Experimental setup used to induce single DSBs in spermatogonia.

A) Schematic representation of the DR-white reporter, containing a white gene with an inserted I-SceI site (Sce.white) that introduces a premature stop codon, the dsRed sequence, and a truncated white gene (iwhite) [52]. dsRed and white are driven by the eye tissue-specific 3xP3 promoter and glass multiple reporter (GMR), respectively. HR repair of I-SceI induced breaks using the iwhite sequence in the germline, results in w+, i.e., flies with red eyes (for DR-white insertions in euchromatin) or variegated-red eyes (for DR-white insertions in heterochromatin) in the progeny. SSA repair results in the loss of the fluorescent marker dsRed in the eyes. Mutagenic repair, absence of cut or perfect NHEJ results in white-eyed flies (w-), which can be further characterized by PCR amplification, I-SceI digestion, and sequencing across the repair junction. B) Schematic representation of the testis, highlighting the spermatogonia and the corresponding level of activity of the Bam promoter. C) Schematic representation of the genomic location of the DR-white insertions used in this study. D) Description of some of the most distinctive features of these sites. Colors indicate the chromatin state described by the 9-state model [61]. Blue: Pericentromeric heterochromatin. Grey: ‘background’ state, typically corresponding to large domains between active genes. Het1 is located in a transposable element (TE). TE density and repeat density are qualitatively illustrated by the number of light and dark purple bars, respectively. 1 bar: 2 TEs or 10 repeats in a 20Kb window around the insertion site. Het1 and Het4 insertion sites are also located inside of an annotated lncRNA (+), while Eu, Het3, and Het2 are not (-). E) H3K9me3 levels at the insertion sites were determined by ChIP-seq data available for Kc cells. Enrichments are calculated as the ratio of reads +/- 0.5 Mb from the insertion site per million mapped reads between IP and input samples. ****p<0.0001 by two-sample z-test for means. Error bars: SD of 3 replicates. F) Scheme of the fly crosses set up to analyze repair outcomes in the F2 generation.

Frequency of HR products.

The frequency of HR products for indicated DR-white insertions was calculated as the number of red-eyed flies over the total number of flies (n) in the F2 progeny. *p=0.015; **p=0.007; ***p=0.0002; ****p<0.0001 by one-tailed unpaired t-test. Flies analyzed: n=1352 for Eu; n=1267 for Het1; n=880 for Het2; n=679 for Het3; n=485 for Het4. Error bars: SD of two independent experiments when available.

Frequency of alt-EJ repair pathways.

A) Schematic representation of MMEJ and the main SD-MMEJ mechanisms (adapted from [16]), indicating the position of repeat motifs and the corresponding primers (P1, P2) and microhomologies (MH1, MH2). For the ‘loop-out’ mechanism, the DNA is unwinded before loop formation. For the ‘snap-back’ mechanism, resection occurs before hairpin formation. In both mechanisms, annealing of the break-proximal primer (P2) to the break-distal primer (P1), initiates DNA synthesis (red arrows) that creates new microhomology sequences (MH1) and that can create insertions. P1 and P2 are direct repeats for loop-out and inverted repeats for snap-back. Repair continues through secondary structure unwinding, annealing of the newly-synthesized microhomologies with MH2 sequences on the opposite side of the break, fill-in DNA synthesis, and ligation. When P2 and MH2 are not adjacent to the break site DNA flaps form and get trimmed, resulting in long deletions. When P1 and MH1 are adjacent to each other, insertions do not occur. B) Frequency of indels, MHJ and ABJ products of mutagenic-EJ repair across the different DR-white insertions. Sequences analyzed: n=86 for Eu; n=280 for Het1; n=74 for Het2; n=104 for Het3; n=14 for Het4. All comparisons are not significant. C) Frequency of SD-MMEJ-consistent, MMEJ-consistent (i.e., SD-MMEJ-inconsistent MHJ), NHEJ (i.e, SD-MMEJ-inconsistent ABJ with less than 4bp-long alterations of the repair junction), or ‘other’ repair outcomes across different DR-white insertions. The ‘other’ category includes point mutations as well as ABJ and SD-MMEJ-inconsistent deletions or indels greater than 4bp. **p<0.005; ****p<0.0005 by two-proportion z-test, n=86 for Eu; n=280 for Het1; n=74 for Het2; n=104 for Het3; n=14 for Het4. D) Frequency of loop-out and snap-back-consistent SD-MMEJ outcomes across different DR-white insertions. n=59 for Eu; n=211 for Het1; n=61 for Het2. n=52 for Het3; n=9 for Het4; *p<0.05 by two-proportion z-test for the comparison between loop-out-consistent events of indicated heterochromatic DR-white insertions relative to the euchromatic one. All other comparisons are not significant.

Distribution of deletion boundaries.

Deletion boundaries associated with different alt-EJ pathways for Eu (A), Het1 (B), Het2 (C), Het3 (D), Het4 (E) DR-white insertions are shown. Deletion boundaries were defined as the first base proximal and first base distal to the deleted bases, with ambiguous bases aligned to the left of the cut site. + indicates all deletion boundaries to the left or right of the sequence shown. *p<0.05; **p<0.01; ***p<0.0005; ****p<0.0001 by two-proportion z-test for comparisons with the euchromatic DR-white insertion. n=86 for Eu; n=280 for Het1; n=74 for Het2; n=104 for Het3; n=14 for Het4 total repair events. Red text highlights the TTAT/AATA overhangs produced by I-SceI cutting.

