Sgs1 regulates gene conversion tract lengths and crossovers independently of its helicase activity

Abstract

RecQ helicases maintain genome stability and suppress tumors in higher eukaryotes through roles in replication and DNA repair. The yeast RecQ homolog Sgs1 interacts with Top3 topoisomerase and Rmi1. In vitro, Sgs1 binds to and branch migrates Holliday junctions (HJs) and the human RecQ homolog BLM, with Top3alpha, resolves synthetic double HJs in a noncrossover sense. Sgs1 suppresses crossovers during the homologous recombination (HR) repair of DNA double-strand breaks (DSBs). Crossovers are associated with long gene conversion tracts, suggesting a model in which Sgs1 helicase catalyzes reverse branch migration and convergence of double HJs for noncrossover resolution by Top3. Consistent with this model, we show that allelic crossovers and gene conversion tract lengths are increased in sgs1Delta. However, crossover and tract length suppression was independent of Sgs1 helicase activity, which argues against helicase-dependent HJ convergence. HJs may converge passively by a "random walk," and Sgs1 may play a structural role in stimulating Top3-dependent resolution. In addition to the new helicase-independent functions for Sgs1 in crossover and tract length control, we define three new helicase-dependent functions, including the suppression of chromosome loss, chromosome missegregation, and synthetic lethality in srs2Delta. We propose that Sgs1 has helicase-dependent functions in replication and helicase-independent functions in DSB repair by HR.

FIG. 1.

DSB-induced allelic HR in WT and sgs1Δ cells. (A) Fates of DSBs in allelic HR substrate. The parent structure is shown at the top, with shading in the recipient ura3 allele indicating silent RFLP markers (see Fig. S1 in the supplemental material). DSBs are created at the HO recognition site. The HO site and X764 (black bars) are inactivating, frameshift mutations. HIS3 is near the telomere, 110 kbp from ura3. Short- and long-tract gene conversions (GC) without crossovers are predominant in wild-type cells. Crossovers in G₂ lead to the gain/loss of HIS3, producing His⁻ and His⁺⁺ products in 50% of subsequent mitotic divisions; the other 50% remain heterozygous at HIS3. BIR leads only to the His⁻ product shown above the dashed line. Loss of the broken chromosome produces Ura⁻ His⁻ products (not shown). (B) DSB-induced HR frequencies, defined as the number of products per YPD colony scored. The remaining colonies were mostly parental but included rare BIR and chromosome loss events (Fig. 5). Values are averages ± standard deviations (SD) (error bars) for four determinations. Short- and long-tract gene conversions include His⁺ (mostly noncrossover [NCO]) and His⁻ (crossover [CO]) products. His⁻ values exclude BIR and chromosome loss products. Ura^+/− includes sectored colonies of several types with different His phenotypes; most Ura^+/− colonies arise from independent HR events in G₂ cells (47). Significant differences (P < 0.01) are indicated by double asterisks. Increased fractions of Ura⁻ products in sgs1Δ indicate longer tract lengths, and increased fractions of His⁻ (H⁻) products indicate increased crossovers.

FIG. 2.

Sgs1 regulation of crossovers is helicase independent. (A) Percentage of His⁻ crossovers including short- (Ura⁺) and long-tract (Ura⁻) HR products. Values are averages ± SD (error bars) for four determinations. Because only 50% of crossovers become His⁻, these crossover values are twice the raw values in Fig. 1B. The double asterisks indicate significant differences at a P value of <0.0001; the single asterisk indicates a P value of 0.0015. All other comparisons showed no significant differences (P > 0.05). These values exclude His⁻ products arising by chromosome loss or BIR. (B) Percentage of His⁺⁺ crossover products (n ≥ 40 per strain).

FIG. 3.

Sgs1 regulation of gene conversion tract lengths is helicase independent. Conversion frequencies for individual markers, 5′ or 3′ of the DSB, are indicated for the wild-type, sgs1Δ, sgs1KR, and sgs1KA strains. See Fig. S1 in the supplemental material for data for the wild type and sgs1Δ. X764 conversion frequencies were determined as fractions of Ura⁻ colonies among an average of 5,600 colonies scored per strain. R5′ and B3′ conversion frequencies were determined by PCR analysis of 30 to 68 products per strain. Significant differences from the wild type are indicated by the single asterisk (P < 0.02) or double asterisks (P < 0.005); Fisher exact tests were used to compare all markers except X764 (t tests).

FIG. 4.

Sgs1 does not suppress crossovers or tract lengths in direct repeats (A) Direct repeat HR substrate and products. ura3 alleles, as shown in Fig. 1A, flank LEU2 and pUC19. Two gene conversion products and two deletion products are diagrammed. (B) DSB-induced direct repeat HR frequencies, reported as described in the legend for Fig. 1. Values are averages ± SD (error bars) of four determinations. Short- and long-tract gene conversions not associated with crossovers are Leu⁺. Only the Ura⁺ Leu⁺ category was significantly different (P < 0.01). Deletions (Leu⁻), which can arise by single-strand annealing or crossovers, were similar in the WT and sgs1Δ.

FIG. 5.

Modest role for Sgs1 in HR initiation. (A) Chromosome loss was determined in >30 Ura⁻ His⁻ colonies per strain, and the resulting values were used to calculate percent loss values (±SD [error bars]) for each of four determinations. Double asterisks indicate P values of <0.0005. (B) Percent BIR was calculated by subtracting G₂ crossover and HR frequencies from total His⁻ frequencies. Values are averages ± SD for four determinations per strain. The asterisk indicates a P value of 0.05.

FIG. 6.

Sgs1 helicase is required for MMS resistance, srs2Δ viability, and proper chromosome segregation. (A) Percent survival after 6 days of growth on YPD with 0.01% MMS for two determinations with the wild type and three determinations for sgs1 mutants. Significant differences are indicated by single asterisks (P < 0.04) or double asterisks (P < 0.002). (B) Meiotic products from diploids that are heterozygous for sgs1 and srs2Δ mutations. Genotypes were determined by PCR: square, WT; circle, sgs1; diamond, srs2Δ; triangle, inferred sgs1 srs2Δ double mutant. (C) Representative phenotypes of large-budded wild-type and sgs1KA mutant cells. DAPI fluorescence and light microscopic images are shown in the left and right panels, respectively, for each strain. (D) Percentage of abnormal nuclear segregation after 0 or 5 h of treatment with 0.1% MMS. For each value, we scored at least 100 large-budded cells.

FIG. 7.

Helicase-dependent and -independent roles for Sgs1. (A) Replication fork restart. A fork blocked by a DNA lesion (triangle) regresses to produce a “chicken foot” structure that allows replication past the lesion (dashed line) using the newly synthesized strand as template. Fork reversal catalyzed by the Sgs1 helicase restores the replication fork. (B) Reverse branch migration catalyzed by Sgs1 helicase converges double HJs for final resolution by Top3; modeled after Ira et al. (26). (C) Sgs1 helicase-independent reduction of crossovers and tract lengths. Double HJs migrate by random walk, and Sgs1 (plus Rmi1) stimulates Top3 activity at converged HJs. The processes depicted in panels B and C resolve intermediates without an associated crossover by HJ dissolution.