1. Cell Biology

Selective dephosphorylation by PP2A-B55 directs the meiosis I-meiosis II transition in oocytes

  1. S Zachary Swartz
  2. Hieu T Nguyen
  3. Brennan C McEwan
  4. Mark E Adamo
  5. Iain M Cheeseman  Is a corresponding author
  6. Arminja N Kettenbach  Is a corresponding author
  1. Whitehead Institute for Biomedical Research, United States
  2. Department of Biology, Massachusetts Institute of Technology, United States
  3. Department of Biochemistry and Cell Biology, Geisel School of Medicine at Dartmouth, United States
  4. Norris Cotton Cancer Center, Geisel School of Medicine at Dartmouth, United States
https://doi.org/10.7554/eLife.70588
Share this article
Cite this article
  1. S Zachary Swartz
  2. Hieu T Nguyen
  3. Brennan C McEwan
  4. Mark E Adamo
  5. Iain M Cheeseman
  6. Arminja N Kettenbach
(2021)
Selective dephosphorylation by PP2A-B55 directs the meiosis I-meiosis II transition in oocytes
eLife 10:e70588.
https://doi.org/10.7554/eLife.70588
  2. Download BibTeX
  3. Download .RIS

Abstract

Meiosis is a specialized cell cycle that requires sequential changes to the cell division machinery to facilitate changing functions. To define the mechanisms that enable the oocyte-to-embryo transition, we performed time-course proteomics in synchronized sea star oocytes from prophase I through the first embryonic cleavage. Although we found that protein levels were broadly stable, our analysis reveals that dynamic waves of phosphorylation underlie each meiotic stage. We found that the phosphatase PP2A-B55 is reactivated at the meiosis I/meiosis II (MI/MII) transition, resulting in the preferential dephosphorylation of threonine residues. Selective dephosphorylation is critical for directing the MI/MII transition as altering PP2A-B55 substrate preferences disrupts key cell cycle events after MI. In addition, threonine to serine substitution of a conserved phosphorylation site in the substrate INCENP prevents its relocalization at anaphase I. Thus, through its inherent phospho-threonine preference, PP2A-B55 imposes specific phosphoregulated behaviors that distinguish the two meiotic divisions.

Introduction

Animal reproduction requires that oocytes undergo a specialized cell cycle called meiosis, in which two functionally distinct divisions occur in rapid succession to reduce genome ploidy (Kishimoto, 2018). Following fertilization, the oocyte must then transition to a third division strategy, mitosis, for early embryonic development. This oocyte-to-embryo transition occurs in a short temporal window, but must achieve high fidelity to ensure that heritable information is accurately transmitted from the parents to the developing embryo. At the center of this progression is a suite of cell cycle regulatory proteins and molecular machines that drive and integrate processes such as chromosome segregation, fertilization, and pronuclear fusion. An important goal is to unravel the complex regulatory mechanisms that precisely coordinate these divisions in time and space within the oocyte. In particular, it remains unknown how phosphorylation and dephosphorylation drive the meiotic divisions allowing oocytes to rewire the cell division machinery at the meiosis I/meiosis II (MI/MII) transition to facilitate differing requirements.

The female meiotic cell cycle is distinct from mitosis in several ways, necessitating a unique regulatory control. First, oocytes remain in an extended primary arrest in a cell cycle state termed prophase I until receiving an extrinsic hormonal signal (Conti and Chang, 2016; Jaffe and Norris, 2010; Kishimoto, 2018; Von Stetina and Orr-Weaver, 2011). Second, the meiotic divisions use a small asymmetrically positioned spindle to partition chromosomes into polar bodies, which do not contribute to the developing embryo (Severson et al., 2016). In addition, the first meiotic division segregates bivalent pairs of homologous chromosomes, whereas for MII, this configuration is reversed and instead sister chromatids are segregated (Figure 1A; Watanabe, 2012). Finally, meiosis lacks a DNA replication phase between the polar body divisions, which enables the reduction of ploidy to haploid. How the cell division machinery is specialized to perform the distinct functions of MI, and then is rapidly reorganized for the unique requirements of MII while remaining in meiosis and not exiting into gap or S-phase, is an important open question.

Figure 1 with 3 supplements see all
Download asset Open asset
The proteome during the oocyte-to-embryo transition is broadly stable.

(A) Schematic of meiotic progression in sea star oocytes, representing the six stages collected for mass spectrometry analysis. (B) Proteomics workflow diagram, in which protein samples were collected in biological triplicates, digested, tandem mass tag (TMT) labeled, fractionated, and analyzed by liquid chromatography with mass spectrometry (LC-MS). (C) Hierarchical clustering of the relative abundance of 4635 proteins detected across three replicates. Individual proteins are clustered (vertically) by the six isolated meiotic stages (horizontally). (D) Histogram of proteins binned by their maximum fold change in abundance, indicating 98.99% of all proteins undergo a fold change of less than 2. (E) Abundance histogram of proteins identified in our analysis reveals a normal distribution. (F) Relative abundances of selected proteins across stages (Pro, prophase I; GVBD, germinal vesicle breakdown; MI, meiosis I; MII, meiosis II; 2-PN, two pronucleus; FC, first cleavage). Light blue shading represents standard deviation.

In this study, we define phosphoregulatory mechanisms that drive the MI/MII transition. We undertook a proteomic and phosphoproteomic strategy using oocytes of the sea star Patiria miniata, which undergoes meiosis with high synchrony (Swartz et al., 2019). Prior analyses have revealed proteome-wide changes in animal models including Xenopus, Drosophila, and sea urchins (Guo et al., 2015; Krauchunas et al., 2012; Presler et al., 2017; Zhang et al., 2019). However, the biology of these organisms limits access to a comprehensive series of time points spanning prophase I through the embryonic divisions, including the critical MI/MII transition. Our sea star proteomics dataset spans the entire developmental window from prophase I arrest through both meiotic divisions, fertilization, and the first embryonic division (Figure 1A). We identified a surprising differential behavior between serine and threonine dephosphorylation at the MI/MII transition that we propose to underlie key regulatory differences between these meiotic divisions. This regulatory switch is driven by PP2A-B55, which is reactivated after MI to preferentially dephosphorylate threonine residues, thereby creating temporally distinct reversals of cyclin-dependent kinase (CDK) and mitogen-activated protein kinase (MAPK) phosphorylation. We propose a model in which the usage of threonine vs serine endows substrates with different responsivity to a common set of kinases and phosphatases, temporally coordinating individual proteins with meiotic cell cycle progression to achieve specific behaviors for MI and MII without exiting from meiosis.

Results

Proteomic analysis reveals stable protein abundance during the oocyte-to-embryo transition

The oocyte-to-embryo transition involves an ordered series of events including fertilization, chromosome segregation, polarization, and cortical remodeling. To determine the basis for these cellular transitions and their corresponding physical changes, we analyzed the proteome during the oocyte-to-embryo transition using quantitative tandem mass tag (TMT)-multiplexed mass spectrometry (MS) (Thompson et al., 2003). We obtained prophase I-arrested oocytes from the sea star P. miniata and treated them with the maturation-inducing substance 1-methyladenine (1-MeAd) to trigger the resumption of meiotic progression in seawater culture (Kanatani et al., 1969). In an initial trial, we collected oocytes for immunofluorescence every 10 min after 1-MeAd stimulation and determined that oocytes proceed through meiosis with high synchrony (over 90% synchrony of oocytes at germinal vesicle breakdown (GVBD), MI, MII, and pronuclear stages; Figure 1—figure supplement 1). Live-imaging experiments further supported synchronous progression through meiosis and early development in this species (Figure 1—video 1). Leveraging these features, we cultured isolated oocytes and collected biological triplicate samples at the following stages: (1) prophase I arrest (Pro), (2) GVBD, (3) metaphase of MI, (4) prometaphase of MII, (5) just prior to pronuclear fusion (2-PN), and (6) metaphase of the first embryonic cleavage (FC) (Figure 1A).

We first tested whether protein abundance changes could regulate the oocyte-to-embryo transition (Figure 1B). Using MS, we identified 8026 total proteins, of which 6212 were identified in two independent time-course series and 4635 in all three series (Figure 1—figure supplement 2, supplementary file 1). Surprisingly, only a limited number of proteins changed in abundance across these different stages (Figure 1C). In fact, 99% of the 4635 proteins reproducibly identified in all three time-course series displayed a maximum fold change of less than 2 from prophase I to the first embryonic cleavage, with 74.8% of proteins displaying less than a 1.2-fold change (Figure 1D). The absence of changes in protein levels was not due to a bias in our analysis as protein abundance followed a normal distribution (Figure 1E). We also note that this limited protein turnover during meiosis is consistent with prior reports in other organisms (Kishimoto, 2018; Peuchen et al., 2017; Presler et al., 2017). However, despite this broad stability of protein levels, there were several notable exceptions (Figure 1F). Gene ontology (GO) analysis (Liao et al., 2019) of significantly regulated proteins with a fold change of 1.2 or more revealed an enrichment in cytoskeletal proteins, proteins involved in RNA binding, and ribosomal components (Supplementary file 2). Moreover, cyclin B levels were high in prophase I, GVBD, and MI, but declined sharply in MII, before being partially restored in the first cleavage stage (Figure 1F). These dynamics are consistent with Anaphase Promoting Complex/Cyclosome (APC/C) mediated destruction of cyclin B during cell cycle progression (Evans et al., 1983; Kishimoto, 2018; Okano-Uchida et al., 1998). In addition, we identified the serine/threonine kinase Proviral integration site for Moloney murine leukemia virus-1 (PIM-1) (Figure 1F), previously proposed to be a potential mitotic regulator (Bachmann et al., 2006), as a protein that is absent in prophase I oocytes but translated de novo following meiotic resumption. Thus, although the overall proteome is highly stable, selected proteins are translationally or proteolytically regulated with meiotic cell cycle progression.

New translation of selected proteins is required for meiotic progression

Although protein levels are largely constant across the oocyte-to-embryo transition, de novo translation could act to maintain steady-state levels or may be required to produce a limited set of factors involved in meiotic progression. To test this, we globally prevented translation using the 40S ribosomal inhibitor emetine (Jimenez et al., 1977). Emetine-treated oocytes responded to 1-MeAd stimulation to initiate MI, consistent with prior work (Houk and Epel, 1974). However, instead of progressing to MII, the maternal DNA decondensed and formed a pronucleus with oocytes remaining arrested even at time points when control oocytes had initiated first cleavage (Figure 2A,B). Based on a proteomic analysis of MII and pronuclear stage oocytes, we found that only 108 out of 7610 proteins identified in our analysis significantly changed with emetine treatment (Figure 2—figure supplement 1, supplementary file 3). These emetine-sensitive proteins fell into diverse categories but were overrepresented for cytoskeletal elements and actomyosin-related proteins (Supplementary file 4). Notably, we identified the DNA replication factor Cdt1, whose nascent translation may be required for the initiation of DNA replication after meiotic exit. Moreover, we again identified PIM-1 as a factor sensitive to emetine treatment, consistent with our steady-state analysis (Figure 1F). In summary, most proteins are insensitive to translational inhibition, indicating a general lack of turnover between MI and MII, but new protein synthesis is required for progression past MI.

Figure 2 with 2 supplements see all
Download asset Open asset
Protein synthesis is required for the MI/MII transition.

(A) Immunofluorescence of control or emetine-treated oocytes, with DNA provided in single-channel grayscale images. While controls proceed to meiosis II (MII) and first cleavage, emetine-treated oocytes decondense DNA after meiosis I (MI), arrest in a pronuclear state, and fail to incorporate the male DNA. Microtubules were scaled nonlinearly. Scale bars = 10 μm. (B) Fraction of oocytes that successfully extruded both polar bodies and underwent first cleavage (MI: control n = 107, emetine n = 113; MII: control n = 107, emetine n = 114; first cleavage: control n = 31, emetine n = 47 oocytes; ****p<0.0001 by Fisher’s exact test). (C) Immunofluorescence of oocytes in which nascent synthesis of cyclin A, cyclin B, or both was blocked. Control oocytes extruded both polar bodies and initiated first cleavage. Blocking cyclin A synthesis did not affect the meiotic divisions but caused an arrest prior to the first cleavage. Blocking cyclin B instead selectively disrupted the second mitotic division, but the first meiosis and initiation of first cleavage proceeded normally. Combined translational inhibition of both cyclin A and cyclin B resulted in an interphase arrest following the first meiotic division. Microtubules were scaled nonlinearly. Scale bars = 10 μm. (D) Fraction of oocytes that successfully extruded polar bodies and underwent first cleavage (cyclin A n = 82, cyclin B n = 66, cyclin A + B n = 52 oocytes).

The requirement of protein synthesis for meiotic progression could reflect the need to translate selected cell cycle factors. To test whether established cell cycle regulators must be translated de novo, we used morpholino injection to specifically prevent new translation of cyclin B, one of the few proteins that varies in abundance (Figure 2C,D), as well as cyclin A, which is synthesized in late MI in the related sea star species Patiria pectinifera (Hara et al., 2009; Okano-Uchida et al., 1998). Morpholino injection specifically reduced cyclin B protein levels and reduced total CDK-consensus phosphorylation (Figure 2—figure supplement 2). We stimulated oocytes with 1-MeAd immediately following morpholino injection to ensure that pre-existing cyclin protein was unaffected. When new cyclin A synthesis was blocked, oocytes underwent both meiotic divisions normally and the maternal and paternal pronuclei fused, but these zygotes then arrested with a single fused pronucleus and failed to progress to the first cleavage (Figure 2C,D). This is consistent with a role for cyclin A in mitotic entry in cultured cells, and the transition to embryogenesis in P. pectinifera (Hara et al., 2009; Okano-Uchida et al., 2003; Pagano et al., 1992). In contrast, preventing cyclin B synthesis resulted in a normal MI division but failure to extrude the second polar body and retention of an additional centriole. Surprisingly, these oocytes successfully underwent pronuclear fusion, entered the first cleavage, and formed a mitotic spindle. This suggests that cyclin B must be translated de novo following anaphase I to drive meiosis II but is dispensable for the initial transition from meiosis to embryonic mitosis. Finally, when we simultaneously prevented the new translation of both cyclin A and B, oocytes completed MI and then arrested in a pronuclear-like state without conducting MII (Figure 2C,D), similar to the effect of translational inhibition by emetine (Figure 2A,B). We further attempted to prevent translation of PIM-1 but did not observe a substantial phenotype, which could reflect technical challenges in its knockdown. Taken together, our results suggest that the proteome during the oocyte-to-embryo transition is highly stable but that the de novo translation of cyclins is required for meiotic progression.

Defining the phosphorylation landscape of the oocyte-to-embryo transition

The two meiotic divisions, fertilization, pronuclear fusion, and the first mitotic cleavage all occur within less than 3 hr in the absence of substantial changes in protein abundance (Figure 1C,D). This suggests that there are alternative mechanisms to rapidly re-organize the cell division apparatus during these transitions. Therefore, we next assessed phosphorylation across the oocyte-to-embryo transition using phosphopeptide enrichment followed by MS (Figure 3A). Our analysis identified a total of 25,228 phosphopeptides across three multiplexed time courses. Among those phosphopeptides, 16,691 were identified in two, and 11,430 in three, multiplexes (Figure 3—figure supplement 1A, supplementary file 5). We detected 79.3% of phosphorylation on serine residues, 19.6% on threonine, and 1.1% on tyrosine residues based on a phosphorylation localization probability of ≥0.9 (Figure 3—figure supplement 1B). Prior work found similar ratios of S:T:Y phosphorylation based on autoradiographic measurements in chicken cells (92%:7.7%:0.3%) and based on phosphoproteomics in human cells (84.1%:15.5%:0.4%) (Hunter and Sefton, 1980; Sharma et al., 2014).

Figure 3 with 7 supplements see all
Download asset Open asset
Identification of dynamic phosphorylation changes during the oocyte-to-embryo transition.

(A) Proteomics workflow diagram, in which a phosphopeptide enrichment step was performed prior to tandem mass tag (TMT) labeling. (B) Hierarchical clustering of 10,645 phosphorylation events (a localization score of 0.9 or higher, a p-value of 0.05 or less, and stage-specific peak phosphorylation) clustered by phosphosite (rows) and meiotic stage (columns). (C) Percentage of sites that reach their peak phosphorylation levels across stages, revealing peaks at meiosis I (MI) and first cleavage. (D) Average abundance of all phosphorylation events per stage. (E, G) Temporal phosphorylation levels of conserved sites on the kinase Cdk1/2 and the MAP kinase Erk (blue trace, light blue areas are standard deviations; orange trace represents total relative protein abundance). (H, F) Western blots with antibodies recognizing the inhibitory phosphorylation on Cdk1Y15 and activating phosphorylation on ERK Y193, respectively.

Figure 3—source data 1

Identification of dynamic phosphorylation changes during the oocyte-to-embryo transition.

Western blots with antibodies recognizing the inhibitory phosphorylation on Cdk1Y15 and activating phosphorylation on ERK Y193, respectively.

https://cdn.elifesciences.org/articles/70588/elife-70588-fig3-data1-v2.pdf

Hierarchical clustering of the dynamic phosphorylation behavior from prophase I to the first embryonic cleavage revealed several striking transitions in global phosphorylation status (Figure 3B). First, prophase I-arrested oocytes are distinct from those in the other stages in that they not only display a limited number of phosphorylation sites at the relative maximum phosphorylation levels (Figure 3C) but also have the lowest overall phosphorylation state of the samples tested (Figure 3D). Second, more than half of the total phosphorylation sites identified were maximally phosphorylated in MI, whereas phosphorylation was substantially reduced in MII (Figure 3C). These patterns of phosphorylation imply a critical role for phosphoregulation in specializing the two meiotic and first cleavage divisions and suggest a role for a low phosphorylation state in maintaining the prophase I arrest.

