Tousled-like kinase 2 regulates recovery from a DNA damage-induced G2 arrest

Abstract

In order to maintain a stable genome, cells need to detect and repair DNA damage before they complete the division cycle. To this end, cell cycle checkpoints prevent entry into the next cell cycle phase until the damage is fully repaired. Proper reentry into the cell cycle, known as checkpoint recovery, requires that a cell retains its original cell cycle state during the arrest. Here, we have identified Tousled-like kinase 2 (Tlk2) as an important regulator of recovery after DNA damage in G2. We show that Tlk2 regulates the Asf1A histone chaperone in response to DNA damage and that depletion of Asf1A also produces a recovery defect. Both Tlk2 and Asf1A are required to restore histone H3 incorporation into damaged chromatin. Failure to do so affects expression of pro-mitotic genes and compromises the cellular competence to recover from damage-induced cell cycle arrests. Our results demonstrate that Tlk2 promotes Asf1A function during the DNA damage response in G2 to allow for proper restoration of chromatin structure at the break site and subsequent recovery from the arrest.

Keywords: Asf1A; DNA damage; Tlk2; cell cycle; checkpoint recovery.

Figure EV1. Primary screen details

1. Protocol used for the siRNA‐based kinome screen. U2OS cells were transfected for 24 h, synchronized at the G1/S border by thymidine for 24 h, and then either released into nocodazole for 16 h (unperturbed mitotic entry) or released for 7 h, treated with 0.5 μM adriamycin for 1 h, and after 16 h, cells were induced to recover by addition of caffeine for 8 h (checkpoint recovery). Cells were fixed and stained with DAPI and antibodies directed at histone H3‐pS10. Finally, automated analysis of the percentage of histone H3‐pS10‐positive DAPI‐stained nuclei was carried out.

2. Representative images of positive control (siWip1)‐ and negative control (siGAPDH)‐transfected cells treated as in (A). Calculated percentages are indicated. Scale bar indicates 100 μm.

3. Plots of duplicate screens. U2OS cells were treated as in (A). Black diamonds indicate individual siRNA‐targeted genes from the library, light gray diamonds indicate positive (Wip1‐depleted) and negative (untransfected and GAPDH‐depleted) controls, dark gray diamond indicates the genes that met the criteria in both screens, and white diamonds indicate Tlk1‐depleted conditions. Dotted lines indicate selection criteria for recovery‐specific genes.

Figure 1. An siRNA‐based screening approach identifies potential regulators of recovery

Plot of one of the screens (the duplicates are shown in Fig EV1C). U2OS cells were transfected, synchronized at the G1/S border by thymidine for 24 h, and then either released into nocodazole for 16 h (unperturbed mitotic entry) or for 7 h followed by treatment with 0.5 ?M adriamycin for 1 h, and after a 16‐h G2 arrest, cells were induced to recover by addition of caffeine (checkpoint recovery). Mitotic index was scored based on the percentage of histone H3‐pS10‐positive DAPI‐nuclei and normalized to the untransfected controls. Black diamonds indicate individual siRNA‐targeted genes from the library, light gray diamonds indicate positive (Wip1‐depleted) and negative (untransfected and GAPDH‐depleted) controls, and dark gray diamonds indicate the hits based on the two screens. Dotted lines indicate selection criteria for recovery‐specific genes.

Figure 2. Tlk2 kinase activity is required for recovery from a DNA damage‐induced arrest

1. U2OS cells were transfected with four independent siRNAs from the pools used in the screen, treated as in Fig 1, and analyzed for mitotic index. Error bars represent SD, n = 3. Statistical significance was tested using a paired two‐tailed t‐test (NS P > 0.05, *P ≤ 0.05, **P ≤ 0.01).

2. U2OS cells were synchronized with a single thymidine block, released into G2, and damaged with 0.5 μM adriamycin for 1 h. After a 16‐h G2 arrest, cells were harvested for Western blot analysis and analyzed for Tlk2 protein levels.

3. Live cell imaging of thymidine‐synchronized unperturbed or damaged G2 cells. Cumulative percentage of cells entering mitosis were scored and plotted.

