DOI: 10.1126/science.1241779, 239 (2013);342 Science

et al.Jean-Baptiste Vannierand Genome-Wide ReplicationRTEL1 Is a Replisome-Associated Helicase That Promotes Telomere

This copy is for your personal, non-commercial use only.

clicking here.colleagues, clients, or customers by , you can order high-quality copies for yourIf you wish to distribute this article to others

  here.following the guidelines

can be obtained byPermission to republish or repurpose articles or portions of articles

  ): February 10, 2014 www.sciencemag.org (this information is current as of

The following resources related to this article are available online at

http://www.sciencemag.org/content/342/6155/239.full.htmlversion of this article at:

including high-resolution figures, can be found in the onlineUpdated information and services,

http://www.sciencemag.org/content/suppl/2013/10/09/342.6155.239.DC1.html can be found at: Supporting Online Material

http://www.sciencemag.org/content/342/6155/239.full.html#ref-list-1, 9 of which can be accessed free:cites 26 articlesThis article

http://www.sciencemag.org/cgi/collection/molec_biolMolecular Biology

subject collections:This article appears in the following

registered trademark of AAAS. is aScience2013 by the American Association for the Advancement of Science; all rights reserved. The title

CopyrightAmerican Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by theScience

on

Feb

ruar

y 10

, 201

4w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

o

n F

ebru

ary

10, 2

014

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

on

Feb

ruar

y 10

, 201

4w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

o

n F

ebru

ary

10, 2

014

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

on

Feb

ruar

y 10

, 201

4w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

RTEL1 Is a Replisome-AssociatedHelicase That Promotes Telomere andGenome-Wide ReplicationJean-Baptiste Vannier,1* Sumit Sandhu,2* Mark IR. Petalcorin,1 Xiaoli Wu,2 Zinnatun Nabi,2

Hao Ding,2,3† Simon J. Boulton1†

Regulator of telomere length 1 (RTEL1) is an essential DNA helicase that disassembles telomereloops (T loops) and suppresses telomere fragility to maintain the integrity of chromosomeends. We established that RTEL1 also associates with the replisome through binding toproliferating cell nuclear antigen (PCNA). Mouse cells disrupted for the RTEL1-PCNA interaction(PIP mutant) exhibited accelerated senescence, replication fork instability, reduced replicationfork extension rates, and increased origin usage. Although T-loop disassembly at telomeres wasunaffected in the mutant cells, telomere replication was compromised, leading to fragile sitesat telomeres. RTEL1-PIP mutant mice were viable, but loss of the RTEL1-PCNA interactionaccelerated the onset of tumorigenesis in p53-deficient mice. We propose that RTEL1 plays acritical role in both telomere and genome-wide replication, which is crucial for genetic stability andtumor avoidance.

The stability of the genome is critically de-pendent on the coordinate action of DNArepair pathways during the cell cycle (1).

The DNA helicase regulator of telomere length1 (RTEL1) is an anti-recombinase that dismantlesD-loop recombination intermediates to counter

toxic DNA repair (2). RTEL1 also functions inmeiosis to limit excessive crossing over (3) anddisassembles T loops and suppresses telomerefragility to maintain integrity of the chromosomeend (4). The mechanism(s) by which RTEL1 exe-cutes its function at sites of DNA repair and attelomeres remains unclear. BecauseRTEL1−/−miceare embryonic lethal (5) and die mid-gestation, itis also possible that RTEL1 possesses other es-sential functions that remain to be defined.

