BLM helicase unwinds lagging strand substrates to assemble the ALT telomere damage response

SUMMARY

The Bloom Syndrome helicase (BLM) is critical for Alternative Lengthening of Telomeres (ALT), a homology directed repair (HDR) mediated telomere maintenance mechanism that is prevalent in cancers of mesenchymal origin. The DNA substrates that BLM engages to direct telomere recombination during ALT remain unknown. Here, we determine that BLM helicase acts on lagging strand telomere intermediates that occur specifically in ALT positive cells to assemble a replication-associated DNA damage response. Loss of ATRX was permissive for BLM localization to ALT telomeres in S and G2 commensurate with the appearance of telomere C-strand specific single-stranded DNA. DNA2 nuclease deficiency increased 5’-flap formation in a BLM dependent manner, while telomere C-strand, but not G-strand, nicks promoted ALT. These findings define the seminal events in the ALT DNA damage response, linking aberrant telomeric lagging strand DNA replication with a BLM directed HDR mechanism that sustains telomere length in a subset of human cancers.

eTOC Blurb

Jiang & Zhang et al. found BLM helicase promotes alternative lengthening of telomeres (ALT) by unwinding unligated Okazaki fragments to generate 5’-flaps. These events initiate the DNA damage response that underlies homology-directed repair mediated telomere maintenance.

Graphical Abstract

INTRODUCTION

Alternative Lengthening of Telomeres (ALT) is a telomerase independent mechanism of telomere maintenance that is prevalent in cancers of mesenchymal origin^1–4. ALT utilizes several different homology-directed repair (HDR) pathways to lengthen telomeres^2,5–9. Approximately 10–15% of human cancers exclusively rely on ALT for telomere maintenance¹⁰. Telomeres in ALT dependent tumors and cell lines display characteristic features that differentiate them from telomeres in telomerase positive cells, including the presence of extensive single-stranded DNA with 5’-overhangs, colocalization of DNA damage response proteins, and clustered telomeres that associate with PML proteins in subnuclear bodies (ALT-associated PML bodies, APBs) that undergo HDR- mediated DNA synthesis outside of S-phase^{2,5,9,11–13}. These events contribute to telomere lengthening by HDR.

Replication stress and elevated chromosome instability are present in ALT positive cancers^11,14. p53 mutations are also frequent in ALT cancers and cell lines, likely as a permissive event to tolerate the high degree of overall genomic instability and DNA damage responses at telomeres^11,14. Histone chaperones ATRX and DAXX load histone H3.3 onto telomeres^15,16. Loss of either ATRX or DAXX leads to increased expression of the telomere-associated noncoding RNA TERRA, which forms RNA-DNA hybrid R-loops with telomeres that are thought to promote recombination in ALT positive cells¹⁷. ATRX or DAXX deficiency frequently occurs in ALT positive cancers, and ATRX loss was shown to promote ALT during immortalization of primary fibroblast cultures and following differentiation of ATRX null pluripotent human stem cells^14,18–21. In addition, ATRX loss in combination with telomere damage stimulates telomere clustering and other features of ALT in telomerase positive cells²². Conversely, reintroduction of ATRX suppresses ALT in U2OS cells, reducing telomere clustering, and DNA damage response protein colocalization with telomeres^23,24.

Despite the link between ATRX-DAXX deficiency and ALT, little is understood regarding the basis for telomere selective increases in HDR. Extensive telomere recombination in ALT positive cells occurs spontaneously in the absence of exogenous DNA damaging agents¹¹. BLM helicase has been implicated as a key factor in orchestrating the ALT damage response as it is required for telomere length maintenance and telomere clustering into APBs^25–29. Telomere DNA double-strand breaks (DSBs) rapidly initiate ALT through a break induced replication- like mechanism that entails loading of the polymerase processivity replication clamp PCNA complex and DNA polymerase δ at damaged telomeres to perform unidirectional HDR telomere synthesis over many kilobases^9,30,31. However, telomere DSBs bypass requirements for BLM, suggesting they are unlikely to be the primary source of genomic lesion that attracts the ALT dependent damage response.

Here, we demonstrate that BLM helicase orchestrates the ALT telomere DNA damage response through its helicase dependent genesis of 5’-single-stranded DNA flaps on lagging strand telomeres. The specific association of BLM with ALT telomeres was independent of DSB induction, and instead required loss of ATRX. BLM deficiency eliminated the basal telomere damage response in ALT positive cells commensurate with reductions in 5’-single-stranded DNA on the telomere C-rich strand but not on the G-rich strand. The source of this C-rich single-stranded DNA is 5’-flaps on unligated Okazaki fragments or nicked DNA. These findings implicate BLM helicase activity at lagging strand telomere lesions as the initiating event that orchestrates HDR during ALT.

RESULTS

BLM is required to assemble the ALT telomere DNA damage response

BLM deficiency reduced APBs, non-S phase EdU (5-Ethynyl-2’-deoxyuridine), and RPA2 colocalization with telomeres (Figures S1A–F), and led to telomere shortening in ALT cell lines in long-term cultures (Figure S1G), as expected^{6,8,25,26,29,32}. To investigate how BLM helicase orchestrates telomere HDR during ALT, we purified the telomere-associated proteomes from U2OS cells that expressed CRISPR-Cas9 with sgRNAs targeting BLM (sgBLM) or control (sgCtrl). Telomere proteomes were obtained from three biological replicates using the Proteomics of Isolated Chromatin Segments (PICh) method³³ (Figure 1A). Consistent results were observed in each experiment, demonstrating a nearly complete elimination of the BTR (BLM-Topo IIIα-RMI1-RMI2) complex, ATR, Fanconi Anemia (FA), and HDR damage response pathway proteins at U2OS telomeres in BLM deficient cells (Figures 1B and 1C). Approximately 67% reductions in peptide numbers occurred for all three subunits of the trimeric single-stranded DNA binding protein, RPA1-RPA2-RPA3, in agreement with prior reports that BLM deficiency reduces single-stranded DNA at ALT telomeres^25–27. Notably Shelterin complex components remained at similar levels in sgCtrl and sgBLM U2OS telomeres (Figure 1C), demonstrating intact telomere protection in the absence of BLM.

Figure 1. BLM helicase is required for assembly of the ALT telomere DNA damage response.

(A) Schematic of Proteomics of Isolated Chromatin segments (PICh) to define the telomere proteome in control (sgRosa) and BLM knockout U2OS cells. Fractionated and pre-cleared chromatin was hybridized to a biotinylated telomeric probe and then captured on magnetic beads. Telomere-associated proteins were analyzed by Western blot and/or mass spectrometry. (B) Scatter plot comparing the telomere-associated proteomes of control and BLM knockout U2OS cells. The x-axis shows the Log₂FC (Fold change) (1+ total peptide number of mass spectrometry of sgBLM samples)/(1+ total peptide number of mass spectrometry of sgRosa samples). The y-axis shows the −Log₁₀(p-value) from three independent experiments. (C) Tabulated differences in telomere-associated DNA repair protein in control and BLM knockout U2OS from three independent experiments. Shelterin components are included to verify similar amounts of telomeric chromatin was captured in each replicate. (D) Venn diagram comparing the proteins with Log₂FC(U2OS/HeLaS3) > 1 from Zhang. et al.³⁴ or Log₂FC(sgBLM/sgRosa) < −1 from the PICh results. The common proteins were selected for the STRING analysis. (E) STRING analysis of the common proteins in (D). The lines between the circles indicate protein interaction, or the proteins are involved in the same complex or the same pathway.

Telomeres in ALT positive cells exhibit a robust baseline replication stress-associated DNA damage response^11,13,34. Indeed, plotting of recently reported telomere proteomes reveals enrichment of ATR, FA, and HDR repair factors at ALT positive U2OS telomeres in comparison to telomere-associated proteins in telomerase positive HeLa S3 cells (Figure S2A)³⁴. Interestingly, BLM deficiency converted the telomere damage response proteomes in ALT positive U2OS cells to resemble that of telomeres in undamaged telomerase positive HeLa S3 cells (Figure S2B). U2OS telomere proteomes were significantly enriched for DSB repair, replication fork processing, and other factors associated with HDR compared to either HeLa S3 or BLM deficient U2OS telomeres (Figures 1D, 1E and S2). STRING analyses reveal BLM dependent networks of factors that recognize and resolve replication-associated DNA damage by HDR (Figure 1E). These findings are consistent with a basal, BLM dependent HDR network at telomeres in ALT positive cells.

ATRX deficiency promotes BLM recognition of ALT telomeres in S/G2

Analysis of reported PICh derived telomere proteomes revealed BLM specific association with ALT telomeres (Figure S2A)^33,34. Notably, BLM remained absent from telomerase positive telomeres after DSB induction and did not increase at ALT telomeres, indicating that DSBs are not the lesion that attracts BLM to direct ALT³⁴. We confirmed that BLM telomere localization is not stimulated by DSB by PICh coupled with Western blot in U2OS cells following induction of TRF1-FokI nuclease or its nuclease inactive D450A mutant (Figure 2A). We reasoned that ATRX deficiency could account for the specific association between BLM and telomeres in ALT positive cells. In agreement, Doxycycline inducible ATRX expression in U2OS cells diminished BLM localization to ALT telomeres along with reductions in telomere clustering into APBs, RPA2 colocalization, and single-stranded DNA generation assessed by telomere native FISH (also known as ALT-FISH³⁵ (Figures 2B, 2C and S3A–C). Moreover, ATRX depletion by CRISPR/Cas9 in HeLa S3 cells resulted in detectable BLM colocalization with telomeres and increased RPA2 association (Figures 2D, 2E and S3D).

