Ever shorter telomere 1 (EST1)-dependent reverse transcription by Candida telomerase in vitro: Evidence in support of an activating function

Abstract

Telomerase is an RNA–protein complex responsible for the extension of one strand of the telomere terminal repeats. Analysis of the telomerase complex in the budding yeast Saccharomyces cerevisiae has revealed the presence of one catalytic protein subunit (Est2p/TERT) and at least two noncatalytic components (Est1p and Est3p). The TERT subunit is essential for telomerase function, both in vitro and in vivo. In contrast, the Est1p and Est3p subunits, although required for telomere extension in vivo, have not been shown to affect enzyme activity in vitro. We recently identified orthologues of the Saccharomyces telomerase subunits in Candida albicans (named CaTERT, CaEst1p, and CaEst3p). Analysis of telomerase from the Candida Caest1-Δ strains revealed a primer-specific defect in its activity in vitro: The mutant enzyme was impaired in its ability to extend some, but not all, telomeric primers. The CaEst1p-responsive primers have 3′ ends that are clustered in two loci within the 23-bp Candida telomere repeat. The degree of CaEst1p-dependence was modulated by the length and sequence of the 5′ ends. For CaEst1p-dependent primers, the wild-type enzyme consistently exhibited a greater V_max than the mutant enzyme in kinetic studies. These results suggest that CaEst1p augments the ability of telomerase to reverse-transcribe through selected barriers in the telomere repeat by acting as an allosteric activator and provide the basis for a functional in vitro assay for a noncatalytic protein component of the telomerase complex.

Telomeres are specialized nucleoprotein structures that maintain the integrity of chromosomal termini from fusion and recombination and promote chromosomal end replication (1, 2). In most organisms, telomeric DNA consists of short, repetitive sequences that are rich in G residues on the 3′ end-containing strand. These repeats are maintained by a ribonucleoprotein known as telomerase, which acts as an unusual reverse transcriptase, using a small segment of an integral RNA component as template to effect the extension of the G-rich strand of telomeres (3–6). The regulation of telomere length and telomerase activity has been shown to be pivotal in the control of cellular life span.

Components of the telomerase enzyme complex have been analyzed in a variety of organisms including ciliated protozoa, yeast, and mammals (5–7). Two components are essential for the polymerization activity: an RNA that acts as template and facilitates enzymatic function and a protein subunit that catalyzes nucleotide addition. The protein subunit, generically known as TERT (telomerase reverse transcriptase), is an ≈100- to 130-kDa, evolutionarily conserved polypeptide that contains essential reverse transcriptase-like motifs (8, 9). Extensive mutagenesis analysis of TERT using a variety of systems provides compelling support for the notion that TERT mediates nucleotide addition through a reverse transcriptase-like mechanism (10–16).

Many telomerase-associated polypeptides have been identified by using either biochemical or genetic tools, including (i) p80 and p95 from Tetrahymena and related proteins in mammals, (ii) molecular chaperones such as Hsp90 and p23, and (iii) RNA-binding proteins such as La proteins, Sm proteins, dyskerin, a Staufen homologue, and the ribosomal L22 protein (17–26). Several of these proteins have been shown to be required for telomerase ribonucleoprotein formation/stability and telomere length maintenance in vivo. However, with few exceptions, these “noncatalytic” components of telomerase do not appear to participate directly in the telomere-extension function of telomerase in vivo. Two notable exceptions to this generalization are Est1p and Est3p in the budding yeast Saccharomyces cerevisiae (27, 28). Both were identified through genetic screens and shown to act in the same pathway as telomerase RNA and TERT and to be subunits of the telomerase complex but not essential for in vitro activity (29, 30). Further studies implicate Est1p in the recruitment of the telomerase complex to telomere ends through an interaction with Cdc13p, a G strand telomere-binding protein (31, 32). In addition, genetic and chromatin immunoprecipitation studies suggest that Est1p may mediate a “postrecruitment” or “activation” function for telomerase (33, 34).

