|Home | About | Journals | Submit | Contact Us | Français|
Aurora A kinase localizes to centrosomes and is required for centrosome maturation and spindle assembly. Here, we describe a microtubule-independent role for aurora A and centrosomes in nuclear envelope breakdown (NEBD) during the first mitotic division of the C. elegans embryo. Aurora A depletion does not alter the onset or kinetics of chromosome condensation, but dramatically lengthens the interval between the completion of condensation and NEBD. Inhibiting centrosome assembly by other means also lengthens this interval, albeit to a lesser extent than aurora A depletion. By contrast, centrosomally-nucleated microtubules and the nuclear envelope-associated motor dynein are not required for timely NEBD. These results indicate that mitotic centrosomes generate a diffusible factor, which we propose is activated aurora A, that promotes NEBD. A positive feedback loop, in which an aurora A-dependent increase in centrosome size promotes aurora A activation, may temporally couple centrosome maturation to NEBD during mitotic entry.
Aurora A is a mitotic kinase that choreographs events during mitotic entry. Interest in aurora A has been stimulated by its connection to tumorigenesis. Aurora A resides in a genomic region often amplified in tumors (Bar-Shira et al., 2002) and its overexpression can transform cells in culture and in vivo (Bischoff et al., 1998; Wang et al., 2006; Zhou et al., 1998). Aurora A is overexpressed in a high proportion of breast, colorectal and gastric cancers and a specific allele of aurora A, F31I, has been linked to increased cancer susceptibility in humans (Andrews, 2005; Crane et al., 2004; Marumoto et al., 2005; Meraldi et al., 2004).
Several demonstrated functions of aurora A are connected to centrosomes (Crane et al., 2004; Ducat and Zheng, 2004; Dutertre et al., 2002; Marumoto et al., 2005). Centrosomes consist of a pair of centrioles surrounded by pericentriolar material that promotes microtubule assembly. During cell division, centrosomal microtubule asters contribute to the formation and positioning of the mitotic spindle. In preparation for these functions, centrosomes “mature” during mitotic entry, recruiting additional pericentriolar material to increase ~5-fold in size and nucleating capacity (Palazzo et al., 2000). Aurora A localizes to the pericentriolar material and is required for maturation (Berdnik and Knoblich, 2002; Blagden and Glover, 2003; Brittle and Ohkura, 2005; Hannak et al., 2001). Centrosomal aurora A is in dynamic equilibrium with a cytoplasmic pool, turning over rapidly (half-life of ~3s in human cells; Stenoien et al., 2003). This rapid turnover indicates that aurora A has a signaling rather than structural role in centrosome assembly, and that events at centrosomes have the potential to influence the state of the cytoplasmic pool of aurora A.
In addition to centrosome maturation, aurora A has been implicated in regulating cell cycle progression. In cycling Xenopus extracts, depletion of aurora A delays both the activation of Cdk1 and chromosome condensation (Liu and Ruderman, 2006). A delay in Cdk1 activation has also been documented following RNAi-mediated depletion of aurora A in human cells (Hirota et al., 2003). The connection between the role of aurora A in centrosome assembly and cell cycle progression is less clear. Although postulated to be inter-connected in human cells (Hirota et al., 2003), the effect of depleting aurora A on Cdk1 activation is independent of the presence of centrosomes in Xenopus extracts (Liu and Ruderman, 2006).
Subsequent to its involvement in Cdk1 activation and centrosome maturation, both of which occur prior to NEBD, aurora A promotes spindle assembly in conjunction with its activator TPX2. TPX2 is regulated by the Ran pathway after NEBD, and inhibition of TPX2 blocks spindle assembly without apparent effects on centrosome structure or cell cycle progression (Crane et al., 2004; Ducat and Zheng, 2004; Eyers and Maller, 2003; Garrett et al., 2002; Kufer et al., 2003; Özlü et al., 2005).
Here, we capitalize on the highly stereotypical first division of the C. elegans embryo to explore the role of aurora A in the coordination of mitotic events during the period leading up to NEBD. We show that following aurora A inhibition chromosomes initiate and complete condensation with normal timing, suggesting that Cdk1 is activated normally. However, aurora A depleted embryos exhibit a specific delay between the completion of chromosome condensation and NEBD. Inhibition of centrosome assembly via other means also delays NEBD, but to a lesser extent than depletion of aurora A. By contrast, inhibition of microtubule assembly or depletion of dynein does not alter NEBD timing, indicating that the role of centrosomes and aurora A is not mediated by centrosomal microtubules. Our results demonstrate an important role for centrosomes and aurora A in NEBD, and suggest a model in which positive feedback between increasing centrosome size and aurora A activation temporally couples centrosome maturation to NEBD.
