Abstract
A recent study of 17 men with decapitated spermatozoa found that 8 carried two rare SUN5 alleles, and concluded that loss of SUN5 function causes the acephalic spermatozoa syndrome. Consistent with this, the SUN5 protein localises to the head-tail junction in normal spermatozoa, and SUN proteins are known to form links between the cytoskeleton and the nucleus. However, six of the ten SUN5 variants reported were missense with an unknown effect on function, and only one man carried two high confidence loss-of-function (LOF) alleles: p.Ser284* homozygozity. One potential exonic splice mutation, homozygous variant p.Gly114Arg, was not tested experimentally. Thus, definitive proof that loss of SUN5 function causes the acephalic spermatozoa syndrome is still lacking. Based on these findings, we determined the sequence of the SUN5 gene in three related men of North African origin with decapitated spermatozoa. We found all three men to be homozygous for a deletion-insertion variant (GRCh38 - chr20:32995761_32990672delinsTGGT) that removes 5090 base pairs including exon 8 of SUN5, predicting the frameshift, p.(Leu143Serfs*30), and the inactivation of SUN5. We therefore present the second case where the acephalic spermatozoa syndrome is associated with two LOF alleles of SUN5. We also show that the p.Gly114Arg variant has a strong inhibitory effect on splicing in HeLa cells, evidence that homozygozity for p.Gly114Arg causes acephalic spermatozoa syndrome through loss of SUN5 function. Our results, together with those of the previous study, show that SUN5 is required for the formation of the sperm head-tail junction and male fertility.
Introduction
It has been estimated that 14% of couples fail to conceive a child naturally after one year, and compromised male fertility is identified in 60% of these couples (1). Assisted reproductive technology (ART) provides a therapeutic solution for half of these couples, and today accounts for 1–3% of live births in developed countries (2). Despite this, the genetic causes of male infertility and their pathophysiological consequences for gamete quality are known in only a few cases. The identification of genetic causes will be essential to the development of personalised treatments, the extension of treatment to more infertile couples, and informed risk assessment for individuals conceived by ART and their descendants. Presently, the most solid progress has been made in uncovering the genetic basis of teratozoospermia.
Monomorphic teratozoospermia is a group of rare conditions in which almost all the spermatozoa share a specific malformation that renders them non-functional. Over the last decade, homozygous loss-of-function (LOF) mutations have been identified in four genes for three teratozoospermia phenotypes: AURKC (MIM 603495) in large-headed multiflagellar polyploid spermatozoa (MIM 243060) (3), DPY19L2 (MIM 613893) (4,5) in globozoospermia (MIM 613958), SPATA16 (MIM 609856) (6) in globozoospermia (MIM 102530), and DNAH1 (MIM 603332) in MMAF (multiple morphological abnormalities of the flagellum) (7).
In a recent study of 17 Chinese men with acephalic spermatozoa syndrome (MIM 617187), rare coding variants in the spermatid-specific SUN5 gene (MIM 613942) were found to be homozygous or compound heterozygous in eight cases (8). In this syndrome, the sperm head frequently separates from the flagellum, because the sperm head-tail junction is absent (9,10). SUN5 is an attractive candidate for a gene involved in the attachment of the flagellum to the sperm head because the protein localises to the vicinity of the head-tail junction in spermatozoa (8,11), and SUN family proteins are known to be part of links between the inner nuclear membrane and the cytoskeleton (12).