We recently identified putative orthologues of TERT as well as EST1 and EST3 (ever shorter telomeres 1 and 3) in the genome of the pathogenic fungus Candida albicans. Analysis of knockout strains indicates that the Candida homologues (named CaTERT, CaEST1, and CaEST3) are all required for normal telomere maintenance, implying that the conservation of telomerase proteins extends beyond the catalytic subunit (35). CaTERT was found to be essential for telomerase activity in vitro, consistent with a catalytic role for this protein (35). Initial studies did not reveal a large effect of CaEst1p loss on telomerase function in vitro. However, subsequent analysis performed with an extensive series of primers uncovered a primer-specific impairment of function for the telomerase derived from a Caest1-Δ strain. (For the sake of simplicity, the mutant strain was designated by one iteration of the disrupted locus, even though both alleles are deleted.) The mutant enzyme appears to have reduced abilities to catalyze nucleotide addition at two loci within the Candida telomere repeat. The magnitude of the CaEst1p effect was modulated by the length and sequence of the primer. In addition, for CaEst1p-dependent primers, the wild-type enzyme consistently exhibited a greater V_max than the mutant enzyme in kinetic studies. These results suggest that CaEst1p can act as an allosteric activator of telomerase.

Materials and Methods

Strains and Plasmids.

The C. albicans strain BWP17 [ura3Δ∷λimm434/ura3Δ∷λimm434 his1∷hisG/his1∷hisG arg4∷hisG/arg4∷hisG; a gift from A. Mitchell (Columbia University, New York)] was derived from CAI4 as described (36). The Candida Caest1-Δ strain was constructed from BWP17 by using a UAU1 cassette as described (35).

Purification of and Assay for C. albicans Telomerase.

Whole-cell extracts of C. albicans and active telomerase fractions were prepared essentially as described for S. cerevisiae (30, 37). Briefly, C. albicans cultures were grown in yeast extract/peptone/dextrose + uri (80 μg/ml) to an OD of 1.0. Cells were harvested, resuspended in trimethylglycine (TMG)-15 15(0), and lysed by vortex mixing with glass beads. Extracts were clarified by centrifugation, and the soluble fraction was loaded onto a DEAE-agarose column. The column was washed with TMG-10(400), and active telomerase fractions were obtained by eluting the column with TMG-10(900). The fractions were concentrated to ≈1 mg/ml by filtration through Microcon-30 before use. A typical telomerase reaction was carried out in a 30-μl volume containing 10 mM Tris⋅HCl (pH 8.0), 2 mM magnesium acetate, ≈150 mM sodium acetate (contributed by the protein fraction), 1 mM spermidine, 1 mM DTT, 2.5–5% glycerol (contributed by the protein fraction), 5 μM of primer oligodeoxynucleotides, 5–10 μl of column fractions, and combinations of labeled and unlabeled nucleotides (present at 0.2 and 50 μM, respectively). Primer extension products were processed and analyzed by gel electrophoresis as described (37). Signals were obtained from phosphor screens and quantified by using a PhosphorImager system (Molecular Dynamics). Kinetic analysis was performed by varying the primer concentrations and by using a time point within the linear range of product accumulation. The amounts of products were quantified by comparing the “volume” derived from reaction products with those from known amounts of radioactivity. {In control scanning experiments, we spotted known quantities of radioactive label on filter paper, exposed a PhosphorImager plate to the paper for 16 h, and scanned the plate to obtain the volumes yielded by the radioactive spots. From such experiments, we estimate that 1 fmol of label [3,000 Ci/mmol (1 Ci = 37 GBq)] gives rise to a volume of 650,000 in 16 h. Thus, we were able to calculate the molar amount of the product from the scan.} The reaction rates then were calculated (by dividing the product against the duration of the reaction) and plotted against primer concentrations, and the data points were fitted to the Michaelis–Menton equation by using a nonlinear regression algorithm (prism, GraphPad, San Diego).

Results

CaEst1p-Dependent Reverse Transcription by Candida telomerase in Vitro.

