Pot1b −/− tumors activate G-quadruplex-induced DNA damage to promote telomere hyper-elongation

Abstract Malignant cancers must activate telomere maintenance mechanisms to achieve replicative immortality. Mutations in the human Protection of Telomeres 1 (POT1) gene are frequently detected in cancers with abnormally long telomeres, suggesting that the loss of POT1 function disrupts the regulation of telomere length homeostasis to promote telomere elongation. However, our understanding of the mechanisms leading to elongated telomeres remains incomplete. The mouse genome encodes two POT1 proteins, POT1a and POT1b possessing separation of hPOT1 functions. We performed serial transplantation of Pot1b−/− sarcomas to better understand the role of POT1b in regulating telomere length maintenance. While early-generation Pot1b−/− sarcomas initially possessed shortened telomeres, late-generation Pot1b−/− cells display markedly hyper-elongated telomeres that were recognized as damaged DNA by the Replication Protein A (RPA) complex. The RPA-ATR-dependent DNA damage response at telomeres promotes telomerase recruitment to facilitate telomere hyper-elongation. POT1b, but not POT1a, was able to unfold G-quadruplex present in hyper-elongated telomeres to repress the DNA damage response. Our findings demonstrate that the repression of the RPA-ATR DDR is conserved between POT1b and human POT1, suggesting that similar mechanisms may underly the phenotypes observed in human cancers harboring human POT1 mutations.


INTRODUCTION
Telomer es ar e DNA-protein complex es that cap the ends of eukaryotic linear chromosomes ( 1 , 2 ).Mammalian telomer es ar e composed of mostly duplex 5 -TTAGGG-3 DNA hexameric repeats ending in a 3 single-stranded overhang.Due to the end r eplication problem, telomer es shorten with each round of cell division, ultimately leading to activation of p53-dependent replicati v e senescence ( 3 , 4 ).In stem cells and most cancers, telomeres are elongated by the ribonucleoprotein complex telomerase, composed of the telomerase re v erse transcriptase (TERT) catalytic component and the RNA templating component ( Terc ) (5)(6)(7).The shelterin complex, comprised of six specialized proteins, mediates telomere end protection by r epr essing the activation of DNA damage response (DDR) pathways.TRF1 (Telomeric repeat-binding factor 1) and TRF2 (Telomeric repeat-binding factor 2) bind to the double-stranded telomeric DN A, w hile PO T1 (Protection of telomeres 1) binds to the single-stranded (ss) telomeric DNA and interacts with TPP1 (Adrenocortical dysplasia protein homolog).TIN2 (TRF1-interacting nuclear protein 2) bridges TPP1-POT1 with TRF1 and TRF2.POT1 forms a functional heterodimer with TPP1, and in turn, TPP1 tethers POT1 to telomeres by interacting with the TIN2-TRF1 and TIN2-TRF2 complexes ( 1 ).
PO T1 homolo gs have been identified in all eukaryotes.While most vertebrates , including humans , possess a single POT1 gene, the rodent genome encodes two Pot1 genes, Pot1a and Pot1b , due to recent gene duplication (8)(9)(10).All POT1 proteins contain two highly conserved oligosaccharide-oligonucleotide (OB) folds that interact with the 3 terminus of the telomere ss DNA overhang to ex ert telomer e end protection (9)(10)(11)(12).We and others discovered that POT1a and POT1b possess separate functions at telomeres.POT1a protects telomeres from being recognized as damaged DNA by binding to the telomere ss DNA, pre v enting Replication protein A (RPA) access and the initiation of an ATR-CHK1 mediated DDR signaling (11)(12)(13)(14)(15). POT1b binds telomere ss DNA as robustly as POT1a, but it is unclear why POT1b does not play a major role in telomere end protection.Rather, POT1b is involved in the formation of the 3 ss telomere overhang ( 16 ).POT1b also recruits the CTC1-STN1-TEN1 (CST) complex to promote DN A Pol ymerase-␣ fill-in synthesis of the C-strand and modulates exonuclease activities at the Cstrand (17)(18)(19)(20).Consequently, deletion of POT1b in mice increases G-overhang length while accelerating overall telomere shortening ( 21 , 22 ).We discovered recently that POT1b, but not POT1a, plays important roles in promoting telomerase recruitment to telomeres to extend the G-overhang ( 23 ).This result suggests that human POT1 (hPOT1) might also be involved in the regulation of telomere length.Indeed, hPOT1 and hTPP1 recruit the CST-complex to telomeres to regulate telomerase activity at the G-strand and promotes fill-in synthesis of the telomeric C-strand ( 16 , 19 , 24 ).
Next generation sequencing of a wide variety of cancers, including melanoma, chronic lymphocytic leukemia, angiosarcomas and gliomas, has identified r ecurr ent somatic m utations in hPO T1, suggesting that hPO T1 functions as an important tumor suppressor ( 25 , 26 ).Many hPOT1 cancer mutations cluster in its OB folds to disrupt binding to ss-telomeric DNA (27)(28)(29)(30)(31)(32).One of the most striking phenotypes observed in patients bearing hPOT1 cancer mutations is incr eased telomer e elongation, likely an important cancer promoting mechanism ( 28-31 , 33-37 ).Recent data on a population of hPO T1-m utant carriers re v eal that the long telomeres generated sustain clonal evolution to promote clonal hematopoiesis ( 38 ).Sustained clonal evolution allows for the accumulation of additional pro-oncogenic mutations to allow initiated pre-cancer cells to escape replicati v e senescence in favor of tumor progression.While telomere elongation may explain why hPOT1 mutations promote oncogenesis, our understanding of how hPOT1 dysfunction leads to the generation of these long telomeres remains incomplete.While inhibition of the DDR can r epr ess the telomere elongation observed in hPOT1 OB mutants, telomere elongation has also been observed in tumors bear-ing hPOT1 cancer mutations without overt DDR activation ( 34 , 39 ).
The two important functions of hPOT1, r epr ession of ATR-DDR activa tion and regula tion of telomere length, are split between mouse POT1a and POT1b respecti v ely.In this study, we utilized these 'separation-of-functions' POT1 mouse models to better understand how cancer mutations in hPOT1 promote telomere elongation.We generated Pot1b −/ − sarcomas and e xtensi v ely passaged them through serial transplantation in the flanks of SCID mice.While telomere lengths in these tumors initially shortened, we found that the e xtensi v e passaging of these tumors e v entually led to telomere hyper-elongation to lengths that exceeded those observed in Pot1b + / − cells.We show that la te-genera tion Pot1b −/ − sar coma telomer es contain Gquadruple xes, which acti vate an ATR-dependent DDR to recruit telomerase to hyper-elongate telomeres.Increased G-quadruplex accumulation in the telomeric overhang would normally be resolved by POT1b or hPOT1 but not by POT1a.Our results uncover a unique DNA damage protecti v e role for POT1b and provide mechanistic insights into how hPOT1 regulates telomere length.

