Preferentially localized dynein and perinuclear dynactin associate with nuclear pore complex proteins to mediate genomic union during mammalian fertilization

The molecular motor cytoplasmic dynein, together with its cofactor dynactin, transports cargo organelles towards the minus ends of microtubules (Karki and Holzbaur, 1999). Both are large multisubunit complexes that coordinate many events throughout the cell cycle, including interphase trafficking of vesicles from the endoplasmic reticulum (ER) to the Golgi (Presley et al., 1997), mitotic spindle assembly (Vaisberg et al., 1993) and prophase nuclear envelope breakdown (Salina et al., 2002; Beaudouin et al., 2002). Association of dynein and dynactin with the nuclear envelope has been observed in somatic cells (Salina et al., 2002; Busson et al., 1998) and localization of dynein heavy chain to pronuclei has been detected in Caenorhabditis elegans zygotes (Gönczy et al., 1999), but specific nuclear envelope binding partners have not yet been identified.

Pronuclear migration and apposition, the events concluding the fertilization process, are hypothesized to involve dynein and dynactin based upon the directionality of female pronuclear movement and the dynamics of its motion along sperm aster microtubules (Schatten, 1994; Rouvière et al., 1994; Reinsch and Karsenti, 1997; Reinsch and Gönczy, 1998). The sperm aster, a radial array of microtubules nucleated by the sperm centrosome, is essential to unite the female with the male pronucleus in most mammalian zygotes (Navara et al., 1994). Interestingly, rodents do not use a sperm aster during fertilization, precluding their use for investigating sperm-aster-mediated motility in mammals (Schatten, 1994). Bovine and non-human primate zygotes, by contrast, both rely upon sperm asters to mediate genomic union and serve as ideal models for studying the dynamics of human fertilization (Navara et al., 1994; Hewitson et al., 2000). Synthetic nuclei assembled in Xenopus egg extracts have been shown to require dynein to move along centrosome-anchored microtubules towards their minus ends in vitro (Reinsch and Karsenti, 1997), but a role for dynein and dynactin in mediating pronuclear movement has not yet been demonstrated in vivo.

Embedded within the envelopes of somatic nuclei and zygotic pronuclei, nuclear pore complexes (NPCs) are macromolecular assemblies involved in nucleocytoplasmic transport, cell cycle progression and gene expression (Allen et al., 2000; Takizawa and Morgan, 2000; Cronshaw et al., 2002). Recent characterization of mammalian NPCs using mass spectrometry has identified approximately 30 core nucleoporins and additional NPC-associated proteins (Cronshaw et al., 2002). One of these NPC-associated proteins, vimentin, not only corresponds to the 10 nm filaments attached to the ring structure at the nuclear pore cytoplasmic face (Fujitani et al., 1989) but also associates with microtubules, dynein and dynactin (Helfand et al., 2002). Given this evidence of interaction among microtubules, motor proteins, intermediate filaments and NPCs, we asked whether dynein and dynactin associate with pronuclear NPCs and vimentin during mammalian fertilization. Our results identify a novel physical interaction between motor proteins and NPCs, and suggest that dynein and dynactin bind to nucleoporins and vimentin at the cytoplasmic surface of the female pronuclear envelope to mediate pronuclear migration along sperm aster microtubules.

Bovine and rhesus monkey in vitro maturation and in vitro fertilization were carried out according to standard protocols (Sirard et al., 1988; Wolf et al., 1989), with bovine oocytes obtained either from a local abattoir or from BOMED (Madison, WI), and rhesus oocytes obtained from the Assisted Reproductive Technology Cores at the Oregon National Primate Research Center (ONPRC; Beaverton, OR) and the Pittsburgh Development Center (PDC; Pittsburgh, PA). Briefly, immature bovine oocytes were recovered from ovaries by aspiration and matured for 24 hours in drops of TC199 medium modified with 10% fetal calf serum, 5 μg ml^–1 follicle stimulating hormone, 1 μg ml^–1 estrogen and 25 μg ml^–1 gentamycin at 39°C in 5% CO₂ under mineral oil. Mature rhesus oocytes were recovered by laparoscopy from gonadotropin-stimulated macaques (Hewitson et al., 1996).

