α-endosulfine (ENSA) regulates exit from prophase I arrest in mouse oocytes

Abstract

Mammalian oocytes in ovarian follicles are arrested in meiosis at prophase I. This arrest is maintained until ovulation, upon which the oocyte exits from this arrest, progresses through meiosis I and to metaphase of meiosis II. The progression from prophase I to metaphase II, known as meiotic maturation, is mediated by signals that coordinate these transitions in the life of the oocyte. ENSA (α-endosulfine) and ARPP19 (cAMP-regulated phosphoprotein-19) have emerged as regulators of M-phase, with function in inhibition of protein phosphatase 2A (PP2A) activity. Inhibition of PP2A maintains the phosphorylated state of CDK1 substrates, thus allowing progression into and/or maintenance of an M-phase state. We show here ENSA in mouse oocytes plays a key role in the progression from prophase I arrest into M-phase of meiosis I. The majority of ENSA-deficient oocytes fail to exit from prophase I arrest. This function of ENSA in oocytes is dependent on PP2A, and specifically on the regulatory subunit PPP2R2D (also known as B55δ). Treatment of ENSA-deficient oocytes with Okadaic acid to inhibit PP2A rescues the defect in meiotic progression, with Okadaic acid-treated, ENSA-deficient oocytes being able to exit from prophase I arrest. Similarly, oocytes deficient in both ENSA and PPP2R2D are able to exit from prophase I arrest to an extent similar to wild-type oocytes. These data are evidence of a role for ENSA in regulating meiotic maturation in mammalian oocytes, and also have potential relevance to human oocyte biology, as mouse and human have genes encoding both Arpp19 and Ensa.

Introduction

The progression of the mammalian oocyte through meiosis is carefully regulated. This regulation is essential to ensure the coordination of cell cycle progression (also known as nuclear maturation) with a series of events that mediate the acquisition of competence to undergo the egg-to-embryo transition upon fertilization (also known as egg activation) and to progress into embryogenesis.^1-4 Mammalian female meiosis occurs in a very staggered fashion. The oocyte in the ovarian follicle is arrested at prophase of meiosis I; this arrest occurs before birth and lasts for weeks to years, depending on the species. Upon ovulation, the oocyte undergoes a transition known as meiotic maturation (also called oocyte maturation),⁵ characterized by exit from prophase I arrest and progression through meiosis I. This process, particularly prometaphase and metaphase, takes several hours. The oocyte then emits the first polar body and proceeds to an arrest at metaphase II. The second meiotic division occurs after this metaphase II arrest if fertilization occurs; the fertilizing sperm triggers exit from M-phase, progression to the first embryonic mitosis, and other events of the egg-to-embryo transition.

The phosphorylation status of key substrates of cyclin-dependent kinase 1 (CDK1; also known as CDC2 or CDC2A) is crucial for the M-phase state. Active CDK1, associated with its regulatory subunit cyclin B, phosphorylates various substrates, leading to events associated with entry into M phase. Inhibition of the dephosphorylation of CDK1 substrates during M phase is critical, and then the dephosphorylation of these CDK1 substrates is required for exit from M phase, allowing a “reset” back to an interphase state. Such regulation of M-phase phosphoproteins is likely highly significant for mammalian oocytes, which have to regulate extended metaphase states during meiosis I and II, with these multi-hour M-phase states sharply contrasting the minutes-long M phase in the typical somatic cell.

In mammalian oocytes, meiotic resumption from prophase I arrest is analogous to a G₂-to-M transition in mitotic cells. CDK1 activity is low in prophase I-arrested oocytes as a result of protein kinase A (PKA) action on the kinases MYT1 and WEE1 (in the mouse, previously WEE1B, now known as WEE2) and the phosphatase CDC25 (CDC25B in mouse).^6-9 WEE2 and MYT1 phosphorylate CDK1 on 2 inhibitory residues, T14 and Y15. Inhibition of CDC25 by PKA reinforces this inactive state of CDK1.⁷ Also downstream from PKA is the phosphatase CDC14B and the anaphase-promoting complex/cyclosome (APC/C), with functions in maintaining prophase I arrest by regulating cyclin B levels.^9-14 Exit from prophase I is mediated by decreased PKA activity, contributing to activation of CDC25B, which, in turn, dephosphorylates CDK1. Active CDK1 then phosphorylates its M-phase substrates, and progression into M phase is observed with the occurrence of breakdown of the nuclear envelope of the germinal vesicle (known as GVBD).

A key phosphatase involved in the dephosphorylation of M-phase substrates is PP2A, as identified by immunodepletion studies in Xenopus egg extracts and genetic studies in Drosophila.^15,16 PP2A refers to a broad class of heterotrimeric holoenzymes, composed of a catalytic/C subunit, a scaffold/A subunit, and a regulatory/B subunit.¹⁷ The kinase known as MASTL or Greatwall is at the intersection between CDK1 and PP2A. (Note: MASTL [microtubule-associated serine/threonine kinase-like] is the official name in mouse [MGI:1914371] and human [HGNC:19042]; this kinase is Greatwall in Drosophila, and Greatwall is the name used in most Xenopus studies. We attempt to use pan-species terminology in most places in this document, although in some cases, the mouse/MGI gene and protein symbols are used as a default.) Immunodepletion studies in Xenopus egg extracts identified the specific PP2A form that is inhibited by MASTL/Greatwall as PP2A with the regulatory/B subunit B55δ (MGI symbol, PPP2R2D),^18-21 consistent with data showing that Drosophila mutants lacking this regulatory/B subunit of PP2A have low phosphatase activity toward certain CDK1 substrates.¹⁵ Taken together, these data have produced the model that MASTL/Greatwall activity, through its inhibition of PP2A activity, contributes to maintenance of phosphorylated M-phase substrates.^19,20,22,23

