Research ArticleMOLECULAR BIOLOGY

Phosphoregulation of Rad51/Rad52 by CDK1 functions as a molecular switch for cell cycle–specific activation of homologous recombination

See allHide authors and affiliations

Science Advances  07 Feb 2020:
Vol. 6, no. 6, eaay2669
DOI: 10.1126/sciadv.aay2669
  • Fig. 1 Rad51 and Rad52 are phosphorylated by Cdc28 in the G2/M phase.

    (A) Results from the kinase assay for the phosphorylation of Rad51 (left) and Rad52 (right) using the Cdc28-as1 mutant in vitro. Cdc28-as1-TAP was purified by anti-TAP immunoprecipitation from asynchronous cells. GST-purified Rad51 and Rad52 were used as substrates, and 1NM-PP1 (5 μM) was added to cells to suppress the kinase activity of Cdc28-as1. (B) Results from the kinase assay for the phosphorylation of Rad51 (left) and Rad52 (right) using Cdc28 extracted from cell cycle–arrested cells in vitro. The cell cycle was arrested for 3 hours with α-factor (150 μM) to arrest G1 and nocodazole (15 μg ml−1) to arrest G2/M. Async, asynchronous cells. (C) Analysis of the in vivo phosphorylation level of Rad51 and Rad52 through the cell cycle. The cell cycle was arrested for 3 hours with α-factor (150 μM) to arrest G1. Synchronized cells were released to fresh YPD media with nocodazole (15 μg ml−1) for 45 min. After 45 min of incubation, 5 μM 1NM-PP1 or DMSO was directly added to cell cultures for 1 hour. The DNA content data were analyzed by flow cytometry (left). 1C and 2C indicate single and double DNA haploid content, respectively. Phos-tag (50 μM) and 100 μM MnCl2 were mixed with 6% separating gel for analysis of phospho-Rad51 (right panel). The relative ratio of phosphorylated proteins to total proteins is shown below each lane.

  • Fig. 2 Cdc28 combined with Clb2 or Clb3 phosphorylates Ser125 and Ser375 of Rad51 and Thr412 of Rad52.

    (A) Results from the kinase assay for the phosphorylation of Rad51 using purified Cdc28 and cyclins in vitro. Each cyclin was purified by anti-TAP immunoprecipitation from asynchronous cells. Red asterisks indicate bands of corresponding cyclins. (B) Results from the kinase assay for the phosphorylation of Rad52 using purified Cdc28 and cyclins in vitro. (C) Results from the kinase assay using alanine substitution mutants of Rad51 in vitro. Cdc28-as1-TAP was purified by anti-TAP immunoprecipitation from asynchronous cells. GST-tagged Rad51 variants were expressed in E. coli and purified by GST pull down. S125A, S375A, and 2A indicate Rad51 mutants with alanine substitutions at Ser125, at Ser375, and at both Ser125 and Ser375, respectively. WT, wild type. (D) Results from the serial dilution assay used to assess MMS sensitivity of rad51Δ cells expressing Rad51 variants. Cells were spotted in 10-fold serial dilutions on SC medium in the absence or presence of 0.01% MMS. rad51-S125A and rad51-S375A indicate rad51Δ cells expressing the Rad51 mutant with alanine substitutions at Ser125 and Ser375, respectively. (E) Results from the kinase assay using an alanine substitution mutant of Rad52 in vitro. Cdc28-as1-TAP and GST-tagged Rad52 were purified as described in (C). T412A indicates a Rad52 mutant with an alanine substitution at Thr412.

  • Fig. 3 The G2/M-phase CDK1-dependent phosphorylation promotes the DNA binding affinity of Rad51.

