Inhibition of neddylation causes meiotic arrest in mouse oocyte

ABSTRACT

Mammalian oocyte meiosis is a special form of cell division that provides haploid gametes for fertilization. Unlike in mitosis, post-translational modifications (PTMs) are more crucial during meiosis because of the absence of de novo mRNA transcription. As a classic PTM, protein neddylation is a biological process that mediates protein degradation by modifying cullin proteins and activating the Cullin-Ring E3 ligases. This process plays important roles in various biological processes such as autophagy and tumorigenesis. However, the function of neddylation in germ cells is unknown. In this study, we observed that the inhibition of neddylation by its specific inhibitor MLN4924 significantly arrests mouse oocyte at the stage of metaphase during meiosis. The arrested oocytes display impaired spindles with over-activation of spindle assembly checkpoint (SAC). Accordingly, we identified early mitosis inhibitor 1 (Emi1), a key inhibitor of anaphase-promoting complex/cyclosome (APC/C^Fzr1), as a substrate of neddylation-mediated protein degradation. Thus, our study uncovered an unknown role of neddylation in female germ cells and suggests that proper neddylation is essential for oocyte maturation.

KEYWORDS:

• Neddylation
• Nedd8
• spindle assembly checkpoint
• oocyte maturation

Introduction

Infertility and sterility are currently a global health issue affecting 12–28% of couples [1]. Errors in oocyte meiosis comprise one of the major causes of infertility. During meiosis, oocytes arrest at the germinal vesicle (GV) phase for decades and enter germinal vesicle breakdown (GVBD) in response to a surge of luteinizing hormone [2,3]. Subsequently, the oocyte enters the prophase of metaphase I (Pro-MI), during which spindles assemble, centrosomes move to the poles and chromosomes are condensed. Next, during the metaphase of meiosis I (MI) homologous chromosomes align on the metaphase plate. Spindle microtubules attach to the kinetochores, preparing for chromosome segregation in the following anaphase I. Accurate spindle assembly and chromosome segregation are indispensable for successful meiosis [4]. During the metaphase of meiosis II (MII), the oocyte excludes the first polar body and awaits fertilization. Any abnormality in oocyte maturation can lead to female infertility, birth defects, and failures in assisted reproductive technology [5,6].

As de novo mRNA transcription does not occur during oocyte maturation [7], protein post-translational modifications (PTMs) become obviously critical for oocyte maturation. Ubiquitination is a classical PTM that facilitates the degradation of target proteins [8]. Analogous to ubiquitination, neddylation is also catalyzed through a three-step enzymatic cascade. Specifically, it is a process wherein neural precursor cell expressed developmentally down-regulated 8 (Nedd8) is covalently conjugated to its substrates [9], relying on its own Nedd8-activating enzyme (E1), Nedd8-conjugating enzyme (E2) and specific Nedd8 ligases (E3s) [10–13]. As a ubiquitin-like protein, Nedd8 shares 59% amino acid identity and 76% similarity with ubiquitin [14,15] and is ubiquitously expressed in adult tissues. Nedd8 contains a carboxy-terminal tail and covalently binds target proteins via the Gly-Gly sequence of the tail [16]. It is first synthesized as a precursor with no function, then the precursor is hydrolyzed at the G76 residue and becomes mature Nedd8 [17,18]. Mature Nedd8 is activated by Nedd8 E1 activating enzyme (NAE1), then transferred from E2 (including UBE2M and UBE2F) to a lysine residue, and finally covalently conjugated to its substrates via isopeptide bonds with the help of E3 during neddylation process [17,19]. Unlike ubiquitination, the Nedd8 pathway regulates the activity rather than the abundance of its substrates [20]. Although neddylation plays important roles in various biophysical and pathological processes of somatic cells [11,13], whether it functions during meiosis is still unknown.

It is known that the cullin proteins are the most important substrates of neddylation [21]. They are core structures of the Cullin-RING E3 complex and are involved in the ubiquitination of many important proteins. Cullin proteins belong to a superfamily including eight proteins (Cullin1, Cullin2, Cullin3, Cullin4A, Cullin4B, Cullin5, Cullin7, and Cullin9) in mammals [22]. Among these, Cullin1 exerts multiple biological effects on processes such as cell cycle progression [23]. Inhibition of neddylation leads to diminished activity of cullin proteins and an increase of their substrates because of impaired ubiquitination [24]. However, whether neddylation functions during oocyte maturation is unclear.

