Ca²⁺_cyt negatively regulates the initiation of oocyte maturation

Mammalian and amphibian oocytes arrest at the G2/M border of the cell cycle after oogenesis (Yamashita et al., 2000). Before these oocytes become fertilization competent, they undergo a so-called “oocyte maturation” period during which they acquire the ability to activate in response to sperm entry, and to support the early stages of embryonic development (Yamashita et al., 2000). During maturation, oocytes progress through meiosis and arrest at metaphase of meiosis II until fertilization.

Xenopus oocyte maturation provides a valuable model to elucidate the signal transduction cascade mediating meiosis entry and progression. In Xenopus, oocyte maturation is triggered by the hormone progesterone, which binds to a cell surface receptor and not the classical nuclear receptor/transcription factor. Progesterone leads to inhibition of cAMP-dependent PKA and translation of the proto-oncogene c-Mos, which induces the MAPK cascade culminating in the activation of maturation promoting factor (MPF; for review see Nebreda and Ferby, 2000). MPF is the central kinase that regulates meiotic progression, and consists of a catalytic p34^cdc2 serine/threonine kinase subunit (Cdk 1), and a regulatory cyclin B subunit (Coleman and Dunphy, 1994). MPF is also activated by the removal of inhibitory phosphorylations by the Cdc25C phosphatase, which is induced by the polo-like kinase cascade (Nebreda and Ferby, 2000).

A variety of genetic and biochemical evidence support a role for Ca²⁺, and its downstream effectors CaM and Ca²⁺-CaM–dependent protein kinase II, in mitosis initiation and progression (Means, 1994; Whitaker, 1995). Ca²⁺ signals are required for nuclear envelope breakdown (NEBD), and chromosome condensation during mitosis (Steinhardt and Alderton, 1988; Twigg et al., 1988; Kao et al., 1990; Tombes et al., 1992; Ciapa et al., 1994). Furthermore, the cytoplasmic Ca²⁺ (Ca²⁺_cyt) rise observed at fertilization is the universal signal for egg activation in all species investigated (Stricker, 2000). Ca²⁺ signals at fertilization release the metaphase II arrest by activating proteolytic degradation of cytostatic factor, thus inducing completion of meiosis II and entry into the mitotic cell cycle (Tunquist and Maller, 2003). In addition, Ca²⁺ release at fertilization induces both the fast and slow blocks to polyspermy in Xenopus eggs (Machaca et al., 2001). In contrast to the well-defined roles for Ca²⁺ signals in mitosis and after fertilization, the role of Ca²⁺ signals during oocyte maturation remains contentious.

There has been a long-standing debate in the literature as to whether Ca²⁺ signals are required for Xenopus oocyte maturation/meiosis (Duesbery and Masui, 1996). Early reports argued that a Ca²⁺_cyt rise is sufficient to induce oocyte maturation (Wasserman and Masui, 1975; Moreau et al., 1976; Schorderet-Slatkine et al., 1976). Furthermore, oocytes injected with high concentrations of Ca²⁺ buffers were unable to mature (Moreau et al., 1976; Duesbery and Masui, 1996). However, injection of IP₃, which induces Ca²⁺ release, did not stimulate meiotic maturation (Picard et al., 1985). Additional support for a Ca²⁺ role in oocyte maturation comes from reports that measured a Ca²⁺_cyt rise after progesterone addition using ⁴⁵Ca²⁺ as a tracer, Ca²⁺ imaging, or Ca²⁺-sensitive electrodes (O'Connor et al., 1977; Moreau et al., 1980; Wasserman et al., 1980). In contrast, others could not detect Ca²⁺_cyt changes downstream of progesterone addition using similar techniques (Robinson, 1985; Cork et al., 1987). A role for CaM in Xenopus oocyte maturation has also been postulated (Wasserman and Smith, 1981) and challenged (Cicirelli and Smith, 1987). These conflicting reports argue that the relationship between Ca²⁺ and oocyte maturation is complex.

We decided to revisit the role of Ca²⁺ during oocyte maturation/meiosis by framing the problem in terms of the spatially distinct sources of Ca²⁺ signals. Ca²⁺_cyt signals can be generated either due to Ca²⁺ release from intracellular Ca²⁺ stores (ER) or Ca²⁺ influx from the extracellular space. In fact, these two Ca²⁺ sources are mechanistically linked through the store-operated Ca²⁺ entry (SOCE) pathway, which is activated in response to intracellular Ca²⁺ stores depletion. Therefore, Ca²⁺_cyt is regulated by the balance between Ca²⁺ release and Ca²⁺ influx. By manipulating Ca²⁺ store load and the extent of Ca²⁺ influx through SOCE, we show here that Ca²⁺ signals are not required for meiosis entry. On the contrary, high Ca²⁺_cyt delays meiosis entry. However, in the absence of Ca²⁺_cyt signals oocytes arrest in meiosis I, form abnormal spindles, and do not extrude a polar body. Surprisingly, MAPK and MPF kinetics in oocytes deprived of Ca²⁺ signals are normal. We further mapped the site of action of Ca²⁺_cyt on meiosis entry and show that Ca²⁺_cyt negatively regulates the cell cycle machinery by acting downstream of PKA and upstream of Mos. These data argue that Ca²⁺ signals regulate the timing of meiosis entry, and that they are required for the completion of meiosis I. The dual role of Ca²⁺_cyt revealed by these studies help explain some of the controversy surrounding the role of Ca²⁺ in oocyte maturation, and provides a framework to explore the role of Ca²⁺-dependent signaling cascades in meiosis.