MASTL/Greatwall achieves this inhibition of PP2A through intermediary proteins, ENSA (α-endosulfine), and ARPP19 (cyclic adenosine monophosphate-regulated phosphoprotein 19) (Fig. 1). ENSA and ARPP19 are substrates of MASTL/Greatwall.^24,25 The phosphorylated forms of these MASTL/Greatwall substrates bind to PP2A-B55δ/PPP2R2D, inhibiting PP2A-mediated dephosphorylation of M-phase phosphoproteins (Fig. 1).^24,25 In fact, based on this, MASTL/Greatwall, in addition to CDK1 and cyclin B, has been proposed to be a component of M phase-promoting factor (MPF) activity.²⁶ This pathway is conserved in a wide range of organisms. The related Drosophila protein Endos is a substrate of Greatwall,^27,28 and the starfish Patiria pectinifera has a similar, single ENSA/ARPP19 ortholog.²⁶ Saccharomyces cerevisiae also has components of this system. The yeast endosulfines Igo1 and Igo2 are substrates of Rim15, the yeast MASTL/Greatwall; this Rim15-Igo1/2 pathway regulates entry into G₀.^29-31 Additionally, the yeast proteins Zds1 and Zds2 (unrelated to Igo1/2) function as inhibitors of PP2A.^32-35

Figure 1. Regulation of cell cycle progression by the MASTL/Greatwall substrates ENSA and ARPP19. Schematic diagram illustrating fundamental aspects of the functions of MASTL (also known as Greatwall), and ENSA and ARPP19 in cell cycle regulation. In interphase, CDK1 is inactive, and levels of M-phase phosphoproteins are low, whereas the phosphorylated status of these CDK1 substrates needs to be established and maintained for M phase. This is facilitated by inhibition of dephosphorylation of these CDK1 substrates, which is achieved by MASTL phosphorylated ENSA and/or ARPP19, which can bind to and inhibit to the phosphatase PP2A.^37,70-72

This work on mammalian meiotic maturation addresses aspects of this model, building on work in non-mammalian species, and puts the model in context with more recent findings in mammalian cells. Studies presented here address the fundamental questions of whether ENSA has a role in murine female meiosis, and, additionally, if ENSA is a physiologically relevant protein in cell cycle regulation in this cellular context and in a species that has the 2 MASTL/Greatwall substrates. Specifically, we tested the hypothesis that mouse ENSA would play a role in some aspect(s) of regulation of meiotic M phase in oocytes. This is of significance to the field, as there are some differences in data regarding which protein, ENSA or ARPP19, is phosphorylated by Greatwall/MASTL in various types of Xenopus egg extracts, and on which protein functions in cells.^24,25,36-38 Results on the effects of depletion of ARPP19 or ENSA differ in studies using HeLa cells.^25,38 In the context of oocyte meiotic maturation, ARPP19 has been implicated as a key regulator in Xenopus oocytes,³⁶ as has Endos in Drosophila oocytes.^28,39 The kinase MASTL/Greatwall has functions in starfish and porcine oocytes,^26,40 although these species only have one MASTL/Greatwall substrate protein, as does Drosophila. (The pig genome has genes encoding Ensa and Arpp19, but the Ensa gene homolog is a non-functional pseudogene, with a premature stop codon.⁴⁰) The data presented here on mouse oocyte meiotic maturation contribute to this body of knowledge, as well as address some key issues in the field, having revealed a function for ENSA in mammalian oocyte meiotic maturation in a specific pathway with PP2A with the regulatory subunit PPP2R2D/B55δ.

Results

Immunoblotting was used to assess the specificity of the anti-ENSA antibody used in these studies (Fig. 2A). This anti-ENSA antibody cross-reacted with recombinant mouse ENSA (GST-ENSA; Fig. 2A, lanes 1–3) and did not cross-react with recombinant mouse ARPP19 (GST-ARPP19; Fig. 2A, lanes 4–6); mouse ARPP19 has 76% amino acid identity with mouse ENSA. This anti-ENSA antibody detected ENSA in lysates of prophase I mouse oocytes by immunoblotting (Fig. 2B, lanes 1–2; lysates of 26 and 40 oocytes, respectively), as well as positive control purified GST-ENSA (Fig. 2B, lanes 3–5). We also were eager to undertake similar analyses of ARPP19 expression in oocytes, but we thus far have been unable to find an anti-ARPP19 antibody that is specific for ARPP19 and does not cross-react with ENSA (data not shown).

Figure 2. ENSA is expressed by mouse oocytes. (A) Anti-ENSA immunoblot of 25 ng, 50 ng, and 75 ng of GST-ENSA (lanes 1–3) and GST-ARPP19 (lanes 4–6). This anti-ENSA antibody appears to be specific for mouse ENSA and did not cross-react with mouse ARPP19 under these conditions. (B) Anti-ENSA immunoblot of lysates of 26 and 40 mouse oocytes (lanes 1–2) alongside positive control samples of 30 ng, 40 ng, and 50 ng of GST-ENSA (lanes 3–5). The lower M_r band in the GST-ENSA sample is likely either a product of proteolysis or incomplete translation.

To assess the function of ENSA in mouse oocytes, we utilized RNAi-mediated knockdown to generate Ensa-deficient oocytes. Oocytes were microinjected with negative control siRNA (hereafter referred to as “control oocytes”) or with Ensa-targeting siRNA, and then cultured in conditions to maintain prophase I arrest, allowing RNAi-mediated knockdown and protein turnover to occur. Following a 44–48 h culture period, Ensa mRNA levels were decreased by 67 ± 11% in Ensa siRNA-injected oocytes compared with controls (5 experiments; range 39–91%; Fig. 3A). Arpp19 mRNA levels were unaffected by the Ensa-targeting siRNA (Fig. 3B), in agreement with our BLAST searches revealing that the Ensa-targeting siRNA did not have any regions of homology to Arpp19. ENSA protein levels were decreased by 80 ± 6% in Ensa siRNA-injected oocytes as compared with controls, as detected by immunoblotting (5 experiments; range 71–97%; Fig. 3C). In testing this anti-ENSA antibody in immunofluorescence experiments, we observed staining of prophase I oocytes, and this staining was largely lost in oocytes incubated with anti-ENSA antibody that was pre-incubated with a 20-fold molar excess of GST-ENSA (Fig. S1), but oocytes that were injected with the Ensa-targeting siRNA still were labeled by the anti-ENSA antibody (data not shown). With this result raising questions about the utility of this particular antibody in immunofluorescence applications, we did not pursue immunofluorescence studies further here.

