To investigate the biological relevance of YAP1 and NOA, whole-exome sequencing (WES) was performed with 777 patients diagnosed with NOA (Fig. 1A). Interestingly, we found that 16 patients (16/777) had YAP1 (NM_001282101.2) SNVs among 777 NOA patients, and this incidence of YAP1 by 2.06% was remarkably higher when compared to the average gene mutation rate by 0.3%. The minor allele frequencies (MAFs) of all the variants were lower than 0.01 in the Genome Aggregation Database (gnomAD), 1000 Genomes Project (1000G), and ExAC. Notably, among these 16 patients, there were 12 SNVs, including 6 novel SNVs with absent (YAP1 c.42G>A, YAP1 c.55C>A, YAP1 c.134C>T, YAP1 c.1116C>G, YAP1 c.1482G>A, and YAP1 c.689-10T>C) and 6 SNVs with extremely low frequency (YAP1 c.57G>A, YAP1 c.387A>G, YAP1 c.567C>T, YAP1 c.680C>T, YAP1 c.879C>T, and YAP1 c.1016C>T) in the above databases. Among all SNVs, there was one homozygous variant YAP1 (YAP1 c.55C>A) and the remaining variants were heterozygous (Table S1). Next, we conducted the correlation analyses between the SNV allele frequency of YAP1 gene and the risk of NOA. According to the American College of Medical Genetics and Genomics (ACMG) guidelines, 6 variants were categorized as likely pathogenic (PM2). For the other 6 SNVs with extremely low frequency, statistical analyses were performed. The allele frequency of the same locus in the GnomAD database was used as the control, and odds ratios (ORs) and the confidence interval (CI) around the OR were chosen as the indicators for risk assessment. Statistical analysis demonstrated that the carriage of G allele at the rs773991529 locus (OR = 40.445, 95% CI = 4.518 to 362.059), T allele at the rs757399180 locus (OR = 23.059, 95% CI = 2.835 to 187.526), T allele at the rs756175059 locus (OR = 26.926, 95% CI = 3.240 to 223.780), and T allele at the rs527326391 locus (OR = 9.685, 95% CI = 1.007 to 93.171) was found to be associated with increasing risk of NOA (PS4), while the carriage of A allele at the rs937455182 locus might decrease the risk of NOA (OR = 0.025, 95% CI = 0.003 to 0.174). There was no correlation between the carriage of T at the rs376161041 locus and NOA risk (P > 0.05). Subsequently, the pathogenicity of missense variants of YAP1 was predicted and evaluated using Sorting Intolerant from Tolerant (SIFT), PolyPhen-2_HDIV, MutationTaster, and M-Cap. As shown in Table S2, c.55C>A, c.134C>T, c.680C>T, and c.1016C>T in the YAP1 gene were potentially deleterious variants (PP3). In addition, analysis of impact of single-nucleotide polymorphisms on protein stability predicted from the protein structure (AF-P46937-F1, AlphaFold) was conducted using the INPS-MD server (Biocomputing Group) and the MAESTROweb server. The S227L mutant appeared to remarkably impair protein stability, with an increase in stability change (DDGpred) in 1.5 and 0.4 kcal/mol, respectively (Fig. 1B). To verify the prediction results, we constructed P45L and S227L mutations of YAP1 (Fig. S2A) in a human SSC line [19]. The human SSC line displayed phenotypic characteristics similar to human primary SSCs (Fig. S3A and B). The protein stability experiment showed that the stability of YAP1 was decreased after YAP1 S227L mutations treated with cycloheximide (CHX) in human SSCs (Fig. 1C).

We determined the cellular distribution of YAP1 in normal human testicular tissues. Immunohistochemistry via 3,3′-diaminobenzidine (DAB) staining indicated that YAP1 was predominantly localized in cell nuclei of human primary SSCs (arrows) around the basement membrane of seminiferous tubules in the normal testicular tissues (Fig. 1D, left panel). A few of pachytene spermatocytes and round spermatids were positive for YAP1 (Fig. 1D, left panel). Immunohistochemistry via immunofluorescence further demonstrated that YAP1 was predominantly located in the nuclei of primary SSCs (Fig. 1E). We compared YAP1 expression levels in testes from NOA patients and OA patients with normal spermatogenesis. Western blots showed a lower level of YAP1 protein in testicular tissues from NOA patients compared to OA patients with normal spermatogenesis (Fig. 1F). Immunohistochemistry displayed that YAP1 immunostaining could not be seen in NOA patients (Fig. 1G), which was distinct from YAP1 immunostaining in OA patients with normal spermatogenesis (Fig. 1E).

