Loss of CEP70 function affects acrosome biogenesis and flagella formation during spermiogenesis

The spermatogenesis process is complex and delicate, and any error in a step may cause spermatogenesis arrest and even male infertility. According to our previous transcriptomic data, CEP70 is highly expressed throughout various stages of human spermatogenesis, especially during the meiosis and deformation stages. CEP70 is present in sperm tails and that it exists in centrosomes as revealed by human centrosome proteomics. However, the specific mechanism of this protein in spermatogenesis is still unknown. In this study, we found a heterozygous site of the same mutation on CEP70 through mutation screening of patients with clinical azoospermia. To further verify, we deleted CEP70 in mice and found that it caused abnormal spermatogenesis, leading to male sterility. We found that the knockout of CEP70 did not affect the prophase of meiosis I, but led to male germ-cell apoptosis and abnormal spermiogenesis. By transmission electron microscopy (TEM) and scanning electron microscopy (SEM) analysis, we found that the deletion of CEP70 resulted in the abnormal formation of flagella and acrosomes during spermiogenesis. Tandem mass tag (TMT)-labeled quantitative proteomic analysis revealed that the absence of CEP70 led to a significant decrease in the proteins associated with the formation of the flagella, head, and acrosome of sperm, and the microtubule cytoskeleton. Taken together, our results show that CEP70 is essential for acrosome biogenesis and flagella formation during spermiogenesis.


Introduction
Infertility is a serious health problem worldwide. According to the latest data from the World Health Organization (WHO), around 10-15% of couples (50-80 million people) of reproductive age are suffering from infertility. Male factors account for about 50% of these cases of infertility [1][2][3] . In addition, recent data showed that the male etiology of~70% of infertile couples remains unexplained 4 . In male infertility patients, there is a type of disease called oligoasthenoteratozoospermia (OAT) that is mainly characterized by a decreased sperm concentration and motility, and a higher rate of abnormal shapes 5 . As a key cause of male infertility, azoospermia or OAT is mainly caused by severe damage to spermatogenesis 6 . Spermatogenesis is a complicated process that involves the proliferation of spermatogonial stem cells (SSCs), meiosis of spermatocytes, haploid differentiation, and spermiogenesis 7,8 . In the process of spermiogenesis, round spermatids undergo a series of dynamic morphological changes, including the formation of acrosomes, mitochondrial sheath and flagella, and cytoplasmic clearance [9][10][11] . The transcriptomic analyses have shown that there are more than 4000 genes involved in the process of human spermatogenesis 12 . Therefore, many unknown genetic factors that affect spermatogenesis require further investigation.
According to our previous transcriptomic data, CEP70, which belongs to the centrosome protein family (CEP family), is highly expressed throughout various stages of human spermatogenesis, especially during the meiosis and deformation stages 13 . Additionally, analysis of the reported proteomic data of human sperm tails found that CEP70 is present in sperm tails 14 . As a centrosome protein, CEP70, which was first discovered in human centrosome proteomics, is located in the centrosome throughout the cell cycle and interacts with γ-tubulin by two coiled-coil domains 15,16 . Within the CEP family, studies have shown that CEP135 and CEP131 are involved in spermatogenesis 17,18 . Moreover, CEP70 participates in the extension and dynamic regulation of microtubules and interacts with histone deacetylase 6 (HDAC6) to regulate the stability of microtubules 19,20 . In zebrafish, knocking down cep70 and cep131 can cause abnormal cilia production 21 . Based on these data, we hypothesize that CEP70 may play important roles in regulating mammalian spermatogenesis and its mutation is related to male infertility.
In this study, we conducted mutation screening of patients with clinical azoospermia and found a heterozygous site of the same mutation on CEP70. To further verify, we deleted CEP70 in mice and found that it caused abnormal spermiogenesis, leading to male sterility. Finally, we used quantitative proteomic analysis to revealed that the absence of CEP70 led to a significant decrease in the proteins associated with the formation of the flagella, head, and acrosome of sperm, and the microtubule cytoskeleton.

