IQUB mutation induces radial spoke 1 deficiency causing asthenozoospermia with normal sperm morphology in humans and mice

Abstract

Background

Asthenozoospermia (ASZ) accounts for about 20-40% of male infertility, and genetic factors, contributing to 30-40% of the causes of ASZ, still need further exploration. Radial spokes (RSs), a T-shaped macromolecular complex, connect the peripheral doublet microtubules (DMTs) to a central pair (CP), forming a CP-RS-DMT structure to regulate the beat frequency and amplitude of sperm flagella. To date, many components of RSs and their functions in human sperm flagella remain unclear.

Methods

We recruited a cohort of 323 infertile males with ASZ between August 2019 and June 2024. Genetic mutations were identified by whole-exome sequencing. Computer-aided sperm analysis, Papanicolaou staining, and electron microscopy were applied to evaluate the motility, morphology, and ultrastructure of spermatozoa, respectively. Protein mass spectrometry, western blotting, and bioinformatic analyses were performed to identify critical components of mammalian RS1 to model its structure and explore the pathological mechanism of IQUB deficiency. Intracytoplasmic sperm injection (ICSI) was applied for the patient and Iqub^−/− mice.

Results

We identified a novel homozygous IQUB mutation [c.842del (p.L281Pfs*28)] in an ASZ male with normal sperm morphology (ANM), which resulted in the complete loss of IQUB in sperm flagella. Deficiency of RS1, but not RS2 or RS3, was observed in both IQUB^842del patient and Iqub^−/− mice, and resulted in the reduction of sperm kinetic parameters, indicating the critical role of IQUB in regulating mammalian RS1 assembly and sperm flagellar beat. More importantly, we identified twelve critical components of RS1 in humans and mice, among which RSPH3, RSPH6A, RSPH9 and DYDC1 constituting the head, DYDC1, NME5, DNAJB13 and PPIL6 assembling into the head-neck complex, AK8, ROPN1L, RSPH14, DYNLL1, and IQUB forming the stalk of RS1. Along with the RS1 defect, the IQUB deficiency caused significant down-regulation of the inner dynein arms of DNAH7 and DNAH12, highlighting their nearby location with RS1. Finally, ICSI can effectively resolve the male infertility caused by IQUB genetic defects.

Conclusions

We demonstrate that IQUB may serve as an adapter for sperm flagellar RS1 in both humans and mice and consolidated the causal relationship between IQUB genetic mutations and ANM, further enriching the genetic spectrum of male infertility.

Background

Infertility affects approximately 17.5% of adults worldwide, and in the past decade, the infertility rate among Chinese couples of childbearing-age has sharply increased from 12 to 18% [1, 2]. Male factors, including abnormalities in sperm count, motility and morphology, account for about half of all causes of infertility, known as azoospermia, oligozoospermia, asthenozoospermia, teratozoospermia, respectively [3, 4]. Asthenozoospermia (ASZ), defined as sperm progressive motility less than 32%, accounts for about 20-40% of male infertility [5]. A growing number of researches on the etiology of ASZ emphasize the crucial role of genetic mutations, but many pathogenic mechanisms remain to be explored [6].

Sperm flagella is a highly organized organelle with “9 + 2” arranged microtubular axoneme [7]. In cross-section, nine peripheral doublet microtubules (DMTs) are linked together by the nexin-dynein regulatory complex and connected to a central pair (CP) via radial spokes (RSs, RS1-3) [8, 9]. On the longitudinal section, the DMTs are organized regularly with a 96-nm-long repeat unit with seven subtypes of inner dynein arms (IDAa, b, c, d, e, f/I1, and g), four outer dynein arms, and three RSs [10,11,12,13]. As a T-shaped macromolecular complex, the stalk of RS is anchored near the IDAs on DMTs with its head pointing towards CP, forming a CP-RS-DMT structure to regulate the beat amplitude and frequency of flagella [14,15,16,17]. So far, at least 23 distinct polypeptides have been identified in Chlamydomonas reinhardtii, known as RS proteins (RSP1 to RSP23) [18], whereas the components of mouse sperm flagellar RS1 were identified recently [19] and those of human sperm remain unclear. Genetic defects of RSPs induced immobility with severely reduced bend amplitude in Chlamydomonas reinhardtii [20, 21]. In humans and mice, genetic mutations in RS-related genes, such as RSPH1, RSPH3, RSPH6A, or RSPH9, induce primary ciliary dyskinesia or ASZ with normal sperm morphology (ANM) or multiple morphological abnormalities of sperm flagella (MMAF) [22,23,24]. The genetic etiology of MMAF is more easily identified clinically due to significant changes in sperm morphology, while the genetic etiology of ANM is easily overlooked, leading to inappropriate treatment. Therefore, identifying the genes encoding RS components and their relationship with ANM is crucial, especially in human sperm flagella.

IQUB, a protein specifically expressed in the mouse testis containing an IQ motif and a ubiquitin-like domain [25], was found to be the ortholog of FAP253, a unique component of Chlamydomonas reinhardtii RS1 [26], and CMUB116, a protein residing with the RS stalks in Ciona intestinalis [27]. A recent study identified the mutation of IQUB gene in an infertile male with ASZ [19]. Additionally, Zhang et al. revealed that IQUB protein was localized to the RS1 stalk of mouse sperm flagella and the Iqub^−/− male mice were infertile as a significant decrease in sperm motility induced by RS1 deficiency [28]. These findings hint that IQUB may be a causative gene for ASZ, but more evidence is still needed to support the gene-disease relationship between IQUB gene mutations and ASZ, particularly concerning the components of human sperm flagellar RS1 and the outcomes of assisted reproductive technology.

