All the mice used were of the C57BL/6J background and were kept in the animal facilities at the Peking University Shenzhen Hospital. The Clpp^fl/fl mice were produced by the CRISPR/Cas9 and homology-directed repair (HDR) techniques. Two LoxP sites were inserted into the mouse genome to target exon 3-exon 5 of Clpp for deletion ( Figure 1A ). Two guide RNAs (gRNA), donor vector containing the LoxP sites, and Cas9 mRNA were co-injected into fertilized mouse eggs to generate the targeted offspring. The F₀ founder animals were identified by PCR and sequencing analysis. The mice were then bred with wild-type (WT) mice to generate the F₁ generation and to test for germline transmission. The gRNA sequences used are listed in Supplementary Table 1 . The Zp3-Cre mice were then crossed with the Clpp^fl/fl mice to generate an oocyte-specific Clpp knockout mouse line. The mice were housed in a 12-hour light-dark cycle with free access to water and food. All experimental protocols were approved by the ethics committee of the University of Hong Kong-Shenzhen Hospital.

The procedures of genotyping are detailed in a previous report (33). A small tissue piece was excised at the tail tip of mice and was lysed with 200 μL of digestion buffer (50 mM KCl,10 mM pH 9.0 Tris-HCl, 0.1% Triton X-100 and 0.4 mg/ml proteinase K) at 55°C overnight, and then at 98°CC for 10 min to denature the proteinase K in the tail lysate. PCR analysis was performed using the Rapid Taq polymerase (Cat. P222-02, Vazyme, Nanjing, China) according to the manufacture’s protocol. The primers used for PCR are shown in Table S1 .

The in vivo experimental design is as follow: the WT mice and cKO mice were divided into four groups: Clpp^fl/fl, Clpp^fl/fl +CTX, Clpp ^fl/fl; Zp3-Cre and Clpp ^fl/fl; Zp3-Cre +CTX. We intraperitoneally injected 100 μL of 75 μM CTX once a time in Clpp^fl/fl +CTX and Clpp ^fl/fl; Zp3-Cre+CTX group. For the corresponding control group, we used the same injection method to administer 100 μL of PBS. All the experiments were conducted at least three independent biological replications. In this part, all the mice were given one-time injection at 8 weeks of age and collecting samples in 14 days later.

For the fertility assessment, the Clpp cKO or WT female mice (age of mice: 8 weeks, n=6 for each group) were mated with WT male mice (age of mice: 8 weeks) in a mating ratio of 1♂:2♀, in 3 cages for each group. Fertility test will document female mice fertility between postnatal day 56 to postnatal day 280 (8 weeks to 40 weeks). The number of pups from each female was recorded for fertility analysis.

The Clpp cKO and WT female mice (8 weeks old, n=3 for each group) were euthanized by cervical dislocation. The ovaries were immediately collected and fixed in 4% (w/v) paraformaldehyde with phosphate-buffered saline (PBS, Cat. C10010500BT, Thermo Fisher Scientific, MA, USA) overnight before dehydration, paraffin embedding, sectioning at 5 μm thick and mounting on glass slides. The samples were then deparaffinized, stained with hematoxylin solution for 90 sec, washed three times with ddH₂O, mounted with neutral balsam (Cat. G8590, Solarbio, Beijing, China) and imaged under a light microscope. Follicles at different developmental stages were counted in the whole ovaries to determine the total number. Primordial, primary, secondary, antral and atretic follicles were classified as described previously (34).

Mouse oocytes and embryos were collected using standard protocol (27). Female mice were injected with 10 IU PMSG (Cat. P9970-1000 Solarbio, Beijing, China) intraperitoneally to stimulate follicular development. An injection of 10 IU of hCG (Cat.NB1122, NSHF, Ningbo, China) was given at 48 hours post-PMSG injection to induce oocyte maturation and ovulation. Follicles, cumulus-oocyte complex (COCs) or embryos were collected at different time points: (1) at 44 hours post-PMSG injection, ovaries were collected into M2 medium (Cat. M7167, Sigma, MO, USA), ovarian follicles were isolated with a disposable syringe, and GV oocytes were collected from the follicles after removal of the granulosa cells using a pipette with an internal diameter of 75 μm; (2) at 14-16 hours after hCG injection, COCs were collected from the oviduct, and MII oocytes(n=10/15) were isolated after removal of the cumulus cells by incubation in M2 medium containing 1 mg/mL hyaluronidase (Cat. H3506, Sigma, MO, USA) followed by aspiration through a pipette; (3) at 44-48 hours and 92 hours post-hCG injection, 2-cell embryos(n=3) and blastocysts(n=6/7), respectively were collected from the oviduct of mated female mice that were caged with 12-week-old WT male mice immediately after hCG injection. Mating was confirmed by the presence of vaginal plugs in the following morning.

