The Homologous Chromosome Is an Effective Template for the Repair of Mitotic DNA Double-Strand Breaks in Drosophila

Corresponding author: Yikang S. Rong, National Cancer Institute, National Institutes of Health, Bldg. 37, Room 6056, 37 Convent Dr., Bethesda, MD 20892., rongy{at}mail.nih.gov (E-mail)

Communicating editor: G. SMITH

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

In recombinational DNA double-strand break repair a homologous template for gene conversion may be located at several different genomic positions: on the homologous chromosome in diploid organisms, on the sister chromatid after DNA replication, or at an ectopic position. The use of the homologous chromosome in mitotic gene conversion is thought to be limited in the yeast Saccharomyces cerevisiae and mammalian cells. In contrast, by studying the repair of double-strand breaks generated by the I-SceI rare-cutting endonuclease, we find that the homologous chromosome is frequently used in Drosophila melanogaster, which we suggest is attributable to somatic pairing of homologous chromosomes in mitotic cells of Drosophila. We also find that Drosophila mitotic cells of the germ line, like yeast, employ the homologous recombinational repair pathway more often than imperfect nonhomologous end joining.

CHROMOSOMAL double-strand breaks (DSBs) are often encountered in living cells. A wealth of evolutionarily conserved mechanisms are used to repair them. These mechanisms can be grossly categorized into two pathways: nonhomologous end joining (NHEJ) and homologous recombinational (HR) repair (for reviews see JEGGO 1998 ; PAQUES and HABER 1999 ). In NHEJ, the broken ends are ligated, requiring little or no homology between the ends. NHEJ is intrinsically mutagenic in that it may lead to loss or addition of sequences at the site of DSB. HR can also lead to net loss of DNA (nonconservative HR repair). One such mechanism is single-strand annealing (SSA). Cells often employ SSA to repair DSBs induced between directly repeated sequences, generating deletions (LIN et al. 1984 ; RUDIN et al. 1989 ; MARYON and CARROLL 1991 ; RONG et al. 2002 ; PRESTON et al. 2002 ). Alternatively, the conservative mode of HR is gene conversion (GC) that does not lead to net loss of sequences. In GC, a DNA template homologous to the DSB region is used in DNA synthesis. Such a template may be located on the sister chromatid in G₂ (sister), on the homologous chromosome (homolog), or at an ectopic genomic position.

The template that is chosen for DSB repair by GC can have significant consequences for genome integrity. GC may homogenize related sequences and, if accompanied by exchange, can lead to loss of heterozygosity or genome rearrangements. For organisms whose genomes abound with repetitive sequences, preventing the use of an ectopic template for DSB repair may be an important strategy to safeguard the genome. Yeast mating type switching is an example of the use of an ectopic template, but it is a highly regulated process with a specific biological purpose—not a general method of repair. DSB repair in meiosis occurs between homologs, but also is regulated for a specific purpose: use of the sister chromatid is minimized so that exchange between homologs helps to ensure their proper segregation in meiosis I. In mitotic cells of yeast and mammals, however, there is mounting evidence that the homologous chromosome is not preferentially used in the repair of DSBs.

KADYK and HARTWELL 1992 presented the only report in which the frequencies of repair using the sister vs. the homolog were directly compared in irradiated mitotic cells of Saccharomyces cerevisiae. They discovered that while G₁ DSBs were repaired by interhomolog GC, close to 100% of the DSBs in G₂ were repaired by GC with the sister. In mammalian cells, studies using the I-SceI endonuclease have led to the conclusion that the sister is also highly preferred (LIANG et al. 1998 ; JOHNSON and JASIN 2000 ). Moreover, allelic recombination was similar in efficiency to ectopic repair using a template on a heterologous chromosome (MOYNAHAN and JASIN 1997 ; DONOHO et al. 1998 ; RICHARDSON et al. 1998 ). Therefore, the homologous chromosome is rarely used for HR repair in mammals. To compensate for inefficient allelic HR repair, mammalian cells are believed to employ NHEJ as the prevailing repair pathway in G₁ cells (LEE et al. 1997 ; TAKATA et al. 1998 ), and yeast cells may also have the highest NHEJ activity in G₁ (KARATHANASIS and WILSON 2002 ).

These studies raise the question of why the sister chromatid is preferred over the homolog. KADYK and HARTWELL 1992 ruled out the possibility that the sister is a better recombinational substrate by virtue of sisters having identical DNA sequences. They proposed that it might be due to the fact that the sisters were in close proximity after replication, making one readily available as a repair template for the other. This was supported by the observation that yeast cells defective in sister chromatid cohesion also had repair defects (SJOGREN and NASMYTH 2001 ). The Kadyk and Hartwell model predicts that any mechanism that allows close association (pairing) of homologous chromosomes would also increase the efficiency of DSB repair using the homolog. We set out to test this prediction by studying allelic GC in Drosophila melanogaster, one of the Dipteran species in which extensive mitotic chromosome pairing has been observed (STEVEN 1908 ; METZ 1916 ; LEWIS 1954 ; HIRAOKA et al. 1993 ; GOLIC and GOLIC 1996 ; FUNG et al. 1998 ).

To induce DSBs in Drosophila we used the I-SceI site-specific endonuclease, which generates a DSB at its 18-bp recognition site (COLLEAUX et al. 1988 ). I-SceI has been widely used in DSB repair studies in a variety of organisms including yeast, plants, mammalian cells, and Drosophila (PLESSIS et al. 1992 ; PUCHTA et al. 1993 ; ROUET et al. 1994 ; BELLAICHE et al. 1999 ). We previously showed that when I-SceI was expressed in flies, it efficiently cut at its target site that had been embedded in a chromosome (RONG and GOLIC 2000 ). In the work presented here we used the precise and efficient induction of a DSB within a chromosome to study how flies repair such a break and, in cases where HR was used, to assess which template was chosen. We find that HR is the predominant repair pathway in the Drosophila premeiotic germline and that, in contrast to mammalian cells, Drosophila frequently use the homologous chromosome as a repair template.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Genetics:
The description of mutations not given here can be found in FlyBase (http://www.flybase.bio.indiana.edu). The 70I-SceI (RONG and GOLIC 2000 ) insertions used in this study were 1A on chromosome III and 2B on chromosome II.

Germline assays were carried out by testcrossing males that carried a reporter construct and 70I-SceI individually to white (w)-null females. The crosses that generated these males are described below. They were heat-shocked early in development (0–3 days after egg laying) for 1 hr at 38°. The progeny of the testcrosses were scored. A portion of the recovered chromosomes was further characterized to identify the molecular nature of the DSB repair events as described in RESULTS.