Here, we investigated a cohort of 323 primary infertile Chinese males with whole exome sequencing (WES) and identified a novel homozygous IQUB mutation [c.842del (p.L281Pfs*28)] in a man suffered from ANM. The Iqub^−/− male mice recapitulated the infertile phenotype of humans, characterized by normal spermatogenesis with ANM. Transmission electron microscopy (TEM) revealed the absence of RS1 in the IQUB^842del and Iqub^−/− sperm flagella. Subsequently, protein mass spectrometry combining with bioinformatic analyses identified critical components in humans and mice to model the structures of RS1, such as RSPH14, RSPH3, RSPH6A, RSPH9, DYDC1, DNAJB13, DYNLL1, PPIL6, ROPN1L, AK8, NME5 and IQUB. Along with the RS1 defects, IQUB deficiency caused significant down-regulation of the IDAs of DNAH7 and DNAH12. Intracytoplasmic sperm injection (ICSI) assisted the IQUB^842del patient to achieve a good fertility outcome. Our study consolidated the gene-disease relationship between IQUB genetic mutations and ASZ, further enriching the genetic spectrum of male infertility.

Materials and methods

Study participants and ethical approval

A total of 323 Chinese infertile men with ASZ were recruited from August 2019 to June 2024 at the Reproductive Medicine Center of the Women’s and Children’s Hospital of Chongqing Medical University. All somatic chromosome karyotypes were normal (46,XY), and no Y-chromosome microdeletions were found. In this study, the ethical approval (No.: (2023) Ethics Review (Research) 030) was obtained from the Ethics Committee of Chongqing Health Center for Women and Children. An informed consent form was signed before the collection of peripheral blood and semen.

WES, Sanger sequencing and mutation analysis

As previously reported [29], genomic DNA was isolated from peripheral blood samples of the patients using the QIAamp DNA Blood Mini Kit (69504, QIAGEN, Dusseldorf, Germany) and then WES was performed on the Illumina HiSeq X system according to the manufacturer’s instructions. The clean reads were mapped to the human reference sequence (GRCh37) and performed functional annotation using ANNOVAR. Next, PolyPhen-2, SIFT, MutationTaster, and CADD were used for functional prediction. The mutations identified by WES were validated by Sanger sequencing with specific PCR primers (Table S1).

Semen analysis and papanicolaou staining

Semen samples were collected via masturbation from patients after 2–7 days of abstinence and incubated at 37℃ for 30 min for liquefaction. Spermatozoa of mice were released from cauda epididymides and incubated in human tubal fluid media (MR-070, Sigma-Aldrich, Mo, USA) at 37℃ for 30 min. Next, a computer-assisted sperm analysis system (CASA, GSA-180, Jiangsu Rich Life Science Instrument Co.Ltd., Nanjing, China) was applied according to the 6th edition of the WHO guidelines [30]. For sperm morphology, spermatozoa were smeared on slides and stained with Papanicolaou staining according to the manufacturer’s protocol (Cariad Medical Technology Co.Ltd., Zhuhai, China) [29]. For flagellar waveform tracing, single frames of sperm high-speed microscopy video (120 fps) were exported and the flagellar waveforms with each ten frames were traced using Adobe Photoshop software.

Animal model and reverse transcription PCR (RT-PCR)

The animal experiments were approved by the Experimental Animal Management and Ethics Committee of Chongqing Health Center for Women and Children (No.: 2023028). All the mice were housed in a pathogen-free environment with a room temperature range of 20℃ to 24℃ under a 12 h light/dark cycle, 50-70% humidity, and free access to food and water. For gene editing, 4-week-old C57BL/6 female mice were super-ovulated and mated to collect zygotes. Then the zygotes were microinjected with CRISPR/Cas9 mRNA and single-guide RNAs (sgRNAs) targeting the exons 2–6 of Iqub (ENSMUST00000052277.5), which were predicted to affect all the transcripts of the IQUB gene as well as the functional domains of the IQUB protein, and cultured in K + simplex optimized medium (KSOM, MR-106, Sigma-Aldrich, MO, USA) to reach 2-cell stage. The embryos were transferred into the oviducts of pseudopregnant female mice to obtain positive F0 mice, which were bred to obtain stable F1 generation. For RT-PCR, total RNA of mouse tissues was extracted with MiniBEST universal RNA extraction Kit (9767, Takara, Osaka, Japan) and converted to complementary DNA (cDNA) using PrimeScript™ RT reagent Kit (RR047A, Takara, Osaka, Japan) according to the manufacturer’s instructions. The sequence of primers used in genotype identification and RT-PCR and sgRNAs were listed in Table S1.

Electron microscopy

Scanning electron microscopy (SEM) and TEM were performed as previously described [29]. Briefly, semen samples from humans and mice were collected, washed, and fixed in 2.5% phosphate-buffered glutaraldehyde at 4 ℃ overnight. For SEM, the samples were dehydrated using progressive ethanol concentrations (35%, 50%, 75%, 90%, 95%, and 100%), dried using a CO₂ critical-point dryer (Eiko HCP-2, Hitachi, Tokyo, Japan), sputter coated by an ionic sprayer meter (ACE200, Leica), and finally observed under the SEM (Nova NanoSEM 450, FEI, Hillsboro, OR, USA) at an accelerating voltage of 5 kV. For TEM, the samples were post fixed in 1% buffered OsO₄ and embedded in Epon 812 (90529-77-4, SPI, West Chester, PA, USA) after dehydration. The ultrathin (70 nm) sections were double stained with lead citrate and uranyl acetate, and observed by a TEM (TECNAI-10, Philips, Eindhoven, The Netherlands) at an accelerating voltage of 80 kV. On the longitudinal sections of TEM, we measured and plotted the gray values of two or three 96-nm-repeats axonemal units through ImageJ software.

