Abstract
SUMMARY Pachytene piRNAs, which comprise >80% of all small RNAs in the adult mouse testis, have been proposed to bind and regulate target RNAs like miRNAs, to cleave targets like siRNAs, or to lack biological function altogether. Although mutants lacking proteins that make pachytene piRNAs are male sterile, no biological function has been identified for any mammalian piRNA-producing locus. Here, we report that loss of piRNA precursor transcription from a conserved pachytene piRNA locus on mouse chromosome 6 (pi6) perturbs male fertility. Loss of pi6 piRNAs has no measurable effect on sperm quantity or transposon repression, yet pi6−/− mice produce sperm with defects in motility, egg fertilization, and embryo development, severely reducing pup production even at the peak of male reproduction. Our data establish a direct role for pachytene piRNAs in spermiogenesis and embryo viability and enable new strategies to identify the RNA targets of individual piRNA species.
Highlights
Normal male mouse fertility and spermiogenesis require piRNAs from the pi6 locus
Normal sperm motility and binding to zona pellucida require pi6 piRNAs
Sperm from pi6 males fail to support embryo development
Defects in pi6 sperm reflect changes in the abundance of specific mRNAs
INTRODUCTION
Only animals produce PIWI-interacting RNAs (piRNAs), 21–35-nt small RNAs that form the most abundant class of small RNA in the germline. In most animals, piRNAs protect the germline genome from transposons and repetitive sequences, and, in many arthropods, piRNAs fight viruses and transposons in somatic tissues (Houwing et al., 2007; Aravin et al., 2008; Batista et al., 2008; Das et al., 2008; Lewis et al., 2018). The mammalian male germline makes three classes of piRNAs: (1) 26–28 nt transposon-silencing piRNAs predominate in the fetal testis (Aravin et al., 2008); (2) shortly after birth 26–27 nt piRNAs derived from mRNA 3′ untranslated regions (UTRs) emerge (Robine et al., 2009); and (3) at the pachytene stage of meiosis, ~30 nt, non-repetitive pachytene piRNAs appear. Pachytene piRNAs accumulate to comprise >80% of all small RNAs in the adult mouse testis, and they continue to be made throughout the male mouse reproductive lifespan. These piRNAs contain fewer transposon sequences than the genome as a whole, and most pachytene piRNAs map only to the loci from which they are produced. The diversity of pachytene piRNAs is unparalleled in development, with >1 million distinct species routinely detected in spermatocytes or spermatids. Intriguingly, the sequences of pachytene piRNAs are not themselves conserved, but piRNA-producing loci have been maintained at the syntenic regions across eutherian mammals (Girard et al., 2006; Chirn et al., 2015), suggesting that the vast sequence diversity of pachytene piRNAs is itself biologically meaningful.
In mice, 100 pachytene piRNA-producing loci have been annotated (Girard et al., 2006; Grivna et al., 2006; Lau et al., 2006; Ro et al., 2007; Li et al., 2013). All are coordinately regulated by the transcription factor A-MYB (MYBL1), which also promotes expression of proteins that convert piRNA precursor transcripts into mature piRNAs, as well as proteins required for cell cycle progression and meiosis (Bolcun-Filas et al., 2011). Of the 100 piRNA-producing loci, 15 pairs of pachytene piRNA-producing genes are divergently transcribed from bidirectional, A-MYB-binding promoters (Li et al., 2013). The contribution of pachytene piRNAs from each piRNA-producing locus is unequal, with just five loci—pi2, pi6, pi7, pi9, and pi17—contributing to >50% of all pachytene piRNA production at 17 days postpartum (dpp).
Loss of proteins required to make pachytene piRNAs, including the pachytene piRNA-binding protein, MIWI (PIWIL1), invariably arrests spermatogenesis and renders males sterile (Deng and Lin, 2002; Reuter et al., 2011; Zheng and Wang, 2012; Li et al., 2013; Castañeda et al., 2014; Wasik et al., 2015). Yet, loss of the chromosome 17 pachytene piRNA-producing locus, 17-qA3.3-27363(–),26735(+) (henceforth, pi17), has no detectable phenotype or impact on male fertility (Homolka et al., 2015), even though pi17 produces ~30% of all pachytene piRNAs. Similarly, mice disrupted in expression of a piRNA cluster on chromosome 2 are viable and fertile (P.-H.W., K.C., and PDZ, unpublished; Xu et al., 2008). Consequently, the function of pachytene piRNAs in mice is actively debated. One model proposes that pachytene piRNAs regulate meiotic progression of spermatocytes by cleaving mRNAs during meiosis (Goh et al., 2015; Zhang et al., 2015). Another model posits that pachytene piRNAs direct degradation of specific mRNAs via a miRNA-like mechanism involving mRNA deadenylation (Gou et al., 2014). A third model proposes that MIWI functions without piRNAs, and that piRNAs are byproducts without a critical function (Vourekas et al., 2012). Compelling evidence exists to support each model.
In fact, direct demonstration of piRNA function in any animal has proven elusive. Only two piRNA-producing loci have been directly shown to have a biological function— both are in flies and were identified genetically before the discovery of piRNAs (Livak, 1984; Livak, 1990; Palumbo et al., 1994; Pélisson et al., 1994; Bozzetti et al., 1995; Prud’homme et al., 1995; Robert et al., 2001; Robert et al., 2001; Mével-Ninio et al., 2007). In male flies, piRNAs from Suppressor of Stellate, a multi-copy gene on the Y chromosome, silence the selfish gene Stellate, and deletion of Suppressor of Stellate leads to Stellate protein crystals in spermatocytes (Aravin et al., 2001; Aravin et al., 2003). In female flies, deletion of the piRNA-producing flamenco gene, which is expressed in somatic follicle cells that support oogenesis, leads to gypsy family transposon expression and infertility (Brennecke et al., 2007; Saito et al., 2009).
Here, we report that a small promoter deletion in the chromosome 6 pachytene piRNA cluster 6-qF3-28913(–),8009(+) (henceforth, pi6) that eliminates pi6 piRNA production disrupts male fertility. The pi6 locus, one of the five most productive piRNA-producing loci in mice, generates 5.8% of pachytene piRNAs in the adult testis and is conserved among eutherian mammals. Mice lacking pi6-derived piRNAs produce normal numbers of sperm and continue to repress transposons. However, pi6 mutant sperm fertilize eggs poorly due to defective sperm motility and zona pellucida penetration. Consistent with these phenotypes, the steady-state abundance of mRNAs encoding proteins crucial for cilial function, zona pellucida proteolysis, and egg binding was significantly decreased in sperm progenitor cells from pi6 males. Our findings provide direct evidence for a biological function for pachytene piRNAs in male mouse fertility, and pi6 promoter deletions provide a new model for the future identification of piRNA targets in vivo.
