The Human Papillomavirus Type 16 E6 and E7 Oncoproteins Dissociate Cellular Telomerase Activity from the Maintenance of Telomere Length*

The “high risk” subgroup of human papillomaviruses (e.g. HPV-16 and HPV-18) infect and induce tumors of mucosal epithelium. These neoplasms, which can progress to malignancy, retain and express the papillomavirus E6 andE7 oncogenes. In vitro, the E6 and E7 proteins associate with the cellular p53 and Rb proteins and interfere with their normal growth-regulatory functions. We report here that primary human keratinocytes transduced with the HPV-16 E6 gene, but not the E7 gene, express significant telomerase activity. However, despite this detectable enzymatic activity,E6-transduced cells continue to shorten their telomeres during in vitro passaging similar to control cells and to cells expressing the E7 and E6+E7 genes. At late passages, however, E7-transduced cells partially restore telomere length, although they lack detectable telomerase activity, demonstrating that E6-independent, telomerase-independent events mediate this change.

The human papillomaviruses (HPVs) 1 associated with cervical cancer are designated as the "high risk" subgroup of HPVs (e.g. HPV-16 and 18) (1) and encode two viral oncogenes, E6 and E7, which exhibit immortalizing and transforming activities in various rodent and human cell types (2-14; reviewed in Refs. 15 and 16). The oncogenic potential of these proteins is due, at least in part, to their ability to interfere with the function of two cellular tumor suppressor proteins, p53 and the retinoblastoma susceptibility gene product, Rb. The E6 protein binds to the cellular p53 protein and promotes its ubiquitin-dependent degradation (17,18). E7 protein associates with Rb and interferes with its binding to E2F, resulting in impaired Rb cell cycle control functions (19 -21). The natural host cell of these viruses, the keratinocyte, is immortalized by the efficient expression of E7 protein (22). Although the E6 protein cannot independently immortalize these cells, it greatly augments the biological activity of the E7 protein (23), induces resistance to signals for terminal differentiation (24), and prolongs precrisis life span (25). Recently, it has been demonstrated that the expression of E6 in primary human keratinocytes leads to an activation of telomerase (26), an enzyme capable of preventing the shortening of telomeres during DNA replication (27)(28)(29)(30)(31).
Telomeres, the ends of human eucaryotic chromosomes, are shortened progressively during cell aging in vivo and in vitro (29,30,(32)(33)(34). The telomeric ends of chromosomes consist of a stretch of tandemly repeated sequences complexed with DNAbinding proteins that position chromosomes within the nucleus, protect chromosomal ends, and prevent chromosomal fusion that can occur in late-passage or senescent cells (27,28,31). Germ-cell telomeres are longer than somatic cell telomeres and are maintained with age (35)(36)(37), probably due to the activated riboprotein telomerase in these cells (38). In contrast to primary somatic cells, immortalized cells generally retain telomere length during in vitro propagation. The DNA polymerase telomerase, a ribonucleoprotein complex, is capable of elongating the 3Ј lagging DNA strand by adding tandemly repeated sequences to this DNA strand. The ribonucleoprotein complex uses parts of its RNA molecule as a template for the polymerization of telomeric tandem repeats, counteracting the shortening of telomeres during the replication of DNA by DNA polymerase ␣.
We observed that early-passage keratinocytes contained detectable telomerase activity that was lost at later passage numbers (between passages 2 and 6). Because both E6 and E7 genes can enhance and prolong human primary keratinocyte growth in vitro, we evaluated whether they might also induce alterations in chromosomal processing and maintain telomeres in a lengthened state. The analysis of primary human foreskin keratinocytes expressing the HPV-16 E6, E7, or E6ϩE7 (E6/7) genes revealed that E6-and E6/7-expressing cells were telomerase-positive, independent of their passage number. E7-expressing cells, like vector-infected control cells, demonstrated an activated telomerase at early-passage numbers that was rapidly lost during cell passaging. The detectable telomerase activity in precrisis E6-and E6/7-expressing cells, however, did not result in the maintenance of telomere length. All cells, independent of their expression of HPV oncogenes and telomerase, equally shortened their telomeres. However, E7-expressing cells demonstrated partial restoration of telomeric length without the concomitant activation of telomerase during their extended life span.