Whether P45L and S227L mutations of YAP1 affect DNA synthesis and apoptosis of human SSC, we performed 5-ethynyl-2′-deoxyuridine (EDU) incorporation assays and Annexin V/propidium iodide (PI) staining via flow cytometry. EDU incorporation assays showed that S227L mutation of YAP1 decreased DNA synthesis of human SSCs (Fig. 1H and I), and Annexin V/PI staining via flow cytometry indicated the increased apoptosis in the human SSCs compared to wild type (WT) of YAP1 (Fig. 1J and K). However, P45L mutation of YAP1 had no effect on DNA synthesis and apoptosis of human SSCs (Fig. S2B and E). Collectively, these results implicate that YAP1 SNVs and/or lower expression level are associated with the NOA and that YAP1 mutations affect the fate determinations of human SSCs.

Subsequently, we investigated the impact of YAP1 silencing on human SSC proliferation and apoptosis using RNA interference (RNAi). Three small interfering RNAs (siRNAs) were designed to target human YAP1 mRNA, and their detailed information was listed in Table S5. Cy3-labeled siRNA was employed to illustrate that the transfection efficiency of siRNAs in human SSCs was more than 90% (Fig. S4A). Compared to control-siRNA, YAP1-siRNA2 had a higher knockdown efficiency in human SSCs via real-time polymerase chain reaction (PCR) (Fig. S4B) and Western blots (Fig. S4C and D). As such, YAP1-siRNA2 was chosen for further studies of YAP1 functions and mechanisms. Our cell counting kit-8 (CCK-8) assay showed a decrease in the proliferation potential of human SSCs at 48 to 120 h after transfection with YAP1-siRNA2 and YAP1-siRNA3 (Fig. 2A). Western blots indicated a reduction in the level of proliferating cell nuclear antigen (PCNA) in human SSCs following treatment with YAP1-siRNA2 (Fig. 2B and C). In EDU incorporation assays, the percentages of EDU-positive cells were lower in human SSCs treated with YAP1-siRNA2 than in control-siRNA (Fig. 2D and E). Annexin V/PI staining with flow cytometry revealed that YAP1-siRNA2 remarkably elevated both early and late apoptosis of human SSCs (Fig. 2F and G). Moreover, the proportion of terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick end labeling (TUNEL)-positive cells was higher in human SSCs with YAP1-siRNA2 treatment than with control-siRNA (Fig. 2H and I). Considered together, these findings imply that YAP1 silencing reduces the proliferation and DNA synthesis and enhances the apoptosis of human SSCs.

To examine the role of YAP1 in controlling human SSC fate determinations in vivo, xenotransplantation with human SSCs with control-siRNA and YAP1-siRNA2 was performed. At 4 weeks after treatment of busulfan, testicular volumes were obviously reduced in mice (Fig. S4E). Hematoxylin and eosin (H&E) staining indicated the complete absence of male germ cells within the seminiferous tubules of recipient mice, whereas all types of male germ cells were observed in the normal mice (Fig. S4F). Transplantation of human SSCs with treatment of control-siRNA and YAP1-siRNA2 was conducted through the efferent ducts (Fig. S4G). Immunohistochemistry of testis sections displayed the percentages of human nuclear antigen (HumNuc)-positive cells and PCNA-positive cells were obviously decreased in the recipient mice transplanted with human SSCs treated with YAP1-siRNA2 compared to the control mice (Fig. 2J and K). The percentages of HumNuc- and UCHL1-postive cells exhibited a similar trend by YAP1-siRNA2 (Fig. S4H). Together, these results indicate that YAP1 silencing suppresses the self-renewal of human SSCs in vivo.

