Cytosolic DNA sensing by cGAS/STING promotes TRPV2-mediated Ca²⁺ release to protect stressed replication forks

SUMMARY

The protection of DNA replication forks under stress is essential for genome maintenance and cancer suppression. One mechanism of fork protection involves an elevation in intracellular Ca²⁺ ([Ca²⁺]_i), which in turn activates CaMKK2 and AMPK to prevent uncontrolled fork processing by Exo1. How replication stress triggers [Ca²⁺]_i elevation is unclear. Here, we report a role of cytosolic self-DNA (cytosDNA) and the ion channel TRPV2 in [Ca²⁺]_i induction and fork protection. Replication stress leads to the generation of ssDNA and dsDNA species that, upon translocation into cytoplasm, trigger the activation of the sensor protein cGAS and the production of cGAMP. The subsequent binding of cGAMP to STING causes its dissociation from TRPV2, leading to TRPV2 derepression and Ca²⁺ release from the ER, which in turn activates the downstream signaling cascade to prevent fork degradation. This Ca²⁺-dependent genome protection pathway is also activated in response to replication stress caused by oncogene activation.

eTOC BLURB

Fork protection under replication stress requires [Ca²⁺]_i elevation, but the underlying molecular mechanism is unknown. Li, Kong et al. discovered that cGAS/STING-mediated sensing of cytosolic self-DNA produced after replication stress triggers TRPV2-dependent Ca²⁺ release from the ER, leading to the activation of the downstream CaMKK2-AMPK-Exo1 signaling cascade for fork protection.

Graphical Abstract

INTRODUCTION

DNA replication stress represents a major source of mutations and genomic instability with key implications for cancer, developmental disorders and other human diseases ^1,2. Replication stress can be induced by environmental factors, such as radiation, genotoxins and cancer drugs, and by intracellular factors, such as difficult-to-replicate DNA regions, certain metabolic byproducts and collisions with the transcription machinery ^1,3. Oncogene activation also frequently induces replication stress during tumorigenesis ^4,5. The challenges caused by these factors necessitate robust mechanisms to protect the replication fork structure to avoid fork collapse and DNA damage. A number of tumor suppressors such as BRCA1, BRCA2 and FANC proteins have been shown to play an important role in fork protection through their functions at replication forks in controlling nucleases such as Mre11, Exo1 and Dna2 to prevent abnormal processing of fork DNA ⁶. Defects in fork protection in cancer cells lacking these tumor suppressors confer sensitivity to chemotherapies such as PARP inhibitors ^7,8. Conversely, mechanisms that restore fork protection in those cancer cells are associated with chemoresistance and poor clinical outcome ^9,10.

In addition to fork-associated factors, cells also activate key signaling pathways for fork protection. These include the ATR/Chk1-dependent replication checkpoint and a newly identified Ca²⁺-dependent signaling pathway, both of which suppress Exo1 function to prevent aberrant fork resection ^11–13. In the Ca²⁺-dependent fork protection pathway, replication stress induces Ca²⁺ release from intracellular stores, resulting in an elevation in intracellular Ca²⁺ ([Ca²⁺]_i), which in turn activates the Ca²⁺/CaM-dependent protein kinase CaMKK2 and the downstream kinase AMPK. Following activation, AMPK then directly phosphorylates Exo1 at S746, leading to the binding of 14–3-3 proteins and inhibition of the recruitment of Exo1 to replication forks. As a result, aberrant fork processing is avoided. Disruption of this signaling pathway causes elevated chromosomal instability and reduced cell viability in the presence of replication stress ¹³. However, despite these findings, it is unclear how [Ca²⁺]_i induction occurs upon replication stress and how the stress signal is transmitted from the nucleus to the cytoplasm to promote Ca²⁺ release and fork protection.

In our effort to address these fundamental questions, we have uncovered a role for TRPV2 and cytosolic self-DNA (cytosDNA) in [Ca²⁺]_i induction and fork protection under replication stress. An evolutionarily conserved and ubiquitously expressed Ca²⁺-selective channel, TRPV2 is localized on both the ER and plasma membrane and regulates a range of cellular processes, including thermosensation, osmosensation, mechanosensation, and phagocytosis ^14–16. Dysregulation of TRPV2 has been associated with multiple cancers such as bladder cancer, prostate cancer and glioma, although the exact role of TRPV2 in tumorigenesis remains to be defined ^17,18. Previous studies suggest that replication stress, like ionizing radiation, causes the generation of cytosDNA, which in turn can induce type I interferon and inflammatory responses ^19–22. The cGAS-STING axis plays a key role in sensing cytosolic DNA and activating the innate immune response ^23,24. The direct binding of cGAS to cytosolic DNA causes the activation of its catalytic activity, leading to the production of cGAMP from ATP and GTP ^25–27. The subsequent binding of cGAMP to STING, which is also localized on the ER, results in its conformational changes and translocation to Golgi to activate TBK1 and downstream innate immune response ^28–36. In addition to its role in the innate immune response, cGAS acts to restrain DNA replication speed and homologous recombination-mediated DNA repair, but these functions are apparently mediated by the nuclear fraction of cGAS and independent of its catalytic activity ^37–39. cGAS has also been shown to protect replication forks from excessive resection after replication stress, although the underlying mechanism has not been defined ³⁹. Results described in the present study indicate that cGAS prevents excessive fork resection by Exo1 after replication stress through the production of cGAMP and subsequent [Ca²⁺]_i release from the ER. Interestingly, unlike cGAS, STING plays an opposing role in Ca²⁺-dependent fork protection pathway, highlighting the mechanistic differences between the genome maintenance and innate immune signaling pathways induced by cytosolic DNA.

RESULT

TRPV2 promotes fork protection and chromosomal stability under replication stress

We have previously shown that replication stress induces an elevation in [Ca²⁺]_i, which in turn activates CaMKK2 and the downstream kinase AMPK to promote replication fork protection and chromosomal stability ¹³. This [Ca²⁺]_i elevation is apparently a result of Ca²⁺ release from intracellular stores ¹³. However, how [Ca²⁺]_i elevation is induced by replication stress remains unknown. To address this question, we first set out to identify ion channel(s) that are responsible for [Ca²⁺]_i induction after replication stress. Given that ER is the main Ca²⁺ store in cells, we first examined the potential involvement of IP3Rs and RYRs, two major Ca²⁺ channels on the ER that control Ca²⁺ release into the cytoplasm ⁴⁰. As shown in Figure S1A, inhibition of IP3Rs or RYRs in a HeLa cell line stably expressing the [Ca²⁺]_i reporter GCaMP6s (whose fluorescence signal correlates with the [Ca²⁺]_i level) did not affect the elevation in [Ca²⁺]_i after HU treatment, suggesting that these channels are not required for [Ca²⁺]_i induction after replication stress. We noted that TRPM2, a member of Transient Receptor Potential (TRP) ion channel family, is activated by ADP-ribose, which is generated after genotoxic stress ^41,42. This raises the possibility that TRPM2 may contribute to Ca²⁺ release after replication stress. To determine whether TRPM2 or other members of TRP channels are responsible for Ca²⁺ release for fork protection, we conducted a siRNA screen on all the 28 members of the human TRP family ^43,44. A nondenaturing BrdU staining assay was used to measure the effects of the depletion of individual TRP channels on fork resection in HU-treated HeLa cells ¹³, with the rationale that Ca²⁺ channels important for fork protection are likely involved in the Ca²⁺-dependent signaling pathway described above (Figure 1A). As shown in Figure 1B, while siRNAs targeting the majority of the TRP channels (including TRPM2) did not exhibit obvious effects, siRNAs targeting TRPV2, TRPP3 or TRPM5 caused a significant increase in BrdU signal after HU treatment, suggestive of their involvement in fork protection. Among these three TRP channels, TRPM5 is not permeable to Ca²⁺, and thus cannot be directly responsible for [Ca²⁺]_i induction in the replication stress response ^45,46. We noticed that TRPP3 depletion already caused a higher level of [Ca²⁺]_i in the absence of HU and extensive cell death, suggesting that it may not be involved in the Ca²⁺-dependent pathway for fork protection (Figure S1B). We next focused on TRPV2, a conserved, ubiquitously expressed Ca²⁺ channel that is localized in both the ER and plasma membranes ¹⁷.

Figure 1. TRPV2 promotes fork protection, chromosome stability and cell viability after replication stress.

(A) Experimental scheme of a siRNA screen for identifying fork protection factors using a nondenaturing BrdU IF assay.

(B) Quantified BrdU signal in ssDNA in siRNA-transfected HeLa cells after treatment with HU (2 mM) or H₂O for 4 h. Cells with a BrdU signal higher than the majority (98%) of H₂O-treated control cells (black dots) were taken as BrdU-positive (blue dots). Red bars represent the median BrdU intensity of BrdU-positive cells. 1,000 cells were analyzed for each sample. n=2, ****, p≤0.0001 (unpaired t-test). The same H₂O-treated siControl sample was used for both upper and lower panels.

(C) Result of DNA fiber analysis for measuring fork resection in Hela cells 48 h after siRNA transfection. Upper left panel: Experimental scheme (see Methods). Bottom left panel: Representative images of DNA fibers for the indicated samples. Right panel: dot plot of the ratio of CIdU/IdU track lengths from two independent experiments. Red bars represent the median. 150 tracks were scored for each sample. ****, p≤0.0001 (unpaired t-test).

(D) Result of metaphase chromosome spreading in HeLa cells stably expressing mCherry or mCherry-tagged TRPV2(DN) after HU treatment (4 mM, 6 h). Left panel: Representative images of metaphase chromosome spreads for the indicated samples. Chromosomal aberrations are marked by arrows. Right panel: Quantified result of the samples depicted in the left panel. 150 metaphases from three independent experiments were examined for each sample. ****, p≤0.0001 (unpaired t-test).

(E) Result of alkaline comet assay in HeLa cells stably expressing mCherry or TRPV2(DN)-mCherry after HU treatment (4 mM, 6h). Left panel: Representative images showing the tail moment. Right panel: Quantified tail moment for the samples depicted in the left panel. 150 cells from three independent experiments were examined for each sample. ****, p≤0.0001 (unpaired t-test).

(F) Result of clonogenic analysis in HeLa cells stably expressing mCherry or TRPV2(DN)-mCherry after treatment with indicated concentrations of HU for 24 h. Data represent mean ± S.D. from triplicates. ***, p≤0.001. **, p≤0.01. *, p≤0.05.

