A Homozygous Ancestral SVA-Insertion-Mediated Deletion in WDR66 Induces Multiple Morphological Abnormalities of the Sperm Flagellum and Male Infertility

Study of Sperm Morphology, Spermograms, and Transmission electron Microscopy

Sperm morphology was first analyzed on slides under a microscope after modified Papanicolaou staining. Sperm analysis was carried out in the source laboratories during the course of the routine biological examination of the subjects according to World Health Organization (WHO) guidelines.²⁰ Then, transmission electron microscopy (TEM) assessment was carried out. Sperm cells from an affected individual with a WDR66 deletion were fixed in PBS with 4% paraformaldehyde (PFA) for 60 min. They were pelleted (1,000 × g for 10 min) and resuspended in fixation buffer (2.5% glutaraldehyde, 2% PFA, 100 mM phosphate buffer [pH 7.4], and 50 mM sucrose) for 2 hr. Fixed cells were washed in buffer for 10 min, post-fixed in 1% OsO₄ for 1 hr, and washed three times in water, and then samples were stained in 2% uranyl acetate in water at 4°C overnight. Samples were next washed in water, dehydrated in ethanol, embedded in Epoxy resin, and polymerized for 60 hr at 60°C. Sections were stained with 2% uranyl and lead citrate and visualized in a transmission electron microscope (JEOL 1200EX). Trypanosomes were prepared for TEM as above with the exception that they were embedded in Spur resin.

WES and Data Processing

Genomic DNA was isolated from saliva with the Oragen DNA Extraction Kit (DNA Genotech). Coding regions and intron-exon boundaries were enriched with the All Exon V5 Kit (Agilent Technologies). WES and initial data analyses were performed as described previously.⁶ CNVs, which had not been analyzed previously, were analyzed with the ExomeDepth software package.¹⁸ We retained only overlapping CNVs found in at least two subjects and not described as benign in Cathagenia. All data are based on GRCh37 (UCSC Genome Browser hg19).

RT-PCR and qRT-PCR

qRT-PCR was performed on a cDNA panel from testis, lung, kidney, and heart tissues (Clontech) with a StepOnePlus Real-Time PCR System (Life Technologies) using Power SYBR Green PCR Master Mix (Life Technologies) according to the manufacturer’s protocol. Quantification of the fold change in gene expression was determined by the relative quantification (RQ) method (2^−ΔΔCT) with GAPDH as a reference. Data are shown as the mean RQ ± standard deviation. Primers are described in Table S1.

Immunofluorescence and Confocal Microscopy on Human Sperm Cells

Immunostaining was carried out on samples from individuals P7 and P9. Sperm was fixed in PBS and 4% PFA for 7 min. After being washed in PBS, the sperm suspension was spotted onto slides pre-coated with 0.1% poly-l-lysine (Thermo Fisher Scientific). After attachment, sperm was permeabilized with 0.1% (v/v) Triton X-100-DPBS (Triton X-100, Sigma-Aldrich) for 15 min at room temperature (RT). Slides were then blocked in 5% normal serum DPBS (normal goat or donkey serum; GIBCO, Invitrogen) and incubated overnight at 4°C with primary antibodies. Polyclonal rabbit anti-WDR66 antibody was purchased from Sigma-Aldrich and diluted at 1:50. Monoclonal mouse anti-acetylated α-tubulin antibody was purchased from Sigma-Aldrich and diluted at 1:2,500. The slides were then washed and incubated with secondary antibody diluted at 1:800 (Dylight-549-conjugated goat anti-mouse immunoglobulin G [IgG] and Dylight-488-conjugated goat anti-rabbit IgG, Jackson ImmunoResearch) and Hoechst 33342 (Sigma-Aldrich) for 1 hr at RT, rinsed, and mounted with Dako Fluorescence Mounting Medium (Life Technology). Images were acquired on a Zeiss LSM 710 confocal microscope and processed with Zen 2009 software. The list of antibodies used in this study is provided in Table S2.

