A Mutant of Tetrahymena Telomerase Reverse Transcriptase with Increased Processivity*

The protein catalytic subunit of telomerase (TERT) is a reverse transcriptase (RT) that utilizes an internal RNA molecule as a template for the extension of chromosomal DNA ends. In all retroviral RTs there is a conserved tyrosine two amino acids preceding the catalytic aspartic acids in motif C, a motif that is critical for catalysis. In TERTs, however, this position is a leucine, valine, or phenylalanine. We developed and characterized a robustin vitro reconstitution system for Tetrahymenatelomerase and tested the effects of amino acid substitutions on activity. Substitution of the retroviral-like tyrosine in motif C did not change overall enzymatic activity but increased processivity. This increase in processivity correlated with an increased affinity for telomeric DNA primer. Substitution of an alanine did not increase processivity, while substitution of a phenylalanine had an intermediate effect. The data suggest that this amino acid is involved in interactions with the primer in telomerase as in other RTs, and show that mutating an amino acid to that conserved in retroviral RTs makes telomerase more closely resemble these other RTs.

The protein catalytic subunit of telomerase (TERT) is a reverse transcriptase (RT) that utilizes an internal RNA molecule as a template for the extension of chromosomal DNA ends. In all retroviral RTs there is a conserved tyrosine two amino acids preceding the catalytic aspartic acids in motif C, a motif that is critical for catalysis. In TERTs, however, this position is a leucine, valine, or phenylalanine. We developed and characterized a robust in vitro reconstitution system for Tetrahymena telomerase and tested the effects of amino acid substitutions on activity. Substitution of the retrovirallike tyrosine in motif C did not change overall enzymatic activity but increased processivity. This increase in processivity correlated with an increased affinity for telomeric DNA primer. Substitution of an alanine did not increase processivity, while substitution of a phenylalanine had an intermediate effect. The data suggest that this amino acid is involved in interactions with the primer in telomerase as in other RTs, and show that mutating an amino acid to that conserved in retroviral RTs makes telomerase more closely resemble these other RTs.
Telomeres form a protective cap on chromosome ends and are usually composed of short G-rich DNA repeats complexed with proteins (1). The telomeres of unicellular eukaryotes and of the germ cells of multicellular organisms are maintained by the enzyme telomerase, which was first identified in the ciliated protozoan Tetrahymena thermophila (2). Telomerase activity has also been detected in many human cancer cells and appears to be necessary for their continued growth (3)(4)(5).
Telomerase is a reverse transcriptase that utilizes an internal RNA molecule as a template for the extension of DNA ends (6). The RNA component of telomerase has been cloned from many different organisms (reviewed in Ref. 7). The protein catalytic subunit of telomerase, known as TERT for telomerase reverse transcriptase, was first identified in the ciliate Euplotes aediculatus and the yeast Saccharomyces cerevisiae (8,9) and has since been identified in humans, mice, fission yeast, plants, and other ciliated protozoa (10 -18). The evolutionary conservation of TERT 1 suggested that it and the telomerase RNA form a core catalytic unit. This was supported by experiments in which in vitro translated TERT and the telomerase RNA were sufficient to reconstitute both human and Tetrahymena telomerase activity (15,19,20), although reconstitution was dependent on the presence of chaperone proteins in the rabbit reticulocyte lysate used for in vitro translation (21,22).
The derived amino acid sequences of the TERT proteins contain motifs common to all reverse transcriptases, consistent with the role of telomerase in RNA-templated DNA polymerization (23). Within these reverse transcriptase motifs is an invariant trio of aspartic acids that are directly involved in catalysis (24,25). Mutation of any of these three aspartic acids in yeast or human TERT abolished telomerase activity in vitro and in vivo, confirming the importance of the RT motifs for telomerase catalysis (9,10,19,26). While the basic RT catalytic mechanism is undoubtedly conserved in telomerase, there are many features of telomerase that distinguish it from other RTs. The current model for the in vitro action of Tetrahymena telomerase (27,28) is shown in Fig. 1. The 3Ј end of the DNA primer first base pairs with the telomerase RNA, and the primer is then extended using the RNA as a template. When the end of the 9-nucleotide (nt) template region is reached, the primer either dissociates from the complex, or translocates on the RNA template and realigns for another round of elongation. In contrast, retroviral RTs do not form stable complexes with the RNA template, and they catalyze processive DNA synthesis that does not normally involve translocation of the template relative to the primer (for an informative exception, see Ref. 29). The amino acids that are unique to telomerase presumably contribute to these differences. Even within the RT motifs, Tetrahymena TERT and the RT from human immunodeficiency virus (HIV RT) share only 16% amino acid sequence identity. Outside these motifs the sequences are even less conserved, and the TERTs have long amino-and carboxyl-terminal regions that are absent in HIV RT. It is therefore of interest to study the details of telomerase action to determine similarities and differences from what is already known of the mechanisms of other RTs.
One amino acid that differs between telomerase and other RTs is located two residues prior to the two conserved catalytic aspartic acids in RT motif C (Fig. 2). This residue is a leucine, valine, or phenylalanine in all TERTs identified to date, while it is a tyrosine in 29 out of 29 retroviral RTs (30). Furthermore, this tyrosine is critical for catalytic activity of the retroviral RTs. Mutation of this residue to alanine or serine in either HIV RT or murine leukemia virus RT reduced reverse transcriptase activity to 1-7% of wild type (wt) levels (31)(32)(33)(34). Mutation of the tyrosine to phenylalanine led to a more modest reduction of HIV RT activity (22-30% of wt levels; Refs. 32 and 33), and had no effect on murine leukemia virus RT activity (34). Examination of the crystal structure of HIV RT provides a plausible explanation for these effects. The shape of the protein resembles a right hand, with a large cleft between the "fingers" and "thumb" domains in which the template and primer nucleic acids bind (35,36). The tyrosine of motif C protrudes into this cleft and forms several contacts with the template and primer: two carbon atoms in its side chain phenyl ring show hydrophobic interactions with the deoxyribose of the terminal primer nucleotide (37), and the hydroxyl group on its side chain hydrogen bonds with atoms in the penultimate primer nucleotide (37,38). It was hence postulated that the tyrosine is involved in precisely positioning the template-primer relative to the active site (37). Thus, substitution of the tyrosine with a phenylalanine (which lacks a hydroxyl group) would eliminate the hydrogen bond to the primer, and substitution of an amino acid lacking a phenyl ring such as alanine or serine would further reduce interaction with the primer.