To further investigate RTEL1 function in cells,we performed mass spectrometry to identify in-teracting proteins from stable cells expressing aC-terminal green fluorescent protein (GFP)–tagged

RTEL1. In addition to methyl methanesulfonatesensitive 19 (MMS19) that functions in Fe-S clusterassembly (6), we detected numerous componentsof the replisome, including replication factor C(RFC), DNA polymerases, minichromosome main-tenance proteins (MCMs), and proliferating cellnuclear antigen (PCNA) (fig. S1A). We furtherinvestigated a putative RTEL1-PCNA interactionas bioinformatic analysis revealed the presenceof a putative PCNA interaction motif (PIP box)located in the C terminus of the RTEL1 protein(fig. S1B). RTEL1-v5 and endogenous PCNAcoimmunoprecipitated from cell extracts, andthis interaction was resistant to benzonase treat-ment, excluding a nonspecific association via DNAbridging (Fig. 1A). By contrast, RTEL1-v5 proteinsharboring five PIP box substitution mutationsfailed to immunoprecipitate PCNA (fig. S1C).The RTEL1-PCNA interaction is likely directas in vitro pull-down experiments with a gluta-thione S-transferase (GST) fusion to a fragmentof RTEL1 (amino acids 1125 to 1225) contain-ing the wild-type PIP box efficiently bound torecombinant PCNA (rPCNA) as well as endog-enous PCNA from human embryonic kidney–293(HEK293) or HeLa cell extracts, whereas a GST-RTEL1 fusion harboring an I1169A PIP box mu-tant (referred to as IA) failed to interact (Fig. 1B).Arrayed biotinylated peptides comprising thewild-type RTEL1 PIP box sequence bound torPCNA, whereas the RTEL1 PIP boxmutant pep-tides did not (fig. S1D). These results demonstratethat the PIP box in RTEL1 confers a direct inter-action with PCNA in vitro and in cells.

PCNA is a processivity factor for DNA poly-merase and an integral component of the replisomeduring S phase. PCNA is present in multiple

1DNA Damage Response laboratory, London Research Institute,Cancer Research UK, Clare Hall, South Mimms EN6 3LD, UK.2Department of Biochemistry and Medical Genetics, University ofManitoba, Winnipeg, Manitoba R3E 3J7, Canada. 3ManitobaInstitute of Child Health, Winnipeg, Manitoba, R3E 3P4, Canada.

*These authors contributed equally to this work.†Corresponding author. E-mail: dingh@cc.umanitoba.ca(H.D.); simon.boulton@cancer.org.uk (S.J.B.)

Fig. 1. RTEL1-PCNA interaction is dependenton RTEL1 PIP motif. (A) Immunoprecipitation ofRTEL1-v5 in ES cells stably overexpressing RTEL1-v5. IgG, immunoglobulin G; WB, Western blot. (B)GST pull-down experiments between wild-type (WT)RTEL1 or IA RTEL1 mutant and recombinant PCNA(rPCNA) extracts from HEK293 or from HeLa cells.(*) Truncated protein. (C) Detection of RTEL1-v5(red) and PCNA (green) in ES cells overexpressingWT or IA RTEL1-v5 and PCNA-flag. DAPI, 4´,6-diamidino-2-phenylindole. (D) Immunoprecip-itation (IP) of RTEL1-v5 and RTEL1-IA-v5 from EScells derived from RTEL1+/+v5 and RTEL1IA/IAv5 mice.

www.sciencemag.org SCIENCE VOL 342 11 OCTOBER 2013 239

REPORTS

discrete replication foci in S-phase cells (Fig. 1C)(7). Wild-type RTEL1-v5 also formed discretefoci, a subset of which colocalized with PCNA-Flag in unperturbed cells (Fig. 1C). The RTEL1IA PIP box mutant is expressed in the nucleus,but failed to colocalize with PCNA (Fig. 1C).Thus, the presence of RTEL1 within replicationfoci is dependent on a PIP box–mediated inter-action with PCNA. This result also suggests thatany additional contacts between RTEL1 and otherreplisome components are likely to be weak orindirect, as they are not sufficient to maintainRTEL1 at replication foci when the RTEL1-PCNAinteraction is abolished.

