Cooperative interaction of CST and RECQ4 resolves G‐quadruplexes and maintains telomere stability

Abstract

Human CST (CTC1‐STN1‐TEN1) is a ssDNA‐binding complex that interacts with the replisome to aid in stalled fork rescue. We previously found that CST promotes telomere replication to maintain genomic integrity via G‐quadruplex (G4) resolution. However, the detailed mechanism by which CST resolves G4s in vivo and whether additional factors are involved remains unclear. Here, we identify RECQ4 as a novel CST‐interacting partner and show that RECQ4 can unwind G4 structures in vitro using a FRET assay. Moreover, G4s accumulate at the telomere after RECQ4 depletion, resulting in telomere dysfunction, including the formation of MTSs, SFEs, and TIFs, suggesting that RECQ4 is crucial for telomere integrity. Furthermore, CST is also required for RECQ4 telomere or chromatin localization in response to G4 stabilizers. RECQ4 is involved in preserving genomic stability by CST and RECQ4 disruption impairs restart of replication forks stalled by G4s. Overall, our findings highlight the essential roles of CST and RECQ4 in resolving G‐rich regions, where they collaborate to resolve G4‐induced replication deficiencies and maintain genomic homeostasis.

CST plays a crucial role in engaging RECQ4 with telomeres, enabling efficient resolution of G4 structures and restarting stalled replication forks. Their cooperative action addresses G4‐induced replication challenges, preserving genomic stability and homeostasis.

Introduction

Telomeres, consisting of repetitive DNA sequences and telomere‐binding proteins, play a crucial role in protecting the ends of chromosomes (Arnoult & Karlseder, 2015). In mammals, telomeres are composed of double‐stranded TTAGGG repeats spanning 2–20 kb, with single‐stranded overhangs ranging from 50 to 500 nt (Zhao et al, 2008). The replication of telomere DNA is primarily carried out by the conventional DNA replication machinery. However, due to the repetitive and heterochromatic nature of telomeres, their replication often encounters obstacles and necessitates additional factors to prevent replication failure (de Lange, 2018). Errors in telomere replication frequently lead to telomere dysfunction or loss, ultimately resulting in cellular senescence or apoptosis (de Lange, 2005).

G‐quadruplexes (G4s) are secondary DNA structures that form in G‐rich regions, including telomeres (Stewart et al, 2018). These structures are also found at gene promoters, DNA replication origins, and the 5′‐UTR of certain mRNAs (Renaud de la Faverie et al, 2014). The presence of G4s can block DNA polymerases, leading to replication fork stalling (Burger et al, 2005). It is crucial to resolve G4 structures to ensure efficient DNA replication and maintain genome stability. Several protein factors, such as FANCJ, BRCA1, BRCA2, PIF1, and the DNA nuclease DNA2, have been implicated in the unwinding of G4 structures at telomeres (Paeschke et al, 2013; Mendoza et al, 2016; Higa et al, 2017; Maestroni et al, 2017).

In our previous work, we demonstrated that mammalian CST (CTC1‐STN1‐TEN1) promotes telomere replication and safeguards genome integrity by preventing the accumulation of G4 structures (Zhang et al, 2019). Initially identified as a DNA polymerase alpha‐primase (pol alpha) accessory factor (AAF), CST enhances the primase and polymerase activities of pol alpha (Goulian & Heard, 1990). Our previous findings also revealed its crucial role in telomere replication by facilitating duplex replication and C‐strand fill‐in following telomerase‐mediated elongation (Wang et al, 2012). Moreover, evidence from our study and others suggests that CST is involved in genome‐wide DNA replication. For instance, CST collaborates with pol alpha to resolve stalled replication forks through dormant origin activation (Stewart et al, 2012). It also promotes RAD51 recruitment to GC‐rich sequences for recombination‐based fork restart (Chastain et al, 2016) and interacts with the MCM2‐7 complex, preventing excessive origin licensing by inhibiting CDT1 binding to MCM2‐7 (Wang et al, 2019).

Although we previously demonstrated that CST has the ability to unfold G4 structures in vitro (Zhang et al, 2019), as a single‐strand binding protein, it lacks the active helicase function required to actively unwind the G4 structure. Therefore, it is likely that CST collaborates with other protein factors to resolve G4 structures in vivo. In this study, we identified RECQ4 as a novel cofactor for CST in G4 unwinding. RECQ4 is a helicase and a member of the RecQ family, which includes RECQ1, WRN, BLM, RECQ4, and RECQ5. RECQ4 is involved in DNA metabolism and maintenance of genomic stability (Croteau et al, 2014).

RECQ4 is known to play important roles in various DNA processes, including homologous recombination (HR), nonhomologous end‐joining (NHEJ)‐mediated double‐strand break (DSB) repair, and replication initiation (Singh et al, 2012). It has also been shown to interact directly with TRF1 and contribute to telomere homeostasis (Ghosh et al, 2012). However, the specific biological function of RECQ4 in relation to G4 structures remains unknown, and further investigation is required to elucidate its significance in DNA replication and tumorigenesis.

In this study, we have made an exciting discovery by identifying RECQ4 as a novel partner of CST with critical implications for telomere and genome integrity. We found that the localization of RECQ4 to telomeres and genomic regions in response to G4 formation largely depends on the presence of CST. Remarkably, the depletion of RECQ4 led to the accumulation of G4 structures at telomeres and resulted in telomere dysfunction. Moreover, through further mechanistic investigations, we unraveled the collaborative role of RECQ4 and CST in unwinding G4 structures and facilitating the replication of stalled forks induced by G4s. This dynamic interaction between RECQ4 and CST enables the resolution of secondary structure‐induced replication deficiencies. Importantly, our findings shed light on the potential therapeutic implications of targeting CST and RECQ4 in cancer cells, providing a promising avenue for developing novel molecular strategies to overcome G4‐induced chemotherapy resistance.

Results

RECQ4 interacts with CST in mammalian cells

Previous studies have suggested that CST plays a role in unfolding G4 structures, thus preventing their accumulation and facilitating efficient telomere and genome‐wide DNA replication (Chastain et al, 2016; Zhang et al, 2019). However, the precise mechanism by which CST actively resolves these structures remains elusive. To unravel this mechanism, we performed coimmunoprecipitation (co‐IP) experiments followed by mass spectrometry analysis in HEK293T cells to identify potential interacting partners of CST. Among the candidates, RECQ4, a member of the RECQ helicase family known for its ability to resolve G4 structures, stood out (Fig 1A and Appendix Table S1). Subsequently, co‐IP and immunoblotting were employed to confirm the interaction (Fig 1B).

Figure 1. RECQ4 interacts with CST in mammalian cells.

• A
  Silver staining of IP from HEK293T cells transfected with the plasmids expressing FLAG‐empty‐vector as a negative control, or FLAG‐STN1 in conjunction with HA‐CTC1.
• B
  Co‐IP of the interaction between CST and RECQ4 with and without replication stress. HEK293T cells were treated with 10 μM PDS or 2 mM hydroxyurea (HU) for 24 h.
• C
  Relative levels of MYC‐RECQ4 and RECQ4 were quantified with ImageJ software and normalized to Input (IP/Input). n = 4 biological replicates.
• D–F
  Pull‐down experiment in vitro was performed with purified FLAG‐tagged CTC1 and His‐tagged RECQ4 from insect Sf9 cells (D). Pull‐down experiments with purified FLAG‐tagged RECQ4 from insect Sf9 cells and His‐tagged STN1 or TEN1 from from E. coli (E, F). FLAG beads were used.
• G
  A diagram of the full‐length RECQ4 protein and its fragments.
• H, I
  GST pull‐down assays were used to investigate interactions between different GST‐fused RECQ4 fragments (N‐terminal, helicase domain, and C‐terminal domain) and FLAG‐CTC1, His‐STN1, or His‐TEN1.
• J
  The interactions of RECQ4 and CST complex is depicted in the schematic diagram.

