Differential replication capacities of G1 and S-phase extracts from sea urchin eggs

The control of DNA replication in eukaryotic cells is not well understood. Most knowledge about the processes of DNA replication and its regulation is based on studies of bacterial and eukaryotic viral DNA replication (reviewed by Bramhill and Kornberg, 1988; Stillman, 1989; Challberg and Kelly, 1989; Tsurimoto et al., 1990). Studies of nuclear DNA replication in eukaryotic cells have been hampered by the lack of cell-free systems that can carry out DNA synthesis. So far, only extracts made from activated Xeno -pus eggs show the capacity to support semi-conservative DNA replication on exogenously added nuclei (Blow and Laskey, 1986; Blow and Watson, 1987; Hutchinson et al., 1987; Laskey et al., 1989). In that system, demembranated sperm nuclei undergo dramatic morphological changes before the start of DNA synthesis, including decondensa-tion of sperm chromosomes, re-formation of the nuclear envelope and swelling of the re-formed nuclei. It appears that the complete formation of the nuclear envelope is required for initiation of DNA synthesis (Lohka and Masui, 1983; Newport, 1987; Blow and Laskey, 1988; Sheehan et al., 1988; Blow and Sleeman, 1990; Leno and Laskey, 1991). Furthermore, extracts of G2-arrested oocytes and M phase-arrested eggs of Xenopus fail to re-form nuclei or support DNA synthesis, indicating that the frog system reproduces the cell cycle stage-specific regulation of DNA synthesis as well (Mechali and Harland, 1981; Cox and Leno, 1990).

Unlike frog oocytes, which arrest naturally at two points during meiosis, once at the G2/M border of meiosis I and later at metaphase of meiosis II, sea urchin eggs are natu-rally arrested in G1 of the first mitotic cell cycle (Hinegard-ner et al., 1964; Longo, 1973). Fertilization triggers a rapid series of metabolic changes and release from cell cycle arrest. DNA replication begins about 20 minutes after fer-tilization, even in the absence of ongoing protein synthesis, indicating that activation of DNA replication is due to the modification of pre-existing components (Hinegardner et al., 1964; Wagenaar, 1983). To study the regulation of this process, we have developed cell-free systems from unfertil-ized, G1-arrested eggs and fertilized, S-phase eggs that repro-duce their stage-specific capacities for DNA replication.

Lytechinus pictus and Strongylocentrotus purpuratus were obtained from Marinus (Long Beach, California) and were kept in a recirculating sea water aquarium at 16°C or 13°C, respec-tively. Lytechinus variegatus were obtained from Susan Decker (Miami, Florida) and kept at 22°C. Spawning was induced by the intracoelonic injection of 0.55 M KCl. Sperm was collected and stored at 4°C until use. Eggs were given several washes in sea water and then cultured in Jamarin artificial sea water (Jamarin Laboratory, Osaka, Japan) and 16°C (L. pictus and S. purpuratus) or 22°C (L. variegatus ) as described by Rosenthal et al. (1980).

L. variegatus eggs were dejellied by gentle swirling in sea water; S. purpuratus eggs were dejellied by brief exposure to a buffer containing 0.027 M KCl, 0.512 M NaCl, 0.002 M Hepes, pH 8.0, 1 mM EDTA. L. pictus eggs were dejellied in the same solution except that the concentration of EDTA was 0.1 mM. The forma-tion of the fertilization envelope was prevented by treating eggs with 10 mM DTT in sea water, pH 9.0, for 5-12 minutes, fol-lowed by three washes with sea water. Sperm were added at con-centrations predetermined to give 100% fertilization with untreated eggs.