CEP70 may play an important function in human spermatogenesis
According to the single-cell transcriptome data of germ cells 13 , CEP70 is highly expressed in various stages of human spermatogenesis, especially during the meiosis and deformation stages (Fig. 1A). To determine whether the CEP70 mutation is associated with clinical azoospermia patients, we designed 14 pairs of primers (listed in Supplementary Table S2) for the coding sequence of CEP70 to investigate the variation of it in 476 infertile patients with azoospermia and 252 men with normal fertility. After sequencing, we found that four subjects with azoospermia had the same heterozygous mutation site (c.269A>T, p. N33I), while there was no such mutation in the control group (Fig. 1B). This mutation was conserved among various species (Fig. 1C) and predicted to be deleterious and possibly damaging by SIFT and PolyPhen2 (https://asia.ensembl.org/Multi/Tools/VEP). Through the clinical pathology report, it was found that two of the four cases had Sertoli cell-only syndrome (SCO), which caused azoospermia (Fig. 1D). However, there were no homozygous mutations or compound heterozygous mutations were found in the azoospermia patients, indicating that the recessive CEP70 may just relate to azoospermia and the deleterious heterozygous mutation of CEP70 may increase the risk of azoospermia.

Knockout of Cep70 causes spermatogenesis disorders and male infertility
A previous study of single-cell transcriptome data during spermatogenesis showed that Cep70 mRNA was highly expressed during the spermatogenesis of mice 22 ( Fig. 2A, up). Then, we sorted mouse spermatogenic cells at different stages by FACS and detected CEP70 protein expression in these samples by western blot. The results showed that CEP70 was highly expressed in spermatocytes and round spermatids ( Fig. 2A, down).
To study the biological functions of CEP70, we used CRISPR/Cas9 technology to knockout the gene in mice (Fig. 2B). We obtained a mouse strain with two nucleotides inserted in exon five of the Cep70 gene, which is located downstream of the translation start site. This insertion caused a reading frameshift and created a Fig. 2 Cep70 knockout caused male mice infertility. A (Up) Relative mRNA levels of Cep70 in mouse spermatocytes. L: leptotene; Z: zygotene; eP: early pachytene; mP: middle pachytene; lP: late pachytene; D: diplotene; MI: metaphase I; MII: metaphase II; RS1o2: steps 1-2 spermatids; RS3o4: steps 3-4 spermatids; RS5o6: steps 5-6 spermatids; RS7o8: steps 7-8 spermatids. (Down) The expression pattern of protein and mRNA of CEP70 at different staged germ cells isolated by FACS during spermatogenesis in mice. LZ: leptotene and zygotene stage; PD: pachytene and diplotene stage; MII: meiosis II spermatocytes; RS: round spermatids. B Schematic diagram of generating Cep70-knockout mouse model (Cep70 −/− ) by CRISPR/ Cas9 system. The genotype analysis results by Sanger sequencing showed a frameshift variant (+1/+1 bp) of Cep70 −/− mice. C Western blotting showed that CEP70 was deleted in 10-week-old Cep70 −/− whole testis lysates compared to wild type. β-Actin was used as the loading control. D Representative images of testis and cauda epididymidis from WT and Cep70 −/− mice. The testis of Cep70 −/− was smaller than that of the WT, the cauda epididymis of Cep70 −/− was more transparent than that of the WT. (E) Testis weight to body weight ratio of WT and Cep70 −/− mice at 10weeks (n = 6). Data are presented as the mean ± SEM. p < 0.05 (*), 0.01 (**), or 0.001 (***). F Number of pups per litter from male mice (>8-weeks old) naturally crossed with WT female mice (>6-weeks old) for 6 months. Cep70 −/− male mice showed complete sterility. Data are presented as the mean ± SD, n = 5, p < 0.001 (***). G Histological analysis of the testis and epididymis seminiferous tubules from the WT and Cep70 −/− mice. No elongated spermatids are present in Cep70 −/− mice. Scale bar: 100 μm (left); 200 μm (middle); 50 μm (right). premature stop codon (Fig. 2B). The deletion of CEP70 was detected by Western blot, and the CEP70 band was only detected in WT testis lysate (Fig. 2C), indicating that the CEP70 protein was successfully depleted in Cep70 −/− (KO) mice. The growth and healthy of Cep70 −/− mice were normal, and the Cep70 −/− females have normal fertility. However, Cep70 −/− males were completely sterile. Compared to WT male mice, the testis size of Cep70 −/− male mice was small, and the testes weight was significantly reduced (107.83 ± 5.36 mg versus 68.92 ± 8.45 mg), as shown in Fig. 2D, E. In a 6-months fertility test, the Cep70 −/− male mice that mated with WT females had no offspring (Fig. 2F). Histological examination showed that the seminiferous tubules of testes and the epididymis of Cep70 −/− male mice were devoid of elongated spermatids. However, both spermatocytes and round spermatids were present in the Cep70 −/− male mice testes (Fig. 2G). These results suggest that CEP70 plays an essential role in spermatogenesis.

