This action might not be possible to undo. Are you sure you want to continue?
1 RESEARCH ARTICLE
The RNA Component of Human Telomerase
Junli Feng, Walter D. Funk, Sy-Shi Wang, Scott L. Weinrich, Ariel A. Avilion, Choy-Pik Chiu, Robert R. Adams, Edwin Chang, Richard C. Allsopp, Jinghua Yu, Siyuan Le, Michael D. West, Calvin B. Harley, William H. Andrews, Carol W. Greider,
Human chromosomes terminate with several kilobases of the simple telomere repeat (TTAGGG)n. Telomeres protect the chromosomes from DNA degradation, end-toend fusions, rearrangements, and chromosome loss (1, 2). Because DNA polymerases synthesize DNA in the 5' to 3' direction and require an RNA primer for initiation, telomeric DNA may be lost at chromosome ends unless the termini are specifically extended by telomerase (3, 4). Telomerase is a specialized ribonucleoprotein polymerase that contains an integral RNA with a short template element that directs the synthesis of telomeric repeats at chromosome ends (5, 6). The deregulation of telomerase may participate in cellular immortality and oncogenesis. Normal human somatic cells express low or undetectable telomerase activity and progressively lose their telomeric sequences with replicative senescence in vitro or with normal, in vivo aging (7-14). In contrast, germline cells and almost all tumor cell lines and tissues express telomerase and maintain telomere length through an indefinite number of cell divisions (12, 15-17). Transfection of normal somatic human cells with viral oncogenes leads to cell division beyond the normal senescence checkpoint and the continued
Feng, W. D. Funk, S.-S. Wang, S. L. Weinrich, C.-P. Chiu, R. R. Adams, E. Chang, R. C. Allsopp, J. Yu, M. D. West, C. B. Harley, W. H. Andrews, and B. Villeponteau are at Geron Corporation, 200 Constitution Drive, Menlo Park, CA 94025, USA. A. A. Avilion, S. Le, and C. W.
J. Greider are at Cold Spring Harbor Laboratory, Cold Spring Harbor, NY 11724, USA. *To whom correspondence should be addressed.
loss of telomere sequences, defined as "crisis." At crisis, telomeres are critically short and genomic instability is marked (12, 15, 18). Although most cells die in crisis, the rare cells that escape this fate are typically aneuploid, express telomerase activity, and have a stable telomere length. Therefore, telomerase activation can be a critical step in cell immortalization. Short telomeres have been detected in a range of tumor cell lines and tissues, including breast cancer, Wilms tumor, colorectal carcinoma, and leukemia (8, 19-26), which suggests that telomere shortening may be involved in oncogenesis. Perhaps because of this telomere shortening, over 90 percent of primary human tumors express telomerase activity (17, 27). Thus, telomerase may be activated in the premalignant cell as telomeres get critically short, and continued telomerase activity may be required for the long-term growth of the fully malignant cancer cell (7-9, 12, 17). We have cloned the RNA component of human telomerase and show here that expression of an antisense telomerase RNA leads to telomere shortening and cell death. Cyclic selection cloning of telomerase RNA candidates. The RNA component of telomerase has been cloned from many ciliates (28-30) and yeast (31). No similarity in RNA size or sequence was found between these species, which indicates that the primary sequence of telomerase RNAs is not conserved. Conventional low-stringency
hybridization techniques, therefore, were unlikely to yield the human telomerase RNA. In Tetrahymena, which contains
>20,000 telomeres, telomerase RNA is
SCIENCE · VOL. 269
1 SEPTEMBER 1995
Downloaded from www.sciencemag.org on September 22, 2013
Eukaryotic chromosomes are capped with repetitive telomere sequences that protect the ends from damage and rearrangements. Telomere repeats are synthesized by telomerase, a ribonucleic acid (RNA)-protein complex. Here, the cloning of the RNA component of human telomerase, termed hTR, is described. The template region of hTR encompasses 11 nucleotides (5'-CUAACCCUAAC) complementary to the human teiomere sequence (TTAGGG),. Germline tissues and tumor cell lines expressed more hTR than normal somatic cells and tissues, which have no detectable telomerase activity. Human cell lines that expressed hTR mutated in the template region generated the predicted mutant telomerase activity. HeLa cells transfected with an antisense hTR lost telomeric DNA and began to die after 23 to 26 doublings. Thus, human telomerase is a critical enzyme for the long-term proliferation of immortal tumor cells.
