Testis-expressed Bbof1 Is Required For Male Fertility
We first assessed the expression pattern of BBOF1 in mice. According to the semi-quantitative RT-PCR results for cDNA samples derived from multiple mouse tissues, Bbof1 was exclusively expressed in mouse testes (Fig. 1a). The expression of Bbof1 in testes was first detected at PD16 (postnatal day 16), the time point when the first wave of spermatocytes enter the pachytene stage, and was increased with testis development from PD18 to PD69 (Fig. 1b). Specifically, we sorted spermatocytes and spermatids at different stages and found that the level of Bbof1 mRNA was markedly increased in pachytene/diplotene spermatocytes and round spermatids (Fig. 1c).
To investigate the physiological functions of BBOF1 in mammals, we generated a Bbof1 knockout mouse model using the CRISPR/Cas9 system. An sgRNA was designed to target exon 6 of Bbof1 and was microinjected into wild-type (WT) mouse zygotes together with Cas9 protein (Fig. 1d). Through PCR amplification and sequencing, we identified a founder mouse with an overall 25 bp deletion within exon 6, which introduces a premature stop codon and produces a putative mutant protein with 212 amino acids (aa) in length (Fig. 1d). This allele (regarded as the knockout allele or null allele) was successfully transmitted to the next generation by backcrossing the founder mouse with WT mice. Bbof1–/– mice were obtained by interbreeding between heterozygous (Bbof1+/–) males and females (Fig. 1e). The ablation of BBOF1 protein was confirmed by Western blotting of the testis lysates of Bbof1–/– mice (Fig. 1f). The percentage of Bbof1–/– progeny mice was in accordance with the Mendelian ratio (25.6%, n = 215), and no apparent defects in their viability, body weight, and external morphology had been observed (Supplementary Fig. 1).
We performed fertility tests on BBOF1-deficient males and females. According to the statistical analysis of progeny numbers in different mating groups, BBOF1 deletion severely impaired male fertility while does not affect female fertility (Fig. 1g). Adult Bbof1–/– males were able to mate with Bbof1+/– and WT females, as indicated by the presence of vaginal plugs; however, successful pregnancy were not guaranteed (4 litters from 38 plugs). Therefore, Bbof1–/– males exhibited a phenotype of subfertility, as demonstrated by the fertility test (0.13 pups per plug and 1.25 pups per litter, Fig. 1g).
Bbof1 Is Dispensable For Spermatogenesis
Next, we investigated the effect of BBOF1 deletion on male fertility. Initially, we analyzed the testes and epididymides of adult WT and knockout males. Unexpectedly, the size and weight of the testes were not influenced by BBOF1 deletion (Fig. 2a-b). Hematoxylin and eosin staining of paraffin-embedded WT and Bbof1–/– testis tissue sections showed no apparent defects (Fig. 2c). All stages of spermatogenesis, spermatogonia, spermatocytes, round spermatids, and elongated spermatids were evident in the Bbof1–/– testis tissue sections (Fig. 2c). Wilms’ tumor 1 (WT1) is a marker of Sertoli cells and is required for the development of testicular cords during embryogenesis, and mouse Vasa homolog (MVH) is a marker of germ cells27. Immunofluorescence staining of WT1 and MVH showed no obvious difference in the number of Sertoli cells or germ cells in Bbof1–/– testes (Supplementary Fig. 2a-b). PSMA8 has been identified as a subunit of the testis-specific 20S core proteasome, which accumulates in spermatocytes from the pachytene stage28. PSMA8 signals indicated normal meiotic prophase I progression in BBOF1-deleted spermatocytes, like γH2AX signals, a marker of the generation and repair of DNA double-strand breaks (Supplementary Fig. 2c-d).
Bbof1 –/– males could produce spermatozoa with a normal morphology as their epididymides sections were full of mature spermatozoa (Fig. 2d). Peanut agglutinin (PNA) staining is often used to assess the Golgi to acrosome transition during spermiogenesis. However, PNA signals were not considerably different between WT and Bbof1–/– testes (Fig. 2e). To determine the effects of BBOF1-deletion on both the quantity and quality of spermatozoa, they were collected from the cauda epididymides for examination in vitro. A slight decrease in sperm count was observed in Bbof1–/– males (Fig. 2f). The morphology of spermatozoa and the incidence of abnormal spermatozoa were not affected by BBOF1 deletion (Fig. 2g-h; Supplementary Fig. 3). Taken together, these results suggested that BBOF1 may be dispensable for spermatogenesis and the morphology of spermatozoa.
