Eukaryotic DNA replication terminates when replisomes from adjacent replication origins converge. Termination involves local completion of DNA synthesis, decatenation of daughter molecules and replisome disassembly. Termination has been difficult to study because termination events are generally asynchronous and sequence nonspecific. To overcome these challenges, we paused converging replisomes with a site-specific barrier in Xenopus egg extracts. Upon removal of the barrier, forks underwent synchronous and site-specific termination, allowing mechanistic dissection of this process. We show that DNA synthesis does not slow detectably as forks approach each other, and that leading strands pass each other unhindered before undergoing ligation to downstream lagging strands. Dissociation of the replicative CMG helicase (comprising CDC45, MCM2-7 and GINS) occurs only after the final ligation step, and is not required for completion of DNA synthesis, strongly suggesting that converging CMGs pass one another and dissociate from double-stranded DNA. This termination mechanism allows rapid completion of DNA synthesis while avoiding premature replisome disassembly.
Subscribe to Journal
Get full journal access for 1 year
only $3.90 per issue
All prices are NET prices.
VAT will be added later in the checkout.
Rent or Buy article
Get time limited or full article access on ReadCube.
All prices are NET prices.
Levene, A. J., Kang, H. S. & Billheimer, F. E. DNA replication in SV40 infected cells. I. Analysis of replicating SV40 DNA. J. Mol. Biol. 50, 549–568 (1970)
Fanning, E. & Zhao, K. SV40 DNA replication: from the A gene to a nanomachine. Virology 384, 352–359 (2009)
Tapper, D. P. & DePamphilis, M. L. Discontinuous DNA replication: Accumulation of simian virus 40 DNA at specific stages in its replication. J. Mol. Biol. 120, 401–422 (1978)
Seidman, M. M. & Salzman, N. P. Late replicative intermediates are accumulated during simian virus 40 DNA replication in vivo and in vitro. J. Virol. 30, 600–609 (1979)
Sundin, O. & Varshavsky, A. Terminal stages of SV40 DNA replication proceed via multiply intertwined catenated dimers. Cell 21, 103–114 (1980)
Ishimi, Y., Sugasawa, K., Hanaoka, F., Eki, T. & Hurwitz, J. Topoisomerase II plays an essential role as a swivelase in the late stage of SV40 chromosome replication in vitro. J. Biol. Chem. 267, 462–466 (1992)
Hiasa, H. & Marians, K. J. Two distinct modes of strand unlinking during theta-type DNA replication. J. Biol. Chem. 271, 21529–21535 (1996)
Espeli, O., Levine, C., Hassing, H. & Marians, K. J. Temporal regulation of topoisomerase IV activity in E. coli. Mol. Cell 11, 189–201 (2003)
Segawa, M., Sugano, S. & Yamaguchi, N. Association of simian virus 40 T antigen with replicating nucleoprotein complexes of simian virus 40. J. Virol. 35, 320–330 (1980)
Tack, L. C. & DePamphilis, M. L. Analysis of simian virus 40 chromosome-T-antigen complexes: T-antigen is preferentially associated with early replicating DNA intermediates. J. Virol. 48, 281–295 (1983)
Chen, M. C., Birkenmeier, E. & Salzman, N. P. Simian virus 40 DNA replication: characterization of gaps in the termination region. J. Virol. 17, 614–621 (1976)
Sundin, O. & Varshavsky, A. Arrest of segregation leads to accumulation of highly intertwined catenated dimers: dissection of the final stages of SV40 DNA replication. Cell 25, 659–669 (1981)