4. U2OS cells were transfected with either a control siRNA or Tlk2 siRNA #3, synchronized, and damaged in G2. Cells were either harvested or treated with caffeine for 8 h before harvest, and cell cycle distribution was analyzed by FACS. Percentages of cells in each quadrant are indicated.

5. Tlk2Δ cells were generated using CRISPR/Cas9 genome editing. Cells were synchronized in G2 by thymidine release and damaged with 0.5 μM adriamycin for 1 h. After a 16‐h G2 arrest, cells were induced to recover by addition of caffeine for 8 h and analyzed by FACS. Error bars represent SD, n = 4. Statistical significance was tested using a paired two‐tailed t‐test (NS P > 0.05, *P ≤ 0.05, ***P ≤ 0.001).

6. U2TR cells stably expressing Tlk2 siRNA #3‐insensitive tetracycline‐inducible FLAG‐Tlk2‐wt or FLAG‐Tlk2‐D613A were thymidine‐synchronized and damaged in G2. Tetracycline was present form the start of the experiment where indicated.

7. Caffeine‐induced recovery assay of cell lines shown in (F). Error bars represent SD, n = 3. Statistical significance was tested using a paired two‐tailed t‐test (NS P > 0.05, *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001).

Figure EV2. Secondary screening of identified kinases

1. A–C

   Single siRNAs which were pooled in the primary screen were deconvolved. U2OS cells were transfected with the siRNAs, and mitotic index is determined as in Fig 1. For each gene identified in the screen, the mitotic index of the single siRNAs is shown in both the unperturbed setup (A), recovery setup (B) and a ratio between the two (C). The gray‐dotted line indicates the cutoff criteria. Error bars represent SD, n = 3.

Figure EV3. Tlk2 is not essential for replication and cell viability

1. A

   Asynchronously cycling U2OS cells were transfected with the indicated siRNAs and analyzed for outgrowth after 7 days using crystal violet staining. Hydroxyurea was added where indicated at a concentration of 1 mM.

2. B

   Quantification of the experiment in (A). Error bars represent SD, n = 3.

3. C

   Asynchronously cycling U2OS cells were treated as in (A). Cells were treated with EdU for 2 h, fixed, and analyzed for Edu incorporation. Scale bar indicates 10 μm.

4. D

   Quantification of the experiment in (C). Error bars represent SD, n = 3.

5. E

   Western blot analysis of Tlk2 knockout cells.

6. F, G

   Incorporation of the Blast^R gene as measured by PCR amplification of genomic DNA.

7. H

   Cell cycle analysis of asynchronous parental and Tlk2Δ cell lines.

8. I

   Asynchronously cycling parental and Tlk2Δ cell lines were treated with EdU for 4 h and analyzed for EdU incorporation similar to (D). Error bars represent SD, n = 3.

9. J

   Population doubling during 3 consecutive days of asynchronously cycling parental and Tlk2Δ cell lines. Error bars represent SD, n = 3.

10. K

    Adriamycin sensitivity in parental and Tlk2Δ cells. Cells were treated with the indicated amounts of adriamycin and analyzed for outgrowth using crystal violet staining after 7 days.

Figure EV4. Analysis of Tlk1 and Asf1A during DNA damage and recovery

1. A

   U2OS cells were transfected with control or Tlk2 siRNAs and synchronized in G2 by a single thymidine block. Cells were subsequently damaged in G2 by 0.5 μM adriamycin for 1 h and harvested at the indicated times after damage. Samples were analyzed by Western blot with the indicated antibodies.

2. B

   Cells were treated as in Fig 2A. Error bars represent SD, n = 3. Statistical significance was tested using a paired two‐tailed t‐test (NS P > 0.05, **P ≤ 0.01).

3. C, D

   U2OS cells were synchronized with a single thymidine block, released into G2, and damaged with 0.5 μM adriamycin for 1 h. After a 16‐h G2 arrest, cells were harvested for Western blot analysis and analyzed for Tlk1 and Tlk2 protein levels.

