Oocytes were obtained from the proband and a normal individual undergoing in vitro fertilization (IVF) or intracytoplasmic sperm injection (ICSI). Morphological assessments of oocyte maturation, fertilization, and embryonic development were performed using time lapse imaging. The study was approved by the Medical Ethics Committee of Foshan Maternity and Child Healthcare Hospital, Guangdong Medical University (No. FSFY-MEC-2025-028). The participants provided their written informed consent to participate in this study.

Genomic DNA (gDNA) was extracted with QIAamp DNA Blood Mini Kit (QIAGEN, Hilden, Germany) from peripheral blood of proband and other family members using standard methods. Candidate variants were identified by whole-exome sequencing (WES). Briefly, DNA was fragmented and processed using the NadPrep DNA Universal Library Preparation Kit (Nanodigmbio, Nanjing, China) for whole-genome and exome library construction. Whole-exome enrichment was performed with the Exome Plus Panel v2.0 (Nanodigmbio, Nanjing, China) targeting exon and known pathogenic regions. Libraries were sequenced on the DNBSEQ-T7 platform (MGI Tech, Shenzhen, China). The sequencing data were converted to FASTQ format. Clean reads were aligned to the UCSC hg19 reference genome using BWA software, with duplicate reads removed via Picard. Variants, including SNVs and InDels, were identified using GATK HaplotypeCaller. Variants were annotated with ANNOVAR. The online software SpliceAI, dbscSNV, MaxEntScan, SPiCE, CADD splice were used to predict the influence on the splicing. Pathogenicity analysis of variants was performed according to the ACMG/AMP practice guidelines.

Sanger sequencing was used to verify the candidate variants in the proband and perform segregation analysis in the other family members. The following Sanger sequencing primer pairs were used: PATL2-c223-F/R and PATL2-c877-F/R.

Total RNA was isolated from the peripheral blood leukocytes of the proband and control individuals using TRIzol reagent (Invitrogen, MA, United States). RNA was reverse-transcribed using a TaKaRa PrimeScript reagent kit (TaKaRa, Dalian, China) according to the manufacturer’s protocol. To investigate the aberrant splicing of the variant of PATL2, we amplified PATL2 cDNA spanning exons 10 to 14 in the proband and a normal control individual (NC) using primers PATL2-E10-F and PATL2-E14-R. Then, the obtained PCR products were analyzed by gel electrophoresis on a 1% agarose gel. Individual bands were excised from the gel, eluted using the Gel Extraction Purification Kit (QIAGEN, Hilden, Germany) and the further Sanger sequencing was performed on the cDNA purified product. Quantitative RT-PCR (qPCR) was conducted using SYBR Premix Ex Taq (Takara, Dalian, China) according to the manufacturer’s protocol. GAPDH served as an endogenous control. The 2^−ΔΔCt method was used to analyze the relative gene expression. All sequencing and qPCR primers used in this study are listed in Supplementary Table S1.

The proband, a 30-year-old female, and her 28-year-old husband had been diagnosed with primary infertility for 1 year. Both had normal karyotypes. Following two unsuccessful intrauterine insemination (IUI) cycles, the couple underwent two consecutive IVF/ICSI cycles, neither of which yielded viable embryos. In these two cycles, we used a time-lapse embryo incubator for cultivation and imaging, with images captured every 15 min for 7 consecutive days of continuous observation. In the first cycle, 16 oocytes were retrieved. Then IVF fertilization was conducted, but total fertilization failure occurred. After 16-18 h in vitro, 1 germinal vesicle (GV) stage oocyte, 2 metaphase I (MI) stage oocytes, and 13 metaphase II (MII) oocytes were observed. Strikingly, all MII oocytes displayed morphological abnormalities, notably enlarged or double polar bodies (Figure 1A). Subsequently, Late rescue ICSI was conducted on 8 MII oocytes, but no pronuclei formation was observed, 1 oocyte showed abnormal cleavage. Eventually, no embryo was obtained, and the cycle was canceled. In the second cycle, 18 oocytes were retrieved, 8 at the GV stage, 4 at the MI stage, and 6 at the MII stage. Similar to the first cycle, all MII oocytes again exhibited polar body defects (Figure 1A). ICSI was performed on 6 MII oocytes, but no pronuclei formation was observed on D1. 4 oocytes showed abnormal cleavage on D2 and D3 (Figure 1C). Eventually, no embryos were obtained, and the cycle was canceled. Notably, 6 oocytes were arrested at the GV stage, 4 oocytes were arrested at the MI stage after 65 h in vitro (Figure 1B). The proband has five sisters and two brothers range from 20 to 40 years old, all of whom exhibit normal fertility without reproductive complications.