EXPERIMENTAL PROCEDURES
Cell Culture-Primary human keratinocytes were derived from neonatal foreskins as described (10) and grown in KSF medium (Life Technologies, Inc.) supplemented with gentamycin. The primary cells were infected with derivatives of the amphotrophic LXSN retrovirus expressing the various HPV-16 open reading frames (E6, E7, and E6ϩE7). The retroviruses were generated as described (39) using existing recombinant vectors (24). Retrovirus-infected cells were selected in G418 (100 g/ml medium) for 10 days. G418-resistant colonies were pooled from each transduction and passaged every 3-4 days (ratio of 1:5).
Telomeric Repeat Amplification Protocol-A modified telomeric amplification protocol (TRAP) was performed as described by Kim et al. (38). Cells were harvested by trypsinization followed by a wash in Dulbecco's modified Eagle's medium containing 10% fetal calf serum to inactivate trypsin. The cells were washed a second time in phosphatebuffered saline and frozen at Ϫ70°C. Cell pellets (Ϸ5 ϫ 10 6 ) were lysed in 800 l of lysis buffer (0.5% 3-((3-cholamidopropyl)-dimethyl-ammonio)-1-propanesulfonate, 10 mM Tris-Cl, pH 7.5, 1 mM MgCl 2 , 1 mM EGTA, 5 mM ␤-mercaptoethanol, and 10% glycerol) for 30 min on ice. The lysates were then centrifuged for 2 min at 14,000 ϫ g (4°C), and 600 l of the supernatant were removed. The protein concentration of the supernatant was determined using a DC Protein Assay kit (Bio-Rad). RNase inhibitor (RNasin from Promega Corp.) was added afterward to a final concentration of 10 units/ml. The remaining nuclear pellet was digested with proteinase K and treated with phenol; chromosomal high molecular weight DNA was isolated by standard techniques for Southern blot analysis. The TRAP assay was performed in a 0.2-ml MicroAmp reaction tube (Perkin Elmer) containing 98 l of a reaction mixture composed of 1 ϫ Taq polymerase buffer (10 mM Tris-HCl, pH 8.8, 1 mM KCl, and 0.02% Tween 20), 1.5 mM MgCl 2 , 100 ng of TS primer (5Ј-AATCCGTCGAGCAGAGTT-3Ј), 200 M of each deoxynucleoside triphosphate, and 0.05-6 g of protein extract. The reaction mixture was incubated for 20 min at 25°C. The samples were then heated to 80°C, and 1 l (100 ng/l) of CX primer (3Ј-AATCCCAT-TCCCATTCCCATTCCC-5Ј) and 0.5 l of Taq polymerase (UlTima DNA polymerase, 6 units/l; Perkin Elmer) were added before the samples were heated to 94°C for 90 s. The polymerase chain reaction was performed for 28 cycles under the following conditions: 94°C for 30 s, 50°C for 30 s, and 72°C for 45 s. Ten % of the polymerase chain reaction product was separated on 8% SDS-polyacrylamide gel (1 ϫ TBE), and the gel was stained with a Gelcode color silver stain kit (Pierce). To demonstrate telomerase specificity, control samples were either digested for 20 min at 37°C with 1 g of RNase A and 1 g of RNase H or the samples were heated for 10 min to 100°C prior to the telomerase reaction.
Southern Blot-Three or 10 g of chromosomal high molecular weight DNA were restricted with the indicated restriction endonucleases for telomere Southern blot analysis or HPV copy number analysis, respectively. The restricted DNA samples were separated on 1 ϫ TAE agarose gels (0.5-1.0%) and blotted onto a nylon membrane. A 32 P kinase-labeled primer, (TTAGGG) 3 , was used under standard high stringency hybridization conditions (T m Ϫ15°C) to detect the telomeric ends. A random-primed 32 P-labeled HPV-16 E6/7 DNA fragment was used to detect the various HPV open reading frames using high stringency conditions.

