Infertility issues are faced by 15% of couples worldwide, and almost half are estimated to have a genetic background. At present, >200 casual genes have been identified for female infertility, and genes related to oocyte maturation defect (OOMD) such as ZP1 (OOMD1), TUBB8 (OOMD2), ZP3 (OOMD3), PATL2 (OOMD4), WEE2 (OOMD5), ZP2 (OOMD6), PANX1 (OOMD7), BTG4 (OOMD8), TRIP13 (OOMD9), REC114 (OOMD10), ASTL (OOMD11), FBXO43 (OOMD12) and TBPL2 have been discovered by high-throughput sequencing (Chen et al., 2017; Sang et al., 2018; Wang et al., 2020; Zhang et al., 2020; Zheng et al., 2020; Cao et al., 2021; Sang et al., 2021; Yang et al., 2021; Capalbo et al., 2022; Loeuillet et al., 2022; Maddirevula et al., 2022). Topoisomerase II homologue 2 (PATL2) is a major pathogenic gene that is associated with germinal vesicle (GV) arrest and oocyte cleavage defects. Since the first case of homozygous mutation of PATL2 gene in a patient who suffered from assisted reproductive technology (ART) failure because of oocytes being arrested at the GV stage (Chen et al., 2017) was reported, subsequent genetic testing was largely carried out and obtained important results in patients with phenotypes of recurrent failure in in vitro fertilisation (IVF) and intracytoplasmic sperm injection (ICSI), empty follicle syndrome (EFS), high proportion of GV-arrested oocytes, extremely low fertilisation rate, abnormal oocyte morphology and abnormally large first polar body. However, OOMD is a rare disease, and scarce variations have been found in suspected patients (Maddirevula et al., 2017; Christou-Kent et al., 2018; Huang et al., 2018; Wu et al., 2019; Liu et al., 2020; Wang et al., 2021; Huang et al., 2022; Huo et al., 2022; Lei et al., 2022; Sun et al., 2022); therefore, limited mutants are available for molecular epidemiological analysis. Furthermore, the phenotype caused by PATL2 gene mutation was found to be insidious among the population and even in infertile women who underwent ART. Men carrying homozygous mutants of PATL2 did not exhibit fertility issues (Huang et al., 2018). Therefore, conducting an epidemiological analysis based on the outcome data and analysing the mutant distribution and carrier frequency of PATL2 using a population-based single nucleotide polymorphism (SNP) database was necessary.

Previous studies have shown that PATL2 is required in oocyte maturation in model organisms. PATL2, as a translation repressor, anchors to the mRNA binding protein (mRNP) granule and is expressed in a spatio-temporal pattern. PATL2 is expressed within a short time when metaphase I-arrested oocytes (referred to as GV oocytes) resume meiosis. Oocyte maturation is characterised by GV breakdown (GVBD) and first polar body extrusion. PATL2 expression gradually decreases with development and is almost undetectable after the extrusion of the first polar body. Subsequently, the function of the translation repressor in the embryo may be compensated by other proteins. Oocytes carrying PATL2 mutants have shown a spectrum of phenotypes, including GV arrest, fertilisation difficulty or early embryo arrest. Some mutants, such as p.Ile318Thr, have displayed considerably less disruptive protein function than some other mutants, such as p.Arg262* and p.Tyr186*, which have led to the complete loss of function [reviewed by Fei and Zhou (2022)]. Therefore, the aim of the study was to explore the database for prospective analysis to provide data for further mutation-phenotype correlation studies in the future.

When a different amino acid being incorporated into the polypeptide, it may have a wide range of effects on protein function. Mutations can lead to changes in the structure of an encoded protein. Changes of protein structure can be used for prediction of pathogenicity. Actually the elucidation of the three-dimensional structure of proteins is one of the biggest challenges in biology. AlphaFold is an ideal online prediction tool for protein structure changes, recently developed based on artificial intelligence. Over 1,700,000 resolved proteins were used for training. Three-dimensional shape of proteins can be accurately predicted with amino acid sequences, using deep learning algorithms combined with tension control algorithms.