In order to identify the targets of YAP1 in human SSCs, RNA sequencing was performed to detect the differentially expressed genes (DEGs) of human SSCs between YAP1-siRNA2 and control-siRNA. To ensure RNA quality, we captured electropherograms using the Agilent Bioanalyzer and utilized the RNA integrity number (RIN) to evaluate RNA quality. The electropherograms of human SSCs with treatment of control-siRNA and YAP1-siRNA2 displayed flat baselines and sharp peaks (Fig. S5), and the RIN of 2 groups was 9.5 and 9.6, respectively, which reflects a high integrity of RNA extracted from human SSCs. RNA sequencing revealed that 14,747 and 14,754 genes were present in human SSCs treated with control-siRNA and YAP1-siRNA2, respectively. Based on fold change < 0.5 or fold change > 2, clustered heatmap and volcano plot indicated the DEGs in human SSCs treated with YAP1-siRNA2 and control-siRNA (Fig. 3A and B). According to the criteria [3 < fragments per kilobase of transcript per million mapped reads (FPKM) values < 50, P < 0.01, and fold change < 0.5], Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses, NEDD4, RNF144B, ARPC5, and UCN were selected as candidate genes for further validation. Real-time PCR showed that mRNA levels of YAP1, NEDD4, RNF144B, ARPC5, and UCN were decreased by YAP1-siRNA2 in human SSCs compared to control-siRNA (Fig. 3C), which was in agreement with our findings using RNA sequencing. Moreover, YAP1-siRNA2 remarkably reduced the protein level of NEDD4 compared to control-siRNA (Fig. 3D). Likewise, the S227L mutation of YAP1 decreased the NEDD4 protein level compared to WT of YAP1 (Fig. 3E). Taken together, NEDD4 was identified as a target of YAP1 in human SSCs.

As a target of YAP1, we investigated whether NEDD4 plays a role in regulating human SSC fate determinations. First, real-time PCR and Western blots were used to evaluate the knockdown efficiencies of NEDD4, which reflects the decreases in mRNA (Fig. S6A) and protein (Fig. S6B and C) levels of NEDD4 in human SSCs treated with NEDD4-siRNA, especially NEDD4-siRNA3.

Next, CCK-8 assay displayed that cell proliferation was suppressed by NEDD4-siRNA2 and NEDD4-siRNA3 in human SSCs at 48 to 120 h after transfection (Fig. 4A). Western blots showed the decrease in PCNA protein level in human SSCs treated with NEDD4-siRNA2 and NEDD4-siRNA3 (Fig. 4B and C). EDU incorporation assays showed a notable reduction in EDU-positive cells in human SSCs upon treatment with NEDD4-siRNA3 (Fig. 4D and E). Moreover, flow cytometry and TUNEL assay were used to evaluate the apoptosis. Flow cytometry indicated that NEDD4-siRNA3 increased the apoptosis in human SSCs compared to control-siRNA (Fig. 4F and G). Similarly, proportion of TUNEL-positive cells in human SSCs was enhanced by NEDD4 silencing compared to control-siRNA (Fig. 4H and I). Collectively, these results suggest that NEDD4 knockdown reduces the proliferation and DNA synthesis of human SSCs and increases their apoptosis.

To probe whether NEDD4 mediated the influence of YAP1 knockdown on human SSCs, we performed rescue experiments. First, we constructed a NEDD4 overexpression plasmid that was verified by DNA sequencing. Real-time PCR and Western blots revealed significant increases in the mRNA (Fig. S7A) and protein (Fig. S7B and C) levels of NEDD4 in human SSCs transfected with the NEDD4 plasmid. Moreover, human SSCs overexpressing NEDD4 exhibited an increased level of PCNA protein compared to the control vector (Fig. S7B and C), which was in contrast with the effect of NEDD4 knockdown on human SSCs. Notably, NEDD4 overexpression in human SSCs reversed the reduction in proliferation caused by YAP1-siRNA2 (Fig. 5A). Similarly, NEDD4 overexpression partially rescued the decrease in PCNA protein level caused by YAP1-siRNA2 in human SSCs (Fig. 5B and C). As shown in Fig. 5D and E, NEDD4 overexpression led to an enhancement in EDU-positive cells of human SSCs, whereas YAP1-siRNA2 inhibited their DNA synthesis. NEDD4 overexpression counteracted the effect of YAP1-siRNA2 on DNA synthesis of the human SSCs. Moreover, flow cytometry (Fig. 5F and G) and TUNEL assay (Fig. 5H and I) displayed that the impact of YAP1-siRNA2 on apoptosis could be reverted by NEDD4 overexpression. Taken together, these findings imply that YAP1 silencing leads to decreases in proliferation and DNA synthesis as well as an enhancement in the apoptosis via NEDD4 in human SSCs.