Figure 2. ATRX suppresses BLM recognition of ALT telomeres.

(A) PICh followed by Western blot to detect BLM localization at telomeres in U2OS cells following TRF1-FokI (D450A and WT) induction for 2 hrs. (B, C) Representative IF-FISH images and quantifications of BLM- (B) and RPA2- (C) telomere localization by IF-FISH, in the U2OS cells with/ without adding 400ng/ml doxycycline for 7 days to induce ATRX expression, respectively. The data represent the median (line in the box), 1st to 3rd quartile (box), and the 75th percentile plus 1.5 times interquartile range (IQR) by the top line. The data were from 8 and 6 independent experiments, respectively. (D, E) Representative IF-FISH images and quantifications of BLM- (D) and RPA2- (E) telomere localization in control or ATRX knockout HeLa S3 cells. The colocalization events were indicated by the arrowheads. The data represent the mean and the standard error of the mean (SEM) from 4 independent experiments. The ATRX knockout HeLa S3 cells were generated by a dual CRISPR/Cas9 strategy, which is shown in Figure S3D. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 (Unpaired two-tailed student’s t-test).

Cell cycle dependency of BLM-telomere colocalization was examined in interphase cells using cyclin A positivity as a marker of S/G2 populations (Figure 3A). BLM localization occurred primarily in cyclin A positive S/G2 cells in contrast to cyclin A negative G1 cells with consistent results across five different ALT positive cell lines. This corresponded to the appearance of C-rich single-stranded DNA at telomeres as assessed by native-FISH with Cy3-labeled telomeric G-probe (Figure 3B).

Figure 3. Requirements for BLM telomere recognition and ALT activity.

(A, B) BLM colocalization with telomeres (A) and the generation of C-rich single-stranded DNA (ssTeloC) at telomeres (B) was monitored in cyclin A positive (S/G2) and negative (G1) interphase cells. The data represent the median (line in the box), 1st to 3rd quartile (box), and the 75th percentile plus 1.5 times IQR (Top line). The data were from 4 independent experiments. (C) Schematic of BLM mutants used to determine requirements for BLM telomere recognition and ability to stimulate ALT. (D, E) Colocalization of GFP-tagged BLM (D) or RPA2 (E) and telomeres was quantified in cells expressing WT BLM or indicated mutants. (F) Cells that have at least 2 GFP-BLM positive telomere foci were considered as BLM-Telo colocalization positive cells. RPA2 colocalization with the GFP-BLM positive telomere foci was counted and cells with at least two colocalization events were considered as RPA2-BLM-Telo positive cells. The data in Tukey box plot represent the median (line in the box), 1st to 3rd quartile (box), and the 75th percentile plus 1.5 times IQR (Top line). The data in scatter plot with bar represent the mean and SEM. The data were from 5 independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 (Unpaired two-tailed student’s t-test).

The BLM helicase has several functions in DNA repair and replication that rely on distinct protein-protein interactions. The amino-terminal 133 amino acids of BLM mediate interaction with RMI1, RMI2, and Topo IIIα as the BTR complex, to execute branch migration dependent dissolution of Holliday junction recombination intermediates. This interaction can be disrupted by the mutation of 3 amino acids (lysines 38, 39, and 40 to alanine, “K3A”)³⁶. The BLM N-terminus (amino acid 1 to 641) also interacts with the RPA complex as a critical part of its function in restarting stalled replication forks³⁷. Enzymatic activity is encoded within the carboxy-terminal half and eliminated by mutation of K695 ATPase function^38,39.

We mapped the domains in BLM required for telomere localization and ALT using the indicated panel of missense and deletion mutants (Figure 3C and S4A). ATPase inactive BLM K695R mutation severely reduced its telomere localization as well as the generation of single-stranded DNA as assessed by RPA signal at telomeres. Disruption of interaction with RMI1/2 and Topo IIIα in the BTR complex also reduced BLM telomere localization and RPA association with telomeres. Depletion of the N-terminus led to a further decrease of BLM localization at telomere, indicating additional regulators of BLM telomere localization (Figures 3C–E and S4B). Tethering the BLM mutants to telomeres by fusing BLM to shelterin component TRF1 restored RPA accumulation at telomeres, except for the K695R ATPase inactive mutant (Figures S4C and S4D). This indicates the ATPase activity of BLM but not assembly into the BTR complex is essential for the generation of single-stranded telomere DNA. Indeed, quantification of RPA2-telomere colocalization at GFP-BLM positive foci also showed robust RPA2 association at telomeres in the BLM mutants (Figure 3F, K3A and Δ1–133) but not the ATPase mutant. Deletion of the C-terminus removed the nuclear localization signal of BLM and led to cytoplasmic accumulation of GFP-BLMΔC and ΔHRDC+C. Despite a modest reduction of BLM-telomere localization (Figures 3C and 3D), these mutants still showed telomere colocalization possibly because association with other BLM-associated proteins such as the BTR complex can facilitate its nuclear entry²⁶. Interestingly, deletion of the HRDC domain significantly enhanced the RPA2 association at telomere (Figures 3E and 3F). The HRDC domain is required for BLM-dependent double Holliday junction (dHJ) dissolvase activity⁴⁰. However, HRDC deletion increases BLM ATPase activity in vitro^41,42, suggesting the ATPase activity rather than the dissolution function of BLM is essential for the generation of single-stranded DNA at ALT telomere.

BLM helicase increases telomere C-rich single-stranded DNA with 5’ overhangs

We investigated the impact of BLM deficiency on telomere single-stranded DNA using a combination of telomere native FISH³⁵ and Southern blot following one-dimensional (1D) and neutral-neutral two-dimensional (2D) gel electrophoresis. CRISPR deletion of BLM yielded reductions in both G- and C- rich telomere ssDNA by telomere native FISH (Figures 4A, 4B, S5A and S5B). We next probed nondenaturing/native Southern blots with ³²P-labeled oligonucleotides that were complementary to the telomere G-rich or C-rich strands. BLM deficiency resulted in a striking decrease in C-rich single-stranded DNA at telomeres that was detected using a ³²P-labeled telomeric G-rich probe (Figures 4C and S5C). No reduction in the G-rich single-stranded DNA was observed on the native telomere Southern blots in LM216J or U2OS cells (Figures 4D and S5C). The C-rich telomere single-strand DNA signal from BLM knockout LM216J cells was also nearly absent in 2D telomere Southern blots, with a notable reduction in the arc (dsDNA region) and in the T-complex region that represents telomere recombination intermediates in ALT positive cells^34,43,44 (Figures 4E and 4F). In contrast, G-rich single-stranded DNA remained present at similar levels in the arc region of BLM depleted LM216J cells but was reduced at the T-complex region (Figures 4E and 4F). This indicates that BLM primarily affects the presence of C-rich single-stranded DNA in unrecombined telomeres, whereas the reductions in G-rich single-stranded DNA (Figures 4A–F and S5A–C) occur due to the loss of recombination in BLM deficient cells. In agreement, G-rich single-stranded DNA was diminished only at the top of the loading well on 1D Southern blots (Figure 4D), which likely represents the complex intermediates that migrate slowly into the gel. We next incubated genomic DNA from control or BLM deficient LM216J cells with either E. coli ExoI (3’-exonuclease) or with RecJf (5’-exonuclease). RecJf markedly reduced C-rich single-stranded DNA in control cells whereas E. coli ExoI had a minimal effect (Figure 4G and S5D). This indicates that BLM is required for the formation of 5’-single-stranded telomere overhangs, which are characteristic of ALT cells^12,45.

Figure 4. BLM is required for the generation of 5’-C-rich single-stranded DNA at ALT telomeres.

(A, B) Representative telomere native FISH images and quantification of telomere native FISH in control (sgRosa) and BLM knockout LM216J cells. Single-stranded telomeric C-rich (A, ssTeloC) or G-rich (B, ssTeloG) DNA was recognized by Cy3-conjugated telomeric G-probe and C-probe, respectively. Two sgRNAs (#6, #7) were used separately to knockout BLM. The data represent the mean and SEM from 4 independent experiments. (C, D) Genomic DNA purified from control and BLM knockout LM216J cells were treated with HinfI and RsaI and subjected to Southern blot under native and denatured conditions. The data in scatter plot with bar represent the mean and SEM of the relative intensity of the native signal compared to the denatured signal from 6 independent experiments. (E) Schematic showing telomeric linear double-stranded DNA, circular DNA, and recombination intermediates in a 2D agarose gel. (F) 2D agarose gel electrophoresis of purified genomic DNA treated with HinfI and RsaI in control and BLM knockout LM216J cells that separates recombination intermediates with linear DNA. (G) Southern blot of purified genomic DNA treated with HinfI and RsaI are subjected to nuclease digestion using either E. coli Exonuclease I (ExoI) (3’-5’ exonuclease) or RecJf (5’-3’ exonuclease) at 37 °C for 2 hrs. For all the Southern blot, either ³²P-labeled telomeric G-probe or C-probe was used as indicated. **p < 0.01, ***p < 0.001, ****p < 0.0001 (Unpaired two-tailed student’s t-test).

