A compound heterozygous mutation in ZP1 and two novel heterozygous cis mutations in ZP3 causes infertility in women presenting with empty follicle syndrome

Abstract

Background

Infertility, as a major human reproductive health problem, affects approximately 17.5% of the global population. The emergence of assisted reproductive technology, particularly in vitro fertilization (IVF), has resolved the reproductive challenges of most infertile couples. In IVF, there exists a group of women who are unable to retrieve oocytes from typically developing follicles despite a positive response to ovarian stimulation, a condition clinically referred to as empty follicle syndrome (EFS). The pathogenesis of EFS is still unclear, but increasing evidence suggests that genetic factors play a very significant role. A clinical and experimental analysis of two pedigrees was performed in this study to investigate the genetic basis of EFS.

Methods

Genomic DNA was extracted from peripheral blood samples for whole-exome sequencing in EFS patients. The identified variants were validated by Sanger sequencing. Computer simulation of protein structure was used to assess the pathogenicity of the variants on the proteins. The effects of the ZP1 and ZP3 variants on protein expression were probed via western blotting, and the effects of the variants on protein localization were probed by immunofluorescence. ELISA and Co-IP were used to detect the effects of variants on protein secretion and interactions.

Results

In this study, we identified a compound heterozygous mutation in ZP1 (c.[2T > A]; [1429G > T]) and two novel ZP3 heterozygous cis mutations (c.[724G > T;815 A > G]) from two EFS patients, respectively. These mutations are highly conserved between different species. Through in vitro experiments, we showed that the ZP1 (p.[Met1?]; [Gly477*]) mutations result in reduced protein expression, whereas the ZP3 (p.[Asp242Tyr; Asn272Ser]) mutations lead to increased protein expression. However, neither mutation affected the subcellular localization of the ZP proteins. Bioinformatic analysis revealed that these mutations disrupt the conformation of the ZP protein, which may affect its stability and binding capability. Functional experiments showed that the ZP1 and ZP3 mutations altered the interaction between themselves and ZP2 proteins; the ZP1 mutation inhibited ZP1 protein secretion, whereas the ZP3 mutation increased the secretion of ZP3 protein, which may affect ZP assembly.

Conclusions

Our study has enriched the mutational spectrum of the ZP gene by identifying mutations in the causative genes ZP1 and ZP3 associated with EFS. In vitro experiments exploring the effects of mutations on ZP protein expression and function confirmed that ZP is an important genetic cause of EFS, thus broadening our understanding of the genetics of female infertility. We emphasize the importance of genetic analysis in the diagnosis and prognosis of “genuine” EFS (GEFS) and recommend that EFS patients strive for a successful pregnancy through an oocyte donation program.

Introduction

In recent years, the problem of reproductive disorders has come to the forefront with the delays in childbearing and changes in the environment and lifestyle. Infertility is a major human reproductive health problem with increasing prevalence yearly, affecting approximately 17.5% of the global population [1]. Assisted reproductive technology (ART), especially in vitro fertilization (IVF), has become the mainstay for infertile couples to solve their fertility problems. As oocytes are the core element of reproduction, the most critical steps in IVF are controlled ovulation and oocyte retrieval by puncture, with the ultimate goal of obtaining oocytes with developmental potential [2, 3]. However, there exists a subset of clinically infertile women with normal monitoring of the ovulation induction process in IVF, but the cumulus-oocyte complex (COCs) retrieved are empty mucus clusters (i.e., they do not contain oocytes), which researchers refer to as empty follicle syndrome (EFS) [4].

EFS is classified into two categories: “false” EFS (FEFS) and “genuine” EFS (GEFS) [5]. FEFS can be improved or prevented by adjusting ART medication regimens, whereas GEFS is defined as the inability to retrieve oocytes after a correct ovulation-triggering operation [6]. Patients with GEFS often experience multiple failed ARTs, which undoubtedly creates a great deal of stress for both doctors and patients. Therefore, understanding the etiology of GEFS is critical to the diagnosis, treatment, and prognosis of patients.

With the widespread use of whole-exome sequencing (WES) technology in recent years, a number of mutations have been identified in patients with disseminated and familial infertility that can lead to EFS. For example, mutations in the luteinizing hormone/chorionic gonadotropin receptor (LHCGR) gene have been found to cause GEFS by resisting luteinizing hormone (LH) stimulation [7]. Mutations in the zona pellucida glycoprotein 1 (ZP1), zona pellucida glycoprotein 2 (ZP2), and zona pellucida glycoprotein 3 (ZP3) genes affect zona pellucida (ZP) assembly and secretion, which leads to GEFS [8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30]. In mammals, ZP proteins wrap around the outside of the oocyte, which play important roles in oocyte growth and development, fertilization, and preimplantation embryo protection. During oogenesis, ZP serves as a connecting channel between the oocytes and cumulus cells, transporting a steady stream of nutrients, metabolites, and other molecules from the external environment to the growing oocytes [31]. The human ZP consists of four glycoproteins (ZP1, ZP2, ZP3, and ZP4), where ZP3 and either ZP2 or ZP4 are alternately repeated to form the ZP filament structure, ZP1 occasionally replaces ZP2 or ZP4 and stabilizes ZP by covalently cross-linking the ZP filaments [32]. Since the first ZP mutation was reported in 2014 [21], 43 ZP1 mutations [8,9,10,11,12,13,14,15,16,17,18,19,20,21,22, 30], 4 ZP2 mutations [9, 23, 24], and 8 ZP3 mutations [9, 16, 25,26,27,28,29] have been sequentially identified in GEFS patients so far (Table 1). Among them, DAI et al. reported normal preantral folliculogenesis but abnormal ZP assembly in patients with EFS by morphological evidence, which reminds us that genetic variation in ZP may be an important etiological factor in EFS [10]. Some advancements have been made in the research of EFS, but most of the previous studies have been limited to phenotypic observations without functional validation, and the correlation between changes in the expression trend of the ZP gene and EFS is still unclear. Therefore, more investigations on new pathogenic variants are needed to further clarify the etiology of EFS.

