Oncogene (1998) 16, 1345±1350  1998 Stockton Press All rights reserved 0950 ± 9232/98 $12.00

Cloning and characterization of human and mouse telomerase RNA gene promoter sequences

Jiang-Qin Zhao1, Stacey F Hoare1, Robert McFarlane2, Sharon Muir1, E Kenneth Parkinson2, Donald M Black2 and W Nicol Keith1

1CRC Department of Medical Oncology, University of Glasgow, CRC Beatson Laboratories, Alexander Stone Building, Garscube Estate, Switchback Road, Bearsden, Glasgow G61 1BD; 2Beatson Institute for Cancer Research, CRC Beatson Laboratories, Garscube Estate, Switchback Road, Glasgow G61 1BD, UK

Variation in telomerase activity is correlated with 1995; Nakamura et al., 1997; Soder et al., 1997b,c). cellular senescence and tumour progression. However, Whilst telomerase expression is detectable in normal although the enzymatic activity of telomerase has been embryonic tissues and germline stem cells, telomerase well studied, very little is known about how expression of expression is repressed in most normal postnatal telomerase genes is regulated in mammalian cells. We somatic cells (Blasco et al., 1995; Feng et al., 1995; have therefore cloned the promoter regions of the human Prowse and Greider, 1995; Soder et al., 1997a; Wright (hTR), and mouse, (terc), telomerase RNA genes in et al., 1996). The lack of telomerase expression may order to identify the regulatory elements controlling be the major reason for the progressive loss of telomerase RNA gene transcription. 1.76 kb encompass- telomeric sequences in somatic cells, which is ing the hTR gene promoter region was sequenced, as was considered to be one regulatory mechanism which 4 kb encompassing the terc promoter. No signi®cant monitors the number of times a cell divides before sequence similarity could be detected in comparisons entering replicative senescence (Campisi, 1997). How- between human and mouse 5'-regions, ¯anking the ever, although telomerase appears to be stringently transcribed sequences. However, both the human and repressed in normal somatic tissues, there is sub- mouse telomerase RNA genes are within CpG islands stantial evidence to suggest that telomerase is and may therefore be under the regulation of DNA expressed in the majority of human cancers and methylation. Transient expression of hTR-reporter gene contributing to the immortal phenotype through the constructs in HeLa and GM847 cells identi®ed the maintenance of telomere integrity (Holt et al., 1997; elements responsible for promoter activity are contained Kim, 1997; Shay and Bacchetti, 1997). in a 231 bp region upstream of the transcriptional start The regulation of telomerase expression is likely to site. Transient expression of terc-reporter gene con- be a complex issue including transcriptional activity of structs in Swiss3T3 and A9 cells identi®ed the elements the telomerase RNA and protein component genes, responsible for promoter activity are contained in a interaction of telomerase with other telomere asso- 73 bp region upstream of the transcriptional start site. ciated proteins and post-translational modi®cation of These studies have implications for novel transcription the enzyme complex. However, at present there are few targeted cancer therapies. studies which directly address the mechanisms regulat- ing telomerase expression in normal and cancer cells Keywords: telomerase; gene expression; hTR; terc; gene (Bodnar et al., 1996; Broccoli et al., 1997; Li et al., promoter; gene transcription 1997; Mandal and Kumar, 1997; Morin, 1997; Nakamura et al., 1997; Soder et al., 1997a). To identify elements important for the regulation of telomerase RNA genes, we have cloned genomic Introduction DNA sequences encompassing both the human and mouse telomerase RNA genes, including 5'-¯anking Telomeres are found at the end of linear chromosomes regions. and consist of short repetitive sequences essential for the maintenance of normal chromosome structure and function (Wellinger and Sen, 1997). With each cell Results division, telomeres shorten due to the inability of DNA polymerases to replicate the ends of linear DNA Cloning of genomic sequences encompassing the human molecules. However, telomere erosion is counteracted and mouse telomerase RNA genes by the activity of the enzyme telomerase, a ribonuclear protein with reverse transcriptase activity, which adds We have previously reported the identi®cation of P1 telomeric repeats to the chromosomal termini (Morin, genomic clones for the human, (hTR), and mouse, 1997; Nakamura et al., 1997). The genes for the (terc), telomerase RNA genes (Soder et al., 1997b,c). In human, (hTR), and mouse, (terc), RNA components order to obtain sequences ¯anking the genes, the P1 have recently been cloned, as has the human protein genomic clones were digested with EcoRI and HindIII, component; (hTRT) (Blasco et al., 1995; Feng et al., subcloned into the plasmid vector, pBluescript, and colonies containing hTR or terc sequences identi®ed by hybridization to PCR generated probes speci®c for the genes. A 1.3 kb genomic clone encompassing hTR was Correspondence: WN Keith Received 17 December 1997; revised 3 February 1998; accepted 5 isolated as was a 4 kb genomic clone emcompassing February 1998 terc (Figure 1). Telomerase RNA gene promoters J-Q Zhao et al 1346 A BLAST search using the 4 kb mouse sequence identi®ed both the published human gene sequences, (HSU85256, HSU86046) and the published sequence for the transcribed region of the mouse gene, MMU33831, (590 bp of sequence) (Blasco et al., 1995; Bryan et al., 1997; Feng et al., 1995). In order to con®rm that the genomic sequence obtained from the P1 subclone was genuine, we cloned 5'-¯anking sequences using genomic DNA from Balb/c mice in PCR reactions. Sequence analysis of the Balb/c clones were identical to the P1 sequence except for minor polymorphisms (data not shown). Thus we have extended the published sequence for the coding region of the mouse telomerase RNA gene in both the 3' and 5' ¯anking regions of terc. A schematic representation of the 4 kb of sequence information encompassing the mouse telomerase RNA gene is shown in Figure 1.

Figure 1 Restriction enzyme map of the genomic clones Analysis of nucleotide sequence encompassing the human encompassing the human and mouse telomerase RNA genes. and mouse telomerase RNA genes The transcribed regions of hTR and terc are depicted as black boxes within the central regions of the genomic sequences and the To investigate the relationship between the human and site of the template sequence within the telomerase RNA genes is mouse genomic clones, sequence comparisons were indicated. The position of the CpG islands are shown as a box carried out. The transcribed regions of the two genes beneath the genomic sequence. Numbers in brackets refer to the showed 67% identity in keeping with the published nucleotide position within the sequence. The 3'-end of all the estimate (Feng et al., 1995). However, no signi®cant human promoter fragments is shown as hProm and fragments extend 5-prime to hProm867, hProm697, hProm341 and sequence identity could be identi®ed in either the 5'-or hProm111. The 3'-end of all the mouse promoter fragments is 3'-regions ¯anking the transcribed sequences. shown as mProm and fragments extend 5-prime to mProm628, Both the human and mouse sequences were analysed mProm458, mProm418, mProm267, mProm208, mProm136. The for CpG islands by GRAIL. CpG islands were de®ned numbers after the pre®x hProm or mProm refer to the number of nucleotides of genomic sequence contained in the promoter as regions larger than 200 bp, with an average GC fragment content greater than 50% and the ratio of observed versus expected CpGs greter than 0.6 (Gardiner- Garden and Frommer, 1987). Interestingly, both the human and mouse genes lie within CpG islands, (see A BLAST search using the 1.3 kb human sequence Figure 1). The human