To explore the functions of aurora A during the period leading up to NEBD, we took advantage of the stereotypical first mitotic division of the C. elegans embryo. After fertilization, the chromosomes in the oocyte nucleus complete their meiotic segregation, generating the oocyte pronucleus and two polar bodies. Subsequently, the embryo enters its first mitotic division. During prophase, the replicated chromosomes condense as the oocyte and sperm pronuclei migrate towards each other. Concurrently, the centrosomes associated with the sperm pronucleus increase ~5-fold in size. After the pronuclei meet, the nuclear envelopes become permeable and the chromosomes interact with spindle microtubules to align and segregate. (Oegema and Hyman, 2005; Fig. 1B). Nuclear envelope permeabilization can be followed by monitoring the diffusion of free GFP:histone out of the nucleus. We refer to the timepoint when the free nuclear GFP:histone fluorescence has equilibrated with the cytoplasm as NEBD (nuclear envelope breakdown). In previous work, we established RNAi conditions that reduce aurora A protein levels by >90% and analyzed the consequences of this depletion (Hannak et al., 2001; Fig. 1A). This analysis revealed that aurora A is required for the increase in centrosome size during mitotic entry. In addition, mitotic entry appeared to take longer in aurora A-depleted embryos (Hannak et al., 2001; Fig. 1A), although the lack of quantifiable cell cycle reference points precluded further analysis of this defect.
In Xenopus extracts, depletion of aurora A results in parallel delays in Cdk1 activation and chromosome condensation (Liu and Ruderman, 2006), a process downstream of Cdk1 activation (Liu and Ruderman, 2006; Potapova et al., 2006; Shimada et al., 1998). To test if the delay in mitotic entry in aurora A depleted C. elegans embryos is due to a defect in Cdk1 activation, we used a recently developed image analysis method (outlined in Fig. 1B; Maddox et al., 2006) to compare chromosome condensation kinetics in control and aurora A depleted embryos. In this method, the progressive shift of the fluorescence intensity distribution of the GFP:histone signal that accompanies condensation is monitored by measuring the percentage of pixels with intensities less than 50% of the image maximum (the condensation parameter) in individually scaled projections of 3D timelapse sequences. The condensation parameter increases monotonically as chromosomes condense and can be used to quantitatively compare condensation kinetics between control and perturbed embryos (Fig. 1B; Maddox et al., 2006). For every tested perturbation, the average of the condensation parameter at each time point was calculated from values measured in 6–12 embryos time-aligned with respect to NEBD. To simplify presentation, the plots of the condensation parameter are displayed aligned with the onset of condensation. No increase in the condensation parameter was detected during prophase in embryos depleted of SMC-4 (Fig. 1B), one of the ATPase subunits of the condensin complex (Hagstrom et al., 2002), and intermediate effects are clearly detected following depletions of other chromosomal proteins (Maddox et al., 2006), validating the method.
In aurora A depleted embryos, chromosomes condensed with kinetics identical to those in control embryos, attaining a state of maximum condensation over a period of ~500s (Fig. 2A). However, there was a striking delay between the end of condensation and NEBD (Fig. 2A,B,D). To determine if the onset of condensation was delayed, we used the independent timer of pronuclear size. We filmed embryos starting at the onset of anaphase of meiosis II, and found that pronuclear diameter steadily increased ~2-fold between the end of meiosis II and NEBD in both control and aurora A-depleted embryos. The average pronuclear diameter when aurora A depleted embryos initiated condensation (7.0 + 0.1 μm) was indistinguishable from controls (7.1 + 0.1 μm), indicating that the timing of condensation onset was not altered by aurora A depletion (Fig. 2C, D). However, in contrast to control embryos where NEBD occurred ~60s after chromosome condensation was complete, chromosomes remained in the condensed state for an additional 7 minutes before NEBD in aurora A depleted embryos.
The normal onset and kinetics of chromosome condensation indicates that aurora A depletion does not result in a global defect in Cdk1 activation. Consistent with this conclusion, the interval between the onset of meiosis II anaphase and regression of the pseudocleavage furrow, a cortical event analogous to the relaxation of surface contractile waves that accompany Cdk1 activation in Xenopus embryos (Rankin and Kirschner, 1997), was also not altered by aurora A depletion (12.9 + 0.3 min in controls vs. 13.4 + 0.5 min in aurora A depleted embryos; n=9 for both). Given that wild-type embryos take only 20 min to progress from the end of meiosis II to NEBD, and an additional 4 minutes to progress from NEBD to the onset of cytokinesis, the magnitude of the NEBD delay following aurora A depletion (~7 min) indicates that aurora A makes a major contribution to timely NEBD.