Mature bovine and rhesus oocytes, arrested in metaphase of second meiosis, were then placed into drops of Tyrode's Albumin-Lactate-Pyruvate (TALP) culture medium under mineral oil. Frozen bull semen (American Breeders Service) was thawed to room temperature, layered over a two-part 45%, 90% Percoll gradient and centrifuged at 700 g for 15 minutes to isolate live sperm. Rhesus-monkey semen was collected, washed and centrifuged at 400 g for 5 minutes. Sperm were added to the drops of TALP medium containing the oocytes to give a final concentration of 110⁶ sperm per ml. Bovine oocytes and sperm were incubated at 39°C under 5% CO₂, and rhesus-monkey gametes were incubated at 37°C under 5% CO₂, until the desired stages in development, after pronuclear formation but before the onset of mitosis (Long et al., 1993; Laurinčík et al., 1998). To examine the effects of microtubule depolymerization on dynein and dynactin localization in pronucleate-stage zygotes, some oocytes and sperm were incubated until 14 hours after insemination, at which time the zygotes were transferred into drops of TALP medium containing 20 μM nocodazole (Sigma-Aldrich). The zygotes were then cultured in the presence of the drug for 1 hour. Some oocytes were parthenogenetically activated with two 5-minute incubations in 5 μM ionomycin, given 4 hours apart (Navara et al., 1994). These activated oocytes extruded a second polar body, formed a female pronucleus in the absence of a sperm aster, and were cultured for 16 hours.

Cumulus cells and zona pellucidae were removed from zygotes and parthenogenotes by short incubations with 1 mg ml^–1 hyaluronidase and 2 mg ml^–1 pronase, respectively. Zona-free parthenogenotes and zygotes were then gently pipetted onto poly-L-lysine-coated coverslips in Ca²⁺-free TALP medium, fixed for 40 minutes in 2% formaldehyde and permeabilized in 10 mM PBS + 0.1% Triton X-100 for an additional 40 minutes. Alternatively, samples were fixed in absolute methanol using previously described methods (Simerly and Schatten, 1993). After fixation and permeabilization, samples were blocked for 1 hour in 10 mM PBS + 0.3% bovine serum albumin (BSA) + 1% fetal calf serum prior to incubation with primary and secondary antibodies. Primary antibodies were applied for 12 hours in a humidified chamber and samples were rinsed four times for 10 minutes each following the addition of primary and secondary antibodies. AlexaFluor-488- and -568-conjugated secondary antibodies were applied to the samples for 1 hour, and DNA was labeled with 10 μg ml^–1 TOTO-3 (Molecular Probes). Coverslips were mounted onto glass microslides in Vectashield anti-fade medium (Vector Laboratories) to retard photobleaching. Zygotes and parthenogenotes were imaged using a Leica TCS SP2 spectral confocal microscope, with laser lines at 488 nm, 568 nm and 633 nm wavelengths.

Three monoclonal antibodies (MAbs) against dynein intermediate chain were used for immunocytochemistry in these studies: MAb1618 at 1:50 (Chemicon), clone 74.1 at 1:200 (Covance/Babco) and clone 70.1 at 1:150 (Sigma-Aldrich). Anti-dynein-heavy-chain clone 440.4 (1:20; Sigma-Aldrich) was also used. Three monoclonal and two polyclonal anti-dynactin-p150^Glued antibodies were used in these experiments: MAb150.1 and MAb150B (1:20 each; kind gifts from T. Schroer), MAb clone 1 (1:20; BD Transduction Labs), and polyclonal N-19 and G-18 (1:50 each; Santa Cruz Biotechnology). Anti-dynactin p62 polyclonal antibody N-17 (1:50; Santa Cruz Biotechnology) and anti-dynactin p50 clone 25 (1:20; BD Transduction Labs) were also used. To identify and target the nuclear pore complex, MAb414 (1:250; Covance/Babco) and anti-nucleoporin p62 clone RL31 (1:50; Affinity BioReagents) and polyclonal N-19 (1:50; Santa Cruz Biotechnology) were used. Anti-vimentin polyclonal antibodies H-84 and C-20 (each at 1:50; Santa Cruz Biotechnology) and anti-β-tubulin clone E7 (1:5; Developmental Studies Hybridoma Bank) and polyclonal ATN02 (1:200; Cytoskeleton) were also used during these studies. Anti-calreticulin polyclonal antibody PA3-900 (1:1500; Affinity BioReagents) and anti-Golgi matrix GM130 clone 35 (1:20; BD Transduction Labs) were used for the transfection experiments. Primary antibodies were detected with AlexaFluor-488- and -568-conjugated secondary antibodies (each at 1:200; Molecular Probes).