Kinase activity across the oocyte-to-embryo transition

To remain arrested in prophase I, CDKs must be kept inactive. As our proteomics analysis indicated that the majority of proteins in the oocytes, including kinases, are present constitutively (Figure 1C), kinase activity must be controlled post-translationally. We, therefore, analyzed the pattern of phosphorylation events on established cell cycle kinases. We first identified inhibitory phosphorylation sites on Cdk1 or Cdk2 (Y21 or Y15, respectively) by proteomics and western blotting using phospho-specific antibodies against these conserved sites (Figure 3E,F). These sites are phosphorylated in prophase I arrest, and to a lesser extent at first cleavage, but not during meiosis. This phosphorylation pattern suggests that Cdk is inactive in prophase I-arrested oocytes but active throughout the meiotic divisions. In addition to Cdk, the meiotic divisions and secondary arrest that occurs in the absence of fertilization require MAP kinase activity downstream of the conserved activator Mos (Dupré et al., 2011; Tachibana et al., 2000). Based on our phosphoproteomics and western blotting, we found that a conserved activating phosphorylation on the MAP kinase p42/ERK phosphorylation (Y204) was undetectable in prophase I-arrested oocytes, high in MI and MII, and low in first cleavage (Figure 3G,H).

Meiotic resumption in sea star oocytes requires the action of kinases downstream of the 1-MeAd hormone G protein-coupled receptor (GPCR) on the oocyte surface (Chiba et al., 1992). Hormonal reception leads to activation of phosphoinositide 3-kinase (PI3K), which in turn phosphorylates and activates the meiotic trigger kinase, which has been previously proposed to be RAC-alpha protein kinase (AKT) kinase (Okumura et al., 2002). However, recent work has instead suggested that SGK kinase is the primary trigger in sea star P. pectinifera, which also acts downstream of PI3K (Hiraoka et al., 2019). In support of this model, we detected an increase in phosphorylation of the SGK activation segment following hormonal stimulation (Figure 3—figure supplement 2A). SGK then phosphorylates the CDK regulators Cdc25 and Myt1 to activate and repress them, respectively, thereby enabling the activation of Cdk1 (Hiraoka et al., 2019). Concordantly, we observed an increase in phosphorylation on Cdc25 S188, an activating SGK site, and Myt1 S75, a repressive SGK site, following hormonal stimulation (Figure 3—figure supplement 2B,C; Hiraoka et al., 2019). Finally, to assess the relative contributions of AKT and SGK to meiotic resumption, we investigated the presence of RxRxxS/T consensus motifs in the dataset for the localized, single, and reproducibly quantified phosphopeptides. Based on known AKT and SGK substrates, the AKT consensus motif shows enrichment in basophilic amino acids in the −4 position compared to SGK (Hornbeck et al., 2015). We then compared the sites that increased in phosphorylation abundance by threefold or more from prophase I to GVBD to sites with less than a threefold increase over the same time period. The sites that increased in abundance by more than threefold were enriched for a SGK-like consensus motif, whereas those that increased less displayed a more AKT-like motif (basophilic residues in the −4 position) (Figure 3—figure supplement 2D). Collectively, these results support a model where SGK acts as the major trigger kinase driving meiotic resumption (Hiraoka et al., 2019).

Finally, meiotic progression also requires Greatwall kinase (Kishimoto, 2018), which acts upstream to inhibit PP2A-B55. Greatwall is sequestered in the germinal vesicle in prophase I sea star oocytes and is activated downstream of Cdk1/cyclin B (Hara et al., 2012). We identified a conserved activating phosphorylation within the activation segment of Greatwall kinase (T194 in humans; T204 in sea star) (Figure 3—figure supplement 2E; Blake-Hodek et al., 2012; Gharbi-Ayachi et al., 2010), indicative of high Greatwall kinase activity during GVBD and MI, and reduced activity in later stages. In summary, our phosphoproteomic time course reveals orchestrated transitions in the activity of regulatory kinases during the oocyte-to-embryo transition.

Prophase I arrest is enforced by high phosphatase activity

Our phosphoproteomics analysis indicated that prophase I is characterized by low global phosphorylation. To determine whether this state is reinforced by phosphatase activity, we next examined modifications to the major cell cycle phosphatases, Protein Phosphatase 1 (PP1) and Protein Phosphatase 2A (PP2A), and their regulators (Nasa and Kettenbach, 2018). PP1 activity is inhibited by phosphorylation of its catalytic subunits (phosphorylation of T320 (human) by Cdk1) (Dohadwala et al., 1994; Kwon et al., 1997). PP1 T316 (corresponding to T320 in humans) is hypo-phosphorylated in prophase I-arrested oocytes (Figure 3—figure supplement 2F), which would result in high PP1 activity. PP1 activity is also controlled through regulatory subunits, which it recognizes through short-linear motifs, most prominently the ‘RVxF’ motif (Bollen et al., 2010; Heroes et al., 2013). Phosphorylation of the ‘x’ position within an RVxF motif or of adjacent residues disrupts the interaction between regulatory and catalytic subunits (Nasa et al., 2018). Analysis of the phosphorylation abundance of RVxF motifs across the phosphoproteomics time course revealed low RVxF phosphorylation occupancy in prophase I, which increased in GVBD and MI, before decreasing in later stages (Figure 3—figure supplement 2G). To validate this observation, we performed western blots using an antibody raised against a phosphorylated RVp[S/T]F epitope (Nasa et al., 2018) and observed similar trends (Figure 3—figure supplement 3). This suggests that, in prophase I, the PP1 catalytic subunit is maximally bound to regulatory subunits to promote substrate dephosphorylation.

PP1-interacting proteins can act as targeting subunits, but can also act as PP1 inhibitors, resulting in a complex PP1 regulatory network (Bollen et al., 2010; Heroes et al., 2013; O'Connell et al., 2012). For example, the association of PP1 with the regulatory factor NIPP1 (nuclear inhibitor of PP1) inhibits its activity but can also recruit some substrates for dephosphorylation. NIPP1 association with PP1 is negatively regulated by phosphorylation of an RVxF motif at the binding interface (O'Connell et al., 2012). Indeed, we found that phosphorylation adjacent to its RVxF is low in prophase I-arrested oocytes, but increases following meiotic resumption (Figure 3—figure supplement 4A–C). Preventing NIPP1 phosphorylation with phospho-inhibitory mutations (P. miniata NIPP1 S199A or S197A S199A double mutants) resulted in a greater than fivefold increase in the binding of sea star NIPP1 to human PP1 when expressed in human 293T cells (Figure 3—figure supplement 4D; Beullens et al., 1992; Beullens et al., 1999; Tanuma et al., 2008; Winkler et al., 2015). Thus, the low NIPP1 phosphorylation in prophase I oocytes would increase the association between PP1 and NIPP1, modulating and reducing its phosphatase activity toward specific substrates. Taken together, these observations indicate that there is complex regulation of PP1 activity in prophase I involving multiple activating and inhibitory holoenzyme complexes.

We also identified phosphorylation events predicted to regulate PP2A activity through its B55 subunit (Figure 3—figure supplement 5A). Upon phosphorylation of ARPP19 by Greatwall kinase (on S67 in humans, S106 in P. miniata), phospho-ARPP19 binds to and inactivates the regulatory PP2A regulatory subunit B55 (Gharbi-Ayachi et al., 2010; Okumura et al., 2014). Indeed, we found that PmARPP19 thio-phosphorylated in vitro by Greatwall kinase resulted in increased Cdc25 phosphorylation when added to mitotic human cell lysates (Figure 3—figure supplement 5B). Thus, phospho-ARPP19 S106 acts as an active B55 inhibitor such that ARPP19 phosphorylation serves as a proxy for assessing PP2A-B55 activity in sea star oocytes. PmARPP19 S106 phosphorylation is low in prophase I, but is high in GVBD and MI, before decreasing again in MII (Figure 3—figure supplement 5C). ARPP19 is additionally activated by Cdk1 phosphorylation on S69 (Figure 3—figure supplement 5C; Okumura et al., 2014). Consistent with this regulation, we observed increasing phosphorylation on S69 following meiotic resumption (Figure 3—figure supplement 5C). Based on the collective behavior of these regulatory phosphorylation events on ARPP19, we conclude that PP2A-B55 activity is high in prophase I, low in MI, but is reactivated at the MI/MII transition.

To test the functional requirement for these phosphatases in maintaining the prophase I arrest, we treated oocytes with the potent dual PP1 and PP2A inhibitor calyculin A. Following calyculin A addition, we found that 100% of oocytes spontaneously underwent GVBD within 70 min (Figure 3—figure supplement 6A,C; also see Tosuji et al., 1991). To determine whether this effect was due to inhibition of PP1 or PP2A, we used the selective PP1 inhibitor tautomycetin (Choy et al., 2017; Swingle et al., 2007). Using in vitro phosphatase assays, we confirmed that tautomycetin potently inhibited PP1 (Figure 3—figure supplement 6B). In contrast to calyculin A treatment, tautomycetin did not induce GVBD across a wide range of concentrations (Figure 3—figure supplement 6C). However, following meiotic resumption induced by hormonal stimulation, tautomycetin treatment did cause meiotic chromosome alignment and segregation errors (Figure 3—figure supplement 6D). This phenotype supports a role for PP1 in regulating kinetochore-microtubule attachments and other activities during meiosis. We therefore conclude that PP1 is dispensable for maintaining a prophase I arrest but is required to achieve high fidelity chromosome segregation during meiosis, with PP2A playing the critical role in maintaining the prophase I arrest.

To test the contributions of PP1 and PP2A, we next conducted a phosphoproteomic analysis of calyculin A-treated oocytes. This analysis revealed a broad upregulation of phosphorylation (Supplementary file 5), with 85% of phosphorylation sites displaying a high level of phosphorylation after 70 min of calyculin A treatment (Figure 3—figure supplement 7A,B). However, phosphorylation sites with a normal maximal phosphorylation occupancy in prophase I (when PP2A displays high activity) decreased after 70 min of calyculin A treatment (Figure 3—figure supplement 7B). Although global phosphorylation levels increased similarly to those observed upon meiotic entry, calyculin A-treated oocytes failed to progress past this GVBD-like state based on the absence of a contractile actin network, chromosome congression, or spindle formation (Figure 3—figure supplement 7C). Therefore, although PP2A phosphatase activity is required to maintain a normal prophase I arrest, its inhibition is not sufficient to recapitulate physiological meiotic resumption. Collectively, our phosphoproteomic and functional analyses reveal high PP2A activity in prophase I-arrested oocytes, which enables a low global phosphorylation state, with additional waves of PP1 and PP2A phosphatase activity controlling subsequent phases of meiotic phosphorylation.

PP2A-B55 drives selective dephosphorylation at the MI/MII transition

During meiosis, oocytes must undergo two consecutive chromosome segregation events without exiting into an interphase state. These rapid divisions occur within 30 min of each other in the sea star, but they each achieve distinct functions. Therefore, a subset of MI-associated phosphorylation events must be reversed to allow progression to MII, whereas others must be maintained to remain in meiosis. Of the stages tested, MI displayed the largest number of sites with maximal phosphorylation (Figure 3A), including a large number of sites with a TP or SP consensus motif indicative of CDK- or MAPK-dependent phosphorylation (Supplementary file 5). Moreover, phosphorylation of these sites increased when oocytes were treated with the calyculin A, indicating that they are putative substrates for PP1 or PP2A (Figure 3—figure supplement 7A,B, supplementary file 5). Hierarchical clustering of peptides with a single phosphorylation site maximally phosphorylated in MI revealed three distinct clusters (Figure 4A). A subset of sites sharply decreased in their phosphorylation after MI (Figure 4B, cluster 3), whereas other sites remained phosphorylated during MII and the first embryonic cleavage (Figure 4B, cluster 2). A third cluster displayed intermediate dephosphorylation kinetics (Figure 4B, cluster 1). These differential behaviors could provide a mechanism by which the meiotic divisions are specified and underlie the transition from MI to MII.

Figure 4 with 2 supplements see all
Download asset Open asset
Serine and threonine display distinct phosphorylation behaviors.

(A) Heatmap representation of a subset of sites that peak in phosphorylation in meiosis I (MI). Hierarchical clustering identifies three phosphorylation clusters with distinct temporal behaviors, indicated by green (cluster 1), cyan (cluster 2), and black (cluster 3) vertical lines. (B) Line graphs of temporal phosphorylation levels of sites within each of the three clusters. Color scale represents the distance from the mean. The number of single phosphorylation sites is indicated. (C) Sequence logos for over and underrepresented motifs within the three clusters. Threonine with proline in the +1 position followed by basic amino acids is overrepresented in cluster 3, which is dephosphorylated after MI. In contrast, cluster 3 is depleted for serine followed by acidic amino acids. Instead, cluster 1, which is more stably phosphorylated, is enriched for serine but depleted for threonine as the phosphoacceptor. (D) Cumulative frequency distribution of phosphopeptides with proline-directed serine or threonine phosphorylation. Significant differences in the population distribution as determined by Kolmogorov-Smirnov (KS) statistics (****p<0.0001). Only cluster 2 shows a significant difference in the dephosphorylation of SP vs TP phosphorylation sites. (E) Cumulative frequency distribution of phosphopeptides with proline-directed threonine phosphorylation with different amino acids in the +2 position. Only the most significant differences are indicated (****p<0.0001). (F) The average phosphorylation abundance of sites that peak in MI, comparing threonine (black bar) vs serine (gray bars) phosphorylation sites across stages as determined by phosphoproteomics. (G) Quantification of western blots with an antibody recognizing phosphorylated TPxK vs K/HpSP, revealing distinct behaviors at the meiosis I/meiosis II (MI/MII) transition for these phosphorylation sites.

To determine the mechanisms controlling these differential phosphorylation behaviors, we analyzed the specific phosphorylation events that are eliminated after MI (Figure 4B, cluster 3). Motif analysis of single phosphorylation sites in cluster 3 revealed that they predominantly occur on threonine with proline in the +1 position (TP sites). Indeed, by directly comparing the phosphorylation of SP vs TP sites in the three clusters, we found that TP phosphorylation declined more substantially than SP after MI (Figure 4D). On average, the mean phosphorylation of TP motifs detected by phosphoproteomics peaked in MI, but then declined substantially in MII. In contrast, SP motif phosphorylation decreased to a lesser extent (Figure 4F). To evaluate differential SP vs TP phosphorylation at other cell cycle stages, we additionally analyzed their relative abundances in prophase I-arrested oocytes and found that TP phosphorylation is lower on average (Figure 4—figure supplement 1B). Thus, our analysis reveals an unanticipated difference in the phosphorylation behavior of serine- vs threonine-containing sites at the MI/MII transition.

The behavior of these sites in our proteomics analysis suggests TP sites are selectively dephosphorylated between MI and MII, whereas SP sites remain phosphorylated. However, an alternative interpretation is that both SP and TP sites are dephosphorylated equally at the MI/MII transition, but then SP sites are selectively and rapidly re-phosphorylated in MII. To distinguish between these models, we collected samples with an increased temporal resolution (every 10 min) throughout meiosis and performed western blots using antibodies against phosphorylated pTP and pSP CDK motifs. Consistent with our MS results, overall TP site phosphorylation peaked in MI, followed by strong reduction in MII. In contrast, SP phosphorylation declined less substantially at the MI/MII transition (Figure 4G, Figure 4—figure supplement 1C). The same trends were observed with antibodies recognizing TPP and SPP phosphorylation motifs (Figure 4—figure supplement 2A,B). Thus, despite the transient inactivation of Cdk1 between the meiotic divisions (Okano-Uchida et al., 1998), SP phosphorylation remains relatively stable during this window. We note that MAPK activity remains high between both meiotic divisions (Figure 3G,H) and could contribute to the relative stability of SP phosphorylation events during this window. However, this model would still require a phosphatase activity to preferentially oppose TP phosphorylation sites to account for the difference in SP vs TP behavior. Taken together, these data reveal unexpected differences in SP vs TP phosphorylation sites at the MI/MII transition, with threonine residues being preferentially dephosphorylated and SP sites remaining phosphorylated.

In addition to differences in the phosphorylated residue (T vs S), we also found that the identity of surrounding residues correlated with phosphorylation behavior. Specifically, we identified an enrichment for basic amino acids starting in the +2 position (e.g., lysine) and a depletion of acidic amino acids (e.g., aspartic and glutamic acid) (Figure 4C). For example, TP sites with a small non-polar (A or G) or basic (K or R) amino acid were significantly more dephosphorylated than TP with acidic amino acids (E or D) or proline (Figure 4E, Figure 4—figure supplement 1A). Importantly, we note that the sites preferentially dephosphorylated at the MI/MII transition (cluster 3, TP followed by basic or nonpolar amino acids) match the sequence preferences of the phosphatase PP2A-B55, which we and others have recently reported (Cundell et al., 2016; Holder et al., 2020; Kruse et al., 2020; McCloy et al., 2015; Touati et al., 2019). These observations raise the possibility that PP2A-B55 drives the striking difference in behavior between TP and SP sites at the MI/MII transition.