Permeabilization of the nuclear envelope has been shown to proceed in two phases (Lenart et al., 2003; Terasaki et al., 2001). During the first phase, the peripheral components of the nuclear pores are dismantled, rendering the envelope permeable to macromolecules up to the size of the open pore, ~40 nm in diameter. In the second phase, the nuclear pores are completely removed, concomitant with the fenestration of the nuclear envelope which allows larger particles to enter the nuclear space (Lenart et al., 2003). To determine whether depletion of aurora A slows the process of permeabilization in addition to delaying its onset, we monitored the interval between the sequential entry into the nuclear space of two different-sized cytoplasmic markers (Fig. 3A,B). As a marker for the first phase, we used a Texas-Red labeled 70 kD dextran with a predicted hydrodynamic radius of 36 nm (Lenart et al., 2003). As a marker for the second phase, we used a GFP fusion with the heavy chain of myosin II (GFP:NMY-2; Nance et al., 2003), which is expected to have a hydrodynamic radius larger than 40 nm (Citi and Kendrick-Jones, 1987). In control embryos (n=10), GFP:NMY-2 entered the nucleus 42 + 4 s after the 70kD dextran. In aurora A depleted embryos (n=7), this interval was increased to 60 ± 9 s (p<0.0004), indicating that, in addition to a dramatic delay in the onset of permeabilization, depletion of aurora A slows the progressive permeabilization of the nuclear envelope once it initiates.
Next, we examined the dynamics of nuclear pores and the nuclear lamina using strains co-expressing RFP-histone and either a GFP fusion with NUP-155, a stable component of the pore wall (Franz et al., 2005), or a YFP fusion with LMN-1, the single C. elegans B-type lamin (Liu et al., 2000; Riemer et al, 1993). This analysis revealed no significant difference between control and aurora A-depleted embryos in the localization of NUP-155 or LMN-1 at the time when the nuclear envelope became permeable to free RFP-histone. This indicates that depletion of aurora A coordinately delays the onset of nuclear envelope permeabilization, nuclear pore removal, and lamin disassembly by ~ 7 minutes.
We next monitored the progressive loss of pores and the lamina following permeabilization onset. Consistent with the slowing of permeabilization observed in the dextran-myosin analysis (Fig. 3A), the progression of nuclear pore removal (Fig. 3C) and disassembly of the lamin meshwork (Fig. 3D,E) were also slowed by aurora A depletion. We conclude that after permeabilization initiates, pore removal, fenestration of the envelope, and lamin disassembly all progress more slowly in aurora A-depleted embryos than in controls.
TPX2 is a well-characterized aurora A activator that plays an important role in spindle assembly (Crane et al., 2004; Ducat and Zheng, 2004; Eyers and Maller, 2003; Kufer et al., 2003). To determine if TPX2 contributes to aurora A-dependent timely NEBD, we analyzed embryos depleted of the C. elegans TPX2-related protein, TPXL-1. Centrosomes mature and separate normally in TPXL-1 depleted embryos. However, after NEBD, the two centrosomal microtubule asters collapse into the chromosomes (Supplemental Fig. 1A; Özlü et al., 2005). There was no significant difference between control and TPXL-1 depleted embryos in the length of the interval between the completion of chromosome condensation and NEBD (Supplemental Figure 1). This result is consistent with the emerging picture of genetically separable functions for aurora A before and after NEBD (Crane et al., 2004; Ducat and Zheng, 2004; Eyers and Maller, 2003; Garrett et al., 2002; Kufer et al., 2003; Özlü et al., 2005). After NEBD, aurora A acts in concert with TPX2-like proteins and the Ran pathway to promote spindle assembly. Prior to NEBD, a different activator(s) mediates the critical functions of aurora A in centrosome maturation and nuclear envelope permeabilization.
