Essential Regions of Saccharomyces cerevisiae Telomerase RNA: Separate Elements for Est1p and Est2p Interaction

Abstract

The Saccharomyces cerevisiae telomerase RNA subunit is encoded by the TLC1 gene. A selection for viable alleles of TLC1 RNA from a large library of random deletion alleles revealed that less than half (∼0.5 kb of the ∼1.3-kb RNA) is required for telomerase function in vivo. The main essential region (430 nucleotides), which contains the template for telomeric DNA synthesis, was required for coimmunoprecipitation with Est1p and Est2p. Furthermore, the subregion required for interaction with Est1p, the telomerase recruitment subunit, differed from those required for interaction with Est2p, the reverse transcriptase subunit. Two regions of the RNA distant from the template in the nucleotide sequence were required for Est2p binding, but the template itself was not. Having the RNA secured to the protein away from the template is proposed to facilitate the translocation of the RNA template through the active site. More generally, our results support a role for the telomerase RNA serving as a scaffold for binding key protein subunits.

Telomerase is a ribonucleoprotein enzyme that catalyzes the addition of telomeric DNA sequence to chromosome ends. Telomerase minimally contains a catalytic subunit, which is a reverse transcriptase (19), and an RNA subunit that provides the template for telomeric DNA addition (9). In the budding yeast Saccharomyces cerevisiae, the telomerase RNA subunit is approximately 1.3 kb and is encoded by the TLC1 gene (38). The function of less than 80 nucleotides (nt) of TLC1 RNA has been reported. The goals of the present study are to determine the regions of TLC1 RNA that are required for in vivo telomerase function and to then use this information to facilitate the study of RNA-protein interactions in this complex.

The S. cerevisiae telomerase catalytic subunit, encoded by the EST2 gene, and TLC1 RNA comprise the catalytic core of the enzyme and are required for activity in vitro (6, 19). Additional proteins encoded by EST1, EST3, and CDC13 are required in vivo for telomere maintenance. Mutations in EST2, TLC1, EST1, EST3, and CDC13 can lead to an est (ever shorter telomeres) phenotype—telomeres progressively shorten, and the cells eventually senesce (14, 21). Est1p and Est3p are telomerase subunits (11), while Cdc13p binds single-stranded telomeric DNA and appears to be a component of the telomere itself (12, 16, 25). Est1p also binds single-stranded G-rich telomeric DNA in vitro (40) and interacts with Cdc13p by two-hybrid analysis (29). The Est1p-Cdc13p interaction recruits telomerase to the telomere (7, 26).

TLC1 RNA has been proposed to bind Est1p via an RNA recognition motif (44). TLC1 RNA contains a 48-nt stem-loop that interacts either directly or indirectly with the DNA end-binding protein Ku (27). In addition, TLC1 RNA contains an Sm binding site near its 3′ end that is important for the stability of the RNA and is bound by Sm proteins, therefore leading to the classification of S. cerevisiae telomerase as an Sm snRNP (small nuclear ribonucleoprotein particle) (35). Because TLC1 RNA interacts with multiple proteins, it may serve as a scaffold for the formation of the telomerase RNP.

At ∼1.3 kb, S. cerevisiae telomerase RNA is about twice the size of the average vertebrate telomerase RNA (4) and about five to seven times the size of the average ciliate telomerase RNA (18, 23). The secondary structures of vertebrate and ciliate telomerase RNAs are known (4, 30, 43); however, the structure of the yeast TLC1 RNA has not been determined.

It has previously been shown in studies with S. cerevisiae that some snRNAs that are much larger than their mammalian counterparts can have large regions deleted without affecting the function of the RNA. For example, 945 nt that separate two conserved domains of the 1,175-nt U2 snRNA of S. cerevisiae can be deleted without having any effect on growth (36). Furthermore, in another budding yeast, Kluyveromyces lactis, approximately half of the telomerase RNA, TER1, is dispensable for in vivo telomerase function (32).

We reasoned that large regions of the telomerase RNA from S. cerevisiae might be nonessential, thus simplifying the problem of determining structure-function relationships. We conducted an in vivo selection for viable tlc1 deletion alleles and found that approximately 40% of TLC1 RNA is required to provide sufficient telomerase function to sustain growth. We also were able to attribute functions to a subset of the essential regions: interaction with telomerase subunits Est1p and Est2p. Determinants for interaction with Est1p were distinct from determinants for Est2p interaction, supporting the notion that these proteins can bind TLC1 RNA independently.

MATERIALS AND METHODS

Strains and plasmids.

