Introduction

The differentiation of round spermatids into sperm during spermiogenesis is a highly organized and spatiotemporally controlled process taking place in the seminiferous epithelium of the testis. Cellular and morphological remodeling involves DNA hypercondensation, establishment of a species-specific head morphology, removal of excess cytoplasm as well as formation of accessory structures like acrosome and flagellum. These structural changes depend on a highly efficient protein trafficking machinery and a unique sperm cytoskeleton ¹. One essential cytoskeletal element is the perinuclear theca (PT), which surrounds the sperm nucleus, except for the caudal edge, at the implantation site of the flagellum. The PT is supposed to serve as a structural scaffold for the sperm nucleus and resembles a rigid cytosolic protein layer, which is resistant to non-ionic detergents and high salt buffer extractions ^{2, 3}. The PT has been subdivided into a subacrosomal and postacrosomal part based on its localization, function, composition, and developmental origin. The subacrosomal part of the PT develops early during spermiogenesis, simultaneously with the formation of the acrosome. It is supposed to emerge from acrosomal vesicles, and presents as a thin cytosolic protein layer between the inner acrosomal membrane and the nuclear envelope ^{4, 5}. The postacrosomal part of the PT, also known as the sperm calyx, originates from cytosolic proteins that are transported via the manchette and assembled during sperm head elongation ⁴. It is located in between the nuclear envelope and the sperm plasma membrane.

Proteomic analyses using murine and bovine PT extracts revealed 500-800 different proteins, highlighting its high molecular complexity ^{6, 7}. Apart from cytoskeletal proteins, signaling molecules and several de novo synthesized core histones were identified. They are proposed to be essential for sperm-egg interaction and chromatin remodeling of the male pronucleus at early post-fertilization stages ^{8, 9, 4, 10, 11}. Many structural proteins of the PT are testis-specific and uniquely expressed in the PT, including Calicin (Ccin) ^{12, 2, 13}, Cylicin 1 (Cylc1) and Cylicin 2 (Cylc2) ^{2, 14, 15}, actin-capping proteins CPβ3 and CPα3 ^16–18 as well as actin-related proteins Arp-T1 and Arp-T2 ¹⁹. Despite being characterized on molecular level, their function remains largely elusive.

Both, Ccin and Cylicins are highly basic proteins with a predominant localization in the calyx of mature sperm ^{14, 15, 12}. In most species, two Cylicin genes, Cylc1 and Cylc2, have been identified. They are characterized by repetitive lysine-lysine-aspartic acid (KKD) and lysin-lysin-glutamic acid (KKE) peptide motifs, resulting in an isoelectric point (IEP) > pH 10 ^{14, 15}. While the CYLC2/Cylc2 gene is encoded on autosomes, the human and murine CYLC1/Cylc1 gene is encoded on the X-chromosome, resulting in hemizygosity in males. Further, in bovine, Cylc2 serves as actin-binding protein of the sperm perinuclear cytoskeleton ²⁰. Cylicins seem to be cytoskeletal regulators and required for proper sperm head architecture.

In this study, we report the CRISPR/Cas9-mediated generation and characterization of Cylc1-, Cylc2-, and Cylc1/2-deficient mice, demonstrating that Cylicins are indispensable for male fertility and play a key function in the formation of the sperm calyx. We show that Cylicins are required for the maintenance of the PT integrity during spermiogenesis and in mature sperm. Their deficiency results in morphological defects of the sperm head, acrosome, and mid-piece. Our analyses revealed that Cylicin genes are evolutionary conserved in rodents and primates. Furthermore, we identified Cylicin variants in an infertile man, demonstrating the conserved role of Cylicins in regulating male infertility.

Results

Cylicins are indispensable for male fertility in mice

Cylicins were first discovered and characterized in the early nineties, and a role in sperm head architecture was postulated due to their subcellular localization ^{2, 14, 15}, however, any functional evidence has been lacking. To address the question of the role of Cylicins during spermiogenesis, we used CRISPR/Cas9-mediated gene-editing to generate Cylc1 and Cylc2 deficient mouse models. First, for Cylc1, a pair of sgRNAs targeting exon 5 was designed, and a mouse line with a frameshift inducing deletion of 1.585 kb, accounting for 85% of the Cylc1 coding sequence, was established. Next, two guide RNAs targeting exon 4/5 of the Cylc2 gene were applied to establish a Cylc2 deficient mouse line with a 1.145 kb frameshift inducing deletion (Fig. 1 A). In both lines, the majority of predicted functional domains with repetitive, lysine-lysine-aspartic acid (KKD) and lysin-lysin-glutamic acid (KKE) peptide motifs were depleted. Depletion was confirmed by PCR-based genotyping (Fig. 1 B).

Figure 1:

Fertility testing of Cylc1^−/y males revealed significantly reduced pregnancy rates (16 %) and mean litter size (2.2) (Fig. 1 C). Cylc2^−/- males were infertile, while Cylc2^+/- males showed no significant difference in fertility parameters compared to wildtype mice (Fig. 1 C). Additionally, established mouse lines were intercrossed to generate Cylc1^−/y Cylc2^+/- and Cylc1^−/y Cylc2^−/- males. Of note, Cylc1^−/y Cylc2^+/-, and Cylc1^−/y Cylc2^−/- males were infertile (Fig. 1 C). This indicates that loss of Cylc1 alone is partially tolerated, as suggested by subfertility of Cylc1^−/y males, whereas the additional loss of one Cylc2 allele renders male mice infertile. Taken together, the results suggest, that two functional Cylicin alleles are required for male fertility.

qRT PCR confirmed the absence of Cylc1 and/or Cylc2 transcripts in Cylicin-deficient animals (Fig. 1 D). For Cylc2^+/- animals a gene-dosage effect was observed, as indicated by a 50 % reduced Cylc2 expression. Neither loss of Cylc1 nor Cylc2 resulted in upregulation of Cylc2 or Cylc1, respectively.