at "20,000 copies per cell (32). Because human cells have only 92 telomeres and telomerase activity in humans is weaker than in Tetrahymena (33), we anticipated that the human telomerase RNA might be a rare transcript requiring special protocols to enrich for this RNA. Enrichment for telomerase genes was initially carried out by subtractive hybridization or by partial telomerase purification (34). RNA was prepared from telomerasepositive cells (293), telomerase-negative cells (IMR90), or partially purified telomerase fractions from 293 cells. Complementary DNA (cDNA) libraries were generated from these RNAs by random-primed reverse transcription of the RNAs and the addition of synthetic linkers to allow for polymerase chain reaction (PCR) amplification. Subtraction of the telomerase-negative library from the telomerase-positive library generated the telomerase-enriched subtracted PCR-cDNA library. In addition, a purified PCR-cDNA library was prepared from partially purified telomerase enzyme fractions, which represented a second independent library enriched for potential telomerase RNAs. Telomerase RNAs typically contain 1.5 repeats of the telomere complementary sequence (29, 30, 35). We designed a PCR cyclic selection method to enrich for candidate telomerase RNAs containing the predicted human telomeric template sequence CUAACCCUA. Both the subtracted and purified PCR-cDNA libraries were enriched for telomerase template-containing sequences by hybridization to a biotinylated 12-nucleotide complement to the predicted template (36). After hybridization, cDNAs bound to the biotinylated oligonucleotide were separated from nontemplate cDNAs with the use of streptavidin-agarose magnetic beads. To maximize the concentration of cDNAs containing CUAACCCUA in the libraries, we used three or four repetitive cycles of positive selection with the biotinylated, template-specific oligonucleotide. After the third cycle of selection, individual PCR bands could be resolved on native polyacrylamide gels, which indicated that some sequences had been highly selected. Individual bands were excised, subcloned into pBluescript, and sequenced. Twelve cDNAs containing the predicted telomere template (CUAACCCUA) were identified as telomerase RNA candidates (TRCs). Preliminary identification of TRC3 as the RNA component of human telomerase. To determine which of the 12 candidates might encode the human telomerase RNA, we initially assayed for the inhibition of telomerase activity using antisense oligonucleotides. Tetrahymena, Euplotes, and Oxytricha telomerases are specifically inhibited by oli-
a genomic RNase would DNA fragment that expresses the entire TRC4 TRC3 gene sequence was subcloned from lambda and mutagenized. ri. er extension PCR and RT-PCR (40). The oligonu. The reaction products were resolved by polyacrylamide gel electrophoresis. and the position of the 17-base insert (ATCGGATCCAGACGAGC) in Fig.TCAGGC 370 GAITCA''i 850 830 840 800 810 820 TTTGAGAGAT CAITTAACAT TTAATGAATA 'ITTAATrAGA AGATCTAAAT 390 410 420 860 CGCAGGAAGA GGAACGGAGC GAGTCCCGCC GCGGCGCGAT TCCCTIGAGCT GTGGGAcGTG 430 440 450 460 470 480 CACCCAGGAC TCGGCTCACA CATGCAGTPC GCTTTCCTGT TGGTGGGGGG AACGCCGATC GAACATIGGA AA'TGTGTT 910 900 870 880 890 CITTAATGGT CATCGGTITTA oTGCCAGAGGT TAGAAGTTITC lTI'ri'rAAA rr 960 940 950 920 930 AAqTAGACCT TGGCGATGAC CTTGAGCAGT AGGATATAAC CCCCACAAGC SCIENCE * VOL. These oligonucle.ase. TRC3 RNA correlates with biochemical purification of telomerase activity. These mutant telomerase activities element (28-30). Therefore.allowed specific detection of the altered sense oligonucleotides distal from the tem. Alter- MuC* is shown at nucleotide 176 by an asterisk.6-kb Hind ITRC1 Ill-Sac I fragment containing TRC3 was TRCll first inserted into a modified version of pBluescript. gradient fractions were collected (fraction 1 = bottom) and assayed for telomerase activity by the PCR method (17). Similar copurification of TR.TRC3 was altered from CUAACCCUA to didates. TRC3.Southern blot was consistent wit]h the re. Fractions containing active telomerase were pooled (DS input) and then loaded onto a preformed glycerol gradient (15 to 42 percent). 35). 1). the pattern of inhibition by copurification and the antisense iinhibition into functional telomerase.candidates tested. 1.upstream sequence. The full TRC3 sequence was then obtained from this genomic subtelomerase template and adjacent sequences to the pattern observed in the cili ates. S100 extracts from 293 cells were fractionated over DEAE-Sepharose and then Sephacryl S300HR resins. elomerase activity ation of the template region of Tetrahymena Mutation of the TRC3 template. generating the genomic TRC3 gonucleotides that are complementary to the plate were not inhibitory.contain an exact copy of the TRC>3 cDNA which a 17-bp insertion was present at man telomerase requires oligonucleotide fragment used to probe the library.C3 with CCAACCCCA (MuC) or CAAACCCAA ty. 2).striction pattern observed in th( TRC3. The +176 bp (asterisks in Figs. Sequence of hTR and the downstream region.T)-PCR which is consistent with its migration on gate each of the antisense oligonucleotides (Fig. Both strands were sequenced from the genomic clone by the Sanger dideoxy method 10 20 30 40 50 60 GGGCCATIT TTGTCTAAC CCTAACIGAG GGGTTGCGGA GGGTGGGCCT GGGAGGGGGT 70 80 AAGGGCGTAG GCGCCGTGCT 490 500 510 520 530 540 GTGCGCATCC GTCACCCCTC GCCGGCAGTG GGGGCTTGTG AACCCCCAAA CCTGACTGAC 550 610 670 90 100 160 T'TGC-ICCC GCGCGC'IGTT TFHCTCGCT'G ACTFTCAGCG 150 110 120 560 620 570 630 580 640 700 590 650 710 600 660 720 'IGGGCCAGTG TCCTGCAAAT TGGCAGGAGA CGTGAAGGCA CCTCCAAAGT CGGCCAAAAT GAATGGGCAG TGAGCGGG 680 130 140 GGCGGAAAAG CCTCGGCCTG CCGCC7TCCA CCGTTCATTC TAGAGCAAAC AAAAATGTC 170 *180 