Nucleoplasmin-mediated chromatin remodelling is required for Xenopus sperm nuclei to become licensed for DNA replication

Abstract

During late mitosis and early G₁, a series of proteins are assembled onto replication origins, resulting in them becoming ‘licensed’ for replication in the subsequent S phase. Four factors have so far been identified that are required for chromatin to become functionally licensed: ORC (the origin recognition complex) and Cdc6, plus the two components of the replication licensing system RLF-M and RLF-B. Here we describe the first steps of a systematic fractionation of Xenopus egg extracts to identify all the components necessary for the assembly of licensed replication origins on Xenopus sperm nuclei (the physiological DNA substrate in this system). We have purified a new activity essential for this reaction, and have shown that it is nucleoplasmin, a previously known chromatin remodelling protein. Nucleoplasmin decondenses the sperm chromatin by removing protamines, and is required at the earliest known step in origin assembly to allow ORC to bind to the DNA. Sperm nuclei can be licensed by a combination of nucleoplasmin, RLF-M and a partially purified fraction that contains ORC, Cdc6 and RLF-B. This suggests that we are likely to have identified most of the proteins required for this assembly reaction.

INTRODUCTION

In eukaryotes the initiation of DNA replication occurs at multiple replication origins scattered throughout the genome. To maintain constant ploidy and ensure genetic stability, DNA must be replicated only once per cell cycle such that no region of the genome remains unreplicated and no region is re-replicated. Chromosome replication can be biochemically analysed in cell-free extracts derived from Xenopus eggs that recapitulate cell cycle events in vitro. Chromatin added to this system is first assembled into a functional nucleus, and then undergoes a single complete round of semi-conservative DNA replication (1).

The precise duplication of chromosomal DNA in the Xenopus cell-free system is regulated by two distinct cell cycle signals (reviewed in 2,3). The first of these, replication licensing factor (RLF), stably binds to replication origins and puts them into an initiation competent state (4). The second signal, S-phase promoting factor (SPF), induces licensed origins to initiate and in doing so removes the license (5–7). The two signals are both temporally and physically separated (3,4,8,9), thus ensuring the precise duplication of chromosomal DNA.

RLF is inactive during metaphase, and is abruptly activated on entry into anaphase (9). The activation of RLF can be inhibited by protein kinase inhibitors such as 6-dimethylaminopurine (6-DMAP) (8) or staurosporine (10). Extracts treated with 6-DMAP contain all activities required for complete DNA replication with the exception of RLF itself. This has provided an assay system for the chromatographic fractionation and identification of RLF from Xenopus. This has shown that RLF consists of two distinct components: RLF-M, comprising all six members of the minichromosome maintenance (MCM/P1) family (11–15, reviewed in 16) and RLF-B, a currently unpurified activity that functions during the loading of RLF-M onto chromatin (11,17). The RLF-M complex is removed from chromatin as replication proceeds, thus ensuring initiation does not occur on replicated DNA.

Two proteins, XORC and XCdc6, are required to be present on chromatin before licensing can occur and RLF-M can bind (18–21, reviewed in 22). XORC is the Xenopus homologue of the Saccharomyces cerevisiae origin recognition complex (ORC). ORC was identified by its ability to bind specifically to yeast origins of replication (23). When sperm nuclei are added to egg extract XORC is rapidly loaded on to the chromatin, and subsequent licensing is dependent on the presence of XORC on the DNA (20,21). XOrc1, the largest XORC subunit, saturates chromatin at ~1 molecule per 10 kb (20,24). This corresponds to the average spacing between replication origins expected in the Xenopus early embryo (25–27), suggesting that the binding of a single molecule of XORC may be sufficient to specify a replication origin.

XCdc6 is the Xenopus homologue of the S.cerevisiae Cdc6, and the Schizosaccharomyces pombe cdc18 proteins, both of which have been implicated in preventing re-replication of DNA in a single cell cycle (28–30). XCdc6 is loaded onto XORC-containing chromatin and is required for RLF-M loading (19). After licensing has occurred and RLF-M has been loaded onto chromatin, both XORC and XCdc6 can be removed from the chromatin and are no longer required for DNA replication (24,31).

In this paper we describe work towards reconstituting the assembly of licensed replication origins using proteins purified from Xenopus egg extracts. We first describe an alternative purification of RLF-M from Xenopus egg extract. This novel purification scheme has permitted the identification of an additional activity required for licensing of Xenopus sperm nuclei. We have used standard chromatographic fractionation to identify this additional activity as the chromatin remodelling protein nucleoplasmin. We show that nucleoplasmin is required to decondense the sperm chromatin and to allow XORC to bind to the DNA.