RESULTS
pi6 Promoter Deletion Eliminates pi6 pachytene piRNAs
To eliminate production of pi6 pachytene piRNAs while minimizing the impact on adjacent genes, we used a pair of single-guide RNAs to delete 227 bp, including the A-MYB-binding promoter sequences, from pi6 (Figure 1, S1A, and S1B, and Table S1; Li et al., 2013). For comparison, we created an analogous promoter deletion in pi17. We established stable mutant lines (pi6em1−1, −2, and −3 in Figure S1B) from three founders whose pi6 promoter deletion sizes range from 219 to 230 bp and differ at their deletion boundaries, reflecting imprecise DNA repair after Cas9 cleavage. All three deletions eliminated pi6 primary transcripts and mature pachytene piRNAs from both arms of the locus (Figure 1). Because these lines were created using the same pair of sgRNA guides, we refer to all as the pi6em1 allele.
pi6 is Required for Male Mouse Fertility
When paired with C57BL/6 females, pi6em1/em1 males between 2 and 8 months old produced fewer pups compared to their littermates, even at peak reproductive age (Figure 2A and S2A). In six months, C57BL/6 males produced 7 ± 1 (n = 5) litters, while pi6eml/em1 males produced 2 ± 2 (n = 6) litters. The significantly smaller number of progeny produced by pi6em1/em1 males over their reproductive lifetime does not reflect fewer pups produced in each litter: pi6em1/em1 males sired 5 ± 2 (n = 4) pups per litter compared to 6 ± 2 (n = 27) pups per litter for C57BL/6 control males (Figure 2A). Moreover, pi6em1/em1 males regularly produced mating plugs, a sign of mating, in cohabiting females. Instead, the reduced progeny from pi6em1/em1 males reflects two abnormal aspects of their fertility (Figure 2B). First, 29% of pi6em1/em1 males never produced pups. Second, the mutants that did sire pups did so less frequently. These defects are specific for the loss of pi6 piRNAs in males, because pi6+/em1 heterozygous males and pi6em1/em1 homozygous mutant females showed no discernable phenotype. As observed previously for a partial-loss-of-function pi17 promoter deletion (Homolka et al., 2015), males and females carrying a ~583-bp promoter deletion in pi17 were fully fertile, despite loss of primary transcripts and mature piRNAs from both arms of the pi17 locus (Figure 1).
To test that the reduced fertility of pi6em1/em1 male mice reflects loss of the pi6 promoter—and not an undetected Cas9-induced off-target mutation elsewhere in the genome—we used Cas9 and a second pair of sgRNAs to generate a 117 bp pi6 promoter deletion, pi6em2 (Figures 1, S1A, and S1C, and Table S1). Like pi6em1/em1 male mice, pi6em2/em2 males produced neither primary pi6 transcripts nor mature pi6 piRNAs and showed reduced fertility (Figure S2A). We conclude that pi6 piRNAs are required for C57BL/6 male fertility in mice.
pi6em1/em1 Males Produce Fewer Embryos
pi6 mutant male matings were less likely to produce fully developed embryos. We examined the embryos produced by natural mating of C57BL/6 females housed with C57BL/6, pi6+/em1, or pi6em1/em1 males at 8.5, 14.5, or 16.5 days after occurrence of a mating plug. At 8.5 days after mating, C57BL/6 females housed with pi6em1/em1 males carried fewer embryos (2 ± 2, n = 3) compared to the females paired with pi6+/em1 (6 ± 5, n = 2) or C57BL/6 control (7 ± 4, n = 1) males (Figure 2C). At 14.5 and 16.5 days postmating, female mice paired with pi6em1/em1 males had even fewer embryos. Consistent with the observation that naturally-born pups sired by pi6em1/em1 males were rare but healthy, the surviving embryos resulting from natural mating showed no obvious abnormalities.
Moreover, pi6 piRNAs appear to play little if any role in the soma of the developing embryo. pi6+/em1 heterozygous males mated to pi6+/em1 heterozygous females yielded progeny at the expected Mendelian and sex ratios. Moreover, the weight of pi6em1/em1 homozygous pups (28.3 ± 0.6 g, n = 8) that developed to adulthood was indistinguishable from their C57BL/6 (26.9 ± 0.3 g, n = 8) or heterozygous littermates (28.6 ± 0.3 g, n = 8) (Figure S2B). We detected no difference in the gross appearance or obvious changes in behavior among these pups.
pi6em1/em1 Males Produce Mature Spermatozoa
Two-to-four months after birth, both pi6+/em1 and pi6em1/em1 testes weighed slightly less than C57BL/6 testes (Figure S2B). Nonetheless, pi6em1/em1 testis gross histology was normal, with all expected germ cell types present in seminiferous tubules and sperm clearly visible in the lumen (Figure 2D). The quantity of caudal epididymal sperm produced by pi6em1/em1 mice (19 ± 10 million sperm per ml; n = 6) was also comparable to that of their pi6+/em1 (23 ± 7 million sperm/ml; n = 4) or C57BL/6 (20 ± 10 million sperm per ml; n = 13) littermates (Figure 2E).
Although pi6em1/em1 mice produce normal numbers of sperm, the sperm showed signs of agglutination compared to C57BL/6 sperm after 90 min of incubation in vitro, and ~10% of pi6em1/em1 caudal epidydimal sperm had abnormal head morphology (Figure S2C). Defects in germ cell chromosomal synapsis, triggering errors in gene expression, have been linked to abnormal sperm head shape (Wong et al., 2008; de Boer et al., 2015). In fact, 22 ± 7 percent of pi6em1/em1 pachytene spermatocytes had unsynapsed sex chromosomes or incompletely synapsed autosomal chromosomes, compared to 7 ± 3 percent for C57BL/6 (n = 4) (Figure S2E).
pi6em1/em1 Sperm Fail to Fertilize
pi6 mutant males produce ordinary numbers of normally shaped sperm (~90%), yet are ineffectual at siring offspring. We used in vitro fertilization (IVF) to distinguish between defects in mating behavior and sperm function, incubating sperm from C57BL/6, pi6+/em1, or pi6em1/em1 males with wild-type oocytes and scoring for the presence of both male and female pronuclei and the subsequent development of the resulting bi-pronuclear zygotes into two-cell embryos 24 h later (Figure 3A). The majority of oocytes incubated with C57BL/6 (91 ± 5%; n = 5) or pi6+/em1 (60 ± 35%; n = 3) sperm developed into two-cell embryos. By contrast, only 7 ± 5% (n = 7) of oocytes incubated with pi6em1/em1 sperm reached the two-cell stage. The majority of these oocytes remained one-cell embryos, and few contained a male pronucleus, suggesting that pi6em1/em1 sperm are defective in fertilization.
pi6em1/em1 Sperm Nuclei Support Fertilization
The best studied piRNA function is transposon silencing, and mouse pi2 has been proposed to be involved in LINE1 element silencing, although pi2 mutant males are fertile (Xu et al., 2008). Moreover, LINE1 transcript abundance increases in mice bearing inactivating mutations in the catalytic site of MIWI (Reuter et al., 2011). Transposon activation can produce DNA damage, and genomic integrity is critical for fertilization (Ahmadi and Ng, 1999; Morris et al., 2002; Bourc’his and Bestor, 2004; Lewis and Aitken, 2005). However, pachytene piRNAs are depleted of repetitive sequences in contrast to other types of piRNA-producing genomic loci (Figure S3A; Aravin et al., 2006; Girard et al., 2006; Gainetdinov et al., 2018).