CEP70 is dispensable to the prophase of meiosis I during spermatogenesis
To determine whether CEP70 deletion affects the number of germ cells, including spermatocytes, we performed immunofluorescence staining of MVH, a germcell marker, to characterize the first wave of spermatogenesis in mice at 9 days postpartum (dpp), 12 dpp, and 14 dpp that represent the generation time of leptotene, zygotene, and pachytene spermatocytes, respectively 23 . The results indicated that the numbers of germ cells had no significant differences in the first wave of spermatogenesis between WT and Cep70 −/− male mice, suggesting that the prophase of meiosis I was not affected in Cep70-deficient mice ( Supplementary Fig. S1A-D). To further confirm this result, we identified different stages of meiotic prophase I by staining of the testes section for phosphorylated H2AX (γH2AX), a marker of unrepaired DNA lesions and the sex body in pachynema, and SYCP3, a component of the synaptonemal complex. The results showed that all stages existed in the spermatocyte nuclei of WT and Cep70 −/− mice testes (Fig. 3A). To determine whether the ratios of spermatocytes at leptotene, zygotene, pachytene, and diplotene stage were abnormal, we performed spermatocyte spreading of WT and Cep70 −/− mice testicular tissues (Fig. 3B). The statistical results showed that the lack of CEP70 did not cause abnormal spermatocyte development at all four stages (Fig. 3C). Additionally, we examined the chromosomal synapsis process during the prophase of meiosis I by staining for SYCP3 and SYCP1, important components of the synaptonemal complex. We found that the signals of SYCP1 were similar between WT and Cep70 −/− mice at pachytene stage ( Supplementary Fig. S1E). These results demonstrated that CEP70 deletion had no effects on the prophase of meiosis I.

CEP70 deletion results in germ-cell apoptosis and abnormal spermiogenesis
Although the number of germ cells at 9 dpp, 12 dpp, and 14 dpp of the WT and Cep70 −/− male mice testis were not affected, the results of MVH staining showed that the loss of CEP70 caused a significant decrease in the number of germ cells at 10-weeks (Fig. 4A, B), probably because the germ cells were undergoing apoptosis. The TUNEL assay results showed that the apoptosis level of germ cells was significantly increased in Cep70 −/− mice testes (Fig. 4C). To identify exactly which stage of spermatogenesis was affected in Cep70-deficient mice, we performed PAS staining to distinguish the 12 stages (I-XII) of spermatogenesis. The results showed that spermatogenesis of Cep70-deficient mice was blocked at stages VII and VIII, and most of the round spermatids could not be transformed into elongated spermatids (Fig.  4D). According to previous reports, Sertoli cells play an important role in the stage of spermiogenesis 24 . In order to determine whether Cep70 deletion affects the development and localization of Sertoli cells, SOX9 (a marker of Sertoli cells) staining results showed that there was no difference in the number and location of Sertoli cells in the testes of WT and Cep70 −/− male mice ( fig. S2). Together, these findings suggest that the absence of CEP70 might lead to abnormal spermiogenesis in male mice and ultimately leads to infertility.

Loss of CEP70 function results in abnormal formation of flagella and acrosomes
To further determine whether there was an abnormality in the process of spermatid head shaping, SEM and TEM were performed. The results of SEM showed that the mature sperm of WT male mice had a complete structure, including obvious equatorial segments, a post-acrosomal sheath, a ventral spur, and a sharp hook rim (Fig. 5A). Although there were spermatozoa with flagella in Cep70 −/− mice epididymis, the tail and head were abnormal compared to WT mice ( Fig. 5B-E). Sperm head defects showed microcephaly and irregular shapes, and tail deformities showed bending, coiling, wrinkling, and shortening. Furthermore, we found that the sperm head of Cep70 −/− mice also had an abnormal acrosome structure, which suggested that the deletion of CEP70 might also affect the acrosome biogenesis. To verify this hypothesis, we conducted a TEM analysis of the testes. TEM results showed that the Golgi, cap and maturation phases in WT mice testes could be identified by their standard characteristics ( Fig. 5F, H, J). However, multiple acrosomal structures could be found in the Golgi phase of Cep70 −/− mice testes (Fig. 5G). Additionally, the vacuolated or irregularly shaped acrosomes were detected in the cap and maturation phases of Cep70 −/− mice spermatids (Fig. 5I,  K). Moreover, we found that WT and Cep70 −/− testis sperm flagellar cross-section (principal and end piece) of each structure were not different, according to the TEM analysis ( Supplementary Fig. S3). This result demonstrated that the absence of CEP70 did not affect the ultrastructure of sperm flagellar.
To further confirm the defects in acrosome formation in Cep70 −/− mice, we performed immunofluorescence staining on testis sections with FITC-conjugated peanut agglutinin (PNA), a marker of the outer acrosomal membrane of spermatids. As expected, the WT spermatocytes had normal acrosome at Golgi, cap and acrosome phases ( Fig. 5L-N). However, the Cep70 null spermatocytes exhibited abnormal acrosome structures (e.g., multiple acrosome centers and discrete structures) at all three acrosomal differentiating phases ( Fig. 5O-Q), which was similar to observations from TEM analysis. Together, these results indicated that the CEP70 deficient affected the process of acrosome biogenesis. In addition, we found that the total sperm, motile sperm and progressive sperm were significantly decreased by CASA ( Supplementary  Fig. S4). The data further indicate that CEP70 was involved in acrosome biogenesis and sperm flagellar formation and its loss can lead to OAT-like phenotypes and further cause male infertility.