Control experiments were performed using pre-immune mouse IgG antibodies at 1:20 (Chemicon). Pre-incubation of antibodies for 1 hour with either their corresponding antigens or human endothelial cell (HEC) lysates was performed as additional controls for both immunocytochemistry and antibody transfection. HEC lysates were provided by the Cell Culture Cores at the ONPRC and PDC. For the transfection experiments, antibodies were dialysed overnight using Slide-A-Lyzer cassettes (Pierce) in multiple changes of 10 mM PBS to remove sodium azide from the storage buffer. Transfection of zygotes at 12 hours after insemination was achieved using the Chariot™ reagent according to the manufacturer's recommendations (Active Motif). Briefly, for each antibody transfection into 20 zygotes, we prepared a 20 μl volume mix containing a 1:10 dilution of Chariot reagent and one of the following dilutions of antibodies: 1:2 of anti-dynactin MAb150.1; 1:5 of anti-dynactin p62 N-17; 1:15 of anti-dynein clone 70.1; 1:2 of anti-dynein clone 440.4; 1:25 of MAb414; 1:5 of anti-nucleoporin p62 N-19; 1:5 of anti-vimentin H-84; 1:150 of anti-calreticulin PA3-900; 1:2 of anti-GM130 clone 35; and 1:2 of pre-immune IgG. These dilutions were determined empirically by performing antibody titrations on oocytes and zygotes and assessing staining specificity and function-blocking ability. The concentrated mixtures of Chariot reagent and antibodies were made in serum-free TALP medium and incubated at room temperature for 30 minutes. Following this Chariot reagent-antibody binding step, each 20 μl volume was added to one well of a 96-well plate containing 20 zygotes, free of cumulus cells and zona pellucidae, in 80 μl of serum-free TALP medium. The samples were incubated at 39°C for 1 hour, after which an additional 100 μl of serum-containing TALP medium was added to the well; samples were cultured for an additional 11 hours and fixed for immunocytochemistry at 24 hours after insemination.

Immunoprecipitations (IPs) of p150^Glued dynactin were performed using Protein A Sepharose 4 Fast Flow medium (AP Biotech). Mature bovine oocytes were either lysed immediately or fertilized in vitro and cultured until 14 hours post-insemination, at which time they were lysed and prepared for antibody incubations. The lysis buffer consists of 50 mM triethanolamine (TEA), 500 mM NaCl, 0.5% Triton X-100, 1 mM DTT, 1 mM PMSF and 1:1000 dilution of CLAP protease inhibitors (10 mg ml^–1 each of chymostatin, leupeptin, antipain and pepstatin A). Lysates were centrifuged at 12,000 g for 10 minutes at 4°C, and protein concentrations were determined using the Bradford assay. Equal amounts of protein were used to load each lane in the gels.

Lysates were incubated with 3 μg of p150^Glued dynactin or pre-immune mouse IgG antibodies for 1 hour at 4°C. Protein A Sepharose 4 Fast Flow medium was then added to the samples for overnight incubation at 4°C. Following centrifugation, the supernatants were removed from the pellets and concentrated through a centrifugal filter device (Millipore) to reduce the volume. The pellets were washed, resuspended in Laemmli buffer and heated to 100°C for 5 minutes. Brief centrifugation of the pellets separated out the protein from the beads. Whole zygote and oocyte lysates, IP supernatants, IP pellets and beads alone were carefully loaded onto 4-20% gradient gels, and processed by SDS-PAGE and western blotting onto PVDF membranes.

Protein bands were visualized using the ECL Plus system and HyperFilm (AP Biotech), and referenced to Kaleidoscope prestained standards (Bio-Rad). Silver Stain Plus system from Bio-Rad was used to verify protein samples isolated from immunoprecipitations. Western blots were initially probed with MAb414 at 1:250, then stripped and reprobed twice: first with anti-vimentin C-20 at 1:500, and then by anti-dynactin p150^Glued MAb clone 1 at 1:250 and anti-dynein MAb1618 at 1:500. Blots were stripped each time with 100 mM β-mercaptoethanol, 2% SDS, 62.5 mM Tris-HCl pH 6.7 for 30 minutes at 50°C, followed by multiple washes, incubations with secondary antibodies and detection with enhanced chemiluminescence (ECL) reagents to verify that primary antibodies had been removed.

