Cep76 is a spermatid-enriched mRNA

Consistent with its putative role in male fertility, human CEP76 and mouse Cep76 are most prominently expressed in the testis (Fig S2A and B). To define which germ cell types Cep76 is expressed in, we used single-cell RNA-sequencing data for mouse testes. This revealed that Cep76 expression is considerably elevated in step 6–9 spermatids (Fig S2C), at the time when the core of the sperm tail is formed. We tested multiple antibodies to investigate CEP76 localisation in male germ cells but found all were non-specific (data not shown). All antibodies reacted with antigens in mutant tissue and at the incorrect size, suggesting they bound non-specifically.

Cep76 is required for male fertility in mice

To directly test the hypothesis that CEP76 is required for sperm tail development and thus male fertility, we generated a mutant mouse model using the CRISPR/Cas9 technology. We removed exon 3 of the mouse gene (Fig S1E), which resulted in a premature stop codon in exon 4 of Cep76 of the full-length protein-coding transcript (ENSMUST00000097542.3). This transcript encodes a protein of identical length (659 AAs, predicted molecular mass of 74.3 kD) and 97.7% identity (Fig S1B) to the human CEP76 principal isoform. In addition, sequence comparison revealed that the amino acid affected in the infertile man is a highly conserved residue from men to zebrafish (Fig S1B and C). We note that Cep76 is predicted to encode a second transcript that generates a smaller protein (54 kD) that only contains the complete TGL domain. This transcript is predicted to be unaffected by the genetic modification because of the presence of an alternative start codon in exon 5. The complete removal of exon 3 from all mRNA was confirmed via qRT–PCR (Fig S1F). Primers targeting mRNA downstream of the deletion revealed a highly significant reduction in mRNA in the mutant to 22% of wild-type levels. This residual mRNA may correspond to the expression of the second transcript. No functional protein was, however, predicted to be expressed.

Although mutant males were free of overt systemic disease and displayed normal mating behaviour, they were sterile (Fig 1A). Mutant females were fertile (not shown). Body, testis, and epididymis weights (Fig 1B–D) of Cep76 mutant males were equivalent to wild-type males. Although the daily sperm production of Cep76 mutant males was similar to wild-type counterparts (Fig 1E), the number of sperm within their epididymides was reduced to ∼35% of wild-type levels (Fig 1F; P < 0.0001). The reduced sperm number in the Cep76 mutant epididymis, in the absence of a reduction in testis weight and daily sperm production, was due at least in part to spermiation failure, as evidenced histologically in stage IX tubules, wherein sperm were retained within the seminiferous epithelium (Fig 1J). Sperm were not observed in wild-type tubules at a comparable stage (Fig 1I). Other processes, including the selective removal of abnormal sperm from the epididymides of mutant males, may also have contributed to the reduced sperm output. With the exception of spermiation failure, spermatogenesis appeared normal at a light microscopic level in Cep76 mutants (Fig 1H) and was comparable to wild type (Fig 1G). Similarly, epididymal histology was comparable between genotypes (Fig 1K and L).

CEP76 is required for normal sperm morphology and motility

Analysis of sperm morphology via light microscopy revealed overt defects in sperm from mutant males (Fig 2A). Total sperm tail length, principal piece length, and midpiece length (measured by mitochondrial sheath staining) of sperm from Cep76 mutant males were significantly shorter (by 15%, 16%, and 25%, respectively) than those from wild-type males (Fig 2B and C; all P < 0.0001), suggesting a fundamental role of CEP76 in tail assembly within the sperm ciliary compartment and mitochondrial loading. In agreement with the defects in sperm morphology, computer-assisted sperm analysis revealed a striking reduction in the percentage of sperm from mutants displaying any form of motility (Fig 2D; 6% in mutant versus 80% in wild type, P < 0.0001) and the virtual absence of forward, progressively motile sperm in comparison with wild-type males (Fig 2E; 0.4% in mutant versus 42.4% in wild type, P < 0.0001). The small population of sperm from Cep76 mutant males that were classified as motile were seen to be simply twitching on the spot.