CEP70 regulates key proteins of acrosome and flagella formation in spermatogenesis
To investigate the molecular mechanism of CEP70 in the development of acrosome biogenesis and sperm flagellar formation, quantitative proteomic analyses of testes of WT and Cep70 −/− male mice were performed by TMT labeling, HPLC fractionation and LC-MS/MS analysis (Fig. 6A). The fold change (FC) in the amount of differential expression exceeding 1.3, and less than 1/1.3 (KO/ WT) was regarded as the change threshold for a significant upregulation and downregulation, respectively. A total of 4608 proteins were quantified, of which 259 proteins were identified to be differentially expressed (Supplementary Table S3, p < 0.05), including 167 downregulated and 92 upregulated proteins (Fig. 6B). GO annotation and enrichment analysis of DEPs in WT and Cep70 −/− testes showed that 61 DEPs are related to the development of spermiogenesis (e.g., sperm part, acrosomal vesicle, motile cilium, sperm flagellum, 9 + 2 motile cilium, cilium, sperm fibrous sheath, and ciliary part) among the downregulated proteins (Fig. 6C). The results of the PPI network indicated that most downregulated proteins interact with each other and play important functions in spermatogenesis. In addition, the absence of CEP70 also caused the upregulated the expression of proteins related to the biological processes of ribosome and protein digestion and absorption, indicating that these two biological processes may be involved in the regulation of spermatogenesis (Fig. 6D). Furthermore, the heatmap of 73 downregulated proteins displayed sharp differences between WT and Cep70 −/− male testes and high-quality repeatability between samples in the same group (Fig. 6E).

