INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Telomeres are the natural ends of linear chromosomes. Telomeres are maintained at a characteristic length by a balance between two forces, loss of telomeres during DNA replication and synthesis of telomeres by an enzyme called telomerase. Telomerase is a special reverse transcriptase which contains not only a reverse transcriptase catalytic subunit but also an RNA molecule which serves as the template for telomere elongation (11). In yeast, the catalytic subunit is called Est2 (7, 20, 25, 26), and the RNA template is called Tlc1 (39).

Telomerases from several organisms have been partially characterized (3, 7, 12, 13, 24, 26, 29, 30). In general, these complexes contain components in addition to the catalytic subunit and the RNA template (10, 12, 24, 30, 36). For the yeast Saccharomyces cerevisiae, genetic screens have identified five genes (EST1, EST2, EST3, EST4/CDC13, and TLC1) (20, 27, 39) whose mutations lead to progressive telomere shortening and eventual loss of viability (i.e., senescence). EST2 and TLC1 encode the reverse transcriptase (25, 26) and the RNA template (39), respectively. The Cdc13 or Est4 protein can bind the single-stranded G-rich telomeric sequence both in vitro and in vivo (2, 23, 32). This protein apparently caps the telomere, protecting it from nucleolytic digestion. The functions of the other two genes, EST1 and EST3, are less clear. Neither of them is required for in vitro telomerase activity (5, 25), even though mutants exhibit the same senescence phenotype as TLC1 or EST2 mutants (20). There is evidence that Est1 is associated with telomerase, since Est1 coprecipitates with Tlc1 and telomerase activity (21, 40). In addition, Est1 may be associated with the telomere since, like Cdc13, Est1 can bind single-stranded G-rich telomeric DNA in vitro (43). However, the affinity of Est1 for such DNA is low, much lower than the affinity of Cdc13. Unlike Cdc13, Est1 requires a free end for binding to DNA (43). We noticed a possible RNA-binding motif in Est1 and have studied the role of this motif with the idea that Est1 might bind Tlc1 directly.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Yeast strains, genetic manipulations, and plasmids. All yeast strains were derived from W303 (MATa/MAT ade2-1/ade2-1 his3-11,15/his3-11,15 leu2-3,112/leu2-3,112 trp1-1/trp1-1 ura3-1/ura3-1 can1-100/can1-100 ssd1-d/ssd1-d [psi⁺]) (41). Yeast media and transformation were as described previously (42).

tlc1/TLC1

View larger version (57K): [in this window] [in a new window]   FIG. 1.   Est1 is a nuclear protein. (Top panels) Est1 tagged with 3HA was expressed from the GAL promoter (plasmid pVL242). Cells were fixed and stained with 4',6'-diamidino-2-phenylindole (DAPI), 12CA5 (anti-HA antibody), and fluorescein isothiocyanate (FITC)-labeled secondary antibody. Cells were visualized by Nomarski optics (NOM) or by fluorescence for DAPI (DAPI) or fluorescence for FITC (FITC). About one-third of all cells examined showed strong tag- and antibody-dependent nuclear FITC staining. (Bottom panel) Control strain containing an empty vector processed for FITC staining.

Yeast cell extracts, immunoprecipitation, total RNA preparation, and Northern blot analysis. For preparation of yeast extracts, yeast cells were grown to mid-log phase (optical density at 600 nm [OD₆₀₀], about 1) with selection for plasmid-borne markers. Cells were harvested by centrifugation and washed once with ice-cold water and once with high-salt lysis buffer (HSLB) (40). All operations thereafter were carried out at 4°C. The cell pellet from 50 ml of cells at an OD₆₀₀ of 1 was suspended in HSLB containing protease inhibitors and an RNase inhibitor, and about 750 µl of acid-washed glass beads was added. The cells were then either quickly frozen to 70°C and stored or broken by shaking in a Mini-Bead-Beater (Biospec). Breaking was accomplished by 30 s of vibration at the top speed twice, separated by 3 min of cooling on ice. The mixture was centrifuged at 14,000 rpm at 4°C for 5 min in an Eppendorf microcentrifuge, and the resulting supernatant was transferred into a fresh tube. The beads were washed with an additional 300 µl of the same buffer and spun to collect additional protein. The two supernatants were combined and spun again under the same conditions. The supernatant was transferred into a clean tube. The protein concentration of such cell extracts was typically 12 to 20 mg/ml, as determined by the Bradford method (Bio-Rad kit). Such extracts were used for immunoprecipitation.