Generation of MEFs and sarcoma cell lines
CAG-Cre ER ; p53 F / F ; Pot1a F / F ; Pot1b + / − mice were generated from multiple-step cross-mating between Pot1a F / F mice ( 10 ), Pot1b −/ − ( 22 ), p53 F / F mice and CAG-Cre ER mice (Jackson Laboratory).Primary Mouse Embryonic Fibroblasts (MEFs) CAG-Cre ER ; p53 F / F ; mPot1a F / F ; mPot1b + / − and CAG-Cre; p53 F / F ; mP ot1a F / F ; mP ot1b −/ − were isolated from embryos generated from cross-mating between CAG-Cre ER ; p53 F / F ; Pot1a F / F ; Pot1b + / − and p53 F / F ; Pot1a F / F ; P ot1b + / − mice .After treatment with 4-hydro xy-Tamo xyfen (4-HT) or Adeno-Cre and passaging for 3 months, p53 was completely deleted and Pot1a was partially removed in immortalized MEFs.Primary Pot1b −/ − MEFs were immortalized with SV40 to generate the Pot1b −/ − ; p53 −/ − MEF cell line.The injection procedure is described in detail in Figure 1 A. G1 sarcoma cell lines were harvested from ICR-SCID mice sarcoma by subcutaneous injections of G0 MEFs into 8-week-old ICR-SCID mice.Subsequent generations of sarcoma cell lines were isolated from the repeated injections of sarcoma cells into ICR-SCID mice.All mice were maintained according to the IACUC-approved protocols of Yale Uni v ersity.All mouse cell lines were cultured in DMEM / high glucose media supplemented with 7% FBS.The 293T cell line used to generate virus was cultured in DMEM / high glucose media supplemented with 7% FBS.The U2OS cell line was cultured in McCoy's 5A media supplemented with 7% FBS.

Telomere signal analysis by telomere PNA-FISH and CO-FISH
Cells were treated with 0.5 g / ml of Colcemid (Invitrogen) for 4 h before harvest.Chromosomes were fixed with 3% formamide and hybridized with a telomere PNA-FISH 5 -Cy3-OO-(CCCTAA) 3 -3 probe (PNAgene) as described in Wu et al .( 10 ).DNA was counterstained with DAPI.Digital images were captured using NIS-Elements BR (Nikon) with a Nikon Eclipse 80i microscope utilizing an Andor CCD camera.The relati v e telomere signals were analyzed with ImageJ software (downloaded from Fiji).For CO-FISH, cells were incubated with 10 M BrdU for 12 h, treated with 0.5 g / ml of Colcemid for 4 h and harvested.Formalinfixed metaphase spreads were stained with 0.5 g / ml of Hoechst 33258 (Sigma) in 2 × SSC for 15 min at room temperatur e befor e being e xposed to UV light equi valent to 5.4 × 10 3 J / m 2 .After digestion with 200 U of Exonuclease III (Promega), the samples were denatured at 85 • C for 3 min and incubated sequentially with 5 -Cy3-OO-(CCCTAA) 3 -3 and 5 -FAM-OO-(TTAGGG) 3 -3 probes as described above.Images were captured as described above.

Immunofluorescence and fluorescent in situ hybridization
Cells were attached to slides by cytospin, fixed for 10 min in 2% (w / v) sucrose and 2% (v / v) paraformaldehyde at room temperature followed by PBS washes and permeabilized with 0.5% NP40 for 10 min followed by 0.5% Tri-tonX for 5 min.Slides were blocked overnight in blocking solution (0.2% (w / v) fish gelatin and 0.5% (w / v) BSA in 1x PBS a t 4 • C .The cells were incuba ted with primary antibodies overnight at 4 • C.After 0.1% Triton-PBS washes, slides were incubated with the appropriate Alexa fluor secondary antibody for 1 h followed by washes in 1 × PBS with 0.1% Triton.PNA-FISH was carried out as above using a PNA telomere probe 5 -Cy3-OO-(CCCTAA) 3 -3 (PNAgene).Digital images were captured as described above.

Cell-cycle-dependent TIF analysis
G7 Pot1b −/ − sarcoma cells were infected with Fucci-CDT1 or Fucci-geminin lentivirus and maintained for 3 days.Then the cells were processed for immunostaining with ␥ -H2AX and mTRF2 antibodies.The cells were mounted under coverslips without DAPI mounting reagent before taking images.

Detection of telomerase RNA (TR) by RNA-FISH in mouse cells
G3 Pot1b −/ − and G7 Pot1b −/ − sarcoma cells were coinfected with r etroviruses expr essing mouse telomerase from vector HA-mTert / pMGIB and human telomerase RNA fr om vector pBABEpur o-U3-hTR-500.After double selection with 10 g / ml of blasticidin and 2 g / ml of puromycin for several days, the super-mTert / hTR cells were established and used for telomerase recruitment assays.Following immunostaining with antibody, cells were incubated in prehybridization solution (0.1% Dextran sulfate, 1 mg / ml of BSA, 2x SSC, 50% formamide, 0.5 mg / ml spermidine DNA, 0.1 mg / ml Escherichia coli tRNA, 1 mM RNase inhibitor VRC) at 37 • C for 1 h , then hybridized with combined PNA-FISH Cy3-OO-(CCCTAA) 3 and Cy5-hTR cDNA probes in prehybridization solution overnight a t 37 • C .The washing conditions are the same as described for PNA-FISH.Digital images wer e captur ed as above.

TERRA FISH
Cells grown on coverslips were treated with cytobuffer (100 mM NaCl, 300 mM sucrose, 3 mM MgCl 2 , 10 mM PIPES pH7.0, 0.1% Triton X-100, 10 mM vanadyl ribonucleoside complex Sigma #R3380) for 7 min at 4 • C. Cells were rinsed briefly, fixed with 4% paraformaldehyde in PBS for 10 min at RT. Cells were then washed three times with PBS for 5 min each and then incubated with hybridization mix (10 nM TERRA FISH probe 5 -(TAACCC) 7 -Alexa488-3 , Integrated DNA Technologies , 50% formamide , 2 × SSC, 2 mg / ml BSA, 10% dextran sulfa te, 10 mM vanad yl ribonucleoside complex) for 18 h in a humidified chamber at 39 • C. Cells were washed with 2 × SSC in 50% formamide three times at 39 • C for 5 min each, three times in 2 × SSC at 39 • C for 5 min each, and finally once in 2 × SSC at RT for 10 min.Coverslips were than mounted on glass microscope slides with DAPI.Digital images were captured as described above.

TRF Southern analyses
A total of 2 × 10 6 cells were suspended in PBS, 1:1 mixed with 1.8% agarose in 1x PBS and cast into plugs.The plugs were digested at 55 • C for 2 days with 1 mg / ml proteinase K (Roche) in 10 mM sodium phosphate (pH7.2) and 0.5 mM EDTA and 1% sodium lauryl sarcosine.After completely washing away proteinase K with TE buffer, the DNA in plugs were subsequently digested with RsaI and HinfI overnight a t 37 • C .Plugs were washed 3 × at 55 • C in ExoI buffer (67 mM Tris-HCl, 6.7 mM MgCl 2 , 10 mM ␤mercaptoethanol, 0.1 mg / ml Bovine Serum Albumin) with 150 mM of either KCl or LiCl.Plugs were cooled and then digested overnight at 37 • C in ExoI buffer.Plugs were loaded onto a 0.8% pulse-field agarose (Bio-Rad) gel in 0.5x TBE and electrophoresed on a CHEF-DRII pulse field electrophoresis apparatus (Bio-Rad).The electrophoresis conditions were as follows: initial pulse 0.3 s, final pulse 16 s, voltage 6 V / cm, run time 14.5 h.The gels were dried and pre-hybridized in Church mix (0.5 M NaH 2 PO 4 , pH7.2, 7% SDS), and then hybridized with telomeric repeat oligonucleotide probes ␥ -32 P-(CCCTAA) 4 or ␥ -32 P-(TTAGGG) 4 in Church mix at 55 • C overnight.Gels were washed with 4 × SSC, 0.1% SDS buffer at 55 • C and exposed to Phosphorimager screens overnight.The screen was scanned on a Typhoon Trio image system (GE Healthcare).For quantification of total telomeric DNA, the gels were de-probed with denaturing solution 0.5 M NaOH, 1.5 M NaCl and neutralized with 3 M NaCl, 0.5 M Tris-Cl, pH7.0, and re-probed with telomeric ␥ -32 P-(CCCTAA) 4 or ␥ -32 P-(TTAGGG) 4 probes.Quantification of signal intensity was performed on ImageJ.Nati v e signals were normalized to total telomeric signal.

