Loss of DRC1 function leads to multiple morphological abnormalities of the sperm flagella and male infertility in human and mouse

Abstract Motile cilia and flagellar defects can result in primary ciliary dyskinesia, which is a multisystemic genetic disorder that affects roughly 1:10 000 individuals. The nexin-dynein regulatory complex (N-DRC) links neighboring doublet microtubules within flagella, serving as a central regulatory hub for motility in Chlamydomonas. Herein, we identified two homozygous DRC1 variants in human patients that were associated with multiple morphological abnormalities of the sperm flagella (MMAF) and male infertility. Drc1−/−, Drc1R554X/R554X and Drc1W244X/W244X mice on the C57BL/6 background suffered from pre-pubertal mortality. However, when the ICR background was introduced, some of these mice were able to survive and recapitulate the MMAF phenotypes detected in human patients. By analyzing these animals, we determined that DRC1 is an essential regulator of N-DRC assembly in cilia and flagella. When DRC1 is absent, this results in the shortening of cilia and consequent impairment of their motility. Damage associated with DRC1 deficiency in sperm flagella was more pronounced than in cilia, as manifested by complete axoneme structural disorder in addition to the loss of the DRC structure. Altogether, these findings suggest that DRC1 is required for the structural stability of flagella but not cilia, emphasizing the key role of this protein in mammalian species.


