A new cohort of 465 infertile female patients with oocyte maturation arrest, fertilization failure, and early embryonic arrest were recruited from the Suzhou Municipal Hospital, the Reproductive Medicine Department of the Third Affiliated Hospital of Zhengzhou University, and the Ninth Hospital affiliated with Shanghai Jiao Tong University and were used in this study. The patient recruit criteria were applied as follows: (1) female patients are younger than 45 years old, failing to conceive after 1 year (or longer) of regular unprotected sex, (2) had undergone ≥2 failed attempts of IVF/ICSI, characterized by oocyte maturation arrest, fertilization failure, or early embryonic arrest in IVF/ICSI attempts, (3) female patients with other known causes of infertility, including male factors, chromosome anomalies, radiotherapy, or chemotherapy, were excluded. Peripheral blood samples were taken for DNA extraction. This study was approved by the Ethics Committee of the Medical College of Fudan University and the Reproductive Study Ethics Committees of the hospitals.

The details for ovarian stimulation, oocyte retrieval, IVF/ICSI procedures, and embryo culture have been described elsewhere (Huang et al., 2019). hMG (Anhui Fengyuan Pharmaceutical) was started on day 3 of the menstrual cycle. Urinary hCG (Lizhu Pharmaceutical Trading Co., China) was administered at a dose of 5,000 IU when two or more follicles measured 18 mm or more. Serum FSH, LH, E₂, and P levels were analyzed on spontaneous menstrual cycle day 3 (MC₃) and the day of hCG triggering. Oocytes retrieval was scheduled at 34–36 h after trigger and IVF or ICSI was performed 4–5 h after oocytes retrieval.

For IVF, collected oocytes were incubated in human tubal fluid (HTF; Irvine Scientific, United States), supplemented with 10% serum substitute supplement (SSS; Irvine Scientific, United States), and 300,000 progressively motile spermatozoa and left overnight; For ICSI, denudated oocytes were injected with a single mechanically immobilized sperm and directly thereafter cultured in fertilization medium (HTF + 10% SSS). All embryos were maintained in Continuous Single Culture of HTF (Irvine Scientific, United States) and were incubated under oil at 37°C and a 5% O₂ and 6% CO₂ humidified incubators with 30 μL of culture media drop throughout the entire duration of in vitro culture.

Genomic DNA was extracted from peripheral blood using the QIAamp DNA Blood Mini Kit (Qiagen). Whole-exome capture was performed using the SeqCap EZ Exome Kit (Roche), and sequencing was performed on the Illumina NovaSeq 6000 platform (Illumina). Sequencing analysis was compared with the human reference sequence (NCBI Genome build GRCh37). Mutations were annotated with GRCh37 and the dbSNP (version 138) and genome Aggregation Database (gnomAD) databases along with our in-house exome database, and functional prediction was performed with the SIFT and MutationTaster programs. For patient from consanguineous family, homozygosity mapping was performed with Homozygosity Mapper (Brazas et al., 2009). All candidate mutations were screened with the following criteria: (1) homozygous mutations with a minor allele frequency of less than 0.1% in gnomAD database and located within homozygous regions greater than 2.0 Mb, (2) predicted to be damaging or disease causing by SIFT or MutationTaster, (3) no homozygous or compound-heterozygous mutations in our control sample, and (4) inherited from both the father and mother if parents’ blood samples were available. The candidate genes were confirmed by Sanger sequencing of the affected probands as well as their parents and siblings.

The sequence of the compound heterozygous mutation (c.1155G > C and c.330 + 1G > A) was amplified from the genomic DNA of the proband (II-1) in family 3 and cloned into the pClone007 vector according to the manufacturer’s protocol (Tsingke).

The full-length coding sequence of wild-type human CDC20 (GenBank: NM_001255.3) was amplified from human metaphase I (MI) oocyte cDNA and cloned into the pCMV6-empty vector. Point mutagenesis was performed with the KOD-Plus Mutagenesis Kit (Toyobo) according to the manufacturer’s protocol. The FLAG tag was inserted into the wild type and mutant vectors at the C-terminus for the localization experiment.

Chinese hamster ovary (CHO) cells were cultured in Dulbecco’s modified Eagle medium supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin (Gibco) and maintained in a humidified incubator at 37°C with 5% CO₂. Wild type and mutant plasmids of CDC20 were transfected into CHO cells using the PolyJet In Vitro DNA Transfection Reagent (SignaGen) according to the manufacturer’s instructions.

The CHO cells were seeded in a six-well plate 1 day prior to transfection and then transfected with 1 μg of the plasmids. Cells lysates were acquired in RIPA lysis buffer (Shanghai Wei AO Biological Technology) with 1% protease inhibitor cocktail (Bimake). The protein concentration was measured by the bicinchoninic acid method, and equal amounts of protein were resolved on 10% SDS-PAGE gels and transferred to nitrocellulose filter membranes. Membranes were blocked in 5% non-fat milk for 1 h and incubated overnight at 4°C with primary antibodies. Antibodies against CDC20 (14866, CST) and cyclin B1 (55004-1-AP, Proteintech) were used at 1:1,000 dilution, and anti-vinculin (1:3,000 dilution, 13901, CST) was used as the internal control. Goat anti-rabbit IgG (1:5,000 dilution, M21001, Abmart) was used as the secondary antibody to detect the primary antibodies. The blots were finally captured using ECL Western Blotting Substrate (Tanon) after incubation with the secondary antibodies (1:5,000 dilution, Abmart).