In addition, we measured the distance of annulus migration from the base of the sperm head, where SEPT4 staining was used to mark the annulus (Fig 2F–H). Measurements revealed a significantly shorter distance between the annulus and the nucleus in sperm from Cep76 mutant males relative to wild type (18.6 μm in mutant versus 22.6 μm in wild type; P < 0.0001). In addition, the portion of sperm with SEPT4-positive staining (Fig 2F) was significantly reduced in the absence of CEP76 (60% in mutant versus 84% in wild type, P = 0.019). Of note in some sperm from Cep76 mutants, we also identified ectopic SEPT4 staining at the sperm neck, suggestive of TZ localisation (Fig 2G). In the same sperm, SEPT4 was also still observed at the annulus. This was never seen in wild-type sperm and suggests CEP76 is required for both annulus formation and migration. It likely also suggests that the annulus is fragmented in sperm from Cep76 mutants.

Light microscopy also revealed defects in sperm head shape (Fig S3A) in at least 40% of sperm from knockouts compared with 5% in wild type (P < 0.0001). We also observed a significant increase in sperm with abnormal acrosomes from Cep76 mutant males (35% in mutant versus 3% in wild type, Fig S3D; P = 0.0002). In addition, we analysed sperm head morphology using nuclear shape analysis software. This revealed a significantly lower proportion of the mutant population classified as normal (Cluster 1), although there were significant increases in sperm with slightly (Cluster 2) and largely (Cluster 4) abnormal head shapes (compared with wild type). This analysis revealed that ∼70% of sperm heads from Cep76 mutant males had abnormal nuclear morphology (Fig S3E; P = 0.0002). In addition, there was a twofold increase in sperm decapitation in samples from mutant males, indicating that CEP76 plays a role in establishing patency of the head–tail coupling apparatus (HTCA) (Fig S3C; P = 0.0002).

Collectively, these results reveal CEP76 is required for the production and release of normal numbers of functional sperm. Shorter sperm tails and annulus defects suggest CEP76 plays a role in TZ function. Defects in nuclear morphology and a weakened HTCA are supportive of a role of CEP76 in manchette function and/or acrosome formation and in the fortification of the neck region (a derivative of the basal body) in late spermiogenesis. It is, however, currently unclear as to how CEP76 would achieve this role. At a functional level, sperm from Cep76 mutant males are unable to reach the site of fertilisation because of highly impaired sperm motility.

CEP76 is required for normal sperm flagella and HTCA ultrastructure

To define the origins of the motility and structural defects, we investigated sperm ultrastructure using transmission electron microscopy (TEM) and scanning electron microscopy (SEM) on isolated epididymal sperm (Figs 3 and 4 and S4). At the midpiece level, transverse TEM sections revealed that axoneme ultrastructure appeared superficially normal, bearing all outer dense fibres and the 9 + 2 microtubule formation with dynein arms in sperm from wild-type and mutant males (Fig 3A and B, respectively). In Cep76 mutant sperm, however, a build-up of mitochondria and membranes was seen (Fig 3C). Longitudinal TEM images of sperm tail sections of the midpiece revealed an intact mitochondrial sheath in wild type (Fig 3G). In sperm from mutant males, abnormal mitochondria with enlarged spacing within mitochondrial matrixes were observed (Fig 3H), including notable mitochondrial aggregation (Fig 3I—arrow).

TEM investigation of the axoneme ultrastructure at the principal piece level revealed absent outer dense fibres and microtubule doublets in many sperm tails from mutant males (Fig 3D–F). In addition, we observed the abnormal formation of the annulus. Specifically, in sperm from Cep76 mutant males the annulus was not clearly distinguishable (Fig 3K), or not identifiable (Fig 3L). SEM of sperm from Cep76 mutant males further emphasised the poorly formed annulus and accessory structures in the midpiece and principal piece (Fig 4). In sperm from wild-type males, the annulus was positioned at the junction between the mitochondrial sheath of the midpiece and the fibrous sheath of the principal piece (Fig 4A and D). In sperm from Cep76 mutant males, although the annulus was often associated with the fibrous sheath, it was rarely found aligned at the distal end of the mitochondrial sheath (Fig 4B, C, E, and F). As a result, a region of exposed ODFs was evident between the mitochondrial sheath and the annulus or fibrous sheath. Quantification of the defects in annulus positioning (Fig 4M) revealed normal positioning and structure of the annulus in only 10% of sperm from mutant males compared with 95% of sperm from wild-type males (P = 0.036). Consequently, we quantified the incidence of ODF-exposed regions at the distal midpiece (Fig 4N). This revealed that ODFs were exposed in 97% of sperm from Cep76 mutants but only in 10% of sperm from wild type (P < 0.0001). In Cep76 mutants, the annulus was often found embedded within the proximal region of the fibrous sheath (e.g., Fig 4K) rather than at the junction of the midpiece, or it was not clearly observed because of large deformations in the fibrous sheath (Fig 4F). Because of the predicted role of CEP76 in TZ function, we anticipate this is due to a defect in TZ/annulus migration rather than the fibrous sheath assembling too far proximally. This phenotype, along with the ectopic localisation of SEPT4 as shown in Fig 2G, suggests that CEP76 plays a role in establishing the formation/structural competency and movement of the annulus.