Table 1 Summarize of the ZP gene variants associated with EFS reported in the literature to date

In this study, one ZP1 compound heterozygous mutation and two novel ZP3 heterozygous cis mutations were identified from EFS patients by WES. In silico analysis and in vitro experiments revealed that the mutations led to abnormal ZP1 and ZP3 protein expression, resulting in conformational abnormalities of the ZP protein, which ultimately lead to the development of EFS. Our findings expand the genetic spectrum of EFS caused by ZP mutations and highlight the importance of genetic analysis in the diagnosis and prognosis of GEFS.

Materials and methods

Clinical samples and ethical approval

The patients in this study were enrolled in the Center for Reproductive Medicine, Affiliated Women’s and Children’s Hospital of Qingdao University. Patients from two different families were diagnosed with primary infertility by their doctors. After all the participants voluntarily signed an informed consent form to participate in the study, we collected clinical information and blood samples from their families. Our study was approved by the ethics committee of the Affiliated Women and Children’s Hospital of Qingdao University (project number QFELL-YJ-2023-120).

Genomic DNA extraction

Genomic DNA was extracted from peripheral blood using TIANamp Blood DNA Kit (Tiangen, DP348), and a NanoDrop One Microvolume UV‒Vis spectrophotometer (Thermo Fisher Scientific, 840-317500) was used to determine the DNA concentration and quality.

WES

We performed WES on probands to identify causative genes in infertile patients. After DNA was extracted from the patient’s peripheral blood samples, genomic DNA was fragmented using enzymatic interruption. The ends were flattened, and junction sequences were added to both ends of the DNA fragments, followed by polymerase chain reaction (PCR) enrichment to complete the library construction. A biotin-carrying probe is artificially designed on the basis of the gene locus, and the target fragment and the probe are directly hybridized. After hybridization was completed, the target fragment was captured via magnetic beads with streptavidin. Free DNA was washed away by thermal elution and room temperature elution, and the enriched DNA was amplified to construct a high-throughput sequencing library. The constructed libraries were upsequenced on a sequencer (MGISEQ-2000RS FluoXpert) based on the MGI2000 platform combined with the probe-anchored polymeric sequencing method. Candidate variants with the following criteria were selected: (a) present in the patient; (b) frequency in public databases (such as 1000 Genomes, gnomAD and ExAC) is less than 1%; (c) screen for known pathogenic genes.

Sanger sequencing

Sanger sequencing further validates mutations in patients and their family members. The primers in Table 2 were used to amplify the exons of the ZP1 and ZP3 genes found by WES to contain the mutation sites. Mutant regions were amplified by PCR using 2X Accurate Taq Master Mix (dye plus) II (Accurate Biology, AG11022). The product was purified and sequenced on an ABI 3730xl DNA Analyzer (Thermofisher).

Table 2 Primer sequences (5’→3’)

Bioinformatic analysis and protein structure prediction

We used PolyPhen-2 [34], Mutation Taster [35] and InterVar [36] to predict the functional effects of mutations. The 3D structures of the wild-type and mutant ZP1 and ZP3 proteins were modeled using homology modeling in SWISS-MODEL [37]. We used PyMOL 3.0 [38] software to analyze the protein structures and predict the effects of mutations on the protein.

Plasmid construction

Wild-type human ZP1, mutant ZP1^G477X, wild-type human ZP2, wild-type human ZP3, mutant ZP3^D242Y and mutant ZP3^N272S were constructed and then recombined with the eukaryotic expression vector pcDNA3.1. An HA-tag, a MYC-tag and a FLAG-tag was fused at the N-terminus of ZP1, ZP2 and ZP3, respectively. In addition, pcDNA3.1-FLAG-ZP1^wt and pcDNA3.1-FLAG-ZP1^M1? were prepared separately for analyzing the effect of start codon site mutations on ZP proteins. The FLAG-tags were located at the N-terminus, an untranslated region was inserted between the FLAG-tag and the ZP1 protein to avoid interference of the tag sequence with the translation initiation process. All vectors were constructed by Sangon Biotech and were confirmed by DNA sequencing.

Cell culture and transfection

HeLa cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM)/High Glucose (Sparkjade, CA0004) supplemented with 10% fetal bovine serum (FBS) Premium (PAN biotech, P30-3302), penicillin-streptomycin mixture (100×) (Sparkjade, CM0001), and the cells were kept at 37 °C with 5% CO2. When cells reached 60% confluence in the culture dish, cells were transiently transfected with 1000 ng of plasmid per ml of medium for 6 h using Lipofectamine™ 3000 Transfection Reagent (Thermo Fisher Scientific, L3000015). The cells were then washed twice with Phosphate-Buffered Saline (PBS) and cultured in serum-free medium for 48 h. No plasmid was added to the blank control group, and the remaining operations were consistent with the transfection group.