Genomic DNA preparation and Southern blot analysis. Genomic DNA was prepared after cells were broken with glass beads. For Southern analysis of telomere length, 1.0 µg of genomic DNA was digested with XhoI or PstI endonuclease at 37°C for 2 h. In some Southern blots (e.g., see Fig. 3 and Fig. 5C, right panel), 3 ng of DNA from an HaeII-NdeI digest of plasmid pBST3 was added to the digested genomic DNA. This digestion produced a fragment of 511 bp and a fragment of 1,436 bp that hybridize to yeast telomeric sequences and so serve as molecular size markers on Southern blots. Electrophoresis was carried out at 80 V for about 5 h. The separated DNA was transferred to NYTRAN PLUS Nylon membranes and probed with a ³²P-labeled yeast telomere sequence (TELPG; Table 1).

Single-stranded DNA-binding assay. The binding reaction was conducted with a total reaction volume of 20 µl. The binding buffer (43) contained 50 µg of poly(dI-dC) per ml. To this binding buffer was added 5.0 µl of the immunoprecipitate that had been washed once with binding buffer after the standard immunoprecipitation protocol (see above). After incubation at room temperature for 8 min, a mixture of 3.0 ng of ³²P-labeled TELPG (27-mer) and 6.6 ng of ³²P-labeled oligo EST1-RNP1C (62-mer) was added to the reaction, and incubation was continued for another 15 min at room temperature. The reaction was stopped by gently adding 1,000 µl of binding buffer. The beads were collected immediately, washed gently three times in binding buffer, and eventually suspended in 10 µl of binding buffer. Four microliters of loading buffer (0.25% bromophenol blue, 0.25% xylene cyanol, 30% glycerol) was added to the suspension. The mixture was heated in boiling water for 3 min before being loaded onto a 7.6% polyacrylamide-7.0 M urea-1× Tris-borate-EDTA sequencing gel. After electrophoresis, the gel was placed on Whatman paper and vacuum dried. The signal was quantified using a Molecular Dynamics PhosphorImager.

Telomerase assay. Telomerase was purified and assayed as previously described (34). About 15 µg of each partially purified fraction was used for each telomerase assay. For the RNase A control, 50 ng of RNase A was added to the partially purified telomerase fraction; this mixture was incubated at room temperature for 5 min, and then telomerase activity was assayed.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Est1 is a nuclear protein. If Est1 is a component of active telomerase at telomeres, then it ought to be a nuclear protein. On the other hand, if Est1 is less directly involved in telomerase function, then it might have some other location. We overexpressed a functional, 3HA-tagged version of Est1 and localized it by immunofluorescence. All the Est1-dependent immunofluorescence appeared to be in the nucleus (Fig. 1).

The association of Est1 with Tlc1 does not depend on EST2 or EST3. We have previously shown that the Est1 protein is associated with the Tlc1 RNA (40). To find out whether this interaction depends on other telomerase components, we examined the Est1-Tlc1 association in est2 and est3 deletion strains. A functional, 3HA-tagged version of Est1 was expressed in yeast and immunoprecipitated. Immunoprecipitates were processed for Northern blotting, and the presence of Tlc1 and control RNAs was assayed. Figure 2 shows the result from a typical experiment. The Tlc1 RNA was detected in the tagged Est1 immunoprecipitates from wild-type, est2, and est3 cells. This association was specific, since control ACT1 mRNA was not detected in the 3HA-tagged Est1 immunoprecipitate, and no RNA was detected in the untagged Est1 immunoprecipitate. The Tlc1 RNA signal revealed by Northern analysis is not due to any DNA contamination, since RNase A treatment completely eliminates the signal (see Fig. 6, lanes 26 and 27). These results indicate that the association of Est1 and Tlc1 does not require Est2 or Est3, the only other known components of the telomerase complex. Thus, the interaction of Est1 with Tlc1 may be direct.