To prepare extracts, eggs were spun briefly in a clinical cen-trifuge, washed once with sea water and washed three times with ice-cold buffer S (40 mM NaCl, 2.5 mM MgCl2, 300 mM glycine, 100 mM potassium gluconate, 2% glycerol, 50 mM Hepes, adjusted to to pH 7.4) as modified from Winkler and Steinhardt (1981). Eggs were transferred to a Dounce homogenizer (Wheaton) and spun briefly. After removing as much buffer as possible, protease inhibitors leupeptin, chymostatin and pepstatin were added to 10 μg/ml and the eggs were homogenized with 5– 10 strokes of a loose-fitting pestle. In some cases, highly con-centrated extracts were made by adding 1 ml Versilube F-50 oil (General Electric Corp., density = 1.03 g/ml) to the eggs and spin-ning, as modified from a procedure for obtaining cycling extracts from frog eggs (Murray, 1992). Excess buffer was forced to the top of the F-50 oil during centrifugation and then removed prior to homogenization. Essentially identical results were obtained with the two methods. Homogenates were centrifuged at 10,000 g for 15 minutes in a Sorvall SS34 rotor at 0-2°C. The clear cytosol fraction between the lipid cap and the orange-yellow pellet was collected, usually about one-third of the total volume of the packed eggs. The extract was supplemented with ATP (2 mM), creatine phosphate (10 mM) and phosphate creatine kinase (50 μg/ml) and then centrifuged at 10,000 g for 15 minutes to remove any remain-ing pigment granules. The extracts were used immediately or frozen in liquid nitrogen and stored at -80°C. For DNA synthe-sis assays, similar results were obtained with both fresh and frozen extracts.

Demembranated Xenopus sperm were prepared according to Lohka and Masui (1983). Briefly, testes were removed and rinsed with buffer X (80 mM KCl, 5 mM EDTA, 15 mM Pipes, pH 7.4, 15 mM NaCl, 0.2 M sucrose, 7 mM MgCl2). Testes were minced with tweezers to release sperm without disintegrating the outer tissue. Sperm were collected by centrifugation in a clinical cen-trifuge and washed several times with buffer X. Sperm were treated with 0.05% lysolecithin (Sigma) in buffer X for 5 minutes at room temperature. The reaction was stopped by adding an equal volume of 6% BSA (Boehringer Mannheim) in buffer X. Demem-branated sperm heads were washed extensively with buffer X, put into 30% glycerol in buffer X, frozen in liquid nitrogen and stored at -80°C. Demembranated sea urchin sperm heads were prepared in the same way except that they were used immediately since the freeze-thaw cycle caused lysis.

DNA synthesis in fertilized embryos was monitored by labelling cells with 50 μCi/ml [ ³H]thymidine (New England Nuclear, NET-027Z, 84.8 Ci/mmol). Samples of the embryo suspension were mixed with an equal volume of 25% TCA. Precipitated DNA was washed with 10% TCA, collected on nitrocellulose filters (0.45 μm pore size, Millipore), and quantitated by liquid scintillation counting. In some experiments, 10 μg/ml aphidicolin or 200 μM emetine (both from Sigma) was added to eggs before fertilization. For measurements of DNA synthesis in extracts, indicated amounts of demembranated frog or urchin sperm nuclei were added to 50 μl extract and incubated with 0.2 mCi/ml [a-³²P]Dctp (New England Nuclear, 3000 Ci/mmol) for 3 hours at 18°C. Reac-tions were stopped by the addition of 50 μl 40 mM EDTA, 1% SDS, 400 μg/ml proteinase K and incubated at 37°C overnight. DNA was purified by phenol extraction and ethanol precipitation. DNA samples were treated with 20 μg/ml RNase A and elec-trophoresed through a 1% agarose gel. Incorporation was visual-ized by autoradiography.

DNA synthesis in individual nuclei was monitored by per-forming incubations in the presence of biotin-dUTP (Boehringer Mannheim). The incorporated nucleosides were visualized by flu-orescence microscopy according to Hutchison et al. (1988).

To determine the periods of S phase, eggs were fertilized, labelled continuously with [³H]thymidine and assayed for the incorporation of label into TCA-precipitable material. [³H]thymidine incorporation began to rise about 15-20 min-utes after fertilization and continued to about 60 minutes (Fig. 1). First mitosis began at about 70 minutes and cell division around 95 minutes after fertilization. The second and third S phases began during the preceding periods of cytokinesis. Treatment with the protein synthesis inhibitor emetine did not interfere with the first round of DNA syn-thesis but, as expected from previous studies (Wilt et al., 1967; Wagenaar, 1983), blocked entry into first mitosis, cell division and the next S phase. In fertilized embryos, [³H]thymidine incorporation was prevented by the presence of 10 μg/ml aphidicolin, a specific inhibitor of a and b DNA polymerases. Prolonged incubation of unfertilized eggs with [³H]thymidine did not result in significant incor-poration above background (not shown). These results suggest that [³H]thymidine incorporation accurately reflected the replication of nuclear DNA.

