Ca²⁺ during oscillations. Recent studies have shown that during fertilization, [Ca²⁺]_ER levels change in parallel with changes in [Ca²⁺]_i (Takahashi et al., 2013; Wakai and Fissore, 2013), although how [Ca²⁺]_ER levels are regulated during oscillations was not examined. Moreover, it is unknown whether common parthenogenetic agonists such as SrCl₂ or thimerosal, which reportedly induce [Ca²⁺]_i oscillations by sensitizing IP₃Rs and without generating IP₃ (Cheek et al., 1993; Galione et al., 1993), also modify [Ca²⁺]_ER levels. Hence, direct measurement of [Ca²⁺]_ER is required to get a better appreciation of Ca²⁺ homeostasis in these cells. Toward this end, advances in genetically encoded Ca²⁺ probes now allow the measurement of Ca²⁺ dynamics in live cells and in targeted organelles (Demaurex, 2005). Cameleon D1ER, a fluorescence resonance energy transfer (FRET)-based Ca²⁺ indicator, was successfully used to report [Ca²⁺]_ER in somatic cells (Palmer et al., 2004) and recently in mouse eggs (Takahashi et al., 2013; Wakai and Fissore, 2013).

Prior to ovulation and following a systemic LH surge, prophase-arrested germinal vesicle (GV) oocytes resume meiosis and progress to the metaphase stage of the second meiosis (MII). This process is commonly referred to as oocyte maturation, and during it, oocytes acquire fertilization-competence and the precise spatio-temporal pattern of fertilization-associated [Ca²⁺]_i responses. In fact, in-vitro-fertilized GV oocytes show fewer [Ca²⁺]_i oscillations and each [Ca²⁺]_i rise exhibits lesser duration and amplitude than those observed in fertilized MII oocytes (henceforth referred to as eggs) (Jones et al., 1995b; Mehlmann et al., 1996). Given that several of the parameters of Ca²⁺ homeostasis progressively change during maturation, including the steady increase of [Ca²⁺]_ER (Jones et al., 1995b; Mehlmann and Kline, 1994; Wakai et al., 2012), it is possible that the molecules responsible for these adjustments in Ca²⁺ homeostasis, which are mostly unknown in mammalian oocytes, experience dynamic modifications such that some of the mechanisms active at the GV stage might not be so at the MII stage and vice versa.

In the present study, using D1ER and mouse oocytes and eggs we measured [Ca²⁺]_ER levels during [Ca²⁺]_i oscillations induced by a variety of agonists. We found simultaneous changes in [Ca²⁺]_ER and [Ca²⁺]_i levels during oscillations, although with unique [Ca²⁺]_ER kinetics according to the agonist applied. Furthermore, we show that [Ca²⁺]_ER can act as a pacemaker of oscillations and is regulated by the action of mechanisms involved in Ca²⁺ influx and in buffering and sequestering [Ca²⁺]_i. Finally, as well as showing that the SERCA2b protein is expressed in oocytes and eggs, we observed that it undergoes dynamic reorganization during maturation, which might contribute to shaping the fertilization-associated [Ca²⁺]_i responses in these cells.

We first examined whether expression of D1ER protein in both mouse oocytes and eggs attained a distribution that replicated that of the ER in these two stages of maturation. To accomplish this, D1ER cRNA was injected into GV oocytes and MII eggs, and fluorescence was examined ∼5 hours after injection. Confocal images showed that D1ER fluorescence displayed the widespread reticular pattern characteristic of the ER in these cell types, including the conspicuous cortical clusters localization distinctive of the ER in MII eggs (FitzHarris et al., 2007; Kline et al., 1999). Together, these results suggest that D1ER was successfully targeted to the ER of oocytes and eggs (Fig. 1A).

To explore the range of FRET signals in our system, which would allow estimating the relative changes in [Ca²⁺]_ER during Ca²⁺ stimulations, we recorded changes in emission ratios of D1ER (YFP/CFP) induced by addition of ionomycin. The studies were performed in the absence of [Ca²⁺]_e, to maximize emptying of [Ca²⁺]_ER and prevent Ca²⁺ influx. Following addition of ionomycin, the emission ratio of D1ER immediately decreased, because the CFP signal increased and the YFP signal decreased (Fig. 1B). In the absence of agonists, emission ratios remained unchanged for at least 3 hours, although they gradually decreased during additional measurement (data not shown). To establish whether the change in [Ca²⁺]_ER closely corresponded with that of [Ca²⁺]_i levels, we performed simultaneous measurements of [Ca²⁺]_ER and [Ca²⁺]_i using Rhod-2 as the Ca²⁺ indicator (Fig. 1C). Rhod-2 is generally used to measure mitochondrial Ca²⁺, although given that in mouse eggs Rhod-2 fails to target to the mitochondria, it can be used to report cytoplasmic [Ca²⁺]_i (Dumollard et al., 2004). As expected, following addition of ionomycin, [Ca²⁺]_ER and [Ca²⁺]_i underwent simultaneous but opposite changes in relative concentrations.