gene is covered by a CpG island identi®ed three high-scoring segment pairs: HSU85256, 733 bp in length, with a GC content of 66% and a HSU86046 and MMU33831. HSU85256 (598 bp of ratio of observed versus expected CpGs of 0.89. The sequence) and HSU86046, (545 bp of sequence), are mouse gene is covered by a CpG island of 659 bp in published sequences for the transcribed region of the length, with a GC content of 64% and a ratio of human telomerase RNA gene and con®rmed that we observed versus expected CpGs of 0.81. had cloned genomic sequences encompassing hTR The 5'-¯anking regions of the human and mouse (Bryan et al., 1997; Feng et al., 1995). MMU33831 is telomerase RNA genes were also analysed for potential the sequence of the transcribed region of the mouse transcription factor recognition sites. As shown in telomerase RNA gene which has previously been Figure 2, a number of potential binding sites can be shown to have homology to the human gene (Blasco identi®ed, including consensus sequences for glucocor- et al., 1995). In addition to the BLAST search, a text ticoid/progesterone/androgen receptor binding, AP1 search for telomerase sequences carried out on the and Ets family members. CCAAT box's are found in NCBI database revealed a 2.4 kb sequence, (I31750), both genes close to the published transcriptional start from US patent number 5583016, (Inventors: Ville- sites (Blasco et al., 1995; Feng et al., 1995), however, ponteau B, Feng J, Funk W and Andrews WH, there is no obvious TATA box in the mouse gene with assigned to the Geron Corporation), which aligned the human gene TATA box consensus sequence being with the HindIII site at the 3'-end of our genomic clone in the reverse orientation. The mouse promoter region and extended 5' to the EcoRI site in our clone and thus also contains a run of CpA dinucleotide repeats which completely overlapped the sequence of our 1.3 kb clone may be of use in developing microsatellite genetic with only minor base di€erences. We extended the markers for this gene. sequence of our P1 clone 5' to the EcoRI site by designing a PCR primer, hProm867 from the patent Transfection assays detect promoter activity in the sequence and cloning and sequencing the PCR product 5'-¯anking regions of the human and mouse telomerase using the P1 clone as a template. In total we have RNA genes generated 1765 bp of sequence information encompass- ing the human telomerase RNA gene and a schematic To identify whether the 5'-¯anking DNA of the representation of this is shown in Figure 1. Thus we telomerase RNA genes exhibited promoter activity, have extended the sequence in both the 3' and 5' sequences were fused to a ®re¯y luciferase reporter ¯anking regions of hTR and con®rmed the unpub- gene, (pGL3-Basic). The transcriptional start sites for lished patent sequence. both the human and mouse telomerase RNA genes Telomerase RNA gene promoters J-Q Zhao et al 1347

Figure 2 Nucleotide sequence of the human, (a), and mouse, (b), telomerase RNA gene 5'-¯anking regions. Putative regulatory motifs are underlined. Arrows indicate the transcriptional start sites (Blasco et al., 1995; Feng et al., 1995) and numbers to the left of each Figure refer to the number of bases upstream of the transcriptional start site. The template regions are in bold and underlined. Sequences contained in promoter constructs are shown by vertical lines and labelled, hProm or mProm (see Figures 1 and 3). The regions containing elements responsible for minimum promoter activity are highlighted in bold (see text for details). The run of CpA dinucleotide repeats in the mouse promoter is shown in bold and italic have been established (Blasco et al., 1995; Feng et al., with fragments containing 208 bp or more (from 1995). Various promoter constructs containing the position 794, see Figure 2b). Construct mProm136, transcriptional start site were therefore generated (see which contains only 136 bp of 5'-¯anking