PP2A-B55 inhibition is required for meiotic resumption from prophase I into MI in response to hormonal stimulation (Hara et al., 2012; Okumura et al., 2014), but a role for PP2A-B55 at the MI/MII transition has not been defined directly. To determine the activation state of PP2A-B55 at the MI/MII transition, and whether it could selectively dephosphorylate TP residues, we assessed the conserved PP2A-B55 regulatory pathway in our MS datasets. Cdk1 phosphorylation activates Greatwall kinase, which then phosphorylates and activates the B55 inhibitor ARPP19 (Figure 5A; Hara et al., 2012; Okumura et al., 2014). Therefore, a drop in cyclin B levels and CDK activity would drive the reactivation of PP2A-B55. Indeed, following MI, our dataset reveals a decrease in cyclin B levels and a corresponding reduction in the activating phosphorylation on Greatwall kinase (T204) and ARPP19 (S106). This would result in the release of PP2A-B55 to dephosphorylate its substrates at the MI/MII transition (Figure 5B, Figure 3—figure supplement 5C). Together, these observations suggest that PP2A-B55 is reactivated at the MI/MII transition to preferentially dephosphorylate specific substrates with TP/basic consensus motifs, through a conserved pathway involving ARPP19, Greatwall, and CDK/cyclin B (Figure 5B).

Figure 5 with 3 supplements see all
Download asset Open asset
PP2A-B55 substrate specificity is required for the MI/MII transition.

(A) Schematic of the Cdk-Gwl-ARPP19 pathway regulating PP2A-B55. (B) Relative phosphorylation levels of Gwl T204 and protein abundance of cyclin B. PP2A-B55 activity trace values are derived from the substrates in Figure 5—figure supplement 3A–C (1-mean phosphorylation levels). (C) Crystal structure of PP2A holoenzyme with the B subunit colored cyan. Mutated residues within the acidic surface are indicated in orange. (D) Immunofluorescence of meiotic time course of oocytes expressing wild-type or mutant B55 constructs. In contrast to wild-type oocytes, DE/A mutants successfully complete MI but fail in MII. The DE/K mutant oocytes successfully form the first meiotic spindle but fail to undergo homologous chromosome or sister centromere separation. The spindle poles do separate, ultimately resulting in two semi-distinct spindles. Microtubules were scaled nonlinearly. (E) Percentage of oocytes successfully completing MI and MII (control n = 65, DE/A n = 65, DE/K n = 49 oocytes; ****p<0.0001 by Fisher’s exact test). (F) Centrin2 staining of B55 wild-type- and DE/K mutant-expressing oocytes stained for Centrin2 reveals centriole separation at the spindle poles. Scale bars = 10 μm.

Selective dephosphorylation of TP vs SP residues is a conserved feature of meiosis

The selective dephosphorylation behavior that we observed suggests an important role for PP2A-B55 at the MI/MII transition. Preferential TP dephosphorylation has been previously reported to create a temporal order to the cellular events during mitotic exit (Cundell et al., 2016; Hein et al., 2017; McCloy et al., 2015; Touati et al., 2019). However, our work suggests that, in addition to temporal regulation of dephosphorylation at mitotic exit, preferential dephosphorylation of these motifs could provide a mechanism to specify the two meiotic divisions in oocytes. To test whether differential dephosphorylation is a conserved feature of meiosis, we reanalyzed a published phosphoproteomics dataset of early development in Xenopus laevis oocytes (Peuchen et al., 2017). Although the samples collected in this previous study did not have the temporal resolution to definitively capture the different stages of meiosis or the MI/MII transition, we used conserved inhibitory phosphorylation events on Cdk2 and PP1 to align these time points to corresponding samples from our analysis of synchronized sea star oocytes (Figure 5—figure supplement 1A,B). With this analysis, we identified time points that appear to correspond to an MI sample and an MII sample. Strikingly, we found that the phosphorylation sites that decrease in abundance between these samples in Xenopus oocytes were also enriched for TP-basic sequences but depleted for SP-acidic sequences (Figure 5—figure supplement 1C). Thus, based on this cross-species comparison, the differential behavior of SP and TP phosphorylation may represent a conserved regulatory mechanism for meiosis in animals and suggests that PP2A-B55 is an important conductor of the MI/MII transition.

Selective dephosphorylation by PP2A-B55 is required for the MI/MII transition

Our data are consistent with a model in which PP2A-B55 reactivation at the MI/MII transition drives the selective dephosphorylation of substrates to reorganize the cell division machinery and distinguish these divisions. We therefore sought to test the functional contribution of PP2A-B55 to the MI/MII transition. As the small molecule used to inhibit PP2A also inactivates other phosphoprotein phosphatases (e.g., Figure 3—figure supplement 7), we instead altered its specificity through mutations designed to disrupt B55 interactions with the downstream basic patch in its substrates (Cundell et al., 2016; Xu et al., 2008; Figure 5C). We generated mutations in complementary acidic residues in B55, changing these to alanine or creating charge-swap substitutions of these sites to lysine. Compared to wild-type B55, ectopic expression of the alanine mutant (DE/A) induced spontaneous GVBD and apoptosis in prophase I-arrested oocytes (Figure 5—figure supplement 2A,B). This result, along with the effect that we observed following calyculin A treatment (Figure 3—figure supplement 6A,C), further supports our model in which PP2A is specifically required to maintain the prophase I arrest.

We next tested the functional contributions of B55 to meiosis. Following hormonal stimulation, the expression of either of the B55 mutants had a potent dominant effect on meiotic progression. Oocytes expressing the alanine mutant (DE/A) successfully assembled the MI spindle and completed cytokinesis of polar body I. However, events after MI did not occur, with the MII spindles failing to form properly and displaying a ball-like morphology. Furthermore, anaphase and cytokinesis did not progress normally, resulting in a reduced rate of polar body II extrusion (Figure 5D,E). Oocytes expressing the charge-swap mutations (DE/K) displayed a more severe phenotype: the first meiotic spindle formed normally, but remained arrested, with the bivalent chromosomes maintaining cohesion and the co-oriented centromeres remaining fused, even at time points where wild-type control oocytes completed MII. However, although normal anaphase I and cytokinesis did not occur, the spindle poles eventually separated, ultimately resulting in two semi-distinct spindles. Staining for Centrin2 revealed that pole fragmentation was due to the separation of the pair of centrioles in the spindle pole (Figure 5F). Therefore, alterations to PP2A-B55 substrate specificity allow MI events to occur, but prevent changes to the cell division machinery that occur at the MI/MII transition. We, therefore, propose that PP2A-B55 serves as an essential regulator of the MI/MII transition by selectively dephosphorylating substrates to achieve exit from MI, but retaining sites that must remain phosphorylated for MII to occur.

Threonine-specific dephosphorylation is essential for the spatiotemporal control of PP2A-B55 substrates

The selective TP dephosphorylation at the MI/MII transition suggests a potential mechanism to encode MI- or MII-specific functions directly into substrates. The evolution and conservation of phosphorylation sites with higher or lower affinity for PP2A-B55 may provide temporal control for individual protein behaviors in meiosis. We identified several conserved phosphorylation sites on PRC1 (T470 in humans, T411 in sea star), TPX2 (T369 in humans, T508 in sea star), and Inner Centromere Protein (INCENP) (T59 in humans, T61 in sea star) (Figure 5—figure supplement 3A–C) that are known substrates of PP2A-B55 (Cundell et al., 2016; Hein et al., 2017; Hümmer and Mayer, 2009; Jiang et al., 1998). Consistent with the reactivation of PP2A-B55, these substrates display a stark decrease in phosphorylation following MI. In contrast to the behavior of these conserved PP2A-B55 substrate sites, we identified an SP phosphorylation event on Prc1 (S571), whose levels remain relatively constant during the meiotic divisions (Figure 5—figure supplement 3D). Furthermore, relative to other TP sites, an SP phosphorylation in Tpx2 (S625) was dephosphorylated to a lesser degree following MI (Figure 5—figure supplement 3E). To test the contributions of these sites, we focused on the chromosome passenger complex (CPC) subunit INCENP, which contains both stable serine phosphorylations as well as MI-specific threonine phosphorylations (Figure 6A). During anaphase of mitosis, the CPC transitions from the inner centromere to the central spindle, where it is required for cytokinesis (Carmena et al., 2012; Kaitna et al., 2000). However, the localization dynamics and impact of phosphorylation are not well defined in meiosis. Using a green fluorescent protein (GFP) fusion construct, we found that INCENP-GFP localized to centromeres in metaphase of MI, but then translocated to the central spindle at anaphase of MI. In MII, INCENP localization to centromeres was reduced, but it then returned at the midzone in anaphase II. As the maternal pronucleus formed at meiotic exit, INCENP localized to nucleolar structures (Figure 5—figure supplement 3F).

Figure 6
Download asset Open asset
Dephosphorylation of a conserved threonine is necessary for INCENP localization.

(A) Heatmap representation of relative phosphorylation levels of sites identified by phosphoproteomics. Phospho-centered amino acid sequences are provided on the right. (B) Immunofluorescence images of oocytes expressing INCENP-GFP in meiosis I along with either wild-type B55 or the DE/A mutant. DE/A expression reduces translocation of INCENP-GFP to the central spindle. Images are scaled individually to aid in visualization of localization changes. (C) Pixel intensity quantification of the unscaled midzone INCENP signal with wild-type or DE/A B55 expression (control n = 22, DE/A n = 13 oocytes; error bars represent mean and standard deviation; **p = 0.0014, Mann-Whitney test). (D) Immunofluorescence images of oocytes expressing wild-type INCENP-GFP or INCENPT61S. Wild-type INCENP translocates from centromeres to the central spindle in anaphase, whereas INCENPT61S increases at centromere localization but not at the central spindle. Images are scaled individually to aid in visualization of localization changes. (E) Pixel intensity quantitation of unscaled oocyte images at centromeres or central spindles in anaphase of meiosis I (WT n = 16, T61S n = 10 oocytes; error bars represent mean and standard deviation; *p = 0.0122, ****p<0.0001, Mann-Whitney test).

To test whether PP2A-B55 activity is required for the localization dynamics of INCENP, we co-expressed INCENP-GFP with either wild-type B55 or the B55 DE/A mutant. In wild-type B55-expressing oocytes, INCENP correctly translocated to the central spindle in anaphase of MI. In contrast, expression of the B55 DE/A mutant significantly reduced translocation of INCENP to the central spindle (Figure 6B,C). Therefore, the selective activity of PP2A-B55 is required for the proper localization behavior of INCENP at the MI/MII transition. Consistent with this model, we found that phosphorylation on INCENP T61 sharply decreased following MI, whereas other proline-directed sites on INCENP remained phosphorylated throughout the oocyte-to-embryo transition (Figure 6A, Figure 5—figure supplement 3C). Sequence alignment of sea star T61 with human INCENP T59 revealed the presence of a conserved downstream basic patch, supporting its potentially conserved dephosphorylation by PP2A-B55 (Figure 5—figure supplement 3C). In contrast, T506 remains relatively stable, likely due to the presence of downstream acidic amino acids, thereby disfavoring PP2A-B55 association (Figure 6A).

Finally, to test the role of INCENP T61 residue in MI, we used GFP fusion constructs with either wild-type PmINCENP or T61 replaced with serine (T61S), which is typically considered to be a conservative change that would preserve protein function. Wild-type INCENP localized to centromeres in metaphase of MI, but then relocalized to the central spindle at anaphase of MI. In contrast, INCENP(T61S)-GFP failed to translocate to the central spindle in anaphase of MI and remained at high levels at centromeres (Figure 6D,E), consistent with prior work in mitosis (Goto et al., 2006; Hümmer and Mayer, 2009). The retention of INCENP at centromeres in anaphase I suggests that this residue remains phosphorylated at the MI/MII transition when substituted with serine, a lower affinity substrate for PP2A-B55. Thus, the usage of serine vs threonine, modulated by adjacent charged amino acids, can directly encode a differential response to a common set of kinases and phosphatases, enabling rewiring events at key cell cycle transitions. Through this paradigm, the behaviors of individual proteins may be temporally coordinated with meiotic cell cycle progression to achieve specific behaviors.

Discussion

In this work, we define an extensive program of phosphorylation changes during the oocyte-to-embryo transition, spanning the complete developmental window from a prophase I arrest to the first embryonic cleavage. Using TMT-based proteomics, phosphoproteomics, and functional approaches, we found that overall protein levels are stable but that selected cell cycle regulators, including cyclins A and B, must undergo new protein synthesis for progression through meiosis. This general stability of the proteome across the oocyte-to-embryo transition in sea stars is consistent with prior results in oocytes of the vertebrate X. laevis (Peuchen et al., 2017; Presler et al., 2017). Therefore, post-translational modification instead serves as the primary and conserved control mechanism for meiosis and the transition to early development. Indeed, our analysis has uncovered a complex landscape of phosphorylation and dephosphorylation that underlies this developmental period. We found that the prophase I arrest is characterized by low overall phosphorylation and that maintaining this arrest requires phosphatase activity. Strikingly, although serine and threonine residues are often considered as conserved and interchangeable, we observed distinct behaviors for TP and SP phosphorylation at the MI/MII transition, with threonine sites being preferentially dephosphorylated. Such differential dephosphorylation suggests a novel paradigm for the regulatory control of oocytes, which must rapidly transition between the two meiotic divisions for the specialized meiotic cell cycle.

Our analysis reveals that, upon the decrease in Cdk activity at the MI/MII transition, selective dephosphorylation of TP vs SP sites by PP2A-B55 is essential for meiotic progression. By analzying the TP sites that are dephosphorylated after MI, we identified an enrichment for basic amino acids starting in the +2 position, matching the known consensus for PP2A-B55 substrates (Cundell et al., 2016; Kruse et al., 2020). Moreover, we found that PP2A-B55 is re-activated at the MI/MII transition, based on phosphorylation signatures of its inhibitory pathway, including Greatwall and ARPP19, as well as conserved PP2A-B55 substrates. Our results are consistent with prior studies in a related sea star reporting complete cyclin B degradation followed by rapid resynthesis to a lower level for MII and pronuclear fusion, resulting in the dephosphorylation of Greatwall kinase following MI and during first cleavage (Hara et al., 2012; Okano-Uchida et al., 1998; Tachibana et al., 2008). Thus, the temporal profile of PP2A-B55 activity leaves it poised to play an essential role in the MI/MII transition for specializing the meiotic cell cycle and division machinery.

Our work supports the emerging picture that threonine and serine phosphorylation is not interchangeable (Cundell et al., 2016; Deana et al., 1982; Deana and Pinna, 1988; Hein et al., 2017; Pinna et al., 1976) but instead represents an important regulatory mechanism to temporally control cellular processes. In cultured mitotic cells, threonine dephosphorylation is important for timely mitotic exit. For example, inhibitory phosphorylations on the APC/C regulator Cdc20 are conserved as threonines, whereas serine substitution mutants display delayed dephosphorylation and delayed Cdc20 activation (Hein et al., 2017). Our work extends this model into a new physiological context, wherein differential dephosphorylation provides not just a kinetic delay but also confers distinct behaviors for the two different meiotic divisions. For example, in our mutant analysis, substitution of threonine 61 for serine on INCENP, which is subject to inhibitory phosphorylation by Cdk1, resulted in its failure to translocate to the central spindle in anaphase of MI. This phosphorylation code is further modulated by the charge of adjacent amino acids, with positively charged residues favoring PP2A-B55 association. These findings reveal the importance for temporally coordinated dephosphorylation during meiotic progression, such that critical events including chromosome segregation and cytokinesis are properly synchronized with other cellular events. Such differential dephosphorylation represents a novel paradigm for the regulatory control of oocytes, which must rapidly transition between the two meiotic divisions for the specialized meiotic cell cycle.

These results point to PP2A-B55 as a critical conductor of cell cycle progression in oocytes. In diverse species, ARPP19/ENSA is required for exit from prophase I arrest in response to hormonal stimulation by inhibiting PP2A-B55 (Dupré et al., 2013; Dupré et al., 2014; Matthews and Evans, 2014; Okumura et al., 2014; Von Stetina et al., 2008). Intriguingly, although there is considerable conservation among the downstream phosphorylation steps associated with Cdk1 activation and meiotic progression, the upstream triggers of meiotic cell cycle resumption have diverged across evolution. For example, numerous species, including mammals, sea stars, and cnidarians (jellyfish), all utilize GPCRs that respond to hormonal signals and undergo communication with surrounding somatic ovarian cells to propagate signals (Jaffe and Norris, 2010; Quiroga Artigas et al., 2020; Tadenuma et al., 1992). However, the hormone/receptor pairs differ amongst species. In mice, luteinizing hormone acts upon receptors expressed in somatic cells and induces them to reduce the production of cyclic guanosine monophosphate (cGMP) and close gap-junction connections with the oocytes. This leads to a decrease in cyclic adenosine monophosphate (cAMP) levels in the oocyte, a drop in protein kinase A (PKA) activity, and the subsequent release from meiotic arrest by activation of Cdc25 (Conti and Chang, 2016; Norris et al., 2009). Similarly, in X. laevis, a drop in PKA activity is associated with meiotic resumption downstream of progesterone stimulation, although the nature of the roles of PKA and cAMP is disputed (Dupré et al., 2014; Nader et al., 2016). In echinoderms, meiotic resumption is stimulated by the hormone 1-MeAd, which is synthesized by surrounding somatic cells and acts upon an unidentified GPCR, leading to the activation of the trigger kinase SGK (Kishimoto, 2018). Although these upstream steps are divergent, they still result in the downstream activation of Cdc25 and inhibition of Myt1, as in vertebrate oocyte systems.

