cGAS guards against chromosome end-to-end fusions during mitosis and facilitates replicative senescence

As a sensor of cytosolic DNA, the role of cyclic GMP-AMP synthase (cGAS) in innate immune response is well established, yet how its functions in different biological conditions remain to be elucidated. Here, we identify cGAS as an essential regulator in inhibiting mitotic DNA double-strand break (DSB) repair and protecting short telomeres from end-to-end fusion independent of the canonical cGAS-STING pathway. cGAS associates with telomeric/subtelomeric DNA during mitosis when TRF1/TRF2/POT1 are deficient on telomeres. Depletion of cGAS leads to mitotic chromosome end-to-end fusions predominantly occurring between short telomeres. Mechanistically, cGAS interacts with CDK1 and positions them to chromosome ends. Thus, CDK1 inhibits mitotic non-homologous end joining (NHEJ) by blocking the recruitment of RNF8. cGAS-deficient human primary cells are defective in entering replicative senescence and display chromosome end-to-end fusions, genome instability and prolonged growth arrest. Altogether, cGAS safeguards genome stability by controlling mitotic DSB repair to inhibit mitotic chromosome end-to-end fusions, thus facilitating replicative senescence. Supplementary Information The online version contains supplementary material available at 10.1007/s13238-021-00879-y.

(A) ChIP analysis of TRF1, TRF2 and POT1 associating with telomeres in asynchronous (Asynchr) or mitotic HeLa cells. Cells were either asynchronous or synchronized at G1/S, released for 8 h and then treated with colcemid for 12 h. ChIP enriched DNA was used for slot blot and hybridization using telomeric G-rich probe or Alu probe.
(B) Quantification of A. The relative amount of enriched telomeric DNA was calculated (ChIP / Input, %). Alu was used as a control. All values are the average ± SEM of three independent experiments.
(C) Immunoblot analysis TRF1, TRF2 or POT1 in asynchronous (Asynchr) and mitotic U2OS cells. Phosphorylated H3Ser10 was used as a marker for mitotic cells.
(D) Visualization of endogenous cGAS and telomeres in asynchronous (Asynchr) or mitotic HeLa cells. Cells were treated as in A and then subjected to metaphase spread followed by IF/FISH. Scale bars, 10 μm. (A) Immunoblot analysis cGAS in U2OS cells. cGAS was depleted by CRISPR/Cas9-based approach (gcGAS-1 or gcGAS-2, see method for details) and scramble sg-RNA (Scr) was used as a control.
(B) Time course of cell synchronization in C, G.
(D) Quantification of Figure 2A. Relative lengths of fused telomeres were compared with that of total telomeres in cGAS-deficient U2OS cells (gcGAS-1).
(F) Immunoblot analysis cGAS in VA13 cells. The experiments were same as in A.
(G) FISH of telomeres on metaphase spreads to detect chromosome end-to-end fusions in VA13. The experiments were same as in Figure 2A. Scale bars, 10 μm.
(H) Quantification of G. The number of fused chromosome ends per metaphase was given (n ≥ 44 metaphase). All values are the average ± SEM of three independent experiments.
(I) Quantification of G. Relative lengths of fused telomeres were compared with that of total telomeres in cGASdeficient VA13 cells (gcGAS-1).
(L) Immunoblot analysis cGAS in 6kb-HeLa cells. The experiments were same as in A.
(M) Immunoblot analysis cGAS in 2kb-HeLa cells. The experiments were same as in A.

A-Scr
(C) ChIP analysis of TRF1, TRF2 and POT1 associating with telomeres in asynchronous (Asynchr) or mitotic U2OS cells. Control (Scr) and cGAS-deficient U2OS cells (gcGAS-1) were either asynchronous or synchronized at G1/S, released for 8 h and then treated with colcemid for 12 h. ChIP enriched DNA was used for slot blot and hybridization using telomeric G-rich probe or Alu probe. (C) Quantification of A. Relative lengths of telomeres colocalized with γH2AX were compared with that of total telomeres.
(D) IF and FISH performed on metaphase spreads of HeLa cells to visualize γH2AX foci and telomeres, respectively. Mitotic TIF were indicated with arrows. Scale bars, 10 μm.
(E) Quantification of A, D and Figure 4A. The number of γH2AX foci colocalized with telomeres/chromosome ends per metaphase between VA13, U2OS and HeLa was given.
(F) Immunoblot analysis γH2AX in asynchronous (Asynchr) and mitotic VA13, U2OS or HeLa cells. (A) Co-IP assay using exogenously expressed GFP-CDK1 to determine the interaction of cGAS with CDK1 in asynchronous (Asynchr) or mitotic U2OS cells. GFP-EV was used as a negative control. Cells were either asynchronized or synchronized in mitosis as described above.
(B) Co-IP assay to determine the interaction of endogenous CDK1 with cGAS in mitotic HeLa cells. IgG was used as a negative control. Cells were synchronized in mitosis as described above.
(C) Co-IP assay using exogenously expressed GFP-cGAS to determine the interaction of CDK1 with cGAS in mitotic U2OS cells. Mitotic cell lysis was treated with DNase I prior to IP. Cells were synchronized in mitosis as described above.
(D) Co-IP assay using exogenously expressed GFP-cGAS to determine the interaction of CDK1 with cGAS in mitotic HeLa cells. Mitotic cell lysis was treated with DNase I prior to IP. Cells were synchronized in mitosis as described above.  (C) Immunoblot analysis of p53 in pre-senescent BJ fibroblast cells sequentially transfected with si-scramble (NC) or sicGAS or siSTING for 12 days and then with sip53 for 4 days.
(D) Quantification of Figure 6G. Relative lengths of fused telomeres were compared with that of total telomeres in cGAS and p53-double knocked down BJ fibroblast cells.
(E) Detection of mRNA levels of cGAS, STING, RAD51, and LIG4 in BJ cells transfected with NC or sicGAS or siSTING in combination with transfection of siRAD51 or siLIG4 as indicated. mRNA levels were quantified by qPCR.
(F) SA-β-Gal staining of pre-senescent BJ fibroblast cells transfected with si-scramble (NC) or sicGAS or siSTING in combination with transfection of siRAD51 or siLIG4 as indicated. After transfection, cells were cultured for 12 days and subjected to SA-β-Gal staining. Scale bars, 100 μm.
(G) Quantification of F. The percentage of SA-β-Gal positive cells was calculated (n ≥ 100 cells). All values are the average ± SEM of three independent experiments.
(H) EdU labeling was used to indicate S phase BJ fibroblast cells. Pre-senescent BJ fibroblast cells sequentially transfected with si-scramble (NC) or sicGAS or siSTING for 24 days. Corresponding siRNA was transfected every 4 days. EdU-labeled cell was indicated with arrow. Scale bars, 10 μm.
(I) Quantification of H. The percentage of cells with EdU was determined (n ≥ 100 cells). All values are the average ± SEM of three independent experiments. The unpaired Student's two-tailed t-test was used to determine the statistical significance (***P<0.001, ****P<0.0001).