Requirement of Cdk2-Cyclin E Activity for Repeated Centrosome Reproduction in Xenopus Egg Extracts

See allHide authors and affiliations

Science  05 Feb 1999:
Vol. 283, Issue 5403, pp. 851-854
DOI: 10.1126/science.283.5403.851


The abnormally high number of centrosomes found in many human tumor cells can lead directly to aneuploidy and genomic instability through the formation of multipolar mitotic spindles. To facilitate investigation of the mechanisms that control centrosome reproduction, a frog egg extract arrested in S phase of the cell cycle that supported repeated assembly of daughter centrosomes was developed. Multiple rounds of centrosome reproduction were blocked by selective inactivation of cyclin-dependent kinase 2–cyclin E (Cdk2-E) and were restored by addition of purified Cdk2-E. Confocal immunomicroscopy revealed that cyclin E was localized at the centrosome. These results demonstrate that Cdk2-E activity is required for centrosome duplication during S phase and suggest a mechanism that could coordinate centrosome reproduction with cycles of DNA synthesis and mitosis.

In animal cells, the interphase centrosome reproduces or duplicates only once per cell cycle, thereby ensuring a strictly bipolar mitotic spindle axis (1). Because there is no cell cycle checkpoint that monitors the number of spindle poles (2), uncontrolled duplication of the centrosome can contribute to genomic instability through the formation of multipolar mitotic spindles. Indeed, many human tumor cells, including those lacking the tumor suppresser protein p53 (3), have abnormally high numbers of centrosomes (4).

Studies of sea urchin and Xenopus embryos and clam oocyte lysates have revealed that the centrosome cycle can be regulated solely by cytoplasmic mechanisms (5–8): The repeated duplication of the centrosome proceeds in the complete absence of either a nucleus (7) or protein synthesis (8). In theory, the cyclical rise and fall in the activity of one or more cyclin-dependent kinases (Cdks) could be the cytoplasmic mechanism that coordinates centrosome reproduction with cell cycle progression. However, the fact that centrosomes repeatedly duplicate in the complete absence of protein synthesis indicates that the activities of those Cdks that are dependent on the translation of their cyclin subunits during each cell cycle (that is, Cdk1–cyclin A or –cyclin B or both) do not regulate centrosome reproduction or assembly (8). Nevertheless, Cdk2–cyclin E (Cdk2-E) remains a potential candidate to control centrosome duplication and coordinate it with nuclear events during the cell cycle (6, 9,10). Cdk2-E activity drives the transition from G1 to S phase in somatic cells (11), which is the time during the cell cycle when daughter centrosome assembly is thought to begin (12). Importantly, in earlyXenopus embryos, Cdk2-E activity is not dependent on the synthesis and degradation of the cyclin E subunit, as the amount of cyclin E remains constant until the mid-blastula transition (MBT) (13).

To investigate whether Cdk2-E activity regulates centrosome duplication, we developed an S phase–arrested Xenopus egg extract that supports repeated centrosome reproduction in vitro. We used an S phase extract because centrosomes will undergo multiple rounds of duplication during S phase arrest in both zygotes and somatic cells (6, 8, 14, 15). Unlike cycling extracts, Cdk2-E activity can be inhibited in S phase–arrested extracts without the concern that this inhibition will block cell cycle progression at a point before centrosomes are normally scheduled to reproduce. To make these extracts, we prepared a cyclingXenopus egg extract (16, 17) and then added aphidicolin, an inhibitor of α-DNA polymerase (18), and demembranated Xenopus sperm nuclei (19). Histone H1 kinase activity in control extracts cycled at least twice with a cell cycle time of ∼50 min; in contrast, H1 activity in aphidicolin-treated extracts remained at a constant, low amount for 6 hours (20, 21). Time-lapse videomicroscopy of aphidicolin-treated extracts revealed that nuclear envelope breakdown did not occur during the 6-hour experiment (20). Thus, our extracts are arrested in S phase because of the activation of the cell cycle checkpoint that monitors the completion of DNA synthesis.