Data information: The immunoblots in (A, D–F, H, and I) are representative examples from three biological replicates. The immunoblots in (B) are representative examples from four biological replicates. Data represent mean ± SD of four biological replicates (C). ****P < 0.0001, **P < 0.01, *P < 0.05. One‐way ANOVA with Fisher's LSD tests was used for (C).

Source data are available online for this figure.

To investigate whether the CST‐RECQ4 interaction is influenced by replication stress or G4 formation, we treated cells with the ribonucleotide reductase inhibitor hydroxyurea (HU) and the G4 stabilizer pyridostatin (PDS) (Petermann et al, 2010). Interestingly, our results revealed an augmented binding between RECQ4 or MYC‐RECQ4 and FLAG‐STN1 following treatment with both HU and PDS (Fig 1B and C), suggesting that replication stress stimulates the interaction between CST and RECQ4.

Next, to further dissect the direct interactions between RECQ4 and specific components of the CST complex, we conducted in vitro pull‐down experiments using purified FLAG‐tagged CTC1, His‐tagged RECQ4, His‐tagged STN1, and His‐tagged TEN1. Purification of CTC1 and RECQ4 was performed using insect Sf9 cells, while STN1 and TEN1 purification was carried out using E. coli. Consistently, our results demonstrated direct interactions between full‐length RECQ4 and both CTC1 and STN1 (Fig 1D and E). However, no direct interaction was observed between TEN1 and full‐length RECQ4 (Fig 1F).

Subsequently, the specific region of RECQ4 that interacts with distinct components of the CST complex was determined, and we performed a GST pull‐down assay using various GST‐tagged RECQ4 truncated fragments, including the N‐terminal domain, helicase domain, and C‐terminal domain (Fig 1G). Remarkably, the results showed that both the C‐terminus and N‐terminus of RECQ4 could be recognized by STN1, while only the N‐terminus exhibited an interaction with CTC1 (Fig 1H and I). Collectively, our findings unveil RECQ4 as a novel cofactor of the CST complex and provide compelling evidence of the physical interaction between RECQ4 and CST (Fig 1J).

RECQ4 plays an essential role in telomere homeostasis

Given the observed telomere defects and increased γ‐H2AX staining on telomeres upon CST depletion, we hypothesized that RECQ4, a potential cofactor of CST, might also be involved in telomere maintenance (Stewart et al, 2012; Kasbek et al, 2013). To investigate this, we generated stable RECQ4‐knockdown clones in HeLa cells and examined telomere dysfunction hallmarks (Figs 2A and EV1A and B). The results revealed a significant increase in multiple telomere FISH signals (MTSs), signal‐free ends (SFEs), and cells with telomere dysfunction‐induced foci (TIFs) after RECQ4 knockdown (Figs 2B–F and EV1C and D).

Figure 2. RECQ4 plays an essential role in telomere homeostasis.

• A
  The expression of RECQ4 in the RECQ4‐knockdown clone is demonstrated, as is the re‐expression of sh‐resistant FLAG‐RECQ4 (RECQ4sh Res) in RECQ4‐knockdown cells. NT stands for non‐target.
• B
  Telomere FISH on metaphase spreads displaying signal‐free ends telomere (SFEs, white arrows) and multiple telomere FISH signals (MTSs, yellow arrows). Scale bar = 20 μm.
• C, D
  Percentage of chromosomes with SFEs (C) and MTSs (D), respectively. Each experiment examined approximately 1,000 chromosomes.
• E
  Immunofluorescence staining of 53BP1 (green) and telomere (red). DAPI (blue). The yellow arrow represents the colocalized signal. Scale bar = 10 μm.
• F
  The percentage of cells that have TIFs. TIFs for at least 50 cells are determined per experiment.

Data information: The immunoblots (A) and micrographs (B and E) show representative examples from three biological replicates. Data represent mean ± SD of three biological replicates (C, D, and F). ***P < 0.001, **P < 0.01, *P < 0.05, ns, not significant. One‐way ANOVA with Fisher's LSD tests was used for (C, D, and F).

Source data are available online for this figure.

Figure EV1. RECQ4 plays an essential role for telomere homeostasis.

• A, B
  Western blot and qRT–PCR were used to demonstrate the levels of RECQ4 in the RECQ4‐knockdown HeLa clone.
• C
  Telomere FISH on HCT116 cells metaphase spreads revealing MTSs (white arrows) and SFEs (yellow arrows). DAPI (blue). Scale bar = 10 μm.
• D
  The percentage of chromosomes with MTSs and SFEs. More than 1,000 chromosomes are analyzed in each experiment.
• E
  In HCT116 cells, traces of the WT and K508G mutations were sequenced, and altered nucleotide changes are indicated.

Data information: The immunoblots (B) and micrographs (C) show representative examples from three biological replicates. Data represent mean ± SD of three biological replicates (A and D). ****P < 0.0001, ***P < 0.001, *P < 0.05, ns, not significant. One‐way ANOVA with Fisher's LSD tests was used for (A and D).

Source data are available online for this figure.

To further understand the role of RECQ4's helicase function in telomere maintenance, we generated a K508G point mutation helicase‐abolish knock‐in (K508G‐KI) HCT116 cell line (Figs EV1E and 3A). The K508G‐KI HCT116 cells exhibited severe telomere abnormalities, indicating that RECQ4's telomere function is dependent on its helicase activity. Rescue experiments were performed in both RECQ4‐knockdown HeLa cell lines and K508G‐KI HCT116 cell lines to confirm the specific effect of RECQ4. The results demonstrated that re‐expressing wild‐type RECQ4 rescued the telomere defects, whereas overexpression of a helicase‐dead RECQ4 mutant failed to restore telomere stability in either knockdown or knock‐in cells (Figs EV1C and D, and 3B). These findings suggest that RECQ4 plays a significant role in maintaining telomere integrity, possibly by catalyzing the unwinding of telomeric DNA structures.

Figure 3. RECQ4 helicase functions at telomeres.

1. Structure diagrams of RECQ4 wild‐type (WT) and helicase truncation (HT). The helicase‐dead mutation was obtained by replacing lysine in the 508 position of HCT116 cells with Glicine (K508G‐KI).
2. The percentage of chromosomes with MTSs and SFEs. Each experiment examines more than 1,000 chromosomes.
3. Following excitation with 494 nM light, C‐strand, CST, RECQ4 or RECQ4 helicase‐dead mutant was added with or without ATP to produce emission spectra for 5′(FAM)‐Tel21‐(TAMRA)3′.
4. The telomere length distribution is based on the fluorescence intensity of the FISH signal (TFU, telomere fluorescence unit).

Data information: Data represent mean ± SD of three biological replicates (B). ***P < 0.001, **P < 0.01, *P < 0.05, ns, not significant. One‐way ANOVA with Fisher's LSD tests was used for (B).