Populations in different regions have experienced a long history of integration and show a certain genetic distribution. With the development of high-throughput sequencing, an increasing number of human samples have been sequenced, and various infertility-related genes and loss-of-function mutations have been identified and confirmed (Biswas et al., 2021; Houston et al., 2021; Jiao et al., 2021). Particularly, when infertile patients with recurrent IVF/ICSI failure or with a family history sought help from genetic testing, pathogenic genes and loci quickly accumulated, together with a large number of loci of possibly pathogenic or unknown clinical significance, waiting to reveal their potential functions and effects. A prospective assessment based on public databases and analyses is necessary.

The carrying rate of suspected pathogenic alleles of PATL2 in the global population has not been estimated. Based on a public database of more than 300,000 whole genomes and whole exome sequencing, we predicted for the first time that the overall carrier frequency of suspected pathogenic mutation sites is at least 1.14‰, which is 1/877. This study summarised 53 mutations in PATL2, including 31 mutations from reported independent individuals, six nonsense mutations and 16 predicted pathogenic mutations identified from the public database, which expanded the spectrum of genetic causes of oocyte maturation defects. Although 22 loci from the public database was not reported in the infertile population, they might be found in ART patients undergoing genetic testing as the sample size increases. Subsequent studies are recommended to confirm whether these mutations have true influence on oocyte maturation. It is believed that this carrier frequency will continue to be rewritten when additional sequencing results are shared, additional clinical cases are tested and confirmatory experiments are performed. Identifying more casual mutants will help establish the complete mapping of genetic factors in human infertility.

Analysing the distribution of these mutations showed that the highest frequency of mutations detected in the independent patients was c.223-14_223-2del13. The carrier frequency in the population was 0.25‰, which may be a potential threat to fertility. Exons with additional mutation sites and cases were Exon13 (6 sites, 9 patients). This study carried out a molecular epidemiological analysis based on the updated definition of the PAT1 domain of the PATL2 protein within p.252-491. Analysis of 31 PATL2 compound heterozygous and homozygous mutation sites in patients with oocyte maturation defects showed that 70.91% of mutations were located in the PAT1 domain, indicating the important function of the RNA-binding domain. Identifying the regulatory RNA that can bind to this domain is an important future research goal.

We predicted the entire secondary and tertiary structures of the PATL2 protein. To the best of our knowledge, we mapped for the first time 23 helical regions on the secondary and tertiary structures of the PATL2 protein. A percentage of 65.71% of the mutation sites detected in the reported patients were in the helical region, and the rest were in the non-helical region. A percentage of 81.82% of the SNP identified in the public database were located in the helical region and 18.18% were located in the non-helical region, indicating the importance of changes in amino acid in helical regions. According to the polarity of the side chain, amino acids can be divided into four categories: polar with a positive charge, polar with a negative charge, polar without a charge and non-polar. Each amino acid has a specific R side chain, which determines its physical and chemical properties. The solubility of non-polar to polar amino acids in water increases. The mutants causing non-polar to polar changes with negative charges include two sites (p.L189R and p.P416H) and the ones causing changes from non-polar hydrophobic to polar uncharged include two sites (p.I318T and p.P510T). The amino acids in the polar uncharged (hydrophilic) side chain contain neutral polar residues; therefore, when the polar uncharged amino acids are between hydrophobic amino acids and charged amino acids, they can form hydrogen bonds with water molecules to change the protein structure. Contrastingly, the hydrogen bond breaks when polar and uncharged amino acids become non-polar amino acids (such as p.T425P and p.T449A). Additionally, proline is an imino acid with a stable angle; therefore, the mutation of 425T to proline will greatly change the structure of the protein itself. Sites that change from polar uncharged to polar negative charge (e.g., p.G370R), from polar negative charge to polar uncharged (e.g., p.R279C, p.R280Q, p.R307Q and p.R376Q) and from polar positive charged to polar uncharged (p.D293Y) were found. Generally speaking, if wild-type and mutated amino acid do not belong to the same category, the protein structure is unstable.