The c.330 + 1G > A mutation was located at the donor splice-site of intron 2. The full sequence of exon 1, intron 1, exon 2, intron 2, exon 3, and intron 3 was amplified from the genomic DNA of the proband (II-I) in family 3, and the PCR product contained both the wild-type allele and the c.330 + 1G > A mutant allele due to the heterozygous status of this mutation. The PCR product was cloned into the minigene vector, and the wild type and c.330 + 1G > A mutant plasmids were transfected into HeLa cells. After incubation for 36 h, total RNA was extracted using the RNeasy Mini Kit (Qiagen), and cDNA was obtained with the PrimeScript RT reagent kit (Takara).

Germinal vesicle (GV) oocytes were collected from the ovaries of 7–8-week-old ICR mice. For long-term incubation, oocytes were cultured in Minimum Essential Medium (Gibco) with 10% fetal bovine serum and 1% penicillin/streptomycin (Gibco) in a humidified incubator at 37°C with 5% CO₂. GV breakdown was prevented with 2.5 μM milrinone (Sigma). For the localization experiment, wild type and mutant FLAG-tagged CDC20 cRNAs were microinjected into the cytoplasm of each mouse GV oocyte. The injected GV oocytes were released into milrinone-free medium and subsequently fixed for immunofluorescence at 7 h after release to collect the MI oocytes. For the small interference RNA (siRNA) experiment, Cdc20 5’UTR siRNA (siCdc20, 5′-UGU UCG GGA GAG CUG AGU ATT-3′, 40 μM) was injected into GV oocytes and incubated for 24 h to knock down the endogenous Cdc20 level, which led to MI arrest. For the rescue experiment, wild-type CDC20 cRNA (750 ng/μL) and mutant CDC20 cRNA (750 ng/μL) were microinjected into the cytoplasm of each mouse GV oocyte at 6 h after siRNA injection. The first polar body (PB1) extrusion rate was recorded at 12 h after the oocytes were released into milrinone-free medium.

The mouse oocytes were fixed and stained as previously described (Wu et al., 2018). Anti-FLAG antibody (1:500 dilution, Sigma-Aldrich) was used for determining the CDC20 localization. The kinetochores were stained with an anti-Crest antibody (1:500 dilution, Immunovision), and Hoechst 33342 (1:700 dilution, BD) was used to label the DNA. Oocytes were mounted on glass slides, and images were captured on a confocal laser-scanning microscope (Leica).

All three probands had been diagnosed with primary infertility of unknown cause for several years despite having normal menstrual cycles, and their male partners all had normal sperm counts and normal sperm morphology and motility. The detailed clinical information of ovarian stimulation characteristics is shown in Supplementary Table 1. The proband in family 1 was 33 years old and came from a consanguineous family (Figure 1A). She had undergone two failed IVF attempts during her 3 years of infertility, and all seven retrieved oocytes were arrested at the MI stage (Figure 1B). The proband in family 1 thus showed the phenotype of oocyte maturation arrest. The proband in family 2 was 37 years old and had been diagnosed with primary infertility for 11 years. She had undergone one failed IVF attempt and two failed ICSI attempts. In her IVF attempt, two oocytes were retrieved and both were arrested at the MI stage (Figure 1B). In her ICSI attempts, fourteen oocytes were retrieved, nine were arrested at the GV or MI stage, and five were maturated to the metaphase II (MII) stage. Only one MII oocyte was fertilized, but was arrested at the eight-cell stage. The proband in family 3 was 32 years old and had undergone one failed IVF attempt and one failed ICSI attempt. In her IVF attempt, five out of eight oocytes were MII oocytes. Only two of them were fertilized, but were they arrested at the one-cell stage. During her ICSI attempt, twelve out of thirteen oocytes were arrested at the MI stage, only one was matured failed to be fertilized. In brief, both the two probands in families 2 and 3 showed a mixed phenotype of maturation arrest and fertilization failure. The family pedigrees are shown in Figure 1A, and the clinical characteristics of the retrieved oocytes are summarized in Table 1.