Following insemination in bovine and rhesus oocytes, the distinct sperm and egg genomic material becomes enclosed within separate pronuclear envelopes. The sperm-derived male pronucleus and egg-derived female pronucleus occupy predominantly central and cortical locations, respectively, within most zygotes examined (91%; 2486/2724). Male pronuclei can be discerned by identifying attached sperm tails (data not shown). The female then migrates through the egg cytoplasm to join the male pronucleus in the center. During female pronuclear migration, microtubules extend from a region adjacent to the male pronucleus and contact the surface of the female pronucleus (Fig. 1A, left). Subsequent movement of the female to the male pronucleus results in pronuclear apposition (Fig. 1A, right). Previous studies have shown that, when microtubules are either depolymerized with nocodazole or stabilized with Taxol, female pronuclear migration to the male is inhibited (Schatten, 1982).

The organization of microtubules and orientation of female pronuclear motility suggest that the minus-end-directed motor cytoplasmic dynein might facilitate this pronuclear migration. We therefore first examined where dynein localizes in the zygote, and whether it might be able to transport the female towards the male pronucleus. During both pronuclear migration and apposition, dynein intermediate chain concentrates exclusively around the female pronucleus (Fig. 1B, left and center). Weak, diffuse staining is also detected throughout the cytoplasm. When preabsorbed with its antigen, the anti-dynein antibody no longer shows staining within the zygotes (Fig. 1B, right). Dynein heavy chain shows a similar distribution to the intermediate chain, localized around the female but not the male pronucleus (data not shown). The presence of dynein at the surface of the female pronucleus, and its absence from the male, suggests that it might have a specific temporal and spatial function to mediate female pronuclear migration.

In order for dynein to transport cargo on microtubules, the cofactor dynactin is required. Consisting of at least eight distinct subunits, the dynactin complex contains domains that bind directly to dynein, microtubules, and cargo proteins (Schafer et al., 1994; Waterman-Storer et al., 1995; Vaughan and Vallee, 1995; Karki and Holzbaur, 1995; Echeverri et al., 1996; Burkhardt et al., 1997). We therefore examined the distribution of three different dynactin subunits in bovine zygotes to determine whether they also concentrate around the female pronucleus. The p150^Glued subunit, which binds to the dynein intermediate chain, concentrates around not only the female pronucleus but also the male pronucleus (Fig. 2A, far left). Dynactin subunit p50, which maintains the structure of the complex with respect to microtubule organization, also concentrates around both pronuclei during migration and apposition (Fig. 2A, center left). The p62 subunit, which resides in the domain that binds to cargo, localizes around the pronuclei in a pattern similar to the other two subunits (Fig. 2A, center right). When preincubated with human endothelial cell lysates, the primary antibodies against all three dynactin subunits are no longer detected by secondary antibodies in the zygotes; anti-p150^Glued is shown as an example (Fig. 2A, far right). This association of dynactin around both female and male pronuclei is distinct from that of dynein, examined further by double-labeling zygotes with both anti-dynein and anti-dynactin antibodies (Fig. 2B, top). Zygotes stained with antibodies against calreticulin, which distributes throughout the cytoplasm, reveal that not all proteins concentrate around the pronuclei during their migration (Fig. 2B, bottom). The enriched distribution and differential staining of dynein and dynactin at pronuclear surfaces might indicate specific functions performed during fertilization.

To determine whether dynein and dynactin are necessary for pronuclear migration and apposition, we transfected antibodies against the dynein intermediate and heavy chains, and dynactin subunits p150^Glued and p62, into pronucleate-stage bovine zygotes using Chariot reagent (Morris et al., 2001). Transfection of antibodies occurred after pronuclear formation but before pronuclear apposition, optimizing the temporal and spatial effects of the antibodies. Zygotes were allowed to develop until just prior to mitosis, when they were fixed and the distances between the male and female pronuclei scored for inhibition of pronuclear union. Measured from edge-to-edge, inter-pronuclear distances ≥10 μm, the average diameter of a pronucleus, indicate disrupted pronuclear migration.

