The ubiquitin ligase KBTBD8 regulates PKM1 levels via Erk1/2 and Aurora A to ensure oocyte quality

PKM1 is essential for oocyte maturation and quality

Studies addressing PKM1/PKM2 functions in oocytes are lacking. Attending concerns about incomplete depletion of target protein after siRNA-mediated knockdown, and based on the successful use of antibody-mediated protein depletion [18] and our previous experience in oocytes [19], we decided to use antibody transfection to deplete PKM1 (Figure 1A) and other proteins of interest. PKM1 depletion significantly reduced both first polar body (1PB) extrusion (Figures 1B and 1C), and ATP content (Figure 1D). Furthermore, mitochondrial membrane potential (assessed by JC-1 staining) was also reduced (Figures 1E and 1F), while ROS levels increased significantly (Figures 1G and 1H). These results indicated that PKM1 expression crucially affects multiple aspects of oocyte quality. However, unexpectedly, PKM2 depletion caused very similar effects (Supplementary Figures 1A and 1G).

KBTBD8 regulates PKM1 levels and is essential for oocyte quality

Next, we investigated how is PKM1 expression regulated. In another project, we observed that KBTBD8, a female fertility factor with ubiquitin ligase activity, interacted with PKM. So herein we first compared PKM1 localization and abundance in control and KBTBD8-depleted oocytes. In control oocytes PKM1 immunofluorescence showed a scattered punctuated pattern throughout the cytoplasm, as well as enrichment at spindle poles. In contrast, PKM1 staining of both cytoplasm and spindle poles diminished significantly after antibody-mediated depletion of KBTBD8 (Figure 2A). Western blotting confirmed that KBTBD8 depletion dramatically reduced PKM1 expression, but had little effect on PKM2 levels (Figure 2B). Similarly, RT-PCR demonstrated that PKM1 mRNA levels decreased significantly, while PKM2 mRNA levels remained unchanged (Figure 2C), in KBTBD8-depleted oocytes. PKM1 and PKM2 mRNA bands were verified by sequencing (Figure 2D). Next, we examined whether oocyte quality was affected by KBTBD8 depletion. In KBTBD8-depleted oocytes the mitochondria formed large, irregularly distributed aggregates (Figure 2E). Accordingly, both intracellular ATP levels (Figure 2F), and mitochondrial membrane potential (Figures 2G and 2H) were reduced significantly. In addition, apoptosis (Figures 2I and 2J), and ROS generation (Figures 2K and 2L) were considerably elevated after KBTBD8 depletion. These results suggest that KBTBD8 depletion severely compromises oocyte quality, probably due to decreased PKM1.

KBTBD8 is essential for meiosis and regulates multiple key kinases

Due to the significant impact of KBTBD8 inhibition on oocyte quality, we further investigated its expression during meiosis, and the effects of its depletion on meiotic phenotypes and regulatory enzymes. We first compared mRNA expression of all KBTBD family members in oocytes. KBTBD6, KBTBD7, and KBTBD8 were similarly and prominently expressed, while the other KBTBD members were almost undetectable (Figure 3A, and Supplementary Table 1). Furthermore, KBTBD8 was far more abundant in ovaries than in testes (Figure 3B), and in oocytes than in granular cells (GCs) (Figure 3C), indicating that KBTBD8 might function primarily in female gametogenesis. Temporal expression analysis showed that ovarian KBTBD8 expression was very low on post-natal day 1 (PND1), increased abruptly by PND14, and peaked on PND21, when initial follicle activation/maturation occurs (Figure 3D). These results indicated that KBTBD8 functions mainly during meiotic transition from GV to MII, rather than in the meiotic prophase (from leptone to diplotene). Next, we further examined KBTBD8 expression and subcellular localization during meiosis. Whereas KBTBD8 levels remained constant through all stages (Figure 3E), its expression was concentrated within nuclei during GV and at the spindle poles during MI and MII (Figure 3F).

We next examined the meiotic phenotype after KBTBD8 depletion (Figures 4A and 4B). At 8 hours of IVM (in vitro maturation) the spindle assembly and chromosome alignment were severely disrupted (Figure 4C); at 16 hours of IVM, the percentage of mature oocytes (as 1PB) was significantly decreased (Figures 4D and 4E). Furthermore, the fertilization rate and the rate of normally fertilized eggs (2PN) were both significantly decreased after KBTBD8 depletion (Figures 5A and 5B).

Next, we assessed whether KBTBD8 depletion affected the activities of kinases known to be essential for oocyte meiosis. We found that Cyclin B levels were reduced, while phosphorylated Cdk1 (p-Cdk1) was increased, indicating that KBTBD8 depletion decreased maturation promoting factor (MPF) activity. This notion was further supported by the finding that germinal vesicle breakdown (GVBD), which depends on MPF activity, was also decreased in KBTBD8-depleted oocytes after 3 hours of IVM (Supplementary Figures 2A and 2B). Furthermore, phosphorylated Akt (p-Akt) was also markedly diminished (Figure 4F, and Supplementary Figures 2C and 2F).