Yeast extract-peptone-dextrose (YPD) and synthetic liquid and solid media used for cell growth were as described previously (31). Yeast strain TCY43 (mata ura3-52 lys2-801 ade2-101 trp1-Δ1 his3-Δ200 leu2-Δ1 VR::ADE2-TEL Δppr1::HIS3 adh4::URA3-TEL Δtlc1::LEU2 rad52::LEU2, pl[TLC1 LYS2 CEN]) (3) was used in the selection for viable tlc1 deletion alleles. AVL78 (MATa leu2 trp1 ura3-52 prb prc pep4-3) (17) is the parent strain for the strains used in the coimmunoprecipitation studies: TVL288 (HA₃-EST1) (11) and YAL102 (ProA-EST2), which was generated by digesting pKF409 (8) with ClaI and integrating into strain AVL78 by two-step gene replacement (33). Strains HA₃-EST1 tlc1::LEU2 and ProA-EST2 tlc1::LEU2 have nt 128 to 1140 of TLC1 replaced with LEU2 and were maintained with pRS316/TLC1, which contains the BamHI/EcoRI fragment of pSD107[CEN TRP1 TLC1] (3) subcloned into the same restriction sites in the polylinker of pRS316 (37). The tlc1 deletion alleles used in the coimmunoprecipitation assay included alleles from the tlc1 deletion library as well as alleles generated by Kunkel mutagenesis (13) or oligonucleotide-mediated site-directed mutagenesis using PCR (10). Alleles generated by either type of mutagenesis were digested with NcoI and NsiI and subcloned into the same restriction sites of pB7[CEN TRP1 tlc1ΔNcoI-NsiI].

Deletion library construction.

A deletion library of the TLC1 gene from S. cerevisiae was constructed as follows: pUC19 was cleaved with restriction endonucleases AatII and AflIII, removing an 875-bp fragment. Oligonucleotides 5′-CATAGATCTAAACCCGGGAGGACTATAAACCACCAAATGGGTA-3′and 5′-CATGTACCCATTTGGTGGTTTATAGTCCTCCCGGGTTTAGATCTATGACGT-3′ were synthesized and annealed, providing overhangs which were used to ligate the oligonucleotide into gel-purified AatII-AflIII-cut pUC19; the resulting plasmid was named pUCΔAatII-AflIII. BglII, PflMI, and SmaI sites in the oligonucleotide insert facilitated subsequent cloning and screening. The BglII-PflMI fragment containing the TLC1 gene from pSD107 was inserted into the BglII-PflMI sites of pUCΔAatII-AflIII, resulting in the plasmid pUCΔAatII-AflIII-TLC1.

Supercoiled pUCΔAatII-AflIII-TLC1 (50 μg) was treated with 0.035 U of RQ1 RNase-free DNase I (Promega) at 37°C in 0.05 M Tris-HCl (pH 7.5)-0.01 M MnCl₂-0.1 mg of bovine serum albumin (BSA) per ml in a total volume of 100 μl. In the presence of Mn²⁺, DNase I cleaves both strands of duplex DNA to generate molecules that are blunt-ended or have protruding termini of just 1 or 2 nt. Aliquots were removed at 30-s intervals, and the DNase I was stopped by adding EDTA to a final concentration of 10 mM. The extent of digestion was monitored by agarose gel electrophoresis, and the samples containing linearized plasmid were pooled. Since exonuclease III requires blunt ends as substrates, linearized plasmid DNA was subsequently incubated for 15 min at 25°C with 4 U of T4 DNA polymerase (New England Biolabs) in 10 mM Tris-HCl (pH 7.9)-50 mM NaCl-10 mM MgCl₂-1 mM dithiothreitol (DTT)-0.5 mM concentrations of each deoxynucleoside triphosphate (dNTP), followed by incubation with 5 U of Klenow DNA polymerase (New England Biolabs) for 15 min at 25°C. Following phenol and chloroform extractions, the linearized DNA was gel purified using a QIAquick gel extraction kit (Qiagen). In a total volume of 60 μl of 10 mM 1,3-bis[tris(hydroxymethyl)methamino]propane-HCl (pH 7.0)-10 mM MgCl₂-1 mM DTT (NEB buffer no. 1), the purified linear DNA was incubated with 100 U of exonuclease III at 37°C; 4.5-μl aliquots were removed every 5 s and added to tubes on ice that contained 13.5 μl of 0.05 M sodium acetate (pH 4.6), 0.28 M NaCl, 4.5 mM ZnCl₂, and 5 U of S1 nuclease (Roche). Following incubation at 30°C for 30 min, 2 μl of 0.33 M Tris-HCl (pH 7.5)-0.5 mM EDTA was added to each tube and incubated at 70°C for 10 min to inactivate the S1 nuclease. The 12 tubes were combined into 2 tubes of 120 μl each. Two units of Klenow DNA polymerase in 10 μl of 0.01 M Tris-HCl (pH 7.5)-0.2 M MgCl₂ was added to each tube and incubated for 5 min at 37°C. Following the addition of 6 μl of 1 mM concentrations of each dNTP, incubation continued at 25°C for an additional 15 min. The Klenow DNA polymerase was inactivated by a 10-min incubation at 70°C. The DNA was ethanol precipitated and then treated with 5 U of DNA ligase (New England Biolabs) in 100 μl of 50 mM Tris-HCl (pH 7.5)-10 mM MgCl₂-10 mM DTT-1 mM ATP-25 μg of BSA per ml at 16°C overnight. Following ethanol precipitation, the ligated DNA was dissolved in 4 μl of H₂O. The ligated DNA (1 μl) was transformed into Epicurian Coli XL1-Blue electroporation-competent Escherichia coli (Stratagene), followed by plating 2.5% of the transformation on Luria-Bertani-AMP plates. The remaining 97.5% was used in a maxi-DNA prep. To transfer to a vector competent to express in S. cerevisiae, 50 μg of the library of tlc1 deletions was cut with BglII and PflMI. Following separation on a 1% agarose gel, the smear of inserts was cut from the gel in two sections and purified using a QIAquick gel extraction kit (Qiagen). The gel-purified inserts were then cloned into the BglII- and PflMI-cut pSD107 ΔNcoI-NsiI vector.