Most zygotes transfected with antibodies against dynein intermediate and heavy chains show pronuclei ≥10 μm apart (94% and 93%, respectively), whereas all zygotes transfected with pre-immune mouse IgG antibodies and Chariot reagent alone have interpronuclear distances <10 μm (Fig. 3A,B). Antibodies against dynactin subunits p150^Glued and p62 also block pronuclear union, with most pronuclei at distances ≥10 μm from one another (98% and 74%, respectively). Pre-absorption of anti-dynein and anti-dynactin antibodies with either their corresponding antigens or human endothelial cell lysates prior to transfection results in all pronuclei to be <10 μm apart (Fig. 3B).

Zygotes transfected with antibodies against dynein and dynactin display a modest alteration in staining when compared with control zygotes examined at earlier stages: dynein 70.1 is enriched at the surface of the female pronucleus proximal, but not distal, to the sperm aster; and dynactin p150^Glued distribution around both pronuclear surfaces, as well as along microtubules, exhibits decreased concentration (Fig. 3A). These alterations can be attributed to the effects of the different antibodies on the ability of dynein and dynactin to bind to membranes and microtubules. Despite the failure of pronuclear apposition in the zygotes transfected with anti-dynein and anti-dynactin antibodies, the sperm-aster microtubules appear unperturbed between the male and female pronuclei (Fig. 3A). Thus, although both dynein and dynactin are required for female pronuclear motility, disruption of these proteins with antibodies does not reorganize the sperm aster. The stabilizing presence of the centrosome probably precludes such alteration.

We next examined the role of the sperm aster in distributing dynein and dynactin to the female pronuclear surface by parthenogenetically activating oocytes. Parthenogenesis generates polymerized microtubules throughout the ooplasm, but the absence of a sperm centrosome prevents aster formation. In all of the parthenogenotes observed (131 total; n=3), dynactin concentrates around the female pronucleus, whereas dynein redistributes towards the cortex and is strikingly absent from the female pronuclear surface (Fig. 4). Microtubules are also enriched at the cortex, unfocused in their orientation. These results suggest that dynein association with the female pronucleus requires a sperm aster and that polymerized microtubules alone are not sufficient.

Because the localization of dynein, but not dynactin, to the female pronucleus depends upon the formation of a sperm aster, we then asked whether microtubules are necessary to retain dynein and dynactin at the female pronuclear surface once they are there. Nocodazole was administered to pronucleate-stage zygotes to depolymerize microtubules; exposure occurred after the female pronucleus was contacted by the sperm aster but before its apposition with the male pronucleus. In nocodazole-treated zygotes, the association of dynein with the female pronucleus is only 7±2.7%, compared with 82±5.6% in controls (P<0.001; n=4; Fig. 5A,B). By contrast, nocodazole does not cause significant redistribution of dynactin from male and female pronuclear surfaces (i.e. 80±4.3% in controls, compared with 62±8.3% in nocodazole-treated zygotes; Fig. 5A,B), even though the microtubules are clearly disassembled.

Microtubules distribute near the surface of the male pronucleus in controls in the absence of dynein, suggesting that microtubule localization to male pronuclear surfaces is dynein independent. The loss of dynein from the female pronuclear surface after nocodazole supports the hypothesis of a microtubule-dependent distribution of dynein to the female pronucleus. The retention of dynactin at the surfaces of both pronuclei in the absence of microtubules raises the possibility that dynactin associates with structural components found near, at or within pronuclear membranes.

Because dynactin concentration around both pronuclei is retained after microtubules are depolymerized, we examined whether there is a physical interaction between dynactin and pronuclear envelope proteins. Embedded within the pronuclear envelope, NPCs distinguish pronuclear membranes from ER and Golgi membranes, and are attractive candidates for an association with dynein, dynactin and microtubules (Salina et al., 2002; Lénárt and Ellenberg, 2003). The p150^Glued subunit of dynactin was immunoprecipitated from pronucleate-stage zygotes and unfertilized oocytes, and an antibody with wide specificity against nucleoporins was used to probe the dynactin IP. We chose the pan-nucleoporin antibody MAb414 to detect the NPCs, because it has been shown to recognize nucleoporins with molecular mass 270 kDa, 175 kDa and 62 kDa (Davis and Blobel, 1987), and (depending on the cell type being used for SDS-PAGE and western blotting) can detect up to an additional three or four nucleoporins of different molecular masses (Aris and Blobel, 1989; Cronshaw et al., 2002).