Maturing oocytes in media with different Ca²⁺ concentrations ([Ca²⁺]) does not affect the time course or extent of germinal vesicle (nucleus) breakdown (GVBD; Fig. 1, A and C), which is indicative of meiosis entry. The rate of maturation in the population was quantified as the time required for 50% of the oocytes to undergo GVBD (GVBD₅₀). Because the rate and extent of GVBD were unaffected in low Ca²⁺ (L-Ca) medium, this shows that Ca²⁺ influx is not required for entry into meiosis (Fig. 1, C and D).

To test whether intracellular Ca²⁺ levels affect oocyte maturation, we emptied Ca²⁺ stores either by treating cells with thapsigargin, an inhibitor of the ER Ca²⁺ ATPase (SERCA), or with the Ca²⁺ ionophore ionomycin. Thapsigargin leads to Ca²⁺ store depletion because of a poorly defined Ca²⁺ leak pathway from the ER (Camello et al., 2002). Emptying Ca²⁺ stores activates Ca²⁺ influx through SOCE (Parekh and Penner, 1997). Because the extent of Ca²⁺ influx through SOCE depends on [Ca²⁺] in the medium, oocytes incubated in high Ca²⁺ (H-Ca) medium will have more Ca²⁺ influx than those in normal Ca²⁺ (N-Ca) medium, and no Ca²⁺ influx in expected in L-Ca medium (see Fig. 4).

Emptying Ca²⁺ stores with either thapsigargin or ionomycin in N-Ca does not affect the time to GVBD₅₀ (Fig. 1, B and C, Thap-N-Ca and Ion-N-Ca), but decreases maximal levels of GVBD (Fig. 1 D, Thap-N-Ca and Ion-N-Ca). In H-Ca, emptying Ca²⁺ stores results in high percentage of cellular degeneration due to excessive Ca²⁺ influx, thus prohibiting analysis of the rate of meiosis entry because GVBD₅₀ is rarely reached under these conditions (Fig. 1 B, Thap-H-Ca and Ion-H-Ca). In contrast, emptying Ca²⁺ stores in L-Ca medium accelerates the rate of maturation (Fig. 1, B and C, Thap-L-Ca and Ion-L-Ca), without affecting maximal maturation levels (Fig. 1 D, Thap-L-Ca and Ion-L-Ca). In L-Ca medium with Ca²⁺ stores depleted, oocytes are unable to generate Ca²⁺ signals after progesterone addition because Ca²⁺ stores are depleted and Ca²⁺ influx is prevented in L-Ca medium (see Fig. 4); nonetheless they enter meiosis at an accelerated rate. These data show that Ca²⁺ signals are not required for GVBD and argue that Ca²⁺_cyt negatively regulates meiosis entry.

Although Ca²⁺ signals after progesterone addition are dispensable for GVBD, it is conceivable that Ca²⁺_cyt signals generated before progesterone addition are required for meiosis entry. To determine whether this is the case, we depleted Ca²⁺ stores with thapsigargin and waited for extended periods of time before inducing maturation with progesterone. We reasoned that if Ca²⁺ signals are required for GVBD, the longer we wait after depriving oocytes of Ca²⁺ signals the less effective progesterone will be in inducing maturation. Depleting stores for as long as 48 h does not affect the extent of oocyte maturation (Fig. 1 F), but still enhances oocyte maturation rate (Fig. 1 E). These results support the conclusion that Ca²⁺ signals are not required for entry into meiosis.

The more rapid maturation observed in L-Ca medium when Ca²⁺ stores are depleted, argues that Ca²⁺_cyt negatively regulates meiosis entry. It follows then that buffering Ca²⁺_cyt at low levels should also accelerate meiosis entry. This is indeed the case as injection of 500 μM 1,2-bis(2-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid (BAPTA) alone or in combination with thapsigargin in varying orders, results in a more rapid maturation (Fig. 2, A and B). BAPTA and thapsigargin treatments accelerate maturation to a similar extent with no significant additive effect. However, the longer the interval between BAPTA injection and progesterone addition the more rapid the maturation rate (Fig. 2 B). In all treatments the extent of maturation was comparable (Fig. 2 C). Similar results were obtained when BAPTA was injected at 1 mM (unpublished data). These data argue that at resting Ca²⁺_cyt levels some Ca²⁺-dependent pathways are active and negatively regulate meiosis entry.

As reported by others, injecting high BAPTA concentrations (2.5–5 mM) blocks GVBD, arguing that Ca²⁺ is required for GVBD (Moreau et al., 1976; Duesbery and Masui, 1996). However, at such high BAPTA concentrations we observe significant levels of oocyte degeneration, bringing into question the specificity of this treatment. As shown below, injecting BAPTA at 500 μM effectively buffers Ca²⁺_cyt transients (Fig. 4). In addition, depleting Ca²⁺ stores in the absence of extracellular Ca²⁺, thus depriving oocytes of Ca²⁺ signals, does not block GVBD. Furthermore, treating oocytes with the heavy metal chelator TPEN blocks oocyte maturation (unpublished data). Based on these findings, it is possible that high BAPTA concentrations block GVBD in a Ca²⁺-independent manner, either due to a nonspecific effect of BAPTA, or chelation of other metal ions because BAPTA is a potent chelator of transition metals (Arslan et al., 1985).

