The Slx4-Rad1-Rad10 nuclease differentially regulates deletions and duplications induced by a replication fork barrier

Abstract

Genome instability is a hallmark of cancer that can be caused by DNA replication stress. Copy number variation (CNV) is a type of genomic instability that has been associated with both tumorigenesis and drug resistance, but how these structural variants form in response to replication stress is not fully understood. Here, we established a direct repeat genetic reporter in Saccharomyces cerevisiae to detect recombination events that result in either a duplication or a deletion. Using this system, we measured recombination resulting from site-specific replication fork stalling initiated by Tus binding to an array of Ter sites. We found that a Tus/Ter fork block downstream of direct repeats induced CNV by a mechanism involving the Mph1 translocase, Exo1-catalyzed end resection and Rad51-dependent strand invasion. While the Slx4 scaffold protein and its nuclease-binding partner, Rad1-Rad10, were shown to be required for duplications, we found that they suppress deletion formation in this context. These opposing functions suggest that both recombination products arise through a large loop heteroduplex intermediate that is cleaved by Rad1-Rad10 in a manner that promotes duplications and eliminates deletions. Taken together, these studies give insight into the mechanisms governing CNV in the context of replication fork stalling, which may ultimately provide a better understanding of how replication stress contributes to cancer and other diseases characterized by genome instability.

Fig 1. A Tus/Ter block downstream of direct repeats stimulates Rad51 and Rad52-dependent CNV.

A. Schematic of direct repeat reporter consisting of two trp1 fragments (5′Δ-trp1 and trp1-3′Δ) with 426 base pairs of overlapping homologies (indicated by the shaded lines) that are separated by a K. lactis URA3 (URA3) marker. The reporter is integrated ~3.8 kb distal to ARS607 on Chr VI with 21x TerB repeats integrated downstream of the trp1 repeats. Recombination between two truncated trp1 alleles can yield Trp⁺ recombinants (TDs) or Ura^- recombinants (deletions). PvuII restriction sites used for Southern blots in are indicated where appropriate. Created in BioRender. Triplett, M. (2025) https://BioRender.com/loxxdnu. Frequency of Trp⁺ (B) and Ura^- (C) recombinants in WT, rad52Δ, and rad51Δ strains. Statistical significance was determined by one-way ANOVA on log-transformed data with a Bonferroni post-test. p-values are indicated as follows: ns (not significant) p > 0.05, *p < 0.05, **p < 0.005, ***p < 0.001, ****p < 0.0001. D. Southern blot of the non-recombined parental direct repeat reporter strain (P) and the Trp⁺ and Ura^- recombination products under conditions in which the Tus/Ter block has been induced in WT, rad51Δ, and rad52Δ strains. “M” indicates the molecular weight size marker (1 kb ladder).

Fig 2. Loss of MRX components and sister chromatid cohesion increases the frequency of Tus/Ter-induced TDs.

A. The MRX complex and its functions. Key: Mre11-yellow, Rad50-orange, Xrs2-green. Created in BioRender. Triplett, M. (2025) https://BioRender.com/j1vpn7o. B. Frequencies of Trp⁺ recombinants in WT and mre11Δ, xrs2Δ, sae2Δ, dnl4Δ, and tel1Δ strains. C. Frequencies of Trp⁺ recombinants in WT, mre11Δ, rad51Δ and mre11Δ rad51Δ strains. D. Frequencies of Trp⁺ recombinants in WT and scc1-73 and tof1Δ strains. Statistical significance for recombination assays was determined by one-way ANOVA on log-transformed data with a Bonferroni post-test. p-values are indicated as follows: ns (not significant) p > 0.05, *p < 0.05, **p < 0.005, ***p < 0.001, ****p < 0.0001.

Fig 3. Long-range resection by Exo1 is important for Tus/Ter-induced TD formation.