Our data from sea stars indicate that phosphatase activity is critical not only for maintaining the prophase I arrest but also to drive the transition between the two meiotic divisions. In support of a potentially conserved role for PP2A-B55 in the MI/MII transition, prior work in mouse oocytes found that the Greatwall ortholog Mastl is essential to proceed past MI (Adhikari et al., 2014). CDK and MAPK, combined with a wave of opposing PP2A-B55 activity at the MI/MII transition, provide a pacemaker for meiotic cell cycle progression. By encoding PP2A-B55 affinity into substrates, different temporal phosphorylation profiles can be achieved yielding subsets of cell division proteins that are constitutively active but others that are uniquely modified in MI or MII. The rapidity of the two meiotic divisions (separated by only 30 min in sea stars) places a selective pressure for specific substrates to be promptly dephosphorylated. In contrast, other sites may need to remain phosphorylated to prevent exit into interphase, creating evolutionary pressure for conservation as serine. The ability to segregate heritable material through meiosis is essential for organismal fitness. We propose that evolution has fine-tuned substrate dephosphorylation by selecting for amino acid sequences that favor or disfavor the interaction with PP2A-B55, thereby enabling precise temporal coordination of events in the oocyte-to-embryo transition.

Materials and methods

Key resources table
Reagent type
(species) or resource		DesignationSource or referenceIdentifiersAdditional information
Strain, strain background (Patiria miniata)South Coast Bio, LLCWild-caught animals
AntibodyAnti-CENP-C (rabbit polyclonal)Swartz et al., 2019IF(1:5000)
AntibodyAnti-cyclin B (rabbit polyclonal)Ookata et al., 1992WB(1:1000)
AntibodyAnti-alpha tubulin, DM1alpha (mouse monoclonal)Sigma-AldrichCat# T9026, RRID:AB_477593IF(1:5000)
AntibodyAnti-phoshpo-Cdk1 Y15Cell Signaling TechnologiesCat# 4539WB(1:3000)
AntibodyAnti-phospho ERK1/2 T202/Y204Cell Signaling TechnologiesCat# 9101WB(1:3000)
AntibodyAnti-pTPxK CDK consensusCell Signaling TechnologiesCat# 14371WB(1:3000)
AntibodyAnti-(K/H)pSP CDK consensusCell Signaling TechnologiesCat# 9477WB(1:3000)
AntibodyAnti-pTPP CDK consensusCell Signaling TechnologiesCat# 5757WB(1:3000)
AntibodyGFP boosterChromotekgba488-100, RRID:AB_2631386IF(1:500)
Recombinant DNA reagentpZS281
(plasmid)		This paperTo be deposited to AddgeneWild-type B55-GFP (pCS2+eight backbone)
Recombinant DNA reagentpZS282
(plasmid)		This paperTo be deposited to AddgeneDE/A mutant B55-GFP (pCS2+eight backbone)
Recombinant DNA reagentpZS283
(plasmid)		This paperTo be deposited to AddgeneDE/K mutant B55-GFP (pCS2+eight backbone)
Recombinant DNA reagentpZS294
(plasmid)		This paperTo be deposited to AddgeneB55-mCherry (pCS2+eight backbone)
Recombinant DNA reagentpZS295
(plasmid)		This paperTo be deposited to AddgeneDE/A mutant B55-mCherry (pCS2+eight backbone)
Recombinant DNA reagentpZS13
(plasmid)		This paperTo be deposited to AddgeneINCENP-GFP (pCS2+eight backbone)
Recombinant DNA reagentpZS259
(plasmid)		This paperTo be deposited to AddgeneINCENP-GFP T61S mutant (pCS2+eight backbone)
Recombinant DNA reagentGST-ARPP19This paperTo be deposited to AddgeneFor protein expression
Recombinant DNA reagentpCS2+eight backboneGökirmak et al., 2012RRID:Addgene_34952Pol III-based shRNA backbone
Sequence-based reagentStandard Control MorpholinoGene-Tools, LLCCCT CTT ACC TCA GTT ACA ATT TAT
Sequence-based reagentCyclin A morpholinoGene-Tools, LLCMorpholino antisense oligoTTCACTTTGTTCCCGAGATTAAC
Sequence-based reagentCyclin B morpholinoGene-Tools, LLCMorpholino antisense oligoTAACCAATGCGAGTTCCGAGGAG
Commercial assay or kitmMessage mMachine SP6 in vitro transcription kitInvitrogenCat# AM1340
Commercial assay or kitPoly(A) tailing kitInvitrogenCat# AM1350
Commercial assay or kitCan Get Signal Immunoreaction Enhancer SolutionCosmo Bio USACat# TYB-NKB-101T
Chemical compound, drug1-MethyladenineAcros OrganicsAC20131-1000
Chemical compound, drugCalyculin ASanta Cruz BiotechnologySC-24000APP1 and PP2A inhibitor
Chemical compound, drugEmetineSigma AldrichE2375-500MGTranslation inhibitor
Chemical compound, drugTautomycetinGift from Dr. Richard HonkanenPP1 inhibitor
Software, algorithmSPSSSPSSRRID:SCR_002865
OtherHoechst 33342 stainLife TechnologiesCat# H1399(2 µg/ml)
OtherProlong Gold Antifade ReagentLife TechnologiesP36930
Commercial assay or kitTMT10plex Isobaric Label Reagent Set plus TMT11-131C Label ReagentThermo ScientificA34808
Commercial assay or kitHigh-Select Fe-NTA Phosphopeptide Enrichment KitThermo ScientificA32992

Experimental models and subject details

Request a detailed protocol

Sea stars (P. miniata) were wild-caught by South Coast Bio Marine (http://scbiomarine.com/) and kept in artificial seawater aquariums at 15°C. Intact ovary and testis fragments were surgically extracted as previously described (Swartz et al., 2019). Oocyte samples for proteomics were collected from a single female within the same season to maximize synchrony between time courses. Stereotypical meiotic timings were visually confirmed in live oocytes using GVBD and polar body emission as metrics, before snap freezing samples in liquid nitrogen. The synchrony data in Figure S1B were performed with oocytes from a second animal collected at the same season and location that was confirmed to undergo meiosis at a similar rate. 293T cells were obtained from commercial sources and were routinely tested for mycoplasma by polymerase chain reaction (PCR).

Ovary and oocyte culture

Request a detailed protocol

Ovary fragments were maintained in artificial seawater containing 100 units/ml pen/strep solution. Intact ovary fragments were cultured this way for up to 1 week until oocytes were needed, with media changes every 2–3 days. Isolated oocytes were cultured for a maximum of 24 hr in artificial seawater with pen/strep. To induce meiotic re-entry, 1-MeAd (Acros Organics) was added to the culture at a final concentration of 10 μM. For fertilization, extracted sperm was added to the culture at 1:1,000,000 dilution prior to emission of the first polar body. For emetine treatments (Figure 2), oocytes were pre-treated with 10 μM emetine (Sigma-Aldrich) for 30 min prior to hormonal stimulation.

Mass-spectrometry sample preparation

Request a detailed protocol

Oocytes were collected by centrifugation at 200 xg and resuspending pellets and washed one time with wash buffer (50 mM 4-(2-hydroxyethyl) -1-piperazineethanesulfonic acid (HEPES), pH 7.4, 1 mM ethylene glycol tetraacetic acid (EGTA), 1 mM MgCl2, 100 mM KCl, 10% glycerol), pelleted with all excess buffer removed, and snap frozen in liquid nitrogen. Samples were lysed and proteins were digested into peptides with trypsin. Oocyte pellets were lysed in ice-cold lysis buffer (8 M urea, 25 mM Tris-HCl, pH 8.6, 150 mM NaCl, phosphatase inhibitors (2.5 mM beta-glycerophosphate, 1 mM sodium fluoride, 1 mM sodium orthovanadate, 1 mM sodium molybdate), and protease inhibitors (one mini-Complete EDTA-free tablet per 10 ml lysis buffer; Roche Life Sciences)) and sonicated three times for 15 s each with intermittent cooling on ice. Lysates were centrifuged at 15,000 xg for 30 min at 4°C. Supernatants were transferred to a new tube and the protein concentration was determined using bicinchoninic acid (BCA) assay (Pierce/ThermoFisher Scientific). For reduction, dithiothreitol (DTT) was added to the lysates to a final concentration of 5 mM and incubated for 30 min at 55°C. Afterwards, lysates were cooled to room temperate and alkylated with 15 mM iodoacetamide at room temperature for 45 min. The alkylation was then quenched by the addition of an additional 5 mM DTT. After sixfold dilution with 25 mM Tris-HCl, pH 8, the samples were digested overnight at 37°C with 1:100 (w/w) trypsin (Promega). The next day, the digest was stopped by the addition of 0.25% trifluoroacetic acid (TFA) (final v/v), centrifuged at 3500 xg for 15 min at room temperature to pellet precipitated lipids, and peptides were desalted (Zecha et al., 2019). Peptides were lyophilized and stored at −80°C until further use.

TMT labeling was carried out as previously described (Zecha et al., 2019). For proteomics analysis, an aliquot containing ~ 100 µg of peptides was resuspended in 133 mM HEPES (SIGMA), pH 8.5, TMT reagent (Thermo Scientific) stored in dry acetonitrile (ACN) (Burdick and Jackson) was added, and vortexed to mix reagents and peptides. After 1 hr at room temperature, an aliquot was withdrawn to check for labeling efficiency by LC-MS3 analysis, while the remaining reaction was stored at −80°C. Once labeling efficiency was confirmed to be at least 95%, each reaction was quenched by addition of ammonium bicarbonate to a final concentration of 50 mM for 10 min, mixed, diluted with 0.1% TFA in water, and desalted. The desalted multiplex was dried by vacuum centrifugation and separated by offline pentafluorophenyl (PFP)-based reverse-phase high-performance liquid chromatography (HPLC) fractionation as previously described (Grassetti et al., 2017).

For phosphoproteomic analysis, ~4 mg of peptides was enriched for phosphopeptides using a nitrilotriacetic acid (Fe-NTA) phosphopeptide enrichment kit (Thermo Scientific) according to instructions provided by the manufacturer and desalted. Phosphopeptides were labeled with TMT reagent and labeling efficiency was determined as described for the protein sample. Once labeling efficiency was confirmed to be at least 95%, each reaction was quenched with 5 mM ammonium bicarbonate, mixed, diluted with 0.1% trifluoroacetic acid (TFA) in water, and desalted. The desalted multiplex was dried by vacuum centrifugation and separated by offline PFP-based reverse-phase HPLC fractionation as previously described (Grassetti et al., 2017).

Proteomic and phosphoproteomic analyses of oocyte-to-embryo time courses were performed on an Orbitrap Fusion mass spectrometer (Thermo Scientific) equipped with an Easy-nLC 1000 (Thermo Scientific) (Senko et al., 2013). Peptides were resuspended in 8% methanol/1% formic acid and separated on a reverse-phase column (40 cm length, 100 μm inner diameter, ReproSil, C18 AQ 1.8 μm 120 Å pore) pulled and packed in-house across a 2-hr gradient from 3% acetonitrile/0.0625% formic acid to 37% acetonitrile/0.0625% formic acid. The Orbitrap Fusion was operated in data-dependent, SPS-MS3 quantification mode (McAlister et al., 2014; Ting et al., 2011). For proteomics analysis, an Orbitrap MS1 scan was taken (scan range = 350–1300 m/z, R = 120K, automatic gain control (AGC) target = 2.5e5, max ion injection time = 100 ms). This was followed by a data-dependent iontrap MS2 scan on the most abundant precursors for 3 s in rapid scan modem (AGC target = 1e4, max ion injection time = 40 ms, collision-induced dissociation (CID) collision energy = 32%), followed by an Orbitrap MS3 scan for quantification (R = 50K, AGC target = 5e4, max ion injection time = 100 ms, higher energy collision dissociation (HCD) collision energy = 60%, scan range = 110–750 m/z, synchronous precursors selected = 10). For phosphoproteomics analysis, an Orbitrap MS1 scan was taken (scan range = 350–1300 m/z, R = 120K, AGC target = 3.5e5, max ion injection time = 100 ms), followed by data-dependent Orbitrap MS2 scans on the most abundant precursors for 3 s. Ion selection (to ensure nominally informative peptide length for unambiguous sequence assignment); charge state = 2: minimum intensity 2e5, precursor selection range 625–1200 m/z; charge state 3: minimum intensity 3e5, precursor selection range 525–1200 m/z; charge states 4 and 5: minimum intensity 5e5 (quadrupole isolation = 0.7 m/z, R = 30K, AGC target = 5e4, max ion injection time = 80 ms, CID collision energy = 32%). This was followed by an Orbitrap MS3 scan for quantification (R = 50K, AGC target = 5e4, max ion injection time = 100 ms, HCD collision energy = 65%, scan range = 110–750 m/z, synchronous precursors selected = 5).

The phosphoproteomic analysis of calyculin A-treated occytes was performed on the Orbitrap Lumos mass spectrometer, wherein an Orbitrap MS1 scan was taken (scan range = 350–1250 m/z, R = 120K, AGC target = 2.5e5, max ion injection time = 50 ms), followed by data-dependent Orbitrap MS2 scans on the most abundant precursors for 2 s. Ion selection (to ensure nominally informative peptide length for unambiguous sequence assignment); charge state = 2: minimum intensity 2e5, precursor selection range 650–1250 m/z; charge state 3: minimum intensity 3e5, precursor selection range 525–1250 m/z; charge states 4 and 5: minimum intensity 5e5 (quadrupole isolation = 1 m/z, R = 30K, AGC target = 5e4, max ion injection time = 55 ms, CID collision energy = 35%). This was followed by Orbitrap MS3 scans for quantification (R = 50K, AGC target = 5e4, max ion injection time = 100 ms, HCD collision energy = 65%, scan range = 100–500 m/z, synchronous precursors selected = 5).

High-resolution tandem mass spectra were searched using COMET with a precursor ion tolerance of +/-1 Dalton and a fragment ion tolerance of +/-0.02 m/z, static mass of 229.162932 on peptide N-termini and lysines and 57.02146 Da on cysteines, and a variable mass of 15.99491 Da on methionines against a target-decoy version of the P. miniata proteome sequence database. For phosphoproteomics analysis, 79.96633 Da on serines, threonines, and tyrosines was included as an additional variable mass. The resulting peptide spectral matches (PSMs) were filtered to a <1% false discovery rate (FDR) using the target-decoy strategy with a typical precursor ion mass filter of +/-1.5 parts-per-million (PPM) mass accuracy and corresponding XCorr and dCn values. Quantification of spectra of liquid chromatography with tandem mass spectrometry (LC-MS/MS) was performed using an in-house-developed software. Probability of phosphorylation site localization was determined by PhosphoRS (Taus et al., 2011).

The P. miniata proteome sequence database was generated using published paired-end RNA-seq data for a P. miniata ovary sample (SRX445851/SRR1138705), with the Agalma transcriptome pipeline (Dunn et al., 2013). Each protein sequence from this transcriptome was assessed for similarity with known proteins in the NCBI nr (nonredundant) sequence database using NCBI BLAST. Transcriptome sequences were annotated with the accession and sequence similarity metrics of their top BLAST hits. This annotation procedure was repeated for an unannotated protein database sourced from EchinoBase to generate a composite database. Duplicate sequences and subsequences within the composite database were removed to reduce redundancy.

Construct generation

Request a detailed protocol

P. miniata INCENP was identified using the genomic resources at echinobase.org and previously published ovary transcriptomes (Kudtarkar and Cameron, 2017; Reich et al., 2015). Wild-type INCENP was amplified from first-strand complementary DNA (cDNA) reverse transcribed from total ovary mRNA. T61S mutants were generated by overlap extension PCR. These cDNAs were then cloned into pCS2+eight as c-terminal GFP fusions (Gökirmak et al., 2012). Wild-type and mutant versions of B55 were synthesized (Twist Biosciences) and cloned into pCS2+eight with standard restriction enzyme methods.

Oocyte microinjection

Request a detailed protocol

For expression of constructs in oocytes, plasmids were linearized with NotI to yield the linear template DNA. mRNA was transcribed in vitro using mMessage mMachine SP6 and the polyadenylation kit (Life Technologies), then precipitated using lithium chloride solution. Prophase I-arrested oocytes were injected horizontally in Kiehart chambers with approximately 10–20 picoliters of mRNA solution in nuclease-free water. B55 constructs were injected at 500 ng/μl, and INCENP constructs were injected at 1000 ng/μl. After microinjection, oocytes were cultured for 18–24 hr to allow time for the constructs to translate before 1-MeAd stimulation. Custom-synthesized cyclin morpholinos, or the Gene Tools standard control, were injected at 500 μM immediately before 1-MeAd stimulation (Gene Tools).

