Dear Editor,

Telomeres are highly organized DNA-protein structures at the ends of linear eukaryotic chromosomes 1. In budding yeast Saccharomyces cerevisiae, each chromosome end has TG1–3/C1–3A double-stranded telomeric DNA and a G-rich single-stranded tail 2. The replication of telomere involves both telomerase and recombination pathways 3, 4. When telomerase is absent, telomeric DNA will progressively shorten, and the colony will eventually undergo senescence 4. Some survivors will escape senescence by maintaining telomeres through a Rad52-dependent homologous recombination mechanism 3. Our previous work had identified a conserved non-OB-fold single-stranded telomeric (ssTG) DNA binding protein Sua5p, which facilitates telomerase activity 5. Here we report that Sua5p is also involved in the telomeric DNA recombination in telomerase-negative cells, and type II survivors were completely diminished in the SUA5 deletion telomerase-negative cells. These two distinct functions of Sua5p are independent of each other, but both require its DNA-binding ability.

We performed a genetic interaction analysis of SUA5 and other telomere regulation genes including Rad52 epistasis group genes, in which RAD genes were replaced by a KanMX4 cassette (conferring G418 resistance to the cells). Notably, sua5 deletion mutants grew relatively slowly and formed small colonies compared to the wild-type colonies (Figure 1A). Failure of recovering small colonies on YPD-G418 plates indicated that sua5Δ and rad52Δ are synthetically lethal (Figure 1A). However, double deletion of other genes in the Rad52 epistasis group (i.e. RAD50, RAD51, RAD57 and RAD59) with SUA5 did not affect cell viability (Figure 1A left and Supplementary information, Figure S1A), supporting the argument that the genetic interaction of SUA5 and RAD52 was not a “sick” phenotype. In addition, a SUA5 CEN plasmid, containing the Sua5 ORF and Sua5 endogenous promoter and terminator (Supplementary information, Figure S1B), can rescue the synthetic lethality of rad52Δsua5Δ (Figure 1A, right). These data suggested that Sua5 may affect recombination pathway(s), e.g. telomere recombination.

Figure 1
figure 1

(A) Left, Tetrad dissection of RAD/radΔ SUA5/sua5Δ strains on YPD and YPD G418 plates. Rad52 epistasis group genes were replaced with KanMX4 maker, conferring G418 resistance, and SUA5 deletion cells grew slowly and formed small colonies. Eight asci labeled as 1 to 8 are shown for each genotype. No small colony was seen in the tetrad dissection result of RAD52/rad52Δ SUA5/sua5Δ strains on YPD G418 plate. Right, Tetrad dissection of RAD52/rad52Δ SUA5/sua5Δ strain, which harbored a plasmid-borne SUA5 gene. The genotypes were confirmed by auxotrophic makers. The rad52Δsua5ΔpSUA5 clones were indicated by open triangles. Eight asci labeled as 1 to 8 are shown. (B) Deletion of SUA5 eliminates type II survivor production. The survivor types of isogenic strains were determined by telomere southern blot using XhoI digested genomic DNAs and hybridized with a TG probe. Individual strains were continuously passaged in YPD liquid medium until survivors arose. Two independent clones of survivors are shown for each telomerase-deficient strain. The cdc13-2 mutants were constructed in the YPH499 background, and the corresponding wild-type and sua5Δ controls in YPH499 were shown in the right panel. The left panel shows strains derived from BY4743. Telomere lengths of first- and sixteenth-passage cultures are shown for wild-type and sua5Δ strains. “+” and “−” indicate wild-type SUA5 and sua5Δ, respectively. (C) Senescence assay of deletion mutants lacking telomere recombination genes. Serial passages were initiated with O.D600 ∼0.05, and the cell numbers were measured by spectrometry. The X-axis is normalized to the passaged times as sua5Δ grew slowly. No survivor was recovered in the triple-deletion strain of est2Δsua5Δrad51Δ after the seventh subculturing. (D) Type II survivor switches to type I survivor upon deletion of SUA5. Sibling spores of SUA5/sua5Δ type II cells were restreaked on the YPD plate for 16 times. The genomic DNAs were digested with XhoI, and telomeres were examined by southern blot with a TG probe. Restreak times are labeled on the top of the panel. The Y' band is indicated on the right. (E) Loss-of-DNA-binding mutants (K93A, R95A, D98K, R248A and R114D) cannot produce type II survivor cells. The genomic DNAs were digested with XhoI, and telomeres were examined by southern blot with a TG probe. (F) Deletion of SUA5 in telomere-recombination deficient cells further shortens telomere length. Telomere lengths of single- and double-deletion mutants of RAD51, RAD57, RAD50 or RAD59 with SUA5 are shown. “+” and “−” indicate wild-type SUA5 and sua5Δ, respectively. The genomic DNAs were digested with XhoI, and telomeres were examined by Southern blot with a TG probe.