The specific requirement of BLM for the C-rich strand suggests an asymmetry in how BLM generates single-stranded DNA. Notably, 5’-flap DNA is generated on C-rich strand Okazaki fragments during lagging strand synthesis due to unidirectional telomere replication that originates in subtelomeric DNA⁴⁶. Flap removal involves 5’-flap endonucleases FEN1 and DNA2⁴⁷. We had previously observed increased telomere single-stranded C-rich DNA in DNA2 depleted U2OS cells³⁴. To determine if BLM could act on these unprocessed substrates, we deleted DNA2 using three independent CRISPR sgRNAs in ALT positive LM216J cells (Figure 5A). In control cells, DNA2 loss strongly increased telomere ssDNA as assessed by RPA2 localization and telomere clustering (Figures 5B and 5C). Notably, BLM depletion compromised the RPA2 localization and telomere clustering that are induced by DNA2 loss (Figures 5B and 5C). Accordingly, BLM deficiency prevented the increased ALT-associated telomere synthesis in DNA2-depleted LM216J cells (Figure 5D), suggesting that BLM acts upstream of DNA2 during ALT.

Figure 5. BLM helicase activity creates 5’-flap DNA substrates for DNA2 during ALT.

(A) Western blot verifying depletion of BLM and DNA2 protein by CRISPR/Cas9 in LM216J cells. (B, C) Representative IF-FISH images and quantification of RPA2 at telomere upon DNA2 knockout in WT and BLM knockout LM216J cells. The data represent the mean and SEM from 3 independent experiments. (D) Quantification of non-S phase EdU incorporation at telomere upon DNA2 knockout in WT and BLM knockout LM216J cells. The data represent the mean and SEM from 4 independent experiments. (E) Southern blot of HinfI and RsaI treated genomic DNA from control and BLM knockout LM216J cells with or without treatment of 2.5 nM recombinant WT or nuclease-dead (D277A) DNA2 at 37 °C for 2 hrs. (F) HinfI and RsaI treated genomic DNA from indicated U2OS cell lines were further incubated with recombinant BLM or helicase inactivated (K695R) BLM and DNA2, then subjected to Southern Blot with either ³²P-labeled telomeric G-probe (Left) or C-probe (Right). **p < 0.01, ***p < 0.001, ****p < 0.0001 (Unpaired two-tailed student’s t-test).

We next determined if recombinant DNA2 protein could specifically reduce 5’-C-rich telomere ssDNA on native Southern blots. Genomic DNA was isolated from control or BLM-depleted LM216J cells and incubated with recombinant DNA2 following restriction enzymes digestion (Figure 5E). This treatment with recombinant DNA2, but not its nuclease inactive mutant (D277A), reduced telomeric C-rich ssDNA in the control cell, while left telomeric G-strand DNA levels unaffected. Telomere C-rich ssDNA in sgBLM-expressing cells was strongly reduced at baseline and showed minimal further reductions with DNA2 incubation (Figure 5E). We then determined if BLM helicase activity could generate telomeric 5’-C-rich single-stranded flap substrates for recombinant DNA2 on isolated genomic DNA. Purified genomic DNA was incubated with recombinant wild-type (WT) BLM or the BLM K695R ATPase/helicase inactive mutant following restriction enzymes digestion (Figures 5F and S6), and telomere single-stranded DNA was detected by native Southern blot. Recombinant WT BLM, but not the helicase inactive mutant, increased C-rich single-stranded telomere DNA isolated from either BLM proficient or deficient LM216J cells but not the telomerase positive HeLa S3 cells (Figures 5F and S6). Notably, incubation with recombinant DNA2 reduced the BLM helicase dependent C-rich single-stranded DNA signal but did not affect the G-rich telomere ssDNA signal (Figure 5F). Collectively, this reveals a hierarchical relationship of BLM-dependent generation of a 5’-flap substrate for cleavage by DNA2. The balance of these activities regulates the level of C-rich single-stranded DNA and HDR at ALT telomeres.

BLM acts specifically at nicks in the telomeric C-rich strand to induce ALT

Telomere DSBs induced by either TRF1-FokI or Cas9 elicit a process termed Break-Induced Telomere Synthesis that recapitulates many of the features of ALT^9,30,31,48. However, telomere DSB induction bypasses requirements for BLM, suggesting that it is not the endogenous substrate for BLM directed HDR (Figure 2A). Given the BLM dependent increases in C-rich single-stranded DNA at lagging strand telomeres, we sought to understand if BLM actions at these lesions could promote ALT. In principle, BLM would act differentially at C- vs. G-rich telomere single-stranded nicks to affect telomere HDR and ALT. We designed sgRNAs that would be used in combination with the Cas9 D10 and H480 mutations that introduce single-strand nicks rather than DSBs in U2OS cells (Figures 6A and 6B). C-rich strand specific nicks produced by a sgRNA targeting telomeres and the D10A nickase increased BLM at telomeres, while G-rich strand specific nicks by H840A or WT Cas9-induced DSBs yielded similar levels of BLM association as the dCas9 nuclease inactive control and the scramble sgRNA (Figures 6C and 6D). Moreover, D10A nickase strongly stimulated telomeric C-rich single-stranded DNA, telomere lengthening and telomere length heterogeneity in contrast to other CRISPR/Cas9 constructions (Figures 6E and 6F).

Figure 6. Telomeric C-strand nicks initiate BLM telomere recognition and ALT.

(A) Schematic of CRISPR/Cas9 nickases. sgTelo is used to target the telomere C-rich sequence. In this context, WT Cas9 generates double strand breaks; Cas9 D10A mutant cleaves “targeting strand” (C-strand); H840A mutant cleaves “non-targeting strand” (G-strand); while dCas9 (catalytically dead Cas9) does not cleave DNA. (B) Western blot of U2OS cells expressing indicated sgRNA and Cas9 protein. (C, D) Representative IF-FISH images (C) and quantification (D) of BLM-telomere colocalization in U2OS cells expressing the indicated sgRNAs and Cas9 proteins. The data represent the mean and SEM from 4 independent experiments. (E) Genomic DNA purified from cells expressing indicated sgRNA and Cas9 were digested with HinfI and RsaI and subjected to Southern blot under native and denatured conditions with ³²P-labeled telomeric G-probe. The data in scatter plot with bar represent the mean and SEM from 5 independent experiments. (F) TRF assay showing telomere length from cells expressing indicated sgRNA and Cas9 protein with ³²P-labeled telomeric C-probe. (G, H) Representative IF-FISH images (G) and quantification (H) of ADP-ribose-telomere colocalization in U2OS cells expressing indicated sgRNA and Cas9 protein. The data represent the mean and SEM from 4 independent experiments. SCR, scrambled sgRNA. dCas9, catalytically dead Cas9. ns: p > 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 (Unpaired two-tailed student’s t-test).

PARP1 and PARP2-dependent PARylation arises at single- and double-stranded DNA breaks. It also primarily occurs at unligated Okazaki fragments in undamaged cells^49,50. Interestingly, PARylation on C-strand specific nicks was increased in comparison to G-strand nicks (Figures 6G and 6H), raising the possibility that processing of the Cas9 nickase induced single-strand breaks on the C-strand accounts for this difference. To determine if these observations were representative of BLM activity at lagging strand telomeres, we monitored PARylation at telomeres and its colocalization with PCNA (Figures 7A–C). This revealed that telomere-associated PARylation during replication also required BLM activity (Figures 7A–C), while conversely, DNA2 loss dramatically increased telomere-associated PARylation during replication and was suppressed by BLM knockout in LM216J cells (Figures 7D and 7E). These findings support a model that lagging strand replication-associated nicks on telomere C-rich DNA represent the substrates that BLM acts on to assemble the ALT DNA damage response (Figure 7F).

Figure 7. BLM promotes a replication-associated ALT telomere damage response.

(A-C) Representative IF-FISH images of PCNA and ADP-ribose-telomere colocalization (A), quantification of ADP-ribose levels at telomere (B), and quantification of PCNA colocalization with ADP-ribose-telomere (C) in WT and BLM knockout LM216J cells. The data represent the mean and SEM from 4 independent experiments. (D, E) Representative IF-FISH images (D) and quantification (E) of ADP-ribose levels at telomeres upon DNA2 and BLM double knockout in LM216J cells. The data represent the median (line in the box), 1st to 3rd quartile (box), and the 75th percentile plus 1.5 times IQR (Top line) from 4 independent experiments. (F) Model. Inefficient Okazaki fragment maturation at lagging strand ALT telomeres provides DNA nicks/gaps as substrates for BLM recognition. BLM helicase activity on these lagging strand substrates generates telomeric C-rich strand 5’-flaps during strand displacement DNA synthesis by the replisome (indicated by Polδ-PCNA). The telomeric C-rich 5’-flaps induce replication-associated DNA damage responses to promote ALT-associated HDR. **p < 0.01, ***p < 0.001, ****p < 0.0001 (Unpaired two-tailed student’s t-test).

DISCUSSION

An extensive telomere-associated DNA damage response in unperturbed ALT-positive cells suggest the presence of endogenous lesions that elicit HDR-dependent telomere maintenance. ATRX deficiency and BLM helicase activity were known drivers of ALT. However, the nature of the DNA substrate that BLM acts on to direct ALT and its dependency on ATRX loss was undefined. Additionally, it was unclear how BLM helicase activity directs the necessary assembly of a damage response to promote ALT. Our findings provide details underlying many of these knowledge gaps. BLM recognizes single-stranded nicks on the C-rich telomere strand and acts on these lesions to promote 5’-flap formation. These substrates are increased at telomeres in ATRX deficient cells during S-phase. These findings support a model that BLM helicase acts on nicks/gaps in the C-rich telomere strand to generate 5’- single strand flaps that seed an extensive replication stress-associated damage response during ALT.