Immunofluorescence, imaging, and immunoblots

Request a detailed protocol

Oocytes were fixed at various stages in a microtubule stabilization buffer as described previously (2% paraformaldehyde, 0.1% triton X-100, 100 mM HEPES, pH 7.0, 50 mM EGTA, 10 mM MgSO4, 400 mM dextrose) for 24 hr at 4° C (von Dassow et al., 2009). The oocytes were then washed three times in phosphate-buffered saline (PBS) with 0.1% triton X-100 and blocked for 15 min in AbDil (3% bovine serum albumin (BSA), 1 X tris-buffered saline (TBS), 0.1% triton X-100, 0.1% sodium azide). Primary antibodies diluted in AbDil were then applied and the oocytes were incubated overnight at 4°C. Anti-CENP-C and alpha-tubulin (DM1α, Sigma) antibodies were used at 1:5000 ratio. DNA was stained with Hoechst. GFP booster (Chromotek) was used at 1:500 to improve the signal from GFP fusion constructs. The oocytes were compressed under coverslips in ProLong Gold antifade Mountant (Thermo Fisher). Images were collected with a DeltaVision Core microscope (Applied Precision/GE Healthsciences) with a CoolSnap HQ2 CCD camera and a x100 1.40 NA Olympus U-PlanApo objective. Confocal images (Figure 2G) were collected with a Yokogawa W1 spinning disk microscope. Images were processed with Fiji and scaled equivalently across conditions unless otherwise specified (Schindelin et al., 2012). For western blot analyses, oocytes were lysed in Laemmli samples buffer, separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), and analyzed by western blotting. Antibodies were diluted in 3% milk in TBST (TBS with Tween 20) and phosphatase inhibitors anti-phospho Cdk1 Y15 (Cell Signaling, #4539), anti-phospho ERK1/2 T202/Y204 (Cell Signaling, #9101; this antibody detects singly and dually phosphorylated ERK1/2 and the total intensity of singly and dually phosphorylated ERK1/2 in sea stars suggests that Y193 is the dominant phosphoform), phospho CDK substrate [pTPXK] (Cell Signaling, #14371), [(K/H)pSP] (Cell Signaling, #9477), [pSPP] (Cell Signaling, #14390), and [pTPP] (Cell Signaling, #5757). Antibodies raised against phosphorylation sites on human proteins were evaluated for cross-reactivity and judged to do if (1) they only recognized one band, (2) the size of the band migrated at the expected molecular weight, and (3) the phosphorylation pattern matched the MS quantification. Anti-P. pectinifera cyclin B antibody was a gift from the Kishimoto laboratory (Ookata et al., 1992). Cyclin B western blotting was conducted by diluting the primary and secondary antibodies into Can Get Signal Immunoreaction Enhancer Solution as described by the manufacturer’s instructions (Cosmo Bio USA).

Protein expression

Request a detailed protocol

P. miniata ARPP19 was identified using the genomic resources at echinobase.org and previously published ovary transcriptomes (Kudtarkar and Cameron, 2017; Reich et al., 2015). Wild-type ARPP19 was amplified from first-strand cDNA reverse transcribed from total ovary mRNA and cloned into pCS2+8. ARPP19 was sub-cloned into pGEX4T3 and transformed into Rosetta (DE3). Expression, purification, and thiophosphorylation were performed as previously described (Gharbi-Ayachi et al., 2010).

In vitro phosphatase assay

Request a detailed protocol

PP1c was diluted in assay buffer (0.15 M NaCl, 30 mM HEPES, pH 7.0, 1 mM DTT, 0.1 mg/ml BSA, 1 mM ascorbate, 1 mM MnCl2) with a final assay concentration of 0.5 nM PP1c and 75 μM 6,8-difluoro-4-methylumbelliferyl phosphate (DiFMUP) (Invitrogen, D6567). 6,8-difluro-4-methylumbelliferone (DiFMU) was measured every minute for 45 min (Ex 360 nm, Em 460 nm) in the presence of indicated concentrations of calyculin A and tautomycetin.

Quantification and statistical analysis

Request a detailed protocol

For the time-course proteomic analysis, peptide TMT intensities were summed on a per protein basis. Replicate multiplexes were normalized to each other based on an internal standard consisting of a mixture of oocytes from all investigated stages. Protein intensities were adjusted based on the total TMT reporter ion intensity in each channel. Individual protein intensity across each time course was scaled between 0 and 1. For the time-course phosphoproteomic analysis, replicate multiplexes were normalized to each other based on the internal standard. Phosphopeptide intensities were adjusted based on the total amount of protein input in the respective channel to account for differences in input and the median total TMT reporter ion intensity of the respective condition. Individual phosphopeptide intensity across each time course was scaled between 0 and 1. A Pearson correlation coefficient for the time-course data was calculated as pairwise correlation for proteins and phosphopeptides identified in two time-course series and as multiple correlations for proteins identified in three time-course series in Excel. A p-value was calculated from the Pearson correlation coefficient. Proteins quantified in all three time-course series were analyzed by hierarchical clustering using Euclidean distance and average linkage in Perseus. Phosphorylation sites with a localization score of 0.9 or higher and quantified with a correlation coefficient of 0.8 more (corresponding to a p-value of 0.05 or less) were analyzed by hierarchical clustering using Euclidean distance and average linkage in Perseus. For the calyculin A treatment, the 0-min and 30-min and the 0-min and 70-min time points were compared. p-values were calculated using a two-tailed Student’s t-test assuming unequal variance. Average phosphorylation abundances of phosphorylation site upon 0-, 30-, and 70-min calyculin A treatment were compared between prophase I and other phases by 2x3 Fisher's exact test.

For GO analysis of proteins with significant changes in abundance during oocyte-to-embryo transition, human homologs of P. miniata proteins were identified using NCBI's BLAST+. A custom R script was then used to filter the BLAST results to only high-confidence matches (E-value<0.01) and to remove redundant matches (https://github.com/BrennanMcEwan/starfish-phos-pub-code; copy archived at McEwan, 2021). The resulting list was analyzed using WebGestaltR to find enriched GO terms (Liao et al., 2019). Motif analysis for selected and deselected amino acids surrounding phosphorylation sites in each cluster was performed using the Icelogo web interface using single localized phosphorylation sites. Phosphorylation sites in each cluster (experimental set) were compared to all single localized phosphorylation sites maximally phosphorylated in MI (reference set). Statistical analysis to determine population difference of dephosphorylation preferences at the MI/MII transition was performed using an unpaired nonparametric Kolmogorov-Smirnov (KS) t-test. RVxF motifs based on the definition [KRL][KRSTAMVHN][VI] [^FIMYDP][FW] (^ indicates excluded) with an S or T at the ‘x’ position or within two amino acids before and after the motif that were phosphorylated were identified in the phosphoproteomics time-course dataset and the average and standard deviation of the phosphorylation abundance were plotted.

Graphing and statistical analyses involving the scoring of meiotic phenotypes (e.g., Figure 5E) were performed using Prism (GraphPad). The statistical tests used to calculate p-values are indicated in the figure legends. Pixel intensity quantifications of INCENP fluorescence images (i.e., Figure 6D) were conducted using Fiji. Circular regions of interest (ROIs) were placed over inner centromeres or the central spindle (7 px and 30 px diameter, respectively), along with an adjacent ROI to measure the background intensity of equivalent size by RawIntDen. The background measurements were then subtracted from the centromere or midzone measurement, and the resulting value was normalized to the selection area to allow comparison between inner centromere and midzone intensity. The crystal structure of the PP2A-B55 holoenzyme (Figure 5C) was displayed using UCSF ChimeraX (Goddard et al., 2018).

Data availability

Raw MS data for the experiments performed in this study are available at MassIVE and ProteomeXchange, accession number: PXD020916. Plasmids generated from this study are deposited to Addgene. Custom R script is available at Github (https://github.com/BrennanMcEwan/starfish-phos-pub-code; copy archived at https://archive.softwareheritage.org/swh:1:rev:7d81808b1697cf470dcd1127725e8a94c8752659).

The following data sets were generated
    1. Swartz SZ
    2. Nguyen HT
    3. McEwan BC
    4. Adamo ME
    5. Cheeseman IM
    6. Kettenbach AN
    (2020) ProteomeXchange
    ID PXD020916. Selective dephosphorylation by PP2A-B55 directs the meiosis I - meiosis II transition in oocytes.
    http://proteomecentral.proteomexchange.org/cgi/GetDataset?ID=PXD020916

References

  36. Book
    1. Jaffe LA
    2. Norris RP
    (2010)
    Initiation of the Meiotic Prophase-to-Metaphase Transition in Mammalian Oocytes
    Oogenesis.
    1. Tosuji H
    2. Kato Y
    3. Fusetani N
    4. Nakazawa T
    (1991)
    Induction of oocyte maturation by calyculin-a in starfish
    Zool Sci 8:201–205.

Decision letter

  1. Jon Pines
    Reviewing Editor; Institute of Cancer Research Research, United Kingdom
  2. Anna Akhmanova
    Senior Editor; Utrecht University, Netherlands
  3. Peter Lenart
    Reviewer; Max Planck Institute for Biophysical Chemistry, Germany

In the interests of transparency, eLife publishes the most substantive revision requests and the accompanying author responses.

Acceptance summary:

This study undertakes a comprehensive proteomics/phosphoproteomics analysis of meiosis and the beginning of embryogenesis in the sea star, Pateria miniata. This dataset should be a valuable resource for the community. The study focuses on the regulatory circuits that drive the oocyte to embryo transition and identifies a divergent pattern of dephosphorylations consistent with an increased activity of the phosphatase PP2A-B55 that prefers to dephosphorylate phospho-threonine over phospho-serine. The differential phosphorylation observed provides a novel angle that may provide the framework for a better understanding of the steady states stabilizing meiosis I and meiosis II in the oocytes.

Decision letter after peer review:

[Editors’ note: the authors submitted for reconsideration following the decision after peer review. What follows is the decision letter after the first round of review.]

Thank you for submitting your work entitled "Selective dephosphorylation by PP2A-B55 directs the meiosis I – meiosis II transition in oocytes" for consideration by eLife. Your article has been reviewed by 2 peer reviewers, and the evaluation has been overseen by a Reviewing Editor and a Senior Editor. The following individual involved in review of your submission have agreed to reveal their identity: Peter Lenart (Reviewer #1).

Our decision has been reached after consultation between the reviewers. Based on these discussions and the individual reviews below, we regret to inform you that your work will not be considered further for publication in eLife.

The referees agreed on the quality of your data, and that this will be a useful resource for the community. The main concern was that a similar data set has already been published for Xenopus meiosis and that the conclusions from your study do not provide the insights that would warrant publication in a general audience publication such as eLife. If, however, you are able to address the substantive questions raised by Reviewer 2 – in particular, validating the role of NIPP1 in vivo – then we would be prepared to consider a further submission, but we judge that this will take more than two months and that the fairest decision is to return the paper to you now for you to consider your next step.

Reviewer #1:

The manuscript presents the, to my knowledge, first complete 'proteomics time course' through oocyte meiotic divisions up to the first embryo cleavage. This is made possible by taking advantage of the uniquely suited model system, starfish oocytes combined with state-of-the-art quantitative TMT (tandem mass tag) phosphoproteomics. The data is high quality and is a very valuable resource for the oocyte community with potential relevance for improving assisted reproduction procedures. Analysis of the data confirms several known mechanisms of phospho-regulation, and reveals new mechanisms; in particular, the authors show that the inherent phospho-threonine preference of PP2A-B55 phosphatase plays a role in differential regulation of first and second meiotic divisions.

Starfish oocytes are uniquely suited for the study of meiosis, because oocytes are available in large quantities and proceed through meiosis synchronously upon addition of the maturation hormone, and if fertilized, even continue into embryonic development. The authors took advantage of this system to carry out a 'proteomics time course': they have taken samples at 6 time points and analyzed them in a multiplexed TMT phosphoproteomics experiment. They have carefully validated the specific stages by immunostaining and microscopic analysis of oocytes from each time point.

The data appears overall high quality with over 4600 proteins and over 10 000 phosphopeptides identified in 3 replicates, comparable to studies in other species (e.g. on the range of 6000 proteins identified in Xenopus oocytes). This renders this dataset, which is made publicly available in its entirety, an extremely valuable resource to the community.

The data reveals an overall rather constant protein composition over the entire process of two meiotic and one mitotic cleavage division, with less than 1% of the proteins showing more than a 2-fold change. This was assumed to be the case before, and shown here for the first time. The authors validate a couple of the newly translated proteins, and show that -- consistent with earlier findings in starfish and other species -- the second meiotic M-phase require production of cyclin B protein, whereas the first mitotic cleavage division requires both cyclin A and B synthesis.

Phosphoproteomics reveals massive waves of phosphorylation and dephosphorylation, again consistent with seminal earlier studies, for example using in vivo P32 isotope labeling. The authors identify many of the previously know activating/inactivating phosphosites on regulatory proteins such as cdk1, Greatwall, ARPP19 and others. These show that before meiotic entry the high activity of phosphatases keeps phosphorylation levels low. Most excitingly, the data show that PP2A-B55 is reactivated in the MI/MII transition and drives a wave of dephosphorylation specifically affecting threonine residues, which is confirmed experimentally by overexpressing dominant negative forms of B55. Intriguingly, this preference of PP2A-B55 to TP (threonine) vs SP (serine) sites appears to have key regulatory importance. To this end, the authors show using INCENP as a model substrate that swapping threonine to serine residue indeed affects the timing of meiotic progression (that in the case of INCENP is timely relocalization from the centromeres to the spindle midzone).

Taken together, the manuscript is (i) a highly valuable resource for the community; (ii) it directly confirms several earlier indirect observations about phosphoregulation of meiotic progression; (iii) as a new result, it reveals the key importance of PP2A-B55 in meiosis-specific regulation of the cell cycle.

Overall, I am happy to recommend the manuscript for publication with minor changes:

1. I find it very important to validate the quality and synchronicity of samples. Therefore, it would be important to state the number of oocytes analyzed for Figure S1B, and also more clearly state whether these samples have been taken from the same starfish animal and season.

2. I greatly appreciate that the authors are describing the methods in detail and provide the raw proteomics data for download. Just one more detail: could they provide the MS1 and MS2 errors?

3. I would be curious to see the localization of INCENP also in MII and AnaII. Could the authors provide these additional immunofluorescence data?

Reviewer #2:

Swartz et al. applies MS analyses to investigate protein abundance and modification during maturation of Patiria miniata oocytes. MS analyses revealed that 99% of the app. 5000 proteins reproducibly identified displayed a max fold change of less than a 1.2-fold change indicating that meiotic cell cycle transitions are not associated with bulk changes in protein levels. Using Emetine, the authors show that progression to MII requires protein synthesis, in particular synthesis of cyclin A and B. Phospho-proteome analyses revealed that the majority of phosphorylation events occur at serine residues, followed by threonine and tyrosine residues. Prophase-I arrested oocytes displayed the lowest overall phosphorylation state. Consistent with what is known from other organisms, inhibitory phosphorylations of Cdk were found in Prophase-I arrested oocytes, but not during oocyte maturation. Then, the authors focus on PP1 and PP2A. Calyculin A treatment induced spontaneous GVBD. However, treated oocytes did not progress beyond a GVBD-like state. Cluster analyses identified TP sites as being more likely to be efficiently dephosphorylated upon exit from MI compared to SP sites. Consistent with studies in other organisms, activities of Cdk and B55 are inversely regulated such that either Cdk or B55 are active. Expression of B55 charge-swap mutations induce severe phenotypes at the MI/MII transition. T to S replacement in INCEP prevented the relocalization of the GFP-tagged fusion protein from centromeres to the spindle midzone at anaphase I.

The decision on this manuscript is not an easy one. On the one hand, the authors have done a lot of experiments and the collected data are of high quality. On the other hand, the provided insight are not particular novel or exciting and often the authors leave interesting insights without further following up. It does not come to a surprise that meiotic progression in Patiria miniata is not mediated by bulk protein turnover, but by few changes in key cell cycle regulators such as cyclins (Figures 1 and 2). Likewise, it is not surprising that phosphorylation plays a major role in meiotic cell cycle progression and that changes in Cdk and PP2A-B55 activity mediate those (Figure 3).

Given this mixed overall impression – not much new insights and some loose ends of potentially interesting insights – I think that the manuscript is more suitable for a more specialized journal rather than eLife. Apart from the aforementioned shortcomings, I have the following concerns.

– The authors show by Emetine treatment that protein synthesis is required for meiotic progression, but apart from cyclin B the authors don´t provide a correlation to their MS data on changes in protein abundance. Are there other interesting (cell cycle-related) candidates among the strongly upregulated proteins in the MS data that could be required for meiotic progression?

– The authors provide many indirect, but no direct evidence that PP1 activity is indeed high in Prophase I. They investigate phospho-regulation of NIPP1 in vitro without translating these insights to the oocyte. How meaningful / physiological relevant are these in vitro insights for the Prophase I arrest. Additionally, the analysis is focused on a single PP1 inhibitor, NIPP1, but PP1 activity could be regulated by several other inhibitors as well. Are there data on the phosphorylation of other known PP1 inhibitors, e.g. Inhibitor-1 or Inhibitor-2, in the MS experiments.

– Figure 4C: Do the three clusters analyzed here comprise all phosphorylated sites that have their maximum in MI? Which cluster would then contain most SP sites? Cluster 1 and 3 are depleted of Proline in the +1 position and Cluster 2 is depleted of phosphorylated Ser.

– Figure 4G: the dephosphorylation pattern for the SP and the TP sites are very similar here. It would be helpful to have a quantification of several biological replicates of this experiment to judge if there is a significant difference. Additionally, any difference might not be just to the Ser/Thr identity, because the antibodies also have a different sensitivity for adjacent amino acids.