Oocytes were fixed in 4% (w/v) paraformaldehyde in PBS for 30 min, and permeabilized in 0.5% Triton X-100 for 5 min before incubation with 2 μg/mL Alexa Fluor 488 conjugated-anti-α-tubulin antibody (Cat.16-232, Millipore, Billerica, MA, USA) for 1 h, washed three times for 5 min in PBS and stained with 4’,6-diamidino-2-phenylindole (Cat#G1012-10ML DAPI, Servicebio) prior to being examined under a Zeiss LSM 900 confocal microscope. The tubulin staining was examined with an excitation at 488 nm and an emission at 530 nm while that for DAPI with an excitation at 350 nm and an emission at 470 nm and female mice (8 weeks old, n=5 for each group) were used for oocyte sample collection.

Oocytes were incubated in M2 medium containing 500 nM cell permeant MitoTracker™ Red CMXRos (Cat. M7512, Invitrogen, MA, USA) for 30 min at 37 °C in a dark environment with 5% CO₂ in air. After washing three times with fresh M2 medium (2 min each time), the oocytes were mounted on non-fluorescent glass slides for imaging. The images were captured by a Zeiss LSM 900 confocal microscope.

Female mice (8 weeks old, n=3 for each group) were used for oocyte sample collection for this assay. To quantify mtDNA copy number in MII oocytes, the Cox3 fragment was amplified and subcloned into the pCR™2.1-TOPO^® - cloning vector (Invitrogen, MA, USA) as previously described (35). The primer sequences used are listed in Supplementary Table S1 . The One Shot TOP10 competent cells were used for plasmid transformation. After overnight incubation at 37 °C, the recombinant plasmids were extracted using the Qiagen plasmid isolation kit (Cat.12145, Qiagen, Hilden, Germany). The inserted mtDNA fragment was confirmed by DNA sequence analysis. The plasmid DNA was quantified using a NanoDrop 2000 spectrophotometer (Thermo Scientific, MA, USA). A standard curve from 10⁸ to 10¹ plasmid molecules was generated by serial 10-fold dilutions. Single MII oocyte was lysed in 10 μL of lysis solution before incubation at 55°C for 2 hours. The proteinase K in the lysis solution was denatured by heating at 95°C for 10 min and the mixture was used directly for downstream PCR in triplicates for each group. Each 10 μL reaction mixture contained 5 μL of Taq Pro Universal SYBR qPCR Master Mix (Cat. Q712-02, Vazyme, Nanjing, China), 0.3 μM primers, and oocyte DNA. The mtDNA copy number of each oocyte was then extrapolated from the standard curve.

6-carboxy-2’, 7’-dichlorodihydrofluorescein diacetates (carboxy-H2DCFDA; Cat. C-400, Life Technologies, Thermo Fisher Scientific, MA, USA) was used to assess the ROS levels in mouse oocytes (36). The MII oocytes were pre-treated with 10 mM H₂O₂ in M2 medium for 5 min to induce ROS generation. They were then washed and incubated with 30 µM H2DCFDA in M2 medium for 20 min. The oocytes were washed 3 times with M2 medium and then immediately imaged under a Zeiss LSM 900 confocal microscope. The Image J software was used to quantify the fluorescence intensity.

Mitochondrial membrane potential was evaluated with the JC-1 probe (Cat. T3168, Invitrogen, MA, USA). In brief, oocytes were incubated with M2 medium containing 2 μg/mL JC-1 probe for 30 min at 37°C in the dark and then washed 3 times in M2 medium for 3 mins each time. The oocytes were immediately examined under a Zeiss LSM 900 confocal microscope. The JC-1 dye exhibits a potential-dependent accumulation in mitochondria as indicated by an emission shift of fluorescence from green (~529 nm) to red (~590 nm). Mitochondrial depolarization was determined by the ratio of red-to-green fluorescence intensity.

