Rad52 sumoylation prevents the toxicity of unproductive Rad51 filaments independently of the anti-recombinase Srs2

Abstract

The budding yeast Srs2 is the archetype of helicases that regulate several aspects of homologous recombination (HR) to maintain genomic stability. Srs2 inhibits HR at replication forks and prevents high frequencies of crossing-over. Additionally, sensitivity to DNA damage and synthetic lethality with replication and recombination mutants are phenotypes that can only be attributed to another role of Srs2: the elimination of lethal intermediates formed by recombination proteins. To shed light on these intermediates, we searched for mutations that bypass the requirement of Srs2 in DNA repair without affecting HR. Remarkably, we isolated rad52-L264P, a novel allele of RAD52, a gene that encodes one of the most central recombination proteins in yeast. This mutation suppresses a broad spectrum of srs2Δ phenotypes in haploid cells, such as UV and γ-ray sensitivities as well as synthetic lethality with replication and recombination mutants, while it does not significantly affect Rad52 functions in HR and DNA repair. Extensive analysis of the genetic interactions between rad52-L264P and srs2Δ shows that rad52-L264P bypasses the requirement for Srs2 specifically for the prevention of toxic Rad51 filaments. Conversely, this Rad52 mutant cannot restore viability of srs2Δ cells that accumulate intertwined recombination intermediates which are normally processed by Srs2 post-synaptic functions. The avoidance of toxic Rad51 filaments by Rad52-L264P can be explained by a modification of its Rad51 filament mediator activity, as indicated by Chromatin immunoprecipitation and biochemical analysis. Remarkably, sensitivity to DNA damage of srs2Δ cells can also be overcome by stimulating Rad52 sumoylation through overexpression of the sumo-ligase SIZ2, or by replacing Rad52 by a Rad52-SUMO fusion protein. We propose that, like the rad52-L264P mutation, sumoylation modifies Rad52 activity thereby changing the properties of Rad51 filaments. This conclusion is strengthened by the finding that Rad52 is often associated with complete Rad51 filaments in vitro.

Figure 1. Characterization of rad52-L264P, a suppressor of the MMS sensitivity of srs2Δ mutants that avoid the formation of toxic recombination intermediates.

(A) Serial 10-fold dilutions of haploid strains with the indicated genotypes were plated onto rich media (YPD) containing different MMS concentrations. rad52-L264P* denotes the original isolated mutant strain and rad52-L264P** denotes a strain in which the RAD52 gene was replaced by the mutant newly generated by directed mutagenesis. (B) Conservation of a motif comprising L264. The primary structure of Rad52 is schematized showing the conserved N-terminus moiety containing the major DNA binding and self-association domains (black, amino acids 1 to 179) as well as the C-terminus part (white, amino acids 180 to 471) containing the RPA (amino acids 275 to 278) and the Rad51 (amino acids 376 to 379) binding domains. The alignment of the Rad52 protein in Hemiascomycetes species shows the conservation of a domain containing the non-essential L264 residue and the QDDD residues essential for RPA binding and consequently for Rad52 mediator activity. The color code used in the alignment follows the default ClustalX color scheme as implemented in JalView (see Material and Methods). Cyan is for fully hydrophobic (I, L, V, M, F), turquoise for aromatic residues containing polar moieties (Y, H), green for small polar (T, S), purple for acidic (D, E), orange for glycine (G) and dark yellow for proline (P) residues.

Figure 2. rad52-L264P does not affect DNA repair or HR, whereas it completely suppresses srs2Δ phenotypes.

(A, B) Survival curves of haploid cells grown in log phase culture exposed to γ-ray or UV light. (C, D, E) Survival curves and heteroallelic HR frequency of diploid cells grown in log phase culture exposed to γ-ray or UV light. The results shown are the average of at least 3 independent experiments.

Figure 9. Rad52-L264P behaves like Rad52-SUMO.