Introduction
Flagella and motile cilia are evolutionarily ancient structures present in prokaryotic and eukaryotic cells that are involved in sensation and movement (1). These organelles also play essential roles in the development and functionality of key systems including the respiratory, nervous and reproductive systems (2). Genetic defects in the motility of flagella and cilia can cause a multisystem disorder known as primary ciliary dyskinesia (PCD) (3,4), which affects roughly 1:10 000 individuals globally (5,6). The nexin-dynein regulatory complex (N-DRC) functions by linking neighboring doublet microtubules within motile cilia and flagella, stabilizing the axonemal core structure and thereby regulating ciliary motility (7)(8)(9)(10)(11)(12)(13). In the flagella of Chlamydomonas cells, at least 11 subunit proteins compose the N-DRC (7,14), and many homologs of these proteins are present in humans and mice. The deletion of DRC genes in mice can result in an array of distinct phenotypes. For example, Drc7 (15) and Drc9 (16) knockout mice exhibit multiple morphological abnormalities of the sperm flagella (MMAF) and male infertility, whereas Drc5 deletion (17) causes asthenospermia and the deletion of Drc6 (15) does not induce significant flagellar abnormalities. A single mutation at amino acid position 89 (Leu89Pro) in DRC3 (LRRC48) can result in classical symptoms of PCD such as mucus accumulation, male sterility and an increased risk of postnatal death (18).
Analyses of Chlamydomonas have identified DRC1, DRC2 and DRC4 as core N-DRC structural components (11). In mice, the loss of DRC4 expression is associated with severe hydrocephaly and lethality within 14-21 days after birth (19). This suggests that this core N-DRC protein is essential for murine survival. No mice harboring knockout mutations for core DRC genes that survive into adulthood have been described to date. However, one case report published in 2013 described the case of a PCD patient exhibiting a homozygous nonsense mutation (c.2056A > T) in DRC1 that was predicted to cause the pre-mature translational arrest of the nascent DRC1 peptide (p.Lys686 * ) (20). That patient suffered from symptoms including chronic otitis media, sinusitis, recurrent pneumonia and neonatal respiratory distress, all of which are consistent with the diagnosis of PCD. In another recent study, a patient with diffuse panbronchiolitis was found to harbor a large homozygous deletion spanning exons 1-4 of DRC1 (21). In both of these cases, the patients survived to adulthood. No Drc1 knockout mice have yet been developed to explore the functional role of this protein in mammals to date.
Herein, we found that two homozygous DRC1 mutations resulted in MMAF and male infertility phenotypes in humans without causing any concomitant respiratory symptoms. While Drc1 −/− , Drc1 R554X/R554X and Drc1 W244X/W244X mice on the C57BL/6 background suffered from pre-pubertal death in all cases, the introduction of the ICR background into these strains enabled some of the mice to survive and to exhibit development and behavior similar to that of wild-type (WT) control mice. The surviving mice recapitulated the MMAF phenotypes observed in the context of human DRC1 deficiency. Interestingly, we further found that damage to sperm flagellum was more severe than ciliary damage in Drc1 R554X/R554X and Drc1 W244X/W244X mice, resulting in damage beyond just the loss of the DRC structure. Altogether, our data suggest that DRC1 mutations can result in dysregulated cilia and flagella formation in humans and mice while also indicating that this protein regulates the structural homeostasis of sperm flagella.
We next generated antisera specific for the homologous Nterminal (1-146 aa) region of human and murine DRC1, which we used to determine that no DRC1 protein was detectable in the sperm of patient aIV-1 ( Fig. 2A), consistent with nonsensemediated mRNA decay (NMD) (41)(42)(43). Immunofluorescence analyses similarly confirmed the absence of DRC1 in sperm from this patient, while Ac-tubulin staining revealed that DRC1 R554X/R554X sperm presented with short, coiled, absent or irregular flagella ( Fig. 2B and C, Supplementary Material, Fig.  S1A and B). The sperm of patient bIV-1 exhibited a phenotype similar to that of Patient aIV-1 (Supplementary Material, Fig.  S1C). TEM analyses revealed the disordered structure of the flagellar axoneme in DRC1 R554X/R554X sperm, with microtubules being scattered in the cytoplasm and with normal centriole implantation and implantation nest formation ( Fig. 2D-I). Axoneme structural disorder was also evident in DRC1 R80X/R80X sperm (Supplementary Material, Fig. S1C). Following ICSI treatment, Patient aIV-1 was able to obtain fertilized embryos, and one of which was successfully implanted (Table 2), while     S3B) was generated. Unexpectedly, mice harboring both of these mutations experienced pre-puberal death, usually before postnatal day 12 (Table 3). These mice exhibited clear signs of growth retardation and hydrocephaly (Supplementary Material, Fig. S4). As the Cas9 system has the potential to introduce off-target genetic changes and these two murine lines were constructed  using the same sgRNA, we additionally attempted to generate an additional murine line simulating human DRC1 mutations using base editing technology. After selecting mutation sites in the gnomAD database capable of causing a loss of DRC1 function and analyzing potential sgRNA designs and conserved residues, we determined that the W244X model would be wellsuited to this base editing approach (Supplementary Material, Fig. S3). However, as with these first two mouse models, all Drc1 W244X/W244X mice died before puberty (Table 3).