Spread of deletions for Mutagenic EJ repair.

Heat map plots show the spread of deletions for SD-MMEJ-consistent (A) and MMEJ-consistent (B) repair events across the different DR-white insertions. The frequency of deletion occurring at each bp was calculated as the number of events with that bp deleted divided by the total number of events. Red text highlights the TTAT/AATA overhangs produced by I-SceI cutting. *p<0.05; **p<0.005; ***p<0.001 by two-proportion z-test at each bp for comparisons with the euchromatic DR-white insertion. (A) n=58 for Eu; n=198 for Het1; n=58 for Het2; n=52 for Het3; n=9 for Het4 total SD-MMEJ-consistent repair events. (B) n=19 for Eu; n=56 for Het1; n=7 for Het2; n=48 for Het3; n=5 for Het4 total MMEJ-consistent repair events.

Position of repeat motifs for SD-MMEJ.

Heat map plots show the frequency of usage of different repeat motifs 2 (A) and 1 (B) in SD-MMEJ repair resulting in deletions, for each DR-white insertion. Arrows highlight motifs typically used more frequently (red arrows) or less frequently (black arrows) in heterochromatic DR-white insertions, relative to the euchromatic one. n=58 for Eu; n=198 for Het1; n=58 Het2; n=52 for Het3; n=9 for Het4. Red text highlights the TTAT/AATA overhangs produced by I-SceI cutting. Underlined text highlights some of the most frequently used repeat motifs. (C) Schematic representation of the distance of P2/MH2 (left) and repeat motif 1 (right) from the cut site for the most common repair events in euchromatin and heterochromatin, sorted by frequency. The most common repair products are shown at the top of the list.

Model for HR and alt-EJ repair in heterochromatin.

In heterochromatin (A) HR repair is promoted, and polθ binding close to the cut site facilitates the use of proximal primers for HR repair, resulting in smaller deletions. In euchromatin (B) polθ binding further from the cut site and more extensive probing for homology allows larger deletions to occur.

Mutagenic repair frequency.

Frequency of A) SSA, or B) mutagenic-EJ and non-mutagenic repair/uncut, in white-eyed F2 flies is shown. ***p<0.001 by two proportion z-test. n.s.: not significant. Mutagenic repair results in a PCR product across the cut site that cannot be digested by I-SceI. Non-mutagenic repair results in a PCR product across the cut site that can be digested by I-SceI. In A) n=1352 for Eu; n=1267 for Het1; n=880 for Het2; n=679 for Het3; n=485 for Het4. Error bars: SEM. In B) n=570 for Eu; n=193 for Het1; n=166 for Het2; n=312 for Het3; n=85 for Het4.

Average size of deletions.

Violin plots show the distribution of deletion sizes across all DR-white insertions. Green line: mean. Red line: median. Dashed black lines: quartiles. **p<0.01; ***p=0.0003 by one-tailed Mann-Whitney test. n=86 for Eu; n=280 for Het1; n=72 for Het2; n=104 for Het3; n=14 for Het4.

Common SD-MMEJ repair events in euchromatin and heterochromatin.

The sequences of the most common SD-MMEJ repair events shown in Fig. 6C, have been sorted from top to bottom by frequency of occurrence. Red: microhomologies. Blue: primers. The minimum primer sequence is shown for all repair outcomes.

Common SD-MMEJ intermediates in euchromatin and heterochromatin.

Examples of snap-back SD-MMEJ repair intermediates (1-7) for two of the most frequent repair events detected in euchromatic (A) and heterochromatic (B) DR-white insertions, highlighting: the deleted tract (violet square); the cut site (black line); the position of primers (P1, P2) and microhomologies (MH1, MH2) as defined in Figure 3A; templated DNA syntheses (red arrows) and flap removals (black arrowheads).

miRNA-seq and RNA-seq enrichments at euchromatic and heterochromatic insertion sites.

Average reads per kilobase per million reads (RPKM) values of A) RNA-seq and B) miRNA-seq data from Drosophila testes at +/- 10Kb from DR-white insertion sites [63]. How and mir-9a are spermatogonia-enriched RNAs, used as a reference [100, 101]. ***p<0.005, ****p<0.0001 by two-tailed two-sample z-test for means. Error bars: SD of 3 replicates.

Common SD-MMEJ and MMEJ repair outcomes.

The most frequent deletions across all DR-white insertions are shown (B,C), relative to the original sequence around the I-SceI site of the SceI.white (A). The red rectangle highlights the region containing overhangs from I-SceI cut. B) Three examples (1-3) of SD-MMEJ-consistent repair outcomes are shown, each followed by the possible deletions and repeat motifs (red, underlined) associated with it. Brackets mark bases that could align to the left or to the right of the break, and cannot be unequivocally assigned to one position. In the analysis, all possible outcomes are assigned a weight assuming equal probability of occurrence. The respective SD-MMEJ repair mechanism (snap-back or loop-out) is also indicated. C) The microhomology used for MMEJ repair is shown in brackets.