We next examined changes in [Ca²⁺]_ER levels during [Ca²⁺]_i oscillations. [Ca²⁺]_i oscillations during mammalian fertilization are thought to be triggered following fusion of the gametes by release from the sperm of a male-specific phospholipase C (PLC) enzyme, PLCζ (Nomikos et al., 2013; Saunders et al., 2002). Therefore, we initiated oscillations by injecting PLCζ cRNA, which evokes [Ca²⁺]_i responses that closely and reproducibly resemble fertilization-induced [Ca²⁺]_i oscillations (Saunders et al., 2002). The results show that during oscillations the sharp upstroke of each [Ca²⁺]_i rise coincided with an equally fast drop in [Ca²⁺]_ER levels, as reported by a change in the YFP/CFP ratio (Fig. 2A). Remarkably, it appears that each of the first ∼6–10 [Ca²⁺]_i rises occurred before [Ca²⁺]_ER levels had recovered to levels similar to those prior to the rise, which resulted in a progressive downturn of [Ca²⁺]_ER levels (Fig. 2B,C). In addition, the transient recovery of [Ca²⁺]_ER levels exhibited a two-step process, an initial rapid increase followed by a more gradual increase. Afterwards, basal [Ca²⁺]_ER stabilized and each [Ca²⁺]_i rise seemed to occur from the same [Ca²⁺]_ER level (Fig. 2D), which suggests that by then, the rate of refilling of the stores determines the initiation of the next [Ca²⁺]_i response.

[Ca²⁺]_i oscillations in eggs can also be induced by a variety of pharmacological agents, including strontium chloride (SrCl₂) and thimerosal. Nevertheless, the precise mechanism whereby these agents promote oscillations or the impact on [Ca²⁺]_ER levels are unknown. We first examined the role of SrCl₂, because it is widely used to induce artificial activation of mouse eggs following somatic cell nuclear transfer or round spermatid injection (Loren and Lacham-Kaplan, 2006; Wakayama et al., 1998). To accomplish oscillations, medium devoid of CaCl₂ was supplemented with SrCl₂; this treatment generates repetitive [Ca²⁺/Sr²⁺]_i rises that are characteristically of longer duration than those induced by other procedures (Fig. 3A). Remarkably, although a decrease in the overall [Ca²⁺]_ER levels was noticeable, the magnitude of the YFP/CFP ratio change was small (approximately less than half of the change observed in PLCζ cRNA-induced [Ca²⁺]_i oscillations) and, in actuality, after a few [Ca²⁺/Sr²⁺]_i rises, the levels of [Ca²⁺]_ER started to creep up, suggesting that oscillations by SrCl₂ rely less on intracellular [Ca²⁺]_i mobilization. Supporting this notion, it is worth noting that the changes in [Ca²⁺]_ER levels were slow and clearly behind the changes in [Ca²⁺]_i, because by the time [Ca²⁺/Sr²⁺]_i had peaked, [Ca²⁺]_ER had not changed (Fig. 3A, right panel). The [Ca²⁺]_i rises induced by thimerosal, a thiol-oxidizing agent known to induce oscillations in mammalian oocytes (Swann, 1991), caused greatly different effects on [Ca²⁺]_ER levels. For example, the first and subsequent [Ca²⁺]_i rises were not accompanied by a reduction but rather by an increase in [Ca²⁺]_ER levels (Fig. 3B). Further, in accordance with the high frequency of the oscillations, [Ca²⁺]_ER was rapidly refilled and eventually overloaded, as demonstrated by higher [Ca²⁺]_ER levels at the end of the monitoring period (Fig. 3B, right panel). Collectively, our results show that monitoring [Ca²⁺]_ER reveals specific agonist regulation of [Ca²⁺]_ER, which is consistent with the unique pattern of oscillations by these agonists.

To gain insight into the contribution of Ca²⁺ influx to the refilling of [Ca²⁺]_ER, we simultaneously measured [Ca²⁺]_ER and [Ca²⁺]_i in eggs induced to oscillate by injection of PLCζ cRNA while they were maintained in Ca²⁺-free conditions. As expected, [Ca²⁺]_i oscillations ceased prematurely and, on average, eggs only displayed 2.2±0.32 [Ca²⁺]_i rises, supporting the notion that Ca²⁺ influx is required to maintain oscillations (Fig. 4A). [Ca²⁺]_ER levels also declined, which was confirmed by addition of ionomycin after the oscillations had ceased, because both [Ca²⁺]_i and [Ca²⁺]_ER responses were greatly reduced in PLCζ-injected eggs compared with untreated controls cultured in Ca²⁺-free medium (Fig. 4B). Furthermore, the absence of [Ca²⁺]_e slowed down the refilling of [Ca²⁺]_ER during oscillations, as reflected by the lower mean slope of the second [Ca²⁺]_ER increases in these eggs compared with those oscillating in the presence of [Ca²⁺]_e (Fig. 4C; Fig. 2A compare with Fig. 4A), which prolonged the interval between the first and second [Ca²⁺]_i rises in these eggs (Fig. 4D; Fig. 2A compare with Fig. 4A). The absence of [Ca²⁺]_e also compromised the duration of the first rise (Fig. 4E; Fig. 2A compare with Fig. 4A), confirming that Ca²⁺ influx also contributes to the robust first [Ca²⁺]_i rise (Halet et al., 2004). Collectively, our results therefore show that Ca²⁺ influx is required to replenish [Ca²⁺]_ER and shape [Ca²⁺]_i, oscillations, because without it, [Ca²⁺]_i oscillations cease prematurely and show abnormal parameters.