Drc1 is necessary for male fertility and sperm flagellum formation in C57BL/6 × ICR mice
Neither patient harboring DRC1 stop-gain mutations identified in this study exhibited respiratory symptoms or evidence of hydrocephaly. While humans exhibit a complex genetic background, inbred C57BL/6 mice present with a simplified background that may influence the penetrance of particular mutations. We noted that the knockout of other PCD-related genes including Pcdp1 (44), Camsap3 (45) and Cep290 (46) in mice was lethal in inbred strains, whereas survival rates rose when the underlying genetic background was altered. We therefore speculated that the observed lethality associated with our Drc1 mutant lines may have been attributable to the underlying genetic background of these animals. We thus hybridized Drc1 mutant mice onto the C57BL/6 × ICR background and found that this significantly ameliorated the post-natal survival of these animals ( Table 3). Mice that survived to adulthood did not exhibit hydrocephaly or decreased size (Fig. 3A). The fertility of individual male mice from these different strains (Drc1 +/+ , Drc1 R554X/R554X and Drc1 W244X/W244X ) was assessed by housing these animals with Drc1 +/+ (WT) females and recording the number of offspring per litter. Both Drc1 R554X/R554X and Drc1 W244X/W244X males failed to sire any offspring despite copulation with females ( Fig. 3B and C). No DRC1 protein expression was detected in the testes of Drc1 R554X/R554X and Drc1 W244X/W244X mice, with no truncated protein being evident therein (Fig. 3D). Levels of Drc1 mRNA were also significantly lower in these tissue samples, consistent with an NMD phenotype associated with these mutations mRNA ( Fig. 3E and F). No differences in gross testis appearance or weight were noted when comparing Drc1 +/+ and Drc1 R554X/R554X littermates ( Fig. 3G and H), and the same was true when comparing Drc1 +/+ and Drc1 W244X/W244X samples ( Fig. 3I and J). Following PAS staining, we found that both Drc1 R554X/R554X and Drc1 W244X/W244X lacked flagella of normal length within the lumen of seminiferous tubules ( Almost all spermatozoa collected from the epididymal cauda in Drc1 R554X/R554X and Drc1 W244X/W244X mice exhibited abnormalities including short, bent, curled, thick or absent flagella (Supplementary Material, Fig. S7). Immunofluorescent staining confirmed DRC1 to be completely absent in these Drc1 R554X/R554X and Drc1 W244X/W244X spermatozoa (Supplementary Material, Fig. S9). Normal sperm acrosome morphology was detected in Drc R554X/R554X and Drc1 W244X/W244X mice, whereas sperm tails were clearly disordered (Supplementary Material, Fig. S8). These serious tail defects coincided with a complete absence of sperm motility for both murine strains (Supplementary Material, Movies S1-S3). Altogether, these findings indicate that DRC1 is essential for sperm flagellum formation and male fertility.

DRC1 is an essential N-DRC assembly mediator in both flagella and cilia
We next conducted a co-IP experiment confirming the ability of DRC1 to interact with DRC2-5 (Fig. 4A). We then evaluated the expression of other DRC proteins in sperm to understand the impact of DRC1 knockout on N-DRC assembly. While β-Tubulin signals were largely unchanged, DRC1-4 could not be detected in Drc1 R554X/R554X mature sperm (Fig. 4B). The radial spoke (RS) component RSPH9 was also not detectable in these cells (Fig. 4B). Immunofluorescent staining similarly indicated that CCDC65 (DRC2), GAS8 (DRC4) and RSPH9 signals were largely absent in Drc1 R554X/R554X spermatozoa ( Fig. 4C-K). The same was true for Drc1 W244X/W244X spermatozoa (Supplementary Material, Fig. S10). Scanning electron microscope (SEM) analyses revealed clear distortion of the flagella of Drc1 R554X/R554X and Drc1 W244X/W244X sperm ( Fig. 5A-H). Transmission electron microscope (TEM) further indicated that Drc1 R554X/R554X and Drc1 W244X/W244X sperm exhibited flagellum appendices such as outer dense fibers separated by microtubules, whereas N-DRC, RSs and dynein arms were not detectable ( Fig. 5I-L).
We additionally assessed respiratory cilia changes in Drc1 R554X/R554X (Fig. 6) and Drc1 W244X/W244X mice (Supplementary Material, Fig. S12), revealing them to be significantly shortened ( Fig. 6A-E, Supplementary Material, Fig. S12). While these respiratory cilia from Drc1 R554X/R554X and Drc1 W244X/W244X mice lacked N-DRC structure formation, they still maintained the 9 + 2 microtubule, RS and dynein arm structural arrangements in contrast to findings in sperm flagella (Fig. 6F-I). In B6 background, respiratory cilia were found no obvious different from mix background by light microscopy (Supplementary Material, Movie S7), while TEM analyses further indicated that the 9 + 2 microtubule was also maintained normally in Drc1 −/− (Supplementary Material, Fig. S11). Immunofluorescent staining confirmed the absence of DRC1, DRC2 and DRC4 in respiratory tract cilia from Drc1 R554X/R554X and Drc1 W244X/W244X mice ( Fig. 7 and Supplementary Material, Fig. S13), whereas RSPH9 signal intensity was unchanged in these cilia, suggesting that the RS structure was unaffected by DRC1 deletion in respiratory tract cilia. Cilia lacking the N-DRC structure exhibited abnormal motility (Supplementary Material, Movies S4-S6). These data indicate that DRC1 is necessary for N-DRC complex assembly in both flagella and cilia.