Discussion
Our previous transcriptomic data demonstrated that CEP70 is highly expressed during meiosis and spermiogenesis, suggesting that CEP70 may play an indispensable function during spermatogenesis 13 . The human CEP70 is located at 3q22.3, with 18 exons that encodes a 597 amino acid protein (~70 kD), and the N terminus (1-326 amino acid) contains two coiled-coil domains, and the C terminus (327-597 amino acid) contains a tetratricopeptide repeat (TPR) domain according to the database of UniProt (https:// www.uniprot.org/uniprot/Q8NHQ1#structure). In this study, we investigated the potential role and functional importance of CEP70 in the regulation of acrosome biogenesis and sperm flagellar formation during spermatogenesis by generating Cep70 −/− mice using CRISPR/Cas9 technology. Cep70-deficient male mice show complete infertility and severely impaired spermatogenesis. To the best of our knowledge, this is the first report on the involvement of CEP70 in the regulation of spermatogenesis.
By comparing and analyzing the exon sequences of CEP70 in azoospermia patients and normal males, we found that four subjects had the same heterozygous mutation site, which was predicted to be a deleterious mutation. Unfortunately, there were no homozygous mutations or compound heterozygous mutations were found in the azoospermia patients, indicating that the recessive CEP70 may just relate to azoospermia and the deleterious heterozygous mutation of CEP70 may increase the risk of azoospermia. Therefore, we hypothesized that CEP70 is involved in the regulation of spermatogenesis, and used CRISPR/Cas9 technology to construct a global knockout model of Cep70 to study its biological functions. By analyzing the histological changes of the testis and epididymis, we found that CEP70 deficiency resulted in impaired spermatogenesis. Fertility testing revealed that the absence of CEP70 caused complete infertility in male mice.
Then, we found that the loss of CEP70 function caused the spermatogenesis to be blocked at stages VII and VIII, leading to no sperm flagella formation through PAS staining. The concentration of sperm in the epididymis and the number of motile sperm in Cep70 −/− mice was comparable to that in WT mice by CASA, and there was no motility sperm in the epididymis of Cep70 −/− mice. The main manifestation of oligospermia is a significant decrease in the number of spermatids due to the apoptosis of germ cells or the failure of sperm release from the seminiferous epithelium 44 . The TUNEL assay showed that germ cells had obvious apoptosis in Cep70 −/− mice. By comparing the testicular proteome data of WT and Cep70 −/− mice, the loss of CEP70 caused a significant downregulation of the expression of DYNLL1, leading to the high expression of pro-apoptotic proteins that cause germ-cell apoptosis 45,46 . More remarkably, through SEM analysis we found that sperm lacking CEP70 had abnormal acrosomes and irregular tails; thus, the poor progressive motility in Cep70 −/− spermatozoa could be attributed to the abnormal head and flagellar morphology. An earlier study showed that the zebrafish homolog of CEP70 can promote cilia assembly by controlling the length of the axoneme in zebrafish embryos 21 . Through this series of phenotypic analyses, we determined that the loss of CEP70 caused the OAT phenotype (low spermatozoa count, poor sperm motility, and abnormal sperm head structure) of male mice 47,48 . OAT, partially caused by defects in spermiogenesis, is the most common clinical profile in infertile men, including oligozoospermia, asthenozoospermia, and teratozoospermia 49,50 . However, non-obstructive azoospermia (NOA) is caused by a critical failure in the early stage of spermatogenesis (abnormal sperm self-renewal or meiotic arrest), which is different from OAT symptom 49 . Overall, we determined that CEP70 participates in the regulation of spermiogenesis, and the genetic alteration of Cep70 might related to OAT phenotype. According to the results of TEM and PNA immunofluorescence staining, CEP70 plays an important role in acrosome formation. Analysis of the proteomic data revealed that the components of the acrosome matrix (ZP3R and ZAN) in Cep70 −/− mice were significantly downregulated compared to WT mice 51 . After capacitation, ZP3R (also named SP56), a zona pellucida-binding protein, is located on the sperm surface and released at the time of the fusion of outer acrosomal and plasma membranes 52 . Additionally, the expression of another two membrane proteins (IZUMO1 and SPACA1) required for head shaping and oocyte fusion in Cep70 −/− mice testes were significantly decreased compared to WT testes 53,54 . Therefore, CEP70 plays an important regulatory function in the process of acrosome formation.
Additionally, 61 downregulated proteins associated with the formation of sperm tail according to GO annotation were identified in the Cep70 −/− mice compared with the WT mice. Among them, it has been reported that the deletion of some genes (Spata6, Ropn1, and Cabyr) in a mouse model showed abnormalities and deformities of the sperm tail structure 55 . Studies have shown that the ROPN1 and CABYR expression level decreased remarkably in asthenospermia, and they can bind to AKAP3 to regulate the PKA signaling pathway and affect sperm flagella formation 28,56 . The depletion of spermatogenesis associated 6 (SPATA6) disrupts the proper formation of the sperm connecting piece and head-tail conjunction 57 . Furthermore, testis-specific serine/threonine-protein kinase 1, 2, and 4 (TSSK1, 2, and 4) regulate the development of sperm flagellum 58,59 . In our analysis of the proteomics data, it was found that the expression of ROPN1, CABYR, SPATA6, TSSK1, TSSK2, TSSK4, and AKAP3 decreased significantly after CEP70 deficiency, so the sperm of Cep70 −/− mice showed a phenotype of flagella formation failure. Moreover, the mutation of CAPZA3 can cause male infertility and show an OAT-like phenotype, which may also be one of the important reasons for the OAT phenotype caused by CEP70 deletion 60 .
In summary, our genetic and functional data, based on human subjects and mouse models, strongly suggest that heterozygous deleterious mutations of CEP70 may be a novel genetic cause of OAT. Additionally, the underlying mechanism that CEP70 regulates sperm flagellar development was elucidated in this study. The screening of the deleterious mutations of CEP70 could be important for clinical molecular diagnosis of male infertility. As the findings about CEP70 in this study are novel, further research is required on this unfamiliar gene to explore the importance of CEP70 in male fertility.