Source data are available online for this figure.

Given the previous evidence of CST and RECQ4 binding to G4 structures and their involvement in stalled replication re‐initiation (Keller et al, 2014; Zhang et al, 2019), we proposed that RECQ4 may work together with CST in telomere regulation by controlling G4 formation. To compare the ability of RECQ4 and CST to unwind G4 structures, we performed an in vitro fluorescence resonance energy transfer (FRET) experiment using a double‐labeled FAM‐(GGGTTA)₃GGG‐TAMRA Tel21 oligo. RECQ4, in the presence of ATP, exhibits active and rapid unwinding of G4 structures, demonstrating superior binding and resolution capabilities compared with CST. It efficiently resolves 50% of the G4 structure within 1.8 s, whereas CST requires approximately 4.6 s (Fig 3C). On the contrary, the helicase‐dead mutant of RECQ4 and RECQ4 without ATP lack the ability to actively unwind G4 structures such as RECQ4 with ATP. However, they contribute to the reduction of G4 structures by influencing the chemical equilibrium. This is similar to how the addition of the C‐complementary strand, which binds to the G‐rich single‐stranded DNA, gradually reduces the abundance of G4 structures, albeit at a slower rate. Hence, RECQ4, with ATP, plays a crucial role in the active resolution of G4 structures, while CST and the helicase‐dead mutant of RECQ4 contribute to G4 reduction by affecting the chemical equilibrium. These findings are depicted in Fig 3C. The observation that CST not only facilitates the incorporation of RECQ4 but also enhances its unwinding activity provides support for the hypothesis that CST may transiently capture the G‐rich single‐stranded state that arises from the G4 structure. This transient capture by CST is likely responsible for facilitating RECQ4's engagement and enabling it to fully resolve and unwind the G4 structure.

To further investigate the in vivo effect of RECQ4 on counteracting G4 formation, RECQ4‐depleted cells were treated with G4 stabilizers. The results showed a dramatic increase in telomere dysfunction phenotypes, such as SFEs, MTSs, and TIFs, upon G4 stabilization. Additionally, RECQ4 disruption exacerbated the telomere homeostasis defects synergistically, which could be rescued by inducing sh‐resistant RECQ4 (Fig 2B–F), indicating that RECQ4 is necessary for responding to G4 formation in vivo.

Telomeric quantitative fluorescence in situ hybridization (Q‐FISH) was next employed to examine the role of RECQ4 in telomere length regulation. Interestingly, RECQ4 depletion significantly reduced the mean telomere length, consistent with the increase in the number of chromosomes with telomere loss upon RECQ4 knockdown (Fig 3D). Collectively, our findings suggest that RECQ4 is essential for maintaining telomere homeostasis, possibly through its ability to unwind G4 structures.

RECQ4 functions at telomeres via its interaction with CST

To investigate whether RECQ4's recruitment or stabilization at telomeres is mediated by its interaction with CST, we performed immunofluorescence experiments using TRF2 as a marker for telomeres (Fig 4A and B). We observed a significant decrease in the colocalization between RECQ4 and telomeres upon knockdown of STN1 or CTC1 (Figs 4C and D, and EV2A and B), which could be rescued by re‐expression of shRNA‐resistant STN1 (Fig 4C). The data revealed that approximately 44% of cells exhibited more than four RECQ4‐telomere colocalization foci (Fig EV2C). However, in the absence of STN1, the percentage of cells with RECQ4 localized to telomeres (> 4 foci) was reduced by approximately 40% (to ~17.7%), and this was restored to ~30.4% upon re‐expression of shRNA‐resistant STN1 (Fig EV2C). These findings suggest that CST is required for the recruitment of RECQ4 to telomeres. We also examined RECQ4‐telomere colocalization foci after treatment with PDS to further assess the impact of G4 stabilizers on RECQ4 recruitment to telomeres (Fig 4B–D). The results demonstrated a significant increase in RECQ4‐telomere colocalization foci upon PDS treatment, consistent with the findings in wild‐type cells. Moreover, in the absence of STN1 or CTC1, the number of cells with RECQ4 localized to telomeres was reduced, but this reduction was restored by re‐expression of shRNA‐resistant STN1 (Figs 4B–D and EV2A–C). These results suggest that RECQ4 plays a crucial role in resolving telomeric G4 structures through its interaction with CST.

Figure 4. CST is required for the recruitment of RECQ4 to binding on telomeres.

• A
  STN1 expression was demonstrated in the STN1 knockdown clone, as well as re‐expression of sh‐resistant FLAG‐STN1 (STN1sh Res).
• B
  RECQ4 (green) and TRF2 (red) coimmunofluorescence staining in HeLa cells. DAPI (blue). Colocalizations (yellow) were marked with arrows and displayed in enlarged images. Scale bar = 20 μm.
• C, D
  Number of RECQ4 and TRF2 colocalization per cell after STN1 or CTC1 knock down, as indicated. HeLa cells were treated with 10 μM PDS for 24 h.
• E
  Representative gels showing overhang signal from STN1 or RECQ4 knockdown cells. In‐gel hybridization with (AT2C3)₄ probe before and after denaturation.
• F
  Quantification of overhang signal from STN1 or RECQ4 knockdown cells.

Data information: The immunoblots (A and E) and micrographs (B) are representative examples from three biological replicates. Data represent mean ± SD of three biological replicates (C, D, and F). ****P < 0.0001, ***P < 0.001, **P < 0.01, *P < 0.05, ns, not significant. One‐way ANOVA with Fisher's LSD tests was used for (C, D, and F).

Source data are available online for this figure.

Figure EV2. CST is required for the recruitment of RECQ4 to telomeres.

1. IF‐FISH was used to investigate the colocalization of RECQ4 and telomere in HeLa cells. RECQ4 (red) in U2OS cells; telomere (green); DAPI (blue). Colocalizations are indicated by arrowheads. Scale bar = 20 μm.
2. Quantification of the RECQ4 and telomere colocalization signal of (A) grouped in NC (n = 103 data points) and CTCsi (n = 122 data points).
3. Percentage of cells with RECQ4 and TRF2 colocalization in Hela cells, as indicated.
4. qRT‐PCR were used to demonstrate the relative levels of CTC1 in the CTC1‐knockdown U2OS cells.
5. Examination of RECQ4 and TRF2 colocalization of in CTC1‐depleted U2OS cells. RECQ4 (red); TRF2 (green); DAPI (blue). Colocalization events are indicated by arrowheads. Scale bar = 10 μm.
6. Quantification of the RECQ4 and TRF2 colocalization signal of (E) grouped in NC (n = 190 data points) and CTCsi (n = 191 data points).
7. IF‐FISH was used to investigate the colocalization of RECQ4 and telomere. RECQ4 (red) in U2OS cells; telomere (green); DAPI (blue). Colocalizations are indicated by arrowheads. Scale bar = 10 μm.
8. Quantification of the RECQ4 and telomere colocalization signal of (G) grouped in NTsh (n = 243 data points) and STN1sh (n = 245 data points).

Data information: The micrographs (A, E, and G) show representative examples from three biological replicates. Data represent mean ± SD of three biological replicates (B–D, and F). ****P < 0.0001, *P < 0.05. One‐way ANOVA with Fisher's LSD tests was used for (C). Unpaired two‐tailed Student's t‐tests were used for (B, D, F, and H).

Source data are available online for this figure.