Immunofluorescence and Hematoxylin and eosin (H&E) staining

Spermatozoa were washed, fixed in 4% paraformaldehyde, then permeabilized with 0.5% Triton X-100 (X100, Sigma-Aldrich, MO, USA), blocked with 3% bovine serum albumin (BSA, B2064, Sigma-Aldrich, MO, USA), and finally incubated with primary antibodies overnight at 4℃, respectively. On next day, the samples were washed and incubated with the secondary antibodies, and the nuclei were labeled with 4′,6-diamidino-2-phenylindole (DAPI, P0131, Beyotime, Shanghai, China). Finally, images were captured under a confocal laser scanning microscope (TCS SP8, Leica, Wetzlar, Germany). The used antibodies were listed in Table S2. Testicular and epididymal tissues from 9-week-old male mice were fixed with animal testicular tissue fixation solution (G1121, Servicebio, Hubei, China) for 24 h. After dehydration by ethanol, they were embedded in paraffin and sectioned at 5 μm. The sections were stained with H&E, and images were captured using a digital biopsy scanner (Pannoramic 250, 3DHISTECH, Hungary).

Tandem mass tag mass spectrometry (TMT-MS)

Spermatozoa samples were analyzed with a standard protocol of TMT-MS for proteomic analysis. Briefly, the sample was sonicated on ice using a high intensity ultrasonic processor (Scientz) in lysis buffer. The supernatant was collected after centrifugation, and the concentration was determined with BCA kit (P0010, Beyotime, Shanghai, China). Protein samples were trypsinized and purified, and peptides were reconstituted in 0.5 M TEAB and processed using the TMT kit according to the manufacturer’s protocol. The sample was fractionated into fractions by high pH reverse-phase HPLC using Agilent 300 Extend C18 column (5 μm particles, 4.6 mm ID, 250 mm length) and subsequently analyzed in Orbitrap Exploris 480 with a nano-electrospray ion source. The raw data were processed using Domain Annotation (InterProScan), GO Annotation (http://www.ebi.ac.uk/GOA/), and KEGG Pathway Annotation (KEGG online service tools KAAS mapper).

Co-immunoprecipitation (Co-IP) and Western blotting

HEK293T cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin at 37 °C under 5% CO₂. A full-length of Iqub, Calm, Rsph14, Ak8 and Morn3 cDNAs cloned in pCDNA3.1-FLAG or pCDNA3.1-HA vectors were constructed by Youbo Biotechnology (Zhengzhou, China). Lipofectamine™ 8000 Transfection Reagent (C0533, Beyotime, Shanghai, China) was used for transfecting and the transfected HEK293T cells or mouse sperm were lysed with RIPA buffer (P0013B, Beyotime, Shanghai, China) to obtain protein extract. For Co-IP, cell lysates were treated with IP buffer (P2181, Beyotime, Shanghai, China), and then incubated with BeyoMagTM anti-Flag magnetic beads (P2181, Beyotime, Shanghai, China) overnight at 4 °C. The beads were washed and then the immunoprecipitated proteins were eluted from the beads by FLAG peptide. Then the proteins were electrophoresed in SDS-PAGE gels and transferred to polyvinylidene difluoride membrane (1620177, BIO-RAD, CA, USA), followed by blocking with skimmed milk and subsequent incubation with primary and secondary antibodies. Finally, the enhanced chemiluminescence reagent (G2020, Wuhan, Servicebio Wuhan, China) and a chemiluminescence imaging system (FluorChem M, Protein Simple, CA, USA) were used for imaging. The related antibodies were listed in Table S2.

Radial spoke 1 (RS1) structure remodeling

We retrieved the RS1 structure of Chlamydomonas reinhardtii from the Protein Data Bank (PDB) (accession codes 7JTK and 7JTS) and performed a BLAST search to identify homologous proteins in Homo sapiens and Mus musculus. This enabled the selection of candidate proteins for homology modeling based on sequence similarity, with all homologous proteins showing over 20% sequence identity (Table S3). To predict the three-dimensional structures of RS1 in humans and mice, we used AlphaFold3 (https://alphafold.com), a state-of-the-art deep learning model known for its high accuracy in protein structure prediction [31]. For the AlphaFold3 predictions, we utilized the default settings outlined in the AlphaFold3 documentation (v3.0), focusing on the latest version. The predicted structures were then aligned, altered, and fitted using PyMol (http://www.pymol.org) to optimize the models. We assessed the reliability of the predicted structure using pIDDT scores, which represent confidence in each residue, flagging those with scores below 50 as potentially uncertain. During this process, we removed amino acid residues with pIDDT scores lower than 50, as these residues were predicted with low confidence and could contribute to structural inaccuracies. Most of the predicted models exhibited pTM scores greater than 0.5, indicating a high level of structural confidence. For reproducibility, we adhered to the guidelines provided in the AlphaFold3 GitHub repository (https://github.com/deepmind/alphafold), where the detailed protocol and all necessary configuration files for running AlphaFold3 are available. The full computational environment, including version numbers of dependencies, can be found in the supplementary materials of the referenced AlphaFold3 publication.

