|Home | About | Journals | Submit | Contact Us | Français|
Centrioles are lost during oogenesis and inherited from the sperm at fertilization. In the zygote, the centrioles recruit pericentriolar proteins from the egg to form a mature centrosome that nucleates a sperm aster. The sperm aster then captures the female pronucleus to join the maternal and paternal genomes. Because fertilization occurs before completion of female meiosis, some mechanism must prevent capture of the meiotic spindle by the sperm aster. Here we show that in wild-type Caenorhabditis elegans zygotes, maternal pericentriolar proteins are not recruited to the sperm centrioles until after completion of meiosis. Depletion of kinesin-1 heavy chain or its binding partner resulted in premature centrosome maturation during meiosis and growth of a sperm aster that could capture the oocyte meiotic spindle. Kinesin prevents recruitment of pericentriolar proteins by coating the sperm DNA and centrioles and thus prevents triploidy by a non-motor mechanism.
Centrioles play essential roles in controlling the number of spindle poles during mitosis and in assembly of cillia and are thus subject to strict number control (Leidel and Gönczy, 2005). In Caenorhabditis elegans, centrioles are eliminated early during oogenesis and reintroduced by sperm during fertilization. This model is supported by electron microscopy showing an absence of centrioles at the poles of oocyte meiotic spindles (Albertson and Thomson, 1993) and the presence of centrioles in mature sperm (Ward et al., 1981) and at the male pronucleus of the zygote (Pelletier et al., 2006). In addition, GFP fusions to centriolar and pericentriolar proteins label distinct foci associated with nuclei in the distal gonad (early oogenesis), no foci in the proximal gonad (later oocytes) (Zhou et al., 2009), then again label discrete foci in fertilized zygotes (Pelletier et al., 2006). Failure to eliminate maternal centrioles has been reported to result in multipolar mitotic spindles in the C. elegans zygote (Kim and Roy, 2006) and fertilization by more than one sperm results in multipolar mitotic spindles in several species (Harris et al., 1980; Navara et al., 1994).
Sperm-derived centrioles recruit centriolar (Pelletier et al., 2006) and pericentriolar (Hamill et al., 2002) proteins from the egg cytoplasm to form a mature centrosome that nucleates a sperm aster. Recruitment of maternal pericentriolar proteins to sperm centrioles has been reconstituted in Xenopus egg extracts and occurs in as little as 7 min (Stearns and Kirschner, 1994). The radial array of microtubules that constitutes the sperm aster captures the female pronucleus after meiosis to bring maternal and paternal chromosomes into close proximity before formation of the first mitotic spindle (Oegema and Hyman, 2006). The fact that fertilization occurs during female meiosis in most animal species (prometaphase I in C. elegans) necessitates mechanisms to prevent interference with the female meiotic spindle by the sperm aster. Little is known about these mechanisms.
Earlier studies have shown that in Urechis (Gould and Stephano, 1999) and Crassostrea (Stephano and Gould, 2000), sperm aster formation is completely suppressed until after completion of meiosis. In contrast, the size of the sperm aster is limited during meiosis in starfish zygotes (Stephano and Gould, 2000). In Spisula, maternal and paternal centrioles share the same cytoplasm during female meiosis; however, the ability of the paternal centrioles to form a sperm aster is suppressed during meiotic metaphase I when maternal centrioles drive assembly of two microtubule asters utilized in meiotic chromosome segregation (Wu and Palazzo, 1999). Wu and Palazzo suggested that failure to suppress the sperm aster during metaphase I might result in chromosome misegregation by a tripolar meiotic spindle. The mechanism that prevents sperm aster formation during metaphase I in Spisula has not been addressed and the timing of sperm aster formation during female meiosis has not been addressed in most species.
In this study, we demonstrate that recruitment of maternal pericentriolar proteins to sperm-derived centrioles is completely suppressed from prometaphase I through anaphase II in C. elegans zygotes. We show that this suppression is mediated by kinesin-1 and a kinesin-binding protein and that these proteins form a shell surrounding the sperm DNA/centriole complex in the zygote. We also demonstrate that premature sperm asters can capture female meiotic spindles and prevent polar body extrusion.