TCCIGAG 690 CCGGTCCTGC GTGGGTTC'IC CCGTCTTCCG 190 200 210 220 230 240 AGCTGCTGGC CCGTICGCCT CCCGGGGACC TGCGGCGGGT CGCCTGCCCA GCCCCCGAAC CTI'TFI'I7TG CCTTITATGG TTGTA1TACA ACTTAGTTCC TGC'CIVGCAG ATTTIGTGA 730 740 770 780 750 760 GGITrITCT CTCCCAAGG TAGATCTCGA CCAGTCCCTC AACGGGGTGT GGGGAGAACA 790 250 310 260 320 380 270 330 280 340 400 290 350 300 360 CCCGCCTGGA GCCGCGGTCG GCCCGGGGCT TCTCCGGAGG CACCCACTGC CACCGCGAAG AGT'GGGCTC TGTCAGCCGC GGGTCT1CG GGGCGAGGG CGAGGTTCAC CGT. ity were then tested for their ability to elon. Stratagene).Fig. only TRC3 hadI a signal fragment in pGRN33 and contains 1. we designed oligonucleotides that would hybridize to the template was copurification with telomerase activity. tive telomerase primer (17)].candidates did not copurify with telomer. and cDNA was made with random hexamers as primers. which is known to serve as an effec. In ciliates. both genomic plasmid pGRN33. (56). Southern (DNA) blot analysi s and in than that of the wild-type repeat TTAdependent on their targeting to specific sequences around the telomeric template situ hybridization demonstrated th at TRC3 GGG.and fractions were tested for t he TRC size of the RNA transcript was "450 bases.RNAs by reverse transcriptase (R. The data a 15-kb insert was isolated and found to double mutant (MuC*) was also prepared in indicated that antisense inhibition of hu.(MuA) by in vitro mutagenesis (41) of the ent purification scheme (38). Extracts with telomerase activ. we designed a is a single-copy gene localized to 1the distal could be distinguished from that of the wildseries of oligonucleotides complementary to quarter of the long arm of chrorrlosome 3 type because they no longer require deoxyvarious sequences near or distal to the pu. TI mus. RNA was extracted from each fraction. Of the gion for TRC3 was central to the hTR gene [oligonucleotide AATCCGTCGAGCAG. 31.that correlated with telomerase ac tivity.bonuclease (RNase) P and three ot her TRC The predicted template sequence of cleotides complementary to one of the can. and yeast telomerase RNA genes leads to corresponding changes in the telomere repeat synthesized (6. This mutant around the TRC3 template and that anti.RNA with a probe to the 17-bp insertion.Telomerase activity was partially purified. A inhibit telomerase activity (37). RNA transcript were mapped with 5' primand adjacent regions for each of the candi. 2 and 3A) in complementarity to sequences in and restriction pattern deduced from tlhie TRC3 addition to the MuC template. The 5' and 3' ends of the (28-30). whereas oligonucleotides designed for the telomerase activity was observed iria differ. The transcribed reand the nontelomeric telomerase primer TS tions 7 to 9 of the glycerol gradienit. After centrifugation. Therefore. To determine if TRC3 alteration of the TRC3 template region produce a similar result. If incorporated other 11 candidates did not. which confirmed that this lambda clone contained TRC3 genomic DNA (39). TRC3 cDNA was used to probe a adenosine triphosphate (dATP) for activity tative template of TRC3.(39). The 2. these mutant antisense oligonucleotides that were based experiments suggested that TRC3 encodes RNAs should direct the synthesis of TTGGGG (MuC) or TTTGGG (MuA) rather on the telomerase RNA component is the human telomerase RNA comi )onent. The date clones. The resulting cDNA was included in PCR reactions with the use of sequence-specific oligonucleotides derived from the TRC candidates or from RNase P (55) as indicated. inhibited telomerase activi. DS 1 3 5 7 9 11 13 15 17 19 21I selected lambda clone. An additional test of the TR4C clones clone (Fig.Northern (RNA) blots.human genomic DNA library (lamLbda ZAP and only the wild-type activity is sensitive to otides were again tested for their ability to II. The telomerase template is underlined. 2. 269 · 1 SEPTEMBER 1995 1237 .4 kb of AGTT. A single positive ci lone with termination by dideoxyATP (ddATP). which is similar plasmid pGRN33. Telomerase activity peaked in frac.
13.-. Telomerase is representative of the predicted sequence of the MuC telomerase prodproducts with specific DNA sequence were then detected by PCR amplifi. respectively.lll. 1T. labeled C). 3A). Telomerase activity from cells expressing temrnplate-mutated TRC3.produce any phates (dNTPs). normal minus dTTP plus 0. primer (AACCCC) 3 tained 8 ~M total dGTP. In (B). cell extracts were first incubated with the TS sub. 12. which indicates that telomerase reconstitution is sensitive to certain alterations in the TRC3 To confirm that the sequence synthesized by the mutant MuA was (GGGTTT)n. and 18. [at-32P]deoxyguano- TRC3 RNA. the synthetic telomerase a or C [AT[a-32P]dGTP (800 Ci/mmol). normal plus 0.) primers and with all four dNTPs 1238 SCIENCE synthetic C products permutation primer. (A) Extracts from mutant TRC3-expressing stable transformants were frac- A Normal ddC ddA Mu_ I MuC - WTI MuA ll A -+-+ -- C* C A C* C A C* C A C* A CA C* C C*CA RNase - + -+- B C tionated by DEAE-Sepharose and were assayed for telomerase activity by conventional assays under various reaction conditions. we determined whether the 2. or normal minus -41 (CCCTAA)4 MuQ (ACCCAA) 4 CMMIQ. amd MuC cells in the presence of all four deoxynucleoside triphos.5 mM ddCTP (lanes 7 to 9). B to E) added onto a unique telomerase primer (17). verse primer.5 mM ddATP (lanes 13 to 18). 10. primer bp 65 expressing MuC* TRC3 (lanes 1.l. and 17. whereas (TTAGGG)n oligonucleotides did not wild-type telomerase activity was detected in cell extracts from wild-type. labeled A) were assayed under normal reaction conditions (lanes 1 to 6). Flanking lanes contain DNA markers with sizes in nucleotides (nt) as tect the MuA or MuC products.ucts and is further supported by the comigration of MuC products with the of the reverse with each cation with the appropriate reverse (Rev. This result residual telomerase activity by boiling the extracts for 5 min. or MuA plasmids into HT1080 cells along with pCMVneo as a selection marker. 