We first set out to further validate the role of TRPV2 in fork protection by performing a DNA fiber assay that measures resection of nascent DNA at forks. In this experiment, HeLa cells transfected with TRPV2 siRNAs were pulse-labeled with thymidine analogs IdU and CldU sequentially followed by HU treatment to induce replication stress. Chromosomal DNA was then spread on slides and the IdU (red) and CldU (green) tracks were detected by immunofluorescence staining. A decrease in the ratio of CldU/IdU track length represents fork resection ⁴⁷. Using this assay, we detected a marked increase in fork resection in cells treated with TRPV2 siRNAs, further confirming its role in fork protection (Figures 1C and S1C). Consistent with the role of TRPV2 in fork protection, the result of metaphase chromosome spreading indicates that cells overexpressing a mCherry-tagged (TRPV2(DN)), a dominant negative mutant of TRPV2 that contains two mutations (E599K/E609K) in the pore region of the channel ⁴⁸, exhibited a higher level of chromosomal abnormalities after HU treatment, including breaks and fusions, compared to mCherry-expressing cells (Figures 1D and S1D). The result of comet assay also showed increased tail moment, indictive of a higher level of DNA damage, in TRPV2(DN)-expressing cells after HU treatment (Figure 1E). Furthermore, overexpression of TRPV2(DN) or shRNA-mediated knockdown of TRPV2 also resulted in reduced viability in both HeLa and Molm13 cell lines (Figures 1F, S1E and S1F). Taken together, these data indicate that the Ca²⁺ channel TRPV2 plays a key role in promoting fork protection, chromosomal stability and cell survival after replication stress.

Upon replication stress, TRPV2 mediates Ca²⁺ release from the ER to promote AMPK activation and Exo1 suppression, thereby preventing aberrant fork processing

As a Ca²⁺ channel, TRPV2 may prevent aberrant fork processing after replication stress through the aforementioned Ca²⁺-CaMKK2-AMPK-Exo1 pathway. To determine whether this is the case, we examined the effects of functional disruption of TRPV2 on HU-induced [Ca²⁺]_i elevation in HeLa cells expressing the pan-cell Ca²⁺ reporter GCaMP6s. We found that shRNA-mediated knockdown of TRPV2, overexpression of TRPV2(DN), or inhibition of TRPV2 activity with an inhibitor Tranilast ⁴⁹, all blocked [Ca²⁺]_i elevation in S-phase synchronized HeLa cells after HU treatment (Figures 2A–2C, S2A and S2B). Disruption of TRPV2 function also abrogated HU-induced [Ca²⁺]_i elevation in S phase-synchronized U2OS cells (Figure S2C and S2D). These observations strongly suggest that TRPV2 is indeed a major calcium channel responsible for [Ca²⁺]_i induction after replication stress.

Figure 2. TRPV2 mediates Ca²⁺ release to promote AMPK activation and Exo1 phosphorylation for fork protection.

(A) Effects of shRNA-mediated knockdown of TRPV2 on HU-induced [Ca²⁺]_i elevation. Left panel: Representative images of GCaMP6s signal. Right panel: Quantified GCaMP6s signal in S phase-synchronized HeLa cells after treated with HU (2 mM, 4 h) or H₂O. 250 cells were scored for each sample. Red bars represent the median. n=3, ****, p≤0.0001 (unpaired t-test). The same numbers of experiments performed and cells scored, and statistical analysis apply to Figure 2B–2F below.

(B-C) Effects of TRPV2(DN) expression (B) or Tranilast treatment (50 μM) (C) on HU-induced [Ca²⁺]_i elevation in S phase-synchronized HeLa cells treated with HU (2 mM, 4 h).

(D) Effects of shRNA-mediated knockdown of TRPV2 on HU-induced Ca²⁺ release from the ER. Left panel: Representative images of GCaMPer signal. Right panel: Quantified GCaMPer signal in S phase-synchronized HeLa cells after treated with HU (2 mM, 4 h) or H₂O.

(E-F) Effects of TRPV2(DN) expression (E) or Tranilast treatment (50 μM) (F) on the ER release of Ca²⁺ in S phase-synchronized HeLa cells treated with HU (2 mM, 4 h).

(G-H) Effects of TRPV2 knockdown (G) or TRPV2-DN expression (H) on T172-phosphorylation of AMPKα in HeLa cells treated with HU (2 mM) for 4 h.

(I) Effects of Tranilast treatment (50 μM) on T172-phosphorylation of AMPKα in HeLa cells treated with HU (2 mM, 4 h). γH2AX IF signal marks the cells with replication stress.

(J) Effects of Tranilast treatment (50 μM) on S746-phosphorylation of Exo1-GFP in HeLa cells treated with HU (2 mM) for 4 h.

(K) Effects of Exo1 knockdown on fork resection in TRPV2-depleted cells treated with HU (4 mM) for 2 h. HeLa cells were transfected with the indicated siRNAs followed by DNA fiber analysis 48 h after transfection. Top panel: Experimental scheme (see Methods). Bottom panel: dot plot of the ratio of CIdU/IdU track lengths from two independent experiments. Red bars represent the median. 150 tracks were scored for each sample. ****, p≤0.0001 (unpaired t-test).

Because the origin of Ca²⁺ induced by replication stress is intracellular and because TRPV2 is known to be localized in the ER membrane ^13,14, we next determined whether Ca²⁺ is released from the ER after replication stress. To this end, we used an ER lumen-targeting GCaMPer reporter to measure potential Ca²⁺ level changes inside the ER before and after replication stress ⁵⁰. A decrease in GCaMPer signal represents Ca²⁺ level reduction in the ER. We found that HU treatment of GCaMPer-expressing HeLa cells caused a marked decrease in Ca²⁺ in the ER in a time- and dose-dependent manner (Figure S2E and S2F). Importantly, this HU-induced Ca²⁺ decrease in the ER was abrogated by shRNA-mediated TRPV2 knockdown, overexpression of TRPV2(DN) or Tranilast treatment (Figure 2D–2F). These results further demonstrate that replication stress triggers TRPV2-mediated Ca²⁺ release from the ER into the cytoplasm, resulting in [Ca²⁺]_i elevation.

In agreement with our previous finding that [Ca²⁺]_i elevation is required for the activation of downstream CaMKK2-AMPK-Exo1 pathway after replication stress ¹³, TRPV2 knockdown or TRPV2(DN) overexpression inhibited HU-induced AMPKα T172 phosphorylation (an AMPK activation marker) by CaMKK2 in HeLa cells (Figure 2G and 2H). Disruption of TRPV2 function apparently did not affect Chk1 phosphorylation by ATR after HU treatment, in agreement with the observation that the Ca²⁺/AMPK-dependent pathway operates separately from the replication checkpoint in fork protection (Figure 2G and 2H) ¹³. We have previously shown that activated AMPK induced by replication stress is primarily localized in the nucleus ¹³. Consistently, treating cells with Tranilast abrogated HU-induced nuclear signal of AMPKα T172-phosphorylation (Figure 2I). Tranilast treatment also partially inhibited Exo1 phosphorylation at S746 after HU treatment, consistent with the finding that this site is phosphorylated by both AMPK and Chk1 after replication stress (Figure 2J) ¹³.

As Exo1 is the main target of the Ca²⁺-dependent fork protection pathway ¹³, we next determined whether removal of Exo1 can rescue the fork resection phenotype of TRPV2-depleted cells under replication stress. Indeed, results of DNA fiber and nondenaturing BrdU IF assays show that siRNA-mediated knockdown of Exo1 completely reversed the fork resection phenotype of TRPV2-depleted cells after HU treatment (Figures 2K and S2G). Together, the results described above strongly suggest that TRPV2 is the major ion channel responsible for Ca²⁺ release from the ER and the activation of the downstream CaMKK2-AMPK-Exo1 signaling cascade for fork protection after replication stress.

CytosDNA promotes TRPV2-mediated Ca²⁺ release after replication stress

How does the cell activate TRPV2 for Ca²⁺ release upon replication stress? It is conceivable that a factor that senses, or is generated from, stressed replication forks in the nucleus translocates into the cytoplasm and triggers TRPV2 activation and ER release of Ca²⁺. To test this idea, we examined the potential role of cytosDNA, which is generated after genotoxic stress and can activate an innate immune response ^19–22. Using a ssDNA antibody and PicoGreen to detect ssDNA and dsDNA, respectively ²², we observed an increased level of both ssDNA and dsDNA species in the cytoplasm in Hela cells after HU treatment (Figure 3A). To determine whether this cytosDNA is responsible for triggering the TRPV2-mediated Ca²⁺ release, we first tested whether overexpression of TREX1, which degrades both ssDNA and dsDNA in the cytoplasm ^51,52, prevents [Ca²⁺]_i induction after replication stress. Indeed, overexpression of mCherry-tagged TREX1 blocked HU-induced [Ca²⁺]_i elevation in both HeLa and U2OS cells (Figures 3B, S3A and S3B). Next we examined whether TREX1 depletion, which is known to cause cytosDNA accumulation, can promote [Ca²⁺]_i elevation by TRPV2, like HU treatment. Indeed, shRNA-mediated knockdown of TREX1 caused cytosDNA accumulation and [Ca²⁺]_i elevation in the absence of replication stress (Figures 3C, 3D and S3C). Moreover, this [Ca²⁺]_i elevation phenotype was rescued by re-expressing of wild-type TREX1, but not by a nuclease-dead TREX1(D18N) mutant (Figure 3E) ⁵². Importantly, [Ca²⁺]_i accumulation in TREX1-depleted cells was abolished by TRPV2(DN) overexpression (Figures 3F and S3D). These observations indicate that constitutive degradation of cytosDNA by TREX1 prevents unscheduled Ca²⁺ release by TRPV2 in the absence of exogenous replication stressors. In further support the role of cytosolic DNA in [Ca²⁺]_i induction, we found that direct transfection of ssDNA or dsDNA fragments also resulted in rapid [Ca²⁺]_i elevation in HeLa, U2OS and nontransformed human pancreatic ductal epithelial H6c7 cells (Figures 3G, S3E and S3F). Importantly, this [Ca²⁺]_i elevation was also TRPV2-dependent, as it was blocked by TRPV2(DN) overexpression or by Tranilast treatment (Figures 3H, 3I and S3G). Moreover, like HU treatment, direct ssDNA or dsDNA transfection also caused an elevated level of T172-phosphorylated AMPKα in the nucleus (Figure 3J). Consistent with the results of GCaMP6s reporter, we found that TREX1 knockdown or direct DNA transfection induced a reciprocal decrease in GCaMPer signal in the ER in a TRPV2-dependent manner (Figure S3H–S3J). Taken together, these data strongly suggest that cytosDNA generated after replication stress triggers TRPV2-dependent Ca²⁺ release from the ER and subsequent AMPK activation and signaling.