Next, due to the lack of commercial antisera, polyclonal antibodies against murine CYLC1 and CYLC2 were raised to visualize the localization of Cylicins during spermiogenesis. Immunofluorescence staining of wildtype testicular tissue showed presence of both, CYLC1 and CYLC2 from the round spermatid stage onward (Fig. 1D, E). The signal was first detectable in the subacrosomal region as a cap like structure, lining the developing acrosome (Fig. 1D, E). As the spermatids elongate, CYLC1 and CYLC2 move across the PT towards the caudal part of the cell (Fig. S1 A). At later steps of spermiogenesis, the localization in the subacrosomal part of the PT faded while it intensified in the postacrosomal calyx region. (Fig. 1 D, E). Of note the localization of CYLC1 and CYLC2 in the calyx of mature sperm has been reported in bovine and human ^{14, 15, 20}. The generated antibodies did not stain testicular tissue and mature sperm of Cylc1- and Cylc2-deficient males demonstrating i) specificity of the antibodies and ii) validating the knockout mice (Fig. 1 E).

Next, spermiogenesis of Cylicin-deficient males was analyzed in detail. Gross testicular morphology as well as testicular weight were not significantly altered (Fig. 2 A, B). The testicular morphology appeared unaltered, with all stages of spermatogenesis being detectable in PAS-stained testicular sections (Fig. 2 C). However, a strong decline in cauda epididymal sperm counts was observed in all Cylicin-deficient males. For Cylc1^−/y and Cylc2^+/- males, a moderate reduction of 40-47% was determined, whereas Cylc2^−/- and Cylc1^−/y Cylc2^+/- displayed an approx. 65% reduction in epididymal sperm counts compared to wildtype mice (Fig. 2 D). In Cylc1^−/y Cylc2^−/- males, spermiogenesis was most impaired, as indicated by an 85% reduction of the sperm count (Fig. 2 D). Eosin-nigrosine staining revealed that the viability of epididymal sperm from all genotypes was above the lower reference limit of 58% (Fig. 2 E, Fig. S1 B). However, viability of Cylc2^−/- and Cylc1^−/y Cylc2^−/- sperm was reduced by approx. 15% compared to wildtype sperm. (Fig. 2 E). The motility of Cylc1^−/y and Cylc2^+/- sperm remained unchanged compared to wildtype sperm (around 60% motile cells), whereas motility of Cylc2^−/- sperm was drastically reduced to only 7% motile sperm (Fig. 2 F) and the motility of Cylc1^−/y Cylc2^−/- sperm was reduced to 2% motile sperm. In addition, the few motile sperm cells were not progressive but were swimming in circular trajectories. Interestingly, in Cylc1^−/y Cylc2^+/- mice, sperm motility was reduced as well, but less drastically, with 27% of sperm cells being motile (Fig. 2 F).

Figure 2:

Cylc2^−/- sperm cells have altered flagellar beat

The SpermQ software was used to analyze the flagellar beat of Cylc2^−/- sperm in detail ²¹. Cylc2^−/- sperm showed stiffness in the neck and a reduced amplitude of the initial flagellar beat, as well as reduced average curvature of the flagellum during a single beat (Fig. S4 A). Interestingly, we observed that the flagellar beat of Cylc2^−/- sperm cells was similar to wildtype cells with - however – interruptions during which midpiece and initial principal piece appeared stiff whereas high frequency beating occurs at the flagellar tip (Fig. 2G, supplementary Video 1, supplementary Video 2). These interruptions occurred only on the open-hook side and the duration of such interruptions varied from beat to beat. As these anomalies introduce a high asymmetry to the flagellar beating, swimming of such sperm is very likely perturbed limiting the capability to reach the egg ²². Of note similar phenotypes have been observed for sperm with structural defects in the axoneme ²². Thus, we hypothesized that there may be structural defects in Cylc2-deficient sperm.

Sperm morphology is severely altered in Cylicin deficient mice

Indeed, transmission electron microscopy (TEM) of epididymal sperm revealed that Cylc2-deficiency causes severe structural defects (Fig. 3 A, Fig. S2 A): Cylc2^−/- sperm showed coiling of the tail and dislocation of the tail connecting piece from the basal plate, resulting in parallel position of head and tail (Fig. 3 A, white arrowheads). The coiling of the flagellum was also observed using bright field microscopy (Fig. S2 B). Furthermore, in Cylc2^−/- sperm, excess of cytoplasm was observed, located around the nucleus and coiled tail (Fig. 3 A). Anomalies of the head were observed at the level of the PT, while the nuclei appeared unaltered. In all Cylc2^−/- sperm cells, the posterior portion of PT – calyx was absent (Fig. 3 A, red arrowheads). Instead of surrounding the nucleus entirely, the PT in Cylc2^−/- sperm appeared interrupted, missing completely its caudal part. Further, we observed loosening of the peri-acrosomal region, which is not compact and adherent to the nucleus (Fig. 3 A, green arrowheads). Although the axoneme structure appeared unaltered, presenting typical 9+2 microtubular composition, many Cylc2^−/- sperm cells had damaged mitochondria (Fig. S2 A).

Figure 3:

Next, we used PNA-FITC lectin immunofluorescence staining to analyze acrosome localization in mature sperm, as well as MITOred to visualize mitochondria in the flagellum (Fig. 3 B, C). Loss of Cylc1 alone caused malformations of the acrosome in around 42% of mature sperm, while their flagellum appeared unaltered and properly connected to the head. While Cylc2^+/- males showed normal sperm morphology, 94% of Cylc2^−/- mature sperm cells displayed morphological alterations of head and mid-piece (Fig. 3 B, C). Cylc2^−/- sperm cells showed acrosome malformations, bending of the neck region, and/or coiling of the flagellum, occasionally resulting in its wrapping around the sperm head (Fig. 3 B). While 75% of Cylc1^−/Y Cylc2^+/- sperm showed these morphological alterations, 99% of Cylc1^−/Y Cylc2^−/- sperm presented with coiled tail and abnormal acrosome (Fig. 3 B, C).