Furthermore, we examined the localization of RECQ4 foci at telomeres in telomerase‐negative U2OS cells (ALT‐positive cells) following disruption of CTC1 or STN1 (Fig EV2D–H). We found that depletion of CST led to a reduction in RECQ4 colocalization with telomeres (Fig EV2D–H), indicating that CST is required for the incorporation of RECQ4 at telomeres, regardless of the presence of telomerase.

Since CST is known to play a role in telomere overhang processing and overhangs can form G4 structures, we conducted an investigation to examine the impact of RECQ4 knockdown on the overhang signal. Interestingly, our results demonstrated a significant increase in the overhang signal following STN1 knockdown, indicating a disruption in overhang processing. However, no significant change in the overhang signal was observed upon RECQ4 knockdown, suggesting that RECQ4 is not necessary for the proper formation of overhangs (Fig 4E and F). These findings imply that RECQ4's role at telomeres is primarily related to its interaction with CST and the resolution of G4 structures, rather than direct involvement in overhang processing.

Subsequent experiments were performed to determine the dependency of RECQ4's role in telomere maintenance on CST. We examined the telomeric status in single and double knockdown cell lines of STN1 and RECQ4 (Fig 5A and B). Consistent with previous studies and the data presented earlier (Fig 2B–D), individual knockdown of STN1 or RECQ4 resulted in an increase in fragile telomeres and telomere loss (Figs 5C and D, and EV3A–D). Interestingly, no additional increase was observed when both RECQ4 and STN1 were simultaneously knocked down (Figs 5A–D and EV3A and B). This suggests that RECQ4 and CST primarily function together in the same pathway to maintain telomere integrity.

Figure 5. RECQ4 functions at telomeres via its interaction with CST.

• A
  Western blot of RECQ4 and STN1 protein expression in control (NC), RECQ4 knockdown, STN1 knockdown, and RECQ4/STN1 double knockdown HeLa cells.
• B
  Telomere FISH on metaphase spreads in HeLa cells. MTSs (red arrows) and SFEs (yellow arrows). DAPI (blue). Scale bar = 20 μm.
• C, D
  Percentage of chromosomes with MTSs (C) and SFEs (D). Each experiment examines approximately 1,000 chromosomes.
• E
  Representative images of G4 (red) and telomere (green) colocalization in HeLa cells. DAPI (blue). Colocalizations (yellow) are shown in enlarged images and are indicated by arrows. Scale bar = 10 μm.
• F
  Percentage of cells with G4 and telomere colocalization in control (NC) and RECQ4 knockdown and STN1 knockdown or RECQ4/STN1 double knockdown HeLa cells.
• G
  Percentage of cells with G4 and telomere colocalization in HCT116 cells.

Data information: The immunoblots (A) and micrographs (B) show representative examples from three biological replicates. The micrographs (E) show representative examples from four biological replicates. Data represent mean ± SD of three biological replicates (C, D, and G). Data represent mean ± SD of four biological replicates (F). ***P < 0.001, **P < 0.01, *P < 0.05, ns, not significant. One‐way ANOVA with Fisher's LSD tests was used for (C, D, F, and G).

Source data are available online for this figure.

Figure EV3. RECQ4 functions at telomeres via its interaction with CST.

1. Telomere FISH on HCT116 cells in metaphase, showing MTSs (white arrows) and SFEs (yellow arrows). Scale bar = 10 μm.
2. The percentage of chromosomes with MTSs and SFEs in HCT116 cells.
3. Telomere FISH on U2OS cells in metaphase showing MTSs (white arrows) and SFEs (yellow arrows). Scale bar = 20 μm.
4. The percentage of chromosomes with SFEs and MTSs. More than 1,000 chromosomes are analyzed in each experiment.
5. Representative images of D1 (red) and telomere (green) colocalization in HeLa cells. DAPI (blue). Colocalizations are indicated by arrowheads. Scale bar = 20 μm.
6. Percentage of cells with D1 and telomere colocalization in control (NC) and RECQ4 knockdown and STN1 knockdown or RECQ4/STN1 double knockdown HeLa cells.

Data information: The micrographs (A, C, and E) show presentative examples from three biological replicates. Data represent mean ± SD of three biological replicates (B, D, and F). **P < 0.01, *P < 0.05, ns, not significant. One‐way ANOVA with Fisher's LSD tests was used for (B, D, and F).

Source data are available online for this figure.

Considering that STN1 deficiency leads to an accumulation of G4 structures at telomeres, we proposed that the interaction between RECQ4 and CST might facilitate G4 unwinding in vivo. To investigate the role of RECQ4 in G‐quadruplex unwinding, we utilized immunostaining with a G4 structure‐specific antibody (BG4) and D1 antibody to detect the colocalization of G4 with telomeres (Figs 5E and EV3E). Our results demonstrated a significant increase in the number of telomeric G4 structures in RECQ4 knockdown and helicase‐dead mutant knock‐in cells compared with the nontargeting control cells (Figs 5E–G and EV3E and F). Subsequently, we examined G4 formation in the double knockdown cell line of RECQ4 and STN1. Interestingly, no additional increase in the number of total or telomeric G4 structures was observed when both RECQ4 and STN1 were depleted (Figs 5E and F, and EV3E and F), indicating that RECQ4 and CST collaborate to resolve G4 structures in vivo.

In summary, our findings underscore the critical role of RECQ4 in telomere maintenance and G4 resolution through its interaction with CST, indicating their interdependence and further supporting the essential nature of their interaction in ensuring proper telomere function.

RECQ4 contributes to genome‐wide DNA replication dependent on CST

Our investigation extended beyond telomeres, as G4 structures are also present in other GC‐rich regions of the genome, and CST is essential for genome‐wide DNA replication (Stewart et al, 2012). To explore the potential genome‐wide role of RECQ4, we conducted various experiments to assess the impact of RECQ4 and STN1/RECQ4 knockdown on genome‐wide DNA replication. We observed a significant increase in double‐stranded DNA breaks and micronuclei formation in RECQ4‐depleted cells indicating genomic DNA damage (Fig EV4A–D). Furthermore, RECQ4 depletion resulted in a marked reduction in cells exhibiting high levels of EdU incorporation (> 40 AFU; arbitrary fluorescence units), indicating impaired conventional double‐strand DNA replication (Figs 6A and B, and EV4E). Re‐expression of sh‐resistant RECQ4 rescued the reduced number of nuclei with high EdU uptake in RECQ4‐knockdown cells (Fig 6A and B), suggesting a crucial role for RECQ4 in general DNA replication. In STN1/RECQ4 double knockdown cells, the number of EdU‐positive cells with > 40 AFU did not exhibit an additional decrease compared with single knockdown cells (Fig 6A and B), indicating that RECQ4's function in genomic DNA replication is dependent on the presence of CST. We quantified the average number of G4 structures per nucleus in our study and found that the depletion of RECQ4 or STN1 led to an increase in the number of G4 structures in cells. To further explore the relationship between RECQ4 and STN1 in the context of genome‐wide G4 structures, we examined the average number of G4 structures in STN1/RECQ4 double knockdown cells. Surprisingly, we did not observe an additional increase in the number of G4 structures in the double knockdown cells compared with the single knockdown cells (Fig 6C). This suggests that RECQ4's function in regulating genomic G4 structures is also dependent on the presence of CST, similar to its role in telomeres. Consistent with this, the number of EdU‐positive cells decreased by 50% upon treatment with PDS, which induces G4 formation (Fig 6D). Furthermore, the re‐expression of RECQ4‐WT partially rescued genomic DNA replication, whereas the re‐expression of the helicase‐dead RECQ4 mutant did not (Fig 6D), suggesting that RECQ4 may contribute to genome‐wide DNA replication by unwinding G4 structures through its helicase activity.