Figure 3. Cytosolic DNA triggers TRPV2-mediated Ca²⁺ release.

(A) Detection of ssDNA (by immunostaining) and dsDNA (by PicoGreen staining) in HeLa cells after treating with HU (4 mM) or H₂O for 6 h.

(B) Quantified GCaMP6s signal in S phase-synchronized HeLa cells expressing mCherry or TREX1-mCherry that were treated with HU (2 mM, 4 h) or H₂O. 250 cells were scored for each sample. Red bars represent the median. n=3, ****, p≤0.0001 (unpaired t-test). The same numbers of experiments performed and cells scored, and statistical analysis apply to Figure 3D–3I below.

(C) Left panel shows TREX1 knockdown. Right panel shows ssDNA and dsDNA detected in TREX1-depleted HeLa cells.

(D) Effects of TREX1 knockdown on [Ca²⁺]_i elevation in HeLa cells.

(E) Quantified GCaMP6s signal in TREX1-depleted HeLa cells that were reconstituted with shRNA-resistant TREX1(WT) or a nuclease dead mutant TREX1(D18N).

(F) Effects of TRPV2(DN) expression on [Ca²⁺]_i elevation induced by TREX1 knockdown in HeLa cells.

(G) Effects of TREX1 expression on [Ca²⁺]_i elevation in HeLa cells induced by transfected ssDNA or dsDNA fragments (2 μg/ml). Cells were imaged 6 hours after transfection. ssDNA used was a 45 bp oligo, and dsDNA was generated by annealing the corresponding sense and anti-sense oligos.

(H) Effects of TRPV2(DN) expression on [Ca²⁺]_i elevation in HeLa cells transfected with ssDNA or dsDNA fragments (2 μg/ml). Cells were imaged 6 h after transfection.

(I) Effects of Tranilast on [Ca²⁺]_i. in HeLa cells transfected with ssDNA or dsDNA fragments (2 μg/ml). Cells were imaged 6 h after transfection, and Tranilast (50 μM) or DMSO was added 4 h before imaging.

(J) Effects ssDNA or dsDNA fragment transfection (2 μg/ml) on T172-phosphorylation of AMPKα in HeLa cells detected by western blot (upper panel) and immunofluorescence (lower panel). Samples were prepared 6 h after transfection.

The cytosolic DNA sensor cGAS and its enzymatic activity are required for TRPV2-mediated Ca²⁺ release after replication stress

We next asked how cytosDNA triggers TRPV2-mediated Ca²⁺ release after replication stress. To address this question, we tested the involvement of cGAS/STING, a major cytosolic DNA sensing and signaling axis that promotes an innate immune response. Upon binding to cytosDNA or pathogen DNA, cGAS synthesizes cGAMP from ATP and GTP, which in turn binds to STING on the ER to induce downstream interferon and inflammatory responses ^23,24. We found that shRNA-mediated knockdown or pharmacological inhibition of cGAS with RU.521 or G150 abrogated [Ca²⁺]_i elevation in HU-treated HeLa cells, suggesting that cGAS and its enzymatic activity are required for [Ca²⁺]_i induction after replication stress (Figure 4A and 4B) ^53,54. In further support of this idea, CRISRP/Cas9-mediated knockout of cGAS also abolished [Ca²⁺]_i elevation in HU-treated HeLa cells (Figure 4C). This phenotype was rescued by ectopic expression of Flag-tagged WT cGAS, but not cGAS(EDAA) (containing E225A and D227A mutations), a mutant that is defective in cGAMP synthesis (Figure 4C)³⁵. The role of cGAS in [Ca²⁺]_i induction is apparently independent of its nuclear function, because expression of NES-cGAS, a fusion protein that is constitutively cytoplasmic due to the fusion of two nuclear export signal (NES) sequences to the N-terminus of cGAS, also rescued HU-induced [Ca²⁺]_i elevation in cGAS-knockout cells (Figures 4C and S4A).

Figure 4. cGAS promotes cytosolic DNA-induced Ca²⁺ release.

(A) Upper left panel: shRNA-mediated knockdown of cGAS in Hela cells. Lower left panel: Representative images of GCaMP6s signal in S phase-synchronized, cGAS-depleted HeLa cells that were treated with HU (2mM, 4h) or H₂O. Right panel: Quantified GCaMP6s signal. 250 cells were scored for each sample. Red bars represent the median. n=3, ****, p≤0.0001 (unpaired t-test). The same numbers of experiments performed and cells scored, and statistical analysis apply to Figure 4B–4G below.

(B) Quantified GCaMP6s signal in S phase-synchronized HeLa cells treated with HU (2mM, 4h) in the presence of G150 (2.5 μM), RU.521 (20 μM) or DMSO.

(C) Upper panel: Expression of Flag-tagged cGAS(WT), cGAS(EDAA, E225A/D227A) or NES-cGAS in cGAS knockout HeLa cells. Lower panel: Quantified GCaMP6s signals in S phase-synchronized, cGAS knockout HeLa cells reconstituted with the cGAS variants. Cells were treated with HU (2mM) or H₂O for 4 h.

(D) Effects of cGAS knockdown on [Ca²⁺]_i levels in TREX1-depleted HeLa cells.

(E) Effects of cGAS inhibition on [Ca²⁺]_i.elevation in HeLa cells induced by ssDNA or dsDNA fragment transfection (2 μg/ml, 6 h). G150 (2.5 μM), RU.521 (20 μM) or DMSO were added 4 h before imaging.

(F) Effects of cGAMP transfection (at indicated concentration, 6 h) on [Ca²⁺]_i in HeLa cells.

(G) Effects of TRPV2(DN) expression on [Ca²⁺]_i in HeLa cells induced by cGAMP transfection (0.5 μM, 6 h).

(H) Effects of cGAS inhibition by G150 (2.5 μM) on T172-phosphorylation of AMPKα in HeLa cells after HU treatment (2 mM, 4 h). γH2AX signal marks the cells with replication stress.

(I) Quantified BrdU signal in ssDNA in cGAS-depleted HeLa cells treated with HU (2 mM) or H₂O for 4 h. 1,000 cells were scored for each sample. n=2, ****, p≤0.0001 (unpaired t-test).

(J) Quantified BrdU signal in ssDNA in HeLa cells depleted of cGAS, Exo1, or both, after treatment with HU (2 mM) or H₂O for 4 h. 1,000 cells were scored for each sample. n=2, ****, p≤0.0001 (unpaired t-test).

Consistent with our finding that cytosDNA triggers [Ca²⁺]_i elevation after replication stress, knockdown or inhibition of cGAS also blocked [Ca²⁺]_i elevation induced by TREX1 depletion or DNA transfection (Figure 4D and 4E). In further support the role of the enzymatic activity of cGAS in [Ca²⁺]_i induction, we found that direct transfection of cGAMP into cells caused [Ca²⁺]_i elevation in the absence of replication stressors (Figures 4F and S4B). Importantly, this [Ca²⁺]_i elevation was blocked by TRPV2(DN) overexpression (Figure 4G). Taken together, the results described above indicate that cGAS and its enzymatic activity play a key role in TRPV2-mediated Ca²⁺ release in response to cytosDNA generated after replication stress.

Consistent with the requirement of cGAS for [Ca²⁺]_i induction after replication stress, inhibition of cGAS activity prevented AMPKα T172-phosphorylation in the nucleus after HU treatment (Figure 4H). Furthermore, knockdown of cGAS caused aberrant fork resection in HU-treated cells, and importantly, this fork resection phenotype was fully rescued by Exo1 knockdown (Figure 4I and 4J) ³⁹. In agreement with its role in fork protection, depletion of cGAS also caused an elevated level of chromosomal aberrations after HU treatment (Figure S4C). These data further demonstrate that cGAS promotes TRPV2-mediated [Ca²⁺]_i induction and downstream CaMKK-AMPK-Exo1 signaling cascade to prevent abnormal fork processing after replication stress.

cGAMP binding to STING promotes [Ca²⁺]_i elevation through TRPV2 derepression

In the innate immune response elicited by cytosolic DNA, STING functions downstream of cGAS and cGAMP to promote TBK1 activation and downstream type I interferon and inflammatory responses ⁵⁵. To determine whether STING, like cGAS, is also required for [Ca²⁺]_i elevation after replication stress, we depleted STING using shRNAs in GCaMP6s-expressing HeLa cells and measured [Ca²⁺]_i levels in the presence or in the absence of HU treatment. Surprisingly, in contrast to cGAS, STING depletion did not block HU-induced [Ca²⁺]_i elevation in S phase-synchronized cells (Figure 5A). Remarkably, we found that STING depletion itself caused [Ca²⁺]_i elevation in the absence of HU in S phase-synchronized or -asynchronized cells (Figures 5A, S5A and S5B). Moreover, this [Ca²⁺]_i elevation was abolished by TRPV2(DN) overexpression (Figure 5B). Consistent with the GCaMP6s reporter results, we also observed a reciprocal decrease in GCaMPer signal in HeLa cells after STING depletion, which was also reversed by TRPV2(DN) overexpression (Figure S5C and S5D). These data suggest that TRPV2 is normally repressed by STING in the absence of replication stress to prevent Ca²⁺ efflux from the ER and that this repression is released after replication stress, resulting in [Ca²⁺]_i elevation.

Figure 5. cGAMP binding to STING promotes [Ca²⁺]_i elevation through TRPV2 derepression.

(A) Left panel: shRNA-mediated knockdown of STING in Hela cells. Right panel: Quantified GCaMP6s signal in S phase-synchronized, STING-depleted HeLa cells that were treated with HU (2mM, 4h) or H₂O. Middle panel: Representative images of GCaMP6s. 250 cells were scored for each sample. Red bars represent the median. n=3, ****, p≤0.0001 (unpaired t-test). The same numbers of experiments performed and cells scored, and statistical analysis apply to Figure 5B, 5C and 5E below.

(B) Effects of TRPV2(DN) expression in HeLa cells induced by STING knockdown.