Ivessa, A. S., Zhou, J. Q. & Zakian, V. A. The Saccharomyces Pif1p DNA helicase and the highly related Rrm3p have opposite effects on replication fork progression in ribosomal DNA. Cell 100, 479–489 (2000)
Steinacher, R., Osman, F., Dalgaard, J. Z., Lorenz, A. & Whitby, M. C. The DNA helicase Pfh1 promotes fork merging at replication termination sites to ensure genome stability. Genes Dev. 26, 594–602 (2012)
Fachinetti, D. et al. Replication termination at eukaryotic chromosomes is mediated by Top2 and occurs at genomic loci containing pausing elements. Mol. Cell 39, 595–605 (2010)
DiNardo, S., Voelkel, K. & Sternglanz, R. DNA topoisomerase II mutant of Saccharomyces cerevisiae: topoisomerase II is required for segregation of daughter molecules at the termination of DNA replication. Proc. Natl Acad. Sci. USA 81, 2616–2620 (1984)
Baxter, J. & Diffley, J. F. Topoisomerase II inactivation prevents the completion of DNA replication in budding yeast. Mol. Cell 30, 790–802 (2008)
Lucas, I., Germe, T., Chevrier-Miller, M. & Hyrien, O. Topoisomerase II can unlink replicating DNA by precatenane removal. EMBO J. 20, 6509–6519 (2001)
Gaggioli, V., Le Viet, B., Germe, T. & Hyrien, O. DNA topoisomerase II controls replication origin cluster licensing and firing time in Xenopus egg extracts. Nucleic Acids Res. 41, 7313–7331 (2013)
Ilves, I., Petojevic, T., Pesavento, J. J. & Botchan, M. R. Activation of the MCM2–7 helicase by association with Cdc45 and GINS proteins. Mol. Cell 37, 247–258 (2010)
Pacek, M., Tutter, A. V., Kubota, Y., Takisawa, H. & Walter, J. C. Localization of MCM2–7, Cdc45, and GINS to the site of DNA unwinding during eukaryotic DNA replication. Mol. Cell 21, 581–587 (2006)
Moyer, S. E., Lewis, P. W. & Botchan, M. R. Isolation of the Cdc45/Mcm2–7/GINS (CMG) complex, a candidate for the eukaryotic DNA replication fork helicase. Proc. Natl Acad. Sci. USA 103, 10236–10241 (2006)
Kang, Y. H., Galal, W. C., Farina, A., Tappin, I. & Hurwitz, J. Properties of the human Cdc45/Mcm2–7/GINS helicase complex and its action with DNA polymerase ε in rolling circle DNA synthesis. Proc. Natl. Acad. Sci. USA 109, 6042–6047 (2012)
Maric, M., Maculins, T., De Piccoli, G. & Labib, K. Cdc48 and a ubiquitin ligase drive disassembly of the CMG helicase at the end of DNA replication. Science 346, 1253596–1253596 (2014)
Priego Moreno, S., Bailey, R., Campion, N., Herron, S. & Gambus, A. Polyubiquitylation drives replisome disassembly at the termination of DNA replication. Science 346, 477–481 (2014)
Hiasa, H. & Marians, K. J. Tus prevents overreplication of oriC plasmid DNA. J. Biol. Chem. 269, 26959–26968 (1994)
Rudolph, C. J., Upton, A. L., Stockum, A., Nieduszynski, C. A. & Lloyd, R. G. Avoiding chromosome pathology when replication forks collide. Nature 500, 608–611 (2013)
Hook, S. S., Lin, J. J. & Dutta, A. Mechanisms to control rereplication and implications for cancer. Curr. Opin. Cell Biol. 19, 663–671 (2007)
Czajkowsky, D. M., Liu, J., Hamlin, J. L. & Shao, Z. DNA combing reveals intrinsic temporal disorder in the replication of yeast chromosome VI. J. Mol. Biol. 375, 12–19 (2008)
McGuffee, S. R., Smith, D. J. & Whitehouse, I. Quantitative, genome-wide analysis of eukaryotic replication initiation and termination. Mol. Cell 50, 123–135 (2013)