It is therefore curious that the TERT proteins do not need a tyrosine at this site for activity. In order to further investigate the role of this amino acid (Leu-813) in Tetrahymena telomerase action, we mutated it to tyrosine, phenylalanine, and alanine. The tyrosine mutant (L813Y), while not showing any change in telomerase activity level, did show an increase in enzyme processivity. This is an unusual case where a single amino acid substitution provides an "improvement" in enzyme action. Because this substitution does not occur in any of the nine TERTs that have been identified, we propose that it may be advantageous to telomerases to synthesize rather small numbers of repeats in a single enzymatic cycle.

EXPERIMENTAL PROCEDURES
Construction of a Synthetic Tt_TERT Gene-The cloning of the T. thermophila TERT gene has been described (14). A synthetic Tt_TERT gene was constructed in order to correct the 66 Tetrahymena Glu codons (TAA and TAG) to the universal genetic code and to change the codon usage to more closely resemble that of both Escherichia coli and rabbit (Oryctolagus cuniculus). 469 of the 3351 base pairs of the gene were changed without changing any encoded amino acids. The sequence of the synthetic gene is available from the authors upon request. A set of 42 100-mer oligonucleotides was synthesized on an ABI DNA synthesizer. These oligonucleotides spanned the Tt_TERT gene, with each alternate oligonucleotide representing the sense and antisense strands of the gene, respectively, and a 20-nt overlap between them. A combination of two approaches was used to assemble the gene. In the first approach, adjacent pairs of 100-mer oligonucleotides were amplified by polymerase chain reaction (PCR) utilizing short (24 nt) oligonucleotides complementary to the ends of the 100-mers, and the enzyme Pwo polymerase (Roche Molecular Biochemicals). The resulting PCR products were combined in pools of adjacent pairs and subjected to another round of PCR using the short primers. This process was repeated for 5 rounds of PCR, resulting in a single product of 2 kilobases representing the 5Ј-half of the gene, which was cloned into the BamHI and HindIII sites of the vector pCR-Script TM (Stratagene). In the second approach, all 42 100-mers were combined in one tube, extended with Klenow polymerase (New England Biolabs), ligated, and run on a 1% agarose gel. The smear at around 600 -800 base pairs was cut out of the gel, purified with a gel extraction kit (Qiagen), and amplified with short PCR primers, spaced about 600 base pairs apart in the 3Ј-half of the gene sequence. The resulting fragments were digested with restriction enzymes, ligated, and cloned into the HindIII and HincII sites of pCR-Script TM . The inserts from the two plasmid constructs were combined to form the complete gene. Sequencing of the entire insert revealed 19 PCR-induced errors, which were repaired by site-directed mutagenesis (39 -41). The insert was then subcloned into the BamHI and XhoI sites of the pET-28a expression vector (Novagen).
Mutant versions of synthetic Tt_TERT in pET-28a were constructed by site-directed mutagenesis (39 -41). The presence of the mutation and the lack of any second site mutations were confirmed by sequencing.
In Vitro Translation of Tt_TERT-Tt_TERT was expressed in rabbit reticulocyte lysates using the TNT Coupled Reticulocyte Lysate kit (Promega). In a typical 400-l reaction, 8 g of pET-28a-TERT DNA was incubated with 432 ng (8 pmol) in vitro transcribed telomerase RNA (see below), 32 l of [ 35 S]methionine (1175 Ci/mmol, 10 Ci/l; FIG. 1. A model for the mechanism of Tetrahymena telomerase action. A, TERT and its associated RNA subunit. The sequence of the template portion of T. thermophila telomerase RNA is shown. B, telomerase binds to a telomere end (in vivo) or a telomere-like DNA primer (in vitro). Binding involves the nucleotides at the 3Ј end of the primer base pairing with the RNA template. Telomerase then catalyzes addition of nucleotides onto the primer until the end of the template is reached. C, after extension, the newly elongated product may either dissociate from telomerase, or translocate to the beginning of the template (D) allowing another round of elongation. Solid arrows indicate processive synthesis; primer dissociation steps indicated by dashed arrows decrease processivity. HIV-1 is human immunodeficiency virus 1 retroviral reverse transcriptase, Sc_a1 is a reverse transcriptase involved in yeast mitochondrial group II intron transposition, and Bm_R1 is a reverse transcriptase involved in non-LTR retrotransposition (references in Ref. 62). The TERT consensus sequence includes amino acids that occur in at least 7 of the 9 TERTs. * represents amino acid similarity. The two consecutive aspartates are necessary for telomerase catalysis. The boxed amino acid was the one mutated in Tt_TERT in this study (Leu to either Tyr, Ala, or Phe). NEN Life Science Products Inc.), and other kit components as recommended by the manufacturer, at 30°C for 60 min. Translation reactions were flash-frozen in liquid nitrogen and stored at Ϫ80°C. The telomerase complex was either used in a telomerase assay directly from the translation reaction, or first purified on agarose beads coupled to antibodies to the T7-tag encoded by the pET-28a expression vector. Immunopurification was carried out as follows. T7-tag antibody-agarose beads (Novagen; 100 l) were washed 4 times in 1.5 ml of Wash Buffer-100 (20 mM Tris acetate, pH 7.5, 10% glycerol, 1 mM EDTA, 5 mM MgCl 2 , 0.1% Nonidet P-40, 1 mM DTT, 100 mM potassium glutamate), centrifuging at 1,500 ϫ g for 2 min between washes. The beads were incubated twice with 1 ml of Blocking Buffer (20 mM Tris acetate, pH 7.5, 10% glycerol, 1 mM EDTA, 5 mM MgCl 2 , 0.1% Nonidet P-40, 1 mM DTT, 100 mM potassium glutamate, 0.5 mg/ml lysozyme, 0.5 mg/ml bovine serum albumin, 0.05 mg/ml glycogen, 0.1 mg/ml yeast RNA) for 15 min at 4°C with agitation. Reticulocyte lysate translation reaction (350 l) was added to 350 l of Blocking Buffer and centrifuged at 16,000 ϫ g for 10 min at 4°C to remove any particulates. The supernatant from that spin was added to the 100 l of blocked beads and agitated at 4°C for 2 h. The beads were washed 4 times in 1.3 ml of Wash Buffer-300 (20 mM Tris acetate, pH 7.5, 10% glycerol, 1 mM EDTA, 5 mM MgCl 2 , 0.1% Nonidet P-40, 1 mM DTT, 300 mM potassium glutamate), 2 times in 1.3 ml of TMG (10 mM Tris acetate, pH 8.0, 1 mM MgCl 2 , 10% glycerol), and resuspended in 100 l of TMG to make a 1:1 slurry. To check protein recovery, 1.5 l of the slurry or 2.5 l of the original translation reaction was added to Laemmli's sample buffer (125 mM Tris-Cl, pH 6.8, 4% SDS, 0.005% bromphenol blue, 20% glycerol, 0.72 M ␤-mercaptoethanol), heated to 100°C for 3 min, and electrophoresed in an 8% polyacrylamide/SDS gel (42). The gel was fixed in 25% isopropyl alcohol, 10% acetic acid for 30 min, dried at 80°C, and exposed to a PhosphorImager screen overnight. The amount of fulllength protein was calculated by comparison to [ 35 S]methionine standards spotted onto the dried gel (spots were placed on the front and back of the gel and their intensity averaged). Immunopurification typically recovered 15-25% of the full-length TERT protein present in the original translation reaction, resulting in a final full-length TERT concentration of 2-10 nM in the 1:1 bead slurry.