(C) Quantified GCaMP6s signals in S phase-synchronized, STING-depleted HeLa cells that were reconstituted with V5-tagged STING(WT) or STING(RYAA). Cells were treated with HU (2 mM) or H₂O for 4 h. STING(RYAA): STING(R238A/Y240A).

(D) Quantified result of alkaline comet assay in STING-depleted HeLa cells that were reconstituted with STING(WT) or STING(RYAA). Cells were treated with HU (2 mM) or H₂O for 6 h. 500 cells from three independent experiments were examined for each sample. p≤0.0001 (unpaired t-test).

(E) Left panel: Functional relationship between cGAS and STING in the Ca²⁺-dependent fork protection pathway and predicted effects of cGAS/STING double-depletion on [Ca²⁺]_i elevation and fork resection after replication stress. Right panel: Quantified GCaMP6s signals in HeLa cells depleted of STING, cGAS, or both.

(F) Quantified BrdU signal in ssDNA in HeLa cells depleted of STING, cGAS, or both after HU (2 mM, 4h) or H₂O treatment. Red bars represent the median BrdU intensity of BrdU-positive cells. 1,000 cells were analyzed for each sample. n=3, **, p≤0.01; ***, p≤0.001 (unpaired t-test).

The role of cGAMP described above for [Ca²⁺]_i induction after replication stress suggests that its binding to STING is important for TRPV2 derepression and Ca²⁺ release. To directly test this, we ectopically expressed wild-type (WT) STING or a cGAMP binding-deficient mutant STING(RYAA) (containing R238A and Y240A mutations) in STING-depleted cells, and then examined [Ca²⁺]_i induction after HU treatment in these replacement cells ^56,57. As shown in Figure 5C, while expression of WT STING rescued [Ca²⁺]_i elevation in STING-depleted cells after HU treatment, STING(RYAA) expression failed to do so. Consistent with its inability to mediate [Ca²⁺]_i induction, cells expressing STING(RYAA) exhibited a higher level of chromosomal aberrations after HU treatment, compared with cells expressing WT STING (Figure 5D). Together, these results strongly suggest that following cGAS activation, the binding of cGAMP to STING causes TRPV2 derepression and Ca²⁺ release to preserve chromosome stability after replication stress.

The opposing role of STING downstream of cGAS described above in TRPV2 regulation suggests that double-depletion of cGAS and STING would cause [Ca²⁺]_i elevation, like STING depletion alone, in the absence of replication stress (Figure 5E). This is indeed the case. In the presence of replication stress, cGAS depletion caused elevated fork resection, due to the lack of [Ca²⁺]_i induction (Figures 4A, 4I and 5F). By contrast, STING depletion had no obvious effects on fork resection, as [Ca²⁺]_i was already elevated (Figure 5F). Remarkably, double-depletion of cGAS and STING rescued the fork resection phenotype observed in cells depleted of cGAS alone (Figure 5F). These data further support the idea that cGAS facilitates [Ca²⁺]_i elevation and fork protection upon replication stress by liberating STING-mediated TRPV2 repression.

Replication stress induces STING-TRPV2 dissociation, leading to TRPV2 derepression and Ca²⁺ release

The functional relationship between TRPV2 and STING described above and their shared ER localization raise a possibility that these two proteins associate with each other in the absence of replication stress to suppress TRPV2 and that they undergo dissociation after replication stress, leading to TRPV2 derepression and Ca²⁺ release. In support of this idea, we found that endogenous and ectopically expressed STING was co-immunoprecipitated (co-IP) with TRPV2 in cells in the absence of replication stress (Figure 6A and 6B). STING and TRPV2 also exhibited colocalization and a strong proximity ligation assay (PLA) signal in cells (Figure S6A–S6B). Furthermore, affinity-purified STING-Flag protein bound to TRPV2-Flag (but not GFP-Flag) immobilized on the blot in a far-western experiment (Figure S6C). These data strongly suggest a physical interaction between these two proteins. Interestingly, treating cells with HU attenuated the co-IP and PLA signals of STING and TRPV2 (Figures 6A, 6B and S6B), suggesting that replication stress induces their dissociation. Consistent with the observation that replication stress induces [Ca²⁺]_i elevation via cytosDNA, DNA transfection also reduced the co-IP and PLA signals of these two proteins (Figures 6C and S6C). In contrast to WT STING, the STING(RYAA) mutant failed to dissociate from TRPV2 upon HU treatment, suggesting that cGAMP binding induces a conformational change in STING that leads to its dissociation from TRPV2 (Figure 6B).

Figure 6. Replication stress induces STING-TRPV2 dissociation to trigger Ca²⁺ release.

(A) Result of Co-IP experiment for detecting the association between endogenous TRPV2 and STING in Molm-13 cells treated with HU (2 mM) or H₂O for 6 h.

(B) Result of Co-IP experiment for detecting the association between stably expressed TRPV2-Flag and STING(WT)-V5 or STING(RYAA)-V5 in HeLa cells that were treated or untreated with HU (2 mM, 6 h).

(C) Effects of ssDNA or dsDNA transfection (2 μg/ml, 6 h) on the association between TRPV2-Flag and STING-V5 in HeLa cells.

(D) Diagram of human STING mutants with truncations or point mutations. Del-C, STING without C-terminal domain; Del-(C+TM4), STING without C-terminal domain and the 4^th transmembrane domain; TM, transmembrane domain. Amino acid sequence of STING (107–135 aa) was shown at the bottom, and point mutations were generated in full-length STING.

(E) Result of Co-IP experiment for detecting the association between TRPV2-Flag and WT or truncation mutants of STING-V5 in 293T cells.

(F) Result of Co-IP experiment for detecting the association between TRPV2-Flag and WT or mutant STING-V5 with the point mutations depicted in Figure 6D in 293T cells.

(G) Quantified GCaMP6s signals in S phase-synchronized, STING-depleted HeLa cells that were reconstituted with STING(WT), STING(M5), STING(RYAA) or STING(CCSS). Cells were treated with HU (2 mM) or H₂O for 4 h. STING(M5), STING(F117A/T118A/W119S); STING(RYAA), STING(R238A/Y240A); STING(CCSS), STING(C88S/C91S).

(H) TBK1 phosphorylation in the STING-replacement HeLa cells depicted in Figure 6G after MnCl₂ treatment (2.5 mM, 4 h).

To further characterize the interaction between STING and TRPV2 and its role in [Ca²⁺]_i regulation, we mapped the functional domains in STING that are important for TRPV2 association using truncation mutants. As shown in Figure 6D and 6E, a STING (1–140) mutant encompassing the N-terminal 4 transmembrane segments still retained TRPV2 association, indicating that the C-terminal cytoplasmic domain is dispensable for TRPV2 interaction ^56,57. Interestingly, further deletion of the 4^th transmembrane segment (TM4) and its flanking linker residues completely abolished TRPV2 interaction (Figure 6D and 6E). To identify key residues in this region for TRPV2 association, we generated a series of mutants of STING (M1-M10) with point mutations in 2–3 consecutive residues and tested their ability to associate with TRPV2. As shown in Figures 6F, S6E, S6F, while most of the mutants still co-immunoprecipitated with TRPV2-Flag, a STING mutant (M5) with FTW→AAS substitutions exhibited much reduced association, although this mutant was still localized on the ER and still associated with another ER protein STIM1 in the absence of replication stress ⁵⁸. This result suggests that the FTW residues, which are located at the N-terminus of TM4 and which are conserved in mammals (Figure S6G), are important for the interaction. In supporting the idea that STING interaction suppresses TRPV2 activity, we found that expression of the STING(FTWAAS) mutant could not rescue [Ca²⁺]_i elevation caused by STING depletion, in contrast to WT STING (Figure 6G).

In the innate immune signaling induced by cytosolic DNA, cGAMP binding to STING promotes its translocation from the ER to Golgi where STING undergoes palmitoylation and oligomerization for downstream TBK1 activation. Subsequent STING phosphorylation at S366 by TBK1 leads to IRF3 recruitment and activation for type I interferon production ^23,24. In the Ca²⁺-dependent fork protection pathway, STING-mediated molecular events occurring after its ER exit presumably are not required for TRPV2 activation. The differential functions of STING in TRPV2 and TBK1 activation suggest that separation-of-function mutants of STING can be identified to distinguish between these two signaling pathways. To test this idea, we expressed a palmitoylation-deficient mutant STING(CCSS) (with C88S and C91S mutations) in HeLa cells that were depleted of endogenous STING ⁵⁹. Consistent with what was reported before, this mutant was deficient in supporting TBK1 phosphorylation induced by MnCl₂ that was previously shown to directly activate cGAS (Figure 6H) ^59,60. However, this mutant was still proficient in the ER-to-Golgi translocation and in supporting [Ca²⁺]_i induction in HU-treated cells, similar to WT STING (Figures 6G and S6E) ⁵⁹. Similarly, STING(S366A), a mutant that is deficient in supporting the type I interferon response due to the lack of phosphorylation by TBK1, was also proficient in supporting [Ca²⁺]_i induction after replication stress (Figure S6H). These data also suggest that the STING/TBK1-mediated innate immune signaling is not required for the activation of the Ca²⁺-dependent fork protection pathway. In contrast to STING(CCSS) and STING(S366A), STING(RYAA) could not support HU-induced [Ca²⁺]_i elevation or MnCl₂-induced TBK1 phosphorylation, consistent with the observations that cGAMP binding is required for both TRPV2 derepression and TBK1 activation (Figures 5C and 6H) ^56,57. Interestingly, the TRPV2 interaction-deficient mutant STING(FTWAAS), which cannot support [Ca²⁺]_i induction after HU treatment, was still proficient in promoting TBK1 phosphorylation after cGAS activation by MnCl₂ (Figure 6H). These results further demonstrate that STING plays distinct roles in the TRPV2- and TBK1-dependent signaling pathways and that these pathways are mechanistically separable.