If Ca²⁺_cyt negatively regulates meiosis entry it is expected that raising Ca²⁺_cyt levels would lead to a slower rate of oocyte maturation. To test whether this is the case we induced different levels of Ca²⁺ influx through SOCE by depleting Ca²⁺ stores with thapsigargin and incubating oocytes with solutions containing 5 mM, 3 mM, 1.5 mM, 0.6 mM, and 50 μM Ca²⁺ (H-5, H-3, H-1.5, N-Ca, and L-Ca, respectively). Inducing maturation in solutions with different [Ca²⁺] has no effect on the rate of oocyte maturation, except for L-Ca medium where maturation rate was more rapid (Fig. 3, A and C). Although this enhancement was more pronounced in this set of experiments, Fig. 1 shows a similar tendency toward a more rapid maturation in L-Ca medium. More importantly, after store depletion with thapsigargin, the higher the concentration of extracellular Ca²⁺ the slower the rate of maturation (Fig. 3, B and C). Furthermore, the extent of maturation was reduced in H-Ca containing solutions (H-5, H-3, and H1.5). At both 3 and 5 mM of extracellular Ca²⁺ (H-5 and H-3) some cellular degeneration was observed, but at 1.5 mM of extracellular Ca²⁺ the oocytes were healthy, but the rate of oocyte maturation was slower. These data show that the higher the level of Ca²⁺ influx the slower the rate of maturation, supporting the conclusion that high Ca²⁺_cyt levels negatively regulate meiosis entry.

To confirm that the different treatments are modulating Ca²⁺_cyt levels as predicted, we used endogenous Ca²⁺-activated Cl⁻ current (I_Cl,Ca), as an in situ marker of Ca²⁺_cyt levels (Fig. 4). We have shown previously that I_Cl,Ca provides an accurate measure of both Ca²⁺ release and influx (Machaca and Hartzell, 1999). During Ca²⁺ release I_Cl,Ca is activated as a sustained current (I_Cl1) at depolarized voltages (+40 mV; Fig. 4 A, left, trace t). I_Cl1 is sustained because during Ca²⁺ release Ca²⁺_cyt levels remain high for the duration of the voltage pulse. In contrast, during Ca²⁺ influx I_Cl,Ca is activated as a transient current (I_ClT) only when the +40 mV pulse is preceded by a hyperpolarization step (−140 mV) to induce Ca²⁺ influx (Fig. 4 A, right, traces w–z). I_ClT is transient because Ca²⁺ flows into the cell during the preceding −140 mV pulse, and then dissipates rapidly resulting in current inactivation (Machaca and Hartzell, 1999). We have shown using simultaneous electrical recording and Ca²⁺ imaging that I_Cl,Ca faithfully reports the levels and kinetics of Ca²⁺ release and influx (Machaca and Hartzell, 1999). However, it is important to note that although I_Cl,Ca provides an accurate measure of Ca²⁺ release and Ca²⁺ influx, it does not directly reflect Ca²⁺ levels deep in the cytosol as these channels localize to the plasma membrane.

To determine store Ca²⁺ load in the different treatments, we incubated oocytes in Ca²⁺-free Ringer (F-Ca), and depleted Ca²⁺ stores with ionomycin. Because no Ca²⁺ influx is possible in Ca²⁺-free solution, the level of I_Cl1 activated in response to ionomycin provides a measure of the extent of store Ca²⁺ load (Fig. 4, A and B). After the dissipation of the Ca²⁺ release transient indicated by the return of I_Cl1 to baseline (Fig. 4 B, squares), oocytes were sequentially exposed to L-Ca, N-Ca, H1.5-Ca, H3-Ca, and H5-Ca to determine the extent of Ca²⁺ influx (Fig. 4, B, D, and F). This protocol was applied to control untreated oocytes (Fig. 4 B) or to oocytes incubated in thapsigargin to fully deplete Ca²⁺ stores (Fig. 4 D), or injected with 500 μM BAPTA (Fig. 4 F). The levels of Ca²⁺ release as indicated by I_Cl1 and the levels of Ca²⁺ influx as indicated by I_ClT were quantified in the different treatments (Fig. 4, C, E, and G). In control oocytes ionomycin activates a large I_Cl1, indicating that Ca²⁺ stores are fully loaded (Fig. 4 B, squares; Fig. 4 C, Ca Rel.). Ca²⁺ release leads to store depletion which activates SOCE. As expected, no I_ClT is detected in either Ca²⁺-free solution (F-Ca) or in L-Ca medium (L-Ca) confirming our prediction that at 50 μM of extracellular Ca²⁺ no Ca²⁺ influx occurs (Fig. 4 B, circles; Fig. 4 C). Increasing levels of Ca²⁺ influx (indicated by I_ClT) are observed in media with increasing [Ca²⁺] (Fig. 4 B, circles; Fig. 4 C; Fig. 4 A, right).

Oocytes treated with thapsigargin did not release Ca²⁺ in response to ionomycin as no I_Cl1 was activated (Fig. 4 D, squares; Fig. 4 E, Ca Rel.), showing that Ca²⁺ stores were depleted. Because thapsigargin depletes Ca²⁺ stores, it activates Ca²⁺ influx through SOCE. As for control oocytes no Ca²⁺ influx is observed in F-Ca or L-Ca solutions, and higher levels of Ca²⁺ influx, as indicated by I_ClT, are detected in solutions containing increasing Ca²⁺ (Fig. 4 D, circles; Fig. 4 E). Ca²⁺ influx levels in thapsigargin-treated cells were similar to those in control cells (Fig. 4, C and E).