Real-time quantitative polymerase chain reaction (RT-qPCR)

ZP1^WT, mutant ZP1^G477X, mutant ZP1^M1?, ZP3^WT, mutant ZP3^D242Y and mutant ZP3^N272S plasmids were transfected into each well of a 6-well plate according to the cell transfection. Forty-eight hours after transfection, RNA in Hela cells was extracted strictly according to the instructions of RNA-Quick Purification Kit (Yishan Biotech, RN001), and the NanoDrop One Microvolume UV-Vis spectrophotometer (Thermo Fisher Scientific, 701-058112) was used to determine the RNA concentration. RNA was reverse transcribed into cDNA using the SPARKscript II RT Plus Kit (SparkJade, AG0304), RT-qPCR was performed using the SYBR Green Premix Pro Taq HS qPCR Kit (Accurate Biology, AG11701). The primers used are listed in Table 2.

Protein extraction and Western blot (WB) analysis

Forty-eight hours after transfection of Hela cells, cells were lysed with sodium dodecyl sulfate (SDS) lysis buffer (Beyotime, P0013G) and centrifuged at 12,000 g for 10 min to collect the total cell protein supernatant. 5 × Loading Buffer (Beyotime, P0015) was added to the extracted total proteins and heated at 97.5 °C for 10 min.

Total proteins extracted from cells were separated in 7.5% sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) gels and then transferred to polyvinylidene difluoride (PVDF) transfer membranes (Millipore, IPVH00010). Non-specific binding sites were blocked for 1 h at room temperature with 5% nonfat milk in Tris-buffered saline containing 0.05% Tween-20. Membranes were incubated overnight at 4 °C with dilutions of the following antibodies: GAPDH (1:1000, Proteintech, 10494-1-AP and 60004-1-Ig), HA (1:1000, Cell Signaling Technology, #3724), MYC (1:1000, Cell Signaling Technology, #2276), FLAG (1:1000, Sigma, F3165), ZP1 (1:10000, Santa Cruz, sc-365435) and ZP3 (1:1000, Santa Cruz, sc-398359). After incubation with HRP-conjugated Affinipure goat anti-mouse/anti-rabbit IgG(H + L) (1:5000, Proteintech, SA00001-1 and SA00001-2) for 1 h at room temperature, the immune complexes were detected by enhanced chemiluminescence (Tanon, 4600). For densitometric analyses, protein bands on the blots were measured by ImageJ software.

IF staining

We seeded Hela cells transfected with the desired plasmids for 48 h onto coverslips. 24 h later, observing that the cell density did not exceed 50%, the cells were fixed and membrane-broken, and then incubated with primary antibodies for 12 h at 4 °C. The primary antibodies used included FLAG (Sigma, F3165), ZP1 (Santa Cruz, sc-365435), and ZP3 (Proteintech, 21279-1-AP). Cells were washed with Phosphate-Buffered Saline with Tween (PBST) 3 times for 5 min each. Goat anti-rabbit IgG H&L (Alexa Fluor^® 594) and Goat anti-mouse IgG H&L (Alexa Fluor^® 488) (Abcam, ab150080 and ab150113) were used as secondary antibodies to incubate the cells for 1 h in a wet box protected from light. Nuclei were stained and blocked with mounting medium, antifading agent (with DAPI) (Solarbio, S2110) and finally visualized under a confocal microscope (Leica, SP8).

Coimmunoprecipitation (Co-IP) assays

Immunoprecipitation was performed according to the Pierce Classic Magnetic IP/Co-IP Kit (Thermo Fisher Scientific, 88805) user guide. After cells were transfected for 48 h, PBS was used to wash the cells once, and the cells were lysed by adding IP lysis/wash buffer for 30 min on ice. Adherent cells were collected with a spatula, and cell debris was settled by centrifugation at 13,000 × g for 10 min. After centrifugation the supernatant was split 100 ug for the INPUT group and the remaining 500 ug protein sample was used for Co-IP. After taking the Pierce Protein A/G magnetic beads and pre-washing them, the antigen sample/antibody mixture was added to them and incubated overnight at 4 °C with rotation. The magnetic beads were washed three times repeatedly and then incubated with the elution buffer for ten minutes at room temperature to obtain IP group samples. 5 × Loading Buffer (Beyotime, P0015) was added proportionally, heated at 97.5 °C for 10 min, and protein interactions were analyzed by WB.

Elisa

Cell culture medium was collected 48 h after plasmid transfection and centrifuged at 12,000 × g for 10 min at 4 °C to get rid of cell debris. The untransfected plasmid Hela cell culture medium was used as a blank control group to detect the effect of gene mutation on ZP1 and ZP3 secretion by using the Human Zona Pellucida Protein ELISA kit (Nanjing SenBeiJia Biological Technology, H2263 and H2304). All experiments were performed in triplicate and repeated three times.

Statistical analysis

The data in this study are presented as the means ± SEMs, and comparisons of means between groups were analyzed by one-way ANOVA. All the statistical analyses were performed using Prism 9.0.0 (GraphPad). All experiments in this study were repeated three times.

