The DNA damage checkpoint pathway promotes extensive resection and nucleotide synthesis to facilitate homologous recombination repair and genome stability in fission yeast

Abstract

DNA double-strand breaks (DSBs) can cause chromosomal rearrangements and extensive loss of heterozygosity (LOH), hallmarks of cancer cells. Yet, how such events are normally suppressed is unclear. Here we identify roles for the DNA damage checkpoint pathway in facilitating homologous recombination (HR) repair and suppressing extensive LOH and chromosomal rearrangements in response to a DSB. Accordingly, deletion of Rad3(ATR), Rad26ATRIP, Crb2(53BP1) or Cdc25 overexpression leads to reduced HR and increased break-induced chromosome loss and rearrangements. We find the DNA damage checkpoint pathway facilitates HR, in part, by promoting break-induced Cdt2-dependent nucleotide synthesis. We also identify additional roles for Rad17, the 9-1-1 complex and Chk1 activation in facilitating break-induced extensive resection and chromosome loss, thereby suppressing extensive LOH. Loss of Rad17 or the 9-1-1 complex results in a striking increase in break-induced isochromosome formation and very low levels of chromosome loss, suggesting the 9-1-1 complex acts as a nuclease processivity factor to facilitate extensive resection. Further, our data suggest redundant roles for Rad3ATR and Exo1 in facilitating extensive resection. We propose that the DNA damage checkpoint pathway coordinates resection and nucleotide synthesis, thereby promoting efficient HR repair and genome stability.

Figure 1.

Rad3^ATR suppresses break-induced extensive LOH. (A) Schematic of the minichromosome Ch¹⁶-RMGAH. The relative positions of the arg3 marker (diagonal stripes), centromeres (ovals), the MATa site (black), the kanMX6 resistance marker (gray), the complementary ade6 heteroalleles (ade6-M216 and ade6-M210; white) and the his3 marker (vertical stripes) on Ch¹⁶-RMGAH and ChIII are shown as described previously (35). The sizes of the ChIII and Ch¹⁶ are shown. In Ch¹⁶-RMYAH, kanMX6 is replaced by hph. Derepression of pREP81X-HO (not shown) generates a DSB at the MATa target site (scissors). Possible outcomes resulting from DSB induction, together with schematics of the minichromosome, and expected phenotypes are shown. (B) Colony sectoring of wild-type or loh1–1 arg⁺ G418^S ade⁻ his⁻ colonies grown on Edinburgh minimal medium (EMM) plus uracil, histidine and low adenine (5 mg/l) without arginine (arg- plates) thus facilitating detection of extensive LOH (LOH) in the presence (HO off) or absence (HO on) of thiamine. (C) Ten-fold serial dilutions of wild-type (WT) Ch¹⁶-RMGAH (TH2130) or loh1–1 (TH4089) strains on Ye5S plates, Ye5S plates exposed to 300 Gy IR, or 0.005% MMS as indicated. (D) 4',6-diamidino-2-phenylindole (DAPI) stained wild-type Ch¹⁶-RMGAH (TH2130) or loh1–1 (TH4089) strains either untreated or following exposure to 5 mM HU for 6 h. ‘Cut’ phenotypes indicated (yellow arrows). (E) Serial dilutions of wild-type Ch¹⁶-RMGAH (TH2130), loh1–1 (TH4089) with pREP41X empty vector or pREP41X-rad3 (TH4093) on Ye5S and 10 mM HU EMM plates without thiamine, to derepress pREP41X expression. (F) Percentage DSB-induced marker loss of Ch¹⁶-RMGAH in wild-type (TH2130) and rad3Δ (TH2941) backgrounds. The levels of NHEJ/sister chromatid conversion (SCC), GC, Ch¹⁶ loss and extensive LOH are shown. Data are the mean of three experiments and standard errors of the mean are indicated. The asterisk (*) represents significant difference compared to wild-type.

Figure 2.