The Ca²⁺-dependent fork protection pathway is activated in response to oncogene-induced replication stress

To further illustrate the physiological significance of the Ca²⁺-dependent fork protection pathway, we determined whether intrinsic replication stress induced by oncogene activation also activates the signaling cascade. To this end, we tested whether expression of Cyclin E1 or KRAS(G12D), both of which are known to induce replication stress ^61–66, causes Ca²⁺ release and AMPK activation. By using a doxycycline-inducible expression system, we found that Cyclin E1 overexpression in nontransformed H6c7 cells indeed caused an increase in [Ca²⁺]_i concentration over time (Figures 7A and S7A). Like Cyclin E1, stable expression of KRAS(G12D) (but not WT KRAS) also caused a higher level of [Ca²⁺]_i in cells (Figures 7B and S7B). Importantly, [Ca²⁺]_i elevation induced by Cyclin E1 or KRAS(G12D) was abrogated by the TRPV2 inhibitor Tranilast or the cGAS inhibitor G150 (Figure 7C and 7D ). Furthermore, both Cyclin E1 and KRAS(G12D) overexpression resulted in T172-phosphorylation of AMPKα, indicative of AMPK activation (Figure 7E and 7F). These data suggest that intrinsic replication stress induced by oncogenes also activates the same Ca²⁺-dependent fork protection pathway that is activated by exogenous replication stressors. In further support of this idea, we found that supplementation of dNTPs, which has been shown to alleviate replication stress induced by Cyclin E1 overexpression, prevented [Ca²⁺]_i induction in Cyclin E1-expressing cells (Figure S7C)^64,67. Disruption of this Ca²⁺-dependent fork protection pathway through shRNA-mediated knockdown of cGAS or TRPV2 resulted in a heightened level of chromosomal abnormalities in cells expressing Cyclin E1 or KRAS(G12D) (Figures 7G, 7H, S7D and S7E). Taken together, these data strongly suggest that oncogene-induced replication stress also activates the Ca²⁺-dependent signaling pathway for genome protection.

Figure 7. Oncogene-induced replication stress activates the Ca²⁺-dependent fork protection pathway.

(A) [Ca²⁺]_i induction in H6c7 cells by Dox-induced expression of Cyclin E1 for the indicated times. 250 cells were scored for each sample. Red bars represent the median. n=3, ****, p≤0.0001 (unpaired t-test). The same numbers of experiments performed and cells scored, and statistical analysis apply to Figure 7B-7D below.

(B) Quantified GCaMP6s signal in H6c7 cells stably expressing empty vector (Vec), wild-type KRAS (WT) or KRAS (G12D) mutant. Cells were imaged 7 days after lentiviral infection.

(C-D) Effects of Tranilast (50 μM, 6 h) or G150 (2.5 μM, 6 h) treatment on [Ca²⁺]_i elevation induced by Cyclin E1 overexpression (Dox, 48 h) or by KRAS (G12D) overexpression (7 days post-infection).

(E-F) AMPKα T172-phosphorylation in H6c7 cells induced by Cyclin E1 overexpression (Dox, 1, 2, 3 days) or by KRAS (G12D) overexpression (7 days post-infection).

(G-H) Effects of cGAS depletion or TRPV2 depletion on chromosomal stability (alkaline comet assay) in H6c7 cells expressing Cyclin E1 (Dox, 96 h) or KRAS(G12D) (7 days post-infection). 500 cells from three independent experiments were examined for each sample. p≤0.0001 (unpaired t-test).

(I) A model for the role of the Ca²⁺-dependent signaling pathway for the protection of stressed replication forks and genome stability. See text for details. Model was drawn using BioRender and Photoshop.

DISCUSSION

This study has established a cytosDNA- and TRPV2-dependent signaling pathway in the replication stress response crucial for [Ca²⁺]_i regulation and genome maintenance. Our results indicate that cytosDNA generated after replication stress promotes Ca²⁺ release from the ER through the ion channel TRPV2, which in turn promotes CaMKK2 and AMPK activation and Exo1 suppression to prevent deleterious fork processing and chromosomal instability. The cGAS-STING axis links cytosDNA to [Ca²⁺]_i induction via a unique mechanism. While cGAS mediates cytosDNA sensing and cGAMP production to promote TRPV2-mediated Ca²⁺ release from the ER after replication stress, STING normally represses TRPV2 activity to prevent unscheduled activation of the pathway in the absence of replication stress. Upon replication stress, the binding of cGAMP to STING causes its dissociation from TRPV2, leading to TRPV2 derepression and Ca²⁺ release for fork protection (Figure 7I).

The involvement of cytosDNA in [Ca²⁺]_i regulation and fork protection sheds new light on the replication stress response and physiological functions of cytosolic DNA. CytosDNA induced by replication stress could originate from normal fork processing prior to fork reversal. In addition, retroelement activation, which can occur after genotoxic stress, may also contribute to cytosDNA accumulation and cGAS activation ^68–71. Regardless of its origin, our data indicate that cytosDNA triggers TRPV2-mediated Ca²⁺ release for fork protection. Although the relative transfection efficiency could vary between different experiments, we noted that ssDNA fragment (45 nt) generally had a lower efficiency in activating the Ca²⁺-dependent fork protection pathway (Figures 3G–3I, S3F and S3G), which is consistent with the observation that ssDNA has a lower efficiency in activating cGAS in cells ^35,36. Exactly how the replication stress-induced cytosDNA translocates into the cytoplasm remains to be defined, but micronucleus formation is likely not required for the activation of the Ca²⁺-dependent fork protection pathway in this context ⁷².

The unique functional relationship between cGAS and STING in the Ca²⁺-dependent fork protection pathway differs from their roles in the innate immune response. Both cGAS and STING promote the activation of the type I interferon response to cytosolic DNA ^23,24,55. By contrast, they play opposing roles in the replication stress response. While cGAS and its enzymatic product cGAMP are required for [Ca²⁺]_i elevation after replication stress, STING normally associates with and represses TRPV2 to prevent unscheduled Ca²⁺ release in the absence of replication stress (Figures 4 and 5). Upon replication stress, the binding of cGAMP to STING causes its dissociation from TRPV2, leading to TRPV2 derepression and Ca²⁺ efflux from the ER (Figures 5 and 6). STING undergoes ER-to-Golgi translocation in response to cytosolic DNA, which may in part be responsible for its dissociation from TRPV2 on the ER ^33,73. Regardless of the mechanism of dissociation, it is likely that the biochemical events (e.g. STING palmitoylation, oligomerization and phosphorylation) that occur after STING’s ER exit during the innate immune response are not required for [Ca²⁺]_i induction after replication stress. Indeed, consistent with this idea, the palmitoylation-deficient mutant STING(CCSS) and the phosphorylation-deficient mutant STING(S366A) can support the TRPV2-dependent fork protection pathway, but not TBK1 activation (Figures 6G, 6H and S6H) ⁷⁴. Conversely, the TRPV2 interaction-deficient STING(FTWAAS) mutant can promote TBK1 activation, but cannot repress TRPV2 in the absence of replication stress (Figure 6F–6H). These separation-of-function mutants provide a valuable tool for future further dissection of the relationship between these two signaling pathways. The role of cGAS in replication fork protection is also separable from its function in replication speed regulation. cGAS has been shown to restrain replication speed in the nucleus, but this function does not require its enzymatic activity and is independent of STING ³⁹. In contrast, cGAS promotes fork protection through cytosDNA sensing, cGAMP synthesis and the release of STING-mediated TRPV2 repression in the cytoplasm (Figures 4 and 5). Consistently, NES-cGAS, which is constitutively cytoplasmic, is proficient in supporting fork protection (Figures 4C and S4A).

The role of TRPV2 in replication fork protection further establishes the importance of Ca²⁺ signaling for genome maintenance and sheds new light on the association of TRPV2 with a number of malignancies. Besides its derepression upon STING dissociation, additional factors and steps may be required for TRPV2 activation for Ca²⁺ release and fork protection after replication stress. In addition to TRPV2, other ion channels may also be involved in [Ca²⁺]_i induction, either directly or indirectly, in the replication stress response. For instance, although TRPM5 is not directly responsible for Ca²⁺ release after replication stress, it may promote [Ca²⁺]_i elevation and subsequent fork protection indirectly by modulating the levels of Na⁺ or K⁺ in cells (Figure 1B).

Our data indicate that in addition to exogenous replication stressors, the Ca²⁺-dependent genome protection pathway is also activated by replication stress induced by oncogene activation, which occurs frequently during early stages of tumorigenesis (Figures 7 and S7) ^4,5. As is the case for the ATR/Chk1-dependent checkpoint, the role of the Ca²⁺-dependent signaling pathway in tumorigenesis is likely to be context-dependent ⁷⁵. By protecting stressed replication forks in normal cells, the Ca²⁺-dependent pathway likely prevents genomic instability and suppresses cancer initiation. However, activation of this pathway in cancer cells with intrinsic replication stress may promote cell survival and cancer progression. Thus, inhibition of this pathway may be beneficial for treating cancer with intrinsic or induced replication stress. The opposite roles of STING revealed here in regulating TRPV2 and TBK1 activity in response to cytosolic DNA suggest that STING modulators may have differential effects on the Ca²⁺-dependent replication stress response and the innate immune response. Since Ca²⁺, CaMKK2, AMPK and other pathway components may target additional proteins in cells, the Ca²⁺-dependent signaling cascade described here may also regulate other cellular processes. Further characterization of the intersections between replication stress, cytosolic DNA sensing, Ca²⁺ signaling and the innate immune response may aid the understanding and treatment of cancer and other diseases.

LIMITATIONS OF STUDY

This study identifies multiple key components in the Ca²⁺-dependent fork protection pathway, including cytosolic self-DNA, cGAS, cGAMP, STING and TRPV2, that promote Ca²⁺ release from the ER for the protection of stressed replication forks. Although TRPV2 apparently plays a major role in Ca²⁺ release after replication stress, we cannot rule out the involvement of other ion channels in the pathway. Similarly, we cannot rule out that other organelles (e.g. mitochondria and lysosomes) also contribute to [Ca²⁺]_i elevation during replication stress. Additionally, the limitations of current patch clamping techniques preclude a direct recording of TRPV2 activation on the ER after replication stress, and it remains to be determined whether it is the local Ca²⁺ activity or the global [Ca²⁺]_i accumulation that activates CaMKK2 and downstream signaling for fork protection.

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, Zhongsheng You (zyou@wustl.edu).

Materials availability

• All cell lines and plasmids generated in this study will be shared by the lead contact upon request.

• This study did not generate new unique reagents.

Data and code availability

• Original immunofluorescence images, live cell confocal images and immunoblot images have been deposited to Mendeley and are publicly available as of the date of publication. DOIs are listed in the key resources table.

• This paper does not report original code.