Previous work in human cells suggested that centrosomal microtubules interact with dynein on the nuclear envelope to generate tension that accelerates NEBD (Beaudouin et al., 2002; Salina et al., 2002). Since aurora A is required for the increase in centrosome size and nucleating capacity that normally occurs prior to NEBD, such a mechanism could explain the NEBD delay in aurora A-inhibited embryos. To investigate this possibility, we characterized embryos depleted of dynein or treated with nocodazole to depolymerize centrosomal microtubules. Depletion of dynein resulted in the expected phenotypes (Gönczy et al., 1999) —multiple oocyte pronuclei as a consequence of defects in female meiosis (Fig. 4A; arrowheads), and failure of centrosome separation and pronuclear migration. However, the interval between the completion of chromosome condensation and breakdown of the sperm pronuclear envelope was not significantly different from controls (Fig. 4B, E). Nocodazole treatment, which eliminated detectable microtubule polymers, prevented pronuclear migration, and caused the centrosomes to dissociate from the nuclei (Fig. 4A), also did not cause a significant delay (Fig. 4C, E). To more rigorously test the consequences of disrupting centrosome-nucleus attachment, we analyzed embryos depleted of ZYG-12, a hook domain-containing protein required for the association of centrosomes with the nuclear envelope (Malone et al., 2003). Prior to NEBD, the centrosomes in ZYG-12 depleted embryos are randomly positioned within the embryo (Fig. 4A). Although the interval between the completion of chromosome condensation and NEBD that we measured in ZYG-12 depleted embryos was ~50s longer than that in controls (Fig. 4D, E), this delay is an order of magnitude less than that following depletion of aurora A (Fig. 2A) and we cannot be certain if this difference is significant as it is close to the typical error for this measurement (between ~10–40s). We conclude that a defect in mechanical interactions between centrosomally-nucleated microtubules and the nuclear envelope cannot account for the dramatic delay in nuclear envelope breakdown caused by aurora A depletion.
The NEBD delay resulting from inhibition of aurora A, while independent of centrosomal microtubules, could be a consequence of its effect on centrosome structure. To explore this possibility, we analyzed embryos in which centrosome assembly was perturbed by depletion of SPD-2 or SPD-5. Like aurora A, SPD-2 localizes to centrosomes and is required for their mitotic maturation. Depletion of SPD-2 results in mitotic centrosomes that are much smaller than in wild-type (Fig. 5A,B; Kemp et al., 2004; Pelletier et al., 2004). SPD-5 is the major scaffold protein for centrosome assembly in C. elegans. Depletion of SPD-5 completely inhibits the recruitment of pericentriolar material by the centrioles; in SPD-5 depleted embryos no foci of the centrosomal marker γ-tubulin or centrosomal microtubule asters are observed at any cell cycle stage (Fig. 5A,B; Hamill et al., 2002). Chromosomes in embryos depleted of SPD-2 (Fig. 5C, E) or SPD-5 (Fig. 5D, E) condensed with kinetics similar to controls. However, both perturbations increased the interval between the completion of chromosome condensation and NEBD by ~3 min. This delay, while significant, is less than the 7 min delay observed for aurora A depletion under identical conditions. Simultaneous depletion of SPD-5 and aurora A resulted in an NEBD delay identical to that in embryos depleted of aurora A alone (Supplemental Figure 2). We conclude that centrosomes play an important role in promoting timely NEBD. This function is intrinsic to the pericentriolar material and independent of centrosomally-nucleated microtubules. In aurora A depleted embryos, the centrosomes that remain do not accelerate NEBD. However, cytoplasmic aurora A can accelerate NEBD, albeit less effectively, in the absence of centrosomes.
Our results demonstrate that centrosomes promote nuclear envelope permeabilization and that it is the presence of centrosomes, and not mechanical interactions between centrosomal microtubules and motor proteins on the nuclear envelope, that is critical for timely NEBD. Cumulatively, these findings suggest that mitotic centrosomes generate a diffusible factor that promotes NEBD. If this hypothesis is correct, then increasing the distance between the centrosome and the nucleus should delay nuclear envelope permeabilization by the amount of time it takes the signal to diffuse the additional distance. We tested this idea by measuring the interval between permeabilization of the oocyte-derived and sperm-derived pronuclei in embryos that fail to undergo pronuclear migration (Fig. 6A,B). Migration of the oocyte-derived pronucleus towards the sperm-pronucleus is mediated by interactions between microtubules (nucleated by the centrosomes associated with the sperm pronucleus), and dynein (on the envelope of the oocyte-derived pronucleus). Consequently, pronuclei fail to migrate in nocodazole-treated embryos as well as in embryos in which centrosome assembly (spd-5(RNAi)) or dynein is inhibited (dhc-1(RNAi)). In nocodazole-treated embryos, the sperm pronucleus, which is immediately adjacent to the centrosomes, became permeable to GFP:histone ~80s before the oocyte pronucleus, which is on the other side of the embryo (Fig. 6A,B). A similar asynchrony in the permeabilization of the sperm and oocyte pronuclei was observed in dynein-depleted embryos. This asynchrony is absent in SPD-5 depleted embryos (which lack functional centrosomes) and in ZYG-12 depleted embryos (in which functional centrosomes are present but not in preferential proximity to either pronucleus). We conclude that increasing the distance between the centrosomes and the maternal pronucleus by preventing pronuclear migration can delay its permeabilization by as much as 80s. In the absence of centrosomes, both pronuclei breakdown synchronously about ~200s after the pronuclei in controls. Cumulatively, these data strongly argue that centrosomes generate a diffusible factor that promotes NEBD.