Figure 3. Ability to exit from prophase I arrest is impaired in ENSA-deficient oocytes. Prophase I-arrested oocytes were injected with 20 μM negative control siRNA or Ensa-targeting siRNA and cultured for 44–48 h in the presence of dbcAMP to maintain prophase I arrest and allow time for knockdown. (A) RT-PCR analysis of Ensa transcript levels. Lanes 1–2 show the positive control PCR product (Plat) for negative control siRNA-injected oocytes (lane 1) and Ensa siRNA-injected oocytes (lane 2). Lanes 3–4 show the Ensa PCR product for negative control siRNA-injected oocytes (lane 3) and Ensa siRNA-injected oocytes (lane 4). Ensa mRNA levels were decreased by 67 ± 11% in Ensa siRNA-injected oocytes compared with controls (5 experiments, with one representative experiment shown here; range 39–91%). (B) RT-PCR analysis of Arpp19 transcript levels. Lanes 1–2 show the positive control Plat PCR product for negative control siRNA-injected oocytes (lane 1) and Ensa siRNA-injected oocytes (lane 2). Lanes 3–4 show the Arpp19 PCR product for negative control siRNA-injected oocytes (lane 3) and Ensa siRNA-injected oocytes (lane 4). (C) Anti-ENSA immunoblot of negative control siRNA-injected oocytes (lane 1) and Ensa siRNA-injected oocytes (lane 2) (20 oocytes per lane). GST-ENSA (20 ng) was loaded in lane 3 as a positive control. ENSA protein levels were decreased by 80 ± 6% in Ensa siRNA-injected oocytes as compared with controls (5 experiments, with one representative experiment shown here; range 71–97%). (D) Assessment of germinal vesicle breakdown in negative control siRNA-injected oocytes (n = 149 oocytes; gray triangles) and Ensa siRNA-injected oocytes (n = 174 oocytes; open circles). These data are pooled from four replicate experiments, and error bars show standard errors of the mean.

ENSA-deficient oocytes were examined for the ability to progress through meiotic maturation. To assess exit from prophase I arrest, oocytes were placed in culture conditions that support progression from prophase I arrest (i.e., absence of dbcAMP) and were observed at 30 min intervals over a 5 h culture period for germinal vesicle breakdown (also called nuclear envelope breakdown). Over this time period, an average of 60% of control oocytes underwent germinal vesicle breakdown compared with only 23% of ENSA-deficient oocytes (Fig. 3D). These data suggested that ENSA deficiency impaired the ability of mouse oocytes to exit from prophase I arrest.

To assess the meiotic status of these oocytes in more detail, we examined lamin B immunofluorescence staining of Ensa siRNA-injected and control oocytes after the 5 h of culture in conditions that support exit from prophase I arrest (Fig. 4A). After this 5 h of culture, oocytes from the negative control siRNA and Ensa siRNA experimental groups were sorted into 2 classifications for these immunofluorescence studies: prophase I (based on the presence a germinal vesicle) or pro-metaphase I (based on the absence of a germinal vesicle). In the oocytes classified as prophase I, anti-lamin B immunofluorescence showed strong labeling of the nuclear lamina, and no apparent differences were detected between Ensa siRNA-injected (Fig. 4A, panels vii–ix) and control (Fig. 4A, panels i–iii) oocytes. While the majority of Ensa siRNA-injected oocytes were arrested at the prophase I stage after 5 h (average 77%, Fig. 3D), a subset of Ensa siRNA-injected oocytes did show evidence of germinal vesicle loss. Normal nuclear lamina breakdown and chromosome condensation was observed in these Ensa siRNA-injected oocytes (Fig. 4A, panels x–xii), with prometaphase I control oocytes having very similar appearances (Fig. 4A, panels iv–vi). Thus, the subset of Ensa siRNA-injected oocytes that exited from prophase I arrest appeared comparable to control oocytes at this stage, able to progress normally from prophase I arrest into meiosis I.

Figure 4. Progression through meiosis in control and Ensa siRNA-injected oocytes. (A) Immunofluorescence analysis of nuclear lamina in oocytes fixed at 5 h after initiation of meiotic maturation (i.e., removal of dbcAMP). After this 5 h of culture, oocytes from the negative control siRNA and Ensa siRNA experimental groups were separated into 2 classifications: prophase I (based on the presence a germinal vesicle) or pro-metaphase I (based on the absence of a germinal vesicle). As shown in Figure 3 and addressed in the text, while the majority of Ensa siRNA-injected oocytes were arrested at the prophase I stage after 5 h, a subset of oocytes showed evidence of germinal vesicle loss. Images show prophase I oocytes from the control group (panels i-iii) and the Ensa siRNA-injected group (panels vii-ix) and pro-metaphase I oocytes from the control group (panels iv-vi) and the Ensa siRNA-injected group (panels x-xii) stained with an anti-lamin B antibody (ii, v, viii, xi) and DAPI (iii, vi, ix, xii) to label DNA. Scale bar (shown in Panel i) = 10 μm. (B) Meiotic progression at 14 h after initiation of meiotic maturation (i.e., removal of dbcAMP). Control and Ensa siRNA-injected oocytes were classified as being prophase I (ProI, based on the presence a germinal vesicle; open bar), GVBD/no PB1 (based on the absence of a germinal vesicle and first polar body; gray bar), or having the first polar body (PB1; black bar). (C) Meiotic progression at 14 h after initiation of meiotic maturation in the subset of oocytes that exited from prophase I arrest. Control and Ensa siRNA-injected oocytes were classified as being GVBD/no PB1 (absence of a first polar body; gray bar), or having the first polar body (PB1; black bar). (D) Immunofluorescence analysis of the metaphase II spindle in oocytes fixed at 14 h after initiation of meiotic maturation. Images show oocytes from the control group (panels i-iii) and the Ensa siRNA-injected group (panels iv-vi) stained with an anti-α-tubulin antibody (ii, v) and DAPI (iii, vi) to label DNA. Scale bar (shown in panel i) = 10 μm.