Loss of Cylicins prohibits calyx assembly during spermiogenesis

Since TEM micrographs revealed absence of calyx in Cylc2^−/- epididymal sperm (Fig. 3 A, red arrowheads), we analyzed the localization of other calyx specific proteins, such as CCIN and CapZα3. In Cylc1^−/y and Cylc2^+/- sperm, CCIN localization remained unchanged, being present in the calyx and in the ventral portion of PT as described previously ⁷. However, in Cylc2^−/- sperm, CCIN localizes to the tail or in random parts of the sperm head (Fig. 3 D). In Cylc1^−/Y Cylc2^+/- and Cylc1^−/Y Cylc2^−/- sperm, the localization of CCIN was also altered, with the signal mainly being present in the sperm tail. CapZα3 forms a heterodimer with CapZβ3, creating a functional complex that localizes in the calyx ²³. Immunofluorescence stainings revealed that the localization of CapZα3 remained unchanged in Cylc1^−/Y and Cylc2^+/-mice compared to wildtype mice. In many Cylc2^−/- sperm cells, CapZα3 localizes in the caudal portion of the head but without regular calyx localization (Fig. 3 D). Interestingly, Cylc1^−/Y Cylc2^+/- mice showed less severe anomalies of the calyx and although CCIN is located almost exclusively in the tail, CapZα3 maintains the correct calyx localization (Fig. 3 D). Finally, Cylc1^−/Y Cylc2^−/- spermatozoa showed CapZα3 localization across the sperm head without regular calyx shape. These results suggest that the loss of Cylicins impairs the formation of calyx and the correct localization of its components, which might cause further morphological anomalies of the sperm (Fig. 3 D).

Loss of Cylc2 impairs sperm nuclear morphology

To analyze in detail the sperm head, we used Nuclear Morphology software on DAPI stained sperm samples. Cylc1^−/Y and Cylc2^+/- sperm showed no alterations of the nuclear shape when compared to wildtype (Fig. 3 E, Fig. S2C, D). However, heads of Cylc2^−/- and Cylc1^−/Y Cylc2^−/- sperm appeared smaller, with shorter hooks and increased circularity of the nuclei as well as reduced elongation (Fig. 3 E, Fig. S2D). This indicates that loss of Cylc2 causes a mild globozoospermia. Interestingly, Cylc1^−/Y Cylc2^+/- sperm heads were not drastically altered, suggesting that only Cylc2 has a crucial role for sperm head shaping, and one functional Cylc2 allele is sufficient to maintain the correct shape of the nucleus (Fig. 3 E, Fig. S2D). Furthermore, the head of many Cylc2-deficient sperm appeared less concave at the region proximal to the connecting piece (Fig. 3 E).

Cylicins are required for acrosome attachment to the nuclear envelope

PNA lectin staining of testicular tissue sections was performed to investigate acrosome biogenesis. During Golgi phase, the acrosome first starts to appear as an aggregation of proacrosomal vesicles into a single granule. This premature acrosomal structure was unaltered in all genotypes, with PNA signal appearing as a small dot on one pole of round spermatids (Fig. 3 F). During cap phase, acrosome grows to cover the apical part of the nucleus. In wildtype and Cylc2^+/- mice, the forming acrosome appears equally smooth and shows a regular cap structure on the perinuclear region of round spermatids. However, in some of the round spermatids from Cylc2^−/- and Cylc1^−/Y mice, gaps in the forming acrosome are observed, as well as an irregular shape of the cap. In Cylc1^−/Y Cylc2 ^+/- and Cylc1^−/Y Cylc2^−/- mice, most of the round spermatids were deformed or displayed irregularly localized caps (Fig. 3 F, Fig. S3A). At acrosome phase, many elongating spermatids of Cylc1^−/Y, Cylc2^−/-, Cylc1^−/Y Cylc2^+/- and Cylc2^−/- Cylc1^−/Y mice have irregular acrosome (Fig. 3 F, Fig. S3A). Detachment of the acrosome from the nuclear envelope is evident in testis samples of Cylc2^−/- and Cylc1^−/Y Cylc2^−/- male mice. These results suggest that Cylicins are required for the attachment of the developing acrosome to the nuclear envelope during spermiogenesis.

Cylicin-deficiency results in abnormal manchette elongation and disassembly

Next, we investigated the role of Cylicins during formation and development of the manchette – a sperm-specific, transient structure that represents a microtubular platform for protein transport. Transport of the intracellular vesicles is crucial for the formation of the flagellum, acrosome assembly and removal of excess cytoplasm. The manchette is first detected at step 8 at the perinuclear ring of round spermatids, just prior to their elongation. During the next steps of spermiogenesis, as the spermatids elongate, manchette moves towards the neck region in a skirt-like structure and starts disassembling at step 13 when the elongation is complete ²⁴. We used immunofluorescence staining of acetylated α-tubulin on squash testis samples containing spermatids at different stages of spermiogenesis to investigate whether the altered head shape, calyx structure, and tail-head connection anomalies originate from possible defects of the manchette structure. Spermatids starting from stage 8 were observed individually for step-to-step comparison. In all genotypes, a cap-like shape of the manchette was observed in step 8 round spermatids, suggesting that the manchette assembles properly and starts elongating towards the neck region during step 9 (Fig. S3 B). In all genotypes, the typical skirt-like structure was observed at the caudal region of the spermatids at step 10 and 11, suggesting that the manchette assembles correctly even in Cylicin-deficient sperm (Fig. 4 A). In spermatids from Cylc1^−/Y and Cylc2^+/- mice, regular manchette development was observed in further steps of spermiogenesis (Fig. 4 A). However, starting from step 12, spermatids from Cylc2^−/-, Cylc1^−/Y Cylc2^+/-, and Cylc1^−/Y Cylc2^−/- mice showed abnormal manchette elongation, which became more prominent at step 13 (Fig. 4 A). Interestingly, some of Cylc2-deficient spermatids showed shifting of the manchette to the ventral side of the nucleus along with excessive elongation. At step 16, the manchette was normally disassembled in wildtype, Cylc1^−/Y, and Cylc2^+/- spermatids (Fig. 4 A). However, Cylc2 ^−/-, Cylc1^−/Y Cylc2^+/-, and Cylc1^−/Y Cylc2^−/- spermatids showed a persistent α-tubulin signal, indicating that disassembly of the manchette is delayed or incomplete (Fig. 4 A).