– Does the expression of the B55 charge-swap mutants cause spontaneous GVBD? Importantly, with this mutant in hand, the authors could investigate by WB if known B55 substrates show altered dephosphorylation kinetics.

– Figure 6A, the authors want to make the point that phosphorylation of T61 sharply decreased following MI. S1158 shows an even sharper decline in its phosphorylation level, while other phosphorylated T sites seem not to be dephosphorylated, e.g. T506, indicating that the dephosphorylation code is more complex than being encoded in the nature of the phosphorylated residue. Why did the authors not investigate the phosphorylation state of T61 in their charge-swap experiment?

– In their analysis the authors talk a lot about what is happening during the transition from MI to MII, although their MS data just provide information about metaphase of MI and metaphase of MII, but not for the situation in between. In the data from previous publications (e.g. Okano-Uchida et al., 1998) it seems as if Cdk1 activity is very high in MI, then drops almost completely between MI and MII before rising again towards metaphase of MII (although not as high as in MI). So, in theory, it could be that there is much more B55-dependent dephosphorylation happening between MI and MII than was measured here, but a specific subset of sites (eg SP sites) got preferentially rephosphorylated for MII (e.g. defined by different Cdk activity thresholds). The INCENP data suggest that Thr sites are earlier dephosphorylated than Ser sites, but it might just be a matter of timing and not if a site is at all dephosphorylated or not between MI and MII as suggested here.

[Editors’ note: further revisions were suggested prior to acceptance, as described below.]

Thank you for resubmitting your work entitled "Selective dephosphorylation by PP2A-B55 directs the meiosis I – meiosis II transition in oocytes" for further consideration by eLife. Your revised article has been evaluated by Anna Akhmanova (Senior Editor) and a Reviewing Editor.

All three reviewers recommend acceptance of your paper but only after the inclusion of controls for the morpholine experiments (see Referee 3, point 3), and some extensive re-writing.

In your rewriting, the referees would like you to place your confirmatory data in proper context (eg: PMID: 29643273 and note referee 3, points 1 and 2), and mine your data further (referee 3 points 6 and 7). These results should then be discussed with reference to other systems (referee 3 point 8).

Reviewer #2:

As before, I remain supportive of the publication of the revised manuscript. Indeed, the authors carried out additional experiments and added data and explanations that together significantly strengthened their conclusions, and in my opinion these addressed most of the criticism raised by me and Reviewer 2.

Specifically, the added data for translationally regulated proteins is another important resource for the community. More importantly, the authors performed additional experiments to demonstrate the major role of PP2A-B55 in meiotic regulation, in contrast to PP1 having a minor contribution. Further, the authors provide additional Western blots using phospho-specific antibodies in support of their conclusion of differential TP/SP phosphorylation in MII.

Together, besides providing a very useful resource to the community, the authors present a conceptually new mode of meiotic regulation based on the preferential TP vs SP dephosphorylation by PP2A-B55 required for successful completion of the second meiotic division.

Reviewer #3:

This study focuses on the phosphoproteomics analysis during the two specialized meiotic divisions and beginning of embryogenesis in the sea star Pateria miniata. Dr. Swartz and collaborators have used a comprehensive proteomics/phosphoproteomics approach to investigate the regulatory circuits that drive the oocyte to embryo transition in this species. The first part of the study includes largely confirmatory observations of widely accepted concepts on regulation of the meiotic cell cycle. In the second part, the authors identify a divergent pattern of dephosphorylations consistent with an increased activity of the phosphatase PP2A-B55. Even though a PP2A role at this transition has been reported, the differential phosphorylation observed provides a novel angle that may provide the framework for a better understanding of the steady states stabilizing MI and MII in the oocytes.

The authors should be commended for the large amount of data they have generated and that certainly will be a useful resource for the meiosis/cell cycle community. The authors have responded to the previous review in a constructive way by including new data in support of their conclusions. Unfortunately and instead of using this large wealth of data to explore new features of meiosis or settle the numerous outstanding issues, they spend a good part of the study for largely confirmatory experiments. Nevertheless, the observation of the global differential dephosphorylation of TP and SP sites would be an interesting finding worth reporting (even though the PP2A- B55 preferential dephosphorylation of TP sites has already been described). Some missing controls would strengthen the authors' conclusions.

1. By using a proteomic approach the authors conclude that protein expression does not change substantially throughout the entire sea star cell cycle, which is already well established (see PMID: 29643273). However, it is also established that a number of critical proteins are synthesized during these transitions in the sea star. In addition to cyclin B and cyclin A and Pim-1 mentioned by the authors, Mos, cyclin E, Cdk2 and Wee1 accumulate during the two-cell cycle and pronuclear stages. The authors should report whether their proteomics approach confirms these previous findings. This would confirm the quality of the proteomics data collected. A similar question had already been posed during the previous review. In the discussion, it would be useful to comment that the stability of the proteome associated with minimal new protein synthesis is a peculiarity of the sea star, as large changes in protein synthesis drive the meiotic division in other animal models like the frog and mammals.

2. Page 5-12 of the manuscript report experiments that explore basic changes in phosphorylation and the involvement of phosphatases and essentially confirm widely established concepts for sea star meiosis and are consistent with basic views established for most species studied. All these properties have been widely described and well summarized, for instance in PMID: 29643273. In this comprehensive review, Figure 3 is used to summarize all these temporal changes in activity driving meiosis; it is hard to discern how this section of the manuscript adds any significant new information. This is puzzling because the data collected by the authors could have been mined to provide substantial and more relevant information and they have not used the data generated to resolve several contentious issues. For instance, the authors have the phosphorylation data for the prophase1 to GVBD transition to distinguish the state of phosphorylation of Myt 1 and Cdc25 at Cdk1 and SGK sites, not to mention the PI-3K dependent phosphorylation sites of SGK. There are reports suggesting differences, albeit subtle, in the consensus AKT/SGK sites. This analysis would have helped to consolidate the recent reports that SGK, and not Akt, functions distally to 1-methyladenine receptor. Moreover, the data could have been used to better tease apart at GVBD/MI what is usually called "initial activation" versus "auto-amplification" components required for switch/like maximal MPF activity. The authors could have reported and discussed the dual ARPP19 phosphorylations by Greatwall and CDK1.

3. Page 7. The authors use morpholino oligonucleotides to block the translation of cyclin A and B. However, they provide no control data on the extent and selectivity of the knockdown. This control is necessary to draw any conclusion.

4. Results, page 13. In the revised manuscript, the authors have included additional data with a more informative sampling rate between MI and MII to include points at anaphase. Using TP and SP specific antibodies the authors confirm that dephosphorylation at the TP site has faster kinetics than dephosphorylation at SP sites. The authors conclude that "despite the transient inactivation of Cdk1 between the meiotic division, SP phosphorylation remains relatively stable during this window". They use this finding to rule out the possibility of a kinase rephosphorylating at the SP sites. However, the authors do not discuss the possibility that the divergent MAPK activity, which remains constantly high, while Cdk1 activity is declining, at least in part contributes to divergent SP/TP phosphorylation state. This should be further discussed and possibly addressed experimentally.

5. To disrupt PP2A-B55 function, the authors use a dominant-negative approach with point mutations in B55 that disrupt interactions with substrates. The authors provide some data on the effect on meiotic progression. However, they do not investigate whether disruption of PP2A-B55 anchoring indeed blocks the divergent TP/SP dephosphorylations. This could be explored using phospho-specific antibodies.

6. The authors impute the divergent dephosphorylation of TP and SP sites to the selectivity of PP2A-B55. If this were correct, one should find a mirror image of dephosphorylation observed at the MI/MII transition during prophase to GVBD transition. PP2A is active in prophase 1, but it becomes inactivated at GVBD/MII. One should then see a preferential dephosphorylation of TP sites in prophase with little change in the SP sites. The authors should discuss this possibility and provide the data they have available; a hint of this is reported in Figure 6A.

7. INCEMP is a highly phosphorylated protein with a large number of sites to choose from. The authors report changes in the TP and SP sites, which clearly diverge. This is considered a major strength of the study, as it verifies that spatial sequestration of the protein is not an issue in the differential TP/SP dephosphorylation. The authors should include the SP phosphorylation patterns also for PRC1 and TPX2. Now only TP patterns are reported in Figure 5-S3. This would confirm that the TP/SP divergent dephosphorylation is of wide applicability.

8. In the discussion the authors make no attempt to compare and contrast their data with the wealth of information reported for other species. The authors do not specify the species used in the title. In the same vein, in the opening statement of the discussion the authors claim to have defined an extensive program of phosphorylation during the oocyte-to-embryo transition. However, they do not mention that they do so in the sea star. This has already been reported for other species. In addition, the authors do not discuss that the sea star provides a simplified experimental model to investigate regulation of meiosis and that many additional layers of regulation are present in frogs and mammals. Thus, the mechanisms of meiotic resumption vary considerably among species including sea star, frog and mouse. This important concept is not captured in the discussion.

Reviewer #4:

As outlined in my comments on the initial manuscript, I was not sure if the novelty of the findings suffice publications in eLife. For the revisions, the authors have done a great job to address all the points I raised. The manuscript significantly improved and I therefore recommend publication.

https://doi.org/10.7554/eLife.70588.sa1

Author response

[Editors’ note: the authors resubmitted a revised version of the paper for consideration. What follows is the authors’ response to the first round of review.]

The referees agreed on the quality of your data, and that this will be a useful resource for the community. The main concern was that a similar data set has already been published for Xenopus meiosis and that the conclusions from your study do not provide the insights that would warrant publication in a general audience publication such as eLife. If, however, you are able to address the substantive questions raised by Reviewer 2 – in particular, validating the role of NIPP1 in vivo – then we would be prepared to consider a further submission, but we judge that this will take more than two months and that the fairest decision is to return the paper to you now for you to consider your next step.

We thank the reviewers for their evaluation of our work. Importantly, we believe that our study represents a substantial advance from prior proteomic studies in oocytes. In particular, our work provides improved temporal resolution over these prior studies by unambiguously correlating cell cycle state to the phosphorylation landscape of the oocyte, as well as testing time points that span the entire oocyte-to-embryo transition. To accomplish this, we leveraged the unique biology of the sea star, which provides excellent synchrony and cell biological tractability. We have carefully cited these prior studies and provided a comparison to our work, but it is important to highlight that we were unable to do this clearly in many cases due to incomplete time courses, missing information, and ambiguity in cell cycle stages. Furthermore, Presler et al. PNAS 2017 only reported ~3500 phosphorylation events. Peuchen et al. Scientific Reports 2017 reported 8971 phosphorylation events identified across the two biological and technical replicates that they performed, but with only limited statistical analysis to assess the reproducibility of the data (the authors comment that “For the phosphoproteome experiments, biological duplicates and technical duplicates provided reasonable correlation (Pearson coefficient > 0.7)”). For our time course analyses, we identify a total of 25,227 phosphorylation sites across three biological replicates, of which 16,691 are identified in two or more replicates. The average Pearson coefficient of our dataset is 0.84 and 12,481 sites are reproducibly quantified (p-value 0.05 or less).

Thus, we believe that our proteomics analysis provides a definitive picture of the phosphorylation landscape across the entire oocyte-to-embryo transition. In addition to the value of our study as a resource for future work, our work suggests a new paradigm by which the MI / MII transition is regulated through differential serine vs. threonine conservation, and a critical role for the phosphatase PP2A-B55. To our knowledge, this has not been presented before and we believe this represents an important advance.

In our revised manuscript, we provide further support for the central role of PP2A-B55 as a conductor of meiotic prophase I arrest, and the transition between the two meiotic divisions. In response to the reviewer comments and discussion, we have undertaken a substantial effort to further test the role of PP1. Through proteomic and cell biological analysis, we find that PP1 plays a modest role relative to PP2A-B55 in the developmental transitions tested in our study. We believe these data substantially improve the manuscript by providing deeper insights into the relative contributions of these phosphatases.

Reviewer #1:

[…]

1. I find it very important to validate the quality and synchronicity of samples. Therefore, it would be important to state the number of oocytes analyzed for Figure S1B, and also more clearly state whether these samples have been taken from the same starfish animal and season.

Thank you for raising this important point. We have now included the number of oocytes analyzed in the legend for Figure 1—figure supplement 1B (“each time point represents at least 49 oocytes”). To better document the synchrony, we have added the following to the methods section: “Oocyte samples for proteomics were collected from a single female within the same season to maximize synchrony between time courses. Stereotypical meiotic timings were visually confirmed in live oocytes using germinal vesicle breakdown and polar body emission as metrics, before snap freezing samples in liquid nitrogen. The synchrony data in Figure S1B were performed with oocytes from a second animal collected at the same season and location that was confirmed to undergo meiosis at a similar rate.”

2. I greatly appreciate that the authors are describing the methods in detail and provide the raw proteomics data for download. Just one more detail: could they provide the MS1 and MS2 errors?

We are pleased that the methods section and raw data were helpful. We have now included the additional requested details. We have adjusted the methods section to include additional information on the data filtering:

“High resolution tandem mass spectra were searched using COMET with a precursor ion tolerance of +/- 1 Dalton and a fragment ion tolerance of +/- 0.02 m/z, static mass of 229.162932 on peptide N-termini and lysines and 57.02146 Da on cysteines, and a variable mass of 15.99491 Da on methionines against a target-decoy version of the Patiria miniata proteome sequence database. For phosphoproteomics analysis 79.96633 Da on serines, threonines and tyrosines was included as an additional variable mass. The resulting peptide spectral matches (PSMs) were filtered to a < 1% false discovery rate (FDR) using the target decoy strategy with a typical precursor ion mass filter of +/- 1.5 parts-per-million (PPM) mass accuracy and corresponding XCorr and dCn values. Quantification of LC-MS/MS spectra was performed using in house developed software. Probability of phosphorylation site localization was determined by PhosphoRS (Taus et al., 2011).”

3. I would be curious to see the localization of INCENP also in MII and AnaII. Could the authors provide these additional immunofluorescence data?

Thank you for this suggestion. Based on this comment and related suggestions from Reviewer #2, we have now conducted multiple additional experiments to analyze the role of PP2A-B55 and differential dephosphorylation in controlling INCENP localization. First, we have analyzed the localization for wild type INCENP across Meiosis I and Meiosis II. The localization behavior for INCENP in Meiosis II is now included in Figure 5—figure supplement 3D. In contrast to Meiosis I, we detect reduced INCENP localization to centromeres in Meiosis II, but INCENP additionally localizes to the Anaphase II midzone, and nucleolar structures in the female pronucleus. Second, we tested INCENP localization in oocytes expressing the inhibitory B55 mutants that alter its substrate specificity. Under these conditions, INCENP localization to metaphase centromeres is increased, and the translocation to the midzone in anaphase is strongly reduced. These data are now included in Figure 6. This localization is strongly consistent with our model that the selective dephosphorylation of INCENP T61 in Meiosis II by B55 governs the localization of INCENP to the centromere versus the midzone.

Reviewer #2:

Swartz et al. applies MS analyses to investigate protein abundance and modification during maturation of Patiria miniata oocytes. MS analyses revealed that 99% of the app. 5000 proteins reproducibly identified displayed a max fold change of less than a 1.2-fold change indicating that meiotic cell cycle transitions are not associated with bulk changes in protein levels. Using Emetine, the authors show that progression to MII requires protein synthesis, in particular synthesis of cyclin A and B. Phospho-proteome analyses revealed that the majority of phosphorylation events occur at serine residues, followed by threonine and tyrosine residues. Prophase-I arrested oocytes displayed the lowest overall phosphorylation state. Consistent with what is known from other organisms, inhibitory phosphorylations of Cdk were found in Prophase-I arrested oocytes, but not during oocyte maturation. Then, the authors focus on PP1 and PP2A. Calyculin A treatment induced spontaneous GVBD. However, treated oocytes did not progress beyond a GVBD-like state. Cluster analyses identified TP sites as being more likely to be efficiently dephosphorylated upon exit from MI compared to SP sites. Consistent with studies in other organisms, activities of Cdk and B55 are inversely regulated such that either Cdk or B55 are active. Expression of B55 charge-swap mutations induce severe phenotypes at the MI/MII transition. T to S replacement in INCEP prevented the relocalization of the GFP-tagged fusion protein from centromeres to the spindle midzone at anaphase I.

The decision on this manuscript is not an easy one. On the one hand, the authors have done a lot of experiments and the collected data are of high quality. On the other hand, the provided insight are not particular novel or exciting and often the authors leave interesting insights without further following up. It does not come to a surprise that meiotic progression in Patiria miniata is not mediated by bulk protein turnover, but by few changes in key cell cycle regulators such as cyclins (Figures 1 and 2). Likewise, it is not surprising that phosphorylation plays a major role in meiotic cell cycle progression and that changes in Cdk and PP2A-B55 activity mediate those (Figure 3).

We appreciate this feedback on the quality of our data and the questions of novelty. We agree that it comes as no surprise that phosphorylation is an important driver of cell cycle progression building on prior work across multiple organisms. However, the differential dephosphorylation of threonine vs. serine as a paradigm to drive the MI / MII transition has not been proposed to our knowledge. Our work provides answers to critical questions including how oocytes complete Meiosis I, and then proceed directly in to Meiosis II without undergoing a gap phase or DNA replication, and how Meiosis I-specific functions are reversed while maintaining an overall meiotic state. The differential behavior of TP versus SP phosphorylation, mediated by PP2A-B55, provides an important, new conceptual framework to understand this process – findings that establish this paradigm and create new directions for future work in this area. In addition, we believe our study provides a uniquely holistic view of the oocyte-to-embryo transition, enabled by the powerful experimental advantages of the sea star. Together, we feel that our work provide a substantial conceptual advance, as well as a valuable resource that will be broadly useful for researchers interested in exploring this specialized cell cycle program.