Results

Clinical characteristics

This study included two female patients from different families who had been suffering from infertility for several years. Each patient had a regular menstrual cycle and no abnormalities on ultrasound. The patient’s basal sex hormone levels were within the normal range. In addition, the couples had no significant family history of disease and no abnormalities on chromosomal karyotype analysis.

The proband of Family 1 (Family 1 II-3) was 42 years old and had not been pregnant for 4 years since her first marriage. She had unprotected sexual intercourse with her current husband for 13 years and was unable to get pregnant. The patient underwent hysterosalpingography, which suggested bilateral fallopian tubal patency. Ovulation occurred for 3 cycles of ovulation induction therapy, but no pregnancy was achieved after instructed intercourse. After a failed attempt at one cycle of artificial insemination by husband (AIH), the patient underwent her first IVF treatment. A luteal-phase long protocol was used to stimulate follicular development with rhFSH for 10 days during the course of the treatment. At the trigger day, a total of 20 follicles were observed, of which 8 follicles had diameters of ≥ 14 mm. Then, human chorionic gonadotropin (hCG) 8000 IU was injected, eggs were retrieved 36 h later, which obtained a total of 10 empty COCs, but no oocytes. In 2022, the couple underwent the second IVF treatment at our hospital. Gonadotropin (Gn) stimulation of the ovaries for nine days was performed with recombinant human follicle stimulating hormone (rhFSH) β injection 300 IU for the first seven days and rhFSH for injection in conjunction with human menopausal gonadotropin (HMG) 75 IU each for the last two day, Gn for a total dose of 2550 IU. 3 follicles ≥ 14 mm in diameter were obtained on the day of triggering, and oocyte retrieval operation was performed 36 h after a dual trigger using hCG (10000 IU) and gonadotropin-releasing hormone agonist (GnRH-a, 0.1 mg). The serum β-hCG concentration was 189.9 IU/L 12 h after trigger; however, after repeated aspiration and flushing, only three empty COCs were observed, and no oocytes were obtained.

The proband of family 2 (family 2 II-1) was 32 years old and had been married without contraception for 4 years without pregnancy. The couple underwent IVF treatment in our hospital due to incomplete bilateral fallopian tubal obstruction. In the first cycle of IVF, the total dose of Gn was 2025 IU to stimulate follicular development, i.e. rhFSH for Injection 225 IU for nine days. 5 leading follicles (≥ 14 mm) were observed on trigger day. Oocyte extraction was performed 36 h after hCG (8000 IU) triggered ovulation. The serum β-hCG concentration was 298 IU/L 12 h after trigger, but puncture of bilateral ovarian follicles and repeated rinsing showed only a few granulosa cells, and small mucus clusters were unwrapped without oocytes. The second IVF treatment was performed using a natural cycle regimen. Monitoring of follicle size by ultrasound, combined with observations of hormone levels and other indicators, indicated that the patient’s LH level failed to reach the threshold required for normal ovulation induction, and the urine LH test remained weakly positive. In order to improve the success rate of ovulation, an exogenous ovulation induction drug (hCG 8000 IU) was eventually used to compensate for the lack of endogenous LH. Only one follicle of 15.5 mm in diameter was observed on the trigger day. Repeated rinses during ovulation retrieval did not reveal any oocytes but only a few granulosa cells. Table 3 summarizes the IVF cycle protocols and the clinical characteristics of the probands. Gonadotropin stimulation and follicular response in the patients’ IVF cycles are shown in Table 4.

Table 3 IVF cycle protocols and clinical characteristics of probands from two families treated at our hospital

Table 4 The probands from two families with gonadotropin stimulation and follicle response during IVF cycles

Identification of ZP1 and ZP3 variants

To verify whether the patients’ EFS is associated with genetic mutations, we performed WES on the probands and confirmed the mutations by Sanger sequencing. We identified only mutations in the ZP1 or ZP3 genes but not in other female infertility-causing genes. The family 1 proband (Family 1 II-3) carried a compound heterozygous mutation (c.[2T > A]; [1429G > T]) in ZP1, both mutations followed autosomal recessive inheritance. The heterozygous mutation c.2T > A(p.Met1?) was inherited from her mother, and the patient’s sisters all carried the mutation. The heterozygous mutation c.1429G > T(p.Gly477*) was inherited from her father, and the patient’s brother carried this variant. Two novel ZP3 heterozygous cis mutations (c.[724G > T;815 A > G]) was identified in the Family 2 proband (Family 2 II-1), and both mutations followed autosomal dominant inheritance. The patient’s mother carried a heterozygous mutation of c.815 A > G(p.Asn272Ser), and no variation in ZP3 was found in her father, suggesting that c.724G > T(p.Asp242Tyr) was a de novo variant in the family. The patient’s family line and Sanger sequencing results are shown in Fig. 1A. Table 5 shows an overview of the ZP1 and ZP3 variants observed in the patients. Neither the ZP1 nor ZP3 mutations appeared in the 1000 Genomes Project Database and the gnomAD database.