(A to E) Spot assay of haploid cells with the indicated genotype on rich medium (YPD) containing increasing MMS concentrations. Note that the deletion of the SIZ2 gene (A) and the rad52-3KR allele (B) cannot suppress the MMS sensitivity of srs2Δ cells. Therefore, the MMS resistance of rad52-L264P siz2Δ srs2Δ and rad52-3KR-L264P srs2Δ cells is only related to rad52-L264P. (D) Note that the haploid strain spotted at the bottom contains both rad52-3KR and RAD52 alleles.

Figure 3. rad52-L264P suppresses mutations that are synthetically lethal with srs2Δ.

(A) Tetrad analysis of crosses between haploid rad52-L264P srs2Δ strains and haploid mutants synthetically lethal with srs2Δ. Double mutant spores, which do not contain rad52-L264P, are indicated by white squares. The white circles mark triple mutants. (B) siz2Δ and rad52-3KR do not suppress the synthetic lethality of srs2Δ sgs1Δ and srs2Δ rrm3Δ mutants. In crosses involving siz2Δ, white squares display spores of srs2Δ sgs1Δ or srs2Δ rrm3Δ genotypes and white circles indicate triple mutants. To analyze the genetic interaction between rad52-3KR inserted at the URA3 locus and the synthetically lethal rrm3Δ srs2Δ double mutant, diploids homozygous for rad52Δ were sporulated, in order to avoid the co-expression of RAD52 and rad52-3KR. The white square indicates srs2Δ rrm3Δ rad52Δ triple mutants and white circles indicate unviable srs2Δ rrm3Δ rad52Δ ura3::rad52-3KR monosporic colonies.

Figure 4. rad52-L264P can only suppress srs2Δ deficiencies in the management of unproductive Rad51 filaments.

Schematic representation and Southern blot analysis of the two HO-induced DSB repair systems involving SSA between two direct repeats 25 kb apart (A) and gene conversion between ectopic copies of MAT (D). The kinetic of repair in the SSA system after HO induction by addition of galactose to the medium was monitored by probing a Southern blot of KpnI (K) digested genomic DNA of cells harvested at the indicated time with a PCR fragment complementary to the 3′end of LEU2 (bold line). Quantification of the product band relative to the parental band (leu2::cs) measured at 24 hours is indicated (see Material and Methods for more information). To follow gene conversion of the MATα allele after DSB induction, DNA was digested with ClaI (C1) and HindIII (H3), and probed with a MAT distal PCR fragment (bold line). The two possible outcomes, gene conversion associated with a CO or not (NCO), are indicated. The proportion of repaired products (NCO+CO) relative to the parental band (MATα) and the proportion of CO among repair products measured at 10 hours are indicated (see Material and Methods for more information). (B and E) Cell viability after DSB formation in both assays. (C and F) Western blot analysis of Rad53 phosphorylation after HO induction in both systems.

Figure 5. ChIP analysis of Rad51 filament formation at a DSB created by the HO endonuclease in cells that express Rad52-FLAG, Rad52-L264P-FLAG or Rad52-SUMO-FLAG.

The HO endonuclease was induced in WT or srs2Δ cells that express Rad52-FLAG, Rad52-L264P-FLAG or Rad52-SUMO-FLAG to create a single DSB that can be repaired by SSA. Samples were taken before induction and at 2, 4, 6 and 8 hours after galactose addition. Antibodies against the RPA complex, the FLAG epitope or Rad51 were used to precipitate protein-bound chromatin. Quantitative PCR was performed with primers located at 0.6 kb or 7.6 kb from the DSB site, using the immunoprecipitated chromatin (IP) and input DNA as template. As a control, primers specific for the ARG5,6 locus were used. The relative enrichment represents the ratio of the PCR enrichment in the IP fraction to the input fraction. The median value of at least 3 experiments is shown and error bars represent the upper and lower values observed.

Figure 6. Rad52-L264P still interacts with RPA and Rad51.