In Vitro Transcription of Telomerase RNA-A plasmid containing the Tetrahymena telomerase RNA gene, a promoter for T7 RNA polymerase and a hammerhead ribozyme self-cleavage domain to process the 5Ј end of the RNA was constructed as described (43). This plasmid was digested with the enzyme Ear1, which cuts at the 3Ј end of the telomerase RNA gene. Linearized plasmid (25 g) was transcribed in a 500-l reaction containing 40 mM Tris-Cl, pH 7.5, 12 mM MgCl 2 , 10 mM DTT, 2 mM spermidine, 1 mM each rNTP, and 25 l of T7 RNA polymerase (1 mg/ml), at 37°C for 60 min. Hammerhead cleavage was stimulated by the addition of an extra 10 mM MgCl 2 and incubation at 50°C for 30 min. The reaction was centrifuged briefly at 16,000 ϫ g to remove the precipitate, ethanol precipitated, and resuspended in 100 l of loading buffer (0.05% bromphenol blue, 0.05% xylene cyanol in formamide). The entire reaction was heated to 70°C for 10 min, loaded on a 1.5-mm thick 4% polyacrylamide, 8 M urea gel and electrophoresed at 25 W for 1 h. UV shadowing allowed visualization of the band representing processed RNA, which was then excised and the RNA was eluted in 1 ml of TEN (10 mM Tris-Cl, pH 7.5, 1 mM EDTA, 0.25 M NaCl) at 4°C overnight before ethanol precipitation.
Tetrahymena Extract Preparation-T. thermophila strain B7 (from the American Type Culture Collection) was grown vegetatively and harvested as described (14). S100 extracts were made from the cells as described (6). The extracts were partially purified on a 2-ml DEAEagarose (Bio-Rad) column. The column was equilibrated with TMG before loading 5 ml of S100 extract, washed with 6 ml of TMG plus 0.2 M potassium glutamate, and eluted with 12 ml of TMG plus 0.35 M potassium glutamate in 1-ml fractions. Fractions containing active telomerase were pooled, dialyzed against TMG at 4°C overnight, flashfrozen, and stored at Ϫ80°C.
Telomerase Assay-A standard telomerase assay included 10 l of partially purified Tetrahymena extract, reticulocyte lysate translation reaction, or 1:1 slurry of immunopurification beads (see above). Where different translation reactions or immunopurifications were being assayed in the same experiment, the amounts used were adjusted in order to equalize their TERT protein contents. Typical reactions contained full-length TERT to a concentration of 1-5 nM. A standard 20-l telomerase assay included 1 ϫ Telomerase Buffer (50 mM Tris-Cl, pH 8. The reaction was incubated at 30°C for 60 min and terminated by the addition of 80 l of TES (50 mM Tris-Cl, pH 8.0, 20 mM EDTA, 0.2% SDS). The reaction was phenol/chloroform extracted, ethanol precipitated, and resuspended in 5 l of loading buffer (0.05% bromphenol blue, 0.05% xylene cyanol in formamide). A 2.5-l portion was loaded on a 12% polyacrylamide, 8 M urea sequencing gel and electrophoresed in 1 ϫ TBE at 50 watts for 2 h. The gel was dried at 80°C and exposed overnight to a PhosphorImager screen. As a control for recovery and loading, a 100-mer oligonucleotide (or an 18-mer for assays of Tetrahymena extracts) was end-labeled with [␥-32 P]ATP and polynucleotide kinase (New England Biolabs), and 5,000 cpm was added to the reaction prior to phenol/chloroform extraction. As a size marker, (G 4 T 2 ) 3 primer was labeled with terminal deoxynucleotidyltransferase (Amersham Pharmacia Biotech) and either [␣-32 P]dGTP or [␣-33 P]ddGTP, and 5,000 or 10,000 cpm, respectively, were loaded on the gel.
For the pulse-chase telomerase assays (Figs. 4 A and E, and 7, A and D) a 5Ј biotinylated primer ((G 4 T 2 ) 3 ) (Integrated DNA Technologies) was used as a substrate. The biotinylated primer (250 nM) was incubated with immunopurified telomerase (the volume adjusted to result in equal protein concentrations) under standard assay conditions (1 ϫ Telomerase Buffer (see above), 100 M dTTP, and 10 M [␣-32 P]dGTP, 80 Ci/mmol) for 2 min at 30°C. A "chase" of 40 M non-biotinylated (G 4 T 2 ) 3 or (G 4 T 2 ) 5 G primer was then added, and incubation at 30°C continued for a total of 10, 20, or 30 min. A control with no chase and a control in which non-biotinylated chase primer was added prior to incubation were both incubated for 30 min at 30°C. All reactions were terminated with 80 l of TES, extracted with 100 l of phenol/chloroform, and the biotinylated primer was recovered on magnetic Streptavidin Dynabeads (Dynal) as follows. Dynabead slurry (50 l; 10 mg/ml) was prewashed twice in 50 l of 0.1 M NaOH, 0.05 M NaCl, once in 50 l of 0.1 M NaCl, once in 50 l of 2 ϫ Binding Buffer (10 mM Tris-Cl, pH 7.5, 1 mM EDTA, 2 M NaCl) and then resuspended in 100 l 2 ϫ Binding Buffer. The 100 l of phenol-extracted telomerase reaction was added to the Dynabeads and incubated at room temperature for 10 min with rotation. The beads were then washed 3 times with 100 l of 2 ϫ Binding Buffer, resuspended in 5 l of formamide loading buffer, and heated at 95°C for 5 min to release the biotinylated products; 2.5 l was electrophoresed as above. As a recovery and loading control, a 3Ј biotinylated 12-mer primer was end-labeled with polynucleotide kinase and 5,000 cpm was added to the reaction prior to phenol/chloroform extraction.