Selection for viable tlc1 deletion alleles.

A single colony of TCY43 was streaked onto plates containing α-aminoadipate to counterselect for the complementing wild-type TLC1 plasmid pL317/Tg2[TLC1 LYS2 CEN] as previously described (31). Following growth for an additional 50 generations on YPD plates to allow telomeres to shorten, a single colony was picked and used to inoculate liquid YPD media and allowed to grow overnight to an optical density at 600 nm of 0.6. The cells were prepared for transformation with lithium acetate as previously described (34). Cells were next transformed with the TLC1 deletion library pSD107[CEN TRP1 tlc1.del.library] and subsequently plated on plates lacking tryptophan to select for alleles of the TLC1 gene competent to rescue senescing yeast.

Mapping deletion alleles of TLC1.

Yeast colonies from the transformation with the deletion library pSD107[CEN TRP1 tlc1.del.library] were isolated and used to inoculate 10-ml cultures of media without tryptophan. After overnight growth at 30°C, cells were harvested and DNA was isolated by using a DNA-Pure yeast genomic kit (CPG, Inc.). The DNA was subsequently used to transform electroporation-competent E. coli. The plasmid was isolated and digested with BglII and PflMI. The DNA fragment containing the tlc1 deletion was separated by gel electrophoresis and isolated by using a QIAquick gel extraction kit (Qiagen) and then digested with TaqI. The resulting fragments were run on a 4% Nusieve agarose gel (FMC Corporation) and visualized by ethidium bromide staining.

Unselected tlc1 deletion alleles were identified as follows. The tlc1 deletion library was used to transform electroporation-competent E. coli. Following growth on Luria-Bertani-AMP plates, colonies were lifted from the plates with flat toothpicks and resuspended in 300 μl of 10 mM Tris-HCl (pH 8.3)-1.5 mM MgCl₂-50 mM KCl-0.2 mM concentrations of each dNTP. Deletions were initially identified by PCR using 100 μl of the resuspended colony, 0.625 μM YTBgl5′R (5′-GGTTACATATAGATCTCAAGGTTCTC-3′), 0.625 μM YT1287 (5′-GGGCTTTATTATAAATATTAAGAGG-3′), and 2.5 U of Taq polymerase (Boehringer Mannheim). Following 30 cycles of three-temperature PCR amplification (1 min at 95°C, 1 min at 50°C, 1 min at 72°C), products were run on a 1% agarose gel and visualized by ethidium bromide staining. In those cases where the amplification product was smaller than the wild-type TLC1 product , the position of the deletion was further characterized by PCR with a set of primers near the 5′ end of the gene (YTBgl5′R, 5′-GGTTACATATAGATCTCAAGGTTCTC-3′, and YT164, 5′-CAATCCGAAATCCGACACTATC-3′) or a set of primers in the central region of the gene (YT415R, 5′-GTACCGATCCTCTTCTCGACCTAACC-3′, and YT885, 5′-GGACACCCTTGCCTTTTGGCTTATTTACC-3′). Tlc1 deletion alleles identified by this method as well as alleles identified by restriction endonuclease mapping were then tested for rescue of senescing yeast as described above. The exact endpoints of the deletions were identified by DNA sequencing.

Extract preparation and immunoprecipitations.

Yeast strains were grown at 30°C in synthetic complete media without tryptophan to an optical density at 600 nm of ∼1. Cells were lysed with glass beads in TMG (10 mM Tris-Cl at pH 8, 1 mM MgCl₂, 10% glycerol) plus 200 mM NaCl, 0.1 mM DTT, and 0.1 mM EDTA. One Complete EDTA-free protease inhibitor tablet (Boehringer) was added for each 50 ml of TMG used in extract preparations or immunoprecipitations. Immunoprecipitation of ProA-Est2p was done using a slightly modified version of a protocol obtained from Friedman and Cech (8). The extract was adjusted to 0.5% Tween 20 and 400 U of RNasin (Promega) per ml and then incubated overnight at 4°C with immunoglobulin G Sepharose beads (Pharmacia) equilibrated with TMG (plus 200 mM NaCl and 0.5% Tween 20). Beads were washed three times with TMG (plus 200 mM NaCl, 0.5% Tween 20, and 0.1 mM DTT), once with TMG (plus 50 mM NaCl and 0.1 mM DTT) and resuspended in TMG (plus 50 mM NaCl, 0.5 mM DTT, and 80 U of RNasin per ml). Immunoprecipitation of HA₃-Est1p was done using a modified version of a protocol from Seto et al. (35). Protein G Plus-Agarose (Oncogene) was equilibrated in phosphate-buffered saline plus 0.1% NP-40 and then incubated overnight at 4°C with HA.11 antibody (Babco), after which it was added to extract plus 400 U of RNasin (Promega) per ml and allowed to bind 5 to 12 h at 4°C. Beads were washed in TMG (plus 200 mM NaCl, 0.1 mM DTT, and 0.1 mM EDTA).

RNA isolation and Northern blots.