ICSI procedures

The female mice (ICR, 4-week-old) were intraperitoneally injected with 5 IU Pregnant Mare Serum Gonadotropin (PMSG, Ningbo Sansheng, Ningbo, China) followed with 5 IU human Chorionic Gonadotropin (hCG, Ningbo Sansheng, Ningbo, China) 48 h later for superovulation. 13–14 h after hCG administration, oocytes at metaphase II (MII) stage were collected from the oviducts and incubated in M2 medium (MR-015, Sigma-Aldrich, MO, USA). The single spermatozoa collected from cauda epididymides of wild-type or KO male mice was microinjected into a MII oocyte under a NIKON inverted microscope (Tokyo, Japan) through a Piezo system (PrimeTech, Osaka, Japan). The injected oocytes were then cultured in KSOM at 37℃ in an atmosphere of 5% CO₂ in air for development into blastocyst.

Statistical analysis

Statistical analyses were conducted using GraphPad Prism software. Statistical significance of the differences between two groups was measured by Student’s t-test with paired, two-tailed distribution. All data were presented as the mean ± SEM and P < 0.05 were considered statistically significant.

Results

Identification of a homozygous IQUB mutation in a male with ASZ

To identify genetic factors associated with male infertility, we recruited a cohort of 323 Chinese infertile men with ASZ and performed WES following a standardized workflow (Fig. S1A). A novel homozygous IQUB mutation [c.842del (p.L281Pfs*28)] was identified in a 37-year-old man (L019-II-5) from a consanguineous family who suffered from infertility for 3 years (Fig. 1A). The IQUB^842del patient has normal chromosome karyotype (46,XY), bilateral testicular size, and hormone levels without any Y chromosomal microdeletions. Sanger sequencing showed the homozygous pattern of c.842del in the proband (L019-II-5) and the wild type of his wife (L019-II-6) (Fig. 1B). This mutation was absent in the gnomAD database and predicted to be disease-caused by MutationTaster. IQUB gene (NM_178827.5) is mapped to chromosome 7q31.3 and consists of 13 exons encoding the IQUB protein with a ubiquitin-like domain at 131–207 aa and an IQ domain at 338–367 aa (Fig. 1C). The mutation c.842del was located in Exon 5 and induced the premature termination of protein translation as p.L281Pfs*28 without the IQ domain and C-terminus (Fig. 1C; Fig. S1B). Immunofluorescence staining showed that IQUB protein was located along the sperm flagella of normal control, but no IQUB signals were observed in the sperm flagella of IQUB^842del patient (Fig. 1D). Western blotting also confirmed the absence of IQUB protein in the IQUB^842del sperm without any truncated protein detected (Fig. 1E). Moreover, the expression and location of TOM20 (a marker of mitochondrial sheath), DNALI1 (a light intermediate chain protein of IDA) and DNAI1 (an intermediate chain protein of outer dynein arms) were comparable in the spermatozoa of normal control and the IQUB^842del patient (Fig. S1C-E). These findings suggested that the ASZ phenotype of the proband may attribute to the loss-of-functional mutation in IQUB independent of mitochondrial sheath or dynein arms.

Fig. 1

Identification of a IQUB^842del mutation in an infertile patient with asthenozoospermia. A Pedigree of the consanguineous family with the infertile patient. The black arrow points to the proband, the “〧” sign indicates infertility. B Sanger sequencing chromatograms of the family. The red arrow indicates mutation loci. C The position of the identified IQUB^842del mutation at the chromosome, transcript (ENST00000324698.6) and protein (Q8NA54) levels. D Immunofluorescence staining of spermatozoa from the normal control and IQUB^842del patient against β-TUBULIN (green) and IQUB (red), DNA (blue) was counterstained with DAPI. Scale bars, 5 μm. E Western blotting of spermatozoa revealed that IQUB was absent in IQUB^842del patient. Three times independently collected sperm samples from normal control and the IQUB^842del patient

IQUB deficiency causing the absence of radial spoke 1 in sperm flagella

As shown in Table 1, the semen volume, semen pH, sperm concentration, and total sperm count of the IQUB^842del patient were all normal. The sperm motility, especially the progressive motility, decreased dramatically to near 0 (Video S1 and S2). The kinetic parameters, such as curvilinear velocity, straight-line velocity, average path velocity, and beat-cross frequency, were significantly reduced. The results of Papanicolaou staining and SEM showed that sperm morphology of the IQUB^842del patient was similar to that of normal control, with an oval head and smooth flagella of appropriate length (Fig. 2A and B). The ultrastructural analysis by TEM revealed no obvious abnormalities in the head, head-tail connecting apparatus, mitochondrial sheaths, fibrous sheaths, outer dense fibers, as well as “9 + 2” structures of the IQUB^842del sperm (Fig. 2C). However, cross-sections of the sperm flagella displayed absence of RS structures in about 47.4% of the IQUB^842del patient and about 4.3% of the normal control (Fig. 2C and D). Combining TEM and gray value analysis, we found that in the longitudinal-sections, there were three RSs in each 96-nm axoneme unit of the sperm flagella from normal control, with a spacing of 32.4 nm between RS1 and RS2, 23.6 nm between RS2 and RS3, and 40.5 nm between RS3 and RS1, respectively (Fig. 2E). Intriguingly, there were only two RSs in each axoneme unit of sperm flagella from the IQUB^842del patient, and the spacing between adjacent RSs was 24 nm and 72.4 nm (Fig. 2E). Therefore, we inferred that the missing RS of the IQUB^842del patient was RS1, since missing RS1, RS2, or RS3 would result in a long gap with a length of roughly 72 nm, 56 nm, or 64 nm, respectively (Fig. S2A). Regarding the important role of RSs in controlling the beat amplitude of flagella, we used CASA and Photoshop to respectively trace the movement path and flagellar beat of sperm. As shown in Fig. 2F, the amplitude of the curvilinear path of sperm movement as well as the range of flagellar beat were evidently larger in the normal control than those in the IQUB^842del patient. These data indicated that IQUB may be critical for the RS1 assembly and sperm flagellar beat.