Candida telomeres have an unusual characteristic; instead of an irregular or short repeat typical of other organisms, Candida telomeres consist of a 23-bp invariant repeat (38, 39) (Fig. 1B). To determine the effect of CaEst1p loss on enzyme function, we prepared extracts from the parental strain (BWP17) and a Caest1-Δ strain and partially purified telomerase through ion-exchange chromatography. The resulting fractions were analyzed in primer extension assays by the use of oligonucleotides with telomere-like sequences (35). Initial studies did not reveal a strong effect of CaEst1p loss on activity, in accordance with earlier results from Saccharomyces (29, 30, 35). Unexpectedly, when certain primers were used in subsequent experiments, the fraction from the Caest1-Δ strain became almost completely inert (e.g., Fig. 1A, P6 and P7 primers). To rule out the presence of primer-specific inhibitors in the mutant fraction, we added both BWP17 and Caest1-Δ fractions to an assay that included P6 as the substrate. The level of DNA synthesis was found to be the same as that of the BWP17 fraction alone, suggesting that the mutant enzyme has an intrinsic defect in nucleotide incorporation (data not shown).

Figure 1.

Candida telomerase from a Caest1-Δ strain manifests a primer-specific defect in reverse transcription in vitro. (A) Telomerase fractions derived from BWP17 and a Caest1-Δ strain were tested for primer-extension activity in vitro by using a series of 12-nt primers that correspond to different regions of the Candida telomere repeat. Four primers that exhibit strong Est1p dependence are boxed. (B) The ratio of the activity for BWP17 telomerase to Caest1-Δ telomerase was determined for 23 different primers and plotted. Each primer has a sequence that corresponds to part of the telomere repeat and was designated by the position of the first nucleotide added. Thus, e.g., primer P5 has a 3′ end that corresponds to position 4 in the figure and the first nucleotide added is at position 5. Each pair of assays was performed two to three times, and the averages and average deviations of the ratios are presented in the plot. For primers P6 and P7, the ratios cannot be determined accurately, because the Caest1-Δ enzyme essentially was inert. One copy of the 23-nt repeat is boxed in the figure. Primers P1–P12 have sequences that extend beyond the 5′ end of the boxed repeat such that the 5′ region of these primers contains sequences that correspond to the 3′ region of the boxed repeat (indicated by shaded background). The sequences of the primers are also illustrated by horizontal bars at the bottom of the figure.

To analyze in detail the effect of 3′ end sequence on CaEst1p dependence, we designed a series of 12-mer oligonucleotides (each with a different 3′ end) that correspond to all of the different permutations of the Candida repeat. Each primer has a sequence that corresponds to part of the 23-nt telomere repeat and was designated by the position of the first nucleotide addition. (Thus, primer P5 has a 3′ end that corresponds to position 4 in Fig. 1B, and the first nucleotide added is dC at position 5.) To simplify the analysis, we assayed wild-type and mutant fractions in the presence of just one labeled nucleotide such that only the primer + 1 (or primer + 2) product is generated. As shown in Fig. 1, analyses using fractions derived from the parental and Caest1-Δ strains revealed the existence of two loci within the telomere repeat (at positions 3–7 and 16–18), where the loss of CaEst1p significantly reduced enzyme activity. At most of the remaining positions, the loss of CaEst1p had minor effects on DNA synthesis. At four scattered positions (positions 9, 13, 21, and 23), the mutant fraction consistently displayed greater polymerization activity, suggesting that Est1p also can inhibit enzyme function at defined positions. The selective effect of CaEst1p implies that the defect of the mutant fraction is not due to a reduction in the amount of telomerase. Curiously, the majority of the CaEst1p-dependent positions call for the incorporation of dT. However, stimulation of dT incorporation cannot account for the function of CaEst1p: position 5 calls for the incorporation of dC, and other dT positions such as 10 and 12 do not manifest an effect of CaEst1p. Multiple fractions were prepared from both the BWP17 and Caest1-Δ strain. Although there were small variations in the quality of the enzyme preparation, the preferential utilization of specific primers by the wild-type enzyme relative to the mutant enzyme was reproducibly observed.

The Effects of 5′ End Sequence on CaEst1p-Dependent Reverse Transcription.