For experiments shown in Fig 1, flies that carried an Iw insertion on the X chromosome or on a dominantly marked autosome were mated to flies with 70I-SceI on chromosome II or III. For experiments shown in Fig 2, females with wIw insertion 4A on II were mated to 70I-SceI insertion 1A on III, and females with wIw insertion 2 on III were mated to 70I-SceI 2B on II. For experiments in Fig 5, males with Iw insertion 7-13b and 70I-SceI 1A were mated to females with Iw insertion 7. For experiments in Fig 6, males with 70-I-SceI 2B and wIw insertion 8z were mated to wIw insertion 2.

View larger version (24K): In this window In a new window Download PPT slide   Figure 1. Repair of DSBs generated in the context of the Iw construct. The solid arrow represents the w+ gene. The solid box represents the 18-bp I-SceI cut site (not to scale). The small arrowheads represent P-element inverted repeats 5 kb apart. The positions and directions for three primers are shown: 1, PE5'; 2, break3'; and 3, w7926u. The repair of the DSB can give rise to three different classes of progeny. Possible repair mechanisms are given for each class of event. The progeny were first divided on the basis of whether they retained a functional w+ gene. The mechanism for w+ loss was not further distinguished, and only the NHEJ model is illustrated here. For this classification, six different transformant lines were used; chromosomal location is in parentheses. The w+ progeny were further divided into two classes on the basis of whether they retained an intact I-SceI cut site. The open box represents a mutated cut site. Three lines were tested for the recut +/- classification. The overall percentage was calculated by combining data from the two sets of experiments. N, number of progeny scored that have inherited the reporter chromosome.

View larger version (27K): In this window In a new window Download PPT slide   Figure 2. Repair of DSBs generated in the context of the wIw construct. The larger arrow at the right side of the I-SceI cut site represents the w+ gene. The smaller arrow at the left represents the 3' portion of w. The open portion in the large arrow represents the 5' part of w+ that is not present in the smaller arrow on the left of the cut site. For a more detailed structure of this construct, see Fig 4. The positions and directions for three primers are shown: 2, break3'; 3, w7926u; and 4, w14178d. Line 4A on chromosome II and line 2 on III were used for the heat-shock experiment (+HS). Only line 2 was used for the no heat-shock experiment (-HS). The numbers in parentheses (N) are the numbers of progeny that inherited the reporter chromosome. The progeny were first divided on the basis of whether they retained a functional w+ gene. The w+ progeny were further scored for recut phenotypes according to the text description. The open box represents a mutated I-SceI cut site.

View larger version (90K): In this window In a new window Download PPT slide   Figure 3. Somatic cutting by I-SceI. The 9-kb fragment from the pug locus is shown at the top along with the I-SceI cut site position. It is flanked by NotI (N) sites that are not present in the endogenous locus (86C). The probe was the entire 9-kb fragment. The expected sizes of fragments from I-SceI, NotI double digest are indicated in kilobases. Lane m, molecular markers in kilobases; lane 0, DNA from flies taken at 0 hr after heat-shocking; lane 4, 4 hr after; lane 8, 8 hr after; lane 16, 16 hr after; lane c, control DNA from flies with the 9-kb fragment but not the 70I-SceI gene.

View larger version (62K): In this window In a new window Download PPT slide   Figure 4. Southern blot analysis of possible SSA events. The upper diagram represents a chromosome II insertion of the wIw construct (Fig 2). Hatched boxes represent the white gene with the first intron specifically indicated in which an FLP recombination target (FRT) resides. The lower diagram represents an SSA event. The sizes for EcoRV (R5) fragments are shown in kilobases. The probe is also indicated, which hybridizes to both copies of the 3' w duplication. Lane m, markers with sizes shown in kilobases; lane 1, control DNA from w- flies without the reporter construct; lane 2, DNA from flies carrying the original insertion; lanes 3 and 4, two independent SSA events. The 2-kb band is from the endogenous w gene.

View larger version (29K): In this window In a new window Download PPT slide   Figure 5. Repair in Iw homozygotes. (A) The solid box represents a wild-type I-SceI cut site in line 7. The shaded box represents the mutated cut site in line 7-13b derived from line 7 on a Sco-marked chromosome II (for the sequence of the mutated cut site, see Table 1). The DSB on the Sco+ chromosome can be repaired to give rise to the same three classes of progeny shown in Fig 1. In addition, a fourth class is allelic GC events in which the I-SceI cut site was converted to the exact sequence in line 7-13b. An allelic event would be positive for a PCR-based test (see main text). The numbers are percentages. A detailed presentation of the data is given in Table 2. HS, heat-shock treatment. (B) Examples of the allelic PCR assay. DNA from 18 repair events was PCR amplified using the primers w7926u (for location see Fig 1) and 13b-minus, which annealed only to the mutated cut site (shaded) on the Sco chromosome. Lane m, marker with size in base pairs shown; lane c, control DNA from the NHEJ line Iw 7-13b. All lanes except 1 and 4 represent allelic GC events.

View larger version (16K): In this window In a new window Download PPT slide   Figure 6. Repair in wIw homozygotes. The solid box is the wild-type I-SceI cut site. The shaded box represents the mutated cut site in line 8z (Table 1). Genetic markers along the chromosome are also given. The DSB can be repaired to give rise to the same three classes of progeny shown in Fig 2. In addition, a fourth class is allelic GC events in which the I-SceI cut site was converted to the exact sequence as in 8z. The numbers are percentages. A detailed presentation of the data is given in Table 2.

Molecular analyses:
For experiments in Fig 3, flies with 70I-SceI 2B and a chromosome III insertion of the 9-kb fragment from the pug locus (RONG et al. 2002 ) were heat-shocked as adults for 1 hr at 37°. DNA from 15 flies was taken as described (RONG and GOLIC 1998 ) at different time points and digested with NotI prior to Southern blot analysis.

The PCR primers w7926u (5'-atagcgagcacagctaccag) and PE5' (5'-gatagccgaagcttaccgaagt) were used to amplify NHEJ junctions of DSB repair in the context of the Iw construct. A 1.4-kb fragment was gel-purified and sequenced using the primer break3' (5'-cgcgatgtgttcactttgct). For the wIw construct, the primers w7926u and w14178d (5'-tgtgtgtttggccgaagtat) were used. A 1.1-kb fragment was gel-purified and sequenced with break3'. The relative positions of the primers are shown in Fig 1 and Fig 2.