RESULTS
The E6 and E7 Genes Independently Extend the Life Span of Neonatal Foreskin Keratinocytes-Primary human foreskin keratinocytes were infected with amphotropic retroviruses expressing either the HPV-16 E6, E7, E6/7 genes or the neomycin resistance gene (control). Cells were selected in serum-free keratinocyte medium containing G418 as described under "Experimental Procedures" and passaged at a ratio of 1:5. Keratinocytes infected with control retrovirus ceased proliferation at passages 10 -12, whereas those transduced with E6/7 established into cell lines. Keratinocytes expressing either E6 or E7 alone exhibited extended life span and continued to proliferate beyond passage 26.
E6, but not E7, Generates Keratinocyte Cultures with Telomerase Activity-We observed, in three independent experiments, that E6 retroviruses generated human keratinocyte FIG. 1. Telomerase activity and telomere length in keratinocytes before E6-or E7-induced "extended life span." In A, E6 induces telomerase in early-passage keratinocytes. Primary human embryonic foreskin keratinocytes were isolated and infected at second passage with LXSN-based retroviruses encoding the HPV-16 E6-, E7-, or E6/7 genes. As a control, cells were infected with the LXSN vector virus (vector). Following infection, cells were selected in G418, and samples were harvested at sequential early passages (up to passage 10) for a TRAP assay as described by Kim et al. (38). Ten % of the TRAP reaction product was separated on an 8% nondenaturing polyacrylamide gel, which was then stained with silver. A typical DNA ladder (6-base pair difference per band) was generated by cellular telomerase in the cell extract of keratinocytes transduced with the E6 or E6/7 genes. Vector-transduced and E7-transduced cells showed a weak telomerase activity at the earliest passages, which disappeared by passage 6. Differences between E6-and E7-transduced cells were most prominent at passages 8 and 10. In B, the absence of TRAP assay activity in E7 cell extracts is not due to the presence of inhibitors of the telomerase assay. Mixing experiments with extracts derived from telomerase-negative, E7-expressing cells and telomerase-positive, E6and E6/7-expressing cells (as well as an HPV-16 immortalized cell line) were carried out to determine if the telomerase-negative E7 extract contained any inhibitors of the telomerase reaction or the polymerase chain reaction. 0.5 g of the indicated cell extract was either assayed directly in the TRAP assay (ϪE7 extract) or after mixing with 0.5 g of telomerase-negative E7 cell extract (ϩE7 extract). The telomerase as-say was performed as described in Fig. A. There was no detectable inhibition of telomerase activity. In C, telomere length gradually decreases in early-passage keratinocytes expressing E6, E7, or E6/7. Southern blot analysis of telomere length was determined using 3 g of chromosomal DNA (cleaved with the HinfI and RsaI restriction enzymes) from cells at the indicated passage number. The DNA was separated on 0.5% agarose gels (1 ϫ TAE) and blotted onto a nylon filter, which was then hybridized with a 32 P 5Ј-labeled primer (TTAGGG) 3 . A gradual loss of telomeric length (1.5-2.0 kb) was observed over the course of 10 passages, and there was no difference observed in the telomere length of the cells expressing vector DNA or the E6, E7, or E6/7 genes at any given passage. strains with significant telomerase activity at early passage (e.g. passages 4 and 6, Fig. 1A), similar to the recently published results of Klingelhutz et al. (26). The concomitant ex-pression of the E6/7 genes gave similar results. Although low levels of telomerase were observed in control keratinocytes and in E7-expressing cells at early passages (passages 4 and 6), this activity was absent at later passages (passages 8 and 10), presumably reflecting the loss of basal epithelial "stem" cells in the neonatal human foreskin keratinocyte population. The detection of telomerase activity in normal keratinocytes was dependent upon the isolation procedures and the number of cell divisions needed to establish the primary culture and was normally detected up to passages 2-3 but never after passage 6.
Mixing experiments (Fig. 1B) using a telomerase-negative E7 cell extract ( Fig. 1A; E7 passage 8) and telomerase-positive cell extracts ( Fig. 1A; E6, E6/7 passage 8, and a cell extract from an HPV-16 immortalized keratinocyte line) demonstrated that the loss of TRAP assay activity in E7 extracts was not due to the presence of an inhibitor of telomerase or polymerase activity. Thus, the gradual passage-dependent loss of telomerase activity in E7-expressing cells represents a true loss of enzymatic activity.