Highly concentrated 10,000 g supernatants were prepared from embryos collected at 30 minutes post-fertilization, a time when cells were in S phase of the first cell cycle. Demembranated, lysolecithin-treated Xenopus sperm nuclei (Lohka and Masui, 1983; Blow and Laskey, 1986) were added to the extract and labeled with [a-³²P]dCTP. DNA was purified by phenol extraction and analysed by agarose gel electrophoresis followed by autoradiography. At low nuclear concentrations, incorporation was dose-dependent; below about 800 nuclei/μl, most incorporation was aphidi-colin-sensitive (Fig. 2B), suggesting that most incorpora-tion under these conditions was due to semi-conservative DNA replication. At high nuclear densities, aphidicolin-insensitive incorporation increased, probably reflecting increasing contributions from repair synthesis (Blow and Sleeman, 1990). Extracts from eggs that had been partheno-genetically activated by the calcium ionophore A23187 or by ammonia also showed high levels of DNA synthesis (not shown).

To examine the kinetics of DNA synthesis in fertilized eggs extracts, nuclei (500/μl) were added, samples of the incubation mix were removed at 10 minute intervals and replication was allowed to proceed for another 10 minutes in the presence of [α-³²P]dCTP. As shown in Fig. 2C, incor-poration began after a lag of 20–30 minutes and then con-tinued, peaking around 100 minutes. This initial lag is sim-ilar to that seen in studies using Xenopus egg extracts, where formation of the nuclear envelope is a general requirement for the initiation of DNA synthesis (Blow and Laskey, 1986; Blow and Sleeman, 1990; Leno and Laskey, 1991).

To ask if the G1 arrest of unfertilized eggs was maintained in vitro, extracts of unfertilized eggs were prepared and assayed as above. As shown in Fig. 3B, incorporation of [a-³²P]dCTP into nuclear DNA was markedly lower in extracts of unfertilized eggs. High densities of sperm nuclei did stimulate incoporation in unfertilized egg extracts but this incorporation was mostly insensitive to aphidicolin (not shown), suggesting that it was due to repair-type synthesis. Similar results were obtained with all three species of sea urchins tested: L. pictus, L. variegatus and S. purpuratus.

Preliminary attempts to activate DNA synthesis in unfer-tilized egg extracts by raising pH or calcium levels, events that are triggered by fertilization, were not successful. Sim-ilarly, the addition of recombinant cyclin, which can acti-vate SV40 DNA replication in extracts of human G1 cells (D’Urso et al., 1990), had no effect (not shown).

In the intact cell, there is a 15–20 minute lag between fer-tilization and the onset of DNA replication (Fig. 1). To ask when during this interval the fertilized egg becomes com-petent to initiate DNA synthesis, we prepared an extract from eggs taken 3 minutes after fertilization and compared its replication capacity to those of unfertilized eggs and cells taken at 30 minutes post-fertilization, when cells are in S phase. Unexpectedly, the 3 minute extract displayed about the same incorporation as the extract of S phase embryos (Fig. 4). This result suggests that fertilization is rapidly followedby switching on the capacity for DNA syn-thesis.

The inability of unfertilized eggs and their extracts to support DNA synthesis could be due to the presence of sup-pressors of DNA replication or the absence of activators. To test for these, extracts of unfertilized and 30 minute post-fertilization eggs were mixed in various proportions and then assayed for replication of added sperm nuclei. As shown in Fig. 5, the level of [a-³²P]dCTP incorporation was roughly proportional to the contribution of the fertil-ized egg extract. This result suggests the absence of dom-inant inhibitors in the unfertilized egg extract. It further sup-ports the idea that the potential for turning on DNA synthesis after fertilization is due to the activation of the DNA synthetic machinery.

Fertilization could switch on DNA replication by activat-ing initiation or by turning on elongation of previously ini-tiated origins of replication. To investigate this, the repli-cation efficiencies of unfertilized and S phase extracts were tested using single-stranded bacteriophage M13 DNA. This DNA can be used efficiently by a-polymerase and primase in a reaction that is in some ways similar to the elongation step of DNA replication (Mechali and Harland, 1982; Mechali et al., 1983). Both extracts gave high levels of incorporation using M13 DNA as template, with the fertil-ized egg extract showing slightly higher levels (Fig. 6), sug-gesting that pre-and post-fertilization extracts have about the same capacities for elongation. This result implies that the block to DNA replication in unfertilized eggs acts at the level of initiation.