To score interhomolog GC events using Iw, the primers w7926u and 13b-minus (5'-cggtacattaccctgttatgcggtt) were used in PCR. For wIw, the primers w7926u and 8z-minus (5'-ggcgggtacattaccctgttatctgttata) were used. The stringency of these allele-specific PCRs was tested on all imperfect NHEJ events that had been sequenced (Table 1). DNA templates for PCR were prepared from single flies as described (GLOOR et al. 1993 ). To minimize contamination, only male and virgin female flies were used in the PCR assay. The reactions were performed with a 60° annealing temperature for higher specificity. A positive reaction with a 0.9-kb product indicated allelic GC; a negative reaction indicated an imperfect NHEJ event. On templates yielding negative results for the first PCR test, a second PCR was performed with a pair of control primers outside of the region of interest. All were able to generate a product (data not shown).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

To examine the repair of a chromosomal DSB we used the two transgenic constructs diagrammed in Fig 1 and Fig 2. In the first construct, called Iw (I-SceI-white), the 18-bp I-SceI recognition site was placed 30 bp upstream of the white⁺ (w⁺) reporter gene. In the second construct, called wIw (white-I-SceI-white), the cut site was at a similar position, and a 3' nonfunctional portion of w was duplicated, in the same orientation, on the opposite side of the I-SceI cut site. In a w-null background, both transgenic constructs produce flies with pigmented eyes. We generated flies that carried one of these constructs and the heat-inducible 70I-SceI transgene (RONG and GOLIC 2000 ) and heat-shocked them early in their development to generate a DSB at the I-SceI cut site. For the somatic assays, these heat-shocked flies were examined directly; for the germline assays, they were mated individually to recover chromosomes that had been exposed to I-SceI. Since no male meiotic division occurred at the stages in which the animals were heat-shocked (LINDSLEY and TOKUYASU 1980 ), we are assaying DSB repair in mitotic cells of the germ line.

Efficient repair of DSBs:
If a break were left unrepaired, a part of the broken chromosome would be lost in subsequent cell division. Cells with such deficient chromosomes or progeny derived from such cells are unlikely to be viable. If failure to repair an I-SceI -induced DSB in the germline were a significant outcome, we would expect sterility or preferential recovery of the unbroken homolog in the progeny. This was not apparent in any of the experiments discussed below (data not shown). Additionally, examination of adults revealed no apparent deleterious effects from the high frequency of DSBs induced by I-SceI (>90% in some cases: see experiments to follow). These observations suggested that the DSBs generated by I-SceI were efficiently repaired.

We also directly assessed the presence of DSBs in genomic DNA by Southern blotting. We generated flies that carried a P element with 9 kb of DNA from the pugilist (pug) locus with an I-SceI cut site placed in the middle (RONG et al. 2002 ). The pug DNA was flanked by NotI sites that were not present at the endogenous pug locus (Fig 3). Adults that carried this construct and 70I-SceI were heat-shocked, and genomic DNA was prepared at different time points after the heat-shock treatment and digested with NotI for Southern blot analysis. We observed I-SceI cutting up to 8 hr after heat-shock induction, but at any time point only a small fraction of the DNA exhibited a DSB at the I-SceI cut site (Fig 3 and data not shown), suggesting that the repair process was rapid. Since I-SceI can induce very efficient cutting (see below), we conclude that most DSBs were successfully and rapidly repaired. Similar results were obtained in a previous report (GONG and GOLIC 2003 ).

Repair pathways:
To more closely examine how the I-SceI -generated DSBs were repaired we recovered the chromosomes carrying a reporter construct that had been exposed to I-SceI.

We first examined the fate of hemizygous insertions of Iw. Initially the progeny were categorized with respect to whether or not they had pigmented eyes (Fig 1). The white-eyed progeny may result from extensive exonucleolytic degradation of the broken ends that has removed at least a part of the w⁺ gene, followed by NHEJ. Alternatively, in the case of autosomal insertions, GC with the homolog that lacks the insertion could remove the entire P element, including w⁺. A PCR test was able to detect sequences from the starting Iw construct in some of the white-eyed progeny, indicating that NHEJ could account for at least some of the w⁺-loss events (data not shown). Since w⁺-loss events were such a small fraction of the total (2.3%), we did not estimate the relative contribution from these two modes of repair.

The w⁺ progeny were further characterized to determine whether the I-SceI cut site was intact. An easy test for an intact cut site can be carried out by crossing the w⁺ individuals to flies that carry the 70I-SceI transgene. Their progeny are heat-shocked during early development, and when the adults eclose they are examined for somatic white mosaicism in the eye. Since I-SceI cutting causes occasional loss of w sequences, such mosaicism indicates that the cut site is intact. When an Iw chromosome that had not been exposed to I-SceI cutting was tested in this fashion 40% of adult eyes exhibit mosaicism. In the current test, the presence of a similarly large fraction of progeny with mosaicism indicated that the I-SceI cut site was not altered, and we classified such a chromosome as recut⁺. Recut⁺ chromosomes could be produced by three mechanisms: the I-SceI cut site was unchanged because it escaped cutting by I-SceI; the site was cut and then repaired by GC with the intact sister chromatid; or it was cut and the ends were rejoined without sequence alteration. The w⁺ flies that did not produce mosaic offspring in this somatic test must have an altered or missing I-SceI cut site and were classified as recut^-. These progeny could result from a variety of imperfect NHEJ events that deleted or inserted a small number of bases at the cut site.

Six different Iw insertions were tested in the first set of experiments. A total of 138 heat-shocked male parents were outcrossed individually to measure loss of w⁺. After exposure to I-SceI, only 2.3% of the chromosomes lost w⁺ (Fig 1). We then tested 553 of the w⁺ chromosomes in males, derived from 34 heat-shocked 70I-SceI-bearing fathers, for alteration of the I-SceI cut site. We classified 129 as recut^-, with over 50 progeny scored for each male. For some of these recut^- chromosomes, the region surrounding the cut site was sequenced to reveal alterations (see below). Overall, 23.2% of the F₁ progeny retained w⁺, but were recut^-, and we interpreted that they harbored an imperfect NHEJ event. The remaining 74.5% retained the original P element.

Results presented below show that at least 75% of the chromosomes carrying this construct experienced a DSB under our experimental conditions (Fig 5). We conclude that inaccurate joining of DSB ends occurred no more than one-third of the time. But this experiment cannot distinguish whether this was because there was efficient end joining of unaltered ends, or whether the intact sister chromatid was frequently used as a template for GC, restoring the original I-SceI cut site. Therefore, we could not compare the relative contributions of imperfect NHEJ vs. HR to DSB repair.

This comparison was possible when we used the second construct (wIw) to investigate repair by SSA of a DSB generated between directly repeated sequences. Chromosomes carrying this construct were recovered from males that were hemizygous for a wIw insertion, after induction of I-SceI synthesis under the same conditions used for the experiments of Fig 1. Two different lines were used in which a total of 69 male parents were individually crossed. After heat shock only 15% of the chromosomes with wIw retained w⁺ (Fig 2). These w⁺ progeny were further divided using the somatic mosaic test described previously to assay the integrity of the I-SceI cut site. One-third of the w⁺ progeny showed evidence of NHEJ because they had lost the cut site. The remaining w⁺ recut⁺ chromosomes may have resulted from GC using an intact sister, from accurate religation of the cut ends, or from a failure of cutting. We conclude that I-SceI cuts in at least 90% of the germ cells at the time of heat-shock induction.