MATERIALS AND METHODS

Preparation of egg extracts and chromatin templates

Metaphase-arrested Xenopus egg extracts were prepared as described (32). For licensing assays, extracts were supplemented with 100 µg ml^–1 cycloheximide, 25 mM phosphocreatine, 15 µg ml^–1 creatine phosphokinase, 3 mM 6-DMAP and [α-³²P]dATP, and were then released into interphase with 0.3 mM CaCl₂.

Licensing factor extract (LFE) which was used as a source of protein for fractionation studies was prepared as described (32). Briefly, eggs were activated for 5 min by the calcium ionophore A23187, before being spin-crushed in buffer lacking EGTA. Recovered cytoplasm was diluted 1:5 with LFB1 [40 mM HEPES KOH pH 8.0, 20 mM K₂HPO₄/KH₂PO₄ pH 8.0, 2 mM MgCl₂, 1 mM EGTA, 2 mM DTT, 10% (w/v) sucrose and 1 µg ml^–1 each of leupeptin, pepstatin and aprotinin] supplemented with 50 mM KCl (i.e. LFB1/50) and re-centrifuged at 20 000 r.p.m. for 30 min in an SW41 swinging bucket rotor (Beckman). The clarified supernatant (LFE) was frozen and stored at –80°C until required.

Nucleoplasmin-depleted extract was prepared from LFE. Monoclonal antibody PA3C5 against nucleoplasmin (33) was prepared from a cell line kindly provided by Dr S. Dilworth. The immunodepletion was performed as described (32), with minor alterations. Briefly, the quantity of anti-nucleoplasmin antibody needed to saturate protein-A beads (Amersham-Pharmacia Biotech) was first determined. Antibody-bound beads were then extensively washed in 100 mM HEPES KOH pH 8.0 and LFB1/50. For the depletion LFE was supplemented with 0.12 vol of protein A–antibody beads and incubated for 40 min at 4°C on a flatbed roller. Depleted extract was recovered by centrifugation through a 25 µm nylon filter. This process was then repeated and the recovered extract was frozen in aliquots in liquid nitrogen.

Demembranated Xenopus sperm nuclei were prepared as described (32) and frozen in aliquots in liquid nitrogen. Unlicensed 6-DMAP chromatin was assembled by incubating sperm nuclei for 15 min at 23°C in 6-DMAP treated extracts at 30 000 nuclei µl^–1. The extract was then diluted 10-fold in nuclear isolation buffer (32) (NIB: 50 mM KCl, 50 mM HEPES KOH pH 7.6, 5 mM MgCl₂, 2 mM DTT, 0.5 mM spermidine 3HCl, 0.15 M spermine 4HCl, 1 µg ml^–1 each of leupeptin, pepstatin and aprotinin) supplemented with 0.01% Triton X-100, and then underlayered with the same buffer containing 15% sucrose. The chromatin was pelleted at 6000 g in a swinging bucket rotor at 4°C for 5 min. The diluted extract was then removed and the chromatin pellet was resuspended in NIB and frozen in aliquots in liquid nitrogen.

RLF-M purification

All chromatographic procedures were performed at 4°C. First, crude extract was fractionated by differential precipitation with polyethylene glycol (PEG) (11,32). Briefly, LFE was supplemented with 50% PEG 6000 solution to a final concentration of 3.5% PEG, incubated on ice for 30 min and precipitated proteins removed by spinning at 12 500 g for 15 min in a fixed-angle rotor. The pellet (‘PEG-B’ fraction) was resuspended in LFB1/150 at 5× with respect to undiluted egg extract. Further fractionation of PEG-B by chromatography on phosphocellulose and by ammonium sulphate precipitation to generate the BPAS fraction was performed as described (17,32). The PEG-B supernatant was adjusted to 9.0% PEG and proteins were precipitated as before to generate the ‘PEG-M’ fraction. The pellet was resuspended at 1× in LFB3/10 [20 mM HEPES KOH pH 8.0, 2 mM DTT, 10% (w/v) sucrose and 1 µg ml^–1 each of leupeptin, pepstatin and aprotinin, supplemented with 10 mM KCl].

Five millilitres of 1× PEG-M fraction were applied to 10 ml High Performance SP-Sepharose (Amersham-Pharmacia Biotech) packed into an XK 16/20 column (Amersham-Pharmacia Biotech) equilibrated in LFB3/10. Bound protein was eluted with a linear gradient to LFB3/500 over 10 column vol with a flow rate of 0.16 ml min^–1 and 1.8 ml fractions were taken. XMcm peak fractions were pooled, precipitated with 17.5% PEG and resuspended in LFB2/50 (LFB1/50 supplemented with 2.5 mM ATP) to a concentration of 10× (500 µl). The 10× concentrate was then applied to a 24 ml Superose 6 gel filtration column (Amersham-Pharmacia Biotech) equilibrated in LFB1 lacking EGTA and run at a flow rate of 0.5 ml min^–1. RLF-M active fractions were pooled and applied to a 100 µl SMART chromatography system (Amersham-Pharmacia Biotech) MonoQ column equilibrated in LFB1/70 (–EGTA). Bound proteins were eluted with a linear gradient to LFB1/350 (–EGTA) over 6 column vol and 25 µl fractions were collected. 2.5 µl of each fraction was taken for SDS–PAGE Coomassie analysis and 0.1 µl was used for immuno-blotting. When reconstituting RLF-M activity from the MonoQ fractions, MQ1 and MQ3 peak fractions were mixed in a ratio 1.25:0.75 and then diluted 1:4 in LFB2.