Yardimci, H., Loveland, A. B., Habuchi, S., van Oijen, A. M. & Walter, J. C. Uncoupling of sister replisomes during eukaryotic DNA replication. Mol. Cell 40, 834–840 (2010)
Loveland, A. B., Habuchi, S., Walter, J. C. & van Oijen, A. M. A general approach to break the concentration barrier in single-molecule imaging. Nature Methods 9, 987–992 (2012)
Santamaría, D. et al. Bi-directional replication and random termination. Nucleic Acids Res. 28, 2099–2107 (2000)
Zhang, J. et al. DNA interstrand cross-link repair requires replication fork convergence. Nature Struct. Mol. Biol. 22, 242–247 (2015)
Duxin, J. P. P., Dewar, J. M. M., Yardimci, H. & Walter, J. C. C. Replication-coupled repair of a DNA-protein crosslink. Cell 159, 346–357 (2014)
Walter, J., Sun, L. & Newport, J. Regulated chromosomal DNA replication in the absence of a nucleus. Mol. Cell 1, 519–529 (1998)
Sofueva, S. et al. Ultrafine anaphase bridges, broken DNA and illegitimate recombination induced by a replication fork barrier. Nucleic Acids Res. 39, 6568–6584 (2011)
Charbin, A., Bouchoux, C. & Uhlmann, F. Condensin aids sister chromatid decatenation by topoisomerase II. Nucleic Acids Res. 42, 340–348 (2014)
Laskey, R. A., Mills, A. D. & Morris, N. R. Assembly of SV40 chromatin in a cell-free system from Xenopus eggs. Cell 10, 237–243 (1977)
Räschle, M. et al. Mechanism of replication-coupled DNA interstrand crosslink repair. Cell 134, 969–980 (2008)
Long, D. T. T., Joukov, V., Budzowska, M. & Walter, J. C. C. BRCA1 promotes unloading of the CMG helicase from a stalled DNA replication fork. Mol. Cell 56, 174–185 (2014)
Fu, Y. V. et al. Selective bypass of a lagging strand roadblock by the eukaryotic replicative DNA helicase. Cell 146, 931–941 (2011)
Costa, A. et al. The structural basis for MCM2–7 helicase activation by GINS and Cdc45. Nature Struct. Mol. Biol. 18, 471–477 (2011)
Kaplan, D. L., Davey, M. J. & O’Donnell, M. Mcm4,6,7 uses a ‘pump in ring’ mechanism to unwind DNA by steric exclusion and actively translocate along a duplex. J. Biol. Chem. 278, 49171–49182 (2003)
Prokhorova, T. A., Mowrer, K., Gilbert, C. H. & Walter, J. C. DNA replication of mitotic chromatin in Xenopus egg extracts. Proc. Natl Acad. Sci. USA 100, 13241–13246 (2003)
Lebofsky, R., Takahashi, T. & Walter, J. C. DNA replication in nucleus-free Xenopus egg extracts. Methods Mol. Biol. 521, 229–252 (2009)
Budzowska, M., Graham, T. G. W., Sobeck, A., Waga, S. & Walter, J. C. Regulation of the Rev1-pol ζ complex during bypass of a DNA interstrand cross-link. EMBO J. 34, 1971–1985 (2015)
Magnaghi, P. et al. Covalent and allosteric inhibitors of the ATPase VCP/p97 induce cancer cell death. Nature Chem. Biol. 9, 548–556 (2013)
Walter, J. C. & Newport, J. Initiation of eukaryotic DNA replication: origin unwinding and sequential chromatin association of Cdc45, RPA, and DNA polymerase α. Mol. Cell 5, 617–627 (2000)
Long, D. T., Räschle, M., Joukov, V. & Walter, J. C. Mechanism of RAD51-dependent DNA interstrand cross-link repair. Science 333, 84–87 (2011)
We thank C. Richardson and members of the Walter laboratory for feedback on the manuscript. We thank K. J. Marians and J. T. Yeeles for plasmids and the LacI purification protocol. J.C.W. was supported by NIH grants GM62267 and GM80676. J.C.W. is an investigator of the Howard Hughes Medical Institute.
The authors declare no competing financial interests.
