Annealing of Complementary DNA Sequences During Double-Strand Break Repair in Drosophila Is Mediated by the Ortholog of SMARCAL1

Abstract

DNA double-strand breaks (DSBs) pose a serious threat to genomic integrity. If unrepaired, they can lead to chromosome fragmentation and cell death. If repaired incorrectly, they can cause mutations and chromosome rearrangements. DSBs are repaired using end-joining or homology-directed repair strategies, with the predominant form of homology-directed repair being synthesis-dependent strand annealing (SDSA). SDSA is the first defense against genomic rearrangements and information loss during DSB repair, making it a vital component of cell health and an attractive target for chemotherapeutic development. SDSA has also been proposed to be the primary mechanism for integration of large insertions during genome editing with CRISPR/Cas9. Despite the central role for SDSA in genome stability, little is known about the defining step: annealing. We hypothesized that annealing during SDSA is performed by the annealing helicase SMARCAL1, which can anneal RPA-coated single DNA strands during replication-associated DNA damage repair. We used unique genetic tools in Drosophila melanogaster to test whether the fly ortholog of SMARCAL1, Marcal1, mediates annealing during SDSA. Repair that requires annealing is significantly reduced in Marcal1 null mutants in both synthesis-dependent and synthesis-independent (single-strand annealing) assays. Elimination of the ATP-binding activity of Marcal1 also reduced annealing-dependent repair, suggesting that the annealing activity requires translocation along DNA. Unlike the null mutant, however, the ATP-binding defect mutant showed reduced end joining, shedding light on the interaction between SDSA and end-joining pathways.

Keywords: DSB repair; Drosophila; SDSA; SMARCAL1; homologous recombination.

Figure 1

DSB repair strategies. Blue, double-stranded DNA (dsDNA) molecule; orange, dsDNA template (sister chromatid or homologous chromosome). (A) A DSB occurs in the blue DNA molecule. (B) 5′ resection marks the first step of HDR and results in 3′ ssDNA tails. (C) Rad51-coated ssDNA tail invades a template duplex, displacing one strand to create a D-loop, and primes synthesis. (D) The D-loop is disassembled and a complementarity test between the opposing ends of the break occurs. (E) SDSA is defined by annealing between complementary sequences, followed by trimming and/or gap filling. (F) Ligation restores an intact duplex DNA molecule. Alternative strategies (dotted arrows): (G) cNHEJ can occur instead of resection, which directly ligates the ends and can generate small insertions and deletions (pink segment). (H) MMEJ (catalyzed by DNA polymerase θ in metazoans) can occur prior to strand exchange or after failure to find or anneal at complementary sequences. (I) MMEJ/TMEJ will usually generate a deletion or insertion (pink segment). (J) If the DSB occurs between two direct repeats, complementary sequences may be exposed during resection, and annealing can occur without synthesis, called SSA. (K) SSA results in deletion of one repeat. (L) Second-end capture (annealing of the opposing strand to the D-loop, allowing for extension of that strand) can occur during synthesis. (M) Ligation to the opposing 5′ ends creates a dHJ. (N) Dissolution of the dHJ involves migration of the junctions toward each other and decatenation via topoisomerase activity to (O) restore the DNA molecule. (P) Resolution involves endonucleolytic cleavage of the junctions which can be cut in either orientation, resulting in both (O) noncrossover (restoration of the DNA molecule) and (Q) crossover (recombinant) products.

Figure 2

Marcal1 mutants are sensitive to killing by DSB-inducing agents. Flies heterozygous for null mutations in Marcal1 were mated in two broods of at least 10 vials, with each vial representing a biological replicate. Brood one was unexposed; brood two received a dose of MMS, HN2, HU, ETS, CPT, or IR during larval feeding. Relative survival was calculated as the ratio of homozygous mutant to heterozygous control adults in treated vials, normalized to the same ratio in the corresponding unexposed vial. Dotted line represents 100% relative survival. Dosage, number of replicates, and total progeny counts are in Table S1 in File S1. **** P < 0.0001 in paired t-tests between unexposed and exposed vials.

Figure 3

The P{w^a} assay for SDSA. (Inset, top right) The construct is a 14-kb P element inserted into the second intron of the essential gene sd (blue) in reverse orientation (diagram is relative to genome coordinates on X chromosome). Black segments represent P-element sequences needed for excision. Red segments are exons (boxes) and introns (lines) of a w gene, the product of which loads eye pigments when functional. This copy of w is interrupted by a copia retrotransposon (orange central box) which is flanked by two 276-bp LTRs (green directional boxes), resulting in partial loss of w function and an apricot-eyed phenotype. (A) Line representations of the construct on two sister chromatids in the male germline. Exposure to inefficient P transposase results in excision of the construct from one sister, leaving 17-nt noncomplementary overhangs on each side, and the ends are resected. (B) One of the 3′ ssDNA tails invades the intact sister to initiate synthesis. If D-loop dissociation is defective, the D-loop is cleaved, creating a deletion into an sd exon on one or both sides of the construct. When mated to a homozygous P{w^a} female, the progeny with flanking deletions will have a yellow eye (due to the full copy of the construct from the mother) but the event will be male lethal in subsequent generations (∼0% of progeny from wild-type males have this phenotype). (C) Premature EJ after two-ended strand exchange and some synthesis results in complete loss of w function; progeny will have yellow eyes and viable males in subsequent generations (∼8% of progeny from wild-type males). (D) Synthesis of the LTRs followed by annealing, tail clipping, and gap filling restores w function and progeny have a red eye (∼5% of progeny from wild-type males). (E) Synthesis to the LTRs followed by inappropriate EJ in copia results in an apricot eye in the progeny and is indistinguishable from nonexcision or full gene conversion events via dHJ intermediates (∼87% of progeny from wild-type males). These progeny are scored but not counted as repair events.