The ATP content of individual oocyte was determined using the ATP Bioluminescent Somatic Cell Assay Kit (Sigma, MO, USA). Each oocyte was mixed with 100 µL of somatic ATP release reagent and 100 μL of ATP mix working solution (diluted 1:25 from ATP Assay Mix stock solution) before loading onto the 96-well plate and incubation at room temperature for 3-5 min to allow hydrolysis of endogenous ATP. The amount of emitted light was immediately measured with a BioTek luminometer (Cat. SYNERGY H1, BioTek, Vermont, USA). Background luminescence was subtracted from all readings. ATP in individual oocytes was determined by comparison to a standard curve generated in the range z.

Total RNA was extracted from 20 oocytes per mice using RNeasy^® Micro Kit (Cat. 74004, Qiagen, Hilden, Germany) and reverse transcription was performed using the HiScript III RT SuperMix for qPCR (+gDNA wiper) kit (Cat. R323-01, Vazyme, Nanjing, China) in two steps. First, template RNA and 4 × gDNA wiper mix were incubated at 42°C for 2 min to remove genomic DNA contamination. The 5X HiScript III qRT SuperMix was then added and the reverse transcription was carried out at 37°C for 15 min and end at 85°C for 5 sec. RT-qPCR was carried out on a Life 7500 (Cat. 4351107, Applied Biosystems™, Thermo Fisher Scientific, MA, USA). Complementary DNA (cDNA) was prepared as described above. Each experiment was repeated at least three times using different animals. The reaction mixture contained 10 µL of the Taq Pro Universal SYBR qPCR Master Mix (Cat. Q712-02, Vazyme, Nanjing, China), 7 µL of H₂O, 1 µL of primers, and 2 µL of cDNA. The 2^-ΔΔCT (CT: cycle threshold) method was used to calculate relative expression levels after normalization to β-actin mRNA expression. The primers used are listed in Table S1 .

Oocytes were incubated in Tyrode’s buffer (pH 2.5) for 30 s at 37°C to remove the zona pellucida. After washing 4 times in M2 medium, the oocytes were fixed in a drop of 1% paraformaldehyde with 0.25% Triton X-100 (Cat. 85111, Thermo Fisher Scientific, MA, USA) and 6 µM DTT (Cat. R0862, Thermo Fisher Scientific, MA, USA) on a glass slide. After air drying, chromosomes were counterstained with DAPI and examined under a confocal microscope.

The sample of GV oocytes were collected from WT and ClpP cKO female C57BL/6 mice and suspended in lysis buffer [50mM HEPES-KOH (pH 7.5), 100mM KCl, 2mM EDTA, 10% glycerol, 0.1% NP-40, 10mM NaF, 0.25mM Na3VO4, and 50mM ßglycerophosphate] supplemented with complete protease inhibitor (Cat. 04693116001, Roche, Basel, Switzerland). The samples were homogenized and centrifuged at 20,000 g for 20min at 4°C, after which the supernatant was retained for western blotting analysis. The proteins in each sample were separated using 8–16% Bis-Tris gels (Cat. M00659, SurePAGE™, GenScript, Nanjing, China) and a mini protein electrophoresis system (Cat. 1658034, BIO-RAD, CA, USA) following the manufacturer’s instructions. The protein bands were then transferred to polyvinylidene fluoride (PVDF) membranes (Cat. IPVH00010, Immobilon, Millipore, MA, USA) via a Mini Trans-Blot Electrophoretic Transfer Cell (Cat. 1703930, BIO-RAD, CA, USA). The immunoreactive bands were detected and analyzed with a Bio-Rad ChemiDoc MP imaging System (Cat. 12003154, BIO-RAD, CA, USA) in conjunction with the Image Lab Software (Bio-Rad, CA, USA). The relative protein levels in each sample were normalized to ß-Actin to standardize the loading variations. The obtained images were analyzed by image J software for gray value analysis.

The results displayed by numerical values are combined with the results of three independent repeated experiments for t-test; for the data displayed by ratios, the ratio of each group is calculated separately after each independent repeated experiment, and then the t-test is performed to evaluate the statistic difference between two groups. Results showed in the figures were given as the mean ± SEM. At least three independent samples were repeated in all experiments. Groups were compared using the two-tailed unpaired Student’s t-tests. Statistically significant P-values (< 0.05, < 0.01 and < 0.001) were indicated by asterisks (*, ** and ***, respectively), while ‘ns’ represents not significantly different. Graphs were generated using the Microsoft Excel and GraphPad prism 8. Figures were prepared with the CorelDraw version X8 (Corel Corp., Ottawa, ON, Canada).