A separate series of experiments assessed completion of meiosis I (i.e., emission of the first polar body) and progression to metaphase II arrest in control and Ensa siRNA-injected oocytes (Fig. 4B–D). At 14 h after initiation of meiotic maturation (i.e., culture in conditions that support exit from prophase I arrest), 76% (81/106) of control oocytes and 33% (34/104) of Ensa siRNA-injected oocytes had exited from prophase I arrest and undergone GVBD (Fig. 4B). Of this subset that exited from prophase I arrest, 56% (45/81) of control oocytes and 50% (17/34) of ENSA-deficient oocytes progressed through meiosis I, as evidenced by the presence of the first polar body (Fig. 4C). These oocytes that emitted the first polar body were confirmed to have metaphase II spindles, with comparable spindle morphology and chromosome alignment in both the negative control siRNA-injected oocytes and the Ensa siRNA-injected oocytes (Fig. 4D). There was no evidence of parthenogenetic exit from metaphase II arrest during this in vitro culture period in oocytes injected with either the negative control siRNA or the Ensa siRNA.

To test the hypothesis that ENSA was functioning to inhibit PP2A activity in mouse oocytes, we first used Okadaic acid treatment to inhibit PP2A activity in control and Ensa siRNA-injected oocytes, assessing the abilities of these oocytes to exit from prophase I arrest (Fig. 5A). As also shown above (Fig. 3D), Ensa siRNA-injected oocytes were impaired in their ability to exit from prophase I arrest as compared with controls. However, Ensa siRNA-injected oocytes that were treated with 2.5 µM Okadaic acid progressed out of prophase I arrest to extents comparable to control oocytes (Fig. 5A).

Figure 5. Suppression of PP2A activity restores the ability of Ensa siRNA-injected oocytes to exit from prophase I arrest and progress through meiosis I. (A) Assessment of germinal vesicle breakdown in negative control siRNA-injected oocytes, untreated (gray triangles) and treated with 2.5 μM Okadaic acid (black squares), and Ensa siRNA-injected oocytes, untreated (open circles) and treated with 2.5 μM Okadaic acid (open squares). These data are pooled from 2 replicate experiments (63–77 total oocytes per experimental group), and error bars show the high and low values for these 2 experiments. (B) RT-PCR analysis of Ppp2r2d transcript levels. Lanes 1–2 show the positive control Plat PCR product for negative control siRNA-injected oocytes (lane 1) and Ppp2r2d siRNA-injected oocytes (lane 2). Lanes 3–4 show the Ppp2r2d PCR product for negative control siRNA-injected oocytes (lane 3) and Ppp2r2d siRNA-injected oocytes (lane 4). Ppp2r2d mRNA levels were decreased by 80 ± 5% in Ppp2r2d siRNA-injected oocytes as compared with controls (3 experiments, with one representative experiment shown here; range 73–90%). (C) Assessment of germinal vesicle breakdown in negative control siRNA-injected oocytes (gray triangles), Ensa siRNA-injected oocytes (open circles), Ppp2r2d siRNA-injected oocytes (open squares), and Ensa + Ppp2r2d siRNA-injected oocytes (black diamonds). These data are pooled from 3 replicate experiments (115–128 oocytes per experimental group), and error bars show standard errors of the mean.

This result was rationale for more detailed examination of PP2A. We targeted the PP2A B55δ regulatory subunit, Ppp2r2d, for siRNA-mediated knockdown in mouse oocytes. Following a 44–48 h culture period to allow for knockdown, Ppp2r2d mRNA levels were decreased by 80 ± 5% in Ppp2r2d siRNA-injected oocytes as compared with controls (3 experiments; range 73–90%; Fig. 5B). These Ppp2r2d-siRNA-injected oocytes did not exit prophase I arrest during this 44–48 h culture period in conditions that support maintenance of prophase I arrest (i.e., inclusion of 0.25 mM dbcAMP in the culture medium). Considering that prophase I oocytes treated with or injected with Okadaic acid will exit from prophase I arrest even in the presence of agents that maintain high protein kinase A activity,^41-43 this result suggested that the PPP2R2D/B55δ regulatory subunit is not the only PP2A regulatory subunit required for the G₂–M progression in mouse oocytes.

Subsequent experiments assessed whether the Ppp2r2d deficiency would rescue the ability of Ensa-siRNA-injected oocytes to exit from prophase I arrest. As shown above (Fig. 3D), Ensa-siRNA-injected oocytes in these experiments showed reduced exit from prophase I arrest as compared with controls (Fig. 5C). Ppp2r2d-deficient oocytes exited from prophase I arrest to extents similar to control oocytes. Most significantly, the extent of germinal vesicle breakdown was similar to controls in oocytes injected with both Ensa-targeting and Ppp2r2d-targeting siRNA. Thus, the ability of Ensa-siRNA-injected oocytes to exit from prophase I arrest was rescued when Ppp2r2d was also knocked down, indicating that ENSA acts in a pathway with PP2A containing the PPP2R2D/B55δ regulatory subunit during meiosis I in mouse oocytes.