The w^- flies arose at a much higher frequency than in the previous experiment (Fig 1) suggesting that they were the result of a different mode of DSB repair. The result of an SSA event would be the deletion of all sequences flanked by the w repeats as well as one unit of the repeat. This would give rise to a nonfunctional w gene (Fig 2). To determine whether the w progeny had the predicted repair product, lines were established from 40 independent repair events. Genomic DNA was digested with EcoRV and probed with the repeated 3' portion of w in Southern blot analyses. All 40 events produced the structure that was predicted by SSA repair (Fig 4 and data not shown). Lane 2 in Fig 4 shows insertion 4A on chromosome II with three characteristic bands of 4.1, 4.3, and 5.0 kb. Lanes 3 and 4 are two w⁺-loss lines showing only the outside 4.1- and 4.3-kb bands that indicate deletion of all repeat-flanked sequences and one of the repeats. Therefore, all or nearly all of the w^- progeny in this experiment were the result of recombinational repair. It is clear that, given the opportunity, SSA can be a highly preferred pathway of repair. The proportion of imperfect NHEJ events among w⁺ progeny was similar to that observed in the previous experiment (approximately one-third), suggesting that breaks not subject to SSA repair were treated similarly in the two situations.

Imperfect DSB repair by NHEJ:
Thirteen independent w⁺ recut^- chromosomes from the first experiment and 11 from the second were selected, and the region surrounding the I-SceI cut site was PCR amplified and sequenced to reveal the end-joining junction (Table 1). Three kinds of sequence alteration were found at the junction: deletion, insertion, or a combination of both. All are typical of NHEJ repair (DUCAU et al. 2000 ; BIBIKOVA et al. 2002 ; ADAMS et al. 2003 ). Simple deletion products were the most abundant (18/24) with sizes ranging from 1 to 22 bp. In the three simple insertions, the inserted sequences apparently originated within a few base pairs of the DSB. In the cases of 5z and 8z, short stretches of sequence were duplicated more than once. For example, the sequence ATAA was duplicated three times in 8z. This may indicate repeated attempts at repair before final success. The remaining three cases were compound rearrangements that involved both deletion and insertion of sequences.

The I-SceI enzyme leaves a 3' overhang of ATAA (COLLEAUX et al. 1988 ). In 16 of the 24 cases of NHEJ that we sequenced the ATAA was preserved, indicating that at least one end was not subject to exonuclease degradation of both strands. In a previous study in vitro, I-SceI protection of a specific end was observed (PERRIN et al. 1993 ). The signature of such preferential protection would be for deletions to extend in only one direction from the cut site. In cases where the ATAA was retained, we did not observe a significant preference for deletions extending in one direction. If I-SceI does protect an end by continued association after cutting, it does not appear to prefer a particular end in vivo.

Recombinational repair using the homologous chromosome:
In the above experiments, the P-element insertion was hemizygous so that the sequence homology between homologs lies outside of the P element. For the DSB to be repaired by allelic recombination, the entire P element would need to be removed by exonucleolytic digestion. This would represent degradation of 5 kb for the Iw insertions and 9 kb for wIw. Since we observed only 2% w⁺ loss with the Iw construct, repair using the homolog must have been very rare. This is also likely to be the case for the wIw construct since the w⁺-loss events were the result of intrachromosomal recombination.

It seems reasonable that the homolog is more likely to be used as a repair template if it is homologous to the reporter chromosome in the immediate vicinity of the DSB. To obtain a more realistic estimation of the participation of the homolog in normal DSB repair, we generated flies with near-identical P elements at allelic positions (Fig 5 and Fig 6). One element was the reporter P element with an intact I-SceI cut site. The homolog carried the same P element insertion except that the cut site was mutated by the deletion or insertion of a few bases (produced by NHEJ in the first sets of experiments). Since we know the sequence of the mutant cut site, we can distinguish interchromosomal GC events from further imperfect NHEJ events by allele-specific PCR. Because we are studying repair in the male germline, which lacks meiotic recombination, and since mitotic GC events were seldom associated with crossing over (see below), we can correctly identify the recipient chromosome of a conversion event by using linked chromosomal markers.

For the Iw construct, line 7 on chromosome II was chosen. The mutant derivative 7-13b (Table 1) was placed on the homolog marked by the dominant Scutoid (Sco) mutation. Males were produced that carried both elements and 70I-SceI. They were heat-shocked and testcrossed. The results are presented in Table 2 and summarized in Fig 5. The Sco⁺ progeny were scored for eye color, yielding a w⁺-loss frequency of 0.5%. The w⁺ progeny consist of two classes that are distinguished by whether they have an intact or a mutated I-SceI cut site. We estimated the proportion of each by testing some of the chromosomes for the recut phenotypes as described previously. The cut site was intact in only 24% of the progeny. These results differ from the previous experiment with a hemizygous insertion, where we observed 2.3% w⁺ loss and the I-SceI cut site was found intact in 75% of progeny. Therefore, providing matching homology directly opposite the DSB has a strong influence on repair.

The fraction of w⁺ recut^- offspring that had an imperfect NHEJ event and those that resulted from allelic GC was determined by PCR using a pair of primers, one that was designed to specifically anneal to the mutated I-SceI cut site in 7-13b and another outside primer from w. Examples of this allelic specific PCR test are shown in Fig 5B. We observed a high frequency of interhomolog repair events, accounting for 65% of the total repair events. This suggests that allelic GC is a very efficient process in Drosophila. Surprisingly, this class even outnumbers the w⁺ recut⁺ progeny, some of which could result from GC templated from the sister chromatid. Since I-SceI can cut with an efficiency of 90% or better, a possible explanation is that in this experiment both of the sister chromatids were cut frequently, discouraging intersister events. This may lead to a higher frequency of allelic events.

To better determine whether the homolog could be frequently used in the presence of an intact sister chromatid, we took advantage of the low level of noninduced activity of the 70I-SceI gene (data not shown). We repeated the experiment without heat-shocking the male parents, thus providing a lower frequency of cutting. In this case, a DSB may have an uncut sister chromatid to use for GC repair. For results with or without heat shock (+HS or -HS), we compare the frequency of allelic GC to that of imperfect NHEJ repair, which is not a major repair pathway in the Drosophila germline. Even with low levels of cutting, allelic events occurred at a significant rate (9.1%), and were still manyfold higher than the products of imperfect NHEJ. This is very different from the situations in mammalian cells in which imperfect NHEJ outcompetes allelic GC (STARK and JASIN 2003 ).