4. E

   Inducible FLAG‐Asf1A‐wt cell lines were synchronized by a single thymidine block and released for the hours indicated. Adriamycin was added for 1 h at 0.5 μM and washed out for the indicated time. FLAG‐Asf1A was immunoprecipitated and blotted with the indicated antibodies.

5. F

   Cells were synchronized in G2 phase by release from a single thymidine block. Cells were left untreated or were irradiated with 6 Gy of IR. Samples were fixed 3 h afterward and stained for Asf1A‐pS166. Nuclear fluorescence intensity was measured relative to DAPI. Whiskers represent 5–95% of data points. Statistical significance was tested using an unpaired two‐tailed t‐test (****P ≤ 0.0001).

6. G

   U2OS cells were synchronized in G2 by a single thymidine block and were left untreated or were irradiated with 6 Gy of IR. Samples were pre‐extracted at the indicated time points and were fixed for analysis. Cells were stained for total levels of histone H3 and quantified relative to DAPI. Whiskers represent 5–95% of data points. Statistical significance was tested using an unpaired two‐tailed t‐test (NS P > 0.05, ****P ≤ 0.0001).

Figure 3. Depletion of Tlk2 leads to altered kinetics of the DNA damage response

1. U2OS cells were grown on glass coverslips and synchronized in G2 after thymidine release. Cells were fixed at the indicated times after adriamycin‐induced DNA damage and stained for DAPI and γH2AX. Scale bar indicates 10 μm.

2. Quantification of cells in (A). Each dot represents the total signal of γH2AX in the nucleus of a single cell. Bars indicate the mean and error bars indicate the SD of the data points in each condition. Whiskers represent 5–95% of data points. Statistical significance was tested using an unpaired two‐tailed t‐test (NS P > 0.05, ****P ≤ 0.0001).

3. U2OS cells were synchronized in G2 and harvested at the indicated times after induction of adriamycin‐induced DNA damage with or without Tlk2 depletion. Samples were analyzed by Western blotting for the indicated proteins.

4. U2OS cells were synchronized in G2 and damaged with 6 Gy of IR. Samples were taken at the indicated times after induction of damage with or without Tlk2 depletion. Samples were analyzed by Western blot for the indicated proteins. Band signal intensity for Asf1A and Asf1B was measured and corrected for Ponceau S staining.

5. U2OS cells were transfected with the indicated siRNAs and treated as in Fig 2A. Mitotic index was analyzed by FACS. Error bars represent SD, n = 3. Statistical significance was tested using a paired two‐tailed t‐test (NS P > 0.05, *P ≤ 0.05).

6. U2OS cells were synchronized with a single thymidine block, released into G2, and damaged with 0.5 μM adriamycin for 1 h. After a 16‐h G2 arrest, cells were harvested for Western blot analysis and analyzed for the indicated protein levels. Lanes were all present on the same blot but pasted together for comparison as indicated by the lines.

7. Cells were synchronized in S or G2 phase by a single thymidine block. Damage was induced by 0.5 μM adriamycin for 1 h, and cells were allowed to incubate for 16 h afterward. Samples were fixed at the indicated cell cycle stages and stained for Asf1A‐pS166. Nuclear fluorescence intensity was measured relative to DAPI. Whiskers represent 5–95% of data points.

8. Cells were treated and analyzed as in Fig 2E. Induction of FLAG‐Asf1A‐4D was induced using doxycycline. Error bars represent SD, n = 5. Statistical significance was tested using a paired two‐tailed t‐test (*P ≤ 0.05).

9. U2OS cells were synchronized in G2 by a single thymidine block and where indicated treated with 0.5 μM adriamycin for 1 h and after the indicated time points were pre‐extracted and fixed. Cells were stained for total levels of histone H3 and quantified relative to DAPI. Scale bar indicates 10 μm. Whiskers represent 5–95% of data points.

10. FLAG‐Asf1A‐4D‐inducible cells were treated as in (I), and expression was induced using doxycycline where indicated. Whiskers represent 5–95% of data points. Statistical significance was tested using an unpaired two‐tailed t‐test (****P ≤ 0.0001).