In wild-type C. elegans zygotes, the pericentriolar proteins SPD-5 and γ-tubulin are not recruited to the sperm centrioles until after the second polar body has been extruded and the male pronucleus has formed and moved to the cortex (see Figure S1 and Table S1 available online). Thus, some mechanism must prevent recruitment of SPD-5 and γ-tubulin to the sperm-derived centrioles from the time of fertilization at prometaphase I through completion of meiosis II, a period lasting 35 min (McNally and McNally, 2005). In striking contrast, both SPD-5 and γ-tubulin are recruited to the sperm centrioles as early as anaphase I after depletion of the kinesin-1 heavy chain, UNC-116, or the kinesin cargo adaptor, KCA-1 (Figure 1; Table S1). In addition, sperm asters consisting of long microtubules grew from the premature centrosomes in kca-1(RNAi) meiotic embryos (Figure 1) but not in wild-type meiotic embryos. The GFP::γ-tubulin recruited to sperm centrosomes in kca-1(RNAi) embryos (Table S1) was expressed in oocytes but not in sperm before fertilization. Identical results were obtained with GFP::TAC-1 that was expressed in oocytes but not in sperm (data not shown). Thus, the premature centrosome maturation observed involves recruitment of maternal pericentriolar proteins to the sperm centrioles. Because worms were exposed to RNAi treatment at the L4 stage when spermatogenesis is already complete, premature centrosome maturation must be caused by depletion of UNC-116 or KCA-1 from the egg cytoplasm rather than defects in spermatogenesis. Consistent with this interpretation, we previously showed that fertilization by wild-type sperm does not decrease the embryonic lethality of unc-116(RNAi) or kca-1(RNAi) (Yang et al., 2005).
In an attempt to expand the list of gene products involved in suppression of sperm centrosome maturation, we examined a number of candidate proteins. The breast cancer associated ubiquitin ligase BRCA-1/BRD-1 inhibits centrosome maturation in mitotic mammalian cells (Parvin, 2009) but loss-of-function mutations in these genes did not cause premature centrosome maturation during C. elegans meiosis (Table S1). The conserved protein, SZY-20, suppresses sperm centrosome maturation in C. elegans mitotic embryos (Song et al., 2008) but a szy-20(ts) mutant did not cause premature centrosome maturation during meiosis (Table S1). MAP kinase has been implicated in suppression of sperm aster formation during meiosis in marine invertebrates (Gould and Stephano, 1999; Stephano and Gould, 2000). However, we observed no SPD-5 staining at the sperm centrioles in meiotic embryos dissected from mpk-1(ga111ts) (Lee et al., 2007) worms after a 24 hr shift to nonpermissive temperature (Table S1). KCA-1 was reported to interact with the heterochromatin protein, HPL-2 (Li et al., 2004), which binds to nucleosomes containing trimethyl K9 histone H3. Although the sperm DNA stained with an anti-trimethyl K9 H3 antibody in 20/20 meiotic embryos, we did not observe premature centrosome maturation in hpl-2(tm1489) (Schott et al., 2006) worms (Table S1). A loss-of-function mutation in unc-14 (Table S1), a Caenorhabditis-specific regulator of kinesin-1 in the nervous system (Sakamoto et al., 2005), did not result in SPD-5 accumulation at sperm centrioles. Loss-of-function mutations in the conserved kinesin-1 regulators UNC-16 (Sakamoto et al., 2005), UNC-76 (Toda et al., 2008), or UNC-33 (Tsuboi et al., 2005) caused accumulation of a very small focus of SPD-5 in a small fraction of meiotic embryos but these were never associated with a microtubule aster (Table S1). These results support a specific role for kinesin-1 heavy chain and KCA-1 in suppression of sperm centrosome maturation.