Break-induced extensive LOH in rad3Δ results from extensive resection, and predominantly isochromosome formation (A). Left panel: PFGE analysis from rad3Δ Ch¹⁶-RMGAH parental strain (TH2941; lane 1), individual arg⁺ G418^S ade⁻ his⁻ (LOH) colonies from wild-type (a CGH confirmed isochromosome I(Ch^16L); lane 2) and rad3Δ (lanes 3–5) backgrounds following DSB induction are shown. Right panel: Southern blot analysis of the PFGE, probed with Spcc4b3.18, which anneals directly distal the centromere on Ch¹⁶-RMGAH and ChIII (as indicated) (B). CGH of wild-type Ch¹⁶-RMGAH (TH2125) and an arg⁺ G418^S ade⁻ his⁻ (LOH) strain (TH8399) carrying a truncated minichromosome that is shorter than the known isochromosome (TH4313) (Figure 2A, lane 1) previously characterized by CGH (35). The Log2 of the LOH:parental signal ratio across the and chromosome III (from which the minichromosome is derived) is shown. (C) A schematic of the structure of the smaller chromosomal element arising following DSB induction in a rad3Δ background as related to the CGH data. CGH analysis of an isochromosome with a duplicated left arm is presented in Supplementary Figure S2 for comparison.

Figure 3.

The DNA damage checkpoint promotes HR and suppresses break-induced LOH. (A) Percentage DSB-induced marker loss of Ch¹⁶-RMGAH in wild-type (TH2130), rad26Δ (TH3410), crb2Δ (TH3383) and OPcdc25 (TH3395) backgrounds. (B) The DNA replication checkpoint does not suppress break-induced LOH. Percentage DSB-induced marker loss of Ch¹⁶-RMGAH in wild-type (TH2130), mrc1Δ (TH3253) and cds1Δ (TH3256) backgrounds. (C) An additional role for Chk1 activation in promoting HR and suppressing break-induced LOH. Percentage DSB-induced marker loss of Ch¹⁶-YAMGH in wild-type (TH3317), chk1Δ (TH3153), rad9-T412A (TH5381), rad4.110 (TH4481) and rad3Δchk1Δ (TH3623) backgrounds. For (A), (B) and (C) the levels of NHEJ/SCC, GC, Ch¹⁶ loss and extensive LOH are shown. Data are the mean of three experiments and standard errors of the mean are indicated. The asterisk (*) represents P < 0.05 compared to wild-type.

Figure 4.

An additional role for Rad17 and the 9-1-1 complex in promoting HR and suppressing break-induced LOH. (A) Percentage DSB-induced marker loss of Ch¹⁶-RMYAH in wild-type (TH4104, TH4121, TH4122, TH4125), rad17Δ (TH7427-TH7430), rad9Δ (TH7588-TH7591), rad1Δ (TH7493, TH7494, TH7495) and hus1Δ (TH7431-TH7434) backgrounds. (B) Percentage DSB-induced marker loss of Ch¹⁶-RMGAH in rad17Δ rad9Δ (TH3454), rad17Δ rad3Δ (TH3455) and rad17Δ crb2Δ (TH3529) backgrounds. (C) Rad3^ATR and Exo1 function redundantly to suppress break-induced LOH. Percentage DSB-induced marker loss of Ch¹⁶-RMGAH in wild-type (TH2130), exo1Δ (TH3378), rad3Δ (TH2941), rad3Δexo1Δ (TH3382) and rad17Δ exo1Δ (TH3701) backgrounds. For (A), (B) and (C) the levels of NHEJ/SCC, GC, Ch¹⁶ loss and extensive LOH are shown. Data are the mean of three experiments and standard errors of the mean are indicated. The asterisk (*) represents significant difference compared to wild-type.

Figure 5.

