DNA replication in quiescent cell nuclei: regulation by the nuclear envelope and chromatin structure

Quiescent nuclei from differentiated somatic cells can reacquire pluripotence, the capacity to replicate, and reinitiate a program of differentiation after transplantation into amphibian eggs. The replication of quiescent nuclei is recapitulated in extracts derived from activated Xenopus eggs; therefore, we have exploited this cell-free system to explore the mechanisms that regulate initiation of replication in nuclei from terminally differentiated Xenopus erythrocytes. We find that these nuclei lack many, if not all, pre-replication complex (pre-RC) proteins. Pre-RC proteins from the extract form a stable association with the chromatin of permeable nuclei, which replicate in this system, but not with the chromatin of intact nuclei, which do not replicate, even though these proteins cross an intact nuclear envelope. During extract incubation, the linker histones H1 and H1(0) are removed from erythrocyte chromatin by nucleoplasmin. We show that H1 removal facilitates the replication of permeable nuclei by increasing the frequency of initiation most likely by promoting the assembly of pre-RCs on chromatin. These data indicate that initiation in erythrocyte nuclei requires the acquisition of pre-RC proteins from egg extract and that pre-RC assembly requires the loss of nuclear envelope integrity and is facilitated by the removal of linker histone H1 from chromatin.     

Reactivation of DNA Replication in Nuclei from Terminally Differentiated Cells: Nuclear Membrane Permeabilization Is Required for Initiation inXenopusEgg Extract

We have usedXenopusegg extract to investigate the requirements for reactivation of DNA replication in nuclei isolated from terminally differentiated chicken erythrocytes. Previous work has shown that reactivation of erythrocyte nuclei in egg extract is accompanied by chromatin decondensation, nuclear envelope reformation, and the accumulation of egg lamin, L_III. However, in those studies, erythrocyte nuclei were prepared by methods that were not designed to maintain the selective permeability of the nuclear membrane, and as such, it is not clear if loss of nuclear membrane integrity played a role in the reactivation process. Therefore, the purpose of this study was to determine if changes in nuclear membrane permeability are required for reactivation of erythrocyte nuclei in egg extract. Nuclei with intact nuclear membranes were prepared from erythrocytes with streptolysin O and permeable nuclei by treatment of intact nuclei with the detergent Nonidet-P40. Like permeable nuclei, most intact nuclei decondensed, imported nuclear protein, and accumulated lamin L_IIIfrom the extract. However, unlike permeable nuclei, which replicated extensively in the extract, few intact nuclei initiated replication under the same conditions. These data demonstrate that permeabilization of the nuclear membrane is required for reactivation of DNA replication in terminally differentiated erythrocyte nuclei by egg extract and suggest that loss of nuclear membrane integrity may be a general requirement for replication of quiescent cell nuclei by this system.     

The roles of the MCM, ORC, and Cdc6 proteins in determining the replication competence of chromatin in quiescent cells

Most eukaryotic cell types can withdraw from proliferative cell cycles and remain quiescent for extended periods. Intact nuclei isolated from quiescent murine NIH3T3 cells fail to replicate in vitro when incubated in Xenopus egg extracts, although intact nuclei from proliferating cells replicate well. Permeabilization of the nuclear envelope rescues the ability of quiescent nuclei to replicate in the extract. We show that origin replication complex (ORC), minichromosome maintenance (MCM), and Cdc6 proteins are all present in early quiescent cells. Immunodepletion of Cdc6 or the MCM complex from Xenopus egg extract inhibits replication of permeable, quiescent, but not proliferating, NIH3T3 nuclei. Immunoblotting results demonstrate that mouse homologues of Mcm2, Mcm5, and Cdc6 are displaced from chromatin in quiescent cells. However, this absence of chromatin-bound Cdc6 and MCM proteins from quiescent cells appears not to be due to the absence of ORC subunits as murine homologues of Orc1 and Orc2 remain chromatin-bound in quiescent cells. Surprisingly, intact quiescent nuclei fail to bind exogenously added XCdc6 or to replicate in Xenopus egg extracts immunodepleted of ORC, even though G1- or S-phase nuclei still replicate in these extracts. Our results identify Cdc6 and the MCM complex as essential replication components absent from quiescent chromatin due to nonfunctional chromatin-bound ORC proteins. These results can explain why quiescent mammalian nuclei are unable to replicate in vivo and in Xenopus egg extracts.     

Licensing of DNA replication by a multi-protein complex of MCM/P1 proteins in Xenopus eggs.