In addition, we observed rare examples of mitochondria ectopically incorporated within the principal piece in Cep76 knockout sperm (Fig 3K), but never in sperm from wild-type controls, that is, the presence of overlapping fibrous sheath and mitochondria. Mitochondrial packing in epididymal sperm from Cep76 mutants was irregular and characterised by misaligned and poorly compacted mitochondria along the mitochondrial sheath (Fig 4H versus Fig 4G). Mitochondrial aggregation defects identified via TEM were confirmed via SEM and were characterised by collections of poorly formed mitochondria and a build-up around the midpiece (Fig 4I). Scoring of mitochondrial sheath normality (as defined in the Materials and Methods section; Fig 4O) revealed a significant reduction in mitochondrial sheath quality in sperm from Cep76 mutant males compared with wild type (3.4 in mutant versus 4.6 in wild type; P < 0.0001).

Moreover, although in wild-type sperm, the fibrous sheath appeared intact and continuous via TEM (Fig 3J), in mutant sperm, we observed fragmented fibrous sheath components (Fig 3K and L). SEM confirmed that fibrous sheath formation was notably impaired in the absence of CEP76 (Fig 4C, F, K, and L). In wild-type sperm, the fibrous sheath is comprised of two longitudinal columns linked by circumferential ribs. Gaps in the fibrous sheath were evident at regular intervals (Fig 4D and J). In the absence of CEP76, however, gaps/slits between circumferential ribs were rare (Fig 4K), and in many cases, the ribs appeared to be unanchored and projected from the body of the tail (Fig 4L). Notably, the proximal portion of the fibrous sheath appeared to be the most disorganised (e.g., Fig 4C). Fibrous sheath normality was scored (as detailed in the Materials and Methods section; Fig 4P), which revealed a highly significant reduction in fibrous sheath quality in sperm from Cep76 mutants (1.6 in mutant versus 2.9 in wild type; P < 0.0001).

Examination of the sperm neck region by TEM (Fig S4D–F) revealed a number of abnormal structures, including imperfect capitulum structures and a build-up of mitochondria and ectopic vesicles in sperm from Cep76 mutant males. Furthermore, an excess of what we predicted to be granulated bodies was seen throughout the cytoplasm (Fig S4E). Granulated bodies are transported into the ciliary compartment to form the outer dense fibres, which is continuous with the HTCA (43). This finding strongly supports a role of CEP76 in the selective entry of proteins and vesicles into the ciliary compartment during spermiogenesis and, by extension, in building of the ODF and the fortification of the HTCA. Of relevance, we did not observe a difference in the proportion of sperm positive for SUN5 staining between genotypes (Fig S4J; P = 0.41). SUN5 is a protein required to fortify the connection between the capitulum and the basal plate (44), that is, the junction between the tail and the nucleus. As this structure is proximal to where the TZ exists, these data are consistent with a role of CEP76 in the TZ and the assembly of more distal regions of the tail, and that the capitulum defects are secondary effects of CEP76 loss.

To assess whether the defects described above were associated with structural abnormalities in the TZ as sperm tails are forming, we undertook TEM on early elongating spermatids. As shown in Fig 3M–O, TZ structures were comparable between genotypes.

Collectively, these data reveal CEP76 is a key determinant in the development of multiple aspects of sperm tail development, including the fibrous sheath, annulus positioning, and the mitochondrial sheath.