How ATRX deficiency leads to the genesis of this BLM specific telomere substrate remains unclear and an important area for future investigation given its causal role in ALT. ATRX binds to G-rich DNA to suppress G4-quadruplex structures⁵¹. ATRX deficiency may impede DNA replication, causing aberrant Okazaki fragment maturation and nicks/gaps in the C-strand that attract BLM. It is also unclear how the ensuing BLM-dependent damage response mediates ALT. BLM plays several important roles in telomere integrity in both ALT and telomerase positive cells. It escorts telomere replication by suppressing G4 structures at both the leading strand and the template of the lagging strand^52,53. An interesting distinction is the ALT specific phenomenon of BLM dependent elevations in C-rich 5’-flaps. ALT represents a gain of function, non-canonical HDR mechanism that substantially differs from classical homologous recombination. BLM appears to promote a DNA damage response that is dominated by replication-associated factors that recognize single-stranded DNA, including ATR and FA proteins (Figures 1 and S2). Such factors mediate DNA strand annealing type HDR when single-stranded homology donors are employed in CRISPR/Cas9 experimental settings. Type II survivors in yeast are also thought to use annealing-mediated recombination for HDR synthesis⁵⁴. We speculate similar mechanisms exist in human cells that utilize ALT following a BLM-directed single-stranded DNA damage response.

Our findings provide a framework to systematically address the question of BLM-dependent telomere recombination during ALT. BLM ATPase and helicase activities were required for ALT as was its amino-terminus, which was necessary for telomere recognition. Interestingly, interaction with the BTR complex was dispensable as was its HRDC domain, suggesting that dHJ dissolution is not a critical BLM function in generating telomere single-stranded DNA. Instead, BLM uses its DNA motor activities to generate substrates for DNA2 endonuclease activity, which in turn limits ALT recombination. In addition to BLM helicase activity, SUMOylation may also play an important role in directing ALT. Targeting SUMO interacting motifs (SIM domains) to telomeres through a TRF1-SIM interaction enacted telomere clustering specifically in ALT cells independent of the DNA damage response⁵⁵. BLM may play a critical role in ALT through a combination of its helicase activity and via its known modification by SUMO^25–27.

Aberrations in lagging strand synthesis activate PARP enzymes through recognition of unligated Okazaki fragments and are responsible for the majority of chromatin associated PARylation in replicating cells^49,50. PARP inhibition is proposed to trap PARP1 and PARP2 at such endogenous lesions as the underlying basis for synthetic lethality in homologous recombination deficient cells. Interestingly, PARylation is abundant at telomeres in ALT positive cells and PARP inhibition has been reported to increase ALT activity⁵⁶. ALT positive cells are also suggested to have aberrations in lagging strand telomere synthesis⁵⁷. Our findings link inefficiencies in Okazaki fragment maturation to BLM-dependent ALT telomere recombination and implicate opposing functions of BLM helicase and DNA2 endonuclease activities on their steady-state levels.

Data mining on the Cancer Dependency Map (DepMap) Portal revealed that ALT cells are hypersensitive to replication stress and highlighted 14 genes that are preferentially essential in ALT cell lines⁵⁸. Eleven out of the fourteen genes relied on BLM for localization at ALT telomeres (Figure S7). The BLM dependency of this response may indicate hierarchical events that determine the survival of ALT cells. FANCM deficiency is lethal specifically in ALT positive cells^58–60. This occurs concomitantly with an increase in ALT activity and the generation of excessive extrachromosomal telomeres that are thought to arise from nucleolytic cleavage of telomere recombination intermediates. BLM deficiency restores viability upon FANCM loss⁵⁹ in ALT positive cells, predicting that BLM loss would be a resistance mechanism to inhibitors of FANCM. We observed complete BLM dependency for FANCM or SMARCAL1 telomere association in PICh experiments (Figures 1B, 1C and S7). A parsimonious explanation is that ATRX deficiency increases the abundance of ALT telomere-specific lesions that attract a BLM helicase-dependent DNA damage response that executes ALT. We propose that FANCM and SMARCAL1 are required to prevent the accumulation of complex BLM-dependent recombination intermediates that become toxic in their absence. Deficiency in either FANCM or SMARCAL1 would lead to compensatory nucleolytic resolution of unresolved telomere recombination intermediates and cell death due to rampant genomic instability. How these BLM-dependent intermediates arise and what constitutes their unique dependencies on FANCM and SMARCAL1 will be an important question for future study.

Limitations of the Study

Our findings reveal that BLM helicase acts on unligated Okazaki fragment intermediates during lagging strand telomere synthesis to generate long 5’-flaps that seeds the ALT telomere DNA damage response. ATRX deficiency, a hallmark of ALT reliant cells, stimulated BLM recruitment to telomeres. A logical extension of these findings is that ATRX loss impairs Okazaki fragment maturation at telomeres. However, we have not formally tested this possibility. Therefore, how ATRX deficiency affects BLM telomere recruitment to telomeres is a subject for future investigation. We also propose that BLM generates complex recombination intermediates by producing C-rich 5’-flaps, which in turn initiate a DNA damage response. Indeed, BLM is required for the formation of telomere recombination intermediates and for telomere association. We do not currently know the structure of these putative BLM dependent intermediates. A deeper understanding of the structure of BLM-dependent recombination intermediates will be necessary to definitively test the proposed models.

STAR★Methods

RESOURCE AVAILABILITY

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Roger A. Greenberg (rogergr@pennmedicine.upenn.edu).

Materials availability

All cell lines and DNA constructs generated in this study will be made available upon request.

Data and code availability

• PICh, Western blot, Southern blot, and immunofluorescence quantification raw data have been deposited at Mendeley and are publicly available as of the date of publication. The DOI is listed in the key resources table.

• This paper does not report original code.

• Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
BLM	Bethyl Laboratories	Cat No. A300-110A-M; RRID:AB_2779016
Poly/Mono-ADP Ribose	Cell Signaling Technology	Cat No. 83732; RRID:AB_2749858
PCNA	Cell Signaling Technology	Cat No. 2586S; RRID:AB_2160343
Anti-phospho-Histone H2A.X (Ser139) Antibody	EMD Millipore	Cat No. 05-636-I; RRID:AB_2755003
PML	Santa Cruz Biotechnology	Cat No. sc-966; RRID:AB_628162
GFP	Thermo Fisher Scientific	Cat No. A-11122; RRID:AB_221569
Cyclin A	Santa Cruz Biotechnology	Cat No. sc-271682; RRID:AB_10709300
GAPDH	Cell Signaling Technology	Cat No. 2118S; RRID:AB_561053
mCherry	Abcam	Cat No. ab183628; RRID:AB_2650480
DNA2	Abcam	Cat No. ab96488; RRID:AB_10677769
RPA2 (Anti-RPA, clone RPA34-20)	EMD Millipore	Cat No. MABE285; RRID:AB_11213221
α-Tubulin	Developmental Studies Hybridoma Bank	Cat No. 12G10; RRID:AB_528498
KU86 (H-300)	Santa Cruz Biotechnology	Cat No. sc-9034; RRID:AB_2218743
β-Actin	Santa Cruz Biotechnology	Cat No. sc-47778; RRID:AB_626632
TRF1	GeneTex	Cat No. GTX70304; RRID:AB_377594
ATRX	Santa Cruz Biotechnology	Cat No. sc-15408; RRID:AB_2061023
ATRX	Santa Cruz Biotechnology	Cat No. sc-55584; RRID:AB_831012
Cas9	Epigentek	Cat No. A9000-050; RRID:AB_2828022
Bacterial and Virus Strains
DH10Bac	Invitrogen	10361012
Chemicals, Peptides, and Recombinant Proteins
Puromycin	Thermo Fisher	A1113803
Hygromycin	Invivogen	ant-hg-1
G418 Sulfate	Thermo Fisher Scientific	11811031
Blasticidin	Invivogen	ant-bl-1
Polyethylenimine	Polysciences	23966
polybrene	Sigma-Aldrich	H9268
RO-3306	Selleck Chemicals	S7747
Q5 Site-Directed Mutagenesis Kit	New England Biolabs	E0554S
NEBuilder® HiFi DNA Assembly Master Mix	New England Biolabs	E2621S
Avalanche®-Omni Transfection Reagent	EZ Biosystems	EZTOMNI-1
Formaldehyde solution	Sigma-Aldrich	252549
RNAseA	Qiagen	19101
RNase H	New England Biolabs	M0297L
RsaI	New England Biolabs	R0167L
HinfI	New England Biolabs	R0155S
BsmBI-v2	New England Biolabs	R0739S
Streptavidin Agarose	EMD Millipore	69203-3
Sephacryl S-400 HR column	GE Healthcare Life Sciences	17060901
Dynabeads™ MyOne™ Streptavidin C1	Thermo Fisher Scientific	65002
D-biotin	Invitrogen	B20656
poly-L-Lysine	Sigma-Aldrich	P4832
PDD00017273	Tocris	59-521-0
Click-iT EdU Cell Proliferation Kit	Thermo Scientific	C10337
DAPI	Sigma-Aldrich	D9542
ProLong Gold Antifade Mountant	Thermo Scientific	P36934
cOmplete protein inhibitor cocktail	Roche	11873580001
Benzonase	EMD Millipore	70746
Protein Assay Dye Reagent	Bio-Rad	5000006
Western Lightning Plus-ECL	Perkins Elmer	NEL105001EA
Restore Western Blot Stripping Buffer	Thermo Fisher Scientific	21059
DNeasy Blood & Tissue Kit	Qiagen	69506
Qubit™ dsDNA Quantification Assay Kits	Thermo Fisher Scientific	Q32851
Pulse field certified agarose	BioRad	1620137
Wizard Genomic DNA Purification Kit	Promega	A1120
protease K	Qiagen	19133
TelC-Cy3	PNA Bio	F1002
TelG-Cy3	PNA Bio	F1006
Deposited Data
PICh of TRF1-Fork1D450A expressing U2OS vs HelaS3	Zhang et al.34	N/A
PICh of sgRosa vs sgBLM U2OS	This Paper	DOI: 10.17632/hbm2k47rnx.1
Uncropped Images	This Paper	DOI: 10.17632/hbm2k47rnx.1
Immunofluorescence quantification raw data	This Paper	DOI: 10.17632/hbm2k47rnx.1
Experimental Models: Cell Lines
U2OS	ATCC	HTB-96
LM216J	Dilley et al.9	N/A
GM847	Dilley et al.9	N/A
WI-38 VA13	ATCC	CCL-75.1
HelaS3	ATCC	CCL-2.2
HEK293T	ATCC	CRL-3216
Oligonucleotides
sgRosa sequence, see Method details	Zhang et al.34	N/A
sgBLM sequence, see Method details	This paper	N/A
sgDNA2 sequence, see Method details	Zhang et al.34	N/A
sgATRX sequence, see Method details	This paper	N/A
sgSCR sequence, see Method details	Zhang et al.45	N/A
sgTelo sequence, see Method details	Zhang et al.45	N/A
Primers for GFP-BLM cloning, see Method details	This paper	N/A
Primers for BLM-mCherry-TRF1 cloning, see Method details	This paper	N/A
Recombinant DNA
pLenti CMV TRE3G Puro FLAG-DD-ER-mCherry-TRF1-FokI WT	Dilley et al.9	N/A
pLenti CMV TRE3G Puro FLAG-DD-ER-mCherry-TRF1-FokI D450A	Dilley et al.9	N/A
GFP-BLM	Addgene	80070
GFP-BLM-EV	This paper	N/A
GFP-BLM-K3A	This paper	N/A
GFP-BLM-Δ133	This paper	N/A
GFP-BLM-ΔN	This paper	N/A
GFP-BLM-ΔC	This paper	N/A
GFP-BLM-ΔHRDC+C	This paper	N/A
plenti-CMV-BSD	Addgene	17486
plenti-CMV-BLM-mCherry-TRF1-BSD	This paper	N/A
plenti-CMV-BLMK695R-mCherry-TRF1-BSD	This paper	N/A
plenti-CMV-BLMK3A-mCherry-TRF1-BSD	This paper	N/A
plenti-CMV-BLMΔ133-mCherry-TRF1-BSD	This paper	N/A
plenti-CMV-BLMΔN-mCherry-TRF1-BSD	This paper	N/A
lenti-SpCas9-hygro	Addgene	104995
lentiGuide-Puro	Addgene	52963
lentiGuide-sgRosa-Puro	This paper	N/A
lentiGuide-sgBLM#6-Puro	This paper	N/A
lentiGuide-sgBLM#7-Puro	This paper	N/A
lentiGuide-GFP	Addgene	104374
lentiGuide-sgRosa-GFP	This paper	N/A
lentiGuide-sgBLM#6-GFP	This paper	N/A
lentiGuide-sgBLM#7-GFP	This paper	N/A
lentiCRISPRv2-blast	Addgene	98293
lentiCRISPRv2-sgRosa-blast	This paper	N/A
lentiCRISPRv2-sgDNA2#7-blast	This paper	N/A
lentiCRISPRv2-sgDNA2#8-blast	This paper	N/A
lentiCRISPRv2-sgDNA2#9-blast	This paper	N/A
lentiSaCas9-sgRosa-neo	This paper	N/A
lentiSaCas9-sgATRX#2-neo	This paper	N/A
lentiCRISPRV2-sgRosa-hygro	This paper	N/A
lentiCRISPRV2-sgATRX#1-hygro	This paper	N/A
lentiCRISPRV2-sgATRX#11-hygro	This paper	N/A
pMD2.G	Addgene	12259
psPAX2	Addgene	12260
Software and Algorithms
Image J	NIH	https://imagej.nih.gov/ij/
Graphpad Prism 9	GraphPad software Inc.	https://www.graphpad.com/
Nikon NIS-Elements	Nikon	https://www.microscope.healthcare.nikon.com/products/software/nis-elements
EPSON Scan	EPSON	https://www.epson.com/usa
Cell Profiler	Broad Institute	https://cellprofiler.org/
Snapgene	Snapgene	https://www.snapgene.com/

EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS

Cell lines

U2OS, HeLa S3, HEK 293T cell lines were grown in DMEM (Thermo Fisher) with 10% calf serum (Thermo Fisher) and 1% Penicillin/ Streptomycin (Thermo Fisher). VA13, GM847, LM216J, IIICF E6/E7 cell lines were grown in DMEM (Thermo Fisher) with 10% FBS (Bio-Techne) and 1% Penicillin/Streptomycin (Thermo Fisher). Cell lines were cultured in an incubator at 37°C and 5% CO₂. Mycoplasma contaminations were tested to be negative using MycoAlert Plus Mycoplasma Detection Kit (Lonza).

METHOD DETAILS

Cloning and transient transfection

The BLM mutants were constructed by PCR mediated mutagenesis using Q5^® Site-Directed Mutagenesis Kit (New England Biolabs, E0554S) or NEBuilder^® HiFi DNA Assembly Master Mix (New England Biolabs, E2621S) and the GFP-BLM (a gift from Nathan Ellis, Addgene plasmid #80070) as the template. Plasmid sequences were confirmed by whole plasmid sequencing (Plasmidsaurus). The PCR primers are listed as follows:

HJBLMmutBTR_FWD: 5’-TGCTACATCTTCAGATAACAATGTATC-3’

HJBLMmutBTR_REV: 5’-GCAGCAAAAGTGAAACCTGAAAATTTTG-3’

HJBLMK695R_FWD: 5’-TGGAGGTGGTAGAAGTTTGTGTTAC-3’

HJBLMK695_REV: 5’-GTCGGCATCAGGATAAAAC-3’

GBLMD133_FWD: 5’-ACTGCTCTCAAGAAATTAG-3’

GBLM1_REV: 5’-TCTACCCTCGATCTCG-3’

GBLMD641_FWD: 5’-GAGCGTTTCCAAAGTC-3’

GBLM1417FWD: 5’-TAACAACCGAATCTCAATG-3’

GBLMd1207_REV: 5’-TGCTACTAACGCTTTTTG-3’

GBLMd1291_REV: 5’-CGATGTCCATTCAGAG-3’

GB_BLMsg7RN _FWD: 5’-TACAAGTCCGGACTCAGATCTCGAGCTCAAGCTTCGAATTC-3’

GB_BLMsg7RN _REV: 5’-CCATTGTATCGCACTCTCGCCTGGAGAGGC-3’

GB_BLMsg7RC_ FWD: 5’-GCGAGAGTGCGATACAATGGCTGACACGTTACAG-3’

GB_BLMsg7RC_ REV: 5’-CCGGTGGATCCCGGGCCCGCGGTACCGTCGAGTTTTTTTTTTTTTTTTTTTG-3’

GBLMemp_FWD: 5’-TAACAACCGAATCTCAATG-3’

GBLMemp_REV: 5’-TCTACCCTCGATCTCG-3’

Plasmids expressing BLM-TRF1 fusion protein were constructed by assembling BLM (or mutants) fragments with mCherry-TRF1 fragments generated from pLVX-Ptuner-mCherry-TRF1-FokI and cloned into the pLenti CMV Blast (a gift from Eric Campeau and Paul Kaufman, Addgene plasmid #17486) vector using NEBuilder^® HiFi DNA Assembly Master Mix (New England Biolabs,E2621s) or through PCR mediated mutagenesis using Q5^® Site-Directed Mutagenesis Kit (New England Biolabs, E0554S). Plasmid sequences were confirmed by whole plasmid sequencing (Plasmidsaurus). The PCR primers or synthesized DNA fragments are listed as follows:

BLM_N_ FWD: 5’-CACCATGGGAACCAATTCAGATGGCTGCTGTTCCTCAAAATAATC-3’

BLM_N_REV: 5’-CTATTGGCTCCTGATGTCGCTTGAGAAGTATGTGAGGCTG-3’

BLM_mCheTRF1C_FWD: 5’-GCGAGGGCCGCCCCTACGAGGGCACCCAGACCGCCAAG-3’

BLM_mCheTRF1C_REV: 5’-CGGCCGCGAATTCGGTACCGTTAGTCTTCGCTGTCTGAGGAAATCAG-3’

LinkerFrag (synthesized): 5’-GCGACATCAGGAGCCAATAGCAAATTGGGGATTATGGCTCCACCGAAGCCTATAAATAGACCGTTTCTTAAGCCTTCATATGCATTCTCAGGAGGGAGCGGGGGAGGCTCTGGCGGATCCGTGAGCAAGGGCGAGGAGGATAACATGGCCATCATCAAGGAGTTCATGCGCTTCAAGGTGCACATGGAGGGCTCCGTGAACGGCCACGAGTTCGAGATCGAGGGCGAGGGCGAGGGCCGCCCCTACGAG-3’

BLMmutBTR_ FWD: 5’-TGCTACATCTTCAGATAACAATGTATC-3’

BLMmutBTR_REV: 5’-GCAGCAAAAGTGAAACCTGAAAATTTTG-3’

BLMmutK695R_FWD: 5’-TGGAGGTGGTAGAAGTTTGTGTTAC-3’

BLMmutK695_REV: 5’-GTCGGCATCAGGATAAAAC-3’

BLMmut1417_ FWD: 5’-ACGCGTGTGAGCAAGGGC-3’

BLMmutd12071417_ REV: 5’-TGCTACTAACGCTTTTTGTTTTTTCACAC-3’