In eukaryotes, chromosomal DNA is licensed for a single round of replication in each cell cycle. Xenopus MCM3 protein has been implicated in the licensing of replication in egg extract. We have cloned cDNAs encoding five immunologically distinct proteins associated with Xenopus MCM3 as members of the MCM/P1 family. Six Xenopus MCM proteins formed a physical complex in the egg extract, bound to unreplicated chromatin before the formation of nuclei, and apparently displaced from replicated chromatin. The requirement of six XMCM proteins for the replication activity of the egg extract before nuclear formation suggests that their re-association with replicated chromatin at the end of the mitotic cell cycle is a key step for the licensing of replication.     

Reactivation of DNA replication in erythrocyte nuclei by Xenopus egg extract involves energy-dependent chromatin decondensation and changes in histone phosphorylation.

Reactivation of chicken erythrocyte nuclei for DNA replication in Xenopus egg extracts involves two phases of chromatin remodelling: a fast decondensation leading to a small volume increase and chromatin dispersion occurring within a few minutes (termed stage I decondensation), followed by a slower membrane-dependent decondensation and enlargement of up to 40-fold from the initial volume (stage II decondensation). Chromatin decondensation as measured by nuclear swelling and micrococcal nuclease digestion required ATP. We observed a characteristic change in the phosphorylation pattern of erythrocyte proteins upon incubation in egg extract. While histones H5, H2A, and H4 became selectively phosphorylated during decondensation, the phosphorylation of histone H3 and of several nonhistone proteins was prevented. Furthermore, histone H5 was selectively released from erythrocyte nuclei in an energy-dependent reaction. These molecular changes already occurred during stage I decondensation and they persisted during stage II decondensation. DNA replication was confined to nuclei of stage II decondensation which incorporated lamin LIII from the egg extract. These results show that initiation of DNA replication in chicken erythrocytes requires in addition to ATP-dependent chromatin remodelling (stage I), further changes in chromatin structure that correlates with lamin LIII incorporation, and stage II decondensation.     

Nuclear proteins of quiescent Xenopus laevis cells inhibit DNA replication in intact and permeabilized nuclei

Quiescent cells from adult vertebrate liver and contact-inhibited or serum-deprived tissue cultures are active metabolically but do not carry out nuclear DNA replication and cell division. Replication of intact nuclei isolated from either quiescent Xenopus liver or cultured Xenopus A6 cells in quiescence was barely detectable in interphase extracts of Xenopus laevis eggs, although Xenopus sperm chromatin was replicated with approximately 100% efficiency in the same extracts. Permeabilization of nuclei from quiescent Xenopus liver or cultured Xenopus epithelial A6 cells did not facilitate efficient replication in egg extracts. Moreover, replication of Xenopus sperm chromatin in egg extracts was strongly inhibited by a soluble extract of isolated Xenopus liver nuclei; in contrast, complementary-strand synthesis on single-stranded DNA templates in egg extracts was not affected. Inhibition was specific to endogenous molecules localized preferentially in quiescent as opposed to proliferating cell nuclei, and was not due to suppression of cdk2 kinase activity. Extracts of Xenopus liver nuclei also inhibited growth of sperm nuclei formed in egg extracts. However, the rate and extent of decondensation of sperm chromatin in egg extracts were not affected. The formation of prereplication centers detected by anti-RP-A antibody was not affected by extracts of liver nuclei, but formation of active replication foci was blocked by the same extracts. Inhibition of DNA replication was alleviated when liver nuclear extracts were added to metaphase egg extracts before or immediately after Ca++ ion-induced transition to interphase. A plausible interpretation of our data is that endogenous inhibitors of DNA replication play an important role in establishing and maintaining a quiescent state in Xenopus cells, both in vivo and in cultured cells, perhaps by negatively regulating positive modulators of the replication machinery.     

Cyclin A-dependent kinase activity affects chromatin binding of ORC, Cdc6, and MCM in egg extracts of Xenopus laevis.