To test whether a complex of kinesin-1 and KCA-1 might act directly at the sperm centrioles during female meiosis, we carefully examined the localization of the kinesin-1 heavy chain, UNC-116, the kinesin-1 light chain, KLC-1, and KCA-1 in meiotic embryos. Anti-KCA-1 antibody labeled an irregular ring (average maximal diameter 5.18 ± 0.17 µm, n = 22) surrounding the sperm DNA (Figures 2B and 2D). This ring was absent from kca-1(RNAi) meiotic embryos (Figure S2E; Figure 2Q) and an irregular ring of similar diameter (5.8 ± 0.33 µm, n = 5) was labeled in living mCherry::KCA-1 meiotic embryos in utero (Figure 2E). Yolk granules are transported away from the cortex and packed in a central mass in meiotic embryos (McNally et al., 2010). In most meiotic embryos, the sperm DNA was outside the zone of centrally packed yolk granules (Figure S2F). In those embryos where the sperm DNA was within the packed mass of yolk granules, yolk granules were excluded from an irregular zone around the sperm DNA (Figures 2C and 2P) with an average maximal diameter (5.24 ± 0.19 µm, n = 20) similar to that of the KCA-1 ring. Antibodies specific for the kinesin-1 heavy chain, UNC-116, labeled a similar ring around the sperm DNA (Figures 2H and 2I) with a diameter (4.52 ± 0.12 µm, n = 20) slightly smaller than that of KCA-1. No anti-UNC-116 staining around the sperm DNA was observed in unc-116(RNAi) embryos (Figure 2Q; Figure S2D). Antibodies specific for the kinesin-1 light chain, KLC-1, also labeled a ring surrounding the sperm DNA in meiotic embryos (Figures 2K and 2M). This labeling was absent in klc-1(RNAi) meiotic embryos (Figure 2Q) and an identical ring was observed in living mCherry::KLC-1 meiotic embryos in utero (Figure 2L). The KLC-1 ring had a slightly smaller diameter than the KCA-1 ring (anti-KLC-1: 4.06 ± 0.12 µm, n = 14; mCherry::KLC-1: 4.63 ± 0.15 µm, n = 7). The rings of UNC-116, KLC-1, and KCA-1 around the sperm DNA in meiotic embryos were reminiscent of SPD-2, which is found in a ring around the DNA of mature sperm before fertilization and in the centrioles of mature sperm (Pelletier et al., 2004). SPD-2 antibodies labeled a much smaller ring around the sperm DNA in meiotic embryos (2.20 ± 0.05 µm, n = 12) and the centrioles (Figure 2O). Because identical SPD-2 localization has been previously reported in mature sperm (Pelletier et al., 2004), we infer that the SPD-2 in this ring is paternally derived.
The smaller diameter of the SPD-2 around the sperm DNA suggested that paternal SPD-2 might be required to recruit maternal KCA-1 to the sperm DNA and centrioles. Because SPD-2 is required for spermatogenesis and RNAi is not effective on mature sperm, we utilized a fast-acting ts allele of spd-2 (O’Rourke et al., 2011) to inactivate SPD-2 just before fertilization. Anti-KCA-1 staining of meiotic embryos dissected from spd-2(or293) worms raised at the permissive temperature of 16°C revealed rings of KCA-1 around the sperm DNA identical to those seen in wild-type. When spd-2(ts) worms were shifted to the nonpermissive temperature of 25°C for 6 hr, the intensity of anti-KCA-1 staining around the sperm DNA decreased. The enrichment of KCA-1 around the sperm DNA was quantified as the ratio of fluorescence intensity adjacent to the sperm DNA divided by the fluorescence intensity of the adjacent cytoplasm and expressed as percent. At permissive temperature, KCA-1 was enriched around the sperm DNA by 209.9 ± 9.6% (n = 24) or 2-fold relative to the adjacent cytoplasm. At nonpermissive temperature this enrichment was reduced to 139.5 ± 5.2% (n = 30) or 1.4-fold. The results of Student’s t test indicated that these numbers are significantly different (p < 0.0001) indicating that SPD-2 is required to recruit KCA-1 to the sperm DNA and centrioles. SPD-2 is likely only partially inactivated by this ts allele as SPD-2 antibody staining at the sperm DNA and centrioles of meiotic embryos appeared wild-type; thus, only partial reduction in the KCA-1 ring would be expected. Our results are consistent with a model in which KCA-1 and kinesin-1 are recruited to the sperm DNA and centrioles by paternal SPD-2. SPD-2 is required for recruitment of SPD-5 and γ-tubulin to centrioles (Pelletier et al., 2004; Kemp et al., 2004), therefore we propose that kinesin-1 and KCA-1 are components of a structure that might prevent interaction between maternal SPD-5 and paternal SPD-2 and thus prevent centrosome maturation.