Figure EV4. RECQ4 contributes to genome‐wide DNA replication dependent on CST.

1. IF staining for 53BP1 (green) per nucleus (blue). Scale bar = 10 μm.
2. The percentage of cells containing 53BP1 foci.
3. Representative images of cells with (bottom) and without (top) micronuclei. Scale bar = 10 μm.
4. Micronuclei quantification.
5. FACS analysis of DNA content in NTsh and RECQ4‐knockdown cells.

Data information: The micrographs (A and C) show representative examples from three biological replicates. Data represent mean ± SD of three biological replicates (B and D). *P < 0.05. unpaired two‐tailed Student's t‐tests were used for (B and D).

Source data are available online for this figure.

Figure 6. RECQ4 contributes to genome‐wide DNA replication dependent on CST.

1. EdU incorporation in NTsh, RECQ4sh, STN1sh, and STN1sh/RECQ4si cells as indicated. EdU (red); DAPI (blue). Scale bar = 250 μm.
2. EdU uptake quantification using mean fluorescence intensity (AFU). Each bar represents the percentage of nuclei that fall within the specified AFU range. AFU, arbitrary fluorescence units. Nuclei with less than 10 AFU are considered EdU negative, while those more than 10 AFU are considered EdU positive.
3. Average number of G4 in control (NC) and RECQ4 knockdown and STN1 knockdown or RECQ4/STN1 double knockdown HeLa cells.
4. Percentage of cells that have more than 40 AFU EdU intensity with or without PDS treatment. Each bar represents the percentage of nuclei fall within the indicated AFU range. AFU: arbitrary fluorescence units.

Data information: The micrographs (A) show representative examples from three biological replicates. Data represent mean ± SD of three biological replicates (B and D). Data represent mean ± SD of four biological replicates (C). ***P < 0.001, **P < 0.01, *P < 0.05, ns, not significant. One‐way ANOVA with Fisher's LSD tests was used for (B–D).

Source data are available online for this figure.

RECQ4 collaborates with CST to respond to G4‐induced stalled replication forks

To investigate the impact of RECQ4 and CST on DNA replication, particularly when encountering different replicational barriers, DNA fiber analysis was performed to examine replication events at the single‐molecule level. HeLa cells were labeled with IdU for 30 min and then treated with PDS (50 μM) for 2 h to induce G4s or HU (2 mM) for 2 h to induce fork stalling (Fig 7A). Subsequently, the cells were released into the media containing CldU for 30 min, harvested, and lysed on microscope slides (Fig 7A). DNA tracks labeled with IdU and CldU were stained with IdU and CIdU antibodies and visualized using confocal microscopy. Replication events were recorded and quantified (Figs 7B and C, and EV5A). In the absence of PDS, depletion of RECQ4 or STN1 did not show any differences in red‐only tracks (fork stalling), green‐only tracks (origin firing), or red‐green tracks (restarted forks) compared with the nontargeting control. However, PDS treatment resulted in a significant decrease in the number of elongating forks (red‐green fibers) and an increase in the number of stalled replication forks (red‐only tracks) in RECQ4 or STN1‐depleted cells, indicating the involvement of both RECQ4 and STN1 in DNA replication in response to G4 formation (Figs 7C and EV5A and B). To further clarify whether RECQ4 functions through G4 unwinding rather than solely responding to replication fork stalling induced by PDS, we utilized HU as a control condition that induces replication fork stalling without affecting G4 structure formation. HU treatment stimulated stalled replication forks (red only), but no additional effects were observed in RECQ4 or STN1 knockdown cells (Figs 7C and EV5A and B). Furthermore, the double knockdown of STN1/RECQ4 did not lead to additional changes (Figs 7C and EV5A and B). Hence, we concluded that while STN1 primarily regulates origin firing in the context of conventional replication stress (induced by HU), RECQ4 likely collaborates with STN1 to restart replication forks specifically in response to G4‐induced replication stalling (induced by PDS).

Figure 7. RECQ4 collaborates with CST to respond to G4‐induced stalled replication forks.

1. Experimental approach diagram and types of DNA fibers scored. To cause replication fork stalling, HeLa cells were pulse labeled with IdU and CldU, with or without PDS.
2. Representative images of DNA fibers in NTsh, RECQ4sh, STN1sh, and STN1sh/RECQ4si cells treated with or without PDS. IdU (Red); CldU (green). Scale bar = 10 μm.
3. Graph illustrating the percentage of red‐green tracts (replication elongation). Five hundred fibers were counted for each condition.
4. The impact of G4 structure formation on the forward movement of replication forks toward telomeres was assessed. Stalled telomeres are identified by the absence of colocalization between the labeled DNA and telomeres. Partially replicated telomeres display partial colocalization, indicating replication stalling within the telomere region. Completely replicated telomeres are visualized when the labeled DNA completely covers the entire telomere region. Replication tracts labeled with IdU are depicted in red, while telomere tracts are highlighted in green. Scale bar = 5 μm.
5. Quantification of telomere replication forks that have been arrested, partially replicated, or completely replicated.
6. The fluorescence intensity of RECQ4 in each nucleus was displayed as dot plots grouped in NTsh (n = 222), STN1sh (n = 288), NTsh PDS (n = 247), STN1sh (n = 226), NTsh PDS ATRi (n = 288), STN1sh PDS ATRi (n = 247), NTsh PDS ATMi (n = 169), STN1sh PDS ATMi (n = 226), as indicated.
7. Model of how RECQ4 is recruited to stalled forks in response to G4 stabilization.

Data information: The micrographs (B) show representative examples from three biological replicates. Data represent mean ± SD of three biological replicates (C, E, and F). ****P < 0.0001, **P < 0.01, *P < 0.05, ns, not significant. One‐way ANOVA with Fisher's LSD tests was used for (C, E, and F).

Source data are available online for this figure.

Figure EV5. RECQ4 collaborates with CST to respond to G4‐induced stalled replication forks.

1. Total number of tracks scored. In parenthesis is the percentage of the total number of tracks scored.
2. The percentage of tracts that are red only (fork stalling) is represented graphically. Data represent mean ± SD of three biological replicates. **P < 0.01. One‐way ANOVA with Fisher's LSD tests was used for (B).
3. Chromatin‐bound RECQ4 were detected in pre‐extracted HeLa cells. RECQ4 (green), DAPI (blue). Scale bar = 20 μm. The micrographs show representative examples from three biological replicates.

Source data are available online for this figure.

Telomere replication poses inherent challenges due to its G‐rich and repetitive nature, rendering them prone to replication fork stalling. Fiber combined with telomere FISH was performed in HeLa cells to determine whether the formation of G4 structures impedes the forward movement of replication forks toward telomeres, examining completely replicated telomeres, partially replicated telomeres, or stalled telomeres (Fig 7D). Disrupting RECQ4 in HeLa cells led to a twofold increase in stalled replication forks in the telomere region and an approximately 30% increase in partially replicated telomeric DNA, consistent with the effect of STN1 depletion on telomere replication deficiency. This suggests that both RECQ4 and STN1 contribute to maintaining normal telomere DNA replication (Fig 7E). Moreover, the double knockdown of RECQ4 and STN1 did not further increase the number of stalled telomeric replication forks beyond the levels observed with RECQ4 knockdown alone, indicating that RECQ4 likely aids in telomere DNA replication through its interaction with CST (Fig 7E).