The loss of DRC1 disrupts flagellar axoneme assembly and causes nuclear deformation during spermiogenesis
While DRC1-deficient sperm exhibited disordered microtubular arrangement and RS formation, this was not observed in cilia. We therefore hypothesized that DRC1 plays a specific role in the context of sperm flagellum assembly or stabilization. To test this possibility, we examined flagellum formation during spermiogenesis (Fig. 8A-T), revealing that axoneme assembly occurred normally in early-stage spermatids from Drc1 R554X/R554X and Drc1 W244X/W244X mice ( Fig. 8G and N), but that with spermatid differentiation, the microtubules within these cells became increasingly disordered such that a normal axoneme structure was no longer present (Fig. 8H-L and O-T). Flagellum lacking DRC1 appeared disordered, whereas the 9 + 2 microtubule structure was still evident in respiratory cilia from these animals, emphasizing that DRC1 plays a key role in regulating the structural stability of sperm flagellar axonemes.
Sperm from patients aIV-1 and bIV-1 exhibited a high frequency of head deformities in addition to the aforementioned tail deformities, and such head deformities were also common in sperm from Drc1 R554X/R554X and Drc1 W244X/W244X mice (Supplementary Material, Fig. S8). This may be attributable to the microtubular structural disorder observed in the cytoplasm of spermatids during spermiogenesis. When we further analyzed dynamic changes in the perinuclear manchette microtubule structure during spermiogenesis, we observed a long and narrow manchette distribution around the nucleus in the sperm of Drc1 R554X/R554X and Drc1 W244X/W244X mice (Supplementary Material, Fig. S14A-O), resulting in the amorphous head phenotypes of these cells. Overall, these data suggest that sperm flagellar microtubule disorder as a consequence of DRC1 deletion can additionally impact dynamic cytoplasmic microtubule changes in spermatids, thereby causing complex head and tail defects.