sequence Figures 2 and 3). (position 722, see Figure 2b), produced dramatically Human promoter constructs containing truncated reduced levels of luciferase activity (Figure 2b). Thus, a portions of the 5'-¯anking region were transiently minimal promoter sequence can be de®ned as transfected into HeLa and GM847 cells (Figure 3a). extending 94 bp upstream of the transcription start HeLa is a telomerase positive cervical carcinoma cell site, and that elements responsible for promoter line, GM847 is a SV40-immortalized skin ®broblast cell activity must be contained in a 73 bp region between line which expresses the telomerase RNA component 794 bp and 722 bp (Figures 2b and 3b). but is telomerase-negative (Bryan et al., 1997). As Transfection of the human promoter construct shown in Figure 3a, promoter activity was observed in hProm867 into mouse cells gave very strong promoter both cell lines with fragments containing 341 bp or activity, with up to twice that of the strongest mouse more (from position 7272, see Figures 2a and 3a). The construct, and transfection of the mouse promoter highest luciferase activity was observed with construct construct, mProm628 into human cells also showed hProm505 which contains a 505 bp fragment, (position luciferase activity at around 25% of the strongest 7463, see Figure 2). Construct, hProm111, which human construct (data not shown). contains only 111 bp of 5'-¯anking sequence (position 742, see Figures 2a and 3a), produced a dramatically reduced level of luciferase activity (Figure 3a), and is Discussion not sucient on its own for promoter activity. Thus, a minimal promoter sequence can be de®ned as The levels of telomerase RNA gene expression vary extending 272 bp upstream of the transcription start during normal development and between normal and site, and that elements responsible for promoter cancerous cells and tissues (Avilion et al., 1996; Bestilny activity must be contained in a 231 bp region between et al., 1996; Blasco et al., 1995, 1996; Bodnar et al., 1996; 7272 bp and 742 bp, (Figures 2a and 3a). Broccoli et al., 1996; Feng et al., 1995; Kuniyasu et al., Mouse promoter constructs containing various 1997; Soder et al., 1997a). Knowledge of telomerase truncated portions of the 5'-¯anking region were RNA gene expression should therefore aid our under- transiently transfected into Swiss3T3 and A9 cells standing of the signal transduction pathways linking (Figure 3b). Swiss3T3 cells are an embryo derived line telomere attrition to proliferation, cellular senescence, and A9 cells are of areolar and adipose origin. Both di€erentiation and oncogenesis. As a ®rst step towards cell lines are telomerase positive. As shown in Figure this goal, we have cloned the promoter regions of the 3b, promoter activity was observed in both cell lines human (hTR) and mouse (terc), telomerase RNA genes Telomerase RNA gene promoters J-Q Zhao et al 1348 of transcription factors implicated in haematopoiesis and leukaemogenesis such as GATA-1, PU.1, PEA2/ PEBP2, C/EBP and c-Ets-2 (Tenen et al., 1997). These data will therefore be of interest in examining the basis for the detection of telomerase activity in normal and malignant haematopoietic cells (Bodnar et al., 1996; Cheng et al., 1997; Holt et al., 1997; Norrback and Roos, 1997; Pan et al., 1997). The human and mouse telomerase RNA genes do share an interesting similarity, in that they both lie in CpG islands, and thus their expression may be regulated by methylation. DNA methylation is thought to be important for gene regulation during normal development and cellular senescence, and abnormal methylation patterns may be a fundamental change in tumour progression (Baylin et al., 1991; Bird, 1996; Laird and Jaenisch, 1996; Vertino et al., 1994; Wilson and Jones, 1983). Thus it has been suggested that aberrant CpG island methylation during normal ageing process, could contribute to immortali- zation by interfering with expression