To characterize the effect of primer length on the observed CaEst1p dependence, we tested a series of primers with different lengths but the same 3′ end as P6. As shown in Fig. 2A, both the wild-type and the mutant fractions became progressively active with increasing primer lengths. However, the mutant fraction generated detectable signals only with the 14- and 16-nt primers, whereas the wild-type fraction was somewhat active on the 9- and 12-nt primers as well. As a consequence, the degree of CaEst1p dependence was particularly pronounced for the shorter primers. A series of primers with different lengths but the same 3′ end as P18 were subjected to the same analysis (Fig. 2B). For these primers, the wild-type fraction became increasingly active when the primer length was increased from 9 to 16 nt, whereas the Caest1-Δ fraction mediated approximately the same level of DNA synthesis on all three primers. As a consequence, the degree of CaEst1p dependence was greater for the longer primers. Although the results for the two primer series were somewhat disparate, they both suggest a functional interaction between the 5′ region of the primer DNA and CaEst1p.

Figure 2.

The magnitude of CaEst1p dependence is modulated by the length and sequence of the primer. (A) Telomerase fractions derived from BWP17 and a Caest1-Δ strain were tested for primer extension activity in vitro by using a series of primers with the same 3′ end as P6 but different overall length. Each pair of assays was performed two to three times, and the averages and average deviations of the activities were obtained from PhosphorImager analysis and are plotted at the bottom. The sequences for the primers and levels of activity for the wild-type enzyme are listed at the top. The 12-nt P6 primer is boxed. Also listed are the degrees of Est1p response, which denotes the average ratio of the activity for BWP17 telomerase to Caest1-Δ telomerase. (B) Telomerase fractions derived from BWP17 and a Caest1-Δ strain were tested for primer extension activity in vitro by using a series of primers with the same 3′ end as P18 but different overall length. Each pair of assays was performed two to three times, and the averages and average deviations of the activities were obtained from PhosphorImager analysis and are plotted at the bottom. The sequences for the primers and levels of activity for the wild-type enzyme are listed at the top. The 12-nt P18 primer is boxed. Also listed are the degrees of Est1p response, which denotes the average ratio of the activity for BWP17 telomerase to Caest1-Δ telomerase. (C) Telomerase fractions derived from BWP17 and a Caest1-Δ strain were tested for primer extension activity in vitro by using a series of primers with sequences that deviate from the Candida telomere repeat at selected regions. The names and sequences for the primers and the degree of Est1p response (the ratio of the activity for BWP17 telomerase to Caest1-Δ telomerase) are listed on the left. Regions of the primer that deviate from Candida telomere repeat are underlined. Examples of the assays are shown on the right.

The sequence requirement in the 5′ region of the P6-related primers was investigated (Fig. 2C). Interestingly, for the 16-nt primer [named P6(16)], mutating the 5′-endmost sequence of the primer accentuated the effect of CaEst1p loss [primer P6(16)mut1 and P6(16)mut2], whereas mutating an internal region abolished the effect [P6(16)mut3]. These results further suggest that CaEst1p may interact functionally with a DNA site located ≈5–8 nt upstream of the 3′ end of the P6 primer and that this interaction is sequence-dependent. We compared the sequences of the two sets of primers (P3–P7 and P16–P18) that supported Est1p-dependent reverse transcription but were unable to discern any obvious common features.

The Selective Effect of Est1p Loss at Specific Positions Can Be Observed During Both Initiation and Elongation.

Because our standard assay allows the incorporation of mostly 1 nt, the defect observed for the Caest1-Δ enzyme at particular positions can be considered an “initiation” defect. To determine whether Est1p is required for maximal activity at the same positions during elongation, we compared the ability of the wild-type and mutant enzymes to add multiple nucleotides onto primer P2(8). This primer has the same 3′ end sequence as P2 but is only 8 nt long. As illustrated in Fig. 3A, in the presence of dCTP and dTTP, telomerase is expected to add up to 6 nt (corresponding to position 2–7 of the repeat) to P2 (8). Consistent with a requirement for Est1p at most of these positions during elongation, the synthesis of long extension products was reduced greatly in the Caest1-Δ fraction (Fig. 3A, compare lane 1 with 2 and lane 3 with 4). The difference was particularly striking when labeled dTTP (present at 0.2 μM) was used in combination with unlabeled dCTP (present at 50 μM). With this combination of nucleotides, the amount of “primer + 5” and “primer + 6” product generated by the wild-type enzyme was >50-fold that of the mutant enzyme. In contrast, with the use of labeled dCTP (present at 0.2 μM) and unlabeled dTTP (present at 50 μM), the difference in the amount of primer + 6 product was ≈10-fold. This discrepancy can be explained by the preponderance of dT residues from position 3–7: When dTTP is labeled and present at low concentrations, elongation may be especially inefficient in the absence of CaEst1p. We conclude that CaEst1p is required for maximal telomerase activity at positions 3–7 during elongation.