Table 1 Semen analysis using CASA in the patient with IQUB^842del mutation

Fig. 2

Morphological and ultrastructural evaluations of the spermatozoa from IQUB^842del patient. A and B Papanicolaou staining and scanning electron microscopy both showed normal sperm morphology of IQUB^842del patient. Scale bars, 5 μm. C Spermatozoa cross-sections of transmission electron microscopy (TEM) showed nine radial spokes (RSs) arranged in normal control and partial loss of RSs (red arrow) in IQUB^842del patient. D Percentage of cross-sections with RSs loss in normal control and IQUB^842del patient. E Spermatozoa longitudinal-sections of TEM showed that RS1-3 were arranged in 96-nm repeats in normal control, whereas in IQUB^842del patient there were only two intact RSs per 96-nm repeat without RS1. Red arrow indicates RS1, green arrow indicates RS2, and cyan arrow indicates RS3. Yellow rectangles represent the region where the gray values are measured. F Sperm movement paths and flagellar swing amplitudes in normal control and IQUB^842del patient

Characterization of the infertile phenotype and RS1 loss of Iqub ^−/− male mice

In mice, the Iqub was abundantly expressed in the testis (Fig. S2B), and initially expressed in the testis at postnatal day 14 (P14) and peaking at P56 (Fig. S2C). To further investigate the mechanism by which IQUB mutation leads to ANM and RS1 loss observed in humans, we constructed the Iqub knockout mouse model by deleting exons 2 to 6 through CRISPR/Cas9 technique (Fig. S2D). PCR and Sanger sequencing confirmed the gene editing effect (Fig. S2E and F), while immunofluorescence staining and western blotting affirmed the complete ablation of IQUB in the Iqub^−/− spermatozoa (Fig. S3A and B). The expressions of TOM20, DNALI1 and DNAI1 did not differ significantly between the Iqub^+/+ and Iqub^−/− spermatozoa (Fig. S3C and D). The adult male mice with different genotypes, including Iqub^+/+, Iqub^+/−, and Iqub^−/−, were mated with wild-type female mice for 2–3 months, respectively. Normal mounting behaviors were observed with all males, but only the Iqub^+/+ and Iqub^+/− males consistently produced offsprings whereas Iqub^−/− males were completely infertile (Fig. 3A). The fertility of Iqub^−/− female mice was not affected (Fig. S4A and B). The size of testis and epididymis and the ratio of testis weight to body weight were all normal (Fig. 3B and C), and further histological analysis with H&E staining showed no significant difference of spermatogenesis in testis and epididymis between the Iqub^+/+ and Iqub^−/− male mice (Fig. S4C and D). The total count of spermatozoa released from epididymis were comparable between the Iqub^+/+ and Iqub^−/− male mice (Fig. 3D). Nevertheless, the sperm motility, especially the progressive motility, was obviously reduced in the Iqub^−/− mice (Fig. 3E and F; Video S3 and S4).

Fig. 3

Iqub knockout causes asthenozoospermia with RS1 loss in mice. A Pup birth quantification per vaginal plug of Iqub^+/+, Iqub^+/−, and Iqub^−/− mice. No pregnancy occurred in females mated with Iqub^−/− male mice. Data are presented as means ± SEM (n = 6, *** P < 0.001). B Testes and epididymis from 9-week-old Iqub^+/+, Iqub^+/−, and Iqub^−/− mice. C The ratios of testis/body in Iqub^+/+, Iqub^+/−, and Iqub^−/− mice were not significantly different. n = 4, P > 0.05. D Sperm number obtained from cauda epididymal of Iqub^+/+ and Iqub^−/− mice. n = 4. E and F Percentage of motile spermatozoa and progressively motile spermatozoa from Iqub^+/+ and Iqub^−/− mice. n = 4, *** P < 0.001. G and H Both Papanicolaou staining and scanning electron microscopy showed normal sperm morphology of Iqub^−/− mice. Scale bars, 5 μm. I Spermatozoa cross-sections of transmission electron microscopy (TEM) showed nine radial spokes (RSs) regularly arranged in Iqub^+/+ mice and partial loss of RSs (red arrow) in Iqub^−/− mice. J Percentage of cross-sections with RSs loss in Iqub^+/+ and Iqub^−/− mice. K Spermatozoa longitudinal-sections of TEM showed that RS1-3 were arranged in 96-nm repeats in Iqub^+/+ mice, while in Iqub^−/− mice there were only two intact RSs per 96-nm repeat without RS1. Red arrow indicates RS1, green arrow indicates RS2, and cyan arrow indicates RS3. Blue rectangles represent the region where the gray values are measured. L Sperm movement paths and flagellar swing amplitudes in Iqub^+/+ and Iqub^−/− mice