Six of the ten SUN5 variants reported are missense with an unknown effect on the protein. Although one variant, p.Gly114Arg (c.340G > A), was predicted to affect splicing, this was not confirmed experimentally. The other four variants were certainly LOF variants: intronic splice site (c.425 + 1G > A), frameshift (p.Val128Serfs*7) and two nonsense mutations (p.Trp72* and p.Ser284*). However, only one man was homozygous for a certain LOF variant, p.Ser284*, and the variants identified in SUN5 were designated as causal, based on the high frequency of men carrying two rare SUN5 alleles in the study group. In three cases with a missense variant, the SUN5 protein was not detected at the head-tail junction, although it was present in spermatozoa from two control patients with acephalic spermatozoa syndrome but without a variant in SUN5 (8). This is consistent with these missense variants being LOF, but the possibility that SUN5 absence is secondary to junction loss cannot be completely excluded, because the pathological mechanism in the acephalic controls is unknown, and could differ from that in the SUN5 variant cases. Furthermore, six of the SUN5 variants are present at a low frequency in the gnomAD database (13), and four predominate in, or are exclusive to, the East Asian group, including the most frequent, the p.Val128Serfs*7 LOF variant, which is present in 1:140 individuals. Indeed, overall in gnomAD, approximately 1:100 individuals from the East Asian group are heterozygous for one of the rare variants designated as causal of acephalic spermatozoa syndrome. If all these variants were LOF, this sperm phenotype would be expected to have a high incidence among infertile men in East Asia, but this has never been reported, raising a doubt about the pathogenicity of some SUN5 missense variants.
Thus, although the case that the loss of SUN5 function causes a failure of the sperm head-tail junction is persuasive, definitive proof requires a second case where two high confidence LOF SUN5 alleles are found to be associated with acephalic spermatozoa syndrome. We therefore searched for pathogenic variants in the SUN5 gene in three men with decapitated spermatozoa.
Results
Homozygous deletion of exon 8 of SUN5
Two brothers (13-1662 and 13-2016) and their cousin (13-4335) were diagnosed with acephalic spermatozoa syndrome following consultation for primary couple infertility at our clinic. They are members of a consanguineous family from Algeria (Fig. 1). We previously described the phenotype of the two brothers (14), and clinical details for all three men are summarized in Table 1. Based on the recent report of the association of SUN5 variants with acephalic spermatozoa (8), we sequenced the coding region of SUN5 in two of the affected individuals (brother 13-1662 and cousin 13-4335) to determine if mutation in SUN5 was responsible for their infertility. We found no single nucleotide variants, but we failed to amplify exon 8. An additional marker tested in the middle of intron 8 was also negative in the three men, confirming the existence of a homozygous deletion in the three affected men (Fig. 2A). To further define the deletion, we amplified from upstream of exon 7 to downstream of exon 9 (7 kb) and upstream of exon 8 to downstream of exon 9 (6.2 kb), and obtained shorter than expected bands of, respectively, 2 kb (not shown) and 1.2 kb (Fig. 2A). The 1.2 kb fragment was sequenced across the deletion junction, revealing the deleted segment to cover 5090 bp that include the entirety of exon 8 (Fig. 2 and Supplementary Material, Fig. S1). There is an insertion of four bases, TGGT, at the junction of the breakpoint. There are no blocks of homology at the breakpoints, showing that the deletion was not mobilised by unequal crossing-over and is unlikely to be a recurrent event. The deletion of exon 8 is predicted to inactivate SUN5, as it will induce a frameshift, p.(Leu143Serfs*30), making the transcript lacking exon 8 a substrate for NMD (nonsense mediated decay). Using a PCR test that amplifies the deletion allele specifically, we did not detect the SUN5 deletion in 150 infertile men without the acephalic sperm phenotype recruited from our clinic, showing it to be a rare variant.
Family pedigree of the three infertile men presenting with acephalic spermatozoa syndrome (MIM: 617187). Infertile men: 1) 13-1662; 2) 13-2016; 3) 13-4335.