– The authors show by Emetine treatment that protein synthesis is required for meiotic progression, but apart from cyclin B the authors don´t provide a correlation to their MS data on changes in protein abundance. Are there other interesting (cell cycle-related) candidates among the strongly upregulated proteins in the MS data that could be required for meiotic progression?

Thank you for this helpful point. We have now addressed this in two ways. First, we now specifically highlight additional proteins whose steady-state levels fluctuate in our time-course proteomics. This includes proteins such as the kinase PIM-1 and Importin 8, which are translated de novo after meiotic resumption. Second, we have now conducted additional mass spectrometry experiments in which we compare control Meiosis II oocytes to those in which new translation was blocked with emetine. With this approach, we identified 108 proteins whose levels are sensitive to translational inhibition. These factors include PIM-1, as well as the DNA replication factor Cdt1, a potential Cdk regulatory subunit, and other proteins that are now discussed in the Results and Discussion sections. We include these data as an additional supplemental figure (Figure 2—figure supplement 1) and excel sheet (Supplementary file 3). We believe that this dataset provides a valuable resource and additional useful insights into the proteins for which nascent translation is necessary for meiotic progression.

– The authors provide many indirect, but no direct evidence that PP1 activity is indeed high in Prophase I. They investigate phospho-regulation of NIPP1 in vitro without translating these insights to the oocyte. How meaningful / physiological relevant are these in vitro insights for the Prophase I arrest. Additionally, the analysis is focused on a single PP1 inhibitor, NIPP1, but PP1 activity could be regulated by several other inhibitors as well. Are there data on the phosphorylation of other known PP1 inhibitors, e.g. Inhibitor-1 or Inhibitor-2, in the MS experiments.

Thank you for this interesting question regarding the physiological contribution of PP1 and its inhibitory partners. This comment prompted us to undertake a substantial investigation of PP1 behavior by mass spectrometry, Western blotting, and in vivo experiments in oocytes. First, in addition to our previous analysis of NIPP1, we identified a sea star ortholog of the PP1 regulatory protein Inhibitor-2 (I2). We find that I2 also becomes phosphorylated on several sites following hormonal stimulation and meiotic resumption. These phosphorylation sites could regulate the inhibitory activity of I2 (Author response image 1). However, we generated both wild-type and phosphomimetic mutations for these sites, and found no dominant effect when the constructs were expressed in oocytes.

Author response image 1
Download asset Open asset
IPP2 phosphorylation sites.

Second, we analyzed our phosphoproteomic data for a well-characterized PP1 binding motif, RVxF (in which “x” is S or T). Phosphorylation of these motifs by Aurora and other kinases negatively regulates the recruitment of PP1 to its regulatory partners and substrates. We find that RVxF phosphorylation is low in Prophase I, but increases in meiosis, consistent with a corresponding downregulation of PP1 activity. We further tested this trend by Western blotting with antibodies specific to the RVp[S/T]F motif, which revealed consistent results. We have included these data in Figure 3—figure supplement 2C, and figure supplement 3.

Having defined these temporal phosphorylation behaviors indicative of PP1 activity across meiosis, we next generated multiple constructs to evaluate the in vivo contributions of PP1 dephosphorylation in oocytes. Under the same conditions in which we tested the PP2A-B55 mutant constructs and the inhibitor Calyculin A, we evaluated the effects of expressing the following constructs in oocytes: 1) Wild-type NIPP1, 2) the NIPP1 S197,199A mutant, 3) wildtype I2, 4) a phosphomimetic I2 mutant, and 5) direct injection of RVSF peptides to compete off PP1 from endogenous substrates. Notably, none of these perturbations had a substantial effect on the maintenance of the Prophase I arrest, meiotic resumption, or the successful extrusion of polar bodies.

Finally, to directly test the role of PP1, we used the inhibitor Tautomycetin, which is highly selective for PP1 vs. PP2A. Using in vitro assays, we first confirmed that Tautomycetin is indeed a highly potent and selective inhibitor of PP1. Next, we tested whether Tautomycetin could induce spontaneous germinal vesicle breakdown, using the same assay as we previously conducted with Calyculin A. Strikingly, we found no effect even after a 24 hour treatment with this inhibitor under identical conditions as those used for Calyculin A. Finally, we found that Tautomycetin-treated oocytes progressed through meiosis with normal kinetics including undergoing polar body extrusion. However, Tautomycetin treatment did result in significant chromosome alignment and mis-segregation defects, consistent with an inability to correct kinetochore-microtubule attachment errors. These data have been added to Figure 3—figure supplement 6.

In summary, based on the reviewer’s valuable suggestion, we have now monitored and perturbed the activity of PP1 and its inhibitory partners using multiple orthogonal approaches. This analysis allows us to now conclude that PP1 plays a comparatively modest role in maintaining a Prophase I arrest and driving meiotic progression relative to PP2A. We feel that this data provides a substantial conceptual advance for the paper, and further emphasizes the central function of PP2A-B55 in orchestrating meiosis.

– Figure 4C: Do the three clusters analyzed here comprise all phosphorylated sites that have their maximum in MI? Which cluster would then contain most SP sites? Cluster 1 and 3 are depleted of Proline in the +1 position and Cluster 2 is depleted of phosphorylated Ser.

For the motif analysis in Figure 4C, we only included singly phosphorylated peptides. In addition to commenting on this in the Methods section, we have now included this information in the Results section. We made this choice because, for peptides with more than one phosphorylation site, we cannot discern which sites are responsible for the changing abundance. However, in the previous version of Figure 4A, all sites peaking in MI were clustered. For the revised paper, we now repeated the clustering analysis using only peptides with single phosphorylation sites and have adjusted Figures 4A and B accordingly. The conclusions for these new analyses are highly comparable with our previous findings.

For the analysis in Figure 4C evaluating the over-and under-representation of specific residues, each position is investigated independently and compared against the amino acid composition of all phosphorylation sites at this position. Although there are more SP than TP sites in Clusters 1 and 2, the number of proline-directed sites in these clusters is not overrepresented (Author response image 2). However, in Cluster 3, nearly all threonine phosphorylation sites are proline-directed.

Author response image 2
Download asset Open asset
Distribution of S and T residues, and SP and TP motifs in Cluster 1 3.

– Figure 4G: the dephosphorylation pattern for the SP and the TP sites are very similar here. It would be helpful to have a quantification of several biological replicates of this experiment to judge if there is a significant difference. Additionally, any difference might not be just to the Ser/Thr identity, because the antibodies also have a different sensitivity for adjacent amino acids.

Thank you for this suggestion. To increase the confidence in our time course Western blot experiments, we have now performed multiple independent experiments and present the results as the mean and standard deviation of three biological replicates. These new data are included in Figure 4G and Figure 4—figure supplement 1B, and emphasize the reproducibility of the differences we discovered in SP vs. TP behavior.

In addition, we have now tested antibodies that recognize a different phosphorylation consensus as an additional orthogonal metric. We obtained commercially-available antibodies that recognize pTPP and pSPP, and found similar temporal phosphorylation profiles as those generated with the previous pTPxK and [K/H]pSP antibodies. These data are now included in Figure 4—figure supplement 2.

– Does the expression of the B55 charge-swap mutants cause spontaneous GVBD? Importantly, with this mutant in hand, the authors could investigate by WB if known B55 substrates show altered dephosphorylation kinetics.

Thank you for pointing out this interesting possibility. We have now tested the effect of the B55 DE/A mutant on maintaining the Prophase I arrest. Similar to Calyculin A treatment, expression of this B55 mutant induces spontaneous GVBD, and results in the subsequent apoptosis of the oocytes within 3 days of culture. These data are now included in Figure 5— figure supplement 2. This indicates that PP2A-B55 perturbation alone is sufficient to phenocopy Calyculin A treatment, and supports a model in which PP2A is the primary phosphatase that acts to maintain the Prophase I arrest. We believe this is a valuable conceptual advance for the regulation of oocyte arrest.

We agree that it would be very valuable to directly test the dephosphorylation kinetics of known B55 substrates by Western blot. Unfortunately, we are technically unable to do this experiment as we currently lack antibodies that recognize specific B55 substrates in the sea star. As an alternative approach, we now report the cellular behavior of INCENP-GFP, a PP2A-B55 substrate, following expression of either wild-type or mutant B55. We find that altering the substrate specificity of PP2A with this B55 mutant substantially reduces the ability of wild-type INCENP to translocate to the central spindle in anaphase of Meiosis I. In addition, this perturbation increases the localization of INCENP to centromeres in Metaphase. These localization defects are consistent with a failure in the timely dephosphorylation of INCENP T61, based on our analysis of this mutant construct (Figure 6D,E). These data have now been added to Figure 6.

– Figure 6A, the authors want to make the point that phosphorylation of T61 sharply decreased following MI. S1158 shows an even sharper decline in its phosphorylation level, while other phosphorylated T sites seem not to be dephosphorylated, e.g. T506, indicating that the dephosphorylation code is more complex than being encoded in the nature of the phosphorylated residue. Why did the authors not investigate the phosphorylation state of T61 in their charge-swap experiment?

Thank you for pointing this out. In our discussion of the results, we did not fully capture the nuance of the phosphorylation code. It is not simply the threonine vs. serine that drives these kinetics, but likely also depends on the identity of the adjacent amino acids. In the case of T506, the presence of acidic amino acids (E508 and D510) are disfavored for B55 association, as supported by our motif analysis (Figure 4C-E). We agree that the behavior of S1158 is unique, and the reason for this difference are less obvious, but we suggest that perhaps this residue’s position at the extreme C-terminus of the protein may influence its accessibility to B55.

To better describe and explain these behaviors, we have added several lines to the Results and Discussion section to convey that the dephosphorylation code is more complex than the single phosphorylated residue. (Results: “In contrast, T506 remains relatively stable, likely due to the presence of downstream acidic amino acids, thereby disfavoring PP2A-B55 association (Figure 6A).”, Discussion: “This phosphorylation code is further modulated by the charge of adjacent amino acids, with positively charged residues favoring PP2A-B55 association”).

Unfortunately, it is not technically possible at this time to directly evaluate the phosphorylation state of INCENP T61 following the expression of B55 mutants. We lack phospho-specific antibodies recognizing this residue, and testing this by mass spectrometry would require the injection of a prohibitive number of oocytes to achieve sufficient quantities for this analysis. As an alternative approach to this, we now report the localization behavior of INCENP-GFP following expression of mutant B55. As our model predicts, INCENP translocation to the central spindle in anaphase is strongly reduced when B55 activity is disrupted, whereas the localization to centromeres is increased. These results are consistent with the behavior governed by delayed dephosphorylation of T61.

– In their analysis the authors talk a lot about what is happening during the transition from MI to MII, although their MS data just provide information about metaphase of MI and metaphase of MII, but not for the situation in between. In the data from previous publications (e.g. Okano-Uchida et al., 1998) it seems as if Cdk1 activity is very high in MI, then drops almost completely between MI and MII before rising again towards metaphase of MII (although not as high as in MI). So, in theory, it could be that there is much more B55-dependent dephosphorylation happening between MI and MII than was measured here, but a specific subset of sites (eg SP sites) got preferentially rephosphorylated for MII (e.g. defined by different Cdk activity thresholds). The INCENP data suggest that Thr sites are earlier dephosphorylated than Ser sites, but it might just be a matter of timing and not if a site is at all dephosphorylated or not between MI and MII as suggested here.

Thank you for suggesting this interesting model. We agree that rephosphorylation of selected residues is an alternative possibility that could explain the phosphorylation kinetics that we observe. To test this, we have conducted several additional experiments. First, we have now conducted a higher temporal resolution analysis of phosphorylation behaviors across the oocyte-to-embryo transition using Western blotting with multiple pTP and pSP CDK consensus antibodies, in biological triplicate. We find that serine phosphorylation is substantially more stable between MI and MII than threonine phosphorylation (Figure 4G, Figure 4—figure supplement 1B). The increased sampling at multiple time points during the MI/MII transitions allows us to more confidently define these differences, and importantly, indicates that there is not an apparent decrease in phosphorylation followed by rephosphorylation. As the reviewer points out, this is the window of meiosis during which cyclin B is fully degraded and is prior to Cdk1 reactivation. We therefore believe that this would disfavor a model in which selective rephosphorylation of serine-proline by Cdk explains this difference in behavior. We now discuss these alternative possibilities with the following lines in the Results section: “However, an alternative interpretation is that both SP and TP sites are dephosphorylated equally at the MI/MII transition, but then SP sites are selectively and rapidly re-phosphorylated in MII. To distinguish between these models, we performed Western blots using antibodies against phosphorylated pTP and pSP CDK motifs on samples collected with increased temporal resolution…”

Finally, prior work suggests that Cdk1 preferentially phosphorylates threonine residues (Miller et al., Science Signaling, 35, 2008), further arguing against this model. This is further supported by the S/T distribution of 1576 Cdk1 substrate phosphorylation sites listed on Phosphosite.org. Both sources indicated that ~30% of Cdk1 phosphorylation sites are threonine, which is substantially higher than the ~15.5% phosphothreonine that would be expected for a non-S/T selective kinase based on the global occurrence of phosphorylated threonine residues (Sharma et al., Cell Reports, 8, 2014).

[Editors’ note: what follows is the authors’ response to the second round of review.]

Reviewer #3:

This study focuses on the phosphoproteomics analysis during the two specialized meiotic divisions and beginning of embryogenesis in the sea star Pateria miniata. Dr. Swartz and collaborators have used a comprehensive proteomics/phosphoproteomics approach to investigate the regulatory circuits that drive the oocyte to embryo transition in this species. The first part of the study includes largely confirmatory observations of widely accepted concepts on regulation of the meiotic cell cycle. In the second part, the authors identify a divergent pattern of dephosphorylations consistent with an increased activity of the phosphatase PP2A-B55. Even though a PP2A role at this transition has been reported, the differential phosphorylation observed provides a novel angle that may provide the framework for a better understanding of the steady states stabilizing MI and MII in the oocytes.

The authors should be commended for the large amount of data they have generated and that certainly will be a useful resource for the meiosis/cell cycle community. The authors have responded to the previous review in a constructive way by including new data in support of their conclusions. Unfortunately and instead of using this large wealth of data to explore new features of meiosis or settle the numerous outstanding issues, they spend a good part of the study for largely confirmatory experiments. Nevertheless, the observation of the global differential dephosphorylation of TP and SP sites would be an interesting finding worth reporting (even though the PP2A- B55 preferential dephosphorylation of TP sites has already been described). Some missing controls would strengthen the authors' conclusions.

1. By using a proteomic approach the authors conclude that protein expression does not change substantially throughout the entire sea star cell cycle, which is already well established (see PMID: 29643273). However, it is also established that a number of critical proteins are synthesized during these transitions in the sea star. In addition to cyclin B and cyclin A and Pim-1 mentioned by the authors, Mos, cyclin E, Cdk2 and Wee1 accumulate during the two-cell cycle and pronuclear stages. The authors should report whether their proteomics approach confirms these previous findings. This would confirm the quality of the proteomics data collected. A similar question had already been posed during the previous review.

We have further highlighted in the text that our observations on protein stability are consistent with previous observations of the limited turnover of specific proteins. Because of the global scale of our analysis, we can now extend the previous observations for specific proteins to a more generalizable paradigm.

A challenge in proteomic versus genomic analyses is the dynamic range problem, particularly when combining proteomics with quantitative approaches such as tandem mass tagging. Although modern mass spectrometry instrumentation has come far, it is still not at the point where all expressed proteins are easily identified and quantified to the same extent regardless of their expression level. Although our analysis is, to the best of our knowledge, one of the most comprehensive investigations into a meiotic proteome to date, and the only one that has been conducted in the sea star, a subset of low expression level proteins are absent from this data. The abundance of many cell cycle regulators, including Mos, Cyclin E, Cdk2, and Wee1, is low and tightly controlled, and therefore often not accessible yet by proteomics approaches without additional enrichment strategies. Although we do detect phosphorylation sites on these four proteins, we were only able to identify and quantify Wee1 on the proteome level in 2 of the 3 biological replicates with a small number of peptides (3 in replicate 2, and 1 in replicate 3). The detection of phosphorylation sites is likely due to the 100-fold enrichment upon phospho-peptide purification. Although this limits our ability to make statements about these proteins, it is not a data quality problem, but rather a depth of analysis problem typical of all standard mass spectrometry approaches.

In the discussion, it would be useful to comment that the stability of the proteome associated with minimal new protein synthesis is a peculiarity of the sea star, as large changes in protein synthesis drive the meiotic division in other animal models like the frog and mammals.