For the quantitation of protein recovery total protein concentration was determined using the Coomassie Plus System (Pierce Warriner) and specific bands were quantified from western blots using the LAS-1000 Intelligent Dark Box (Fuji).

Nucleoplasmin purification

Five millilitres of 1× PEG-M fraction were applied to an SP Sepharose column as described above for RLF-M purification. The 280 nm UV absorbance peak of the flow through fraction from this column was collected, diluted to 0.5× with LFB1/0 and supplemented with 3 M KCl to a final concentration of 50 mM and stored at –80°C until required. Two millilitres of this material was applied to a 1 ml Hi Trap Q Sepharose column (Amersham-Pharmacia Biotech) in LFB1/50 at a flow rate of 0.25 ml min^–1 on a SMART chromatography system (Amersham-Pharmacia Biotech). The column was washed in LFB1/325 and nucleoplasmin was eluted with a step to LFB1/1000, collecting 250 µl fractions. The fractions containing the 280 nm UV absorbance peak from the 325 mM–1 M KCl eluate were pooled, precipitated with 17.5% PEG and resuspended in LFB2/50 to a concentration of 5× (200 µl). The sample was then incubated in an 80°C water bath for 10–15 min with regular shaking and then incubated on ice for 20 min. Precipitated proteins were then pelleted by centrifugation at 12 500 g for 20 min at room temperature in a bench-top microfuge. The supernatant contained apparently pure nucleoplasmin. Nucleoplasmin was also purified from whole extract using the protocol of Sealey et al. (34).

Replication and decondensation assays

RLF-M assays were performed as described (32) with minor alterations. Briefly, fractions to be analysed were first diluted in LFB2/50 to the appropriate concentration. A total of 3 µl containing the different licensing fractions were first incubated with 0.3 µl of either 6-DMAP chromatin (50 ng DNA/µl) or demembranated Xenopus sperm nuclei (50 ng DNA/µl) for 30 min at 23°C (the ‘licensing reaction’). The degree of licensing was assessed by subsequent addition of 5.7 µl 6-DMAP treated extract containing [α-³²P]dATP and incubation for a further 90 min at 23°C (the ‘replication reaction’). Total DNA synthesised was measured by precipitation with trichloroacetic acid. The volume of each fraction used in the assay was either 3 µl LFE/buffer control, or 1 µl 1.5× BPAS plus 1 µl ‘RLF-M’ fraction plus 1 µl ‘nucleoplasmin’ fraction or LFB2/50.

Decondensation assays were performed as above except that the sperm chromatin was prepared for microscopy instead of being subject to a replication assay. For microscope analysis 1 µl sperm chromatin was mixed with 1 µl Hoechst 33258 fluorescent dye (20 mg ml^–1 in H₂O) and visualised under UV or phase-contrast optics on an Axioskop microscope (Zeiss). Photographs were taken using the MC 80DX camera system (Zeiss) on p3200 135 TMax film (Kodak Eastman Co.).

Chromatin isolation and blotting

To investigate chromatin binding, licensing assays were performed as above except that they were scaled up 10-fold to a total of 33 µl. After the 30 min licensing incubation each reaction was diluted in 200 µl of NIB supplemented with 0.1% Triton X-100, and underlayered with the same buffer containing 10% sucrose. The chromatin was pelleted at 6000 g in a swinging bucket rotor at 4°C. The diluted extract was removed and the chromatin pellet was resuspended in SDS–PAGE loading buffer. The samples were run on 10% SDS–PAGE gels and analysed by immuno-blotting. Polyclonal rabbit antibodies against XOrc1 and XCdc6 were as previously described (17,20).