Breakpoint Investigation

We conducted a walking PCR approach to get closer to the deleted fragment. We performed long-range PCR to amplify the CNV breakpoint junctions and analyzed the product by Sanger sequencing to refine the breakpoints. To investigate the potential mechanism causing the deletion, we aligned the sequences around the breakpoints with each other by using the Basic Local Alignment Search Tool (BLAST) and then analyzed them for different repetitive elements by using RepeatMasker software. PCR primers and protocols used for each individual are listed in Table S3. Sequencing reactions were carried out with BigDye Terminator v.3.1, sequence analysis was carried out on ABI 3130XL, and sequences were analyzed with SeqScape software (all Applied Biosystems).

T. brucei Cell Lines, Culture, Transfection, and Western Blotting

The trypanosome cell lines used in this study were derived from the parental (wild-type [WT]) procyclic form (PCF) T. brucei 427 29-13 (present in the insect host) and bloodstream form (BSF) T. brucei 427 90-13 strains (present in the mammalian host), both co-expressing the T7 RNA polymerase and tetracycline repressor.²¹ Cells were cultured, transfected, and cloned as previously described^{22, 23} in medium supplemented with hygromycin (25 μg/mL for PCF and 5 μg/mL for BSF), neomycin (10 μg/mL for PCF and 2.5 μg/mL for BSF), and blasticidin (20 μg/mL for PCF and BSF) for constitutive translation of 10TY1-tagged proteins and with phleomycin (5 μg/mL for PCF and 2.5 μg/mL for BSF) for RNAi-inducible cells. RNAi was induced with tetracycline (10 μg/mL).

To produce transgenes at their endogenous loci encoding proteins tagged at the C terminus with a 10TY1 tag (TbWDR66_TY1 or TbWDR66-ΔCter_TY1), we transfected WT cells with tagging cassettes obtained by PCR by using (1) a forward primer containing 80 nucleotides from the 3′ end of the TbWDR66 or TbWDR66-ΔCter open reading frame (ORF) and a sequence specific to the pPOTv7-10TY1-neo vector²⁴ to amplify the TY1 tag and (2) a reverse primer containing 80 nucleotides from the 3′ UTR of TbWDR66 and a sequence specific to the vector to amplify a selection gene.²⁴ For RNAi, a fragment of the TbWDR66 ORF (base pairs 2,719–3,070, corresponding to amino acids 903–1,023 in the C-terminal domain) was cloned into p2T7tiB²⁵ and transfected into the TbWDR66_TY1 and TbWDR66-ΔCter_TY1 background cell lines. The primer sequences are presented in Table S4, and a schematic of the constructs is provided in Figure S1. RNAi was induced by tetracycline (10 μg/mL). For western blotting, samples were prepared and treated as previously described.⁶

Immunofluorescence of T. brucei Cells

Cells were collected, washed, and processed for immunolabeling as previously described.³ In brief, cells were pelleted and washed, loaded on poly-l-lysine-coated slides, and air dried. After rehydration, cells were fixed and permeabilized in methanol at −20°C, rehydrated in PBS, and then incubated with primary antibodies in two steps: (1) anti-TY1 (BB2) followed by Alexa-594-conjugated anti-mouse IgG1 (Thermo Fisher Scientific, A21125) and (2) anti-PFR2 (rabbit polyclonal) and anti-tubulin TAT1 mouse IgG2a primary antibodies followed by anti-rabbit ATTO-N647-conjugated (Sigma 40839) and anti-mouse IgG2a Alexa-conjugated (Thermo Fisher Scientific A21131) secondary antibodies. Nuclei and kinetoplasts were stained with DAPI (10 μg/mL) in epifluorescence experiments. Epifluorescence images were acquired on a Zeiss Imager Z1 microscope with a Photometrics Coolsnap HQ2 camera equipped with a Zeiss 100× objective (numerical aperture [NA] 1.4) and Metamorph software (Molecular Devices) and were analyzed with ImageJ. Stimulated emission depletion microscopy (STED) images were acquired on a Leica DMI6000 TCS SP8 X-STED microscope with an HC PL APO CS2 100×/1.40 oil objective and de-convolved with Huygens Pro 16.10. All antibodies used are listed in Table S5.

Video Microscopy of T. brucei Cells

Video microscopy was carried out as previously described.²⁶ In brief, BSF cells (WT cells and cells induced with RNAi for 48 hr) were washed, and cell mobility was recorded by phase contrast on a Zeiss AxioImager with a 40× objective (NA 1.3). Fifty seconds of digital video from separate regions was captured and analyzed with Metamorphoh software (Molecular Devices) and video edited with ImageJ (setting frame rate was 10 fps).