The 5Ј biotinylated primer ((G 4 T 2 ) 3 ) was also used as a substrate for the bind-and-chase telomerase assay (Figs. 4C and 7C). It was prebound at 250 nM with immunopurified telomerase in 1 ϫ Telomerase Buffer for 15 min at 30°C. Non-biotinylated ((G 4 T 2 ) 3 ) chase primer (40 M) and nucleotides dTTP and [␣-32 P]dGTP (concentrations as in standard assay) were then added (or nucleotides alone for the no-chase controls). The reaction was incubated for a further 10 or 30 min, terminated, and the biotinylated products were isolated as above.
Primer Dissociation Assay-A modified telomerase assay was used to measure the rate of dissociation of primer from Tt_TERT (Fig. 8). An 18-mer oligonucleotide (GT 2 (G 4 T 2 ) 2 G 3 ) was preincubated at a concentration of 100 nM with 10 l of immunopurified telomerase (1:1 bead slurry) in 1 ϫ Telomerase Buffer at 30°C for 15 min. A 21-mer competitor primer (GT 2 (G 4 T 2 ) 3 ) was added (final concentration 20 M) and incubated with the enzyme/primer mixture at 30°C for varying lengths of time (see Fig. 8). Telomerase activity was initiated by the addition of nucleotides (ddTTP at 100 M and [␣-32 P]dGTP at 10 M) and allowed to extend the primers at 30°C for 10 min. As a control to ensure that the excess of 21-mer did not allow the enzyme to rebind 18-mer, both primers were preincubated with enzyme simultaneously prior to telomerase extension. The reactions were terminated and electrophoresed as above. To determine k off , the intensity of the major (larger) extension band (indicated by * in Fig. 8A) was quantitated, normalized to the combined intensities of all three major extension products (18-and 21-mer) and plotted against time. The data were fit to the equation y ϭ ae Ϫkt ϩ b, where k ϭ k off and b represents a background in the experiment of about 10%.

Reconstitution and Characterization of Telomerase Activity
in Vitro-It has previously been demonstrated that Tetrahymena TERT protein, when coexpressed with Tetrahymena telomerase RNA in rabbit reticulocyte lysates, forms a complex capable of telomerase activity (15) and that formation of this complex is dependent on factor(s) present in the reticulocyte lysate (22). We were also able to reconstitute telomerase activ-ity in vitro (Fig. 3A). In ciliates the codons TAA and TAG encode glutamine, whereas these are stop codons in the "universal" genetic code. A synthetic Tt_TERT gene was constructed both to correct these Glu codons and to adjust the codon usage to more closely resemble that of rabbit. The synthetic gene was cloned into the pET-28a expression vector. When expressed in rabbit reticulocyte lysates in the presence of [ 35 S]methionine, a major protein band of ϳ130 kDa was observed, corresponding to full-length protein. In addition, several smaller proteins were visible, representing either products of proteolytic degradation or proteins resulting from internal translation start sites (Fig. 3B). The protein was expressed in the presence of in vitro transcribed telomerase RNA, and telomerase activity was assayed in the presence of a (G 4 T 2 ) 3 telomeric primer, dTTP and [␣-32 P]dGTP. A ladder of primer extension products with the 6-nt pausing pattern characteristic of Tetrahymena telomerase was observed upon denaturing polyacrylamide gel electrophoresis (Fig. 3A, lane 1). Activity was abolished by the use of either a telomerase RNA lacking the template domain (lane 2) or a TERT protein with the catalytic aspartate in motif A substituted by alanine (lane 3), demonstrating that the activity observed is indeed telomerase activity.
The pET-28a-TERT plasmid encodes an 11-amino acid tag (T7-tag) at the NH 2 terminus of the TERT protein. Agarose beads coupled to antibodies to this tag were used to immunoaffinity purify the telomerase complex from the reticulocyte lysate after expression, which resulted in an enhancement of the proportion of full-length TERT (Fig. 3B, Beads). This immunopurification step resulted in an increase in apparent processivity of telomerase (Fig. 3A, lanes 4 and 5), as observed by the increase in intensity of larger products relative to the ϩ1 product. This increase in apparent processivity was eliminated by the addition of rabbit reticulocyte lysate to the immunopurified telomerase, indicating that factor(s) in the reticulocyte lysate inhibit telomerase processivity (Fig. 3A, lane 6).