2D-TRF Southern assay
TRF Southern plugs were prepared as described above and digested with Hinf1 / Rsa1 befor e r esolving them in the first dimension in 0.4% agarose gel in 0.5 × TBE at 26 V for 15 h.Ethidium bromide-stained gel slices were rotated 90 • from the first gel and cast in a 1% agarose gel and resolved at 115 V for 4 h.Dried gels wer e denatur ed and subsequently hybridized as described above.After exposure, hybridization signals were analyzed with a Typhoon Trio imager system and ImageQuant TL software.

TRAP assay
TRAP assa ys were perf ormed as described in the manufacturer's protocol (Millipore #S7700).Briefly, 1 × 10 6 cells were lysed with 200 l of CHAPS buffer and the soluble fractions wer e fr eshly frozen in liquid nitrogen and stored a t −80 • C .The telomerase reaction was carried out in 25 l TRAP buffer containing: 20 mM Tris-Cl (pH8.0),63 mM KCl, 1.5 mM MgCl 2 , 1 mM EGTA, 2 ng / l of ␥ -32 P labeled TS primer (AAT CCGT CGA GCA GA GTT), 50 M 4dNTP and 1x TRAP primer mix at 30 • C for 40 min and quenched at 95 • C for 2 min.Immediately following the addition of Taq DN A pol ymerase, the PCR was performed as follows: 94 • C 30 s, 59 • C 30 s repeating for 28 times.The PCR reaction was separated on 10% acrylamide gel in 0.5 × TBE (Invitrogen EC62752) and electrophoresed for 81 min under 150 V.The gel was dried and exposed to phosphorimager screen.The radioacti v e signals were captured and quantified as described for TRF Southern blotting.

C-circle assay
The C-circle assay was performed as described in Henson et al. ( 44 ).Briefly, genomic DNA was digested by RsaI and HinfI overnight at 37 • C. The digested DNA was precipitated using ethanol and resuspended in H 2 O. Digested DNA was then amplified using Phi29 polymerase (New England Biolabs M0269S).The amplified product was then immobilized to an Amersham Hybond N + membrane using a slot dot blot apparatus.C-circle products were then detected by hybridization with ␥ -32 P-(CCCTAA) 4 telomeric probes.

Western blot analyses
Trypsinized cells were lysed in urea lysis buffer (8 M urea, 50 mM Tris-HCl, pH 7.4, and 150 mM ␤merca ptoethanol).The l ysates were denatured and then resolved on an SDS-PAGE gel.The separated proteins were then blotted on a nitrocellulose plus membrane (Amersham), blocked with blocking solution (5% non-fat dry milk in PBS / 0.1% Tween-20) for at least 1 h and incubated with the appropriate primary antibody in blocking solution at least 2 h at room temperature or overnight at 4 • C. The membranes were washed 3 × 5 min with PBS / 0.1% Tween-20 and incubated with appropriate secondary antibody in blocking solution for 1 h at room temperature.Chemiluminescence detection was performed using an ECL Western Blotting Detection kit from GE Healthcare.
In contrast to P ot1b + / − sarcomas, G1-G3 P ot1b −/ − sarcomas display chromosomal fusions, with up to 14% of chromosomes containing Robertsonian chromosomal fusions without telomeric signals at the fusion site (Figure 1 C, E).These types of end-to-end chromosomal fusions are indicati v e of the critical telomere shortening observed in the absence of Pot1b ( 21 , 22 , 47 ).Howe v er, the number of chromosomal fusions did not increase in G4 to G7 Pot1b −/ − sar comas, r emaining at ∼14% of all chromosomes (Figure 1 C, E).We speculate that telomere hyper-elongation acquired in later generations is counteracting the initial telomere shortening to pre v ent further accumulation of fused chromosomes.The Robertsonian translocations accrued during G1-G3 would be stable and heritable as they contain only one centromere and so the chromosomal fusions would be retained in la te-genera tion Pot1b −/ − cells.To evaluate the ability of hyper-elongated telomeres from G7 Pot1b −/ − cells to inhibit DNA repair, we expressed the dominant negative shelterin proteins TPP1 RD and TRF2 B M ( 12 , 48 ).TPP1 RD and TRF2 B M expression had no effect on the number of chromosomal fusions without telomeric signals but led to a 15% increase of chromosomal fusions with telomeric signals, suggesting that expressing these two proteins deprotected telomeres in both G7 P ot1b + / − and P ot1b −/ − cells (Supplementary Figure S2a, d-f).These results suggest that hyperelongated telomeres in G7 Pot1b −/ − cells r epr ess aberrant chromosomal fusions similar to those in Pot1b + / − cells.We next set out to determine whether hyper-elongated telomer es ar e able to protect against activation of a DDR.G7 Pot1b −/ − cells had detectable ␥ -H2AX-positive telomere dysfunction induced foci (TIFs) in ∼22% of nuclei (Supplementary Figure S2a-c).Altogether, these results suggest tha t hyper-elonga ted telomer es in Pot1b null sar comas r emain protecti v e against DNA r epair at telomer es and pr event chromosomal fusions but are unable to fully r epr ess DDR activation.

Telomer e hyper -elongation in late-gener ation Pot1b −/ − tumors is mediated by telomerase
Telomere hyper-elongation can be mediated either by telomerase or via the Alternati v e Lengthening of Telomeres (ALT) pathway ( 49 ).To determine if telomerase is responsible for the elongating telomeres in late-generation Pot1b −/ − cells, we utilized a G3 Pot1b −/ − cell line with hypomorphic telomerase activity ( Pot1b −/ − ; Tert hypo ) that we had previously generated using CRISPR / Cas9 ( 23 ).TRF Southern analysis re v ealed that serial transplantation of G7 Pot1b −/ − ; Tert hypo cells failed to induce telomere hyper-elongation (Figure 2 A).These results suggest that WT le v el of telomerase activity is r equir ed for Pot1b −/ − cells to undergo telomere hyper-elongation.To further confirm a role for telomerase in telomere hyper-elongation, we used CRISPR / Cas9 to knockout Tert in G7 Pot1b −/ − cells (G7 Pot1b −/ − ; Tert −/ − ) (Supplementary Figure S3ac).Telomere Q-FISH analysis following serial transplantation of G7 Pot1b −/ − ; Tert −/ − cells re v ealed that deletion of telomerase leads to rapid telomere erosion, with the average telomere intensity decreasing by 36% (Figure 2 B, C).In addition, both telomer e-fr ee chr omosome ends and chr omosome fusions without telomeric signals increased by 1.5fold, further re v ealing the necessity of telomerase in lategeneration Pot1b −/ − cells for telomere hyper-elongation (Figure 2 C-E).
Two phenotypes associated with high specificity for ALT are extrachromosomal telomeric C-circles and ALTassociated PML nuclear bodies (APBs) ( 44 , 50 ).Lategeneration Pot1b −/ − cells displayed no detectable C-circles or APBs (Figure 2 F-H).In addition, 2D-gel analysis of la te-genera tion Pot1b −/ − cells had no detectab le e xtrachromosomal telomeric DNA such as telomere (T)-circles or recombina tion intermedia tes such as telomer e (T)-complex es (Supplementary Figure S3d, e) ( 51 ).We also did not detect any increase in the rate of telomere sister chromatid exchange (T-SCE) another prominent phenotype observed in ALT cells (Supplementary Figure S4a, b) (52)(53)(54).Furthermor e, ther e was no evidence at telomeres in G7 Pot1b −/ − cells of increased replication stress or telomeric repeat-containing RNA (TERRA) transcription, phenotypes thought to dri v e ALT acti vity (Supplementary Figure S4c-e) (55)(56)(57).The absence of ALT phenotypes strongly suggests that the telomere hyper-elongation of lategeneration Pot1b −/ − cells is mediated by telomerase rather than through the activation of ALT.