RESULTS

An activity required for sperm chromatin licensing is lost as RLF-M is purified

We embarked on a fractionation–reconstitution programme to determine all the factors in Xenopus egg extract that are required to assemble licensed replication origins on demembranated Xenopus sperm nuclei, the natural substrate for DNA replication in egg cytoplasm. The first step in our standard fractionation scheme of Xenopus egg extract (Fig. 1A) is a differential PEG precipitation which produces two fractions termed ‘PEG-B’ and ‘PEG-M’ (11). PEG-B contains XORC, XCdc6 and RLF-B (11,17,20), whilst PEG-M contains RLF-M (11,13,15). Unlicensed chromatin (‘6-DMAP chromatin’) can be prepared by incubating Xenopus sperm nuclei in 6-DMAP-treated Xenopus extracts that lack RLF (8,9). This 6-DMAP chromatin can be efficiently licensed by a combination of PEG-M and PEG-B fractions (Fig. 1B and ref. 11). Demembranated Xenopus sperm nuclei were also efficiently licensed by a combination of these two fractions (Fig. 1B). PEG-B can be further purified by chromatography on phosphocellulose and ammonium sulphate precipitation to yield a fraction termed “BPAS“, containing XORC enriched ~100-fold, XCdc6 enriched ~36-fold and RLF-B enriched ~230-fold (Fig. 1C and refs 17,20). BPAS could replace PEG-B for the licensing of both 6-DMAP chromatin and sperm nuclei without significant loss of activity (Fig. 1B). The next step in the purification of RLF-M from PEG-M is chromatography on Q Sepharose (Fig. 1A and refs 11,13). Although the eluate from this step yielded significant RLF-M activity as evidenced by its ability to license 6-DMAP chromatin, this fraction was unable to support significant licensing of sperm nuclei (Fig. 1B). This suggests that as RLF-M is purified an essential activity is lost from the PEG-M fraction that is required for the licensing of sperm nuclei. We therefore sought to identify this activity.

Figure 1.

As RLF-M is purified an activity is lost that is required for the licensing of Xenopus sperm nuclei. (A) Fractionation scheme for the purification of licensing activities from Xenopus egg extract. Purification steps are shown in boxes, elution conditions in plain text and fraction names in bold. (B) Xenopus sperm nuclei and unlicensed 6-DMAP chromatin were licensed with putative licensing fractions. Licensing was assayed by quantifying replication in 6-DMAP treated Xenopus egg extract. The dashed line indicates the amount of replication obtained with chromatin incubated in buffer alone for the licensing reaction. (C) The recovery of total protein, RLF-B activity and XOrc1 and XCdc6 in the BPAS fraction.

Novel purification scheme for RLF-M

Fractionation of RLF-M on Q Sepharose separates the constituent MCM/P1 proteins into three different subcomplexes that elute at different salt concentrations (15). The separation of these subcomplexes not only hindered reconstitution of RLF-M activity, but also complicated our search for the missing additional activity due to the partial co-elution of the essential activity with RLF-M subcomplexes (data not shown). To facilitate the further characterisation of the licensing reaction we therefore devised an alternative RLF-M purification strategy that optimised purification and avoided the use of Q Sepharose at the first column stage.

PEG-M was first subjected to SP Sepharose chromatography, followed by gel filtration on Superose 6 (Fig. 2A). The peak gel filtration fractions were then applied to a MonoQ column (Fig. 2C) which separated the MCM/P1 proteins of RLF-M into three distinct subcomplexes, MQ1 (containing XMcms 3 and 5), MQ2 (containing XMcms 3 and 7) and MQ3 (containing XMcms 2, 4, 6 and 7) (15). In our new fractionation scheme, MQ1 and MQ3 peak fractions, which can be combined to provide RLF-M activity (Fig. 2B and ref. 15), were essentially pure (>80%). Although rapid, the new purification gave a low yield of MCM/P1 proteins and a correspondingly low yield of RLF-M activity (Fig. 2B and D). XMcm3 is present in undiluted Xenopus extract (total protein concentration ~60 mg/ml) at ~100 µg/ml (9), and this figure is expected to apply to the other five XMCM/P1 proteins (15). Full purification is therefore expected at ~100-fold enrichment, close to the value of 76-fold that we obtained (Fig. 2D).

Figure 2.

New purification scheme for RLF-M. (A) Schematic diagram of the new fractionation scheme for the PEG-M fraction. (B) RLF-M activity of fractions during the course of the purification, using 6-DMAP chromatin as template. Apart from whole extract, all fractions were assayed in combination with BPAS to provide RLF-B activity. The dashed line indicates the amount of replication obtained with chromatin incubated in buffer alone for the licensing reaction. (C) Details of the final MonoQ step: salt gradient showing MQ1, MQ2 and MQ3 fractions (top), immunoblot for XMcms 2, 3 and 7 (middle) and Coomassie-stained gel (bottom). Molecular weight markers (kDa) are shown to the left of the Coomassie-stained gel. (D) The recovery of RLF-M activity (left) and XMcm3 protein (right) at each step of the new purification scheme.