View larger version (40K): [in this window] [in a new window]   FIG. 2.   The association of Tlc1 RNA with Est1 does not depend on EST2 or EST3. The amount of Tlc1 RNA (top panel) or ACT1 mRNA (bottom panel) in various experiments was assayed by Northern blotting. "Total RNA" shows the total RNA from wild-type, est2, or est3 strains, each of which either carries (+) or does not carry () tagged Est1 expressed from the ADH1 promoter (pJZ-3HA-EST1). "IP-RNA" shows the RNA associated with 3HA-tagged Est1 after the latter was immunoprecipitated from tagged (+) or untagged control () strains with monoclonal antibody 12CA5. The amount of Tlc1 precipitated was somewhat smaller in the est2 and est3 strains than in the wild-type strain in all three of three experiments.

Ebs1 is a homolog of Est1. Database searches show that the Est1 protein has 48% similarity (27% identity) to the protein encoded by the poorly characterized gene EBS1 (open reading frame YDR206w). This similarity is spread throughout the two proteins but is most pronounced in a 100-amino-acid region centered on a putative RNA recognition motif (RRM) in Ebs1 (see below). Interestingly, disruption of EBS1 resulted in slightly shortened telomeres (Fig. 3), although there was no senescence phenotype. Furthermore, there was an indication that est1 ebs1 double mutants had a slightly stronger senescence phenotype than est1 single mutants alone. est1 single-mutant spore clones show a variable time to senescence, but est1 ebs1 double mutants always senesce as fast as the fastest-senescing est1 single mutants. These results suggest that EBS1 is slightly redundant with EST1 for telomere maintenance.

View larger version (55K): [in this window] [in a new window]   FIG. 3.   The ebs1 mutant has short telomeres. Genomic DNA was isolated from two ebs1 null mutants, a wild-type (wt) strain, and an est1 null mutant. DNA was digested with PstI, mixed with 511- and 1,436-bp size markers, and fractionated by agarose gel electrophoresis. After Southern blotting, DNA was hybridized with a 32P-labeled telomeric sequence.

View larger version (78K): [in this window] [in a new window]   FIG. 4.   The EBS1 gene is not essential for telomerase activity in vitro. Telomerase was partially purified (34) from a wild-type strain and from an est1 ebs1 double mutant. Telomerase activity was assayed (34) in the absence () or presence (+) of 50 ng of RNase A. Two to three times as much protein was used in the est1 ebs1 assay as in the wild-type assay, accounting for the apparently higher level of telomerase activity. Lane M, end-labeled 28-mer which serves as a marker for the input telomeric primer oligo.

Est1 has a functionally important RRM. Computer analysis of the Ebs1 protein showed that it contained a good match to an RRM (Fig. 5). The RRM in Ebs1 led us to a putative RRM in Est1 (Fig. 5). Although this putative Est1 RRM has only a poor match to a consensus RRM, it is also true that RRMs of this class can share structural similarity without highly conserved sequence similarity, and it is the structural similarity that is important (1, 4, 17). An RRM consists of 70 to 90 amino acid residues (1, 4, 17). RNP1 and RNP2 are two important and more highly conserved segments within an RRM. The RRM crystal structures of human hnRNP C and U1 snRNP A show that the RRM forms a module of four antiparallel beta sheets on the surface and two alpha helices underneath (a secondary structure). RNP1 and RNP2 form the two central antiparallel beta sheets and are directly involved in single-stranded RNA interactions.