Compared with the quick return to baseline that cytosolic [Ca²⁺]_i transients experience during oscillations, the recovery of [Ca²⁺]_ER is more gradual, suggesting that other Ca²⁺-buffering systems contribute to cytosolic [Ca²⁺]_i clearance. We therefore investigated the role of PMCA, one of the mechanisms known to mediate Ca²⁺ efflux, on the refilling of [Ca²⁺]_ER. To accomplish this, we took advantage of the knowledge that PMCA is inhibited by millimolar concentrations of gadolinium (Gd³⁺), which at these concentrations generates a Ca²⁺ insulation system, because both Ca²⁺ influx and efflux are greatly reduced (Bird and Putney, 2005); this approach was used in mouse eggs to maintain [Ca²⁺]_i oscillations in the absence of [Ca²⁺]_e (Miao et al., 2012). Using this experimental system (5 mM Gd³⁺), we found that although the initiation of periodical oscillations was delayed by Gd³⁺, injection of PLCζ cRNA induced long-lasting [Ca²⁺]_i oscillations that were unattenuated despite the absence of [Ca²⁺]_e (Fig. 5A). Consistent with the high frequency of [Ca²⁺]_i rises, [Ca²⁺]_ER was rapidly refilled, and unlike [Ca²⁺]_i oscillations under normal [Ca²⁺]_e, basal [Ca²⁺]_ER levels remained largely unchanged. These results suggest that PMCA shapes the pattern of [Ca²⁺]_i oscillations and inhibits its function, leading to retention of Ca²⁺ in the cytosol, which increases the refilling of [Ca²⁺]_ER by SERCA.

Mitochondrial function is closely linked to Ca²⁺ homeostasis, because inhibition of its function in mouse eggs disrupts [Ca²⁺]_i oscillations and causes an increase in [Ca²⁺]_i (Dumollard et al., 2004; Liu et al., 2001). To ascertain whether inhibition of mitochondrial function involved ER Ca²⁺ homeostasis, we exposed oscillating eggs to an inhibitor of mitochondrial ATP synthase (Wolvetang et al., 1994) and monitored the effects on the oscillations and [Ca²⁺]_ER refilling. Addition of oligomycin reduced [Ca²⁺]_ER levels, impaired its refilling and induced a sustained elevation in [Ca²⁺]_i in the majority of eggs (Fig. 5B). We interpret these results to mean that without ATP synthesis, the refilling of [Ca²⁺]_ER is compromised and oscillations cannot be maintained.

The active presence of SERCA proteins in mammalian eggs can be surmised by the alteration of [Ca²⁺]_i levels caused by their exposure to thapsigargin or cyclopiazonic acid (CPA), two SERCA inhibitors; these inhibitors also prevent continuation of [Ca²⁺]_i oscillations in oscillating eggs (Kline and Kline, 1992; Lawrence and Cuthbertson, 1995; Macháty et al., 2002). Nevertheless, the precise rate of change of [Ca²⁺]_ER caused by SERCA inhibition has not been elucidated in these cells. We found that addition of thapsigargin to oscillating eggs immediately reduced basal [Ca²⁺]_ER levels and prevented their recovery thereafter (Fig. 5C). Thus, the results demonstrate that SERCA function is required for long-lasting [Ca²⁺]_i oscillations by maintaining [Ca²⁺]_ER.

Despite their pivotal role in Ca²⁺ homeostasis, the molecular properties of SERCA proteins and the dynamic modifications that they might undergo during maturation have not yet been examined in mammalian eggs. Three different SERCA genes (ATP2A1–ATP2A3) encode three main isoforms (SERCA1–SERCA3), each of which undergoes tissue-specific splicing (Brini and Carafoli, 2011). SERCA1a and SERCA1b are expressed in skeletal muscle, SERCA2a is found in cardiac muscle, whereas SERCA2b is ubiquitous and considered the housekeeping isoform. SERCA3 is instead expressed in a limited number of non-muscle cells. We thus investigated the expression of SERCA2b protein and its subcellular distribution during in vitro maturation of oocytes.