Full size image

In budding yeast, telomerase-deficient cells will undergo senescence after approximately 120 population doublings 4. Most of the telomerase-deficient cells are subjected to cell cycle arrest at G2/M phase and cell death 4. Rare survivors, which use homologous recombination to maintain telomere, could recover at a low frequency 4. Type I survivors have tandem arrays of the subtelomeric Y' element, whereas type II survivors have very long terminal tracts of C1–3A/TG1–3 DNA. As Rad52 epistasis group genes are required for telomere recombination maintenance, we wondered whether Sua5 had any role in survivor production. The estΔ and est Δsua5Δ cells were continuously cultured in liquid medium until the survivor arising. Double deletions of SUA5 and EST genes showed similar senescence rates with the EST single-deletion mutants (Supplementary information, Figure S2). As Type II survivors grow much faster than Type I survivors in liquid medium, the est2Δ post-senescent survivor cells under this condition were typical Type II according to the telomere pattern (Figure 1B). Interestingly, the est2Δsua5Δ post-senescent survivor cells in liquid medium were Type I with amplified Y' subtelomeric elements (Figure 1B). The same results were observed in est1Δsua5Δ, est3Δsua5Δ and tlc1Δsua5Δ post-senescent cells (Figure 1B, left), and cdc13-2sua5Δ post-senescent cells of YPH499 background (Figure 1B, right). To confirm the result of liquid senescence assay, we continuously passaged the est2Δsua5Δ cells on solid medium. The single-colony survivors were picked at the fifth restreak and subjected to telomere southern blot. Consistent with the result of liquid senescence assay, all the 58 independent est2Δsua5Δ post-senescent clones were Type I, while the est2Δ post-senescent clones showed a significant fraction of Type II survivors (Supplementary information, Figure S3).

The high-frequency appearance of Type I survivor could be attributed to an elevated rate of Type I survivor formation or a diminished rate of Type II survivor production in sua5Δ cells. In the Rad52 epistasis group, RAD51 and RAD57 are essential for Type I survivor formation, RAD50 and RAD59 are essential for Type II survivor production, and RAD52 is required for both types arising 6. Survivors are almost eliminated in rad50Δrad51 Δest2Δ triple mutant cells similar to rad52Δest2Δ cells 7. If SUA5 increases Type I survivor production, rad51Δ sua5Δ est2 Δ or rad57Δ sua5Δ est2Δ cells will produce Type II survivors. On the other hand, if SUA5 is essential for Type II survivor production, rad51Δ sua5Δ est2Δ or rad57Δ sua5Δ est2Δ cells will not produce survivors. The isogenic spores were dissected and subjected to senescence assay. The rad51Δ sua5Δ est2Δ and rad57Δ sua5Δest2Δ triple mutant cells underwent senescence and no survivors were recovered (Figure 1C and Supplementary information, Figure S4). This result suggests that Sua5 is required for Type II survivor formation in post-senescent cells.

To further address whether Sua5 is essential for the maintenance of Type II survivors, we deleted SUA5 in diploid Type II survivor cells (est2Δ::LEU2/est2Δ::URA3). The sua5Δ haploid Type II survivor cells were obtained after tetrad dissection. Continuous passages of SUA5 Type II survival cells resulted in gradual telomere shortening (Figure 1D) 8. Notably, the sua5Δ Type II survival cells exhibited accelerated telomere shortening and extensive Y' amplification (Figure 1D and Supplementary information, Figure S5). The telomere pattern of the 16th-restreaked sua5Δ Type II cells (about 400 generations after sporulation) is similar to that of Type I cells (shown by XhoI digestion of Figure 1D, and AluI-HaeIII-HinfI-MspI digestion of Supplementary information, Figure S5), indicating that SUA5 deletion abolishes the telomere-telomere recombination in the Type II survivors, and the post-senescent cells have to elongate their telomeres through Y'-element recombination.

Our previous data have shown that Sua5p is a single-stranded telomeric DNA binding protein, and mutations at its DNA-binding region affect its function at telomeres 5. We then asked whether the recombinational role of Sua5p also requires its DNA-binding activity. Mutant sua5 alleles that are defective in DNA-binding were transformed into the pre-senescent telomerase-deficient cells, and the cells were continuously passaged until survivors arose. The cells harboring a loss-of-DNA-binding sua5 allele could only produce Type I survivors in liquid-medium senescence assay (Figure 1E), indicating that the function of Sua5p in recombinational replication also requires its telomeric DNA-binding ability. Our earlier report 5 and the current study showed that Sua5 affected telomerase activity in the wild-type cells and recombination-dependent telomere elongation in telomerase-deficient cells. It was interesting to know whether the telomere shortening phenotype in sua5Δ telomerase-positive cells was also caused by the telomere recombination defect. In rad51Δ, rad57Δ, rad50Δ or rad59Δ mutant cells, which were defective in either sub-telomere or telomere-telomere recombination and could not produce certain type of survivors, SUA5 deletion further shortened the telomeres (Figure 1F), suggesting that SUA5 functions independently in telomerase- and recombination-dependent telomere maintenance pathways.

In conclusion, our study on Sua5 function showed that ssTG-binding protein Sua5p also plays a role in telomere recombination in the post-senescent telomerase-negative cells, and its ssTG-binding activity is indispensable for these functions.

(Experimental materials and methods are depicted in the Supplementary information, Data S1)