Figure 4:

Other than α-tubulin, we also used HOOK1 as manchette marker. HOOK1 is a member of a family of coiled-coil proteins, which bind to microtubules and organelles and regulate microtubule trafficking during endocytosis and spermiogenesis ²⁵. Co-staining of the spermatids with antibodies against PNA lectin and HOOK1 revealed that abnormal manchette elongation and acrosome anomalies simultaneously occurred in elongating spermatids of Cylc2^−/- male mice (Fig. 4 B): While manchette was moving excessively towards the tail, the developing acrosome lost its compact adherence to the nuclear envelope (Fig. 4 B), causing the concomitant anomalies in the mature sperm. Schematic representation shows acrosome biogenesis and manchette development in wildtype and Cylc2^−/- spermatids (Fig. 4 B). While round spermatids of Cylc2^−/- mice elongated as those of the wildtype sperm, the manchette elongated abnormally and the acrosome became loosened (Fig. 4 B).

Transmission electron microscopy of testicular tissues of wildtype and Cylc2^−/- mice showed detailed structure of spermatids at different steps of spermiogenesis (Fig. 4 C). Round spermatids at step 6 showed correct assembly of the proacrosomal granule typical for Golgi phase, as well as normal looking PT in both wildtype and Cylc2^−/- mice (Fig. 4 C). However, in Cylc2^−/- elongating spermatids already at step 8, we observed wide gaps in the PT at the perinuclear ring level, between the growing acrosome (Fig. 4 C, green arrowheads) and the developing manchette (Fig. 4 C, red arrowheads; gaps are depicted by dashed line). In wildtype spermatids, these structures were in close contact (Fig. 4 C). Acrosome biogenesis proceeded regularly with acrosomal caps forming at the apical pole of the spermatids. At step 9-10 spermatids of Cylc2^−/- mice we already observed manchette anomalies. In many cells, microtubules appeared longer on one side of the nucleus than on the other, displacing the acrosome to the side. In turn, this might result in abnormal attachment of the developing flagellum to the basal plate and consequently flipping of the head and coiling of the tail in mature spermatozoa (Fig. 4 C). At step 11-12, excessive elongation of the manchette in Cylc2^−/- spermatids as well as the gaps in PT were observed. Whereas elongated spermatids at step 14-15 in wildtype sperm already disassembled their manchette and the PT appeared as a unique structure that compactly surrounds nucleus, in Cylc2^−/- spermatids, remaining microtubules failed to disassemble, and the acrosome detached from the nuclear envelope. In addition, at step 16, the calyx was absent, and an excess of cytoplasm surrounded the nucleus and flagellum (Fig. 4 C, yellow arrowhead). Furthermore, many damaged and degrading cells were observed in testicular tissue TEM samples, having perforated nuclei and detached structures (Fig. 4 D). Interestingly, phagosomes with cellular remains were observed far away from the lumen and sometimes even at the basal membrane of the tubuli, suggesting that the cells that suffer most severe structural damage are being removed. This mechanism of removing malformed cells explains the reduction of epididymal sperm count in Cylicin-deficient genotypes.

We conclude that loss of two Cylicin alleles does not disturb the initial phase of manchette formation but causes the delocalization of the manchette towards the tail and abolishes timely manchette disassembly. The localization of manchette on the caudal portion of spermatids coincides with the localization of the calyx in the mature sperm, so these results indicate that the manchette might be maintained longer to compensate for the missing formation of the calyx structure.

Cylc2 coding sequence is slightly more conserved among species than Cylc1

To address why Cylc2 deficiency causes more severe phenotypic alterations than Cylc1-deficiency in mice, we performed evolutionary analysis of both genes. Selective pressures on protein level were represented by calculation of the nonsynonymous/synonymous substitution rate ratio (ω = dN/dS). It distinguishes between purifying selection (ω<1), neutral evolution (ω=1) and positive selection (ω>1). Analysis of the selective constrains on Cylc1 and Cylc2 across rodents and primates revealed that both genes’ coding sequences are under purifying selection in general, although conservation is weaker in Cylc1 trending towards a more relaxed constraint (Fig. S4 A, Table S1). For Cylc1, the evolutionary rate (ω) of the whole sequence for the whole tree analyzed was significantly lower than, but close to 1 at ω=0.85. For Cylc2, it was lower at ω=0.49 (see Table S1). A model allowing for separate calculation of the evolutionary rate for primates and rodents, did not detect a significant difference between the clades, neither for Cylc1 nor for Cylc2, indicating that the sequences are equally conserved in both clades.

To analyze the selective pressure across the coding sequence in more detail, we calculated the evolutionary rates for each codon site across the whole tree. The model allowing for positive selection on codon sites explained the data significantly better than the null model according to likelihood ratio analysis (see Table S1). For Cylc1, 34% of codon sites were assigned to the conserved site class, 51% to the neutral/relaxed constraint site class, and 15% to the positive selection site class. For Cylc2, 47% of codon sites were assigned to the conserved site class, 44% to the neutral/relaxed constraint site class, and 9% to the positive selection site class. Of the codon sites assigned to the positive selection site class three sites were significantly positively selected for Cylc1 and 11 for Cylc2 (Table S1). The distribution of evolutionary rates across the sequence revealed largely relaxed constraint in the N-terminal region with several sites of positive selection and purifying selection with positive selection hot spots in the C-terminal region of Cylc1. The N-terminal region of Cylc2 seems to be largely conserved with a pattern of conservation and stronger positive selection in the internal and C-terminal region. Interestingly, codon sites encoding lysine residues, which are proposed to be of functional importance for Cylicins, are mostly conserved. For Cylc1, 17% of lysine residues are significantly conserved and 35% of significantly conserved codons encode for lysine. For Cylc2, this pattern is even more pronounced with 27.9% of lysine codons being significantly conserved and 24.3% of all conserved sites encoding for lysine (Fig. S4).