of `mortality' genes, of which hTR and terc can be included (Vertino et al., 1994; Wilson and Jones, 1983). We are currently Figure 3 Detection of promoter activity in the 5'-¯anking testing the relationship between DNA methylation and regions of human and mouse telomerase RNA genes. For each telomerase RNA gene expression. construct, the length of sequence upstream from the transcrip- tional start site is shown to the left and the luciferase activity to Turning to the functional analysis of the cloned the right. (a) Diagram comparing luciferase activity from human sequences, the minimal promoter for hTR resides promoter constructs in GM847 and HeLa cells. Data for each within a region of 272 bp upstream of the published construct is plotted as a percentage of the hProm505 luciferase transcriptional start site (Feng et al., 1995) (Figures 2a activity as this construct consistently gave the highest activity in and 3a). There are a number of potential transcription human cells. For each construct the mean and standard deviation for duplicate transfected wells is shown. (b) Diagram comparing factor binding sites in this region including consensus luciferase activity from mouse promoter constructs in Swiss3T3 sequences for AP1, Sp1, PEA2/PEBP2, PEA3 and and A9 cells. Data for each construct is plotted as a percentage of PU.1. Interestingly, the expression in the fos/jun family the mProm458 luciferase activity as this construct consistently of proteins, which determine AP1 activity, are gave the highest activity in mouse cells. For each construct the mean and standard deviation for duplicate transfected wells is suppressed during the onset of senescence and would shown be predicted to lead to a reduction in AP1 activity in senescent cells (Campisi, 1997; Irving et al., 1992; Riabowol et al., 1992; Seshadri and Campisi, 1990). AP1 also responds to protein kinase C, and it has recently been demonstrated that hTR expression is in order to identify the regulatory elements controlling induced by protein kinase C during T-cell activation telomerase RNA gene transcription. (Bodnar et al., 1996). Extending the promoter region to In general, there are no signi®cant sequence 463 bp upstream of the transciptional start site, homologies between the promoter regions of the increases the luciferase activity to its maximum level human and mouse telomerase RNA genes. Indeed, (Figures 2a and 3a). This region contains several there is considerable debate as to whether telomere consensus binding sites for glucocorticoid/progester- length is regulated in a similar fashion in humans and one/androgen receptor binding, which may contribute mouse (Blasco et al., 1997; Kipling, 1997a,b; Zakian, to the maximal activity demonstrted by hProm505. A 1997). However, mouse models represent a valuable reduction in promoter activity is observed on extending resource with which to study the role of telomerase in the promoter fragments to include more 5'-sequence cellular senescence and tumour progression and mouse (Figure 3a), suggesting that sequences towards the 5'- models are likely to be required to investigate new end of the clone may in¯uence promoter activity in a therapies based on telomerase inhibition. In addition, negative fashion. the developmental regulation of telomerase will be The minimal promoter for terc resides in a 94 bp more easily approached in mice (Bestilny et al., 1996; region upstream of the published transcriptional start Blasco et al., 1995, 1996, 1997; Broccoli et al., 1996; site (Blasco et al., 1995), (Figures 2b and 3b). A Prowse and Greider, 1995). Thus, any di€erences striking feature of this region is the presence of three between the two species may in fact aid our under- AP-2 consensus sites, two of which are coupled to standing of the function for telomerase in maintaining c-Ets-2 sites and all these elements are contained in the genome stability and may be important in developing 73 bp