Fig. 1

Pedigree chart of EFS patients and genetic analysis of ZP gene mutations. (A) Pedigrees of the two EFS patients and Sanger sequencing validation plots. Squares represent male family members, circles represent female family members, solids indicate affected members, and equal signs indicate infertility. The black arrows point to the probands in the families. Genotypes are shown below the family members. (B) The locations of mutations in ZP1 and ZP3 are indicated in the genomic structure (top) and functional domains (bottom). SP (gray): signal peptide; ZP-N1 (azure blue): N-terminal isolated ZP-N; TD (green): trefoil domain; ZP-N (light blue): ZP-N domain; ZP-C (navy blue): ZP-C domain; CFCS (orange): consensus furin cleavage site; TMD (brown): transmembrane-like domain; IHP (magenta): internal hydrophobic patches; EHP (red): external hydrophobic patches. (C) The ZP1 and ZP3 mutation sites are highly conserved across different species. The red color indicates mutated amino acids. * indicates the stop codon

Table 5 Overview of the ZP1 and ZP3 variants observed in the patients

The ZP1 missense mutation c.2T > A is located at the start codon of ZP1, Mutation Taster predicts that it will result in a translation initiation codon 141 bases downstream. The ZP1 nonsense mutation c.1429G > T was positioned on exon 8 of ZP1 with the glycine at position 477 replaced by a stop codon. The two novel missense mutations (c.[724G > T;815 A > G]) in ZP3 were both located in exon 5. In this case, c.724G > T caused the aspartic acid at position 242 to be replaced by tyrosine, and c.815 A > G caused the asparagine at position 272 to be replaced by serine. Most of the amino acid substitution sites identified in this study were located in the ZP domain (ZPD), which is the most conserved of all ZP proteins. The positions of the variant sites in the gene structure and functional domains of the ZP1 and ZP3 genes are shown in Fig. 1B. In addition, multiple sequence comparisons showed that the altered amino acids in diseased individuals are highly conserved across different species (Fig. 1C).

Bioinformatic analysis of variants

Bioinformatic analysis was performed using PolyPhen-2 [34], Mutation Taster [35] and InterVar [36]. The results indicated that the mutations ZP1 (c.2T > A and c.1429G > T) and ZP3 (c.724G > T) in this study were potentially pathogenic, whereas the significance of ZP3 (c.815 A > G) was uncertain (Table 5). We modeled wild-type and mutant proteins using SWISS-MODEL, and PyMOL was used to predict the effects of mutations on protein structure (Fig. 2A). Compared with wild-type ZP1, the nonsense mutation (p.Gly477*) in ZP1 disrupted the 3D structure from β-sheet to termination, with the ZPD incomplete and a deletion of the C-terminus of the protein. ASP-242 in wild-type ZP3 interacted with HIS-231, ILE-233, THR-245, and SER-249 to form hydrogen bonds, respectively. The missense mutation (p.Asp242Tyr) in ZP3 altered several hydrogen bond positions. The TYR-242 substitution clashed with neighboring residues (HIS-231, SER-249), leading to hydrogen bond breakage and interaction with ASP-266 to form a new hydrogen bond. The SER-272 substitution in the ZP3 missense mutation (p.Asn272Ser) altered the loop structure of ZP3. These mutations lead to conformational changes in the ZP protein that may affect its stability and binding ability.

Effect of ZP1 and ZP3 variants on mRNA levels

We selected samples transfected with wild-type and different mutant plasmids, extracted total RNA and performed reverse transcription to generate cDNA, RT-qPCR was carried out in order to demonstrate the effect of mutation on mRNA levels. Compared with wild-type ZP1, there was no significant difference in the expression of the ZP1 gene in the p.Gly477* group, demonstrating that the mutation generates the termination codon early but does not trigger nonsense-mediated mRNA decay (NMD). In contrast, the expression of the ZP1 gene in p.Met1? was significantly increased relative to the wild-type group, confirming increased mRNA levels (Fig. 2B). Against the ZP3 wild-type group, p.Asp242Tyr resulted in increased ZP3 gene expression and p.Asn272Ser in decreased expression (Fig. 2C).

Fig. 2

(A) Protein structure modeling of the ZP1 and ZP3 proteins. We used PyMOL software to predict the protein structures of ZP1 wild-type, ZP1(p.Gly477*), ZP3 wild-type, ZP3(p.Asp242Tyr) and ZP3(p.Asn272Ser), respectively. These 3D models are shown in cartoon form, where the cyan structure indicates the α-helix, the purple structure indicates the β-sheet and the wheat-colored structure indicates the loop structure. Amino acids are shown as sticks form in a close-up view of the ZP3 protein. Red arrows point to mutated amino acids. The yellow dashed lines indicate hydrogen bonds. (B-C) RT-qPCR demonstrates the effect of ZP1 and ZP3 mutations on the mRNA expression level of target genes. Data are represented as the means ± SEMs, n = 3. ****P < 0.0001

Effects of ZP1 and ZP3 variants on protein expression

To investigate the effects of ZP1 and ZP3 mutations on protein expression and localization, we transfected plasmids carrying wild-type or mutant cDNA into Hela cells. WB analysis confirmed that premature introduction of a termination codon in mRNA by ZP1 (p.Gly477*) resulted in truncated mutant proteins, where low expression levels of the truncated proteins could be caused by partial protein degradation (Fig. 3A). ZP1 (p.Met1?) mutant protein expression level was significantly reduced, as evidenced by the production of low-intensity bands (Fig. 3B). We repeated the WB experiments using the ZP1 antibody to obtain the same results (Supplementary Fig. 1), speculating that it might be associated with the mutation being located at the start codon. For the ZP3 missense mutations (p.Asp242Tyr and p.Asn272Ser), the levels of the mutant ZP3 proteins in cell lysates were all prominently higher than those of the wild-type ZP3 proteins (Fig. 3C). The transfected cells were subjected to immunofluorescence staining. Confocal fluorescent microscopy revealed that the wild-type ZP1 and ZP3 proteins expressed in the cytomembrane and cytoplasm. The mutations did not alter the localization of the proteins in Hela cells. Compared with the wild-type proteins, the signals of ZP1 (p.Gly477* and p.Met1?) mutant proteins were significantly decreased (Fig. 3D), and the signals of ZP3 (p.Asp242Tyr and p.Asn272Ser) mutant proteins were markedly increased (Fig. 3E). The immunofluorescence results were consistent with those of WB.