WDR66 Description, Transcript Expression, and Protein Localization

WDR66, also known as CFAP251, is located approximately 11 Mb from the telomere of 12q (chr12: 122,355,768–122,441,833 [Ensembl: ENSG00000158023]) and contains 22 exons. Transcription of WDR66 leads to two isoforms: T1 (GenBank: NM_144668), which encodes 1,149 amino acids (UniProt: Q8TBY9-1), and T2 (GenBank: NM_001178003), which encodes 941 amino acids (UniProt: Q8TBY9-2). Because the deleted exons 20 and 21 are present only in T1, it is important to measure the expression of both transcripts in tissues of interest to confirm that T1 is expressed in the testis (Figure 2B). We therefore measured the expression levels of the transcripts by performing qRT-PCR in different human tissues by using specific primers for each isoform. Results showed that T1 was highly and predominantly expressed in the testis and moderately expressed in other ciliated tissues (especially the lung), whereas T2 had negligible expression in all tested tissues (Figure 2C). These results indicate that T1, the transcript affected by the identified deletion, is the main physiological transcript and is highly expressed in the testis (Figure 2C). In this article, we will therefore refer to this transcript (Ensembl: ENST00000288912.8) and to the protein it encodes. The consequence of the deletion on the transcript is c.3007_3337del.

Using the InterPro domain annotation program, we found that nine WD40-repeat domains are present in both isoforms T1 and T2, and one calcium-binding EF-hand domain is present only in isoform T1 in the C terminus (Figure 2D and Table S7), indicating that the activity of this isoform could be modulated by calcium. In the protein encoded by T1 (GenBank: NP_653269), the first amino acid affected by the identified deletion is predicted to be amino acid 1,003. Deletion of exons 20 and 21 induces a frameshift, and translation of exon 22 is predicted to be incorrect for the first 25 amino acids before a stop codon is induced 26 nucleotides after the first abnormally coded amino acid: p.Ile1003Lysfs26. Data on genetic variants described here are available under accession number ClinVar: SUB3375641.

We then wanted to determine whether the deleted transcript is present in sperm from individuals carrying the genomic deletion or subjected to mRNA decay. Unfortunately, we could not obtain enough biological material to perform a reliable RT-PCR. However, mRNA decay has been shown to be elicited by abnormally long 3′ UTRs and to be more efficient when a premature stop codon is introduced near the 5′ end of the transcript.³⁰ The deletion identified here removes the two penultimate exons and creates a new stop codon in the last exon near the transcript’s 3′ end. We therefore do not expect this mechanism to be active here. Out of the four commercially available antibodies, we observed that only one provided a specific staining (Table S2). This antibody has been raised against a 101 amino acid epitope overlapping with the subject’s deletion. We observed that, although the staining was weaker than in control sperm, the subject’s short and abnormal flagella were immunodecorated along the flagella, suggesting that the truncated protein is present in the affected subjects and localized to the flagella (Figure 2E). The intensity of the signal in sperm from affected individuals is, however, very weak, and in the absence of a negative control we cannot exclude the possibility that it could be due to non-specific staining and thus that the protein could be completely absent in subjects with this deletion.

Breakpoint Characteristics and CNV Mutational Mechanisms

Then, to unequivocally confirm the exact size and position of the deletion, we performed a walking PCR approach. We validated a PCR protocol that specifically amplified a 950 bp fragment in affected individuals by using the primers F1 (located 5,209 bp upstream of exon 20) and R1 (positioned 1,597 bp downstream of exon 21) (Figures 3A and 3B). As could be expected, the 8,541 bp interval in control individuals could not be amplified and yielded no signal, whereas a shorter band (950 bp) was obtained from all affected individuals (Figure 3C). Sanger analysis of the PCR sequence covering the breakpoint region in the affected individuals allowed the characterization of the breakpoint at single-nucleotide resolution (Figure 3D). We observed that the deletion spanned 7,591 bp and exons 20 and 21 of WDR66 were completely deleted (chr12: 122,432,883–122,440,629). All affected individuals carried the same deletion with the same nucleic acid sequence at the breakpoint locus, suggesting that they had all inherited the same deleted allele from a common ancestor. The loss of exons 20 and 21 induces a frameshift leading to a premature stop codon at the beginning of the last exon (exon 22). Therefore, the 147 amino acids encompassing the C-terminal region of the protein (12.8% of the full-length protein) are predicted to be lost.