Telomer ase r ecruitment to telomer es is incr eased in lategeneration Pot1b −/ − cells
We next set out to identify the mechanism underlying telomerase-media ted telomere hyper-elonga tion.We first examined whether la te-genera tion Pot1b −/ − cells have increased telomer ase tr anscription.TER T promoter m utations are known to increase telomer ase tr anscription in many cancers (58)(59)(60)(61)(62). Howe v er, sequencing the Tert promoters in G1 and G7 Pot1b −/ − cells did not re v eal any muta tions (da ta not shown).To evalua te if telomerase enzymatic activity is increased in G7 Pot1b −/ − sarcomas, we performed a quantitati v e Telomere Repeat Amplification Protocol (TRAP) assay ( 7 ).Compared to G1-G3 Pot1b −/ − sarcomas, telomerase enzymatic activity did not increase appreciably in G7 Pot1b −/ − cells (Figure 3 A, B).We ne xt e xamined whether telomerase recruitment to telomeres was increased in G7 Pot1b −/ − sarcomas ( 40 ).Compared to G3 Pot1b −/ − cells, we detected a 1.8-fold increase in telomerase co-localization with telomeres in G7 Pot1b −/ − cells (Figure 3 C, D), suggesting that the increased telomerase recruitment contributed to telomere hyper-elongation.Since telomerase recruitment is mediated through its interaction with TPP1's TEL patch and NOB domains ( 63-65 ), we sequenced these domains for mutations that might promote telomerase recruitment but did not detect any (data not shown).
Previous data re v eal that there are different requirements for POT1-mediated telomere protection in the G1 and S / G2 phases of the cell cycle ( 13 ).We utilized the FUCCI system to determine whether the source of the DNA damage observed in G7 Pot1b −/ − cells was limited to a particular stage of the cell cycle ( 41 ).␥ -H2AX TIFs were detected in 40% of nuclei in S / G2, marked by Geminin expression, w hile onl y 4% of nuclei in the G1 phase, indicated by CDT1 expr ession, wer e TIF-positi v e (Figure 4 G, H).These da ta indica te tha t the G7 Pot1b −/ − telomeres are recognized as damaged ss-DNA in the S / G2 phase of the cell cycle and that hyper-elongated telomeres remain protecti v e during G1.This suggests processing of the overhang during S / G2 in la te-genera tion Pot1b −/ − cells likely underlies this dysfunction.
The N-terminus of POT1b r epr esses a DDR in lategeneration Pot1b −/ − cells PO T1a and hPO T1 r epr ess ATR activa tion a t telomeres by binding to the G-strand ss-telomeric DNA to exclude RPA ( 13 , 14 , 75 ).Since p-RPA TIFs are present in G7 Pot1b −/ − cells, we reasoned that endo genous PO T1a cannot adequately protect the elongated ss-telomeric DNA.PCR genotyping re v ealed that Pot1a expression is intact in G7 Pot1b −/ − cells (Supplemental Figure S1a).We next overex-pressed POT1a or POT1b in the G4 Pot1b −/ − cells to evaluate if increasing POT1 proteins could repress the DNA damage.Une xpectedly, ov ere xpression of POT1b, but not POT1a, in G4 Pot1b −/ − cells abolished both ␥ -H2AX and p-RPA TIF formation by ∼96% and ∼93%, respecti v ely (Figure 5 A, B, Supplementary Figure S5d-g).It has ne v er been demonstrated previously that POT1b has a unique ability to r epr ess a DDR that POT1a cannot.Howe v er, this r esult r e v ealed that the DDR observ ed in G4 Pot1b −/ − cells are uniquely protected by POT1b but not by POT1a.We next set out to determine how POT1b is r epr essing TIFs.Unlike PO T1a, PO T1b is critical for the recruitment of the CST complex to telomeres to mediate C-strand synthesis ( 16 ).Howe v er, e xpression of POT1b CST , a mutant unable to recruit the CST complex to telomeres ( 16 ), also repressed TIFs in la te-genera tion Pot1b −/ − cells (Figure 5 A, B, Supplementary Figure S5d-g).We additionally confirmed that PO T1b's DN A binding activity is r equir ed to r epr ess TIFs through expression of POT1b F62A , an N-terminal mutant unable to bind to ss-telomeric DNA ( 10 ) (Figure 5 A, B, Supplementary Figure S5d-g).To further dissect which domains of POT1b ar e r equir ed to r epr ess TIFs, we over expressed the PO T1a 1-350 -PO T1b 351-640 chimeric protein (abbreviated POT1ab) or the POT1b 1-350 -POT1a 351-640 protein (abbreviated POT1ba) ( 23 ) in G7 Pot1b −/ − cells (Figure 5 C).POT1ba but not POT1ab r epr essed ␥ -H2AX TIFs by ∼73%, confirming that the N-terminus but not the Cterminus of POT1b is r equir ed for POT1b's ability to repr ess ␥ -H2AX TIFs (Figur e 5 D, E, Supplementary Figure S6a, b).Importantl y, the PO T1ba chimera lacks critical amino acids r equir ed for r ecruitment of the CST-complex to telomeres ( 16 ) (Supplementary Figure S6c, d), further supporting the notion that POT1b's telomere protecti v e function is independent of its role in CST recruitment.We found that ov ere xpressing hPOT1 repressed ␥ -H2AX TIFs in G7 Pot1b −/ − cells, suggesting that POT1b's protecti v e function is conserved in hPOT1.We confirmed these results by overexpressing the chimeric POT1a / b and hPOT1 constructs in independently generated Pot1b −/ − ; p53 −/ − sarcomas, demonstra ting tha t POT1ba and hPOT1 r epr ess ␥ -H2AX TIFs to nearly undetectable levels (Supplementary Figure S6e-g).These results suggest that the DNA-protecti v e ability of POT1b in G7 Pot1b −/ − cells is distinct from that of POT1a.