    (A) Results from the serial dilution assay used to assess MMS sensitivity of rad51Δ cells expressing Rad51 variants. Cells were spotted in 10-fold serial dilutions on SC medium in the absence or presence of 0.01% MMS. rad51-2A indicates rad51Δ cells expressing the Rad51 mutant with alanine substitutions at Ser125 and Ser375. rad51-2E indicates rad51Δ cells expressing the Rad51 mutant with glutamate substitutions at Ser125 and Ser375. (B) Homologous recombination efficiency test of rad51Δ cells expressing Rad51 variants. Genomic DNA was extracted every 1 hour after 2% galactose addition and analyzed by PCR. Arrowheads indicate the PCR products of the homologous recombination intermediates. Asterisks indicate the PCR products of the control region (ARG5,6). (C) Accumulated GFP-tagged Rad51 and Rad51-2A at the DNA damage sites. To make DSB, HO endonuclease was induced during incubation with 2% galactose for 3 hours. DIC, differential interference contrast. Scale bars, 4 μm. The percentages of cells with Rad51-GFP foci are shown on the right panel. Data are presented as the means ± SD of triplicate experiments. P values were determined by Student’s t test (***P < 0.005). (D) ChIP assay was used to assess the DNA binding affinity of Rad51 and Rad51-2A. HO endonuclease was induced during incubation with 2% galactose for 3 hours. Immunoprecipitated DNA with Rad51 variants was analyzed for the MAT locus by PCR (left panel) and by quantitative real-time PCR (right panel). CUP1 was used as a negative control. Data are presented as the means ± SD of triplicate experiments. P values were determined by a one-sample t test (***P < 0.005). ns, not significant. (E) Results from the EMSA used to assess the ssDNA binding affinity of Rad51 and Rad51-2A. EMSA was performed using a binding buffer that includes 35 mM tris-Cl (pH 7.5), 5 mM ATP, 5 mM MgCl2, 50 mM KCl, bovine serum albumin (100 μg/ml), and 1 mM dithiothreitol. (F) Results from the EMSA used to assess the DNA binding affinity of the G1- and the G2/M-phase Rad51. α-Factor (150 μM) and nocodazole (15 μg ml−1) were treated to synchronize cell cycle to the G1 and G2/M phases, respectively.

  • Fig. 4 The G2/M-phase CDK1-dependent phosphorylation does not affect Rad52 functions in the initial step of homologous recombination.

    (A) Results from the serial dilution assay used to assess MMS sensitivity of rad52Δ cells expressing Rad52 variants. Cells were spotted in 10-fold serial dilutions on SC medium in the absence or presence of 0.01% MMS. rad52-T412A indicates rad52Δ cells expressing the Rad52 mutant with an alanine substitution at Thr412. rad52-T412E indicates rad52Δ cells expressing the Rad52 mutant with a glutamate substitution at Thr412. (B) Results from the homologous recombination efficiency test for rad52Δ cells expressing Rad52 variants. Genomic DNA was extracted every 1 hour after 2% galactose addition and analyzed by PCR. Arrowheads indicate the PCR products of the homologous recombination intermediates. Asterisks indicate the PCR products of the control region (ARG5,6). (C) Results from the analysis of RPA focus formation and Rad52 accumulation. Images of GFP-tagged Rfa1 and RFP-tagged Rad52 were taken every 5 min after the addition of 2% galactose (top). Scale bar, 2 μm. The percentages of cells with RPA foci (green) and Rad52 foci (red) are shown in the bottom. (D) The elapsed time between RPA foci formation and Rad52 accumulation. A box plot is shown with whiskers from the 5th to the 95th percentile, and the data were normalized to the median of measures from the RAD52 cells (n = 200). P values were determined by the Mann-Whitney U test. (E) Coimmunoprecipitation assay used to assess the binding affinity between Rad52 and Rfa1. Protein complexes with Rad52-HA were precipitated using anti-HA agarose beads. Rfa1 was detected by anti-Rfa1 antibody. The relative ratio of Rfa1 to Rad52, normalized against that of cells with WT Rad52, is shown below each lane. (F) Coimmunoprecipitation assay used to assess the binding affinity between Rad51 and Rad52. Rad51 was detected by anti-Rad51 antibody. The relative ratio of Rad51 to Rad52, normalized against that of cells with WT Rad52, is shown below each lane.

  • Fig. 5 The G2/M-phase CDK1-dependent phosphorylation facilitates the physical interaction between Rad52 proteins.