BLMmutd12911417_ REV: 5’-CGATGTCCATTCAGAGTATTTCTGTAATACTG-3’

mCherryMluI_FWD: 5’-ACGCGTGTGAGCAAGGGCGAGGAG-3’

pLentiCMVMulI_REV: 5’-CATCTGAATTGGTTCCCATGGTG-3’

MluIBLMWT_FWD: 5’-ATGGGAACCAATTCAGATGACGCGTATGGCTGCTGTTCCTCAA-3’

MluIBLMWT_REV: 5’-CCTCCTCGCCCTTGCTCACACGCGTTGAGAATGCATATGAAGGCTTAAG-3’

MluIBLMd1–133_FWD: 5’-ATGGGAACCAATTCAGATGACGCGTACTGCTCTCAAGAAATTAGAATTT-3’

MluIBLMd1–641_FWD: 5’-ATGGGAACCAATTCAGATGACGCGTGAGCGTTTCCAAAGTCTTAG-3’

HJBLMSg7RFrag1_FWD: 5’-GCCCGACTACGCCGGAGGACATGGCTGCTGTTCCTCAAAATAATC-3’

HJBLMSg7RFrag1_REV: 5’-CCATTGTATCGCACTCTCGCCTGGAGAGGC-3’

HJBLMSg7RFrag2_FWD: 5’-GCGAGAGTGCGATACAATGGCTGACACGTTACAG-3’

HJBLMSg7RFrag2_REV: 5’-GTCGAGCGGCCGCCACTGTGTTATGAGAATGCATATGAAGG-3’

The constructs were transiently transfected into target cell lines with Avalanche^®-Omni Transfection Reagent (EZT-OMNI-1) according to the manufacture’s instruction.

CRISPR/Cas9 mediated gene deletions, lentivirus generation and transduction

SpCas9 was expressed with lenti-SpCas9-hygro (a gift from Brett Stringer, Addgene plasmid # 104995), while sgRNAs targeting candidate genes of interest were generated with lentiGuide-neo³⁴ or lentiGuide-Puro (a gift from Feng Zhang, Addgene plasmid #52963) or lentiGuide-GFP (a gift from Richard Young, Addgene plasmid #104374). For U2OS cells, three BLM knockout clones expressing sgRNA#7 were isolated and validated. For LM216J cells, two individual sgRNAs (#6, #7) were used separately to knockout BLM. For BLM and DNA2 double knockout, BLM and DNA2 were sequentially knocked out with sgRNA cloned into the lentiGuide-GFP (Addgene plasmid #104374) vector and lentiCRISPRv2-blast (a gift from Brett Stringer, Addgene plasmid #98293) vector, respectively. SaCas9 and the sgRNAs targeting ATRX were cloned into a SaCas9 all-in-one vector as described⁶¹. The sgRNAs targeting telomere was cloned as described⁴⁵. In brief, the vectors were digested with BsmBI-v2 (New England Biolabs, R0739S) and ligated with annealed sgRNA oligos. sgRNA sequences were listed as follows:

sgRosa26_FWD: 5’-CACCGGAAGATGGGCGGGAGTCTTC-3’

sgRosa26_REV: 5’-AAACGAAGACTCCCGCCCATCTTCC-3’

sgBLM#6_FWD: 5’-CACCGGGAACGAACTGCTTCAGCAG-3’

sgBLM#6_REV: 5’-AAACCTGCTGAAGCAGTTCGTTCCC-3’

sgBLM#7_FWD: 5’-CACCGCAGGCGAGAATGTGACACCA-3’

sgBLM#7_REV: 5’-AAACTGGTGTCACATTCTCGCCTGC-3’

sgDNA2#7_FWD: 5’-CACCGCCTGCCCATTTACAAAACGA-3’

sgDNA2#7_REV: 5’-AAACTCGTTTTGTAAATGGGCAGGC-3’

sgDNA2#8_FWD: 5’-CACCGTTGGTGTGAAAATACATCGA-3’

sgDNA2#8_REV: 5’-AAACTCGATGTATTTTCACACCAAC-3’

sgDNA2#9_FWD: 5’-CACCGGTTGGTGTGAAAATACATCG-3’

sgDNA2#9_REV: 5’-AAACCGATGTATTTTCACACCAACC-3’

sgATRX#1_FWD: 5’-CACCGGTGAATCCGAAGATGAACAG-3’

sgATRX#1_REV: 5’-AAACCTGTTCATCTTCGGATTCACC-3’

sgATRX#11_FWD: 5’-CACCGCAGGATCGTCACGATCAAAG-3’

sgATRX#11_REV: 5’-AAACCTTTGATCGTGACGATCCTGC-3’

sgATRX#2_FWD: 5’-CACCGACAAATAAGAACTTGCAATG-3’

sgATRX#2_REV: 5’-AAACCATTGCAAGTTCTTATTTGTC-3’

sgSCR _FWD: 5’-CACCGTGCTCCGTGCATCTGGCATC-3’

sgSCR _REV: 5’-AAACGATGCCAGATGCACGGAGCAC-3’

sgTelo_FWD: 5’-CACCGGTTAGGGTTAGGGTTAGGGTTA-3’

sgTelo_REV: 5’-AAACTAACCCTAACCCTAACCCTAACC-3’

Viral particles were used to deliver plasmids for gene depletion experiments. Briefly, plasmids and lentiviral packaging plasmids, psPAX2 and pMD2.G, were transfected into HEK 293T cells with Polyethylenimine (Polysciences, 23966). Fresh medium was changed 8hrs after transfection. The lentiviral supernatant was collected and combined at 48hrs and 72 hrs post transfection, then filtered through 0.45 μm PES filter. Spin-infection were done with 8 μg ml⁻¹ polybrene (Sigma-Aldrich, H9268). Antibiotic selection for lentivirus infected cells were done using puromycin (2 μg ml⁻¹, Thermo Fisher, A1113803), blasticidin (10 μg ml⁻¹, Invivogen, ant-bl-1), hygromycin (100 μg ml⁻¹, Invivogen, ant-hg-1) or neomycin (800 μg ml⁻¹, Thermo Fisher, 11811031). Alternatively, when lentiGuide-GFP construction was used, the cells were sorted with flow cytometry.

Proteomics of isolated chromatin segments (PICh)

Telomere-associated proteins were isolated by PICh as described before ^33,62 in sgRosa and sgBLM U2OS cells. Eighty-five 15 cm dishes of cells were fixed with 4% formaldehyde (Sigma-Aldrich, 252549) for 45 min at room temperature then washed three times with ice-cold 1× PBS and scrapped in pre-chilled 1× PBS+ 0.05% Tween 20. Cell pellets were then dounced in sucrose solution (0.3 M Sucrose, 10 mM HEPES-NaOH pH 7.9, 1% Triton-X 100, 2 mM Magnesium Acetate). Cell pellets were resuspended in 1× PBS, 0.5% Triton X-100, 1 mM PMSF and 0.25 mg ml⁻¹ RNase A (Qiagen, 19101) and incubated on a rotator for 16 hrs at 4 °C. Cell pellets were then washed three times with ice-cold 1× PBS and two more times with LB4 buffer (50mM Tris-HCl pH 8.0, 200mM NaCl, 20 mM EDTA-NaOH pH 8.0, 1% SDS). Then cell pellets were resuspended in LB4 buffer supplemented with 1 mM PMSF (~2.5 folds to nuclei volume) and sonicated with BRANSON Digital Sonifier (Amplitude: 70%; Pulse cycle: 15 sec On, 45 sec Off; Total process time: 7.5 min). Soluble chromatin fractions were heated at 58 °C for 5min and centrifuged at 16,000 g for 10 min at room temperature. Soluble fraction was incubated with 1 mL Streptavidin Agarose (EMD Millipore, 69203–3) for 3hrs at room temperature on a rotator, and then collected by passing through Sephacryl S-400 HR column (GE Healthcare Life Sciences, 17060901) with centrifuge at 750 g for 5 min. Optical Density (OD260 and OD280) was measured with NanoDrop 1000 to confirm efficient RNA removal and quality of soluble chromatin. Soluble fractions were hybridized with 1500 pmol desthiobiotin labeled 2’-Fluor-RNA telomeric probes (t-Desthiobiotin TEG-Spacer18-Spacer18-UUAGGGUUAGGGUUAGGGUUAGGGUUAGGGUUAGGGUUAGGGt) in a PCR machine (25°C for 3 min; 80°C for 5 min; 37°C for 60 min; 60°C for 3 min; 37°C for 30 min; 60°C for 3 min; 37°C for 30 min; then keep at 25°C). Pooled hybridized chromatin was centrifuged at 16,000 g for 15 min at room temperature; and then incubated with 900 μL Dynabeads^™ MyOne^™ Streptavidin C1 (Thermo Fisher, 65002) overnight on a rotator at room temperature. Then bound chromatin was immobilized on a magnetic stand, followed by five times washes with LB3JD buffer (10 mM HEPES-NaOH pH 7.9, 100 mM NaCl, 2 mM EDTA-NaOH pH 8.0, 1 mM EGTA-NaOH pH 8.0, 0.2% SDS, 0.1% Sodium Sarkosyl), followed by washing with LB3JDS buffer (10 mM HEPES-NaOH pH 7.9, 30 mM NaCl, 2 mM EDTA-NaOH pH 8.0, 1 mM EGTA-NaOH pH 8.0, 0.2% SDS, 0.1% Sodium Sarkosyl) at 42 °C and 1000 rpm for 5 min. Telomere-associated proteins on beads were eluted twice with 450 μL elution buffer (75% LB3JD buffer + 25% D-biotin (Invitrogen, B20656)) by the following steps (37 °C and 65 °C for 30 min each). The eluted samples were precipitated with TCA precipitation protocol. Crosslinked proteins were reversed with decrosslinking buffer (250 mM Tris-HCl pH 8.8, 2% SDS, 0.1 M 2-Mercaptoethanol). NuPAGE LDS Sample Buffer (4×, Invitrogen) was added to boil samples. Then samples were analyzed by silver staining, and mass spectrometry (Taplin Biological Mass Spectrometry Facility at Harvard Medical School).