In addition to premature centrosome maturation, the position of the sperm DNA/centriole complex during meiosis was dramatically different in kca-1(RNAi) embryos than in wild-type embryos. In wild-type meiotic embryos, the sperm DNA was found deep in the cytoplasm, usually in the future posterior quadrant of the embryo from metaphase I through anaphase II (Figures 1A, 2A–2C, and and3A;3A; Figures S1A, S1B, and S1I). The male pronucleus moved into close proximity to the future posterior cortex around the time of centrosome maturation (Figures S1F and S1I, 22 min) as previously reported (Rappleye et al., 2002). In contrast, the sperm DNA was observed in close contact with the future posterior cortex at all stages of meiosis in fixed kca-1(RNAi) embryos (Figures 1B and and3A).3A). One interpretation of these results is that in wild-type embryos, the sperm DNA/centriole complex is transported inward from the site of fertilization in a KCA-1-dependent manner and that the sperm DNA/centriole complex moves back to the cortex after meiosis. In this scenario, the sperm DNA remains at the site of fertilization in kca-1(RNAi) embryos. Alternatively, the sperm DNA might move inward from the site of fertilization in a KCA-1-independent manner but move back to the cortex prematurely in kca-1(RNAi) embryos, possibly due to premature sperm aster formation.
If the sperm DNA moved inward normally in kca-1(RNAi) embryos but moved back to the cortex prematurely, then the sperm DNA would be expected to be further from the cortex early in meiosis than late in meiosis. The data in Figure 3A, however, indicate that the distance of the sperm DNA from the cortex does not vary between metaphase I and metaphase II. If premature centrosome maturation caused premature return of the internalized sperm to the cortex, then sperm DNA with SPD-5-positive centrosomes should be close to the cortex while sperm DNA without SPD-5-positive centrosomes in kca-1(RNAi) embryos should be as far from the cortex as in wild-type embryos. As shown in Figure 3B, sperm DNA without SPD-5-positive centrosomes was still much closer to the cortex than in wild-type. Sperm DNA with SPD-5-positive centrosomes was closer to the cortex than SPD-5-negative sperm DNA in kca-1(RNAi) embryos (Figure 3B), possibly indicating that sperm asters can stabilize the position of the sperm DNA at the cortex. In addition, kca-1(RNAi) decreased the distance between sperm and cortex in spd-2(ts) embryos which should be defective in centrosome maturation and aster formation (Figure 3B). These data support a model in which the sperm DNA/centriole complex moves inward from the site of fertilization by a KCA-1-dependent mechanism.
We previously presented evidence that yolk granules are transported inward, away from the cortex by kinesin-1/KCA-1 on cytoplasmic microtubules that extend diagonally inward from the cortex (McNally et al., 2010). If the sperm DNA/centriole complex were also transported inward as direct cargo of kinesin-1/KCA-1, this transport should be blocked when kinesin-1 heavy chain, UNC-116, is depleted or when cytoplasmic microtubules are depleted by tubulin (RNAi). Indeed, the sperm DNA remained close to the cortex in unc-116(RNAi), tubulin (RNAi) or unc-116(f130) meiotic embryos (Figure 3C). We also found that very early in wild-type meiosis, sperm DNA without kinesin-1 or KCA-1 were closer to the cortex than sperm DNA with a kinesin-1/KCA-1 shell (Figure 2Q). This suggests that a shell of maternal kinesin-1/KCA-1 must assemble around the sperm DNA/centriole complex before it is transported inward on microtubules.
Premature centrosome maturation did not occur in tubulin(RNAi) or unc-116(f130) meiotic embryos (Figures 3D and 3F; Figure S3A). Cytoplasmic microtubules were absent or greatly reduced in tubulin(RNAi) meiotic embryos (Figure 3E) and maternal chromosomes were in a mass in the center of the meiotic embryo as previously reported (Yang et al., 2003), presumably because both kinesin-dependent (Yang et al., 2005) and dynein-dependent (Ellefson and McNally, 2009, 2011) meiotic spindle positioning mechanisms require cytoplasmic microtubules. Importantly, tubulin(RNAi) did not prevent centrosome maturation in postmeiotic embryos (Figure S3). Thus, kinesin-dependent suppression of centrosome maturation does not appear to require kinesin-dependent transport on microtubules. The unc-116(f130) mutant, which has a single amino acid substitution in kinesin heavy chain, causes a severe meiotic spindle positioning defect (Figure 3F) indicating that it lacks transport function. Thus, the ability of unc-116(f130) meiotic embryos to suppress centrosome maturation during meiosis (Figure 3F) further supports the hypothesis that microtubule-based transport is not required for suppression of centrosome maturation during meiosis. The findings that the sperm DNA is closely opposed to the plasma membrane in unc-116(f130) embryos and tubulin(RNAi) embryos (Figure 3C) but centrosomes fail to mature, rule out the possibilities that premature centrosome maturation is caused by cortical contact of the sperm centrioles or that inward transport of the sperm centrioles suppresses maturation. Both UNC-116 (Figure 3H) and KCA-1 (Figure 3G) were still present in a ring around the sperm DNA during meiosis in unc-116(f130) embryos but not in unc-116(RNAi) embryos (Figure 3I). These results further support a model in which the ring of kinesin-1/KCA-1 surrounding the sperm DNA/centriole complex directly inhibits recruitment of pericentriolar proteins to the sperm centrioles.