Figure 3.

Comparison of nucleotide addition at positions 2–7 and 18 during elongation. (A) Telomerase fractions derived from BWP17 and a Caest1-Δ strain were tested for primer extension activity in vitro by using the P2(8) primer in the presence of different combinations of labeled and unlabeled nucleotides (with and without asterisk, respectively). The sequence of P2(8) and the expected nucleotide additions are listed at the top. The signals for the products that are marked with solid circles are determined. The ratio of the wild-type to Caest1-Δ signal for each product was calculated and is listed at the bottom. The positions of the primer + 1 to primer + 6 products are indicated by ticks to the right. For ease of visualization, a shorter exposure of lane 1 is shown at the far right. (B) Telomerase fractions derived from BWP17 and a Caest1-Δ strain were tested for primer extension activity in vitro by using two 12-nt primers (P17 and P18) in the presence of different combinations of labeled and unlabeled nucleotides (with and without asterisk, respectively). The sequences of the primers and the expected nucleotide additions are listed at the top. The signals for the products that are marked with solid circles or triangles are determined. The ratio of the wild-type to Caest1-Δ signal for each product was calculated and is listed at the bottom.

To determine whether the same elongation defect can be observed at the other Est1p-dependent locus, we compared the activity of the enzymes on the P17 and P18 primers in the presence of different nucleotide combinations (Fig. 3B). In accordance with earlier results, the addition of dT at the initiating position to P18 is CaEst1p-dependent (Fig. 3B, lanes 1 and 2). The inclusion of unlabeled dG in the reaction allowed efficient extension up to the primer + 3 position and slightly diminished the CaEst1p effect (Fig. 3B, lanes 3 and 4). Importantly, when the same labeled dT is added after the incorporation of unlabeled dA during elongation, the reaction is still CaEst1p-dependent (lanes 9 and 10). Even when dTTP is unlabeled and present at a much higher concentration (50 μM instead of 0.2 μM), the elongation from position 17 to 18 is still less efficient for the Caest1-Δ fraction, as evidenced by the amount of residual primer + 1 product (marked by triangles, Fig. 3B, lanes 7 and 8). The defect of Caest1-Δ telomerase at position 18 therefore is evident both for an initiating and an elongating enzyme.

Kinetic Analysis of Reverse Transcription.

To probe the mechanistic basis for the observed defect, we analyzed the kinetics of the reaction with respect to primer concentration. As shown in Fig. 4 and Table 1, for P6(13), which is Est1p-responsive, the wild-type enzyme exhibits a slightly lower K_m but a much higher V_max than the mutant enzyme. [The P6(13) primer is identical to the P6 primer except that the former is longer by 1 nt on the 5′ end. We used the 13-nt P6(13) primer instead of the 12-nt P6 primer for the kinetic analysis because the Caest1-Δ enzyme was completely inert on the shorter primer (Fig. 1A).] For P18, another Est1p-responsive primer, the wild-type enzyme has a higher K_m and V_max than the mutant enzyme (by 8- and 7-fold, respectively). The consistently higher V_max for the wild-type enzyme suggests that CaEst1p can increase the catalytic rate of telomerase in vitro. In contrast, for P12, a nonresponsive primer, the K_m and V_max differ by 2-fold or less between the wild-type and mutant enzymes (Table 1). Interestingly, for the same enzyme, the absolute K_m values for different primers can vary by two to three orders of magnitude, implying significant variability in the strength of enzyme–DNA interactions.

Figure 4.

Kinetic analysis of the wild-type and Caest1-Δ enzymes in reactions using the P6(13) and P18 primers. Telomerase fractions derived from BWP17 and a Caest1-Δ strain were tested for primer-extension activity in vitro by using increasing concentrations of the P6(13) and P18 primers. For these assays, a labeled 46-mer oligonucleotide was precipitated along with the reaction products as a recovery control. The signals were quantified, converted to reaction rates (see Materials and Methods for a more detailed explanation), and plotted against primer concentration at the bottom. The data points for the wild-type and Caest1-Δ reactions were fitted to the Michaelis–Menton equation, and the resulting curves are shown.