Nucleoplasmin provides a novel activity required for licensing sperm nuclei

We used this new RLF-M purification scheme to identify the additional factor(s) required to license sperm nuclei in the presence of BPAS and RLF-M. When sperm nuclei were incubated with BPAS (to provide XORC, XCdc6 and RLF-B) and the SP-Sepharose eluate (to provide RLF-M), very little licensing occurred (Fig. 3A), consistent with our previous results (Fig. 1B). However, licensing activity could be restored to this mixture when supplemented with the flow-through fraction from the SP-Sepharose column (Fig. 3A). No restoration of activity was obtained with any other fractions from the SP-Sepharose column, showing that the additional activity was present only in the flow-through fraction (Fig. 3A).

Figure 3.

Nucleoplasmin is required for the licensing of sperm chromatin. (A) Assay for the licensing of sperm nuclei, using different combinations of fractions. The dashed line indicates the amount of replication obtained with chromatin incubated in buffer alone for the licensing reaction. (B) Immunoblot for nucleoplasmin (top) and Coomassie stained gel (bottom) for different fractions during the purification of nucleoplasmin from PEG-M; molecular weight markers (kDa) are shown to the left.

We further purified the essential activity present in the SP-Sepharose flow-through fraction by chromatography on Q Sepharose (Fig. 3B). A Coomassie-stained gel of this fraction revealed prominent bands at ~31–33 kDa (Fig. 3B, lower panel). Since the size and chromatographic behaviour of this protein matched that of the chromatin-remodelling protein nucleoplasmin (34,35), we next heat-treated the Q Sepharose eluate, since this is an effective step in the purification of nucleoplasmin. The heat-treated fraction retained all of the activity essential for replication licensing (Fig. 3A), and on Coomassie staining showed that the ~31–33 kDa bands had been almost completely purified (Fig. 3B, lower panel). Immunoblotting confirmed the presence of nucleoplasmin in the final heat treated fraction (Fig. 3B, upper panel). Nucleoplasmin purified by a published protocol (34) also supported efficient licensing of sperm nuclei (Fig. 3A), and showed a pattern of ~31–33 kDa bands that closely matched our most purified fraction (Fig. 3B). We therefore conclude that nucleoplasmin is required for the efficient licensing of demembranated Xenopus sperm nuclei in our reconstituted system.

Nucleoplasmin is a phosphopentamer present in Xenopus eggs and early embryos that was originally identified as a chromatin assembly factor capable of loading histones onto naked DNA to form nucleosomes (36). In the Xenopus egg it binds predominantly to the large stockpile of histones H2A and H2B and allows them to be assembled onto DNA when necessary (33,37,38). At fertilisation it also decondenses sperm chromatin by exchanging sperm protamines for histones H2A and H2B (39–41). The purification of nucleoplasmin from the PEG-M fraction recovers ~50% of the total cellular pool (Fig. 3B and data not shown). Thirty-four percent of cellular nucleoplasmin partitions in the PEG-B fraction, which is lost in our purification scheme as no nucleoplasmin is present in BPAS (Fig. 3B). Virtually all of the nucleoplasmin present in the PEG-M cut is recovered in the SP Sepharose flow-through fraction, and nucleoplasmin is absent from the RLF-M-containing eluate at 110–200 mM KCl, as well as the other parts of the elution gradient (Fig. 3B, 10–110 mM KCl and 200–500 mM KCl). Recovery of nucleoplasmin over the next two steps (Q Sepharose and heat treatment) is also >50% (Fig. 3B), so this scheme therefore provides a good alternative purification strategy for nucleoplasmin.

Further reconstitution of replication licensing

To determine if there were any other factors present in the PEG-M fraction that are required for sperm chromatin licensing we assayed licensing activity with BPAS, nucleoplasmin and successive RLF-M purification fractions (Fig. 4A). We found that a combination of the BPAS fraction with purified RLF-M and purified nucleoplasmin could support significant licensing of sperm nuclei (Fig. 4A). As RLF-M is purified there is a decrease in the efficiency with which the sperm nuclei is licensed. However, total RLF-M activity also drops significantly in the course of the purification (Fig. 2B). We therefore compared the recovery of RLF-M activity with the efficiency of the successive fractions in supporting the licensing of sperm nuclei. Figure 4B shows that the activity of the fractions in the two assays remained virtually the same. This suggests that the loss of activity on sperm nuclei (Fig. 4A) as RLF-M is purified is due only to the loss of RLF-M activity and not due to the loss of any other essential factor. RLF-M and nucleoplasmin are therefore the only activities required for the licensing of sperm nuclei that are present in the PEG-M fraction but not in BPAS. Since BPAS is already 230-fold enriched this suggests that we are likely to have identified most of the activities required to support this reaction.

Figure 4.

Licensing sperm nuclei with purified RLF-M. (A) Assay for the licensing of sperm nuclei using different combinations of fractions. The fractions used were exactly the same as those used to determine RLF-M activity in Figure 2B. The dashed line indicates the amount of replication obtained with chromatin incubated in buffer alone for the licensing reaction. (B) Comparison of the ability of fractions to provide RLF-M activity (in combination with BPAS) and to support the licensing of sperm chromatin (in combination with BPAS and nucleoplasmin). The data were taken from the experiments shown in Figures 2B and (A), and were normalised to the extent of replication obtained in whole extract.