Loss of CEP76 leads to aberrant sperm composition

To explore the hypothesis that CEP76 is involved in the selective entry of proteins into the ciliary lobe, we performed quantitative mass spectrometry on sperm from the cauda epididymis of Cep76 mutant and wild-type males (Table 1A). Here, we identified 63 differentially expressed proteins in sperm from mutant males (35 up-regulated and 28 down-regulated), including multiple mitochondrial and apoptotic proteins. This difference in mitochondrial protein content is consistent with the abnormal accumulation of mitochondria around the sperm tail as detailed above. In agreement with the observed reduced tail length, alpha-tubulin content was significantly reduced in sperm from Cep76 mutant males. In a separate analysis, and to account for differences in tail length, the content of known (or predicted) sperm tail proteins was normalised to alpha-tubulin. As shown in Table 1B, after this normalisation, 12 proteins were identified as significantly altered in sperm from Cep76 mutant males compared with wild type including several that are essential for sperm tail function and male fertility. We also identified 23 proteins that were only present in a single genotype (i.e., wild-type or mutant sperm). No direct sperm motility–regulating proteins of interest were identified in this last cohort. Candidates of interest to explain sperm dysfunction included the following: axonemal protein DNAH2 (1.65-fold wild-type levels, P = 0.041), the actin-based motor protein myosin 9 (3.5-fold wild-type levels, P = 0.017), and AKAP3 and AKAP4, which are major components of the fibrous sheath (24) (1.36-fold and 1.46-fold wild-type levels, P = 0.032 and P = 0.05, respectively).

To investigate this counterintuitive increase in several motility proteins, we defined the localisation of DNAH2 and AKAP4 proteins in sperm from Cep76 mutant males using immunofluorescence (Fig 5). As expected, in wild-type sperm, the dynein arm protein DNAH2 was localised throughout the midpiece and principal piece of the tail (Fig 5A) (45, 46). In contrast, although DNAH2 was found within the sperm tail in many sperm from Cep76 mutants, it was notably accumulated in the neck region. In other sperm, DNAH2 was seen nearly exclusively localised to the neck and appeared to be almost absent from the tail (Fig 5A). Specifically, 22% of sperm from Cep76 mutant males displayed an accumulation of DNAH2 in the neck region compared with 5.1% of sperm from wild types (4.3 x wild-type levels, P = 0.0078; Fig 5C). Tail pixel intensity analysis (per area) revealed a 29% reduction in DNAH2 content across the entire flagellum of sperm from mutant mice compared with wild-type controls (P = 0.0040; Fig 5E).

Also as expected for a component of the fibrous sheath, AKAP4 was primarily localised to the principal piece sperm from wild type (Fig 5B). As above, in Cep76 mutant sperm, AKAP4 accumulated in the neck region in 13% of sperm compared with 4% of sperm from wild-type mice (Fig 5D). Quantification of neck AKAP4 staining revealed 3.1-fold levels in Cep76 mutant sperm relative to wild type (P = 0.040; Fig 5D). An analysis of pixel intensity revealed a significant, 47% reduction in AKAP4 content (per area) within the principal piece of sperm from Cep76 mutant males compared with wild type (P < 0.0001; Fig 5F).

Collectively, these data underscore an essential role of CEP76 in TZ function and the regulated entry of key sperm tail proteins into the ciliary compartment and sperm tail development.

CEP76 is not required for manchette formation and migration

SEM confirmed the magnitude of head deformation in sperm from Cep76 mutant males (Fig S4A–C). The manchette is critically involved in shaping the distal half of the sperm nucleus and further acts as a transport freeway for protein delivery to the basal body and into the sperm tail (26). Despite the highly irregular sperm head shape, the formation of the manchette appeared normal in the absence of CEP76 as identified by alpha-tubulin staining of isolated germ cells (Fig S3F) and testis sections (Fig S3G). Electron microscopy reinforced that the manchette structure was overtly normal (Fig S3H), including a normal perinuclear ring and manchette microtubules. This analysis does not, however, exclude the possibility of subtle defects in manchette structure and/or kinetics.

CEP76 is required for the maintenance of centriole number in male germ cells

CEP76 has previously been linked to a role in the suppression of centriole duplication in 293T cells (39), thus raising the possibility of a similar role in germ cells. To explore this hypothesis, we stained purified round and elongating spermatids with the centriole marker centrin (Fig S4G and H) and quantified the number of centrioles per cell (Fig S4I). As expected, two distinct centrin foci were observed in 84% of wild-type spermatids and only 1% of wild-type cells exhibited 3+ centrin foci. In contrast, in spermatids from Cep76 mutant males, 18% of cells possessed 3 or more centriole structures (P = 0.046) and only 70% of cells contained 2 centrin foci (P = 0.0076). These data confirm a role of CEP76 in germ cell centriole duplication suppression. Despite this finding, sperm from Cep76 mutants contained only single tails and axonemes; that is, centriole overduplication did not lead to multiple basal bodies and axoneme growth.

