Recombination-mediated lengthening of terminal telomeric repeats requires the Sgs1 DNA helicase

Abstract

The Saccharomyces cerevisiae SGS1 gene encodes a RecQ-like DNA helicase, human homologues of which are implicated in the genetic instability disorders, Bloom syndrome (BS), Rothmund-Thomson syndrome (RTS), and Werner syndrome (WS). Telomerase-negative yeast cells can recover from senescence via two recombinational telomere elongation pathways. The "type I" pathway generates telomeres with large blocks of telomeric and subtelomeric sequences and short terminal repeat tracts. The "type II" pathway generates telomeres with extremely long heterogeneous terminal repeat tracts, reminiscent of the long telomeres observed in telomerase-deficient human tumors and tumor-derived cell lines. Here, we report that telomerase-negative (est2) yeast cells lacking SGS1 senesced more rapidly, experienced a higher rate of telomere erosion, and were delayed in the generation of survivors. The est2 sgs1 survivors that were generated grew poorly, arrested in G(2)/M and possessed exclusively type I telomeres, implying that SGS1 is critical for the type II pathway. The mouse WS gene suppressed the slow growth and G(2)/M arrest phenotype of est2 sgs1 survivors, arguing that the telomeric function of SGS1 is conserved. Reintroduction of SGS1 into est2 sgs1 survivors restored growth rate and extended terminal tracts by approximately 300 bp. Both phenotypes were absolutely dependent on Sgs1 helicase activity. Introduction of an sgs1 carboxyl-terminal truncation allele with helicase activity restored growth rate without extending telomeres in most cases, demonstrating that type II telomeres are not necessary for normal growth in the absence of telomerase.

Figure 1

A two plasmid Cre-loxP system for the inducible deletion of EST2. (A) One plasmid carries full-length EST2 gene and ADE2, a marker gene required for adenine biosynthesis. In addition, two loxP sites flank the ARS/CEN cassette required for plasmid replication and efficient segregation during mitosis. A second URA3-based plasmid contains the Cre recombinase gene under the control of the GAL10 (galactose-inducible) promoter. (B) Growth of strains on galactose-containing medium resulted in Cre-mediated excision of the ARS/CEN sequence and the loss of EST2 from the strain, as indicated by the conversion of the strains from Ade⁺ to Ade⁻ and the inability of cells to grow on medium lacking adenine.

Figure 2

Senescent phenotypes and telomere loss rates of sgs1, est2, and rad52 mutant combinations. (A) Five cultures of est2 (●), est2 sgs1 (□), est2 rad52 (■), and est2 sgs1 rad52 (▿) and two cultures of wild type (▵) carrying the Cre-loxP EST2 recombination system were pregrown in glucose-containing medium and then serially passaged each day in galactose/raffinose-containing medium. Cultures were inoculated to 1 × 10⁵ cells/ml and grown until the wild-type culture reached a cell density of 1.4 × 10⁸/ml (≈10 generations of the wild-type strain/passage). We estimated that the strains had undergone ≈30 generations before day 1. To determine the rates of telomere loss, glucose-grown cells (generation 0) were resuspended in galactose/raffinose-containing medium and harvested for telomere analysis exactly 4 and 11 generations later (B and C, respectively). XhoI-digested genomic DNA was probed with a Y′ sequence after Southern blotting. Lanes 1–3, est2; lanes 4–6, est2 sgs1; and lanes 7–9, est2 sgs1 rad52. (D) Mean terminal fragment lengths of est2 (●), est2 sgs1 (□), est2 rad52 (■), and est2 sgs1 rad52 (▿) strains after growth in galactose/raffinose-containing medium. For each strain, nine independent cultures were assayed, and each culture was assayed by using three independent blots. Mean terminal fragment lengths (±SD) are shown relative to the glucose-grown control (0 generations).