Telomere Length Decreases during Early Passages of All Transduced Keratinocytes-To determine whether the observed activation of telomerase in E6-transduced keratinocytes (Fig. 1A) was accompanied by altered processing of the chromosome telomeres, we harvested nuclei from keratinocytes at the indicated passages and performed Southern blotting on isolated cellular DNA with a telomeric probe to evaluate average telomere length (Fig. 1C). Cells were evaluated until passage 10 when the HPV-negative control cells reached crisis. Southern blot analysis demonstrated that regardless of the HPV genes transfected, all cells exhibited shortened telomeres similar to control-transduced cells. During the first 10 passages, the average telomere length decreased from approximately 10 to 8.5 kb, which represents an average loss of 60 -80 base pairs per population doubling. This is in agreement with previous studies in primary human fibroblasts (32,34) and human embryonic kidney cells. The correlation between the loss of keratinocyte growth potential (at 30 -40 population doublings) and the shortening of telomeres by 1.5-2.0 kb also resembles the differences in telomere length seen in vivo between fibroblasts from human embryos and late-age individuals (30,33).
Telomerase Activity Is Maintained in E6-transduced Keratinocytes during Extended Life Span-We also compared the telomerase activity of retrovirus-infected keratinocytes at passage 26, far past the time when the HPV-negative control cells had undergone crisis and lost their proliferative potential (passage 10). The TRAP assay depicted in Fig. 2A demonstrates that E6-transduced cells continued to display telomerase activity throughout passaging, whereas the vector-and E7-transduced cells were negative.
During Extended Life Span, Telomere Length Is Partially Restored in E7-transduced, but not E6-transduced, Keratinocytes-The expression of the E6 or E7 protein leads to an extended keratinocyte life-span in vitro in comparison to noninfected or vector-infected control cells (25). In one of our experiments, E6-transduced cells proliferated beyond passage 26. Despite the continued presence of telomerase activity in these late-passage cells ( Fig. 2A), we observed a further shortening of the mean telomere length to Ϸ5 kb (decreased from Ϸ8.5 kb at passage 10) (Fig. 2B). E6-transduced cells lost an estimated 60 -80 base pairs per population doubling. The corresponding E7-expressing cells also demonstrated a similar loss of telomeric length between passages 10 and 21 (Fig. 2B). However, from approximately passage numbers 21 to 26, the E7-transduced keratinocytes showed an increase in telomere length (Fig. 2B), although there was no detectable telomerase activity

FIG. 2. Telomerase activity and telomere length in keratinocytes after E6-or E7-induced "extended life span."
In A, telomerase activity is maintained in E6-transduced extended life span keratinocytes. Keratinocytes expressing E6 or E7 were analyzed at passages 10 and 26 for telomerase activity as described above. At passage 26, keratinocytes expressing E6 remained positive for telomerase, and E7transduced cells remained negative. Vector-transduced keratinocytes did not proliferate beyond passages 10 -12 and could not be evaluated for enzymatic activity. In B, late passage, E7-transduced keratinocytes lengthen telomeric DNA. Keratinocytes expressing E6 or E7 were analyzed for telomere length at passages 21 and 26 as described above. E6-transduced cells continued to lose telomeric DNA until the average telomere length was approximately 5.0 kb. At passage 26, the E7transduced cells gained telomeric DNA, which was also more heterogeneous in size. The vector-transduced keratinocytes failed to proliferate beyond passage 10, despite containing telomeres (Ϸ8.5-kb size; Fig. 1C) that were longer than E6-or E7-transduced cells at passage 21 (Ϸ5.0-kb size). in these cells ( Fig. 2A). In contrast, E6-transduced cells at passage 26 continued to shorten their telomeres. The increase in telomere length of E7-expressing cells might indicate, as observed in other tissue culture systems (40), that the cells have reached immortalization crisis.