Nucleoplasmin is required for licensing sperm nuclei

In order to characterise the role that nucleoplasmin plays in licensing sperm nuclei, we first wanted to check whether there are any other components in whole extract that can substitute for its function in licensing. Nucleoplasmin was immunodepleted from whole egg extract using an anti-nucleoplasmin monoclonal antibody (33) (Fig. 5A, inset). Untreated and immunodepleted extracts were then titrated into a reconstituted licensing reaction containing BPAS (to provide XORC, XCdc6 and RLF-B activities) and the SP-Sepharose eluate (to provide RLF-M) (Fig. 5A). In the absence of added nucleoplasmin, very little licensing was observed (Fig. 5A, lower dashed line). When untreated extract was titrated into this reaction, efficient licensing was observed with 10 nl extract per ng sperm DNA (Fig. 5A, open squares). This figure is an agreement with the capacity of egg extract to license up to 100 ng sperm DNA/µl (9,20). In contrast, an equivalent quantity of the nucleoplasmin-depleted extract gave only a low level of activity (Fig. 5A, filled circles). This suggests that in whole egg extract, nucleoplasmin provides the majority (~90%) of this essential licensing function. Since nucleoplasmin is present in egg extracts at ~500 ng/µl (42), this suggests that ~5 ng nucleoplasmin should be sufficient for licensing 1 ng of sperm DNA.

Figure 5.

Nucleoplasmin is required for Xenopus sperm chromatin decondensation. (A) Sperm chromatin was incubated with BPAS, the SP-Sepharose RLF-M fraction and varying amounts of either extract or nucleoplasmin depleted extract. The extent of DNA replication per ng template DNA is shown. The lower dashed line indicates the amount of licensing with buffer alone and the upper dashed line indicated the amount of licensing with BPAS, SP-Sepharose eluate and purified nucleoplasmin. Inset, extract (1) and nucleoplasmin depleted extract (2) were immunoblotted for nucleoplasmin. (B) Xenopus sperm nuclei were incubated for 30 min in a mock licensing assay with varying amounts of extract and nucleoplasmin-depleted extract. DNA was stained with Hoechst 32258 and visualised under UV illumination. The percentage of sperm nuclei showing visible decondensation was recorded. (C and D) Licensing assays were performed on sperm nuclei incubated with BPAS, SP-Sepharose eluate and various quantities of either nucleoplasmin, or poly-glutamate. The extent of DNA replication per ng template DNA is shown. The dashed line indicates the amount of replication obtained with chromatin incubated in buffer alone for the licensing reaction. (E) Sperm nuclei were incubated for 30 min in either (i) whole extract (LFE); (ii) nucleoplasmin (4 ng/ng DNA); (iii) poly-glutamate (40 ng/ng DNA); (iv) buffer; (v) the BPAS fraction or (vi) the SP eluate. DNA was stained with Hoechst 33258 and viewed by fluorescent microscopy. Scale bar = 10 µm.

We confirmed this result by determining the optimal concentration of nucleoplasmin required for licensing sperm nuclei. Figure 5C shows that, as predicted, ~5 ng of purified nucleoplasmin was required to efficiently license 1 ng of sperm DNA. At concentrations below this, the degree of licensing declined in proportion with the quantity of nucleoplasmin. Providing nucleoplasmin in excess (up to 17 ng/ng DNA) did not inhibit the licensing reaction.

Nucleoplasmin is required for sperm decondensation and XORC binding

The nucleoplasmin-mediated removal of protamines from Xenopus sperm nuclei causes a visible decondensation of the chromatin (39,40) (Fig. 5E, i, ii and iv). We observed that when licensing was limited by insufficient nucleoplasmin (Fig. 5A and C), there was a proportionate decrease in the number of sperm nuclei showing visible chromatin decondensation. When we quantified this effect using whole and nucleoplasmin-depleted extracts, we found reasonable agreement between the proportion of sperm nuclei showing visible decondensation (Fig. 5B) and the degree of replication licensing (Fig. 5A). As previously reported (39), extract immunodepleted of nucleoplasmin displayed only limited sperm decondensation (Fig. 5B, filled circles). This correlation suggests that the essential function of nucleoplasmin in replication licensing is to decondense the sperm chromatin. Consistent with this interpretation, neither BPAS nor the SP-Sepharose eluate fractions used in the reconstituted licensing reactions were themselves able to promote decondensation (Fig. 5E, iv–vi). Therefore sperm chromatin is only efficiently licensed when a fraction with decondensing activity is provided in addition to XORC, XCdc6, RLF-B and RLF-M (Figs 1B, 3A, 4A, 5A and C).