Papanicolaou staining and SEM showed no obvious morphological abnormalities in sperm head and flagella of the Iqub^+/+ and Iqub^−/− mice (Fig. 3G and H). The ultrastructure of head, head-tail connecting apparatus, mitochondrial sheath, fibrous sheath, outer dense fiber, as well as “9 + 2” structure of the Iqub^−/− spermatozoa observed with TEM were all normal (Fig. 3I). As expected, the deficiency of RS structures was found in approximately 42.3% of the cross-sections from Iqub^−/− spermatozoa versus 3.9% in Iqub^+/+ spermatozoa (Fig. 3I and J). Furthermore, in the longitudinal-sections, there were three RSs (RS1-3) residing in each 96-nm axoneme unit of the Iqub^+/+ spermatozoa flagella, whereas only two in the Iqub^−/− spermatozoa (Fig. 3K). Gray value analysis found that the distances between RS1 and RS2, RS2 and RS3, and RS3 and RS1 were 32.8 nm, 24.1 nm, and 39.6 nm, respectively, in the Iqub^+/+ sperm flagella. Whereas the distances between adjacent RSs were 24.5 nm and 71.9 nm in the Iqub^−/− sperm flagella, hinting the deficiency of RS1 (Fig. 3K). Ulteriorly, both the amplitude of the curvilinear path of sperm movement and the range of flagellar beat in the Iqub^−/− mice were remarkably smaller than those in the Iqub^+/+ mice (Fig. 3L). In conclusion, the Iqub^−/− mice recapitulated the infertile phenotype of ANM in humans characterized by RS1 deficiency without affecting spermatogenesis.

Identification and modeling of the RS1 complex in humans and mice

The complete loss of RS1 in both IQUB^842del patient and Iqub^−/− mice makes them ideal samples for studying the composition and structure of RS1. TMT-MS analysis was thus performed with spermatozoa of the IQUB^842del patient and normal control as well as the Iqub^+/+ and Iqub^−/− mice to identify the differential proteins as critical components of RS1 in humans and mice (Fig. 4A). A total of 1013 up-regulated and 543 down-regulated proteins were identified in the IQUB^842del patient (Fig. S5A), while a total of 361 up-regulated and 201 down-regulated proteins were identified in the Iqub^−/− mice (Fig. S5B). Proteins involved in microtubule-based movement and motile cilium were significantly down-regulated in the IQUB^842del patient and Iqub^−/− mice (Fig. S5C and D). Firstly, we identified the homologs of RSPs belong to RS1 in Chlamydomonas reinhardtii from Homo sapiens and Mus Musculus through alignment analysis with BLAST (Table S3) [26]. Secondly, we intersected the down-regulated proteins in IQUB^842del and Iqub^−/− spermatozoa, as well as the RSPs of RS1 in Chlamydomonas reinhardtii, and identified RSPH14, RSPH3, RSPH6A, RSPH9, DYDC1, DNAJB13, DYNLL1, PPIL6, ROPN1L, AK8, NME5 and IQUB as the key components of sperm flagellar RS1 (Fig. 4B). RSPH1, CALM and CYB5D1 were observed with down-regulation in the IQUB^842del patient but not in the Iqub^−/− mice. Interestingly, DNAH7 and DNAH12, the components of IDA, were significantly reduced both in the spermatozoa of IQUB^842del patient and Iqub^−/− mice.

Fig. 4

The identification of sperm RS1 components in human and mice. A Schematic showing the procedures of comparative proteomics performed in the spermatozoa of human and mice. n = 3. B Venn diagram showing the significantly down regulated proteins involving in sperm RS1 components (yellow and white font) and inner dynein arms (IDAs, blue font) in human and mice. C and D Western blotting showed a significant decrease of RS1 proteins in the spermatozoa of IQUB^842del patient and Iqub^−/− mice. For humans, three times independently collected sperm samples from normal control and the IQUB^842del patient. For mice, three individual animals were analyzed as biological replicates. NC, normal control. E-H Co-IP revealed that IQUB interacted with CALM, RSPH14 and AK8, but not MORN3. I and J The IDAs of DNAH7 and DNAH12 were significantly decreased in the spermatozoa of IQUB^842del patient and Iqub^−/− mice. NC, normal control. Data are presented as means ± SEM (* P < 0.05, ** P < 0.01)

Western blotting further confirmed the significantly down-regulated expression of RSPH14, RSPH3, RSPH9, CALM, DNAJB13, NME5, ROPN1L, AK8 and DYNLL1 in the spermatozoa of the IQUB^842del patient and Iqub^−/− mice (Fig. 4C and D, Fig. S5E and F). Immunofluorescence staining showed the signals of RSPH14, RSPH3, RSPH9, CALM, DNAJB13, NME5, ROPN1L, AK8 and DYNLL1 were distributed along the entire flagella of the spermatozoa, consistent with the signal distribution of IQUB (Fig. S6). Moreover, we co-expressed IQUB with CALM, RSPH14, or AK8, in HEK293 cells, respectively, and corroborated the interaction between these proteins through Co-IP (Fig. 4E, F and G). However, we found no interaction between IQUB and MORN3 (Fig. 4H). Furthermore, the IDAs of DNAH7 and DNAH12 were also obviously decreased in the IQUB^842del patient and Iqub^−/− mice, meaning their nearby location with RS1 (Fig. 4I and J). To visualize the spatial relationship of RSPs identified in humans and mice, we modeled the RS1 structure referring to the reported data on RS1 of Chlamydomonas reinhardtii [26]. As shown in Fig. 5A and B, the RS1 heads of humans and mice were composed of RSPH1, RSPH3, RSPH6A, RSPH9 and DYDC1. Different from Chlamydomonas reinhardtii, the RS1 heads of humans and mice were more constricted [26]. The RSPH3 acted as a scaffolding structure extending from the head of RS1 to the base of the stalk, while DYDC1, NME5, DNAJB13 and PPIL6 constituted the head-neck complex, and AK8, CYB5D1, ROPN1L, RSPH14, DYNLL1, CALM and IQUB constituted the stalk of RS1. Taken together, IQUB may serve as an RS1 adaptor, and IQUB deficiency may result in failure of recruiting RS1 components.