Homozygous deletion of exon 8 from SUN5 in two brothers and a cousin with acephalic sperm syndrome. (A) Mapping of the SUN5 deletion using locus-specific PCR. Results are shown for the three affected men, a fertile man and H2O, following electrophoresis on ethidium bromide-stained agarose gels. Sample order for all gels is as specified on the gel at the left. Affected men: 1) - 13-1662, 2) - 13-2016, 3) - 13-4335. Primers are shown for each test (oNNNN) and their approximate position in SUN5 is indicated by a white or black shape. (B) Schematic representation of the SUN5 deletion and the sequence of the deletion junction. Deletion coordinates are based on the NCBI assembly GRCh38. The corresponding sequence electropherogram is available in Supplementary Material, Fig. S1.
Sperm parameters of the three related infertile men carrying a homozygous deletion within the SUN5 gene
| Patients | 13-1662 | 13-2016 | 13-4335 | |||
|---|---|---|---|---|---|---|
| Age | 36 | 25 | 34 | |||
| Spermogram | A | B | A | B | A | B |
| Sperm concentration (106/ml) | 4.5 | 0.5 | 1.7 | 1.0 | 2.3 | 2.3 |
| Motility a/b/ca(%) | 0/20/80 | 0/10/90 | 0/15/85 | 0/30/70 | 5/5/90 | 5/20/75 |
| Acephalic sperm (sperm flagella) (106/ml) | 45 | 83 | 7.9 | 16.8 | 41 | 64 |
| Normal formsb | 0 | 2 | 0 | 0 | 0 | 0 |
| Vitality (%) | 40 | 40 | 53 | 53 | 72 | 40 |
| Patients | 13-1662 | 13-2016 | 13-4335 | |||
|---|---|---|---|---|---|---|
| Age | 36 | 25 | 34 | |||
| Spermogram | A | B | A | B | A | B |
| Sperm concentration (106/ml) | 4.5 | 0.5 | 1.7 | 1.0 | 2.3 | 2.3 |
| Motility a/b/ca(%) | 0/20/80 | 0/10/90 | 0/15/85 | 0/30/70 | 5/5/90 | 5/20/75 |
| Acephalic sperm (sperm flagella) (106/ml) | 45 | 83 | 7.9 | 16.8 | 41 | 64 |
| Normal formsb | 0 | 2 | 0 | 0 | 0 | 0 |
| Vitality (%) | 40 | 40 | 53 | 53 | 72 | 40 |
WHO criteria (2010).
David criteria.
Sperm parameters of the three related infertile men carrying a homozygous deletion within the SUN5 gene
| Patients | 13-1662 | 13-2016 | 13-4335 | |||
|---|---|---|---|---|---|---|
| Age | 36 | 25 | 34 | |||
| Spermogram | A | B | A | B | A | B |
| Sperm concentration (106/ml) | 4.5 | 0.5 | 1.7 | 1.0 | 2.3 | 2.3 |
| Motility a/b/ca(%) | 0/20/80 | 0/10/90 | 0/15/85 | 0/30/70 | 5/5/90 | 5/20/75 |
| Acephalic sperm (sperm flagella) (106/ml) | 45 | 83 | 7.9 | 16.8 | 41 | 64 |
| Normal formsb | 0 | 2 | 0 | 0 | 0 | 0 |
| Vitality (%) | 40 | 40 | 53 | 53 | 72 | 40 |
| Patients | 13-1662 | 13-2016 | 13-4335 | |||
|---|---|---|---|---|---|---|
| Age | 36 | 25 | 34 | |||
| Spermogram | A | B | A | B | A | B |
| Sperm concentration (106/ml) | 4.5 | 0.5 | 1.7 | 1.0 | 2.3 | 2.3 |
| Motility a/b/ca(%) | 0/20/80 | 0/10/90 | 0/15/85 | 0/30/70 | 5/5/90 | 5/20/75 |
| Acephalic sperm (sperm flagella) (106/ml) | 45 | 83 | 7.9 | 16.8 | 41 | 64 |
| Normal formsb | 0 | 2 | 0 | 0 | 0 | 0 |
| Vitality (%) | 40 | 40 | 53 | 53 | 72 | 40 |
WHO criteria (2010).
David criteria.