In this study, we found that both Nedd8 and Cullin1 were expressed in mouse oocytes and the oocytes were arrested at the MI phase after the inhibition of neddylation with MLN4924 (a selective inhibitor of NAE1) treating and by knocking down the Nedd8 of mouse oocytes. The arrested oocytes displayed impaired spindles with over-activation of spindle assembly checkpoint (SAC) together with failed chromosome segregation. Based on proteomics analysis, we identified Emi1, a key inhibitor of anaphase-promoting complex/cyclosome (APC/C^Fzr1), as a substrate of Nedd8-mediated protein degradation. Thus, our study uncovers an unknown role of neddylation in oocytes and suggests that proper neddylation is essential for meiotic maturation.

Materials and methods

Antibodies

Rabbit monoclonal anti-Nedd8 antibody, rabbit monoclonal anti-Bub3 antibody, goat polyclonal anti-BubR1 antibody, and rabbit polyclonal anti-Mad2 were purchased from Abcam (USA; Cat#: ab81264, ab133699, ab28193, ab70383, respectively). Mouse monoclonal anti-Emi1 antibody was purchased from Santa Cruz Biotechnology (USA; Cat#: sc-365,212). Rabbit polyclonal anti-Fzr1 antibody was purchased from Novus (USA; Cat#: NBP2-15,840). Rabbit polyclonal anti-Cullin1 antibody was purchased from Proteintech (USA; Cat#: 12,895–1-AP). Mouse monoclonal anti-α-tubulin-FITC antibody was purchased from Sigma (USA; Cat#: F2168). Alexa Fluor 555 goat anti-rabbit IgG (H + L) (USA; Cat#: A21429) and Alexa Fluor 555 donkey anti-mouse IgG (H + L) were purchased from Invitrogen (USA; Cat#: A31570).

Oocyte collection and culture

Animal care and handling were conducted in accordance with the Institutional Animal Welfare and Ethics Committee Policies of Peking University Health Science Center. Female ICR mice (4–6 weeks) were hormonally stimulated with 5 IU pregnant mare serum gonadotropin for 44–46 hours and sacrificed by cervical dislocation. GV-stage oocytes were collected from ovaries and transferred into warmed M2 medium (USA, Sigma). To determine the optimal dose of MLN4924 for oocyte cultivation, gradient concentrations of MLN4924 (0.1 μM, 0.5 μM, 1 μM and 5 μM) were utilized. These oocytes were cultured at 37°C in a humid atmosphere of 5% CO₂. For subsequent analysis, oocytes were harvested at GV, GVBD, Pro-MI, MI, and MII stages, respectively at 0, 2, 6, 8, and 12 h.

Dot blotting

Mouse oocytes of GV, GVBD, Pro-MI, MI, and MII stages were collected, and 30 oocytes were used for each sample. After lysis in RIPA buffer (China, Beyotime), the lysate was spotted on a nitrocellulose membrane. The membrane was dried in air and blocked with 5% skim milk (USA, Becton Dickinson) in TBST (China, Applygen) for 1 h at room temperature, and then washed three times with TBST. Then, the membrane was incubated in anti-Nedd8 antibody (1:1000) at 4°C overnight. After three washes in TBST, the membrane was incubated with horseradish peroxidase (HRP)-conjugated secondary antibody for 1 h at room temperature (RT). The dot blot was detected using an enhanced chemiluminescence (ECL) detection system (China, Tanon).

Immunoblotting

All samples were lysed in RIPA buffer containing a proteinase inhibitor on ice for 15 min. Proteins were extracted after centrifuging (12,000 × g, 10 min, 4°C). After the proteins were boiled for 5 min, proteins were loaded and separated using sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Proteins were electrically transferred to polyvinylidene fluoride membranes (USA, Millipore). These membranes were blocked in TBST containing 5% skim milk for 1 h at RT, and then incubated with specific primary antibodies (diluted by 1% skim milk in TBST) overnight at 4°C. After three washes with TBST, membranes were further incubated with corresponding HRP-conjugated secondary antibodies for 1 h at RT. The membranes were incubated with ECL buffer and visualized using an automatic ECL system.

