Polyubiquitylation drives replisome disassembly at the termination of DNA replication

See allHide authors and affiliations

Science  24 Oct 2014:
Vol. 346, Issue 6208, pp. 477-481
DOI: 10.1126/science.1253585

How to stop after copying the genome

Replication is highly regulated: Failure to copy any part of the genome or copying parts of it more than once can cause genome instability with potentially disastrous consequences. Maric et al. and Priego Moreno et al. show that the DNA replication machinery, which stably encircles DNA during the duplication process, is actively disassembled once replication is complete (see the Perspective by Bell). The protein ring encircling the DNA is covalently modified, which allows it to be opened and the whole replication complex to be removed from DNA by a special disassembly complex.

Science, this issue 10.1126/science.1253596, p. 477; see also p. 418


Resolution of replication forks during termination of DNA replication is essential for accurate duplication of eukaryotic genomes. Here we present evidence consistent with the idea that polyubiquitylation of a replisome component (Mcm7) leads to its disassembly at the converging terminating forks because of the action of the p97/VCP/Cdc48 protein remodeler. Using Xenopus laevis egg extract, we have shown that blocking polyubiquitylation results in the prolonged association of the active helicase with replicating chromatin. The Mcm7 subunit is the only component of the active helicase that we find polyubiquitylated during replication termination. The observed polyubiquitylation is followed by disassembly of the active helicase dependent on p97/VCP/Cdc48. Altogether, our data provide insight into the mechanism of replisome disassembly during eukaryotic DNA replication termination.

During replication fork termination, the replisomes disassemble by an unknown mechanism, and topoisomerase II resolves the daughter DNA molecules (1, 2). If not resolved efficiently, terminating forks result in genomic instabilities (3, 4). In eukaryotic cells, the Mcm2-7 complex forms the core of the replicative DNA helicase (5). Its regulated loading onto chromatin (68) leads to replication origins licensing. The Mcm2-7 complexes stably associate with DNA and require replication progression to unload (9). A small proportion of chromatin-loaded Mcm2-7 is activated during S phase by the formation of the CMG complex [with Cdc45 and go ichi ni san (GINS)], which acts as a helicase and as an organizing center for the replisome (10, 11). In higher eukaryotes, minichromosome maintenance complex–binding protein (MCM-BP) facilitates the removal of the bulk of Mcm2-7 complexes from chromatin (12, 13), but the mechanism of active helicase disassembly upon termination of replication is unknown.

Our aim was to determine the role of ubiquitylation during DNA replication. Recombinant histidine (His6)–tagged ubiquitin added to cell-free Xenopus laevis egg extract is efficiently used by the extract without affecting the replication process (fig. S1, A to C). To inhibit polyubiquitylation, we used a chain-terminating mutant of ubiquitin, which has all lysine residues mutated to alanine (His6-UbiNOK). Addition of His6-UbiNOK to the replicating extract resulted in increased and prolonged association of the active replicative helicase with chromatin, whereas nascent DNA synthesis was not affected (Fig. 1A and figs. S1D and S2D) (14). The majority of Mcm2-7 complexes (90 to 95%), which represent the inactive origins, were removed from chromatin with normal timing (fig. S1, D and E), and the accumulated CMG components formed large-molecular-size complexes (fig. S1F).

Fig. 1 Partial inhibition of polyubiquitylation leads to active helicase accumulation on chromatin.

(A) DNA synthesis in interphase extract supplemented with His6-Ubi or His6-UbiNOK was assessed at indicated times (left) [average of three experiments in independent extracts with standard error of the mean (SEM)], whereas chromatin was isolated and subjected to Western blotting (right). For comparison, chromatin was also isolated from samples treated with aphidicolin (aph) or aphidicolin and caffeine (aph + caff). (B) Cdt1 activity in the extracts supplemented with His6-Ubi or His6-UbiNOK was blocked with geminin 3 min after sperm nuclei DNA addition. DNA synthesis in both extracts and in control extract (buffer control with no geminin addition) was assessed (left), whereas isolated minimally licensed chromatin was analyzed by Western blotting (right). (C) Extract was supplemented with 5-bromo-2′deoxyuridine 5′-triphosphate (BrdUTP) (visualized in red) and optionally with buffer, His6-UbiNOK, or ICRF-193. Chromatin was isolated at late S phase (90 min) and incubated in a second extract supplemented with geminin and digoxigenin-11-2′-deoxyuridine 5′-triphosphate (Dig-dUTP) (visualized in green). The number of short green-labeled fragments of replicated DNA (tracks) was scored per 100 kb of DNA (top) and the lengths of 70 green incorporation fragments were measured from three independent experiments each (bottom). The means of three independent experiments were plotted with SEM.