YAP1 is a transcriptional coactivator, and it could not bind to DNA promoter due to lacking DNA-binding domains. Thus, we speculated that YAP1 could interact with transcription factors to regulate the downstream gene NEDD4. In order to screen potential transcription factors binding to YAP1 in human SSCs, we further conducted co-immunoprecipitation (Co-IP) and mass spectrometry (MS). Our MS for YAP1 was shown in Fig. 6A, and our Co-IP and MS identified 956 and 890 proteins in human SSCs, respectively. Among them, 518/437 were up-regulated proteins, whereas 239/254 were down-regulated proteins and 199/207 were unchanged proteins (Fig. 6B). Our GO analysis showed that proteins pulled down by YAP1 were mainly enriched in the translation activity (Fig. 6C). Meanwhile, the promoter region of NEDD4 was determined to be from transcription start site (TSS) to position TSS+ 2,000 base pairs in human chromosome 15 (Chr15: 55993612-55995612, GRCh38.p14). Next, human transcription factor database was employed to predict the potential transcription factors binding to the promoter of NEDD4. Subsequently, we took the intersection of both the up-regulated proteins for Co-IP/MS and the potential transcription factors of NEDD4, and a Venn diagram was plotted. There were 9 transcription factors, including RAD21 (RAD21 cohesin complex component), PRKDC, SMC1A, LMNB1, GLYR, TFCP2, MAFG, SMARCC1, and SF1 (Fig. 6D). To uncover the potential binding proteins, we screened the above transcription factors based on the localization, expression, and bioinformatics analysis with Cistrome Data Brower. Notably, RAD21 was identified as the potential intermediate factor that links YAP1 and NEDD4 in human SSCs. Our MS for RAD21 was shown in Fig. 6E.

In order to demonstrate the interaction between YAP1 and RAD21, we performed the Co-IP analysis. Endogenous Co-IP showed that RAD21 was co-precipitated with YAP1 (Fig. 6F). Double immunostaining revealed that YAP1 (green fluorescence, nuclei-cytoplasm) and RAD21 (red fluorescence, nuclei) are coexpressed in the human SSCs (Fig. 6G). Intriguingly, immunohistochemistry of human testes displayed that RAD21 was specially located in the nuclei of human primary SSCs along with the basement membrane of seminiferous tubules rather than other male germ cells in human testicular tissues (Fig. 6I, upper panel). Notably, double immunostaining demonstrated that RAD21 was co-expressed with UCHL1, a marker for human SSCs, in primary SSCs of human testes (Fig. 6H). RAD21 has been known to be involved in DNA damage repair. We designed 3 RAD21-siRNAs that could effectively knock down the mRNA (Fig. S8A) and protein (Fig. S8B and C) levels of RAD21. CCK-8 assay exhibited that RAD21 silencing inhibited the proliferation of human SSCs (Fig. 6J). Considered together, these findings imply that RAD21 exerts a significant role in maintaining the stemness of human SSCs mediated by YAP1.

Next, we analyzed NEDD4 promoter sequences using the hTFtarget database to predict transcription factor binding sites. We found that multiple binding sites in the NEDD4 promoter sequences may bind to RAD21 (Table S3). We chose 4 binding sites with high scores to perform dual-luciferase reporter assay (Fig. 6K). Compared to control plasmid, the NEDD4 promoter activity was enhanced by RAD21 in human SSCs (Fig. 6L). When site 2 was mutated, NEDD4 activity enhanced by RAD21 was reduced in these cells (Fig. 6L). Collectively, these data suggest that YAP1 binds to RAD21 and targets NEDD4 to mediate the proliferation, DNA synthesis, and apoptosis of human SSCs.