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

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Rabbit polyclonal anti-mCherry	BioVision	Cat#5993; RRID: AB_1975001
Rabbit polyclonal anti-Exo1	EMD Millipore	Cat# ABE1354
Rabbit polyclonal anti-phospho Exo1 (Ser746)	EMD Millipore	Cat# ABE1066
Mouse mAb anti-Chk1	Santa Cruz	Cat# sc-8408; RRID: AB_627257
Rabbit mAb anti-Phospho-Chk1 (Ser345) (133D3)	Cell Signaling Technology	Cat# 2348; RRID: AB_331212
Mouse mAb Phospho-Histone H2A.X (Ser139) (D7T2V)	Cell Signaling Technology	Cat# #80312; RRID: AB_2799949
Rabbit mAb anti-GM130 (D6B1)	Cell Signaling Technology	Cat# #12480; RRID: AB_2797933
Mouse mAb anti-ssDNA, clone F7–26	Millipore	Cat# MAB3299; RRID: AB_94795
Rabbit polyclonal anti-TREX1	ABclonal	Cat#A6778; RRID: AB_2767361
Rabbit polyclonal anti-cGAS	Cell Signaling Technology	Cat#15102; RRID: AB_2732795
Rabbit mAb anti-STING (D2P2F)	Cell Signaling Technology	Cat#13647; RRID: AB_2732796
Rabbit mAb anti-Phospho-AMPKα (Thr172) (40H9)	Cell Signaling Technology	Cat#2535; RRID: AB_331250
Rabbit polyclonal anti-AMPKα	Cell Signaling Technology	Cat# 2532; RRID: AB_330331
Rabbit polyclonal anti-TRPV2	Sigma-Aldrich	Cat# HPA044993; RRID: AB_10960889
Rabbit mAb anti-Cyclin E1(D7T3U)	Cell Signaling Technology	Cat# 20808; RRID: AB_2783554
Mouse mAb anti-V5 Tag	Thermo Fisher Scientific	Cat# R960–25; RRID: AB_2556564
Rabbit mAb anti-HA (C29F4)	Cell Signaling Technology	Cat#3724; RRID: AB_1549585
Mouse mAb anti-FLAG M2	Cell Signaling Technology	Cat#8146; RRID: AB_10950495
Mouse mAb anti-β-Actin (8H10D10)	Cell Signaling Technology	Cat#3700; RRID: AB_2242334
Rabbit polyclonal anti-α-Tubulin	Abcam	Cat# ab4074; RRID: AB_2288001
Mouse mAb anti-BrdU, Clone 3D4	BD Pharmingen	Cat# 555627; RRID: AB_10015222
Mouse mAb anti-BrdU, Clone B44	BD Pharmingen	Cat# 347580; RRID: AB_400326
Rat mAb anti-BrdU [BU1/75 (ICR1)]	Abcam	Cat# ab6326; RRID: AB_305426
Goat anti-Rabbit IgG (H+L) Secondary Antibody, DyLight 800	ThermoFisher	Cat# SA5–10036; RRID: AB_2556616
Goat anti-Mouse IgG (H+L) Secondary Antibody, DyLight 800	ThermoFisher	Cat# SA5–10176; RRID: AB_2556756
Goat anti-Rabbit IgG (H+L) Secondary Antibody, DyLight 680	ThermoFisher	Cat# 35568; RRID: AB_614946
Goat anti-Rabbit IgG (H+L) Secondary Antibody, Alexa Fluor 488	ThermoFisher	Cat# A-11008; RRID: AB_143165
Goat anti-Mouse IgG (H+L) Secondary Antibody, Alexa Fluor 568	ThermoFisher	Cat# A-11004; RRID: AB_2534072
Goat anti-Rabbit IgG (H+L) Secondary Antibody, Alexa Fluor 568	ThermoFisher	Cat# A-11011; RRID: AB_143157
Goat anti-Rat IgG (H+L) Secondary Antibody, Alexa Fluor 568	ThermoFisher	Cat# A-11077; RRID: AB_2534121
Goat anti-Mouse IgG1 Secondary Antibody, Alexa Fluor 546	ThermoFisher	Cat# A-21123; RRID: AB_2535765
Goat anti-Mouse IgG (H+L) Secondary Antibody, Alexa Fluor 488	ThermoFisher	Cat# A-11001; RRID: AB_2534069
Chemicals, peptides, and recombinant proteins
Hydroxyurea	Sigma-Aldrich	H8627, CAS: 127–07-1
PicoGreen	YEASEN	12641ES01
G150	MedChemExpress	HY-128583, CAS: 2369751–30-2
RU.521	MedChemExpress	HY-114180, CAS: 2262452–06-0
Tranilast	MedChemExpress	HY-B0195, CAS: 53902–12-8
Nocodazole	Sigma-Aldrich	Cat# M1404, CAS: 31430–18-9
Thymidine	EDM Millipore	Cat# AC226740050, CAS: 50–89-5
5-Iodo-2′-deoxyuridine (IdU)	Sigma-Aldrich	I7125, CAS: 54–42-2
5-Chloro-2′-deoxyuridine (CldU)	Sigma-Aldrich	C6891, CAS: 50–90-8
5-BrdU	ThermoFisher	B9285, CAS: 59–14-3
U73122	MedChemExpress	HY-13419, CAS: 112648–68-7
2-Aminoethyl diphenylborinate (2-APB)	MedChemExpress	HY-W009724, CAS: 524–95-8
Hoechst 33342	ThermoFisher	H3570, CAS: 2,5’-Bi-1H-benzimidazole, 2’-(4-ethoxyphenyl)-5-(4-methyl-1-piperazinyl)-23491–52-3
DAPI	Sigma-Aldrich	Cat# D9542, CAS: 28718–90-3
Protease inhibitor cocktail	Thermo	Cat# A32953
Phosphatase inhibitor cocktail	GOLDBIO	Cat# GB-451
Anti-FLAG® M2 Magnetic Beads	Sigma-Aldrich	Cat# M8823
Critical commercial assays
Alkaline comet assay	Cell Biolabs	Cat# STA-351–5
Lipofectamine™ RNAiMAX Transfection Reagent	ThermoFisher Scientific	Cat# 13778100
TransIT®-LT1 Transfection Reagent	Mirus	Cat# MIR2300
PureLink™ HiPure Plasmid Filter Midiprep Kit	ThermoFisher Scientific	Cat# K210015
Deposited data
Raw and analyzed data	This paper	http://dx.doi.org/10.17632/45zxhmx37y.1
Experimental models: Cell lines
HeLa	ATCC	CCL-2
U2OS	ATCC	HTB-96
HEK 293T	ATCC	CRL-11268
MOLM-13	Leibniz Institute DSMZ	ACC 554
H6c7	Kerafast	ECA001-FP
Oligonucleotides
siRNA targeting luciferase (Control)	Thermo Fisher	siRNA ID: 12935146
siRNA for TRPM1 #1	Thermo Fisher	siRNA ID: s8841
siRNA for TRPM1 #2	Thermo Fisher	siRNA ID: s534609
siRNA for TRPM2 #1	Thermo Fisher	siRNA ID: s14424
siRNA for TRPM2 #2	Thermo Fisher	siRNA ID: s14425
siRNA for TRPM3 #1	Thermo Fisher	siRNA ID: s36862
siRNA for TRPM3 #2	Thermo Fisher	siRNA ID: s224843
siRNA for TRPM4 #1	Thermo Fisher	siRNA ID: s29447
siRNA for TRPM4 #2	Thermo Fisher	siRNA ID: s224316
siRNA for TRPM5 #1	Thermo Fisher	siRNA ID: s26620
siRNA for TRPM5 #2	Thermo Fisher	siRNA ID: s26621
siRNA for TRPM6 #1	Thermo Fisher	siRNA ID: s44402
siRNA for TRPM6 #2	Thermo Fisher	siRNA ID: s44403
siRNA for TRPM7 #1	Thermo Fisher	siRNA ID: s29516
siRNA for TRPM7 #2	Thermo Fisher	siRNA ID: s29517
siRNA for TRPM8 #1	Thermo Fisher	siRNA ID: s35489
siRNA for TRPM8 #2	Thermo Fisher	siRNA ID: s35490
siRNA for TRPV1 #1	Thermo Fisher	siRNA ID: s14818
siRNA for TRPV1 #2	Thermo Fisher	siRNA ID: s14819
siRNA for TRPV2 #1	Thermo Fisher	siRNA ID: s28081
siRNA for TRPV2 #2	Thermo Fisher	siRNA ID: s28082
siRNA for TRPV3 #1	Thermo Fisher	siRNA ID: s46346
siRNA for TRPV3 #2	Thermo Fisher	siRNA ID: s46347
siRNA for TRPV4 #1	Thermo Fisher	siRNA ID: s34001
siRNA for TRPV4 #2	Thermo Fisher	siRNA ID: s34002
siRNA for TRPV5 #1	Thermo Fisher	siRNA ID: s32135
siRNA for TRPV5 #2	Thermo Fisher	siRNA ID: s32136
siRNA for TRPV6 #1	Thermo Fisher	siRNA ID: s30899
siRNA for TRPV6 #2	Thermo Fisher	siRNA ID: s30900
siRNA for TRPA1 #1	Thermo Fisher	siRNA ID: s17148
siRNA for TRPA1 #2	Thermo Fisher	siRNA ID: s17149
siRNA for TRPC1 #1	Thermo Fisher	siRNA ID: s14409
siRNA for TRPC1 #2	Thermo Fisher	siRNA ID: s14410
siRNA for TRPC2 #1	Thermo Fisher	siRNA ID: n270079
siRNA for TRPC2 #2	Thermo Fisher	siRNA ID: n270080
siRNA for TRPC3 #1	Thermo Fisher	siRNA ID: s14413
siRNA for TRPC3 #2	Thermo Fisher	siRNA ID: s14414
siRNA for TRPC4 #1	Thermo Fisher	siRNA ID: s14416
siRNA for TRPC4 #2	Thermo Fisher	siRNA ID: n229619
siRNA for TRPC5 #1	Thermo Fisher	siRNA ID: s14418
siRNA for TRPC5 #2	Thermo Fisher	siRNA ID: s14419
siRNA for TRPC6 #1	Thermo Fisher	siRNA ID: s14421
siRNA for TRPC6 #2	Thermo Fisher	siRNA ID: s14423
siRNA for TRPC7 #1	Thermo Fisher	siRNA ID: s32703
siRNA for TRPC7 #2	Thermo Fisher	siRNA ID: s32704
siRNA for TRPML1 #1	Thermo Fisher	siRNA ID: s32875
siRNA for TRPML1 #2	Thermo Fisher	siRNA ID: s32876
siRNA for TRPML2 #1	Thermo Fisher	siRNA ID: s48633
siRNA for TRPML2 #2	Thermo Fisher	siRNA ID: s48634
siRNA for TRPML3 #1	Thermo Fisher	siRNA ID: s30633
siRNA for TRPML3 #2	Thermo Fisher	siRNA ID: s30634
siRNA for TRPP1 #1	Thermo Fisher	siRNA ID: s10563
siRNA for TRPP1 #2	Thermo Fisher	siRNA ID: s502577
siRNA for TRPP2 #1	Thermo Fisher	siRNA ID: s10566
siRNA for TRPP2 #2	Thermo Fisher	siRNA ID: s10567
siRNA for TRPP3 #1	Thermo Fisher	siRNA ID: s17214
siRNA for TRPP3 #2	Thermo Fisher	siRNA ID: s17215
siRNA for Exo1	This paper	GCCTGAGAATAATAT GTCT
Negative control shRNA	This paper	GAATCGTCGTATGCA GTGAAA
shRNA for TRPV2 #1	This paper	AGCCGGATCCAAAC CGATTTG
shRNA for TRPV2 #2	This paper	CCTAGTGATGATCTC GGACAA
shRNA for cGAS #1	This paper	GATGCTGTCAAAGTT TAGGAA
shRNA for cGAS #2	This paper	CGTGAAGATTTCTGC ACCTAA
shRNA for STING #1	This paper	GCCCGGATTCGAAC TTACAAT
shRNA for STING #2	This paper	GTCCAGGACTTGACATCTTAA
shRNA for TREX1 #1	This paper	AACACGGCCCAAGG AAGAGCT
shRNA for TREX1 #2	This paper	AAGACCATCTGCTGT CACAAC
ssDNA for transfection to induce calcium	This paper	TTTGACCTCCATAGA AGATTCTAGAATGAAAGCCCACCCCAAGGA
qPCR primers for human TRPV2	This paper	Forward: GGAGGTGAACTGGGCTTCATG Reverse: GCACCATCCTCATCCTCCTTG
qPCR primers for human GAPDH	This paper	Forward: AACAGCGACACCCACTCCTC Reverse: GGAGGGGAGATTCAGTGTGGT
Recombinant DNA
PSPAX	Addgene	Cat#12260; RRID: Addgene_12260
pMD2.G	Addgene	Cat#12259; RRID: Addgene_12259
pMOS003-lenti-CMV-GCaMPer	Addgene	Cat#65227; RRID: Addgene_65227
pBOB-GCaMP6s	This paper	N/A
pCDH-TREX1-mCherry (wild-type or mutants)	This paper	N/A
pLVX-STING-V5 (wild-type or mutants)	This paper	N/A
pcDNA3.1-STING-Flag	This paper	N/A
pCDH-TRPV2-mCherry	This paper	N/A
pCDH-TRPV2-Flag	This paper	N/A
pLenti-EF1a-cGAS-Flag (wild-type or mutants)	This paper	N/A
pLenti-EF1a-NES-cGAS-Flag (wild-type or mutants)	This paper	N/A
pLKO.1-shRNA	This paper	N/A
pCDH-STING-HA	This paper	N/A
Software and algorithms
GraphPad Prism	GraphPad	https://www.graphpad.com/
ImageJ	ImageJ	https://imagej.nih.gov/ij/
FlowJo	FlowJo	https://www.flowjo.com/
Image Studio™ Lite	LI-COR Biosciences	https://www.licor.com/bio/products/software/image_studio_lite/