Polyubiquitylation with ubiquitin chains linked through lysine 48 (K48) and lysine 11 (K11) often drives proteasomal degradation of modified substrates (15), yet inhibition of proteasomal degradation with the small-molecule inhibitor bortezomib (16) (fig. S2, A and B) did not affect replicative helicase on chromatin (fig. S2, C and D). Ubiquitin-driven proteasomal degradation of Cdt1 is important for preventing rereplication during S phase (17). After adding His6-UbiNOK, we observed only a ∼15-min delay of its degradation (fig. S2E), whereas inhibition of proteasome results in complete stabilization of Cdt1 (fig. S2E). The inhibition of polyubiquitylation was thus not complete with His6-UbiNOK, which explained the temporary effect on helicase chromatin accumulation (Fig. 1A).

We next aimed to establish the underlying reason behind our observation. Addition of His6-UbiNOK did not affect the efficiency of DNA synthesis (Fig. 1A), the length of individual tracks of nascent DNA (fig. S3, A and B), or induced Chk1 phosphorylation (fig. S3C). These findings suggested that progression of the replicating forks was not affected by blocked polyubiquitylation.

We did not see helicase accumulation upon proteasomal inhibition and complete stabilization of Cdt1 (fig. S2C), but to exclude the possibility of Cdt1-driven rereplication (18) upon polyubiquitylation inhibition, we supplemented the extract with the Cdt1 inhibitor geminin shortly after adding DNA sperm and His6-UbiNOK (19). The amount of Mcm2-7 complexes loaded onto chromatin was severely reduced (fig. S3D), whereas CMG subunits still accumulated on chromatin when polyubiquitylation was inhibited (Fig. 1B). Moreover, as the majority of Mcm2-7 complexes became activated under the conditions of minimal licensing and thus behaved similarly to other CMG components, we could now observe also the accumulation of Mcm5 on chromatin (Fig. 1B). Blocking late-origin firing by the addition of a CDK inhibitor, p27(KIP1), in early S phase had no effect on the accumulation of CMG on chromatin when polyubiquitylation was inhibited (fig. S3E), nor was the interorigin distance affected by the addition of His6-UbiNOK (fig. S3F), which suggested that the number of activated origins was not affected by polyubiquitylation inhibition.

We then asked whether the inhibition of polyubiquitylation resulted in CMG accumulation on chromatin through a defect in helicase disassembly at termination. Unloading of helicase is an irreversible process (20), and CMG complexes stably encircle DNA without any dynamic exchange of molecules (fig. S3G), which leads to a requirement for a regulated process to remove them when forks terminate. Blocking polyubiquitylation also led to a replication termination defect as seen by synthesis of short fragments of DNA after release from His6-UbiNOK inhibition, similar to topoisomerase II inhibition [with the bisdioxopiperazine ICRF-193 (21)] replication termination defect (2) (Fig. 1C and fig. S3H). This defect did not create long stretches of single-stranded DNA, as Chk1 was not activated (fig. S3C).

Using mutants of ubiquitin containing only one lysine or having all but one lysine mutated, we found that K48 linkage of ubiquitin chains directed the disassembly defect (fig. S4, A to C), although proteasomal degradation was not involved (fig. S2, C and D). The cullin family of ubiquitin ligases was required for this disassembly, as the neddylation inhibitor MLN4924 (fig. S5, A to C) (22) also resulted in prolonged CMG association (fig. S5D).

Nickel-mediated pull-down of chromatin-bound proteins modified by His6-tagged ubiquitin under denaturing conditions showed that Mcm7 was the only CMG subunit detectably modified by multiple ubiquitins during S phase (Fig. 2A). Mcm7 ubiquitylation on chromatin coincided with the presence of replication forks on chromatin (Fig. 2B and fig. S6A) and was represented by mostly polyubiquitylated forms, as addition of His6-UbiNOK reduced their size (fig. S6A). Mcm7 was polyubiquitylated with K48-linked ubiquitin chains (Fig. 2C), and its ubiquitylation depended on the cullin family of E3s (Fig. 2D).

Fig. 2 Mcm7 is polyubiquitylated with K48-linked chains during DNA replication.

(A) Egg extract was supplemented with His6-Ubi and optional CDK inhibitor p27(KIP1); ubiquitylated chromatin proteins were pulled down and analyzed by Western blotting. (B) Ubiquitylated proteins were pulled down from replicating chromatin as in (A) at indicated times and analyzed by Western blotting. (C) Ubiquitylated proteins were pulled down as in (A) from S-phase chromatin, assembled in extract supplemented with indicated His6-Ubi mutants and analyzed by Western blotting. Two complementary sets of ubiquitin mutants were tested. (D) Ubiquitylated proteins were pulled down as in (A) from early S-phase chromatin, assembled in extract supplemented with indicated inhibitors, and analyzed by Western blotting.

Any inhibition of replication by DNA damage greatly reduced Mcm7 ubiquitylation (Fig. 3A). Blocking the B family of DNA polymerases with aphidicolin inhibited Mcm7 ubiquitylation, as did the addition of aphidicolin and caffeine combined, which resulted in accumulation of blocked forks on chromatin (Fig. 3B), which in turn suggested that Mcm7 is ubiquitylated only when forks can terminate. Indeed, blocking fork termination with the topoisomerase II inhibitor ICRF-193 (fig. S6B) greatly reduced Mcm7 ubiquitylation (Fig. 3C).