To investigate the functional importance ofthe RTEL1-PCNA interaction in vivo, RTEL1IA PIP mutation marked with a v5 epitope wasknocked into the RTEL1 locus in mice (fig. S2).The resultant mouse allele, termed RTEL1IA/IAv5,abolished RTEL1-PCNA binding as comparedto the control mouse allele (RTEL1+/+v5), whichretains an intact RTEL1 PIP motif (Fig. 1D).Loss of the RTEL1-PCNA interaction inducedgrowth arrest and senescence in primary mouseembryonic fibroblasts (MEFs) derived from em-

bryonic day 14.5 (E14.5) RTEL1IA/IAv5 embryoswhen compared to RTEL1+/+v5 control MEFs(Fig. 2A and fig. S3A), which we attribute toincreased levels of spontaneous DNA damagein the mutant cells (fig. S5, E and F). Cell cycleanalysis revealed a significant reduction in 5-bromo-2´-deoxyuridine (BrdU) incorporation andan accumulation of cells in late S/G2 in primaryRTEL1IA/IAv5 MEFs when compared to control,suggestive of a defect in DNA replication (Fig. 2Band fig. S3B). Indeed, measurement of the pro-gression of sister replication forks by molecularcombing revealed that loss of the RTEL1-PCNAinteraction caused forks to move considerablymore slowly in a genome-wide manner comparedto control cells (0.87 and 1.8 kb min−1, respective-ly; Fig. 2C). Of the replication forks, 52.9% werehighly asymmetrical in RTEL1IA/IAv5 cells, com-pared to 7.8% asymmetric tracts in RTEL1+/+v5control cells (Fig. 2, D and E), which suggeststhat loss of the RTEL1-PCNA interaction in cellsresults in increased replication fork stalling and/orcollapse. Interorigin distances were also signif-icantly shorter in the RTEL1IA/IAv5 cells whencompared to control cells (87 T 28 and 122 T 28 kb,

respectively; Fig. 2F), indicative of increased ori-gin firing inmutant cells. Similar replication defectswere also observed after conditional inactivationof RTEL1 following adenovirus-Cre treatmentof RTEL1F/F MEFs (fig. S3, C to F). Blockingreplication origin firing in the RTEL1IA/IAv5 cellswith the CDC7 inhibitor PHA-767491 (fig. S4A)(8, 9) rescued both interorigin distance and repli-cation fork extension rates to wild type but failedto rescue fork asymmetry (fig. S4, B to D). Thissuggests that increased origin usage and the re-duction in fork speeds in RTEL1IA/IAv5 cells occuras a secondary consequence of a defect in prevent-ing fork stalling and/or collapse. We propose thatin the absence of the RTEL1-PCNA interaction,increased fork stalling and/or collapse or a failureto repair and/or restart a subset of replication forkstriggers dormant origin firing, which in turn leadsto a global reduction in replication fork exten-sion rates.

RTEL1 is proposed to maintain telomere in-tegrity in part by catalyzing T-loop disassemblyduring S phase (4). Failure to dismantle T loopsafter inactivation of RTEL1 is associated withrapid telomere shortening and loss of the T loop

Fig. 2. RTEL1-PCNAinteractionpromotesnormalgenome replication. (A) Quantification of growth cellarrest after consecutive passages (P1 to P9) of eachgenotype. (B) BrdU incorporation of RTEL1IA/IAv5and RTEL1+/+v5 primary MEFs. (C) Replication forkdynamics in RTEL1IA/IA and control MEFs pulse-labeledwith iododeoxyuridine (IdU) or chlorodeoxyuridine(CldU) and subjected to DNA combing. One hundredfibers were measured per genotype, and replicationfork speed was measured in kb min−1. (D) Represent-ative images of sister replication forks in RTEL1+/+v5and RTEL1IA/IAv5 primary MEFs. Arrows mark un-stable and/or stalled forks. (E) Quantification offork asymmetry in RTEL1IA/IAv5 and control primaryMEFs. (F) Representative images and quantifica-tion of inter-origin distances (kb) in RTEL1IA/IAv5and RTEL1+/+v5. Fifty fibers were measured pergenotype. ****P < 0.0001 (two-tailed Mann-Whitneytest). Error bars show the SD.