Extended data figures and tables
a, To investigate whether a LacR array blocks replication forks, a plasmid containing a tandem array of 16 lac operator (lacO) sequences, p[lacO16] (or p[lacOx16]), was incubated with buffer or LacR and then replicated in egg extract containing [α-32P]dATP. Radiolabelled replication intermediates were cleaved with XmnI (far left cartoon) and separated according to size and shape by 2D gel electrophoresis (see schematic of 2D gel). As replication neared completion at 4.5 min, mainly linear molecules were produced in the presence of buffer (orange arrowhead). In contrast, in the presence of LacR, a discrete spot appeared on the double-Y arc (blue arrowhead), demonstrating that converging replication forks accumulate at a specific locus on p[lacO16]. These data indicate that 16 copies of LacR block replication forks. b–f, To test whether the double-Y structures observed in panel a arose from replication forks stalling at the outer edges of the lacO array, we tested whether LacR specifically inhibited replication of lacO sequences. To this end, p[lacO16] (c) and the parental plasmid lacking lacO repeats, p[empty] (b), were incubated in the presence of buffer or LacR and replicated using Xenopus egg extracts containing [α-32P]dATP. Radiolabelled replication intermediates were cleaved with AflIII and PvuII to release the 2,354-bp plasmid backbone (b and c) and a 294-bp control fragment from p[empty] (b) or a 794-bp lacO fragment from p[lacO16] (c). The plasmid backbone and the respective inserts were separated on a native gel and detected by autoradiography (d). A longer exposure of the small fragments is shown, since they are less intense than the large fragments. The results in panel d were quantified in e and f. Notably, LacR specifically inhibited replication of the lacO-containing fragment in p[lacO16] (f, blue circles) but not the control fragment in p[empty] (e, green circles). We conclude that LacR prevents replication of the lacO array and that the double-Ys in panel a represent forks converged on the outer edges of the array. Importantly, synthesis within the 2,354-bp backbone fragment (f, orange circles) of p[lacO16] was not inhibited in the presence of LacR, indicating that no global structural changes occur that inhibit replication.
a, The gel shown in Fig. 1e was overexposed and shown in its entirety so that the smaller leftward strands (LWS, Fig. 1d) could be detected. As observed for the rightward strands (RWS, Fig. 1e), LWS rapidly increased in size and then disappeared as they were ligated to produce full-length strands (FLS, Fig. 1e). b–e, To determine whether the heterogeneity of LWS (a) and RWS (Fig. 1e) was due to delayed extension of lagging strands, or because a significant fraction of leading strands did not restart upon IPTG addition, we specifically monitored leading strand progression upon IPTG addition on p[lacO16]. To this end, DNA samples were treated with Nt.BspQI or Nb.BsrDI to specifically liberate the rightward or leftward leading strands, respectively (b), and DNAs were separated on a denaturing agarose gel (c). Before IPTG addition, discrete leading strand products of the expected size were observed (lanes 2 and 10). The presence of two stall products reflects the fact that at a slow rate, the replisome bypasses LacR (see also Fig. 3). Upon IPTG addition, these species rapidly and completely shifted up the gel, indicating that rightward and leftward leading strands restarted efficiently. Therefore, the heterogeneity of the LWS (a) and RWS (Fig. 1e) is probably due to delayed ligation of a new Okazaki fragment to the lagging strands. Quantification of leading strands that had not reached the midpoint of the array (rightward and leftward strands smaller than 550 and 500 nt, respectively, b) revealed that by 6.25 min, 90% of rightward and leftward leading strands passed the midpoint of the array (d, e). This demonstrates that leading strands pass each other when forks meet. KB, kilobase ladder, with the length of each band (in kilobases) labelled.
a, The autoradiograph in primary Fig. 1g is reproduced with cartoons indicating the structures of the replication and termination intermediates n–n, n–sc, sc–sc, n and sc (see Fig. 1 for definitions). The order of appearance of the different catenanes matches previous work5 (n-n, then n-sc, then sc-sc). b–d, To determine the role of Topo II during termination within a lacO array, termination was monitored in mock- or Topo-II-depleted extracts. To confirm immunodepletion of Topo II, mock and Topo-II-depleted NPE was blotted with MCM7 and Topo II antibodies (b). p[lacO16] was incubated with LacR, then replicated in either mock- or Topo-II-depleted egg extracts in the presence of [α-32P]dATP, and termination was induced with IPTG (at 7 min). Untreated DNA intermediates were separated by native gel electrophoresis (c). In the mock-depleted extract, nicked and supercoiled monomers were readily produced (as in panel a, albeit with slower kinetics due to nonspecific inhibition