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Cell culture, transfection and generation of stable cell lines

HeLa, U2OS, HEK293T cells were grown in Dulbecco’s Modified Eagle’s Medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 100 U/mL penicillin and 100 μg/mL streptomycin at 37°C with 5% CO₂. RPMI 1640 medium was used for culturing Human Pancreatic Duct Epithelial Cell Line (H6c7) and Molm-13 cells. Plasmids were transfected into cells using TransIT-LT1 (Mirus, HEK293T) or Lipofectamine^® 3000 (Life Technologies, for HeLa, U2OS) according to the protocols of the manufacturers. siRNA transfection was done using TransIT-siQUEST transfection reagent (Mirus).

Gene knockdown and overexpression were done through lentiviral transduction. GCaMP6s-expressing lentivirus was generated in HEK293T cells through co-transfection of pMDLg/pRRE, pRSV.Rev, pMD2.G and pBOB-GCaMP6s. The packaging plasmids pSPAX2 and pMD2.G were used for the production of all the other lentiviruses used in this study. Viral supernatants were collected 48 h and 72 h after transfection and filtered using Minisart Syringe Filter (0.45 μm, Sartorius). Target cells were infected with filtered viral supernatant in the presence of polybrene (10 μg/ml). Cells stably expressing GCaMP6s were obtained through cell sorting and single clones were screened by imaging and western blot. Cells infected with Cyclin E1-expressing lentivirus were selected with G418 (400 μg/ml). Cells infected with shRNA-, KRAS-, TRPV2-, TREX1- or STING-expressing lentiviruses were selected with puromycin (2 μg/ml). All the stable cell lines were used for experiments after at least 10 days post-infection to ensure that no cytosolic DNA is generated by lentiviral RNA through reverse transcription.

METHOD DETAILS

DNA constructs and mutagenesis

Polymerase chain reaction (PCR)-amplified human TRPV2, TREX1, STING, cGAS and KRAS were cloned into pCDH, pLenti-EF1a-EGFP-3FLAG-PGK-Puro or pLVX-IRES-Puro-3xV5 vector. TRPV2(E599K/E609K) (TRPV2 DN), TREX1(D18N), STING(C88S/C91S, CCSS), STING(R238A/Y240A, RYAA), STING M1-M10, STING(S366A), KRAS (G12D) were generated through site-directed mutagenesis. pInducer20-Cyclin E1 was purchased from Addgene (#109348). Cytoplasmic retained cGAS protein was produced by fusing two copies of the nuclear export sequence (LELLEDLTL) in tandem to the N-terminus of cGAS-Flag fusion protein. For shRNA-mediated gene knockdown, shRNAs were cloned into lentivirus plasmid pLKO.1-puro. shRNAs used in this study include: Non-target control (Ctrl), GAATCGTCGTATGCAGTGAAA; TRPV2 shRNA#1, AGCCGGATCCAAACCGATTTG; TRPV2 shRNA#2, CCTAGTGATGATCTCGGACAA; Trex1 shRNA#1, AACACGGCCCAAGGAAGAGCT; Trex1 shRNA#2, AAGACCATCTGCTGTCACAAC; cGAS shRNA#1, GATGCTGTCAAAGTTTAGGAA; cGAS shRNA#2, CGTGAAGATTTCTGCACCTAA; STING shRNA#1, GCCCGGATTCGAACTTACAAT; STING shRNA#2, GTCCAGGACTTGACATCTTAA (targeting 3’ UTR). siRNA used in this study include: siExo1 (5’ CTTTTGAACAGATCGATGA 3’), siTRPV2#1 (5’ GAACCTGCTTTACTATACA), siTRPV2#2 (5’ CCCTAGT GATGATCTCGGA) and siLuc (negative control, Thermo Fisher, 12935146). siRNAs for TRP channels are listed in Supplementary Table 1. shTREX1#2-resistant shRNA containing mutations of A732G, A735T, T738A, T741A, C744G, and A747T was generated through site-directed mutagenesis. All constructs were verified by sequencing.

Nondenaturing BrdU immunofluorescence staining for measuring fork resection

A procedure of BrdU incorporation followed by nondenaturing BrdU immunofluorescence staining was used to measure intracellular ssDNA levels, as previously described ¹³. Briefly, cells cultured on glass-bottomed dishes were firstly incubated with BrdU (10 μM) for 36 h for incorporation and then with HU to induce replication stress. Cells were then permeabilized with freshly made extraction buffer (10 mM PIPES pH 6.8, 100 mM NaCl, 300 mM sucrose, 3 mM MgCl₂, 1 mM EGTA and 0.2% Triton X-100) for 5 min. For fixation, cells were incubated with 3% paraformaldehyde in PBS (room temperature) for 20 min, cold methanol (−20°C) for 20 min, and then ice-cold acetone (4 °C) for 30 seconds. Subsequently, cells were blocked with blocking buffer (PBS containing 0.05 % Tween-20 and 2% BSA) for 1 h at room temperature before immunostaining with anti-BrdU antibodies (1:1000, BD Pharmingen, 555627) overnight at 4°C. Both primary and secondary antibodies were diluted in PBS containing 2% BSA. After nuclear staining with Hoeschst 33342 (1 μg/ml), images were captured using an inverted microscope (Nikon Ti-E) and Metamorph software (Molecular Devices). The BrdU signal in individual nuclei (defined by Hoeschst-stained area) was determined using ImageJ. Cells with a BrdU signal above that in the majority (98%) of HU-untreated control cells were taken as BrdU-positive. Images of 1,000 randomly selected cells for each sample were analyzed. Statistical analysis was performed in GraphPad Prism using unpaired t-test.

Single-molecular DNA fiber assay for measuring fork resection

Single-molecular DNA fiber assay was performed to measure the resection of nascent DNA, as previously described⁷⁶. Briefly, cells were incubated sequentially with prewarmed fresh media containing thymidine analogs iododeoxyuridine (IdU, 20 μM) and chlorodeoxyuridine (CldU, 200 μM) for 20 min each before treatment with HU (4 mM) for 2 h. Cells were then trypsinized and resuspended in ice-cold PBS to a final concentration of 10⁶/ml. A drop of 2 μl of this cell solution was added on one edge of a glass slide and set for 1 min, and 8 μl of lysis buffer (200 mM Tris-HCl pH7.5, 50 mM EDTA, 0.5 % SDS) was then added in drops onto the cells. After 6 min-incubation till the edges of the lysis buffer were slightly dried out, slides were tilted (20–45°) to allow the liquid to slowly run down the length of the slide. The resulting DNA spreads were air dried, fixed in methanol-acetic acid (3:1) for 10 min, and then denatured with 2.5 N HCl for 1 hr. After washing with PBS, the slides were incubated with blocking buffer (5% BSA in PBS with 0.1% Tween-20) for 1 h. Immunodetection of DNA fibers was performed with rat anti-BrdU antibody (1:50, Abcam, ab6326) for CldU and mouse anti-BrdU antibody (1:50, Becton Dickson, 347580) for IdU in a humid chamber at 37°C for 1 h. Slides were then incubated with secondary antibodies (anti-rat Alexa 488 (Molecular Probes, A21470, 1:100) and anti-mouse Alexa 546 (Molecular Probes, A21123, 1:100) at 37°C for 45 min in the dark. After washing with PBST (0.1% Tween 20), excess liquid was drained from the slides followed by mounting with Prolong Gold Antifade reagent (Life Technologies). Images of the DNA fibers were captured by using 63× oil immersion objective of a fluorescence microscope (inverted Nikon Ti-E microscope) with the appropriate filters. 150 fiber tracts were scored for each sample. The DNA track lengths were measured using ImageJ and the pixel length values were converted into micrometers using the scale bars created by the microscope. Statistical analysis was done using GraphPad Prism.