BAPTA injection dramatically reduces both Ca²⁺ release (I_Cl1) and Ca²⁺ influx (I_ClT) transients (Fig. 4, F and G). Small levels of Ca²⁺ release are observed in BAPTA-injected cells (Fig. 4 F, squares; Fig. 4 G, Ca Rel.), indicating that Ca²⁺ stores still contain Ca²⁺, but that as Ca²⁺ is released it is chelated by BAPTA, thus drastically reducing the levels of free Ca²⁺ available to activate I_Ca,Cl. The same is true during the Ca²⁺ influx phase in different [Ca²⁺] (Fig. 4, F and G). No Ca²⁺ influx can be detected in L-Ca solution, and small levels of I_ClT are observed in N-Ca through H3-Ca. Only H5-Ca produced evident, but small I_ClT consistently (Fig. 4 G). This indicates that the primary effect of BAPTA injection is to buffer Ca²⁺_cyt at low levels. The fact that BAPTA injection enhances the rate of meiosis entry in a similar fashion to store depletion argues that this enhancement is due to a reduction of Ca²⁺_cyt levels. It is noteworthy that Ca²⁺ store depletion has been shown to alter ER protein expression (Soboloff and Berger, 2002), however, based on the BAPTA data and the delayed maturation rate with high Ca²⁺_cyt, it is unlikely that this is affecting meiosis entry.

We assayed the rate and extent of oocyte maturation above based on the GVBD time course. GVBD marks entry into meiosis but does not provide any information about meiosis progression. Oocyte maturation is considered complete once oocytes reach metaphase of meiosis II. Although, the data presented so far show that Ca²⁺ signals are not required for entry into meiosis, they do not address whether meiosis/oocyte maturation can progress normally in the absence of Ca²⁺ signals. To determine whether interfering with Ca²⁺ signaling pathways affects meiosis progression, we tested the activation kinetics of the MAPK-MPF kinase cascade, which regulates meiosis transitions (Nebreda and Ferby, 2000). As described above (Figs. 1 and 2), treating cells with either thapsigargin or injecting BAPTA accelerates meiosis entry (Fig. 5 A). In N-Ca medium, MAPK phosphorylation is first detected 2 h before GVBD, peaks at GVBD, and remains high for the remainder of maturation (Fig. 4 B, N-Ca). MAPK activates with similar kinetics in L-Ca, or after the thapsigargin (Thaps) or BAPTA treatments (Fig. 5 B). However, consistent with the GVBD time course, MAPK activated 1 h earlier in L-Ca medium and 2 h earlier after thapsigargin or BAPTA treatments (Fig. 5 B). These data show that in the absence of Ca²⁺ signals MAPK is induced earlier, but has normal kinetics after GVBD.

MPF activates in a characteristic fashion during oocyte maturation with a sharp peak at GVBD followed by a decline to an intermediate level that is indicative of the meiosis I to meiosis II transition, and rises again as oocytes progress through meiosis II (Fig. 4 C, N-Ca). MPF activity in oocytes matured in L-Ca medium follows the same kinetics except that peak MPF activity (GVBD) occurs ∼1 h earlier than in N-Ca. This is consistent with the MAPK activation kinetics (Fig. 5 B) and the rate of GVBD (Fig. 5 A). MPF kinetics in the thapsigargin and BAPTA treatments are normal, except that, as for GVBD and MAPK, MPF activity peaks 2 h earlier than in the N-Ca treatment (Fig. 5 C, Thaps and BAPTA). Interestingly, an increase in MPF activity is detected as early as 1 h before GVBD (Fig. 5 C, Thaps and BAPTA, arrows). Such a premature activation of MPF is never observed in N-Ca or L-Ca. Therefore, MAPK and MPF activation kinetics (Fig. 5, B and C) correlate well with each other and with the rate of GVBD (Fig. 5 A), and show that reducing Ca²⁺_cyt transients leads to premature activation of the MAPK-MPF kinase cascade. This premature activation explains the accelerated maturation rate in treatments that reduce Ca²⁺_cyt transients. Therefore, Ca²⁺_cyt modulates meiosis entry by negatively regulating the MAPK-MPF cascade.

Kinase data in oocytes deprived of Ca²⁺ signals (Thaps and BAPTA) suggest that in the absence of Ca²⁺_cyt signals meiosis proceeds normally after GVBD. To determine whether this is the case we imaged meiotic spindle structure and chromosome dynamics in oocytes matured in N-Ca, L-Ca, and oocytes treated with thapsigargin or BAPTA (Fig. 6). This allowed us to assess the progression of nuclear maturation, and directly compare it to MPF, MAPK, and GVBD kinetics because all three experiments were performed on the same batch of oocytes.