The Clpp ^fl/fl; Zp3-Cre mice were identified by genotyping ( Figure 1B ). The 213-base pair (bp) band (higher band in Figure 1B ) of the heterozygous (Het.) control indicated that the LoxP sites were inserted into one strand of the genomic DNA. The unedited strand of the genomic DNA produced a 150 bp band (lower band in Figure 1B ) by PCR. The homozygous floxed Clpp mice exhibited only a single 213 bp band ( Figure 1B ). The mice were also genotyped for a Cre expression, which was indicated by a 293 bp band in the gel ( Figure 1B ). The homozygous floxed Clpp mice that expressed Cre (Clpp ^fl/fl; Zp3-Cre) represented the ClpP conditional knockout (cKO) mice. The wild-type (WT) control mice used were with the genotype of Clpp ^fl/fl or Clpp ^fl/+ and were from the same litter of the cKO mice.

We determined the mRNA and protein levels of Clpp in the oocytes of the cKO mice to confirm the knockout efficacy. The result of RT-qPCR demonstrated a decreased Clpp mRNA level in the Clpp cKO oocytes, and the Clpp mRNA level of the control was approximately 4 times that of the cKO mice ( Figure 1E ). Western blot analysis demonstrated a band for ClpP expression in the control oocytes, and no band was detected in the cKO oocytes ( Figures 1Ci, Cii ). The results confirmed successful knockout of ClpP in the oocytes of the Clpp^fl/fl; Zp3-Cre mouse line. Next, we evaluated the fertility of the mice. At 8-week-old, the morphology and size of the ovary were comparable between the control and the Clpp cKO mice ( Figure 1D ). The Clpp cKO mice were fertile ( Figure 1F ) but the number of pups per litter in the Clpp cKO female mice decreased with aging ( Figure 1F ). At 10 months of age, the litter size was significantly lower in the cKO mice than the controls ( Figure 1G ).

We further studied the ovarian histology of the 8-week-old Clpp cKO mice. The overall morphology of the cKO ovary was quite comparable with that of the control ( Figure 2A ). When the number of follicles at different stage of folliculogenesis in the whole ovary tissue was counted, we found a significant decrease in the number of primary follicles and a significant increase of the atretic follicles relative to that of the controls ( Figure 2B ). There was no difference in the number of follicles at other stages of folliculogenesis and corpora lutea between the two groups. In addition, the control and the Clpp cKO mice at 8-week-old produced similar number of MII oocytes, 2-cell embryos and blastocysts ( Figures 2C–E ).

We next evaluated the impact of ClpP deficiency on meiosis. The spindle of MII oocytes was visualized by α-tubulin immunostaining. The control oocytes showed well aligned spindle (green) and chromosomes (blue with DAPI), while the Clpp cKO oocytes exhibited a disrupted spindle formation and misalignment of spindle-chromosomes structure ( Figure 2F ). Quantitatively, the rate of abnormal spindles in the cKO oocytes was 64.78%, which was significantly higher than that of the control oocytes (9.36%, Figure 2G ). We counted the number of chromosomes in these MII oocytes using the chromosome spreading assay ( Figure 2H ). Significantly, more aneuploid oocytes were found in the Clpp ^fl/fl; Zp3-Cre mice than the Clpp ^fl/fl mice ( Figure 2I ; p=0.0021).

We hypothesized that Clpp deletion induced mitochondrial dysfunction and thereby causing spindle formation defect and high aneuploidy in oocytes. Therefore, we assessed the mitochondrial membrane potential, the driving force for mitochondrial ATP synthesis, by JC-1 staining. Mitochondria with high membrane potential exhibited red fluorescence, whereas those with low membrane potential exhibited green fluorescence ( Figure 3A ). The ratio of red to green signals in the Clpp ^fl/fl; Zp3-Cre oocytes was significantly lower than that of the Clpp ^fl/fl oocytes ( Figure 3D ). We also evaluated the recovery ability from ROS stress in the Clpp cKO oocytes. The MII oocytes were pre-treated with H₂O₂ for 5 minutes before measurement of their ROS level by carboxy-H2DCFDA ( Figure 3B ). Compared with the Clpp ^fl/f oocytes, the ROS level in the Clpp ^fl/fl; Zp3-Cre oocytes was significant increased (p<0.001, Figure 3E ). We visualized the distribution of mitochondria using the Mito-tracker staining ( Figure 3C ), and found that the mitochondria in the Clpp cKO oocytes exhibited a strong aggregated pattern rather than an evenly distributed pattern as in the control oocytes ( Figure 3C ). The pattern was significantly higher in the cKO oocytes than the control oocytes ( Figure 3F ). Interestingly, the ClpP deficient oocytes had more mitochondrial DNA (mtDNA) copies than the control oocytes ( Figure 3G ). Furthermore, we measured the mRNA expression of five genes (Atp5a1, Ndufv1, Cox1, Sdhb and Uqcrc2) representing complex I-V in the mitochondrial respiratory chain by RT-qPCR. The results revealed a significant decrease of the Atp5a1, Ndufv1, Sdhb mRNA expression in the cKO oocytes when compared with the controls ( Figure 3H ). Not surprisingly, the ATP synthesis level was significantly decreased in the cKO oocytes ( Figure 3I ). These results demonstrated a disrupted oxidative phosphorylation in the ClpP cKO oocytes.