Mutation screening and sanger sequencing
To scan for CEP70 mutations in infertile male patients, we used DNA samples collected from 476 infertile patients with azoospermia and 252 men with normal fertility, from the our reproductive Center between 2013 and 2016 to perform CEP70 exon amplification. The collection of these samples was approved in 2012 by the Reproductive Medicine Ethics Committee of Peking University Third Hospital (Item number: 2012sz018). The OneTaq® Hot Start Quick-Load® 2X Master Mix (New England Biolabs, M0488S) was used for PCR amplification, and the 14 coding exon sequences of CEP70 were analyzed by Sanger sequencing using an ABI 3730xl DNA sequencer (Applied Biosystems™) by BeiJing Ruibio Biotech Co., Ltd. The sequencing results were compared with the reference genome sequence of CEP70. The primers used for the PCR are listed in Supplementary Table S2.

Generation of the Cep70-knockout mouse model
According to experimental procedures similar to those reported previously 61,62 , a Cep70 (NCBI: NM_023873.4) knockout mouse model (termed Cep70 −/− ) with the frameshift variant was obtained using the CRISPR/ Cas9 system, as shown in Fig. 1A. Two pairs of singleguide RNAs (sgRNAs) were designed for Cep70 on exon five. Sanger sequencing was used to identify the genotype of the founder mouse and its offspring with a frameshift variant. The sequence information of the sgRNAs and genotyping primers are listed in Supplementary Table S1. Mice, with a C57BL/6J genetic background, were maintained under controlled temperature (20-22°C), appropriate humidity (50-70%), lighting conditions (12/12 h light/dark cycle), and food and water provided ad libitum. The Cep70-knockout mice used in the experiment were 8-weeks of age or older. All experimental procedures and animal care were approved by the Animal Care and Use Committee of the Peking University Health Science Center. During the experiment, the mice in the WT group and the KO group were randomly selected without any subjective factors, and the number of animal samples in each experiment was guaranteed to be at least 3 or more.

Fertility assessment of mice
To assess the fertility of the Cep70 −/− mice, sexually mature male (>8-weeks old, n = 5) and female (>6-weeks old, n = 5) Cep70 −/− mice were mated with two wild-type (WT) female (>6-weeks old) and male (>8-weeks old) mice for 6 months. During this period, the two WT female mice were replaced every other gestation cycle. Meanwhile, WT female (>6-weeks old) and male mice (>8-weeks old) were mated 1:2, as the control group (n = 5). Litter sizes were recorded during the fertility assessment of Cep70 −/− and WT mice.

Meiotic nuclear spreading and immunofluorescence staining
Testes were collected from adult male mice, and spermatocyte surface spreading were made as previously reported 63 . In brief, seminiferous tubules were treated with hypotonic extraction buffer (HEB, pH 8.2) for 30-60 min, and then pulverized in 100 mM sucrose buffer (pH 8.2) and gently pipetted to form a cell suspension. The suspension was loaded onto adhesive slides containing 1% PFA and 0.15% Triton X-100 fixative (pH 9.2). After 2 h of incubation, the slides were air-dried overnight. Finally, the slides were washed with PBS (pH 7.4) three times (5 min each) and immunostained with primary antibodies, including anti-SYCP3 (a gift from Hengyu Fan lab, 1:400, rat), anti-SYCP1 (Abcam, ab15090, 1:200, rabbit), and anti-γH2AX (Abcam, ab11174, 1:400, rabbit), according to the above immunofluorescence protocol.

Sperm motility assays
Sperm motility measurements were performed as previously described 64 . Briefly, the cauda epididymis was dissected from an adult mouse and cut into pieces in 1 mL of human tubal fluid (HTF), and exuded using the "swimout" method at 37°C for 30 min to collect sperm. Then, 10 μL of the exudate was put into a glass cell chamber (Leja Products BV, Nieuw-Vennep, The Netherlands) and kept on a heating platform at 37°C. The spermatozoa was further observed through a ×20 objective lens (Olympus BX51 microscope). A CCD camera (Olympus) was used to image the observation area of each chamber. Spermatozoa samples were analyzed using computer-aided sperm analysis (CASA, CEROS v.12, Hamilton Thorne Research), which were implemented using Minitube sperm visual digital semen assessment system (12500/ 1300, Minitube Group, Tiefenbach, Germany). The total number of sperm, motile sperm and the proportion of progressive sperm were analyzed.