Discussion
PCD is a multisystem disease that typically arises due to mutations in abnormal genes associated with ciliary movement, with dynein arms, RS, N-DRC and related motility-associated structural defects having been linked to this condition in humans. PCD patients often exhibit a range of genetic defects in genes such as DNAH1 (22), CCDC114 (47) and RSPH4A (48), but phenotypic differences still exist even among individuals that share a given mutation. For example, CEP290 mutations can cause a spectrum of phenotypes that range from retinal degeneration (Leber congenital amaurosis) to embryonic lethality (Meckel-Gruber syndrome) (49)(50)(51)(52)(53)(54)(55), and there are often no clear correlations between patient genotype and phenotype (46). Studies of knockout mice have similarly shown that the phenotypes of mice harboring mutations in different PCD-related genes vary depending on the parental background strain (44)(45)(46)56). In the present report, we detected no respiratory symptoms harboring (acetylated-tubulin, green) and for the RS (RSPH9, red). RSPH9 signal did not differ between WT and Drc1 R554X/R554X samples. DRC1 mutations, but we did identify a previously unreported MMAF phenotype in these patients. To confirm the relationship between DRC1 mutations and these phenotypes, we generated three Drc1 mutant mouse strains. The phenotypic manifestations of this mutation were dependent upon murine genetic background such that Drc1 −/− and Drc1 R554X/R554X mice on the C57BL/6 background exhibited postnatal death. To determine whether this mutation was mutation-specific, we also assessed an additional DRC1 mutation recorded in gnomAD predicted to cause premature stop codon generation. Prepubertal death was also observed for Drc1 W244X/W244X mice on the C57BL/6 background and as such we not able to use these mice to model male infertility. When we instead introduced the ICR background into these Drc1 −/− , Drc1 R554X/R554X and Drc1 W244X/W244X strains, we found that some of these mice survived and exhibited weight and behaviors comparable to those of WT control mice. Importantly, these mice recapitulated the MMAF phenotypes detected in humans with a DRC1 deficiency. These results confirmed the ability of DRC1 mutations to directly cause MMAF while also reaffirming the fact that genetic background is a key determinant of PCD pathogenesis. While we did not explore the genomic basis for this finding, it nonetheless underscores the contribution of individual genetic background to PCD patient phenotypes.
Few prior studies have explored spermiogenesis in the context of MMAF pathogenesis. We found that sperm flagellum damage was more severe than ciliary damage in the present study, as evidenced by axoneme structural disorder in addition to DRC structural loss. During the early stages of spermiogenesis, axonemes appeared similar in Drc1 +/+ , Drc1 R554X/R554X and Drc1 W244X/W244X round spermatids. However, axoneme structural disorder was evident in Drc1 R554X/R554X and Drc1 W244X/W244X elongating spermatids and spermatozoa, and these phenotypic changes cannot be explained by the known functional role of DRC, suggesting that DRC1 to additionally be necessary for flagellar structural stability. We simultaneously detected manchette structural abnormalities in elongating spermatids from mice harboring these DRC1 mutations, further contributing to sperm nuclear transformation and disorder. Similar findings have also been observed in mice exhibiting sperm flagellar dysplasia including Drc7 −/− , Rsph6a −/− and Cfap43 −/− animals (15,57,58). These data suggest that abnormal microtubule organization in sperm flagella is associated with abnormal microtubule organization in the elongating spermatid manchette, potentially explaining the mixed head and tail malformations observed in the context of teratospermia.
In summary, we herein identified a novel DRC1 mutation that causes MMAF and male infertility, and we found that genetic background profoundly influences the phenotypic manifestation of this mutation. These data offer new insights regarding the genetic basis for phenotypic diversity in PCD patients, emphasizing the fact that differences in genetic background can influence the penetrance of different PCD-related mutations. We also found that DRC1 mutations resulted in decreased flagellum axoneme stability, thereby causing flagellar structural disorder while also affecting the manchette of spermatids. These DRC1 mutations also led to distinct phenotypic manifestations in cilia and flagella, providing a foundation for the future study of flagellar and ciliary structural stability (Fig. 9). Overall, our data emphasize the important functions of the N-DRC core component DRC1 in mammals.

Study participants
A cohort of 100 Han Chinese men with MMAF-related male infertility were enrolled from the First Affiliated Hospital of Anhui Medical University and the Affiliated Suzhou Hospital of Nanjing Medical University in China. All enrolled patients exhibited primary infertility, and patients with PCD were excluded. To be eligible for enrollment, patients had to present with a standard MMAF phenotype characterized by severe asthenozoospermia (total sperm motility < 10%; normal: 40%) with >40% of spermatozoa exhibiting the following flagella abnormality: short, absent, coiled, bent or irregular flagella. Karyotypic analyses for all enrolled patients were normal (46, XY), as were hormone levels, bilateral testicular size distributions and secondary sex characteristics. The samples of parental DNA were obtained from 84 of the 100 enrolled MMAF patients. An EX20 kit (AGCU ScienTech Incorporation, Wuxi, China) was used to confirm parental relationships for the enrolled patients. The Institutional review boards of the School of Basic Medical Sciences, Nanjing Medical University, School of Life Sciences at Fudan University, the First Affiliated Hospital of Anhui Medical University and the Affiliated Suzhou Hospital of Nanjing Medical University approved this study. All patients provided informed consent to participate, and the study was performed as per the Declaration of Helsinki.