To probe the function of YAP1 in controlling SSCs and spermatogenesis in vivo, we generated Yap1 cKO mice by crossing Yap1 floxed mice and Stra8-EGFP Cre mice. For the Yap1 cKO mice, the exon 2 of Yap1 was deleted (Fig. 7A). Eventually, Yap1 cKO mice was successfully generated as demonstrated by the absence of Yap1 transcription (Fig. S9A) and Yap1 protein (Fig. S9B) in testes of mice as shown by reverse transcription PCR (RT-PCR) and Western blots, respectively. The adult testis/weight ratio in the Yap1 cKO mice was lower than that in the WT mice (Fig. 7B and Fig. S9C). Meanwhile, in adult Yap1 cKO mice, the count of sperm in the cauda epididymis was remarkably fewer compared to control mice (Fig. 7C, lower and left panels). There was almost no active spermatogenesis in the seminiferous tubules of Yap1 cKO mice at 6 weeks (Fig. 7C, lower and middle panels). Similarly, H&E staining of the testicular sections in Yap1 cKO mice revealed a decrease in spermatozoa count at 10 weeks (Fig. 7C, lower and right panels). Together, these data indicate that Yap1 cKO results in spermatogenesis disorder.

To explore the impact of Yap1 knockout on the fate determinations of SSCs in Yap1 cKO mice, we determined the expression of SSC markers (PLZF and UCHL1) and spermatogonial differentiation hallmark (c-Kit) utilizing immunofluorescence. Compared with control WT mice, the percentages of PLZF-positive cells (Fig. 7D and H) at 7 days after birth (P7), c-Kit-positive cells (Fig. 7E and H) at P56, and UCHL1-positive cells (Fig. 7F and H) in Yap1 cKO mice were lower. Meanwhile, we detected proliferation and apoptosis of SSCs using immunofluorescence and TUNEL assay. The proportion of PCNA-positive cells (Fig. 7G and H) was obviously decreased in the seminiferous tubules of Yap1 cKO mice, whereas the percentages of TUNEL-positive cells were dramatically increased in Yap1 cKO mice (Fig. 7I and J). Taken together, these data implicate that Yap1 knockout leads to notable reduction in the proliferation and a remarkable increase in apoptosis of SSCs in vivo.

We further analyzed the roles of Yap1 in mediating late stage of spermatogenesis in Yap1 cKO mice. Interestingly, computer-assisted sperm analysis (CASA) revealed significant decreases in sperm count and sperm motility parameters, e.g., progressive motility, curvilinear velocity, straight-line velocity, and average path velocity, in the cauda epididymis of Yap1 cKO mice (Fig. 8A). Sperm morphology was assessed by Papanicolaou staining. As seen in Fig. 8B and C, there were a number of sperm with aberrant morphology, including abnormal head formations, acephalic sperm, bent tails, shorted tails, and absent tails. The ultrastructural analysis of epididymal sperm was performed with transmission electron microscopy (TEM). In the midpiece of most flagella of the Yap1 cKO mice, a disordered structure of mitochondrial sheath was observed (Fig. 8D). In the principal piece of numerous sperm tails, we found the disorganized microtubule doublets (OD) and outer dense fiber (ODF) (Fig. 8D). Collectively, these findings indicate that Yap1 depletion causes oligoasthenoteratospermia in mice and Yap1 is involved in the regulation of spermiogenesis.

Testicular tissues were collected from patients diagnosed with OA and NOA undergoing testicular biopsy or surgical resection at The Third Xiangya Hospital of Central South University and Hunan Cancer Hospital. The tissues were rinsed with phosphate-buffered solution (PBS) supplemented with 4% penicillin and streptomycin. Some testicular tissues were preserved in Bouin's fixative (Servicebio) for paraffin-embedded sections, while the remaining tissues were utilized for RNA and protein extraction. Human testicular tissues were used for research only and approved by the Ethics Committee of Hunan Normal University.

