Proteolysis and DNA Replication: The CDC34 Requirement in the Xenopus Egg Cell Cycle

See allHide authors and affiliations

Science  12 Sep 1997:
Vol. 277, Issue 5332, pp. 1672-1676
DOI: 10.1126/science.277.5332.1672


The cell division cycle gene, CDC34, is required for ubiquitin-mediated degradation of G1 regulators and cell cycle progression through the transition from G1 to S phase in budding yeast. A CDC34 requirement for S phase onset in higher eukaryotes has not been established. Studies of the simple embryonic cell cycle of Xenopus laevis eggs demonstrated that Cdc34p in a large molecular size complex was required in the initiation of DNA replication. Cdc34p appears to regulate the initiation function of Cdk2–cyclin E, perhaps through the degradation of the Xenopus cdk inhibitor, Xic1.

Protein ubiquitination and degradation were linked to cell cycle control by the discovery that Cdc34p, an essential G1- to S-phase regulator in budding yeast, is a ubiquitin conjugating enzyme (UBC3) (1). The Cdc34p-dependent degradation of a single substrate, p40sic1p [an inhibitor of cyclin-dependent kinase (cdk) Cdc28p complexed with cyclins Clb5p or Clb6p], appears to be sufficient to trigger the transition from G1 to S phase (2). This degradation requirement in yeast is a key element coupling extrinsic control of cell proliferation to the cell cycle. The human homolog ofCDC34 can complement the yeast cdc34-2temperature-sensitive strain (3). Cdc34p in mammalian cells may also function to couple extrinsic control by growth factors to cell cycle progression. However, Cdc34p may play a more basic role in the intrinsic control of each cell cycle. A requirement for Cdc34p in the early embryonic cell cycle of the frog Xenopus laevis, where cell division is independent of extrinsic controls, may reveal a more fundamental role for protein degradation in regulating the initiation of DNA replication.

We tested whether Cdc34p is required for the onset of DNA replication in higher eukaryotes and examined the nature of such a requirement using an in vitro DNA replication system derived fromXenopus eggs (4). When nuclei are added toXenopus interphase egg extracts, they faithfully reproduce initiation events preceding DNA replication and undergo semi-conservative chromosomal replication, using all of the known eukaryotic replication initiation factors (5). Despite a lack of G1 control before the midblastula transition (MBT),Xenopus embryos may still share a core regulatory requirement for DNA replication, and hence, the same Cdc34p requirement.

To test whether general proteolysis via the proteasome is required for DNA replication, we measured chromosomal replication in interphase egg extracts in the presence of inhibitors of proteasome function. Methyl ubiquitin inhibited DNA replication of added sperm nuclei by 80% at 1 mg/ml, whereas the peptide inhibitorN-acetyl-leu-leu-norleucinal inhibited DNA replication by 70% at 0.3 μM (6). This suggests that degradation via ubiquitination and the proteasome may be required for efficient DNA replication in Xenopus.

We cloned the Xenopus and mouse homologs of the humanCDC34 gene and found that human, Xenopus, and mouse Cdc34p shared >92% amino acid identity (6). In budding yeast, a mutant yeast Cdc34p, in which the active site cysteine and downstream leucine are changed to serine, blocks cell growth and in vitro ubiquitination of Cln2p (7). We generated an analogous pair of mutations in human Cdc34p (hCdc34p). This mutant hCdc34p (CL→S) inhibits the in vitro stability and ubiquitination of the mammalian cdk inhibitor, p27, and thus acts as a dominant negative mutant (8).

To determine whether Cdc34p is required for the onset of S phase, recombinant CL→S was added to Xenopus interphase egg extracts (also referred to as low-speed supernatant or LSS), and DNA replication of added sperm nuclei was measured. The addition of CL→S to LSS inhibited chromosomal DNA replication by ∼90% at 3.3 μM (Fig. 1A). This concentration is approximately eight times higher than the endogenous Cdc34p concentration (400 nM) in LSS. At 3.3 μM, the wild-type (WT) hCdc34p showed little inhibition of DNA replication. Addition of CL→S and WT hCdc34p had little effect on DNA replication when an M13 single-stranded DNA (ssDNA) template was used (Fig. 1A). M13 ssDNA replication is considered to be a model for the elongation step in DNA replication because it requires neither nuclear formation around the DNA nor initiation events (9).