Whole-exome sequencing
Whole-exome sequencing (WES) was conducted for the 100 enrolled MMAF patients using gDNA isolated from peripheral blood with a DNeasy Blood and Tissue kit (QIAGEN, Duesseldorf, Germany). An Agilent SureSelectXT Human All Exon Kit was utilized to isolate and enrich exonic sequences, after which sequencing was performed on the Illumina HiSeq X-TEN platform. Standard assembly (Burrows-Wheeler Aligner), calling (Genome Analysis Toolkit) and annotation (ANNOVAR) were performed for sequencing analyses as detailed previously (37), and DRC1 mutations identified via this approach were confirmed via Sanger sequencing using primers listed in Supplementary Material, Table S1.

Animals
Mice were housed in a standard animal facility (20-22 • C; 50-70% humidity; 12 h light/dark cycle) with free food and water access. The Institutional Animal Care and Use Committees of Nanjing Medical University approved this study (Approval No. IACUC-1810020), and all experiments were performed as per the Guide for the Care and Use of Laboratory Animals and institutional guidelines.

Fertility test
Fertility tests were conducted for adult mice of each genotype. Males were mated with three WT ICR mice, with vaginal plug inspections being conducted every morning. Dates of birth and number of pups per litter were recorded.

Sperm analysis
Sperm were collected from the cauda epididymis, extruded and suspended in modified HTF Medium (Irvine Scientific, CA, USA) containing 10% FBS. After a 5 min incubation at 37 • C, sperm samples (10 μl) were subjected to computer-assisted semen analysis (Hamilton-Thorne Research, Inc., MA, USA). Remaining sperm samples were fixed for 30 min with and spread onto slides, after which H&E staining was performed as per standard methods to assess sperm morphology.

Histological analysis
Mouse testes, epididymal and tracheal tissues were collected from a minimum of three mice per genotype. Modified Davidson's fluid was used to fix testis and epididymis samples for up to 24 h, whereas 4% PFA was used to fix tracheal samples overnight followed by storage in 70% ethanol. Samples were then dehydrated via ethanol gradient, paraffin-embedded and 5 μm thick tissue sections were mounted on glass slides. H&E staining was conducted as per standard protocols while Periodic Acid-Schiff (PAS) staining was conducted with the PAS staining kit (395B, Sigma-Aldrich, St. Louis, MO, USA).

Murine tracheal epithelial cell isolation
A tracheal brushing approach was used to isolate murine tracheal epithelial cell (mTECs) as in prior studies (60). Briefly, tracheal brushing was conducted to isolate multiciliated airway cells, which were fixed for 30 min with 4% PFA, spread on glass slides and allowed to air-dry.

Western blotting
A lysis buffer (50 mM Tris-HCl pH 8.2, 75 Mm NaCl, 8 M urea) containing a 1× cOmplete™ EDTA-free Protease Inhibitor Cocktail (Roche, Basel, Switzerland) was used to extract proteins, which were then separated via SDS-PAGE and transferred onto PVDF membranes that were blocked for 2 h with 5% non-fat milk in TBS at room temperature, followed by overnight incubation with appropriate primary antibodies at 4 • C. Blots were then washed thrice with TBST, probed for 2 h at room temperature with appropriate secondary antibodies and protein bands were then detected with the high-sig ECL western blotting substrate (Tanon, Shanghai, China).