We appreciate the comment, but disagree with this point. A previous analysis by Presler et al. (PNAS 2017, Figure 1 C and D) of Xenopus oocytes found that only 48 out of 8,641 proteins were dynamic, whereas the majority of proteins remain constant during a time course of 0 – 20 minutes after fertilization and during progression from metaphase II to anaphase II. Similarly, Peuchen et al. (Scientific Reports 2017, Figure 1B) report that there is “not a substantial change in the Xenopus proteome” in a time course of oocyte maturation to first zygotic cleavage. They find that only 486 of 6428 proteins change in expression by more than 5%. Based on these observations, we believe that, at least in frogs, there are also not significant changes in protein levels on a global scale. Although some selected critical regulatory proteins do change in their protein levels or undergo new synthesis (as described in this paper), thereby driving meiotic progression, the majority of proteins remain unchanged during this developmental window. It is possible that this reviewer is referring to the maternal to zygotic transition, during which time there are likely to be more substantial changes.

2. Page 5-12 of the manuscript report experiments that explore basic changes in phosphorylation and the involvement of phosphatases and essentially confirm widely established concepts for sea star meiosis and are consistent with basic views established for most species studied. All these properties have been widely described and well summarized, for instance in PMID: 29643273. In this comprehensive review, Figure 3 is used to summarize all these temporal changes in activity driving meiosis; it is hard to discern how this section of the manuscript adds any significant new information. This is puzzling because the data collected by the authors could have been mined to provide substantial and more relevant information and they have not used the data generated to resolve several contentious issues. For instance, the authors have the phosphorylation data for the prophase1 to GVBD transition to distinguish the state of phosphorylation of Myt 1 and Cdc25 at Cdk1 and SGK sites, not to mention the PI-3K dependent phosphorylation sites of SGK. There are reports suggesting differences, albeit subtle, in the consensus AKT/SGK sites. This analysis would have helped to consolidate the recent reports that SGK, and not Akt, functions distally to 1-methyladenine receptor.

We thank the reviewer for their helpful comments. As suggested, we have now further mined our data. Although we do not detect the activation loop phosphorylation site (T312) in SGK in our time course dataset, we do observe phosphorylation of the neighboring TP (T316) site, which is part of the P+1 loop and activation segment.

Concomitantly, we observe an increase in the phosphorylation of Cdc25 S188, an SGK site (Hiraoka et al. JCB, 2019), which serves to activate Cdc25. Furthermore, we observe an increase in phosphorylation of a double and triple phosphopeptide of Myt1, which includes S75, an SGK site (Hiraoka et al. JCB, 2019), which functions to inactivates Myt1. Activation of Cdc25 and inactivation of Myt1 trigger activation of Cdk1-Cyclin B (Hiraoka et al. JCB, 2019).

Based on an analysis of the known human substrates for SGK1 and Akt1 (Phosphosite.org), we propose that the most dramatic difference in the SGK1 and Akt1 consensus motifs is a preference in basophilic residues in position -4 for Akt1 (see Author response image 3).

Author response image 3
Download asset Open asset

Based on the reviewer’s suggestion, we have now investigated the presence of RxRxxS/T consensus motifs in the dataset for the localized, single, and reproducibly quantified phosphopeptides that increased in phosphorylation abundance by 3-fold or more from Prophase 1 to GVBD versus sites with less than a 3fold increase over the same time period.This analysis revealed a stronger preference for basic amino acids in the -4 position, indicative of Akt phosphorylation for phosphorylation sites with a minimal increase in abundance in the Prophase 1 to GVBD position. Although we find this data intriguing and suggestive of a model where there is stronger SGK activation than Akt, we do not feel that this point represents a major conclusion that merits specific highlighting in the manuscript. We hope that this information will be useful to readers who are reading this as part of the transparent review process, and also believe that this highlights the value of this broad dataset for diverse researchers interested in phospho-regulation downstream of diverse kinases.

Moreover, the data could have been used to better tease apart at GVBD/MI what is usually called "initial activation" versus "auto-amplification" components required for switch/like maximal MPF activity. The authors could have reported and discussed the dual ARPP19 phosphorylations by Greatwall and CDK1.

This is a great point. We have now included an additional discussion of the S69 Cdk1 and S165 PKA sites on ARPP19 in manuscript (see Figure 3-S5). Briefly, we find that the phosphorylation sites S106 (Greatwall) and S165 (PKA) display similar temporal behavior, whereas S69 (Cdk1) phosphorylation increase at the Prophase 1 to GVBD transition, remain high during MI to 2-PN, and further increase during first cleavage.

3. Page 7. The authors use morpholino oligonucleotides to block the translation of cyclin A and B. However, they provide no control data on the extent and selectivity of the knockdown. This control is necessary to draw any conclusion.

Thank you for this suggestion. We have now included a new experiment with Western blotting using Cyclin B antibodies (a gift from the Kishimoto lab) for first cleavage embryos following injection of either control, Cyclin A, or Cyclin B morpholinos. We find that Cyclin B protein is robustly and specifically depleted following Cyclin B morpholino injection. Unfortunately, we lack antibodies against starfish Cyclin A. In addition, as these are start site-targeting morpholinos, they have no measurable effect on mRNA levels that could be assessed by RT-PCR. Therefore, to increase confidence in the Cyclin A depletion phenotype, we have also included a Western blot for TPxK phosphorylation. As predicted, individual depletion of either Cyclin A or Cyclin B results in a decrease in phosphorylation levels relative to controls. We believe that these added experiments, combined with the robust and specific phenotypes that are consistent with prior expectations for the roles of these cyclins during meiotic progression, provide high confidence in our morpholino results. These results have been added to Figure 2—Figure Supplement 2.

4. Results, page 13. In the revised manuscript, the authors have included additional data with a more informative sampling rate between MI and MII to include points at anaphase. Using TP and SP specific antibodies the authors confirm that dephosphorylation at the TP site has faster kinetics than dephosphorylation at SP sites. The authors conclude that "despite the transient inactivation of Cdk1 between the meiotic division, SP phosphorylation remains relatively stable during this window". They use this finding to rule out the possibility of a kinase rephosphorylating at the SP sites. However, the authors do not discuss the possibility that the divergent MAPK activity, which remains constantly high, while Cdk1 activity is declining, at least in part contributes to divergent SP/TP phosphorylation state. This should be further discussed and possibly addressed experimentally.

We agree with this point. We have now added a statement to the Discussion section stating that MAPK activity remains high until the MII to 2PN transition and could be responsible for continuous phosphorylation of proline-directed serine and threonine phosphorylation. However, even if this is the case, the presence of a phosphatase activity with preferential threonine dephosphorylation activity is still needed to explain our results for differential dephosphorylation, since MAPKs do phosphorylate both serine and threonine residues.

5. To disrupt PP2A-B55 function, the authors use a dominant-negative approach with point mutations in B55 that disrupt interactions with substrates. The authors provide some data on the effect on meiotic progression. However, they do not investigate whether disruption of PP2A-B55 anchoring indeed blocks the divergent TP/SP dephosphorylations. This could be explored using phospho-specific antibodies.

6. The authors impute the divergent dephosphorylation of TP and SP sites to the selectivity of PP2A-B55. If this were correct, one should find a mirror image of dephosphorylation observed at the MI/MII transition during prophase to GVBD transition. PP2A is active in prophase 1, but it becomes inactivated at GVBD/MII. One should then see a preferential dephosphorylation of TP sites in prophase with little change in the SP sites. The authors should discuss this possibility and provide the data they have available; a hint of this is reported in Figure 6A.

This is a fantastic suggestion. For the revised manuscript, we have compared the abundance of single, localized, and reproducibly-quantified phosphopeptides in the dataset for TP and SP consensus motifs. We found that the average abundance of SP phosphorylation sites in Prophase I was 0.3 compared to 0.14 for TP sites (p-value <0.0001). This indicates that TP sites are more readily opposed in Prophase I. Intriguingly, this trend is reversed in GVBD, where the average abundance of SP phosphorylation sites is Prophase I was 0.66 compared to 0.74 for TP sites (p-value <0.0001). We have added the data showing the relative SP vs. TP phosphorylation abundance in Prophase I to Figure 4-S1.

7. INCEMP is a highly phosphorylated protein with a large number of sites to choose from. The authors report changes in the TP and SP sites, which clearly diverge. This is considered a major strength of the study, as it verifies that spatial sequestration of the protein is not an issue in the differential TP/SP dephosphorylation. The authors should include the SP phosphorylation patterns also for PRC1 and TPX2. Now only TP patterns are reported in Figure 5-S3. This would confirm that the TP/SP divergent dephosphorylation is of wide applicability.

We appreciate this helpful suggestion. We have now added this data to Figure 5-S3.

8. In the discussion the authors make no attempt to compare and contrast their data with the wealth of information reported for other species. The authors do not specify the species used in the title. In the same vein, in the opening statement of the discussion the authors claim to have defined an extensive program of phosphorylation during the oocyte-to-embryo transition. However, they do not mention that they do so in the sea star. This has already been reported for other species. In addition, the authors do not discuss that the sea star provides a simplified experimental model to investigate regulation of meiosis and that many additional layers of regulation are present in frogs and mammals. Thus, the mechanisms of meiotic resumption vary considerably among species including sea star, frog and mouse. This important concept is not captured in the discussion.

We now added a more detailed comparison of our findings to those in other organisms to the discussion.

https://doi.org/10.7554/eLife.70588.sa2

Article and author information

Author details

  1. S Zachary Swartz

    1. Whitehead Institute for Biomedical Research, Cambridge, United States
    2. Department of Biology, Massachusetts Institute of Technology, Cambridge, United States
    Contribution
    Conceptualization, Data curation, Formal analysis, Funding acquisition, Validation, Investigation, Writing - original draft, Writing - review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-3264-6880

  2. Hieu T Nguyen

    Department of Biochemistry and Cell Biology, Geisel School of Medicine at Dartmouth, Hanover, United States
    Contribution
    Data curation, Formal analysis, Investigation, Writing - review and editing
    Competing interests
    No competing interests declared

  3. Brennan C McEwan

    Department of Biochemistry and Cell Biology, Geisel School of Medicine at Dartmouth, Hanover, United States
    Contribution
    Data curation, Formal analysis, Investigation
    Competing interests
    No competing interests declared

  4. Mark E Adamo

    Norris Cotton Cancer Center, Geisel School of Medicine at Dartmouth, Lebanon, United States
    Contribution
    Software, Formal analysis
    Competing interests
    No competing interests declared

  5. Iain M Cheeseman

    1. Whitehead Institute for Biomedical Research, Cambridge, United States
    2. Department of Biology, Massachusetts Institute of Technology, Cambridge, United States
    Contribution
    Conceptualization, Data curation, Supervision, Funding acquisition, Project administration, Writing - review and editing
    For correspondence
    icheese@wi.mit.edu
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-3829-5612

  6. Arminja N Kettenbach

    1. Department of Biochemistry and Cell Biology, Geisel School of Medicine at Dartmouth, Hanover, United States
    2. Norris Cotton Cancer Center, Geisel School of Medicine at Dartmouth, Lebanon, United States
    Contribution
    Conceptualization, Data curation, Supervision, Funding acquisition, Project administration, Writing - review and editing
    For correspondence
    Arminja.N.Kettenbach@dartmouth.edu
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-3979-4576

Funding

National Institute of General Medical Sciences (R35GM126930)

  • Iain M Cheeseman

National Institute of General Medical Sciences (R35GM119455)

  • Arminja N Kettenbach

Eunice Kennedy Shriver National Institute of Child Health and Human Development (K99HD099315)

  • S Zachary Swartz

Gordon and Betty Moore Foundation

  • Iain M Cheeseman

Buck Institute (GCRLE-1220)

  • Iain M Cheeseman

National Institute of Child Health and Human Development (5K99HD099315)

  • S Zachary Swartz

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

SZS was supported by a K99 fellowship from NIH/NICHD (5K99HD099315). This work was supported by grants from the NIH/National Institute of General Medical Sciences to IMC (R35GM126930) and ANK (R35GM119455), and grants to IMC from the Gordon and Betty Moore Foundation, and a Pilot award from the Global Consortium for Reproductive Longevity and Equity (GCRLE-1220). The Orbitrap Fusion Tribrid mass spectrometer was acquired with support from NIH (S10-OD016212). Molecular graphics and analyses were performed with UCSF ChimeraX, developed by the Resource for Biocomputing, Visualization, and Informatics at the University of California, San Francisco, with support from National Institutes of Health R01-GM129325 and the Office of Cyber Infrastructure and Computational Biology, National Institute of Allergy and Infectious Diseases. We thank Brooke Brauer for reading the manuscript. We thank Dr. Wolfgang Peti for purified PP1 and Dr. Richard Honkanen for tautomycetin. The RVp[S/T]F antibody was a gift from Dr. Greg Moorhead. We thank Takeo Kishimoto and Ei-Ichi Okumura for valuable comments on the manuscript and for sharing antibodies.

Senior Editor

  1. Anna Akhmanova, Utrecht University, Netherlands

Reviewing Editor

  1. Jon Pines, Institute of Cancer Research Research, United Kingdom

Reviewer

  1. Peter Lenart, Max Planck Institute for Biophysical Chemistry, Germany

Publication history

  1. Preprint posted: August 21, 2020 (view preprint)
  2. Received: May 23, 2021
  3. Accepted: August 2, 2021
  4. Accepted Manuscript published: August 3, 2021 (version 1)
  5. Version of Record published: August 17, 2021 (version 2)

Copyright

© 2021, Swartz et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 1,447
    Page views
  • 244
    Downloads
  • 4
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. S Zachary Swartz
  2. Hieu T Nguyen
  3. Brennan C McEwan
  4. Mark E Adamo
  5. Iain M Cheeseman
  6. Arminja N Kettenbach
(2021)
Selective dephosphorylation by PP2A-B55 directs the meiosis I-meiosis II transition in oocytes
eLife 10:e70588.
https://doi.org/10.7554/eLife.70588

Categories and tags

Further reading

    1. Cell Biology

    Global analysis of contact-dependent human-to-mouse intercellular mRNA and lncRNA transfer in cell culture

    Sandipan Dasgupta, Daniella Y Dayagi ... Jeffrey E Gerst
    Research Article Updated

    Full-length mRNAs transfer between adjacent mammalian cells via direct cell-to-cell connections called tunneling nanotubes (TNTs). However, the extent of mRNA transfer at the transcriptome-wide level (the ‘transferome’) is unknown. Here, we analyzed the transferome in an in vitro human-mouse cell co-culture model using RNA-sequencing. We found that mRNA transfer is non-selective, prevalent across the human transcriptome, and that the amount of transfer to mouse embryonic fibroblasts (MEFs) strongly correlates with the endogenous level of gene expression in donor human breast cancer cells. Typically,<1% of endogenous mRNAs undergo transfer. Non-selective, expression-dependent RNA transfer was further validated using synthetic reporters. RNA transfer appears contact-dependent via TNTs, as exemplified for several mRNAs. Notably, significant differential changes in the native MEF transcriptome were observed in response to co-culture, including the upregulation of multiple cancer and cancer-associated fibroblast-related genes and pathways. Together, these results lead us to suggest that TNT-mediated RNA transfer could be a phenomenon of physiological importance under both normal and pathogenic conditions.

    1. Cell Biology
    2. Neuroscience

    Repair of noise-induced damage to stereocilia F-actin cores is facilitated by XIRP2 and its novel mechanosensor domain

    Elizabeth L Wagner, Jun-Sub Im ... Jung-Bum Shin
    Research Article

    Prolonged exposure to loud noise has been shown to affect inner ear sensory hair cells in a variety of deleterious manners, including damaging the stereocilia core. The damaged sites can be visualized as ‘gaps’ in phalloidin staining of F-actin, and the enrichment of monomeric actin at these sites, along with an actin nucleator and crosslinker, suggests that localized remodeling occurs to repair the broken filaments. Herein, we show that gaps in mouse auditory hair cells are largely repaired within 1 week of traumatic noise exposure through the incorporation of newly synthesized actin. We provide evidence that Xin actin binding repeat containing 2 (XIRP2) is required for the repair process and facilitates the enrichment of monomeric γ-actin at gaps. Recruitment of XIRP2 to stereocilia gaps and stress fiber strain sites in fibroblasts is force-dependent, mediated by a novel mechanosensor domain located in the C-terminus of XIRP2. Our study describes a novel process by which hair cells can recover from sublethal hair bundle damage and which may contribute to recovery from temporary hearing threshold shifts and the prevention of age-related hearing loss.

    1. Biochemistry and Chemical Biology
    2. Cell Biology

    Autophagosome membrane expansion is mediated by the N-terminus and cis-membrane association of human ATG8s

    Wenxin Zhang, Taki Nishimura ... Sharon A Tooze
    Research Article

    Autophagy is an essential catabolic pathway which sequesters and engulfs cytosolic substrates via autophagosomes, unique double-membraned structures. ATG8 proteins are ubiquitin-like proteins recruited to autophagosome membranes by lipidation at the C-terminus. ATG8s recruit substrates, such as p62, and play an important role in mediating autophagosome membrane expansion. However, the precise function of lipidated ATG8 in expansion remains obscure. Using a real-time in vitro lipidation assay, we revealed that the N-termini of lipidated human ATG8s (LC3B and GABARAP) are highly dynamic and interact with the membrane. Moreover, atomistic MD simulation and FRET assays indicate that N-termini of LC3B and GABARAP associate in cis on the membrane. By using non-tagged GABARAPs, we show that GABARAP N-terminus and its cis-membrane insertion are crucial to regulate the size of autophagosomes in cells irrespectively of p62 degradation. Our study provides fundamental molecular insights into autophagosome membrane expansion, revealing the critical and unique function of lipidated ATG8.