To determine if an ~80s delay is consistent with diffusion of a cytoplasmic signal across the embryo, we introduced a 10 kD photoactivatable dextran by gonad injection into embryos. 10 kD dextran has an effective hydrodynamic diameter of 10.6 nm (Lenart et al., 2003), similar to that of a globular protein complex of ~250 kDa. A signal produced by the sperm centrosomes in the nocodazole/dynein-inhibition experiment (Fig 6A,B) was simulated by using a pulse of UV light to photoactivate the dextran on one side of the embryo (Supplemental Fig. 3A). Diffusion of the activated dextran was monitored by plotting the equilibrium ratio (Supplemental Fig. 3B, C), which declines from 1 to 0 as the activated dextran comes to diffusional equilibrium. The equilibrium ratio dropped by 50%, over ~50s. This value is similar to the ~80s asynchrony between the breakdown of the centrosome proximal and distal pronuclei, indicating that diffusion of a centrosomally-generated cytoplasmic signal is a feasible mechanism for explaining this asynchrony.
The fact that depletion of aurora A delays NEBD to a significantly greater extent than completely inhibiting centrosome assembly, leads us to speculate that the diffusible factor generated by centrosomes that promotes NEBD is activated aurora A. This idea is consistent with the rapid turnover of aurora A previously documented in human cells (half-life = 3s; Stenoien et al., 2003). To confirm that centrosomal aurora A also turns over rapidly in the C. elegans embryo, we monitored the fluorescence recovery after photobleaching of centrosomal GFP:AIR-1 prior to NEBD (Fig. 6C,D). Centrosomal GFP:AIR-1 recovered to 95 + 24% of its initial value, with a half time of 11.8 + 2.1 seconds (n=8; errors are 95% confidence interval). We conclude that the centrosomal and cytoplasmic populations of aurora A are in rapid equilibrium, making aurora A an excellent candidate for the diffusible centrosomally-generated signal that promotes NEBD.
Previous work demonstrated that aurora A depletion appears to lengthen mitotic entry during the first division of the C. elegans embryo (Hannak et al., 2001). However, the lack of quantifiable cell cycle reference points prevented determination of the nature of this delay. Here, we use a recently developed method to monitor the kinetics of chromosome condensation (Maddox et al., 2006) to show that depletion of aurora A specifically lengthens the interval between the completion of chromosome condensation and NEBD. By contrast, the onset and kinetics of chromosome condensation and regression of the pseudocleavage furrow, a cortical event analogous to the Cdk1-stimulated relaxation of surface contractile waves in Xenopus embryos (Rankin and Kirschner, 1997), were not altered by aurora A depletion. These results indicate that the delay in nuclear envelope permeabilization that we describe here is distinct from the delay in global Cdk1 activation observed following depletion of aurora A in human cells and Xenopus extracts (Hirota et al., 2003; Liu and Ruderman, 2006). It is important to note that our data do not rule out the possibility that aurora A might also have a role in Cdk1 activation in C. elegans. When the dsRNA against aurora A is introduced into L4 stage hermaphrodites, embryo production ceases after ~26 hours. By contrast, control hermaphrodites (as well as hermaphrodites in which other essential cell division proteins are depleted) continue embryo production for more than 48 hours. We therefore analyzed aurora A depleted embryos produced between 22 and 26 hours after dsRNA injection. These embryos are ~90% depleted of aurora A by western blotting (Hannak et al., 2001), but are hypomorphic and not null for aurora A function. As RNAi of the C. elegans homolog of Cdk1 also leads to cessation of embryo production (Boxem et al., 1999), further work will be needed to determine if the sterility following depletion of aurora A reflects an additional role for this kinase in Cdk1 activation during oocyte maturation.
Our results provide three lines of evidence supporting the idea that centrosomes generate a diffusible factor that promotes nuclear envelope permeabilization. First, analysis of nocodazole-treated and dynein depleted embryos indicates that it is the presence of the mitotic centrosome scaffold rather than mechanical interactions between centrosomal microtubules and the nuclear envelope that is critical for timely NEBD. Second, depletion of ZYG-12, which causes the centrosomes to dissociate from the nuclear envelope and be randomly positioned within the embryo, does not delay NEBD to the same extent as completely inhibiting centrosome assembly. This result indicates that centrosomes can act at a distance to promote NEBD. Third, analysis of the asynchrony in NEBD between the sperm and oocyte nuclei when pronuclear migration is inhibited indicates that the breakdown of the oocyte nucleus, which is further from the centrosomes, is significantly delayed relative to breakdown of the sperm nucleus, which is proximal to the centrosomes. A comparison with the rate of diffusion of photoactivated 10 kD dextran indicates that the magnitude of this delay is consistent with the time required for diffusion of a centrosomally-generated factor that promotes NEBD.