We characterized the pattern of centrosome reproduction in aphidicolin-treated extracts with time-lapse videomicroscopy. We used polarization optics to directly visualize the astral microtubules organized by the sperm centrosomes and found that aster number increased over 6 hours (Fig. 1A). BecauseXenopus extracts do not spontaneously assemble microtubule asters or centrosomes in the absence of added sperm nuclei (22), this increase in aster number indicates centrosome doubling. The total number of asters in the field eventually declined, because some asters moved out of the plane of focus or off the field of view (Fig. 1A, panel d). Figure 1B shows an individual aster from another extract that doubled three times, its daughters remaining fortuitously close together and in the plane of focus.

Figure 1

(A) Repeated rounds of aster doubling in an aphidicolin-treated extract. Frames from a time-lapse video sequence, showing the increase in aster number over time in a microscope field. The decrease in aster number in (d) is due to the migration of asters from the plane of focus and field of view. Minutes after addition of sperm nuclei are seen in the lower right corner of each frame. Polarization optics. Ten micrometers per scale division. (B) Time-lapse sequence showing an individual aster from another aphidicolin-treated extract undergoing three rounds of doubling. (a) The aster at the start of the time-lapse sequence. (b and c) Doubling of this aster and separation of the daughter asters. (d) Doubling of these daughters. (e) Third round of doubling yields eight asters. Minutes after addition of sperm nuclei are seen in the lower right corner of each frame. Polarization optics. Bar in (e), 10 μm.

We analyzed the pattern of doubling for all asters in a given field in aphidicolin-treated extracts and scored those that had at least one first, second, or third generation daughter aster visible for the duration of the 6-hour experiment. Of the individual asters for which we obtained a complete lineage analysis (N = 59), 5% completed four rounds of duplication, 69% underwent three rounds, 14% showed two rounds, 2% showed a single round, and 10% did not duplicate. Asters whose progeny were all lost from view during the experiment were not included. Nevertheless, Table 1 shows the complete data set for aster duplication at each successive round, regardless of whether or not individual asters subsequently became lost from the field. The multiple rounds of aster doubling in a strict one-to-two fashion at each cycle (Fig. 1B) are characteristic of complete centrosome reproduction and cannot be explained by either centrosome splitting or fragmentation of the microtubule organizing center (23). Together, our results define a cell-free system that allows the real-time observation of the complete centrosome reproductive cycle in vitro.

Table 1

Aphidicolin-treated extracts followed for 6 hours.

View this table:

To directly test whether Cdk2-E activity was required for repeated centrosome reproduction during S phase arrest, we selectively inactivated Cdk2-E by addition of recombinant Δ34Xic1, a 34–amino acid NH2-terminus truncated variant of Xic1p27, a Xenopus cdk inhibitor (24, 25). Δ34Xic1 inhibits the activity of Cdk2-E with a median inhibitory concentration (IC50) value of 10 nM and only affects Cdk1–cyclin A and –cyclin B activity at higher concentrations (IC50 = 5 μM) (24). Because Cdk2 does not complex with cyclin A until after the MBT (9,13, 26), Cdk2–cyclin A activity was not a factor in this study. Inhibition of Cdk2-E does not drive the cell cycle out of S phase because the majority of S phase–promoting activity is provided by Cdk1–cyclin A (26), which is not inhibited in vitro by Δ34Xic1 at the concentration used here (9, 24).

Extracts were prepared and split into two portions. To the first of these, C-Xic1 (the COOH-terminal half of Xic1) was added as a control because it does not affect the activity of Cdk2-E in vitro (9, 24). To the second portion, we added Δ34Xic1 (175 nM final concentration) and then separated this sample into two further portions. To the first of these, buffer was added, and to the second, we added an equal volume of baculovirus-expressed, affinity-purified active Cdk2-E complex (245 nM final concentration, equal to 1.4 times the molar amount of Δ34Xic1 added) (9,13, 24, 27). The pattern of aster doubling under each of these three conditions was then simultaneously recorded with separate videomicroscopy systems (Fig. 2, A and B).

