Sgs1 function in the repair of DNA replication intermediates is separable from its role in homologous recombinational repair

Abstract

Mutations in human homologues of the bacterial RecQ helicase cause diseases leading to cancer predisposition and/or shortened lifespan (Werner, Bloom, and Rothmund-Thomson syndromes). The budding yeast Saccharomyces cerevisiae has one RecQ helicase, Sgs1, which functions with Top3 and Rmi1 in DNA repair. Here, we report separation-of-function alleles of SGS1 that suppress the slow growth of top3Delta and rmi1Delta cells similar to an SGS1 deletion, but are resistant to DNA damage similar to wild-type SGS1. In one allele, the second acidic region is deleted, and in the other, only a single aspartic acid residue 664 is deleted. sgs1-D664Delta, unlike sgs1Delta, neither disrupts DNA recombination nor has synthetic growth defects when combined with DNA repair mutants. However, during S phase, it accumulates replication-associated X-shaped structures at damaged replication forks. Furthermore, fluorescent microscopy reveals that the sgs1-D664Delta allele exhibits increased spontaneous RPA foci, suggesting that the persistent X-structures may contain single-stranded DNA. Taken together, these results suggest that the Sgs1 function in repair of DNA replication intermediates can be uncoupled from its role in homologous recombinational repair.

Figure 1

Unlike sgs1Δ yeast, sgs1-AR2Δ yeast is not sensitive to DNA damage but can suppress top3Δ slow growth similar to sgs1Δ. Yeast with the acidic regions (AR1 or AR2) of Sgs1 deleted either separately or together (sgs1-AR1–2Δ) were analysed for SGS1 function. (A) Schematic representation of the functional motifs and domains of Sgs1. The grey bars above the protein represent the Sgs1 segments that mediate two-hybrid interactions with the indicated proteins. (B) sgs1-AR2Δ and sgs1-AR1–2Δ suppress top3Δ slow growth similarly to sgs1Δ, whereas sgs1-AR1Δ only slightly improves the growth rate of the top3Δ strain. (C) Unlike sgs1Δ or sgs1-AR1–2Δ, yeast with sgs1-AR1Δ or sgs1-AR2Δ is not sensitive to HU or MMS. Yeast were grown to log phase and plated in 10-fold serial dilutions onto YPD or YPD with 100 mM HU or 0.03% MMS.

Figure 2

Deletion of aspartic acid residue 664 in SGS1 causes a separation of function in SGS1. (A) Ten-fold serial dilutions were plated onto YPD or YPD with 0.02% MMS or 100 mM HU. Unlike sgs1Δ yeast, sgs1-D664Δ and sgs1-D664A yeast are not sensitive to HU or MMS. (B) sgs1-D664Δ suppresses the slow doubling times of top3Δ or rmi1Δ yeast. (C) sgs1-D664Δ suppresses the HU and MMS sensitivity of top3Δ. Ten-fold serial dilutions were plated onto YPD, YPD with 0.002% MMS, or YPD with 5 mM HU. Note, lower drug concentrations than in (A) were used due to the hypersensitivity of top3Δ.

Figure 3

Unlike sgs1Δ yeast, sgs1-D664Δ yeast do not have recombination defects. (A) A schematic representation of the SUP4-o locus with a URA3 insertion is shown, not to scale (Wallis et al, 1989). The rate of recombination events generating ura3 colonies was assayed and plotted for WT, sgs1Δ, sgs1-D664Δ, and sgs1-AR2Δ yeast strains. The recombination rate of sgs1-AR2Δ and WT were determined separately as indicated by the dotted line. Standard deviations are graphed. t-test analysis revealed that sgs1Δ has a significantly different recombination rate (P⩽0.001) when compared with WT, sgs1-D664Δ, or sgs1-AR2Δ, which were not significantly different from each other (P⩾0.05). (B) A schematic representation of the rDNA (35S repeated sequence) with ADE2 and CAN1 gene insertions is shown. Frequency of the generation of CAN^R, ade2 recombinants was measured in WT, sgs1Δ, sgs1-D664Δ, and sgs1-AR2Δ yeast strains. Standard deviations are graphed.