(A) To test the interaction with RPA, Rad52 or Rad52-L264P was immunoprecipitated with a rabbit anti-Rad52 polyclonal antibody from 1 mg of whole cell extracts (without DNAse treatment) prepared from RAD52, rad52-L264P or rad52Δ strains. To test the robustness of the interaction, increasing NaCl concentrations were added to the cell extracts. Proteins from the whole extracts (50 µg) and from the immunoprecipitated fractions were separated by SDS-PAGE and immunoblotted with rabbit anti-Rad52 polyclonal antibody or rabbit anti-RPA polyclonal antibody (allowing detection of the Rfa1 subunit of RPA). The signals corresponding to immunoprecipitated Rad52 or Rad52-L264P were quantified in three independent experiments and plotted as a fraction of the signal intensity measured in the 150 mM NaCl experiment. (B) To assess the interaction between Rad52-L264P and Rad51, Rad51 was immunoprecipitated from 1 mg of whole cell extracts (without DNAse treatment) from cells expressing Rad52-FLAG or Rad52-L264P-FLAG. The strength of the interaction was also evaluated against increasing NaCl concentrations. The proteins from whole cell extracts (50 µg) and from immunoprecipitated fractions were separated by SDS-PAGE and immunoblotted with rabbit anti-Rad51 polyclonal antibody and mouse anti-FLAG monoclonal antibody. The presence of Rad51 in the immunoprecipitated fraction cannot be detected to validate the efficiency of the immunoprecipitation because it migrates at the same level as the IgG anti-Rad51 used for the immunoprecipitation. However, the absence of Rad52 in the rad51Δ immunoprecipitate confirmed that the Rad52-FLAG signals observed are related to the Rad52-Rad51 interaction. The signals corresponding to immunoprecipitated Rad52 or Rad52-L264P were quantified in three independent experiments and plotted as in (A).

Figure 7. Biochemical analysis of Rad52-L264P-mediated filaments.

(A) The purity of recombinant Rad52 and Rad52-L264P (2 µg/each) was assessed by separation on 8% SDS-PAGE and staining with Coomassie blue. (B) Binding of Rad52-L264P to ssDNA. Protein titration reactions were performed by incubating 0.27 µM of a 62-nucleotides long Cy5-labeled ssDNA fragment with various amounts of Rad52-L264P at 37°C for 10 min (protein/bases: 1/10, 1/5, 1/2.5, 1/1.25, 1/0.6, 1/0.3, 1/0.15). Quantification of free ssDNA is shown. The data were fitted into a sigmoidal curve by using the PRISM software (GraphPad). (C) Rad52-L264P-mediated DNA annealing. Representative gels of Rad52 or Rad52-L264P-promoted DNA annealing are shown in the upper panel (Rad52/bases: 1/100, 1/42, 1/14, 1/6; same DNA as in (B) with reverse-complement, 340 nM each). The dsDNA/total DNA ratio at 10 min is shown in the lower panel. (D) RPA bound to ssDNA inhibits equally Rad52- and Rad52-L264P-catalyzed annealing reactions. Reactions were carried out with primers 25 and 26 (200 nM each, see Material and Methods) that were first incubated with 30 nM RPA (1/13 bases) at 30°C for 5 minutes, followed by addition of 40 nM Rad52 (1/10 bases). Self-annealing of the primers incubated without proteins and reactions performed without RPA or Rad52 are also shown. (E) Over-stimulation of DNA strand exchange by Rad52-L264P. Upper panel, diagram of the reaction substrates and products. Middle panel, ethidium bromide-stained DNA gel. As shown by the standard reaction (st), Rad51 efficiently catalyzes the formation of nicked circular products. Pre-bound RPA inhibits this reaction (line 2). Increasing amounts of Rad52 (lines 3–6, Rad52/bases: 1/55, 1/27, 1/18, 1/14) or Rad52-L264P (lines 7–10, Rad52/bases: 1/880, 1/220, 1/110, 1/55) overcome the inhibitory effect of pre-bound RPA. In line 1, only RPA was added to the reaction. Lower panel, the ratio of the nicked circular product over the sum of the linear dsDNA substrate and the nicked circular product is shown. (F) Salt titration of Rad51-Rad52/Rad52-L264P-ssDNA complex formation. The nucleoprotein complexes were assembled by incubating 0.8 µM Rad51 with 0.09 µM Rad52 or Rad52-L264P and 0.08 µM RPA pre-bound to 2.5 µM Cy5-labeled ssDNA (400 nucleotides) in the presence of the indicated NaCl concentrations at 37°C for 15 min. The Cy5 signals after nucleoprotein gel eletrophoresis are shown. Quantifications are shown below. Data were fitted into a third order polynomial curve. Western blot analysis of the gel using antibodies against Rad51 or Rad52 is also shown. Stars indicate signals corresponding to proteins not bound to ssDNA. (G) Transmission electron microscopy images of protein-DNA complexes showing the association of Rad52 with complete Rad51 filaments. Positive (left) and negative (right) staining images are shown for each type of filaments. The proportion of each type of complete Rad51 filaments formed by Rad52 or Rad52-L264P at different NaCl concentrations is shown and compared to a control reaction without Rad52. 100 molecules were examined in each experiment.