SUN5 exonic variant c.340G > a (p.Gly114Arg) inhibits splicing of exon 5
To further strengthen the evidence that loss of SUN5 function is a cause of acephalic spermatozoa syndrome, we investigated the effect of the c.340G > A (p.Gly114Arg) variant on splicing. This rare variant affects the last base of exon 5 and was previously reported in the homozygous state in a man with decapitated spermatozoa (8). We made two expression constructs, one with the variant (SUN5-G114R) and the other without (SUN5-WT), in which the genomic region from exon 4 to exon 6 of SUN5 was under the transcriptional control of the CMV promoter. We transfected SUN5-G114R and SUN5-WT into HeLa cells and assessed the splicing of exons 4, 5 and 6 by RT-PCR with primers in exon 4 and in exon 6. Our results show that both constructs produce transcripts lacking exon 5, but that only SUN5-WT produces transcripts that include exon 5 (Fig. 3). The transcripts without exon 5 from the SUN5-WT construct are certainly artefacts of our cellular assay system, since the same transcript was not amplified from human testis cDNA (Fig. 3). The absence of transcripts that include exon 5 from the SUN5-G114R construct shows that p.Gly114Arg greatly diminishes splicing efficiency, and will lead to exclusion of exon 5 (62 bases) from the mature SUN5 transcript. This will cause a frameshift preventing expression of the SUN5 protein.
The SUN5 p.Gly114Arg variant reduces splicing efficiency in HeLa cells. HeLa cells were transfected with two constructs in which the genomic region of SUN5 from exon 4 to exon 6 (3.6 kb) was under the control of the CMV promoter in pcDNA3.1+. There is one nucleotide difference between these constructs: SUN5-G114R carries the p.Gly114Arg variant, and SUN5-WT has the wild-type sequence. 24 h after transfection, RNA was extracted from cells, treated with DNase-I and RT-PCR performed with primers o5336 and o5337 from exon 4 and exon 6, respectively. Normal testis RNA (Testis-WT) and H2O were included as controls. The expected sizes are indicated to the right of the gel: 206 bp for normal splicing including exon 5, and 143 bp if exon 5 is excluded. The asterisk indicates the 200 bp band of the size marker (M) (1 kb plus - Life Technologies).
Discussion
Here, we present the second homozygous LOF mutation in SUN5, in three related men of North African origin who produce decapitated spermatozoa. We also provide functional evidence that the previously published variant p.Gly114Arg strongly affects splicing and will result in a reduction of SUN5 protein levels. Taken together with the recent report of biallelic SUN5 variants in Chinese men with acephalic sperm syndrome, our findings establish that loss of SUN5 function causes a failure of head-tail junction formation during spermiogenesis, and show that LOF mutation of SUN5 is also a cause of acephalic sperm phenotype in non-Asian populations.
The two brothers in our study have had treatment of their infertility by ICSI (14), using spermatozoa with the head and tail still joined. Despite the severity of the sperm phenotype in these men in whom the proximal centriole is not anchored to the nucleus (14), ICSI resulted in the development of good quality embryos in both cases. Following embryo transfer, pregnancy was obtained for each brother, resulting in the birth of a healthy girl and triplets, including identical twins. Thus for infertile men who produce decapitated spermatozoa consequent to the loss of SUN5 function, ICSI can be envisaged as a therapeutic solution. However a full evaluation of the efficiency of ICSI with SUN5-deficient spermatozoa, and the associated risks, will require additional cases and the long-term follow-up of the children conceived.
Several knockout mouse models share key features with acephalic spermatozoa syndrome in human: isolated male infertility, the presence of epididymal spermatozoa and an extreme fragility of the head-tail junction. The mouse genes inactivated in these models are Hook1 (15), Oaz3 (16), Odf1 (17) and Spata6 (18). Like SUN5, all are expressed specifically in the testis, and represent good candidates for genes that could carry causal mutations in unresolved cases of human male infertility with decapitated spermatozoa.