As depletion of aurora A delays NEBD to a significantly greater extent than completely inhibiting centrosome assembly, we speculate that the diffusible factor generated by centrosomes that promotes NEBD is activated aurora A. We propose that the pericentriolar material of the centrosome catalyzes aurora A activation (Fig. 7A) and activated aurora A in turn, either directly or indirectly, promotes NEBD. The idea that events at mitotic centrosomes could affect the cytoplasmic pool of aurora A is supported by the rapid turnover of aurora A at centrosomes. Such a mechanism requires a centrosomally-localized aurora A activator. TPXL-1, the well-characterized aurora A activator, is important for spindle assembly after NEBD, but does not contribute to the functions of aurora A prior to NEBD (Özlü et al., 2005; this study). In human cells, the LIM protein ajuba is proposed to activate aurora A during mitotic entry (Hirota et al., 2003). However, homologs of ajuba have not been identified outside of vertebrates. A recently described aurora A activator, Bora, does not localize to centrosomes (Hutterer et al., 2006) and its inhibition in multiple genome-wide RNAi screens is reported to not affect C. elegans embryo viability (Kamath et al., 2003; Rual et al., 2004; Sönnichsen et al., 2005). We therefore suspect that an unidentified activator on the centrosome scaffold is important for aurora A function prior to NEBD.
Overall, our analysis supports the emerging picture in which the functions of aurora A prior to and after NEBD are mediated by distinct sets of activators/effectors. It will be interesting to see if the same is true for the different events that require aurora A in the period leading up to NEBD: Cdk1 activation (Hirota et al., 2003; Liu and Ruderman, 2006), centrosome maturation (Berdnik and Knoblich, 2002; Blagden and Glover, 2003; Brittle and Ohkura, 2005; Hannak et al., 2001), the establishment of cell polarity (Berdnik and Knoblich, 2002; Hutterer et al., 2006; Schumacher et al., 1998), and nuclear envelope breakdown itself (this study).
Depletion of aurora A delays the onset of nuclear envelope permeabilization, lamin disassembly, and nuclear pore removal to an essentially identical extent (Fig. 3). Although it is possible that aurora A regulates all three events independently, the coordinate delay suggests that aurora A might regulate one event that is a prerequisite for the other two. Since nuclear pore disassembly is accompanied by phosphorylation of a subset of the nucleoporins (Belgareh et al., 2001; Favreau et al., 1996; Ganeshan and Parnaik, 2000; Macaulay et al., 1995; Onischenko et al., 2005), an attractive possibility is that aurora A either directly or indirectly triggers phosphorylation of nuclear pore components, promoting envelope permeabilization and subsequent pore removal and lamin disassembly. Alternatively, the coordinate regulation by aurora A could be explained by its control of an upstream event required for all three aspects of nuclear envelope breakdown. One mechanism is suggested by the observation that cyclin B1 accumulates in the nucleus after chromosome condensation but prior to changes in nuclear envelope permeability (Hagting et al., 1999, Terasaki et al., 2003). This nuclear accumulation is regulated by phosphorylation (Hagting et al., 1998, 1999; Toyoshima et al., 1998; Yang et al., 1998) and may reflect a role for nuclear-localized Cdk1-cyclin B1 in triggering NEBD, although direct evidence for this idea is currently lacking. Assuming that the translocation of Cdk1-cyclin B1 into the nucleus is important for NEBD, aurora A could promote NEBD by regulating the balance of Cdk1-cyclin B1 import and export during prophase. Distinguishing between these and other possibilities will require future work.
Mitotic progression is directed by the sequential activation of cyclin dependent kinases. However, how the many events that occur during mitotic entry are coordinated remains an important question. Aurora and polo family kinases refine the broad strokes of Cdk regulation to ensure that mitotic events occur in their proper sequence (Barr et al., 2004; Crane et al., 2004; Ducat and Zheng, 2004; Dutertre et al., 2002; Lowery et al., 2005; Marumoto et al., 2005). The fact that aurora A is important for both centrosome maturation and NEBD suggests the existence of a positive feedback loop between an aurora A-dependent increase in centrosome size during mitotic entry and the cellular pool of activated Aurora A that, either directly or indirectly, promotes NEBD. This idea provides an attractive explanation for how the temporal coupling between centrosome maturation, (an event that occurs in preparation for chromosome segregation on the spindle), and NEBD, (which provides centrosomal microtubules access to the replicated chromosomes), is achieved. Our findings also lend support to the emerging idea that the centrosome acts as a signaling scaffold (Doxsey et al., 2005), coordinating the progression of mitotic events independently of its role as a microtubule nucleating and organizing center.