11 OCTOBER 2013 VOL 342 SCIENCE www.sciencemag.org240

REPORTS

as a circle (4). Phi29 polymerase–dependent telo-mere circles accumulate in cells after conditionalinactivation ofRTEL1 inRTEL1F/FMEFs (Fig. 3A)(4). By contrast, telomere circles (TCs) were un-detectable in either RTEL1+/+v5- or RTEL1IA/IAv5-complemented embryonic stem (ES) cells (Fig. 3A).We also found no evidence of telomere loss inmeta-phase spreads ofRTEL1+/+v5 orRTEL1IA/IAv5 cells(Fig. 3, B and C). RTEL1+/+v5 or RTEL1IA/IAv5cells also exhibited a wild-type telomere lengthdistribution (fig. S5A) and did not show any evi-dence of DNA damage at telomeres (fig. S5B),increased sister chromatid exchanges (SCEs), orincreased telomere SCEs (fig. S5, C and D). Be-causeRTEL1IA/IAv5 cells do not exhibit a significantinduction of DNAdamage at telomeres (fig. S5B),the elevated spontaneous levels of the histonegH2AX in these cells (fig. S5, E and F) are notcaused by dysfunctional telomeres. These resultsestablish that the RTEL1-PCNA interaction isdispensable for T-loop disassembly and prevent-ing loss of the telomere as a circle.

RTEL1 is also responsible for the suppressionof telomere fragility (4, 10). Loss of RTEL1 inES cells resulted in enhanced telomere fragility(24 T 10 fragile telomeres per metaphase, com-pared to 8.8 T 3.6 and 9.9 T 4.2 in wild-type EScells and RTEL1−/− ES cells complemented withRTEL1+v5; Fig. 3B and table S1). RTEL1−/− EScells complemented with RTEL1IAv5 PIP boxmutant and RTEL1IA/IAv5 MEFs exhibited high

levels of telomere fragility (20 T 7.9 and 10 T 4.9,respectively; Fig. 3, B and C, and table S1), com-parable to those detected in RTEL1-null ES cellsor RTEL1−/− MEFs [24 T 10 and (4), respec-tively]. After conditional inactivation of RTEL1in RTEL1F/F MEFs, telomere fragility is exacer-bated by treatment with aphidicolin or the G4-DNA stabilizer TMPyP4 (4). Similarly, treatmentof RTEL1IA/IAv5MEFs with aphidicolin or TMPyP4resulted in a significant increase in telomere fra-gility, corresponding to 15 T 5.5 and 14 T 6.2 fragiletelomeres per metaphase compared to 10 T 4.9fragile telomeres per metaphase without treatment(Fig. 3C and table S1). Collectively, these resultsestablish that the RTEL1-PCNA interaction isessential for suppressing telomere fragility but isdispensable for T-loop disassembly (fig. S5G).

RTEL1-deficient cells are sensitive to G4-DNAstabilizer TMPyP4 (11) and telomere fragility isexacerbated by telomeric G4-DNA stabilization(4), suggesting a strong correlation between frag-ile telomeres andG4-DNA secondary structures. Toaddress whether RTEL1 can disassemble telomericG4-DNA, recombinant RTEL1wild-type and aden-osine triphosphatase (ATPase) dead (RTEL1K48R)mutant proteins were purified and tested for ac-tivity toward telomericG4-DNA structures. RTEL1wild-typeefficientlydisassembled telomericG4-DNAstructures in anATPase-dependentmanner in vitro(fig. S6A). Addition of recombinant PCNA to thereaction neither stimulated nor inhibited the re-

action. However, addition of TMPyP4 to the re-action inhibited the ability of RTEL1 wild type todisassemble the substrate (fig. S6B), suggestingthat TMPyP4 exacerbates telomere fragility in vivoby blocking the ability of RTEL1 to dismantletelomeric G4-DNA structures.