Discussion

This study identifies a physiological role for ENSA in mouse oocytes, with knockdown of ENSA in mouse oocytes, resulting in defective progression out of prophase I arrest. The defect in progression into M phase of meiosis I in ENSA-deficient mouse oocytes has similarities to Drosophila oocytes, with endos mutant oocytes having a prolonged prophase I arrest and delayed progression to metaphase I.³⁹ However, differences between Ensa siRNA-injected mouse oocytes and endos mutant Drosophila oocytes should be noted as well. Endos mutant Drosophila oocytes that progressed past prophase I show abnormalities in metaphase I spindle organization and chromosome congression.³⁹ Moreover, Endos deficiency has more deleterious effects on Drosophila female meiosis than in other cell divisions, including male meiosis, leading to speculation that Endos has additional functions in female meiosis in Drosophila.²⁸ In our study, while the majority of Ensa siRNA-injected mouse oocytes remain arrested at the prophase I stage, a subset of Ensa siRNA-injected oocytes undergo normal meiotic maturation, progressing into meiosis I and to the metaphase II arrest stage. An explanation for our results is that there is oocyte-to-oocyte variability in the extent of knockdown of Ensa RNA and/or ENSA protein. A subset of Ensa siRNA-injected oocytes here may have been able to undergo germinal vesicle breakdown, because in these particular oocytes, sufficient amounts of ENSA persisted to support exit from prophase I arrest. We attempted to test this hypothesis and examine knockdown in individual oocytes by immunofluorescence (e.g., see Fig. S1), but found that Ensa siRNA-injected oocytes still showed some labeling with the anti-ENSA antibody (data not shown). Since immunoblotting shows clear knockdown of ENSA protein in Ensa siRNA-injected oocytes (Fig. 3), this result makes us question the utility of this particular anti-ENSA antibody in immunofluorescence applications. An additional possibility is that in the context of ENSA deficiency, this polyclonal anti-ENSA antibody is able to cross-react with ARPP19 (even though this anti-ENSA antibody does not cross-react with purified ARPP19 in immunoblots; Fig. 2).

Demonstration of this functionality of ENSA in mouse oocytes is relevant to discussion of whether ENSA and/or ARPP19 are physiologically relevant substrates for MASTL/Greatwall. This question is obviously not an issue in the species that have only one MASTL/Greatwall substrate, such as Drosophila with Endos. The case has been made, based on data in Xenopus related to protein concentrations, phosphorylation dynamics, and effects of depletion, that ARPP19 is the physiological MASTL/Greatwall substrate.³⁷ Consistent with this model, expression of a non-phosphorylatable form of ARPP19 in Xenopus prophase I oocytes results in defects in exit from prophase I arrest, suggesting that this non-phosphorylatable form of ARPP19 acts as a dominant-negative in this system,³⁶ although it is unknown if this dominant-negative ARPP19 would have any effects on the function of ENSA. Additional studies using injected ARPP19 show how exogenously introduced ARPP19 can affect normal Xenopus oocyte maturation parameters; these actions were characterized in wild-type Xenopus oocytes with endogenous ARPP19 and ENSA.⁴⁴ With regard to the physiological relevance of ENSA vs. ARPP19 in mammalian cells, one study observes that ARPP19 knockdown in HeLa cells produces defects in mitotic progression,²⁵ whereas another study reports that siRNA-mediated depletion of ARPP19 had little effect in one specific molecular aspect of cell cycle progression, and, additionally, that ARPP19 was not detected in HeLa cells.³⁸ This is potentially explained by lab-to-lab differences in HeLa lines, but nonetheless highlights variability in pathways utilizing ENSA or ARPP19. Budding yeast is another system in which there are 2 substrates of the yeast MASTL/Greatwall (Rim15), known as Igo1 and Igo2.^29-31 This pathway in yeast regulates entry into G₀, and deletion of both Igo1 and Igo2 is required to produce a defect in G₀ entry, with igo1Δ and igo2Δ single mutants not having this phenotype.²⁹ Budding yeast has 2 proteins, Zds1 and Zds2, with no sequence homology to Igo1/2 but with functional similarity to ENSA and ARPP19, serving as inhibitors of PP2A.^34,35 Like Igo1 and Igo2, Zds1 and Zds2 also appear to be redundant to each other, with zds1Δ/zds1Δ double mutants, but not zds1Δ or zds1Δ single mutants, showing a defect in G₂ delay.^32,33 This is suggestive of functional specialization between cell types arising in metazoans that have both ENSA and ARPP19.

There is strong evidence for the MASTL/Greatwall pathway functioning in oocytes of multiple species, including Xenopus, Drosophila, starfish, and pig.^{26,28,36,39,40,45-47} Starfish and Drosophila have only one MASTL/Greatwall substrate, and the pig ENSA gene has a premature stop codon, such that porcine cells would not express functional ENSA protein.⁴⁰ Xenopus oocytes may rely solely or primarily on ARPP19 to regulate exit from prophase I arrest,^36,37 while the data here clearly indicate that ENSA is important for exit from prophase I arrest in mouse oocytes. Such differences between species are not unprecedented, with one example in female meiosis being the phosphatase calcineurin. Calcineurin is important for exit from metaphase II arrest in Xenopus, exit from metaphase I arrest in an ascidian species, and progression of Drosophila oocytes through meiosis,^48-50 whereas there is no evidence for calcineurin functioning in exit from metaphase II arrest in mouse.⁵¹

The defect in meiotic resumption with ENSA deficiency in mouse oocytes, analogous to a G₂-to-M transition in mitotic cells, differs from what is observed in mouse embryonic fibroblasts (MEFs) depleted of MASTL through a conditional knockout approach.⁵² G₀-arrested MEFs with a loxP-flanked Mastl allele were infected with adenovirus to drive expression of Cre recombinase, then allowed to re-enter the cell cycle. These Mastl-deficient MEFs showed no differences in several markers of mitotic entry as compared with control cells, although defects in progression through mitosis were detected.⁵² These differences between MASTL-deficient MEFs and ENSA-deficient mouse oocytes may reflect fine-tuned functions of this pathway in these 2 cellular contexts.