Ethics statement

Experimental procedures involving mice followed animal ethics guidelines generated by the Australian National Health and Medical Research Council (NHMRC). All animal experiments were approved by the Animal Experimentation Ethics Committee (BSCI/2017/31) at Monash University, or The University of Melbourne Animal Ethics Committee (application 20640). Mice were housed on a 14-h light cycle with temperature set to 20°C and standard environmental enrichment.

Mutant mouse production

As described previously (36), exome sequencing of infertile men identified CEP76 as a high-confidence candidate male fertility gene. The patient carried a homozygous c.607G>C (p.Gly203Arg) missense variant in a conserved residue of CEP76, which was predicted to be intolerant to variation (Fig S1). Pathogenicity assessments using SIFT and PolyPhen-2 (66, 67) predicted the missense variant to affect function (SIFT) and to be possibly damaging (PolyPhen). To test the requirement for CEP76 in male fertility, Cep76 loss-of-function mice were generated on the C57BL/6J background through the Monash University Genome Modification Platform (a partner of the Australian Phenomics Network) using the CRISPR/Cas9 technology. Excision of exon 3 of Cep76 was undertaken with CRISPR guide sequences targeting regions flanking exon 3: upstream—TTTAAAACTCAGTTCGTGGT; and downstream—GGTCTACATAGTAAAGTTCT. This was predicted to lead to a premature stop codon in exon 4 and a truncated protein in the only full protein-coding transcript of Cep76 (ENSMUST00000097542.3) but did not target the smaller second transcript encoding exons 5-12. Changes in the gene sequence were identified with Sanger sequencing (Supplemental Data 1). Mice heterozygous for the Cep76 deletion were intercrossed to generate mutant individuals and wild-type controls. Genotyping was performed by Transnetyx. A reduction in the Cep76 truncated transcript level was investigated using qRT–PCR on testis cDNA (primers F—GCGGCTCGATTTGTTAATGT; and R—AGTCCCCACACAGACAAAGG) relative to Ppia (primers F—GTCTCCTTCGAGCTGTTT; and R—ACCCTGGACATGAATCCT). Similarly, confirmation that all transcripts contained the Cep76 deletion was done using primers in the deleted region (F—CCCTTCTTCTCCAAAGCAAACCG; and R—CGAGCAAATCTGTCCAGGCAA).

CEP76 species alignments

CEP76 protein alignments were conducted using the protein Basic Local Alignment Search Tool (NCBI). Sequences used were Homo sapiens ENSP00000262127; Pan troglodytes ENSPTRP00000016812; Macaca mulatta ENSMMUP00000069818; Rattus norvegicus ENSRNOP00000034590; Mus musculus ENSMUSP00000095149; Danio rerio ENSDARP00000075595. We compared entire protein identity across species and focused on the conservation of amino acid 203G, which was mutated in the infertile patient (Fig S1B).

Analysis of Cep76 expression

Whole organ RNA was extracted from adult mouse brain, epididymis, heart, liver, lung, spleen, and testis, to investigate the expression of Cep76 across different tissues. Each tissue was homogenised in TRIzol Reagent to isolate RNA, which was converted to cDNA using 50 μM oligo(dt)15 primers with SuperScript III enzyme, and used for qRT–PCR with SYBR Green Master Mix as previously described (68). Primers used to detect Cep76 were F—CTCGGTCACCAGCAATGAAA; and R—CAGACAGTGGTGAGGCCAAG, and the housekeeping gene Ppia is denoted above. We also used published RNA-seq data for human organs to assess CEP76 expression (40) and single-cell RNA-sequencing data contained within FertilityOnline (mcg.ustc.edu.cn/bsc/spermgenes2.0/index.html) to investigate Cep76 expression in mouse testes (69).