Metaphase chromosome spreading assay

Detection of chromosomal aberrations was performed by DAPI-stained metaphase spreads, as described previously ^13,77. Briefly, cells treated with 4 mM HU for 6h were released into fresh medium to recover for 20h. Cells were then treated with 10 μM nocodazole for 4h before harvest. Trypsinized cells were resuspended in 10 ml of pre-warmed hypotonic solution (10 mM KCl, 10% FBS) for 10-minute incubation at 37°C, and then in ice-cold fixation buffer (acetic acid 1: 3 methanol) for 30-minute fixation on ice. Cells were washed with ice-cold fixation buffer for another 4 times and dropped onto pre-chilled slides. The slides were air dried thoroughly and mounted with Prolong Gold Antifade reagent with Hoechst. Images were acquired with a Zeiss Axio Observer A1 Inverted Phase Contrast Fluorescence Microscope. 150 randomly selected metaphases were scored per sample. Statistical analysis was performed in GraphPad Prism.

Alkaline comet assay

The alkaline comet assay was performed according to the manufacturer’s instructions (Cell Biolabs). Briefly, cells after treatment as depicted in figure legends were collected and resuspended in PBS at a concentration of 5 X 10⁵ cells per ml. Ten thousand cells were then mixed with 150 μl low melting-point agarose kept at 37 °C. Seventy microliteres of this mixture was added to each well of comet slides, and then placed on ice for 30 min to accelerate gelling of the agarose layer. Subsequently, cells were incubated with lysis buffer (1 h) and then with alkaline buffer (30 min) at 4 °C. Slides were then subject to electrophoresis for 20-min at 0.81 V/cm. After dehydration and drying, the DNA was stained with SYBR^®Green and images were acquired through fluorescence microscopy. Tail moment of cells was measured using ImageJ software with a plugin ‘OpenComet’ ⁷⁸. 250 cells were scored per sample. Statistical analysis was performed in GraphPad Prism.

Clonogenic assay and CCK8 assay for measuring cell viability

To measure cell survival after HU treatment, clonogenic assay and CCK8 assay were used for adherent cells and suspension cells, respectively. HeLa cells were plated on 6-well dishes at a density of 500 cells per well. 16 hours after plating, cells were treated with HU at the indicated concentrations for 24 h. Cells were cultured in fresh medium for 14 days to allow colony formation. Cell colonies were then stained with 0.2% Crystal Violet in 50% methanol and counted. The Cell Counting Kit-8 assay (CCK8, Dojindo Molecular Technologies, Inc., Rockville, MD) was performed according to the manufacturer’s protocol. Absorbance at 450 nm was determined by an CLARIOstar plate reader and Gen5 version 2.09 software (BMG LABTECH). Statistical analysis was performed in GraphPad Prism.

Live cells imaging of Ca²⁺ using the GCaMP6s and GCaMPer reporters

For Ca²⁺ imaging in live cells, genetically-encoded calcium indicators GCaMP6s and GCaMPer were used to measure Ca²⁺ levels in the cytoplasm and ER, respectively ^50,79. GCaMP6s or GCaMPer was stably expressed in different cell lines via lentiviral infection. To measure the [Ca²⁺]_i elevation or Ca²⁺ ER release after HU treatment, cells were first synchronized in S phase by releasing from a double-thymidine block. Briefly, cells cultured on 35 mm glass-bottomed dishes were treated first with 2 mM thymidine for 18 h, followed by 9 h release in normal cultural medium, then treated again with 2 mM thymidine for 17 h. After the double-thymidine block, cells arrested at G1/S boundary were released into in normal cultural medium for 2 h, followed by HU (4 mM) treatment for 4 h before imaging. The fluorescence signals of GCaMP6s or GCaMPer were acquired using a confocal Microscope (Leica TCS SP8) with an objective of 20x, and a live-cell imaging chamber capable of maintaining temperature, humidity, and CO₂ levels. GCaMP6s or GCaMPer fluorescence signals in cytoplasm in individual cells were quantified using ImageJ after subtracting the background (defined as the signal in an area in each image without cells). [Ca²⁺]_i imaging was also performed for cell cycle-asynchronized cells expressing Cyclin E1 or KRAS, depleted of TREX1 or transfected with ssDNA or dsDNA fragments, as indicated in figure legends.

Immunofluorescence staining and immunoprecipitation

Immunofluorescent staining was performed as previously described ¹³. Briefly, cells cultured on glass-bottomed dishes were fixed with 4% paraformaldehyde for 15 min at room temperature. After washing 3 times with PBS, cells were permeabilized for 15 min in PBS containing 0.2% Triton X-100, and then blocked with PBS containing 3% BSA. Cells were incubated with primary antibodies (in PBS with 0.1% Triton X-100 and 3% BSA) overnight at 4°C and then with secondary antibodies (in PBS with 0.1% Triton X-100 and 3% BSA) for 1h at room temperature, followed by DAPI (Sigma) staining for 10 min. Images were captured using a Leica TCS SP8 confocal microscope.

To immunoprecipitate (IP) endogenous TRPV2 protein, 2 μg of anti-TRPV2 antibody or control normal rabbit IgG was incubated with 20 μl of Protein A-Sepharose beads at room temperature for 1 h with rotation. Antibody-bound beads were then incubated with cell lysates at 4°C for 3 h. After washing 3 times with lysis buffer (10 mM NaKPO4, pH 7.2, 150 mM NaCl, 1% sodium deoxycholate, 1% Triton-100, protease inhibitor cocktail and phosphatase inhibitor cocktail) ⁸⁰, bead-bound proteins were dissolved in SDS-PAGE sample buffer. For co-IP in Figure 6 to detect the interaction between STING-V5 (WT or mutants) and TRPV2-Flag, cell lysates were incubated with pre-conjugated anti-Flag M2 affinity beads (Sigma) for 3 h. After washing 3 times with lysis buffer, bead-bound proteins were dissolved in SDS sample buffer and incubated at 37 °C for 30 min before gel loading.

Proximity ligation assay (PLA)

A proximity ligation assay (PLA) was performed for HeLa cells stably expressing TRPV2-Flag and STING-HA to detect STING-TRPV2 association in cells in situ. HeLa cells expressing TRPV2-Flag only or STING-HA only were used as controls. Cells were seeded onto Lab-Tek II CC2 chamber slides (MilliporeSigma, S6815) and treated as indicated in the figure legends. After treatment, cells were fixed with 4% paraformaldehyde in PBS for 15 min and then permeabilized with 0.2% Triton X-100 in PBS for 15 min. After permeabilization, cells were incubated in the Blocking Solution (MilliporeSigma, DUO82007) for 1 hour at 37°C. Primary antibodies were then incubated with cells in the Antibody Diluent (MilliporeSigma, DUO82008) at 1:1K overnight at 4°C. Cells were then washed in Wash Buffer A (MilliporeSigma, DUO82046) for 10 min at room temperature (RT) and incubated with PLUS and MINUS PLA probes (MilliporeSigma, DUO82002, DUO82004) at 1:5 in the Antibody Dilute for 1 hour at 37°C. Cells were again washed in Wash Buffer A for 10 min at RT and then incubated with the Ligase (MilliporeSigma, DUO82027) in the Ligation buffer (MilliporeSigma, DUO82009) for 30 min at 37°C, as described by the manufacturer. Next, cells were washed in Wash Buffer A for 10 min at RT and incubated with the Polymerase (MilliporeSigma, DUO82028) in the Amplification Buffer (MilliporeSigma, DUO82011) for 100 min at 37°C. Finally, cells were washed in Wash Buffer B (MilliporeSigma, DUO82048) for 20 min at RT and then stained for 10 min in Wash Buffer B containing Hoeschst 33342 (5ug/ml). Cells were then washed in 0.01×Wash Buffer B for 1 min at RT followed by mounting with Prolong Gold Antifade reagent (Thermofisher, P36930). Fluorescent images were acquired using Nikon spinning disk confocal X-Light V3 with 60X oil immersion objective. 150 cells were quantified for each sample.

Far-western blot

The Far-western blot procedure was modified based on the protocol previously described ⁸¹. STING-Flag, TRPV2-Flag, and GFP-Flag fusion proteins as a control were expressed and affinity-purified from HEK293T cells using anti-Flag M2 affinity gel. Equal amount (200 ng) of the purified TRPV2-Flag and GFP-Flag proteins were resolved by SDS-PAGE and transferred onto Immobilon-P PVDF membrane (Millipore Corporation). The blot was blocked with 5% BSA in PBST (1XPBS with 0.1% Tween20) for 1 h at room temperature, and then incubated with 500 ng/ml of STING-Flag in PBST with 5% BSA overnight at 4 °C. After washing, the blot was incubated with anti-STING primary antibody (2 h at room temperature) and then with HRP-labeled anti-rabbit secondary antibody for 1h at room temperature before ECL detection.

QUANTIFICATION AND STATISTICAL ANALYSIS

Data were tested for statistical significance with GraphPad Prism software. The tests performed, the sample size (n) and the number of independent replicates for each experiment are depicted in the figure legends.

HIGHLIGHTS.

• Replication stress induces TRPV2-mediated Ca²⁺ release from the ER

• STING binds to and represses TRPV2 on the ER in the absence of replication stress

• Cytosolic self-DNA-induced cGAMP promotes STING dissociation from TRPV2 and Ca²⁺ efflux

• Oncogene-induced replication stress also activates the Ca²⁺-dependent fork protection