The prophase I arrest phenotype in ENSA-deficient mouse oocytes was rescued by inhibition of PP2A using Okadaic acid or knockdown of the PP2A regulatory subunit Ppp2r2d. Thus, the data here are consistent with the model coming from Xenopus studies that ENSA functions to inhibit PP2A with the regulatory subunit PPP2R2D/B55δ, and, in turn, suggest that this activity is important during the exit from prophase I arrest in mouse oocytes. An additional observation in our studies is that Ppp2r2d siRNA-injected oocytes maintain prophase I arrest for ~48 h in the presence of dbcAMP, and then undergo germinal vesicle breakdown only upon removal of dbcAMP. In contrast, Okadaic acid-treated oocytes exit from prophase I arrest and progress into meiosis I even in the presence of dbcAMP or the phosphodiesterase inhibitor IBMX.^41-43 Since Okadaic acid treatment would inhibit all PP2A forms, these results suggest that multiple PP2A forms function in maintenance of prophase I arrest, and that the PPP2R2D/B55δ regulatory subunit is dispensable during prophase I arrest. Our data on the involvement of PPP2R2D in this ENSA-dependent pathway in oocytes are also interesting in light of findings from various mammalian cell types, showing that 2 regulatory subunits, PPP2R2D/B55δ and PPP2R2A/B55α, are involved in the regulation of mitotic exit in 2 different mammalian systems,^53,54 and that mitotic progression defects in Mastl-deficient MEFs are rescued by depletion of all 4 B55 regulatory subunits.⁵² Our data here identify PP2A-B55δ specifically as the phosphatase functioning in the MASTL-ENSA/ARPP19 pathway in this particular mammalian cell type, the prophase I mouse oocyte.

The metaphase II arrest stage and exit from this arrest are likely to involve diverse phosphatase activities of various types of PP2A heterotrimers. The model summarized in Figure 1 and other data^53,54 predict that exit from metaphase II arrest would involve ceasing inhibition of PP2A to allow PP2A-mediated dephosphorylation of CDK1 M-phase substrates. With regard to data here, the subset of Ensa siRNA-injected oocytes that progress to metaphase II arrest do not show evidence of parthenogenetic exit from metaphase II arrest. This may hint that ENSA is dispensable for metaphase II arrest, but this result should be interpreted conservatively, as these Ensa siRNA-injected oocytes were not subjected to extended culture, such as was used in studies of parthenogenesis in Mos-null oocytes or of parthenogenetic exit from metaphase II, with increased times after ovulation or start of in vitro maturation.^55-58 Knockdown of MASTL in metaphase II pig eggs results in a moderate extent of exit from metaphase II arrest over an 18-h culture period.⁴⁰ Further studies will be valuable to assess the roles of MASTL and ENSA and/or ARPP19 in regulating PP2A during exit from metaphase II arrest. Metaphase II arrest also entails specific PP2A functions, although with different PP2A types. PP2A with a B56 regulatory subunit localizes to centromeres in metaphase II mouse eggs, with function in protection of centromeric cohesion.⁵⁹ Fine-tuning of CDK1 activity during metaphase II arrest is achieved through a feedback loop with combined action of CDK1 and PP2A (in Xenopus eggs, with a B56β/ε regulatory subunit⁶⁰) on EMI2, an inhibitor of the anaphase-promoting complex/cyclosome.^60-62 Data from studies using Okadaic acid to inhibit PP2A and FTY720 to activate PP2A provide evidence that this same mechanism functions in metaphase II mouse eggs.⁶³

In summary, this study demonstrated that exit from prophase I arrest is impaired in ENSA-deficient oocytes. Our data further indicate that ENSA functions to inhibit PP2A-B55δ during exit from prophase I arrest in mouse oocytes. This adds a new dimension to our understanding of the molecular regulation of mammalian female meiosis, extending recent work showing that MASTL deficiency in pig oocytes also impairs meiotic progression,⁴⁰ with added potential relevance to human oocyte biology, as mouse and human have genes encoding both Arpp19 and Ensa.

Materials and Methods

Oocyte collection, maturation, and culture

Animals were used in accordance with the guidelines of the Johns Hopkins University Animal Care and Use Committee. Germinal vesicle-intact (GVI) (prophase I-arrested) oocytes were collected from 6–8-wk-old CF-1 mice (Harlan). Mice were injected with 5 or 10 I.U. of pregnant mare serum gonadotropin (PMSG); 46–48 h prior to oocyte collection for experiments, oocytes were matured to the metaphase II stage. Oocytes were collected in Whitten–HEPES medium (109.5 mM NaCl, 4.7 mM KCl, 1.2 mM KH₂PO₄, 1.2 mM MgSO₄, 5.5 mM glucose, 0.23 mM pyruvic acid, 4.8 mM lactic acid hemicalcium salt⁶⁴) supplemented with 7 mM NaHCO₃, 15 mM HEPES (hereafter referred to as Whitten-HEPES), and 0.05% polyvinyl alcohol (PVA; Sigma). Dibutyryl cAMP (dbcAMP, 0.25 mM) was added to culture medium to maintain prophase I arrest.⁶⁵ Ovaries were punctured with syringe needles to release oocyte–cumulus complexes from ovarian follicles, and cumulus cells were dissociated from oocytes by pipetting oocyte–cumulus complexes through a thin-bore pipette. Oocytes were transferred to Whitten medium with 22 mM NaHCO₃ (hereafter referred to as Whitten-Bicarbonate medium) containing 0.05% PVA and 0.25 mM dbcAMP for culture. Oocytes were cultured in a humidified atmosphere of 5% CO₂ in air. For microinjection of siRNA (details below), oocytes were transferred to EmbryoMax^® KSOM + amino acids with D-Glucose (MR-106-D, Millipore) (hereafter referred to as KSOM) supplemented with 0.25 mM dbcAMP. Microinjected oocytes were cultured in KSOM medium following microinjection and during meiotic maturation.