RNA was isolated from immunoprecipitation beads by incubation in proteinase K buffer (0.4 mg of proteinase K/ml, 150 mM NaCl, 100 mM Tris-HCl [pH 7.4], 12.5 mM EDTA, 1% sodium dodecyl sulfate [SDS]) at 37°C for 15 min. RNAs were phenol-chloroform extracted followed by chloroform extraction and ethanol precipitation. Isolated RNAs were separated by electrophoresis on 4% polyacrylamide-8 M urea gels and then transferred onto a Hybond N+ membrane (Amersham) with a Hoefer electroblot apparatus at 1.5 A for 2 to 3 h in 0.5× Tris-borate-EDTA (TBE) buffer (0.05 M Tris base, 0.042 M boric acid, 0.0005 M EDTA). RNAs were fixed to the nylon membrane by UV cross-linking (0.12 J, UV Stratalinker; Stratagene). Membranes were prehybridized 0.5 to 1 h at 55°C in 0.5 M NaPO₄ (pH 7.2)-0.001 M EDTA-7% SDS-1% BSA (5). Hybridization was performed in the same buffer at 55 to 65°C overnight with 0.5 × 10⁷ to 1.0 × 10⁷ cpm of probes specific for TLC1 RNA ([³²P]dCTP labeled StuI-NsiI fragment of TLC1) and U1 snRNA (³²P-5′-end-labeled complementary oligonucleotide). Subsequently the membrane was rinsed three times with 50 ml of 0.1 M SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate)-0.1% SDS and then washed three times for 20 min in the same solution at the hybridization temperature.

Analysis of coimmunoprecipitations.

To be used in this coimmunoprecipitation assay, the alleles of interest had to be at levels in the cell sufficient for detection by Northern analysis. Deletion of some regions of TLC1 resulted in little or no RNA detectable by Northern blotting: Δ1002-1167, which contains the Sm site, and Δ795-912 each had no detectable RNA by Northern blotting. For tlc1 deletion alleles sufficiently resolved from the wild-type RNA on Northern analysis, binding was analyzed more quantitatively by comparing the amount of deletion allele bound [Bound(del)] relative to the amount in the input extract [Input(del)] and normalizing this to the same ratio for the wild-type (wt) RNA: coimmunoprecipitation of tlc1 deletion allele relative to wt TLC1 RNA = [Bound(del)/Input(del)]/[Bound(wt)/Input(wt)].

Quantitation of band intensity was done by Phosphorimager analysis using Imagequant software. The binding level relative to that of the wild type could not be directly determined for the tlc1Δ template allele, as it was almost the same length as wild-type TLC1 RNA. Therefore the coimmunoprecipitations were done in a tlc1::LEU2 strain that expressed a plasmid-borne copy of Δ885-1125, which served as a positive control for the coimmunoprecipitations.

RESULTS

Selection for functional deletion alleles of TLC1.

To identify regions of the TLC1 RNA that are nonessential for growth of S. cerevisiae, we constructed random deletion libraries of the TLC1 gene as described in Materials and Methods. One library had deletions ranging in size from 0 to 270 bp, and the second library had deletions of 240 to 400 bp (Fig. 1).

FIG. 1.

Creation of a library of random deletions within the TLC1 gene. Each deletion library and the parent plasmid pSD107 were cut with BglII and PflMI restriction endonucleases to free the TLC1 insert from the 5,974-bp vector. DNA was run on a 1% agarose gel and stained with ethidium bromide. Lane 1, library with deletions ranging from 0 to 270 bp; lane 2, library with deletions ranging from 240 to 410 bp; lane 3, pSD107; lane M, φX174 double-stranded DNA cut with HaeIII; lane M′, phage λ DNA cut with HindIII.

A tlc1Δ::LEU2 disruption strain was utilized to allow identification of complementing tlc1 deletion alleles. This strain also carries a disruption of the RAD52 gene to prevent senescence of the tlc1Δ strain from being suppressed by RAD52-dependent amplification of telomere sequences (14, 20). Cell growth was maintained by wild-type TLC1 carried on a LYS2 plasmid (Fig. 2). Loss of the complementing plasmid was selected for by plating on α-aminoadipate, which is toxic in the presence of the LYS2 gene product (42). The resulting tlc1::LEU2 rad52::LEU2 strain was then grown on YPD for an additional 50 to 75 generations, during which time telomeres progressively shortened and the cells began to senesce.

FIG. 2.

Scheme for selection of functional deletion alleles. A strain with a chromosomal deficiency of the TLC1 gene, tlc1Δ::LEU2 rad52::LEU2, was plated on α-aminoadipate to select for the loss of the complementing plasmid pL317/Tg2[TLC1 LYS2 CEN]. A colony was picked from the α-aminoadipate plate and streaked onto YPD. From that plate (first streak after α-aminoadipate), a single colony was picked and grown in liquid YPD. These senescing cells were transformed with various forms of pSD107: a plasmid harboring a wild-type copy of the TLC1 gene, pSD107[CEN TRP1 TLC1], resulting in a large number of transformants (1); a plasmid carrying a 690-bp deletion of the TLC1 gene, pSD107[CEN TRP1 tlc1ΔNcoI-NsiI], resulting in no colonies (2); or the TLC1 deletion library, pSD107[CEN TRP1 tlc1.del.library], resulting in an intermediate number of transformants (3).