Western blot analysis revealed that SERCA2b protein is expressed throughout maturation and comparable levels of protein are also found in in vivo matured MII eggs, as estimated by the intensity of SERCA2b-reactive bands (artificial units) at different stages of maturation or after ovulation that showed values of 32.9, 39.6, 39.4, 44.6, 40.6 and 41.4 for oocytes at 0, 2 4, 8, 12 hours after initiation of in vitro maturation and 14 hours after administration of hCG (in vivo), respectively (Fig. 6). To examine the subcellular distribution of SERCA2b, we used EGFP-tagged fusion proteins. We found that SERCA2b is diffusely distributed in GV oocytes, whereas it accumulates around the chromosomes at the GV breakdown (GVBD) stage (Fig. 7A, top panel). With progression of maturation, during the transition from the metaphase of first meiosis (MI) to MII, SERCA2b migrates toward the cortical regions surrounding the meiotic spindle. Time-lapse imaging of single oocytes also demonstrated the reorganization of SERCA2b during maturation (supplementary material Movie 1). Given that SERCA2b is an ER-resident protein, we used ER-tagged DsRed cRNA to determine whether it colocalized with the ER. We found that SERCA2b displayed identical localization as ER-DsRed (Fig. 7A, middle panel). Importantly, expression of higher concentration of RNA (1 µg/µl) to attain expression levels of SERCA2b-EGFP comparable to endogenous SERCA2b (Fig. 7B, bottom panel) revealed that SERCA2b displayed dense accumulations or clusters in the subcortical areas of MII eggs (Fig. 7B, top panel), which were not observed in GV oocytes. Furthermore, these clusters were more prominent in in vivo matured MII eggs and curiously remained detectable after extrusion of the second polar body extrusion (2PB) but disappeared by the PN stage after fertilization and did not reappear at the subsequent embryonic stage (2-cell) (Fig. 7C).

Microtubules and microfilaments in a step-wise manner are responsible for the redistribution that the ER undergoes during oocyte maturation (FitzHarris et al., 2007). In accordance with the ER reorganization, inhibition of microtubule polymerization by colcemid abrogated the accumulation of SERCA2b around the chromosomes at the GVBD stage (Fig. 7D), but not the subsequent migration to the cortex. However, treatment with Latruculin A, which depolymerizes actin microfilaments, did not inhibit migration of SERCA2b toward the GV or spindle area, but prevented its migration to the cortex. Taken together, these results suggest that SERCA2b moves along cytoskeletal tracks during oocyte maturation and is actively re-organized into the cortical area in preparation for fertilization.

The increase in [Ca²⁺]_ER during oocyte maturation is a well-documented phenomenon (Jones et al., 1995b; Mehlmann et al., 1996; Wakai et al., 2012) that is likely to contribute to the robust Ca²⁺ release during fertilization (Wakai et al., 2012). To address the functional importance of SERCA2b for ER Ca²⁺ homeostasis during maturation, we investigated whether enhanced SERCA2b expression altered [Ca²⁺]_i responses in oocytes and eggs. This overexpression caused a significant increase in [Ca²⁺]_ER in GV oocytes, because the amplitude of ionomycin-evoked Ca²⁺ release was higher than in control GV oocytes (P<0.05), albeit far lower than in control MII eggs (Fig. 8A); [Ca²⁺]_ER levels in MII eggs were not increased by SERCA2b overexpression. The contribution of SERCA2b to the increase in [Ca²⁺]_ER was further demonstrated by inhibiting SERCA2b function using the pharmacological inhibitor CPA, which prevented the increase of [Ca²⁺]_ER in control or SERCA2b-expressing MII eggs. It should be noted that oocytes overexpressing SERCA2b or matured in the presence of CPA advanced to the MII stage without complications, suggesting that the change in [Ca²⁺]_ER levels is dispensable for meiotic progression (Fig. 8B). Taken together, our results suggest that SERCA2b is required for the increase in [Ca²⁺]_ER during maturation.

We next examined the effects of SERCA2b overexpression on IP₃-mediated Ca²⁺ release. To do this, IP₃ responses were produced by caged-IP₃, which was photo-released by a flash of UV light as previously described (Wakai et al., 2012). To exclude the participation of Ca²⁺ influx, experiments were performed in Ca²⁺-free medium. Consistent with our previous data, release of IP₃ in control oocytes generated a single [Ca²⁺]_i rise that gradually returned to baseline (Fig. 8C). Nevertheless, in the majority of SERCA2b-overexpressing oocytes, although IP₃ release caused a similar [Ca²⁺]_i rise, it returned to basal levels considerably faster and displayed repetitive [Ca²⁺]_i transients before reaching baseline; time-lapse imaging shows the regenerative propagation of [Ca²⁺]_i responses (supplementary material Movie 2). Collectively, these results demonstrate that SERCA2b actively sequesters [Ca²⁺]_i during IP₃-induced Ca²⁺ release.