In conclusion, our results establish that SUN5 plays a key role in the attachment of the flagellum to the sperm head and demonstrate the prognostic value of testing for SUN5 mutations in men with decapitated spermatozoa seeking to father a child.
Materials and Methods
Patients
Patient 13-4335 and his wife were 34 and 31 years old, respectively, and had been trying to have a child for three years without success. Both had a normal karyotype, the patient was 46,XY,21ps+ and his wife’s was 46, XX. Both were healthy with no history of significant illness. His wife had regular menses, and her hysterosalpingography and hormone assessment was normal. Two semen samples from patient 13-4335, collected after three days of abstinence, were analysed and revealed severe oligozoospermia and complete teratozoospermia (Table 1). While there were only 2.3 × 106 intact spermatozoa/ml in each sample, there were, respectively, 41 × 106 and 64 × 106 headless flagella/ml. These characteristics are similar to those of his cousins (13-1662 and 13-2016) and are typical of acephalic spermatozoa syndrome.
The clinical details of the two brothers, 13-1662 et 13-2016, have been described elsewhere (14), but are summarised for comparison with those of 13-4335 in Table 1. The three men gave their informed consent for their samples to be used in the search for the genetic cause of their infertility.
SUN5 sequencing and deletion interval characterisation
The SUN5 gene was sequenced with the BigDye v1.1 terminator cycle sequencing kit (Applied Biosystems) following amplification of the coding exons with specific flanking primers (Supplementary Material, Table S1). PCR products for sequencing SUN5 and deletion mapping were amplified with Q5 Taq polymerase (New England Biolabs). The allele-specific PCR assay used to screen controls was performed as a duplex PCR with primers o5226 and o5340 that flank the deleted segment around exon 8 (deletion allele = 490 base pairs) and o5216 and o5217 that flank exons 2 and 3 of SUN5 (positive control = 802 bp).
Transfection of SUN5-WT and SUN5-G114R constructs
The segment from exon 4 to exon 6 of the SUN5 gene was amplified from genomic DNA from a fertile man using primers o5336 (exon 4) and o5337 (exon 6) with, respectively NheI and XhoI sites added at their 5 ´ end. To obtain the SUN5-WT construct the genomic fragment was cloned into the NheI and XhoI sites of pcDNA3.1-cGFP (pcDNA3.1+ - Life Technologies – with a C-terminal GFP cassette cloned between XhoI and ApaI sites). For the SUN5-G114R construct, a SUN5-G114R insert was cloned into pcDNA3.1-cGFP, following its amplification from SUN5-WT using two-step PCR-fusion with flanking primers o5336 and o5337 and central primers o5338 and o5339 specifying the p.Gly114Arg variant. The sequence of both constructs was determined and showed that neither contained any errors. All PCR steps used the Q5 DNA polymerase (New England Biolabs).
Constructs were transfected into HeLa cells using JetPrime (Polyplus-transfection) and the RNAs harvested after 24 h. RNA was reverse-transcribed to cDNA using the GFP-specific primer o5374 and the PCR performed using primers o5336 and o5337.
Supplementary Material
Supplementary Material is available at HMG online.
Acknowledgements
We are grateful to the patients who gave their informed consent to the use of their samples for research. The Biobank of the Department of Medical Genetics at La Timone Children's Hospital, Marseille for providing DNA samples. RAE was funded by the Islamic Development Bank and the Lebanese Association for Scientific Research (LASeR).
Conflict of Interest statement. None declared.
Funding
Agence de la biomédecine, AOR “AMP, diagnostic prenatal et diagnostic génétique” 2013, Inserm and Aix-Marseille Université, the Centre Hospitalier Universitaire of Toulouse and APHM (Assistance Publique Hôpitaux de Marseille), and Islamic Development Bank and the Lebanese Association for Scientific Research (LASeR).