Silver staining and mass spectrum preparation

Total proteins of mouse oocytes were extracted as previously described. Protein samples were separated by SDS-PAGE. Silver staining of all protein bands was performed according to the standard protocol of the Pierce ^TM Silver Stain for Mass Spectrometry kit (USA, Thermo Scientific). Bands of predicted location in the gel was cut and performed in-gel trypsin digestion for the following mass spectrum assay.

RNA extraction and quantitive real-time PCR (qRT-PCR)

Total RNA was extracted from mouse oocytes (200 oocytes/sample) using the RNeasy Micro Kit (Germany, QIAGEN) according to the manufacturer’s instructions. Then, 1 μg of total RNA was subjected to reverse transcription and converted to cDNA using a RevertAid First Strand cDNA Synthesis Kit (USA, Thermo Fisher Scientific) following the manufacturer’s instructions. qRT-PCR was performed using SYBR Green master mix with a QuantStudio 3 Real-Time PCR System (USA, Applied Biosystems). Each experiment was performed three times with triplicate samples. The primers sequences were as follows: Cul1 (5′-AGTGGGAAGATTACCGATTCTCC; 3′-CACGGCGAACCCAATGTCTA); Cul2 (5′-ATGTTGGAATATGTCGAGCGAG; 3′- AGTGGTTCAGGATAGGCCACA); Cul3 (5′-GAGATCAAGTTGTACGGTATGGG; 3′- GCTCTCTGGCAATCATATCCAAT); Cul4A (5′-TCTCACAAAGTCTCCCCAACG; 3′-AGGACGTAGGTTCGATCCAGA); Cul4B (5′-CGACTCAGAAGATAACAGTTCGG; 3′-GGCAGACACTTTTCATAGCTCC); Cul5 (5′-CGCCAGGAATCTGTAACAAAACA; 3′-CAGTACACGAGCCTGTGCTT); Cul7 (5′-GACCCAGATCCTTCTGTCATTG; 3′-GTGTTCCGTGAGAGTAGCCTG); Cul9 (5′-AGCGAAGAGGGCAAAGTGAG; 3′-CCTTTAGGTATTGGCCGCTC).

Immunofluorescence staining and image analysis

For immunofluorescence analysis, the zona pellucida was removed using hyaluronidase (1 mg/mL, 5 min). To detect protein expression, eggs were fixed in 4% paraformaldehyde in PBS (pH 7.4) for 30 min at RT and permeabilized using 0.5% triton X-100 in PBS for 30 min at RT. The oocytes were then incubated in blocking buffer (1% BSA in PBS) for 1 h at RT followed by incubating with primary antibodies (diluted in blocking buffer) at 4°C overnight, after washing three times (10 min each) in washing buffer (0.1% tween-20, 0.01% triton X-100 in PBS). Appropriate fluorescent secondary antibodies were used to label the oocytes for 1 h at RT. To label the DNA, oocytes were incubated in 10 μg/mL Hoechst 33342 for 15 min at RT, or were labeled with a direct-fluorescence α-tubulin antibody (1:200) for 1 h at RT. The immunostaining was visualized using a confocal microscope (Germany, Carl Zeiss 710). To ensure comparable results, all settings among different groups and replications were kept constant.

Chromosome spreading

According to previous reports [25], the zona pellucida was removed and oocytes were transferred to M2 medium for a 5-min recovery. Oocytes were lysed and the chromosomes were fixed in a drop of hypotonic fix buffer (1% paraformaldehyde in H₂O, 0.15% triton X-100, 3 mM dithiothreitol, pH 9.2) on a glass slide for 30 min in a drying oven at 37°C. The immunofluorescence process and image analysis were continued as described previously herein.

Microinjection and RNAi experiment

Fully grown oocytes of GV-stage were microinjected with non-targeting (control) or targeting small interfering RNA (siRNA) by a Diaphot ECLIPSE Ti-S inverted microscope (Japan, Nikon) equipped with MM-93B hydraulic three-dimensional (3-D) micromanipulators (USA, Narishige). Microinjected oocytes were cultured in M2 medium with 2.5 μM milrinone for 16 h to facilitate the targeting mRNA depletion. Next, the GV-stage oocytes were washed and cultured with fresh M2 medium. The siRNA sequences were as follows: Nedd8 (5ʹ-GAUUGAGAUAGACAUCGAATT-3ʹ, 5ʹ-UUCGAUGUCUAUCUCAAUCTT-3ʹ; China, GenePharma), Cullin1 (5ʹ-GCCGCAUGUAUAAUCUUGUTT-3ʹ, 5ʹ- ACAAGAUUAUACAUGCGGCTT-3ʹ; China, GenePharma).

