Protein Composition of Catalytically Active Human Telomerase from Immortal Cells

See allHide authors and affiliations

Science  30 Mar 2007:
Vol. 315, Issue 5820, pp. 1850-1853
DOI: 10.1126/science.1138596


Telomerase is a ribonucleoprotein enzyme complex that adds 5′-TTAGGG-3′ repeats onto the ends of human chromosomes, providing a telomere maintenance mechanism for ∼90% of human cancers. We have purified human telomerase ∼108-fold, with the final elution dependent on the enzyme's ability to catalyze nucleotide addition onto a DNA oligonucleotide of telomeric sequence, thereby providing specificity for catalytically active telomerase. Mass spectrometric sequencing of the protein components and molecular size determination indicated an enzyme composition of two molecules each of telomerase reverse transcriptase, telomerase RNA, and dyskerin.

Telomeres, repetitive nucleoprotein structures at the ends of linear chromosomes (1), shorten during each cycle of cell division (2), providing a counting mechanism to limit the number of times a cell can divide (3). Many cancer cells escape limits on proliferation by activating the ribonucleoprotein enzyme telomerase to catalyze the synthesis of telomeric repeats (4). The protein component, human telomerase reverse transcriptase (hTERT), contains conserved catalytic reverse transcriptase motifs (5, 6), and the human telomerase RNA component (hTR) (7) directs the addition of deoxynucleotide triphosphates (dNTPs) by means of an internal template complementary to the telomeric repeat sequence TTAGGG.

Telomerase has previously been purified only from the ciliate Euplotes aediculatus as a complex of TERT, RNA, and associated protein p43 (8). At least 32 distinct proteins have been proposed to associate with human telomerase (table S1). Size measurements of human telomerase have indicated a complex larger than expected for a composition of one hTERT (127 kD) and one hTR (153 kD) (9, 10) but smaller than the sum of all proposed protein associations (∼2.6 MD). Nonetheless, the precise composition of the active enzyme complex within the cell has remained undefined.

We measured the size of the active human telomerase complex in a panel of immortal cell lines (MCF-7, A2182, HCT-116, TE-85, HT-1080, and HEK-293, derived from cancers of the breast, lung, colon, bone, and connective tissue, and from embryonic kidney cells, respectively). Quantification of telomerase was performed with a direct (non–polymerase chain reaction) primer-extension activity assay (fig. S1). Whole-cell lysates (11) from all cell lines exhibited a similar sedimentation profile, with ≥60% of total activity eluting in fractions 9 and 10 (Fig. 1, A to C). Thyroglobulin (669 kD) peaked in fraction 9 (Fig. 1B), indicating that telomerase exists as an enzyme complex of ∼650 to 670 kD.

Fig. 1.

Size measurements of the active human telomerase enzyme complex. (A) Glycerol gradient sedimentation profiles of HEK-293 telomerase activity from extracts of whole cells in lysis buffer with or without Triton X-100 (0.1% v/v). Gradients of 10 to 40% glycerol were collected into 22 0.5-mL fractions; fraction 1 is from the bottom of the gradient at 40% glycerol. *, 100-nucleotide oligomer DNA recovery/loading standard included for quantitation. (B) Black and red bars, quantitation of data from (A); the lane with the highest activity is arbitrarily given a value of 100. Blue line, sedimentation profile of thyroglobulin (669 kD). (C) Glycerol gradient sedimentation profiles of telomerase activity for all cell lines examined. (D) Tracking HEK-293 telomerase activity through the purification. Lane 1, immunoaffinity-purified telomerase (defined as 100%; subsequent yields are quantified relative to lane 1). Lane 2, immunoaffinity-purified telomerase after incubation at room temperature for 5 hours. Lane 3, immunoaffinity-purified telomerase after preclearing with unmodified neutravidin beads. Lane 4, telomerase left in solution after treatment with (TTAGGG)3-modified neutravidin beads. Lane 5, telomerase eluted in 1 hour from (TTAGGG)3 beads in the absence of dTTP/dATP. Lane 6, telomerase eluted in 10 min from (TTAGGG)3 beads in the presence of dTTP/dATP. The volumes of the samples represented in lanes 5 and 6 are reduced by factors of 5 and 10, respectively, relative to those in lanes 1 to 4.