Open in a separate window

Figure 3

Fine Characterization of the Deletion and Theoretical Description of the Initial Deletion Event

(A) Representation in control subjects of part of WDR66 (between exons 19 and 22) and localization of the deletion-specific primers F1 and R1.

(B) Representation of the same region in affected subjects. Detailed analysis of the deleted region within WDR66 highlights the presence of a type F SINE-VNTR-Alu (SVA) retrotransposon (SVA-F) positioned at the centromeric border of the identified WDR66 deletion (orange box).

(C) PCR of the deleted region. In control subjects (C1–C3), F1 and F2 primers are too distant (8,541 bp) and do not yield any amplification. PCR amplification with F1 and F2 in affected subjects (P1–P7) permits the amplification of a 950 nucleotide fragment.

(D) Sanger sequencing through the deletion breakpoints permits localization of the exact position of the breakpoints, indicating that a total of 7,591 nucleotides were deleted. It also highlights the presence of a 165 nucleotide sequence (yellow) inserted at the junction point.

(E) Blast analysis of the 165 nucleotide inserted sequence shows that an identical sequence (100% identity) is present within an SVA-D sequence located 400 kb telomeric to SVA-F between exons 5 and 6 of CLIP1 (yellow box).

(F) SVA-D was most likely retrotransposed between WDR66 exons 21 and 22.

(G) An intra-chromatidien non-allelic homologous recombination (NAHR) was then initiated.

(H) The formation of a hybrid SVA-F/SVA-D sequence preceding the WDR66 exon 20 and 21 deletions.

To estimate the allelic frequency of this variant and to confirm that it is not a frequent CNV in the North African population, we carried out a multiplex PCR allowing the co-amplification of the deleted allele (950 bp of product size) with the normal allele (220 bp), thus including a positive control in all reactions. A total of 100 North African individuals were analyzed (Figure S3). None of the tested control individuals amplified the deletion, indicating that the variant is also rare (allelic frequency < 1/200) within the North African population. We then re-analyzed the whole data from WES analysis (SNPs and indels) of the 78 individuals, but we did not identify any other WDR66 deleterious variants.

To investigate the origin of this complex rearrangement, we carried out a series of bioinformatic analyses focusing on the genomic region surrounding the deletion. Clustering the deleted sequence by using RepeatMasker software indicated that the deleted region is rich in transposable elements and prone to genomic rearrangements. In addition, the centromeric breakpoint of the identified deletion was localized at the exact ending of an SVA_F retrotransposable element (Figures 3A, 3B, and 3E). Intriguingly, detailed sequence analysis revealed that in the affected subjects, the telomeric fragment of this element (122,432,706–122,432,883) did not perfectly match the expected SVA-F sequence (Figure S4). Alignment of this sequence with BLAT showed 100% similarity to a distal sequence of an SVA-D element (Figure S4) located in intron 5 of CLIP1, a gene located 400 kb telomeric to WDR66 (Figure 3E). In control individuals, the abnormal telomeric end of the WDR66 SVA_F sequence perfectly matched the reference sequence (GRCh37, UCSC Genome Browser hg19), indicating that the short SVA_D sequence observed only in affected subjects was specific to the deletion event and most likely came from the CLIP1 SVA_D retrotransposable element. We conclude that SVA-D was most likely retrotransposed into WDR66 intron 21 (Figure 3F) and initiated non-allelic homologous recombination (NAHR) with the SVA-F element in intron 19 (Figure 3G) to induce the formation of a single chimeric SVA-F/SVA-D sequence, which preceded the WDR66 deletion (Figure 3H). This indicates that WDR66-SVA-F recombined at the following localization: 122,432,691–122,432,706. Such a mechanism has already been observed and has been described as an SIMD.¹⁹