Fig. 5

The RS1 models and ICSI results of human and mice. A and B The predicted RS1 complexes of human and mice, which were based on the RS1 structure of Chlamydomonas reinhardtii. C The development of two embryos from the IQUB^842del patient underwent ICSI. D The ICSI results of Iqub^+/+ and Iqub^−/− males, with no significant differences in the percentage of 2 PN (pronucleus), 2 cell, and blastocyst. Data are presented as means ± SEM, n = 3

Favorable prognosis of ICSI treatment for IQUB-related infertility

The couple affected by IQUB mutation accepted ICSI treatment in our clinic center after signing an informed consent form. As shown in Table 2, the IQUB^842del patient’s wife had normal basal hormone levels and underwent superovulation with a short-acting GnRH-agonist long protocol. Subsequently, fifteen oocytes were retrieved and fourteen MII oocytes were microinjected with the sperm from IQUB^842del patient to reach a fertilization rate, cleavage rate and usable embryo rate of 85.7% (12/14), 100% (12/12) and 33.3% (4/12), respectively (Table 2). The IQUB^842del patient did not choose to culture the usable embryos into blastocysts and the developmental process of two usable embryos was displayed in Fig. 5 C. Promisingly, the IQUB^842del patient’s wife successfully conceived and delivered twins live births after transferring two frozen-thawed embryos. After ICSI treatment, there were no significant differences in the rates of fertilization, cleavage and blastocyst formation between Iqub^+/+ and Iqub^−/− spermatozoa, which were 94.2% ± 4.1%, 92.6% ± 1.8% and 45.2% ± 1.9% for Iqub^+/+ spermatozoa and 88.4% ± 2%, 88.6% ± 2% and 42.6% ± 2.6% for Iqub^−/− spermatozoa, respectively (Fig. 5D). These data suggested that ICSI treatment is an effective strategy for the male infertility caused by IQUB mutation.

Table 2 Clinical characteristics of the IQUB^842del patient’s wife with ICSI

Collectively, we suggest that IQUB may serve as an adapter for sperm flagellar RS1 in humans and mice, and the RS1 deficiency caused by IQUB mutations can reduce flagellar swing amplitude, leading to male infertility with ANM (Fig. 6).

Fig. 6

The proposed model of IQUB mutation in male infertility. IQUB mutation induced the loss of RS1 in sperm flagella, further reduced flagella swing amplitude and led to asthenozoospermia with normal sperm morphology (ANM) in human and mice

Discussion

RSs are critical for sperm motility and fertilization, but many components of RSs and their functions are still unclear, especially in human sperm flagella. In this study, we identified an infertile man with ANM affected by IQUB mutation and constructed Iqub^−/− mice to explore the mechanism. IQUB deficiency induced ultrastructural loss of RS1 in the sperm flagella of both humans and mice, and we further identified key components of mammalian RS1 complex to model its structure. Moreover, ICSI can overcome the male infertility caused by IQUB deficiency.

Currently, positive genetics research based on collection of clinical rare cases, screening gene mutations with WES, and animal model analysis, has been an effective strategy to identify genetic factors of infertility, which has identified more than 40 pathogenic genes related to MMAF and explained 60 -70% of MMAF patients [29, 32, 33]. The CP-RS system generated the large-amplitude, asymmetric waveforms required for forward swimming in Chlamydomonas reinhardtii, with the RS serving as a mechanochemical signal transducer between the CP and IDAs [34]. Mutations of RS-related genes have been reported to be associated with primary ciliary dyskinesia, ANM, and MMAF in humans. For example, the sperm from the individuals carrying RSPH3, DNAJB13, RSPH6A and RSPH1 mutations exhibited obviously decreased motility and MMAF [22,23,24, 35]. However, the sperm from the individuals affected by AK8 mutations exhibited ANM [36], which is the same as what we found in the IQUB^842del patient. The rarity of clinical cases, as well as the heterogeneity and obscurity of clinical phenotypes, limit the research and understanding of RSs. We speculated that the normal morphology of the spermatozoa with AK8 or IQUB defects might be because AK8 and IQUB are localized uniquely to RSs especially RS1 whose deficiency can be partially compensated by RS2 and RS3. In addition to constituting RS1, DNAJB13 has also been found to be involved in the migration of sperm annulus [37], thus the functional diversity during spermatogenesis may be the reason why its defects cause severe sperm morphological abnormalities. Collectively, positive genetics research can provide theoretical basis for gene diagnosis and genetic counseling of male infertility patients in clinical practice and will promote in-depth exploration of the molecular basis of sperm flagella.