It has been previously reported that poly-anions such as poly-glutamate and RNA can decondense chromatin and mediate the assembly of histones into nucleosomes (43–45). We therefore investigated if poly-glutamate could replace nucleoplasmin in the sperm licensing reaction. Figure 5D shows that poly-glutamate was able to support efficient licensing of sperm DNA, though this required ~40 ng poly-glutamate per ng DNA, ~10 times more than the required amount of nucleoplasmin. At these concentrations, poly-glutamate caused visible decondensation comparable to that seen with whole extract (Fig. 5E, iii). Unlike nucleoplasmin, the titration of poly-glutamate gave a sharp peak of activity with excess peptide being inhibitory to the reaction (Fig. 5D).

The assembly of licensed replication origins consists of three distinct stages, marked by the sequential assembly of XORC, XCdc6 and RLF-M onto chromatin (19–21,24). We next investigated at which stage of this assembly reaction nucleoplasmin was required (Fig. 6). Sperm nuclei were incubated with various combinations of fractions; chromatin was then isolated and immunoblotted with antibodies against XOrc1 (to indicate XORC binding), XCdc6 and XMcm7 (to indicate RLF-M binding). In whole extract, XOrc1, XCdc6 and XMcm7 were all assembled onto chromatin, whilst addition of geminin, an inhibitor of the licensing reaction (46), allowed XOrc1 and XCdc6 to bind, but not XMcm7 (Fig. 6, lanes 1 and 2). Incubation of sperm nuclei with the BPAS fraction alone (containing XORC, XCdc6 and RLF-B) led to neither XOrc1 nor XCdc6 binding to chromatin (Fig. 6, lane 3). The PEG-M fraction (containing nucleoplasmin) was required in addition to BPAS for the assembly of XOrc1 and XCdc6 onto chromatin (Fig. 6, lane 5). Importantly, purified nucleoplasmin was able to substitute for PEG-M in allowing the binding of XOrc1 and XCdc6 (Fig. 6, lane 7). In contrast, the SP eluate (containing RLF-M but lacking nucleoplasmin) could not support the loading of XOrc1 or XCdc6 (Fig. 6, lane 8). In order for XMcm7 to be assembled onto chromatin (an indication that the origin has become functionally licensed) a combination of nucleoplasmin, BPAS and the SP eluate was required (Fig. 6, lane 9). These results demonstrate that nucleoplasmin is required for XORC to bind to Xenopus sperm nuclei, the earliest known step in replication origin assembly.

Figure 6.

Nucleoplasmin is required for XORC binding. Sperm nuclei were incubated for 30 min in various combinations of fractions, the chromatin was then isolated and immunoblotted for XOrc1, XCdc6 and XMcm7.

DISCUSSION

Recent work, both from yeast genetics and the Xenopus cell-free system, has shown that a functional replication origin is formed by the ordered assembly of proteins onto chromosomal DNA. We describe here a fractionation scheme for Xenopus egg extracts that should allow us to identify all the proteins required to reconstitute licensed replication origins on its physiological substrate, Xenopus sperm nuclei. We describe the identification of the chromatin remodelling protein nucleoplasmin as a component of this reaction.

A new fractionation scheme for licensing activities

Previous work has identified four distinct activities in Xenopus egg extracts that are required for DNA to become functionally licensed (Fig. 7). XORC (18,20,21), XCdc6 (19), RLF-M (11–15) and RLF-B (11,17) interact with chromatin in a specific order, to yield functionally licensed replication origins (Fig. 7C–E). Differential PEG precipitation of egg extract produces two fractions that are sufficient to reconstitute licensing: PEG-B (containing XORC, XCdc6 and RLF-B) and PEG-M (containing RLF-M). In the present work we have examined the PEG-M fraction for other essential components. Because the MCM/P1 proteins that comprise RLF-M are separated into subcomplexes following anion exchange chromatography (15), it was necessary to change our existing RLF-M fractionation scheme, which used Q Sepharose chromatography as a the step directly after PEG precipitation. The new fractionation scheme described here provides a rapid purification of active RLF-M in three steps, with a comparable yield to our previous protocol (11,32). This new scheme also enabled us to purify another essential activity present in the PEG-M fraction and to show that it was Xenopus nucleoplasmin.

Figure 7.

Schematic diagram of the steps required for Xenopus sperm nuclei to become licensed. A small region of chromosomal DNA around a single replication origin is shown. (A) Sperm DNA is initially covered with H3/H4 tetramers (shaded discs) and protamines (open discs). (B) Nucleoplasmin removes the protamines, thus decondensing the chromatin. (C) A future replication origin is bound by a single molecule of XORC. (D) XCdc6 binds to the XORC-containing DNA. (E) In the presence of RLF-M (M) and RLF-B (B), replication licensing then occurs, which results in the loading of multiple copies of RLF-M onto the adjacent chromatin.