of the extracts by the immunodepletion procedure), while in the Topo-II-depleted extracts, a discrete species was produced. DNA from the last time point in each reaction (lanes 4 and 8 in panel c) was purified and treated with XmnI, which cuts p[lacO16] once, or Nt.BspQI, which nicks p[lacOx16] once, or recombinant Topo II, and then separated by native gel electrophoresis (d). Cleavage of the mock- and Topo-II-depleted products with XmnI yielded the expected linear 3.15-kb band (lanes 2 and 6), demonstrating that in both extracts all products were fully dissolved topoisomers of each other. Relaxation of the mock-depleted products by nicking with Nt.BspQI yielded a discrete band corresponding to nicked plasmid (lane 3), while the Topo-II-depleted products were converted to a ladder of discrete topoisomers (lane 7), which we infer represent catenated dimers of different linking numbers, since the mobility difference cannot be due to differences in supercoiling. Importantly, the mobility shift after Nt.BspQI treatment (lane 5 versus lane 7) demonstrated that the Topo-II-depleted products (lane 5) were covalently closed and thus in the absence of Topo II, ligation of the daughter strands still occurred. Treatment of the mock- and Topo-II-depleted products with recombinant human Topo II produced the same relaxed monomeric species (lanes 4 and 8), further confirming that the Topo-II-depleted products contained catenanes. Collectively, these observations demonstrate that termination within a lacO array in Topo-II-depleted extracts produces highly catenated supercoiled–supercoiled dimers, as seen in cells lacking Topo II16,17. These data confirm that Topo II is responsible for decatenation and argue that termination within a lacO array reflects physiological termination. e, n–n, n–sc, sc–sc, n and sc products were also detected when plasmid lacking lacO sequences (pBlueScript) was replicated in the absence of LacR without the use of cyclin A to synchronize replication. Therefore, these intermediates arise in the course of unperturbed DNA replication in Xenopus egg extracts.
a, Cartoon depicting intermediates detected in the dissolution assay. b, To determine the ability of different-sized LacR arrays to inhibit termination, the earliest stage of termination, dissolution (a), was monitored in plasmids containing 0, 12, 16, 32, or 48 lacO repeats. Plasmids were incubated with LacR, and replicated in the presence of [α-32P]dATP. To measure dissolution, radiolabelled termination intermediates were cut with XmnI. Cleaved products were separated on a native agarose gel and detected by autoradiography. c, Quantification of dissolution in b. When 12 or more lacO repeats were present in the array, dissolution was robustly inhibited for at least 5 min. Potent inhibition lasted 10 min when 32 lacO sequences were present, and 20 min in the presence of 48 lacO sequences. In the absence of lacO sequences, dissolution was essentially complete by 5 min. Therefore, 12 lacO repeats are sufficient to inhibit termination for 5 min.
a–c, To test further whether replication stalls or slows before dissolution, p[lacO12] was pre-incubated with LacR and replicated in Xenopus egg extracts. Termination was then induced by addition of IPTG after 5 min. Simultaneously, [α-32P]dATP was added to specifically radiolabel DNA synthesized after IPTG addition (a). Radiolabelled DNA was then separated on a native agarose gel and total signal was measured by autoradiography (b). Total signal was quantified, normalized to peak signal, and graphed alongside the rate of dissolution, which was also measured in the same experiment (c). This approach gives a highly sensitive measure of DNA synthesis without manipulation of DNA samples. DNA synthesis should occur primarily within the lacO array (see Extended Data Fig. 1). Upon IPTG addition, there was an approximately linear increase in signal, which plateaued by 5.83 min. Importantly, dissolution was 65% complete by 5.83 min. Therefore, the large majority of dissolution occurs without stalling of DNA synthesis. d, e, Experimental repeats of b, c. f, The experiments shown in c–e were graphed together with mean ± s.d. Synthesis data were normalized so that for each experiment, synthesis at 1 min was assigned a value of 84.4%, since this was the average value from c, d, where synthesis was allowed to plateau. Given the rate of replication fork progression in these egg extracts (260 bp min−1 (ref. 32)) and the size of the array (365 bp), forks should require, on average, 0.7 min to converge if no stalling occurs ((365 bp/2)/260 bp min−1 = 0.7 min). The time required for dissolution was not appreciably longer than this (dissolution was 50% complete by 0.67 min after IPTG addition, f), consistent with a lack of stalling. g, h, The experiment shown in b, c was repeated using p[lacO16]. Synthesis was approximately linear until 6.17 min, at which point 81% of molecules had dissolved, further demonstrating that the majority of dissolution occurs without stalling of DNA synthesis.