In the standard telomerase assay, the multiple repeat products are typically thought to result from processive extension (i.e. the product remains associated with the same enzyme molecule), and this has been demonstrated to be the case for Tetrahymena telomerase from extracts (27). However, such a pattern could also arise from a distributive mode of action (i.e. the products dissociate from the enzyme and rebind for further addition). To distinguish between these possibilities for reconstituted telomerase, we used a pulse-chase variation of the telomerase assay similar to that described by Maine et al. (44). A biotinylated (G 4 T 2 ) 3 substrate primer was incubated with immunopurified telomerase for 2 min under standard reaction conditions. A 160-fold excess of non-biotinylated (G 4 T 2 ) 3 competitor primer was then added, and the reaction continued for a total of 10, 20, or 30 min. The biotinylated products were then isolated using magnetic streptavidin beads and electrophoresed (Fig. 4A). Only the biotinylated products were recovered; a control using only non-biotinylated substrate primer showed no recovered products (data not shown). When competitor primer was added at the beginning of the reaction it almost completely prevented extension of biotinylated primer molecules (Fig. 4A,  lane 6), demonstrating that the concentration of competitor primer was sufficient to prevent rebinding of biotinylated primer during the chase period. Thus, the products visible on the gel represent only processive synthesis. The size of products increased between the 2-min pulse reaction and the chased reactions, indicating that the enzyme is truly processive, albeit to a modest extent (Fig., 4A, lanes 2-4). A control in which no competitor primer was added during a 30-min reaction showed a greater intensity of products, indicating that the enzyme can carry out multiple turnovers of the processive elongation reaction (lane 5). A graphical representation of the processivity in this experiment is shown in Fig. 4B, where the intensity of each successive repeat is expressed relative to the intensity of the first repeat. (It should be noted that longer products are less abundant than they appear from their intensities, since they incorporate increasing amounts of radiolabeled nucleotides and hence have a higher specific activity. Thus, the data in Fig. 4B and all subsequent such graphs have been adjusted for the specific activity of each repeat band, i.e. the intensity of the second repeat was divided by 5, that of the third repeat divided by 9, and so on.) The data are linear on a semi-log plot, with the slope of the line being inversely related to the processivity. The processivity increased between the 2-min pulse and the 10-min FIG. 3. In vitro telomerase activity assay. A, the telomerase protein-RNA complex was reconstituted in vitro in rabbit reticulocyte lysates (reconstituted) or partially purified from cellular extracts of vegetatively growing Tetrahymena (extract) and was incubated with telomeric primer ((G 4 T 2 ) 3 ), dTTP, and [␣-32 P]dGTP. Lane 1, wt TERT protein was coexpressed with wt telomerase RNA. Lane 2, a control in which the telomerase RNA lacked its template region. Lane 3, a control in which TERT amino acid 618 (the catalytic aspartate in motif A) was mutated to alanine. Lanes 4 -6, TERT protein was coexpressed with telomerase RNA in reticulocyte lysates and then subjected to immunopurification using an antibody to a T7-tag present at the N terminus of TERT. Lane 4, prior to immunopurification (input). Lane 5, immunopurified telomerase (beads). Lane 6, 5 l of rabbit reticulocyte lysate was added back to the beads after immunopurification. Lanes 7 and 8, activity from Tetrahymena extract, without and with 5 l of added reticulocyte lysate, respectively. Lane M, ddG: size marker, primer (G 4 T 2 ) 3 labeled with terminal deoxynucleotidyltransferase and [␣-33 P]ddGTP. Lane M, dG: size marker, primer (G 4 T 2 ) 3 labeled with terminal deoxynucleotidyltransferase and [␣-32 P]dGTP. LC, loading control; end-labeled 100-mer oligonucleotide for reconstituted telomerase, 12-mer for extracts. B, SDS-polyacrylamide gel electrophoresis showing the in vitro expressed TERT protein used in the experiments in A. TERT was expressed in reticulocyte lysates in the presence of [ 35 S]methionine and the telomerase complex was immunopurified. Equivalent portions of the translation reaction (input), supernatant (sup), and immunopurified material (beads) were electrophoresed on a 8% SDS-polyacrylamide electrophoresis gel which was then dried and exposed to a PhosphorImager screen. Full-length TERT protein is indicated. chased reaction, but showed no further increase and was identical to that of the unchased reaction (Fig. 4B).
A variation on the pulse-chase assay was carried out to confirm that reconstituted telomerase is processive. In this bind-and-chase experiment, immunopurified enzyme was allowed to bind to biotinylated primer at 30°C for 15 min. After binding, a 160-fold excess of competitor non-biotinylated primer and nucleotides were added, and the reaction was allowed to proceed for a further 10 or 30 min (Fig. 4C, ϩ chase  lanes). The pattern of extension was compared with controls in which competitor primer was omitted (Fig. 4C, Ϫ chase lanes). Since the chase was almost completely effective (Fig. 4C, control lane), any products larger than ϩ1 appearing in the chased reactions represent processive synthesis. Once again, reconstituted telomerase displayed true processivity.
The results in Fig. 4, A and C, could also be obtained if telomerase utilized an "efficiency enhanced" distributive mechanism (45), i.e. if partially elongated primers had a greater affinity for telomerase and thus became preferentially elongated even in the presence of excess chase primer. To investigate this possibility, we incubated telomerase with the original 18-mer primer together with various ratios of a 31-mer primer ((G 4 T 2 ) 5 G) that is identical to the largest product visible from the 2-min pulse reaction in Fig. 4A. Primers with four or more G-rich repeats have the potential to form G-quartet secondary structures which can inhibit telomerase activity (46). To minimize formation of these structures, primers were heated to 95°C prior to use, and the reaction buffer contained no monovalent cations, which induce G-quartet formation (47). When the two primers were incubated at a ratio of 1:1 and extended by telomerase with [␣-32 P]dGTP and ddTTP, both primers were elongated (Fig. 4D). The presence of ddTTP terminates the reaction after the addition of 4 and 5 nucleotides to the 31-and 18-mer, respectively, allowing the extension products of each to be distinguished. The fact that the 18-mer can be extended in the presence of an equal concentration of the 31-mer indicates that the latter does not have a greater affinity for telomerase. When the 18-mer is present at a 160-fold excess over the 31-mer, a situation mimicking the chase in Figs. 4, A and C, the 31-mer is not visibly extended, indicating that the extension products seen in the chased reactions are not due to a greater affinity for longer substrates. The 31-mer in this experiment probably partially exists in a G-quartet structure, since it was extended less efficiently than the 18-mer in a 1:1 ratio (Fig.  4D). However, the conditions in this experiment and the pulsechase experiments are identical and hence any long products would be expected to form the same structure as the 31-mer primer.
To confirm the above conclusion, we also carried out a pulsechase experiment using the 31-mer as the chase primer (Fig.  4E). A biotinylated 18-mer was extended for 2 min, followed by incubation with a 160-fold excess of nonbiotinylated 31-mer for a total of 30 min and recovery of the biotinylated products. The chase was effective (lane 4), and once again the chased reaction showed an increase in size of products over the 2-min pulse (lanes 1 and 2). Thus we conclude that the processivity observed on telomerase assay gels, while limited in extent, does represent true processivity.
An increase in dGTP concentration in the telomerase assay has been shown to increase processivity in Tetrahymena, Euplotes, and mammalian systems (15,44,48,49), an effect that added simultaneously at the start of the reaction. The reaction products in all lanes were subsequently isolated on magnetic streptavidin beads and hence represent extension products of the biotinylated primer only. LC, biotinylated 12-mer DNA loading and recovery control.

FIG. 4. Telomerase assays with chase primers to test processivity of wt reconstituted telomerase.