During the past few decades, cryo-ET has enabled us to gain new insights into the structures of RS in various organisms, such as Trypanosome flagella [38], Chlamydomonas reinhardtii flagella [39], mouse ependymal cilia [40, 41], mouse sperm flagella [42, 43], bovine sperm flagella [44], bovine oviduct cilia [44], porcine brain ventricle cilia [44], human respiratory cilia [11, 44] and human sperm flagella [43]. The near atomic resolution structure of RS1 in Chlamydomonas reinhardtii flagella has been illuminated and 23 components had been identified [26]. Among mammals, the structure of the murine RS head was the first to be reconstituted and was formed primarily by the polymerization of RSPH1-RSPH4a-RSPH9-RSPH3b subunits [41]. A recent study shed light on the near-atomic models of RS1, RS2 and RS3 of mouse ependymal cilia, showing that the RS1 consisted of fifteen proteins forming two-fold symmetric brake pad-shaped head, head-neck complex and stalk [40]. Among the fifteen proteins, the RSPH4a was replaced by RSPH6a in our predicted RS1 model, and the remaining 14 proteins were consistent with our model. Additionally, Leung MR et al. constructed the near atomic models of bovine sperm flagellar RS1 consisting of twenty proteins [44], which contains five more proteins, CAMK4, RGS22, EFCAB10, SPA17 and LRP2BP, than the RS1 model we constructed. The remaining fifteen proteins are consistent with the RS1 components we identified. Moreover, by performing TMT-MS and IP-MS analysis on the mice carrying Iqub variant, Zhang et al. identified the components and predicted the structure of RS1 in mouse sperm flagella [28], which is very similar to the model we constructed (Fig. 5B). Compared with the model predicted by Zhang et al., our mouse sperm RS1 model had an additional PPIL6 protein, which we believed to be a homologous protein of RSP12. In addition, MORN3 was absent from our model. On the one hand, no expression changes of MORN3 were detected in either human or mouse sperm samples, and on the other hand, Co-IP results showed no interaction between IQUB and MORN3. Likewise, MORN3 was considered to be a component of RS2 rather than RS1 in the previously reported near atomic model of mouse ependymal cilia and bovine sperm [40, 44]. To date, no high-resolution model of human sperm RS1 complex has been reported. We identified the possible RS1 components of human sperm by conducting comparative proteomics between the IQUB^842del patient and the normal control, and then predicted the RS1 structural model in human sperm based on the Chlamydomonas reinhardtii RS1 structure (Fig. 5A), which was extremely similar to that of mouse sperm, except that the RS1 model of human was slightly more constrictive than that of the mouse. It is noteworthy that for the predicted RS1 models, the interfaces between key functional regions, such as those involved in protein-protein interactions or structural stability, were supported by strong homology, particularly in regions with high sequence conservation. However, some flexible or less-conserved regions showed higher uncertainty in the model, especially in areas with lower pIDDT scores. In the human complex model, Chain R successfully assembled with the adjacent chains B, L, Q, and j, without causing any spatial clashes, following the template structure of 7JTK. However, due to the lower confidence and homology of Chain s, the model exhibited spatial clashes with neighboring chains e, j, and t. Similarly, in the mouse model, Chain R also assembled well with the adjacent chains B, L, Q, and j without significant spatial clashes. Chain s, however, still exhibited spatial clashes with neighboring chains e and t due to its lower confidence and homology.

In each 96-nm repeat unit there are six different single-headed IDAs (IDAa–IDAe and IDAg) and one double-headed IDA (IDAf). The molecular identity of each IDAs in Chlamydomonas was previously verified [45, 46]. In humans, however, the molecular identities of only IDAf (DNAH2 and DNAH10), IDAd (DNAH1), and IDAg (DNAH6) were inferred on the basis of sequence conservation [47]. The identities of IDAa, IDAb, IDAc and IDAe remain incompletely understood. Based on sequence conservation analysis, Martin deduced DNAH7 as IDAa, DNAH3 as IDAb, and DNAH12 as IDAc, respectively [47]. Additionally, Lucie et al. considered DNAH3, DNAH7, DNAH12, and DNAH14 to be IDAa-c and IDAe, but had not been able to assign them as specific subtypes [48]. Nevertheless, a recent study revealed by cryo-ET that IDAa was DNAH12, IDAc was DNAH3, and IDAb and IDAe were both DNAH7 [49]. Our study found that the expression of DNAH7 and DNAH12 was significantly decreased in humans and mice sperm lacking RS1. In addition to unique microtubule adapters, each RS in the 96-nm repeat unit is linked to a different set of neighboring structures, participating in the intricate connectivity among multiple force generators (inner and outer dynein arms) and signaling hubs such as nexin-dynein regulatory complex, intermediate chain and light chain complex, calmodulin spoke complex and IDAf [50, 51]. Based on our predicted RS1 model and the reported high-resolution models of the axoneme [44, 49], IQUB functions as an adapter that anchors RS1 to the A-tubule of DMT and also connects to the base of IDAa and IDAb, implying that defects in IQUB may cause instability of IDAa and IDAb. Hence our data provides further evidence of DNAH12 as IDAa and DNAH7 as IDAb.

The current study also has some limitations. First, only one patient harboring IQUB mutation were identified in our study, more cases are needed to consolidate the genetic contribution of IQUB mutations to asthenozoospermia. Second, despite the high accuracy of AlphaFold3, the models may still contain some error, especially in regions with low sequence conservation or flexible structures. Additionally, since the homology search was based on sequence similarity, our structural models rely on the assumption that the structural features of RS1 in Chlamydomonas reinhardtii are conserved across species, which could introduce potential discrepancies in the predicted structures of human and mouse RS1 proteins. Consequently, expanding the genetic analysis in larger cohorts and collecting more spermatozoa from patients with IQUB mutations for cryo-ET to refine the RS1 structure seems to be a very promising direction for subsequent research. Moreover, it would be worthwhile to perform rescue experiments such as re-expression of IQUB in Iqub^−/− mice in the future to confirm the causality and mechanistic details of IQUB mutations leading to asthenozoospermia.

Conclusions

In conclusion, our study functionally shed light on the role of a RS1 constitutive protein IQUB in male reproduction. These results may have important ramifications for enhanced understanding of sperm RS1 structures in humans and mice and provide diagnostic target for genetic counseling and future individualized treatment of male infertility in clinical practice.