Telomer e hyper -elongation in G7 Pot1b −/ − cells is due to the activation of ATR-dependent DDR at telomeres
Both telomere hyper-elongation and TIFs were not observed in ear ly-gener ation but only in la te-genera tion Pot1b −/ − cells (Figure 1 B-D, Figure 4 A, B, Supplementary Figure S5a), suggesting that these phenotypes are likely associated to each other.Telomerase recruitment in yeast is promoted through activation of homologues of the ATR and ataxia-telangiectasia mutated (ATM) DDR pathways ( 69 , 70 , 72 ).ATR and ATM pathways are critical for facilitating telomerase recruitment in mammalian systems ( 39 , 68 ).We postula ted tha t the elevated DDR observed in G7 Pot1b −/ − cells (Figure 3 C, D) led to increased telomerase recruitment.To test this notion, we repressed the DDR at telomeres of G7 Pot1b −/ − cells by  ov ere xpressing POT1b, and POT1ba.Ov ere xpression of POT1b r epr essed TIF formation in G7 Pot1b −/ − cells but did not lead to a decrease in telomerase recruitment (Figure 6 A-C, Supplementary Figure S7a, b).Howe v er, the C-terminus of full-length POT1b would not only r epr ess the DDR at telomeres but would also promote telomerase recruitment through its interaction with TPP1 ( 23 ), likely conf ounding an y effect r epr ession of DDR by full-length POT1b might have on telomerase r ecruitment.Over expr ession of POT1ba in G7 Pot1b −/ − cells r epr essed TIF formation and decreased telomerase recruitment by approximately 28% (Figure 6 A-C, Supplementary Figure S7a, b).To interrogate the role of ATR-dependent DDR in telomerase r ecruitment mor e dir ectly, we inhibited ATR in G7 Pot1b −/ − cells using ATRi AZ20 and measured telomerase r ecruitment.ATRi decr eased TIF formation and telomerase recruitment rates by ∼26%, further suggesting that the DDR activation is r equir ed to promote telomerase recruitment to telomeres in G7 Pot1b −/ − cells (Figure 6 A-C).
A TR / A TM pathways are critical for telomerase comple x assemb l y ( 39 ).Potentiall y, inhibition of telomerase recruitment by ATRi could be the result of decreased telomerase complex biogenesis instead of telomerase recruitment to telomeres.Similarl y, PO T1ba might be inhibiting telomerase recruitment through similar mechanisms as POT1a (Figure 3 C, D) and not through r epr ession of TIFs ( 23 ).We ther efor e combined POT1ba + ATRi tr eatments to determine if these conditions are epistatic, resulting in diminished telomerase r ecruitment.When compar ed to POT1ba + DMSO (28% decrease) or Vector + ATRi (26% decrease), POT1ba + ATRi did not further reduce telomerase recruitment, with a 29% decrease in TR / telomere colocalizations per nucleus compared to Vector + DMSO (Figure 6 A-C).This result suggests that both POT1ba and ATRi treatment inhibited telomerase recruitment through the same pathway.Similarl y, PO T1b WT ov ere xpression prevents ATRi inhibitory effect on telomerase recruitment as POT1b WT + ATRi did not decrease telomerase recruitment compared to POT1b WT + DMSO (Figure 6 A-C).Gi v en that POT1b WT r epr esses TIFs, this finding is consistent with the hypothesis that elevated ATR-dependent DDR at telomeres is promoting telomerase recruitment in G7 Pot1b −/ − cells.To confirm that it is the r epr ession of the ATR-dependent DDR at telomeres that is reducing telomerase recruitment, we measured the effect of POT1ba and ATRi in G3 Pot1b −/ − cells without robust DDR activa tion a t telomeres (Figure 6 D-F, Supplementary Figure S7c, d).Neither POT1ba nor ATRi decreased telomerase recruitment in G3 Pot1b −/ − cells, re v ealing that incr eased telomerase r ecruitment to dysfunctional telomeres in G7 Pot1b −/ − cells is responsible for its telomere hyper-elongation phenotype (Figure 6 D-F, Supplementary Figure S7c, d).
Telomerase is required for DNA damage at telomeres in lategeneration Pot1b −/ − cells An important question to address is how damaged telomeric DNA is generated in G7 Pot1b −/ − cells.POT1b loss would pre v ent CST r ecruitment to telomer es, leading to elevated telomere replication-dependent defects detected as fragile telomeres.Howe v er, telomere fragility was indistinguishable between Pot1b + / − and G7 Pot1b −/ − cells (Supplementary Figure S4a, b).Furthermore, POT1b WT overexpr ession, which r eadily r epr essed TIF formation, did not decrease the number of fragile telomeres (Figure 5 A, B, Supplementary Figure S4a, b).These results indicate that r eplication str ess is unlikely to be the sour ce of telomere damage.Another possible source of telomere damage is telomeric repeat-containing RNA (TERRA), which promotes RPA displacement and POT1 binding following telomer e r eplication ( 75 ).Howe v er, TERRA was not detectable by RNA-FISH in la te-genera tion Pot1b + / − and Pot1b −/ − cells (Supplementary Figure S4d, e).Another potential source of damaged telomeric DNA could arise from aberrant telomeric repeats generated by telomerase ( 76 , 77 ) as it hyper-elongates telomeres.To address this hypothesis, we deleted Tert using CRISPR / Cas9.While the parental G7 Pot1b −/ − cells displayed ≥5 ␥ -H2AX TIFs in ∼25% of nuclei, deletion of Tert completely abolished the formation of ␥ -H2AX-positi v e TIFs in two independent CRISPR / Cas9 Tert KO clones (Figure 7 A, B).TERT WT ov ere xpression in G7 Pot1b −/ − ; Tert −/ − cells r eadily r estor ed ␥ -H2AX TIFs (Figure 7 C, D, Supplementary Figure S8a-d), re v ealing that TIF forma tion in la te-genera tion Pot1b −/ − cells r equir es telomerase.Ov ere xpression of TPP1 RD in G7 Pot1b −/ − ; Tert −/ − cells lead to robust TIF formation, suggesting that DDR pathways are not impaired by Tert deletion (Figure 7 C, D).To understand the mechanism through which telomerase promotes telomeric damage in G7 Pot1b −/ − cells, we ov ere xpressed TERT K78E (which disrupts telomerase recruitment to telomeres but not catalytic activity) ( 78 , 79 ) and TERT K560N (which abolishes telomerase catalytic activity but not recruitment) ( 79 , 80 ) in G7 Pot1b −/ − ; Tert −/ − cells (Figure 7 C-E, Supplementary Figure S8ad).Neither TER T K78E nor TER T K560N ov ere xpression restored TIF formation in G7 Pot1b −/ − ; Tert −/ − cells (Figure 7 C-E, Supplementary Figure S8a-d), indicating that telomerase recruitment to telomeres and its catalytic activity are both r equir ed f or TIF f orma tion.Our da ta suggests tha t the telomeric repeats synthesized by telomerase in G7 Pot1b −/ − cells are recognized as damaged DNA.