View larger version (98K): [in this window] [in a new window]   FIG. 5.   Phenotypes caused by altering the RRM region of Est1. (A) The RRM consensus sequence (1, 17) is compared to RRM-like regions in Ebs1 and Est1. The most highly conserved regions (RNP1 and RNP2) are shown in bold lettering. U, hydrophobic amino acid (aa). The amino acid changes made in the four est1 mutants are indicated by lowercase letters. To the right of RNP1 in Est1 is the sequence FNDDY; F is F511, and the first D is D513. These amino acids are altered in two previously characterized alleles of est1 (43) (see the Discussion). (B) A presenescent est1 strain (about 35 generations removed from the original spore clone) was transformed with plasmids based on pJZ-3HA-EST1 harboring the empty vector, wild-type (wt) EST1, or the NC, C, ga, or gs mutant allele of EST1. Two independent colonies from each transformation were streaked on YEPD plates and photographed after 60 h of growth. (C) Genomic DNA was extracted from each transformant, digested with XhoI (left panel) or PstI (right panel), and fractionated by agarose gel electrophoresis. After Southern blotting, telomeric DNA was detected with 32P-labeled oligo TELPG. V, transformants containing an empty vector; WT, wild type. In the right panel, the two sharp bands just below and above the fuzzy telomere bands are molecular size markers of 511 and 1,436 bp, respectively (see Materials and Methods).

View larger version (54K): [in this window] [in a new window]   FIG. 6.   Association of Tlc1 RNA with Est1 RRM mutant proteins in vivo. Total RNA and immunoprecipitated RNA (IP-RNA) were made from est1 cells expressing wild-type or mutant forms of either untagged () or 3HA-tagged (+) Est1 from the ADH1 promoter (plasmid pJZ-3HA-EST1 and derivatives). The amounts of Tlc1 RNA and ACT1 mRNA in total RNA or coimmunoprecipitated with Est1 were assayed by Northern blotting. Two independent transformants of Est1-RNP1-NC were assayed. Controls included RNA from a tlc1 deletion mutant (lane 8), RNA from a strain carrying 3HA-tagged Est1 expressed from the genomic EST1 locus (lanes 1 and 11), and IP-RNA treated with RNase A before loading (lanes 26 and 27).

The EST1-RNP1-NC mutant has dominant negative phenotypes. The failure of the triple substitution, est1-rnp1-nc, to complement an est1 mutation raised the question of whether the mutations might have grossly disturbed protein structure (although the fact that Est1 and Est1-RNP1-NC proteins were found equally abundant by Western analysis argues against this notion). However, we found that wild-type EST1 cells transformed with overexpressed est1-rnp1-nc had at least three phenotypes. First, they grew slowly (data not shown). Second, as assayed with a Coulter Channelyzer, average cell size was abnormally large, and microscopic examination showed that some cells were large dumbbells, typical of DNA damage arrest (data not shown). Third, these cells had relatively short telomeres (Fig. 7). These dominant phenotypes suggest that the mutant Est1-RNP1-NC protein retains some functions of wild-type Est1 and consequently disturbs normal telomere maintenance. This notion in turn suggests that the disruption of the putative RRM causes a fairly specific defect in the protein as opposed to a general defect in protein folding.

View larger version (78K): [in this window] [in a new window]   FIG. 7.   The est1 rnp1 nc mutant has a dominant short-telomere phenotype. Genomic DNA from an EST1 W303 strain also expressing plasmid-borne EST1 or est1-rnp1-nc or est1-rnp2-ga from the ADH1 promoter (pJZ-3HA-EST1 and derivatives) was digested with XhoI and analyzed by Southern blotting with 32P-labeled TELPG as a probe.