WES was performed to elucidate the genetic basis of infertility in the proband (individual II-4). Bioinformatic analysis identified two heterozygous variants in the PATL2 gene (NM_001387263.1): a splicing variant c.223-14_223-2del (p. R75Vfs*21) and a missense variant c.877G>T (p.D293Y). These variants may be responsible for the phenotypes of the proband. No pathogenic variants were detected in other known female infertility or oocyte development-associated genes. The c.223-14_223-2del variant was classified as pathogenic based on prior literature and database annotations (Chen et al., 2017; Huang et al., 2018). While the c.877G>T variant was initially classified as a VUS according to ACMG/AMP guidelines (PM2_Supporting, PM3_Supporting, PP3), the supporting evidence is as follows: 1. The c.877G>T variant has an allele frequency of 0% in population databases and has been undetected in all major control cohorts (e.g., gnomAD, ExAC, 1,000 Genomes, NHLBI ESP) (PM2_Supporting); 2. In a female infertility patient, it cannot be determined whether this variant and another variant c.223-14_223-2del (a known pathogenic variant) are in cis or trans configuration (PM3_Supporting) (Wu et al., 2019); 3. Multiple in silico prediction tools (e.g., SpliceAI, dbscSNV, MaxEntScan, SPiCE, CADD splice) suggest that this variant may disrupt mRNA splicing (PP3).

Then, Sanger sequencing was used to verify the candidate variants in the proband and to perform segregation analysis among other family members (Figure 2A). The Sanger sequencing results indicated that the proband’s two sisters (II-3, II-6) exhibited wild-type genotypes (WT/WT) at both loci, two other sisters (II-1, II-5) were heterozygous carriers of the c.223-14_223-2del variant, the youngest sister (II-7) was heterozygous carriers of the c.877G>T variant (Figure 2B). All the sisters of proband retained fertility, consistent with an autosomal recessive inheritance pattern and supporting co-segregation of the compound heterozygous variants with the infertility phenotype. Following validation, the variant c.877G>T lies in trans to the pathogenic variant c.223-14_223-2del, meeting PM3 criteria for pathogenic evidence. Combining segregation data from affected and unaffected individuals yielded a LOD score of 0.62, meeting PP1 criteria for pathogenic evidence (Oza et al., 2018).

The missense variant c.877G>T was located on the first base of exon 12 in the PATL2 gene (NM_001387263.1), adjacent to the splicing acceptor site of intron 11-a region critical for splicing fidelity (Figure 3A). Multiple in silico prediction tools (e.g., SpliceAI, dbscSNV, MaxEntScan, SPiCE, CADD splice) suggest that this variant may disrupt mRNA splicing, SpliceAI suggested the score of acceptor loss was 0.43 (threshold ≥ 0.2), indicated that the variant (c.877G>T) was likely to change the acceptor spot for mRNA splicing. To assess the impact of the variant on PATL2 transcript processing, total RNA was isolated from the proband and normal control individual. The RT-PCR amplification of PATL2 transcripts spanning exons 10-14 revealed two distinct bands in the lane of proband (bands A and B), contrasting with a single product (band A) observed in the lane of control sample (Figure 3B). Sanger sequencing of gel-purified products demonstrated that band A corresponded to the canonical splicing pattern (exon11-exon12-exon13), but band B was an aberrant splicing band which skipped complete exon12 (exon11-exon13) (Figure 3C). Exon12 contain 54 base pair. So, this splicing defect led to 18-amino acid in-frame deletion (p.293_310del) in N-terminus of PAT1 domain of PATL2 (Figure 3A). We subsequently used qPCR to investigate the abundance of aberrant splicing variants at the mRNA level. Primers targeting exon7-exon9 (shared by both alleles) were used to evaluate the total expression of PATL2, while mutant-specific primers (forward: upstream of the deletion region [exon 9]; reverse: within the deletion region [exon12]) specifically amplified the other allele. The results showed that the total PATL2 mRNA level in the proband was slightly increased compared with normal controls, but the difference was not statistically significant. The mRNA level of the exon 9-exon12 region in the proband was approximately half that of normal control (Figure 3D).These findings indicated that aberrant splicing may not result in degradation of PATL2 mRNA. Collectively, RNA assays demonstrated that the c.877G>T variant results in truncation of <10% of the full-length PATL2 protein, meeting the criteria for PVS1_Moderate according to the ACMG/AMP guidelines.

To illustrate the structural location of the deleted amino acids within the protein, we employed AlphaFold3 (Abramson et al., 2024) for structural modeling of wild-type and mutant PATL2 proteins, which revealed that this variant caused a deletion of an alpha-helix in the PAT1 domain. Further analysis demonstrated that a critical salt bridge interaction between Lys310 and Glu350 in the wild-type protein was disrupted by the Asp293-Lys310 deletion (Figure 3E). While these protein structure predictions have not been experimentally validated and serve only as supporting evidence, they can provide valuable insights for subsequent research.