Finally, we investigated whether SERCA2b overexpression influences [Ca²⁺]_ER homeostasis during oscillations. The overall patterns of [Ca²⁺]_i oscillations in SERCA2b-overexpressing eggs were indistinguishable from controls, because the duration, amplitude and frequency of [Ca²⁺]_i rises were not statistically different (data not shown). Nevertheless, after the fourth or fifth [Ca²⁺]_i rise, basal [Ca²⁺]_ER levels remained higher in SERCA2b-overexpressing eggs, suggesting that SERCA2b is one of the molecules that maintains [Ca²⁺]_ER during oscillations (Fig. 8D).

Long-lasting [Ca²⁺]_i oscillations are a hallmark of mammalian fertilization and are important to establish the step-wise completion of all events of egg activation. The remarkable features presented here regarding Ca²⁺ homeostasis in mouse eggs include: (1) direct [Ca²⁺]_ER measurements demonstrate that refilling of [Ca²⁺]_ER underpins the sustained and periodical configuration of [Ca²⁺]_i oscillations; (2) the regulation of [Ca²⁺]_ER is closely associated with Ca²⁺ influx, Ca²⁺ buffering and sequestering mechanisms; (3) SERCA2b undergoes spatial redistribution during oocyte maturation, which is likely to optimize [Ca²⁺]_i responses during fertilization. Collectively, our results are the first to describe the regulation of [Ca²⁺]_ER in mammalian oocytes and set the stage to investigate how [Ca²⁺]_ER levels affect developmental competence.

Understanding the regulation of [Ca²⁺]_ER is crucial to elucidate the mechanisms that underlie [Ca²⁺]_i oscillations. Using D1ER, we show for the first time that [Ca²⁺]_ER undergoes agonist-specific patterns of oscillations. For example, PLCζ-induced [Ca²⁺]_i oscillations triggered [Ca²⁺]_ER oscillations that displayed two notable characteristics: (1) basal [Ca²⁺]_ER levels progressively decreased during the initial [Ca²⁺]_i rises, although they settled into steady state after certain time; (2) the recovery of [Ca²⁺]_ER after each [Ca²⁺]_i rise was accomplished in two phases, an initial rapid phase and a slower second phase. The progressive decrease in basal [Ca²⁺]_ER during initiation of oscillations might be explained at least in part by the large magnitude of the first rises coupled to the apparent inactivation of the Ca²⁺ influx mechanisms at the MII stage (Cheon et al., 2013). Furthermore, the presence of [Ca²⁺]_i oscillations despite decreasing [Ca²⁺]_ER levels might be associated with mouse eggs having a highly sensitized IP₃R1. MII eggs contain the highest [Ca²⁺]_ER levels of any stage, which, combined with increased ambient IP₃ induced by PLCζ, enhances IP₃R1 sensitivity, making Ca²⁺ release possible when [Ca²⁺]_ER is decreasing. In this context, it can be surmised that during these early stages of fertilization, the refilling of [Ca²⁺]_ER does not seem to set the pace of oscillations. Nevertheless, the decrease in [Ca²⁺]_ER eventually came to a halt and thereafter [Ca²⁺]_ER levels became stable so that after each [Ca²⁺]_i rise [Ca²⁺]_ER levels fully recover in advance of the next [Ca²⁺]_i rise. This stabilization led to high synchrony between initiation of the upstroke of the [Ca²⁺]_i rise with initiation of the downstroke of [Ca²⁺]_ER, which suggests that by this time the refilling of [Ca²⁺]_ER becomes the pacemaker of the oscillations. Moreover, at this point, Ca²⁺ efflux and influx is in balance with Ca²⁺ uptake and release from the ER. Concerning the refilling phases of [Ca²⁺]_ER, the rapid phase appears to consist of Ca²⁺ reuptake from the cytosol by SERCA proteins, whereas the gradual phase might depend on Ca²⁺ influx from the external medium, because in the absence of [Ca²⁺]_e, the latter was markedly extended.

In contrast to PLCζ-induced oscillations, [Ca²⁺]_ER levels remain largely unchanged or even enhanced during Sr²⁺- or thimerosal-induced oscillations. In the case of SrCl₂, only the initial, prolonged rise, possibly caused by sensitization of IP₃R1 by SrCl₂ (Cheek et al., 1993; Zhang et al., 2005), led to a slow decline in [Ca²⁺]_ER. After that, however, Ca²⁺/Sr²⁺ mobilization from the ER appeared to be modest, because by the time [Ca^2+/Sr²⁺]_i levels peaked, [Ca²⁺]_ER levels did not show a corresponding reduction, which suggests that Sr²⁺ from the external milieu is directly contributing to the cytosolic oscillations. Nevertheless, additional studies are needed to ascertain with higher degree of resolution the timing of [Ca²⁺]_i increases and [Ca²⁺]_ER decreases during oscillations. Furthermore, we are unaware of the ability of D1ER to bind Sr²⁺ and the affinity of SERCA for Sr²⁺. Thimerosal produced the most striking results in [Ca²⁺]_ER, because overall [Ca²⁺]_ER levels seemed to increase with progression of oscillations. Thimerosal has been proposed to initiate high-frequency [Ca²⁺]_i oscillations by sensitizing IP₃Rs (Cheek et al., 1993; McGuinness et al., 1996). Moreover, it has been speculated that by acting on cortical IP₃Rs, it might more effectively deplete peripheral stores, thereby promoting near-constitutive Ca²⁺ entry (McGuinness et al., 1996). This persistent Ca²⁺ entry might cause rapid saturation of the cytoplasmic stores, thereby promoting premature Ca²⁺ release and reducing the interval between spikes. Thimerosal could also directly promote Ca²⁺ influx or reduce Ca²⁺ efflux, further contributing to the rapid refilling of [Ca²⁺]_ER and increased basal [Ca²⁺]_ER with progression of oscillations, as observed in our studies. Collectively, our results show that oscillations induced by different agonists cause distinct changes in [Ca²⁺]_ER levels in mouse eggs and Ca² influx-coupled [Ca²⁺]_ER refilling might control the frequency of oscillations.