The initiation of DNA replication in eukaryotes requires the loading of the origin recognition complex (ORC), Cdc6, and minichromosome maintenance (MCM) proteins onto chromatin to form the preinitiation complex. In Xenopus egg extract, the proteins Orc1, Orc2, Cdc6, and Mcm4 are underphosphorylated in interphase and hyperphosphorylated in metaphase extract. We find that chromatin binding of ORC, Cdc6, and MCM proteins does not require cyclin-dependent kinase activities. High cyclin A-dependent kinase activity inhibits the binding and promotes the release of Xenopus ORC, Cdc6, and MCM from sperm chromatin, but has no effect on chromatin binding of control proteins. Cyclin A together with ORC, Cdc6 and MCM proteins is bound to sperm chromatin in DNA replicating pseudonuclei. In contrast, high cyclin E/cdk2 was not detected on chromatin, but was found soluble in the nucleoplasm. High cyclin E kinase activity allows the binding of Xenopus ORC and Cdc6, but not MCM, to sperm chromatin, even though the kinase does not phosphorylate MCM directly. We conclude that chromatin-bound cyclin A kinase controls DNA replication by protein phosphorylation and chromatin release of Cdc6 and MCM, whereas soluble cyclin E kinase prevents rereplication during the cell cycle by the inhibition of premature MCM chromatin association.     

In vitro nuclear assembly with affinity-purified nuclear envelope precursor vesicle fractions, PV1 and PV2.

Nuclear envelope precursor vesicles were affinity purified from a Xenopus egg extract by a chromatin binding method. Vesicles bound to chromatin at 4 degrees C were dissociated with a high salt buffer and further fractionated into nuclear envelope precursor vesicle fractions 1 (PV1) and 2 (PV2) by differential centrifugation. PV1 contained larger vesicles. When chromatin was incubated in a Xenopus egg cytosol fraction supplemented with PV1, vesicles bound to chromatin, fused with each other, formed a bilayered nuclear envelope, and assembled into spherical small nuclei. However, the thus assembled nuclei did not grow to the normal size. Nuclear pore complexes were not found on the thus assembled nuclei. On the other hand, PV2 contained smaller vesicles. PV2 vesicles bound to chromatin, fused little with each other in the Xenopus egg cytosol fraction, and no nuclei were assembled. When PV1 supplemented with PV2 was used for the nuclear assembly reaction, the assembled nuclei grew to the normal size. Nuclear pore complexes existed in the thus assembled nuclear envelopes. These results suggested that 1) two vesicle populations, PV1 and PV2, are necessary for the assembly of normal sized nuclei, 2) PV1 contains a chromatin targeting molecule(s) and membrane fusion machinery, 3) PV2 contains a chromatin targeting molecule(s) and a molecule(s) necessary for nuclear pore complex assembly, and 4) PV1 has the ability to assemble a nuclear membrane, and PV2 is necessary for the assembly of nuclear pore complexes and for nuclei to grow to the normal size. An in vitro nuclear assembly system constituted with affinity-purified vesicle fractions, PV1 and PV2, was established.     

利用非洲爪蟾精子染色质和卵提取物在体外重建细胞核

应用非洲爪蟾去膜精子染色质和卵提取物成功地进行了细胞核本外重建。当精子染色质加入卵提取物后,首先发生染色质去浓缩作用,染色质整体结构膨胀;膜泡在膨胀的染色质外周聚集并逐渐彼此融合成双层膜;核孔复合体以某种未知方式组装入双层膜而形成核膜结构,并逐渐完全覆盖膨大的染色质,最终形成典型的间期核结构。     

Initiation of DNA replication in nuclei from quiescent cells requires permeabilization of the nuclear membrane

We have investigated the replication capacity of intact nuclei from quiescent cells using Xenopus egg extract. Nuclei, with intact nuclear membranes, were isolated from both exponentially growing and contact- inhibited BALB/c 3T3 fibroblasts by treatment of the cells with streptolysin-O. Flow cytometry showed that > 90% of all contact- inhibited cells and approximately 50% of the exponential cells were in G0/G1-phase at the time of nuclear isolation. Intact nuclei were assayed for replication in the extract by incorporation of [alpha- 32P]dATP or biotin-dUTP into nascent DNA. Most nuclei from exponential cells replicated in the egg extract, consistent with previous results showing that intact G1 nuclei from HeLa cells replicate in this system. In contrast, few nuclei from quiescent cells replicated in parallel incubations. However, when the nuclear membranes of these intact quiescent nuclei were permeabilized with lysophosphatidylcholine prior to addition to the extract, nearly all the nuclei replicated under complete cell cycle control in a subsequent incubation. The ability of LPC-treated quiescent nuclei to undergo DNA replication was reversed by resealing permeable nuclear membranes with Xenopus egg membranes prior to extract incubation demonstrating that the effect of LPC treatment is at the level of the nuclear membrane. These results indicate that nuclei from G1-phase cells lose their capacity to initiate DNA replication following density-dependent growth arrest and suggest that changes in nuclear membrane permeability may be required for the initiation of replication upon re-entry of the quiescent cell into the cell cycle.     