    (A) Stimulation of DNA ligation by Rad52 variants in vitro. The left panel presents the schematic of the in vitro DNA ligation test. The efficiency of multimeric product formation was evaluated by measuring the band intensity of the linear multimer (right panel). The relative amount of multimeric product, normalized against that of WT Rad52, is shown below each lane. (B) Results from the BiFC assay used to assess the effect of CDK1-dependent phosphorylation on the physical interaction between Rad52 proteins. Representative BiFC images for the interaction between Rad52 proteins are shown in the left panel. RAD52 indicates cells expressing both VN-Rad52 and VC-Rad52, and rad52-T412A indicates cells expressing both VN-Rad52-T412A and VC-Rad52-T412A. Nuclei were visualized by DAPI (1 μg ml−1) staining. Scale bars, 4 μm. The relative BiFC signal intensity in the RAD52 and rad52-T412A cells is shown in the right panel. A box plot is shown with whiskers from the 5th to the 95th percentile, and the data were normalized to the median of the measures from RAD52 cells (n > 200). P values were determined by the Mann-Whitney U test (***P < 0.005). a.u., arbitrary units. (C) Recovery of the physical interactions between Rad52-T412A proteins by the FKBP-FRB system. The illustration represents a schematic diagram of the artificial recovery of the physical interaction between Rad52-T412A proteins by the FKBP-FRB system. Genomic DNA was extracted every 1 hour after 2% galactose addition and analyzed by PCR. Rapamycin (1 μM) or DMSO was added to cell cultures 3 hours after galactose addition. Arrowheads indicate the PCR products of the homologous recombination intermediates. Asterisks indicate the PCR products of the control region (ARG5,6).

  • Fig. 6 Summary diagram for cell cycle–dependent regulation of homologous recombination by the G2/M-phase CDK1.

    In the G1 phase, Rad51 and Rad52 are not phosphorylated by Cdc28 combined with the G1 phase cyclins. In the G2/M phase, Cdc28 combined with either Clb2 or Clb3 phosphorylates both Rad51 and Rad52. When a DSB occurs, it is recognized by the RPA complex and Rad52. Subsequently, because Cdc28-dependent phosphorylation promotes DNA binding of Rad51, phosphorylated Rad51 proteins form a nucleoprotein complex with ssDNA at the DNA damage site. Rad51 binding allows strand invasion to the homologous DNA region. After DNA synthesis for repair of the DSB, the newly synthesized DNA strand dissociates from its template strand. Because Cdc28-dependent phosphorylation facilitates protein-protein interaction between Rad52 rings, the dissociated strand is annealed to the other recessed end of the DSB by superstructure formation of Rad52 rings that are associated with each strand of broken DNA. Subsequently, the DSBR pathway is completed by the gap filling and ligation process.

Supplementary Materials

  • Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/6/6/eaay2669/DC1

    Fig. S1. Deletion of either RAD51 or RAD52 impairs DNA damage repair process.

    Fig. S2. Deletion of either CLB2 or CLB3 affects the phosphorylation of Rad51 and Rad52 in cells.

    Fig. S3. CDK1 phosphorylates S125 and S375 of Rad51 and T412 of Rad52 in cells.

    Fig. S4. Nonphosphorylatable mutation of Rad51 impairs the strand invasion process even in the G2/M phase–arrested cells.

    Fig. S5. CDK1-dependent phosphorylation regulates the DNA binding affinity of Rad51.

    Fig. S6. Nonphosphorylatable mutation of Rad52 impairs ligation process even in the G2/M phase–arrested cells.

    Fig. S7. Both wild-type Rad52 and Rad52-T412A form ring structures.

    Fig. S8. Rapamycin treatment does not restore the defect in the ligation process in cells that express Rad52-T412A protein without FKBP or FRB attachment.

    Fig. S9. Full images of Western blots.

    Table S1. Yeast strains used in this study.

  • Supplementary Materials

    This PDF file includes:

    • Fig. S1. Deletion of either RAD51 or RAD52 impairs DNA damage repair process.
    • Fig. S2. Deletion of either CLB2 or CLB3 affects the phosphorylation of Rad51 and Rad52 in cells.
    • Fig. S3. CDK1 phosphorylates S125 and S375 of Rad51 and T412 of Rad52 in cells.
    • Fig. S4. Nonphosphorylatable mutation of Rad51 impairs the strand invasion process even in the G2/M phase–arrested cells.
    • Fig. S5. CDK1-dependent phosphorylation regulates the DNA binding affinity of Rad51.
    • Fig. S6. Nonphosphorylatable mutation of Rad52 impairs ligation process even in the G2/M phase–arrested cells.
    • Fig. S7. Both wild-type Rad52 and Rad52-T412A form ring structures.
    • Fig. S8. Rapamycin treatment does not restore the defect in the ligation process in cells that express Rad52-T412A protein without FKBP or FRB attachment.
    • Fig. S9. Full images of Western blots.
    • Table S1. Yeast strains used in this study.

    Download PDF

    Files in this Data Supplement:

Stay Connected to Science Advances

Navigate This Article