Accumulating evidence shows that Ca²⁺ influx is required for the persistence of [Ca²⁺]_i oscillations after fertilization (Lee et al., 2013; Lee et al., 2012; Shirakawa and Miyazaki, 1995; Wang et al., 2012). In the present study, we show that one role of Ca²⁺ influx is to replenish [Ca²⁺]_ER levels, because in the absence of [Ca²⁺]_e the rate of [Ca²⁺]_ER refilling was slowed, which led to the premature termination of PLCζ-induced oscillations. Remarkably, by the time oscillations ceased in these eggs, [Ca²⁺]_ER levels were still substantial, because application of ionomycin caused a larger reduction in [Ca²⁺]_ER levels, which suggests that IP₃-sensitive stores are a comparatively small part of the total Ca²⁺ stored in the cell. Ca²⁺-free medium also reduced the duration of the first PLCζ-induced [Ca²⁺]_i rise, although it did not affect the [Ca²⁺]_ER recovery slope (Fig. 4C), which suggests that Ca²⁺ influx prolongs the first rise, possibly by causing Ca²⁺-induced Ca²⁺ release. Ca²⁺-free medium did not affect the [Ca²⁺]_ER recovery slope after the first [Ca²⁺]_i rise (Fig. 4E), suggesting that Ca²⁺ influx does not contribute to refill [Ca²⁺]_ER at this stage, which is consistent with the sharp loss of [Ca²⁺]_ER levels following this rise. Importantly, we are still unaware of the channels that mediate Ca²⁺ influx during maturation and fertilization, and thus, future studies should examine the type of plasma membrane channel(s) responsible for the influx.

The lag time between the rapid recovery of [Ca²⁺]_i and the slow refill of [Ca²⁺]_ER during oscillations suggests that additional Ca²⁺-buffering mechanisms besides SERCA participate in returning [Ca²⁺]_i to basal levels. There are several possible mechanisms including the Na⁺/Ca²⁺ exchanger, which is known to be active in mouse eggs (Carroll, 2000; Pepperell et al., 1999), although here we focused on the roles of PMCAs and the mitochondria, because mouse eggs maintained normal oscillatory patterns in Na⁺-free medium (Carroll, 2000). We used high concentrations of Gd³⁺ to examine the contribution of PMCA. In its presence, the duration of the initial [Ca²⁺]_i rises induced by injection of PLCζ cRNA was increased and the interval between spikes widened. We interpret the broadening of the initial [Ca²⁺]_i rises to be due at least in part to the sensitization of IP₃R1 caused by the retention of Ca²⁺ in the cytosol and/or the enhanced SERCA-mediated Ca²⁺ reuptake, because the decline of [Ca²⁺]_ER levels was also greatly prolonged. The significant initial depletion of [Ca²⁺]_ER might require longer refilling, thereby prolonging the initial interspike intervals. Interestingly, ∼3 hours after initiation of the [Ca²⁺]_i responses, the oscillations became very frequent and of small amplitude, which might be due to faster than normal Ca²⁺ reuptake into the ER and rapid refilling of [Ca²⁺]_ER caused by the modest [Ca²⁺]_i increases. Therefore, our results suggest active participation of PMCA in shaping the pattern of [Ca²⁺]_i oscillations and future studies should identify the molecular presence of PMCAs in mouse eggs.

The mitochondria also contribute to shape [Ca²⁺]_i rises during oscillations (Duchen, 2000; Rizzuto et al., 2000), because they can uptake Ca²⁺ into the matrix, thereby alleviating the overall cytosolic Ca²⁺ load (Rizzuto et al., 1998). In support of this role, inhibition of mitochondrial function in mouse eggs disrupted oscillations and caused a sustained increase in [Ca²⁺]_i (Dumollard et al., 2004; Liu et al., 2001), although this effect was reportedly more related to the ability of mitochondria to produce ATP than their ability to uptake Ca²⁺ (Dumollard et al., 2004). Our results support this view, because inhibition of mitochondrial function with oligomycin reduced [Ca²⁺]_ER levels, inhibited refilling and terminated oscillations. The increase in basal [Ca²⁺]_i is possibly due to a malfunction of SERCA and PMCA pumps. Thus, mitochondrial function seems to be indispensable for [Ca²⁺]_i oscillations, because it is required both to maintain Ca²⁺ levels in the cytosol and in the ER through ATP-driven Ca²⁺-pumping mechanisms.