KBTBD8 has been shown to act as a ubiquitin ligase and to interact with cul3 in neurons [20]. Through CoIP experiments, we found that KBTBD8 also interacts with cul3 in oocytes (Figure 4G).

The KBTBD8→Erk1/2→Aurora A axis regulates PKM1 levels

To uncover the mechanism by which KBTBD8 maintains oocyte quality, we initially conducted a comparative RNA-seq analysis in control and KBTBD8-depleted (by sgRNA transfection) NIH3T3 cells. We observed a string of interacting proteins from different families, where KBTBD8 and PKM seemed to be at the upstream and downstream ends, respectively, while Aurora A appeared to be in the middle. In addition, Erk1/2 could also be fitted in the cascade, upstream of Aurora A according to available references.

Supporting this model, we found that KBTBD8 depletion led to decreases in phosphorylated Erk1/2 (p-Erk1/2) and phosphorylated Aurora A (p-Aurora A) (Figures 6A and 6B), suggesting that KBTBD8 acts upstream of these two proteins. Next, we found that Erk1/2 inhibition decreased p-Aurora A levels (Figure 6C), confirming that Aurora A is a downstream target of Erk1/2. Therefore, the final integrated string seemed to be KBTBD8→Erk 1/2→Aurora A→PKM.

Next, we investigated whether the kinases above regulated the expression of PKM isoforms. We found that similar to KBTBD8 depletion, inhibition of either Erk1/2 or Aurora A decreased PKM1 but not PKM2 expression (Figures 6D-G). These results indicated that the KBTBD8→Erk 1/2→Aurora A pathway specifically regulates PKM1 levels (Figure 6H).

Transcriptome-wide characterization of KBTBD8-regulated pathways

Next, we compared mRNA sequencing data between control and KBTBD8-depleted oocytes to further uncover whether KBTBD8 might regulate oocyte meiosis and maintain oocyte quality via other mechanisms. Overall, there were 2304 differentially expressed genes (KBTBD8/control log2>2; Supplementary Datasets 1 and 2). Of those, 69% (1584) were downregulated (Figure 7A, and Supplementary Dataset 1), indicating that the primary effect of KBTBD8 was to stabilize mRNAs. Although the log2 ratio for PKM was less than 2 (90.83/50.51=1.84), its downregulation is consistent with that seen in KBTBD8-depleted oocytes, and agrees well with our results (Figure 2A). Thus, PKM was included among other downregulated genes (Figure 7A, individual small red region).

We then searched for differentially expressed splicing proteins that may participate in PKM splicing. Of 44 RBM members, RBM10, 38, and 43 were downregulated (log2 ratios ≈2) (Supplementary Dataset 3). RBM4 has been reported to increase the PKM1/PKM2 ratio during the differentiation of MSCs into neurons [14]. Although the log2 ratio of RBM4 was just 0.73, (Supplementary Dataset 3), changes at the mRNA level do not always reflect those occurring at the protein level, so we included this RBM as a candidate for PKM splicing as well. Thus, we propose that these RBMs might be the splicing enzymes that convert pre-PKM into PKM1 (Figure 7A).

Next, we focused on the top 200 (log2 ratios > 5) differentially expressed genes and analyzed their interactions. We found that 18 of those genes could interact with the proteins in our pathway model (Figure 7B), suggesting that KBTBD8 might affect the stability or function of other proteins besides those in our model.

Next, we completed an overall pathway classification of all the differentially expressed genes, which revealed their involvement in multiple pathways (e.g. AMPK, HIF-1, Notch, PI3K-Akt, etc) essential to the regulation of oocyte meiosis and quality (Figure 7C and Supplementary Table 2). Finally, q-PCR analysis of 7/200 genes identified above verified the credibility of our sequencing data (Figure 7D and Supplementary Table 3). These results indicated that, in addition to the signaling pathway highlighted in the present study, KBTBD8 might also function in other pathways in oocytes.