Figure 3

Analysis of SGS1 function during senescence and recovery. Seventeen spore colonies of est2 (●), est2 sgs1 (□), est2 rad52 (■), and est2 sgs1 rad52 (▿) strains and two cultures of wild type (▵) were subcultured for 13 days. Each culture was inoculated to 1 × 10⁵ cells/ml and grown until the wild-type culture reached a cell density of 4 × 10⁸/ml. Each passage represented 11 generations of the wild-type strain. We estimated that strains had undergone ≈40 generations before day 1. (A) Average daily cell densities for each genotype. Error bars represent standard deviations. (B) Typical time course for individual cultures showing the typical loss of viability of est2 sgs1 survivors. (C) Representative flow cytometric analyses of wild-type, sgs1, est2, and est2 sgs1 cultures, 1 and 11 days after survivors were generated (mid- and post-recovery, respectively). Percentage of cells with a G₁ or G₂ DNA content are shown. (D) Poly[AC/TG]-probed Southern blot XhoI-digested genomic DNA from est2 (lanes 2–5), est2 sgs1 survivors (lanes 7–10), and wild-type (lanes 1 and 6) cultures on day 13. (E) Equivalent blot as for D but probed for Y′ sequences. (F) Y′-probed Southern blot of XhoI-digested genomic DNA from est2 and est2 sgs1 survivors 1, 2, 3, and 4 days after survivors were first generated. Lane 1, wild type; lanes 2–5, est2; and lanes 6–9, est2 sgs1.

Figure 4

Murine WRN can suppress telomere-related phenotypes of est2 sgs1 strains. (A) The average rate of senescence and recovery for six wild-type (▵) est2 (●), est2 sgs1(□), and est2 sgs1 mWRN (⋄) cultures. (B) Survivors from A passaged in liquid complete (YPD) medium for 10 days after survivors were generated. Cells were then spotted to YPD plates in serial 10-fold dilutions and grown for 24 h. (C) Flow cytometric analysis of strains in B. (D) Representative telomere structures of mWRN-expressing survivors in B. Y′-probed Southern blot of XhoI-digested genomic DNA isolated from strains. Lane 1, wild type; lane 2, est2 survivor; lane 3, est2 sgs1 survivor; and lanes 4 and 5, representative est2 sgs1 mWRN survivors.

Figure 5

Reintroduction of SGS1 into est2 sgs1 survivors restores growth rate and lengthens C_1–3A/TG_1–3 repeat tracts. The sgs1-hd (helicase-defective) allele carries a single amino acid substitution in the ATP binding domain (3). The sgs1-ct (Δ200) allele is a carboxyl-terminal truncation that removes the conserved RecQ carboxyl-terminal (ct) and RNaseD homology (HRD) domains (3). (A) Low copy plasmids carrying the SGS1, sgs1-hd, or sgs1-ct allele, or no insert, were introduced by transformation into est2 sgs1 survivors, 1 day after they were generated (indicated by arrow). Five random transformants were resuspended in medium lacking leucine and serially passaged each day between 1 × 10⁵ cells/ml and 1.4 × 10⁸ cells/ml. Each passage represented 10 generations of the wild-type strain. (B) Telomere structure analysis of strains in A on day 8. XhoI-digested genomic DNA was probed with a poly[AC/TG] sequence. Identical terminal band sizes were detected by probing blots for Y′ sequences (not shown). Lane 1, wild type; lane 2, est2 survivor; lane 3, est2 sgs1 survivor; lanes 4–6, vector only; lanes 7–9, SGS1; lanes 10–12, sgs1-hd; and lanes 13–15, sgs1-ct. (C) Telomere structure analysis of two independently derived est2 sgs1 type II survivors after the loss of plasmid-borne SGS1 (pSGS1). Lanes 1–3, survivor 1; and lanes 4–6, survivor 2. Lanes 1 and 4, telomeres before loss of pSGS1; lanes 2 and 5, 250 generations with selection for pSGS1; and lanes 3 and 6, 250 generations after loss of pSGS1. Genomic DNA was probed with a poly[AC/TG] sequence as in B.