Bbof1 Deletion Leads To Reduced Sperm Motility
However, the percentages of motile spermatozoa and progressive motile spermatozoa were significantly decreased in Bbof1–/– mice, which are common characteristics of asthenozoospermia (Fig. 3a; Supplementary Movie 1 and 2). Computer-assisted sperm analysis (CASA) further revealed the defective motility of BBOF1-deleted spermatozoa (Fig. 3b). Therefore, BBOF1 deletion led to typical asthenozoospermia with decreased motility but a normal morphology.
To understand the underlying mechanisms of subfertility, we superovulated WT females and mated them with adult WT or Bbof1–/– males. The majority of embryos from females mated with WT males exhibited a normal morphology at the 2-cell stage at 40 h after hCG administration (Fig. 3c-d). However, only a few embryos/oocytes from females mated with Bbof1–/– males were fertilized (1 of 137). Spindle and F-actin staining showed that most embryos were arrested at the MII stage with intact spindles and condensed chromosomes, i.e., not fertilized (Fig. 3e).
The subfertility of Bbof1–/– males could be attributed to obstructive azoospermia, in which the vas deferens, epididymis, and ejaculatory ducts are obstructed, or asthenozoospermia, where sperm motility is insufficient for MII oocyte fertilization. We performed in vitro fertilization (IVF) experiments to rule out the possibility of ductal obstruction. Spermatozoa obtained from the Bbof1–/– and WT cauda epididymides were incubated with WT cumulus-oocyte complexes (COCs) for IVF. In comparison with the WT group, in which most oocytes were successfully fertilized (24 h post-hCG administration), oocytes in the Bbof1–/– group failed to fertilize and remained at the MII stage (Fig. 3f-h, ZP intact). Only a small proportion of oocytes (12 of 258; 4.35%) were fertilized with BBOF1-null spermatozoa and developed to the 2-cell stage, which is consistent with the subfertility of males observed during fertility test.
In Xenopus, bbof1 has been reported to be associated with motile cilia. Considering that the sperm flagellum is a specialized type of motile cilium and that Bbof1 is specifically expressed in the testes, we hypothesized that mammalian BBOF1 functions mainly in the sperm flagellum and is exclusively required for sperm motility. When zona pellucida is removed with acidic medium (ZP free), MII oocytes were comparably fertilized with WT and BBOF1-null spermatozoa (96.9% vs. 92.5%, P = 0.25) were fertilized (Fig. 3f-h, ZP intact), suggesting reduced motility of BBOF1-null spermatozoa.
Bbof1 Is Required For Axonemal Assembly In The Sperm Flagellum
Sperm motility relies largely on an intact flagellar structure. Therefore, we investigated the ultrastructure of WT and BBOF1-deleted sperm flagella by transmission electron microscopy (TEM). Representative images of the cross sections of the mid-piece, principal piece, and end piece are shown in Fig. 4a and Supplementary Fig. 4. In the mid-piece, BBOF1-deleted spermatozoa had an intact “9 + 2” axonemal structure with intact MTDs, radial spokes, and dynein arms, as well as ODFs. However, the absence of one to three MTD(s) were frequently observed in the principal piece and end piece (Fig. 4a; Supplementary Fig. 4). Abnormal axonemes accounted for approximately 30–40% of all cross sections of the principal piece and end piece (Fig. 4b). Based on statistical analysis, MTD-1, 2, 7, 9, or their combinations were likely to be absent, as indicated by the representative images (Fig. 4a and c). MTD-9, MTD-1, and MTD-7/9 were the top three most frequently absent MTDs (Fig. 4c). Together with the absence of MTDs, the corresponding ODFs were often missing (Fig. 4a), suggesting the role of BBOF1 in the maintenance of the sperm flagellar ultrastructure.