Fertility analysis

Mutant males (Cep76^−/−) and wild-type male littermates (Cep76^+/+) were aged 10–14 wk, and their fertility was assessed using the pipeline outlined previously (70). In brief, five mice of each genotype were set up to mate with two independent wild-type females each (6–12 wk old). The presence of a copulatory plug was recorded as an indication of successful mating. Litter sizes were recorded as the number of pups generated per plug. Males were subsequently culled and weighed (10–14 wk of age), and one testis and epididymis were dissected and processed for histological assessment. Additional testes and epididymides were snap-frozen on dry ice for calculation of daily sperm production and epididymal sperm counts as described previously (71). In brief, testes were weighed and resuspended in a detergent-based solution and sonicated to burst the nuclei of cells with non-compacted DNA while leaving resistant elongated spermatids and sperm nuclei intact. Detergent-resistant nuclei were counted on a haemocytometer (71), and the total number of cells within the testis was calculated. To calculate daily sperm production, this number was divided by 4.84—the length of time in days step 14–16 spermatids (detergent-resistant) spent in the testis (72). To measure the total epididymal sperm count, the same method was used without the calculation to obtain a daily count. Before incubation in the detergent solution, the epididymides were finely minced.

In addition, sperm were collected from the cauda of the epididymis through backflushing, then resuspended in MT6 medium at 37°C for motility assessment via computer-assisted semen analysis (73). Residual sperm were washed in PBS, then dried onto SuperFrost slides overnight. Sperm were then fixed in 4% PFA for 10 min, washed in PBS, and stained with Mayer’s haematoxylin for 10 min and eosin (Amber Scientific) for 1 min to allow an assessment of sperm morphology and tail length. Alternatively, fixed sperm were permeabilised in 0.1% Triton X-100/PBS (Sigma-Aldrich) for 10 min, washed in PBS, and stained with 10 μg/ml DAPI (Thermo Fisher Scientific) to allow an assessment of sperm head morphology, or 1 μg/ml peanut agglutinin conjugated to Alexa Fluor 488 to label the acrosome. Head morphology assessment was performed using Nuclear Morphology Analysis software version 1.17.1 (74) via an ImageJ plugin (National Institutes of Health, USA). All length measurements (midpiece length, principal piece length, full tail length, and annulus positioning) were measured using ImageJ (version 1.52k).

Electron microscopy

To investigate germ cell ultrastructure, testes were processed for electron microscopy as outlined previously (75). Similarly, caudal sperm were backflushed into MT6 medium and processed for electron microscopy as outlined previously (75). Images were taken either on a Jeol 1400 Plus electron microscope at the Vera and Clive Ramaciotti Centre for Electron Microscopy (Monash University, Australia), or on a Talos L120C or a FEI Teneo VolumeScope at the Ian Holmes Imaging Center (The University of Melbourne). To view the mitochondria, annulus, and fibrous sheath structure of sperm via SEM, sperm were isolated from the cauda epididymis and incubated in 100 μl of 1× PBS for 30 min to strip the plasma membrane, then processed as outlined in reference (76).

Scoring of mitochondrial sheath, fibrous sheath, and annulus normality

To quantify the degree of ultrastructural defects within sperm tail structures, scoring of each structure (mitochondrial sheath, fibrous sheath, and annulus) was performed on SEM images. For each biological replicate, at least 25 sperm were assessed. Mitochondrial sheath normality was scored (1–5): 1—missing mitochondria; 2—many abnormally oriented and/or thick mitochondria; 3—few abnormally oriented and/or thick mitochondria; (4)—broadly normal with very few abnormally oriented mitochondria; and (5)—no defects, where “normal” was defined as homogeneously coiled mitochondria, with none missing, thick, or abnormally oriented. Fibrous sheath normality was scored from ranks (1–3): 1—no slits and/or massive aggregation of circumferential ribs or longitudinal columns; 2—reduced number of slits or slight aggregation; and 3—normal structure, where normal was defined as slits occurring at regular intervals and no aggregation. Annulus normality was scored as follows: normal—intact, not shrunken, and localised to the junction between the midpiece and the principal piece; or abnormal—small, ectopically placed, or unidentifiable.

Mass spectrometry to define sperm protein composition

Sperm were backflushed from cauda epididymides of adult wild-type (one mouse per replicate, n = 3 replicates) and Cep76 mutant (two mice per replicate because of reduction in epididymal sperm content) males into MT6 medium for 15 min at 37°C. Sperm were washed three times in 1× Tris-buffered saline, pelleted, and then stored at −80°C. Sperm pellets were then dried in a speed vacuum, prepared for liquid chromatography–tandem mass spectrometry, and run on a SCIEX QTRAP6500 as described previously (77).