To test whether the sperm asters that form in kca-1(RNAi) meiotic embryos are capable of capturing the female meiotic spindle, we monitored microtubules and chromosomes by time-lapse imaging in GFP:tubulin, mCherry:histone, kca-1(RNAi) meiotic embryos in utero. In 6/24 time-lapse sequences, the sperm asters captured the meiosis II spindle (Figure 4A). Following three of the six spindle-capture events, three pronuclei (rather than two) were detected in the embryos by DIC microscopy. The other three examples of spindle capture resulted in arrested embryos. This result indicates that capture of the meiosis II spindle by a premature sperm aster can lead to triploidy by preventing extrusion of the second polar body.
We hypothesized that the low frequency of spindle capture in time-lapse sequences of kca-1(RNAi) embryos might be because the sperm derived centrosomes are too far away from the meiotic spindle in the majority of embryos. In all six examples of spindle capture, the sperm asters were initially stationary at the future posterior cortex, consistent with sperm position in fixed kca-1(RNAi) embryos (Figure 3). In each example of spindle capture, the meiosis II spindle initiated directed migration toward the sperm asters only after the sperm asters moved inward to within an average distance of 15.7 ± 0.8 µm (n = 5) of the spindle. In kca-1(RNAi) time-lapse sequences where spindle capture did not occur, the sperm aster remained close to the future posterior cortex and the closest distance between spindle and sperm aster was 23.5 ± 2.0 µm (n = 14). These results suggest that spindle capture is rare in kca-1(RNAi) embryos because the sperm aster is stationary at the future posterior end of the embryo where it is beyond a threshold capture distance of 17 µm. All six spindle capture events occurred at times indicating that the cell cycle was being delayed by photodamage relative to previous measurements of cell-cycle timing in kca-1(RNAi) and wild-type embryos (Yang et al., 2005). If photodamage-induced cell-cycle delay causes dissociation of the sperm aster from the future posterior cortex, then spindle capture events should be rare among fixed kca-1(RNAi) embryos. As expected, only one spindle capture event was observed among 42 fixed kca-1(RNAi) embryos (Figure 4B).
To further test the hypothesis that spindle capture was promoted by a prolonged time in meiosis, we monitored meiotic spindles in embryos arrested in metaphase I by depletion of an anaphase promoting complex subunit with a mat-2(ts) mutant. We analyzed fixed metaphase I embryos from mat-2(ts);kca-1(RNAi) worms that correspond to an arrest period of 0–92 min (see Experimental Procedures). As shown in Figures 5A and 5B, spindle capture events occurred in 22% of metaphase-arrested, kca-1(RNAi) embryos. This result confirms that a delay in meiosis allows time for spindle capture, possibly because the sperm aster dissociates from the posterior cortex after an extended period in meiosis.
SPD-5 was also recruited to sperm centrioles in the majority of mat-2(ts), metaphase-arrested embryos that were otherwise wild-type (Figures 5A and 5B). The majority of SPD-5-positive sperm centrosomes in mat-2(ts) embryos did not have an associated sperm aster, indicating a partial derepression of sperm centrosome maturation. In two time-lapse sequences of GFP:tubulin in metaphase-arrested mat-2(ts) embryos filmed in utero, sperm asters formed at 25 and 30 min after ovulation. These times are close to the times that wild-type meiotic embryos would complete meiosis and mature their centrosomes. These observations led to the hypothesis that KCA-1 might be degraded at the end of meiosis in wild-type embryos and that this degradation might still occur in APC-depleted embryos.