In the initial experiments with a hemizygous insertion of Iw, a minority of chromosomes showed evidence of cutting after HS: 75% of the progeny that received the reporter chromosome had an unaltered Iw element (Fig 1). However, in this second experiment, with copies of the insertion on both chromosomes, just <25% of the chromosomes appeared unaltered. Most of the repair events in this experiment were detectable only because the homolog was used as a repair template and it had a slightly altered sequence at the site of the DSB. We conclude that efficient use of the homolog requires sequence homology in the immediate region of the break. When this homology is absent, the channeling of a DSB into different repair pathways (NHEJ vs. HR) may be altered so that precise end joining is a more efficient process, or the choice of template (homolog vs. sister) may be altered so that an uncut sister chromatid is more frequently used.

We wish to compare allelic GC to another mode of HR repair. The SSA type of intrachromosomal repair is very efficient and readily assayed. So, we next asked whether the homolog would be used as a repair template in competition with the highly efficient SSA pathway. We used the wIw construct on III and the NHEJ line 8z (Table 2) as a template on a Stubble-marked (Sb) homolog. The experiment was conducted similarly to the one that used the Iw construct, with one difference: the assay for recut phenotypes did not involve a second cross with 70I-SceI flies. Half of the relevant progeny also inherited the 70I-SceI gene on a Sco-marked chromosome II. This particular 70I-SceI insertion had a high constitutive activity so that flies carrying this gene and a wIw insertion showed somatic eye mosaicism in 100% of the eyes even without heat-shock treatment (data not shown). We were then able to score recut phenotypes directly in the F₁ progeny. To test for allelic GC, a PCR primer that is specific to the mutant I-SceI cut site in 8z was used. To confirm the specificity of the PCR assay, we sequenced the region surrounding the I-SceI cut site for nine PCR-positive samples and one PCR-negative sample. In the nine positive samples, sequencing showed an exact match to the cut site in 8z. The negative sample had a mutant cut site with a different sequence (data not shown).

The results are summarized in Fig 6, while the data are presented in Table 2. The frequency of repair using the homolog was slightly higher than that of repair by SSA, indicating that the homolog is preferred even in competition with the highly efficient SSA pathway. Again in this experiment it is possible that a high frequency of cutting greatly reduces the use of the sister chromatid as a template, thus driving repair into SSA or allelic conversion pathways, at the expense of what would normally be a more efficient sister-templated conversion. We repeated the experiment with no heat-shock induction and found more than twice as many repair events that used the homolog as there were intrachromosomal SSA events. As before, we conclude that the homolog was efficiently utilized. However, since we could not estimate the frequency at which both sister chromatids were cut by I-SceI under these conditions, we cannot provide an accurate assessment on how effectively allelic GC can compete with intersister GC.

When the hemizygous vs. the homozygous experiments are compared (Fig 2 vs. Fig 6), it seems apparent that some of the perfect end-joining or sister-templated conversion events may have been redirected to the allelic GC pathway when there is matching homology on the homolog. With a hemizygous insertion and no heat-shock treatment, we observed 13.2% repair events that altered the sequence of the element (Fig 2). With homozygous insertions, there were twice as many such repair events. Since the same 70I-SceI insertion was used in both experiments to produce the same level of noninduced cutting, it seems that some repair events that were "invisible" before could be observed in the second experiment only because they now used the homolog in preference. We conclude that Drosophila use the homologous chromosome very effectively as a template for DSB repair, at least when an intact sister chromatid is not available for repair.

Rare GC-associated crossing over:
It has been shown in yeast, in Drosophila, and in mammalian cells that mitotic GC events are seldom associated with crossing over (reviewed in PAQUES and HABER 1999 ). We tested whether this was also true for repair of an I-SceI-generated DSB in Drosophila. The homozygous males used in the experiment of Fig 6 were also heterozygous for the markers roughoid (ru), hairy (h), and ebony (e), with the P-element insertion lying between h and e. When the male parents were heat-shocked and subsequently testcrossed, we recovered 20 recombinants, all between h and e, from 6282 progeny (Table 2), yielding a frequency of 0.3% for crossover events. If we assume that these crossover events were associated with GC at the P-element insertion site, then a maximum of 0.7% (0.003/0.457) of the allelic GC events were associated with a crossover. In the no heat-shock experiment recombinants were recovered at a slightly higher rate of 1.2% (64/5384), allowing for up to 7% (0.2/0178) of the allelic GC events to be associated with crossovers. In both cases, a very minor fraction of GC events are associated with exchange, confirming the results of an earlier study (ENGELS et al. 1990 ).

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

In this study, we used I-SceI to induce site-specific DSBs in Drosophila cells and found that multiple DSB repair mechanisms were employed in the male germline. By recovering repair events in individual offspring, we were able to estimate the contribution of different DSB repair pathways in the germline. An important advantage of using a site-specific DSB system is that the immediate genomic context of the DSB can be controlled so that specific modes of repair can be studied. By using these methods we found that allelic recombination is a prominent repair pathway in Drosophila and that it competes very effectively with intrachromosomal recombination.

On the other hand, we could not control repeated I-SceI cutting at its recognition site. DSB studies using site-specific endonucleases may not provide an accurate estimate for the frequency of repair events that restore the cut site (LIN et al. 1999 ). For example, perfect ligation of the broken ends, as well as GC with an intact sister chromatid, would both give rise to DNA targets that can be cut again. Perfect rejoining is a significant DSB repair activity in both yeast and mammalian cells (LEE et al. 1999 ; LIN et al. 1999 ; KARATHANASIS and WILSON 2002 ). Moreover, intersister GC is believed to be one of the most prominent HR repair mechanism. In this study, I-SceI might persist in Drosophila cells for some time after induction so that repeated cutting occurs. In mammalian DSB studies, evidence of I-SceI stability has been observed, with cutting evident in cell cycles subsequent to I-SceI induction (JOHNSON and JASIN 2000 ). Therefore, the frequency of repair events that eliminate the cut site is likely an overestimate. We tried to minimize the problem by comparing results from experiments in which I-SceI was induced to different extents. While one needs to be mindful of the possibility that a DSB may be repaired to generate a site that can be cut again, important conclusions can still be drawn about the biology of DSB repair from studies that use site-specific endonucleases.

Choice dependence of repair pathways on genomic context:
We found that the outcome of DSB repair depended on the immediate genomic environment of the DSB. When a DSB was generated in the Iw construct (Fig 1), nearly one-quarter of the progeny inherited a repaired chromosome in which the few nucleotides around the I-SceI cut site were altered due to imperfect NHEJ. Similar repair products were recovered in only 5% of the progeny when the wIw construct was used (Fig 2). In addition, repair led to the loss of w⁺ in only 2.3% of the progeny carrying Iw, but in 85% of the progeny carrying wIw. These different outcomes in the repair of identical DSBs are certainly attributable to the differing chromosomal contexts of each break. The DSB was flanked by a duplication of 3 kb in wIw but not in Iw. The duplication promoted a repair pathway, probably SSA, leading to deletions.