3-D modeling

Oocytes were imaged using the Zeiss program as a z-series of 0.5-μm intervals to capture the entire spindle and chromosome structures based on three-dimensional confocal imaging. The Imaris program (USA, Bitplane) was used to further investigate oocyte meiosis status. Surface Creation Wizard was used to reconstruct Emi1 localization and chromosome status.

Statistical analysis

All experiments were repeated at least three times to ensure consistency. Means and standard deviations were plotted. Two-tailed Student’s t-test and ANOVA was performed for statistical analysis. The level of significance was accepted as p < 0.05.

Results

Expression and subcellular localization of nedd8 during mouse oocyte meiosis

To investigate the function of Nedd8 during oocyte meiosis, we first examined its expression level at each stage of oocyte maturation by dot blotting. Oocytes were collected after culture for 0, 2, 6, 8, and 12 h, corresponding to GV, GVBD, Pro-MI, MI, and MII stages, respectively. As shown in Figure 1a, Nedd8 was constantly expressed from GV to MII stages (p > 0.05; Figure 1a). Furthermore, we examined the subcellular localization of Nedd8 by immunofluorescence staining. Results showed that in GV stage, Nedd8 mainly accumulated in the nucleus (Figure 1b), whereas during GVBD and Pro-MI stages, it was localized to the cytoplasm (Figure 1c-d). During the following MI and MII stages when spindles had been already assembled, Nedd8 was found to be distributed on the spindle (Figure 1e-f). These results suggested that proteins involved in MI and MII stages might exhibit Nedd8 modifications and that the neddylation pathway might play a potential role in oocyte maturation.

Figure 1. Expression and subcellular localization of Nedd8 during mouse oocyte meiosis.

(a) Mouse oocytes were collected after culture for 0, 2, 6, 8, and 12 h, corresponding to germinal vesicle (GV), GV breakdown (GVBD), prophase of metaphase I (pro-MI), metaphase I (MI) and metaphase II (MII) stages, respectively. Whole lysates from 30 oocytes were loaded in each lane for dot blotting. Nedd8 levels were determined using an anti-Nedd8 antibody. The relative staining intensity of Nedd8 was assessed by densitometry in the histogram. Error bars represent the standard deviation. (b-f) Oocytes of different stages were collected for immunofluorescence staining. Blue: DNA; Green: α-tubulin; Red: Nedd8. Scale bar, 10 μm.

Inhibition of neddylation causes MI arrest and spindle disorder during oocyte maturation

To further investigate the role of neddylation during oocyte meiosis, we used MLN4924, a first-in-class NAE1 inhibitor [24], to disrupt the Nedd8 pathway. Mouse oocytes were cultured with gradient concentrations (0, 0.1 μM, 0.5 μM, 1 μM and 5 μM) of MLN4924 for 12 h. Results showed that oocytes exhibited MI arrest and the polar body exclusion (PBE) rate dose-dependently decreased when the MLN4924 concentration was higher than 0.5 μM (47.75 ± 5.17%, 0.5 μM MLN4924 vs 80.49 ± 2.58%, control, p < 0.01; 8.52 ± 1.11%, 1 μM MLN4924 vs 80.49 ± 2.58%, control, p < 0.001; Figure 2a-b). The inhibitory effect reached the maximum at 1 μM MLN4924, which is similar to that of 5 μM MLN4924 (8.52 ± 1.11%, 1 μM MLN4924 vs 7.11 ± 2.05%, 5 μM MLN4924, p > 0.05; Figure 2b). Accordingly, the expression level of Nedd8 significantly decreased at the doses of 0.5 μM (0.56 ± 0.02 vs 1.03 ± 0.05, p < 0.001), 1 μM (0.29 ± 0.01 vs 1.03 ± 0.05, p < 0.001; Figure 2c-d). Therefore, 1 μM MLN4924 was selected for subsequent study. To explore the subcellular oocyte phenotype after inhibition of neddylation, immunofluorescent staining was performed. Results revealed that in the MLN4924-treated group, the oocytes were arrested at the MI stage. Meanwhile, the spindle could not form bipolar structures and chromosomes could not segregate in the MI stage (Figure 2e).