Comparing the pellets and supernatants of the dynactin IP, along with protein-A/Sepharose beads alone and whole cell lysates from both zygotes and unfertilized oocytes, we detect five bands corresponding to nucleoporins with molecular masses 270 kDa, 175 kDa, 62 kDa, 35 kDa and 18 kDa in the IP pellets isolated from zygotes (Fig. 6A). These nucleoporins are also detected in the whole zygote lysates, but not in the IP supernatants or beads alone. This observation contrasts with the samples isolated from oocytes, in which the nucleoporins are identified in the IP supernatants as well as the whole oocyte lysates, but not in the IP pellets or beads alone (Fig. 6A). Thus, nucleoporins immunoprecipitate with dynactin in pronucleate-stage zygotes, but not in unfertilized oocytes.

Vimentin is also detected in the IP pellets isolated from zygotes, and is identified with molecular mass 58 kDa (Fig. 6B). Found at low abundance in the IP supernatants, vimentin is absent from the beads-alone samples. In the oocyte-derived samples, vimentin is detected in the IP supernatants and the whole cell lysates, but not in the IP pellets or beads alone (Fig. 6B). The low level of vimentin detected in the zygote IP supernatants suggests that not all of the vimentin protein associates with dynactin during pronuclear migration We conclude that, like nucleoporins, vimentin co-immunoprecipitates with dynactin specifically under zygotic conditions. When the western blots are then probed for dynein intermediate chain and dynactin p150^Glued, both proteins are detected in the IP pellets and whole cell lysates of both zygotes and oocytes, and are identified with molecular masses 74 kDa and 150 kDa, respectively (Fig. 6C).

The distribution of nucleoporins and vimentin during pronuclear migration and apposition was then characterized by confocal microscopy. Nucleoporin p62, recognized in the dynactin immunoprecipitate by MAb414, localizes predominantly to the surfaces of the male and female pronuclei, with additional punctate cytoplasmic foci (Fig. 6D). Co-distribution of this NPC protein with dynactin p150^Glued is observed when the images are merged. Vimentin, meanwhile, displays a branched staining pattern between the two pronuclei, concentrating around their surfaces and extending throughout the cytoplasm (Fig. 6E). Co-localization of vimentin and dynactin p150^Glued at pronuclear surfaces is detected when the images are merged, with additional distinct patterns of vimentin observed between pronuclei.

The association of nucleoporins and vimentin with both dynactin and dynein suggests that they might also be required for female pronuclear migration. To determine whether nucleoporins and vimentin are necessary for pronuclear apposition, we transfected antibodies against them into pronucleate-stage zygotes once again using the Chariot reagent. Interpronuclear distances were scored as before to determine the level of inhibition of pronuclear union. When transfected with MAb414 and anti-nucleoporin p62 antibodies, the majority of zygotes show pronuclei ≥10 μm apart (83% and 91%, respectively), whereas all zygotes transfected with preimmune mouse IgG antibodies have interpronuclear distances <10 μm (Fig. 7A,B). Antibodies against vimentin also block pronuclear union, with most pronuclei at distances ≥10 μm from one another (98%; Fig. 7A,B). Preabsorption of anti-nucleoporin and anti-vimentin antibodies with their corresponding antigens before transfection results in all pronuclei to be <10 μm apart (Fig. 7B).

Nucleoporins show normal distribution around pronuclei in both control IgG-transfected and anti-nucleoporin-transfected zygotes, despite the arrest in female pronuclear migration seen in the second group (Fig. 7A). Vimentin, however, branches around and between the two pronuclei in zygotes transfected with anti-vimentin antibodies. Dynactin p150^Glued shows normal localization to pronuclei in control IgG-transfected and anti-nucleoporin-transfected zygotes, whereas its distribution around pronuclear surfaces is reduced after transfection with anti-vimentin antibodies (Fig. 7A). This reduction could be due to an alteration in binding between vimentin and dynactin in the presence of the antibodies, or to a masking effect revealed during the immunocytochemistry. Dynein distributes around female pronuclei in transfected zygotes, with decreased concentration at pronuclear surfaces following transfection with anti-vimentin antibodies (data not shown). Antibodies against the ER-resident protein calreticulin and the Golgi apparatus protein GM130 do not inhibit pronuclear union, because all zygotes transfected with these antibodies have interpronuclear distances <10 μm (Fig. 7B). Thus, specifically, nuclear pore complex proteins and vimentin appear to be necessary for pronuclear migration and apposition.