Control oocytes matured in N-Ca medium progress normally through meiosis (Fig. 6 A, Table I, N-Ca). At GVBD oocytes were at the late prophase I stage (Fig. 6 A, N-Ca, P), which refers to oocytes that have undergone GVBD, have condensed chromosomes, and organized microtubules around the chromosomes, but have not yet formed a bipolar spindle (Fig. 6 A, N-Ca, P). At 0.5 h after GVBD prometaphase I structures (30%; Table I) are observed with a typical bipolar spindle and associated chromosomes (Fig. 6 A, N-Ca, PM I). This is followed by metaphase I with chromosome lined up at the metaphase plate (Fig. 6 A, N-Ca, M I) at ∼1 h after GVBD (Table I). Between 2–4 h after GVBD oocytes progress from M I to metaphase II (Table I) at which stage they arrest. Examples of anaphase I (A), prometaphase II (PM II), and metaphase II spindles are shown in Fig. 6 A (N-Ca).

Surprisingly, oocytes matured in L-Ca medium alone or treated with thapsigargin or BAPTA formed abnormal spindles. The progression through meiosis was not significantly different between the three treatments which will be discuss as a group. At GVBD a large percentage of these oocytes (≥57%; Table I) had condensed chromosomes, but the microtubule were still dispersed over a large area. We refer to this stage as early prophase (EP; Fig. 6 A), because eventually these oocytes do progress to the late prophase stage (P) as described for the control group (N-Ca) above (Table I). However, oocytes matured in L-Ca medium or treated with BAPTA or thapsigargin rarely progress to prometaphase I and never reach metaphase I (Fig. 6 A; Table I). Instead, they form abnormal structures at different rates depending on the treatment as detailed in Table I. Based on the severity of defects we divided abnormal spindles into three groups: (1) We refer to small and/or slightly disorganized spindles as prometaphase like (PM-L; Fig. 6 A). These spindles are the least disorganized and are observed throughout the time period studied (Table I). (2) The second group represents completely disrupted spindles (Ab) with no clear structure and with the microtubules highly condensed and/or spread over a large area. In most but not all instances condensed chromosomes were still associated with the disrupted spindle (Fig. 6 A, Ab). (3) The last and most interesting group we refer to as the double spindle (DS) group (see Fig. 6 A). These double spindles were most common in the L-Ca group (Table I), but were also observed in the thapsigargin treatment (Fig. 6 A, Thaps, DS).

It is interesting that the highest percentage of disrupted spindles (Ab) in the experimental groups (L-Ca, Thaps, and BAPTA) is observed at 2–3 h after GVBD which corresponds to the meiosis I to meiosis II transition in control oocytes. At later time points, spindle structures are improved (PM-L), as if the cells are attempting to go through meiosis II. This is supported by the appearance of DS structures at 3–4 h after GVBD arguing that although these oocytes do not progress normally through meiosis I, they attempt to form a meiosis II spindle, that in some cases lead to the observed double-spindle structures. This might be expected with normal MAPK and MPF kinetics past GVBD in all the treatments (Fig. 5), which will drive these cells into meiosis II despite the abnormal progression through meiosis I. The fact that MPF and MAPK kinetics are normal past GVBD in the L-Ca, thapsigargin, and BAPTA treatments (Fig. 5), but that meiotic spindles are disorganized show that interfering with Ca²⁺ signaling during meiosis uncouples the MAPK-MPF cascade from spindle structure regulation. Furthermore, these results show that although Ca²⁺ signals are not required for meiosis entry (GVBD and chromosome condensation), they are necessary for progression through meiosis I and for bipolar spindle formation.

Spindle structure data show that interfering with Ca²⁺ signaling leads to abnormal spindles, arguing that oocytes do not complete meiosis I. To directly confirm this, we assessed polar body emission from 1–4 h after GVBD in the different treatments (Fig. 6, B and C). In N-Ca medium, polar bodies were observed in the majority of the cells, but as expected from the spindle structure experiments, oocytes matured in L-Ca medium or treated with thapsigargin or BAPTA rarely extrude a polar body (Fig. 6 C). This shows that interfering with Ca²⁺ signaling inhibits the completion of meiosis I.

The fact that oocytes matured in L-Ca medium had abnormal spindle structure and could not finish meiosis I argues that either Ca²⁺ influx from the extracellular space is required during the early stages of oocytes maturation for normal meiosis progression, or that oocytes are unable to form a polar body at low extracellular Ca²⁺. To differentiate between these possibilities we assessed polar body emission in oocytes incubated in N-Ca solution until GVBD and then switched to L-Ca solution (N-L); or in oocytes incubated in L-Ca solution until GVBD and then switched to N-Ca medium (L-N). Oocytes in the N-L group, but not the L-N group, emitted polar bodies to the same extent as the N-Ca control group (Fig. 6, B and C), arguing that Ca²⁺ influx in the early stages of oocyte maturation is required for meiosis I progression.