C. elegans strains expressing GFP:histone (AZ212; Praitis et al., 2001), GFP:NMY-2 (JJ1473; Nance et al., 2003), and co-expressing GFP:γ-tubulin and GFP:histone (TH32; Cheeseman et al., 2004) were maintained at 20°C. The strains OD139, co-expressing RFPmCherry:histone H2B and YFP:LMN-1 (unc-119(ed3) III; ltIs37[pAA64; unc-119(+) pie-1/RFPmCherry::HIS-58]; qals3507 [unc-119(+) pie-1/YFP::LMN-1]), and OD141, co-expressing RFPmCherry:histone H2B and GFP:NUP-155 (unc-119(ed3) III; ltIs37[pAA64; unc-119(+) pie-1/RFPmCherry::HIS-58]; [unc-119(+); pie-1/GFP::NUP-155]), were generated by mating previously described strains expressing fluorescent fusions with LMN-1 (Galy et al., 2003), NUP-155 (Franz et al., 2005), and histone H2B (McNally et al., 2006) and were maintained at 25°C. The strain OD142 expressing a GFP fusion with AIR-1 was generated by cloning the spliced AIR-1 coding sequence into the Spe I site of pIC26 (Cheeseman et al., 2004), and integrating the construct into DP38 (unc-119 (ed3)) by ballistic bombardment (Praitis et al., 2001).
dsRNA was prepared as described (Oegema et al., 2001). DNA templates were generated using the following gene-specific primers: smc-4 (TAATACGACTCACTATAGGCTCCAAAACAAGCCGAACTT, AATTAACCCTCACTAAAGGTGCATCTTCTTCTTTCCCTACA), air-1 (TAATACGACTCACTATAGGGCCTCTCGGAAAAGGAAAGT, AATTAACCCTCACTAAAGGCCTTGATTCTGGCGATCAAT), spd-2 (AATTAACCCTCACTAAAGGTGCATGCGAATAAGACGAAG, TAATACGACTCACTATAGGTTGCGGACACAGAAAACAAA), spd-5 (AATTAACCCTCACTAAAGGTGTCGCAACCAGTTCTGAAT, TAATACGACTCACTATAGGATGGAGGCAAATTGTTGCTG), dhc-1 (AATTAACCCTCACTAAAGGGAAGGAAGGAGCTCAACGACA, TAATACGACTCACTATAGGCCTTTCCTTCCTGGGTCTTC), zyg-12 (AATTAACCCTCACTAAAGGGACGGCTGGCTTGAAACAATG, TAATACGACTCACTATAGGGCAACTGAGCAATCCCATTT) to amplify regions of genomic N2 DNA. For depletion of aurora A, the previously described RNAi conditions that led to >90% depletion were used (Hannak et al., 2001). For depletion of DHC-1 and ZYG-12, worms were incubated at 20°C for 24–28 hours post injection. For all other depletions, injected L4 larvae were incubated at 20°C for 48 hours prior to analysis of their embryos.
Embryos expressing GFP:histone were imaged at 20°C using a Nikon E800 upright microscope (Nikon Instruments, Melville, NY) equipped with a 60x 1.4NA Plan Apo objective lens and an Orca ER CCD camera (Hamamatsu Photonics, Bridgewater, NJ) without binning. For condensation analysis, embryos were dissected and mounted in M9 (22 mM KH2PO4, 42 mM Na2HPO4, 86 mM NaCl, 1 mM MgSO4) on a 2% agarose pad under a coverslip. At 10s intervals, 5 z-sections at 2 μm steps were acquired using a 250 ms exposure and a 12.5% transmission neutral density filter. Presented images are maximum intensity projections of the z-series for the indicated time points. Condensation kinetics were analyzed as described previously (Maddox et al., 2006). For measurement of pronuclear diameter, embryos were mounted prior without compression (Monen et al., 2005). Prior to anaphase of meiosis II, 3 fluorescence z-sections at 2 μm steps were collected at 30 s intervals. After anaphase of meiosis II, 5 DIC (Differential Interference Contrast) z-sections at 2 μm steps were collected at 10s intervals until pronuclear meeting. After pronuclear meeting, 5 fluorescence z-sections at 2 μm steps were collected at 10s intervals until metaphase of mitosis.