3.·lll*msll(·1. assay reactions in lanes and trace amounts of [a-32P]dCTP. and dATP]. and 16. The RNA was detected in MuC*-transfected cells but not in mock-transfected cells (42). 3. 6. To facilitate mutant telomerase detection.primers. -+ kUC C RNase nt + + D . the presence of PCR products with (ACCCAA)4 or a single band with the (CCCCAA)3 rerespectively. In telomerase with template-mutated TRC3 was to use PCR primers specific further tests of the specificity of our PCR amplification conditions. Extracts were treated with DNase-free RNase products A [AATCCGTCGAGCAGAGTT-TGG (TTTGGG)4TTTG] (25 i~g/ml for 10 min at 300C) before telomerase assays (lanes 1 to 3 and 16 CCGTCGAGCAGAGTT-GGGG(TTGGGG)4TTGG] were prepared.l. 14. In contrast. (AACCCC)3. To deto 18).8 actions in lanes 1 to 9 con(CCAACC)3 (CCCCAA) a Rev. Resistant clones were selected in G418. 7. 11.4. Extracts from cells bp 64 46 40 Rev. of which a portion was 2NPM added. synthetic for the repeat sequence (-TTGGG)n.5 mM ddTTP (lanes 10 to 12). Assays of extracts from untransfected cells gave the same result as C*. no telomerase extract was 10 to 18 contained 8 iM total dGTP. whereas the C* extract did not. bp 38 34 dATP plus 0.6 kb of genomic sequence was sufficient for expression of TRC3 in vivo. (E) To confirm the sequence of the MuC products. Stable transformants were then derived by electroporation of MuC*. To test for mutant telomerase activity. Fig. In (C). As a control for MuA or MuC specificity. when ddATP was substituted for dATP. 8. ature to allow for the addition of telomeric repeats. labeled C*). we assayed the extracts from untransfected cells and from three stable transformants with integrated MuC*. 15. 9. MuC.8 bp 36. and (CCAACC)3. C. 5. 269 * 1 SEPTEMBER 1995 . (B to D) The second strategy for analyzing reprogramming of (AACCCC)3 and found them to be specific for the MuA or MuC product. Assay re- 2!2 1. or MuA vectors (C*. In preliminary experiments. · VOL. MuC. or A in Fig. found in A or C telomerase oligonucleotides generated the appropriate 6-nt ladder MuA. and RNA expressed from the DNA was detected by RT-PCR with the 17-bp insert sequence as the primer in the reverse transcription step. The simplest interpretation of these results is that the ddATP-resistant activities represent telomerase reprogrammed with MuC or MuA triphosphate (dTTP). PCR products with the (ACCCAA)4 or (CCCCAA)3 reverse MuA. we used the reverse primers (ACCCAA)4 or indicated. Using synthetic telomerase products.6-kb genomic clone was sufficient for TRC3 expression.reverse primers: (CCCCAA)3. various assay conditions were used to distinguish between them. or wild-type (WT) TRC3. (-TTGGGG)n. we modified the existing PCR methodology for amplifying telomerase repeats (Fig. which indicates that the 2. Cells were transiently transfected with the MuC* plasmid. We then destroyed which gave PCR products with the corresponding mobility shifts. MuC TRC3 (lanes 2. Because the mutant extracts were expected to contain both wild-type and mutant telomerase activities. the C and A extracts still displayed RNase-sensitive telomerase WT ddT activity. all three extracts from the mutant construct series showed telomerase activity that was sensitive to RNase. this activity was unaffected when dideoxycytosine triphosphate (ddCTP) was included in the reactions and was abolished by dideoxyTTP (ddTTP). As expected.11. we identified reaction conditions where a 3' return primer with the sequence d(CCCAAACCCAAACCCAA) would ampliN sequence. or (-TTAGGG). we used three strate in the presence of only dTTP and dGTP for 10 min at room temper. Under normal reaction conditions [with deoxythymidine sine triphosphate (dGTP). In lane N.131. the 17-bp insertion in C* inhibits reconstitution. In contrast. or MuA TRC3 (lanes 3. respectively. MuC. and expression of the mutant RNAs was verified by RT-PCR (43). (AACCCC)3 161718 1 23 45 678 9 101112 of which a portion was 1 131415 pM [a-32P]dGTP (800 Ci/ rmmol).
96 C 0/48 0/36 C 1239 .u-. with the highest expression seen in testis and ovary tissue.. These methods were also used to analyze the MuC mutant (Fig. lane 8.£ n .59 11/11 2. Testis tissue was arbitrarily given the value of 100 to compare these data with those from the Northern blot in (B). Eleven to 18 stable clones were isolated from each transfection series. colon tumor SW620 without RT. Relative steady-state transcription was nomalized to GAPDH as a control for loading and is shown below each lane. respectively. which differs from that in both the ciliates [153 to 192 nucleotides (nt)] and Saccharomyces cerevisiae (1. Telomerase RNA is divergent between species. The length of the mature hTR transcript is approximately 450 nucleotides.96 0/11 2. The steady-state concentrations of hTR transcripts were two. lane 9.19 3. o 8 >' _U . 4A). which indicates that telomerase from MuA-containing cells generated (FTTGGG)n repeats that were similar to synthetically made (TTTGGG)n repeats ("A" in Fig.5 No 7. foreskin fibroblasts. only U3 has been reported to be transcribed by different RNA polymerases in different species (46. primary fetal hand skin. the blot was washed and reprobed for 18S ribosomal RNA.parrr. 