Fig. 3

Characterization of mutant ZP1 and ZP3 proteins. (A-C) Effects of mutants on ZP1 and ZP3 proteins levels in Hela cells transfected with wild-type or mutant plasmids were detected by WB. The bar graph (below) represents an optical densitometric plot of relative protein expression, with GAPDH used as a standard. Data are represented as the means ± SEMs, n = 3. **P < 0.01, ****P < 0.0001. (D) Effects of ZP1 mutations (p.Gly477* and p.Met1?) on intracellular localization and expression of ZP1 proteins by IF staining. Confocal microscopy image showing ZP1 (red) and DAPI (blue) imaging. Scale bar: 20 μm. (E) Effects of ZP3 mutations (p.Asp242Tyr and p.Asn272Ser) on intracellular localization and expression of ZP3 proteins by IF staining. Confocal microscopy image showing images of ZP3 (red), anti-FLAG (green), and DAPI (blue). Scale bar: 25 μm

ZP1 and ZP3 mutations affect ZP protein interactions

We performed Co-IP experiments to analyze the interaction between the ZP gene mutation and the ZP2 glycoprotein. When HA-ZP1^WT or HA-ZP1^G477X plasmids were co-transfected with MYC-ZP2^WT plasmid into Hela cells respectively, ZP1^G477X protein truncation and reduced protein expression could be observed in the input group. Notably, the interaction between ZP1 and ZP2 transfected with ZP1^G477X was enhanced compared to ZP1^WT in the IP group (Fig. 4A).

When FLAG-ZP3^WT, FLAG-ZP3^D242Y or FLAG-ZP3^N272S were co-transfected with MYC-ZP2^WT, respectively, an increase in overall ZP3 protein expression could be observed in input group for ZP3^D242Y and ZP3^N272. However, when we co-transfected the two plasmids ZP3^D242Y and ZP3^N272S, ZP3 protein expression in input group was unaffected. A similar effect occurs in IP group, where a small amount of ZP3^D242Y can interact with a large amount of ZP2^WT, but a large amount of ZP3^N272S can only pull down a small amount of ZP2^WT. When ZP3^D242Y and ZP3 ^N272S were co-transfected indicated that their interaction with ZP2^WT was significantly weakened compared to ZP3^WT (Fig. 4B). This shows that ZP3^N272S has a greater effect on ZP2 binding and that the results when simulating a compound heterozygous mutation are not as simple as the neutralization of the effects of the two mutations.

Fig. 4

Effects of ZP1 and ZP3 mutations on ZP protein interactions and secretion. (A) Interaction of ZP1^WT/ZP1^G477X with ZP2 was evaluated by Co-IP with an anti-HA antibody. (B) Interaction of ZP3^WT/ZP3^D242Y/ZP3^N272S/ZP3^D242Y + ZP3^N272S with ZP2 was evaluated by Co-IP with an anti-FLAG antibody. (C-D) The ZP1 and ZP3 levels in the medium were measured by ELISA. Data are represented as the mean ± SEM; *P < 0.05

ZP1 and ZP3 mutations affect ZP protein secretion

In addition, we added ELISA experiments to explore whether the mutations affect ZP protein secretory function. We co-transfected HA-ZP1^WT or HA-ZP1^G477X with MYC-ZP2^WT and FLAG-ZP3^WT to demonstrate the effect of ZP1 nonsense mutations on secretion. ELISA results demonstrate that ZP1^G447X inhibits ZP1 protein secretion (Fig. 4C).

We co-transfected HA-ZP1^WT and MYC-ZP2^WT with FLAG- ZP3^WT, FLAG-ZP3^D242Y or FLAG-ZP3^N272S, respectively, to demonstrate the effect of ZP3 missense mutations on secretion. We also set up a set of samples transfected with FLAG-ZP3^D242Y and FLAG-ZP3^N272S to simulate compound heterozygous mutation. ZP3 ELISA Kit assessment of secreted factors concluded that the level of ZP3 expression was increased in the presence of ZP3^D242Y or ZP3^N272S. However, when ZP3^D242Y and ZP3^N272S were co-transfected, the secretion level of ZP3 was almost unchanged (Fig. 4D).