Mutational screening of CDC20 (GenBank: NM_001255.3) was performed in our whole exome-sequenced cohort of patients with abnormalities in oocyte maturation, fertilization, and early embryonic development. We identified four affected individuals from three independent families with homozygous and compound heterozygous mutations in CDC20 (Figure 1A). The proband (II-1) of family 1 came from a consanguineous family, and had the homozygous missense mutation c.887G > A (p. Arg296Gln) in CDC20. CDC20 localized within the homozygous region (4.22M) of the proband in family 1 (Figure 1C), and the parents were both heterozygous carriers (Figure 1A). The functional impact of this mutation was assessed to be damaging by SIFT and disease causing by Mutation Taster (Table 2). The proband (II-1) of family 2 had a compound heterozygous CDC20 mutation consisting of the missense mutation c.965G > A (p.Arg322Gln) and the nonsense mutation c.964C > T (p.Arg322^∗). The parents were both heterozygous carriers, and the infertile sibling (II-2) carried the same compound heterozygous mutation as the proband (II-1) (Figure 1A). The proband (II-1) in family 3 carried a compound heterozygous mutation consisting of the missense mutation c.1155G > C (p.Trp385Cys) and the splicing mutation c.330 + 1G > A. Peripheral blood samples from the parents in family 3 were not available. To determine whether the compound mutations were inherited from both her father and mother separately, TA clone was constructed and Sanger sequencing was performed. As indicated in Supplementary Figure 1, the two mutations were located on different alleles, which confirms a recessive inheritance pattern. Information about the mutations is shown in Table 2, and the positions of the corresponding mutations and their conservation analysis in different species are shown in Figure 2.

The effect of the c.330 + 1G > A splicing mutation was studied with a minigene assay. Agarose gel electrophoresis showed a different-size band for the c.330 + 1G > A compared with the wild type, and the larger band indicated the abnormal alternative splicing isoform resulting from this mutation (Figure 3A). The cDNA sequence analysis of the alternative splicing isoform is shown in Figure 3B. The c.330 + 1G > A splicing mutation caused the canonical donor splice site missing, which resulted in the retention of intron 2 and the premature termination of CDC20 protein. This indicated that the splicing mutation c.330 + 1G > A affected the normal splicing of CDC20 and resulted in an abnormal transcript.

To assess the impact of the missense mutations (p.Arg296Gln, p.Arg322Gln, and p.Trp385Cys) and the nonsense mutation (p.Arg322^∗) on the function of CDC20, we first detected the CDC20 protein level in CHO cells. Compared with wild-type CDC20, the missense mutations p.Arg296Gln and p.Trp385Cys resulted in a reduction in the CDC20 protein level, while the nonsense mutation p.Arg322^∗ resulted in a truncated protein (Figures 3C,D). Although the missense mutation p.Arg322Gln had no obvious effect on the CDC20 protein level in CHO cells, it was previously reported that the p.Arg322Gln mutation caused a reduction in the CDC20 protein level in the patient’s lymphoblastoid cell line in vivo (Zhao et al., 2020b).

During the metaphase to anaphase transition, Cdc20 binds to and activates the ubiquitin ligase activity of APC/C and enables the ubiquitination and degradation of cyclin B1, thus promoting the onset of anaphase (Nasmyth and Haering, 2005). We therefore further assessed the effect of the CDC20 mutations on cyclin B1 degradation. As shown in Figures 3C,D, overexpression of wild-type CDC20 significantly decreased the endogenous protein level of cyclin B1, while most of mutants affected the degradation of cyclin B1.

It has been reported that Cdc20 localizes to the kinetochore in MI mouse oocytes (Tsurumi et al., 2004). To determine the effect of the mutations on human CDC20 localization, we expressed FLAG-tagged wild type and mutant CDC20 to monitor its subcellular localization in mouse oocytes. Wild-type CDC20 showed a strong kinetochore signal (Figure 4A), which was consistent with previous studies. For the two missense mutations (p.Arg296Gln and p.Trp385Cys), CDC20 also showed normal kinetochore localization similar to wild type. In contrast, CDC20 with the nonsense mutation (p.Arg322^∗) failed to localize to the kinetochore (Figure 4A), indicating the functional impairment of CDC20 protein.

To further explore the effect of the mutations on CDC20 function in mouse oocytes, we performed a rescue experiment as described in our previous studies (Reis et al., 2007). The MI arrest phenotype in mouse oocyte was constructed by injecting Cdc20 siRNA, and the PB1 extrusion could be rescued by supplementation with human wild-type CDC20 cRNA (Figure 4B). Compared with wild-type, the homozygous or compound heterozygous mutations significantly reduced the ability of CDC20 to rescue the PB1 extrusion (Figure 4B). To assessed the further reason of reduced ability, we detected the CDC20 protein level of wild-type and mutations in mouse oocytes. Compared with wild-type CDC20, the missense mutation p.Arg296Gln, p.Trp385Cys and p.Arg322Gln resulted in a reduction on CDC20 protein level. The mutation corresponding to the splicing mutant (c.330 + 1G > A) p.Glu111Ilefs^∗36 resulted in truncated protein, and the nonsense mutation p.Arg322^∗ showed no detectable signal which indicates the severe impairment on protein level (Supplementary Figure 2). To further assess the effect of each individual mutant independently, the wild-type and mutant cRNA were injected separately into the MI arrest mouse oocyte which caused by injecting Cdc20 siRNA. Compared with wild-type CDC20, all individual mutants significantly reduced the ability of CDC20 to rescue the PB1 extrusion, which indicates that each individual mutant affected the normal function of CDC20 (Supplementary Figure 3). All these results demonstrated that all the identified mutations affected the normal function of CDC20.