A, pulse-chase assay. Reconstituted and immunopurified wt telomerase was incubated with biotinylated (G 4 T 2 ) 3 primer, dTTP, and [␣-32 P]dGTP for a 2-min pulse at 30°C (lane 1). An excess of non-biotinylated (G 4 T 2 ) 3 competitor primer was added and the reactions continued for a total of 10, 20, or 30 min (lanes 2-4). Lane 5, 30 min incubation with no competitor primer. Lane 6, control for effectiveness of the chase; competitor primer and biotinylated primer were added simultaneously at the start of the reaction. The reaction products in all lanes were subsequently isolated on magnetic streptavidin beads and hence represent extension products of the biotinylated primer only. LC, biotinylated 12-mer DNA loading control, which also controls for the efficiency of binding to the streptavidin beads. B, quantitation of processivity of wt telomerase in pulse-chase assay. The intensity of the first 5 major repeat bands in lanes of gel in A were normalized to the intensities of the 1st repeat, adjusted for specific activity (the intensity of the second repeat was divided by 5, the third by 9, the fourth by 13, the fifth by 17, and the sixth by 21) and plotted versus repeat number. ⌬, 2 min unchased; ࡗ, 10 min chased; OE, 20 min chased; q, 30 min chased; Ⅺ, 30 min unchased. C, bind and chase assay. Immunopurified wt telomerase was prebound to biotinylated (G 4 T 2 ) 3 primer for 15 min. An excess of non-biotinylated competitor primer was added (ϩ chase lanes) along with dTTP and [␣-32 P]dGTP, and the reaction incubated for a further 10 or 30 min. Ϫ, chase lanes: no competitor primer added. Control lane, competitor primer and biotinylated primer added simultaneously at the start of the reaction. LC, biotinylated 12-mer loading and recovery control. D, competition between two primers. An 18-mer ((G 4 T 2 ) 3 ) and a 31-mer ((G 4 T 2 ) 5 G) primer were combined at the indicated ratios (using 250 nM, 4 M, or 40 M primer) and incubated for 60 min with immunopurified wt telomerase, [␣-32 P]dGTP, and ddTTP. LC, 100-mer loading and recovery control. E, pulse-chase with 31-mer ((G 4 T 2 ) 5 G) as the chase primer. Immunopurified wt telomerase was incubated with biotinylated (G 4 T 2 ) 3 primer, dTTP, and [␣-32 P]dGTP for a 2-min pulse at 30°C (lane 1). An excess of non-biotinylated ((G 4 T 2 ) 5 G) competitor primer was added and the reaction continued for a total of 30 min (lane 2). Lane 3, 30-min incubation with no competitor primer. Lane 4, control for effectiveness of the chase; competitor primer and biotinylated primer were may be due to stimulation of enzyme translocation by dGTP (50). Since we observed that immunopurification of telomerase from reticulocyte lysates increased its processivity, we therefore determined whether immunopurified telomerase was still responsive to changes in dGTP concentration. Our immunopurified reconstituted Tetrahymena telomerase did demonstrate an increase in processivity with dGTP concentration, as did native activity from a Tetrahymena cell extract (Fig. 5A). A graphical representation of the processivity of reconstituted telomerase is shown in Fig. 5B, where the intensity of each successive repeat is expressed relative to the intensity of the first repeat after adjusting for specific activity. A dGTP concentration of 10 M, which gives maximum processivity with the reconstituted enzyme, was chosen as the standard concentration for all other experiments in this study. The processivity at 10 M dGTP varies slightly in different experiments, but on average each band is about 30% as intense as the previous one, indicating that the efficiency of continuing extension from one repeat to the next is about 30%. The average product length synthesized before one-half of the enzyme molecules have dissociated or stalled is approximately 0.3 repeats, or about 2 nt.
Telomerase activity from extracts was much more processive than the reconstituted enzyme at all dGTP concentrations (Fig.  5A). Previous studies have estimated that telomerase from Tetrahymena extracts synthesizes about 521 nt before one-half of the enzyme dissociates (27), although this calculation did not take into account the greater specific activity of longer products and hence is an overestimate. Addition of reticulocyte lysate did reduce the processivity of telomerase from extracts, but the processivity remained much higher than that of reconstituted telomerase (Fig. 3A).
In addition to affecting the processivity of telomerase, reticulocyte lysate also inhibits total telomerase activity (data not shown). Therefore, due to the impurity of the reticulocyte lysate system, most of the following experiments have been carried out using immunopurified telomerase.
Leucine-Tyrosine Mutant at Position 813 Has Increased Pro-cessivity-The L813Y mutant was constructed by site-directed mutagenesis and expressed in reticulocyte lysates. The amount of protein produced and its stability at 30°C were approximately equal to wt (Fig. 6E). In the in vitro telomerase assay, the L813Y mutant showed levels of activity equal to wt (Fig.  6A). Quantitation of the intensity of all radiolabeled products on the gel and comparison to standards of known specific activity demonstrate that both wt and L813Y enzymes synthesize approximately 0.2-0.4 fmol of extended primer per min per fmol of enzyme. However, the apparent processivity of L813Y telomerase was increased (Fig. 6A).
Since immunopurification of wt TERT increased its processivity, we next examined whether the L813Y mutant remained more processive than wt after immunopurification (Fig. 6B, compare "beads" lanes). The results are represented graphically in Fig. 6C, which shows the intensity of each repeat band relative to the first repeat after adjusting for specific activity. Mutant L813Y was more processive than wt both before and after immunopurification. After immunopurification, the average product length synthesized before one-half of the enzyme molecules had dissociated or stalled was approximately 0.6 repeats, or about 4 nt. Thus, this mutant is about twice as processive as wt telomerase.
Position Leu-813 was also mutated to alanine and phenylalanine. Fig. 6, B and D, show that L813A is less processive than wt, while L813F has a processivity intermediate between wt and L813Y. The expression levels and stability of all mutant proteins were approximately equal to wt (Fig. 6E).