region required for promoter activity (Figures 2b good murine models for human disease or develop- and 3b). Oncogenic Ras gene signalling has been mental processes. Whether the sequence divergence shown to operate through c-Ets-2 binding sites, thus between the human and mouse gene promoters has any there is a testable relationship between oncogene functional consequences can now be studied. activation during tumour progression and telomerase Despite the lack of sequence similarity between the RNA gene transcriptional activity (Galang et al., 1994; human and mouse telomerase RNA gene promoter Wasylyk et al., 1994). A reduction in promoter activity regions, they both have consensus sites for the binding is observed on extending the promoter fragments to Telomerase RNA gene promoters J-Q Zhao et al 1349 include more 5'-sequence, Figure 3b, mProm628), and the mouse P1 clone, 11792, is derived from a mouse suggesting that sequences towards the 5'-end of the C127 ®broblast P1 library. Brie¯y, in order to subclone the clone may in¯uence promoter activity in a negative promoter regions, the P1 clones were digested with EcoRI fashion. and HindIII and ligated into pBluescript. Colonies contain- These studies have a number of implications for the ing telomerase RNA gene sequences were identi®ed by hybridization with PCR generated probes as previously development of new transcription based therapies for described (Soder et al., 1997b,c). Plasmid DNA was cancer (Cai et al., 1996; Connors, 1995; Miller and prepared from positively hybridizing colonies, and inserts Whelan, 1997; Peterson and Baichwal, 1993). Directly sequenced on both strands by dideoxy chain termination down-regulating expression of the telomerase RNA using the ABI PRISM dye teminator cycle sequencing kit gene through manipulation of transcription factors (PE Applied Biosystems, Warrington, UK) and 25 ng should be an e€ective anticancer therapy and the oligonucleotide primers. Dye labelled products were cloning of the hTR gene promoter will allow the resolved and detected using the Applied Biosystems DNA analysis of therapeutic molecules which modulate hTR sequencer ABI373. Sequence was analysed using the promoter activity (Cai et al., 1996; Peterson and Sequencing Analysis program v3.0. Homology searches Baichwal, 1993; Sharma et al., 1997). Indeed, by using carried out using BLAST (Basic Local Alignment Search Tool), National Centre for Biotechnology Information a human cell line which has telomerase activity (NCBI): http://www.ncbi.nlm.nih.gov/. Sequence was ana- (HeLa) and one which expresses the hTR gene but lysed for potential transcription factor binding sites by is telomerase negative (GM847) (Bryan et al., 1997), TESS: Transcription Element Search Software on the we have established a system in which the speci®city WWW, Jonathan Schug and G Christian Overton, of transcriptional manipulation of hTR can be Technical Report CBIL-TR-1997-1001-v0.0, of the Compu- examined. In comparison to HeLa, the growth of tational Biology and Informatics Laboratory, School of GM847 does not appear to be dependant on Medicine, University of Pennsylvania, 1997, http:// telomerase expression, thus transcriptional targeting agave.humgen.upenn.edu/tess/index.html. Identi®cation of of hTR in GM847 should have no cellular e€ects, CpG islands was carried out using GRAIL: Gene Recogni- whereas HeLa should be sensitive to the predicted tion and Assembly Internet Link, http://compbio.ornl.gov/ Grail-1.3/. The full sequences have been submitted to anti-proliferative e€ects of the transcriptional target- GenBank, accession numbers AF047386, AF047387. ing. In addition, we have recently shown tumour- speci®c patterns of hTR gene expression with clear di€erentials in expression between cancerous and Construction of luciferase reporter gene constructs adjacent normal tissue (Soder et al., 1997a,b). There- The structures of the telomerase RNA gene-luciferase fore, the hTR gene promoter may be of considerable constructs used in the study are shown in Figures 2 and use in genetic therapies requiring selective expression 3. Promoter-luciferase constructs were made by inserting of therapeutic genes in cancer cells expressing high PCR products into pGL3-Basic (Promega). Orientation levels of hTR (Connors, 1995; Miller and Whelan, and sequence of each insert was checked by sequencing. 