Human SSC line was cultured in Dulbecco's modified Eagle's medium (DMEM)/F12 (Gibco), supplemented with 10% fetal bovine serum (FBS) (Gibco), 100 units/ml penicillin (Gibco), and 100 g/ml streptomycin (Gibco) in a humidified atmosphere incubator of 5% CO₂ at 34 °C. This cell line was characterized using RT-PCR and immunocytochemistry to identify as human primary SSCs.

BALB/c male mice at 6 weeks old were obtained from Hunan Slake Jingda Experimental Animal Co. Ltd. (Changsha, China), and they were housed in the specific pathogen-free (SPF) animal facility. Busulfan (40 mg/kg body weight) was intraperitoneally injected into mice to remove male germ cells. After 4 weeks for busulfan injection, human SSC line transfected with YAP1-siRNA2 or control-siRNA was transplanted into seminiferous tubules of mice pursuant to the previously described method [19]. Four weeks after cell transplantation, testicular tissues were obtained from recipient mice for further studies. The animal experiments were ethically approved by the Ethics Committee of Hunan Normal University.

Yap1 floxed mice were provided by J. Wu from Xiamen University, and Stra8-EGFP-Cre mice with C57BL/6J background were obtained from Saiye Biotechnology Co. Ltd. (Jiangsu, China). First, Yap1^flox/wt mice were bred with each other to obtain Yap1^flox/flox mice. Next, the Yap1 homozygous mice were crossed with Stra8-Cre mice to generate Yap1^flox/wt;Stra8-Cre mice that were then crossed with Yap1^flox/flox mice to acquire Yap1 ^flox/;Stra8-Cre mice (designated as Yap1 cKO mice). Genotyping of Yap1 cKO mice was conducted via PCR using genomic DNA extracted from tail samples. The gene primers for genotyping were shown in Table S6. The animal experiments were conducted in accordance with the ethical standards and guidelines approved by the Ethics Committee of Hunan Normal University.

For immunocytochemistry, human SSCs treated with different siRNAs were harvested with trypsinization and resuspended in PBS. These cells were attached to the glass slides using cytospin (Cence) at 1,700 rpm for 1 min, and fixation was performed with 4% paraformaldehyde (PFA) for 15 min. After cold PBS washing, cells underwent 15 min of permeabilization at room temperature with 1% Triton. Subsequently, they were blocked with 5% bovine serum albumin for 1 h at room temperature, followed by overnight incubation with primary antibodies at 4 °C and 1-h incubation with fluorescent secondary antibodies at room temperature. Cell nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI), and the images were acquired using a confocal microscope (Leica TCS SP8 SR) or fluorescence microscope (Leica DM3000LED).

For immunohistochemistry, paraffin sections of testes were dewaxed at 65 °C for 90 min and dehydrated with gradient alcohol. Antigen retrieval was conducted through a 20-min boiling process in 1× sodium citrate buffer, and blocking of endogenous peroxidase was conducted with 3% hydrogen peroxide. The procedures of blocking, antibody incubation, and imaging capture were similar to the method as described in immunocytochemistry. The detailed information on the antibodies used in immunocytochemistry and immunohistochemistry, including the dilution and concentrations, was shown in Table S4.

The siRNAs for silencing PDK1, YAP1, NEDD4, and RAD21 were synthesized by GenePharma, while a nontargeting control-siRNA was employed as the negative control. The sequences of all siRNAs were listed in Table S5. The NEDD4-overexpressing plasmid carrying a red fluorescent (Cherry) was constructed by Shanghai Jikai Gene Technology Company, while empty vector was utilized for the negative control. According to the manufacturer's protocol, cells were transfected with siRNAs or plasmids using Lipofectamine 3000 (Invitrogen). The relative gene and protein levels were analyzed via real-time PCR and Western blots at 48 h after transfection, respectively.

Total RNA was extracted using Trizol reagent (Invitrogen) in terms of the manufacturer's protocol. Subsequently, the concentration and quality of total RNA were determined with NanoDrop One (Thermo Fisher Scientific). Next, Evo M-MLV RT Premix (Accurate Biology) was used to reverse transcribe the total RNA into cDNA, which was then used for PCR and real-time PCR of genes, respectively. ACTB, a housekeeping gene, was chosen to detect the loading RNA, while negative control did not contain any cDNA.

