Drosophila brca2 is required for mitotic and meiotic DNA repair and efficient activation of the meiotic recombination checkpoint

Abstract

Heterozygous mutations in the tumor suppressor BRCA2 confer a high risk of breast and other cancers in humans. BRCA2 maintains genome stability in part through the regulation of Rad51-dependent homologous recombination. Much about its precise function in the DNA damage responses is, however, not yet known. We have made null mutations in the Drosophila homolog of BRCA2 and measured the levels of homologous recombination, non-homologous end-joining, and single-strand annealing in the pre-meiotic germline of Drosophila males. We show that repair by homologous recombination is dramatically decreased in Drosophila brca2 mutants. Instead, large flanking deletions are formed, and repair by the non-conservative single-strand annealing pathway predominates. We further show that during meiosis, Drosophila Brca2 has a dual role in the repair of meiotic double-stranded breaks and the efficient activation of the meiotic recombination checkpoint. The eggshell patterning defects that result from activation of the meiotic recombination checkpoint in other meiotic DNA repair mutants can be strongly suppressed by mutations in brca2. In addition, Brca2 co-immunoprecipitates with the checkpoint protein Rad9, suggesting a direct role for Brca2 in the transduction of the meiotic recombination checkpoint signal.

Figure 1. brca2 Mutations Dramatically Decrease HR-h Repair, Increase SSA Repair, and Result in Increased Deletion Formation

(A) The Rr3 assay measures the relative repair pathway usage for repair of I-SceI endonuclease-cut breaks in the pre-meiotic germline of Drosophila males. The Rr3 transgene contains an I-SceI site flanked by two partial copies of dsRed. One dsRed copy contains a 147 bp duplication. Repair by SSA reconstitutes dsRed and results in red fluorescent progeny. Repair by NHEJ modifies the I-SceI site and results in progeny that are dark regardless of the presence of the endonuclease. The EJ1 transgene is a variant of Rr3 which has a non-functional I-SceI site and a 16 bp deletion 156 bp from the I-SceI site. Progeny resulting from repair by HR-h are also dark even when the endonuclease is present and can be distinguished from NHEJ by PCR. PCR analysis also distinguishes between short and long tract HR-h; long-tract HR-h results in the inclusion of the deletion on EJ1 and a smaller product. Deletions are measured by the loss of the w ⁺ marker. (B) In Cross 2, the EJ1 chromosome was present, and repair by HR-h could occur. Error bars represent the standard errors between individual males; P-values were calculated using a permutation test described in [31]. Number of males/total Rr3-carrying progeny counted for OreR, brca2, and okra, respectively: 72/3861, 72/3163, and 78/3486. (C) In Cross 1, the EJ1 chromosome was not present, and repair by HR-h was not available. Number of males/total Rr3-carrying progeny counted for OreR, brca2, and okra, respectively: 66/7103, 67/6054, and 65/5546. ^aThe deletion class was not isolated prior to sorting and therefore represents a small portion of the NHEJ class.

Figure 2. Persistent Double-Stranded Breaks in brca2 Mutants Activate the Meiotic Recombination Checkpoint

(A) Egg chambers stained for DNA (green) and membranes (red). Percentage shown is the average number of wild-type-like, spherical karyosomes from two independent experiments with a total of 175–200 nuclei counted. Insert shows the region containing the oocyte nucleus. (B) Eggshells laid by wild-type and brca2 ^KO/56 mutant females. (C) Frequency of the different classes of eggshells laid by females fed yeast for 3–4 days. Class 1 eggs have two separate dorsal appendages and appear wild-type with respect to dorsal patterning. Class 2 eggs have 2 dorsal appendages that are fused at the base. Class 3 eggs have a single dorsal appendage. Class 4 eggs have no dorsal appendages and represent the complete loss of dorsal fate. Error bars represent the standard error between 3 independent experiments. Total number of eggs counted per genotype was between 500 and 3,000 eggs. P(CG30169}i23 and P(CG30169}i48 are genomic rescue constructs inserted at two different chromosomal locations.

Figure 3. brca2 Mutants Suppress the Patterning Defects of DNA Repair Mutants That Activate the Meiotic Recombination Checkpoint

(A) γ-H2AV(red) foci in the germarium, with the oocyte in region 3 marked by the arrow. Pro-oocytes are marked by C(3)G staining (blue). Images represent a projection of stacks taken through whole germaria. Number of γ-H2AV foci in region 3 is the average number of foci +/− standard error from a total of 14, 14, 11, 7, 10, and 10 region 3 oocytes counted for OreR, brca2, spnA, okra, brca2; spnA, and brca2 okra, respectively. (B) Frequency of the different classes of eggshells laid by females fed yeast for 5–7 days. Classes are as described in Figure 2. Note the increase in the wild-type like eggshells (blue) in the double mutants. Error bars represent the standard error between 3 independent experiments. Total number of eggs counted per genotype were: 1149, 73, 1687, 456, 645, 1663, 922, 1199, 341, 339, 455, 268, and 196, respectively.

Figure 4. brca2 Is Not Required for the Checkpoints That Respond to Irradiation Damage

Climbing third instar larvae were irradiated and then dissected prior to fixation. Error bars represent the standard error from a total of 6–10 samples from 2–3 separate experiments. Representative images can be found in Figure S1. (A) Larvae were fixed 1 hour after a 4,000 Rad IR exposure and stained for the mitotic marker phospho-histone H3. n = 6. (B) Larvae were fixed 4 hours after a 4,000 Rad IR exposure and stained for activated Caspase-3. n = 6. (C) Larvae were fixed 1 hour after a 500 Rad IR exposure and stained for the mitotic marker phospho-histone H3. n = 11. (D) Larvae were inverted 1.5 hours after a 4,000 Rad IR exposure and incubated in Schneider's media with BrdU for 30 minutes. n = 7–10.

Figure 5. Brca2 Co-Immunoprecipitates with Rad9, a Member of the 9-1-1 Checkpoint Complex

Western blots from FLAG-Rad9 immunoprecipitations probed for FLAG and HA epitopes. Approximately 1/20 of the lysate prior to immunoprecipitation and the entire immunoprecipitated protein fraction was loaded. Ovaries from females carrying single copies of P(UAS-Brca2-HA} and/or P(UAS-FLAG- Rad9} and the germline driver nanos-Gal4-VP16 were harvested for the immunoprecipitation.