Figure 1

Requirement of Cdc34p for initiation of chromosomal replication. (A) WT and CL→S hCdc34p (8) were added to LSS (16, 17) to concentrations of 0.9, 1.7, 3.3, or 6.3 μM and incubated for 20 min at 23°C before the addition of sperm nuclei (left panel) (18). WT and CL→S were also added to LSS to concentrations of 1.1, 2.8, or 6.3 μM 20 min before the addition of ssM13 DNA (13) (right panel). DNA replication was normalized to 100% of that in the buffer control (Buffer) (18). (B) Replication of nuclei was measured over 240 min in LSS with no additions (left) or after 240 min with XB– (23) (□) or 3.5 μM CL→S (•) (right). (Right panel) XB– or CL→S was added at time points between –45 and +70 min. Samples were incubated at 0°C (–45 min) or at 23°C (all other samples), and nuclei were added at time zero. (C) Left panel: LSS was not depleted (No depl) or depleted with pA-seph coupled to preimmune Ig (Preimm), rabbit Ig (RIg) (Ctrl), rabbit 1 anti-hCdc34p Ig (R1 Cdc34), or rabbit 2 anti-hCdc34p Ig (R2 Cdc34), and either sperm nuclei or ssM13 DNA was added (18,19). Replication is normalized to 100% of that in the control-depleted sample (Ctrl). (Upper right) Anti-Cdc34p immunoblot of depleted LSS (20). (Lower right) Sperm nuclei visualized by phase contrast (a), Hoechst staining (b), or rhodamine-strepavidin staining of biotinylated deoxyuridine triphosphate incorporation (c) after incubation in LSS depleted with pA-Seph beads coupled to RIg (RIg pA-Seph) (Ctrl depl) or anti-hCdc34p Ig (Cdc34 depl) (21). Images were acquired on a Zeiss Axiophot with an integrating color video camera, using a Northern Exposure Software Phase 3 Imaging System (Milford, Massachusetts). (D) (Left) Relative histone H1 kinase activity (16) of fertilized embryos over time where embryos were fertilized at time zero (Fert), placed into cycloheximide buffer at 68 min (CHX), and harvested at 95 min (Harvest). (Right) DNA replication in extracts prepared from fertilized embryos with sperm nuclei and added XB– (Buffer) or CL→S to 3.5 μM (CL-S 3.5 μM) after 2 or 4 hours (18).

To further distinguish a requirement for Cdc34p in the initiation of replication from a requirement in elongation, we varied the time of addition of CL→S. The CL→S mutant was added to LSS to a concentration of 3.5 μM before and after the addition of sperm nuclei (time zero). We began to observe incorporation of [α-32P]deoxyadenosine triphosphate (dATP) into DNA 30 min after the addition of sperm nuclei, and by 40 min, about 7% of the total [α-32P]dATP was incorporated. When the CL→S was added after 30 min, it did not inhibit DNA replication, despite the completion of only a small fraction of the total replication (Fig. 1B).

Because LSS is prepared from eggs released from metaphase arrest, they are in the interphase of the first embryonic cell cycle which might exhibit special requirements for replication. We prepared extract from fertilized eggs in the interphase of the second cell cycle to test whether they too exhibited a Cdc34p requirement for DNA replication. Results showed that addition of 3.5 μM CL→S to this extract inhibited DNA replication (Fig. 1D), indicating that the requirement for Cdc34p is not limited to the first Xenopus cell cycle.

