Rnr1, but not Rnr3, facilitates the sustained telomerase-dependent elongation of telomeres

Abstract

Ribonucleotide reductase (RNR) provides the precursors for the generation of dNTPs, which are required for DNA synthesis and repair. Here, we investigated the function of the major RNR subunits Rnr1 and Rnr3 in telomere elongation in budding yeast. We show that Rnr1 is essential for the sustained elongation of short telomeres by telomerase. In the absence of Rnr1, cells harbor very short, but functional, telomeres, which cannot become elongated by increased telomerase activity or by tethering of telomerase to telomeres. Furthermore, we demonstrate that Rnr1 function is critical to prevent an early onset of replicative senescence and premature survivor formation in telomerase-negative cells but dispensable for telomere elongation by Homology-Directed-Repair. Our results suggest that telomerase has a "basal activity" mode that is sufficient to compensate for the "end-replication-problem" and does not require the presence of Rnr1 and a different "sustained activity" mode necessary for the elongation of short telomeres, which requires an upregulation of dNTP levels and dGTP ratios specifically through Rnr1 function. By analyzing telomere length and dNTP levels in different mutants showing changes in RNR complex composition and activity we provide evidence that the Mec1ATR checkpoint protein promotes telomere elongation by increasing both dNTP levels and dGTP ratios through Rnr1 upregulation in a mechanism that cannot be replaced by its homolog Rnr3.

Fig 1. The loss of Rnr1 causes short telomeres that are stably maintained.

A) Mec1^ATR regulates RNR activity and abundance via its downstream effector kinases Rad53 and Dun1. Dun1 phosphorylates Sml1 and releases its inhibitory effect on Rnr1 activity. Upon phosphorylation by Dun1, the repressive effect of Crt1 on Rnr2/3/4 transcription is abolished. The major RNR complex present in S-phase of the cell cycle consists of a tetramer containing two large Rnr1 subunits. Rnr3 is usually expressed at low levels but is upregulated in response to DNA damage. Sequencing of the CRT1 gene identified a premature stop codon that overcame the slow growth defect of rnr1Δ mutants, indicating that an RNR complex containing only Rnr3 as large subunits can compensate for the loss of Rnr1 in terms of cell viability. Tetrad analysis of spores derived from RNR1/rnr1Δ CRT1/crt1 (left) and RNR1/rnr1Δ CRT1/crt1Δ (right) heterozygous diploids indicate that the crt1 suppressor mutation behaves as a single Mendelian trait and phenocopies the full deletion of CRT1 in the suppression of impaired growth. For each heterozygous diploid four tetrads (1–4) have been analyzed by dissection. B) Telomere length analysis by Southern blotting. Following meiotic segregation heterozygous diploid RNR1/rnr1Δ CRT1/crt1Δ mutants have been dissected and the ability to maintain Y´ telomeres has been analyzed in the indicated spore colonies (2 biological replicates over 2 serial dilutions in liquid culture; approximately 30 and 35 generations after meiotic segregation). The telomere shortening of rnr1Δ crt1Δ double mutants resembles the one of rnr1Δ single mutants. C) dNTP pools and dGTP ratios (in %) were measured in the indicated strains. rnr1Δ mutants show lower dNTP levels and dGTP ratios when compared to wild type cells. CRT1 mutation results in an elevation of dNTP levels of rnr1Δ mutants but dGTP levels remain below wild type levels. Mean values +/-SEM are indicated. All values are shown as fold change over the individual dNTP levels of the wild type (set to 1). n(wild type) = 9; n(rnr1Δ crt1Δ) = 3; n(rnr1Δ) = 3; n(crt1Δ) = 2. D) Telomere length analysis by Southern blotting showing that rnr1Δ crt1Δ mutant telomeres are short but stably maintained. The plasmid-born expression of wild type Rnr1 for 200 generations restored telomere length of rnr1Δ crt1Δ mutants to a large extent. The expression of the catalytic inactive Rnr1-C428A did not affect telomere length in rnr1Δ mutants. Represented are biological replicates of the indicated strains.

Fig 2. The Mec1^ATR–checkpoint promotes telomere elongation by regulating Rnr1 through Sml1.