Senescing cells were next transformed with various forms of pSD107. As a positive control, transformation with a plasmid carrying the wild-type TLC1 gene (pSD107) produced many colonies on a selective plate (Fig. 2). As a negative control, a plasmid carrying a 690-bp deletion spanning the NcoI-NsiI sites of the TLC1 gene (pSD107ΔNcoI-NsiI), which includes the 16-nt template (38), gave rise to few or no colonies. Transformation of senescing cells with either of the libraries of deletions (pSD107 tlc1.del.library) resulted in an intermediate number of colonies, as seen in Fig. 2, indicating that a subset of the tlc1 deletion alleles in the library can complement a tlc1Δ strain.

Mapping deletion alleles.

To map the position of each deletion in TLC1 that complemented the tlc1Δ strain, DNA isolated from yeast colonies that successfully recovered from senescence was subsequently transformed into E. coli to amplify the plasmid carrying the tlc1 deletion allele. Plasmid containing the deletion was digested with BglII and PflMI, both unique restriction sites in pSD107. In the case of the wild-type plasmid with no deletion, digestion generates a 1,375- and a 5,974-bp fragment. The entire TLC1 gene is included in the smaller of these two fragments (Fig. 3A). Following gel purification, the BglII-PflMI fragment was digested with TaqI restriction endonuclease, which in the absence of a deletion generates six fragments ranging in size from 80 to 424 bp. The banding pattern seen on the agarose gel upon TaqI digestion of each tlc1 deletion allele could be compared to that of wild-type TLC1 (Fig. 3B, pSD107) to identify the approximate position of the deletion. As seen in Fig. 3C, the deletions were restricted to three regions of the gene, with the gaps between them defining three putative essential regions.

FIG. 3.

Mapping deletions of TLC1 that complement a senescing tlc1Δ strain. (A) The RNA transcribed from the TLC1 gene is diagrammed as a solid line with the 5′ and 3′ ends indicated (3, 38). Regions of the RNA with known functions are indicated by boxes: a Ku domain (nt 288 to 335) (27), the template (nt 468 to 483) (38), and the Sm site (nt 1152 to 1160) (35). Below is a diagram of the BglII-PflMI restriction endonuclease fragment of pSD107 showing the location of the TaqI sites used to map the positions of the deletions. The BglII-PflMI fragment includes 65 bp upstream and 50 bp downstream of the coding region for the TLC1 RNA; subsequent digestion withTaqI produces fragments of the indicated sizes. (B) Agarose gel showing the TaqI digests of BglII-PflMI fragments obtained from tlc1 deletion alleles that complement a tlc1Δ strain. Lanes: M, HaeIII fragments of φX174 double-stranded DNA; M′, MspI fragments of pBR322; ∗, wild-type alleles recovered in the selection. (C) Summary of deletion alleles competent to rescue senescing yeast. The approximate position of each deletion is indicated below the diagram of the BglII-PflMI fragment of the TLC1 gene. The numbers in parentheses indicate the number of different deletions found in each region. (D) Unselected deletion alleles span the entire TLC1 gene. Summary of deletion alleles isolated from a culture of E. coli transformed with the deletion library. The numbers in parentheses indicate the number of different deletions found in that region of the gene. (E) Three candidate essential regions of TLC1 were inferred from DNA sequence analysis of complementing alleles shown in panel C. The 5′ region (nt 100 to 138) was directly tested; tlc1 Δ101-138 did complement senescing tlc1Δ cells, and therefore this region was reclassified as not essential. The central essential region, which consists of nt 455 to 884 and contains the template (38), was directly tested with a number of selected alleles as described in Materials and Methods. The Sm region (35) was not directly tested in the present study.

Since deletions that complemented the tlc1Δ strain were confined to such limited regions of the gene, we were concerned about the randomness of the deletion library. To test this, the unselected deletion library was transformed into E. coli. Plasmid was isolated from colonies and analyzed as above with a BglII and PflMI digest followed by a TaqI digest and agarose gel electrophoresis. Thirty-three deletions were analyzed and found to span the entire TLC1 gene (Fig. 3D). Thus, deletions in the unselected library were not confined to limited regions of the gene.

Identification of essential regions of the gene.

The restriction fragment analysis described above gave only the approximate position of each deletion. To identify the exact deletion endpoints, the plasmid carrying the largest deletion in each region of the TLC1 gene was sequenced across the site of the deletion. In addition, the deletion allele was retested for complementation of a tlc1Δ strain (data not shown). This information was used to generate the diagram shown in Fig. 3E, specifying endpoints for nonessential regions of the TLC1 gene. From the deletion analysis, it was found that approximately 800 of the 1,261 bp of the TLC1 gene are absent from deletion alleles that rescue senescing tlc1Δ yeast and are therefore defined as nonessential (indicated as the unshaded regions in Fig. 3E). None of the complementing deletions spanned the template or the entire Sm binding site of the TLC1 gene. Several different tlc1 alleles containing deletions within the central region were tested for complementation of senescing tlc1Δ yeast, and each failed, so the central region was shown to be essential by both positive and negative results. The in vivo selection analysis implicated nt 100 to 138 as candidates for an essential region. Although direct testing subsequently showed that a Δ101-138 tlc1 allele was able to complement senescing tlc1Δ yeast, it will be shown below that the region of nt 101 to 138 makes a significant contribution to Est2p binding. No further analysis has yet been done for deletions in the Sm region.