Characterization of the membrane binding and fusion events during nuclear envelope assembly using purified components

At the end of mitosis membrane vesicles are targeted to the surface of chromatin and fuse to form a continuous nuclear envelope. To investigate the molecular mechanisms underlying these steps in nuclear envelope assembly, we have developed a defined cell-free system in which the binding and fusion steps in nuclear envelope assembly can be examined separately. We have found that extensively boiled Xenopus egg extracts efficiently promote the decondensation of demembranated Xenopus sperm chromatin. When isolated membranes are added to this decondensed chromatin a specific subfraction of membrane vesicles (approximately 70 nM in diameter) bind to the chromatin, but these vesicles do not fuse to each other. Vesicle binding is independent of ATP and insensitive to N-ethylmalamide. Quantitative analysis of these sites by EM suggests that there is at least one vesicle binding site per 100 kb of chromosomal DNA. We show by tryptic digestion that vesicle-chromatin association requires proteins on both the vesicle and on the chromatin. In addition, we show that the vesicles bound under these conditions will fuse into an intact nuclear envelope when incubated with the soluble fraction of a Xenopus egg nuclear assembly extract. With respect to vesicle fusion, we have found that vesicles prebound to chromatin will fuse to each other when ATP and GTP are present in the boiled extract. These results indicate that nuclear envelope assembly is mediated by a subset of approximately 70-nM-diam vesicles which bind to chromatin sites spaced 100 kb apart and that fusion of these vesicles is regulated by membrane-associated GTP-binding proteins.     

Mitotic remodeling of the replicon and chromosome structure

Animal cloning by nuclear-transfer experiments frequently fails due to the inability of transplanted nuclei to support normal embryonic development. We show here that the formation of mitotic chromosomes in the egg context is crucial for adapting differentiated nuclei for early development. Differentiated erythrocyte nuclei replicate inefficiently in Xenopus eggs but do so as rapidly as sperm nuclei if a prior single mitosis is permitted. This mitotic remodeling involves a topoisomerase II-dependent shortening of chromatin loop domains and an increased recruitment of replication initiation factors onto chromatin, leading to a short interorigin spacing characteristic of early developmental stages. It also occurs within each early embryonic cell cycle and dominantly regulates initiation of DNA replication for the subsequent S phase. These results indicate that mitotic conditioning is crucial to reset the chromatin structure of differentiated adult donor cells for embryonic DNA replication and suggest that it is an important step in nuclear cloning.     

The RLF-B component of the replication licensing system is distinct from Cdc6 and functions after Cdc6 binds to chromatin

Replication licensing factor (RLF) is an essential initiation factor that can prevent re-replication of DNA in a single cell cycle [1] [2]. It is required for the initiation of DNA replication, binds to chromatin early in the cell cycle, is removed from chromatin as DNA replicates and is unable to re-bind replicated chromatin until the following mitosis. Chromatography of RLF from Xenopus extracts has shown that it consists of two components termed RLF-B and RLF-M [3]. The RLF-M component consists of complexes of all six Xenopus minichromosome maintenance (MCM/P1) proteins (XMcm2-7), which bind to chromatin in late mitosis and are removed as replication occurs [3] [4] [5] [6] [7]. The identity of RLF-B is currently unknown. At least two factors must be present on chromatin before licensing can occur: the Xenopus origin recognition complex (XORC) [8] [9] and Xenopus Cdc6 (XCdc6) [10]. XORC saturates Xenopus sperm chromatin at approximately one copy per replication origin whereas XCdc6 binds to chromatin only if XORC is bound first [9] [10] [11]. Although XORC has been shown to be a distinct activity from RLF-B [9], the relationship between XCdc6 and RLF-B is currently unclear. Here, we show that active XCdc6 is loaded onto chromatin in extracts with defective RLF, and that both RLF-M and RLF-B are still required for the licensing of XCdc6-containing chromatin. Furthermore, RLF-B can be separated from XCdc6 by immunoprecipitation and standard chromatography. These experiments demonstrate that RLF-B is both functionally and physically distinct from XCdc6, and that XCdc6 is loaded onto chromatin before RLF-B function is executed.     