a–d, To test whether replisomes meet later in a lacO32 array than a lacO12 array, we monitored dissolution. LacR block-IPTG release was performed on p[lacO12] and p[lacO32] and radiolabelled termination intermediates were digested with XmnI to monitor the conversion of double-Y molecules to linear molecules (dissolution). Cleaved molecules were separated on a native agarose gel, detected by autoradiography (a, c), and quantified (b, d). Upon IPTG addition, dissolution was delayed by at least 1 min within the 32 lacO array compared to the 12 lacO array (b, d). Moreover, by 6 min, 92% of forks had undergone dissolution on p[lacO12] while only 9% had dissolved on p[lacO32] (b, d). e, Stall products within the 12 lacO array (Fig. 3b, lane 2) were quantified, signal was corrected based on size differences of the products, and the percentage of stall products at each stall point was calculated. 78% of leading strands stalled at the first three arrest points (red columns), 19% stalled at the fourth to tenth arrest points (yellow columns) and the remaining 3% stalled at the tenth to fourteenth arrest points (grey columns). The appearance of fourteen arrest points is reproducible but surprising, given that the presence of only 12 lacO sequences was confirmed by sequencing in the very preparation of p[lacO12] that was used in Fig. 3. The thirteenth and fourteenth arrest points cannot stem from cryptic lacO sites beyond the twelfth lacO site, as this would position the first leftward leading strand stall product ∼90 nucleotides from the lacO array, instead of the observed ∼30 nucleotides (see f, g). At present, we do not understand the origin of these stall products. f, g, Progression of leftward leading strands into the array. The same DNA samples used in Fig. 3 were digested with the nicking enzyme Nb.BsrDI, which released leftward leading strands (f), and separated on a denaturing polyacrylamide gel (g). The lacO sites of p[lacOx12] are highlighted in blue on the sequencing ladder (g), which was generated using the primer JDO109 (green arrow, f). Green circles indicate two nonspecific products of digestion. These products arise because nicking enzyme activity varies between experiments, even under the same conditions. There was no significant difference in the pattern of leftward leading strand progression between the 12 lacO and 32 lacO arrays, as seen for the rightward leading strands (Fig. 3b). Specifically, by 5.67 min, the majority of leading strands had extended beyond the seventh lacO repeat within lacO12 (lane 6) and the equivalent region of lacO32 (lane 18). Therefore, progression of leftward leading strands is unaffected by the presence of an opposing replisome, suggesting that converging replisomes do not stall when they meet.
a, Cartoon depicting the LAC, FLK2 and FAR loci, which were used for ChIP. Their precise locations relative to the leftward edge of the lacO array are indicated. The LAC amplicon is present in four copies distributed across the lacO16 array and three copies distributed across the lacO12 array. b–e, p[lacOx12] was incubated with buffer or LacR and termination was induced at 5 min by IPTG addition. MCM7, RPA, CDC45 and Polε ChIP was performed at different time points after IPTG addition but also in the buffer control and no IPTG control. Recovery of FLK2 was measured as a percentage of input DNA. Upon IPTG addition, ChIP signal declined and by 9 min was comparable to the buffer control, demonstrating that unloading of replisomes was induced within 4 min of IPTG addition. f, To test whether movement of the replisome into and out of the lacO array could be detected upon IPTG addition, termination was monitored within a lacO array, and we performed ChIP of the leading strand polymerase Polε, which was inferred to move into and out of the array based on the behaviour of leading strands during termination (Extended Data Fig. 2b–e). It was predicted that Polε ChIP at the LAC locus should increase slightly as Polε enters the lacO array and decline again as converging