Data were assessed as (1) total/raw spectral counts or (2) total spectral counts were normalised to the alpha-tubulin content (tubulin alpha-8 chain) for a comparison of sperm tail protein content in recognition of the shorter sperm tails measured from Cep76 null mice. Sperm tail proteins were identified via proteomics or localisation studies. For completeness, proteins with an unknown localisation or that are localised throughout the head and tail were included in the tail protein group. Proteins associated exclusively with the sperm head or neck and mitochondrial proteins were not included here as the normalisation process (to alpha-tubulin) as they would not be influenced by processes involved in selective entry into the ciliary compartment. A two-tailed t test was used to determine significant differences in protein content between genotypes. All data are available in Table S1. As we have not corrected for multiple comparisons in our proteomics data and have used zero values for proteins not detected in individual biological replicates, the biological significance of proteins identified via this screen as being significantly modulated in mutant sperm should be interpreted with caution.

Immunofluorescence and protein localisation

To define centriole number and manchette structure in male germ cells, spermatids were isolated from the testes of males using the STAPUT method (78). Cells were settled onto poly-L-lysine–coated SuperFrost slides for 15 min and then fixed in ice-cold methanol at −20°C for a maximum of 7 min. Slides were immediately washed and rehydrated in PBS. Fixed elongating spermatids were permeabilised in 0.2% Triton X-100/PBS for 10 min, blocked in CAS-Block (Dako), incubated overnight in 0.1 μg/ml centrin antibody (04-1624; Merck) to stain centriole components and beta-tubulin (ab21057; Abcam) to stain the manchette for spermatid staging, then counterstained with DAPI. For manchette imaging, spermatids were incubated overnight with an alpha-tubulin antibody (T5168, 1:1,000; Sigma-Aldrich), then stained with secondary goat anti-mouse 488 (Abcam) and DAPI. Z-stacks of labelled cells were taken using a Leica SP8 confocal and flattened to capture all slices in a single frame. The number of centriole components per spermatid was then counted for 100 cells per replicate, per genotype.

To define the localisation of a subset of differentially expressed proteins, fixed sperm were permeabilised in 0.2% Triton X-100/PBS, blocked in CAS-Block, incubated overnight in primary antibodies (0.5 μg/ml DNAH2 [64309; Invitrogen], 2.5 μg/ml SEPT4 [166788; Abcam]) at 4°C, stained with relevant fluorescent secondary antibodies (Thermo Fisher Scientific) for 1 h, then counterstained with DAPI. For other antibodies (2.8 μg/ml AKAP4 [24986-1-AP; ProteinTech], 0.8 μg/ml SUN5 [17495-1-AP; ProteinTech]), sperm were permeabilised in 0.5% Triton X-100/PBS. For the assessment of sperm mitochondrial sheath length, live sperm were loaded with 5 μM MitoTracker Red CMXRos in MT6 solution for 30 min as per the manufacturer’s instructions. Sperm were then washed in PBS, fixed in 4% PFA, and allowed to settle onto slides, as described above. All images were taken using cellSens software (Olympus) on an Olympus BX-53 microscope (Olympus) equipped with an Olympus 392 DP80 camera or a Leica SP8 confocal. Ectopic localisation of DNAH2 and AKAP4 was scored in at least 100 sperm per biological replicate. Immunofluorescence staining intensity was used as a surrogate of protein content in sperm tails. FIJI v2.1.0 was used to trace sperm tails (freehand line tool) and measure the average pixel intensity for AKAP4 and DNAH2 staining in at least 20 sperm, to account for differences in tail length.

CEP76 localisation in testis sections, germ cells, and sperm was attempted with three antibodies: ab86613; Abcam, #A302-326A; Bethyl Laboratories, and HPA039395; Sigma-Aldrich (Human Protein Atlas). All antibodies reacted with antigens in mutant tissue, suggesting they bound non-specifically. Equally, all antibodies bound non-specifically to non-CEP76 proteins when used in Western blotting (data not shown).

Statistical analysis

Statistical analyses to determine the significance between wild-type and Cep76 mutant data were performed in Prism 9 (GraphPad). For all data that were normally distributed with sufficient sample size (Shapiro–Wilk test), we used a t test with Welch’s correction to account for variance. For data not normally distributed, we used a Mann–Whitney U test. A P-value less than 0.05 was considered significant. No data were excluded.