We have shown that mammalian fertilization requires both dynein and dynactin to mediate genomic union, and that dynein concentrates exclusively around the female pronucleus. Dynactin, by contrast, localizes around both pronuclei and associates with nucleoporins and vimentin in addition to dynein. The findings that a sperm aster is required for dynein to localize to the female pronucleus and that microtubules are necessary to retain dynein, but not dynactin, at its surface suggests that nucleoporins, vimentin and dynactin might associate upon pronuclear formation, and that subsequent sperm aster contact with the female pronuclear surface allows dynein to interact with these proteins.

The dependence of dynein on the sperm aster to localize to the female pronuclear surface is consistent with earlier observations that female pronuclear movement does not begin until sperm aster contact has been established (Schatten, 1982; Navara et al., 1994). We conclude from the parthenogenetic activation experiments that the preferential distribution of dynein around the female pronucleus is due to sperm-aster-mediated delivery of dynein to the female pronuclear surface on the plus ends of microtubules. Recent studies in budding yeast provide evidence supporting this mechanism for dynein transport to its target membrane, showing that dynein is delivered to the cell cortex on the plus ends of polymerizing astral microtubules (Sheeman et al., 2003). Such accumulation of dynein at microtubule plus ends would be expected to occur in the absence of dynactin and, indeed, Sheeman and colleagues saw this when they examined dynactin mutants (Sheeman et al., 2003). Given the different observed distributions of dynein and dynactin during mammalian fertilization and parthenogenesis, and the need for a sperm aster for proper dynein concentration, this mechanism could explain how dynein localizes to the female pronuclear surface independently of dynactin.

The preferential localization of dynein probably depends on the spatiotemporal control of the cell cycle. Fertilized eggs enter S-phase after the formation of the sperm aster and during the migration of the pronuclei, as detected with 5-bromo-2′-deoxyuridine-5′-monophosphate (BrdU) (Nomura et al., 1991; Eid et al., 1994). Recent evidence has shown that centrosome-anchored microtubule arrays in fibroblasts do not accumulate dynein at their minus ends until after entering S-phase (Quintyne and Schroer, 2002). Because maintenance of the sperm aster is also likely to be dynein independent until after S-phase entry, proximity of the centrosome to the male pronucleus might preclude dynein from both associating with the male pronuclear surface and accumulating at the centrosome until after pronuclear migration commences.

Dynactin concentration around zygotic pronuclei, however, resembles its localization to prophase nuclear envelopes in somatic cells, where it is thought to facilitate nuclear envelope breakdown (Salina et al., 2002; Beaudouin et al., 2002). Neither dynein nor dynactin, to our knowledge, has previously been reported to associate with interphase nuclear envelopes. Because migrating pronuclei are in S-phase, dynactin concentration around pronuclear envelopes is likely to persist throughout zygotic interphase, revealing a unique functional association during fertilization.

Retention of dynactin around both pronuclei in the presence of nocodazole demonstrates a microtubule-independent, dynein-independent association of dynactin with pronuclear membranes. Although they are not required to retain dynactin at pronuclei once they have formed, microtubules are necessary for recruiting membrane material around paternal and maternal chromatin, bringing nuclear pore complexes to these membranes for insertion, and allowing pronuclei to form properly, as shown through the use of Xenopus egg extracts (Sutovsky et al., 1998; Ewald et al., 2001). Microtubules involved in these roles do not require the formation of a sperm aster, however, because female pronuclear membranes complete their formation around the maternal chromatin in parthenogenetically activated oocytes as well as in zygotes prior to contact by the sperm aster (C.P., unpublished). For phyla in which oocytes complete meiosis before fertilization, as in many echinoderms, a female pronucleus is fully formed before sperm entry, demonstrating its independence from the sperm aster (Gilbert, 1997).