Figure 2

(A) Selectively inhibiting Cdk2-E activity blocks repeated centrosome reproduction. Time-lapse sequence showing one round of aster doubling in an aphidicolin-treated extract containing 175 nM Δ34Xic1. Minutes after addition of sperm nuclei are seen in the lower right corner of each frame. Polarization optics. Ten micrometers per scale division. (B) Time-lapse sequence showing the restoration of multiple rounds of aster doubling in an aphidicolin-treated extract containing 175 nM Δ34Xic1 plus 245 nM Cdk2-E. Minutes after addition of sperm nuclei are seen in the lower right corner of each frame. Polarization optics. Ten micrometers per scale division.

Analysis of the behavior of the asters in extracts treated with Xic1-C (N = 31 asters) revealed that 77% of the asters completed three rounds of doubling and 23% doubled twice. In the Δ34Xic1-treated extracts, 6% of the asters (N = 53) failed to double at all, 79% doubled only once, 15% doubled twice, and none doubled three times over 6 hours.

When purified active Cdk2-E was added back to Δ34Xic1-treated extracts, 50% of the asters (N = 52) underwent two and 42% underwent three rounds of duplication, whereas only 8% of the asters doubled just once over 6 hours. Table 2 shows the complete data set for all asters followed under these three conditions regardless of whether any daughter asters subsequently became lost from view. These results reveal that the activity of Cdk2-E is required for centrosomes to undergo repeated reproduction during S phase arrest.

Table 2

Aphidicolin-treated extracts followed for 6 hours.

View this table:

We characterized the localization of cyclin E in Xenopusembryonic blastomeres by confocal immunofluorescence microscopy, using an affinity-purified antibody to cyclin E (9,28). Cyclin E was found to be distributed diffusely throughout the cytoplasm but showed maximal concentration at the centrosome region (Fig. 3). Thus, the Cdk2-E complex may become accumulated at the centrosome, as has been reported for other Cdk-cyclin complexes (29).

Figure 3

Localization of cyclin E in Xenopusembryonic blastomeres by immunofluorescence microscopy. Shown are three separate cells double immunostained with antibody to cyclin E (A through C) and antibody to α tubulin (α-Tub) (a and b) or antibody to γ tubulin (γ-Tub) (c). In (B), the antibody to cyclin E was preabsorbed to the initial cyclin E antigen. A + a, B + b, and C + c represent merged images. Centromeres are marked by arrows (A and C). Confocal microscopy. Bars in (A), (B), and (C), 10 μm.

Previously, it was demonstrated that only a single round of daughter centrosome assembly can occur in sea urchin zygotes arrested before the onset of S phase, whereas repeated rounds of duplication occur during S phase arrest (6). Thus, the morphological events of daughter centrosome assembly can occur before the G1-S transition, but entry into S phase appears to be necessary for the centrosome to duplicate again. This suggests that a “licensing” event during S phase restores the reproductive capacity to the daughter centrosomes, thereby permitting them to duplicate again during the next cell cycle. Here, our finding that the inactivation of Cdk2-E does not inhibit the first round of aster doubling in vitro, but does block further rounds (Fig. 2A and Table 2), suggests that the daughter centrosomes cannot reduplicate in the absence of Cdk2-E activity. Perhaps Cdk2-E is the “licensing” factor that restores the reproductive capacity to the daughter centrosomes. In this regard, it is important to determine if the abnormal centrosome number observed in both mouse embryonic fibroblasts lacking p53 (3) and many human tumor cells (4) is due to the misregulation of Cdk2-E activity at the G1-S transition and during S phase.

  • * These authors contributed equally to this work.

  • To whom correspondence should be addressed. E-mail: Greenfield.Sluder{at}


View Abstract

Stay Connected to Science

Navigate This Article