A) Sml1 inhibits Rnr1 but not Rnr3 activity. dNTP pools and dGTP ratios (in %) were measured in the indicated strains. The co-deletion of SML1 upregulates dNTP synthesis in rnr3Δ crt1Δ double mutants but not in rnr1Δ crt1Δ mutants. The co-deletion of CRT1 increases dNTP synthesis in rnr1Δ but not substantially in rnr3Δ mutants. Mean values +/-SEM are indicated. All values are shown as fold change over the individual dNTP levels of the wild type (set to 1). The dNTP levels of rnr1Δ and rnr1Δ crt1Δ mutants were taken from the experiment in Fig 1C and are shown as a comparison. n(wt) = 2; n(rnr3Δ) = 2; n(rnr3Δ crt1Δ) = 2; n(rnr3Δ crt1Δ sml1Δ) = 2; n(rnr1Δ) = 3; n(rnr1Δ crt1Δ) = 3; n(rnr1Δ crt1Δ sml1Δ) = 2; n(crt1Δ) = 2; n(crt1Δ sml1Δ) = 2. B) Telomere length analysis by Southern blotting. Following meiotic segregation heterozygous diploid MEC1/mec1Δ SML1/sml1Δ RNR1/rnr1Δ CRT1/crt1Δ mutants were dissected and the ability to maintain telomeres was analyzed in the indicated spore colonies over 3 serial dilutions in liquid culture (generations 30G-40G). mec1Δ crt1Δ mutants show a short telomere length phenotype that resembles the one of rnr1Δ crt1Δ mutants. The short telomeres of mec1Δ crt1Δ could be rescued by a co-deletion of SML1. C) dNTP pools and dGTP ratios (in %) were measured in the indicated strains. mec1Δ sml1Δ, mec1Δ sml1Δ crt1Δ and mec1Δ crt1Δ mutants show dNTP levels above wild type. Mean values +/-SEM are indicated. All values are shown as fold change over the individual dNTP levels of the wild type which were taken from the experiment in (A) (set to 1). n(wt) = 2; n(mec1Δ sml1Δ) = 2; n(mec1Δ sml1Δ crt1Δ) = 2; n(mec1Δ crt1Δ) = 2. D) Telomere length has been analyzed as a function of dGTP ratios (dGTP levels relative to overall dNTP levels). An R-square test has been performed to investigate the goodness of fit. n(dNTPs/length): wt (9/3), rnr1Δ (3/2), crt1Δ (2/2), rnr1Δ crt1Δ (3/3), sml1Δ crt1Δ (2/1), mec1Δ sml1Δ (2/3), mec1Δ crt1Δ (2/3), mec1Δ sml1Δ crt1Δ (2/3). E) The Mec1^ATR checkpoint pathway promotes telomere elongation specifically via its regulation of Rnr1 activity through Sml1 degradation. High Rnr1 activity promotes telomere maintenance by keeping high cellular dGTP ratios. Shortening of telomere length correlates with Rnr1 inhibition or inactivity and therefore lower dGTP ratios that are produced by Rnr3.

Fig 3. The loss of Rnr1 activity interferes with the maintenance of short telomeres and accelerates replicative senescence and survivor formation in telomerase-negative cells.

A) Experimental approach: Following meiotic segregation and tetrad dissection spore colonies were diluted in liquid medium to an OD₆₀₀ value of 0.02. After every 24h of growth at 30°C the optical cell density has been measured and cells were re-diluted to an OD₆₀₀ of 0.02. The number of cell divisions/generations was calculated as log₂ (OD₆₀₀24h/0.02). B) Senescence curves reflecting the cell density reached after 24 hours of growth as a function of cell divisions (generations). In rnr1Δ crt1Δ est2Δ mutants the onset of replicative senescence and survivor formation is accelerated when compared to crt1Δ est2Δ and est2Δ controls. The number of cell divisions the spore colony went through before the first dilution has been estimated as 25 generations (starting point of the curves). Data is shown as mean +/- SEM (n = 3–5). C) Telomere length analysis by Southern blotting of one biological replicate of each genotype shown in (B). Telomeres of est2Δ and crt1Δ est2Δ controls shortened progressively with ongoing generations (G) and showed a heterogeneous length indicative of survivor formation at 65 and 62 cell divisions, respectively. The telomeres of rnr1Δ crt1Δ est2Δ mutants could not be maintained after the first dilution and showed a heterogeneous length already at generation 40.

Fig 4. In the absence of Rnr1 telomeres cannot become elongated by increased telomerase activity.