To test whether degradation of KCA-1 occurs at the end of meiosis to allow wild-type centrosome maturation, we captured spinning disk confocal images of mCherry:KCA-1 in living embryos in utero (Figure 5C). Embryos in the +1 position (adjacent to the spermatheca) revealed KCA-1 packed in the central cytoplasm. This pattern matched that observed in fixed metaphase I embryos by anti-KCA-1 immunofluorescence (Figure 2A; McNally et al., 2010). The cytoplasmic fluorescence intensity of mCherry:KCA-1 dropped in the +2 and +3 embryos (Figures 5C and 5D) suggesting proteolysis after meiosis. This proteolysis was specific for KCA-1 as GFP:calmodulin (CMD-1) showed no decrease in cytoplasmic fluorescence intensity in the +2 relative to the +1 embryo (Figures 5C and 5D). In worms depleted of the APC subunit, MAT-1, by RNAi, the +1 embryo exhibited a wild-type metaphase I packing of cytoplasmic mCherry:KCA-1 (Figure 5C). However, the fluorescence intensity dropped in the +2 and +3 embryos even though they were arrested at metaphase I. The small amount of KCA-1 remaining in +2 and +3 embryos of APC-depleted embryos was sequestered in structures that were never observed in wild-type (Figure 5C). In contrast, degradation of GFP:cyclin B1 was completely blocked by mat-1(RNAi) (Figures 5C and 5D). This result strongly suggests that KCA-1 is degraded late in meiosis and that a majority of KCA-1 is degraded in an APC-independent manner. The partial degradation of KCA-1 in APC-depleted embryos explains why partial centrosome maturation occurs in metaphase-arrested, APC-depleted embryos (Figures 5A and 5B). These results suggest that wild-type centrosome maturation may be initiated by KCA-1 degradation late in meiosis. Cowan and Hyman (2006) reported a delay in centrosome maturation in embryos depleted of cyclin E or CDK-2. Consistent with previous findings, we found that only 2/13 cdk-2(RNAi) embryos had a focus of GFP:TAC-1 associated with the male pronucleus in contrast with 9/11 wild-type embryos. However, no enrichment of KCA-1, UNC-116, or KLC-1 was observed around the male pronuclei of cdk-2(RNAi) embryos just as they were absent from the male pronuclei of wild-type embryos (Figures 2G, 2J, 2N, and 2Q). Thus, cyclin E/CDK-2 does not appear to promote sperm centrosome maturation by initiating degradation of KCA-1.
Because fertilization occurs during female meiosis and because sperm-derived centrioles form sperm asters equipped to capture the female pronucleus, all animals must actively prevent sperm asters from interfering with meiotic spindles until after extrusion of the second polar body. C. elegans uses at least two mechanisms to accomplish this: (1) the meiotic spindle and sperm are positioned at opposite ends of the zygote and (2) recruitment of pericentriolar proteins to the sperm centrioles is suppressed so that asters do not form until after extrusion of the second polar body. Kinesin-1 and its binding partner, KCA-1, are involved in both of these mechanisms.
Early in diakinesis, the prophase nucleus or germinal vesicle is centered within the oocyte. The germinal vesicle then migrates away from the end of the oocyte that is closest to the sperm in the spermatheca. This directed migration requires kinesin-1/ KCA-1 (McNally et al., 2010). Fertilization then occurs at the end of the oocyte that first enters the spermatheca (Samuel et al., 2001). This system results in the rough placement of the meiotic spindle and sperm at opposite ends of the oval embryo long before establishment of embryonic polarity, which is induced by the sperm aster after completion of meiosis (Gönczy and Rose, 2005). During meiotic metaphase I and II, the spindle is maintained at the future anterior cortex by kinesin-1/KCA-1 (Yang et al., 2005), whereas during anaphase I and II the spindle is anchored at the cortex by cytoplasmic dynein (Ellefson and McNally, 2009; 2011). Positioning of the sperm DNA during female meiosis is more enigmatic. Kinesin-1/KCA-1 appears to transport the sperm DNA inward from the site of fertilization at the plasma membrane but the sperm DNA is maintained in the future posterior third of the embryo by an unknown mechanism. This migration of the sperm DNA is counterintuitive because it places the sperm DNA closer to the meiotic spindle. In fact, the average minimum distance between sperm DNA and meiotic spindle is not significantly different (p = .16, unpaired t test) between wild-type (27.8 ± 1.7 µm, n = 24) and kca-1(RNAi) (24.2 ± 1.6 µm, n = 15). Even though the meiotic spindle is further from the future anterior cortex in kca-1(RNAi) embryos, the sperm DNA is also closer to the future posterior cortex. Because cortical contact by the sperm aster initiates establishment of embryonic polarity in C. elegans (Gönczy and Rose, 2005), the inward transport of the sperm might be important in preventing premature polarity establishment. Two time-lapse studies of human fertilization reported that the male pronucleus always forms in the center of the zygote (Mio, 2006; Payne et al., 1997). This result indicates that the sperm DNA must be transported inward from the site of fertilization in human zygotes, which do not use the sperm as a cue for establishing embryonic polarity (Cooke et al., 2003). Depending on the sizes of the oocyte, spindle and sperm aster, centering the sperm while anchoring the spindle at the cortex might serve the general purpose of keeping the spindle and sperm apart.