CST promotes RECQ4 recruitment to chromatin in response to G4 stabilization

We examined the localization of RECQ4 to genome‐wide chromatin by harvesting cells and removing soluble proteins before measuring the fluorescence intensity of chromatin‐bound RECQ4 levels (Fig EV5C). As anticipated, the levels of chromatin‐bound RECQ4 were significantly reduced in shSTN1 cells compared with nontarget control cells (Fig 7F), confirming that STN1 depletion diminishes chromatin‐bound RECQ4. Moreover, when cells were treated with PDS, the levels of chromatin‐bound RECQ4 increased approximately twofold (Fig 7F), indicating that G4 stabilization stimulates RECQ4 recruitment to DNA. Importantly, this enrichment was abolished by STN1 depletion, suggesting that the association of RECQ4 with chromatin requires CST and further supporting the idea that CST facilitates RECQ4 to resolve G4s (Fig 7F).

Previous studies have shown that PDS induces G4 formation and replication fork stalling, leading to the activation of the DNA damage response (McLuckie et al, 2013; Zimmer et al, 2016). To investigate whether inhibiting the DNA damage response through ATM or ATR affects RECQ4 binding to DNA, we added an ATM inhibitor (KU55933) or an ATR inhibitor (ETP‐46464) to PDS‐treated cells for 24 h and measured the levels of chromatin‐bound RECQ4 (Fig EV5C). Interestingly, we found that treatment with the ATR inhibitor restored RECQ4 enrichment on PDS‐stimulated chromatin, while the ATM inhibitor had no significant effect (Fig 7F). Furthermore, when STN1 was depleted prior to ATR inhibition, RECQ4 recruitment to DNA was reduced (Fig 7F), suggesting that ATR regulates RECQ4 recruitment to DNA in a CST‐dependent manner.

Discussion

Previous studies have demonstrated the role of mammalian CST in facilitating telomere replication and maintaining genome integrity through G4 unfolding. However, the specific mechanism by which CST resolves these structures in vivo and whether other factors are involved have remained unclear. In our study, we have made advancements in understanding this process. We have discovered that CST plays a crucial role in engaging RECQ4 with chromatin, including telomeres, to effectively resolve G4 structures and restart stalled replication forks (Fig 7G). This finding highlights the collaborative nature of CST and RECQ4 in resolving G4s and preserving genome and telomere stability.

CST has been shown to facilitate efficient recovery of telomere and genomic replication by activating dormant origins (Stewart et al, 2012; Wang et al, 2014). Previous studies have demonstrated that purified mammalian CST can melt G4 structures in a concentration‐dependent, self‐displacement manner (Bhattacharjee et al, 2017; Zhang et al, 2019). However, resolving G4 structures in the context of the cellular environment may be more challenging due to factors such as supercoiling and torsional stress within genomic DNA. As a single‐strand binding protein, CST may lack the ability to actively unwind the quadruplex structure like a helicase. Additionally, similar to RPA, once the G4 structures are resolved, CST needs to be removed for fork restart or other processing events to occur. Therefore, it is likely that CST recruits and cooperates with other factors to effectively resolve G4 structures in vivo. These additional factors may provide the necessary helicase activity or other functions required for efficient G4 resolution.

In our study, we made the discovery that CST interacts with RECQ4 in vitro, establishing a direct interaction between these two proteins (Fig 1). RECQ4, as one of the five RecQ helicases found in human cells, has been shown to bind to G4 structures but not actively unwind them (Keller et al, 2014). However, our more sensitive FRET assay revealed that the addition of ATP enabled RECQ4 to actively unwind the intramolecular G4 structures. Furthermore, the presence of CST significantly enhanced the reaction kinetics (Fig 3C), indicating that RECQ4 can effectively resolve G4 structures in vitro and that this activity can be stimulated by the CST complex.

In our study, an accumulation of G4 structures was observed upon depletion of RECQ4, which is consistent with its documented capacity to bind and resolve G4 structures in vitro. However, it is important to acknowledge the potential involvement of RECQ4 in processing stalled replication forks induced by G4 stabilizers. Nevertheless, our findings suggest that this role may have a relatively modest impact, as no significant perturbation was observed in DNA replication or replication restart upon RECQ4 depletion under conditions of replication stress induced by hydroxyurea (HU) treatment, which is known to induce replication fork stress. Therefore, our proposed hypothesis is that RECQ4 primarily functions in unwinding G4 structures during this process.

In addition to its role in G4 resolution, RECQ4 is known to play an essential role in telomere maintenance (Croteau et al, 2014). Consistently, our study showed that depletion of RECQ4 led to significant telomeric abnormalities, such as multiple telomere FISH signals and signal‐free end telomeres (Fig 2B–D). Furthermore, RECQ4 interacted with STN1 to facilitate the restart of replication forks stalled by G4 structures (Fig 7C). Other DNA helicases, such as WRN and BLM, have been reported to resolve telomeric G4 structures and promote telomere replication by interacting with shelterin components TRF2 and TRF1 (Opresko et al, 2002; Zimmermann et al, 2014; Drosopoulos et al, 2015). RTEL1 and DNA2 helicases also contribute to G4 resolution to protect telomere integrity by preventing telomere loss and fragility (Vannier et al, 2012; Lin et al, 2013). Although RECQ4's G4 unwinding function may not be as robust as that of other helicases (Keller et al, 2014), our findings suggest that CST may stimulate RECQ4's activity to improve G4 unwinding by unknown reasons, thereby contributing to telomere maintenance and genome stability.

Moreover, upon analyzing our data, we observed a synergistic increase in the formation of telomere defects, specifically signal‐free ends, in the STN1‐RECQ4 double knockdown cell lines. Although the difference is not statistically significant, the trend strongly suggests that RECQ4 and CST individually contribute to telomere dynamics beyond their cooperative role in G4 resolution. Furthermore, our telomere combing analysis revealed a significant rise in the percentage of stalled telomeres in RECQ4‐depleted cells compared with STN1 knockdown cell lines, indicating that RECQ4 may influence telomere replication fork dynamics independently of STN1. However, when considering the overall impact of single knockdown or double knockdown of STN1 and RECQ4 on telomere defects and G4 formation, we observed comparable effects, which supports the notion that their primary functional pathway involves the regulation of G4 formation and the subsequent response to telomere damage repair.

CST, similar to RPA, functions as a crucial player in DNA repair and synthesis processes in eukaryotic cells (Chen & Wold, 2014). It shares structural and functional similarities with RPA, including DNA‐binding activity and the ability to stimulate DNA polymerases and helicases (Bhattacharjee et al, 2017). However, CST has a more specialized role in G‐rich DNA. In contrast, RECQ1 has been reported to regulate the RPA‐mediated replication stress response for maintaining genome homeostasis (Banerjee et al, 2015). In our study, we observed that RECQ4 is recruited to chromatin via CST in response to telomeric G4 structures and genome‐wide replication stress (Fig 7G). Furthermore, CST is involved in recruiting polα for C‐strand synthesis, as well as RAD51 for homologous recombination and alternative telomere lengthening (Wang et al, 2012; Chastain et al, 2016; Huang et al, 2017). Thus, we propose that CST, similar to RPA, acts as a central hub for recruiting specific protein factors to DNA, facilitating additional processing steps at G4 structures and stalled replication forks in cells.