Figure 4

Sgs1-D664Δ protein levels are reduced but the protein levels of Top3 and Rmi1 are unchanged. The protein levels were calculated using the ImageJ gel analysis tool by comparing the loading control (an unrelated abundant protein Adh1) to the protein of interest. Protein was extracted from equal cell numbers of the untagged parent strain (WT) and strains expressing Sgs1-3 × HA, Sgs1-D664Δ-3 × HA, Sgs1-AR1Δ-3 × HA, and Sgs1-AR2Δ-3 × HA, were analysed by protein blots using anti-HA antibodies. The expression of the mutant sgs1 alleles were 0.49 for Sgs1-D664Δ, 0.03 for Sgs1-AR1Δ, and 0.29 for Sgs1-AR2Δ to that of wild-type Sgs1. Top3–TAP and Rmi1–TAP protein levels were analysed by protein blot in SGS1, sgs1-D664Δ, and sgs1Δ strain backgrounds. Protein extracts were made from equal cell numbers and analysed for Top3 and Rmi1 levels by protein blot using anti-protein A antibodies and protein levels were quantitated from three experiments. For Top3–TAP strains, sgs1-D664Δ was 0.90±0.35 and sgs1Δ was 1.19±0.56 and for Rmi1–TAP strains, sgs1-D664Δ was 0.70±0.28 and sgs1Δ was 0.92±0.30 to that of SGS1. t-test analysis did not detect any significant difference between Top3 or Rmi1 protein levels in SGS1, sgs1Δ, or sgs1-D664Δ strains.

Figure 5

The number of cells with RPA foci, but not Rad52 foci, increases in sgs1-D664Δ. Spontaneous formation of Rfa1–YFP and Rad52–CFP nuclear foci was analysed in SGS1, sgs1Δ, and sgs1-D664Δ backgrounds. Images of Rfa1 and Rad52 are shown with white arrowheads indicating a focus. The average of three independent experiments is plotted with standard error bars.

Figure 6

Replication-associated X-structures accumulate in sgs1-D664Δ that are independent of top3Δ. (A) Two-dimensional (2D) gel electrophoresis was conducted to visualize replication fork progression and processing at ARS305. Cells were first arrested in G2/M with nocodazole at 25°C and then released into YPD medium with 0.033% MMS at 30°C. ARS305 was analysed for replication intermediates after 60, 120, 180, and 240 min after release from nocodazole by Southern blotting. A cartoon of the replication intermediates observed is shown. X-structure accumulation was quantitated and graphed. Cells with sgs1-D664Δ accumulate X-structures. (B) Replication intermediates formed at ARS305 were analysed by 2D gel electrophoresis in WT, sgs1-D664Δ, top3Δ, and sgs1-D664Δ top3Δ strains as described for (A). Replication-associated X-structures accumulate in sgs1-D664Δ top3Δ to the same extent as in sgs1-D664Δ.

Figure 7

Replication-associated X-structure accumulation in sgs1Δ and sgs1-D664Δ cells is Rad51 dependent. Replication fork progression and processing at ARS305 was visualized by 2D gel electrophoresis and X-structure accumulation was quantitated.

Figure 8

Model of sgs1-D664Δ function in DNA repair and DNA replication. During DNA replication, DNA damage can occur that produces a DNA intermediate, the exact structure of which is unknown, but can be resolved in an Sgs1-dependent manner. In wild-type cells, the DNA intermediate is acted on by Top3, ultimately leading to resolution of the DNA damage (left arrows). When TOP3 is disrupted, Sgs1 creates a ‘toxic intermediate' that leads to hyper-recombination and slow growth. In the sgs1-D664Δ background, the replication lesions are repaired through a replication-associated repair pathway that produces replication-associated X-shaped structures as DNA intermediates but which can ultimately be resolved without the need for Top3 activity (right arrows). This model is in agreement with the findings that sgs1-D664Δ yeast is not sensitive to DNA-damaging agents but can suppress top3Δ slow growth as sgs1-D664Δ avoids shunting the DNA intermediate through the Top3-mediated DSB repair pathway.