The Est1-RNP1-NC mutant can bind single-stranded telomeric DNA. Virta-Pearlman et al. reported that Est1 protein can bind to yeast single-stranded TG-rich telomeric DNA, albeit with low affinity (43). In some proteins, RRMs can bind to single-stranded DNA (8, 14, 16, 31, 35). In fact, single-stranded TG-rich telomeric DNA may be particularly good at interacting with RRMs (8, 16, 31). Thus, we wished to know whether the RRM of Est1 was responsible for the DNA-binding activity detected by Virta-Pearlman et al. (43). We developed a single-stranded DNA-binding assay (see Materials and Methods). This assay compares the ability of immunoprecipitated 3HA-tagged Est1 protein to bind two different single-stranded DNAs, one with a telomeric sequence and one with an irrelevant sequence. The assay has a high background (i.e., there is significant binding of telomeric sequences even in the absence of Est1); nevertheless, immunoprecipitates from tagged Est1 strains reproducibly bind more telomeric sequence than do immunoprecipitates from untagged control strains. Figure 8 shows the results of one typical experiment. Immunoprecipitates from the tagged Est1 strain retained two- to threefold more telomeric sequence (oligomer TELPG; 27-mer) than immunoprecipitates from the untagged control strain (Fig. 8). Seven independent experiments (Table 3) were evaluated by analysis of variance (ANOVA), and the difference between the tagged and untagged strains was significant at the 0.05 level. The immunoprecipitated Est1 protein from tlc1 cells gave similar results (data not shown), indicating that the Tlc1 RNA bound to Est1 in vivo does not significantly interfere with the DNA-binding capacity of Est1.

View larger version (70K): [in this window] [in a new window]   FIG. 8.   The Est1-RNP1-NC mutant protein can bind single-stranded telomeric DNA. Monoclonal antibody 12CA5 and protein A beads were mixed with extract from tagged (+) or untagged () Est1 strains (lanes 1 to 4). After washing was done, immunoprecipitates were challenged with a mixture of two end-labeled oligos, a 62-mer (EST1-RNP1C) of irrelevant sequence (Table 1) and a TG-rich, telomeric 27-mer (TELPG) (Table 1). Immunoprecipitates were then washed, and the bound oligos were assayed by phosphorimaging after gel electrophoresis. Lane 5 shows 1/20 the amount of the 27-mer-62-mer mixture that was added to the immunoprecipitates in the other lanes. The ratio of 27-mer to 62-mer is shown beneath the lanes.

Two other alleles of est1 separate the single-stranded DNA-binding function from the RNA-binding function. There are two previously known nonfunctional alleles of est1 that happen to fall within the RRM. These are est1_F511S and est1_D513I (F511 and D513 are shown in Fig. 5A) (43). These alleles are phenotypically similar to est1-rnp1-nc in that they fail to complement an est1 deletion, and they cause telomere shortening when overexpressed. However, both mutant proteins, when purified from Escherichia coli, have wild-type ability to bind single-stranded DNA in vitro (43). We have tested both of these mutant proteins for RNA binding by immunoprecipitation. Comparable amounts of wild-type and mutant proteins were immunoprecipitated (data not shown), but only the wild-type protein caused coimmunoprecipitation of substantial amounts of Tlc1 (Fig. 9). Thus, these alleles, like est1-rnp-nc, separate the RNA-binding function of Est1 from the single-stranded DNA-binding function.

View larger version (31K): [in this window] [in a new window]   FIG. 9.   The Est1 F511S and D513I mutant proteins are defective in binding Tlc1 RNA. Total RNA (left) and IP-RNA (right) were made from est1 cells expressing wild-type or mutant forms of either untagged () or 3HA-tagged (+) Est1 from the ADH1 promoter (plasmid pJZ-3HA-EST1 and derivatives). The amount of Tlc1 RNA in total RNA or coimmunoprecipitated with Est1 was assayed by Northern blotting. The amount of immunoprecipitated protein was assayed by Western blotting (data not shown).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

We have shown previously that Est1 is in a complex with active telomerase. Here we report that Est1 binds to Tlc1 RNA. This binding is very likely direct, as it requires neither EST2 nor EST3. When mutations are made in the RRM of Est1, Est1 partially or wholly loses its ability to associate with Tlc1. Furthermore, the severity of this association defect is correlated with the severity of the telomere shortening and senescence phenotypes of each mutant.