Immunoprecipitation
Lipofectamine 2000 was used to transfect HEK293T cells with DRC expression plasmids. At 2 days post-transfection, cells were lysed for 40 min using RIPA buffer (P0013C, Beyotime, Shanghai, China) containing 1× cOmplete EDTA-free Protease Inhibitor Cocktail (Roche, Basel, Switzerland) at 4 • C, after which samples were spun down for 20 min at 12000 × g. Supernatants from these lysates were then precleared for 1 h with 10 μl Protein G magnetic beads (10004D, ThermoFisher, Waltham, MA, USA) at 4 • C, after which they were combined with anti-HA-tag antibodies overnight at 4 • C. They were then mixed with 50 μl of Protein G magnetic beads for 3 h at 4 • C. Beads were washed thrice with RIPA buffer, boiled for 10 min in 1× SDS loading buffer and proteins were then subjected to SDS-PAGE analysis.

Immunofluorescence
Immunofluorescent staining of testis and tracheal tissue sections was conducted as detailed previously (61). Spermatozoa and mTECs were obtained as detailed above. Samples were washed thrice with PBS (10 min/wash), and antigen retrieval was conducted for 10 min in 10 mM citrate buffer (pH 6.0) in a microwave oven. Following three additional PBS washes, 5% BSA was used to block slides for 2 h, after which they were stained overnight with appropriate primary antibodies at 4 • C. Following secondary antibody staining for 2 h at room temperature, Hoechst 33342 counterstaining was performed for 5 min. Slides were then rinsed with PBS and mounted using glycerol prior to imaging with an LSM800 confocal microscope (Carl Zeiss AG, Jena, Germany) or a TCS SP8X confocal microscope (Leica Microsystems, Wetzlar, Germany).

Analysis of tracheal ciliary length
Tracheas were excised from three 12-day-old Drc1 +/+ , Drc1 W244X/W244X and Drc1 R554X/R554X mice, and isolated mTECs were stained with anti-acetylated α-tubulin as above. Cells were then imaged via confocal microscope (Leica TCS SP8X), and LAS X was used to measure cilia length by assessing the ciliary tuft length for each cell, with 20 cells being analyzed per animal.

Assessment of tracheal ciliary motility
Murine tracheal tissues were dissected, added to high-glucose DMEM containing 10% FBS (Gibco, Grand Island, NY,USA), opened on the dorsal side and minced under stereoscopic magnification to yield ∼5 mm tissue fragments. These tissues were then transferred to a confocal dish (BDD012035, BIOFIL, Guangzhou, China) and a scotch tape spacer was used to facilitate their imaging under a 40× objective (CFI S Plan Flour ELWD NAMC) with an inverted microscope (Eclipse Ti2-U, Nikon, Tokyo, Japan).

Transmission electron microscopy
For ultrastructural analyses, 2.5% glutaraldehyde was used to fix tracheal and epididymal tissue samples overnight followed by 2% OsO4 post-fixing and embedding in Araldite. Ultrathin 80 nm sections were then stained using uranyl acetate and lead citrate, followed by analysis with an electron microscope (JEM.1010, JEOL, Tokyo, Japan).

Scanning electron microscopy
Spermatozoa and tracheal samples were fixed for 2 h with 2.5% phosphate-buffered glutaraldehyde at 4 • C. Spermatazoa were then allowed to attach to coverslips coated with poly-L-lysine. Both sample types were then washed with PBS, dehydrated with a chilled ethanol gradient (30,50,70,80, 90 and 100%) and subjected to critical point drying with a Lecia EM CPD300 Critical Point Dryer (Wetzlar, Germany). Samples were then attached to appropriate specimen holders and coated with gold particles via the use of an ion sputter coater (EM ACE200, Leica, Wetzlar, Germany). A Helios G4 CX scanning electron microscopy (SEM) (Thermo Scientific, Waltham, MA, USA) was then used to image samples.

Statistical analysis
Experiments were conducted in triplicate. Data are given as means ± standard error and were compared by one-way ANOVAs and unpaired two-tailed t-tests. P < 0.05 was the significance threshold. Microsoft Excel and GraphPad Prism 6.0 were utilized for all statistical testing.

Supplementary Material
Supplementary Material is available at HMG online.