We developed a purification scheme that achieved ∼108-fold enrichment of active telomerase in three steps (11). The first step was immunoaffinity purification with a sheep polyclonal antibody generated against the peptide antigen ARPAEEATSLEGALSGTRH (hTERT amino acids 276 to 294). HEK-293 lysate was incubated with antibody, and the antibody-enzyme complex was immobilized onto protein G agarose beads. Excess antigenic peptide was added to the resuspended immunoprecipitate to allow dissociation of the enzyme from immobilized antibody into solution (Fig. 1D, lane 1). Immunopurified telomerase displayed excellent stability (Fig. 1D, lane 2), allowing the purification to be continued at room temperature.

The specificity of the second and third steps is provided by thermodynamic and kinetic properties of the telomerase enzyme. The dissociation rate of the telomeric DNA substrate primer 5′-(TTAGGG)3-3′ from human telomerase is known to be very slow (t1/2 ≥ 10 hours) (12). We exploited this stable binding between enzyme and substrate to develop a substrate-directed affinity purification. The synthetic DNA 5′-Biotin-CTAGACCTGTCATCA(TTAGGG)3-3′ was immobilized onto neutravidin beads, providing the affinity reagent.

Immunopurified telomerase from the first step was precleared with unmodified neutravidin beads and 5′-CTAGACCTGTCATCA-3′; loss of telomerase activity was negligible (Fig. 1D, lane 3). The precleared solution was then treated with (TTAGGG)3-modified beads to capture telomerase. This affinity purification consistently proceeded to >90% yield, as evidenced by the low level of telomerase activity remaining in solution (Fig. 1D, lane 4). Activity from the captured enzyme was detected by suspending the beads in assay buffer (dNTPs), resulting in extension of the immobilized DNA substrate (fig. S2).

The stable binding between telomerase and immobilized DNA primer (t1/2 ≥ 10 hours) required that we develop a rapid elution procedure specifically for catalytically active enzyme complexes under mild conditions. We exploited a distinctive relationship between human telomerase and DNA primer: Kd and koff between enzyme and primer change as a function of the nucleotide at the 3′ end (12). A primer ending in GGG, such as 5′-(TTAGGG)3-3′, affords the most stable binding interaction; primers ending in TTA display the weakest binding between enzyme and primer, with dissociation occurring in minutes. An active enzyme immobilized onto 5′-(TTAGGG)3-3′, in the presence of only deoxythymidine triphosphate (dTTP) and deoxyadenosine triphosphate (dATP), should catalyze the addition of TTA, thereby transforming a stable enzyme-substrate complex (t1/2 ≥ 10 hours) into a relatively unstable complex (t1/2 < 5 min), promoting rapid elution (Fig. 2A).

Fig. 2.

Activity-dependent elution of telomerase. (A) Addition of TTA to immobilized (TTAGGG)3 destabilizes the telomerase-DNA complex. (B) Dissociation of telomerase from (TTAGGG)3-modified beads was followed in the absence or presence of dTTP and dATP. (C) Quantitation of data from (A).

To demonstrate activity-dependent elution, (TTAGGG)3-modified beads bearing telomerase were divided into two suspensions. Excess free 5′-(TTAGGG)3-3′ was added to trap the enzyme in solution phase after dissociation. One suspension was left as a control to follow dissociation from DNA with GGG at the 3′-end; to the other was added dTTP and dATP. Samples of each suspension were removed over time, and the solution phase was separated from beads to measure the amount of dissociated telomerase (Fig. 2, B and C). The presence of dTTP and dATP promoted a dramatic increase in the rate of enzyme dissociation (control, k < 0.002 min–1; +dTTP/dATP, k = 0.25 ± 0.05 min–1). At 10 min, the selectivity for activity-dependent elution is >50:1 (Fig. 2C). Such strong dependence on dTTP/dATP provides exquisite specificity in the final elution and illustrates an elegant structure-activity relationship of human telomerase: Transformation to a less stable enzyme-substrate complex as nucleotide addition approaches the end of the template facilitates efficient translocation.

For the purification, the (TTAGGG)3 beads bearing telomerase were suspended in buffer containing free (TTAGGG)3 for 1 hour as a further washing step, during which time just 4 to 5% of telomerase had dissociated (Fig. 1D, lane 5). The beads were suspended in buffer containing free (TTAGGG)3 for 10 min; this “control elution,” with < 1% activity, was collected for mass spectrometry. Finally, the beads were suspended in buffer containing free (TTAGGG)3 and dTTP/dATP, and the product solution was collected after 10 min to provide purified telomerase (Fig. 1D, lane 6).