polymerases pass each other, but persist at FLK2 while the polymerases move out of the array. Before IPTG addition, Polε was enriched at LAC and FLK2 compared to FAR, consistent with the leading strands being positioned on either side of the lacO array (Extended Data Fig. 2c and Fig. 3). Upon IPTG addition, Polε became modestly enriched at LAC compared to FLK2 (5.5 min) but then declined to similar levels at both LAC and FLK2 by 6.5 min. These data are consistent with the leading strand polymerases entering the lacO array and passing each other. g, h, To test whether CMG exhibited the same ChIP profile as Polε, MCM7 and CDC45 ChIP was performed using the same samples. After IPTG addition, MCM7 and CDC45 were enriched at LAC compared to FLK2 (5.5 min), then declined to similar levels at both LAC and FLK2 by 6.5 min, as seen for Polε (f). These data are consistent with a model in which CMGs enter the array and pass each other during termination. A caveat of these experiments is the relatively high recovery of the FAR locus in MCM7, CDC45 and Polε ChIP. Specifically, signal was at most only ∼2-fold enriched at LAC compared to FAR. This was not due to high background binding, because by the end of the experiment (10 min time point, not shown) we observed a decrease in signal of ∼5–7-fold. Furthermore, we observed ∼5–7-fold enrichment in binding (ChIP) of replisome components to p[lacO12] that had been incubated in LacR compared to a buffer control (see g–i, below). Instead, the high FAR signal was probably due to poor spatial resolution of the ChIP. Consistent with this, when a plasmid containing a DNA interstrand cross-link (ICL) was replicated, essentially all replisomes converged upon the ICL but the ChIP signal for MCM7 and CDC45 was only ∼3–4-fold enriched at the ICL compared to a control locus41. We speculate that the higher background observed at the control locus in our experiments is due to the decreased distance of the control locus from the experimental locus (1.3 kb for p[lacO16] and p[lacO12] versus 2.4 kb for the ICL plasmid) and possibly due to increased catenation of the parental strands during termination. The high signal at FAR should not complicate interpretation of the MCM7, CDC45 and Polε ChIP (f), as signal at FAR was essentially unaltered between 5 and 6.5 min. Further evidence that the high signal seen at the FAR locus emanates from forks stalled near the lacO array is presented in panel k. i, ChIP of RPA was performed on the same chromatin samples used in b–d. As seen for Polε, MCM7 and CDC45, enrichment of RPA at LAC compared to FAR was relatively low, consistent with poor spatial resolution. j, Predicted binding of CMGs to the LAC, FLK2 and FAR loci before and after IPTG addition if converging CMG pass each other. k, To determine whether most forks stalled at the array and not elsewhere in the plasmid, we performed a time course in which p[lacO16] undergoing termination was examined by 2D gel electrophoresis at various time points. p[lacO16] was pre-bound to LacR and replicated in Xenopus egg extract containing [α-32P]dATP. Termination was induced by IPTG addition and samples were withdrawn at different times. Radiolabelled replication intermediates were cleaved with XmnI (as in Extended Data Fig. 1a) and separated according to size and shape on 2D gels50. A parallel reaction was performed in which samples were analysed by ChIP, which was one of the repeats analysed in b–e. In the presence of LacR, a subset of double-Y molecules accumulated (blue arrowhead), demonstrating that 83% of replication intermediates (signal in dashed blue circle) contained two forks converged at a specific locus. After IPTG addition, linear molecules rapidly accumulated (orange arrowhead) as dissolution occurred. Importantly, the vast majority of signal was present in the discrete double-Y and linear species (blue and orange arrowheads), demonstrating that the relatively high ChIP signal observed at FAR in panels f–i was derived from forks present at the lacO16 array and not elsewhere.