Dating of the Identified Mutation

To compute the age of the deletion, we use the size of the shared ancestral haplotypes surrounding the WDR66 deletion locus (Table S6); if we assume a recombination rate of 1.30 cM/Mb,³¹ the size is equal to 0.26 cM. The probability of observing an ancestral haplotype of size c (measured in morgans) or larger around the WDR66 deletion is equal to (1 − c)^g, where g is the deletion’s age measured in generations.³² The probability that the 2n haplotypes carried by the seven individuals is larger than c is (1 − c)2^ng, which can be approximated by e^−2ngc. Therefore, the length of the shared ancestral haplotype can be approximated by an exponential distribution of rate 2ng. The parameter g can be estimated by 1/2nL, which provides an estimate of g = 27 generations. However, Labuda et al. have shown that dating arising from genetic clock methods is biased downward because these methods do not account for population expansion.³³ To account for an exponential expansion of rate r per generation, the mathematical expression (1/r)ln(c.e^r/(e^r − 1)) should be added to the estimates provided by genetic clock methods. If we assume that an upper bound for the expansion rate is given by the expansion r = 0.4, as was calculated for the Ashkenazi Jewish populations,³⁴ the age of the mutation is in the range of 27–43 generations. Given 25 years for the generation time, the mutation would have then arisen 675–1,075 years ago. Heterozygous individuals have apparently not been negatively selected, and the deleted allele now seems relatively frequent in Tunisia given that 7 of the 46 (15%) Tunisian MMAF subjects analyzed here had a homozygous WDR66 deletion.

Phenotypic Analysis of MMAF Individuals Carrying a Genomic WDR66 Deletion

The analysis of semen samples from the seven individuals carrying the same mutation showed normal semen volume and concentration, although with large inter-individual variations (Table 1). A total absence of sperm motility was observed for all individuals carrying the WDR66 deletion (Table 1). Sperm morphology appeared seriously altered under light microscopy, wherein 100% of spermatozoa were abnormal, as illustrated in Figure 1. The most common morphological defects observed in the semen of all individuals concerned the flagellum dimensions (reduced length and irregular caliber). A high rate (56%–95%) of acrosomal abnormalities was also observed in many affected subjects. We compared the sperm parameters of individuals with different genotypes (mutations in DNAH1, CFAP43, CFAP44, and WDR66) but observed no significant differences among the groups (Table S8).

Because an insufficient number of sperm cells were collected from the affected individuals, only sperm cells from one individual carrying the WDR66 deletion could be analyzed by TEM (Figure 1). Longitudinal sections showed severe axonemal and peri-axonemal defects affecting the outer dense fibers (ODFs), the fibrous sheath (FS), and the mitochondrial sheath (MS), which appeared completely disorganized, resulting in aborted flagella or their replacement by a cytoplasmic mass englobing unassembled axonemal components (data not shown). Because oligo-teratozoospermia was observed in this individual, fewer than ten cross-sections were of a sufficient quality for analysis. Among these sections, 100% were abnormal, and the main defect constantly observed concerned the axonemal and the peri-axonemal structures, such as the unassembled FS (Figure 1F).

Efficiency of WES to Detect Pathological CNVs and Importance of SVA Retrotransposons Elements

CNVs are a common cause of disease but until now have not been explored in WES data in the field of male infertility, particularly the MMAF phenotype. In the current study, we analyzed WES data by using ExomeDepth software, which is based on the assessment of sequencing read depth. We analyzed the breakpoints to elucidate the mechanisms underlying the deletion. Results showed a high density of Alu retrotransposons in the region flanking the deletion breakpoints, suggesting that this region exhibits a high degree of genomic instability. Interestingly, we observed a single chimeric SVA element located precisely adjacent to the centromeric breakpoint, strongly suggesting that the creation of this abnormal SVA and the deletion event are the consequence of the same NAHR between SVA_F in intron 19 of WDR66 and the SVA_D retrotransposed from intron 5 of CLIP1 to intron 21 of WDR66. This event resulted in the deletion of a portion of each SVA element and of the intervening DNA sequences (Figure 3). This mechanism was previously described as SIMD.¹⁹ SVAs, originally named SINE-R, are composite transposons that make up 0.2% of the human genome and represent the youngest and currently the most active retrotransposable elements of the human genome.⁴² SVAs are mobilized by L1 enzymatic machinery throughout the genome by a “copy-and-paste” mechanism.⁴³ It has been shown that retrotransposition of SVA elements is a source of insertions, deletions, and rearrangements within the target site and leads to various pathologies.⁴⁴ The spectrum of diseases that have been reported in relation to SVA transposition, including Marfan syndrome, hemophilia B, neurofibromatosis types 1 and 2, leukemia, and congenital muscular dystrophy, has been increasing progressively.^{45, 46, 47, 48, 49, 50} In this study, we report a homozygous CNV that is generated by an SVA transposition event and causes MMAF, leading to male infertility.