Scanning and transmission electron microscopy
For scanning electron microscopy (SEM), epididymal sperm of WT and Cep70 −/− male mice (10-weeks old) were collected and shredded in HTF medium using the "swim-out" method at 37°C, shaken every 10 min three times. Then, the supernatant containing sperm was placed in a clean centrifuge tube and pelleted at 300 × g for 5 min. After that, the sperm was resuspended in an electron microscopy fixative (Servicebio, G1102), fixed for 2 h at room temperature, and transferred to 4°C for preservation. Sperm samples were blocked with 1% OsO4 in 0.1 M PBS (pH = 7.4) for 1-2 h at room temperature. After washing in PBS three times (15 min each), the sperm samples were gradually dehydrated in increasing concentrations of ethanol (30%, 50%, 70%, 80%, 90%, 95%, and 100% twice; 15 min each) and in isoamyl acetate (Sinaopharm Group Chemical Reagent Company, 10003128) for 15 min. After that, the samples were dried with a critical point dryer (Quorum, K850). Specimens were attached to metallic stubs using carbon stickers and sputter-coated with gold for 30 s using a Lon sputtering apparatus (Hitachi, MC1000). Finally, the samples were imaged using a scanning electron microscope (Hitachi, SU8100) by the Wuhan Servicebio Technology Company.
For transmission electron microscopy (TEM), the size of testes should be no more than 1 mm 3 . The testes samples were fixed in an electron microscopy fixative at room temperature for 1 h and then transferred into an Eppendorf (EP) tube with fresh TEM fixative for further fixation. Then, the tissues were washed with 0.1 M PBS (pH 7.4) three times for 15 min each. The tissues were fixed with 1% OsO4 in PBS for 2 h at room temperature, and then rinsed in PBS three times for 15 min each. The samples were gradually dehydrated in increasing concentrations of ethanol (30%, 50%, 70%, 80%, 95%, and 100% twice; 20 min each) and in 2 Â acetone (Sinaopharm Group Chemical Reagent Co. LTD, 10000418) for 15 min. The samples were then infiltrated and embedded in resin according to the following steps: acetone: EMBed 812 (SPI, 90529-77-4) = 1:1 for 2-4 h at 37°C; acetone: EMBed 812 = 1:2 overnight at 37°C; and pure EMBed 812 for 5-8 h at 37°C. After that, pure EMBed 812 was poured into the embedding model, the tissues were inserted into the pure EMBed 812, and then kept at 37°C overnight. The embedding models with resin and samples were moved into a 65°C oven to polymerize for more than 48 h. Then, the resin blocks were removed from the embedding models for standby application at room temperature. The resin blocks were cut to 60-80 nm thickness using an ultra-microtome (Leica, Leica UC7), and the tissues were fished onto 150 mesh cuprum grids with formvar film. The copper mesh was stained in the dark with 2% uranium acetate saturated alcohol solution for 8 min, rinsed in 70% ethanol three times, and then rinsed in ultrapure water three times, 2.6% lead citrate was used to avoid CO2 staining for 8 min, and then rinsed with ultrapure water three times. After drying with filter paper, the cuprum grids were placed into the grid board and dried overnight at room temperature. The cuprum grids were observed under a transmission electron microscope (Hitachi, HT7800) and images were taken by the Wuhan Servicebio Technology Company.