We asked whether the defect in fertilization by pi6em1/em1 might reflect DNA damage or epigenetic dysregulation of the pi6em1/em1 sperm genome. pi6+/em1 or pi6em1/em1 sperm heads were individually injected into the cytoplasm of wild-type oocytes (intracytoplasmic sperm injection, or ICSI) (Figure 3B), bypassing the requirement for sperm motility, acrosome reaction, egg binding, or sperm-egg membrane fusion (Kuretake et al., 1996). pi6em1/em1 sperm heads delivered by ICSI fertilized the oocyte at a rate similar to that of pi6+/em1 sperm: 66% of oocytes injected with homozygous mutant pi6em1/em1 sperm heads reached the two-cell stage, compared to 79% for pi6+/em1. Thus, most pi6em1/em1 nuclei are capable of fertilization.
The steady-state abundance of transposon RNA in pi6em1/em1 testicular germ cells further supports the view that the fertilization defect caused by loss of pi6 piRNAs does not reflect a failure to silence transposons. We used RNA-seq to measure the abundance of RNA from 1,007 transposons in four distinct germ cell types, purified by fluorescence-activated cell sorting: pachytene spermatocytes (4C), diplotene spermatocytes (4C), secondary spermatocytes (2C), and spermatids (1C). pi6 piRNAs are plentiful in pachytene spermatocytes onwards (Figure S3B), yet when pi6 piRNAs were eliminated, we found no significant changes in steady-state RNA abundance (i.e., an increase or decrease ≥ 2-fold and FDR ≤ 0.05) for any transposon family compared to C57BL/6 cells (Figure S3C). We also note that, similar to C57BL/6 testis, γH2AX expression is confined to meiotic spermatocytes in pi6em1/em1 testis, indicating absence of DNA damage (data not shown). Together with the rescue of the fertilization defects of pi6em1/em1 sperm by ICSI, these data suggest that transposon silencing is unlikely to be the biological function of pi6 piRNAs.
Impaired Motility in pi6 Mutant Sperm
To assess whether abnormal sperm motility might contribute to pi6em1/em1 male subfertility, we observed freshly extracted caudal epididymal sperm from pi6em1/em1 or C57BL/6 mice for 5 h. Ten minutes after sperm extraction, most pi6em1/em1 sperm moved more slowly than C57BL/6 control sperm (Movies S1 and S2). With time, pi6em1/em1 sperm motility declined more rapidly than C57BL/6 sperm (Movies S3–S10). At 4 and 5 h, most pi6em1/em1 sperm only moved in place and showed signs of agglutination (Movies S8 and S10).
To quantify the differences between pi6 mutant and control sperm, we used computer-assisted sperm analysis (CASA) to measure pi6em2/em2 sperm motility 10 min after isolation (Mortimer, 2000). While control sperm swam at a path velocity comparable to previously reported (110 ± 50 μm/sec for 221 ± 75 cells measured; n = 3; Ren et al., 2001), pi6em2/em2 sperm moved at a lower average path velocity (80 ± 60 μm/sec for 232 ± 57 cells measured; n = 3) (Table 1). Similarly, The pi6em2/em2 sperm also showed less forward, progressive movement (progressive velocity = 50 ± 60 μm/sec for 232 ± 57 cells measured; n = 3) compared to control sperm (progressive velocity = 70 ± 50 μm/sec for 221 ± 75 cells measured; n = 3). For comparison, knockout of CatSper1 leads to ~65% reduction in path velocity and ~62% reduction in progressive velocity (Ren et al., 2001). As a population, the speed and progressivity of pi6 mutant sperm motility patterns varied more widely than control sperm (Movies S1–S10 and Table 1). Lower average path and progressive velocity in sperm populations is linked to worse outcomes in fertilization and pregnancy in IVF (Donnelly et al., 1998). Thus, the slower and less progressive movement in pi6em1/em1 sperm likely contributes to the subfertility of pi6em1/em1 males.
pi6 Mutant Sperm Struggle to Penetrate the Zona Pellucida
Mammalian spermatozoa stored in the epididymis are dormant. Sperm “capacitate,” i.e., resume maturation, only upon entering the female reproductive tract (de Lamirande et al., 1997). Upon capacitation, sperm become capable of undergoing the acrosome reaction, which is required to bind and penetrate the outer oocyte glycoprotein layer, the zona pellucida (Florman and Storey, 1982; de Lamirande et al., 1997; Jin et al., 2011). To test whether the defect in fertilization by pi6 mutant sperm was due to impaired binding to or penetration of zona pellucida, we compared IVF using wild-type oocytes with their zona pellucida either intact or removed (Figure 4A). As before, 10 ± 6% (n = 3) of intact oocytes incubated with pi6em1/em1 sperm reached the two-cell stage, compared to 94 ± 5% (n = 3) for C57BL/6 sperm (Figure 4B). Strikingly, removing the zona pellucida from the wild-type oocytes fully rescued the fertilization rate of pi6 mutant sperm: 92 ± 7% (n = 3) of zona pellucida-free oocytes incubated with pi6em1/em1 sperm reached the two-cell stage, compared to those with intact zona pellucida (10 ± 6%; n =3)
Ex vivo, the acrosome reaction occurs spontaneously in some sperm and can be further triggered by inducing Ca2+ influx using the ionophore A23187 (Talbot et al., 1976), which results in an acrosome reaction visually indistinguishable from that triggered by natural ligands such as progesterone (Osman et al., 1989) or ZP3 (Arnoult et al., 1996), while bypassing signaling pathways essential for acrosome reaction in vivo (Tateno et al., 2013) (Figure 4C and 4D). The spontaneous acrosome reaction rates for C57BL/6 (19 ± 3%; n = 3) and pi6 mutant sperm were similar (17 ± 8%; n = 3). Acrosome reaction triggered by ionophore-induced Ca2+ influx differed between the two genotypes: 45 ± 14% of pi6 mutant sperm (n = 3) underwent partial or complete reaction, compared to 66 ± 6% (n = 3) for C57BL/6 (Figure 4C). Our data suggest that pi6 mutant sperm less effectively undergo an acrosome reaction triggered by ionophore-induced Ca2+ influx, a defect expected to impair binding and penetrating the zona pellucida.
Potential Role of Paternal pi6 piRNAs in Embryo Development
Even when pi6 sperm successfully fertilize the oocyte, the resulting heterozygous embryos are less likely to complete gestation. Two-cell embryos generated by IVF using heterozygous or homozygous pi6 mutant or C57BL/6 control sperm were transferred to C57BL/6 surrogate mothers (Figure 5A). At least half of embryos from pi6+/em1 (50 ± 10%; n = 3) or C57BL6 control (70 ± 10%; n = 3) sperm developed to term (Figure 5B), a rate typical for the C57BL/6 background (González-Jara et al., 2017).