a, Cartoon depicting the XmnI and AlwNI sites on p[empty], which are used for the dissolution and ligation assays, respectively, and the FLK2 locus, which is used for ChIP. b, Plasmid DNA without a lacO array (p[empty]) was replicated and at different times chromatin was subjected to MCM7 and CDC45 ChIP. Per cent recovery of FLK2 was quantified and used to measure dissociation of MCM7 and CDC45 (see Methods). Dissolution and ligation were also quantified in parallel. Mean ± s.d. is plotted (n = 3). The MCM7 and CDC45 dissociation data are obtained from the vehicle controls in Fig. 5b, c, while the dissolution and ligation data are obtained from the vehicle controls in Fig. 5d, e. c, To seek independent evidence for the conclusions of the ChIP data presented in Fig. 5b, c, we used a plasmid pull-down procedure. p[empty] was replicated in egg extracts treated with vehicle or Ub-VS. At the indicated times, chromatin-associated proteins were captured on LacR-coated beads (which binds DNA independently of lacO sites) and analysed by western blotting for CDC45, MCM7 and PCNA. CDC45 and MCM7 dissociated from chromatin by 8 min in the vehicle control, but persisted following Ub-VS treatment. d, To test whether the MCM7 modifications detected in panel c represented ubiquitylation, extracts were incubated with His6-ubiquitin in the absence of cyclin A, and in the absence or presence of plasmid DNA. After 15 min, His6-tagged proteins were captured by nickel resin pull down and blotted for MCM7. DNA replication greatly increased the levels of ubiquitylated MCM7, with the exception of a single species that was ubiquitylated independently of DNA replication (*). These data show that MCM7 is ubiquitylated during plasmid replication in egg extracts, as observed in yeast and during replication of sperm chromatin after nuclear assembly in egg extracts24,25. e, In parallel to the plasmid pull downs performed in c, DNA samples were withdrawn for dissolution, ligation and decatenation assays, none of which was perturbed by Ub-VS treatment. These data support our conclusion, based on ChIP experiments (Fig. 5), that defective CMG unloading does not affect dissolution, ligation, or decatenation. f, Decatenation was measured in the same reactions used to measure dissolution and ligation (Fig. 5d, e), mean ± s.d. is plotted (n = 3). g–i, Given the experimental variability at the 4 min time point in Fig. 5d–f, the primary data and quantification for dissolution (g), ligation (h) and decatenation (i) for one of the three experiments summarized in Fig. 5d–f is presented. This reveals that Ub-VS does not inhibit dissolution, ligation, or decatenation at the 4 min time point. The same conclusion applies to two additional repetitions of this experiment (data not shown). j, The primary ChIP data used to measure dissociation of MCM7 and CDC45 in Fig. 5b, c is shown. Recovery of FLK2 was measured. Mean ± s.d. is plotted (n = 3).
a, b, To synchronize DNA replication in Xenopus egg extracts, we treated extracts with cyclin A, which probably accelerates replication initiation45. Plasmid DNA was incubated in High Speed Supernatant for 20 min, then either buffer or cyclin A was added for a further 20 min. NucleoPlasmic extract was added to initiate DNA replication, along with [α-32P]dATP to label replication intermediates. Replication products were separated on a native agarose gel, detected by autoradiography (a), and quantified (b). In the presence of vehicle, replication was not complete by 9.5 min, but in the presence of cyclin A, replication was almost complete by 4.5 min (b). Thus, cyclin A treatment approximately doubles the speed of DNA replication in Xenopus egg extracts. c–f, To test whether cyclin A affects the ability of LacR to inhibit termination, we monitored dissolution of plasmids containing a 12 or 16 LacR array in the presence and absence of cyclin A. p[lacO12], p[lacO16], and the parental control plasmid p[empty] were incubated with LacR, and then treated with buffer or cyclin A before replication was initiated with NPE in the presence of [α-32P]dATP. Samples were withdrawn when dissolution of p[empty] plateaued (9.5 min in the presence of buffer, 4.5 min in the presence of cyclin A). Given that cyclin A treatment approximately doubles the speed of replication (see b), samples were withdrawn from these reactions twice as frequently as the buffer-treated samples. To measure dissolution, radiolabelled termination intermediates were cut with XmnI to monitor the conversion of double-Y molecules to linear molecules. Cut molecules were separated on a native agarose gel and detected by autoradiography (c, e). By the time the first sample was withdrawn, dissolution of p[empty] was essentially complete, in the absence (9.5 min, d) or presence (4.5 min, f) of cyclin A. Importantly, dissolution of p[lacO12] and p[lacO16] was prevented in the absence (9.5 min, d) or presence (4.5 min, f) of cyclin A. Moreover, dissolution occurred approximately twice as fast in the presence of cyclin A (note the similarity between d and f even though samples are withdrawn twice as frequently in f) consistent with replication being approximately twice as fast in the presence of cyclin A. Therefore, cyclin A does not affect the ability of a LacR array to block replication forks.
About this article
Cite this article
Dewar, J., Budzowska, M. & Walter, J. The mechanism of DNA replication termination in vertebrates. Nature 525, 345–350 (2015). https://doi.org/10.1038/nature14887
Journal of Biological Chemistry (2020)
Characterization of the dimeric CMG/pre-initiation complex and its transition into DNA replication forks
Cellular and Molecular Life Sciences (2020)
Biochemical Society Transactions (2020)
Frontiers in Microbiology (2020)
EMBO reports (2020)