Bbof1 Interacts With Mns1 And Odf2
To elucidate the molecular functions of BBOF1 in axonemal assembly in the sperm flagellum, yeast two-hybrid (Y2H) screening was performed to identify BBOF1-interacting proteins. Full-length mouse BBOF1 (aa 1–533 or 1–533) was fused to Gal4-BD as the bait protein. The Gal4-AD plasmid library carrying the total PD30 testis cDNA was used for large-scale screening. Combining the results of two rounds of screening, 1014 positive colonies and 52 BBOF1-interacting candidates were sorted and listed based on the number of colonies identified (Fig. 5a, only candidates interacting at least twice are shown). The top four candidates were SPATA24, RP9, MNS1, and ODF2, among which MNS1 is an axonemal component, and ODF2 is an ODF component (Supplementary Fig. 5a-b). The minimal BBOF1-interacting regions of the four candidates were identified based on the cDNA fragments of AD plasmids from positive colonies (Supplementary Fig. 5a).
BBOF1 is identified as a CCDC protein (CCDC176) with two conserved coiled-coil domains on its N- and C-terminal domains, aa 81–195 and aa 277–366 (Fig. 5b). Domain mapping results demonstrated that the C-terminal domain of BBOF1 could mediate its interaction with various proteins, such as ODF2, MNS1, SPATA24, RP9, and IFT74 (Fig. 5c-d; Supplementary Fig. 5). However, in some cases, the whole sequence of BBOF1 was required, such as during interaction with SASS6, CFAP57, and CCDC136 (Supplementary Fig. 5). Notably, the putative mutant protein BBOF1-KO expressed in Bbof1–/– mice could not bind to all of the examined BBOF1 interactors (Fig. 5c-d; Supplementary Fig. 5).
Y2H analysis demonstrated that BBOF1 could interact with both the axonemal component and ODFs, suggesting the possible role of BBOF1 in mediating the association between ODFs and the axoneme. Therefore, we investigated whether BBOF1 deletion affects the stability and localization of MNS1 and ODF2. In Bbof1–/– mouse testes, the endogenous protein levels of ODF2 and MNS1 were decreased (Fig. 6a), indicating that BBOF1 is required for the stabilization of ODF2 and MNS1. However, the localization of MNS1 and ODF2 in the sperm flagellum were not affected by BBOF1 deletion (Fig. 6b-c). Furthermore, we lysed WT and BBOF1-deleted spermatozoa with Triton X-100 and SDS-containing buffer, which gave rise to three protein fractions: Triton X-100-soluble (TS), SDS-soluble (SS), and SDS-resistant (SR). Similar to MNS1, BBOF1 was detected mainly in the SS fraction and partially in the SR fraction from WT spermatozoa (Fig. 6d). However, the protein level of MNS1 was markedly decreased in the SS fraction from BBOF1-deleted spermatozoa. Moreover, ODF2 was detected in the SR fraction from both WT and BBOF1-deleted spermatozoa; however, its level was decreased in BBOF1-deleted spermatozoa (Fig. 6d). These results suggested that BBOF1 may preferentially localize to the axonemal structure and may be required for the stability of MNS1 and ODF2.
Bbof1 Is Dispensable For The Function Of Motile Cilia
As the “9 + 2” axonemal structure is commonly found in all types of motile cilia, we further investigated whether subfertility is a symptom of PCD. Bbof1–/– mice did not show typical PCD symptoms, such as hydrocephalus, respiratory tract infections, or ectopic positioning of internal organs (situs inversus) during breeding, fertility testing, and maintenance (Supplementary Fig. 1). Brain samples obtained from Bbof1–/– mice exhibited a normal external morphology (Fig. 7a), and the ventricles were not enlarged, as observed in the tissue sections (Fig. 7b). The differentiation of ependymal cells and motile cilia on their surface were not affected by BBOF1 deletion (Fig. 7c; Supplementary Fig. 6a-b). Similarly, motile cilia were normally distributed on the epithelial cells of the trachea of Bbof1–/– mice (Fig. 7d; Supplementary Fig. 6c-d). We also examined the ultrastructure of the tracheal cilia and found that the axonemal structure was intact in Bbof1–/– mouse tracheal sections (Fig. 7e). Taken together, BBOF1 is dispensable for the assembly and function of motile cilia and thus may play a role in regulating the structural integrity of the sperm flagellum.