PCR was carried out in accordance with the protocol as described previously [53], and the reaction conditions comprised an initial denaturation at 95 °C for 5 min and followed by 35 cycles of denaturation at 95 °C for 30 s, annealing at 60 °C for 30 s, and elongation at 72 °C for 5 min. The PCR products were separated on 2% agarose gels stained with GoldView.

Real-time PCR was conducted utilizing 2× SYBR Green Premix Pro Taq HS qPCR Kit (Accurate Biology) in 20-μl reaction via the Bio-Rad CFX system. The gene expression was normalized to ACTB using the 2^−ΔΔCT method. The sequences of gene primers synthesized by Shenggong Biotech were listed in Table S6.

Protein extraction was performed by radioimmunoprecipitation assay (RIPA) supplemented with 0.1 mM phenylmethylsulfonyl fluoride and protease inhibitor cocktail (Thermo Fisher Scientific). The cytoplasmic and nuclear proteins were extracted using a nucleocytoplasmic protein extraction kit (Beyotime). The protein supernatant was combined with 5× sodium dodecyl sulfate (SDS)-loading buffer and denatured by boiling at 99 °C for 10 min.

Following separation via 10% SDS-polyacrylamide gel electrophoresis gel electrophoresis, proteins were transferred onto polyvinylidene difluoride membranes with 0.45-μm pore sizes (Millipore). After blocking with QuickBlock Blocking Buffer (Beyotime) for 30 min, primary antibodies were applied and incubated overnight at 4 °C. Horseradish peroxidase-conjugated secondary antibodies were used, and protein bands were visualized with enhanced chemiluminescence reagents and analyzed using ImageJ software. The detailed information on antibody dilution concentrations and specifics was shown in Table S7.

Cells were seeded in 6-well plates, and CHX was added into each well at a concentration of 200 μg/ml to block new protein synthesis. Cells were harvested at 0, 6, and 12 h after CHX treatment. Total protein was extracted using RIPA lysis buffer. Protein samples were used for Western blotting pursuant to the method as described above.

The Pierce IP/Co-IP kit (Thermo Fisher Scientific) protocol was utilized for proteins' interaction assay. Human SSCs were cultured in 10-cm dishes until they reached 85% confluence, and they were lysed using 500 μl of binging/washing buffer. The cell lysates were added with 10 μg of antibodies, including YAP1 or RAD21, and incubated overnight at 4 °C. Next, the antibody–lysate complexes were subjected to IP using 25 μl of protein A/G beads and constantly rotated for 2 h at room temperature. The magnetic rack was employed to gather the magnetic bead–immune complex, and washing buffer and rinsing with ultrapure water were conducted. Subsequently, the antigen–antibody–bead complexes were eluted with elution buffer and denatured for Western blots.

DNA extraction kit (QIAGEN) was utilized to extract genomic DNA from peripheral blood of OA and NOA patients according to the manufacturer's instructions. DNA concentration and purity were evaluated, and construction of the whole exome library was conducted, followed by sequencing on the Illumina HiSeq 2000 or NovaSeq 6000 sequencing platforms (Illumina). The raw reads underwent alignment to GRCh37/hg19 using the Burrows–Wheeler Aligner. Genetic variations, e.g., SNVs, deletions, and small insertions, were analyzed, annotated, and filtered by various public databases and in silico tools, including 1000G, gnomAD, ExAC, SIFT, PolyPhen-2, and MutationTaster.

Human SSCs were seeded into 96-well plates, with subsequent treatment with various kinds of siRNAs or control-siRNA. CCK-8 assays were conducted using CCK-8 (Dojindo) after siRNA transfection from 24 to 120 h. Following the manufacturer's guidelines, each well was received 10 μl of CCK-8 solution and 90 μl of DMEM/F12 with a 4-h incubation at 34 °C. Absorbance measurements were performed at a wavelength of 450 nm.

