The effect of repeat length on Marcal1-dependent single-strand annealing in Drosophila

Abstract

Proper repair of DNA double-strand breaks is essential to the maintenance of genomic stability and avoidance of genetic disease. Organisms have many ways of repairing double-strand breaks, including the use of homologous sequences through homology-directed repair. While homology-directed repair is often error free, in single-strand annealing homologous repeats flanking a double-strand break are annealed to one another, leading to the deletion of one repeat and the intervening sequences. Studies in yeast have shown a relationship between the length of the repeat and single-strand annealing efficacy. We sought to determine the effects of homology length on single-strand annealing in Drosophila, as Drosophila uses a different annealing enzyme (Marcal1) than yeast. Using an in vivo single-strand annealing assay, we show that 50 base pairs are insufficient to promote single-strand annealing and that 500-2,000 base pairs are required for maximum efficiency. Loss of Marcal1 generally followed the same homology length trend as wild-type flies, with single-strand annealing frequencies reduced to about a third of wild-type frequencies regardless of homology length. Interestingly, we find a difference in single-strand annealing rates between 500-base pair homologies that align to the annealing target either nearer or further from the double-strand break, a phenomenon that may be explained by Marcal1 dynamics. This study gives insights into Marcal1 function and provides important information to guide the design of genome engineering strategies that use single-strand annealing to integrate linear DNA constructs into a chromosomal double-strand break.

Keywords: DNA repair; DSB repair; annealing.

Fig. 1.

P{wIw} assay design and outcomes. a) Homologies used in the P{wIw} assay. The original assay (Rong and Golic 2003) uses a nonfunctional mini-white gene with part of exon 1 deleted. This is upstream (5′) of the I-SceI site. Our assay uses different upstream homology lengths corresponding to different regions of the functional, downstream (3′) mini-w gene (yellow boxes). b–e) Outcomes of the assay. b) Imprecise nonhomologous end-joining (NHEJ/EJ) results in a mutated I-SceI cut site and leads to a red eye in progeny. c) Resection followed by DNA polymerase TMEJ results in a mutated or deleted I-SceI cut site, preventing further cutting and resulting in a red eye in progeny. d) Resection followed by SSA results in a distinct deletion for each homology class (SSA product) and white eyes in progeny. e) Some deletions are larger than expected by canonical NHEJ or TMEJ; we refer to these are cryptic end-joining since their origins have not been determined.

Fig. 2.

Length of homology affects SSA efficiency in wild-type and Marcal1 mutant flies. a) Percent of total progeny with white eyes in wild-type P{wIw} crosses. *P < 0.05 vs 3.5-kb homology, ^#P < 0.05 vs 2-kb homology (5′ and 3′), ^{^}P < 0.05 vs 500-bp homology (5′ and 3′), ***P < 0.0001 between 500 bp 5′ and 500 bp 3′ (ANOVA with Tukey’s post hoc test). b) Pie charts showing the white-eye vs red-eye progeny for each wild-type assay. While the 3.5- and 2-kb homologies provide good templates for SSA (top row), as homology length is reduced, SSA efficiency diminishes (bottom row). c) Percent of total progeny with white eyes in Marcal1 P{wIw} crosses. *P < 0.05 vs 3.5-kb homology, ^#P < 0.05 vs 2-kb homology (5′ and 3′), ^&P < 0.05 vs 2-kb 3′ homology, ^P < 0.05 vs 500-bp homology (5′ and 3′) (ANOVA with Tukey’s post hoc test). For 50 bp 3′, n = 1,286 progeny; other n values are in Table 1. d) Pie charts showing the white-eyed vs red-eyed progeny for each assay in Marcal1 mutant flies.

Fig. 3.

Molecular analysis of white-eyed progeny. In white-eyed progeny, the repaired region was amplified by PCR to determine whether region was repaired by SSA, producing a distinct product size or cryptic EJ, producing a larger or smaller product. Data from wild-type flies and Marcal1 mutants were compared by Fisher’s exact test. a) Summed data from all assays except 50 bp. n = 144 for wt; n = 117 for Marcal1; ****P < 0.0001. b) 3.5-kb homology. n = 29 for wild type, 24 for Marcal1; **P < 0.01. c) 2-kb 5′ and 3′ homologies. For 5′ n = 30 for wild type, 15 for Marcal1; for 3′ n = 26 for wild type, 43 for Marcal1; ***P < 0.001. d) 500-bp 5′ and 3′ homologies. For 5′ n = 30 for wild type, 25 for Marcal1; for 3′ n = 29 for wild type, 25 for Marcal1; n.s., P = 0.7165; **P <0.01.

Fig. 4.

Possible mechanism for Marcal1 SSA based on differences between 5′ and 3′ 500-bp homologies. a) Proposed Marcal1 mechanism for the 3′ 500-bp homology. Upon I-SceI cutting (blue) and 5′ resection, Marcal1 (purple ring) localizes to the ssDNA–dsDNA interface created by resection. Marcal1 translocates along the ssDNA (indicated by purple arrow) until it finds a region of complementarity (yellow) and promotes annealing. b) Proposed Marcal1 mechanism for the 5′ 500-bp homology. Marcal1 (purple ring) localizes to the ssDNA–dsDNA interface created by resection and then translocates (purple arrow) toward the initial cut site (blue). However, since the 5′ 500-bp homology (yellow) of the functional mini-white gene is further from the ssDNA–dsDNA resection end point, Marcal1 must translocate a greater distance to reach homology and a higher probability of dissociation before reaching the region of complementarity. This may provide a longer window during which DNA polymerase theta can engage the ends to carry out TMEJ, resulting in a reduction in the SSA outcome.