To further investigate the Cdc34p requirement for the onset of DNA replication, we attempted to immunodeplete egg extracts of Cdc34p. Antibodies generated against hCdc34p recognized a single protein band in extracts from mammalian cells and Xenopus. Protein immunoblotting of the depleted extracts showed >95% removal of Cdc34p (Fig. 1C). Immunodepletion of Cdc34p from LSS reduced chromosomal replication by 90 to 95% with no effect on M13 ssDNA replication (Fig.1C). Localized sites of biotin-labeled nucleotide incorporation (replication foci) were visible in control-immunodepleted LSS, whereas such foci were not observed in extracts depleted of Cdc34p (Fig. 1C). We conclude that there is an early requirement for Cdc34p, closely associated with the initiation of DNA replication, and that Cdc34p has no role in the elongation phase of replication.

Attempts to rescue DNA replication in Cdc34p-immunodepleted extracts with recombinant Cdc34p were unsuccessful (Fig.2A), indicating that other proteins may have been depleted along with Cdc34p. Cdc34p-immunodepleted extracts could be rescued by readdition of LSS or high-speed supernatant (HSS; supernatant of LSS after a high-speed centrifugation) (Fig. 2A), indicating that extracts were not irreversibly inhibited by the immunodepletion process. Anti-Cdc34p–coupled beads incubated in LSS, washed with 500 mM KCl buffer, and added to Cdc34p-immunodepleted extracts restored DNA replication to 60% of that in control samples. Control antibody-coupled beads treated similarly exhibited no rescuing activity. These results suggest that either the human Cdc34p does not functionally replace the Xenopus Cdc34p or additional components are required for initiation of DNA replication.

Figure 2

Presence of Cdc34p in a large complex required for the onset of DNA replication. (A) LSS was not depleted (No depl), control-depleted with RIg pA-Seph (Ctrl), or pA-Seph coupled to anti-hCdc34p Ig (Cdc34). LSS (3 μl) or HSS (3 μl) control-depleted with RIg pA-Seph was added to LSS (10 μl) depleted with pA-Seph coupled to anti-hCdc34p Ig (Cdc34 depl). Bacterially expressed (Bact) hCdc34p was added (50, 100, or 200 nM) to Cdc34p-depleted LSS. pA-affiprep (2 or 5 μl) coupled to RIg (Ctrl beads) or anti-hCdc34p Ig (Cdc34 beads) were incubated in LSS, washed, and added to Cdc34p-depleted LSS (22). DNA replication was normalized to 100% of that in the control-depleted sample (Ctrl). (B) LSS (10 μl) was not depleted (No depl), control-depleted (Ctrl), or CDC34p-depleted (Cdc34) and added to 8 μl pA-affiprep coupled to RIg (Ctrl beads) or anti-hCdc34p Ig (Cdc34 beads), incubated in LSS, and washed (22, 23). Cdc34p-depleted LSS (Cdc34 depl) was also incubated with pA-affiprep coupled to anti-hCdc34p Ig and incubated with column fractions (23). DNA replication was normalized to 100% of that in the control-depleted sample (Ctrl). Photographed gel strip (Anti-Cdc34) represents fractions immunoblotted with anti-Cdc34p Ig. The molecular sizes listed above represent the elution of standards used to calibrate the column. (C) LSS was incubated with pA-affiprep coupled to either RIg (Ctrl beads) or anti-hCdc34p Ig (Cdc34 beads), and the beads were washed, eluted, subjected to SDS-PAGE, and silver stained (22). Sizes of protein standards are listed to the left and Cdc34p is indicated. (Bottom panel) Immunoblot of eluted material with anti-CDC34 Ig.

To determine whether the Cdc34p-rescuing activity required for the onset of DNA replication is part of a multiprotein complex, S100 (diluted and centrifuged LSS) was fractionated by gel filtration, and the fractions were analyzed for their ability to rescue the Cdc34p-immunodepleted LSS. Immunoblot analysis showed that >90% of the Cdc34p eluted between 42 and 64 kD, whereas a small amount eluted at a larger apparent molecular size between 70 and 450 kD (Fig. 2B). After concentration on anti-Cdc34p beads, washing, and addition to immunodepleted extracts, all the rescuing activity was found to fractionate between 320 and 440 kD. These fractions contained about 5% of the Cdc34p. No rescuing activity was detected in the fractions containing the bulk of the Cdc34p. Therefore, Cdc34p in a complex of large molecular size is the functional Cdc34p required for chromosomal replication in egg extracts.