1997; Soder et al., 1997a). The constructs described in Details of the primers used for cloning are available from the present study will be ideal for more focused gene the authors on request, ([email protected]). therapy studies and the development of transcription based therapeutics. Transfection and luciferase assays In conclusion, by de®ning the human and mouse telomerase RNA gene promoters we are now in a All transfections were carried out in duplicate in 6-well 4 position to identify the factors which repress or plates, (35 mm diameter). Cells were seeded at 6610 cell per well and cultured overnight. Transfection was carried activate hTR and terc expression during normal out using SuperFect Transfection Reagent (Qiagen), development and senescence and determine whether according to the manufacturers instructions. Cells were these pathways di€er in the regulation of telomerase exposed to the transfection mix for 3 h and harvested for RNA gene expression in cancer progression. We can analysis after 48 h. Equivalent amounts of cellular protein now begin to address the question of whether signals as determined by Bio-Rad assay (BioRad), were used in the arising from critically short telomeres directly lead to luciferase assay. Luciferase assays were performed accord- telomerase activation and whether telomerase RNA ing to the manufacturers protocols (Promega). To ensure gene expression is directly linked to the oncogenic reproducibility in the assays, particular care was taken over process. These studies are critical to our understanding the following: DNA used for transfection was quanti®ed by of how telomerase activity is repressed or activated and spectrophotometry and direct visualisation by gel electro- phoreses. All transfections were carried out in duplicate to the identi®cation of the cell signalling pathways wells and we found this to be a good measure of the which are utilized to co-ordinately regulate the reproducibility of transfection. In each experiment, all expression of the genes encoding the telomerase deletion constructs were analysed together with both the enzyme complex. basic cloning vector, pGL3-Basic and the positive control vector, pGL3-Control, which contains SV40 promoter and enhancer sequences. Each extract was measured for luciferase activity at least twice. All transfections were Materials and methods carried out at least three times. Initial transfection conditions were determined by using promoter fragments Cloning of sequences encompassing the human and mouse liked to a green ¯uorescent protein reporter gene telomerase RNA genes (Clontech), as this allowed direct visualisation of promo- ter activity in live cells (data not shown). We found it We have previously reported the identi®cation of genomic important to transfect and analyse the cells at sub- clones in P1 vectors containing hTR and terc transcribed con¯uence and that it was important to avoid harsh sequences (Soder et al., 1997b, c). The human P1 clone, transfection protocols such as electroporation, resulting 9913, is derived from a human foreskin ®broblast P1 library in poor cell viability. Telomerase RNA gene promoters J-Q Zhao et al 1350 Acknowledgements equipment used in this study was purchased as a result of This work was supported by the Cancer Research grants from the University of Glasgow Medical Research Campaign (UK), and Glasgow University. Some of the Funds and donations from the White Lily Group.

References