To better define the mechanism by which Ca²⁺_cyt negatively regulates meiosis entry, we mapped the site of action of Ca²⁺_cyt on the cell cycle machinery using an epistatic approach. For these experiments thapsigargin-treated oocytes represented the experimental group deprived of Ca²⁺ signals. Control oocytes were activated in N-Ca solution. Our approach was to activate the cell cycle machinery at different points along the MAPK-MPF signal transduction cascade, and determine whether the enhancing effect of Ca²⁺ depravation on the rate of maturation is still observed. Oocytes were activated with progesterone, or by injection of an inhibitor of PKA inhibitor (PKI), Mos RNA (Mos), cyclin B1 RNA (Cy) and Δ87cyclin B1 protein (CyP; Fig. 7 A). Thapsigargin treatment enhanced the rate of oocyte maturation to a similar extent in oocytes activated with progesterone or PKI (Fig. 7 A, PKI). Thapsigargin-treated oocytes activate faster than controls after Mos RNA injection, but the effect on Ca²⁺ deprivation on the rate of maturation is smaller than in the case of progesterone (Fig. 7 A, Mos). Because the kinetics of the kinase cascade downstream of Mos are similar in control and thapsigargin-treated oocytes (Fig. 7 C; Fig. 5), we wondered whether the faster maturation rate in thapsigargin-treated Mos-injected oocytes is due to an effect of Ca²⁺ deprivation on RNA translation. To determine whether this is the case we induced meiosis by directly activating MPF through the expression of cyclin B1 to activate the free cdc2 pool in the oocyte. Similar to Mos RNA injection, thapsigargin-treated oocytes activated faster than controls after cyclin B1 RNA injection (Fig. 7 A, Cy). In contrast, injecting oocytes with cyclin B1 protein induces GVBD with a similar time course in both thapsigargin-treated and control oocytes (Fig. 7 A, CyP). The rate of oocyte maturation with the different activators is summarized in Fig. 7 B. Because the rate of maturation varies between activators we normalized maturation rate for each activator to the rate of activation in N-Ca (Fig. 7 B) to allow a better visualization of the relative effect of Ca²⁺ deprivation on maturation. For example, although cyclin B activates maturation faster than Mos (Fig. 7 A), the relative enhancement in the rate of maturation (∼20% faster) is similar between the two activators in the absence of Ca²⁺ signals (Fig. 7 B). This argues that the more rapid maturation in oocytes deprived of Ca²⁺ signals after both Mos and cyclin RNA injections is due to an effect of Ca²⁺_cyt on translation of the injected RNAs. This conclusion is supported by the fact that control and thapsigargin-treated oocytes mature at similar rates when activated with Δ87cyclin B1 protein. However, the different responses after cyclin RNA or protein injections could be due to an effect of Ca²⁺_cyt on protein turnover because we injected full-length cyclin B1 RNA and Δ87cyclin B1 protein, which is missing the first 87 aa and is thus nondegradable because it lacks the destruction box (Kumagai and Dunphy, 1995). Nonetheless, we favor an effect of Ca²⁺_cyt on RNA translation because maturation is enhanced to a similar level when oocytes are activated with Mos or Cyclin B1, two activators that induce the cell cycle kinase cascade at different points.

These data show that Ca²⁺_cyt negatively regulates meiosis entry by acting on at least two sites between PKA inhibition and Mos activation. One site is downstream of PKA inhibition and the other site appears to be mRNA translation, which is required for the induction of the cell cycle machinery (Fig. 7 D). Furthermore, the fact that the rate of maturation is enhanced to a similar extent in oocytes activated with progesterone and PKI (Fig. 7, A and B, PKI) argues that Ca²⁺_cyt acts downstream of PKI (Fig. 7 D).

To confirm that Ca²⁺_cyt acts upstream of Mos we analyzed in more details the steps of the cell cycle machinery downstream of Mos in both control and thapsigargin-treated oocytes (Fig. 7 C). As described above MAPK activates significantly earlier in thapsigargin-treated oocytes consistent with the GVBD time course (Fig. 7 C). p90RSK, the downstream substrate of MAPK is activated with a similar time course to MAPK. For these experiments we analyzed lysates form oocytes at GVBD and at GVBD₅₀. For the latter time point lysates from oocytes that have undergone GVBD (w) or not (nw) were collected. Interestingly, p90RSK was phosphorylated to higher levels in the GVBD₅₀-nw group in thapsigargin-treated oocytes as compared with controls, thus confirming the earlier activation of the MAPK cascade in thapsigargin-treated oocytes.

Xenopus oocytes contain two pools of cdc2, the catalytic subunit of MPF: the preMPF and the free cdc2 pool. The preMPF pool which is activated at GVBD, contains cdc2 associated with cyclin B. Pre-MPF is kept inactive by phosphorylation on Tyr15 of cdc2 (Nebreda and Ferby, 2000). The free cdc2 pool is activated after association with B-type cyclins synthesized during meiosis I (Hochegger et al., 2001). To determine which pool of cdc2 is activated in thapsigargin-treated oocytes we probed Western blots with a phosphospecific antibody against Tyr15 of cdc2 (Fig. 7 C, P-Y15-cdc2), and determined MPF activity as the H1-kinase activity from p13^suc1 pulldowns (Fig. 7 C, MPF). In both control and thapsigargin-treated oocytes the disappearance of P-Y15 immunoreactivity coincides with an increase in MPF activity indicating that as in control oocytes the preMPF pool is activated in thapsigargin-treated oocytes. Interestingly, a small level of MPF activity is detected at the GVBD₅₀ time point in thapsigargin-treated oocytes that have not undergone GVBD (Fig. 7, Thaps, G50 nw), confirming the early activation of MPF before GVBD observed in Fig. 5.

These data show that in the absence of Ca²⁺_cyt signals the cell cycle kinase cascade downstream of Mos activates normally. Therefore, the more rapid entry into meiosis (GVBD) observed in oocytes deprived of Ca²⁺ signals is due to a negative regulation of Ca²⁺_cyt on the initiation of this kinase cascade upstream of Mos. Blocking Ca²⁺_cyt signals relieves this negative regulation thus allowing more rapid induction of the MAPK-MPF cascade and GVBD.