When microinjected into Xenopus eggs, closed circular double-stranded plasmid DNAs shows correct cell cycle dependence of DNA replication, suggesting the absence of any stringent origin-like sequence requirements in Xenopus eggs (Harland and Laskey, 1980). Similar experiments were carried out with sea urchin eggs but the plasmid DNAs were replicated only after forming high molecular weight con-catemers (McMahon et al., 1985). To examine this capac-ity in the sea urchin in vitro system, closed circular double-stranded plasmid DNA was treated with RNase H to remove potential primers and then added to extracts. Neither unfer-tilized nor fertilized egg extracts supported significant incorporation. The labeled DNA was sensitive to digestion with the restriction enzyme DpnI (which should not cut semi-conservatively replicated, hemi-methylated DNA), suggesting that the observed low level of incorporation was not due to semi-conservative replication (not shown).

When demembranated sperm nuclei are added to Xenopus egg extracts, they decondense to form nucleus-like struc-ture, a step essential for the initiation of DNA replication (see above). To determine whether the marked differences in the DNA synthetic capacities of pre-and post-fertiliza-tion extracts were due to differences in their abilities to sup-port the formation of nuclei, we examined the morphology of nuclei after incubation in each type of extract. Demem-branated Xenopus sperm nuclei added to extracts of fertil-ized eggs displayed more swelling and decondensation than those added to extracts of unfertilized eggs (Fig. 7). It is worth noting that in all preparations of frog sperm nuclei, usually only about 5-10% displayed large, well-rounded shapes; the rest were partially decondensed, forming irreg-ularly shaped nuclei with localized regions of swelling at the perimeter. The input sperm nuclei contained a similar percentage of small, round nuclei, complicating the com-parisons between nuclear changes and DNA synthetic capacities.

To get around this problem, we examined DNA synthe-sis in individual nuclei. Biotin-dUTP was used as precur-sor and incorporation into nuclei was visualized by react-ing the nuclei with Texas Red-tagged streptavidin followed by fluorescence microscopy (Hutchison et al., 1988). Only nuclei incubated in the fertilized egg extracts were labelled and, of these, it was only the large and well-rounded ones that showed incorporation (not shown). Sea urchin sperm nuclei, which decondense only slightly to form very small rounded nuclei, failed to give any significant incorporation of the biotin-dUTP and [a-³²P]dCTP in either extract (not shown). Taken together, these results suggest that partially decondensed nuclei are not capable of DNA synthesis.

To examine the extent of nuclear envelope formation, we included high relative molecular mass FITC-tagged dextran (Mr > 150 × 10³) in the incubations. This reagent is excluded by the nuclear envelope, serving as an indicator for nuclear envelope formation (Finlay and Forbes, 1990) and has been used extensively. Previous studies of nuclear changes occurring in Xenopus extracts indicated that demembranated sperm nuclei added to G2-arrested oocyte extracts, which fail to turn on DNA synthesis, also fail to form nuclear envelopes around the sperm chromosomes (Cox and Leno, 1990). When frog sperm nuclei were added to extracts of fertilized sea urchin eggs, they went on to round up, decondense and exclude dextran, indicating the formation of nuclear envelopes (Fig. 7E, F). In contrast, the same nuclei added to unfertilized egg extracts failed to round up, decondense significantly or exclude dextran (Fig. 7B, C), suggesting that nuclear membranes were not formed in these nuclei. Furthermore, DNA staining using Hoechst 33342 gave images consistent with the idea that nuclei incu-bated in eggs extracts did not contain a complete nuclear envelope. Nuclei incubated for up to 6 hours in egg extracts showed partially decondensed thread-like structures with fuzzy edges and no clear border, whereas those incubated in the fertilized egg extracts showed bright, rounded and homogeneously stained structures with well-defined borders (not shown). It appears, therefore, that the ability to form nuclear envelopes is stage-dependent and might partially account for the low DNA synthetic capacity of unfertilized egg extracts.