TUNEL assays
Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay was carried out according to the manufacturer instructions of In Situ Cell Death Detection Kit (Roche, 11684817910). Images were captured using a laser scanning confocal microscope (3D HISTECH, Pannoramic MIDI).
Proteomics analysis of testes of WT and Cep70 −/− male mice After protein extraction and trypsin digestion of testis of WT (n = 3) and Cep70 −/− (n = 3) male mice, the peptide was desalted by Strata X C18 SPE column (Phenomenex) and vacuum-dried. The peptide was reconstituted in 0.5 M TEAB and processed according to the manufacturer protocol for tandem mass tag (TMT) kit/isobaric tags for relative and absolute quantitation (iTRAQ) kit. Briefly, one unit of TMT/iTRAQ reagent were thawed and reconstituted in acetonitrile. The peptide mixtures were then incubated for 2 h at room temperature and pooled, desalted, and dried by vacuum centrifugation. After TMT/ iTRAQ labeling, the tryptic peptides were fractionated into fractions by high pH reverse-phase high-performance liquid chromatography (HPLC) using Thermo Betasil C18 column (5 μm particles, 10 mm ID, 250 mm length). After HPLC fractionation, the tryptic peptides were analyzed by LC-tandem mass spectrometry (MS/MS). The resulting MS/MS data were processed using Maxquant search engine (v.1.5.2.8). False discovery rate (FDR) was adjusted to <1% and minimum score for modified peptides was set >40.
According to the data quality control and search, a Gene Ontology (GO) annotation proteome was derived from the UniProt-GOA database (http://www.ebi.ac.uk/ GOA/). Then proteins were classified by GO annotation based on three categories: biological process, cellular component and molecular function. Identified proteins domain functional description were annotated by Inter-ProScan (a sequence analysis application, http://www.ebi. ac.uk/interpro/) based on protein sequence alignment method, and the InterPro domain database was used. The Kyoto Encyclopedia of Genes and Genomes (KEGG) database was used to annotate protein pathway. For functional enrichment of GO, pathway, and protein domain, a two-tailed Fisher's exact test of was employed to test the enrichment of the differentially expressed proteins (DEPs) against all identified proteins based on GO annotation, the KEGG database, and the Interpro domain database. A corrected p-value < 0.05 was considered as significant.
Further hierarchical clustering was based on DEP functional classifications (such as GO, domain, pathway, and complex). We first collated all the categories obtained after enrichment along with their p-values, and then filtered for those categories which were at least enriched in one of the clusters with a p-value < 0.05. This filtered pvalue matrix was transformed by the function x = −log10 (p-value). Finally, the x values were z-transformed for each functional category. These z-scores were then clustered by one-way hierarchical clustering (Euclidean distance and average linkage clustering) in Genesis. Cluster membership was visualized by a heatmap using the "heatmap.2" function from the "gplots" R-package.
Protein-protein interaction (PPI) networks of all the DEPs, database accession, or sequence, were searched against the STRING database (version 10.1). Only interactions between the proteins belonging to the searched data set were selected, thereby excluding external candidates. STRING defines a metric called "confidence score" to define interaction confidence; we fetched all interactions that had a confidence score ≥ 0.7 (high confidence). The interaction network from STRING was visualized in the R-package "networkD3." The TMT proteomics analysis in our research was supported by Jingjie PTM BioLabs.

Parallel reaction monitoring (PRM)
The trypsin-digested peptides of WT and Cep70 −/− male mice testes were dissolved in an aqueous solution containing 0.1% formic acid and 2% acetonitrile and then subjected to gradient treatment with an aqueous solution containing 0.1% formic acid and 90% acetonitrile, all at a constant flow rate of 500 nL/min on an EASY-nLC 1000 ultra-performance liquid chromatography (UPLC) system (ThermoFisher Scientific). After treatment, the peptides were subjected to a Nanospray ionization (NSI) source followed by MS/MS in a Q ExactiveTM Plus (Thermo-Fisher Scientific) coupled online to the UPLC. The ion source voltage was set to 2.1 kV, and the peptide precursor ions and their secondary fragments were detected and analyzed by high-resolution Orbitrap. The m/z scan range was 350-1000 for the full MS at a resolution of ×70,000, and peptides were then selected for MS/MS using normalized collision energy (NCE) setting of 27, and the fragments were detected in the Orbitrap at a resolution of ×17,500. The data acquisition mode uses the dataindependent acquisition (DIA) program that alternated between one MS scan followed by 20 PRM scans, and the fragmentation energy of the higher energy collisional dissociation (HCD) was set to 27. The target of automatic gain control (AGC) and maximum injection time (Max IT) for full MS and MS/MS were set at 3E6/50 ms and 1E5/200 ms, respectively. The isolation window for MS/ MS was set at 1.6 m/z. PRM data were manually curated within Skyline (version 3.6) 65 . The peptide parameters were set as follows: enzyme was set as Trypsin [KR/P], and max missed cleavage was set as 2 and the peptide length was set as 7-25 amino acid residues, and variable modification was set as alkylation on cysteine. The transition parameters were set as follows: precursor charges were set as 2, 3, ion charges were set as 1, ion types were set as b, y. The product ions were set from ion 3 to last ion, the ion match tolerance was set as 0.02 Da.

Statistical analysis
Data are presented as the mean ± standard deviation (SD) of at least three independent replicates, and error bars indicate SD. Statistical analyses were performed using a two-tailed Student's t-test (n ≥ 3). Differences were considered significant at p < 0.05 (*), p < 0.01 (**), and p < 0.001 (***).