Immunofluorescence and Immunofluorescence- fluorescence in situ hybridization (IF-FISH)

For immunofluorescence, cells were grown on circular coverslips, poly-L-Lysine (Sigma-Aldrich, P4832) treatment was performed if necessary. Cells were first pre-extracted with CSK buffer+0.5% Triton X-100 (20mM HEPES pH8.0, 100mM NaCl, 3mM MgCl₂ and 300mM Sucrose, 0.5% Triton X-100) for 10min on ice if necessary; then fixed with 3% (w/v) formaldehyde (Sigma-Aldrich) for 15 min. Cells were then permeabilized with pre-chilled 0.5% Triton X-100 in 1× PBS for 10 min on ice and blocked with blocking buffer (1× PBS, 0.1% Tween-20 and 5% goat serum) for 1 hr at room temperature. Primary antibodies were prepared in blocking buffer and incubated overnight at 4 °C in wet chamber. The coverslips were washed and incubated with corresponding secondary antibodies for 1 hr at room temperature in wet chamber.

Immunofluorescence of PARylation was done as described⁵⁰. In brief, cells were treated with 10 μM PARG inhibitor (PDD00017273, Tocris, 59–521-0) for 30 min before processing. Then cells were rinsed with PBS and pre-extracted in CSK + 0.5% Triton X-100 (20mM HEPES pH8.0, 100mM NaCl, 3mM MgCl₂ and 300mM Sucrose, 0.5% Triton X-100) supplemented with 10μM PARPi (Olaparib, Selleck Chemicals, S1060) and PARGi inhibitor (PDD0017273, Tocris, CAS: 1945950–21-9) on ice for 5 min. Then cells were fixed with cold 3% (w/v) formaldehyde for 15 min. Cells were then permeabilized using ice-cold methanol/acetone (1:1) mix for 5 min followed by another 5 min in pre-chilled 0.5% Triton X-100 in 1× PBS and blocked in 3% BSA in PBS for 1 hr at room temperature. Then primary antibodies and secondary antibodies were incubated as regular immunofluorescence.

EdU was labeled using the Click-iT EdU Cell Proliferation Kit (Thermo Scientific, C10337) following the manufacture’s instruction.

For IF-FISH, following immunofluorescence, coverslips were re-fixed with 3% (w/v) formaldehyde at room temperature for 10 min, washed with 1× PBS and dehydrated with ethanol (75%, 95%, 100%) and then air-dried. Coverslips were then heat-denatured at 80°C for 7 min and hybridized with a telomeric probe (TelC-Cy3) (PNA Bio, F1002) in hybridization solution (70% deionized formamide, 0.5% Roche blocking reagent, 10mM Tris-HCl pH 7.4) for 2 hrs (U2OS and LM216J) or overnight (HeLa S3, GM847, VA13 and IIICF E6/E7) at room temperature in wet chamber. Coverslips were then washed, stained with DAPI (Sigma-Aldrich, D9542) and mounted using ProLong Gold Antifade Mountant (Thermo Scientific, P36934). Images were acquired with a QImaging RETIGA-SRV camera connected to a Nikon Eclipse 80i microscope at 63x oil objective, or the Leica TCS SP5 II confocal microscope at 63x oil objective.

Primary antibodies used for immunofluorescence: RPA2 (Anti-RPA, clone RPA34–20, MABE285, EMD Millipore), BLM (A300–110A, Bethyl Laboratories), Poly/Mono-ADP Ribose (#83732, Cell Signaling Technology), PCNA (2586S, Cell Signaling Technology), GFP (A-11122, Thermo Fisher Scientific), mCherry (ab183628, Abcam), Cyclin A (sc-271682, Santa Cruz Biotechnology). Secondary antibodies used for immunofluorescence: Goat anti-Rabbit IgG (H+L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor^™ 488 (A-11034, Thermo Fisher Scientific); Goat anti-Mouse IgG (H+L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor^™ 488 (A-11029, Thermo Fisher Scientific); Goat anti-Mouse IgG (H+L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor^™ 647 (A-21236, Thermo Fisher Scientific); Goat anti-Rabbit IgG (H+L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor^™ 647 (A21245, Thermo Fisher Scientific).

Telomere native FISH

Telomere native FISH was done as described³⁵. In brief, cells were rinsed with PBS and fixed in pre-chilled (−20°C) 70% ethanol for 20 min at room temperature. The cells were then washed twice in washing buffer (100 mM Tris–HCl pH 8,150 mM NaCl, 0.05% Tween-20) and incubated at 37°C for 30 min with 50 ug/ml RNase A (Qiagen, 19101) and 10U/ml RNase H (New England Biolabs, M0297L) in PBS. Then cells were washed again in washing buffer. Followed by hybridization in the hybridization solution (2× SSC buffer, 8% de-ionized formamide, 5 nM fluorescent telomeric probe (TelC-Cy3, PNA Bio, F1002) or (TelG-Cy3, PNA Bio, F1006)) for 40 min at 37 °C. After hybridization, cells were rinsed twice in filtered 2xSSC buffer and one time in PBS. Then the nucleus was stained in 0.8ug/ml DAPI in 1× PBS for 10 min and dehydrated with ethanol (75%, 95%, 100%), air-dried and mounted using ProLong Gold Antifade Mountant (Thermo Fisher Scientific, P36934).

SDS-PAGE and Western blot

Cell pellets were lysed with RIPA buffer supplemented with cOmplete protein inhibitor cocktail (Roche) and Benzonase (EMD Millipore, 70746) for 30 min on ice. Soluble protein concentration was measured with Protein Assay Dye Reagent (Bio-Rad, 5000006). Proteins were separated by SDS-PAGE using 4–12% Bis–Tris gel (Invitrogen) and transferred to an Amersham Protran 0.2 μm nitrocellulose membrane (Amersham, 10600004). Membranes were blocked with 5% milk and sequentially incubated with primary antibodies overnight at 4 °C, corresponding secondary antibodies for 1 hr at room temperature, and developed using Western Lightning Plus-ECL (Perkins Elmer, NEL105001EA). When necessary, primary and secondary antibodies were stripped with Restore Western Blot Stripping Buffer (Pierce, 21059). Primary antibodies used for Western blot: BLM (A300–110A-M, Bethyl Laboratories), ATRX (sc-15408 or sc-55584, Santa Cruz Biotechnology), DNA2 (ab96488, Abcam), GFP (A-11122, Thermo Fisher Scientific), Cas9 (A9000–050, Epigentek), gamma-H2AX (05–636-I, EMD Millipore), KU86 (sc-9034, Santa Cruz Biotechnology), TRF1 (GeneTex, GTX70304), GAPDH (2118S, Cell Signaling Technology), α-Tubulin (12G10, Developmental Studies Hybridoma Bank), β-Actin (sc-47778, Santa Cruz Biotechnology). Secondary antibodies used for western blot: ECL Anti-Mouse IgG, horseradish peroxidase-linked whole antibody (from sheep) (Sigma-Aldrich/GE, NA931) or ECL Anti-Rabbit IgG, horseradish peroxidase-linked whole antibody (from donkey) (Sigma-Aldrich/GE, NA934).

Telomere restriction fragment (TRF) Assay

Telomere restriction fragment (TRF) assay with pulsed-field gel electrophoresis was performed as described³⁴. Genomic DNA was extracted using DNeasy Blood & Tissue Kit (Qiagen, 69506), and concentration was measured by Qubit Fluorometric Quantification (Qubit^™ dsDNA Quantification Assay Kits, Thermo Fisher Scientific, Q32851; Qubit 3.0 Fluorometer, Thermo Fisher Scientific). Five micrograms of genomic DNA were digested with HinfI (New England Biolabs, R0155S) and RsaI (New England Biolabs, R0167L) overnight at 37°C. The digested DNA was subjected to electrophoresis with the CHEF-DRII system (Bio-Rad). Samples were loaded in 1% pulse field (PFGE) certified agarose (1620137, BioRad) in pre-chilled 0.5× TBE buffer using following parameter: 4 V/cm; initial switch time 5 sec, final switch time 5 sec, for 20 hrs at 14 °C. Alternatively, constant-field gel electrophoresis was utilized to separate the digested DNA using the Owl A5 Large Gel System (Thermo Fisher) with 0.7% agarose gel in 1× TAE buffer at 2 V/cm for 20 hrs in the cold room.

The gel was dried for 4 hrs at 42 °C, and stained with Ethidium Bromide (Sigma-Aldrich, 2375), and subjected to denatured in-gel hybridization with ³²P-labeled telomere C-probe as shown below.

Neutral-neutral two-dimensional gel electrophoresis

Neutral-neutral two-dimensional gel electrophoresis was performed as previously described^43,63. Ten micrograms of genomic DNA were digested as for TRF assay. Restriction enzyme digested gDNA were first resolved with 0.4% agarose gel in 1× TBE at 1 V cm⁻¹ at room temperature for 7 hrs (6 samples). The lanes with samples were cut and soaked in 1× TBE with 0.3 μg ml⁻¹ Ethidium Bromide (EB) for 30 min, then casted into 1% agarose gel containing 0.3 μg ml⁻¹ EB in 1× TBE. Second dimensional gel electrophoresis was operated at 3 V/cm for 3–3.5 hrs at 4°C. The gel was dried for at least 2 hrs at 42 °C and subjected to native and denatured in-gel hybridization with ³²P-labeled telomere C-/ G-probe.