Cylicins are required for normal sperm morphology in human

As loss of two Cylicin alleles causes fertility defects in mice, we next addressed whether infertile men also display variants in CYLC1/CYLC2.

Exome sequencing within the MERGE (Male Reproductive Genomics study) cohort identified one patient (M2270) carrying rare (MAF < 0.01, gnomAD) missense variants in both CYLC1 and CYLC2. The hemizygous variant c.1720G>C in CYLC1 leads to an amino acid exchange from glutamic acid to glutamine (p.(Glu574Gln)) which is predicted to be deleterious or possibly damaging by in silico tools (SIFT ²⁶ and PolyPhen ²⁷, respectively) and has a CADD score of 11.91. It is located in exon four out of five and affects a region that is predicted to be intolerant to such substitutions (Figure S5, metadome). It occurs only twice in the gnomAD database (v2.1.1) comprising 141,456 individuals (67,961 XY), once identified in a hemizygous male and once in a female carrier, and is absent from our in house database.

The heterozygous variant c.551G>A in CYLC2 is predicted to be tolerated (SIFT) or benign (PolyPhen) which is in accordance with a low CADD score of 0.008. It is located in exon five out of eight and affects a region in which variants are likely to be tolerated (Figure S5, metadome). However, it is a rare variant occurring with an allele frequency of 0.0035 in the general population, according to the gnomAD database. Importantly, only three XX individuals are reported to be homozygous for the variant within CYCL2.

Segregation analyses revealed maternal inheritance of the X-linked CYLC1 variant c.1720G>C p. (Glu574Gln), whereas the father carries the heterozygous CYLC2 variant c.551G>A p.(Gly184Asp) (Fig. 5A). According to ACMG-AMP criteria ²⁸ both variants are classified as variants of uncertain significance (Table S2). No other potentially pathogenic variants in genes associated with sperm morphological defects were identified by exploring the exome data of M2270.

Figure 5:

Semen analysis performed following WHO guidelines ²⁹ is shown in Table 1. The sperm concentration in the semen was slightly reduced, while significant reduction of motile spermatozoa (12.5%) was observed. Interestingly, only 2% of the sperm cells appeared morphologically normal, while 98% of sperm showed head defects (Table 1). Immunofluorescence staining of CYLC1 revealed that while in healthy donor’s sperm, CYLC1 localizes in the calyx, in M2270, CYLC1 labeling was absent (Fig. 5B). Bright field microscopy demonstrated that M2270’s sperm flagella coiled similarly to that in Cylicin deficient murine sperm samples. Interestingly, DAPI staining revealed that many flagella carry cytoplasmatic bodies without nuclei and could be defined as headless spermatozoa (Fig. 5C). CCIN staining demonstrated that while spermatozoa of a healthy donor showed a typical, funnel-like calyx structure in the posterior region of PT, spermatozoa from M2270 had CCIN localized in irregular manner throughout head and tail, suggesting that Cylicins have a role in maintenance of the calyx structure and composition in both mice and human spermatozoa (Fig. 5D).

Table 1:

Discussion

Spermiogenesis is a highly organized process that is dependent on a unique cytoskeletal organization of the sperm cells. The PT has a role of structural scaffold that surrounds sperm nucleus, and its protein composition is crucial for correct sperm development. In this study, we used CRISPR/Cas9 gene editing to establish Cylc1-, Cylc2- and Cylc1/2-deficient mouse lines to analyze the role of Cylicins. We demonstrated that the loss of Cylicins impairs male fertility in mice by severely disturbing sperm head architecture. The significance of our findings is supported by the identification of CYLC1/2 variants in an infertile patient who presents similar structural anomalies in sperm cells.

Detailed analysis of sperm development and morphology demonstrated that infertility of Cylicin-deficient male mice is caused by anomalies of different sperm structures. The prominent loss of PT integrity and calyx structure in the mature sperm lead to acrosome loosening and detachment from the nuclear envelope. Furthermore, the shifting of the basal plate caused damage in the head-tail connecting piece resulting in coiling of the flagellum and impaired swimming. Murine Cylicins are first detected in the periacrosomal region of round spermatids and move to the postacrosomal region of PT during spermatid elongation to finally localize to the calyx of mature sperm. This dynamic localization pattern of Cylicins has been described in human, boar and bovine sperm as well ^{2, 14, 15}. Interestingly, other PT enriched proteins CCIN ^{12, 30, 7} and CPβ3/CPα3 move in the similar manner across PT during spermiogenesis and are present in the calyx of mature sperm of various mammalian species. This co-localization suggests the potential interaction between calyx proteins. Furthermore, CCIN ^{2, 12, 13} and CPβ3-CPα3 complex ^16–18 are described as actin-binding proteins and porcine Cylc2 has been shown to have a high affinity for F-actin as well ²⁰. The potential roles of F-actin during spermiogenesis in mammals involve biogenesis of the acrosome ³¹ and its correct attachment to the outer nuclear membrane of the spermatids ³² as well as removal of excess cytoplasm ³³.

The loss of Cylicins caused acrosome detachment from NE starting from cap phase of the acrosome biogenesis. Interestingly, loss of CCIN results in similar loosening of the acroplaxome from the outer nuclear membrane ⁷. Calicin is shown to be necessary for the IAM-PT-NE complex by establishing bidirectional connections. Based on our findings, we conclude that both CCIN and Cylicins need to interact in order to provide the ‘molecular glue’ necessary for the acrosome anchoring. Furthermore, Cylicin-deficiency resulted in cytoplasmatic retention and bending of the midpiece, similarly to the CPα3 mutant phenotype repro32/repro32 ³⁴. The CAPZ complex has been shown to regulate the actin dynamics by preventing addition or loss of G-actin subunits to the ends of F-actin filaments ³⁵. Since we demostrated that the loss of Cylicins resulted in mislocalization of both CCIN and CPα3, we speculate that Cylicins have a crucial role in maintenance of the integrity of PT structure and thus are required for proper function of PT proteins.