The Transduced Keratinocyte Cultures Contain, on Average, at Least One Copy of the HPV Gene Per Cell and Express Telomerase at a Level Equivalent to Clonally Derived, HPVimmortalized Cell Lines-Although all keratinocyte strains were isolated following G418 selection and should consist only of cells transduced with HPV genes, we performed Southern blot analysis of chromosomal DNA from the indicated keratinocyte strains to verify the presence of viral genes. Keratinocytes transduced with E6, E7, or E6/7 were assayed at passages 4 and 10 to determine the average copy number of HPV genes in the cell population. Fig. 3A demonstrates that all keratinocyte strains contained at least one copy of the transduced HPV gene per cell at both early and late passages. This result suggests that the observed shortening of telomeres in E6-transduced cells is not the consequence of inefficient gene transduction or HPV gene instability.
Evidence that most or all of the E6-transduced keratinocytes are expressing telomerase is derived from the comparative analysis of telomerase activity in transduced cells and clonally FIG. 3. HPV gene copy number and telomerase activity of transduced keratinocytes. In A, transduced keratinocytes contain an average of at least one copy of the E6, E7, or E6/7 genes per cell. The average number of the HPV-16 E6, E7, or E6/7 genes, which were integrated into transduced keratinocytes, was estimated by Southern blot analysis. The copy number standards (left three lanes) were obtained via dilution of HPV-16 E6/7 DNA fragments to the indicated copy number equivalents, and the calculations were based upon the estimation that 10 g of chromosomal DNA correspond to 10 6 cells. 10 g of chromosomal DNA were restricted with enzymes that released the E6 (XhoI/HindIII), E7 (EcoRI/XhoI), or E6/7 (BamHI) genes from vector DNA. The DNA was then separated on 1% agarose and blotted onto nylon filters. Hybridization with a 32 P-labeled E6/7 probe and subsequent washing were performed under high stringency conditions. The E6-, E7-, and E6/7-transduced cultures contained more than one copy of the indicated HPV oncogene per cell. B, comparison of the telomerase activity of nonimmortalized HPV-16 E6-expressing cells with the telomerase activity of immortalized keratinocyte cell lines. One g of cytoplasmic protein from each of the indicated keratinocyte cultures (passage 8) was used for a TRAP assay as described under "Experimental Procedures." The keratinocyte cell lines established by HPV-16, HPV-18, and SV40 and the two keratinocyte strains generated with E6-containing retroviruses (E6 and E6/7) exhibited similar expression of telomerase activity. The negative control is a culture of E7-transduced keratinocytes at passage 8.

FIG. 4. Specificity and sensitivity of the TRAP assay. In A,
TRAP assay activity is abolished by RNase treatment and heat denaturation. To verify the dependence of the observed DNA ladder upon the enzymatic activity of the riboprotein telomerase complex, we treated the cytoplasmic extracts of keratinocytes with RNase A (1 g) and H (1 g) for 20 min at 37°C or heated the extracts at 100°C for 10 min prior to the TRAP assay. The activity of cell extracts was completely inactivated by either RNase or heat treatment. The assays used 1 g of protein extract from passage 6 keratinocytes. B, titration of telomerase activity in E6-and E6/7-transduced keratinocytes. The relative telomerase activity of the indicated keratinocyte cultures was estimated by serial dilution analysis. TRAP assays, using 10-fold dilutions of cytoplasmic protein, indicated that E6-and E6/7-transduced keratinocytes exhibited similar telomerase activity that was detectable with as little as 60 ng of protein (equivalent to approximately 100 keratinocytes). derived cell lines immortalized by either HPV-16 or HPV-18 DNA or by SV40 virus (Fig. 3B). The telomerase activities of the tested HPV-immortalized cell lines were equivalent to the telomerase activity of HeLa cells (data not shown). The extracts of nonimmortalized, transduced cell strains exhibited telomerase activity that was similar to that observed in several independent clonal cell lines, suggesting that the transduced keratinocyte population (like the clonal cell lines) consists mostly of cells expressing telomerase.