In contrast to the established role of Ca²⁺ signaling in mitosis (Whitaker and Larman, 2001), the requirement for Ca²⁺ in both Xenopus and mammalian oocyte meiotic maturation has been difficult to define (Homa et al., 1993; Duesbery and Masui, 1996). To delineate the function of Ca²⁺ during Xenopus oocyte meiosis we manipulated Ca²⁺_cyt, extracellular Ca²⁺, and store Ca²⁺ load and tested the effect on nuclear maturation and the cell cycle machinery. Our data show that Ca²⁺ has two opposing roles during Xenopus oocyte maturation: It negatively regulates meiosis entry by delaying the activation of the cell cycle machinery, and it is required for completion of meiosis I (Fig. 7 D).

Progesterone leads to lower cAMP levels and PKA inhibition within 10 min (Sadler and Maller, 1981), but the next known step in the pathway, that is polyadenylation of maternal RNAs to induce their translation, does not occur until much later (Sheets et al., 1995). The molecular steps during this time lag are not known. Our data show that Ca²⁺_cyt is an important regulator of the transition between PKA inhibition and mRNA translation. Ca²⁺_cyt negatively regulates the activation of the cell cycle machinery by acting on at least two sites between PKA and Mos (Fig. 7 D). One site of action appears to be mRNA translation. Therefore, the level of Ca²⁺_cyt provides a timing mechanism for entry into meiosis by regulating the initiation of the MAPK-MPF cascade downstream of PKA inhibition (Fig. 7 D). It is tempting to propose that by acting in this capacity Ca²⁺_cyt could synchronize morphological and biochemical changes during oocyte maturation. Under such a scenario, which is completely speculative at this point, Ca²⁺_cyt levels could signal the physiological preparedness of the oocyte to begin maturation. Relatively low Ca²⁺_cyt levels would be indicative of proper functioning of the Ca²⁺ signaling machinery, and thus a healthy oocyte that is ready to mature. In contrast, relatively high Ca²⁺_cyt levels would indicate a compromised oocyte where Ca²⁺_cyt would negatively regulate initiation of maturation. It is interesting in that context that Ca²⁺_cyt acts upstream of Mos, that is before the oocyte activates the cell cycle machinery and commits to maturation. The proposed role of Ca²⁺_cyt in synchronizing morphological and biochemical changes in the oocyte during maturation is further supported by the fact that disrupting Ca²⁺_cyt signaling uncouples the nuclear cell cycle from the MAPK-MPF kinase cascade (Fig. 6).

Oocytes deprived of Ca²⁺ signals do not complete meiosis I as they do not extrude a polar body. Rather, they form abnormal spindles early in meiosis I despite normal MAPK and MPF kinetics. This shows that progression through meiosis I requires Ca²⁺, possibly Ca²⁺ influx before GVBD because oocytes are dependent on extracellular Ca²⁺ only before GVBD (Fig. 6). A Ca²⁺ influx requirement before GVBD fits nicely with the regulation of SOCE during oocyte maturation because SOCE inactivates at the GVBD stage due to MPF activation (Machaca and Haun, 2000, 2002).

Interestingly, others have shown that inhibition of the MAPK cascade (Gross et al., 2000), down-regulation of MPF (Nakajo et al., 2000), or inhibition of protein synthesis (Kanki and Donoghue, 1991) block completion of meiosis I. These treatments lead to a decrease in MPF activity, and induce an interphase-like state that is usually absent between meiosis I and II. In contrast, interfering with Ca²⁺ signaling blocks meiosis I completion, but is not associated with an interphase-like state. Rather, spindle structure is disrupted and polar body formation inhibited, but the chromosomes remain condensed. Furthermore, MPF activity cycles normally with a dip in activity between 1–2 h after GVBD, the expected time for meiosis I to meiosis II transition. These data show that disruption of Ca²⁺ signaling uncouples the cell cycle machinery (MAPK-MPF) from nuclear maturation (i.e., bipolar spindle formation and completion of meiosis I).

It is interesting that the disrupted meiosis I spindle does not activate a spindle checkpoint to arrest the cell cycle. However, there is good evidence against the existence of a meiosis I spindle checkpoint in Xenopus oocytes. Blocking the activity of the APC/C or the checkpoint protein Mad2 does not affect progression through meiosis I (Peter et al., 2001; Taieb et al., 2001). This is consistent with our observation of a lack of cell cycle arrest in the absence of Ca²⁺ signals despite disrupted meiosis I spindles.

Recently, Castro et al. (2003) described a similar block of meiosis I after inhibition of Aurora A kinase or its substrate Eg5 (a kinesin-like protein) in Xenopus oocytes. Unfortunately, these authors did not assess spindle morphology, but they show that blocking Aurora A inhibits polar body formation with chromosomes maintaining their condensed state long after GVBD (Castro et al., 2003). Members of the Aurora kinase family associate with the spindle and have been shown to be important for both meiosis and mitosis transitions. Therefore, it is possible that Ca²⁺-dependent pathways somehow modulate Aurora A kinase activity which in turn regulates spindle structure. This possibility remains to be explored.