Other than morphological defects of the mature sperm PT, during spermiogenesis Cylicin-deficiency results in excessive manchette elongation, its delayed disassembly as well as formation of abnormal gaps in the PT at the level of perinuclear ring. The intra-manchette transport (IMT) of proteins from the apical pole of the head to the base of the developing tail occurs during spermiogenesis ³⁶. The malfunctions of the IMT can cause delay in manchette clearance and morphological defects of mature sperm. Mouse models deficient for IMT proteins such as HOOK1 ²⁵ CEP131 ³⁷ and IFT88 ³⁸ show abnormal manchette elongation and delay in its clearance resulting in aberrant nuclear shape of the sperm. Furthermore, the loss of HOOK1 results in head-tail connection and mid-piece anomalies such as flipping of the head, basal plate defects and coiling of the tail similar to Cylc2-deficiency ³⁹. Furthermore, we observed wide gaps in the perinuclear ring during manchette elongation, suggesting that Cylicins might have a role in maintenance of the contact between caudal and apical PT region and its integrity.

Our evolutionary analysis of Cylc1 and Cylc2 genes across rodents and primates revealed that both coding sequences are under purifying selection. Overall, the results reveal that Cylc1 is under slightly less conserved constraint than Cylc2 leading to the hypothesis that loss of function in Cylc1 might be less severe and could be compensated for by Cylc2 due to partial redundancy. This hypothesis is supported by our finding that Cylc1-deficiency causes subfertility in male mice while the loss of both Cylc2 alleles results in male infertility. Furthermore, in Cylc2^+/- male mice fertility was preserved while Cylc1^−/Y Cylc2^+/- males were unable to sire offspring leading to the conclusion that the loss of one Cylc2 allele could be compensated by Cylc1, however at least two functional Cylicin alleles are required for male fertility in mice.

In general, the evolutionary rate of C-terminal lysine rich region of both Cylicins seems to be highly volatile between conserved and positively selected codon sites, while the lysine residues seem to be strongly conserved. Changes in the C-terminal region, potentially affecting the length of the lysine rich domain, might have an adaptive advantage. Targeted positive selection on codon sites could also be a signature of co-evolution with a rapidly evolving interactor.

Sperm morphological defects and infertility observed in one patient with variants in both Cylicin genes points towards a requirement for human spermiogenesis and fertility. A defect in acrosome formation and the existence of variants in two out of three alleles in both CYLC genes is in line with the observations made in mice. The absence of CYLC1 confirmed through immunofluorescent staining indicates a functional impact of the missense variant on the CYLC1 protein. Furthermore, impaired CCIN localization was observed in patient M2270 as well, suggesting that PT has similar roles in human and rodents despite the differences in sperm head shape ⁴⁰. These results suggest that establishing mouse models deficient for Cylicin and other PT proteins might provide insights into mechanisms of human spermiogenesis and cytoskeletal regulation as well. However, with our data we cannot exclude the possibility that there is a discrepancy between mice and men and that the missense variant in CYLC1 might alone be sufficient to cause the observed phenotype of M2270. The o/e ratios of both genes calculated within the gnomAD database rather indicate a stronger selective pressure on human CYLC1 (0.08) than on CYLC2 (0.84). Furthermore, the CYLC1 variant affects an intolerant region within the protein sequence according to metadome ⁴¹, whereas the CYLC2 variant is located in a region that is more tolerant to variation (Fig. S5). Therefore, we cannot definitively validate the hypothesis of an oligogenic disease in men as well.

The CYLC2 missense variant is inherited by the father of M2270 (Fig. 5A) and, thus, not sufficient to cause infertility. However, the father reported difficulties to conceive naturally. Based on this one family, we cannot exclude an effect of pathogenic heterozygous CYLC2 variants.

Overall, the identification and detailed characterization of further patients with variants in CYLC1 and CYLC2 is warranted to draw firm conclusions on the effect of variants in these genes on spermiogenesis and infertility.

Acknowledgements

This study was supported by a grant from the German Research Foundation (DFG) to HS (SCHO 503/23-1), SS (SCHN 1668/1-1) and FT (Clinical Research Unit ‚Male Germ Cells’, CRU326). We are grateful to Gaby Beine, Angela Egert, Andrea Jäger, Greta Zech, Luisa Meier and Christina Burhöi for excellent technical assistance. We would like to thank the Core Facility for Microscopy of the Medical Faculty at the University of Bonn for providing support and instrumentation funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation, project number: 388169927)

Author Contributions

S.S. and H.S. conceived and designed the experiments. S.S. generated gene-edited mouse models. S.S., A.K. and M.M. analyzed Cylc1 and Cylc2 deficient mice. L.A. performed evolutionary analyses. A.K.D., S.A.K., C.B. and F.T. analyzed and provided exome sequencing data of infertile men. S.S., A.K. and H.S. were major contributors in writing the manuscript. All authors read and approved the final manuscript.