spd1Δ suppresses the repair defect of rad3Δ and rad26Δ. (A) Five-fold serial dilutions of wild-type (TH2094), spd1Δ (TH4355), rad3Δ (TH7329), rad3Δspd1Δ (TH8295), rad26Δ (TH7330) and rad26Δspd1Δ (TH8194) strains (top panel) and wild-type (TH2094), spd1Δ (TH4355), rad17Δ (TH7331), rad17Δspd1Δ (TH7794), rad9Δ (TH7414), rad9Δspd1Δ (TH7146), rad1Δ (TH7333), rad1Δspd1Δ (TH8249), hus1Δ (TH8296) and hus1Δspd1Δ (TH8195) strains (bottom panel) grown on Ye5S (untreated) and Ye5S + 0.2 μg/ml bleocin. (B) Percentage DSB-induced marker loss in wild-type (TH4121, TH4122, TH4104), spd1Δ (TH4077-TH4079) rad26Δ (TH7424-TH7426) and rad26Δspd1Δ (TH7585-TH7587) backgrounds. Means ± standard errors of three experiments are shown. Asterisk (*) represents significant difference compared to rad26Δ and rad26Δspd1Δ mutants. (C) Percentage DSB-induced marker loss in wild-type (TH4121, TH4122, TH4104), spd1Δ (TH4077-TH4079), rad17Δ (TH7429-TH7430), rad17Δspd1Δ (TH7566-TH7568), rad9Δ (TH7589-TH7591) and rad9Δspd1Δ (TH7464-TH7466) backgrounds. Means ± standard errors of three experiments are shown.

Figure 6.

A role for Rad17 and the 9-1-1 complex in SSA repair. (A) A schematic of a resection and SSA assay as previously described (37). (B) Graph of HOcs-HIS SSA genetic colony assay showing loss of his3+ marker following induction of Purg1lox-HO-endonuclease in wild-type (TH7184), rad3Δ (TH8091) rad17Δ (TH8040) and rad9Δ (TH8050) backgrounds. The genetic assay was repeated independently at least three times. Error bars are ± standard deviation of the mean. (C) Physical analysis of HO-endonuclease cutting and repair by Southern hybridization in wild-type (TH7184), rad3Δ (TH8091) rad17Δ (TH8040) and rad9Δ(TH8050) cells. Genomic DNA extracted after Purg1lox induction at intervals shown, digested with PvuI and NruI, blotted and hybridized to probe as indicated in (A). Marker lane (M) and band sizes (kb) are indicated. The 6.2 kb pre-SSA fragment (*) and 3.1 kb post-SSA fragment (**) are indicated. (D) Graph of band intensities at 360 min without HO induction (OFF) or with HO induction (ON) for blots shown in (C). Blots were scanned using a personal molecular imager^TM (PMITM) and Quantity One Software (Bio-rad). Relative intensities of 6.2 kb pre-SSA fragment and 3.1 kb post-SSA fragments are shown, and were normalized by calculating the intensities of pre- and post-SSA bands as a percentage of the total intensities for these bands for each time point. M indicates DNA size marker and kb sizes of marker bands shown. 360 OFF refers to cells grown in EMM+L+H.

Figure 7.

(A) Model for roles for the DNA damage checkpoint pathway in suppressing extensive LOH and chromosomal rearrangements associated with failed DSB repair. The DNA damage checkpoint pathway promotes efficient HR repair. Failed HR leads to extensive end processing and to chromosome loss or rearrangements. Rad17 and the 9-1-1 complex further suppress break-induced LOH by promoting extensive end processing through the centromere, resulting in loss of the broken chromosome. This is supported by the findings that Rad17 and the 9-1-1 complex are required for extensive resection, removal of the unrepaired broken minichromosome and suppression of extensive LOH. (B) Model for the roles of the DNA damage checkpoint proteins and Exo1 in facilitating extensive resection in S. pombe. Following DSB induction, the 9-1-1 complex (ring) is loaded by Rad17. The 9-1-1 complex facilitates processivity of Exo1 and nuclease X. Rad3^ATR, together with other checkpoint proteins (not shown), promotes dNTP synthesis, promotes nuclease X and additionally inhibits Exo1. This model is supported by the findings that the rad3Δ exo1Δ double mutant phenocopies the DSB repair profile of rad17Δ, leading to high levels of extensive LOH and low levels of minichromosome loss, while rad3Δ or exo1Δ do not; as exo1Δ was not equivalent to rad17Δ or loss of the 9-1-1 complex, this suggests that the 9-1-1 complex additionally provides processivity to another nuclease (X), which requires Rad3 for activity. All checkpoint genes tested are required for transactivating Cdt2 expression, an initial step in damage-induced dNTP synthesis. See the text for details.