Immunoprecipitation of dynactin p150^Glued from pronucleate-stage zygotes identifies a physical interaction between dynactin and nucleoporins of molecular masses 270 kDa, 175 kDa, 62 kDa, 35 kDa and 18 kDa, as well as vimentin and dynein. The nucleoporins are enriched in the zygote immunoprecipitate compared with whole zygote lysates, showing a slight increase in intensity as analysed by western blotting. Similar enrichment might also be seen for vimentin, dynein and dynactin. Localization of the different immunoprecipitate components to pronuclear surfaces, including the co-distribution of dynactin with nucleoporin p62 and vimentin, reinforces the association of dynactin with NPCs and intermediate filaments. Given that nucleoporin p62 resides within the center of the nuclear pore, which has an estimated maximum channel diameter of 40 nm (Ryan and Wente, 2000; Panté and Kann, 2002), the interaction between nucleoporins and dynactin is probably indirect. Recent evidence of a direct interaction between the dynein-dynactin complex and vimentin at the ultrastructural level (Helfand et al., 2002), combined with observations demonstrating an association between intermediate filaments and nuclear pores (Jones et al., 1982; Fujitani et al., 1989), supports the hypothesis that dynactin interacts with nucleoporins through vimentin.

Transfection of antibodies into pronucleate-stage zygotes has allowed us to determine that dynein, dynactin, nucleoporins and vimentin are specifically required during pronuclear migration and apposition. Inhibition of the intermediate and heavy chains of dynein, and the p150^Glued and p62 subunits of dynactin dramatically blocks genomic union. Perturbation of the motor complex, however, does not significantly alter the microtubule network between the pronuclei. Blocking nucleoporins and vimentin also inhibits pronuclear apposition, with a slight decrease in dynein and dynactin concentration at pronuclear surfaces. Mammalian fertilization could thus provide distinct roles for dynein and dynactin in the zygote: dynein, together with dynactin, trafficking the female pronucleus towards the centrosome attached to the male pronucleus; and dynactin, together with vimentin, associating with NPC proteins to provide attachment sites for dynein and sperm-aster microtubules at the female pronucleus.

We propose a model for pronuclear formation, migration and apposition in which dynactin, nucleoporins and vimentin interact together with microtubules and dynein to mediate genomic union (Fig. 8). Nuclear pore complex assembly and insertion into the envelopes of newly forming pronuclei bring dynactin and vimentin filaments to the cytoplasmic face of nuclear pores, where they interact as a macromolecular complex (Fig. 8B). This formation probably depends upon sperm-aster-independent microtubules, appearing within the ooplasm after egg activation and contacting the surfaces of the pronuclei during their assembly. The sperm aster, meanwhile, develops as a focused set of microtubules radiating out from the centrosome, now serving as the dominant microtubule-organizing center in the zygote (Fig. 8B). Enlargement of the sperm aster extends the microtubule plus ends away from the male pronucleus, some of which then reach the female pronuclear envelope. Growth of the sperm aster towards the cortex could deliver dynein to the surface of the female pronucleus on these microtubule plus ends, allowing it to bind to dynactin and vimentin at nuclear pores and to activate its motor activity (Fig. 8C). The dynein-dynactin complex would then be able to transport the female pronucleus to the sperm aster minus ends, culminating in pronuclear apposition (Fig. 8D).

Understanding these protein interactions might shed light on certain cases of clinical idiopathic infertility in which inseminated eggs arrest development after the pronuclei fail to unite (Simerly et al., 1995). It has been observed that, in approximately 6% of human fertilization failures, characterized in embryos discarded from infertility clinics, the sperm aster shows abnormal morphology within the zygote (Asch et al., 1995). Both incomplete assembly and disarrayed organization of sperm-aster microtubules would compromise the association of dynein with the female pronucleus. Should microtubules fail to bind to the female pronucleus, preventing dynein from localizing to its surface, genomic union would be unsuccessful.

From these studies, we conclude that dynactin associates with nucleoporins and vimentin at the surfaces of both pronuclei upon their formation and throughout the fertilization cell cycle. Dynein, by contrast, depends upon sperm-aster microtubules to associate with dynactin, nucleoporins and vimentin exclusively at the female pronucleus to facilitate motility. The interaction between dynactin and nuclear envelope proteins common to both pronuclei, together with the spatial distribution of dynein to these proteins at the female pronucleus, might ensure successful genomic union that completes mammalian fertilization.

We thank T. Schroer for the gift of antibodies and C. Chace, A. Chan, T. Dominko, E. Jacoby, L. Hewitson, R. Moreno, P. Sutovsky and D. Takahashi for experimental assistance. We thank K.-Y. Chong, H. Gray, L. Kauffman, C. Navara and P. Sammak for their critical reading of the manuscript and for helpful discussions. J.R.-S. is supported by a grant from FCT, Portugal (POCTI/ESP/38049/2001). This work was funded by NIH research grants to G.S.