Figure 4

Marcal1 mutants have reduced SDSA capacity in the P{w^a} assay. (A) SDSA events are measured as the percentage of scored progeny with red eyes. Mean and SEM are indicated. Marcal1 null mutant, Brca2 mutant, and Marcal1 Brca2 double mutant frequencies were all significantly reduced compared to wild type. The numbers of single males (biological replicates) and total progeny scored are listed below the graph. (B) EJ events were measured as the percentage of scored progeny with yellow eyes. Brca2 and Marcal1 Brca2 mutants had significantly elevated EJ compared to wild type and Marcal1 single mutants. P-values: **** P < 0.0001, ** P < 0.002, * P < 0.05, based on parametric ANOVA. (C) Synthesis tracts in repair events recovered in yellow-eyed progeny were measured using a series of PCRs. Each interval was measured independently and quantified as a percentage of total independent events analyzed. x-axis denotes distance (in nucleotides) from each end of the gap, on the same scale as the schematic of P{w^a} below. y-axis is percent of events analyzed that had a positive PCR and therefore synthesized at least as far as the most internal primer. Marcal1 (n = 90) was not significantly different from wild type (n = 48) when corrected for multiple intervals (Materials and Methods). Blm (n = 75) mutants were significantly different (P < 0.0001) from both wild type and Marcal1.

Figure 5

Marcal1 mutants have reduced annealing capacity in the P{wIw} assay. (Inset, top right) The construct is on chromosome 3 and consists of two copies of the mini-w gene which are tandemly arrayed and separated by a linker containing an I-SceI recognition site. The upstream copy is missing exon 1 and intron 1, rendering it nonfunctional, while the downstream copy is functional (represented by a red glow in the schematic) and produces a wild-type red eye in a w null background. The upstream mini-w has a 5′ FRT site and the downstream copy has an FRT insertion in intron 1. PCR amplification of the FRT anchored in mini-w yields different size products for each gene, which is used to identify the presence or absence of each copy. (A) The assay is performed in the male germline with P{wIw} heterozygous. I-SceI is expressed via heat shock during larval development, resulting in DSBs with 4-nt overhangs on both sister chromatids. (B) Mutational cNHEJ can occur at the cut site, yielding red-eyed progeny and the presence of both FRTs but a mutated I-SceI cut site. (C) If resection is insufficient to expose complementary sequences, MMEJ/TMEJ can give white-eyed progeny with a deletion that includes the downstream FRT. These events cannot be differentiated from products of SSA unless the deletion is sufficiently large (>1000 bp); if no difference in size was observed, these events were classified as SSA. (D) Full resection of at least 3.6 kb reveals complementarity between the ssDNA ends that can be annealed. This results in white-eyed progeny with only the upstream FRT site; such events were categorized as SSA. (E) Excessive resection can result in deletion of the entire construct, resulting in white-eyed progeny with no amplification products. (F) Percentage of total progeny with white or red eyes for wild type and Marcal1 mutants. P < 0.0001 using χ² test. (G) Percentage of repair products that involved annealing, after molecular analysis correction (Figure S2 in File S1). Marcal1 null mutants had significantly reduced annealing compared to wild type, **** P < 0.0001. Error bars are SD. Biological replicates (single males) and total progeny scored are denoted below each genotype.

Figure 6

Marcal1^K275M mutants have reduced SDSA and EJ capacity. (A) SDSA and (B) EJ events were measured in Marcal1^K275M mutants as described in Figure 4. SDSA was similar between Marcal1^K275M and Marcal1 null, but EJ events were significantly reduced in Marcal1^K275M compared to both wild type and Marcal1 null mutants. (C) Synthesis-tract lengths measured as described in Figure 4C. No significant differences were found between Marcal1^K275M (n = 52) and wild type (n = 48) or Marcal1 null mutants (n = 90). (D) Marcal1^K275M heterozygotes and Marcal1^K275M mutants were tested in the P{wIw} SSA assay as described in Figure 5. Heterozygotes had 96% white-eyed progeny, which is not statistically different from wild type (Figure 5F) based on a parametric ANOVA test. Marcal1^K275M mutants had 67% white-eyed progeny, which was significantly reduced compared to heterozygotes but not significantly different from Marcal1 null mutants (Figure 5F). * P = 0.0185; **** P < 0.0001.