The Cell-Light EDU Apollo Kit (Ribobio) was utilized to conduct EDU incorporation assays pursuant to the manufacturer's instructions. Briefly, EDU solution was diluted at a ratio of 1:2,000 with DMEM/F12 medium and added into wells of human SSCs (100 μl/well) for overnight. The cells were subjected to fixation with 4% PFA for 30 min at room temperature and followed by neutralization with 2 mg/ml glycine for 5 min and permeabilization with 0.5% Triton X-100 for 10 min. Staining of EDU was conducted by incubating the cells with Apollo staining buffer in the dark for 30 min. After staining the nuclear DNA with DAPI or Hoechst 33342, the images were captured under a fluorescence microscope (Olympus). The ratio of EDU-positive cells was calculated from at least 500 cells with ImageJ.

Flow cytometry with APC Annexin V/PI staining was used to analyze apoptosis in human SSCs treated with different siRNAs in terms of a previously described method [20]. Briefly, human SSCs in 6-well plates were digested with trypsin without EDTA after siRNA transfection for 48 h. The cells were harvested, rinsed with PBS, centrifuged, and then suspended in 100 μl of Annexin V Binding Buffer (BioLegend). Next, 5 μl of APC Annexin V and 10 μl of PI (BioLegend) were added to cell suspension and followed by a 15-min incubation at room temperature in the dark. Flow cytometry (BD Biosciences) was utilized to quantify apoptotic cells.

Human SSCs were received 24-h treatment of either control-siRNA or YAP1-siRNAs and followed by total RNA extraction for RNA sequencing. RNA quality was assessed with the Agilent Bioanalyzer, while construction and sequencing of RNA libraries were conducted using the Illumina HiSeq X Ten platform. For data analysis, raw reads were performed for quality control, aligned, and annotated, and FPKM values were calculated using cufflinks2 software. The count of each sample was normalized using DESeq software. The fold changes were calculated, and a significance test was conducted to identify the DEGs based on the criteria of fold change >2.0 or <0.5 and P < 0.05.

IP with anti-PDK1, anti-YAP1, anti-RAD21 antibodies and normal immunoglobulin G (IgG) (Table S7) were used for MS. The magnetic bead samples were eluted, reduced with dithiothreitol, alkylated with iodoacetamide, and digested into polypeptides with trypsin. After desalting, the polypeptides were measured on a liquid chromatography–MS system comprising UltiMate 3000 RSLCnano (Thermo Fisher Scientific) and Q-Exactive Plus (Thermo Fisher Scientific). Raw data were analyzed with ProteomeDiscover software and identified via the UniProt-Human database. The identified proteins were employed for subsequent MS analysis.

The plasmids of pGL3-NEDD4, pcDNA3.1, pcDNA3.1-RAD21, pGL3-NEDD4 MUT1, pGL3-NEDD4 MUT2, pGL3-NEDD4 MUT3, and pGL3-NEDD4 MUT4 (Miaoling Biology, China) were transfected into human SSCs using Lipofectamine 3000. After 48 h of transfection, the cells were treated and subjected to luciferase activity detection using the Dual-Glo Luciferase Reporter Assay System (E2920, Promega) by the tube luminometer (Berthold, Germany).

The cauda epididymis was isolated from Yap1 cKO mice, sectioned into multiple pieces, and immersed in preheated G-IVF Plus (Vitrolife, Gothenburg, Sweden, LOT509784) at 34 °C for 30 min. In total, 10 μl of the supernatant was used to detect sperm motility by the CASA system. For estimation of sperm morphology, 5 μl of the supernatant was spread over slides, dried, and fixed in 95% ethanol for the modified Papanicolaou staining.

After fixation with 2.5% glutaraldehyde and osmium tetroxide, the cauda epididymis underwent dehydration in acetone followed by embedding in Epon 812. The morphology of sperm from Yap1 cKO mice was observed under a light microscope after staining the semithin sections with methylene blue. After being stained with uranyl acetate and lead citrate, the ultrathin sections underwent ultrastructure examination under a JEM-1400-FLASH TEM.

SPSS (IBM SPSS Statistics, version 23.0) was utilized for all statistical analyses. The mean ± standard deviation was calculated from data from 3 independent experiments. One-way analysis of variance (ANOVA) was conducted to assess the significance between groups.