During mitosis NEBD has been shown to be dependent on Ca²⁺ (Poenie et al., 1985; Steinhardt and Alderton, 1988; Twigg et al., 1988; Kao et al., 1990; Wilding et al., 1996), and studies in sea urchin embryos suggest that Ca²⁺ exerts its effect through CaMKII activation (Baitinger et al., 1990). In contrast, during meiosis GVBD is Ca²⁺-independent as shown here for Xenopus oocytes, and in both mouse (Carroll and Swann, 1992; Tombes et al., 1992) and starfish oocytes (Witchel and Steinhardt, 1990). One exception to this rule are some bivalve molluscs where GVBD has been shown to require Ca²⁺ (Deguchi and Osanai, 1994), but unlike amphibian and mammalian oocytes, in this case oocyte maturation and GVBD occur after fertilization which invariably induces a Ca²⁺_cyt rise. Nonetheless, the differential requirement of NEBD on Ca²⁺ signals during meiosis and mitosis is surprising because both NEBD and GVBD require the activation of MPF (Lenart and Ellenberg, 2003), and it is reasonable to assume that the basic structural properties of the nuclear envelope are similar in mitotic and meiotic cells. It has been argued that a Ca²⁺ signal is still required for GVBD but occurs very early or even before the initiation of oocyte maturation in some species (Tombes et al., 1992; Homa et al., 1993). This does not seem to be the case for Xenopus, because as shown in Fig. 1 F, eliminating Ca²⁺ signals for as long as 48 h before inducing oocyte maturation has no effect on GVBD, strongly arguing that GVBD is Ca²⁺ independent. Therefore, the differential requirement for Ca²⁺ during the breakdown of the nuclear envelope suggests that NEBD and GVBD are mechanistically distinct.

In conclusion, our results show that Ca²⁺ signals are dispensable for GVBD and chromosome condensation, that Ca²⁺_cyt controls the timing of meiosis entry by negatively regulating the initiation of cell cycle machinery, and that N-Ca²⁺ homeostasis is important for bipolar spindle formation and completion of meiosis I. These results provide a framework to further explore and better define the role of Ca²⁺-dependent signaling pathways in meiosis and oocyte maturation.

Xenopus oocytes were obtained as described previously (Machaca and Haun, 2002). The control l-15 solution contains 0.63 mM Ca²⁺. Ca²⁺ was buffered at 50 μM in the low solution as calculated using the MaxChelator program (http://www.stanford.edu/~cpatton/maxc.html) by the addition of 0.58 mM EGTA. For the H-Ca l-15 solutions Ca²⁺ was added to the indicated concentration as CaCl₂. In all experiments, GVBD was visually confirmed by fixing oocytes in methanol and bisecting them in half.

Recording of the I_Ca,Cl was performed as described previously (Machaca and Haun, 2000). I_Ca,Cl were recorded in the following solutions: F-Ca contains in micromolars: 96 NaCl, 2.5 KCl, 5 MgCl₂.6H₂O, 0.1 EGTA, 10 Hepes, pH 7.4. Low Ca²⁺ Ringer solution (50 μM free Ca²⁺) contains in micromolars: 96 NaCl, 2.5 KCl, 4.37 MgCl₂.6H₂O, 0.63 CaCl₂.2H₂O, 0.58 EGTA, 10 Hepes, pH 7.4. Normal Ringer (N-Ca, 0.63 mM Ca²⁺) and H-Ca solutions (1.5, 3, and 5 mM Ca²⁺) had the following composition: 96 NaCl, 2.5 KCl, 10 Hepes, pH 7.4; with Ca²⁺ and Mg²⁺ concentrations adding up to 5 mM.

MPF kinase activity assay and phospho-MAPK Western blots were prepared as described previously (Machaca and Haun, 2002), except that α-tubulin was used as the loading control for MAPK Western blots. The activation of p90RSK and preMPF was assessed using phosphospecific antibodies against phospho-Thr573 of p90RSK and phospho-Tyr15 of cdc2 (Cell Signaling). MPF activity was also assayed from lysates affinity purified on p13^suc1 beads using histone-H1 kinase as a substrate essentially as described previously (Howard et al., 1999).

Heat stable PKI was purchased from Calbiochem. Cyclin B1 and Mos RNAs were synthesized from a pXen-GST-Mos and pSP64-cyclinB1^xen plasmids provided by A. Macnicol (University of Arkansas for Medical Sciences; Freeman et al., 1991; Howard et al., 1999) using the mMessage Mmachine transcription kit (Ambion). The His₆-tagged Δ87cyclin B1 protein was used as described previously (Machaca and Haun, 2002).

Oocytes were fixed in 100% methanol, bisected in half, and incubated in DM1A an antitubulin mAb (Sigma-Aldrich) in TBS containing 2% BSA, followed by a Cy2-conjugated donkey anti–mouse secondary (Jackson ImmunoResearch Laboratory) for 24 h each. The oocytes were washed, dehydrated, stained in 1 μM Sytox^® orange (Molecular Probes), and cleared in benzyl alcohol/benzyl benzoate (1:2). Images were collected using a Fluoview confocal (Olympus) coupled to a microscope (model IX70; Olympus) at RT, using a UPlanApo 40× oil objective with an NA of 1.00. The acquisition software was Fluoview 2.1 and figures were compiled using Adobe Photoshop 7.0. For each spindle a z section was obtained and projected onto a single plane to visualize the entire spindle. For polar body emission studies oocytes were fixed in methanol, stained with Sytox^® orange, and visualized by confocal microscopy as described for the spindle staining.

We thank members of the Machaca Lab for critical reading of the manuscript, A. Macnicol for providing plasmid constructs, and A. Charlesworth for help with the p13^suc1 MPF assay.

This work was supported by grant GM-61829 from the National Institutes of Health.