The analysis of 944 oocytes reported in independent individuals diagnosed with oocyte maturation defects in ART cycles showed a phenotype spectrum: 53.81% of the oocytes were arrested at the GV stage, 9.22% at the MI stage, 15.36% at the first polar body stage and 21.61% exhibited atresia and had empty follicles. Even those carrying the same compound heterozygous or homozygous mutations at the same locus showed different phenotypes (Wu et al., 2019). The expression of PATL2 in granulosa cells of polycystic ovary patients was reported to have significantly increased (Peng et al., 2021). Although studies on the factors that influence PATL2 are rare, PATL2 mutants may be sensitive to other factors, such as the follicular microenvironment and external environment, at the early stage of meiotic recovery, resulting in differences in phenotypic profiles. ICSI oocytes failed to establish pregnancy, even though they were fertilised and cleaved. Thus, even if development reaches the first polar body stage, ICSI may not alter embryo fate. Although PATL2 expression is almost undetectable after extrusion of the first polar body in normal oocytes, early dysfunction has a subsequent effect on oocyte outcome. Recently, the results of the study by Hatirnaz and team provided the hope of pregnancy for patients with oocyte maturation abnormalities (Hatirnaz et al., 2022). According to the Hatirnaz and Dahan Classification System, the ongoing pregnancy in women with type II (MI arrest) and type V (GV, MI and MII mixed arrest) oocyte seemed to be achieved by dual stimulation in in vitro maturation (Duastim IVM) in the Hartinaz clinic; this did not include patients with Type I (GV arrest), Type III (MII arrest) and Type IV (GV-MI arrest) oocyte abnormalities. One patient had two failed IVF attempts because of GV arrest by PANX1 mutation and was treated with letrozole priming IVM, and one fertilised embryo was obtained, giving the PANX1 mutation oocyte a living potential. Therefore, this led to an important step forward in therapy (Hatirnaz et al., 2022). But at present the patients with pathogenic mutations are still suggested to stop the recurrent self oocyte attempts and refer to a oocyte donation strategy. For those with only one heterozygous mutation, it should also be noted, since biallelic gene effect may also reduce the fertility capacity largely.

Mouse is a good model for functional analysis when a novel mutation is found. But the phenotypes are not always same. It was found that Patl2^−/− female mice exhibited influence on fertility not as severe as that in human, the Patl2 −/− oocytes could fertilise and develop to blastocyst stage. But the absence of Patl2 affected oocyte growth which showed abnormal small size of the GV oocytes and MII oocytes. Abnormal spindle morphology, misalignment of chromosomes on the spindle and numerous cytoplasmic asters were also notably found in the arrested oocytes (Christou-Kent et al., 2018). There may exist the other repair mechanism to overcome the barrier in oocyte development in mice.

Recently, it was reported that an induced pluripotent stem cell (iPSC) model was constructed from CD34⁺ cells in the peripheral blood of a 28-year-old patient with PATL2 compound heterozygous mutation, and was reprogrammed (Xue et al., 2022). iPSCs provide the possibility for in-depth research on their mechanisms and subsequent targeted therapies. Additionally, the PATL2^Y217N mutant was transfected into 293T cells to explore the possible mechanism by which PATL2 mutations cause oocyte maturation defects. The mutant seemed to be more stable than PATL2^WT, resulting in abnormal accumulation of the PATL2 protein. It is worth noting that mouse oocytes microinjected with PATL2^Y217N cRNA showed decreased ubiquitination levels together with abnormally large polar bodies similar to those in patients. This phenomenon is similar to the phenotype observed in mouse MOS −/− oocytes, with anomalous polar bodies, partial symmetrical cleavage and abnormal spindles (Cao et al., 2021). Spatiotemporal and specific expression of PATL2 at the meiosis initiation stage and proper functioning of the downstream gene inhibition at an appropriate time (after the first polar body is extruded) are particularly important. Too little or too much untimely expression or accumulation has a severe influence on oocyte maturation. The interference of the above mutations in the process of early maturation allows us to consider the method to choose the appropriate time to correct the defect when designing a targeted therapy. Therefore, the choice of time, mode and dose may become a bottleneck for successful treatment and intervention.

In summary, bioinformatics analysis of PATL2 gene mutations will help reveal the molecular epidemiological characteristics, help identify the prevalence of PATL2 gene mutations and provide important references for new mutation assessment, function prediction, phenotype prediction, disease diagnosis, genetic counselling, drug research and incubation and formulation of intervention strategies. Furthermore, it provides a reference for a reasonable sample size involved in epidemiological studies of PATL2 gene mutations. Using the existing database for prediction analysis can reduce the number of discovery and confirmation cycles for novel loci. Patient-derived mutated animal models are recommended to be used for establishing whether these features also reappear in the model animal and for exploring the mechanisms underlying the phenotypes involved in structural defects and infertility.