Native and denatured in-gel hybridization

For native hybridization, the dried gel was first prehybridized with Denhart’s hybridization buffer (5×SSC, 0.5% SDS, 10× Denhart’s buffer) then hybridized overnight at 42°C with ³²P-labeled C- or G- telomeric probe prepared as shown before⁶⁴. The hybridized gel was then washed 3 times with 2× SSC and 0.5% SDS and 3 times with 2× SSC + 0.1% SDS, then exposed to PhosphorImager screen (GE Healthcare) and scanned on Typhoon FLA 7000 with ImageQuant (Molecular Dynamics). For denatured hybridization, gel was first denatured with 0.5 M NaOH + 1.5 M NaCl for half hour; neutralized with 2 washes in 0.5 M Tris–HCl pH 8.0 + 1.5 M NaCl for half hour each; and then prehybridized and hybridized as shown above.

Purification of Human Recombinant Proteins

WT and nuclease-deficient (D277A) hDNA2 with C-terminal Flag epitope tag was introduced into 438A MacroBAC vector to create pJD177 and pJD73 respectively and bacmid DNA was generated in the Escherichia coli strain DH10Bac (Invitrogen). The bacmid was used to generate a recombinant virus in SF9 insect cells. HighFive insect cells at the density of 1 × 10⁶ cells/ml were infected with the baculovirus, grown in shaking flask at 25°C and harvested after 48 hrs. All subsequent steps were carried between 0 and 4°C. An extract was prepared by sonication of a cell pellet derived from 800 ml of culture in 60 ml of K buffer (20 mM KH₂PO₄, 10% glycerol, 0.5 mM EDTA, 0.01% IGEPAL CA-630 (MilliporeSigma), 1 mM DTT, 1 mM phenylmethylsulfonyl fluoride (PMSF) and 5 μg/ml each of aprotinin, chymostatin, leupeptin and pepstatin) containing 500 mM KCl. After ultracentrifugation (100 000 × g for 60 min), the clarified lysate was incubated with 3 ml of anti-Flag M2 agarose (Sigma) for 1 hr with constant agitation. The resin was washed with 15 ml of K buffer with 500 mM KCl, 100 ml of K buffer with 500 mM KCl, 1 mM ATP and 10 mM MgCl₂ and 100 ml of K buffer with 150 mM KCl. To elute the bound proteins, the resin was treated for 10 min with 3 ml of K buffer containing 150 mM KCl and 250 ng/μl of Flag peptide (Sigma). This elution procedure was repeated four additional times. The pooled eluate was loaded onto a 1 ml HP Q column and eluted with a 50 ml gradient of 150–500 mM KCl. Fractions containing DNA2 were pooled, concentrated to 500 μl using Amicon centrifugal filter tubes (MilliporeSigma) and further fractioned on 24 ml Suprdex S200 column (GE Healthcare) in K buffer with 250 mM KCl. Fractions containing the DNA2 peak were combined, concentrated using Amicon filter tubes and stored in small portions at −80°C.

BLM with C-terminal His epitope tag was introduced into 438A MacroBAC vector and a bacmid was generated in the Escherichia coli strain DH10Bac (Invitrogen). The bacmid was used to generate a recombinant virus in SF9 insect cells. HighFive insect cells at the density of 1 × 10⁶ cells/ml were infected with the baculovirus, grown in shaking flask at 25°C and harvested after 48 hrs. All subsequent steps were carried between 0 and 4°C. Cell extract was prepared by sonication of a cell pellet derived from 800 ml of culture in 60 ml of K buffer (20 mM KH₂PO₄, 10% glycerol, 0.5 mM EDTA, 0.01% IGEPAL CA-630, 1 mM DTT, 1 mM phenylmethylsulfonyl fluoride and 5 μg/ml each of aprotinin, chymostatin, leupeptin and pepstatin) containing 300 mM KCl and 25 mM imidazole. After ultracentrifugation (100 000 × g for 60 min), the clarified lysate was incubated with 3 ml of Ni-NTA resin (Qiagen) for 2 hrs with constant agitation. The resin was washed with 20 ml of K buffer with 300 mM KCl and 25 mM imidazole, 1000 ml of K buffer with 1M KCl and 25 mM imidazole, 1 mM ATP and 10 mM MgCl₂ and 20 ml of K buffer with 150 mM KCl and 25 mM imidazole. To elute the bound proteins, the resin was treated for 10 min with 3 ml of K buffer containing 150 mM KCl and 250mM imidazole. This elution procedure was repeated four additional times. The pooled eluate was loaded onto a 1 ml heparin column (Amersham) and eluted with a 50 ml gradient of 150–600 mM KCl. Fractions containing BLM were pooled, concentrated to 500 μl using Amicon filter tubes and further fractioned on 24 ml Superdex 200 column (GE Healthcare) in K buffer with 300 mM KCl. Fractions containing the BLM peak were combined, concentrated using Amicon centrifugal filter tubes and stored in small portions at −80°C.

pYES2-BLM K695R was introduced into the protease-deficient S. cerevisiae strain JEL-1 and protein expression was induced by 2% galactose. Cells were harvested by centrifugation and stored at –80 °C. All the subsequent steps were carried out at 4 °C. To prepare extract, 50 g of frozen yeast paste was thawed in 50 ml of cell breakage buffer (50 mM Tris-HCl,pH 7.5, 500 mM KCl, 10% sucrose, 0.5 mM EDTA, 1 mM 2-mercaptoethanol, and the following protease inhibitors: aprotinin, chymostatin, leupeptin, and pepstatin A each at 5 μg/ml and 1 mM PMSF). Cells were lysed in a French press and the crude extract was subject to centrifugation (100,000 × g for 90 min). The supernatant was mixed gently with 5 ml of Ni-NTA resin (Qiagen) for 1 hr. The resin was poured into a column and washed with 50 ml of buffer K (20 mM K2HPO4, pH 7.4, 10% glycerol, 0.5 mM EDTA, 0.01% Igepal, and 1 mM DTT, and protease inhibitors) containing 500 mM KCl and 15 mM imidazole. Bound proteins were eluted with a 40 ml gradient of 15–250 mM imidazole in buffer K containing 500 mM KCl, with BLM eluting at ~100 mM imidazole. The peak fractions were pooled, diluted with 1.5 volumes of buffer T (20 mM Tris-HCl, pH 7.5, 10% glycerol, 0.5 mM EDTA, 0.01% IGEPAL CA-630, 1 mM DTT, and the protease inhibitors listed above), and applied onto an 8 ml Source Q column (GE Healthcare), which was washed with 20 ml of buffer T containing 200 mM KCl and then developed with a 65 ml gradient of 200–500 mM KCl in buffer K. Fractions containing BLM, which eluted at ~360 mM KCl, were pooled and applied onto a 1 ml column of macro-hydroxyapatite (Bio-Rad), which was eluted with a 30 ml gradient of 80–350 mM KH2PO4 in buffer K, with BLM eluting at ~250 mM KH2PO4. The peak fractions from this step were pooled, diluted with an equal volume of buffer K, and applied onto a 1 ml Mono S column, which was eluted with a 15 ml 100–700 mM KCl gradient in buffer K. Fractions that contained BLM, were concentrated using Amicon filter tubes and stored in small aliquots at −80°C.

In vitro recombinant BLM and DNA2 activity on telomeric DNA assay

Genomic DNA, from control or BLM knockout LM216J cells, was extracted using Wizard Genomic DNA Purification Kit (Promega, A1120), and the concentration was measured by Qubit Fluorometric Quantification (Qubit^™ dsDNA Quantification Assay Kits, Thermo Fisher Scientific, Q32851; Qubit 3.0 Fluorometer, Thermo Fisher Scientific). gDNA was digested with HinfI (New England Biolabs, R0155S) and Rsa I (New England Biolabs, R0167L) as for TRF assay. Then 1 μg of the restriction enzyme-digested genomic DNA were aliquoted for the following assay. 15 nM WT or K695R human recombinant BLM, and 2.5 nM human DNA2 were added into the aliquoted reactions for another 2 hrs at 37°C, followed by adding 0.2% SDS and 0.25 μg/μl protease K (Qiagen, 19133) and incubate at 37°C for another 15 min. The samples were then subjected to agarose electrophoresis and native and denatured in-gel hybridization with ³²P-labeled telomere C-/ G-probe as shown above separately.

QUANTIFICATION AND STATISTICAL ANALYSIS

All statistical analyses were performed using GraphPad Prism 9 software. Significance was calculated by the two-tailed Student’s t-test, unless otherwise specified. The scatter plot shows each data together with the mean and the SEM highlighted with red lines. In the scatter plot with bar, all data points are shown in red circles. The mean was indicated by the bar and the SEM was indicated by the lines. The Tukey whiskers plot displays the 25th to the 75th percentiles as a box with the median highlighted. The top line shows the 75th percentile plus 1.5 times interquartile range (IQR) and the bottom line shows the minimum value of the data set. All experiments were biologically repeated three times or more for confirmation. The significance was defined as follows: ns: p > 0.05, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. All of the statistical details of experiments can also be found in the figure legends.

Highlights.

• BLM helicase is required for assembly of the ALT telomere DNA damage response

• BLM promotes a replication-associated ALT telomere damage response.

• BLM helicase activity generates 5’-flaps at telomere during lagging strand synthesis

• DNA2 acts on BLM generated telomere 5’-flaps to limit ALT