269 · 1 SEPTEMBER 1995 rRNA anoma LOX. For each series. lane Relative hTR concentration 10.15 ±0. The P values were calculated by unpaired t test. To determine if hTR is polyadenylated [poly(A)+] . TRC3 was renamed hTR for human telomerase RNA. which indicates that hTR is not polyadenylated. lane 2. Replicative senescence for the BJ line typically occurs at >PD 90. Clones from all isolates displayed identical morphology and growth profiles until 20 PDs. SCIENCE · VOL. '2 0._ 0 0 0=o _ o mined with quantitative RT-PCR (43). and the hTR was normalized to the 18S loading control. Euplotes. RT-PCR was analyzed for five normal telomerase-negative cell lines (lanes 1 to5) andfivetumor telomerase-positive cells lines (lanes 6 to 10): lane 1. comparison of hTR with the mouse homolog (48) reveals blocks of conserved regions with little similarity around the template. The TRF analysis on the foreskin clones was done at PD 60.46 0/13 4/4 0/8 No Yes No pBBS pBBS-hTR None 3. PCR products were labeled with 32p. lane 12. By PDs 23 to 26.09 Foreskin control No 7. As further controls for these experiments. Cells from the clones were harvested at PD 23. extracts from MuA but not from wild-type cells generated products in the modified telomerase assay (Fig. 4. Although higher concentrations of hTR were Table 1. lane 3. lane 6.71 2. The hTR remained in the poly(A).23 ± 0.40 + 0. However. NCIH23 lung carcinoma.5 Experiment 3 2.S I Plasmid Crisis Ratio 4. 3) and preliminary results indicating sensitivity to o-amanitin (38) suggest that hTR might be an RNA polymerase II transcript. eight clones expressing antisense hTR continued growing in a manner similar to that of the control cultures. and the average mean TRF lengths were determined. Testes and most cancer cells have telomerase activity. these results verify that TRC3 encodes the RNA ence of runs of T residues within the transcribed region of hTR (Fig. these cells underwent crisis. 47).- _. cerevisiae. not all of the clones generated from the pBBS-hTR-transfected HeTe7 cells underwent crisis. 3C). poly(A)+ transcripts were selected on oligo(dT) beads (38). 3D). as was the case with synthetically made (TTGGGG)n oligomer. Telomerase RNA was detected in all tissues. lane 11. To determine if hTR RNA is elevated in testis tissue and immortal cancer cell lines. Thus. Under these conditions. the pres- for glyceraldehyde phosphate dehydrogenase (GAPDH) was enriched in the poly(A)+ fraction. 3C).03 ± 0. The Northern hybridization signals were quantified on a Phosphorlmager. However. 28-30).11 2.13 1. B ~ R hT R concentration Relative ) c . c. The MuC product was shifted in register by two or four nucleotides. the ciliate telomerase RNAs are conserved in the template region (30). So m H - - o quantitations GAPDH . lane 7.22 0. leukemia U25x. at which point the growth of most of the antisense-expressing clones (p10-3hTR and pBBS-hTR) slowed. This putative difference in transcription may reflect the complexity of regulation of telomerase in mammals compared to that in the ciliates. control. the ratio of clones that underwent crisis is indicated in relation to the total.3 kb). Controls show that all PCR were in the linear range up to 23 cycles and that the signal was abolished if reverse transcriptase was left out (right two lanes). Taken together. so that this threefold difference was factored into our normalization value for testis tissue. Expression of hTR.20 Control No 3.61 1. None of the foreskin clones underwent crisis even after 38 doublings (PD 22 to PD 60). lane hTR §i p10-3 Mean Short - by 6% polyacrylamide gel electrphoresis. resolved p10O-3-hTR p10-3 plO-3-hTR pBBS pBBS-hTRa pBBS-hTRb None No Yes Experiment 1 - 0/7 18/18 No Yes Experiment 2 3. _ . 0£ L CC |. The relative signal is indicated below the bands with testis tissue again given the value of 100.14 2. antisense hTR or pBBS vector only were transfected into normal human foreskin BJ fibroblast cells at PD 22. 3D). which synthesized (TTGGGG)n repeats but typically generates a single band during the PCR amplification. Little or no identity was found between the primary RNA sequence of hTR and that of the ciliates or . s -. Expression of hTR in normal and immortalized cells and normal _ ' < i .5 No 7. whereas in most normal somatic human cells in adults telomerase is not detected (17).'" := E . whereas the message S 5 AST~i component of telomerase. breast tumor MCF7. Finally. mel- hTR . (A) The steady-state level of hTR and GAPDH RNA was deter- Fig. 3E). In contrast. (B) Northern RNA blot of hTR RNA prepared from human tissues (56). TRF size tissues. primary sinovial fibroblasts.! fy only (TTTGGG)n repeats and would not amplify repeats containing either wild-type (TTAGGG)n or MuC (TTGGGG)n (44). characterized by the appearance of enlarged and rounded cells.to sevenfold greater in the tumor cell lines than in primary cells when normalized to GAPDH messenger RNA (Fig.. The lack of polyadenylation and apparent transcription by RNA polymerase II suggests that hTR may be regulated by the specialized transcription complex characteristic of small nuclear RNA promoters (45). testes. H -U . C. primary fetal lung. and Oxytricha telomerase RNAs are apparently transcribed by RNA polymerase III (6. This confirms the sequence of the MuC products.30 ±0.fraction. ("C" in Fig. S. breast tumor MCF7 without RT. we analyzed hTR and GAPDH transcripts from five mortal primary cell strains that lack detectable telomerase activity and from five immortal cancer cell lines and testis tissue that have high telomerase activity. and lane 13. adult primary prostate. lane 5. The sequence of the MuC product was further confirmed by permutation of the MuC return primer (CCCCAA)3 in the PCR reaction by two or four nucleotides (Fig. As a control for loading. The Tetrahymena. colon tumor SW620. Of the small RNAs characterized to date. Note that testis tissue typically expresses one-third of the GAPDH with equal RNA loading.$apoPr. as does synthetic (TTGGGG). and quantified with a Phosphorlmager (Molecular Dynamics). .