Table 1.

Kinetic analysis of Candida telomerase

	P6(13)		P18		P12
	BWP17	est1-Δ	BWP17	est1-Δ	BWP17	est1-Δ
Km, μM	0.71	1.73	24.4	3.3	0.011	0.017
Vmax, fmol/min	0.075	0.0081	0.31	0.051	0.00082	0.0016

Discussion

To our knowledge, this is one of the first instances in which a noncatalytic component of the telomerase complex is shown to affect activity (rather than assembly or stability) of the enzyme in vitro. Overall, our results suggest that CaEst1p facilitates reverse transcription by telomerase during parts of the extension cycle. Even though earlier analyses of Saccharomyces Est1p did not reveal a significant effect of this protein on enzyme activity in vitro, all of these studies used primers with the same 3′ end (29, 30). A position-specific effect of Est1p therefore could have been missed. Further investigation will be necessary to determine whether the activity of Est1p in vitro is conserved between these two fungi.

The mechanistic basis for the in vitro effect of CaEst1p is complex. On one hand, the effect of primer length and sequence (for the P6- and P18-related primers) on CaEst1p dependence suggests a functional (and sequence-specific) interaction between CaEst1p and the DNA primer (Fig. 2). Consistent with this notion, recombinant Saccharomyces Est1p has been shown to bind single-stranded telomeric DNA with moderate affinity (27). On the other hand, the kinetic experiments for P6(13) and P18 suggest that CaEst1p does not always improve the affinity of telomerase for DNA but consistently enhances the catalytic rate of the enzyme. In addition, the greater activity of the wild-type enzyme at positions 3–7 and 18 during elongation (Fig. 3) rules out faster dissociation and enzyme turnover as a means of increasing rate. Thus, we are left with the notion that CaEst1p may both interact with primer DNA and act as an allosteric activator of telomerase function in vitro. Consistent with this notion, the sequence of the primer upstream region (outside of the RNA–DNA hybrid) has been reported to influence the catalytic rate of Tetrahymena telomerase (40).

The mechanisms of Est1p in vivo are being investigated. A recruitment model for Est1p, in which Est1p brings the telomerase complex to telomeres through its association with Cdc13p (a telomere-binding protein), has been proposed based on a series of fusion and suppression analyses (31, 32). Alternatively, an activation model in which Est1p stimulates the activity of telomere-bound telomerase has been put forward based on chromatin-immunoprecipitation studies (33). The two hypotheses are not mutually exclusive. Indeed, genetic analysis has uncovered “separation-of-function” mutations in EST1 that appear to abolish selectively its recruitment or activation function (34). In serving both as a recruitment target and as an allosteric activator, Est1p appears to resemble the “mediator” complex of the RNA polymerase II holoenzyme (reviewed in ref. 41). Mediator interacts with DNA-bound transcriptional activators, operating as a bridge to recruit the core polymerase subunits to regulated promoters, where transcription takes place. Mediator also can activate basal transcription, probably by modulating the function of the Rpb3/Rpb11 subunits of the core polymerase, with which it makes direct contact. Another interesting parallel can be drawn between Est1p and transcriptional elongation factors (reviewed, e.g., in ref. 42): Both bind stably to the core polymerase subunits but enhance polymerase elongation only at selective sites along the templates. Moreover, some transcriptional elongation factors have been shown to bind nascent products (transcripts), much as Est1p is believed to bind the primer upstream region, which can be considered the product of telomerase extension. It would be interesting to determine in the future whether these superficial similarities are reflections of deeper mechanistic resemblance.

The in vitro stimulatory effect of CaEst1p is likely to be related to its activation function, given the consistent alterations in V_max. The magnitude of the effect should enable detailed biochemical investigation of the underlying mechanisms. Because potential orthologues of Est1p have been identified in phylogenetically diverse organisms (35, 43), continued analysis of CaEst1p may have broad implications for telomerase mechanisms and regulation.

Abbreviations

TERT

telomerase reverse transcriptase

EST1 and EST3

ever shorter telomeres 1 and 3