Bacterially expressed recombinant ENSA and ARPP19

cDNA fragments encoding the entire coding region of mouse Ensa (NM_019561.2) and Arpp19 (NM_021548.4) were generated by polymerase chain reaction (PCR) from oocyte cDNA using Phusion™ High-Fidelity DNA polymerase (New England Biolabs). Cloning primer sequences are provided in Table S1. The resulting PCR products were gel-purified (QIAquick Gel Extraction kit; Qiagen), digested with BamHI and SalI (New England Biolabs), and cloned into pGEX-4T-1 (Amersham-Pharmacia Biotech) according to standard protocols. The resulting plasmid (pGEX-4T-ENSA) was verified by DNA sequencing. Plasmids were transformed into E. coli BL21 cells (Stratagene). Transformed cells were induced to express GST fusion proteins with 0.5 mM IPTG (Sigma) at 37 °C for 4 h. Cell lysates were prepared by sonication in cold PBS containing 1 μg/ml leupeptin, 1 μg/ml aprotinin and 1 mM AESBF (Sigma). The lysate was centrifuged twice (20 000 × g, 10 min each), and GST fusion proteins were purified by affinity chromatography on a gluthathione column, eluting with 10 mM of reduced glutathione in 10 mM TRIS-HCl, pH 7.5. The purified protein was dialyzed against PBS. Protein concentration was determined using the micro BCA assay (Thermo Scientific), and purity was confirmed by examining the protein on silver-stained SDS-PAGE gels.

Immunoblotting

For immunoblot analysis, purified recombinant proteins or oocyte lysates were suspended in SDS-PAGE sample buffer (3% SDS, 10% glycerol, 0.02% bromophenol blue, 4% 2-mercaptoethanol, 65 mM TRIS-HCl, pH 6.8) and heated for 5 min at 95 °C. Proteins were separated on a 12.5% SDS-gel and transferred to an Immobilon-P PVDF membrane (Millipore). The membrane was blocked overnight in 10% cold-water fish gelatin (Sigma) in PBS containing 0.1% Tween-20 (Sigma) (hereafter referred to as PBS-T). The membrane was then incubated with an anti-ENSA antibody (Santa Cruz Biotechnology, catalog #sc-135145; made against full-length human ENSA) diluted to a concentration of 0.5 μg/ml in PBS-T containing 3% BSA and 0.02% NaN₃) for 3 h, washed 3 times with PBS-T, then incubated with 0.3 μg/ml goat anti-rabbit IgG-horseradish peroxidase (HRP) (Jackson Immunoresearch) in PBS-T plus 3% BSA for 1.5–2 h, followed by 3 washes with PBS-T. For detection, the membrane was incubated with Supersignal^® West Pico Chemiluminescent substrate (Pierce Chemical Company) and exposed to X-ray film. X-ray film was scanned using the HP LaserJet 3390 scanner, and images were saved as TIFF files. ImageJ software (http://rsb.info.nih.gov/ij/) was used to analyze band intensity. The rectangular selection tool was used to select each band, and peak intensity was determined. The area under each peak was calculated as a measure of band intensity. Alternatively, for some experiments, the membrane was then incubated with the anti-ENSA antibody (0.3 μg/ml, diluted 1:685 in a 1:1 solution of PBS and Odyssey^® Block Buffer) for 1 h, washed 4 times in PBS-T (5 min each wash), incubated with goat anti-rabbit IRDye^® 800CW secondary antibody (diluted 1:20,000 in a 1:1 solution of PBS and Odyssey^® Block Buffer) for 1 h, washed 4 times in PBS-T (5 min each wash), and washed once in PBS. The membrane was then scanned using the Odyssey^® CL_X Infrared Imaging System (Li-Cor) at an intensity setting between 6.5 and 7.5.

Immunofluorescence

Immunofluorescence experiments used a mouse monoclonal anti-α-tubulin antibody (clone 12G10⁶⁶; developed by Joseph Frankel and E Marlo Nelsen, obtained from the Developmental Studies Hybridoma Bank developed under the auspices of the NICHD and maintained by The University of Iowa, Department of Biology) to label the meiotic spindle, and a rabbit polyclonal anti-lamin B antibody (ref. 67, gift of the lab of Mike Matunis, Department of Biochemistry and Molecular Biology, Johns Hopkins School of Public Health) to label the nuclear envelope. For immunoflurorescence analysis, zona pellucida (ZP)-free oocytes were prepared by briefly incubating the oocytes in acidic culture medium-compatible buffer (~10 s; 10 mM HEPES, 1 mM NaH₂PO₄, 0.8 mM MgSO₄, 5.4 mM KCl, 116.4 mM NaCl, pH 1.5). ZP-free oocytes were fixed for 30 min at 37 °C in 4% paraformaldehyde (Sigma) in 130 mM KCl, 25 mM HEPES, 3 mM MgCl₂, 0.06% Triton-X (pH 7.4). Fixed oocytes were briefly washed in PBS, then permeablized (PBS containing 0.1% Triton X-100 [Sigma]; 15 min), then incubated for 1 h or overnight in immunofluorescence (IF) blocking buffer (PBS containing 1% BSA). Following blocking, oocytes were incubated with primary antibody diluted in IF blocking buffer for 1–1.5 h (anti-tubulin, 10.6 µg/ml; anti-lamin B, diluted 1:1500). The oocytes were then washed in IF blocking buffer and incubated for 1–1.5 h in secondary antibody diluted in blocking buffer (7.5 μg/ml goat-anti-mouse IgG-FITC; 7.5 μg/ml goat-anti-rabbit IgG-FITC) (Jackson Immunoresearch) followed by 3 washes in IF blocking buffer. Oocytes were mounted on slides in VectaShield mounting medium (Vector Laboratories) containing 0.75 µg/ml DAPI. Microscopic imaging was performed on a Zeiss Axio Observer Z1 microscope with AxioCam MRm Rev3 camera, ApoTome optical sectioning, and AxioVision software (Carl Zeiss, Inc).