Although these noncomplementing deletions could in principle simply destabilize the RNA, it will be shown later that numerous deletions tested accumulate significant amounts of RNA (∼15 to >100% of the wild-type level by Northern analysis). Furthermore, there are examples of tlc1 mutants with more-drastically reduced RNA levels that still complement a tlc1Δ strain (35).

Attempts were made to create a minimal functional telomerase RNA by combining deletions that by themselves were tolerated. tlc1 alleles containing a deletion of either nt 148 to 440 or nt 885 to 1125 are still viable, although telomere length is reduced (data not shown). However, the Δ148-440 Δ885-1125 double mutant did not rescue senescing yeast, and substitution of an unrelated sequence from nt 148 to 440 and 885 to 1125 also did not restore activity. Thus, the putative essential regions of TLC1 RNA are not sufficient by themselves for function.

Analysis of binding of Est proteins.

Numerous tlc1 alleles with deletions of approximately 16 to 300 nt, spanning almost the entire region of TLC1 from nt 101 to 1125, were tested for interaction with telomerase protein subunits Est1 and Est2 by a coimmunoprecipitation assay. Each tlc1 deletion allele, in a TRP1-marked CEN plasmid, was transformed into cells containing wild-type genomic TLC1 and an epitope-tagged allele either of EST1 (HA₃-EST1) or of EST2 (ProA-EST2) (Fig. 4). Following immunoprecipitation of either HA₃-Est1p or ProA-Est2p, any coimmunoprecipitated RNA was examined by Northern blot hybridization to determine if the tlc1 deletion allele was associated with the Est protein of interest. Samples of the data are shown in Fig. 5A and 6A. U1 snRNA, which does not bind either Est protein, was not immunoprecipitated; telomerase RNA from a yeast strain without epitope-tagged proteins was also not immunoprecipitated (or in some cases, to a very limited extent), confirming the specificity of the assay.

FIG.4.

Method for determining regions of TLC1 RNA that interact with Est protein(s). A strain with wild-type genomic TLC1 and either protein A-tagged EST2 (ProA-EST2) or HA₃-tagged EST1 (HA₃-EST1) is transformed with a plasmid expressing a deletion allele of TLC1 RNA. The epitope-tagged Est protein is subsequently immunoprecipitated from a whole-cell extract. RNA is isolated from the immunoprecipitation beads and analyzed by Northern blot hybridization.

FIG.5.

Coimmunoprecipitation of tlc1 deletion alleles with HA₃-tagged Est1 protein. (A) Northern analysis of HA₃-Est1p coimmunoprecipitations. Lanes: M, φX174 DNA HaeIII fragments; I, input RNA; B, RNA coimmunoprecipitated with HA₃-Est1p; S, supernatant RNA. Approximately 10% of the total input and supernatant RNAs and 25% of the bound RNA were loaded. The region of TLC1 RNA deleted in the allele is listed above. (Upper panel) Wild-type TLC1 RNA and faster-migrating tlc1 deletion alleles. (Lower panel) U1 snRNA, which is not a telomerase subunit and is not coimmunoprecipitated with HA₃-Est1p. (B) Summary of results. Bars represent the region of the gene deleted in each tlc1 deletion allele tested; a filled bar indicates that the allele coimmunoprecipitated with HA₃-Est1p (20 to 95% of wild-type level), and an empty bar indicates that the allele was not substantially coimmunoprecipitated with HA₃-Est1p (<10% of wild-type level). ∗, coimmunoprecipitation of the allele was not normalized to the wild type. The cross-hatched bar at the bottom indicates the region within which determinants for Est1p binding reside.

FIG. 6.

Coimmunoprecipitation of tlc1 deletion alleles with ProA-tagged Est2 protein. (A) Northern analysis of ProA-Est2p coimmunoprecipitations. Lanes: M, φX174 DNA HaeIII fragments; I, input RNA; B, RNA coimmunoprecipitated with ProA-Est2p; S, supernatant RNA. Approximately 10% of the total input and supernatant RNA and 17% of the bound RNA were loaded. The region of TLC1 RNA deleted in the allele is listed above. Upper and lower panels are as for Fig. 5. (B) Summary of results. Bars represent the region of the gene deleted in each tlc1 deletion allele tested; a filled bar indicates that the allele coimmunoprecipitated with ProA-Est2p (20 to 95% of wild-type level), a shaded bar for tlc1Δ616-727 indicates intermediate binding (∼15% of wild-type level), and an empty bar indicates that the allele was not substantially coimmunoprecipitated with ProA-Est2p (<10% of wild-type level). ∗, coimmunoprecipitation of the allele was not normalized to the wild type. Cross-hatched bars at the bottom indicate regions within which determinants for Est2p binding reside.