The choice between repair pathways was also strongly influenced by whether or not matching sequences were present on the homologous chromosome immediately opposite the site of the DSB. When the Iw construct was present as a hemizygous insertion, 75% of chromosomes showed no evidence of having been cut, with 25% repaired by imperfect NHEJ. Moreover, the frequency of repair by homolog-templated GC could not have been >2.3% since such an event would delete the entire P element and be manifested as white-eyed progeny (Fig 1). In contrast, when another Iw element was placed at the allelic position to provide sequence homology, 75% showed evidence of cutting and repair, with >65% of the repair products being the result of interhomolog GC (Fig 5). Thus, many repair events were redirected to use GC from the homolog when allelic homology was provided.

Efficient use of the homologous chromosome:
The contribution of the homologous chromosome to recombinational repair has been evaluated in several systems. In S. cerevisiae and mouse ES cells, the homologous chromosome is generally not a preferred template for repair when other homologous templates exist. What factors would affect the template choice for recombinational repair? KADYK and HARTWELL 1992 suggested that it was simply the proximity of the DSB to its repair template. The hypothesis predicts that in Diptera in which homologous chromosomes are intimately associated during mitotic phases, allelic GC should occur at a high frequency. In this study, we confirmed this prediction.

We first studied repair from the homolog with the Iw construct. This was intended to mimic the most common genomic environment that a chromosomal DSB would encounter, that is, matching homology is provided by the sister or the homolog. We observed that in 65% of the germ cells, a DSB in Iw was at least eventually repaired by homolog-templated GC (Fig 5). Even when the wIw construct was tested, which provided for efficient intrachromosomal repair by SSA, repair that used allelic homology was more frequent than repair by SSA (Fig 6). This result differs significantly from that of mammalian studies in which repair by intrachromosomal recombination was often not a major event, but was still a few orders of magnitude more efficient than allelic GC even without direct competition between the two modes of HR repair (MOYNAHAN and JASIN 1997 ; DONOHO et al. 1998 ; LIANG et al. 1998 ; RICHARDSON et al. 1998 ; JOHNSON and JASIN 2000 ).

On the other hand, our results from homozygous Iw experiments (Fig 5) did not differ significantly from results obtained in S. cerevisiae. In diploid yeast cells, the homolog was used to repair G₁ breaks (ESPOSITO 1978 ; FABRE 1978 ). Moreover, MALKOVA et al. 1996 showed that a DSB induced by the HO endonuclease was repaired by allelic GC 98% of the time. This efficient allelic GC would be consistent with the results by BURGESS et al. 1999 who showed by using fluorescence in situ hybridization that the homologous chromosomes were associated in vegetative S. cerevisiae cells. However, the homologous chromosome is clearly not a preferred GC template when compared to an ectopic one in S. cerevisiae. LICHTEN and HABER 1989 showed that an ectopic template in cis and 20 kb away was utilized in spontaneous mitotic recombination 6- to 13-fold more efficiently than the homolog. Allelic recombination achieved a frequency similar to ectopic recombination only when the ectopic template was placed at a more distant site in cis or on a heterologous chromosome. This led to the conclusion that in yeast homologous chromosomes interacted with each other no more than with nonhomologous chromosomes. Collectively, these two studies suggest that "somatic pairing" seen in S. cerevisiae does not render the homolog any advantage over ectopic templates. Interestingly, mitotic allelic GC in Schizosaccharomyces pombe can be manyfold more efficient than ectopic GC from a template on a heterologous chromosome (VIRGIN et al. 2001 ). However, templates in cis were not included in that study. In Drosophila, Engels' group showed that ectopic GC induced by P-element transposition could be efficient. In the female germline, allelic GC was generally more efficient than ectopic GC from a cis template (NASSIF and ENGELS 1993 ; ENGELS et al. 1994 ). However, repair templates with sequence mismatches were used in those studies. Therefore, it remains unclear whether Drosophila cells are different from yeast cells in choosing the homologous chromosome vs. an ectopic template.

Previous studies of DSB repair in Drosophila failed to reveal the highly efficient use of the homologous chromosome that we observed. ENGELS et al. 1990 studied the repair of DSBs induced by the P transposase, which causes DSBs at the terminal inverted repeats of a P element. They observed that, in females with a P element inserted in one of their X chromosomes, up to 15% of the germ cells had undergone precise loss of the element in the presence of transposase, presumably as the result of GC templated from the homolog that did not carry the insertion. In two other similar studies, the same group observed only 7% precise loss of an autosomal hemizygous P insertion (PRESTON and ENGELS 1996 ; PRESTON et al. 2002 ). These frequencies for allelic GC are much lower than those we observed in this study for the Iw construct (65%) or the wIw construct (45%), given that the efficiency of inducing DSBs was likely to be similar in experiments using I-SceI and P transposase: both were able to induce SSA in >80% of the germ cells (Fig 2; PRESTON et al. 2002 ). The difference may lie in the structures of homologous chromosomes in the region of the DSB.

We were able to construct a genotype wherein, in the vicinity of the DSB, the uncut chromosome differed by only a few base pairs from the chromosome that was cut by I-SceI. We showed that this matching homology promoted the efficient use of the homolog in GC. On the other hand, an entire P element had to be removed to facilitate GC from the homolog in the prior studies from Engels' group. This could have been accomplished by two means. Transposase might induce a DSB simultaneously at both P-element termini, liberating the entire element. Alternatively, the tranposase might cut at only one terminus, followed by exonucleolytic degradation traveling to the other end of the P element. We propose that the first gap-forming mechanism rarely happens. As shown in vitro, P tranposase behaves as a site-specific endonuclease (BEALL and RIO 1997 ). It cuts efficiently only when both termini are present in the reaction mixture, which suggests that tranposase molecules bound at the P ends interact with each other. However, the experiments were not designed to address how often transposase cuts at both ends simultaneously. In fact, the results showed preferential cutting at the 3' terminus over the 5' end (BEALL and RIO 1997 ), suggesting that the tranposase probably cuts independently at the ends. Finally, DSBs generated by I-SceI appear to be repaired rapidly in Drosophila (this work; GONG and GOLIC 2003 ). Two simultaneous I-SceI-generated DSBs 8 kb apart could be detected in only 2–7% of cells using a heat-shock protocol similar to ours (GONG and GOLIC 2003 ). It is likely that DSBs at the P-element ends are also repaired rapidly so that a P element is infrequently excised by double cutting. This is consistent with the observation that a P element transposed very poorly to a new position, even when it experienced 82% precise loss due to SSA (PRESTON et al. 2002 ). Therefore, allelic GC induced by P transposition may often require exonucleolytic degradation of the entire element. This could at least in part explain the fact that a P element of 625 bp permitted more frequent allelic conversion (15%; ENGELS et al. 1990 ) than a P element that was 5 kb in size (7%; PRESTON et al. 2002 ). One may ask then why SSA could be induced to 82% with the same larger P element. We are led to conclude that a single DSB at either terminus would be sufficient to initiate the SSA process.