The low number of fertilized two-cell embryos produced in IVF using pi6em1/em1 speed precluded transferring the standard number of embryos to surrogate mothers. For example, in two IVF experiments using pi6em1/em1 sperm, only 5 or 7 embryos could be transferred; the surrogate females failed to become pregnant (Figure 5B and S4A, Trials 1 and 2). In theory, this result might suggest a paternal role for pi6. A more mundane explanation is that the low number of embryos transferred reduced the yield of live fetuses, as reported previously (McLaren, 1955; Johnson et al., 1996; González-Jara et al., 2017). We conducted additional experiments to distinguish between these two possibilities. Oocytes were again fertilized by IVF with pi6em1/em1 or C57BL/6 control sperm, and two-cell embryos transferred to surrogate females, but matching the number of embryos transferred to each surrogate for the two sperm genotypes. We used two strategies. First, similar numbers of embryos derived from pi6em1/em1 sperm and filler embryos derived from control sperm were transferred to separate oviducts (Figure 5B, Trials 3 and 4). Again, fewer embryos developed to term for pi6em1/em1 (17%) compared to control sperm (37%). Second, embryos were mixed before transfer and then equal numbers of embryos, selected randomly, were implanted in each oviduct (Figure 5B, Trial 5). Pups isolated by cesarean section 18.5 days after transfer were genotyped by PCR. In this experiment, only 40% of embryos derived from pi6em1/em1 sperm developed to term, compared to 80% of filler embryos. Finally, in one experiment (Trial 6) where we obtained sufficient numbers of embryos derived from pi6em1/em1 sperm, 10 pi6em1/em1- derived two-cell embryos were transferred to each oviduct of the surrogate female. Nevertheless, only 15% of the pi6em1/em1-derived embryos developed to term, compared to 85% of the control.
We also monitored pre-implantation development ex vivo for up to 96 h, a period during which the one-cell embryo develops into a blastocyst. Of all the oocytes incubated with pi6em1/em1 sperm, 40% remained one cell without evidence of a male pronucleus, presumably because they were not fertilized by pi6em1/em1 mutant sperm. Among the remaining 60% oocytes that progressed to at least two-cell stage, which indicated successfully fertilization by pi6em1/em1 sperm, 82% showed delayed development, requiring 48 h to reach the two-cell stage. None of these developed further. Only 3% of fertilized oocytes progressed to the blastocyst stage by 96 h, compared to 98% of oocytes fertilized by C57BL/6 sperm (Figure 5C).
Further support for this idea comes from transfer of embryos generated by ICSI (Figure 5D). ICSI with pi6em1/em1 or pi6+/em1 sperm yielded comparable normal numbers of fertilized oocytes (Figure 3B), so no filler embryos were used; all embryos were transferred into a single oviduct of the surrogate female. In two independent experiments in which embryos generated by ICSI were transferred to surrogate mothers, only 19% of two-cell embryos derived from pi6em1/em1 sperm heads developed to term, compared to 34% for embryos fertilized with pi6+/em1 (Figure 5C). Only four of seven (57%) surrogate mothers carrying embryos derived from pi6em1/em1 sperm became pregnant. All three surrogate mothers receiving embryos derived from pi6+/em1 sperm became pregnant (Figure S4B).
We note that the live fetuses generated using pi6em1/em1 sperm in IVF or sperm heads in ICSI, like those produced by natural mating using pi6em1/em1 males, showed no obvious morphological abnormalities and grew to adulthood normally when fostered by host mothers. This suggests a direct or indirect requirement for paternal pi6 piRNAs in early embryogenesis.
Changes in Spermatocyte mRNA Abundance Accompany Loss of pi6 piRNAs
To characterize the molecular phenotypes of pi6 and pi17 mutants, we used RNA-seq to measure steady-state RNA abundance in pachytene spermatocytes, diplotene spermatocytes, secondary spermatocytes, and spermatids purified from pi6em1/em1, pi17−/−, and C57BL/6 adult testis (Figure 6A). pi6 and pi17 precursor transcripts are abundant in meiotic pachytene spermatocytes (tetraploid), decrease in diplotene spermatocytes, and fall to low levels in post-meiotic spermatids (haploid) (Figure S5B). Compared with C57BL/6 controls, pi6em1/em1 mutants had widespread changes in mRNA abundance in pachytene spermatocytes—481 mRNAs more than doubled, while 394 fell by more than half (FDR ≤ 0.05; Figure 6B and S5A, and Table S2)—but caused little alteration in mRNA abundance in diplotene spermatocytes, secondary spermatocytes, or spermatids. In contrast, pi17−/− mutants showed significant changes in mRNA abundance in diplotene (10 mRNAs increased, 267 decreased) and secondary spermatocytes (103 mRNA increased, 400 decreased) but not in pachytene spermatocytes or spermatids (Figure S5A). Among the mRNAs that changed in the diplotene spermatocytes of pi17−/− mutants, 56% remained different from controls in secondary spermatocytes in these mutants. These data suggest that, despite similar temporal expression, pi6 piRNAs function primarily in pachytene spermatocytes, while pi17 piRNAs may be more important at a later stage of spermatogenesis. Furthermore, 734 (84%) of mRNAs with altered abundance in pi6em1/em1 pachytene spermatocytes were unchanged in any pi17−/− sorted germ cell type we examined, suggesting that distinct sets of genes are dysregulated in pi6em1/em1 and pi17−/− mutants.
The abundance of piRNAs from the other four major pachytene piRNA clusters, including pi17, was unaffected by loss of pi6 piRNAs, and loss of neither pi6 nor pi17 piRNAs had any significant effect on the abundance of mRNAs encoding piRNA pathway proteins (Table S3), suggesting that the changes in mRNA abundance in pi6em1/em1 or pi17−/− cells reflect direct regulation of target genes by pi6 or pi17 piRNAs or the downstream regulation through the direct targets of these piRNAs.
Gene Ontology (GO) analysis of the 481 up-genes found over 354 significantly enriched GO biological processes (FDR ≤ 0.01 and enrichment ≥ 2). Curiously, 106 of these GO terms correspond to developmental processes that do not normally occur in testis, suggesting a failure to suppress inappropriate programs without pi6 piRNAs. Similarly, pi6em1/em1 mutants show increased mRNA abundance for 20 transcription factors that normally act in undifferentiated spermatogonia or spermatogonial stem cells or the stem cells of other tissues (Table S4).
The mRNA abundance of several miRNA pathway genes also increased in pi6em1/em1 pachytene spermatocytes, including Lin28a (5.6-fold), Zc3h7b (5-fold), and Ajuba (5.3-fold; Figure S5C) (Dresios et al., 2005; James et al., 2010; Pilotte et al., 2011; Piskounova et al., 2011). LIN28A inhibits let-7 biogenesis by binding to the loop of pre-let-7, blocking its processing by DICER (Piskounova et al., 2008; Hagan et al., 2009; Heo et al., 2009), and let-7 promotes Lin28a degradation by binding two conserved sites in the Lin28a 3′ untranslated region (Reinhart et al., 2000; Agarwal et al., 2015) predicting that let-7 levels should fall and let-7 targets should rise in pi6em1/em1. Indeed, in pi6em1/em1 adult testis, the aggregate abundance of let-7a, let-7b, let-7c, let-7e, let-7f, let-7g, and let-7i, the seven most abundant let-7 family members (≥ 10 ppm in wild-type testis) fell to less than half of wild-type, suggesting pi6 regulation of downstream target genes via let-7. Moreover, 48 predicted let-7 targets (Agarwal et al., 2015) increased in the absence of pi6em1/em1, including Lin28a and the mRNAs encoding three transcription factors: Sall4 (increased 8.7-fold), Elf4 (increased 7-fold), and Pbx2 (increased 6.7-fold). SALL4 is normally expressed in undifferentiated spermatogonia where it represses genes that specify somatic gene expression programs (Gassei and Orwig, 2013; Yamaguchi et al., 2015; Chan et al., 2017). ELF4 has been implicated in regulation of quiescence in hematopoietic stem cells (Lacorazza et al., 2006). Our data suggest that piRNAs, miRNAs, and transcription factors collaborate to ensure precise regulation of gene expression in spermatogenesis.