Wild-type TLC1 RNA present in the cells served as an internal control for the coimmunoprecipitations; binding of tlc1 deletion alleles was compared to binding of wild-type TLC1 RNA (see Materials and Methods). It was important to determine if the presence of wild-type TLC1 RNA affects the extent of tlc1 deletion allele coimmunoprecipitation, for instance by competing with the deletion allele for Est protein binding or mediating indirect binding through dimerization. In the presence of wild-type TLC1 RNA, Δ411-534 coimmunoprecipitated with HA₃-Est1p but not with ProA-Est2p, and Δ517-637 coimmunoprecipitated with ProA-Est2p but not with HA₃-Est1p. When these tlc1 deletion alleles were tested in strains that did not express wild-type TLC1 RNA, the same results were obtained with one exception: Δ411-534, which was not detectably coimmunoprecipitated with ProA-Est2p in the presence of wild-type TLC1 RNA, resulted in a weak signal in the absence of wild-type TLC1 RNA (data not shown).

Regions of TLC1 RNA required for Est1p and/or Est2p binding.

A summary of the results obtained from the coimmunoprecipitation studies is shown in Fig. 5B for HA₃-Est1p and in Fig. 6B for ProA-Est2p. The region of the gene deleted in each allele examined is shown as a bar. In addition to visual inspection, for the majority of alleles we also quantitated the extent of coimmunoprecipitation relative to that of the wild type by Phosphorimager analysis of the Northern blots as described in Materials and Methods. Deletion alleles classified as binders gave signals substantially above background on a Northern blot and when quantitated had levels of binding 20 to 95% of that of the wild type. The majority had levels of binding that were 20 to 40% that of the wild type. However, a few alleles had levels of binding ≥70% that of the wild type: tlc1Δ148-440 and tlc1Δ483-534 with both HA₃-Est1p and ProA-Est2p and tlc1Δ237-467 with HA₃-Est1p. The tlc1 deletion allele Δ616-727 was intermediate in the ProA-Est2p coimmunoprecipitations, showing a binding level which was 15% that of the wild type. The tlc1 deletion alleles classified as nonbinders had no detectable coimmunoprecipitated RNA, or a faint signal and binding <10% that of the wild type. We suspect that some of these alleles may still interact weakly with the protein of interest. For example, a tlc1 allele with nt 101 to 138 deleted failed to detectably coimmunoprecipitate with ProA-Est2p; however, this same allele rescued tlc1Δ yeast, so presumably it can still interact with the catalytic subunit to a limited extent.

Any region of TLC1 RNA that was absent from a tlc1 deletion allele that substantially coimmunopurified with the Est protein of interest was classified as not required for coimmunoprecipitation. Regions of TLC1 RNA that whenever deleted resulted in very little to no detectable coimmunopurification with the Est protein of interest were classified as required for coimmunoprecipitation and thus were implicated as containing determinants for interaction with the Est protein of interest. We found that nt 535 to 707 of the central essential region of TLC1 RNA were required for coimmunoprecipitation with HA₃-Est1p (Fig. 5B) while nt 101 to 138 and 728 to 864 were required for coimmunoprecipitation with ProA-Est2p (Fig. 6B).

Regions of TLC1 RNA flanking the template could be deleted without eliminating coimmunoprecipitation with ProA-Est2p: tlc1Δ237-467 deletes up to the 5′ end of the template, and tlc1Δ483-534 deletes from the last nucleotide of the template to nt 534. However, tlc1Δ411-534 does not detectably coimmunoprecipitate with ProA-Est2p in the presence of wild-type TLC1 RNA, and the only region of the RNA that it does not have in common with the two alleles mentioned previously is nt 468 to 482 of the template, suggesting a potential interaction in this region (Fig. 6B). Therefore, we tested a tlc1 allele that had a deletion of the template (Δ468-483) for the ability to coimmunoprecipitate with ProA-Est2p (as described in Materials and Methods). As shown in Fig. 7, the template is not required for coimmunoprecipitation with ProA-Est2p. Perhaps the deletion in tlc1Δ411-534 happens to result in an RNA structure inhibitory to Est2p binding, or perhaps there are several weaker binding determinants in the 411 to 534 region, no one of which is essential.

FIG. 7.

The template is not required for coimmunoprecipitation of TLC1 RNA with ProA-Est2p. Northern analysis of ProA-Est2p coimmunoprecipitations. Lane I, input RNA; lane B, RNA coimmunoprecipitated with ProA-Est2p; lane S, supernatant RNA; lane M, φX174 DNA HaeIII fragments. Approximately 10% of the total input and supernatant RNA and 17% of the bound RNA were loaded. RNAs analyzed are indicated to the left of the gel.

This coimmunoprecipitation analysis did not cover the entire TLC1 RNA. Therefore, it is possible that additional regions of TLC1 RNA not tested in this study contribute to the association of Est1p and/or Est2p with TLC1 RNA—separately or by long-range interaction with the regions that we have shown to be necessary for interaction.

DISCUSSION

Our results indicate that approximately 460 nt of the ∼1.3-kb S. cerevisiae telomerase RNA are required for in vivo telomerase function, divided into two noncontiguous regions of the gene. The main essential region of 430 nt includes the template and contributes to binding Est2p, the telomerase catalytic subunit, and Est1p, another protein required for telomere maintenance in vivo (14). Another putative essential region is 32 nt and contains the majority of the Sm binding site. The regions that we show not to be required in vivo, although nonessential, cannot be assumed to be nonfunctional. For instance, the 48-nt stem-loop implicated in Ku interaction (27) is in one of our nonessential regions. Also, nonessential nt 101 to 138 are required for detectable coimmunopurification of the S. cerevisiae telomerase RNA with Est2p.