The telomeric 3 overhangs of G7 Pot1b −/ − sarcomas contain G-quadruple x es
We used nati v e and denaturing TRF Southerns to characterize the nature of the telomeric DNA damage in G7 Pot1b −/ − cells.G7 Pot1b −/ − telomeres possess only single-stranded TTAGGG overhangs while ss CCCTAAoverhangs were not detected (Supplementary Figure S9a).Surprisingly, the ss G-overhang in G7 Pot1b −/ − cells was completely resistant to the E. coli 3 -5 Ex onuclase I (Ex oI) digestion (Figure 8 A, Supplementary Figure S9b), suggesting that it contains secondary structures that pre v ents ExoI hydrolysis.The G-rich telomeric strand is known to fold into G-quadruplexes (GQ), four-stranded DNA helices formed from stacks of square-planar arrays known as Gquartets which are stabilized by guanine hydrogen bonding and monovalent cations such as K + or Na + but destabilized by Li + cations ( 81 , 82 ).GQs inhibit E. coli ExoI mediated digestion of telomeric DNA in vitro ( 83 ).We   postula ted tha t the long ss G-overhangs in hyper-elonga ted G7 Pot1b −/ − telomeres might be prone to form GQ structures inhibitory to ExoI digestion.To test this hypothesis, we performed ExoI digestion in a GQ-stabilizing (150 mM KCl) or a GQ-destabilizing (150 mM LiCl) buffer (Figure 8 B, Supplementary Figure S9c).We again observed a complete resistance to ExoI digestion in DNA from the G7 Pot1b −/ − cells treated with 150 mM KCl (Figure 8 B, Supplementary Figure S9c).Howe v er, treatment of G7 Pot1b −/ − telomeres with 150 mM LiCl allowed ExoI to almost completely digest the ss G-rich overhang (Figure 8 B, Supplementary Figure S9c).We confirmed this phenotype in an independent Pot1b −/ − ; p53 −/ − sarcoma where telomeric overhangs were also resistant to ExoI digestion in the presence of KCl but sensitized with LiCl (Supplementary Figure S9d, e).Previous in vitro and structural works have suggested that GQs form during telomerase activity ( 84 , 85 ).To evaluate if the increased telomerase elongation is contributing to GQ formation, we performed ExoI digestion in two G7 Pot1b −/ − ; Tert −/ − cell lines we had generated using CRISPR / Cas9.While the telomeric overhangs in the parental line were nearly completely resistant to ExoI digestion, Tert KO sensitized the ss G-rich overhang by ∼37% in both CRISPR / Cas9 clones (Supplementary Figur e S9f, g).Over expr essing TERT in G7 Pot1b −/ − ; Tert −/ − cells reduced the sensitivity to ExoI hydrolysis (Supplementary Figure S9f, g).These data suggest that the increased telomerase-mediated telomere elongation is promoting GQ formation in the telomeric ov erhang.Howe v er, the ss G-rich overhang is still partially resistant to ExoI digestion in G7 Pot1b −/ − ; Ter t −/ − DNA suggesting tha t GQ resolution is still impaired e v en in the absence of telomerase.
We incubated POT1 proteins at various concentrations with GGG(TTAGGG) 3 GQs, removed proteins with SDS and Proteinase K treatment and performed nati v e gel electrophoresis to detect GQs and unfolded GQs.While 50 nM of hPOT1 and POT1b were able to unfold GQs, POT1a was unable to unfold GQs at this protein concentration.Compared to hPOT1 and PO T1b, PO T1a was also less efficient at unfolding GQs at all protein concentrations (Figure 8 E).These results re v ealed that POT1b, but not POT1a, can efficiently interact with and unfold GQs similar to hPOT1.

DISCUSSION
Serial transplantation of Pot1b −/ − sarcomas into SCID mice has uncovered a role for POT1b in the regulation of telomere elongation and repression of an ATR-dependent DDR.Our speculati v e model is shown in Figure 8 F. Telomere shortening in mid-generation Pot1b −/ − cells leads to activation of a DDR, promoting telomerase recruitment and telomere hyper-elongation.Increased telomere hyperelongation leads to GQ accumulation in the telomeric overhang that would otherwise be resolved by POT1b.Without POT1b, the GQ structures are unfolded by RPA, leading to activation of an aberrant ATR-dependent DDR.The ATR pathway then promotes further telomerase recruitment, resulting in a positi v e feed-back loop to generate further telomere hyper-elongation and telomere DDR activa tion.We demonstra te tha t the r epr ession of this DDR at telomeres is conserved between POT1b and hPOT1, suggesting that similar mechanisms may underly the phenotypes observed in tumors harboring hPOT1 cancer mutations.

Telomere elongation in POT1 mutant cancers
Next generation sequencing identified hPOT1 mutations in a wide variety of cancers, making hPOT1 the most frequentl y m utated shelterin component in human cancers ( 27 , 28 , 30 , 36 , 86 ).Howe v er, gi v en the di v erse functions that hPOT1 plays in telomere regulation, our understanding of its functions in healthy and diseased tissues remain incomplete.Two phenotypes commonly associated with cancer hPO T1 m utations ar e telomer e elongation and genomic instability ( 25 ).Through long-term serial passaging of Pot1b −/ − sarcomas in SCID mice, we were able to genera te la te-genera tion Pot1b −/ − cell lines displaying both telomere hyper-elongation as well as an ATR-dependent DDR at telomer es, ther eby mimicking the most prominent phenotypes observed in cancers with hPOT1 mutations.We provide experimental evidence in Pot1b −/ − cells that the telomere hyper-elongation and DDR activation are inextricably linked, with each phenotype promoting the other.POT1b has a role distinct from POT1a in pre v enting ATR activa tion a t telomer es and is specifically r equir ed to r esolve GQs that are produced during telomere elongation by telomerase (Figure 8 F).

Incr eased telomer ase r ecruitment and telomer e hyperelongation mediated by ATR activation
POT1b is critical for telomere length maintenance by recruiting the CST complex to modulate C-strand fill-in and promoting telomerase recruitment to the G-strand ( 16 , 23 ).Consequentl y, PO T1b loss has been associated with accelerated telomere shortening rather than telomere elongation ( 21 , 22 ).Consistent with this notion, early generation Pot1b −/ − sarcomas also undergo progressi v e telomere shortening (Supplementary Figure S1b, c).Howe v er, after undergoing e xtensi v e passaging in vivo , telomere shortening in Pot1b −/ − sarcomas is re v ersed and telomerasemediated telomere hyper-elongation phenotype begins to dominate in two independently generated Pot1b −/ − sarcoma serial transplantations (Figure 1 B-D, Supplementary Figure S1d, e).Because telomere hyper-elongation is only observed after extensively passaging of Pot1b −/ − cells in vivo , we speculate that global telomere shortening is critical to begin the process of telomere hyper-elongation.Previous work has suggested that the shortest telomeres are preferentially elongated by telomerase (87)(88)(89)(90).Consequently, it is likely that a few shortened telomeres of mid-generation Pot1b −/ − cells promote telomerase recruitment.Elongation of the shortest telomeres is well characterized in S. cerevisiae which r equir es the yeast ATM homolog Tel1p, suggesting a role for the DDR to promote telomerase recruitment to critically shortened telomeres ( 66 , 67 , 71 ).Extensi v e work in both fission and budding yeast also demonstrates that Mec1p / Rad3p, yeast homologs of ATR, share functional redundancy with Tel1p and are both critical in facilitating telomerase recruitment to telomeres ( 69 , 70 , 72 , 91 ).A TM and A TR ar e r equir ed for telomerase r ecruitment in mammalian cells, suggesting functional conservation of DDR regulation of telomerase-mediated telomere length maintenance ( 39 , 68 ).Ther efor e, critically short telomer es may activa te DDR pa thways as an initiating e v ent, leading to increased telomerase recruitment in mid-generation Pot1b −/ − sarcomas.Howe v er, we cannot e xclude the possibility that these Pot1b −/ − sarcomas may have accumula ted soma tic muta tions during passaging tha t contribute to telomere elongation.Once critically short telomeres elongate in a POT1b-proficient setting, shelterin function would normally be r estor ed to r epr ess the DDR and maintain telomere lengths.Howe v er, we demonstra te tha t telomerase activity in the absence of POT1b produces a DDR at telomer es (Figur e 7 ).This DDR then promotes telomerase recruitment leading to further telomere elongation in a positi v e feedback loop, resulting in sustained telomere hyperelongation (Figure 8 F).We speculate that early generation Pot1b −/ − cells have not entered this positive feedback loop as the DDR is substantially weaker, leading to continued telomere shortening.
Pre vious inv estigations of hPO T1 and PO T1a functions re v ealed that disruption of their OB-folds leads to incr eased telomer e r eplication str ess and telomer e elongation ( 33 , 35 , 39 , 92 ).Incr eased r eplication str ess, ATR activation and rapid telomere elongation are also reminiscent of the phenotypes observed with TRF1 loss ( 93 ).Howe v er, we are unable to detect any signs of replication stress in lategeneration Pot1b −/ − cells (Figure 1 B-D, Supplementary Figure S4a, b).Our data suggest that an ATR-dependent DDR alone could promote telomerase-mediated telomere hyper-elongation without requiring activation of replication stress at telomeres.