Genes that Function in the Cilium Assembly, Cilium Motility, and Fertilization Pathways Decrease in mRNA Abundance upon Loss of pi6 piRNAs
GO analysis of the 394 down-genes revealed only 36 significantly enriched GO biological processes (FDR ≤ 0.01 and fold enrichment ≥ 2), of which 34 are related to the production and function of sperm and can be organized into four sets (Table S5). One set encompasses broad spermatogenesis terms (e.g., male gamete generation, 4.6-fold enriched, FDR = 5.8 × 10−11; sperm capacitation, 12-fold enriched, FDR = 7.4 × 10−3) while three sets are highly specific and match the in vivo phenotypes of pi6 mutant males. The first specific set includes cilium assembly (6.2-fold enriched, FDR = 4.1 × 10−9) and axonemal dynein complex assembly (18-fold enriched, FDR = 1.1 × 10−5). The second set contains sperm motility (13-fold enriched, FDR = 6.0 × 10−10) and cilium movement involved in cell motility (27-fold enriched, FDR = 2.0 × 10−3). The third set involves fertilization (6.2-fold enriched, FDR = 1.7 × 10−5) and binding of sperm to zona pellucida (12-fold enriched, FDR = 2.3 × 10−3). None of these three sets of GO terms is enriched in the 481 genes whose mRNA levels increased in pi6em1/em1 pachytene spermatocytes. The three sets of specific GO terms contain 28, 36, and 22 genes whose mRNAs decreased (63 total and 23 shared between sets; Figure 6C and Table S6). The last two general GO terms—microtubule-based process (GO:0007017; with 27 genes whose mRNA abundance declined) and organelle assembly (GO:0070925; with 28 genes whose mRNA abundance decreased)—likely gained their enrichment from the large number of genes they share with Cilium assembly and Sperm motility processes (23 and 25 genes for the two GO terms, respectively).
Master Regulators of Cilium Assembly and Sperm Motility
The 63 Cilium Assembly, Sperm Motility, or Fertility genes with reduced mRNA abundance in pi6 mutants include two transcription factors, Rfx2 and Foxj1, that act as master regulators of ciliogenesis (Figure 6C). Like pi6 itself, Rfx2 transcription is activated by A-MYB, and RFX2 also binds its own promoter (Horvath et al., 2009). Of the genes with decreased mRNA abundance in pi6em1/em1 pachytene spermatocytes, 31 both bind RFX2 and have reduced mRNA abundance in Rfx2−/− testis, suggesting they are direct targets of RFX2 (Figure 6C and Table S7) (Kistler et al., 2015). Intriguingly, 23 of these 31 RFX2-regulated genes also bind A-MYB (Table S7). A-Myb mRNA levels are normal in pi6em1/em1, which may account for the relatively modest decreases in the mRNA abundance of these 23 genes. Unlike RFX2, the role of FOXJ1 in sperm flagellar assembly has not been extensively studied but its role in general ciliogenesis is well established: FoxJ1−/− mouse died at or soon after birth due to absence of cilia in multiple organs (Chen et al., 1998; Blatt et al., 1999; Brody et al., 2000; Yu et al., 2008). Six genes—Tekt4, Spa17, Drc1, Rsph1, Meig1, and Tsnaxip1—out of the 394 genes with reduced mRNA abundance in pi6em1/em1 pachytene spermatocytes are regulated by FOXJ1 in ciliogenesis in other tissues (Yu et al., 2008; Stauber et al., 2017). Fourteen genes whose mRNA abundances decrease in pi6em1/em1 are uniquely annotated with the GO term Fertilization (Figure 6C and Table S6). Several are required for sperm to bind the zona pellucida or for acrosome function, including Acrosin (halved in pi6em1/em1 pachytene spermatocytes), Adam3 (decreased 2.5-fold), Zpbp2 (decreased 3.3-fold), and the FOXJ1-regulated gene Spa17 (decreased 5-fold). Among the genes with decreased or increased mRNA abundance in pi6em1/em1 cells, 28 have been reported to disrupt mouse or human male fertility or to play a role in spermatogenesis, spermiogenesis, or sperm function (Table S8).
DISCUSSION
Deletion of the mouse pachytene piRNA pi6 locus results in specific, quantifiable defects in male fertility. These include impaired sperm mobility and failure in sperm to bind and penetrate the zona pellucida. The male fertility defects accompanying loss of pi6 piRNAs are specific to this locus, as deletion of the promoter of pi17, which eliminates pi17 piRNAs, had no detectable effect on male or female fertility or viability, as reported previously (Homolka et al., 2015). The phenotypic defects of pi6 mutants reflect the molecular changes—decreased steady-state abundance of mRNAs encoding proteins that function in cilial motility and fertilization. Mutations in four of these genes also cause infertility in men. The molecular changes were detected only in pachytene spermatocytes but not in diplotene spermatocytes, secondary spermatocytes, or spermatids. By contrast, RNA-seq for 17.5 dpp or adult pi6em1/em1 testes revealed no changes in mRNA abundance compared to controls. These results underscore the power of analyzing sorted germ cells.
Pachytene piRNAs have been proposed to act collectively in meiotic spermatocytes or post-meiotic spermatids to target mRNAs for destruction (Gou et al., 2014; Goh et al., 2015), but the extent to which piRNAs from different pachytene piRNA loci regulate overlapping sets of targets is unknown. Transcriptome analysis of sorted germ cells from pi6em1/em1 and pi17−/− mutant mice revealed distinct changes in mRNA abundance, suggesting that, despite the coordinate temporal expression of pachytene piRNAs, individual pachytene piRNA loci regulate distinct sets of genes. Given that pi6 produces 95,677 distinct piRNA sequences, the phenotypic specificity of the pi6 mutant is extraordinary. For both miRNAs and siRNAs, the seed sequence plays a central role in determining a small RNA’s regulatory target. Assuming that pachytene piRNAs find their target RNAs by a similar mechanism, the sequence diversity of the small RNAs produced by individual loci is enormous: pi6 piRNAs encompass 9,880 distinct seed (g2–g8 or 7mer-m8; Bartel, 2009) and 17,304 distinct extended seed sequences (g2–g9) in adult mouse testis, while pi17 generates 134,358 distinct piRNA sequences, encompassing 11,324 distinct g2–g8 seed and 21,972 distinct g2–g9 seed sequences. Yet, the g2–g9 seed sequences of the 100 most abundant pi6 piRNAs are not found among the 100 most abundant pi17 piRNAs. Furthermore, 97 of these pi6 g2–g9 seed sequences are not found among any of the 100 most-abundant piRNAs produced by pi2, pi7, pi9, or pi17. Together with pi6, these loci produce more than half of all pachytene piRNAs. The unique seed sequences of the most abundant pi6 piRNAs are consistent with the lack of compensation of loss of pi6 piRNAs by other piRNA-producing loci.