Perhaps the protein compo. Proc. Because telomerase activity with vector alone or the telomerase-negais not detected in most somatic tissues.22. Similar tissue-specific differences in telo. none of the control clones (HeTe7 human tissues. 10. H. none of the vector control cell lines had any change in growth or mortality over 50 doublings. K. express. they were 27 to 31 percent shorter than those of the vector control clones. Biol. All p1 0-3-hTR B Antisense Control and pBBS-hTR clones were shown 4to express antisense transcripts.5Hinf and Rsa I. B. 458 (1990). Green. Harley. Cell crisis in the hTR antisense HeTe7 ure) that hTR is inactive in these tissues.0004) in the four clones that went into crisis. However. 181 (1973). G.22 to 2.5% c 2agarose gel. All colonies containing the control vector had mean TRF (terminal restriction fragment) lengths (3. J. D. 25. Blackburn.week. 345.30 and 3. Greider and E. E.23 kb (P = 0.3-~ tomegalovirus minimal promoter (54) or pBBS-hTR. 9. The DNA was probed Z 1.15 kb). The TRF gel was scanned with a Phosphorlmager. 4A) and indicate that telo. Nuclear DNA from the HeTe7 clones was purified. Net 13. Olovnikov. L. L. small ly immortalize in culture.with the control vector. Thus. we assayed telomerase activity in 14 of the clones in experiment 3. McClintock. Natl. Greider. 2. and telomerase activity was similar to that of controls. as the mean TRF fell from 3. and run on a 0.4-ing antisense hTR under control of the tetracycline-VP16-induced cy. though not as large as that ing antisense or control plasmids were sefound in testis tissue. [In these experiments. Mutat. B. D. Testis tissue. 197 (1972). we assayed telomere length (11) and telomerase activity (17) in several precrisis control and experimental colonies at 23 PDs after transfection. McGill. Res. W. 5. W. these results support the hypothesis that telomere loss leads to crisis and cell death once telomeres are shortened to a critical length. a number of lected in three separate experiments (Table other tissues also expressed hTR (Fig.pression) went into crisis. C.355).sense hTR coding sequence were also elecern blot analysis. To examine the functivity. 126 (1990). Nature 344. A. and tisense hTR grew the same as those of cells adult liver tissue.revertant colonies were observed when culnents of telomerase are stringently regulat. Taken together. In parallel. J. HeTe7 cells were also transfected with the parental vector under 1-~ ?i the same conditions. Size markers are 0. RNA. [ i mR (1939). 26 population doublings (PDs) after transThese results confirm the data from cell fection. Lindsey. Telomerase activity was generally low but detectable in many of the antisense clones (53). B. A horizontal line marks the average mean TRF for the antisense and control vector groups.. REFERENCES AND NOTES 1. ibid. H. Watson. D. 7. In addition. The revertants may represent Fig. In 28 out of 33 cases in which the merase RNA concentrations in mice. 346. In conmerase RNA is present in a number of trast.A.22 kb for experiments 2 and 3 in Table 1) similar to those in the parent cell line (3. many normal and then the rounding up and detachment mouse tissues have telomerase activity (49. 41. 3.present in the immortal cancer cells that could explain why mouse cells spontaneousexpressed high telomerase activity. P = 0. 269 · 1 SEPTEMBER 1995 . l I 1240 SCIENCE · VOL. These data suggest that telomere repeats were lost in the antisense-containing clones as a result of inhibition of telo- *I To test this directly. 1). C.cultures underwent crisis (Fig. Antisense hTR leads to a A Antisense Control shorter mean TRF and cell crisis. Blackburn. 271 (1991). H. Acad. Futcher. Nature 239. whereas human but readily detectable amounts of hTR cells do not (51). as expected.] As a control.5in the gel with an [a-32P]-labeled 10 (TTAGGG)3 oligonucleotide to label the telomeric TRF.5TRF analysis is shown for experi3ment 2. J. 405 activity. the 41 cultures expressing anThese include normal kidney. Ovary tissue had large troporated into HeLa cells. the tive foreskin cells with hTR antisense expresence of hTR suggests (barring assay fail. C. p10-3 pl0-3-hTR M C 1 2 3 4 5 6 7891011 1 2 3 4 5 6 78 91011C Three independent experiments were done in which HeTe7 cells (HeLa cells that express large io amounts of the tetracycline repres.cultures was characterized by a marked inmerase RNA expression can be seen in hibition in cell growth from 20 to 26 PDs mouse tissues (48). whereas clones containing the antisense vector constructs that underwent crisis had mean TRF lengths of 2. of cells from the plate over a period of 1 50). Lindsey. RNA were also present in mortal primary Expression of antisense hTR trancells that lacked detectable telomerase ac. Harley. no regulation by tetracycline was observed (57). B.23 kb-that is. 256. L. of hTR in was also examined control expression vectors lacking the antiExpression a variety of normal human tissues by North. telomerase activity parallels telo. To determine if telomere length and telomerase inhibition correlated with cell merase crisis in the antisense-expressing clones. expressing antisense hTR under the control of the MPSV promoter (54). Attardi. A. 8. H. N. Clones containamounts of hTR. which has high telo. U. Cell 43. A. which crisis event. Theor. but cells underwent crisis. However. Sci. at 23 to able amounts of telomerase activity (17). all of which lack detect. 4. Initially.S. and the mean TRF quantified as described previously (11). rare (<1 percent) not in humans. Muller. 181 (1938). 5. 4B).tures were maintained for 3 weeks after the ed in human but not mouse cells. although the presence of shortened telomeres suggests that the amount of activity was not sufficient to maintain telomere length. Collect. N.scripts in HeLa cells.tion of telomerase in an immortalized cell. 866 (1990). 5A). normal young foreskin fibroblasts with long telomeres and no detectable telomerase showed no effect of the antisense constructs for at least 38 doublings after transfection. merase activity.03 versus 3. J. ibid. cut with c 2. had large amounts of hTR we introduced antisense hTR expression constructs into HeLa cells (52). D. C. Yu.n 3. (A) . telomere length was not significantly changed (3. In the eight clones that contain the antisense vector (pBBS-hTRb) but did not enter crisis. (B) l l Jill I II II I I II l IlJl The mean TRF is shown for each 1 2 3 4 5 6 7 891011 1 234567891011 antisense and control clone in experiment 2.40 and 2. 33 out of 41 antisense-expressing lines (Fig. J. Hastie et al. The induction of cell crisis in HeLa cells expressing antisense hTR demonstrates the potential of telomerase inhibition as a therapeutic approach for treating human cancer.5shown on the left in kilobases. 405 6.6sor-VP16 chimeric protein) were transfected by electroporation with 5-A either plasmid p10-3-hTR. In contrast to the antisense clones. However. prostate. I. Bradley. variants that escape the inhibiting effect of the antisense hTR construct. (1985).