POT1b r epr esses ATR activ ation and GQ f ormation at telomeres
Pr evious r eports suggest that POT1b does not play a significant role in protecting ss-telomeric DNA from activating a DDR ( 9 , 22 ).Howe v er, our data re v eal a uncharacterized role for POT1b in protecting telomeres from activating a DDR at telomer es (Figur e 5 A, B).Ear ly-gener ation Pot1b −/ − cells do not have a detectable DDR at telomeres and TIF formation only appeared after e xtensi v e in viv o passa ging (Figure 4 A, B).Activation of this DDR in the absence of POT1b has ne v er been characterized.We also discovered that telomerase activity at telomeres is required f or TIF f ormation in the absence of POT1b (Figure 7 A-E).Our data suggest that telomerase activity is generating the GQs (Supplementary Figure S9f, g) that ar e r eco gnized as DN A damage by RPA, w hich explain why telomerase is r equir ed for DDR activation.In vitro e xperiments hav e demonstra ted tha t GQs form within acti v ely e xtending telomeres ( 84 ).Recent structural data supports this finding by re v ealing that hPOT1 and telomerase form a cavity that is sterically large enough to accommodate folding of telomeric DNA into a GQ ( 85 ).This suggests that nascent repeats polymerized by telomerase form GQs which then must be unfolded by hPOT1 ( 84 ).POT1b but not POT1a is pr efer entially incorporated into the telomerase / TPP1 / POT1b / telomere complex to initiate telomere elongation, as POT1b promotes telomerase recruitment to telomeres ( 23 ).POT1b would then be in a prime position to resolve GQ structures that fold as telomerase synthesizes new telomeric repeats.GQs have been strongly implicated to be obstacles for replication fork progression and are a source for increased replication stress ( 94 , 95 ).Howe v er, la te-genera tion Pot1b −/ − cells do not display increased fragile telomere and show no overt signs of r eplication str ess (Supplementary Figur e S4a, b).Since GQs gener ated by telomer ase would be contained within the 3 telomeric overhang, it is possible that the replication machinery ne v er r eaches the GQs as the eukaryotic r eplicati v e 3 -5 helicase, minichromosome maintenance (Mcm2-7) complex, may fall off the 5 end at the ds-ss telomere junction ( 96 ).
In the absence of POT1b, unresolved GQs in telomeres may be unfolded by other proteins such as RPA, a known GQ resolving protein complex ( 97 ).In this setting, GQs would aberrantly activate an ATR-mediated DDR.Consequentl y, PO T1b's resolution of GQs may be the mechanism through which POT1b is r epr essing the activation of a DDR in telomeres of la te-genera tion Pot1b −/ − cells (Figure 5 ).This model is also consistent with POT1a's inability to r epr ess TIF formation in G7 Pot1b −/ − sar comas (Figure 5 ), as POT1a is less proficient at resolving GQ than POT1b (Figure 8 D, E).Additionally, POT1b's r epr ession of TIF formation in G7 Pot1b −/ − cells r equir es its N-terminus (Figure 5 C-E), which contains the ss-telomeric DNA binding OB folds critical for GQ resolution (Supplementary Figure S9h, i).Interestingly, the CST complex also has GQ resolving properties ( 98 ) and it is possible that POT1b recruitment of the CST complex ( 16 ) may further facilitate GQ unfolding.Howe v er, our data re v eal that PO T1b alone is full y

Figure 1 .
Figure 1.Telomere hyper-elongation in late-generation Pot1b −/ − cells.( A ) Schematic of serial transplantation through subcutaneous injections in SCID mice.( B ) TRF Southern blot detection of G-overhang in native gel and total telomere length in denatured gel hybridization with ␥ -32 P-(CCCTAA) 4 telomere probe.Numbers indicate relati v e G-ov erhang and total telomer e signals, with telomer e signals set to 1.0 for Pot1b + / − MEFs.Molecular weight markers as indica ted.( C ) Representa tive images of metaphase spreads of the indicated cell lines visualized with PNA-FISH probe Cy3-OO-(CCCTAA) 3 and DAPI.Scale bar: 5 m. ( D ) Q-FISH analysis showing the median telomeric signal intensity from metaphases in (C).30 metaphases wer e scor ed per cell line.( E ) Quantification of chromosomal fusions in metaphases in (C).Data show the mean ± standard deviation from three independent experiments with at least 1000 chromosomes analyzed for each cell line per experiment.P -values are shown and generated from one-wa y ANOVA analysis f ollowed by Tukey's multiple comparison.

Figure 2 .
Figure 2. Telomere hyper-elongation is mediated by telomerase.( A ) TRF Southern blot detection of G-overhang in native gel and total telomere length in denatured gel hybridized with ␥ -32 P-(CCCTAA) 4 telomere probe.Numbers indica te rela ti v e G-ov erhang and total telomere signals, with telomere signals set to 1.0 for WT MEFs.Molecular weight markers as indicated.( B ) Q-FISH analysis showing the median telomeric signal intensity from metaphases in (C).30 metaphases were scored per cell line.( C ) Representati v e images of metaphase spreads of the indicated cell lines visualized with PNA-FISH probe Cy3-OO-(CCCTAA) 3 and DAPI.Scale bar: 5 m. ( D ) Quantification of signal-free ends in metaphases in (C).Data show the mean ± standard deviation from two independent experiments with 30 metaphases analyzed for each cell line per experiment.P -values are shown and generated from one-way ANOVA analysis followed by Tukey's multiple comparison.( E ) Quantification of chromosomal fusions from metaphases in (C).Data show the mean ± standard deviation from two independent experiments with 30 metaphases analyzed for each cell line per experiment.P -values are shown and generated from two-way ANOVA analysis followed by Tukey's multiple comparison.Blue P -values are used for chromosomal fusions without telomeric signals.Red P -values are used for chromosomal fusions with telomeric signals.( F ) C-circle detection in the indicated cell lines.Slot-blot of C-circle assay performed on genomic DNA with and without Phi29 DNA polymerase.U2OS was used as a positi v e control.( G ) Representati v e images of APBs in the indicated cell lines.Cells were immunostained with ␣-PML antibody (gr een), telomer e probe Cy3-(CCCTAA) 3 (red) and DAPI to stain nuclei (blue).Scale bar: 5 m. ( H ) Quantification of the indicated cell lines in (G) with ␣-PML foci colocalizing with telomeres.Data show the mean ± standard deviation from two independent experiments with 150 nuclei analyzed for each cell line per experiment.P -values are shown and generated from one-way ANOVA analysis followed by Tukey's multiple comparison.U2OS was used as a positi v e control.

Figure 3 .
Figure 3. Increased telomerase recruitment to telomeres in la te-genera tion Pot1b −/ − cells.( A ) TRAP assay of the indicated cell lines.WT MEFs were used as a positi v e control and heat denatured lysates as negati v e controls.IC: Internal Control.( B ) Quantification of relati v e TRAP acti vity in (A).Values are normalized to WT le v els and show the mean ± standard deviation.p-values are shown and generated from one-way ANOVA analysis followed by Tukey's multiple comparison.( C ) Representati v e images of telomerase recruitment assay.Flag-PO T1b or Myc-PO T1a (blue) were detected by immunostaining with anti-Flag or anti-Myc antibodies.hTR RNA was detected by hybridization with Cy5-hTR cDNA probes (green) and telomeres visualized by hybridization with PNA probe Cy3-OO-(CCCTAA) 3 (r ed).White arrows: TR / Telomer e foci colocalization.Scale bar: 5 m. ( D ) Quantification of telomerase and telomere colocalization in (C).Data show the mean ± 95% CI of the mean of one experiment with 150 nuclei analyzed per cell line.