We envision that piRNAs from distinct loci target overlapping sets of genes, ensuring robust control of mRNA abundance across spermatogenesis. Our data show that pi6 piRNAs regulate—directly or by regulating upstream factors—a specific set of mRNAs whose protein products must be eliminated for successful spermiogenesis. In this view, pi6 piRNAs target mRNAs whose expression must decline at the onset of the pachynema in order to allow new sets of mRNAs to accumulate, such as the RFX2-regulated genes required for ciliogenesis. While we cannot exclude a direct role for piRNAs in activating gene expression or increasing mRNA stability, we note that the overwhelming majority of siRNAs and miRNAs in plants and animals act as repressors not activators.
The phenotypic and molecular specificity of pi6 may reflect a lower degree of redundancy with other piRNA clusters. Nonetheless other piRNA clusters may partially rescue the pi6 phenotype, accounting for the incomplete penetrance of the pi6 sterility phenotype. Conversely, the lack of a phenotype for other pachytene piRNA clusters may simply reflect greater redundancy with their piRNA-producing peers. Loss of regulation of the targets of pi17 piRNAs may be compensated by piRNAs from other loci. Testing this hypothesis is clearly a prerequisite for explaining why loss of pi6 and not pi17 piRNAs has a measurable biological consequence.
Beyond the requirement for pi6 piRNAs to produce fully functional sperm, pi6 piRNAs appear to play an additional role in embryo development. Our data suggest that the arrested development and reduced viability of embryos derived from pi6 mutant sperm reflects a paternal defect and not the embryonic genotype. Damaged sperm DNA, abnormal sperm chromatin structure, and failure to form a male pronucleus in fertilized embryos have been reported to be linked to retarded embryo development (Sakkas et al., 1998; Borini et al., 2006). Our analysis of transposon RNA abundance in pi6 mutant germ cells argues against a role for pi6 piRNAs in transposon silencing during spermatogenesis, but we cannot currently exclude a direct or indirect role for pi6 piRNAs in silencing transposons in the early embryo (Peaston et al., 2004). Of course, DNA damage might reflect incomplete repair of the double-stranded DNA breaks required for recombination, rather than transposition or transposon-induced illegitimate recombination.
How piRNAs identify their targets remains poorly understood, in part because suitable biochemical or genetic model systems are not available. The availability of a mouse mutant missing a specific set of piRNAs whose absence causes a readily detectable phenotype should provide an additional tool for understanding the base-pairing rules that govern the binding of piRNAs to their RNA targets and for unraveling the regulatory network created by pachytene piRNAs.
SUPPLEMENTAL INFORMATION
Supplemental Information includes Extended Experimental Procedures, Figures S1–5, Tables S1–S7, and Movies S1–S10.
AUTHOR CONTRIBUTIONS
P.H.W., K.C., Y.F., Z.W., and P.D.Z. conceived and designed the experiments. P.H.W. and K.C. performed the experiments. Y.F. analyzed the sequencing data. D.M.Ö generated A-MYB ChIP-seq datasets. P.H.W., Y.F., and P.D.Z. wrote the manuscript.
ACKNOWLEDGEMENTS
We thank P. Cohen and K. Grive at Cornell University for generously sharing protocols and advice on germ cell sorting and meiotic chromosome studies; H. Florman and P. Visconti for sharing protocols and advice on sperm studies; the UMMS Transgenic Animal Modeling Core for advice on fertility test and embryo phenotype; the UMMS FACS core for advice on and help with germ cell sorting; and members of our laboratories for critical comments on the manuscript. This work was supported in part by National Institutes of Health grants GM65236 to P.D.Z. and P01HD078253 to P.D.Z. and Z.W.
STAR METHODS
Mouse mutants
Mice were maintained and sacrificed according to guidelines approved by the Institutional Animal Care and Use Committee of the University of Massachusetts Medical School (A-2222-17).
Small guide RNAs (sgRNAs) flanking piRNA promoters were designed using CRISPR design tools (crispr.mit.edu/). DNA oligos containing guide sequences were cloned into pX330 vectors (Cong et al., 2013), and their cleavage activity tested in NIH3T3 cells by co-transfecting pX330 constructs containing sgRNA sequences and puromycin-resistant plasmid (pPUR) using TransIT-X2 (Mirus Bio, Madison, WI). Puromycin (3 μg/μl) was added 24 h after transfection and DNA extracted 48 h afterwards. Promoter deletions were detected by PCR using primers flanking the predicted Cas9 cleavage sites.
For mice, sgRNAs were generated by in vitro transcription and purified by electrophoresis on 8% (w/v) polyacrylamide gels. To generate the pi6em1/em1 and pi17−/− lines used in this study, in vitro transcribed sgRNAs (10 ng/μl each) targeting pi6 and pi17 were mixed with Cas9 mRNA (40 ng/μl) and injected together into the cytoplasm of one-cell C57BL/6 zygotes (RNA only). For some founders, the sgRNA and Cas9 mRNA mixture was combined with pX330 plasmids expressing the same four sgRNAs and Cas9 and injected into both the cytoplasm and pronuclei of one-cell C57BL/6 zygotes (RNA + DNA). For pi6em2/em2, in vitro transcribed sgRNAs and Cas9 mRNA were injected into the cytoplasm of one-cell C57BL/6 embryos. Embryos were transferred to pseudopregnant females using standard methods. To screen for mutant founders, DNA was extracted from small pieces of tail clipped from three-week-old pups (Truett et al., 2000). Deletions were detected by PCR, and PCR products purified and cloned into TOPO blunt vectors. Mutant sequences were determined by Sanger sequencing.
Mouse fertility test
Each 2–8 month-old male mouse was housed with one 2–4 month-old C57BL/6 female, who was examined for the presence of a vaginal plug the following morning. When a plug was observed, the female was housed separately. For male mice who did not produce pups after 3 months (~3 cycles), the original female was replaced with a new female and the fertility test continued.
Testis histology, sperm count, and sperm morphology
Mouse testes were fixed in Bouin’s solution overnight, washed with 70% ethanol, embedded in paraffin, and sectioned at 5 μm thickness. Sections were stained with hematoxylin solution, countered stained with eosin solution, and imaged using Leica DMi8 brightfield microscope equipped with an 20× 0.4 N.A. objective (HC PL FL L 20×/0.40 CORR PH1, Leica Microbiosystems, Buffalo Grove, IL). To quantify sperm abundance, the cauda epididymides were collected from mice and placed in phosphate-buffered saline (PBS) containing 4% (w/v) bovine serum albumin. A few incisions were made in the epididymides with scissors to release the sperm, followed by incubation at 37°C and 5% CO2 for 20 min. A 20 μl aliquot of sperm suspension was diluted in 480 μl of 1% (w/v) paraformaldehyde (PFA), and sperm cells counted at 10× by brightfield microscopy. To assess sperm morphology, caudal epididymal sperm were fixed in 1% (w/v) PFA, stained with trypan blue, and a Leica DMi8 brightfield microscope equipped with an 63× 1.4 N.A. oil immersion objective (HC PL APO; Leica Microbiosystems, Buffalo Grove, IL). Sperm stained with Alexa 488-conjugated PNA (see below) were also used to assess sperm morphology.