Consistent with this notion, we found that RECQ4's binding to chromatin depends on STN1 (CST) and ATR signaling, particularly in response to G4 stabilization (Fig 7F). Notably, ATR also regulates CST's interaction with RAD51 following replication stress, suggesting a connection between CST and the ATR‐CHK1 pathway (Chastain et al, 2016; Ackerson et al, 2020). However, we could not determine whether the ATR pathway directly regulates the interaction between RECQ4 and DNA or RECQ4 and CST.

Mutations in CST lead to Coats plus syndrome, a condition that shares clinical similarities with Rothmund–Thomson syndrome (RTS) type II, caused by RECQ4 mutations. Both disorders are associated with telomere dysfunction and premature aging (Anderson et al, 2012; Simon et al, 2016). Our findings highlight the requirement of CST and RECQ4 for the stability of GC‐rich repetitive sequences, such as telomeres. Exploring whether mutations in CTC1 or STN1 occur in Coats plus patients, affecting the recovery of stalled replication forks through CST recruitment together with RECQ4, may shed light on the molecular mechanisms underlying these diseases and contribute to the development of effective therapeutic approaches for Coats plus or RTS.

Materials and Methods

Cell culture and transfections

RPMI 1640 (Corning, USA) was used to culture HeLa and HCT116 cells. In DMEM (Corning), U2OS and HEK293T cells were grown. Grace's Insect Medium (Gibco, USA) was used to culture sf9 cells. The lentivirus viral shRNA sequences used for cell transduction are STN1 shRNA 5′‐CAAGGCAATTCATAGTATA‐3′, and the rescue cell line was created as previously reported (Stewart et al, 2012). The lentivirus viral shRNA sequences used for cell transduction are RECQ4 shRNA (5′‐CCTCGATTCCATTATCATT‐3′) and Scramble shRNA (5′‐CCTAAGGTTAAGTCGCCCTCG‐3′). The sequences of siRNA used are as follows: siNC (negative control): 5′‐UUCUCCGAACGUGUCACGUTT‐3′; siRECQ4: 5′‐CCUCGAUUCCAUUAUCAUUdTdT‐3′; siCTC1: 5′‐GAAAGUCUUGUCCGGUAUUdTdT‐3′. The primer sequences used for qPCR are RECQ4‐F: 5′‐TGGGCAAAGTGAAGAAGGG‐3′, RECQ4‐R: 5′‐AGGGAGTCAAGGGCGAAGG‐3′; GAPDH‐F: 5′‐GCCACATCGCTCAGACAC‐3′; GAPDH‐R: 5′‐GCCCAATACGACCAAATCC‐3′.

Transfections were carried out in the Opti‐MEM medium (Gibco, USA) using polyethyleneimine (PEI) (Polysciences, USA) with a DNA:PEI ratio of 1:6. HEK293T cells were cultured in a 10‐cm dish with 10 μg of expressing vectors, 7.5 μg of pSPAX2, and 2.5 μg of pMD2G. The supernatant was collected in preparation for viral infection.

HEK293T cells were treated for 24 h with 10 μM pyridostatin (PDS; Sigma‐Aldrich) or 2 mM hydroxyurea (HU; Sigma‐Aldrich) for co‐IP experiments. G4s were formed in HeLa cells treated for 24 h with 10 or 50 μM TMPyP4/PDS. To inhibit ATM or ATR, 10 μM KU55933 (Selleck Chemical) or 2 μM ETP‐46464 (Sigma) was used, respectively.

Antibodies

The following antibodies were used: anti‐RECQ4 (SDIX, T00417‐A1), anti‐TRF2 (Millipore, clone4A794), anti‐GAPDH (Affinity, AF0911), anti‐HA‐tag (Cell Signaling Technology, 3724), anti‐FLAG‐tag (Sigma, F1804), anti‐FLAG‐tag (Cell Signaling Technology, 2368S), anti‐OBFC1 (Abcam, ab89250), anti‐OBFC1 (Abcam, ab251855), anti‐β‐Tubulin (Abclonal, AC008), anti‐His‐tag (Proteintech, 66005), anti‐GST‐tag (Proteintech, 10000‐0‐AP), anti‐Myc‐tag (Cell Signaling Technology, 2276S), Alexa Fluor 488 (Thermo Fisher, A11029, A11034), Alexa Fluor 555 (Thermo Fisher, A‐21428, A‐21422).

Silver staining and mass spectrometry

FLAG‐empty‐vector or FLAG‐tagged‐STN1 were transfected into HEK293T cells in an instant. After that, the cell lysate was incubated with anti‐FLAG beads (Sigma). Following binding, the beads were washed gently with cold NP‐40 buffer containing 50 mM Tris–HCl (PH8.0), 150 mM NaCl, and 1% NP‐40 supplemented with protease inhibitors. Elution buffer containing 0.2 mg/ml FLAG peptide (Sigma) was used to elute the FLAG protein complex. SDS–PAGE was used to separate the eluted protein samples, which was then stained with silver. Specific band excision for mass spectrum sequencing and data analysis (Bai et al, 2016).

Western blotting and co‐immunoprecipitation (co‐IP)

Cell extracts were harvested using RIPA lysis buffer containing protease inhibitors, electrophoresed on SDS–PAGE gel, and then transferred onto PVDF membranes (Millipore). Membranes were washed with TBST buffer containing 0.1% Tween‐20 and incubated with secondary antibodies after being blocked with 5% nonfat milk and incubated with specific antibodies overnight at 4°C. Finally, the antibodies were immunodetected using an enhanced chemiluminescence system (Millipore). Cells were lysed on ice for 30 min with an NP‐40 buffer containing 50 mM Tris–HCl (PH8.0), 150 mM NaCl, and 1% NP‐40 supplemented with protease inhibitors for co‐immunoprecipitation (co‐IP) assay. The lysates were then centrifuged at 4°C for 20 min at 13,523 g. At 4°C overnight, the protein supernatant was mixed with anti‐FLAG beads (Sigma). After washing the beads five times in NP‐40 buffer, they were boiled in SDS loading buffer and analyzed by western blotting. ImageJ was used to calculate the relative protein levels.

Recombinant proteins purification and GST pulldown

STN1 and TEN1 with 6 × His‐tag were cloned in PET28a and purified with Ni‐beads from E. coli. CTC1 with FLAG‐tag, RECQ4 with FLAG‐tag, and RECQ4 with His‐tag were purified from infected Sf9 cells. GST‐tagged RECQ4 were cloned pGEX‐6P‐1 and expressed in BL21, including GST‐RECQ4 (aa 1–449), GST‐RECQ4 (aa 449–830), and GST‐RECQ4 (aa 830–1,208).