Mutant L813Y Also Demonstrates True Processivity-It seemed unlikely that the increased apparent processivity of mutant L813Y could result from a switch to a distributive mode of action. Nevertheless, pulse-chase and bind-and-chase assays identical to those already described for wt telomerase were performed to test for processivity. A 2-min pulse with a biotinylated (G 4 T 2 ) 3 primer (Fig. 7A, lane 1) was followed by a chase with a 160-fold excess of nonbiotinylated (G 4 T 2 ) 3 primer (lanes 2-4) and recovery of the biotinylated products on streptavidin magnetic beads. Since the chase was effective (lane 6), the increase in size of products is due to a processive mode of action. Graphical representation of the processivity of the reactions indicates that, as for wt enzyme, the chased reactions have a processivity identical to the unchased control (Fig. 7B). A bind-and-chase assay in which the L813Y enzyme was prebound to biotinylated (G 4 T 2 ) 3 primer followed by extension in the presence of an excess of non-biotinylated (G 4 T 2 ) 3 primer (Fig. 7C) confirmed the processivity of the reaction. Finally, to confirm that these results were not due to an increased affinity for longer substrates, a pulse-chase experiment was performed with an excess of nonbiotinylated 31-mer ((G 4 T 2 ) 5 G) as the chase primer (Fig. 7D). Once again, an increase in size of products was seen after 30 min in the presence of the long chase primer (lane 2).
Primer Dissociates More Slowly from Mutant L813Y Than from wt Telomerase-We postulated that the greater processivity of mutant L813Y might result from an increased affinity for primer, resulting in a greater probability of the enzyme translocating for another round of synthesis rather than dissociating (Fig. 1). As a measure of primer affinity for enzyme, we used a modified telomerase assay to determine the rate constant for primer dissociation (k off ). An 18-mer primer (GT 2 (G 4 T 2 ) 2 G 3 ) was preincubated with immunopurified wt or L813Y telomerase at 30°C for 15 min. A 200-fold excess of 21-mer competitor primer (GT 2 (G 4 T 2 ) 3 ) was added and incubated with the enzyme/primer mixture at 30°C for varying lengths of time. Telomerase activity was then initiated by the addition of nucleotides (ddTTP and dGTP) and allowed to proceed at 30°C for 10 min. The inclusion of ddTTP resulted in termination of extension after incorporation of the first T, i.e. after 2 nucleotides for the 18-mer, and after 5 nucleotides for the 21-mer (Fig.  8A). As a control to ensure that the excess of 21-mer did not allow the enzyme to rebind 18-mer, both primers were preincubated with enzyme simultaneously prior to telomerase extension; there was almost no detectable extension of the 18-mer (Fig. 8A, chase control lanes). Thus, the amount of addition to the 18-mer after varying lengths of chase time is a reflection of the amount of the original primer still bound to enzyme at that time. Fig. 8A shows that the 18-mer dissociates rapidly from the wt enzyme (it is almost entirely dissociated after 10 min), but more slowly from the L813Y enzyme. The intensity of the major (larger) extension band (indicated by * in Fig. 8A) was quantitated, normalized to the combined intensities of all three major extension products (18-and 21-mer) to control for differences in enzyme activity between lanes, and plotted against time (Fig. 8B). The data fit well to a single exponential, giving a k off of 0.68 Ϯ 0.08 min Ϫ1 for wt and 0.21 Ϯ 0.02 min Ϫ1 for L813Y. These correspond to half-lives of the primer-enzyme complexes of 1.0 Ϯ 0.2 min for wt and 3.3 Ϯ 0.3 min for L813Y. Thus primer dissociates approximately 3-fold more slowly from the L813Y mutant than from wt enzyme. DISCUSSION Telomerase reverse transcriptases are related to retrotransposon and retroviral RTs (9,23), yet telomerase has a distinct mode of action including repetitive copying of its template RNA and relatively low processivity. The protein structural features that account for these differences are completely unknown. We now identify an amino acid in the conserved motif C of TERT that affects the intrinsic processivity of the enzyme.
Mutating a leucine to a tyrosine at a position that is likely to be close to the active site caused Tetrahymena telomerase to increase in processivity. We demonstrated that both wt telomerase and the mutant show true processivity rather than distributive action. The 2-fold increase in processivity correlated with a 3-fold slower dissociation rate of primer from the enzyme. Although these are modest increases, it is rare for a site-directed mutation to "improve" an enzyme that has undergone natural selection over evolutionary time. Indeed, none of 28 other site-directed amino acid substitutions we have tested in Tetrahymena telomerase increased either activity or processivity. 2 Our results, together with the crystal structure of the homologous region from HIV RT, suggest a possible mechanism for the increase in processivity. The side chain of tyrosine 813 could be hydrogen bonding with bases at the end of the primer as well as engaging in hydrophobic interactions with the primer, as the equivalent tyrosine 183 does in HIV RT (37). These interactions may increase the affinity of the enzyme for primer and make it more likely that the primer will undergo translocation for another round of extension rather than dissociating (see Fig. 1). This conclusion is supported by the observation that changing the same amino acid to phenylalanine resulted in intermediate processivity and changing it to alanine resulted in reduced processivity, since this emulates the pattern of activity displayed by these amino acids at the equivalent position in HIV RT.
All variants of telomerase tested showed approximately equal amounts of activity and were affected only in their pro-2 T. Bryan, K. Goodrich, and T. Cech, manuscript in preparation.
FIG. 6. Telomerase activity of mutants at position 813 of Tt_TERT. A, telomerase reconstituted with either wt or L813Y TERT in reticulocyte lysates was assayed with or without substrate primer. Lane M, size marker, (G 4 T 2 ) 3 labeled with [␣-32 P]dGTP and terminal deoxynucleotidyltransferase. LC, 100-mer loading control. B, telomerase was reconstituted with wt TERT or the mutants indicated and assayed for telomerase activity prior to (input) or after (beads) immunopurification from the reticulocyte lysates. The assay was also carried out with immunopurified telomerase with 5 l of reticulocyte lysate added (Bϩretic). C, comparison of processivity of wt and L813Y telomerase before and after immunopurification. The intensities of the first 6 major repeat bands in lanes of gel in B were normalized to the intensity of the first repeat, adjusted for specific activity (the intensity of the second repeat was divided by 5, the third by 9, the fourth by 13, the fifth by 17, and the sixth by 21) and plotted versus repeat number. f, wt before immunopurification; q, wt after immunopurification; Ⅺ, L813Y before immunopurification; E, L813Y after immunopurification. D, quantitation of processivity of L813 mutants after immunopurification, as in C. f, wt; Ⅺ, L813Y; ࡗ, L813A; ⌬, L813F. E, SDS-polyacrylamide electrophoresis of wt and mutant TERT proteins. Each protein was translated in rabbit reticulocyte lysate for 3 h at 30°C and an aliquot was removed after 1, 2, or 3 h. cessivity. In contrast, mutants at this position in other RTs were greatly affected in their activity levels (31)(32)(33)(34). This difference may be due to additional stabilization of the enzymeprimer complex provided by other regions of telomerase such as the anchor region (51,52). Thus, loss of the primer interaction provided by the tyrosine in HIV RT is catastrophic for enzyme activity, while in telomerase this loss may be compensated by other primer-enzyme interactions.
