Proximity labeling of axonemal protein CFAP91 identifies EFCAB5 that regulates sperm motility

Abstract

Radial spokes (RSs) are conserved multimolecular structures attached to the axonemal microtubule doublets and are essential for the motility control of both cilia and sperm flagella. CFAP91, an RS3 protein, is implicated in human male infertility, yet its molecular function remains poorly understood. Here, we demonstrate that Cfap91 knockout (KO) mice exhibit impaired sperm flagellum formation and male infertility. Using a transgenic rescue model expressing FLAG- and BioID2-tagged CFAP91, we reveal that CFAP91 immunoprecipitates with RS3 proteins CFAP251 and LRRC23, whose localization is disrupted in Cfap91 KO sperm flagella. In addition, proximity labeling in mature spermatozoa identifies EFCAB5 as a sperm-specific CFAP91-proximal component. We show that Efcab5 KO males exhibit reduced sperm motility and fertility. Our findings establish CFAP91 as an essential scaffolder of RS3 assembly and EFCAB5 as a sperm-specialized movement regulator, advancing understanding of axonemal specialization in mammalian spermatozoa and its relevance to male infertility.

Fig. 1. Ablation of Cfap91 leads to defects in spermiogenesis.

a A schematic diagram of the radial spoke in mouse sperm flagella and human respiratory cilia. CP central pair, DMT doublet microtubule, RS radial spoke. b Atomic model of RS3 components. Data was retrieved from the paper by Leung et al.. c, d RT-PCR of Cfap91 utilizing cDNA from multiple mouse organs (c) or mouse postnatal testes (d). Actb was used as a loading control. BR brain, TH thymus, LU lung, HE heart, LI liver, SP spleen, KI kidney, TE testis, OV ovary. e A schematic drawing of the KO strategy of Cfap91. f Ratio of Cfap91^+/+, Cfap91^+/-, and Cfap91^-/- mice obtained from the mating of Cfap91^+/- mice in the B6D2 or B6D2/129 background. n = 89 for B6D2 background and n = 107 for B6D2/129 background. g Fertility tests of Cfap91^+/+ and Cfap91^-/- males. Data from each male mouse was individually color-coded. n = 30 plugs examined over 3 males for WT mice, n = 27 plugs examined over 3 males for Cfap91^-/- mice. Data were presented as mean ± SD. h Sections of stage VII-VIII seminiferous tubules from Cfap91^+/- and Cfap91^-/- males. i Immunohistochemistry of stage II-III, stage VII-VIII, and stage X-XI seminiferous tubules from Cfap91^+/- and Cfap91^-/- males.

Fig. 2. Cfap91^-/- males show oligo-astheno-teratozoospermia.

a Phase contrast images of cauda epididymal spermatozoa from Cfap91^+/- and Cfap91^-/- males. b Sperm tail length of cauda epididymal spermatozoa from Cfap91^+/- and Cfap91^-/- males. Data from each male mouse was individually color-coded. n = 267 spermatozoa examined over 3 males for Cfap91^+/- mice, n = 300 spermatozoa examined over 3 males for Cfap91^-/- mice. Data were presented as mean ± SD. An unpaired two-tailed t-test was performed for statistical analysis. P < 1.0E-15. c Ratio of the abnormal head of cauda epididymal spermatozoa from Cfap91^+/- and Cfap91^-/- males (n = 3 males for each genotype). Data were presented as mean ± SD. An unpaired two-tailed t-test was performed for statistical analysis. P = 1.5E-8. d Ratio of motile spermatozoa from the cauda epididymis of Cfap91^+/- and Cfap91^-/- males. Cfap91^-/- males showed no motile spermatozoa (n = 5 males for each genotype). Data were presented as mean ± SD. e, f A schematic diagram of the EGFP-Tuba3a transgene driven by a Clgn promoter (e). The results of genomic PCR performed with the corresponding primers are shown (f). g Immunoblotting of EGFP in testes and cauda epididymal spermatozoa of WT and EGFP-Tuba3a TG males. ACTB was used as a loading control. h Imaging of manchette using EGFP-Tuba3a TG mice with Cfap91^+/- and Cfap91^-/- genotypes. Hoechst 33342 (H33342) was used to visualize the nuclei of spermatids. Images are arranged from left to right with the progression of spermiogenesis.

Fig. 3. Ultrastructural defects in Cfap91 KO spermatozoa.

a–f Cross-sections of spermatids from the testes (a–c), and spermatozoa in the cauda epididymis (d–f) of Cfap91^+/- and Cfap91^-/- males. g, h Longitudinal sections of spermatozoa in the cauda epididymis of Cfap91^+/- and Cfap91^-/- males. Genotypes of Cfap91 are indicated on the upper right of each figure. Abnormal proximal annulus found in Cfap91 KO spermatozoa is shown with black triangles.

Fig. 4. CFAP91 is localized in sperm tails.

a A schematic diagram of the Cfap91-BioID2-3×FLAG transgene. Mice carrying this transgene are referred to as Cfap91 TG mice. b Immunoblotting of CFAP91-BioID2-3×FLAG with an anti-FLAG antibody using testes or cauda epididymal spermatozoa of WT and Cfap91^-/- TG males. c Fertility tests on Cfap91^+/+ and Cfap91^-/- TG males. Data from each male mouse was individually color-coded. n = 23 plugs examined over 3 males for WT mice, n = 15 plugs examined over 3 males for Cfap91^-/- TG mice. Data were presented as mean ± SD. An unpaired two-tailed t-test was performed for statistical analysis. d Immunohistochemistry on the sections of seminiferous tubules with an anti-FLAG antibody in WT and Cfap91^-/- TG males. e Immunocytochemistry on spermatids obtained from WT and Cfap91^-/- TG males with an anti-FLAG antibody. f Immunoblot analyses on sperm head-tail separated lysates. IZUMO1 and acetylated tubulin (Ac-TUB) served as controls of the head and tail fractions, respectively. g Immunoblot analyses on fractionated sperm lysates. BASIGIN (BSG), Ac-TUB, and AKAP4 served as controls for Triton-soluble, SDS-soluble, and SDS-resistant fractions, respectively.