Purified telomerase was assessed for size and yield. The sedimentation profile of purified telomerase was conserved relative to that of the crude lysate (Fig. 3, A and B). Yield of telomerase was determined by Northern blot analysis against hTR, with gel-purified in vitro transcribed hTR as a quantitation standard (fig. S3). From 50 fmol HEK-293 cells (∼100 g), we obtained ∼250 to 300 fmol purified telomerase (5 to 6 molecules per cell, ∼100 ng). These data were used to determine the cellular abundance of telomerase, factoring in the percent yield of purification, for which an upper limit is estimated at ≤30% [immunopurification, ≤60%; (TTAGGG)3 immobilization, ≤90%; +dNTP elution, ≤60%]. This corresponds to a lower limit of ∼20 telomerase molecules per HEK-293 cell. Allowing for error in these analyses, and given the variability of telomerase levels in different cell subpopulations (13), we conclude that an average HEK-293 cell has ∼20 to 50 molecules of telomerase. The data also indicate a purification factor of ∼108 (∼100 ng telomerase from 100 g cells, adjusted for ≤30% yield) (11).

Fig. 3.

Characterization of the purified HEK-293 telomerase enzyme complex. (A) Glycerol gradient sedimentation profile of telomerase activity from purified HEK-293 telomerase. (B) Elution profile of purified enzyme [data from (A)] was compared with that of whole cell lysate (data from Fig. 1B). (C) MS/MS spectrum for hTERT amino acids 249 to 261: TPVGQGSWAHPGR. (D) MS/MS spectrum for dyskerin amino acids 65 to 80: TTHYTPLACGSNPLKR.

Purified telomerase was digested with trypsin in solution, and the resulting peptides were captured on cation exchange resin at pH 4 to remove buffers, nucleic acid, and detergent (11). The peptide products were eluted and analyzed by nano–liquid chromatography–tandem mass spectrometry (nanoLC-MS/MS). Protein identification from peptide sequencing was performed on three independent purifications, each from 100 g HEK-293 cells. The 10-min control elution, containing only (TTAGGG)3 and essentially devoid of telomerase activity, was analyzed in parallel with the dTTP/dATP-dependent elution of telomerase. Comparison of these samples identified proteins dependent on the presence of dTTP/dATP and telomerase activity. In the control sample, typically 6 to 8 proteins were observed at low levels in each run, 5 of which were observed consistently: tubulin, actin, Y-box, heterogeneous nuclear ribonucleoprotein (hnRNP) A1, and hnRNP M. These proteins represent background and result from nonspecific adhesion inherent in protein purification. In the telomerase sample, in addition to the background proteins tubulin and actin, two new proteins were consistently observed: hTERT and dyskerin (Fig. 3, C and D, figs. S4 and S5, and Table 1). Only hTERT and dyskerin were specifically enriched in the active telomerase fraction.

Table 1.

hTERT and human dyskerin peptides sequenced by nanoLC-MS/MS (21).


We conclude that the active human telomerase enzyme complex is composed solely of two protein components, hTERT and dyskerin, and the RNA component hTR. Dyskerin is a putative pseudouridine synthase within the class of H/ACA box ribonucleoproteins (14). The H/ACA sequence motif is present in hTR (15). A relationship between telomere length, telomerase, and dyskerin has been documented in dyskeratosis congenita, a syndrome characterized by reduced proliferative capacity attributed to telomerase deficiency. Mutations in dyskerin resulted in poor telomere maintenance and lower telomerase activity, and the evidence supported an association of dyskerin with telomerase (16). It is noteworthy that mutations in hTERT (17), hTR (18), and dyskerin (16), the three components of the active telomerase enzyme complex, are the only known genes whose mutation has been shown to cause dyskeratosis congenita.

The combined mass of hTERT (127 kD), hTR (153 kD), and dyskerin (57 kD) make up about half the observed size (650 to 670 kD). There is considerable evidence from in vitro reconstitution studies that telomerase exists as a dimer (9, 19, 20). We propose that human telomerase exists as a complex of two molecules each of hTERT, hTR, and dyskerin. It should be emphasized that the purification is specific for catalytically active enzyme complexes and therefore represents a restricted view of the overall biology of telomerase. Other proteins reported to associate with telomerase (table S1) may be involved in its biogenesis, trafficking, recruitment to the telomere, and degradation. However, from the analyses described here, it can be concluded that these proteins are not required for nucleotide addition, nor do they constitute integral components of the catalytically active enzyme complex.

Supporting Online Material

Materials and Methods

Figs. S1 to S5


References and Notes

View Abstract

Stay Connected to Science

Navigate This Article