Despite the functional evidence that SERCA2b is associated with [Ca²⁺]_i oscillations in mouse eggs (Kline and Kline, 1992), molecular evidence of SERCA2b and its contribution to the filling and refilling of [Ca²⁺]_ER in these cells has not been confirmed. Here, we show that inhibition of SERCA2b during maturation prevents the increase in [Ca²⁺]_ER during this process. Furthermore, we show that the protein remains uniformly expressed from the GV stage, which suggests that mechanisms other than protein expression must account for the increase in [Ca²⁺]_ER during maturation.

SERCA2b undergoes major cellular redistribution during maturation. The overall reorganization is similar to that described for the ER and IP₃Rs, both of which form marked clusters in subcortical areas in MII eggs. As a consequence of this reorganization, SERCA2b and IP₃R1 are closely apposed, thereby facilitating the refilling of ER Ca²⁺ stores rich in IP₃R1, which are probably frequently depleted during fertilization. It is worth noting that SERCA2b clusters remain past the 2PB stage, ∼3 hours after fertilization, but disappear by the PN stage, ∼8 hours after fertilization, coinciding with the natural duration of the [Ca²⁺]_i oscillations in these species (Jones et al., 1995a).

Our results also show that inhibition of SERCA2b by pharmacological inhibitors disrupts the refilling of [Ca²⁺]_ER, lowering [Ca²⁺]_ER levels and causing premature termination of oscillations. Remarkably, despite lower [Ca²⁺]_ER levels, most eggs continued to oscillate for an additional 20–30 minutes, a finding previously noted by others, which suggests that mouse eggs contain thapsigargin-insensitive or -resistant stores that can support oscillations (Kline and Kline, 1992). The contribution of SERCA2b to oscillations was also addressed in overexpression studies. Expression of exogenous SERCA2b reduced the duration of [Ca²⁺]_i rises generated by caged IP₃ while at the same time stimulating the presence of oscillations. This result is consistent with data in Xenopus oocytes where SERCA overexpression increased the frequency of IP₃-induced [Ca²⁺]_i oscillations and repetitive [Ca²⁺]_i rises were observed during the decaying phase of photo-release IP₃-induced responses (Camacho and Lechleiter, 1993). We also observed that SERCA2b overexpression lessened the loss of [Ca²⁺]_ER levels experienced by eggs injected with mPLCζ cRNA.

In conclusion, the present study provides novel insights into the function of [Ca²⁺]_ER during oscillations in mouse eggs, and shows that agonists have distinct impact on the refilling and overall load of [Ca²⁺]_ER. We also found that Ca²⁺ influx and the function of Ca²⁺-ATPase pumps and mitochondria are all key regulators of [Ca²⁺]_ER during maturation and fertilization. Future studies should identify the channels that regulate Ca²⁺ influx into oocytes and eggs and the regulatory mechanisms that modulate the function of the Ca²⁺ pumps and the mitochondria, because this knowledge will make it possible to manipulate [Ca²⁺]_ER levels to ascertain its impact on maturation and development.

Ionomycin, cyclopiazonic acid, thapsigargin, oligomycin and Latrunculin A were purchased from Calbiochem (San Diego, CA). Fura-2AM, Fluo-4AM, Rhod2AM, pluronic acid and caged-IP₃ (cIP₃) were purchased from Invitrogen (Carlsbad, CA). Other all chemicals were from Sigma (St Louis, MO) unless otherwise specified.

GV oocytes and MII eggs were collected from the ovaries of 4- to 8-week-old and oviducts of 6- to 10-week-old CD-1 female mice, respectively. Females were injected with 5 IU pregnant mare serum gonadotropin (PMSG). Cumulus-cell-enclosed GV oocytes were recovered 42–46 hours post-PMSG and the cumulus cells were removed by repeated pipetting. Ovulated MII eggs were recovered 12–14 hours after injection of 5 IU hCG, which was administered 44–48 hours after PMSG stimulation. All procedures were performed according to research animal protocols approved by the University of Massachusetts Institutional Animal Care and Use Committee.