We have developed extracts from unfertilized, G1-arrested eggs and fertilized, S-phase embryos of sea urchins that maintain their stage-specific capacities for DNA replication. Extracts prepared from fertilized eggs supported high levels of incorporation of deoxynucleotides into DNA, whereas extracts of unfertilized eggs did not. Furthermore, when demembranated Xenopus sperm nuclei were added, extracts of fertilized eggs supported nuclear envelope re-formation and nuclear decondensation, whereas extracts of unfertil-ized eggs did not.

While we have not directly tested for semi-conservative DNA replication, several pieces of evidence suggest the bulk of incorporation is of this type. First, most incorpora-tion in fertilized egg extracts was aphidicolin-sensitive at appropriate nuclear concentrations, indicating the involve-ment of a-and/or d-polymerases. Second, only nuclei, but not the same concentration of double-stranded closed cir-cular DNA, showed high levels of DNA synthesis in the fertilized egg extracts, suggesting that incorporation is not due simply to the nick-translation type of DNA replication. Third, kinetic analysis revealed a significant lag preceding the onset of incorporation; such a lag might represent the time necessary for nuclear envelope assembly and forma-tion of initiation complexes (Sheehan et al., 1988). The find-ing that extracts prepared from eggs were not capable of incorporation supports the notion that incorporation in the fertilized extracts reflects the high DNA synthesis activities present in the S-phase extracts.

Nuclei incubated with extracts prepared from unfertilized eggs showed a very low capacity to synthesize DNA. Inter-estingly, this low efficiency was only restricted to double-stranded DNA; the DNA synthesis on single-stranded DNA templates, which probably reflects the capacity for elongation, was about the same in both fertilized and unfer-tilized egg extracts. The single-stranded DNA-dependent synthesis was also aphidicolin-sensitive, suggesting that in both extracts the activities of a-and/or d-polymerases were not rate-limiting, strengthening the argument that the aphidicolin-sensitive high incorporation observed in the fer-tilized egg extracts might be due to initiation on double-stranded DNA. This result also suggests that activation of DNA replication in sea urchin eggs is regulated at the level of initiation. Since the DNA synthesis capacitites of mixed lysates containing both fertilized and unfertilized egg extracts were proportinal to the contribution of fertilized egg extract, the pre-fertilization block to DNA synthesis does not appear to be due to the presence of dominant sol-uble inhibitors in the egg extracts. These studies also suggest that the process that removes the block to DNA replication in eggs is initiated soon after fertilization, since extracts prepared from eggs at just 3 minutes post-fertil-ization gave rates of DNA synthesis as high as those from mid-S phase extracts.

Extracts of unfertilized eggs supported very little chro-mosome decondensation and were incapable of re-forming nuclear envelopes around the demembranated nuclei. Sim-ilar phenomena were also observed in extracts prepared from Xenopus oocytes (Cox and Leno, 1990). Since sea urchin eggs are arrested in G1 of the first mitotic cell cycle, whereas Xenopus oocytes are arrested at the G2/M border of meiosis I, it seems that the arrested state rather than the specific cell cycle arrest point might define this incapacity for nuclear envelope formation. Whether this block is due to changes in properties of a membrane or cytosolic frac-tion is not clear. The establishment of a nuclear assembly system from fertilized sea urchin eggs provides a useful means to study these changes.

In addition to the well-characterized mitotic roles of cyclin and the protein kinase cdc2 (reviewed by Ruderman et al., 1991), there is a growing number of indications that these or related proteins may be involved in the regulation of initiation of DNA synthesis as well. When extracts of Xenopus eggs are depleted of cdk2, either by exposure to antibodies or p13^suc1 beads, they lose the ability to initiate DNA replication (Blow and Nurse, 1990; Fang and New-port, 1991). When extracts of human G1 cells are supple-mented with recombinant cyclin, they gain the ability to initiate replication of SV40 DNA (D’Urso et al., 1990). Finally, microinjection of cyclin A or cdc2 antisense oligonucleotides into G1 human cells blocked them from entering S phase (Furukawa et al., 1990; Girard et al., 1991). While we have not yet tried the effects of depleting cdc2, preliminary experiments testing the effect of adding recombinant, biologically active cyclin A or B (Luca et al., 1991) to sea urchin egg extracts failed to detect any stim-ulation of DNA synthesis (Zhang and Ruderman, unpub-lished observations). However, in the absence of more extensive efforts to test the effects of cyclins on DNA syn-thesis in the extracts, important roles cannot yet be excluded.