Figure 2. Inhibition of neddylation causes oocyte MI arrest.

(a) Images of oocytes in the control group and MLN4924-treated groups. Oocytes were cultured in medium with different concentrations of MLN4924 (0.1 μM, 0.5 μM, 1 μM and 5 μM) or without MLN4924. Scale bar, 100 μm. (b) Germinal vesicle breakdown (GVBD) rate and polar body exclusion (PBE) rate of control and MLN4924-treated oocytes. Error bars represent the standard deviation. ns: no statistical significance, *p < 0.05, **p < 0.01, ***p < 0.001. (c-d) Expression of Nedd8 after treated with different concentrations of MLN4924. Oocytes were cultured at MI stage and total lysate of 80 oocytes per group was loaded for dot blotting and western blotting. Nedd8 levels were determined using an anti-Nedd8 antibody. β-actin was used as a loading control. Relative intensities of bands are shown in the histograms. Error bars represent the standard deviation. ns: No statistical significance, *p < 0.05, **p < 0.01, ***p < 0.001. (e) Spindle morphology in control and MLN4924-treated oocytes. Blue: DNA; Green: α-tubulin; Red: Nedd8. Scale bar, 10 μm.

As Nedd8 is the most critical factor during neddylation process, siRNA mediated knockdown of Nedd8 was performed to verify the importance of neddylation during mouse oocyte maturation. Oocytes of Pro-MI stage were collected for dot blotting. The results confirmed an effective depletion of Nedd8 (Figure 3a-b). Control and Nedd8 RNAi mouse oocytes were cultured in vitro for 12 h. The results revealed that depletion of Nedd8 had no effect on the GVBD stage (87.53 ± 1.42% vs 82.55 ± 2.047%, p > 0.05; Figure 3d), but induced a significant decrease of PBE rate (81.00 ± 2.34% vs 4.69 ± 1.74%, p < 0.001; Figure 3c-d). Moreover, the immunofluorescent staining results showed that homologous chromosomes could not segregate and spindle failed to form bipolar structure in the Nedd8 RNAi oocytes (Figure 3e). All these results demonstrated that neddylation was critical for meiosis and that inhibiting the neddylation pathway caused MI arrest during oocyte maturation.

Figure 3. Depletion of Nedd8 blocks mouse oocytes maturation process.

(a-b) Control and Nedd8 RNAi oocytes were cultured to Pro-MI stage, and were collected for dot blotting assay and western blotting. β-actin was used as a loading control. Relative intensities of bands are shown in the histograms. Error bars represent the standard deviation. ***p < 0.001. (c) Images of control and Nedd8 RNAi oocytes. Mouse oocytes were cultured to MII stage. Scale bar, 100 μm. (d) GVBD rate and PBE rate of control and Nedd8 RNAi oocytes. Error bars represent the standard deviation. ns: no statistical significance, ns: no statistical significance. ***p < 0.001. (e) Spindle morphology of control and Nedd8 RNAi oocytes. Blue: DNA; Green: α-tubulin; Red: Nedd8. Scale bar, 10 μm.

The SAC is over-activated upon neddylation inhibition

SAC proteins (such as Bub3, Mad2, and BubR1) function during the MI phase to ensure the accuracy of homologous chromosome segregation [26,27]. To analyze SAC status, we performed a chromosome spreading assay by chromosome immunofluorescence staining using anti-Bub3, anti-Mad2, and anti-BubR1 antibodies. In the control group, SAC was activated when the oocytes were cultured for 7 h, while the signal of SAC disappeared after 9 h culture (Figure 4a-c). However, when oocytes were cultured for 9 h with MLN4924, Bub3, Mad2, and BubR1 were still detected at the kinetochores of chromosomes (Figure 4a-c). Besides, the protein level of SAC proteins were similar between control and MLN4924-treated groups (Figure 4d). This indicates that impaired neddylation might affect chromosome segregation as manifested by the over-activation of SAC.

Figure 4. Inhibition of neddylation induces over-activation of the spindle assembly checkpoint (SAC).