Discussion

This study recruited two patients with primary infertility from different families who were diagnosed with GEFS after only null COCs without oocytes were obtained in repeated IVF cycles. We undertook WES and Sanger sequencing to identify ZP1 and ZP3 variants associated with female infertility. Family 1 proband carried a ZP1 compound heterozygous mutation (c.[2T > A]; [1429G > T]). In this case, the p.Gly477* can lead to premature termination of the protein, generating a truncated protein. The p.Met1? caused the initiation codon for translation to emerge 141 bases downstream. The proband of family 2 carried two new heterozygous cis mutations in ZP3 (c.[724G > T;815 A > G]). We verified by WB that the ZP1 mutations resulted in a decrease in protein expression, whereas the ZP3 mutations resulted in an increase in protein expression. Additionally, we used confocal fluorescence microscopy to observe that both ZP1 and ZP3 proteins are localized in the cell membrane and cytoplasm, demonstrating that mutations do not alter the localization of the proteins. Co-IP assays revealed that ZP1^G477X and ZP3^D242Y had increased binding capacity to ZP2, but ZP3^N272S had weakened interactions with ZP2, which may affect ZP formation and assembly. ELISA demonstrated that ZP1^G477X inhibits ZP1 protein secretion, ZP3^D242Y or ZP3^N272S resulted in an increase in the secretion level of ZP3. However, co-transfection of ZP3^D242Y and ZP3^N272S did not affect the level of ZP3 in the culture medium, and Co-IP also showed that the binding of ZP3 and ZP2 was significantly weakened by co-transfection of the two missense mutations. This study advances the investigation of the genetic etiology of EFS and broadens the spectrum of ZP gene mutations associated with EFS. In vitro experiments demonstrate the effect of mutations on ZP protein expression and function, and speculate on the pathogenic mechanisms of EFS. This study also provides evidence for the development of molecular diagnostics and therapy for ZP gene defects.

The prevalence of GEFS in patients undergoing IVF was 0.016% [39]. Although cases of GEFS are rare, patients with this disease often present with repeated IVF failures, which undoubtedly puts tremendous pressure on doctors and patients. The main causative factors of GEFS include ovarian aging, folliculogenesis dysfunction, and genetic defects [40]. With increasing genetic studies, four genes have been found to be associated with GEFS: LHCGR [7], ZP1 [8,9,10,11,12,13,14,15,16,17,18,19,20,21,22, 30], ZP2 [9, 23, 24], and ZP3 [9, 16, 25,26,27,28,29]. Previously, Yang et al. reported that ZP gene variants accounted for 51.43% (18 of 35) of women with specific types of GEFS [9]. We summarized patients with EFS that have been reported in studies to date and identified 43 ZP1 mutations [8,9,10,11,12,13,14,15,16,17,18,19,20,21,22, 30], 4 ZP2 mutations [9, 23, 24], and 8 ZP3 mutations [9, 16, 25,26,27,28,29] (Table 1). Therefore, we speculate that mutations in the ZP gene, as one of the genetic factors, play a very significant role in the etiology of GEFS.

ZP is vital for oocyte growth and development, fertilization, and preimplantation embryo protection. During oogenesis, ZP stabilizes the intercellular connections or gap junctions between the oocytes and cumulus cells, allowing the proper transport of nutrients, metabolites, and other molecules from the external environment to growing oocytes [31]. Human ZP consists of four glycoproteins, ZP1-4. ZP2, ZP3, and ZP4 act as building blocks for ZP to assemble into ZP filaments, ZP1 covalently cross-links ZP2/3/4 filaments to form a robust extracellular matrix around the oocyte [41].

Differences in the position of mutations in the ZP gene and associated changes in protein structure often give rise to phenotypic heterogeneity. Mutations in the consensus furin cleavage site (CFCS) affect ZP protein secretion [24]. Mutations located in the transmembrane-like domain (TMD) or resulting in loss of the TMD produce ZP-free oocytes. Variants located in the N-terminal structural ZP-N domain (ZP-N1) and ZPD (containing ZP-N and ZP-C) can have severe consequences in terms of oocyte degeneration and GEFS [9, 22]. Yang et al. reported five nonsense mutations in ZP1 located in the ZP-C; either homozygous or heterozygous mutations increased ZP1 protein expression [9]. However, in this study, ZP1 (c.1429G > T, p.Gly477*) was located in the ZP-C, and mutant protein expression was lower than that of the wild type, which differs from the trend reported in previous studies. WB demonstrated that ZP1 (p.Gly477*) triggers partial degradation of truncated ZP proteins, and Co-IP suggests that ZP1^G477X binds excessively to ZP2. Therefore, we hypothesize that the remaining mutated ZP1 proteins are incapable of assembling into normal ZPs. ZP deletion disrupts the communication pathway between cumulus granulosa cells and oocytes. The metabolism of oocytes and cumulus granulosa cells is impaired, which triggers the production of reactive oxygen species [42, 43]. Without a supply of nutrients, the oocyte becomes more brittle, or there is a risk of accelerated granulosa cell apoptosis [10], which ultimately leads to the emergence of EFS. ZP1 (c.2T > A, p.Met1?) was first reported as a de novo variant carried by patients with EFS from LIU et al. They did not detect mutant ZP1 protein expression by WB; hence, they presumed that the mutation disrupted ZP1 expression [8]. However, Yang et al. presented a different opinion, indicating that ZP1 (p.Met1?) produces low-intensity bands, which is consistent with our experimental results [9]. In this study, we further confirmed by immunofluorescence that p.Met1? decreases the level of protein expression but does not affect the localization. Combined with the significantly increased expression of p.Met1? mRNA in RT-qPCR, we hypothesized that p.Met1? as a mutation located at the initiation codon may have affected the mRNA splicing process. Although mRNA synthesis was increased at the transcriptional level, the aberrant splicing may have generated erroneous isoforms that may contain code-shifting mutations, preventing translation from proceeding properly and ultimately resulting in no functional protein production.