A) Telomere length analysis by Southern blotting shows that mutations of the long TLM genes PIF1 or ELG1 cause telomere elongation when introduced in a wild type background but do not affect the short telomere length of rnr1Δ crt1Δ mutants. Telomere blots were performed approximately 200 generations after the long tlm mutation has been introduced. Represented are biological replicates of the indicated strains. B) Telomere length analysis by Southern blotting of cells expressing Cdc13-Est1 and Cdc13-Est2. Expression of the fused proteins results in telomere elongation in wild type cells and crt1Δ mutants but not in rnr1Δ crt1Δ mutants or rnr1Δ crt1Δ elg1Δ mutants. A plasmid expressing Cdc13 was used as a control and did not cause elongation of wild type cell telomeres. Represented are biological replicates of the indicated strains approximately 200 generations after transformation of the indicated plasmid.

Fig 5. The Mec1^ATR-Dun1 pathway promotes telomere over-elongation via Rnr1 activation through Sml1 degradation.

A) dNTP pools and dGTP ratios (in %) were measured in the indicated strains. dun1Δ mutants show low dNTP levels, which can be elevated above wild type levels by a co-deletion of SML1. In sml1Δ elg1Δ dun1Δ mutants dNTP levels are elevated approximately 4 fold relative to wild type. Mean values +/-SEM are indicated. All values are shown as fold change over the individual dNTP levels of the wild type which were taken from the experiment in Fig 1C (set to 1). n(wt) = 9; n(elg1Δ) = 2; n(sml1Δ) = 2; n(sml1Δ elg1Δ) = 2; n(dun1Δ) = 2; n(sml1Δ dun1Δ) = 2; n (dun1Δ elg1Δ) = 2; n(sml1Δ dun1Δ elg1Δ) = 2. B) Telomere length analysis by Southern blotting. A deletion of DUN1 results in short telomeres, which only slightly become elongated by a co-deletion of the long TLM gene ELG1. The simultaneous deletion of SML1 restores telomere length in dun1Δ mutants and rescues the over-elongation of telomeres in dun1Δ elg1Δ double mutants. Represented are biological replicates of the indicated strains after approximately 200 generations. C) dNTP pools and dGTP ratios (in %) were measured in the indicated strains. dun1Δ mutants show low dNTP levels, which can be elevated above wild type levels by a deletion of CRT1 or simultaneous deletions of CRT1 and ELG1. Mean values +/-SEM are indicated. All values are shown as fold change over the individual dNTP levels of the wild type which were taken from the experiment in Fig 1C (set to 1). The dNTP levels of elg1Δ and dun1Δ mutants were taken from the experiment in (A) and are shown as a comparison. n(wt) = 9; n(elg1Δ) = 2; n(elg1Δ crt1Δ) = 2; n(crt1Δ) = 2; n(dun1Δ) = 2; n(dun1Δ crt1Δ) = 2; n(dun1Δ elg1Δ) = 2; n(dun1Δ elg1Δ crt1Δ) = 2. D) Telomere length analysis by Southern blotting. A co-deletion of CRT1 does not restore telomere length in dun1Δ mutants nor does it rescue the over-elongation of telomeres in dun1Δ elg1Δ double mutants. Represented are biological replicates of the indicated strains after approximately 200 generations.

Fig 6. Model: Rnr1 but not Rnr3 provides telomerase with sustained activity.

Rnr1 is essential for sustained telomerase-dependent elongation of telomeres. In the absence of Rnr1 (1) telomeres shorten and the expression of the alternative large subunit, Rnr3, only allows a basal activity of telomerase due to reduced dGTP ratios. This activity is sufficient to compensate for the "end replication problem" but not for re-elongation of the short telomeres even when long TLM genes become mutated (2). In wild type cells Mec1^ATR upregulates Rnr1 activity through Sml1 degradation and therefore ensures that telomeres are maintained at a wild type length by keeping high dGTP ratios and supplying dNTPs (3). Alike, in long tlm mutants telomere over-elongation is promoted by Rnr1 activation through Mec1^ATR (4). In both cases Rnr3 expression cannot compensate for this function of Rnr1 suggesting that Rnr1-produced dNTPs are provided in a local or temporal manner to promote telomere elongation. Stalled replication forks at telomeres (not displayed) might promote telomerase activity by triggering the S-phase checkpoint, thus maintaining high Rnr1 activity when telomeres become replicated.