Purified GST‐tagged truncated forms of RECQ4 fusion proteins immobilized on glutathione agarose were incubated in PBS binding buffer with equal amounts of FLAG‐tagged CTC1, His‐tagged STN1, or His‐tagged TEN1. Purified FLAG‐tagged CTC1 proteins immobilized on FLAG agarose beads were incubated in PBS binding buffer with an equal amounts of His‐tagged CTC1. Purified FLAG‐tagged RECQ4 proteins immobilized on FLAG agarose were incubated in PBS binding buffer with an equal amount of His‐tagged STN1 or TEN1.After that, agarose beads were added to the binding reaction and mixed for 3 h at 4°C. The beads were washed five times with binding buffer before boiled in SDS loading buffer to release the bound proteins for SDS–PAGE separation and western blotting.

Telomere FISH and 53BP1 staining

FISH was performed on methanol/acetic acid fixed cells using a TelC‐Alexa 488 PNA G‐strand probe (5′‐CCCTAACCCTAACCCTAA, Panagene) or TelG‐Cy3 PNA C‐strand probe (5′‐GGGTTAGGGTTAGGGTTA, Panagene) (Feng et al, 2017). Multiple Telomere Signals (MTSs) and Signal Free Ends (SFEs) were displayed and manually analyzed. TFL‐TELO was used to measure telomere fluorescence intensity for quantitative telomere length (qFISH) (Poon et al, 1999).

Cells were harvested for combined 53BP1 and telomere staining, fixed with 2% paraformaldehyde, and permeabilized in 0.5% Triton X‐100 at room temperature before being dehydrated with an ethanol series. Hybridization was carried out for 3 min at 85°C in the presence of a Cy3‐labeled CCCTAA PNA probe (Panagene, KR). The coverslips were then washed and blocked for 1 h in PBG buffer containing 0.1% gelatin and 0.1% bovine serum album. Anti‐53BP1 antibody was used to stain the slides for 1 h. After washing with PBST, the slides were incubated for 0.5 h with an Alexa Fluor 488 secondary antibody before being dehydrated in 70, 90, and 100% ethanol. DAPI was used to adhere the coverslips (Life Technologies). Fluorescence was detected using a Nikon fluorescence microscope.

Immunofluorescence‐fluorescence in situ hybridization (IF‐FISH)

Cells were harvested and gently extracted (1 M HEPES, 1.5 M sucrose, 1 M MgCl₂, 1 M NaCl, and NP‐40) before being washed four times with PBST and fixed in 2% paraformaldehyde. Coverslips were then treated with 1 mg/ml pepsin, permeabilized in 0.5% Triton X‐100, and treated with 50 μg/ml RNase overnight at 37°C before being dehydrated with an ethanol series and hybridized with a TelC‐Alexa 488 PNA G‐strand probe, as described above. A BG4 antibody was used for G4s staining and incubated on the coverslips for 1 h. The coverslips were washed four times before being incubated for 1 h at room temperature with a FLAG‐tag (MA4) mouse monoclonal antibody. Finally, the coverslip was washed with PBST, probed for 0.5 h at room temperature with an Alexa Fluor 555 secondary antibody, and dehydrated using an ethanol series. DAPI was used to adhere the coverslips (Life Technologies). A Nikon fluorescence microscope was used to detect fluorescence.

Co‐IF

Cells were grown in chamber slides and treated with 1 × CSK buffer (1 M Hepes, 1 M NaCl, 1 M MgCl₂, 0.75 M sucrose, 0.1% Triton X‐100) on ice. Coverslips were fixed with 100% methanol on ice, then blocked in PBG buffer containing 0.1% gelatin and 0.1% bovine serum album. Coverslips were incubated with RECQ4 and TRF2 antibodies, which were then diluted in PBG buffer at a ratio of 1:1,000. Finally, the coverslip was washed with PBST, probed with an Alexa Fluor 555 secondary antibody, and dehydrated using an ethanol series, as described above. The coverslips were stained with DAPI (Life Technologies). Fluorescence was detected using a Nikon fluorescence microscope.

Fluorescence resonance energy transfer (FRET)

Tel21 DNA was marked at the 5′‐end with fluorescein (FAM) and the 3′‐end with tetramethylrhodamine (TAMRA). Takara Biotech (Dalian, China) supplied the samples. FRET was carried out as described (Mergny & Maurizot, 2001), and 1 M of DNA was mixed with 25 or 100 nM of protein in a buffer with 150 mM NaCl and 10 mM Tris (pH 7.4) for 2 h at 4°C Fluorescence was looked at and measured. At 25°C, the Spex Fluorolog‐3 spectrofluorometer (HORIBA Jobin, France) was used to measure the fluorescence. The slits for both excitation and emission were 5 nm. The excitation was set to 480 nm, and the emission was measured between 490 and 650 nm. Using the reference channel, changes in the lamp were fixed. For each spectrum, a buffer blank was taken away.

Analysis of genomic DNA replication rates

The uptake of EdU revealed the rates of genome replication. Cells were stained with a Click‐it Alexa Fluor 555 EdU Imaging Kit (Invitrogen, C10338) after 30 min of incubation with 10 μM EdU, as previously reported (Stewart et al, 2012; Zhang et al, 2019).

DNA fiber analysis

Before harvesting, HeLa cell lines were plated and labeled with IdU for 30 min. PDS (50 μM) G4s was used to induce G4s for 2 h, after which the cells were washed and released into media containing CldU for 30 min before harvested and lysed on microscope slides. The DNA fibers were then prepared in the manner previously described (Stewart et al, 2012).

Cell cycle analysis

PBS was used to wash the cells after they were harvested. Cells were washed and resuspended in PBS containing 50 μg/ml RNase A and 100 μg/ml propidium iodide (PI) after being fixed in ice‐cold 70% ethanol. A FACS Calibur was used to determine cell cycle distribution.

G‐overhang analysis

DNA was purified and restriction digested then briefly run on 1% nondenaturing agarose gels. Gels were hybridized with a 32P‐labeled (AT2C3)₄ probe to the G‐overhang after drying. Finally, gels were then denatured to rehybridized with the same probe and the signal used to normalize for gel loading. The G‐overhang analysis was performed as previously described (Wang et al, 2012).

Statistical analysis

All experiments were repeated three times or more. The error bars represent the mean ± SD. P‐values were calculated using the two‐tailed test or one‐way ANOVA with Fisher's LSD tests, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. GraphPad was used to analyze the data.

Author contributions

Tingfang Li: Conceptualization; investigation; writing – original draft. Miaomiao Zhang: Conceptualization; investigation. Yanjing Li: Conceptualization; investigation. Xinyu Han: Investigation. Lu Tang: Investigation. Tengfei Ma: Investigation. Xiaotong Zhao: Investigation. Rui Zhao: Investigation. Yuwen Wang: Investigation. Xue Bai: Investigation. Kai Zhang: Investigation. Xin Geng: Investigation. Lei Sui: Investigation. Xuyang Feng: Investigation. Qiang Zhang: Investigation. Yang Zhao: Investigation. Yang Liu: Conceptualization; investigation; writing – review and editing. Jason A. Stewart: Conceptualization; investigation; writing – review and editing. Feng Wang: Conceptualization; supervision; funding acquisition; investigation; writing – review and editing.

Disclosure and competing interests statement

The authors declare that they have no conflict of interest.

Supporting information

Appendix

Expanded View Figures PDF

Source Data for Expanded View

PDF+

Source Data for Figure 1

Source Data for Figure 2

Source Data for Figure 3

Source Data for Figure 4

Source Data for Figure 5

Source Data for Figure 6

Source Data for Figure 7

EMBO reports (2023) 24: e55494

Data availability

No primary datasets have been generated or deposited.