Meiotic chromosome spreads
Meiotic chromosome spreads were prepared as described (Holloway et al., 2014). Mouse testes were incubated in hypotonic buffer (30 mM Tris-Cl, pH 8.2, 50 mM sucrose, 17 mM sodium citrate, 5 mM EDTA, 0.5 mM DTT) for 30 min on ice, then small fragments of seminiferous tubules were moved to 100 mM sucrose solution and pulled apart with forceps to release germ cells. A drop of sucrose solution containing germ cells was pipetted onto a glass slide with a thin layer of 1× PBS containing 1% PFA and 0.15% (v/v) Triton-X100 (pH 9.2) and spread by swirling. Slides were placed in a humidifying chamber for 2.5 h, air-dried, and washed twice with 1× PBS with 0.4% Photo-Flo 200 (Kodak, Rochester, NY) and once with water with 0.4% Photo-Flo 200, and air-dried. For immunostaining of meiotic chromosomes, slides were sequentially washed with (1) 1× PBS with 0.4% Photo-Flo 200, (2) 1× PBS containing 0.1% (v/v) Triton-X, and (3) blocked with PBS containing 3% (w/v) BSA, 0.05% (v/v) Triton X-100, and 10% (v/v) goat serum in 1× PBS at room temperature. The slides were then incubated with primary antibodies, anti-SCP1 (1:1000 dilution) and anti-SCP3 (1:1000 dilution), in a humidifying chamber overnight at room temperature. Washing and blocking steps were repeated the next day, and the slides were incubated with Alexa 488- or Alexa 594-conjugated secondary antibodies (1:10,000 dilution) for 1 h at room temperature. Slides were washed thrice with 1× PBS containing 0.4% (v/v) Photo-Flo 200, once with water containing 0.4% Photo-Flo 200 mixture, air-dried in the dark, mounted by incubation in ProLong Gold Antifade Mountant with DAPI (4′,6′-diamidino-2-phenylindole; Thermo Fisher Scientific, Waltham, MA) overnight in the dark, and imaged using a Leica DMi8 fluorescence microscope equipped with an 63× 1.4 N.A. oil immersion objective (HC PL APO; Leica Microbiosystems, Buffalo Grove, IL).
Cell sorting by FACS
Testicular cell sorting was performed as described (Cole et al., 2014). Testes were collected, decapsulated, and incubated in 0.4 mg/ml collagenase type IV (Worthington LS004188) in 1× Grey’s Balanced Salt Solution (GBSS, Sigma, G9779) at 33°C rotating at 150 rpm for 15 min. Separated seminiferous tubules were washed with 1× GBSS and incubated in 0.5 mg/ml Trypsin and 1 μg/ml DNase I in 1× GBSS at 33°C rotated at 150 rpm for 15 min. Tubules were dissociated on ice by gentle pipetting, and then 7.5% (v/v) fetal bovine serum (f.c.) was added to inactivate trypsin. The cell suspension was filtered through a pre-wetted 70 μm cell strainer, and cells pelleted at 300 × g for 10 min at 4°C. Cells were resuspended in 1× GBSS containing 5% (v/v) FBS, 1 μg/ml DNase I, and 5 μg/ml Hoechst 33342 (Thermo Fisher Scientific, Waltham, MA) and rotated at 150 rpm at 33°C for 45 min. Propidium iodide (0.2 μg/ml, f.c.; Thermo Fisher Scientific, Waltham, MA) was added, and cells strained through a pre-wetted 40 μm cell strainer. Cell sorting was performed on a FACSAria II (BD Biosciences, Franklin Lakes, NJ). The purity of sorted fractions was assessed by immunostaining. Secondary spermatocyte and spermatid populations were >90% pure, and the pachytene spermatocytes and diplotene spermatocytes were >80% pure.
In vitro fertilization (IVF) and embryo transfer
In vitro fertilization was performed as previously described (Nagy et al., 2003) using spermatozoa from caudal epididymis of either C57BL/6, pi6+/em1, or pi6em1/em1 mice. Spermatozoa were incubated in human tubal fluid (HTF; 101.6 mM NaCl, 4.69 mM KCl, 0.37mM KH2PO4, 0.2 mM MgSO47H2O, 21.4 mM Na-lactate, 0.33 mM Na-pyruvate, 2.78 mM glucose, 25 mM NaHCO3, 2.04 mM CaCl2-2H2O, 0.075 mg/ml Penicillin-G, 0.05 mg/ml streptomycin sulfate, 0.02% (v/v) phenol red, 4 mg/ml BSA) with oocytes (98–146 for control sperm and 120–293 for pi6em1/em1 sperm) from B6SJLF1/J mice for 3–4 h at 37°C with constant 5% O2, 90% N2, and 5% CO2 concentration. Oocyte viability and the presence of pronuclei were assessed under a Nikon SMZ-2B (Nikon, Tokyo, Japan) dissecting microscope. To observe embryo development, embryos were moved into potassium-supplemented simplex optimized media (KSOM; 95 mM NaCl, 2.5 mM KCl, 0.35 mM KH2PO4, 0.2 mM MgSO4-7H2O, 10 mM Na-lactate, 0.2 mM Na-pyruvate, 0.2 mM glucose, 25 mM NaHCO3, 1.71 mM CaCl2-2H2O, 1 mM L-glutamine, 0.01 mM EDTA, 0.075 mg/ml Penicillin-G, 0.05 mg/ml streptomycin sulfate, 0.02% (v/v) phenol red, 1 mg/ml BSA; Millipore Sigma, Burlington, MA) after IVF and assessed every 24 h. To measure birth rates, two-cell embryos were transferred to Swiss Webster pseudopregnant females, and fetuses isolated by cesarean section 18.5 d after embryo transfer.
For zona-free IVF, the zona pellucida of oocytes was removed with acid Tyrode’s solution as described (Yanagimachi et al., 1976; Johnson et al., 1991).
Intracytoplasmic sperm injection (ICSI)
Frozen caudal epididymal spermatozoa were thawed, the sperm tails detached (Nagy et al., 2003), and individual pi6+/em1 or pi6em1/em1 sperm heads injected into B6D2F1/J oocytes in Chatot-Ziomek-Bavister media (CZB; 81.62 mM NaCl, 4.83 mM KCl, 1.18 mM KH2PO4, 1.18 mM MgSO4VH2O, 25 mM Na2HCO3, 1.70 mM CaCl2-2H2O, 0.11 mM Na2-ETDA-2H2O, 1 mM L-glutamine, 28 mM Na-lactate, 0.27 mM Na-pyruvate, 5.55 mM glucose, Penicillin-G 0.05 mg/ml, 0.07 mg/ml streptomycin sulfate, 4 mg/ml BSA) (Millipore Sigma, Burlington, MA) using the PiezoXpert (Eppendorf, Hamburg, Germany; Cat#5194000024). Surviving oocytes were counted, collected, and cultured in KSOM (Millipore Sigma, Burlington, MA) at 37°C and 5% CO2 for 24 h. Two-cell embryos were surgically transferred unilaterally into the oviducts of pseudopregnant Swiss Webster females. At 16.5 days after the surgery, live fetus isolated by cesarean section.