Previously, it has been shown that Est1 is a single-stranded DNA-binding protein in vitro with specificity for TG-rich telomeric sequences (43). In many cases, RRMs can bind both single-stranded RNA and single-stranded DNA. In particular, it seems quite common to find RRMs capable of binding TG-rich single-stranded DNA (8, 15, 16, 18, 19, 22, 28, 31, 33, 35). This fact raises the issue of whether the previously observed binding to TG-rich single-stranded DNA was due to the RRM of Est1. One argument that this might be the case is that the deletion of amino acids 435 to 565 of Est1 greatly diminished single-stranded DNA binding (43) and removed the RRM.

However, there are a number of strong arguments that single-stranded DNA binding probably is not dependent on the RRM. First, Est1 binds d(TGTGTGGG)₃ but does not bind the corresponding polyribonucleotide, r(UGUGUGGG)₃ (43), contrary to the expectation if an RRM were responsible for binding single-stranded DNA (8, 16). Second, a 1,000- to 10,000-fold molar excess of Tlc1 RNA fails to prevent single-stranded telomeric DNA from binding to Est1 (43). Third, two missense alleles that fall within the RRM of EST1, est1_F511S and est1_D513I, lack the ability to bind Tlc1 RNA (Fig. 9) but retain the ability to bind single-stranded DNA (43). These mutations thus separate the single-stranded DNA-binding function from the Tlc1 RNA-binding function. These two alleles have phenotypes strikingly similar to that of est1-rnp1-nc, as expected if their defect were specific for binding Tlc1. Fourth, Est1-RNP1-NC immunoprecipitated from yeast cells (possibly together with associated proteins) can specifically bind single-stranded telomeric DNA (Fig. 8 and Table 3), even though this protein is defective for Tlc1 binding (Fig. 6).

Thus, it appears that Est1 binds to the yeast telomerase RNA, Tlc1, via an RRM, and binds to single-stranded telomeric DNA via some other nearby motif. Therefore, as suggested by Virta-Pearlman et al. (43), Est1 could serve as a bridge between telomerase and telomere ends. That is, by simultaneously binding telomerase RNA through the RRM and binding telomeric DNA through a second motif, it could help anchor telomerase at the telomere. Recently, strong evidence for this bridge model has been provided by Evans and Lundblad (9), who showed that a Cdc13-Est2 fusion could entirely bypass the need for Est1.

The bridge model is consistent with the observation that est1 mutants have telomerase activity in vitro (where substrate telomeric oligos are provided at high concentrations) but nevertheless suffer telomere shortening in vivo (where telomere concentrations are low). That is, the need for Est1 may be apparent only at low substrate concentrations; the effect of Est1 may be to reduce the K_m of the telomerase reaction. The affinity of Est1 for telomeric DNA is, however, quite low (43); perhaps this interaction is regulated by cell cycle position, telomere length, or other proteins (such as Cdc13) to help regulate steady-state telomere length.

Homologs of Est1 have not been found in other organisms. However, telomerases from other organisms do have protein components in addition to the reverse transcriptase subunit. In Tetrahymena, proteins p80 and p95 are associated with telomerase, and p80 can be cross-linked to the RNA component in vivo (6). A human protein called TP1 (also called TLP1) is associated with human telomerase and has significant homology to Tetrahymena p80 (12, 30). Furthermore TP1 can interact with the RNA component of human telomerase in the yeast-based three-hybrid assay, suggesting that TP1 binds to RNA directly. Thus, Tetrahymena, human, and perhaps other telomerases may contain RNA-binding proteins with functions analogous to the function of Est1. Since RNA-binding motifs are often short and are not all of the RNP1 or RNP2 type, functional homologs of Est1 will not necessarily have any protein sequence similarity to Est1. Moreover, it is not clear that tethering of telomerase to telomeres must occur through the RNA template, as it apparently does in S. cerevisiae. Instead, in some organisms, tethering might occur through the reverse transcriptase catalytic subunit.