In summary, the absence of homology immediately opposite to the DSB in prior studies may have hidden the highly efficient process of allelic GC in Drosophila. To definitively attribute this high efficiency to the avid pairing of homologs in Drosophila, one has to study allelic GC in a situation in which mitotic pairing is disrupted, similar to GC studies by ENGELS et al. 1990 . Moreover, because a direct comparison between GC templated from the homolog vs. the sister could not be made by our assays, it may well be that intersister GC is still the most efficient HR pathway in Drosophila, and the efficient allelic GC that we observed was due to some combination of the following reasons: (1) the majority of the male mitotic germ cells were in G₁ at the time of heat shock, (2) the chromosomes repaired by intersister GC were often cut again by I-SceI, and (3) both sister chromatids were often cut simultaneously inhibiting intersister GC.

	ACKNOWLEDGMENTS

We thank Simon Titen and Su-chin Wei for technical assistance. This work was supported by National Institutes of Health grant GM-5604 to K.G.G. and the intramural research program of the National Cancer Institute.

Manuscript received April 15, 2003; Accepted for publication August 7, 2003.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ADAMS, M. D., M. MCVEY, and J. J. SEKELSKY, 2003  Drosophila BLM in double-strand break repair by synthesis-dependent strand annealing. Science 299:265-267.[Abstract/Free Full Text]

BEALL, E. L. and D. C. RIO, 1997  Drosophila P-element transposase is a novel site-specific endonuclease. Genes Dev. 11:2137-2151.[Abstract/Free Full Text]

BELLAICHE, Y., V. MOGILA, and N. PERRIMON, 1999  I-SceI endonuclease, a new tool for studying DNA double-strand break repair mechanisms in Drosophila. Genetics 152:1037-1044.[Abstract/Free Full Text]

BIBIKOVA, M., M. GOLIC, K. G. GOLIC, and D. CARROLL, 2002  Targeted chromosomal cleavage and mutagenesis in Drosophila using zinc-finger nucleases. Genetics 161:1169-1175.[Abstract/Free Full Text]

BURGESS, S. M., N. KLECKNER, and B. M. WEINER, 1999  Somatic pairing of homologs in budding yeast: existence and modulation. Genes Dev. 13:1627-1641.[Abstract/Free Full Text]

COLLEAUX, L., L. D'AURIOL, F. GALIBERT, and B. DUJON, 1988  Recognition and cleavage site of the intron-encoded omega transposase. Proc. Natl. Acad. Sci. USA 85:6022-6066.[Abstract/Free Full Text]

DONOHO, G., M. JASIN, and P. BERG, 1998  Analysis of gene targeting and intrachromosomal homologous recombination stimulated by genomic double-strand breaks in mouse embryonic stem cells. Mol. Cell. Biol. 18:4070-4078.[Abstract/Free Full Text]

DUCAU, J., J. C. BREGLIANO, and C. DE LA ROCHE SAINT-ANDRE, 2000  Gamma-irradiation stimulates homology-directed DNA double-strand break repair in Drosophila embryo. Mutat. Res. 460:69-80.[Medline]

ENGELS, W. R., D. M. JOHNSON-SCHLITZ, W. B. EGGLESTON, and J. SVED, 1990  High-frequency P element loss in Drosophila is homolog dependent. Cell 62:515-525.[Medline]

ENGELS, W. R., C. R. PRESTON, and D. M. JOHNSON-SCHLITZ, 1994  Long-range cis preference in DNA homology search over the length of a Drosophila chromosome. Science 263:1623-1625.[Abstract/Free Full Text]

ESPOSITO, M. S., 1978  Evidence that spontaneous mitotic recombination occurs at the two-strand stage. Proc. Natl. Acad. Sci. USA 75:4436-4440.[Abstract/Free Full Text]

FABRE, F., 1978  Induced intragenic recombination in yeast can occur during the G1 mitotic phase. Nature 272:795-798.[Medline]

FUNG, J. C., W. F. MARSHALL, A. DERNBURG, D. A. AGARD, and J. W. SEDAT, 1998  Homologous chromosome pairing in Drosophila melanogaster proceeds through multiple independent initiations. J. Cell Biol. 141:5-20.[Abstract/Free Full Text]

GLOOR, G. B., C. R. PRESTON, D. M. JOHNSON-SCHLITZ, N. A. NASSIF, and R. W. PHILLIS et al., 1993  Type I repressors of P element mobility. Genetics 135:81-95.[Abstract]

GOLIC, M. M. and K. G. GOLIC, 1996  A quantitative measure of the mitotic pairing of alleles in Drosophila melanogaster and the influence of structural heterozygosity. Genetics 143:385-400.[Abstract]

GONG, W. J. and K. G. GOLIC, 2003  Ends-out, or replacement, gene targeting in Drosophila. Proc. Natl. Acad. Sci. USA 100:2556-2561.[Abstract/Free Full Text]

HIRAOKA, Y., A. F. DERNBURG, S. J. PARMELEE, M. C. RYKOWSKI, and D. A. AGARD et al., 1993  The onset of homologous chromosome pairing during Drosophila melanogaster embryogenesis. J. Cell Biol. 120:591-600.[Abstract/Free Full Text]

KADYK, L. C. and L. H. HARTWELL, 1992  Sister chromatids are preferred over homologs as substrates for recombinational repair in Saccharomyces cerevisiae.. Genetics 132:387-402.[Abstract]

KARATHANASIS, E. and T. E. WILSON, 2002  Enhancement of Saccharomyces cerevisiae end-joining efficiency by cell growth stage but not by impairment of recombination. Genetics 161:1015-1027.[Abstract/Free Full Text]

JEGGO, P. A., 1998  DNA breakage and repair. Adv. Genet. 38:185-218.[Medline]

JOHNSON, R. D. and M. JASIN, 2000  Sister chromatid gene conversion is a prominent double-strand break repair pathway in mammalian cells. EMBO J. 19:3398-3407.[Medline]

LEE, S. E., R. A. MITCHELL, A. CHENG, and E. A. HENDRICKSON, 1997  Evidence for DNA-PK-dependent and -independent DNA double-strand break repair pathways in mammalian cells as a function of the cell cycle. Mol. Cell. Biol. 17:1425-1433.[Abstract]

LEE, S. E., F. PAQUES, J. SYLVAN, and J. E. HABER, 1999  Role of yeast SIR genes and mating type in directing DNA double-strand breaks to homologous and non-homologous repair paths. Curr. Biol. 9:767-770.[Medline]

LEWIS, E. B., 1954  The theory and application of a new method of detecting chromosomal rearrangements in Drosophila melanogaster.. Am. Nat. 88:225-239.