In studies of ZP3, five families of patients with recurrent EFS were screened for the ZP3 heterozygous missense mutation (c.400G > A, p.Ala134Thr), which emphasizes the high-frequency pathogenicity of this mutation for EFS [9, 26]. Chen et al. demonstrated that ZP3 (p.Ala134Thr) can reduce the interaction between wild-type ZP3 and ZP2 through dominant-negative effect inhibitory binding, which prevents ZP assembly, ultimately leading to EFS [26]. Following this, Zhou et al. discovered another heterozygous mutation in ZP3 (c.763 C > G, p.Arg255Gly) with a dominant-negative effect; it can compete with the wild-type ZP3 protein for binding to the other three wild-type ZP proteins in vivo, preventing the formation of stable ZPs around the patient’s oocytes [16]. Subsequent studies reported four more ZP3 variants (c.176T > A; c.502_504delGAG; c.518 C > G; c.565_579del), among which c.502_504delGAG and c.518 C > G do not alter the expression of ZP3 in the cell lysates, but hinder the ZP3-ZP2 dimer formation [25, 27,28,29]. All six cases of EFS-related ZP3 mutations reported up to now were in the ZPD. The ZPD is a structural domain common to ZP proteins, which is responsible for mediating the homodimerization of ZP3 in the dormant intracellular state and the heterodimerization of ZP2 and ZP3 in the extracellular space, ultimately enabling their integration into the ZP protofibrils [32].

In our study, two ZP3 heterozygous cis mutation (c.[724G > T;815 A > G]), which is also located in the ZPD, was identified for the first time in GEFS patients. In vitro experiments revealed that both mutations increased protein expression but did not change the localization of the ZP protein, and that the mutations increase the level of secretion of the ZP3 protein. It is surprising that when co-transfecting ZP3^D242Y and ZP3^N272S plasmids, we can observe that the expression of ZP3 protein is unchanged and the level of ZP3 secretion tends to be normal, but Co-IP experiment show that the ability of ZP3 and ZP2 binding is reduced. This demonstrates that the mechanism of the two missense mutations affecting the target protein is complex. Through bioinformatic analysis, we found that ZP3 (p.Asp242Tyr) could disrupt the function of hydrogen bonding, so we presume that the mutation might affect the stability and binding ability of ZP protein, ZP could not be successfully assembled, ultimately leading to oocyte degeneration and EFS.

The phenotypes of ZP gene mutations are diverse, including thin ZP, ZP deletion, oocyte degeneration, and even severe enough to transform into EFS. Knowing the genotype of the ZP gene mutation and its corresponding phenotype helps us formulate the most appropriate treatment protocol for the patient. Today, ART is assisting a growing number of people suffering from infertility. Previous studies have shown that ZP abnormalities in patients with ZP1, ZP2, and ZP3 mutations can be treated with ICSI to achieve a successful pregnancy [16, 17, 19, 33, 44, 45]. This finding suggests that if ZP-free oocytes can be obtained from patients, we may be able to promote pregnancy by using a combination of unique culture systems and ICSI methods. However, for most patients with EFS, the inability to obtain oocytes in IVF cycles leads to multiple ART failures, which are the norm. A case of two infertile sisters diagnosed with GEFS conceived and became mothers through oocyte donation [27]. Therefore, we recommend early genetic analysis to detect any genetic pathogenic factors in patients with recurrent EFS. Suppose that a mutation in the ZP gene was found. In such cases, doctors should advise patients with EFS to consider donor oocytes as a first option. In short, genetic analysis plays an irreplaceable role in the diagnosis and prognosis of GEFS. Nevertheless, the study of the etiology of GEFS is far from complete, so the full spectrum of genes and mutations involved remains to be expanded.

It must be recognized that our research has some limitations. First, the study’s small sample size is not convincing, we still need other reproduction centers working together to identify the same mutation sites to confirm the harm further. Second, we were unable to keep a photo of the patient’s COCs at the time of her visit, which would have helped us to better understand the patient’s condition. If we could have retained the samples for experimental staining, we could have also investigated the effect of the mutation in depth. Third, this study only simulated the effects of the mutation models on RNA and protein functions through in vitro cellular experiments, and no animal studies were conducted. As a result, it is impossible to replicate the in vivo changes fully. Furthermore, we only individually examined the two cis-heterozygous mutations of the ZP3 gene and did not consider the scenario in which the mutations occurred on the same allele. Therefore, it remains unclear whether the combined presence of these mutations on the same allele would lead to synergistic or antagonistic effects on ZP3 function. Future research with more comprehensive methods, including animal models and allele-specific investigations, is necessary to fully clarify these mutations’ pathological mechanisms and biological implications in EFS.

Conclusion

This study identified pathogenic ZP1 and ZP3 mutations in two families of EFS patients. In particular, we report two novel ZP3 heterozygous cis mutation associated with EFS and female infertility, which enriches the spectrum of genetic etiological variants associated with EFS. We verified the effect of mutations on ZP protein expression and function by in vitro experiments, further elucidating that ZP is an essential genetic etiology of EFS. Moreover, mutations located in the ZPD may have more severe consequences. Finally, on the basis of the failure of most EFS patients to achieve pregnancy outcomes, we highlight the importance of genetic analysis in the diagnosis and prognosis of GEFS and recommend that patients strive for a successful pregnancy through oocyte donation.