Specificity and Sensitivity of the TRAP Assay-To demonstrate that the polymerase chain reaction-based TRAP assay was indeed measuring telomerase activity rather than the contamination of cytoplasmic extracts with chromosomal telomeric DNA, we treated the cell extracts with RNase A and H for 20 min at 37°C or with 100°C exposure for 10 min. In both cases, telomerase activity was completely abolished, indicating that the riboprotein telomerase was responsible for the observed results (Fig. 4A).
Serial dilution of cell extracts demonstrated that E6-transduced cells had telomerase activity similar to E6/7-expressing cells and that the sensitivity of the TRAP assay was sufficient to detect activity in approximately 100 cells (Fig. 4B). This is consistent with the sensitivity of the assay using established cell lines (e.g. HeLa cells), further supporting our previous findings that the E6-transduced keratinocyte population was composed predominantly of cells expressing telomerase (Fig. 3B). DISCUSSION The ability of the E6 protein to extend the precrisis life span of human keratinocytes and to activate telomerase is in agreement with the recent independent finding of Klingelhutz et al. (26). However, by simultaneously measuring telomere length and telomerase activity during sequential passages, we have demonstrated that the E6-dependent activation of telomerase has no detectable effect on the normal shortening of telomeres during in vitro cell passaging. In addition, we have also demonstrated that E7-transduced cells lengthen telomeres after prolonged passaging (potentially at cellular immortalization) and that this telomere elongation occurred independent of detectable telomerase activity. The finding that E7-transduced cells elongate telomeres without detectable telomerase activity is in agreement with a study of Bryan et al. (41) describing immortal human cell lines without detectable telomerase activity. Although the elongation of telomeres during and after cell crisis normally coincides with a high degree of chromosomal aberrations, keratinocytes immortalized by E6/7 typically display few or limited chromosomal abnormalities (42,43).
The current study demonstrates that the E6-dependent increase in telomerase activity is rapid and occurs by the time the retrovirus-infected keratinocytes have been selected in G418. However, it is unclear whether this increase in enzymatic activity represents the induction of telomerase, the activation of telomerase, or possibly the selection of telomerase-positive stem cells. Thus, it remains a possibility that, following the infection of primary keratinocytes by the E6 retrovirus, the E6 oncogene selectively augments the growth of telomerase-positive cells. This selection would result in the amplification of a keratinocyte population containing high telomerase activity. It is also possible that all of the primary keratinocytes express telomerase activity and that E6 functions to maintain this phenotype. However, it is believed that telomerase activity is, in general, not present in somatic cells but is restricted to germ-line cells and tumor cells (30,38).
The telomerase activity detected in E6-transduced keratinocytes is presumed to represent the activity of all cells in the population rather than a minor subpopulation for two reasons: (a) Southern blot analysis indicates that, on the average, each keratinocyte contains at least one copy of the transduced HPV gene(s). Thus, it is likely that most of the transduced cells are expressing E6 and are, therefore, altered in telomerase activity. Unfortunately, however, there are no in situ techniques to evaluate whether E6 protein (or telomerase) is expressed in each cell; and (b) the level of telomerase detected in the G418selected, E6-transduced keratinocyte population is equivalent to the telomerase level observed in a clonal keratinocyte cell line that was immortalized by E6/7 (Fig. 3).
Although only the E6 protein increases cellular telomerase activity, both E6 and E7 are able to extend the keratinocyte life span in vitro, suggesting that prolongation of keratinocyte proliferation is not necessarily dependent upon telomerase activity. In addition, our findings indicate that telomere length is not the critical determinant for the cessation of keratinocyte proliferation in vitro because E6-and E7-transduced cells have shorter telomeres than control (vector-transduced) cells. Finally, our results demonstrate that increased or sustained telomerase activity is insufficient to induce cellular immortalization because E6-transduced genital keratinocytes do not establish into cell lines.
The experimental finding that E7-transduced cells elongate telomeres after prolonged passaging indicates that there must be E6-independent, telomerase-independent mechanisms for the elongation of telomeres in E7-transduced cells. Our results, therefore, support the findings of Bryan et al. (41), who postulate that a mechanism exists for the elongation of telomeres that is not accompanied by detectable levels of telomerase. It is interesting that the observed lengthening of telomeres in E7expressing cells appears to coincide with the onset of cell crisis, suggesting a potential role in cell immortalization.