Human SERCA2b was subcloned into a pcDNA6 vector (pcDNA6/Myc-His B; Invitrogen, Carlsbad, CA) between the XhoI and PmeI restriction sites. To analyze the subcelluar distribution, SERCA2b was N-terminally tagged with EGFP between the EcoRV and XhoI restriction sites. Cameleon D1ER and pDsRed2-ER were kindly provided by R. Tsien (UCSD) and M. Trebak (Albany Medical College), respectively. The ER-targeting sequence of calreticulin, DsRed2 and the KDEL ER retention sequence were ligated to pcDNA6. Because the original D1ER construct somehow failed to target ER compartment in oocytes/eggs, the cameleon portion was amplified by PCR and inserted between calreticulin targeting and KDEL sequences in place of DsRed2. Mouse PLCζ was a kind gift from K. Fukami (Tokyo University of Pharmacy and Life Science, Japan) and was subcloned into a PCS2+ vector, as previously described (Kurokawa et al., 2007).

Plasmids were linearized with a restriction enzyme downstream of the insert to be transcribed. cDNA was in vitro transcribed using the T7 or SP6 mMESSAGE mMACHINE Kit (Ambion, Austin, TX) according to the promoter that is contained in the constructs. A poly(A) tail was added to the mRNAs using a Tailing Kit (Ambion) and poly(A)-tailed RNAs were eluted with RNase-free water and stored in aliquots at −80°C. For microinjection, cRNA solution was loaded into glass micropipettes and delivered into oocytes/eggs by pneumatic pressure (PLI-100, Harvard Apparatus, Cambridge, MA). Each oocyte received 7–12 pl (∼1–3% of the total volume of the egg). The volumes injected typically range from 2 to 10 pl, which is 1–5% of the egg (∼70–75 µm).

To estimate relative changes in [Ca²⁺]_ER, emission ratio imaging of the D1ER (YFP/CFP) was performed using a CFP excitation filter, dichroic beamsplitter, CFP and YFP emission filters (Chroma technology, Rockingham, VT; ET436/20X, 89007bs, ET480/40m and ET535/30m). To measure [Ca²⁺]_ER and [Ca²⁺]_i simultaneously, eggs that had been injected with Cameleon D1ER cRNA were loaded with 1 µM Rhod-2AM supplemented with 0.02% pluronic acid for 20 minutes at room temperature. Eggs then were attached on glass-bottom dishes (MatTek Corp., Ashland, MA) and placed on the stage of an inverted microscope. CFP, YFP and Rhod-2 intensities were collected every 20 seconds by a cooled Photometrics SenSys CCD camera (Roper Scientific, Tucson, AZ). The rotation of excitation and emission filter wheels was controlled using the MAC5000 filter wheel/shutter control box (Ludl) and NIS-elements software (Nikon). Fura-2AM (1.25 µM) and Fluo-4AM (1 µM) were used to analyze ionomycin-induced and cIP₃-induced Ca²⁺ rises, respectively. Fura-2 fluorescence was excited with 340 nm and 380 nm wavelengths every 20 seconds and emitted light was collected at wavelengths above 510 nm. Fluo-4 was excited with 480 nm wavelengths every 5 seconds and emitted light was collected at wavelengths greater than 510 nm. cIP₃ (0.25 mM) was injected into oocytes and the photolysis was accomplished with a 360 nm wavelength.

Live-cell imaging of oocytes/eggs expressing fluorescent-tagged proteins was performed using a laser-scanning confocal microscope (LSM 510 META, Carl Zeiss Microimaging Inc., Germany) fitted with a 63× 1.4 NA oil-immersion objective lens. Images were acquired with LSM software (Carl Zeiss Microimaging Inc., Germany) and psueudocolored in Photoshop CS (Adobe). Time-lapse imaging of SERCA2b-EGFP was performed using a Nipokow disk confocal unit (CV-1000, Yokogawa Electric Corp., Japan) outfitted with a 30× silicone immersion objective lens.

Cell lysates from mouse oocytes/eggs were prepared by adding 2× sample buffer. Proteins were separated by SDS-PAGE and transferred to PVDF membrane (Millipore, Bedford, MA). After blocking with 6% skimmed milk dissolved in TBS supplemented with 0.05% Tween-20 (TBST) for 2 hours at 4°C, membranes were probed with the rabbit polyclonal antibody specific to SERCA2b (Vangheluwe et al., 2007) (1∶2000) in TBST with 3% skimmed milk overnight at 4°C. Goat anti-rabbit antibody conjugated to horseradish peroxidase (HRP) was used as a secondary antibody (1∶2000, Bio-Rad, Hercules, CA) for detection of chemiluminescence (NEN Life Science Products, Boston, MA) according to the manufacturer's instructions. The signal was digitally captured using a Kodak Imaging Station (440 CF, Rochester, NY).

Values from three or more experiments performed on different batches of oocytes/eggs were analyzed by the Student's t-test, Chi-squared test or one-way ANOVA, as appropriate. Differences were considered significant at P<0.05. Values are presented as means ± s.e.m. Significance among groups or treatments is denoted in bar graphs by different superscripts or by the presence of asterisks.

The authors want to thank Roger Tsien's lab (UCSD) and M. Trebak (Albany Medical College) for sharing the D1ER and pDsRed2-ER constructs, respectively. We also thank Changli He and Banyoon Cheon for technical assistance.