Mouse oocytes were cultured in M2 medium for 7 h and 9 h, or in MLN4924 for 9 h, and then collected for chromosome spreading. (a-c) Immunofluorescence staining was performed after chromosome spreading. Core SAC proteins (Bub3, Mad2, and BubR1) were stained using specific primary and secondary antibodies. In control oocytes, the SAC was activated at 7 h, but the positive signal at kinetochores disappeared after MI at 9 h. However, they retained at kinetochores after MI in MLN4924-treated oocytes. Blue: DNA; Red: SAC core proteins. Scale bar, 10 μm. (d) Two hundred oocytes in each sample were lysed, and followed by western blotting for Bub3, Mad2, and BubR1 detection. β-actin was used as a loading control.

Cullin1 is the main substrate of Nedd8 in mouse oocytes

Since neddylation regulates the cellular biochemical functions through modulating the activity of its substrates, we investigated the substrates of Nedd8 in mouse oocytes [22]. Cullins are the most important substrates of Nedd8 to date. To identify their expression in mouse oocytes, we measured the levels of eight cullin transcripts by qRT-PCR. Results showed that Cul1 was the most abundant cullin in mouse oocytes (1.63 ± 0.08; Figure 5a). To demonstrate if Cullin1 has Nedd8 modification in mouse oocytes, we used specific Cullin1 antibody to examine its expression. The results confirmed that Cullin1 was modified by Nedd8 as evidenced by the higher molecular weight bands (Figure 5b,d). Moreover, a detection of Cullin1-Nedd8 complex by mass spectrum further confirmed this speculation (Figure S1). In addition, the expression level of Cullin1-Nedd8 complex in each stage of oocyte meiosis was examined by western blotting. The results revealed that Nedd8 modified Cullin1 was continuously expressed during mouse oocyte (Figure 5b-c). As for the control and MLN4924 treated mouse oocytes, the expression of unmodified Cullin1 in both groups was similar (0.891 ± 0.02 vs 0.87 ± 0.02, p > 0.05; Figure 5d-e), whereas the Cullin1-Nedd8 level decreased in the MLN4924-treated group (0.59 ± 0.02 vs 0.26 ± 0.02, p < 0.001; Figure 5d-e). This indicated that Cullin1 might be the main substrate of Nedd8 in mouse oocytes. MLN4924 treatment suppressed the Nedd8 modification of Cullin1 and downregulated Cullin1 activity in oocytes.

Figure 5. Cullin1 neddylation decreased after neddylation inhibition.

(a) qRT-PCR was performed to examine the expressions of eight cullin transcripts in mouse oocytes (200 oocytes/sample). The relative expression levels are shown in the histograms. Error bars represent the standard deviation. (b-c) The expression of Cullin1 and Cullin1-Nedd8 during mouse oocyte maturation. β-actin was used as a loading control. Relative intensities of bands are shown in the histograms. Error bars represent the standard deviation. （d-e） Anti-Cullin1 antibody was used to examine Cullin1 and Cullin1-Nedd8 in the control and MLN4924-treated mouse oocytes. β-actin was used as a loading control. Relative intensities of bands are shown in the histograms. Error bars represent the standard deviation. ns: no statistical significance, ***p < 0.001. (f-g) Expression of Cullin1 in control and Cullin1 RNAi oocytes. β-actin was used as a loading control. Relative intensities of bands are shown in the histograms. Error bars represent the standard deviation. ***p < 0.001. (h-i) Images of control and Cullin1 RNAi oocytes after cultured for 12 h. GVBD and PBE rates are shown in the histograms. Error bars represent the standard deviation. ns: no statistical significance, ***p < 0.001. Scale bar, 100 μm.

To gain insight into the importance of Cullin1 during mouse oocyte meiosis, gene-targeting siRNA microinjection was performed to knockdown Cullin1 in oocytes. Western blotting showed a successful depletion of Cullin1 (0.96 ± 0.06 vs 0.07 ± 0.02, p < 0.001; Figure 5f-g). Furthermore, we cultured the control and Cullin1 RNAi oocytes in vitro for 12 h to investigate the maturation status in both groups. Results confirmed that the PBE rate of Cullin1 RNAi oocytes was much lower than that of control groups (80.24 ± 2.37 vs 7.17 ± 1.52, p < 0.001; Figure 5h-i). In conclusion, as one of the most critical substrates of Nedd8, Cullin1 involves in the oocyte maturation process. When Cullin1 is depleted, oocyte meiosis is arrested at MI stage. This phenotype is consistent with that of neddylation inhibition.