Less than half of TLC1 RNA is essential in vivo.

Our in vivo selection for functional tlc1 deletion alleles revealed that less than half of TLC1 is required for sufficient telomerase function to maintain cell growth. The essential regions comprise ∼460 nt of TLC1 RNA, which is still considerably larger than the proposed essential portion of human telomerase RNA, two regions totaling about 290 nt (24). The evolutionarily conserved structure of vertebrate telomerase RNAs has a core region with a structure similar to that of ciliate telomerase RNAs, and three additional vertebrate-specific elements (4). It may be that a portion of yeast telomerase RNA adopts a structure similar to this evolutionarily conserved core region. The increased size of the TLC1 RNA may reflect additional function(s), or it may simply be required for structural stability.

Est2p interacts with RNA sequences distant from the template.

The ability of ProA-Est2p to coimmunoprecipitate a template-deleted RNA demonstrates that the template is not required for binding the catalytic subunit, although it remains possible that it contributes some binding energy. In Tetrahymena thermophila, binding of the telomerase catalytic subunit to template-proximal elements has been observed (15); until the secondary structure is elucidated, it is not clear whether any of the RNA sequences which we have identified might be proximal to the template in the structured RNA.

We show that the two regions of S. cerevisiae telomerase RNA required for interaction with the reverse transcriptase subunit Est2p are distant from the template in linear sequence. It has already been reported for both yeast and Tetrahymena that regions of the telomerase reverse transcriptase responsible for RNA binding reside in the noncatalytic N-terminal domain (2, 8). Thus, there is a striking similarity between results for the RNA and the protein: the portions that interact during catalysis are not tightly bound, leaving the RNA template free to translocate through the protein active site, while interactions removed from the template and from the reverse transcriptase motifs maintain the integrity of the RNP complex by preventing dissociation.

Regions of TLC1 RNA required for Est1p and Est2p interaction are separable.

In the present work, we provide evidence for an Est1p-TLC1 RNA interaction and an Est2p-TLC1 RNA interaction. It is not possible to determine exact endpoints of binding determinants from such large deletions. Regardless, in linear sequence the regions that contain the determinants for Est1p and Est2p interaction are nonoverlapping (Fig. 8). Our observations that some deletion alleles coimmunoprecipitate with one protein but not the other suggest that these proteins can bind TLC1 RNA independently. This is consistent with the fact that Est1p can coimmunoprecipitate TLC1 RNA in an Est2p-independent manner (11, 44). However, in an intact, active telomerase complex there may be important interactions between these two proteins.

FIG. 8.

Regions of TLC1 RNA required for binding Est1p and Est2p are nonoverlapping. Determinants for Est1p binding reside within nt 535 to 707 of the central essential region of TLC1 RNA, while determinants for Est2p binding reside within nt 101 to 138 and 728 to 864.

RNA sequences required for binding a particular protein could make direct RNA-protein interactions. On the other hand, they could participate indirectly, for example by contributing to the correct folding of the portion of the RNA that makes direct contact. Despite this caveat, we predict that if there are discrete RNA structures necessary and sufficient for Est1p and Est2p binding, they will reside within the regions which we have shown to be required for coimmunoprecipitation.

In conclusion, TLC1 RNA appears to function as a scaffold for various proteins to bind and be assembled into the telomerase complex. Proteins include the previously described Sm proteins, which may be important for complex formation and/or nuclear localization (22), and telomerase subunits Est1p and Est2p. This does not preclude an additional, more direct role for regions of the RNA other than the template in telomerase activity. There are examples of regions of telomerase RNA distant from the template being required for active site function, but not for reverse transcriptase binding per se (15, 32, 39).

Is yeast telomerase a dimer in vivo?

Telomerase complexes with more than one active site have been observed for S. cerevisiae telomerase assembled in vivo (28) and for human telomerase assembled from recombinant components in vitro (41). Functional multimerization between two inactive human telomerase catalytic subunits has also been demonstrated (1). The strains used for our coimmunoprecipitation assay contained both a wild-type and a deletion variant of TLC1 RNA. We found that deletion of certain regions of the RNA would disrupt coimmunoprecipitation with one Est protein but not the other. Such results seem improbable if a tlc1 deletion allele was dimerized with a wild-type TLC1 RNA-containing complex. For example, a tlc1 deletion allele lacking the region required for Est1p interaction but still containing the region required for Est2p interaction, if in a monomeric complex, should coimmunoprecipitate with Est2p but not Est1p. If this same deletion allele were in a multimeric complex, it would most likely still coimmunoprecipitate with Est1p, albeit through an indirect interaction (i.e., a tlc1 deletion allele-containing complex which lacked Est1p could dimerize with a wild-type TLC1 RNA-containing complex which contained Est1p). Of course, some of our deletions could in principle prevent dimerization, but that seems unlikely for all of the multiple nonoverlapping deletions we examined.

Thus, our binding studies suggest that yeast telomerase is not an obligate multimeric complex in the steady state. Higher-order complexes may occur during specific points in the cell cycle or certain subcellular locations, e.g., the telomere. This is not inconsistent with the identification of multimeric complexes by Prescott and Blackburn (28). Alternatively, if interactions between monomers are significantly weaker than those within monomers, multimeric complexes could have been disrupted during one of the steps of our coimmunoprecipitation assay.