Figure 8. Over-expression of the SIZ2 SUMO-ligase coding gene suppresses the MMS sensitivity of srs2Δ by stimulating Rad52 sumoylation.

(A) Spot assay of haploid cells over-expressing SIZ2. Serial 10-fold dilutions were plated on minimal media lacking uracil with or without MMS. Strains of the indicated genotype were transformed with an empty vector or with the same plasmid containing the SIZ2 gene. (B) Over-expression of SIZ2 stimulates Rad52 sumoylation. Proteins conjugated with a His₇-SUMO radical were pull-down on Ni-NTA from 5 mg of extracts of RAD52-FLAG cells over-expressing His7-SMT3. Pull-downs were carried out from strains transformed with a SIZ2-containing multi-copy vector or with an empty vector. Cells treated with 0.3% of MMS were also tested as a positive control of sumoylation. rad52-L264P-FLAG strains were also subjected to pull-down analysis. Proteins from the whole extracts (3 µg) and from the pull-down fractions were separated by SDS-PAGE and immunoblotted with an anti-FLAG mouse monoclonal antibody.

Figure 10. Rad52 sumoylation prevents the toxicity of unproductive Rad51 filaments.

(A) Schematic representation of the two kinds of toxic recombination intermediates eliminated by Srs2 in WT cells. We found that srs2Δ haploid cells sensitivity to DNA damage is related to recombination-deficient Rad51 nucleoprotein filaments. The toxicity of such filaments disappears in rad52-L264P srs2Δ cells, or alternatively, the intermediates themselves are not formed. However, srs2Δ cells sensitivity to DNA damage related with toxic intertwined HR intermediates cannot be suppressed by this allele. Unproductive Rad51 filaments can be formed after resection of a DSB located in a unique sequence in the genome. In a situation where homologous dsDNA cannot be found by the recombinase, Srs2 is essential to remove Rad51 filaments to allow alternative repair pathways such as SSA. Srs2 could also edit Rad51 filaments improperly nucleated by Rad52. In srs2Δ cells, nonrecombinogenic Rad51 filaments could also accumulate on ssDNAs generated from the uncoupling between the helicase complex opening replicative dsDNA and the DNA synthesis machinery in replicative mutants such as mrc1Δ. Finally, when a stable paranemic joint cannot be processed from a plectonemic joint because of mutations in genes involved in late recombination steps, Srs2 is necessary to address lesions to other DNA repair pathways. Intertwined recombination intermediates that occur between homologous chromosomes in diploid cells or between ectopic chromosomes in haploid cells cannot be suppressed by rad52-L264P. (B) Unproductive Rad51 filaments mediated by Rad52-SUMO are not toxic even in srs2Δ cells. When mediated by Rad52, Rad51 filaments that cannot complete strand invasion have to be removed by Srs2 in order to allow SSA or post-replication repair processes (PRR). Conversely, Rad52-SUMO (or Rad52-L264P) might lower (or shorten) Rad51 filaments. These modified mediators might also change Rad51 filament properties as indicated by Rad52 occupancy on Rad51 filaments. These changes might suppress Rad51 filaments toxicity, thereby bypassing the need for Srs2. Rad51 filaments might be removed by a Srs2-independent process.