Replication of Xenopus erythrocyte nuclei in a homologous egg extract requires prior proteolytic treatment

Reactivation and reinitiation of DNA replication in quiescent frog erythrocyte nuclei has been analyzed following incubation in extracts prepared from activated Xenopus eggs. Nuclear decondensation and DNA synthesis only occurred if nuclei were pretreated with low doses of trypsin. This protease treatment did not digest histones, but did degrade several nonhistone proteins. Activated erythrocyte nuclei swell and begin DNA synthesis by 30 min after being mixed with the egg extract. In some extracts virtually complete genome replication was achieved in all nuclei after 2-3 hr. Addition of several protease inhibitors during sperm nuclear isolation significantly reduced the template efficiency of these preparations. We concluded that proteolytic alteration of nonhistone nuclear structural proteins may be a general mechanism which permits quiescent nuclei to reenter the replication cycle. Erythrocyte nuclei and egg extracts provide an excellent experimental system in which to investigate the processes of nuclear reactivation.     

Reversible effects of nuclear membrane permeabilization on DNA replication: evidence for a positive licensing factor

We have investigated the mechanism which prevents reinitiation of DNA replication within a single cell cycle by exploiting the observation that intact G2 HeLa nuclei do not replicate in Xenopus egg extract, unless their nuclear membranes are first permeabilized (Leno et al., 1992). We have asked if nuclear membrane permeabilization allows escape of a negative inhibitor from the replicated nucleus or entry of a positive activator as proposed in the licensing factor hypothesis of Blow and Laskey (1988). We have distinguished these possibilities by repairing permeabilized nuclear membranes after allowing soluble factors to escape. Membrane repair of G2 nuclei reverses the effects of permeabilization arguing that escape of diffusible inhibitors is not sufficient to allow replication, but that entry of diffusible activators is required. Membrane repair has no significant effect on G1 nuclei. Pre-incubation of permeable G2 nuclei in the soluble fraction of egg extract before membrane repair allows semiconservative DNA replication of these nuclei when incubated in complete extract. Addition of the same fraction after membrane repair has no effect. Our results provide direct evidence for a positively acting "licensing" activity which is excluded form the interphase nucleus by the nuclear membrane. Nuclear membrane permeabilization and repair can be used as an assay for licensing activity which could lead to its purification and subsequent analysis of its action within the nucleus.     

The nuclear membrane prevents replication of human G2 nuclei but not G1 nuclei in Xenopus egg extract.

We have used synchronized HeLa cells to investigate the role of the nuclear membrane in preventing rereplication in a single cell cycle. Nuclei were prepared with intact nuclear membranes using streptolysin-O or digitonin and assayed for replication in Xenopus egg extracts. Intact G1 nuclei replicate semiconservatively, but intact G2 nuclei do not replicate in egg extract. However, permeabilizing the nuclear membranes of G2 nuclei by treatment with NP-40 allows them all to replicate in egg extract under cell cycle control, suggesting that integrity of the nuclear membrane is required to distinguish G2 from G1 human nuclei and to prevent rereplication within a single cell cycle. The results are discussed in terms of the previously proposed licensing factor model.     

Cyclin E uses Cdc6 as a chromatin-associated receptor required for DNA replication

Using an in vitro chromatin assembly assay in Xenopus egg extract, we show that cyclin E binds specifically and saturably to chromatin in three phases. In the first phase, the origin recognition complex and Cdc6 prereplication proteins, but not the minichromosome maintenance complex, are necessary and biochemically sufficient for ATP-dependent binding of cyclin E--Cdk2 to DNA. We find that cyclin E binds the NH(2)-terminal region of Cdc6 containing Cy--Arg-X-Leu (RXL) motifs. Cyclin E proteins with mutated substrate selection (Met-Arg-Ala-Ile-Leu; MRAIL) motifs fail to bind Cdc6, fail to compete with endogenous cyclin E--Cdk2 for chromatin binding, and fail to rescue replication in cyclin E--depleted extracts. Cdc6 proteins with mutations in the three consensus RXL motifs are quantitatively deficient for cyclin E binding and for rescuing replication in Cdc6-depleted extracts. Thus, the cyclin E--Cdc6 interaction that localizes the Cdk2 complex to chromatin is important for DNA replication. During the second phase, cyclin E--Cdk2 accumulates on chromatin, dependent on polymerase activity. In the third phase, cyclin E is phosphorylated, and the cyclin E--Cdk2 complex is displaced from chromatin in mitosis. In vitro, mitogen-activated protein kinase and especially cyclin B--Cdc2, but not the polo-like kinase 1, remove cyclin E--Cdk2 from chromatin. Rebinding of hyperphosphorylated cyclin E--Cdk2 to interphase chromatin requires dephosphorylation, and the Cdk kinase-directed Cdc14 phosphatase is sufficient for this dephosphorylation in vitro. These three phases of cyclin E association with chromatin may facilitate the diverse activities of cyclin E--Cdk2 in initiating replication, blocking rereplication, and allowing resetting of origins after mitosis.