RAD51 is required for the repair of plasmid double-stranded DNA gaps from either plasmid or chromosomal templates

Abstract

DNA double-strand breaks may be induced by endonucleases, ionizing radiation, chemical agents, and mechanical forces or by replication of single-stranded nicked chromosomes. Repair of double-strand breaks can occur by homologous recombination or by nonhomologous end joining. A system was developed to measure the efficiency of plasmid gap repair by homologous recombination using either chromosomal or plasmid templates. Gap repair was biased toward gene conversion events unassociated with crossing over using either donor sequence. The dependence of recombinational gap repair on genes belonging to the RAD52 epistasis group was tested in this system. RAD51, RAD52, RAD57, and RAD59 were required for efficient gap repair using either chromosomal or plasmid donors. No homologous recombination products were recovered from rad52 mutants, whereas a low level of repair occurred in the absence of RAD51, RAD57, or RAD59. These results suggest a minor pathway of strand invasion that is dependent on RAD52 but not on RAD51. The residual repair events in rad51 mutants were more frequently associated with crossing over than was observed in the wild-type strain, suggesting that the mechanisms for RAD51-dependent and RAD51-independent events are different. Plasmid gap repair was reduced synergistically in rad51 rad59 double mutants, indicating an important role for RAD59 in RAD51-independent repair.

FIG. 1

Physical map of MET17, the molecular structure of the gap, and the phenotype conferred to cells by the met17-s mutation. (A) The hatched box indicates the 1.5-kb ORF (arrow, start codon; ∗, stop codon). MET17 plasmids to be used as substrates in the gap repair assay were digested with BspEI and EcoNI to produce a 238-bp gap, indicated by a black box. The plasmid and chromosomal DNA donor sequences contain a nonsense mutation that destroyed a SnaBI site (met17-s) 216 bp downstream of the EcoNI site. (B) The gap produced by BspEI-EcoNI digests consists of noncomplementary 5′ overhangs that overlap in one nucleotide (C · C) and is expected to provide a poor substrate for ligation. However, degradation or melting of the EcoNI end could provide microhomologies (C · G and/or GG or CC) for annealing to the overhang produced by BspEI digestion. (C) The first sequence shows the SnaBI site in the MET17 allele; the second line of sequence indicates the A insertion at the SnaBI site, which creates a stop codon and disrupts the recognition site for SnaBI. The met17-s mutation confers to cells a dark brown phenotype when grown on medium supplemented with lead (Pb).

FIG. 2

Rationale of the gap repair assay. (A) The double-strand gap in MET17 on either ARS (open square), CEN ARS (open square and open circle), or integrating (NON) plasmid is repaired from homologous chromosomal or plasmid met17-s sequences. (B) Repair of gapped ARS plasmids without a crossover produces a repaired MET17 plasmid and an unchanged donor sequence (chromosome or plasmid); repair associated with a crossover results in an integrated ARS plasmid. Repair of a gapped CEN ARS plasmid has to occur by a noncrossover mechanism because integration results in a dicentric chromosome or plasmid, which is inviable. Repair of the gapped plasmid that contains no ARS element has to occur by integration to yield a stable transformant. The products were drawn based on the assumption that the gap in the plasmid is not extended by nucleases over the MET17 SnaBI site. (C) The products expected from gap repair unassociated and associated with crossover were distinguished by monitoring the selective and colony color phenotypes conferred by the URA3 and MET17 markers to the recombinants and by the mitotic stability phenotype of these markers. For the identification of single patches by numbers, a grid is included in C6. Patches on independent plates are shown; note that plates 4 and 5 are not from the same master plate. Confluent growth on 5-FOA medium indicated that the Ura⁺ phenotype was mitotically unstable (i.e., C1 no. 2, and C2, no. 2). Secondary pop-out recombination between duplicated MET17 alleles delineating URA3 leads to the formation of papillae on 5-FOA due to the excision and loss of URA3 (i.e., C1, no. 8 and C3, no. 2). Cells displaying a dark brown phenotype when grown on lead (Pb) plates indicated that the Met⁺ phenotype was unstable in cells transformed with ARS plasmid (i.e., C4, no. 10) or absent (i.e., C5, no. 2). During further incubation of plate C5 for 4 to 5 days at room temperature, patches that were previously white turned a beige color and showed dark brown papillae, indicating the progressive loss of CEN ARS plasmids (i.e., C5, no. 8). White cells were diagnostic for a stable Met⁺ phenotype (i.e., C4 no. 2, and C5, no. 46 and 50).