Figure 4. Depletion of Tlk2 leads to a loss of recovery competence

1. U2OS cells were treated as in Fig 2B after which mRNA was isolated and analyzed by qPCR. Error bars represent SD, n = 3. Statistical significance was tested using a paired two‐tailed t‐test (*P ≤ 0.05).

2. Tlk2Δ 9.3.1 cells were treated and analyzed as in Fig 3A. Error bars represent SD, n = 3. Statistical significance was tested using a paired two‐tailed t‐test (*P ≤ 0.05, **P ≤ 0.01).

3. U2OS cells were synchronized in G2 and harvested or synchronized in G2 and harvested 16 h after adriamycin‐induced DNA damage. Samples were analyzed by Western blot for the indicated proteins.

4. U2OS cells were transfected with the indicated siRNAs, synchronized in G2, and damaged with 6 Gy of IR. Cells were harvested 24 h later, and the mitotic index was analyzed by FACS. Error bars represent SD, n = 3. Statistical significance was tested using a paired two‐tailed t‐test (**P ≤ 0.01).

5. Cells transfected with the indicated siRNAs were treated and analyzed as in (C).

6. U2OS cells were transfected with the indicated siRNAs, arrested in G2 with adriamycin, and induced to recover by addition of caffeine. Mitotic index was analyzed by FACS. Error bars represent SD, n = 3. Statistical significance was tested using a paired two‐tailed t‐test (**P ≤ 0.01).

7. U2OS cells were synchronized in G2 and harvested 16 h after adriamycin‐induced DNA damage. Samples were analyzed by Western blot for the indicated proteins.

8. U2TR cells stably expressing Tlk2 siRNA #3‐insensitive tetracycline‐inducible FLAG‐Tlk2‐wt were thymidine‐synchronized, damaged in G2, and harvested 24 h afterward. Tetracycline was either absent (off), added from the start of the experiment (early) or 16 h after induction of DNA damage (late). Samples were analyzed by Western blot for the indicated proteins.

9. Cells were treated as in (G) and recovery was induced by addition of caffeine for 8 h. Mitotic index was analyzed by FACS and normalized to the control. Error bars represent SD of three independent experiments. Error bars represent SD, n = 3. Statistical significance was tested using a paired two‐tailed t‐test (NS P > 0.05, **P ≤ 0.01).

10. Model of Tlk2/Asf1A‐dependent recovery competence. Red circle denotes phosphorylation event at S166, and white circles denote histones. See text for further details.

Figure EV5. Analysis of recovery competence

1. HCT116‐p53^+/+ and HCT116‐p53^−/− cells were transfected with the indicated siRNAs, synchronized in G2, and damaged with 10 Gy of IR. Cells were harvested 24 h later, and the mitotic index was analyzed by FACS. Error bars represent SD, n = 3. Statistical significance was tested using a paired two‐tailed t‐test (*P ≤ 0.05).

2. HCT116‐p53^+/+ and HCT116‐p53^−/− cells were treated as in (A) and analyzed by Western blot to determine the knockdown efficiency of the indicated siRNAs and levels of p53.

3. HCT116‐p53^−/− and HCT116‐p53^+/+ cells were either released for 24 h from a 24‐h thymidine block into medium containing nocodazole or released for 7 h into G2, irradiated with 10 Gy IR, and released for 24 h into medium containing nocodazole. Mitotic index was analyzed by FACS and normalized to the control‐treated cells. Error bars represent SD, n = 3. Statistical significance was tested using a paired two‐tailed t‐test (NS P > 0.05, *P ≤ 0.05).

4. FLAG‐Asf1A‐4D‐inducible U2OS cells were treated, synchronized in G2 by a single thymidine block and damaged in G2 by 0.5 μM adriamycin for 1 h, and were released for 16 h. Expression was induced by addition of doxycycline. mRNA was isolated and analyzed by qPCR. Error bars represent the SD of three independent experiments, n = 2.

5. Cells were treated as in (D), and samples were analyzed by Western blot with the indicated antibodies.