Kinesin-1/KCA-1 is also essential for blocking recruitment of the maternal pericentriolar proteins, SPD-5 and γ-tubulin, to the sperm-derived centrioles. Unlike other kinesin-1/KCA-1-dependent processes like germinal vesicle positioning (Yang et al., 2003; McNally et al., 2010), meiotic spindle positioning (Yang et al., 2003; 2005), and sperm DNA positioning (this study), inhibition of pericentriolar protein recruitment to sperm centrioles does not require cytoplasmic microtubules and therefore cannot involve transport by kinesin on cytoplasmic microtubules. The requirement for the centriolar protein, SPD-2, in assembly of a shell of kinesin-1/KCA-1 around the sperm DNA and centrioles suggests a direct or indirect interaction between kinesin-1/KCA-1 and SPD-2. The shell of kinesin-1/KCA-1 might directly block diffusion of large complexes of pericentriolar material (PCM) from the cytoplasm to the centrioles or act as a scaffold for phosphatases that oppose PCM recruitment. Alternatively kinesin-1/KCA-1 might act more globally on the large cytoplasmic pool of SPD-2. Disruption of this pathway only resulted in spindle capture by the sperm aster in cases where the sperm aster and spindle ended up in close proximity.
The importance of sperm aster suppression in human zygotes is difficult to discern from the literature. The time between sperm-egg fusion and second polar body extrusion after in vitro fertilization is difficult to interpret from published studies (Mio, 2006) in part because of the long and heterogeneous time required for the sperm to traverse the zona pellucida. Intracytoplasmic injection of sperm eliminates these uncertainties and the second polar body is extruded on average 2 hr after injection of human sperm into human oocytes (Mio, 2006; Payne et al., 1997). This would appear to be sufficient time for a sperm aster to form and interfere with anaphase II. The invasive methods required to unambiguously score sperm aster formation are not compatible with assisted reproduction in humans. However, when human sperm is injected into bovine oocytes, sperm aster formation is suppressed during anaphase II and does not occur until decondensation of both male and female pronuclei (Nakamura et al., 2001). It is likely that the use of advanced polarization microscopy (Kilani et al., 2011) will allow noninvasive analysis of sperm aster formation in human zygotes.
The finding that KCA-1 is degraded during meiosis suggests that normal centrosome maturation after meiosis is initiated by regulated proteolysis of KCA-1. Numerous proteins are degraded by different ubiquitin ligases in a highly ordered temporal sequence during or shortly after meiosis in C. elegans. For example, CAV-1 is degraded during meiosis in an endocytosis-dependent manner (Sato et al., 2006). MEI-1 is degraded at the exit from meiosis by a CUL-3-dependent mechanism (Pintard et al., 2003). OMA-1 is degraded after the first mitotic cleavage (Nishi and Lin, 2005). Although each of these proteins is degraded at a different time by a different mechanism, degradation of CAV-1 (Sato et al., 2006), MEI-1 and OMA-1 (Pellettieri et al., 2003) are also blocked in embryos arrested in meiotic metaphase I by depletion of anaphase promoting complex subunits (Pellettieri et al., 2003). The surprising finding that KCA-1 is degraded and that centrosomes mature in metaphase-arrested, APC-depleted embryos indicates that multiple, synchronized but independent clocks control the transition from oocyte to embryo.