FIG. 3

Structural analysis of Ura⁺ transformants. Yeast DNA isolated from Ura⁺ transformants was digested with BamHI (B) and SnaBI (S), and Southern blot analysis was performed with a MET17 DNA fragment as the hybridization probe. (A) Schematic representation (not drawn to scale) of the expected DNA repair product to generate an unstable (u) Ura⁺ Met⁺ phenotype; also shown is the chromosomal met17-s allele. The sizes of DNA fragments that hybridize to the probe are shown. The lower panel shows a Southern blot of this class of events. (B) Schematic representation of integration events to produce a stable Ura^+s Met^+s phenotype. The three simplest classes are shown, although multiple integration events to produce fragments of 1.5 and/or 7.2 kb also occur. The lower panel shows a representative Southern blot of this class of events. (C) Schematic representation of events to produce a Ura^+u Met⁻ phenotype. Conversion of the plasmid MET17 to met17-s is monitored by the appearance of a 7.2-kb fragment. A 1.3-kb BamHI-SnaBI fragment that hybridizes to the probe is diagnostic for a nonhomologous end joining of the gapped plasmid substrate. For both classes, the 6.7-kb met17-s allele is unchanged. Events from a wild-type strain (RAD) are shown on the left Southern blot, and those from a rad51 strain are shown on the right. (D) Schematic representation of an integration event to produce a stable Ura^+s Met⁻ phenotype. The 9.9- and 3.0-kb fragments are diagnostic of two copies of met17-s; multiple integration results in an additional fragment of 7.2 kb. Southern blots of DNA from RAD and rad51 strains are shown below the schematic. Included as size markers (M) are SnaBI-BamHI-digested plasmid pSB110 (MET17) and genomic DNA, isolated from an untransformed tester strain (met17-s), which produce signals of 1.5 and 6.7 kb, respectively. Fragment sizes are given in kilobase pairs on the left and were determined relative to HindIII-digested lambda DNA run as a standard.

FIG. 4

Model for RAD51-independent recombination of inverted repeats. Following introduction of a DSB in one of the repeats, one end is resected to produce a 3′ single-stranded tail that invades the other repeat, possibly through the action of Rad52. DNA synthesis is primed from the invading strand and proceeds to the end of the DNA molecule (the other side of the break). The linear molecule formed contains short sequences corresponding to the 5′ end of the inverted repeat at its ends. If the linear intermediate is degraded by a 5′-3′ exonuclease so that complementary single-stranded regions are revealed, then strand annealing can occur to form two types of products. One has the same structure as a reciprocal exchange, and the other has the same structure as the parental plasmid. The inverted repeats are shown by thick arrows, DNA synthesized during repair is shown by dashed lines, and sequences between the inverted repeats are designated A and B.

FIG. 5

Models for the recombinational repair of DSBs. (A) The DSBR model. After formation of a DSB, the 5′ ends are resected to form 3′ single-stranded tails. One of the 3′ single-stranded tails invades a homologous duplex and primes DNA synthesis. The displaced strand from the donor duplex pairs with single-stranded DNA at the other side of the break and is the template for DNA synthesis. After ligation, a double Holliday structure is formed and can be resolved to yield noncrossover or crossover products. (B) The synthesis-dependent strand-annealing (SDSA)/migrating D-loop model. The first two steps are the same as those in the DSBR model, but most of the time a double Holliday junction intermediate is not formed. Instead, the invading strand that has been extended by DNA synthesis is displaced from the donor duplex and can anneal to the single-stranded tail on the other side of the break. The resulting gaps are filled by DNA synthesis and ligation to yield a noncrossover product. To account for the crossover products recovered from plasmid gap repair, we propose that 40% of the events form a Holliday junction intermediate and, of these, 50% resolve to generate crossover products. Dashed lines represent newly synthesized DNA.

FIG. 6

Models for inversions between ade2 inverted repeats (9, 46). (A) In G2 cells, the inverted repeats on different chromatids (open arrows) can pair such that the intervening DNA sequences (shown as solid arrows) are in an antiparallel configuration. A gene conversion event initiated within one repeat that extends to the other repeat could result in the inversion of the intervening DNA. (B) A reciprocal crossover between inverted repeats located on the same chromatid inverts the intervening DNA sequences, but the product is indistinguishable from a G2 conversion event. The inverted repeats are shown by open boxes with arrowheads, the short repeat corresponds to a truncation of the 5′ end, and the long repeat contains a point mutation shown by the vertical line.