Inhibition of neddylation increases the expression of Emi1

Emi1 is a known substrate of Cullin1 [28] and plays a critical role in the cell cycle progression of somatic cells [29]. To investigate the role of Emi1 in oocyte meiosis, we performed immunofluorescence assays and western blotting to explore its subcellular localization and expression changes when the Nedd8 pathway was inhibited. The immunofluorescence results showed that Emi1 mainly localized to the mouse oocyte spindle (Figure 6a-b). This result implied that Emil may play similar function in oocytes as in somatic cells. Upon treatment with MLN4924, the Emil expression appeared to be stronger than that in the control group (Figure 6a-b). Western blotting results further confirmed that the expression level of Emi1 increased after treated with MLN4924 (0.81 ± 0.05, n = 3 vs 0.54 ± 0.04, n = 3, p < 0.01; Figure 6c-d). Similarly, the expression of Emi1 increased in both Nedd8 RNAi oocytes (0.81 ± 0.03, n = 3 vs 0.54 ± 0.04, n = 3, p < 0.01) and Cullin1 RNAi oocytes (0.85 ± 0.03 vs 0.54 ± 0.04, n = 3, p < 0.01; Figure 6c-d). These results implied that Emi1 might be regulated by neddylation.

Figure 6. Emi1 is a potential substrate of neddylation-mediated ubiquitination.

(a) Control and MLN4924-treated mouse oocytes were cultured and collected at MI stage for immunofluorescence staining. Blue: DNA; Green: α-tubulin; Red: Emi1. Scale bar, 10 μm. (b) 3-D modeling of control and MLN4924-treated oocytes. 3-D models were constructed using the Imaris program. Isosurface renderings of the oocytes were created and rotated in 3-D space to assess morphology. (c-f) Western blotting detection of Emi1 and Fzr1 protein levels in control, Nedd8 RNAi and Cullin1 RNAi oocytes. β-actin was used as loading control. Relative intensities of bands are shown as bar graph in the histograms. Error bars represent the standard deviation. *p < 0.05, ***p < 0.001.

Since Emi1 is an inhibitor of APC/C^Fzr1 [30] in which Fzr1 is a crucial activator of APC/C and forms a complex with it [31], we tested the expression of Fzr1. To examine if the expression level of Fzr1 could be affected by Nedd8 pathway inhibition, we performed western blotting to detect its protein level. Results showed that compared to the control group, the expression of Fzr1 was decreased in MLN4924-treated group (0.30 ± 0.01, n = 3 vs 0.79 ± 0.04, n = 3, p < 0.001; Figure 6e-f), Nedd8 RNAi (0.27 ± 0.02, n = 3 vs 0.79 ± 0.04, n = 3, p < 0.001) and Cullin1 RNAi groups (0.29 ± 0.04, n = 3 vs 0.79 ± 0.04, n = 3, p < 0.001; Figure 6e-f). These results suggested that neddylation is essential for APC/C activation. Thus, we concluded that when neddylation was inhibited, the expression of Emi1 was increased and that APC/C^Fzr1 was inhibited, resulting in MI arrest in oocytes.

Discussion

Stringent regulation is indispensable for correct mouse oocyte meiosis otherwise it is highly error-prone [6,32,33]. Since transcription does not occur during oocyte meiosis, PTMs appear to be crucial. As a ubiquitin-like PTM, neddylation is implicated in many biophysical regulatory processes such as autophagy, cell proliferation and signal transduction [34–36]. It is a process in which Nedd8 covalently binds to its substrates. Neddylation regulates the activity of its substrates rather than their abundance [20]. Cullins are the most important substrates of Nedd8 and are essential for G1 to S cell cycle transition [21,37]. Moreover, they comprise a family of proteins containing a conserved Cullin domain that can be modified by Nedd8 [19]. In this manner, neddylation regulates the activity of cullin proteins and activates the subsequent ubiquitin transfer [38]. Thus, Cullin1 neddylation regulates the degradation of a set of proteins through the Cullin1-RING type E3 ligases [19]. However, whether this process functions during oocyte meiosis has not been well studied. In this study, we first identified that Nedd8 was constitutively expressed from GV to MII stages. In addition, Nedd8 localized to nucleus during the GV stage and translocated to the cytoplasm during GVBD and Pro-MI stages, after which it finally localized to the spindle. This suggested that proteins involved in spindle assembly and meiosis might be modified by Nedd8.