WDR66 Function and the Effect of the Identified Deletion

The axoneme, the internal cytoskeleton of motile cilia and flagella, is a highly evolutionary conserved structure that consists of nine DMTs circularly arranged around the central-pair complex of microtubules (9 + 2 structure).⁵¹ The beating of cilia and flagella is orchestrated by multiprotein-ATPase complexes that are located on the peripheral doublets, which provide the sliding force for sperm motility. In addition, the sperm flagellum harbors specific peri-axonemal structures that are not found in other motile cilia: a helical MS in the midpiece, the FS in the principal piece (PP), and ODFs in the midpiece and the proximal part of the PP.⁵²

Previous reports have shown that the WDR66 ortholog (FAP251) in Chlamydomonas localizes to the CSC (calmoduline- and spoke-associated complex) at the base of RS3.⁵³ The CSC was recently identified as an axonemal complex that associates the bases of RS2 and RS23 and the N-DRC (nexin-dynein regulatory complex) and is involved in the modulation of dynein activity and therefore in the regulation of flagellar motility.⁵³ Recent work in Tetrahymena has shown that, in the absence of CFAP251, RS3 was either missing or incomplete, and the protozoa had an altered swimming ability with an abnormal ciliary waveform.⁴⁰ Here, we studied the localization of the WDR66 ortholog in T. brucei and used this model to confirm the deleterious effect of the identified deletion. Immunofluorescence and STED microscopy confirmed that TbWDR66 is an axonemal protein and that in T. brucei, knockdown by RNAi affects flagellar structure and motility. We then generated cell lines with a deletion of the TbWDR66 C-terminal domain, corresponding to the deletion identified in MMAF individuals. We observed that the absence of the C-terminal domain did not affect protein localization (Figure 4A and Figure S6) but severely affected cell mobility, as was observed for the sperm of affected individuals. This work confirms that the WDR66 C-terminal domain, which contains a calcium-regulating EF-hand domain, is critical for maintaining flagellar function probably by regulating flagellar beating; we also observed that WDR66 plays a critical structural role, perhaps by consolidating the links between RS3 and the microtubule doublets.

We thank the affected and control individuals for their participation. This work was mainly supported by the Agence Nationale de la Recherche (ANR; programme DS0401-2014, project ANR-14-CE15-0002) through the “MAS-Flagella” project, the Direction Générale de l’Offre de Soins through the 2014 Programme de Recherche Translationnelle en Santé (to P.F.R., M.B., A.T., C.A., and N.T.M.), the Fondation Maladies Rares through “Whole genome sequencing of individuals with Flagellar Growth Defects (FGD)” of the 2012 Séquençage à Haut Débit program (to P.F.R.), and ANR LabEx through the ParaFrap grant (ANR-11-LABX-0024 to D.R.R.; http://www.labex-parafrap.fr/fr). Stimulated emission depletion microscopy was performed at the Bordeaux Imaging Centre, a service unit of the CNRS-INSERM and Bordeaux University and a member of France-BioImaging, supported by the ANR (ANR-10-INBS-04). The help of Patrice Mascalchi is acknowledged. Some of the electron microscopy was performed at the Bordeaux Imaging Centre. We thank Nicolas Biteau (Bordeaux University) for the rabbit anti-PFR2 antibody, Sam Dean (Oxford University) for the pPOTv7 plasmid, Keith Gull (Oxford University) for the anti-tubulin (TAT1) antibody, and Philippe Bastin (Institut Pasteur, Paris) for the anti-TY1 antibody. We also thank the Grenoble Institut des Neurosciences electron microscopy platform and Anne Bertrand as well as the Institut pour l’Avancée des Biosciences microscopy platform, Alexei Grichine, and Jacques Mazzega for their technical help.