Avilion AA, Piatyszek MA, Gupta J, Shay JW, Bacchetti S Laird PW and Jaenisch R. (1996). Annu. Rev. Genet., 30, 44± and Greider CW. (1996). Cancer Res., 56, 645 ± 650. 464 Baylin SB, Makos M, Wu J, Yen RW, De BA, Vertino P and Li H, Zhao LL, Funder JW and Liu JP. (1997). J. Biol. Nelkin BD. (1991). Cancer Cells, 3, 383 ± 390. Chem., 272, 16729 ± 16732. Bestilny LJ, Brown CB, Miura Y, Robertson LD and Mandal M and Kumar R. (1997). J. Biol. Chem., 272, Riabowol KT. (1996). Cancer Res., 56, 3796 ± 3802. 14183 ± 14187. Bird AP. (1996). Cancer Sur., 28, 87±101 Miller N and Whelan J. (1997). Human Gene Ther., 8, 803 ± Blasco MA, Funk W, Villeponteau B and Greider CW. 815. (1995). Science, 269, 1267 ± 1270. Morin GB. (1997). Eur. J. Cancer Part A, 33, 750 ± 760. Blasco MA, Lee H, Hande MP, Samper E, Lansdorp PM, Nakamura TM, Morin GB, Chapman KB, Weinrich SL, DePinho RA and Greider CW. (1997). Cell, 91, 25 ± 34. Andrews WH, Lingner J, Harley CB and Cech TR. (1997). Blasco MA, Rizen M, Greider CW and Hanaham D. (1996). Science, 277, 955 ± 959. Nature Genetics, 12, 200 ± 204. Norrback KF and Roos G. (1997). Eur. J. Cancer Part A, 33, Bodnar AG, Kim NW, E€ros RB and Chiu CP. (1996). 774 ± 780. Exper. Cell Res., 228, 58 ± 64. Pan C, Xue BH, Ellis TM, Peace DJ and Diaz MO. (1997). Broccoli D, Chong L, Oelmann S, Fernald AA, Marziliano Exper. Cell Res., 231, 346 ± 353. N, Van SB, Kipling D, Le BMM and De LT. (1997). Peterson MG and Baichwal VR. (1993). Trends in Biotech., Human Mol. Genet., 6, 69 ± 76. 11, 11 ± 18. Broccoli D, Godley LA, Donehower LA, Varmus HE and De Prowse KR and Greider CW. (1995). Proc. Nat. Acad. Sci. LT. (1996). Mol. Cell. Biol., 16, 3765 ± 3772. USA, 92, 4818 ± 4822. Bryan TM, Marusic L, Bacchetti S, Namba M and Reddel Riabowol K, Schi€ J and Gilman MZ. (1992). Proc. Nat. RR. (1997). Human Mol. Genet., 6, 921 ± 926. Acad. Sci. USA, 89, 157 ± 161. Cai W, Hu L and Foulkes JG. (1996). Curr. Opin. Seshadri T and Campisi J. (1990). Science, 247, 205 ± 209. Biotechnol., 7, 608 ± 615. Sharma S, Raymond E, Soda H, Sun D, Hilsenbeck SG, Campisi J. (1997). Euro. J. Cancer Part A, 33, 703 ± 709. Sharma A, Izbicka E, Windle B and Von Ho€ DD. (1997). Cheng AJ, Liao SK, Chow SE, Chen JK and Wang TC. Ann. Oncol., 8, 1063 ± 1074. (1997). Biochem. Biophys. Res. Comm., 237, 438 ± 444. Shay JW and Bacchetti S. (1997). Eur. J. Cancer Part A, 33, Connors TA. (1995). Gene Therapy, 2, 702 ± 709. 787 ± 791. Feng J, Funk WD, Wang SS, Weinrich SL, Avilion AA, Chiu Soder AI, Going JJ, Kaye SB and Keith WN. (1997a). CP, Adams RR, Chang E, Allsopp RC, Yu J, Le S, West Oncogene, in press. MD, Harley CB, Andrews WH, Greider CW and Soder AI, Hoare SF, Muir S, Going JJ, Parkinson EK and Villeponteau B. (1995). Science, 269, 1236 ± 1241. Keith WN. (1997b). Oncogene, 14, 1013 ± 1021. Galang CK, Der CJ and Hauser CA. (1994). Oncogene, 9, Soder AI, Hoare SF, Muire S, Balmain A, Parkinson EK and 2913 ± 2921. Keith WN. (1997c). Genomics, 41, 293 ± 294. Gardiner-Garden M and Frommer M. (1987). J. Molec. Tenen DG, Hromas R, Licht JD and Zhang DE. (1997). Biol., 196, 261 ± 282. Blood, 90, 489 ± 519. Holt SE, Wright WE and Shay JW. (1997). Euro. J. Cancer Vertino PM, Issa JP, Pereira-Smith OM and Baylin SB. Part A, 33, 761 ± 766. (1994). Cell Growth Di€er., 5, 1395 ± 1402. Irving J, Feng J, Wistrom C, Pikaart M and Villeponteau B. Wasylyk C, Maira SM, Sobieszczuk P and Wasylyk B. (1992). Exper. Cell Res., 202, 161 ± 166. (1994). Oncogene, 9, 3665 ± 3673. Kim NW. (1997). Eur. J. Cancer Part A, 33, 781 ± 786. Wellinger RJ and Sen D. (1997). Eur. J. Cancer Part A, 33, Kipling D. (1997a). Hum. Mol. Genet., 6, 1999 ± 2004. 735 ± 749. Kipling D. (1997b). Eur. J. Cancer Part A, 33, 792 ± 800. Wilson VL and Jones PA. (1983). Science, 220, 1055 ± 1057. Kuniyasu H, Domen T, Hamamoto T, Yokozaki H, Yasui Wright WE, Piatyszek MA, Rainey WE, Byrd W and Shay W, Tahara H and Tahara E. (1997). Jap. J. Cancer Res., JW. (1996). Dev. Genet., 18, 173 ± 179. 88, 103 ± 107. Zakian VA. (1997). Cell, 91, 1±3.