Fig. 3 Mcm7 is ubiquitylated at termination of replication forks.

(A) Extract was supplemented with indicated inhibitors and DNA damaging agents. The ability of each extract to replicate DNA (top) and chromatin-bound proteins (at 45 min) (bottom) were analyzed. (B) Extract was supplemented with indicated inhibitors, and the level of replicated DNA and chromatin-bound proteins were analyzed at indicated times as in (A). (C) Extract was supplemented with topoisomerase II inhibitor ICRF-193 or control dimethyl sulfoxide (DMSO) and analyzed as in (A).

Finally, we wished to establish the mechanism by which polyubiquitylation could drive the disassembly of the replicative helicase during termination. Blocking proteasomal degradation did not affect the level of ubiquitylated Mcm7 on chromatin (fig. S7A). Mass spectrometry analysis of factors interacting with Mcm3 during DNA replication identified 29 peptides of the protein remodeler p97/valosin-containing protein (VCP)/Cdc48 (hereafter, p97). p97 complexes recognize proteins modified with K48 ubiquitin chains and extract them either from endoplasmic reticulum (by endoplasmic reticulum–associated degradation) or chromatin (23). We confirmed the interaction of chromatin-associated Mcm3 with p97 during S phase (fig. S7B). Treatment of extracts with MLN4924 leads to the accumulation of the replicative helicase and lower-mobility ubiquitylated forms of Mcm7 on chromatin in later S phase. This treatment also led to an increase in Cdc45 and Psf1 (but not p97) interacting with Mcm3, which suggested that p97 interacted preferentially with fully ubiquitylated helicase (fig. S7B).

To test whether p97 is indeed required for active helicase disassembly during DNA replication termination, we supplemented extract with recombinant wild-type or inactive p97 mutant (24). Addition of either protein did not affect nascent DNA synthesis (Fig. 4A) or Cdt1 degradation (fig. S7C). However, supplementation with mutant p97 led to the prolonged association of replicative helicase with chromatin (Fig. 4A), accumulation of the ubiquitylated forms of Mcm7 on chromatin (Fig. 4C and fig. S7D), and replication forks termination defect (Fig. 4D). This suggests that p97-mediated remodeling acts after ubiquitylation of Mcm7 but before disassembly of the replisome at the termination of replication.

Fig. 4 p97-dependent unloading of active helicase from chromatin.

(A) Extract was supplemented with recombinant wild-type p97 (p97wt), adenosine triphosphatase (ATPase)–dead p97D1D2 mutant (p97mut) or buffer, and extract’s ability to replicate DNA was assessed (left) (average of three experiments), whereas chromatin was isolated at indicated times and analyzed by Western blotting (right). (B) The accumulation on chromatin of Psf2 and nonhelicase replisome component proliferating cell nuclear antigen (PCNA) at 90 min was quantified, normalized to loading control, and compared with buffer control (average of three experiments with SEM). (C) Extract was supplemented with mutant p97 or buffer; chromatin was isolated in the middle of S phase; and ubiquitylated proteins were pulled down as in Fig. 2A and analyzed by Western blotting. (D) Termination fibers were examined as in Fig. 1C but with buffer, p97wt, or p97mut in the first extract. The means of three independent experiments were plotted with SEM.

p97-deficient Caenorhabditis elegans embryos have previously been shown to exhibit persistent chromatin association of Cdc45/GINS during mitosis, which has been suggested as a consequence of Cdt1 stabilization (25). However, although p97 plays a role in Cdt1 degradation in specific situations [e.g., mitotic chromatin or after UV irradiation (25, 26)], it is less critical during unperturbed S phase (fig. S7C) (27), and the effect it has on the CMG complex disassembly is not driven through Cdt1 but by direct removal of ubiquitylated helicase from chromatin. Moreover, data obtained from budding yeast presented in the accompanying paper by Maric et al. (28) indicate that this mechanism of replisome disassembly at the termination of replication is conserved throughout evolution. Polyubiquitylation of Mcm7 has also been observed in human embryonic kidney 293 cells, although a function for it has not yet been reported (29).


Materials and Methods

Supplementary Text

Figs. S1 to S8

References (3040)

References and Notes

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 38.
  40. 40.
  41. Acknowledgments: We would like to thank K. Labib for sharing his data and critical discussions, J. Blow for support and help with egg extract, R. Hay for advice, O. Stemmann for vectors to express p97, and E. Petermann for help with fibers. A.G. would like to thank D. Tennant for support. This work was supported by U.K. Medical Research Council Career Development Award MR/K007106/1 to A.G. and Cancer Research UK Birmingham Centre award DF/270312 to A.G. S.P.M. was funded by ERASMUS scholarship and Wellcome Trust strategic award ISSFPP47; N.C. was funded by Cancer Research UK Birmingham Centre award CAICS/05/12.
View Abstract

Navigate This Article