167. S. Sci. 17 (1993)] with the use of a TRC3-selected genomic lambda clone labeled with biotin-16-deoxyuridine triphosphate (dUTP) as a probe. 50. U. ibid. N. C. Vaziri etal. B. S. S.A. Nature Med. 34. M. Counter. Chem. C. 42. C. Siegel. J. Natl. Science 269. M. With the use of the specific MuA (ACCCAA)4 or MuC (AACCCC)3 return primers. S.6) were observed in the genomic Southern blot. 58. 80 (1992). (1994). 281-313. Eds. B. Proc. then by sedimentation through a glycerol gradient. made into double-stranded cDNA with reverse transcriptase. H. D. F. Dahir. A. respectively. S. Natl. 1 are as follows: TRC3.Am. Total 293 RNA was prepared by guanidinium thiocyanate and phenol chloroform extraction.5% agaroseformaldehyde gel and transferred onto Hybond N.A. 15. Pikaart.A..4. G. An oligonucleotide (NotA: 5'pATAGCGGCCGCTTGCCTTCA) was ligated to these first-strand products with T4 RNA ligase. 30 s. 10. MuC. E. 1992).B. R. Cell 59. W. Plasmid p10-3-hTR expresses the antisense of hTR under the control of the tetracycline-regulated cytomegalovirus minimal promoter [M.. HeTe7 cells that expressed the antisense vector and went into crisis had some onefifth of the telomerase activity found in vector controls. The cDNAs that hybridized to this biotinylated oligonucleotide were captured with streptavidin-magnetic beads (Promega) and reamplified by PCR. W. We estimate that in our highest expressing clones -10 percent of the wild- 45. Marshallsay. Cancer Genet.A. Prowse and C. and tissue RNAs were obtained from Clonetech. J. total RNA from telomerase-positive 293 cells or telomerase-negative IMR90 cells was size-selected on an agarose gel. 41.j. TRC3 runs at -450 bases on Northem blots. All PCR conditions were checked for specificity with synthetic MuA. TRC1. and C. M. G. C. Genet. C. Gossen and H. which places the putative 3' end of the mature TRC3 transcript at +450. and PCR was then done with a nested primer R3c (5'GTTTGCTCTAGAATGAACGGTGGAAG) and an oligonucleotide (NotB: 5'-TGAATTCTTGCGGCCGCTAT) complementary to the NotA oligonucleotide. 5. Broccoli. C. PCR products were resolved on agarose gels. Cancer Res. Southern blotted. Bacchetti.. or Spe plus Xba I. 91. S. p. Gynecol. unpublished results. Genes Chromosomes Cancer 6. Kim et al. K. F. 4818 (1995). telomerase from S100 extracts of 293 cells was partially purified with DEAE ion exchange and sephacryl sizing columns. Botelho. J. W. 30 s). Genet. Morser.S. which has limited use as a quantitative assay. M. except that the annealing temperature was 50°C. Proc. 0. J. Genes Dev. Blackburn. C.S. 16. B. 2 May 1995. Eco RI plus Hind III. H. NY.3. J. P. Greider. The PCR primers for Fig. Natl. N. 38. 2315 (1995). H. Proc. Perez. Lingner. A 200-bp Eco RI DNA fragment containing nucleotides 1 to 185 of TRC3 was inserted into the Eco RI sites of p10-3 and pBBS212 to make the plasmids p10-3-hTR and pBBS-hTR. King. J. Wistrom. ibid. U. Le. Roos. 23. B. 13.. Gottschling. 11. 269. L. J. 53. or growth rate.W. J. Acad. J. Exp. 47. A. 554 (1995). D. J.97. C. ACCCTAACAGCTAAATCAAGGACC and GGAGTGCAAGACCAAGGGTTCTGG. Acad. Sci. Nature Genet. 44. Funk. Acad. In the second approach. 33. Andrews. Because of the small cell populations. 12. 35. 29. The restricted DNA was run on an agarose gel. Windle. (Cold Spring Harbor Laboratory Press. Holzmann et al. Harley. Hiyama et al. The PCR conditions cells.S. Young. G.0. 14. U. A. Proc. and probed with an [a-32P]-labeled TRC3 cDNA fragment of 186 bp. B. In this case. The method for the 3' end mapping was as follows. Barber. Schwartz. 249 (1995).404 (1994). Natl. Feng. 55. Am. 51. A. in press. D. We thank G. 37. Avilion. C. W. Smith. J. 331 178(1993). 3410 (1994). 27. Cell 65. L. W. Cell 67. J. 60°C.S. 161(1992). T. Cell.-H. Cold Spring Harbor. Harley. 36. P. R. C. D. H. Feng. 2496 (1991). M. 1267 (1995). Cech. Coulsen. U. the PCR products specifically indicated that these were diagnostic for MuA and MuC DNA sequences. Hiyamal et al. (1989). J. Pst I. identified by autoradiography. As expected with such a low level of mutant telomerase. W. To enrich for potential telomerase RNAs. no mutant telomerase products were generated from extracts of parental cells or cells transfected with the wild-type TRC3 gene. Counter. 59. Cooke. For MuA. the cells still underwent crisis at 23 to 26 PDs in the presence of tetracycline. 48. Blackburn. 21. TRC4. S. Feng. H. J. The method of the 5' end mapping was as follows. a PCR-amplifiable cDNA library from telomerase-positive cells was hybridized to an excess of PCR-amplified cDNAs produced from telomerasenegative cells. J. ligated to oligonucleotide arms. C. The blot was probed with an [a-32P]-labeled RNA complementary to the human telomerase RNA. de Lange.G. McEachern and E. Smith and G. 32. Acad. 52. used for MuC were the same as those for MuA. The expression level of the inactive 17-base insert RNA in the analyzed clones (MuC*) was >10-fold higher than the endogenous RNA level in other clones that were analyzed. B. L. 517 (1991). Lin. 1883 (1992). Biol. 39. Greider. H. Thus. 13. 18. C. A. we followed two procedures: subtractive hybridization to remove high-abundance common sequences or partial telomerase purification. J. Connelly. R. ibid. and wild-type telomerase products.661 (1993). Shippen-Lentz and E. McKnight and K. 563 (1994). RNA was extracted from purified telomerase preparations and was then used in the following experiments. Hum. and extracted. Sci.841 (1995).S. J. H. K. 1. Sanger. and 1. Virol. Science 247. Harrington. Filipowicz. P. Plant Mol. Dev. accepted 10 August 1995 SCIENCE · VOL. 36. 