Methods

Animals

All animal experiments were conducted according to German law of animal protection and in agreement with the approval of the local institutional animal care committees (Landesamt für Natur, Umwelt und Verbraucherschutz, North Rhine-Westphalia, approval IDs: AZ84-02.04.2013.A429, AZ81-02.04.2018.A369). Cylc1- and Cylc2-deficient mice were generated by CRISPR/Cas9 mediated gene-editing in zygotes of the hybrid strain B6D2F1. Guide sequences are depicted in Table 2. For Cylc1, in vitro transcribed sgRNAs (50 ng/µl each) and Cas9 mRNA (100 ng/µl) (Sigma-Aldrich, Taufkirchen, Germany) were microinjected into the cytoplasm of zygotes as described previously ⁴². Cylc2 deficient mice were generated by electroporation of ribonucleoprotein (RNP) complexes into zygotes using a GenePulser II electroporation device (BioRad, Feldkirchen, Germany). For RNP formation, crRNA and tracrRNA (IDT, Leuven, Belgium) were combined in duplex buffer (IDT) to a final concentration of 50 mM each and annealed (95°C, 5 min; cool down to room temperature with -0,2°C/sec.). For RNP assembly, 4 µM Cas9 protein (IDT) and 2 µM of each annealed cr/tracr RNA were combined in Opti-MEM media (Gibco 11058-021, Thermo Fisher, Carlsbad, USA) and incubated for 10 min at room temperature. RNPs complexes were diluted 1:2 in Opti-MEM, supplemented with 30-40 zygotes in a 0.1 cm gene-pulser cuvette (BioRad) and electroporated (two 30 V square wave pulses, 2 ms pulse length, 100 ms pulse interval). Recovered embryos were cultured over night at 37°C, 5% CO₂ in G-TL medium (Vitrolife, Göteborg, Sweden) and transferred into the oviduct of pseudopregnant CB6F1 foster mice at 2-cell stage. Offspring were genotyped and gene-edited alleles were separated by backcrossing of founder animals with C57Bl/6J mice. Established mouse lines were registered with Mouse Genome Informatics: Cylc1^em1Hsc (MGI: 7341368), Cylc2^em1Hsc (MGI:6718280). Mouse lines were maintained as congenic strains on C57Bl/6J background. For all analyses sexually mature males at the age of 8-15 weeks, backcross generation ≥N3 were used.

Table 2:

Genomic DNA extraction and genotyping

Genomic DNA was extracted from biopsies using the HotShot method ⁴³. PCR reactions were assembled according to the manufacturers protocol of the DreamTaq Green DNA Polymerase (Thermo Fisher, EP0712) using gene specific primers listed in Table 3.

Table 3:

Fertility analysis

Males were mated 1:2/1:1 with C57Bl/6J females, which were checked daily for presence of a vaginal plug indicative for successful copulation. Plug positive females were separated and pregnancies as well as litter size were recorded.

Sampling

Mature sperm were obtained from the cauda epididymis in M2 medium (Merck) or PBS preheated at 37°C. Following several incisions of the cauda, sperm were retrieved by flush-out for 15-30 min. The extracted sperm samples were washed in PBS, centrifuged (4000 rpm, 5 min, 4°C) and snap frozen. For squash tubule samples, fresh testes were transferred to PBS and tubules were dissected as described by ⁴⁴. Tubule sections were pressed on a slide, quickly frozen in liquid nitrogen and fixed in 90% ethanol for 5 min.

Semen analysis

Sperm concentrations for at least 4 animals of each genotype were determined using a Neubauer hemocytometer. Viability of sperm was determined by Eosin-Nigrosine staining as described previously ⁴⁵ for at least 3 animals per genotype. Sperm motility was analyzed referring to the WHO guidelines for analysis of human semen. For all analyses at least 100 sperm per individual were analyzed and the percentage of motile/immotile, viable/inviable sperm was calculated.

Quantitative reverse transcription-polymerase chain reaction (qRT-PCR)

RNA was extracted from the testis tissue with TRIzol (Life Technologies, Carlsbad, USA). RNA concentrations and purity ratios were determined by NanoDrop ONE (Thermo Scientific) measurements. qRT-PCR was performed as described previously ⁴², on ViiA 7 Real Time PCR System (Applied Biosystems) using Maxima SYBR Green qPCR Mastermix (Thermo Fisher). Beta-actin was used as a housekeeping gene for normalization. Primers used are shown in table 4.

Table 4:

High-resolution microscopy of the flagellar beat

Image sequences of mouse sperm were acquired using dark-field at an inverted microscope (IX71; Olympus, Hamburg, Germany), equipped with a dark-field condenser and a high-speed camera (ORCA-Flash4.0 V3, C13220-20 CU, Hamamatsu, Hamamatsu City, Japan). A 10x objective (NA 0.4, UPlanFL; Olympus, Hamburg, Germany) with an additional 1.6x magnifying lens (Olympus, Hamburg, Germany) that was inserted into the light path (final magnification: 16x) was applied. Image sequences were recorded at a rate of 200 frames per second (fps). A custom-made observation chamber was used ²¹. Sperm samples were diluted in THY buffer shortly before insertion of the suspension into the observation chamber.

Sperm nuclear morphology

For the analysis of sperm nuclear morphology, epididymal sperm samples from 3 animals of each genotype were fixed in methanol and acetic acid (3:1). The samples were spread onto a slide and stained with 4’,6-diamidino-2-phenylindole (DAPI) containing mounting medium (ROTImount FluorCare DAPI (Carl Roth GmbH, Karlsruhe, Germany)). The sperm cells were imaged at 100x magnification, using a Leica DM5500 B fluorescent microscope. At least 200 pictures were taken from each group and analyzed using Nuclear Morphology software by ⁴⁶ according to the developer’s instructions. The minimum detection area was set to 1.000 pixels while the maximum detection area was 7.000 pixels.

Histology

Bouin fixed testis tissues were paraffinized, embedded and sectioned at 3-5 µm. For histological analysis, the sections were deparaffinized and stained with Periodic acid Schiff (PAS staining). Stained sections were imaged at 20x and 63x magnification under bright field using 5500 B microscope.

Immunofluorescence

Bouin fixed testis tissue sections were deparaffinized in Xylene and rehydrated in decreasing alcohol to be used for immunohistochemical staining. Squash testis samples fixed in 90% EtOH were used for the staining of a-Tubulin and HOOK1. For sperm immunofluorescence, mature sperm cells isolated from cauda epididymis were fixed with methanol acetic acid (3:1), dropped on glass slides and air dried. After washing in PBS twice, all samples were permeabilized using 0.1% Triton x100 for 10 min at room temperature. The samples were then blocked with 5% BSA for 30 minutes, followed blocking with normal horse serum (Vectorlabs, Burlingame, USA) for 30 minutes at room temperature. For tissue sections, heat-activated antigen retrieval was performed using buffers at pH6. All primary antibodies were incubated over night at 4°C. Antibodies and dilutions are listed in table 5. The respective secondary antibodies were incubated for 1 h at room temperature using VectaFluor Labeling Kit DyLight 488 and DyLight 594 (Vectorlabs, Burlingame, USA). All slides were mounted with DAPI containing mounting medium (ROTImount FluorCare DAPI, Carl Roth).