RTEL1 was recently implicated in Hoyeraal-Hreidarsson syndrome, a severe form of thecancer predisposition and bone marrow fail-ure syndrome dyskeratosis congenita (12–14).A genome-wide association study also identi-fied RTEL1 as a susceptibility locus for glioma(15, 16), raising the possibility that RTEL1 mayfunction as a tumor suppressor. Despite thegenome-wide replication problems in primarycells, RTEL1IA/IAv5 mutant mice are born at ex-pected Mendelian ratios and exhibit normalbody weight and morphology at adult stage. Theviability of these mice is potentially explainedby increased origin usage that likely compen-sates for the replication defects in cells. To ex-amine the role of the RTEL1-PCNA interactionduring tumor development, we established acohort of compound Trp53−/−RTEL1+/IAv5 andTrp53−/−RTEL1IA/IAv5 mutant mice and monitoredtheir survival. Homozygous Trp53−/−RTEL1IA/IAv5mice exhibited substantially shorter life spanscompared with control Trp53−/−RTEL1+/IAv5mice (P < 0.0002; Fig. 4A). Tumor formationoccurred in the Trp53−/−RTEL1IA/IAv5mice within115 T 31 days compared with 147 T 35 days for

Fig. 3. RTEL1-PCNA interac-tion is required to suppresstelomere fragility but is dis-pensable to T-loop disassem-bly. (A) Phi29-dependent TCsamplification and quantificationin RTEL1+/+, RTEL1−/−, and EScells complemented with RTEL1-v5 (+/+) and RTEL1-IA-v5 (IA/IA).(B) Quantification of telomereloss and fragile telomeres permetaphase in RTEL1+/+, RTEL1−/−,andEScells complemented(Comp.)with RTEL1-v5 (Wt) and RTEL1-IA-v5 (IA/IA). (C) Representativeimages and quantification offragile telomeres (arrow) andtelomere loss (*) per metaphasein RTEL1+/+v5 and RTEL1IA/IAv5primary MEFs subjected to theindicated treatments (–, no treat-ment; APD, aphidicolin; statis-tical analysis with two-tailedMann-Whitney test). Error barsshow the SD.

www.sciencemag.org SCIENCE VOL 342 11 OCTOBER 2013 241

REPORTS

Trp53−/−RTEL1+/IAv5mice.Histological analysisshowed that both cohorts of Trp53−/−RTEL1+/IAv5and Trp53−/−RTEL1IA/IAv5 mutant mice devel-oped predominantly lymphomas (61.5 and 70.6%,respectively; Fig. 4B), although sarcomas andteratomas were also observed (15.4 versus 14.7%and 10.3 versus 14.7%, respectively; Fig. 4B), as de-scribed previously in Trp53−/− mice (17). Telomereanalysis of the tumors from Trp53−/−RTEL1IA/IAv5mice revealed extensive sister chromatid fusions,end-to-end fusions with telomeric repeats, and ahigh frequency of fragile telomeres when com-pared to nontumor cells from the samemice (Fig. 4,C and D, and fig. S7I). Furthermore, tumors fromTrp53−/−RTEL1+/IAv5mice orTrp53−/−RTEL1+/+

mice exhibited no evidence of telomere fragility ordysfunction (Fig. 4, C andD).Trp53−/−RTEL1IA/IAv5mutant mice also developed medulloblastomas(12.8%; Fig. 4B and fig. S7, A to F), which werenot observed in the Trp53−/−RTEL1+/IAv5 miceor in Trp53−/−mice described previously (17). Be-cause RTEL1 is expressed in the cerebellar ex-ternal granular layer (EGL) cells (fig. S7, G andH), which are the source of medulloblastomas(18), these data suggest that the RTEL1-PCNAinteraction is important for protecting theseEGL cells from transformation.

Our data reveal a critical interaction betweenRTEL1 and PCNA that is essential for telomereand genome-wide replication and suppression oftumorigenesis in vivo. Suppression of telomerefragility, which we attribute to the telomeric G4-

DNA unwinding activity of RTEL1, is abso-lutely dependent on the RTEL1-PCNA interac-tion. This suggests that replisome associationis required for RTEL1 to counteract telomericG4-DNA structures that arise at the replicationfork. By contrast, the role for RTEL1 in T-loopdisassembly and the suppression of telomere lossas a circle does not require the RTEL1-PCNAinteraction. The RTEL1-PCNA interaction is alsonecessary to prevent replication fork stallingand/or collapse, which affects genome-wide rep-lication. Finally, the accelerated rate of tumor for-mation and predisposition to medulloblastomasconferred by the RTEL1IA/IAv5 mutation in theTrp53−/− background implicates RTEL1 as a tumor-suppressor gene.