Sperm motility
Cauda epidydimal sperm were collected from mice and placed in 37°C HTF media in an incubator with 5% CO2. A drop of sperm was removed from the suspension and pipetted into a sperm counting glass chamber, then assayed by CASA or video acquisition. CASA was conducted using an IVOS II instrument (Hamilton Thorne, Beverly, MA) with the following settings: 100 frames acquired at 60 Hz; minimal contrast = 50; 4 pixel minimal cell size; minimal static contrast = 5; 0%straightness (STR) threshold; 10 μm/s VAP Cutoff; prog. min VAP, 20 μm/s; 10 μm/s VSL Cutoff; 5 pixel cell size; cell intensity = 90; static head size = 0.30–2.69; static head intensity = 0.10–1.75; static elongation = 10–94; slow cells motile = yes; 0.68 magnification; LED illumination intensity = 3000; IDENT illumination intensity = 3603; 37°C. Agglutination of pi6em1/em1 sperm prevented CASA measurements at later times. A Nikon Diaphot 200 microscope (Nikon, Tokyo, Japan) with darkfield optics equipped with Nikon E Plan 10×/0.25 160/-Ph1 DL objective (Nikon, Tokyo, Japan), ZWO ASI 174mm Monochrome CMOS Imaging camera (ZWO, SuZhou, China), and the SharpCap software (https://docs.sharpcap.co.uk/2.9/) using darkfield at 10× magnification were used to record sperm movement at 37°C.
In vitro acrosome reaction assay
Acrosome reaction was assessed as described (Talbot et al., 1976). Cauda epididymides were collected from mice, placed in HTF media pre-warmed for at least 2 h in a 37°C incubator at 5% CO2. A few incisions were made in the epididymides with scissors to release the sperm, followed by incubation at 37°C in 5% CO2 for 90 min. Calcium ionophore A23187 (10 μm f.c. in DMSO) was added, and incubation continued for 30 min. Sperm were fixed at room temperature for 10 min by adding two volumes of 4% (w/v) PFA, pelleting at 1,000 × g for 5 min, washed with 1× PBS, resuspended in fresh 1× PBS, spotted on a glass slide, and air-dried. Methanol was pipetted onto the sperm to permeabilize the cells, followed by washing with 1× PBS. Slides were incubated overnight in 10 μg/ml Alexa Fluor 488-conjugated peanut agglutinin (PNA) in 1× PBS (Mortimer D., 1987), washed with 1× PBS, air-dried, and mounted with ProLong Gold Antifade Mountant with DAPI (Thermo Fisher Scientific, Waltham, MA). Sperm were imaged using a Leica DMi8 fluorescence microscope equipped with a 63× 1.4 N.A. oil immersion objective (HC PL APO; Leica Microbiosystems, Buffalo Grove, IL) and analyzed using ImageJ (version 2.0.0-rc-68/1.52e; https://fiji.sc/).
Chromatin Immunoprecipitation (ChIP) and sequencing
Frozen testes were cross-linked with 2% (w/v) formaldehyde at room temperature for 30 min using an end-over-end tumbler. Fixed tissues were homogenized in buffer containing 1% (w/v) sodium lauryl sulfate (SDS), 10mM EDTA, and 50mM Tris-HCl (pH 8.1) by 40 strokes in a Dounce tissue grinder with Pestle B (Kimble-Chase, Rockwood, TN). Lysed samples were sonicated using the E220 Covaris ultrasonicator (Covaris, Woburn, MA) to shear the chromatin to 150–200 bp fragments and diluted 1:10 with a buffer containing 0.01% (w/v) SDS, 1.1% (v/v) Triton X-100, 1.2 mM EDTA, 16.7 mM Tris-HCl (pH 8.1), 167 mM NaCl. Immunoprecipitation was performed using 5.5 μg of rabbit anti-A-MYB antibody (Sigma, St. Louis, MO), DNA was extracted with phenol:chloroform:isoamyl alcohol (25:24:1) (pH 8), and ChIP-seq libraries were prepared as previously described (Li et al., 2013). Libraries were sequenced using paired-end reading on NextSeq500 (Illumina, San Diego, CA), and reads were mapped to mouse genome assembly mm10 using Bowtie2 (v2.2.5). ChIP-seq peaks were determined using MACS2 (v2.1.1) and unique mapping reads were reported in this study as fold enrichment over input.
RNA-seq and small RNA-seq
Small RNA-seq and RNA-seq libraries were constructed and sequenced using NextSeq 500 (Illumina, San Diego, CA) as described (Fu et al., 2018). To sequence mature piRNAs, small RNA was oxidized with 25 mM NaIO4 in 30 mM sodium borate, 30 mM boric acid (pH 8.6; Sigma Aldrich, St. Louis, MO) at 25°C for 30 min. RNA was precipitated with ethanol before adapter ligation. Small RNA-seq and RNA-seq reads were mapped to mouse genome assembly mm10 using piPipes (Han et al., 2015).
Transcript abundance between pi6+/em1 and C57BL/6 testes were indistinguishable (< 2fold change and FDR > 0.05). Transcripts with low abundance (< 1 fpkm) in both C57BL/6 and pi6em1/em1 cells were excluded.
Transposon mapping
RNA-seq reads were intersected using BEDtools (Quinlan and Hall, 2010) with Repeat Masker annotation from UCSC (downloaded from https://genome.ucsc.edu/cgi-bin/hgTables). Reads mapping to multiple genomic locations were apportioned. Reads for individual repeats were aggregated to obtain reads counts for repeat families.
Statistics
All statistics were performed using R (https://www.rstudio.com/) and graphs were generated using Igor Pro v7.08 (WaveMetrics) or ggplot2 v3.0.0 (https://ggplot2.tidyverse.org/). Unless otherwise stated, Mann-Whitney-Wilcoxon test was used to calculate p values.
ACCESSION NUMBERS
All sequencing data are available through the NCBI Sequence Read Archive using accession number PRJNA480354.
Movies S1-10. pi6em1/em1 sperm motility
Movie S1. C57BL/6 sperm motility at 10 minute time point.
Movie S2. pi6em1/em1 sperm motility at 10 minute time point.
Movie S3. C57BL/6 sperm motility at 90 minute time point.
Movie S4. pi6em1/em1 sperm motility at 90 minute time point.
Movie S5. C57BL/6 sperm motility at 3 hour time point.
Movie S6. pi6em1/em1 sperm motility at 3 hour time point
Movie S7. C57BL/6 sperm motility at 4 hour time point.
Movie S8. pi6em1/em1 sperm motility at 4 hour time point.
Movie S9. C57BL/6 sperm motility at 5 hour time point.
Movie S10. pi6em1/em1 sperm motility at 5 hour time point.
Footnotes
↵5 Lead contact