202. 10 s. H. Cytogenet. EMBO J. Science 266. DNA was isolated from 293 cells and digested with Bgl II plus Eco RI.2. 8. R. Plasmid pBBS-hTR expresses the antisense of hTR under the control of the MPSV promoter (54). This labeled TRC3 probe also specifically hybridized to chromosome 3 in a somatic cell hybrid panel. 8. Oncogene 11. Chadeneau. K. 4. K. 25. Avilion. R. Both transient and stable transformants were generated. T. T. G. and sequenced directly with oligonucleotide as a primer. Blackburn. Counter. telomerase activity was assayed for the samples in experiment 3 by the PCR method (17). J. L. Am. Control experiments showed that subtraction hybridization of the 293 cell library by the IMR90 cell library depleted the telomerase-positive 293 cell library of most high-abundance common sequences.. McDougall. in Proc. Dyer. S. Klingelhutz. A. A. A. 40. U. and RNase P. DeLange et al. C. Natl. 52.A. B. Genes Dev. Reverse transcriptase-sensitive PCR bands of the predicted size were seen with cDNA made with all primers designed to the interval +96 to +446 but not with any of the primers designed to +554 to +946. Huang and P. and amplified by PCR. ligated to oligonucleotide arms. 5463 (1977).. Nicklen. K. 518 (1990). excised. A cDNA library was prepared from the total RNA of the peak gradient fractions. Genet. >2. G1 (5'-AGAAAAACAGCGCGCGGGGAGCAAAGCA) 54. B. Vilander. 61. A. Biol.A. W. Morin. Science 266. Wang. Wang. J. Hernandez. C. N. 22485 (1994). Bacchetti. Holmes. M. Greider and E. Singer and D. because control experiments with luciferase or antisense hTR under the control of the tetracycline-VP16-induced cytomegalovirus minimal promoter showed that the presence or absence of tetracycline had little effect on luciferase or antisense hTR expression in HeTe7 type telomerase activity was converted to mutant activity. Total RNA (30 jg) was loaded onto a 1. D. Hirte. First-strand cDNA was made with antisense TRC3 primers (R3B: 5'-CAACGAACGGCCAGCAGCTGACATT) in the presence of [a-32P]dATP. Nature. 20. Blackburn.. 43. 17. In contrast. Obstet. 56. 22. The experiments with p10-3-hTR were done in the absence of tetracycline. Supported by Geron Corporation and by grant AG09383 (C. in press. Parra and B. Funk and J. Mol. W. 2011 (1994). W. A. Sci. 19. Villeponteau. Bacchetti. 1. Buiard. R. In early experiments with antisense hTR constructs in HeTe7 cells. W. R. Stony Brook (1995). Yu and E. in Transcriptional Regulation. Mehle. Counter et al. CTAACCCTATCTGAGTTGGGCGTA and CAACGGACAGACAGCAGCTGACAT.. Muller. Yeh. Allsopp etal. W. Natl. Greider. Andrews. W. L. Eco RI alone. Cancer S * &m 10114(1992).. Blasco. Nilsson. Autexier and C. Villeponteau. R. and 72°C. C. 71. 546(1990).S. colony number. 11. 2900 (1994). An RT-PCR strategy was first used in which a series of antisense primers complementary to the genomic DNA sequence primed first-strand cDNA synthesis. 269 · 1 SEPTEMBER 1995 1241 . Filipowicz. Ho for skilled technical assistance. TTGCTCATGAGGTCATGTCCCCGGAA and GGAGGAATTGCCTTAGGGTTAG. Kiss. S. RTPCR was used to measure endogenous and mutant TRC3 amounts. the PCR went for 27 cycles (94°C. B. As to the levels of mutant template hTR (MuC or MuA) in transfected cells compared with the endogenous hTR levels.S. Hendrick. Oncogene 9. Harley. 287 (1994).) from NIH. Hind III alone. 28. Sci. E. Adamson. Greider. W. The primers in this series were spaced at approximately 150-bp intervals starting from the 5' end of the transcript. excised. 30. 45. 973 (1992) M. pp. M. The chromosome 3 location of the TRC3 was narrowed to the distal end of 3q by chromosome in situ suppression hybridization [I. A. Harley. 9857 14. 961 (1994). 1984 (1994). Proc. State University of New York.. 74. Remes. Gupta. Leber. Telomerase RNA cDNA candidates were selected by hybridization to the biotinylated 2'-O-methyl RNA oligonucleotide (5'-UUAGGGUUAGII-3' biotin. 1921 (1992). in which is inosine) that contained two full copies of the predicted telomeric repeat in the human telomerase template. Acad. 49. 57. 68. 3043 (1994). Butler. Irving. Villeponteau. K. although longer TRC3 transcripts out to +541 have been detected by RT-PCR. Sci. Sizes expected for TRC3 based on the lambda clone (1. Yamamoto. 31. Extension products were resolved by polyacrylamide gel electrophoresis.132 (1993). H.-L.. W. Wydro. we did not notice any effect on HT1080 viability. K. P. W. 26. M. C. 92. 5547 (1992)]. Chromosome 3 was identified with a chromosome 3-specific centromeric DNA (which was labeled with digoxigenin-11-dUTP) as a probe. 91. C. 823 (1991). T. Haites. H. Vaziri et al. Bacchetti. Proc. M. U. Cancer 68. S. Assoc. 89. 19. those antisense clones that did not go into crisis averaged three-fifths of the telomerase activity found in the control vector clones. Miller. H. 46. Nature 337. Greider. 521 (1989). Marshallsay. 24. C. G. GGAAGGTCTGAGACTAG and ATCTCCTGCCCAGTCTC. M. 89. Cytogenet. E. Acad.H. C. Blood 85. Cell Res. Gene 147. unpublished data. PCR was then done with the first-strand primer and a primer whose sequence was internal to the known TRC3 transcript (F3b: 5'-TCTAACCCTAACTGAGAAGGGCGTAG). Biol. 204 (1992). thesis. we did not observe much change in the overall amount of hTR in any of our mutant cells (with the exception of the inactive MuC* mutants). In the subtraction approach. and amplified by PCR. P..