References and Notes1. J. R. Chapman, M. R. Taylor, S. J. Boulton, Mol. Cell 47,

497–510 (2012).2. L. J. Barber et al., Cell 135, 261–271 (2008).3. J. L. Youds et al., Science 327, 1254–1258 (2010).4. J. B. Vannier, V. Pavicic-Kaltenbrunner,

M. I. Petalcorin, H. Ding, S. J. Boulton, Cell149, 795–806 (2012).

5. H. Ding et al., Cell 117, 873–886 (2004).6. K. Gari et al., Science 337, 243–245 (2012).7. J. Essers et al., Mol. Cell. Biol. 25, 9350–9359

(2005).8. A. Montagnoli et al., Nat. Chem. Biol. 4, 357–365

(2008).9. L. C. Chuang et al., Mol. Cell 35, 206–216 (2009).10. A. Sfeir et al., Cell 138, 90–103 (2009).11. E. J. Uringa et al., Mol. Biol. Cell 23, 2782–2792

(2012).

12. B. J. Ballew et al., Hum. Genet. 132, 473–480(2013).

13. A. J. Walne, T. Vulliamy, M. Kirwan, V. Plagnol,I. Dokal, Am. J. Hum. Genet. 92, 448–453(2013).

14. T. Le Guen et al., Hum. Mol. Genet. 22, 3239–3249(2013).

15. K. M. Egan et al., J. Neurooncol. 104, 535–542(2011).

16. S. Shete et al., Nat. Genet. 41, 899–904 (2009).17. T. Jacks et al., Curr. Biol. 4, 1–7 (1994).18. R. J. Gilbertson, D. W. Ellison, Annu. Rev. Pathol. 3,

341–365 (2008).

Acknowledgments: We thank the Protein Analysis andProteomics group at London Reasearch Institute formass spectrometry and J. Mendez for MCM2 antibody.Research in the DNA damage response lab of S.J.B. isfunded by Cancer Research UK and by a EuropeanResearch Council (ERC) Advanced Investigator Grant(RecMitMei). S.J.B. is a recipient of a Royal SocietyWolfson Research Merit Award. The laboratory of H.D.is supported by Canada Research Chair program,Canada Institute of Health Research, Manitoba Instituteof Child Health, and Terry Fox Research Institute.J.-B.V. is funded by a long-term fellowship from ERC,and S.S. and Z.N. are supported by the fellowships fromNatural Sciences and Engineering Research Council ofCanada (to S.S.) and the Manitoba Health ResearchCouncil (to Z.N.).

Supplementary Materialswww.sciencemag.org/content/342/6155/239/suppl/DC1Materials and MethodsFigs. S1 to S7Table S1References (19–26)

11 June 2013; accepted 11 September 201310.1126/science.1241779

Fig. 4. RTEL1IA/IA mutation acceleratestumorigenesis inmice. (A) Kaplan-Meier survivalcurves of Trp53−/−RTEL1IA/+v5 and Trp53−/−RTEL1IA/IAv5mice with the number of mice indicated for eachgenotype. (B) Tumor spectrum in RTEL1IA/+v5 andRTEL1IA/IAv5 mice in Trp53−/− background. (C) Fre-quency of fragile telomeres and chromosome fu-sions in Trp53−/−RTEL1IA/IAv5 tumorigenic andnontumorigenic tissues. P value was calculatedwith two-tailed Mann-Whitney test. (D) Represen-tative images of metaphase spreads derived fromTrp53−/−RTEL1IA/+v5 and Trp53−/−RTEL1IA/IAv5tumors subjected to telomere fluorescence in situhybridization (red, telomeres; blue, DNA). Whiteasterisk (*), sister chromatid fusions; yellow as-terisk, end-to-end fusions; arrows, fragile telomeres.

11 OCTOBER 2013 VOL 342 SCIENCE www.sciencemag.org242

REPORTS