The role of nucleoplasmin in replication licensing

Nucleoplasmin performs two known operations on chromatin (reviewed in 47). First, it can serve as a molecular chaperone and assemble free histones into nucleosomes on DNA. In purified systems it can assemble all four core histones onto DNA (36), but in vivo its function is probably restricted to the assembly of histones H2A and H2B, whilst histones H3 and H4 are assembled by another protein, N1 (33,37,48). Secondly, nucleoplasmin can remodel chromatin by removing a range of proteins such as protamines (40) and linker histones (49). The removal of protamines from sperm nuclei is accompanied by a visible decondensation of the chromatin that normally happens within a few minutes after fertilisation (39). Removal of nucleoplasmin from Xenopus extract by immunodepletion significantly reduces this decondensation (39). Consistent with this result we show here by chromatographic fractionation and immunodepletion, that nucleoplasmin represents the major sperm decondensing activity present in Xenopus egg extract, accounting for ~90% of total.

Two observations suggest that the requirement for nucleoplasmin in the licensing of sperm nuclei corresponds to a requirement for chromatin decondensation, rather than for assembly of H2A and H2B. First, histones H2A and H2B are absent from the reconstituted reaction since they are lost from nucleoplasmin during the purification and are not expected to be present in the BPAS fraction given their known chromatographic properties. Secondly, concentrations of nucleoplasmin and poly-glutamate that are optimal for replication licensing are also optimal for sperm decondensation. We have shown that nucleoplasmin is required for XORC to be able to bind to the sperm chromatin, which represents the first known step in the establishment of replication origins. We therefore suggest that the licensing of sperm chromatin requires that it becomes decondensed by the removal of protamines to permit the binding of XORC and other licensing components (Fig. 7A–C).

Full reconstitution of licensed origins on defined DNA templates

Sperm nuclei can be licensed by a combination of purified nucleoplasmin, purified RLF-M and the BPAS fraction (containing XORC, XCdc6 and RLF-B enriched ~230-fold). Since RLF-B has yet to be purified to homogeneity (17), we are not in a position to determine whether there are any other essential licensing proteins present in the BPAS fraction. However, given the degree of enrichment of BPAS, it seems likely that these five proteins (nucleoplasmin, RLF-M, XORC, XCdc6 and RLF-B) represent most of the proteins required for assembly of functionally licensed replication origins on sperm nuclei.

Although Xenopus sperm nuclei represent a relatively simple chromatin template, there remains the possibility that they also contain proteins necessary for the establishment of replication origins. Of particular note is the observation that purified RLF-M interacts only weakly with naked DNA in the absence of other proteins (unpublished data and ref. 15), so it is possible that RLF-M binds to licensed DNA by interacting with other DNA-binding proteins. Neither XORC nor XCdc6 can play this role, since the binding of RLF-M to licensed DNA no longer depends on the presence of XORC or XCdc6 (24,31,50). One possible candidate for an RLF-M binding partner is histone H3, since Xenopus sperm nuclei contain their full complement of histones H3 (40,51) and certain MCM/P1 proteins have been shown to bind tightly to histone H3 in vitro (52). Reconstitution of functional replication origins on naked DNA templates will be necessary to investigate this point.

Role of chromatin structure in restricting ORC binding

It is currently unclear what influences the selection of sites where ORC binds to DNA in higher eukaryotes. In S.cerevisiae, ORC binds specifically to a conserved 12 bp motif in replication origins, the ARS consensus sequence (23). However, no such sequence specificity has been detected with Xenopus ORC (A.Rowles and J.J.Blow, unpublished data). The lack of sequence specificity of XORC may relate to the previously reported lack of sequence specificity of replication origins in the early Xenopus embryo (25–27,53). Despite this lack of sequence specificity, there is some evidence that replication origins are spaced at ~10 kb intervals in this system (25,27). Interestingly, XORC also saturates chromatin at ~1 molecule per 10 kb (20), and if the quantity of XORC is reduced, there is a proportionate decrease in replication rate (24). One possible explanation for these results is that other chromatin-bound proteins are required to direct the correct binding of XORC to DNA.

Our demonstration that nucleoplasmin is required to allow XORC to bind to sperm nuclei provides an example of how the association of ORC and DNA might be modulated by chromatin structure. In condensing the chromatin, it appears that protamines prevent XORC from being able to bind properly to the DNA. Other structural features of the chromatin, such as the overall organisation of nucleosomes, might limit the number of sites where XORC is capable of binding to the DNA. This type of effect has been seen when somatic linker histones were added to the Xenopus extract, causing a reduction in the number of replication origins assembled on sperm nuclei (54,55). Further investigation of the features that specify the position of replication origins on chromosomal DNA is needed to understand this.