Chemicals, antibodies, and inhibitors

All chemicals and reagents were obtained from Sigma (St. Louis, MO, USA) unless otherwise stated. Primary antibodies were mouse anti-β-actin (Cat#: A5316-100; Santa Cruz, Dallas, TX, USA); mouse anti-GAPDH (Cat#: 30201ES60; YEASEN, Shanghai, China); human anti-centromere CREST (Cat#: 15-234; Antibodies Incorporated, Davis, CA, USA); rabbit polyclonal anti-phospho Aurora Kinase (p-Thr288) (Cat#: SAB4503898; Sigma); rabbit polyclonal anti-cyclin B1 (Cat#: 55004-1-AP; Proteintech, Rosemont, IL, USA); rabbit polyclonal anti-Kbtbd8 (Cat#: 1096; ZoonBio Biotechnology, Nanjing, China); rabbit polyclonal anti-CDK1 (p-Thr14) (Cat#: OAAN02724; Aviva Systems Biology, San Diego, CA, USA); mouse anti-diphosphorylated Erk-1&2 (Cat#: M9692; Sigma); rabbit polyclonal anti-PKM1 (Cat#: ab116271; Abcam, Cambridge, UK); and rabbit polyclonal anti-PKM2 (Cat#: BS6443; Bioworld Technology, Dublin, OH, USA). Secondary antibodies were Cy2-conjugated donkey anti-mouse IgG (Code: 715-225-150), Rhodamine (TRITC)-conjugated donkey anti-human IgG (Code: 709-025-149), Cy2-conjugated donkey anti-human IgG (Code: 709-225-149), and Cy2-conjugated donkey anti-rabbit IgG (Code: 711-225-152), all purchased from Jackson ImmunoResearch Laboratory (West Grove, PA, USA). Horseradish peroxidase (HRP)-conjugated rabbit anti-goat IgG and HRP-conjugated goat anti-mouse IgG were purchased from Vazyme (Nanjing, Jiangsu, China). Aurora A Inhibitor I (Cat#: S1451) and the Erk1/2 inhibitor SCH772984 (Cat#: S7101), both from Selleck Chemicals (Houston, TX, USA), were dissolved in dimethyl sulfoxide at a concentration of 5 or 10 mM. GV oocytes were treated with 10 μM Aurora A Inhibitor or ERK1/2 inhibitor for 24 h.

Oocyte collection and culture

GV oocytes were obtained from the ovaries of 3-4 weeks-old ICR mice supplied by the Animal Core Facility of Nanjing Medical University. The mice were euthanized by CO₂ inhalation and cervical dislocation, and ovaries were isolated and placed in operation medium (Hepes) with 2.5 nM milrinone and 10% fetal bovine serum (FBS) (Gibco, Waltham, MA, USA). Oocytes were removed from the ovary by puncturing the follicles with a hypodermic needle. Cumulus cells were washed off the cumulus-oocyte complexes, and groups of 50 isolated denuded oocytes were placed in 100 μl droplets of culture medium under mineral oil in plastic dishes. The culture medium was MEM+ (MEM with 0.01 mM EDTA, 0.23 mM Na-pyruvate, 0.2 mM penicillin/streptomycin, 3 mg/ml BSA) containing 20% FBS. Oocytes were grown at 37.0°C, 5% O₂, 5% CO₂ in a humidified atmosphere. Before in vitro maturation (IVM), all culture media included 2.5 nM milrinone to prevent the resumption of meiosis. All experiments were approved by the Animal Care and Use Committee of Nanjing Medical University and were performed following institutional guidelines.

Antibody transfection

Chariot™ Protein Delivery Reagent (Active Motif, Carlsbad, CA, USA) was used for antibody transfection. Briefly, two tubes, one containing 1 µl Chariot reagent (1mg/ml in 50% DMSO) in 5 µl sterile water, and the other 1 µg antibody in PBS (final volume = 6 µl) were first set up. Next, both solutions were mixed together gently and incubated at room temperature for 30 min to allow the formation of Chariot-IgG complexes, which were then added into the 100 µl MEM+ droplet containing 50 oocytes. After 12–14 h, the oocytes were washed to remove the complex-containing MEM+, and after 1–2 h another two rounds of antibody transfection were repeated to ensure the effectiveness of antibody-mediated inhibition. During the whole procedure, typically 40–44 h long, 2.5 nM milrinone was always included to prevent the resumption of meiosis. Next, oocytes were transferred into milrinone-free MEM+ and cultured for 8 or 16 h, then subjected to the experiments described below. Antibodies for transfection were thoroughly buffer-exchanged (over 10⁴ dilutions of original buffer) into PBS/50% glycerol with size-exclusion spin columns (Millipore, cutoff, 100KDa; spin speed, 5000 rpm) to remove antiseptics (usually NaN3) contained in the original formulation.

Mitochondrial staining and ATP measurements

For mitochondrial staining, oocytes were incubated in Hepes containing 100 nM Mito Tracker (Cat#: M7521, Invitrogen, Carlsbad, CA, USA) and 10 µg/ml Hoechst 33342 (Sigma) for 30 min. Images were taken with an Andor Revolution spin disk confocal workstation.