To investigate how neddylation affects meiosis, we treated oocytes with MLN4924, which is an adenosine sulfamate analogue that can form a covalent adduct with Nedd8, thereby suppressing the Nedd8-AMP step and selectively locking NAE1 into an inactive state [24,39]. Moreover, it has anti-tumor activity and is widely used for the treatment of many types of cancers [39,40]. Using MLN4924 or Nedd8 siRNA to inhibit neddylation, we revealed that oocytes arrested at the MI stage, exhibiting disorders in spindle assembly and chromosome segregation. Thus, we showed for the first time that neddylation was essential for oocyte maturation. As previously mentioned, cullins are core substrates of Nedd8 [21]. Given that eight members of the cullin family exist in cells, we performed qRT-PCR and identified Cullin1 as the major cullin member expressed in mouse oocytes. Since Cullin1 neddylation is required for the activation of CRL E3 ubiquitination ligase, we examined the Cullin1 neddylation status using the specific anti-Cullin1 antibody. In the Cullin1 western blotting, we confirmed that Cullin1 was modified by Nedd8 as evidenced by the two bands representing Cullin1 and Cullin1-Nedd8 respectively. Moreover, there was no significant difference in the Cullin1 expression in either groups. However, compared to the MLN4924-treated group, the level of Cullin1-Nedd8 decreased. These results verified that Cullin1 was the major neddylation substrate in mouse oocytes and when neddylation was inhibited, Nedd8 modified Cullin1 diminished. In this manner, neddylation could further regulate the following ubiquitination mediated by Cullin1 in mouse oocytes.

Based on our hypothesis, the inhibition of neddylation might suppress the degradation of Cullin1 substrates. Emi1, also known as Fbxo5, is a substrate of Cullin1 [28] and an inhibitor of APC/C [41–43]. In contrast, Fzr1 acts as an activator of APC/C and forms a complex with APC/C. During the cell cycle, APC/C can be activated to ensure that cells can exit metaphase [44,45]. Previous studies have shown that Emi1 inactivates APC/C by competing with APC/C for Fzr1 binding, thereby stabilizing APC/C substrates [41,46]. When APC/C^Fzr1 is completely inhibited, cells irreversibly revert to a quiescent state [47]. However, it is still unclear whether and how Emi1 regulates meiosis through neddylation. First, we identified that Emi1 was mainly localized to the spindle of mouse oocytes. It implied that Emi1 might function in spindle assembly. In addition, we examined the expression of Emi1 in the control group and the MLN4924-treated group. In well accordance with previous data, we found that the inhibition of neddylation suppressed the activity of Cullin1, following the increased expression level of Emi1, leading to APC/C^Fzr1 inactivation and subsequent MI arrest. Moreover, the spindle could not form bipolar structures and chromosomes were not segregated. Consistent with our hypothesis, a pathway underlying the regulation of oocyte maturation by neddylation is presented.

Taken together, this study identified for the first time that neddylation plays a key role in mouse oocyte maturation. We propose a model wherein Cullin1 neddylation regulates oocyte meiosis. As a substrate of Cullin1 mediated ubiquitin E3 ligase, Emi1 is possibly ubiquitinated and downregulated to allow oocyte meiosis by activating APC/C^Fzr1. When neddylation is selectively inhibited, Emi1 upregulation blocks the oocyte meiosis. In this manner, oocyte meiosis is disturbed, manifesting as MI arrest, a deficiency of chromosome segregation, and spindle disorders. Errors in meiosis are a major cause of female infertility and in vitro fertilization failure. Accordingly, this study might provide a potential marker for the diagnosis of infertility and offer new clues for the development of novel therapeutic strategies.

Acknowledgments

We are grateful to Dr. Bingteng Xie for the insightful comments and suggestions on the manuscript.

Disclosure statement

The authors declare that they have no competing interests.

Supplementary material

Supplemental data for this article can be accessed here.

Additional information

Funding

This work was supported by the Key Special Fund of Ministry of Science and Technology of China [2017YFC1001501, 2018YFC1003802], the National Natural Science Foundation of China (NSFC) for Excellent Young Scholar [81622035], and the Foundation for Innovative Research Groups of NSFC [81521002].