In this study, wild-type was N2 Bristol. The following transgenic strains were used: GFP:γ-tubulin/OD44: [pie-1:GFP:tbg-1 + unc-119(+)] (provided by Karen Oegema, UCSD), DE90: oxIs318 [spe-11::mCherry::histone + unc-119(+)] II, ruIs32 [pie-1::GFP::histone H2B + unc-119(+)] III, ddIs6 [tbg-1::GFP + unc-119(+)], dnIs17 [pie-1::GFP::PLC-PH domain + unc-119(+)] (Johnston et al., 2010), GFP:TAC-1/JA1334 (Le Bot et al., 2003), GFP:tubulin + mCherry:histone/FM125 (Ellefson and McNally, 2011), mat-2(ts) + GFP: tubulin/FM157 (Ellefson and McNally, 2011), HR399: unc-36(e251), unc-116(f130) III, FM106: Is[pie-1:mCherry:KLC-1 + unc-119(+)], FM105: Is[pie-1:mCherry:KCA-1 + unc-119(+)], WH280/GFP:calmodulin (Batchelder et al., 2007), FM145: unc-36(e251), unc-116(f130) III, pwls23 [GFP:vit-2], YP170 /RT368: unc-119(ed3) III; pwIs98 [YP170::tdimer2 + unc-119(+)] (provided by Barth Grant, Rutgers University).
All of the RNA interference (RNAi) experiments performed were carried out by feeding bacteria (HT115) induced to express double-stranded RNA corresponding to each gene as described by Timmons et al. (2001). L4 hermaphrodites were transferred to RNAi plates and allowed to feed on the RNAi bacterial lawn for 24–36 hr. All bacterial strains have been described previously (McNally et al., 2010; Ellefson and McNally, 2011). RNAi of unc-116 was carried out on strain CB91 [rol-1(e91)II] as N2 was resistant to RNAi by feeding for this gene.
Briefly, adult hermaphrodites were anesthetized with tricaine/tetramisole and immobilized between a coverslip and agarose pad on a slide as previously described (Ellefson and McNally, 2011). Time-lapse imaging was done on an Olympus IX71 inverted microscope equipped with a 603 PlanApo NA 1.42 objective and a Hamamatsu Orca R2 deep cooled C10600-10B digital mono CCD camera. Excitation light was shuttered with a Sutter shutter controlled by a Sutter Lambda 10-3 controller and MetaMorph Imaging software through a Sutter excitation filter wheel. For DE90 (Figure S1) and FM125 (Figure 4), pairs of single focal plane red and green fluorescence images were captured at 10 s intervals. Manual focus control was used to switch between the spindle and sperm. For imaging mCherry:KCA-1 and mCherry:KLC-1 (Figure 2), trans-genes were first desilenced with mut-16(RNAi), worms were mounted as above and single time point, single focal plane images were acquired with a Perkin Elmer Ultraview spinning disk confocal microscope.
Anti-SPD-5 and anti-SPD-2 rabbit antibodies were provided by Karen Oegema (UCSD). Rabbit antisera for UNC-116, KCA-1, and KLC-1 were described previously (McNally et al., 2010). For this study, Escherichia coli expressed UNC-116, KLC-1, and KCA-1 were immobilized on Affigel 15 beads (Biorad) and antibodies were bound, washed and eluted with 200 mM glycine pH 2.4 then neutralized with Tris 8.0 before use. Anti-γ-tubulin was from Sigma (LL17) and anti-α tubulin was from Sigma (DM1A).
Meiotic embryos were dissected from adult hermaphrodites, freeze cracked and fixed with cold methanol as previously described (Ellefson and McNally, 2011). Complete Z stacks were captured for each meiotic embryo with an Olympus DSU spinning disk confocal using a 603 Plan Apo 1.42 objective and a Hamamatsu Orca R2 CCD. Maximum intensity projections were generated to include the meiotic spindle, polar bodies and sperm DNA/centrosomes.
In living mat-2(ts); GFP:histone worms shifted to nonpermissive temperature, chromosomes are decondensed in the embryos starting at the +5 position from the spermatheca. Since oocytes ovulate on average every 23 min (McCarter et al., 1999), the embryo in the +4 position has been arrested between 69 and 92 min. Because embryos must be dissected from the worm for fixed immunofluorescence, we used the condensation state of the chromosomes to restrict the analysis in Figure 6 to embryos arrested at metaphase I between 0 and 92 min.
We thank Karen Oegema for antibodies, Bruce Bowerman, Kevin O’Connel, and The Caenorhabditis Genetics Cener for strains, and Yuji Kohara for cDNA clones. This work was supported by grant 1R01GM-079421 from the National Institute of General Medical Science to F.J.M.
Supplemental Information includes three figures and one table and can be found with this article online at doi:10.1016/j.devcel.2012.01.010.