RNAi-mediated knockdown

ON-TARGETplus SMARTpool siRNA targeting mouse Ensa or Ppp2r2d (Dharmacon) was resuspended according to manufacturer’s instructions, to a concentration of 100 μM in 4 volumes of RNase-free water and 1 volume of 5× siRNA Buffer (Dharmacon, B-002000-UB-100). The ON-TARGETplus control pool (Dharmacon, D-001810-10-20) was used as a negative control. The 100 μM siRNA stocks were diluted to a 20 μM in 1× siRNA buffer for microinjection, or, for double knockdown of Ensa and Ppp2r2d, siRNA stocks were 10 µM each. ZP-intact oocytes were injected using a Nikon Eclipse TE 2000–5 microscope equipped with an Eppendorf FemtoJet^®, using injection pressure (p_i) of 100–200 hPa, injection time (t_i) of 0.1–0.2 s, and compensation pressure (p_c) of 0 hPa. Injection was considered successful upon observation of cytoplasmic recoiling following needle insertion and post-injection dispersal of the siRNA solution into the oocyte cytoplasm. siRNA-injected oocytes were cultured in KSOM medium containing 0.25 mM dbcAMP for 44–48 h, with transfer to fresh medium after 24 h. For experiments assessing meiotic maturation to metaphase II, siRNA-injected oocytes were cultured for 36 h prior to the initiation of maturation.

RNA isolation, cDNA synthesis, and semi-quantitative RT-PCR

Total RNA was isolated from siRNA-injected oocytes by lysing 20 prophase I oocytes in 200 μl Trizol (Invitrogen) for 5 min at room temperature. This mixture was chloroform-extracted, then the RNA-containing aqueous phase was supplemented with a final concentration of 1 μg/ml glycogen (Invitrogen) and subjected to isopropanol precipitation overnight at −20 °C, after which the RNA was pelleted by centrifugation (15 min, 14 000 × g, 4 °C). The pellet was washed with 70% ethanol, air-dried, and then resuspended in 8 μl of RNase-free water. The RNA concentration was determined by measuring the OD₂₆₀. First-strand cDNA was synthesized from total RNA with random hexamer primers, SuperScript Reverse Transcriptase III (Invitrogen), and RNaseOUT RNase Inhibitor (Invitrogen), according to the manufacturer’s protocol, after which the cDNA was treated with E. coli RNase H (New England Biolabs).

To assess RNA knockdown, RT-PCR using gene-specific primers was performed; primer sequences are provided in Table S1. Tissue plasminogen activator (gene symbol Plat) was used as a positive control, as previously described for mouse oocytes.⁶⁸ For semi-quantitative PCR, the amount of PCR product was assessed for a range of cycle numbers, ensuring that the reaction was in the linear range and not saturated. PCR products were separated on agarose gels, which were scanned using the FujiFilm FLA-7000 imaging system (FujiFilm). FLA-7000 image files were exported to the MultiGauge program (FujiFilm) and saved as TIFF files. ImageJ software (http://rsb.info.nih.gov/ij/) was used to analyze band intensity. The rectangular selection tool was used to select each band, and peak intensity was determined. The area under each peak was calculated as a measure of band intensity. Band intensity for Ensa, Arpp19, or Ppp2r2d primer-amplified products was normalized to those for the positive control (Plat) primer set. Transcript levels in oocytes injected with target-specific siRNAs (Ensa, Ppp2r2d) were expressed relative to those levels in oocytes injected with negative control siRNA.

In vitro meiotic maturation

siRNA-injected oocytes were washed through 6 drops of KSOM to wash away the dbcAMP.⁶⁵ For observation of progress through germinal vesicle breakdown (also known as nuclear envelope breakdown), oocytes were cultured in KSOM, with observation at 30-min intervals for 5 h by dissecting microscope for the presence of an intact germinal vesicle. Data are presented as the percentage of oocytes having undergone germinal vesicle breakdown (GVBD) over time. For in vitro maturation to the metaphase II stage, oocytes were cultured for 14–15 h and assessed for first polar body emission by observing with a dissecting microscope.

For a subset of experiments, oocytes were treated with the phosphatase inhibitor Okadaic acid. Okadaic acid (Santa Cruz Biotechnology, sc-3513) was re-suspended in dimethyl sulfoxide (DMSO) to a 2 mM stock concentration and stored at −80 °C; this 2 mM stock was diluted to 2.5 μM in KSOM medium. This dose was chosen based on past studies of Okadaic acid treatment of mouse oocytes.^41,43,69 These studies had 4 experimental groups: negative control siRNA-injected oocytes, negative control siRNA-injected oocytes treated with 2.5 µM Okadaic acid, Ensa siRNA-injected oocytes, and Ensa siRNA-injected oocytes treated with 2.5 µM Okadaic acid. For treatment with Okadaic acid, oocytes in KSOM medium containing 0.25 mM of dbcAMP were washed through 6 drops of KSOM and then transferred to KSOM with 2.5 µM Okadaic acid. Oocytes were assessed at 30 min intervals for the presence of an intact germinal vesicle as described above.

Glossary

Abbreviations:

ENSA

alpha-endosulfine

ARPP19

cAMP-regulated phosphoprotein-19

cyclin-dependent kinase 1

CDK1

MASTL

microtubule associated serine/threonine kinase-like

protein kinase A

PKA

protein phosphatase 2A

PP2A

ZP

zona pellucida