When Cdc34p was immunopurified from LSS onto anti-CDC34 beads, washed, eluted, and examined by SDS–polyacrylamide gel electrophoresis (SDS-PAGE), a 34-kD band corresponding to Cdc34p was observed by immunoblotting (Fig. 2C). In addition, two other bands of ∼180 and 220 kD were eluted from the anti-Cdc34p beads but not control beads. These two bands were also observed when S100 and gel filtration fractions 11 and 12 (Fig. 2B) were immunopurified on anti-Cdc34p beads (6). These two proteins copurify and coimmunoprecipitate with Cdc34p and are potential components of the complex.

Given that the p40sic1p inhibitor of the Cdc28-Clb5,6 kinase must be degraded to allow entry into S phase in yeast (2), we looked for evidence of a similar process in vertebrate cells. The analogous S phase regulator in Xenopus is probably the Cdk2–cyclin E kinase, which is required for the initiation of DNA replication in Xenopus egg extracts (10). When mutant hCdc34p was added to extracts to a concentration of 2.3 μM, DNA replication was reduced by ∼80% and 92% in experiments 1 and 2, respectively (Fig. 3). Baculovirus-expressed and affinity-purified XenopusCdk2–cyclin E restored replication to 85% (experiment 1, 63 nM) and 75% (experiment 2, 53 nM) of the control amount (Fig. 3), whereas addition of kinase inactive Xenopus Cdk2 with or without cyclin E did not restore DNA replication activity. Control extract with added Cdk2–cyclin E (34 to 63 nM) exhibited no increase in replication activity. When DNA replication was inhibited by 95% or more by immunodepletion of Cdc34p, addition of Cdk2–cyclin E only partially rescued replication activity. From these results, it appears likely that at least one function of Cdc34p is to directly or indirectly regulate Cdk2–cyclin E activity, perhaps by targeting an inhibitor of Cdk2–cyclin E for destruction.

Figure 3

DNA replication inhibited by mutant Cdc34p and restored by Cdk2–cyclin E. CL→S was added to a concentration of 2.3 μM, and baculovirus-expressed Cdk2–cyclin E (500 nM) (11, 24) was added to a concentration of 23, 34, 44, 53, and 63 nM. DNA replication was normalized to 100% of that in the control sample, which contained no CL→S and was equalized to the same final volume as all other samples with XB– (23).

The Xenopus cdk inhibitor Xic1 is related to both p21 and p27 in mammals, inhibits Cdk2–cyclin E activity, and inhibits DNA replication when added to Xenopus egg extracts (11). To test whether Xic1 may be a substrate of Cdc34p, Xic1 was in vitro translated (ivt) and added to LSS. Contrary to expectations, ivt Xic1 was stable in both control- and Cdc34p-depleted extracts (Fig. 4A). However, when sperm nuclei (2500 nuclei/μl) were added to extracts, ivt Xic1 was efficiently degraded in control-depleted extracts, with a half-life of 60 to 90 min (Fig. 4, A and B). Xic1 was not degraded in similar extracts that had been depleted of Cdc34p (Fig. 4A). The half-life of ivt Xic1 decreased with increasing numbers of nuclei. When 5000 nuclei/μl were added (the approximate concentration at the MBT), the half-life of total ivt Xic1 was 45 min (Fig. 4B). This result suggests that Xic1 may be a bona fide substrate of Cdc34p inXenopus eggs, but that its degradation may occur in a nucleus-dependent manner, perhaps requiring localization to the nucleus for Cdc34p-dependent degradation. We postulate that localized nuclear degradation by Cdc34p may target specific regulatory proteins for degradation at sites of initiation. Permeabilization of the nucleus during each mitosis could allow a pool of Cdc34p substrates to reenter the nucleus after each division without a requirement for protein synthesis.