Figur e 4 .
Figur e 4. Elevated DN A damage detected in la te-genera tion Pot1b −/ − cells.( A ) Representa ti v e image of ␥ -H2AX TIFs in the indicated cell lines.Cells were immunostained with ␥ -H2AX antibody (gr een), telomer e probe Cy3-(CCCTAA) 3 (red) and DAPI to stain nuclei (blue).Scale bar: 5 m. ( B ) Quantification of the indicated cell lines with ≥5 ␥ -H2AX TIFs in (A).Data show the mean ± standard deviation from three independent experiments in which 150 nuclei were analyzed per cell line.P -values are shown and generated from one-way ANOVA analysis followed by Tukey's multiple comparison.( C ) Representati v e images of p-RPA TIFs in the indicated cell lines.Cells were immunostained with p-RPA antibody (gr een), telomer e probe Cy3-(CCCTAA) 3 (r ed) and DAPI to stain nuclei (blue).Scale bar: 5 m. ( D ) Quantification of the indicated cell lines with ≥5 p-RPA TIFs in (C).Data show the mean ± standard deviation from three independent experiments in which 150 nuclei were analyzed per cell line.p-values are shown and generated from one-way ANOVA analysis followed by Tukey's multiple comparison.( E ) Representati v e images of ␥ -H2AX TIFs in G7 Pot1b −/ − cells treated with DMSO or AZ20 150 nM for 24hr.Cells were immunostained with ␥ -H2AX antibody (green), telomere probe Cy3-(CCCTAA) 3 (red) and DAPI to stain nuclei (blue).Scale bar: 5 m. ( F ) Quantification of the indicated cell lines with ≥5 ␥ -H2AX TIFs in (E).Data show the mean ± standard deviation from two independent experiments in which 150 nuclei were analyzed per cell line.P -values are shown and generated from an unpaired t -test.( G ) Representati v e images of G7 Pot1b −/ − cells e xpressing cell-cy cle sensors Geminin or CDT1 and immunostained with ␥ -H2AX antibody and TRF2 antibody.Scale bar: 5 m. ( H ) Quantification of G7 Pot1b −/ − cells expressing the indicated construct in (G) with ≥5 ␥ -H2AX TIFs.Data show the mean ± standar d de viation from three independent experiments in which 150 nuclei were analyzed per cell line.P -values are shown and generated from one-way ANOVA analysis followed by Tukey's multiple comparison.

Figur e 5 .
Figur e 5. PO T1b protects against telomeric DN A damage in la te-genera tion Pot1b −/ − cells.( A ) Representa ti v e images of ␥ -H2AX TIFs in G4 Pot1b −/ − cells expressing the indicated construct.Cells were immunostained with ␥ -H2AX antibody (gr een), telomer e probe Cy3-(CCCTAA) 3 (r ed) and DAPI to stain nuclei (blue).Scale bar: 5 m. ( B ) Quantification of nuclei with ≥ 5 ␥ -H2AX TIFs from (A).Data show the mean ± standard deviation from two independent experiments in which 150 nuclei were analyzed per cell line.P -values are shown and generated from one-way ANOVA analysis followed by Tukey's multiple comparison.( C ) Schematic of POT1a, POT1b, POT1ab and POT1ba chimera constructs.( D ) Representati v e images of ␥ -H2AX TIFs in G7 Pot1b −/ − cells expressing the indicated construct.Cells were immunostained with ␥ -H2AX antibody (gr een), telomer e probe Cy3-(CCCTAA) 3 (r ed) and DAPI to stain nuclei (blue).Scale bar: 5 m. ( E ) Quantification of nuclei with ≥5 ␥ -H2AX TIFs from (D).Data show the mean ± standard deviation from three independent experiments in which 150 nuclei were analyzed per cell line.P -values are shown and generated from one-way ANOVA analysis followed by Tukey's multiple comparison.

Figure 6 .
Figure 6.Increased telomeric damage in G7 Pot1b −/ − cells promotes telomerase recruitment to telomeres.( A ) Representati v e images of telomerase recruitment assay in G7 Pot1b −/ − cells treated with / without 150nM AZ20, and with / without e xpressing v ector / PO T1b / PO T1ba.␥ -H2AX (blue) were detected by immunostaining with anti-␥ H2AX antibodies.hTR RNA was detected by hybridization with Cy5-hTR cDNA probes (green) and telomeres visualized by hybridization with PNA probe Cy3-OO-(CCCTAA) 3 (red).Scale bar: 5 m. ( B ) Quantification of telomerase and telomere colocalization in (A).Data show the mean ± 95% CI of the mean from at least two experiments with 150 nuclei analyzed per cell line.p-values are shown and generated from one-way ANOVA analysis followed by Tukey's multiple comparison.( C ) Quantification of the indicated conditions with ≥ 5 ␥ -H2AX TIFs in (A).Data show the mean ± standard deviation from at least two independent experiments in which 150 nuclei were analyzed per cell line.P-values are shown and generated from one-way ANOVA analysis followed by Tukey's multiple comparison.( D ) Representati v e images of telomerase recruitment assay in G3 Pot1b −/ − cells treated with / without 150nM AZ20, with / without expressing POT1ba.␥ -H2AX were detected by immunostaining with anti-␥ H2AX antibodies (blue).hTR RNA was detected by hybridization with Cy5-hTR cDNA probes (gr een) and telomer es visualized by hybridization with PNA probe Cy3-OO-(CCCTAA) 3 (red).Scale bar: 5 m. ( E ) Quantification of telomerase and telomere colocaliza tion in (D).Da ta show the mean ± 95% CI of the mean with 150 nuclei analyzed per cell line.P-values are shown and generated from one-way ANOVA analysis followed by Tukey's multiple comparison.( F ) Quantification of the indicated conditions with ≥5 ␥ -H2AX TIFs in (D).Data show the mean from 150 nuclei analyzed per cell line.

Figure 7 .
Figure 7. Telomerase activity is r equir ed to promote telomeric damage in la te-genera tion Pot1b −/ − cells.( A ) Representa ti v e images of ␥ -H2AX TIFs in the indicated cell lines.Cells were immunostained with ␥ -H2AX antibody (green), TRF2 antibody (red) and DAPI to stain nuclei (blue).Scale bar: 5 m. ( B ) Quantification of the indicated cell lines with ≥5 ␥ -H2AX TIFs in (A).Data show the mean ± standar d de viation from two independent experiments in which 150 nuclei were analyzed per cell line.P -values are shown and generated from one-way ANOVA analysis followed by Tukey's multiple comparison.( C ) Representati v e images of ␥ -H2AX TIFs in the indicated cell lines expressing the indicated constructs.Cells were immunostained with ␥ -H2AX antibody (green), TRF2 antibody (red) and DAPI to stain nuclei (blue).Scale bar: 5 m. ( D ) Quantification of the indicated cell lines with ≥ 5 ␥ -H2AX TIFs in (C).Data show the mean ± standard deviation from three independent experiments in which 150 nuclei were analyzed per cell line.P -values are shown and generated from one-way ANOVA analysis followed by Tukey's multiple comparison.( E ) Schematic of mTERT showing the K78 and K560 amino acids.