RPA coordinates DNA end resection and prevents formation of DNA hairpins

Abstract

Replication protein A (RPA) is an essential eukaryotic single-stranded DNA binding protein with a central role in DNA metabolism. RPA directly participates in DNA double-strand break repair by stimulating 5'-3' end resection by the Sgs1/BLM helicase and Dna2 endonuclease in vitro. Here we investigated the role of RPA in end resection in vivo, using a heat-inducible degron system that allows rapid conditional depletion of RPA in Saccharomyces cerevisiae. We found that RPA depletion eliminated both the Sgs1-Dna2- and Exo1-dependent extensive resection pathways and synergized with mre11Δ to prevent end resection. The short single-stranded DNA tails formed in the absence of RPA were unstable due to 3' strand loss and the formation of fold-back hairpin structures that required resection initiation and Pol32-dependent DNA synthesis. Thus, RPA is required to generate ssDNA, and also to protect ssDNA from degradation and inappropriate annealing that could lead to genome rearrangements.

Figure 1. RPA is essential for both extensive resection pathways

A. Schematic representation of the band disappearance assay for DSB end resection. Red bars show the location of the two probes used for hybridization. B. Southern blot analysis of the StyI-XhoI digested genomic DNA from indicated strains taken at different time points after HO induction, showing resection 0.7 kb and 2.6 kb from the DSB. The bracket indicates the smeared products observed in the td-RFA1 derivatives and the sgs1Ä exo1Ä mutant. C. Quantification of the 2.4 kb fragment located 2.6 kb from the DSB by the ratio at each time point to 0 h. The mean values from 3 independent trials are plotted and error bars show SD. D. Schematic representation of the quantitative PCR method to monitor HO-induced DSB end resection. E. Plot showing the ratio of resected DNA amongst HO cut DNA at each time point by qPCR analysis. The mean values from 3 independent trials are plotted and error bars show SD. See also Figures S1 and 2

Figure 2. MRX-Sae2 and RPA act sequentially in end resection

A synergistic defect of td-RFA1 and mre11Δ or sae2Δ in resection was seen by both the band disappearance assay (A) and qPCR (B). Plot showing the ratio of resected DNA (0.7 kb fragment) amongst the HO cut DNA at each time point by qPCR analysis. The mean values from 3 independent trials are plotted and error bars show SD.

Figure 3. RPA is required to shield the 3’ ssDNA tails from nucleolytic degradation

A. Both 3’ and 5’ end degradation were observed (R1a) when Rfa1 was depleted. Alkaline electrophoresis of XbaI-digested genomic DNA from different time points after HO induction. ssDNA intermediates were detected by 3’ or 5’ strand specific RNA probes spanning the region from 0.2 kb to 0.7 kb distal of the HO cut site. B. Schematic showing alkaline gel analysis of genomic DNA digested by XbaI. Possible DNA intermediates formed after HO cutting and resection are shown. The black bar shows the location of the probe used. The thin gray lines represent fragments detected by 3’ strand specific probe. C. Only 3’ tails were detected in sgs1 exo1 and td-RFA1 cells. StyI digested genomic DNA was run on a neutral gel, transferred without denaturing and hybridized with the indicated strand specific probes. See also Figure S3

Figure 4. RPA is required for Dna2 but not for Exo1 localization to DSBs

A. Epifluorescence microscopy showing DSB induced foci formation of YFP-tagged Rfa1 or td-Rfa1 and Dna2-CFP in G2/M arrested cells. Arrowheads indicate foci. Scale bar is equal to 3 m. B. Quantitation of Dna2-CFP foci. The percentage of cells with Dna2 foci was quantitated for the experiment shown in panel A, of which more than 90% co-localized with RPA. C. Exo1 focus formation is independent of Rfa1. Exo1-YFP and Rfa1-CFP or td-Rfa1-CFP were monitored before (t = 0) and 2 or 3 h after HO expression in G2-M-arrested cells. The left panel shows Rfa1-CFP cells and the right panel shows td-Rfa1-CFP cells. For the CFP channel, an 8-fold contrast enhanced version of the images from td-Rfa1-CFP cells are shown to illustrate that some cells contain weak residual Rfa1 foci that co-localize with Exo1 (t = 2), while other cells exhibit an Exo1 focus even in the complete absence of detectable td-Rfa1-CFP. D. Quantitation of Exo1 and Rfa1 foci. For the experiment in panel C, 200-600 cells were examined for Exo1 and Rfa1 foci at each time point. Error bars indicate 95% confidence intervals. Significance was determined by Fisher's exact test. See also Figure S4

Figure 5. RPA prevents the formation of fold-back structures

A. Schematic showing alkaline gel analysis of genomic DNA digested by StyI. Possible DNA intermediates formed after HO cutting and resection are shown. The red bars show the locations of the probes used. The red lines represent fragments detected by the 3’ strand specific probe close to the HO cut site. B. Four different dsDNA probes were used to determine the range of fragments r1b and r2b. C. Alkaline electrophoresis of the StyI-digested genomic DNA from different time points after HO induction. ssDNA intermediates were detected by the 3’ strand specific RNA probe. r1b and r2b bands can be eliminated by mung bean nuclease treatment of StyI-digested DNA. D. Schematic showing the formation of hairpins. Alkaline electrophoresis of the StyI-digested genomic DNA from different mutants showing their contribution to hairpin formation. See also Figures S3 and S5

Figure 6. Model for RPA-mediated coordination of DNA end processing

Mre11 and Sae2 initiate resection to form short 3’ tails. In the absence of RPA Dna2 is not recruited to the break site, extensive resection is eliminated, and the 3’ overhangs tend to form secondary structures by annealing between short repeats. Following annealing between inverted repeats, the 3’ unpaired flap is trimmed and after DNA synthesis is ligated to the 5’ end to form a hairpin. The 3’ strand loss could occur by Mre11-dependent degradation of 3’ ends or by hairpin cleavage before (shown) or after DNA synthesis and ligation.