Fig. 5. CFAP91 immunoprecipitates with multiple RS3 proteins.

a A volcano plot of the results from IP-MS studies using an anti-FLAG antibody. When comparing Cfap91^-/- TG males to WT males, proteins with fold change >2 and P < 0.05 are considered as significantly upregulated, and dots are color-coded in magenta. An unpaired two-tailed t-test was performed for statistical analysis. b Immunoblotting was performed after IP with an anti-FLAG antibody. Signals of CFAP251, LRRC23, IFT140, and BBS2 were found in Cfap91^-/- TG but not WT. IZUMO1 served as a loading control of inputs. c Immunoblotting was performed after IP with an anti-LRRC23 antibody or rat IgG, on Cfap91^-/- TG testicular lysate. A band of FLAG was found in the IP product using the anti-LRRC23 antibody. d Immunoblotting was performed after IP using P15 testes from WT and Cfap91^-/- TG males. A band of CFAP251 was found in Cfap91^-/- TG but not WT, while bands of LRRC23 were not found in either Cfap91^-/- TG or WT. Acetylated tubulin (Ac-TUB) served as a loading control for input lysates. e Immunohistochemistry of Cfap91^+/- and Cfap91^-/- testicular sections. LRRC23 and CFAP251 were co-localized with Ac-TUB, while this co-localization was not found in Cfap91^-/- testicular sections. f Immunoblot analyses of testes and cauda epididymal spermatozoa from WT and Cfap91^-/- males. Alpha-tubulin (α-TUB) served as a loading control.

Fig. 6. Identification of EFCAB5 as a sperm axoneme-specific RS3 protein.

a A schematic drawing of the application of BioID2 in mature spermatozoa. Spermatozoa from WT and Cfap91^-/- TG males were incubated in a medium supplemented with biotin for 16 h. Collected spermatozoa were lysed, and biotinylated proteins were pulled down by streptavidin (SA). b Immunoblotting was performed after the pull-down of biotinylated proteins in WT and Cfap91^-/- TG males. EFCAB5 was not detected in the input, likely due to low protein abundance and/or low solubilization efficiency (0.4% SDS compared to 1% SDS or 6 M urea in other experiments). IZUMO1 served as a loading control for inputs. c Molecular weight and total spectra of twenty-four proteins only identified in Cfap91 KO TG spermatozoa. CFAP91 and EFCAB5 were identified with the most peptides among all twenty-four proteins. A linear regression line with 95% confidence bands is shown. Slope = 0.06 and Y-intercept = 5.38. d Venn diagram of CFAP91 immunoprecipitates, proximity labeling-identified proteins, and mammalian sperm RS3 proteins. e RT-PCR of Efcab5 utilizing cDNA from multiple mouse organs. Actb was used as a loading control. f Immunoblot analyses of human spermatozoa. g immunoblotting using testes and cauda epididymal spermatozoa from WT and Cfap91^-/- males with an anti-EFCAB5 antibody. Alpha-tubulin (α-TUB) served as a loading control. h Fractionation using WT spermatozoa. Signals of EFCAB5, LRRC23, and CFAP251 were found in the SDS-soluble fraction.

Fig. 7. EFCAB5 is vital for sperm motility.

a A schematic drawing of the KO strategy of Efcab5. b Fertility tests of Efcab5^+/+ and Efcab5^-/- males. Data from each male mouse was individually color-coded. n = 46 plugs examined over 5 males for WT mice, n = 48 plugs examined over 5 males for Efcab5^-/- mice. Data were presented as mean ± SD. An unpaired two-tailed t-test was performed for statistical analysis. P = 8.5E-8. c Phase contrast images of cauda epididymal spermatozoa from WT and Efcab5^-/- males. d Immunoblotting of testes and cauda epididymal spermatozoa from WT and Efcab5^-/- males, with anti-EFCAB5, anti-CFAP251, anti-LRRC23, and anti-RGS22 antibodies. ACTB served as a loading control. e Ratio of motile and progressive spermatozoa in WT and Efcab5^-/- males after 10 min and 120 min of incubation in a capacitation medium. n = 5 males for each genotype. Data were presented as mean ± SD. An unpaired two-tailed t-test was performed for statistical analysis. P = 2.5E-04 for progressive motility after 120 min of incubation. f VCL, VSL, and VAP of cauda epididymal spermatozoa of WT and Efcab5^-/- males after 10 min and 120 min of incubation in a capacitation medium. n = 5 males for each genotype. Data were presented as mean ± SD. An unpaired two-tailed t-test was performed for statistical analysis. g α-angle of cauda epididymal spermatozoa in WT and Efcab5^-/- males after 10 min and 120 min of incubation in a capacitation medium. Data from each male mouse was individually color-coded. n = 45 spermatozoa examined over 3 males for each genotype and time point. Data were presented as mean ± SD. One-way ANOVA and Tukey’s multiple comparisons test with adjustment were used for statistical analysis. P = 7.0E-14 for WT mice between 10 min and 120 min of incubation and P = 7.1E-14 for 120 min of incubation between WT and Efcab5^-/- mice.