Embryos from the strains OD139 and OD141 were imaged using a spinning disk confocal (McBain Instruments, Los Angles, CA) mounted on a Nikon TE2000e inverted microscope (Nikon Instruments, Melville, NY). Images were acquired using a 60x 1.4NA Plan Apo objective lens with 1.5x auxiliary magnification using an Orca ER CCD camera (Hamamatsu Photonics, Bridgewater, NJ) with 2x2 binning. Acquisition parameters, shutters and focus were controlled by MetaMorph software (Universal Imaging, Downingtown, PA). YFP:Lamin at the nuclear periphery was quantified by averaging the total fluorescence intensity measured in six separate 5x5 pixel regions at the periphery of the sperm pronucleus for each timepoint.
Texas-Red conjugated, lysine fixable 70kD dextran (Molecular Probes) at 0.08 mg/mL in injection buffer (1 mM potassium citrate, 6.7 mM KPO4, pH 7.5, 0.67% PEG) was injected into the gonads of control and air-1(RNAi) worms expressing GFP:NMY-2. After 4 hours, dextran-containing embryos were dissected from the mothers and imaged using a spinning disk confocal as described for imaging of embryos from the strains OD139 and OD141 except 4x4 binning was used. Entry was defined as the point when the nuclear and cytoplasmic fluorescence of the probe had equilibrated.
To introduce nocodazole into embryos, 3 worms were dissected in 8μL egg salts (48mM NaCl, 118 mM KCl) on a 24x60 coverslip on which a drop of 4μL polylysine (1 mg/mL) had been dried in an oven for 10 minutes. The buffer was removed with a mouth pipet and replaced with 8μL of 9:1 ddH2O:bleach by volume. After 2 minutes, the bleach solution was replaced with 8μL egg salts buffer, followed sequentially by 8μL L-15 blastomere culture medium (Edgar, 1995) and 8μL chitinase (5U/mL in L-15 blastomere culture medium). After 4 minutes, the chitinase solution was replaced with 8μL L-15 blastomere culture medium containing 10μg/mL nocodazole. Embryos were filmed as described above for measurement of pronuclear size except 5 fluorescence sections were acquired at 2μm z steps every 10s.
For photobleaching of centrosomal GFP:AIR-1, embryos were mounted without compression (Monen et al., 2005) in L-15 blastomere culture medium containing 10μg/mL nocodazole to minimize centrosome movement. Embryos were imaged using a spinning disk confocal (McBain Instruments, Los Angles, CA) mounted on a Nikon TE2000e inverted microscope (Nikon Instruments, Melville, NY). Images were acquired by using a 60 × 1.4 N.A. Plan Apo objective lens with a Xion electron multiplication back-thinned CCD camera (Andor technologies) with no binning. Photobleaching was performed by selecting the 488nm line from a Krypton-Argon mixed gas 2.5W water cooled laser (Spectra-Physics, Mountain View, CA) which was steered into a custom modified epi-fluorescence port creating a single diffraction limited spot at the field diaphragm which is projected to the objective focal plane (full width half max at the focal point ~800nm). Exposure times for bleaching were 1–2 seconds. Acquisition parameters, shutters, and focus were controlled by MetaMorph software (Universal Imaging, Downingtown, PA). The fluorescence intensity of centrosomal GFP:AIR-1 was calculated by measuring the total intensity in a box containing the centrosome and subtracting the camera background. The signal at the first postbleach timepoint (typically 15–25% the pre-bleach value) was subtracted from all postbleach measurements and the fraction of fluorescence recovered at each timepoint was calculated by dividing by the difference between the prebleach and first postbleach measurements. Kaleidagraph software (Synergy Software, Reading PA) was used to fit the data to an equation of the form y= m(1−exp(−k*t)) where m is the maximum fractional recovery and the half time for recovery t1/2= −ln(0.5)/k.
KO is a Pew Scholar in the Biomedical Sciences; AD is the Connie and Bob Lurie Scholar of the Damon Runyon Cancer Research Foundation (DRS 38-04). AA is a fellow of the Helen Hay Whitney Foundation; PSM is the Fayez Sarofim Fellow of the Damon Runyon Cancer Research Foundation (DRG-1808-04); RAG is a fellow of the American Cancer Society. Some nematode strains used in this work were provided by the Caenorhabditis Genetics Center, which is funded by the NIH National Center for Research Resources (NCRR). KO and AD receive salary and additional support from the Ludwig Institute for Cancer Research.