For ATP measurements, the oocytes were first lysed with 100 µl ATP lysis solution (Cat#: S0026, Beyotime) on ice. The samples were then detected by enzyme-labeled instrument Synergy2 (BioTek, Winooski, VT, USA) to evaluate ATP level.

Assay of mitochondrial transmembrane potential

Oocytes were incubated at 37°C for 20 min with the fluorescent potentiometric indicator JC-1 (Cat#: 40706ES60, YEASEN) diluted 1:200, then washed twice with PBS and added to droplets (50 µl) of culture medium. Images of green fluorescence (JC-1 as monomers at low membrane potentials) and red fluorescence (JC-1 as aggregates at higher membrane potentials) were captured using confocal microscopy as above. Mitochondrial depolarization is indicated by a decrease in the red/green fluorescence intensity ratio.

Detection of ROS generation

The ROS Assay Kit (Cat#: S0033, Beyotime) was used to detect ROS generation in oocytes. Briefly, oocytes were incubated with dichlorofluorescein diacetate (DCFH-DA) probe for 20 min at 37°C in the dark, washed, and mounted on slides for confocal imaging.

Apoptosis detection

Apoptosis was detected with an Annexin V-FITC/PI Apoptosis Detection Kit (Cat#: 40302ES20, YEASEN). Oocytes were stained with 5 μl annexin V‐FITC and 10 μL PI Staining Solution diluted in 100 μl binding buffer for 15 min in the dark at RT. After washing, oocytes were mounted onto glass slides, and images were obtained as above.

Immunofluorescence staining

Oocytes were briefly washed in PBS with 0.05% polyvinylpyrrolidone (PVP), permeabilized in 0.5% Triton X-100/PHEM (60 mM PIPES, 25 mM Hepes, pH 6.9, 10 mM EGTA, 8 mM MgSO₄) for 5 min and washed three times rapidly in PBS/PVP. Next, the oocytes were fixed in 3.7% paraformaldehyde (PFA)/PHEM for 20 min, washed three times (10 min each) in PBS/PVP, and blocked with blocking buffer (1% BSA/PHEM with 100 mM glycine) at RT for 1 h. Oocytes were then incubated at 4°C overnight with primary antibody diluted in blocking buffer, washed three times (10 min each) in PBS with 0.05% Tween-20 (PBST), incubated at room temperature for 45 min with secondary antibody diluted in blocking buffer (1:750 in all cases), and washed three times (10 min each) in PBST. Finally, DNA was stained with 3 µg/ml Hoechst 33258, and the oocytes were mounted onto a slide with mounting medium (0.5% propyl gallate, 0.1M Tris-HCl, pH 7.4, 88% glycerol) and covered with a cover glass (0.13–0.17 µm thick). To maintain oocytes’ structure, two strips of double-stick tape (90 µm thick) were positioned between the slide and cover glass. Images were obtained as above.

Co-Immunoprecipitation (Co-IP)

For Co-IP experiments, 5 μg of rabbit anti-KBTBD8 or rabbit anti-Cul3 antibody was first coupled to 30 μl protein-A/G beads (Macgene, Beijing, China) for 4 h at 4 °C on a rotating wheel in 250 μl IP buffer (20 mM Tris-HCl, pH 8.0, 10 mM EDTA, 1 mM EGTA, 150 mM NaCl, 0.05% Triton X-100, 0.05% NP-40, 1 mM phenylmethylsulfonyl fluoride) with 1:100 protease inhibitor (Sigma) and 1:500 phosphatase inhibitor (Sigma). Meanwhile, 600 oocytes were lysed and ultra-sonicated in 250 µl IP buffer and then pre-cleaned with 30 µl protein A/G beads for 4 h at 4 °C. Then, protein A/G-coupled KBTBD8 or Cul3 antibody was incubated overnight at 4 °C with pre-cleaned oocyte lysate supernatant. Finally, after three washes (10 min each with 250 ul IP buffer), the beads with bound immune complexes were subjected to western blotting with KBTBD8 or Cul3 antibody in parallel.

RNA-sequencing

RNA sequencing was performed by BGI-Wuhan Co. Ltd. in 3 control and 3 KBTBD8-depleted oocyte batches (each containing 200 oocytes) that were freshly frozen in liquid nitrogen immediately at the end of treatment. RNA was amplified using the SMARTer Ultra Low Input RNA Kit, the resulting cDNA fragmented using the Agilent2100 chip, and paired-end libraries constructed using the Nextera XT DNA Library Preparation Kit. The libraries were sequenced on an Illumina HiSeq.

Data analysis and statistics

All experiments were repeated at least three times. Measurements on confocal images were performed with ImageJ (NIH). Data are presented as mean ± SEM or percentages. Statistical comparisons between two groups were made with Student’s t test. Statistical significance was set at P < 0.05.