LIANG, F., M. HAN, P. J. ROMANIENKO, and M. JASIN, 1998  Homology-directed repair is a major double-strand break repair pathway in mammalian cells. Proc. Natl. Acad. Sci. USA 95:5172-5177.[Abstract/Free Full Text]

LICHTEN, M. and J. E. HABER, 1989  Position effects in ectopic and allelic mitotic recombination in Saccharomyces cerevisiae.. Genetics 123:261-268.[Abstract/Free Full Text]

LIN, F. L., K. SPERLE, and N. STERNBERG, 1984  Model for homologous recombination during transfer of DNA into mouse L cells: role for DNA ends in the recombination process. Mol. Cell. Biol. 4:1020-1034.[Abstract/Free Full Text]

LIN, Y., T. LUKACSOVICH, and A. S. WALDMAN, 1999  Multiple pathways for repair of DNA double-strand breaks in mammalian chromosomes. Mol. Cell. Biol. 19:8353-8360.[Abstract/Free Full Text]

LINDSLEY, D. L., and K. T. TOKUYASU, 1980 Spermatogenesis, pp. 225–294 in The Genetics and Biology of Drosophila, Vol. 2d, edited by M. ASHBURNER and T. R. F. WRIGHT. Academic Press, New York.

MALKOVA, A., E. L. IVANOV, and J. E. HABER, 1996  Double-strand break repair in the absence of RAD51 in yeast: a possible role for break-induced DNA replication. Proc. Natl. Acad. Sci. USA 93:7131-7136.[Abstract/Free Full Text]

MARYON, E. and D. CARROLL, 1991  Characterization of recombination intermediates from DNA injected into Xenopus laevis oocytes: evidence for a nonconservative mechanism of homologous recombination. Mol. Cell. Biol. 11:3278-3287.[Abstract/Free Full Text]

METZ, C. W., 1916  Chromosome studies on the Diptera II: the paired association of chromosomes in the Diptera, and its significance. J. Exp. Zool. 21:213-279.

MOYNAHAN, M. E. and M. JASIN, 1997  Loss of heterozygosity induced by a chromosomal double-strand break. Proc. Natl. Acad. Sci. USA 94:8988-8993.[Abstract/Free Full Text]

NASSIF, N. and W. ENGELS, 1993  DNA homology requirements for mitotic gap repair in Drosophila. Proc. Natl. Acad. Sci. USA 90:1262-1266.[Abstract/Free Full Text]

PAQUES, F. and J. E. HABER, 1999  Multiple pathways of recombination induced by double-strand breaks in Saccharomyces cerevisiae.. Microbiol. Mol. Biol. Rev. 63:349-404.[Abstract/Free Full Text]

PERRIN, A., M. BUCKLE, and B. DUJON, 1993  Asymmetrical recognition and activity of the I-SceI endonuclease on its site and on intron-exon junctions. EMBO J. 12:2939-2947.[Medline]

PLESSIS, A., A. PERRIN, J. E. HABER, and B. DUJON, 1992  Site-specific recombination determined by I-SceI, a mitochondrial group I intron-encoded endonuclease expressed in the yeast nucleus. Genetics 130:451-460.[Abstract]

PRESTON, C. R. and W. R. ENGELS, 1996  P-element-induced male recombination and gene conversion in Drosophila. Genetics 144:1611-1622.[Abstract]

PRESTON, C. R., W. ENGELS, and C. FLORES, 2002  Efficient repair of DNA breaks in Drosophila: evidence for single-strand annealing and competition with other repair pathways. Genetics 161:711-720.[Abstract/Free Full Text]

PUCHTA, H., B. DUJON, and B. HOHN, 1993  Homologous recombination in plant cells is enhanced by in vivo induction of double strand breaks into DNA by a site-specific endonuclease. Nucleic Acids Res. 21:5034-5040.[Abstract/Free Full Text]

RICHARDSON, C., M. E. MOYNAHAN, and M. JASIN, 1998  Double-strand break repair by interchromosomal recombination: suppression of chromosomal translocations. Genes Dev. 12:3831-3842.[Abstract/Free Full Text]

RONG, Y. S. and K. G. GOLIC, 1998  Dominant defects in Drosophila eye pigmentation resulting from a euchromatic-heterochromatic fusion gene. Genetics 150:1551-1566.[Abstract/Free Full Text]

RONG, Y. S. and K. G. GOLIC, 2000  Gene targeting by homologous recombination in Drosophila. Science 288:2013-2018.[Abstract/Free Full Text]

RONG, Y. S., S. W. TITEN, H. B. XIE, M. M. GOLIC, and M. BASTIANI et al., 2002  Targeted mutagenesis by homologous recombination in D. melanogaster.. Genes Dev. 16:1568-1581.[Abstract/Free Full Text]

ROUET, P., F. SMIH, and M. JASIN, 1994  Expression of a site-specific endonuclease stimulates homologous recombination in mammalian cells. Proc. Natl. Acad. Sci. USA 91:6064-6068.[Abstract/Free Full Text]

RUDIN, N., E. SUGARMAN, and J. E. HABER, 1989  Genetic and physical analysis of double-strand break repair and recombination in Saccharomyces cerevisiae.. Genetics 122:519-534.[Abstract/Free Full Text]

SJOGREN, C. and K. NASMYTH, 2001  Sister chromatid cohesion is required for postreplicative double-strand break repair in Saccharomyces cerevisiae.. Curr. Biol. 11:991-995.[Medline]

STARK, J. M. and M. JASIN, 2003  Extensive loss of heterozygosity is suppressed during homologous repair of chromosomal breaks. Mol. Cell. Biol. 23:733-743.[Abstract/Free Full Text]

STEVEN, N. M., 1908  A study of the germ cells of certain Diptera, with reference to the heterochromosomes and the phenomena of synapsis. J. Exp. Zool. 5:359-383.

TAKATA, M., M. S. SASAKI, E. SONODA, C. MORRISON, and M. HASHIMOTO et al., 1998  Homologous recombination and non-homologous end-joining pathways of DNA double-strand break repair have overlapping roles in the maintenance of chromosomal integrity in vertebrate cells. EMBO J. 17:5497-5508.[Medline]

VIRGIN, J. B., J. P. BAILEY, F. HASTEH, J. NEVILLE, and A. COLE et al., 2001  Crossing over is rarely associated with mitotic intragenic recombination in Schizosaccharomyces pombe.. Genetics 157:63-77.[Abstract/Free Full Text]

This article has been cited by other articles:

				J. P. Blumenstiel, R. Fu, W. E. Theurkauf, and R. S. Hawley Components of the RNAi Machinery That Mediate Long-Distance Chromosomal Associations Are Dispensable for Meiotic and Early Somatic Homolog Pairing in Drosophila melanogaster Genetics, November 1, 2008; 180(3): 1355 - 1365. [Abstract] [Full Text] [PDF]

G. Gao, C. McMahon, J. Chen, and Y. S. Rong A powerful method combining