Table 5:

For the analysis of acrosome biogenesis peanut agglutinin (PNA)-fluorescein isothiocyanite (FITC) Alexa Fluor 488 conjugate (Molecular Probes, Invitrogen, Waltham, USA) was used on the Buin fixed testis tissues. After permeabilization and blocking, the tissues were incubated with PNA-FITC 5 μg/ml for 30 minutes at room temperature. Mature sperm were fixed with paraformaldehyde (4%) for 20 min at room temperature. After PBS washing, the sperm samples were incubated with 5 μg/ml PNA-FITC and 5 nM Mito Red (53271; Sigma Aldrich) for 30 min at room temperature. The slides were then mounted with DAPI mount.

Transmission electron microscopy

For Transmission Electron Microscopy fresh epididymal sperm and testis tissue were used. After washing with PBS, the samples were fixed in 3% glutaraldehyde at 4°C overnight. The samples were then washed in 0.1 M cacodylate buffer and fixed again in 2% osmium tetroxide at 4% for 2 h. After dehydration in an ascending ethanol row, the samples were contrasted in 70% ethanol 0.5% uranyl acetate for 1.5h at 4°C. Samples were then washed in propylenoxide, three times for 10 minutes at room temperature before embedding in Epon C at 70°C for 48h. Ultrathin sections were examined using a Verios 460L microscope (FEI) with a STEM III-detector.

Evolutionary analysis

Evolutionary rates of mammalian Cylicin genes were analyzed according to ⁴⁷. Briefly, Cylc1 and Cylc2 nucleotide sequences were obtained from NCBI genbank and Ensembl genome browser. Phylogenetic trees of considered species were built according to the “Tree of Life web project”. The webPRANK software was applied for codon-based alignment of orthologous gene sequences and results were visualized using the ETE toolkit. To determine evolutionary rates of gene sequences across mammals, for different clades and for individual codon sites, the codeml application implemented in the PAML software was used ^{48, 49}. The M0 model served as basis for all performed analyses. Different codon frequency settings were tested for the M0 model of each gene and the setting with the highest likelihood was chosen. To test whether alternative models describe the selective constraints within a dataset more precise than the M0 model, likelihood ratio-tests (LRT) were performed. Applied models and LRTs are specified in Table S1.

Study cohort and ethical approval

The MERGE (Male Reproductive Genomics study) cohort currently comprises over 2,030 men, mainly recruited at the Centre of Reproductive Medicine and Andrology (CeRA) in Münster. The large majority has severe spermatogenic failure, i.e. severe oligozoospermia (<5 Mill./ml sperm concentration), crypto- or azoospermia. So far, only 35 cases were included because of notable sperm morphological defects (≥5 Mill./ml sperm concentration, <4 % normal forms). Common causes for infertility such as oncologic diseases, AZF deletions or chromosomal aberrations were ruled out in advance. Patients with aetiologically still unexplained infertility underwent whole exome sequencing.

All participants gave written informed consent according to the protocols approved by the Ethics Committee of the Medical Faculty Münster (Ref. No. MERGE: 2010-578-f-S) in accordance with the Declaration of Helsinki in 1975.

Whole exome sequencing and data analysis

After DNA extraction from patients’ peripheral blood lymphocytes WES was performed as previously described ⁵⁰. WES data obtained from 2,066 infertile men was filtered for rare (≤ 0.01 minor allele frequency, gnomAD) variants located within the coding sequence or the adjacent 15 bp of flanking introns in CYLC1 and CYLC2. Patients carrying only one variant in either of the gene were excluded. In case of either bi-allelic CYLC2 variants or a combination of CYLC1 and CYLC2 variants, the whole exome dataset was screened to rule out other potential genetic causes. Variants detected in this study were classified according to the guidelines by the American College of Medical Genetics and Genomics-Association for Molecular Pathology (ACMG-AMP) ²⁸ adapted to recent recommendations as outlined in ⁵¹.

To rule out an alternative genetic cause for the patient’s condition, his exome data was screened for rare (MAF ≤ 0.01, gnomAD and in-house database), homozygous, X-linked or potentially compound heterozygous high-impact variants (stop gain, start lost, stop lost, frameshift, splice site and splice region as well as missense variants with CADD > 20) and rare, heterozygous LoF variants (stop gain, start lost, stop lost, frameshift, splice site) without filtering for a specific set of genes. Respective genes were screened for testis-expression and reports of infertility.

Sanger sequencing

Sanger sequencing with variant specific primers was conducted for validation and segregation purposes. The primers to confirm the CYLC1 (NM_021118.3) variant c.1720G>C p.(Glu574Gln) are 5’-ACTGATGCTGACTCTGAACCG-3’ (forward) and 5’-CCTTCGAGTCACAAAGTTGTCA-3’ (reverse). To confirm the CYLC2 (NM_001340.5) variant c.551G>A p.(Gly184Asp), the primers 5’-CTGTCGAGGTGGATTCTAAAGC-3’ (forward) and 5’-TGCATCCTTCTTTGCATCCT-3’ (reverse) were used.

Analysis of the human sperm samples

Human ejaculate samples from healthy donor and patient M2270, were analyzed according to WHO guidelines prior to washing in buffer and centrifuged (1000 rpm, 20 min). The cells were fixed in methanol and acetic acid (3:1) and used for immunofluorescence staining. Samples were dropped on slides and permeabilized with 0.1% Triton x100. After blocking with 5% BSA for 30 min, slides were incubated with primary CCIN or CYLC1 antibodies (concentrations shown in Table 3) for 3 h at room temperature. Secondary antibodies were incubated for 1 h, followed by mounting with DAPI containing medium.

Statistics

For all analyses mean values +/- SD were calculated. Statistical significance was determined by One-way ANNOVA analyses using Bonferroni correction.