A. Frequencies of Trp⁺ recombinants in WT, exo1Δ, sgs1Δ, and exo1Δ sgs1Δ strains. B. Frequencies of Ura^- recombinants in WT, exo1Δ, sgs1Δ, and exo1Δ sgs1Δ strains. C. Frequencies of Trp⁺ recombinants in WT and mre11Δ strains in which Exo1 is overexpressed compared to strains without Exo1 overexpression. Statistical significance for recombination assays was determined by one-way ANOVA on log-transformed data with a Bonferroni post-test. p-values are indicated as follows: ns (not significant) p > 0.05, *p < 0.05, **p < 0.005, ***p < 0.001, ****p < 0.0001.

Fig 4. Tus-induced TDs and deletions have specific genetic requirements.

A. Frequencies of Trp⁺ recombinants in WT and strains lacking genes involved in replication fork reversal and regression (rad5Δ, mph1Δ, and rad5Δ mph1Δ). B. Frequencies of Trp⁺ recombinants in WT and strains lacking SSEs (mus81Δ yen1Δ, rad1Δ, slx1Δ, slx4Δ). C. Frequencies of Ura^- recombinants in WT and strains lacking Rad1, Slx4 or Msh3. Statistical significance was determined by one-way ANOVA on log-transformed data with a Bonferroni post-test. p-values are indicated as follows: ns (not significant) p > 0.05, *p < 0.05, **p < 0.005, ***p < 0.001, ****p < 0.0001. D. Fold change in the Tus/Ter-induced recombination frequency relative to WT for rad51Δ, exo1Δ, rad1Δ and slx4Δ mutants.

Fig 5. Loss of Rad52 or Rad51 further reduces TDs in the rad1

Δ background. Frequencies of Trp⁺ (A) and Ura^- (C) recombinants in WT, rad1Δ, rad52Δ and the rad1Δ rad52Δ double mutant. Frequencies of Trp⁺ (B) and Ura^- (D) recombinants in WT, rad1Δ, rad51Δ and the rad1Δ rad51Δ double mutant. Statistical significance was determined by one-way ANOVA on log-transformed data with a Bonferroni post-test. p-values are indicated as follows: ns (not significant) p > 0.05, *p < 0.05, **p < 0.005, ***p < 0.001, ****p < 0.0001.

Fig 6. Cas9^D10A-induced fork breakage bypasses the need for Mph1 and HJ resolvases in formation of TDs.

A. Schematic of a replication fork stalled at the Tus/Ter barrier and fork collapse induced by a nick on the leading strand template. Frequencies of Trp⁺ (B) and Ura^- (C) recombinants under non-induced (no β-estradiol) and induced (+β-estradiol) conditions in WT, mph1Δ, mus81Δ yen1Δ, and rad1Δ strains expressing Cas9^D10A and gRNA6. Statistical significance was determined by one-way ANOVA on log-transformed data with a Bonferroni post-test. p-values are indicated as follows: ns (not significant) p > 0.05, *p < 0.05, **p < 0.005, ***p < 0.001, ****p < 0.0001.

Fig 7. Model for CNV formation at a stalled replication fork.

Upon encountering a Tus/Ter barrier, the replication fork could undergo reversal mediated by Mph1. The reversed fork could be resected through the nuclease activity of Exo1 to form a 3’ ssDNA overhang that includes the 3′-truncated repeat. The end of the reversed nascent strand could invade the annealed parental strands at the 5′-truncated repeat, mediated by Rad52 and Rad51. Following DNA synthesis to copy the 5′ end of TRP1, the D-loop could be resolved by an incoming replication fork resulting in the formation of a large heterologous loop. Cleavage of the strand opposite the loop heterology by Rad1-Rad10, followed by fill-in synthesis would generate a full-length copy of TRP1 and duplicate most of the reporter. A deletion could occur if invasion initiated from the 5′ truncated repeat, copying sequence downstream of the 3′ truncated repeat. Resolution by an incoming replication fork would loop out the parental strand. Cleavage of the nascent strand opposite the loop heterology by Rad1-Rad10, followed by fill-in synthesis would preserve the original structure of the reporter. A deletion product would require cleavage of the ssDNA loop, or segregation of the strands at the next division cycle. The region of homology shared by the repeats is shown in mid-blue, while the 3′ end and 5′ end of TRP1 are shown in dark blue. Parental strand is indicated by black lines, nascent strand by gray lines, and repair-associated DNA synthesis by a dashed gray line.