It is difficult to directly compare the levels of processivity reached by the tyrosine mutant of telomerase and HIV RT since in the former case processive synthesis involves repeated translocations on a short region of template, whereas HIV RT normally acts continuously along a template molecule. The processivity of HIV RT also varies greatly depending on the template sequence (53, 54) but can reach up to 300 nt on an RNA template in vitro (53). In the in vitro reconstitution assay described here, the tyrosine mutant of TERT more closely resembles classical RTs in this respect than does wt TERT.
Neither the wt nor the tyrosine mutant of reconstituted telomerase was as processive as telomerase from Tetrahymena extracts. The reason for the difference in processivity of reconstituted and native Tetrahymena telomerase is unknown, but may be due to factors that copurify with telomerase from extracts that are missing in reticulocyte lysates. Indeed, differences in purification schemes have resulted in differences in processivity of telomerase from S. cerevisiae and Euplotes crassus (55,56). There is also precedence for processivity increasing factors associated with other polymerases, e.g. the processivity of T7 DNA polymerase is increased by the associated protein thioredoxin (57). Tetrahymena telomerase is not very processive in vivo, however, synthesizing only a few repeats before dissociating (58), suggesting that the extreme processivity of Tetrahymena telomerase from extracts may itself be an in vitro artifact. FIG. 7. Telomerase assays with chase primers to test processivity of L813Y reconstituted telomerase. A, pulse-chase assay. Reconstituted and immunopurified L813Y telomerase was incubated with biotinylated (G 4 T 2 ) 3 primer, dTTP, and [␣-32 P]dGTP for a 2-min pulse at 30°C (lane 1). An excess of non-biotinylated (G 4 T 2 ) 3 competitor primer was added and the reactions continued for a total of 10, 20, or 30 min (lanes 2-4). Lane 5, 30-min incubation with no competitor primer. Lane 6, control for effectiveness of the chase; competitor primer and biotinylated primer were added simultaneously at the start of the reaction. The reaction products in all lanes were subsequently isolated on magnetic streptavidin beads and hence represent extension products of the biotinylated primer only. LC, biotinylated 12-mer DNA loading control, which also controls for the efficiency of binding to the streptavidin beads. B, quantitation of processivity of L813Y telomerase in pulse-chase assay. The intensities of the first 5 major repeat bands in lanes of gel in A were normalized to the intensity of the 1st repeat, adjusted for specific activity (the intensity of the second repeat was divided by 5, the third by 9, the fourth by 13, the fifth by 17, and the sixth by 21) and plotted versus repeat number. ⌬, 2 min unchased; ࡗ, 10 min chased; OE, 20 min chased; q, 30 min chased; Ⅺ, 30 min unchased. C, bind and chase assay. Immunopurified L813Y telomerase was prebound to biotinylated (G 4 T 2 ) 3 primer for 15 min. An excess of non-biotinylated competitor primer was added (ϩ chase lanes) along with dTTP and [␣-32 P]dGTP, and the reaction incubated for a further 10 or 30 min. Ϫ chase lanes, no competitor primer added. Control lane, competitor primer and biotinylated primer added simultaneously at the start of the reaction. LC, biotinylated 12-mer loading and recovery control. D, pulse-chase with 31-mer ((G 4 T 2 ) 5 G) as the chase primer. Immunopurified L813Y telomerase was incubated with biotinylated (G 4 T 2 ) 3 primer, dTTP, and [␣-32 P]dGTP for a 2-min pulse at 30°C (lane 1). An excess of non-biotinylated ((G 4 T 2 ) 5 G) competitor primer was added and the reaction continued for a total of 30 min (lane 2). Lane 3, 30-min incubation with no competitor primer. Lane 4, control for effectiveness of the chase; competitor primer and biotinylated primer were added simultaneously at the start of the reaction. The reaction products in all lanes were subsequently isolated on magnetic streptavidin beads and hence represent extension products of the biotinylated primer only. LC, biotinylated 12-mer DNA loading and recovery control.
FIG. 8. Rate of dissociation of primer from telomerase measured by a telomerase assay. A, reconstituted and immunopurified wt or mutant L813Y telomerase was incubated for 15 min with 18-mer oligonucleotide (GT 2 (G 4 T 2 ) 2 G 3 ). An excess of 21-mer competitor primer (GT 2 (G 4 T 2 ) 3 ) was added and incubated with the enzyme/primer mixture for 0, 0.5, 1, 2, 5, or 10 min prior to addition of nucleotides (ddTTP and [␣-32 P]dGTP). Extension was allowed to proceed for 10 min. The patterns of extension of the 21-mer primer or the 18-mer primer alone are shown (21-mer and 18-mer lanes). As a control to ensure that the excess of 21-mer did not allow the enzyme to rebind 18-mer, both primers were preincubated with enzyme simultaneously prior to telomerase extension (chase control). B, quantitation of data from experiment in A; data from two such experiments are shown. The intensity of the upper 18-mer extension band (*) was normalized to the combined intensities of the three major extension bands, plotted versus time of chase, and the data fit with a single exponential as described under "Experimental Procedures." f, wt experiment 1; q, wt experiment 2; Ⅺ, L813Y experiment 1; E, L813Y experiment 2.
The in vivo effects of this tyrosine mutation of telomerase in Tetrahymena are unknown. We predict that it would be more processive in vivo, i.e. each enzyme molecule would synthesize more telomeric repeats before dissociating from the telomere. It is unclear whether this would have an effect on steady-state telomere length, however, since there are likely to be many factors regulating telomere length besides telomerase, and there may be selective pressure in Tetrahymena cells against long telomeres (59,60).
To investigate the effects of amino acid substitutions in TERT we have used an in vitro reconstitution system based on translation in rabbit reticulocyte lysates, as has been described by others (15,19,20). Chaperone proteins or other factors in the reticulocyte lysates contribute to the assembly of the telomerase ribonucleoprotein complex (21,22). We have shown that the reticulocyte lysates also have a negative effect on telomerase activity, impacting both processivity and the total levels of in vitro activity. Immunopurification of the telomerase complex from the reticulocyte lysates abrogated these effects and thus will provide a useful system for investigation of other sitedirected mutants of telomerase.