Figure 4

Degradation of in vitro translated Xic1 in a Cdc34p- and nuclei-dependent manner. (A) LSS was either control-depleted with RIg pA-Seph (Ctrl depl) or pA-seph coupled to anti-hCdc34p Ig (Cdc34 depl). Sperm nuclei were not added (–Nuclei) or added to a concentration of 2500 nuclei/μl (+Nuclei). Ivt Xic1 was added to LSS and samples were incubated for 0 to 4 hours (25). (B) LSS was incubated without nuclei (0 nuclei/μl) or with 1000, 2500, or 5000 nuclei/μl with ivt Xic1, and the relative amount of Xic1 remaining was determined by phosphoimager (26). (C) LSS (Interphase) was incubated with ivt Xic1 for 3 hours with 2500 nuclei/μl, in XB– (Buffer) (23), XB– with 1 mg/ml ubiquitin (Ub), 1 mg/ml methyl ubiquitin (MeUb), or a mixture of the two (25). Ivt Xic1 was also incubated in LSS with 2500 nuclei/μl with XB– with 10% ethanol (Buffer) or 50 μg/ml aphidicolin in XB– with 10% ethanol (Aphid) for 3 hours (middle left panel) or with XB– with 5% dimethylsulfoxide (DMSO) (Buffer) or 4 mM 6-DMAP in XB– with 5% DMSO for 3 hours (middle right panel). LSS was treated with cyclin Δ90 for 40 min (Mitotic), and ivt Xic1 was added without (–Nuc) or with nuclei to 2500 nuclei/μl (Buffer). The samples were incubated for 0 to 3 hours, and one sample was treated with CIP (right panel).

The degradation of ivt Xic1 in LSS was inhibited by methyl ubiquitin and reversed by the addition of ubiquitin (Fig. 4C), confirming that ivt Xic1 is degraded by the ubiquitin degradation pathway. The degradation of ivt Xic1 appeared to be inhibited by 6-dimethylaminopurine (6-DMAP), an inhibitor of Cdk's and the initiation of DNA replication (12), but not by aphidicolin, an inhibitor of the elongation phase of DNA replication (Fig. 4C). This suggests that Cdk2 activity or initiation events, or both, may be required for Xic1 degradation, whereas elongation is not. Ivt Xic1 was not degraded in mitotic extracts with or without nuclei, indicating that its degradation may be cell cycle regulated (Fig. 4C). Ivt Xic1 from mitotic extracts had a lower mobility during SDS-PAGE which was reversed by treatment with calf intestinal phosphatase (CIP) (Fig. 4C).

These studies indicate that Cdc34p in a large molecular size complex is required for events involved in the initiation of DNA replication and that localized degradation of an inhibitor of Cdk2–cyclin E such as Xic1 and perhaps other S phase regulators may be required in each Xenopus cell cycle. We do not rule out the possibility that Cdc34p may also regulate the function of Cdk2–cyclin A2 in Xenopus interphase egg extracts. Genetic studies in yeast suggest that Cdc34p function may require a complex of Cdc34p, Cdc53p, Skp1p, and Cdc4p, although biochemical characterization is still incomplete (13). However, all the yeast components thus far identified are smaller than the polypeptides associated with Xenopus Cdc34p.

In the early Xenopus embryo, the cell cycle is specifically modified such that materials are stockpiled for at least 12 divisions and there is no G1 or G2 phase and no growth (14). Between fertilization and late cleavage stages, there is no external control of the cell cycle by mitogens and the pace is set by the complete degradation and resynthesis of cyclin B (15). Under these circumstances, a Cdc34p degradation requirement for DNA replication was not necessarily expected. Our results indicate that there are one or more proteolytic steps in the intrinsic control of DNA replication and that degradation is not simply a feature of extrinsic control, but that in higher eukaryotes it may be an obligatory step in each round of initiation of DNA replication.


View Abstract

Navigate This Article