High-dose of CTX treatment causes premature ovarian insufficiency (POI) in mice ovaries (37). Here, we planned to use a reduced concentration of CTX to induce a mild stress to the ovary and test the role of Clpp in protecting the oocytes upon encountering environmental stress. We pre-tested the dosage of CTX and found 75 µM CTX induction will not affect the follicle development in the control Clpp ^fl/fl mice ( Figures 4E , F ). Thus, we decided to use this dosage and injected 75 µM CTX into the Clpp ^fl/fl and the Clpp ^fl/fl; Zp3-Cre mice ( Figure 4A ). The number of follicles in the ovary of the Clpp ^fl/fl; Zp3-Cre mice was obviously decreased in histological sections when compared with the PBS treated Clpp ^fl/fl; Zp3-Cre mice or the CTX treated Clpp ^fl/fl mice ( Figure 4E ). Counting of the follicles in the whole ovary showed that the CTX treatment severely affected follicle development in the ClpP cKO mice; the numbers of primary, secondary and antral follicles were significant decreased, and the number of atretic follicles increased dramatically when compared with that of other groups ( Figure 4F ). The CTX treatment did not affect the development of the follicles of the Clpp ^fl/fl mice ( Figure 4F ).

Next, we studied the development competence of the treated oocytes. MII oocytes were collected from the treated mice 14 days later for parthenogenetic activation. The Clpp ^fl/fl; Zp3-Cre+CTX oocytes failed to form blastocyst, whereas ~97% of the control oocytes from Clpp ^fl/fl+CTX mice were able to form blastocytes ( Figures 4B, D ). Data analysis of three biological replicates showed that the CTX treatment significantly reduced the activation rate ( Figure 4C ) and blastulation rate ( Figure 4D ) of the Clpp ^fl/fl; Zp3-Cre oocytes in the parthenogenetic activation assay.

We evaluated whether the CTX-induced phenotypes were due to a further deterioration of mitochondrial functions in terms of ROS level, mitochondrial membrane potential, ATP level and mtDNA copy number. The results showed a remarkably increase in ROS level ( Figures 5A, C ) and a decrease in red/green fluorescence intensity ratio ( Figures 5B, D ) in the Clpp cKO oocytes upon treatment with 75 µM CTX when compared with the PBS treated Clpp cKO oocytes. The treatment also significantly decreased the ATP level ( Figure 5E ) and the mtDNA copy number in the Clpp cKO oocytes ( Figure 5F ), indicating that the CTX treatment not only impaired mitochondrial functions but might also decreased the number of mitochondria in the Clpp cKO oocytes.

Although both the PBS and CTX treated Clpp cKO oocytes showed aggregation of mitochondria, the CTX treatment aggravated the abnormal distribution ( Figure 5H ) and the percentage of oocytes with the abnormal mitochondrial distribution was increased significantly in the CTX treated Clpp cKO oocytes relative to the PBS treated control oocytes ( Figure 5J ). The CTX treated Clpp cKO oocytes also exhibited significantly lower mRNA expression of Ndufv1 and Sdhb when compared with the PBS treated counterparts ( Figure 5K ), suggesting that the treatment further affected mitochondrial oxidative phosphorylation in the Clpp cKO oocytes. As expected with a deterioration of mitochondrial function, the CTX treatment significantly increased the aneuploidy rate of the Clpp deficient oocytes but not the Clpp ^fl/fl oocytes ( Figures 5G, I ).